Datasets:

TomTBT

/

pmc_open_access_xml

like 5

[ "<title>Introduction</title>", "<p>Coccidioides species, which are dimorphic fungi, are endemic to the Southwest United States and focal regions in Central and South America. With increased domestic and international travel, physicians must take a thorough travel history to consider coccidioides infection, given its non-specific presenting symptoms. Fortunately, infection with coccidioides does not always lead to clinical disease and may even result in lifelong cellular immunity. Typical clinical manifestations of this fungus include malaise, fever, cough and other non-specific symptoms that are indistinguishable from an influenza infection. We present a case of primary pleural coccidioidomycosis to add to the literature [##REF##10319068##1##, ####REF##414573##2##, ##REF##16482953##3####16482953##3##] and discuss the diagnostic tools that can assist in confirming the presence of a sole pleural effusion as a rare manifestation of this disease.</p>" ]

[]

[]

[ "<title>Discussion</title>", "<p>Even in endemic areas, primary pleural coccidioidomycosis is rare. When it has been reported, it is mainly right sided and can be present in all sizes. A pleural fluid lymphocyte-predominant exudative effusion is commonly associated with tuberculosis pleurisy, other fungal infections and lymphoma. In general, exudative pleural effusions present a diagnostic challenge because a wide differential of organisms may potentially cause the effusion, because numerous organs may serve as the foci of infection (Table ##TAB##0##1##), and because of limitations in commercially available confirmatory studies. This is particularly true when a rare presentation of an infectious organism is observed, as in our case.</p>", "<p>The incidence of coccidioidomycosis is increasing, with the majority of reports from the states of Arizona and California. There is particular risk associated with outdoor activity owing to seasonal precipitants and aerosolization of fine sand and silt, particularly in Filipinos and African-Americans.</p>", "<p>A number of methods assist in the diagnosis of symptomatic coccidioidomycosis infections, but recognition of its existence is of primary importance. A positive skin reaction can occur as early as 6 to 48 hours after injection, but a positive test is not entirely diagnostic of an active infection; it merely raises suspicion of cellular immunity [##REF##14682448##4##]. The limitations of this test include false-positive test results in individuals vaccinated against or previously exposed to coccidioides. In addition, cross-reactivity with histoplasma capsulatum can occur. The expression of this delayed-type hypersensitivity is lower with disseminated disease.</p>", "<p>It is the active and passive immunity that are used in detection and monitoring strategies. Serum immunoglobulin levels are used to detect the presence of an acute infection or immunity, depending on the immunoglobulin type: IgM or IgG. Coccidioides IgM levels may be persistently elevated for up to 6 months in acute infections. Complement fixation is another useful tool for monitoring both the extent of coccidioidomycosis and the response to treatment. Low serum quantified titers of between two and four have been encountered in early-phase coccidioidal infection, limited dissemination and late-stage disease as the titer is declining [##REF##2200605##5##]. A serum complement fixation titer level greater than 32 is generally a sign of disseminated disease. If dissemination is considered, then a lumbar puncture needs to be performed because of the risk of coccidioidomycosis meningitis. Cerebrospinal infection occurs in approximately 35% of patients with disseminated disease, even in the absence of meningeal signs. A cerebrospinal fluid complement fixation titer of 1:2 or greater usually indicates the presence of meningitis [##UREF##0##6##].</p>", "<p><italic>Coccidioides immitis </italic>can be cultured from tissue and body fluids. Saubolle et al. showed that the respiratory tract has the highest yield of recovery [##REF##17108067##7##]. Cultures may take 3 to 4 weeks to grow, delaying diagnosis. The distinguishing feature is the presence of a thick-walled spherule with endospores. A more rapid approach to identification involves real-time polymerase chain reaction (PCR) [##REF##17108077##8##]. Cross-reactivity comparisons with bacteria, other fungi, mycobacteria and viruses demonstrate 100% specificity to coccidioides. Biopsies and surgical specimen cultures are more likely to result in a positive culture than microscopic examination. Between 25% and 50% of sputum samples, bronchial washings, spinal fluid and urine specimens yield positive cultures [##REF##17108067##7##]. Blood cultures are unlikely to yield the presence of coccidioides, but when positive, they are associated with acute infection, dissemination and a high mortality.</p>", "<p>The Infectious Disease Society of America's recommendation for initial therapy of non-meningeal extrapulmonary infection is with an oral azole agent [##UREF##1##9##]. Clinical trials have shown that fluconazole daily dosage eradicates the disease in a majority of patients. In cases of clinical deterioration, amphotericin B 0.5 to 1.5 mg/kg per day should be administered. The newer extended spectrum azoles voriconazole and posaconazole appear to be effective in small clinical trials but are not yet suitable to be considered as first-line therapy. In small clinical trials, the use of posaconazole in the treatment of refractory coccidioidomycosis shows promising results with minimal side effects [##REF##15909265##10##]. Measuring the response to treatment can be a slow and challenging process. To establish adequate therapy, patients should be routinely followed up every 3 to 6 months for up to 2 years.</p>" ]

[ "<title>Conclusion</title>", "<p>With increased domestic and international travel, coccidioidomycosis will likely be encountered in nonendemic regions. A parapneumonic effusion from pulmonary coccidioidomycosis is seen in up to 50% of cases; primary pleural coccidioidomycosis, however, is a rare clinical feature of an endemic infectious disease. This lymphocytic-predominant effusion mimics other diseases, therefore recognition by physicians is critical for a timely diagnosis and therapy. Tissue culture can assist in the diagnostic approach and PCR analysis shows potential as a possible addition.</p>" ]

[ "<p>This is an Open Access article distributed under the terms of the Creative Commons Attribution License (<ext-link ext-link-type=\"uri\" xlink:href=\"http://creativecommons.org/licenses/by/2.0\"/>), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.</p>", "<title>Introduction</title>", "<p>Community-acquired pneumonia is the most common manifestation in primary coccidioides infections (<italic>Coccidioides immitis, C. posadasii</italic>). It is essential that this endemic dimorphic fungus be considered in order to proceed with the most appropriate diagnostic tools and therapy.</p>", "<title>Case presentation</title>", "<p>We present a rare case of primary pleural coccidioides and a review of the current literature for optimal diagnostic methods and therapeutic strategies.</p>", "<title>Conclusion</title>", "<p>With increased domestic and international travel, coccidioidomycosis will likely be encountered in nonendemic regions. Recognition by physicians is critical for a timely diagnosis and therapy. Tissue culture can assist in the diagnosis and polymerase chain reaction analysis shows potential as a possible addition.</p>" ]

[ "<title>Case presentation</title>", "<p>A 39-year-old man was admitted in October 2006 with a 2-week history of sharp, non-radiating pain of the right shoulder blade with associated dyspnea upon exertion and 5 kg loss of weight. He denied fever, chills, night sweats or cough. His symptoms did not interfere with his occupation as a gardener. Vitals demonstrated a normotensive, afebrile 155 cm, 100 kg man with an oxygen saturation of 96% on room air.</p>", "<p>Physical examination was normal with the exception of decreased breath sounds half way up the right lung field along with dullness to percussion and without tactile fremitus. A chest radiograph showed a moderately sized, right pleural effusion (Figure ##FIG##0##1##). The right thoracentesis fluid analysis showed a slightly cloudy and yellow fluid. Cell count results were 1,164 nucleated cells, 12% polymorphonuclear leukocytes, 80% lymphocytes, 7% monocytes, 1% eosinophils, glucose 123 mg/dl, lactate dehydrogenase (LDH) 103 units/l, and protein 5.3 g/dl (pleural to serum protein ratio, 0.8; pleural to serum LDH ratio, 0.7). Bacterial Gram stain and culture, acid-fast bacilli smears, fungal culture and cytology were all negative. Histopathological evaluation of the pleural biopsy noted granulomatous inflammation and fungal elements consistent with coccidioides (Figure ##FIG##1##2##). Cultures for tuberculosis remained negative even after 7 weeks. Purified protein-derivative skin test to the right forearm produced a 0 mm induration. Tests for human immunodeficiency virus were negative by both enzyme-linked immunosorbent assay and Western blot. Serum coccidioidomycosis complement fixation was normal (&lt; 1:2). A post-thoracentesis chest radiograph did not reveal any evidence of parenchymal infiltrate. After several weeks of fluconazole therapy, the patient improved clinically, and follow-up chest radiograph showed near-complete resolution of the pleural effusion.</p>", "<title>Abbreviations</title>", "<p>LDH: lactate dehydrogenase; PCR: polymerase chain reaction.</p>", "<title>Competing interests</title>", "<p>The authors declare that they have no competing interests.</p>", "<title>Consent</title>", "<p>Written informed consent was obtained from the patient for publication of this case report and accompanying images. A copy of the written consent is available for review by the Editor-in-Chief of this journal.</p>" ]

[]

[ "<fig position=\"float\" id=\"F1\"><label>Figure 1</label><caption><p>Chest radiograph showing moderate right pleural effusion.</p></caption></fig>", "<fig position=\"float\" id=\"F2\"><label>Figure 2</label><caption><p>Closed pleural biopsy showing coccidioidomycosis with evidence of endospores.</p></caption></fig>" ]

[ "<table-wrap position=\"float\" id=\"T1\"><label>Table 1</label><caption><p>Differential diagnosis for lymphocytic pleural effusion</p></caption><table frame=\"hsides\" rules=\"groups\"><tbody><tr><td align=\"left\">Tuberculosis</td></tr><tr><td align=\"left\">Non-tuberculosis Mycobacterium</td></tr><tr><td align=\"left\">Fungal pleurisy</td></tr><tr><td align=\"left\">Viral pleurisy</td></tr><tr><td align=\"left\">Malignancy</td></tr><tr><td align=\"left\"> Lymphoma</td></tr><tr><td align=\"left\"> Solid tumors</td></tr><tr><td align=\"left\">Sarcoidosis</td></tr><tr><td align=\"left\">Chylothorax</td></tr><tr><td align=\"left\">Post-Coronary Bypass Graft</td></tr><tr><td align=\"left\">Yellow-Nail Syndrome</td></tr></tbody></table></table-wrap>" ]

[]

[]

[]

[]

[]

[]

[]

[ "<graphic xlink:href=\"1752-1947-2-291-1\"/>", "<graphic xlink:href=\"1752-1947-2-291-2\"/>" ]

[]

{ "acronym": [], "definition": [] }

[ "<title>Introduction</title>", "<p>Thyroid carcinoma sometimes shows a microscopic vascular invasion, but gross angioinvasion with intraluminal thrombosis is extremely rare. Very few cases about metastasis of thyroid cancer to the internal jugular vein, and fewer cases about metastasis to the parotid gland have been separately reported. Our patient has both these organs involved by direct spread from a thyroid follicular carcinoma.</p>" ]

[]

[]

[ "<title>Discussion</title>", "<p>Invasion of the parotid gland and the great cervical veins from a thyroid cancer is extremely rare, and is mostly detected at autopsy [##REF##9492155##1##, ####REF##15229908##2##, ##REF##11213825##3####11213825##3##]. Both of these organs were involved in our patient following direct spread from a thyroid follicular carcinoma. Two general types of metastases should be distinguished in metastatic salivary gland tumours: regional metastases (head and neck) and distant metastases [##REF##3562340##4##].</p>", "<p>Involvement of the parotid gland by invasion or spread by metastases from malignant tumours in the head and neck is uncommon, with the exception of melanoma of the temple, scalp and ear, and anaplastic squamous cell carcinoma of the ear and ear canal [##REF##14058839##5##]. Seifort et al. [##REF##3562340##4##] reported three cases of a metastatic thyroid cancer to the parotid in their analysis of 108 cases of secondary metastases to salivary glands. Another case was found by the Pack Medical Group among 81 cases of parotid gland involvement as a secondary extension of malignant tumours [##REF##14058839##5##]. It is more common for the parotid gland to be involved as an incidental part of a generalized metastatic disease rather than a site of isolated metastasis. This gland contains 20 to 30 lymph follicles and lymph nodes connected with a rich interlacing network of lymph vessels. Lymph entrance to the gland may be direct, without involvement of the paraglandular lymph nodes, may be secondarily deposited from paraglandular lymph nodes, or may contaminate the system by retrograde extension from massive metastases in the neck [##REF##12561024##6##]. Clinically and pathologically, secondary spread to the parotid manifests itself as a primary salivary gland tumour that may mislead clinicians, radiologists and pathologists [##REF##12561024##6##].</p>", "<p>The cytological recognition of a thyroid metastasis to different body sites may pose a diagnostic difficulty, especially when a thyroid cancer presents initially at the metastatic site. Immunohistochemical thyroglobulin positivity is a useful tool in distinguishing between a thyroid primary and other metastatic lesions, as this marker is specific for thyroid tumours [##REF##12561024##6##]. Once the parotid has become a focus of metastasis in malignant tumours of the head and neck, the prognosis is grave [##REF##14058839##5##].</p>", "<p>Thyroid carcinoma sometimes shows a microscopic vascular invasion, but rarely causes tumour thrombus in the internal jugular vein or the great veins of the neck [##REF##15453533##7##]. The tumour thrombus is the result of a tumour extension from the thyroid gland to the IJV or the result of occult vascular spreading. The most common clinical manifestation is a dilated vein. Findings on neck palpation are usually non-specific and may reveal oedema and tenderness of the sternocleidomastoid muscle and the surrounding soft tissues [##REF##15453533##7##].</p>", "<p>The primary management of an advanced disease with vascular invasion would be radical surgery to remove a macroscopic disease. This is followed by high-dose radioiodine ablative therapy with or without external beam radiotherapy and suppression of thyroid stimulating hormone [##REF##11213825##3##]. The role of chemotherapy in these cases remains unproven.</p>" ]

[ "<title>Conclusion</title>", "<p>This rare case of a thyroid follicular carcinoma presenting as a metastasis in the parotid gland serves to highlight the importance of remaining clinically vigilant to the possibility that a salivary gland lesion may be a metastasis from another site. The necessity of communication between clinicians, histopathologists and radiologists is also well illustrated by this case. This very rare presentation of a thyroid follicular carcinoma could easily have been reported incorrectly as benign thyroid follicular cells if there was poor communication and the reporting pathologist was not made aware that the initial aspirate was from the parotid gland and not from the thyroid gland.</p>" ]

[ "<p>This is an Open Access article distributed under the terms of the Creative Commons Attribution License (<ext-link ext-link-type=\"uri\" xlink:href=\"http://creativecommons.org/licenses/by/2.0\"/>), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.</p>", "<title>Introduction</title>", "<p>The parotid gland and the great cervical veins are very rarely involved in a metastatic thyroid cancer.</p>", "<title>Case presentation</title>", "<p>We report an interesting case of an unusual metastasis of a thyroid follicular carcinoma including the histopathological and radiological findings. A woman was seen in the otolaryngology clinic with a mass at the angle of the left side of her jaw. Clinical examination and investigations confirmed a thyroid follicular carcinoma with metastases to the parotid gland and the internal jugular vein.</p>", "<title>Conclusion</title>", "<p>This is an educational case which highlights the importance of close communication between clinicians, histopathologists and radiologists to ensure that such rare cases are not missed.</p>" ]

[ "<title>Case presentation</title>", "<p>A 78-year-old woman was seen in the otolaryngology clinic in June 2006 with a painless swelling at the angle of the left side of her jaw which had been present for 9 months. The mass had slightly increased in size over this period. The patient had tinnitus but no other complaints. Her weight was stable. Clinical examination revealed a smooth, soft lesion in the tail of the left parotid gland. There was no cervical lymphadenopathy. The ears, nose and throat were normal and the facial nerve was intact.</p>", "<p>Ultrasound of the neck showed swellings in the left parotid gland and the left thyroid lobe. Fine needle aspiration (FNA) of the left parotid gland showed thyroid follicular cells. A magnetic resonance imaging (MRI) scan of the neck confirmed both soft tissue masses with extensive thrombosis of the left internal jugular vein contiguous with the primary tumour (Figure ##FIG##0##1A## and ##FIG##0##1B##). A computed tomography (CT) scan of the chest was normal. Subsequent FNA of the left thyroid lobe and the internal jugular vein (IJV) revealed thyroid follicular cells similar to those seen in the first FNA. The cells were positive for thyroglobulin and thyroid transcription factor 1 and negative for chromogranin and synaptophysin on immunohistochemistry, confirming the diagnosis of a thyroid follicular carcinoma (Figure ##FIG##1##2A, B## and ##FIG##1##2C##). Although the patient was not fit for aggressive surgery, she was given two courses of radioiodine. An uptake scan performed approximately 14 months after diagnosis (6 weeks after her last course of radioiodine) showed no further significant iodine uptake. At that time she was clinically well with no palpable residual or recurrent disease. She is still on routine follow-up in the oncology clinic.</p>", "<title>Abbreviations</title>", "<p>CT: computed tomography; FNA: fine needle aspiration; IJV: internal jugular vein; MRI: magnetic resonance imaging; TTF-1: Thyroid Transcription Factor 1.</p>", "<title>Competing interests</title>", "<p>The authors declare that they have no competing interests.</p>", "<title>Consent</title>", "<p>Written informed consent was obtained from the patient for publication of this case report and accompanying images. A copy of the written consent is available for review by the Editor-in-Chief of this journal.</p>", "<title>Authors' contributions</title>", "<p>AA performed the literature search, and drafted and revised the manuscript. JS evaluated the histological slides. GW evaluated the radiological images. IA assisted with the literature search. MQ edited the manuscript. All authors have read and approved the final manuscript.</p>" ]

[]

[ "<fig position=\"float\" id=\"F1\"><label>Figure 1</label><caption><p>Coronal T2 weighted image <bold>(A) </bold>and STIR sequence <bold>(B) </bold>showing left thyroid tumour extending directly into the left internal jugular vein.</p></caption></fig>", "<fig position=\"float\" id=\"F2\"><label>Figure 2</label><caption><p>Parotid aspirate (A &amp; B) showing thyroid follicular cells. Nucleus positive immunohistochemistry for Thyroid Transcription Factor-1 confirms thyroid origin (C).</p></caption></fig>" ]

[]

[]

[]

[]

[]

[]

[]

[]

[ "<graphic xlink:href=\"1752-1947-2-297-1\"/>", "<graphic xlink:href=\"1752-1947-2-297-2\"/>" ]

[]

{ "acronym": [], "definition": [] }

[ "<title>Introduction</title>", "<p>Inter-trochanteric fractures have traditionally been treated by closed reduction and internal fixation with a dynamic hip screw or an intramedullary device (gamma nail, reconstruction nail, proximal femoral nail or intramedullary hip screw) [##REF##13680275##1##, ####REF##9611022##2##, ##REF##11886906##3##, ##REF##9602800##4####9602800##4##]. Reduction is usually achieved by positioning the patient on a fracture table with the foot secured to a boot to aid in traction and rotation. These fractures and positioning for their surgical treatment pose a difficult problem when encountered in patients with below-knee amputations. Absence of the foot and part of the leg in these patients makes positioning on the fracture table challenging. We highlight the difficulties encountered in a patient with bilateral below-knee amputations undergoing fixation of an inter-trochanteric fracture and the various techniques available to overcome this problem.</p>" ]

[]

[]

[ "<title>Discussion</title>", "<p>Inter-trochanteric fractures of the femur are quite common in the elderly. Management of these fractures is essentially surgical and the various techniques used include dynamic hip screw fixation, intramedullary nailing and dynamic condylar screw fixation [##REF##13680275##1##, ####REF##9611022##2##, ##REF##11886906##3##, ##REF##9602800##4####9602800##4##]. Patients with bilateral below-knee amputations with inter-trochanteric fractures pose a special problem as positioning them on the fracture table is difficult due to the absence of the foot and part of the leg. The problem is accentuated when there is a need to apply traction for obtaining reduction of the fracture. There is little information in the literature on techniques to deal with this problem. We describe a few methods that can be used when this rare and unusual problem is encountered.</p>", "<p>If the fracture is undisplaced or minimally displaced, the limb can be placed on a radiolucent leg support (Figure ##FIG##0##1##) with the opposite hip kept abducted to allow access for the image intensifier. Traction and rotation of the hip can be performed by an assistant. An alternative is to fit the patient's prosthesis onto the stump and secure the foot of the prosthesis to the boot on the traction table (Figure ##FIG##1##2##). A radiolucent leg support should be placed under the limb for safety. These techniques cannot be used when the fracture is displaced and more traction is needed.</p>", "<p>If the fracture is displaced and greater traction is anticipated, the method of shortening the traction arm and inverting the boot to accommodate the flexed knee (Figure ##FIG##2##3##) and stump, as described by Al-Harthy <italic>et al. </italic>[##REF##9616404##5##], can be used. A standard boot should be used and the stump should be 12 cm or more (below the tibial tuberosity). If the stump is long, the boot tongue can be inverted for the stump to protrude. Upper tibial skeletal traction can be used if the stump is short but this method has some drawbacks. The skeletal pins may 'cut out' of the bone, which is usually osteoporotic, on applying traction. The other option is to use a distal femoral skeletal traction which would assist in traction.</p>" ]

[ "<title>Conclusion</title>", "<p>Hip fracture fixation surgery in patients with below-knee amputations is a difficult and challenging problem for the surgeon. The dilemma is on how to provide the traction and rotation required for reduction of the fracture. We believe that the techniques mentioned here to overcome this problem are safe and give the surgeon various options to handle this situation.</p>" ]

[ "<p>This is an Open Access article distributed under the terms of the Creative Commons Attribution License (<ext-link ext-link-type=\"uri\" xlink:href=\"http://creativecommons.org/licenses/by/2.0\"/>), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.</p>", "<title>Introduction</title>", "<p>Hip fracture fixation surgery in patients with below-knee amputations poses a challenging problem to the surgeon in terms of obtaining traction for reduction of the fracture. The absence of the foot and part of the leg in these patients makes positioning on the fracture table difficult. We highlight this difficult problem and suggest techniques to overcome it.</p>", "<title>Case presentation</title>", "<p>A 73-year-old man with bilateral below-knee amputations presented with a history of fall. Radiographs revealed an inter-trochanteric fracture of the femur. A dynamic hip screw fixation was planned for the fracture but the dilemma was on how to position the patient on the fracture table for the surgery. Special attention was needed in positioning the patient and in surgical fixation of the fracture.</p>", "<title>Conclusion</title>", "<p>Hip fracture fixation in patients with below-knee amputations poses a special problem in positioning for fracture reduction and fixation. In this case report, we share our experience and suggest techniques to use when encountering this difficult problem.</p>" ]

[ "<title>Case presentation</title>", "<p>A 73-year-old man presented to our department with a history of fall. He complained of pain in the right hip especially on movement of his hip. He had bilateral below-knee amputations following peripheral vascular disease and had below-knee suction prostheses fitted to his lower limbs for mobility. Radiographs of his pelvis and right hip revealed an undisplaced inter-trochanteric fracture of the femur. A dynamic hip screw fixation was planned for the fracture but the dilemma was how to position the patient on the fracture table for the surgery.</p>", "<p>The patient was positioned on a fracture table with a perineal post and the affected limb supported on a radiolucent leg support (Figure ##FIG##0##1##). The opposite below-knee stump was strapped securely to a leg support with the limb placed in abduction to allow easy access for the image intensifier (Figure ##FIG##0##1##). As the fracture was undisplaced, fixation of the fracture was performed with rotation of the hip by the assistant. The procedure was completed satisfactorily and postoperatively the patient was mobile with full weight-bearing after fitting prostheses to his lower limbs.</p>", "<title>Competing interests</title>", "<p>The authors declare that they have no competing interests.</p>", "<title>Consent</title>", "<p>Written informed consent was obtained from the patient for publication of this case report and any accompanying images. A copy of the written consent is available for review by the Editor-in-Chief of this journal.</p>", "<title>Authors' contributions</title>", "<p>UR was involved in collecting patient details, reviewing the literature and drafted the manuscript as the main author. RSY was involved in reviewing the literature and proofreading of the manuscript. AS was involved in critically revising the manuscript for important intellectual content. TKR was involved in conception of the study and revising the manuscript. All authors have read and approved the final manuscript.</p>" ]

[]

[ "<fig position=\"float\" id=\"F1\"><label>Figure 1</label><caption><p>Limb placed on radiolucent leg support with unaffected limb abducted for easy access of the image intensifier.</p></caption></fig>", "<fig position=\"float\" id=\"F2\"><label>Figure 2</label><caption><p>Prosthesis fitted onto the stump and the limb secured on the boot of the traction table.</p></caption></fig>", "<fig position=\"float\" id=\"F3\"><label>Figure 3</label><caption><p>Boot piece inverted to accommodate the flexed knee of the stump.</p></caption></fig>" ]

[]

[]

[]

[]

[]

[]

[]

[]

[ "<graphic xlink:href=\"1752-1947-2-296-1\"/>", "<graphic xlink:href=\"1752-1947-2-296-2\"/>", "<graphic xlink:href=\"1752-1947-2-296-3\"/>" ]

[]

{ "acronym": [], "definition": [] }

[]

[]

[]

[]

[]

[ "<p>This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.</p>", "<title>Background</title>", "<p>Coenzyme A (CoA) is an essential metabolite, synthesized from vitamin B₅by the subsequent action of five enzymes: PANK, PPCS, PPCDC, PPAT and DPCK. Mutations in <italic>Drosophila dPPCS </italic>disrupt female fecundity and in this study we analyzed the female sterile phenotype of <italic>dPPCS </italic>mutants in detail.</p>", "<title>Results</title>", "<p>We demonstrate that <italic>dPPCS </italic>is required for various processes that occur during oogenesis including chorion patterning. Our analysis demonstrates that a mutation in <italic>dPPCS </italic>disrupts the organization of the somatic and germ line cells, affects F-actin organization and results in abnormal PtdIns(4,5)P₂localization. Improper cell organization coincides with aberrant localization of the membrane molecules Gurken (Grk) and Notch, whose activities are required for specification of the follicle cells that pattern the eggshell. Mutations in <italic>dPPCS </italic>also induce alterations in scutellar patterning and cause wing vein abnormalities. Interestingly, mutations in <italic>dPANK </italic>and <italic>dPPAT-DPCK </italic>result in similar patterning defects.</p>", "<title>Conclusion</title>", "<p>Together, our results demonstrate that <italic>de novo </italic>CoA biosynthesis is required for proper tissue morphogenesis.</p>" ]

[ "<title>Findings</title>", "<p>Coenzyme A (CoA), the major acyl carrier in all organisms, constitutes an essential cofactor to support cellular metabolism [##REF##11153265##1##]. Synthesis of CoA occurs via a conserved route in which vitamin B₅is subsequently modified by five enzymes: PANK, PPCS, PPCDC, PPAT and DPCK [##REF##11278255##2##, ####REF##15014152##3##, ##REF##12860978##4##, ##REF##11923312##5####11923312##5##]. Although CoA biosynthesis is well characterized in bacteria and in <italic>in vitro </italic>systems [##REF##15893380##6##], only recently has the impact of abnormal CoA biosynthesis on animals been investigated [##REF##11238410##7##, ####REF##18407920##8##, ##REF##15525657##9##, ##REF##11479594##10####11479594##10##].</p>", "<title>Mutations in <italic>dPPCS </italic>impair female fecundity and fertility</title>", "<p>Previously, we isolated a <italic>Drosophila dPPCS </italic>mutant as a female sterile, neurologically impaired mutant and we demonstrated that CoA metabolism is required to maintain DNA integrity during development of the central nervous system [##REF##18407920##8##]. Here, we analyzed the female sterile phenotype of a hypomorphic allele of <italic>dPPCS </italic>(<italic>dPPCS</italic>¹) in detail. <italic>dPPCS</italic>³³mutants (a null allele) are homozygous lethal [##REF##18407920##8##], and in <italic>dPPCS</italic>^1/33mutants, no vitellogenic egg chambers were observed. Using immunohistochemistry and confocal laser scanning microscopy (supplement) we have analyzed the defects that occur during oogenesis (see for recent reviews [##REF##15704134##11##,##REF##12879448##12##]).</p>", "<p>At 48 h after eclosion (AE), the ovaries from <italic>dPPCS</italic>^1/1females were small compared to wild-type (wt) ovaries and mutant ovaries did not contain mature eggs (Fig. ##FIG##0##1Aa–b##). In wt, the oldest egg chambers found in newly eclosed females are at stage 7, and upon food intake, hormones are produced which trigger the egg chambers to proceed into vitellogenesis, a process whereby the oocyte accumulates nutrients and increases in size [##REF##11180967##13##]. At 72 h AE, 100% (n = 35) of the wt ovaries contained vitellogenic egg chambers, while only 11% of the <italic>dPPCS</italic>^1/1ovaries (n = 36) contained vitellogenic egg chambers (Fig. ##FIG##0##1Ac–d##). At 120 h AE, 80% of the <italic>dPPCS</italic>^1/1ovaries (n = 26) contained vitellogenic egg chambers; however, the two lobes were frequently different in size and displayed features of degenerating egg chambers (Fig. ##FIG##0##1Ae##).</p>", "<p>Between 144–192 h AE, <italic>dPPCS</italic>^1/1females deposited 0.03 (± 0.02 SEM) eggs/24 h, none of which hatched (n = 142 eggs), while wt females produced 10.0 (± 1.4 SEM) eggs/24 h, of which 90% hatched (n = 1005 eggs). It has been reported that a mid-oogenesis checkpoint monitors the integrity of pre-vitellogenic egg chambers, and that activation of this checkpoint results in the removal of abnormal egg chambers [##REF##11180967##13##]. A Tunnel assay was performed, which revealed that in <italic>dPPCS</italic>^1/1ovaries at 144 h AE, prior to vitellogenesis, a 6-fold increase of ovariols containing apoptotic egg chambers was observed, compared to wt ovaries (see additional file ##SUPPL##0##1##). Approximately 32% of <italic>dPPCS</italic>^1/1ovariols (n = 222) contained stage 5–7 egg chambers that displayed packaging defects (abnormal amount of germ line cells), while 4% of the wt ovariols (n = 109) contained egg chambers with packaging defects. When we expressed a <italic>dPPCS </italic>transgene (<italic>P[dPPCS]</italic>) in the <italic>dPPCS</italic>^1/1background, 11% (n = 166) of the ovariols displayed defects, demonstrating that <italic>dPPCS </italic>is required for early egg chamber development. Within <italic>dPPCS</italic>^1/1germaria, aberrant separation of the developing egg chambers by the intercyst cells likely results in production of egg chambers with abnormal interfollicular stalk cell and/or polar follicle cell formation, egg chambers with mispositioned oocytes, or egg chambers that display packaging defects (see additional file ##SUPPL##0##1##). Thus, the reduced fecundity of the <italic>dPPCS</italic>^1/1females is most likely due to production of aberrant egg chambers that did not pass the mid-oogenesis checkpoint and were absorbed.</p>", "<title><italic>dPPCS </italic>is required for F-actin remodeling during cytoplasmic dumping</title>", "<p>In addition to impaired fecundity, 80% of the eggs deposited by <italic>dPPCS</italic>^1/1females displayed a dumpless phenotype [##REF##15922834##14##] and a wide array of chorion patterning defects (Fig. ##FIG##0##1B##). Since patterning defects can arise from aberrant actin fiber formation within the nurse cells [##REF##8163553##15##,##REF##8044841##16##], we analyzed actin formation during cytoplasmic dumping. In stage 10 wt egg chambers, an elaborate network of F-actin bundles is assembled inside the nurse cells which is a prelude to cytoplasmic dumping. These bundles anchor the nurse cell nuclei to prevent them from entering the oocyte when the remaining nurse cell material is actively squeezed into the oocyte [##REF##12429700##17##]. Assembly of this F-actin network requires the Quail protein, which colocalizes with the F-actin fibers (Fig. ##FIG##1##2A##) [##REF##8044841##16##,##REF##10572041##18##]. In <italic>dPPCS</italic>^1/1egg chambers, assembly of the cytoplasmic F-actin fibers was disrupted, and the Quail protein failed to associate with the F-actin bundles and remained diffuse throughout the nurse cell cytoplasm (Fig. ##FIG##1##2B##). As a result of aberrant F-actin assembly, nurse cell nuclei were trapped inside ring canals during dumping (Fig. ##FIG##1##2D##, Table ##TAB##0##1##). Interestingly, we also found oocyte nuclei that were encapsulated by bundles of actin (Fig. ##FIG##1##2E##, Table ##TAB##0##1##). Furthermore, large actin fibers were assembled at the cortical membrane of the oocyte, and the follicular epithelium of the oocytes was frequently disorganized (Fig. ##FIG##1##2H–I##). Mutant oocytes also contained large clumps of F-actin (Fig. ##FIG##1##2H–I##, Table ##TAB##0##1##) and we frequently found nurse cell nuclei inside the oocyte compartment (Fig. ##FIG##1##2F, J##, Table ##TAB##0##1##).</p>", "<p>We stained freshly dissected ovaries with Nile red, which has fluorescent properties in the presence of triacylglycerol and sterol esters [##REF##3972906##19##], to determine if neutral lipid synthesis and transport of these lipids to the oocyte was disrupted during cytoplasmic dumping. In wt, synthesis of these neutral lipids increases in the germ line and somatic cells when egg chambers proceed into late stage oogenesis, and these neutral lipids are transported to the oocyte where they accumulate uniformly near the oocyte membrane (Fig. ##FIG##2##3A##). In <italic>dPPCS</italic>^1/1egg chambers, neutral lipid synthesis was reduced compared to wt, suggesting that the synthesis of neutral lipids is affected in <italic>dPPCS </italic>mutants (Fig. ##FIG##2##3B##). Furthermore, accumulation inside the oocyte of these lipids appeared abnormal compared to wt ovaries (compare Fig. ##FIG##2##3A## and ##FIG##2##3B##).</p>", "<p>Next, we investigated whether a mutation in <italic>dPPCS</italic>^1/1affects cell migration events due to defective F-actin organization. During stages 8–10, the border cells, which include the anterior polar cells and part of the main body epithelium, migrate through the nurse cell compartment towards the anterior end of the oocyte [##REF##12185849##20##]. In wt, when the border cells reach the oocyte and the centripetal follicle cells start migrating, Fasciclin III (FasIII) is expressed in the follicle cells of the dorsoanterior corner (Fig ##FIG##3##4A##). After centripetal follicle cells finished their migration, FasIII expressing cells form two distinct cell populations at the dorsoanterior surface of the oocyte. Here, formation of the dorsal appendages is initiated (Fig ##FIG##3##4B##) [##REF##15922834##14##]. In <italic>dPPCS</italic>^1/1egg chambers, centripetal migration was finished before the border cells reached the anterior of the oocyte (Fig ##FIG##3##4C##), indicating that these two cell migration events are not properly synchronized. Together, these data demonstrate that <italic>dPPCS </italic>is required for F-actin organization and cell migration events during oogenesis.</p>", "<title>Grk and Notch localization is disrupted in <italic>dPPCS^1/1</italic>egg chambers</title>", "<p>We hypothesized that disorganized tissue integrity may also affect the signaling routes required for specification of follicle cells that pattern the chorion. To investigate this, we stained ovaries with antibodies against Notch and Grk, which both are required for specification of the follicle cell populations that pattern the eggshell [##REF##15922834##14##,##REF##16828735##21##]. Although we cannot conclude that Grk or Notch signaling was disrupted in <italic>dPPCS</italic>^1/1ovaries, the localization of both proteins was frequently impaired compared to wt ovaries. In wt egg chambers, when the border cells reach the centripetal follicle cells, Notch is highly expressed at the dorsoanterior corner, where it is required for the specification of the dorsal appendage producing cells, while Notch expression is restricted to the nurse cell membranes during cytoplasmic dumping (Fig. ##FIG##3##4Ba##, see additional file ##SUPPL##0##1##) [##REF##1451667##22##]. In <italic>dPPCS</italic>^1/1stage 11 egg chambers, Notch localization was more diffuse throughout the nurse cells and not restricted to the membranes (Fig. ##FIG##3##4Ca##). Notch localization was also severely affected during late stage oogenesis (see additional file ##SUPPL##0##1##) and FasIII staining revealed that the dorsal appendage/operculum forming follicle cells were not properly organized (see additional file ##SUPPL##0##1##).</p>", "<p>In wt stage 9–10 egg chambers, Grk is localized at the dorsoanterior corner of the oocyte compartment. Although in <italic>dPPCS</italic>^1/1egg chambers, Grk was present at the dorsoanterior corner, the distribution of the protein was frequently impaired in stage 8–9 egg chambers (see additional file ##SUPPL##0##1##) and progressively worsened when egg chambers proceeded into later stages of oogenesis (see additional file ##SUPPL##0##1##).</p>", "<p>These findings imply that <italic>dPPCS </italic>is not required for cell specification or signaling per se, but merely required for cell organization and morphology. This is supported by the finding that aberrant intercyst cell migration/organization likely underlies the observed packaging and follicle cell specification defects during early oogenesis (see additional file ##SUPPL##0##1##).</p>", "<title>Membrane localization of PtdIns(4,5)P₂is impaired in <italic>dPPCS</italic>^1/1</title>", "<p>The levels of phospholipids are reduced in <italic>dPPCS </italic>mutant flies, indicating a general defect in phospholipid biosynthesis [##REF##18407920##8##]. Therefore, it is plausible to assume that phosphatidylinositol (PtdIns) production, the precursor for all phosphoinositides [##REF##3029593##23##], is also reduced. Although levels and localization of PtdIns have not been determined during <italic>Drosophila </italic>oogenesis, it is generally accepted that actin remodeling processes depend on PtdIns signaling [##REF##16704377##24##].</p>", "<p>To investigate whether PtdIns signaling was affected in <italic>dPPCS </italic>mutant ovaries, we expressed a PLCδ-PH-GFP fusion protein, which is able to bind to PtdIns(4,5)P₂[##REF##15743877##25##]. We used an Act5C-GAL4 driver to analyze PLCδ-PH-GFP expression and thus PtdIns(4,5)P₂localization in all cells. During wt cytoplasmic dumping, PtdIns(4,5)P₂is abundant at the cell membranes of the border cells and the apical membranes of the follicle cells that encapsulate the oocyte, while low levels of PtdIns(4,5)P₂can be detected at the nurse cell membranes (Fig. ##FIG##4##5A,B,E##). In contrast, PtdIns(4,5)P₂localization at the apical membranes of the follicle cells that encapsulate the oocyte was hardly detectable or absent in <italic>dPPCS</italic>^1/1egg chambers (Fig. ##FIG##4##5C,D,F,H##). Moreover, large patches of follicle cells that encapsulate the oocyte did not accumulate PtdIns(4,5)P₂at their membranes (Fig. ##FIG##4##5C,D,F##). Because aberrant apical localization of PtdIns(4,5)P₂at the follicle cell membranes coincides with impaired oocyte cortex integrity and abnormal F-actin organization (Fig. ##FIG##4##5H##), this suggests that altered PtdIns(4,5)P₂signaling could underlie the F-actin defects in <italic>dPPCS</italic>^1/1egg chambers.</p>", "<p>Although the F-actin/PtdIns(4,5)P₂connection should be investigated in more detail, we propose that F-actin remodeling within the <italic>Drosophila </italic>ovary likely depends on PtdIns(4,5)P₂signaling and that this lipid derived signaling route is disrupted in <italic>dPPCS</italic>^1/1. Abnormal cytoskeletal organization in <italic>dPPCS</italic>^1/1disrupts the overall shape of all membranous structures and the organization of the cells during morphogenesis. Disorganized tissue integrity could affect Notch and Grk localization and possibly signaling, which is required for specification of the follicle cells that pattern the eggshell, and causes severe chorion patterning defects.</p>", "<title><italic>dPPCS </italic>is required for patterning of various tissues</title>", "<p>Next, we wondered whether dPPCS is also required for morphogenesis of other tissues. Hereto, we closely investigated <italic>dPPCS</italic>^1/1flies for other morphological abnormalities. A stereotypical pattern of four scutellars exists on the dorsal surface of the wt scutellum, and <italic>dPPCS </italic>mutants displayed ectopic formation of scutellars (see additional file ##SUPPL##0##1##). Furthermore, <italic>dPPCS</italic>^1/1flies also developed ectopic wing veins (see additional file ##SUPPL##0##1##). Mutants initiated longitudinal vein formation between L3–L4 and L4–L5. These results show that <italic>dPPCS </italic>is required for morphogenesis of various tissues during <italic>Drosophila </italic>development.</p>", "<title>Mutations in <italic>de novo </italic>CoA synthesis disrupt morphogenesis</title>", "<p>Next, we investigated whether mutations in other CoA biosynthesis enzymes give rise to similar defects. Indeed, mutations in <italic>dPANK/fumble </italic>and the bifunctional enzyme <italic>dPPAT-DPCK </italic>result in similar characteristics compared to the <italic>dPPCS </italic>mutant phenotype. <italic>dPANK/fumble </italic>and <italic>dPPAT-DPCK </italic>mutant females have poorly developed ovaries, have fecundity defects, produce eggs that exhibit polarity defects, synthesize abnormal neutral lipids (droplets), and these mutants display scutellar and wing vein patterning defects (see additional file ##SUPPL##0##1##). As in <italic>dPPCS</italic>^1/1, a mutation in <italic>dPPAT-DPCK </italic>disrupts actin localization and results in plugging of the ring canals by nurse cell nuclei during dumping (see additional file ##SUPPL##0##1##). <italic>dPANK/fumble </italic>mutants produce small ball-shaped eggs, which are typically due to a loss of actin regulatory elements that control the polarized arrangement of F-actin fibers at the basal cortex of follicle cells required to establish planar cell polarity [##REF##15704134##11##]. These findings imply that impaired CoA synthesis in general disrupts morphogenesis, possibly due to aberrant F-actin organization. Because the biosynthesis route towards the production of CoA is conserved amongst species it would be interesting to explore the significance of CoA during processes that involve actin/PtdIns dynamics such as chemotaxis, axon growth cone guidance, endocytosis/exocytosis, cell division or actin dependent chromatin remodeling.</p>", "<title>Competing interests</title>", "<p>The authors declare that they have no competing interests.</p>", "<title>Authors' contributions</title>", "<p>FB, AR and OCMS conceived and designed the experiments. FB, AR and WL performed the experiments. FB, OCMS and HHK analyzed the data. FB and OCMS wrote the manuscript.</p>", "<title>Supplementary Material</title>" ]

[ "<title>Acknowledgements</title>", "<p>We thank L. Cooley and A. Wodarz for the <italic>UAS-PLCδ-PH-GFP </italic>line and S. Wasserman for the <italic>P[dPANK] </italic>line. This work was supported by a VIDI grant from the Netherlands Organization for Scientific Research (NWO; 971-36-400) to O.C.M.S and by a Topmaster grant from the Graduate School GUIDE to A.R.</p>" ]

[ "<fig position=\"float\" id=\"F1\"><label>Figure 1</label><caption><p><bold>Egg chamber development and eggshell patterning is disrupted in <italic>dPPCS</italic>^1/1</bold>. (A) Morphology of <italic>dPPCS</italic>^1/1ovaries was analyzed and compared with wt using light microscopy. (Aa) Bright field microscopy revealed that at 48 h after eclosion (AE), Wt ovaries are well developed and contain mature eggs (arrowheads). (Ab) At 48 h after eclosion, <italic>dPPCS</italic>^1/1ovaries were small in size compared to wt. (Ac-e) Ovaries were labeled with rhodamin-phalliodin to detect F-actin and stained with DAPI to visualize DNA. (Ac) At 72 h AE, wt ovaries contain vitellogenic egg chambers, as determined by the increased size of the oocyte compartment (asterisk). (Ad) <italic>dPPCS</italic>^1/1ovaries remained small in size (1 entire lobe is shown) and no vitellogenic egg chambers were observed. (Ae) At 120 h AE, <italic>dPPCS</italic>^1/1ovaries contained vitellogenic egg chambers (asterisks) and exhibited features of degenerating egg chambers (arrowheads). The 2 lobes were frequently different in size. (B) Chorion patterning was analysed in <italic>dPPCS</italic>^1/1. (Ba, Bd) Wt embryos have 2 dorsal appendages. (mp = micropyle; p = paddle; s = stalk). (Bb-c, Be-f) Embryos deposited by <italic>dPPCS</italic>^1/1mothers showed a dumpless phenotype and had a wide range of patterning defects, which were classified in 5 groups: (Bb) embryos with opercula positioned in a different angle in relation to the stalks (bracket, compare with Ba) and 4 appendages; (Bc) abnormal stalks (arrows); (Be) fused appendages (bracket, compare with Bd); and (Bf) embryos without dorsal appendages. Percentages are indicated (n = 142). The remaining 22.8% had either 2 dorsal appendages of different lengths, missing paddles, or a shift of the dorsal appendages posteriorly. Scale bars: 500 μm (Aa-b), 150 μm (Ac-e).</p></caption></fig>", "<fig position=\"float\" id=\"F2\"><label>Figure 2</label><caption><p><bold>Cytoplasmic F-actin filament assembly and dumping is disrupted in <italic>dPPCS</italic>^1/1egg chambers</bold>. To examine the morphology of <italic>dPPCS</italic>^1/1mutant ovaries, rhodamin-phalloidin was used to visualize the F-actin network in combination with DAPI to stain nuclei and various other antibodies as described. (Aa-Ac) Wt nurse cells assemble an elaborate network of transverse F-actin filaments prior to cytoplasmic dumping. Labeling using antibodies against Quail revealed colocalization of Quail with F-actin filaments in wt. (Ba-Bc) The cytoplasmic F-actin network is not properly formed inside <italic>dPPCS</italic>^1/1nurse cells, and Quail localization is diffuse inside the cytoplasm. To visualize nurse cell nuclei an antibody against lamin D_owas used. (C) In wt ovaries, F-actin bundles anchor the nurse cell nuclei during dumping. (D) <italic>dPPCS</italic>^1/1nurse cells fail to assemble F-actin filaments, and nurse cell nuclei trapped inside the ring canals during dumping were observed (arrows in D, F). (E) Example of a <italic>dPPCS</italic>^1/1oocyte nucleus encapsulated by F-actin fibers. (F) In <italic>dPPCS </italic>mutant egg chambers, nurse cell nuclei were found inside the oocyte compartment (arrow marks a nurse cell nucleus trapped inside a ring canal). (G) During cytoplasmic dumping, a tight array of F-actin is present at the subcortical membrane of wt oocytes. (H-I) The subcortical F-actin fibers at the membrane of the <italic>dPPCS</italic>^1/1oocyte compartment were increased in size and thickness (boxed arrowhead), and large clumps of F-actin were found within the oocyte compartment (arrowheads). (J) An antibody against <italic>D</italic>E-cadherin was used to visualize centripetal migrating follicle cells because these cells express high levels of <italic>D</italic>E-cadherin. In <italic>dPPCS </italic>mutants, migration of these cells occurred normally, but nurse cells are observed within the oocyte compartment after these cells finished their migration. An example of a nurse cell nucleus in the dorsoanterior corner of the <italic>dPPCS</italic>^1/1oocyte compartment is shown. Asterisks mark the oocyte nuclei. (oo) oocyte compartment. Scale bars: 100 μm (A-B), 20 μm (C-E), 50 μm (F-J).</p></caption></fig>", "<fig position=\"float\" id=\"F3\"><label>Figure 3</label><caption><p><bold>Synthesis and transport of neutral lipids is hampered in <italic>dPPCS </italic>mutant egg chambers</bold>. Freshly dissected wt and <italic>dPPCS</italic>^1/1ovaries were stained with Nile red to visualize neutral lipids and dissected ovaries were directly analyzed by CLSM. Images represent single confocal scans. (A) Wt nurse cells produce high levels of neutral lipids, which are transported towards the oocyte, and are uniformly accumulated near the oocyte membrane. (B) <italic>dPPCS</italic>^1/1nurse cells produced less neutral lipids compared to wt nurse cells. No uniformly accumulated lipids were observed within the mutant oocyte compartment compared to wt. (oo) oocyte compartment. Scale bars: 100 μm.</p></caption></fig>", "<fig position=\"float\" id=\"F4\"><label>Figure 4</label><caption><p><bold>A mutation in <italic>dPPCS </italic>affect follicle cell migration and patterning</bold>. <italic>dPPCS</italic>^1/1and wt egg chambers were stained with rhodamin falloidin (to visualize F-actin) and DAPI to stain nuclei and labeled with antibodies against FasIII and Notch to analyze follicle cell migration and patterning. (Aa-Ad) Wt stage 10 egg chamber. During stages 8–10, the border cells migrate through the nurse cell compartment towards the anterior end of the oocyte. When the border cells (arrow) reach the oocyte (stage 10) and the centripetal follicle cells start migrating (arrowheads), FasIII is expressed in the follicle cells of the dorsoanterior corner. At this stage, Notch is expressed at the membranes of the follicle cells of the dorsoanterior corner, where it is required for the specification of the dorsal appendage producing cells. (Ba-Bd) During stage 11, after the centripetal cells finished their migration, two patches of follicle cells can be found at the dorsoanterior corner of the epithelium of wt egg chambers (arrowheads). These follicle cells express high levels of FasIII and will initiate the production of the dorsal appendages. At this stage, Notch expression is restricted to the nurse cell membranes. (Ca-Cd) <italic>dPPCS</italic>^1/1egg chamber at stage 11. The centripetal follicle cells finished migration (arrowheads), but the border cells (arrow) failed to reach the centripetal follicle cells, while follicle cells of the dorsoanterior corner were already expressing FasIII. The border cell cluster is surrounded by an elaborate network of F-actin. Notch localization is not restricted to the nurse cell membranes and shows a more diffuse pattern. Boxed arrowheads point to two nurse cell nuclei that seem to contact (push against) the centripetal follicle cells. Asterisks mark the position of the oocyte. Scale bars: 50 μm.</p></caption></fig>", "<fig position=\"float\" id=\"F5\"><label>Figure 5</label><caption><p><bold>PtdIns(4,5)P₂localization and expression is affected in <italic>dPPCS</italic>^1/1</bold>. We overexpressed a PLCδ-PH-GFP fusion protein [##REF##15743877##25##] under the control of a ubiquitously expressed Act5C-GAL4 driver to analyze PtdIns(4,5)P₂localization and expression in a wt (A,B,E,G) and <italic>dPPCS</italic>^1/1(C,D,F,H) background. (A,B) During wt oogenesis, the PLCδ-PH-GFP fusion protein is present at the border cells (asterisk) and the apical membranes of the follicle cells that encapsulate the oocyte compartment, while the nurse cell membranes do not accumulate the fusion protein. (C,D) In <italic>dPPCS</italic>^1/1egg chambers, the PLCδ-PH-GFP fusion protein is not present at the apical membranes of the follicle cells that encapsulate the oocyte (arrowheads) and large patches of follicle cells are not labeled (dashed lines). In mutant ovaries, the follicular epithelium is sometimes disrupted (arrows in C), and the nurse cell membranes accumulate patches of high levels of the PLCδ-PH-GFP fusion protein (arrow in D). The PLCδ-PH-GFP fusion protein was also frequently detected at membranes of cells that are (based on their localization) most likely border cells (boxed arrowhead in D). (E-H) For further analysis, F-actin organization was analyzed in combination with localization of the PLCδ-PH-GFP fusion protein. (Ea-Ec) In wt egg chambers, the oocyte cortex is in close contact with the apical membranes of the follicle cells, which accumulate the PLCδ-PH-GFP fusion protein. (F) In <italic>dPPCS</italic>^1/1egg chambers, the oocyte cortex is disrupted (arrowheads in Fa) and the follicle cells do not accumulate the PLCδ-PH-GFP fusion protein (arrowheads and dashed lines in Fb). Arrows point to defects in cell organization of the follicular epithelium. (G) Higher magnification of a wt egg chamber, showing that the PLCδ-PH-GFP fusion protein accumulates at the apical membranes of follicle cells in close contact with the oocyte cortex (boxed arrowheads). (H) Higher magnification of a <italic>dPPCS</italic>^1/1egg chamber, showing that the apical membranes of follicle cells in close contact with the oocyte cortex do not accumulate the PLCδ-PH-GFP fusion protein (arrowheads in Ha). This coincides with impaired oocyte cortex morphology and aberrant F-actin nucleation (arrow in Hb). DAPI was used to visualize DNA. Scale bars: 150 μm (A,C), 100 μm (B,D), 50 μm (E-H).</p></caption></fig>" ]

[ "<table-wrap position=\"float\" id=\"T1\"><label>Table 1</label><caption><p>Mutations in <italic>dPPCS </italic>affect egg chamber development, stage 10–11 F-actin organization &amp; cytoplasmic dumping</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td/><td align=\"center\" colspan=\"3\">% of egg chambers</td></tr><tr><td/><td colspan=\"3\"><hr/></td></tr><tr><td/><td align=\"left\"><italic>wild-type</italic></td><td align=\"left\"><italic>dPPCS</italic>^1/1</td><td align=\"left\"><italic>P[dPPCS];dPPCS</italic>^1/1</td></tr></thead><tbody><tr><td align=\"left\">Nurse cells trapped inside the oocyte</td><td align=\"left\">0.0 (n = 100)</td><td align=\"left\">20.9 (n = 67)</td><td align=\"left\">1.8 (n = 56)</td></tr><tr><td align=\"left\">F-actin clumps in ooplasm</td><td align=\"left\">3.0 (n = 100)</td><td align=\"left\">50.1 (n = 53)</td><td align=\"left\">7.0 (n = 57)</td></tr><tr><td align=\"left\">Aberrant F-actin in nurse cells</td><td align=\"left\">0.0 (n = 100)</td><td align=\"left\">92.2 (n = 51)</td><td align=\"left\">29.5 (n = 61)</td></tr><tr><td align=\"left\">Nurse cells plugging ring canals</td><td align=\"left\">0.0 (n = 100)</td><td align=\"left\">71.0 (n = 62)</td><td align=\"left\">15.5 (n = 58)</td></tr><tr><td align=\"left\">Oocytes with disorganized subcortical F-actin</td><td align=\"left\">0.0 (n = 100)</td><td align=\"left\">43.7 (n = 55)</td><td align=\"left\">0.0 (n = 53)</td></tr><tr><td align=\"left\">Oocyte nuclei with F-actin fibers</td><td align=\"left\">0.0 (n = 100)</td><td align=\"left\">18.8 (n = 48)</td><td align=\"left\">0.0 (n = 46)</td></tr></tbody></table></table-wrap>" ]

[]

[]

[]

[]

[]

[ "<supplementary-material content-type=\"local-data\" id=\"S1\"><caption><title>Additional file 1</title><p>dPPCS, dPANK/fumble and dPPAT-DCPK mutants show comparable defects during oogenesis, and abnormal vein and scutellar patterning. The data provided show additional morphological information concerning defective egg chamber development and abnormal vein and scutellar patterning in dPPCS, dPANK/fumble and dPPAT-DCPK mutants.</p></caption></supplementary-material>" ]

[ "<table-wrap-foot><p>Nurse cells were stained with DAPI to detect DNA and labeled with rhodamin-phalloidin to visualize the F-actin network.</p><p><italic>P[dPPCS] </italic>is a FLAG-tagged dPPCS cDNA under control of a ubiquitin promoter.</p></table-wrap-foot>" ]

[ "<graphic xlink:href=\"1756-0500-1-75-1\"/>", "<graphic xlink:href=\"1756-0500-1-75-2\"/>", "<graphic xlink:href=\"1756-0500-1-75-3\"/>", "<graphic xlink:href=\"1756-0500-1-75-4\"/>", "<graphic xlink:href=\"1756-0500-1-75-5\"/>" ]

[ "<media xlink:href=\"1756-0500-1-75-S1.pdf\" mimetype=\"application\" mime-subtype=\"pdf\"><caption><p>Click here for file</p></caption></media>" ]

{ "acronym": [], "definition": [] }

[ "<title>Introduction</title>", "<p>Epstein-Barr virus (EBV) is widely recognized as a causative agent of nasopharyngeal carcinoma (NPC), as most NPC tumors are monoclonal proliferations of latently EBV-infected cells ##REF##12450729##[1]##. Latent EBV infection in NPC involves expression of four viral proteins; two latent membrane proteins (LMP1 and LMP2), one nuclear protein (EBNA1) and one secreted protein (BARF1) ##REF##12450729##[1]##,##REF##15778977##[2]##. LMP1, LMP2 and BARF1 have all been reported to have cellular effects that may contribute to the development of NPC, although LMP1 and BARF1 are not consistently detected in all NPC tumors ##REF##15498875##[3]##–##REF##15064715##[5]##. EBNA1 is required for the replication and stable persistence of EBV episomes in proliferating cells and is the only EBV protein that is expressed in all EBV-associated tumors ##UREF##0##[6]##. EBNA1 enables the expression of the other EBV latency proteins, however, whether or not EBNA1 directly contributes to the development of tumors has not been clear.</p>", "<p>A number of observations in the literature are consistent with a role for EBNA1 in the proliferation of EBV-positive cells. For example, interference with EBNA1 function in EBV-positive Burkitt's lymphoma cells, by overexpression of a dominant-negative EBNA1 mutant, increased cell death ##REF##14603034##[7]##. Similarly, down-regulation of EBNA1 in Raji Burkitt's lymphoma or EBV-positive epithelial cells by RNA interference decreased cell proliferation ##REF##16180023##[8]##,##REF##16343579##[9]##. However, since EBNA1 is needed to maintain the EBV episomes and to enhance expression of other latency proteins, it is not clear from the above observations whether EBNA1 is directly affecting cell proliferation or is functioning indirectly by enabling expression of other EBV gene products.</p>", "<p>Other studies have investigated whether expressing EBNA1 in various EBV-negative cancer cells affects tumorgenicity. EBNA1 expression in HONE-1 NPC cells was found to increase primary tumor formation as well as metastases in nude mice ##REF##8958799##[10]##. EBNA1 expression in Hodgkin's lymphoma cells enhanced their ability to form tumors in non-obese diabetic-SCID mice but not in regular SCID mice ##REF##9882370##[11]##. In addition, Kaul et al ##REF##17634231##[12]## found that expression of EBNA1 in a breast carcinoma cell line promoted the rate of tumor growth in nude mice, reversed the growth inhibitory effect of the cellular Nm23-H1 protein and increased lung metastases.</p>", "<p>The molecular basis for the observed effects of EBNA1 on cell proliferation are largely unknown, although an interaction between EBNA1 and the cellular ubiquitin specific protease, USP7 or HAUSP, has been proposed to be partially responsible ##REF##12783858##[13]##. USP7 binds and stabilizes p53 ##REF##11923872##[14]##, and EBNA1 was found to block the USP7-p53 interaction <italic>in vitro</italic> by competing for the same binding pocket on USP7 ##REF##14506283##[15]##–##REF##16474402##[17]##. In keeping with these findings, expression of EBNA1 (but not a USP7-binding mutant of EBNA1) in U2OS cells was shown to protect these cells from apoptosis in response to DNA damage by interfering with p53 stabilization ##REF##15808506##[16]##. However, USP7 is likely to have multiple cellular roles and the functional significance of the EBNA1-USP7 interaction remains to be determined in the context of latent EBV infection.</p>", "<p>Few studies have investigated the role of EBNA1 in NPC. Studies on the contribution of EBV proteins to NPC in general have been hampered by the lack of EBV-positive NPC cell lines, since NPC cells tend to rapidly lose the EBV genomes when propagated in culture. The isolation of a NPC cell line (C666-1) that stably maintains EBV episomes ##REF##10449618##[18]## has greatly facilitated NPC studies, enabling a comparison to EBV-negative NPC cell lines. We have compared C666-1 cells to the EBV-negative NPC cell lines CNE2 ##REF##1631151##[19]## and HK1 ##REF##6259064##[20]## in order to better understand cellular alterations caused by EBV infection that may contribute to cell transformation. Here we show that EBV latent infection in NPC cells is associated with the disruption of host PML nuclear bodies (NBs) and that EBNA1 is entirely responsible for this effect. Consistent with the known importance of PML NBs in p53 activation and DNA damage responses, we also show that EBNA1 expression in NPC impairs both of these processes.</p>" ]

[ "<title>Materials and Methods</title>", "<title>Cell lines</title>", "<p>The Saos-2 p53-negative, human osteoblast cell line was cultured in DMEM (Sigma) supplemented with 10% fetal calf serum. The EBV-positive NPC C666-1 cell line and the EBV-negative NPC cell lines HK1 and CNE2Z (CNE2) (both of which lost the EBV genomes after growth in culture) have been previously described ##REF##10449618##[18]##–##REF##6259064##[20]##. C666-1 and HK1 were grown in HEPES-modified RPMI 1640 (Sigma), while CNE2 was maintained in alpha minimal essential media (αMEM, Gibco), in all cases supplemented with 10% fetal calf serum. To generate the CNE2E and HK1E cell lines, EBNA1 cDNA was cloned into pcDNA3.1/hygro (−/−; Invitrogen), and CNE2 and HK1 cells were transfected with 10 µg of linearized plasmid. Transfected cells were maintained at low densities in medium supplemented with hygromycin B (Invitrogen; 0.75mg/mL for CNE2 and 0.5mg/mL for HK1) to allow growth of individual colonies. Colonies were examined for EBNA1 expression by IF microscopy, then picked and propagated for further studies in media containing 0.5 mg/mL (CNE2E) and 0.35 mg/mL (HK1E) of hygromycin B, respectively.</p>", "<title>Transfections and RNA interference</title>", "<p>To generate CNE2 or Saos-2 cells transiently expressing EBNA1, 1.5×10⁵ cells were transfected with 2 µg of an EBNA1 expressing plasmid containing EBV oriP, pc3OriPE ##REF##11836426##[53]##, unless otherwise indicated, using lipofectamine 2000 (Invitrogen). Where indicated, the same plasmid expressing EBNA1 mutants Δ325–376 or Δ395–450 was used ##REF##12783858##[13]##,##REF##10074103##[54]##, and the same plasmid lacking EBNA1 cDNA (pc3OriP) was used as a negative control ##REF##11836426##[53]##. 48 hrs later, cells were fixed for IF imaging as described below. For RNA interference experiments, 1.5×10⁵ CNE2E or C666-1 cells were transfected with 50 pmol of siRNA against GFP (<named-content content-type=\"gene\">GCAAGCUGACCCUGAAGUUCAU</named-content>), against EBNA1 (<named-content content-type=\"gene\">GGAGGUUCCAACCCGAAAUTT</named-content>) or against USP7 (<named-content content-type=\"gene\">UCAAGAUGACUACCAGCUG</named-content>) using 2 µL of lipofectamine 2000. For C666-1 and one CNE2E sample (##FIG##2##Figure 3A and B##), cells underwent an identical second round of siRNA transfection 72 hours after the first transfection. For ##FIG##4##Figure 5C##, cells were subjected to three rounds of transfection with siRNA against USP7 or GFP prior to transfection with pc3OriPE.</p>", "<title>Immunofluorescence microscopy</title>", "<p>Cells grown on coverslips were fixed with 3.7% formaldehyde in phosphate buffered saline (PBS) for 20 min, rinsed twice in PBS and permeabilized with 1% Triton X-100 in PBS for 5min. Samples were blocked with 4% BSA in PBS followed by incubation with primary antibodies against EBNA1 (R4 rabbit serum at 1∶300 dilution ##REF##12783858##[13]##) and PML (Santa Cruz PG-M3 at 1∶50 dilution) and incubation with the secondary antibodies goat anti-rabbit Alexafluor 555 (Molecular Probes) and goat anti-mouse Alexafluor 488 (Molecular Probes) in 4% BSA. Coverslips were mounted onto slides using ProLong Gold antifade medium containing DAPI (Invitrogen). Images were obtained using the 40 x oil objective on a Leica inverted fluorescent microscope and processed using OpenLAB (ver.X.0) software. PML nuclear bodies were quantified by counting all visible PML foci in 100 cells.</p>", "<title>Western blots</title>", "<p>Cells were lysed in 9 M urea, 5 mM Tris-HCl (pH 6.8) and briefly sonicated. 100 µg of total protein was subjected to 10% SDS-PAGE and transferred to nitrocellulose. Where indicated, CNE2E cells were treated with 10 µM MG132 for 8 or 10 hours prior to lysis. Membranes were blocked in 5% non-fat dry milk in PBS, then incubated with antibodies against PML (Chemicon AB1370; 1∶2000 dilution), EBNA1 (OT1X at 1∶2000; kindly supplied by Japp Middeldorp), actin (Ab-1, Oncogene Research Products; 1∶20 000), Sp100 (Chemicon 1380, 1∶1000 dilution) or USP7 (rabbit serum against full-length USP7). After washing, blots were probed with goat anti-mouse peroxidase (1∶3000) or goat anti-rabbit peroxidase (1∶5000) from Santa Cruz, then developed using chemiluminescence reagents (ECL, Perkin Elmer). Membranes were stripped in 0.1 M glycine pH 2.9 for 30 min, washed in PBS-Tween, blocked and re-probed with the next antibody as described above.</p>", "<p>Experiments in ##FIG##6##Figure 7## were performed as above except that some cells were treated with 10 µg/mL of etoposide 24 hrs prior to harvesting and 80 µg (p53 blot) or 60 µg (acetyl-p53 blot) of total protein was analyzed by Western blotting using the following antibodies: PAb 1801 for p53 ##REF##2428616##[55]## (a gift from Sam Benchimol) and antibody 2525 for acetyl-p53 K382 (Cell Signaling Technologies). For acetyl-p53 blots, membranes were blocked and incubated with primary antibody in 5% BSA, 50 mM Tris pH7.4, 150 mM NaCl, 0.1% Tween-20. The same experiment was also performed in Hela cells transfected with pc3OriP or pc3OriPE. In the later case, EBNA1 expression was confirmed in approximately 80% of the cells by IF prior to etoposide treatment (10 µg/mL) 48 hours post-transfection. 100 µg of cell lysate was Western blotted as described above except that the anti-p53 antibody was DO-1 from Santa Cruz (sc-126).</p>", "<title>Co-immunoprecipitation</title>", "<p>Log phase C666-1 cells were lysed in IP buffer (20 mM Tris-HCl pH 7.5, 150 mM NaCl, 1 mM MgCl₂, 10% glycerol, 1% Triton X-100, and protease inhibitors) on ice for 30 min. After centrifugation, the supernatant was pre-cleared with protein A/G agarose (Santa Cruz) and then equal amounts (4 mgs) were incubated for 4 hr at 4°C with either mouse IgG coupled to agarose (Santa Cruz; negative control) or with OT1X EBNA1 monoclonal antibody followed by protein A/G agarose overnight at 4°C. Beads were collected by centrifugation, washed in IP buffer then boiled in SDS loading buffer. Immunoprecipitated proteins were separated by SDS-PAGE and Western blotted as described above. The positions of PML I and PML IV isoforms on Western blots was determined by transfecting C666-1 cells with constructs expressing FLAG-PML I or FLAG-PML IV ##REF##15922731##[56]## and analyzing 30 µg of lysate by Western blotting using anti-FLAG antibody (Abcam AB21536; 1∶10,000 dilution).</p>", "<title>FACS analysis</title>", "<p>CNE2 and CNE2E cells (before and after siRNA treatment) seeded at 60% confluence in 10 cm dishes were treated with UV (50×10² µJ/cm²) or 10 µg/mL of etoposide. 24 hours later, adherent cells were harvested, fixed in 70% ethanol, treated with RNAse (50 µg/mL) at 37^oC for 1 hour and stained with propidium iodide. DNA content was analyzed immediately after propidium iodide treatment on a FACScalibur (Becton Dickinson, USA) and cell cycle analysis was performed using Modfit LT 3.1 (Verity Software House).</p>", "<title>Apoptosis Assay</title>", "<p>CNE2 and CNE2E cells were treated with etoposide as stated above. 48 hours later, cells were processed for TUNEL staining using the APO-BrdU TUNEL Assay Kit (Invitrogen, MP23210) according to the manufacturer's instructions. Cells were then mounted on coverslips, counter-stained with DAPI and analyzed by fluorescence microscopy. Apoptotic index was calculated as the number of TUNEL-positive cells divided by the total number of cells. The experiment was done in triplicate and at least 100 cells were counted for each sample.</p>", "<title>Cell viability assay</title>", "<p>Cell viability was measured using Trypan blue (Gibco) exclusion assay as follows: CNE2 and CNE2E cells were seeded in 12-well plates such that they reached 80–90% confluence at time of harvesting. 24 hours later, cells were treated with etoposide (10 µg/mL) or UV (50×10² µJ/cm²). After growing for 24, 48 and 72 hours, floating and adherent cells were harvested, washed in PBS, stained with 0.04% Trypan blue in PBS and counted immediately. Experiments were done in triplicate and at least 200 cells were counted for each replicate.</p>" ]

[ "<title>Results</title>", "<title>EBNA1 disrupts PML NBs in NPC cells</title>", "<p>In initial studies comparing EBV-positive and EBV-negative NPC cell lines, we examined the host PML NBs, which have the PML protein as the main constituent and have been implicated in many important cellular processes ##UREF##1##[21]##. We were particularly interested in examining PML NBs in the context of NPC because the disruption of PML NBs, or lack of the PML protein, is a factor in the development of several types of cancer ##REF##14970276##[22]##,##REF##9488655##[23]## and because DNA viruses are known to have mechanisms to disrupt PML NBs ##REF##11704855##[24]##. Immunofluorescence (IF) microscopy for PML revealed that C666-1 cells have considerably fewer PML NBs than do CNE2 or HK1 cells (average number per cell of 4, 16 and 11 respectively; ##FIG##0##Figure 1A and 1B##), suggesting that some aspect of EBV infection disrupts PML NBs. We investigated whether this involved the EBNA1 protein by down-regulating EBNA1 expression with siRNA. This treatment greatly decreased EBNA1 expression in some but not all of the C666-1 cells allowing a direct comparison of silenced and non-silenced cells in the same culture (##FIG##0##Figure 1A## bottom row). The number of PML NBs was found to increase 2–3 fold upon EBNA1 silencing, as compared to non-treated cells or cells treated with siRNA against green fluorescence protein (GFP), indicating that EBNA1 contributes to PML disruption in C666-1.</p>", "<p>To determine whether EBNA1 expression was sufficient to disrupt PML NBs, we generated CNE2E and HK1E cell lines in which EBNA1 was constitutively expressed in CNE2 and HK1 NPC cells from an integrated cassette. EBNA1 expression levels in HK1E and CNE2E were shown by immunoblot to be approximately 2-fold and 3-fold higher than in C666-1, respectively (##SUPPL##0##Figure S1##). EBNA1 expression in both cell lines resulted in a notable decrease in the number of PML NBs per cell (##FIG##0##Figure 1B##). To further verify that this effect was caused by EBNA1, CNE2E cells were treated with siRNA to down-regulate EBNA1 expression. EBNA1 silencing restored the number of PML NBs to the level seen in the parent CNE2 cells (##FIG##0##Figure 1B##). We also examined whether transient expression of EBNA1 was sufficient to cause PML disruption. To this end, CNE2 cells were transfected with an EBNA1 expression plasmid and examined by IF 48 hrs later (##FIG##1##Figure 2A##). EBNA1 expression lowered the number of PML NBs in a dose-dependent manner, where PML NBs were decreased 5-fold in cells staining brightly for EBNA1 and decreased 3-fold in those with lighter EBNA1 staining (##FIG##1##Figure 2B##). Therefore EBNA1 expression has an immediate effect on PML NBs.</p>", "<title>EBNA1 lowers PML protein levels</title>", "<p>Viral proteins have been found to decrease PML NBs either by inducing the degradation of the PML protein or by disrupting the interaction of PML proteins required to form NBs ##REF##11704855##[24]##. To address the mechanism by which EBNA1 expression disrupts PML NBs, we compared the levels of PML isoforms in CNE2 before and after stable (CNE2E) or transient expression of EBNA1 by Western blotting. PML is known to exist as several isoforms (comprised of alternative spliced and modified forms), resulting in multiple bands migrating between 60 and 200 Kda on PML immunoblots ##UREF##2##[25]##,##REF##16778193##[26]##. Down-regulation of EBNA1 expression in C666 cells resulted in increased expression of all PML isoforms (##FIG##2##Figure 3A##). Similarly, EBNA1 expression in CNE2E resulted in a dramatic decrease in all PML isoforms (##FIG##2##Figure 3B## compare lanes 1 and 2), which was restored by silencing EBNA1 expression (##FIG##2##Figure 3B## lanes 3 and 4). In addition, transient EBNA1 expression in CNE2 decreased the level of all PML isoforms in a dose-dependent manner (##FIG##2##Figure 3C##). Therefore EBNA1 expression results in the loss of PML protein, as opposed to dispersal of PML from the foci. However the level of another PML NB component, Sp100, was unaffected by EBNA1 demonstrating the specificity of this effect (##FIG##2##Figure 3C##). Effects of EBNA1 expression on the level of PML mRNA was also examined by RT-PCR to rule out potential effects on PML transcription. As expected, no change in the level of PML transcripts was evident, indicating that the EBNA1-mediated PML effects were occurring at the protein level (##SUPPL##1##Figure S2##).</p>", "<p>The EBNA1-induced loss of PML protein suggests that EBNA1 might be increasing the degradation of PML isoforms by the proteasome. We tested this possibility by examining the effect of blocking the proteasome in CNE2E cells with MG132 (##FIG##2##Figure 3D##). This treatment was found to increase the levels of all PML isoforms, and higher molecular weight forms suggestive of polyubiquitination also became visible. Therefore the loss of PML protein caused by EBNA1 is proteasome dependent.</p>", "<title>EBNA1 associates with PML NBs through a specific PML isoform</title>", "<p>In most cells, EBNA1 is found throughout the nucleus making it difficult to assess whether some of the EBNA1 localizes to PML NBs. However, some of the transiently transfected CNE2 cells expressed very low levels of EBNA1 and, in these cells, discreet EBNA1 foci were observed, many of which localized to PML NBs (##FIG##3##Figure 4A##). In addition, EBNA1 foci are frequently seen in C666-1, which naturally express low levels of EBNA1, and these foci often correspond to or overlap with PML NBs, even though few PML NBs are present in C666-1 (##FIG##3##Figure 4A##).</p>", "<p>To further assess the interaction of EBNA1 with PML in the context of a latent infection, EBNA1 was immunoprecipitated from C666-1 and co-precipitating proteins were analysed for PML (##FIG##3##Figure 4B##). One PML isoform was consistently found to co-immunoprecipitate with EBNA1 (##FIG##3##Figure 4B## lane 3). Interestingly this did not correspond to the most prevalent PML band in the lysate (presumably isoforms I and II according to its size and abundance) but rather corresponded to a less abundant form consistent with the size of PML isoform IV ##REF##16778193##[26]##. EBNA1 was also found to preferentially bind PML IV over PML I when FLAG-tagged versions of these proteins were over-expressed in CNE2 cells along with EBNA1 (##SUPPL##2##Figure S3##).</p>", "<title>EBNA1-mediated disruption of PML NBs involves USP7</title>", "<p>To gain insight into the mechanism by which EBNA1 induces loss of PML NBs and protein, we tested the ability of EBNA1 mutants to disrupt PML NBs after transfection in CNE2 cells (##FIG##1##Figure 2A and 2C##). Initially we tested the EBNA1 Δ325–376 mutant, as this mutation disrupts the interaction of EBNA1 with cellular chromatin and abrogates the transcriptional activation function of EBNA1 ##REF##10799567##[27]##,##REF##11265753##[28]##. However EBNA1 Δ325–376 disrupted PML NBs to the same degree as wildtype EBNA1 indicating that neither transcriptional activation nor strong chromatin interactions are required for the observed effects. An EBNA1 mutant, Δ395–450, that is fully functional for all of the known functions of EBNA1 (replication, segregation and transcriptional activation) but fails to bind the cellular USP7 protein was also tested for PML effects ##REF##12783858##[13]##. USP7 is known to be partially associated with PML NBs and associates with another herpesvirus protein (ICP0 or Vmw110 from herpes simplex virus) that also disrupts PML NBs through loss of PML protein ##REF##9130697##[29]##–##REF##9658103##[31]##. Unlike wildtype EBNA1, Δ395–450 caused no obvious change in the number of PML NBs or the level of PML protein, suggesting that USP7 binding might be important for PML disruption.</p>", "<p>The role of USP7 in EBNA1-mediated disruption of PML NBs was further investigated by silencing USP7 by siRNA treatment in the CNE2E cells that are stably expressing EBNA1. USP7 silencing restored the number of PML NBs and the level of the PML protein (##FIG##4##Figure 5A and 5B##), indicating that EBNA1 does not disrupt PML NBs in the absence of USP7. Similar experiments were conducted in which CNE2 cells were transfected with siRNA against USP7 (or GFP as a negative control) prior to transient expression of EBNA1. The siUSP7 treatment was confirmed by IF to silence USP7 expression in virtually all of the CNE2 cells prior to transfection of the EBNA1 expression plasmid (##FIG##4##Figure 5C, left panel##). Cells pretreated with siUSP7 had numerous PML NBs regardless of whether or not EBNA1 was expressed, whereas EBNA1 continued to diminish PML NBs in cells pretreated with siGFP (##FIG##4##Figure 5C, right panel##). Similarly, pretreatment with siUSP7 but not siGFP interfered with EBNA1-induced loss of PML protein as determined by Western blotting (##FIG##4##Figure 5D##). Therefore EBNA1-mediated disruption of PML NBs requires USP7.</p>", "<p>Since USP7 can alter p53 levels ##REF##11923872##[14]##,##REF##15053880##[32]## and p53 induces PML transcription ##REF##14992722##[33]##,##REF##9398618##[34]##, we wanted to ensure that the observed effects of EBNA1 on PML were not due to interference with p53 stabilization by USP7. Therefore we examined whether the EBNA1 effects on PML were independent of p53 by expressing EBNA1 in the p53-null Saos-2 cells. As shown in ##FIG##5##Figure 6##, EBNA1 reduced the number of PML NBs per cell and the level of PML protein to a similar degree as in the NPC cells lines. Therefore the disruption of PML NBs by EBNA1 does not involve p53.</p>", "<title>EBNA1 interferes with p53 activation, DNA repair and apoptosis</title>", "<p>Considerable data indicate that PML NBs are important for p53 activation, apoptosis and DNA repair which would have important consequences for the development of NPC. Therefore we asked whether the effect of EBNA1 on PML NBs was sufficient to disrupt these processes. PML NBs are required for the activation of p53 through acetylation ##REF##11025664##[35]##,##REF##10910364##[36]##, and therefore we compared p53 activation in CNE2 and CNE2E cells after treatment with the DNA damaging agent etoposide (##FIG##6##Figure 7A##). We consistently observed that the EBNA1-expressing cells had an impaired ability to acetylate p53 (at K382) while the induction of p53 was affected to a lesser degree. In Hela cells, EBNA1 has no obvious effect on PML NBs (data not shown) and therefore we examined the effect of EBNA1 expression on p53 activation in Hela cells to verify that this effect involved PML NB disruption. As shown in ##FIG##6##Figure 7B##, p53 acetylation in Hela cells occurred in response to etoposide treatment at least as efficiently in the presence of EBNA1 as in its absence. Therefore PML disruption by EBNA1 appears to be responsible for the lack of p53 acetylation.</p>", "<p>We examined the effect of EBNA1 expression on DNA repair by comparing FACS profiles of CNE2 and CNE2E after inducing DNA damage with UV or etoposide treatment (##FIG##7##Figure 8A##). Previous studies have shown that unrepaired DNA damage is reflected by the accumulation of cells in S-phase, while cells that have repaired the damage pass through S and accumulate either in G2/M or G1 depending on which DNA damage checkpoint has been activated ##REF##17483520##[37]##–##REF##16868026##[39]##. Hence silencing of PML or a number of DNA repair proteins has been found to increase the percentage of cells in S phase after DNA damage ##REF##17483520##[37]##,##REF##16868026##[39]##. Similarly, we consistently observed that CNE2E cells had a higher fraction of cells in S phase after UV or etoposide treatment as compared to CNE2 cells (compare profiles ii and v, and profiles iii and vi in ##FIG##7##Figure 8A##) even though the cell cycle distribution of the two cell lines was indistinguishable prior to treatment. In multiple experiments the percentage of CNE2 cells in S phase after UV or etoposide treatment was 55.8±2.5 and 55.4±1.5, respectively, while the same treatments in CNE2E cells resulted in S-phase percentages of 66.5±2.2 and 91.4±0.2, respectively. In both cases, differences with and without EBNA1 are statistically significant with p values &lt;0.001. This effect was confirmed to be due to EBNA1 expression, as down-regulation of EBNA1 in CNE2E with siRNA reduced the S-phase accumulation after DNA damage as compared to the control siRNA treatment against GFP (##FIG##7##Figure 8A##, compare profiles viii and xi, and profiles ix and xii). Therefore EBNA1 expression results in an impaired ability to repair DNA damage, consistent with the disruption in PML NBs.</p>", "<p>The effect of EBNA1 expression in CNE2 cells on apoptosis was also examined by TUNEL assay. The percentage of cells that became TUNEL-positive after etoposide treatment was decreased two-fold in the presence of EBNA1 (##FIG##7##Figure 8B##), showing that EBNA1 also interferes with apoptosis.</p>", "<title>EBNA1 increases cell survival after DNA damage</title>", "<p>We compared the viability of CNE2 and CNE2E cells after etoposide or UV treatment and found that CNE2E cells had a somewhat higher survival rate than CNE2 cells (particularly after etoposide treatment), despite their reduced ability to repair DNA damage (##FIG##7##Figure 8C##). This is consistent with the known importance of PML NBs in apoptosis ##REF##15077145##[40]## and the observed inhibition of apoptosis by EBNA1. The results suggest that EBNA1 promotes the survival of cells even though they contain DNA damage, which has important implications for tumorigenesis.</p>" ]

[ "<title>Discussion</title>", "<p>We have identified a major effect of EBNA1 expression on host cell PML NBs in NPC, where EBNA1 expression results in pronounced loss of PML NBs and the PML protein itself. This effect has important biological implications due to the strong association between PML disruption and tumor development. While initially identified as a gene whose rearrangement leads to promyelocytic leukemia, it has since been found that loss of the PML protein is associated with cancer development for a variety of human tumors ##REF##14970276##[22]##. In addition, mice lacking PML develop normally but their cells are more prone to malignant transformation ##REF##9488655##[23]##.</p>", "<p>Unlike the results in NPC cells, we have not seen any notable disruption of PML NBs when EBNA1 is expressed in Hela or 293 cells, suggesting that this effect is specific to particular cell backgrounds. Indeed a previous study examined PML NBs in B-cells latently infected with EBV (both latency and I and latency III forms of infection) and found no obvious difference from uninfected cells ##REF##11090180##[41]##. This ability of EBNA1 to disrupt PML NBs in cells of the nasopharnyx could be part of the reason that these cells are particularly susceptible to malignant transformation by EBV.</p>", "<p>We found that EBNA1 is partly associated with PML NBs in a native latent infection in NPC cells and can physically associate with at least one PML isoform that appears to be PML IV. We have also shown that disruption of the PML NBs by EBNA1 is due to loss of multiple isoforms of the PML protein and that this effect is proteasome-dependent. This suggests that EBNA1 is targeted to PML bodies through an interaction with PML IV but, once there, can promote the degradation of all PML isoforms.</p>", "<p>The disruption of PML NBs by EBNA1 requires EBNA1 binding to USP7, a cellular ubiquitin specific protease that is known to associate with PML NBs ##REF##8178435##[42]##. There are several possible scenarios of the role of the EBNA1-USP7 interaction in PML-disruption as depicted in ##FIG##8##Figure 9##. In scenario I, EBNA1 mediates the interaction between a specific PML isoform (ie. PML IV) and USP7 thereby increasing recruitment of USP7 to PML NBs. USP7 could then promote PML degradation either through its catalytic activity or through recruitment of additional cellular proteins. It is not intuitively obvious why deubiquitination would lead to PML destabilization, however approximately three quarters of USP7 is comprised of protein interaction domains, so recruitment of additional cellular enzymes to PML is a viable possibility ##REF##14506283##[15]##, ##REF##16474402##[17]##, ##REF##15749019##[43]##–##REF##16402859##[46]##. In scenarios II and III, the interaction of EBNA1 with PML NBs depends entirely (scenario II) or partly (scenario III) on USP7. In these cases, loss of PML may require the recruitment of additional cellular proteins by EBNA1, since EBNA1 itself is not known to have any enzymatic activities. For example, we have previously shown that EBNA1 forms a stable complex with casein kinase 2 (CK2) ##REF##12783858##[13]## and Scaglioni et al ##REF##16873060##[47]## showed that phosphorylation of PML by CK2 targets PML proteins for ubiquitination and subsequent degradation. Therefore it is possible that increased recruitment of CK2 to PML NBs by EBNA1might give the observed effect.</p>", "<p>While additional experiments are required to distinguish the above scenarios, the involvement of USP7 in PML disruption by EBNA1 has an interesting parallel with studies of PML disruption by herpes simplex virus type 1 (HSV-1). In HSV-1 infection, ICP0 (also called VMW110) plays a major role in disrupting PML NBs by promoting the degradation of PML protein through its action as an ubiquitin ligase ##REF##11704855##[24]##,##REF##11752173##[48]##. Although unrelated to EBNA1 in its sequence, ICP0 also binds tightly to USP7 (albeit through a different USP7 domain than EBNA1 ##REF##14506283##[15]##) and this interaction is required for PML disruption, at least in part because USP7 stabilizes ICP0 by preventing its autoubiquinitation ##REF##9130697##[29]##,##REF##15247261##[49]##. While EBNA1 and ICP0 may utilize USP7 in different ways, it is striking that the only two viral proteins known to bind USP7 both require this interaction to facilitate PML disruption.</p>", "<p>PML NBs have been shown to be important for p53 activation and DNA repair ##REF##10910364##[36]##,##REF##16868026##[39]##, prompting us to investigate whether the degree of disruption of PML NBs by EBNA1 was sufficient to impair these processes. Indeed both the acetylation of p53 and the repair of DNA lesions after UV or etoposide treatment were found to be impaired by EBNA1. EBNA1-expressing cells, however, survived these treatments as well or better than parental cells, due to an inhibition of apoptosis which also requires PML NBs ##REF##11025664##[35]##,##REF##9806545##[50]##. The data as a whole support a model in which EBNA1 contributes to the development of NPC through the disruption of PML NBs, thereby increasing the accumulation of DNA damage while promoting cell survival. This model is consistent with reports of frequent but varying chromosomal aberrations in NPC tumors ##REF##11712805##[51]##,##REF##16423296##[52]##. PML disruption by EBNA1 also provides a mechanistic basis for the observation that EBNA1 expression increases the tumorigenicity of EBV-negative NPC cells ##REF##8958799##[10]##. While several viral proteins are known to promote lytic viral infection through disruption of PML NBs, our results indicate that viral proteins can also contribute to carcinogenesis through PML disruption.</p>" ]

[]

[ "<p>Conceived and designed the experiments: NS LF. Performed the experiments: NS FS. Analyzed the data: NS FS. Contributed reagents/materials/analysis tools: NS. Wrote the paper: LF.</p>", "<p>Latent Epstein-Barr virus (EBV) infection is strongly associated with several cancers, including nasopharyngeal carcinoma (NPC), a tumor that is endemic in several parts of the world. We have investigated the molecular basis for how EBV latent infection promotes the development of NPC. We show that the viral EBNA1 protein, previously known to be required to maintain the EBV episomes, also causes the disruption of the cellular PML (promyelocytic leukemia) nuclear bodies (or ND10s). This disruption occurs both in the context of a native latent infection and when exogenously expressed in EBV-negative NPC cells and involves loss of the PML proteins. We also show that EBNA1 is partially localized to PML nuclear bodies in NPC cells and interacts with a specific PML isoform. PML disruption by EBNA1 requires binding to the cellular ubiquitin specific protease, USP7 or HAUSP, but is independent of p53. We further observed that p53 activation, DNA repair and apoptosis, all of which depend on PML nuclear bodies, were impaired by EBNA1 expression and that cells expressing EBNA1 were more likely to survive after induction of DNA damage. The results point to an important role for EBNA1 in the development of NPC, in which EBNA1-mediated disruption of PML nuclear bodies promotes the survival of cells with DNA damage.</p>", "<title>Author Summary</title>", "<p>Epstein-Barr virus (EBV) infects most people worldwide and is associated with several types of cancer due to its ability to induce cell proliferation. Only one viral protein, EBNA1, is expressed in all forms of EBV-associated tumors. Here, we have investigated whether EBNA1 directly contributes to the development of nasopharyngeal carcinoma (NPC), the most common EBV-associated tumor. We found that EBNA1 disrupts structures in the cell nucleus, called PML bodies, that are known to inhibit malignant transformation and to be important for cells to repair DNA that has been damaged due to exposure to carcinogenic agents. We show that EBNA1 interacts with and degrades the principal component of PML bodies. As a result, cells expressing EBNA1 are less able to repair their DNA and more likely to survive with DNA damage that could result in malignant transformation.</p>" ]

[ "<title>Supporting Information</title>" ]

[ "<p>We thank Drs. Fei-fei Liu and Kwok Lo for NPC cell lines, Dr. Sam Benchimol for p53 antibody and Dionne White for assistance with FACS analysis. We also thank Reagan Ching and Dr. David Bazett-Jones for antibodies, FLAG-PML constructs and helpful discussions throughout the course of this work.</p>" ]

[ "<fig id=\"ppat-1000170-g001\" position=\"float\"><object-id pub-id-type=\"doi\">10.1371/journal.ppat.1000170.g001</object-id><label>Figure 1</label><caption><title>Immunofluorescence imaging of PML NBs in NPC cell lines.</title><p>Log phase cells were fixed and stained for EBNA1 (red) and PML (green). The number of PML foci seen per cell was counted for 100 cells for each sample in three separate experiments and the average number with standard deviation is shown in the histograms, where *** denotes p values less than 0.0001 relative to the parental cell line. Exposure times of image capture were constant for all samples with the same antibody treatment. (A) EBV-positive C666-1 cells before and after treatment with siRNA against GFP (siGFP) or EBNA1 (siEBNA1) are shown. Arrowheads indicate a siEBNA1 treated cell that continued to express EBNA1 and can be used for comparison to neighboring silenced cells. (B) EBV-negative CNE2 and HK1 cell lines with (CNE2E, HK1E) and without stable EBNA1 expression are shown. CNE2E are also shown after silencing of EBNA1 expression where one of the three cells shown continues to express EBNA1 (arrowhead).</p></caption></fig>", "<fig id=\"ppat-1000170-g002\" position=\"float\"><object-id pub-id-type=\"doi\">10.1371/journal.ppat.1000170.g002</object-id><label>Figure 2</label><caption><title>Transient expression of EBNA1 and EBNA1 mutants in CNE2 cells.</title><p>(A) CNE2 cells were transiently transfected with a plasmid expressing EBNA1 or EBNA1 mutants Δ325–376 or Δ395–450, then stained for EBNA1 and PML. Both EBNA1-expressing and nonexpressing cells are shown 48 hrs post transfection. Exposure times of image capture were constant for all samples with the same antibody treatment. (B) Numbers of PML NBs per cell were counted 48 hours after expression of wildtype EBNA1. Cells were categorized into low and high EBNA1 expression depending on the intensity of EBNA1 staining. ** indicates 0.0001&lt;p&lt;0.001 and *** indicates p&lt;0.0001 relative to untransfected cells. (C) The number of PML NBs were counted for all cells in (A) expressing wildtype or mutant EBNA1 proteins.</p></caption></fig>", "<fig id=\"ppat-1000170-g003\" position=\"float\"><object-id pub-id-type=\"doi\">10.1371/journal.ppat.1000170.g003</object-id><label>Figure 3</label><caption><title>EBNA1 expression diminished PML protein levels.</title><p>(A) C666-1 cells were treated with siRNA against EBNA1 (siEBNA1) or GFP (siGFP) then equal amounts of whole cell lysates were Western blotted and probed with an antibody recognizing all PML isoforms, EBNA1 and actin. (B) Equal amounts of whole cell lysates from CNE2 and CNE2E cells were Western blotted and probed as in A. Lysates from CNE2E cells after one (+) or two (++) rounds of transfection with siRNA against EBNA1 are also shown (lanes 3 and 4). (C) Lysates from CNE2 cells 48 hrs after transfection with the indicated amounts of an EBNA1 expression plasmid (OriPE) or the empty plasmid (OriP) were Western blotted for PML, EBNA1, actin or Sp100. (D) CNE2E cells were treated with MG132 proteasomal inhibitor for 0, 8 or 10 hours then equal amounts of lysates were blotted for PML, EBNA1 and actin.</p></caption></fig>", "<fig id=\"ppat-1000170-g004\" position=\"float\"><object-id pub-id-type=\"doi\">10.1371/journal.ppat.1000170.g004</object-id><label>Figure 4</label><caption><title>Interaction of EBNA1 with PML.</title><p>(A) Immunofluorescence images of CNE2 cells transfected with pc3OriPE and of C666-1 cells are shown after staining for EBNA1 and PML. The transfected CNE2 cells shown are those expressing very low levels of EBNA1. (B) EBNA1 was immunoprecipitated from C666-1 cells with anti-EBNA1 antibody (IP:EBNA1). The starting lysate (Input) and protein remaining after IP (Post IP) are also shown, in each case representing 1/40^th of the lysate used in IP. The same lysate was also treated with IgG beads as a negative control (IP:IgG). All samples were Western blotted using antibodies against EBNA1 or all PML isoforms. In the right panel, the positions of FLAG-tagged PML isoform I (FLAG-PML I) and PML isoform IV (FLAG-PML IV) expressed in C666-1 cells are shown by Western blotting with anti-FLAG antibody.</p></caption></fig>", "<fig id=\"ppat-1000170-g005\" position=\"float\"><object-id pub-id-type=\"doi\">10.1371/journal.ppat.1000170.g005</object-id><label>Figure 5</label><caption><title>Effect of USP7 silencing on PML degradation by EBNA1.</title><p>(A) CNE2E cells expressing EBNA1 were transfected with siRNA against USP7 or GFP (negative control) then stained for USP7 and PML. Exposure times of image capture were constant for all samples with the same antibody treatment. (B) Equal amounts of cell lysates from (A) were analysed by Western blotting with the indicated antibodies. siUSP7-1 and siUSP7-2 are duplicate samples treated with siRNA against USP7. (C) CNE2 cells were transfected with siRNA against GFP or USP7 (left panel) then were transfected with EBNA1 expression plasmid pc3OripE and stained for EBNA1 and PML (right panel). Exposure times of image capture were constant for all samples with the same antibody treatment. (D) Equal amounts of cell lysates from (C) were analysed by Western blotting after pretreatment with siGFP or siUSP7 followed by EBNA1 expression.</p></caption></fig>", "<fig id=\"ppat-1000170-g006\" position=\"float\"><object-id pub-id-type=\"doi\">10.1371/journal.ppat.1000170.g006</object-id><label>Figure 6</label><caption><title>EBNA1-induced disruption of PML NBs in Saos-2 cells.</title><p>p53-null Saos-2 cells were transiently transfected with the expression plasmid with or without the EBNA1 gene as in ##FIG##1##Figure 2##. (A) Cells were stained for EBNA1, PML and DNA (DAPI) and visualized by fluorescence microscopy. The number of PML NBs per cell were counted and average numbers with standard deviations are shown in the histogram, where *** indicates p&lt;0.0001. (B) Equal amounts of lysates from the transfected cells were analysed by Western blotting as in ##FIG##2##Figure 3##.</p></caption></fig>", "<fig id=\"ppat-1000170-g007\" position=\"float\"><object-id pub-id-type=\"doi\">10.1371/journal.ppat.1000170.g007</object-id><label>Figure 7</label><caption><title>Effects of EBNA1 on p53 activation.</title><p>(A) CNE2 and CNE2E cells were treated with etoposide (+) or left untreated (−) and equal amounts of total cell lysates were analysed by SDS-PAGE and Western blotting for p53 acetylated on K382 and total p53. Actin loading controls are also shown. (B) Hela cells were transfected with a plasmid lacking (oriP) or expressing (oriPE) EBNA1 then were treated with etoposide (+) or left untreated (−). Equal amounts of cell lysate were then analysed by Western blotting as in A.</p></caption></fig>", "<fig id=\"ppat-1000170-g008\" position=\"float\"><object-id pub-id-type=\"doi\">10.1371/journal.ppat.1000170.g008</object-id><label>Figure 8</label><caption><title>Effects of EBNA1 on DNA repair, apoptosis and cell survival.</title><p>(A) CNE2 and CNE2E cells, before and after siRNA treatment for GFP (siGFP; negative control) or EBNA1 (siEBNA1), were treated with UV or etoposide or left untreated. 24 hrs later cells were fixed, stained with propidium idodide and analysed for DNA content by FACS. The percentage of cells in each cell cycle stage was determined using Modfit and is shown for each sample. (B) CNE2 and CNE2E cells were treated with etoposide then analysed by TUNEL assay. Average percentage of TUNEL-positive cells are shown from three experiments with standard deviation and 0.0001&lt;p&lt;0.001 (**). (C) CNE2 (grey) and CNE2E (black) cells were treated with etoposide (top graph) or UV (bottom graph) then grown for the indicated number of days. At each time point, cells were incubated with Trypan blue and the percentage of cells that excluded Trypan blue was determined. Experiments were performed in triplicate and average numbers with standard deviations are shown. The difference in cell survival with and without EBNA1 3 days post etoposide treatment is statistically significant with a p value of 0.05.</p></caption></fig>", "<fig id=\"ppat-1000170-g009\" position=\"float\"><object-id pub-id-type=\"doi\">10.1371/journal.ppat.1000170.g009</object-id><label>Figure 9</label><caption><title>Models of the EBNA1, USP7 and PML interactions.</title><p>Three possible interpretations of the data on EBNA1-PML and EBNA1-USP7 interactions at PML nuclear bodies are shown.</p></caption></fig>" ]

[]

[]

[]

[]

[]

[]

[ "<supplementary-material content-type=\"local-data\" id=\"ppat.1000170.s001\"><label>Figure S1</label><caption><p>Quantification of EBNA1 levels in NPC cell lines. (A) C666-1, HK1E, CNE2E and CNE2 cells were lysed in 9 M urea, 10mM Tris pH6.8 and 100 µg of protein from each sample was analysed by Western blotting using antibodies against EBNA1 and actin and the ECL Plus system (Perkin Elmer). Positions of the full length EBNA1 (EBNA FL) expressed endogenously in C666-1 and recombinant EBNA1 lacking most of the Gly-Ala repeat region (EBNA1ΔGA) expressed in HK1E and CNE2E are indicated. (B) Flourescent signals from each band were quantified on a Typhoon Imaging scanner using ImageQuant 5.0 software. EBNA1 levels were normalized to the actin loading control and are shown relative to the C666-1 value.</p><p>(0.20 MB TIF)</p></caption></supplementary-material>", "<supplementary-material content-type=\"local-data\" id=\"ppat.1000170.s002\"><label>Figure S2</label><caption><p>Quantification of PML mRNA levels in NPC cell lines. Total RNA from C666-1, CNE2E and CNE2 cells was harvested using RNeasy mini kit (Qiagen). RNA quality was assessed by confirmation of intact 28S and 18S ribosomal bands following agarose gel electrophoresis and ethidium bromide staining. cDNA was synthesized using 1 µg total RNA and First Strand cDNA sythesis kit (Fermentas). Quantitative real-time PCR was performed with 1/5 to 1/20 of the cDNA template and Platinum SYBR Green qPCR SuperMix-UDG (Invitrogen) in a Rotorgene qPCR System (Corbett Research). The primer pairs used to amplify PML mRNA were <named-content content-type=\"gene\">CGGAGGAGGAGTTCCAGTTT</named-content> and <named-content content-type=\"gene\">CCACAATCTGCCGGTACAC</named-content>. β-actin mRNA was amplified from the same samples using the primers <named-content content-type=\"gene\">CATGTACGTTGCTATCCAGGC</named-content> and <named-content content-type=\"gene\">CTCCTTAATGTCACGCACGAT</named-content>. PML mRNA levels are shown realtive to β-actin mRNA.</p><p>(0.07 MB TIF)</p></caption></supplementary-material>", "<supplementary-material content-type=\"local-data\" id=\"ppat.1000170.s003\"><label>Figure S3</label><caption><p>Interactions of EBNA1 with overexpressed FLAG-tagged PML I and PML IV. CNE2 cells were co-transfected with pc3oriPE (expressing EBNA1) and plasmids expressing either FLAG-PML I, FLAG-PML IV or FLAG-tagged lacZ (negative control). 16 hours later, IPs were performed using anti-FLAG M2 agarose beads. Recovered proteins were immunoblotted for FLAG and EBNA1. Input samples contain 1/50th the amount of lysate used in the FLAG IPs.</p><p>(0.18 MB TIF)</p></caption></supplementary-material>" ]

[ "<fn-group><fn fn-type=\"COI-statement\"><p>The authors have declared that no competing interests exist.</p></fn><fn fn-type=\"financial-disclosure\"><p>This work was funded by the Canadian Cancer Society through an operating grant to L.F. from the National Cancer Institute of Canada (NCIC). F.S. and N.S. were supported by studentships from the NCIC and the Natural Sciences and Engineering Council of Canada, respectively. L.F. is a tier 1 Canada Research Chair in Molecular Virology.</p></fn></fn-group>" ]

[ "<graphic xlink:href=\"ppat.1000170.g001\"/>", "<graphic xlink:href=\"ppat.1000170.g002\"/>", "<graphic xlink:href=\"ppat.1000170.g003\"/>", "<graphic xlink:href=\"ppat.1000170.g004\"/>", "<graphic xlink:href=\"ppat.1000170.g005\"/>", "<graphic xlink:href=\"ppat.1000170.g006\"/>", "<graphic xlink:href=\"ppat.1000170.g007\"/>", "<graphic xlink:href=\"ppat.1000170.g008\"/>", "<graphic xlink:href=\"ppat.1000170.g009\"/>" ]

[ "<media xlink:href=\"ppat.1000170.s001.tif\"><caption><p>Click here for additional data file.</p></caption></media>", "<media xlink:href=\"ppat.1000170.s002.tif\"><caption><p>Click here for additional data file.</p></caption></media>", "<media xlink:href=\"ppat.1000170.s003.tif\"><caption><p>Click here for additional data file.</p></caption></media>" ]

{ "acronym": [], "definition": [] }

[ "<title>Introduction</title>", "<p>The human immunodeficiency virus (HIV-1) exterior envelope glycoprotein, gp120, and the transmembrane glycoprotein, gp41, are non-covalently associated to comprise the trimeric, functional viral spike. These glycoproteins mediate entry and represent the sole virally encoded targets for neutralizing antibodies (nAbs) on the surface of the virus. The HIV-1 envelope glycoproteins, and those from related immunodeficiency viruses, are somewhat unusual in that they mediate target-to-membrane fusion by receptor-triggered conformational changes rather than by low pH-mediated fusion events typified by the influenza virus type 1 viral membrane protein, hemagglutinin (HA) ##REF##12671653##[1]##. The interaction of gp120 with the primary receptor, CD4, induces formation or exposure of a bridging sheet mini-domain that is, along with elements of the gp120 third variable region (V3), involved with binding to the co-receptor, CCR5 ##REF##8906795##[2]##,##REF##8906796##[3]##,##REF##17901336##[4]##.</p>", "<p>As was previously shown, antibodies against this induced co-receptor binding site are abundantly generated during natural HIV infection ##REF##15867093##[5]## and may be in part elicited due to the unique ability of gp120 to undergo receptor-induced conformations required for the sequential entry process. The co-receptor site antibodies are termed CD4-induced (CD4i) because following CD4 binding to gp120 (which functionally induces the co-receptor binding), these antibodies bind with enhanced affinity to gp120. The prototype for the co-receptor-directed, CD4i antibodies is 17b. However, it is less well appreciated that several full-length gp120 proteins actually are recognized by CD4i antibodies like 17b with high affinity (or avidity) even in the absence of the primary receptor ##REF##11327825##[6]##. The co-receptor-directed antibodies do not generally neutralize most circulating isolates ##REF##12970440##[7]##. However, these antibodies have attracted considerable interest due to the remarkable post-translational sulfation of a subset of these antibodies that mimics the functionally important sulfation of the CCR5 co-receptor N-terminus and their selective VH gene usage ##REF##10089882##[8]##,##REF##15220422##[9]##. Viral evasion of the CD4i antibodies likely occurs due to the in vivo selection for viruses that occlude or do not form this highly conserved region until the virus interacts with the primary receptor, CD4 ##REF##12970440##[7]##,##REF##11160708##[10]##. Once formed, the conserved site interacts with the largely invariant HIV co-receptor, CCR5. In contrast to the ability of affinity-matured CD4i antibodies, which can recognize the co-receptor site in the absence of CD4 with high functional affinity, the requirements for the naïve B cell receptor to recognize the same site is not presently understood and may differ from that of a mature CD4i antibody. Therefore, one aim of this study was to determine if previously described soluble envelope glycoprotein trimeric immunogens ##REF##11932429##[11]## might elicit CD4i antibodies in primates that possess a CD4 that is capable of a high-affinity interaction with the viral spike.</p>", "<p>As an immunogen, monomeric gp120 does not elicit broadly nAbs ##REF##8568294##[12]## and has failed as a vaccine in a large clinical trial ##REF##17109337##[13]##. Therefore, much of the field has moved toward the design of soluble trimeric Env immunogens that more closely mimic the functional spike ##REF##11932429##[11]##,##REF##10623724##[14]##,##REF##12163607##[15]##,##REF##10823881##[16]##,##REF##15665645##[17]##,##REF##14512572##[18]##. The gp140 trimers which we have studied are derived from a neutralization resistant primary isolate, YU2, and are stabilized by heterologous trimerization domains (foldon) and somewhat improve the elicitation of neutralizing antibodies when inoculated into small animals possessing CD4 molecules that do not interact with gp120 ##REF##11152489##[19]##,##REF##16415019##[20]##. However, these stabilized trimeric immunogens have not been extensively tested in primates, which possess CD4 molecules capable of high affinity interaction with HIV-1 Env. Here, we demonstrate directly that the elicitation of CD4i antibodies by Env trimers is dependent upon the in vivo presence of high-affinity CD4 found in primates, but not present in the wild-type (WT) rabbits. We definitively demonstrate that the presence of endogenous primate CD4 is sufficient to generate CD4i antibodies following inoculation of these same trimers into rabbits rendered transgenic for human CD4. Consistent with these data, we also show the presence of co-receptor-directed antibodies in sera from a subset of patients who participated in the non-efficacious VaxGen phase III clinical trial using monomeric gp120 as a candidate vaccine. Our findings provide clear evidence that binding of a type-I viral membrane protein to its primary receptor can lead to its in vivo altered immunogenicity. It also illustrates that, in contrast to antibodies that have the undergone affinity maturation, the naïve B cell receptor repertoire does not recognize the co-receptor binding site with sufficient affinity to elicit antibodies against this region in the absence of primate CD4.</p>" ]

[ "<title>Materials and Methods</title>", "<title>Protein production and purification</title>", "<p>Proteins were produced by transient transfection of adherent 293F cells or 293Freestyle suspension cells. The highly glycosylated and His-tag containing YU2gp140-F trimers were captured and purified from the serum-free media in a three-step process. First, the protein was captured via glycans over with lentil-lectin affinity chromatography (GE Healthcare, Uppsala, Sweden). After extensive washing with PBS the protein was eluted and captured in the second step via the His-tag by nickel-chelation chromatography. (GE Healthcare) then washed and eluted with a 300 mM Imidazole containing PBS buffer. In some cases the YU2gp140-F trimers were separated from lower molecular weight forms by the third step of gel filtration chromatography using a superdex200 26/60 prep grade column by the ÄKTA Fast protein liquid chromatography system (GE Healthcare). In contrast, the YU2gp120 and HXBc2 gp120 core proteins were purified by capturing the molecules on an IgG17b affinity column. After extensive washing with PBS, the proteins were eluted from the column with 100 mM glycine/Tris HCl/150 mM NaCl. pH 2.8 and immediately neutralized with Tris base, pH 8.5.</p>", "<title>CD4 solution competition assay</title>", "<p>Env protein was co-incubated at concentrations of 0.4 to 46 nM in PBS with 2, 4 or 9 nM sCD4 at room temperature for 1 h, allowing for CD4-Env trimer complexes to form. Non-Env bound sCD4 was captured on a plate pre-adsorbed with 200 ng/well of the anti-CD4 antibody RPA-T4 (Ebioscience, San Diego, CA). RPA-T4 binds to domain 1 of CD4 and competes with HIV-1 gp120 for binding. RPA-T4 bound sCD4 was probed with a biotin-conjugated, non-competitive anti-CD4 antibody, OKT-4 (Ebioscience). Horseradish peroxidase (HRP) conjugated streptavidin (Sigma) followed by the colorimetric peroxide enzyme immunoassay substrate (3,3′,5,5′-tetramethylbenzidine; Bio-Rad) was added to induce a colorimetric change and the reaction was stopped by adding 1 M H₂SO₄. OD was read at 450 nm.</p>", "<title>Enzyme-linked immunosorbent assay (ELISA)</title>", "<p>High-protein-binding MaxiSorp plates (Nunc) plates were coated with 200 ng/well of mAb 17b in 100 µl of PBS at 4°C overnight after which the wells were blocked for 2 h at room temperature (RT) with PBS-2% fat-free milk. The gp140-F trimers, at concentrations between 3.2 ng/ml to 10 µg/ml, were pre-incubated with 20 µg/ml sCD4 for 1 h at RT and then added to the wells. The wells were then probed for 17b bound gp140-F trimer with rabbit anti-gp140-F polyclonal sera. Addition of HRP conjugated anti-rabbit-Ig (Fc region) (Jackson Laboratories, Bar Harbor, MN) followed by the colorimetric peroxide enzyme immunoassay substrate (3,3′,5,5′-tetramethylbenzidine; Bio-Rad) was used to induce a colorimetric change and the reaction was stopped by adding 1 M H₂SO₄. OD was read at 450 nm. The 17b binding competition assay was performed by coating the ELISA plate with 200 ng/well of lectin from <italic>Galanthus nivalis</italic> (Sigma), followed by 200 ng/well of HXBc2 core protein. After blocking with 2% fat-free milk, serum was incubated for 45 minutes at dilutions between 25 and 6400 in a total volume of 100 ul after which 25 ul of biotin conjugated 17b antibody was added to a final dilution of 2500 and incubated for an additional 45 minutes. The plate was probed with HRP conjugated streptavidin and developed as above.</p>", "<title>SPR kinetic analysis</title>", "<p>To determine the kinetic constants of YU2gp140-F interaction with 17b IgG, we performed binding analysis were in two different formats on a Biacore3000 surface plasmon resonance spectrometer. In one format (##FIG##2##Fig 3A and B##), YU2gp140-F (without or with pre-binding to 20-fold molar excess of D1D2 sCD4 for 1 h) was passed over the 17b IgG surface. Because of the potential oligomeric interaction of the trimers with 17b IgG presented on the surface of the chip, this analysis likely measured avidity rather than simple affinity. However, since the curves approximated single binding kinetics and we did not know how many monomeric subunits within the trimers were capable of 17b interaction, curve fitting was done assuming a 1∶1 interaction. In the reverse format (##FIG##2##Fig. 3C##), the 17b IgG was passed over YU2gp140-F surface and functional (or apparent) affinity was calculated using the bivalent analyte program (Biacore) that derived affinity from the potential bivalent interaction of the 17b with trimers immobilized on the chip surface.</p>", "<p>To prepare binding surfaces, ligands (7 µg/ml in 10 mM NaOAc, pH 5.5 buffer) were immobilized on CM5 chip by the amine coupling method following manufacturer's protocol. One flow cell receiving only NaOAc buffer was used as reference control for correction of background binding. For binding experiments, analytes were serially diluted at concentrations ranging from 4.6 nM to 600 nM in the HEPES-EP reaction buffer. To determine the rate of association, each analyte was allowed to flow over the activated surfaces at a rate of 30 µl/min for 3 minutes. Dissociation was determined by washing off bound analyte for the next 5 min. Likely due to avidity of the oligomeric analytes, especially in ##FIG##2##Fig 3A##, the rate of dissociation was difficult to determine and likely represents and over estimate of the actual 1∶1 binding kinetics. The surface was regenerated by removing any unbound analyte with two injections (60 sec each) of 10 mM Glycine, pH 3.0. All procedures were performed at RT.</p>", "<title>Flow cytometry</title>", "<p>Monkey, human and rabbit PBMCs were analyzed by flow cytometry using a modified LSR I system (BD Biosciences). Data analysis was performed using FlowJo software (Tree Star, San Carlos, CA). Staining and gating strategies to detect YU2 trimer binding to live, CD3⁺/CD4⁺/CD8⁻ cells from primates is described in ##SUPPL##1##Fig. S2##. Staining and gating strategy for trimer binding to rabbit cells is described in ##SUPPL##3##Fig. S4##. The mAb 447-52D was a kind gift from Susan Zolla-Pazner (New York University School of Medicine).</p>", "<title>Animals</title>", "<p>Five female cynomolgus macaques (<italic>Macaca fascicularis</italic>) of Chinese origin, 5–6 years old, were housed in the Astrid Fagraeus laboratory at the Swedish Institute for Infectious Disease Control. Housing and care procedures were in compliance with the provisions and general guidelines of the Swedish Animal Welfare Agency. The animals were housed in pairs in 4 m³ cages, enriched to give them possibility to express their physiological and behavioural needs. They were habituated to the housing conditions for more than 6 weeks before the start of the experiment, and subjected to positive reinforcement training in order to reduce the stress associated with experimental procedures. All immunizations and blood sampling were performed under sedation with ketamine 10 mg/kg intramuscularly (i.m.). (Ketaminol 100 mg/ml, Intervet, Sweden) The macaques were weighed and examined for swelling of lymphnodes and spleen at each immunization or sampling occasion. Before entering the study, all monkeys were confirmed negative for simian immunodeficiency virus (SIV), simian T-cell lymphotropic virus and simian retrovirus type D. Female New Zealand White Rabbits and male huCD4 New Zealand White transgenic rabbits were housed at BioCon, Inc (Rockville, MD) or at an animal facility at the National Institutes of Health according to current regulations. Cynomolgus macaques were injected once with 200 ug followed by three injections with 100 µg YU2gp140-F trimer. Rabbits were injected four times with 50 ug YU2 trimer. All proteins were formulated in the GSK-AS01B adjuvant system (GlaxoSmithKline, Rixensart, Belgium) prior to injection unless otherwise stated and all injections were administered i.m. at an interval of one month. Sera were collected before the first injection as well as two weeks after each injection. All procedures were approved by the Local Ethical Committee on Animal Experiments.</p>", "<title>Neutralization Assays</title>", "<p>Analysis for HIV-1 and HIV-2 neutralization in serum samples were performed as previously described ##REF##17721546##[25]##,##REF##17113201##[26]##. Briefly, Env pseudoviruses were prepared by co-transfecting 293T cells with an Env expression plasmid containing a full gp160 env gene and an <italic>env</italic>-deficient HIV-1 backbone vector (pSG3 Env). For screening, a single dilution of sera or plasma was used and the percent neutralization was calculated compared to controls with no sera added. To determine the dilution of the sera that resulted in a 50% reduction in RLU against selected viruses, serial dilution assays were performed and the neutralization dose-response curves were fit by non-linear regression using a 4-paremeter hill slope equation programmed into JMP statistical software (JMP 5.1, SAS Institute Inc., Cary, NC). The results are reported as the serum neutralization ID₅₀, which is the reciprocal value of the serum dilution resulting in a 50% reduction in viral entry. Dana Gabuzda (Dana Farber Cancer Institute) provided the Env plasmid for YU2. Env plasmids for SF162 and JRFL were provided by Leonidas Stamatatos (Seattle Biomedical Research Institute) and James Binley (Torrey Pines Institute), respectively. The Clade A Env-pseudovirus DJ263.8 was cloned from the original PBMC derived virus provided by Francine McCutchan and Vicky Polonis (U.S. Military HIV Research Program). Env plasmids BaL.01 was recently described by our laboratory ##REF##17113201##[26]## and the Env used to generate the pseudovirus SS1196.1 was previously described ##REF##16051804##[39]##. The HIV-2 Env-pseudovirus 7312 containing the V343M mutation have previously been described ##REF##15867093##[5]##. The remaining functional Env plasmids were obtained from the NIH ARRRP.</p>", "<title>Human serum samples</title>", "<p>Twenty randomly chosen serum samples were obtained via an MTA with the Global Solutions for Infectious Diseases. These sera were derived from volunteers from the VaxGen Inc phase III clinical trial. At the time of sampling (month 12.5) the participants had received four injections (month 0, 1, 6 and 12) with the AIDSVAX B/B vaccine containing 300 ug each of recombinant HIV-1_MN or HIV-1_GNE8 derived gp120 in Alum adjuvant ##REF##17109337##[13]##.</p>" ]

[ "<title>Results</title>", "<title>Purified soluble Env trimers bind soluble and cell-surface CD4</title>", "<p>The highly glycosylated gp140-F trimers derived from the primary isolate YU2; previously referred to as YU2gp140(-)/FT ##REF##11932429##[11]## were purified from the supernatant of transiently transfected mammalian cells by lentil lectin affinity chromatography followed by chelation chromatography. In most cases, size exclusion chromatography was used to isolate the predominant trimeric peak fraction (##SUPPL##0##Fig. S1##). To confirm binding of the trimers to sCD4 independent of avidity effects, a solution-based binding assay was developed. To begin, 1 to 137 nM of the gp140-F molecules were co-incubated with 2, 4 or 9 nM soluble human 4-domain CD4 (sCD4) in solution. Next, non-Env-bound sCD4 was captured by RPA-T4 and detected in an ELISA format to evaluate the relative binding (##FIG##0##Fig. 1A##). The gp140-F trimers bind to sCD4 in a concentration dependent manner with half-maximal binding at approximately 7, 13 and 26 nM respectively. To confirm the specificity of the binding, we introduced a mutation at position 368 of gp140 such that 368 Asp was changed to Arg. This mutation (368D/R) was previously shown to specifically reduce or abrogate CD4 binding of monomeric gp120 ##REF##2243375##[21]## and as expected in this soluble CD4 reporter assay, the gp140 368D/R trimers did not bind sCD4 at any concentration tested. Since <italic>in vivo</italic>, abundant cell-surface CD4 is a potential source for high-affinity binding of Env, we sought to confirm that the YU2 trimers could bind to CD4-positive cells derived from non-human primates before initiating immunogenicity studies. We co-incubated primate peripheral blood mononuclear cells (PBMCs) with 2 µg/ml, 10 µg/ml or 20 µg/ml gp140-F trimers and stained the cells for CD3, CD4, CD8 and a marker for dead cells. Trimer binding to cell-surface-expressed CD4 was detected with the V3-directed antibody 447-52D on live, CD3⁺/CD4⁺/CD8⁻ cells by flow cytometry (##FIG##0##Fig. 1B##; for staining and gating strategy, see ##SUPPL##1##Fig. S2##). Similar to the results obtained in the CD4 solution assay, the gp140-F trimers bound to the CD4⁺ T cells in a concentration- dependent manner. Trimer binding to the CD4⁺ cells could be fully abrogated by pre-incubation of 20 µg/ml of the gp140-F molecules with an excess of sCD4. Further, no binding could be detected after incubation of the PBMCs with the gp140 368D/R CD4 binding-defective trimers. Together, the data confirmed that the YU2 gp140-F trimers used in this study bind both soluble, and importantly, cell-surface CD4 in a dose-dependent and specific manner.</p>", "<title>Co-receptor binding site directed antibody recognition of the gp140-F trimers</title>", "<p>To confirm that the highly purified trimers used in this study were competent for recognition by 17b, as well as competent for induction of the CD4i epitope by CD4, we performed both ELISA-based and surface plasmon resonance (SPR) binding assays. First, we incubated 3.2 ng/ml to 10 µg/ml of the YU2 gp140-F trimers in solution, without or with an excess of sCD4, after which gp140-F was captured by 17b on a plate (##FIG##1##Fig. 2A##). Consistent with previous data with monomeric YU2 gp120 ##REF##11327825##[6]##,##REF##7685405##[22]##,##REF##17360741##[23]##, the trimers were well recognized by 17b in the absence of CD4. However, under the conditions of this assay, the relative binding increases approximately 2 to 5-fold in the presence of 1 ug/ml or 20 ug/ml sCD4, confirming that sCD4 induces better exposure, by formation or stabilization, of the CD4i site on the gp120 moieties present in the soluble trimeric context.</p>", "<p>We next determined the recognition of the trimers by the protypic CD4i antibody 17b by Biacore SPR in two formats (##FIG##2##Fig 3##). In the first format, the gp140-F trimers were flowed over 17b immobilized on the surface of chip as the analyte. Because the trimers are oligomeric, this binding analysis detects avidity rather than strict affinity. However, this would be the case if the trimers were in solution and recognized by the bivalent BCR in multi-valent array on the surface of a B cell or if the trimers were displayed on the surface of a CD4⁺ lymphocyte. By this means, we determined that the avidity of the trimers for 17b was nanomolar to subnanomolar regardless if the trimers were in complex or not with CD4 (see ##FIG##2##Fig 3A##). To better approximate the actual affinity of interaction, the 17b antibody was flowed over the gp140-F trimers immobilized on the chip and the binding was analyzed by bivalent curve fitting. This analysis also confirmed that recognition of the trimers by 17b in the absence of CD4 was a high-affinity interaction in the low nanomolar range (##FIG##2##Fig 3B##).</p>", "<title>Env trimer immunogenicity in monkeys and in rabbits</title>", "<p>To assess if <italic>in vivo</italic> interaction of primate CD4 with HIV-1 Env is a requirement for elicitation of CD4i antibodies, we utilized the previously published observations that WT rabbit CD4 is unable to bind HIV-1 Env ##REF##7598907##[24]##. We immunized cynomolgus macaques and rabbits four times with the YU2 gp140-F trimers formulated in the GSK Adjuvant System AS01B and confirmed that the ELISA titers saturated by three inoculations and were roughly equivalent (data not shown). For neutralization, we first analyzed the serum samples from individual animals for their ability to inhibit viral entry against a panel of selected HIV-1 isolates. The rationale for this analysis was to assess the over-all neutralization capacity of responses elicited in monkeys versus the rabbits against HIV-1 to make a comparison of other responses more meaningful (i.e. CD4i-directed HIV-2 neutralizing antibodies, see below). The sera were analyzed at a 1 to 5 dilution against a panel of nine HIV-1 Env pseudotyped viruses in a standardized neutralization assay using TZM-bl target cells ##REF##17721546##[25]##,##REF##17113201##[26]## (##FIG##3##Fig. 4A##). Sera derived from animals of both species potently neutralized the three sensitive viruses, the lab-adapted HxBc2 (clade B), SF162 (clade B) and MW.965 (clade C) with values between 90 and 100%. Overall, the potency and breadth of neutralization for this panel of viruses were very similar between the monkey and rabbit sera. Subtle cross-species differences in neutralization were observed, but these were not statistically significant. For example, sera from the immunized rabbits tended to display more consistent animal-to-animal neutralization capacity against the homologous YU2 strain. In contrast, the sera derived from the monkeys displayed a trend of greater potency against BaL and the tier 2 isolate SS1196 (clade B). Sera derived from both species of animals inoculated with the YU2 trimers sporadically neutralized the DJ293 isolate (clade A), but poorly neutralized JRFL (clade B), as well as TRO1.1 (clade B; not shown, done only with monkey sera), demarking the limits of the neutralization activity elicited by the current immunogen design. Perhaps the subtle differences observed in the neutralization potency against some of the isolates between sera derived from the monkeys versus the rabbits is due to slight differentials in the elicited antibody repertoires, however, in general, the data highlights the overall similarities in the elicited neutralization capacity.</p>", "<p>To detect if there is a species-difference in specific antibody elicitation against the co-receptor site of gp140-F trimers, we analyzed the sera in the same assay format as above but against virus pseudotyped with Env from an HIV-2 isolate, 7312 (containing a V434M amino acid change). While this virus is relatively insensitive to antibodies raised against HIV-1 Env it becomes highly sensitive to anti-HIV-1 CD4i-antibodies in the presence of sub-inhibitory concentrations of sCD4, facilitating the specific detection of such antibodies ##REF##15867093##[5]##. Consistent with data from HIV-1 infected individuals ##REF##15867093##[5]##,##REF##17409164##[27]## and gp140-inoculated humans (GMS, unpublished observations), CD4i-antibodies detected by this assay were abundant in sera from all five monkeys (ID₅₀ titers: 55, 91, 268, 479 and 2618; ##FIG##3##Fig 4B##). We could detect low-level cross-neutralization of HIV-2_7312/V434M in sera from four of the five monkeys, consistent with what had been observed previously from some HIV-1 infected humans ##REF##15867093##[5]##. Following these results, we analyzed three cynomolgus macaques that had been inoculated with the YU2 gp140-F trimers in Ribi adjuvant 2 times and at a similar dose, and detected CD4i antibodies in these animals with ID 50 values of 21, 27 and 198 (##SUPPL##2##Fig S3##). The lower levels of CD4i antibodies relative to those elicited by trimers formulated in AS01B, correlated with similarly lower potency of neutralization against the HIV-1 isolate, MN (##SUPPL##2##Fig S3##). We also detected CD4i antibodies in 5 out of 5 cynomolgus monkeys primed with Semliki Forest virus (SFV) particles expressing the YU2 gp140-F trimers and boosted with trimeric protein with ID₅₀ values of 57, 1619, 44 and 335 and in 3 out of 3 baboons immunized with YU2 gp140 molecules rendered trimeric with a modified GCN4 motif ##REF##10823881##[16]## in Ribi adjuvant (data not shown). Taken together, these data clearly demonstrate that elicitation of CD4i antibodies by Env trimers in non-human primates is a highly reproducible and commonly elicited response and can occur at low levels when Env is expressed from a viral vector (SFV; not shown).</p>", "<p>In stark contrast, CD4i-antibodies could not be detected in the serum from any of the WT rabbits (##FIG##3##Fig 4B##), suggesting that the elicitation of CD4i-antibodies is dependent on the <italic>in-vivo</italic> presence of CD4 with affinity for the YU2 trimers in the non-human primates. Alternatively, it might be that rabbits lack B cell receptors (BCR) in their naïve repertoire with the ability to recognize the HIV-1 co-receptor binding site while the monkeys possess such a capacity.</p>", "<p>While the HIV-2 assay detects antibodies specific for the co-receptor binding site, we wanted to confirm these results by performing an ELISA based assay where serum from immunized animals were tested for their ability to compete with a biotinylated 17b antibody for binding to gp120. It is known that antibodies not directly directed against the co-receptor site are capable of competing with 17b for binding to gp120 ##REF##8627711##[28]##. To minimize such indirect effects, we analyzed the ability of sera to compete for biotinylated 17b binding to an HXBc2 gp120 core protein capable of binding 17b with high affinity (Dey et al, manuscript in preparation). In this assay format serum samples from monkeys were able to inhibit 17b binding at an approximately 10-fold higher dilution than that of serum samples from the rabbits (##FIG##3##Fig 4C##). The weak, but detectable, level of antibodies capable of competing with 17b in serum from rabbit serum may be due to antibodies recognizing the co-receptor binding elements in the pre-CD4 induced state or other antibodies that can inhibit 17b binding, such as CD4 binding site antibodies ##REF##8627711##[28]##.</p>", "<p>Direct interaction of the trimers with primate CD4 might be expected to expose as well V3, the other element of gp120 involved in co-receptor interaction ##REF##17901336##[4]##. To address this issue, we performed a binding assay to determine the proportion of V3-specific antibodies compared to antibodies against intact gp120 in sera derived from the 3 types of test animals. We observed a slight 1.6-fold average higher proportion of antibodies against V3 in serum samples from primates than in WT rabbits that was not statistically significant (##SUPPL##2##Fig S3##). The huCD4 rabbits, although possessing lower gp120 and V3 binding titers compared to WT animals, also displayed a slightly greater proportion of V3-specific antibodies. The lack of a significant increase in V3-directed antibodies may be due to maximal exposure of V3 on gp140 as we could see no enhancement of binding of a V3-specific antibody to the trimers following addition of CD4 (not shown) consistent with a previous study ##REF##8627711##[28]##.</p>", "<title>Immunization of human-CD4-transgenic rabbits with trimers</title>", "<p>To address if the presence of primate CD4 was required for the elicitation of CD4i antibodies from the gp140-F trimeric immunogens, we used rabbits previously engineered to be transgenic for human CD4 (huCD4) ##REF##7598907##[24]##. These rabbits were generated from the New Zealand White (NZW) background, and are relatively similar in their genetic background to the out-bred NZW WT rabbits used for the initial immunogenicty analysis above. The huCD4 transgenic rabbits allowed us to perform controlled immunogenicity experiments to determine if the <italic>in vivo</italic> presence of primate CD4 allows for BCR recognition of the YU2 gp140-F co-receptor site and subsequent elicitation of CD4-induced antibodies in rabbits. Before initiating the immunogenicity experiment, we confirmed that the huCD4 transgenic animals, ranging from 2 to 5 years of age, still expressed human CD4 on their PBMCs as follows. Incubation of 20 µg/ml gp140-F trimers with PBMCs from WT and huCD4 rabbits and analysis by flow cytometry using species-specific cellular makers (for staining and gating strategy, see ##SUPPL##4##Fig. S5##) confirmed that only PBMCs from the transgenic animals can bind the gp140-F, and that binding occurred only on cells co-expressing human and rabbit CD4 (rCD4; ##FIG##4##Fig. 5A##). These results are consistent with the initial design of the transgenic rabbits to restrict expression of huCD4 only to cells also co-expressing rCD4 by use of a cell-type-specific promoter ##REF##7598907##[24]##.</p>", "<p>Five huCD4 rabbits were inoculated with the YU2 gp140-F trimers formulated in AS01B adjuvant by an identical regimen as the WT rabbits and monitored for the appearance of CD4-induced antibodies by the <italic>in vitro</italic> HIV-2 assay. CD4-induced antibodies could be detected in the sera from four out of five huCD4 rabbits after three immunizations with gp140-F trimers (##FIG##4##Fig 5B##). The ID₅₀ titers detected were 84, 127, 190 and 4507, which is comparable with the levels detected in serum samples derived from immunized monkeys. These data demonstrate that rabbits have the capacity to induce antibodies against the HIV-1 co-receptor site, but that the <italic>in vivo</italic> presence of primate CD4 is required for the elicitation of these antibodies. This most likely occurs by a direct interaction with primate CD4 and induction of the co-receptor binding site, consistent with a recent study that detects CD4i antibodies following inoculation of monkeys with a CD4-gp120 fusion protein ##REF##17956985##[29]##. We also monitored for elicitation of neutralizing antibodies against HIV-1 pseudotyped virus in serum samples of these huCD4 rabbits and observed that the titers were not as consistent and potent for as for the WT rabbits (data not shown). These results might be due to the relatively advanced age of the huCD4 rabbits or as a consequence of unanticipated immune-related effects of the huCD4 transgene. However, as a slightly diminished immune response in these animals would only bias the results against the elicitation of CD4i antibodies, this remains a stringent model to assess the dependence upon primate CD4 for elicitation of these antibodies.</p>", "<title>The 17b antibody can recognize cell-surface, CD4-bound soluble Env trimers</title>", "<p>The elicitation of CD4i antibodies in monkeys and huCD4 rabbits after immunization with YU2 trimers suggests that the co-receptor site of gp120 is accessible for BCR recognition: likely on the surface of CD4⁺ PBMCs. Therefore, we investigated if the prototypic, co-receptor-site-directed mAb, 17b, could bind gp140-F trimers once they were bound cell-surface CD4. We sought to confirm that there was adequate accessibility of the CD4-induced co-receptor binding site on the trimers, once they were removed from the context of the virus. In the viral spike context, the induced co-receptor site is apparently not accessible to most CD4-induced antibodies due to steric constraints. To approximate the <italic>in vivo</italic> scenario in which trimers formulated in adjuvant would likely drain to proximal lymph nodes to encounter abundant CD4⁺ cells, we incubated cynomolgus macaques PBMCs with 20 µg/ml of the gp140-F trimers and detected CD4-specific binding to live CD3⁺/CD4⁺/CD8⁻ cells by the V3-directed antibody 447-52D or 17b using flow cytometry (##FIG##5##Fig. 6A##). Binding of the gp140-F trimers could be detected with both antibodies. Recognition of the trimers by 447-52D verifies that the YU2 gp140-F molecules bind to the CD4⁺ cells, while cell-surface recognition of the trimers by 17b confirms that the co-receptor site is accessible after trimer binding to membrane-bound CD4. Similar results were obtained when the cell-surface binding assay was repeated using human PBMC targets as shown in ##FIG##3##Fig 4B##. Monomeric gp120 bound to the human CD4⁺ cells was used as a control and displayed the 17b epitope at levels higher than that of the gp140-F trimers (##FIG##5##Fig 6B##).</p>", "<p>Sera from humans immunized with gp120 possess CD4i antibodies. Following the observation that the gp120 monomers bound to cell-surface CD4 displayed the CD4i epitope, we obtained serum samples from the VaxGen Inc phase III clinical trial now licensed to the Global Solutions for Infectious Diseases. Twenty randomly selected sera from trial volunteers that had been inoculated four times with recombinant gp120 (MN/GNE8 mixture) formulated with alum were assessed for gp120 binding antibodies by ELISA. All sera exhibited detectable binding titers to the unmatched YU2 gp120 ranging from 5000 to 100,000 endpoint titers (not shown). We assessed the ability of the sera to inhibit the entry of MN and, in the presence of CD4, the HIV-2 virus 7312. As shown in ##TAB##0##Table 1##, all sera neutralized not only the homologous virus, MN, but displayed detectable, relatively high-titer, cross-neutralizing, co-receptor-directed antibodies against the CD4-triggered HIV-2 isolate.</p>" ]

[ "<title>Discussion</title>", "<p>In this study, we demonstrate that the elicitation of co-receptor site directed antibodies by the YU2 gp140-F trimers requires the presence of primate CD4. We show that the relatively homogenous, soluble, stable YU2 trimers bind to human CD4 with high affinity in a solution-based assay that, by design, should be independent of oligomeric influences on functional affinity by avidity-dependent interactions. Analysis of the interaction between the prototypic co-receptor antibody, 17b, and the trimers demonstrated that high-affinity and high-avidity binding is detectable even in the absence of CD4. Binding of the trimers to primate cell-surface CD4, but not to cell-surface rabbit CD4 was also shown. WT rabbits and monkeys inoculated with the CD4-binding YU2 Env trimers formulated in the same adjuvant system elicited an overall similar pattern of HIV-1 <italic>in vitro</italic> neutralization against the viruses tested. However, cross neutralization of HIV-2 in presence of sCD4, an assay that is diagnostic for the detection of CD4i antibodies, was observed initially in sera derived from monkeys inoculated with the YU2 trimers but not in WT rabbits. Taken together, these data strongly suggest that Env-CD4 complexes generated <italic>in vivo</italic> upon inoculation are the source for elicitation of the CD4-induced antibodies following vaccination. This observation was confirmed by the inoculation of Env trimers into rabbits transgenic for human CD4 and the detection of CD4i antibodies in the sera of these animals, in contrast to WT rabbits, revealing conclusively the mechanism of their generation (see schematic ##FIG##6##Fig 7##). Consistent with this observation, high levels of co-receptor-directed, HIV-2 (+CD4) cross neutralizing antibodies were detected in 20 of 20 human serum samples from the VaxGen phase III clinical trial using monomeric gp120 ##REF##17109337##[13]##, suggesting that during natural infection shed, soluble gp120 can elicit these antibodies ##REF##15165814##[30]##.</p>", "<p>The induction of co-receptor site directed antibodies in non-human primates and humans is consistent with previous reports that detected the presence of CD4i, co-receptor directed antibodies in gp140-immunized humans (GMS, unpublished observations), as well as in naturally infected humans ##REF##15867093##[5]##,##REF##16720981##[31]## and following SHIV challenge of naïve non-human primates ##REF##17956985##[29]##. However, the mechanistic basis for the elicitation of CD4i antibodies was not previously addressed in a direct manner. In this study, we present a controlled experiment, which demonstrates that the elicitation of the co-receptor binding site antibodies by Env alone requires the presence of, and likely direct interaction with, primate CD4. This requirement has not previously been defined, despite numerous Env trimer immunogenicity experiments performed to date in both monkeys and non-primate species ##REF##15665645##[17]##,##REF##14512572##[18]##,##REF##16415019##[20]##,##REF##17540729##[32]##,##REF##15932765##[33]##. This is in part, because the HIV-2-based neutralization assay diagnostic for CD4i antibodies was a relatively recent development and is more definitive for neutralizing capacity directed at the co-receptor binding site then are binding assays employed by us and others previously ##REF##17360741##[23]##,##REF##17580087##[34]##,##REF##16160142##[35]##. Elicitation of CD4i antibodies in primates by Env trimers also nicely illustrates another potential mechanism of immune escape by HIV-1. Not only does Env binding to CD4 obscure a conserved surface that, if it was highly immunogenic, might elicit antibodies capable neutralizing a broad array of isolates (essentially antibodies mimicking the soluble primary receptor), but the binding event induces a second conserved and apparently immunogenic region: the co-receptor binding site. The CD4i antibodies directed against this region are not generally able to neutralize primary isolates in vitro ##REF##12970440##[7]##. This is likely due to a commonly elicited selection pressure that renders the co-receptor binding inaccessible to most antibodies of this type ##REF##11160708##[10]##. The inability of the CD4i antibodies to control HIV-1 infection is supported by data from a recent study where elicitation of CD4i antibodies can be detected prior to the detection of autologous virus neutralization capacity in sera derived from HIV-1 infected patients ##REF##17409164##[27]##, as well as the data here, which indicates that they were elicited in the phase III Vaxgen clinical trial where no protection was observed. However, a recent study in non-human primates suggests that the presence of CD4i antibodies (as determined <italic>in vitro</italic> by the same HIV-2 detection assay as used here) is associated with more rapid viral clearance following SHIV162P3 challenge ##REF##17956985##[29]##, illuminating that this is an area worthy of further investigation.</p>", "<p>In the present study, the most likely <italic>in vivo</italic> source for presentation of the CD4i region to the humoral immune system is by direct interaction of the trimers with cell-surface CD4 displayed on CD4-expressing T cells or other CD4-positive cells of the hematopoetic lineage. It is also possible that low levels of CD4 are shed from CD4⁺ cells into interstitial spaces and soluble complexes are formed. Detection of low levels of sCD4 has been previously reported in humans ##REF##1939640##[36]##,##REF##1914226##[37]##, although in the assay used here we could not detect soluble CD4 in the sera of animals examined. It is also possible that trimer binding to cell-surface CD4 induces shedding of complexes, but we could find no such evidence for soluble Env-CD4 complexes in the sera.</p>", "<p>The implications of inducing the co-receptor binding site has been discussed extensively at the level of entry, but less so at the level of antibody elicitation. That the CD4i antibodies are not elicited by trimers in the absence of CD4, even though the gp140-F molecules are well recognized by 17b, and that CD4 induction of the epitope in the trimer context is not a requirement for 17b binding, may seem to be a bit of a paradox. However, we interpret these data to indicate that the conformational fixation imparted by CD4 binding to gp120 is a critical requirement for the naïve, germline B cell repertoire to efficiently recognize the site as opposed to the affinity-matured 17b antibody, which can likely induce the fit of its epitope in the absence of CD4 (see Schematic ##FIG##6##Fig 7##). This is more likely to be an affinity limitation of the non-affinity matured BCR repertoire, although it is possible that it is somehow related to binding limitations of Ig molecules on the surface of B cells and epitope accessibility issues.</p>", "<p>Another possible interpretation of the data is that elements of the pre-CD4 co-receptor binding site are immunogenic, but do not elicit antibodies that cross react efficiently with the fully formed site induced by CD4. However, the 17b blocking assay, using a form of the gp120 core with the potential to be recognized by any antibodies to the co-receptor site indicated that the pre-CD4 state of the trimers did not elicit many antibodies directed toward this region compared to those elicited in the presence of primate CD4 (##FIG##3##Fig 4C##). Also, we cannot rule out that array of gp140 or gp120 on the surface of CD4+ cells might also enhance the elicitation of CD4i antibodies in a manner independent of conformational fixation. However, the study by DeVico <italic>et al</italic>\n##REF##17956985##[29]## clearly demonstrates that covalent gp120-CD4 complexes incapable of binding cell-surface CD4 efficiently elicit CD4i antibodies, so CD4-dependent cell-surface array cannot be the only explanation for their elicitation in the presences of primate CD4.</p>", "<p>The implications of the data presented here are also an important consideration for vaccine candidates designed to elicit neutralizing antibodies against the conserved gp120 CD4 binding site. The Env CD4bs likely remains fully accessible in animals without human or primate CD4, however the elicitation of the CD4i antibodies in animals with primate CD4 indicates that this is likely not the case in species harboring CD4 molecules with a high affinity to Env. These results suggest that a fraction of the population of a CD4-binding-competent immunogen will interact with primate CD4 and thereby occlude the CD4 binding region on this protein subset. It is possible that the subtle differences detected in the neutralization profile between WT rabbits and monkeys occur as a result of such an interaction, partially altering the spectrum of antibodies that are elicited. However, the fractional component of the inoculate which binds to CD4 as yet remains to be determined, and may not be absolute as the overall HIV-1 neutralization profile elicited by the trimers used in this study was similar between the rabbits and the non-human primates.</p>", "<p>It was also previously shown that HIV Env-CD4 interaction resulted in altered CD4⁺ T cell function <italic>in vitro</italic>\n##REF##17158230##[38]## and it was suggested that elimination of Env interaction with CD4 in the context of vaccination might be beneficial to better elicit functional T cell help and more potent neutralizing antibody responses. From that study and the data presented here it will be interesting to assess if Env variants that do not bind CD4, but still retain the ability to bind CD4 binding site antibodies might make better immunogens than do unmodified YU2 gp140-F proteins. Alternatively, redirecting the immunogen more efficiently to B cell and antigen presenting cells might also overcome any potential detrimental effects of Env-based immunogens interacting with primate CD4. Follow up immunogen trimer design, characterization and immunogenicity studies are warranted to clarify these issues further in the near future.</p>" ]

[]

[ "<p>Conceived and designed the experiments: MNF GBKH RTW. Performed the experiments: MNF BD AM KS SO. Analyzed the data: MNF BD JRM GBKH RTW. Contributed reagents/materials/analysis tools: MNF CMH GV RT GMS JRM. Wrote the paper: MNF JRM GBKH RTW.</p>", "<p>The surface HIV-1 exterior envelope glycoprotein, gp120, binds to CD4 on the target cell surface to induce the co-receptor binding site on gp120 as the initial step in the entry process. The binding site is comprised of a highly conserved region on the gp120 core, as well as elements of the third variable region (V3). Antibodies against the co-receptor binding site are abundantly elicited during natural infection of humans, but the mechanism of elicitation has remained undefined. In this study, we investigate the requirements for elicitation of co-receptor binding site antibodies by inoculating rabbits, monkeys and human-CD4 transgenic (huCD4) rabbits with envelope glycoprotein (Env) trimers possessing high affinity for primate CD4. A cross-species comparison of the antibody responses showed that similar HIV-1 neutralization breadth was elicited by Env trimers in monkeys relative to wild-type (WT) rabbits. In contrast, antibodies against the co-receptor site on gp120 were elicited only in monkeys and huCD4 rabbits, but not in the WT rabbits. This was supported by the detection of high-titer co-receptor antibodies in all sera from a set derived from human volunteers inoculated with recombinant gp120. These findings strongly suggest that complexes between Env and (high-affinity) primate CD4 formed <italic>in vivo</italic> are responsible for the elicitation of the co-receptor-site-directed antibodies. They also imply that the naïve B cell receptor repertoire does not recognize the gp120 co-receptor site in the absence of CD4 and illustrate that conformational stabilization, imparted by primary receptor interaction, can alter the immunogenicity of a type 1 viral membrane protein.</p>", "<title>Author Summary</title>", "<p>A major goal of HIV-1 vaccine research is to design novel candidates capable of neutralizing the vast array of viruses circulating in the human population. One approach is to base the vaccine upon the HIV-1 outer surface envelope glycoproteins to generate antibodies. However, during persistent infection in humans, the HIV-1 envelope glycoproteins have evolved structural features that limit the elicitation of broadly neutralizing antibodies. These immune “decoys” divert the antibody response resulting in virus subpopulations that can escape the host response. A potential means by which the virus elicits these decoy responses comes as a by-product of the entry process. Binding of the HIV-1 envelope glycoproteins to the primary receptor, human CD4, induces the formation of a second co-receptor binding site on the envelope glycoproteins, which then binds to another protein required for viral entry. Antibodies to the co-receptor binding site are generally ineffective at neutralizing HIV-1 patient isolates. Here, we demonstrate the mechanism by which antibodies to the HIV-1 co-receptor binding site are elicited in animals and humans injected with HIV-1 envelope glycoproteins and describe the implications of their formation regarding natural HIV-1 infection and vaccine design.</p>" ]

[ "<title>Supporting Information</title>" ]

[ "<p>We thank Brenda Hartman and Michael Cichanowski, for help with the figures, Dr Mats Spångberg, Dr Helene Fredlund and personnel at the Astrid Fagraeus laboratory at the Swedish Institute for Infectious Disease Control for expert assistance with the non-human primates and JP Todd and Alida Ault for expert assistance with the transgenic rabbits. We would like to thank Global Solutions for Infectious Diseases and their scientists for providing the VaxGen Inc phase III clinical serum samples. We thank James Robinson (17b) and Susan Zolla-Pazner (447-52D) for providing antibodies, Adhuna Phogat for help with trimer protein production and purification and Ralph Pantophlet, Emily Carrow and Alan Schultz for baboon sera.</p>" ]

[ "<fig id=\"ppat-1000171-g001\" position=\"float\"><object-id pub-id-type=\"doi\">10.1371/journal.ppat.1000171.g001</object-id><label>Figure 1</label><caption><title>Soluble Env trimers bind to soluble and cell-surface CD4.</title><p>(A) sCD4 at concentrations of 2, 4 and 9 nM was co-incubated with YU2 gp140-F trimers (filled circles) or CD4-binding-defective gp140 368D/R trimers (open circles) at concentrations shown. Subsequently, non-trimer-bound sCD4 was captured in an ELISA format by the anti-CD4 antibody RPA-T4, competing with HIV-1 Env for CD4 binding. RPA-T4-captured sCD4 was detected with the non-competing anti-CD4 antibody, OKT-4 and plotted as shown. (B) Cynomolgus macaque PBMCs were incubated with 2, 10 or 20 µg/ml of YU2 gp140-F trimers, 20 µg/ml gp140-F trimers in the presence of an excess sCD4 (100 µg/ml) or with 20 µg/ml of gp140-F 368D/R trimers. Cells were stained for CD3, CD4 and CD8 expression and analyzed by flow cytometry. The CD3⁺/CD8⁻ populations are shown (see supplemental ##SUPPL##1##FigS2## for gating strategy). The y-axis indicates Env binding, as detected with mAb 447-52D, and the x-axis shows CD4 expression. The CD4⁺ cell population was defined as cells detected to the right of the dotted line. The median fluorescence intensity (MFI) of Env binding to the CD3⁺/CD4⁺/CD8⁻ cell population is shown.</p></caption></fig>", "<fig id=\"ppat-1000171-g002\" position=\"float\"><object-id pub-id-type=\"doi\">10.1371/journal.ppat.1000171.g002</object-id><label>Figure 2</label><caption><title>ELISA analysis of the 17b:trimer interaction.</title><p>The YU2 gp140-F trimers, at concentrations of 3.2 ng/ml to 10 µg/ml, were incubated in the presence (closed circles) or absence (filled circles) of 20 µg/ml sCD4, after which recognition of the trimers by the co-receptor-binding-site-directed antibody, 17b, was analyzed by ELISA. Because of the oligomeric state of the trimers, this assay likely assesses binding avidity, however, the precise stoichiometry of CD4 occupancy of each trimer or the functional ability of each gp120 subunit within the trimer to bind CD4 is not yet known. The addition of sCD4 to the trimers resulted in an approximate 5-fold shift in the levels of ½ maximal binding, from 9 nM, in the absence of CD4, to 2 nM in the presence of CD4.</p></caption></fig>", "<fig id=\"ppat-1000171-g003\" position=\"float\"><object-id pub-id-type=\"doi\">10.1371/journal.ppat.1000171.g003</object-id><label>Figure 3</label><caption><title>Biacore analysis of 17b:trimer interactions.</title><p>The recognition of the gp140-F trimers by 17b IgG was assessed by SPR in two formats as shown. (A) 17b IgG was immobilized on the sensor surface and the trimers were then flowed over the Biacore chip either with or without pre-incubation with a 60-fold molar excess of D1D2 CD4. An approximation of the “functional affinity” of each interaction is presented as calculated by 1∶1 Langmuir binding curve fitting. These values likely represent avidity due to the oligomeric state of the gp140-F trimers as the analyte. In this format, the off-rate value was difficult to accurately assess due to the very slow rate of dissociation and may overestimate the functional affinity or avidity. Nevertheless, this is clearly an avid reaction and the on-rates shown here agree with previous reports of the affinity values of monomeric gp120 recognition by monomeric 17b Fab, which is likely the best estimate of affinity ##REF##11327825##[6]##. (B) In the other orientation, the trimers were immobilized on the chip surface and 17b was used as the analyte. The affinity was calculated using bivalent curve fitting since the 17b IgG is dimeric. The value derived by this analysis indicates a high-affinity, low nanomolar interaction of the 17b IgG with the trimers in the absence of CD4, as shown.</p></caption></fig>", "<fig id=\"ppat-1000171-g004\" position=\"float\"><object-id pub-id-type=\"doi\">10.1371/journal.ppat.1000171.g004</object-id><label>Figure 4</label><caption><title>HIV-1 and HIV-2 neutralization.</title><p>(A) Overall neutralization activity against nine different HIV-1 Env pseudotyped viruses in serum samples from monkeys and rabbits following 4 immunizations with the YU2 gp140-F trimers. Sera were screened at a 1∶5 dilution (except against MW.965, which was done at a 1∶10 dilution of the sera) for neutralization activity. Neutralization between 70 and 79% is indicated in yellow, while neutralization between 80 and 100% is indicated in red. Pre-bleed serum sample effects on viral entry were negligible; BSA adjuvant control animals and MuLV pseudotyped virus were also included as negative controls. (B) Detection of CD4i antibodies in serum samples from immunized cynomolgus macaques and rabbits after 4 immunizations. Serum samples from monkeys and rabbits were titrated for neutralization activity against HIV-2_7312/V434M. Data are presented as the reciprocal dilution of immune serum resulting in a 50% inhibition of entry (ID₅₀) in the absence (white bars) or presence (blue bars) of 9 nM sCD4. (C) A 17b blocking assay was performed using sera elicited from monkeys (blue) and WT rabbits (black). For the assay, the sera were serially diluted in a 96 well format in which the wells were pre-coated with HXBc2 core gp120 glycoproteins. Following washing, biotinylated 17b was added to each well and the ability of the sera to block 17b binding was assessed by ELISA.</p></caption></fig>", "<fig id=\"ppat-1000171-g005\" position=\"float\"><object-id pub-id-type=\"doi\">10.1371/journal.ppat.1000171.g005</object-id><label>Figure 5</label><caption><title>Characterization of huCD4 rabbit PBMCs and HIV-2 neutralization.</title><p>(A) CD4 expression and YU2 gp140-F trimer binding to PBMCs from WT and huCD4 rabbits (T4-19, T4-24, T4-41, T4-44 and T4-45) was analyzed by flow cytometry. PBMCs were incubated with 20 µg/ml of the gp140-F trimers and stained with an anti-rabbit CD4 antibody (x-axis) followed by detection of trimer binding with the mAb 447-52D (y-axis) (see supplemental ##SUPPL##4##Fig S5## for gating strategy). rCD4⁺/Env⁻ cells are located in the lower right quadrant while rCD4⁺/Env⁺ cells are located in the upper right quadrant. The percentage of cells detected in each quadrant is indicated. (B) ID₅₀ HIV-2_7312/V434M neutralization in the presence of 9 nM sCD4 by serum samples from WT and huCD4 rabbits (blue) after 3 immunizations with the gp140-F trimers.</p></caption></fig>", "<fig id=\"ppat-1000171-g006\" position=\"float\"><object-id pub-id-type=\"doi\">10.1371/journal.ppat.1000171.g006</object-id><label>Figure 6</label><caption><title>Accessibility of the HIV-1 co-receptor site after CD4 binding.</title><p>(A). PBMCs from cynomolgus macaques and humans were co-incubated with 20 µg/ml YU2 trimers or (B) with 5 µg/ml YU2gp120 (only human PBMCs). CD4 specific trimer binding (blue) to CD3⁺/CD4⁺/CD8⁻ cells was detected with either the V3-directed mAb, 447-52D (left panels), or with the co-receptor site-directed mAb, 17b (right panels), by flow cytometry analysis. Negative control staining (red, both panels) shows the fluorescence signal obtained in the absence of 447-52D or 17b.</p></caption></fig>", "<fig id=\"ppat-1000171-g007\" position=\"float\"><object-id pub-id-type=\"doi\">10.1371/journal.ppat.1000171.g007</object-id><label>Figure 7</label><caption><title>Schematic representations of the pre-CD4 and post-CD4 conformations of Env.</title><p>(A) Based upon the structures of the unliganded SIV core gp120 ##REF##15729334##[40]## and the CD4-bound HIV-1 core ##REF##9641677##[41]##, the movements of the co-receptor site bridging sheet beta-strands are highlighted in green. Note that 17b binds to the unliganded HIV-1 core with low affinity (∼1 uM) consistent with the “split orientation” of the bridging sheet revealed by the unliganded SIV core structure. 17b can also recognize full-length gp120 with high affinity ##REF##11327825##[6]## and the gp140 trimers as demonstrated in this study (see Figs 2 and 3). These observations suggest that the bridging sheet may not be in exactly the same conformation in the full-length protein context as it is in the original gp120 core protein structures with and without CD4. (B) A model of how the co-receptor region (green) is recognized by the naïve B cell receptor repertoire. In the left and middle panel, the co-receptor binding site is not formed. Following binding to high-affinity primate CD4 (likely on the surface of CD4⁺ cells), the bridging sheet is formed and locked into a single, CD4-dependent conformation, which then allows elicitation of co-receptor-site-directed, CD4i antibodies (right panel). The somatically mutated and affinity-matured 17b antibody can recognize either conformation, presumably inducing a proper fit of the unliganded bridging sheet conformation.</p></caption></fig>" ]

[ "<table-wrap id=\"ppat-1000171-t001\" position=\"float\"><object-id pub-id-type=\"doi\">10.1371/journal.ppat.1000171.t001</object-id><label>Table 1</label><caption><title>Neutralization values present in sera elicited by gp120 inoculated into humans</title></caption><alternatives><table frame=\"hsides\" rules=\"groups\"><colgroup span=\"1\"><col align=\"left\" span=\"1\"/><col align=\"center\" span=\"1\"/><col align=\"center\" span=\"1\"/><col align=\"center\" span=\"1\"/><col align=\"center\" span=\"1\"/></colgroup><thead><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">Serum sample</td><td colspan=\"4\" align=\"left\" rowspan=\"1\">VIRUS</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\"/><td colspan=\"2\" align=\"left\" rowspan=\"1\">HIV<sc>-2</sc>\n<sc>7312/</sc>V<sc>4334</sc>M<sc>+9n</sc>M <sc>s</sc>CD<sc>4</sc>\n</td><td colspan=\"2\" align=\"left\" rowspan=\"1\">HIV-1 MN</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\"/><td align=\"left\" rowspan=\"1\" colspan=\"1\">ID₅₀\n</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">ID₈₀\n</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">ID₅₀\n</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">ID₈₀\n</td></tr></thead><tbody><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">\n<bold>001</bold>\n</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">4,178</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">737</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">24,410</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">1,925</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">\n<bold>002</bold>\n</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">74,789</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">10,124</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">141,655</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">12,343</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">\n<bold>003</bold>\n</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">8,890</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">1,145</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">11,180</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">1,051</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">\n<bold>004</bold>\n</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">2,452</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">365</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">1,181</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">216</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">\n<bold>005</bold>\n</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">15,110</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">2,279</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">9,134</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">1,912</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">\n<bold>006</bold>\n</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">1,286</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">309</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">1,116</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">91</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">\n<bold>007</bold>\n</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">1,664</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">406</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">8,496</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">577</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">\n<bold>008</bold>\n</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">2,002</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">391</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">2,340</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">378</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">\n<bold>009</bold>\n</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">2,974</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">642</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">46,508</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">1,485</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">\n<bold>010</bold>\n</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">3,226</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">666</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">20,787</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">1,619</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">\n<bold>011</bold>\n</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">1,229</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">226</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">8,962</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">369</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">\n<bold>012</bold>\n</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">630</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">94</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">1,532</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">14</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">\n<bold>013</bold>\n</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">3,601</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">548</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">20,572</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">1,143</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">\n<bold>014</bold>\n</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">1,864</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">339</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">21,759</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">1,313</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">\n<bold>015</bold>\n</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">292</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">80</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">1,351</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">282</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">\n<bold>016</bold>\n</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">5,415</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">1,162</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">6,687</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">967</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">\n<bold>017</bold>\n</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">1,904</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">376</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">31,204</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">2,083</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">\n<bold>018</bold>\n</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">3,571</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">661</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">2,579</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">383</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">\n<bold>019</bold>\n</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">430</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">115</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">302</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">87</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">\n<bold>020</bold>\n</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">762</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">132</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">6,406</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">644</td></tr></tbody></table></alternatives></table-wrap>" ]

[]

[]

[]

[]

[]

[ "<supplementary-material content-type=\"local-data\" id=\"ppat.1000171.s001\"><label>Figure S1</label><caption><p>Biochemical analysis of the YU2 gp140-F trimers.</p><p>(3.17 MB TIF)</p></caption></supplementary-material>", "<supplementary-material content-type=\"local-data\" id=\"ppat.1000171.s002\"><label>Figure S2</label><caption><p>FACS gating strategy for cynomolgus macaque and human PBMCs.</p><p>(2.29 MB TIF)</p></caption></supplementary-material>", "<supplementary-material content-type=\"local-data\" id=\"ppat.1000171.s003\"><label>Figure S3</label><caption><p>HIV-2_7312/V434M (+9 nM sCD4) and HIV-1_MN neutralization by sera from monkeys immunized 2 times with gp140-F trimers formulated in the GSK Adjuvant System AS01B or in Ribi adjuvant.</p><p>(1.12 MB TIF)</p></caption></supplementary-material>", "<supplementary-material content-type=\"local-data\" id=\"ppat.1000171.s004\"><label>Figure S4</label><caption><p>V3 peptide or gp120 binding reactivity in serum samples from cynomolgus macaques, WT rabbits or huCD4 rabbits after immunization with gp140-F trimers.</p><p>(2.12 MB TIF)</p></caption></supplementary-material>", "<supplementary-material content-type=\"local-data\" id=\"ppat.1000171.s005\"><label>Figure S5</label><caption><p>Staining and gating strategy for flow cytometry analysis of rabbit PBMCs.</p><p>(2.12 MB TIF)</p></caption></supplementary-material>" ]

[ "<fn-group><fn fn-type=\"COI-statement\"><p>Gerald Voss is an employee of GlaxoSmithKline Biologicals. All other authors have declared that no competing interests exist.</p></fn><fn fn-type=\"financial-disclosure\"><p>This study was supported by a grant from Swedish International Development Agency (Sida)/Department of Research Cooperation (SAREC) (GBKH and RT) and by the National Institute of Allergy and Infectious Diseases, National Institutes of Allergies and Infectious Diseases intramural research program (RTW and JRM), the International AIDS Vaccine Initiative (GBKH and RTW) and the Bill and Melinda Gates Foundation (GMS, JRM, RTW).</p></fn></fn-group>" ]

[ "<graphic xlink:href=\"ppat.1000171.g001\"/>", "<graphic xlink:href=\"ppat.1000171.g002\"/>", "<graphic xlink:href=\"ppat.1000171.g003\"/>", "<graphic xlink:href=\"ppat.1000171.g004\"/>", "<graphic xlink:href=\"ppat.1000171.g005\"/>", "<graphic xlink:href=\"ppat.1000171.g006\"/>", "<graphic id=\"ppat-1000171-t001-1\" xlink:href=\"ppat.1000171.t001\"/>", "<graphic xlink:href=\"ppat.1000171.g007\"/>" ]

[ "<media xlink:href=\"ppat.1000171.s001.tif\"><caption><p>Click here for additional data file.</p></caption></media>", "<media xlink:href=\"ppat.1000171.s002.tif\"><caption><p>Click here for additional data file.</p></caption></media>", "<media xlink:href=\"ppat.1000171.s003.tif\"><caption><p>Click here for additional data file.</p></caption></media>", "<media xlink:href=\"ppat.1000171.s004.tif\"><caption><p>Click here for additional data file.</p></caption></media>", "<media xlink:href=\"ppat.1000171.s005.tif\"><caption><p>Click here for additional data file.</p></caption></media>" ]

{ "acronym": [], "definition": [] }

[ "<title>Introduction</title>", "<p>Hematogenous dissemination of pathogenic organisms is an important feature of disease progression. However, dissemination is poorly understood, in large part because of the difficulty in studying this process directly in living organisms under the shear stress conditions that characterize the host vasculature. One such disseminating pathogen is the spirochete <italic>Borrelia burgdorferi</italic>, a primarily extracellular bacterium causing Lyme disease, also referred to as Lyme borreliosis ##REF##17160600##[1]##.</p>", "<p>Pathogenic spirochetes cause a number of emerging and re-emerging diseases, including syphilis, leptospirosis, relapsing fever and Lyme disease ##REF##15085185##[2]##–##REF##16704771##[5]##. <italic>B. burgdorferi</italic> is transmitted to the dermis of vertebrate hosts during the blood meal of <italic>Ixodes</italic> ticks, and subsequently disseminates to other tissues and organs during the hematogenous phase of infection ##REF##17160600##[1]##. <italic>B. burgdorferi</italic> and other spirochetes interact with endothelial cells under static conditions <italic>in vitro</italic>\n##REF##8106781##[6]##–##REF##3285346##[8]##. However, until recently, spirochete-vascular interactions have never been directly examined in the host itself, or under the fluid shear forces that are present at dissemination sites ##REF##18566656##[9]##.</p>", "<p>To facilitate direct study of hematogenous dissemination we recently generated a fluorescent infectious strain of <italic>B. burgdorferi</italic>, and used intravital microscopy (IVM) to directly visualize its interaction with and extravasation from the microvasculature of living murine hosts (as summarized in ##FIG##0##\nFig. 1A–C\n##) ##REF##18566656##[9]##. IVM is a powerful tool for studying the dissemination and transmigration of tumor and immune cells in living hosts, but it is only recently that this technique has begun to be applied to the study of host-pathogen interactions ##REF##17662072##[10]##,##REF##17983749##[11]##.</p>", "<p>The results of our recent study indicated that <italic>B. burgdorferi</italic> dissemination from the host microvasculature <italic>in vivo</italic> is a progressive, multi-stage process consisting of several successive steps: transient and dragging interactions (collectively referred to as short-term interactions), followed by stationary adhesion and extravasation. Short-term interactions constitute the majority of spirochete-endothelial associations (89% and 10% for transient and dragging interactions, respectively), take less than one second (transient interactions) or 3–20s (dragging interactions) to travel 100 µm along the vessel wall, and occur primarily on the surface of endothelial cells and not at endothelial junctions ##REF##18566656##[9]##. Transient interactions are characterized by a tethering-type attachment-detachment cycle of association in which part of the spirochete adheres briefly to the endothelium before being displaced by blood flow, whereas dragging spirochetes adhere along much of the length of the bacterium, and creep more slowly along the vessel wall ##REF##18566656##[9]##. In contrast, stationary adhesions (1% of interactions) do not move along the vessel wall, occur chiefly, but not exclusively, at endothelial junctions, and entail a more intimate association with the endothelium than short-term interactions ##REF##18566656##[9]##. Finally, spirochete extravasation (&lt;0.12% of interactions) also occurs primarily, but not exclusively, at endothelial junctions, and is a triphasic process consisting of a rapid, end-first initial penetration of the endothelium, followed by a prolonged period of reciprocating movement, and ending with a rapid exit phase in which the bacterium bursts out of the vessel and migrates rapidly into the surrounding tissue ##REF##18566656##[9]##.</p>", "<p>\n<italic>In vitro</italic> studies have shown that <italic>B. burgdorferi</italic> binds several host molecules that might mediate endothelial interactions <italic>in vivo</italic>, including fibronectin (Fn), integrins, heparan sulfate-type glycosaminoglycans (GAGs) and regulators of the complement cascade ##REF##2332509##[12]##-##REF##11385611##[20]##. A broad array of pathogens have been shown to interact with these ubiquitous host molecules in direct binding assays and tissue culture models <italic>in vitro</italic>; most of these studies have been performed in the absence of shear forces, microvascular endothelium or a functioning immune system, and so the potential contribution of such interactions to hematogenous dissemination in the living host is unknown. To date, 19 candidate adhesin genes have been identified in <italic>B. burgdorferi</italic>, two of which are known to interact with integrins (P66 and BBB07), and two of which can associate with heparan sulfate GAGs (BBK32 and Bgp) ##REF##9988477##[16]##, ##REF##10594819##[17]##, ##REF##17822440##[19]##, ##REF##17784908##[21]##–##REF##16368999##[25]##. <italic>B. burgdorferi</italic> encodes one characterized Fn binding protein, BBK32, and appears to express a number of others ##REF##2332509##[12]##,##REF##9988477##[16]##,##REF##9685613##[24]##. Five <italic>B. burgdorferi</italic> CRASP proteins that interact with complement cascade regulating proteins factor H, FHL-1 and FHR-1 have also been identified ##REF##11385611##[20]##,##REF##11705962##[26]##,##REF##17538892##[27]##, but their potential contributions to endothelial cell adhesion are unknown.</p>", "<p>In the work described here we used IVM to explore the mechanistic basis for <italic>B. burgdorferi</italic> interactions with the microvasculature of living mice. We found that the initiating and stationary adhesion stages of microvascular interactions were mechanistically distinct but inter-dependent events, and that BBK32, Fn and GAGs played a substantial role in initiation events. These findings and the methodology described here provide a framework for investigating the role of Fn and GAGs in vascular interactions during hematogenous dissemination by <italic>B. burgdorferi</italic> and possibly other pathogens.</p>" ]

[ "<title>Materials and Methods</title>", "<title>Construction of the BBK32 expression plasmid pTM170</title>", "<p>Plasmid pTM170 was constructed by PCR amplification of the P<italic>ospC</italic>-driven <italic>bbk32</italic> cassette from pBBK32 ##REF##16368999##[25]## with flanking <italic>Kpn</italic>I and <italic>Fsp</italic>I sites, using primers B1093 (<named-content content-type=\"gene\">5′-GGTACCTTAATTTTAGCATATTTGGCTTTG-3′</named-content>) and B1094 (<named-content content-type=\"gene\">5′-GGCCTGCGCATTAGTACCAAACGCCATTCTTG-3′</named-content>). The PCR product was cloned into the GeneJet plasmid, using the Gene Jet blunt cloning kit (MBI Fermentas) to generate pTM169. The <italic>Kpn</italic>I/<italic>Fsp</italic>I-digested pTM169 insert was cloned into the KpnI/FspI sites of the GFP-encoding plasmid pTM61 ##REF##18566656##[9]## to yield pTM170.</p>", "<title>\n<italic>B. burgdorferi</italic> transformations and screening</title>", "<p>\n<italic>B. burgdorferi</italic> strains used in this study were GCB705 (non-infectious strain B31-A transformed with pTM61) ##REF##18566656##[9]##,##REF##10762244##[62]##, GCB726 (infectious <italic>B. burgdorferi</italic> strain B31 5A4 NP1 transformed with pTM61) ##REF##18566656##[9]##,##REF##15557639##[63]## and GCB769 (non-infectious B31-A transformed with pTM170). The plasmid content of these strains is noted in ##SUPPL##0##\nTable S1\n##. All strains were grown in BSK-II medium prepared in-house ##REF##6393604##[64]##. Electrocompetent <italic>B. burgdorferi</italic> strains were prepared as described ##REF##18566656##[9]##,##REF##7550741##[65]##. Liquid plating transformations were performed with 50 µg pTM61 or pTM170 in the presence of 100 µg/ml gentamycin as described ##REF##17714442##[66]##,##REF##14981112##[67]##. Gentamycin-resistant <italic>B. burgdorferi</italic> clones were screened for: 1) the presence of <italic>aacC1</italic> sequences by colony screening PCR performed with primers B348 and B349 as described ##REF##16936037##[68]##; 2) GFP expression by conventional epifluorescence microscopy, and 3) BBK32 expression, as detected by immunoblotting (for <italic>bbk32</italic> complementation strains). The presence of plasmids in non-integrated form in fluorescent strains was confirmed by agarose gel electrophoresis of total genomic DNA prepared on a small scale as described ##REF##15948955##[69]##. PCR screening for native plasmid content was performed as described ##REF##16936037##[68]##,##REF##11106398##[70]##.</p>", "<title>PCR amplification and sequencing of candidate genes</title>", "<p>Gene sequences were amplified from genomic DNA preparations of GCB705 and GCB726. PCR was performed with Phusion DNA polymerase (NEB, Pickering, Ontario, Canada), according to manufacturer's instructions. Sequencing was performed by the University of Calgary DNA Services. All primers used for PCR amplification and sequencing are provided in ##SUPPL##0##\nTable S1\n##.</p>", "<title>Protein expression and localization studies</title>", "<p>The expression and outer membrane localization of adhesins P66 and Bgp were analyzed as previously described ##REF##9988477##[16]##,##REF##16368999##[25]##. Briefly, for each strain, two pellets containing 5×10⁷ spirochetes were washed twice with PBS+2% BSA. PBS containing 5 mM MgCl₂ was added without dislodging pellets. Proteinase K was added to one pellet to a concentration of 4 mg/ml. After 30 min incubation at room temperature, reactions were stopped with 150 µg phenylmethylsulfonyl fluoride, spirochetes were pelleted and washed twice with PBS+0.2% BSA, and pellets were lysed using SDS-PAGE loading dye. Proteins were resolved by electrophoresis on 12% SDS-PAGE gels, transferred to nitrocellulose membranes, followed by immunoblotting with antibodies to P66, Bgp or BBK32, as previously described ##REF##10594819##[17]##,##REF##10712702##[22]##.</p>", "<title>Preparation of fluorescent <italic>B. burgdorferi</italic> for direct bloodstream injection, surgical preparations and intravital microscopy conditions</title>", "<p>These conditions have been described in detail previously ##REF##18566656##[9]##. Quantification of spirochete interactions was performed as recently described ##REF##18566656##[9]##. All animal studies were carried out in accordance with the guidelines of the University of Calgary Animal Research Centre.</p>", "<title>Quantification of leukocyte recruitment in dermal microvasculature of mice in the presence and absence of infectious <italic>B. burgdorferi</italic>\n</title>", "<p>Leukocyte recruitment studies were carried out as previously described ##REF##18097053##[28]##. Briefly, animals were injected with 50 µl of 0.05% (i.v.) rhodamine 6G (Sigma-Aldrich). Fluorescence was visualized by epi-illumination using 510 and 560 filters. Leukocytes were considered adherent to the venular endothelium if they remained stationary for 30 s or longer. Experiments were performed in mice that had been intravenously inoculated with 4×10⁸ infectious <italic>B. burgdorferi</italic> grown for 48h in 1% mouse blood, as previously described, and also with mice that were not inoculated with spirochetes. Leukocyte adhesions were counted in the dermal postcapillary venules of infected and non-infected mice from 5 minutes after injection of spirochetes and/or rhodamine until at least 1 hour from injection, in order to monitor leukocyte recruitment during the time frame that is used for all experiments reported in this study.</p>", "<title>Fibronectin antibody experiments</title>", "<p>GCB726 spirochetes prepared as described above were resuspended to 2×10⁹/ml in PBS. The IgG fraction of polyclonal goat anti-rabbit plasma Fn serum or non-specific goat IgGs (Cappel/MP Biomedicals, Solon, OH) were added to 1 mg/ml final. After mixing for 30 min at room temperature, spirochetes were directly injected into the mouse bloodstream, along with 2 mg of Fn antiserum IgGs or non-specific IgGs.</p>", "<title>Fn peptide experiments</title>", "<p>GCB726 spirochetes were prepared and injected as described above, together with 100 µg of GRGDS or FN-C/H II peptide (Sigma Canada, Oakville, ON; catalogue numbers G4391 and F7049, respectively), injected via the femoral vein. Peptides injected at this amount are known to disrupt leukocyte adhesion and recruitment <italic>in vivo</italic>\n##REF##8040319##[38]##.</p>", "<title>Dalteparin and dextran sulfate experiments</title>", "<p>Two hundred µl of a 25 I.U./µl solution of dalteparin (Fragmin: Pfizer Canada, Kirkland, PQ) were injected via the femoral vein 15 minutes before intravenous inoculation with infectious spirochetes. This concentration has previously been shown to inhibit leukocyte rolling <italic>in vivo</italic>\n##UREF##0##[32]##. Dextran sulfate-treated spirochetes were incubated with 20 µg/ml dextran sulfate (500 kDa; Fisher Scientific Canada, Ottawa, ON) in a final volume of 100 ml PBS for 30 min at RT°C, followed by 2 100ml washes with PBS. Spirochetes were resuspended to 2×10⁹/ml in PBS, and injected as previously described ##REF##18566656##[9]##. The concentration of dextran sulfate used in these preincubations (20 µg/ml) is the dose that maximally inhibits <italic>B. burgdorferi</italic> interaction with endothelial cells <italic>in vitro</italic>, and does not affect spirochete morphology or motility ##REF##7532628##[30]##,##REF##9488387##[31]##.</p>", "<title>Integrin-targeting antibody experiments</title>", "<p>One hundred µg anti-CD41 monoclonal Ab (Clone MwReg30; Becton Dickinson, San Diego, CA), or 20 µg CD49e monoclonal Ab (clone 5H10-27; Pharmingen, Oxford, UK) were intravenously administered immediately prior to injection of spirochetes. These quantities of anti-CD41 and CD49e antibodies are those that respectively protect against <italic>Plasmodium berghei</italic> infection <italic>in vivo</italic>\n##REF##15494426##[71]##, and which inhibit neutrophil migration <italic>in vivo</italic>\n##REF##11207307##[72]##.</p>", "<title>Statistics</title>", "<p>For quantitative analysis, average and standard error values for different variables were calculated and plotted graphically for all vessels from all mice using GraphPad Prism 4.03 (GraphPad Software, Inc., San Diego, CA). Statistical significance was calculated in GraphPad Prism using a two-tailed non-parametric Mann Whitney t-test with a 95% confidence interval.</p>" ]

[ "<title>Results</title>", "<title>Host GAGs promote <italic>B. burgdorferi</italic> interactions with the microvasculature <italic>in vivo</italic>\n</title>", "<p>To quantitatively analyze <italic>B. burgdorferi</italic> interactions with the microvascular endothelium <italic>in vivo</italic> we employed conventional epifluorescence IVM to examine interactions in the flank skin of mice after intravenous inoculation with 4×10⁸ spirochetes (##FIG##0##\nFig. 1\n## and ##SUPPL##1##\nVideos S1\n## and ##SUPPL##2##\nS2\n##). Conventional epifluorescence IVM was used instead of spinning disk confocal IVM because it is more effective for imaging the rapid associations that constitute the bulk of <italic>B. burgdorferi</italic> microvascular interactions (##FIG##0##\nFig. 1\n##). Analysis was performed in post-capillary venules, where interactions could be most accurately quantified. For all experiments reported in this manuscript, data describing the numbers of recorded vessels and mice for each experimental condition, as well as the average time after spirochete injection at which recordings were made are provided in the <bold>Figure Legends</bold>. As we have recently reported, during the experimental observation period no signs of endothelial or leukocyte activation were detected ##REF##18566656##[9]##. In addition, leukocyte adhesion in dermal postcapillary venules is an indicator of local activation, and can be measured by using the dye rhodamine 6G to fluorescently label all circulating leukocytes and then counting the number of adherent leukocytes in a 100 µm length of vessel ##REF##18097053##[28]##. The presence of infectious <italic>B. burgdorferi</italic> in the vasculature for as long as 70 minutes after injection of spirochetes did not significantly alter leukocyte adhesion from the baseline levels observed in the absence of <italic>B. burgdorferi</italic> (1.56−/+0.29 vs 1.75−/+1.15 adhered leukocytes/100 µm, respectively; P = 0.797; N = 50 vessels from 5 mice). The observed number of leukocyte adhesions was normal for the dermal microvasculature of mice ##REF##18097053##[28]##.</p>", "<p>As shown in ##FIG##1##\nFig. 2\n##, the ability to interact with the microvascular endothelium was specific to infectious spirochetes. When mice were injected with non-infectious <italic>B. burgdorferi</italic> exhibiting the same fluorescence intensity as infectious spirochetes (##FIG##1##\nFig. 2A and B\n##), transient interactions were reduced by 94% (##FIG##1##\nFig. 2C\n##). Furthermore, non-infectious <italic>B. burgdorferi</italic> did not form many dragging interactions and no detectable stationary adhesions (##FIG##1##\nFig. 2B and C\n##). Non-infectious spirochetes were never observed escaping the microvasculature. These observations indicated that early-stage interaction events were essential for sustained association and vascular escape. These observations also demonstrated that microvascular interactions were dependent on <italic>B. burgdorferi</italic> proteins expressed only in the infectious strain.</p>", "<p>Many bacterially-encoded proteins interact with host cells via GAGs ##REF##11971262##[29]##, and a number of previous <italic>in vitro</italic> studies performed under static conditions have found that <italic>B. burgdorferi</italic> can bind to GAGs and that exogenously applied GAGs can competitively inhibit interaction of <italic>B. burgdorferi</italic> with cell monolayers ##REF##8113413##[13]##,##REF##9573074##[15]##,##REF##7532628##[30]##,##REF##9488387##[31]##. Therefore, we investigated the potential role of endothelial host cell GAGs in spirochete microvascular association <italic>in vivo</italic>. Interaction rates were first examined in the presence and absence of a therapeutic low molecular weight heparin compound, dalteparin (Fragmin, average molecular weight 5 kDa). Dalteparin was used at a concentration previously shown to block leukocyte rolling <italic>in vivo</italic>\n##UREF##0##[32]##. Dalteparin (200 µl of a 25 I.U./µl solution) was injected via the femoral vein 15 minutes before intravenous inoculation with infectious spirochetes (see <xref ref-type=\"sec\" rid=\"s4\">\n<bold>Materials and Methods</bold>\n</xref>). As shown in ##FIG##2##\nFig. 3A\n##, dalteparin treatment did not cause any significant change in transient interactions between fluorescent, infectious <italic>B. burgdorferi</italic> and the vascular endothelium. However, dragging interactions were significantly reduced by 72% (##FIG##2##\nFig. 3B\n##) while the number of stationary adhesions were also reduced by dalteparin treatment to a similar extent (##FIG##2##\nFig.\n3C\n##, 76% of controls).</p>", "<p>Similar experiments were performed with dextran sulphate (##FIG##2##\nFig. 3D–F\n##), a 500 kDa high molecular weight GAG analogue, which interacts <italic>in vitro</italic> with infectious but not non-infectious <italic>B. burgdorferi</italic>\n##REF##7532628##[30]##. In these experiments, dextran sulfate was incubated with infectious <italic>B. burgdorferi</italic> for 30 minutes, followed by extensive washing, prior to <italic>B. burgdorferi</italic> administration to the animal, as this compound was toxic when injected directly into the mouse bloodstream. The concentration of dextran sulfate used in preincubations (20 µg/ml) was the dose that has been previously shown <italic>in vitro</italic> to maximally inhibit <italic>B. burgdorferi</italic> interaction with endothelial cells without altering spirochete morphology or motility ##REF##7532628##[30]##,##REF##9488387##[31]##. Preincubation of infectious <italic>B. burgdorferi</italic> with dextran sulfate caused a slight (30%) reduction in the number of transient interactions, but dragging interactions were reduced by 80% (##FIG##2##\nFig. 3D and E\n##). A similar reduction in the number of stationary adhesions was also observed (##FIG##2##\nFig. 3F\n##).</p>", "<p>The results from the dalteparin and dextran sulphate experiments indicated that host GAGs play an important role in dragging interactions between <italic>B. burgdorferi</italic> and the microvascular endothelium <italic>in vivo</italic>, and that competition with a high molecular weight GAG analogue (dextran sulphate) also inhibited transient interactions. The similar levels of inhibition of dragging interactions and stationary adhesions caused by treatment with GAGs suggested that reductions in stationary adhesion were the result of inhibition of dragging. This in turn implied that additional host and spirochete molecules might contribute to stationary adhesion. However, the results of these experiments alone did not rule out the possibility that GAGs played a role in stationary adhesion.</p>", "<title>The <italic>B. burgdorferi</italic> GAG- and Fn-binding protein BBK32 is sufficient for transient and dragging microvascular interactions <italic>in vivo</italic>\n</title>", "<p>Many bacterial adhesins can interact with GAGs, either directly, or indirectly via their association with host molecules such as fibronectin, regulators of the complement cascade and components of the coagulation system; furthermore, GAGs can act as bridging molecules that facilitate interactions between pathogen adhesins and host receptors ##REF##11971262##[29]##. In an effort to identify spirochete adhesins mediating GAG-dependent microvascular interactions <italic>in vivo</italic>, we PCR-amplified and sequenced all candidate <italic>B. burgdorferi</italic> adhesin genes identified to date ##REF##9988477##[16]##, ##REF##10594819##[17]##, ##REF##17822440##[19]##, ##REF##17784908##[21]##–##REF##16368999##[25]##, using genomic DNA extracted from our fluorescent infectious and non-infectious strains (##SUPPL##0##\nTable S1\n##). This approach indicated that the genes encoding BBK32, VlsE, OspF, ErpL and ErpK were absent or mutated in the non-infectious strain (##SUPPL##0##\nTable S1\n##). It is possible that VlsE, OspF, ErpL and ErpK could mediate GAG-dependent host interactions directly or through recruitment of host molecules such as complement cascade regulators; however, interaction of these proteins with GAGs has not been directly demonstrated. In contrast, BBK32 has recently been shown to bind to host GAGs and to rescue the ability of non-infectious <italic>B. burgdorferi</italic> to interact with endothelial cells <italic>in vitro</italic>\n##REF##16368999##[25]##. It was, therefore, of interest to investigate a possible role for BBK32 in <italic>B. burgdorferi</italic> interactions with the microvasculature in the living mouse. The <italic>bbk32</italic> coding sequence, under the control of the <italic>ospC</italic> promoter ##REF##16368999##[25]##, was cloned into the GFP expression construct and the resulting plasmid was used to transform the parental non-infectious <italic>B. burgdorferi</italic> strain. Both parental and complemented strains had the same endogenous plasmid content (data not shown). Expression of BBK32 in the complemented strain was lower than the expression observed in the infectious strain (8.0−/+1.8%), but even this reduced expression was sufficient to restore transient and dragging interactions to the level observed with the infectious strain (##FIG##3##\nFig. 4A and B\n##). However, stationary adhesion rates in the <italic>bbk32</italic> complementation strain did not reach the same levels as in the infectious strain (##FIG##3##\nFig. 4C\n##), implying either a greater dependence upon BBK32 or a dependence upon additional spirochete factors that were missing in the complemented non-infectious strain. Attempts to genetically disrupt the <italic>bbk32</italic> locus in the infectious strain were successful, but did not result in usable constructs due to loss of endogenous plasmids (lp28-1 and others) in all recovered strains.</p>", "<p>PCR amplification and sequencing of all candidate <italic>B. burgdorferi</italic> adhesin genes identified to date also indicated that other candidate adhesin genes (<italic>bbf32, bbk2.10, bbO39 and bbm38</italic>, encoding VlsE, OspF, ErpL and ErpK, respectively) were absent or mutated in the non-infectious parental strain (##SUPPL##0##\nTable S1\n##). Hence, these genes were not essential for transient and dragging interactions with the microvasculature of murine skin <italic>in vivo</italic>, since expression of BBK32 alone in this strain was sufficient to restore transient and dragging interactions. However, the possibility still exists that some of these genes play a role in stationary adhesion.</p>", "<p>Examination of the sequence, expression and localization of two other major adhesins, P66 and Bgp, which have been shown to associate with integrins and GAGs respectively under static conditions <italic>in vitro</italic>, indicated that these proteins were expressed and localized similarly in non-infectious and infectious strains, and were not mutated (##FIG##3##\nFig. 4D\n##; ##SUPPL##0##\nTable S1\n##). Therefore, neither P66 nor Bgp expression nor localization was sufficient for transient or dragging interactions in the absence of BBK32 expression. Although the genomic sequence of the P66- and Bgp-encoding genes was identical <italic>in</italic> both infectious and non-infectious strains, it remains possible that secondary mutations elsewhere in the genome of non-infectious <italic>B. burgdorferi</italic> could have negatively affected transient and dragging interactions.</p>", "<title>Plasma Fn is essential for <italic>B. burgdorferi</italic> transient and dragging microvascular interactions <italic>in vivo</italic>\n</title>", "<p>Because BBK32 binds Fn in addition to GAGs, we also investigated a possible role for Fn in the adhesion of <italic>B. burgdorferi</italic> to the endothelium <italic>in vivo</italic>. Rabbit serum, which contains fibronectin, is an important component of the BSK-II medium used to propagate <italic>B. burgdorferi</italic>; therefore, we investigated whether antibodies to rabbit plasma Fn could disrupt <italic>B. burgdorferi</italic> microvascular interactions <italic>in vivo</italic>. Anti-Fn IgGs did not alter spirochete morphology or motility <italic>in vitro</italic>, implying that they were not toxic to <italic>B. burgdorferi</italic>. The tethering, dragging and stationary interactions/min for infectious <italic>B. burgdorferi</italic> treated with αFn IgGs were compared to the interaction rates of untreated spirochetes, and of spirochetes treated with nonspecific goat IgGs (##FIG##4##\nFig. 5\n##). Preincubation of infectious spirochetes for 20 minutes with the IgG fraction of goat antiserum to rabbit plasma Fn, together with intravenous injection of this IgG fraction into the blood stream of mice, reduced transient and dragging microvascular interactions by 92% and 99%, respectively (##FIG##4##\nFig. 5A and B\n##). When the same treatment regimen was performed using nonspecific goat IgGs, no effect on interaction rates was observed, indicating that the reduction in interactions following treatment with αFn IgGs was specific. Although stationary adhesions were essentially abolished by the αFn treatment (##FIG##4##\nFig. 5C\n##), the reduction in transient and dragging interaction rates was so great that we could not determine if stationary adhesion rates were specifically affected by treatment with anti-Fn IgGs. Interestingly, interaction rates returned to normal levels 15–20 minutes after injection of the spirochetes and antibody (data not shown), suggesting that antibody-blocked rabbit Fn bound to spirochetes might have been replaced by mouse Fn <italic>in vivo</italic>, thus restoring microvascular interactions. The long population doubling time of <italic>B. burgdorferi</italic> (6–8h) precludes the possibility that restored interaction rates were caused by spirochete replication. The dramatic reduction in transient and dragging interactions resulting from Fn antibody treatment suggested that <italic>B. burgdorferi</italic> exploits host Fn for these interactions with the host microvasculature <italic>in vivo</italic>.</p>", "<title>Fibronectin sequences that bind heparin but not RGD-dependent integrins mediate transient and dragging microvascular interactions <italic>in vivo</italic>\n</title>", "<p>Although experiments performed with anti-Fn IgGs suggested that Fn played a major role in the initiation of microvascular interactions, it was possible that IgG-dependent inhibition was partly a result of factors such as steric hindrance of interactions by bulky IgGs. Therefore, we also investigated the Fn dependence of interactions using Fn peptides. Fn is a structurally and functionally complex molecule (reviewed in ##REF##12244123##[33]##). Briefly, the N-terminal Type I Fn repeats and gelatin-binding region interact with Fn-binding proteins from <italic>B. burgdorferi</italic>, <italic>Staphylococci</italic> and <italic>Streptococci in vitro</italic>\n##REF##9988477##[16]##,##REF##15101971##[34]##,##REF##15737988##[35]##. The central cell-binding domain contains multiple integrin-binding sites, including the canonical RGD sequence, which binds to most integrins that have been implicated in <italic>B. burgdorferi</italic>-host cell interactions to date ##REF##10594819##[17]##,##REF##17822440##[19]##,##REF##12244123##[33]##. Finally, the Fn C-terminus contains a high affinity heparin-binding domain that also interacts with host cell GAGs ##REF##12244123##[33]##.</p>", "<p>To investigate endothelial cell molecules associating with spirochete-bound Fn, we used peptides derived from the C-terminal heparin-binding domain and the integrin-interacting cell-binding domain in an attempt to block <italic>B. burgdorferi-</italic>microvascular interactions <italic>in vivo</italic> (##FIG##5##\nFig. 6\n##). The heparin domain peptide (FN-C/H II: KNNQKSEPLIGRKKT) inhibits Fn-mediated cell adhesion and heparan sulfate binding ##REF##8340411##[36]##, and the G<underline>RGD</underline>S cell-binding domain peptide is a well-studied competitive antagonist of integrin binding ##REF##7522656##[37]## that also inhibits <italic>B. burgdorferi</italic> interactions with integrins α_IIbβ₃, α_vβ₃ and α₅β₁\n<italic>in vitro</italic>\n##REF##9573074##[15]##. Peptides were injected via the femoral vein immediately before inoculation with infectious spirochetes, at concentrations (∼50 µg/ml of circulating blood) that disrupt leukocyte adhesion and recruitment <italic>in vivo</italic>\n##REF##8040319##[38]##. Microvascular interactions in dermal postcapillary venules were recorded for no longer than 20 minutes after injection of peptide as the effect of peptide treatment on interaction rates was diminished at later time points, presumably because linear peptides are rapidly cleared from the mouse circulation ##UREF##1##[39]##.</p>", "<p>Intravenous injection of 100 µg of the heparin-binding domain peptide reduced transient interaction rates by 52%, and impaired both dragging interactions and stationary adhesion levels by 84%, confirming the role of GAGs in early stages of microvascular interaction (##FIG##5##\nFig. 6A–C\n##). In contrast, competition with the RGD peptide did not significantly inhibit any class of interaction (##FIG##5##\nFig. 6D–F\n##), even though the estimated final concentration of this peptide in the mouse circulation (100 µM) was twice as high as the dose known to reduce <italic>in vitro B. burgdorferi</italic>-integrin interactions by at least 75% <italic>in vitro</italic>\n##REF##9573074##[15]##. Administration of twice as much RGD peptide (∼200 µM final concentration) did not inhibit interactions, nor did intravenous administration of anti-CD41 and CD49e antibodies that respectively target RGD-dependent <italic>B. burgdorferi</italic>-interacting integrins containing α_IIb or α₅ chains (platelet glycoprotein α_IIbβ₃ and the α₅β₁ Fn receptor; data not shown). The effect of treatment with antibodies to α_vβ₃ was not examined; however, since treatment with RGD peptide <italic>in vitro</italic> has been shown to strongly inhibit <italic>B. burgdorferi</italic> binding to this integrin as well as glycoprotein α_IIbβ₃ and integrin α₅β₁, it seems unlikely that integrin α_vβ₃ mediated the early stages of microvascular interactions. The conclusion that RGD-dependent integrins are not required for microvascular recruitment is consistent with the localization of known <italic>B. burgdorferi</italic>-associating RGD-dependent integrins, which are found at sites of endothelial attachment to extracellular matrix, and not in the lumen ##REF##9573074##[15]##. Collectively, these results implied that Fn-dependent transient and dragging interactions <italic>in vivo</italic> were mediated by host GAGs and not by RGD-dependent integrin interactions.</p>" ]

[ "<title>Discussion</title>", "<p>In this study we employed intravital microscopy, a live cell imaging technique commonly used to analyze leukocyte recruitment and tumor dissemination <italic>in situ</italic>\n##REF##15245733##[40]##,##REF##12001988##[41]##, to investigate the molecular basis of <italic>B. burgdorferi</italic> dissemination <italic>in vivo</italic>. Our results demonstrate that IVM can provide critical insight into the mechanisms of pathogen dissemination. ##FIG##6##\nFig. 7\n## provides a summary of the features of <italic>B. burgdorferi</italic> dissemination which we have identified using IVM, based on data described in this study and in a recent companion report ##REF##18566656##[9]##.</p>", "<p>This study revealed pivotal roles for the <italic>B. burgdorferi</italic> adhesin BBK32 as well as host GAGs and fibronectin in the initiation of spirochete-microvascular interactions (see below) (##FIG##6##\nFig. 7\n##). Although other host and <italic>B. burgdorferi</italic> molecules may also contribute to microvascular interactions <italic>in vivo</italic>, the involvement of GAGs and Fn in this process is especially interesting, since a broad array of pathogens are known to interact <italic>in vitro</italic> with these host molecules in direct binding assays and tissue culture models (reviewed in ##REF##11971262##[29]##, ##REF##16782385##[42]##–##REF##17906137##[45]##. However, the potential contribution of these interactions to processes such as hematogenous dissemination has not been directly examined in living hosts.</p>", "<title>Evidence for mechanistically distinct stages in <italic>B. burgdorferi</italic> microvascular dissemination <italic>in vivo</italic>\n</title>", "<p>We found that BBK32 and its host ligands Fn and GAGs played major roles in transient and dragging interactions. Although we cannot rule out the possibility that these molecules also contribute to stationary adhesion, the results of the <italic>bbk32</italic> complementation experiments indicate that additional spirochete molecules are likely required for stationary adhesion. This implies that stationary adhesion is a mechanistically distinct step in <italic>B. burgdorferi</italic> dissemination. This conclusion is supported by our previous observations that: 1) stationary adhesions form primarily at endothelial junctions, whereas short-term interactions occur chiefly on endothelial cells themselves; and 2) stationary adhesions associate more intimately with the endothelium than short-term interactions and appear to traverse the surface of the endothelium when these cells are labeled with PECAM-1 (##FIG##6##\nFig. 7\n##) ##REF##18566656##[9]##. These observations imply that <italic>B. burgdorferi</italic> dissemination shares functional similarities with the sequence of events that constitute the leukocyte recruitment cascade ##REF##16217160##[46]##,##REF##16917509##[47]##, as well as the events associated with dissemination of circulating tumor cells ##REF##17498748##[48]##. Leukocyte recruitment is initiated by selectin-mediated tethering and rolling interactions that permit firm adhesion, which is mediated by integrins. The initiation phase of leukocyte recruitment is a rate-limiting step, as it is essential for all subsequent events in the recruitment cascade. Similarly, we propose that transient and dragging interactions mediated by GAGs and Fn together constitute the corresponding initiation phase of <italic>B. burgdorferi</italic> dissemination, while other host and spirochete molecules become essential at the stationary adhesion phase.</p>", "<p>Though our data indicated that transient and dragging associations were mediated by the same host and spirochete molecules, the observation that the low molecular weight heparin dalteparin inhibited only dragging interactions was surprising. The reason for this is currently unknown, but may result from differences in total charge, chain length and chemical composition of the carbohydrate moieties.</p>", "<title>A role for Fn, GAGs and BBK32 in the initiation of <italic>B. burgdorferi</italic> microvascular interactions</title>", "<p>This study identified a central role for the <italic>B. burgdorferi</italic> protein BBK32, host GAGs and Fn in the initiation of microvascular interactions. This observation was unexpected, since previous studies have shown that genetic disruption of <italic>bbk32</italic> attenuates but does not abolish infectivity ##REF##16468997##[49]##,##REF##16714558##[50]##. However, <italic>bbk32</italic> disruption mutants still bind Fn ##REF##16468997##[49]##,##REF##16714558##[50]##, implying that other functionally redundant Fn-binding proteins in <italic>B. burgdorferi</italic> might also mediate the initiation of dissemination. The simplest interpretation of our data is that initiation is mediated by BBK32 interactions with GAGs, either independently or via a fibronectin bridge. It is possible that initiation might also be mediated by RGD-independent integrins such as α₃β₁, which interacts with Fn, GAGs and the <italic>B. burgdorferi</italic> protein BBB07 ##REF##16785564##[18]##,##REF##17822440##[19]##; however, this integrin is expressed at endothelial junctions ##REF##9566977##[51]## implying that it is more likely to mediate stationary adhesion or extravasation than initiation interactions. Furthermore, the activation of adhesive properties by endothelial integrins generally requires endothelial activation ##REF##16217160##[46]##, which is not detected in the short time frame of our experiments ##REF##18566656##[9]##. Taken together, these data make it unlikely that integrins play a role in the initiation of vascular adhesion.</p>", "<p>All molecules to date implicated in tethering under shear force conditions (selectins, von Willebrand factor the <italic>E. coli</italic> FimH adhesin) interact with sugar-containing ligands ##REF##17000873##[52]##,##REF##16369083##[53]##, suggesting that Fn-dependent or -independent interactions between BBK32 and GAGs might promote <italic>B. burgdorferi</italic> tethering by a similar mechanism. The affinity of BBK32 for GAGs is unknown, but in the absence of shear forces BBK32 associates with high specificity and probable high affinity to the Fn N-terminus via a tandem β-zipper mechanism shared with Fn-binding proteins of <italic>Staphylococcus aureus</italic> and <italic>Streptococcus pyogenes</italic>\n##REF##15101971##[34]##,##REF##15737988##[35]##,##REF##12736686##[54]##,##REF##15292204##[55]##. In the absence of shear forces, the affinity of plasma Fn for heparin (K_d = 0.1–1.0 µM) is within the affinity range of P- and E-selectins for their ligands (K_d = 1.5 µM and 109 µm, respectively) ##REF##10671486##[56]##,##REF##18250165##[57]##, suggesting that BBK32-, GAG- and Fn-dependent initiation interactions may be mechanistically feasible.</p>", "<p>Although under shear stress conditions Fn does not bind to the leukocyte Fn receptor VLA-4, which mediates tethering to endothelial VCAM-1 under flow ##REF##9160691##[58]##, previous reports indicate that both platelets and <italic>Mycobacterium tuberculosis</italic> can bind to immobilized Fn <italic>in vitro</italic> under shear stress conditions that mimic those found in postcapillary venules ##REF##16714591##[59]##,##REF##7949128##[60]##; interestingly, platelet-Fn interactions are almost completely blocked by treatment with unfractionated or high molecular weight heparin ##REF##7949128##[60]##. This suggests the possibility that Fn-dependent tethering interactions entail cooperative GAG binding, a conclusion that is consistent with our observation that expression of the Fn- and GAG-binding BBK32 protein was sufficient to restore initiation interactions to wild-type levels.</p>", "<p>Another possibility is that BBK32-induced conformational changes in Fn might facilitate Fn- and GAG-dependent tethering interactions. This hypothesis stems from recent data from the Höök laboratory indicating that BBK32 binding to Fn induces the formation of superfibronectin (S. Prabhakaran and M. Höök, personal communication), a high molecular weight Fn complex that substantially enhances adhesion of cells to Fn by integrin-dependent and independent mechanisms ##REF##8114919##[61]##. Further analysis of the precise mechanisms underlying BBK32-, Fn- and GAG-dependent dissemination under shear force conditions will be required.</p>", "<p>The results of this study emphasize the importance of directly investigating host-pathogen interactions in a native context where major regulators of interaction such as fluid shear stress are present. The methodology and observations presented here provide the first direct insight into the role of host GAGs, Fn and a <italic>B. burgdorferi</italic> protein that binds both of these host components, in host microvascular interactions <italic>in situ</italic>. These results may have broad-reaching implications for our understanding of processes underlying the dissemination of a variety of other bacterial pathogens that interact with Fn and GAGs.</p>" ]

[]

[ "<p><bold>¤:</bold> Current address: Centre for Inflammatory Diseases, Department of Medicine, Monash University, Victoria, Australia</p>", "<p>Conceived and designed the experiments: MUN TJM GC. Performed the experiments: MUN TJM ARD BM. Analyzed the data: MUN TJM GC. Wrote the paper: MUN TJM PK GC.</p>", "<p>Hematogenous dissemination is important for infection by many bacterial pathogens, but is poorly understood because of the inability to directly observe this process in living hosts at the single cell level. All disseminating pathogens must tether to the host endothelium despite significant shear forces caused by blood flow. However, the molecules that mediate tethering interactions have not been identified for any bacterial pathogen except <italic>E. coli</italic>, which tethers to host cells via a specialized pillus structure that is not found in many pathogens. Furthermore, the mechanisms underlying tethering have never been examined in living hosts. We recently engineered a fluorescent strain of <italic>Borrelia burgdorferi,</italic> the Lyme disease pathogen, and visualized its dissemination from the microvasculature of living mice using intravital microscopy. We found that dissemination was a multistage process that included tethering, dragging, stationary adhesion and extravasation. In the study described here, we used quantitative real-time intravital microscopy to investigate the mechanistic features of the vascular interaction stage of <italic>B. burgdorferi</italic> dissemination. We found that tethering and dragging interactions were mechanistically distinct from stationary adhesion, and constituted the rate-limiting initiation step of microvascular interactions. Surprisingly, initiation was mediated by host Fn and GAGs, and the Fn- and GAG-interacting <italic>B. burgdorferi</italic> protein BBK32. Initiation was also strongly inhibited by the low molecular weight clinical heparin dalteparin. These findings indicate that the initiation of spirochete microvascular interactions is dependent on host ligands known to interact <italic>in vitro</italic> with numerous other bacterial pathogens. This conclusion raises the intriguing possibility that fibronectin and GAG interactions might be a general feature of hematogenous dissemination by other pathogens.</p>", "<title>Author Summary</title>", "<p>Many bacterial pathogens can cause systemic illness by disseminating through the blood to distant target sites. However, hematogenous dissemination is still poorly understood, in part because of an inability to directly observe this process in living hosts in real time and at the level of individual pathogens. We recently engineered a fluorescent strain of <italic>Borrelia burgdorferi,</italic> the Lyme disease pathogen, and visualized its dissemination from the microvasculature of living mice using intravital microscopy. We found that dissemination was a multistage process that included tethering, dragging, stationary adhesion and extravasation. In the study described here, we used quantitative real-time intravital microscopy to investigate the mechanistic features of the vascular interaction stage of <italic>B. burgdorferi</italic> dissemination in living hosts. We found that tethering and dragging interactions (collectively referred to as initiation interactions) were mechanistically distinct from stationary adhesion. Initiation of microvascular interactions required the <italic>B. burgdorferi</italic> protein BBK32, and host ligands fibronectin and glycosaminoglycans. Initiation interactions were also strongly inhibited by the low molecular weight clinical heparin dalteparin. Since numerous bacterial pathogens can interact with fibronectin and glycosaminoglycans <italic>in vitro</italic>, these observations raise the intriguing possibility that fibronectin and glycosaminoglycan recruitment might be a feature of hematogenous dissemination by other pathogens.</p>" ]

[ "<title>Supporting Information</title>" ]

[ "<p>We thank John Leong and Nikhat Parveen for providing pBBK32 and Bgp antibodies, pBBK32 and for helpful advice on use of the antibodies and examination of surface localization of <italic>B. burgdorferi</italic> proteins. We are also grateful to Jenifer Coburn for providing P66 antibody and to Dean Brown, Genevieve Chaconas and Derrice Knight for technical support.</p>" ]

[ "<fig id=\"ppat-1000169-g001\" position=\"float\"><object-id pub-id-type=\"doi\">10.1371/journal.ppat.1000169.g001</object-id><label>Figure 1</label><caption><title>Intravital microscopy to study <italic>B. burgdorferi</italic> microvascular interactions <italic>in vivo.</italic>\n</title><p>A) Table summarizing the properties of spinning disk confocal and conventional epifluorescence intravital microscopy (IVM) with respect to visualization of <italic>B. burgdorferi</italic> vascular interactions <italic>in situ</italic>. B) Spinning disk confocal and C) conventional epifluorescence micrographs of fluorescent infectious <italic>B. burgdorferi</italic> in the skin microvasculature of a living mouse. Blood vessels in b) were visualized using AlexaFluor555-conjugated antibody to PECAM-1 (red). Spinning disk confocal and conventional epifluorescence IVM videos from the microvasculature are presented in ##SUPPL##1##Videos S1## and ##SUPPL##2##S2##, respectively.</p></caption></fig>", "<fig id=\"ppat-1000169-g002\" position=\"float\"><object-id pub-id-type=\"doi\">10.1371/journal.ppat.1000169.g002</object-id><label>Figure 2</label><caption><title>Interaction of infectious and non-infectious fluorescent <italic>B. burgdorferi</italic> with the microvasculature of a living murine host.</title><p>Micrographs of infectious (A) and non-infectious (B) GFP-expressing <italic>B. burgdorferi</italic> visualized on slides by epifluorescence microscopy. (C) Graphical summary of infectious and non-infectious <italic>B. burgdorferi</italic> interactions in postcapillary venules of the skin microvasculature. The number of interactions/minute in each interaction class was determined by measuring interactions from video footage of conventional epifluorescence IVM. The percentages above the bars for non-infectious spirochetes indicate the non-infectious interaction rate expressed as a percentage of the interaction rate of infectious spirochetes. Sample footage of the videos used to measure spirochete interactions is presented in ##SUPPL##2##Video S2##. A total of 8,343 spirochete interactions in 85 venules from 17 mice (n = 8 and n = 9, respectively, for experiments performed with non-infectious and infectious <italic>B. burgdorferi</italic>) were analyzed. Standard error bars are indicated for each interaction class. P-values for this figure and all others were determined using a two-tailed non-parametric Student's t-test. Microvascular interactions were measured between 5 and 45 minutes after spirochete injection. The average time after injection for all recordings made with infectious and non-infectious <italic>B. burgdorferi</italic> was 15.9−/+8.1 and 14.1−/+9.3 min, respectively.</p></caption></fig>", "<fig id=\"ppat-1000169-g003\" position=\"float\"><object-id pub-id-type=\"doi\">10.1371/journal.ppat.1000169.g003</object-id><label>Figure 3</label><caption><title>Role of host GAGs in spirochete microvascular interactions <italic>in vivo</italic>.</title><p>The role of host GAGs was examined by conventional IVM performed using infectious <italic>B. burgdorferi</italic> preincubated with dextran sulfate (D–F) or with mice pre-injected with dalteparin, a therapeutic heparin compound (A–C). The percentages above the bars for dalteparin- or dextran sulfate-treated spirochetes indicate the treatment group interaction rate expressed as a percentage of the interaction rate of untreated spirochetes. For the dextran sulfate experiments, a total of 7,987 interactions in 96 venules from 14 mice (n = 8 and n = 6, respectively, for experiments performed with untreated and dextran sulfate-treated <italic>B. burgdorferi</italic>) were analyzed. Microvascular interactions were measured between 5 and 45 minutes after spirochete injection. Average time after injection: untreated (17.8−/+9.8 min), treated (18.1−/+9.5 min). For the dalteparin experiments, a total of 13,951 interactions in 77 venules from 13 mice (n = 7 and n = 6, respectively, for experiments performed with untreated and dalteparin-treated mice) were analyzed. Microvascular interactions were measured between 5 and 45 minutes after spirochete injection. Average time after injection: untreated (15.7−/+6.8 min), treated (16.3−/+7.1 min).</p></caption></fig>", "<fig id=\"ppat-1000169-g004\" position=\"float\"><object-id pub-id-type=\"doi\">10.1371/journal.ppat.1000169.g004</object-id><label>Figure 4</label><caption><title>Role of <italic>B. burgdorferi</italic> GAG- and Fn-binding protein BBK32 in spirochete-microvascular interactions <italic>in vivo</italic>.</title><p>(A–C) Interaction rates of three <italic>B. burgdorferi</italic> strains, infectious, non-infectious and the non-infectious strain with GFP and BBK32 expressed from the same plasmid (<italic>bbk32</italic> knock-in), as analyzed using conventional IVM. A total of 19,380 interactions in 81 venules from 16 mice (n = 9 infectious; n = 3 non-infectious; n = 4 non-infectious+<italic>bbk32</italic>) were analyzed. Microvascular interactions were measured between 5 and 45 minutes after spirochete injection. Average time after injection: infectious (13.4−/+9.9 min), non-infectious (12.1−/+4.8 min), <italic>bbk32</italic> knock-in (13.0−/+5.8 min). D) Expression and surface localization of adhesins P66 and Bgp in non-infectious and infectious strains, as determined by immunoblotting. The top panel shows total protein loaded for each strain, detected by Coomassie staining of the SDS-PAGE gel. To identify cell surface localized proteins, cell pellets were incubated in the presence (+) or absence (−) of proteinase K (ProK) before lysis. Proteinase K treatment resulted in a dramatic decrease in the level of P66 in both the infectious and non-infectious strains used here. Bgp showed a lesser, but similar reduction in both the infectious and non-infectious strain following treatment with Proteinase K.</p></caption></fig>", "<fig id=\"ppat-1000169-g005\" position=\"float\"><object-id pub-id-type=\"doi\">10.1371/journal.ppat.1000169.g005</object-id><label>Figure 5</label><caption><title>Role of host Fn in spirochete-microvascular interactions <italic>in vivo</italic>.</title><p>(A–C) The role of host Fn in microvascular interactions was examined by conventional IVM performed using infectious <italic>B. burgdorferi</italic> preincubated with goat IgGs against rabbit Fn (αFn) or with non-specific goat IgGs. Non-specific or αFn IgGs were also injected directly into the mouse blood stream immediately before inoculation with spirochetes, and the effect of antibody on <italic>B. burgdorferi</italic> interactions was measured for up to 20 minutes after spirochete injection. The percentages above the bars for Fn antibody treatments indicate the interaction rate expressed as a percentage of the interaction rate of untreated spirochetes. A total of 2,614 interactions in 49 venules from 12 mice (n = 4 each for experimental group) were analyzed. Average time after injection: untreated (9.4−/+4.4 min), IgG (10.4−/+4.9 min), αFn (10.5−/+4.8 min).</p></caption></fig>", "<fig id=\"ppat-1000169-g006\" position=\"float\"><object-id pub-id-type=\"doi\">10.1371/journal.ppat.1000169.g006</object-id><label>Figure 6</label><caption><title>Contributions of GAG-binding sequences and integrin-binding sequences in Fn to spirochete-microvascular interactions <italic>in vivo</italic>.</title><p>The roles of host glycosaminoglycans (GAGs) and integrins in Fn-mediated spirochete interactions with the microvasculature <italic>in vivo</italic> were examined by injecting mice with peptides corresponding, respectively, to a portion of the Fn heparin-binding domain (A–C) or the Fn RGD sequence (D–F). Microvascular interactions of infectious <italic>B. burgdorferi</italic> were examined by conventional IVM, for up to 20 minutes after intravenous inoculation of spirochetes. The percentages above the bars for peptide treatments indicate the interaction rate expressed as a percentage of the interaction rate of untreated spirochetes. For the heparin-binding Fn peptide experiments, a total of 8,346 interactions in 64 venules from 11 mice (n = 5 and n = 6, respectively, for experiments performed with untreated and peptide-treated mice) were analyzed. Average time after injection: untreated (10.8−/+3.6 min), treated (10.0−/+4.2 min). For the integrin-blocking Fn RGD peptide experiments, a total of 22,094 interactions in 89 venules from 16 mice (n = 9 and n = 7, respectively, for experiments performed with untreated and peptide-treated mice) were analyzed. Average time after injection: untreated (13.2−/+9.7 min), treated (12.4−/+6.0 min).</p></caption></fig>", "<fig id=\"ppat-1000169-g007\" position=\"float\"><object-id pub-id-type=\"doi\">10.1371/journal.ppat.1000169.g007</object-id><label>Figure 7</label><caption><title>Schematic summarizing the stages of infectious <italic>B. burgdorferi</italic> interaction with and escape from the microvasculature.</title><p>Based upon our previous work ##REF##18566656##[9]## and this study, we propose that transient and dragging interactions together constitute the essential first step (initiation) of microvascular interactions. We also propose that initiation interactions are mechanistically distinct from downstream interaction events (described in our recent paper ##REF##18566656##[9]##) for two reasons: 1) stationary adhesions and transmigrating spirochetes localize to different sites on the endothelium than transient and dragging interactions, and 2) stationary adhesion appears to require host and/or spirochete molecules in addition to or other than BBK32, GAGs and Fn. The results reported in this study do not indicate whether BBK32-mediated initiation events are entirely dependent on BBK32-GAG interactions bridged by Fn, or whether direct BBK32-GAG interactions also contribute to initiation.</p></caption></fig>" ]

[]

[]

[]

[]

[]

[]

[ "<supplementary-material content-type=\"local-data\" id=\"ppat.1000169.s001\"><label>Table S1</label><caption><p>PCR amplification and sequencing of candidate <italic>B. burgdorferi</italic> adhesin genes.</p><p>(0.05 MB DOC)</p></caption></supplementary-material>", "<supplementary-material content-type=\"local-data\" id=\"ppat.1000169.s002\"><label>Video S1</label><caption><p>Spinning disk confocal IVM video footage of fluorescent <italic>B. burgdorferi</italic> interacting with a postcapillary venule of the skin vasculature. Elapsed time is shown at the top right and the scale at bottom left. Direction of blood flow is down and to the left.</p><p>(1.1 MB SWF)</p></caption></supplementary-material>", "<supplementary-material content-type=\"local-data\" id=\"ppat.1000169.s003\"><label>Video S2</label><caption><p>Conventional epifluorescence IVM video footage of fluorescent <italic>B. burgdorferi</italic> interacting with a postcapillary venule of the skin vasculature. The video is shown in real time (time indicated at the bottom). Blood flow direction is to the right and up.</p><p>(1.9 MB SWF)</p></caption></supplementary-material>" ]

[ "<fn-group><fn fn-type=\"COI-statement\"><p>The authors have declared that no competing interests exist.</p></fn><fn fn-type=\"financial-disclosure\"><p>This work was supported by grants from the Canadian Institutes of Health Research (CIHR) to P.K. and G.C. (MOP-53086), and by a CIHR group grant. P.K. and G.C. are Scientists of the Alberta Heritage Foundation for Medical Research (AHFMR), and Canada Research Chairs in, respectively, Leukocyte Recruitment in Inflammatory Disease, and Molecular Biology of Lyme Disease. M.U.N. and T.J.M. were each supported by postdoctoral fellowships from CIHR and AHFMR, and M.U.N is an Australian NHMRC CJ Martin Fellow (284394).</p></fn></fn-group>" ]

[ "<graphic xlink:href=\"ppat.1000169.g001\"/>", "<graphic xlink:href=\"ppat.1000169.g002\"/>", "<graphic xlink:href=\"ppat.1000169.g003\"/>", "<graphic xlink:href=\"ppat.1000169.g004\"/>", "<graphic xlink:href=\"ppat.1000169.g005\"/>", "<graphic xlink:href=\"ppat.1000169.g006\"/>", "<graphic xlink:href=\"ppat.1000169.g007\"/>" ]

[ "<media xlink:href=\"ppat.1000169.s001.doc\"><caption><p>Click here for additional data file.</p></caption></media>", "<media xlink:href=\"ppat.1000169.s002.swf\"><caption><p>Click here for additional data file.</p></caption></media>", "<media xlink:href=\"ppat.1000169.s003.swf\"><caption><p>Click here for additional data file.</p></caption></media>" ]

{ "acronym": [], "definition": [] }

[ "<title>Introduction</title>", "<p>Innate immunity is the first line of defense in multicellular organisms, and effectively prevents or limits infection after exposure to microbes ##REF##10334979##[1]##. The innate immune response to microbes triggers diverse humoral and cellular activities via signal transduction pathways that exhibit transphyletic conservation in animals ##REF##10652463##[2]##–##REF##12495727##[4]##. In mammals, the adaptive immune system is recruited for complete elimination of microbes or microbial debris after initial neutralization or clearance by the innate immune system. However, <italic>Drosophila</italic> relies on humoral and cellular innate immune responses for protection against the barrage of microbes that thrive in its habitats ##REF##11834378##[3]##–##REF##11812988##[6]##.</p>", "<p>A hallmark of the humoral response in <italic>Drosophila</italic> is the massive synthesis of antimicrobial peptides (AMPs) after immune challenge. AMPs are produced primarily by the fat body, the anatomical equivalent of the mammalian liver, and are secreted into the hemolymph where they directly kill invading microorganisms ##REF##11812988##[6]##. Genetic analysis has shown that AMP genes are regulated by various immunogenes through the Toll and Imd pathways ##REF##11834378##[3]##,##REF##11812988##[6]##. The Toll pathway is activated by both Gram-positive bacteria and fungi. Recognition of microbial components triggers proteolytic cleavage of the Toll ligand Spatzle (Spz) leading to activation of the Rel proteins, Dif and Dorsal ##REF##9405661##[7]##–##REF##11823479##[10]##. In contrast, the Imd pathway mainly responds to Gram-negative bacteria and controls the expression of specific AMP genes by activating Relish ##REF##11269502##[9]##,##REF##10619029##[11]##,##REF##11485985##[12]##.</p>", "<p>In addition to strong antimicrobial activities provided by the humoral response, cell-mediated defenses also play an important role in the elimination of apoptosed cells and invading microbes or parasites ##REF##8630729##[13]##–##REF##1499832##[18]##. The <italic>Drosophila</italic> hemocyte population consists of three cell types: plasmatocytes, crystal cells, and lamellocytes ##UREF##1##[19]##,##REF##14602069##[20]##. Plasmatocytes represent 90–95% of all mature <italic>Drosophila</italic> hemocytes and function in the phagocytic removal of dead cells and microbial pathogens ##UREF##0##[15]##,##REF##7924990##[16]##. Crystal cells, which constitute approximately 5% of the hemocyte population, are non-phagocytic cells that facilitate innate immune responses and promote wound healing through the process of melanization ##UREF##0##[15]##,##REF##11161576##[17]##,##REF##8587797##[21]##. Lamellocytes are relatively large (15–40 µm), flat, adherent cells that facilitate the encapsulation and neutralization of objects too large to be engulfed by plasmatocytes ##REF##1499832##[18]##. These hemocytes are activated by microbial molecules through the same pattern recognition receptors as in the fat body, but the mechanisms leading to the activation of cellular immune responses are not fully understood.</p>", "<p>Significant effort has focused on identifying components of the signaling pathways involved in regulating the innate immune response. Previous studies have identified a number of genes that are differentially regulated in hematocytes during microbial infection ##REF##15679837##[22]##,##REF##15777795##[23]##. However, the role of these genes in the immune response is only known for a few of them. To evaluate the role of these genes in antifungal immune responses, we examined the effect of individual mutations on the immune response of flies against <italic>Beauveria bassiana</italic> infection, and identified 16 mutants with increased sensitivity to <italic>B. bassiana</italic>. Examination of the sensitivities of these mutants to infection with several types of bacteria identified several mutants that were required mainly for defense against fungal infection. Examination of cellular immune responses revealed that transcription factors involved in chromatin remodeling or lineage specific differentiation were required for proper hemocyte development. Mutation of genes involved in cytoskeletal remodeling caused a strong defect in phagocytosis, while Trx-2 and DDB1 were required for development of functional crystal cells. The screen also identified several novel genes required for activation of antimicrobial peptide genes, indicating their involvement in signaling during pathogen specific immune responses. The distinct requirement of these genes for defense against different microbial infections also reveals the complexity of innate immune responses designed to compete with diverse offensive mechanisms used by microbes. In this paper, we present new findings on the regulation of cellular and humoral immune responses of <italic>Drosophila</italic> against fungal infection.</p>" ]

[ "<title>Materials and Methods</title>", "<title>\n<italic>Drosophila</italic> stocks</title>", "<p>\n<italic>Drosophila melanogaster</italic> strains were cultured on a standard cornmeal-yeast medium at 25°C and 60% humidity. Mutant flies containing a <italic>P</italic>-element at the translated/untranslated region of the candidate genes (##SUPPL##0##Table S1##) were purchased from GenExel (Daejeon Korea). Because the GenExel EP lines contain Gal4 binding sites, overexpression of Gal4 can induce strong expression of adjacent endogenous genes in which an EP element is inserted at the 5′ UTR in a forward orientation ##REF##8901596##[47]##. To activate transcription of <italic>P</italic>-element inserted genes from the EP promoter, we crossed mutant flies containing a <italic>P</italic>-element at the 5′ UTR in a forward orientation with <italic>hs-Gal4</italic> driver (Bloomington Stock Center). For homozygous viable lines, we generated flies carrying a homozygous <italic>P</italic>-element inserted chromosome in addition to a <italic>hs-Gal4</italic> driver. Overexpression of target genes was achieved by heat shocking the adult flies for 1 h at 37°C, and these flies were used for infection one day after a heat shock. <italic>W¹¹¹⁸</italic> was used as a wild type stock and <italic>P</italic>[<italic>ry</italic>+Δ2–3]<italic>sb</italic>/<italic>TM6B</italic>, <italic>TB</italic> was used as a genomic transposase source. The <italic>Imd</italic> and <italic>spz^rm7</italic> were a gift from Dr. Won-Jae Lee, and <italic>Dif ²</italic> was a gift from Dr. Kwang-Min Choe.</p>", "<title>Infection and survival experiments</title>", "<p>\n<italic>Beauveria bassiana</italic> from three day cultures (per 1.0 L distilled water: Dextrose 10 g, Peptone 2.5 g, Yeast extract 5 g, 25°C). <italic>Staphylococcus aureus</italic> (per 1.0 L distilled water: Trypticase soy broth 30 g, 37°C), <italic>Micrococcus luteus</italic> and <italic>Erwinia carotovora carotovar-15</italic> (per 1.0 L distilled water: Beef extract 3.0 g, Peptone 5.0 g pH 6.8, 30°C) from overnight cultures were recovered by centrifuging at 6,000 rpm for 10 min at 25°C. The supernatants were discarded and the pellets were resuspended in corresponding fresh culture media. Septic injury was performed by pricking the leg disc of adult flies with a tungsten needle previously dipped into a concentrated <italic>B. bassiana</italic> or by injecting diluted bacteria (OD = 0.1, 55 nl) into the ventral lateral side with a thin needle using a Picospritzer III injector (Parker Hannifin, USA). Natural infections with <italic>B. bassiana</italic> were performed by shaking anesthetized flies for 60 sec in a Petri dish containing a sporulating fungal culture ##REF##9405661##[7]##. Survival rates of flies after pathogen infection were measured under identical conditions for each genotype tested. Groups of 30 adults, aged 2–4 days, were septically injured, maintained at 25°C, and transferred to a fresh vial every three days. Fewer than five percent of the total flies tested died within three hours after infection and these flies were not considered in the analyses.</p>", "<title>\n<italic>P</italic>-element excision</title>", "<p>Revertants for each <italic>P</italic>-element insertion mutant were generated through precise excision of the <italic>P</italic>-element by crossing with flies containing the Δ2–3 transposase, as described by Robertson et al. ##REF##2835286##[48]##. Excision allele identity was confirmed by PCR and direct sequencing of the excision sites.</p>", "<title>Preparation of genomic DNA and PCR</title>", "<p>Approximately 10–15 adult flies were placed in a 1.5 ml centrifuge tube and frozen in liquid nitrogen for 5 min. The frozen flies were homogenized with a small pestle and genomic DNA was isolated with a G-spin™ Genomic DNA Extraction kit (Intron, Gyeonggi-do, Korea). The oligonucleotide primers used in PCR amplifications, with each sequence shown in 5′ to 3′ orientation, are described in ##SUPPL##2##Table S3##. The standard thermal profile for PCR amplifications was 30 cycles of denaturation at 95°C for 1 min, annealing at 50°C for 1 min, and extension at 72°C for 1 min.</p>", "<title>Quantitative real time PCR</title>", "<p>Adult males were challenged with live <italic>B. bassiana</italic> spores and incubated at 25°C for 6 h. Total RNA was isolated from 8–10 adult flies with TRIzol (Invitrogen, Carlsbad, CA) and used for cDNA synthesis with Superscript II reverse transcriptase (Invitrogen, Carlsbad, CA). Target cDNAs were measured by real time PCR using a LightCycler 480 (Roche, Basel, Switzerland). PCR reactions contained 1×SYBR Green mix (Applied Biosystems, Foster City, CA) and were analyzed with LightCycler 480 software 4 (Roche). All results were normalized to the level of <italic>RpL32</italic> mRNA in each sample. Primers used are shown in ##SUPPL##2##Table S3##.</p>", "<title>\n<italic>In vivo</italic> phagocytosis</title>", "<p>\n<italic>In vivo</italic> phagocytosis assays of adult flies were performed following the procedure of Elrod-Erickson et al. and Brandt et al. ##REF##10898983##[28]##,##REF##15562316##[49]##. Groups of 3–5 day-old adult males were injected with Alexa Fluor 488-labeled heat killed spores of <italic>B. bassiana</italic>, fluorescein conjugated <italic>E. coli</italic> (K-12) BioParticles, and fluorescein conjugated <italic>S. aureus</italic> BioParticles (1 mg/ml, 50–60 nl) (Molecular Probes, Invitrogen) on the ventral lateral side with a thin needle using a Picospritzer III injector. Flies were incubated for 1 h at 25°C to permit phagocytosis of the spores or bacteria, followed by injection of excess Trypan blue (0.4%, 220 nl) to quench extracellular fluorescence. Phagocyte ablation experiments were performed as described by Kocks et al. ##REF##16239149##[50]##. CML latex beads (1.0 µm diameter, Molecular Probes) were washed in PBS and concentrated in PBS to 8% solids. Beads (100 nl) were injected 24 hours before the phagocytosis test.</p>", "<p>Phagocytosis of India ink was observed as described in Rutschmann et al. ##REF##11823479##[10]##. India ink carbon particles (Pébéo, Gemenos, France) (diluted 1/50 in PBS, 90 nl) were injected on the ventral lateral side with a thin needle using a Picospritzer III injector (Parker Hannifin). The phagocytosis of India ink by the sessile blood cells was observed 2 h later. Phagocytosed signals were observed under a Zeiss Axioplan 2 microscope (Zeiss). Fluorescence particles and Indian ink around the dorsal vessel was quantified from raw unaltered pictures using Image J software (NIH, Bethesda, MD). Before the software was used to count the area of particles, each image was converted to a 32-bit grey scale image and was thresholded to highlight the particles. The phagocytic index was expressed as area of the signal corresponding to the sum of the encircled areas.</p>", "<title>Analysis of larval hematopoiesis</title>", "<p>Larvae were staged according to procedures described in Zettervall et al. ##UREF##3##[32]##. Emptying of the gut marks the difference between early- and late-wandering third instar larvae, therefore a red household food dye was added to the food to allow visualization of the gut contents. The six homo-lethal alleles were maintained as heterozygotes balanced with either the second chromosome balancer <italic>CyO</italic> or with the third chromosome balancer <italic>Ubx</italic>. Precisely staged late-wandering third instar larvae were rinsed well in PBS (137 mM NaCl, 2.7 mM KCl, 6.7 mM Na₂HPO₄, and 1.5 mM KH₂PO₄) and blotted on Kimwipes to remove excess PBS before bleeding. The larval cuticle was ripped gently near the posterior end while submerging the larva in 20 µl PBS. The hemocytes were transferred to a Neubauer improved hemocytometer (Marienfeld) to determine plasmocyte number. To quantify crystal cells, late-wandering third instar larvae were heated at 60°C for 10 min in a water bath to induce blackening of mature crystal cells and blackened crystal cells in the last two posterior dorsal segments of third instar larvae were counted under a dissecting microscope. For melanization reactions, third instar larvae were pricked with a clean standard needle and the reaction was observed 2 h after injury. Melanization signals were quantified from raw unaltered pictures using Image Pro Plus 4.5 software (Media Cybernetics, Silver Spring, USA). The melanization index was expressed as [area]×[mean intensity] of encircled areas.</p>" ]

[ "<title>Results</title>", "<title>Screening of immune defective mutant flies</title>", "<p>Previously we identified genes that were differentially induced in SL2 cells after treatment with LPS/PGN or curdlan using <italic>Drosophila</italic> cDNA microarrays ##REF##18460901##[24]##. These LPS/PGN-or curdlan-induced genes are probably involved in diverse immune responses, such as activation of signaling pathways downstream of pathogen associated molecular pattern recognition receptors, induction of phagocytosis, and differentiation into a specialized immune effector cell type. Because these immune responses require crosstalk between different cell types in a physiological condition, expression profile analysis of SL2 cells alone may not provide a complete picture of gene regulation during infection. However, because SL2 cells display important characteristics of macrophages in an in vitro assay, we assumed that their expression pattern may reflect regulatory mechanism underlying some immune responses of macrophages. To identify key regulators of innate immunity, we obtained mutants of the genes that are differentially regulated following treatment with microbial components, and monitored their requirement for defense against infection. Out of 5,405 genes screened on the microarray, 231 and 1,151 genes were induced more than 1.6 fold after the LPS/PGN or curdlan treatment of SL2 cells, respectively. A search for congenic EP (Enhancer-Promoter) lines in which these differentially regulated genes were disrupted by a <italic>P</italic>-element insertion identified 130 lines (110 and 20 lines with a <italic>P</italic>-element inserted at the untranslated and coding regions of the differentially regulated genes, respectively) from the GenExel library (Daejeon, Korea). The <italic>P</italic>-element insertion positions of all the GenExel EP lines were confirmed twice independently by direct sequencing of the inverse PCR fragment amplified with <italic>P</italic>-element specific primers (data not shown). These results suggest that most of the defects associated with the EP lines are related to disruption of the candidate genes. About one-third (47 lines) of the EP lines obtained were homozygous lethal, indicating that the <italic>P</italic>-element insertion effectively disrupted function of the target genes. None of the 83 homozygote viable EP lines showed obvious developmental abnormalities. These results indicated that the EP lines could be used to screen for genes involved specifically in defense against microbial infection. Therefore, adult homozygote flies were screened for survival after infection with entomopathogenic fungi (<italic>B. bassiana</italic>) (##SUPPL##0##Table S1##). Although a developmental defect caused by heterozygocity of a gene is rare, functional insufficiency of a heterozygote is often observed under strong environmental stress such as infection, and can influence survival of the heterozygotes as shown in the study of <italic>Dif ¹</italic> heterozygotes ##REF##10843389##[8]##. Based on this assumption, adult heterozygote flies were monitored for survival after fungal infection in the case of the homozygous lethal lines. To identify EP lines with a compromised defense against fungal infection, 30 adult flies from each of the 130 lines were pricked on the leg disc with a needle dipped into a concentrated solution of live <italic>B. bassiana</italic>, and the survival rate was followed over a six day period at 25°C. The septic infection with <italic>B. bassiana</italic> resulted in approximately 10% mortality in the wild type flies. Under the same infection conditions, most of the mutant flies showed similar levels of survival (##FIG##0##Figure 1A##, ##SUPPL##0##Table S1##). However, 16 mutant lines, including six heterozygote flies (<italic>Pcl</italic>, <italic>DDB1</italic>, <italic>shg</italic>, <italic>Rab6</italic>, <italic>CG6181</italic>, <italic>and CG7263</italic>), were significantly more sensitive to fungal infection (<italic>p</italic>&lt;0.002) (##FIG##0##Figure 1A##). In these cases, death was clearly associated with uncontrolled fungal growth, as the dead flies were covered with fungal hyphae (##SUPPL##3##Figure S1##).</p>", "<p>To confirm the defects of the 16 lines, we first compared their survival rates after fungi infection with wild type and <italic>spz</italic> mutant as negative and positive controls, respectively, in three independent experiments. The repeated experiments revealed that the 16 EP lines had a clear defect in survival (##FIG##0##Figure 1B##). We next examined the survival rates after natural infection with <italic>B. bassiana</italic> to rule out the possibility that reduced viability resulted from septic injury rather than from fungal infection. When the flies were raised after being covered with spores for 1 min, the 16 mutant lines showed remarkably less survival comparable to that of the <italic>spz</italic> mutant, while wild type showed only a minor decrease in survival (##FIG##0##Figure 1C##). This result indicated that we have identified <italic>Drosophila</italic> mutants that have a reduced ability to defend against <italic>B. bassiana</italic> infection.</p>", "<title>Rescue of the mutant phenotype by precise <italic>P</italic>-element excision or by overexpressing the disrupted genes from an EP promoter</title>", "<p>To confirm that the increased sensitivity of these mutants to fungal infection is caused by specific disruption of the candidate genes by the <italic>P</italic>-element, we excised the <italic>P</italic>-element from the mutant flies by crossing with <italic>P</italic>[<italic>ry</italic>\n⁺Δ2–3](99B)<italic>Sb</italic>/<italic>TM6B</italic>, <italic>TB</italic>. After excising the <italic>P</italic>-element from the germ cells, white-eye progeny were established as homozygous lines for all mutants. Excision of the <italic>P</italic>-element in each line was confirmed by PCR with primers specific to one end of the <italic>P</italic>-element (PF) and to target sequences surrounding the <italic>P</italic>-element insertion sites (F and R) (##FIG##1##Figure 2A##). The PF and R primer pairs amplified specific fragments (fragment II) from the <italic>P</italic>-element mutant lines confirming the mutation sites, but failed to amplify this fragment in any of the excised lines. On the other hand, F and R primer pairs that amplify the undisrupted target gene sequences (fragment I) failed to amplify specific fragments from the homozygous mutant flies and produced reduced levels of the amplification products from the heterozygous mutants (<italic>Pcl</italic>, <italic>DDB1</italic>, <italic>shg</italic>, <italic>Rab6</italic>, <italic>CG6181</italic>, <italic>and CG7263</italic>). These PCR primers specifically amplified products from all of the excised lines (##FIG##1##Figure 2B##). To confirm that these excised lines did not contain a small deletion or insertion at the <italic>P</italic>-element insertion sites, we cloned fragment I amplified from each excision line and sequenced them together with fragment II obtained from the corresponding <italic>P</italic>-element insertion mutants. This sequencing analysis confirmed that precise excision lines had been obtained for all mutants except <italic>CG12004</italic> and <italic>CG6181</italic> (data not shown).</p>", "<p>After obtaining the precise excision lines for all mutants, we examined whether excision of the <italic>P</italic>-element from the mutants could revert their sensitivity to fungal infection to that of wild type flies. As shown in ##FIG##1##Figure 2C##, similar survival rates (90%) were observed in both the wild type and the precise excision homozygous lines following fungal infection that caused complete death of the <italic>spz</italic> mutant (##FIG##1##Figure 2C##).</p>", "<p>In addition to rescuing the increased lethality following infection by precise excision of the <italic>P</italic>-element, we tested whether overexpression of disrupted genes with the EP promoter inserted in front of the coding region could reverse the mutant phenotype. Half of the mutants contained a Gal4-dependent promoter (EP element) at the 5′ UTR in a forward orientation to the disrupted gene. Heat shock in combination with an <italic>hs-Gal4</italic> driver induced overexpression of the disrupted gene in this half of the EP mutant lines (##SUPPL##1##Table S2##). Thus, we generated flies carrying a copy of <italic>hs-Gal4</italic> driver and homo- or heterozygous <italic>P</italic>-element insertions, depending on the corresponding mutant configuration used for the screen. Quantitative RT-PCR analysis of the mutant EP lines revealed that the disrupted gene transcript was significantly less than that in the wild type. However, heat shock treatment (1 h at 37°C) in the presence of the <italic>hs-Gal4</italic> driver activated transcription of the target genes above the level observed in wild type flies (##SUPPL##4##Figure S2##). Consistent with this observation, lethality of the mutant lines reverted completely to wild type levels (##FIG##1##Figure 2D##). These results demonstrate that the genes identified from our screen are required for <italic>Drosophila</italic> antifungal immunity.</p>", "<p>These genes identified in our screen encode proteins from many different functional classes including transcription factors involved in chromatin remodeling or lineage specific transcription (<italic>spen</italic>, <italic>Pcl</italic>, <italic>CG12744</italic>, <italic>jumeaux</italic>, <italic>inv</italic>, <italic>and Lmpt</italic>), cytoskeletal regulation (<italic>coro</italic>, <italic>shg</italic>, <italic>loco</italic>, <italic>and Rab6</italic>), DNA fragmentation, apoptosis and redox signaling (<italic>CG7263</italic>, <italic>DDB1 and Trx-2</italic>), along with a few genes (<italic>CG6181</italic>, <italic>CG12004 and JhI-21</italic>) of unknown function. Therefore, genes involved in immune responses ranging from development to cell movement were identified in this fungal defense screen.</p>", "<title>Specificity of genes for antifungal defense</title>", "<p>To determine whether reduced survival of the mutant flies resulted from defective immune responses specifically to fungal infection, we examined the effect of these mutations on wound healing and defense against bacterial infection. When mutant flies were pricked with a sterile tungsten needle, the majority of the flies survived the wounding and only the <italic>spen</italic> mutant showed a minor decrease in survival (##FIG##2##Figure 3A##), suggesting that the reduced survival rates of these mutants, except <italic>spen</italic>, were caused by a defective defense against microbial infection. Septic infection with Gram-negative bacteria does not normally affect the viability of wild type flies. However, loss of a major antibacterial gene, such as <italic>imd</italic>, severely reduces survival following infection with Gram-negative bacteria. When the mutant flies were tested for susceptibility to <italic>Ecc-15</italic> infection, most showed no significant defect in survival. However, <italic>spen</italic> and <italic>imd</italic> mutants were highly sensitive to infection. Interestingly, <italic>imd</italic> was not required for defense against <italic>Micrococcus luteus</italic> (Gram-positive bacteria) ##REF##11269502##[9]##,##REF##14985331##[25]##. On the other hand loss of <italic>spz</italic> caused a minor defect in immune response against <italic>M. luteus</italic> infection as was shown ##REF##11269502##[9]##,##UREF##2##[26]##. Similar infection analysis with <italic>M. luteus</italic> showed significantly more lethality in <italic>spen</italic>, <italic>CG12744</italic>, and <italic>CG12004</italic> than in <italic>spz</italic> mutants without affecting survival in most of the other mutants (##FIG##2##Figure 3B, 3C##). These results indicate that most of the genes except <italic>spen</italic> are not required to defend against Gram-negative bacteria, while two novel genes (<italic>CG122744</italic> and <italic>CG1200</italic>4) are required to defend against Gram-positive bacterial infection. We also tested survival of the mutant flies after <italic>Staphylococcus aureus</italic> infection. In addition to the three mutants susceptible to <italic>M. luteus</italic> infection, <italic>jumeaux</italic>, <italic>Lmpt</italic>, <italic>shg</italic>, and <italic>Trx-2</italic> mutants were highly susceptible to <italic>S. aureus</italic> infection (##FIG##2##Figure 3D##). This result indicated that more sophisticated immune responses are required to control the highly pathogenic <italic>S. aureus</italic>. Therefore, of the 16 genes found to be essential for anti-fungal defense, <italic>spen</italic> appears to be required for general immune responses, while nine genes (<italic>Pcl</italic>, <italic>inv</italic>, <italic>DDB1</italic>, <italic>coro</italic>, <italic>loco</italic>, <italic>Rab6</italic>, <italic>JhI-21</italic>, <italic>CG6181</italic>, and <italic>CG7263</italic>) are specifically required for anti-fungal defense. The other six genes (<italic>CG12744</italic>, <italic>jumeaux</italic>, <italic>Lmpt</italic>, <italic>Trx-2</italic>, <italic>shg</italic>, and <italic>CG12004</italic>) are differentially required, to defend against Gram-positive bacteria, depending on the pathogenic activities of the infecting bacteria. Because flies utilize several defense mechanisms against microbial infection, Gram-negative bacteria may be easily cured even if some mechanisms are not functional, while both cellular and humoral defenses may be needed to eradicate highly pathogenic microbes such as <italic>S. aureus</italic> and fungi.</p>", "<title>Effects on antimicrobial peptide gene expression</title>", "<p>To determine whether the immune response was defective in each mutant, particularly in adults, we first examined the synthesis of diverse antimicrobial peptides (AMPs) in response to fungal infection. Quantitative RT-PCR analysis of five major AMP transcripts (<italic>AttA</italic>, <italic>CecA2</italic>, <italic>Dpt</italic>, <italic>Drom</italic>, and <italic>Def</italic>) revealed very low AMP transcript levels that are comparable to those in wild type flies prior to fungal infection in all of the mutants, indicating no major defect in the regulation of basal AMP expression in the mutants (data not shown). When the flies were challenged with fungal spores, all the five AMP genes were highly induced in wild type flies, and the expression of these genes was strongly reduced or abolished by mutation of the Toll-dependent transcription factor, <italic>Dif</italic>. Under the same infection condition, most of the mutant flies were defective in activation of certain types of AMP gene expression, and different AMP genes appear to require different genes for their activation in response to fungal infection (##FIG##3##Figure 4##). <italic>AttA</italic>, <italic>Drom</italic>, and <italic>Dpt</italic> synthesis in response to fungal infection was not affected in most of the mutants. However, mutations in <italic>Trx-2</italic>, <italic>coro</italic>, <italic>CG6181</italic> and <italic>spen</italic> caused moderate defects in their activation. In contrast, the induction of <italic>CecA2</italic> and <italic>Def</italic> by fungal infection was significantly reduced in most of the mutants analyzed. <italic>CecA2</italic> expression was defective in most of the mutants except <italic>DDB1</italic>. In particular, <italic>CecA2</italic> expression was completely abolished in <italic>JhI-21</italic> and <italic>CG6181</italic> mutants, and was highly repressed in <italic>spen</italic> and <italic>jumeaux</italic> mutants. Activation of <italic>Def</italic> expression was affected in most of the mutants except <italic>CG12744</italic>, <italic>jumeaux</italic>, and <italic>CG7263</italic>, with the most severe defects found in <italic>Trx-2</italic>, <italic>CG12004</italic>, and <italic>JhI-21</italic> mutants. Therefore, <italic>spen</italic>, <italic>Trx-2</italic>, <italic>coro</italic>, and <italic>CG6181</italic> appear to be required to activate most of the antimicrobial peptide genes upon fungal infection, while <italic>CG12004</italic> and <italic>JhI-21</italic> appear to be required to activate <italic>Def</italic> and <italic>CecA2</italic>, respectively. However, <italic>DDB1</italic> does not seem to be required to activate AMP expression induced by fungal infection.</p>", "<title>\n<italic>In vivo</italic> assessment of phagocytosis</title>", "<p>Along with the humoral response, which is mediated mainly by the synthesis of specific antimicrobial peptides, the phagocytosis of invading microbes by hemocytes is another major defense mechanism of adult flies. Hemocytes are mostly sessile and cannot easily be removed from adult flies. However, these cells can be observed through the cuticle, and clusters of hemocytes are present under the dorsal surface of the abdomen, along the dorsal vessel ##REF##12761139##[27]##,##REF##10898983##[28]##. To assay the phagocytic activities of mutant hemocytes in vivo, wild type and mutant adult male flies were infected with Alexa Fluor 488-labeled spores of <italic>B. bassiana</italic>, and the level of fluorescence from phagocytosed spores was measured after quenching the signal from spores outside the hemocytes (##FIG##4##Figure 5, A and C##). Wild type flies showed a strong fluorescence signal from the phagocytosed spores; however, eleven (<italic>spen</italic>, <italic>Pcl</italic>, <italic>CG12744</italic>, <italic>Lmpt</italic>, <italic>coro</italic>, <italic>shg</italic>, <italic>loco</italic>, <italic>Rab6</italic>, <italic>CG12004</italic>, <italic>JhI-21</italic>, <italic>and CG7263</italic>) of the sixteen EP mutants had a weak fluorescence signal, indicating that the mutant hemocytes were defective in uptake of the spores. To determine whether the reduction in phagocytosed spores in some of the mutant flies resulted from the reduced hemocytes, we measured the number of hemocytes present under the dorsal surface of the abdomen of each of the mutant flies. Hemocytes were visualized by injecting India ink, and the amount of black particles taken up by each mutant hemocyte was quantified. India ink staining revealed that most of the mutants contained hemocytes that were comparable to or even higher (<italic>spen</italic>, <italic>jumeaux</italic>, <italic>CG12004</italic>, <italic>JhI-21</italic>, <italic>and CG7263</italic>) than wild type (##FIG##4##Figure 5B##). Therefore, the reduced fluorescent signals appear to reflect defective phagocytosis rather than fewer hemocytes in the mutants. When the fluorescent signal of the phagocytosis assay was normalized to the number of hemocytes estimated by India ink staining, we observed a moderate defect in <italic>jumeaux</italic> mutant in addition to the eleven mutants that showed clear phagocytic defects (##FIG##4##Figure 5D##). In addition, these fluorescent signals appeared to depend on the phagocytotic machinery of the hemocytes since injection of excessive latex beads competed out the signal completely (##FIG##4##Figure 5C##). Therefore, in addition to obvious phagocytotic components (cytoskeletal regulators; <italic>coro</italic>, <italic>shg</italic>, <italic>loco</italic>, and <italic>rab6</italic>), genes in diverse categories, such as transcription factors (<italic>spen</italic>, <italic>Pcl</italic>, <italic>CG12744</italic>, and <italic>Lmpt</italic>), cell death regulators (<italic>CG7263</italic>), and other novel factors (<italic>CG12004</italic> and <italic>JhI-21</italic>), appear to be required to phagocytose fungal spores.</p>", "<p>We next examined whether similar genes are required to phagocytose bacteria. The <italic>E. coli</italic> phagocytosis signal was strongly reduced in flies carrying a mutation in the cytoskeletal regulators (<italic>coro</italic>, <italic>shg</italic>, <italic>loco</italic>, and <italic>rab6</italic>) or in some of the genes required to phagocytose fungal spores (<italic>spen</italic>, <italic>Pcl</italic>, <italic>JhI-21</italic>, <italic>and CG7263</italic>). In addition, <italic>CG6181</italic> appeared to be required specifically for <italic>E. coli</italic> phagocytosis (##FIG##4##Figure 5, E and F##). In addition to genes required to phagocytose <italic>E. coli</italic>, phagocytosis of <italic>S. aureus</italic> requires additional genes that function as transcription factors (<italic>CG12744</italic>, <italic>jumeaux</italic>, and <italic>Lmpt</italic>) or as a redox regulator (<italic>Trx-2</italic>) (##FIG##4##Figure 5G, 5H##). These results indicate that genes involved in cytoskeletal and cell death regulation, along with <italic>spen</italic> (chromatin regulator) and <italic>JhI-21</italic> (transporter induced by juvenile hormone), are generally required for phagocytosis of diverse microorganisms. In contrast, <italic>jumeaux</italic> and <italic>Trx-2</italic> are required to specifically phagocytose <italic>S. aureus</italic>, which is known to utilize diverse immune evading mechanisms ##REF##12586708##[29]##–##REF##15627982##[31]##. Therefore, hemocytes appear to require genes involved in diverse cellular functions to mediate a proper cellular immune response against fungal and bacterial infection.</p>", "<title>Analysis of <italic>Drosophila</italic> larval hematopoiesis</title>", "<p>The analysis of hemocytes in adult flies revealed that some mutants are defective in the activation of both phagocytosis and AMP synthesis, and showed an abnormal number of hemocytes. This observation suggested that some of the immune defects were caused by inappropriate hematopoiesis. To test this idea we examined whether hemocyte development in these mutants occurred normally. We first compared the number of circulating plasmatocytes in third instar larvae of mutant and wild type flies. Since the number of circulating hemocytes increases rapidly during development, we staged the wandering larvae according to the presence or absence of food in the gut ##UREF##3##[32]##. Because the mutants showed no obvious developmental defects or delay, this method enabled us to measure the circulating hemocytes of each mutant at a comparable developmental stage. However, we cannot rule out the possibility that hemocyte development in certain mutant larvae was affected in some degree by the mutations. When we counted the circulating hemocytes, late third instar larvae of six mutants (<italic>spen</italic>, <italic>Pcl</italic>, <italic>jumeaux</italic>, <italic>CG12004</italic>, <italic>JhI-21</italic>, <italic>and CG7263</italic>) displayed a 2- to 6-fold increase in the number of plasmatocytes (##FIG##5##Figure 6A##), which is consistent with the higher number of hemocytes observed in the adult of the same mutants. Therefore, the defective immune responses observed in many mutants appear to be related to abnormal plasmatocyte proliferation.</p>", "<p>The defects in plasmatocyte proliferation in some of the immune compromised mutant flies prompted us to examine the effect of the mutations on crystal cell development. To measure the number of crystal cells in the larvae of each mutant, third instar larvae were heated to 60°C for 10 min to induce blackening of mature crystal cells. <italic>spen</italic>, <italic>Pcl</italic>, <italic>CG12744</italic>, <italic>Trx-2</italic>, <italic>and DDB1</italic> mutant larvae showed fewer crystal cells than did wild type larvae (##FIG##5##Figure 6B##). We also tested the functional activity of crystal cells in each mutant by injuring third instar larvae with a clean needle and measuring the level of melanization in each mutant larva. Strong melanization at the injury site was observed in wild type larvae and most of the mutant larvae. However, <italic>spen</italic>, <italic>Pcl</italic>, <italic>CG12744</italic>, <italic>Trx-2</italic>, <italic>and DDB1</italic> mutant larvae showed much less melanization induced by injury, consistent with their defects in crystal cell proliferation (##FIG##5##Figure 6C##, ##SUPPL##5##Figure S3##). Therefore, <italic>spen</italic> and <italic>Pcl</italic>, which are involved in chromatin regulation, appear to function in the development of both plasmocytes and crystal cells. It is intriguing that genes involved in the recognition of damaged DNA (<italic>DDB1</italic>), redox regulation (<italic>Trx-2</italic>), and a novel transcription factor (<italic>CG12744</italic>) are also required for proper crystal cell development.</p>" ]

[ "<title>Discussion</title>", "<p>The immune system employs multiple layers of defense against pathogens and it is difficult for most invading bacteria to overcome these redundant host defense barriers. However, fungi are largely opportunists, causing infection when any of host defenses are breached. <italic>Beauveria bassiana</italic> is an entomopathogenic fungus that causes a disease in insects known as white muscadine disease. Unlike bacterial pathogens, once inside the insect it produces a toxin that weakens the host immune system. To search for important factors within the entire <italic>Drosophila</italic> immune system that are required for antifungal defense, we screened for genes specifically required for survival following <italic>B. bassiana</italic> infection and identified several genes involved in diverse aspects of cellular and humoral immune responses (summarized in ##TAB##0##Table 1##). Although some of the mutants showed general immune defects and were susceptible to both fungal and bacterial infection, most of the other mutants exhibited distinct immune defects and were susceptible only to fungal or to highly pathogenic bacterial infection. This increased susceptibility specifically to fungal infection might result from defects in defenses against fungal-specific pathogenic molecules, but it is also possible that anti-fungal responses require more diverse immune defense mechanisms than bacterial infection, such that mutants with specific defects could overcome bacterial infection using other functional immune responses.</p>", "<p>Inappropriate development of plasmocytes and crystal cells appears to be one of the main causes of the immune defects in the mutants identified in this screen. Spen and Pcl play essential roles in the chromatin modification needed for hemocyte development ##REF##11331609##[33]##–##REF##12697833##[35]##. Mutations in these genes must prevent progenitor hemocytes from differentiating into functional plasmocytes or crystal cells, and cause pleiotrophic defects in diverse aspects of immune function. Pcl appears to be less important for bacterial infection than does Spen, but the difference may be due to different degrees of gene inactivation in the <italic>Pcl</italic> heterozygotes vs. <italic>spen</italic> homozygotes, rather than from differences in regulatory function. A similar explanation could be applied to fungal specific defects of the other heterozygote mutants.</p>", "<p>In addition to chromatin regulators, it is intriguing that CG12744 and Jumeaux are required specifically for the development of crystal cells and plasmocytes, respectively. CG12744 is a novel transcription factor; in contrast, Jumeaux is a transcription factor expressed in embryonic CNS and is required in neuronal development ##REF##10887088##[36]##. How these transcription factors regulate the development of specific hemocytes is not known, but their expression pattern and requirement in a specific blood cell type suggest a role in the maturation of distinct types of hemocytes.</p>", "<p>Crystal cell differentiation also requires Trx-2 (thoredoxin-2) and DDB1 (Damaged DNA Binding protein 1). Trx-2 regulates redox signaling, which is essential for the activation of immune effector functions ##REF##7577807##[30]##,##REF##15627982##[31]##,##REF##3896121##[37]## and the melanization reaction. The misregulation of redox signals by the loss of Trx-2 may affect early steps in the signal transduction pathway induced by pathogen recognition, causing diverse defects in immune function. DDB1 is involved in the recognition of damaged DNA in dying cells or in invading pathogens and is required for plasmocyte development ##REF##15381102##[38]##. However, how DDB1 affects crystal cell function is not known.</p>", "<p>In addition to transcription factors, cytoskeletal regulators are another major group of genes required to defend against infection. Coro has F-actin binding activity and is required for membrane trafficking ##REF##15090595##[39]##. Shg is a <italic>Drosophila</italic> Cadherin and is required for cell motility and adhesion ##REF##8598295##[40]##,##REF##9012491##[41]##. Loco and Rab6 are involved in asymmetric cell division and vesicle transport, respectively ##REF##15937221##[42]##,##REF##9261061##[43]##. Therefore, these proteins must be required for cytoskeletal rearrangement during phagocytosis. It is interesting that these mutants also showed defects in AMP synthesis. Efficient recognition of pathogens or subsequent signaling may require cytoskeletal rearrangement.</p>", "<p>We also identified several novel genes, whose function in innate immunity has not been previously suggested. CG12004 is a novel protein without known protein motifs, but it appears to play an important role in plasmocyte development. JhI-21 is a cationic amino acid transporter induced by juvenile hormone ##REF##12530223##[44]##. It is required for plasmocyte development and affects their phagocytosis and AMP synthesis. CG6181 is a novel protein and CG7263 are known to be involved in apoptosis ##REF##10908589##[45]##. Recently, endocytic degradation by apoptosis was suggested to play essential roles in defense against pathogenic microbes that can escape from endosomes to cytoplasm ##REF##9423876##[46]##.</p>", "<p>Several novel genes identified from this screen appear to have essential roles in defense against both fungi and bacteria, indicating their roles in the regulation of primary immune responses. The putative functions of these newly identified genes (Spen, CG12744, Jumeaux, Lmpt, Trx-2, Shg, and CG12004) as transcription factors, redox regulator, or cell adhesion molecule hints at their role in regulating immune responses. Therefore, further study of these genes will provide important insight into regulatory mechanism of the <italic>Drosophila</italic> immune system.</p>", "<p>Our results showed that complex immune reactions are required to defend against fungal infection in <italic>Drosophila</italic>, and identified key regulatory components involved in these immune reactions. These findings increase our understanding of the mechanisms underlying cellular and humoral aspects of <italic>Drosophila</italic> antifungal immunity, and have significant implications in the treatment of human diseases caused by fungi.</p>" ]

[]

[ "<p>Conceived and designed the experiments: LHJ YJK. Performed the experiments: LHJ JS JSY BK JK. Analyzed the data: LHJ YJK. Contributed reagents/materials/analysis tools: LHJ JKH. Wrote the paper: LHJ YJK.</p>", "<p>Essential aspects of the innate immune response to microbial infection appear to be conserved between insects and mammals. Although signaling pathways that activate NF-κB during innate immune responses to various microorganisms have been studied in detail, regulatory mechanisms that control other immune responses to fungal infection require further investigation. To identify new <italic>Drosophila</italic> genes involved in antifungal immune responses, we selected genes known to be differentially regulated in SL2 cells by microbial cell wall components and tested their roles in antifungal defense using mutant flies. From 130 mutant lines, sixteen mutants exhibited increased sensitivity to fungal infection. Examination of their effects on defense against various types of bacteria and fungi revealed nine genes that are involved specifically in defense against fungal infection. All of these mutants displayed defects in phagocytosis or activation of antimicrobial peptide genes following infection. In some mutants, these immune deficiencies were attributed to defects in hemocyte development and differentiation, while other mutants showed specific defects in immune signaling required for humoral or cellular immune responses. Our results identify a new class of genes involved in antifungal immune responses in <italic>Drosophila</italic>.</p>", "<title>Author Summary</title>", "<p>The innate immune response is the first line of defense against microbial infections in insects and mammals. In <italic>Drosophila</italic>, multiple defense mechanisms that contribute to the innate immune response include antimicrobial peptides (AMP), reactive oxygen species, phagocytosis and melanization. A search for genes involved in these immune processes identified sixteen mutants that exhibited increased lethality after infection. The diverse functions annotated to these genes indicate the complexity of the regulatory mechanisms required for defense against fungal infection. Lineage specific transcription factors and chromatin modifiers appeared to be required for proper development of functional hemocytes, while cytoskeletal regulators were required for phagocytotic activities of hemocytes. In addition, we identified several genes involved in the immune signaling required for AMP synthesis or melanization. These results may lay the foundation for defining a new class of genes that are involved in humoral and cellular antifungal immune responses.</p>" ]

[ "<title>Supporting Information</title>" ]

[ "<p>We thank Dr. Won-Jae Lee for providing <italic>Ecc-15</italic> bacterial strains and <italic>Imd</italic> and <italic>spz^rm7</italic>, and Dr. Kwang-Min Choe for <italic>Dif ²</italic> mutant flies.</p>" ]

[ "<fig id=\"ppat-1000168-g001\" position=\"float\"><object-id pub-id-type=\"doi\">10.1371/journal.ppat.1000168.g001</object-id><label>Figure 1</label><caption><title>Screening of immune defective mutant flies.</title><p>(A) The survival rate on day six after <italic>Beauveria bassiana</italic> (<italic>B. bassiana</italic>) septic infection is shown for the 130 mutant lines examined. Of 130 mutant lines, 16 had dramatically reduced resistance to <italic>B. bassiana</italic> infection, with a survival rate of less than 50% after six days (<italic>p</italic>&lt;0.002). (B, C) Survival rate kinetics from three independent experiments are shown for the 16 mutants with increased sensitivity, (B) septic infection (C) natural infection. The <italic>spz^rm7</italic> mutant was used as a positive control for <italic>B. bassiana</italic> infection. WT is wild type, and D3 and D6 stands for the survival at day 3 and day 6, respectively, post-infection.</p></caption></fig>", "<fig id=\"ppat-1000168-g002\" position=\"float\"><object-id pub-id-type=\"doi\">10.1371/journal.ppat.1000168.g002</object-id><label>Figure 2</label><caption><title>Confirmation of precise excision lines from <italic>P</italic>-element inserted mutant flies.</title><p>(A) Relative positions and orientations of PCR primers used to confirm the position of <italic>P</italic>-element insertions and their precise excision. Primers F and R are complementary to the genomic DNA surrounding each <italic>P</italic>-element insertion site, while the PF primer is complementary to the <italic>P</italic>-element. PCR fragments amplified by the gene-specific primers (F and R) or by the <italic>P</italic>-element- and gene-specific primers (PF and R) are indicated as (I) and (II), respectively. (B) PCR fragments (I) and (II) amplified from wild type (W) or mutant (M) flies are shown in the left panel. The right panel shows corresponding PCR fragments amplified from the precise excision lines (M') of each mutant. Names of the specific precise excision alleles used in the analysis are indicated at the right. Survival kinetics of the precise excision lines (C) and the flies overexpressing wild type transgenes in mutant background (D). The <italic>spz^rm7</italic> mutant was used as a positive control for <italic>B. bassiana</italic> infection.</p></caption></fig>", "<fig id=\"ppat-1000168-g003\" position=\"float\"><object-id pub-id-type=\"doi\">10.1371/journal.ppat.1000168.g003</object-id><label>Figure 3</label><caption><title>Survival rates of wild type and mutant flies following bacterial infection.</title><p>(A). Survival rate kinetics for the 16 mutants with respect to wound healing. (B, C, D) Survival rates of the 16 mutants following septic infection with bacteria are shown, <italic>Ecc-15</italic> (Gram-negative) (B), <italic>M. luteus</italic> (Gram-positive) (C), and <italic>S. aureus</italic> (Gram-positive) (D). <italic>Imd</italic> and <italic>spz</italic> mutants were used to control for sensitivity to bacterial infections. The color code of each line is the same as in right side of figure.</p></caption></fig>", "<fig id=\"ppat-1000168-g004\" position=\"float\"><object-id pub-id-type=\"doi\">10.1371/journal.ppat.1000168.g004</object-id><label>Figure 4</label><caption><title>Expression profiles of antimicrobial peptide genes in wild type and mutant flies.</title><p>Quantitative real time PCR analysis of antimicrobial peptide genes is presented. Wild type and mutant adult flies were infected with live spores of <italic>B. bassiana</italic>, total RNA was isolated from adult flies, and subjected to real time PCR analysis six hours after injection. The amount of transcripts in each sample were normalized to <italic>RpL32</italic> transcripts (AMP transcripts = normalized target mRNA expression in sample×1000). The mean and standard deviation of three independent experiments are shown. <italic>AttA</italic>, <italic>Attacin A</italic>; <italic>CecA2</italic>, <italic>Cecropin A2</italic>; <italic>Drom</italic>, <italic>Drosomycin</italic>; <italic>Dpt</italic>, <italic>Diptericin</italic>; <italic>Def</italic>, <italic>Defensin</italic>.</p></caption></fig>", "<fig id=\"ppat-1000168-g005\" position=\"float\"><object-id pub-id-type=\"doi\">10.1371/journal.ppat.1000168.g005</object-id><label>Figure 5</label><caption><title>\n<italic>In vivo</italic> phagocytosis in adult flies.</title><p>(A) Adult males of the indicated genotypes were injected with Alexa Fluor 488-labeled heat killed spores of <italic>B. bassiana</italic>. (B, C, E, G) Quantitation of in vivo phagocytosis of India ink, spore and bacteria. (B) India ink, (C) Alexa Fluor 488-labeled heat killed spores of <italic>B. bassiana</italic>, (E) Fluorescein conjugated <italic>E. coli</italic> (K-12), (G) Fluorescein conjugated <italic>S. aureus</italic>. Phagocytosed signals were observed under a Zeiss Axioplan 2 microscope. Phagocytic index was derived by multiplying the area of the India ink and fluorescence signal measured. Phagocytosis was inhibited by prior injection of latex beads in wild type. LXB, CML latex beads. (D, F, H) A phagocytic index was obtained by multiplying phagocytosing signals with the mean area of internalized India ink. (D) spore, (F) <italic>E. coli</italic>, (H) <italic>S. aureus</italic>. The mean and standard deviation of 10–16 adult flies were analyzed for each genotype. <italic>p</italic>-values were calculated by Student's t-test. India ink, **<italic>p</italic>&lt;0.1. Fungi and bacteria, **<italic>p</italic>&lt;0.007.</p></caption></fig>", "<fig id=\"ppat-1000168-g006\" position=\"float\"><object-id pub-id-type=\"doi\">10.1371/journal.ppat.1000168.g006</object-id><label>Figure 6</label><caption><title>Analysis of <italic>Drosophila</italic> larval hematopoiesis.</title><p>(A) Analysis of circulating plasmatocytes. Plasmatocytes were counted from at least six third instar larvae of each genotype; error bars represent the SD of the mean from 3–6 independent experiments. (B, C) Analysis of crystal cell development. (B) Third instar larvae were heated to 60°C for 10 min in a water bath to visualize crystal cells through the cuticle. Crystal cell counts from the sessile population of the last two posterior dorsal segments of third instar larvae of the indicated genotype are shown. (C) Third instar larvae pricked with a clean standard needle led to a hemocoelic melanization reaction. Melanization index was derived by the area with the mean intensity of melanization signal measured. 12–20 larvae were analyzed for each genotype, error bars represent standard deviation. In each panel, <italic>p</italic>-values were calculated by Student's t-test. **<italic>p</italic>&lt;0.01.</p></caption></fig>" ]

[ "<table-wrap id=\"ppat-1000168-t001\" position=\"float\"><object-id pub-id-type=\"doi\">10.1371/journal.ppat.1000168.t001</object-id><label>Table 1</label><caption><title>Summary of the cellular and humoral immune responses of the sixteen mutants.</title></caption><alternatives><table frame=\"hsides\" rules=\"groups\"><colgroup span=\"1\"><col align=\"left\" span=\"1\"/><col align=\"center\" span=\"1\"/><col align=\"center\" span=\"1\"/><col align=\"center\" span=\"1\"/><col align=\"center\" span=\"1\"/><col align=\"center\" span=\"1\"/><col align=\"center\" span=\"1\"/><col align=\"center\" span=\"1\"/><col align=\"center\" span=\"1\"/><col align=\"center\" span=\"1\"/><col align=\"center\" span=\"1\"/><col align=\"center\" span=\"1\"/><col align=\"center\" span=\"1\"/><col align=\"center\" span=\"1\"/><col align=\"center\" span=\"1\"/><col align=\"center\" span=\"1\"/><col align=\"center\" span=\"1\"/></colgroup><thead><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\"/><td align=\"left\" rowspan=\"1\" colspan=\"1\">\n<italic>spen</italic>\n</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">\n<italic>Pcl</italic>\n</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">\n<italic>CG12744</italic>\n</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">\n<italic>jumeaux</italic>\n</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">\n<italic>inv</italic>\n</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">\n<italic>Lmpt</italic>\n</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">\n<italic>Trx-2</italic>\n</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">\n<italic>DDB1</italic>\n</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">\n<italic>coro</italic>\n</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">\n<italic>shg</italic>\n</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">\n<italic>loco</italic>\n</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">\n<italic>Rab6</italic>\n</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">\n<italic>CG12004</italic>\n</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">\n<italic>JhI-21</italic>\n</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">\n<italic>CG6181</italic>\n</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">\n<italic>CG7263</italic>\n</td></tr></thead><tbody><tr><td colspan=\"17\" align=\"left\" rowspan=\"1\">\n<bold>Survival</bold>\n</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">Fungi</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">---</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">---</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">---</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">---</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">---</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">---</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">---</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">---</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">---</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">---</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">---</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">---</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">---</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">---</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">---</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">---</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">\n<italic>Ecc-15</italic>\n</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">---</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">\n<italic>M. luteus</italic>\n</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">---</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">---</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">---</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">\n<italic>S.aureus</italic>\n</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">---</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">---</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">---</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">---</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">---</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">---</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">---</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td></tr><tr><td colspan=\"17\" align=\"left\" rowspan=\"1\">\n<bold>Phagocytosis</bold>\n</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">\n<italic>Fungi</italic>\n</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">---</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">---</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">---</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">---</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">---</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">---</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">---</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">---</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">---</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">---</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">---</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">\n<italic>E.coli</italic>\n</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">---</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">---</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">---</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">---</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">---</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">---</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">---</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">---</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">---</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">\n<italic>S.aureus</italic>\n</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">---</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">---</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">---</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">---</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">---</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">---</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">---</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">---</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">---</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">---</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">---</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">---</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">---</td></tr><tr><td colspan=\"17\" align=\"left\" rowspan=\"1\">\n<bold>Proliferation</bold>\n</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">Plasmatocytes</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">---</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">---</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">---</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">---</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">---</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">---</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">Crystal cells</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">---</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">---</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">---</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">---</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">---</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td></tr><tr><td colspan=\"17\" align=\"left\" rowspan=\"1\">\n<bold>AMP genes</bold>\n</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">\n<italic>AttA</italic>\n</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">-</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">-</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">-</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">\n<italic>CecA2</italic>\n</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">-</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">-</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">-</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">-</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">-</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">-</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">-</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">-</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">-</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">-</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">-</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">-</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">---</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">---</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">-</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">\n<italic>Dpt</italic>\n</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">-</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">-</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">\n<italic>Drom</italic>\n</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">-</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">\n<italic>Def</italic>\n</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">-</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">-</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">-</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">-</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">---</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">-</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">-</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">-</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">-</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">-</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">---</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">---</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">-</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+</td></tr></tbody></table></alternatives></table-wrap>" ]

[]

[]

[]

[]

[]

[ "<supplementary-material content-type=\"local-data\" id=\"ppat.1000168.s001\"><label>Table S1</label><caption><p>\n<italic>P</italic>-element insertion lines used and their survival after septic infection of <italic>B. bassiana</italic>.</p><p>(0.26 MB DOC)</p></caption></supplementary-material>", "<supplementary-material content-type=\"local-data\" id=\"ppat.1000168.s002\"><label>Table S2</label><caption><p>GenExel EP lines isolated from the antifungal screen.</p><p>(0.05 MB DOC)</p></caption></supplementary-material>", "<supplementary-material content-type=\"local-data\" id=\"ppat.1000168.s003\"><label>Table S3</label><caption><p>Primer sequences used for PCR analyses.</p><p>(0.06 MB DOC)</p></caption></supplementary-material>", "<supplementary-material content-type=\"local-data\" id=\"ppat.1000168.s004\"><label>Figure S1</label><caption><p>Germinating hyphes of <italic>B. bassiana</italic> on dead <italic>Drosophila</italic>.</p><p>(0.45 MB TIF)</p></caption></supplementary-material>", "<supplementary-material content-type=\"local-data\" id=\"ppat.1000168.s005\"><label>Figure S2</label><caption><p>Overexpression of the disrupted genes using the Gal4-dependent promoter of the <italic>P</italic>-element that was inserted at the 5′ UTR of the gene in a forward orientation.</p><p>(0.63 MB TIF)</p></caption></supplementary-material>", "<supplementary-material content-type=\"local-data\" id=\"ppat.1000168.s006\"><label>Figure S3</label><caption><p>Melanization induced by a clean injury in the 16 mutant larvae.</p><p>(1.26 MB TIF)</p></caption></supplementary-material>" ]

[ "<table-wrap-foot><fn id=\"nt101\"><label/><p>---, severe defects; -, defect; +, no defect.</p></fn></table-wrap-foot>", "<fn-group><fn fn-type=\"COI-statement\"><p>The authors have declared that no competing interests exist.</p></fn><fn fn-type=\"financial-disclosure\"><p>This work is supported by the GRL grant from the Korean Ministry of Education, Science and Technology to Y.J.K.</p></fn></fn-group>" ]

[ "<graphic xlink:href=\"ppat.1000168.g001\"/>", "<graphic xlink:href=\"ppat.1000168.g002\"/>", "<graphic xlink:href=\"ppat.1000168.g003\"/>", "<graphic xlink:href=\"ppat.1000168.g004\"/>", "<graphic xlink:href=\"ppat.1000168.g005\"/>", "<graphic xlink:href=\"ppat.1000168.g006\"/>", "<graphic id=\"ppat-1000168-t001-1\" xlink:href=\"ppat.1000168.t001\"/>" ]

[ "<media xlink:href=\"ppat.1000168.s001.doc\"><caption><p>Click here for additional data file.</p></caption></media>", "<media xlink:href=\"ppat.1000168.s002.doc\"><caption><p>Click here for additional data file.</p></caption></media>", "<media xlink:href=\"ppat.1000168.s003.doc\"><caption><p>Click here for additional data file.</p></caption></media>", "<media xlink:href=\"ppat.1000168.s004.tif\"><caption><p>Click here for additional data file.</p></caption></media>", "<media xlink:href=\"ppat.1000168.s005.tif\"><caption><p>Click here for additional data file.</p></caption></media>", "<media xlink:href=\"ppat.1000168.s006.tif\"><caption><p>Click here for additional data file.</p></caption></media>" ]

{ "acronym": [], "definition": [] }

[ "<title>Introduction</title>", "<p>Human cytomegalovirus (HCMV) is a widespread pathogen which is mostly asymptomatic in immune competent individuals, but pathogenic in the immune compromised such as post-transplant or AIDS patients ##UREF##0##[1]##. Following primary infection, HCMV establishes a latent infection for life which is largely controlled by the cellular immune system. Immune control of HCMV requires enormous immunological resources with often more than 10% of the T cell pool being CMV-specific, a number that might further increase with age ##REF##16147978##[2]##. However, these immunological efforts are unable to eliminate the virus and do not prevent super-infection ##REF##11333993##[3]##. Thus, HCMV is a master in surviving in the face of a constant immunological onslaught.</p>", "<p>As one of the largest human viruses, with well over 200 open reading frames (ORFs), HCMV uses only about a third of its coding potential for “essential” functions whereas the majority of its genes are non-essential for growth <italic>in vitro</italic>\n##REF##14519856##[4]##,##REF##14623981##[5]##. Many of these “non-essential” genes encode modulators of innate or adaptive immune responses including inhibitors of apoptosis, interferon-induction, T cell and NK cell recognition ##REF##14690857##[6]##–##REF##17892221##[9]##. However, the importance of these immune modulators for viral pathogenesis and immune escape <italic>in vivo</italic> is not known since HCMV does not infect immunocompetent experimental animals. Such restricted species specificity is a hallmark of CMVs and, as a result, CMVs have co-evolved with their hosts ##REF##7714900##[10]##. Chimpanzee CMV is most closely related to HCMV ##REF##12533697##[11]##. However, chimpanzees are a protected species and unsuitable as an animal model. Although more distantly related to humans, rhesus macaques (RM) are readily available for experimentation. Sequence analysis of rhesus CMV (RhCMV) revealed that approximately 60% of the open reading frames (ORFs) are homologous to HCMV ORFs including most of the aforementioned immune modulators ##REF##12767982##[12]##,##REF##16571834##[13]##. In order to study the importance of some of the immune regulatory functions <italic>in vivo</italic>, we have begun to characterize several of the conserved immune modulators of RhCMV.</p>", "<p>The US2-US11 genomic region of HCMV encodes multiple proteins that interfere with several MHC and MHC-like molecules. Among the best studied of these is the US6-family which contains four genes that inhibit MHC class I (MHC-I)-mediated antigen presentation to T cells: US2, US3, US6 and US11 ##REF##12224515##[14]##–##REF##12224516##[16]##. These proteins are type I transmembrane glycoproteins that reside in the endoplasmic reticulum and show clear homology to each other and structural features resembling the IG-superfamily fold ##REF##11391001##[17]##. Despite these structural similarities, each protein interferes in its own unique way with the assembly of MHC-I with peptides at a post-translational level. Upon completion of heavy chain (HC) translation and translocation into the lumen of the ER, but prior to assembly with the light chain β2-microglobulin (β2-m), US2 and US11 mediate the retro-translocation of MHC-I molecules to the cytosol ##REF##16098592##[18]##. There, the HC is deglycosylated by N-glycanase and degraded by the proteasome ##REF##8625414##[19]##. US6 inhibits peptide translocation by the TAP thus preventing the MHC-I heterodimers from obtaining viral peptides ##REF##12224516##[16]##. Finally, US3 prevents ER exit of peptide-loaded MHC-I molecules ##REF##11208115##[15]##, both by directly interacting with MHC-I molecules and by interfering with tapasin and protein-disulfide isomerase, both chaperones of the peptide loading complex ##REF##17055437##[20]##.</p>", "<p>We previously demonstrated that the US2-11 orthologues of RhCMV are also functionally equivalent in that Rh182 (US2) and Rh189 (US11) mediate proteasomal destruction of MHC-I, Rh183 (US3) retains MHC-I and Rh185 (US6) inhibits TAP ##REF##15827193##[21]##. Thus, it seemed likely that eliminating the genomic region spanning <italic>RhUS2-11</italic> from RhCMV would restore MHC-I assembly and transport in RhCMV-infected cells as previously observed for <italic>US2-11</italic>-deleted HCMV ##REF##7609050##[22]##. Surprisingly however, we discovered that in addition to these conserved mechanisms, RhCMV contains an additional ORF, <italic>rh178</italic>, that targets the MHC-I assembly pathway. Interestingly, this ORF does not display any homology to the US6 gene family and acts by a novel mechanism that operates post-transcriptionally, but prior to completion of translation/translocation.</p>" ]

[ "<title>Materials and Methods</title>", "<title>Cells and viruses</title>", "<p>Telomerized rhesus fibroblasts (TRFs) ##REF##12088823##[49]## and telomerized human fibroblasts (THFs) were obtained from Jay Nelson and maintained in Dulbecco's modified eagle's medium (DMEM) with 10% fetal bovine serum, 100U/mL penicillin and 100ug/mL streptomycin. RhCMV strain 68.1 was obtained from Scott Wong ##REF##12767982##[12]## and propagated in TRFs. Recombinant RhCMVs were created as described in the supplemental methods using the RhCMV BAC obtained from Peter Barry ##REF##12692210##[23]##. Recombinant rh178 adenoviruses were created using the AdEasy vector system according to the manufacturers protocol (Stratagene). Adenoviruses AdTrans and AdUS11 were obtained from David Johnson.</p>", "<title>Plasmids and Nucleofection</title>", "<p>HLA-A3 and CD4 constructs were expressed from a modified version of pCDNA3.1(-) (Invitrogen, Carlsbad, CA) in which the CMV promoter was replaced with the EF1α promoter (obtained from Jay Nelson) to create pEF1α. HLA-A3 was obtained by PCR from Jurkat T-cell cDNA using the forward primer <named-content content-type=\"gene\">5′ctggaattcatggccgtcatggcgccccgaac</named-content> and the reverse primer <named-content content-type=\"gene\">5′gtcggatcctcacactttacaagctgtgag</named-content> to amplify the coding region only or the reverse primer <named-content content-type=\"gene\">5′gtcggatccttaggaatcttctcc</named-content> to include the 3′UTR. pEF1α expression plasmids were electroporated into TRFs using the AMAXA Nucleofector II (AMAXA Biosystems, Gaithersburg, MD) using cell line solution L and the T-030 program. 1e6-2e6 TRFs were resuspended in 100 µl AMAXA solution and 2 µg expression plasmid. After electroporation cells were recovered in 500 µl RPMI for 45min at 37°C, and then plated in prewarmed complete DMEM. Transfection efficiency was monitored with a GFP reporter and was consistently &gt;90%. Infections with RhCMV were performed 24 hours after electroporation.</p>", "<title>Metabolic labeling and immunoprecipitation</title>", "<p>Cells were starved for 30-min, except where noted, using DMEM without serum, methionine (Met) or cysteine (Cys). Labeling was performed for indicated times using Pro-mix ³⁵S-Met/Cys (GE Healthcare) at 400 µCi/mL. To chase the label, cells were washed 3× in phosphate buffered saline (PBS) followed by incubation at 37°C in DMEM with 10% FBS containing 90 µg/mL Met and 188 µg/mL Cys. For NP-40 lysis, cells were lysed for 30 minutes at 4°C in 1% NP-40 in PBS with complete protease-inhibitor cocktail (Roche). For SDS lysis, cells were lysed for 10 minutes at 25°C in 0.6% SDS in PBS with complete protease-inhibitor cocktail, then diluted in 3× volume of 1.2% triton X-100 in PBS prior to immunoprecipitation. For glycosidase treatment, PNGase was obtained from NEB and used according to the manufacturers protocol after NP-40 lysis.</p>", "<title>Antibodies</title>", "<p>Polyclonal sera K455 recognizes both chains of the MHC-I heterodimer, assembled and unassembled (obtained from Per Peterson) ##REF##2934137##[24]##. HC-10 only recognizes free MHC-I heavy chains ##REF##2088481##[25]##. HLA-A3 antibody was purified from the GAP A3 hybridoma, obtained from ATCC (HB-122). Antibodies to Calreticulin, Transferrin, Vimentin, HA and FLAG were obtained, respectively, from Stressgen (Victoria, BC), Zymed (S. San Francisco, CA), Biomeda (Burlingame, CA), Santa Cruz, and Sigma. Human CD4 antibody (AHS0412) was obtained from Invitrogen. Secondary Alexa Fluor-conjugated antibodies 594 goat anti-rabbit and 488 goat anti-mouse were obtained from Invitrogen.</p>", "<title>Polyribosome fractionation and northern blots</title>", "<p>Approximately 5×10⁶ TRFs were either Mock-infected or RhCMV-infected for 24 hours. Fresh media was placed on the cells for 45-minutes, and cells were placed on ice and washed 2× with cold PBS containing 0.1 mg/ml cycloheximide (Sigma). All subsequent steps were performed at 4°C. Cells were lysed for 10 min using 600 µl of polysome lysis buffer (15mM Tris, pH 7.4, 15mM MgCl₂, 0.3M NaCl, 1% Triton x-100, 0.1 mg/mL cycloheximide, 1 mg/mL heparin). Lysates were cleared at 12,000× g for 10 min. The supernatant was layered onto the top of a 10–50% sucrose gradient composed of sucrose in polysome lysis buffer excluding Triton x-100. The gradients were centrifuged at 35,000 rpm in a Sorvall SW-41 rotor for 3 hours. 750 µl fractions were collected from the top of the gradient. After adding 4.25ml of 5.65M guanidine HCl, each fraction was ethanol precipitated (−20°C overnight). RNA was pelleted at 15,000× g for 30 min, washed with 70% ethanol, dried at 25°C, and resuspended in 400 µl RNAse-free water. RNA was then re-precipitated by adding 40 µl 0.3M sodium acetate and 900 µl 100% ethanol, washed with 70% ethanol and resuspended in 50 µl RNAse-free water.</p>", "<p>For Northern blotting, 10 µl of each fraction was separated on a denaturing 1% agarose gel containing 1× MESA (Boston BioProducts, Worcester, MA) and 3.7% formaldehyde and transferred to Immobilon-Ny+ nylon membrane (Millipore) by capillary blotting in 20×SSC. RNA was fixed by air drying at 25°C for 30 min and baking at 80°C for 2 hours. Radiolabeled probes were generated by random priming. After denaturing at 100°C for 10 min, the probe was chilled on ice and added to 5mL ExpressHyb hybridization solution (Clontech) for hybridization. Membranes were pre-hybridized for 30 min at 68°C followed by probe hybridization for 2 hours, rinsed and washed twice with 2× SSC, 0.05%SDS followed by two washes in 0.1× SSC, 0.1% SDS.</p>", "<title>Immunofluorescence</title>", "<p>Transfected cells were fixed with 3.7% formaldehyde for 40 minutes, washed twice with PBS, quenched with 50mM NH₄Cl for 10 min, washed twice with PBS, and permeabilized with 0.1% Triton X-100 in PBS for 7 min prior to staining.</p>", "<title>RACE</title>", "<p>Total RNA from TRFs infected with WT RhCMV (or RhCMV lacking <italic>rh175–178</italic> as a negative control) for 24 hours was used. For 3′ RACE, cDNA was synthesized using an oligo-dT anchor (<named-content content-type=\"gene\">5′gaccggatccgaattcgtcgacttttttttttttttttv</named-content>). PCR was performed from cDNA using a PCR anchor primer (<named-content content-type=\"gene\">5′-gaccggatccgaattcgtcgac</named-content>) and a gene specific primer. For 5′ RACE, cDNA was synthesized with a gene specific primer (<italic>rh178</italic>\n<named-content content-type=\"gene\">5′-catttgcatgcagctgtgcg</named-content>). 10 µg cDNA was then treated with terminal deoxynucleotidyl transferase and 0.5mM dATP at 37°C for 30 min, followed by purification and PCR using a nested gene specific primer (<italic>rh178</italic>\n<named-content content-type=\"gene\">5′-gcgcgaaacacgcgtttgc</named-content>) and the oligo-dT anchor.</p>" ]

[ "<title>Results</title>", "<title>Inhibition of MHC-I expression despite deletion of RhUS2-US11</title>", "<p>Deletion of the genomic region encoding US2-US11 restores MHC-I expression in HCMV-infected cells ##REF##7609050##[22]##. To determine if deletion of the homologous region in the RhCMV genome would likewise restore MHC-I expression we created a recombinant RhCMV lacking RhUS2-11 using a RhCMV-derived bacterial artificial chromosome (BAC) (Protocol S1) ##REF##12692210##[23]##. Similar to recombinant HCMV lacking US2-11, a growth defect was not observed for ΔRhUS2-11 ##REF##7609050##[22]##. However, unlike US2-11-deleted HCMV, ΔRhUS2-11 retained some ability to reduce MHC-I steady state levels in infected TRFs as shown by immunoblot (##FIG##0##Fig. 1A##). At 48 hours post-infection, MHC-I was markedly reduced in ΔRhUS2-11-infected TRFs.</p>", "<p>To determine whether the reduced steady state levels were due to interference with newly synthesized MHC-I, we immunoprecipitated MHC-I from radiolabeled TRFs infected with wild-type (WT) or ΔRhUS2-11. When cells were labeled for one or two hours, we recovered dramatically less MHC-I from RhCMV-infected cells despite the use of polyclonal antiserum K455 recognizing all forms of MHC-I (##FIG##0##Fig. 1B##) ##REF##2934137##[24]##. Compared to WT there was an increase in HC recovery from ΔRhUS2-11-infected cells. Such residual HC was also observed in pulse-chase experiments, when ΔRhUS2-11-infected TRFs were pulsed for 10 min and chased from 30 min up to 90 min (##SUPPL##0##Fig. S1A##). However, compared to mock-treated cells, radiolabeled HC was drastically reduced at all time points either during pulse or chase. In contrast to HC, expression of control proteins such as Transferrin-receptor or vimentin was unaffected in RhCMV-infected cells (##FIG##0##Fig. 1D##). Also, we did not observe a general shut-off of host protein expression or a dramatic decrease of glycoprotein recovered with the lectin concanavalin A (data not shown). Moreover, expression of the light chain β2-m was much less affected by RhCMV compared to HC, particularly in short pulse/chase experiments (##FIG##0##Fig. 1C##). These data suggested that in addition to RhUS2-11 inhibiting MHC-I assembly, RhCMV specifically interferes with expression of HC. The residual HC recovered from ΔRhUS2-11-infected cells indicate that this viral inhibition of HC expression (VIHCE) was either incomplete or VIHCE did not equally affect all MHC-I alleles present in TRFs.</p>", "<title>VIHCE does not cause rapid degradation of complete HCs</title>", "<p>Since only minimal amounts of HC are detectable during ΔRhUS2-11 infection, we wanted to examine if VIHCE caused rapid degradation of HCs. In cells infected with HCMV, HC is initially synthesized but then rapidly degraded as shown by pulse-chase (##FIG##0##Fig. 1C##). This observation is consistent with previous reports and is due to the reverse translocation of MHC-I mediated by US2 and US11 followed by proteasomal destruction of MHC-I ##REF##8625414##[19]##. In contrast, during infection with both WT RhCMV and ΔRhUS2-11 only minimal amounts of HC were detectable after a 10-min radiolabel, and remained low during a 30-min chase (##FIG##0##Fig. 1C##). Furthermore, during a radiolabel for only 1-min HC synthesis was markedly reduced during RhCMV infection (##SUPPL##0##Fig. S1B##). To rule out that HC was not recovered due to epitope masking by a viral protein or because HC was in a complex with NP40-insoluble proteins, we lysed cells in SDS to disrupt protein complexes and denature the HC prior to IP. Using either a monoclonal antibody that recognizes only free HC (HC-10) ##REF##2088481##[25]## or K455, we were unable to recover increased amounts of HC under these conditions (##FIG##0##Fig. 1E##). Taken together these data suggest that RhCMV either prevents complete HC synthesis or degrades HC prior to complete protein synthesis.</p>", "<p>Since co-translational degradation is mediated by proteasomes ##REF##11000112##[26]## we wanted to determine whether HC translation could be completed in the presence of proteasome inhibitors. TRFs were infected with ΔRhUS2-11 and treated with the proteasomal inhibitor MG132. However, no significant increase in HC recovery was observed either when total MHC-I was immunoprecipitated with K455 from NP40-lysates or with HC-10 from SDS-lysates (##FIG##0##Fig. 1F##). In contrast, HC was stabilized in cells transduced with Adenovirus expressing HCMV US11. The proteasomal inhibitors Lactacystin and ZL₃VS also failed to stabilize HC in ΔRhUS2-11-infected cells (data not shown). Taken together these data strongly suggest that RhCMV inhibits expression of HC prior to or during polypeptide synthesis. Since this phenotype is observed in the absence of RhUS2-11 and is not present in HCMV, we further conclude that RhCMV contains one or more unique gene(s) encoding VIHCE.</p>", "<title>HC synthesis is restored upon infection with RhCMV lacking <italic>Rh158–180</italic>\n</title>", "<p>Since VIHCE seems to be specific to RhCMV, but absent from HCMV, we examined the RhCMV genome for potential candidate genes. The genomic region spanning ORFs <italic>Rh158</italic> to <italic>rh180</italic>, corresponding to the region between <italic>IE1/IE2 (UL123/UL122)</italic> and <italic>US1</italic> in HCMV, contains a large number of genes that are either specific to RhCMV or are homologous to genes frequently deleted in laboratory strains of HCMV ##REF##12767982##[12]##,##REF##8523595##[27]##. To examine whether this region contains the VIHCE gene, we deleted <italic>Rh158–180</italic> using the BAC-recombination strategy shown in ##FIG##1##Fig. 2A##. Interestingly, Δ158–180 did not show any obvious growth defects despite such a large deletion (data not shown). Moreover, pulse-chase labeling of Δ158–180-infected TRFs revealed initial synthesis of MHC-I followed by degradation (##FIG##1##Fig. 2B##). This degradation could be inhibited by the proteasome inhibitor MG132 (##FIG##1##Fig. 2C##). MG132 also stabilized a smaller, presumably deglycosylated, degradation intermediate (*) which is also observed in cells transfected with RhUS2 ##REF##15827193##[21]##. Thus, it seemed likely that Δ158–180 lacked VIHCE, and that in the absence of VIHCE HC was now degraded by the RhCMV homologues of US2 and US11. To examine whether the combined deletion of <italic>RhUS2-11</italic> and VIHCE would restore HC expression in RhCMV-infected cells, we created a recombinant lacking both <italic>Rh158–180</italic> and <italic>RhUS2-11</italic> (##FIG##1##Fig. 2A##). As expected from the single deletions, the resulting double-deletion virus Δ158–180,ΔRhUS2-11 did not display a growth defect <italic>in vitro</italic> (not shown). When TRFs were infected with Δ158–180,ΔRhUS2-11, HC expression was similar to Mock-infected cells indicating that this recombinant virus no longer interfered with MHC-I expresson (##FIG##1##Fig. 2B##). Taken together, these data indicate that the VIHCE gene is located within the <italic>Rh158–180</italic> region of RhCMV. Furthermore, the fact that HC synthesis is observed in the absence of VIHCE supports our conclusion that VIHCE acts prior to the ER-associated degradation caused by the US2-US11 homologs.</p>", "<title>RhCMV VIHCE maps to <italic>rh178</italic>\n</title>", "<p>To identify the gene(s) coding for VIHCE we systematically deleted fragments of decreasing size within the <italic>Rh158–180</italic> region in an iterative fashion (##FIG##2##Fig. 3A##; ##SUPPL##4##Table S1##). We took advantage of the fact that HC is initially synthesized in cells infected with VIHCE-deleted virus but then degraded by US2 and US11 to distinguish between recombinants encoding or lacking VIHCE. Initially, viruses carrying deletions approximately spanning the left or right half of the <italic>Rh158–180</italic> region were generated (##FIG##2##Fig. 3A##). TRFs were infected with recombinants Δ158–168 and Δ167–180 and pulse-chase was performed. Since HC was expressed in TRFs infected with Δ167–180 and not in TRFs infected with Δ158–168, we concluded that VIHCE was located in the <italic>Rh167–180</italic> region. Similarly, HC was expressed in TRF infected with viruses Δ175–180, Δ175–178, Δ176–178, and Δ177–178, but not Δ167–174, Δ179–180, and Δ175–177 (##FIG##2##Fig. 3A##). These data suggested that <italic>rh178</italic> encodes VIHCE.</p>", "<p>The region encoding <italic>rh178</italic> overlaps with several predicted ORFs and with a previously identified large intron of the <italic>US1</italic>-homologue <italic>Rh181</italic>\n##REF##12186931##[28]## (Gene Bank Accession: AF474179). To exactly determine the mRNAs encoding VIHCE we mapped the transcriptional start and stop sites of the <italic>rh178</italic> ORF and generated additional, smaller deletions and point mutants within the <italic>rh178</italic> coding region (##FIG##2##Figs. 3##–##FIG##3##4##). We performed 5′ and 3′ RACE as well as Northern blot analysis. Sequence analysis of the 5′ RACE product identified a transcription start site downstream of the originally predicted <italic>rh178</italic> start codon (##FIG##2##Fig. 3D##). The identified transcript is predicted to encode a shorter version of rh178. 3′RACE and cDNA cloning further revealed additional splice products in this region: a shorter splice product lacking most of the <italic>rh178</italic> protein encoding region (<italic>rh178.4</italic>; Note that Rivailler et al., (2006) have detailed additional predicted ORFs upstream of <italic>rh178</italic> and denoted them <italic>rh178.1, rh178.2</italic>, and <italic>rh178.3</italic>) and the above mentioned large <italic>Rh181</italic>-transcript which does not contain <italic>rh178</italic> since it is removed by splicing. These three transcripts share the same polyadenylation signal and 3′ terminus (##FIG##2##Fig. 3B##). Northern blot analysis using the predicted <italic>rh178</italic> coding region as probe revealed two transcripts (##FIG##2##Fig. 3C##). A larger predominant transcript of approximately 1600bp corresponds to the expected size of <italic>rh178</italic>. The smaller transcript may correspond to <italic>rh178.4</italic>, a shortened <italic>rh178</italic>, or an unidentified transcript of the opposite sense. These data confirm the expression of the <italic>rh178</italic> transcript during infection and correct the prediction of its protein coding region.</p>", "<p>Kinetic analysis indicates that VIHCE is an early gene that is already expressed within 4 hours of infection (##SUPPL##2##Fig. S3##). To determine whether the protein encoded by <italic>rh178</italic> is responsible for VIHCE, we created a frameshift in the 5′-end of the predicted coding region (rh178FS) (##FIG##3##Fig. 4A##). Since the primer-directed mutagenesis strategy also caused deletion of a portion of the 5′-UTR we generated a control virus (rh178FSCtrl) containing the same modification of the predicted 5′-UTR of <italic>rh178</italic> but no frameshift (##FIG##3##Fig. 4A##). While rh178FSCtrl inhibited HC expression similar to WT (##FIG##3##Fig. 4B##), HC was synthesized in rh178FS-infected TRFs (##FIG##3##Fig. 4C##). Thus, VIHCE is mediated by the <italic>rh178</italic>-encoded protein.</p>", "<p>The rh178 protein (##FIG##4##Fig. 5A##), with a molecular weight of approximately 24 kDa, does not display significant homology with non-RhCMV sequences in the genomic database. A stretch of highly hydrophobic amino-acids beginning at amino acid 14 is predicted to represent a non-cleaved amino-terminal signal anchor (##FIG##4##Fig 5B##). Thus, the most likely topology for this protein is that of a type 1b transmembrane protein, i.e. a large cytoplasmic C-terminus following the signal-anchor. Immunofluorescence analysis of epitope-tagged rh178 indicates that the protein localizes to the ER, suggesting that rh178 is anchored in the ER-membrane (##FIG##4##Fig. 5E##). To obtain better expression of rh178 for further analysis, we constructed replication-defective adenovirus vectors expressing either wild type rh178 (Ad178) or HA-tagged rh178 (Ad178-HA). While there is a predicted glycosylation site at position N101, digestion of whole cell lysate from Ad178-HA transduced cells with peptide:N-Glycosidase F (PNGase) failed to cause a shift in rh178 migration, while a shift was seen with MHC-I HC (##FIG##4##Fig. 5C##). Thus, rh178 does not appear to be glcosylated and this is a further indication that the C-terminus of rh178 is located in the cytosol. To determine if rh178 by itself was capable of VIHCE, we transduced TRFs and performed pulse-chase analysis. Cells transduced with Ad178 exhibited reduced expression of HCs while β2-m was unaffected (##FIG##4##Fig. 5D##), similar to the HC inhibition observed in RhCMV-infected cells (##FIG##0##Fig. 1##). MHC-I HC in cells transduced with a control adenovirus vector, AdTrans, was not affected. Thus, rh178 is both necessary and sufficient for VIHCE.</p>", "<title>RhCMV does not inhibit transcription or ribosome association with HC mRNA</title>", "<p>Our data suggest that VIHCE prevents expression of the majority of HCs prior to completion of protein synthesis. Residual, VIHCE-resistant HCs are eliminated by RhUS2-11. The dramatic reduction of newly synthesized HC observed even in the presence of proteasome inhibitors further suggests that VIHCE either blocks transcription of HC mRNA, completion of HC protein synthesis, or causes HC degradation in a proteasome-independent manner. However, the levels of HC mRNA did not change upon RhCMV-infection as shown by Northern blot (##FIG##5##Fig. 6A##) and by quantitative RT-PCR (data not shown). Additionally, the size of the HC mRNA was unaltered in RhCMV-infected cells suggesting that mRNA is not cleaved, alternative spliced or degraded by RhCMV. We further determined whether HC mRNA is polyadenylated and exported into the cytoplasm by isolating nuclear, cytoplasmic, and polyadenylated RNA fractions from infected cells. We did not observe a significant difference in any of these fractions compared to Mock-infected cells (data not shown). These data indicate that HC mRNA transcription, poly-adenylation, splicing and export to the cytosol is not affected by RhCMV.</p>", "<p>To determine whether the association of HC mRNA with ribosomes is inhibited we analyzed the polyribosome distribution of HC mRNA ##REF##5167019##[29]##. When sucrose-gradient fractions from lysates of Mock-infected or RhCMV-infected TRFs were analyzed by Northen blots, HC mRNA sedimented to the polyribosome fractions 12 and 13 in both Mock- and RhCMV-infected cells (##FIG##5##Fig. 6B##). Small shifts in polyribosome density were observed in RhCMV infection for both HC and GAPDH mRNA, suggesting virus infection causes a slight reduction of ribosomal occupancy on cellular transcripts. Therefore, it seems that VIHCE does not inhibit the association of polyribosomes with HC mRNA.</p>", "<p>While sedimentation to the polyribosome fraction indicates the association of HC mRNA with ribosomes, it was possible that the ribosomes were not active. In order to determine if the ribosomes associated with the HC mRNA are actively translating we incubated cells with puromycin. Puromycin is a polypeptide chain terminator that requires an active peptidyl transferase to cause ribosome dissociation from transcripts. A short (4 min) incubation with puromycin caused a shift in the polyribosome profile of HC mRNA in both RhCMV and Mock-infected cells, indicating ribosome dissociation (##FIG##5##Fig. 6C##). This result indicates that the ribosomes bound to the HC mRNA are actively translating and not simply stalled on the transcript. Taken together these data suggested that HC mRNA is transcribed normally in RhCMV-infected cells and that protein translation is not inhibited at the level of initiation or elongation. However, since full-length HC protein cannot be recovered it seems most likely that HC translation is not completed.</p>", "<title>VIHCE is dependent upon the MHC-I signal peptide</title>", "<p>Observations similar to VIHCE were reported for translation inhibition by microRNAs that bind to the 3′-UTR of target transcripts. Similar to VIHCE, mRNAs that are targeted by a given microRNA are found in an active polyribosomal complex but a translated polypeptide intermediate can not be recovered even in the presence of proteasome inhibitors ##REF##17128272##[30]##. To examine the possibility that VIHCE targets the 3′-UTR of HCs we tested the ability of VIHCE to block synthesis of HC with or without its native 3′-UTR. Since antibodies to rhesus HCs are not available, and VIHCE is able to block expression of human HCs (##FIG##6##Fig. 7A##), we chose to examine VIHCE function on HLA-A3. To determine whether the 3′-UTR was required for this inhibition we transiently expressed HLA-A3 with or without its native 3′-UTR in TRFs. Following transfection we infected cells with either RhCMV containing VIHCE (ΔRhUS2-11) or RhCMV lacking VIHCE (Δrh178,ΔRhUS2-11). Expression of both HLA-A3 carrying the native 3′-UTR and a heterologous vector-derived 3′-UTR sequence was reduced by VIHCE (##FIG##6##Fig. 7B##). The 5′-UTR was vector-derived in both constructs. Therefore, we conclude that VIHCE does not target the UTRs of HC mRNA.</p>", "<p>Translation of type I transmembrane proteins such as HC is dependent upon an N-terminal signal peptide (SP) that mediates translocation across the ER membrane. Upon translation initiation, the SP is recognized by the signal-recognition particle (SRP) which binds to the SP and arrests translation. This is followed by docking of the translation complex to the SRP-receptor which aids the transfer of the ribosomal/mRNA/nascent polypeptide complex to the SEC61 translocon ##REF##10676815##[31]##. Translation then resumes and the nascent polypeptide chain is imported into the lumen of the ER. The fact that VIHCE requires the HC coding sequence suggested that the HC protein might be at least partially translated and that VIHCE acts on the nascent polypeptide. Compared to human HC, we observed that the murine MHC-I molecule H2-K^b was more resistant to VIHCE (data not shown). We hypothesized that this resistance was encoded in the amino-terminus of H2-K^b, specifically the SP. To test this hypothesis we replaced the SP of HLA-A3 with that of H2-K^b. As a further control, we also introduced the SP of CD4 which is more divergent from the HLA-A3 SP (##FIG##6##Fig. 7C##). In both instances we observed that expression of the chimeric protein was much less reduced by virus expressing VIHCE compared to native HLA-A3. Remarkably, the SP of Kb is quite similar to that of HLA-A3 (##FIG##6##Fig. 7C##) yet HLA-A3 expression was restored to almost the same levels as observed for the CD4 SP (##FIG##6##Fig. 7D##). Therefore, we conclude that the SP of primate MHC-I is required for VIHCE to inhibit HC translation. The fact that VIHCE requires the MHC-I SP further suggests that VIHCE interferes with SP-dependent translocation which would lead to translation arrest and rapid, co-translational destruction of the resulting protein fragments.</p>", "<p>We next examined if the MHC-I SP is sufficient for VIHCE recognition. To test this we created a chimeric CD4 molecule with the HLA-A3 signal peptide in place of the native CD4 signal peptide (A3/CD4). When either wild type CD4 or A3/CD4 was expressed in TRFs, neither molecule was significantly affected by the presence of VIHCE, whereas the endogenous MHC-I HC was decreased (##FIG##6##Fig. 7E, 7D##). This indicates that while the MHC-I SP is necessary for recognition by VIHCE, it is not entirely sufficient.</p>" ]

[ "<title>Discussion</title>", "<p>We report here that the ORF <italic>rh178</italic> of RhCMV encodes a novel immune modulatory function, viral inhibitor of heavy chain expression (VIHCE), which prevents the translation of HC in a signal-peptide dependent, but not sufficient, manner. This finding is surprising because RhCMV additionally expresses the HCMV US2-US11 homologs that also interfere with MHC-I stability and assembly. The VIHCE-encoding rh178 is so far unique to RhCMV suggesting that rh178 represents an adaptation to the evolutionary pressure of the non-human primate MHC system. Our previous observations ##REF##15827193##[21]## suggested that the immune evasion mechanisms encoded by the US2-US11 region predate the separation of human and old-world non-human primates which is assumed to have taken place about 25 million years ago ##REF##9668008##[32]##. Recent sequence analysis of the MHC-I locus in RM revealed that the MHC-I has undergone a tremendous change since then. Whereas a typical human or ape haplotype contains “only” six active MHC-I genes, as many as 22 different MHC-I genes are expressed in rhesus. Moreover, the sequence divergence was estimated to be 10-fold higher and genes have been duplicated at an approximately three times greater rate than in humans ##REF##15289473##[33]##,##REF##15269276##[34]##. Thus, it is conceivable that the additional MHC-I genes forced RhCMV to evolve additional countermeasures. It is known that polymorphic MHC-I proteins are differentially affected by US2 and US11 of HCMV ##REF##16361314##[35]##,##REF##9016885##[36]##, although the exact rules of this discrimination still need to be determined. Moreover, each of the US6-family viral immune modulators interferes at a distinct step during the assembly cascade ##REF##10854174##[37]##. Allele-specificity has also been reported for MCMV which contains three genes ##REF##12235213##[38]##, unrelated to either the US6-family or VIHCE, and each of three MCMV-gene products interferes with a different step of MHC-I assembly ##REF##10399073##[39]##. Thus, it seems that CMVs optimize their interference mechanisms, both within a given organism by sequentially attacking MHC molecules during assembly and within a given population by broadening the allele-specificities of these attacks. This conclusion is also supported by our finding that RhCMV lacking either rh178 or RhUS2-11 only partially suppressed MHC-I assembly and transport compared to WT RhCMV. This is either due to differences in allele-specificity within a given animal or an incomplete elimination of all alleles. The finding that RhCMV has a larger number of gene products interfering with MHC-I assembly than either HCMV or MCMV thus correlates with the observation that RM have a larger number of active MHC-I alleles than either human or mouse.</p>", "<p>The extracellular domains of MHC-I, particularly the peptide-binding regions, are highly polymorphic and evolve rapidly. In contrast, the cleaved signal peptide is highly conserved among different MHC-I alleles including the RM MaMu and the human HLA genes ##REF##15289473##[33]##. Many signal peptides for MaMu-I, MaMu-3 and MaMu-A show less than 3 amino-acids difference to either HLA-A, B or C alleles and some MaMu-SPs are identical to HLA-SPs ##REF##10640754##[40]##. A possible reason for the high conservation of HLA signal peptide sequences is the fact that a conserved nona-peptide (VMAPRTLLL in the HLA-A3 sequence) is presented by the non-polymorphic HLA-E molecule to the negative signaling receptor CD94/NKG2A or C of NK cells ##REF##10764619##[41]##. This system seems to be conserved in RM, although some alleles start at the methionine within the peptide ##REF##15289473##[33]##. Interestingly, the SP of the HCMV UL40 glycoprotein contains this nona-peptide which is presented by HLA-E in HCMV-infected cells in a TAP-independent fashion ##REF##10669413##[42]##,##REF##10799855##[43]##. By loading the decoy peptide onto HLA-E, HCMV is thought to prevent the “missing self” stimulation of NK cells by MHC-I downregulation. Importantly, this nona-peptide is also encoded within the SP of Rh67 of RhCMV which otherwise shares only 19% identity with UL40 ##REF##12767982##[12]##. Since VIHCE requires polypeptide sequence beyond the SP in MHC-I HCs, the Rh67 protein is likely resistant to VIHCE despite containing a similar SP sequence.</p>", "<p>The MHC-I SP mimic contained in UL40 sets precedence for CMV taking advantage of the highly conserved SP to escape the cellular immune response. Different from UL40 however, rh178 does not mimic the SP, but seems to rely at least in part on this conserved sequence to broadly eliminate HCs. VIHCE is clearly different from any other previously described immune modulatory mechanism since the ER-localized protein rh178 interferes with HC expression after the onset but prior to the completion of translation. One possible mechanism is that rh178 inhibits translation at a step that occurs after the SRP targets the nascent polypeptide/ribosomal complex to the ER membrane-localized SRP receptor. During this process, translation is arrested until SRP is released upon GTP hydrolysis and SEC61 binding ##REF##10676815##[31]##,##REF##10459008##[44]##. A possible scenario is that rh178 interacts with the SRP/nascent polypeptide/ribosome complex at the ER-membrane thus prolonging translational arrest. Alternatively, rh178 could prevent this complex interaction with the SEC61 translocon in ER-membrane. Conceivably, rh178 could also interfere with the translocation of HC in a manner similar to cotransin, a small molecule translocation inhibitor, which specifically interferes with binding of certain SPs to a SEC61 subunit ##REF##16015336##[45]##. The ensuing translocational stalling results in co-translational degradation by the proteasome, a process that involves cytosolic chaperones ##REF##16923392##[46]##. For non-stop RNA it was recently also shown that translational arrest results in protein fragments that are rapidly degraded by the proteasome ##REF##17344413##[47]##. Therefore, it seems likely that HC translation intermediates are degraded by the proteasome despite the fact that we were unable to detect a degradation intermediate in the presence of proteasome inhibitors. Possible reasons why such breakdown products were not identified are their potentially small and heterogenous size and their extremely rapid degradation. HC-derived intermediates might also lack the epitopes recognized by the HC-specific antibodies used in this study.</p>", "<p>Targeted disruption of protein translation by a viral protein has so far not been described as an immune evasion strategy. However, it was recently shown that the microRNA miR-UL112 of HCMV inhibits the translation of MHC-I-related chain B (MICB), a ligand for the activating NK cell receptor NKG2D ##REF##17641203##[48]##. Thus, CMVs seem to interfere at multiple levels and by multiple strategies with translation of immune stimulatory genes. The virus might thereby employ or mimic cellular pathways of translational or translocational regulation. Further elucidation of the molecular events of VIHCE might thus reveal previously unrecognized host cell mechanisms of translational and translocational control.</p>" ]

[]

[ "<p>Conceived and designed the experiments: CJP KF. Performed the experiments: CJP. Analyzed the data: CJP KF. Wrote the paper: CJP KF.</p>", "<p>The <italic>US2-11</italic> region of human and rhesus cytomegalovirus encodes a conserved family of glycoproteins that inhibit MHC-I assembly with viral peptides, thus preventing cytotoxic T cell recognition. Since HCMV lacking <italic>US2-11</italic> is no longer able to block assembly and transport of MHC-I, we examined whether this is also observed for RhCMV lacking the corresponding region. Unexpectedly, recombinant RhCMV lacking <italic>US2-11</italic> was still able to inhibit MHC-I expression in infected fibroblasts, suggesting the presence of an additional MHC-I evasion mechanism. Progressive deletion analysis of RhCMV-specific genomic regions revealed that MHC-I expression is fully restored upon additional deletion of <italic>rh178</italic>. The protein encoded by this RhCMV-specific open reading frame is anchored in the endoplasmic reticulum membrane. In the presence of rh178, RhCMV prevented MHC-I heavy chain (HC) expression, but did not inhibit mRNA transcription or association of HC mRNA with translating ribosomes. Proteasome inhibitors stabilized a HC degradation intermediate in the absence of rh178, but not in its presence, suggesting that rh178 prevents completion of HC translation. This interference was signal sequence-dependent since replacing the signal peptide with that of CD4 or murine HC rendered human HCs resistant to rh178. We have identified an inhibitor of antigen presentation encoded by rhesus cytomegalovirus unique in both its lack of homology to any other known protein and in its mechanism of action. By preventing signal sequence-dependent HC translocation, rh178 acts prior to US2, US3 and US11 which attack MHC-I proteins after protein synthesis is completed. Rh178 is the first viral protein known to interfere at this step of the MHC-I pathway, thus taking advantage of the conserved nature of HC leader peptides, and represents a new mechanism of translational interference.</p>", "<title>Author Summary</title>", "<p>To avoid immune detection by cytotoxic T lymphocytes, viruses interfere with antigen presentation by major histocompatibility complex class I (MHC-I) molecules. We have discovered a unique cytomegaloviral protein that interferes with the biosynthesis of MHC-I heavy chains and was thus termed viral inhibitor of heavy chain expression (VIHCE). We show that VIHCE does not affect transcription of MHC-I mRNA or the formation of poly-ribosomes. Surprisingly, however, very little MHC-I protein is detected, even when proteasomal protein degradation is inhibited, suggesting incomplete protein translation. Interestingly, VIHCE requires the proper MHC-I signal peptide, suggesting that CMV takes advantage of the high conservation of MHC-I signal peptides and interferes with protein translation by inhibiting signal sequence-dependent protein translocation. This is the first description of a viral protein that specifically targets the translation of a cellular immuno-stimulatory protein.</p>" ]

[ "<title>Supporting Information</title>" ]

[ "<p>We thank Jay Nelson, David Johnson, Scott Wong, Per Peterson and Peter Barry for reagents. We also thank David Johnson for critical reading of the manuscript.</p>" ]

[ "<fig id=\"ppat-1000150-g001\" position=\"float\"><object-id pub-id-type=\"doi\">10.1371/journal.ppat.1000150.g001</object-id><label>Figure 1</label><caption><title>RhCMV inhibits HC expression in the absence of RhUS2-11.</title><p>All experiments were performed at 24 hours post infection at MOI = 3. A) Immunoblot analysis of MHC-I or calreticulin in Mock- or RhCMV-infected TRF lysates. B) IP of total MHC-I upon labeling with ³⁵S-Met/Cys for the indicated time. (*) All IPs from WT and recombinant RhCMV- infected cells contain antibody-binding proteins around 55kDa (see ##SUPPL##1##Fig. S2##) which likely correspond to the RhCMV homologues of the Fc-receptor UL119-118 of HCMV ##REF##12163579##[50]##. Since these viral proteins are not involved in MHC-I inhibition they are not shown in most figures. C) Pulse-chase labeling of 10 min and immunoprecipitation of total MHC-I from Mock-infected, HCMV-infected THFs, or RhCMV-infected TRFs. D) Pulse-labeling of 60 min and IP of MHC-I, Tfn Rec (Transferrin receptor) or Vimentin from Mock-infected or RhCMV-infected TRFs. E) Pulse-chase labeling of 10 min and IP of total MHC-I or HC. Cells were labeled as in 1C, but lysed in SDS buffer prior to IP. F) Pulse-chase labeling and IP of RhCMV-infected TRFs treated with proteasome inhibitor. Where indicated TRFs were incubated with 50 µM MG132 or DMSO during 60-min of Met/Cys starvation, 10-min label, and 30-min chase. For control, TRFs were transduced with AdUS11 (MOI = 25), a recombinant adenovirus expressing HCMV US11, for 24 hours followed by NP40-lysis and IP with K455. Shown for RhCMV-infection is both NP-40 lysis (top panel) and SDS-lysis (bottom panel) prior to IP with the noted antibody.</p></caption></fig>", "<fig id=\"ppat-1000150-g002\" position=\"float\"><object-id pub-id-type=\"doi\">10.1371/journal.ppat.1000150.g002</object-id><label>Figure 2</label><caption><title>Deletion of Rh158–180 restores MHC-I expression during RhCMV infection.</title><p>A) Diagram of the step-wise construction of the ΔRhUS2-11 and Δ158–180,ΔRhUS2-11 viruses. Using the RhCMV BAC the RhUS2-11 region was replaced with a PCR-fragment containing a Kanamycin resistance (Kan^r) cassette flanked by RhCMV homologous regions. The Kan^r cassette was removed by arabinose-induced FLP recombinase prior to replacing the Rh158-180 region with Kan^r. B) Pulse-chase labeling for 10 min of TRFs infected with WT or recombinant RhCMV followed by IP of total MHC-I. In C) 50 µM MG132 or DMSO was included as in ##FIG##0##Fig. 1F##. (*) indicates a deglycosylated cytosolic degradation intermediate stabilized by MG132.</p></caption></fig>", "<fig id=\"ppat-1000150-g003\" position=\"float\"><object-id pub-id-type=\"doi\">10.1371/journal.ppat.1000150.g003</object-id><label>Figure 3</label><caption><title>VIHCE is encoded by <italic>rh178</italic>.</title><p>A) Deletional mapping of VIHCE. Predicted open reading frames between Rh158–180 are shown as open white arrows. Solid black rectangles indicate the region of deletion. Pulse-chase labeling for 10 min with the indicated recombinant virus was performed as in ##FIG##0##Fig. 1C## followed by IP with K455. Lack of VIHCE is readily apparent by the initial synthesis of HC (left lane) followed by US2-11-mediated destruction (right lane). B) Predicted ORFs and experimentally confirmed transcripts in the <italic>rh178</italic> region. The red rectangle indicates the region essential for VIHCE function as determined by deletions in several independent recombinants. Large black arrows indicate positions of ORFs <italic>rh175–178</italic> predicted by ##REF##12767982##[12]##. Transcripts confirmed by RACE and cDNA PCR are shown below. C) Northern blot analysis of total RNA isolated from mock or WT RhCMV-infected TRFs at 24 hours post infection. ORF <italic>rh178</italic> was used to generate ³²P-dCTP labeled DNA probe. D) Complementary sequence of the RhCMV genome from 181921–182060bp. Underlined at 182058bp is the original predicted start codon for rh178 ##REF##12767982##[12]##. Transcription actually begins at 182015bp as determined by 5′ RACE (see sequence chromatogram below genomic sequence). Shaded in gray is the first ATG codon of the transcript. Also noted is the splice donor site for <italic>rh178.4</italic> which is spliced at 181944bp.</p></caption></fig>", "<fig id=\"ppat-1000150-g004\" position=\"float\"><object-id pub-id-type=\"doi\">10.1371/journal.ppat.1000150.g004</object-id><label>Figure 4</label><caption><title>A frameshift mutation of <italic>rh178</italic> restores HC expression.</title><p>A) Sequence of the rh178 frameshift control and frameshift recombinants. Shown is complementary genomic sequence, with transcripts running from right to left. In each recombinant a 20bp sequence in the 5′ UTR of <italic>rh178</italic> (gray boxes) was replaced with 93bp from the recombination vector including the FRT recombination site. (*) indicates the single base insertion causing a frameshift. B) and C) HC expression in TRFs infected with control or frameshift viruses. Pulse-chase and IP was performed as in ##FIG##0##Fig. 1C##.</p></caption></fig>", "<fig id=\"ppat-1000150-g005\" position=\"float\"><object-id pub-id-type=\"doi\">10.1371/journal.ppat.1000150.g005</object-id><label>Figure 5</label><caption><title>rh178 is a 212aa non-glycosylated ER localized protein that is sufficient to block HC synthesis.</title><p>A) Complete polypeptide sequence of rh178. Shaded in gray is the predicted signal anchor sequence. B) Hydrophobicity graph of rh178 (TopPred, <ext-link ext-link-type=\"uri\" xlink:href=\"http://bioweb2.pasteur.fr/\">http://bioweb2.pasteur.fr/</ext-link>). TM refers to a predicted transmembrane domain cutoff value. C) Western blot of lysate from TRFs transduced with replication deficient adenovirus vectors AdTrans (expressing a tetracycline responsive transactivator) or AdTrans together with Ad178-HA (expressing HA-tagged rh178) for 48 hours. Lysate was treated without or with PNGase to remove N-linked sugars and blotted for MHC-I HC using the HC-10 antibody or for rh178-HA with an anti-HA antibody. D) HC expression in TRFs transduced with AdTrans or AdTrans with Ad178 (expressing wild type rh178) for 24 hours, followed by a 10-min pulse label and 30-min chase. HCs were recovered with K455 from NP40 lysates. E) Immunofluorescence analysis of TRFs 24 hours after transfection with HA-tagged rh178 together with FLAG-tagged K5 from KSHV. Primary antibodies were mouse anti-FLAG and rabbit anti-HA. Secondary antibodies were 594 Alexa Fluor conjugated goat anti-rabbit and 488 Alexa Fluor conjugated goat anti-mouse.</p></caption></fig>", "<fig id=\"ppat-1000150-g006\" position=\"float\"><object-id pub-id-type=\"doi\">10.1371/journal.ppat.1000150.g006</object-id><label>Figure 6</label><caption><title>HC mRNA is present, intact, and associates with actively translating ribosomes during RhCMV infection.</title><p>A) Northern blot analysis of HC- or GAPDH-specific mRNA from total RNA isolated at 24 hours after Mock- or RhCMV-infection. The ³²P-dCTP labeled probes were generated using rhesus-derived cDNAs for HC or GAPDH as templates. B) Polyribosome fractionation and northern blot analysis. TRFs were either mock infected or infected with wild-type (WT) RhCMV at MOI = 3 for 24 hours followed by isolation and fractionation of polysomes. Ethidium Bromide (EtBr) staining of a denaturing agarose gel shows the amount and ratio of 18S and 28S rRNA present in each fraction, indicating the presence of ribosomal subunits. Polysomes sediment to higher, denser fractions. Lower panels show northern blots of the gel using the HC and GAPDH-specific probes. C) Cells were infected as in B. However, after 24 hours, cells were incubated for 4 min with either DMSO or 100 µg/ml puromycin prior to polysome harvesting.</p></caption></fig>", "<fig id=\"ppat-1000150-g007\" position=\"float\"><object-id pub-id-type=\"doi\">10.1371/journal.ppat.1000150.g007</object-id><label>Figure 7</label><caption><title>Efficient HC targeting by rh178 is signal-peptide dependent.</title><p>A) Rh178 inhibits expression of human HC. THFs were infected with the indicated virus at MOI = 3 for 24 hours, followed by a 10-min pulse-label, a chase of 30-min and IP with K455. B) UTR-independent inhibition of HLA-A3 expression by rh178. TRFs were electroporated with pEF1α containg the indicated HLA-A3 construct. After 24 hours, cells were either mock infected or infected with recombinant RhCMV (MOI = 3) containing VIHCE (+; ΔRhUS2-11) or lacking VIHCE (−; Δ178, ΔRhUS2-11). After an additional 24 hours, cells were labeled for 30-min, lysed in NP-40, and HLA-A3 was immunoprecipitated. C) Upper panel: Amino acid sequence of the signal peptides used in chimeric HLA-A3 HCs. Gray shading indicates identity with the HLA-A3 signal peptide. Lower panel: TRFs were electroporated with native HLA-A3 (A3) or the indicated SP-chimera (the HLA-A3 signal peptide was replaced with the H2-K^b or the CD4 signal peptide in K^b/A3 or CD4/A3, respectively) prior to infection with RhCMV, metabolic labeling and IP as in 7B. (*) indicates an uncharacterized HC-band that appears prominently in IPs from CD4/A3 transfectants and that could represent a deglycosylated or truncated HC. D) Quantitation of HLA-A3, total HC, or CD4 expression from 7C and 7E shown as a percent relative to HC or CD4 levels in the absence of VIHCE. Bands were quantitated using ImageQuant 5.1 software (Molecular Dynamics). E) TRFs were electroporated with native CD4 or CD4 containing the HLA-A3 SP (A3/CD4) and treated as in 7C. All experiments are representative of several replicates.</p></caption></fig>" ]

[]

[]

[]

[]

[]

[]

[ "<supplementary-material content-type=\"local-data\" id=\"ppat.1000150.s001\"><label>Figure S1</label><caption><p>HC synthesis is not delayed nor rapidly degraded upon synthesis during RhCMV infection. A) HC synthesis is not delayed. Cells were radiolabeled for 10 min followed by chase of indicated times. After SDS lysis, IP was performed using HC-10 antibody, which recognizes free MHC-I HC. (*) A non-MHC-I-specific band indicating protein loading. B) HC is not rapidly degraded upon synthesis. TRFs were infected with the indicated virus, radiolabeled for 1 min, chased for 30 min, lysed with NP-40 lysis buffer and IP performed with K455.</p><p>(1.59 MB TIF)</p></caption></supplementary-material>", "<supplementary-material content-type=\"local-data\" id=\"ppat.1000150.s002\"><label>Figure S2</label><caption><p>RhCMV contains viral antibody binding proteins that are not specific to the immunoprecipitated antigen. Complete autoradiograph from ##FIG##1##Fig 2B## showing pulse-chase and IP during infection with RhCMV Δ158–180 and Δ158–180, ΔRhUS2-11. Indicated on the left side are molecular weight estimates. This indicates the viral antibody binding proteins that are not shown in IPs from other figures since they are non-specific to the immunoprecipitated antigen.</p><p>(1.97 MB TIF)</p></caption></supplementary-material>", "<supplementary-material content-type=\"local-data\" id=\"ppat.1000150.s003\"><label>Figure S3</label><caption><p>\n<italic>rh178</italic> is expressed as an early gene transcript. Northern blot analysis of rh178 and Rh156 (IE1) at 4 and 24 hours post infection. Cyclohexamide (CHX) and phosphonoacetic acid (PAA) were included where indicated. Note that PAA did not inhibit VIHCE expression indicating that VIHCE is not a late gene. In contrast, CHX inhibited VIHCE expression indicating that VIHCE is not an immediate early gene.</p><p>(0.96 MB TIF)</p></caption></supplementary-material>", "<supplementary-material content-type=\"local-data\" id=\"ppat.1000150.s004\"><label>Protocol S1</label><caption><p>Supplemental materials and methods and figure legends.</p><p>(0.04 MB DOC)</p></caption></supplementary-material>", "<supplementary-material content-type=\"local-data\" id=\"ppat.1000150.s005\"><label>Table S1</label><caption><p>Sequences of the recombination portion of the BAC mutagenesis primers.</p><p>(0.03 MB DOC)</p></caption></supplementary-material>" ]

[ "<fn-group><fn fn-type=\"COI-statement\"><p>The authors have declared that no competing interests exist.</p></fn><fn fn-type=\"financial-disclosure\"><p>This work was supported by NIH (RO1 AI 0594457) to KF and by NCRR support for the Oregon National Primate Resource Center (RR00163). CP was supported by a Ruth L. Kirschstein National Research Service Award (T32AI007472) and an OHSU Tartar Trust Fellowship.</p></fn></fn-group>" ]

[ "<graphic xlink:href=\"ppat.1000150.g001\"/>", "<graphic xlink:href=\"ppat.1000150.g002\"/>", "<graphic xlink:href=\"ppat.1000150.g003\"/>", "<graphic xlink:href=\"ppat.1000150.g004\"/>", "<graphic xlink:href=\"ppat.1000150.g005\"/>", "<graphic xlink:href=\"ppat.1000150.g006\"/>", "<graphic xlink:href=\"ppat.1000150.g007\"/>" ]

[ "<media xlink:href=\"ppat.1000150.s001.tif\"><caption><p>Click here for additional data file.</p></caption></media>", "<media xlink:href=\"ppat.1000150.s002.tif\"><caption><p>Click here for additional data file.</p></caption></media>", "<media xlink:href=\"ppat.1000150.s003.tif\"><caption><p>Click here for additional data file.</p></caption></media>", "<media xlink:href=\"ppat.1000150.s004.doc\"><caption><p>Click here for additional data file.</p></caption></media>", "<media xlink:href=\"ppat.1000150.s005.doc\"><caption><p>Click here for additional data file.</p></caption></media>" ]

{ "acronym": [], "definition": [] }

[ "<title>Introduction</title>", "<p>The Retinoblastoma tumor suppressor protein (pRB) and the related proteins p107 and p130 negatively regulate cell proliferation. In a textbook model, the role of pRB family members in cell cycle regulation is explained by their ability to attenuate the activity of E2F transcription factors. E2F is best known for its ability to control the G1/S transition and is rate limiting for S phase entry (for review see: ##REF##17100600##[1]##–##REF##11823794##[4]##). The E2F transcriptional program provides cell cycle dependent expression of a large panel of genes encoding replication proteins, cell cycle regulators and others. In early G1 phase, members of the pRB family are complexed with members of the E2F family and repress expression of E2F regulated genes through recruitment of corepressor complexes to target promoters. In late G1 phase, cyclin dependent kinases phosphorylate pRB family members, thus releasing free E2F proteins to allow induction of E2F-dependent transcription. Since functional inactivation of the pRB pathway occurs in most tumor cells it is thought that unrestrained E2F activity drives inappropriate proliferation in tumors ##REF##10647931##[5]##. Such an idea is further supported by findings that mutations in E2f genes reduce proliferation in <italic>Rb</italic> deficient mouse embryos ##REF##12944480##[6]##–##REF##12498715##[8]##.</p>", "<p>In mammalian cells, E2F activity is a combined output of eight family members, which, in turn, are loosely grouped into a class of repressors (E2F-3b through E2F-8) and a class of activators (E2F-1 through E2F3a). E2F-1 through E2F-6 require a heterodimeric partner of the DP family of proteins to bind to DNA, while E2F-7 and E2F-8 bind to DNA in a DP-independent manner. As a way to dissect the contribution of E2F to cell proliferation, dominant negative forms of DP and E2F, dn-DP and dn-E2F respectively, were used. Expression of dn-E2F, which binds to DNA, but fails to repress or activate, leads to immortalization in mouse fibroblasts and renders cells resistant to senescence induced by p19^ARF, p53 or by RAS^V12\n##REF##12150825##[9]##. However, cells expressing dn-E2F were impaired in the ability to proliferate following serum stimulation. This suggests that E2F activity is not needed during cell proliferation but is required in a specific context, such as cell cycle re-entry from quiescence. In contrast, a reduction of DP function, either by siRNA or by using a dn-DP form, resulted in cell cycle arrest and a senescence-like phenotype, indicating that E2F is in fact needed for cell proliferation ##REF##8668186##[10]##,##REF##15716376##[11]##. One potential explanation for these discrepancies is that reducing DP does not inactivate the total pool of E2Fs, since E2F-7 and E2F-8 repressors bind to DNA in a DP-independent manner, and therefore the two remaining E2Fs may induce the cell cycle arrest. An alternative explanation is that dn-E2F does not completely inhibit E2F activity ##REF##15716376##[11]## and the remaining E2F activity is sufficient to sustain cell proliferation.</p>", "<p>The biological role of E2F in the context of animal development is being extensively studied by using gene targeting approaches in mice. However, interpretation of the phenotypes of individual E2F knockouts is often complicated by the redundancy and compensation among the family members. The impact of genetic ablation of E2f genes on cell proliferation is more profound in compound knockouts. Mouse embryonic fibroblasts (MEFs) lacking a whole class of activator E2Fs, <italic>E2f-1, E2f-2</italic> and <italic>E2f-3</italic>, fail to proliferate due to a high level of p21 ##REF##11719808##[12]##, <italic>E2f-4</italic>; <italic>E2f-5</italic> double knockouts are defective in a p16 mediated cell cycle arrest ##REF##11030352##[13]##, while <italic>E2f-7</italic>; <italic>E2f-8</italic> knockouts have a high level of apoptosis due to deregulation of <italic>E2f-1</italic> expression ##REF##18194653##[14]##. Nevertheless, the large number of E2F genes makes it currently unfeasible to genetically ablate all E2F activity in mammals to determine the consequences of the loss of E2F function on cell proliferation.</p>", "<p>Genetically, <italic>Drosophila</italic> provides a relatively simpler system to study the role of E2F, since the corresponding families are smaller. The <italic>Drosophila</italic> E2F family consists of a single activator, dE2F1, and a lone repressor, dE2F2. Both dE2Fs dimerize with the single dDP protein in order to bind to DNA. Unlike mammalian cells, the <italic>Drosophila</italic> genome lacks orthologs of E2F-7 or E2F-8 that bind to DNA in a dDP independent manner. It is important to note that the loss of <italic>dDP</italic> has been shown to functionally inactivate both dE2F1 and dE2F2 ##REF##15798191##[15]##. A <italic>de2f1</italic> mutation severely reduces cell proliferation, leads to the loss of expression of E2F target genes, and almost complete cessation of DNA synthesis ##REF##7601349##[16]##,##REF##9271122##[17]##. Strikingly, these defects are largely suppressed by a concomitant mutation in <italic>de2f2</italic>. <italic>de2f1 de2f2</italic> double mutant animals can survive until late pupal stages and show normal patterns of cell proliferation and differentiation even though E2F targets are no longer expressed in a cell cycle dependent manner and are likely to be present at suboptimal levels ##REF##11511545##[18]##. A similar phenotype has been observed in <italic>dDP</italic> mutant animals. Thus, the complete loss of E2F function in <italic>dDP</italic> or in <italic>de2f1 de2f2</italic> mutants is permissive for cell proliferation and appears to have a relatively minor impact on animal development; however, whether the loss of E2F affects cell proliferation during oncogenic stimuli has not been studied. Given that most models emphasize the prominent role of E2F in proliferation of tumor cells, this is an important question to be addressed.</p>", "<p>\n<italic>Drosophila</italic> has proven to be an excellent model to investigate cancer-causing genes. This is illustrated, for example, by studies of the recently identified the Hippo tumor suppressor pathway. The cellular functions of the Hippo pathway are to restrict cell proliferation and promote apoptosis (for review see: ##REF##17437995##[19]##–##REF##17174912##[22]##). The core components of the pathway are the protein kinases Warts (Wts), Hippo (Hpo) and Mob as tumor suppressor (Mats). Salvador (Sav) serves a scaffold for Wts and Hpo. Assembly of an active complex of the four negative regulators Wts/Hpo/Mats/Sav is accompanied by mutual phosphorylation and leads to activation of the Wts kinase. Once active, Wts phosphorylates and inactivates the transcriptional co-activator Yorkie (Yki) by excluding Yki from nucleus. Yki is thus far the most downstream component of the pathway. In the absence of Wts-dependent phosphorylation, Yki enters the nucleus where it requires transcription factors to be recruited to the promoter of the Hippo pathway target genes. Thus far, only the lone TEAD/TEF protein family member in <italic>Drosophila</italic>, Scalloped, has been shown to interact with Yki ##REF##18258485##[23]##,##REF##18258486##[24]## while the Yki mammalian homolog YAP binds to a variety of transcriptional factors and modulates their activity. Among Hippo pathway targets are genes that promote cell proliferation and genes that inhibit apoptosis such as <italic>cyclin E</italic>, microRNA <italic>bantam,</italic> and <italic>diap1</italic>. Inactivation of any negative regulator of Hippo pathway signaling, or overexpression of the positive regulator Yki, stimulates additional cell divisions by increasing the proliferation rate of actively dividing cells, delaying the cell cycle exit, and simultaneously protecting cells from apoptosis. Failure to exit the cell cycle on time gives rise to inappropriate proliferation of Hippo pathway mutant cells. Since patterns of cell proliferation are normal in <italic>dDP</italic> mutants and in <italic>de2f1 de2f2</italic> double mutants, these combinations provide us with an opportunity to determine when and where proliferation driven by the potent oncogene Yki is dependent on dE2F family members. In this work, we show that the loss of E2F function produces a distinctly different result in actively dividing cells and in cells that proliferate inappropriately due to the failure to exit the cell cycle. Inactivation of the entire dE2F family in actively dividing cells has only a subtle effect on the ability of Yki to increase the rate of cell division. In contrast, the loss of E2F function fully blocks inappropriate proliferation of these cells. Thus, our work uncovers the <italic>in vivo</italic> requirement for E2F function during oncogenic proliferation driven by Yki, specifically at the point when cells normally exit the cell cycle and enter quiescence.</p>" ]

[ "<title>Materials and Methods</title>", "<title>Fly Stocks and Mosaic Analysis</title>", "<p>For mutant analysis, the following strong loss of function or null alleles were used:</p>", "<p>\n<italic>de2f1⁷²⁹</italic>, <italic>de2f2^c03344</italic>, <italic>wts^X1</italic>, <italic>hpo^MGH4</italic>, <italic>yki^B5</italic>, <italic>ago¹</italic>, and <italic>dDP^a4</italic>.</p>", "<p>Clones of homozygous mutant cells were generated with the <italic>ey-</italic>FLP/FRT technique. For clones of <italic>ago¹</italic> mutant cells the following genotype was used:</p>", "<p>\n<italic>ey-FLP</italic>; <italic>ago¹ FRT80B/P[Ubi-GFP] FRT80B</italic>\n</p>", "<p>For clones of double mutant cells of <italic>de2f2 wts</italic> the following genotype was used:</p>", "<p>\n<italic>ey-FLP</italic>; <italic>de2f2^c03344FRT40A/P[Ubi-GFP] FRT40A</italic>; <italic>FRT82B wts^X1/FRT82B P[Ubi-GFP]</italic>\n</p>", "<p>For clones of double mutant cells of <italic>wts de2f1</italic> the following genotype was used:</p>", "<p>\n<italic>ey-FLP</italic>; <italic>FRT82B de2f1⁷²⁹ wts^X1/ FRT82B P[Ubi-GFP]</italic>\n</p>", "<p>Clones of triple mutant cells of <italic>de2f1 de2f2 wts</italic> were generated in:</p>", "<p>\n<italic>ey-FLP</italic>; <italic>de2f2^c03344FRT40A/P[Ubi-GFP] FRT40A</italic>; <italic>FRT82B de2f1⁷²⁹ wts^X1/FRT82B P[Ubi-GFP]</italic>\n</p>", "<p>Activity at the <italic>de2f1</italic> promoter in <italic>hpo</italic> mutant tissue was determined in larvae of the following genotype:</p>", "<p>\n<italic>ey-FLP</italic>; <italic>FRT42D hpo^MGH4</italic>/<italic>FRT42D P[Ubi-GFP]</italic>; <italic>FRT82B de2f1⁷²⁹/</italic>+</p>", "<p>To determine the expression of the E2F reporter, <italic>PCNA-GFP</italic>, in <italic>wts</italic> mutant and in <italic>de2f1 wts</italic> double mutant cells the following genotypes were used:</p>", "<p>\n<italic>ey-FLP</italic>; <italic>PCNA-GFP/+; FRT82B wts^X1/FRT82B P[arm-LacZ]</italic>\n</p>", "<p>\n<italic>ey-FLP</italic>; <italic>PCNA-GFP/+; FRT82B de2f1⁷²⁹ wts^X1/FRT82B P[arm-LacZ]</italic>\n</p>", "<p>Analysis of <italic>yki</italic> overexpression was done with the MARCM system [25] in larvae of the following genotypes:</p>", "<p>\n<italic>y w hs-FLP70 tub-GAL4 UAS-GFP-6XMyc.NLS</italic>; <italic>FRT42D tub-GAL80/FRT42D dDP^a4</italic>; <italic>UAS-Yki/+</italic>\n</p>", "<p>\n<italic>y w hs-FLP70 tub-GAL4 UAS-GFP-6XMyc.NLS</italic>; <italic>FRT42D tub-GAL80/FRT42D</italic>; <italic>UAS-Yki/+</italic>\n</p>", "<p>\n<italic>y w hs-FLP70 tub-GAL4 UAS-GFP-6XMyc.NLS</italic>; <italic>FRT42D tub-GAL80/FRT42D</italic>\n</p>", "<p>\n<italic>eyFLP UAS-GFP</italic>; <italic>tub-GAL4 FRT82B P[UAS-yki]/FRT82B tub-GAL80</italic>\n</p>", "<p>Analysis of cyclin E overexpression was done with the MARCM system ##REF##11311363##[25]## in:</p>", "<p>\n<italic>y w hs-FLP70 tub-GAL4 UAS-GFP-6XMyc.NLS; FRT42D tub-GAL80/FRT42D; UAS-cyclinE/+</italic>\n</p>", "<p>To determine the cell population doubling time, clones of <italic>dDP</italic> mutant cells were induced 48 hrs AED and discs were dissected and fixed 66–70 hrs later. A standard error of the mean was calculated for each genotype. To determine number of interommatidial cells for the pupal retinae, bristle, secondary, and tertiary cells for one ommatidium were counted. One ommatidium was defined as a single cluster of 4 cone cells, 2 primary cells, 3 bristle cells, 3 tertiary cells, and 6 secondary cells. A standard deviation of the mean was taken to determine significance. To measure the area per clone the histogram function in Adobe Photoshop was used and a standard error of the mean was taken to determine significance.</p>", "<title>Immunohistochemistry</title>", "<p>Antibodies used were as follows: mouse anti-cyclinE 1∶20 (from B. Edgar), guinea pig anti-Expanded 1∶500 (from R. Fehon), mouse anti-Discs Large 1∶400 (DSHB), mouse anti-BrdU 1∶50 (Beckton Dickinson ), rabbit anti-dE2F1 1∶400 (from C. Seum), mouse anti-β-galactosidase 1∶30 (DSHB), rabbit anti-GFP 1∶200 (Invitrogen), rabbit anti-C3 (Cleaved Caspase3) 1∶100 (Cell Signaling), rabbit anti-dE2F2 1∶100, rabbit anti-phosH3 1∶175 (Upstate), and Cy3, Cy5 (Jackson Immunolaboratories) and Alexa488 (Invitrogen) conjugated anti-mouse and anti-rabbit secondary antibodies. Larval and pupal tissues were fixed in 4% formaldehyde for 30 minutes on ice, washed in phosphate-buffered saline, and then incubated with antibodies overnight at 4°C in phosphate-buffered saline, 10% normal goat serum, and 0.3% Triton-X100 as previously described ##REF##17923695##[29]##. To detect dE2F1 protein, fixation was adjusted to 40 minutes on ice and then treated as described above. To detect cyclin E protein in larval eye imaginal disc, PLP fixation was used and then the same protocol described above was used. To detect S phases, dissected larval or pupal eye discs were labeled with BrdU for 2 hrs at room temperature and then the eye discs were fixed overnight in 1.5% formaldehyde at 4°C. Apoptosis was measured in pupal eye discs 30 hrs APF. To measure apoptosis following DNA damage, larvae were exposed to 40 Grays of irradiation and then imaginal discs were dissected 4 hrs later.</p>", "<title>S2 Cell Manipulations</title>", "<p>RNAi and transient transfections were done as described previously##REF##17923695##[29]##. For western blot analysis, S2 cells were lysed in NP40 buffer, frozen at −80°C for 1 hr, thawed, spun down, and then boiled in protein sample buffer. Samples were resolved using SDS-PAGE on 7% gels, transferred to Immobilon-P membrane (Millipore), and incubated with the following antibodies: mouse anti-E7(tubulin) 1∶9000 (DSHB), guinea pig anti-Wts 1∶10,000 (From K. Irvine), rabbit anti-dE2F2 1∶2,000, guinea pig anti-dE2F1 1∶7,000 (from T. Orr-Weaver).</p>" ]

[ "<title>Results</title>", "<title>Yki-Driven Proliferation of Actively Dividing Cells in the Wing Imaginal Disc Does Not Require E2F Mediated Control</title>", "<p>We initially used a <italic>dDP</italic> mutation to determine whether Yki requires dE2F mediated control to drive cell proliferation. dDP is an obligatory heterodimeric partner of both dE2F1 and dE2F2, and the loss of <italic>dDP</italic> has been shown to functionally inactivate both <italic>de2f1</italic> and <italic>de2f2</italic>\n##REF##15798191##[15]##. Importantly, <italic>dDP</italic> single mutant or <italic>de2f1 de2f2</italic> double mutant animals survive until late pupal stages and exhibit normal patterns of cell proliferation ##REF##9271122##[17]##,##REF##11511545##[18]##. Thus, the use of these mutant combinations allowed us to minimize indirect cell cycle effects produced when dE2F1 function alone is inactivated. The MARCM technique ##REF##11311363##[25]## was employed to generate clones of wild-type cells, wild-type cells overexpressing <italic>yki</italic>, <italic>dDP</italic> mutant cells, and <italic>dDP</italic> mutant cells that overexpress <italic>yki</italic> in the larval wing imaginal disc. All clones were marked with GFP, induced simultaneously, and allowed to grow for the same period of time (##FIG##0##Figure 1##). At this stage of development, the majority of cells in the wing disc are asynchronously dividing and therefore the rates of cell proliferation can be accurately measured. As expected, overexpression of <italic>yki</italic> accelerated the cell cycle progression of wild type cells (##FIG##0##Figure 1A–B##). The median population doubling time in <italic>yki</italic> overexpressing clones was 10.6 hr. This was faster than that of the wild type population, which was 13.8 hr. We found that a <italic>dDP</italic> mutation did not significantly affect the ability of <italic>yki</italic> to increase rates of cell division (##FIG##0##Figure 1C–D##). The median population doubling time in clones of <italic>dDP</italic> mutant cells overexpressing <italic>yki</italic> was 11.3 hr, indicating that these cells were still proliferating faster than wild type cells. Thus, this result suggests that E2F mediated control is not required for Yki induced proliferation in actively dividing cells.</p>", "<title>The Loss of E2F–Dependent Control Blocks Yki-Driven Inappropriate Proliferation in the Cells Posterior to the SMW in the Eye Imaginal Disc</title>", "<p>In the eye imaginal disc, Hippo pathway mutant cells delay the cell cycle exit and undergo inappropriate proliferation. To determine the effect of the loss of E2F mediated regulation in these settings, we examined the effect of <italic>yki</italic> overexpression in clones of wild type or <italic>dDP</italic> mutant cells during cell cycle exit. In a wild type eye disc, BrdU labeling reveals a narrow stripe of S phase cells posterior to the morphogenetic furrow (MF), referred to as the Second Mitotic Wave (SMW) (##FIG##1##Figure 2A##). Cells within the MF are arrested in G1 and therefore do not incorporate BrdU. Posterior to the SMW, cells exit the cell cycle and differentiate. In contrast, <italic>yki</italic> overexpressing cells (GFP positive) failed to undergo cell cycle exit and continued proliferating ##REF##16096061##[26]##, as shown by the appearance of BrdU positive cells posterior to the SMW (##FIG##1##Figure 2B##). Strikingly, no ectopic BrdU incorporation was observed when <italic>yki</italic> was overexpressed in clones of <italic>dDP</italic> mutant cells, which are marked by the presence of GFP (##FIG##1##Figure 2C##). This indicates that although E2F function is unnecessary for Yki to increase rates of cell proliferation in actively dividing cells, Yki is dependent on the presence of dE2F/dDP activity to drive cells into inappropriate cell cycles posterior to the SMW.</p>", "<p>To further confirm this conclusion, we examined cell proliferation when both <italic>de2f1</italic> and <italic>de2f2</italic> were genetically ablated while the Hippo pathway was inactivated by a <italic>wts</italic> mutation. Clones of <italic>de2f2</italic> single, <italic>de2f1 wts</italic> double, and <italic>de2f2 de2f1 wts</italic> triple mutant cells were simultaneously generated in the same eye imaginal disc using the <italic>ey-</italic>FLP/FRT technique. The triple mutant tissue could be distinguished from the neighboring wild type tissue by the complete lack of GFP. Similar to <italic>yki</italic> overexpressing cells, <italic>wts</italic> mutant cells failed to exit the cell cycle and proliferated inappropriately posterior to the SMW ##REF##12941273##[27]##,##REF##12941274##[28]## (##FIG##1##Figure 2D##). Additionally, <italic>wts</italic> mutant cells continue cell divisions during early pupal development when wild type cells are quiescent (##FIG##1##Figure 2G##). The inappropriate proliferation during larval and pupal stages gives rise to a surplus of interommatidial cells. During pupal eye development, the excess of interommatidial cells is eliminated by a wave of apoptosis. However, since <italic>wts</italic> mutants are defective in normal apoptosis in the eye, these supernumerary cells remain and can be visualized in 48 hr old pupal retina as extra layers of cells between adjacent ommatidial clusters (##FIG##1##Figure 2J##). Consistent with the results of the overexpression of <italic>yki</italic> in <italic>dDP</italic> mutant cells (##FIG##1##Figure 2C##), clones of <italic>de2f2 de2f1 wts</italic> mutant cells posterior to the SMW were largely devoid of BrdU incorporation and mitoses, the latter were detected by the appearance of phosphorylated histone H3 (phosH3), (##FIG##1##Figure 2E–F##). No S phases were detected in the triple mutant combination at 12 hr after puparium formation either (##FIG##1##Figure 2H##), a time point when <italic>wts</italic> mutant cells continue inappropriate proliferation (##FIG##1##Figure 2G##). Furthermore, examination of pupal retinas revealed that clones of <italic>de2f1 de2f2 wts</italic> triple mutant cells (##FIG##1##Figure 2K##) or <italic>dDP</italic> mutant cells that overexpress <italic>yki</italic> (##FIG##1##Figure 2M##) no longer contain an abnormally large number of supernumerary interommatidial cells which are otherwise found in clones of <italic>wts</italic> mutant cells (##FIG##1##Figure 2J##) or in clones of cells that overexpress <italic>yki</italic> (##FIG##1##Figure 2L##). To measure the extent to which the loss of dE2F function reduced the number of interommatidial cells in <italic>wts</italic> mutant tissue, we counted the number of secondary, tertiary, and bristles cells per each ommatidial hexagon. Clones of <italic>wts</italic> mutant tissue in pupal retinas contained an average of 31.6±2.6 cells, which was significantly higher than 12.0±0.1 cells found in wild type tissue (##FIG##1##Figure 2N##). However, the regions that were triple mutant for <italic>de2f1 de2f2 wts</italic> had only 13.7±1.0 cells. Similarly, a <italic>dDP</italic> mutation significantly reduced the number of cells in clones that overexpress <italic>yki</italic> from 30.3±3.3 down to 15.0±1.0 (##FIG##1##Figure 2N##). These reductions are consistent with the observations that these cells fail to proliferate posterior to the SMW. We further emphasize that the inability of <italic>wts</italic> mutant cells, or <italic>yki</italic> overexpressing cells, to undergo inappropriate proliferation in the absence of E2F control is not merely a consequence of non-specific cell cycle defects due to inactivation of the dE2F family. Most cell proliferation occurs normally in <italic>de2f1 de2f2</italic> double mutants ##REF##11511545##[18]## or <italic>dDP</italic> mutants ##REF##9271122##[17]## and, as shown here, a <italic>dDP</italic> mutation did not prevent <italic>yki</italic> from increasing rates of cell division in asynchronously dividing cells of the wing disc (##FIG##0##Figure 1##).</p>", "<p>Next, we wished to determine whether both dE2F family members are equally important in <italic>wts</italic> mutant cells to undergo inappropriate proliferation. To address this question we compared the S phases posterior to the SMW in clones of <italic>de2f2 wts</italic> double mutant cells with that of <italic>wts</italic> single mutant cells in the same eye imaginal disc. Previous analysis revealed that patterns of cell proliferation are normal in clones of <italic>de2f2</italic> mutant cells ##REF##17923695##[29]##. The <italic>de2f2 wts</italic> double mutant cells were marked by the lack of GFP, <italic>wts</italic> single mutant cells were distinguished by an intermediate GFP signal and by the increased spacing between ommatidial clusters, while the wild type cells had the highest intensity of GFP (##FIG##2##Figure 3A##). The loss of <italic>de2f2</italic> had no effect on inappropriate S phases posterior to the SMW in <italic>wts</italic> mutant cells, as the phenotype of <italic>de2f2 wts</italic> mutant cells was indistinguishable from the phenotype of <italic>wts</italic> mutant cells (##FIG##2##Figure 3B–D##). Additionally, the spacing between adjacent ommatidial clusters (marked with Elav) is increased in clones of <italic>de2f2 wts</italic> double mutant cells when compared to that of wild type cells. Such increase reflects the appearance of additional interommatidial cells in the <italic>wts</italic> mutant tissue ##REF##12941273##[27]##. Taken together with the results described above, this indicates that <italic>de2f2</italic> is not important for the phenotype of <italic>wts</italic> mutant cells and that <italic>de2f1</italic> is required for inappropriate proliferation of <italic>wts</italic> mutant cells. Consistently, no ectopic S phases were detected in <italic>de2f1 wts</italic> double mutant clones (data not shown).</p>", "<p>In addition to delaying the cell cycle exit, the loss of Hippo pathway protects cells from both developmental and stress induced apoptosis ##REF##17437995##[19]##–##REF##17174912##[22]##. Although a mutation in <italic>de2f1</italic> blocked inappropriate proliferation in Hippo pathway mutant cells, <italic>de2f1 wts</italic> double mutant cells, like <italic>wts</italic> single mutant cells, were fully resistant to a wave of apoptosis that normally occurs during pupal eye development (##FIG##3##Figure 4A–B##) or to DNA damage induced apoptosis in the larval eye disc following irradiation (##FIG##3##Figure 4C##). Thus, resistance to apoptosis, a hallmark of inactivation of Hippo pathway, remains unaffected by the loss of <italic>de2f1</italic>. This sustained resistance to apoptosis is a likely explanation for the slight increase of the number of interommatidial cells in <italic>de2f1 de2f2 wts</italic> mutant tissue and in <italic>dDP</italic> mutant cells overexpressing <italic>yki</italic> in comparison to wild type (##FIG##1##Figure 2N##). From these data we concluded that <italic>de2f1</italic> is specifically required in Hippo pathway mutant cells to delay the cell cycle exit and sustain inappropriate proliferation posterior to the SMW.</p>", "<title>The Loss of E2F–Mediated Control Does Not Block Induction of Yki Target Genes</title>", "<p>A trivial explanation for the lack of inappropriate proliferation in <italic>de2f2 de2f1 wts</italic> is that inactivation of dE2F family members renders Yki inactive in these cells. To directly address this question we examined whether Yki is capable of inducing its target genes in <italic>dDP</italic> mutant cells. <italic>Drosophila</italic> inhibitor of apoptosis, dIAP1, has been recently shown to be a direct Yki transcriptional target ##REF##18258485##[23]##,##REF##18258486##[24]## and, together with Expanded ##REF##16341207##[30]##, are commonly used to accurately assess the activity of the Hippo pathway. Notably, upregulation of Expanded and dIAP1 following <italic>yki</italic> overexpression was observed in wild type cells (##FIG##4##Figure 5A, 5C##) and to almost the same extent in clones of <italic>dDP</italic> mutant cells (##FIG##4##Figure 5B, 5D##). This suggests that the loss of E2F regulation does not prevent induction of at least two well established Yki targets.</p>", "<p>Next, we investigated the expression of cyclin E since it is an E2F target ##REF##15838517##[2]## and is also considered to be a critical target of the Hippo pathway ##REF##17437995##[19]##–##REF##17174912##[22]##. Clones of <italic>wts</italic> single and <italic>de2f1 wts</italic> double mutant cells were generated. As shown in ##FIG##5##Figure 6##, the level of cyclin E was elevated in <italic>wts</italic> mutant cells and in <italic>de2f1 wts</italic> double mutant cells. Thus, the observation that induction of multiple Yki target genes is not compromised by the loss of E2F control implies that Yki remains active in dE2F deficient cells. Such a conclusion is in agreement with the resistance to apoptosis of <italic>de2f1 wts</italic> mutant cells, further evidence that Yki is fully functional in the absence of dE2F family.</p>", "<title>Inactivation of the Hippo Pathway Leads to Elevation of E2F Activity</title>", "<p>To further elucidate the role of <italic>de2f1</italic> in Hippo pathway mutant cells, we used a <italic>PCNA</italic>-GFP reporter ##REF##12526745##[31]## to accurately measure E2F activity in clones of <italic>wts</italic> mutant cells. In a wild type eye disc, the <italic>PCNA</italic>-GFP reporter is expressed in a narrow stripe of cells prior to S phase entry in the SMW and is absent in the posterior region of the eye (##FIG##6##Figure 7A##). In contrast, <italic>wts</italic> mutant cells failed to downregulate the expression of the <italic>PCNA-</italic>GFP reporter posterior to the SMW (##FIG##6##Figure 7B##). This indicates that these cells have an abnormally high E2F activity. Importantly, the high E2F activity is due to <italic>de2f1</italic> because the <italic>PCNA</italic>-GFP reporter is no longer expressed in <italic>de2f1 wts</italic> double mutant cells posterior to the SMW (##FIG##6##Figure 7C##). The finding that <italic>wts</italic> mutant cells have a high E2F activity posterior to the SMW is unexpected since the activator dE2F1, which provides the pattern of expression of the <italic>PCNA</italic>-GFP reporter in the eye disc ##REF##12526745##[31]##, is normally downregulated in these cells (##FIG##6##Figure 7D## and ##REF##15798191##[15]##). Therefore, we examined the expression of dE2F1 in clones of <italic>wts</italic> and <italic>hpo</italic> mutant cells using a highly specific dE2F1 antibody (##FIG##6##Figure 7E##). In contrast to the wild type dE2F1 pattern, the expression of dE2F1 was highly abnormal in <italic>wts</italic> and <italic>hpo</italic> mutant cells (##FIG##6##Figure 7F##). First, the level of dE2F1 was elevated in cells that normally express dE2F1 within the MF. Second, dE2F1 was ectopically expressed in cells posterior to the MF. Thus, Hippo pathway mutant cells that inappropriately proliferate posterior to the SMW have an elevated E2F activity which is likely due to a high level of dE2F1.</p>", "<p>To determine whether the high level of dE2F1 is a specific response to inactivation of the Hippo pathway or an indirect consequence of inappropriate cell proliferation posterior to the SMW, we examined the pattern of dE2F1 expression in clones of <italic>archipelago</italic> (<italic>ago</italic>) mutant cells. <italic>ago</italic> mutant cells, like <italic>wts</italic> or <italic>hpo</italic> mutant cells, fail to exit the cell cycle and proliferate posterior to the SMW ##REF##11565033##[32]##. In contrast to the abnormally high expression of dE2F1 in Hippo pathway mutant cells, the level of dE2F1 was not elevated in clones of <italic>ago</italic> mutant cells posterior to the SMW (##FIG##6##Figure 7G##). Furthermore, expression of cyclin E, which is sufficient to drive quiescent cells posterior to the SMW into the cell cycle ##REF##18035529##[33]##,##REF##15084262##[34]##, did not result in an increase of the level of dE2F1 (##FIG##6##Figure 7H##). Finally, the level of another dE2F family member, dE2F2, is unaffected in <italic>wts</italic> mutant cells (##FIG##6##Figure 7I##). Thus, we concluded that dE2F1 is specifically upregulated following inactivation of the Hippo pathway.</p>", "<p>As a further characterization, we tested whether <italic>de2f1</italic> is transcriptionally induced in Hippo pathway mutant cells. The <italic>in vivo</italic> activity of the <italic>de2f1</italic> promoter in <italic>hpo</italic> mutant cells was examined using an enhancer trap allele, <italic>de2f1⁷²⁹</italic>. The <italic>de2f1⁷²⁹</italic> allele has been extensively used as an accurate measurement of <italic>de2f1</italic> transcription ##REF##15084262##[34]##–##REF##18313299##[36]##. This allele carries a <italic>lacZ</italic> gene inserted within the endogenous <italic>de2f1</italic> gene. Therefore, the production of β-galactosidase reflects transcriptional activity at the <italic>de2f1</italic> promoter. Clones of <italic>hpo</italic> mutant cells were induced in the eye disc and the expression of the <italic>lacZ</italic> gene from the <italic>de2f1⁷²⁹</italic> allele was compared between the <italic>hpo</italic> mutant and adjacent wild type cells. Through detection of immunofluorescence, we found a higher level of β-Gal in <italic>hpo</italic> mutant cells than in wild type cells within the MF; thus, indicating that <italic>de2f1</italic> transcription was induced (##FIG##6##Figure 7J##).</p>", "<p>Finally, we depleted SL2 tissue culture cells of Hpo and Wts by RNA interference, (RNAi), and examined what effect this had on the level of dE2Fs by western blot analysis. Consistent with the effects seen in eye imaginal discs, the dE2F1 protein level was markedly increased in cells deficient of either Wts or Hpo, while the dE2F2 protein level remained constant (##FIG##6##Figure 7K##). To further biochemically characterize the properties of Hpo and Wts depleted cells we transiently transfected these depleted cells with an E2F reporter, <italic>PCNA</italic>-luc. Depletion of RBF1, the <italic>Drosophila</italic> pRB homolog, elevated the expression of the reporter by 2.5 fold in comparison to control treated cells (##FIG##6##Figure 7L##). Noticeably, the reporter was induced approximately 2 fold in Hpo and in Wts depleted cells, which is similar to the level seen in cells deficient of RBF1, the endogenous inhibitor of dE2F1. We concluded that inactivation of the Hippo pathway increases the level of dE2F1 and, more importantly, elevates E2F activity both <italic>in vivo</italic> and <italic>in vitro</italic>.</p>" ]

[ "<title>Discussion</title>", "<p>Current models emphasize the importance of the E2F transcription factor in cell cycle control as one of the key downstream targets of the pRB tumor suppressor protein. Although E2F activity is rate limiting for S phase entry in tissue culture cells, ablation of the entire pool of <italic>Drosophila</italic> E2F is permissive for cell proliferation <italic>in vivo</italic> and only marginally affects animal development. However, it is unknown whether oncogene driven cell proliferation would also be insensitive to the loss of the entire dE2F family. This is an important conceptual point because unrestrained proliferation is a central property of a cancer cell, and this unrestrained proliferation is thought to be the result of deregulated E2F activity ##REF##10647931##[5]##.</p>", "<p>In this report, we addressed the role of the dE2F family members in cell proliferation following inactivation of the recently identified Hippo tumor suppressor pathway. In order to minimize any possible non-specific cell cycle effects seen in the presence of a <italic>de2f1</italic> mutation, we have taken advantage of the previous observation that the complete ablation of E2F function in either <italic>de2f1 de2f2</italic> double mutants or <italic>dDP</italic> single mutants is permissive for cell proliferation ##REF##15798191##[15]##,##REF##9271122##[17]##,##REF##11511545##[18]##. Our results strongly argue that the effect of the loss of E2F function on proliferation of Hippo pathway mutant cells is distinctly different in actively dividing cells and in cells undergoing unscheduled proliferation posterior to the SMW. In actively dividing cells that overexpress the pro-oncogene <italic>yki</italic>, a positive effector of the Hippo pathway, inactivation of the dE2F family has a minimal effect on cell proliferation; as Yki is capable of dramatically accelerating the rate of cell cycle progression of <italic>dDP</italic> mutant cells. Similarly, clones of cells which lack <italic>de2f1</italic>, <italic>de2f2</italic>, and <italic>wts</italic> (a negative regulator of Yki) appear to be relatively large; however, the quantification of a population doubling time in these mutant cells is technically inaccurate due to the requirement of two independent recombination events to generate triple mutant clones. Elimination of E2F function does not abolish Yki-dependent transcription, thus, we suggest that an elevated level of Yki target genes such as <italic>cyclin E</italic> and others may account for the accelerated proliferation of <italic>dDP</italic> mutant cells. Indeed, previous studies have shown that transient expression of <italic>cyclin E</italic> is sufficient to increase the rate of DNA synthesis of <italic>dDP</italic> mutant cells in the eye disc ##REF##15798191##[15]##. Thus, another conclusion that we drew from these results is that proliferation defects of <italic>dDP</italic> mutant cells are essentially rescued by Yki overexpression. This idea is consistent with the notion that Yki and the dE2Fs appear to share some common targets such as <italic>cyclin E</italic>. A caveat to this explanation is that for the exception of <italic>diap1</italic>, it is not known what putative targets are directly regulated by Yki. Secondly, since Yki fails to induce an E2F-reporter in the absence of <italic>de2f1</italic>, this suggests that Yki does not generally rescue E2F-dependent transcription in <italic>dDP</italic> mutant cells, but rather increases expression of a limited set of shared targets. Discerning how Yki overexpression accelerates the rate of proliferation in the absence of dE2F and how the interplay between Yki and dE2F occurs at common targets will be important directions in future studies.</p>", "<p>In striking contrast to the results of inactivation of the entire dE2F family in actively dividing cells, we find that E2F function is required during Yki-driven unscheduled proliferation in otherwise quiescent cells posterior to the SMW in the eye imaginal disc. Overexpression of <italic>yki</italic> or inactivation of negative regulators of the Hippo tumor suppressor pathway, such as <italic>wts</italic>, renders cells of the eye imaginal disc refractory to the cell cycle exit signals and, as a result, cells continue to proliferate inappropriately ##REF##17437995##[19]##,##REF##17318211##[21]##. These abnormal cell cycles are fully blocked when E2F function is eliminated either by a mutation in the <italic>dDP</italic> gene or by combined ablation of both <italic>de2f1</italic> and <italic>de2f2</italic> genes. This conclusion is supported by the complete absence of cells in S phase or in mitosis posterior to the SMW in mutant clones. Furthermore, loss of E2F function in clones of <italic>wts</italic> mutant cells or in clones of cells that overexpress <italic>yki</italic> significantly reduces the number of supernumerary interommatidial cells that primarily arise due to inappropriate proliferation during larval and early pupal development. Interestingly, this reduction is very similar to that seen in clones of <italic>expanded</italic> mutant cells, an upstream negative regulator of Yki, albeit the molecular mechanism is distinctly different. Unlike <italic>de2f1 de2f2 wts</italic> triple mutants, <italic>expanded</italic> mutant cells proliferate inappropriately posterior to the SMW ##REF##17258190##[37]##. However, the supernumerary interommatidial cells are largely removed during the wave of developmental pupal apoptosis, while <italic>de2f1 de2f2 wts</italic> triple mutant cells are fully protected from cell death.</p>", "<p>The results described here highlight the specific requirement for dE2F to maintain a proliferation potential in cells with high Yki activity posterior to the SMW. We emphasize that the loss of E2F control is permissive for cell proliferation in actively dividing wild type cells, as well as in actively dividing cells that overexpress <italic>yki</italic>. However, inactivation of the dE2F family fully prevents inappropriate divisions of Hippo pathway mutant cells that have failed to exit the cell cycle. While one could predict this result in the absence of <italic>de2f1</italic> alone, since cell proliferation is severely reduced in <italic>de2f1</italic> mutants, it was perhaps surprising to find that Yki-driven inappropriate proliferation posterior to the SMW is completely blocked by the total inactivation of the dE2F family. In this respect, these results are distinct from the predicted outcome of inactivation of other cell cycle regulators such as <italic>cyclin E</italic> or <italic>cyclin A</italic> on cell proliferation in Hippo pathway mutants. Mutations in these genes are likely to fully abrogate Yki induced proliferation in most, if not all, settings due to their fundamental roles in cell cycle regulation. In support of this distinction we note that the loss of the microRNA <italic>bantam</italic> has been shown to block Yki-driven proliferation in actively dividing cells of the wing disc, as well as cell proliferation during normal development ##REF##16923395##[38]##,##REF##16949821##[39]##.</p>", "<p>Why is Yki unable to drive cells into the cell cycle in the absence of E2F activity? One formal possibility is that Yki function is compromised in dE2F deficient cells posterior to the SMW. However, this seems unlikely since <italic>de2f1 wts</italic> double mutant cells, like <italic>wts</italic> single mutant cells, are fully protected from DNA damage induced apoptosis at larval stage and from developmental apoptosis in the pupal eye. Thus, inhibition of apoptosis, one of the key aspects of Yki function, remains unaltered. Consistently, Yki does induce its target genes, <italic>diap1</italic> and <italic>Expanded</italic>, in <italic>dDP</italic> mutant cells. Hence, Yki activity does not appear to be generally affected by the loss of E2F function. We also note that Yki-dependent induction of cyclin E (this work) and cyclin B (our unpublished observations) still occurs in <italic>de2f1</italic> deficient cells posterior to the SMW, yet these cells fail to proliferate posterior to the SMW. Thus, it remains a likely possibility that high cyclin E activity is capable of driving proliferation in actively dividing cells in the absence of dE2F, but not in cells during the cell cycle exit where cyclin E appears to require an assist from dE2F1 to sustain unscheduled cell proliferation. Although we cannot formally exclude the possibility that expression of some Yki target genes is deregulated in dE2F deficient cells, these results suggest that in the absence of dE2F, the Yki transcriptional program alone is insufficient to drive cell proliferation in otherwise quiescent cells posterior to the SMW. We emphasize that overexpression of dE2F1 is not sufficient to sustain proliferation in cells posterior to the SMW (for example see: ##REF##8670871##[35]##,##REF##17419999##[40]##) and that the phenotype of Hippo pathway mutant cells is likely to be a result of a cumulative effect of deregulation of a panel of Yki target genes. This is consistent with several studies that have shown that upon the cell cycle exit, cells become highly resistant to proliferative signals. For example, co-expression of dE2F1 and cyclin E is needed to bypass the cell cycle exit ##REF##17419999##[40]##. Similarly, combined ablation of two negative regulators of the cell cycle, RBF1 and the cdk2 inhibitor Dacapo, is required to prevent the exit from the cell cycle in the larval eye ##REF##15809036##[41]##. Thus, our results highlight the need for dE2F during inappropriate proliferation at the specific point when cells attempt to exit the cell cycle.</p>", "<p>The Hippo pathway controls epithelial tissue growth by regulating the expression of genes that can promote cell proliferation and genes that can inhibit apoptosis. In humans, loss of expression of Lats1/2 (Wts homolog) ##REF##15746036##[42]## and mutations in <italic>WW45/Sav</italic>\n##REF##12202036##[43]## and <italic>Mob</italic> (homolog of <italic>mats</italic>) ##REF##15766530##[44]## have been found in several tumor cell lines, while YAP expression is frequently elevated in cancers ##REF##17974916##[45]##,##REF##17889654##[46]##. Accordingly, mouse embryos lacking <italic>WW45/Sav</italic> display hyperplasia due to defects in cell cycle exit and terminal differentiation of epithelial progenitor cells ##REF##18369314##[47]##; while <italic>Lats1</italic>\n^−/− knockout animals develop soft-tissue sarcomas and ovarian stromal cell tumors ##REF##9988269##[48]##. Additionally, in a transgenic mouse model, YAP activation in the liver induces hyperplasia followed by tumor formation ##REF##17889654##[46]##,##REF##17980593##[49]##. Thus, the Hippo pathway represents a frequent mutational target and the outcome of its deregulation is tumorigenesis in both mice and humans. Although the status of the pRB pathway has not been determined in these tumors, it is generally thought that inactivation of the pRB pathway is an obligatory event in most, if not all, types of tumors ##REF##10647931##[5]##. In this respect, it is particularly intriguing that the ablation of Hippo function leads to an increase in dE2F1 level and elevation of E2F activity. Since the Hippo pathway is highly conserved between flies and mammals, it would be interesting to determine whether expression of mammalian <italic>E2f</italic>s is also induced following inactivation of the Hippo pathway. In support of this possibility we note that ectopic expression of YAP in transgenic mice increases the level of a well known E2F target gene, <italic>PCNA</italic>\n##REF##17980593##[49]##. A comparison of <italic>de2f2 wts</italic> and <italic>de2f1 de2f2 wts</italic> mutant clones revealed that <italic>de2f1</italic> is more important during inappropriate proliferation than <italic>de2f2</italic>. Importantly, the dE2F1 increase is not a coincidental result of an accelerated rate of proliferation in Hippo pathway mutant cells, since the level of dE2F1 is normal in clones of <italic>ago</italic> mutant cells, which, like <italic>wts</italic> mutants, proliferate posterior to the SMW. Although dE2F1 induction appears to be a result of a transcriptional response in <italic>hpo</italic> mutant cells and following overexpression of Yki (this work and ##REF##18313299##[36]##), whether the <italic>de2f1</italic> promoter is directly regulated by Yki is currently not known. Further experiments will be necessary to decipher the exact mechanism by which Yki exerts its effect on <italic>de2f1</italic> expression.</p>", "<p>Finally, the data described here have a potential implication in cancer research. Inactivation of total E2F activity using a dominant negative form of E2F has been shown to impair re-entry into the cell cycle from quiescence in immortalized murine fibroblasts ##REF##12150825##[9]##. We find that a complete genetic ablation of the dE2F family fully blocks proliferation of Hippo pathway mutant cells posterior to the SMW. However, Hippo pathway mutant cells do not undergo a transient quiescence state during inappropriate proliferation since the mutant cells continuously express proliferation markers, such as phosH3, cyclin B, cyclin A, incorporate BrdU, and have a high level of an E2F reporter posterior to the SMW (this work and ##REF##12941273##[27]##,##REF##12941274##[28]##,##REF##12202036##[43]##). Thus, dE2F is required to sustain inappropriate cell proliferation specifically at the point when cells normally exit the cell cycle and enter quiescence. One implication of this result is that, at least, in the case of the inactivation of the Hippo pathway, the use of pharmacological E2F inhibitors might be beneficial in tumors in which cell cycle exit cues are induced even though these tumors do not necessarily respond to these signals and pass through temporary states of quiescence.</p>" ]

[]

[ "<p>Conceived and designed the experiments: BNN MVF. Performed the experiments: BNN. Analyzed the data: BNN MVF. Wrote the paper: BNN MVF.</p>", "<p>The Hippo pathway negatively regulates the cell number in epithelial tissue. Upon its inactivation, an excess of cells is produced. These additional cells are generated from an increased rate of cell division, followed by inappropriate proliferation of cells that have failed to exit the cell cycle. We analyzed the consequence of inactivation of the entire E2F family of transcription factors in these two settings. In <italic>Drosophila</italic>, there is a single activator, dE2F1, and a single repressor, dE2F2, which act antagonistically to each other during development. While the loss of the activator dE2F1 results in a severe impairment in cell proliferation, this defect is rescued by the simultaneous loss of the repressor dE2F2, as cell proliferation occurs relatively normally in the absence of both dE2F proteins. We found that the combined inactivation of dE2F1 and dE2F2 had no significant effect on the increased rate of cell division of Hippo pathway mutant cells. In striking contrast, inappropriate proliferation of cells that failed to exit the cell cycle was efficiently blocked. Furthermore, our data suggest that such inappropriate proliferation was primarily dependent on the activator, <italic>de2f1</italic>, as loss of <italic>de2f2</italic> was inconsequential. Consistently, Hippo pathway mutant cells had elevated E2F activity and induced dE2F1 expression at a point when wild-type cells normally exit the cell cycle. Thus, we uncovered a critical requirement for the dE2F family during inappropriate proliferation of Hippo pathway mutant cells.</p>", "<title>Author Summary</title>", "<p>The E2F transcription factor family is considered to be the best-characterized downstream target of the retinoblastoma protein (pRB). The pRB pathway is functionally inactivated in most tumor cells, and it is thought that unrestrained activity of E2F drives inappropriate proliferation in tumors. We utilized the relative simplicity of the <italic>Drosophila</italic> model to determine the role of the dE2F family in proliferation of cells following inactivation of the recently identified Hippo tumor suppressor pathway. We found that Hippo pathway mutant cells require the dE2F family to delay the cell cycle exit and to proliferate inappropriately when wild-type cells enter quiescence. This is significant since the loss of the entire dE2F family exerts almost no effect on the ability of Hippo pathway mutations to accelerate proliferation of actively dividing cells. Thus, the importance of the dE2F family in cells with an inactivated tumor suppressor pathway varies in different contexts. This discovery may have implications in designing anti-cancer therapies that inhibit E2F activity.</p>" ]

[]

[ "<p>We are grateful to B. Edgar, R. Fehon, I. Hariharan, K. Harvey, K. Irvine, L. Johnston, A. Katzen, Z. Lai, K. Moberg, T. Orr-Weaver, G. Struhl, N. Tapon, J. Treisman, the Developmental Studies Hybridoma Bank (University of Iowa) and the Bloomington Stock Center for fly stocks and antibodies. We thank V. Rasheva and A. Ambrus for technical assistance. We thank N. Dyson, A. Katzen, G. Ramsey and M. Truscott for critical discussions.</p>" ]

[ "<fig id=\"pgen-1000205-g001\" position=\"float\"><object-id pub-id-type=\"doi\">10.1371/journal.pgen.1000205.g001</object-id><label>Figure 1</label><caption><title>Inactivation of the dE2F family does not block Yki-induced proliferation in actively dividing cells.</title><p>Clones of cells of four different genotypes marked with GFP (green) were induced simultaneously with the MARCM technique and allowed to grow for the same period of time. Cells were visualized by staining with DAPI (blue). Population doubling time (DT) is shown for each genotype. Data were collected from 28 clones for wild type, 28 clones for <italic>tub</italic>&gt;<italic>yki</italic>, 66 clones for <italic>DP^−/−</italic>, and 33 clones for <italic>tub</italic>&gt;<italic>yki</italic> in <italic>DP^−/−</italic>. Average clone areas are 9,206±613 pixels for wild type; 18,491±1,780 pixels for <italic>tub</italic>&gt;<italic>yki</italic>; 2,814±279 pixels for <italic>DP^−/−</italic> and 12,909±1,263 pixels for <italic>tub</italic>&gt;<italic>yki</italic> in <italic>DP^−/−</italic>. (A) Control clones induced with a wild type <italic>FRT42D</italic> chromosome. (B) Clones of wild type cells that overexpress <italic>yki</italic> contain more cells than control in (A). (C) Clones of <italic>dDP</italic> mutant cells. (D) Clones of <italic>dDP</italic> mutant cells that overexpress <italic>yki</italic> contain more cells than clones of <italic>dDP</italic> mutant cells (C) or control clones (A).</p></caption></fig>", "<fig id=\"pgen-1000205-g002\" position=\"float\"><object-id pub-id-type=\"doi\">10.1371/journal.pgen.1000205.g002</object-id><label>Figure 2</label><caption><title>Yki-driven proliferation of cells posterior to the SMW is blocked in the absence of the dE2F family.</title><p>(A) The pattern of S phases in the wild type eye discs as revealed by BrdU labeling. Position of the morphogenetic furrow (MF) is shown by arrowhead. Posterior is to the right. Wild type cells asynchronously proliferate anterior to the MF, arrest in G1 in the MF and undergo a synchronous S phase in the second mitotic wave (SMW) posterior to the MF. (B–C) Clones of wild type (B) and <italic>dDP</italic> mutant (C) cells overexpressing <italic>yki</italic> were generated with the MARCM system and marked with GFP (green). Clones in (B) were generated with the <italic>ey</italic>-FLP while clones in (C) were generated with the <italic>hs</italic>-FLP. (B) Posterior to the SMW, wild type cells that overexpress <italic>yki</italic> fail to exit the cell cycle and proliferate inappropriately as evident by the appearance of BrdU positive cells. (C) In contrast, <italic>yki</italic> is unable to induce inappropriate proliferation of <italic>dDP</italic> mutant cells posterior to the SMW. Note, that <italic>dDP</italic> mutant cells that overexpress <italic>yki</italic> show a normal pattern of BrdU incorporation in the SMW but do not incorporate BrdU posterior to the SMW. (D–M) Clones of mutant cells of different genotypes were generated with <italic>ey</italic>-FLP and the mutant tissue was distinguished by the lack of GFP (green). (D–F) Mosaic larval eye discs were labeled with BrdU (red) to detect the S phases (D and E) or stained with anti-phosH3 (magenta) to visualize mitoses (F). (D) <italic>wts</italic> mutant cells fail to exit the cell cycle posterior to the SMW and undergo inappropriate proliferation, which is evident by the persistence of BrdU incorporation (pointed by arrows). (E–F) In contrast, inappropriate proliferation posterior to the SMW is strongly reduced in <italic>de2f1 de2f2 wts</italic> triple mutant cells as judged by the absence of cells in S phase (red in E) or in mitosis (magenta in F) (pointed by arrows). (G–H) BrdU incorporation (red) in 12 hr pupal eye discs. (G) <italic>wts</italic> mutant cells continue unscheduled proliferation during early pupal development while wild type cells remain fully quiescent as revealed by BrdU labeling. (H) Inappropriate BrdU incorporation is absent in clones of <italic>de2f2 de2f1 wts</italic> triple mutant cells (a clone is pointed by arrow). (I–M) Pupal retina at 48 hr APF stained with anti-Discs large protein (Dlg) (red) to visualize cell outlines. (I) Wild type retina contains a single layer of interommatidial cells between ommatidial clusters. (J) Inappropriate proliferation of <italic>wts</italic> mutant cells posterior to the SMW and resistance of these cells to the developmental apoptosis during the pupal stage gives rise to the dramatic excess of interommatidial cells (pointed by arrow). (K) In contrast, the number of interommatidial cells is significantly reduced in <italic>de2f1 de2f2 wts</italic> triple mutant tissue (indicated by arrows). (L–M) The MARCM technique was used to overexpress <italic>yki</italic> in wild type (L) or in <italic>dDP</italic> mutant cells (M). A <italic>dDP</italic> mutation dramatically reduces supernumerary interommatidial cells which arise when <italic>yki</italic> is expressed. (N) Quantification of the number of interommatidial cells in pupal retina shown in (J–M). Data were collected from 23 ommatidia clusters for wild type, 11 ommatidia clusters for <italic>wts</italic>, 24 ommatidia clusters for <italic>de2f1 de2f2 wts</italic>, 12 ommatidia clusters for <italic>tub</italic>&gt;<italic>yki</italic> and 17 ommatidia clusters for <italic>tub</italic>&gt;<italic>yki</italic> in <italic>DP^−/−</italic>. The following abbreviations were used: <italic>e1 e2 wts</italic> corresponds to <italic>de2f1 de2f2 wts</italic>; <italic>yki</italic> corresponds to <italic>tub</italic>&gt;<italic>yki</italic> and <italic>yki</italic> in <italic>DP</italic> corresponds to <italic>tub</italic>&gt;<italic>yki</italic> in <italic>DP^−/−</italic>. Error bars represent standard deviations. Note that in comparison to the wild type tissue there is a small excess of interommatidial cells in <italic>de2f1 de2f2 wts</italic> triple mutant tissue and in <italic>dDP</italic> mutant tissue that overexpresses <italic>yki</italic>. This is likely due to the failure to execute normal pupal developmental apoptosis in these cells.</p></caption></fig>", "<fig id=\"pgen-1000205-g003\" position=\"float\"><object-id pub-id-type=\"doi\">10.1371/journal.pgen.1000205.g003</object-id><label>Figure 3</label><caption><title>Loss of <italic>de2f2</italic> does not affect the <italic>wts</italic> mutant phenotype in the eye.</title><p>Clones of mutant cells were generated with <italic>ey</italic>-FLP. Position of the morphogenetic furrow (MF) is marked by a white arrowhead in B and D. Posterior is to the right. (A) <italic>de2f2</italic> and <italic>wts</italic> are on two separate chromosomal arms. Therefore following expression of <italic>ey-</italic>FLP, clones of cells of four different genotypes are generated (wild type, <italic>de2f2</italic> mutant, <italic>wts</italic> mutant and <italic>de2f2 wts</italic> double mutant). <italic>de2f2 wts</italic> double mutant tissue is marked by the lack of GFP (green), is labeled <italic>e2f2^−/− wts^−/−</italic> in (A) and is outlined by a white line in (A–C). <italic>wts</italic> single mutant tissue is distinguished by a reduced intensity of GFP (green) and an increased spacing between ommatidial clusters (marked by ELAV). An example of <italic>wts</italic> mutant tissue is labeled <italic>wts^−/−</italic> in (A) and is denoted by yellow outline in (A–C). Wild type tissue is distinguished by the strongest level of GFP (green). An example of the wild type tissue is found between the yellow and white line. (B–D) Mosaic larval eye discs were labeled with BrdU (red) to detect cells in S phase (B,D) and ELAV (blue) to identify position of ommatidial clusters (B–D). Posterior to the MF, wild type cells undergo a single round of S phases in the second mitotic wave (SMW) (denoted by white arrow in B). In contrast, inappropriate BrdU (red) incorporation posterior to the SMW was detected in both <italic>wts</italic> and <italic>de2f2 wts</italic> mutant cells (B,D). (C) As a result of this inappropriate proliferation, spacing between ommatidial clusters is increased in both <italic>wts</italic> and <italic>de2f2 wts</italic> mutant tissue in comparison to wild type tissue. A merged image is shown in D.</p></caption></fig>", "<fig id=\"pgen-1000205-g004\" position=\"float\"><object-id pub-id-type=\"doi\">10.1371/journal.pgen.1000205.g004</object-id><label>Figure 4</label><caption><title>Loss of <italic>de2f1</italic> does not affect resistance to DNA damage and pupal developmental apoptosis in <italic>wts</italic> mutants.</title><p>In all panels, clones were generated with the <italic>ey</italic>-FLP/FRT technique and mutant tissue is distinguished by the absence of GFP (green). (A–B) The pupal eye discs at 30 hr APF containing clones of <italic>wts</italic> mutant (A) and <italic>de2f1 wts</italic> double mutant (B) cells were stained with anti-Cleaved Caspase3 (C3) antibody (red) to detect apoptotic cells. In the pupal eye discs, developmentally regulated apoptosis is abundant in wild type cells (green) but is largely absent in <italic>wts</italic> mutant tissue (lack of green) and in <italic>de2f1 wts</italic> double mutant tissue indicating that <italic>de2f1 wts</italic> double mutant cells, like <italic>wts</italic> mutant cells, are protected from the cell death. (C) DNA damage induced apoptosis following irradiation was detected with anti-C3 antibody (red). There is an extensive apoptosis in wild type tissue. In contrast, <italic>de2f1 wts</italic> double mutant cells are protected from apoptosis after DNA damage. An example of <italic>de2f1 wts</italic> double mutant tissue is pointed by arrow. The morphogenetic furrow is marked by the arrowhead.</p></caption></fig>", "<fig id=\"pgen-1000205-g005\" position=\"float\"><object-id pub-id-type=\"doi\">10.1371/journal.pgen.1000205.g005</object-id><label>Figure 5</label><caption><title>\n<italic>dDP</italic> mutation does not block <italic>yki-</italic>dependent induction of its target genes, dIAP1 and Expanded.</title><p>The MARCM system was used to drive <italic>yki</italic> overexpression in wild type (A, C) or in <italic>dDP</italic> mutant cells (B, D) of the larval eye disc. Cells that express <italic>yki</italic> are marked with GFP (green). Merge images are on the right. <italic>yki</italic> overexpression induces Expanded (A) and dIAP1 (C) expression (pointed by arrows) in wild type cells. Inactivation of the dE2F family in <italic>dDP</italic> mutant cells does not significantly affect <italic>yki</italic>-dependent induction of Expanded (B) and dIAP1 (D) (pointed by arrows). Images in (B and D) show the same clone that was double stained with dIAP1 and Expanded, while images in (A and C) represent two different clones stained singularly with Expanded (A) and dIAP1 (C).</p></caption></fig>", "<fig id=\"pgen-1000205-g006\" position=\"float\"><object-id pub-id-type=\"doi\">10.1371/journal.pgen.1000205.g006</object-id><label>Figure 6</label><caption><title>Loss of <italic>de2f1</italic> does not block induction of cyclin E in <italic>wts</italic> mutant cells.</title><p>Clones of mutant cells were generated with <italic>ey</italic>-FLP and distinguished by the lack of GFP (green). (A) In wild type eye imaginal discs, cyclin E (magenta) expression is elevated within and immediately posterior to the morphogenetic furrow (MF). In <italic>wts</italic> mutant cells (B) and in <italic>de2f1 wts</italic> double mutant cells (C) cyclin E is expressed further posterior. Position of MF is shown by arrowhead. Posterior is to the right.</p></caption></fig>", "<fig id=\"pgen-1000205-g007\" position=\"float\"><object-id pub-id-type=\"doi\">10.1371/journal.pgen.1000205.g007</object-id><label>Figure 7</label><caption><title>Elevated E2F activity in cells with inactivated Hippo pathway.</title><p>Clones of mutant cells were induced with <italic>ey</italic>-FLP. Position of the morphogenetic furrow (MF) is shown by arrowhead. Posterior is to the right. (A–C) Expression of the E2F reporter, <italic>PCNA</italic>-GFP, (red) in the wild type eye disc (A), or in the eye discs containing clones of <italic>wts</italic> mutant tissue (B) and <italic>de2f1 wts</italic> double mutant tissue (C). (A) In wild type disc, the E2F reporter is expressed in a narrow stripe (red) immediately posterior to the MF, preceding S phase entry into the second mitotic wave (SMW). (B) In <italic>wts</italic> mutant cells which are marked by the absence of β-Gal (green), the E2F reporter is inappropriately expressed in the posterior region of the eye disc. Mutant tissue is outlined. (C) Inappropriate expression of the E2F reporter in the posterior region of the eye disc is absent in <italic>de2f1 wts</italic> double mutant cells. Note, that clones of <italic>wts de2f1</italic> double mutant cells were marked with β-Gal (green) produced from the <italic>de2f1⁷²⁹</italic> mutant allele. <italic>de2f1 wts</italic> double mutant tissue is outlined. (D) Endogenous dE2F1 (magenta) is expressed within the MF in a wild type disc as revealed by anti-dE2F1 antibody. (E–G, I) Clones of mutant cells were induced with <italic>ey</italic>-FLP and mutant tissue is identified by the lack of GFP (green). (E) The anti-dE2F1 antibody is highly specific as the staining is absent in <italic>de2f1</italic> mutant tissue (lack of green in E and pointed by the arrow). (F) <italic>wts</italic> and <italic>hpo</italic> mutant cells have an increased level of dE2F1 within the MF and inappropriately express dE2F1 posterior to the MF. Examples are pointed by the arrows. Position of mutant tissue is outlined. (G) Expression of endogenous dE2F1 protein (magenta) is unaffected in <italic>ago</italic> mutant cells in larval imaginal eye discs. (H) cyclin E was expressed ectopically in wild type mitotic clones using the MARCM system. Ectopic expression of cyclin E fails to elevate level of dE2F1 protein (magenta) posterior to the MF. Cells that express cyclin E are marked with GFP (green) and are outlined. (I) Endogenous dE2F2 protein (red) is expressed ubiquitously throughout the eye disc. Level of dE2F2 protein remains the same in both <italic>wts</italic> mutant and wild type tissue. (J) <italic>de2f1</italic> is transcriptionally induced in <italic>hpo</italic> mutant cells as revealed by the <italic>de2f1</italic> enhancer trap allele, <italic>de2f1⁷²⁹</italic>. <italic>de2f1⁷²⁹</italic> contains the <italic>lacZ</italic> insertion into the endogenous <italic>de2f1</italic> gene. The <italic>lacZ</italic> expression reflects transcription from the <italic>de2f1</italic> promoter ##REF##15084262##[34]##,##REF##8670871##[35]##. Staining with anti-β-Gal antibody (magenta) was used to reveal expression of the <italic>lacZ</italic> gene in <italic>de2f1⁷²⁹</italic>. (K) SL2 cells were treated with nonspecific (NS), dE2F1 (E1), Warts (Wts) and Hippo (Hpo) dsRNA to deplete the corresponding proteins by RNAi. Cell extracts were analyzed by Western blot using antibody specific for Wts, dE2F1 and dE2F2. Depletion of Wts and Hpo shows an increase in the level of dE2F1 protein. In contrast, the dE2F2 protein level is not affected. The same blots were re-probed with anti-Tubulin antibody to control for equal loading. (L) Endogenous E2F activity is elevated in Hpo or Wts depleted SL2 cells. SL2 cells were incubated with non-specific (NS), RBF1, Hpo, and Wts dsRNAs for 4 days to deplete the corresponding proteins. On day 4, the E2F reporter (<italic>PCNA</italic>-luc) was transfected into the depleted cells and the luciferase activity was measured 2 days later to determine the level of the endogenous E2F activity in these cells. The pIE-LacZ plasmid was co-transfected to normalize for transfection efficiency. Results depict the mean of three experiments. Unpaired Student's <italic>t</italic>-Test assuming equal variance concluded that the increase of <italic>PCNA-</italic>luc reporter activity in RBF1, Hpo and Wts depleted cells was statistically significant from the NS control. RBF1 and Hpo depleted cells had a p-Value &lt;0.001. Wts depleted cells had a p-Value &lt;0.03.</p></caption></fig>" ]

[]

[]

[]

[]

[]

[]

[]

[ "<fn-group><fn fn-type=\"COI-statement\"><p>The authors have declared that no competing interests exist.</p></fn><fn fn-type=\"financial-disclosure\"><p>This work was supported by grant GM079774 from NIH to MVF.</p></fn></fn-group>" ]

[ "<graphic xlink:href=\"pgen.1000205.g001\"/>", "<graphic xlink:href=\"pgen.1000205.g002\"/>", "<graphic xlink:href=\"pgen.1000205.g003\"/>", "<graphic xlink:href=\"pgen.1000205.g004\"/>", "<graphic xlink:href=\"pgen.1000205.g005\"/>", "<graphic xlink:href=\"pgen.1000205.g006\"/>", "<graphic xlink:href=\"pgen.1000205.g007\"/>" ]

[]

{ "acronym": [], "definition": [] }

[ "<title>Introduction</title>", "<p>The genetic and developmental basis of the formation of organismal shape and form is a long-standing question in biology. The analysis of mutations has been essential in identifying the genes and regulatory networks underlying development. However, while the genetic basis of embryonic development has been extensively studied by systematic mutagenesis screens, we know little of the genes involved in the development of adult morphology. Yet, it is the heritable variation in adult form that natural selection primarily acts on during evolution. In order to understand the basis of variation, we need to know more about the genetic control of the development of adult form: which genes are involved, what are their function, and when are they required in development ##REF##10472187##[1]##,##UREF##0##[2]##. To identify genes important for development of adult structures, we initiated a large-scale mutagenesis screen in zebrafish and scored for mutants affected in the shape and pattern of adult structures. We isolated only adult viable mutants, therefore we selected for genes that have an increased probability to be involved in morphological change during evolution. Identification of zebrafish genes homologous to human genes associated with disease that arise during postnatal development into adulthood is also likely in this screen.</p>", "<p>We focused on mutants that exhibit defects in the dermal skeleton of the adult zebrafish. The dermal skeleton encompasses the external form of the adult fish. The most prominent dermal skeletal elements are the dermocranium of the skull and lateral bones of the opercular series, the scales, and the fin rays (or lepidotrichia). Additionally, the teeth and gill rakers (bones that support the gills in teleosts) are elements of the dermal skeleton ##REF##12803422##[3]##,##REF##9541261##[4]##. Unlike the ossification process that occurs during endochondrial bone development in which organic matrix is deposited by osteoblasts over a chondrogenic scaffold, dermal skeletal elements originate as direct mineralization of a collagenous matrix deposited by dermal fibroblasts. This process occurs in close association with the epidermis. The initiation and patterning of dermal elements are thought to be similar to epidermal appendages (e.g. hairs and feathers) and is controlled by reciprocal signaling between an epithelium and mesenchyme (see ##REF##14023393##[5]##,##REF##15071597##[6]##). Importantly, in zebrafish, as in most teleosts, the majority of dermal skeletal elements are not formed during larval development, rather through juvenile metamorphosis and development of the adult pattern. Those that begin to form in late larval development such as the teeth and gill rakers, do not fully attain their shape and pattern until juvenile metamorphosis.</p>", "<p>Variations in the shape of dermal skeletal elements of the fins, scales, cranium, and teeth play a significant role in adaptations of fish populations to new environments (e.g. dermal plate development and stickleback radiation ##UREF##1##[7]##). Additionally, integumentary appendages, such as hair and feathers, have been essential and defining traits of vertebrate classes. Early vertebrates, the conodonts, ostracoderms and placoderms, possessed a pronounced dermal skeleton, often in the absence of an ossified axial skeleton ##REF##12430166##[8]##. Through vertebrate evolution from fish to tetrapods, dermal structures such as lateral bones of the opercular series, scales, dermal plates and fin rays were either reduced or lost. This evolutionary transition was paralleled with the elaboration of the cartilaginous endoskeleton of the limbs and the evolution of specialized keratinized appendages of the integument such as epidermal scales, feathers and hairs. In contrast, the diversity of form in extant bony fishes involves modification in size, shape and number of the scales/dermal plates, fin rays, cranial dermal bones and teeth.</p>", "<p>Here, we describe a collection of mutants that have shared defects in the formation of the dermal skeletal elements of the skull, fins, scales and teeth of the adult zebrafish. The mutations disrupt the genes <italic>ectodysplasin</italic> (<italic>eda)</italic> and <italic>edar</italic> encoding the <italic>eda</italic> receptor. In mammals the EDA signaling pathway is involved in hair and teeth formation ##REF##12787560##[9]## and mutations affecting this pathway cause the human hereditary disease hypohidrotic ectodermal dysplasia (HED). Loss of Eda signaling in the zebrafish causes a spectrum of phenotypes corresponding to those described for HED in humans, and therefore the zebrafish mutants may serve as a genetic model of this disease. We describe the requirement of Eda signaling in the zebrafish epidermis for the formation of a structure resembling an epidermal placode seen in the early development of other vertebrate integumentary appendages. The mutations also result in defects of skeletal elements unique to fish and suggest an ancestral role of Eda signaling in the formation and patterning of the dermal skeleton. Lastly, whereas loss of function of Eda signaling causes a severe phenotype, the expressivity of dominant alleles is sensitive to background modifiers that buffer the phenotypic consequences of loss of Eda signaling. Additionally, we find that the response to reduction of Eda signaling is dose sensitive and organ specific. We suggest that such alleles may provide a basis for morphological variation in evolution.</p>" ]

[ "<title>Statistical Methods</title>", "<p>Analysis of cranial measurements were performed using Hotelling's T squared test for two dependent variables. For scale counts and size dimensions, a <italic>t</italic>-statistic for differential means was used to assess significance. Calculations and probability assessment were calculated using Biosoft 200 software (<ext-link ext-link-type=\"uri\" xlink:href=\"http://www.biosoft.com\">www.biosoft.com</ext-link>) and Excel statistical package.</p>" ]

[ "<title>Results</title>", "<title>\n<italic>finless</italic> and <italic>Nackt</italic> Mutants Exhibit Defects in the Development of the Dermal Skeleton</title>", "<p>In a mutagenesis screen for mutations affecting adult zebrafish structures, we identified three mutants that showed nearly identical defects in the formation of scales, lepidotrichia, and shape of the skull of homozygous fish. These mutants fell into two complementation groups. The first is allelic to the <italic>finless (fls</italic>\n^te370f\n<italic>)</italic> mutant that was previously isolated in the background of the Tübingen wild type stock (Tü) on the basis of the loss of fins in adults ##REF##24173565##[10]##. We isolated two new alleles of <italic>fls</italic> in the screen and found another in the background of the <italic>TLF</italic> wild type stock. The majority of the <italic>fls</italic> alleles isolated are recessive and have a strong phenotype (see below). However, the <italic>fls</italic>\n^dt3Tpl allele is dominant with a partial scale loss phenotype in heterozygotes (##FIG##0##Figure 1G, J##). One further <italic>fls</italic> allele was isolated in a screen for mutations that failed to complement the <italic>fls</italic>\n^te370f mutant (##FIG##0##Figure 1M##). We named this allele <italic>fang</italic> (<italic>fls</italic>\n^tfng) after its unique dental phenotype in homozygotes of having only one tooth on the fifth ceratobranchial (##FIG##0##Figure 1N##). The <italic>fls</italic>\n^tfng allele shows no effect on fin development and has a slight increase in the number of scales than the other <italic>fls</italic> alleles isolated.</p>", "<p>The second complementation group was comprised of a single gene, which we called <italic>Nackt</italic> (<italic>Nkt</italic>). This allele is dominant causing a slight defect in the patterning and shape of scales as heterozygotes (##FIG##1##Figure 2D##). The homozygous phenotype is more severe than that of strong <italic>fls</italic> alleles (##FIG##1##Figure 2A##).</p>", "<p>Phenotypic defects of <italic>fls</italic> and <italic>Nkt</italic> mutants become apparent in juvenile fish; as larvae, homozygous <italic>fls</italic> and <italic>Nkt</italic> mutants are visibly unaffected. Homozygous adults are viable, and of normal size. With the exception of the <italic>fls</italic>\n^tfng allele, the lepidotrichia that form during juvenile metamorphosis are defective, leading to fin loss in the adult (##FIG##0##Figure 1D##, ##FIG##1##2A##). The dermal bones of the pectoral girdle are present and patterned appropriately in both the <italic>fl</italic>s and <italic>Nkt</italic> mutants. By close examination of the visceral skeleton we found that neither the pharyngeal teeth, nor the bony substrates of the gills, the gill rakers, are formed (##FIG##0##Figure 1E and F##; ##FIG##1##Figure 2B, C##). In addition, scales are largely absent with infrequent formation of inappropriately shaped scales near the dorsal, anal and pectoral fins (##FIG##0##Figure 1D##, ##FIG##1##2A##). <italic>Nkt</italic> homozygous fish exhibit more severe defects in the formation of the dermal skeleton than <italic>fls</italic> alleles in the extent of lepidotrichial growth and number of residual scales formed (compare ##FIG##0##Figure 1D## and ##FIG##1##Figure 2A##). The skull of mutants has a normal appearance with all the bones being present, although the size, shape and relative proportion of the various bones differ compared to wild type individuals (##SUPPL##0##Figure S1##); no change in cranial shape was apparent in larvae.</p>", "<p>\n<italic>Nkt</italic> heterozygous fish exhibit a dominant phenotype as they lack several scales on the flank and those present at the flank are elongated dorso-ventrally. The number of teeth and gill rakers is reduced, however lepidotrichia formation and growth of the fins are not affected (##FIG##1##Figure 2D–F##). The skulls of <italic>Nkt</italic> heterozygotes do not show increased size, but retain altered shape and proportion as seen in homozygotes (##SUPPL##0##Figure S1##).</p>", "<title>The Topless Allele of <italic>fls</italic> Uncovers Background Specific Modulation of <italic>fls</italic> Expressivity</title>", "<p>We isolated a dominant allele of <italic>fls</italic> that exhibits a distinct phenotype in heterozygotes that we named <italic>Topless</italic> (<italic>fls</italic>\n^dt3Tpl). Heterozygous <italic>fls</italic>\n^dt3Tpl have a reduction in the number of scales, teeth and gill rakers, but show little to no effect on lepidotrichia development (##FIG##0##Figure 1G–I##). Mutant <italic>fls</italic>\n^dt3Tpl fish exhibit the strong <italic>fls</italic> phenotype when homozygous or heterozygous with other <italic>fls</italic> alleles. Similar to <italic>Nkt</italic>, <italic>fls</italic>\n^dt3Tpl exhibits a dominant effect on skull shape as well (##SUPPL##0##Figure S1##).</p>", "<p>The expressivity of the dominant <italic>fls</italic>\n^dt3Tpl phenotype depends on the genetic background. Fish heterozygous for <italic>fls</italic>\n^dt3Tpl exhibited either a “strong” or a “weak” phenotype in the Tü background (##FIG##0##Figure 1G and J##, respectively). The “strong” phenotype shows loss of scales regionally in the midflank, loss of medial pharyngeal teeth along the fifth ceratobranchial and loss of posterior gill rakers of the anterior arches (##FIG##0##Figure 1G–I##). In contrast, the “weak” phenotype displays only subtle variation in scale patterning and no effect on the teeth or gill rakers could be detected (##FIG##0##Figure 1J–L##). The segregation pattern of the two phenotypic classes of <italic>fls</italic>\n^dt3Tpl suggests the presence of separate, unlinked, modifier loci in the Tü background affecting the number of scales (##TAB##0##Table 1## and data not shown). Additionally, we found that the <italic>fls</italic>\n^dt3Tpl “strong” phenotype was partially suppressed when crossed with the polymorphic WIK mapping strain indicating the presence of dominant modifier(s) in the WIK line (##TAB##0##Table 1##). The resulting heterozygous progeny showed reduced scale-loss compared to <italic>fls</italic>\n^dt3Tpl heterozygotes in a Tü background, but had similar defects in scale shape (##TAB##0##Table 1##). Therefore, dominant modifier loci are present in the WIK strain that buffer the expressivity of the <italic>fls</italic>\n^dt3Tpl dominant phenotype. None of the other <italic>fls</italic> alleles showed any dominance in the Tü, TLF, or WIK strains.</p>", "<title>Mutations in <italic>fls</italic> Disrupt the Ectodysplasin Receptor in Zebrafish</title>", "<p>We identified the affected loci of the <italic>fls</italic> mutants by positional cloning. The <italic>fls</italic>\n^te370f mutation was linked to SSLP markers on linkage group 9 (LG9). Due to similarity of the <italic>fls</italic> phenotype to ectodermal dysplasia phenotypes in mammals, we mapped several genes of the ectodysplasin pathway to the zebrafish radiation hybrid map to see if any of these genes were linked to <italic>fls</italic>. The <italic>edar</italic> gene is located on LG9 within the determined linkage interval for <italic>fls</italic> (see Methods). We cloned the full-length wild type cDNA of <italic>edar</italic> and found several polymorphisms in the Tü <italic>edar</italic> cDNA when compared to the WIK mapping strain; these polymorphisms were tightly linked with the <italic>fls</italic> mutation and did not show recombination in 238 meioses (##SUPPL##1##Figure S2##).</p>", "<p>The <italic>edar</italic> gene encodes a transmembrane protein with similarity to tumor necrosis factor receptor (TNFR). The Edar protein contains a conserved TNFR extracellular ligand binding domain and a cytoplasmic terminal death domain essential for protein interactions with signaling adaptor complexes. The <italic>fls</italic>\n^te370f mutation is an A to T transversion at a splice acceptor site, resulting in missplicing of the mRNA leading to a frame shift in translation and the generation of a premature stop codon (##FIG##2##Figure 3B## and ##SUPPL##1##Figure S2##). This allele is a likely molecular null mutation as only a fragment of the ligand-binding domain is present while the transmembrane and cytoplasmic death domains, which are essential for function of this protein, are both absent. The spontaneous mutation <italic>fls</italic>\n^t0sp212 was found to have a splicing defect leading to the inclusion of intronic sequence. This is predicted to form a protein with incorrect amino acid sequence after residue 212, at the end of the transmembrane domain leading to a premature termination codon (##FIG##2##Figure 3B##, ##SUPPL##1##Figure S2##). The two alleles generated by the ENU mutagen both have missense mutations resulting in amino acid changes in the death domain (<italic>fls</italic>\n^t3R367W, R367W^(C-T); <italic>fls</italic>\n^dt3Tpl, I428F ^(A-T)). These mutations were found at identical positions as seen in familial cases of HED in humans (##FIG##2##Figure 3B, E##; ##REF##10431241##[11]##,##REF##17125505##[12]##.</p>", "<title>The <italic>fang</italic> Allele Uncovers Dose and Organ Specific Sensitivity to Levels of Eda Signaling</title>", "<p>The <italic>fang</italic> allele of <italic>fls</italic> was isolated in an allele screen for mutants that failed to complement <italic>fls^te370f</italic> (##FIG##0##Figure 1P##). <italic>fls^tfng</italic> homozygotes do not show any observable effect on lepidotrichia development yet have a reduction of scales and teeth/rakers as seen in other <italic>fls</italic> alleles (##FIG##0##Figure 1M–O##). The fang allele in <italic>trans</italic> to the te370f putative null allele shows an intermediate phenotype affecting lepidotrichial growth and a further reduction of teeth and scales suggesting that the <italic>fang</italic> allele is a hypomorph (##FIG##0##Figure 1P–R##); <italic>fls^tfng</italic> heterozygotes do not show any differences compared to wild type. The shape and number of the scales in <italic>fang</italic> is similar to the other homozygous <italic>fls</italic> alleles (##TAB##0##Table 1##). Analysis of <italic>edar</italic> RNA from homozygous <italic>fls</italic>\n^tfng showed the presence of two distinct transcripts with an additional larger isoform than seen in wildtype. Analysis of the sequence of the novel isoform showed the addition of intronic sequence leading to a premature termination codon (##FIG##2##Figure 3C##). The predicted protein would be similar to the <italic>fls</italic>\n^t0sp212 allele having truncation just after the transmembrane domain at amino acid 218 (##FIG##2##Figure 3B## and ##SUPPL##1##Figure S2##). Analysis of the genomic sequence in the mutant revealed that the altered splicing is due to an A to G transition leading to the creation of a new splice donor site in the intron (##FIG##2##Figure 3C##). Given the presence of both isoforms in homozygous individuals, this novel splice site is used in addition to the normal splice junction. Using quantitative real time PCR we found that the <italic>fang</italic>-specific <italic>edar</italic> transcript represents 74% of the total pool of <italic>edar</italic> transcripts in homozygous mutants (##FIG##2##Figure 3D##). The dilution of wild type transcripts can explain the observed hypomorphic effect of the allele. From this unique allele of <italic>fls</italic>, it is clear that the phenotypic effect of loss of Eda signaling is dose dependent and that scales and teeth are more sensitive to alterations in the level of Eda signaling than fins.</p>", "<title>Ectodysplasin Is Mutated in the <italic>Nackt</italic> Mutant</title>", "<p>EDAR and its orthologue XEDAR recognize specific EDA isoforms that vary by two amino acids ##REF##11039935##[13]##–##REF##11309369##[16]##. The receptor-ligand complex signals though NF-κB using several adaptor proteins that are generally specific to each receptor. Together, mutations in <italic>Edar</italic> and <italic>Eda</italic> lead to the majority of cases of human HED in which the development of integumentary appendages (hairs, glands and teeth) are affected (OMIM 300451, ##REF##8696334##[17]##; OMIM 604095 ##REF##16435307##[18]##).</p>", "<p>We reasoned that, because of the similarity in phenotype to <italic>fls</italic>, the <italic>Nkt</italic> gene could be <italic>eda</italic>, encoding the ligand for Edar. We isolated the entire coding region for zebrafish <italic>eda</italic> by RACE (##SUPPL##2##Figure S3##). The <italic>eda</italic> transcript from the <italic>Nkt</italic> mutant shows a precocious stop codon predicting a truncation of the protein at the beginning of the TNF domain, which is involved in ligand-receptor binding (S243X^(C-A)); ##FIG##2##Figure 3G## and ##SUPPL##2##Figure S3##). An analysis of the location of <italic>eda</italic> in the zebrafish radiation hybrid map placed <italic>eda</italic> on LG5<bold>.</bold> Subsequent linkage analysis of the <italic>Nkt</italic> mutation and markers indicated by radiation hybrid analysis demonstrated tight linkage of the mutant to this region (##FIG##2##Figure 3F##); the S243X mutation was always found in fish with the <italic>Nkt</italic> phenotype and served as a consistent genotypic marker.</p>", "<title>Role of Ectodysplasin Signaling in Regulating Epithelial Signaling Centers: Scale Placode Formation</title>", "<p>In fish, scales are bony elements that develop in the dermis underlying the epidermis. In amniotes, most integumentary organs affected by loss of Eda signaling have structural derivatives stemming from the epidermis (<italic>e.g</italic>. specific keratins of hair, feather and nail, secretory cells of glands). These integumentary organs develop from reciprocal signaling interactions between the basal epidermis and subjacent mesenchyme often controlled by a regional epithelial thickening called the epidermal placode. Eda signaling is necessary for the development and patterning of epithelial placodes of many integumentary organs in both the mouse and chick ##REF##10431226##[19]##–##REF##17362907##[22]##. Expression of <italic>Eda</italic> and <italic>Edar</italic> is found predominantly in the basal epidermal cells, but in the case of feathers <italic>Eda</italic> is detected in the subjacent mesenchyme as well ##REF##15673574##[21]##,##REF##17362907##[22]##. Whereas expression of developmental signaling genes such as <italic>sonic hedgehog (shh)</italic> in the development of integumentary appendages are comparable between vertebrates ##REF##15272389##[23]##, evidence for an early developmental role of the epidermis in induction or patterning of the teleost scale is lacking. The formation of an epithelial placode and signaling center in the development of amniote integumentary appendages is associated with histological changes in the basal cells of the epidermis; a similar structure has not been described in fish epidermis ##REF##9183678##[24]##. As early teleost scale development is quite different to that of other vertebrate integumentary organs, such as hairs and feathers, we addressed the question whether Eda signaling had a similar function in the epidermis of zebrafish during scale formation.</p>", "<p>We detected the expression of both <italic>edar</italic> and <italic>eda</italic> in the skin of juvenile fish by whole mount <italic>in situ</italic> analysis (WMISH). The expression of both genes presaged the formation of the initial scale row along the flank just ventral from the midline mysoseptum in the caudal peduncle (arrowheads, ##FIG##3##Figure 4A and C##; ##REF##9183678##[24]##). During scale formation, the expression of <italic>edar</italic> becomes progressively restricted to the posterior margin (##FIG##3##Figure 4B##) while <italic>eda</italic> expression persists throughout the scale primordia (##FIG##3##Figure 4D##). Developmental genes <italic>shh</italic> and <italic>bone morphogenic protein 2b</italic> (<italic>bmp2b</italic>), whose orthologues are known to be essential for placode development in the mouse and chick, show similar placodal expression as seen with <italic>edar</italic> (##FIG##3##Figure 4E and G##). Analysis of <italic>shh</italic> and <italic>bmp2b</italic> expression in <italic>fls</italic>\n^te370f indicated the necessity of <italic>edar</italic> function for their expression (##FIG##3##Figure 4F and H##).</p>", "<p>We investigated the development of scale primordia in wild type and mutant <italic>fls</italic>\n^te370f fish by light and transmission electron microscopy. Previous detailed histological work found evidence for raised signaling activity in the epidermis as measured by increased endoplasmic reticulum (ER), and secretory activity of the basal epidermal cells prior to scale formation ##REF##15272389##[23]##. However these changes in the basal epidermal cells were not associated regionally with sites of scale formation nor was there any indication of histological changes in basal cell morphology that are associated with placode formation in other vertebrates. To our surprise, in wild type juvenile fish, we discovered the formation of histologically defined, localized assemblies of cells of the basal epidermis that resemble early stages of the formation of hair and feather placodes.</p>", "<p>Prior to the development of the scale, the dermis consists of compact collagen layers (stratum compactum) and scattered dermal fibroblasts ##REF##9183678##[24]##. The epidermal basal cells have a uniform elongate morphology (black arrows, ##FIG##4##Figure 5A, D##) with high levels of basally located intermediate fibrils (##FIG##4##Figure 5G##). At the initiation of scale development, there is an accumulation of fibroblasts subjacent to the basal epidermis, associated with a reworking of the collagen strata ##REF##9183678##[24]##. We find a specific alteration in the morphology of the basal epidermal cells in wild type juvenile fish that coincides with the initial accumulation of fibroblasts at the sites of scale development (##FIG##4##Figure 5B, E##). These basal cells become cuboidal and have decreased width, such that they form a unit of closely packed cells (black arrows, ##FIG##4##Figure 5B, E##). This is observed above the localized accumulation of fibroblasts in the dermis (white arrows, ##FIG##4##Figure 5B, E##). In addition, in these placodal-like cells, the ER appears less prominent (data not shown), and hemidesmosomes, the cellular junctions involved in the attachment of the basal epidermal cells to the basal lamina, are almost completely absent (brackets, ##FIG##4##Figure 5H and I##). In contrast, the adjacent lateral epidermal cells show high levels of both ER and hemidesmosomes (data not shown and ##FIG##4##Figure 5G## brackets, respectively).</p>", "<p>In <italic>fls</italic>\n^te370f juvenile fish, at a corresponding site on the flank as in wild type, we detected the formation of similar aggregations of basal epidermal cells (black arrows. ##FIG##4##Figure 5C, F##). However, unlike the structures found in the wild type zebrafish, the epidermal cells of the placode were disorganized and showed histological evidence of cell death (##FIG##4##Figure 5F##). As is the case in wild type, the epidermal basal cells in the placode of <italic>fls</italic> display a reduced ER, however hemidesmosomes are present in the same high numbers as in adjacent cells in wildtype (brackets, ##FIG##4##Figure 5H## compared to brackets ##FIG##4##Fig. 5I##). Lateral basal epidermal cells in <italic>fls</italic>\n^te370f\n<italic>/edar</italic> showed elongate morphology similar to those of their wild type siblings (data not shown). The expression of <italic>edar</italic> is seen in the basal cells of forming scale placodes (##FIG##4##Figure 5J–L##; stages 1–3 ##FIG##4##Figure 5M##) arising during early specification of the scale placode (s1; arrowhead ##FIG##4##Figure 5K##). We were unable to detect <italic>eda</italic> expression in sections due to the weak hybridization signal.</p>", "<p>These data support the notion that an epidermal placode is involved in dermal scale formation. Further we find that the epithelial organization and function of the developing scale placode is dependent on <italic>edar</italic>.</p>", "<title>Role of Ectodysplasin Signaling in Regulating Epithelial Signaling Centers:Maintenance of the Fin Fold and Establishing Anterior-Posterior Polarity of the Developing Fin</title>", "<p>The phenotype of both <italic>Nkt</italic> and <italic>fls</italic> demonstrate that Eda signaling is necessary for fin development. The growth and patterning of lepidotrichia are affected in all fins. Lepidotrichia are specified, however fail to maintain growth and elaboration of the fin rays (##FIG##5##Figure 6A–C, H–P##). Unpaired fins showed no defects in patterning of the endochondrial bones of the proximal and distal radials (##FIG##5##Figure 6K–P##); the dorsal pitch of the caudal fin is an indirect effect of the mutation on swimming without fin rays (amputated fins that fail to regenerate show similar morphology). In <italic>fls</italic> adults, fusions of the distal radials of the pectoral fin are seen at a low penetrance (##FIG##5##Figure 6G##, data not shown). In <italic>Nkt</italic>, there is an increase in the frequency of patterning and growth defects of the endochondrial components of the fin (##FIG##5##Figure 6G##). These alterations include the loss of the fourth proximal radial, altered growth patterns of anterior proximal radials 1 and 2, as well as lack of articulation of the distal radials (##FIG##5##Figure 6## D–F). <italic>Nkt</italic> causes a strong effect on lepidotrichial growth of both the pectoral and pelvic fins (##FIG##5##Figure 6E–F, J##). In contrast, a specific effect on the growth of anterior lepidotrichia of the pelvic fin is seen in <italic>fls</italic> where the dermal rays of the anterior (<italic>e.g.</italic> 1,2) are significantly shorter than rays at equivalent positions in wild type (##FIG##5##Figure 6I##). The asymmetry of lepidotrichial development suggests that, like the proximal endochondrial fin skeleton, the fin rays have a specific regional identity to provide the shape and form of the fin.</p>", "<p>Early limb development is driven by a localized organization of epithelial cells at the distal tip of the forming limb, termed the apical ectodermal ridge (AER) ##REF##18882505##[25]##. In zebrafish, the AER is involved in larval patterning of the paired fins, while the later stages of fin development are organized by an analogous epidermal formation of the fin fold in both paired and unpaired fins ##REF##10349624##[26]##,##REF##4031750##[27]##. In tetrapod limb development anterior-posterior specification is controlled by posterior mesenchyme expressing <italic>Shh</italic>. The function of this zone of polarizing activity (ZPA) expressing <italic>Shh</italic> is maintained by reciprocal signaling interactions between the ZPA and the AER. This interaction is necessary for proper patterning and growth of the tetrapod limb. In the zebrafish, <italic>shh</italic> and signals from the AER also orchestrate patterning and outgrowth of the early fin buds ##REF##7601313##[28]##–##REF##10976049##[30]##. In addition, genes functioning early in fin development, such as <italic>shh</italic>, play important roles during late fin development regulating growth and branching of lepidotrichia growth ##REF##9753672##[31]##.</p>", "<p>We investigated the regulation of <italic>edar</italic> in mid to late fin development focusing on the development of the paired fins. In early fin fold stage of pectoral fin development (8 mm), we detected <italic>edar</italic> expression in both the distal margin of the endochondrial radials (black arrowhead, ##FIG##6##Figure 7A##) as well as more distally in the forming lepidotrichial rays (##FIG##6##Figure 7A##). The expression of <italic>edar</italic> in the fin fold had a posterior bias in wild type fins (##FIG##6##Figure 7A##, white arrow). The pelvic fin showed similar expression of <italic>edar</italic> in forming lepidotrichial rays (##FIG##6##Figure 7E##). <italic>shh</italic> and <italic>bmp2b</italic> expression was observed in the forming lepidotrichia of both the pectoral and pelvic fins of wild type juveniles, having a similar distal bias in the leading margin of all rays (##FIG##6##Figure 7C, G and I, K##, respectively). Expression of all three genes in <italic>fls</italic> was decreased in the anterior portion of the pectoral fins (##FIG##6##Figure. 7B, D, J##). However, residual expression of all three genes was found in the posterior margin of the fin (##FIG##6##Figure 7B, D, J## arrow). In the pelvic fins, similar loss of anterior expression of <italic>edar</italic> (##FIG##6##Figure 7F##) and <italic>shh</italic> (##FIG##6##Figure 7H##) was seen in the <italic>fls</italic> mutant. We did not detect any difference in <italic>bmp2b</italic> expression in the pelvic fins even though obvious morphological differences in the developing rays of the samples could be seen (##FIG##6##Figure 7L##).</p>", "<p>We asked if the alteration of polarity of gene expression in the <italic>fls</italic> mutant was associated with regional cell death. Using acridine orange uptake as an assay for cell death (<italic>e.g.</italic>\n##REF##8223253##[32]##), we analyzed fins of <italic>fls</italic> and siblings at size matched stages (7–9 mm) for regional patterns of cell death. In 7 mm juveniles, we detect a differential retention of acridine orange between <italic>fls</italic> and siblings in the anterior, and anterior-distal margin of the developing pectoral fin (##FIG##6##Figure 7M–N##). Similarly, in the pelvic fin of 8–9 mm juveniles, retention of the label was seen in the anterior distal margin of the fin (##FIG##6##Figure 7O–P##). Consistent with these data suggesting asymmetrical loss of the fin fold epidermis, we find that <italic>msxa</italic>, a marker of the distal epithelium, is differentially expressed in the mutant (##FIG##6##Figure 7Q and R##).</p>", "<p>We next analyzed gene expression during late development of the lepidotrichia. The expression of <italic>edar</italic> during late fin development was observed in forming lepidotrichia of all fins with a distal bias in its expression (##FIG##7##Figure 8C##). The expression in the forming dermal ray was similar in both location and timing to that of <italic>bmp2b</italic> and <italic>shh</italic> (##FIG##7##Figure 8A–B##, asterisk). In addition, <italic>edar</italic> was found expressed proximally between forming rays and at the distal margin (##FIG##7##Figure 8C## arrow). We were unable to resolve a clear signal for <italic>eda</italic> in the forming fins using WMISH. The expression of <italic>edar</italic> in the distal lepidotrichial tips suggests a late developmental role of Eda signaling in regulating formation of the lepidotrichia in concert with <italic>shh</italic> and <italic>bmp2b</italic>.</p>", "<p>Histological analysis of <italic>fls</italic> mutant fins at an early stage of lepidotrichial formation reveal a general deficiency of the development of the entire mesodermal component of the fin such as cartilage and muscle in both the paired (pectoral fin, ##FIG##7##Figure 8D, E##) and unpaired fins (anal and caudal fins, ##FIG##7##Figure 8F–G##; H–I, respectively). In contrast, the epidermis of the fin is formed and is similar to that of size matched siblings. However, close inspection of the distal tip of the fins showed disorganization of the epidermis and degeneration of distal epidermal nuclei (insets ##FIG##7##Figure 8F, J##). From these analyses, we hypothesize that loss of <italic>edar</italic>-mediated signaling leads to a defect in mesenchymal cell proliferation, muscle cell migration and defective lepidotrichial growth in the fin that correlates with degenerative defects seen in the distal epidermal fin fold.</p>" ]

[ "<title>Discussion</title>", "<p>We used a forward mutagenesis approach in the zebrafish to investigate the developmental mechanisms that underlie changes in adult form. Here, we identified a role of Eda signaling in the development of the dermal skeleton in the adult zebrafish. Mutations in either <italic>Eda</italic> or <italic>Edar</italic> have been shown previously to cause defects in integumentary appendages in several mammalian species. Additionally, Eda signaling genes have been associated with variation in morphology that occurred during the evolution of teleost fishes <italic>(eda)</italic> and in variation of human populations <italic>(Edar)</italic>\n##REF##15790847##[33]##,##REF##17943131##[34]##; see below). Thus, through a forward genetic approach, we were successful in identifying genes that are important for the development and variation in adult form. We further show that the ENU generated alleles of <italic>fls</italic>, <italic>t3R357W</italic> and <italic>dt3Tpl</italic>, affect similar residues as those mutated in familial cases of HED ##REF##10431241##[11]##,##REF##17125505##[12]##,##REF##16435307##[18]## supporting the utility of adult zebrafish mutants as models for the investigation of heritable human disease.</p>", "<title>Conserved and Ancestral Role of Eda Signaling in Vertebrate Development</title>", "<p>We show that Eda signaling is necessary for the development and patterning of the dermal bones of the skull, scales, fin rays as well as teeth of the adult zebrafish. The correlated effect in these zebrafish structures is due to a developmental role of Eda signaling in organizing epithelial cells into signaling centers. In the case of scale development, Eda signaling is necessary for the basal epidermal cells to form a functional placode. Epidermal placodes are involved in the formation of integumentary appendages of other vertebrates such as hair, glands, feathers and teeth. These structures have been shown to act as signaling centers to orchestrate appendage development. We speculate that a primary function of Eda signaling in scale development is to promote cell-cell adhesion within the placode and that the coordinated signaling of the placode induces fibroblast assembly in the underlying dermis, an early step in scale formation.</p>", "<p>Schmidt-Ullrich et al. documents the formation of the hair placode and outline a stage series of placode formation in the mouse ##REF##16481354##[35]##. They report that the <italic>downless</italic> mouse mutant, which has a mutation that disrupts the mouse <italic>Edar</italic> gene ##REF##10431242##[36]##, causes arrest of placode formation at a pre-placode stage of development (P0–P1). This stage closely resembles the stage of scale placode formation that is affected in <italic>fls</italic> shown here. In agreement with our findings, Schmidt-Ullrich et al. further note a reduction of cell-to-cell adhesion within the placode and find increased apoptosis in the absence of Edar function. This suggests that there is a conserved developmental role of Edar between dermal scales of fish and mammalian hairs. During normal hair development, the hair placode invaginates to form the hair bulb. By contrast, the post-placodal events of scale formation in fish do not involve morphogenetic changes of the epidermis, rather the accumulation of mesenchymal cells subjacent to the epidermal placodal cells to form the scale pocket. Thus, Eda signaling in mammals and teleosts is conserved in the early phases of placode formation in controlling the functional continuity and signaling of the epidermal placode to orchestrate appendage formation. However, the downstream interpretation of the epithelial-mesenchymal signaling differs beyond this point leading to altered morphogenetic responses and histological differentiation to form diverse appendages such as scales and hair.</p>", "<p>In the fin, Eda signaling directs late stages of fin development such as the formation and growth of the dermal rays. The effect of loss of <italic>edar</italic> function on fin development uncovers an intrinsic developmental polarity of the late developing fish fin. This is seen both in the development of the proximal endochondrial bones as well as in the formation of the fin rays. We find that the change in patterning in the mutants is correlated with asymmetrical cell death of the distal marginal fin fold as well as a reduction of <italic>shh</italic> expression. This finding is similar to the effect of loss of AER function resulting in anterior-distal cell death and reduction of Shh activity in tetrapod limbs ##REF##18265010##[37]##–##REF##11476582##[39]##. While there has not been any previous indication of a role of Eda signaling in tetrapod limb development, both the expression of <italic>Edar</italic> and related receptor, <italic>Troy</italic>, have been detected in the AER of mice ##REF##12972005##[40]##,##REF##11023871##[41]##.</p>", "<p>A second developmental role of Eda signaling in the developing fin is observed in the outgrowth and patterning of the individual lepidotrichial rays evidenced by expression of <italic>edar</italic> in the distal tip of the forming rays and distal epidermis. The expression of <italic>edar</italic> is again associated with that of <italic>shh</italic> and <italic>bmp2b</italic>. The expression of <italic>shh</italic> and <italic>bmp2b</italic> has been shown to be within the basal epidermis overlying the forming lepidotrichia ##REF##9753672##[31]##. Given the expression of <italic>edar</italic> during fin development and the defects observed in the distal epidermis in the mutant, it is likely that the function of Eda signaling is to maintain the growth permissive function of the fin fold through its regulation of a distal signaling center of individual rays. The concomitant expression of <italic>edar</italic>, <italic>shh</italic> and <italic>bmp2b</italic> in both distal lepidotrichia development and during placode specification suggest that they work in concert to mediate the inductive and/or permissive effects of the epidermis – thus organizing signaling centers for the development of the dermal skeleton.</p>", "<p>While the nature of the defect in tooth formation or dermal bone patterning of the skull in the <italic>fls</italic> and <italic>Nkt</italic> mutants has not been characterized in detail, there is evidence that inductive signaling from the pharyngeal epithelium or cranial epidermis is necessary for appropriate development of both tooth ##REF##14550785##[42]## and skull ##REF##15071597##[6]##, respectively; Eda signaling likely shares a common role in inductive signaling in each of these diverse organs.</p>", "<title>Genetic Commonalities of Eda Signaling: From Fish to Man</title>", "<p>Mutations in the <italic>EDAR</italic> and <italic>EDA</italic> genes underlie a large percentage of autosomal and X-linked HED in humans, respectively ##REF##16435307##[18]##,##REF##11378824##[43]##. In the case of <italic>EDAR</italic>, both recessive and dominant mutations are associated with the HED phenotype in humans, however dominant mutations are found only within the death domain of the protein. These mutations are believed to act in a dominant negative fashion, although by unknown mechanisms ##REF##11570810##[44]##. We see similar dominance of a <italic>fls</italic> allele that affects the death domain of <italic>edar</italic> while all <italic>fls</italic> mutations outside this region do not show a dominant phenotype. Autosomal dominant HED in humans caused by mutation of <italic>EDAR</italic> within the death domain displays a large degree of phenotypic variability ##REF##17125505##[12]##. For example, the I418T mutation in human, which affects the same amino acid as <italic>fls</italic>\n^dt3Tpl (I327F), shows distinct phenotypic variability depending on genetic background ##REF##16435307##[18]##. Interestingly, the <italic>fls</italic>\n^dt3Tpl zebrafish mutant displays similar dominance and variation as the human allele affecting the same residue. These findings suggest that the molecular mechanisms of Edar function are similar between fish and humans.</p>", "<p>X-linked HED caused by mutations in the human <italic>EDA</italic> gene represents the majority of cases of this disease ##REF##11378824##[43]##,##UREF##2##[45]##. The zebrafish <italic>Nkt</italic> mutation described here is affected in the TNF domain and shows a mild dominant phenotype (S243X). As the <italic>EDA</italic> gene is sex linked in humans the molecular nature of different alleles can not be analyzed since the allele will be hemizygous in males and mosaic in female carriers. The zebrafish <italic>eda</italic> gene is autosomal in the zebrafish. Thus, <italic>Nkt</italic> exposes previously unknown dominant function of mutations in this gene since a true heterozygous condition is formed. Hemizygous wildtype condition in humans indicates that the dominance we see in <italic>Eda</italic> is probably not due to haploinsufficiency. Since EDA functions as a homotrimeric protein ##REF##10484778##[46]##, a plausible mechanistic explanation for the observed dominance of <italic>Nkt</italic> is that the C-terminal truncation inhibits the function of the wild type protein in binding to Edar.</p>", "<title>Eda Signaling and the Development and Variation of Adult Form</title>", "<p>Mutations affecting Eda signaling lead to impaired development of integumentary appendages of fish, birds, and man. These changes lead to viable changes in adult morphology. Mutations disrupting Eda signaling have been described for another teleost species. The spontaneous <italic>rs-3</italic> mutant in medaka (<italic>Oryzias latipes</italic>), is shown to have a transposon insertion in the 5′ UTR of <italic>edar</italic> resulting in the reduction of scales but no effect on fin or teeth development ##REF##11516953##[47]##. The zebrafish mutations described here show a previously undescribed role of Eda signaling in the development of the fins, teeth, as well as dermal bones of the skull – phenotypic traits observed in human alleles but not reported in the medaka mutant. As the phenotype of the <italic>rs-3</italic> mutant is similar to the <italic>fls</italic>\n^dfang allele in the zebrafish, it is likely that the more subtle phenotypes observed in the medaka mutant is due to partial loss of function of <italic>edar</italic> caused by a hypomorphic <italic>rs-3</italic> allele ##REF##11516953##[47]##.</p>", "<p>The graded effects seen in the expressivity of mutations affecting <italic>eda</italic> and <italic>edar</italic> points to a dose sensitive readout of the Eda signaling pathway that affects different organ systems with varied expressivity. In the dominant <italic>fls</italic>\n^dt3Tpl or <italic>Nkt</italic> heterozygotes, the shape and number of scales and teeth as well as patterning of the skull are affected, however there is no change in fin ray development. Similarly, the <italic>fang</italic> allele of <italic>fls</italic> clearly demonstrates this dose sensitivity as functional copies of Edar are titrated by the concomitant use of a new splice site in the mutant leading to the reduction in the amount of wild type transcript made (##FIG##2##Figure 3D##). This reduction in the amount of <italic>edar</italic> transcripts cause defects in scale and tooth development, however fins are normal. <italic>fang/te370f</italic>, in which the <italic>fang</italic> allele is in <italic>trans</italic> to a presumed null, further reduces the relative levels of wild type <italic>edar</italic> transcripts leading to further reduction of both teeth as well as fin lepidotrichia. Similar dose sensitive responses to levels of EDA signaling are seen in tooth development of the mouse regulating the number and shape of teeth ##REF##15031115##[48]##,##REF##15538367##[49]##. There are several reports of hypodontia in humans resulting from altered EDA function that do not show other phenotypes such as hypothrichosis or nail defects ##REF##18545687##[50]##–##REF##17256800##[52]##. Given our findings, it is likely that these particular alleles are hypomorphic and this is sufficient to explain the differential organ sensitivity to levels of EDA signaling during development. These data indicate that control of the level of Eda signaling in post-embryonic development is an essential component for the determination of the number and form of many different organ systems of the adult.</p>", "<p>Supporting this finding, we observed significant modification of expressivity of <italic>fls</italic>\n^dt3Tpl in different genetic backgrounds indicating the existence of genetic modifiers of Eda signaling. This sensitivity of Eda signaling to genetic modifiers occurs in other teleost fish as well. In our analysis of the medaka <italic>rs-3/edar</italic> mutant, we find a high degree of variability in the extent of scale formation (##SUPPL##3##Figure S4##) suggesting the existence of background modifiers of Edar function in this species. Additionally, evidence from the stickleback, <italic>Gasterosteus aculeatus</italic>, suggests that genetic variance at the <italic>eda</italic> locus underlies differences in the extent of dermal plate formation in diverged populations of this species ##REF##15790847##[33]##. A quantitative trait analysis (QTL) of lateral plate formation in a low-plated form of the stickleback indicates a significant modification of the reduced plate phenotype (<italic>eda</italic> locus) with modifying effects within and between loci affecting plate number and size ##REF##11780061##[53]##,##REF##15069472##[54]##. Interestingly, recent evidence also shows a significant association between the <italic>edar</italic> locus and dermal plate number in sticklebacks in addition to the predominant <italic>eda</italic> locus ##REF##17371397##[55]##. Thus variation at these gene loci may act in concert to regulate number of dermal plates/scales.</p>", "<p>Thus, while loss of Eda signaling can lead to severe phenotypes, the phenotypic consequences of variation in Eda signaling are graded and canalization of Eda signaling is prevalent. Therefore, buffering of the phenotypic outcome that results from defective Eda signaling could be a common mechanism that permits viable and diverse phenotypes. These viable phenotypic variations then could serve as a basis for selection. The lack of a coding change at the <italic>eda</italic> locus in sticklebacks that is associated with the loss of dermal plates has lead to the argument that, in this case, evolution of this trait is due to changes at <italic>cis</italic>-regulatory elements controlling <italic>eda</italic> expression ##REF##15790847##[33]##. Our findings on the dose and organ specific sensitivity of Eda signaling in different structures of the zebrafish argues that evolution of this trait could result from a regulation of absolute levels of expression.</p>", "<p>Interestingly, recent analysis of single nucleotide polymorphism (SNP) frequency in human populations supports the role of Eda signaling in causing phenotypic variation. Analysis of SNP variation between diverse human populations shows evidence of selection of the <italic>EDAR</italic> locus in East Asian and American populations ##REF##17356696##[56]##,##REF##17542651##[57]##. A defined allelic variant of <italic>EDAR</italic> within these populations leads to a coding change in the death domain of EDAR and is a candidate allele for altered gene function that could have lead to the region being fixed in these populations ##REF##17943131##[34]##. There is evidence from association data that this allele is associated with thick hair in these populations ##REF##18065779##[58]##, however the full extent of phenotypes that are affected in these populations that are related to EDA signaling has not been analyzed. It is interesting to note that recent work has identified this allele of <italic>EDAR</italic> as having an enhanced effect on Eda signaling in mouse models containing the altered human residue ##UREF##3##[59]##. Given that variation in the number and shape of integumentary derivatives of the dermal skeleton are a common morphological change in teleost evolution <italic>e.g.</italic>\n##UREF##4##[60]##, it will be important to further investigate the prevalence and type of genetic changes in Eda signaling genes in cases of natural variation of these adult characters.</p>", "<title>Mutagenesis and Allele Designation</title>", "<p>Zebrafish mutagenesis was performed following ##UREF##5##[61]## with 5 treatments of 3.3–3.5 mM ethylnitrosourea. Screen design was similar to that described ##UREF##6##[62]##. Allele designation was determined using standard nomenclature with the addition of the molecular lesion or phenotypic description (when appropriate) to the designation. The serial numbers of the mutants found in the ZF models screen are as follows: <italic>fls</italic>\n^t3R367W (#0621); <italic>fls</italic>\n^dt3Tpl (#1248); <italic>Nkt</italic>\n^dt3S243X (#1261). Information on the screen can be found at <ext-link ext-link-type=\"uri\" xlink:href=\"http://www.zf-models.org/\">http://www.zf-models.org/</ext-link>. The screen for additional <italic>fls</italic> alleles used mutagenized <italic>TLF</italic> founder males treated similarly as Tü males used in the screen.</p>", "<title>Mapping</title>", "<p>Rough mapping of F2 progeny against a reference panel of SSLP markers ##UREF##7##[63]## indicated that <italic>fls</italic> was located on linkage group (chromosome) 9 (LG9) with loose linkage to z20031 (61.3cR; ##FIG##2##Figure 3A##). We found <italic>fls</italic> to be closely linked to markers z7001 and z11672. Results from radiation hybrid screening indicated linkage of zebrafish <italic>edar</italic> to markers positioned on LG9 in the region predicted by initial mapping analysis. Analysis of flanking markers and internal polymorphisms in <italic>edar</italic> showed tight linkage of the <italic>fls</italic> mutation to the <italic>edar</italic> gene. Using the defined molecular differences between WIK and Tü strains, we did not find recombination in 238 meioses indicating that the mutation was located less than 0.4 cM away.</p>", "<title>Cloning and Sequence Analysis</title>", "<p>We isolated the full-length cDNA of zebrafish <italic>edar</italic> and <italic>eda</italic> by reverse transcription (RT) PCR using sequences provided from genomic alignments and subsequent amplification of the cDNA ends by rapid amplification of DNA ends (RACE). cDNA was generated from RNA from blastemas of amputated caudal fins that had been allowed to regenerate for two days. cDNA sequences of zebrafish <italic>edar</italic> and <italic>eda</italic> genbank accession numbers are EF137867 and EF137866, respectively. Protein alignment of Edar and Eda were generated by ClustalW alignment (<ext-link ext-link-type=\"uri\" xlink:href=\"http://www.ebi.ac.uk/clustalw/\">http://www.ebi.ac.uk/clustalw/</ext-link>) and Box Shade software (<ext-link ext-link-type=\"uri\" xlink:href=\"http://www.ch.embnet.org/software/BOX_form.html\">http://www.ch.embnet.org/software/BOX_form.html</ext-link>) using a 0.4 identity threshold. <italic>Edar</italic> and <italic>eda</italic> sequences of other species were obtained from genomic databases at NCBI (<ext-link ext-link-type=\"uri\" xlink:href=\"http://www.ncbi.nlm.nih.gov/\">http://www.ncbi.nlm.nih.gov/</ext-link>), Sanger (<ext-link ext-link-type=\"uri\" xlink:href=\"http://www.ensembl.org/index.html\">http://www.ensembl.org/index.html</ext-link>), and Tigr (<ext-link ext-link-type=\"uri\" xlink:href=\"http://www.tigr.org/tdb/tgi/\">http://www.tigr.org/tdb/tgi/</ext-link>).</p>", "<p>Real Time PCR was performed on cDNA obtained from blastemas from two day old caudal fin regenerates. Calculations were made from three biological replicates and three technical replicates according to ##REF##11328886##[64]##. Primers were designed for wild type specific transcripts by using sequence from the neighbouring exon borders that are not adjacent to each other in Fang-transcripts. Fang specific primers were designed against the fang specific transcript sequence, which is spliced out in wild type. Crossing points of the control reaction, which were higher than in the water control, were set to the value for water. Normalization was done against the efficiency of primers to β-actin.</p>", "<title>Bone Stain and Measurements</title>", "<p>Adult bones were stained with alizarin red. Embryos were fixed in formalin (3.7% formaldehyde), briefly dehydrated in 70% ethanol, and placed in 1 g/l alizarin red; 0.5% KOH until bones suitably stained. Fish were destained in 1% KOH until background stain was lost and subsequently cleared in glycerol for analysis. For analysis of forming cartilage, fish were prestained in alcian blue from 4–24 hours. The fish were then destained, lightly trypsinized (3 g/l; 37°C) and processed for alzarin red staining.</p>", "<p>Skeletal measurements were made using digitizing software from Zeiss using a dissecting microscope. Measurements were made from fixed landmarks on each axis of the skull that did not vary depending on position of the suspensorium: the premaxilla was used for the distal most point on the length (L) axis, while the quadrate-anguloarticular joint was used as a ventral landmark for the height (H) axis. Raw measurement values are represented as normalized ratios of the distance along each axis in relation to the position of the center of the eye; values are normalized for standard length of the fish.</p>", "<title>Whole Mount In Situ Hybridisation and Immunostaining</title>", "<p>Probes for whole mount <italic>in situ</italic> hybridisation were generated by reverse transcription from cDNA made from regenerating caudal fin tissue. Digoxigenin labeled RNA probes were purified using P-30 micro bio-spin columns before use (BioRad). WMISH protocol was performed as described ##UREF##8##[65]## , at 70°C and with the addition of 0.1% CHAPS to hybridization and post hybridization wash buffers. Reactions were stopped in PBS, post-fixed and placed in methanol overnight to reduce non-specific staining.</p>", "<p>Acridine orange (Sigma) was used as a marker of apoptosis in developing tissue ##REF##8223253##[32]##. Juveniles were immersed in fish water containing 5 μg/ml acridine orange for 5 minutes and then washed with fish water, anesthetized and post-fixed in formalin to assist visualization of staining.</p>", "<title>Statistical Methods</title>", "<p>Analysis of cranial measurements were performed using Hotelling's T squared test for two dependent variables. For scale counts and size dimensions, a <italic>t</italic>-statistic for differential means was used to assess significance. Calculations and probability assessment were calculated using Biosoft 200 software (<ext-link ext-link-type=\"uri\" xlink:href=\"http://www.biosoft.com\">www.biosoft.com</ext-link>) and Excel statistical package.</p>", "<title>Electron Microscopy</title>", "<p>Specimens of 8–9 mm juvenile fish were fixed with a mixture of 4% formaldehyde in PBS and 1–2.5% glutaraldehyde at room temperature and subsequently placed at 4°C. After post-fixation with 1% osmium tetroxide in 100 mM PBS for 1 h on ice, samples were washed with H₂O, treated with 1% aqueous uranyl acetate for 1 h at 4°C, dehydrated through a graded series of ethanol and embedded in Epon. Ultrathin sections were stained with uranyl acetate and lead citrate and viewed in a Philips CM10 electron microscope. In addition, toluidine blue stained Epon sections of 0.5 or 3 μm thickness were prepared for light microscopy.</p>" ]

[]

[ "<p>Conceived and designed the experiments: MPH. Performed the experiments: MPH NR HS SP. Analyzed the data: MPH NR HS SP PK. Contributed reagents/materials/analysis tools: MPH NR. Wrote the paper: MPH NR CNV.</p>", "<p>The genetic basis of the development and variation of adult form of vertebrates is not well understood. To address this problem, we performed a mutant screen to identify genes essential for the formation of adult skeletal structures of the zebrafish. Here, we describe the phenotypic and molecular characterization of a set of mutants showing loss of adult structures of the dermal skeleton, such as the rays of the fins and the scales, as well as the pharyngeal teeth. The mutations represent adult-viable, loss of function alleles in the <italic>ectodysplasin</italic> (<italic>eda</italic>) and <italic>ectodysplasin receptor</italic> (<italic>edar</italic>) genes. These genes are frequently mutated in the human hereditary disease hypohidrotic ectodermal dysplasia (HED; OMIM 224900, 305100) that affects the development of integumentary appendages such as hair and teeth. We find mutations in zebrafish <italic>edar</italic> that affect similar residues as mutated in human cases of HED and show similar phenotypic consequences. <italic>eda</italic> and <italic>edar</italic> are not required for early zebrafish development, but are rather specific for the development of adult skeletal and dental structures. We find that the defects of the fins and scales are due to the role of Eda signaling in organizing epidermal cells into discrete signaling centers of the scale epidermal placode and fin fold. Our genetic analysis demonstrates dose-sensitive and organ-specific response to alteration in levels of Eda signaling. In addition, we show substantial buffering of the effect of loss of <italic>edar</italic> function in different genetic backgrounds, suggesting canalization of this developmental system. We uncover a previously unknown role of Eda signaling in teleosts and show conservation of the developmental mechanisms involved in the formation and variation of both integumentary appendages and limbs. Lastly, our findings point to the utility of adult genetic screens in the zebrafish in identifying essential developmental processes involved in human disease and in morphological evolution.</p>", "<title>Author Summary</title>", "<p>A major goal of the study of developmental genetics is to understand the genes and developmental mechanisms underlying the formation of organismal complexity and diversity. Here, we focus on genes controlling postembryonic development and describe mutations in genes of the ectodysplasin (Eda) pathway in regulating the formation of the scales, skull, fins, and teeth. Mutations in genes of this signaling pathway are common in humans with defects in ectodermal structures such as hair, glands, and teeth. We show that the similar phenotypes of loss of Eda signaling in fish and human are due to a conserved early developmental stage in the development of mammalian hair and fish scales; subsequent development of these two structures diverge. Our findings show that the Eda signaling pathway has an ancestral role in regulating the developmental interactions involved in patterning and growth of the dermal skeleton of fish. Recent work has shown that these genes are associated with morphological variation between humans and evolution within fish populations, suggesting that alteration in the function of these genes permits viable morphological change. Our data support the value of forward genetic studies on postembryonic development to reveal the genetic and developmental basis of both human disease and morphological evolution.</p>" ]

[ "<title>Supporting Information</title>" ]

[ "<p>The authors would like to thank Drs. M. Sonawane, K. Siegfried, C. Koehler, M. Levesque, and JY Sire for helpful comments on the manuscript, Dr. M. Hargrave for contributing the <italic>fls</italic>\n^t0sp213 background allele, M. Akimenko for <italic>msxa</italic> cDNA, and Dr. Vu Nguyen for discussions and sharing of results before publication. We are grateful to Dr. Hiroshi Mitani in his gracious offer to provide medaka/<italic>rs-3</italic> mutant for anatomical analysis. In addition, the authors would like to thank the assistance of Dr. Robert Geisler and Ines Gehring for radiation hybrid analysis, Jennifer Zenker for assistance with characterization of the <italic>fls</italic>\n^te370f mutant, and Brigitte Sailer and Iris Koch for their expertise in histology. Lastly, we would like to thank the assistance and advice of three anonymous reviewers.</p>" ]

[ "<fig id=\"pgen-1000206-g001\" position=\"float\"><object-id pub-id-type=\"doi\">10.1371/journal.pgen.1000206.g001</object-id><label>Figure 1</label><caption><title>The formation of the adult dermal skeleton and pharyngeal teeth is affected in <italic>fls</italic> mutant zebrafish.</title><p>A) Alizarin red-stained wild type adult zebrafish shows staining of the scales, fin rays, dermal bones of the skull as well as the pharyngeal teeth along ceratobranchial 5 (bracket, B) and (C) gill rakers along both the anterior (antGr) and posterior edge (postGr) of non-teeth bearing ceratobranchials. D) <italic>fls</italic>\n^te370f shows loss of dermal skeletal structures of the fin rays, scales and alteration in the shape of the skull. Additionally, <italic>fls</italic>\n^te370f shows loss of pharyngeal teeth (E) and gill rakers (F). G–I) The Topless allele (<italic>fls^dt3Tpl</italic>) shows a dominant effect on scalation and tooth/gill raker formation while not affecting lepidotrichial growth. J–L) Expressivity of <italic>fls</italic>\n^dt3Tpl is sensitive to a modifier in the Tü strain leading to a “weak” <italic>fls</italic>\n^dt3Tpl phenotype; <italic>fls</italic>\n^dt3Tpl homozygotes were phenotypically identical to <italic>fls</italic>\n^te370f (not shown). The fang allele of <italic>fls</italic>, <italic>fls</italic>\n^tfang, isolated in a non-complementation screen with <italic>fls</italic>\n^te370f, shows no effect on fin development while exhibiting partial loss of scales (M), teeth (N), and gill rakers (O). Transallelic tfang/te370f zebrafish exhibit an intermediate phenotype between homozygous <italic>fls</italic>(te370f) and <italic>fls</italic>(tfang) (P–R).</p></caption></fig>", "<fig id=\"pgen-1000206-g002\" position=\"float\"><object-id pub-id-type=\"doi\">10.1371/journal.pgen.1000206.g002</object-id><label>Figure 2</label><caption><title>The dominant gene <italic>Nkt</italic> is phenotypically similar, however complements <italic>fls</italic> mutants.</title><p>\n<italic>Nkt</italic> homozygotes show complete loss of scales, teeth and gill rakers resembling the <italic>fls</italic> phenotype (A–C). Heterozygous <italic>Nkt</italic> zebrafish show an intermediate phenotype of scale loss and patterning defect (arrows) while no effect on fin development is seen (D). Heterozygous <italic>Nkt</italic> also show a dominant effect on the number of teeth (arrows, E) and gill rakers (F), showing deficiencies along the posterior branchial arches and formation of rudimentary rakers along ceratobranchial 1 and 2 (arrows, F). <italic>Cb1-5</italic>, ceratobranchial bones.</p></caption></fig>", "<fig id=\"pgen-1000206-g003\" position=\"float\"><object-id pub-id-type=\"doi\">10.1371/journal.pgen.1000206.g003</object-id><label>Figure 3</label><caption><title>\n<italic>fls</italic> and <italic>Nkt</italic> are mutations in genes encoding ectodysplasin receptor (<italic>edar</italic>) and its ligand ectodysplasin (<italic>eda)</italic>.</title><p>A) Mapping of <italic>fls</italic> using SSLP markers and placement of the <italic>edar</italic> gene within the candidate region on LG 9 by radiation hybrid mapping. The insert shows genetic linkage of the <italic>fls</italic> gene to local markers on LG 9. The numbers on the right of the insert indicate the number of recombinants seen in identified mutants per the number of meioses tested. B) Schematic of wild type Edar protein and mutant alleles. Polymorphisms seen in the WIK strain are noted above the wild type gene. Mutations that lead to premature termination are represented as truncated proteins showing the predicted residual fragment and position of the identified mutation. C) Analysis of the mutation in <italic>fls</italic>\n^dfang. A unique splice donor site (red) is generated leading to inclusion of additional coding sequence encoding a premature termination codon (underlined). D) Quantitative analysis of different <italic>edar</italic> transcript levels in <italic>fls</italic>\n^dfang homozygotes compared with wildtype. E) Similarity of altered residues in <italic>fls</italic>\n^t3R367Wand <italic>fls</italic>\n^dt3Tpl with human HED shown in the death domain. The position of the mutated residues in <italic>fls</italic>\n^t3R367W (blue box) and in <italic>fls</italic>\n^dt3Tpl (red box) is identical to ones changed in cases of human autosomal dominant HED although the substitution is different. F) Linkage between the <italic>Nkt</italic> allele and <italic>eda</italic> on LG5. G) Schematic of wild type Eda protein and position of <italic>Nkt</italic> mutation. Numbers on gene diagrams represent amino acid length. <italic>TNF</italic>, tumor necrosis factor domain; <italic>TNFR</italic>, tumor necrosis factor receptor domain; <italic>TM</italic>, transmembrane domain; <italic>DD</italic>, death domain; the blue box in <italic>Eda</italic> is the furin binding site.</p></caption></fig>", "<fig id=\"pgen-1000206-g004\" position=\"float\"><object-id pub-id-type=\"doi\">10.1371/journal.pgen.1000206.g004</object-id><label>Figure 4</label><caption><title>The role of <italic>edar</italic> in expression of developmental patterning genes during early scale development.</title><p>Expression of <italic>edar</italic> (A, B) and <italic>eda</italic> (C, D) in early forming scales; arrowheads indicate site of expression of initial forming scales. A) <italic>edar</italic> expression above site of scale formation in 8 mm long (approximately 30 dpf juvenile fish) and in larger juveniles (9 mm; 30 dpf). C) <italic>eda</italic> expression during early scale development on the flank (8 mm) and in forming scales of older juvenile fish (10 mm, D). Expression of developmental genes <italic>bmp2b</italic> and <italic>shh</italic> in early scale development in wildtype (E, G) and <italic>fls</italic>\n^te370f (F, H) juveniles (9 mm).</p></caption></fig>", "<fig id=\"pgen-1000206-g005\" position=\"float\"><object-id pub-id-type=\"doi\">10.1371/journal.pgen.1000206.g005</object-id><label>Figure 5</label><caption><title>Eda signaling regulates the formation of an epidermal placode during scale development.</title><p>Histological analysis of wild type (A, D, B, E) and <italic>fls</italic>\n^te370f (C, F, I) integument of 8 mm standard length. In wild type juveniles (B, E), basal epidermal cells (black arrow heads) show a heightened, and cuboidal morphology at sites of scale development as indicated by an accumulation of migrating fibroblast-like cells (white arrowheads). (H) This morphology of the epidermis is associated with a reworking of the collagen layer of the stratum compactum (<italic>cpt</italic>; ##REF##9183678##[24]##). This is in contrast to the flattened morphology of basal epidermal cells lateral to those of the scale placode (A, D) and underlying dense stratum compactum (G). In <italic>fls</italic>\n^te370f this basal epidermal structure is disorganized and cell morphology is disrupted (C, F) including evidence of cell death (asterisk). The lack of reworking of the collagen of the stratum compactum in the <italic>fls</italic>\n^te370f mutant is associated with retention of hemidesmosomes (horizontal bracket G–I). <italic>edar</italic> is expressed in cells of the wildtype epidermis (J, K, L). Counterstaining of the same sections confirms the expression in basal cells overlying initial accumulating fibroblasts (white arrowheads; J′, K′, L′). Expression of <italic>edar</italic> is observed prior to organization of the placode and fibroblast aggregation and maintained in cells of the epidermal placode through early scale development (J–L). M) Schematic depicting scale development and <italic>edar</italic> expression. The stages of scale development are modeled using analogous stages as described for hair development ##REF##16481354##[35]##; stage 0, nascent epidermis; stage 1, placode specification; stage 2, scale pocket; stage 3, matrix deposition and ossification. Blue, <italic>edar</italic> expression; red, scale formation. <italic>ep</italic>, epidermis; <italic>cpt</italic> stratum compactum. The vertical bracket demarcates the extent of the epidermis in the sections. Measurement bar equals 10 μm.</p></caption></fig>", "<fig id=\"pgen-1000206-g006\" position=\"float\"><object-id pub-id-type=\"doi\">10.1371/journal.pgen.1000206.g006</object-id><label>Figure 6</label><caption><title>Fin development is defective in <italic>fls</italic> and <italic>Nkt</italic> mutant zebrafish.</title><p>Alizarin red stained adult zebrafish fins show a drastic effect of <italic>fls</italic>\n^te370f and <italic>Nkt</italic> on development of the lepidotrichial dermal rays of both the paired and unpaired fins. A–C,F) Pectoral fins, anterior-dorsal view; D, E) double staining developing pectoral fins with alcian blue and alizarin red show early patterning of the endochondrial bones of the pectoral fin of size matched wild type and <italic>Nkt</italic> homozygotes (asterisk indicates loss of fourth proximal radial). G) effect of <italic>fls</italic> and <italic>Nkt</italic> mutants on the patterning of the pectoral fin skeleton scored as the number of specimens showing alteration in pattern or form over total analyzed. The identity of the proximal radials is noted (I–IV). H–J) Analysis of pelvic fin development in <italic>fls</italic> and <italic>Nkt</italic> mutants. Numbers denote anterior-posterior identity of the lepidotrichia. K–M) Defects in the formation of the lepidotrichia in adult anal and (N–P) caudal fin of <italic>fls</italic> (L and O) and <italic>Nkt</italic> (M and P). <italic>ap</italic>, ascending process; <italic>cl</italic>, cleithrum; <italic>co</italic>, corticoid; <italic>dr</italic>, distal radial; <italic>sc</italic>, scalpula, <italic>le</italic>, lepidotrichia; <italic>pcl</italic>, postcleithrum; <italic>pr</italic>, proximal radial.</p></caption></fig>", "<fig id=\"pgen-1000206-g007\" position=\"float\"><object-id pub-id-type=\"doi\">10.1371/journal.pgen.1000206.g007</object-id><label>Figure 7</label><caption><title>Eda signaling and the maintenance of anterior-posterior pattern in late paired fin development.</title><p>Analysis of <italic>edar</italic> (A–B, E–F), <italic>shh</italic> (C–D, G–H), and <italic>bmp2b</italic> (I–L) expression in developing pectoral (A–D, –J) and pelvic fins (E–H, K–L) from 8 mm juvenile fish of wild type (A, E; C, G; I, K) and <italic>fls</italic>\n^te370f mutant fish (B, F; D, H; J, L). A–B) Arrowheads indicate two distinct patterns of <italic>edar</italic> expression in the pectoral fin: an expression that marks the posterior edge and distal region of the development of the proximal radials (black); and a posterior bias of <italic>edar</italic> expression in the forming lepidotrichia (white). Arrows point out the remaining posterior expression in mutant fins. Brackets in all panels outline anterior deficiencies in gene expression in <italic>fls</italic> mutant fins. M–P, analysis of patterns of cell death in the developing paired fins by retention of acridine orange stain. N, P) Arrows point out anterior distal regions of cell death in both pectoral and pelvic fins from the mutant; (M, N) pectoral fin and (O, P) pelvic fin respectively. Q, R) Expression of <italic>msxa</italic> in wild type and mutant pectoral fins. Region of expression outlined with brackets; asterisk marks an ectopic site of expression.</p></caption></fig>", "<fig id=\"pgen-1000206-g008\" position=\"float\"><object-id pub-id-type=\"doi\">10.1371/journal.pgen.1000206.g008</object-id><label>Figure 8</label><caption><title>Eda signaling is required for the function of the fin fold during late fin development.</title><p>Expression of <italic>bmp2b</italic> (A), <italic>shh</italic> (B) and <italic>edar</italic> (C) transcripts in developing juvenile (8 mm, 30 dpf) fin rays of the caudal fin; asterisks indicate regional expression within distal tip of developing ray; arrows in (C), expression in distal epidermis of the fin fold. D–I) Histological analysis of both paired (pectoral, D, E) and unpaired fins (anal F, G; and caudal, H, I) from <italic>fls</italic>\n^te370f and wild type siblings. <italic>fls</italic>\n^te370f fins showed a general deficiency in the maturation of the muscle and dermis of the fin (arrow G; acellular debris in anal fin of <italic>fls</italic>). Insets (D, E), tip of fin at higher magnification showing degeneration of the nuclei of the epidermis in the mutant fin. <italic>ffd</italic>, fin fold; <italic>le</italic>, lepidotrichia of the fin rays.</p></caption></fig>" ]

[ "<table-wrap id=\"pgen-1000206-t001\" position=\"float\"><object-id pub-id-type=\"doi\">10.1371/journal.pgen.1000206.t001</object-id><label>Table 1</label><caption><title>Quantitative effect of <italic>fls</italic> on scale number and shape and the effect of background modifiers in <italic>Danio rerio</italic> strains on <italic>fls</italic>\n^dt3Tpl.</title></caption><alternatives><table frame=\"hsides\" rules=\"groups\"><colgroup span=\"1\"><col align=\"left\" span=\"1\"/><col align=\"center\" span=\"1\"/><col align=\"center\" span=\"1\"/><col align=\"center\" span=\"1\"/><col align=\"center\" span=\"1\"/></colgroup><thead><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">Phenotype/Genotype</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">Scale #/ stl</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">n</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">Scale DV/AP</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">n</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\"/><td align=\"left\" rowspan=\"1\" colspan=\"1\"/><td align=\"left\" rowspan=\"1\" colspan=\"1\">fish</td><td align=\"left\" rowspan=\"1\" colspan=\"1\"/><td align=\"left\" rowspan=\"1\" colspan=\"1\">scales</td></tr></thead><tbody><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">+/+</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">6.8±0.18</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">4</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">1.14±0.15</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">13</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">\n<italic>fls</italic>\n^dt3Tpl / Tü</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">3.0±0.20 ##</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">2</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">1.52±0.29 #</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">8</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">\n<italic>fls</italic>\n^dt3Tpl / Tü; <italic>mod</italic>\n</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">5.6±0.44 #</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">3</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">1.4±0.3 #</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">12</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">\n<italic>fls</italic>\n^dt3Tpl / WIK</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">5.84±0.66 #</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">9</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">1.43±0.35 #</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">32</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">\n<italic>fls</italic>\n^tfang / <italic>fls^t</italic>\n^fang\n</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">0.97±0.50 ###</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">2</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">1.57±0.18 #</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">7</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">\n<italic>fls</italic>\n^te370f / <italic>fls</italic>\n^te370f\n</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">0.41±0.39 ###</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">6</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">1.8±0.64 #</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">16</td></tr></tbody></table></alternatives></table-wrap>" ]

[]

[]

[]

[]

[]

[ "<supplementary-material content-type=\"local-data\" id=\"pgen.1000206.s001\"><label>Figure S1</label><caption><p>\n<italic>fls</italic> and <italic>Nkt</italic> alter size and proportion of the adult zebrafish skull. In addition to variations in integumentary structures, <italic>fls</italic> and <italic>Nkt</italic> exhibited a distinct change of the shape and size of the adult skull. Measurements of the absolute proportions of the adult skull, normalized for overall growth of the fish as determined by standard length, demonstrate that both <italic>fls</italic> and <italic>Nkt</italic> homozygous mutations result in overall larger skulls of the fish (<italic>fls</italic>\n^{te370f/te370f}, n = 16, T2 = 23.7, p&lt;0.001; <italic>Nkt</italic>, n = 10, T2 = 64.1, p&lt;0.001; <italic>fls</italic>\n^{te370f/dt3Tpl}, n = 6, T2 = 22.8, p&lt;0.005). The dominant effect of <italic>Nkt</italic> and <italic>fls</italic>\n^dt3Tpl seen in development of the scale pattern was not observed in the formation of skull size. However, an analysis of changes in the proportional development of the skull by measurements of the relative positioning of the eye within the skull (L1/L2, H1/H2; Panel A) showed a significant and dominant effect of <italic>Nkt</italic>, <italic>fls</italic>\n^dt3Tpl on the patterning of the skull (Panel C). This effect was seen in <italic>fls</italic>\n^te370f homozygotes as well and was not specific to particular alleles of <italic>fls</italic>. The alteration in skull size and shape in the mutants does not involve loss of a particular organ structure or specific bone, rather a change in proportions of the developing skull.</p><p>(3.39 MB TIF)</p></caption></supplementary-material>", "<supplementary-material content-type=\"local-data\" id=\"pgen.1000206.s002\"><label>Figure S2</label><caption><p>A comparison of <italic>edar</italic> sequence in representative vertebrates. Edar alleles <italic>fls</italic>\n^dt3Tpl and <italic>fls</italic>\n^t3R367W positioned above sequence. Sites of splicing defects of <italic>fls</italic>\n^te370f and <italic>fls</italic>\n^t0sp213 alleles demarcated with ˆ marker. Yellow, TNFR domain; Grey, transmembrane domain; Green, death domain; Red, polymorphic sites in WIK mapping strain.</p><p>(0.04 MB DOC)</p></caption></supplementary-material>", "<supplementary-material content-type=\"local-data\" id=\"pgen.1000206.s003\"><label>Figure S3</label><caption><p>A comparison of <italic>eda</italic> sequence of representative vertebrates. Blue, transmembrane domain; Green, furin cleavage site; Yellow, TNF domain; asterisk <italic>Nkt</italic>\n^dtS238X allele; |, deleted residues in alternate spliced form of Eda-2.</p><p>(0.04 MB DOC)</p></caption></supplementary-material>", "<supplementary-material content-type=\"local-data\" id=\"pgen.1000206.s004\"><label>Figure S4</label><caption><p>Scale formation and variation in the rs3/<italic>edar</italic> medaka mutant on the cs-2 background. (A) Alizarin-red stained rs3 medaka showed substantial scale formation and variation of the extent of scalation. (B) Wild type cs-2 strain scalation pattern.</p><p>(5.72 MB TIF)</p></caption></supplementary-material>" ]

[ "<table-wrap-foot><fn id=\"nt101\"><label/><p>The total number of scales on one side of alizarin red stained adults of different genotypes were counted and measured. Counts were normalized for standard length (stl) of individual fish as shape and number of scales in the mutants may vary as a measure of size. Shape characteristics of scales were quantified by measuring three to four scales from set positions across the flank of each fish and comparing the height (dorsal-ventral; DV) to length (anterior-posterior; AP) ratios. Results are presented as sample average and standard deviation around the mean. <italic>mod</italic>, inferred genotype of a modifier in Tü background leading to “weak” phenotype. The numerical symbol (#) indicates significant difference compared to wild type values (students <italic>t</italic>, p&lt;0.05). The different number of symbols signifies a significantly different phenotypic classes of scale development (#, ##, ###).</p></fn></table-wrap-foot>", "<fn-group><fn fn-type=\"COI-statement\"><p>The authors have declared that no competing interests exist.</p></fn><fn fn-type=\"financial-disclosure\"><p>MPH was supported by a Max Planck Postdoctoral Fellowship.</p></fn></fn-group>" ]

[ "<graphic xlink:href=\"pgen.1000206.g001\"/>", "<graphic xlink:href=\"pgen.1000206.g002\"/>", "<graphic id=\"pgen-1000206-t001-1\" xlink:href=\"pgen.1000206.t001\"/>", "<graphic xlink:href=\"pgen.1000206.g003\"/>", "<graphic xlink:href=\"pgen.1000206.g004\"/>", "<graphic xlink:href=\"pgen.1000206.g005\"/>", "<graphic xlink:href=\"pgen.1000206.g006\"/>", "<graphic xlink:href=\"pgen.1000206.g007\"/>", "<graphic xlink:href=\"pgen.1000206.g008\"/>" ]

[ "<media xlink:href=\"pgen.1000206.s001.tif\"><caption><p>Click here for additional data file.</p></caption></media>", "<media xlink:href=\"pgen.1000206.s002.doc\"><caption><p>Click here for additional data file.</p></caption></media>", "<media xlink:href=\"pgen.1000206.s003.doc\"><caption><p>Click here for additional data file.</p></caption></media>", "<media xlink:href=\"pgen.1000206.s004.tif\"><caption><p>Click here for additional data file.</p></caption></media>" ]

{ "acronym": [], "definition": [] }

[ "<title>Introduction</title>", "<p>The genomic era has provided us with hundreds of complete microbial genome sequences, and has allowed us to generate genome sequences from whole environments using a “metagenomics” approach ##REF##16339357##[1]##,##REF##15590779##[2]##. Taken together, the sequencing of individual genomes and whole communities has enabled the realization of a level of genetic diversity and complexity that was previously unappreciated. This massive volume of data has led to an increasing reliance on bioinformatic predictions, since the traditional experimental approach of characterizing gene function one at a time can not keep pace with the sequence-based discovery of novel putative genes. Automated bioinformatic pipelines together with manual curation by expert human annotators typically allow the functional predictions for 50–70% of the genes of a newly sequenced microorganism ##REF##15980861##[3]##,##REF##17468768##[4]##.</p>", "<p>Bioinformatic predictions are largely based on sequence similarity to known proteins based on BLAST, Hidden Markov Model or other searches, along with supporting evidence based on genome context methods ##REF##16931121##[5]##. <italic>Ab initio</italic> prediction methods remain highly speculative in the absence of other evidence, hence bioinformatic gene function predictions are essentially limited to what we already know experimentally from other systems. Furthermore, the accuracy of bioinformatic predictions remains largely undetermined, i.e., the likelihood of any single gene functional assignment being correct is at best an educated guess.</p>", "<p>Our group has focused on bioinformatic predictions of membrane transporter function, developing a pipeline for annotation of membrane transport genes and a relational database, TransportDB, describing the predicted transporter content of all sequenced genomes ##REF##14681414##[6]##,##REF##16118665##[7]##. Hence we have been interested in finding approaches to functionally characterize transporter genes in a high throughput fashion to assess the accuracy of our bioinformatic predictions. In some aspects, membrane transport genes are good candidates for high throughput phenotypic screens, since in many cases individual knockout mutants might be expected to give relatively simple phenotypes, e.g., loss of a glucose transporter causing a defect on growth on glucose as a sole carbon source. Of course, the presence of multiple transporters with overlapping specificities, indirect effects from loss of a transporter and other phenomenon, have the capacity to complicate such a simplistic scenario and must be taken into account when analyzing data.</p>", "<p>One technology that has the potential to accelerate the functional characterization of genes is Biolog phenotype MicroArrays, a respiration-based assay system that can test up to 2000 phenotypic traits simultaneously ##UREF##0##[8]##. This system uses 96 well plates where each well tests a separate phenotype using a tetrazolium redox dye that produces a color change in response to cellular respiration. The detection system is a Biolog OmniLog incubator/reader that cycles each plate in front of an imaging head every 15 minutes, measuring and recording the color change from reduction of the tetrazolium dye in each well, providing a quantitative kinetic plot of color formation against time. This technology has been used previously to facilitate both characterization of transporters ##REF##15489430##[9]##–##REF##16164569##[11]## and the testing of bioinformatic predictions ##REF##17593909##[12]##,##REF##17573341##[13]##.</p>", "<p>In this study, we have focused on characterizing a collection of knockout mutants of integral cytoplasmic membrane transporter genes in the ecologically and metabolically diverse bacterium <italic>Pseudomonas aeruginosa</italic> using Biolog phenotype MicroArrays in conjunction with more traditional experimental approaches. <italic>P. aeruginosa</italic> is known to be capable of growth on a broad range of substrates such as amino acids, carboxylates, aromatic compounds, but on only a narrow selection of carbohydrates ##REF##5963505##[14]##. Consistent with this, more than 300 predicted cytoplasmic membrane transport systems were identified in <italic>P. aeruginosa</italic> in the initial characterization of its genome ##REF##10984043##[15]##.</p>" ]

[ "<title>Materials and Methods</title>", "<title>Bacterial Strains and Media</title>", "<p>All strains employed in this study are listed in supporting ##SUPPL##2##Table S3##. The 384 <italic>P. aeruginosa</italic> Mini-Tn<italic>5</italic>-Tc^r transporter gene knockout mutants were obtained from PathoGenesis Corporation (Seattle, WA). The liquid minimal growth medium was M9 salts plus 20 mM sodium succinate, 2 µM ferric citrate, 1 mM magnesium sulfate, 0.1 mM calcium chloride. Minimal medium plates were made with 1.5% agar.</p>", "<title>Bioinformatic Analysis</title>", "<p>Functional predictions of membrane transporter genes were assigned by running the complete <italic>P. aeruginosa</italic> gene set through our automated annotation pipeline ##REF##14681414##[6]##, which utilizes a series of BLAST, HMM and COG-based searches in conjunction with other analyses such as transmembrane segment prediction, followed by careful manual curation. Additionally, phylogenetic trees for every transporter protein were generated, and a comparative analysis of the genomic context (i.e., the flanking genes) was undertaken, in order to update the annotation of <italic>P. aeruginosa</italic> transporter genes.</p>", "<title>Phenotype MicroArray Assays</title>", "<p>Strains to be tested were plated on Biolog Universal Growth medium plus Blood agar plates and incubated overnight at 37°C. Cells were swabbed from the plates after overnight growth and suspended in appropriate medium containing Dye Mix C; 100 µL of a 1∶200 dilution of an 85% transmittance suspension of cells were added to each well of the PM plates. Plates 1–8, which test for catabolic pathways for carbon, nitrogen, phosphorus, sulfur, as well as for biosynthetic pathways, and plates 9–10, which test for osmotic/ion and pH effects, were utilized in this study. IF-0 GN Base was used for PM plates 1 and 2. IF-0 GN Base plus 20 mM sodium succinate, pH 7.1, and 2 µM ferric citrate was used for plates 3–8. IF-10 Base was used for plates 9 and 10. Plates were incubated in the OmniLog for 48 hours with readings taken every 15 minutes. Data analysis was performed using Kinetic and Parametric software (Biolog). Phenotypes were determined based on the area difference under the kinetic curve of dye formation between the mutant and wild type. Data points for the entire 48 hours were used for PM1 through PM8, and area differences were mean-centered by plate. PM4 was subdivided into phosphorus utilization and sulfur utilization sections, and these were mean-centered individually. Only data points from the first 24 hours were used for PM9 and PM10, and area differences were not mean-centered. PAK and PAO1 strains were grouped separately, and mean values were determined for each well. Values beyond two standard deviations from the mean were considered for further analysis.</p>", "<title>Minimal Media Growth Assays</title>", "<p>Wild type and mutant strains were grown on minimal medium agar plates containing either the predicted substrate substituted for sodium succinate and ferric citrate when testing carbon sources, or the predicted substrate substituted for ammonium chloride when testing nitrogen sources. Each strain was also grown on minimal media plates without substitutions as a control. Growth phenotypes were determined based on isolated colony sizes, with assays performed in triplicate.</p>", "<title>Quantitative RT-PCR</title>", "<p>The ability of putative substrates to induce transporter gene expression in the wild type PAK and PAO1 strains was assessed by quantitative real-time polymerase chain reaction (qRT-PCR). Overnight liquid minimal medium cultures were grown for an additional two hours in the presence of 0.1% (w/v) of the potential inducer, and RNA was extracted using Trizol and was further purified using the RNeasy Mini Kit (Qiagen). RNA underwent DNase digestion using the DNA-free Turbo DNase Digestion Kit (Ambion). qRT-PCR reactions were carried out using the Superscript III Platinum SYBR Green One-Step qRT-PCR Kit (Invitrogen) in an ABI PRISM 7900(HT). Primers were designed with Primer3 ##REF##10547847##[32]## for products of 100–250 bp in length. As an internal control, RT-PCR was carried out using primers for the amplification of <italic>rplU</italic> which encodes 50S ribosomal protein L21. Reactions were performed at least once from each of three biological replicates. Cycle threshold (CT) values for each reaction were determined with ABI PRISM SDS 2.1 software. Analysis of relative RNA abundance was performed using the ΔΔC_T method (PE Applied Biosystems).</p>" ]

[ "<title>Results</title>", "<title>Transporter Annotation Update</title>", "<p>One of the objectives of this study was to systematically compare the bioinformatic predictions of transporter function with the high throughput functional characterization of transporters. To provide a good basis for these comparisons we decided to update the original annotation of the <italic>P. aeruginosa</italic> membrane transporters (more than seven years has passed since the original genome annotation was performed ##REF##10984043##[15]##), as well as undertake a subjective estimate of our bioinformatic annotation.</p>", "<p>First, the complete <italic>P. aeruginosa</italic> gene set was run through our TransportDB automated annotation pipeline ##REF##14681414##[6]##. Second, all of the transporters identified by this approach or by the original genome annotation were analyzed phylogenetically to investigate their evolutionary histories and to determine whether or not they could clearly be identified as orthologues of known membrane transporters. Third, we analyzed their comparative genome context, looking at their flanking genes in <italic>P. aeruginosa</italic>, but also examining the flanking genes of homologous transporters encoded in other genomes.</p>", "<p>Based on these analyses, a total of 427 predicted membrane transport genes were identified and assigned functions and family groupings (##SUPPL##0##Table S1##). Using this approach, 124 “hypothetical proteins” from the original <italic>P. aeruginosa</italic> PA01 genome annotation ##REF##10984043##[15]##, were annotated as membrane transporters with specific substrates or general functions. These include newly identified paralogues in major transporter families such as ATP binding cassette (ABC), major facilitator superfamily (MFS), and drug/metabolite transporter (DMT) superfamily. In addition, our analyses detected transporters belonging to several new transporter families that were identified after the genome annotation was published, for example, the tricarboxylate transporter (TTT) family ##REF##12562821##[16]## and the aromatic acid exporter (ArAE) family ##REF##15489430##[9]##.</p>", "<p>One of the key objectives of this study was to compare bioinformatic predictions with high throughput functional data, in order to obtain an estimate of the accuracy of our bioinformatic predictions. Based on current knowledge of transporter systems, of the 427 predicted transporters, 16 (4%) were previously experimentally characterized transporters in <italic>P. aeruginosa</italic>, 116 (27%) were clear orthologues to experimentally characterized transporters from other species, such as <italic>E. coli</italic>, and other <italic>Pseudomonas</italic> species, 229 (54%) were transporters whose specificities were predicted with high bioinformatic confidence, while the remaining 66 (15%) had weak bioinformatic evidence supporting their function as transporters.</p>", "<p>Additionally we compared the transporter complement of <italic>P. aeruginosa</italic> PA01 with the recently sequenced clinical isolate <italic>P. aeruginosa</italic> UCBPP-PA14 ##REF##17038190##[17]##. These two strains are very similar in their overall transporter complement with the most notable differences being in respect to additional iron and nickel uptake systems present in the clinical PA14 strain. There was almost no difference in terms of predicted uptake capabilities for amino acids and sugars. The complete list of predicted PA14 transporters from this analysis is available on <ext-link ext-link-type=\"uri\" xlink:href=\"http://www.membranetransport.org\">www.membranetransport.org</ext-link>.</p>", "<title>Biolog Phenotype MicroArray Analysis</title>", "<p>We obtained a subset of 384 <italic>P. aeruginosa</italic> transporter gene knockout mutants from a larger collection of <italic>P. aeruginosa</italic> Mini-Tn<italic>5</italic>-Tc^r gene knockout mutants. This collection of mutants was constructed in two different genetic backgrounds. The isogenic parent of the majority of the Tn5 knockouts was strain PAK, the parent of the remainder of these was the sequenced strain PA01 ##REF##10984043##[15]##,##REF##17114261##[18]##. Knockout mutants from the larger collection have also been used in studies by other research groups ##REF##15375113##[19]##,##REF##16352820##[20]##.</p>", "<p>Biolog phenotype MicroArrays were employed to identify phenotypes for a subset of the transporter gene mutants. In this initial study we chose to focus on characterizing the subset of knockout mutants which were predicted to encode amino acid and sugar phosphotransferase system (PTS) transporters. Amino acid transporters were a focus as there is very good coverage of amino acids and derivatives as carbon and/or nitrogen sources on the Biolog phenotype MicroArrays. Also, previous experimental studies have described arginine/ornithine ##REF##9791103##[21]## and branched chain amino acid transporters ##REF##2120183##[22]## in <italic>P. aeruginosa</italic>. The (PTS) transporter mutants were chosen for further analysis as the bioinformatic predictions for these tended to have a high degree of confidence, in contrast to the amino acid transporters.</p>", "<p>Based on the updated annotation, our mutant collection included 78 knockout mutants of putative amino acid transporter or sugar PTS genes (##TAB##0##Table 1##). This included 46 transporters with a precise specificity and 32 transporters with a more generic prediction. The transposon insertion site of each of these mutants had been sequenced by PathoGenesis Corporation. We additionally confirmed the identities of each of these mutants by PCR with primers specific to each gene and the transposon insert (data not shown).</p>", "<p>We initially optimized conditions for utilizing phenotype MicroArrays with <italic>P. aeruginosa</italic>, including modifying the initial growth medium to diminish production of pigmented compounds by the bacterium that interfered with recording of the colourimetric reaction in each plate by the Biolog OmniLog incubator/reader. Reproducibility of the <italic>P. aeruginosa</italic> PM data was investigated by running multiple replicates of each of the parental strains PAK and PAO1 on separate days. ##FIG##0##Figure 1## displays an overlay of multiple experiments for these strains on the PM5 plate, where it can be seen that, overall, there is a fairly high level of reproducibility. However, reproducibility did vary somewhat between specific wells/compounds, for example, well D5 on the PM5 plate, containing homoserine lactone, showed a substantially higher level of variation than other wells on the same plate (##FIG##0##Figure 1##). A comparison of the substrate utilization capabilities of the two different background strains, PAK and PAO1 indicated that these two strains had highly comparable utilization patterns that differed primarily in terms of different dipeptides.</p>", "<p>The 78 predicted amino acid or carbohydrate transporters were phenotypically screened on PM1-10 plates and compared with their isogenic parental strains. An example of the data generated via the Biolog OmniLog reader is shown in ##FIG##1##Figure 2##, which compares a mutant (PA0220 knockout) with its isogenic parent, PAK, on PM1-10 plates, showing evidence of a significant defect in histamine utilization (well E1 on PM3 plate). Other minor differences in substrate utilization may be noted in this example, whose significance is not readily apparent (##FIG##1##Figure 2##).</p>", "<p>In order to minimize problems arising from the higher variability in some substrate wells, and to make use of the large dataset, the data from all 78 mutant and parent strain comparisons were analyzed concurrently (##FIG##2##Figure 3##). This facilitated identification of statistically significant phenotypic differences. A statistical analysis of this combined data identified phenotypes that showed substantial defects in substrate utilization compared to the parental strain, differing from the mean by more than two standard deviations. Based on this criterion we were able to assign significant phenotypes to 27 of the 78 genes (35%) tested (##TAB##1##Table 2##).</p>", "<p>This pooled analysis allows the amount of noise in the data to be gauged more accurately (as visualized in ##FIG##2##Figure 3## by the central band extending from approximately −5000 to 5000 depending on the particular well/plate) and highlights substrates which show systematic variability. For instance, the wells that show extended vertical lines (examples include, PM4, well E8 methylene diphosphonic acid; PM4, well A7, hypophosphite; PM3, well A8, L-cysteine) represent wells that show greater variability and whose data is probably not reliable for any individual mutant. Another benefit of this type of analysis is that anomalous results are easily detected, such as in one instance where a particular plate was defective for one set of experiments (see the cluster of brown dots highlighted above the noise band in ##FIG##2##Figure 3##, a repeat of this experiment produced a more typical result).</p>", "<title>Confirmatory Assays</title>", "<p>Independent growth studies on minimal media were undertaken to confirm the phenotypes of the mutants with discernable phenotypes from the Biolog PM screen. Growth data for each mutant in minimal media with the appropriate carbon and nitrogen sources are shown in ##TAB##1##Table 2##. These growth assays confirmed the phenotypes of 27 mutants, with some minor discrepancies with regard to the ability of specific substrates to act as both sole carbon and nitrogen source.</p>", "<p>Expression of many transporter genes is regulated in response to their transported substrate ##UREF##1##[23]##. The ability of the putative substrates of each of the 27 transporter genes to induce expression of the respective genes in <italic>P. aeruginosa</italic> PAK was examined by qRT-PCR. Expression of 11 out of the 27 genes was induced by between 5- to 86-fold by their transporter substrates (##TAB##2##Table 3##). Expression of the other 17 genes was observed, but appeared to be constitutive under the conditions tested.</p>", "<title>Transporter Functions</title>", "<p>There are two clusters of sugar PTS genes encoded within the <italic>P. aeruginosa</italic> PAO1 genome ##REF##10943558##[24]##. The first of these, PA3560 and PA3562 encode a putative fructose enzyme IIBC and a fused putative enzyme I/HPr/enzyme IIA fructose. These are encoded in a putative operon with PA3561, a predicted 1-phosphofructokinase. Knockouts in either PA3560 or PA3562 were defective in growth on D-fructose as a sole carbon source, based on both Biolog and growth data (##TAB##1##Table 2##). Expression of both of these genes is induced by D-fructose, supporting the notion that these genes form a dedicated, inducible system for fructose uptake (##TAB##2##Table 3##).</p>", "<p>The second cluster of PTS genes encodes a putative fused enzyme I/HPr/enzyme IIA and a putative enzyme IIBC. These two genes are located at the end of a putative operon encoding glucosamine utilization genes, suggesting N-acetylglucosamine as a potential substrate for the PTS transporter. Knockouts in either PA3760 or PA3761 were defective in growth on N-acetylglucosamine as a sole carbon source, based on both Biolog and growth data (##TAB##1##Table 2##). As for the fructose PTS genes, expression of the N-acetylglucosamine PTS genes was induced by their apparent transported substrate (##TAB##2##Table 3##).</p>", "<p>We were able to identify specific phenotypes for eight predicted amino acid transporter genes that had only generalized bioinformatic predictions of function. PA5504 encodes a predicted membrane component of an ABC amino acid transporter, and is encoded in a putative two gene operon along with an ATP-binding component gene. A PA5504 knockout affected growth on L-histidine as a carbon source (##TAB##1##Table 2##), however its expression was not induced by this compound. A noninducible histidine outer membrane porin OpdC has been previously identified ##REF##16352820##[20]##. However, since it is encoded in a different region of the genome, the interdependence of these two systems is unclear.</p>", "<p>PA1256, PA1257 and PA1260 are all components of a predicted ABC amino acid transporter encoded in a putative operon along with two hypothetical genes. Knockouts in these genes affected growth on hydroxy-L-proline as a carbon source, and expression of all three genes was induced by hydroxy-L-proline (##TAB##1##Table 2## and ##TAB##2##3##). There was no growth defect on L-proline, nor was expression affected by L-proline as an inducer, suggesting specificity for hydroxy-L-proline. Although annotated as a hypothetical protein, the product of PA1255 showed sequence similarity to proline racemases. To our knowledge these would be the first genes identified for a hydroxy-L-proline transporter.</p>", "<p>PA1339, PA1340 and PA1341 all encode components of a predicted ABC amino acid transporter, and are located in a putative three gene operon. All three mutants were defective for utilization of L-glutamic acid as a carbon source, and expression of two of these genes was induced by L-glutamic acid (##TAB##1##Table 2## and ##TAB##2##3##). Additionally, one of these mutants (PA1340) also showed significantly decreased utilization of the related compounds, L-glutamine, D-glutamic acid, and N-acetyl-L-glutamic acid. Examination of the Biolog data indicated that the other two mutants showed reduced utilization of these three compounds that was below the two standard deviation significance cutoff employed, suggesting that this transport system is probably specific for all four of these substrates.</p>", "<p>The last predicted amino acid transporter with a generic specificity was PA0220, and disruption of this gene affected utilization of histamine as a nitrogen source. Expression of PA0220 was strongly induced by histamine (##TAB##2##Table 3##) and as discussed below may be co-encoded with two genes that potentially form a novel histamine utilization pathway. This is particularly interesting as no transporter for histamine has been described at a molecular level previously.</p>", "<p>Fifteen mutants were phenotypically characterized for which precise substrate specificity predictions had been made. PA0783, located in a putative monocistronic operon, was predicted to encode a proline uptake system, and a knockout mutant affected utilization of L-proline as both a carbon and nitrogen source (##TAB##1##Table 2##). PA0783 also displayed L-proline-inducible expression (##TAB##2##Table 3##).</p>", "<p>The predicted glutamate transporter gene PA3176 is the last gene in a three gene operon, also encoding a putative regulator and a homologue of formiminoglutamate hydrolase (PA3175). Disruption of PA3176 had no effect on utilization of glutamate but did disrupt utilization of N-acetyl-L-glutamate (##TAB##1##Table 2##). This presents the possibility that the role of PA3175 is to cleave off the acetyl group to yield L-glutamate. These two genes may therefore represent a novel transporter and catabolic enzyme for utilization of N-acetyl-L-glutamate.</p>", "<p>Genes PA4910 and PA4912 comprise two genes of a five gene operon predicted to be involved in transport of branched chain amino acids. PA4910 encodes a putative ATP-binding protein, while PA4912 encodes a predicted membrane component of this ABC amino acid transporter. Disruption of PA4910 resulted in reduced utilization of D-Alanine (as a nitrogen source), while PA4912 mutants had disrupted D-Alanine and D-Valine utilization (##TAB##1##Table 2##).</p>", "<p>PA5153 and PA5155 are two genes of a four gene operon proposed to encode an arginine/ornithine transport system. Growth experiments indicated that both these mutants showed reduced growth on L-Ornithine. This operon has previously been shown to be arginine-inducible under control of the argR regulator ##REF##15175299##[25]##, however under our experimental conditions using qRT-PCR we did not observe any induction by arginine.</p>", "<p>Genes PA0888, PA0889, PA0890 and PA0892, comprise four of the six genes which make up the <italic>aot</italic> operon (<italic>aotJ</italic>, <italic>aotQ</italic>, <italic>aotM</italic> and <italic>aotP</italic>, respectively), previously shown to be involved in arginine/ornithine uptake ##REF##9791103##[21]##. PA0888 is thought to encode an arginine/ornithine transport system substrate binding protein, PA0889 and PA0890 encode putative permease proteins while PA0892 is predicted to encode the associated ATP-binding component gene. Disruption of each of these four genes resulted in mutants defective in arginine/ornithine uptake, as expected (##TAB##1##Table 2##).</p>", "<p>PA1070, PA1071, PA1072, PA1073, PA1074, which correspond to <italic>braG</italic>, <italic>braF</italic>, <italic>braE</italic>, <italic>braD</italic> and <italic>braC</italic>, comprise a five gene operon which has also been characterized experimentally. The <italic>braC</italic> has been shown to encode the binding protein for branched-chain amino acids ##REF##6767701##[26]##, <italic>braF</italic> and <italic>braG</italic> genes are thought to encode ATP-binding proteins, while <italic>braE</italic> and <italic>braD</italic> encode very hydrophobic proteins ##REF##2120183##[22]##. Complementation experiments have shown that each of these genes is essential for correct functioning of the high affinity branched-chain amino acid transport system encoded by the <italic>bra</italic> operon ##REF##2120184##[27]##. In this work, mutants for each of these genes were found to be defective in Alanine metabolism, while utilization of L-Valine and L-Isoleucine was also altered in some cases (PA1074 and PA1071 were also defective in d-Amino Valeric Acid uptake, ##TAB##1##Table 2##).</p>", "<title>Mapping Phenotype MicroArray Data onto PseudoCyc</title>", "<p>The 27 transporter genes with confirmed phenotypes appear to be responsible for the transport of a total of 16 different substrates. We were interested in the correlation of these transporter substrates with the predicted metabolic network of <italic>P. aeruginosa</italic>. Mapping of these substrates onto the PseudoCyc database ##REF##15608211##[28]## indicated that 15 out of the 16 substrates were starting inputs of predicted metabolic pathways in <italic>P. aeruginosa</italic>.</p>", "<p>The one exception was histamine. Various soil microorganisms are capable of breaking down histamine, typically via histamine oxidase ##REF##16233600##[29]##. However, there are no homologues of this enzyme present in <italic>P. aeruginosa</italic> or any other sequenced pseudomonad. A monoamine oxidase has been reported previously in <italic>P. aeruginosa</italic>, however, histamine was not a substrate for this enzyme, nor was it induced by histamine ##REF##120132##[30]##. The putative histamine transporter gene PA0220 is encoded in a gene cluster with putative aldehyde dehydrogenase and aminotransferase genes of unknown specificity (##FIG##3##Figure 4a##). Speculatively this gene cluster may represent an operon encoding a transporter for histamine, and a two-step catabolic pathway that would convert histamine to imidazole-4-acetate (##FIG##3##Figure 4b##). Imidazole-4-acetate might then conceivably be fed into aspartate metabolism via conversion to imidazolone acetate and subsequently to N-formyl-L-aspartate. Supporting this hypothesis, qRT-PCR analysis indicated that expression of both PA0219 and PA0221 are induced 40-fold by histamine. PA0218 is a putative transcriptional regulator gene that is divergently transcribed from this putative histamine utilization operon and might be responsible for its regulation.</p>" ]

[ "<title>Discussion</title>", "<p>The use of Biolog phenotype MicroArray plates PM1-10 allowed us to determine almost 1000 phenotypic traits for 78 individual knockout mutants. These experiments in <italic>P. aeruginosa</italic> PAO1 identified a total of 136 different compounds that could be utilized as carbon sources and 351 different compounds that could be utilized as nitrogen sources (##SUPPL##1##Table S2##). Comparisons of substrate utilization capabilities of mutants and isogenic parent strains allowed significant phenotypes to be assigned to 27 of the genes tested. For each mutant the phenotype was subsequently confirmed by independent growth studies on minimal media plates.</p>", "<p>Based on these results it is possible to assess the accuracy of the initial bioinformatic predictions. In some cases the experimental data and the bioinformatic predictions essentially correlate. In other cases bioinformatics was able to make only a very general prediction, and experimental data was necessary to clarify the precise substrate specificity of the transporter. Of the 27 successfully characterized genes predicted to encode transporter-related functions, only five annotations were revised based on experimental data. Of these, three genes are part of a single operon (PA1256, PA1257 and PA1260) which was predicted to be involved in amino acid transport and instead found to be involved in Hydroxy-L-Proline transport. The predicted function of PA3176 (gltS) was likewise modified slightly, with bioinformatic predictions suggesting a role as a glutamate/sodium symporter, while experimental data indicated transport of N-acetyl-glutamate. Perhaps the most interesting deviation from the bioinformatic prediction involves PA0220, which, while originally predicted to function in amino acid uptake, based on the Biolog data probably functions in histamine transport. To our knowledge this is the first identification of a histamine transporter gene in a bacterial species.</p>", "<p>In the remaining 22 instances, bioinformatic predictions correlated well with the characterization results (12/27) or provided a correct start point, which experimental findings added to, giving a more precise functional assignment (10/27). These findings indicate that bioinformatic predictions are a valuable starting point in determining gene functions. However, such approaches are not infallible and in many cases can provide only a generalized function prediction, highlighting the need for high throughput experimental approaches to confirm and add detail to phenotypic predictions.</p>", "<p>The Biolog PM system has been proposed to provide a rapid means of evaluating phenotypic predictions for large sets of genes ##UREF##0##[8]##. To date, however, there has only been one published account of a study using the PM system to undertake a large scale functional screen for a specific subset of genes. Zhou et al (2003) used this system to investigate two-component regulatory systems in <italic>Escherichia coli</italic> K-12, screening mutants in 37 different two-component genes for altered growth on PM1-20 plates ##REF##12897016##[31]##. Altered phenotypes were observed for 22 of the 37 different two-component mutants, in most instances these were as expected, however several new phenotypes were revealed. This was a higher rate of mutant phenotype identification than what we observed for <italic>P. aeruginosa</italic> transporter mutants, but this is probably due to the likelihood that regulatory gene mutants have more pleiotropic effects than transporter gene knockouts.</p>", "<p>In this study the Biolog PM system allowed phenotypes to be assigned to 27 of the 78 genes tested. While the Biolog assays performed in this study did not assign specific functions to all putative transporter genes, these experiments did provide indirect evidence of the transporter capacity of <italic>P. aeruginosa</italic>. Both studied strains of <italic>P. aeruginosa</italic> were observed to grow on all tested amino acids indicating that specific or generic transporters are likely to exist for each. Our experimental analyses were successful in characterizing transporters for seven of the standard amino acids. A greater degree of success was obtained with characterization of the two clusters of genes predicted to comprise sugar phosphotransferase system transporters, with their roles confirmed and specific substrates identified for each.</p>", "<p>There are a number of potential reasons why only a proportion of the <italic>P. aeruginosa</italic> amino acid transporters were able to be experimentally identified. We may not have knockout mutants in all <italic>P. aeruginosa</italic> amino acid transporter genes, possibly because they are essential for cell survival or are novel transporter types that were not identified in our bioinformatic screen. For some substrates there may be multiple transporters responsible for their uptake, in which case loss of one of these transport pathways through a knockout mutation may not reduce uptake sufficiently to be detected with these assays. The well-specific variability in the Biolog assays observed for some compounds, such as cysteine, probably precluded the identification of any transporters for those substrates. Another general issue for this type of approach is that the substrates of particular transporter systems may not be one of the 2000 phenotypes currently tested by Biolog plates.</p>", "<p>While there are limitations to what the Biolog PM system can achieve, utilizing this tool allowed many more mutants to be screened (and on vastly more substrates) than would have been possible with any other current approach. This has allowed the characterization of a total of 27 transporter genes in a single study, a considerable undertaking when most characterization studies to date have tackled only one or two such operons at a time. One notable exception is the recent study by Tamber et al (2006), which aimed to characterize 17 genes predicted to encode porins involved in nutrient uptake in <italic>P. aeruginosa</italic>\n##REF##16352820##[20]##. This study characterized knockout mutants, defective in putative <italic>opr</italic>D family genes, obtained from the same mutant collection we utilized. In this case standard growth experiments, rather than Phenotype MicroArrays, were used to determine the associated phenotypes, providing an interesting parallel to the characterization approach presented here. The authors of this study were able to confirm the predicted functions of six of these 17 <italic>opr</italic>D family genes, a similar proportion to what was achieved in this study while a smaller overall number.</p>", "<p>This comparison highlights another advantage of the Biolog system. Standard plate growth assays can be set up easily and cheaply to confirm gene functions where reasonably specific bioinformatic predictions of function are available, limiting the number of growth substrates to be tested. However, this approach is not suitable where there are no bioinformatic “clues” regarding specificity of the putative transporter, an issue overcome with the Biolog system, which allows rapid screening of thousands of potential compounds simultaneously. Such comprehensive screening also makes it possible to detect novel transporters, such as the histamine transporter detected in this study, which would have been unlikely to have been uncovered by more directed growth assays.</p>", "<p>In this study, the Biolog PM system was used to screen a significant number of mutants for changes in growth on a very large number of substrates. One of the most significant advantages of this approach was the capacity to rapidly characterize transporters for which bioinformatic predictions provide only a general indication of function, a feat which previously would have required a much greater expenditure of time and research effort.</p>" ]

[]

[ "<p>Conceived and designed the experiments: ITP. Performed the experiments: DAJ KP JC. Analyzed the data: DAJ SGT QR ITP. Wrote the paper: DAJ SGT ITP.</p>", "<p>The deluge of data generated by genome sequencing has led to an increasing reliance on bioinformatic predictions, since the traditional experimental approach of characterizing gene function one at a time cannot possibly keep pace with the sequence-based discovery of novel genes. We have utilized Biolog phenotype MicroArrays to identify phenotypes of gene knockout mutants in the opportunistic pathogen and versatile soil bacterium <italic>Pseudomonas aeruginosa</italic> in a relatively high-throughput fashion. Seventy-eight <italic>P. aeruginosa</italic> mutants defective in predicted sugar and amino acid membrane transporter genes were screened and clear phenotypes were identified for 27 of these. In all cases, these phenotypes were confirmed by independent growth assays on minimal media. Using qRT-PCR, we demonstrate that the expression levels of 11 of these transporter genes were induced from 4- to 90-fold by their substrates identified via phenotype analysis. Overall, the experimental data showed the bioinformatic predictions to be largely correct in 22 out of 27 cases, and led to the identification of novel transporter genes and a potentially new histamine catabolic pathway. Thus, rapid phenotype identification assays are an invaluable tool for confirming and extending bioinformatic predictions.</p>", "<title>Author Summary</title>", "<p>Genome sequencing has led to the identification of literally millions of new genes, for which there is no experimental evidence concerning their function. This limits our knowledge of these genes to computational predictions; however, the accuracy of such bioinformatic predictions is essentially unknown. We have focused on investigating the accuracy of bioinformatic predictions for a specific class of genes—those encoding membrane transporters. Our approach used Biolog phenotype MicroArrays to screen transporter gene knockout mutants in the bacterium <italic>P. aeruginosa</italic> for the ability to metabolize hundreds of different compounds. We were able to identify functions for 27 out of 78 genes, all of which were confirmed through independent growth assays. For 80% of these genes, the computationally predicted and experimentally determined functions were either identical or generically similar. Additionally, this led to the discovery of entirely new types of transporters and a novel potential histamine metabolic pathway.</p>" ]

[ "<title>Supporting Information</title>" ]

[ "<p>We would like to thank B. Bochner and J. Carlson from Biolog Corporation, Hayward, CA, for technical assistance and B. Bochner for critical reading of the manuscript.</p>" ]

[ "<fig id=\"pgen-1000211-g001\" position=\"float\"><object-id pub-id-type=\"doi\">10.1371/journal.pgen.1000211.g001</object-id><label>Figure 1</label><caption><title>Reproducibility of Biolog Phenotype MicroArray data with wild type <italic>P. aeruginosa</italic> strains, PA01, and PAK.</title><p>An overlay of data from plate PM5, representing multiple replicates of assays with each of the parental strains PAK and PAO1, illustrates the degree of reproducibility possible with the plate assay system. Note that for most substrates there is minimal variability on the PM5 plate, with the exception of well D5, containing the substrate homoserine lactone.</p></caption></fig>", "<fig id=\"pgen-1000211-g002\" position=\"float\"><object-id pub-id-type=\"doi\">10.1371/journal.pgen.1000211.g002</object-id><label>Figure 2</label><caption><title>Data for Biolog Phenotype MicroArray PM1-10 plates comparing one mutant (PA0220 knockout) with its isogenic parent, PAK.</title><p>This mutant shows evidence of a significant defect in histamine utilization (well E1 on PM3 plate, circled). Other minor differences in substrate utilization may be noted in this example (reduced utilization of the mutant compared to parent strain indicated by red areas in the growth curve for each substrate, increased utilization in the mutant by green areas).</p></caption></fig>", "<fig id=\"pgen-1000211-g003\" position=\"float\"><object-id pub-id-type=\"doi\">10.1371/journal.pgen.1000211.g003</object-id><label>Figure 3</label><caption><title>Graphical display of substrate utilization data from Biolog plates PM1-10 for all mutant and parent strain comparisons.</title><p>Each point on the horizontal axis represents a particular well on a particular plate, the vertical axis represents the area difference (in arbitrary units) under the kinetic curve of dye formation between the mutant and wild type over a 48 hour time period (see <xref ref-type=\"sec\" rid=\"s4\">Materials and Methods</xref>). The black arrows highlight some examples of wells where substrate utilization showed a high degree of variability across the strain tested. The method of pooled analysis used here draws attention to anomalous data, highlighting potentially problematic plates or substrates, which may require further confirmatory assays. The brown arrow highlights the outlier data points produced by a faulty plate, retesting produced more typical data for this mutant.</p></caption></fig>", "<fig id=\"pgen-1000211-g004\" position=\"float\"><object-id pub-id-type=\"doi\">10.1371/journal.pgen.1000211.g004</object-id><label>Figure 4</label><caption><title>Putative histamine uptake and utilization operon in <italic>P. aeruginosa</italic>.</title><p>A) Diagrammatic representation of the genetic vicinity of the putative histamine transporter gene PA0220. Each gene is labeled with its predicted function, and the predicted operon structure is indicated with grey shading. B) Illustration of the proposed biochemical pathway for histamine utilization encoded for by this putative operon.</p></caption></fig>" ]

[ "<table-wrap id=\"pgen-1000211-t001\" position=\"float\"><object-id pub-id-type=\"doi\">10.1371/journal.pgen.1000211.t001</object-id><label>Table 1</label><caption><title>Details of knockout mutants of <italic>Pseudomonas aeruginosa</italic> putative transporter genes.</title></caption><alternatives><table frame=\"hsides\" rules=\"groups\"><colgroup span=\"1\"><col align=\"left\" span=\"1\"/><col align=\"center\" span=\"1\"/><col align=\"center\" span=\"1\"/></colgroup><thead><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">Gene</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">Host Strain</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">Predicted Function</td></tr></thead><tbody><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA0129</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">PAK</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">4-amino butyrate APC family transporter (<italic>gabP</italic>)</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA0220</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">PAK</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">amino acid APC family transporter</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA0313</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">PAK</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">amino acid ABC transporter membrane protein</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA0314</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">PAK</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">amino acid ABC transporter periplasmic binding protein</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA0322</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">PAK</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">cationic amino acid APC family transporter</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA0783</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">PAK</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">proline/sodium transporter</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA0789</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">PAK</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">proline APC family transporter</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA0866</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">PAK</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">aromatic amino acid APC family transporter</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA0888</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">PAK pili-</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">arginine/ornithine ABC transporter periplasmic binding protein (<italic>AotJ</italic>)</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA0889</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA01</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">arginine/ornithine ABC transporter membrane protein (<italic>AotQ</italic>)</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA0890</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">PAK</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">arginine/ornithine ABC transporter membrane protein (<italic>AotM</italic>)</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA0892</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">PAK</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">arginine/ornithine ABC transporter ATP binding protein (<italic>AotP</italic>)</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA1070</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">PAK</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">branched chain amino acid ABC transporter ATP binding protein (<italic>BraG</italic>)</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA1071</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">PAK</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">branched chain amino acid ABC transporter ATP binding protein (<italic>BraF</italic>)</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA1072</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">PAK</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">branched chain amino acid ABC transporter membrane protein (<italic>BraE</italic>)</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA1073</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">PAK</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">branched chain amino acid ABC transporter membrane protein (<italic>BraD</italic>)</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA1074</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">PAK</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">branched chain amino acid ABC transporter periplasmic binding protein (<italic>BraC</italic>)</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA1147</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">PAK</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">amino acid APC family transporter</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA1194</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">PAK</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">arginine/ornithine APC family transporter</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA1256</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">PAK</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">amino acid ABC transporter ATP binding protein</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA1257</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA01</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">amino acid ABC transporter membrane protein</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA1258</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">PAK</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">amino acid ABC transporter membrane protein</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA1260</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">PAK</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">amino acid ABC transporter periplasmic binding protein</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA1339</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">PAK</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">amino acid ABC transporter ATP binding protein</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA1340</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">PAK</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">amino acid ABC transporter membrane protein</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA1341</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">PAK</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">amino acid ABC transporter membrane protein</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA1418</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">PAK</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">sodium∶solute symporter</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA1485</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA01</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">amino acid APC family transporter</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA1590</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">PAK</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">branched chain amino acid sodium ion symporter (<italic>BraB</italic>)</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA1819</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">PAK</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">amino acid APC family transporter</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA1958</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">PAK</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">nicotinamide mononucleotide transporter</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA1971</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA01</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">branched chain amino acid∶cation symporter (<italic>BraZ</italic>)</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA2041</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">PAK</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">amino acid APC family transporter</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA2079</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA01</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">amino acid APC family transporter</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA2202</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">PAK</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">amino acid ABC transporter membrane protein</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA2203</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">PAK</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">amino acid ABC transporter membrane protein</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA2204</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA01</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">amino acid ABC transporter periplasmic binding protein</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA2252</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA01</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">alanine/glycine/sodium symporter</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA2307</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">PAK</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">nitrate/sulfonate/taurine ABC transporter membrane protein</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA2533</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">PAK</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">alanine/glycine/sodium symporter</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA2923</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">PAK</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">histidine ABC transporter periplasmic binding protein (<italic>HisJ</italic>)</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA2924</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">PAK</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">histidine ABC transporter membrane protein (<italic>HisQ</italic>)</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA2925</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">PAK</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">histidine ABC transporter membrane protein (<italic>HisM</italic>)</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA2926</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">PAK</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">histidine ABC transporter ATP binding protein (<italic>HisP</italic>)</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA3000</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">PAK</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">aromatic amino acid APC family transporter (<italic>AroP1</italic>)</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA3176</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA01</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">glutamate/sodium ion symporter (<italic>GltS</italic>)</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA3560</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">PAK</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">phosphotransferase system transporter fructose-specific IIBC component (<italic>FruA</italic>)</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA3562</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">PAK</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">phosphotransferase system transporter enzyme I (<italic>FruI</italic>)</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA3597</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">PAK</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">amino acid APC family transporter</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA3641</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">PAK</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">alanine/sodium symporter</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA3760</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">PAK</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">N-Acetyl-D-Glucosamine acphosphotransferase system transporter</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA3761</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">PAK</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">N-Acetyl-D-Glucosamine phosphotransferase system transporter</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA3865</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">PAK</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">arginine/ornithine ABC transporter periplasmic binding protein</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA3889</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">PAK</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">glycine ABC transporter periplasmic binding protein</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA4023</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">PAK</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">ethanolamine APC family transporter</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA4072</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">PAK</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">amino acid APC family transporter</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA4096</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">PAK</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">MFS transporter</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA4192</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">PAK</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">amino acid ABC transporter ATP binding protein</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA4193</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">PAK</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">amino acid ABC transporter membrane protein</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA4194</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">PAK</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">amino acid ABC transporter membrane protein</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA4195</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">PAK</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">amino acid ABC transporter periplasmic binding protein</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA4233</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA01</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">MFS transporter</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA4628</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">PAK</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">lysine APC family transporter</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA4804</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">PAK</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">amino acid APC family transporter</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA4909</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA01</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">branched chain amino acid ABC transporter ATP binding protein</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA4910</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">PAK</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">branched chain amino acid ABC transporter ATP binding protein</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA4911</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA01</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">branched chain amino acid ABC transporter membrane protein</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA4912</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">PAK</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">branched chain amino acid ABC transporter membrane protein</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA4981</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA01</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">amino acid APC family transporter</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA5074</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">PAK</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">glutamine ABC transporter ATP binding protein</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA5075</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA01</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">glutamine ABC transporter membrane protein</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA5076</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">PAK</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">glutamine ABC transporter periplasmic binding protein</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA5097</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">PAK</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">proline APC family transporter</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA5153</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">PAK</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">amino acid (lysine/arginine/ornithine/histidine/octopine) ABC transporter periplasmic binding protein</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA5155</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">PAK</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">amino acid (lysine/arginine/ornithine/histidine/octopine) ABC transporter membrane protein</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA5170</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">PAK</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">arginine/ornithine APC family antiporter (<italic>ArcD</italic>)</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA5504</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA01</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">D-methionine ABC transporter membrane protein</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA5510</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">PAK</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">amino acid APC family transporter</td></tr></tbody></table></alternatives></table-wrap>", "<table-wrap id=\"pgen-1000211-t002\" position=\"float\"><object-id pub-id-type=\"doi\">10.1371/journal.pgen.1000211.t002</object-id><label>Table 2</label><caption><title>Phenotypes of <italic>P. aeruginosa</italic> transport mutants, based on Biolog phenotype MicroArray and minimal media growth assays.</title></caption><alternatives><table frame=\"hsides\" rules=\"groups\"><colgroup span=\"1\"><col align=\"left\" span=\"1\"/><col align=\"center\" span=\"1\"/><col align=\"center\" span=\"1\"/><col align=\"center\" span=\"1\"/></colgroup><thead><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">Gene</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">Predicted function</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">Biolog Phenotypes<xref ref-type=\"table-fn\" rid=\"nt101\">a</xref>\n</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">Plate growth<xref ref-type=\"table-fn\" rid=\"nt102\">b</xref>\n</td></tr></thead><tbody><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA0220</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">amino acid APC family transporter</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">Histamine – N</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">-</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA0783</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">proline/sodium transporter</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">L-Proline - C+N</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">-</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA0888</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">arginine/ornithine ABC transporter periplasmic binding protein (<italic>AotJ</italic>)</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">L-Arginine – C</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">-</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\"/><td align=\"left\" rowspan=\"1\" colspan=\"1\"/><td align=\"left\" rowspan=\"1\" colspan=\"1\">L-Ornithine – C</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">-</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA0889</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">arginine/ornithine ABC transporter membrane protein (<italic>AotQ</italic>)</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">L-Arginine – C</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+/−</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\"/><td align=\"left\" rowspan=\"1\" colspan=\"1\"/><td align=\"left\" rowspan=\"1\" colspan=\"1\">L-Ornithine – C</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+/−</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA0890</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">arginine/ornithine ABC transporter membrane protein (<italic>AotM</italic>)</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">L-Arginine – C</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">-</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\"/><td align=\"left\" rowspan=\"1\" colspan=\"1\"/><td align=\"left\" rowspan=\"1\" colspan=\"1\">L-Ornithine – C</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">-</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA0892</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">arginine/ornithine ABC transporter ATP binding protein (<italic>AotP</italic>)</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">L-Arginine – C</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">-</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\"/><td align=\"left\" rowspan=\"1\" colspan=\"1\"/><td align=\"left\" rowspan=\"1\" colspan=\"1\">L-Ornithine – C</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">-</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA1070</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">branched chain amino acid ABC transporter ATP binding protein (<italic>BraG</italic>)</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">D-Alanine – C</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+/−</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\"/><td align=\"left\" rowspan=\"1\" colspan=\"1\"/><td align=\"left\" rowspan=\"1\" colspan=\"1\">L-Alanine – C</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+/−</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA1071</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">branched chain amino acid ABC transporter ATP binding protein (<italic>BraF</italic>)</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">D-Alanine – C</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+/−</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\"/><td align=\"left\" rowspan=\"1\" colspan=\"1\"/><td align=\"left\" rowspan=\"1\" colspan=\"1\">L-Alanine – C</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+/−</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\"/><td align=\"left\" rowspan=\"1\" colspan=\"1\"/><td align=\"left\" rowspan=\"1\" colspan=\"1\">L-Isoleucine – C</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+/−</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\"/><td align=\"left\" rowspan=\"1\" colspan=\"1\"/><td align=\"left\" rowspan=\"1\" colspan=\"1\">L-Valine – N</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+/−</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\"/><td align=\"left\" rowspan=\"1\" colspan=\"1\"/><td align=\"left\" rowspan=\"1\" colspan=\"1\">d-Amino Valeric Acid – C</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+/−</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA1072</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">branched chain amino acid ABC transporter membrane protein (<italic>BraE</italic>)</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">D-Alanine – C</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+/−</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA1073</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">branched chain amino acid ABC transporter membrane protein (<italic>BraD</italic>)</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">D-Alanine – C</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+/−</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\"/><td align=\"left\" rowspan=\"1\" colspan=\"1\"/><td align=\"left\" rowspan=\"1\" colspan=\"1\">L-Alanine – C</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">-</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\"/><td align=\"left\" rowspan=\"1\" colspan=\"1\"/><td align=\"left\" rowspan=\"1\" colspan=\"1\">L-Isoleucine – C</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+/−</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA1074</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">branched chain amino acid ABC transporter periplasmic binding protein (<italic>BraC</italic>)</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">D-Alanine – C</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+/−</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\"/><td align=\"left\" rowspan=\"1\" colspan=\"1\"/><td align=\"left\" rowspan=\"1\" colspan=\"1\">L-Alanine – C</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">-</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\"/><td align=\"left\" rowspan=\"1\" colspan=\"1\"/><td align=\"left\" rowspan=\"1\" colspan=\"1\">L-Isoleucine – C</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+/−</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\"/><td align=\"left\" rowspan=\"1\" colspan=\"1\"/><td align=\"left\" rowspan=\"1\" colspan=\"1\">L-Valine – N</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">-</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA1256</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">amino acid ABC transporter ATP binding protein</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">Hydroxy-L-Proline -C</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+/−</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA1257</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">amino acid ABC transporter membrane protein</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">Hydroxy-L-Proline -C</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">-</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA1260</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">amino acid ABC transporter periplasmic binding protein</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">Hydroxy-L-Proline -C</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+/−</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA1339</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">amino acid ABC transporter ATP binding protein</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">L-Glutamic Acid - C</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+/−</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA1340</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">amino acid ABC transporter membrane protein</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">L-Glutamic Acid - C</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+/−</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\"/><td align=\"left\" rowspan=\"1\" colspan=\"1\"/><td align=\"left\" rowspan=\"1\" colspan=\"1\">L-Glutamine – C</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+/−</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\"/><td align=\"left\" rowspan=\"1\" colspan=\"1\"/><td align=\"left\" rowspan=\"1\" colspan=\"1\">L-Glutamic Acid - N</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+/−</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\"/><td align=\"left\" rowspan=\"1\" colspan=\"1\"/><td align=\"left\" rowspan=\"1\" colspan=\"1\">D-Glutamic Acid - N</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">-</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\"/><td align=\"left\" rowspan=\"1\" colspan=\"1\"/><td align=\"left\" rowspan=\"1\" colspan=\"1\">N-Acetyl-L-Glutamic Acid – N</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+/−</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA1341</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">amino acid ABC transporter membrane protein</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">L-Glutamic Acid -C</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+/−</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA3176</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">glutamate/sodium ion symporter (<italic>GltS</italic>)</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">N-Acetyl-L-Glutamic Acid - N</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+/−</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA3560</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">phosphotransferase system transporter fructose-specific IIBC component (<italic>FruA</italic>)</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">D-Fructose – C</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">-</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA3562</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">phosphotransferase system transporter enzyme I (<italic>FruI</italic>)</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">D-Fructose – C</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">-</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA3760</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">N-Acetyl-D-Glucosamine phosphotransferase system transporter</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">N-Acetyl-D-Glucosamine – N</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">-</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA3761</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">N-Acetyl-D-Glucosamine phosphotransferase system transporter</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">N-Acetyl-D-Glucosamine – C+N</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">-</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA4910</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">branched chain amino acid ABC transporter ATP binding protein</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">D-Alanine – N</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+/−</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA4912</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">branched chain amino acid ABC transporter membrane protein</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">D-Alanine – N</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+/−</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\"/><td align=\"left\" rowspan=\"1\" colspan=\"1\"/><td align=\"left\" rowspan=\"1\" colspan=\"1\">D-Valine – N</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+/−</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA5153</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">amino acid (lysine/arginine/ornithine/histidine/octopine) ABC transporter periplasmic binding protein</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">L-Ornithine – C</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+/−</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA5155</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">amino acid (lysine/arginine/ornithine/histidine/octopine) ABC transporter membrane protein</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">L-Ornithine – C</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">-</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA5504</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">D-methionine ABC transporter membrane protein</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">L-Histidine – C</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">+/−</td></tr></tbody></table></alternatives></table-wrap>", "<table-wrap id=\"pgen-1000211-t003\" position=\"float\"><object-id pub-id-type=\"doi\">10.1371/journal.pgen.1000211.t003</object-id><label>Table 3</label><caption><title>Putative transporter mutants found to be induced in qRT-PCR experiments.</title></caption><alternatives><table frame=\"hsides\" rules=\"groups\"><colgroup span=\"1\"><col align=\"left\" span=\"1\"/><col align=\"center\" span=\"1\"/><col align=\"center\" span=\"1\"/></colgroup><thead><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">Gene</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">qRT-PCR inducer<xref ref-type=\"table-fn\" rid=\"nt103\">a</xref>\n</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">Ratio<xref ref-type=\"table-fn\" rid=\"nt104\">b</xref>\n</td></tr></thead><tbody><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA0220</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">Histamine</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">35.8</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA0783</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">Proline</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">7.0</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA1256</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">Hydroxy-L-Proline</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">86.3</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA1257</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">Hydroxy-L-Proline</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">15.5</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA1260</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">Hydroxy-L-Proline</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">43.9</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA1339</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">L-Glutamic Acid</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">8.0</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA1340</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">L-Glutamic Acid</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">9.4</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA3560</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">D-Fructose</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">4.3</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA3562</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">D-Fructose</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">27.1</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA3760</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">N-Acetyl-D-Glucosamine</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">5.2</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">PA3761</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">N-Acetyl-D-Glucosamine</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">5.1</td></tr></tbody></table></alternatives></table-wrap>" ]

[]

[]

[]

[]

[]

[ "<supplementary-material content-type=\"local-data\" id=\"pgen.1000211.s001\"><label>Table S1</label><caption><p>List of bioinformatic predictions for possible transporters.</p><p>(0.75 MB DOC)</p></caption></supplementary-material>", "<supplementary-material content-type=\"local-data\" id=\"pgen.1000211.s002\"><label>Table S2</label><caption><p>\n<italic>Pseudomonas aeruginosa</italic> PA01 substrate utilization profiles from Biolog phenotype MicroArrays.</p><p>(1.42 MB DOC)</p></caption></supplementary-material>", "<supplementary-material content-type=\"local-data\" id=\"pgen.1000211.s003\"><label>Table S3</label><caption><p>Strain list.</p><p>(0.49 MB DOC)</p></caption></supplementary-material>" ]

[ "<table-wrap-foot><fn id=\"nt101\"><label>a</label><p>Knockout mutants with a significant reduction in substrate utilization as assayed by Biolog phenotype MicroArray plates. Substrates used as carbon sources are labeled with C, substrates used as nitrogen sources are labeled with N.</p></fn><fn id=\"nt102\"><label>b</label><p>Growth on minimal media containing specified sole carbon or nitrogen source, - indicates no growth or pinpoint colonies of mutant, +/− indicates growth of mutant was reduced compared to wild type.</p></fn></table-wrap-foot>", "<table-wrap-foot><fn id=\"nt103\"><label>a</label><p>The compounds used to induce expression were the probable transported substrates as determined by Biolog phenotype and plate growth assays.</p></fn><fn id=\"nt104\"><label>b</label><p>The ratio of gene expression for each knockout mutant compared to its isogenic parent strain, under induced conditions. The differences were significant, with p-values of &lt;0.05 recorded for each (the highest was 0.016).</p></fn></table-wrap-foot>", "<fn-group><fn fn-type=\"COI-statement\"><p>The authors have declared that no competing interests exist.</p></fn><fn fn-type=\"financial-disclosure\"><p>ITP is supported by a Life Science Research Award, provided by the New South Wales Office of Science and Medical Research. This funding body had no further role in any aspect of this study.</p></fn></fn-group>" ]

[ "<graphic id=\"pgen-1000211-t001-1\" xlink:href=\"pgen.1000211.t001\"/>", "<graphic xlink:href=\"pgen.1000211.g001\"/>", "<graphic xlink:href=\"pgen.1000211.g002\"/>", "<graphic xlink:href=\"pgen.1000211.g003\"/>", "<graphic id=\"pgen-1000211-t002-2\" xlink:href=\"pgen.1000211.t002\"/>", "<graphic id=\"pgen-1000211-t003-3\" xlink:href=\"pgen.1000211.t003\"/>", "<graphic xlink:href=\"pgen.1000211.g004\"/>" ]

[ "<media xlink:href=\"pgen.1000211.s001.doc\"><caption><p>Click here for additional data file.</p></caption></media>", "<media xlink:href=\"pgen.1000211.s002.doc\"><caption><p>Click here for additional data file.</p></caption></media>", "<media xlink:href=\"pgen.1000211.s003.doc\"><caption><p>Click here for additional data file.</p></caption></media>" ]

{ "acronym": [], "definition": [] }

[ "<title>Introduction</title>", "<title>Hierarchical Biological Networks</title>", "<p>The analysis of biological networks presents an intriguing challenge, due to the complex, non-random organization of these systems and the diverse dynamic behaviors that they express. The topology of several biological networks has been shown to be based on a scale-free degree distribution, which implies the existence of highly connected network hubs ##REF##10521342##[1]##,##UREF##0##[2]##. Biological systems were also found to be organized in network modules ##REF##12060727##[3]##,##UREF##1##[4]##, or to contain characteristic circuits (motifs) that do not occur as frequently in other types of networks ##UREF##2##[5]##. Hub nodes, which have been identified in several biological networks, such as protein-protein interaction networks or metabolic networks, may serve as central distributing elements or linkage point for many regions of a network ##UREF##0##[2]##,##UREF##3##[6]##,##REF##11333967##[7]##. Such hubs might also be present in neural systems networks ##UREF##4##[8]##. A hub, for our purposes, can either be a node with a high degree or with a high centrality (i.e. with many shortest paths between nodes passing through). For our purposes, the latter definition is dynamically more relevant. Modules or network clusters, which are characterized by a higher frequency or density of connections within than between node clusters ##REF##8577827##[9]## have been identified in biological metabolic networks ##UREF##5##[10]##,##REF##15729348##[11]##, as well as neural networks at the cellular level ##UREF##6##[12]## or the systems level ##REF##10703046##[13]##. These modules often represent a specific function, e.g. a specific synthesis pathway in a metabolic reaction network ##REF##12202830##[14]##, and may shape the functional interactions within the networks at different scales ##REF##10703047##[15]##–##REF##17548818##[17]##. It has also been argued that motifs may represent specific functional circuits ##REF##15001784##[18]##–##UREF##8##[20]##.</p>", "<p>In addition to the mentioned features, the organization of biological systems is often described as hierarchical. However, no formal definition of hierarchical topology appears to exist. Typical descriptions of hierarchical organization use a modules-within-modules view ##UREF##5##[10]##,##UREF##9##[21]##, others focus on the coexistence of modules and central (hub) nodes ##REF##15729348##[11]##,##REF##15190252##[22]## or relate the concept of hierarchy to fractality ##REF##16757100##[23]##. The distinction between hubs which organize modules around them and hubs which connect modules on a higher topological level has been productive for understanding the functional roles of these hub categories in various empirical networks ##REF##15729348##[11]##,##REF##15190252##[22]##,##UREF##10##[24]##. Note: (1) In ##UREF##5##[10]## the algorithm for generating modules within modules, leading to a hierarchical network, also produces a hierarchy of hubs in the network; (2) it is not immediately clear, whether the fractal graphs discussed in ##REF##16757100##[23]## are also “fractal” from the perspective of the box-counting formalism developed in ##REF##15674285##[25]##,##REF##16486532##[26]##. Particularly the latter concept of fractality has interesting implications for the organization of dynamic processes on the graph ##REF##17470793##[27]##.</p>", "<p>In the present paper, we attempt to summarize current topological concepts, condense the spectrum of different network arrangements into a few salient topological features and, using a simple three-state model of excitable dynamics on graphs, study how these topological features organize dynamic behavior. While this approach and our findings are valid for a wide range of networks, we investigate the question and the implications of our findings particularly in the context of neural networks, which most clearly express diverse patterns of excitable dynamics.</p>", "<title>Models of Network Topology</title>", "<p>From a combination of modular and hub features, various types of network topologies can arise. Classical Erdös-Rényi (ER) random graphs do not contain hubs or modules and may thus serve as a general null model. Scale-free Barabási-Albert (BA) graphs, on the other hand, contain only hubs and no modules. Within such graphs, projections from the hubs can reach many network regions, and the hub nodes thus have a more privileged role than nodes with fewer connections and a more restricted reach. On the other hand, networks that do not contain hubs, but are modular, may arise from linking many distributed, dense clusters with a small number of inter-cluster connections. Such clusters could exist at different levels (representing clusters of sub-clusters of sub-sub-clusters ##UREF##9##[21]##), resulting in a hierarchical network organization, which has recently been termed “fractal” ##REF##16757100##[23]##. Finally, networks may be modular and also contain hubs, which are either contained within the modules serving as local hubs, or may form global hubs that integrated network modules at different scales of organization ##UREF##5##[10]##,##REF##12202830##[14]##,##UREF##10##[24]##,##UREF##11##[28]##. The two latter networks combine features of scale-free and modular networks. ##FIG##0##Figure 1## summarizes the topology of modular and hub features and their combination in complex networks. While all feature combinations provide networks of complex organization, we are particular interested in the hierarchical networks shown in the last row of ##FIG##0##Figure 1##, which form modular arrangements, with or without hubs, at different network scales.</p>", "<title>Models of Network Dynamics</title>", "<p>For discussing the link between network topology and dynamics we use a simple three-state model of an excitable medium. The model consists of three discrete states for each node (susceptible <italic>S</italic>, excited <italic>E</italic>, refractory <italic>R</italic>), which are updated synchronously in discrete time steps according to the following rules: (1) A susceptible node becomes an excited node, if there is at least one excitation in its direct neighborhood. If not, spontaneous firing occurs with the probability <italic>f</italic>, which is the rate of spontaneous excitation; (2) an excited node enters the refractory state; (3) a node regenerates (<italic>R</italic>→<italic>S</italic>) with the recovery probability <italic>p</italic> (the inverse of which is the average refractory time of a node). This minimal model of an excitable system has a rich history in biological modeling. It has been first introduced in a simpler variant under the name “forest fire model” ##UREF##12##[29]## and subsequently expanded by Drossel and Schwabl ##REF##10046273##[30]## who also introduced the rate of spontaneous excitations (the “lightning probability” in their terminology). In this form it was originally applied on regular architectures in studies of self-organized criticality. Other variants of three-state excitable dynamics have been used to describe epidemic spreading ##UREF##13##[31]##–##UREF##16##[34]##. As discussed previously ##UREF##17##[35]##,##UREF##18##[36]##, this general model can readily be implemented on arbitrary network architectures. It has been shown that short-cuts inserted into a regular (e.g., ring-like) architecture can mimic the dynamic effect of spontaneous excitations ##UREF##17##[35]##. Using a similar model setup we have recently shown ##UREF##18##[36]## that the distribution pattern of excitations is regulated by the connectivity as well as by the rate of spontaneous excitations. An increase in each of these two quantities leads to a sudden increase in the excitation density accompanied by a drastic change in the distribution pattern from a collective, synchronous firing of a large number of nodes in the graph (spikes) to more local, long-lasting and propagating excitation patterns (bursts). Further studies on the activity of integrate-and-fire neurons in the classical small-world model from ##REF##9623998##[37]## also revealed a distinct dependency of the dynamic behavior on the connectivity of the system ##REF##15169447##[38]##.</p>", "<p>Here, we take this investigation one step further by analyzing which topological properties determine the distribution patterns of excitations. In order to study these patterns, we consider the individual time series of all nodes and for each pair of nodes (<italic>s</italic>,<italic>t</italic>) compute the number <italic>C</italic> = <italic>C_st</italic> of simultaneous firing events. When applied to the whole network the resulting matrix <italic>C</italic> essentially represents the distribution pattern of excitations which we now can compare with a corresponding distribution pattern of some topological property.</p>", "<p>Examining hub and modular aspects of topology separately we first investigate which of them explains best the observed pattern of simultaneous firing events. In particular, we show that in different parameter regimes (characterized by the rate of spontaneous excitations) different topological properties determine the observed synchronization patterns. Moreover, we show that small systematic changes in the graph architecture, designed to enhance or decrease the selected topological property, are reflected in the dynamics. In a second step, we extend our study to hierarchically structured artificial graphs and then to biological networks, in order to demonstrate that the distribution patterns of excitations change dramatically when both properties are represented to different degrees in the respective graphs. Finally, we summarize our results and discuss limitations of the present approach, and extend our observations to describe general principles of pattern formation on graphs.</p>" ]

[ "<title>Methods</title>", "<title>Simulated Network Topologies and Network Modification</title>", "<title>Overview</title>", "<p>This work is based on a variety of network architectures, topological parameters, and dynamic techniques. The basic artificial network types and methods presented in the first part best suit our objective to rule out the individual impact of modularity and hub dominance on dynamic pattern formation. The second part deals with hierarchically structured networks and with real-world topologies, that is, biological neural networks, which will be analyzed concerning both topological properties. The third part contains the analysis tools to probe the networks topologically and dynamically.</p>", "<title>Scale-free network</title>", "<p>This basic network type is constructed via preferential attachment following the Barabási-Albert (BA) model ##REF##10521342##[1]##. The generation of this network starts with a small set (we use <italic>n</italic>\n₀ = 2) of completely connected nodes. Then, new nodes are added to the graph and connected with <italic>m_A</italic> edges (we use <italic>m_A</italic> = 1.25) preferentially to the nodes with the highest degrees (details on the BA algorithm for non-integer values of <italic>m_A</italic> are given in ##UREF##18##[36]##). A typical network of this type is shown in ##FIG##3##Figure 4##. It consists of <italic>n</italic> = 250 nodes, <italic>m</italic> = 313 edges, and a connectivity of <italic>z</italic> = 0.01 (with <italic>z</italic> = 2<italic>m</italic>/(<italic>n</italic>\n²−<italic>n</italic>)). The nodes in this network are hierarchically distributed in the following sense: during the growth process the hubs are more likely connected to each other than to other nodes, thus forming the center of the graph, while the nodes with small degrees contribute to the periphery for the most part.</p>", "<title>Scale-free modular network</title>", "<p>This network type consists of several modules of approximate identical size. We used a modification of the community model ##UREF##1##[4]##,##UREF##23##[55]## to generate graphs with 5 modules (<italic>n</italic> = 250, <italic>m</italic> = 496, and <italic>z</italic> = 0.016 [<italic>n</italic> = 250, <italic>m</italic>≈515, and <italic>z</italic>≈0.0165 for the analysis of the randomized topologies]). According to the BA graph generation rule, each module starts with a small number of fully connected nodes (<italic>n</italic>\n₀ = 2). All further nodes are attached preferentially with <italic>m_A</italic> = 2 edges until the average size of each module is reached. At last, each module pair is connected with <italic>m_E</italic> = 1 (<italic>m_E</italic> = 3 for the analysis of the randomized topologies) random edge on the average. In contrast to the BA graph, the hubs are distributed among the modules.</p>", "<title>Randomized network topologies</title>", "<p>We use a systematic randomization process to modify an existing network topology in a directed way. In this procedure two linked pairs of nodes are randomly selected and rewired (i.e. the two edges are reassigned among the four nodes) as long as neither network fragmentation occurs nor double or self-edges form. In the course of the first variant of this randomization procedure (process 1), the topological modularity <italic>Q_top</italic> (determined as described, e.g., ##UREF##1##[4]##,##UREF##24##[56]##) of a graph is reduced by randomly selecting pairs of nodes in different modules and cross-linking them, thus increasing the amount of inter-modular edges. To avoid the formation of a hierarchical structure we ensured cross-linking between nodes with a degree <italic>k_s</italic>&lt;<italic>median</italic>(<italic>k_N</italic>) with <italic>N</italic> = (1,2,3,‥,<italic>n</italic>). During the other variant of the randomization procedure (process 2), the influence of the hubs, specified by the betweenness centrality <italic>B</italic>, is reduced by first selecting an edge connecting two nodes with <italic>B</italic>&gt;0.4·max(<italic>B_N</italic>) and then selecting a second edge with <italic>B</italic>&gt;0.2·max(<italic>B_N</italic>). The BA graph exhibits only a small amount of nodes with betweenness values above this threshold. The elimination of the most important edges ensures a drastic degradation of the central hub significance with increasing randomization depth.</p>", "<title>Hierarchical scale-free network</title>", "<p>Hierarchical scale-free graphs ##UREF##5##[10]##,##UREF##11##[28]## have been introduced to account for both, hub dominance and modular clustering. The graph generation is based on a fractal algorithm which uses multiplication and cross-linking of existing graph structures to produce a deterministic scale-free network with self-similar elements. Compared to a BA graph, the degree <italic>k_h</italic> of the central node <italic>h</italic> of the hierarchical graph is notedly high (<italic>k_h</italic> = (−3+3<italic>^it</italic>\n⁺¹)/2 with <italic>it</italic> denoting the number of iterations in the generation rule). Such a network would still display an unbalance in both levels of dynamic organizations. To reduce <italic>k_h</italic> we added the probability <italic>g</italic> for an edge to form between <italic>h</italic> and the respective other node during the generation process. Starting with a set of 4 completely connected nodes and applying the rules in ##UREF##0##[2]## one would yield a graph with <italic>n</italic> = 256 nodes and <italic>m</italic> = 780 edges after 4 iterative steps. For <italic>g</italic> = 0.5 the resulting networks possessed <italic>m</italic>≈650 edges.</p>", "<title>Fractal modular network</title>", "<p>The fractal modular network displays some basic properties of the hierarchical scale-free network, but its fractal connection scheme disagrees completely. This network has been introduced by Sporns et al. ##REF##16757100##[23]## for the analysis of the cerebral cortex which is also characterized by multiple hierarchical levels. We constructed a mapped fractal graph with six hierarchical levels according to the following parameter constellation. We preset parameter <italic>E_s</italic> which in principle regulates the connectivity of the graph to a value of <italic>E_s</italic> = 2. Starting with a complete graph of 8 nodes (<italic>m_S</italic> = 3 and <italic>n_S</italic> = 8) the resulting graph comprises <italic>n</italic> = 256 nodes and <italic>m</italic> = 3456 edges (for details on the generation of the mapped fractal networks see Sporns et al. 2006 ##REF##16757100##[23]##; the index <italic>S</italic> denotes the variables which are used in ##REF##16757100##[23]##).</p>", "<title>Biological Neural Network Data</title>", "<p>We applied the analysis approach to two sets of neural network data at different scales of organization. The first data set describes systems level connections between different areas of the cat cerebral cortex, and is based on a global collation of cat cortical connectivity (892 interconnections of 55 areas). This collation was developed from the data set described in Scannell et al. (1995) ##REF##7869111##[57]## and forms part of a larger database of thalamo-cortical connectivity of the cat ##REF##10355908##[39]##. The database was created by the interpretation of a large number of reports of tract-tracing experiments from the anatomical literature.</p>", "<p>The second data set represents cellular neuronal connectivity of the nematode <italic>C. elegans</italic> (277 neurons and 2,105 synaptic connections). This data set was adapted from Achacoso and Yamamoto (1992) ##UREF##19##[43]##. That compilation is largely based on the dataset of White et al. ##UREF##25##[58]## in which connections were identified by electron microscope reconstructions. The previously presented connectivity data ##UREF##19##[43]## was modified in the following way. Neurons of the pharyngeal ring, for which there was no internal connection information, were removed from the network, leaving 280 neurons. In addition, three neurons (AIBL, AIYL, and SMDVL) were removed, because of lacking spatial information. Eventually 277 neurons were included in the analysis. The size of the global and local <italic>C. elegans</italic> datasets analyzed here was comparable to that used in previous studies. For example, studies of the small-world properties ##REF##9623998##[37]## or characteristic motifs ##UREF##6##[12]## of <italic>C. elegans</italic> considered 282 and 187 neurons, respectively. Both chemical and electric synapses (gap junctions) were included as connections in the analysis.</p>", "<title>Topological References</title>", "<p>In order to understand how topological properties and dynamic observations are related, we will address our quantification schemes for topology and dynamics separately at first.</p>", "<p>We determine two topological references which are both based on the pairwise distances of all nodes within a network. Let the distance <italic>L_st</italic> be the shortest path connecting node <italic>s</italic> with node <italic>t</italic> The first reference is based on the topological modules (topological module reference, TM, see ##FIG##1##Figure 2 top##). It is computed from the distance matrix <italic>L</italic> = <italic>L_st</italic> which is then analyzed with a standard hierarchical clustering method. We tested single-linkage, complete-linkage and average-linkage approaches and found basically no differences between these methods for the task at hand. In the following, we used UPGMA (Unweighted Pair Group Method with Arithmetic mean) clustering, that is, the pair-wise combination of nodes or groups of nodes with minimal distance which is determined by the arithmetic means of the respective groups. The relative positions of the nodes which are the leaves of the topological reference tree obtained in this fashion are a condensed representation of all distance relations within the network. A similar way of analyzing the module structure uses the topological overlap ##REF##12202830##[14]##. The modules predicted with this method can be recovered from the topological reference tree by horizontally cutting the tree at a certain hight. The tree fragments resulting from this thresholding procedure serve as module predictions. In principle one has to analyze the dependence of the module predictions on threshold variation or conversely one can determine the threshold by prescribing the number of modules <italic>μ</italic>. Assigning a label (e.g. a color) to each node within a particular module leads to the final result, the TM reference, for which agreement with the distribution patterns of excitations can be checked.</p>", "<p>The second topological reference is based on the central node <italic>h</italic> of the network (central node reference, CN, see ##FIG##2##Figure 3 top##). Although many properties can in principle contribute to the centrality of a node, we will here select node <italic>h</italic> to be the one displaying the highest node betweenness <italic>B</italic>\n##UREF##26##[59]##–##UREF##28##[61]##. The distances between <italic>h</italic> and all other nodes form a distance vector. All nodes with the same entry in the distance vector (e.g. equidistant nodes from <italic>h</italic>) are taken to form a cluster, representing this topological reference (CN clusters). Resorting the distance vector accordingly yields the color-coded CN reference. Here, the number of clusters <italic>μ</italic> is given by the maximal distance from node <italic>h</italic>.</p>", "<title>Dynamic Models</title>", "<p>Dynamics were simulated on the graph architectures using the discrete excitable (DE) model described in the introduction. We used 35000 update steps (first 10000 updates were discarded) with the following parameter constellation: the rate of spontaneous excitations <italic>f</italic> was varied in the range of 10⁻⁶&lt;<italic>f</italic>&lt;1 to systematically study the impact of noise on the formation of the excitation patterns; recovery probability <italic>p</italic> was set to a constant value of <italic>p</italic> = 0.1; the initial condition was a random equipartition of the states <italic>E</italic> and <italic>T</italic>. This parameter constellation will be used in all of the studies presented here.</p>", "<p>In the basic DE model highly connected networks are in principle characterized by burst dynamics. Indeed, spikes emerge at very low values of <italic>f</italic> even here, but with a sufficiently high simulation time they are outbalanced by burst dynamics. We solved this problem by introducing parameter <italic>ω</italic> in our excitable model system. This threshold depends on the degree <italic>k_s</italic> of a node <italic>S</italic> and determines the number of excitations necessary to turn a susceptible node into the excited state. In this variant all incoming excitations are stored in node <italic>S</italic> until <italic>ω_s</italic> = <italic>k_s</italic>·<italic>κ</italic> (with a minimum value of <italic>ω</italic> = 1) is reached.</p>", "<title>Comparison between Topological and Dynamic Organization</title>", "<p>In order to allow for a direct comparison with topology, we base our analysis of the dynamics on pairwise node comparisons: for each pair of nodes we count the number of simultaneous excitations <italic>σ_st</italic> in the given time interval. Properly normalizing these quantities to arrange between 0 and 1 (<italic>σ</italic> ˜<italic>_st</italic>) and converting the corresponding matrix into a distance matrix <italic>C</italic> = <italic>C_st</italic> = 1−<italic>σ</italic> ˜<italic>_st</italic> leads to the correlation matrix <italic>C</italic> which represents the distribution patterns of excitations for a given graph and a given parameter constellation of the DE model. We aimed at understanding to what extent a selected topological reference is capable of explaining the patterns in the correlation matrix. To this end, the matrix can now be converted into a clustering tree (again by using UPGMA: see Topological references). The idea is now to rearrange the branches in the tree to best fit a given reference vector. The corresponding sequence of nodes constitutes the final result for the dynamics, namely the vector of dynamically detected clusters (DDC vector). The reference of the sorting vector can be any of the two topological references discussed above. ##FIG##1##Figures 2## and ##FIG##2##3## summarize our analysis strategy. For the sorting we use an alignment algorithm which switches two neighboring branches at any position in the tree (obtained from the excitation patterns) as long as the similarity to the topological reference is increased. The decisive factor concerning the comparison of a pair of branches is the individual module composition of the respective leaves indicated by the mixture of (color) labels. A similar technique for the comparison of clustering trees has been introduced in ##UREF##29##[62]##.</p>", "<title>A Measure of Dynamic Modularity</title>", "<p>For computation of our new quantity assessing the match between topology and dynamics, the dynamic modularity <italic>Q_dyn</italic>, we compare two clustering trees, one coming from topology (with the clusters in the tree matching the modules in the graph), the other coming from the dynamics (more specifically: the matrix of simultaneous excitations). Cutting the first tree at a certain height (given by the module number, which is a parameter in our analysis) yields a set of modules, which we label by colors. Copying these node labels in the topological tree to the dynamic tree, and sorting for as many matching colors as the tree structure allows, permits us to quantify the color matches and mismatches between the two trees. Our null model is randomly distributing color labels on the graph (i.e. a sorting task of the dynamics tree to a random topological reference). As all these quantities depend strongly on the numbers of nodes in each module (or reference class), we normalize them to these sizes. In practice, this normalization is only important when we have very different sizes of modules in a graph. In this way we can assess whether the matching between a topological feature (here: the modules) and the dynamics (represented by the matrix of simultaneous excitations) is higher (or, in principle, even lower) than expected at random.</p>", "<p>The same holds for the other topological reference, the CN reference, where the labels are provided not by a clustering tree, but by the distance from the central node. The possible values of for a topological reference <italic>R</italic> lie between zero and unity with indicating the strongest agreement to the topological reference. Values below unity hint at a deviating distribution of nodes in the dynamic cluster tree.For both the topological reference and each DDC vector the distribution values <italic>θ</italic> are determined via comparison of the scattering of nodes <italic>π</italic> belonging to the same topological module <italic>i</italic> (as indicated by the color) with a null-hypothesis of this color distribution which is the average standard deviation (in <italic>l</italic> = 1,000 realizations) of the same amount of nodes randomly scattered over the whole network size <italic>n</italic>. The resulting quotient is normalized to the size of each module <italic>n_mod</italic>.\n\n</p>" ]

[ "<title>Results</title>", "<title>Overview</title>", "<p>In this study, we focus on two structural properties of networks and use them in terms of topological references. These properties are modularity and node centrality and they are represented by the topological modularity (TM) reference and the central-node based (CN) reference, respectively. To highlight the individual impact of each topological property on dynamic pattern formation we first probe different types of artificial networks dynamically and compare the results with the respective topological reference. We then validate our results with modified versions of these networks (see ##SUPPL##0##Figure S1## and ##SUPPL##1##Figure S2## in ##SUPPL##5##Text S1##: Analysis of randomized network topologies) and with different types of hierarchically structured graphs, which represent the two topological properties to different extents. We finally transfer our analysis to more densely connected networks and to different hierarchically structured real-world topologies (see <xref ref-type=\"sec\" rid=\"s4\">Methods</xref> for details on the construction of the respective references, the dynamic models and the different types of graph architectures and graph randomization processes). ##FIG##1##Figures 2## and ##FIG##2##3## summarize our strategy of comparing the pattern of simultaneous excitations (correlation matrix <italic>C</italic>) with the corresponding topological feature, namely the topological modules (TM, ##FIG##1##Figure 2##) and the central-node based reference (CN, ##FIG##2##Figure 3##). Both, the graph and the simulated “space-time” pattern are converted into matrices giving the pairwise distances and the number of simultaneous excitations, respectively. The two matrices are processed further to yield the respective clustering trees, which then are sorted, color-coded and systematically compared (see <xref ref-type=\"sec\" rid=\"s4\">Methods</xref> for a detailed description of this procedure.)</p>", "<title>Analysis of the Modular Topology</title>", "<p>We start our analysis with the modular scale-free network in order to test the explanatory power of the TM reference. As a first step we visualize for a single value of <italic>f</italic> how well the dynamically detected clusters follow the topological modules. We can map the clustering tree obtained from the correlation matrix onto the graph by thresholding it to yield the same number of modules <italic>μ</italic> as detected topologically and assign colors as labels to the modules. ##FIG##3##Figure 4## displays the corresponding graph with the modules colored exclusively on the basis of the dynamically detected clusters (DDCs), resulting from a simulation with <italic>f</italic> = 0.01. In this case, the dynamic clusters have a large overlap with the modules found topologically.</p>", "<p>As a next step, we analyze the whole range of the parameter <italic>f</italic>. This is summarized in ##FIG##4##Figure 5##. The color bar on the left-hand side represents the color-coded TM reference. The sequence of color bars from left to right are the color-coded DDC vectors for increasing values of <italic>f</italic>. There are three distinct ranges in <italic>f</italic> characterized by different patterns of the DDC vectors. Above a value of <italic>f</italic> = 0.1 any regularity is replaced by a random distribution of colors. Here, the random excitation events dominate the dynamics, thus leading to uncorrelated excitations and to a formation of unsystematic dynamically detected clusters. For lower values of <italic>f</italic> two different forms of node integration into dynamic clusters can be discriminated. Up to a value of <italic>f</italic> = 10⁻³ the DDC vectors are a mixture of homogeneous regions (representing well detected topological modules) on the one hand (in the bottom part of each DDC vector in this <italic>f</italic> range) and regions with smaller scale homogeneities on the other (top part of the DDC vectors). In this range the topological modules coincide partly with the dynamic clusters, but the dynamic integration fails to comply with the topological hierarchy of the modules. The middle range in <italic>f</italic> = 0.01 (10⁻³&lt;<italic>f</italic>&lt;0.1) is characterized by a very high order of the DDC vectors and an almost perfect agreement with the TM reference. Besides this systematic dynamic retrieval of the topological modules the DDC vectors in this <italic>f</italic> -range are also characterized by a strong consistency with the hierarchy of the modules on the level of the whole graph. The separation of the DDC vectors into two regimes with respect to <italic>f</italic> (omitting here the noise-driven high <italic>f</italic>-regime) is basically driven by the three-state model's behavior under spontaneous excitations. As pointed out in our previous work ##UREF##18##[36]##, the model displays a transition in the distribution patterns of excitations from a global (spike) to a more local (burst) regime with an increasing rate of spontaneous excitations <italic>f</italic>. While a spike (low-<italic>f</italic> regime) is able to reach most of the system (depending on the excess of nodes in the excitable state <italic>S</italic>), the burst (higher-<italic>f</italic> regime) is characterized by one or more excitation spots which propagate through the system on a localized level due to a more balanced distribution of the states <italic>S</italic> and <italic>R</italic> (Video S1 in ##SUPPL##5##Text S1## illustrates the propagation of excitations on a modular graph architecture during the burst regime). Consequently, the DDC vectors separate rather precisely at the position where the burst dynamics outbalances the spike dynamics. In this sense the burst dynamics provides a suitable tool for the dynamic retrieval of topological modules.</p>", "<title>Analysis of the Hub Dominance</title>", "<p>The results for <italic>f</italic>&lt;10⁻³ suggest that another form of dynamic integration of nodes takes place beyond the module level. Groups of nodes which belong to different topological modules (see e.g. the blue and red labels in ##FIG##4##Figure 5##) are placed in close dynamic proximity (that is, they are integrated into the same dynamic cluster). For testing this new principle of dynamic integration we repeat this simulation with a non-modular scale-free BA graph (see <xref ref-type=\"sec\" rid=\"s4\">Methods</xref>) and the CN reference discussed in ##FIG##2##Figure 3##. In ##FIG##5##Figure 6## the BA graph representation has been color-coded according to the dynamically detected clusters (with a preset value of 7 clusters, which determines the threshold applied to the corresponding clustering tree) at <italic>f</italic> = 10⁻⁵. One observes a rather clear ring-like arrangement of colors around a central node which is one of the hubs in the graph. This distribution of the dynamic clusters around a central node <italic>h</italic> (displayed in black) confirms our hypothesis that another topological feature is shaping the distribution of excitations in this low-<italic>f</italic> regime.</p>", "<p>Studying the agreement between the CN reference and the DDC vectors for the BA graph over a whole range in <italic>f</italic> leads to the result shown in ##FIG##6##Figure 7##. The CN reference (left-hand side) clusters all nodes <italic>t</italic> according to their distances <italic>d</italic> to the central node <italic>h</italic> with <italic>d</italic> = <italic>L_ht</italic>. Up to a value of <italic>f</italic> = 10⁻³ all equidistant nodes assemble more or less in the same dynamic cluster and even the distance order is maintained (except for <italic>d</italic> = 1 and <italic>d</italic> = 2). Above <italic>f</italic> = 10⁻³ the homogeneity of the DDC vector drops rapidly finally reaching a random composition. Again this decrease of dynamic order is accompanied by a decrease of the spike regimes in the overall dynamics. The recurrent simultaneous excitations which lead to the observed pattern are caused by global properties of the graph's topology. We assume that such networks are able to channel the excitations produced by random events into their centers, which are composed of one or a few nodes displaying the highest betweenness centrality (as given by the number of shortest paths leading to the node; see <xref ref-type=\"sec\" rid=\"s4\">Methods</xref>). From there, the excitation waves pass through the rest of the system reaching all equidistant nodes (seen from the center) at about the same time and thus integrate them dynamically. The dynamics in Video S2 in ##SUPPL##5##Text S1## contains several spike events which demonstrate the typical propagation of excitations in a BA graph. In addition ##SUPPL##4##Figure S5## illustrates the consistency between the sequential arrangement of ring-shaped modules (as seen from the central node) and the chronology of excitations showing the fraction of simultaneous excitations within each of these modules at the same time.</p>", "<title>Analysis of Hierarchical Network Topologies</title>", "<p>An integration of both topological properties (modularity and hub dominance) into one system has been accomplished via the introduction of the hierarchical scale-free graph ##UREF##5##[10]##,##UREF##11##[28]##. We expect from the previous discussion that both levels of dynamic organization are present in such a network. As other network designs exhibit hierarchical properties as well, we contrasted different types of hierarchical graphs, also considering densely connected graph structures which, for instance, characterize many neuronal systems. To allow for the analysis of highly connected networks we extended our dynamic model with the additional node degree-dependent parameter <italic>ω</italic> (which regulates the excitability of a node, i.e. the number of excitations needed in order to trigger a firing event; see <xref ref-type=\"sec\" rid=\"s4\">Methods</xref> for the exact definition of <italic>ω</italic>).</p>", "<p>All hierarchical networks introduced here share a hierarchical fashion of linking the modules, but some of them lack the hubs and the scale-free degree distribution. One would expect that such graphs are not able to produce consistent ring-like excitation patterns as observed in the BA graph. In the following we will investigate how these topological properties determine the distribution pattern of excitations. We checked, however, that this general phenomenon does not depend on the exact method of generating a particular topological property.</p>", "<p>We tested four different hierarchical networks, i) the hierarchical scale-free graph ##UREF##5##[10]##,##UREF##11##[28]##, ii) a variant of the hierarchical scale-free model (which permits the construction of densely connected graphs), iii) the fractal modular network ##REF##16757100##[23]##, and iv) the hierarchical cluster network ##UREF##9##[21]##. We generated 10 networks of each graph type, simulated the dynamics, and computed <italic>Q_dyn</italic> from the resulting dynamic clustering trees, as before. Densely connected networks were simulated with a threshold of <italic>κ</italic> = 0.1, as described in <xref ref-type=\"sec\" rid=\"s4\">Methods</xref>.</p>", "<p>In the following the results are limited to the hierarchical scale-free graph ##UREF##5##[10]## and the mapped fractal graph ##REF##16757100##[23]## as both other results agree well with their respective counterparts. ##FIG##7##Figure 8## displays averaged over all networks as a function of <italic>f</italic> for the TM reference (blue ▵) and the CN reference (red ○). In the hierarchical graph (##FIG##7##Figure 8A##) the dynamic detection of the topological modules based on the TM reference works very well for high values of <italic>f</italic>. Increase and decline of depend on the transition from spike dynamics to burst dynamics and on the increasing noise intensity <italic>f</italic>, respectively (##SUPPL##2##Figure S3## displays the corresponding time series of the excitation density <italic>ρ_F</italic> for three different values of <italic>f</italic>). This increase is accompanied by decreasing values of for the CN-dependent results which display their maximum in the low <italic>f</italic>-regime. Here, the high values of indicate a strong dominance of the hubs and their importance for the formation of the excitation waves. Indeed, this graph structure facilitates the emergence of both forms of dynamic organization. This observation, that certain types of hierarchical graphs can host both dynamic patterns with the rate of spontaneous excitations inducing a switch from one to the other, will be discussed in detail elsewhere.</p>", "<p>In the mapped fractal graph (##FIG##7##Figure 8B## the absence of hubs prevents the generation of ring-like excitation patterns (as seen in the low values of ) with the effect that the range of dynamically detected topological clusters () enlarges towards low values of <italic>f</italic>.</p>", "<p>By an adjustment of the dynamic model the consistency to the more sparsely connected networks demonstrates that (i.e. by rescaling the excitability; see <xref ref-type=\"sec\" rid=\"s4\">Methods</xref>) it is still possible to retrieve both dynamic regimes even in densely connected graph architectures, similarly to the more sparsely connected networks. Rescaling the excitability (by requiring more than one excitation in the neighborhood for exciting a node) thus provides a consistent extension of our original dynamics to higher connectivities.</p>", "<title>Analysis of Biological Neural Topology</title>", "<p>Compared to metabolic reaction networks or protein-protein interaction networks, the architecture of many neuronal systems is characterized by a high density of connections ##REF##10703046##[13]##,##REF##10355908##[39]##,##REF##10703044##[40]##. We studied neuronal networks of two organisms at two fundamentally different levels of organization, namely the cortical systems network of the cat and the cellular neuronal network of the nematode <italic>C. elegans</italic>.</p>", "<p>First, we analyzed the cortical network of the cat, which has a well-characterized topology ##UREF##4##[8]##,##REF##10703046##[13]## and has been the basis of previous dynamical simulations ##REF##17280251##[16]##,##REF##10703050##[41]##,##REF##10703050##[42]##. We focused at connectivity at the systems level, which is more reliably established than cellular cortical connectivity. At the systems level, all the neurons of a cortical area are integrated into a single node. This coarse-graining approach scales the cortical network representation down to <italic>n</italic> = 55 nodes and 238 directed edges and 327 undirected edges which originate from 892 cortico-cortical connections.</p>", "<p>Second, we considered the cellular neuronal connectivity of the nematode <italic>C. elegans</italic>, which has also been studied extensively. Due to the fixed number of nodes, the neuronal network of <italic>C. elegans</italic> serves as an excellent neuronal model system ##UREF##19##[43]##. This version of the cellular neuronal network of <italic>C. elegans</italic> contained <italic>n</italic> = 277 nodes and 1731 directed edges and 187 undirected edges.</p>", "<p>The connection density of the cat cortex representation is comparatively high (<italic>z</italic> = 0.3), while the connection density of the neuronal network of <italic>C. elegans</italic> is about tenfold smaller (<italic>z</italic> = 0.028). Therefore, we decided to use the modified DE model for the cat cortex with <italic>κ</italic> = 0.15 and <italic>p</italic> = 0.1 and the original DE model for <italic>C. elegans</italic> with <italic>p</italic> = 0.01. We analyzed both networks in the range of 10⁻⁶&lt;<italic>f</italic>&lt;1. The TM references consist of 4 modules (cat) and 8 modules (<italic>C. elegans</italic>), respectively. The four modules in the cat systems network correspond to those previously identified by other clustering approaches ##REF##10703046##[13]##, and represent sets of visual, auditory, sensory-motor and fronto-limbic cortical areas. The diagrams (##FIG##8##Figure 9 top##) display the analysis of the dynamic modularity for both topological references. The diagrams on the bottom show corresponding curves with highlighted markers on the top. They display the TM-dependent DDC vectors for the Cat (##FIG##8##Figure 9A bottom##) and the CN-dependent DDC vectors for <italic>C. elegans</italic> (##FIG##8##Figure 9B bottom##).</p>", "<p>Examining the relation between topology and dynamic properties independently of the organism, both networks show certain characteristics of a hierarchical scale-free network ##UREF##5##[10]##,##UREF##11##[28]##, that is, the typical differences in the dynamic dominance of modular and hub features for different levels of spontaneous activation (as indicated in the <italic>f</italic>-dependent course of <italic>Q_dyn</italic> in ##FIG##8##Figure 9 top##), which implicate the existence of a complex hierarchical structure. However, both organisms also exhibit great differences in their dynamic regimes.</p>", "<p>For low levels of spontaneous excitation in the cat cerebral network (##FIG##8##Figure 9A top##), the CN and TM references are equally well related to the network's dynamic behavior. The strong correlation between dynamics and the modular topology is reflected in a high consistency between the TM reference and DDC vectors in the high <italic>f</italic>-regime (##FIG##8##Figure 9A bottom##) also indicated in ##FIG##8##Figure 9A top## in , while there seems to be only a marginal influence of hubs. If we exchange the TM reference by the modules previously identified for the cat cortical network ##REF##10703046##[13]##, the general features of <italic>Q_dyn</italic>(<italic>f</italic>) remain intact (in particular the clear peak in <italic>f</italic>; see ##SUPPL##3##Figure S4##).</p>", "<p>On the other hand, the dynamic behavior of the cellular network of <italic>C. elegans</italic> is for all but the highest levels of activation dominated by the distance to a central node (##FIG##8##Figure 9B##). Betweenness analysis revealed two nodes in direct neighborhood, which display the highest node degrees of the neuronal network, and which may serve as an initial point of circular excitation waves. Nodes 52 (AVAL) and 53 (AVAR) display the highest node betweenness (and the highest node degrees). The distance between both nodes is 1, as they are mirror-symmetric versions of the same neuron, AVA, on the L and R sides of the nematode's body.</p>" ]

[ "<title>Discussion</title>", "<title>Overview</title>", "<p>The current paper presents some aspects of a pattern-based computational approach for linking network topology and dynamics. This approach proved useful in probing the functional organization of complex biological networks. The comparison of topological features and simulated network dynamics demonstrated that features such as central hub nodes and network modularity can strongly and systematically shape a network's dynamic behavior. Moreover, in hierarchical modular networks, where multiple of these features were present, the network dynamics exhibit a functional switch for different levels of spontaneous network activation between the dynamic organization through a central node or through modular features.</p>", "<p>The method also reveals the dynamic impact of different topological characteristics in biological neural networks. In particular, the dynamics in the cellular neuronal network of <italic>C. elegans</italic> appears organized by the topological distance to a central hub node, whereas the dynamic behavior of the cat cerebral cortical network appears more strongly influenced by network modularity. Both topological features, however, contribute to the organization of the networks synchronization dynamics. Given the restricted size of the biological networks, the functional implications of the features would have been difficult to derive from a conventional analysis of the networks' degree distributions. These findings have implications for understanding the relationship of network topology and dynamics in complex neural networks, as detailed in the following sections.</p>", "<title>Model Limitations</title>", "<p>The presented approach draws on a simple dynamic model for describing excitable elements. This model only represents node activation, inactivation, as well as a refractory period, with discrete time steps. Given the complex dynamic behavior of neurons and neuronal systems, the model may appear overly simplistic. However, we believe that the model captures essential features of excitable elements, such as the principal activation cycle of neurons. Moreover, at the moment it is far from clear how much detail is required to realistically describe the interaction of excitable elements in networks. A good starting point for analyzing such pattern-formation aspects also in more sophisticated models could be built upon the parallel to a recent simulation study of the cat cortical network, which uses a more sophisticated population oscillator model to describe the activity of individual cells within the cortical areas ##REF##17280251##[16]##. This study led to a similar finding of a modular dynamic organization that strongly followed the modular topological organization. There are also precedents for the successful application of highly simplified models of cortical networks. For example ##REF##10703050##[41]## used a simple spreading model to infer basic properties of the relationship between node lesions and network activity in the thalamo-cortical network of the cat. Similarly, ##REF##10703050##[42]## replicated epileptiform steady-state activation patterns in the cat cortical network with the help of a simple thresholded spreading model. In addition, in the present work the model parameters were varied over a wide range; however, the different simulations resulted in similar principal behavior.</p>", "<title>Topology and Dynamics of Neural Systems</title>", "<p>When applied to biological neural networks, our approach revealed that the dynamic behavior of neural networks may be coordinated via different topological features. While activity in the neuronal network of <italic>C. elegans</italic> is shaped by excitation spreading from central hub nodes, the dynamic behavior of the cat cortical network is largely dominated by the network's modular organization. Moreover, the cortical network may switch from modular to hub dominance for low levels of spontaneous activation.</p>", "<p>The current analysis applies to network dynamics with spontaneous node activations, as observed in tonic neural activity, but without explicit external (sensory) input. This description corresponds to the experimental case of so-called resting state connectivity, a type of functional connectivity that persists in the absence of specific external stimulation. Resting state networks have been studied intensively over the last years and have been considered as default frameworks of neural dynamics ##REF##17719799##[44]##. Resting state connectivity can be derived experimentally from time-series correlations between large-scale brain regions, such as cortical areas. The regions' activity is estimated from different functional imaging techniques (e.g., EEG, fMRI); and typically, the coupling occurs at very low frequencies, around or below 0.1 Hz ##REF##17919927##[45]##. The slow-frequency coupling may be a reflection of faster electrophysiological coupling among distributed neuronal populations ##REF##17548818##[17]##. Experimental resting state data are currently available for cortical networks in humans and non-human primates, but not for the cat cortical network studied here. However, the present theoretical findings largely agree with what is known from the available experimental data. For instance, resting state data for human and primate cortical networks at the systems level show a strongly modular organization ##REF##15635061##[46]##,##REF##17476267##[47]##. Earlier experimental findings, based on activity spreading after local cortical disinhibition, also suggest that primate cortical areas co-activate, in groups that closely match the known topological clusters ##REF##10703047##[15]##. In addition, previous theoretical studies also support the conclusion that the dynamic organization of large-scale cortical networks in the absence of external stimuli is strongly shaped by the networks' modular structural connectivity ##REF##17280251##[16]##.</p>", "<p>However, it was also suggested that hub-like areas exist in cortical networks which possess a relatively large number of connections and which can be identified implicitly from the networks' behavior after simulated node lesions ##UREF##4##[8]##,##UREF##10##[24]##,##REF##16399673##[48]##. The leading central nodes identified here for the cat cortical network by node betweenness, multimodal areas 35 and AES, are also among those suggested previously by degree and lesion impact ##UREF##4##[8]##,##UREF##10##[24]##. For low rates of spontaneous activation, the cortical dynamics became somewhat more strongly correlated to hub distance than network modules. This dynamic switch characterizes the cortical connectivity as a complex hierarchical network and indicates the possibility that particular cat cortical areas act as hub-like nodes for the organization of low-noise dynamic regimes. This point still needs to be investigated in more detail. Importantly, only coarse large-scale activations can be resolved with the current neuroimaging techniques. Nonetheless, it is clear that cortical networks have a multi-level modular organization (forming clusters of sub-clusters of excitable nodes ##UREF##9##[21]##, with modules spanning from cellular cortical circuits and columns to clusters of strongly interlinked areas). Therefore, it can be speculated that, once data for additional scales of cortical networks are available, switches of the dynamic behavior between different topological features become more clearly apparent.</p>", "<p>In contrast to the cortical network the dynamic behavior of the <italic>C. elegans</italic> network was dominated by central node distance for all levels of spontaneous activation. Experimental findings also indicate that neuronal dynamics in <italic>C. elegans</italic> are coordinated by central pattern organizers ##REF##10571229##[49]##,##UREF##20##[50]## rather than through network modules. Indeed, the pair of AVA neurons, which have the highest degree and highest node betweenness in the <italic>C. elegans</italic> network, and which therefore may be considered as network hubs, have been implicated as a component in a central pattern generator responsible for locomotion control ##REF##10571229##[49]##. Specifically, AVA is thought to be responsible for backward movements. The present results suggest that this node may also have a more general function in coordinating dynamic activity in the nematode nervous system.</p>", "<p>The finding of dynamic organization through network modules in large-scale cortical networks, versus organization through few central nodes in cellular neuronal networks, makes intuitive sense. Given the small size of its nervous system, the functional specialization in <italic>C. elegans</italic> occurs at the level of individual cells, which exert their roles globally across the network. On the other hand, specialization in the mammalian cortex arises for whole brain regions (e.g., visual cortex, sensory-motor cortex) comprising several cortical areas which are closely cooperating within modules to perform the various aspects of their functional subdivision.</p>", "<title>Conclusions and Outlook</title>", "<p>When studying dynamics on networks, the synchronization behavior of each single node is a suitable indicator to estimate the dynamic scope provided by a graph's topology. Different forms of synchronization require different structural properties. By the application of a simple excitable medium (the DE model) we were able to generate two distinct forms of synchronization via the regulation of a single dynamic parameter, the amount of spontaneous excitations <italic>f</italic>. This noise level <italic>f</italic> also defines the (length) scale on which a specific dynamic process will predominantly be situated. Consequently the (larger-scale) wave-like propagation (consistency with CN reference) is dominant at lower levels of <italic>f</italic>, while the local module-based synchronization (consistency with the TM reference) is situated preferentially at higher <italic>f</italic>.</p>", "<p>Via comparison to two different topological references representing the elementary graph properties modularity and hub dominance the dynamic results were attributed to the respective synchronization behavior. In the burst range of <italic>f</italic>, networks exclusively featuring modular properties with decentralized hubs display synchronization behavior predominantly within their communities as indicated by the consistency to a module-based topological reference. If a graph is dominated by one or a few hubs in its center (a feature of the BA graph) a global (ring-like) synchronization phenomenon is visible due to the formation of excitation waves which reach the whole system from the graph's center. In contrast to our modularity definition it is more difficult to decide whether a node is the center of a graph or not. Here, we used the betweenness centrality (<italic>B</italic>) definition, but the results indicate that <italic>B</italic> does not alone account for the unifying topological quantity for different networks. The analysis of different hub categories ##UREF##5##[10]##,##REF##15729348##[11]##,##UREF##6##[12]## and their involvement in organizing the dynamics ##UREF##10##[24]## is an important next step of the study described here. We did not do this so far, because it would require simulating substantially larger networks to obtain reliable results. We would also like to point out that the prototypes of pattern formation we identify, might serve as minimal models of the brain activity regimes reported by Izhikevich and Edelman in their model of mammalian thalamocortical systems, which emerge spontaneously as a result of interactions between architectural features and the dynamics ##REF##18292226##[51]##. An important challenge for the future will be to activate modeled neural networks more selectively with patterns representing functional inputs, and to observe the interactions of stimulus-related activity with default activity.</p>", "<p>In summary, by using a simple dynamic model we could determine a “network equivalent” of pattern formation, where patterns are represented by correlations between topology and dynamics. Specific topological features give rise to and regulate quantitatively certain elementary forms of patterns. We believe that this correspondence is not restricted to the specific dynamics considered here. The recent findings on synchronization of phase oscillators ##REF##16605825##[52]##,##UREF##21##[53]## show similar matches between topology and dynamics as the results reported for an excitable system. In this light a comparison of these systems in detail (our discrete excitable three-state model and the continuous phase oscillator model) would be very interesting and could point towards common links between topology and dynamics far beyond individual dynamical systems. It is particularly interesting that the authors employ phase oscillators and their synchronization properties also to determine functional groups in the neural system of <italic>C. elegans</italic>\n##UREF##22##[54]##.</p>" ]

[ "<title>Conclusions and Outlook</title>", "<p>When studying dynamics on networks, the synchronization behavior of each single node is a suitable indicator to estimate the dynamic scope provided by a graph's topology. Different forms of synchronization require different structural properties. By the application of a simple excitable medium (the DE model) we were able to generate two distinct forms of synchronization via the regulation of a single dynamic parameter, the amount of spontaneous excitations <italic>f</italic>. This noise level <italic>f</italic> also defines the (length) scale on which a specific dynamic process will predominantly be situated. Consequently the (larger-scale) wave-like propagation (consistency with CN reference) is dominant at lower levels of <italic>f</italic>, while the local module-based synchronization (consistency with the TM reference) is situated preferentially at higher <italic>f</italic>.</p>", "<p>Via comparison to two different topological references representing the elementary graph properties modularity and hub dominance the dynamic results were attributed to the respective synchronization behavior. In the burst range of <italic>f</italic>, networks exclusively featuring modular properties with decentralized hubs display synchronization behavior predominantly within their communities as indicated by the consistency to a module-based topological reference. If a graph is dominated by one or a few hubs in its center (a feature of the BA graph) a global (ring-like) synchronization phenomenon is visible due to the formation of excitation waves which reach the whole system from the graph's center. In contrast to our modularity definition it is more difficult to decide whether a node is the center of a graph or not. Here, we used the betweenness centrality (<italic>B</italic>) definition, but the results indicate that <italic>B</italic> does not alone account for the unifying topological quantity for different networks. The analysis of different hub categories ##UREF##5##[10]##,##REF##15729348##[11]##,##UREF##6##[12]## and their involvement in organizing the dynamics ##UREF##10##[24]## is an important next step of the study described here. We did not do this so far, because it would require simulating substantially larger networks to obtain reliable results. We would also like to point out that the prototypes of pattern formation we identify, might serve as minimal models of the brain activity regimes reported by Izhikevich and Edelman in their model of mammalian thalamocortical systems, which emerge spontaneously as a result of interactions between architectural features and the dynamics ##REF##18292226##[51]##. An important challenge for the future will be to activate modeled neural networks more selectively with patterns representing functional inputs, and to observe the interactions of stimulus-related activity with default activity.</p>", "<p>In summary, by using a simple dynamic model we could determine a “network equivalent” of pattern formation, where patterns are represented by correlations between topology and dynamics. Specific topological features give rise to and regulate quantitatively certain elementary forms of patterns. We believe that this correspondence is not restricted to the specific dynamics considered here. The recent findings on synchronization of phase oscillators ##REF##16605825##[52]##,##UREF##21##[53]## show similar matches between topology and dynamics as the results reported for an excitable system. In this light a comparison of these systems in detail (our discrete excitable three-state model and the continuous phase oscillator model) would be very interesting and could point towards common links between topology and dynamics far beyond individual dynamical systems. It is particularly interesting that the authors employ phase oscillators and their synchronization properties also to determine functional groups in the neural system of <italic>C. elegans</italic>\n##UREF##22##[54]##.</p>" ]

[ "<p>Conceived and designed the experiments: MTH. Performed the experiments: MML. Contributed reagents/materials/analysis tools: CCH. Wrote the paper: MML.</p>", "<p>This study investigates the contributions of network topology features to the dynamic behavior of hierarchically organized excitable networks. Representatives of different types of hierarchical networks as well as two biological neural networks are explored with a three-state model of node activation for systematically varying levels of random background network stimulation. The results demonstrate that two principal topological aspects of hierarchical networks, node centrality and network modularity, correlate with the network activity patterns at different levels of spontaneous network activation. The approach also shows that the dynamic behavior of the cerebral cortical systems network in the cat is dominated by the network's modular organization, while the activation behavior of the cellular neuronal network of <italic>Caenorhabditis elegans</italic> is strongly influenced by hub nodes. These findings indicate the interaction of multiple topological features and dynamic states in the function of complex biological networks.</p>", "<title>Author Summary</title>", "<p>Many complex biological networks are characterized by the coexistence of topological features such as modules and central hub nodes. What are the relative contributions of these structural features to the networks' dynamic behavior? We used a computational model to simulate the general activation and inactivation behavior of excitable nodes in neural networks and studied the spread of activity in hierarchically organized networks as well as specific biological neural networks. We then evaluated the impact of modules and hub nodes on the network dynamics by correlating the patterns of node activity with the network architecture at difference levels of spontaneous network activation. Two dynamic regimes were observed: waves propagating from central nodes and module-based synchronization. Remarkably, the dynamic behavior of hierarchical modular networks switched between these modes as the level of spontaneous activation changed. We also found that the two dynamic regimes have different significance in the neuronal network of <italic>C. elegans</italic>, where activity is mainly organized by hub nodes, and the systems network of the cat cerebral cortex, which is dominated by the network's modular organization. Our approach can be used to dynamically explore the organization of complex neural networks, beyond the structural characterizations that were available previously.</p>" ]

[ "<title>Supporting Information</title>" ]

[]

[ "<fig id=\"pcbi-1000190-g001\" position=\"float\"><object-id pub-id-type=\"doi\">10.1371/journal.pcbi.1000190.g001</object-id><label>Figure 1</label><caption><title>Basic graph models representing different combinations of both modular and hub characteristics.</title><p>The degree of a node (as an example of a hub characteristic) is indicated by its size, while the grouping of the nodes reveals the modular structure. (A) The Erdös-Rényi (ER) random graph lacks both hubs and modules; (B) the scale-free Barabási-Albert (BA) graph displays a center of interlinked hubs only; the (C) random modular graph and the (D) scale-free modular graph consist of planarly linked modules, which are composed of smaller ER graph and BA graphs, respectively. The hubs in the BA graph version are distributed among the modules. The hierarchical graphs in (E) and (F) are featured by modules consisting of modules. In contrast to the hierarchical cluster graph in (E), the hierarchical scale-free graph (F) is additionally characterized by a hierarchical structure of hubs with one hub dominating the center.</p></caption></fig>", "<fig id=\"pcbi-1000190-g002\" position=\"float\"><object-id pub-id-type=\"doi\">10.1371/journal.pcbi.1000190.g002</object-id><label>Figure 2</label><caption><title>Construction of a color-coded topological reference based on the TM of a network (Top), and formation of a dynamic clustering tree on the basis of the dynamic model simulation (Bottom).</title><p>(Top) The distance relations between all nodes are converted into a distance matrix <italic>L</italic>. (The color label encodes the distances between pairs of nodes.) The matrix <italic>L</italic> is then translated into a topological reference tree via UPGMA clustering (see <xref ref-type=\"sec\" rid=\"s4\">Methods</xref>). The node indices in the graph correspond to the ones in the tree, the circles in the graph representation denote the modules found in the cluster tree after assigning a threshold (dotted line) which separates the downstream branches. Next, color labels are assigned (TM reference). (Bottom) The model produces a space-time pattern of excitations (white lines) which is then converted into a correlation matrix <italic>C</italic>. (The color labels encode the number of simultaneous excitations.) The matrix <italic>C</italic> is translated into a clustering tree (from the excitation patterns). The color labels of the leaves are copied from the TM reference.</p></caption></fig>", "<fig id=\"pcbi-1000190-g003\" position=\"float\"><object-id pub-id-type=\"doi\">10.1371/journal.pcbi.1000190.g003</object-id><label>Figure 3</label><caption><title>Construction of a color-coded topological reference which is based on the location of the CN in the network (top row), and computation of the dynamic clustering tree is carried out as described in ##FIG##1##Figure 2## (bottom row).</title><p>(Top row) The central node <italic>h</italic> (inner circle in the graph representation) displays the highest betweenness centrality <italic>B</italic> (see <xref ref-type=\"sec\" rid=\"s4\">Methods</xref>: betweenness). It is surrounded by modules of equidistant nodes (from <italic>h</italic>). The nodes of the resulting distance vector are re-sorted according to their distance to <italic>h</italic>.</p></caption></fig>", "<fig id=\"pcbi-1000190-g004\" position=\"float\"><object-id pub-id-type=\"doi\">10.1371/journal.pcbi.1000190.g004</object-id><label>Figure 4</label><caption><title>Graph representation of the modular scale-free network.</title><p>The nodes are colored according to the dynamic clustering tree (resulting from a simulation with <italic>f</italic> = 0.01) after assigning a threshold for 5 modules (the number of topological modules). The dynamic clustering agrees with the topological modules almost completely.</p></caption></fig>", "<fig id=\"pcbi-1000190-g005\" position=\"float\"><object-id pub-id-type=\"doi\">10.1371/journal.pcbi.1000190.g005</object-id><label>Figure 5</label><caption><title>Dynamically detected cluster (DDC vectors) for 10⁻⁶&lt;<italic>f</italic>&lt;1 (right) re-sorted and colored according to the TM reference (left), as described in ##FIG##1##Figure 2##.</title><p>The region of the image displaying the highest consistency between the TM reference and the DDC vectors (10⁻³&lt;<italic>f</italic>&lt;0.1) marks the range, where the dynamics is able to exploit the given topological modules rather precisely. In this range of <italic>f</italic> the distribution patterns of excitations are dominated by burst regimes (as discussed in ##UREF##18##[36]##. The pattern formation for <italic>f</italic>&gt;0.1 is strongly influenced by random firing events, while for <italic>f</italic>&lt;10⁻³ the modular boundaries are followed only partly by the dynamics, hinting at another form of correlation between dynamics and topology, which acts on a larger topological scale.</p></caption></fig>", "<fig id=\"pcbi-1000190-g006\" position=\"float\"><object-id pub-id-type=\"doi\">10.1371/journal.pcbi.1000190.g006</object-id><label>Figure 6</label><caption><title>Network representation of the BA graph.</title><p>The nodes are colored according to the dynamic cluster tree (resulting from a simulation with <italic>f</italic> = 10⁻⁵) after assigning a threshold for 7 modules (the maximal distance to the hub). Most of the dynamically detected clusters are arranged in a ring-like fashion around the central hub highlighted in black.</p></caption></fig>", "<fig id=\"pcbi-1000190-g007\" position=\"float\"><object-id pub-id-type=\"doi\">10.1371/journal.pcbi.1000190.g007</object-id><label>Figure 7</label><caption><title>Dynamically detected clusters (DDC vectors) for 10⁻⁶&lt;<italic>f</italic>&lt;1 (right) re-sorted and colored according to the CN reference (left), as described in ##FIG##2##Figure 3##.</title><p>In this reference, the nodes sharing the same color have the same distance <italic>d</italic> to the central node <italic>h</italic> (see ##FIG##2##Figure 3 top row##). Up to a value of <italic>f</italic> = 10⁻³, the equidistant nodes are almost completely integrated dynamically according to this topological reference. In this <italic>f</italic>-regime the dynamics is characterized by excitation waves (spikes), which cover the whole system and which emerge from <italic>h</italic> preferentially and independently of the location of the accidentally excited node. The increasing scattering of colors for higher values of <italic>f</italic> indicates a change of the dynamic regime, the spike dynamics is increasingly replaced by burst dynamics.</p></caption></fig>", "<fig id=\"pcbi-1000190-g008\" position=\"float\"><object-id pub-id-type=\"doi\">10.1371/journal.pcbi.1000190.g008</object-id><label>Figure 8</label><caption><title>Levels of dynamic organization in different graphs with a hierarchical distribution of modules.</title><p>The dynamic modularity <italic>Q_dyn</italic> for both the TM reference (blue ▵) and the CN reference (red ○) is depicted as a function of the rate of spontaneous excitations <italic>f</italic>. (A) The hierarchical scale-free graph displays properties of both, the modular and the BA graph. Thus, the two levels of dynamic integration are visible within the same network for the respective values of <italic>f</italic>. The transition between these two levels corresponds to the transition from spike to burst dynamics. (B) The mapped fractal graph from ##UREF##22##[54]## lacks a scale-free degree distribution and, consequently, hubs, which is reflected in low values of . The absence of ring-like excitation patterns also explains the extension of the high-value range of towards low values of <italic>f</italic>.</p></caption></fig>", "<fig id=\"pcbi-1000190-g009\" position=\"float\"><object-id pub-id-type=\"doi\">10.1371/journal.pcbi.1000190.g009</object-id><label>Figure 9</label><caption><title>Levels of dynamic organization in two different neuronal networks.</title><p>The highlighted curves (bigger symbols; top row) correspond to the respective DDC vector results (bottom row). (A) The dominance of modular elements in the cortical network of the cat is reflected by a distinct increase of <italic>Q_dyn</italic> for the TM-dependent results in the high-<italic>f</italic> regime (blue ▵; top) as well as by the homogeneous clustering of the DDC vectors (TM-dependent results; bottom), while central node effects seem to play only a marginal role (see the slight superelevation in the low-<italic>f</italic> regime [red ○]; top). (B) By contrast, the cellular network of <italic>C. elegans</italic> displays a strong dependency on two adjoining central nodes which dominate the dynamics in a wide range of <italic>f</italic>. The drastic increase of the CN-dependent results for <italic>Q_dyn</italic> in the low-<italic>f</italic> region (red ○; top) reflects the high order of the DDC vectors (CN-dependent results; bottom) with a conserved distance ranking of the topologically detected node clusters. Even here, there exists a noticeable but comparatively subordinate influence of the module-based excitation patterns (blue ▵; top).</p></caption></fig>" ]

[]

[ "<inline-formula></inline-formula>", "<inline-formula></inline-formula>", "<inline-formula></inline-formula>", "<inline-formula></inline-formula>", "<inline-formula></inline-formula>", "<inline-formula></inline-formula>", "<inline-formula></inline-formula>", "<inline-formula></inline-formula>", "<inline-formula></inline-formula>", "<inline-formula></inline-formula>", "<inline-formula></inline-formula>", "<inline-formula></inline-formula>", "<disp-formula></disp-formula>", "<disp-formula></disp-formula>", "<disp-formula></disp-formula>" ]

[]

[]

[]

[]

[ "<supplementary-material content-type=\"local-data\" id=\"pcbi.1000190.s001\"><label>Figure S1</label><caption><p>Computation of the average dynamic modularity &lt;<italic>Q_dyn</italic>&gt; as a function of the topological modularity <italic>Q_top</italic> for different network realizations of the modular scale-free graph. Depicted are the TM results (blue △) which have been averaged over the range of 0.01&lt;<italic>f</italic>&lt;0.1 and the CN results (red ○), averaged in the respective range of 10⁻⁶&lt;<italic>f</italic>&lt;10⁻⁵. The modular graphs (<italic>n</italic> = 250, <italic>m</italic>≈515 with <italic>m_E</italic> = 3) were randomized in several steps producing networks with similar graph statistics but a decreased modularity. (A) Average randomization path of 10 different randomizations of the same network. The strong correlation between the TM dependent values of &lt;<italic>Q_dyn</italic>&gt; and the topological modularity <italic>Q_top</italic> proves the assumption that this level of dynamic organization has to be regarded as a consequence of the particular exploitation of modular network structures via burst dynamics. The respective exploitation via spike dynamics remains small and comparatively constant. (B) A similar behavior is also true for different network realizations and their respective randomization paths. These networks display the same correlation between &lt;<italic>Q_dyn</italic>&gt; and <italic>Q_top</italic>. One randomization path from (A) has been highlighted.</p><p>(0.87 MB EPS)</p></caption></supplementary-material>", "<supplementary-material content-type=\"local-data\" id=\"pcbi.1000190.s002\"><label>Figure S2</label><caption><p>Computation of the average dynamic modularity &lt;<italic>Q_dyn</italic>&gt; as a function of the hub dominance <italic>B̃</italic> for different network realizations of the BA graph. Corresponding to ##SUPPL##0##Figure S1A and Figure S1B##, different BA graphs and their randomized versions have been examined. (A) The randomization procedure causes a decrease of the hub dominance and, accordingly, a reduction of the CN-dependent values of &lt;<italic>Q_dyn</italic>&gt;. These results confirm the assumption that the whole graph structure and the central node in particular are responsible for the emergence of ring-shaped excitation waves, whose regularity is more and more disturbed with increasing randomization steps. (B) The randomization versions of the different networks are separated across the decreasing curve, but show nevertheless the same correlation as in (A).</p><p>(0.89 MB EPS)</p></caption></supplementary-material>", "<supplementary-material content-type=\"local-data\" id=\"pcbi.1000190.s003\"><label>Figure S3</label><caption><p>Sections of time series of the excitation density <italic>ρ_F</italic> of the hierarchical scale-free graph (see <xref ref-type=\"sec\" rid=\"s4\">Methods</xref>) for different rates of spontaneous excitations <italic>f</italic>. (From top to bottom) Increasing parameter <italic>f</italic> induces a change of the dynamic behavior from spike dynamics to burst dynamics with a transition region of <italic>f</italic> displaying a mixture of both dynamic regimes.</p><p>(0.74 MB EPS)</p></caption></supplementary-material>", "<supplementary-material content-type=\"local-data\" id=\"pcbi.1000190.s004\"><label>Figure S4</label><caption><p>Dynamic organization within the modular structure of the cortical network of the cat for two different definitions of the individual module composition. The curve indicated by the blue triangles corresponds to the TM-dependent results obtained from a UPGMA cluster analysis of the graph's distance information (using a threshold for 4 modules; see also ##FIG##8##Figure 9A##). Similar results (green curve; errors are of the size of the other results) were obtained from simulations using a different TM-reference consisting of 5 modules which have been identified in a work of Hilgetag et al. (2000). The additional module contains three nodes which could not be assigned to the remaining modules. Concerning the individual composition, both references display a high consistency (75%).</p><p>(0.74 MB EPS)</p></caption></supplementary-material>", "<supplementary-material content-type=\"local-data\" id=\"pcbi.1000190.s005\"><label>Figure S5</label><caption><p>Average fraction of excited nodes within each module resulting from previous excitations in a simulation of the scale-free (BA) graph in the spike-regime (<italic>p</italic> = 0.1 and <italic>f</italic> = 10⁻⁴). In the presence of excitations within the BA graph at a given time <italic>t</italic> the respective module (which is the concentric arrangement of nodes resulting from the CN-reference) with the strongest excitation density <italic>ρ_F</italic>, i.e. the biggest fraction of excited nodes compared to its module size, has been identified. As a function of this module (the numbers on the abscissa denote the distance of the modules to the central node) the distribution of excitations over all modules has been computed for the following time step t+1 and depicted on the ordinate as the module-specific fraction of excited nodes. Based on the central node and the resultant concentric modules there is an apparent propagation of the excitations in the spike-regime from the center of the graph to its periphery including an average module-specific excitation of 45 to 65 percent of the respective nodes.</p><p>(0.67 MB EPS)</p></caption></supplementary-material>", "<supplementary-material content-type=\"local-data\" id=\"pcbi.1000190.s006\"><label>Text S1</label><caption><p>Supporting Information</p><p>(1.30 MB ZIP)</p></caption></supplementary-material>" ]

[ "<fn-group><fn fn-type=\"COI-statement\"><p>The authors have declared that no competing interests exist.</p></fn><fn fn-type=\"financial-disclosure\"><p>The authors received no funding for this work.</p></fn></fn-group>" ]

[ "<graphic xlink:href=\"pcbi.1000190.g001\"/>", "<graphic xlink:href=\"pcbi.1000190.g002\"/>", "<graphic xlink:href=\"pcbi.1000190.g003\"/>", "<graphic xlink:href=\"pcbi.1000190.g004\"/>", "<graphic xlink:href=\"pcbi.1000190.g005\"/>", "<graphic xlink:href=\"pcbi.1000190.g006\"/>", "<graphic xlink:href=\"pcbi.1000190.g007\"/>", "<inline-graphic xlink:href=\"pcbi.1000190.e001.jpg\" mimetype=\"image\"/>", "<inline-graphic xlink:href=\"pcbi.1000190.e002.jpg\" mimetype=\"image\"/>", "<inline-graphic xlink:href=\"pcbi.1000190.e003.jpg\" mimetype=\"image\"/>", "<inline-graphic xlink:href=\"pcbi.1000190.e004.jpg\" mimetype=\"image\"/>", "<inline-graphic xlink:href=\"pcbi.1000190.e005.jpg\" mimetype=\"image\"/>", "<inline-graphic xlink:href=\"pcbi.1000190.e006.jpg\" mimetype=\"image\"/>", "<graphic xlink:href=\"pcbi.1000190.g008\"/>", "<inline-graphic xlink:href=\"pcbi.1000190.e007.jpg\" mimetype=\"image\"/>", "<inline-graphic xlink:href=\"pcbi.1000190.e008.jpg\" mimetype=\"image\"/>", "<inline-graphic xlink:href=\"pcbi.1000190.e009.jpg\" mimetype=\"image\"/>", "<graphic xlink:href=\"pcbi.1000190.g009\"/>", "<inline-graphic xlink:href=\"pcbi.1000190.e010.jpg\" mimetype=\"image\"/>", "<inline-graphic xlink:href=\"pcbi.1000190.e011.jpg\" mimetype=\"image\"/>", "<inline-graphic xlink:href=\"pcbi.1000190.e012.jpg\" mimetype=\"image\"/>", "<graphic xlink:href=\"pcbi.1000190.e013.jpg\" mimetype=\"image\" position=\"float\"/>", "<graphic xlink:href=\"pcbi.1000190.e014.jpg\" mimetype=\"image\" position=\"float\"/>", "<graphic xlink:href=\"pcbi.1000190.e015.jpg\" mimetype=\"image\" position=\"float\"/>" ]

[ "<media xlink:href=\"pcbi.1000190.s001.eps\"><caption><p>Click here for additional data file.</p></caption></media>", "<media xlink:href=\"pcbi.1000190.s002.eps\"><caption><p>Click here for additional data file.</p></caption></media>", "<media xlink:href=\"pcbi.1000190.s003.eps\"><caption><p>Click here for additional data file.</p></caption></media>", "<media xlink:href=\"pcbi.1000190.s004.eps\"><caption><p>Click here for additional data file.</p></caption></media>", "<media xlink:href=\"pcbi.1000190.s005.eps\"><caption><p>Click here for additional data file.</p></caption></media>", "<media xlink:href=\"pcbi.1000190.s006.zip\"><caption><p>Click here for additional data file.</p></caption></media>" ]

{ "acronym": [], "definition": [] }

[ "<title>Background</title>", "<p>The prevalence of back pain among high school students has been regularly reported to be an international public health concern [##REF##17978666##1##]. However, given the high frequency of research into adult back pain, adolescent back pain has a much lower research profile. In the small amount of available research, a high prevalence of back pain has been reported in the early teenage years, which then increases each year until the late teens [##REF##10975661##2##, ####REF##8915066##3##, ##REF##12160340##4####12160340##4##]. In developed countries, the lifetime prevalence of back pain in 15-year-olds exceeds 50% [##REF##8915066##3##]. Whether there is a similar trend among high school students in developing countries such as South Africa is yet to be established.</p>", "<p>There is limited but consistent research which indicates that many adolescents reporting frequent back pain become adults reporting back pain, perhaps because causal mechanisms and pain sensitisation become established during the formative years [##REF##17978666##1##,##REF##17976240##5##]. Given the high costs to the individual and to society of adult back pain [##REF##16982001##6##], minimising its prevalence by understanding causal mechanisms of precursor adolescent back pain would seem to be a public health priority.</p>", "<p>A number of causal mechanisms have been proposed for adolescent back pain, including carriage of heavy school bags, rapid bony growth, inadequate fit of furniture to body size, poor muscle strength, poor motor control, balance and coordination, and poor posture [##REF##17976240##5##,##REF##17429701##7##, ####REF##17302855##8##, ##REF##17521435##9####17521435##9##]. However, despite the interest in adolescent back pain, its causes are far from well understood. Sustained and poor sitting postures have been identified as important risk factors for back pain in adults [##UREF##0##10##,##REF##15074366##11##]. A common reason for adults to sit for long periods of time in poor postures is when using computers [##REF##9127560##12##]. Healthy computer use involves good workstation design features such as appropriate fit of body size to chair and desk height, screen angle and height, and keyboard arrangement, as well as the amount of time spent at the computer. Extended computer use has thus been proposed as a reason for adult back pain [##REF##16137636##13##,##REF##15177516##14##]. Computer use is increasingly common among high school students around the world, although whether it constitutes a risk for adolescent back pain has not been established [##REF##15996890##15##,##REF##11831210##16##]. Prior to testing any association between computer use and back pain, a reliable and valid measure of sitting posture is required. Any valid posture measurement tool must be able to detect postural abnormalities that could place abnormal stress on spinal structures.</p>", "<p>We undertook a systematic review of published research, which reported on sitting posture measurement tools. We identified nine relevant papers describing only three measurement approaches (goniometer [##UREF##1##17##, ####REF##1989013##18##, ##REF##10703104##19####10703104##19##], inclinometer [##UREF##2##20##, ####REF##14595170##21##, ##REF##11673934##22##, ##UREF##3##23####3##23##] and flexicurve [##REF##16226628##24##,##UREF##4##25##]). None of these approaches has been validated for adolescents (high school students). We found no papers in this systematic review on use of photographs to measure sitting posture, although photographs have been reported as a measure of adult, adolescent and children's standing posture [##REF##11960561##26##, ####REF##16096037##27##, ##REF##15985186##28####15985186##28##]. Given the reported reliability and efficiency of photographs, the longevity of digital records, and the cost-effectiveness of digital photographs to measure standing posture in adults and children, it is feasible that they would also be appropriate to measure sitting posture in adolescents.</p>", "<p>All the reported measures of sitting posture, as well as photographs, have one flaw. These measures are external to the body, that is, they use calculations from external bony landmarks to estimate spinal posture, on the understanding that what is being measured externally reflects the shape, health and performance of structures of the underlying spine. Without a comparative measure of the relative position of the structures of the spine, the validity of any external spinal posture in humans is often difficult to establish and may not give an accurate interpretation of true spinal alignment. The only trustworthy measure of the position of spinal structures is Radiography [##REF##16226628##24##,##REF##9608381##29##]. To date, however, little research has been undertaken to validate external posture measurement methods with Radiographs into healthy spinal posture, and this may largely be because of the ethical and health implications of subjecting healthy spines to irradiation [##REF##11595369##30##]. These concerns are, if possible, even more important for adolescents, given the potential influence of irradiation on growing systems and organs [##REF##11595369##30##].</p>", "<p>Recently a low dose Radiograph was developed in South Africa. The LODOX (LODOX (Pty) (Ltd) (a digital radiography device) was developed by De Beers as a safe Radiograph security scanner for the detection of smuggled diamonds. LODOX conducts a full body scan in 13 seconds, with smaller areas requiring proportionately less time [##REF##15290526##31##]. On average, the mean conventional dose of radiography is 0.573 R (5.73 mGy) while the mean digital dose (LODOX) is 0.033 R (0.33 mGy), 5.9% of the dose of the conventional Radiograph [##REF##15290526##31##]. Low dose radiograph systems provide population-applicable, 'Gold Standard' radiographic approach for measuring spinal segmental posture in healthy individuals.</p>", "<p>This paper reports on a study which tests the validity of photographs to measure adolescent sitting posture. The aim of the project is to correlate the postural angles of the photographs with LODOX images, for three types of adolescent sitting spinal postures (slouched, upright or normal).</p>" ]

[ "<title>Methods</title>", "<title>Ethics</title>", "<p>Ethical approval was obtained from the Committee for Human Resources at Stellenbosch University and the Western Cape Department of Education. Written informed consent was obtained from all students, and their parents or legal guardians.</p>", "<title>Setting</title>", "<p>The study was conducted in a laboratory at the Department of Human Biology, University of Cape Town.</p>", "<title>Sample size</title>", "<p>Sample size calculations were based on previously reported variability in sitting posture angles in healthy adults [##REF##10975665##32##,##REF##17419060##33##]. As little normative data was available on healthy adolescent sitting posture, this sample calculation was an estimate only. A sample of 40 was proposed (power 80%, alpha 5%) to detect differences of three degrees or more between the repeated measures of upright, normal or slouched posture (photographs) for the reliability study and between the posture photographs and LODOX measures for the validity study.</p>", "<title>Sample</title>", "<p>The population, from which the study sample was selected, comprised high school students from the Cape Metropolitan Region, Cape Town, South Africa. The Cape Metropolitan Region is divided into four educational management regions. One school from each region was selected by a statistician independent of the study, using a random numbers table. Eligible participants were healthy male and female subjects aged 15 or 16 years old, in Grade 10, and who were undertaking Computer or Computype studies. Eligible subjects in the selected schools were asked to volunteer to participate in the study. Subjects were excluded if they experienced any recent musculoskeletal pain or illness, which could compromise their ability, to assume upright, slouched or normal sitting posture on the day of data collection. These subjects were identified using a pain questionnaire that has been extensively validated for this population of high school students using computers. This questionnaire was administered prior to the commencement of validity and reliability testing [##UREF##5##34##].</p>", "<title>Measurement tools</title>", "<p>Two posture measurement tools were used in the study.</p>", "<p><italic>1. Photographs </italic>were taken using the Photographic Posture Analysis Method (PPAM). This method consisted of a digital camera (Fujifilm Finepix X5100), Intellect software (Version 1.1.4), reflective markers (see later section for details) and a computer for downloading images (Windows 2000 or XP compatible).</p>", "<p><italic>2. </italic>Radiograph<italic>s </italic>were taken using the LODOX (LODOX (Pty) Ltd) system (see Figure ##FIG##0##1##).</p>", "<title>Test purposes and set-up</title>", "<p>To test the validity of the photograph compared with the Radiograph, the testing station consisted of the LODOX surrounding the computer workstation, and one digital camera mounted on a tripod outside the LODOX. The LODOX system captured an image of the upper part of the body (T8 to head). The digital camera was positioned to capture the same spinal area as the LODOX. To test the reliability of sitting postures using photographs, the same workstations were set up outside the LODOX.</p>", "<title>Posture measurement set-up</title>", "<p>Subjects' posture was assessed in simulated computer workstations. The chair height and seat pan depth were selected based on the findings of an evaluation of school workstations in order for the chairs to simulate the typical chairs used in the schools [##UREF##6##35##]. Subjects could not adjust the chair position to suit their personal preference, as the current chairs in the schools are not adjustable. The chair height was between 440 mm and the seat pan depth was between 380 mm.</p>", "<title>Data collection procedures for the reliability and validitystudies</title>", "<p>The same 39 subjects participated in both the reliability and validity studies. The subjects and both studies were conducted on the same day for specific subject. Figure ##FIG##0##1## outlines the data collection procedure for both the reliability and validity studies.</p>", "<title>Subject preparation and positioning</title>", "<p>Anatomical markers were placed on all subjects by the one researcher, to identify seven external landmarks in photographs. These landmarks were randomly checked by another researcher (QL) to confirm their accuracy of placement. Prior to placement, the relevant areas of the subjects' skin were wiped with alcohol to facilitate good contact between the reflective markers and the skin. Golem retro-reflective markers were applied to the lateral canthus of the eye, the tragus of the ear, the spinous process of C7 [##UREF##6##35##], the midpoint of the superior border of the manubrium, T8 and the lateral epicondyle of the elbow [##REF##15996890##15##]. Both C7 and T8 markers were placed on extension sticks to allow for better visibility by the camera. All markers were placed on the subjects' dominant side and were not removed until testing was completed. Photographs and radiographs were taken from the dominant side. The markers were checked between each photograph and radiograph measure to ensure that they were in place, and accurate.</p>", "<title>Posture estimation</title>", "<p>Five postural angles were calculated from the LODOX images and the photographs (outlined below and illustrated in Figure ##FIG##1##2##).</p>", "<p>a) The <underline>sagittal head angle</underline> indicates the position of the head relative to the neck [##REF##15985186##28##]. This angle is commonly affected by computer usage [##UREF##1##17##]. McEvoy and Grimmer (2006) reported that a decrease in this angle reflects a \"poking-chin\" posture [##REF##15985186##28##,##REF##15681266##36##].</p>", "<p>b) The <underline>cervical angle</underline> is the measure of the forward-head position, which is a useful clinical marker of mid/lower cervical spine posture [##REF##15985186##28##].</p>", "<p>c) The <underline>protraction/retraction shoulder angle</underline> was measured using the method by Szeto <italic>et al. </italic>(2002) [##REF##11831210##16##].</p>", "<p>d) The <underline>arm angle</underline> was not measured in previously published studies. However, we have decided to measure the arm angle as it may be associated with the degree of shoulder protraction/retraction angle.</p>", "<p>e) The <underline>thoracic angle</underline> was also measured in the manner described by Szeto <italic>et al. </italic>(2002) [##REF##15996890##15##,##REF##11831210##16##]. Unfortunately few measures for this spinal section are reported in the literature.</p>", "<title>Camera positioning</title>", "<p>For all tests, the digital cameras (flash on) were mounted on tripods and placed 2 metres away from the chair on which the subject was positioned. The cameras were positioned so that all anatomical markers were detectable in the one image.</p>", "<title>Test protocols</title>", "<p>Approximately 12 subjects were tested per day. When they attended the testing session, subjects were randomly allocated to one of three sitting postures (slouched, straight or 'normal' (normal) sitting), as outlined in Figure ##FIG##0##1##. Subjects who had to assume the slouched posture were given the following instructions to \"lean with your arms on the table with your back bend forwards\", subjects who had to assume the straight posture were given the instructions to \"sit up straight with head, shoulders and hips in line\", while subjects who assumed the normal posture were given the following instructions \"sit as you would normally sit in front of a desktop computer\" subjects were given two to three practice opportunities to accommodate to the assigned posture. Subjects were instructed to assume the same allocated sitting posture for all tests. The use of three postures served to ascertain whether the PPAM could validly and reliably test postural angles through sitting posture range.</p>", "<title>Data capture from images</title>", "<p>The photographic and radiographic data was imported to a laptop via a USB data-transfer cable and Intellect 1.1.4 software (DVT Corporation). The principal researcher digitized all photographic and radiological data in order to calculate the angles. The Intellect 1.1.4 software functions are 'detecting and following a marker', 'circle fitting', 'constructing lines' and 'measuring angles'.</p>", "<title>Step 1</title>", "<p>To digitize the information of the Lodox images (see figure ##FIG##2##3##), the operator electronically placed a marker on the respective bony landmarks (spinous processes) of C7 and T8 in order for the software to detect the bony point. The rest of the markers could be used as they were as they were already placed on bony landmarks. Detecting and following a marker was the most complex function during the digitizing process. The software recognized the markers by defining the edges of the image. The user was required to 'teach' the software how to recognize the marker. The shape of the marker is 'learned' by the software by defining the edges in the image. Software learning refers to an automated memory of the software to process the information as done before.</p>", "<title>Step 2</title>", "<p>Once the software detected the marker, the next step in the photographic digitising process was to calculate the centre of the marker. This was done by applying edge detection on the border of the marker and thereafter a circle was fitted through the edge points. Provided that the markers could be detected accurately, the calculation of the angles for a series of images could be automated (See Figure ##FIG##3##4##)</p>", "<title>Step 3</title>", "<title>Calculation of the angles</title>", "<p>The system was programmed for the first image of each participant, and additional software (DVT Reader) was developed to apply the digitizing process described above to the full set of photographic images, instead of only a single image, in order to calculate the angles much faster than with the original Intellect 1.1.4 software. A co-author (KS) and another engineer developed the additional software. The angles were calculated using basic trigonometry. The (X, Y) positions of the markers are provided, as well as the length of the stems, where applicable. An example of how an angle was calculated is provided in Appendix 1.</p>", "<title>Statistical comparisons</title>", "<p>Descriptive and comparative statistics were calculated to determine differences and correlations between measurements. The mean and standard deviation for each angle from photographs and radiographs was calculated using Microsoft Excel (2002) and SPSS Viewer Version 14 software, for each of the sitting postures. Concurrent validity was estimated as Pearson r correlation coefficients, calculated between the mean angles from the two photographs, and the LODOX measures for each posture (normal, slumped and upright). Bland Altman, with the 95% limits of agreement equivalent to the mean difference ± 2 SD was also calculated to compare the angle values of the photographs and radiographs. Reliability was calculated between the angles from the five repeated photographs. Five repeated photos were taken for reliability study, this excluded photos taken for the validity study. Reliability was determined from the interclass correlation coefficients (ICCs) by means of the 2-way model and Standard Error Measurement (SEM) [##UREF##7##37##], with the strength of the ICCs interpreted as &lt;0.50 = poor, 0.50 &lt; 0.75 = moderate, 0.75 &lt; 0.90 = good and &gt; 0.90 = excellent. The ICC and SEM convey different information about reliability of a measure.</p>", "<p>The ICC provides information about a measure's capacity to differentiate change within subjects, whereas the SEM quantifies the error in the same units as the original measurement and therefore provides meaningful information about the reliability of the measurement.</p>" ]

[ "<title>Results</title>", "<p>Although 40 subjects consented to participate, one subject refused to undress her right (dominant) side due to burn scars. The data from 39 subjects (19 males and 20 females) were thus used for analysis for the reliability and validity studies. A total of seventeen 15-year-olds (7 boys and 10 girls) and twenty-two 16-year-olds (12 boys and 10 girls) were examined. Table ##TAB##0##1## reports participants' age, gender and posture allocations.</p>", "<p>The findings of this study suggest that photographs provide valid and reliable indicators of the position of the underlying spine in sitting.</p>", "<title>Validity</title>", "<p>Table ##TAB##1##2## reports the Pearson r values comparing the LODOX measures with the mean values from two photographs, of the five PPAM angles in each of the randomly allocated postures. All photographically captured angles (except for the protraction/retraction angle) demonstrated strong correlation with LODOX angles, with Pearson r correlation values of at least 0.84. The protraction/retraction angle in the normal sitting posture demonstrated the lowest Pearson r correlation value overall, even thought this was still a moderate correlation.</p>", "<title>Bland Altman limits of agreement</title>", "<p>Bland Altman limits of agreement are demonstrated in Table ##TAB##2##3##. The Bland Altman method revealed a small bias of -1.12° for the cervical angle, -1.56° for the head angle, -1.98° for the shoulder protraction/retraction angle, -3.76° for the arm angle and -1.12° for the thoracic angle.</p>", "<title>Reliability results</title>", "<p>Table ##TAB##3##4## reports the descriptive statistics for each of the angles measured in the three sitting postures. The protraction/retraction angle demonstrated the largest variability (largest SD) in each sitting posture.</p>", "<p>All angles calculated from the repeated photographs demonstrated moderate to good agreement (See Table ##TAB##4##5##). Neither sitting posture nor gender significantly influenced the reliability of the angles calculated from the repeated photographs.</p>" ]

[ "<title>Discussion</title>", "<p>This paper reports the first known research to report on the validity of photographs of adolescent sitting posture, based on comparison with 'Gold Standard' Radiograph measures, the LODOX. The LODOX measures in this study provide unique information on the position of the spine in healthy adolescents sitting in a range of positions at computer workstations. The comparison between photographs and LODOX establishes, for the first time, the validity of photographs of external landmarks in measuring posture. Prior to this study, photographs have only been assumed to be representative of underlying spine position. Bland Altman analysis demonstrated a small bias and relatively wider limits for the shoulder protraction/retraction and arm angles. We have proposed that the circle fit function of the software may explain these variations. Given this explanation and the moderate to strong correlations between angles calculated from LODOX and digital photographs for all sitting postures, angles calculated from anatomical landmarks from photographs may be usefully proposed as a measure for gross estimates of spinal curvature. However spinal geometry still cannot be inferred from external postural analysis and should also be addressed in future studies.</p>", "<title>Photographs</title>", "<p>Based on the strong correlations between angles calculated from LODOX and digital photographs for all sitting postures, angles calculated from anatomical landmarks from photographs can be proposed as an alternative 'Gold Standard' for estimating sitting posture, on the assurance that they allow gross estimates of spinal curvature. Repeated angles calculated from photographs of subjects in the three sitting postures were also reliable, which suggests that one photograph only, taken in any sitting position, would provide an accurate representation of spinal posture for that individual. The PPAM method is cost and time-efficient, is non-invasive and incurs no exposure to radiation. Thus it is an ideal tool for use in large epidemiological studies of sitting posture in school settings.</p>", "<title>Measurement issues</title>", "<p>The researchers experienced difficulty in detecting the sticks on which the external markers for C7 and T8 were placed, with the Intellect 1.1.4 Software. We recommend that these sticks be covered in retro-reflective material in future studies, as this will ensure easier detection of the angle against which the marker is positioned on the body on the Radiograph image.</p>", "<title>Choice of angles</title>", "<p>All angles assessed in this study appear to be useful indicators of different aspects of stresses on the cervical and thoracic spine in sitting. The variability in the five angles across the three sitting postures was sufficiently large to enable future research to investigate issues such as the association between reported pain, muscle strength and length, and low, medium and high angles in each anatomical area.</p>", "<title>Angles</title>", "<p>The values for the sagittal head angle, cervical angle and protraction/retraction angle were similar to those published by Szeto et al [##REF##15996890##15##,##REF##11831210##16##], which suggested that adolescent angles were similar to adult angles, and that different sitting postures could be captured by the range of angles from photographs. The cervical angle demonstrated moderate reliability in the normal sitting posture with the second highest SEM value of all angles measured. The range for normal sitting posture is very wide compared to the upright and slouched postures which are more repeatable as they represent end of range positions. Thus, students were more likely to resume an extreme postural position (such as slouch or upright), than to accurately repeat the precise position of the spine and body segments in the normal posture range. The arm angle however, has not been reported on in the current published literature. We believed that it was an important angle as it may confound the values of the shoulder protraction/retraction angles. Shoulder protraction/retraction may be biomechanically affected by the position of the arm in glenohumeral flexion and extension [##UREF##8##38##]. This functional link could occur because of the structural linkage of multiple ligaments and muscles crossing the shoulder girdle complex [##UREF##8##38##]. The arm angle was thus measured to understand potential confounding effects in shoulder protraction/retraction angle reliability values. Both the arm angle and the protraction/retraction angle showed large variation in range in all three of the measured postures. This might be because subjects were not given a standardised position for their hands on the desks. We propose that for future studies, subjects keep their hands on actual keyboards for the duration of testing, as this might decrease the large variance in the arm angle and the protraction/retraction angle range. The thoracic angle showed very little change in the angle between postures. This may be because the thoracic region is the most inflexible region of the vertebral column. Therefore, since the body usually follows the path of least resistance, it may explain why relatively less movement was noted between the three postures.</p>", "<title>Clinical application</title>", "<p>Clinically it is important to know whether a patient is showing true progression in relation to a postural intervention. Based on the results of this study, the PPAM can be used in practice as a valid measure of sitting posture.</p>" ]

[ "<title>Conclusion</title>", "<p>The findings of this study suggest that photographs provide valid and reliable indicators of the position of the underlying spine in sitting. Clinically it is important to know whether a patient is showing true progression in relation to a postural intervention. Based on the results of this study, the PPAM can be used in practice as a valid measure of sitting posture.</p>" ]

[ "<p>This is an Open Access article distributed under the terms of the Creative Commons Attribution License (<ext-link ext-link-type=\"uri\" xlink:href=\"http://creativecommons.org/licenses/by/2.0\"/>), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.</p>", "<title>Background</title>", "<p>All the reported measures of sitting posture, as well as photographs, have one flaw, as these measures are external to the body. These measures use calculations from external bony landmarks to estimate spinal posture, on the understanding that what is being measured externally reflects the shape, health and performance of structures of the underlying spine. Without a comparative measure of the relative position of the structures of the spine, the validity of any external spinal posture measure cannot be established. This paper reports on a study which tests the validity of photographs to measure adolescent sitting posture.</p>", "<title>Methods</title>", "<p>The study was conducted in a laboratory at the Department of Human Biology, University of Cape Town. A random sample of 40 adolescents were recruited from the Cape metropolitan schools, to detect differences of three degrees or more between the repeated measures of upright, normal or slouched posture (photographs) and between the posture photographs and LODOX measures. Eligible participants were healthy male and female subjects aged 15 or 16 years old, in Grade 10, and who were undertaking Computer or Computype studies at their schools. Two posture measurement tools were used in the study, namely: Photographs were taken using the Photographic Posture Analysis Method (PPAM) and Radiograph<italic>s </italic>were taken using the LODOX (LODOX (Pty) Ltd) system. Subjects' posture was assessed in simulated computer workstations. The following angles were measured: the sagittal head angle, cervical angle, protraction/retraction angle, arm angle and the thoracic angle.</p>", "<title>Results</title>", "<p>Data from 39 subjects (19 males, 20 females) was used for analysis (17 15-year-olds (7 boys and 10 girls), 22 16-year-olds (12 boys and 10 girls)). All but one photographic angle showed moderate to good correlation with the LODOX angles (Pearson r values 0.67–0.95) with the exception being the shoulder protraction/retraction angle Pearson r values. Bland Altman limits of agreement illustrated a slight bias for all angles. The reliability study findings from repeated photographs demonstrated moderate to good correlation of all angles (ICC values 0.78–0.99).</p>", "<title>Conclusion</title>", "<p>The findings of this study suggest that photographs provide valid and reliable indicators of the position of the underlying spine in sitting. Clinically it is important to know whether a patient is showing true progression in relation to a postural intervention. Based on the results of this study, the PPAM can be used in practice as a valid measure of sitting posture.</p>" ]

[ "<title>Appendix 1</title>", "<p>The first step is to calculate of the actual position of C7, T8 and the manubrium. The positions are as follows:</p>", "<p></p>", "<p></p>", "<p></p>", "<p></p>", "<p></p>", "<p></p>", "<p>Now the angles can be calculated. We denote vectors in bold. The dot product is denoted with \"·\". The vector norm is denoted with \"|| ||\".</p>", "<p>Thoracic Angle</p>", "<p>Let <bold>T1 </bold>be the vector from the manubrium to C7:</p>", "<p></p>", "<p>Let <bold>T2 </bold>be the vector from the manubrium to T8:</p>", "<p></p>", "<p>Then the thoracic angle is: acos(<bold>T1·T2</bold>/(||<bold>T1</bold>|| × ||<bold>T2</bold>||))</p>", "<title>Study limitations</title>", "<p>The height and weight of the students were not measured in this study, but may be useful in future studies which also incorporate chair compatibility. A further limitation was that markers were placed manually on the C7 and T8 spinous processes of the spine and reliability of the manual placement of these markers were not tested. The circle fit process is not always accurate and therefore we recommend further development of the data processing software where this aspect of the data processing is standardised electronically.</p>", "<title>Recommendations for future studies</title>", "<p>Photographs measured using the PPAM system are valid indicators of adolescent sitting posture. When given standard instructions regarding assuming a sitting posture, subjects' posture is also reliable, when measured by repeated photographs.</p>", "<title>Competing interests</title>", "<p>The authors declare that they have no competing interests.</p>", "<title>Authors' contributions</title>", "<p>SVN conducted the study and drafted the manuscript. QL conceptualised the project, assisted with data collection and revised the manuscript. KGS assisted with statistical advice, study design assisted in writing the manuscript. CV assisted with project conceptualisation and LODOX imaging. KS wrote the software for angle analysis and wrote the data processing sections.</p>", "<title>Pre-publication history</title>", "<p>The pre-publication history for this paper can be accessed here:</p>", "<p><ext-link ext-link-type=\"uri\" xlink:href=\"http://www.biomedcentral.com/1471-2474/9/113/prepub\"/></p>" ]

[ "<title>Acknowledgements</title>", "<p>The authors thank the MRC for financial support and the Department of Education for assisting in the sample collection process. The authors also thank the research assistants for their help during the data collection phase.</p>" ]

[ "<fig position=\"float\" id=\"F1\"><label>Figure 1</label><caption><p>Data collection procedure for each subject (reliability and validity studies).</p></caption></fig>", "<fig position=\"float\" id=\"F2\"><label>Figure 2</label><caption><p>Diagrammatic representation of the angles measured.</p></caption></fig>", "<fig position=\"float\" id=\"F3\"><label>Figure 3</label><caption><p>LODOX machine.</p></caption></fig>", "<fig position=\"float\" id=\"F4\"><label>Figure 4</label><caption><p>LODOX image.</p></caption></fig>" ]

[ "<table-wrap position=\"float\" id=\"T1\"><label>Table 1</label><caption><p>The students' age, gender and posture</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td/><td align=\"left\" colspan=\"2\"><bold>15-year-olds</bold></td><td align=\"left\" colspan=\"2\"><bold>16-year-olds</bold></td><td/></tr><tr><td/><td colspan=\"4\"><hr/></td><td/></tr><tr><td align=\"left\"><bold>Posture</bold></td><td align=\"left\"><bold>Male</bold></td><td align=\"left\"><bold>Female</bold></td><td align=\"left\"><bold>Male</bold></td><td align=\"left\"><bold>Female</bold></td><td align=\"left\"><bold>Total</bold></td></tr></thead><tbody><tr><td align=\"left\"><bold>Slouched</bold></td><td align=\"left\">2</td><td align=\"left\">5</td><td align=\"left\">4</td><td align=\"left\">2</td><td align=\"left\"><bold>13</bold></td></tr><tr><td align=\"left\"><bold>Upright</bold></td><td align=\"left\">3</td><td align=\"left\">2</td><td align=\"left\">4</td><td align=\"left\">4</td><td align=\"left\"><bold>13</bold></td></tr><tr><td align=\"left\"><bold>Normal</bold></td><td align=\"left\">2</td><td align=\"left\">3</td><td align=\"left\">4</td><td align=\"left\">4</td><td align=\"left\"><bold>13</bold></td></tr><tr><td align=\"left\"><bold>Total</bold></td><td align=\"left\"><bold>7</bold></td><td align=\"left\"><bold>10</bold></td><td align=\"left\"><bold>12</bold></td><td align=\"left\"><bold>10</bold></td><td align=\"left\"><bold>39</bold></td></tr></tbody></table></table-wrap>", "<table-wrap position=\"float\" id=\"T2\"><label>Table 2</label><caption><p>Validity findings (Pearson r values)</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td align=\"left\"><bold>Validity</bold></td><td align=\"left\"><bold>Sagittal head angle</bold></td><td align=\"left\"><bold>Cervical angle</bold></td><td align=\"left\"><bold>Protraction/retraction angle</bold></td><td align=\"left\"><bold>Thoracic angle</bold></td><td align=\"left\"><bold>Arm angle</bold></td></tr></thead><tbody><tr><td align=\"left\">All angles measures from validity photos and LODOX</td><td align=\"center\">0.84</td><td align=\"center\">0.89</td><td align=\"center\">0.89</td><td align=\"center\">0.92</td><td align=\"center\">0.79</td></tr><tr><td align=\"left\"><bold>Upright (n = 13)</bold></td><td align=\"center\">0.73</td><td align=\"center\">0.89</td><td align=\"center\">0.88</td><td align=\"center\">0.81</td><td align=\"center\">0.76</td></tr><tr><td align=\"left\"><bold>Normal(n = 13)</bold></td><td align=\"center\">0.97</td><td align=\"center\">0.85</td><td align=\"center\">0.48</td><td align=\"center\">0.93</td><td align=\"center\">0.86</td></tr><tr><td align=\"left\"><bold>Slouched (n = 13)</bold></td><td align=\"center\">0.84</td><td align=\"center\">0.79</td><td align=\"center\">0.90</td><td align=\"center\">0.93</td><td align=\"center\">0.66</td></tr><tr><td align=\"left\"><bold>Female (n = 20)</bold></td><td align=\"center\">0.67</td><td align=\"center\">0.90</td><td align=\"center\">0.73</td><td align=\"center\">0.95</td><td align=\"center\">0.75</td></tr><tr><td align=\"left\"><bold>Male (n = 19)</bold></td><td align=\"center\">0.92</td><td align=\"center\">0.89</td><td align=\"center\">0.87</td><td align=\"center\">0.86</td><td align=\"center\">0.87</td></tr></tbody></table></table-wrap>", "<table-wrap position=\"float\" id=\"T3\"><label>Table 3</label><caption><p>Bland Altman Limits of Agreements</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td/><td align=\"center\"><bold>Limits of agreement ± 2 SD</bold></td></tr></thead><tbody><tr><td align=\"left\"><bold>Sagittal head angle</bold></td><td align=\"center\">-7.04–3.93</td></tr><tr><td align=\"left\"><bold>Cervical angle</bold></td><td align=\"center\">-8.04–6.73</td></tr><tr><td align=\"left\"><bold>Protraction/retraction angle</bold></td><td align=\"center\">-11.45–15.41</td></tr><tr><td align=\"left\"><bold>Thoracic angle</bold></td><td align=\"center\">-8.61–6.37</td></tr><tr><td align=\"left\"><bold>Arm angle</bold></td><td align=\"center\">-10.84–3.32</td></tr></tbody></table></table-wrap>", "<table-wrap position=\"float\" id=\"T4\"><label>Table 4</label><caption><p>The mean, SD and range values of the angles</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td align=\"left\"><bold>Angles</bold></td><td align=\"left\" colspan=\"3\"><bold>Normal</bold></td><td align=\"left\" colspan=\"3\"><bold>Upright</bold></td><td align=\"left\" colspan=\"3\"><bold>Slouched</bold></td></tr><tr><td/><td colspan=\"9\"><hr/></td></tr><tr><td/><td align=\"left\"><bold>Mean</bold></td><td align=\"left\"><bold>SD</bold></td><td align=\"left\"><bold>Range (degrees)</bold></td><td align=\"left\"><bold>Mean</bold></td><td align=\"left\"><bold>SD</bold></td><td align=\"left\"><bold>Range (degrees)</bold></td><td align=\"left\"><bold>Mean</bold></td><td align=\"left\"><bold>SD</bold></td><td align=\"left\"><bold>Range (degrees)</bold></td></tr></thead><tbody><tr><td align=\"left\"><bold>Sagittal head angle</bold></td><td align=\"left\">20.05</td><td align=\"left\">7.84</td><td align=\"left\">0 – 34.70</td><td align=\"left\">19.99</td><td align=\"left\">8.15</td><td align=\"left\">0.90 – 34.40</td><td align=\"left\">10.28</td><td align=\"left\">10.68</td><td align=\"left\">(-)15.90 – 34.20</td></tr><tr><td align=\"left\"><bold>Cervical angle</bold></td><td align=\"left\">47.66</td><td align=\"left\">9.75</td><td align=\"left\">21.90 – 62.90</td><td align=\"left\">52.72</td><td align=\"left\">11.18</td><td align=\"left\">22.30 – 71.30</td><td align=\"left\">21.49</td><td align=\"left\">27.57</td><td align=\"left\">(-)34.10 – 53.40</td></tr><tr><td align=\"left\"><bold>Protraction/retraction angle</bold></td><td align=\"left\">130.21</td><td align=\"left\">25.77</td><td align=\"left\">65.30 – 178.70</td><td align=\"left\">124.76</td><td align=\"left\">20.36</td><td align=\"left\">76.50 – 159.80</td><td align=\"left\">145.68</td><td align=\"left\">20.62</td><td align=\"left\">103.70 – 208.70</td></tr><tr><td align=\"left\"><bold>Thoracic angle</bold></td><td align=\"left\">63.25</td><td align=\"left\">8.57</td><td align=\"left\">49.50 – 89.20</td><td align=\"left\">61.37</td><td align=\"left\">11.76</td><td align=\"left\">40.80 – 97.60</td><td align=\"left\">61.46</td><td align=\"left\">8.88</td><td align=\"left\">39.30 – 78.10</td></tr><tr><td align=\"left\"><bold>Arm angle</bold></td><td align=\"left\">23.46</td><td align=\"left\">12.75</td><td align=\"left\">(-)5.00 – 50.30</td><td align=\"left\">24.21</td><td align=\"left\">12.09</td><td align=\"left\">3.30 – 60.90</td><td align=\"left\">32.72</td><td align=\"left\">10.34</td><td align=\"left\">14.50 – 48.80</td></tr></tbody></table></table-wrap>", "<table-wrap position=\"float\" id=\"T5\"><label>Table 5</label><caption><p>Reliability Findings: ICC's and SEM values of all angles, postures and genders</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td align=\"center\"><bold>Reliability</bold></td><td align=\"center\"><bold>Sagittal head angle</bold></td><td align=\"center\"><bold>Cervical angle</bold></td><td align=\"center\"><bold>Protraction/retraction angle</bold></td><td align=\"center\"><bold>Thoracic angle</bold></td><td align=\"center\"><bold>Arm angle</bold></td></tr></thead><tbody><tr><td align=\"center\">95% Lower and Upper interval</td><td align=\"center\">0.82 – 0.96</td><td align=\"center\">0.86 – 0.96</td><td align=\"center\">0.74 – 0.93</td><td align=\"center\">0.60 – 0.97</td><td align=\"center\"><bold>0.95–0.94</bold></td></tr><tr><td align=\"center\">All angles (5 reliability photos)</td><td align=\"center\">0.98</td><td align=\"center\">0.98</td><td align=\"center\">0.94</td><td align=\"center\">0.96</td><td align=\"center\">0.99</td></tr><tr><td align=\"center\"><bold>Upright (n = 13)</bold></td><td align=\"center\">0.97 (0.93–0.99)</td><td align=\"center\">0.98 (0.55–0.93)</td><td align=\"center\">0.92 (0.79–0.93)</td><td align=\"center\">0.97 (0.83–0.99)</td><td align=\"center\">0.99 (0.97–0.99)</td></tr><tr><td align=\"center\"><bold>Normal (n = 13)</bold></td><td align=\"center\">0.97 (0.92–0.97)</td><td align=\"center\">0.78 (0.56–0.94)</td><td align=\"center\">0.91 (0.78–0.92)</td><td align=\"center\">0.92 (0.84–0.98)</td><td align=\"center\">0.98 (0.96–0.98)</td></tr><tr><td align=\"center\"><bold>Slouched (n = 13)</bold></td><td align=\"center\">0.98 (0.99–0.95)</td><td align=\"center\">0.98 (0.96–0.98)</td><td align=\"center\">0.99 (0.97–0.99)</td><td align=\"center\">0.97 (0.93–0.99)</td><td align=\"center\">0.98 (0.95–0.98)</td></tr><tr><td align=\"center\"><bold>Female (n = 20)</bold></td><td align=\"center\">0.96 (0.92–0.96)</td><td align=\"center\">0.99 (0.98–0.99)</td><td align=\"center\">0.94 (0.88–0.97)</td><td align=\"center\">0.94 (0.89–0.97)</td><td align=\"center\">0.98 (0.97–0.99)</td></tr><tr><td align=\"center\"><bold>Male (n = 19)</bold></td><td align=\"center\">0.99 (0.97–0.99)</td><td align=\"center\">0.96 (0.91–0.98)</td><td align=\"center\">0.95 (0.88–0.96)</td><td align=\"center\">0.97 (0.94–0.98)</td><td align=\"center\">0.97 (0.95–0.98)</td></tr><tr><td align=\"center\"><bold>SEM </bold>(in degrees)</td><td align=\"center\">3.50</td><td align=\"center\">8.06</td><td align=\"center\">11.09</td><td align=\"center\">4.04</td><td align=\"center\">3.33</td></tr></tbody></table></table-wrap>", "<table-wrap position=\"float\" id=\"T6\"><label>Table 6</label><caption><p>Declaration of Symbols</p></caption><table frame=\"hsides\" rules=\"groups\"><tbody><tr><td align=\"left\">C7'_x, C7'_y</td><td align=\"left\">X coordinate of C7 marker</td></tr><tr><td align=\"left\">T8'_x, T8'_y</td><td align=\"left\">X coordinate of T8 marker</td></tr><tr><td align=\"left\">M'_x, M'_y</td><td align=\"left\">X coordinate of Manubrium marker</td></tr><tr><td align=\"left\">θ_C7</td><td align=\"left\">Smallest angle between horizontal and C7 marker stem</td></tr><tr><td align=\"left\">θ_T8</td><td align=\"left\">Smallest angle between horizontal and T8 marker stem</td></tr><tr><td align=\"left\">θ_M</td><td align=\"left\">Smallest angle between horizontal and Manubrium marker stem</td></tr><tr><td align=\"left\">L_C7</td><td align=\"left\">Length of C7 stem</td></tr><tr><td align=\"left\">L_T8</td><td align=\"left\">Length of T8 stem</td></tr><tr><td align=\"left\">L_m</td><td align=\"left\">Length of Manubrium stem</td></tr></tbody></table></table-wrap>" ]

[ "<disp-formula>C7_x= C7'_x+ L_C7cos(θ_C7)</disp-formula>", "<disp-formula>C7_y= C7'_y- L_C7sin(θ_C7)</disp-formula>", "<disp-formula>T8_x= T8'_x+ L_T8cos(θ_T8)</disp-formula>", "<disp-formula>T8_y= T8'_y- L_T8sin(θ_T8)</disp-formula>", "<disp-formula>M_x= M'_x- L_mcos(θ_C7)</disp-formula>", "<disp-formula>M_y= M'_y- L_msin(θ_C7)</disp-formula>", "<disp-formula><bold>T1 </bold>= {C7_x- M_x; C7_y- M_y}</disp-formula>", "<disp-formula><bold>T2 </bold>= {T8_x- M_x; T8_y- M_y}</disp-formula>" ]

[]

[]

[]

[]

[]

[]

[ "<graphic xlink:href=\"1471-2474-9-113-1\"/>", "<graphic xlink:href=\"1471-2474-9-113-2\"/>", "<graphic xlink:href=\"1471-2474-9-113-3\"/>", "<graphic xlink:href=\"1471-2474-9-113-4\"/>" ]

[]

{ "acronym": [], "definition": [] }

[ "<title>Background</title>", "<p>Osteoarthritis (OA) is the most common joint disorder and is responsible for substantial economic, social and psychological costs. While the prevalence of OA is variable based on how the disease is defined, it has been said that the majority of individuals over the age of 65 years living in the western world demonstrate radiographic evidence of disease [##REF##9588729##1##, ####REF##10750883##2##, ##REF##2746583##3####2746583##3##]. Conventional radiography has been, and continues to be, the primary imaging modality used in the evaluation of OA, both in terms of diagnosis and monitoring of disease progression. In the knee joint, osteoarthritic features visible on radiographs include joint space narrowing, osteophytosis, subchondral osteosclerosis and subchondral cysts.</p>", "<p>The measurement of the separation between the distal femur and the proximal tibia, joint space width (JSW), has become the standard tool for the assessment of knee OA progression [##REF##12192261##4##]. Both the fluoroscopic and non-fluoroscopic acquisition of radiographs have allowed for the evaluation of JSW revealing good precision of measurement [##REF##15880526##5##, ####REF##12571847##6##, ##REF##7799359##7##, ##REF##10381056##8##, ##REF##12605275##9####12605275##9##]. Standard X-rays acquired using the non-fluoroscopic fixed-flexion technique can be as reproducible as fluoroscopic techniques (root-mean-square standard deviation = 0.1 mm) with the added advantages of lower costs and considerably less radiation dose [##REF##12605275##9##,##REF##15150666##10##], although they have been shown to be less sensitive to change in knee OA patients than the Lyon-Schuss fluoroscopic technique [##UREF##0##11##]. Also, it has been shown that the reproducibility of measurement of minimum joint space width (mJSW) is better when using an automated computer algorithm as compared to manual methods such as a hand-held lens [##REF##12605275##9##,##REF##15150666##10##,##REF##10757609##12##,##REF##7492235##13##].</p>", "<p>Cartilage thinning is the signature feature of knee OA and JSW measurements should be an indirect measure of the articular cartilage thickness in the joint. The emergence of magnetic resonance imaging as an important tool in the visualization of articular cartilage together with dedicated image analysis software permits one to quantify cartilage volume and thickness directly [##REF##9678042##14##, ####REF##15643573##15##, ##REF##17762608##16##, ##REF##16713720##17##, ##REF##16515760##18##, ##REF##17506024##19####17506024##19##].</p>", "<p>Despite the advances in techniques used to evaluate joint space width and cartilage volume and thickness, their use in clinical studies has been limited to the longitudinal measurement of disease progression (i.e. change over time) rather than the diagnosis of OA in which values are compared to a standard as has been done in the diagnosis of osteoporosis where there are age dependent normal values of bone mineral density [##REF##14558088##20##, ####UREF##1##21##, ##REF##11710712##22##, ##REF##15479900##23####15479900##23##]. The potential for age dependent normal values of mJSW and cartilage morphometry may also be the case, as investigated here. In addition, since joint space width measures are thought to be a surrogate measure of cartilage thickness, one might hypothesize that these variables are correlated with one another. Although one study investigating the relationship between these variables has been previously conducted in osteoarthritic individuals, the strength of this relationship has not been explored in healthy individuals. A healthy reference is needed in order to determine how these relationships compare or change under osteoarthritic conditions. Thus, the purpose of this pilot study was to estimate reference values of medial minimum joint space width and cartilage morphometry in healthy males and females between the ages of 20 and 69 years using standard radiography and peripheral magnetic resonance imaging (pMRI) and to further investigate the correlation between medial tibial cartilage thickness and medial minimum joint space width.</p>" ]

[ "<title>Methods</title>", "<p>Healthy volunteers between 20 and 69 years of age with no known bone or joint disease were recruited to undergo a knee X-ray and a peripheral magnetic resonance scan of the same knee via locally posted advertisements and word of mouth. Individuals were asked to respond to the advertisements by telephone and were posed a number of screening questions to ensure that study inclusion and exclusion criteria were met. Those excluded from participating (N = 8) were those who a) stated they were currently experiencing knee pain, b) stated they had been previously diagnosed with a bone or joint disease (i.e. rheumatoid arthritis, osteoporosis), or c) had previously sustained a knee injury and/or had undergone any knee surgery (i.e. arthroscopy, menisectomy, etc.). Only those with radiographically normal X-rays were included in the analyses. All participants were required to sign a consent form that had been approved by the Research Ethics Board at St. Joseph's Healthcare. In addition to completing the study consent form, individuals were required to complete a questionnaire which asked questions pertaining to his/her medical history, medications and exercise activity.</p>", "<title>Plain X-ray</title>", "<p>Study participants underwent a single knee X-ray of the non-dominant knee acquired in the fixed-flexion position [##REF##12605275##9##,##REF##15150666##10##]. In this position, participants are required to stand, with their weight distributed equally between their legs, on a piece of cardboard such that both great toes are touching the vertical X-ray table and feet are externally rotated by approximately 10°. Both feet are traced onto cardboard should the foot map be needed for use in successive X-rays. Facing the vertical X-ray table and holding the sides for balance and support, subjects are asked to bend their knees slightly such that both their patellas and thighs are pressed tightly against the table. In doing so, the position of the femur and tibia are fixed, and thus, so is the degree of knee flexion. The posteroanterior X-ray beam is directed parallel to the tibial plateau at a 10° caudal beam alignment.</p>", "<p>Radiographs were graded independently by two musculoskeletal radiologists according to the Kellgren-Lawrence (K-L) scoring system [##REF##13498604##24##]. This grade was used to confirm or refute the presence of knee OA. Those assigned a grade of 0 or 1 on the scale were included. Individuals whose X-rays scored ≥ 2 on the K-L scale were excluded from the analyses because they had evidence of knee OA.</p>", "<p>X-ray films were subsequently digitized using a Sierra plus™ digitizer (Vidar Systems Corporation, Herndon, VA, USA) at an isotropic pitch of 84.7 μm and a 12 bit grey scale resolution. The digitized images were further analyzed for mJSW in the medial compartment of the knee using a automated computer algorithm, details of which have been described previously [##REF##10757609##12##]. The reproducibility of this analysis technique has been shown to be very good (RMSSD = 0.15 mm; CV = 3.31% in healthy individuals). An analyzed radiograph is depicted in Figure ##FIG##0##1##. This program delineates the bony margins of the femoral condyles and the tibial plateau. In approximately 3% of radiographs analyzed for mJSW, user intervention was required to slightly alter the delineations drawn by the computer algorithm.</p>", "<title>pMRI</title>", "<p>Peripheral MR scans were acquired using a 1.0 Tesla peripheral MRI (pMRI) system (OrthOne™, ONI Inc., Wilmington, MA, USA). Subjects were seated in the scanning chair with their knee fully extended and centred within the iso-centre of the 180 mm removable quadrature volume transmit-receive coil. Padding was placed around the knee, thigh and leg to limit the potential for movement inside the magnet. All study participants were positioned and scanned by the same technologist.</p>", "<p>Sagittal gradient-echo and axial fast spin-echo localizer scans were performed (total scan time 2–3 minutes). Following this, a fat-saturated spoiled gradient recalled acquisition in the steady state (SPGR) was performed in the sagittal plane using the following parameters: TR 60 ms; TE 12.4 ms (or minimum); flip angle 40°; bandwidth 30 kHz; matrix 512 × 256 (frequency × phase); 1 excitation; field of view 150 mm; slice thickness 1.5 mm; 56 to 64 partitions depending on patient size; scan time 15–16 minutes. Images were transferred to an independent workstation where they were saved in DICOM format. Upon completing the acquisition of all images, one trained technician conducted analyses to quantify the cartilage morphology of the medial tibia (MT) using a reproducible, validated proprietary segmentation software program (Chondrometrics GmbH, Ainring, Germany) [##REF##16126797##25##, ####REF##16978886##26##, ##UREF##2##27##, ##REF##17260363##28##, ##REF##17654593##29####17654593##29##]. Cartilage segmentation was conducted on a slice-by-slice basis (number of slices was dependent upon patient size) by manual tracing the bone-cartilage interface and the cartilage surface of the entire MT [##REF##16126797##25##,##REF##16978886##26##]. This segmentation algorithm has previously been validated in the pMRI [##REF##16978886##26##]. After segmenting all MT plates, images were reviewed a second time for the purposes of quality assurance and adjustments in segmentation were made if deemed necessary. Total volume of MT cartilage (VC), cartilage volume normalized to medial tibial bone size (VCtAB), and cartilage thickness over the total area of medial tibial subchondral bone (ThCtAB) were computed [##REF##16126797##25##,##REF##16730462##30##]. An example of a single slice of segmented MT cartilage is shown in Figure ##FIG##1##2##.</p>" ]

[ "<title>Results</title>", "<p>In total, 119 healthy individuals with no history of knee pain or injury and without a bone or joint disease agreed to participate in the X-ray portion of the study. Of these, 73 were female and 46 were male and all but 3 were Caucasian. Demographic data are presented in Table ##TAB##0##1##.</p>", "<p>K-L grading of X-rays revealed that, of the women, 49 had radiographic scores of 0 while the remaining 24 had scores of 1. Grading for males revealed that 31 had K-L scores of 0 and 15 had a score of 1. Of the 119 individuals who participated in the X-ray portion of the study, 86 also underwent a pMRI scan of the same knee, 50 of whom were female and 36 of whom were male. It should be noted that the mean age and BMI of individuals who received pMR scans was very similar to the entire study population (mean age 38.3 yrs, mean BMI 25.3 kg/m²) suggesting that results may be generalizable to the overall study group.</p>", "<p>Subjects were subdivided according to decade of life for the analysis of mJSW, thus treating age as a categorical variable. The purpose of doing so was to determine if there was an identifiable decade where initial changes (i.e. decreases) in mJSW could be detected. In addition, other groups performed analyses by age decades thus allowing comparisons to be made between studies [##REF##9893570##31##, ####REF##8915618##32##, ##REF##1747696##33####1747696##33##]. The mean (SD), range, minimum and maximum mJSW data for each of these age groups were calculated and are presented in Table ##TAB##1##2##.</p>", "<p>The descriptive statistics do not appear to show any differences in mean mJSW values between decades in either males or females. In fact, these cross-sectional data suggest that a mean (SD) \"normal\" value of mJSW for healthy women is 4.8 (0.7) mm while in healthy men this value is larger at 5.7 (0.8) mm. This was supported by results from an ANOVA analysis in which no significant differences in mJSW were found between age groups in either men or women even after considering BMI as a covariate (p &gt; 0.05). The only significant difference identified was that between genders where an ANOVA analysis performed with BMI, age and gender as covariates revealed that males have significantly larger mJSW values than females (p &lt; 0.05).</p>", "<p>Analyses were repeated with age as a continuous rather than a categorical variable with age, BMI and gender considered independent variables. While age and BMI were not predictive of mJSW, gender was again found to be significant with healthy males having significantly larger mJSW values compared to healthy females (β regression coefficient = 0.84, p &lt; 0.001).</p>", "<p>Cartilage analyses were also conducted in an attempt to determine \"normal\" cartilage volume and thickness values in healthy males and females of different age groups. These data are presented in Table ##TAB##2##3##. The mean (SD) medial tibial cartilage volume normalized to the area of subchondral bone was 1.50 (0.19) μL/mm²in females while in males it was 1.77 (0.24) μL/mm². Corresponding mean values for cartilage thickness over the entire bone surface were 1.45 (0.19) mm and 1.71 (0.24) mm in females and males, respectively.</p>", "<p>Just as with mJSW values, medial tibial cartilage volume and thickness data did not appear to differ significantly between age groups for healthy males or females. This observation was confirmed by ANOVA analyses which revealed no significant differences in VCtAB or ThCtAB values between different age groups (p &gt; 0.05). However, significant differences were found between genders with males consistently having thicker cartilage than their female counterparts.</p>", "<p>While investigating the relationship between medial tibial cartilage morphometry and age as a continuous variable in males, the relationship did not change from that considering age as a categorical variable. Regression analyses with BMI as a covariate showed that medial tibial cartilage volume, normalized to bone area, and thickness did not decrease significantly with age (p &gt; 0.05). However, this was not the case for females. Both cartilage volume and thickness in the medial tibia appeared to decrease with ageing showing standardized regression coefficients of -0.41 (p = 0.008) and -0.37 (p = 0.015), respectively.</p>", "<p>Analyses were also conducted to investigate the relationship between medial tibial cartilage morphometry and medial mJSW, since mJSW is considered to be a surrogate measure of cartilage thickness. Between cartilage volume (VC) and mJSW, correlation analyses revealed a correlation coefficient of 0.67 while the correlation between VCtAB and mJSW was 0.69 and the correlation between ThCtAB and mJSW was also 0.69. These results suggest that approximately 47% of the variation in mJSW can be explained by the variation in cartilage thickness of the medial tibia.</p>" ]

[ "<title>Discussion</title>", "<p>The primary purpose of establishing normal values of mJSW in a healthy population of males and females is to provide age-specific references to which osteoarthritic values can be compared. In addition, it is important to determine if mJSW values appear to decrease with age in a healthy population or if, indeed, this is characteristic of only those affected by knee OA. We also investigated the correlation between mJSW and medial tibial cartilage morphometry in this healthy population. Results of this pilot study appear to suggest that mJSW values are not significantly different between younger and older individuals without radiographic evidence of OA as shown by mean (SD) values of 4.8 (0.7) mm and 5.7 (0.8) mm in females and males, respectively. Individuals included in these analyses were those with K-L grades of 0 and 1 as was the case in a recent study by Conrozier et al. [##REF##15880526##5##]. While it may be argued that a K-L grade of 1 may correspond to early OA, the definition of this categorization states the doubtful presence of osteophyte without regard for joint space narrowing. This would support the notion that cartilage thickness measurements would not be affected by the inclusion of those with K-L grade 1. To verify this, additional analyses including only those with K-L grades of 0 (grade 1 excluded) were performed and results did not differ significantly from those which included both K-L grades 0 and 1.</p>", "<p>Results from this study suggest that there is no identifiable decade of life when one might expect joint space width to narrow. When considered as a continuous variable, age was not found to be significantly related to mJSW in either males or females again supporting the notion that joint space narrowing may not simply be a consequence of aging. To confirm this, however, we recognize that a longitudinal study collecting data over decades would be required and therefore was not feasible at this point in time.</p>", "<p>Despite the fact that there are few studies which are longitudinal in nature, there are a few cross-sectional studies which have investigated the relationship between JSW measurements and age with methodologies slightly different than the ones used in the present study. For example, a study of healthy young adults 16–22 years of age reported mean (SD) medial mJSW values of 4.74 (0.94) and 5.65 (0.93) in females and males respectively, results much like those of similar aged participants in this study [##REF##17216807##34##]. Also like our study, JSW values in 125 healthy individuals between 40 and 75 years of age were not found to decrease with increasing decade of life with mean values ranging from 4.6 – 5.0 mm in females and 5.0 – 5.5 mm in males [##REF##9893570##31##]. In both of these studies, males were generally found to have larger JSWs compared to females, results which are also consistent with those reported here [##REF##9893570##31##,##REF##17216807##34##]. In contrast, studies conducted by Dacre et al. and Sargon et al. showed that JSW decreased with increasing age group, although both of these studies were cross sectional in nature and methodological differences existed including X-ray acquisition technique (weight bearing vs. non-weight bearing), joint space analysis (manual vs. automated, joint space area (mm²) vs. mJSW [##REF##1747696##33##]) and the symptomatic nature of patients [##REF##8915618##32##,##REF##1747696##33##]. While our results and those of other cross-sectional studies of healthy individuals suggest mJSW values remain constant, other results suggesting the opposite justify the need for large-scale, cross-sectional and longitudinal population-based data of healthy individuals acquired using the most reproducible techniques [##REF##11710712##22##,##REF##7966075##35##].</p>", "<p>Although there is a paucity of radiographic data from healthy individuals conducted over time, a study conducted by Conrozier et al. examined longitudinal changes in mJSW in individuals reporting chronic knee pain (&gt;3 months) but lacking radiographic evidence of knee OA (K-L grade ≤ 1), as was considered in the present study. These authors reported a mean (SD) annual rate of joint space narrowing of 0.05 (0.22) mm [##REF##15880526##5##]. However, the symptomatic nature of the participants may be indicative of cartilage lesions that may not be radiographically detectable, as reported by Ding et al. in patients with K-L grade 1, thereby questioning the status of this sample as a \"healthy\" population [##REF##15727885##36##]. In addition, this study did not report whether this change was statistically significant from baseline to one-year follow-up. In fact, such small changes in joint space width are often within the range of reproducibility error of measurement [##REF##15880526##5##].</p>", "<p>Other studies of medial JSW values in healthy individuals have reported average values for the entire populations under investigation but have not analyzed these measurements as they varied with age or sex [##REF##8915618##32##,##REF##9876394##37##]. Dacre et al., for instance, reported the mean medial JSW acquired from non-weight-bearing radiographs to be 5.73 (0.15) mm in females and 7.03 (0.12) mm in males, results which are 17% and 19% larger than those of the present study, respectively [##REF##1747696##33##]. However, joint space width values acquired from non weight-bearing X-rays may be larger than those acquired from weight-bearing ones, suggesting that these results may, indeed, be consistent with those of the present study [##REF##16483906##38##].</p>", "<p>Given that mJSW is a surrogate measure of cartilage thickness, one would hypothesize that these variables would be correlated with one another. However, it is widely understood that joint space width measurements reflect only a thickness measure at one specified location in the joint and may include tissues such as menisci and synovial fluid, findings that are supported by previously published studies [##REF##9394671##39##, ####REF##10484216##40##, ##REF##12681955##41##, ##REF##16508930##42##, ##REF##16868968##43##, ##REF##16801692##44####16801692##44##]. Our results revealed that the variance in medial tibial cartilage thickness, normalized to bone area, can explain less than half of the variation in medial mJSW. It must be noted here, however, that medial femoral cartilage thickness was not analyzed in this study population. This variable would certainly also account for some of the variation in joint space, although cartilage thickness in one plate is not highly correlated with cartilage in another plate [##REF##15975965##45##]. Analyses previously conducted in an osteoarthritic population where both medial femoral and tibial cartilage were examined suggested the variation in cartilage thickness accounted for 54% of mJSW [##UREF##3##46##].</p>", "<p>The issue of whether sex and age are significantly related to cartilage volume and thickness has been the subject of many studies [##REF##16861710##47##, ####REF##12810930##48##, ##REF##11357452##49##, ##REF##9862027##50##, ##REF##17321168##51####17321168##51##]. Mixed results have been reported with respect to gender differences in cartilage volume and thickness, although our results revealed that males have significantly larger mean tibial cartilage volume and thickness just as was the case with mJSW. Similarly, studies by Faber et al. and Cicuttini et al. reported significantly larger mean cartilage volume values in healthy males compared to healthy females [##REF##11357452##49##,##REF##17321168##51##, ####REF##10329301##52##, ##REF##11908572##53####11908572##53##]. While Cicuttini et al. also reported significantly larger medial tibial cartilage thickness values in males of the same population, although thickness was not assessed directly but calculated as volume per unit area, Faber et al. did not find such significant differences in thickness between genders [##REF##11357452##49##,##REF##17321168##51##,##REF##11908572##53##]. For instance, in our study, men had 18% more medial tibial cartilage thickness compared to women while Faber demonstrated that men had 13.3% thicker cartilage than women, although this difference was not statistically significant. Discrepancies in results from these studies likely exist because of differences in sample populations (i.e. age, definition of \"healthy\") and the relatively small sample size.</p>", "<p>In the current study, in contrast to results investigating age as a categorical variable, increasing age (as a continuous variable) was found to be associated with less cartilage volume normalized to total bone area (β = -0.41) and thickness (β = -0.37) in females after adjusting for BMI. Such age-related differences in medial tibial cartilage volume and thickness were not observed in males. While this may be related to the relatively small number of healthy males over 50 years of age in our study sample, it is also possible that inconsistencies between males and females may be related to hormonal changes which occur during menopause of which there are no comparable changes that occur in men. This is similar to the BMD findings in osteoporosis [##REF##15640270##54##,##REF##17303218##55##]. Other cross-sectional studies which have investigated the relationship between age and cartilage volume and thickness have shown inconsistent results with one reporting a significant decrease in medial cartilage thickness, but not volume, with age [##REF##15769915##56##], while another reports no significant changes in tibial cartilage thickness with age [##REF##11710712##22##]. However, one should be cautious about the interpretation of these results since these data are cross-sectional in nature and do not reflect changes in a single person over time but comparisons between different individuals.</p>", "<p>Three studies have reported longitudinal changes in cartilage volume for healthy individuals and have shown that cartilage volume does, indeed, decrease with aging [##REF##15640270##54##,##REF##12595619##57##,##REF##15020341##58##]. In healthy males (N = 28, mean age 52 years), the mean annual reduction in tibial cartilage volume was found to be 2.8% (95% CI = 0.2% to 5.5%) [##REF##15640270##54##]. In healthy postmenopausal females, the average annual decrease in total tibial cartilage volume was similar at 2.4% (3.2%) [##REF##15020341##58##]. What is notable in these two studies is the mean age of subjects being investigated was over 50 years. To this point, there is only one study investigating longitudinal changes in a population including younger adults. Ding et al. demonstrated a significant association between age and loss of cartilage volume by approximately 1.5 – 4.2% per annum in individuals between the ages of 26 and 60 years, with a higher rate of loss in females as compared to males [##REF##16861710##47##]. However, despite these seemingly age-related declines, it is still plausible that these values lie within what may be considered to be a \"normal\" or \"healthy\" range.</p>", "<p>It is important to recognize that there are a number of methodological limitations to this study including the small sample sizes, particularly in some age groups, the cross-sectional nature of the data and the lack of medial femoral cartilage analyses. Despite these limitations, results suggest that mJSW values do not decrease with increasing age group in males or females between the ages of 20 and 69 years. This information may be helpful in defining radiographic joint space width references for comparison with those suspected of having knee OA. These results also suggest that there is no defined decade at which point joint space width decreases. Cartilage volume and thickness did not decrease with increasing age in males as was the case with mJSW. However, the observation that cartilage volume and thickness decreased with ageing in females may support the role of estrogen in cartilage physiology, although the exact mechanism remains unknown. It is also possible that tissue other than medial tibial cartilage may play a more significant role in joint space narrowing than in males, although this has not yet been shown.</p>" ]

[ "<title>Conclusion</title>", "<p>The results of this cross-sectional pilot study investigating the knees of healthy individuals suggest that mJSW measures from plain radiographs remain relatively constant through the third to seventh decades of life. The lack of significant declines associated with ageing also suggests that mJSW values may be helpful for comparisons with those suspected of having knee OA. In males, these results are supported by cartilage volume and thickness data which also remain fairly constant throughout the middle ages. However, decreasing values of cartilage thickness and volume in females over the ages suggest that discrepancies with mJSW results may be due to tissue other than medial tibial cartilage or another mechanism yet to be fully elucidated.</p>" ]

[ "<p>This is an Open Access article distributed under the terms of the Creative Commons Attribution License (<ext-link ext-link-type=\"uri\" xlink:href=\"http://creativecommons.org/licenses/by/2.0\"/>), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.</p>", "<title>Background</title>", "<p>The clinical use of minimum joint space width (mJSW) and cartilage volume and thickness has been limited to the longitudinal measurement of disease progression (i.e. change over time) rather than the diagnosis of OA in which values are compared to a standard. This is primarily due to lack of establishment of normative values of joint space width and cartilage morphometry as has been done with bone density values in diagnosing osteoporosis. Thus, the purpose of this pilot study is to estimate reference values of medial joint space width and cartilage morphometry in healthy individuals of all ages using standard radiography and peripheral magnetic resonance imaging.</p>", "<title>Design</title>", "<p>For this cross-sectional study, healthy volunteers underwent a fixed-flexion knee X-ray and a peripheral MR (pMR) scan of the same knee using a 1T machine (ONI OrthOne™, Wilmington, MA). Radiographs were digitized and analyzed for medial mJSW using an automated algorithm. Only knees scoring ≤1 on the Kellgren-Lawrence scale (no radiographic evidence of knee OA) were included in the analyses. All 3D SPGRE fat-sat sagittal pMR scans were analyzed for medial tibial cartilage morphometry using a proprietary software program (Chondrometrics GmbH).</p>", "<title>Results</title>", "<p>Of 119 healthy participants, 73 were female and 47 were male; mean (SD) age 38.2 (13.2) years, mean BMI 25.0 (4.4) kg/m². Minimum JSW values were calculated for each sex and decade of life. Analyses revealed mJSW did not significantly decrease with increasing decade (p &gt; 0.05) in either sex. Females had a mean (SD) medial mJSW of 4.8 (0.7) mm compared to males with corresponding larger value of 5.7 (0.8) mm. Cartilage morphometry results showed similar trends with mean (SD) tibial cartilage volume and thickness in females of 1.50 (0.19) μL/mm²and 1.45 (0.19) mm, respectively, and 1.77 (0.24) μL/mm²and 1.71 (0.24) mm, respectively, in males.</p>", "<title>Conclusion</title>", "<p>These data suggest that medial mJSW values do not decrease with aging in healthy individuals but remain fairly constant throughout the lifespan with \"healthy\" values of 4.8 mm for females and 5.7 mm for males. Similar trends were seen for cartilage morphology. Results suggest there may be no need to differentiate a t-score and a z-score in OA diagnosis because cartilage thickness and JSW remain constant throughout life in the absence of OA.</p>" ]

[ "<title>Competing interests</title>", "<p>Dr. Felix Eckstein is the founder and CEO of Chondrometrics GmbH. This is the company responsible for providing the cartilage segmentation software that was used in this study. There are no other conflicts of interest to mention.</p>", "<title>Authors' contributions</title>", "<p>KB, PB, JDA and CW conceived of the study, and participated in its design and coordination and helped to draft the manuscript. JD assisted in the automated analyses of the radiographs while MP and JO read and scored radiographs using the Kellgren-Lawrence scale while FE assisted in the cartilage morphometry analyses of the pMR images. All authors read and approved the final manuscript.</p>", "<title>Pre-publication history</title>", "<p>The pre-publication history for this paper can be accessed here:</p>", "<p><ext-link ext-link-type=\"uri\" xlink:href=\"http://www.biomedcentral.com/1471-2474/9/119/prepub\"/></p>" ]

[ "<title>Acknowledgements</title>", "<p>The authors of the study would like to acknowledge the McMaster Institute for Applied Radiation Sciences and the Father Sean O'Sullivan Research Centre</p>" ]

[ "<fig position=\"float\" id=\"F1\"><label>Figure 1</label><caption><p>A digitized radiograph analyzed for minimum joint space width using an automated computer algorithm (medial compartment on left).</p></caption></fig>", "<fig position=\"float\" id=\"F2\"><label>Figure 2</label><caption><p>Medial tibial cartilage segmented from a single sagittal slice of an MR image acquired from a healthy knee.</p></caption></fig>" ]

[ "<table-wrap position=\"float\" id=\"T1\"><label>Table 1</label><caption><p>Study population demographic statistics</p></caption><table frame=\"hsides\" rules=\"groups\"><tbody><tr><td align=\"left\"><italic>N (females)</italic></td><td align=\"left\"><italic>119 (73)</italic></td></tr><tr><td align=\"left\">Age (SD) (yrs)</td><td align=\"left\">38.2 (13.2)</td></tr><tr><td align=\"left\">BMI (SD) (kg/m²)</td><td align=\"left\">25.0 (4.4)</td></tr><tr><td align=\"left\">K-L grades (N):</td><td/></tr><tr><td align=\"left\"> 0</td><td align=\"left\">80</td></tr><tr><td align=\"left\"> 1</td><td align=\"left\">39</td></tr></tbody></table></table-wrap>", "<table-wrap position=\"float\" id=\"T2\"><label>Table 2</label><caption><p>mJSW data per sex and decade in healthy individuals</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td/><td align=\"center\"><bold>Age Group (yrs)</bold></td><td align=\"center\"><bold>N</bold></td><td align=\"center\"><bold>K-L Grade 0 (%)</bold></td><td align=\"center\"><bold>Mean (mm)</bold></td><td align=\"center\"><bold>SD (mm)</bold></td></tr></thead><tbody><tr><td align=\"left\"><bold>Females</bold></td><td align=\"center\">20 – 29</td><td align=\"center\">22</td><td align=\"center\">91</td><td align=\"center\">5.06</td><td align=\"center\">.56</td></tr><tr><td/><td align=\"center\">30 – 39</td><td align=\"center\">15</td><td align=\"center\">80</td><td align=\"center\">4.62</td><td align=\"center\">.66</td></tr><tr><td/><td align=\"center\">40 – 49</td><td align=\"center\">14</td><td align=\"center\">43</td><td align=\"center\">4.84</td><td align=\"center\">.69</td></tr><tr><td/><td align=\"center\">50 – 59</td><td align=\"center\">17</td><td align=\"center\">47</td><td align=\"center\">4.75</td><td align=\"center\">.93</td></tr><tr><td/><td align=\"center\">60 – 69</td><td align=\"center\">5</td><td align=\"center\">60</td><td align=\"center\">4.61</td><td align=\"center\">.44</td></tr><tr><td align=\"left\"><bold>Males</bold></td><td align=\"center\">20 – 29</td><td align=\"center\">18</td><td align=\"center\">89</td><td align=\"center\">5.55</td><td align=\"center\">.51</td></tr><tr><td/><td align=\"center\">30 – 39</td><td align=\"center\">13</td><td align=\"center\">46</td><td align=\"center\">5.76</td><td align=\"center\">.71</td></tr><tr><td/><td align=\"center\">40 – 49</td><td align=\"center\">7</td><td align=\"center\">86</td><td align=\"center\">5.35</td><td align=\"center\">1.08</td></tr><tr><td/><td align=\"center\">50 – 59</td><td align=\"center\">6</td><td align=\"center\">33</td><td align=\"center\">5.43</td><td align=\"center\">.71</td></tr><tr><td/><td align=\"center\">60 – 69</td><td align=\"center\">2</td><td align=\"center\">50</td><td align=\"center\">5.41</td><td align=\"center\">.56</td></tr></tbody></table></table-wrap>", "<table-wrap position=\"float\" id=\"T3\"><label>Table 3</label><caption><p>Medial tibial cartilage data per sex and decade in healthy individuals</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td/><td align=\"center\"><bold>Age Group (yrs)</bold></td><td align=\"center\"><bold>N</bold></td><td align=\"center\"><bold>Mean (SD)</bold><break/><bold>VCtAB (μL/mm²)</bold></td><td align=\"center\"><bold>Mean (SD)</bold><break/><bold>ThCtAB (mm)</bold></td></tr></thead><tbody><tr><td align=\"left\"><bold>Females</bold></td><td align=\"center\">20 – 29</td><td align=\"center\">13</td><td align=\"center\">1.52 (0.15)</td><td align=\"center\">1.45 (0.14)</td></tr><tr><td/><td align=\"center\">30 – 39</td><td align=\"center\">11</td><td align=\"center\">1.60 (0.25)</td><td align=\"center\">1.54 (0.24)</td></tr><tr><td/><td align=\"center\">40 – 49</td><td align=\"center\">11</td><td align=\"center\">1.49 (0.16)</td><td align=\"center\">1.44 (0.15)</td></tr><tr><td/><td align=\"center\">50 – 59</td><td align=\"center\">11</td><td align=\"center\">1.42 (0.20)</td><td align=\"center\">1.38 (0.19)</td></tr><tr><td/><td align=\"center\">60 – 69</td><td align=\"center\">4</td><td align=\"center\">1.40 (0.12)</td><td align=\"center\">1.35 (0.11)</td></tr><tr><td align=\"left\"><bold>Males</bold></td><td align=\"center\">20 – 29</td><td align=\"center\">13</td><td align=\"center\">1.78 (0.28)</td><td align=\"center\">1.71 (0.27)</td></tr><tr><td/><td align=\"center\">30 – 39</td><td align=\"center\">11</td><td align=\"center\">1.78 (0.26)</td><td align=\"center\">1.73 (0.28)</td></tr><tr><td/><td align=\"center\">40 – 49</td><td align=\"center\">3</td><td align=\"center\">1.67 (0.23)</td><td align=\"center\">1.61 (0.24)</td></tr><tr><td/><td align=\"center\">50 – 59</td><td align=\"center\">6</td><td align=\"center\">1.82 (0.18)</td><td align=\"center\">1.75 (0.13)</td></tr><tr><td/><td align=\"center\">60 – 69</td><td align=\"center\">2</td><td align=\"center\">1.85 (0.19)</td><td align=\"center\">1.80 (0.14)</td></tr></tbody></table></table-wrap>" ]

[]

[]

[]

[]

[]

[]

[]

[ "<graphic xlink:href=\"1471-2474-9-119-1\"/>", "<graphic xlink:href=\"1471-2474-9-119-2\"/>" ]

[]

{ "acronym": [], "definition": [] }

[]

[]

[]

[]

[]

[]

[ "<p>It is generally\nrecognized that the connecting peptide (C-peptide) of proinsulin fulfills an important function in the biosynthesis of\ninsulin. It brings together the A- and B-chains such that the initial folding\nand interchain disulfide bonds can be formed. Evolutionary considerations\nsuggest that a length of approximately 30 residues for the connecting segment,\nas is the case for human C-peptide, is optimal for the efficient further processing of the molecule (i.e., its\ncleavage into insulin and C-peptide). Following this, the two are stored in\nsecretory granules and eventually coreleased into the circulation. Because of\nits intimate connection to the insulin biosynthesis, C-peptide has been used as\na marker of insulin secretion. As such, it has contributed importantly to our\nunderstanding of the pathophysiology of several metabolic disorders, notably\ntype 1 and type 2 diabetes.</p>", "<p>The possibility\nthat C-peptide may possess biological effects of its own was considered but\nreceived relatively little attention at the time of its discovery in 1968. No\ndetectable influence on glucose metabolism or on lipolysis of isolated fat\ncells could be observed. In the absence of any insulin-like effect by C-peptide\nin isolated cell systems or when administered to healthy individuals, it was\nconcluded that C-peptide was without biological effect other than its role in\nthe biosynthesis of insulin; for a review see [##REF##403392##1##]. Consequently, C-peptide as a\nbioactive peptide left the scientific limelight and the interest was focused\ninstead on its usefulness as a marker of insulin secretion.</p>", "<p>It was not until\nthe early 1990s that direct C-peptide effects were re-evaluated. A series of\nstudies was undertaken involving administration of the peptide in type 1\ndiabetes patients, who lack C-peptide [##REF##1547915##2##]. This proved a useful\napproach and it became apparent that replacement of physiological\nconcentrations of C-peptide in this patient group results in significant\namelioration of diabetes-induced abnormalities of regional blood flow as well\nas improvements in peripheral nerve and kidney function. These surprising\nfindings, subsequently confirmed and extended by several laboratories, prompted\na renewed interest in C-peptide as a bioactive peptide in its own right. Since\nthen, a steadily increasing number of reports on new aspects of C-peptide\nphysiology have been presented. Today, a vast body of scientific evidence is\navailable comprising in vitro studies of the peptide's membrane interaction and\ncellular effects, in vivo studies in animal models of type 1 diabetes defining\nC-peptide's influence on functional and structural abnormalities of the kidneys\nand the peripheral nerves as well as clinical trials on nerve and kidney\nfunction in patients with type 1 diabetes, all of which attest to a wide\nspectrum of physiological effects being mediated by C-peptide. In addition, the\nfindings provide a basis for the notion that C-peptide administration, in\ncombination with regular insulin therapy, may be beneficial in the prevention\nand treatment of microvascular complications of type 1 diabetes.</p>", "<p>In the present,\nspecial issue of <italic>Experimental Diabetes\nResearch</italic>, most of the\nrecent developments in C-peptide research are being reviewed including an\nauthoritative review of the history and diagnostic aspects of C-peptide. A\nhighly qualified attempt is made to sort out the multitude of intracellular\neffects of C-peptide, seemingly contradictory when studied in different cell\nsystems and under varying experimental conditions. Perhaps the most compelling\nend effect of C-peptide is its stimulatory influence on the microcirculation in\na number of tissues, achieved via both activation and induction of endothelial\nnitric oxide synthase. These events are reviewed as are the beneficial effects\nof C-peptide and its C-terminal hexa- and pentapeptide segments on the\ndiabetes-induced reduction of red blood cell deformability. A possible\nstimulatory effect by C-peptide on glucose uptake is discussed on the basis of\nboth in vitro experiments and findings in type 1 diabetes patients. It is,\nhowever, noted that interpretation of such results is confounded by the recent\nobservation that C-peptide may elicit disaggregation of insulin hexamers,\nthereby augmenting the availability of bioactive insulin monomers [##REF##16845606##3##].</p>", "<p>C-peptide and\nits influence on renal physiology, particularly tubular function, are discussed. Likewise, \nC-peptide effects on the peripheral and central nervous system are reviewed. Much new and valuable\ninformation in this central area of C-peptide research has been presented from\nAnders Sima's laboratory. The comprehensive findings now point towards a need\nfor clinical trials and the current situation regarding clinical studies in\npatients with diabetic neuropathy is described. Finally, the possibility that\nC-peptide may serve as a mediator in the development of atherosclerotic lesions\nis discussed. Is the peptide guilty as charged or wrongly accused? Only future\nstudies can tell but, attesting to the rapid developments in the field of\nC-peptide physiology, a study just published reports that physiological as\nopposed to elevated concentrations of C-peptide serve to diminish\nhyperglycemia-induced vascular smooth muscle proliferation [##UREF##0##4##].</p>", "<p>The purpose of\nthis issue is to provide an update of our understanding of C-peptide physiology\nand the role of C-peptide deficiency in the development of microvascular\ncomplications of type 1 diabetes. Clearly, there is much more to be learned\nabout C-peptide. Identification of a receptor or the mechanism whereby\nC-peptide interacts with the cell membrane has a high priority. On the clinical\nside, further trials of long duration are needed to define the possible role\nfor C-peptide, together with insulin, in the treatment of type 1 diabetes. A\nmajor obstacle for extended clinical trials has been the lack of GMP-produced\nC-peptide suitable for human use. It is hoped that the evidence summarized in\nthis issue will convey the urgency with which clinical studies are needed and\nstimulate the interest of funding organizations and the pharmaceutical industry\nto become involved in this rapidly developing field.</p>", "<p content-type=\"signature-group\">\n<named-content content-type=\"signature\"><italic>Thomas Forst</italic>\n<italic>Thomas Forst</italic>\n</named-content>\n</p>", "<p content-type=\"signature-group\">\n<named-content content-type=\"signature\"><italic>John Wahren</italic>\n<italic>John Wahren</italic>\n</named-content>\n</p>" ]

[]

[]

[]

[]

[]

[]

[]

[]

[]

[]

[]

[]

{ "acronym": [], "definition": [] }

[]

[]

[]

[]

[]

[ "<p>Recommended by Dipak Panigrahy</p>", "<p>Today, there is increasing evidence that PPAR<italic>γ</italic> agonists, including thiazolidinediones (TDZs) and nonthiazolidinediones, block the motility and invasiveness of glioma cells and other highly migratory tumor entities. However, the mechanism(s) by which PPAR<italic>γ</italic> activators mediate their antimigratory and anti-invasive properties remains elusive. This letter gives a short review on the debate and adds to the current knowledge by applying a PPAR<italic>γ</italic> inactive derivative of the TDZ troglitazone (Rezulin) which potently counteracts experimental glioma progression in a PPAR<italic>γ</italic> independent manner.</p>" ]

[ "<p>Gliomas are the most common primary tumors in the central nervous system, with\nglioblastomas as the most malignant entity [##REF##11150363##1##]. Despite multimodal therapy regimens incl uding\nneurosurgical resection, radio- and polychemotherapy, the prognosis of glioma\npatients remains poor. Less than 3% of affected patients survive more than five\nyears after diagnosis [##REF##15685439##2##]. Rapid proliferation, tumor-induced neurodegeneration,\nand brain edema [##UREF##0##3##] as well as diffuse brain invasion are\npathological hallmarks of these tumors and are likely to determine unfavorable\nprognosis. Because local invasion of neoplastic cells into the surrounding\nbrain is perhaps the most important aspect in the biology of gliomas that preclude successful\ntreatment, pharmacological inhibition of glioma cell migration and brain\ninvasion is considered as a highly promising strategy for adjuvant glioma\ntherapy.</p>", "<p>Today, there is increasing evidence that PPAR<italic>γ</italic>\nagonists, including thiazolidinediones (TDZs) and nonthiazolidinediones, block the\nmotility and invasiveness of glioma cells and other highly migratory tumor\nentities. GW7845, an investigational non-TDZ PPAR<italic>γ</italic> ligand, binds and activates human PPAR<italic>γ</italic>\nat low nanomolar concentrations and thus possesses a higher potency than TDZs such\nas pioglitazone (Actos), troglitazone (Rezulin), rosiglitazone (Avandia), and the\nexperimental PPAR<italic>γ</italic> agonist ciglitazone, respectively, which require submicromolar doses [##REF##9836622##4##]. Grommes et al. [##REF##15665144##5##] demonstrated that 30 <italic>μ</italic>M concentrations\nof GW7845 reduced the viability of rat (C6) and human glioma cells (U-87 MG,\nA172), which could be attributed to a G₁ cell cycle arrest and\nincreased cell death. Besides its antiproliferative and cytotoxic properties, the authors demonstrated for\nthe first time that GW7845 counteracts migration and invasion of C6 rat glioma\ncells in vitro (spheroid outgrowth, Boyden chamber assay). A subsequent study revealed that the FDA-approved\nTDZ pioglitazone exhibits antiglioma properties similar to GW7845 [##REF##16887936##6##]. Alike GW7845, micromolar doses of\npioglitazone (30 <italic>μ</italic>M) counteract C6 rat glioma cell invasiveness in vitro (Boyden chamber assay). In\nthis study, Grommes et al. [##REF##16887936##6##] demonstrated profound in vivo antiglioma properties of pioglitazone. Following C6\nglioma cell implantation into the striata of adult rats, oral or intracerebral\ndrug application effectively decelerated glioma progression, resulting in an\nimproved clinical outcome and 80% reduction of tumor volume at 3 weeks after\ntumor implantation. Immunohistochemical analyses of pioglitazone-treated\nanimals revealed that protein levels of MMP-9 (<italic>matrix metalloproteinase 9</italic>), which has shown to be intimately\ninvolved in glioma migration and invasion [##REF##12439751##7##], were substantially reduced in the bulk tumor\nand the tumor margins. However, the data regarding the antiglioma properties of\npioglitazone are somewhat contradictory in a mouse glioma model (GL261 glioma\ncells, C57Bl/6 mice). Grommes et al. demonstrated that oral application of\npioglitazone increased the number of surviving animals after 30 days of\ntreatment. By employing the same model, Spagnolo and coworkers observed no\neffect on survival following oral drug application, while intracerebral\ninjection of pioglitazone increased the mean survival time [##REF##17410715##8##]. We have recently shown that the TDZ troglitazone\nreduces the viability and proliferation of rat (F98), mouse (SMA-560), and\nhuman (U-87 MG) glioma cells slightly but significantly more potent than the remaining TZDs tested \n(troglitazone &gt; pioglitazone &gt; rosiglitazone &gt; ciglitazone) [##REF##17541035##9##]. By employing an ex vivo glioma invasion \nmodel [##REF##15871520##10##], troglitazone effectively blocked glioma progression and brain invasion and consistent with the\nin vitro data presented by Grommes and coworkers, we confirmed that troglitazone (30 <italic>μ</italic>M) antagonized rat\nF98 glioma cell migration (scratch wound healing, Boyden chamber assay).</p>", "<p>Inhibition of cell motility and invasiveness by PPAR<italic>γ</italic>\nactivators has also been described for other neoplastic cells and thus appears not\nto be restricted to glioma. Liu et\nal. [##REF##12779083##14##] showed that GW7845 (5 <italic>μ</italic>M) as well as the FDA-approved\nTDZs pioglitazone and rosiglitazone (both 25 <italic>μ</italic>M) inhibits the invasive properties of human MDA-MB-231\nbreast cancer cells. In this study, treatment with PPAR<italic>γ</italic>\nagonists was associated with increased <italic>tissue inhibitor of matrix metalloproteinase 1</italic> (TIMP-1) mRNA and\nprotein levels, which are likely to contribute to the anti-invasive effects observed. Recently, Yang et al. [##REF##18021457##15##] \ndemonstrated that troglitazone (10–30 <italic>μ</italic>M)\ninhibits migration and invasiveness of a human ovarian carcinoma ES-2 cells.\nAnti-invasive properties were also shown for the TDZ ciglitazone, although with\na lower potency. Extended analyses by Yang et al. revealed that troglitazone\n(20 <italic>μ</italic>M) inhibits focal adhesion formation associated with reduced focal adhesion\nkinase (FAK) activity. FAK, an ubiquitously expressed\nnonreceptor tyrosine kinase, has shown to be a vitally important regulator of\ncancer cell migration and invasion. FAK is highly expressed in many tumor\nentities and activated by autophosphorylation [##REF##16997283##16##], which has shown to be reduced by more than 80% in troglitazone-treated\nES-2 cells [##REF##18021457##15##]. Based on these data, the authors concluded that troglitazone may\ninhibit ES-2 cell migration and invasion by preventing FAK activation. Concordantly,\ninhibition of FAK kinase activity by the investigational small molecule TAE226 reduced the invasive properties of human U-87 MG, U251, and LN18 glioma cells by more than 50% [##REF##17431114##17##], suggesting that FAK activation decisively\npromotes migration and invasion also of glioma cells. In all, these data demonstrate that PPAR<italic>γ</italic>\nactivators belonging to different chemical classes effectively diminish glioma\nprogression in vitro, ex vivo, and in vivo [##REF##15665144##5##, ##REF##16887936##6##, ##REF##17541035##9##], which occurs at least in part by the inhibition of glioma \ncell migration and invasiveness.</p>", "<p>Given the fact that cancer cell migration and invasion are highly complex processes [##REF##15674479##18##], the mechanism(s) by which PPAR<italic>γ</italic> agonists exert their\nantimigratory and anti-invasive properties requires further investigation. Besides MMP-9,\nTIMP-1, and FAK, which have been shown to be involved in the antimigratory\nactivities of PPAR<italic>γ</italic> agonists, we recently demonstrated that already low doses of troglitazone\nblock <italic>transforming growth factor beta</italic> (TGF-<italic>β</italic>)\nrelease [##REF##17541035##9##], a cytokine which plays a pivotal role in glioma\ncell motility [##REF##11716069##19##]. Several in vitro studies revealed that exogenously added TGF-<italic>β</italic>\n₁ and TGF-<italic>β</italic>\n₂ elicit a strong stimulation of migration in a variety of\nglioma cells [##REF##8054266##20##–##REF##15520202##23##], while TGF-<italic>β</italic> gene silencing has shown to reduce glioma cell motility and invasiveness [##REF##15492287##24##]. In agreement with these findings, inhibition\nof TGF-<italic>β</italic> signaling by the investigational type I TGF-<italic>β</italic> receptor antagonist,\nSB-431542 reduced the invasive properties of human D-54 MG and rat F98 glioma\ncells by approximately 70% [##REF##17541035##9##, ##REF##15210860##25##]. The role of TGF-<italic>β</italic> as molecular target for\nglioma therapy has been facilitated by studies using surgically resected glioma\ntissues, which revealed an intriguing correlation between tumor grade and the\nexpression of TGF-<italic>β</italic> ligands and their corresponding receptors I and\nII. High-grade gliomas express high levels of TGF-<italic>β</italic>RI, TGF-<italic>β</italic>RII, and TGF-<italic>β</italic>\nligands, while the expression levels of these molecules have been shown to be\nweak in low-grade gliomas and normal brain tissue [##REF##10861501##26##–##REF##7635563##28##]. A comprehensive transcriptome-wide study by\nDemuth et al. [##UREF##1##29##] using 111 glial tumor samples and 24 normal\nbrain specimens identified the TGF-<italic>β</italic> signaling pathway to be predominantly enriched\nin glial tumors compared to normal brain. In all, these data implicate that\nglioma cells release TGF-<italic>β</italic> ligands at high doses and fortify their promigratory\nand proinvasive properties in an autocrine manner, thus promoting glioma\nprogression. Given the fact that 10 <italic>μ</italic>M doses of troglitazone allay TGF-<italic>β</italic>\nrelease of glioma cells (F98, SMA-560, U-87 MG) by more than 50% [##REF##17541035##9##], we hypothesized that the abrogation of glioma\ncell motility and invasiveness by troglitazone and other PPAR<italic>γ</italic>\nactivators is primarily\ndriven by the inhibition of TGF-<italic>β</italic> signaling and thus, troglitazone and related\ncompounds may be considered for adjuvant glioma therapy to counteract TGF-<italic>β</italic>-mediated\nbrain invasion.</p>", "<p>However, the mechanism(s) by which PPAR<italic>γ</italic> activators mediate their antimigratory and anti-invasive properties remains \nelusive. We have shown that PPAR<italic>γ</italic> inhibition by the investigational antagonist GW9662, either alone or in combination\nwith troglitazone, does not affect rat F98 glioma cell invasiveness in a Boyden\nchamber assay, suggesting that the effects observed are not mediated by PPAR<italic>γ</italic> [##REF##17541035##9##]. Simultaneously, Yang and coworkers [##REF##18021457##15##] have shown that PPAR<italic>γ</italic> knockdown by siRNA did not counteract the\nanti-invasive features of troglitazone using human ovarian carcinoma ES-2\ncells, underscoring the idea that the PPAR<italic>γ</italic>\nagonists counteract cancer cell migration by a yet unknown off-target activity.\nTo validate these preliminary findings, we analyzed the effects of a troglitazone\nderivative, Δ2-troglitazone, which has been shown to be PPAR<italic>γ</italic>-inactive [##REF##15109648##11##, ##REF##15735046##31##, ##REF##16728570##32##]. In case the antiglioma properties of\ntroglitazone are solely or predominantly due to PPAR<italic>γ</italic> activation, Δ2-troglitazone\nshould display no or a considerably lower inhibitory potency on glioma cell\nviability than troglitazone. Initially, concentration-dependent\ninhibition of glioma cell viability by troglitazone and Δ2-troglitazone was investigated using glioma cell lines derived from mouse\n(SMA-560), rat (F98), and human (U-87 MG, U-373 MG). As shown by MTT assay, both\ncompounds inhibited glioma cell growth in a concentration-dependent manner with\nsimilar potencies (Figures ##FIG##0##1(a)##, ##FIG##0##1(c)##). Even though numerous PPAR<italic>γ</italic>-dependent mechanisms\nhave been identified (for review see Tatenhorst et al., this issue), these data\nsuggest that PPAR<italic>γ</italic> activation is not an imperative prerequisite for the inhibition of glioma cell viability in vitro, which is in line with\nprevious studies using human PC-3 and LNCaP prostate cancer and human A549 lung\ncarcinoma cells [##REF##15109648##11##, ##REF##15735046##31##]. Next, we analyzed the reduction of glioma cell viability using IC₉₀ doses of Δ2-troglitazone and\nequimolar doses of troglitazone. In all four cell lines tested, Δ2-troglitazone displays\na slightly but significantly higher potency compared with troglitazone.\nHowever, with IC₉₀ doses ranging from 93 <italic>μ</italic>M (SMA-560) to 132 <italic>μ</italic>M (U-87\nMG), the antiproliferative properties of Δ2-troglitazone can be\nregarded as moderate.</p>", "<p>Next, we analyzed the effects of troglitazone and Δ2-troglitazone\non TGF-<italic>β</italic> release by glioma cells. Hjelmeland et al. [##REF##15210860##25##] have shown that secretion of activated TGF-<italic>β</italic>\n₁ is a common attribute of glioma cells (U-87 MG, U-373 MG, D-54 MG, D-270 MG, D-423\nMG, D-538 MG), while simultaneous release of TGF-<italic>β</italic>\n₂ was found only\nsporadically (D-54 MG, U-373 MG, D-423 MG). In accordance with these findings, quantification of TGF-<italic>β</italic>\n₁ and TGF-<italic>β</italic>\n₂ transcript\nlevels by real-time PCR revealed that U-373 MG and SMA-560 glioma cells express\nboth TGF-<italic>β</italic>\n₁ and TGF-<italic>β</italic>\n₂, respectively, \nwhile TGF-<italic>β</italic>\n₁ is clearly\nthe predominant isoform in F98 and U-87 MG glioma cells (data not shown). Repeated\nquantification (<italic>n</italic> ≥ 7) of absolute TGF-<italic>β</italic>\n₁ levels following cultivation of glioma cells for 48 hours in serum-free medium\nrevealed that all cell lines investigated secrete TGF-<italic>β</italic>\n₁ (F98: 8.45 ± 1.59 ng/mL; SMA-560: 2.7 ± 0.54 ng/mL; U-87 MG: 2.55 ± 0.68 ng/mL, U-373 MG: 0.43 ± 0.08 ng/mL), while both troglitazone and Δ2-troglitazone inhibit\nTGF-<italic>β</italic>\n₁ release in a dose-dependent manner (Figures ##FIG##0##1(b)##, ##FIG##0##1(d)##). The\nfinding that Δ2-troglitazone\ncounteracts TGF-<italic>β</italic>\n₁ release indicates that this effect is not PPAR<italic>γ</italic> dependent. Again, Δ2-troglitazone displays\na significantly higher potency as compared with troglitazone (##FIG##0##Figure 1(f)##). In\ncase of Δ2-troglitazone, 90% inhibition\nof TGF-<italic>β</italic>\n₁ release was found at concentrations ranging from 5 <italic>μ</italic>M\n(F98) to 14 <italic>μ</italic>M (U-87 MG, U-373 MG), whereas troglitazone required 11 <italic>μ</italic>M (F98)\nto 30 <italic>μ</italic>M (U-373 MG) to achieve the same effects. Strikingly, troglitazone as\nwell as Δ2-troglitazone is approximately 10 fold more potent inhibitors of TGF-<italic>β</italic>\n₁ release than\nof glioma cell proliferation, suggesting that both effects may not be essentially\ninterlinked.</p>", "<p>In agreement with the finding that TGF-<italic>β</italic>\n₁ promotes glioma cell migration and brain\ninvasion, treatment of glioma cells with micromolar doses of Δ2-troglitazone\neffectively blocks their migrative properties (##FIG##1##Figure 2##). Already 10 <italic>μ</italic>M doses\nof Δ2-troglitazone inhibit F98 glioma cell migration in a Boyden chamber assay, while migration was completely\nsuppressed at 20 <italic>μ</italic>M. An intriguing question is whether inhibition of glioma\ncell migration alone is sufficient to counteract glioma progression. To address\nthis issue we employed rat organotypic hippocampal brain slice cultures (OHSCs)\nto monitor glioma progression and brain invasion in the organotypic brain\nenvironment [##REF##15857402##12##]. Here, eGFP-labelled F98 glioma cells were implanted into the\nentorhinal cortex of OHSCs (##FIG##2##Figure 3(a)##). The tumor infiltration area was\nquantified up to 12 days by fluorescence microscopy. A continuous increase of\nthe bulk tumor mass was observed in solvent-matched control experiments\nat all time periods. 12 days after glioma cell implantation, the tumor\ninfiltration area increased approximately 4.5 fold compared to the initial\ntumor size at day 1 after implantation (Figures ##FIG##2##3(b)##, ##FIG##2##3(c)##). In contrast, the\ntumor infiltration size remained stable over the period of 12 days after\ntreatment with 10 <italic>μ</italic>M Δ2-troglitazone. This\nfinding indicates that Δ2-troglitazone is not\nable to reduce existing tumor masses, but effectively inhibits tumor\nprogression and brain invasion in an organotypic environment. Given the fact\nthat 10 <italic>μ</italic>M doses of Δ2-troglitazone\nsignificantly affect TGF-<italic>β</italic>\n₁ release (##FIG##0##Figure 1(d)##) and glioma cell\nmotility (##FIG##1##Figure 2##) but not glioma cell viability (##FIG##0##Figure 1(c)##), these data suggest that glioma cell\nmigration is an essential requirement for glioma progression in a system closely resembling extracellular matrix environment present\nin the brain.</p>", "<p>TGF-<italic>β</italic> antagonism is considered as a therapeutic strategy\nincluding the development of antisense regimens, inhibition of pro-TGF-<italic>β</italic> processing, scavenging of TGF-<italic>β</italic> by the TGF-<italic>β</italic>-binding proteoglycan decorin, and blocking of\nTGF-<italic>β</italic> receptor I kinase activity [##REF##16454748##33##]. The finding that troglitazone and its derivative Δ2-troglitazone effectively\ninhibit TGF-<italic>β</italic> release suggests readily available PPAR<italic>γ</italic> activators and structurally related PPAR<italic>γ</italic> inactive compounds as\ncandidate drugs for adjuvant glioma therapy. Besides its promigratory and proinvasive activities, TGF-<italic>β</italic> is considered\nas one of the most potent immunosuppressive factors released by gliomas allowing\nglioma cells to escape from immune surveillance [##REF##10795886##34##, ##REF##17414317##35##]. Friese et al. [##REF##15492287##24##] demonstrated that combined TGF-<italic>β</italic>\n₁ \nand TGF-<italic>β</italic>\n₂ knock down in human LNT-229 glioma cells results in a\nloss of tumorigenicity when xenografted into CD1 nude mice, and natural killer\ncells isolated from these animals show an activated phenotype. More than 10\nyears ago, Ständer et al. [##REF##9930319##36##] have shown that inhibition of TGF-<italic>β</italic> signaling\nby decorin increases the number of B and T cells (CD45+), T helper cells\n(CD4+), cytotoxic T cells (CD8+), and, most prominently, of activated T cells\n(CD25+) infiltrating the tumor in an intracerebral C6 rat glioma model. By employing\nan SMA-560 mouse glioma model, Tran et al. [##REF##17522330##37##] have shown that inhibition of TGF-<italic>β</italic> signaling by the TGF-<italic>β</italic> RI\nkinase inhibitor SX-007 increased T-cell (CD3+) infiltration into the tumor. Due\nto the fact that inhibition of TGF-<italic>β</italic> signaling has been shown to enhance\nantiglioma immune responses in vivo [##REF##15492287##24##, ##REF##9930319##36##, ##REF##17522330##37##] it appears likely that troglitazone, inhibiting\nTGF-<italic>β</italic>\n₁ release at clinically achievable doses [##REF##17541035##9##, ##REF##10496299##38##], restores immune\nsurveillance. However, the yet-unknown protein/proteins mediating the inhibition of glioma progression by troglitazone\nand Δ2-troglitazone remain(s) to be identified and may represent\nfuture targets for structure-relationship studies. Moreover, PPAR<italic>γ</italic> inactive derivatives of\nknown PPAR<italic>γ</italic> agonists which retain their propensity to counteract glioma progression might be\nfurther developed to minimize potential PPAR<italic>γ</italic>\nmediated side effects in glioma patients.</p>" ]

[ "<title>ACKNOWLEDGMENTS</title>", "<p>SMA-560 mouse glioma cells [##REF##9402588##39##] were kindly provided by D. D. Bigner (Durham, USA). Thanks are due to Anita Betz, \nUniversity of Würzburg for her skilful assistance in synthesizing Δ2-troglitazone. This study was\nsupported by the Köln Fortune Program (Faculty of Medicine, University of Cologne)\nand the Wilhelm Sander-Stiftung (2008.010.1) to E. H.</p>" ]

[ "<fig id=\"fig1\" position=\"float\"><label>Figure 1</label><caption><p>\n<italic>Troglitazone (TRO) and the PPAR<italic>γ</italic> inactive </italic>Δ<italic>2-troglitazone (</italic>Δ<italic>2-TRO) reduce \nglioma cell viability and TGF-<italic>β</italic>₁ release</italic>. Δ2-TRO was synthesized as previously described in [##REF##15109648##11##]. (a), (c) Concentration-dependent\ninhibition of glioma cell viability by TRO (a) or Δ2-TRO (c) in the\nindicated cell lines are given as mean ± SEM percentage relative to time- and\nsolvent-matched controls. Cell viability assays (MTT\nassay, 96 hours) were performed as described earlier [##REF##15857402##12##, ##REF##6606682##13##]. Inhibitory\nconcentrations IC₅₀ and IC₉₀, defined as concentrations\nshown to inhibit tumor cell viability by 50% or 90%,\nrespectively, were determined by nonlinear regression data analysis: TRO: F98 (62 <italic>μ</italic>M, 166 <italic>μ</italic>M), SMA-560 (26 <italic>μ</italic>M, 407 <italic>μ</italic>M), U-87 MG (120 <italic>μ</italic>M, 324 <italic>μ</italic>M), and U-373 MG (123 <italic>μ</italic>M,\n331 <italic>μ</italic>M); Δ2-TRO: F98 (46 <italic>μ</italic>M, 95 <italic>μ</italic>M), SMA-560 (23 <italic>μ</italic>M, 93 <italic>μ</italic>M), U-87 MG (78 <italic>μ</italic>M, 132 <italic>μ</italic>M), and U-373 MG (71 <italic>μ</italic>M, 126 <italic>μ</italic>M). <italic>Troglitazone and the PPAR<italic>γ</italic> inactive </italic>Δ<italic>2-troglitazone\nreduce TGF-<italic>β</italic>₁ release at low micromolar doses</italic>: (b), (d) quantification\nof TGF-<italic>β</italic>\n₁ release by F98, SMA-560, U87-MG, and U-373 MG glioma cell\nculture supernatants following TRO (b)\nor Δ2-TRO (d) treatment for 48 hours. TGF-<italic>β</italic>\n₁ protein levels in glioma cell culture supernatants were determined as described\nin [##REF##17541035##9##] using the mouse/rat/porcine/canine or the human quantikine\nTGF-<italic>β</italic>\n₁ ELISA Kit (R&amp;D Systems, Minneapolis, Minn, USA), respectively. Each\nexperiment was repeated at least 3 times (<italic>n</italic> ≥ 3). Drug concentrations shown\nto inhibit TGF-<italic>β</italic>\n₁ release by 50% or 90%, respectively, were\ndetermined by nonlinear regression data analysis: TRO: F98 (7 <italic>μ</italic>M, 11 <italic>μ</italic>M),\nSMA-560 (8 <italic>μ</italic>M, 15 <italic>μ</italic>M), U-87 MG (8 <italic>μ</italic>M, 28 <italic>μ</italic>M), and U-373 MG (10 <italic>μ</italic>M, 30 <italic>μ</italic>M); Δ2-TRO:F98 (3 <italic>μ</italic>M, 5 <italic>μ</italic>M), SMA-560\n(3 <italic>μ</italic>M, 8 <italic>μ</italic>M), U-87 MG (4 <italic>μ</italic>M, 14 <italic>μ</italic>M), and U-373 MG (4 <italic>μ</italic>M, 14 <italic>μ</italic>M). Δ<italic>2-Troglitazone\ndisplays higher potencies than troglitazone</italic>. Using IC₉₀ concentrations\nof Δ2-TRO and equimolar concentrations of TRO,\nthe PPAR<italic>γ</italic> inactive Δ2-TRO displays a significantly stronger\neffect in both experimental paradigms (*** = <italic>P</italic> &lt; .001, <italic>t</italic>-test) (e), (f).</p></caption></fig>", "<fig id=\"fig2\" position=\"float\"><label>Figure 2</label><caption><p>\n<italic>The PPAR<italic>γ</italic> inactive</italic> Δ <italic>2-troglitazone (</italic>Δ<italic>2-TRO) inhibits glioma cell \nmigration</italic>. The glioma cell\nmigration assay (Boyden chamber; QCM-FN Migration Assay, Chemicon, Temecula, Calif, USA) was performed as described recently [##REF##17541035##9##]. Briefly, F98 rat\nglioma cells, pretreated with the test compound or solvent for 24 hours, were\ntransferred into each Boyden chamber. After 24 hours of incubation, cells which\nmigrated through the fibronectin-coated\nchamber membranes (8 micron pore diameter) were quantified according to\nthe manufacturer’s protocol. Experiments were repeated 3 times (<italic>n</italic> = 3). (** = <italic>P</italic> &lt; .01; *** = <italic>P</italic> &lt; .001; <italic>t</italic>-test). Right panel: representative\nmicrophotographs of F98 glioma cells which migrated though the\nfibronectin-coated chamber membranes after treatment with Δ2-TRO (20 <italic>μ</italic>M) or solvent only.</p></caption></fig>", "<fig id=\"fig3\" position=\"float\"><label>Figure 3</label><caption><p>\nΔ<italic>2-Troglitazone inhibits\nglioma progression in an organotypic glioma transplantation model</italic>. (a) Organotypic hippocampal glioma\ninvasion assay was performed as described earlier [##REF##15871520##10##, ##REF##15857402##12##, ##REF##16731757##30##]. In brief, enhanced green\nfluorescent protein (eGFP) positive F98 rat glioma cells were transplanted into\nthe entorhinal cortex of organotypic rat brain slice cultures one day after preparation. DAI\n= days after implantation. DG = dentate gyrus. EC = entorhinal cortex. (b) Tumor progression was monitored by fluorescent\nmicroscopy over the time course of 12 days. Quantification of the tumor\ninfiltration area at day 1 to day 12 after transplantation derived from 3\nindependent experiments is shown. For each experiment, the tumor infiltration\narea at DAI 1 was defined as 100%. Data are given as mean ± SD percentage. At\nDAI 12, the tumor infiltration area significantly increased to 448 ± 71 % (<italic>P</italic> = .002, <italic>t</italic>-test) in solvent-matched controls but\nremained unchanged following Δ2-TRO treatment (75 ± 22 %; <italic>P</italic> = .18, <italic>t</italic>-test). Starting from DAI 2,\ndifferences in tumor progression (TRO versus Δ2-TRO) reached\nstatistical significance (<italic>P</italic> &lt; .01, <italic>t</italic>-test) <bold>(c)</bold> A\ncontinuous increase of the bulk tumor masses was observed in solvent-matched\ncontrols while 10 <italic>μ</italic>M concentrations of Δ2-TRO effectively\nblocked tumor progression. Right column: magnification of the indicated border\narea between bulk tumor mass and rat brain tissue. In controls, F98 glioma\ncells have diffusely migrated into the adjacent brain parenchyma, while a sharp\ntumor border was observed following Δ2-TRO treatment (scale bar: 200 <italic>μ</italic>m).</p></caption></fig>" ]

[]

[]

[]

[]

[]

[]

[]

[]

[ "<graphic xlink:href=\"PPAR2008-513943.001\"/>", "<graphic xlink:href=\"PPAR2008-513943.002\"/>", "<graphic xlink:href=\"PPAR2008-513943.003\"/>" ]

[]

{ "acronym": [], "definition": [] }

[ "<title>1. INTRODUCTION</title>", "<p>Peroxisome proliferator-activated receptor-<italic>β</italic>/<italic>δ</italic>\n(PPAR<italic>β</italic>/<italic>δ</italic>)\nis a transcription factor that is activated by endogenous fatty acid ligands\nand by synthetic agonists [##REF##15888456##1##, ##REF##17132851##2##]. Major functions of\nPPAR<italic>β</italic>/<italic>δ</italic> are associated with the regulation of glucose, energy, and lipid\nmetabolism [##REF##16601267##3##], and the control of\ninflammatory responses [##REF##16378501##4##, ##REF##18465655##5##]. PPAR<italic>β</italic>/<italic>δ</italic>, therefore, represents a promising drug target for the treatment of\ncommon diseases such as obesity, metabolic syndrome, chronic inflammation, and\narteriosclerosis, which has led to the development of synthetic drug agonists\nwith subtype selectivity and high-affinity binding [##REF##16322072##6##]. Mice lacking PPAR<italic>β</italic>/<italic>δ</italic> show an aberrant development of the placenta and exhibit a defect in\nwound healing associated with alterations in cell proliferation,\ndifferentiation, and cellular survival [##REF##10866668##7##–##REF##16581799##10##]. Experimental\nevidence obtained with cultured cells has provided additional strong evidence\nfor a role of PPAR<italic>β</italic>/<italic>δ</italic> in cell cycle regulation and differentiation in different cell types (see\n##TAB##0##Table 1##). Consistent with these physiological functions, there is also clear\nevidence for a role of PPAR<italic>β</italic>/<italic>δ</italic> in oncogenesis and tumor growth.\nThese findings might provide a basis for the development of novel strategies\nfor the treatment of proliferative diseases, but also demand some caution with\nrespect to the clinical use of PPAR<italic>β</italic>/<italic>δ</italic>-directed dugs. A detailed knowledge of the role of PPAR<italic>β</italic>/<italic>δ</italic> in cell proliferation and its effects on tumor growth are therefore of\nparamount importance.</p>" ]

[]

[]

[]

[ "<title>6. CONCLUSIONS</title>", "<p>Studies addressing the role of PPAR<italic>β</italic>/<italic>δ</italic> in differentiation have yielded a consistent picture and point to a\ndifferentiation promoting in a wide spectrum of different cell types. Numerous reports have also clearly documented a\nrole for PPAR<italic>β</italic>/<italic>δ</italic>\nin cell proliferation and tumorigenesis, yet\ndifferent studies have produced controversial results, even though the majority\nof studies describe antiproliferative effects by PPAR<italic>β</italic>/<italic>δ</italic> (see ##TAB##0##Table 1##).</p>", "<p>One reason for the apparently discrepant data may be associated\nwith the use of different experimental strategies. Since the precise mechanisms\nof PPAR<italic>β</italic>/<italic>δ</italic>-mediated gene regulation are often not known, the results from\ngain-of-function and loss-of-function are not always easy to interpret. Thus, ligand activation and genetic\ninactivation of PPAR<italic>β</italic>/<italic>δ</italic> may\nhave opposite effects, as in the case of classical PPRE-driven genes, but may\nalso give similar results in other regulatory settings. The latter has been\ndescribed, for instance, for PPAR<italic>β</italic>/<italic>δ</italic>-mediated gene repression through direct interaction with the transcriptional\nrepressor BCL-6 in macrophages [##REF##12970571##54##]. This aspect has not been thoroughly analyzed to date so that it is\ndifficult to judge its contribution to the deviant results published in\ndifferent studies.</p>", "<p>To help explain the discrepant published\ndata, we would therefore like to put forward another hypothesis. This model\npostulates that PPAR<italic>β</italic>/<italic>δ</italic> is not a <italic>bona fide</italic> cell cycle\nregulator with a defined function but rather affects the expression of both\ninducers and inhibitors of cell proliferation (e.g., regulators of the AKT\npathway and PDGF versus the cell cycle inhibitors p57^KIP2 and <italic>G0S2</italic>; see ##TAB##0##Table 1##). This is conceivable both in view of the\nlarge number of potential PPAR target genes, estimated at several thousand for\nthe human genome [##UREF##7##55##]. Depending on the particular cell\ntype, the metabolic or proliferative state of the cell or other experimental\nconditions, positive or negative regulators of the cell cycle may prevail\nresulting in opposite effects. This suggests that the precise effects of PPAR<italic>β</italic>/<italic>δ</italic> on cell proliferation are highly context-dependent and not predictable\non the basis of our current knowledge. Clearly, a better and detailed understanding of the effects of PPAR<italic>β</italic>/<italic>δ</italic>\non cell cycle regulation and differentiation will be a prerequisite for the\ndevelopment of PPAR<italic>β</italic>/<italic>δ</italic>\ndirected drugs and their clinical application.</p>" ]

[ "<p>Recommended by Francine M. Gregoire</p>", "<p>Peroxisome proliferator-activated receptor-<italic>β</italic>/<italic>δ</italic> (PPAR<italic>β</italic>/<italic>δ</italic>) is a ligand-activated transcription factor with essential functions in the regulation of lipid catabolism, glucose homeostasis, and inflammation, which makes it a potentially relevant drug target for the treatment of major human diseases. In addition, there is strong evidence that PPAR<italic>β</italic>/<italic>δ</italic> modulates oncogenic signaling pathways and tumor growth. Consistent with these observations, numerous reports have clearly documented a role for PPAR<italic>β</italic>/<italic>δ</italic> in cell cycle control, differentiation, and apoptosis. However, the precise role of PPAR<italic>β</italic>/<italic>δ</italic> in tumorigenesis and cell proliferation remains controversial. This review summarizes our current knowledge and proposes a model corroborating the discrepant data in this area of research.</p>" ]

[ "<title>2. PPAR<italic>β</italic>/<italic>δ</italic> AFFECTS TUMORIGENESIS</title>", "<p>The role of PPAR<italic>β</italic>/<italic>δ</italic> in tumorigenesis has been explored predominantly in epithelial tumors\nof the skin, lung, and intestine and in the tumor stroma. PPAR<italic>β</italic>/<italic>δ</italic> inhibits chemically induced skin carcinogenesis, since an enhancement\nof chemically induced skin tumor growth is seen in mice with a global\ndisruption of <italic>Pparb</italic> [##REF##15033975##38##]. However, no effect on skin\ncarcinogenesis is observed in mice lacking PPAR<italic>β</italic>/<italic>δ</italic> specifically in basal keratinocytes [##REF##17301838##39##], suggesting that the\ntumor suppressive effect of PPAR<italic>β</italic>/<italic>δ</italic> is due to a function in other cell types. A tumor suppressive role for\nPPAR<italic>β</italic>/<italic>δ</italic> has also been described for a transgenic mouse model of Raf\noncogene-induced lung adenoma formation, but similar to skin carcinogenesis the\nprecise mechanisms and cell types involved are not known [##REF##17671688##40##]. Effects of PPAR<italic>β</italic>/<italic>δ</italic> have also been reported in different mouse models of intestinal\ncarcinogenesis, that is, the Apc/Min mouse lacking functional APC protein and\nchemically induced intestinal carcinogenesis, but these studies differ in their\nconclusions [##UREF##3##41##]. Thus, PPAR<italic>β</italic>/<italic>δ</italic> has been reported to have either no effect on intestinal tumorigenesis [##REF##11756685##9##] to attenuate tumor\ngrowth by promoting terminal differentiation of colonocytes [##REF##17693664##33##, ##UREF##4##42##–##REF##17893232##45##] or to potentiate\ntumorigenesis [##UREF##5##46##–##REF##17148604##48##]. The reason for\nthese discrepancies remains unclear at present [##UREF##6##49##], but may be in part\nrelated to a function of PPAR<italic>β</italic>/<italic>δ</italic>\nin host cells recruited by the tumor, such as endothelial cells, fibroblasts,\nand macrophages [##REF##11900251##50##]. Indeed, recent work\nshowed that PPAR<italic>β</italic>/<italic>δ</italic> is indispensable for the formation of functional tumor microvessels [##REF##17641685##29##, ##REF##17652168##51##], suggesting that PPAR<italic>β</italic>/<italic>δ</italic>\nmay have different functions in the tumor stroma and in tumor cells with\nopposing effects on tumor growth. The role of PPAR<italic>β</italic>/<italic>δ</italic>\nin tumor stroma cells is further discussed below.</p>", "<title>3. ATTENUATION OF TUMOR STROMA CELL\nPROLIFERATION BY PPAR<italic>β</italic>/<italic>δ</italic>\n</title>", "<p>The inhibition of syngeneic tumor growth in\nmice lacking PPAR<italic>β</italic>/<italic>δ</italic>\nstrongly correlates with a lower density of functional tumor microvessels [##REF##17641685##29##, ##REF##17652168##51##], which is associated with a striking increase\nin the proliferation of tumor endothelial cells and an inhibition of their\nmaturation [##REF##17641685##29##]. The immature microvascular structures are also frequently surrounded\nby perivascular cells expressing vast amounts of <italic>α</italic>-smooth\nmuscle actin, giving rise to an overall picture characteristic of tumor\nendothelial hyperplasia. In vivo\nmicroarray analysis led to the identification of PPAR<italic>β</italic>/<italic>δ</italic>\ntarget genes with known inhibitory functions in angiogenesis, including <italic>Cd36</italic> and <italic>Cdkn1c</italic> [##REF##17641685##29##]. A crucial function of CD36 is to serve as a\nreceptor for thrombospondins which are known to attenuate the proliferation of\nendothelial cells [##REF##12714043##52##], and <italic>Cdkn1c</italic> codes for the cyclin-dependent kinase inhibitor p57^KIP2 [##REF##7729683##53##]. Consistent with the existence of a PPAR<italic>β</italic>/<italic>δ</italic> – p57^KIP2 pathway in stroma cell\ntypes, it was shown that the forced expression of PPAR<italic>β</italic>/<italic>δ</italic>\nin <italic>Pparb</italic> null fibroblasts results in\na <italic>Cdkn1c</italic>-dependend inhibition of cell\nproliferation [##REF##17641685##29##]. Other PPAR<italic>β</italic>/<italic>δ</italic>\ntarget genes with potential functions in cell proliferation and differentiation\nwere identified in the same study, suggesting that PPAR<italic>β</italic>/<italic>δ</italic>\nregulates multiple genes with functions in cell proliferation in the context of\ntumor stroma development and tumor angiogenesis.</p>", "<p>An antiproliferative effect of PPAR<italic>β</italic>/<italic>δ</italic>\nagonists in fibroblasts and vascular smooth muscle cells has also been observed\nin two other studies [##REF##17543901##27##, ##UREF##1##30##], while opposite effects have been described\nfor endothelial cells [##UREF##2##31##]. At present, it is difficult to explain these\napparent discrepancies, since they cannot be narrowed down to a single\nparameter, such as experimental strategy, cell type, expression level of PPAR<italic>β</italic>/<italic>δ</italic>,\nor state of the cell (e.g., metabolic activity, proliferative status, stage of\ndifferentiation, exogenous factors). This issue is discussed further in the\nConclusions section below.</p>", "<title>4. ROLE OF PPAR<italic>β</italic>/<italic>δ</italic> IN WOUND HEALING AND\nKERATINOCYTE PROLIFERATION</title>", "<p>\n<italic>Pparb</italic> null\nmice exhibit a defect in wound healing by inhibiting apoptosis in keratinocytes\n[##REF##11514592##8##]. This survival function of PPAR<italic>β</italic>/<italic>δ</italic> has been explained by an induction of AKT/protein kinase B (PKB)\nactivity by PPAR<italic>β</italic>/<italic>δ</italic> resulting from an upregulation of the <italic>Pdk1</italic> and <italic>Ilk</italic> genes and a\ndownregulation of <italic>Pten</italic> [##REF##12419217##11##]. Increased AKT signaling is generally\nassociated with enhanced proliferation, yet others have reported that PPAR<italic>β</italic>/<italic>δ</italic> inhibits cell proliferation [##REF##10866668##7##, ##REF##16109478##15##]. In this case, however, AKT activity was not\naffected by PPAR<italic>β</italic>/<italic>δ</italic> activation. Instead, a downregulation of protein kinase C and MAP\nkinase signaling was observed [##REF##15632134##14##]. The reason for these discrepancies is not\nclear at present, however, in light of the relatively small effects of PPAR<italic>β</italic>/<italic>δ</italic> on the signaling pathways discussed above it is possible that subtle\ndifferences in the experimental settings account for the apparent lack of\nconsistency.</p>", "<title>5. ROLE OF PPAR<italic>β</italic>/<italic>δ</italic> IN DIFFERENTIATION</title>", "<p>Mice lacking PPAR<italic>β</italic>/<italic>δ</italic> show a very high degree of embryonic lethality due to an aberrant\ndevelopment and malfunction of the placenta [##REF##10866668##7##, ##REF##11756685##9##, ##REF##16581799##10##]. Consistent with this finding, the\ndifferentiation and metabolic functions of trophoblast giant cells in vitro are dependent on PPAR<italic>β</italic>/<italic>δ</italic> [##REF##16581799##10##]. In the same model, stimulatory effect of PPAR<italic>β</italic>/<italic>δ</italic> on AKT signaling was observed. Another tissue where PPAR<italic>β</italic>/<italic>δ</italic> plays a role in differentiation is the digestive tract, where PPAR<italic>β</italic>/<italic>δ</italic> promotes the differentiation of Paneth cells in the intestinal crypts\nby down-regulating the hedgehog signaling pathway [##REF##16890607##24##]. A differentiation promoting effect of PPAR<italic>β</italic>/<italic>δ</italic> has also been described for keratinocytes, adipocytes, endothelial\ncells, and oligodendrocytes (see ##TAB##0##Table 1## for details).</p>" ]

[ "<title>ACKNOWLEDGMENT</title>", "<p>Work in the authors' laboratory was supported by grants\nfrom the Deutsche Forschungsgemeinschaft (SFB-TR17/A3 and Mu601-12).</p>" ]

[]

[ "<table-wrap id=\"tab1\" position=\"float\"><label>Table 1</label><caption><p> Effects of PPAR<italic>β</italic>/<italic>δ</italic>\non cell proliferation and differentiation.</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><th align=\"left\" rowspan=\"2\" colspan=\"1\"> Cell type </th><th align=\"left\" rowspan=\"2\" colspan=\"1\">\nExp. approach </th><th align=\"center\" colspan=\"2\" rowspan=\"1\"> Role of PPAR <italic>β</italic>/<italic>δ</italic> in</th><th align=\"center\" rowspan=\"2\" colspan=\"1\"> Affected pathway </th><th align=\"center\" rowspan=\"2\" colspan=\"1\"> References </th></tr><tr><th align=\"center\" rowspan=\"1\" colspan=\"1\"> Prolif. </th><th align=\"center\" rowspan=\"1\" colspan=\"1\"> Diff. </th></tr></thead><tbody><tr><td align=\"center\" colspan=\"6\" rowspan=\"1\">\n<italic>Epithelial\ncells</italic>\n</td></tr><tr><td align=\"center\" colspan=\"6\" rowspan=\"1\">\n<hr/>\n</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">Keratinocyte</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">Agonist, wt versus null</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">↘</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">↗</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">AKT</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">[##REF##12419217##11##]</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">Keratinocyte</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">Agonist, wt versus null</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">↘</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">↗</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">ERK</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">[##REF##11348458##12##–##REF##17254750##17##]</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">Keratinocyte</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">Agonist,\nRNAi, wt versus null</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">↗</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">\n</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">\n</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">[##REF##17713572##18##, ##REF##17637826##19##]</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">Adipocyte</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">Agonist,\nwt versus null</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">\n</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">↗</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">PPAR<italic>γ</italic>\n</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">[##REF##10991946##20##–##REF##15247146##22##]</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">Trophoblast</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">wt versus null</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">\n</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">↗</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">AKT</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">[##REF##16581799##10##, ##UREF##0##23##]</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">Paneth cells (in vivo) </td><td align=\"left\" rowspan=\"1\" colspan=\"1\">wt versus null</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">\n</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">↗</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">Hedgehog</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">[##REF##16890607##24##]</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">Hepatic stellate cell</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">Agonist</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">↗</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">\n</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">\n</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">[##REF##12512042##25##]</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">Oligodendrocyte</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">Agonist</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">\n</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">↗</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">\n</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">[##REF##11241737##26##]</td></tr><tr><td align=\"center\" colspan=\"6\" rowspan=\"1\">\n<hr/>\n</td></tr><tr><td align=\"center\" colspan=\"6\" rowspan=\"1\">\n<italic>Mesenchymal\ncells</italic>\n</td></tr><tr><td align=\"center\" colspan=\"6\" rowspan=\"1\">\n<hr/>\n</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">Fibroblast</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">Agonist</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">↘</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">(↘*)</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">\n<italic>G0S2**, PTEN</italic>\n</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">[##REF##17543901##27##]</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">Fibroblast</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">wt versus null, re-expression in null</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">↘</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">\n</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">p57^KIP2\n</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">[##REF##17380536##28##, ##REF##17641685##29##]</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">Vascular smooth muscle cells</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">Agonist</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">↘</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">\n</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">PDGF</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">[##UREF##1##30##]</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">Tumor endothelium</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">wt versus null</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">↘</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">↗</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">\n</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">[##REF##17641685##29##]</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">Endothelial cells</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">Agonist</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">↗</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">↗</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">\n</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">[##UREF##2##31##]</td></tr><tr><td align=\"center\" colspan=\"6\" rowspan=\"1\">\n<hr/>\n</td></tr><tr><td align=\"center\" colspan=\"6\" rowspan=\"1\">\n<italic>Human\ntumor cell lines</italic>\n</td></tr><tr><td align=\"center\" colspan=\"6\" rowspan=\"1\">\n<hr/>\n</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">MCF-7 breast carcinoma; UACC903 melanoma</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">Agonist</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">↘</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">\n</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">\n</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">[##REF##18054822##32##]</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">HT29, HCT116, LS-174T colon carcinoma;\nHepG2, HuH7 hepatoma</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">Agonist</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">↘</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">\n</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">\n</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">[##REF##17693664##33##]</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">HCT116 colon carcinoma</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">RNAi</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">↘</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">\n</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">\n</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">[##REF##18214615##34##]</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">SH-SY5Y neuroblastoma</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">Agonist</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">\n</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">↗</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">\n</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">[##REF##17390299##35##]</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">NSC lung carcinoma</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">Agonist</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">↗</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">\n</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">AKT, NF<italic>κ</italic>B</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">[##REF##18390835##36##]</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">A549 NSC lung ca.</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">Agonist</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">↘</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">\n</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">\n</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">[##REF##15978581##37##]</td></tr></tbody></table></table-wrap>" ]

[]

[]

[]

[]

[]

[]

[ "<table-wrap-foot><fn><p>*transdifferentiation\ninto myofibroblasts.</p></fn><fn><p>\n**G0S2: <italic>G0/G1 switch gene 2</italic> (cell cycle\ninhibitor).</p></fn></table-wrap-foot>" ]

[]

[]

{ "acronym": [], "definition": [] }

[ "<title>1. INTRODUCTION</title>", "<p>Peroxisome proliferator-activated receptor gamma (PPAR<italic>γ</italic>) is\na member of\nthe ligand-inducible nuclear receptor superfamily. After activation, PPAR<italic>γ</italic> associates\nwith the 9-cis retinoic acid receptor (RXR) to form functional heterodimers, which\nbinds to the PPAR response element of the target genes and regulates the\nexpression of these genes. Previous documents have shown that the PPAR<italic>γ</italic>/RXR\nsignal pathway plays critical role in a variety of biological processes, including\nadipogenesis, glucose metabolism, inflammation as well as inhibition of normal\nand tumor cells growth [##REF##10549292##1##].</p>", "<p>Thiazolidinediones (TZDs) are synthetic agonists for PPAR<italic>γ</italic>. These PPAR<italic>γ</italic> ligands\nwere clinically used as antidiabetic drugs which could attenuate the insulin resistance\nassociated with obesity, hypertension, and impaired glucose tolerance in humans\n[##REF##8922349##2##]. Recent studies have suggested that PPAR<italic>γ</italic> is a potential molecular target for\nanticancer drug development, due to the increased expression of PPAR in several\ncancer cells. It has been reported that TZDs could inhibit growth and induce\napoptosis in a variety of cancer cell lines. More importantly, TZDs exhibited\nantitumor activities in vivo in the prevention of prostate, liver, and\npituitary cancers. Although increasing evidence\nshowed that TZDs are potential anticancer agents [##REF##15231248##3##], the mechanisms underlying\nthe antitumor effects are not well understood. TZDs were initially thought\nto inhibit the cancer cells proliferation through regulation of expression of PPAR<italic>γ</italic>-mediated\ntarget genes. However, recent evidence\nrevealed that the antitumor effects of TZDs exist via PPAR<italic>γ</italic>-independent\nmechanisms in various types of cancers [##REF##16390557##4##–##REF##14563954##6##].</p>", "<p>We previously found the expression\nof PPAR<italic>γ</italic> decreased in primary and metastatic\ngastric carcinoma, compared with normal gastric tissues [##REF##17908458##7##]. Recent studies in\ngastric cancer cells demonstrated that TZDs treatment resulted in significant growth\narrest both in cultured cell and in nude mice models [##REF##11044367##8##–##REF##10558888##12##]; however, the\neffects of PPAR<italic>γ</italic> ligands on invasiveness and angiogenesis of gastric cancer are\nstill unclear. Therefore, this work was undertaken to investigate the effects\nof PPAR<italic>γ</italic> agonists, such as rosiglitazone, on cell growth and the invasiveness in\nhuman cell line SGC-7901, as well as on angiogenesis in vitro. </p>" ]

[ "<title>2. METHODS</title>", "<title>2.1. Cell culture</title>", "<p>Human gastric cancer cell line, SGC-7901, was obtained from the Type Culture Collection\nof Chinese Academy of Sciences (Shanghai, China). Human\numbilical vein endothelial cells (HUVECs) were purchased from the Keygen Technology\nCompany (Najing, China). SGC7901 cells and HUVECs were\ncultured in RPMI-1640 medium (GIBCO, Carlsbad, Calif, USA)\ncontaining 10% fetal bovine serum (FBS) and 1% antibiotics (100 U/mL penicillin G, 100 <italic>μ</italic>g/mL streptomycin\nsulfate, Sigma-Aldrich, Mo, USA).</p>", "<title>2.2. RT-PCR</title>", "<p>Total RNA was isolated using TRIzol\nReagent (Invitrogen, Carlsbad, Calif, USA)\naccording to the manufacturer's instructions. Reverse transcription reaction\nwas performed with random hexamer primers and a SuperScript Reverse\ntranscriptase kit (Invitrogen, Carlsbad, Calif, USA).\nThe sequences of specific primers were as follows: PPAR<italic>γ</italic> mRNA, forward, 5′-TCT CTC CGT AAT GGA AGA CC-3′, and reverse, 5′-GCA TTA TGA GAC ATC CCC AC-3′. MMP-2 mRNA, forward, 5′-GGC CCT GTC ACT CCT GAG AT-3′, and reverse, 5′-GGC ATC CAG\nGTT ATC GGG GA-3′.\nVEGF mRNA, forward,\n5′-GAC AAg AAA ATC CCT GTG GGC-3′,\nand reverse 5′-AAC GCG AGT CTG TGT TTT TGC-3′. <italic>β</italic>-actin mRNA, forward, 5′-CTT CTA CAA TGA GCT\nGCG TA-3′, and reverse, 5′-TCA\nTGA GGT AGT CAG TCA GG-3′.\nPCR conditions were 94°C, 30 seconds, 55–57°C (depending on the primer set), 30 seconds,\nand 72°C, 1 minute with 35 cycles using\nTaq PCR MasterMix (Tianwei, Beijing, China). The resultant PCR products were\n474 bp (PPAR<italic>γ</italic>), 243 bp (<italic>β</italic>-actin), 474 bp (MMP-2), and 169 bp\n(VEGF). PCR products were electrophoresed on a 1.2% agarose gel and visualized\nby ethidium bromide staining.</p>", "<title>2.3. Quantitative\nreal-time RT-PCR analysis</title>", "<p>The PCR\nreactions were performed in a Brilliant\nSYBR Green QPCR master mix (Stratagene, Calif, USA) according\nto the manufacturer's instructions. The sequences of specific primers\nwere the same as for RT-PCR. After 10 minutes at 95°C to denature the cDNA, the cycling conditions were 95°C, 1 minute, 55–57°C (depending on the primer\nset), 30 seconds, and 72°C, 1 minute with 40\ncycles. The LightCycler software constructed the calibration curve by plotting\nthe crossing point (Cp), and the numbers of copies in unknown samples were\ncalculated by comparison of their Cps with the calibration curve. To correct\ndifferences in both RNA quality and quantity between samples, the data were formalized\nto those for <italic>β</italic>-actin.</p>", "<title>2.4. Western blotting</title>", "<p>The cells proteins were extracted\naccording to NE-PER Nuclear and Cytoplasmic Extraction Reagents kit (Pierce, Rockford, Ill, USA). Protein\nconcentration of each sample was assayed using BCA Protein Assay Reagent\naccording to manufacturer's instructions (Pierce Biotechnology, Rockford, Ill,\nUSA). Twenty\nmicrograms of proteins of different groups were separated in 10% SDS-PAGE, and\ntransferred onto PVDF membrane (Invitrogen, Carlsbad, Calif, USA). Five percent of milk\n(blocking solution) was loaded over the membrane and incubated for 1 hour at\nroom temperature with agitation. The membranes were then incubated with the mouse\nantihuman PPAR<italic>γ</italic> antibody at a dilution of 1:200 (Santa\nCruz, Calif, USA), the mouse antihuman\nMMP-2 antibody (1:400, Neomarker, Calif, USA), the rabbit antihuman VEGF\nantibody (1:200, Zymed, Calif, USA), and\nthe mouse antihuman <italic>β</italic>-actin\n(1:200, Xiaxin, China) overnight at 4°C with agitation.\nAfter being washed with 0.1% Tween 20 in Tris-saline, three times, the membranes were incubated with biotin-labeled\nantirabbit or mouse IgG for 1 hour at room temperature with agitation. Reactive\nprotein was detected using ECL chemiluminescence system (Pierce, Rockford, Ill, USA).</p>", "<title>2.5. ELISA of secreted VEGF</title>", "<p>The effect of RGZ on VEGF release\nin tumor cells was measured by ELISA. Cells grown in 90 mm plates were exposed to various concentrations\nof RGZ (1–20 <italic>μ</italic>M) or\nvehicle with or without GW9662 (2.5 <italic>μ</italic>M, pretreated 1 hour) for 24 hours. VEGF concentration in the supernatant was\nmeasured using a VEGF ELISA kit (R &amp; D systems, Minneapolis, Minn, USA).</p>", "<title>2.6. Cell viability</title>", "<p>The viability of the cells was\nassessed by MTT assay. Briefly, cells grown in 96-wells\nwere exposed to various concentrations of RGZ with or without GW9662 (2.5 <italic>μ</italic>M, pretreated 1 hour), for 24, 48, or 72 hours. Then, 20 <italic>μ</italic>L of MTT (5 mg/mL) was\nadded to each well, and cells were incubated continuously at 37°C for 4 hours. After\nremoval of medium, the crystals were dissolved in DMSO, and absorbance was\nassessed at 570 nm with a microplate reader.</p>", "<title>2.7. Cell cycle and apoptosis analysis</title>", "<p>Cells treated with RGZ (1–20 <italic>μ</italic>M) or vehicle with or\nwithout GW9662 (2.5 <italic>μ</italic>M,\npretreated 1 hour) for 48 hours were collected and fixed in cold 70% ethanol.\nThen, the samples were treated with RNase, stained with 50 mg/mL propidium\niodide (PI), and analysed by EPICS Elite flow cytometer (Coulter Electronics, Fla, USA).</p>", "<title>2.8. Invasion assay</title>", "<p>The ability of cells to invade through a\nMatrigel-coated filter was measured in transwell chambers (Corning, NY, USA). Polyvinylpyrrolidone-free\npolycarbonate filters (pore size 8 <italic>μ</italic>m)\nwere coated with basement membrane Matrigel (50 <italic>μ</italic>L/filter) (BD, Bedford, Ohio, USA).\nThe membrane was washed in PBS to remove excess ligand, and the lower chamber\nwas filled with 0.6 mL of RPMI-1640 medium containing 10% fetal bovine serum\n(FBS). Cells were serum-starved overnight (0.5% FBS), harvested with\ntrypsin/EDTA, and washed twice with serum-free RPMI-1640 medium. Cells were\nresuspended in migration medium (RPMI-1640 medium with 0.5% FBS), and 0.1 mL migration\nmedium containing 1 × 10⁵ cells was added to the upper chamber. After incubation with RGC (1–20 <italic>μ</italic>M) with or without GW9662\n(2.5 <italic>μ</italic>M, pretreated\n1 hour) at 37°C for 24 hours, the cells on the upper surface of the membrane were removed using\na cotton swab. The migrant cells attached to the lower surface were fixed in\n10% formalin at room temperature for 30 minutes and stained with hematoxylin.\nThe numbers of migrated cells were counted under a microscope.</p>", "<title>2.9. Scratch wound-healing motility\nassays</title>", "<p>Gastric cancer cells were seeded on 60 mm plates and allowed to grow to\nconfluence. Confluent monolayers were scratched with a pipette tip and\nmaintained under RGZ (1–20 <italic>μ</italic>M) with or without GW9662\n(2.5 <italic>μ</italic>M, pretreated\n1 hour) for 24 hours. Plates were washed once with fresh medium to remove nonadherent\ncells and then photographed. The cell migration was evaluated by counting cells\nthat migrated from the wound edge.</p>", "<title>2.10. In vitro Angiogenesis assay</title>", "<p>The angiogenesis assays were performed as per the manufacturer's\ninstructions, that is, transfer 50 <italic>μ</italic>L\nof ECMatrixTM solution to each well of a precooled 96-well tissue culture plate\non ice. Incubate at 37°C for 1 hour to\nallow the matrix solution to solidify. Harvest human umbilical vein endothelial\ncells (HUVECs) resuspend and Seed 5 × 10³ cells per well onto the\nsurface of the polymerized ECMatrixTM. Incubate with RGC (1–20 <italic>μ</italic>M) with or without GW9662\n(2.5 <italic>μ</italic>M, pretreated\n1 hour) at 37°C for 12 hours. Inspect tube formation under an inverted light microscope at 100 X\nmagnification.</p>", "<title>2.11. Zymography</title>", "<p>Cells were cultured for 24 hours in serum-free\nmedium, washed twice, and finally treated with RGZ (1–20 <italic>μ</italic>M) with or without GW9662\n(2.5 <italic>μ</italic>M, pretreated\n1 hour) for a further 48 hours. The supernatants were collected and concentrated,\nusing centrifugal filter devices (Millipore Corp., Bedford, Mass, USA)\nand the protein content was determined using BCA Protein Assay Reagent. Equal\namounts of protein (20 <italic>μ</italic>g) were mixed with SDS sample buffer without reducing\nagents and incubated for 40 minutes at 37°C. For gelatinolytic activity, the assay\nsamples were separated on polyacrylamide gels containing 1mg/mL gelatin. After\nelectrophoresis, the gels were stained for 1 hour in a 45% methanol/10%\nacetic acid mixture containing coomassie brilliant blue G250 and destained.\nZymograms were photographed after 10 hours of incubation at 37°C.</p>", "<title>2.12. Statistical analysis</title>", "<p>Data are expressed as mean ± standard\ndeviation (SD) of three independent experiments, each done in triplicate. Differences\nbetween control and experiment groups were analyzed using the <italic>t</italic>-test. <italic>P</italic> &lt; .05 was considered\nstatistically significant.</p>" ]

[ "<title>3. RESULTS</title>", "<title>3.1. RGZ inhibited proliferation and\ninduced apoptosis in SGC-7901 cells through PPAR<italic>γ</italic>-dependent mechanism</title>", "<p>In SGC-7901 cells, the expression of PPAR<italic>γ</italic> was observed by RT-PCR and western blot (not\nshown).</p>", "<p>RGZ (0.1–100 <italic>μ</italic>M) treatment for 24, 48, and\n72 hours inhibited cells growth in a dose- and time-dependent manners in\nSGC-7901 gastric cancer cell line as determined by MTT assay. Pretreatment with\nthe highly selective PPAR<italic>γ</italic> antagonist GW9662\n(2.5 <italic>μ</italic>M) reversed the\neffect of RGZ on cell viability (see ##FIG##0##Figure 1(a)##).</p>", "<p>To explore whether the growth inhibition of RGZ in SGC-7901 cells was\ncaused by apoptosis, we analyzed the sub-G1 population of the cells after\ntreatment with RGZ (1–20 <italic>μ</italic>M) for 48 hours. RGZ induced\napoptosis in a dose-dependent manner, which was also reversed completely by 2.5 <italic>μ</italic>M GW9662 treatment (see ##FIG##0##Figure 1(b)##).</p>", "<p>Furthermore, to determine whether the inhibitory effect of RGZ on cell\nviability is associated with the arrest of the cell cycle, we analyzed the cell\ncycle progression after treatment with RGZ (1–20 <italic>μ</italic>M) for 48 hours. RGZ\ntreatment increased the number of cells in the G1-G0 and decreased\nthe number of cells in the S phases in dose-dependent manner. The effects of\nRGZ on cell cycle of SGC-7901 cells were also reversed by 2.5 <italic>μ</italic>M GW9662 (see ##FIG##0##Figure 1(c)##).</p>", "<title>3.2. RGZ inhibited SGC-7901 cells\nmigration and invasiveness through PPAR<italic>γ</italic>-independent mechanism</title>", "<p>After treatment with RGZ (1–20 <italic>μ</italic>M) for 48 hours, the number\nof cells migrated to the scratched area was 60 ± 3.1 cells/mm², 58 ± 2.7 cells/mm², 49 ± 2.8 cells/mm², 27 ± 2.9 cells/mm², and 20 ± 1.9 cells/mm², respectively, which were\nsignificantly lower than those in control group (84 ± 3.4 cells/mm²). GW9662 treatment had no\neffects on the cells migration with inhibition induced by RGZ. The number of\nthe cells migrated to the scratched area treated with GW9662 and RGZ (1–20 <italic>μ</italic>M) for 48 hours was 61 ± 1.8 cells/mm², 53 ± 3 cells/mm², 47 ± 2.5 cells/mm²,\n29 ± 2.8 cells/mm², 18 ± 3.2 cells/mm², respectively, which\nwere not different from those in the groups treated with RGZ alone (see ##FIG##1##Figure 2(a)##).</p>", "<p>The effect of RGZ on the cells invasion through reconstituted basement\nmembranes was analyzed using Matrigel-coated invasion chambers. After treatment with RGZ (1–20 <italic>μ</italic>M) for 48 hours,\nthe cells attached to the lower surface of the filters were 256 ± 9 cells/mm², 248 ± 7 cells/mm², 219 ± 12 cells/mm², 174 ± 11 cells/mm², and 154 ± 10 cells/mm², respectively,\nwhich were significantly lower than those in control group (279 ± 9 cells/mm²). After cotreatment of the cells with\nGW9662 and RGZ, the cells attached to the lower surface were 251 ± 29 cells/mm², 238 ± 12 cells/mm², 220 ± 7 cells/mm², 166 ± 16 cells/mm², and 148 ± 12 cells/mm², respectively,\nwhich were not different from those in the groups treated with RGZ alone\n(see ##FIG##1##Figure 2(b)##).</p>", "<p>Metalloproteases (MMPs) have been\ndemonstrated to play a significant role in tumor cell invasion [##REF##11322833##13##]. In this\nstudy, our results showed that RGZ inhibited the mRNA and protein expression\nlevels of MMP-2 in a\ndose-dependent manner (see Figures ##FIG##2##3(a)##, ##FIG##2##3(c)##, and Tables ##TAB##0##1##, ##TAB##1##2##). Moreover, the gel zymography\nresults demonstrated that the activity of MMP-2 decreased after RGZ (1–20 <italic>μ</italic>M) treatment\nfor 48 hours in dose-dependent manner (see ##FIG##3##Figure 4(a)##). The inhibitory effects of\nRGZ on MMP-2 were not affected by GW9662 treatment (see Figures ##FIG##2##3(b)##, ##FIG##2##3(c)##, and ##FIG##3##4(b)##).</p>", "<title>3.3. Effects of RGZ on angiogenesis in vitro</title>", "<p>Matrigel-plated HUVECs elongated and migrated in the presence\nof VEGF and formed tubular networks. RGZ markedly suppressed the formation of the\ntube-like structures of HUVEC cells in a dose-dependent manner (see ##FIG##4##Figure 5(a)##),\nwhich was completely antagonized by GW9662 (see ##FIG##4##Figure 5(b)##). These results suggested\nthat rosiglitazone exhibits antiangiogenic activity via PPAR<italic>γ</italic>-dependent mechanism.</p>", "<p>To further determine whether the effect of RGZ on\nangiogenesis is due to the down regulation of the tumor-secreted growth\nfactors, we measured the expression levels of VEGF in SGC-7901 cell cultured\nmedium, after treatment with various concentrations of RGZ. Our results\ndemonstrated that RGZ (1–20 <italic>μ</italic>M) did not\nchange the expression of mRNA and protein of VEGF in SGC-7901 cells (see Figures ##FIG##2##3(a)##, ##FIG##2##3(c)##, and ##TAB##0##Table 1##), but also the\nresults were confirmed by ELISA (see ##FIG##5##Figure 6##).</p>" ]

[ "<title>4. DISCUSSION</title>", "<p>As a\npotential molecular target for anticancer drug development, PPAR<italic>γ</italic> and its\nligands have been extensively studied in the past several years. Previous\nstudies have shown that PPAR<italic>γ</italic> is expressed in several human gastric-cancer cell\nlines, including MKN-7, MKN-28, MKN-45, and AGS. TZDs could inhibit these cancer cell lines\ngrowths in vitro and in vivo [##REF##14960510##9##, ##REF##10558888##12##]. Also,\nthe growth inhibitory effects of TZDs on MKN45 cells depend on the PPAR<italic>γ</italic>\nexpression levels. The growth inhibition\nof TDZs was more significant in the higher PPAR<italic>γ</italic> expressing cells. Moreover, Lu et al.\n[##REF##15930296##10##] found that PPAR<italic>γ</italic> (+/ − ) mice were more susceptible to MNU-induced gastric\ncancer than wild-type (+/+) mice, and troglitazone significantly reduced the\nincidence of gastric cancer in PPAR<italic>γ</italic> (+/+) mice but not in PPAR<italic>γ</italic> (+/ − ) mice. All\nthese results indicated that TZDs inhibit the cancer cells growth via PPAR<italic>γ</italic>-dependent\nmechanism. Our results demonstrated that RGZ, the most potent and selective\nsynthetic ligand of PPAR<italic>γ</italic>, inhibited SGC-7901\ngastric cancer cells growth, caused G1 cell cycle arrest, and induced apoptosis in\na dose-dependent manner. The effects of RGZ on SGC-7901 cancer cells were\ncompletely reversed by treatment with PPAR<italic>γ</italic> antagonist GW9662. These results\nindicated that RGZ suppressed the SGC-7901 cancer cells growth in a PPAR<italic>γ</italic>-dependent\nmechanism.</p>", "<p>In this study, we found that the RGZ inhibited invasion, migration, and\nthe secretion of MMP-2 of SGC-7901 cells. The inhibitory effects of RGZ on metastases and\nMMP-2 activity were not directly mediated by PPAR<italic>γ</italic> activation,\nsince these effects were not reversed by GW9662 treatment. Our results were\nconsistent with the previous works on human adrenocortical cancer cell line\nH295R [##REF##15585569##14##], pancreatic cancer cells [##REF##15479693##15##], and human myeloid leukemia cells [##REF##15838654##16##], which\nshowed that PPAR<italic>γ</italic> ligands act independently of PPAR<italic>γ</italic> activation in the invasion\nsuppression and down-regulation of MMP-2 activity. Recent papers\nshowed that PPAR<italic>γ</italic> regulated E-cadherin expression and inhibited\ngrowth and invasion of prostate cancer [##REF##17015477##17##], and PPAR<italic>γ</italic> ligand\ntroglitazone inhibited transforming growth factor-beta-mediated glioma cell\nmigration and brain invasion [##REF##17541035##18##]. But\nsome studies have contrasting results that the PPAR<italic>γ</italic>, ciglitazone, induced cell invasion, through\nactivation of Pro-MMP-2, activation via the generation of ROS, and the activation\nof ERK [##REF##17597617##19##], and that PPAR<italic>γ</italic> antagonists induced vimentin cleavage and inhibited\ninvasion in high-grade hepatocellular carcinoma [##REF##17786342##20##]. Further studies are\nneeded on the mechanism of PPAR<italic>γ</italic> in cancer and invasion.</p>", "<p>Recent\ninvestigations suggested that PPAR<italic>γ</italic> ligands had inhibitory effects on tumor\ncell lines, but the effects appear not to be entirely elicited by the direct\naction on tumor cells. Inhibition of the neovascularization may be another\ntarget of TZDs to suppress the growth of cancers. PPAR<italic>γ</italic> is expressed in\nendothelial cells, and the PPAR<italic>γ</italic> ligands can inhibit the proliferation of these\ncells induced by growth factors, or cause their apoptosis in vitro [##REF##10358055##21##–##REF##10854212##23##]. It has been\nreported that PPAR<italic>γ</italic> ligands could inhibit choroidal, retinal, and corneal\nneovascularization when administered intraocularly [##REF##10892878##24##–##REF##15721625##26##]. In addition,\nsystemic administration of rosiglitazone and troglitazone inhibits FGF2-induced\nangiogenesis; thereby inhibiting primary tumor growth and metastasis [##REF##12370270##27##]. We observed\nthat RGZ inhibited\nthe angiogenesis of HUVECs in dose-dependent manner via PPAR<italic>γ</italic>\npathway. The effects RGZ on the endothelium suggest that RGZ may\nregulate tumor growth by targeting non-cell-autonomous mechanisms.</p>", "<p>Previous studies [##REF##12475986##5##]\nshowed that suppression of angiogenesis could result from a decrease in the\nlocal levels of stimulators (e.g., VEGF and FGF2) and/or an increase of\nendogenous inhibitors of angiogenesis (e.g., thrombospondin) produced by tumor\ncells. PPAR<italic>γ</italic> ligands suppressed\nVEGF production in colon carcinoma [##REF##15289320##28##], human\nbreast cancer [##REF##17384282##29##], and human renal cell carcinoma cells [##REF##15780399##30##]. However, contradictory\nresults have also been reported in bladder and prostate cancer cells in which PPAR<italic>γ</italic> ligands\nincreased VEGF production [##REF##11980898##31##, ##REF##12239635##32##]. Inconsistent with the above documents, our\nresults showed that RGZ did not change the secretion of VEGF from SGC-7901.</p>", "<p>Taken together, our results demonstrated that RGZ inhibited growth and invasiveness\nof SGC-7901 gastric cancer cells and angiogenesis in vitro via PPAR<italic>γ</italic>-dependent or -independent pathway. Further\nstudy is needed to elucidate the\nmechanisms by which RGZ exhibits different manner.</p>" ]

[]

[ "<p>Recommended by Dipak Panigrahy</p>", "<p>Although thiazolidinediones (TZDs) were found to be ligands for peroxisome proliferators-activated receptor<italic>γ</italic> (PPAR<italic>γ</italic>), the mechanism by which TZDs exert their anticancer effect remains unclear. Furthermore, the effect of TZDs on metastatic and angiogenesis potential of cancer cells is unknown. Our results in this paper show that rosiglitazone inhibited SGC-7901 gastric cancer cells growth, caused G1 cell cycle arrest and induced apoptosis in a dose-dependent manner. The effects of rosiglitazone on SGC-7901 cancer cells were completely reversed by treatment with PPAR<italic>γ</italic> antagonist GW9662. Rosiglitazone inhibited SGC-7901 cell migration, invasiveness, and the expression of MMP-2 in dose-dependent manner via PPAR<italic>γ</italic>-independent manner. Rosiglitazone reduced the VEGF induced angiogenesis of HUVEC in dose-dependent manner through PPAR<italic>γ</italic>-dependent pathway. Moreover, rosiglitazone did not affect the expression of VEGF by SGC-7901 cells. Our results demonstrated that by PPAR<italic>γ</italic> ligand, rosiglitazone inhibited growth and invasiveness of SGC-7901 gastric cancer cells and angiogenesis in vitro via PPAR<italic>γ</italic>-dependent or -independent pathway.</p>" ]

[]

[ "<title>ACKNOWLEDGMENTS</title>", "<p>This work was supported by the National Natural\nScience Foundation of China (Grant no. 30671904, 30670949), China Postdoctoral\nScience Foundation (no. 2004035181), and The Doctor Station of Ministry of\nEducation of China (no. 20060558010).</p>" ]

[ "<fig id=\"fig1\" position=\"float\"><label>Figure 1</label><caption><p>(a) RGZ (0.1–100 <italic>μ</italic>M) treatment for 24, 48, and\n72 hours inhibited cell growth in a dose- and time-dependent manners in\nSGC-7901 gastric cancer cell line, as determined by MTT assay, which was reversed\ncompletely by 2.5 <italic>μ</italic>M GW9662 pretreatment for 1 hour. Cell\nviability was expressed as the percentage of cells under control conditions (0 <italic>μ</italic>M\nof RGZ or GW9662). (b) RGZ induced apoptosis in a dose-dependent manner,\nwhich was also reversed completely by 2.5 <italic>μ</italic>M GW9662 pretreatment for 1 hour. (c) RGZ treatment increased the number of\ncells in the G1-G0 and decreased\nthe number of cells in the S phases in dose-dependent manner, which was reversed\ncompletely by 2.5 <italic>μ</italic>M GW9662 pretreatment for 1 hour. Values\nare the means ± SD of three representative experiments.*Statistical significance (<italic>P</italic> &lt; .05 or\nhigher degree of significance) versus vehicle-treated controls.</p></caption></fig>", "<fig id=\"fig2\" position=\"float\"><label>Figure 2</label><caption><p>(a) Effect of RGZ on the migration and (b) invasion\nof SGC-7901 gastric cancer cells, which was reversed completely by 2.5 <italic>μ</italic>M GW9662\npretreatment for 1 hour. Values are the means ± SD of three representative experiments.*Statistical significance (<italic>P</italic> &lt; .05 or\nhigher degree of significance) versus vehicle-treated controls.</p></caption></fig>", "<fig id=\"fig3\" position=\"float\"><label>Figure 3</label><caption><p>(a) RGZ (1–20 <italic>μ</italic>M) inhibited\nthe mRNA and (c) protein expression levels of MMP-2 in a dose-dependent manner, which were not affected by 2.5 <italic>μ</italic>M GW9662 pretreatment\nfor 1 hour (b), (c). RGZ (1–20 <italic>μ</italic>M) did not\nchange the expression of VEGF in SGC-7901 cells (a), (c).</p></caption></fig>", "<fig id=\"fig4\" position=\"float\"><label>Figure 4</label><caption><p>(a)The activity of MMP-2 was\ndecreased after RGZ (1–20 <italic>μ</italic>M) treatment\nfor 48 hours in dose-dependent manner. (b) The inhibitory effects of RGZ on MMP-2\nwere not affected by 2.5 <italic>μ</italic>M GW9662 pretreatment for 1 hour.</p></caption></fig>", "<fig id=\"fig5\" position=\"float\"><label>Figure 5</label><caption><p>(a) RGZ markedly suppressed the formation of the\ntube-like structures of HUVEC cells in a dose-dependent manner, (b) which was\ncompletely antagonized by 2.5 <italic>μ</italic>M GW9662 pretreatment\nfor 1 hour.</p></caption></fig>", "<fig id=\"fig6\" position=\"float\"><label>Figure 6</label><caption><p>RGZ had no effect on the secretion of VEGF of SGC-7901 cell.</p></caption></fig>" ]

[ "<table-wrap id=\"tab1\" position=\"float\"><label>Table 1</label><caption><p>Expression of MMP-2 and VEGF after RZD treatment\nin SGC-7901 gastric cancers by real-time PCR.</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><th align=\"left\" rowspan=\"1\" colspan=\"1\">Rosiglitazone (<italic>μ</italic>mol/L)</th><th align=\"center\" rowspan=\"1\" colspan=\"1\">PPAR<italic>γ</italic>\n</th><th align=\"center\" rowspan=\"1\" colspan=\"1\">MMP-2</th><th align=\"center\" rowspan=\"1\" colspan=\"1\">VEGF</th></tr></thead><tbody><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">0</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">0.132127 ± 0.045513</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">0.008912 ± 0.000133</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">0.61132 ± 0.078921</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">1</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">0.121878 ± 0.034219</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">0.006003 ± 0.000331*</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">0.620255 ± 0.054671</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">3</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">0.130134 ± 0.0521137</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">0.005486 ± 0.000541*</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">0.60728 ± 0.036799</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">5</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">0.137778 ± 0.046222</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">0.005048 ± 0.000346*</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">0.599438 ± 0.076541</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">10</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">0.141171 ± 0.038741</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">0.001924 ± 0.000189*</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">0.624165 ± 0.038966</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">20</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">0.143889 ± 0.061237</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">0.001298 ± 0.000267*</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">0.604246 ± 0.065679</td></tr></tbody></table></table-wrap>", "<table-wrap id=\"tab2\" position=\"float\"><label>Table 2</label><caption><p>Expression of MMP-2 and VEGF after RZD and GW9662\ncotreatment in SGC-7901 by real-time PCR.</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><th align=\"left\" rowspan=\"1\" colspan=\"1\">Rosiglitazone (<italic>μ</italic>mol/L)</th><th align=\"center\" rowspan=\"1\" colspan=\"1\">PPAR<italic>γ</italic>\n</th><th align=\"center\" rowspan=\"1\" colspan=\"1\">MMP-2</th></tr></thead><tbody><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">0</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">0.14161 ± 0.055389</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">0.00975 ± 0.000533</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">1</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">0.137738 ± 0.030102</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">0.008974 ± 0.000113*</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">3</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">0.134614 ± 0.029881</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">0.006003 ± 0.000401*</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">5</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">0.141156 ± 0.564569</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">0.00564 ± 0.000246*</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">10</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">0.135666 ± 0.034887</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">0.002182 ± 0.000364*</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">20</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">0.129278 ± 0.019262</td><td align=\"center\" rowspan=\"1\" colspan=\"1\">0.001712 ± 0.000178*</td></tr></tbody></table></table-wrap>" ]

[]

[]

[]

[]

[]

[]

[ "<table-wrap-foot><fn id=\"TF1\"><p>*Statistical significance (<italic>P</italic> &lt; .05\nor higher degree of significance) versus\nvehicle-treated controls.</p></fn></table-wrap-foot>", "<table-wrap-foot><fn id=\"TF2\"><p>*Statistical significance (<italic>P</italic> &lt; .05\nor higher degree of significance) versus vehicle-treated controls.</p></fn></table-wrap-foot>" ]

[ "<graphic xlink:href=\"PPAR2008-649808.001\"/>", "<graphic xlink:href=\"PPAR2008-649808.002\"/>", "<graphic xlink:href=\"PPAR2008-649808.003\"/>", "<graphic xlink:href=\"PPAR2008-649808.004\"/>", "<graphic xlink:href=\"PPAR2008-649808.005\"/>", "<graphic xlink:href=\"PPAR2008-649808.006\"/>" ]

[]

{ "acronym": [], "definition": [] }

[ "<title>Introduction</title>", "<p>The fiber fuse effect was first reported in the late 1980's ##UREF##0##[1]##, ##REF##19746030##[2]##. It is initiated by the local heating of an optical fiber, which delivers a few watts of light, and generates an optical discharge running along the fiber to the light source at a speed of about 1 m/s (see the movie in the online version of ##REF##19498651##[3]## (##FIG##0##Fig. 1##) that shows a macroscopic view of fiber fuse propagation). This results in the catastrophic destruction of the core region. Thus, it has posed a real threat to every application where high power light is delivered through optical waveguides. ##SUPPL##0##Text S1## is a full list of related 70 papers.</p>", "<p>The behavior of the rapidly moving optical discharge is the key to recognizing the destruction mode. Thus, ultra-high speed videography at nearly half a million frames per second is the most suitable tool for observing this behavior. With the aid of this technique, I investigated a collection of micrographs showing the optical discharge and the voids that remained after quenching to determining the periodic void formation mechanism in a single mode silica glass fiber pumped at 1480 nm ##REF##19498651##[3]##, ##REF##16208896##[4]##, ##REF##19503125##[5]##.</p>", "<p>In this previous work, the videography only showed that the optical discharge moved at a constant speed under a time resolution of 4 µs. In this paper, a state-of-the-art camera is used to observe the oscillating light emission of an optical discharge, which is found to be related to an irregular void track that remains over hetero-core splice point. In addition, a bullet-like void pointing in the opposite direction to the rest of voids was accidentally observed near the splice point although the propagation direction of the optical discharge had not changed. This information is useful for analyzing fiber fuse accidents involving high-power fiber lasers.</p>", "<p>Japanese translation of this paper is available in ##SUPPL##1##Text S2##.</p>" ]

[ "<title>Methods</title>", "<p>\n##FIG##0##Figure 1\n## shows the experimental setup for observing fiber fuse propagation over a hetero-core splice. One end of a commercial single-mode silica glass optical fiber (SMF-28e, Corning, see ##TAB##0##Table 1##) was connected to a Yb fiber laser (PLM-10-1070, IPG Laser, 1.07 µm, 9 W) with an HI 1060, Corning, output fiber (see ##TAB##0##Table 1##). The other end of the fiber was spliced to another HI 1060 fiber. These two points are referred to as (a) and (b), respectively, in this paper. Hetero-core splicing was performed with a fusion splicer (S183, Furukawa Electric) containing a program for this pair.</p>", "<p>A fiber fuse was initiated at the end of a fiber attached to a metallic plate. The propagation over one of the two splice points was observed through an ultra-high speed CCD camera (FASTCAM SA1.1, monochrome version, Photron Ltd., sensitivity range: 380–790 nm) with an appropriate zoom lens. Pictures with a resolution of 256×32 with a 1024-step gradation were taken every 2.78 µs (360,000 frames per second) with an exposure time of 0.37 µs through a neutral density (ND) filter (x8). The damaged sites were examined with an optical microscope.</p>" ]

[ "<title>Results</title>", "<p>Ten examples of fiber-fuse propagation were recorded and they all showed a similar tendency. Typical recordings are shown in ##FIG##1##Fig. 2\n## and ##FIG##2##\n3\n##. The top row of the figures show visible light being emitted from the optical discharge running over one of the splice points. Since the fiber acted as a cylindrical lens, these images were expanded in the vertical direction. The middle row (B) shows a time-varying intensity profile of the emission along the dashed line shown in the top row.</p>", "<p>The bottom photograph (C) shows the damage train that remained after the propagation. Each splice point is estimated to be near the solid white circle. The void train shows that the propagation mode of the optical discharge is modulated at around the splice point. Accordingly, its velocity changed twice, namely before and after it reached the splice point. The inflection points of its velocity were calculated by analyzing the video image and are indicated as two dashed vertical lines in the middle row (B), which represent the peak of the intensity profile at the moment of speed change.</p>", "<p>The averaged speed, <italic>v</italic>, in each segment is listed in ##TAB##1##Table 2##. The speed in HI 1060 on the downstream side is smaller than that of the upstream side because the pumping energy is reduced through two hetero-core splice points with an inevitable insertion loss.</p>", "<p>At the tail of the intensity profile shown in ##FIG##1##Fig. 2B## and ##FIG##2##\n3B##, a small discrete peak appeared periodically and moved to rearward with decreasing height unless the discharge stayed near the splice point. This sub-peak oscillation was recorded in my previous work ##REF##19498651##[3]## but was not analyzed precisely owing to the poor resolution (256-step gradation taken every 4 µs) of the video camera. The oscillation cycle of this peak coincides with the period of one void formation, τ, listed in ##TAB##1##Table 2##, which is calculated from the average speed and void interval.</p>", "<p>Sometimes a single reverse bullet appeared just before the discharge running into the splice point (a) (see ##FIG##1##Fig. 2C##). Its expanded image is shown in ##FIG##3##Fig. 4A## together with other micrographs of the samples obtained under the same condition.</p>" ]

[ "<title>Discussion</title>", "<title>Change in void pattern over splice point</title>", "<p>Since the pump light (1070 nm) propagates in a multimode in SMF-28e, its energy density is lower than that in HI 1060. Thus, the propagation speed of the optical discharge is slower in SMF-28e than in HI 1060 as shown in ##TAB##1##Table 2##. The reason for its speed changing twice, namely before and after passing over the splice point, is that a discharge with a length more than 100 µm running over the border remains in both HI 1060 and SMF-28e, i.e., in a transitional state.</p>", "<p>In this transitional segment (shown between two vertical dashed lines in ##FIG##1##Fig. 2\n## and ##FIG##2##\n3\n##), the splice point (shown as a solid white circle in these figures) is located on the downstream side of the pump laser. This is because the intensity profile of the fiber fuse emission along the fiber length is unsymmetrical.</p>", "<p>Outside the transitional segment, the optical discharge generates a periodic void train. On passing through the splice point (b), the optical discharge reduces the void formation frequency (see ##TAB##1##Table 2##). However, the interval of the periodic voids varies discontinuously around the border to form a void-free segment.</p>", "<p>Considering the energy balance of fiber fuse propagation, the input is the pump laser energy and the outputs are the heating of materials, the emission of light and heat, the movement of the optical discharge and the creation of the void surface. Thus, in the transitional segment, these balance is modified to establish another equilibrium state. During the passage across the border (b), the optical discharge suspends void formation and slightly enhances the light emission. A similar phenomenon is observed in fiber fuse self-termination ##REF##19503125##[5]##.</p>", "<p>On the other hand, the frequency of periodic void formation increases after the optical discharge passed through the border (a). Then, the discharge in the transitional segment enhances the surface formation to construct a row of long voids with a reduction in light emission.</p>", "<p>In both transitional segments, the optical discharge suspends the periodic oscillation of the sub-peak at its tail. This is clearly shown in the superimposed intensity profile patterns shown in ##FIG##1##Fig. 2B## and ##FIG##2##\n3B##. The peak oscillation is recorded as a striped pattern, which is absent from these transitional segments. The stripe interval agrees with that of the periodic void train shown in ##FIG##1##Fig. 2C## and ##FIG##2##\n3C##. In fact, the video images contain no information about the splice point. However, these striped patterns can be accurately superimposed on the void photograph.</p>", "<p>Consequently, it is obvious that the sub-peak oscillation is related to the periodic void formation.</p>", "<title>Direction of bullet-shape in the damage train</title>", "<p>It is well known that the bullet-like voids formed in ordinary fiber point in the pump light propagation direction. The formation mechanism of this shape is explained by the combined effect of the internal pressure of the optical discharge and the temperature gradient along the fiber ##REF##16208896##[4]##. Once a void is pinched off from the tail of the optical discharge, the cool side surface solidifies first leaving a spherical shape and the other hot side is pressed to form a plane.</p>", "<p>However, Bufetov et al. reported that voids with a reversed direction are formed when there is a light intensity modulation along a fiber core introduced by an interference between LP and LP modes ##UREF##1##[6]##. In this study, such a void sometimes appeared near the splice point (a) as shown in ##FIG##3##Fig. 4\n##. Since the video showed the discharge propagates without any change in direction, its formation mechanism is different from that described above. In fact, the striped pattern in ##FIG##1##Fig. 2B## for this reversed void is different from that for regular periodic voids. Thus, this phenomenon also seems to be caused by the light intensity modulation of hetero-core splicing.</p>" ]

[]

[ "<p>Conceived and designed the experiments: SiT. Performed the experiments: SiT. Analyzed the data: SiT. Contributed reagents/materials/analysis tools: SiT. Wrote the paper: SiT.</p>", "<title>Background</title>", "<p>Fiber fuse is a process of optical fiber destruction under the action of laser radiation, found 20 years ago. Once initiated, opical discharge runs along the fiber core region to the light source and leaves periodic voids whose shape looks like a bullet pointing the direction of laser beam. The relation between damage pattern and propagation mode of optical discharge is still unclear even after the first in situ observation three years ago.</p>", "<title>Methodology/Principal Findings</title>", "<p>Fiber fuse propagation over hetero-core splice point (Corning SMF-28e and HI 1060) was observed in situ. Sequential photographs obtained at intervals of 2.78 µs recorded a periodic emission at the tail of an optical discharge pumped by 1070 nm and 9 W light. The signal stopped when the discharge ran over the splice point. The corresponding damage pattern left in the fiber core region included a segment free of periodicity.</p>", "<title>Conclusions</title>", "<p>The spatial modulation pattern of the light emission agreed with the void train formed over the hetero-core splice point. Some segments included a bullet-shaped void pointing in the opposite direction to the laser beam propagation although the sequential photographs did not reveal any directional change in the optical discharge propagation.</p>" ]

[ "<title>Supporting Information</title>" ]

[ "<p>I am grateful to Mr. Keisuke Aizawa, Mr. Yoshihiro Kondou and Mr. Kazuhide Hanaka (Photron Ltd.) for helping with the ultrahigh-speed videography experiment.</p>" ]

[ "<fig id=\"pone-0003276-g001\" position=\"float\"><object-id pub-id-type=\"doi\">10.1371/journal.pone.0003276.g001</object-id><label>Figure 1</label><caption><title>Experimental setup for observing fiber fuse propagation over hetero-core splice.</title></caption></fig>", "<fig id=\"pone-0003276-g002\" position=\"float\"><object-id pub-id-type=\"doi\">10.1371/journal.pone.0003276.g002</object-id><label>Figure 2</label><caption><title>Visible light emission of a fiber fuse and generated voids near the splice point (a).</title><p>A: Photographs of the visible light emission of a fiber fuse pumped by 1070 nm 9W light (original gray-scale image is converted to color-scale image), B: their intensity profiles along the dashed lines in these photographs taken every 2.78 µsec, and C: optical micrographs of the damage pattern generated at corresponding segments immersed in matching oil. The splice point is located near the solid white circle. The cladding diameter is 125 µm. The two dashed vertical lines represent the position at which the optical discharge changed its speed. See also ##SUPPL##2##Movie S1##.</p></caption></fig>", "<fig id=\"pone-0003276-g003\" position=\"float\"><object-id pub-id-type=\"doi\">10.1371/journal.pone.0003276.g003</object-id><label>Figure 3</label><caption><title>Visible light emission of a fiber fuse and generated voids near the splice point (b).</title><p>See also ##SUPPL##3##Movie S2##.</p></caption></fig>", "<fig id=\"pone-0003276-g004\" position=\"float\"><object-id pub-id-type=\"doi\">10.1371/journal.pone.0003276.g004</object-id><label>Figure 4</label><caption><title>Optical micrographs showing fiber fuse damage generated near the splice point (a).</title><p>The pump laser operates at 1070 nm and 9.0 W. The top is an expanded image of ##FIG##1##Fig. 2C##.</p></caption></fig>" ]

[ "<table-wrap id=\"pone-0003276-t001\" position=\"float\"><object-id pub-id-type=\"doi\">10.1371/journal.pone.0003276.t001</object-id><label>Table 1</label><caption><title>Specifications of the fibers used in this study.</title></caption><alternatives><table frame=\"hsides\" rules=\"groups\"><colgroup span=\"1\"><col align=\"left\" span=\"1\"/><col align=\"center\" span=\"1\"/><col align=\"center\" span=\"1\"/></colgroup><thead><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\"/><td align=\"left\" rowspan=\"1\" colspan=\"1\">HI 1060</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">SMF-28e</td></tr></thead><tbody><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">Cutoff wavelength</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">920±50 nm</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">&gt;1280 nm</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">Mode-field diameter</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">6.2±0.3 µm @ 1060 nm</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">9.2±0.4 µm @ 1310 nm</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">Core diameter</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">n. a.</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">8.2 µm</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">Cladding diameter</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">125 µm</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">125 µm</td></tr></tbody></table></alternatives></table-wrap>", "<table-wrap id=\"pone-0003276-t002\" position=\"float\"><object-id pub-id-type=\"doi\">10.1371/journal.pone.0003276.t002</object-id><label>Table 2</label><caption><title>Average speed of optical discharge, <italic>v</italic>, and average time of one void formation, τ, around the spliced points shown in ##FIG##1##Fig. 2\n## and ##FIG##2##\n3\n##.</title></caption><alternatives><table frame=\"hsides\" rules=\"groups\"><colgroup span=\"1\"><col align=\"left\" span=\"1\"/><col align=\"center\" span=\"1\"/><col align=\"center\" span=\"1\"/><col align=\"center\" span=\"1\"/><col align=\"center\" span=\"1\"/><col align=\"center\" span=\"1\"/><col align=\"center\" span=\"1\"/><col align=\"center\" span=\"1\"/></colgroup><thead><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\"/><td align=\"left\" rowspan=\"1\" colspan=\"1\">HI 1060</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">(a)</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">SMF-28e</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">...</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">SMF-28e</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">(b)</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">HI 1060</td></tr></thead><tbody><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">\n<italic>v</italic> (m/s)</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">2.08</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">1.69</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">1.42</td><td align=\"left\" rowspan=\"1\" colspan=\"1\"/><td align=\"left\" rowspan=\"1\" colspan=\"1\">1.40</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">1.51</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">1.88</td></tr><tr><td align=\"left\" rowspan=\"1\" colspan=\"1\">τ (µs)</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">6.6</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">-</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">12.1</td><td align=\"left\" rowspan=\"1\" colspan=\"1\"/><td align=\"left\" rowspan=\"1\" colspan=\"1\">16.7</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">-</td><td align=\"left\" rowspan=\"1\" colspan=\"1\">8.4</td></tr></tbody></table></alternatives></table-wrap>" ]

[]

[]

[]

[]

[]

[ "<supplementary-material content-type=\"local-data\" id=\"pone.0003276.s001\"><label>Text S1</label><caption><p>List of papers on fiber fuse.</p><p>(0.07 MB PDF)</p></caption></supplementary-material>", "<supplementary-material content-type=\"local-data\" id=\"pone.0003276.s002\"><label>Text S2</label><caption><p>Japanese translation of this paper.</p><p>(1.11 MB PDF)</p></caption></supplementary-material>", "<supplementary-material content-type=\"local-data\" id=\"pone.0003276.s003\"><label>Movie S1</label><caption><p>Movie version of ##FIG##1##Figure 2##.</p><p>(0.36 MB MPG)</p></caption></supplementary-material>", "<supplementary-material content-type=\"local-data\" id=\"pone.0003276.s004\"><label>Movie S2</label><caption><p>Movie version of ##FIG##2##Figure 3##.</p><p>(0.40 MB MPG)</p></caption></supplementary-material>" ]

[ "<table-wrap-foot><fn id=\"nt101\"><p>Source: Product information sheets at <ext-link ext-link-type=\"uri\" xlink:href=\"http://www.corning.com/opticalfiber/\">http://www.corning.com/opticalfiber/</ext-link>\n</p></fn></table-wrap-foot>", "<fn-group><fn fn-type=\"COI-statement\"><p><bold>Competing Interests: </bold>The author has declared that no competing interests exist.</p></fn><fn fn-type=\"financial-disclosure\"><p><bold>Funding: </bold>National Institute for Materials Science</p></fn></fn-group>" ]

[ "<graphic xlink:href=\"pone.0003276.g001\"/>", "<graphic id=\"pone-0003276-t001-1\" xlink:href=\"pone.0003276.t001\"/>", "<graphic xlink:href=\"pone.0003276.g002\"/>", "<graphic xlink:href=\"pone.0003276.g003\"/>", "<graphic id=\"pone-0003276-t002-2\" xlink:href=\"pone.0003276.t002\"/>", "<graphic xlink:href=\"pone.0003276.g004\"/>" ]

[ "<media xlink:href=\"pone.0003276.s001.pdf\"><caption><p>Click here for additional data file.</p></caption></media>", "<media xlink:href=\"pone.0003276.s002.pdf\"><caption><p>Click here for additional data file.</p></caption></media>", "<media xlink:href=\"pone.0003276.s003.mpg\"><caption><p>Click here for additional data file.</p></caption></media>", "<media xlink:href=\"pone.0003276.s004.mpg\"><caption><p>Click here for additional data file.</p></caption></media>" ]

{ "acronym": [], "definition": [] }

[ "<title>Introduction</title>", "<p>Trinidad and Tobago has high prevalence of diabetes mellitus that the International Diabetes Federation has projected that by the year 2025, 11.8% of the population will be diagnosed with type 2 diabetes [##UREF##0##1##]. Although the projected prevalence rate represents one of the highest prevalence rates in the North American region [##UREF##0##1##], of greater concern are the several reports of poor glycaemic control amongst type 2 diabetic patients at the primary care settings in this population [##REF##11699733##2##, ####REF##12918799##3##, ##REF##15816267##4####15816267##4##]. Poor glycaemia control, obesity, sedentary lifestyle etc are some of the factors that have been implicated in the increased risk of cardiovascular disease amongst diabetic patients in this population [##REF##11699733##2##,##UREF##1##5##]. Unfortunately, most of the published data from this population have not assessed the impact of anaemia in the risk of cardiovascular disease in type 2 diabetic patients [##REF##11699733##2##, ####REF##12918799##3##, ##REF##15816267##4##, ##UREF##1##5####1##5##]. Yet, there are pathophysiologic reasons why the presence of anaemia may lead to adverse cardiovascular consequences especially in diabetic patients. For instance, it has been demonstrated that patients with chronic anaemia had a high cardiac output and a low systemic vascular resistance [##REF##8217445##6##]. In the long term, this may result in maladaptive left ventricular hypertrophy (LVH), which is a known risk factor for adverse cardiovascular outcome and all-cause mortality [##REF##11583864##7##,##REF##12570955##8##]. Furthermore, anaemia has been shown to be a risk factor for adverse cardiovascular outcomes in non-diabetic [##REF##8712222##9##,##UREF##2##10##] and diabetic patients [##REF##16162813##11##] with chronic kidney disease. Therefore, given that diabetes is a leading cause of kidney disease and kidney failure, assessment of anaemia and kidney dysfunction in diabetic patients should be a regular laboratory tests. This is particularly important given that anaemia presents earlier and often more severe in diabetic patients with chronic kidney disease compared with chronic kidney disease patients without diabetes [##REF##15452405##12##,##REF##15103543##13##]. Although there is scanty reports on the prevalence of kidney disease in type 2 diabetic patients in this population, report of end-stage renal disease that presented with chronic anaemia has been documented [##REF##9924564##14##]. Thus, the presence of anemia in diabetic patients with undetected renal dysfunction may be particularly dangerous especially at the primary care setting where routine laboratory investigations are infrequent and haematological tests not usually included as part of the laboratory tests for the patients' management. In this regard, the present study is aimed to assess the prevalence of anaemia and kidney dysfunction in a cross-section of Afro-Caribbean type 2 diabetic patients that were previously shown to have a high prevalence of the metabolic syndrome [##REF##17852050##15##].</p>" ]

[ "<title>Subjects and methods</title>", "<title>Type 2 diabetic patients</title>", "<p>One hundred and fifty-five (46 males, 109 females) type 2 diabetic patients visiting, consecutively, eight lifestyle disease clinics at primary care setting in Tobago (between June and November 2006) participated in the study. Patients were considered as type 2 diabetic if they had been managed on oral hypoglycaemic medication and/or diet/exercise since diagnosis (except on occasions when patients took insulin to control hyperglycaemia). The protocol was as previously published [##REF##17852050##15##]. Briefly, the patients received information leaflets and posters explaining the objectives and protocol of the research study and a member of the research group addressed the patients during the clinics to explain the rationale for the study. Subsequently, the Research Assistant or Clinic Nurse took names and other information of patients who volunteered to participate in the study. The patients were subsequently reminded of the importance of overnight fasting a day preceding the study.</p>", "<title>Non-diabetic subjects</title>", "<p>Fifty-one (22 males, 29 females) apparently healthy non-diabetic subjects living in the same environment/city (Tobago) with the diabetic patients qualified to participate in the study as non-diabetic control subjects. Each volunteer underwent compulsory oral glucose tolerance tests (OGTT, 75 g anhydrous glucose in 250–300 ml water) to detect undiagnosed diabetes (fasting plasma glucose &gt; 7.0 with 2-hour plasma glucose &gt; 11.1 mmol/L) and impaired glucose tolerance (2-hour plasma glucose &gt; 7.8 mmol/L) [##REF##10466661##16##]. Subjects with plasma glucose values diagnostic of diabetes were excluded from the study and data analysis while subjects with plasma glucose values diagnostic of IGT were included in the study and data analysis.</p>", "<title>Study protocol</title>", "<p>Informed consent was obtained from both the diabetic and non-diabetic subjects that participated in the study. The study protocol conforms to the ethical guidelines of the 1975 Declaration of Helsinki as reflected in a <italic>priori </italic>approval by our institutional Ethics Review Committee. All subjects were studied at the clinics in the morning after an overnight (12–14 hours) fast. Details of self-reported ethnic origin and age were directly ascertained from the patients and recorded. Then, waist (cm), at the level of the umbilicus with the patient standing and breathing normally, and hip circumferences (cm), at the level of the largest projection of the buttocks, were obtained by tape measure while weight (kg), with standard hospital balance, and height (m), with metal rule, were measured (in light clothing, without shoes). Then, fasting blood sample was collected from each subject and, for the non-diabetic subjects alone, 75 g anhydrous glucose in 250–300 ml water was given to each subject and blood samples collected every 30 minutes for 120 minutes. Blood samples were preserved in appropriate tubes for complete blood count, creatinine, glucose and glycated haemoglobin measurements.</p>", "<title>Laboratory analysis</title>", "<p>Plasma glucose and serum creatinine were measured in multi-channel auto-analysers using dry slide kits (Johnson &amp; Johnson Vitros 250, Ortho-Clinical Diagnostics Inc., Rochester NY, USA). Glycated haemoglobin (HbA_1c) was determined using boronate affinity assay (Axis-Shield PoC AS, N-0504 Oslo, Norway) while complete blood count was measured using the routine Coulter counter machine (Sysmex XE 2100, Germany).</p>", "<title>Definitions, calculations and statistics</title>", "<p>The results are expressed as mean ± SE. The Statistical Package for the Social Sciences (SPSS Inc., Chicago, USA) software was used in all analyses. Anemia was defined as haemoglobin &lt; 12 g/dl in females and &lt; 13 g/dl in males [##UREF##3##17##]. Kidney function was assessed using glomerular filtration rate (GFR) as estimated by the four-variable Modification of Diet in Renal Disease (MDRD) study equation as follows: GFR = 186.3 × (serum creatinine^-1.154) × (age^-0.203) × 1.212 (if black) × 0.742 (if female). GFR was expressed in ml/min/1.73 m²[##REF##10075613##18##] and patients were considered to have chronic kidney disease when the estimated GFR was &lt; 60 ml/min per 1.73 m²[##REF##16162813##11##]. Comparisons within- and between-gender, between diabetic and non-diabetic subjects were performed using Student's t-test while chi-square test was employed for categorical variables. A p-value &lt; 0.05 was considered statistically significant on two-tailed testing for all analysis.</p>" ]

[ "<title>Results</title>", "<p>Table ##TAB##0##1## shows the clinical characteristics of all the subjects studied. The diabetic patients were significantly older than the non-diabetic subjects (p &lt; 0.001). Compared with non-diabetic subjects, lower percentage of the diabetic patients smoke cigarettes and drink alcoholic beverages (Table ##TAB##0##1##, p &lt; 0.05). The male non-diabetic subjects had significantly higher red blood cell count (RBC), hemoglobin (Hb) and hematocrit levels than their diabetic counterparts (Table ##TAB##1##2##, p &lt; 0.001). However, the female diabetic and non-diabetic subjects had similar Hb concentrations (12.1 ± 0.1 vs. 12.5 ± 0.2 g/dl, p &gt; 0.05). While male non-diabetic subjects had significantly higher RBC, Hb and hematocrit than non-diabetic female subjects (p &lt; 0.001), the RBC and hematocrit concentrations were similar in male and female diabetic patients (Table ##TAB##1##2##, p &gt; 0.05). The prevalence of anemia and chronic kidney disease in the subjects are shown on Table ##TAB##2##3##. Irrespective of gender, diabetic patients had significantly higher prevalence rate of anemia than non-diabetic subjects (Table ##TAB##3##4##, p &lt; 0.05). Again, diabetic patients had significantly higher prevalence of kidney dysfunction compared with the non-diabetic subjects (Table ##TAB##3##4##, p &lt; 0.05). The diabetic patients with anemia had significantly higher serum creatinine levels (1.4 ± 0.1 vs. 1.0 ± 0.03 mg/dl, p &lt; 0.001) and lower GFR (67.1 ± 3.0 vs. 87.9 ± 5.4 ml/min per 1.73 m², p &lt; 0.001) than diabetic patients without anemia. Similarly, diabetic patients with anaemia had significantly higher levels of glycated hemoglobin (index of blood glucose control), creatinine and higher prevalence of kidney dysfunction than non-diabetic subjects with anaemia (Table ##TAB##2##3##, p &lt; 0.05).</p>" ]

[ "<title>Discussion</title>", "<p>This study has shown that anaemia was more prevalent in diabetic than non-diabetic subjects irrespective of gender, and diabetic patients with anaemia had the lowest kidney function compared with patients without anaemia or non-diabetic subjects with anaemia. The implications of these findings in relation to diabetes management are further discussed.</p>", "<p>All the new frontiers in the management of type 2 diabetes are aimed at achieving optimal blood glucose control so as to prevent macro- and micro-vascular complications [##REF##10938048##19##, ####REF##15487977##20##, ##REF##16915799##21##, ##REF##17496356##22####17496356##22##]. In the primary care setting in the clinics studied, assessment of haematological parameters is no part of routine diabetic evaluation. Therefore, the finding in this study of high prevalence rates of anaemia amongst the diabetic patients studied, irrespective of gender, is perhaps, the first report in the Caribbean diabetic population. This finding is significant given that anaemia has previously been shown to be associated with cardiovascular disease and all-cause mortality in diabetic patients [##REF##16162813##11##]. However, other workers have reported an association between type 2 diabetes and elevated haematocrit, which is thought to negatively affect nitric oxide availability resulting in cardiovascular disease [##REF##15691863##23##]. Furthermore, increase of whole blood viscosity due to increased levels of haematological parameters such as fibrinogen has been reported in type 2 diabetic patients with arteriosclerosis obliterans [##REF##1485469##24##]. We are not aware of any study at the primary care settings in the Caribbean population that evaluated the prevalence of anaemia, high haematocrit or increased whole blood viscosity in individuals with type 2 diabetes. In contrast to the elevated haematocrit report [##REF##15691863##23##], our study found high prevalence rates of anaemia in diabetic patients, irrespective of gender. This finding may be indicative of early renal disease given that previous studies showed earlier appearance of anaemia in diabetic patients; with an inverse relationship between creatinine concentration and haemoglobin levels [##REF##15452405##12##,##REF##15103543##13##]. However, it should be noted that macro-vascular, and not micro-vascular, complication is commoner in type 2 diabetes [##REF##10938048##19##] and as such there appears to be little or no documented report of kidney complications in type 2 diabetes in our population. Although there was a report of two Afro-Caribbean teenagers with end-stage renal disease who presented in clinics with anaemia [##REF##9924564##14##], incidence or prevalence of diabetic patients with anaemia are scanty.</p>", "<p>In this study, anaemia was more prevalent amongst the diabetic patients than non-diabetic subjects and the former group of subjects were previously shown to have a high prevalence of the metabolic syndrome [##REF##17852050##15##]. The limitation is that the study was not designed to determine the type of anaemia the patients had, though iron-deficiency anaemia is common in some other developing countries [##REF##11379460##25##,##REF##17952232##26##]. Thus, diabetic patients and the senior citizens, with limited food choices, would be more vulnerable to iron-deficiency anaemia and all cause mortality [##REF##16162813##11##]. The findings in the present study have implications for diabetes management in that they appear to indicate a need for routine full blood counts in primary care diabetes management. Intervention with erythropoietin has been shown to improve the quality of life for anaemic patients in chronic renal failure [##REF##11379460##25##]. For diabetic patients, it is well documented that reducing blood glucose levels and targeting acceptable glycated haemoglobin (HbA1c) levels have been the focus so as to prevent the risk of micro- and macro-vascular complications [##REF##10938048##19##,##REF##8366922##27##,##REF##9742976##28##]. Even then, the reality is that most patients with type 2 diabetes in developed and developing countries have glycated haemoglobin levels above the recommended target levels and are prone to macro-vascular complications [##REF##11699733##2##,##REF##12918799##3##]. Inadequate or absence of laboratory facilities at the primary care settings in many developing countries is a major limitation in routine laboratory assessment of diabetic patients. We believe that early detection and management of anaemia in diabetic patients at the primary care setting would be cost effective in so far as it would reduce hospital admissions and maintain optimum health. This could be achieved through the provision of adequate laboratory facilities and expansion of the scope of laboratory investigations used in the management of diabetic patients. Therefore, it is suggested that all diagnostic laboratories in developing countries and elsewhere should include complete blood count as one of the routine laboratory tests required in the management of diabetic patients.</p>" ]

[]

[ "<p>This is an Open Access article distributed under the terms of the Creative Commons Attribution License (<ext-link ext-link-type=\"uri\" xlink:href=\"http://creativecommons.org/licenses/by/2.0\"/>), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.</p>", "<title>Background</title>", "<p>Anaemia has been shown in previous studies to be a risk factor for cardiovascular disease in diabetic patients with chronic kidney disorder. This study was aimed to assess the prevalence of anaemia and kidney dysfunction in Caribbean type 2 diabetic patients that have been previously shown to have a high prevalence of the metabolic syndrome.</p>", "<title>Methods</title>", "<p>155 type 2 diabetic patients and 51 non-diabetic subjects of African origin were studied. Anthropometric parameters were measured and fasting blood samples were collected for glucose, creatinine, glycated hemoglobin and complete blood count. Anaemia was defined as haemoglobin &lt; 12 g/dl (F) or &lt; 13 g/dl (M). Kidney function was assessed using glomerular filtration rate (GFR) as estimated by the four-variable Modification of Diet in Renal Disease (MDRD) study equation. Subjects were considered to have chronic kidney disease when the estimated GFR was &lt; 60 ml/min per 1.73 m². Comparisons for within- and between-gender, between diabetic and non-diabetic subjects were performed using Student's t-test while chi-square test was employed for categorical variables.</p>", "<title>Results</title>", "<p>The diabetic patients were older than the non-diabetic subjects. While male non-diabetic subjects had significantly higher red blood cell count (RBC), haemoglobin and hematocrit concentrations than non-diabetic female subjects (p &lt; 0.001), the RBC and hematocrit concentrations were similar in male and female diabetic patients. Furthermore, irrespective of gender, diabetic patients had significantly higher prevalence rate of anemia than non-diabetic subjects (p &lt; 0.05). Anaemic diabetes patients had significantly lower GFR (67.1 ± 3.0 vs. 87.9 ± 5.4 ml/min per 1.73 m², p &lt; 0.001) than non-anaemic patients.</p>", "<title>Conclusion</title>", "<p>A high prevalence of anaemia was identified in this group of type 2 diabetic patients previously shown to have a high prevalence of the metabolic syndrome. It is therefore recommended that diagnostic laboratories in developing countries and elsewhere should include complete blood count in routine laboratory investigations in the management of diabetic patients.</p>" ]

[ "<title>Competing interests</title>", "<p>The authors declare that they have no competing interests.</p>", "<title>Authors' contributions</title>", "<p>CEE – involved in conception, design, acquisition of data, data analysis, interpretation and drafting of manuscript. AJ–L – involved in conception, interpretation and drafting of manuscript. EN – acquisition of data, interpretation and drafting of manuscript. DS – acquisition of data, data analysis, interpretation and drafting of manuscript. FO – acquisition of data, interpretation and drafting of manuscript.</p>" ]

[ "<title>Acknowledgements</title>", "<p>This study was supported by a Research Grant from the University of the West Indies, St Augustine Campus. The Nursing staff of Lifestyle Disease Clinics in Tobago assisted professionally while the Tobago Regional Health Authority granted permission for study.</p>" ]

[]

[ "<table-wrap position=\"float\" id=\"T1\"><label>Table 1</label><caption><p>Clinical characteristics of the diabetic and non-diabetic subjects studied</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td align=\"left\"><bold>Variables</bold></td><td align=\"center\"><bold>Type 2 diabetic patients</bold></td><td align=\"center\"><bold>Non-diabetic subjects</bold></td></tr></thead><tbody><tr><td align=\"left\">Gender (m/f ratio)</td><td align=\"center\">46/109</td><td align=\"center\">22/29</td></tr><tr><td align=\"left\">Age (yr)</td><td align=\"center\">65.9 ± 0.9</td><td align=\"center\">48.8 ± 1.2**</td></tr><tr><td align=\"left\">Duration of diabetes (yr)</td><td align=\"center\">10.6 ± 0.7</td><td align=\"center\">-</td></tr><tr><td align=\"left\">Weight (kg)</td><td align=\"center\">77.9 ± 1.2</td><td align=\"center\">83.3 ± 2.0*</td></tr><tr><td align=\"left\">Height (m)</td><td align=\"center\">1.64 ± 0.01</td><td align=\"center\">1.68 ± 0.01*</td></tr><tr><td align=\"left\">Body mass index (kg/m²)</td><td align=\"center\">28.6 ± 0.4</td><td align=\"center\">30.2 ± 0.8</td></tr><tr><td align=\"left\">Waist circumference (cm)</td><td align=\"center\">99.1 ± 1.0</td><td align=\"center\">95.5 ± 1.4</td></tr><tr><td align=\"left\">Hip circumference (cm)</td><td align=\"center\">105.0 ± 0.8</td><td align=\"center\">108.8 ± 1.4*</td></tr><tr><td align=\"left\">Unemployed (%)</td><td align=\"center\">112 (72.3)</td><td align=\"center\">3 (5.9)**</td></tr><tr><td align=\"left\">Cigarette smokers (%)</td><td align=\"center\">6 (3.9)</td><td align=\"center\">6 (11.8)*</td></tr><tr><td align=\"left\">Alcohol users (%)</td><td align=\"center\">29 (18.7)</td><td align=\"center\">23 (45.1)**</td></tr></tbody></table></table-wrap>", "<table-wrap position=\"float\" id=\"T2\"><label>Table 2</label><caption><p>Between-gender comparison of the haematological tests of the diabetic and non-diabetic subjects</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td align=\"left\">Haematology profile</td><td align=\"center\" colspan=\"2\"><bold>Type 2 diabetic patients</bold></td><td align=\"center\" colspan=\"2\"><bold>Non-diabetic subjects</bold></td></tr><tr><td/><td align=\"center\"><bold>Males</bold></td><td align=\"center\"><bold>Females</bold></td><td align=\"center\"><bold>Males</bold></td><td align=\"center\"><bold>Females</bold></td></tr></thead><tbody><tr><td align=\"left\">Age (years)</td><td align=\"center\">^§68.9 ± 1.2</td><td align=\"center\">64.6 ± 1.2^¶</td><td align=\"center\">48.9 ± 1.4</td><td align=\"center\">48.8 ± 1.8^╫╫</td></tr><tr><td align=\"left\">Whole Blood Count (× 10³/uL)</td><td align=\"center\">5.5 ± 0.3</td><td align=\"center\">6.0 ± 0.1</td><td align=\"center\">5.5 ± 0.2</td><td align=\"center\">5.5 ± 0.3</td></tr><tr><td align=\"left\">Red Blood Cell (× 10⁶/uL)</td><td align=\"center\">^§§4.6 ± 0.1</td><td align=\"center\">4.4 ± 0.1</td><td align=\"center\">5.1 ± 0.1**</td><td align=\"center\">4.6 ± 0.1^╫</td></tr><tr><td align=\"left\">Haemoglobin (g/dl)</td><td align=\"center\">^§§12.9 ± 0.2</td><td align=\"center\">12.1 ± 0.1^¶</td><td align=\"center\">14.6 ± 0.3**</td><td align=\"center\">12.5 ± 0.2</td></tr><tr><td align=\"left\">Hematocrit (%)</td><td align=\"center\">^§39.4 ± 0.5</td><td align=\"center\">37.5 ± 0.3</td><td align=\"center\">44.2 ± 0.7**</td><td align=\"center\">39.0 ± 0.5^╫</td></tr><tr><td align=\"left\">Mean Corpuscular Volume (FL)</td><td align=\"center\">87.0 ± 0.7</td><td align=\"center\">85.3 ± 0.6^¶</td><td align=\"center\">87.4 ± 1.4</td><td align=\"center\">84.4 ± 1.3</td></tr><tr><td align=\"left\">Mean Corpuscular Haemoglobin (pg)</td><td align=\"center\">28.5 ± 0.3</td><td align=\"center\">28.2 ± 0.6</td><td align=\"center\">28.8 ± 0.5*</td><td align=\"center\">27.1 ± 0.5*</td></tr><tr><td align=\"left\">Platelet Count (× 10³/uL)</td><td align=\"center\">224.2 ± 9.9</td><td align=\"center\">244.4 ± 5.7</td><td align=\"center\">237.9 ± 13.8</td><td align=\"center\">262.2 ± 10.1</td></tr><tr><td align=\"left\">Mean Platelet Volume (Fl)</td><td align=\"center\">11.1 ± 0.1</td><td align=\"center\">11.1 ± 0.1</td><td align=\"center\">11.2 ± 0.2</td><td align=\"center\">10.9 ± 0.2</td></tr></tbody></table></table-wrap>", "<table-wrap position=\"float\" id=\"T3\"><label>Table 3</label><caption><p>Glycaemic control and levels of kidney function in patients with and without anemia</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td/><td align=\"center\" colspan=\"2\"><bold>Type 2 diabetic patients</bold></td><td align=\"center\" colspan=\"2\"><bold>Non-diabetic subjects</bold></td></tr><tr><td/><td align=\"center\"><italic>Anemia</italic><break/><italic>N = 72</italic></td><td align=\"center\"><italic>No Anemia</italic><break/><italic>N = 83</italic></td><td align=\"center\"><italic>Anemia</italic><break/><italic>N = 8</italic></td><td align=\"center\"><italic>No Anemia</italic><break/><italic>N = 43</italic></td></tr></thead><tbody><tr><td align=\"left\">Glycated hemoglobin (%)</td><td align=\"center\">7.7 ± 0.3</td><td align=\"center\">7.6 ± 0.2</td><td align=\"center\">5.7 ± 0.2^¶</td><td align=\"center\">5.9 ± 0.2</td></tr><tr><td align=\"left\">Fasting plasma glucose (mmol/L)</td><td align=\"center\">9.3 ± 0.6</td><td align=\"center\">9.1 ± 0.4</td><td align=\"center\">4.6 ± 0.1^¶</td><td align=\"center\">5.0 ± 0.2</td></tr><tr><td align=\"left\">Creatinine (mg/dl)</td><td align=\"center\">1.4 ± 0.1</td><td align=\"center\">1.0 ± 0.03**</td><td align=\"center\">1.0 ± 0.1^¶</td><td align=\"center\">1.1 ± 0.02</td></tr><tr><td align=\"left\">Glomerular filtration rate (ml/min/1.73 m²)<break/>(index of kidney function)</td><td align=\"center\">67.1 ± 3.0</td><td align=\"center\">87.9 ± 5.4**</td><td align=\"center\">86.2 ± 5.1^¶</td><td align=\"center\">82.6 ± 2.1</td></tr><tr><td align=\"left\">Chronic kidney disease (%)<break/>(estimated GFR &lt; 60 ml/min per 1.73 m²)</td><td align=\"center\">27 (38.6)</td><td align=\"center\">9 (10.8)**</td><td align=\"center\">1 (12.5)</td><td align=\"center\">1 (2.4)</td></tr></tbody></table></table-wrap>", "<table-wrap position=\"float\" id=\"T4\"><label>Table 4</label><caption><p>Prevalence of Anemia and chronic kidney disease (estimated by glomerular filtration rate) in diabetic and non-diabetic subjects</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td/><td align=\"center\" colspan=\"2\"><bold>Type 2 diabetic patients</bold></td><td align=\"center\" colspan=\"2\"><bold>Non-diabetic subjects</bold></td></tr><tr><td/><td align=\"center\"><italic>Males</italic></td><td align=\"center\"><italic>Females</italic></td><td align=\"center\"><italic>Males</italic></td><td align=\"center\"><italic>Females</italic></td></tr></thead><tbody><tr><td align=\"left\">Anemia (%)<break/>(Hb &lt; 12 g/dL [F] or &lt; 13 g/dl [M])</td><td align=\"center\">22 (47.8)</td><td align=\"center\">50 (45.9)</td><td align=\"center\">2 (9.1)*</td><td align=\"center\">6 (20.7)^¶</td></tr><tr><td align=\"left\">Creatinine (mg/dl)</td><td align=\"center\">1.6 ± 0.2</td><td align=\"center\">1.0 ± 0.03</td><td align=\"center\">1.2 ± 0.04</td><td align=\"center\">0.96 ± 0.03</td></tr><tr><td align=\"left\">Glomerular filtration rate (ml/min/1.73 m²)<break/>(index of kidney function)</td><td align=\"center\">71.9 ± 2.7</td><td align=\"center\">81.0 ± 4.5</td><td align=\"center\">83.9 ± 3.1*</td><td align=\"center\">82.6 ± 2.5</td></tr><tr><td align=\"left\">Chronic kidney disease (%)<break/>(estimated GFR &lt; 60 ml/min per 1.73 m²)</td><td align=\"center\">12 (26.7)</td><td align=\"center\">24 (22.2)</td><td align=\"center\">1 (4.5)**</td><td align=\"center\">1 (3.6)^¶</td></tr></tbody></table></table-wrap>" ]

[]

[]

[]

[]

[]

[]

[ "<table-wrap-foot><p>*P &lt; 0.05, **P &lt; 0.001 for comparisons between diabetic and non-diabetic subjects</p></table-wrap-foot>", "<table-wrap-foot><p>*P &lt; 0.05, **P &lt; 0.001 for comparisons between male and female non-diabetic subjects</p><p>^¶P &lt; 0.05 for comparisons between male and female diabetic patients</p><p>^§P &lt; 0.05, ^§§P &lt; 0.001 for comparisons between male diabetic and male non-diabetic subjects</p><p>^╫P &lt; 0.05, ^╫╫P &lt; 0.001 for comparisons between female diabetic and female non-diabetic subjects</p></table-wrap-foot>", "<table-wrap-foot><p>**P &lt; 0.001 for comparisons between diabetic patients with and without anaemia</p><p>^¶P &lt; 0.05 for comparisons between diabetic patients with anaemia and non-diabetic subjects with anaemia</p></table-wrap-foot>", "<table-wrap-foot><p>*P &lt; 0.05, **P &lt; 0.001 for comparisons between male diabetic and non-diabetic subjects</p><p>^¶P &lt; 0.05 for comparisons between female diabetic and non-diabetic subjects.</p></table-wrap-foot>" ]

[]

[]

{ "acronym": [], "definition": [] }

[ "<title>Background</title>", "<p>The \"Understanding Aging: Biomedical and Bioengineering Approaches\" conference was held from June 27–29, 2008 at UCLA organized by Aubrey de Grey to discuss and talk about possible intervention in ageing.</p>", "<p>During the past century, mankind has gained more years of average life expectancy than in the last 10,000 years. We are now living in a rapidly ageing world. The sharp rise in life expectancy, coupled to a steady decline in birth rates in all developed countries, has led to an unprecedented demographic revolution characterized by an explosive growth in the numbers and proportion of older persons. Nowadays, people are living much longer than they used to and the longer they live, the longer their bodies are exposed to environmental factors which increase the risk of age-associated diseases. The reduction of the response to environmental stimuli is associated with an increased predisposition to illness and death. This progression causes a reduction of the response to environmental stimuli and, in general, is associated with an increased predisposition to illness and death. In Western countries, the mortality rate increases in people over 65 years, if compared with individuals between 25 and 44 years old, by 100-fold for stroke, as well as chronic lung disease, 92-fold for heart disease, 89-fold for pneumonia and influenza, 43-fold for cancer [##REF##12516005##1##]. On the contrary, ageing in good condition seems directly correlated with a good functioning of the immune system, suggesting that there are genetic determinants of longevity in genes regulating the immune inflammatory response [##REF##16608411##2##,##REF##17703905##3##].</p>", "<p>In senescence alterations of innate and instructive immunity have been described. The modifications of the immune system in the elderly are generally evaluated as a deterioration of the immune system, this is the origin of the term immunosenescence. A good immune system in the elderly is tightly correlated to health status, and some immunological parameters are often notably reduced in the elderly. On the other hand infectious diseases, tumors, autoimmune phenomenona and inflammatory chronic diseases like atherosclerosis and Alzheimer's disease, are frequent in this phase of the life course [##REF##17703905##3##, ####REF##15679921##4##, ##REF##18442326##5####18442326##5##].</p>", "<p>A body of experimental and clinical evidence has suggested that the immune system is implicated, with a variable degree of importance, in almost all age related or associated diseases. Both innate and instructive immune systems are usually involved in the pathogenesis of these chronic diseases. However, innate immunity appears to be the prevalent mechanism driving tissue damages associated with different age-related diseases [##REF##10911963##6##]. So, ageing is accompanied by an age-dependent up-regulation of the inflammatory response, due to the chronic antigenic stress that impinges throughout life upon innate immunity, and has potential implications for the onset of inflammatory diseases [##REF##17118425##7##].</p>", "<p>Here is an extract of the talks and posters presented:</p>", "<title>Mitochondrial damage</title>", "<p>Perturbation of mitochondrial Fe homeostasis cause a decline in mitochondrial function that causes neuromuscular degenerative disease and other tissue dysfunction. C. Leeuwenburgh suggested that mitochondrial non-heme Fe represents a potential novel target for targeted interventions to slow ageing (C. Leeuwenburgh, University of Florida, USA) [##REF##18395385##8##].</p>", "<title>Micronutrient inadequacy</title>", "<p>It was proposed that inadequate micronutrient intake leads to metabolic modification that has long term consequences such as cancer (DNA damage), severe infection (immune dysfunction) and cognitive dysfunction and accelerated ageing (mitochondrial decay). Much evidence supports the idea that micronutrient shortage accelerate ageing (B.N. Ames, University of California, Berkeley, CA, USA) [##REF##18056830##9##].</p>", "<title>Telomeres</title>", "<p>The shortening of telomeres is supposed to be the molecular clock of ageing; indeed there is a strong correlation between age and telomere length and shorter telomeres directly correspond to shorter human life expectancy. For this purpose, several biotech organizations have accepted the challenge of finding ways to prevent telomere shortening by transiently inducing the activity of telomerase (L.A. Briggs, Reno, NV, USA). Another research has shown that therapy acting on the catalytic component of human telomerase, such as TAT2, a small molecule telomerase activator, could stabilize the telomere length and retard the loss of the immune control over viral infection (R. Effros, UCLA, Los Angeles, CA, USA) [##REF##18426956##10##].</p>", "<title>Immunological Point of view</title>", "<p>On the immunological side Dr Z. Cui has shown that cancer cells in vitro could be killed by the effector cells of the innate immune system such as macrophages and neutrophils. A similar activity was discovered in some healthy people concerning granulocytes and monocytes (Z. Cui, Winston-Salem, NC, USA) [##REF##16682640##11##].</p>", "<p>According to the fact that the immune system plays an important role in ageing, the group of Prof. C. Caruso, actively involved in immunosenescence studies [##REF##16608411##2##, ####REF##17703905##3##, ##REF##15679921##4##, ##REF##18442326##5####18442326##5##], has demonstrated that B naïve lymphocytes, are increased in the offspring of healthy old centenarians. It has been demonstrated that the children of centenarians, who are in their 70s and 80s, have a survival advantage when compared with control subject of the same age range whose parents died at an average life expectancy [##REF##17703905##3##]. The main lymphocyte differences observed between the two groups concern B cells. Indeed naïve B cells are more abundant as well as double negative B cells in centenarian children. These data are similar to that found in previously experiment on young subjects. So, B cell compartment of the offspring of centenarians seems to be more similar to that of young respect to the old one (S. Vasto, University of Palermo, Italy) [##REF##18442327##12##].</p>", "<p>It is well known that change in immune function are hallmarks of ageing and the group of Dr. A. Agrawal (University of California, Irvine, CA, USA) has shown that the reactivity of dendritic cells to self-antigens can be characteristic of ageing features. Furthermore, this over-reactivity induces lymphocyte T proliferation with subsequent higher risk of autoimmune diseases [##REF##17828583##13##].</p>", "<p>Interestingly, Effros's group suggests a possible involvement of hyper-activated T cells in bone loss associated with vascular disease in aged mice. The increased proportion of CD8 T cells lacking expression of the co-stimulatory receptor CD28 leads to decreased vaccine responsiveness and early mortality [##REF##17014937##14##]. ST Parish found that loss of CD28 expression is caused by increased Caspase-3 activity that can be induced by Tumor necrosis factor-alpha and suggested possible strategies for retarding the generation of senescent CD8 T cells during ageing (L.S. Graham, UCLA, Los Angeles, CA, USA).</p>" ]

[]

[]

[]

[ "<title>Conclusion</title>", "<p>Ageing is a complex process that negatively impacts the development of the immune system and its ability to function. Progressive changes in the T and B cell systems over the life span have a major impact on the capacity to respond to immune challenge. These cumulative age-associated changes in immune competence are termed immunosenescence. A better understanding of immunosenescence and the development of new strategies to counteract it are essential for improving the quality of life of the elderly population [##REF##18442326##5##]</p>" ]

[ "<p>This is an Open Access article distributed under the terms of the Creative Commons Attribution License (<ext-link ext-link-type=\"uri\" xlink:href=\"http://creativecommons.org/licenses/by/2.0\"/>), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.</p>", "<p>During the past century, humans have gained more years of average life expectancy than in the last 10,000 years; we are now living in a rapidly ageing world. The sharp rise in life expectancy, coupled to a steady decline in birth rates in all developed countries, has led to an unprecedented demographic revolution characterized by an explosive growth in the number and proportion of older people. Ageing is a complex process that negatively impacts the development of the immune system and its ability to function. Progressive changes in the T and B cell systems over the life span have a major impact on the capacity to respond to immune challenge. These cumulative age-associated changes in immune competence are termed Immunosenescence: some immunological parameters are commonly notably reduced in the elderly and, reciprocally good function is tightly correlated to health status. Hence, a better understanding of Immunosenescence and the development of new strategies to counteract it are essential for improving the quality of life of the elderly population.</p>" ]

[ "<title>Competing interests</title>", "<p>The authors declare that they have no competing interests.</p>", "<title>Authors' contributions</title>", "<p>All authors contributed equally to the paper and read and approved the final manuscript</p>" ]

[ "<title>Acknowledgements</title>", "<p>The authors thank Dr. A. De Grey (Methuselah Foundation, UK) and Prof. C. Caruso (University of Palermo, I) for revising critically the manuscript. Original work of Palermo's group was supported by Italian Ministry of University and Research PRIN 2006 Project to C. Caruso.</p>" ]

[]

[]

[]

[]

[]

[]

[]

[]

[]

[]

[]

{ "acronym": [], "definition": [] }

[ "<title>Background</title>", "<p>Matsumoto et al. [##REF##1915324##1##] and Moritani et al. [##REF##8514689##2##] have proposed an incremental cycle ergometer test utilizing fatigue curves to identify the maximal power output at which an individual can maintain without evidence of fatigue, described as the electromyographic fatigue threshold (EMG_FT). The EMG_FTtest is an adaptation to deVries' [##REF##7173165##3##] original monopolar physical working capacity at the fatigue threshold (PWC_FT) test, using a bipolar supramaximal protocol. The EMG_FTinvolves determining the rate of rise in electrical activity from the vastus lateralis during four, two-minute work bouts on a cycle ergometer, with varying power outputs. It has been suggested that the rise in electrical activity is a result of progressive recruitment of additional motor units (MU) and/or an increase in the firing frequency of MUs that have already been recruited. Several investigations have used surface electromyography to characterize the fatigue-induced increase in EMG amplitude, as well as to identify the power output associated with the onset of neuromuscular fatigue during cycle ergometry [##REF##1915324##1##,##REF##8514689##2##,##REF##10378070##4##, ####REF##9754971##5##, ##REF##12914560##6##, ##REF##17149988##7##, ##REF##7328909##8####7328909##8##]. Matsumoto et al. [##REF##1915324##1##] described the EMG_FTas the highest intensity sustainable on a cycle ergometer without signs of neuromuscular fatigue. In addition, Moritani et al. [##REF##8514689##2##] suggested a strong physiological link between myoelectrical changes at fatigue and anaerobic threshold. Furthermore, the EMG_FTmethod has been reported as a valid and reliable technique for examining the transition from aerobic to anaerobic metabolism during exercise [##REF##10378070##4##,##REF##12914560##6##,##REF##17149988##7##]. Identifying a reliable, non-invasive way to measure and predict the onset of fatigue has potential use in clinical populations, as well as serving as a training tool for those with minimal testing equipment. Therefore, the purpose of this study was to examine the metabolic relationship between VO_2PEAK, ventilatory threshold (VT), and the EMG_FT, as well as to compare the power output at VO_2PEAK, VT, and EMG_FT.</p>" ]

[ "<title>Methods</title>", "<title>Participants</title>", "<p>Thirty-eight recreationally trained (1–5 hours/week), college-aged men (Table ##TAB##0##1##) volunteered to participate in this study. All procedures were approved by the University of Oklahoma Institutional Review Board for Human Subjects, and written informed consent was obtained from each participant prior to any testing.</p>", "<title>Determination of VO_2PEAKand Ventilatory Threshold</title>", "<p>Participants performed a continuous graded exercise test (GXT) on an electronically-braked cycle ergometer (Corival Lode 400, Groningen, The Netherlands) to determine maximal oxygen consumption (VO_2PEAK) and ventilatory threshold (VT). Following a five-minute warm-up (50 W), the workload was increased 25 W every two minutes until the participants were unable to maintain 70 rpm, or until volitional fatigue.</p>", "<p>Ventilatory threshold was determined as a plot of ventilation (V_E) vs. oxygen consumption (VO₂), as described previously [##REF##7096157##9##]. Two linear regression lines were fit to the lower and upper portions of the V_Evs. VO₂curve before and after the break points, respectively. The intersection of these two lines was defined as VT.</p>", "<title>Gas Exchange Analysis</title>", "<p>Open circuit spirometry was used to analyze the gas exchange data using the Parvo-Medics TrueOne 2400^®Metabolic Measurement System (Sandy, Utah, United States). Oxygen and carbon dioxide were analyzed through a sampling line after the gases passed through a heated pneumotach and mixing chamber. The data were averaged over 15-second intervals. The highest average VO₂value during the GXT was recorded as the VO_2PEAKif it coincided with at least two of the following criteria: (a) a plateau in heart rate (HR) or HR values within 10% of the age-predicted HRmax, (b) a plateau in VO₂(defined by an increase of no more than 150 ml·min^-1), and/or (c) an RER value greater than 1.15 [##REF##12857763##10##].</p>", "<title>Electromyography</title>", "<p>Pre-gelled bipolar (2.54 cm center-to-center) surface electrodes (Ag-Ag Cl, Quinton Quick Prep, Quinton Instruments Co., Bothell, WA) were placed over the lateral portion of the vastus lateralis muscle, midway between the greater trochanter and the lateral condyle of the femur. A reference electrode was placed over the 7^thcervical vertebrae. The raw EMG signals were pre-amplified ((gain × 1,000) EMG 100C, Biopac Systems, Inc., Santa Barbara, CA), sampled at 1,000 Hz and bandpass filtered from 10–500 Hz (zero-lag 8^thorder Butterworth filter). All EMG amplitude values were stored on a personal computer (Dell Inspiron 8200, Dell, Inc., Round Rock, TX) and analyzed off-line using custom-written software (LabVIEW v 7.1, National Instruments, Austin, TX).</p>", "<title>Determination of the EMG_FT</title>", "<p>Participants returned 24–48 hours after the GXT to perform the EMG_FTtest. Following a five-minute warm-up on an electronically-braked cycle ergometer (Quinton Corival 400), participants completed four two-minute cycling bouts at incrementally ascending workloads (75 W–300 W). The initial workload corresponded with the workload at which VT occurred, determined during the GXT. Adequate rest was given between bouts to allow for participants' heart rate to drop within 10 beats of their resting heart rate. The rate of rise in EMG amplitude values (EMG slope) from the four workloads were plotted over 120 seconds (Figure ##FIG##0##1a##). The EMG slope values for each of the four power outputs were then plotted to determine EMG_FT(Figure ##FIG##0##1b##). The line of best fit was extrapolated to the y-axis, and the power output at which it intersected the y-axis was defined as the EMG_FT. The participants completed the EMG_FTprotocol two times; familiarization trial and baseline.</p>", "<p>Test-rest reliability for the EMG_FTprotocol, determined at the University of Oklahoma, resulted in an intraclass correlation coefficient (ICC) of 0.935 (SEM 5.03 W). The ICC from this lab was higher than previously reported using the vastus lateralis (ICC = 0.65) [##REF##8299596##11##].</p>", "<title>Statistical Analysis</title>", "<p>Each participant's power outputs from the EMG_FTand the VO_2PEAKcorresponding to the outputs during the GXT were regressed. A linear equation was developed to predict the VO₂value that corresponded to the EMG_FT(EMG_FTVO₂). A one-way repeated measures ANOVA was used to determine differences between the EMG_FTVO₂, VT, and VO_2PEAK. When appropriate, follow-up dependent t-test analyses were run. Correlation analyses were run to determine the strength of the relationship between EMG_FTvs. VT (watts) and EMG_FTVO₂vs. VT (l·min^-1). All data are reported as mean ± S.E.</p>" ]

[ "<title>Results</title>", "<p>A one-way repeated measures analysis of variance (ANOVA) indicated a significant (p &lt; 0.001) difference among metabolic parameters for EMG_FTVO₂, VT, and VO_2PEAK. Table ##TAB##1##2## presents the mean metabolic and power output values for EMG_FTand VT, as well as the correlation coefficients for these variables. Dependent t-test analyses resulted in no significant differences (p = 0.794) between the power output at which EMG_FTand VT occurred, as well as no significant differences (p = 0.204) between the EMG_FTVO₂and VT. However, the VO_2PEAKvalues were significantly different from both parameters. Furthermore, power output and metabolic parameters for EMG_FTand VT were strongly correlated (r = 0.766 and r = 0.750, respectively). Figure ##FIG##1##2## displays the relationship between EMG_FTand VT parameters for mean power output (W) and metabolic values (l·min^-1). Based on significant correlation analysis (Table ##TAB##1##2##), a regression equation was developed to predict VT from EMG_FTwhich resulted in a strong relationship with a low (less than 4% of mean) standard error of estimate (SEE):</p>", "<p></p>" ]

[ "<title>Discussion</title>", "<p>The results of the present study demonstrated support for previous work verifying the use of the EMG_FTas a reliable and non-invasive method for identifying the onset of neuromuscular fatigue [##REF##1915324##1##, ####REF##8514689##2##, ##REF##7173165##3##, ##REF##10378070##4##, ##REF##9754971##5##, ##REF##12914560##6##, ##REF##17149988##7####17149988##7##]. In addition, a highly significant relationship between power output values at EMG_FTand VT was found. Furthermore, no significant difference between metabolic values at EMG_FTVO₂and VT was found. Several studies have suggested the use of the EMG_FTas a simple and attractive alternative to identify the onset of fatigue [##REF##1915324##1##, ####REF##8514689##2##, ##REF##7173165##3####7173165##3##,##REF##12914560##6##,##REF##17149988##7##,##REF##3119335##12##]. The results of the current study further support the myoelectrical and physiological similarities proposed between the EMG_FTand VT.</p>", "<p>The EMG_FTtheoretically represents the highest power output that can be sustained without electromyographic evidence of neuromuscular fatigue [##REF##1915324##1##,##REF##8514689##2##]. In addition, the VT has been proposed to correlate with a workload that theoretically can be maintained without evidence of fatigue [##REF##17149988##7##]. The VT may be an indicator of the ability of the cardiovascular system to adequately supply oxygen to the working muscles to prevent muscle anaerobisis [##REF##2403868##13##]. Performing exercise at an intensity greater than the VT would result in an inadequate supply of oxygen to the working muscle, resulting in the recruitment of Type II muscle fibers, quickly leading to fatigue [##REF##2403868##13##]. The fatigued state of a muscle has been associated with changes in motor unit recruitment and/or changes in the frequency of motor unit firing resulting in an increase in EMG activity [##REF##7328909##8##]. Several studies have proposed a strong physiological relationship between VT and the onset of neuromuscular fatigue, with both measures representing recruitment of Type II muscle fibers due to the transition from aerobic to anaerobic metabolism [##REF##7173165##3##,##REF##10378070##4##,##REF##12914560##6##,##REF##7328909##8##,##REF##7394286##14##]. As a result, there would be an increase in muscle lactate concentration corresponding to a decrease skeletal muscle pH, which may further signal arterial chemoreceptors that alter ventilatory regulating mechanisms [##REF##632175##15##, ####REF##3663177##16##, ##REF##8820885##17####8820885##17##]. The evidence presented in this study suggests that the EMG_FTand VT may reflect similar acute physiological adaptations that occur during exercise.</p>", "<p>The data in the present study are in agreement with previous investigations that have reported VT and EMG_FTto occur at similar power outputs during cycle ergometry [##REF##1915324##1##,##REF##7173165##3##,##REF##17149988##7##,##REF##7328909##8##,##REF##3119335##12##]. In addition, the current study provides new data indicating no significant difference between the VT and EMG_FTVO₂. In contrast, Moritani et al. determined EMG_FTVO₂by calculating each participant's delta mechanical efficiency values [##REF##8514689##2##], as described by Gaesser and Brooks [##REF##1141128##18##], during the incremental exercise test. Although Moritani et al. reported a significant difference between VT and EMG_FTVO₂using the delta mechanical efficiency technique, Gaesser and Brooks determined that this technique was not valid. However, the significant relationships (Table ##TAB##1##2##) between VT vs. EMG_FTand VT vs. EMG_FTVO₂found in the present study suggest the possibility of using EMG_FT, rather than gas analysis, to predict VT. Based on this assumption, a regression equation was developed to predict VT from EMG_FT: VT (W) = 0.665(EMG_FT) + 41.53; SEE = 13 W. The strong correlation and low prediction error (SEE &lt; 4.0%) indicate that the EMG_FTtest may be an alternative and salient method to predict VT.</p>" ]

[ "<title>Conclusion</title>", "<p>In summary, the relationship between VT and EMG_FTVO₂suggests a possible attractive alternative to measuring VT via gas analysis. Determining VT using gas analysis requires participants to reach volitional fatigue during a graded exercise test, and, therefore, the results may be influenced by motivation. The EMG_FTtest consists of submaximal workloads which should eliminate the influence of participant motivation. In addition, due to the submaximal nature of the test, it may provide a safe alternative to determining VT for clinical populations in which maximal exertion may not be safe. Furthermore, the EMG_FTtest may reduce or eliminate discomfort experienced during gas analysis due to the gas measurement equipment. However, additional studies are needed to validate the regression equation proposed in the present study to predict VT using EMG_FT. In addition, future studies are warranted to determine whether the regression equation can accurately track changes in VT over time with training.</p>" ]

[ "<p>This is an Open Access article distributed under the terms of the Creative Commons Attribution License (<ext-link ext-link-type=\"uri\" xlink:href=\"http://creativecommons.org/licenses/by/2.0\"/>), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.</p>", "<title>Background</title>", "<p>The use of surface electromyography has been accepted as a valid, non-invasive measure of neuromuscular fatigue. In particular, the electromyographic fatigue threshold test (EMG_FT) is a reliable submaximal tool to identify the onset of fatigue. This study examined the metabolic relationship between VO_2PEAK, ventilatory threshold (VT), and the EMG_FT, as well as compared the power output at VO_2PEAK, VT, and EMG_FT.</p>", "<title>Methods</title>", "<p>Thirty-eight college-aged males (mean ± SD = 22.5 ± 3.5 yrs) performed an incremental test to exhaustion on an electronically-braked cycle ergometer for the determination of VO_2PEAKand VT. Each subject also performed a discontinuous incremental cycle ergometer test to determine their EMG_FTvalue, determined from bipolar surface electrodes placed on the longitudinal axis of the vastus lateralis of the right thigh. Subjects completed a total of four, 2-minute work bouts (ranging from 75–325 W). Adequate rest was given between bouts to allow for subjects' heart rate to drop within 10 beats of their resting heart rate. The EMG amplitude was averaged over 10-second intervals and plotted over the 2-minute work bout. The resulting slopes from each successive work bout were used to calculate EMG_FT.</p>", "<title>Results</title>", "<p>Power outputs and VO₂values from each subject's incremental test to exhaustion were regressed. The linear equations were used to compute the VO₂value that corresponded to each fatigue threshold. Two separate one-way repeated measure ANOVAs indicated significant differences (p &lt; 0.05) among metabolic parameters and power outputs. However, the mean metabolic values for VT (1.90 ± 0.50 l·min^-1) and EMG_FTVO₂(1.84 ± 0.53 l·min^-1) were not significantly different (p &gt; 0.05) and were highly correlated (r = 0.750). Furthermore, the mean workload at VT was 130.7 ± 37.8 W compared with 134.1 ± 43.5 W at EMG_FT(p &gt; 0.05) with a strong correlation between the two variables (r = 0.766).</p>", "<title>Conclusion</title>", "<p>Metabolic measurements, as well as the power outputs at VT and EMG_FT, were strongly correlated. The significant relationship between VT and EMG_FTsuggests that both procedures may reflect similar physiological factors associated with the onset of fatigue. As a result of these findings, the EMG_FTtest may provide an attractive alternative to estimating VT.</p>" ]

[ "<title>Competing interests</title>", "<p>The authors declare that they have no competing interests.</p>", "<title>Authors' contributions</title>", "<p>JG, AS, and KK contributed in writing and editing the manuscript along with concept and design, data acquisition, and data analysis and interpretation. AW and CL contributed in concept and design, data acquisition, and data analysis and interpretation. JM, TB, JC, and JS contributed in writing and editing the manuscript, as well as concept and design. All authors have read and approved the final manuscript.</p>" ]

[ "<title>Acknowledgements</title>", "<p>None</p>" ]

[ "<fig position=\"float\" id=\"F1\"><label>Figure 1</label><caption><p><bold>Determination of EMG_FT</bold>. <bold>a</bold>. Describes the relationship between EMG amplitude and time for the four power outputs used in the EMG_FTtest. The greatest slope was a result from the highest power output. <bold>b</bold>. Depicts the relationship for the power outputs versus slope coefficients with the y-intercept defined as the EMG_FT.</p></caption></fig>", "<fig position=\"float\" id=\"F2\"><label>Figure 2</label><caption><p><bold>Comparison of EMG_FTand VT</bold>. The relationship between differences in EMG_FTand VT mean power outputs (W) and metabolic values (l·min^-1).</p></caption></fig>" ]

[ "<table-wrap position=\"float\" id=\"T1\"><label>Table 1</label><caption><p>Descriptive statistics (mean ± SD) of the subjects.</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td/><td align=\"center\">Subjects (n = 35)</td></tr></thead><tbody><tr><td align=\"center\">Age (yrs)</td><td align=\"center\">22.6 ± 3.5</td></tr><tr><td align=\"center\">Height (cm)</td><td align=\"center\">177.1 ± 7.1</td></tr><tr><td align=\"center\">Weight (kg)</td><td align=\"center\">77.0 ± 11.0</td></tr></tbody></table></table-wrap>", "<table-wrap position=\"float\" id=\"T2\"><label>Table 2</label><caption><p>Mean ± standard error (SE) values and correlations for EMG_FTand VT.</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td/><td/><td/><td align=\"center\" colspan=\"2\"><bold>Correlation analysis</bold></td></tr><tr><td/><td/><td/><td colspan=\"2\"><hr/></td></tr><tr><td/><td align=\"center\"><bold>Mean ± SEM (l-min^-1)</bold></td><td align=\"center\"><bold>Mean ± SEM (W)</bold></td><td align=\"center\"><bold>EMG_FT(l-min^-1)</bold></td><td align=\"center\"><bold>EMG_FT(W)</bold></td></tr></thead><tbody><tr><td align=\"left\">Electromyographic Fatigue Threshold</td><td align=\"center\">1.84 ± 0.09</td><td align=\"center\">134.11 ± 7.06</td><td align=\"center\">1.000</td><td align=\"center\">1.000</td></tr><tr><td align=\"left\">Ventilatory Threshold</td><td align=\"center\">1.89 ± 0.08</td><td align=\"center\">130.71 ± 6.13</td><td align=\"center\">0.750*</td><td align=\"center\">0.766*</td></tr></tbody></table></table-wrap>" ]

[ "<disp-formula>VT (W) = 0.665(EMG_FT) + 41.53; SEE = 13 W</disp-formula>" ]

[]

[]

[]

[]

[]

[ "<table-wrap-foot><p>*p &lt; 0.01</p></table-wrap-foot>" ]

[ "<graphic xlink:href=\"1476-5918-7-15-1\"/>", "<graphic xlink:href=\"1476-5918-7-15-2\"/>" ]

[]

{ "acronym": [], "definition": [] }

[ "<title>Background</title>", "<p>Subsidies of energy and nutrients across habitat boundaries can affect the behaviour and life history in a broad array of taxa [##UREF##0##1##, ####UREF##1##2##, ##REF##10603500##3##, ##UREF##2##4##, ##UREF##3##5##, ##REF##16681850##6####16681850##6##]. The return of salmon (<italic>Oncorhynchus </italic>spp.) from oceanic environments to terrestrial spawning areas provides a striking example of such cross-boundary resource subsidy. Offering a predictable, nutritiously valuable, and spatially and temporally constrained food, salmon attract a diversity of terrestrial predators and scavengers [<italic>e.g. </italic>[##UREF##4##7##, ####UREF##5##8##, ##UREF##6##9##, ##REF##11914157##10##, ##UREF##7##11##, ##UREF##8##12##, ##UREF##9##13##, ##UREF##10##14##, ##REF##18509510##15##, ##UREF##11##16##, ##UREF##12##17####12##17##]].</p>", "<p>Among multiple terrestrial users, only few capture salmon, transferring nutrients to adjacent shorelines and subsequent consumers. Although river otters (<italic>Lontra canadensis</italic>) and flooding activity contribute [##UREF##6##9##], critical to this process are bears (<italic>Ursus </italic>spp.), which partially consume salmon and deposit carcass remains as well as their urine and faeces (containing salmon-derived nutrients) throughout riparian areas [<italic>e.g. </italic>[##UREF##4##7##,##UREF##7##11##,##UREF##13##18##,##REF##14673639##19##]]. This behaviour directly and indirectly provides nutrients to multiple trophic levels, including vegetation, through scavenging of carcasses followed by decay and subsequent fertilization of riparian vegetation [##UREF##13##18##,##UREF##14##20##, ####REF##12729462##21##, ##UREF##15##22##, ##UREF##16##23####16##23##]. If bears are the primary and most widely distributed vectors linking salmon to terrestrial environments, one might predict the ecological consequences based solely on these and similar studies of bear-salmon interactions.</p>", "<p>Another terrestrial carnivore has been linked to salmon but their ecological relationship is not well understood. There are tangential observations of salmon as a food resource for wolves (<italic>Canis lupus</italic>), but none describing them as frequent prey [##UREF##17##24##, ####UREF##18##25##, ##UREF##19##26####19##26##]. In wolves of coastal and interior Alaska, however, Szepanski et al. [##UREF##20##27##] identified marine-enriched stable isotope signatures and suggested the dominant source was spawning salmon. With complementary results, Darimont and Reimchen [##UREF##21##28##] sampled chronologically segmented portions of guard hair from wolves across British Columbia (BC) and demonstrated that seasonal marine isotopic enrichment occurred during fall, when salmon became available. Subsequently, a survey of prey remains in wolf faeces across 60,000 km²of coastal BC detected the presence of salmon in about 7% of samples, even though sampling primarily occurred before the spawning season began [##UREF##22##29##]. Finally, behavioural evidence from coastal BC suggested that wolves are not simply scavengers but can efficiently prey on salmon [##UREF##23##30##]. Collectively, these observations suggest that wolves might also be a frequent and widespread predator of salmon and biological vector of salmon-derived nutrients.</p>", "<p>Such a wolf-salmon association would depart from the dominant pattern defining this terrestrial carnivore. Recent reviews concluded that, although wolves are flexible and opportunistic predators, they primarily prey on ungulates – or hoofed animals – and ungulate presence and density in an area determines the distribution, behaviour, and ultimately reproduction and survival of wolves [##UREF##24##31##, ####UREF##25##32##, ##UREF##26##33####26##33##]. Consistent with this conclusion, Szepanski et al. [##UREF##20##27##] reasoned that the greater salmon consumption among wolves of mainland southeast Alaska they estimated was likely related to reduced ungulate (black-tailed deer; <italic>Odocoileus hemionus</italic>) availability; previous research had shown that deer densities were lower on the mainland compared with the islands. This supported the hypothesis that salmon were alternate prey to which wolves switch under conditions of low ungulate abundance. Likewise, it is consistent with broader theory that predators will switch to alternative prey when preferred foods are less available [##UREF##27##34##,##UREF##28##35##].</p>", "<p>We address this hypothesis with resource use data from eight groups of wolves for three seasons over four years across a landscape that varies in availability of ungulates and salmon. We estimate resource use using two methods: i) identification of prey remains in wolf faeces and, ii) stable isotope analysis [<italic>review in </italic>[##UREF##29##36##]]. Relevant here, marine resources like salmon have higher carbon and nitrogen isotopic signatures compared with terrestrial foods, making a signal of marine resource use detectable in the tissues of consumers [##UREF##30##37##]. Primarily, we test whether wolves use salmon as a function of deer or salmon availability. Additionally, we examine how consistent faecal and isotopic data sets might be in estimating seasonal and intrapopulation variation in foraging. We show strong concordance between these data sets and that wolves target salmon as a function of salmon availability, not deer availability.</p>" ]

[ "<title>Methods</title>", "<title>Study Area</title>", "<p>BC's central coast is a remote area, accessible only by boat or air, and only minimally modified by industrial activity [##UREF##22##29##]. Our study area is roughly 3,300 km², and is centred on Bella Bella (52° 10' N, 128° 09' W; Figure ##FIG##0##1##).</p>", "<title>Assessing resource availability</title>", "<p>To assess variation in resource availability among wolf groups, we first estimated home ranges using data on re-sightings of individuals. After hundreds of hours of direct observations that included videography and photography [<italic>e.g. </italic>[##UREF##23##30##,##UREF##31##38##]], differences among wolves in pelage and other morphological characters allowed us to identify repeatedly at least one member of each group over the 4 years of study. We used ArcView 3.2 to plot these re-sightings and used the 'Home Range' application to estimate 95% kernel home ranges [##UREF##32##39##]. This method might be limited by different probabilities of observing wolves among and within packs. Additionally, estimates cannot account for potential variation in home ranges among years. Nonetheless, we assume our results yield an adequate estimate of home range sizes and configurations to estimate relative resource availability for each pack (below). Indeed, microsatellite data extracted from wolf faeces collected in 2003 across five of these putative home ranges are yielding similar estimates for home range sizes and configurations (Erin Navid, University of Calgary, <italic>unpublished data</italic>).</p>", "<p>To estimate deer availability, we applied a model [##UREF##33##40##,##UREF##34##41##] we previously developed that was based on the relationship between topographical slope and deer pellet density [##UREF##35##42##], derived from 110 km of transects conducted across our study area. Model output was converted to a spatial probability layer with which we calculated a relative deer density estimate (DEER), ranging from 0 to 1, for each home range.</p>", "<p>To estimate salmon availability (SALMON), we extracted data from the Pacific Salmon Escapement Database (nuSEDS), maintained by Fisheries and Oceans Canada for each creek in each year. We then converted escapement numbers to biomass available in each 95% kernel home range using published weights for each species [##UREF##36##43##,##UREF##37##44##], and assuming a 1:1 sex ratio. The number of inventoried salmon creeks in each home range varied from 1 to 8. We assumed that wolves have equal access to each creek within their territories.</p>", "<title>Assessing resource use</title>", "<p>During spring (May/early June), summer (late July), and fall (late September/early October), we collected wolf faeces on established transects. In 2001, we sampled the home ranges of four groups, and in 2002 and 2003, we added four more to total eight groups sampled each season and year. Within home ranges, sites were well-distributed and on average about half included creeks with spawning salmon.</p>", "<p>During spring and summer 2001 to 2004, we also collected wolf hair that had been shed in resting beds on established transects or at 'homesites' (reproductive areas; [##UREF##38##45##]). Wolves have one annual moult that begins in late spring when the old coat sheds and a new one grows until late fall [##UREF##17##24##]. Therefore each hair sample's isotopic datum provides an integrated record of individual diet for roughly half the previous year. We assume each sample originated from one wolf, as they were collected from resting beds and on most occasions we sampled hair directly after viewing wolves.</p>", "<p>Identification of prey remains used dichotomous keys [<italic>e.g. </italic>[##UREF##39##46##]] and followed protocols in Darimont et al. [##UREF##23##30##]. To eliminate inter-observer variability, only one person identified prey remains, and only after a lengthy training period (~60 hours). We estimated her precision by having an independent volunteer select 141 scats (~6%) for re-sampling, as well as administer and score the results. The primary prey item was consistently identified in 139 (98.6%).</p>", "<p>Isotopic analysis of hair followed Darimont and Reimchen [##UREF##21##28##]. Isotopic signatures are expressed in delta notation (δ) as ratios relative to PeeDee limestone (carbon) and atmospheric N₂(nitrogen) standards as follows:</p>", "<p></p>", "<p>where X is ¹³C or ¹⁵N, and R is the corresponding ratio ¹³C/¹²C or ¹⁵N/¹⁴N. Isotopic data are expressed in delta notation (δ) in ‰ units [##UREF##29##36##].</p>", "<title>Assessing resource use in context of resource availability</title>", "<p>We used faecal data to document seasonal and intra-group differences in resource use but focus on isotopic data to test hypotheses of resource selection. For scat data, we report occurrence per faeces (O/F) for comparison with published literature but use occurrence per item (O/I) in statistical tests because the former can be problematic, as it exceeds unity when summed (because some faeces contain multiple items). O/F is the frequency item occurrence in all <italic>faeces</italic>, whereas O/I is the item's frequency among all <italic>items </italic>identified in all faeces. We also estimated mammalian biomass consumed using a regression equation created by Weaver [##UREF##40##47##]: Y = 0.439 + 0.008 X, where Y is the estimated biomass of prey consumed per faecal sample and X is the mass of prey. We used mean masses of adults [##UREF##40##47##, ####UREF##41##48##, ##UREF##42##49####42##49##], and assumed a 1:1 sex ratio. For deer, however, we distinguished between adults and fawns hair using diagnostic diameter and colour characters [##UREF##43##50##] and assigned fawn mass as 25% of adult mass. By necessity, biomass estimates excluded non-mammalian prey (n = 404 of 2692 items).</p>", "<p>For isotopic data, we report signatures from whole hairs as well as in approximately equal distal and proximal segments (relative to root), which – given known moult chronology – provide proxies for summer and fall diets, respectively. We calculated any 'seasonal isotopic shifts' by subtracting summer from fall values [##UREF##21##28##]. In wolves that received δ¹³C and δ¹⁵N enrichment from salmon, which are available only during fall, one would expect positive seasonal isotopic shifts.</p>", "<p>We used information theory to distinguish among competing hypotheses. Specifically, we developed a simple set of candidate models [weighted least squares general linear models (GLMs)] to examine how the availability of deer (DEER), salmon (SALMON), and their interaction might influence salmon use by wolves, and included year (YEAR) as a random term. We used the average δ¹³C seasonal isotopic shift of each group in each year as the dependent variable (n = 15 'group years') and proxy for salmon use for several reasons. First, faecal analyses might be sensitive to numerical and spatial sampling biases; faecal sample sizes varied considerably among 'pack seasons' (n = 9 to 132 scats) and might be biased to contain resources most available at the location of defaecation. In contrast, isotopic signatures incorporate many months of foraging behaviour. Second, we focused on δ¹³C because it is a much better tracer of dietary 'source' (<italic>i.e. </italic>marine versus terrestrial) than δ¹⁵N, which also reflects trophic position [##UREF##43##50##,##UREF##44##51##]. Third, if wolves used salmon, they should show elevated δ¹³C signatures in the fall-grown hair compared to summer-grown hair [##UREF##21##28##].</p>", "<p>For each candidate model, we calculated Akaike Information Criteria (AIC), adjusted for small sample sizes, following the formula: AIC_c= n log(<italic>o</italic>²) + 2K + 2K(K + 1)/(n - K - 1), where <italic>o</italic>²= Sum (e_i²/n), K is the number of parameters (including intercept and error term), n the numbered of 'group years' and e_ithe residuals for each candidate model [[##UREF##45##52##], p. 63]. We then evaluated ΔAIC_cto select best approximating model(s) and make appropriate inference, using ΔAIC_c&lt; 4 to describe the top model set. Finally, we summed Akaike weights (Σω_i) across the top model set for each variable to rank them by importance [##UREF##45##52##]. δ¹³C seasonal isotopic shift data were normally distributed (Kolmogorov-Smirnov Z test; P = 0.35). Models were weighted by the square root of sample size for each 'group year'. Each candidate model had errors that were normally distributed (Kolmogorov-Smirnov Z tests, all P &gt; 0.05). Tests were performed using SPSS 11.0 (SPSS Inc., Chicago, USA).</p>" ]

[ "<title>Results</title>", "<title>Resource availability</title>", "<p>Deer and salmon availability differed among groups. Average probabilities of detecting deer pellets across home ranges (our proxy for relative deer density) varied from 0.06 to 0.26 among the 8 social groups. More variation existed in salmon availability among 'group years', which ranged from approximately 1 to over 220 metric tonnes per group per year.</p>", "<p>Resource availability can also be expressed in terms of nutrients in different foods groups. Although comparable in protein, salmon provide roughly 30% more fat than deer and more than four times the energetic content per unit mass (Table ##TAB##0##1##).</p>", "<title>Resource use among seasons</title>", "<p>Faecal data (n = 2203 scats) collected over spring, summer, and fall showed strong seasonal patterns in resource use. Over all seasons combined and at the population level, resource use was broad but deer dominated diet, occurring in 90 to 95% of faeces during spring and summer (See additional file ##SUPPL##0##1##: Prey items identified in the faeces of wolves of coastal British Columbia). During fall, when salmon become available, however, the population diverged from a deer-dominated diet; for years pooled, population-level occurrence per item (O/I) of deer was significantly lower in the fall (ANOVA; F_2,21= 26.54, P &lt; 0.001; Tamhane's T2 comparing fall with spring and summer, both P &lt; 0.001). This difference was also significant in individual years (ANOVAs; all P &lt; 0.005). Estimates of salmon occurrence per faeces (O/F) during fall averaged 40% and approached 70% for some groups (Figure ##FIG##1##2##; See additional file ##SUPPL##0##1##: Prey items identified in the faeces of wolves of coastal British Columbia). This pronounced shift in foraging behaviour to declining use of deer during fall was strongly related to salmon use; using 'group years' as cases, there was a strong inverse relationship between O/I of salmon and O/I of deer during fall (r = -0.77, n = 20, P &lt; 0.001, Figure ##FIG##2##3##).</p>", "<p>Isotopic data also showed seasonal variation in resource use, much of it related to salmon use. Whole hair δ¹³C values, indexing diet from the summer to fall, ranged from -24.4 to -16.8 (mean = -21.4, SD = 2.0), and δ¹⁵N ranged from 6.4 to 14.3 (mean = 9.5, SD = 2.1). Reflecting the marine nature of this variation, δ¹³C and δ¹⁵N were strongly correlated (r = 0.95, n = 60, P &lt; 0.001).</p>", "<p>Three tests revealed that most marine-derived isotopic enrichment was incorporated during fall and associated with salmon. First, δ¹³C values in whole hair samples were correlated with seasonal (fall minus summer) isotopic shifts in δ¹³C in the same hair (r = 0.54, n = 15, P = 0.038). Second, most individuals showed positive seasonal isotopic shifts between summer and fall, occupy the region of isotopic niche space defined by greater δ¹³C and δ¹⁵N signatures during fall compared with summer (χ²= 56.13, df = 3, n = 60, P &lt; 0.001; Figure ##FIG##3##4##). Third, we examined 'group years' with both faecal data during fall (n range: 9 to 92; mean = 54.0 faeces/group) and group-averaged δ¹³C seasonal isotopic shifts from wolf hair grown during that same year among members of those same groups (n range: 1 to 6; mean = 3.4 hair samples/group); cases with higher O/I salmon during fall showed greater average seasonal isotopic shifts in δ¹³C (r = 0.78, n = 10, P = 0.008, Figure ##FIG##4##5##).</p>", "<title>Inter-group variation in resource use</title>", "<p>Groups varied in salmon use as assessed by both faecal and isotopic data. In a GLM, weighted by the square root of the number of items in all scats in each 'group season', variation among groups in O/I salmon during autumn approached significance (P = 0.051, Figure ##FIG##1##2##). In similar designs, but weighted by the square root of the number of hair samples used to compute averages for each 'group year', δ¹³C signatures in un-segmented wolf hair also differed among groups (P = 0.034), and approached significance for seasonal isotopic shifts in δ¹³C (P = 0.059).</p>", "<title>Resource use in context of resource availability</title>", "<p>This variation in resource use among groups was relatively insensitive to estimated deer availability but correlated positively to salmon availability. In our first evaluation, using the entire dataset, SALMON and YEAR had the greatest utility in predicting the δ¹³C seasonal isotopic shift, our proxy for salmon use. Summing weights among top models ranked SALMON (Σω_i= 0.61) marginally above YEAR (Σω_i= 0.44), whereas DEER and DEER × SALMON were ranked much lower (both Σω_i= 0.14). DEER occurred in only one top model, and with a positive parameter coefficient, suggesting – not consistent with either hypothesis – greater salmon use with greater deer availability. However, a bi-variate plot showed no linear relationship between DEER and δ¹³C seasonal isotopic shift (Figure ##FIG##5##6a##). Examination of parameter coefficients for SALMON revealed a strongly positive and significant effect, but only in the third model (See additional file ##SUPPL##1##2##: Top model sets to predict the use of salmon by wolves). A bi-variate plot between SALMON and δ¹³C seasonal isotopic shift showed how an outlier SALMON datum (Mosquito group 2002) influenced results (Figure ##FIG##5##6b##).</p>", "<p>We re-evaluated candidate models but modified the Mosquito 2002 outlier datum to exclude the contribution of the Neekas watershed to the home range's SALMON estimate. During years studied, the Neekas yielded an average of 179.9 tonnes of salmon per year, representing 81% of biomass in their home range and alone doubled the total biomass available to any group. Yet, Mosquito wolves infrequently used this watershed; the proportion of Mosquito group's faeces collected there during autumn was very low; we collected only 2 faeces in 2003 and none in 2002. With this modification, our analysis revealed that SALMON was clearly the best predictor of salmon use. The preferred model (lowest ΔAIC_c) included only SALMON (and the intercept; ω_i= 0.57, See additional file ##SUPPL##1##2##: Top model sets to predict the use of salmon by wolves). SALMON (Σω_i= 0.75) outranked YEAR (Σω_i= 0.23) and DEER (Σω_i= 0.09) by factors of about 3.3 and 8.3 respectively. A bi-variate plot (Figure ##FIG##5##6c##) revealed a significant and positive correlation between δ¹³C seasonal isotopic shift and SALMON to which a linear (r = 0.67, n = 15, P = 0.006) and quadratic form (r = 0.74, n = 15, P = 0.009) could be fit.</p>" ]

[ "<title>Discussion</title>", "<p>Determining which resources are used in the context of their availability provides fundamental life history information and can yield insight into the ecological relationships among consumer, prey, and the ecosystem. Consistent with prevailing knowledge about wolf-prey systems, for much of the year wolves of coastal BC are closely tied to ungulate prey. During autumn, however, an alternate predator-prey system emerged with previously undocumented ecological detail. When salmon became available seasonally, we observed a population-level shift in resource use as indicated by two independent datasets. Associations between the occurrence of salmon in fall faeces and seasonal isotopic shifts were significant and moderately strong. This suggests the intra-hair methodology [##UREF##21##28##] offers an accurate proxy for salmon consumption, and perhaps also for tracking seasonal dietary shifts in other predator-prey systems.</p>", "<p>Many systems receive pulsed food resources, which decay in abundance over time. Because there are long durations between pulses, theory predicts that few consumers will be specialists on such resources. Instead, generalist consumers should be most likely to respond [##REF##10802548##53##]. Across their remaining Holarctic distribution, although wolves are opportunistic and able to subsist on alternate foods such as beaver, livestock or even garbage, close ecological and evolutionary associations with ungulate prey are the norm [##UREF##24##31##, ####UREF##25##32##, ##UREF##26##33####26##33##]. With this perspective, it follows that any departure from a diet dominated by ungulates might occur only during times or in areas of low ungulate availability.</p>", "<p>In contrast, our data suggest salmon are a targeted resource. Salmon availability clearly outperformed deer availability in predicting use of salmon. Although not a highly important variable, there was variation in salmon use among years. This could represent many conditions that might change yearly, including climate (and deer vulnerability) and competitive interactions (below).</p>", "<p>How we estimated resource availability influences interpretation of results. Manly et al. [##UREF##46##54##] cautioned researchers to carefully consider difference between resource <italic>availability </italic>and <italic>abundance</italic>. Our deer model yielded a coarse estimate of relative deer abundance across large home ranges, and one that does not vary among years. Actual availability (<italic>i.e. </italic>numbers and vulnerability) might be different. For example, coastal black-tailed deer have phenotypes that are resident at low elevations year round and those that seasonally migrate to higher elevations during summer [##UREF##47##55##]. Differences among home ranges in the proportion of these phenotypes might influence the availability of deer to wolves. Regardless, the positive correlation between salmon availability and use is straightforward, and alone provides support to differentiate between hypotheses.</p>", "<title>Adaptive explanations for use of salmon</title>", "<p>Whereas this wolf-prey association during fall departs from a 'wolf-ungulate' model, it is consistent with adaptive explanations based on safety, nutrition, and energetics. Selecting benign prey such as salmon over potentially dangerous ungulate prey follows predictions of foraging theory [##UREF##48##56##]. While hunting ungulates, wolves commonly incur serious and often fatal injuries [##UREF##24##31##].</p>", "<p>In addition to safety benefits, we show here that salmon also provides enhanced nutrition over deer, especially in fat and energy. Moreover, strict comparisons might underestimate the nutritional value of salmon. Wolves selectively consume lipid-rich heads [##UREF##23##30##] and potentially benefit from docosahexaenoic acid, an omega-3 fatty acid, which is critical for nervous system function, can be manufactured only from dietary sources, and occurs at high levels in brain and optic tissue [##REF##10479465##57##]. Finally, for equivalent energetic intake, wolves face less handling time and need to travel far less for salmon compared with searching for vulnerable ungulate prey [<italic>e.g. </italic>[##UREF##49##58##]]. If we consider energetic content as a central currency, and given a ratio of its value per mass of pink salmon compared with deer (4.4:1, calculated from Table ##TAB##0##1##) and an estimated daily requirement of 2.7 kilograms of deer per wolf of average mass per day among coastal populations [##UREF##50##59##], wolves that forgo deer would on average require only 0.62 kg of pink salmon each day. If wolves consume exclusively salmon heads that comprise (a conservatively estimated) 10% of the average mass of pink salmon in the area [1.3 kg; [##UREF##36##43##,##UREF##37##44##]], these energetic requirements would be satisfied by capturing only 4.6 salmon per day.</p>", "<title>Processes that might constrain use of salmon</title>", "<p>These safety, nutritional, and energetic benefits conferred in a spatially-constrained food resource would promote competition with other salmon consumers. Brown and black bears have been observed in several competitive interactions with wolves over resources [<italic>e.g. </italic>[##UREF##51##60##]], including salmon [##UREF##52##61##]. Such interactions might be most intense under conditions of high resource density, and could explain why wolves avoid the Neekas River, which hosts extraordinarily high salmon density (in fact the highest on the entire BC coast [biomass/km]). Likewise, such competitive interactions across the study area might also explain the decline in slope in seasonal isotopic shift at higher salmon abundances (<italic>i.e. </italic>fit to a quadratic form).</p>", "<p>Additional processes might also limit the use of salmon by wolves. First, wolves might be compelled to partition their diet, perhaps requiring a particular suite of micronutrients in deer or avoiding the accumulation of others in salmon. Disease, specifically 'salmon-poisoning disease' (<italic>Neorickettsia helminthoeca</italic>), which in high quantities is fatal to canids, might also play a role [[##UREF##23##30##]<italic>and references therein</italic>]. Third, focusing on a spatially-constrained resource might create opportunity costs of not patrolling and defending larger portions of their territories.</p>", "<title>Ecological implications</title>", "<p>Based on relationships we show between availability and use, we predict salmon consumption is widespread wherever wolves and salmon still exist [<italic>see also </italic>[##UREF##22##29##,##UREF##23##30##]]. Accordingly, we expect higher-order ecological implications, similar to those initiated by wolves in other systems. For example, by preying on large ungulates, wolves indirectly provide a considerable proportion of carcasses to a diversity of scavengers, including coyotes (<italic>C. latrans</italic>), bears, and ravens (<italic>Corvus corax</italic>) [##UREF##53##62##,##UREF##54##63##]. Notable differences, however, exist between unused portions of ungulate and salmon carcasses. First, remains of salmon are not defended by wolves [##UREF##23##30##], and thus the carrion is immediately available. Second, because carcasses are relatively small and can be more readily dispersed, more individual (vertebrate) scavengers likely gain access to salmon compared with large (ungulate) carcasses over which multiple scavengers might compete. As a consequence, this subsidy might be more evenly and broadly dispersed. Finally, the resource subsidy offered by this terrestrial carnivore is one transported across a boundary of land and sea.</p>", "<p>This wolf-provided subsidy of salmon to terrestrial ecosystems also differs from that provided by bear vectors. In contrast to wolves, which often forage among or near family members, carcass transport by bears is thought to be mediated by intra-specific competition. As a consequence, one might expect different spatial patterns of nutrient subsidy. In a black bear system, Reimchen [##UREF##13##18##] observed that about 80% of salmon were transferred up to 100 m into the forest, with larger and fresher male carcasses transported further. In contrast, in 70% of previously observed transport events by wolves, carcasses were deposited on estuarine grasses, within a few metres of the creek [##UREF##23##30##]. Moreover, tissue content in abandoned carcasses also differs. Whereas wolves target head tissue, bears target brains and eggs, and under conditions of relatively low salmon abundance also consume musculature [##UREF##13##18##]. Consequently, on average more tissue (of greater energetic content) would be available to scavengers of wolf-provided carcasses.</p>", "<p>The most notable difference between wolves and bears is the distribution of these vectors across the landscape of coastal BC. Brown bears occur on the mainland, and in low densities and frequencies on inner islands; black bears commonly inhabit mainland and inner islands, but are largely absent on outer islands [##UREF##55##64##]. In contrast, wolves occur on all landmasses [##UREF##33##40##]. Therefore, wolves might be the primary biological vector on some islands, particularly isolated outer islands. Given the behavioural differences among vectors, this distributional pattern would increase and alter the 'resource shed' into which salmon are transported by terrestrial vectors [##UREF##56##65##].</p>", "<p>Wolf-salmon associations might have additional ecological implications, namely in disease ecology and terrestrial predator-prey dynamics. In addition to <italic>N. helminthoeca</italic>, our pilot work on diseases [H. Bryan, University of Saskatoon, <italic>unpublished data</italic>] has shown that wolves in areas and periods of greater salmon consumption have higher prevalence of eggs from <italic>Dyphyllobothrium </italic>spp. This fish tapeworm uses piscivorous terrestrial mammals as final hosts in its life-cycle, which crosses the marine-terrestrial boundary. Additionally, we suspect that wolves subsidized by marine prey such as salmon might limit deer populations [##UREF##22##29##]. Under allochthonous resource supply, densities (and ecological influence) of consumers can be greater than predicted by <italic>in situ </italic>productivity [##UREF##0##1##]. This hypothesis would be especially plausible on islands where deer productivity and/or immigration from other landmasses might not offset predation [##UREF##57##66##].</p>" ]

[ "<title>Conclusion</title>", "<p>Our data suggest that salmon are a targeted resource in our study area and likely wherever wolves and salmon still co-occur. This coupled with the adaptive explanations we present argue for an historical predator-prey association with broad ecological implications. The future and nature of this (formally geographically widespread) wolf-salmon association is uncertain, however, given multiple threats posed to salmon systems. These include overexploitation by fisheries and destruction of spawning habitat [##UREF##58##67##], as well as diseases from exotic salmon aquaculture [##REF##18079401##68##] that collectively have lead to coast-wide declines up to 90% over the last century [##UREF##59##69##].</p>" ]

[ "<p>This is an Open Access article distributed under the terms of the Creative Commons Attribution License (<ext-link ext-link-type=\"uri\" xlink:href=\"http://creativecommons.org/licenses/by/2.0\"/>), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.</p>", "<title>Background</title>", "<p>As a cross-boundary resource subsidy, spawning salmon can strongly affect consumer and ecosystem ecology. Here we examine whether this marine resource can influence a terrestrial wolf-deer (<italic>Canis lupus</italic>-<italic>Odocoileus hemionus</italic>) predator-prey system in coastal British Columbia, Canada. Data on resource availability and resource use among eight wolf groups for three seasons over four years allow us to evaluate competing hypotheses that describe salmon as either an alternate resource, consumed in areas where deer are scarce, or as a targeted resource, consumed as a positive function of its availability. Faecal (n = 2203 wolf scats) and isotopic analyses (n = 60 wolf hair samples) provide independent data sets, also allowing us to examine how consistent these common techniques are in estimating foraging behaviour.</p>", "<title>Results</title>", "<p>At the population level during spring and summer, deer remains occurred in roughly 90 and 95% of faeces respectively. When salmon become available in autumn, however, the population showed a pronounced dietary shift in which deer consumption among groups was negatively correlated (r = -0.77, P &lt; 0.001) with consumption of salmon, which occurred in 40% of all faeces and up to 70% of faeces for some groups. This dietary shift as detected by faecal analysis was correlated with seasonal shifts in δ¹³C isotopic signatures (r = 0.78; P = 0.008), which were calculated by intra-hair comparisons between segments grown during summer and fall. The magnitude of this seasonal isotopic shift, our proxy for salmon use, was related primarily to estimates of salmon availability, not deer availability, among wolf groups.</p>", "<title>Conclusion</title>", "<p>Concordance of faecal and isotopic data suggests our intra-hair isotopic methodology provides an accurate proxy for salmon consumption, and might reliably track seasonal dietary shifts in other consumer-resource systems. Use of salmon by wolves as a function of its abundance and the adaptive explanations we provide suggest a long-term and widespread association between wolves and salmon. Seasonally, this system departs from the common wolf-ungulate model. Broad ecological implications include the potential transmission of marine-based disease into terrestrial systems, the effects of marine subsidy on wolf-deer population dynamics, and the distribution of salmon nutrients by wolves into coastal ecosystems.</p>" ]

[ "<title>Authors' contributions</title>", "<p>CTD lead fieldwork, conducted statistical analyses, and drafted the manuscript. All authors participated in study design and manuscript revision, and approved the final manuscript. This work was part of the Salmon Forest Project, conceived by TER.</p>", "<title>Supplementary Material</title>" ]

[ "<title>Acknowledgements</title>", "<p>We thank C. Aries, P. Clement, G. Gladstone, J. Gordon-Walker J. Housty, G. Pfleuger, and C. Starr for valuable assistance in the field, the Raincoast Conservation Foundation for primary funding, and J. Gordon-Walker for scat analysis. We also thank the National Geographic Society, Patagonia, and the following foundations for support: Bullitt, McCaw, Summerlee, Vancouver, and Wilburforce. CTD was supported by NSERC Graduate and Postdoctoral Fellowships, PCP by WWF Canada, and TER by NSERC operating grant A2354 and the David Suzuki Foundation.</p>" ]

[ "<fig position=\"float\" id=\"F1\"><label>Figure 1</label><caption><p><bold>Study area and home ranges of wolf social groups</bold>. Study area in which wolves (<italic>Canis lupus</italic>) were sampled for hair and faeces on the central coast of British Columbia, 2001 to 2004. Home ranges estimated as 95% kernels based on re-sightings of individual wolves.</p></caption></fig>", "<fig position=\"float\" id=\"F2\"><label>Figure 2</label><caption><p><bold>Inter-group and -season variability in prey remains identified in wolf faeces</bold>. Wolf (<italic>Canis lupus</italic>) faeces collected during spring, summer and fall, pooled across 2001 to 2003 in coastal British Columbia. Local, Ochre, Mosquito and Mystery groups sampled in 2002 and 2003 only. Remaining groups were sampled in all 3 years. 'Other' are prey as identified in Table 1. Occurrence per faeces measures the frequency of occurrence of an item among total faeces of each group.</p></caption></fig>", "<fig position=\"float\" id=\"F3\"><label>Figure 3</label><caption><p><bold>Relationship between consumption of deer and salmon by wolves during autumn</bold>. Occurrence per item (O/I) of deer (<italic>Odocoileus hemionus</italic>) and salmon (<italic>Oncorhynchus </italic>spp.) in wolf (<italic>Canis lupus</italic>) faeces collected during fall 2001 (n = 4 groups), 2002 (n = 8), and 2003 (n = 8) in coastal British Columbia. O/I measures the frequency of occurrence of an item among total items identified in a group's faeces during a given period (in this case, fall).</p></caption></fig>", "<fig position=\"float\" id=\"F4\"><label>Figure 4</label><caption><p><bold>Seasonal isotopic shifts in wolf hair</bold>. Seasonal isotopic shifts in δ¹³C and δ¹⁵N in wolf (<italic>Canis lupus</italic>) hair, collected in coastal British Columbia, 2001 to 2004. Seasonal shifts calculated by subtracting values in distal (summer-grown) hair segments from basal (fall-grown) hair segments.</p></caption></fig>", "<fig position=\"float\" id=\"F5\"><label>Figure 5</label><caption><p><bold>Relationship between faecal and isotopic data to detect salmon use by wolves</bold>. Salmon (<italic>Oncorhynchus </italic>spp.) remains in wolf (<italic>Canis lupus</italic>) faeces expressed as occurrence per item in each group during fall and the mean seasonal isotopic shift, which is the fall minus summer δ¹³C values in wolf hair, averaged among individuals of the same groups grown during the same year. Samples collected in coastal British Columbia, 2001 to 2004.</p></caption></fig>", "<fig position=\"float\" id=\"F6\"><label>Figure 6</label><caption><p><bold>Relationships between salmon use by wolves and the availability of deer and salmon</bold>. Mean group δ ¹³C seasonal isotopic shift in wolf (<italic>Canis lupus</italic>) hair – a proxy for salmon use – as a function of estimated: a) deer (<italic>Odocoileus hemionus</italic>) availability, b) salmon (<italic>Oncorhynchus </italic>spp.) availability, and c) salmon availability in a data set in which the Mosquito group 2002 salmon estimate datum excluded the Neekas River, the most productive in the study area, but one where wolf sign was rarely observed during fall. Samples collected in coastal British Columbia, 2001 to 2004.</p></caption></fig>" ]

[ "<table-wrap position=\"float\" id=\"T1\"><label>Table 1</label><caption><p>Mean nutritional content in 100 grams of raw black-tailed deer and pink salmon.</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td align=\"left\"><bold>Content</bold></td><td align=\"center\"><bold>Deer</bold></td><td align=\"center\"><bold>Salmon</bold></td></tr></thead><tbody><tr><td align=\"left\">Protein (g)</td><td align=\"center\">19.94</td><td align=\"center\">21.5</td></tr><tr><td align=\"left\">Fat (g)</td><td align=\"center\">2.66</td><td align=\"center\">3.45</td></tr><tr><td align=\"left\">Energy (kj)</td><td align=\"center\">111</td><td align=\"center\">485</td></tr></tbody></table></table-wrap>" ]

[ "<disp-formula>δX = [(R_sample/R_standard) - 1] * 1000,</disp-formula>" ]

[]

[]

[]

[]

[ "<supplementary-material content-type=\"local-data\" id=\"S1\"><caption><title>Additional file 1</title><p>Prey items identified in the faeces of wolves of coastal British Columbia.</p></caption></supplementary-material>", "<supplementary-material content-type=\"local-data\" id=\"S2\"><caption><title>Additional file 2</title><p>Top model sets to predict the use of salmon by wolves.</p></caption></supplementary-material>" ]

[ "<table-wrap-foot><p>Data from raw muscle tissue. Source: United States Department of Agriculture Nutrient Database <ext-link ext-link-type=\"uri\" xlink:href=\"http://www.nal.usda.gov/fnic/foodcomp/search//\"/>.</p></table-wrap-foot>" ]

[ "<graphic xlink:href=\"1472-6785-8-14-1\"/>", "<graphic xlink:href=\"1472-6785-8-14-2\"/>", "<graphic xlink:href=\"1472-6785-8-14-3\"/>", "<graphic xlink:href=\"1472-6785-8-14-4\"/>", "<graphic xlink:href=\"1472-6785-8-14-5\"/>", "<graphic xlink:href=\"1472-6785-8-14-6\"/>" ]

[ "<media xlink:href=\"1472-6785-8-14-S1.doc\" mimetype=\"application\" mime-subtype=\"msword\"><caption><p>Click here for file</p></caption></media>", "<media xlink:href=\"1472-6785-8-14-S2.doc\" mimetype=\"application\" mime-subtype=\"msword\"><caption><p>Click here for file</p></caption></media>" ]

{ "acronym": [], "definition": [] }

[ "<title>Background</title>", "<p>Maternal factors such as age and parity are known to influence birth outcomes. Thus advanced maternal age is associated with preterm birth [##REF##17289684##1##, ####REF##15932837##2##, ##REF##17239177##3####17239177##3##], fetal loss and stillbirth [##REF##10864550##4##, ####REF##17937740##5##, ##REF##16291475##6####16291475##6##], pregnancy complications [##REF##17289684##1##], higher risk of perinatal mortality and low birthweight [##REF##10405851##7##]. Higher risks of adverse outcomes are reported for both primiparous [##REF##17289684##1##,##REF##17118520##8##] and multiparous women of advanced maternal age (≥ 35 years) [##REF##17289684##1##]. Birthweight and gestational age are, in turn, important predictors of perinatal and infant mortality [##REF##17015548##9##,##REF##3880598##10##], childhood morbidity and disability [##REF##2378516##11##,##UREF##0##12##], and also health in later life [##REF##15640511##13##,##REF##15911642##14##]. The mutual interplay of the range of risk factors is complex and not yet fully understood.</p>", "<p>While gestational age has been acknowledged as a major determinant of birthweight, it has not been collected as part of routine vital perinatal statistics in many countries, for example the UK [##UREF##1##15##]. Even when it has been included, it has been criticised for being inaccurate, in particular for singleton preterm births [##REF##9690268##16##,##REF##7406096##17##]. There is, therefore, a lack of information on long-term trends in gestational age alongside birthweight, making it impossible to meaningfully interpret temporal changes in birthweight. Other essential covariate information such as parity, mode of delivery and paternal and maternal occupation are also not routinely collected in the UK as part of national data.</p>", "<p>The UK <bold>Pa</bold>rticulate <bold>M</bold>atter and <bold>P</bold>erinatal <bold>E</bold>vents <bold>R</bold>esearch (PAMPER) study offers the unique opportunity to describe temporal changes in key maternal and fetal factors affecting birth outcomes in a single conurbation over three decades, from 1961 to 1992. More specifically, we describe trends in maternal age, parity, aggregate level socioeconomic status, birthweight and gestational age and also demonstrate a reduction in stillbirth and infant mortality by decade.</p>" ]

[ "<title>Methods</title>", "<title>Study setting</title>", "<p>Newcastle upon Tyne, located within the Northern Region of England, has a current population of approximately 260,000 inhabitants. The population structure of the Northern Region is characterised by the low percentage of ethnic minorities, about 2% [##UREF##2##18##], and its relative stability with low levels of in and out migration. For example, among nearly 5,000 children aged between 1 and 11 years recruited into a study from 1996 to 1997, over 85% had lived at their address for most of their lives [##REF##9485483##19##]. Residential mobility in pregnancy is also low: only 9% of cases notified to the population-based Northern Congenital Abnormality Survey (NorCAS) [##UREF##3##20##,##REF##16167937##21##] moved from the time of booking to delivery (Rankin J, personal communication).</p>", "<p>During the 50 years following the end of the Second World War, the economy of Newcastle transformed from one dominated by heavy industry and coal production and trade to a service based economy by the early 1990s. This paralleled remarkable changes in societal factors; for example, the 1967 Abortion Act, the National Health Service (Family Planning) Act (1967), availability of free family planning services irrespective of age or marital status from April 1974, the Sex Discrimination Act (1975) and the Employment Protection Act (1975) were introduced during the study period.</p>", "<title>PAMPER birth population</title>", "<p>The PAMPER database contains birth details on all singletons born during 1961–92 to mothers resident within the city of Newcastle upon Tyne in Northern England. Information on multiple births was also collected, however it was excluded from these analyses as multiplicity is a known risk factor for the outcomes of interest of the PAMPER study, i.e. preterm birth and low birthweight. The boundaries of the PAMPER study area are shown in Figure ##FIG##0##1## with the river Tyne forming the southern boundary of the study area. The PAMPER computer database of birth records was constructed using information from a number of sources (Figure ##FIG##1##2##). The primary source was paper-based neonatal records from the two major maternity hospitals at the time (Princess Mary Maternity, PMMH, and Newcastle General Hospitals, NGH). From the PMMH, delivery and neonatal records were available for the whole study period, 1961–92; the NGH records were available from May 1967 onwards. These neonatal records contained information on important maternal and fetal/infant characteristics and clinical information about the delivery (Table ##TAB##0##1##). Socioeconomic information included paternal and maternal occupation, marital status and housing tenure.</p>", "<p>To capture home births, we additionally abstracted data from 'birth ledgers' (1961–1973), containing limited information on all births (Table ##TAB##0##1##). This data allowed us to obtain complete denominator information and to consider the changing proportion of home births (Figure ##FIG##2##3A##). We also used NGH birth records stored in the Tyne &amp; Wear Archives (available from 24th April 1961), to complement information on key variables unavailable in birth ledgers (Table ##TAB##0##1##).</p>", "<p>Each birth was georeferenced by postcode and/or grid reference. For births between 1961 and 1970 (prior to the introduction of postcodes), the address at birth was assigned a postcode from the 1991 postcode book or a grid reference. This allowed us to locate enumeration district (ED) of mothers' place of residence and hence to obtain the Townsend Deprivation Score (TDS), an area-based measure of material deprivation [##UREF##4##22##], at ED level (about 450 people in 200 households). TDS were calculated based on the 1971 (birth years 1961 to 1976), 1981 (1977 to 1986) and 1991 (1987 to 1992) Census data on unemployment, car ownership, owner occupation and overcrowding.</p>", "<title>Stillbirths and infant deaths</title>", "<p>We linked the dataset to information on stillbirths and infant deaths (including causes of death) from the Office for National Statistics (ONS) and to death data from the Northern Perinatal Mortality Survey (PMS) (available from 1981 onwards) [##REF##6428512##23##]. Multiple births were retained in the PAMPER database for the linkage procedure, but subsequently removed from the singletons database. Among a total of 1,248 eligible stillbirths provided by the ONS, we were able to match 1,222 cases (98%) to the PAMPER database. Among the total of 1,532 eligible ONS infant deaths, 1,510 (99%) were matched to the PAMPER database.</p>", "<p>As Gosforth in the north and some western residential parts of the PAMPER study areas were not part of the city of Newcastle upon Tyne prior to 1974, the ONS could not provide us with all stillbirths and infant deaths for these areas for this earlier period. However, we obtained death certificates and causes of stillbirth for cases known to us to be stillbirths and infant deaths. This may still have resulted in some missing infant deaths if a postneonatal death was not recorded in the hospital notes.</p>", "<title>PAMPER database completeness and accuracy</title>", "<p>Data entry staff (twelve individuals working 3-hour shifts) were trained in the medical terms/abbreviations used in the neonatal records and thus the percentage of errors was minimised. SVG was responsible for completing a descriptive 'summary' field, which contained the medical diagnosis and causes of death. In addition to the ONS and PMS data, stillbirth and infant death data were validated using birth record sources mentioned above.</p>", "<p>At the initial stage of data entry, we double entered approximately 1% of the estimated total of 120,000 birth records for different decades of the study period (n = 1,474) to assess accuracy of the data entry results. At the final stage of data entry, the data were validated by checking for implausible values (e.g. implausible difference between date of discharge and date of birth, implausible birthweight by gestation combinations).</p>", "<p>Table ##TAB##1##2## shows that data derived from hospital records (97,809) had low percentage of missing values for the key variables. Table ##TAB##1##2## also gives the number of births and percentages of maternal age, parity, birthweight, gestational age and mode of delivery categories by decade.</p>", "<p>For data capture we used the 4D database software suitable for a simultaneous data entry by several people, for data manipulation we used Microsoft Office Access 2003.</p>", "<title>Definitions</title>", "<p><italic>Stillbirths </italic>included were all babies born dead at 28 or more completed weeks of gestation. There were 12 cases (1%) recorded as stillbirths by the ONS with uncertain gestational age which were also included. Stillbirths with birthweight less than 500 g were excluded if gestational age was unknown. <italic>Infant death </italic>was defined as a death, following live birth, of an infant under one year of age. We defined <italic>preterm birth </italic>as birth at a gestational age less than 37 completed weeks and <italic>term birth </italic>as birth at a gestational age ≥ 37 weeks.</p>", "<title>Data analysis</title>", "<p>For descriptive statistical analysis we used the statistical software package SPSS for Windows, version 14.0. We used chi-square tests to test differences in proportions and independent-sample t-tests for comparison of means.</p>", "<title>Ethical approval</title>", "<p>The study received a favourable ethical opinion from the Sunderland Local Research Ethics Committee (SLREC 1071).</p>" ]

[ "<title>Results</title>", "<p>The number of births was highest in the early 1960s, followed by a steady decline until the mid 1970s and a further increase in the 1980s (Figure ##FIG##2##3A##). Home births constituted about a third of all births in the early 1960s, their proportion reduced to less than 0.5% by 1973 and remained low until the end of the study period (Figure ##FIG##2##3A##).</p>", "<p>Figure ##FIG##2##3B## shows that the trends in the number of hospital births from the PAMPER data were in line with regional trends.</p>", "<p>There was a dramatic decline in both stillbirth and infant mortality over the three decades (Table ##TAB##1##2##).</p>", "<p>Between 1961 and 1992 the average family size decreased, mainly due to a decline in the proportion of families with ≥ 3 children (Table ##TAB##1##2##).</p>", "<p>We considered mean maternal age by year in all and primiparous women (Figure ##FIG##3##4A##) and the percentages of teenage (≤ 19 years) and older (≥ 35 years) mothers over time (Figure ##FIG##3##4B##) alongside a chronology of key legislative changes, which may have contributed to the observed temporal changes. The lowest mean maternal age corresponded to a peak in the proportion of teenage mothers in 1973. The proportion of older mothers declined until the late 1970s (from 16.5% to 3.4%) but this was followed by a steady increase.</p>", "<p>Mean birthweight was lowest in the early 1960s, averaging around 3267 g in the second decade, followed by a gradual increase during the second half of the study period (Figure ##FIG##4##5A## and Table ##TAB##1##2##). The increase in mean birthweight for term births mostly accounted for the overall increase in mean birthweight, in particular in the last decade (Figure ##FIG##4##5A##). Thus during 1981–92 the mean birthweight at term [3373 g (SD ± 472)] was significantly higher than during the first two decades [3333 g (SD ± 497) in 1961–70 and 3332 g (SD ± 465) in 1971–80, <italic>p </italic>&lt; 0.001], whereas the mean birthweight in preterm births did not change in 1981–92 [2309 g (SD ± 664)] compared to 1971–80 [2308 g (SD ± 683)] in contrast to the first decade [2170 g (SD ± 732), <italic>p </italic>&lt; 0.001].</p>", "<p>The proportion of preterm births declined from 7% in 1961–70 to 6% in 1971–80 (Figure ##FIG##4##5B## and Table ##TAB##1##2##), but it increased again to 7% in 1981–92. In the last decade mean birthweight in all births increased despite the parallel increase in the percentage of preterm births. There was a two-fold increase in the percentage of caesarean section among preterm births from the early 1970s to the early 1990s, which partly accounted for this increase (Figure ##FIG##4##5B##).</p>", "<p>Table ##TAB##1##2## demonstrates that the gap between the most affluent and the most deprived groups of the population widened over the three decades.</p>" ]

[ "<title>Discussion</title>", "<p>Our study using population-based birth data in a single conurbation over three decades reported that between 1961 and 1992, when stillbirth and infant mortality rates declined dramatically, maternal age, parity, birthweight and gestational age changed substantially.</p>", "<title>Comparison with other studies</title>", "<p>National trends on the total fertility rates for 1960–1990 mirror temporal trends shown in our study, where we used parity as a measure of fertility; during the 1960s 'baby boom', the national total fertility rates peaked in 1964 followed by a subsequent decline with a lowest level in the mid 1970s and a slight increase afterwards [##REF##11618378##24##]. It has been suggested that the reduction in total fertility is attributable to improved means of fertility control (1967 Abortion Act and improved contraception efficacy) between 1967–68 and 1975. We also believe that the National Health Service (Family Planning) Act (1967), availability of free family planning services irrespective of age or marital status from April 1974, the Equal Pay Act 1970, the Sex Discrimination Act 1975 and the Employment Protection Act 1975, all contributed to women's reproductive decisions. This resulted in a decline in the proportion of teenage mothers and a parallel increase in the proportion of older mothers after the mid 1970s, as well as the overall increase in the mean maternal age in all and primiparous women. The increase in maternal age from the early 1980s was reported locally [##REF##16291475##6##], nationally [##REF##10909103##25##,##REF##15704383##26##], in Europe [##REF##9447350##27##] and in the United States [##REF##10585972##28##,##REF##10728230##29##]. Our data show that the mean maternal age in all and primiparous women was U-shaped with a declining trend from 1961 to the mid 1970s followed by a steady increase, repeating the national trend [##REF##11618378##24##]. As advanced maternal age is associated with a higher risk of preterm birth and low birthweight [##REF##15932837##2##,##REF##17118520##8##], its rise from the mid 1970s reported here may have contributed to the observed increase in the percentage of preterm birth in the last decade. Thus a study suggested that delayed childbearing may play an increasingly important role in low-birthweight trends in the United States [##REF##16571716##30##].</p>", "<p>We report a steady increase in the overall mean birthweight starting from the mid 1970s, which we observed for term births only and despite the increase in the proportion of preterm births in the second half of the study period. Hence, the observed rise in the total mean birthweight is likely to reflect the increase in birthweight for gestational age for term infants. This was also observed in Norway, where an increasing trend was reported for term births for 1967–1998 [##REF##10857867##31##], but not preterm (22–32 weeks) which were heavier in the first decade compared to the last, in contrast to our findings. Similar trends were also observed in Canada from 1981 to 1997 where the increase in mean birthweight was restricted to term infants [##REF##14629316##32##]. A study based on the Northern Region of England population, with Newcastle as part of this population, reported that the increasing trend in higher birthweights continued in the 1990s [##REF##16291475##6##]. An increase in mean birthweight has been also observed in other parts of England [##REF##10448184##33##], nationally [##REF##2041000##34##] and in other Western countries [##REF##10857867##31##,##REF##14629316##32##,##REF##12780422##35##].</p>", "<p>The proportion of preterm births declined in the second decade compared to the first, but it was followed by a steeper increase in 1981–92. To our knowledge, there are no population-based studies from the UK for comparison. Studies from other countries also reported the increase in the percentage of preterm birth from the 1980s [##REF##10857867##31##,##REF##9811918##36##,##REF##11856451##37##]. Several factors may have contributed to this increase. Thus there was a two-fold increase in the percentage of caesarean section among preterm births from the early 1970s to the 1990s, as with advances in neonatal technology, survival of extremely preterm infants dramatically increased, which justified interventions for fetal or maternal indications at earlier gestational ages. Similarly, in Norway the increase in the percentage of preterm births was attributable to a dramatic increase in the percentage of caesarean section among births delivered between 28 and 35 weeks in the late 1980s-1990s compared to the 1960s-1970s [##REF##10857867##31##]. The increase in births to older mothers, which are associated with a higher risk of preterm birth and a higher percentage of caesarean section due to a higher rate of complications of pregnancy, may also have contributed to this increase. Another factor may be a wider use of assisted reproductive technology in the UK from the late 1980s [##REF##8081085##38##], which is associated with a higher risk of preterm birth in singletons [##REF##15745640##39##,##REF##14742347##40##] and is more widely used among older women.</p>", "<p>Townsend deprivations scores, which we calculated for each birth in the database to measure neighbourhood socioeconomic status, also changed over time: the scores seemed to improve for the most affluent quintile and deteriorate for the most deprived, thus making the gap between the affluent and deprived groups wider. This is in line with the widening socio-economic and health inequalities which are now well documented in the UK.</p>", "<title>Strengths and limitations of the PAMPER birth record database</title>", "<p>The population-based PAMPER birth record database contains historical high-quality birth data in a defined compact geographical setting over a 32 year period during which there have been significant changes in obstetric and neonatal services. The completeness of the PAMPER database both for the number of births and information collected for each birth is a major strength. National and local trends in the number of births in the UK confirm the temporal fluctuations also observed in the PAMPER study: the highest number of births at the beginning of the 1960s (a so-called 'baby boom'), followed by a decline in the 1970s and a further increase in the number of births during the 1980s [##UREF##5##41##]. The completeness of the data for the key variables described here is expressed in the low percentages of missing data for these variables.</p>", "<p>The availability of accurate population-based gestational age, a major determinant of birthweight, is one of the leading strengths of the PAMPER database, as gestational age was not available in national birth statistics during the study period. Further, birthweight for live births was not collected in the UK at national level until 1975 (as part of the Child Health Births Notifications System). Without gestational age, interpretation of trends in birthweight could be misleading, as it is not possible to disentangle whether changes in birthweight are attributable to changes in rates of preterm birth or to changes in actual fetal growth. However, in the UK and elsewhere in the world there is a lack of information on the incidence of premature birth using accurate data by gestation [##UREF##1##15##].</p>", "<p>The accuracy of the data for the key variables was ensured by multiple checking, internal (within the database) and external (with national and regional death data, and other local sources of birth record data) validation of the data.</p>", "<p>The PAMPER database also has several limitations. The lack of information on some important determinants of fetal weight at birth such as maternal height, maternal smoking and exposure to environmental tobacco smoke, which have changed over time thereby affecting changes in birthweight, is disappointing. For example, an increasing trend in maternal height was reported in Scotland for 1980–2000 [##REF##16115285##42##]. In the UK, the prevalence of smoking in women increased sharply during and after the Second World War, reaching the level of about 42–44% in the 1960s – early 1970s [##UREF##6##43##,##UREF##7##44##] followed by a gradual decrease thereafter [##UREF##7##44##]. However, adjustment for year of birth should be able to control for the effect of temporal changes in any factors influencing birth outcomes.</p>", "<p>The accuracy of gestational age estimates is important for epidemiologic studies of pregnancy outcomes. Different methods for gestational age assessment (based on the last normal menstrual period (LMP) or early ultrasound measurements) throughout the study period may introduce bias in gestational age estimation over time. Thus it has been suggested that higher rates of preterm birth may be reported if determination of gestational age is based on ultrasonographic dating alone [##REF##8562631##45##,##REF##15663577##46##]. In the 1960s and 1970s, when gestational age estimate was based on LMP and, if the dates were uncertain, on the paediatric examination of the baby, it may have more uncertainty. However, while creating our birth record database, we made the recording of gestational age as objective and accurate as possible by accepting gestational age calculated from the recorded estimated date of delivery (EDD) (i.e. LMP based) for the majority of births rather than by entering gestational age recorded in the neonatal notes or birth records. For example, the percentage of gestational age records based on the recorded EDD for 1961–70 was about 87% of records with known gestational age. In this study the ultrasound age estimate has been used since the early 1980s only for pregnancies with uncertain date of LMP or if there was a significant discrepancy between the two estimates, therefore it should not bias gestational age estimates over time. Moreover, gestational age seems to be accurate in our study as birthweight distribution at early gestational ages has a single mode in contrast to other studies reporting bimodal birthweight distributions at early gestations with implausible high birthweights for gestational age [##REF##9690268##16##,##REF##7406096##17##].</p>" ]

[ "<title>Conclusion</title>", "<p>This historical population-based study documents substantial temporal changes in key maternal and fetal factors affecting birth outcomes over a 32-year period during which much social change has taken place. The availability of accurate gestational age is extremely important for correct interpretation of trends in birthweight.</p>" ]

[ "<p>This is an Open Access article distributed under the terms of the Creative Commons Attribution License (<ext-link ext-link-type=\"uri\" xlink:href=\"http://creativecommons.org/licenses/by/2.0\"/>), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.</p>", "<title>Background</title>", "<p>The link between maternal factors and birth outcomes is well established. Substantial changes in society and medical care over time have influenced women's reproductive choices and health, subsequently affecting birth outcomes. The objective of this study was to describe temporal changes in key maternal and fetal factors affecting birth outcomes in Newcastle upon Tyne over three decades, 1961–1992.</p>", "<title>Methods</title>", "<p>For these descriptive analyses we used data from a population-based birth record database constructed for the historical cohort <bold>Pa</bold>rticulate <bold>M</bold>atter and <bold>P</bold>erinatal <bold>E</bold>vents <bold>R</bold>esearch (PAMPER) study. The PAMPER database was created using details from paper-based hospital delivery and neonatal records for all births during 1961–1992 to mothers resident in Newcastle (out of a total of 109,086 singleton births, 97,809 hospital births with relevant information). In addition to hospital records, we used other sources for data collection on births not included in the delivery and neonatal records, for death and stillbirth registrations and for validation.</p>", "<title>Results</title>", "<p>The average family size decreased mainly due to a decline in the proportion of families with 3 or more children. The distribution of mean maternal ages in all and in primiparous women was lowest in the mid 1970s, corresponding to a peak in the proportion of teenage mothers. The proportion of older mothers declined until the late 1970s (from 16.5% to 3.4%) followed by a steady increase. Mean birthweight in all and term babies gradually increased from the mid 1970s. The increase in the percentage of preterm birth paralleled a two-fold increase in the percentage of caesarean section among preterm births during the last two decades. The gap between the most affluent and the most deprived groups of the population widened over the three decades.</p>", "<title>Conclusion</title>", "<p>Key maternal and fetal factors affecting birth outcomes, such as maternal age, parity, socioeconomic status, birthweight and gestational age, changed substantially during the 32-year period, from 1961 to 1992. The availability of accurate gestational age is extremely important for correct interpretation of trends in birthweight.</p>" ]

[ "<title>Competing interests</title>", "<p>The authors declare that they have no competing interests.</p>", "<title>Authors' contributions</title>", "<p>SG carried out the statistical analysis and drafted the paper. All authors were co-investigators on the Wellcome Trust grant, contributed to the initiation of the project and study design, and commented on the drafts of the paper. All authors have read and approved the final version of the manuscript.</p>", "<title>Pre-publication history</title>", "<p>The pre-publication history for this paper can be accessed here:</p>", "<p><ext-link ext-link-type=\"uri\" xlink:href=\"http://www.biomedcentral.com/1471-2393/8/39/prepub\"/></p>" ]

[ "<title>Acknowledgements</title>", "<p>The study was supported by the Wellcome Trust, grant No 072465/Z/03/Z. JR is funded by a Personal Award Scheme Career Scientist Award from the National Institute of Health Research (UK Department of Health). We would like to express our gratitude to the data entry staff for their hard work and to Mr Richard Hardy, our PAMPER database manager. We are grateful to staff at the ONS and the local Register Offices (Newcastle, Gateshead, North Tyneside). We also thank the very helpful staff of the Tyne &amp; Wear Archives Service and Mrs Marjorie Renwick, data manager at the Northern Region Maternity Survey Office.</p>" ]

[ "<fig position=\"float\" id=\"F1\"><label>Figure 1</label><caption><p><bold>Map of Newcastle upon Tyne with the PAMPER study area boundaries (black line)</bold> (^©Crown Copyright/database right 2007. An Ordnance Survey/EDINA supplied service).</p></caption></fig>", "<fig position=\"float\" id=\"F2\"><label>Figure 2</label><caption><p><bold>Data sources used to construct the PAMPER dataset.</bold> footnote: NGH = Newcastle General Hospital; PMMH = Princess Mary Maternity Hospital; ONS = Office for National Statistics; PMS = Northern Perinatal Mortality Survey; NorCAS = Northern Congenital Abnormality Survey.</p></caption></fig>", "<fig position=\"float\" id=\"F3\"><label>Figure 3</label><caption><p><bold>(A) Number of births and percentage of home births by year of delivery (PAMPER dataset 1961–92) and (B) Number of hospital births by year of delivery in the PAMPER dataset and all births from two hospitals based on the Northern Region Health Authority data, 1961–92.</bold> footnote: home births are recorded from both birth ledgers and hospital records for 1961–73 and from hospital records only thereafter; NGH = Newcastle General Hospital, PMMH = Princess Mary Maternity Hospital.</p></caption></fig>", "<fig position=\"float\" id=\"F4\"><label>Figure 4</label><caption><p>(A) Mean maternal age in all and primiparous women and (B) percentage of teenage (&lt; 20 years) and older (≥ 35 years) mothers during 1961–1992.</p></caption></fig>", "<fig position=\"float\" id=\"F5\"><label>Figure 5</label><caption><p>(A) Mean birthweight in term (≥ 37 weeks), preterm (&lt;37 weeks) and all births by year of birth; (B) Percentages of preterm birth and caesarean section (CS) among preterm births by year of birth and respective 3-year moving averages of the percentage.</p></caption></fig>" ]

[ "<table-wrap position=\"float\" id=\"T1\"><label>Table 1</label><caption><p>Key variables available across different data sources used for the construction of the PAMPER database</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td/><td align=\"center\"><bold>NGH and PMMH neonatal records</bold></td><td align=\"center\"><bold>Tyne &amp; Wear Archives birth records</bold></td><td align=\"center\"><bold>Birth ledgers</bold></td></tr></thead><tbody><tr><td align=\"left\">Mother's current surname</td><td align=\"center\">√</td><td align=\"center\">√</td><td align=\"center\">√</td></tr><tr><td align=\"left\">Residential address</td><td align=\"center\">√</td><td align=\"center\">√</td><td align=\"center\">√</td></tr><tr><td align=\"left\">Baby's sex</td><td align=\"center\">√</td><td align=\"center\">√</td><td align=\"center\">√</td></tr><tr><td align=\"left\">Date of birth</td><td align=\"center\">√</td><td align=\"center\">√</td><td align=\"center\">√</td></tr><tr><td align=\"left\">Vital status at birth</td><td align=\"center\">√</td><td align=\"center\">√</td><td align=\"center\">√</td></tr><tr><td align=\"left\">Place of birth</td><td align=\"center\">√</td><td align=\"center\">√</td><td align=\"center\">√</td></tr><tr><td align=\"left\">Plurality</td><td align=\"center\">√</td><td align=\"center\">√</td><td align=\"center\">√</td></tr><tr><td align=\"left\">Birthweight</td><td align=\"center\">√</td><td align=\"center\">√</td><td align=\"center\">-</td></tr><tr><td align=\"left\">Gestational age</td><td align=\"center\">√</td><td align=\"center\">√</td><td align=\"center\">-</td></tr><tr><td align=\"left\">Maternal age</td><td align=\"center\">√</td><td align=\"center\">√</td><td align=\"center\">-</td></tr><tr><td align=\"left\">Parity</td><td align=\"center\">√</td><td align=\"center\">√</td><td align=\"center\">-</td></tr><tr><td align=\"left\">Mode of delivery</td><td align=\"center\">√</td><td align=\"center\">√</td><td align=\"center\">-</td></tr><tr><td align=\"left\">Baby's surname</td><td align=\"center\">√</td><td align=\"center\">-</td><td align=\"center\">-</td></tr><tr><td align=\"left\">Paternal occupation</td><td align=\"center\">√</td><td align=\"center\">-</td><td align=\"center\">-</td></tr><tr><td align=\"left\">Maternal occupation</td><td align=\"center\">√ (for 1976–92)</td><td align=\"center\">-</td><td align=\"center\">-</td></tr><tr><td align=\"left\">Admission to Special Care Baby Unit</td><td align=\"center\">√</td><td align=\"center\">-</td><td align=\"center\">-</td></tr><tr><td align=\"left\">Resuscitation</td><td align=\"center\">√</td><td align=\"center\">-</td><td align=\"center\">-</td></tr><tr><td align=\"left\">Early mortality data with cause of death</td><td align=\"center\">√</td><td align=\"center\">-</td><td align=\"center\">-</td></tr><tr><td align=\"left\">Hospital morbidity data</td><td align=\"center\">√</td><td align=\"center\">-</td><td align=\"center\">-</td></tr></tbody></table></table-wrap>", "<table-wrap position=\"float\" id=\"T2\"><label>Table 2</label><caption><p>Basic description of the PAMPER birth population 1961–92</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td align=\"left\"><bold>Variable</bold></td><td align=\"center\"><bold>1961–70</bold></td><td align=\"center\"><bold>1971–80</bold></td><td align=\"center\"><bold>1981–92</bold></td><td align=\"center\"><bold>N missing (%) </bold></td></tr><tr><td/><td/><td/><td/><td align=\"center\"><bold>1961–92</bold></td></tr></thead><tbody><tr><td align=\"left\"><bold>Maternal age (years)</bold></td><td/><td/><td/><td align=\"center\">1796 (1.8)</td></tr><tr><td align=\"left\">Mean (± SD)</td><td align=\"center\">26.4 (6.2)</td><td align=\"center\">24.9 (5.2)</td><td align=\"center\">25.8 (5.3)</td><td/></tr><tr><td align=\"left\">≤ 19 [n (%)]</td><td align=\"center\">3037 (11.6)</td><td align=\"center\">4510 (15.1)</td><td align=\"center\">4913 (12.3)</td><td/></tr><tr><td align=\"left\">20–24 [n (%)]</td><td align=\"center\">8694 (33.2)</td><td align=\"center\">10676 (35.6)</td><td align=\"center\">12075 (30.3)</td><td/></tr><tr><td align=\"left\">25–29 [n (%)]</td><td align=\"center\">6772 (25.9)</td><td align=\"center\">9420 (31.4)</td><td align=\"center\">13123 (32.9)</td><td/></tr><tr><td align=\"left\">30–34 [n (%)]</td><td align=\"center\">4386 (16.8)</td><td align=\"center\">3917 (13.1)</td><td align=\"center\">7290 (18.3)</td><td/></tr><tr><td align=\"left\">35–40 [n (%)]</td><td align=\"center\">2426 (9.3)</td><td align=\"center\">1149 (3.8)</td><td align=\"center\">2175 (5.5)</td><td/></tr><tr><td align=\"left\">40–44 [n (%)]</td><td align=\"center\">807 (3.1)</td><td align=\"center\">259 (0.9)</td><td align=\"center\">287 (0.7)</td><td/></tr><tr><td align=\"left\">45+ [n (%)]</td><td align=\"center\">52 (0.2)</td><td align=\"center\">25 (0.1)</td><td align=\"center\">20 (0.1)</td><td/></tr><tr><td colspan=\"5\"><hr/></td></tr><tr><td align=\"left\"><bold>Parity </bold>[n (%)]</td><td/><td/><td/><td align=\"center\">1290 (1.3)</td></tr><tr><td align=\"left\">Parity = 0 (primipara)</td><td align=\"center\">10753 (41.1)</td><td align=\"center\">13659 (45.4)</td><td align=\"center\">17888 (44.4)</td><td/></tr><tr><td align=\"left\">Parity = 1</td><td align=\"center\">5673 (21.7)</td><td align=\"center\">9798 (32.6)</td><td align=\"center\">13209 (32.8)</td><td/></tr><tr><td align=\"left\">Parity = 2</td><td align=\"center\">3803 (14.5)</td><td align=\"center\">4121 (13.7)</td><td align=\"center\">5794 (14.4)</td><td/></tr><tr><td align=\"left\">Parity = 3</td><td align=\"center\">2101 (8.0)</td><td align=\"center\">1478 (4.9)</td><td align=\"center\">2114 (5.2)</td><td/></tr><tr><td align=\"left\">Parity = 4</td><td align=\"center\">1458 (5.6)</td><td align=\"center\">589 (2.0)</td><td align=\"center\">799 (2.0)</td><td/></tr><tr><td align=\"left\">Parity = 5</td><td align=\"center\">1004 (3.8)</td><td align=\"center\">225 (0.7)</td><td align=\"center\">278 (0.7)</td><td/></tr><tr><td align=\"left\">Parity = 6+</td><td align=\"center\">1383 (5.3)</td><td align=\"center\">201 (0.7)</td><td align=\"center\">191 (0.5)</td><td/></tr><tr><td colspan=\"5\"><hr/></td></tr><tr><td align=\"left\"><bold>Baby's birthweight (g)</bold></td><td/><td/><td/><td align=\"center\">1360 (1.4)</td></tr><tr><td align=\"left\">Mean (± SD)</td><td align=\"center\">3244.2 (603.7)</td><td align=\"center\">3266.5 (540.3)</td><td align=\"center\">3297.0 (558.0)</td><td/></tr><tr><td align=\"left\">&lt;1000 [n (%)]</td><td align=\"center\">150 (0.6)</td><td align=\"center\">75 (0.2)</td><td align=\"center\">128 (0.3)</td><td/></tr><tr><td align=\"left\">1000–1499 [n (%)]</td><td align=\"center\">270 (1.0)</td><td align=\"center\">181 (0.6)</td><td align=\"center\">238 (0.6)</td><td/></tr><tr><td align=\"left\">1500–1999 [n (%)]</td><td align=\"center\">481 (1.8)</td><td align=\"center\">354 (1.2)</td><td align=\"center\">496 (1.2)</td><td/></tr><tr><td align=\"left\">2000–2499 [n (%)]</td><td align=\"center\">1452 (5.6)</td><td align=\"center\">1445 (4.8)</td><td align=\"center\">1790 (4.4)</td><td/></tr><tr><td align=\"left\">2500–2999 [n (%)]</td><td align=\"center\">5132 (19.6)</td><td align=\"center\">6033 (20.1)</td><td align=\"center\">7504 (18.6)</td><td/></tr><tr><td align=\"left\">3000–3499 [n (%)]</td><td align=\"center\">9920 (37.9)</td><td align=\"center\">12175 (40.5)</td><td align=\"center\">15722 (39.1)</td><td/></tr><tr><td align=\"left\">3500–3999 [n (%)]</td><td align=\"center\">6719 (25.7)</td><td align=\"center\">7642 (25.4)</td><td align=\"center\">10852 (27.0)</td><td/></tr><tr><td align=\"left\">4000–4499 [n (%)]</td><td align=\"center\">1714 (6.6)</td><td align=\"center\">1907 (6.3)</td><td align=\"center\">3065 (7.6)</td><td/></tr><tr><td align=\"left\">4500+ [n (%)]</td><td align=\"center\">310 (1.2)</td><td align=\"center\">235 (0.8)</td><td align=\"center\">459 (1.1)</td><td/></tr><tr><td colspan=\"5\"><hr/></td></tr><tr><td align=\"left\"><bold>Gestational age (weeks)</bold></td><td/><td/><td/><td align=\"center\">3562 (3.6)</td></tr><tr><td align=\"left\">Mean (± SD)</td><td align=\"center\">39.5 (2.2)</td><td align=\"center\">39.4 (1.9)</td><td align=\"center\">39.1 (2.0)</td><td/></tr><tr><td align=\"left\">&lt; 32 [n (%)]</td><td align=\"center\">288 (1.1)</td><td align=\"center\">224 (0.8)</td><td align=\"center\">387 (1.0)</td><td/></tr><tr><td align=\"left\">32–36 [n (%)]</td><td align=\"center\">1533 (6.0)</td><td align=\"center\">1557 (5.3)</td><td align=\"center\">2415 (6.1)</td><td/></tr><tr><td align=\"left\">37+ [n (%)]</td><td align=\"center\">23769 (92.9)</td><td align=\"center\">27351 (93.9)</td><td align=\"center\">36723 (92.9)</td><td/></tr><tr><td colspan=\"5\"><hr/></td></tr><tr><td align=\"left\"><bold>Infant gender</bold></td><td/><td/><td/><td align=\"center\">10 (0.01)</td></tr><tr><td align=\"left\">Male/Female ratio</td><td align=\"center\">1.08</td><td align=\"center\">1.06</td><td align=\"center\">1.07</td><td/></tr><tr><td colspan=\"5\"><hr/></td></tr><tr><td align=\"left\"><bold>Mode of delivery</bold></td><td/><td/><td/><td align=\"center\">1763 (1.8)</td></tr><tr><td align=\"left\">Normal vertex delivery [n (%)]</td><td align=\"center\">20000 (76.8)</td><td align=\"center\">22517 (75.1)</td><td align=\"center\">29182 (72.9)</td><td/></tr><tr><td align=\"left\">Assisted (forceps/vacuum extraction) [n (%)]</td><td align=\"center\">3675 (14.1)</td><td align=\"center\">4502 (15.0)</td><td align=\"center\">5635 (14.1)</td><td/></tr><tr><td align=\"left\">Caesarean section [n (%)]</td><td align=\"center\">1822 (7.0)</td><td align=\"center\">2484 (8.3)</td><td align=\"center\">4670 (11.7)</td><td/></tr><tr><td align=\"left\">Breech extraction [n (%)]</td><td align=\"center\">543 (2.1)</td><td align=\"center\">470 (1.6)</td><td align=\"center\">506 (1.3)</td><td/></tr><tr><td colspan=\"5\"><hr/></td></tr><tr><td align=\"left\"><bold>Stillbirth </bold>[n (rate per 1000)]</td><td align=\"center\">688 (18.2)</td><td align=\"center\">325 (10.5)</td><td align=\"center\">227 (5.6)</td><td/></tr><tr><td colspan=\"5\"><hr/></td></tr><tr><td align=\"left\"><bold>Infant mortality </bold>[n (rate per 1000)]</td><td align=\"center\">770 (20.8)</td><td align=\"center\">453 (14.7)</td><td align=\"center\">281 (7.0)</td><td/></tr><tr><td colspan=\"5\"><hr/></td></tr><tr><td align=\"left\"><bold>Quintiles of ED TDS</bold></td><td/><td/><td/><td align=\"center\">2095 (2.1)</td></tr><tr><td align=\"left\">1 (most affluent)</td><td align=\"center\">≤ 0.04</td><td align=\"center\">≤ -1.40</td><td align=\"center\">≤ -1.49</td><td/></tr><tr><td align=\"left\">2</td><td align=\"center\">&gt; 0.04 to ≤ 3.01</td><td align=\"center\">&gt;-1.40 to ≤ 2.43</td><td align=\"center\">&gt;-1.49 to ≤ 2.18</td><td/></tr><tr><td align=\"left\">3</td><td align=\"center\">&gt;3.01 to ≤ 4.9</td><td align=\"center\">&gt;2.43 to ≤ 4.56</td><td align=\"center\">&gt;2.18 to ≤ 5.06</td><td/></tr><tr><td align=\"left\">4</td><td align=\"center\">&gt;4.9 to ≤ 6.2</td><td align=\"center\">&gt;4.56 to ≤ 6.33</td><td align=\"center\">&gt;5.06 to ≤ 7.04</td><td/></tr><tr><td align=\"left\">5 (most deprived)</td><td align=\"center\">&gt;6.2</td><td align=\"center\">&gt; 6.33</td><td align=\"center\">&gt; 7.04</td><td/></tr><tr><td colspan=\"5\"><hr/></td></tr><tr><td align=\"left\"><bold>Address grid referenced</bold></td><td/><td/><td/><td align=\"center\">1110 (1.1)</td></tr><tr><td colspan=\"5\"><hr/></td></tr><tr><td align=\"left\"><bold>Paternal occupation (coded)</bold></td><td/><td/><td/><td align=\"center\">29419 (30.1)</td></tr></tbody></table></table-wrap>" ]

[]

[]

[]

[]

[]

[]

[ "<table-wrap-foot><p><bold>Note: </bold>The following additional data on maternal and child characteristics are available in the PAMPER database either for the whole study period or for shorter periods: time of birth, date of discharge, discharge weight, date of death (in case of infant deaths), maternal blood group, marital status, housing (for the 1960s), details of previous births, placental weight, onset of labour (spontaneous vs induced), Apgar score, type of feeding on discharge, estimated date of delivery.</p></table-wrap-foot>", "<table-wrap-foot><p><bold>Note: </bold>Number of births used for denominator = 109,086 (for calculation of stillbirth (per 1000 total births) and infant mortality rates (per 1000 live births); number of births from hospital records with information on covariates listed in the table = 97,809 (percentages of missing data are given using 97,809 as a total).</p><p>Percentages of the categories were calculated from the total with known data for a variable.</p><p>ED TDS = Townsend Deprivation Score at the enumeration district level.</p></table-wrap-foot>" ]

[ "<graphic xlink:href=\"1471-2393-8-39-1\"/>", "<graphic xlink:href=\"1471-2393-8-39-2\"/>", "<graphic xlink:href=\"1471-2393-8-39-3\"/>", "<graphic xlink:href=\"1471-2393-8-39-4\"/>", "<graphic xlink:href=\"1471-2393-8-39-5\"/>" ]

[]

{ "acronym": [], "definition": [] }

[ "<title>Background</title>", "<p>Breast cancer is still the most common public health problem in women worldwide [##REF##8347848##1##]. Cancer cells may invade the surrounding by tissue remodeling and angiogenesis [##UREF##0##2##, ####UREF##1##3##, ##UREF##2##4##, ##UREF##3##5##, ##UREF##4##6##, ##UREF##5##7####5##7##]. They may spread through the bloodstream and lymphatic system to other parts of the body [##UREF##0##2##, ####UREF##1##3##, ##UREF##2##4##, ##UREF##3##5##, ##UREF##4##6##, ##UREF##5##7####5##7##]. The majority of invasive breast cancer from the epithelium of lobules and ducts of the glands [##REF##5236189##8##, ####REF##1747227##9##, ##REF##3020723##10##, ##REF##7440180##11##, ##REF##10349394##12##, ##REF##3623942##13##, ##REF##10352436##14##, ##REF##3790289##15##, ##REF##3316887##16##, ##REF##194130##17##, ##REF##2823241##18##, ##REF##3744850##19##, ##REF##9775407##20##, ##REF##10211217##21####10211217##21##]. Metastasis is considered as the spreading of tumor cells from the primary neoplasm to distant sites [##UREF##6##22##,##UREF##7##23##]. In spite of significant advancement in early diagnosis, surgical intervention as well as local and systemic adjuvant therapies, the majority of cancer deaths are attributable to metastasis that are resistant to available therapies [##UREF##6##22##].</p>", "<p>The metastasis process is not a random process but consist of a complex series of linked and interrelated steps involving multiple host-tumor interaction [##REF##9673300##24##]. Many proteins including proteases, adhesion molecules, angiogenesis, and growth factor are involved in metastasis [##UREF##8##25##]. Therefore, understanding the gene and protein expression changes in metastatic cancer cells and nearby cells of the microenvironment may aid in early diagnosis and therapeutic intervention. During the last decade, considerable progress has been made in understanding these changes at the molecular level. In fact most deaths of women with breast cancer arise not as a result of primary tumor but from its metastatic spread to distant sites in the body [##UREF##9##26##,##UREF##10##27##]. Once spread and secondary masses are formed, breast cancers are usually incurable [##REF##12466733##28##]. Yet a sensitive and reliable method for early metastasis in breast cancer is still not available.</p>", "<p>Breast cancer is considered to be a systemic disease this would mean that the most breast carcinoma metastasize before diagnosis of the primary lesion [##UREF##11##29##]. Therefore, early detection of metastasized lesion and identification of more effective therapeutic modalities for metastatic disease are necessities if the prognosis for patients with advanced breast cancer is to improve.</p>", "<p>The S100 gene family located on chromosome 1q21, comprises more than 20 members whose protein sequences encompass at least one EF-hand Ca++ binding motif [##REF##8701470##30##,##REF##12756252##31##]. S100 is a 21 Kd highly acidic and water soluble calcium binding protein [##REF##6849880##32##]. S100A4 gene occurs in cluster of 13 S100 genes on chromosome 1 [##REF##8341667##33##], which are also often amplified in cancer of the breast and which contains jumping elements [##REF##1670997##34##]. The expression of individual family members is not permanent for all tissues and appears to be an element of tissue specific expression. S100A4 is composed of an alpha and beta chain with molecular weight of 10 – 12 Kd [##REF##8204608##35##]. It is a small molecule and can pass through the nuclear pores without any active transport mechanism being involved. S100A4 binds and inhibits phosphorylation of the p53 C-terminal peptide by protein kinase C. The tumor suppressor protein p53 has also been identified as an S100A4 interacting protein and may provide a link between S100A4 and apoptosis [##REF##14606958##36##]. p53 is a critical tumor suppressor that is involved in most if not all tumorigenesis. Almost 30–50% of breast cancers contain a p53 mutation [##REF##8632491##37##]. Other reported S100A4 interacting proteins include tropomyosin, methionine aminopeptidase, and CCN3 (cysteine-rich 61/connective tissue growth factor/nephroblastoma overexpressed) [##REF##8120097##38##, ####REF##11994292##39##, ##REF##12147716##40####12147716##40##].</p>", "<p>Studies to determine the mechanistic basis for S100A4 function have shown a potential role for S100A4 in several different facts of tumor progression including motility, invasion, and apoptosis [[##REF##14606958##36##,##REF##7928629##41##], and [##REF##11498791##42##]]. It has also been reported that extracelluar secreted S100A4 can affect cell differentiation and migration [##REF##11498791##42##, ####REF##11018041##43##, ##REF##12445462##44####12445462##44##]. Elevated levels of immunocytochemically detected S100A4 are associated with the more malignant carcinomatous regions of the primary tumor and with liver metastasis [##REF##9815629##45##]. An increase in S100A4 protein expression has been correlated with a worse prognosis for patients with different types of cancer including colorectal, gallbladder, bladder, esophageal, breast, and non small lung cancer [##REF##12404222##46##, ####REF##11956586##47##, ##REF##11857492##48##, ##REF##11251165##49##, ##REF##10754500##50##, ##REF##10749128##51##, ##REF##10811984##52####10811984##52##]. The main purpose of the study was to find the expression pattern of S100A4 proteins in different types of breast cancer with or without lymph node involvement and to determine its role in breast cancer using tissue microarray.</p>" ]

[ "<title>Methods</title>", "<p>Formalin-fixed paraffin-embedded tissue microarrays from 188 lymph node negative breast cancer patients, 50 breast cancer patients with lymph node metastasis (50 malignant tissues and 50 matched lymph node tissue cores with metastasis) and 8 normal breast tissue cores were analyzed by immunohistochemistry for the expression S100A4 protein. Included in this study were patients with infiltrating ductal carcinoma, infiltrating lobular carcinoma, normal breast tissue and lymph node metastasis. The final number was 122 tissue cores of node negative infiltrating ductal carcinoma, 41 node negative infiltrating lobular carcinoma, seven normal breast tissue, 40 node positive (38 infiltrating ductal carcinoma, 2 infiltrating lobular carcinoma) and 46 tissue cores of lymph node metastasis. We did not evaluate infiltrating lobular carcinoma node positive because of the low sample size (2 cases). Forty six tissue cores were excluded from statistics because very little cancer cells or no breast cancer tissue were seen. The total number after exclusion was 254 tissue cores (37 matched tissues) of 216 patients.</p>", "<title>Immunohistochemistry (IHC)</title>", "<p>Rabbit Anti-Human polyclonal primary antibody against S100A4 protein (Code No. A5114, from DakoCytomation, Denmark) was used on deparaffinized tissue microarray slides (Cat. No. BR 2001, BR 1001 from Biomax, USA). A secondary detection system (DAKO Envision) enhanced with conjugated polymer was used to bind with the primary antibody. DAB chromogen was used for permanent color development and detection under microscope.</p>", "<p>The percentage of carcinoma cells with cytoplasmic/membranous/nuclear staining was recorded on each specimen at 200× magnification, using light microscope. The expression of S100A4 was scored in all tumors as: positive ≥ 5% and negative &lt; 5% stained cells. Also the intensity of staining was categorized into three groups: weak, moderate and strong. This was ascertained by a single qualified pathologist.</p>", "<p>A tissue section of breast cancer was used as positive control for S100A4. Rabbit IgG isotype (Sigma-Aldrich, USA) was used instead of primary antibody in the immunohistochemical technique on a tissue section each of breast cancer and normal breast as negative control (Figure ##FIG##0##1##).</p>", "<p>The tissue microarray slides were placed on hot plate at 60°C for 30 minutes. The slides were immersed in two changes of xylene. Slides were then immersed in 3 different concentrations of ethanol. Slides were rinsed with distilled water to remove ethanol. The slides were then placed in target retrieval solution EDTA buffer pH 9.0 (DAKO) and heated on microwave. Slides were allowed to cool at room temperature and rinsed with Tris Buffered saline (TBS) mixed with tween 20. Slides were covered with peroxidase blocking solution (DAKO), followed by rinsing with TBS buffer mixed with tween 20. Then 200 μl of primary antibody (dilution 1:200) was added on the tissue microarray slides, followed by rinsing with TBS mixed with tween 20. Two drops of DAKO Envision/HRP, Rabbit/Mouse (secondary antibody) were added on the slides, followed by rinsing with TBS mixed with tween 20. After that DAB substrate (DAKO) was added on section slides followed by rinsing, immersion into hematoxylin and 4 different concentrations of ethanol. After that slides were immersed in two changes of xylene and cover slipped. All incubation steps after heat induced epitope retrieval were carried at room temperature.</p>" ]

[ "<title>Results</title>", "<p>The S100A4 protein was expressed in the cell cytoplasm without evidence of nuclear staining.</p>", "<p>There were a total of 122 cases of infiltrating ductal carcinoma node negative and 41 cases of infiltrating lobular carcinoma node negative. Infiltrating ductal carcinoma node positive consisted of 38 cases. Thirty seven cases had paired primary infiltrating ductal carcinoma tissue with its matched lymph node core. There were also 9 cases of unrelated lymph node cores containing metastatic deposits.</p>", "<p>A positive expression of S100A4 was observed in 45.1% (55/122) cases of infiltrating ductal carcinoma node negative (Figure ##FIG##1##2##) and 48.8% (20/41) cases of infiltrating lobular carcinoma node negative (Figure ##FIG##2##3##). S100A4 staining was not observed in normal breast tissues (Figure ##FIG##3##4##).</p>", "<p>Five of 37 (13.5%) cases in the paired samples (primary breast carcinoma and matched lymph nodes) showed presence of S100A4 protein in the primary site, while 13/37 (35.1%) cases showed S100A4 in lymph node (Figure ##FIG##4##5##).</p>", "<p>It was observed that 12 of 17 (70.5%) paired cases (primary breast cancer with matched lymph node) showed expression of S100A4 in the lymph node metastasis with absent expression in the primary site. Four (23.5%) cases showed S100A4 expression in the primary breast carcinoma with absent expression in its matched lymph node metastasis. One case showed S100A4 expression in both sites.</p>" ]

[ "<title>Discussion</title>", "<p>Cancer is a disease or disorder characterized by uncontrolled growth (division) of the cells [##UREF##12##53##, ####UREF##13##54##, ##UREF##14##55####14##55##] and their ability to invade other organ or tissue, either by invasion or metastasis [##UREF##13##54##, ####UREF##14##55##, ##UREF##15##56####15##56##]. Although there are many types of cancer, all cancer types begin with uncontrolled growth of abnormal single cells in the body [##UREF##14##55##]. Oncogene amplification usually occurs late in tumor progression and correlates well with aggressiveness of tumor [##REF##10752688##57##]. The function of the S100A4 protein and its role in metastasis is unclear at present. It has been reported that S100A4 may affect the function of cytoskeletal proteins including actin and non muscle myosin [##REF##8455951##58##,##REF##8051043##59##] so it is possible that S100A4 may regulate cell shape and or motility. It was reported that over expression of S100A4 protein is closely correlated with many functions for tumor aggressiveness, such as lymph node metastasis [##REF##12887505##60##].</p>", "<p>In the present study, S100A4 protein expression was examined by immunohistochemistry in infiltrating ductal, infiltrating lobular carcinoma and lymph node metastasis and their relation to tumor promotion and progression. Positive expression of S100A4 was observed in 45.1% of the infiltrating ductal carcinoma node negative cases, while in infiltrating lobular carcinoma with node negative, the expression of S100A4 protein was observed in 48.8%. This shows that S100A4 has a similar expression level in both infiltrating ductal carcinoma and infiltrating lobular carcinoma with node negative. S100A4 staining was not observed in normal breast tissues. S100A4 was expressed in a higher percentage in breast cancer tissue compared to normal tissue which showed direct correlation of S100A4 protein in infiltrating breast cancer node negative. This suggests that S100A4 is over expressed in infiltrating breast carcinoma compared to normal breast. Positive expression of S100A4 protein was observed only in 13.5 % of cases of infiltrating ductal carcinoma node positive while positive expression of S100A4 protein was observed in 35.1% of matched lymph node metastasis. These results showed there was a decrease in expression of S100A4 in infiltrating ductal carcinoma node positive (13.5%) compared with IDC node negative (45.1% positive staining), but interestingly there is an increase of expression of S100A4 protein at metastatic lymph node site. This suggests S100A4 may play a role in advanced breast cancer especially with lymph node metastasis. It was also interesting to note that the expression was seen in one site i.e. either primary tumor or its metastatic lymph node only.</p>", "<p>The majority (70%, 12/17) of paired samples showed that when there was a positive expression of S100A4 protein in metastatic lymph node, there was associated negative expression in the primary tumor (infiltrating ductal carcinoma) of the same patient. One case showed positive expression in both primary tumor and its metastatic lymph node at the same time. This was comparable with another study which showed that S100A4 over expression directly correlated with tumor progression [##REF##12887505##60##]. This study found that S100A4 was expressed in more cases of metastatic lymph node (35.1% cases) compared to matched infiltrating ductal carcinoma node positive (13.5% cases). These results show that there was a decrease expression of S100A4 protein in the primary tumor and increase expression in the metastatic lymph node for the same patient. This suggests that S100A4 is highly expressed in newly growing cancer either in the primary or metastatic sites. S100A4 protein was expressed in the cytoplasm of node negative IDC and ILC, whereas decreased in more advanced cancer (node positive). The reduction in S100A4 expression in the primary site may probably be related to cancer cells in the process of migration to distant sites.</p>" ]

[ "<title>Conclusion</title>", "<p>In conclusion S100A4 protein expression appears to be expressed widely in early and advanced stages of breast cancer compared with normal breast. This study indicates a complex role of S100A4 in breast cancer of different types and stages. The difference in the expression of S100A4 protein suggests it may be useful as an independent marker of breast cancer which appears to be down regulated in more advanced stages of breast cancer. However a larger study with more ILC and metastatic cases may clarify the role and function of S100A4 in breast cancer progression.</p>" ]

[ "<p>This is an Open Access article distributed under the terms of the Creative Commons Attribution License (<ext-link ext-link-type=\"uri\" xlink:href=\"http://creativecommons.org/licenses/by/2.0\"/>), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.</p>", "<title>Background</title>", "<p>Breast cancer is the most common cancer and cause of deaths in women around the world. Oncogene amplification usually occurs late in tumor progression and correlates well with aggressiveness of tumor. In fact the function of the S100A4 protein and its role in metastasis is unclear at present. The purpose of the study was to determine the expression of S100A4 protein in the invasion status and metastatic potential of breast cancer by using tissue microarray and to determine its role in breast cancer based on the expression of S100A4 gene product.</p>", "<title>Methods</title>", "<p>S100A4 protein expression was examined by immunohistochemistry (IHC) using commercially available tissue microarray containing malignant and normal breast tissue cores from 216 patients.</p>", "<title>Results</title>", "<p>S100A4 was absent in normal breast tissues while positive in 45.1% of infiltrating ductal carcinoma (IDC) node negative and 48.8% of infiltrating lobular carcinoma node negative. In paired samples, S100A4 protein was expressed in 13.5% of IDC node positive cases and 35.1% of matched lymph node metastasis.</p>", "<title>Conclusion</title>", "<p>S100A4 protein expression appears widely expressed in early and advanced breast cancer stages compared with normal breast. Our study suggests S100A4 may play a role in breast cancer progression and may prove to be an independent marker of breast cancer which appears to be down regulated in more advanced stages of breast cancer.</p>" ]

[ "<title>Abbreviations</title>", "<p>IHC: Immunohistochemistry; IDC: Infiltrating ductal carcinoma; ILC: Infiltrating lobular carcinoma.</p>", "<title>Authors' contributions</title>", "<p>NII carried out Immunohistochemical part of the study and lab work, participated in drafting the manuscript. GK carried out the pathological part of the study, participated in drafting the manuscript. HH performed the statistical analysis. MSH initiated the project, participated in drafting the manuscript. All authors read and approved this manuscript.</p>" ]

[ "<title>Acknowledgements</title>", "<p>I would like to thank Advanced Medical and Dental Institute (AMDI) for financial assistance and use of laboratory facilities. Also I'd like to extend my thanks to the AMDI diagnostic laboratory staff especially Puan Mariam Azmi and Encik Yahaya Osman for their technical expertise. Finally I'd like to thank all of my family for their support in happiness and sorrow.</p>" ]

[ "<fig position=\"float\" id=\"F1\"><label>Figure 1</label><caption><p>Negative control, S100A4 protein immunohistochemical staining in breast cancer tissue, showing absent staining (×200).</p></caption></fig>", "<fig position=\"float\" id=\"F2\"><label>Figure 2</label><caption><p>Positive expression of S100A4 within infiltrating ductal carcinoma node negative (×200).</p></caption></fig>", "<fig position=\"float\" id=\"F3\"><label>Figure 3</label><caption><p>Positive expression of S100A4 within infiltrating lobular carcinoma node negative (×200).</p></caption></fig>", "<fig position=\"float\" id=\"F4\"><label>Figure 4</label><caption><p>Negative expression of S100A4 protein within normal breast tissues (×200).</p></caption></fig>", "<fig position=\"float\" id=\"F5\"><label>Figure 5</label><caption><p>Positive expression of S100A4 within matched lymph node metastasis (×200).</p></caption></fig>" ]

[]

[]

[]

[]

[]

[]

[]

[]

[ "<graphic xlink:href=\"1475-2867-8-12-1\"/>", "<graphic xlink:href=\"1475-2867-8-12-2\"/>", "<graphic xlink:href=\"1475-2867-8-12-3\"/>", "<graphic xlink:href=\"1475-2867-8-12-4\"/>", "<graphic xlink:href=\"1475-2867-8-12-5\"/>" ]

[]

{ "acronym": [], "definition": [] }

[ "<title>Background</title>", "<p>Xerostomia, mucositis and dysphagia are known serious adverse reactions associated with radiotherapy (RT) or radiochemotherapy (CT-RT) of tumors localized in the head and neck region. Of particular concern is xerostomia, which develops acutely, but persists chronically afterwards and may lead to serious complications. Another serious toxic reaction to RT is xerophthalmia (dry eye syndrome/keratoconjunctivitis sicca), which can develop after local irradiation of the fronto-orbital region and can lead to severe visual impairment. Generally, the burden caused by RT-induced toxic reactions is not only detrimental for the patient's quality of life and the compliance with treatment, but may also have considerable pharmaco-economic consequences because of the costs incurred by loss of productivity, hospitalization and, frequently, expensive supportive measures [##UREF##0##1##]. It is therefore a major and desirable goal of the RT to achieve tumor remission without compromising the general well-being of the patient.</p>", "<p>Amifostine is a unique drug selectively protecting non-affected tissues from CT-RT induced toxicity. Its efficacy in preventing toxic damages induced by irradiation of tumors localized in the head and neck region, particularly in the naso-oropharynx, has been demonstrated in several clinical trials [##REF##11849797##2##, ####REF##9653491##3##, ##UREF##1##4##, ##REF##11013273##5##, ##REF##11984063##6##, ##REF##11072160##7####11072160##7##].</p>", "<p>We report here our experience with amifostine, which was used as a prophylactic measure during RT of malignant tumors localized in the fronto-orbital region, involving the ipsilateral lacrimal gland.</p>" ]

[ "<title>Methods</title>", "<title>Patient sample</title>", "<p>Five Patients, who were admitted to the department of radiation oncology, Inselspital Bern with histopathologically characterized unilateral malignant tumors of the orbital region and who gave informed consent for treatment with RT or CT-RT combined with amifostine, where included in this study. None of the patients had a history of dry eye, sarcoidosis, or thyroid associated eye disease.</p>", "<title>Treatment</title>", "<p>According to the diagnosis and histopathological findings for each patient a 3D RT planning was performed. Using dose volume histograms (DVH) minimum and maximum dose exposure of the lacrimal gland was determined. Two patients received chemotherapy according to the standard CHOP-protocol. In one patient, chemotherapy preceded RT and in another it was given concurrently with RT.</p>", "<p>Amifostine in a dose of 500 mg was given subcutaneously 30 min prior to each RT-fraction to all patients</p>", "<title>Lacrimal gland assessment</title>", "<p>In order to check for possible existence of dry eye syndrome in the post – treatment period lacrimal gland function was systematically assessed in all patients. The objective ophthalmic tests were performed at different time intervals after the termination of RT. The longest interval between RT and the tests was 88 weeks (pat. No 4) and the shortest 9 weeks (pat. No 5), median interval was 57.5 weeks.</p>", "<p>The following objective tests were performed: a) Schirmer I (with and without local anesthesia), which measures basic and reflex-secretion of the lacrimal gland, b) Schirmer II (nasal stimulation) test, which measures stimulated secretion of the lacrimal gland c) Tear-film break up time (BUT), which measures the time of onset of a random appearance of the first black spots on the cornea after one single palpebral (eyelid-blinking) closure under standardized conditions, d) corneal and conjunctival fluorescein staining and e) Rose bengal staining of the conjunctiva.</p>", "<p>A single examiner performed all of the tests and interviews during the examination in the department of ophthalmology. In all tests of the non-irradiated, contralateral eye served as control except in patient 1. The procedures are briefly described below.</p>", "<title>Dry eye symptom questionnaire</title>", "<p>Patients were first interviewed to survey the frequency of occurrence of various dry eye symptoms including dryness, grittiness, redness, excess tearing or watery eyes, sensitivity (to smoke, wind, air conditioning) and soreness. Response categories used for analyses included never, seldom (two to three times per week), often (four to five times per week), and always (everyday).</p>", "<title>Schirmer I test</title>", "<p>The Schirmer I test (Laboratoires H. Faure, Annonay, France) was performed without anesthetic (5 min, open eye) and the strip was placed over the inferior lid margin towards the lateral canthus. Abnormal values were defined as &lt; 5 mm in 5 min for the Schirmer test. Baseline secretion is determined essentially as described above, except that the eye is anesthetized 2 minutes before the exam with one drop of local anesthetic (Novocaine 2%, Inselspital, Bern).</p>", "<title>Schirmer II test</title>", "<p>The procedure is similar to Schirmer I test, except that after the suspension of the paper strip in the inferior fornix of a non-anaesthetized eye, the nasal mucosa is stimulated for 2 minutes with a cotton tip frotting on the nasal mucosa between the concha medalis and the concha inferior [##REF##1985472##8##].</p>", "<title>Tear-film break up time (BUT)</title>", "<p>The tear-film break-up time measurement was taken using the cobalt blue illumination on the slit-lamp and a 3 mm wide scanning beam in each eye. Fluorescein sodium was instilled on the inferior palpebral conjunctiva using a Fluorescein Ophthalmic Strip (Haag-Streit, Köniz, Switzerland). The latency to the first, random appearance of a black spot, in the otherwise yellow-green colored surface of the eye after complete closure of the eye-lid, is measured. The normal latency to the first black-spot appearance is 15 – 35 sec and the values below 10 indicate an insufficient lubrification of the corneal surface.</p>", "<title>Corneal and conjunctival fluorescein staining</title>", "<p>Fluorescein sodium was instilled on the inferior palpebral conjunctiva using a Fluorescein Ophthalmic Strip (Haag-Streit, Köniz, Switzerland). Following instillation of fluorescein, the patient was instructed to blink several times. The slit-lamp cobalt blue was used in the assessment of tear break-up time and fluorescein staining to enhance the appearance of the fluorescein. Staining was recorded for the cornea and the conjunctiva. The cornea and adjacent exposed bulbar conjunctiva was graded using a 0–3 scale resulting in a total staining score. Abnormal staining (fluorescein and rose bengal) was classified as greater than or equal to a score of 3 [##REF##4183019##9##].</p>", "<title>Rose bengal staining of the conjunctiva</title>", "<p>Rose bengal staining of the conjunctiva was performed by using a Rosets TM Rose Bengal Ophthalmic Strip (Chauvin Pharmaceuticals Ltd, Brentwood, UK) wetted with non-preserved buffered saline and instilled on the inferior bulbar conjunctiva. Grading for rose bengal staining was described previously [##REF##4183019##9##]. A score 3 or higher was considered pathological [##REF##8565190##10##].</p>", "<title>Adverse events</title>", "<p>Acute and late adverse effects were recorded using the Radiation Therapy Oncology Group (RTOG) score <ext-link ext-link-type=\"uri\" xlink:href=\"http://www.rtog.org/members/toxicity/late.html\"/>. Spontaneously reported signs of eye discomfort such as for instance tears, redness or dryness of the eye, were regularly recorded. It is to note, that there were no subjective or objective signs of eye discomfort in any patient prior to CT-RT or RT.</p>" ]

[ "<title>Results</title>", "<title>Patients characteristics</title>", "<p>Demographic characteristics of patients are presented in Table ##TAB##0##1##. Incisional histopathology confirmed malignant tumors of different tumor entities (Non-Hodgkin B-Cell Lymphoma, n = 4; neuroendocrine carcinoma, n = 1) with expansion into the orbita. Two patients were male, three were female, median age was 63 years (range, 28 – 79 years).</p>", "<title>Clinical outcome</title>", "<p>Table ##TAB##1##2## illustrates the details of the treatment modalities and the outcome of treatment in each patient. Combined CT-RT was given in two and RT alone to 3 patients. The median RT treatment period was 29 days (range, 23 – 39 days). The median total RT dose applied to the reference point according to the International Comission on Radiation Units and Measurements (ICRU) 10 was 40 Gy (range, 36 – 60 Gy). Median lacrimal gland exposure on the treated side was 35.9 Gy (range, 16.8–42.6 Gy). All patients completed the treatments according to the planned protocol. In all patients the response to therapy was excellent with complete response in 2 patients and a partial remission in the remaining three patients.</p>", "<p>A brief description of the histories is given below.</p>", "<title>Patient No 1</title>", "<p>Presented himself with progressive protrusion of his left eye, anosmia and difficulties with nasal breathing. CT and MRI examinations showed a frontobasal tumorous formation with extension from frontal sinus to sphenoidal sinus with protrusion to orbital and nasal cavity. There were also large osseous destructions of the lamina cribrosa and the medial orbital and maxillar sinus walls. Histopathological findings confirmed the diagnosis of a malignant neuroendocrine small cell carcinoma in advanced stage.</p>", "<p>The patient was hospitalized and sequential CT-RT initiated. He was treated according to the CHOP-protocol with a good and rapid response after 1 cycle. Radical, frontobasal radiotherapy with a total dose of 60 Gy, applied simultaneously with the first and second out of totally six cycles of chemotherapy with Cisplatin 60 mg iv and Etopophos 240 mg iv led to almost complete regression of the tumorous tissue.</p>", "<p>In the course of the RT the patient developed an acute grade II enoral mucositis, conjunctivitis and rhinitis, which regressed without further consequences after topical treatment.</p>", "<title>Patient No 2</title>", "<p>Underwent surgical intervention and subsequent external beam RT (total dose 50 Gy) of the Th-3-5 region because of metastatic, epidural tumors with compression of the spinal cord. 3 years later she presented herself again for consultation because of a massive palpebral edema, ptosis and exophthalmus of the right eye. Histopathological findings confirmed the diagnosis of a low-grade Non-Hodgkin B-Cell Lymphoma, stage IAE, with orbital expansion. A RT was performed with a total dose of 36 Gy. There were no toxic reactions found. One and a half year later the patient was in complete remission.</p>", "<title>Patient No 3</title>", "<p>Suffered from a progressive exophthalmus in her right eye for several years. The MRI showed an expansive tumorous mass in the right orbital cavity with intracranial spreading. Histopathologically findings confirmed the diagnosis of a Non-Hodgkin B-Cell Lymphoma, stage IAE, from the mucosa associated lymphatic tissue (MALT) type. Local radiotherapy of the right orbital region was performed to a total dose of 40 Gy. 27 weeks later the patient showed a complete remission of the tumor.</p>", "<p>Radiotherapy was well tolerated. The only adverse reactions were grade I – skin reactions, a mild transitory conjunctivitis and a localized alopecia without consequences.</p>", "<title>Patient No 4</title>", "<p>Presented herself for a progressive edema in the right temporal region. A biopsy revealed a diffuse Non-Hodgkin B-Cell Lymphoma, stage IAE, with infiltration of the temporal muscle and expansion into the orbit. CHOP-protocol was initiated and completed in 3 cycles. Additionally, she received local external beam RT of the temporal region for consolidation with a total dose of 46 Gy</p>", "<p>Until the end of the RT, except for a mild skin erythema, no other toxic reactions or complications were found. However, after 7 applications of amifostine (12 days after the start of RT) an allergic reaction to amifostine developed in the form of a generalized skin rash and edema of the face and neck. After stopping amifostine, the patient recovered without consequences. The post RT-control of the patient showed partial remission of the tumor.</p>", "<title>Patient No 5</title>", "<p>Presented himself with a progressive, persistent edema of the right eyelid, which had started 3 years ago. Histopathology confirmed the diagnosis of a low grade Non-Hodgkin B-Cell Lymphoma, stage IAE. Local RT of the right palpebral region was performed to a total dose 36 Gy.</p>", "<p>There were no major toxicities during the RT treatment. Mild skin erythema of the right eyelid was the only adverse reaction. The post RT-control of the patient showed partial remission of the tumor.</p>", "<title>Adverse reactions</title>", "<p>As seen from the case histories there were no major toxicities during the RT treatment and only few, mostly mild and transitory adverse events were recorded. One patient however showed, an allergic reaction to amifostine with generalized skin rash and edema of the face and neck which disappeared after halting amifostine.</p>", "<title>Post-treatment ophthalmologic control</title>", "<p>The ophthalmic exams were performed at different time intervals after the termination of RT. The longest interval between finished RT and the exams was 88 weeks (pat. No 4) and the shortest 9 weeks (pat. No 5). The results of the post treatment-control of lacrimal gland function are illustrated in Table ##TAB##2##3##. As seen there were no clinically relevant impairments of glandular function, which would indicate serious dry-eye syndrome. Seldom feeling of a dryness of the exposed eye (n = 1), often tears (n = 1) and often eye-burning (n = 1) were the only subjective and reported complaints.</p>" ]

[ "<title>Discussion</title>", "<p>Selective protective effects of amifostine against RT induced side effects localized in various body organs (lungs, rectum, cervix, ovaries) have been demonstrated in numerous clinical trials [##REF##11984063##6##]. Its efficacy in patients with malignancies located in the head and neck region is, however, of particular importance. Ionizing radiation of this region impairs salivary and lacrimal gland functions, which has acute and long-term consequences for the patient. Proper lubrification of oral mucosa is a physiological prerequisite for normal chewing, swallowing, speaking, dental health and, last but not least, sleep [##REF##2019691##11##,##REF##11236308##12##]. Dryness of the eye, besides the discomfort, may lead to severe visual impairment due to the corneal damage. One study has reported the occurence of mild dry eye syndrome in 21% of patients after radiotherapy for stage IAE orbital lymphoma after a total dose of 30–51 Gy [##REF##12377334##13##]. However in this study no dose volume histogram analysis for dose exposure of the lacrimal gland was performed and the data is thus not directly comparable to our study.</p>", "<p>The experience with amifostine in patients with head and neck cancer is at present based on several studies. In a large (N = 315), randomized, comparative trial in patients with predominantly (≥75%) parotid gland carcinoma, Brizel et al. [##REF##11013273##5##] have shown that by comparison to RT alone the concurrent amifostine-RT treatment reduced the overall incidence of grade ≥2 xerostomia from 78% to 51%. In another comparative study in 50 patients with mainly naso-oro-pharyngeal tumor localization, Antonadou et al. have shown considerable reduction of acute (mucositis and dysphagia) and late (xerostomia) RT-toxicities in the amifostine treated group (n = 22). In the control group 73.9% of patients (n = 23) developed grade 2 xerostomia, whereas in the amifostine-RT group there were only 27.2% patients with mild grade 2 xerostomia. Similar results were also reported in an earlier trial by Büntzel et al. In none of these studies amifostine treatment compromised the outcome of RT. In contrast, by comparison to the controls, the tumor remission rate in amifostine treated patients was, as a rule, higher (up to 90%) [##REF##11849797##14##, ####REF##9653491##15##, ##REF##11013273##16####11013273##16##]. To the best of our knowledge, no clinical reports with amifostine in protecting lacrimal gland function from the irradiation of the orbital region have been reported.</p>", "<p>In a rabbit model Beutel et al. evaluated the effect of amifostine in the tear gland and found morphological and functional evidence of its radioprotective properties [##REF##17318564##17##].</p>", "<p>Even though limited to five patients only, our observation is suggesting high success of combined amifostine-RT-treatment regarding tumor remission. Two out of five patients have achieved complete remission and the remaining three a partial remission of the tumors. Further, none of the patients experienced any major toxic reaction to RT. In particular, neither subjective complaints nor objective findings of the post-treatment ophthalmologic tests have revealed signs of a dry-eye syndrome.</p>", "<p>As a matter of fact, we can not exclude the possibility that the lacrimal gland in our patients was exposed to a insufficiently high irradiation dose to induce any serious functional damage of the gland. Systematic investigations of the irradiation dose-effect relationship in cases of orbital tumors have not yet been done. In his review about the tolerance of the normal tissue to therapeutic radiation Emami [##UREF##2##18##] refers to the experiences related to the eye and parotid gland exposure to radiation. These observations seem to suggest a very steep dose-response curve at least for the eye. The quoted study of the exposure of the retina to 45–50 Gy led to detectable damages and that of 65 Gy and more to visual loss. For the parotid gland his estimate of the tolerance dose (TD) 5/5 (normal tissue complication probability at 5% within 5 years after radiotherapy), based on the reviewed studies was 32 Gy.</p>" ]

[ "<title>Conclusion</title>", "<p>Our results indicate that the addition of Amifostine to RT-/CH-RT in patients with tumors localized in orbital region is safe and associated with absence of dry eye syndrome. Our data encourage further studies of concurrent amifostine therapy.</p>", "<p>This result is in accordance with the results of other clinical studies showing radioprotective effect of amifostine in patients with malignancies located in the head and neck region and radiation exposure involving the parotid gland.</p>" ]

[ "<p>This is an Open Access article distributed under the terms of the Creative Commons Attribution License (<ext-link ext-link-type=\"uri\" xlink:href=\"http://creativecommons.org/licenses/by/2.0\"/>), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.</p>", "<title>Background</title>", "<p>To use amifostine concurrently with radiochemotherapy (CT-RT) or radiotherapy (RT) alone in order to prevent dry eye syndrome in patients with malignancies located in the fronto-orbital region.</p>", "<title>Methods</title>", "<p>Five patients (2 males, 3 females) with diagnosed malignancies (Non-Hodgkin B-cell Lymphoma, neuroendocrine carcinoma) involving the lacrimal gland, in which either combined CT-RT or local RT were indicated, were prophylactically treated with amifostine (500 mg sc). Single RT fraction dose, total dose and treatment duration were individually adjusted to the patient's need. Acute and late adverse effects were recorded using the RTOG score. Subjective and objective dry eye assessment was performed for the post-treatment control of lacrimal gland function.</p>", "<title>Results</title>", "<p>All patients have completed CT-RT or RT as indicated. The median total duration of RT was 29 days (range, 23 – 39 days) and the median total RT dose was 40 Gy (range, 36 – 60 Gy). Median lacrimal gland exposure was 35.9 Gy (range, 16.8 – 42.6 Gy). Very good partial or complete tumor remission was achieved in all patients. The treatment was well tolerated without major toxic reactions. Post-treatment control did not reveal in any patient either subjective or objective signs of a dry eye syndrome.</p>", "<title>Conclusion</title>", "<p>The addition of amifostine to RT/CT-RT of patients with tumors localized in orbital region was found to be associated with absence of dry eye syndrome.</p>" ]

[ "<title>Competing interests</title>", "<p>The authors have no personal financial or non-financial interests in any of the mentioned substances, companies or competing companies related to the study.</p>", "<title>Authors' contributions</title>", "<p>DG, PG and JC performed the acquisition of data and helped to draft the manuscript. RG and DA made substantial contributions to conception and design of the study and helped to draft the manuscript. All authors have given final approval of the version to be published.</p>" ]

[]

[]

[ "<table-wrap position=\"float\" id=\"T1\"><label>Table 1</label><caption><p>Characteristics of the patient sample</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td/><td align=\"center\" colspan=\"4\">Clinical factors</td></tr></thead><tbody><tr><td align=\"center\">Pat No</td><td align=\"center\">Gender</td><td align=\"center\">Age</td><td align=\"center\">Diagnosis</td><td align=\"center\">Tumor localization</td></tr><tr><td colspan=\"5\"><hr/></td></tr><tr><td align=\"center\">1</td><td align=\"center\">Male</td><td align=\"center\">28</td><td align=\"center\">Neuroendocrine carcinoma</td><td align=\"center\">Frontobasal part of the right nasal sinus</td></tr><tr><td align=\"center\">2</td><td align=\"center\">Female</td><td align=\"center\">60</td><td align=\"center\">Non-Hodgkin B-cell Lymphoma, low grade (St. IAE)</td><td align=\"center\">Right lacrimal gland</td></tr><tr><td align=\"center\">3</td><td align=\"center\">Female</td><td align=\"center\">69</td><td align=\"center\">Non-Hodgkin B-cell Lymphoma, low grade, (St. IAE; Malt-Type)</td><td align=\"center\">Right orbita</td></tr><tr><td align=\"center\">4</td><td align=\"center\">Female</td><td align=\"center\">63</td><td align=\"center\">Non-Hodgkin-B-cell Lymphoma, high grade, (St. IAE)</td><td align=\"center\">Right M. temporalis, with orbital expansion</td></tr><tr><td align=\"center\">5</td><td align=\"center\">Male</td><td align=\"center\">79</td><td align=\"center\">Non-Hodgkin B-cell Lymphoma, low grade, (St. IAE)</td><td align=\"center\">Right palpebra</td></tr></tbody></table></table-wrap>", "<table-wrap position=\"float\" id=\"T2\"><label>Table 2</label><caption><p>Treatment modalities and outcome</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td/><td align=\"center\" colspan=\"5\">Treatment factors</td></tr></thead><tbody><tr><td align=\"center\">Pat No</td><td align=\"center\">Radio chemotherapy CT/RT*</td><td align=\"center\">RT (total duration days)</td><td align=\"center\">Total RT Dose† Gy</td><td align=\"center\">Median RT dose Lacrymal gland Gy right/left</td><td align=\"center\">Clinical Outcome</td></tr><tr><td colspan=\"6\"><hr/></td></tr><tr><td align=\"center\">1</td><td align=\"center\">6/39</td><td/><td align=\"center\">60</td><td align=\"center\">16.8/16.5</td><td align=\"center\">partial remission</td></tr><tr><td align=\"center\">2</td><td/><td align=\"center\">35</td><td align=\"center\">36</td><td align=\"center\">35.9/0</td><td align=\"center\">complete remission</td></tr><tr><td align=\"center\">3</td><td/><td align=\"center\">29</td><td align=\"center\">40</td><td align=\"center\">42.6/5.4</td><td align=\"center\">complete remission</td></tr><tr><td align=\"center\">4</td><td align=\"center\">3/34</td><td/><td align=\"center\">46</td><td align=\"center\">35.4/0</td><td align=\"center\">partial remission</td></tr><tr><td align=\"center\">5</td><td/><td align=\"center\">23</td><td align=\"center\">36</td><td align=\"center\">36.3/0.54</td><td align=\"center\">partial remission</td></tr></tbody></table></table-wrap>", "<table-wrap position=\"float\" id=\"T3\"><label>Table 3</label><caption><p>Post-treatment lacrimal gland function</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td/><td align=\"center\" colspan=\"6\">Opthalmologic tests</td></tr></thead><tbody><tr><td align=\"center\">Pat No</td><td align=\"center\">Follow-up time (weeks after RT)</td><td align=\"center\">Schirmer I with anesthetic OD/OS</td><td align=\"center\">Schirmer I (without anesthetic) OD/OS</td><td align=\"center\">Schirmer II (nasal stimulation) OD/OS</td><td align=\"center\">BUT OD/OS</td><td align=\"center\">staining score</td></tr><tr><td colspan=\"7\"><hr/></td></tr><tr><td align=\"center\">1</td><td align=\"center\">57.5</td><td align=\"center\">35/35</td><td align=\"center\">10/9</td><td align=\"center\">2/8</td><td align=\"center\">10/10</td><td align=\"center\">1 (OS)</td></tr><tr><td align=\"center\">2</td><td align=\"center\">77</td><td align=\"center\">5/10</td><td align=\"center\">5/7</td><td align=\"center\">7/11</td><td align=\"center\">15/15</td><td align=\"center\">1 (OD)</td></tr><tr><td align=\"center\">3</td><td align=\"center\">27</td><td align=\"center\">19/11</td><td align=\"center\">24/25</td><td align=\"center\">17/17</td><td align=\"center\">4/9</td><td align=\"center\">1 (OD)</td></tr><tr><td align=\"center\">4</td><td align=\"center\">87.7</td><td align=\"center\">5/18</td><td align=\"center\">1/7</td><td align=\"center\">9/17</td><td align=\"center\">15/15</td><td align=\"center\">2 (OD)</td></tr><tr><td align=\"center\">5</td><td align=\"center\">8.7</td><td align=\"center\">20/10</td><td align=\"center\">17/10</td><td align=\"center\">not performed</td><td align=\"center\">6/6</td><td align=\"center\">0</td></tr></tbody></table></table-wrap>" ]

[]

[]

[]

[]

[]

[]

[ "<table-wrap-foot><p>Abbreviations: St. = stage; IAE = staging according to the Ann-Arbor-classification; M. = musculus</p></table-wrap-foot>", "<table-wrap-foot><p>*CT = chemotherapy; RT = radiotherapy; CT/RT = total cycles/total duration days; † for the RT reference point</p></table-wrap-foot>", "<table-wrap-foot><p>Abbreviations: RT = radiotherapy; OD = right eye; OS = left eye; BUT = break up time</p></table-wrap-foot>" ]

[]

[]

{ "acronym": [], "definition": [] }

[ "<title>Background</title>", "<p>Precise timing of both circadian and seasonal changes in physiology and behavior are important for animals [##REF##11506375##1##]. Biologically appropriate daily and seasonal timing depends upon normal and precise function of the circadian system [##REF##9333206##2##, ####REF##10707315##3##, ##REF##16940279##4####16940279##4##]. Genetic variation for circadian clock function can affect circadian physiological rhythms, daily cyclical behaviors, onset and offset of activity, melatonin rhythms, and regulation of seasonal physiological changes in energetics and reproduction [##REF##16940279##4##, ####REF##11561508##5##, ##REF##14582852##6####14582852##6##]. Such genetic variation may affect animal and human health and function through effects on biological rhythms and related physiological systems [##REF##16940279##4##,##REF##12622845##7##, ####REF##9406025##8##, ##REF##15155003##9####15155003##9##]. Particularly important for the circadian system may be pleiotropic variation that could cause correlated variation in more than one trait, thereby affecting multiple body systems [e.g., examples of gene knockout studies reviewed in [##REF##16940279##4##]].</p>", "<p>Studies on laboratory colonies of hamsters have described pleiotropic genetic variation in the circadian clock that also causes variation in photoperiodism [##REF##3746726##10##, ####REF##3351788##11##, ##REF##9376641##12##, ##UREF##0##13##, ##UREF##1##14####1##14##]. In Siberian hamsters, for example, genetically nonphotoperiodic individuals in short photoperiod have a delayed, lower amplitude nocturnal rise in pineal melatonin, a 4.5 hour delay in the onset of nightly activity, a longer free-running activity period (<italic>tau</italic>), a shorter duration of running wheel activity (<italic>alpha</italic>), and some nonphotoperiodic individuals are arrhythmic, relative to photoperiodic individuals [##REF##3746726##10##,##REF##3351788##11##]. In <italic>tau </italic>mutant Syrian hamsters, the free running rhythm is too short for proper entrainment to a short photoperiod, leading to an inability to produce a short photoperiod melatonin pattern [##REF##8612567##15##]. In contrast, studies on natural populations of rodents have reported genetic variation in seasonal photoperiodic traits which have not been found to be related to circadian variation [##REF##2624861##16##, ####UREF##2##17##, ##UREF##3##18##, ##REF##3946637##19##, ##REF##9151428##20####9151428##20##], suggesting independent sources of variation. It is not clear how commonly variation in photoperiodic seasonality is correlated with circadian rhythms, and additional models that relate genetic variation in circadian and seasonal function would be useful.</p>", "<p>Strains of laboratory rats vary both in circadian rhythms [##REF##2333353##21##, ####REF##2065757##22##, ##REF##1798768##23##, ##REF##10198388##24##, ##REF##10516259##25####10516259##25##] and in photoperiodic responses that include reproduction, food intake, and body mass [##REF##3651546##26##, ####REF##9209091##27##, ##REF##11705766##28##, ##REF##12135532##29##, ##REF##15579603##30####15579603##30##]. The F344/NHsd strain of rats suppresses reproduction, food intake, and somatic growth in short photoperiods, while Harlan Sprague Dawley (HSD) rats do not. F344 rats are becoming a model for mechanisms of photoresponsiveness [##REF##17110533##31##], and may be useful models for the study of correlated genetic variation in rhythms, regulation of appetite and body mass, and reproduction. A recent comparison of the rhythm of excretion of the major metabolite of melatonin, 6-sulfatoxymelatonin, between photoperiodic F344 rats and nonphotoperiodic HSD rats suggests that both strains have long duration melatonin secretion in short photoperiod and short duration melatonin secretion in long photoperiod [##REF##16162292##32##]. Thus, both photoresponsive F344 and nonphotoresponsive HSD rats were similar in pattern of melatonin production to photoresponsive rather than nonphotoresponsive Siberian hamsters [##REF##3746726##10##]. However, HSD rats excreted only about half as much 6-sulfatoxymelatonin per unit body mass as F344 rats, and some individual HSD rats had little or no nocturnal rise in excretion, similar to nonphotoresponsive Siberian hamsters [##REF##3746726##10##].</p>", "<p>For this study, a companion study to Price et al. [##REF##16162292##32##], we tested two competing hypotheses. One hypothesis is that differences in photoresponsiveness between F344 and HSD rats are caused by differences in circadian clock function and/or clock outputs cause differences in photoresponsiveness. A competing hypothesis is that differences in photoresponsiveness are caused by differences at the level of melatonin secretion and responses to melatonin [##REF##16162292##32##], rather than by differences in clock function. We tested for circadian differences between young male F344 and young male HSD rats by measuring running wheel activity in short photoperiod, long photoperiod, and constant dark (DD) to assess circadian traits and the degree of nocturnality. Intrinsic tendencies for nocturnality may be most apparent in DD, when individuals must rely entirely on their circadian clock to indicate subjective night and day. Circadian differences could occur in short or long photoperiods as well, though direct effects of light and dark might mask effects of the endogenous circadian rhythm on nocturnality.</p>" ]

[ "<title>Methods</title>", "<p>Under the hypothesis that differences in clock function cause nonphotoresponsiveness in HSD rats, we predicted that HSD rats would be more likely to be arrhythmic or have poorly defined circadian rhythms of activity, because a damaged or altered clock or clock output pathways that result in an inability to track time-of-day or to pass time-of-day information to other areas of the brain would result in poor circadian regulation of activity. For similar reasons, we predicted that HSD rats would have lower nocturnality than F344 rats because HSD rats would assess day and night inaccurately. Low nocturnality in HSD rats might be most extreme in constant dark, when light cues are not available to mask circadian outputs. In addition, if HSD rats were found to be able to entrain activity to the dark period at least in long photoperiod, we predicted a free-running rhythm greater than 24 hours in HSD rats (which results in entrainment even if there is a deficit in the phase delay portion of the phase response curve) along with inability to lengthen the activity period when moved from long days to short days (which might result from inability to properly phase-delay in response to short photoperiod). This prediction follows Puchalski and Lynch [##REF##3351788##11##], who used the nonparametric theory of entrainment [##REF##10885874##33##] in developing a model for nonresponsiveness of Siberian hamsters.</p>", "<p>The study followed international standards for animal care and welfare, and was approved by the institutional animal care and use committee at the College of William and Mary (IACUC 0018).</p>", "<title>Experiment 1: Comparison of HSD to F344 male rats in LD</title>", "<p>This experiment was designed to compare circadian traits and nocturnality in a photoperiod treatment with light cues that would not normally trigger photoperiodic responses. Activity patterns of young male F344/NHsd rats (breeders from Harlan, Indianapolis, Indiana) and HSD rats (Hsd:Sprague Dawley; breeders from Harlan, Indianapolis, Indiana) were compared in long days (LD), 16 L: 8 D with lights on at 0500 EST (N = 15 F344 rats and 18 HSD rats). Rats were only tested during the four weeks after weaning in order to match the period when F344 rats and other strains are known to show variation in photoperiodic responses [##REF##9209091##27##,##REF##11705766##28##,##REF##15579603##30##,##REF##9687308##34##]. In order to control for potential changes related to age, each individual rat was tested in only one photoperiod treatment.</p>", "<p>All rats were gestated and raised in LD (16 L: 8 D with lights on at 0500 EST) prior to weaning at 23 ± 2 days of age. They were then transferred to individual cages with activity wheels (Harvard Apparatus, Holliston, Massachusetts, Rodent Activity Wheel and Cage, Catalog No. 60-1943) and placed in environmental and photoperiod chambers (Revco, Asheville, North Carolina) in groups of up to 12 rats/chamber. Magnetic switches on the running wheels signaled revolutions to an event recorder sending output in 6-minute data collection periods, or 'bins', to a personal computer. Rats were fed a laboratory diet (Harlan Teklad LM – 485 Sterilizable Mouse/Rat Diet 7012, Madison, WI) and tap water ad libitum. Temperature was maintained at 22.5 ± 1°C. Data collection ended after four weeks ± 4 days. The study followed international standards for animal care and welfare, and was approved by the institutional animal care and use committee at the College of William and Mary (IACUC 0018).</p>", "<title>Experiment 2: Comparison of HSD to F344 male rats in SD</title>", "<p>This experiment was designed to compare circadian traits and nocturnality in a photoperiod with light cues that would normally trigger reproductive suppression, reduced food intake, and slowed somatic growth in a photoresponsive rat. Activity patterns of young male F344 and HSD rats (N = 16 per group) were compared for short days (SD), 8 L: 16 D with lights on at 0900 EST. Except for the photoperiod treatment, procedures and data collection were as in Experiment 1.</p>", "<title>Experiment 3: Comparison of HSD to F344 male rats in DD</title>", "<p>This experiment was designed to compare circadian traits and nocturnality of F344 and HSD rats. Activity patterns of young male F344 rats were compared with those of young male HSD rats (N = 10 per group) in constant darkness (DD). Except for the photoperiod treatment, procedures and data collection were as in Experiment 1.</p>", "<title>Data Analysis</title>", "<p>Data collection and analysis of nocturnality and diurnality was carried out using a software package, Tau, generously provided by Roberto Refinetti <ext-link ext-link-type=\"uri\" xlink:href=\"http://www.circadian.org\"/>. Wheel revolutions were recorded in 6-minute bins and plotted as activity records. Preliminary analyses were conducted to test the effect of removing bins with small numbers of wheel rotations. Because variation in nocturnality was most apparent without removing bins, analyses were conducted using all activity. For all statistical analyses, data were taken only from the last 15 days of activity in order to allow rats to adjust to the treatment photoperiod in SD and DD. For rats in LD, data collection was also restricted to the final 15 days of activity in order to match ages across all three experiments. The nocturnality index is defined here as the ratio of time active in the night (or subjective night) to the total time active over 24 hours. Thus, individuals that were more active during the night had a higher nocturnality index. The Tau program also calculated the <italic>tau </italic>(free-running period) by the chi-square periodogram method for each rat in DD. Phase angle of entrainment, activity onset, activity offset, and <italic>alpha </italic>(duration of activity) were calculated following methods slightly modified from Majoy and Heideman [##UREF##2##17##] and Sullivan and Lynch [##REF##3952177##35##] as described below.</p>", "<p>Activity onset was defined as the first bout of running activity lasting at least two 6-min recording bins that was preceded by at least 2 hours with no sustained activity (e.g., activity in no more than one consecutive bin) and followed by additional activity within the next hour. Eye-fitted lines were drawn through these daily activity onsets on an actogram to obtain a mean daily activity onset for each rat. The activity offset was defined as the end of the last bout of activity followed by at least 2 hours of no sustained activity. Again, eye-fitted lines were drawn through the daily offset times on an actogram to obtain a mean offset for each rat. <italic>Alpha </italic>was defined as the difference between activity offset and activity onset. Phase angle of entrainment was defined as the difference between activity onset and lights off (calculated only in SD and LD photoperiods). Unpaired t-tests were performed to test for statistical significance between the two strains of rats in each photoperiod treatment for each circadian parameter.</p>", "<p>Mean nocturnality index, <italic>tau</italic>, phase angle of entrainment, activity onset, activity offset, and <italic>alpha </italic>were compared between strains in each photoperiod by unpaired t-tests, with the criteria for statistical significance set at P &lt; 0.05. Because data were taken in multiple runs, run was included as a factor in the analyses. Run did not affect significance tests, and was not considered further.</p>" ]

[ "<title>Results and discussion</title>", "<p>In all three photoperiods, there were strain specific differences in activity patterns. Representative rhythms from individual rats with circadian parameters near the means for their treatment group are shown in Figure ##FIG##0##1a–f##. Summary data showing the proportion of rats active in each hour are presented in Figure ##FIG##1##2##.</p>", "<p>In LD, SD, and DD, F344 rats had higher nocturnality indices than HSD rats (LD: t = 6.85, p &lt; 0.0001; SD: t = 4.83, p &lt; 0.0001; DD: t = 5.05, p &lt; 0.0001; Figure ##FIG##2##3a##), with a higher proportion of their activity during the night period than HSD rats (Figure ##FIG##1##2##). In all three photoperiod treatments, <italic>alpha </italic>was significantly longer in HSD than in F344 rats (LD: t = 5.60, p &lt; 0.0001; SD: t = 4.92, p &lt; 0.0001; DD: t = 3.86, p = 0.0005; Figure ##FIG##2##3b##). There were differences in activity onset in LD (t = 5.01, p &lt; 0.0001; Figure ##FIG##2##3c##) but not in SD (t = 1.12, p = 0.27; Figure ##FIG##2##3c##), and differences in activity offset in both LD (t = 4.28, p = 0.0002; Figure ##FIG##2##3d##) and SD (t = 8.37, p &lt; 0.0001; Figure ##FIG##2##3d##). There was a significant difference in phase angle of entrainment in LD (t = 4.95; p &lt; 0.0001; Figure ##FIG##2##3e##), but not in SD (t = 1.26; p = 0.22; Figure ##FIG##2##3e##). HSD rats also displayed <italic>tau </italic>longer than 24 h, significantly longer than <italic>tau </italic>of F344 rats (t = 3.01; p = 0.0048; Figure ##FIG##2##3f##).</p>", "<p>Overall, HSD rats tended to extend activity into the light period and also had a higher amount of activity during the light period than F344 rats (Figures ##FIG##0##1##, ##FIG##1##2##, &amp;##FIG##2##3##). Some HSD rats had activity spread so evenly through the constant dark period as to appear nearly arrhythmic (Figure ##FIG##0##1g##).</p>", "<p>Young male HSD and F344 rats differed in multiple circadian parameters of running wheel activity (Figure ##FIG##2##3##). HSD rats were more likely to begin and end activity during the light period or subjective day in SD, LD, and DD. In addition, HSD rats had more activity during day or subjective day than F344 rats in all three treatments (Figs. ##FIG##1##2## &amp;##FIG##2##3a##). Consistent with Aschoff's prediction for more diurnal animals [##REF##386643##36##], HSD rats also had <italic>tau </italic>&gt; 24 hours, significantly longer than the &lt; 24 hour <italic>tau </italic>of F344 rats (Figure ##FIG##2##3f##). However, HSD rats had significantly longer <italic>alpha </italic>than F344 rats (Figure ##FIG##2##3b##), and both strains entrained well to SD (Figure ##FIG##1##2c,d##), which is consistent with the hypothesis that differences in melatonin secretion or responses to melatonin cause variation in photoresponsiveness, and inconsistent with the hypothesis that circadian deficits in HSD rats cause nonphotoresponsiveness.</p>", "<p>It has been proposed that nonphotoperiodism may occur in Siberian hamsters because they fail to produce an SD pattern of melatonin secretion due to a failure to integrate photoperiodic information by the circadian system [##REF##3746726##10##]. The cause of this failure was associated with a long-duration <italic>tau </italic>(24.04 +/- 0.05), inability to undergo phase delay in response to a light pulse, and a failure to increase <italic>alpha </italic>or entrain properly in SD. While HSD rats have some similarities to nonphotoperiodic Siberian hamsters, the results are not identical. HSD rats have relatively low amplitude 6-sulfatoxymelatonin excretion patterns when adjusted for body mass [##REF##16162292##32##] and a long duration <italic>tau </italic>(24.05 +/- 0.02; Figure ##FIG##2##3f##), similar to nonphotoperiodic Siberian hamsters [##REF##3746726##10##,##REF##3351788##11##]. However, HSD rats produce 6-sulfatoxymelatonin excretion rhythms that adjust to night length and are similar to those of F344 rats in SD [##REF##16162292##32##], and HSD rats successfully decompress activity to a long-duration pattern when transferred from LD to SD (Figure ##FIG##2##3b##). We have not tested directly the ability of HSD rats to phase delay, but successful entrainment to SD suggests that HSD rats have the ability to phase delay to achieve entrainment to SD. Thus, while HSD rats show some similarities to nonphotoperiodic Siberian hamsters, the evidence from this study, together with results of Price et al. [##REF##16162292##32##], suggests that the circadian system of HSD rats is able to respond appropriately to photoperiodic information (this study) to cause a short-day melatonin secretion pattern [##REF##16162292##32##], but HSD rats are unable to respond reproductively to a short-day melatonin pattern [##REF##16162292##32##].</p>", "<p>In combination with the companion study [##REF##16162292##32##], our results support the hypothesis that differences in how F344 and HSD rats secrete and/or respond to melatonin are responsible for differences in photoresponsiveness. We found significant differences between strains in circadian aspects of wheel running, and these differences indicate that HSD rats tend to have more activity during the day, have a greater tendency toward arrhythmia, and differ in some circadian parameters from F344 rats. However, the differences reported here in circadian parameters and nocturnality do not prevent similar timing of 6-sulfatoxymelatonin excretion rhythms in F344 and HSD, albeit potentially lower in amplitude or even arrhythmic in HSD rats [##REF##16162292##32##]. Rats are normally nocturnal [e.g., [##REF##2065757##22##,##REF##10198388##24##]], and our results suggest that F344 rats have the more species-typical pattern of strong nocturnality, while HSD rats differ from that pattern. Variation in activity and 6-sulfatoxymelatonin excretion has also been reported between other strains of rats [##REF##10198388##24##], but our results suggest there is no necessary relationship between photoresponsiveness and circadian rhythm characteristics in rats. For example, some of the strain variation we report in wheel running activity might be related to restlessness, fearfulness, or intrinsic rewards of wheel running that differ among rat strains.</p>" ]

[ "<title>Results and discussion</title>", "<p>In all three photoperiods, there were strain specific differences in activity patterns. Representative rhythms from individual rats with circadian parameters near the means for their treatment group are shown in Figure ##FIG##0##1a–f##. Summary data showing the proportion of rats active in each hour are presented in Figure ##FIG##1##2##.</p>", "<p>In LD, SD, and DD, F344 rats had higher nocturnality indices than HSD rats (LD: t = 6.85, p &lt; 0.0001; SD: t = 4.83, p &lt; 0.0001; DD: t = 5.05, p &lt; 0.0001; Figure ##FIG##2##3a##), with a higher proportion of their activity during the night period than HSD rats (Figure ##FIG##1##2##). In all three photoperiod treatments, <italic>alpha </italic>was significantly longer in HSD than in F344 rats (LD: t = 5.60, p &lt; 0.0001; SD: t = 4.92, p &lt; 0.0001; DD: t = 3.86, p = 0.0005; Figure ##FIG##2##3b##). There were differences in activity onset in LD (t = 5.01, p &lt; 0.0001; Figure ##FIG##2##3c##) but not in SD (t = 1.12, p = 0.27; Figure ##FIG##2##3c##), and differences in activity offset in both LD (t = 4.28, p = 0.0002; Figure ##FIG##2##3d##) and SD (t = 8.37, p &lt; 0.0001; Figure ##FIG##2##3d##). There was a significant difference in phase angle of entrainment in LD (t = 4.95; p &lt; 0.0001; Figure ##FIG##2##3e##), but not in SD (t = 1.26; p = 0.22; Figure ##FIG##2##3e##). HSD rats also displayed <italic>tau </italic>longer than 24 h, significantly longer than <italic>tau </italic>of F344 rats (t = 3.01; p = 0.0048; Figure ##FIG##2##3f##).</p>", "<p>Overall, HSD rats tended to extend activity into the light period and also had a higher amount of activity during the light period than F344 rats (Figures ##FIG##0##1##, ##FIG##1##2##, &amp;##FIG##2##3##). Some HSD rats had activity spread so evenly through the constant dark period as to appear nearly arrhythmic (Figure ##FIG##0##1g##).</p>", "<p>Young male HSD and F344 rats differed in multiple circadian parameters of running wheel activity (Figure ##FIG##2##3##). HSD rats were more likely to begin and end activity during the light period or subjective day in SD, LD, and DD. In addition, HSD rats had more activity during day or subjective day than F344 rats in all three treatments (Figs. ##FIG##1##2## &amp;##FIG##2##3a##). Consistent with Aschoff's prediction for more diurnal animals [##REF##386643##36##], HSD rats also had <italic>tau </italic>&gt; 24 hours, significantly longer than the &lt; 24 hour <italic>tau </italic>of F344 rats (Figure ##FIG##2##3f##). However, HSD rats had significantly longer <italic>alpha </italic>than F344 rats (Figure ##FIG##2##3b##), and both strains entrained well to SD (Figure ##FIG##1##2c,d##), which is consistent with the hypothesis that differences in melatonin secretion or responses to melatonin cause variation in photoresponsiveness, and inconsistent with the hypothesis that circadian deficits in HSD rats cause nonphotoresponsiveness.</p>", "<p>It has been proposed that nonphotoperiodism may occur in Siberian hamsters because they fail to produce an SD pattern of melatonin secretion due to a failure to integrate photoperiodic information by the circadian system [##REF##3746726##10##]. The cause of this failure was associated with a long-duration <italic>tau </italic>(24.04 +/- 0.05), inability to undergo phase delay in response to a light pulse, and a failure to increase <italic>alpha </italic>or entrain properly in SD. While HSD rats have some similarities to nonphotoperiodic Siberian hamsters, the results are not identical. HSD rats have relatively low amplitude 6-sulfatoxymelatonin excretion patterns when adjusted for body mass [##REF##16162292##32##] and a long duration <italic>tau </italic>(24.05 +/- 0.02; Figure ##FIG##2##3f##), similar to nonphotoperiodic Siberian hamsters [##REF##3746726##10##,##REF##3351788##11##]. However, HSD rats produce 6-sulfatoxymelatonin excretion rhythms that adjust to night length and are similar to those of F344 rats in SD [##REF##16162292##32##], and HSD rats successfully decompress activity to a long-duration pattern when transferred from LD to SD (Figure ##FIG##2##3b##). We have not tested directly the ability of HSD rats to phase delay, but successful entrainment to SD suggests that HSD rats have the ability to phase delay to achieve entrainment to SD. Thus, while HSD rats show some similarities to nonphotoperiodic Siberian hamsters, the evidence from this study, together with results of Price et al. [##REF##16162292##32##], suggests that the circadian system of HSD rats is able to respond appropriately to photoperiodic information (this study) to cause a short-day melatonin secretion pattern [##REF##16162292##32##], but HSD rats are unable to respond reproductively to a short-day melatonin pattern [##REF##16162292##32##].</p>", "<p>In combination with the companion study [##REF##16162292##32##], our results support the hypothesis that differences in how F344 and HSD rats secrete and/or respond to melatonin are responsible for differences in photoresponsiveness. We found significant differences between strains in circadian aspects of wheel running, and these differences indicate that HSD rats tend to have more activity during the day, have a greater tendency toward arrhythmia, and differ in some circadian parameters from F344 rats. However, the differences reported here in circadian parameters and nocturnality do not prevent similar timing of 6-sulfatoxymelatonin excretion rhythms in F344 and HSD, albeit potentially lower in amplitude or even arrhythmic in HSD rats [##REF##16162292##32##]. Rats are normally nocturnal [e.g., [##REF##2065757##22##,##REF##10198388##24##]], and our results suggest that F344 rats have the more species-typical pattern of strong nocturnality, while HSD rats differ from that pattern. Variation in activity and 6-sulfatoxymelatonin excretion has also been reported between other strains of rats [##REF##10198388##24##], but our results suggest there is no necessary relationship between photoresponsiveness and circadian rhythm characteristics in rats. For example, some of the strain variation we report in wheel running activity might be related to restlessness, fearfulness, or intrinsic rewards of wheel running that differ among rat strains.</p>" ]

[ "<title>Conclusion</title>", "<p>In laboratory populations of rodents, circadian variation has been reported to be a cause of variation in responsiveness to photoperiod [##REF##3746726##10##, ####REF##3351788##11##, ##REF##9376641##12##, ##UREF##0##13##, ##UREF##1##14####1##14##]. In contrast, in HSD and F344 rats as well as in wild-source populations of mammals, variation in photoresponsiveness may not be linked to variation in circadian characteristics [##REF##2624861##16##, ####UREF##2##17##, ##UREF##3##18##, ##REF##3946637##19##, ##REF##9151428##20####9151428##20##]. Uncoupling of the circadian system and seasonal rhythms can be caused experimentally [##REF##17185605##37##] or may occur naturally during part of the year in polar regions [##REF##16371996##38##]. Natural populations may be under strong selection against circadian mutations that have pleiotropic effects on traits such as seasonality. In these natural populations, mutations that act specifically on outputs, such as melatonin sensitivity of the reproductive axis, pelage, body mass or feeding, are more likely to be adaptive or neutral. In contrast, laboratory populations are not under selection to eliminate mutations that alter circadian clock function in ways that affect multiple output systems, including seasonal photoperiodic responses [e.g., [##REF##8612567##15##]], which may explain differences reported previously between wild-source and laboratory populations.</p>", "<p>The pattern of differences in photoresponsiveness between HSD and F344 rats, which appear to have little relationship to differences in circadian organization, may indicate greater similarity to wild populations containing natural genetic variation in photoresponsiveness. However, we cannot entirely rule out the possibility that HSD rats are nonphotoresponsive due to differences from F344 rats in circadian organization. Knock-out studies of single circadian clock genes in mice have been reported to cause changes in rhythm parameters, reduced nocturnality or loss of rhythmicity, and also changes in reproductive traits [##REF##16940279##4##]. Nevertheless, it appears more likely that differences between HSD and F344 rats in running wheel activity rhythms, nocturnality, and reproductive photoresponsiveness are merely incidentally correlated. Differences in activity and nocturnality may be caused by subtle differences in circadian organization, while differences in photoresponsiveness are likely due to an independent trait, insensitivity of the reproductive axis, body mass, and food intake to a short-day melatonin pattern [##REF##16162292##32##].</p>" ]

[ "<p>This is an Open Access article distributed under the terms of the Creative Commons Attribution License (<ext-link ext-link-type=\"uri\" xlink:href=\"http://creativecommons.org/licenses/by/2.0\"/>), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.</p>", "<title>Background</title>", "<p>Variation in circadian rhythms and nocturnality may, hypothetically, be related to or independent of genetic variation in photoperiodic mediation of seasonal changes in physiology and behavior. We hypothesized that strain variation in photoperiodism between photoperiodic F344 rats and nonphotoperiodic Harlan Sprague Dawley (HSD) rats might be caused by underlying variation in clock function. We predicted that HSD rats would have more activity during the day or subjective day, longer free-running rhythms, poor entrainment to short day length, and shorter duration of activity, traits that have been associated with nonphotoperiodism in other laboratory rodent species, relative to F344 rats. An alternative hypothesis, that differences are due to variation in melatonin secretion or responses to melatonin, predicts either no such differences or inconsistent combinations of differences.</p>", "<title>Methods</title>", "<p>We tested these predictions by examining activity rhythms of young male F344 and HSD rats given access to running wheels in constant dark (DD), short day length (L8:D16; SD), and long day length (L16:D8; LD). We compared nocturnality (the proportion of activity during night or subjective night), duration of activity (alpha), activity onset and offset, phase angle of entrainment, and free running rhythms (tau) of F344 and HSD rats.</p>", "<title>Results</title>", "<p>HSD rats had significantly greater activity during the day, were sometimes arrhythmic in DD, and had significantly longer tau than F344 rats, consistent with predictions. However, HSD rats had significantly longer alpha than F344 rats and both strains entrained to SD, inconsistent with predictions.</p>", "<title>Conclusion</title>", "<p>The ability of HSD rats to entrain to SD, combined with longer alpha than F344 rats, suggests that the circadian system of HSD rats responds correctly to SD. These data offer best support for the alternative hypothesis, that differences in photoresponsiveness between F344 and HSD rats are caused by non-circadian differences in melatonin secretion or the response to melatonin.</p>" ]

[ "<title>Competing interests</title>", "<p>The authors declare that they have no competing interests.</p>", "<title>Authors' contributions</title>", "<p>CDS participated in design of the study, carried out the data collection and preliminary data analyses, and wrote initial drafts of the manuscript. CEJ helped complete data analysis and helped to draft the manuscript. PDH conceived the study, participated in design of the study and data analysis, and developed the final draft of the manuscript. All authors read and approved the final manuscript.</p>" ]

[ "<title>Acknowledgements</title>", "<p>Work was supported by NIH Grant R15 DK51334 to PH and by a Minor Research Grant from a Howard Hughes Medical Institute Undergraduate Sciences Education Program grant to the College of William &amp; Mary.</p>" ]

[ "<fig position=\"float\" id=\"F1\"><label>Figure 1</label><caption><p><bold>Representative actograms of F344 and HSD rats.</bold> (a) F344 rat in LD, (b) HSD rat in LD, (c) F344 rat in SD, (d) HSD rat in SD, (e) F344 rat in DD, (f), HSD rat in DD, (g) arrhythmic HSD rat in DD.</p></caption></fig>", "<fig position=\"float\" id=\"F2\"><label>Figure 2</label><caption><p><bold>Proportion of rats active in relation to time of day (Mean +/- SEM).</bold> (a) F344 in LD, (b) HSD in LD, (c) F344 in SD, (d) HSD in SD, (e) F344 in constant dark, and (f) HSD in constant dark. The profiles show the proportion of individual rats active on the running wheel during each hour, averaged over the final four days of activity monitoring. For rats in constant dark, activity was fit to the 24 hour profile by placing the zero time point at the midpoint of activity.</p></caption></fig>", "<fig position=\"float\" id=\"F3\"><label>Figure 3</label><caption><p><bold>Circadian parameters for F344 (F) and HSD (H) rats in LD, SD, and DD (Mean +/- SEM).</bold> (a) nocturnality index, (b) <italic>alpha</italic>, the duration of activity, (c) activity onset, with dashed lines indicating the time of lights off, (d) activity offset, with dashed lines indicating the time of lights on, (e) phase angle of entrainment, and (f) <italic>tau</italic>, the free-running rhythm in DD. Asterisks indicate significant differences between F344 and HSD rats.</p></caption></fig>" ]

[]

[]

[]

[]

[]

[]

[]

[]

[ "<graphic xlink:href=\"1740-3391-6-8-1\"/>", "<graphic xlink:href=\"1740-3391-6-8-2\"/>", "<graphic xlink:href=\"1740-3391-6-8-3\"/>" ]

[]

{ "acronym": [], "definition": [] }

[ "<title>Background</title>", "<p>The Hedgehog (Hh) signaling pathway is a critical regulator of diverse biological processes including developmental patterning and organogenesis. The pathway is initiated upon Hh ligand binding to the transmembrane receptor Patched (Ptc1). This relieves the Ptc1-mediated suppression of Smoothened (Smo), triggering a complex downstream signal cascade [Reviewed in [##REF##11731473##1##]]. <italic>Gli1 </italic>and <italic>Ptc1 </italic>are conserved Hh target genes and their expression levels are considered reliable indicators of pathway activity. Most biological effects of Hh signaling appear to be mediated through transcriptional regulation of Hh target genes, although a non-transcriptional response was recently identified [##REF##17884337##2##,##REF##18292210##3##].</p>", "<p>Null mouse models have been critical in determining the role of Hh signaling in the growth and morphogenesis of tissues and organs. These models have also proved valuable in gleaning the function of individual Hh signal mediators in pathway regulation. In cell-based assays, <italic>Gli1 </italic>over-expression has been found to induce Hh target gene expression. The finding that <italic>Gli1</italic>^-/-mice develop normally, [##REF##10725236##4##] however, infers that Gli1 function is dispensable for normal development. <italic>Gli2</italic>^-/-mice exhibit neural tube defects and demonstrate diminished Hh target gene expression in several tissues [##REF##9636069##5##, ####REF##9655799##6##, ##REF##16707121##7####16707121##7##]. This supports findings from cell-based assays [##REF##10433919##8##] that Gli2 functions as a critical target gene activator. Increased target gene expression in tissues derived from <italic>Gli3 </italic>null mice relative to tissues from wild type mice [##REF##14602680##9##,##REF##16396903##10##] suggests that Gli3 functions to repress transcription.</p>", "<p>Numerous studies have utilized transgenic MEFs to investigate diverse gene and protein properties. However, the experimental utility of primary cells is limited by a finite propagation and culture period. We showed previously that mouse embryonic fibroblasts (MEFs) from Gli null mice provide a tractable cell-based system in which to quantitatively examine the regulation of Hh target gene expression by the Gli transcription factors [##REF##16571352##11##]. We now describe the generation of immortalized Gli null MEFs (iMEFs) and characterize their transcriptional and migratory response to Hh ligand stimulation.</p>" ]

[ "<title>Methods</title>", "<title>Animals</title>", "<p>This work was conducted with the approval of the University of Wisconsin Animal Care and Use Committee. <italic>Gli1</italic>^{<italic>zfd </italic>}and <italic>Gli2</italic>^{<italic>zfd </italic>}mice were generously provided by Alexandra Joyner and maintained on an outbred CD-1 background. <italic>Gli3</italic>^{<italic>Xtj </italic>}mice were obtained from Jackson Laboratories (Bar Harbor, ME) and were maintained on a C57/C3H background. Primary MEFs [##REF##16571352##11##] were derived from crosses of <italic>Gli1</italic>^{<italic>zfd </italic>}and <italic>Gli2</italic>^{<italic>zfd </italic>}transgenic mice and <italic>Gli3</italic>^{<italic>Xtj </italic>}mutant mice. <italic>Gli1</italic>^{<italic>zfd </italic>}and <italic>Gli2</italic>^{<italic>zfd </italic>}transgenic mice were produced by homologous recombination replacing exons 2–5 and 3–5, respectively with neo cassettes [##REF##10725236##4##,##REF##9006072##12##]. <italic>Gli3</italic>^{<italic>Xtj </italic>}mutant mice from Jackson Laboratories (Bar Harbor, ME) lack <italic>Gli3 </italic>expression due to a deletion mutation in the 3' end of the gene [##REF##8387379##13##].</p>", "<title>Cell immortalization</title>", "<p>Primary MEFs were grown as described previously (Lipinski et al., 2006) in 10% fetal calf serum (FCS) DMEM [with L-glutamine, 4.5 g/L glucose, without sodium pyruvate] with 1% Pen/Strep and propagated following the 3T3 protocol for spontaneous immortalization [##REF##13985244##14##]. 3.0 × 10⁵cells in 4.0 mls media were plated in 60 mm plates and passed at three day intervals. After 8–12 passes, proliferation rates decreased and cells were allowed to grow to confluence before subsequent passing. After 15–25 passes, proliferation rates increased, suggesting spontaneous immortalization. Following, cells were grown for an additional 10–12 passes to ensure stable immortalization. The absence of expression of <italic>Gli1</italic>, <italic>Gli2</italic>, and <italic>Gli3 </italic>in corresponding null iMEF cell lines was confirmed by Real Time-RT-PCR of isolated cDNA [##REF##16571352##11##] as well as standard genotyping of genomic DNA [##REF##10725236##4##,##REF##9006072##12##].</p>", "<title>Generation of stable over-expresser cell lines by retroviral gene delivery</title>", "<p>A pIRES shuttle vector carrying coding sequences for <italic>hShh</italic>, <italic>hGli1</italic>, <italic>ΔNmGli2</italic>, <italic>hSmo* </italic>and independently translated GFP [##REF##17296441##15##] was used to retrovirally infect WT iMEFs. iMEFs were plated at subconfluence in DMEM with 10% FCS 100 mm plates. Cells were then incubated with viral-conditioned media at 4°C for 6 hrs. Following a 72 hr propagation period, GFP-sorting was used to isolate over-expressing populations.</p>", "<title>Cell treatment and Real Time RT-PCR</title>", "<p>iMEFs were plated in Multiwell Primaria™ 24 well plates (Falcon, Franklin Lakes, NJ) at 2.0 × 10⁵cells per well in 400 μl media. Cells were allowed to attach overnight and media were replaced with DMEM containing 1% FCS ± 1 nM octylated Shh peptide (Curis/Genentech). At 24 hrs RNA was harvested and gene expression was determined by real time RT-PCR as described, [##REF##16571352##11##] using gene specific primers as listed: <italic>GAPDH</italic>: 5'-AGCCTCGTCCCG TAGACAAAAT-3' and 5'-CCGTGAGTG GAGTCATACTGGA-3', <italic>Ptc1</italic>: 5'-CTCTGGAGCAGATTTCCAAGG-3' and 5'-TGCCGCAGTTCTTTTGAATG-3', <italic>Gli1</italic>: 5'-GGAAGTCCTATTCACGCCTTGA-3' and 5'-CAACCTTCTTGCTCACACATG TAAG-3', <italic>Gli2</italic>: 5'-CCTTCTCCAATGCCT CAGAC-3' and 5'-GGGGTCTGTGTACCT CTTGG-3', <italic>Gli3</italic>: 5'-AGCCCAAGTATTATT CAGAACCTTTC-3' and 5'-ATGGATAGG GATTGGGAATGG-3'.</p>", "<title>Migration assays</title>", "<p>Cells were grown to 70% confluence in 6-well plates and labeled with 10 μM CellTracker Green (Invitrogen) in serum-free medium for 1 hr. The dye was fixed by adding 10% FCS for 1 h, and subsequently cells were washed and detached with 5 mM ethylenediaminetetraacetic acid (EDTA) in PBS. After complete detachment, cells were resuspended in serum-free medium, pipetted through a 70 μM cell strainer (BD Falcon, Franklin Lakes, NJ), and 100 μl suspension was transferred to 8 μM pore size HTS FluoroBlok Cell Culture Inserts from BD Falcon which were inserted in fitting 24-well plates. In the bottom wells, 600 μl medium was supplemented with indicated chemoattractant. Promptly, fluorescence values representing the number of cells on the bottom side of the insert were read four times every two minutes on a Series 4000 CytoFluor Multi-Well Plate Reader (Perseptive Biosystems, Framingham, MA). The raw fluorescence data were corrected for background fluorescence. No-attractant controls were subtracted at each measured time point to correct for any effects not due to active migration to the chosen attractant. Migration start points were set to zero. For comparison of the different cell lines from multiple experiments, total migration of wild type cells was set to one. For migration assays, Shh peptide (R&amp;D Systems) was used at given concentrations. Preincubation with inhibitors was performed during 10 minutes following detachment and inhibitors were also added to the bottom wells to exclude chemorepellent artifacts. Transfection of iMEFs with <italic>SuFu </italic>overexpression construct (a kind gift of Dr. Toftgård) was performed with Effectene transfection reagent (Qiagen, Hilden, Germany) according to manufacturer's recommendations 16 h before start of migration assay. Western blot analysis revealed a 10-fold increase in SuFu levels following transfection (not shown).</p>" ]

[ "<title>Results and discussion</title>", "<title>MEF immortalization, morphological characterization, and ploidy analysis</title>", "<p><italic>Gli3</italic>^+/+(WT), <italic>Gli1</italic>^-/-, <italic>Gli2</italic>^-/-, <italic>Gli3</italic>^-/-, <italic>Gli1</italic>^-/-<italic>2</italic>^-/-, and <italic>Gli2</italic>^-/-<italic>3</italic>^-/-primary MEFs were propagated by described 3T3 protocols for spontaneous immortalization [##REF##13985244##14##]. Each non-clonal immortalized cell line demonstrated a fibroblast-like morphological appearance in monolayer culture although individual lines exhibited subtle morphological differences (Figure ##FIG##0##1##). Each iMEF line was determined to be tetraploid by flow cytometry analysis (data not shown).</p>", "<title>Characterization of iMEF transcriptional Hh responsiveness</title>", "<p>iMEFs were treated ± Shh ligand, and Hh target gene expression was determined by real time RT-PCR. Figure ##FIG##1##2A## shows the expression of reliable Hh target gene <italic>Ptc1 </italic>following stimulation with Shh or vehicle and Figure ##FIG##1##2B## shows the fold induction (Shh/Veh) of <italic>Ptc1 </italic>expression. <italic>Gli3</italic>^-/-iMEFs demonstrated elevated basal and Shh-induced expression of <italic>Ptc1 </italic>(p = 0.03 and p = 0.02 respectively) relative to WT cells. Shh ligand stimulation induced <italic>Ptc1 </italic>expression in each iMEF line except that lacking expression of both <italic>Gli2 </italic>and <italic>Gli3</italic>, which are essential for a transcriptional Hh response. While loss of Gli1 alone had no effect on target gene expression, <italic>Ptc1 </italic>induction was reduced in both <italic>Gli2</italic>^-/-and <italic>Gli1</italic>^-/-<italic>2</italic>^-/-iMEFs relative to WT.</p>", "<title>Characterization of iMEF non-transcriptional Hh responsiveness</title>", "<p>While Hh signaling effects are thought to be exerted primarily through transcriptional regulation, a novel pathway was recently identified which is Smo-dependent but does not require transcription [##REF##17884337##2##,##REF##18292210##3##]. This alternative pathway triggers cytoskeletal rearrangement, driving a cellular migratory response toward Hh ligand. When activation of this pathway was investigated in wild type iMEF cells, a dose-dependent migratory response to recombinant Shh was observed (Figure ##FIG##2##3A##). In the absence of Shh ligand (no-attractant control), a low level of baseline migration was observed and subsequently subtracted from the migratory responses in all other experiments.</p>", "<p>The generated Gli null iMEFs provide a valuable tool to assess transcription factor dependence of specific biological responses. When <italic>Gli3</italic>^+/+(WT), <italic>Gli1</italic>^-/-<italic>2</italic>^-/-, and <italic>Gli2</italic>^-/-<italic>3</italic>^-/-iMEF cells were allowed to migrate to 2 μg/ml Shh peptide, migration was observed for each genotype (Figure ##FIG##2##3B##). Remarkably, increased migration to Shh was observed for the null cells (<italic>Gli1</italic>^-/-<italic>2</italic>^-/-and <italic>Gli2</italic>^-/-<italic>3</italic>^-/-) that was inversely correlated with their respective transcriptional Hh-responsiveness (Figure ##FIG##2##3C##). This inverse correlation may be explained by competition for shared pathway components between the two different signal transduction mechanisms but no data to suggest such a competition have so far been presented.</p>", "<p>To confirm that the observed migration of <italic>Gli2</italic>^-/-<italic>3</italic>^-/-iMEFs to Shh peptide is Hh-pathway specific, we used the Smo agonist purmorphamine as chemoattractant. Robust migration was observed that was comparable in magnitude to migration to Shh as well as the positive control, FCS (Figure ##FIG##3##4A##). Conversely, treatment with the Smo inhibitor cyclopamine abrogated the migration response to Shh peptide. These data indicate that the migratory response of <italic>Gli2</italic>^-/-<italic>3</italic>^-/-iMEFs is Smoothened dependent and thus Hh-pathway dependent. An artifactual response to endotoxin contamination of Shh peptide [##REF##18207025##16##] was excluded by demonstrating the inability of the lipopolysaccharide (LPS) inhibitor polymixin B (PMB) to reduce the cellular migratory response to Shh peptide.</p>", "<p>While the Hh-inhibitory protein suppressor of fused (SuFu) mitigates the Gli-mediated Hh transcriptional response, it appears to have no effect on migratory response [##REF##17884337##2##]. Accordingly, when SuFu was overexpressed in <italic>Gli2</italic>^-/-<italic>3</italic>^-/-iMEFs, Shh-induced migration was not changed (Figure ##FIG##3##4B##). The translation independence of the migratory response was confirmed by demonstrating the ineffectiveness of cycloheximide in altering the migration response to Shh peptide.</p>", "<p>To confirm the previously demonstrated requirement for intact leukotriene synthesis machinery in the migratory response to Shh [##REF##17884337##2##,##REF##18292210##3##], we used the lipoxygenase inhibitor MK-886 to block leukotriene production. When <italic>Gli2</italic>^-/-<italic>3</italic>^-/-iMEFs were preincubated with 5 μM MK-886 migration to Shh was markedly reduced, indicating that in cells without a functional transcriptional Hh signaling pathway, leukotriene synthesis is required for Shh-mediated migration (Figure ##FIG##3##4B##).</p>", "<p>As several studies have demonstrated that either Gli2 and Gli3 are required for the Hh signaling transcriptional response [##REF##14602680##9##,##REF##16571352##11##,##REF##15604102##17##], the Hh-induced migration of <italic>Gli2</italic>^-/-<italic>3</italic>^-/-cells is affirmative evidence that the migratory response is independent of Gli transcription factor activity. Also important for this study, the observed migration data indicate that the iMEFs have functional Hh-sensing machinery and that the diminished Hh-responsiveness of <italic>Gli1</italic>^-/-<italic>2</italic>^-/-and <italic>Gli2</italic>^-/-<italic>3</italic>^-/-iMEFs is due to the absence of Gli proteins, rather than ablation of the Ptch1/Smo receptor pair or other artifacts.</p>", "<title>Stable over-expression of Hh components drives constitutive pathway activation</title>", "<p>Immortalized cells allow for retroviral-mediated, stable expression of vectors for gene-knockdown or over-expression. We generated WT iMEFs with stable over-expression of several pathway components and assessed pathway activity by measuring the expression of <italic>Ptc1</italic>, a reliable Hh target gene. We found that iMEFs over-expressing <italic>hShh</italic>, <italic>hGli1</italic>, or constitutively active forms of <italic>mGli2 </italic>(<italic>ΔNmGli2</italic>) or <italic>hSmo </italic>(<italic>Smo*</italic>) demonstrated increased pathway activity relative to iMEFs expressing only GFP (Figure ##FIG##4##5##). This demonstrates that over-expression of pathway components at multiple levels including ligand and transcription factor is sufficient to drive constitutive pathway activity.</p>" ]

[ "<title>Results and discussion</title>", "<title>MEF immortalization, morphological characterization, and ploidy analysis</title>", "<p><italic>Gli3</italic>^+/+(WT), <italic>Gli1</italic>^-/-, <italic>Gli2</italic>^-/-, <italic>Gli3</italic>^-/-, <italic>Gli1</italic>^-/-<italic>2</italic>^-/-, and <italic>Gli2</italic>^-/-<italic>3</italic>^-/-primary MEFs were propagated by described 3T3 protocols for spontaneous immortalization [##REF##13985244##14##]. Each non-clonal immortalized cell line demonstrated a fibroblast-like morphological appearance in monolayer culture although individual lines exhibited subtle morphological differences (Figure ##FIG##0##1##). Each iMEF line was determined to be tetraploid by flow cytometry analysis (data not shown).</p>", "<title>Characterization of iMEF transcriptional Hh responsiveness</title>", "<p>iMEFs were treated ± Shh ligand, and Hh target gene expression was determined by real time RT-PCR. Figure ##FIG##1##2A## shows the expression of reliable Hh target gene <italic>Ptc1 </italic>following stimulation with Shh or vehicle and Figure ##FIG##1##2B## shows the fold induction (Shh/Veh) of <italic>Ptc1 </italic>expression. <italic>Gli3</italic>^-/-iMEFs demonstrated elevated basal and Shh-induced expression of <italic>Ptc1 </italic>(p = 0.03 and p = 0.02 respectively) relative to WT cells. Shh ligand stimulation induced <italic>Ptc1 </italic>expression in each iMEF line except that lacking expression of both <italic>Gli2 </italic>and <italic>Gli3</italic>, which are essential for a transcriptional Hh response. While loss of Gli1 alone had no effect on target gene expression, <italic>Ptc1 </italic>induction was reduced in both <italic>Gli2</italic>^-/-and <italic>Gli1</italic>^-/-<italic>2</italic>^-/-iMEFs relative to WT.</p>", "<title>Characterization of iMEF non-transcriptional Hh responsiveness</title>", "<p>While Hh signaling effects are thought to be exerted primarily through transcriptional regulation, a novel pathway was recently identified which is Smo-dependent but does not require transcription [##REF##17884337##2##,##REF##18292210##3##]. This alternative pathway triggers cytoskeletal rearrangement, driving a cellular migratory response toward Hh ligand. When activation of this pathway was investigated in wild type iMEF cells, a dose-dependent migratory response to recombinant Shh was observed (Figure ##FIG##2##3A##). In the absence of Shh ligand (no-attractant control), a low level of baseline migration was observed and subsequently subtracted from the migratory responses in all other experiments.</p>", "<p>The generated Gli null iMEFs provide a valuable tool to assess transcription factor dependence of specific biological responses. When <italic>Gli3</italic>^+/+(WT), <italic>Gli1</italic>^-/-<italic>2</italic>^-/-, and <italic>Gli2</italic>^-/-<italic>3</italic>^-/-iMEF cells were allowed to migrate to 2 μg/ml Shh peptide, migration was observed for each genotype (Figure ##FIG##2##3B##). Remarkably, increased migration to Shh was observed for the null cells (<italic>Gli1</italic>^-/-<italic>2</italic>^-/-and <italic>Gli2</italic>^-/-<italic>3</italic>^-/-) that was inversely correlated with their respective transcriptional Hh-responsiveness (Figure ##FIG##2##3C##). This inverse correlation may be explained by competition for shared pathway components between the two different signal transduction mechanisms but no data to suggest such a competition have so far been presented.</p>", "<p>To confirm that the observed migration of <italic>Gli2</italic>^-/-<italic>3</italic>^-/-iMEFs to Shh peptide is Hh-pathway specific, we used the Smo agonist purmorphamine as chemoattractant. Robust migration was observed that was comparable in magnitude to migration to Shh as well as the positive control, FCS (Figure ##FIG##3##4A##). Conversely, treatment with the Smo inhibitor cyclopamine abrogated the migration response to Shh peptide. These data indicate that the migratory response of <italic>Gli2</italic>^-/-<italic>3</italic>^-/-iMEFs is Smoothened dependent and thus Hh-pathway dependent. An artifactual response to endotoxin contamination of Shh peptide [##REF##18207025##16##] was excluded by demonstrating the inability of the lipopolysaccharide (LPS) inhibitor polymixin B (PMB) to reduce the cellular migratory response to Shh peptide.</p>", "<p>While the Hh-inhibitory protein suppressor of fused (SuFu) mitigates the Gli-mediated Hh transcriptional response, it appears to have no effect on migratory response [##REF##17884337##2##]. Accordingly, when SuFu was overexpressed in <italic>Gli2</italic>^-/-<italic>3</italic>^-/-iMEFs, Shh-induced migration was not changed (Figure ##FIG##3##4B##). The translation independence of the migratory response was confirmed by demonstrating the ineffectiveness of cycloheximide in altering the migration response to Shh peptide.</p>", "<p>To confirm the previously demonstrated requirement for intact leukotriene synthesis machinery in the migratory response to Shh [##REF##17884337##2##,##REF##18292210##3##], we used the lipoxygenase inhibitor MK-886 to block leukotriene production. When <italic>Gli2</italic>^-/-<italic>3</italic>^-/-iMEFs were preincubated with 5 μM MK-886 migration to Shh was markedly reduced, indicating that in cells without a functional transcriptional Hh signaling pathway, leukotriene synthesis is required for Shh-mediated migration (Figure ##FIG##3##4B##).</p>", "<p>As several studies have demonstrated that either Gli2 and Gli3 are required for the Hh signaling transcriptional response [##REF##14602680##9##,##REF##16571352##11##,##REF##15604102##17##], the Hh-induced migration of <italic>Gli2</italic>^-/-<italic>3</italic>^-/-cells is affirmative evidence that the migratory response is independent of Gli transcription factor activity. Also important for this study, the observed migration data indicate that the iMEFs have functional Hh-sensing machinery and that the diminished Hh-responsiveness of <italic>Gli1</italic>^-/-<italic>2</italic>^-/-and <italic>Gli2</italic>^-/-<italic>3</italic>^-/-iMEFs is due to the absence of Gli proteins, rather than ablation of the Ptch1/Smo receptor pair or other artifacts.</p>", "<title>Stable over-expression of Hh components drives constitutive pathway activation</title>", "<p>Immortalized cells allow for retroviral-mediated, stable expression of vectors for gene-knockdown or over-expression. We generated WT iMEFs with stable over-expression of several pathway components and assessed pathway activity by measuring the expression of <italic>Ptc1</italic>, a reliable Hh target gene. We found that iMEFs over-expressing <italic>hShh</italic>, <italic>hGli1</italic>, or constitutively active forms of <italic>mGli2 </italic>(<italic>ΔNmGli2</italic>) or <italic>hSmo </italic>(<italic>Smo*</italic>) demonstrated increased pathway activity relative to iMEFs expressing only GFP (Figure ##FIG##4##5##). This demonstrates that over-expression of pathway components at multiple levels including ligand and transcription factor is sufficient to drive constitutive pathway activity.</p>" ]

[ "<title>Conclusion</title>", "<p>The full complement of Gli genes in most Hh ligand-responsive cell models mitigates their utility in investigations of molecular regulation and biological activity of the individual Gli transcription factors. Here we demonstrated the unique transcriptional and non-transcriptional responses of a battery of Gli-null iMEFs. Moving forward, these cell lines should prove a useful tool in a wide range of the Hh signaling field and have already been distributed to several investigators for a wide range of purposes including; studies of transcriptional co-regulators and Gli-binding partners; chemical pathway inhibitor site of action studies; anti-Gli antibody specificity studies; and several studies of Gli dependence in specific Hh-related biological function.</p>" ]

[ "<p>This is an Open Access article distributed under the terms of the Creative Commons Attribution License (<ext-link ext-link-type=\"uri\" xlink:href=\"http://creativecommons.org/licenses/by/2.0\"/>), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.</p>", "<title>Background</title>", "<p>Hedgehog (Hh) signaling is a conserved morphogenetic pathway which plays critical roles in embryonic development, with emerging evidence also supporting a role in healing and repair processes and tumorigenesis. The Gli family of transcription factors (Gli1, 2 and 3) mediate the Hedgehog morphogenetic signal by regulating the expression of downstream target genes. We previously characterized the individual and cooperative roles of the Gli proteins in Hh target gene regulation using a battery of primary embryonic fibroblasts from Gli null mice.</p>", "<title>Results</title>", "<p>Here, we describe the establishment of spontaneously immortalized mouse embryonic fibroblast (iMEF) cell lines lacking single and multiple Gli genes. These non-clonal cell lines recapitulate the unique ligand mediated transcriptional response of primary MEFs. While loss of Gli1 had no effect on target gene induction, Gli2 null cells demonstrated reduced target gene induction while Gli3 null cells exhibited elevated basal and ligand-induced expression. Target gene response in <italic>Gli1</italic>^-/-<italic>2</italic>^-/-iMEFs was severely reduced while <italic>Gli2</italic>^-/-<italic>3</italic>^-/-iMEFs were incapable of ligand-induced transcriptional response. However, we found that both <italic>Gli1</italic>^-/-<italic>2</italic>^-/-and <italic>Gli2</italic>^-/-<italic>3</italic>^-/-iMEFs exhibited robust leukotriene synthesis-dependent migration responses to Hh ligand, demonstrating that this response is not transcriptionally-dependent.</p>", "<title>Conclusion</title>", "<p>This study provides fundamental characterizations of the transcriptional and non-transcriptional Hh responsiveness of a battery of Gli-null iMEFs. Moving forward, these cell lines should prove a valuable tool set to study the unique functional regulation of the Gli proteins in a Hh-responsive cell-type.</p>" ]

[ "<title>Authors' contributions</title>", "<p>RJL and DJP isolated primary cells, established immortalized cell lines and confirmed proper genotypes. JJG performed transcriptional response assays and MFB performed migration assays. Each was responsible for data acquisition and analysis. WB participated in the design and interpretation of experiments. All authors contributed to preparation of the manuscript and have approved the final form.</p>" ]

[ "<title>Acknowledgements</title>", "<p>The authors wish to thank Curis/Genentech for their kind gift of Shh peptide. We also thank Alejandro Muñoz-del-Rio for assistance with statistical analysis and Mary Bushman for critical review of the manuscript. RJL was supported by National Institutes of Health (DK065303-03); National Institute of Environmental Health Sciences (T32-ES00715). This work was in part supported by National Institutes of Health (DK056238).</p>" ]

[ "<fig position=\"float\" id=\"F1\"><label>Figure 1</label><caption><p><bold>Gli-null iMEF morphology in monolayer cell culture</bold>. Indicated iMEFs were grown to confluence in monolayer culture and imaged at 40× magnification.</p></caption></fig>", "<fig position=\"float\" id=\"F2\"><label>Figure 2</label><caption><p><bold>Transcriptional Hh-responsiveness of generated iMEFs</bold>. Indicated iMEFs were plated at confluence and treated ± Shh ligand. After 24 hrs, expression of <italic>Ptc1 </italic>was determined by Real-Time RT-PCR. A. Basal and Shh-induced expression of <italic>Ptc1</italic>. Values represent the mean ± SEM of 3–5 replicate experiments, * indicates P ≤ 0.05 (paired t-test). B. <italic>Ptc1 </italic>expression plotted as fold induction (Shh/Veh). Values represent the mean ± SEM of three replicate experiments. The letters above the bars denote the groups produced by the ANOVA pair-wise differences. Genotypes sharing a letter are not statistically significant at p ≤ 0.05 (Fisher's LSD).</p></caption></fig>", "<fig position=\"float\" id=\"F3\"><label>Figure 3</label><caption><p><bold>Non-transcriptional Hh-responsiveness of the generated iMEFs</bold>. A. Example of a migration assay using wild type (<italic>Gli3</italic>^+/+) iMEFs in a Transwell system with varying concentrations of Shh as chemoattractant. Fluorescence was read every two minutes and expressed as relative fluorescence unit (RFU). B. Example of a migration assay using wild type (<italic>Gli3</italic>^+/+), <italic>Gli1</italic>^-/-<italic>2</italic>^-/-and <italic>Gli2</italic>^-/-<italic>3</italic>^-/-iMEFs with 2 μg/ml Shh as chemoattractant. No-attractant condition was subtracted and migration starting points were set to t = 0. Robust migration was observed for each cell line. C. Total migration data from several experiments as performed for B, pooled and expressed as fraction of wild type iMEF migration (<italic>Gli3</italic>^+/+, set to 1, n = 5). To determine whether the migration response was significantly different relative to wild type, a 95% confidence interval was calculated based on the mean and standard deviation of the observations. Reported significant differences thus have a P value of &lt; 0.05.</p></caption></fig>", "<fig position=\"float\" id=\"F4\"><label>Figure 4</label><caption><p><bold>Hedgehog pathway specificity in migratory response of Gli2^-/-3^-/-iMEFs</bold>. A. Migration of <italic>Gli2</italic>^-/-<italic>3</italic>^-/-iMEFs was performed using the Smo agonist purmorphamine or 20% FCS as attractants as well as 2 μg/ml Shh in the presence of no inhibitor (control), the Smo antagonist cyclopamine or the LPS inhibitor PMB. Preincubation time with inhibitors following detachment was 10 minutes. B. Migration responses of <italic>Gli2</italic>^-/-<italic>3</italic>^-/-iMEFs transfected with a SuFu or control vector or preincubated with the leukotriene synthesis inhibitor MK-886 or translation inhibitor, cycloheximide. No-attractant condition was subtracted and migration starting points were set to t = 0.</p></caption></fig>", "<fig position=\"float\" id=\"F5\"><label>Figure 5</label><caption><p><bold>Stable over-expression of several Hh components drives constitutive pathway activity</bold>. WT iMEFs were infected with retrovirus encoding, <italic>hShh-GFP</italic>, <italic>hGli1-GFP</italic>, <italic>ΔNmGli2-GFP</italic>, <italic>hSmo-GFP </italic>or an empty <italic>GFP </italic>IRIS vector. Stable over-expresser cell lines produced by GFP sorting were plated at confluence. Following 24 hrs, expression of <italic>Ptc1 </italic>was determined by Real-Time RT-PCR. Values represent the mean ± SEM of three replicate experiments, * indicates P ≤ 0.05 (paired t-test) vs. GFP iMEFs.</p></caption></fig>" ]

[]

[]

[]

[]

[]

[]

[]

[]

[ "<graphic xlink:href=\"1471-2121-9-49-1\"/>", "<graphic xlink:href=\"1471-2121-9-49-2\"/>", "<graphic xlink:href=\"1471-2121-9-49-3\"/>", "<graphic xlink:href=\"1471-2121-9-49-4\"/>", "<graphic xlink:href=\"1471-2121-9-49-5\"/>" ]

[]

{ "acronym": [], "definition": [] }

[ "<title>Background</title>", "<p>There is a widespread belief that there is a high discordance between clinical and radiographic knee osteoarthritis (OA) [##REF##1877852##1##, ####REF##2287948##2##, ##REF##2241266##3####2241266##3##]. However, this belief contrasts with the assumption that osteoarthritis is the commonest knee pathology in older people and the commonest reason for knee pain and disability in this age-group, and that radiographs appropriately identify moderate and severe osteoarthritis. Therapeutic options such as surgery for knee pain are considered in the presence of radiographic abnormalities [##UREF##0##4##]. In previous work we have shown that the presence of radiographic knee osteoarthritis can influence the decision of general practitioners in their management strategies, particularly leading to increased levels of referral to secondary care [##REF##12695159##5##]. It is therefore important to understand the apparent lack of association between pain and knee x rays, particularly if the best clinical choices for patients are to be made and the basis for these choices clearly established. There has been no previous review of all studies which have investigated the association between pain and x rays at the knee. This paper seeks to fill that gap.</p>", "<p>The features revealed by a knee x ray serve its main purpose as a diagnostic tool. However, in common with other diagnostic tests, the x ray also supports several other potential applications. These include estimating prognosis, guiding treatment, post therapeutic evaluation, giving reassurance to patient or physician and helping to preserve the doctor-patient relationship [##UREF##1##6##]. One study has indicated that GPs use x rays as a part of their management strategy because they perceive them as being helpful in making management decisions, such as avoiding unnecessary referrals to specialists, and in providing a useful aid to discussing management with patients [##REF##9166050##7##]. However this means that the relationship between x ray findings and clinical complaints is crucial to understand if decisions based on x ray findings are going to appropriately influence what the patient considers important, namely reducing pain and disability.</p>", "<p>As previously stated, population studies have suggested that the 'fit' between x rays and symptoms at the knee is not perfect. This paper describes a systematic search of the literature to identify the extent of these discrepancies and the possible reasons why they might arise. In general there are two possible reasons for the discrepancy. Firstly the way in which radiographic osteoarthritis is defined will affect the number of cases classed as having radiographic disease or not, and therefore the prevalence of radiographic OA disease. In the knee for example the joint has three compartments. If the only x rays considered are the antero-posterior view, then only osteoarthritis in the medial and lateral compartments would be identified and up to 24% of patients with radiographic knee OA would be missed by not visualising the patello-femoral joint [##REF##9893570##8##]. Secondly clinical symptoms and signs may arise from sources other than the contents of the knee joint or the underlying subchondral bone, and so the ways in which the clinical syndrome of osteoarthritis is defined will influence the extent to which it is linked with osteoarthritis defined on a knee X-ray.</p>", "<p>The first objective of this systematic review was to identify studies which provide an estimate of the prevalence of radiographic knee OA in older people with knee pain. The second objective was to determine what influences this prevalence and therefore might be a source of error or variation in the observed associations between x rays and symptoms: namely the definition of x ray osteoarthritis, the definition of symptoms, and the effect of demographic factors such as age and ethnicity.</p>" ]

[ "<title>Methods</title>", "<p>The strategy and keywords for the search are given in appendix I. The first step of the strategy was to identify papers that included reference to knee osteoarthritis in the various forms by which it can be referred to in the literature, and all papers relating to diagnosis and clinical signs in knee OA. The next step was to filter these papers to extract those which included radiographic investigation and to limit them to extract those which concerned population-based observational studies and not intervention studies. Other exclusions at this stage were papers about arthritic conditions other than osteoarthritis, modes of investigation other than x rays, such as MRI, and papers in non-English languages. This search strategy therefore identified papers which combined x rays, diagnosis, clinical signs and symptoms in knee osteoarthritis in population studies.</p>", "<p>Two databases were used, EMBASE and Medline. The initial search identified 134 papers. These were then examined by title to include papers in the review which specifically related to knee pain, knee symptoms, knee x rays or the prevalence of any these factors. This limited the papers to 60. The abstracts for these papers were then assessed to determine if the paper contained at least one x-ray view of the knee and mentioned at least one knee related symptom. Applying these criteria limited the review to 20 papers.</p>", "<title>Analysis</title>", "<p>The first analysis of the results considers papers from which estimates of the prevalence of radiographic OA in people with knee pain can be derived. The second analysis considers those papers from which the prevalence of knee pain can be derived in populations of people defined as having radiographic knee OA.</p>", "<p>The analysis considered three factors that might potentially influence these associations. There are a range of factors that might explain or influence discordance between radiographs and symptoms. We chose three of these – age, gender, ethnicity – to test the hypothesis that discordance might vary between population sub-types.</p>", "<title>1. The nature and extent of radiographic views</title>", "<p>One potential factor that might lead to apparent lack of association between x rays and symptoms is that in the studies conducted, there were insufficient numbers of x rays – for example of persons with very severe pain – to provide the power to detect strong associations overall. To overcome this, we decided to include all radiographic views of the knees used in these papers. The associations of clinical features with different radiographic views were either drawn directly from the results in the paper or were calculated if the raw data allowed. Sensitivities and specificities for the varying views and grades in relation to symptoms were examined. Where appropriate, odds ratios were examined for the relationships.</p>", "<title>2. The definition of symptoms</title>", "<p>A second factor that might be related to the lack of association was the pain itself. Pain comes in many forms, and is an individual experience. However research has to attempt to standardise the approach to this experience so that it can be measured. We therefore explored all levels and definitions of pain. Papers that examined symptoms were included and their definition of pain identified and catalogued. The prevalence of radiographic knee OA in relation to these definitions was identified and used to estimate the proportions of knee pain sufferers who have radiographic OA. Where appropriate, odds ratios were examined for the associations between pain and radiographic OA. Also included were papers which used the Western Ontario and McMaster Universities Arthritis Index (WOMAC) as an alternative methodology for studying the relationship between symptoms and radiographic knee OA [##REF##3068365##9##]. The WOMAC has several sections which cover knee pain, stiffness and function. As an example, the pain-specific subset for the knee assesses the severity of pain during five activities including walking on a flat surface, going up or down stairs, at night while in bed, sitting or lying and standing upright. Each category is assigned a numerical score of 1 to 5, corresponding to the severity of pain (none, mild, moderate, severe and extreme). Such a scale allows for a more detailed analysis of the relationship between symptoms and x rays.</p>", "<title>3. The nature of the study group</title>", "<p>A third important factor which could influence the association within these studies is the population under scrutiny. If this lack of association is real, then it should be true for all groups equally. We have tested this by selecting three population characteristics – age, gender and ethnicity – and investigated whether they should be taken into account in estimating discordance between pain and radiographic knee OA. Papers were included which examined the differences in prevalence of radiographic knee OA and knee symptoms according to age, gender and ethnicity as examples of external factors that might influence the nature and extent of the association between symptoms and radiographic features. Prevalence estimates for knee pain and radiographic knee OA according to age bands and ethnic groups were collated, and odds ratios calculated for the associations.</p>" ]

[ "<title>Results</title>", "<title>Section 1: The prevalence of radiographic osteoarthritis in people with knee pain</title>", "<p>Table ##TAB##0##1## summarises the estimates of prevalence from the studies reviewed of persons with knee pain found to have x ray abnormalities consistent with radiographic knee OA. Knee pain was the most frequent marker symptom and has been used to construct this main table. However other symptoms were reported in the different studies, but definitions varied and could not be used to compare the study results. Pain, by contrast, featured in all the studies and therefore provided a common factor to which to relate estimates of the frequency of radiographic knee osteoarthritis. The figures are shown stratified by age, with the youngest group first and older age groups further down the table. The different radiographic views used are highlighted, as are the definitions used to classify an abnormal radiograph as showing osteoarthritis.</p>", "<p>The proportion of those with knee pain found to have radiographic osteoarthritis ranged from 15–76%. One study which encompassed a wide age range (19 – 92 years) found that 53% of current knee pain sufferers had radiographic knee osteoarthritis [##REF##7654803##10##].</p>", "<title>The x ray view</title>", "<p>Table ##TAB##0##1## indicates the various x ray views employed in the individual studies. The antero-posterior (A/P) view was employed in all except one and in most of the studies the weight-bearing view was used. Additional views of the joint were employed in several of the studies reviewed including lateral (mediolateral), lateral flexed and skyline views (inferosuperior). Which views are used in the varying studies appears to have some impact upon the relationship of pain to radiographic knee OA. Claessens uses only the A/P weight bearing view and identifies 36% of patients with knee pain as having radiographic knee OA [##REF##2241266##3##]. Lanyon uses the A/P weight bearing in conjunction with the lateral and identifies 53% [##REF##9893570##8##]. Cittucini meanwhile observes that 53% of patients with knee pain have radiographic knee OA when using the skyline in isolation [##REF##8806116##11##]. Knee studies that include x rays of the patello-femoral joint (PFJ), improve the sensitivity with which symptoms such as pain can identify radiographic knee OA to a potential 51–67% [##REF##9893570##8##,##REF##1632657##12##,##REF##11520166##13##]. Excluding this view drops the sensitivity to 24–38% [##REF##2241266##3##,##REF##9893570##8##,##REF##9709175##14##]. It appears that discrepancy between knee symptoms such as pain and radiographic knee OA is due in part to not employing x rays of all three compartments of the knee. However this does not explain all the discrepancy, since even when all compartments are x rayed the highest proportion of patients with pain who have radiographic knee OA is 76% [##REF##1540789##15##]. A recent paper from our unit not included in the review suggests that systematically searching all three X-ray views of the knee for evidence of OA in persons over 50 years with knee pain identifies OA in 70% [##UREF##2##16##].</p>", "<title>Grading the x ray</title>", "<p>Grading an x ray entails defining the level of abnormality found in an x ray considered to represent knee osteoarthritis. Increasingly abnormal features may be added to this base level to define increasing severity. Table ##TAB##0##1## shows the x ray knee OA definitions used in the studies. Common to all the studies was the use of osteophytes at some point in the 'baseline' definition. Ciccutini and Lanyon both used Grade 1 osteophytes (minute) [##REF##9893570##8##,##REF##8806116##11##], all the others used grade 2 (definite) or 'definite osteophytes' as the main defining feature. The use of grade 1 versus grade 2 appears to make little difference between studies. However, the association of knee pain and x ray grade was investigated by McAlindon who found a limited but positive correlation between knee pain and x ray grade (Pearson's correlation coefficient = 0.43) [##REF##8484690##17##].</p>", "<p>Classically the Kellgren and Lawrence grading scale has rated joint space narrowing as grade 3 with osteophytes occurring at grade 1 or more. Cicuttini found that knee pain was significantly associated with osteophytes in all x ray views but not with joint space narrowing. As an example in the A/P view the odds ratio for the association of osteophytes and ever having had knee pain (episodes lasting more than 15 days) was stronger and significant (OR 5.0;95% CI 3.01,11.33) when compared with the odds ratio for pain and joint space narrowing alone (OR 2.13; 95% CI 0.78,5.87) [##REF##8806116##11##]. The association of knee pain with osteophytes was also examined by Lanyon who estimated that of knee pain positive subjects, 12% were K/L osteophyte grade 3, whilst 30% were grade 2 or above. When the lowest grade of osteophyte was included (grade 1), 63 % of knee pain sufferers were classified as having radiographic osteoarthritis [##REF##9893570##8##]. Lethbridge also found increased levels of radiographic OA when using more inclusive grades with 53% of current knee pain sufferers having K/L grade 2 or more, but only 22 % of those with pain had K/L grade 3 and above [##REF##7654803##10##]. In addition, Hart, analysed the sensitivity and specificity for the association of knee joint pain with K/L grade 1 or more and compared this to grade 2 or more and found no difference (23% sensitivity, 88% specificity) [##REF##1877852##18##].</p>", "<title>Defining knee symptoms</title>", "<p>Table ##TAB##1##2## demonstrates how the proportion who have radiographic knee OA varies with the definition of knee pain. There are 10 different definitions used. These vary considerably, from 'ever having an episode of pain lasting 15 days or more' [##REF##8806116##11##] to 'knee pain during the past month' [##REF##9709175##14##]. There is corresponding variation in the prevalence of radiographic knee OA depending upon the definition of pain used. Where the definition is one that involves recalled pain over a specific period, such as in Petersson's study, the prevalence is lower (15%) [##REF##9306873##19##] than for pain \"ever\", as in Ciccutini's study (37%) [##REF##8806116##11##] or recent pain as in Odding's (39%) [##REF##9709175##14##]. Even when the same question is used for different studies, a wide variation in prevalence is evident [##REF##9893570##8##,##REF##7654803##10##,##REF##1540789##15##,##REF##8484690##17##]. As table ##TAB##1##2## details, Felson and colleagues used similar definitions of pain to these [##REF##9404469##20##], but only part of them were used to define a patient as knee pain positive. Felson's results indicated only 16% of patients with knee pain had radiographic knee OA, compared with a range of 30–76% in the other studies [##REF##9893570##8##,##REF##7654803##10##,##REF##1540789##15##,##REF##8484690##17##].</p>", "<p>Other studies have employed the WOMAC to examine knee symptoms but in different ways as shown in table ##TAB##2##3##[##UREF##3##21##, ####REF##11922194##22##, ##REF##12784407##23##, ##REF##10955336##24####10955336##24##]. A direct comparison is not possible due to the variation in the definition of a knee pain positive patient, but despite this variation, there is overall no significant difference in WOMAC pain score between knee pain positive patients with radiographic knee OA and those without it.</p>", "<title>The nature of the study group</title>", "<p>Younger age groups with knee pain have a lower prevalence of radiographic knee osteoarthritis than older persons [##REF##11520166##13##,##REF##9306873##19##]. Restricting analysis to persons aged between 40 and 80, the proportion of knee pain sufferers with radiographic osteoarthritis is 19–30% [##REF##1877852##1##,##REF##9893570##8##,##REF##10852280##25##]. For all those aged over 45 this rises to 36–50% [##REF##2241266##3##,##REF##8806116##11##,##UREF##3##21##,##REF##10955336##24##,##REF##8712864##26##], and over 55 the range is 40–76% [##REF##9709175##14##,##REF##1540789##15##,##REF##8484690##17##]. Several studies support the age-related nature of the changes found in radiographic knee osteoarthritis in those with knee pain [##REF##2241266##3##,##REF##7654803##10##,##REF##1632657##12##,##REF##10852280##25##]. As an example, in Hannan's study, the prevalence was 2 % in those aged 25 – 40, clearly less than the 21% estimate among those aged 51–74 [##REF##10852280##25##]. In only one study was this trend not evident [##REF##1294744##27##].</p>", "<p>Two studies investigated the prevalence of radiographic knee osteoarthritis in both Caucasian and African American subjects with knee pain [##REF##11520166##13##,##REF##12784407##23##]. Lachance identified higher levels of radiographic knee osteoarthritis in African American (AA) than Caucasian (CA) women with knee pain (40% vs. 15%) [##REF##11520166##13##]. The overall level of radiographic osteoarthritis was higher for the African Americans (23.2%) compared to the Caucasians (8.5%). The age range in this study was 40–53. In Ang's study of men and women over 50 (average 65 years) [##REF##12784407##23##], the overall prevalence of knee osteoarthritis was similar for both ethnic groups (AA 39.4%; CA 38.7%), but the severity of the K/L grading was significantly higher in the presence of larger osteophytes in African Americans compared to Caucasians. With respect to the sensitivity with which pain could predict radiographic knee osteoarthritis, Lachance demonstrated that this was higher for African American women (51%) compared to Caucasian women (35%), but the specificity for Caucasian women was higher (CA 85%; AA 77%) [##REF##11520166##13##].</p>", "<title>Section 2: The prevalence of knee pain and clinical osteoarthritis in people with radiographic osteoarthritis</title>", "<p>Table ##TAB##3##4## summarises the studies that give estimates of the prevalence of knee pain for specific age groups from a population found to have abnormal knee radiographs. The different radiographic views are highlighted. There is a large variation in the proportion of those with radiographic knee OA who experienced pain, ranging from 15% – 81%.</p>", "<title>The x ray view</title>", "<p>Considering those studies where an A/P view alone is used, between 24 – 56% of patients with radiographic knee osteoarthritis experience pain [##REF##1877852##1##,##REF##2241266##3##,##REF##7654803##10##,##REF##11520166##13##,##REF##9709175##14##,##REF##9404469##20##,##REF##10852280##25##,##REF##1294744##27##,##REF##2287948##28##]. If lateral views alone are considered then 15% of patients with radiographic OA on this view have pain [##REF##8806116##11##]. Adding a lateral or skyline to the A/P view increases the prevalence of pain in those with radiographic OA to 80% [##REF##9893570##8##,##UREF##3##21##]. Cicuttini's study found that abnormalities in the skyline view were nearly twice as likely to predict knee pain as a lateral view, and were also superior to the A/P view in doing this [##REF##8806116##11##]. Including views of the patello-femoral joint improved the sensitivity of predicting knee pain from 38% to 62% in one study [##REF##9893570##8##], and by 10% to 50% in another [##REF##9404469##20##], but with corresponding reductions in specificity.</p>", "<p>With respect to disability, Odding found that abnormalities in the knee x ray were weak predictors of locomotor disability in women and not at all in men [##REF##9709175##14##]. Davis similarly found no association with disability, even for severe radiographic knee osteoarthritis when controlling for other variables such as age, sex and BMI [##REF##1294744##27##]. McAlindon identified ageing, knee pain and quadriceps weakness as three important factors associated with disability but there was no association with radiographic knee osteoarthritis [##REF##8484690##17##].</p>", "<title>Grading the x ray</title>", "<p>Higher grade of osteoarthritis (K/L 3 or more) is a stronger predictor of the presence of pain than lower grades (K/L 2 or less) [##REF##9893570##8##,##REF##7654803##10##,##REF##8806116##11##,##REF##9709175##14##,##REF##9404469##20##,##REF##1294744##27##]. Table ##TAB##3##4## illustrates three studies in which this is apparent, for example Odding found that knee pain was nearly twice as likely for K/L grade 3 as for lower grades [##REF##9709175##14##]. Felson's study of various definitions for knee osteoarthritis examined the use of different radiographic features and their association with the characteristics of clinical osteoarthritis such as pain. The highest sensitivity found was with any grade one osteophyte (82.5%), but the specificity was low (23.3%). On the other hand, joint space narrowing (K/L grade 3) had a low sensitivity (38.3%), but high specificity (82.9%) [##REF##9404469##20##]. Cicuttini describes higher grades of osteophytes as significantly associated with knee pain in the skyline view, but not in the lateral view [##REF##8712864##26##].</p>", "<title>Defining pain</title>", "<p>Table ##TAB##4##5## examines the proportion of people with varying definitions of knee pain in populations with radiographic knee OA. Definitions which examined 'current' pain found prevalence rates of this symptom in radiographic-positive groups that varied from 59 – 81% [##REF##9893570##8##,##REF##7654803##10##,##REF##9709175##14##,##REF##1294744##27##]; lower prevalence estimates were found in studies of pain 'ever', varying from 20 – 59% [##REF##7654803##10##,##REF##1877852##18##,##REF##9404469##20##,##REF##10852280##25##,##REF##8712864##26##]. Even within studies variations existed between 'ever' and 'current' pain. Lethbridge (see table ##TAB##4##5##) estimated that the prevalence of pain at some time in or around the knee for one month among persons with radiographic OA was 53%, but for the same group, if this was limited to experiencing the pain in the last year, this increased to 64%. Cicuttini describes how osteophytes on any view were better predictors of pain in the knee during the last year than pain in the last month or 'ever' [##REF##8806116##11##]. This provides limited evidence that the type of recalled pain might be linked with radiographic pain.</p>", "<title>Nature of the study group</title>", "<p>There appears to be no consistent relationship between age and prevalence of pain in populations with radiographic knee OA. Table ##TAB##3##4## shows the prevalence of knee pain among patients with radiographic knee osteoarthritis in a specified age-group of the population. Williams and Lanyon looked at older age groups and found that about 80% of patients had knee pain [##REF##9893570##8##,##UREF##3##21##]. Lethbridge considered a much wider age range from 19 – 92 and found lower proportions with pain for both K/L grade 2 (30%) and grade 3 (64%) [##REF##7654803##10##]. However the findings of those studies looking at patients in their 40's and over [##REF##2241266##3##,##REF##8806116##11##,##REF##11520166##13##,##REF##9709175##14##,##REF##1877852##18##,##REF##10852280##25##, ####REF##8712864##26##, ##REF##1294744##27####1294744##27##], were not markedly different to those looking at patients aged in their 60's and over [##REF##9404469##20##,##REF##10955336##24##].</p>", "<p>Two studies considered differences between African Americans and American Caucasians from similar geographic locations [##REF##11520166##13##,##REF##12784407##23##]. Table two shows that American Caucasians with radiographic knee OA were less likely to experience pain compared to African Americans (35% vs. 50%) [##REF##11520166##13##], whereas Ang found no ethnic differences in the WOMAC pain and function score for any given level of radiographic knee osteoarthritis [##REF##12784407##23##].</p>" ]

[ "<title>Discussion and conclusion</title>", "<p>This examination of the literature has revealed a wide variation in the degree to which knee pain relates to radiographic knee osteoarthritis and vice versa. We postulated that there might be three particular reasons as to why discordance between x rays and symptoms might arise, from which we can now draw three main conclusions.</p>", "<p>Firstly there may be insufficient x ray numbers or views used to estimate the association. The studies show that the prevalence of radiographic knee OA will be underestimated in persons with knee pain in studies that do not obtain all potential x ray views of the knee. This is supported by the finding that knee studies including x rays of the patello-femoral joint (PFJ), improve the sensitivity with which symptoms such as pain can identify radiographic knee OA [##REF##9893570##8##,##REF##1632657##12##,##REF##11520166##13##]. By adding a lateral or skyline to the A/P view, overall prevalence of radiographic knee OA in pain positive persons increases to 80% [##REF##9893570##8##,##UREF##3##21##]. A recent paper from our group, subsequent to this review, has confirmed this conclusion by showing directly that the prevalence of overall radiographic OA of the knee increases with the number of radiographic views in a population with knee pain [##UREF##2##16##]. However, much discordance remains between pain and x ray findings, and no combination of views reaches a point where patients with knee pain invariably have radiographic knee OA. This is also true for studies examining the prevalence of pain in populations with radiographic knee OA. There is a great deal of discordance evident amongst these studies as highlighted in Table ##TAB##4##5##. Overall these studies support the conclusion that the lack of association between radiographic knee OA and pain is to some extent a real one.</p>", "<p>Secondly, the way pain is defined (e.g. whether disability is included or not) and the grading of radiographic severity, have important influences upon estimates of association between knee pain and radiographic OA and vice versa. Table ##TAB##1##2## and table ##TAB##3##4## examined this relationship with respect to pain definition and demonstrate the wide variation in pain definitions used, and the correspondingly wide variations in the associations between knee pain and x ray findings. It seems likely that the often observed discrepancy between pain and radiographic knee OA has something to do with this variation in definition of pain, and that, if similar methods of pain definition were used, some consistency in the level of discrepancy might emerge. However, the variation between studies is quite marked, so one cannot be wholly convinced of the idea that using one standard uniform definition will lead to x rays and pain becoming more concordant. Other reasons might play their part here. Figure ##FIG##0##1## shows the sources of chronic knee pain in the older person that as a whole make up the knee 'pain picture' we encounter in general practice. Pain in the knee is more than just the result of the pathological changes reflected in the x ray. Other factors may account for knee pain which will not be evident on the knee x ray. Figure ##FIG##0##1## clearly shows this, indicating that the pain may be the result of other bone problems, not visible on an x ray such as oedema, or non-OA conditions such as ligament injury or tendonitis. Indeed, some chronic knee pain might be more strongly linked to issues of cognitive or emotional state such as depression rather than local pathology at the knee joint. Of course, all these things can coexist at the same time, making up multiple layers of causality of knee pain</p>", "<p>The complementary problem concerns the variation in definitions of radiographic OA in any particular view. Some would argue that an isolated osteophyte is not osteoarthritis, although whether the mildest form of osteophyte is included in the definition of OA or not seems to make little difference to the association with pain. However what is clearer from the papers we reviewed is that, with respect to the x ray grade, at the severe end of the spectrum there is a closer association of pain and x rays as shown in table ##TAB##0##1## and table ##TAB##3##4##, but milder disease is more common and the discordance evident at lower levels of K/L grade is important to consider in studies of knee pain and OA. The way the x ray is taken is also important. Between studies the radiographic technique employed may have differed. This will have encompassed whole protocols which might involve the position of the knee (semi-flexed or straight knee). In addition reading the radiographs requires consistency. The studies described go to great lengths to attain intra-study consistency, but we are unable to comment on inter-study consistency and this must be taken into account when evaluating the findings between studies.</p>", "<p>Thirdly, the nature of the study population is important since variations in the association of knee pain and radiographic knee OA may be influenced by characteristics of the population sampled. Younger age groups with knee pain are less likely to have radiographic knee OA (table ##TAB##0##1##), and there is also some variation with age in the proportion of persons with radiographic knee OA and one study suggests that younger patients with radiographic knee OA are less likely to be symptomatic [##REF##7654803##10##]. Ethnicity also has some influence over the relationship [##REF##11520166##13##,##REF##12784407##23##]. Study populations are of course more diverse than in age, gender and ethnicity alone, and it may be that other characteristics than these may both influence the link between x rays and pain, and vary between the populations studied. We did not investigate the effect of other characteristics in this study.</p>", "<p>The major issue for future research is that commitment to more uniformity and standardisation in definitions is needed to allow comparability between studies, and to remove variability between studies as a factor obscuring accurate estimates of the 'true' association between x rays and symptoms at the knee. This would almost certainly involve x raying multiple views of the knee, in a standardised way using consistent protocols across research groups. Pain analysis needs to be similarly standardised, and as recently used in one paper [##REF##16877532##29##], the WOMAC scale allows detailed analysis of pain and dysfunction. Pain grading is essential and might be achieved through using the von Korff Chronic Pain Grade to allow combined measurement of pain and disability severity [##REF##1408309##30##]. Finally using a sampling frame that identified people with a wide range of severity and duration of knee pain, and unselected for their use of healthcare, would deliver a population that truly would be free of selection bias and comparable across study groups.</p>", "<p>We conclude, inevitably, that knee pain is an imprecise marker of radiographic knee osteoarthritis, even in older age groups, but the extent of this imprecision depends heavily on the extent of radiographic views of the joint obtained. Radiographic knee osteoarthritis is likewise an imprecise guide to the likelihood that knee pain or disability will be present, although the more severe the radiographic osteoarthritis, the more likely there are to be accompanying symptoms. Both associations are affected by the definition of pain used and the nature of the study group. The experience of pain is multi-factorial in its origin, and factors such as patient depression play an important part in its manifestation, and this is as true of osteoarthritis and joint pain in older people as it is for pain of uncertain pathology in younger people [##REF##10952867##31##]. Using x rays as a means for investigating knee pain, particularly in older people, requires these other factors to be taken into consideration, and the results of knee radiographs should not be used in isolation when assessing individual patients with knee pain.</p>" ]

[ "<title>Discussion and conclusion</title>", "<p>This examination of the literature has revealed a wide variation in the degree to which knee pain relates to radiographic knee osteoarthritis and vice versa. We postulated that there might be three particular reasons as to why discordance between x rays and symptoms might arise, from which we can now draw three main conclusions.</p>", "<p>Firstly there may be insufficient x ray numbers or views used to estimate the association. The studies show that the prevalence of radiographic knee OA will be underestimated in persons with knee pain in studies that do not obtain all potential x ray views of the knee. This is supported by the finding that knee studies including x rays of the patello-femoral joint (PFJ), improve the sensitivity with which symptoms such as pain can identify radiographic knee OA [##REF##9893570##8##,##REF##1632657##12##,##REF##11520166##13##]. By adding a lateral or skyline to the A/P view, overall prevalence of radiographic knee OA in pain positive persons increases to 80% [##REF##9893570##8##,##UREF##3##21##]. A recent paper from our group, subsequent to this review, has confirmed this conclusion by showing directly that the prevalence of overall radiographic OA of the knee increases with the number of radiographic views in a population with knee pain [##UREF##2##16##]. However, much discordance remains between pain and x ray findings, and no combination of views reaches a point where patients with knee pain invariably have radiographic knee OA. This is also true for studies examining the prevalence of pain in populations with radiographic knee OA. There is a great deal of discordance evident amongst these studies as highlighted in Table ##TAB##4##5##. Overall these studies support the conclusion that the lack of association between radiographic knee OA and pain is to some extent a real one.</p>", "<p>Secondly, the way pain is defined (e.g. whether disability is included or not) and the grading of radiographic severity, have important influences upon estimates of association between knee pain and radiographic OA and vice versa. Table ##TAB##1##2## and table ##TAB##3##4## examined this relationship with respect to pain definition and demonstrate the wide variation in pain definitions used, and the correspondingly wide variations in the associations between knee pain and x ray findings. It seems likely that the often observed discrepancy between pain and radiographic knee OA has something to do with this variation in definition of pain, and that, if similar methods of pain definition were used, some consistency in the level of discrepancy might emerge. However, the variation between studies is quite marked, so one cannot be wholly convinced of the idea that using one standard uniform definition will lead to x rays and pain becoming more concordant. Other reasons might play their part here. Figure ##FIG##0##1## shows the sources of chronic knee pain in the older person that as a whole make up the knee 'pain picture' we encounter in general practice. Pain in the knee is more than just the result of the pathological changes reflected in the x ray. Other factors may account for knee pain which will not be evident on the knee x ray. Figure ##FIG##0##1## clearly shows this, indicating that the pain may be the result of other bone problems, not visible on an x ray such as oedema, or non-OA conditions such as ligament injury or tendonitis. Indeed, some chronic knee pain might be more strongly linked to issues of cognitive or emotional state such as depression rather than local pathology at the knee joint. Of course, all these things can coexist at the same time, making up multiple layers of causality of knee pain</p>", "<p>The complementary problem concerns the variation in definitions of radiographic OA in any particular view. Some would argue that an isolated osteophyte is not osteoarthritis, although whether the mildest form of osteophyte is included in the definition of OA or not seems to make little difference to the association with pain. However what is clearer from the papers we reviewed is that, with respect to the x ray grade, at the severe end of the spectrum there is a closer association of pain and x rays as shown in table ##TAB##0##1## and table ##TAB##3##4##, but milder disease is more common and the discordance evident at lower levels of K/L grade is important to consider in studies of knee pain and OA. The way the x ray is taken is also important. Between studies the radiographic technique employed may have differed. This will have encompassed whole protocols which might involve the position of the knee (semi-flexed or straight knee). In addition reading the radiographs requires consistency. The studies described go to great lengths to attain intra-study consistency, but we are unable to comment on inter-study consistency and this must be taken into account when evaluating the findings between studies.</p>", "<p>Thirdly, the nature of the study population is important since variations in the association of knee pain and radiographic knee OA may be influenced by characteristics of the population sampled. Younger age groups with knee pain are less likely to have radiographic knee OA (table ##TAB##0##1##), and there is also some variation with age in the proportion of persons with radiographic knee OA and one study suggests that younger patients with radiographic knee OA are less likely to be symptomatic [##REF##7654803##10##]. Ethnicity also has some influence over the relationship [##REF##11520166##13##,##REF##12784407##23##]. Study populations are of course more diverse than in age, gender and ethnicity alone, and it may be that other characteristics than these may both influence the link between x rays and pain, and vary between the populations studied. We did not investigate the effect of other characteristics in this study.</p>", "<p>The major issue for future research is that commitment to more uniformity and standardisation in definitions is needed to allow comparability between studies, and to remove variability between studies as a factor obscuring accurate estimates of the 'true' association between x rays and symptoms at the knee. This would almost certainly involve x raying multiple views of the knee, in a standardised way using consistent protocols across research groups. Pain analysis needs to be similarly standardised, and as recently used in one paper [##REF##16877532##29##], the WOMAC scale allows detailed analysis of pain and dysfunction. Pain grading is essential and might be achieved through using the von Korff Chronic Pain Grade to allow combined measurement of pain and disability severity [##REF##1408309##30##]. Finally using a sampling frame that identified people with a wide range of severity and duration of knee pain, and unselected for their use of healthcare, would deliver a population that truly would be free of selection bias and comparable across study groups.</p>", "<p>We conclude, inevitably, that knee pain is an imprecise marker of radiographic knee osteoarthritis, even in older age groups, but the extent of this imprecision depends heavily on the extent of radiographic views of the joint obtained. Radiographic knee osteoarthritis is likewise an imprecise guide to the likelihood that knee pain or disability will be present, although the more severe the radiographic osteoarthritis, the more likely there are to be accompanying symptoms. Both associations are affected by the definition of pain used and the nature of the study group. The experience of pain is multi-factorial in its origin, and factors such as patient depression play an important part in its manifestation, and this is as true of osteoarthritis and joint pain in older people as it is for pain of uncertain pathology in younger people [##REF##10952867##31##]. Using x rays as a means for investigating knee pain, particularly in older people, requires these other factors to be taken into consideration, and the results of knee radiographs should not be used in isolation when assessing individual patients with knee pain.</p>" ]

[ "<p>This is an Open Access article distributed under the terms of the Creative Commons Attribution License (<ext-link ext-link-type=\"uri\" xlink:href=\"http://creativecommons.org/licenses/by/2.0\"/>), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.</p>", "<title>Background</title>", "<p>Studies have suggested that the symptoms of knee osteoarthritis (OA) are rather weakly associated with radiographic findings and vice versa. Our objectives were to identify estimates of the prevalence of radiographic knee OA in adults with knee pain and of knee pain in adults with radiographic knee OA, and determine if the definitions of x ray osteoarthritis and symptoms, and variation in demographic factors influence these estimates.</p>", "<title>Methods</title>", "<p>A systematic literature search identifying population studies which combined x rays, diagnosis, clinical signs and symptoms in knee OA. Estimates of the prevalence of radiographic OA in people with knee pain were determined and vice versa. In addition the effects of influencing factors were scrutinised.</p>", "<title>Results</title>", "<p>The proportion of those with knee pain found to have radiographic osteoarthritis ranged from 15–76%, and in those with radiographic knee OA the proportion with pain ranged from 15% – 81%. Considerable variation occurred with x ray view, pain definition, OA grading and demographic factors</p>", "<title>Conclusion</title>", "<p>Knee pain is an imprecise marker of radiographic knee osteoarthritis but this depends on the extent of radiographic views used. Radiographic knee osteoarthritis is likewise an imprecise guide to the likelihood that knee pain or disability will be present. Both associations are affected by the definition of pain used and the nature of the study group. The results of knee x rays should not be used in isolation when assessing individual patients with knee pain.</p>" ]

[ "<title>Competing interests</title>", "<p>The authors declare that they have no competing interests.</p>", "<title>Authors' contributions</title>", "<p>JB and PC conceived the study. JB designed and conducted the analysis. All authors contributed to the interpretation and writing of the paper, with prime responsibility taken by JB.</p>", "<title>Appendix – Search protocol for the systematic search and summary of the literature relating to radiographic knee osteoarthritis and knee pain</title>", "<p>Please see Table ##TAB##5##6##</p>", "<title>Pre-publication history</title>", "<p>The pre-publication history for this paper can be accessed here:</p>", "<p><ext-link ext-link-type=\"uri\" xlink:href=\"http://www.biomedcentral.com/1471-2474/9/116/prepub\"/></p>" ]

[ "<title>Acknowledgements</title>", "<p>JB's clinical research fellowship was funded by the NHS R and D Capacity Development Programme through the North Staffordshire Primary Care Research Consortium</p>" ]

[ "<fig position=\"float\" id=\"F1\"><label>Figure 1</label><caption><p>Sources of chronic knee pain.</p></caption></fig>" ]

[ "<table-wrap position=\"float\" id=\"T1\"><label>Table 1</label><caption><p>Proportion (%) of patients who have radiographic osteoarthritis in specified age-groups of populations with knee pain.</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td align=\"left\">Study</td><td align=\"right\">Age Group</td><td align=\"right\">Radiographic View</td><td align=\"center\">Proportion (%)</td><td align=\"center\">OA Definition</td><td align=\"center\">Population</td></tr></thead><tbody><tr><td align=\"left\">Petersson [##REF##9306873##19##]</td><td align=\"right\">31–54</td><td align=\"right\">A/Pwb</td><td align=\"center\">15</td><td align=\"center\">Ahlb ≥ 1</td><td align=\"center\">All</td></tr><tr><td/><td/><td/><td/><td align=\"center\">K&amp;L 2+</td><td/></tr><tr><td align=\"left\">Lachance [##REF##11520166##13##]</td><td align=\"right\">40–53</td><td align=\"right\">A/P</td><td align=\"center\">15</td><td align=\"center\">K&amp;L 2+</td><td align=\"center\">CA</td></tr><tr><td/><td/><td/><td align=\"center\">40</td><td/><td align=\"center\">AA</td></tr><tr><td align=\"left\">Hart [##REF##1877852##18##]</td><td align=\"right\">45–65</td><td align=\"right\">A/Pwb</td><td align=\"center\">19</td><td align=\"center\">K&amp;L 2+</td><td align=\"center\">All</td></tr><tr><td align=\"left\">Hannan [##REF##10852280##25##]</td><td align=\"right\">51–74</td><td align=\"right\">A/P</td><td align=\"center\">15</td><td align=\"center\">Def Ost</td><td align=\"center\">All</td></tr><tr><td align=\"left\">Lanyon[##REF##9893570##8##]</td><td align=\"right\">40–80</td><td align=\"right\">A/Pwb + S/L</td><td align=\"center\">Grade 1+ 63%</td><td align=\"center\">Altman</td><td align=\"center\">All</td></tr><tr><td/><td/><td/><td align=\"center\">Grade 2+ 30%</td><td align=\"center\">≥ Grade I</td><td/></tr><tr><td/><td/><td/><td align=\"center\">Grade 3 12%</td><td align=\"center\">Ost [##REF##8581752##32##]</td><td/></tr><tr><td align=\"left\">Claessens [##REF##2241266##3##]</td><td align=\"right\">&gt; 45</td><td align=\"right\">A/Pwb</td><td align=\"center\">36</td><td align=\"center\">K&amp;L 2+</td><td align=\"center\">All</td></tr><tr><td align=\"left\">Cicuttini [##REF##8806116##11##]</td><td align=\"right\">&gt; 45</td><td align=\"right\">Lateral Flexed wb</td><td align=\"center\">30</td><td align=\"center\">Def</td><td align=\"center\">Female</td></tr><tr><td/><td/><td align=\"right\">S/L</td><td align=\"center\">53</td><td align=\"center\">Ost</td><td align=\"center\">Female</td></tr><tr><td align=\"left\">Cicuttini [##REF##8712864##26##]</td><td align=\"right\">&gt; 45</td><td align=\"right\">A/Pwb</td><td align=\"center\">37</td><td align=\"center\">Ost</td><td align=\"center\">Female</td></tr><tr><td/><td/><td align=\"right\">Lateral</td><td align=\"center\">37</td><td align=\"center\">JSN</td><td align=\"center\">Female</td></tr><tr><td/><td/><td align=\"right\">S/L</td><td align=\"center\">51</td><td align=\"center\">or both</td><td align=\"center\">Female</td></tr><tr><td align=\"left\">Odding [##REF##9709175##14##]</td><td align=\"right\">&gt; 55</td><td align=\"right\">A/Pwb</td><td align=\"center\">39</td><td align=\"center\">K&amp;L 2+</td><td align=\"center\">All</td></tr><tr><td align=\"left\">McAlindon [##REF##8484690##17##]</td><td align=\"right\">&gt; 55</td><td align=\"right\">A/Pwb + Lateral</td><td align=\"center\">76</td><td align=\"center\">K&amp;L 2+</td><td align=\"center\">All</td></tr><tr><td align=\"left\">Brandt [##REF##10955336##24##]</td><td align=\"right\">&gt; 65</td><td align=\"right\">A/Pwb + Lateral</td><td align=\"center\">49</td><td align=\"center\">K&amp;L 2+</td><td align=\"center\">All</td></tr><tr><td align=\"left\">Lethbridge [##REF##7654803##10##]</td><td align=\"right\">19–92</td><td align=\"right\">A/P</td><td align=\"center\">KL 2+ 53%</td><td align=\"center\">K&amp;L 2+</td><td align=\"center\">All</td></tr><tr><td/><td/><td/><td align=\"center\">KL 3+ 22%</td><td/><td/></tr><tr><td/><td/><td/><td align=\"center\">KL 4 2%</td><td/><td/></tr><tr><td align=\"left\">Williams [##UREF##3##21##]</td><td align=\"right\">51–80</td><td align=\"right\">A/Pwb + Lat Flexed</td><td align=\"center\">43</td><td align=\"center\">K&amp;L 2+</td><td align=\"center\">All</td></tr></tbody></table></table-wrap>", "<table-wrap position=\"float\" id=\"T2\"><label>Table 2</label><caption><p>Proportion (%) of people with radiographic knee OA in populations with knee pain according to the definition of knee pain.</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td align=\"left\">Study<break/></td><td align=\"center\">% Radiographic OA in those <break/>with Knee pain</td><td align=\"left\">Definition of knee pain positive subjects<break/></td></tr></thead><tbody><tr><td align=\"left\">Hannan [##REF##10852280##25##]</td><td align=\"center\">15</td><td align=\"left\">Pain, swelling, morning stiffness in or around the knee on most days for one month</td></tr><tr><td colspan=\"3\"><hr/></td></tr><tr><td/><td/><td align=\"left\">Positive response to both parts required:</td></tr><tr><td align=\"left\">Lanyon [##REF##9893570##8##]</td><td align=\"center\">30</td><td align=\"left\">(A) Have you ever had pain in or around the knee on most days for one month?</td></tr><tr><td align=\"left\">McAlindon [##REF##1540789##15##]</td><td align=\"center\">53</td><td/></tr><tr><td align=\"left\">McAlindon [##REF##8484690##17##]</td><td align=\"center\">76</td><td/></tr><tr><td align=\"left\">Lethbridge [##REF##7654803##10##]</td><td align=\"center\">53</td><td align=\"left\">(B) If so, have you experienced pain in the last year?</td></tr><tr><td align=\"left\">Felson [##REF##9404469##20##] (Part A only)</td><td align=\"center\">16</td><td/></tr><tr><td colspan=\"3\"><hr/></td></tr><tr><td align=\"left\">Cicuttinni [##REF##8806116##11##]</td><td align=\"center\">37</td><td align=\"left\">Ever having an episode of knee pain</td></tr><tr><td align=\"left\">Cicuttinni [##REF##8712864##26##]</td><td align=\"center\">30</td><td align=\"left\">Ever having an episode of knee pain lasting more than 15 days</td></tr><tr><td colspan=\"3\"><hr/></td></tr><tr><td align=\"left\">Peterson [##REF##9306873##19##]</td><td align=\"center\">15</td><td align=\"left\">Pain in your knees practically daily for the last 3 months</td></tr><tr><td align=\"left\">Lachance [##REF##11520166##13##]</td><td align=\"center\">15 (CA)<break/>40 (AA)</td><td align=\"left\">Any joint pain in their knees during the last during the last month</td></tr><tr><td align=\"left\">Hart [##REF##1877852##18##]</td><td align=\"center\">19</td><td align=\"left\">Pain, stiffness and swelling lasting more than a month</td></tr><tr><td align=\"left\">Odding [##REF##9709175##14##]</td><td align=\"center\">39</td><td align=\"left\">Knee pain during the past month</td></tr><tr><td align=\"left\">Jordan [##REF##9228135##35##]</td><td align=\"center\">N/A</td><td align=\"left\">Knee pain on most days</td></tr><tr><td align=\"left\">Davis [##REF##1294744##27##]</td><td align=\"center\">N/A</td><td align=\"left\">Knee pain on most days lasting for one month in the past year</td></tr><tr><td colspan=\"3\"><hr/></td></tr><tr><td align=\"left\">Williams [##UREF##3##21##]</td><td align=\"center\">N/A</td><td/></tr><tr><td align=\"left\">Brandt [##REF##10955336##24##]</td><td align=\"center\">N/A</td><td/></tr><tr><td align=\"left\">Ang [##REF##12784407##23##]</td><td align=\"center\">N/A</td><td/></tr></tbody></table></table-wrap>", "<table-wrap position=\"float\" id=\"T3\"><label>Table 3</label><caption><p>Comparison of WOMAC scores between studies employing varying definitons of knee pain positive patients.</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td align=\"left\">Study<break/><break/></td><td align=\"left\">Pain +ve WOMAC definition<break/><break/></td><td align=\"left\">WOMAC score in those with <break/>knee pain but no radiographic knee OA <break/>compared to those with radiographic knee OA</td></tr></thead><tbody><tr><td align=\"left\">Brandt [##REF##10955336##24##]<break/></td><td align=\"left\">Greater than moderate (&gt; 3) for any of the five categories on more than <break/>half the days in the month preceding evaluation</td><td align=\"left\">No significant difference<break/></td></tr><tr><td align=\"left\">Williams [##UREF##3##21##]</td><td align=\"left\">Currently had mild pain or greater (&gt; 0).</td><td align=\"left\">No significant difference</td></tr><tr><td align=\"left\">Ang [##REF##12784407##23##]</td><td align=\"left\">Current or past pain, WOMAC transposed to a scale of 0 – 100</td><td align=\"left\">No significant difference</td></tr></tbody></table></table-wrap>", "<table-wrap position=\"float\" id=\"T4\"><label>Table 4</label><caption><p>Proportion (%) of patients experiencing knee pain in specified age-groups of populations with radiographic osteoarthritis.</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td align=\"left\">Study</td><td align=\"right\">Age Group</td><td align=\"right\">Radiographic View</td><td align=\"right\">Proportion (%)</td><td align=\"center\">OA Definition</td><td align=\"right\">Population</td></tr></thead><tbody><tr><td align=\"left\">Lachance [##REF##11520166##13##]</td><td align=\"right\">40–53</td><td align=\"right\">A/Pwb</td><td align=\"right\">35</td><td align=\"center\">K&amp;L 2+</td><td align=\"right\">CA</td></tr><tr><td/><td/><td/><td align=\"right\">50</td><td/><td align=\"right\">AA</td></tr><tr><td align=\"left\">Hart [##REF##1877852##18##]</td><td align=\"right\">45–65</td><td align=\"right\">A/Pwb</td><td align=\"right\">56</td><td align=\"center\">K&amp;L 2+</td><td align=\"right\">All</td></tr><tr><td align=\"left\">Davis [##REF##1294744##27##]</td><td align=\"right\">45–75</td><td align=\"right\">A/Pwb</td><td align=\"right\">41(KL2)</td><td align=\"center\">K&amp;L I+</td><td align=\"right\">All</td></tr><tr><td/><td/><td/><td align=\"right\">59(KL3)</td><td/><td align=\"right\">All</td></tr><tr><td align=\"left\">Hannan [##REF##10852280##25##]</td><td align=\"right\">51–74</td><td align=\"right\">A/P</td><td align=\"right\">47</td><td align=\"center\">Def Ost</td><td align=\"right\">All</td></tr><tr><td align=\"left\">Claessens [##REF##2241266##3##]</td><td align=\"right\">&gt; 45</td><td align=\"right\">A/Pwb</td><td align=\"right\">24</td><td align=\"center\">K&amp;L 2+</td><td align=\"right\">All</td></tr><tr><td align=\"left\">Cicuttini [##REF##8806116##11##]</td><td align=\"right\">&gt; 45</td><td align=\"right\">Lateral Flexed wb</td><td align=\"right\">16</td><td align=\"center\">Def Ost</td><td align=\"right\">Female</td></tr><tr><td/><td/><td align=\"right\">S/L</td><td align=\"right\">26</td><td/><td align=\"right\">Female</td></tr><tr><td align=\"left\">Cicuttini [##REF##8712864##26##]</td><td align=\"right\">&gt; 45</td><td align=\"right\">A/Pwb</td><td align=\"right\">20</td><td align=\"center\">Ost</td><td align=\"right\">Female</td></tr><tr><td/><td/><td align=\"right\">Lateral</td><td align=\"right\">15</td><td align=\"center\">JSN</td><td align=\"right\">Female</td></tr><tr><td/><td/><td align=\"right\">S/L</td><td align=\"right\">23</td><td align=\"center\">or both</td><td align=\"right\">Female</td></tr><tr><td align=\"left\">Odding [##REF##9709175##14##]</td><td align=\"right\">&gt; 55</td><td align=\"right\">A/Pwb</td><td align=\"right\">30(KL2)</td><td align=\"center\">K&amp;L 2+</td><td align=\"right\">All</td></tr><tr><td/><td/><td/><td align=\"right\">59(KL3)</td><td/><td/></tr><tr><td align=\"left\">Felson [##REF##9404469##20##]</td><td align=\"right\">&gt; 63</td><td align=\"right\">A/Pwb</td><td align=\"right\">40</td><td align=\"center\">K&amp;L2+</td><td align=\"right\">All</td></tr><tr><td/><td/><td/><td/><td align=\"center\">or JSN</td><td/></tr><tr><td align=\"left\">Brandt [##REF##10955336##24##]</td><td align=\"right\">&gt; 65</td><td align=\"right\">A/Pwb + Lateral</td><td align=\"right\">22</td><td align=\"center\">K&amp;L2+</td><td align=\"right\">All</td></tr><tr><td align=\"left\">Lethbridge [##REF##7654803##10##]</td><td align=\"right\">19–92</td><td align=\"right\">A/P</td><td align=\"right\">30(KL2)</td><td align=\"center\">K&amp;L2+</td><td align=\"right\">All</td></tr><tr><td/><td/><td/><td align=\"right\">64(KL3)</td><td/><td/></tr><tr><td align=\"left\">Williams [##UREF##3##21##]</td><td align=\"right\">51–80</td><td align=\"right\">A/P Lat Flexed</td><td align=\"right\">79</td><td align=\"center\">K&amp;L2+</td><td align=\"right\">All</td></tr><tr><td align=\"left\">Lanyon [##REF##9893570##8##]</td><td align=\"right\">40–80</td><td align=\"right\">A/Pwb + S/L</td><td align=\"right\">81</td><td align=\"center\">Altman</td><td align=\"right\">All</td></tr><tr><td/><td/><td/><td/><td align=\"center\">≥ Grade I Ost [##REF##8581752##32##]</td><td/></tr></tbody></table></table-wrap>", "<table-wrap position=\"float\" id=\"T5\"><label>Table 5</label><caption><p>Proportion of knee pain positive patients with radiographic knee OA (A/P views) according to the definition of knee pain.</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td align=\"left\">Study</td><td align=\"center\">% Knee pain positive in those with Radiographic OA</td><td align=\"left\">Definition of knee pain positive subjects</td></tr></thead><tbody><tr><td align=\"left\">Hannan [##REF##10852280##25##]</td><td align=\"center\">47</td><td align=\"left\">Pain, swelling, morning stiffness in or around the knee on most days for one month</td></tr><tr><td colspan=\"3\"><hr/></td></tr><tr><td/><td/><td align=\"left\">Positive response to both parts required:</td></tr><tr><td align=\"left\"><bold>Parts A &amp; B</bold></td><td/><td align=\"left\">(A) Have you ever had pain in or around the knee on most days for one month?</td></tr><tr><td align=\"left\">Lanyon [##REF##9893570##8##]</td><td align=\"center\">81</td><td/></tr><tr><td align=\"left\">Lethbridge [##REF##7654803##10##]</td><td align=\"center\">64</td><td/></tr><tr><td align=\"left\"><bold>Part A only</bold></td><td/><td align=\"left\">(B) If so, have you experienced pain in the last year?</td></tr><tr><td align=\"left\">Felson [##REF##9404469##20##]</td><td align=\"center\">40</td><td/></tr><tr><td align=\"left\">Lethbridge [##REF##7654803##10##]</td><td align=\"center\">53</td><td/></tr><tr><td colspan=\"3\"><hr/></td></tr><tr><td align=\"left\">Cicuttinni [##REF##8712864##26##]</td><td align=\"center\">20</td><td align=\"left\">Ever having an episode of knee pain lasting more than 15 days</td></tr><tr><td align=\"left\">Lachance [##REF##11520166##13##]</td><td align=\"center\">35 (CA)<break/>50 (AA)</td><td align=\"left\">Any joint pain in their knees during the last during the last month</td></tr><tr><td align=\"left\">Hart [##REF##1877852##18##]</td><td align=\"center\">56</td><td align=\"left\">Pain, stiffness and swelling lasting more than a month</td></tr><tr><td align=\"left\">Odding [##REF##9709175##14##]</td><td align=\"center\">59</td><td align=\"left\">Knee pain during the past month</td></tr><tr><td align=\"left\">Davis [##REF##1294744##27##]</td><td align=\"center\">59</td><td align=\"left\">Knee pain on most days lasting for one month in the past year</td></tr></tbody></table></table-wrap>", "<table-wrap position=\"float\" id=\"T6\"><label>Table 6</label><caption><p>Search protocol for the systematic search and summary of the literature relating to radiographic knee osteoarthritis and knee pain </p></caption><table frame=\"hsides\" rules=\"groups\"><tbody><tr><td align=\"right\">1.</td><td align=\"left\">SEARCH:</td><td align=\"left\">KNEE$.TI,AB,SH,DE.</td></tr><tr><td align=\"right\">2.</td><td align=\"left\">SEARCH:</td><td align=\"left\">PATELL$.TI,AB,SH,DE.</td></tr><tr><td align=\"right\">4.</td><td align=\"left\">SEARCH:</td><td align=\"left\">(KNEE ADJ JOINT).TI,AB,SH,DE.</td></tr><tr><td align=\"right\">5.</td><td align=\"left\">SEARCH:</td><td align=\"left\">(GENU ADJ VALGUS).TI,AB,SH,DE.</td></tr><tr><td align=\"right\">6.</td><td align=\"left\">SEARCH:</td><td align=\"left\">(GENU ADJ VARUS).TI,AB,SH,DE.</td></tr><tr><td align=\"right\">7.</td><td align=\"left\">SEARCH:</td><td align=\"left\">1 OR 2 OR 3 OR 4 OR 5 OR 6</td></tr><tr><td align=\"right\">8.</td><td align=\"left\">SEARCH:</td><td align=\"left\">OSTEOARTHR$.TI,AB,SH,DE.</td></tr><tr><td align=\"right\">9.</td><td align=\"left\">SEARCH:</td><td align=\"left\">OA.TI,AB,SH,DE.</td></tr><tr><td align=\"right\">10.</td><td align=\"left\">SEARCH:</td><td align=\"left\">gonarthrosis.TI,AB,SH,DE.</td></tr><tr><td align=\"right\">11.</td><td align=\"left\">SEARCH:</td><td align=\"left\">(DEGENERATIVE ADJ JOINT ADJ DISEASE).TI,AB,SH,DE.</td></tr><tr><td align=\"right\">12.</td><td align=\"left\">SEARCH:</td><td align=\"left\">8 OR 9 OR 10 OR 11</td></tr><tr><td align=\"right\">13.</td><td align=\"left\">SEARCH:</td><td align=\"left\">DIAGNOS$.TI,AB,SH,DE.</td></tr><tr><td align=\"right\">14.</td><td align=\"left\">SEARCH:</td><td align=\"left\">GNOSIS.TI,AB,SH,DE.</td></tr><tr><td align=\"right\">15.</td><td align=\"left\">SEARCH:</td><td align=\"left\">PROGNOS$.TI,AB,SH,DE.</td></tr><tr><td align=\"right\">16.</td><td align=\"left\">SEARCH:</td><td align=\"left\">13 OR 14 OR 14 OR 15</td></tr><tr><td align=\"right\">17.</td><td align=\"left\">SEARCH:</td><td align=\"left\">(X ADJ RAY).TI,AB,SH,DE.</td></tr><tr><td align=\"right\">18.</td><td align=\"left\">SEARCH:</td><td align=\"left\">RADIOGRAPHIC$.TI,AB,SH,DE.</td></tr><tr><td align=\"right\">19.</td><td align=\"left\">SEARCH:</td><td align=\"left\">RADIOLOGICAL$.TI,AB,SH,DE.</td></tr><tr><td align=\"right\">20.</td><td align=\"left\">SEARCH:</td><td align=\"left\">RADIOLOGIST$.TI,AB,SH,DE.</td></tr><tr><td align=\"right\">21.</td><td align=\"left\">SEARCH:</td><td align=\"left\">17 OR 18 OR 19 OR 20</td></tr><tr><td align=\"right\">22.</td><td align=\"left\">SEARCH:</td><td align=\"left\">7 AND 12 AND 16 AND 21</td></tr><tr><td align=\"right\">23.</td><td align=\"left\">SEARCH:</td><td align=\"left\">MRI.TI,AB,SH,DE.</td></tr><tr><td align=\"right\">24.</td><td align=\"left\">SEARCH:</td><td align=\"left\">CT.TI,AB,SH,DE.</td></tr><tr><td align=\"right\">25.</td><td align=\"left\">SEARCH:</td><td align=\"left\">23 OR 24</td></tr><tr><td align=\"right\">26.</td><td align=\"left\">SEARCH:</td><td align=\"left\">22 NOT 25</td></tr><tr><td align=\"right\">27.</td><td align=\"left\">SEARCH:</td><td align=\"left\">ARTHROPLASTY.TI,AB,SH,DE.</td></tr><tr><td align=\"right\">28.</td><td align=\"left\">SEARCH:</td><td align=\"left\">(KNEE ADJ REPLACEMENT).TI,AB,SH,DE.</td></tr><tr><td align=\"right\">29.</td><td align=\"left\">SEARCH:</td><td align=\"left\">(KNEE ADJ SURGERY).TI,AB,SH,DE.</td></tr><tr><td align=\"right\">30.</td><td align=\"left\">SEARCH:</td><td align=\"left\">ARTHROSCOP$.TI,AB,SH,DE.</td></tr><tr><td align=\"right\">31.</td><td align=\"left\">SEARCH:</td><td align=\"left\">27 OR 28 OR 29 OR 30</td></tr><tr><td align=\"right\">32.</td><td align=\"left\">SEARCH:</td><td align=\"left\">26 NOT 31</td></tr><tr><td align=\"right\">33.</td><td align=\"left\">SEARCH:</td><td align=\"left\">gout$.TI,AB,SH,DE.</td></tr><tr><td align=\"right\">34.</td><td align=\"left\">SEARCH:</td><td align=\"left\">rheumatoid$.TI,AB,SH,DE.</td></tr><tr><td align=\"right\">35.</td><td align=\"left\">SEARCH:</td><td align=\"left\">pseudogout$.TI,AB,SH,DE.</td></tr><tr><td align=\"right\">36.</td><td align=\"left\">SEARCH:</td><td align=\"left\">33 OR 34 OR 35</td></tr><tr><td align=\"right\">37.</td><td align=\"left\">SEARCH:</td><td align=\"left\">32 NOT 36</td></tr></tbody></table></table-wrap>" ]

[]

[]

[]

[]

[]

[]

[ "<table-wrap-foot><p>A/P – antero-posterior</p><p>wb – weight bearing</p><p>S/L – skyline view</p><p>Lat – lateral</p><p>CA – Caucasian</p><p>AA – African American</p><p>All – whole population</p><p>K&amp;L – Kellgren &amp; Lawrence Knee OA Grading Scale [##REF##13498604##33##]</p><p>Ahlb – Ahlbäck Knee OA Grading Scale [##UREF##4##34##]</p><p>Ost – Osteophytes</p><p>Def Ost – Definite Osteophytes</p><p>JSN – Joint Space Narrowing</p></table-wrap-foot>", "<table-wrap-foot><p>CA – Caucasian, AA – African American, N/A – not applicable.</p></table-wrap-foot>", "<table-wrap-foot><p>A/P – antero-posterior</p><p>wb – weight bearing</p><p>S/L – skyline</p><p>Lat – lateral</p><p>CA – Caucasian</p><p>AA – African American</p><p>KL – Kellgren Lawrence grade</p><p>All – whole population</p><p>K&amp;L – Kellgren &amp; Lawrence Knee OA Grading Scale [##REF##13498604##33##]</p><p>Ost – Osteophytes</p><p>Def Ost – Definite Osteophytes</p><p>JSN – Joint Space Narrowing</p></table-wrap-foot>", "<table-wrap-foot><p>CA – Caucasian, AA – African American, N/A – not applicable.</p></table-wrap-foot>" ]

[ "<graphic xlink:href=\"1471-2474-9-116-1\"/>" ]

[]

{ "acronym": [], "definition": [] }

[ "<title>Background</title>", "<p>Osteoarthritis (OA) is a common condition that affects 18% women and 10% men (aged &gt; 60 years) worldwide [##REF##14710506##1##]. Treatment for these patients is aimed at controlling pain, improving functional abilities and enhancing health-related quality-of-life [##REF##11014340##2##].</p>", "<p>Non-selective non-steroidal anti-inflammatory drugs (NSAIDs) such as naproxen and ibuprofen are widely used for pain relief in OA. However, upper gastrointestinal (GI) symptoms such as dyspepsia and more importantly ulcer complications occur in 15–60% of NSAID users and frequently necessitate co-therapy with H₂receptor antagonists or proton pump inhibitors [##REF##11179238##3##, ####REF##8969678##4##, ##REF##7781458##5##, ##REF##12528071##6####12528071##6##]. In a prospective cohort study, it was observed that 81% of patients taking NSAIDs and having serious GI complications had no prior GI symptoms [##REF##8687261##7##] and in a survey in the US among NSAID users, it was observed that nearly 75% of those who regularly used NSAIDs did not know about or were unconcerned about NSAID related GI complications [##REF##10225536##8##]. GI adverse events (AEs) are the main factors limiting the use of NSAIDs and represent a significant health burden [##REF##12528071##6##]. Renal impairment, vascular constriction and GI AEs are attributed to inhibition of cyclooxygenase-1 (COX-1), anti-inflammatory and analgesic effect is attributed to inhibition of COX-2. Hence, selective COX-2 inhibitors like celecoxib and rofecoxib provide a more favourable GI safety profile with similar efficacy as compared to non-selective NSAIDs in patients with OA [##REF##11316141##9##,##REF##10580458##10##]. Rofecoxib, however, was withdrawn worldwide on September 30, 2004 due to an increase in the cardiovascular (CV) risk [##UREF##0##11##]. Following this withdrawal, concerns have also been raised regarding CV safety of both selective COX-2 inhibitors and traditional NSAIDs. These concerns arose initially for selective COX-2 inhibitors following the worldwide withdrawal of rofecoxib. Meta-analyses have since reported an increased risk of CV events with both traditional NSAIDs and COX-2 inhibitors and both carry warnings to this effect in their prescribing information [##UREF##1##12##, ####REF##16740558##13##, ##REF##16995929##14####16995929##14##].</p>", "<p>Lumiracoxib is a structurally distinct, selective COX-2 inhibitor for the management of OA and acute pain. Lumiracoxib is effective in treating acute pain conditions such as post-operative dental pain [##REF##15117091##15##], acute gout [##REF##17478464##16##], arthroplasty [##REF##16223396##17##], sprains and strains [##UREF##2##18##] and in treating chronic pain associated with OA [##UREF##3##19##,##UREF##4##20##].</p>", "<p>The 52-week Therapeutic Arthritis Research and Gastrointestinal Event Trial (TARGET) in 18 000 patients with OA investigated the GI, CV and overall safety profile of lumiracoxib 400 mg od (four times the recommended dose for OA) compared to two traditional NSAIDs, naproxen 500 mg bid and ibuprofen 800 mg tid [##REF##15325831##21##,##REF##15325832##22##]. The TARGET study showed that lumiracoxib was associated with a 79% decrease in upper GI complications compared to traditional NSAIDs (non-aspirin population) [##REF##15325831##21##]. The GI benefit with lumiracoxib compared to traditional NSAIDs occurred within 8 days of treatment [##UREF##5##23##]. In TARGET lumiracoxib was also associated with an improved blood pressure (BP) profile as compared to the traditional NSAIDs, already after 4 weeks of treatment [##UREF##6##24##] and the effect was maintained until 52 weeks [##REF##15325832##22##].</p>", "<p>The present short-term safety study assessed the GI tolerability of a 6-week treatment with lumiracoxib 400 mg od (four times the recommended dose for OA) as compared to rofecoxib 25 mg od (therapeutic dose) in patients with OA. In addition, the study also assessed renal effects including the incidence of peripheral oedema and changes in BP in the two treatment groups.</p>" ]

[ "<title>Methods</title>", "<title>Study design</title>", "<p>This study was a 6-week, multicentre, randomised, double-blind, double-dummy, active-controlled, parallel-group, safety study of lumiracoxib 400 mg od (four times the recommended dose for OA) compared to rofecoxib 25 mg od. The study enrolled subjects with primary OA across 51 centres in Europe. This study was performed according to Good Clinical Practice guidelines. Ethics committee approval from all participating institutions was obtained in accordance with the Declaration of Helsinki and all patients gave their written informed consent before enrolment. The study had a 3–7 day wash-out period, 6-week treatment period and a follow-up by phone call 2 weeks after the end of study/early termination.</p>", "<title>Study population</title>", "<p>Symptomatic patients (aged ≥50 years) with OA as defined by the American College of Rheumatology criteria were recruited. The criteria for inclusion were primary OA for at least 3 months in the hip, hand, knee or spine (cervical or lumbar) and pain in the target joint of at least moderate intensity (Likert scale). Patients also needed to be on NSAID or other analgesic therapy or expected to need NSAID treatment for at least 6 weeks.</p>", "<p>The exclusion criteria were secondary OA and/or history/evidence of significant diseases in the affected joints, evidence of active ulceration or bleeding of the upper GI tract, upper GI tract malignancies, diseases of the intestinal tract and bleeding diathesis. Patients were excluded if they had clinically significant hepatic or renal disease, evidence of hepatic, renal or blood coagulation disorders or anaemia, hypertension, type I diabetes or other significant medical problems, used systemic steroids, intra-articular hyaluronic acid injections, H₂receptor antagonists, proton pump inhibitors, sucralfate or prostaglandin analogues in the past month. Pregnant or lactating women and women not on acceptable form of contraception were also excluded.</p>", "<title>Study medication and assessments</title>", "<p>Patients were randomly allocated in the ratio of 1:1 to receive either lumiracoxib 400 mg od (four times the recommended dose for OA) or rofecoxib 25 mg od. Lumiracoxib (Prexige^®Novartis Pharma AG, Basel, Switzerland) was provided as 2 × 200 mg tablets with matching placebos and rofecoxib as 25 mg capsules with matching placebos. Patients were asked to take the medication once every morning at approximately the same time. Compliance with study drug was defined as patients taking ≥80% of the full daily dose. To control GI symptoms, patients were allowed a maximum of eight antacid tablets (calcium carbonate 680 mg/magnesium carbonate 80 mg) per day as rescue medication. Patients received the study medication for 6 weeks.</p>", "<title>Safety assessments</title>", "<p>The key primary assessment was incidence of at least one of the predefined GI AEs: abdominal pain, constipation, diarrhoea, nausea, vomiting, dyspepsia and dysphagia. The other primary assessment was incidence of peripheral oedema: lower limb oedema, upper limb oedema, peripheral swelling and peripheral oedema. The secondary safety assessments were incidence of moderate and severe predefined GI AEs, incidence of each individual predefined GI AE, discontinuations from study because of any AE or GI AE and time to discontinuation, mean sitting systolic and diastolic blood pressure (msSBP and msDBP), and the number of tablets of antacid rescue medication taken. Study assessments were performed at baseline, Weeks 3 and 6.</p>", "<p>Tolerability was evaluated by recording AEs during the entire study period. A follow-up phone call 2 weeks after the end of study was carried out to evaluate serious adverse events (SAEs) after study drug discontinuation. Investigators were requested to report all SAE's which occurred within 4 weeks after last dose of study drug intake. Vital signs including BP measurements and standard laboratory tests were performed at baseline, Weeks 3 and 6. ECG recordings were performed at baseline and Week 6.</p>", "<title>Efficacy assessment</title>", "<p>Efficacy variables were overall pain intensity in the target joint and the global assessments of disease activity by patients and physicians on a 5-point Likert scale at Weeks 3 and 6. For overall pain intensity in the target joint, patients were classified as improved if endpoint assessment was \"none\" or improved by at least two grades from baseline on the Likert scale. For patient's and physician's global assessment of disease activity, patients were classified as improved if endpoint assessment was \"very good\" or improved by at least two grades from baseline.</p>", "<title>Statistical analysis</title>", "<p>The categorical efficacy variables were analysed in the intent-to-treat (ITT) population defined as all randomised patients who received study medication. A multiple logistic model (PROC LOGISTIC in SAS), which considered treatment as main effect was used for the analysis. The treatment contrasts were tested at a two-sided 5% significance level and presented as odds ratios (ORs) together with their 95% confidence intervals (CIs). Missing data for efficacy variables were imputed using the last-observation-carried-forward (LOCF) method. The primary safety endpoints were analysed in the safety population defined as all patients randomised to treatment, who had been exposed to study medication. A multiple logistic model, which took into account country and treatment as the main effect was used for the analysis. The treatment contrasts were tested at a two-sided 5% level of significance and presented as ORs together with their 95% CIs. If the estimated incidence rates were less than 5%, or if the logistic regression model did not converge, Fisher's exact test was used for the comparisons. Analysis was repeated in the per-protocol (PP) population defined as a sub-population of the safety population for sensitivity reasons. Between-treatment comparisons for msSBP and msDBP were performed by means of analysis of covariance (ANCOVA). For the ANCOVA models, treatment and country were taken as fixed effects and the respective baseline values as covariate. Time to discontinuation from study due to any AE or GI AE was analysed using life-table methods. Patient compliance and other categorical safety endpoints were analysed using a multiple logistic model at a two-sided 5% significance level and presented as ORs together with their 95% CIs.</p>", "<title>Sample size and power considerations</title>", "<p>The determination of the sample size was based on the key primary safety variable, the incidence of predefined GI AEs. A two-group continuity corrected chi-squared test with a two-sided 5% significance level had 80% power to detect a clinically relevant difference between the treatment groups, assuming 28% in the rofecoxib group and 14% in the lumiracoxib group when the sample size is 146 patients per treatment arm. Three hundred and four patients (152 each on lumiracoxib and rofecoxib) needed to be randomised to allow for a 4% dropout rate.</p>" ]

[ "<title>Results</title>", "<p>After an initial wash-out period of 3–7 days, a total of 309 patients were randomised to either lumiracoxib 400 mg od (<italic>n </italic>= 154) or rofecoxib 25 mg od (<italic>n </italic>= 155) (Figure ##FIG##0##1##). All randomised patients were included in the ITT and safety populations. At baseline, the treatment groups were comparable in terms of demographic and baseline characteristics (Table ##TAB##0##1##). Medical histories indicated that more patients on lumiracoxib had vascular disorders as compared to rofecoxib (54.5% <italic>vs</italic>. 46.5%, respectively). History of cardiac disorders at baseline was more frequent in lumiracoxib patients (16%) than rofecoxib patients (11%). In both the groups, a similar percentage (42%) of patients had previously undergone surgical and medical procedures. More than 90% of patients in both the treatment groups completed the study. Major protocol violations resulting in exclusion from the PP population occurred in 13 patients receiving lumiracoxib and seven patients receiving rofecoxib.</p>", "<title>Primary safety endpoints</title>", "<p>There was no statistically significant difference in the overall incidence of key primary assessment variables (predefined GI AEs – abdominal pain, constipation, diarrhea, nausea, vomiting, dyspepsia and dysphagia) between the treatment groups (OR: 1.31; 95% CI: 0.82, 2.11, <italic>p </italic>= 0.258). Predefined GI AEs were reported in 43.5% (<italic>n </italic>= 67) of patients in lumiracoxib group and 37.4% (<italic>n </italic>= 58) of patients in rofecoxib group. Thus, overall both the study drugs displayed similar GI safety profiles. The incidence of the other primary assessment variable, peripheral oedema was low in both the treatment groups (<italic>n </italic>= 9, 5.8%) (Figure ##FIG##1##2##).</p>", "<title>Secondary safety endpoints</title>", "<p>There was no statistically significant difference between lumiracoxib and rofecoxib for the incidence of individual predefined GI AEs. Minor differences between lumiracoxib and rofecoxib in the incidence rates for diarrhoea (11.0% <italic>vs</italic>. 5.2%), dyspepsia (26.6% <italic>vs</italic>. 20.6%) and constipation (2.6% <italic>vs</italic>. 0.6%) were observed, but were not statistically significant. When the incidence rates of these predefined GI AEs were analysed based on their severity it was observed that moderate or severe predefined GI AEs associated with lumiracoxib and rofecoxib were comparable with the exception of dyspepsia that occurred more often in the lumiracoxib group (11.0% lumiracoxib <italic>vs</italic>. 4.5% rofecoxib, <italic>p </italic>= 0.035, Fisher's exact test [Table ##TAB##1##2##]).</p>", "<p>The rate of moderate-to-severe peripheral oedema was low in both treatment groups. Only one patient in the lumiracoxib group (0.6%) <italic>versus </italic>three patients (1.9%) in the rofecoxib group had a moderate-to-severe event. This numerical difference was not statistically significant. After 6 weeks of treatment, a significantly lower msSBP and msDBP was observed with lumiracoxib as compared to rofecoxib (least square estimated difference: -3.13 mmHg, 95% CI: – 6.17, -0.10, <italic>p </italic>= 0.043 for msSBP and -1.73 mmHg, 95% CI: -3.43, – 0.03, for msDBP, <italic>p </italic>= 0.046 [Figure ##FIG##2##3##]). The mean number of antacid tablets taken was the same in both treatment groups (0.2 tablets/day).</p>", "<p>The most frequently reported AEs (by preferred term) during this study are listed in Table ##TAB##2##3##. The incidence of AEs was comparable between lumiracoxib and rofecoxib. The most commonly reported AEs by primary system organ class were GI disorders, infections and infestations, and musculoskeletal and connective tissue disorders, which were similar in incidence in both treatment groups. Study-drug related AEs as suspected by the investigator were reported in 40.9% of patients in the lumiracoxib group and 37.4% of patients in the rofecoxib group. As expected in a study focussing on GI safety, AEs were most commonly reported in the GI system. Three rofecoxib-treated patients experienced AEs that led to temporary interruption of study medication (Table ##TAB##3##4##).</p>", "<p>Discontinuations due to GI AEs occurred in 4.5% and 2.6% of the patients treated with lumiracoxib and rofecoxib, respectively (<italic>p </italic>= 0.359). The mean time to discontinuation for patients treated with lumiracoxib as compared to rofecoxib for any AE (23.3 days <italic>vs</italic>. 21.7 days, respectively) and for GI AEs (23.4 days <italic>vs</italic>. 25.3 days, respectively) was comparable. A similar proportion of patients discontinued from the study due to any AE in both groups (5.2% of lumiracoxib and 4.5% of rofecoxib patients).</p>", "<p>No drug-related SAEs or deaths were reported during the course of the study. One SAE (vaginal haemorrhage) was reported in the rofecoxib group.</p>", "<p>Serum chemistry and haematology parameters were in the normal range at baseline for the majority of patients in both the treatment groups and remained so at the end of study. No elevations in alanine amino transferase/aspartate amino transferase &gt; 3 × ULN were observed during the study.</p>", "<title>Efficacy endpoint</title>", "<p>An improvement in target joint pain or disease activity was reported in 30–40% of patients in the lumiracoxib and rofecoxib groups after 6 weeks of treatment (Table ##TAB##4##5##). The differences between the treatment groups were not statistically significant for any efficacy parameter.</p>" ]

[ "<title>Discussion</title>", "<p>In this study, both lumiracoxib and rofecoxib showed similar efficacy in treating pain associated with OA.</p>", "<p>The GI safety profile of lumiracoxib 400 mg od (four times the recommended dose for OA) was comparable to rofecoxib 25 mg od over 6 weeks of treatment. The incidence of individual predefined GI AEs and their severity was also comparable between the treatment groups.</p>", "<p>Lumiracoxib is indicated at a dose of 100 mg once daily for chronic use in OA, and at doses of 200 mg or 400 mg once daily for short-term use in acute pain indications. While liver toxicity is a known rare but serious side effect of all COX-2 inhibitors and traditional NSAIDs [##REF##12842950##25##], there have been some specific concerns from health authorities regarding the hepatic safety profile of lumiracoxib. Lumiracoxib was withdrawn in Australia in August 2007 following reports of severe liver events occurring predominantly at doses higher than the recommended dose of 100 mg od, when taken chronically. The US FDA issued a non-approvable letter in September 2007, citing concerns over the hepatic profile of lumiracoxib. This was followed by withdrawals in Canada, Europe and a few other countries. Assessment of the benefit to risk profile of the drug is currently ongoing by a number of health authorities.</p>", "<p>Liver toxicity is a known rare but serious side effect of all COX-2 inhibitors and traditional NSAIDs and it is not clear the risk is higher with lumiracoxib than other NSAIDs.</p>", "<p>In this 6-weeks study no elevations in liver enzymes were observed with lumiracoxib. This is in agreement with the results from TARGET where the incidence of ALT/AST elevations &gt; 3 × ULN were low with lumiracoxib, comparable to ibuprofen and naproxen and no \"Hy's cases\" (ALT/AST &gt; 3 × ULN and total bilirubin &gt; 3 mg/dL), which are more predictive for severe liver outcome, were observed during the first 49 days of treatment [##REF##18221410##26##].</p>", "<p>Traditional NSAIDs and selective COX-2 inhibitors like rofecoxib and etoricoxib have been shown to increase BP in clinical studies [##REF##8037411##27##,##REF##15710786##28##] and in the recent Multinational Etoricoxib and Diclofenac Arthritis Long-Term (MEDAL) study, discontinuations due to hypertension were observed more frequently with etoricoxib compared with diclofenac [##REF##17113426##29##]. In this study, after 6 weeks of treatment, a statistically significantly better BP profile was observed with lumiracoxib as compared to rofecoxib, with an estimated difference of more than 3 mmHg systolic blood pressure (SBP) in favour of lumiracoxib. Although this difference was small, reports suggest that increases in SBP of 1–5 mmHg have been associated with 7100–35 700 additional ischemic heart disease and stroke events in OA patients over a 1-year period in the USA [##REF##12672188##30##]. These findings are consistent with previous findings where a 2 mmHg decrease in SBP reduced the risk of death due to ischemic heart disease and stroke by approximately 7% and 10%, respectively, in middle age [##REF##12493255##31##]. Hence, maintaining BP control can provide substantial benefits in OA patients [##REF##15545508##32##].</p>", "<p>These results are in agreement with the findings of the 12-month TARGET outcome study with lumiracoxib, where lumiracoxib had an improved BP profile compared with ibuprofen or naproxen [##REF##15325832##22##,##UREF##6##24##]. The improved BP profile with lumiracoxib as compared to ibuprofen was also observed in hypertensive OA patients [##UREF##7##33##]. In addition, results from a meta-analysis involving 9 611 patients on lumiracoxib (100–400 mg od) revealed that lumiracoxib provided a BP profile (both systolic and diastolic) comparable to placebo [##UREF##8##34##].</p>", "<p>Moreover, in TARGET, the incidence of oedema was low and lumiracoxib was not associated with any increase in the incidence of oedema, compared with ibuprofen or naproxen [##UREF##9##35##], while in the VIGOR study, the incidence of oedema was higher in the rofecoxib group as compared to the naproxen group [##UREF##10##36##]. The incidence of peripheral oedema was low and similar in both the groups in this study. A numerical difference for moderate and severe peripheral oedemas was also observed in favour of lumiracoxib, although it did not reach statistical significance.</p>", "<p>The incidence of AEs and discontinuations due to AEs were comparable between the treatment groups. The most common AEs suspected by the investigator to be study-drug related were GI AEs, as expected in a study on GI safety.</p>" ]

[ "<title>Conclusion</title>", "<p>Lumiracoxib 400 mg od (four times the recommended dose in OA) demonstrated comparable GI safety profile to rofecoxib 25 mg od (therapeutic dose) in patients with OA. However, lumiracoxib was associated with a significantly better BP profile as compared to rofecoxib.</p>" ]

[ "<p>This is an Open Access article distributed under the terms of the Creative Commons Attribution License (<ext-link ext-link-type=\"uri\" xlink:href=\"http://creativecommons.org/licenses/by/2.0\"/>), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.</p>", "<title>Background</title>", "<p>Lumiracoxib is a selective cyclooxygenase-2 inhibitor effective in the treatment of osteoarthritis (OA) with a superior gastrointestinal (GI) safety profile as compared to traditional non-steroidal anti-inflammatory drugs (NSAIDs, ibuprofen and naproxen). This safety study compared the GI tolerability, the blood pressure (BP) profile and the incidence of oedema with lumiracoxib and rofecoxib in the treatment of OA. Rofecoxib was withdrawn worldwide due to an associated increased risk of CV events and lumiracoxib has been withdrawn from Australia, Canada, Europe and a few other countries following reports of suspected adverse liver reactions.</p>", "<title>Methods</title>", "<p>This randomised, double-blind study enrolled 309 patients (aged greater than or equal to 50 years) with primary OA across 51 centres in Europe. Patients were randomly allocated to receive either lumiracoxib 400 mg od (four times the recommended dose in OA) (<italic>n </italic>= 154) or rofecoxib 25 mg od (<italic>n </italic>= 155). The study was conducted for 6 weeks and assessments were performed at Weeks 3 and 6. The primary safety measures were the incidence of predefined GI adverse events (AEs) and peripheral oedema. The secondary safety measures included effect of treatment on the mean sitting systolic and diastolic blood pressure (msSBP and msDBP). Tolerability of lumiracoxib 400 mg was assessed by the incidence of AEs.</p>", "<title>Results</title>", "<p>Lumiracoxib and rofecoxib displayed similar GI safety profiles with no statistically significant difference in predefined GI AEs between the two groups (43.5% <italic>vs</italic>. 37.4%, respectively). The incidence and severity of individual predefined GI AEs was comparable between the two groups. The incidence of peripheral oedema was low and identical in both the groups (<italic>n </italic>= 9, 5.8%). Only one patient in the lumiracoxib group and three patients in the rofecoxib group had a moderate or severe event. At Week 6 there was a significantly lower msSBP and msDBP in the lumiracoxib group compared to the rofecoxib group (<italic>p </italic>&lt; 0.05). A similar percentage of patients in both groups showed an improvement in target joint pain and disease activity. The tolerability profile was similar in both the treatment groups.</p>", "<title>Conclusion</title>", "<p>Lumiracoxib 400 mg od (four times the recommended dose in OA) provided a comparable GI safety profile to rofecoxib 25 mg od (therapeutic dose). However, lumiracoxib was associated with a significantly better BP profile as compared to rofecoxib.</p>", "<title>Trial registration number -</title>", "<p>NCT00637949</p>" ]

[ "<title>Abbreviations</title>", "<p>AEs: adverse events; ANCOVA: analysis of covariance; BP: blood pressure; CIs: confidence intervals; COX: cyclooxygenase; CV: cardiovascular; GI: gastrointestinal; ITT: intent-to-treat; LOCF: last-observation-carried-forward; MEDAL: Multinational Etoricoxib and Diclofenac Arthritis Long-Term; msDBP: mean sitting diastolic blood pressure; msSBP: mean sitting systolic blood pressure; NSAIDs: non-steroidal anti-inflammatory drugs; OA: osteoarthritis; Ors: odds ratios; PP: per-protocol; SAEs: serious adverse events; SBP: systolic blood pressure; TARGET: Therapeutic Arthritis Research and Gastrointestinal Event Trial.</p>", "<title>Appendix 1: List of Investigators</title>", "<p><bold>Austria</bold>: Dr. Winfried Graninger, Universitaetsklinik fuer Innere Medizin III, Klin. Abteilung Rheumatologie, Waehringer Guertel 18–20, A-109 Vienna; Dr. Peter Peichl, Kaiser-Franz-Josef-Spital, 2. Medizinische Abteilung mit Rheumatologie und Osteologie der Stadt Wien, Kundratstrasse 3, A-1100 Wien; Dr. Attila Dunky, Wilhelminenspital der Stadt Wien, 5. Mediz. Abteilung m. Rheumatologie, Stoffwechsel, Rehabilitation Montleartstrasse 37, A-11650 Vienna; Dr. Josef Hermann, Medizinische Universitaets Klinik, Universitaet Graz Auenbruggerplatz 15, 8036 Graz</p>", "<p><bold>Belgium</bold>: Prof. P. Geusens, Biomedisch Onderzoeksinstituut – DWI, Limburgs Universitair Centrum, Universitaire Campus – Building C, 3590 Diepenbeek; Prof. Jean-Pierre Devogelaer, Cliniques Universitaire St. Luc, Service de Rhumatologie, Avenue Hippocrate 10, 1200 Bruxelles</p>", "<p><bold>France</bold>: Dr. C. Copere, Private Practice, Roanne; Dr. A. Duplain, Private Practice, Roanne; Dr. D. Estienne, Private Practice, Roanne; Dr. M. Fleury, Private Practice, Roanne; Dr. P.L. Jacquier, Private Practice, Roanne; Dr. J. Richard, Private Practice, Roanne; Dr. J.M. Aupy, Private Practice, Roanne; Dr. S. Benayoune, Private Practice, Roanne; Dr. J.-M. Blot, Private Practice, Roanne; Dr. D. Brechoire, Private Practice, Roanne; Dr. M. Gacioch, Private Practice, Roanne; Dr. G. Etchegary, Private Practice, Niort; Dr. C. Tilly, Private Practice, Niort; Dr. P. Amlard, Private Practice, Niort; Dr. M. Anthony, Private Practice, Niort; Dr. M. Baert, Private Practice, Niort; Dr. J. Marty, Private Practice, Murs Erigne; Dr. J-F. Pascal, Private Practice, Murs Erigne; Dr. D. Tirouflet, Private Practice, Murs Erigne</p>", "<p><bold>Netherlands</bold>: Dr. G.J.M. van Doesburg, Private Practice, Lichtenvoorde; Dr. W.A. de Backer, Private Practice, Rijswijk; Dr. C.P. Buiks, Private Practice, Ewijk; Dr. H.F.C.M. Van Mierlo, General Practice Van Mierlo Rembrandt, van Rijn Singel 37-c, 2371 RB Roelofarendsveen; Dr. A. Veerman, Private Practice, Huizen; Dr. M. Passage, Private Practice, Kerkrade</p>", "<p><bold>Switzerland</bold>: Dr. med. Hans-Ulrich Rentsch, Rheumatologie, Poststrasse 25, 9000 St Gallen; Dr. R Theiler, Kantonsspital Aarau, Buchserstrasse/Haus 1, 5001 Aarau; Dr. med. Hans Schwarz, Rheumatologie Bethesda-Spital, Gellertstrasse 144, 4020 Basel; Dr. med. Michel Pellaton, 2, ruelle du Peyrou, 2000 Neuchâtel; Dr. Heinz Fahrer, Lindenhofspital/Rheumatologische Klinik, Salihaus Bremgartenstrasse 117, 3012 Bern; Dr. Ottmar Gorschewsky, Klinik Permanence Bern West, Orthopädie Bümplizstrasse 83, 3018 Bern; Dr. Thomas Lehmann, Inselspital/Rheumatologische Klinik Eingang-EG-29/Eingang 14a Freiburgstrasse, 3010 Bern; Dr. Paul Hasler, Felix Platter Spital, Rheumatologie Burgfelderstr. 101, 4012 Basel; Dr. med. Pierre-Alain Buchard, Rhumatologie FMH Clinique romande de réadaptation, Avenue Grand Champsec 90, 1951 Sion; Dr. med. Jean Dudler, Hôspital Nestlé, CHUV Rhumatologie FMH Avenue Pierre-Decker 5, 1005 Lausanne; Dr. Pierre-André Guerne, HCUG Rhumatolgoie, FMH Avenue Beau Séjour 25, 1211 Genève 14; Dr. med. Daniel Uebelhart, Universitätsspital Zürich Gloriastrasse 25, 8091 Zürich; Dr. Urs Moser, Rheumatologie Mühlegasse 3, 4410 Liestal; Dr. med. Michel Braun, Rhumatolgoie FMH Rue Gustave-Amweg 21, 2900 Porrentury</p>", "<p><bold>United Kingdom</bold>: Dr Alun George, The Staploe Medical Centre, The Staploe Medical Centre Brewhouse Lane Soham, CB7 5JD Cambridge; Dr Duncan Burwood, Bedgrove Surgery, Bedgrove Surgery Brentwood Way, HP21 7TL Aylesbury; Dr Andrew Cowie, The Porch Beechfield Road, Corsham, SN13 9 Wiltshire; Dr Robert Matthews, The Spa Surgery, The Spa Surgery 6 Spa Road, SN12 7NS Melksham; Dr Anthony Wright, Hathaway Surgery, Hathaway Surgery 32 New Road, SN15 1 Chippenham; Dr Kevin Gruffydd-Jones, Box Surgery, Box Surgery London Road, SN13 8NA Box, Corsham, Wiltshire</p>", "<title>Competing interests</title>", "<p>SY is an employee of Novartis Pharmaceuticals Corporation, East Hanover, NJ. KS and GK are employee of Novartis Pharma AG, Basel, Switzerland (the manufacturer of lumiracoxib). All authors own stocks of the company.</p>", "<title>Authors' contributions</title>", "<p>KS and GK participated in analysis and interpretation of the data. Godehard Hoexter performed the original statistical analysis which was then used by SY to conduct the statistical review of this manuscript. All authors contributed to drafting the manuscript. All authors read and approved the final manuscript.</p>", "<title>Pre-publication history</title>", "<p>The pre-publication history for this paper can be accessed here:</p>", "<p><ext-link ext-link-type=\"uri\" xlink:href=\"http://www.biomedcentral.com/1471-2474/9/118/prepub\"/></p>" ]

[ "<title>Acknowledgements</title>", "<p>The study and statistical analyses were funded by Novartis Pharma AG. The authors would like to thank the investigators (please see Appendix 1 for a list of full names), staff of the centres involved in this work, the clinical trial team for their expert collaboration and the patients who participated in the study. The authors would like to thank Godehard Hoexter, the clinical trial statistician and Elena Ehrsam the clinical trial leader for their expert contribution to the study and data analysis. In addition the authors would like to thank the medical writers Lakshmi Venkatraman and Vikrant Pallapotu (DOC-India, Novartis) for their assistance with drafting the manuscript and incorporating subsequent revisions.</p>" ]

[ "<fig position=\"float\" id=\"F1\"><label>Figure 1</label><caption><p><bold>Patient flow diagram</bold>. †Patients with multiple occurrences of a major protocol violation (PV) were counted only once in that category of PV.</p></caption></fig>", "<fig position=\"float\" id=\"F2\"><label>Figure 2</label><caption><p><bold>Incidence of peripheral oedema in patients treated with lumiracoxib and rofecoxib (safety population)</bold>. The incidence of peripheral oedema at Week 6. Pairwise comparisons tested at the two-sided 5% significance level. <italic>p</italic>-value computed using Fisher's exact test. OA, osteoarthritis.</p></caption></fig>", "<fig position=\"float\" id=\"F3\"><label>Figure 3</label><caption><p><bold>Lumiracoxib shows better blood pressure profile as compared to rofecoxib (safety population)</bold>. msSBP – Mean sitting systolic blood pressure. msDBP – Mean sitting diastolic blood pressure. <italic>p</italic>-value computed from ANCOVA on mean blood pressure at Day 42 with centre, treatment, and baseline blood pressure value. Mean change from baseline at Week 6. OA, osteoarthritis.</p></caption></fig>" ]

[ "<table-wrap position=\"float\" id=\"T1\"><label>Table 1</label><caption><p>Demographic and baseline characteristics (safety population)</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td/><td align=\"center\"><bold>Lumiracoxib 400 mg od</bold><break/><bold>(<italic>n </italic>= 154)</bold></td><td align=\"center\"><bold>Rofecoxib 25 mg od</bold><break/><bold>(<italic>n </italic>= 155)</bold></td></tr></thead><tbody><tr><td align=\"left\">Age (years)^†</td><td align=\"center\">65.3 ± 8.49</td><td align=\"center\">65.5 ± 8.67</td></tr><tr><td align=\"left\">Women, <italic>n </italic>(%)</td><td align=\"center\">94 (61.0)</td><td align=\"center\">100 (64.5)</td></tr><tr><td align=\"left\">BMI (kg/m²)^†</td><td align=\"center\">29.1 ± 5.21</td><td align=\"center\">28.3 ± 4.51</td></tr><tr><td align=\"left\">Race</td><td/><td/></tr><tr><td align=\"left\"> Caucasians, <italic>n </italic>(%)</td><td align=\"center\">154 (100.0)</td><td align=\"center\">153 (98.7)</td></tr><tr><td align=\"left\"> Other</td><td align=\"center\">0 (0.0)</td><td align=\"center\">2 (1.2)</td></tr><tr><td align=\"left\">Disease duration (years)^†</td><td align=\"center\">7.41 ± 7.058</td><td align=\"center\">8.37 ± 8.407</td></tr><tr><td align=\"left\">Physician's global assessment of disease activity <italic>n </italic>(%)</td><td/><td/></tr><tr><td align=\"left\"> Very good</td><td align=\"center\">1 (0.6)</td><td align=\"center\">2 (1.3)</td></tr><tr><td align=\"left\"> Good</td><td align=\"center\">3 (1.9)</td><td align=\"center\">5 (3.2)</td></tr><tr><td align=\"left\"> Fair</td><td align=\"center\">75 (48.7)</td><td align=\"center\">66 (42.6)</td></tr><tr><td align=\"left\"> Poor</td><td align=\"center\">69 (44.8)</td><td align=\"center\">76 (49.0)</td></tr><tr><td align=\"left\"> Very poor</td><td align=\"center\">6 (3.9)</td><td align=\"center\">6 (3.9)</td></tr><tr><td align=\"left\">Patient's global assessment of disease activity <italic>n </italic>(%)</td><td/><td/></tr><tr><td align=\"left\"> Very good</td><td align=\"center\">1 (0.6)</td><td align=\"center\">1 (0.6)</td></tr><tr><td align=\"left\"> Good</td><td align=\"center\">9 (5.8)</td><td align=\"center\">6 (3.9)</td></tr><tr><td align=\"left\"> Fair</td><td align=\"center\">51 (33.1)</td><td align=\"center\">57 (36.8)</td></tr><tr><td align=\"left\"> Poor</td><td align=\"center\">78 (50.6)</td><td align=\"center\">77 (49.7)</td></tr><tr><td align=\"left\"> Very poor</td><td align=\"center\">15 (9.7)</td><td align=\"center\">14 (9.0)</td></tr><tr><td align=\"left\">Pain intensity assessment <italic>n </italic>(%)</td><td/><td/></tr><tr><td align=\"left\"> Moderate</td><td align=\"center\">74 (48.1)</td><td align=\"center\">74 (47.7)</td></tr><tr><td align=\"left\"> Severe</td><td align=\"center\">67 (43.5)</td><td align=\"center\">67 (43.2)</td></tr><tr><td align=\"left\"> Extreme</td><td align=\"center\">13 (8.4)</td><td align=\"center\">14 (9.0)</td></tr><tr><td align=\"left\">Current smokers, <italic>n </italic>(%)</td><td align=\"center\">24 (15.6)</td><td align=\"center\">26 (16.8)</td></tr></tbody></table></table-wrap>", "<table-wrap position=\"float\" id=\"T2\"><label>Table 2</label><caption><p>Incidence of moderate or severe predefined GI AEs (safety population)</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td align=\"left\"><bold>Predefined GI AEs</bold></td><td align=\"center\"><bold>Lumiracoxib 400 mg od</bold><break/><bold>(<italic>n </italic>= 154)</bold><break/><bold><italic>n </italic>(%)</bold></td><td align=\"center\"><bold>Rofecoxib 25 mg od</bold><break/><bold>(<italic>n </italic>= 155)</bold><break/><bold><italic>n </italic>(%)</bold></td></tr></thead><tbody><tr><td align=\"left\">Abdominal pain</td><td align=\"center\">6 (3.9)</td><td align=\"center\">6 (3.9)</td></tr><tr><td align=\"left\">Constipation</td><td align=\"center\">1 (0.6)</td><td align=\"center\">1 (0.6)</td></tr><tr><td align=\"left\">Diarrhoea</td><td align=\"center\">4 (2.6)</td><td align=\"center\">2 (1.3)</td></tr><tr><td align=\"left\">Dyspepsia*</td><td align=\"center\">17 (11.0)</td><td align=\"center\">7 (4.5)</td></tr><tr><td align=\"left\">Dysphagia</td><td align=\"center\">1 (0.6)</td><td align=\"center\">0 (0.0)</td></tr><tr><td align=\"left\">Nausea</td><td align=\"center\">3 (1.9)</td><td align=\"center\">0 (0.0)</td></tr><tr><td align=\"left\">Vomiting</td><td align=\"center\">1 (0.6)</td><td align=\"center\">1 (0.6)</td></tr></tbody></table></table-wrap>", "<table-wrap position=\"float\" id=\"T3\"><label>Table 3</label><caption><p>Incidence of most frequent AEs (≥2% for either group) by preferred term (safety population)</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td/><td align=\"center\"><bold>Lumiracoxib 400 mg od</bold><break/><bold>(<italic>n </italic>= 154)</bold><break/><bold><italic>n </italic>(%)</bold></td><td align=\"center\"><bold>Rofecoxib 25 mg od</bold><break/><bold>(<italic>n </italic>= 155)</bold><break/><bold><italic>n </italic>(%)</bold></td></tr></thead><tbody><tr><td align=\"left\">Dyspepsia</td><td align=\"center\">41 (26.6)</td><td align=\"center\">33 (21.3)</td></tr><tr><td align=\"left\">Abdominal pain NOS</td><td align=\"center\">15 (9.7)</td><td align=\"center\">10 (6.5)</td></tr><tr><td align=\"left\">Diarrhoea NOS</td><td align=\"center\">15 (9.7)</td><td align=\"center\">7 (4.5)</td></tr><tr><td align=\"left\">Nausea</td><td align=\"center\">8 (5.2)</td><td align=\"center\">8 (5.2)</td></tr><tr><td align=\"left\">Abdominal pain upper</td><td align=\"center\">4 (2.6)</td><td align=\"center\">7 (4.5)</td></tr><tr><td align=\"left\">Constipation</td><td align=\"center\">4 (2.6)</td><td align=\"center\">1 (0.6)</td></tr><tr><td align=\"left\">Oedema lower limb</td><td align=\"center\">6 (3.9)</td><td align=\"center\">7 (4.5)</td></tr><tr><td align=\"left\">Fatigue</td><td align=\"center\">5 (3.2)</td><td align=\"center\">4 (2.6)</td></tr><tr><td align=\"left\">Nasopharyngitis</td><td align=\"center\">9 (5.8)</td><td align=\"center\">9 (5.8)</td></tr><tr><td align=\"left\">Influenza</td><td align=\"center\">6 (3.9)</td><td align=\"center\">6 (3.9)</td></tr></tbody></table></table-wrap>", "<table-wrap position=\"float\" id=\"T4\"><label>Table 4</label><caption><p>Incidence of deaths and SAEs (Safety population)</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td/><td align=\"center\"><bold>Lumiracoxib 400 mg od</bold><break/><bold>(<italic>n </italic>= 154)</bold></td><td align=\"center\"><bold>Rofecoxib 25 mg od</bold><break/><bold>(<italic>n </italic>= 155)</bold></td></tr></thead><tbody><tr><td align=\"left\">Patients with serious AEs</td><td/><td/></tr><tr><td align=\"left\"> Death <italic>n </italic>(%)</td><td align=\"center\">0 (0.0)</td><td align=\"center\">0 (0.0)</td></tr><tr><td align=\"left\"> Non-fatal SAEs <italic>n </italic>(%)</td><td align=\"center\">0 (0.0)</td><td align=\"center\">1 (0.6)</td></tr><tr><td align=\"left\">Patients with other significant AEs</td><td/><td/></tr><tr><td align=\"left\"> Pre-specified AEs (GI events or oedema) <italic>n </italic>(%)</td><td align=\"center\">73 (47.4)</td><td align=\"center\">64 (41.3)</td></tr><tr><td align=\"left\"> AEs leading to dose adjustment/interruption <italic>n </italic>(%)</td><td align=\"center\">0 (0.0)</td><td align=\"center\">3 (1.9)</td></tr><tr><td align=\"left\">Discontinuation due to</td><td/><td/></tr><tr><td align=\"left\"> Any AEs including SAEs <italic>n </italic>(%)</td><td align=\"center\">8 (5.2)</td><td align=\"center\">7 (4.5)</td></tr><tr><td align=\"left\"> SAEs <italic>n </italic>(%)</td><td align=\"center\">0 (0.0)</td><td align=\"center\">0 (0.0)</td></tr><tr><td align=\"left\"> AEs (non-serious) <italic>n </italic>(%)</td><td align=\"center\">8 (5.2)</td><td align=\"center\">7 (4.5)</td></tr></tbody></table></table-wrap>", "<table-wrap position=\"float\" id=\"T5\"><label>Table 5</label><caption><p>Efficacy results in patients treated with lumiracoxib and rofecoxib (ITT population)</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td/><td align=\"center\" colspan=\"2\"><bold>Week 3</bold></td><td align=\"center\" colspan=\"2\"><bold>Week 6</bold></td></tr></thead><tbody><tr><td align=\"left\"><bold>Efficacy measures</bold></td><td align=\"center\"><bold>Lumiracoxib 400 mg od</bold><break/><bold>(<italic>n </italic>= 154)</bold></td><td align=\"center\"><bold>Rofecoxib 25 mg od</bold><break/><bold>(<italic>n </italic>= 155)</bold></td><td align=\"center\"><bold>Lumiracoxib 400 mg od</bold><break/><bold>(<italic>n </italic>= 154)</bold></td><td align=\"center\"><bold>Rofecoxib 25 mg od</bold><break/><bold>(<italic>n </italic>= 155)</bold></td></tr><tr><td colspan=\"5\"><hr/></td></tr><tr><td align=\"left\">Patient's pain intensity</td><td/><td/><td/><td/></tr><tr><td align=\"left\"> Improved <italic>n </italic>(%)</td><td align=\"center\">53 (34.4)</td><td align=\"center\">50 (32.3)</td><td align=\"center\">49 (31.8)</td><td align=\"center\">63 (40.6)</td></tr><tr><td align=\"left\"> Non-improved <italic>n </italic>(%)</td><td align=\"center\">101 (65.6)</td><td align=\"center\">105 (67.7)</td><td align=\"center\">105 (68.2)</td><td align=\"center\">92 (59.4)</td></tr><tr><td align=\"left\">Patient's global assessment of disease activity</td><td/><td/><td/><td/></tr><tr><td align=\"left\"> Improved <italic>n </italic>(%)</td><td align=\"center\">49 (31.8)</td><td align=\"center\">53 (34.2)</td><td align=\"center\">57 (37.0)</td><td align=\"center\">65 (41.9)</td></tr><tr><td align=\"left\"> Non-improved <italic>n </italic>(%)</td><td align=\"center\">105 (68.2)</td><td align=\"center\">102 (65.8)</td><td align=\"center\">97 (63.0)</td><td align=\"center\">90 (58.1)</td></tr><tr><td align=\"left\">Physician's global assessment of disease activity</td><td/><td/><td/><td/></tr><tr><td align=\"left\"> Improved <italic>n </italic>(%)</td><td align=\"center\">44 (28.6)</td><td align=\"center\">46 (29.7)</td><td align=\"center\">51 (33.1)</td><td align=\"center\">56 (36.1)</td></tr><tr><td align=\"left\"> Non-improved <italic>n </italic>(%)</td><td align=\"center\">110 (71.4)</td><td align=\"center\">109 (70.3)</td><td align=\"center\">103 (66.9)</td><td align=\"center\">99 (63.9)</td></tr></tbody></table></table-wrap>" ]

[]

[]

[]

[]

[]

[]

[ "<table-wrap-foot><p>BMI = body mass index; Other = Black/African American and Asian or Pacific islander; SD = standard deviation.</p><p>^†Mean ± SD</p></table-wrap-foot>", "<table-wrap-foot><p>GI = gastrointestinal; AEs = adverse events</p><p>*<italic>p </italic>= 0.032</p></table-wrap-foot>", "<table-wrap-foot><p>AEs = adverse events; NOS = not otherwise specified</p></table-wrap-foot>", "<table-wrap-foot><p>AEs = adverse events; SAEs = serious adverse events; GI = gastrointestinal</p></table-wrap-foot>" ]

[ "<graphic xlink:href=\"1471-2474-9-118-1\"/>", "<graphic xlink:href=\"1471-2474-9-118-2\"/>", "<graphic xlink:href=\"1471-2474-9-118-3\"/>" ]

[]

{ "acronym": [], "definition": [] }

[ "<title>Background</title>", "<p>Mexican-Americans (MA) are known to exhibit increased prevalence of various cardiovascular disease (CVD) risk factors – e.g., obesity, hyperlipidemia and diabetes mellitus – compared to non-Hispanic whites (NHW), yet have been reported to have lower CVD case fatality rates than NHW [##REF##7291473##1##, ####REF##1984897##2##, ##REF##11216962##3####11216962##3##]. Although endothelial dysfunction (ED) is known to be an early marker of vascular disease-e.g., atherosclerosis-there is a lack of data examining ethnic differences in ED between <italic>asymptomatic </italic>MA and NHW.</p>", "<p>The Corpus Christi Heart Project investigators suggested that their finding of a greater hospitalized myocardial infarction rate – in the face of reported lower coronary heart disease (CHD) fatality rates among MA in vital statistics studies – is likely due to a misclassification of cause of death and/or ethnicity on death certificates [##REF##8508105##4##]. However, another possible explanation may be differences in degree of subclinical CVD [##REF##15769774##5##].</p>", "<p>Flow-mediated dilatation (FMD) of the brachial artery measured using ultrasound has been shown to be an early predictor of the risk of developing CVD [##REF##1984897##2##]. Microalbuminuria, also hypothesized to be related to vascular endothelial function, has previously been described as the earliest marker for the development of CHD in Type I diabetes mellitus [##REF##11455845##6##]. The aim of this pilot study was to obtain FMD and urinary albumin data in MA and NHW that might help explain the apparent paradox in MA. The hypotheses tested were: 1) There are significant differences in FMD between MA and NHW; these differences relate to ethnic differences in the levels of traditional CHD risk factors; and 2) FMD is associated with subclinical disease another urinary albumin measure, in one or more of these ethnic/gender subgroups.</p>" ]

[ "<title>Methods</title>", "<title>Subjects</title>", "<p>The study was conducted among asymptomatic community-based adult volunteers. To qualify for inclusion, subjects had to be classified as Mexican-American (MA) according to at least one of the following criteria used in the San Antonio Heart Study [##UREF##0##7##]: 1) Father's surname and mother's maiden name are both Spanish, and both parents were born in Mexico; and/or 2) Only one parent has a Spanish surname, but three of four grandparents have Mexican origins.</p>", "<p>We studied 105 adult MA (42 men and 63 women, age 46 ± 14 yrs) and 100 NHW (59 men and 41 women, age 50 ± 11 yrs) volunteers using blood tests, transthoracic echocardiography (echo), and brachial arterial flow-mediated dilatation (FMD) by ultrasound. Subjects were generally healthy: Individuals with hypertension or known CVD, or taking cardiovascular or BP medications were excluded.</p>", "<title>Echocardiography</title>", "<p>This study employed the echo protocol used in both the Cardiovascular Health Study (CHS) [##REF##7882482##8##] and Coronary Artery Risk Development in Young Adults (CARDIA) studies [##REF##7634452##9##]. For each subject, a baseline echo was recorded using a standardized protocol. Two-dimensionally guided M-mode echo measurements of the LV and left atrium were made according to conventions of the American Society of Echocardiography. LV mass was derived from the formula described by Devereux, et al. LV mass was normalized for various body size measures, including height [##REF##2521199##10##].</p>", "<title>Measurement of Flow-mediated Dilatation</title>", "<p>A 7.0 or 10 MHz linear array ultrasound transducer was used to image the right brachial artery 6 cm proximal to the antecubital fossa. Scanning was performed in the longitudinal view with transmit (focal) zone set to the depth of the arterial near-wall. After recording baseline images, a right brachial artery BP cuff was inflated to 30 mmHg above systolic pressure, occluding the artery for 4 minutes. The right brachial cuff was then deflated rapidly to zero pressure, resulting in reactive hyperemia. The arterial segment imaged at baseline was continuously imaged during cuff inflation and for 3 minutes following deflation. Brachial artery diameter after reactive hyperemia was expressed as a percentage of resting diameter.</p>", "<title>Blood/Urine Samples</title>", "<p>Fasting venous blood samples were obtained for analysis of serum electrolytes and creatinine; fasting total, HDL-, and LDL-cholesterol and triglycerides; and glucose. A spot urine was collected for albumin and creatinine determinations.</p>", "<title>Statistical Analysis</title>", "<p>Data are presented as mean ± standard deviations or percentages. Group comparisons were performed using Student's t-test for continuous variables and X²test for categorical variables. Due to a lack of data indicating an appropriate percent cut-point for normal vs. abnormal FMD in the MA population, the reported cut-point for NHW (abnormal FMD &lt; 7%) was examined in this study [##UREF##1##11##, ####REF##14769683##12##, ##REF##11702039##13####11702039##13##].</p>", "<p>Multiple linear regression analyses were performed to identify risk factors simultaneously predictive of FMD. Variables were selected for entry if the probability value was &lt; 0.10 in univariate testing of association. Modeling was done using the selected factors as independent variables and FMD as dependent variables. The full set of potential factors was considered with the forced entry method. To examine whether ethnicity (MA vs. NHW) remained significantly associated with FMD after adjusting for cardiovascular risk factors, multivariate analyses were conducted adjusting for parameters independently related to FMD – e.g., BMI, age, gender, systolic BP, LDL-cholesterol, brachial artery diameter, etc. – with and without urinary albumin. Additional multivariate analyses were performed using analysis of covariance (ANCOVA) with estimated marginal means and main effect options compared to determine the relation of urine albumin and CVD risk factors to FMD. ANCOVA models were created to assess the association between FMD and the various risk factors (and urine albumin) while adjusting for age or age and risk factors. The effect of each risk factor with a significant bivariate relationship to FMD was analyzed by ANCOVA. Models were constructed by adjusting for age (Model 1), age plus CVD risk factors (systolic BP, BMI, LDL-cholesterol, HDL-cholesterol, triglycerides, glucose, and current smoking) (Model 2), and age, CVD risk factors plus either urinary albumin (Model 3), or urine albumin:creatinine (Alb:Cr) ratio (Model 4).</p>" ]

[ "<title>Results and Discussion</title>", "<p>Study participant demographic, risk factor, and subclinical disease characteristics by ethnic group and gender are presented in Table ##TAB##0##1##. BMI and serum triglycerides were significantly higher in the overall MA versus the overall NHW cohort, while age, LDL-cholesterol, serum creatinine, and percent current smokers were all significantly lower in the MA than in the NHW cohort.</p>", "<p>In the overall cohort, MA demonstrated higher FMD compared to NHW (absolute diameter change: 0.0354 ± .02 cm vs 0.0280 ± 0.03 cm and % FMD: 9.1 ± 7.3% vs 7.1 ± 6.3%, respectively, both p &lt; 0.04). Brachial artery baseline diameters were similar in MA and NHW (0.389 ± 0.06 and 0.393 ± 0.07 cm, respectively).</p>", "<title>Bivariate analyses</title>", "<p>Bivariate analyses of the relationship between risk factor and subclinical disease variables and FMD were performed by ethnic-gender subgroup. Inverse relations were noted between FMD and both BMI and height in the overall MA cohort (r = -0.205, p &lt; 0.05 and r = -0.223, p &lt; 0.05, respectively), and also between FMD and height and weight in the overall NHW cohort (r = -0.257, p &lt; 0.01, and r = -0.216, p &lt; 0.05, respectively). In addition, in MA men, FMD was inversely related to all body size measures, whereas in NHW men, FMD was directly related to diastolic BP and HDL-cholesterol.</p>", "<p>Of additional interest, in MA men, an inverse relationship was noted between urinary albumin and FMD (r = -0.26, p &lt; 0.05). In contrast, no significant relationship was found between FMD and urinary albumin in MA women, or in NHW men or women. Furthermore, in women in both ethnic groups, no significant correlation was observed between any of the continuous or categorical variables and FMD-except for an inverse relation between Alb:Cr ratio and FMD in MA women.</p>", "<p>Study participants were further subgrouped based on ethnicity, gender and FMD values (FMD ≥ 7% vs FMD &lt; 7%). LV mass in the overall MA cohort with FMD ≥ 7% was significantly lower than in the overall NHW cohort with FMD ≥ 7% (p = 0.03), and lower in MA with FMD &lt; 7% compared with NHW participants with FMD &lt; 7%. Urinary albumin values were lower in the overall MA cohort than in the overall NHW cohort and in MA women with FMD% ≥ 7 than in MA women with FMD &lt; 7% (0.61 ± 0.7 vs 0.95 ± 0.8 mg/dl and 0.62 ± 0.5 vs 1.1 ± 1.1 mg/dl, p &lt; 0.006 and 0.05, respectively). In contrast, there was no significant relation between urinary albumin levels and FMD &lt; 7 vs ≥ 7% in the NHW cohort (0.76 ± 1.2 vs 0.69 ± 0.5 mg/dl).</p>", "<p>The relationship between FMD and urinary albumin is displayed in Figure ##FIG##0##1## for MA versus NHW men and for MA versus NHW women. Overall, FMD in MA and NHW men and women decreased as urinary albumin levels increased. Among women with urinary albumin levels of 1.1–1.5 mg/dl, FMD was significantly higher in MA than in NHW women (p &lt; 0.05) (Figure ##FIG##0##1##). In MA men with urinary albumin levels of 0–0.5 mg/dl, FMD trended higher than in NHW men with the same albumin levels (p &lt; 0.08). Meanwhile, in MA men with urinary albumin levels of 0.0–0.5 mg/dl, FMD was nearly significantly higher than in MA men with urinary levels of 0.6–1.0 mg/dl (p &lt; 0.06). In NHW men with urinary albumin levels of 0–0.5 mg/dl, FMD trended higher than in NHW men with urinary albumin levels of 1.1–1.5 mg/dl (p = 0.054).</p>", "<title>Multiple linear regression</title>", "<p>In MA men, BMI, systolic BP and urine albumin were predictor variables and inversely related to FMD. After age, brachial artery baseline diameter, BMI, systolic BP, LDL-cholesterol, HDL-cholesterol, triglycerides, glucose, and current smoking were forced into the multiple linear regression model, the <italic>inverse association of FMD and urinary albumin persisted </italic>(Table ##TAB##1##2##). In MA women, BMI, urine albumin and Alb:Cr ratio (β = -0.140, -28.57, and β = -0.211, respectively) were the most important predictor variables, and inversely related to FMD (all p &lt; 0.05). After forcing the other variables listed above into the model, an inverse association of FMD to Alb:Cr ratio persisted in MA women (p &lt; 0.05). In NHW men, only BMI and HDL-cholesterol were positive predictors of FMD, while in NHW women, only age and BMI were predictor variables, and inversely related, to FMD. In both NHW men and women participants, urinary albumin and Alb:Cr ratio were not associated with FMD after entering the above variables (Table ##TAB##1##2##).</p>", "<title>ANCOVA analyses</title>", "<p>FMD was significantly higher in MA compared with NHW individuals. Of interest, after adjustment for age alone, age plus CVD risk factors and brachial artery baseline diameter or age, CVD risk factors plus albumin or plus Alb:Cr ratio, FMD was not found to be significantly different between MA and NHW men or between MA and NHW women. After adjusting for age, CVD risk factors plus urinary albumin, FMD was significantly higher only in MA men compared to NHW men (7.8 ± 6.4% vs. 5.6 ± 4.5%, p &lt; 0.04). Alb:Cr ratio was not significantly associated with FMD in either cohort.</p>", "<p>FMD has been reported to be useful to assess long-term CVD risk in high-risk and lower-risk populations, and to predict short-term postoperative CVD event risk in a high-risk population [##REF##15947345##14##,##REF##11927524##15##]. Endothelial function in normal healthy adults is influenced by variables such as race [##UREF##1##11##,##REF##15947345##14##], age [##REF##11702039##13##,##REF##15947345##14##,##REF##10212168##16##], and prandial state [##REF##12204507##17##]. In healthy people, endothelial function, as measured by FMD, has been reported to be in the range of 7 to 10% (mean) [##UREF##1##11##, ####REF##14769683##12##, ##REF##11702039##13##, ##REF##15947345##14####15947345##14##], but in patients with CVD, FMD is impaired or absent, with FMD values ranging from 0 to 5%.</p>", "<p>In our study, the MA cohort demonstrated significantly higher FMD in comparison with the NHW cohort (9.1 ± 7.3% vs 7.1 ± 6.3%, p &lt; 0.04). These findings are of interest in view of a report from Bild and co-workers that the prevalence of CAC, measured from computed tomographic (CT) scanning, was lower in a Hispanic (55.6%) compared to a non-Hispanic Caucasian (70.4%) cohort⁵. Ethnic differences in endothelium-dependent FMD have also been observed in Chinese subjects, compared to white subjects from Australia [##REF##9207630##18##], and in Indian Asians in the United Kingdom compared with European whites [##REF##10212168##16##].</p>", "<p>Our findings of increased FMD in asymptomatic MA versus NHW, with no significant ethnic differences in baseline brachial artery diameter, may be due to a number of factors – e.g., an increase in eNOS activity in the MA versus the NHW cohort, the presence of cohort differences in other non-evaluated CHD risk factor(s) or biomarkers which also modulate endothelial function, other environmental or dietary factors, or perhaps genetic differences. A consistent, but non-significant increase in % FMD in both MA and NHW women, versus MA and NHW men, may be due to smaller BA diameters in women than men (Table ##TAB##0##1##).</p>", "<p>FMD has been shown previously to be decreased in obese individuals [##REF##16377286##19##]. In our study, FMD was lower in the asymptomatic MA cohort and was inversely associated with BMI. In contrast, there was no such association in NHW. This inverse relation between BMI and FMD in Hispanics is similar to a previous report [##REF##9207630##18##].</p>", "<p>Impaired vasodilatation of vascular endothelium (which predates atherosclerotic deposition) has been observed in apparently healthy patients with risk factors for CHD, including type 2 diabetes mellitus, but normoalbuminuria [##REF##15223221##20##,##REF##10446083##21##]. It has been suggested that endothelial dysfunction may be an earlier predictor than urinary albumin of the development of CHD and cardiovascular risk [##REF##10446083##21##,##UREF##2##22##]. Nonetheless, higher urinary albumin values, even at levels below clinical microalbuminuria, are generally associated with several measures of subclinical CVD in adults without established CVD [##REF##15223221##20##,##REF##1596306##23##,##REF##15956113##24##].</p>", "<p>In our MA, but not our NHW cohort, urinary albumin levels were consistently higher in subjects with FMD &lt; 7% versus those with FMD ≥ 7%. Urinary albumin levels were the only predictor (inverse) in multivariate analyses of FMD in MA men, while urinary Alb:Cr ratio was an inverse predictor of FMD in MA women. Of interest, in patients with persistent microalbuminuria, FMD has been reported to be significantly less in Caucasians than in African-Americans [##REF##12204507##17##].</p>", "<title>Limitations</title>", "<p>The sample size of our pilot study was relatively small, and a substantial proportion of the variability of brachial FMD in our cohort remains unexplained by standard CVD risk factors. A larger sample size might have helped to better characterize the contribution of race/ethnicity to FMD variability. The lack of an automated system for FMD analysis may be an added limitation in the study.</p>" ]

[ "<title>Results and Discussion</title>", "<p>Study participant demographic, risk factor, and subclinical disease characteristics by ethnic group and gender are presented in Table ##TAB##0##1##. BMI and serum triglycerides were significantly higher in the overall MA versus the overall NHW cohort, while age, LDL-cholesterol, serum creatinine, and percent current smokers were all significantly lower in the MA than in the NHW cohort.</p>", "<p>In the overall cohort, MA demonstrated higher FMD compared to NHW (absolute diameter change: 0.0354 ± .02 cm vs 0.0280 ± 0.03 cm and % FMD: 9.1 ± 7.3% vs 7.1 ± 6.3%, respectively, both p &lt; 0.04). Brachial artery baseline diameters were similar in MA and NHW (0.389 ± 0.06 and 0.393 ± 0.07 cm, respectively).</p>", "<title>Bivariate analyses</title>", "<p>Bivariate analyses of the relationship between risk factor and subclinical disease variables and FMD were performed by ethnic-gender subgroup. Inverse relations were noted between FMD and both BMI and height in the overall MA cohort (r = -0.205, p &lt; 0.05 and r = -0.223, p &lt; 0.05, respectively), and also between FMD and height and weight in the overall NHW cohort (r = -0.257, p &lt; 0.01, and r = -0.216, p &lt; 0.05, respectively). In addition, in MA men, FMD was inversely related to all body size measures, whereas in NHW men, FMD was directly related to diastolic BP and HDL-cholesterol.</p>", "<p>Of additional interest, in MA men, an inverse relationship was noted between urinary albumin and FMD (r = -0.26, p &lt; 0.05). In contrast, no significant relationship was found between FMD and urinary albumin in MA women, or in NHW men or women. Furthermore, in women in both ethnic groups, no significant correlation was observed between any of the continuous or categorical variables and FMD-except for an inverse relation between Alb:Cr ratio and FMD in MA women.</p>", "<p>Study participants were further subgrouped based on ethnicity, gender and FMD values (FMD ≥ 7% vs FMD &lt; 7%). LV mass in the overall MA cohort with FMD ≥ 7% was significantly lower than in the overall NHW cohort with FMD ≥ 7% (p = 0.03), and lower in MA with FMD &lt; 7% compared with NHW participants with FMD &lt; 7%. Urinary albumin values were lower in the overall MA cohort than in the overall NHW cohort and in MA women with FMD% ≥ 7 than in MA women with FMD &lt; 7% (0.61 ± 0.7 vs 0.95 ± 0.8 mg/dl and 0.62 ± 0.5 vs 1.1 ± 1.1 mg/dl, p &lt; 0.006 and 0.05, respectively). In contrast, there was no significant relation between urinary albumin levels and FMD &lt; 7 vs ≥ 7% in the NHW cohort (0.76 ± 1.2 vs 0.69 ± 0.5 mg/dl).</p>", "<p>The relationship between FMD and urinary albumin is displayed in Figure ##FIG##0##1## for MA versus NHW men and for MA versus NHW women. Overall, FMD in MA and NHW men and women decreased as urinary albumin levels increased. Among women with urinary albumin levels of 1.1–1.5 mg/dl, FMD was significantly higher in MA than in NHW women (p &lt; 0.05) (Figure ##FIG##0##1##). In MA men with urinary albumin levels of 0–0.5 mg/dl, FMD trended higher than in NHW men with the same albumin levels (p &lt; 0.08). Meanwhile, in MA men with urinary albumin levels of 0.0–0.5 mg/dl, FMD was nearly significantly higher than in MA men with urinary levels of 0.6–1.0 mg/dl (p &lt; 0.06). In NHW men with urinary albumin levels of 0–0.5 mg/dl, FMD trended higher than in NHW men with urinary albumin levels of 1.1–1.5 mg/dl (p = 0.054).</p>", "<title>Multiple linear regression</title>", "<p>In MA men, BMI, systolic BP and urine albumin were predictor variables and inversely related to FMD. After age, brachial artery baseline diameter, BMI, systolic BP, LDL-cholesterol, HDL-cholesterol, triglycerides, glucose, and current smoking were forced into the multiple linear regression model, the <italic>inverse association of FMD and urinary albumin persisted </italic>(Table ##TAB##1##2##). In MA women, BMI, urine albumin and Alb:Cr ratio (β = -0.140, -28.57, and β = -0.211, respectively) were the most important predictor variables, and inversely related to FMD (all p &lt; 0.05). After forcing the other variables listed above into the model, an inverse association of FMD to Alb:Cr ratio persisted in MA women (p &lt; 0.05). In NHW men, only BMI and HDL-cholesterol were positive predictors of FMD, while in NHW women, only age and BMI were predictor variables, and inversely related, to FMD. In both NHW men and women participants, urinary albumin and Alb:Cr ratio were not associated with FMD after entering the above variables (Table ##TAB##1##2##).</p>", "<title>ANCOVA analyses</title>", "<p>FMD was significantly higher in MA compared with NHW individuals. Of interest, after adjustment for age alone, age plus CVD risk factors and brachial artery baseline diameter or age, CVD risk factors plus albumin or plus Alb:Cr ratio, FMD was not found to be significantly different between MA and NHW men or between MA and NHW women. After adjusting for age, CVD risk factors plus urinary albumin, FMD was significantly higher only in MA men compared to NHW men (7.8 ± 6.4% vs. 5.6 ± 4.5%, p &lt; 0.04). Alb:Cr ratio was not significantly associated with FMD in either cohort.</p>", "<p>FMD has been reported to be useful to assess long-term CVD risk in high-risk and lower-risk populations, and to predict short-term postoperative CVD event risk in a high-risk population [##REF##15947345##14##,##REF##11927524##15##]. Endothelial function in normal healthy adults is influenced by variables such as race [##UREF##1##11##,##REF##15947345##14##], age [##REF##11702039##13##,##REF##15947345##14##,##REF##10212168##16##], and prandial state [##REF##12204507##17##]. In healthy people, endothelial function, as measured by FMD, has been reported to be in the range of 7 to 10% (mean) [##UREF##1##11##, ####REF##14769683##12##, ##REF##11702039##13##, ##REF##15947345##14####15947345##14##], but in patients with CVD, FMD is impaired or absent, with FMD values ranging from 0 to 5%.</p>", "<p>In our study, the MA cohort demonstrated significantly higher FMD in comparison with the NHW cohort (9.1 ± 7.3% vs 7.1 ± 6.3%, p &lt; 0.04). These findings are of interest in view of a report from Bild and co-workers that the prevalence of CAC, measured from computed tomographic (CT) scanning, was lower in a Hispanic (55.6%) compared to a non-Hispanic Caucasian (70.4%) cohort⁵. Ethnic differences in endothelium-dependent FMD have also been observed in Chinese subjects, compared to white subjects from Australia [##REF##9207630##18##], and in Indian Asians in the United Kingdom compared with European whites [##REF##10212168##16##].</p>", "<p>Our findings of increased FMD in asymptomatic MA versus NHW, with no significant ethnic differences in baseline brachial artery diameter, may be due to a number of factors – e.g., an increase in eNOS activity in the MA versus the NHW cohort, the presence of cohort differences in other non-evaluated CHD risk factor(s) or biomarkers which also modulate endothelial function, other environmental or dietary factors, or perhaps genetic differences. A consistent, but non-significant increase in % FMD in both MA and NHW women, versus MA and NHW men, may be due to smaller BA diameters in women than men (Table ##TAB##0##1##).</p>", "<p>FMD has been shown previously to be decreased in obese individuals [##REF##16377286##19##]. In our study, FMD was lower in the asymptomatic MA cohort and was inversely associated with BMI. In contrast, there was no such association in NHW. This inverse relation between BMI and FMD in Hispanics is similar to a previous report [##REF##9207630##18##].</p>", "<p>Impaired vasodilatation of vascular endothelium (which predates atherosclerotic deposition) has been observed in apparently healthy patients with risk factors for CHD, including type 2 diabetes mellitus, but normoalbuminuria [##REF##15223221##20##,##REF##10446083##21##]. It has been suggested that endothelial dysfunction may be an earlier predictor than urinary albumin of the development of CHD and cardiovascular risk [##REF##10446083##21##,##UREF##2##22##]. Nonetheless, higher urinary albumin values, even at levels below clinical microalbuminuria, are generally associated with several measures of subclinical CVD in adults without established CVD [##REF##15223221##20##,##REF##1596306##23##,##REF##15956113##24##].</p>", "<p>In our MA, but not our NHW cohort, urinary albumin levels were consistently higher in subjects with FMD &lt; 7% versus those with FMD ≥ 7%. Urinary albumin levels were the only predictor (inverse) in multivariate analyses of FMD in MA men, while urinary Alb:Cr ratio was an inverse predictor of FMD in MA women. Of interest, in patients with persistent microalbuminuria, FMD has been reported to be significantly less in Caucasians than in African-Americans [##REF##12204507##17##].</p>", "<title>Limitations</title>", "<p>The sample size of our pilot study was relatively small, and a substantial proportion of the variability of brachial FMD in our cohort remains unexplained by standard CVD risk factors. A larger sample size might have helped to better characterize the contribution of race/ethnicity to FMD variability. The lack of an automated system for FMD analysis may be an added limitation in the study.</p>" ]

[ "<title>Conclusion</title>", "<p>To our knowledge, this is the first study to analyze, in <italic>asymptomatic </italic>adults, the relation of MA and NHW ethnicity to FMD and urinary albumin levels. Overall, our MA cohort demonstrated higher FMD compared to the NHW cohort. Our results are consistent with growing evidence that risk factors contribute to the development of CVD, at least in part, by impairing endothelial function. Longitudinal follow-up of MA and NHW cohorts should help determine whether endothelial dysfunction, as measured by FMD, and urinary albumin provide additional value in predicting CVD outcomes in additional to traditional risk factors.</p>" ]

[ "<p>This is an Open Access article distributed under the terms of the Creative Commons Attribution License (<ext-link ext-link-type=\"uri\" xlink:href=\"http://creativecommons.org/licenses/by/2.0\"/>), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.</p>", "<title>Background-</title>", "<p>Mexican-Americans (MA) exhibit increases in various cardiovascular disease (CVD) risk factors compared to non-Hispanic Whites (NHW), yet are reported to have lower CVD mortality rates. Our aim was to help explain this apparent paradox by evaluating endothelial function and urine albumin levels in MA and NHW.</p>", "<title>Methods-</title>", "<p>One hundred-five MA and 100 NHW adults were studied by brachial artery flow-mediated dilatation (FMD), blood and urine tests. Participants were studied by ultrasound-determined brachial artery flow-mediated dilatation (FMD), blood and urine tests, at a single visit.</p>", "<title>Results-</title>", "<p>Despite higher BMI and triglycerides in MA, MA demonstrated higher FMD than did NHW (9.1 ± 7.3% vs. 7.1 ± 6.3%, p &lt; 0.04). Among MA, urinary albumin was consistently lower in participants with FMD ≥ 7% FMD versus &lt; 7% FMD (p &lt; 0.006). In multivariate analyses in MA men, urinary albumin was inversely related to FMD (r = -0.26, p &lt; 0.05), as were BMI and systolic blood pressure. In MA women, urinary albumin:creatinine ratio was an independent inverse predictor of FMD (p &lt; 0.05 ).</p>", "<title>Conclusion-</title>", "<p>To our knowledge, this is the first study to analyze, in asymptomatic adults, the relation of MA and NHW ethnicity to FMD and urine albumin levels. The findings confirm ethnic differences in these important subclinical CVD measures.</p>" ]

[ "<title>Competing interests</title>", "<p>The authors declare that they have no competing interests.</p>", "<title>Authors' contributions</title>", "<p>JMD have made substantial contributions to conception and design of the study, interpretation of data and writing of the manuscript and have given final approval of the version to be published. ZA performed the statistical analysis and drafting and revising the manuscript. NDW participated in the design of the study and revising the manuscript critically for important intellectual content. SKS instrumental in the recruitment of study participants and in acquisition of the data. RLB carried out the echocardiography and flow-mediated dilatation studies. MAS participated in the study design and coordination and helped to draft the manuscript. HAP participated in the study design and coordination and helped to draft the manuscript. All authors read and approved the final manuscript.</p>", "<title>Acknowledgements</title>", "<p>This research was supported by a Grant-in-Aid from the American Heart Association, National Association (99505-47N) to Dr. Gardin and by funding from St. John Guild Cardiovascular Research Endowment.</p>" ]

[]

[ "<fig position=\"float\" id=\"F1\"><label>Figure 1</label><caption><p>Flow-mediated dilatation and urine albumin levels in Mexican-American and non-Hispanic white men (top panel) and women (bottom panel).</p></caption></fig>" ]

[ "<table-wrap position=\"float\" id=\"T1\"><label>Table 1</label><caption><p>Distribution of flow-mediated dilatation, and demographic and cardiovascular risk factors among Mexican-American (MA) and non-Hispanic white (NHW) cohorts of men and women</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td/><td align=\"center\" colspan=\"2\"><bold>Overall</bold></td><td align=\"center\" colspan=\"2\"><bold>Men</bold></td><td align=\"center\" colspan=\"2\"><bold>Women</bold></td></tr><tr><td/><td align=\"center\" colspan=\"2\">Mean ± SD</td><td align=\"center\" colspan=\"2\">Mean ± SD</td><td align=\"center\" colspan=\"2\">Mean ± SD</td></tr><tr><td align=\"left\"><bold>Parameter</bold></td><td align=\"center\"><bold>MA </bold>(N = 105)</td><td align=\"center\"><bold>NHW </bold>(N = 100)</td><td align=\"center\"><bold>MA </bold>(N = 39)</td><td align=\"center\"><bold>NHW </bold>(N = 59)</td><td align=\"center\"><bold>MA </bold>(N = 66)</td><td align=\"center\"><bold>NHW </bold>(N = 41)</td></tr></thead><tbody><tr><td align=\"left\"><bold>Risk Factors</bold></td><td/><td/><td/><td/><td/><td/></tr><tr><td align=\"left\">Age (yrs)</td><td align=\"center\">46 ± 13.5</td><td align=\"center\">49.6 ± 10.8*</td><td align=\"center\">46.3 ± 13.6</td><td align=\"center\">47.7 ± 11.3</td><td align=\"center\">45.9 ± 13.6</td><td align=\"center\">52.4 ± 9.5**</td></tr><tr><td align=\"left\">Body Mass Index (kg/m²)</td><td align=\"center\">30.3 + 7.1</td><td align=\"center\">27.8 + 4.2**</td><td align=\"center\">30.9 ± 6.1</td><td align=\"center\">28.4 ± 3.5**</td><td align=\"center\">29.98 ± 7.6</td><td align=\"center\">26.9 ± 4.95*</td></tr><tr><td align=\"left\">Triglycerides (mg/dl)</td><td align=\"center\">172.6 ± 115.9</td><td align=\"center\">124.9 ± 66.6**</td><td align=\"center\">184.7 ± 92</td><td align=\"center\">135.7 ± 71**</td><td align=\"center\">166 ± 126.9</td><td align=\"center\">109.4 ± 56.2**</td></tr><tr><td align=\"left\">LDL cholesterol (mg/dl)</td><td align=\"center\">110.7 ± 29.4</td><td align=\"center\">123.5 ± 30.4**</td><td align=\"center\">112.8 ± 27.7</td><td align=\"center\">127.5 ± 27.7**</td><td align=\"center\">109.5 ± 30.4</td><td align=\"center\">117.8 ± 33.5</td></tr><tr><td align=\"left\">HDL cholesterol (mg/dl)</td><td align=\"center\">53.6 + 13.9</td><td align=\"center\">53.7 + 12.8</td><td align=\"center\">45.5 + 6</td><td align=\"center\">48.8 + 10*</td><td align=\"center\">57.9 + 14.2</td><td align=\"center\">60.9 + 12.6</td></tr><tr><td align=\"left\">Total Cholesterol (mg/dl)</td><td align=\"center\">198.7 ± 35.9</td><td align=\"center\">203.5 ± 37.5</td><td align=\"center\">195.6 ± 33.1</td><td align=\"center\">205.5 ± 37.9</td><td align=\"center\">200.3 ± 37.4</td><td align=\"center\">200.6 + 37.2</td></tr><tr><td align=\"left\">Systolic BP (mmHg)</td><td align=\"center\">124.5 ± 17.4</td><td align=\"center\">125.2 ± 14.4</td><td align=\"center\">129 ± 12.9</td><td align=\"center\">127.4 ± 13.4</td><td align=\"center\">121.9 ± 19</td><td align=\"center\">122.1 ± 15.4</td></tr><tr><td align=\"left\">Diastolic BP (mm Hg)</td><td align=\"center\">72.3 ± 10.6</td><td align=\"center\">74.6 ± 9.8</td><td align=\"center\">79.1 ± 9.9</td><td align=\"center\">77.4 ± 9.9</td><td align=\"center\">68.5 ± 9</td><td align=\"center\">70.5 ± 8.3</td></tr><tr><td align=\"left\">Serum Glucose (mg/dl)</td><td align=\"center\">90.8 ± 13.4</td><td align=\"center\">88.7 ± 21.7*</td><td align=\"center\">91.7 ± 13.1</td><td align=\"center\">91.2 ± 26.6</td><td align=\"center\">90.3 ± 13.7</td><td align=\"center\">85 ± 10.8*</td></tr><tr><td align=\"left\">Serum Creatinine (mg/dl)</td><td align=\"center\">0.8 ± 0.2</td><td align=\"center\">0.9 ± 0.2**</td><td align=\"center\">0.94 ± 0.1</td><td align=\"center\">0.98 ± 0.1*</td><td align=\"center\">0.73 ± 0.1</td><td align=\"center\">0.8 ± 0.1**</td></tr><tr><td align=\"left\">Urinary albumin (mg/dl)</td><td align=\"center\">0.88 ± 0.85</td><td align=\"center\">1.1 ± 2.3</td><td align=\"center\">0.9 ± 0.9</td><td align=\"center\">1.3 ± 2.9</td><td align=\"center\">1.1 ± 1.1</td><td align=\"center\">0.5 ± 0.3**</td></tr><tr><td align=\"left\">Brachial Artery Baseline</td><td/><td/><td/><td/><td/><td/></tr><tr><td align=\"left\">Diameter (cm)</td><td align=\"center\">0.389 ± 0.06</td><td align=\"center\">0.393 ± 0.07</td><td align=\"center\">0.41 ± 0.05</td><td align=\"center\">0.418 ± 0.04</td><td align=\"center\">0.322 ± 0.04</td><td align=\"center\">0.32 ± 0.05</td></tr><tr><td align=\"left\">Non-smokers (%)</td><td align=\"center\">65%</td><td align=\"center\">39%**</td><td align=\"center\">19%</td><td align=\"center\">27%</td><td align=\"center\">46%</td><td align=\"center\">14%**</td></tr><tr><td align=\"left\">Current smokers (%)</td><td align=\"center\">35%</td><td align=\"center\">61%**</td><td align=\"center\">16%</td><td align=\"center\">33%</td><td align=\"center\">19%</td><td align=\"center\">26%</td></tr><tr><td/><td/><td/><td/><td/><td/><td/></tr><tr><td align=\"left\"><bold>Subclinical Disease</bold></td><td/><td/><td/><td/><td/><td/></tr><tr><td align=\"left\">Flow-mediated Dilatation %</td><td align=\"center\">9.1 ± 7.3</td><td align=\"center\">7.1 ± 6.3*</td><td align=\"center\">8.7 ± 6.6</td><td align=\"center\">6.6 ± 4.1</td><td align=\"center\">9.2 ± 7.4</td><td align=\"center\">8.9 ± 7.3</td></tr></tbody></table></table-wrap>", "<table-wrap position=\"float\" id=\"T2\"><label>Table 2</label><caption><p>Multiple linear regression of CHD risk factors and urinary albumin versus flow-mediated dilatation in Mexican-American and non-Hispanic white men and women</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td/><td align=\"center\" colspan=\"2\">Flow-Mediated Dilatation (%)</td><td align=\"center\" colspan=\"2\">Flow-Mediated Dilatation (%)</td></tr><tr><td/><td align=\"center\" colspan=\"2\">Standardized β-Coefficient</td><td align=\"center\" colspan=\"2\">Standardized β-Coefficient</td></tr><tr><td align=\"center\">Variable</td><td align=\"center\">MA men</td><td align=\"center\">MA women</td><td align=\"center\">NHW men</td><td align=\"center\">NHW women</td></tr></thead><tbody><tr><td align=\"left\">Constant</td><td align=\"center\">-4.848</td><td align=\"center\">23.7</td><td align=\"center\">-22.240</td><td align=\"center\">4.3</td></tr><tr><td align=\"left\">Age (years)</td><td align=\"center\">0.042</td><td align=\"center\">-0.090</td><td align=\"center\">-0.045</td><td align=\"center\">-0.230*</td></tr><tr><td align=\"left\">BA Baseline Diameter</td><td align=\"center\">0.053</td><td align=\"center\">0.069</td><td align=\"center\">0.081</td><td align=\"center\">0.074</td></tr><tr><td align=\"left\">BMI</td><td align=\"center\">-15.29*</td><td align=\"center\">-0.140*</td><td align=\"center\">0.539*</td><td align=\"center\">-0.241*</td></tr><tr><td align=\"left\">Systolic BP (mmHg)</td><td align=\"center\">-.311*</td><td align=\"center\">0.013</td><td align=\"center\">0.039</td><td align=\"center\">0.074</td></tr><tr><td align=\"left\">LDL-Cholesterol (mg/dl)</td><td align=\"center\">0.002</td><td align=\"center\">-0.077</td><td align=\"center\">-0.009</td><td align=\"center\">-0.001</td></tr><tr><td align=\"left\">HDL-Cholesterol (mg/dl)</td><td align=\"center\">-0.084</td><td align=\"center\">-0.014</td><td align=\"center\">0.154*</td><td align=\"center\">0.081</td></tr><tr><td align=\"left\">Triglycerides (mg/dl)</td><td align=\"center\">-0.006</td><td align=\"center\">0.002</td><td align=\"center\">0.019</td><td align=\"center\">0.012</td></tr><tr><td align=\"left\">Serum Glucose (mg/dl)</td><td align=\"center\">-0.009</td><td align=\"center\">0.033</td><td align=\"center\">-0.018</td><td align=\"center\">0.026</td></tr><tr><td align=\"left\">Current Smoking (1 = yes)</td><td align=\"center\">-0.031</td><td align=\"center\">0.001</td><td align=\"center\">-0.002</td><td align=\"center\">0.112</td></tr><tr><td align=\"left\">Urine Albumin (mg/dl)</td><td align=\"center\">-26.95**</td><td align=\"center\">-28.57*</td><td align=\"center\">3.425</td><td align=\"center\">-15.973</td></tr><tr><td align=\"left\">Alb: Cr Ratio</td><td align=\"center\">-0.118</td><td align=\"center\">-0.211*</td><td align=\"center\">-0.046</td><td align=\"center\">0.147</td></tr></tbody></table></table-wrap>" ]

[]

[]

[]

[]

[]

[]

[ "<table-wrap-foot><p>*: p &lt; 0.05; **: p &lt; 0.01. P values represent the comparison between MA vs NHW for overall, men and women cohort. Abbreviations: LDL: low density lipoprotein; HDL: high density lipoprotein.</p></table-wrap-foot>", "<table-wrap-foot><p>*p &lt; 0.05, ** p &lt; 0.01, LDL, low density lipoprotein; HDL, high density lipoprotein; Alb:Cr ratio, Albumin:creatinine ratio</p></table-wrap-foot>" ]

[ "<graphic xlink:href=\"1476-7120-6-43-1\"/>" ]

[]

{ "acronym": [], "definition": [] }

[ "<title>Background</title>", "<p>The prevalence of both systolic and diastolic heart failure is high and the prognosis is comparably poor. The prevalence of diastolic heart failure is increasing and the survival rates remain low, whereas the survival rates of systolic heart failure have improved in recent years. Diastolic heart failure is characterized by abnormal myocardial relaxation and increased passive stiffness and is hard to distinguish from systolic heart failure by clinical examination alone [##REF##16855265##1##, ####REF##16003647##2##, ##REF##16386668##3##, ##REF##16488928##4##, ##REF##16401819##5##, ##REF##15356307##6##, ##REF##15659738##7##, ##REF##15128895##8####15128895##8##].</p>", "<p>It was suggested that there is no isolated diastolic dysfunction, but that there is a continuum from normal to impaired diastolic and then systolic dysfunction (concept of \"single syndrome\"). The impairment of the longitudinal systolic function measured by Tissue Doppler Imaging (TDI) in patients with diastolic dysfunction supports this concept [##REF##11796546##9##, ####REF##14594874##10##, ##REF##16387829##11####16387829##11##]. In contrast, the \"two syndromes\" concept claims two different pathological entities in systolic and diastolic heart failure. However, there is no final consensus on the taxonomy and the optimal diagnostic algorithm for the detection of diastolic dysfunction. Symptoms of heart failure, biomarkers, stress tests, and echocardiography are important elements in the diagnostic work up [##REF##17428822##12##].</p>", "<p>The non-invasive methods to diagnose heart failure include echocardiography and cardiac biomakers. Increased left-ventricular filling pressure is believed to induce myocardial wall-stress, the release of natriuretic peptides and increased E/E'-ratio. Diastolic heart failure is unlikely if E/E' is &lt; 8, an E/E' 8–15 is suggestive of but not diagnostic of diastolic heart failure [##REF##14981433##13##,##REF##9362412##14##].</p>", "<p>In addition to myocardial velocity measurements, new sensitive TDI-derived measurements of systolic function were introduced recently: Strain and longitudinal displacement [##REF##15220909##15##]. Strain measures compression and distension of myocardial segments (\"deformation imaging\") and is reduced in systolic dysfunction [##REF##9812093##16##,##REF##17498573##17##]. Longitudinal displacement measures myocardial motion amplitudes in systole and visualizes the segmental velocity-time integrals. It correlates with longitudinal systolic function [##REF##12618734##18##] but has not yet been examined in isolated diastolic dysfunction.</p>", "<p>Natriuretic peptides (BNP and the hormonally inactive NT-proBNP) are significantly elevated in systolic and in less so in diastolic heart failure [##REF##11827925##19##, ####REF##12892964##20##, ##REF##16014646##21##, ##REF##17118955##22####17118955##22##]. NT-proBNP correlates to prognosis in systolic heart failure [##REF##16293638##23##,##REF##14987581##24##], but its diagnostic value in stable asymptomatic patients with diastolic heart failure is still controversial.</p>", "<p>The aim of this study was to assess the diagnostic value of NT-proBNP and the concordance with Tissue Doppler Echocardiography (Strain imaging, longitudinal displacement, E/E') in diastolic and systolic heart failure. We postulate that the integration of these parameters improves the severity estimation of systolic and diastolic heart failure.</p>" ]

[ "<title>Methods</title>", "<title>Patients</title>", "<p>In this prospective monocentric study (enrolment 2004–2005), 137 consecutive clinically stable out- and in-patients with a clinical indication for echocardiography from medical and surgical departments were included. Exclusion criteria: atrial fibrillation, relevant valvular heart disease exceeding mild mitral or aortic valve disease, prosthetic heart valves, pulmonary hypertension, myocardial infarction &lt; 3 months prior to study inclusion, terminal renal failure, creatinine &gt; 2.5 mg/dl, pregnancy, age &lt; 18 years. The blood for NT-proBNP measurements (Elecsys proBNP, Roche Diagnostics, Germany [##REF##15387451##25##] was drawn after echocardiography, centrifuged and frozen at -80°C immediately. The echocardiography examiners were blinded to the NT-proBNP values. The creatinine clearance was calculated as previously described [##REF##11904577##26##,##REF##1244564##27##].</p>", "<p>Written informed consent was obtained from each patient. The ethics committee of the Charité University Hospital approved the protocol.</p>", "<title>Echocardiography</title>", "<p>Transthoracic echocardiography was performed according to the ASE recommendations [##REF##709763##28##] by Vivid 7 Dimension (M3S 1.5–4.0 MHz transducer; GE Vingmed, Horton, Norway). The images were stored digitally and analyzed off-line by EchoPac PC Dimension (GE Vingmed, Horton Norway). Echocardiographic examinations included the trans-mitral inflow profile (E/A), TDI measurement of the ratio of the early-to-late annular velocity (E'/A') in the basal septum and left lateral myocardium at the mitral annulus, the E/E' ratio using the average of the basal septal and basal lateral E', the systolic basal septal and lateral myocardial velocities (S') as well as basal septal and lateral Strain and longitudinal displacement. (For acquisition of the TDI images: see <bold>additional file </bold>##SUPPL##0##1## and <bold>additional file </bold>##SUPPL##1##2##]. All measurements were performed in the apical four chamber view; three beats were stored and analyzed.</p>", "<p>The LVEF was calculated according to Simpson's rule [##REF##2698218##29##]. A normal LVEF was defined as ≥ 55%, 30–55% is mild-moderately abnormal, a LVEF &lt; 30% is severely abnormal according to [##REF##16376782##30##]. LV mass was computed according to the ASE cube method [##REF##2936235##31##].</p>", "<p>Diastolic heart failure was defined as previously described: normal LVEF (≥ 55%) [##REF##16376782##30##], E/E' &gt; 10 [##REF##14981433##13##,##REF##9362412##14##], E/A &lt; 1 [##REF##9247521##32##,##REF##15894236##33##]. The transmital flow and TDI parameters were adjusted to age-related cut-points according to [##REF##15894236##33##,##REF##16640715##34##].</p>", "<p>The patients were classified as normal controls <bold>(group 1)</bold>, diastolic heart failure with preserved left ventricular function (LVEF ≥ 55%, <bold>group 2</bold>), systolic heart failure (LVEF &lt; 55%, <bold>group 3</bold>).</p>", "<title>Statistics</title>", "<p>Statistics were calculated by SPSS (version 12.0, Chicago, Ill, USA). Descriptive statistics of parametric variables are expressed as mean (± SD). Nonparametric variables are expressed as median (inter-quartile range, IQR, of 25 and 75 percentiles).</p>", "<p>The comparison of echocardiographic parameters between groups was calculated by Wilcoxon test for non-parametric data. The Dunnett test was used for comparison to normal findings [##UREF##0##35##]. Dichotomized data were analyzed by the Chi²-test. The level of significance was p = 0.05.</p>", "<p>ROC (Receiver Operator Characteristics) analysis was performed to calculate sensitivity, specificity, negative and positive predictive values and an optimal cut-point of NT-proBNP to detect systolic or diastolic dysfunction. The optimal cut-off point was assessed according to Youden [##REF##15405679##36##].</p>" ]

[ "<title>Results</title>", "<p>137 patients were included. 42 patients had normal systolic and diastolic function (group 1), 43 patients had diastolic dysfunction (group 2), 52 patients had systolic dysfunction with an EF &lt; 55% (group 3).</p>", "<p>The baseline characteristics of the patients are listed in table ##TAB##0##1##; the echocardiographic findings are listed in table ##TAB##1##2##. Patients with reduced LVEF (&lt; 55%) had significantly increased NT-proBNP values compared to the healthy controls. Strain and longitudinal displacement parameters were significantly reduced in severely reduced LVEF compared to controls (table ##TAB##2##3##, figure ##FIG##0##1##).</p>", "<p>The ROC analysis to discriminate between normal LVEF (n = 88) and reduced LVEF (n = 49) had an area under the curve of 0.844, which indicates a good diagnostic accuracy. The best cut-off for this discrimination was 489 pg/ml (sensitivity 81.6%, specificity 85.2%, PPV 75.5% and NPV 89.3%, OR 25.6, Youden Index 0.67). The ROC analysis to discriminate between normal echocardiography (n = 42) and impaired diastolic and/or systolic function (n = 95) had an area under the curve of 0.763, which indicates a fair diagnostic accuracy. The best cut-off for this discrimination was 97 pg/ml (sensitivity 80.4%, specificity 64.3%, PPV 83.5% and NPV 58.7%, OR 7.2, Youden Index 0.44) (figure ##FIG##1##2##)</p>", "<p>Dividing the patients with preserved systolic LV function (groups 1 and 2) by E/E' (cut-point = 8) showed that there were significant differences in NT-proBNP levels (E/E' &lt; 8: median NT-proBNP: 45.8, IQR: 172.6 pg/ml, E/E' &gt; 8: 114.6 (261.7), p = 0.01). Classifying these patients by E/E' &lt; 8, E/E' 8–15 and E/E' &gt; 15 according to [##REF##9362412##14##,##REF##15894236##33##] revealed that those with increased filling pressures (E/E' &gt; 15) had significantly elevated NT-proBNP and reduced longitudinal displacement values compared to patients with E/E' &lt; 8. Strain, in contrast, was not significantly impaired. The patients with an E/E' 8–15 did not differ significantly in NT-proBNP levels compared to patients with E/E' &lt; 8 or E/E' &gt; 15 (figures ##FIG##2##3## and ##FIG##3##4##).</p>", "<p>A division by E/A (cut-point = 1, p = 0.34) or E'/A' (cut-point = 1, p = 0.54) was not associated with significantly different NT-proBNP levels. This indicates that increased E/E' is most closely linked to elevated NT-proBNP.</p>", "<p>There was a correlation of peak systolic velocities (S') and longitudinal displacement with NT-proBNP throughout the spectrum of our patients (Spearman correlation coefficient was -0.578 (p &lt; 0.0001) for longitudinal displacement and -0.605 (p &lt; 0.001) for S').</p>" ]

[ "<title>Discussion</title>", "<p>The main finding of this study is that patients with diastolic and systolic myocardial dysfunction have significantly reduced basal left ventricular septal and lateral strain compared to healthy controls. Patients with normal LVEF and elevated left ventricular filling pressures (E/E' &gt; 15) have significantly reduced longitudinal displacement and significantly elevated NT-proBNP. NT-proBNP can detect patients with a reduced left ventricular systolic function (LVEF &lt; 55%) with a good diagnostic accuracy in accordance with previous studies [##REF##11827925##19##, ####REF##12892964##20##, ##REF##16014646##21##, ##REF##17118955##22##, ##REF##16293638##23##, ##REF##14987581##24####14987581##24##].</p>", "<p>In the patients with normal LV function and an elevated E/E' (8 – 15), the measurement of NT-proBNP does not add a significant diagnostic information. Therefore, the diagnostic accuracy of NT-proBNP to detect diastolic dysfunction in these patients is low. Our findings do not fully support the algorithm recently published consensus statement for the diagnosis of diastolic heart failure that emphasized E/E' and NT-proBNP [##REF##17428822##12##].</p>", "<p>In the grey zone of mildly elevated NT-proBNP and/or an E/E' between 8 and 15, it remains difficult to diagnose diastolic heart failure. This could be due to the weak correlation of invasively measured LA filling pressures and E/E' in this grey zone. The majority of patients with normal LVEF and invasively measured elevated LA filling pressures had an E/E' between 8 and 15 [##REF##11023933##37##].</p>", "<p>This underlines the importance of integrating multiple echocardiographic parameters and an individual interpretation by experienced cardiologists. Especially in the case of elevated NT-proBNP of non-cardiac causes (e.g. renal failure), the diagnostic work-up has to rely on echocardiography alone. In these borderline cases, invasive measurements by a conductance catheter should be considered to exclude relevant diastolic dysfunction [##REF##17646587##38##].</p>", "<p>Why is NT-proBNP not significantly elevated in diastolic dysfunction compared to healthy controls, but correlates to filling pressures (E/E')? We speculate that diastolic dysfunction is a process that includes a variety of mechanisms (LV hypertrophy, abnormal active relaxation, increased stiffness, increased filling pressures). The trigger for NT-proBNP release from cardiomyocytes is primarily wall stress, which is functionally reflected by filling pressures (E/E').</p>", "<p>Our findings support the \"single syndrome\" theory of heart failure. Subtle changes of longitudinal myocardial function (reflected by Strain and longitudinal displacement) begin in diastolic heart failure and are further increased in systolic heart failure. However, the study by Yip [##REF##11796546##9##] has seen a progressive decline of left ventricular long axis function (systolic peak mitral annular velocity) in patients with diastolic heart failure compared to healthy controls. We found that peak systolic velocities are not significantly reduced in diastolic dysfunction. The cut-off of a normal ejection fraction in their study was 45%. According to [##REF##16376782##30##], a normal LVEF is ≥ 55%. For this reason, the results of [##REF##11796546##9##] are possibly explained by inclusion bias of patients with reduced LVEF. In their study neither E/E', Strain or Tracking were measured. In concordance with our results, Dong [##REF##16880097##39##] did not note a reduction of TDI systolic velocity (S') in patients with diastolic dysfunction.</p>", "<p>The cut-off values of LVEF, NT-proBNP and E/E' need to be discussed, because the different studies have used different cut-offs. The definition of a normal LVEF &gt; 50% in the ESC Guidelines [##REF##17428822##12##] is somewhat arbitrary. Lang [##REF##16376782##30##] has suggested 55% as the cut-off for a normal LVEF (used in our study). The lower threshold of 50% will automatically include patients with impairment of longitudinal function and will therefore alter the sensitivity and specificity for the detection diastolic dysfunction.</p>", "<p>The NT-proBNP values physiologically increase with age [##REF##15979602##40##]; therefore the threshold of 220 pg/ml in the Guidelines by Paulus [##REF##17428822##12##] will lead to impaired diagnostic accuracy and false-positives in the elderly patients.</p>" ]

[ "<title>Conclusion</title>", "<p>In conclusion, we found that in patients with systolic and diastolic heart failure, E/E', NT-proBNP as well as Strain and longitudinal displacement add important incremental information for the severity estimation of heart failure. In patients with isolated diastolic dysfunction, Strain is significantly reduced and with increased fillings pressures longitudinal displacement is impaired, paralleled by an increase of NT-proBNP. But in a substantial subset of patients with borderline NT-proBNP and E/E', an individual analysis of all available data has to be performed.</p>" ]

[ "<p>This is an Open Access article distributed under the terms of the Creative Commons Attribution License (<ext-link ext-link-type=\"uri\" xlink:href=\"http://creativecommons.org/licenses/by/2.0\"/>), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.</p>", "<title>Background</title>", "<p>The aim of this prospective study was to assess the diagnostic value of NT-proBNP and the concordance with Tissue Doppler Echocardiography (including strain and longitudinal displacement) in diastolic and systolic heart failure.</p>", "<title>Methods and results</title>", "<p>137 consecutive clinically stable patients were included (42 healthy controls, 43 with diastolic heart failure, 52 with systolic heart failure). In diastolic heart failure, basal septal strain was reduced (-24.8 ± 8.1% vs. controls. -18.5 ± 5.3%, p &lt; 0.0001). In all patients with preserved systolic function, septal basal longitudinal displacement was impaired in patients with increased left-ventricular filling pressures (E/E' &lt; 8: 13.5 mm ± 3.3 mm vs. E/E' &gt; 15: 8.5 mm ± 2.3 mm, p = 0.001) parallel to NT-proBNP elevation (E/E' &lt; 8: 45.8 pg/ml, IQR: 172.5 pg/ml vs. E/E' &gt; 15: 402.0 pg/ml, IQR: 1337.2 pg/ml; p = 0.0007). In ROC analysis, NT-proBNP could detect patients with reduced left ventricular systolic function (LVEF ≥ 55%) with a good diagnostic accuracy. However, the diagnostic accuracy of NT-proBNP to detect diastolic dysfunction was lower.</p>", "<title>Conclusion</title>", "<p>Subtle changes of longitudinal myocardial function begin in diastolic heart failure and are further increased in systolic heart failure. In patients with preserved LV function, a complex approach with the integration of multiple parameters including Tissue Doppler echocardiography and NT-proBNP is necessary to classify patients.</p>" ]

[ "<title>Limitations</title>", "<p>We have excluded patients with atrial fibrillation because of the difficulty to assess certain diastolic function parameters (trans-mitral E/A and myocardial E'/A'). We have only measured NT-proBNP and not BNP, because recent head-to-head studies found that BNP and NT-proBNP can be used comparably [41]. We have not classified diastolic dysfunction according to restrictive, pseudo-normal or impaired relaxation. But previous studies have shown that NT-proBNP is strongly elevated in patients with pseudo-normal and restrictive filling patterns [##REF##11827925##19##]. In the group of patients with normal LVEF, there were no patients with pseudo-normalization or restrictive diastolic dysfunction. No follow-up of the patients or invasive measurements were performed.</p>", "<title>Abbreviations</title>", "<p>NT-proBNP: N-terminal-pro-Brain Natriuretic Peptide; PPV: Positive Predictive Value; NPV: Negative Predictive Value; ROC: Receiver Operator Characteristic; LVMI: Left Ventricular Mass Index; BMI: Body Mass Index.</p>", "<title>Competing interests</title>", "<p>The authors declare that they have no competing interests.</p>", "<title>Authors' contributions</title>", "<p>FK, SE, MB and ACB have designed the study and have acquired the data. SS, VR, SE, IS, MB participated in contributions to conception and, or analysis and interpretation of data. FK has written the manuscript. GB has supervised and commented the study. ACB was the supervisor of echo examinations, is head of the echo lab, contributed by revising the manuscript critically. All authors read and approved the final manuscript.</p>", "<title>Supplementary Material</title>" ]

[ "<title>Acknowledgements</title>", "<p>The authors like to thank Christine Scholz and Sania Mikovic for excellent technical assistance.</p>" ]

[ "<fig position=\"float\" id=\"F1\"><label>Figure 1</label><caption><p><bold>Basal segment strain [%] in patients with severely and moderately reduced left ventricular function and patients with preserved systolic function and diastolic dysfunction</bold>. Right boxplots: healthy controls. White: lateral, grey: septal strain.</p></caption></fig>", "<fig position=\"float\" id=\"F2\"><label>Figure 2</label><caption><p><bold>Receiver Operating Characteristic curve to evaluate the diagnostic accuracy of NT-proBNP to separate patients with diastolic and/or systolic dysfunction (n = 95) from healthy controls (n = 42)</bold>. The area under the curve (AUC) = 0.763 (p &lt; 0.0001), Youden index = 0.44. The optimal cut-off is 97 pg/ml.</p></caption></fig>", "<fig position=\"float\" id=\"F3\"><label>Figure 3</label><caption><p>NT-proBNP [pg/ml] in patients with normal systolic function (n = 85) according to E/E' ratio.</p></caption></fig>", "<fig position=\"float\" id=\"F4\"><label>Figure 4</label><caption><p><bold>Longitudinal displacement in patients with normal left ventricular function according to E/E'</bold>. White: lateral, grey: septal longitudinal displacement.</p></caption></fig>" ]

[ "<table-wrap position=\"float\" id=\"T1\"><label>Table 1</label><caption><p>Baseline characteristics of the patients (mean ± SD, for non-parametric values: median, inter-quartile range) p compared to normal.</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td/><td align=\"left\"><bold>All patients (n= 137)</bold></td><td align=\"left\">Normal (n= 42)</td><td align=\"left\">Diastolic dysfunction (n = 43)</td><td align=\"left\">Systolic dysfunction (n = 52)</td><td align=\"left\"><bold>p (compared to normal)</bold></td></tr></thead><tbody><tr><td align=\"left\">Male sex (%)</td><td align=\"left\"><bold>88 (65)</bold></td><td align=\"left\">23 (55)</td><td align=\"left\">23 (54)</td><td align=\"left\">43 (83)</td><td align=\"left\">1.00/0.0059</td></tr><tr><td align=\"left\">Age [y]</td><td align=\"left\"><bold>53.8 (± 18.1)</bold></td><td align=\"left\">37.8 (± 15.9)</td><td align=\"left\">62.7 (± 12.5)</td><td align=\"left\">59.5 (± 15.2)</td><td align=\"left\"><bold>&lt; 0.0001/&lt; 0.0001</bold></td></tr><tr><td align=\"left\">BMI [kg/m²]</td><td align=\"left\"><bold>25.54 (± 4.36)</bold></td><td align=\"left\">23.8 (± 3.4)</td><td align=\"left\">26.5 (± 3.8)</td><td align=\"left\">26.2 (± 5.1)</td><td align=\"left\"><bold>0.011/0.018</bold></td></tr><tr><td align=\"left\">NT-proBNP [pg/ml]</td><td align=\"left\"><bold>2378 ± 6253, (median 222 (± 1220)</bold></td><td align=\"left\">275.9 ± 519.9 (median <bold>66.8</bold>, ± 185.3)</td><td align=\"left\">255.9 ± 137.4(median <bold>137</bold>, ± 256.7)</td><td align=\"left\">5832 ± 9185(median <bold>1583</bold>, ± 5109)</td><td align=\"left\">0.36/&lt;<bold>0.0001</bold></td></tr><tr><td align=\"left\">Creatinine clearance [ml/min]</td><td align=\"left\"><bold>87.48 (± 37.89)</bold></td><td align=\"left\">107.9 (± 28.6)</td><td align=\"left\">82.1 (± 35.0)</td><td align=\"left\">78.5 (± 41.3)</td><td align=\"left\"><bold>0.003/0.0002</bold></td></tr><tr><td align=\"left\">Heart rate [/s]</td><td align=\"left\"><bold>71.6 (± 13.6)</bold></td><td align=\"left\">68.3 (± 13.9)</td><td align=\"left\">73.1 (± 13.1)</td><td align=\"left\">73.5 (± 13.5)</td><td align=\"left\">0.68/0.17</td></tr><tr><td align=\"left\">Systolic RR [mmHg]</td><td align=\"left\"><bold>123.0</bold></td><td align=\"left\">123.0</td><td align=\"left\">136.8</td><td align=\"left\">115.4</td><td align=\"left\"><bold>0.01</bold>/0.19</td></tr><tr><td align=\"left\">Diastolic RR [mmHg]</td><td align=\"left\"><bold>74.0</bold></td><td align=\"left\">75.4</td><td align=\"left\">80.0</td><td align=\"left\">71.2</td><td align=\"left\">0.21/0.23</td></tr><tr><td align=\"left\">coronary artery disease (%)</td><td align=\"left\"><bold>41 (30)</bold></td><td align=\"left\">1 (2)</td><td align=\"left\">10 (23)</td><td align=\"left\">28 (54)</td><td align=\"left\"><bold>0.0077/0.001</bold></td></tr><tr><td align=\"left\">previous myocardial infarction (%)</td><td align=\"left\"><bold>28 (20)</bold></td><td align=\"left\">0</td><td align=\"left\">5 (12)</td><td align=\"left\">21 (40)</td><td align=\"left\">0.06/<bold>0.0001</bold></td></tr><tr><td align=\"left\">arterial hypertension (%)</td><td align=\"left\"><bold>63 (46)</bold></td><td align=\"left\">11 (26)</td><td align=\"left\">27 (63)</td><td align=\"left\">25 (48)</td><td align=\"left\"><bold>0.0008/0.049</bold></td></tr><tr><td align=\"left\">diabetes mellitus (%)</td><td align=\"left\"><bold>25 (18)</bold></td><td align=\"left\">3 (7)</td><td align=\"left\">3 (7)</td><td align=\"left\">18 (35)</td><td align=\"left\">1.00/<bold>0.0046</bold></td></tr><tr><td align=\"left\">Hyperlipidemia (%)</td><td align=\"left\"><bold>40 (29)</bold></td><td align=\"left\">2 (5)</td><td align=\"left\">15 (35)</td><td align=\"left\">22 (42)</td><td align=\"left\"><bold>0.0018/&lt; 0.0001</bold></td></tr><tr><td align=\"left\">Smoker (%)</td><td align=\"left\"><bold>27 (20)</bold></td><td align=\"left\">10 (24)</td><td align=\"left\">9 (21)</td><td align=\"left\">9 (17)</td><td align=\"left\">0.8035</td></tr><tr><td align=\"left\">ischemic cardiomyopathy (%)</td><td align=\"left\"><bold>30 (22)</bold></td><td align=\"left\">0</td><td align=\"left\">0</td><td align=\"left\">29 (56)</td><td align=\"left\">&lt; 0.0001</td></tr></tbody></table></table-wrap>", "<table-wrap position=\"float\" id=\"T2\"><label>Table 2</label><caption><p>Echocardiographic findings. Mean ± SD</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td/><td align=\"left\"><bold>Normal</bold></td><td align=\"left\"><bold>Diastolic dysfunction</bold></td><td align=\"left\"><bold>Systolic dysfunction</bold></td><td align=\"left\"><bold>P (compared to normal)</bold></td></tr></thead><tbody><tr><td align=\"left\">LVEF (%)</td><td align=\"left\">59.5 (± 2.2)</td><td align=\"left\">59.1 (± 1.9)</td><td align=\"left\">31.5 (± 9.9)</td><td align=\"left\">0.99/<bold>&lt; 0.0001</bold></td></tr><tr><td align=\"left\">Fractional shortening (%)</td><td align=\"left\">0.38 (± 0.1)</td><td align=\"left\">0.39 (± 0.1)</td><td align=\"left\">0.16 (± 0.1)</td><td align=\"left\">0.58/<bold>&lt; 0.0001</bold></td></tr><tr><td align=\"left\">LVEDD (mm)</td><td align=\"left\">46.1 (± 4.2)</td><td align=\"left\">46.8 (± 5.5)</td><td align=\"left\">64.5 (± 12.2)</td><td align=\"left\">0.91/<bold>&lt; 0.0001</bold></td></tr><tr><td align=\"left\">LVESD (mm)</td><td align=\"left\">28.6 (± 5.9)</td><td align=\"left\">46.4 (± 5.7)</td><td align=\"left\">52.1 (± 15.0)</td><td align=\"left\">0.99/<bold>&lt; 0.0001</bold></td></tr><tr><td align=\"left\">PAP (mmHg)</td><td align=\"left\">26.1 (± 11.4)</td><td align=\"left\">27.0 (± 5.8)</td><td align=\"left\">38.4 (± 13.2)</td><td align=\"left\">0.94/<bold>0.0015</bold></td></tr><tr><td align=\"left\">Septum (mm)</td><td align=\"left\">10.3 (± 1.9)</td><td align=\"left\">12.0 (± 2.8)</td><td align=\"left\">11.5 (± 2.0)</td><td align=\"left\">0.02/<bold>0.0007</bold></td></tr><tr><td align=\"left\">Posterior wall (mm)</td><td align=\"left\">10.2 (± 1.7)</td><td align=\"left\">11.7 (± 1.8)</td><td align=\"left\">11.7 (± 1.5)</td><td align=\"left\"><bold>0.001/&lt; 0.0001</bold></td></tr><tr><td align=\"left\">E/A transmitral</td><td align=\"left\">1.5 (± 0.5)</td><td align=\"left\">0.9 (± 0.2)</td><td align=\"left\">1.3 (± 0.8)</td><td align=\"left\"><bold>&lt; 0.0001</bold>/0.36</td></tr><tr><td align=\"left\">Left ventricular mass [mg]</td><td align=\"left\">195.4 (± 59.5)</td><td align=\"left\">236.1 (± 58.3)</td><td align=\"left\">436.7 (175.9)</td><td align=\"left\">0.19/<bold>&lt; 0.0001</bold></td></tr><tr><td align=\"left\">LVMI</td><td align=\"left\">104.1 (± 26.6)</td><td align=\"left\">126.0 (± 28.7)</td><td align=\"left\">222.6 (± 84.8)</td><td align=\"left\">0.14/<bold>&lt; 0.0001</bold></td></tr><tr><td align=\"left\">Strain septal (%)</td><td align=\"left\">-24.8 (± 8.1)</td><td align=\"left\">-18.5 (± 5.3)</td><td align=\"left\">-16.1 (± 7.0)</td><td align=\"left\"><bold>&lt; 0.0001/&lt; 0.0001</bold></td></tr><tr><td align=\"left\">Strain left lateral (%)</td><td align=\"left\">-21.9 (± 11.4)</td><td align=\"left\">-17.6 (± 6.0)</td><td align=\"left\">-14.1 (± 8.3)</td><td align=\"left\"><bold>0.04/&lt; 0.0001</bold></td></tr><tr><td align=\"left\">septal longitudinal displacement (mm)</td><td align=\"left\">12.9 (± 3.0)</td><td align=\"left\">11.8 (± 2.1)</td><td align=\"left\">6.7 (± 3.9)</td><td align=\"left\">0.19/<bold>&lt; 0.0001</bold></td></tr><tr><td align=\"left\">lateral longitudinal displacement (mm)</td><td align=\"left\">12.1 (± 3.4)</td><td align=\"left\">10.9 (± 2.9)</td><td align=\"left\">7.4 (± 4.0)</td><td align=\"left\">0.20/&lt;<bold>0.0001</bold></td></tr><tr><td align=\"left\">TVI velocity E septal (m/s)</td><td align=\"left\">0.09 (± 0.02)</td><td align=\"left\">0.05 (± 0.01)</td><td align=\"left\">0.04 (± 0.02)</td><td align=\"left\"><bold>&lt; 0.0001/&lt; 0.0001</bold></td></tr><tr><td align=\"left\">TVI velocity A septal (m/s)</td><td align=\"left\">0.06 (± 0.02)</td><td align=\"left\">0.08 (± 0.02)</td><td align=\"left\">0.06 (± 0.08)</td><td align=\"left\">0.26/0.98</td></tr><tr><td align=\"left\">TVI velocity S septal (m/s)</td><td align=\"left\">0.06 (± 0.01)</td><td align=\"left\">0.06 (± 0.01)</td><td align=\"left\">0.04 (± 0.01)</td><td align=\"left\">0.14/<bold>&lt; 0.0001</bold></td></tr><tr><td align=\"left\">E/E'</td><td align=\"left\">9.14 (± 4.62)</td><td align=\"left\">11.44 (± 3.14)</td><td align=\"left\">20.56 (± 15.08)</td><td align=\"left\">0.44/&lt;<bold>0.0001</bold></td></tr><tr><td align=\"left\">E'/A'</td><td align=\"left\">1.94 (± 1.17)</td><td align=\"left\">0.88 (± 0.60)</td><td align=\"left\">1.35 (± 0.92)</td><td align=\"left\"><bold>&lt; 0.0001/0.01</bold></td></tr></tbody></table></table-wrap>", "<table-wrap position=\"float\" id=\"T3\"><label>Table 3</label><caption><p>NT-proBNP and Tissue Doppler echocardiography variables according to reduction in systolic function, n = 137. (median ± SD, IQR: inter quartile range for NT-proBNP)</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td/><td align=\"left\"><bold>LVEF &gt; 55%</bold></td><td align=\"left\"><bold>LVEF 30–54%</bold></td><td align=\"left\"><bold>LVEF &lt; 30%</bold></td><td align=\"left\"><bold>p vs normal</bold></td></tr></thead><tbody><tr><td align=\"left\">NT-proBNP</td><td align=\"left\">97.0 (180.5)</td><td align=\"left\">587.6 (2914.9)</td><td align=\"left\">3373.0 (6057)</td><td align=\"left\"><bold>&lt; 0.001/&lt; 0.001</bold></td></tr><tr><td align=\"left\">septal longitudinal displacement [mm]</td><td align=\"left\">11.5 (± 3.2)</td><td align=\"left\">9.3 (± 4.3)</td><td align=\"left\">4.9 (± 2.0)</td><td align=\"left\"><bold>0.015/&lt; 0.001</bold></td></tr><tr><td align=\"left\">lateral longitudinal displacement [mm]</td><td align=\"left\">12.6 (± 2.6)</td><td align=\"left\">8.7 (± 3.6)</td><td align=\"left\">4.1 (± 3.3)</td><td align=\"left\"><bold>&lt; 0.001/&lt; 0.001</bold></td></tr><tr><td align=\"left\">Strain septal [%]</td><td align=\"left\">-21.1 (± 7.5)</td><td align=\"left\">-14.8 (± 6.9)</td><td align=\"left\">-15.4 (± 7.4)</td><td align=\"left\"><bold>0.001/0.002</bold></td></tr><tr><td align=\"left\">Strain lateral [%]</td><td align=\"left\">-19.1 (± 9.3)</td><td align=\"left\">-16.1 (± 9.0)</td><td align=\"left\">-10.0 (± 5.2)</td><td align=\"left\">0.130/<bold>&lt; 0.001</bold></td></tr><tr><td align=\"left\">TVI S' [m/s] septal</td><td align=\"left\">0.09 (± 0.07)</td><td align=\"left\">0.06 (± 0.03)</td><td align=\"left\">0.04 (± 0.02)</td><td align=\"left\"><bold>0.002/&lt; 0.001</bold></td></tr><tr><td align=\"left\">TVI S' [m/s] lateral</td><td align=\"left\">0.06 (± 0.03)</td><td align=\"left\">0.05 (± 0.02)</td><td align=\"left\">0.03 (± 0.02)</td><td align=\"left\"><bold>&lt; 0.001/&lt; 0.001</bold></td></tr><tr><td align=\"left\">E/E'</td><td align=\"left\">9.5 (± 4.1)</td><td align=\"left\">12.3 (± 12.0)</td><td align=\"left\">19.3 (± 16.6)</td><td align=\"left\">0.063/<bold>&lt; 0.001</bold></td></tr></tbody></table></table-wrap>" ]

[]

[]

[]

[]

[]

[ "<supplementary-material content-type=\"local-data\" id=\"S1\"><caption><title>Additional file 1</title><p>TDI image of the interventricular septum. Acquisition of a TDI image in the apical four chamber view.</p></caption></supplementary-material>", "<supplementary-material content-type=\"local-data\" id=\"S2\"><caption><title>Additional file 2</title><p>How to perform the TDI analysis. Illustration of the correct acquisition a TDI region of interest (= ROI) in the basal left-ventricular septum. The size of the ROI is adjusted to the septal diameter and then, the ROI is traced manually in each frame to avoid artifacts. This acquisition applies to Velocity, Strain Rate and Displacement.</p></caption></supplementary-material>" ]

[]

[ "<graphic xlink:href=\"1476-7120-6-45-1\"/>", "<graphic xlink:href=\"1476-7120-6-45-2\"/>", "<graphic xlink:href=\"1476-7120-6-45-3\"/>", "<graphic xlink:href=\"1476-7120-6-45-4\"/>" ]

[ "<media xlink:href=\"1476-7120-6-45-S1.wmv\" mimetype=\"video\" mime-subtype=\"x-ms-wmv\"><caption><p>Click here for file</p></caption></media>", "<media xlink:href=\"1476-7120-6-45-S2.wmv\" mimetype=\"video\" mime-subtype=\"x-ms-wmv\"><caption><p>Click here for file</p></caption></media>" ]

{ "acronym": [], "definition": [] }

[ "<title>Background</title>", "<p>Vaccination compliance will predictably become a significant concern as current schedules approach the limit of public acceptance [##REF##16581162##1##] and new vaccines become available. The development of combination vaccines is a common practice that addresses the concern of repeated visits to the clinic by reducing the total number of injections required compared with administration schedules for the monovalent vaccines. Yet, physical, chemical, and biological interactions between the components of combination vaccines must be considered to avoid detrimental effects on safety or efficacy. For example, when the <italic>Haemophilus influenzae </italic>type b (Hib) vaccine was combined with diphtheria, tetanus, and acellular pertussis vaccine, a decrease in antibody titer for the Hib vaccine was observed [##REF##12744858##2##]. Thus, there is a need to develop new approaches for delivery of multiple vaccines.</p>", "<p>We evaluated delivery of multiple vaccines intradermally (i.d.) to physically isolate each component, thus directly preventing formulation incompatibilities prior to administration. The physiological fate of vaccines administered i.d. is not known. However, vaccination by microneedles [##REF##18023942##3##] permits verification of the physical deposition into the skin while intramuscular (i.m.) injection sites are inaccessible for direct observation. Further, i.d. vaccination using microneedles is less painful [##REF##18023942##3##] than i.m. injection by conventional needles and provides an increased immune response with a lower amount of vaccine than that required by intramuscular (i.m.) methods [##UREF##0##4##,##REF##15525713##5##]. The greater efficacy resulting from i.d. vaccination may permit the administration of an increased number of vaccines compared to i.m. because a smaller volume is required for delivery.</p>", "<p>The pre-clinical phase of vaccine development traditionally focuses on a single disease of concern, often targeting a protein that is critical to pathology. Because emerging infectious diseases and agents of concern to biodefense contribute substantially to the burden of new vaccines, we specifically examined vaccines for anthrax, botulism, toxic-shock syndrome, and plague. The following is a brief description of the diseases and vaccines that were developed for prevention.</p>", "<p><italic>Bacillus anthracis</italic>, the etiological agent of anthrax, produces binary toxins [##REF##7854123##6##, ####REF##10475970##7##, ##REF##6285339##8##, ##REF##10926933##9####10926933##9##] comprised of protective antigen (PA) combined with lethal factor (LF) or edema factor (EF). The vaccine employed in our study was a recombinant form of PA (rPA) that was previously shown to protect rhesus macaques from aerosol challenge with <italic>B. anthracis </italic>spores [##REF##9682372##10##,##REF##11312020##11##]. Antibodies that neutralize PA block the transport of LF and EF to the cytosol, thereby blocking cell death induced by the toxins. Botulinum neurotoxin type A (BoNT/A) causes botulism by blocking the release of acetylcholine at the neuromuscular junction [##REF##3533452##12##]. A recombinant C fragment vaccine of botulinum neurotoxin type A [BoNT/A(H_c)] was developed that does not possess the toxic properties of the wild-type protein [##REF##16730042##13##]. In previous studies, the BoNT/A(H_c) was shown to be effective at protecting vaccinated mice against challenge with the wild-type toxin [##REF##16730042##13##]. Antibodies that prevent botulism are presumed to inhibit binding of the toxin to neurons and thereby impede entry of the toxin into the cell. Staphylococcal enterotoxin B (SEB) is a virulence factor expressed by most isolates of the common human pathogen <italic>Staphylococcus aureus </italic>[##REF##10689316##14##,##REF##8800837##15##]. Secreted SEB binds and cross-links class II molecules of the major histocompatibility complex expressed on antigen-presenting cells to the antigen receptors on T cells, leading to potent activation of the immune system. Life-threatening toxic shock syndrome may result from the rapid release of high levels of IFN-γ, IL-6, TNF-α and other cytokines in response to SEB. The recombinant SEB vaccine (STEBVax) contains three site-specific mutations that collectively alter key protein surfaces, leading to loss of receptor binding and superantigen activity [##REF##9795392##16##]. This vaccine was shown in previous studies to protect rhesus macaques from aerosol challenge with SEB [##REF##12865071##17##] and protection from toxic shock in vaccinated monkeys correlated with SEB neutralization by antibodies [##REF##12865071##17##]. We also examined an experimental plague vaccine (F1-V) consisting of a recombinant fusion protein of the bacterial antigens CaF1 and LcrV, previously shown to protect mice against plague [##REF##11027822##18##,##REF##9682370##19##]. The bubonic form of plague results from <italic>Yersinia pestis </italic>injected into the skin by the bite of infected fleas and is characterized by acute painful swelling of regional lymph nodes. Progression to septicemic or secondary pneumonic plague may also ensue. Primary pneumonic plague may also occur by transfer of bacteria through aerosols produced by coughing. Although mouse data are available [##REF##11027822##18##,##REF##9682370##19##], there are no reports that address protection of non-human primates that were vaccinated with F1-V and challenge with <italic>Y. pestis</italic>. However, we included F1-V in our study to increase the complexity of the vaccine combination and because this high-profile product is ultimately intended for human use.</p>", "<p>All of the vaccines we investigated were developed independently, using buffers and additives that were potentially incompatible if all antigens were directly mixed due to differences in pH, buffers, and stability profiles. For example, STEBVax was maintained in a glycine buffer of pH 8, while a phosphate buffer of pH 7 was used for rPA. Yet, an advantage associated with the vaccines for anthrax, botulism and staphylococcal toxic shock is that all were previously examined in studies using rhesus macaques [[##REF##9682372##10##,##REF##11312020##11##,##REF##12865071##17##], and unpublished observations], allowing us to measure survival from an otherwise lethal sepsis in the same animal disease model. Although co-formulation may ultimately be achievable for many vaccines, physical separation obviates the need for additional costly studies to re-examine safety, stability, and efficacy. We hypothesized that the physical separation of vaccines both in the syringe and at the site of administration will not adversely affect the biological activity of each component.</p>" ]

[ "<title>Methods</title>", "<title>Vaccinations</title>", "<p>The recombinant botulinum neurotoxin serotype A binding domain BoNT/A(H_c), SEB vaccine (STEBVax) and the fusion protein of F1 and V antigens (rF1-V) were prepared as previously described [##REF##9682372##10##,##REF##16730042##13##,##REF##9795392##16##,##REF##9682370##19##]. The recombinant protective antigen (rPA) was obtained from List Laboratories (Wako, TX). Each vaccine was combined with AH adjuvant (Superfos Biosector, Kvistgård, Denmark), before administration using previously optimized ratios (unpublished observations) that in all cases resulted in delivery of &lt; 1 mg of elemental aluminum per animal. Rhesus monkeys were obtained from Primate Products, Inc. (Woodside, CA) and quarantined for 30 d before study initiation. Just before vaccination, anesthetized (ketamine/acepromazine) monkeys were shaved on the deltoid/upper arm region or thigh using electric clippers, and the vaccines were administered i.d. on days 0, 28, and 56. On day 0 the vaccines were administered on the left arm, on day 28 the vaccines were administered on the right arm, and on day 56 the vaccines were administered on the left thigh. Vaccinated animals received 5 μg of the BoNT/A(H_c) vaccine, 150 μg of rF1-V, 50 μg of rPA, and 40 μg of STEBVax. Control animals received injection of AH adjuvant with no antigen. Two 100-μl i.d. injections of each vaccine were administered 2 cm apart with a stainless steel microneedle (1-mm exposed length, 76-μm inner diameter, 178-μm outer diameter) attached to a 1-ml syringe, as previously described [##REF##17012854##20##].</p>", "<title>Serology</title>", "<p>Complete blood counts with white blood cell differential counts as well as serum concentrations of IgM and IgG were determined from blood collected on days 14, 42, and 70. Before each blood draw, animals were anesthetized by injection with ketamine/acepromazine. Antigen-specific serum antibody levels were determined by ELISA. Plastic plates (96 well) were coated (1 h, 37°C) with 100 μl/well of 2 μg/ml of BoNT/A(H_c), rF1-V, rPA, or STEBVax diluted in PBS (pH 7.4) for the sample unknowns, and purified monkey IgM or IgG was serially diluted threefold for the standard curve. The plates were washed three times with PBS/0.1% Tween and blocked (1 h, 37°C) with 0.2% casein/PBS (100 μl/well), washed as above, and then were incubated (1 h, 37°C) with 100 μl of diluted serum samples. Plates were then washed and incubated (1 h, 37°C) with 100 μl/well of goat anti-monkey IgG or goat anti-monkey IgM (1:10,000 dilutions) conjugated to horseradish peroxidase, washed, and developed (30 min, 22°C) with 100 μl of TMB peroxidase substrate (KPL, Gaithersburg, MD). Absorbance was measured at 650 nM and concentrations were determined by comparison to the absorbance of the standard curve.</p>", "<title>Neutralizing antibody assays</title>", "<p>For the anthrax toxin neutralization assay, 100 ng/ml LF and 200 ng/ml of PA, both in high-glucose DMEM with 7.5% fetal bovine serum (FBS), were mixed 1:1 with dilutions of sera and incubated for 1 h (37°C) before being added to J774 cells growing on a 96-well plate (63,000 cells/well in high-glucose DMEM, 7.5% FBS). The cells were incubated at 37°C for 4 h and cell viability was determined by ATP content (Vialight HS, Cambrex, Rockland, ME). The endpoint titer was determined as the serum dilution that gave a response three times greater than background. For the SEB neutralization assay, human peripheral blood mononuclear cells were isolated by density gradient centrifugation and added to a 96-well plate (100,000 cells/well in RPMI, 5% fetal calf serum). After plating, cells were allowed to rest for 2 h at 37°C. Dilutions of the test and control sera were prepared and SEB (200 ng/ml) was added to each dilution. Serum dilutions were then incubated for 1 h. at 37°C. The treatments (50 μl/well) were added to the cells and the plates were incubated at 37°C for 60 h. Finally, 1 μCi of [³H] thymidine (Sigma, St. Louis, MO) was added to each well, the plates were incubated for 9 h at 37°C, and incorporated radioactivity was measured by liquid scintillation. The antibody titer was determined as the highest serum dilution that significantly inhibited (Student's t-test) SEB-induced proliferation of the monocytes compared to the negative control. For the BoNT/A neutralization assay, dilutions of serum from animals in the BoNT/A challenge groups were mixed with 10 LD₅₀of toxin and incubated for 1 h at room temperature. Each dilution was injected intraperitoneally (IP) into four CD-1 mice. The mice were observed for 4 days and the number of deaths in each group was recorded. The neutralizing antibody titer was determined as the reciprocal of the serum dilution that protected 50% of the mice from intoxication with BoNT/A.</p>", "<title>Aerosol challenge</title>", "<p>Animals were split into four separate challenge groups, each containing two controls and six vaccinated monkeys. Each group was challenged with one agent: BoNT/A, Ames strain spores of <italic>B. anthracis</italic>, or SEB, all obtained from USAMRIID. Before challenge, monkeys were anesthetized with ketamine/acepromazine and their breathing rate was determined by plethysmography. For groups challenged with botulinum neurotoxin A (50 LD₅₀), <italic>B. anthracis </italic>(200 LD₅₀), or SEB (25 LD₅₀), each animal was exposed to the agent for 10 min in a head-only exposure chamber. Animals were observed up to two months after challenge. On days 2, 4, and 6 postchallenge, blood was drawn and complete blood counts with white blood cell differential counts were performed on all samples and bacteremia was determined for samples from animals challenged with bacterial agents. Necropsies were performed on animals that did not survive to verify death was a result of exposure to the challenge agent.</p>", "<title>Pathology and necropsy</title>", "<p>A necropsy was performed on all animals, either as soon as death occurred from infection or intoxication or after humane euthanasia of terminally ill or moribund animals by established protocols. Samples of spleen, lymph nodes (mandibular, axillary, tracheobronchial, mesenteric), lung, trachea, mediastinum, and haired skin from the vaccine sites from each monkey were collected for histopathology. Additionally, brain tissue was collected from animals that succumbed due to infection with <italic>B. anthracis</italic>. All tissues were immersion-fixed in 10% neutral buffered formalin.</p>", "<title>Histology and immunohistochemistry</title>", "<p>Formalin-fixed tissues for histology were trimmed, processed, and embedded in paraffin according to established protocols [##UREF##1##21##]. Histology sections were cut at 5–6 μm, mounted on glass slides, and stained with hematoxylin &amp; eosin (H&amp;E). Immunohistochemical staining was performed using the Envision+ method (DAKO, Carpinteria, CA). Briefly, sections were deparaffinized in Xyless, rehydrated in graded ethanol, and endogenous peroxidase activity was quenched in a 0.3% hydrogen peroxide/methanol solution for 30 min at room temperature. Slides were washed in distilled water, placed in a Tris-EDTA Buffer (10 mM Tris Base, 1 mM EDTA Solution, 0.05% Tween 20, pH 9.0) and heated in a vegetable steamer for 30 min. Sections were incubated in the primary antibody, rabbit anti-major histocompatibility complex class II polyclonal antibody (RGU, unpublished), diluted 1:500 for 1 h at room temperature. After the primary antibody incubation, sections were washed in PBS and incubated for 30 min with Envision + System HRP (horseradish peroxidase-labeled polymer conjugated to goat anti-rabbit immunoglobulins) at room temperature. Peroxidase activity was developed with 3,3'-diaminobenzidine (DAB), counterstained with hematoxylin, dehydrated, cleared in Xyless, and coverslips were applied with Permount.</p>", "<title>Adjuvant visualization in tissues</title>", "<p>Adjuvant was localized in tissue samples by detection of aluminum. Five micrometer sections were prepared from formalin fixed, paraffin-embedded tissue blocks, deparaffinized in Xyless, and rehydrated in graded alcohols. Slides were rinsed in distilled water then pretreated in a 1% aqueous solution of hydrochloric acid for 10 min. After rinsing the slides in distilled water for 5 min, we stained them in a 0.2% alcoholic Morin solution (Sigma, Atlanta, GA) for 10 min. After staining with Morin, the sections were incubated for 2 h at 37°C with a 1:20 dilution of Texas Red phalloidin and approximately 1 μg/ml of Hoechst-33258 (Molecular Probes, Eugene Oregon) in PBS. Sections were rinsed twice in PBS and once in water before coverslips were applied with Vecta Shield mounting medium (Vector Labs, Burlingame, CA).</p>", "<title>Confocal microscopy</title>", "<p>Images were collected with a BioRad 2000 MP confocal system attached to a Nikon TE300 inverted microscope fitted with a 60× (1.20 N.A.) water-immersion objective lens. Morin fluorescence was detected with 488 nm laser excitation and a HQ515/30 emission filter. Texas Red phalloidin was imaged with 568 nm laser excitation and an E600LP emission filter. Hoechst dye was visualized with 800 nm 2-photon excitation and a HQ390/70 emission filter. Subsequent contrast enhancement of the resulting images was performed using Adobe PhotoShop software.</p>", "<title>Statistical analysis</title>", "<p>Analysis of variance was used to analyze serology data obtained at various time points after vaccine administration to determine if there were any statistical differences within or between the vaccinated and control groups. The data conformed with the assumptions of the test if plots of the residuals revealed no structure. Comparisons of antibody production and lymphocyte proliferation between vaccinated and control animals were performed using Student's t-test. The data conformed to the assumptions of the t-test if the normal probability plot was a straight line. Historical controls were used to increase the statistical power of the experiment. Uniform lethality was observed in more than 15 untreated control Rhesus exposed to the same strain and route of each agent used in the experiment. Efficacy was evaluated using Fishers exact test comparing the treated group to the control group for each agent consisting of 2 experimental controls and 15 historical controls.</p>" ]

[ "<title>Results</title>", "<title>Intradermal administration of physically separated vaccines</title>", "<p>A simple mixture of the BoNT/A(H_c), F1-V, rPA and STEBVax as currently formulated resulted in formation of a precipitation and a significant change in pH of the solution (data not shown). Because of these apparent chemical incompatibilities we were not able to examine animals vaccinated with simple mixtures of the vaccines. The vaccines BoNT/A(H_c), F1-V, rPA and STEBVax were individually administered three times, 28 d apart, by injection into the shaved dermis of the upper arm or thigh of rhesus macaques using stainless steel microneedles that were the approximate diameter of a human hair, as previously reported [##REF##11027822##18##, ####REF##9682370##19##, ##REF##17012854##20##, ##UREF##1##21####1##21##]. The subject animals received doses of each vaccine that were independently optimized [##REF##11312020##11##,##REF##16730042##13##,##REF##12865071##17##,##REF##9682370##19##] and adsorbed to aluminum hydroxide adjuvant (AH). Control animals received i.d. injections of AH alone. The pattern of vaccinations consisted of an array of 100-μl injections separated by 2 cm, keeping each vaccine isolated from adjacent administrations (Fig. ##FIG##0##1##).</p>", "<p>No visible indications of discomfort were noted in any animal after vaccination. Slight erythema was evident at sites of second or third vaccinations, suggesting a robust recall immune response. Small raised blebs appeared on the skin at each injection site (Fig. ##FIG##0##1A##) immediately after vaccine administration, and the sites were only slightly perceptible on the surface of the skin up to 2 months later (Fig ##FIG##0##1B##). Histology performed on tissue samples obtained from the delivery site showed AH localized within the dermis after administration and a granulomatous response to vaccination in both the controls and vaccinates (Fig. ##FIG##0##1C##). Numerous phagocytes and multinucleated giant cells were present in the dermis and panniculus at the injection site and the phagocytes contained abundant intracytoplasmic blue-gray granular material (Fig. ##FIG##0##1C##). Histochemical staining of the tissue with Morin, a dye that is fluorescent green upon chelation of aluminum, demonstrated positive staining of the intracytoplasmic granular material, which verified the presence of aluminum from the vaccine adjuvant (Fig. ##FIG##0##1C## inset). Immunohistochemical staining of the skin revealed that the phagocytes exhibited expression of MHC-II molecules (Fig. ##FIG##0##1D##). Examination of tissue from the axillary lymph nodes revealed phagocytes that contained a similar intracytoplasmic granular material as the skin sections (Fig. ##FIG##0##1E##). As before, staining the tissue with Morin revealed positive, fluorescent intracytoplasmic granules, verifying the material was aluminum from the vaccine adjuvant (Fig. ##FIG##0##1E## inset). These results suggest that the vaccines were transported from the dermal injection site to the draining lymph nodes.</p>", "<p>Several diagnostic parameters were monitored during the study to evaluate the safety of simultaneous administration of multiple vaccines. Vaccine administration did not significantly affect the white blood cell counts of either the controls or vaccinated animals (Fig. ##FIG##0##1E##). No abnormalities were noted in red blood cell count, platelets, hemoglobin, hematocrit, mean corpuscular volume, mean corpuscular hemoglobin, mean corpuscular hemoglobin concentration, red cell distribution width, or mean platelet volume, and no significant changes were noted in blood chemistries (data not shown). Collectively, these results suggested that i.d. administration of multiple vaccines produced no adverse reactions, as determined by these assays.</p>", "<title>Robust antibody response to individual antigens</title>", "<p>We next examined antibody responses to assess biological compatibility of the vaccines after i.d. administration. Sera were collected after each vaccination and antigen-specific antibodies were measured. All vaccines induced a significant increase in specific IgG compared to control by 14 days after the primary vaccine administration (Table ##TAB##0##1##). Further enhancement of the immune response to each vaccine was observed with each subsequent vaccination (Fig. ##FIG##1##2##). The final recorded antibody levels for BoNT/A(H_c), rPA and STEBVax were similar to previous values for animals receiving individual i.m. vaccinations [##REF##11312020##11##,##REF##16730042##13##,##REF##12865071##17##,##REF##9682370##19##] and F1-V responses were the highest. Serum levels of BoNT/A-specific antibody were lowest compared to all other antibodies except controls, likely as a result of the small amount of BoNT/A(H_c) used for vaccinations. Levels of antigen-specific IgM against all antigens were significantly elevated compared to controls 2 weeks after the final vaccine administrations (Table ##TAB##0##1##). We concluded that levels of serum antibodies against each vaccine were not altered by concurrent i.d. injection to sites that were in close proximity to each other.</p>", "<title>Neutralizing antibody responses</title>", "<p>Standard assays were previously established for determining the level of antibodies present in sera that protect the vaccinated host from SEB-toxic shock, botulism, and anthrax. These neutralizing antibody assays provided an additional parameter for predicting the outcome of exposure to each agent of disease. The BoNT/A neutralizing antibody titers were determined as the reciprocal of the serum dilution that protected 50% of the mice from challenge with 10 LD₅₀of toxin. Serum from vaccinated primates protected CD-1 mice challenged with BoNT/A (Fig. ##FIG##2##3A##); serum from control animals was not protective. Antibodies that neutralized <italic>B. anthracis </italic>were present in all vaccinated animals, but not in controls, as determined by measuring inhibition of J774 cell lysis after exposure to anthrax lethal toxin (Fig. ##FIG##2##3B##). Additionally, serum from vaccinated animals prevented SEB-induced proliferation of human peripheral blood mononuclear cells after addition of the toxin to culture (Fig. ##FIG##2##3C##). We could not determine the titers of neutralizing antibody against plague because there were no previously validated assays available for the rhesus monkey that permitted correlation of antibody titer with protection from disease.</p>", "<title>Protection from multiple bacterial and toxin-mediated diseases</title>", "<p>The results up to this point demonstrated robust antibody responses to all vaccines and these titers were similar or identical to previous studies using monovalent i.m. vaccinations [##REF##11312020##11##,##REF##16730042##13##,##REF##12865071##17##,##REF##9682370##19##]. Therefore, we next evaluated protection of vaccinated animals from disease. The rhesus macaques were healthy with no overt signs of disease or pathology before challenge. The total white blood cell counts and distribution of granulocytes, monocytes, and lymphocytes remained within normal range throughout the study for all vaccinated and control animals prior to disease challenge, indicating minimal systemic inflammatory responses to the multiple vaccines or method of administration (Fig. ##FIG##3##4A–C##). These data were in accordance with the general blood chemistry profiles (described above). This cellular data was collected to follow any potential toxicity resulting from the experimental method and to address the outcome of vaccinations on the inflammatory response occurring during the early stage of disease onset. The animals were divided into four separate challenge groups consisting of two controls and six vaccinated rhesus macaques. Each group was challenged by aerosol with either BoNT/A, SEB, or B. anthracis (Ames) spores and monitored for up to 2 months post-challenge. All disease challenges occurred one month after the final vaccination. Slight to moderate fluctuations in the distribution of white cell populations were noted for all animals within the first 48 h following challenge with toxin or bacteria (Fig. ##FIG##3##4##), perhaps due to a generalized inflammatory response to aerosol challenge. Efficacy was evaluated by comparing the treated group to the control group for each agent consisting of the 2 experimental controls and 15 historical controls. Uniform lethality has been observed in more than 15 untreated control rhesus exposed to the same strain and route of each agent used in the experiment (unpublished observations). Results indicated that the percentage of animals surviving in each treatment group (6/6 or 100%) was significantly higher than the percentage of animals surviving in each pooled control group (0/17 or 0%), p &lt; 0.0001. Further details concerning each disease challenge are described below.</p>", "<p>All vaccinated animals receiving BoNT/A (65 × LD₅₀average) survived (Table ##TAB##1##2##) and exhibited no outward clinical signs of botulism. Both control animals survived for only 2 days after challenge and necropsy findings were suggestive of death due to BoNT/A intoxication, although no specific post-mortem lesions are induced by BoNT/A. These findings included aspiration of foodstuff into the trachea and lungs due to dysphagia secondary to cranial nerve paralysis after exposure to the toxin. White blood cell counts of the vaccinated animals were only slightly affected by challenge. However, the average percentage of lymphocytes and monocytes increased, while granulocytes decreased until about 4 days post-challenge (Fig. ##FIG##3##4A##). Each cell population returned to normal pre-challenge levels by day 55 post-challenge.</p>", "<p>All of the vaccinated animals survived challenge with SEB (23 × LD₅₀average), showing no clinical signs of toxic shock after challenge (Table ##TAB##1##2##). In contrast, control animals survived for only 2 days after challenge. Necropsy and histopathology verified that death of the controls was consistent with toxic shock caused by SEB. Total white blood cells of the vaccinated animals did not significantly change after challenge. Similar to profiles of vaccinated animals surviving botulism, the percentage of lymphocytes and monocytes increased while the percentage of granulocytes decreased until about day 4 (Fig. ##FIG##3##4B##). The percentage of each cell type then returned to prechallenge levels by day 55 postchallenge.</p>", "<p>Control animals exposed to <italic>B. anthracis </italic>spores (377 × LD₅₀) survived 4 days after challenge and death corresponded with an increase in bacteremia detectable by day 4. The control animals exhibited increased blood monocytes (2 d) and granulocytes (4 d), while lymphocytes decreased by 4 days after challenge. Necropsy and histopathology verified that death was consistent with anthrax. All spore-challenged animals that were vaccinated survived with no disease symptoms (Table ##TAB##1##2##), and no significant changes in granulocytes, lymphocytes, or monocytes were observed (Fig. ##FIG##3##4C##).</p>" ]

[ "<title>Discussion</title>", "<p>Our data demonstrates that i.d. vaccination of multiple antigens by a method that physically separates each component circumvents the primary physical, chemical, and biological incompatibilities that are common to combination vaccines prepared by mixing before administration. Our results with four unique diseases suggested that we did not reach a biological limit to the number of vaccines that can be administered at one time and that there was no apparent \"vaccine overload\" [##REF##16581162##1##]. Any injection site trauma appeared to be minor due to the minute size of the needles used, consistent with a previous clinical study [##REF##18023942##3##]. We observed small blebs on the skin of rhesus macaques immediately after vaccination, resulting from the fluid injected, while these sites were barely perceptible by the end of the study and surrounding tissues returned to normal by 3 months. All of the vaccines we examined induced significant levels of serum antibodies (IgM, IgG), equivalent to historic data and neutralizing antibody titers were observed for anthrax, BoNT/A, and toxic shock vaccines. All vaccinated rhesus macaques were protected from an otherwise lethal anthrax, botulism and staphylococcal toxic shock. Our results indicated that the percentage of animals surviving in each treatment group (6/6 or 100%) was significantly higher than the percentage of animals surviving in each pooled control group (0/17 or 0%), p &lt; 0.0001. Collectively, these results indicate that the vaccines were biocompatible by i.d. administration and physical separation. Seroconversion also occurred after the primary dose for each vaccine, though it is not clear if this was dependent on the method of delivery. The rF1-V vaccine was previously shown to be protective against plague in mice [##REF##11027822##18##,##REF##9682370##19##] and this was confirmed with the vaccine used in our study (data not shown). Yet, there is a paucity of published data for efficacy of vaccines based on the LcrV and CaF1 antigens in non-human primates. Antibody levels specific for rF1-V were the highest among all of the vaccinated animals, suggesting that the potency of this vaccine was maintained. Cellular immunity, not addressed in our study, may also be important for protection from plague [##REF##17109349##22##]. We observed that the minor perturbations of blood cell counts occurring within days of challenge returned to normal for all vaccinated animals.</p>", "<p>Notably, the significance of our results should be considered in light of the general benefits of vaccination to society. For example, there are substantial cost savings to the individual and to the public resulting from protection against the 11 diseases preventable by the current routine childhood vaccination schedule [##REF##16330737##23##]. However, there are currently 28 recommended vaccines for children and adults, plus annual influenza vaccinations. Additional vaccines are planned for protection from the nine category A and numerous B-C agents on the Centers for Disease Control and Prevention (CDC) select agent list. Therefore, developing a reasonable vaccination schedule that assures patient compliance is a significant public health objective. Combination vaccines offer one solution, yet these are often difficult and costly to develop due to product incompatibilities that may not be apparent during development of individual component antigens.</p>", "<p>Previous studies demonstrated that vaccine efficacy was improved by targeting the dermis of the skin for delivery [##UREF##0##4##,##REF##15525713##5##,##REF##17012854##20##,##REF##15609239##24##, ####REF##17030580##25##, ##UREF##2##26####2##26##], resulting in dose sparing by a mechanism that is not clearly established. In our study, immune responses to vaccines administered i.d. were not isolated to the skin, though an enhancement of regional tissue immunity may also have been possible. We observed that the vaccines were internalized by dermal antigen-presenting cells and transported to the draining axillary lymph nodes. It is unclear if physiological transport of the vaccines delivered i.d. differs substantially from i.m. vaccination. Regardless of the mechanism, it should also be possible to increase the total number of vaccines that can be administered to a small dermal site by lowering the delivery volume for individual components because reduced amounts of antigen are required for i.d. vaccination.</p>" ]

[ "<title>Conclusion</title>", "<p>The physical separation of vaccines both in the syringe and at the site of administration did not adversely affect the biological activity of any component vaccine. Further, the vaccination method we describe may be scalable to include a greater number of antigens, while avoiding the physical and chemical incompatibilities encountered by combining multiple vaccines together in one product. Our results demonstrate that intradermal delivery of multiple vaccine preparations may provide a practical alternative to traditional combination vaccines and complicated administration schedules.</p>" ]

[ "<p>This is an Open Access article distributed under the terms of the Creative Commons Attribution License (<ext-link ext-link-type=\"uri\" xlink:href=\"http://creativecommons.org/licenses/by/2.0\"/>), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.</p>", "<title>Background</title>", "<p>Combination vaccines reduce the total number of injections required for each component administered separately and generally provide the same level of disease protection. Yet, physical, chemical, and biological interactions between vaccine components are often detrimental to vaccine safety or efficacy.</p>", "<title>Methods</title>", "<p>As a possible alternative to combination vaccines, we used specially designed microneedles to inject rhesus macaques with four separate recombinant protein vaccines for anthrax, botulism, plague and staphylococcal toxic shock next to each other just below the surface of the skin, thus avoiding potentially incompatible vaccine mixtures.</p>", "<title>Results</title>", "<p>The intradermally-administered vaccines retained potent antibody responses and were well- tolerated by rhesus macaques. Based on tracking of the adjuvant, the vaccines were transported from the dermis to draining lymph nodes by antigen-presenting cells. Vaccinated primates were completely protected from an otherwise lethal aerosol challenge by <italic>Bacillus anthracis </italic>spores, botulinum neurotoxin A, or staphylococcal enterotoxin B.</p>", "<title>Conclusion</title>", "<p>Our results demonstrated that the physical separation of vaccines both in the syringe and at the site of administration did not adversely affect the biological activity of each component.</p>", "<p>The vaccination method we describe may be scalable to include a greater number of antigens, while avoiding the physical and chemical incompatibilities encountered by combining multiple vaccines together in one product.</p>" ]

[ "<title>Abbreviations</title>", "<p>AH: aluminum hydroxide adjuvant; BoNT/A: botulinum neurotoxin type A; BoNT/A(H_c): recombinant botulinum neurotoxin type A heavy chain; i.d.: intradermal; rF1-V: recombinant fusion protein of the F1 and V antigens; rPA: recombinant protective antigen; STEBVax: recombinant staphylococcal enterotoxin B vaccine; SEB: staphylococcal enterotoxin B</p>", "<title>Competing interests</title>", "<p>Jason B. Alarcon and John A. Mikszta are employed by Becton Dickinson Technologies, the manufacturer of the micro-needle device used in this study. All other authors declare no potential conflicts of interest</p>", "<title>Authors' contributions</title>", "<p>GLM participated in the design of the study, performed the vaccinations, analyzed data and drafted the manuscript. RFT performed the botulism studies and analyzed the data. BKP performed bacterial challenge studies and analyzed the data. PLW participated in the design of the study and analyzed data from the bacterial challenges. JC carried out the necropsy and histology studies of all animals. LSM contributed the botulinum toxin vaccine and analyzed data from the botulism study. JBA performed the vaccinations and analyzed data. JAM participated in the design of the study, developed the vaccination device and analyzed data. RGU conceived of the study, participated in its design and coordination, and drafted the manuscript.</p>" ]

[ "<title>Acknowledgements</title>", "<p>The authors acknowledge Vicki Pearson, NIAID, for supplying F1-V vaccine, Ms. Gale Krietz and Mr. Neil Davis for histology preparations, Ms. Christine Mech for immunohistochemical and histochemical preparations, and Dr. Gordon Ruthel for confocal imaging and histochemical preparations. This research was conducted in compliance with the Animal Welfare Act and other federal statutes and regulations relating to animals and experiments involving animals and adhered to the principles stated in the <italic>Guide for the Care and Use of Laboratory Animals</italic>, National Research Council, 1996. USAMRIID is fully accredited by the Association for Assessment and Accreditation of Laboratory Animal Care International. Garry L. Morefield was an associate of the National Research Council at the USAMRIID. The views in this paper are those of the authors and do not purport to reflect official policy of the U.S. Government. Support was provided by funding from the Joint Science and Technology Office C.2X00104RDB (RGU) and DAMD17-03-2-0037 (JAM).</p>" ]

[ "<fig position=\"float\" id=\"F1\"><label>Figure 1</label><caption><p><bold>Intradermal administration of the vaccines for anthrax (rPA), botulism [BoNT/A(H_c)], plague (rF1-V), and SEB induced toxic-shock (STEBVax)</bold>. A. Rhesus macaque skin immediately after vaccination (two sites, left to right): BoNT/A, rF1-V, rPA, and STEBVax. B. Rhesus macaque skin two months after vaccine administration. Marks are adjacent to injection sites. C. Skin sections (H&amp;E stain) obtained from the vaccine delivery site exhibited epithelioid macrophages and multinucleated giant cells containing adjuvant (inset, green). Phalloidin staining of actin, red; Hoechst staining of DNA, blue. D. Macrophages at the vaccine delivery site exhibited high expression of MHC-II molecules (brown). Anti-MHC Class II immunohistochemistry (brown). E. Epithelioid macrophages (H&amp;E stain) containing adjuvant (inset) were also present in the axillary lymph nodes of vaccinated animals. F. Vaccination did not significantly alter white blood cell counts of vaccinated animals (solid line) compared to control (dashed line). Mean cell counts ± SD of all animals studied.</p></caption></fig>", "<fig position=\"float\" id=\"F2\"><label>Figure 2</label><caption><p><bold>Concurrent intradermal administration of four independent vaccines resulted in rapid seroconversion of specific IgG</bold>. Mean ± SD (triplicate determinations) of antigen-specific IgG for all vaccinated animals. ■ BoNT/A(H_c) vaccine, □ rF1-V vaccine, △ STEBVax, ▲ rPA vaccine. The arrows indicate the days of vaccine administration.</p></caption></fig>", "<fig position=\"float\" id=\"F3\"><label>Figure 3</label><caption><p><bold>Potent neutralizing antibody responses of rhesus macaques receiving concurrent intradermal administrations of four independent vaccines</bold>. A. Neutralizing antibody titers for animals in: A. botulinum neurotoxin type A challenge group. B. anthrax challenge group. C. SEB challenge group. Individual animals vaccinated with antigens plus AH, Vaccinated 1–6; injected with AH only, Control 1–2. All disease challenges occurred one month after the final vaccination. Geometric mean titers, based on triplicate determinations.</p></caption></fig>", "<fig position=\"float\" id=\"F4\"><label>Figure 4</label><caption><p><bold>Vaccination resulted in rapid recovery of white blood cell populations following disease challenge</bold>. All disease challenges occurred one month after the final vaccination. Peripheral arterial blood was drawn at various time points postchallenge and analyzed for changes in cellular composition. A. Botulinum neurotoxin type A; B. Staphylococcal enterotoxin B. C. <italic>B. anthracis </italic>(Ames) spores.</p></caption></fig>" ]

[ "<table-wrap position=\"float\" id=\"T1\"><label>Table 1</label><caption><p>Robust serum antibody response to simultaneous intradermal vaccination</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td/><td/><td/><td/><td align=\"center\" colspan=\"4\">Antibody concentration (μg/ml) mean ± SD</td></tr><tr><td/><td/><td/><td/><td colspan=\"4\"><hr/></td></tr><tr><td/><td/><td/><td/><td align=\"center\" colspan=\"4\">Vaccine</td></tr><tr><td/><td/><td/><td/><td colspan=\"4\"><hr/></td></tr><tr><td align=\"center\">Isotype</td><td align=\"center\">Day</td><td align=\"center\">Treatment</td><td/><td align=\"center\">BoNT/A(H _c)</td><td align=\"center\">rF1-V</td><td align=\"center\">rPA</td><td align=\"center\">STEBVax</td></tr></thead><tbody><tr><td align=\"center\">IgM</td><td align=\"center\">70</td><td align=\"center\">Control</td><td align=\"center\">(n = 8)</td><td align=\"center\">3.07+/-0.87</td><td align=\"center\">2.99+/-1.47</td><td align=\"center\">6.31+/-3.16</td><td align=\"center\">4.76+/-3.62</td></tr><tr><td/><td align=\"center\">70</td><td align=\"center\">Vaccinated</td><td align=\"center\">(n = 24)</td><td align=\"center\">5.47+/-2.20</td><td align=\"center\">11.2+/-4.04</td><td align=\"center\">13.7+/-9.28</td><td align=\"center\">9.07+/-2.74</td></tr><tr><td/><td/><td align=\"center\"><bold>p-value*</bold></td><td/><td align=\"center\"><bold>0.0001</bold></td><td align=\"center\"><bold>&lt; 0.0001</bold></td><td align=\"center\"><bold>0.002</bold></td><td align=\"center\"><bold>0.012</bold></td></tr><tr><td colspan=\"8\"><hr/></td></tr><tr><td align=\"center\">IgG</td><td align=\"center\">14</td><td align=\"center\">Control</td><td align=\"center\">(n = 8)</td><td align=\"center\">0.31+/-0.15</td><td align=\"center\">2.1+/-3.1</td><td align=\"center\">0.31+/-0.12</td><td align=\"center\">1.25+/-1.76</td></tr><tr><td/><td align=\"center\">14</td><td align=\"center\">Vaccinated</td><td align=\"center\">(n = 24)</td><td align=\"center\">1.4+/-1.1</td><td align=\"center\">421+/-196</td><td align=\"center\">86+/-46</td><td align=\"center\">121+/-109</td></tr><tr><td/><td/><td align=\"center\"><bold>p-value</bold></td><td/><td align=\"center\"><bold>&lt; 0.0001</bold></td><td align=\"center\"><bold>&lt; 0.0001</bold></td><td align=\"center\"><bold>&lt; 0.0001</bold></td><td align=\"center\"><bold>&lt; 0.0001</bold></td></tr><tr><td/><td colspan=\"7\"><hr/></td></tr><tr><td/><td align=\"center\">42</td><td align=\"center\">Control</td><td align=\"center\">(n = 8)</td><td align=\"center\">0.28+/-0.22</td><td align=\"center\">1.95+/-0.98</td><td align=\"center\">2.2+/-1.4</td><td align=\"center\">1.23+/-0.91</td></tr><tr><td/><td align=\"center\">42</td><td align=\"center\">Vaccinated</td><td align=\"center\">(n = 24)</td><td align=\"center\">4+/-2.1</td><td align=\"center\">767+/-382</td><td align=\"center\">689+/-397</td><td align=\"center\">323+/-187</td></tr><tr><td/><td/><td align=\"center\"><bold>p-value</bold></td><td/><td align=\"center\"><bold>&lt; 0.0001</bold></td><td align=\"center\"><bold>&lt; 0.0001</bold></td><td align=\"center\"><bold>&lt; 0.0001</bold></td><td align=\"center\"><bold>&lt; 0.0001</bold></td></tr><tr><td/><td colspan=\"7\"><hr/></td></tr><tr><td/><td align=\"center\">70</td><td align=\"center\">Control</td><td align=\"center\">(n = 8)</td><td align=\"center\">0.65+/-0.37</td><td align=\"center\">1.05+/-1.08</td><td align=\"center\">0.91+/-0.44</td><td align=\"center\">1.93+/-1.25</td></tr><tr><td/><td align=\"center\">70</td><td align=\"center\">Vaccinated</td><td align=\"center\">(n = 24)</td><td align=\"center\">48+/-13</td><td align=\"center\">2331+/-303</td><td align=\"center\">2245+/-1224</td><td align=\"center\">1340+/-215</td></tr><tr><td/><td/><td align=\"center\"><bold>p-value</bold></td><td/><td align=\"center\"><bold>&lt; 0.0001</bold></td><td align=\"center\"><bold>&lt; 0.0001</bold></td><td align=\"center\"><bold>&lt; 0.0001</bold></td><td align=\"center\"><bold>&lt; 0.0001</bold></td></tr></tbody></table></table-wrap>", "<table-wrap position=\"float\" id=\"T2\"><label>Table 2</label><caption><p>Simultaneous intradermal vaccination with four independent vaccines protected Rhesus macaques from fatal infectious or toxin-mediated disease</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td/><td align=\"center\" colspan=\"2\">Bot/A Challenge*</td><td align=\"center\" colspan=\"2\">Spore Challenge</td><td align=\"center\" colspan=\"2\">SEB Challenge</td></tr><tr><td/><td colspan=\"2\"><hr/></td><td colspan=\"2\"><hr/></td><td colspan=\"2\"><hr/></td></tr><tr><td/><td align=\"center\">Dose (LD50s)</td><td align=\"center\">Survival**</td><td align=\"center\">Dose (LD50s)</td><td align=\"center\">Survival</td><td align=\"center\">Dose (LD50s)</td><td align=\"center\">Survival</td></tr></thead><tbody><tr><td align=\"right\">Control 1</td><td align=\"center\">57</td><td align=\"center\">-</td><td align=\"center\">507</td><td align=\"center\">-</td><td align=\"center\">33.5</td><td align=\"center\">-</td></tr><tr><td align=\"right\">Control 2</td><td align=\"center\">100</td><td align=\"center\">-</td><td align=\"center\">412</td><td align=\"center\">-</td><td align=\"center\">18.0</td><td align=\"center\">-</td></tr><tr><td colspan=\"7\"><hr/></td></tr><tr><td align=\"right\">Vaccinated 1</td><td align=\"center\">50</td><td align=\"center\">+</td><td align=\"center\">257</td><td align=\"center\">+</td><td align=\"center\">26.4</td><td align=\"center\">+</td></tr><tr><td align=\"right\">Vaccinated 2</td><td align=\"center\">24</td><td align=\"center\">+</td><td align=\"center\">487</td><td align=\"center\">+</td><td align=\"center\">25.6</td><td align=\"center\">+</td></tr><tr><td align=\"right\">Vaccinated 3</td><td align=\"center\">99</td><td align=\"center\">+</td><td align=\"center\">439</td><td align=\"center\">+</td><td align=\"center\">15.8</td><td align=\"center\">+</td></tr><tr><td align=\"right\">Vaccinated 4</td><td align=\"center\">43</td><td align=\"center\">+</td><td align=\"center\">373</td><td align=\"center\">+</td><td align=\"center\">18.9</td><td align=\"center\">+</td></tr><tr><td align=\"right\">Vaccinated 5</td><td align=\"center\">82</td><td align=\"center\">+</td><td align=\"center\">275</td><td align=\"center\">+</td><td align=\"center\">19.6</td><td align=\"center\">+</td></tr><tr><td align=\"right\">Vaccinated 6</td><td align=\"center\">62</td><td align=\"center\">+</td><td align=\"center\">263</td><td align=\"center\">+</td><td align=\"center\">23.4</td><td align=\"center\">+</td></tr><tr><td colspan=\"7\"><hr/></td></tr><tr><td align=\"center\">Mean+/-SD</td><td align=\"center\">65+/-27</td><td/><td align=\"center\">377+/-101</td><td/><td align=\"center\">23+/-6</td><td/></tr></tbody></table></table-wrap>" ]

[]

[]

[]

[]

[]

[]

[ "<table-wrap-foot><p>*Significance of mean serum IgM and IgG concentrations for control and vaccinated animals were compared using Student's t-test.</p></table-wrap-foot>", "<table-wrap-foot><p>*All disease challenges occurred one month after the final vaccination.</p><p>**Efficacy was evaluated using Fishers exact test comparing the treated group to the control group for each agent consisting of 2 experimental controls and 15 historical controls. Results indicated that the percentage of animals surviving in each treatment group (6/6 or 100%) was significantly higher than the percentage of animals surviving in each pooled (experimental plus historical) control group (0/17 or 0%), p &lt; 0.0001.</p></table-wrap-foot>" ]

[ "<graphic xlink:href=\"1476-8518-6-5-1\"/>", "<graphic xlink:href=\"1476-8518-6-5-2\"/>", "<graphic xlink:href=\"1476-8518-6-5-3\"/>", "<graphic xlink:href=\"1476-8518-6-5-4\"/>" ]

[]

{ "acronym": [], "definition": [] }

[ "<title>Introduction</title>", "<p>Mitogenic signalling induces transcription and translation of the D-type cyclins, the allosteric regulators of CDK4/6, during G1 phase coupling growth stimuli to cell cycle progression [##REF##8513492##1##]. Active cyclin D1/CDK4 complexes translocate to the nucleus and phosphorylate the retinoblastoma protein (Rb) and related pocket proteins, thereby triggering E2F-dependent transcription of genes required for S-phase entry [##REF##7958855##2##, ####REF##10499802##3##, ##REF##12401786##4##, ##REF##11884610##5##, ##REF##8939849##6####8939849##6##]. The timely expression and accumulation of cyclin D1 is ensured through several mechanisms. Initially, <italic>cyclin D1 </italic>expression requires activation of the Raf-Mek-Erk kinase cascade [##REF##7559524##7##, ####REF##8552383##8##, ##REF##9199319##9##, ##REF##9448290##10####9448290##10##]. This increased expression is accompanied by phosphatidylinositol 3-kinase and Akt-dependent increases in cyclin D1 translation and decreased cyclin D1 protein degradation [##REF##8524413##11##, ####REF##9255069##12##, ##REF##9832503##13####9832503##13##]. Following the G1/S transition, cyclin D1 is rapidly phosphorylated by GSK3β on Thr-286, triggering CRM1-dependent nuclear export [##REF##9832503##13##]. Thr-286 phosphorylated cyclin D1 is recognized by Fbx4 and the co-factor αB crystallin and is subsequently poly-ubiquitylated and degraded by the 26S proteasome [##REF##17081987##14##].</p>", "<p>While cyclin D1 overexpression occurs frequently in human cancer and is considered a causative factor in many tumor types, simple over-expression of wild type cyclin D1 is insufficient to drive neoplastic transformation [##REF##11124803##15##]. In contrast, cyclin D1 mutants refractory to phosphorylation and subsequent cytoplasmic proteasomal degradation are acutely transforming <italic>in vitro </italic>and <italic>in vivo </italic>[##REF##11124803##15##,##REF##16247460##16##] implying that compartmentalization of cyclin D1 complexes is essential for cell homeostasis. Indeed, mutations that directly impact on cyclin D1 nuclear export and subsequent proteolysis have been identified in human tumors [##REF##12955092##17##,##REF##16732330##18##]. However, the occurrence of such mutations is rare compared to the frequency of cyclin D1 overexpression. Implicit to this data, if cyclin D1 is a driver oncogene, its overexpression in many cancers must be secondary to tumor-specific alterations that modify its subcellular location during the cell cycle. Here, we discuss recent work to address these questions.</p>", "<title>Cancer-derived cyclin D1 mutations</title>", "<p>Cyclin D1 overexpression occurs in carcinomas of the breast, esophagus, head and neck, and lung; in a majority of these cases, alterations in gene expression cannot account for its overexpression. Perturbations in cyclin D1 degradation have been suggested as the primary contributor in a large percentage of cases. Indeed mutations that interfere with Thr-286 phosphorylation occur, but are rare. Such mutations have been noted in endometrial and esophageal cancer [##REF##12955092##17##,##REF##16732330##18##]. For example, of single-base substitutions in the <italic>CCND1 </italic>gene changing proline-287 (Pro-287) to a serine or threonine residue in endometrioid endometrial carcinoma correlates with overexpression of cyclin D1 in the nucleus of neoplastic cells. Additionally, a 12-base pair in frame deletion corresponding to deletion of amino acids 289–292 was reported with overexpression of cyclin D1 [##REF##12955092##17##]. Significantly, subsequent analyses revealed that disruption of Pro-287 abrogates GSK3β-dependent phosphorylation of Thr-286, resulting in nuclear localization of cyclin D1, and deletion of residues 289–292 impairs cyclin D1 binding to CRM1, also resulting in nuclear accumulation [##REF##16732330##18##,##REF##15513923##19##].</p>", "<p>In accordance with endometrial cancer studies, recently identified cyclin D1 mutations in esophageal cancer and tumor-derived cell lines also disrupt Thr-286 phosphorylation [##REF##16732330##18##]. Sequencing of cyclin D1 (<italic>CCND1</italic>) in a panel of 90 patient esophageal carcinomas revealed mutation of Thr-286 to arginine and a deletion of C-terminal residues 266–295. Additionally, screening of human tumor-derived esophageal carcinoma cell lines also identified a Pro-287 to alanine mutation in three of these lines [##REF##16732330##18##].</p>", "<p>Alternative splicing may also contribute to cancer specific accumulation of cyclin D1 proteins that cannot undergo cytoplasmic degradation [##REF##7675441##20##]. Tumor-specific alternative splicing produces a truncated transcript lacking exon 5, the region encoding the C-terminal Thr-286; this cyclin D1 transcript b (D1b) produces a constitutively nuclear protein refractory to proteasomal degradation [##REF##14612495##21##]. Current work suggests that D1b may be expressed in up to 40% of primary esophageal carcinomas [##REF##14612495##21##]. While cyclin D1b stability is only moderately increased relative to wild type protein, it is refractory to nuclear export and therefore exhibits constitutively nuclear localization. As might be anticipated, expression of cyclin D1b promotes neoplastic transformation <italic>in vitro</italic>, much like the phosphorylation-deficient T286A mutant [##REF##11124803##15##,##REF##14612495##21##]. It is currently unclear what determines alterations in cyclin D1 splicing; polymorphisms that occur in the splice-donor site at the exon 4/intron boundary have been implicated. However, cyclin D1b has been detected in cells that do not exhibit polymorphic residues implicating the occurrence of alternative mechanisms [##REF##14612495##21##].</p>", "<title>Disruption of SCF^{Fbx4-αBcrystallin}function: novel insights into cyclin D1 overexpression</title>", "<p>Cell cycle progression is driven by alternating phases of cyclin expression and destruction. Degradation is coordinated by substrate ubiquitylation and destruction via the 26S proteasome [##REF##6134587##22##,##REF##1846030##23##]. Poly-ubiquitylation of substrate proteins is catalyzed by the sequential activity of a ubiquitin activating enzyme (E1), ubiquitin conjugating enzyme (E2), and ubiquitin ligase (E3) [##REF##7923371##24##]. The Skp1-Cul1-F box (SCF) E3 ubiquitin ligases facilitate polyubiquitylation of phosphorylated substrates; among the substrates are many of the key regulators of G1 progression [##REF##9346238##25##]. We recently identified the E3 ubiquitin ligase that directs phosphorylation-dependent polyubiquitylation of cyclin D1, SCF^{Fbx4-αB crystallin}[##REF##17081987##14##]. Substrate recognition by this ligase is analogous with Skp2/Cks1 in that it is directed by the F box protein Fbx4 in concert with a cofactor, αB crystallin [##REF##17081987##14##].</p>", "<p>Given the low frequency of mutations within cyclin D1 that directly impact is turnover, a logical prediction was the occurrence of inactivating mutations in its E3 ligase. The first clue to cyclin D1 ligase involvement came from analysis of several breast cancer cell lines exhibiting increased cyclin D1 half-life without mutations disrupting phosphorylation. These analyses revealed that MCF-7 and MDA-MB-231 cells lack αB crystallin expression as a consequence of chromosome deletions. In addition, tumor microarray analysis of esophageal carcinomas revealed a reduction in both αB crystallin and Fbx4 mRNA levels in tumor tissues [##REF##17081987##14##]. Collectively, these findings implicated the cyclin D1 ligase as a target in tumorigenesis, with proteins such as αB crystallin or Fbx4 functioning as putative tumor suppressors.</p>", "<p>Significantly, recent assessment of the <italic>Fbx4 </italic>sequence in primary esophageal carcinoma samples identified hemizygous, missense mutations in 14 percent of the tumors; no mutations occurred in α<italic>B crystallin </italic>or <italic>CCND1 </italic>genes in tumors expressing mutated Fbx4. A high percentage of mutations reside within a putative Fbx4 dimerization domain, while others target Ser-12 in the N-terminus. Additionally, one mutation identified within the F box domain results in production of a dominant negative protein incapable of recruiting Skp1 and Cul1[##REF##18598945##26##]. The residues targeted in cancer suggest a model wherein Fbx4 is phosphorylated on Ser-12 and functions as a dimer. While substrate phosphorylation serves as a critical step in regulating SCF ligase function, several studies indicated that F box proteins such as Fbw7 and β-TrCP can form dimers, potentially regulating ligase activity [##REF##17574027##27##, ####REF##17298674##28##, ##REF##17189384##29##, ##REF##10644755##30####10644755##30##].</p>", "<p>Further analysis of the cyclin D1 ligase revealed that GSK3β, the kinase responsible for cyclin D1 Thr-286 phosphorylation, also catalyzes phosphorylation of Fbx4 Ser-12. Fbx4 phosphorylation correlates with low cyclin D1 expression during G2/M and early G1, with a marked decrease during cell cycle entry due to growth factor-dependent GSK3β inhibition [##REF##8524413##11##]. Fbx4 phosphorylation increases again at the G1/S boundary as GSK3β becomes active and temporally controls ligase activation and cyclin D1 phosphorylation. These findings raise the question of how Fbx4 phosphorylation regulates ligase function. Additional work demonstrated that Fbx4 forms homodimers in a phosphorylation- and cell cycle-dependent manner [##REF##18598945##26##]. Fbx4 homodimerization is dependent on Ser-12 phosphorylation; disruption of this residue impairs the ubiquitylation activity of this ligase [##REF##18598945##26##].</p>", "<p>The tumorigenic potential of Fbx4 mutations disrupting phosphorylation and dimerization was assessed <italic>in vitro</italic>, revealing that such mutations are indeed transforming [##REF##18598945##26##]. These findings suggest that mutation of Ser-12 or residues within the Fbx4 dimerization domain impairs ligase function, contributing to cyclin D1 accumulation. Studies over the past several years have identified mutations in both cyclin D1 and its E3 ubiquitin ligase that impair proteasomal degradation and promote nuclear accumulation of cyclin D1/CDK4 complexes (summarized in Table ##TAB##0##1##). Therefore, delineating the mechanism underlying nuclear cyclin D1-driven transformation is critical for understanding the role of cyclin D1 in tumorigenesis and development of therapeutic strategies for treatment of cancers overexpressing this protein.</p>", "<title>Inhibition of cyclin D1 degradation promotes genomic instability</title>", "<p>Aberrant nuclear accumulation of cyclin D1 during S-phase drives cell transformation <italic>in vitro </italic>and B-cell lymphomagenesis in mice [##REF##11124803##15##,##REF##16247460##16##]. Recent evidence links nuclear retention of active cyclin D1/CDK4 complexes with genomic instability, providing a novel mechanism wherein cyclin D1 stabilization initiates tumor formation. Nuclear accumulation of active cyclin D1-dependent kinase was found to interfere with Cdt1 proteolysis [##REF##18006686##31##]. Cdt1 is a component of the pre-replicative complex that promotes loading of the replicative helicase during late G1 phase [##REF##10766247##32##,##REF##10766248##33##]. The failure to degrade Cdt1 during S-phase has been shown to contribute to DNA re-replication in several systems [##REF##15616577##34##,##REF##12718885##35##]. In cells harbouring nuclear cyclin D1, Cdt1 proteolysis was disrupted due to repression of Cul4A and Cul4B expression. Cul4 proteins serve as scaffolds for the E3 ligase that regulates Cdt1 [##REF##16482215##36##,##REF##12815436##37##]. Ultimately, nuclear cyclin D1/CDK4 complexes facilitate stabilization of Cdt1 during S-phase, with concurrent maintenance of the MCM helicase on chromatin, resulting in DNA re-replication and activation of DNA damage checkpoint signalling [##REF##18006686##31##].</p>", "<p>Active cyclin D1/CDK4 complexes influence transcriptional regulation of the Cul4 proteins; however, the precise mechanism of regulation remains to be elucidated. Strikingly, impaired cyclin D1 ligase function results in nuclear accumulation of active cyclin D1/CDK4 complexes, accumulation of Cdt1, and triggers cellular transformation analogous to cyclin D1 T286A [##REF##18598945##26##]. Taken together, data elucidating the role of nuclear cyclin D1 in neoplastic transformation support a model wherein disruption of cyclin D1 phosphorylation or SCF^{Fbx4-αBcrystallin}activation generate genomic instability, ultimately driving tumor formation (Figure ##FIG##0##1##).</p>", "<p>Genomic integrity is monitored by cell cycle checkpoints that promote cell cycle arrest or apoptosis upon detection of damaged DNA [##REF##15549093##38##]. Chronic activation of checkpoints may provide selective pressure for deletion or mutation of critical tumor suppressors such as p53 [##REF##15829956##39##,##REF##15829965##40##]. Consistent with this notion, cyclin D1T286A-dependent tumorigenesis is accompanied by activation of the DNA damage checkpoint pathway and loss of p53 [##REF##18006686##31##]. The deleterious effects of nuclear cyclin D1/CDK4 complexes could then have two different effects on cellular transformation. Chronic checkpoint activation can induce selective pressure for loss of tumor suppressors, thereby providing such cells not only with a growth advantage but also the propensity for additional genomic instability.</p>", "<title>Cyclin D1 protein accumulation in tumors: potential therapeutic strategies</title>", "<p>Provided with this new data suggesting that cyclin D1-dependent kinase contributes to neoplasia at least in part through perturbations in DNA replication and loss of genomic integrity, can we utilize this information for increased therapeutic modalities? One scenario might be to take advantage of the fact that tumors harbouring mutations in cyclin D1 or Fbx4 have a compromised DNA damage checkpoint due to loss of p53. In theory, treatment of normal cells with chemotherapeutic agents that generate DNA crosslinks should trigger an intra S-phase checkpoint, thereby providing an opportunity for repair. Expression of stabilized cyclin D1 should promote maintenance of Cdt1 and MCM complexes and in so doing promote continued origin firing without allowing for repair of damaged DNA, ultimately resulting in mitotic catastrophe. Further work is required to investigate how alterations in cyclin D1 proteolysis might influence cellular responses to DNA damaging therapeutics.</p>", "<p>The alternative is the development of drugs that directly target the kinase subunits that cyclin D1 regulates. If continued activation of CDK4/6 is required for tumor proliferation and survival, such drugs may have significant clinical use. Consistent with this notion of oncogene addiction, treatment of mammary epithelial cells derived from murine tumors harbouring a MMTV-T286A transgene with the CDK4/6 inhibitor PD0332991 [##REF##15542782##41##] triggered G1 arrest [##REF##17724472##42##]. Therefore, CDK4/6 activity is a potential target to prevent cyclin D1-driven proliferation.</p>", "<title>Concluding Remarks</title>", "<p>Phosphorylation-dependent nuclear export and subsequent degradation of cyclin D1 is essential to maintain cellular homeostasis. Disruption of this regulatory pathway has been extensively shown to promote neoplastic transformation; however, the precise mechanism of this event has been elusive. Significantly, recent work revealed that nuclear accumulation of active cyclin D1/CDK4 complexes generates genomic instability through a mechanism of Cdt1 stabilization and DNA re-replication. Furthermore, mutations targeting SCF^{Fbx4-αB crystallin}in human cancers implicate ligase function in cyclin D1 overexpression and subsequent nuclear accumulation of active cyclin D1/CDK4 complexes. Importantly, GSK3β functions as the master switch, turning on cyclin D1 destruction at the G1/S transition by regulating both ligase activation and cyclin D1 phosphorylation. Given the phenotypic outcome of accumulated nuclear cyclin D1/CDK4 complexes, further mechanistic investigation is required for development of novel therapeutic strategies to promote tumor cell death in cancers overexpressing cyclin D1.</p>" ]

[]

[]

[]

[]

[ "<p>This is an Open Access article distributed under the terms of the Creative Commons Attribution License (<ext-link ext-link-type=\"uri\" xlink:href=\"http://creativecommons.org/licenses/by/2.0\"/>), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.</p>", "<p>Mitogenic induction of cyclin D1, the allosteric regulator of CDK4/6, is a key regulatory event contributing to G1 phase progression. Following the G1/S transition, cyclin D1 activation is antagonized by GSK3β-dependent threonine-286 (Thr-286) phosphorylation, triggering nuclear export and subsequent cytoplasmic degradation mediated by the SCF^{Fbx4-αBcrystallin}E3 ubiquitin ligase. Although cyclin D1 overexpression occurs in numerous malignancies, overexpression of cyclin D1 alone is insufficient to drive transformation. In contrast, cyclin D1 mutants refractory to phosphorylation-dependent nuclear export and degradation are acutely transforming. This raises the question of whether overexpression of cyclin D1 is a significant contributor to tumorigenesis or an effect of neoplastic transformation. Significantly, recent work strongly supports a model wherein nuclear accumulation of cyclin D1-dependent kinase during S-phase is a critical event with regard to transformation. The identification of mutations within SCF^{Fbx4-αBcrystallin}ligase in primary tumors provides mechanistic insight into cyclin D1 accumulation in human cancer. Furthermore, analysis of mouse models expressing cyclin D1 mutants refractory to degradation indicate that nuclear cyclin D1/CDK4 kinase triggers DNA re-replication and genomic instability. Collectively, these new findings provide a mechanism whereby aberrations in post-translational regulation of cyclin D1 establish a cellular environment conducive to mutations that favor neoplastic growth.</p>" ]

[ "<title>Competing interests</title>", "<p>The authors declare that they have no competing interests.</p>", "<title>Authors' contributions</title>", "<p>LLP and JAD contributed to the discussion and preparation of this manuscript.</p>" ]

[ "<title>Acknowledgements</title>", "<p>This work was supported by CA93237 (NIH); JAD is a Leukemia &amp; Lymphoma Society Scholar.</p>" ]

[ "<fig position=\"float\" id=\"F1\"><label>Figure 1</label><caption><p><bold>Cyclin D1 Regulatory Pathways are Targeted in Human Cancer</bold>. Cyclin D1 protein accumulation is tightly controlled via phosphorylation-dependent proteolysis. Mutations targeting cyclin D1 phosphorylation or degradation contribute to neoplastic transformation. Specific disruption of Thr-286 phosphorylation occurs in endometrial and esophageal carcinoma, while mutations preventing Crm1 binding occur in endometrial cancer. Mutations targeting Fbx4 have been identified in esophageal cancer, and αB crystallin loss occurs in tumor-derived breast carcioma cell lines. Disruption of cyclin D1 proteolysis promotes accumulation of active cyclin D1/CDK4 kinase, triggering DNA re-replication and subsequent genomic instability necessary to drive neoplastic transformation.</p></caption></fig>" ]

[ "<table-wrap position=\"float\" id=\"T1\"><label>Table 1</label><caption><p>Summary of mutations targeting cyclin D1 phosphorylation or ligase function</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td align=\"left\">Target</td><td align=\"left\">Mutation</td><td align=\"left\">Consequence</td><td align=\"left\">Tumor Type</td><td align=\"center\">Reference</td></tr></thead><tbody><tr><td align=\"left\">Cyclin D1</td><td align=\"left\">T286R</td><td align=\"left\">Constitutively Nuclear</td><td align=\"left\">Esophageal</td><td align=\"center\">[##REF##16732330##18##]</td></tr><tr><td align=\"left\">Cyclin D1</td><td align=\"left\">Δ266–295</td><td align=\"left\">Constitutively Nuclear</td><td align=\"left\">Esophageal</td><td align=\"center\">[##REF##16732330##18##]</td></tr><tr><td align=\"left\">Cyclin D1</td><td align=\"left\">P287A</td><td align=\"left\">Constitutively Nuclear</td><td align=\"left\">Tumor-derived esophageal carcinoma cell lines TE3, TE7, and TE12</td><td align=\"center\">[##REF##16732330##18##]</td></tr><tr><td align=\"left\">Cyclin D1</td><td align=\"left\">P287S/T</td><td align=\"left\">Constitutively Nuclear</td><td align=\"left\">Endometrial</td><td align=\"center\">[##REF##12955092##17##]</td></tr><tr><td align=\"left\">Cyclin D1</td><td align=\"left\">Δ289–292</td><td align=\"left\">Constitutively Nuclear</td><td align=\"left\">Endometrial</td><td align=\"center\">[##REF##12955092##17##]</td></tr><tr><td align=\"left\">αB crystallin</td><td align=\"left\">Chromosome 11 deletion</td><td align=\"left\">Impaired ligase activity</td><td align=\"left\">Tumor-derived breast cancer cell lines (MCF-7, MDA-MB 231)</td><td align=\"center\">[##REF##17081987##14##]</td></tr><tr><td align=\"left\">Fbx4</td><td align=\"left\">S8R</td><td align=\"left\">Impaired ligase activity</td><td align=\"left\">Esophageal</td><td align=\"center\">[##REF##18598945##26##]</td></tr><tr><td align=\"left\">Fbx4</td><td align=\"left\">S12L</td><td align=\"left\">Disrupts phosphorylation</td><td align=\"left\">Esophageal</td><td align=\"center\">[##REF##18598945##26##]</td></tr><tr><td align=\"left\">Fbx4</td><td align=\"left\">P13S</td><td align=\"left\">Disrupts phosphorylation</td><td align=\"left\">Esophageal</td><td align=\"center\">[##REF##18598945##26##]</td></tr><tr><td align=\"left\">Fbx4</td><td align=\"left\">L23Q</td><td align=\"left\">Dimerization-deficient</td><td align=\"left\">Esophageal</td><td align=\"center\">[##REF##18598945##26##]</td></tr><tr><td align=\"left\">Fbx4</td><td align=\"left\">P76T</td><td align=\"left\">Impaired Skp1 binding</td><td align=\"left\">Esophageal</td><td align=\"center\">[##REF##18598945##26##]</td></tr></tbody></table></table-wrap>" ]

[]

[]

[]

[]

[]

[]

[ "<table-wrap-foot><p>Mutations disrupting GSK3β-dependent cyclin D1 phosphorylation and nuclear export include mutation of Thr-286, Pro-287, and deletion of residues corresponding to the CRM1 binding site. Mutations targeting the SCF^{Fbx4-αB crystallin}E3 ubiquitin ligase result in impaired ligase activity and subsequent cyclin D1/CDK4 accumulation in the nucleus.</p></table-wrap-foot>" ]

[ "<graphic xlink:href=\"1747-1028-3-12-1\"/>" ]

[]

{ "acronym": [], "definition": [] }

[ "<title>Introduction</title>", "<p>Remitting Seronegative Symmetrical Synovitis with Pitting oedema (RS3PE) syndrome, a subset of acute onset polyarthritis mainly affects the older people and predominantly males with clinical manifestations of acute onset pitting oedema of the hands. Other notable features include seronegativity for Rheumatoid factor and an excellent response to low dose steroids with long-term remission.</p>", "<p>In 1985, McCarty et al. [##REF##4057484##1##] described the first case of Remitting Seronegative Symmetrical Synovitis with Pitting Oedema (RS3PE) syndrome characterized by symmetrical distal synovitis and tenosynovitis of the mucous sheaths of the flexor and extensor tendons of the hands associated with pitting oedema of the hands and/or feet.</p>", "<p>In this case report we discuss a case of RS3PE syndrome where the initial presentation was bilateral pitting oedema of the extremities without any other systemic cause. Patient showed dramatic response to low dose steroids.</p>" ]

[]

[]

[ "<title>Discussion</title>", "<p>In 1985, Mc Carty [##REF##4057484##1##] described first RS3PE syndrome in a series of 10 patients. Following a retrospective multicenter study of patients Olive <italic>et al</italic>. [##UREF##0##2##] proposed the following diagnostic criteria for this syndrome:</p>", "<p>1. Bilateral pitting oedema of both hands</p>", "<p>2. Sudden onset of polyarthritis</p>", "<p>3. Age more than 50 years</p>", "<p>4. Seronegative rheumatoid factor.</p>", "<p>The exact incidence and prevalence are not known. RS3PE affects more men than women with ratio of 2:1 and more frequently the older people.</p>", "<p>The aetiology and the pathogenic mechanisms are not clear. The syndrome was reported to be associated with HLA B7 [##REF##4057484##1##] and HLA A2 haplotype [##REF##7582734##3##].</p>", "<p>Recently, vascular endothelial growth factor (VEGF) has been implicated as a contributing factor for pathological changes responsible for both hypervascularity (synovitis) and vascular permeability (subcutaneous oedema) [##REF##16227418##4##].</p>", "<p>Aetiology of pitting oedema is not known, but recent MRI studies suggest that marked extensor tenosynovitis is the principle lesion responsible for oedema of subcutaneous and peritendinous soft tissue [##REF##9034979##5##].</p>", "<p>Fever and asthenia could be non-specific manifestations of inflammation, but presence of other systemic signs like weight loss, anorexia and poor response to steroids could indicate a paraneoplastic manifestation.</p>", "<p>Evaluation for pedal oedema should aim at ruling out the other possibilities like congestive cardiac failure, nephritic syndrome and hypothyroidism.</p>", "<p>In older people it is important to distinguish this syndrome from PMR (Polymyalgia Rheumatica) in view of the duration of treatment with steroids that is needed. This is even more pertinent in view of the long-term consequences of the use of steroids in older people, as these patients are more likely to be already having multiple co morbid factors like osteoporosis, hypertension, diabetes and heart failure. RS3PE syndrome can be associated with both solid tumours like gastric [##REF##9150093##6##], pancreatic [##REF##9458232##7##] and haematological malignancies like non-hodgkin's lymphoma [##REF##15474395##8##].</p>", "<p>Patients with idiopathic RS3PE showed an excellent response to low doses of corticosteroids compared to the poor response to RS3PE in association with neoplasia [##UREF##1##9##].</p>", "<p>Clinicians need to be aware of the RS3PE in ageing population and initiate appropriate investigations to exclude any occult malignancy. The search for occult malignancy is particularly crucial in patients whose systemic symptoms are prominent and who are still not responsive to steroids.</p>", "<p>The main differential diagnosis of RS3PE is polymyalgia rheumatica, which can be very difficult in older people. The features helpful in differentiating the two can be seen in Table ##TAB##0##1##.</p>", "<p>Other differential diagnosis with RS3PE in older people includes Rheumatoid arthritis, late onset spondyloarthropathy, mixed connective tissue disease, chondrocalcinosis and amyloid arthropathy.</p>", "<p>Blood tests may typically demonstrate raised inflammatory markers, discrete inflammatory anaemia, and a negative rheumatoid factor. X-rays of the hands and wrists may show soft tissue swelling but absence of erosions is classic. Tenosynovitis of both flexor and extensor tendons at the wrist and the extensor tendons of the feet is the hallmark of RS3PE. Ultrasonograph a reliable and cost effective modality for evaluation of patients with suspected RS3PE this characteristically shows tenosynovitis of flexor and extensor tendons [##UREF##2##10##].</p>", "<p>MRI is useful for monitoring the disease activity in RS3PE syndrome. It can also provide information about soft tissue, cartilage and bony erosions [##UREF##3##11##].</p>", "<p>Whole body Ga-67 scan can show increased uptake in hands and feet and this could be useful in assessing lesion activity [##REF##12973003##12##]. Importantly appropriate investigations need to be carried out if there is any suspicion of malignancy.</p>", "<p>Nonsteroidal anti-inflammatory drugs or Salicylates for pain relief. Most of the patients respond very well to low dose of steroids (10–15 mg prednisolone) and these patients show sustained and complete remission even after withdrawal of steroids. Some patients respond to hydroxyl chloroquine or gold salts.</p>" ]

[ "<title>Conclusion</title>", "<p>RS3PE is a definite syndrome, subset of polyarthritis with favourable outcome and has a good prognosis in the older patients. It may present, as a paraneoplastic manifestation especially in older people who show a poor response to steroids. If suspected, looking for underlying malignancy is recommended.</p>" ]

[ "<p>This is an Open Access article distributed under the terms of the Creative Commons Attribution License (<ext-link ext-link-type=\"uri\" xlink:href=\"http://creativecommons.org/licenses/by/2.0\"/>), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.</p>", "<p>Remitting Seronegative Symmetrical Synovitis with Pitting oedema syndrome, a rare inflammatory arthritis, commonly affects people in the older age group. It can present as an acute onset polyarthritis with associated pitting oedema of the extremities. Patients show excellent response to low dose steroids with complete and sustained remissions. It can also be a paraneoplastic manifestation of an underlying occult malignancy, hence thorough clinical evaluation is warranted.</p>", "<p>We discuss a case of Remitting Seronegative Symmetrical Synovitis with pitting oedema syndrome where the patient presented with acute onset polyarthritis and pitting oedema of the extremities without an underlying systemic cause. Patient showed dramatic response to low dose steroids.</p>" ]

[ "<title>Case presentation</title>", "<p>A 67-year-old male Caucasian patient was admitted with a 2-week history of painful swollen hands and painful knees associated with worsening mobility. This was preceded by a history of swinging low-grade pyrexia for two months and a history of progressive pedal oedema for 4 months. The symptoms were atraumatic in onset and lacked any associated features of connective tissue disease. There was a positive past medical history of systemic hypertension, hypothyroidism, diabetes mellitus, chronic renal failure and Parkinsons disease. The patient had been treated for carcinoma of the prostate gland in the past.</p>", "<p>Examination revealed bilateral pitting oedema of dorsum of hands and legs upto the calves. He also had synovitis at proximal interphalangeal joints, wrists and effusion of both knees and ankles. Initial blood test showed haemoglobin of 9.5 g/l with normochromic and nomocytic anaemia, raised inflammatory markers (ESR 70, CRP 100) and normal WBC. Autoantibody screen and rheumatoid factor were negative. Radiological findings of hands, feet and knees did not show any erosions. The patient was also screened and investigated for associated malignancies. He had normal tumour markers including CEA, AFP, CA19-9 and PSA (Prostrate specific antigen). CT thorax/abdomen and OGD were also reported as normal. In view of low-grade pyrexia, possibility of infective focus was ruled out by repeated blood and urine cultures. A diagnosis of remitting symmetrical seronegative synovitis with pedal oedema was suggested and patient responded extremely well to low dose prednisolone at 7.5 mg daily dosage. Further follow up 8 weeks later on tapering dose of prednisolone showed complete resolution of signs and symptoms without any further flare-ups.</p>", "<title>Abbreviations</title>", "<p>RS3PE: Remitting seronegative symmetrical synovitis with pitting oedema, NSAID: Nonsteroidal anti-inflammatory drug, MRI: Magnetic resonance imaging, HLA: Human leukocyte antigen, PMR: Polymyalgia Rheumatica, WBC: White blood cell, ESR: Erythrocyte sedimentation rate, CRP: C-reactive protein, CEA: Carcinoembryonic antigen, AFP: Alpha-fetoprotein, CA19-9: Carbohydrate antigen 19-9, PSA: Prostate specific antigen, VEGF: Vascular endothelial growth factor.</p>", "<title>Competing interests</title>", "<p>The authors declare that they have no competing interests.</p>", "<title>Authors' contributions</title>", "<p>AS was involved in the initial drafting and formatting of manuscript. RH and TS revised and corrected the manuscript. All authors have read and approved the final manuscript.</p>", "<title>Consent</title>", "<p>Written informed consent was obtained from the patient for publication of this case report and accompanying images. A copy of the written consent is available for review by the Editor-in-Chief of this journal ' [see Additional file ##SUPPL##0##1##]'.</p>", "<title>Supplementary Material</title>" ]

[]

[]

[ "<table-wrap position=\"float\" id=\"T1\"><label>Table 1</label><caption><p>Differentiating features of RS3PE Syndrome and PMR</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td align=\"left\"><bold>RS3PE Syndrome</bold></td><td align=\"left\"><bold>PMR</bold></td></tr></thead><tbody><tr><td align=\"left\">Dramatic response to low dose steroids or NSAIDS and sometimes to hydroxychloroquine.</td><td align=\"left\">Responds only to steroids</td></tr><tr><td colspan=\"2\"><hr/></td></tr><tr><td align=\"left\">Common in males</td><td align=\"left\">More frequent in females</td></tr><tr><td colspan=\"2\"><hr/></td></tr><tr><td align=\"left\">Mainly involves wrist with pitting oedema.</td><td align=\"left\">Involves shoulder and pelvic girdle with associated systemic symptoms. Pitting oedema present rarely.</td></tr><tr><td colspan=\"2\"><hr/></td></tr><tr><td align=\"left\">Association with HLA-B7, B27, A2</td><td align=\"left\">HLA association with HLA-DR4</td></tr><tr><td colspan=\"2\"><hr/></td></tr><tr><td align=\"left\">Excellent long-term prognosis</td><td align=\"left\">Frequent relapses and recurrences.</td></tr></tbody></table></table-wrap>" ]

[]

[]

[]

[]

[]

[ "<supplementary-material content-type=\"local-data\" id=\"S1\"><caption><title>Additional File 1</title><p>Consent form page 1</p></caption></supplementary-material>" ]

[]

[]

[ "<media xlink:href=\"1757-1626-1-132-S1.jpeg\" mimetype=\"image\" mime-subtype=\"jpeg\"><caption><p>Click here for file</p></caption></media>" ]

{ "acronym": [], "definition": [] }

[ "<title>Background</title>", "<p>Pneumothorax is a pulmonary complication that occurs rarely during pregnancy and a few cases of pneumothorax, pneumomediastinum and extensive subcutenous emphysema have been reported [##REF##15321418##1##].</p>", "<p>Tracheo-esophageal fistula (TEF) formation is a rare complication of endotracheal intubation. This complication is generally thought to be iatrogenic and occurs in less than 1% of patients. High-volume, low-pressure cuffs have made TEF an infrequent occurrence; however, it still poses as a potential life-threatening condition. Acquired TEF may cause pneumomediastinum, pneumothorax or subcutaneous emphysema due to the leakage of air from trachea to the neighboring structures [##REF##17706579##2##].</p>", "<p>Here, we present a case of pneumomediastinum, pneumothorax and subcutaneous emphysema developing post-cesarean due to TEF which has been thought to be the result of traumatic intubation.</p>" ]

[]

[]

[ "<title>Discussion</title>", "<p>The differential diagnosis of acute dyspnea in postpartum period should include pulmonary embolism caused by venous thrombosis, amniotic fluid or air embolism as well as pneumothorax and tension pneumothorax, though occurring less frequently [##UREF##0##3##]. Cases of spontaneous pneumothorax in pregnancy were rarely reported in the literature. Miguil and colleagues reported a case of spontaneous pneumothorax, pneumomediastinum and emphysema in a 19 year old primiparous patient during labour [##REF##15321418##1##].</p>", "<p>Tension pneumothorax is a life-threating condition with severe cardiorespiratory compromise. It occurs when air enters the pleural cavity on inspiration but, because of a ball valve mechanism, does not exit on expiration. This progressively enlarges the pleural space and thereby increases intrathoracic pressure. Venous return to the heart is decreased leading to reduced cardiac output and subsequently decreased blood pressure. Hypoxia results from increased shunt caused by continued perfusion of unventilated lung areas. Chest radiography should be performed if the diagnosis is unclear [##UREF##0##3##].</p>", "<p>Risk factors for pneumothorax are respiratory infection, asthma, previous pneumothorax history, the intake of cocaine and ecstasy, insertion of central venous catheter and performing endotracheal anaesthesia with intermittent positive-pressure ventilation(IPPV). General anesthesia with IPPV for caesarean section may cause barotrauma, particularly during operation, since coughing is common when the patient is extubated fully awake [##UREF##0##3##]. Evron and colleagues[##REF##3900168##4##] reported a case of bilateral pneumothorax, subcutaneous emphysema, pneumomediastinum, pneumoretroperitoneum and pneumoperitoneum detected in a 28-year-old healthy pregnant after intubation. They pointed out that this rare complication was the result of positive-pressure ventilation performed under general anaesthesia [##REF##3900168##4##]. Harris presented a case with tension pneumothorax after cesarean section [##REF##10781474##5##]. Acquired TEF is a rare complication of tracheal intubation, and usually results from cuff-related tracheal injury. It is known that high cuff pressure has a detrimental affect. Normal capillary perfusion pressure for tracheal mucous tissue is 20–30 cmH₂O. If the cuff pressure is over 39 cmH₂O, capillary perfusion in mucous tissue may cease [##REF##7826804##6##].</p>", "<p>Problems during endotracheal intubation may cause iatrogenic trauma of the upper airways [##REF##10781474##5##]. The direct causes of the rupture are difficult tracheal intubation, particularly with a stylet inside the tube and overdistension of the cuff of the tracheal tube [##REF##17235268##7##].</p>", "<p>In the presented case, the difficulty of intubation, use of inappropriate stylet, use of high pressure-low volume endotracheal tube, and being a pregnant woman were the risk factors for development of TEF.</p>", "<p>In study of Kalaud and colleagues, use of stylet in intubation in 4 of the 12 cases have been mentioned. The size of endotracheal tube and swelling of cuff may contribute to trauma. Many researchers assert that prevalence of iatrogenic tracheal rupture is higher in females and this assertion leads to conclusion that the membraneous trachea is less firm in women and children as compared with men [##REF##9315814##8##].</p>", "<p>The most common physical finding of TEF is subcutaneous emphysema in the neck or upper chest, especially during positive pressure mask ventilation, which can force gas between fascial planes into the mediastinum and subcutaneous tissue. Diagnostic methods include plain chest radiograph, which often shows subcutaneous emphysema, pneumothorax, pleural effusion, and pneumomediastenium [##REF##12933429##9##]. In our patient, we detected extensive subcutaneous emphysema in the neck and left upper chest on admission. Subsequent chest X-ray revealed subcutaneous emphysema, left pneumothorax, and pneumomediastenium.</p>", "<p>When diagnosed after extubation, the most frequent sign of TEF is coughing after swallowing. A high index of suspicion is required in patients at risk for developing a TEF. The diagnostic evaluation is by bronchoscopy and esophagoscopy [##REF##12755313##10##]. Similarly, TEF was diagnosed by bronchoscopy and esophagoscopy performed due to coughing with swallowing which developed after extubation in our patient.</p>", "<p>When the diagnosis has been made, the immediate goal should be to minimize tracheobronchial soilage by placing the cuff of a tracheostomy tube distal to the fistula. The basic aim of the treatment is to improve airway contamination and insufficient nutrition. Reflux of gastric contents is diminished by placement of a gastrostomy tube, and adequate nutrition is facilitated by inserting a jejunostomy tube. Surgical correction is required because spontaneous closure is rare, but surgery should be postponed until the patient is weaned from mechanical ventilation [##REF##12755313##10##].</p>" ]

[ "<title>Conclusion</title>", "<p>The presence of traumatic intubation attempts is known to constitute a risk to TEF or iatrogenic injuries to trachea. We believe that, TEF observed in our case is secondary to difficult and traumatic intubation which is known as one of the risk factors for TEF. During traumatic intubation, when the possibility of esophageal injury cannot be excluded, urgent endoscopy or water-soluble contrast radiography may be prudent. Our experience confirms that early diagnosis and management is associated with a more favorable outcome.</p>" ]

[ "<p>This is an Open Access article distributed under the terms of the Creative Commons Attribution License (<ext-link ext-link-type=\"uri\" xlink:href=\"http://creativecommons.org/licenses/by/2.0\"/>), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.</p>", "<title>Background</title>", "<p>The non-malignant, acquired tracheoesophageal fistulas (TEF), resulting from tracheal intubation are usually iatrogenic lesions. Tracheal lesions resulting from intubation may occur and pneumomediastinum, pneumothorax or subcutaneous emphysema may develop due to the stream of air.</p>", "<title>Case Presentation</title>", "<p>We present a-39-year old, Caucasian patient, developing severe hypoxia fallowing cesarean section under general anesthesia. The findings of the patient were diffuse subcutaneous emphysema; together with pneumothorax and pneumomediastinum TEF was diagnosed in the patient by bronchoscopy and eusophagoscopy performed due to cough and difficulty in swallowing developing after extubation.</p>", "<title>Conclusion</title>", "<p>It is important to the clinicians to be aware of the TEF can be accompanied to the traumatic intubation and urgent endoscopy or water-soluble contrast radiography may be prudent.</p>" ]

[ "<title>Case presentation</title>", "<p>A 39-year old, Caucasion, multigravida female was admitted for induction of labour at 42 weeks' gestation at obstetric unit of a peripheral hospital. In her past history, she gave birth to her five children by vaginal delivery without any complications. She was a non-smoker and had no chest disease. Artificial rupture of membranes demonstrated heavily meconium-stained liquor, and cardiotocograpy showed repetitive late decelaritons with diminished beat-to beat variability and emergency caesarean section was planned. Preoperatively she was oriented. She had pulse 115 beats/min, respiratory rate 12/min and blood pressure 160/90 mmHg. Airway assessment identified class 3 Mallampati. Examination of the chest revealed normal vesicular breathing without any added pathological sounds. As we learned from her obstetrician, following a rapid sequence induction, intubation was performed in the second attempt by using stylet with number eight low volume-high pressure endotracheal tube by a very junior medical officer and was considered difficult. General anesthesia was maintained by isoflurane end-tidal concentration 0,8–1% in oxygen and N₂O. Although no complication was observed during the operation, she suddenly developed respiratory distress and persistent hypoxia soon after extubation. Within the next 3 min, the patient's saturation declined to 65% in spite of 100% O₂support, and her blood pressure dropped to 65/40 mmHg with a pulse of 158 beats/min. She was reintubated and was refered to our hospital as intubated and ventilated with transport ventilation for further investigation and management.</p>", "<p>On the first examination, she had Glasgow Coma Scale of E₂M₃V₃. Her blood pressure was 70/40 mmHg with a pulse of 145 beats/min, and her O₂saturation was 65% with pulse oximeter. Breath sounds was not heard on auscultation on the left hemithorax while cardiovascular examination was normal except for tachycardia. Extensive subcutaneous emphysema was noted by palpation mainly on the left side of the thorax and neck. Immediately diagnostic pleural puncture was performed and air was aspirated from the left apical region of pleura.</p>", "<p>The tube-thoracostomy was performed and 28 French sized chest tube with an under water seal drain was inserted into the chest wall at the 5th intercostal space in the left mid axillary line. A portable chest X-ray revealed the left pneumothorax together with minimal right pneumothorax, pneumomediastinum and subcutaneous emphysema. The mediastinum was enlarged and thoracostomy tube was observed to reach the left hilar region (Figure ##FIG##0##1##). Computerized tomography (CT) of the thorax revealed the presence of excessive air around cervical soft tissue, in the thorax and mediastinum (Figure ##FIG##1##2##). In the second day, her oxygenation deteriorated suddenly and it was seen on chest radiography that pneumothorax involved the right hemithorax. completely. A second chest-tube was also inserted into the right chest wall. The fiberoptic bronchoscopy was performed due to extensive atelectasia seen on chest X-ray. Bronchoscopic examination revealed diffuse hemorrhage in trachea and thrombothic plug obstructing the left bronchus almost completely. The plug was removed from the left main bronchus. Since clinical and radiological findings of pneumothorax improved, the right and left intercostals drains were removed on the fifth day. The patient was also extubated on the same day, as the oxygen saturation was persistently greater than 97% on room air. The nutrition was provided enterally with nasogastric tube until the 5^thday, and then she was orally fed with water and clear liquids on the first extubation day. She complained of serious throat ache and coughing during intake of food, however throat and indirect larynx examination revealed no problem explaining this condition. So, bronchoscopy and endoscopy was repeated, and TEF was diagnosed. The fistulous tract was located approximately 3 cm inferior to vocal cords, 25 cm distant from the teeth and 5–6-cm over the carina level. Following the diagnosis of TEF, percutaneous endoscopic jejunostomy was performed. Two weeks later, TEF was repaired by cervical approach. Postoperatively, the patient received parenteral nutrition for 7 days and on postoperative 9^thday, she had a gastrografin swallow study, which showed no evidence of TEF. The patient was discharged from the hospital on postoperative 14^thday.</p>", "<title>Abbreviations</title>", "<p>Tracheo-esophageal fistula: TEF, Computerized tomography: CT, intermittent positive-pressure ventilation: IPPV</p>", "<title>Authors' contributions</title>", "<p>HO carried out the patient's diagnosis, drafted the manuscript. NS performed the case management, drafted the manuscript. BZ and ME participated in the patient's management. MFY participated in the writing of the case report. All authors read and approved the final manuscript.</p>", "<title>Consent</title>", "<p>Written informed consent was obtained from the patient for publication of this case report and accompanying images. A copy of the written consent is available for review by the Editor-in-Chief of this journal\"</p>" ]

[ "<title>Acknowledgements</title>", "<p>Thanks to Gulizar Sokmen, MD for English editing.</p>" ]

[ "<fig position=\"float\" id=\"F1\"><label>Figure 1</label><caption><p><bold>The first X-ray graphy in the emergency service after left tube thoracostomy.</bold> Mediastinum was enlarged in chest roentgenograms.</p></caption></fig>", "<fig position=\"float\" id=\"F2\"><label>Figure 2</label><caption><p>Pneumothorax regressed after tube thoracostomy.</p></caption></fig>" ]

[]

[]

[]

[]

[]

[]

[]

[]

[ "<graphic xlink:href=\"1757-1626-1-134-1\"/>", "<graphic xlink:href=\"1757-1626-1-134-2\"/>" ]

[]

{ "acronym": [], "definition": [] }

[ "<title>Background</title>", "<p>Despite technological advances in proteomics, analysis of complex biological samples remains a significant challenge [##REF##12015990##1##]. Some of the most complex biological samples routinely submitted for proteomic profiling include serum or plasma [##REF##17682341##2##, ####REF##17636887##3##, ##REF##17552920##4####17552920##4##]. While disease markers abound in plasma are reflective of ongoing disease, complexity of sample due to post-translational modifications (PTMs) and protein isoforms presents major obstacles for biomarker identification [##REF##17049937##5##, ####REF##16991193##6##, ##REF##16196093##7##, ##REF##16000084##8####16000084##8##]. Reproducibility, sensitivity, resolution, and high throughput analysis are among the developing areas for proteomic platforms [##REF##17288515##9##]. The need for better separation and analytical techniques cannot be overstated [##REF##17166507##10##].</p>", "<p>Improvements in proteomics platforms have been realized in recent years [##REF##17341690##11##,##REF##18162499##12##]. For example, 2D SDS-PAGE when combined with difference in gel electrophoresis (DIGE) was developed as a profiling platform wherein proteins are identified based on electrospray ionization mass spectrometry (ESI-MS/MS) of trypsin-derived peptides. However, resolution of hydrophobic proteins and those within a high molecular mass range is limited [##REF##17503404##13##, ####REF##15900442##14##, ##REF##14708031##15####14708031##15##]. This may be overcome using a combination of molecular sieving chromatography with the multi-dimensional protein identification technology (MudPIT) or protein microarrays [##REF##17034751##16##]. Other \"bottom-up\" tools, such as surface enhanced laser desorption ionization-time of flight (SELDI-TOF) and matrix-assisted labelled desorption/ionization-time of flight (MALDI-TOF), while useful, do not address PTMs in complex body fluids. Another critical need is protein quantification since changes in proteome profiles may be subtle, yet biologically significant [##REF##18162499##12##]. For example, protein glycosylation or phosphorylation leading to functional modification affects only a small percentage of the total protein pool linked to physiological changes. Therefore, fractionating complex protein mixtures while maintaining intact proteins in liquid phase is a most desirable feature for use in further analyses (\"top-down proteomics\"). Fractions collected in liquid phase would provide simpler and more informative secondary analysis, in contrast to gel-embedded proteins in 2D DIGE [##REF##16188874##17##, ####REF##16309368##18##, ##REF##16941567##19####16941567##19##] where intact protein recovery is difficult and associated with greater quantitative loss. Additionally, once protein is enzymatically digested for LC-MS/MS analysis, it cannot be used for other analysis such as Western blot assays.</p>", "<p>ProteomeLab™ PF 2D offers an alternative approach to protein profiling that addresses issues of complexity and utilization of fractions after analysis. To assess the utility of the PF 2D platform for proteomic profiling, we compared plasma samples recovered from amyotrophic lateral sclerosis (ALS) patients who were immunized with glatiramer acetate (GA) to those from non-immunized ALS patients. Strengths and weaknesses of this new proteomic platform for biomarker discovery are discussed.</p>" ]

[ "<title>Methods</title>", "<title>Samples</title>", "<p>Peripheral blood samples from ALS patients used in this study were obtained from Columbia University, New York. ALS patients were treated daily with 20 mg of GA over a six month period [##REF##16606934##22##]. Samples were collected in acid citrate dextrose tubes and after centrifugation at 800 × <italic>g </italic>for 10 min, the plasma was harvested, distributed into 1 mL aliquots, and stored at -80°C. Plasma from 3 ALS patients (collected prior to treatment and 2, 4, and 5 months after initiation of GA treatment) was used in this investigation.</p>", "<title>Sample delipidation</title>", "<p>Plasma samples were centrifuged at 18000 × <italic>g</italic>, 15 min at 4°C. The middle layer was collected and diluted 1:2.5 (0.25 mL plasma + 0.375 mL) in dilution buffer (10 mM Tris-HCl, pH 7.4, 0.15 M NaCl). Next, particles and aggregates were removed from samples by filtration through a 0.45 μm spin filter at 9200 × <italic>g </italic>for 1 min.</p>", "<title>Immunodepletion (partitioning)</title>", "<p>To remove 12 highly abundant proteins, we utilized the ProteomeLab™ IgY-12 High Capacity Proteome Partitioning Kit (Beckman Coulter, Fullerton, CA) according to manufacturer's recommendations. The kit included a LC10 affinity column (12.7 × 79.0 mm) with a capacity of 0.25 mL human plasma per cycle and optimized buffers for sample preparation, loading, washing, and eluting. The LC10 column contains affinity-purified chicken IgY antibodies directed against serum albumin, fibrinogen, IgG, transferin, IgA, IgM, apoA-I, apoA-II, haptoglobin, α1-antitrypsin, α2-macroglobulin, and α1-acid glycoprotein, which are covalently conjugated to polymeric microbeads. After the enriched flow through fractions containing low to medium abundant proteins were collected, the bound and highly abundant proteins were eluted with stripping buffer (0.1 M Glycine-HCl, pH 2.5). The column was then neutralized with 0.1 M Tris-HCl, pH 8.0 buffer. Finally, the column was re-equilibrated with dilution buffer at a flow rate of 2 mL/min. Collected bound fractions were neutralized with 0.1 M Tris-HCl. Flow through and eluted fractions were stored at -80°C until further analysis.</p>", "<title>Sample preparation for PF 2D first dimension (isoelectric focusing)</title>", "<p>Collected flow through fractions were thawed at room temperature and concentrated down to 0.50 mL using an Amicon Ultra-15 centrifugal filter (Millipore, Billerica, MA) previously washed with 3 mL of ProteomeLab™ Start Buffer (Beckman Coulter, pH 8.5). Next, 2 mL of plasma denaturing buffer (7.5 M urea, 2.5 M thiourea, 12.5% glycerol, 62.5 mM Tris-HCl, 2.5% (w/v) n-octylglucoside, 1.25 mM EDTA) was added to the concentrator and left at room temperature for 30 min while shaking. Samples were removed and centrifuged using two 1.5 mL screw-cap microcentrifuge tubes at 15,000 × <italic>g </italic>for 1 hr at 18°C. PD-10 Desalting Columns (GE Healthcare) were prepared by equilibration with 25 mL of Start Buffer. Sample was removed from the plasma denaturing buffer and placed into Start Buffer by placing sample load (2.5 mL) onto PD-10 column, discarding effluent, and collecting the desalted sample with 3.5 mL of Start Buffer. Resulting sample was filtered through a 0.45 μm spin filter previously washed with the Start Buffer. Protein concentration was determined by the Micro BCA Assay Kit (Pierce Biotechnology, Rockford, IL).</p>", "<title>Profiling</title>", "<p>Protein profiling using ProteomeLab™ PF 2D system consists of two steps: first dimension fractionation is chromatofocusing and second dimension is reverse phase HPLC fractionation. The 32 Karat™ Software (Beckman Coulter) was used for data processing and calculation of peak areas and heights.</p>", "<title>First dimension fractionation</title>", "<p>The first dimension was performed at room temperature with a flow rate of 0.2 mL/min using the HPCF column Start Buffer (pH 8.5), Eluent Buffer (pH 4.0), High Ionic Strength Wash + (1 M NaCl in 30% Isopropanol), and water. It was performed at room temperature with a flow rate of 0.2 mL/min. After equilibration with the Start Buffer for 130 min, the samples (1–5 mg of protein) were injected onto the chromatofocusing column and proteins separated based on isoelectric point (pI). Thirty-five min after injection, Eluent Buffer was initiated to generate a pH gradient (8.5–4.0). Shortly after the gradient reached pH 4.0, the column was washed with HISS+ (135–175 min after injection) to remove hydrophobic proteins and proteins with a pI under 4.0. Finally, the column was washed with water for 45 minutes (175–220 min after injection). Proteins were detected by absorbance at 280 nm by a UV detector. Based on the pH gradient generated, fractions were collected with the fraction collector/injector module (FC/I) at 0.3 pH intervals during pH gradient elutriation (8.3–4.0), otherwise fractions were collected every 8.5 min.</p>", "<title>Second dimension fractionation</title>", "<p>The second dimension separations were performed with an RP-HPLC column and two solvents, 0.1% trifluoroacetic acid (TFA) in HPLC water (Solvent A) and 0.08% TFA in acetonitrile (ACN) (Solvent B). Separation was executed at 50°C at a flow rate of 0.75 mL/min and protein containing fractions detected by UV absorbance at 214 nm. Equilibration was achieved with Solvent A for 10 minutes followed by Solvent B for 5 minutes prior to each injection. From selected first dimension fractions, 0.250 mL were injected, run for two minutes, and the column eluted with a linear gradient of 0–100% Solvent B for 30 min. (3.33% change in B solvent/min). Next, Solvent B was continued for four minutes, followed by re-equilibration with 100% Solvent A for eight minutes. Second dimension fractions were collected at 40 second intervals.</p>", "<title>In solution trypsin digestion</title>", "<p>The amount of protein from each fraction used for digestion was determined based on peak UV absorbance at 214 nm. Samples were dried to 10 μL followed by addition of 50 mM ammonium bicarbonate and DTT. After 1 hr incubation in a 60°C water bath, trypsin (0.125 mg/mL) was added and samples incubated at 37°C for 14–16 hours. Trypsinized samples were sonicated using a Branson water bath sonicator for 5 seconds, a second volume of trypsin added, and samples incubated at 37°C for 8–10 hours. Formic acid was added to a final concentration of 0.1% to stop the reaction. Peptides were purified using ZipTip^®(Millipore Corporation, Billerica, MA) according to manufacturer's recommendation. Eluates were dried by vacuum centrifugation and resuspended in 12 μL of HPLC water with 0.1% formic acid prior to mass spectrometer analysis.</p>", "<title>Protein identification by nano-LC-MS/MS</title>", "<p>Peptides were fractionated on a RP-C18 microcapillary column and sequenced using electrospray ionization-liquid chromatography-mass spectrometry system (ESI-LC-MS/MS) (ProteomeX system equipped with LCQDeca<italic>XP</italic>Plus mass spectrometer, ThermoElectron, Inc., San Jose, CA) in a nanospray configuration. The mass accuracy of the LCQDeca<italic>XP</italic>Plus is 500 ppm +/- 100 ppm. Database nr.fasta was retrieved from the NCBI FTP server <ext-link ext-link-type=\"uri\" xlink:href=\"http://www.ncbi.nlm.nih.gov/Ftp/\"/> updated on April 11^th, 2008. The spectra obtained through LC-MS/MS analysis were searched against the protein database narrowed to a subset of human proteins (keywords: Homo sapiens, man, human, primate) using SEQUEST algorithm (BioWorks 3.2 software from ThermoElectron, Inc.). We excluded keratins from our database search based on previous observations that these are contaminants resulting from sample processing. In TurboSEQUEST Search Parameters, threshold for Dta generation was 10000 and precursor mass tolerance for Dta generation was set at 1.4. For Dta Search, peptide tolerance was set at 1.5 and fragment ions tolerance at 0.00. Charge state was set on \"Auto.\" At least two sequenced peptides were required from each protein for high confidence identification.</p>", "<title>SDS-PAGE</title>", "<p>Selected second dimension protein fractions were resolved further by electrophoresis through a NuPAGE^®Novex 4–12% Bis-Tris Gel (Invitrogen, Carlsbad, CA, USA). Fractions were dried, resuspended in 20 μL of 1× sample buffer with reducing agent (Invitrogen), and heated at 95°C for 5 minutes. Electrophoresis was performed at a constant 100 V for 90 min and the gel was fixed in 40% methanol/7% acetic acid. After staining with SYPRO^®Ruby protein gel stain (Invitrogen), destaining was achieved with 10% methanol/7% acetic acid and the gel scanned using a Typhoon 9410 high performance laser scanning system (GE Healthcare/Amersham, Piscata, NJ). Gels were subsequently counterstained with Coomassie Brilliant Blue G-Colloidal (Sigma, St. Louis, MO) and bands excised for in gel tryptic digestion.</p>", "<title>In gel trypsin digestion</title>", "<p>Gel pieces were distained for 1 hr at room temperature using 100 μL of 50% acetonitrile/50 mM NH₄HCO₃. Gel pieces were dried and incubated with trypsin in 10 mM NH₄HCO₃(Promega, Madison, WI) overnight at 37°C. Peptides were extracted from trypsinized gels by washing gel pieces for 2 hours with 0.1% TFA and 60% ACN and were purified using ZipTip (Millipore Corporation). ZipTip eluates were dried and resuspended in 12 μL of HPLC water with 0.1% formic acid prior to mass spectrometer analysis.</p>" ]

[ "<title>Results</title>", "<title>Plasma sample immunodepletion</title>", "<p>Patient serum/plasma represents clinical material that is easily obtained with fewer restrictions compared to other sample types, including cerebrospinal fluid or tissue biopsy; thus, it is one of the most commonly tested patient material from which diagnostic tests are performed. Collectively, protein concentrations span a very broad range (10¹²-fold) [##REF##12488461##20##] in serum/plasma from which the differential concentration in individual protein concentrations presents potential targets for the discovery of clinically important biomarkers. However, the presence of very highly abundant proteins and the complexity of plasma proteins present formidable challenges. Twelve of the most abundant proteins comprise ~96% of the total protein mass from human plasma, with albumin comprising approximately 40–50% of protein. Presence of these abundant proteins in plasma samples masks differential levels of low to medium abundant ones. One strategy is to remove the most abundant proteins prior to profiling. To assess that approach, we used immunoaffinity chromatography with a column that is based on IgY technology to selectively remove 12 of the most abundant proteins in human serum/plasma (Figure ##FIG##0##1##, experimental design). One caveat of immunodepletion is that potential biomarkers that bind to albumin or highly abundant proteins may also be completely or partially depleted from serum samples through protein-protein interactions. However, this possibility can be evaluated with further analyses upon elution of the adherent protein fraction. Although the IgY-12 LC10 affinity column has the highest capacity of commercial immunoaffinity products currently available, only 250 μL of plasma sample can be processed during one chromatographic cycle. Flow through fractions (8–27 minutes) containing unbound proteins (Figure ##FIG##1##2##) were collected, pooled, concentrated, and submitted for analysis on the PF 2D platform. Protein yields from the flow through fractions were between 0.73 and 3.0 mg per mL of plasma.</p>", "<title>Protein fractionation 2-dimensions (PF 2D)</title>", "<p>The proteomic profiling platform ProteomeLab™ PF 2D offers 2-dimensional fractionation in which intact proteins are first separated by chromatofocusing proteins by pI and separated in the second dimension by their hydrophobic properties. The pH profiles from the chromatofocusing absorbencies were obtained from first dimension separation at 280 nm and thirty fractions were selected from each sample to submit for second dimension separation by hydrophobic chromatography. Second dimension absorbance profiles were compiled and displayed as a two-dimensional map using a feature of Mapping Tools software. The map displays pI fractions as lanes with the colour intensity of each band (absorbance at 214 nm) corresponding to protein bands located at their retention time of the second dimension separation (Figure ##FIG##2##3A## and ##FIG##2##3B##, Table ##TAB##0##1##).</p>", "<p>Challenges associated with peak alignment are similar to those found during the alignment of spots in the analysis of 2-dimensional gel electrophoresis. Comparison of two separate UV/pI maps consisting of the entire pH gradient was afforded by a module (DeltaVue) of the Mapping Tools data processing software. A second module (MultiVue) enables the analysis of a selected pH lane from multiple sample runs. In both methods, proper peak alignment between samples provides a critical analytical function. Using the Paired Peak function of MultiVue in Mapping Tools, second dimension chromatograms were aligned by setting the paired peak value at +/- 0.5% to initially determine if peaks should be paired. After confirmation of automatic pairing, minor manual adjustments of retention times (RT) were made to accommodate alignment protocols. For example, raw data from one analyzed peak (fraction corresponding to pH 5.89-5.59 and RT 16.04–16.71) yielded a raw range of 8.4 seconds in RT throughout six samples. However, following alignment adjustments, the range was within 4.2 seconds, providing greater assurance when selecting peaks for quantitative and qualitative analysis.</p>", "<p>Our data indicate that alignment and comparisons of the first dimension separation profiles were consistent. Alignment in second dimension, RP-HPLC, was also precise as differences in retention times between samples were in a range of less than 15 seconds. Figure ##FIG##3##4## shows an example in which differences in retention times are less than 6 sec (0.1 min). Although automatic alignment is a standard software feature utilized in these studies for course alignment, manual alignment has proven necessary to refine peak analysis. Protein identification by LC-MS/MS presented in Table ##TAB##1##2## afforded a level of confidence in our (manual) method of alignment. Nevertheless, alignment, whether automatic or manual, should be used with caution and peak identification must be validated.</p>", "<p>When comparing corresponding fractions from different samples, differences in retention times become critical for peak resolution since fraction collection intervals are determined prior to profile analysis. Therefore, a difference of several seconds may result in one peak being split into two fractions in one sample while collected entirely in one fraction for another. This is an inherent issue with LC based separation, whether stand-alone or in combination with other modes of separation. Thus, methods of separation and collection must be considered at the time when parameters are first set. One of several approaches to analyzing divided peaks involves determining the protein composition of each fraction. Another approach is to pool both adjacent fractions to increase the chances of protein identification from the corresponding split peak. However, pooling fractions may create more complexity in the sample and affect further analysis, such as increasing the number of proteins identified within the pool.</p>", "<title>Protein identification</title>", "<p>High confidence protein identification is an essential step in most proteomic studies. The ProteomeLab™ PF 2D, unlike 2D DIGE, offers the added advantage that collected fractions are in liquid phase and can be utilized directly for any of various analytical procedures, such as mass spectrometer analysis, enzymatic digests, additional fractionation, Western blot, or a combination of analytical tests. Additionally, more material can be fractionated using 2D LC (up to 5 mg with the PF 2D) than with gel electrophoresis, thus significantly increasing the sensitivity of protein identification.</p>", "<p>In the first step of this study, we investigated the correlation between detection of individual plasma fractions by UV absorbance and our ability to identify proteins with high confidence (two or more peptides). We digested proteins in a given plasma sample with trypsin and analyzed the digest for resulting peptides using nano-LC-MS/MS sequencing. Table ##TAB##0##1## summarizes fractions with analogous pH, retention time, and similar peak height absorbance chromatograms. Selected fractions were digested and analyzed by nano-LC-MS/MS. Fractions with a peak height greater than 0.100 absorbance units at 214 nm provided enough material to identify proteins with high confidence using our instrumentation. Proteins in fractions with peak height below 0.050 were either identified with low confidence or remained unidentified. Proteins in fractions within the intermediate peak height range of 0.05 and 0.10 appeared to be the lower limit of identification by nano-LC-MS/MS. Table ##TAB##1##2## displays only the proteins identified with high confidence from examined fractions. Using both automatic and manual alignment protocols, the same proteins were identified among several fractions across all six analyzed samples. These results indicated that using automatic alignment with minor manual adjustments provides enough confidence to pool corresponding sample fractions with low protein content and use them for high confidence protein identification without risk of mixing neighbouring peaks. The capacity to provide peak pooling enhances the utility of this proteomic platform.</p>", "<title>Fourth dimension fractionation</title>", "<p>One of the advantages of the PF 2D profiling platform is a possibility of fractionation in the fourth dimension. In our experimental design, IgY immunodepletion (partitioning) served as first dimension partitioning, isoelectric focusing provided second dimension analysis, and RP-HPLC yielded the third dimension. Several fractions after 3 dimensional analyses were selected for a fourth dimension, 1-dimensional electrophoresis (1DE), to evaluate whether the fraction isolated as a single peak consisted of only one protein. Frequently, fractions (covering a retention time of 0.67 min) with high peak height contain several proteins whose quantities are opposite to each other, thus masking differential expression. We selected one matching fraction from all six samples (pH 5.9 to 5.6, retention time 16.04 to 16.71 min.) with peak height ranging between 1.11 and 1.76 (Table ##TAB##0##1##). Analysis of these six matching fractions based on levels of absorbance after 3-dimensional fractionation did not demonstrate statistical differences between fractions, thus indicating that the total amount of protein in each fraction was the same. We expected and further confirmed by 1 DE that these fractions consisted of multiple, non-separated, proteins. Therefore, in a subsequent step, we analyzed by 1 DE equal amounts of each fraction containing equal amounts of protein based on absorbance at 214 nm. As expected, this analysis showed multiple bands in each fraction (Figure ##FIG##4##5##). A characteristic pattern showed increased intensity of a protein band with a molecular mass above 62 kDa in samples from immunized patients. It is possible that using narrower pH fractionation, e.g. 0.1 units instead of 0.3 units used in this study, would help to further separate proteins. This band was identified as hemopexin based on nano-LC-MS/MS analysis of peptides derived from tryptic digests and demonstrates that levels of this protein are increased in patients immunized with GA. The most extensively studied function of hemopexin is its binding heme, having the highest affinity of any known protein. Also, as a heme scavenger, hemopexin protects organisms from the oxidative damage that can be caused by free heme. Interestingly, histidine-rich glycoprotein precursor, whose function is not understood and which has been found in this study, is also a heme binding protein. Both proteins are made by liver and secreted to plasma.</p>", "<p>Molecular mechanism(s) underlying ALS remain enigmatic and no curative or ameliorative therapy exists for this disorder that leads to degeneration of upper and lower motor neurons and ultimate death. Because ALS appears to be multifactoral, involving interactions among microglia, astrocytes, neurons, and muscles, one can anticipate that efficacious therapy would require a number of therapeutic approaches involving adjunct modalities that target several pathways associated with microglia activation and motor neuron degeneration. One immunomodulatory strategy using a synthetic polymeric immunogen, GA, which cross-reacts with CNS protein epitopes and is currently clinically utilized in the treatment of relapsing/remitting multiple sclerosis, targets several of these pathways [##REF##18223021##21##]. In a recent phase II clinical trial, GA treatment of ALS patients proved to be safe and well tolerated [##REF##16606934##22##]. Whereas significant diminution of lymphocyte proliferation was observed, subsequent studies showed increased concentration of anti-GA antibodies in plasma from treated patients [##REF##17653922##23##]. To fully evaluate plasma differences in immunized patients compared to pre-immunized ALS patients and probe putative treatment targets for ALS, we employed a strategy of protein fractionation that allowed the separation of plasma proteins into several hundred fractions that are amenable to downstream evaluation.</p>" ]

[ "<title>Discussion</title>", "<p>ProteomeLab™ PF 2D is a new technology platform in proteomics [##REF##16188874##17##, ####REF##16309368##18##, ##REF##16941567##19####16941567##19##,##REF##12918970##24##, ####REF##15726440##25##, ##REF##16964509##26####16964509##26##] and literature reports using this platform have continued to emerge, totaling more than 40 to date. Our first approach discovered fewer differences than originally expected. However, the capacity for fourth dimensional separation and additional analysis provided an opportunity to further investigate selected fractions that otherwise would not have been available using the typical \"bottom-up\", non-recoverable tandem methods of separation and analysis used in conventional proteomics. This was possible because intact proteins were initially separated, analyzed and recovered in liquid phase. We were also able to identify isogenous fractions to pool when individual fractions contained insufficient amounts of protein for identification, a feature made possible by the ability to reproduce inter-sample fractionation patterns and precisely align peaks among replicate samples. Another strength of this platform is the flexibility to choose the number of fractions acquired per sample. We selected 30 first dimension fractions for second dimension separation, resulting in fractions covering a wide pH range and producing over 700 fractions per sample. However, the number of analyzed fractions and method parameters are within the investigator's prerogative. First, the collection of fractions based on the pH gradient can be adjusted to increase or decrease fractions collected; this is especially beneficial when the proteins of interest occur in a narrow pH range. Second, larger pH intervals for fraction collection may result in a higher number of proteins contained within each fraction, diminishing the separation of proteins and the potential for biomarker discovery. Narrowing pH intervals increases resolution, but will also increase the number of fractions to be analyzed in the second dimension, taking more time and further decreasing throughput. Similarly, the number of first dimension fractions to be run in the second dimension can be selected. Also, slower gradient elution from RP-HPLC columns can be utilized but will result in longer second dimension runs and requires some initial experimentation to optimize the desired analysis. Adjusting pH parameters and acetonitrile gradients may be useful, but needs to be applied with caution to assure the overall benefit from profiling.</p>", "<p>We have yet to analyze fractions in great depth for PTM differences, although such modifications can be indicative of changes resulting from immunization. In particular, shifts in protein isoelectric point usually indicate changes resulting from PTMs [##REF##15859589##27##]. Therefore, it is possible that the identification of similar proteins from a variety of pH and RT fractions results from PTMs and/or protein fragmentation, as was seen in our study. Other approaches of sample fractionation prior to profiling can be included. For example, use of \"Equalizer^®beads\" [##REF##17321532##28##, ####REF##17877382##29##, ##REF##17276440##30####17276440##30##] might be an advantageous alternative or addition to broadly used immunodepletion of the most abundant proteins. Using lectin columns for isolation of subsets of post-translationally glycosylated proteins offers another approach to address challenges of profiling and discovery of biomarkers in a very complex mixture of proteins, especially when changes are subtle [##REF##17134948##31##,##REF##16782413##32##]. Still, given the sample recovery capacity and sample concentration, those studies are within the realm of this platform.</p>" ]

[ "<title>Conclusion</title>", "<p>In summary, we found the ProteomeLab™ PF 2D to be a useful automated proteomic profiling platform. There are several strengths of this approach. One advantage over the second dimension of 2-dimensional electrophoresis (2-DE) is the separation of proteins, which might be very similar in size but quite different in their biochemical characteristics due to post-translational modification. Also, this platform can be used for profiling basic and hydrophobic proteins that are hard to analyze by 2-DE [##REF##16404720##33##,##REF##16272560##34##]. Another advantage is that fractionated proteins are maintained in a liquid phase, making them available for various assays without loss of material (e.g., extraction from polyacrylamide gel) and/or allowing fractionation by other means. Moreover, due to the automated nature of this platform, the option of using only chromatofocusing for separation affords an attractive advantage over conventional gel-based separation by isoelectric focusing. Weaknesses of the ProteomeLab™ platform include low throughput, allowing 2–3 samples per week per instrument, and relatively large amount of sample required for analysis. Although multiple instruments can be run in parallel to increase throughput, this is very expensive. Large amount of sample is not an issue when serum/plasma samples are analyzed. However, other clinical material such as tissue biopsies might not be available in sufficient amounts. Also, the masking of differential protein expression among samples by opposite quantities of protein in a given fraction may present a problem when looking for potential biomarkers, as seen by the differential expression of hemopexin in our study. We have presented preliminary data regarding the effects of GA treatment for ALS and have found the ProteomeLab™ PF 2D to be a promising platform for protein profiling and a means for biomarker discovery.</p>" ]

[ "<p>This is an Open Access article distributed under the terms of the Creative Commons Attribution License (<ext-link ext-link-type=\"uri\" xlink:href=\"http://creativecommons.org/licenses/by/2.0\"/>), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.</p>", "<title>Background</title>", "<p>The ProteomeLab™ PF 2D platform is a relatively new approach to global protein profiling. Herein, it was used for investigation of plasma proteome changes in amyotrophic lateral sclerosis (ALS) patients before and during immunization with glatiramer acetate (GA) in a clinical trial.</p>", "<title>Results</title>", "<p>The experimental design included immunoaffinity depletion of 12 most abundant proteins from plasma samples with the ProteomeLab™ IgY-12 LC10 column kit as first dimension separation, also referred to as immuno-partitioning. Second and third dimension separations of the enriched proteome were performed on the PF 2D platform utilizing 2D isoelectric focusing and RP-HPLC with the resulting fractions collected for analysis. 1D gel electrophoresis was added as a fourth dimension when sufficient protein was available. Protein identification from collected fractions was performed using nano-LC-MS/MS approach. Analysis of differences in the resulting two-dimensional maps of fractions obtained from the PF 2D and the ability to identify proteins from these fractions allowed sensitivity threshold measurements. Masked proteins in the PF 2D fractions are discussed.</p>", "<title>Conclusion</title>", "<p>We offer some insight into the strengths and limitations of this emerging proteomic platform.</p>" ]

[ "<title>Competing interests</title>", "<p>The authors declare that they have no competing interests.</p>", "<title>Authors' contributions</title>", "<p>JDS has made substantial contributions to data acquisition and analysis. Has been involved in drafting the manuscript. WR has made substantial contributions to data acquisition and analysis. RS has made substantial contributions to conception and design. Has been involved in revising manuscript critically for important intellectual content. RLM has been involved in collecting samples for this study. Has been involved in drafting the manuscript and revising it critically for important intellectual content. HEG has made substantial contributions to conception and design. Has given final approval of the version to be published. PC has made substantial contributions to conception and design, data analysis and interpretation. Has been involved in drafting the manuscript and revising it critically for important intellectual content. All authors have read and approved the final manuscript.</p>" ]

[ "<title>Acknowledgements</title>", "<p>The authors thank Ms. Robin Taylor for outstanding administrative and computer support and for her effort in helping put this manuscript together in a timely manner. This work was funded by NIH grants 1R21 MH075662-01 (P.C.), P01 NS43985, 5R37NS036126 (to H.E.G.).</p>" ]

[ "<fig position=\"float\" id=\"F1\"><label>Figure 1</label><caption><p><bold>Experimental design</bold>. Experimental design of multidimensional fractionation using plasma from ALS patients involved in a phase II clinical trial with GA.</p></caption></fig>", "<fig position=\"float\" id=\"F2\"><label>Figure 2</label><caption><p><bold>Immunodepletion</bold>. Chromatography of immunodepletion of plasma using ProteomeLab™ IgY-12. 250 μL of human plasma was partitioned with LC 10 column at an absorbance of 280 nm. The Flow Through was collected (8–27 min) and used for further analysis with PF 2D. The bound fraction was not analyzed.</p></caption></fig>", "<fig position=\"float\" id=\"F3\"><label>Figure 3</label><caption><p><bold>Results of profiling analysis</bold>. A. PF 2D two-dimensional heat maps of a representative set of samples obtained from one individual before and after GA immunization. PF 2D first dimension separation is based on isoelectric point (pI). PF 2D second dimension separation utilizes reverse phase HPLC fractionation. B. Comparison of two aligned peaks from analyses shown in (A) displaying a quantitative difference in protein contents measured by peak area (volume). Colour scheme ranges from purple (low absorbance) to red (high absorbance). The difference between absorbencies is shown in the middle as either a red or green band, representing the sample with greater absorbance at a specific peak.</p></caption></fig>", "<fig position=\"float\" id=\"F4\"><label>Figure 4</label><caption><p><bold>RP-HPLC</bold>. Differences in retention times (RT) (second dimension) displayed as bands for a specific peak detected and aligned in 6 samples using MultiVue software. The displayed peak corresponds to Patient 3 before treatment and has a RT of 16.62 min. Range of retention times was less than 6 sec.</p></caption></fig>", "<fig position=\"float\" id=\"F5\"><label>Figure 5</label><caption><p><bold>1D Electrophoresis</bold>. 1D electrophoresis (fourth dimension fractionation) showing differential expression of protein band above 62 kDa m.w. marker. Hemopexin was identified as the most prominent protein in this band by LCMS/MS sequencing. Equal amount of protein (based on absorbance at 214 nm) from each sample was loaded per lane. Gel was stained with SyproRuby.</p></caption></fig>" ]

[ "<table-wrap position=\"float\" id=\"T1\"><label>Table 1</label><caption><p>Summary of fractions selected for mass spectrometry analyses.</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td align=\"center\">Fraction ID**<break/>(Fraction Coordinates)</td><td align=\"center\" colspan=\"2\">Patient 1</td><td align=\"center\" colspan=\"2\">Patient 2</td><td align=\"center\" colspan=\"2\">Patient 3</td></tr><tr><td/><td colspan=\"6\"><hr/></td></tr><tr><td/><td align=\"center\" colspan=\"6\">GA Immunization</td></tr><tr><td/><td colspan=\"6\"><hr/></td></tr><tr><td/><td align=\"center\">Before<break/><bold>(B)</bold></td><td align=\"center\">After<break/><bold>(A)</bold></td><td align=\"center\">Before<break/><bold>(B)</bold></td><td align=\"center\">After<break/><bold>(A)</bold></td><td align=\"center\">Before<break/><bold>(B)</bold></td><td align=\"center\">After<break/><bold>(A)</bold></td></tr></thead><tbody><tr><td align=\"center\">F1<break/>(pH: 5.90-5.60, RT:16.04–16.71)</td><td align=\"center\">1.49*</td><td align=\"center\">1.11</td><td align=\"center\">1.76</td><td align=\"center\">1.49</td><td align=\"center\">1.41</td><td align=\"center\">1.46</td></tr><tr><td align=\"center\">F2<break/>(pH: 7.10-6.80, RT:14.03–14.70)</td><td align=\"center\">0.004</td><td align=\"center\">0.006</td><td align=\"center\">0.023</td><td align=\"center\">0.009</td><td align=\"center\">0.008</td><td align=\"center\">0.009</td></tr><tr><td align=\"center\">F3<break/>(pH: 6.20-5.90, RT:14.70–15.37)</td><td align=\"center\">0.043</td><td align=\"center\">0.049</td><td align=\"center\">0.06</td><td align=\"center\">0.049</td><td align=\"center\">0.03</td><td align=\"center\">0.031</td></tr><tr><td align=\"center\">F4<break/>(pH: 6.50-6.20, RT:16.04–16.71)</td><td align=\"center\">0.28</td><td align=\"center\">0.343</td><td align=\"center\">0.57</td><td align=\"center\">0.533</td><td align=\"center\">0.23</td><td align=\"center\">0.11</td></tr><tr><td align=\"center\">F5<break/>(pH: 5.90-5.60, RT:16.71–17.38)</td><td align=\"center\">0.178</td><td align=\"center\">0.171</td><td align=\"center\">0.507</td><td align=\"center\">0.24</td><td align=\"center\">0.134</td><td align=\"center\">0.133</td></tr><tr><td align=\"center\">F6<break/>(pH: 6.20-5.90, RT:15.37–16.04)</td><td align=\"center\">0.01</td><td align=\"center\">0.02</td><td align=\"center\">0.1</td><td align=\"center\">0.09</td><td align=\"center\">0.06</td><td align=\"center\">0.03</td></tr><tr><td align=\"center\">F7<break/>(pH: 4.99-4.88, RT:14.70–15.37)</td><td align=\"center\">0.24</td><td align=\"center\">0.49</td><td align=\"center\">0.48</td><td align=\"center\">0.805</td><td align=\"center\">0.51</td><td align=\"center\">0.46</td></tr><tr><td align=\"center\">F8<break/>(pH: 6.80-6.50, RT:15.37–16.04)</td><td align=\"center\">0.002</td><td align=\"center\">0.016</td><td align=\"center\">0.054</td><td align=\"center\">0.04</td><td align=\"center\">0.01</td><td align=\"center\">0.005</td></tr><tr><td align=\"center\">F9<break/>(pH: 6.80-6.50, RT:16.04–16.71)</td><td align=\"center\">0.11</td><td align=\"center\">0.03</td><td align=\"center\">0.045</td><td align=\"center\">0.065</td><td align=\"center\">0.02</td><td align=\"center\">0.01</td></tr><tr><td align=\"center\">F10<break/>(pH: 6.80-6.50, RT:14.03–14.70)</td><td align=\"center\">0.005</td><td align=\"center\">0.006</td><td align=\"center\">0.024</td><td align=\"center\">0.010</td><td align=\"center\">0.0086</td><td align=\"center\">0.009</td></tr></tbody></table></table-wrap>", "<table-wrap position=\"float\" id=\"T2\"><label>Table 2</label><caption><p>Summary of results from mass spectrometry analyses.</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td align=\"center\"><bold>Protein No</bold></td><td align=\"left\"><bold>Protein</bold></td><td align=\"left\"><bold>Fraction ID: Patient Sample</bold></td><td align=\"left\"><bold>MW</bold></td><td align=\"left\"><bold>NCBI Accession #</bold></td><td align=\"left\"><bold>pI</bold></td></tr></thead><tbody><tr><td align=\"center\">1</td><td align=\"left\">Hemopexin</td><td align=\"left\">F1: 1A,2A,3B,3A<break/>F8: 1B,1A,2B,2A,3B<break/>F6: 1B,1A,2B,2A,3B,3A<break/>F5: 1B,1A,2B,2A,3B,3A<break/>F7: 1B,1A,2B,3B</td><td align=\"left\">51676</td><td align=\"left\">11321561</td><td align=\"left\">6.55</td></tr><tr><td align=\"center\">2</td><td align=\"left\">Plasminogen</td><td align=\"left\">F1: 1A,2A,3B,3A<break/>F4: 1A,2A,3B,3A<break/>F6: 2^a</td><td align=\"left\">90569</td><td align=\"left\">4505881</td><td align=\"left\">7.04</td></tr><tr><td align=\"center\">3</td><td align=\"left\">Anti-thrombin</td><td align=\"left\">F1: 2A,3B</td><td align=\"left\">13788</td><td align=\"left\">23978644</td><td align=\"left\">5.91</td></tr><tr><td align=\"center\">4</td><td align=\"left\">Histidine-rich glycoprotein precursor</td><td align=\"left\">F1: 1B,1A,3B<break/>F4: 1B,1A,2B,2A,3B,3A<break/>F9: 2B</td><td align=\"left\">59578</td><td align=\"left\">4504489</td><td align=\"left\">7.09</td></tr><tr><td align=\"center\">5</td><td align=\"left\">Complement Factor I</td><td align=\"left\">F5: 1A,2B,2°</td><td align=\"left\">65720</td><td align=\"left\">119392081</td><td align=\"left\">7.72</td></tr><tr><td align=\"center\">6</td><td align=\"left\">Complex of the Catalytic Domain of Human Plasmin and Streptokinase</td><td align=\"left\">F1: 1A,2A,3B<break/>F4: 1A,2B,2A,3B,3A<break/>F9: 2B</td><td align=\"left\">27286</td><td align=\"left\">5821850</td><td align=\"left\">8.27</td></tr><tr><td align=\"center\">7</td><td align=\"left\">Prealbumin</td><td align=\"left\">F1: 2A<break/>F5: 1B,2B,2A,3B</td><td align=\"left\">15919</td><td align=\"left\">219978</td><td align=\"left\">5.52</td></tr><tr><td align=\"center\">8</td><td align=\"left\">A Chain A, prealbumin</td><td align=\"left\">F1: 2A<break/>F8: 1B<break/>F5: 1B,2B,2A,3B</td><td align=\"left\">13760</td><td align=\"left\">230651</td><td align=\"left\">5.55</td></tr><tr><td align=\"center\">9</td><td align=\"left\">Beta-2 glycoprotein I apolipoprotein H</td><td align=\"left\">F1: 1A,3B<break/>F4: 1A,2A<break/>F6: 1A,2B,2A,3B</td><td align=\"left\">38312</td><td align=\"left\">28810</td><td align=\"left\">8.34</td></tr><tr><td align=\"center\">10</td><td align=\"left\">Gelsolin isoform a precursor</td><td align=\"left\">F1: 3B<break/>F9: 2B<break/>F6: 2B<break/>F5: 1B,3B,3A</td><td align=\"left\">85697</td><td align=\"left\">4504165</td><td align=\"left\">5.90</td></tr><tr><td align=\"center\">11</td><td align=\"left\">Kringle 2 Domain of Human Plasminogen</td><td align=\"left\">F4: 1A,2A,3B</td><td align=\"left\">9637</td><td align=\"left\">6573460</td><td align=\"left\">7.55</td></tr><tr><td align=\"center\">12</td><td align=\"left\">Complement factor H-related protein 1 precursor (FHR-1)</td><td align=\"left\">F6: 1A,2B,2A,3B<break/>F7: 2A</td><td align=\"left\">37661</td><td align=\"left\">543981</td><td align=\"left\">7.75</td></tr><tr><td align=\"center\">13</td><td align=\"left\">H factor (complement)-like 3</td><td align=\"left\">F6: 2A</td><td align=\"left\">30651</td><td align=\"left\">5031695</td><td align=\"left\">6.00</td></tr><tr><td align=\"center\">14</td><td align=\"left\">Retinol-binding protein 4, plasma precursor</td><td align=\"left\">F1: 2B<break/>F5: 2B</td><td align=\"left\">23010</td><td align=\"left\">55743122</td><td align=\"left\">5.76</td></tr><tr><td align=\"center\">15</td><td align=\"left\">Coagulation factor XII-Mie</td><td align=\"left\">F5: 2A</td><td align=\"left\">67735</td><td align=\"left\">24899162</td><td align=\"left\">8.03</td></tr><tr><td align=\"center\">16</td><td align=\"left\">Glutathione peroxidase 3 precursor</td><td align=\"left\">F5: 2A</td><td align=\"left\">25505</td><td align=\"left\">121672</td><td align=\"left\">8.20</td></tr><tr><td align=\"center\">17</td><td align=\"left\">Factor H</td><td align=\"left\">F7: 2A</td><td align=\"left\">139125</td><td align=\"left\">31965</td><td align=\"left\">6.28</td></tr></tbody></table></table-wrap>" ]

[]

[]

[]

[]

[]

[]

[ "<table-wrap-foot><p>Second dimension fractions representing selected pH ranges and retention times (RT). Corresponding peak heights are listed as absorbance units at 214 nm (*). Although there may be several peaks/fraction, values represent the peak having the greatest absorbance in the corresponding fraction. Fraction IDs are used in Table 2 (**).</p></table-wrap-foot>", "<table-wrap-foot><p>Identification by LCMS/MS of proteins in plasma from three ALS patients before (B) and after (A) GA immunization. Only proteins with two or greater sequenced peptides are listed. Coordinates for Fraction IDs are listed in Table 1.</p></table-wrap-foot>" ]

[ "<graphic xlink:href=\"1477-5956-6-26-1\"/>", "<graphic xlink:href=\"1477-5956-6-26-2\"/>", "<graphic xlink:href=\"1477-5956-6-26-3\"/>", "<graphic xlink:href=\"1477-5956-6-26-4\"/>", "<graphic xlink:href=\"1477-5956-6-26-5\"/>" ]

[]

{ "acronym": [], "definition": [] }

[ "<title>Background</title>", "<p>Venezuelan equine encephalitis virus (VEEV) belongs to the Alphavirus genus within the Togaviridae family and was first isolated from horses in the end of the 1930s [##REF##17818578##1##,##REF##11581380##2##]. These viruses have a natural transmission cycle between rodents and mosquitos [##REF##7968923##3##]. Millions of horses were affected by this arbovirus with a fatality rate of up to 80% in epidemics in Central and South America [##UREF##0##4##].</p>", "<p>Several species of this family are pathogenic to humans and are recognized as potential biological warfare agent (BWA) [##REF##11544355##5##]. VEEV is classified as Bioterrorism Agent Category B by the center of Disease Control (CDC). Alphaviruses do not only have the potential for illness and transmission, but they can also be produced in large quantities and are moderately easy to disseminate. Furthermore, these virus species have the capacity to cause human epidemics [##REF##9261393##6##, ####REF##8709783##7##, ##REF##6438552##8##, ##REF##235212##9##, ##UREF##1##10##, ##REF##9086137##11####9086137##11##]. VEEV causes disease symptoms ranging from mild febrile reactions to fatal encephalitic zoonoses. Outcomes are significantly worse for young and elderly patients, with case fatalities ranging from 4 to 35% [##REF##11791799##12##,##REF##15916288##13##]. These viruses are highly infectious as aerosols [##REF##14999604##14##,##UREF##2##15##] and an intentional release of sufficient quantities as inhalable small-particle aerosol is expected to infect a high percentage of individuals within an area of a least 10,000 km²[##UREF##3##16##]. They can replicate in cell culture to very high titers and are relatively stable to environmental influences [##UREF##4##17##].</p>", "<p>For the surveillance of possible bioterrorism targets and endangered populations, rapid detection and diagnosis of VEEV are of crucial importance. In the past, the generation of monoclonal murine antibodies has improved the fast identification of VEEV infections to locate human and equine outbreaks of encephalitis. On the other hand, monospecific diagnostic monoclonal antibodies (mAbs) against VEEV are either hardly available on the market or too expensive for extensive use. In view of these current limitations the generation of specific high affinity recombinant antibodies may significantly improve the current situation and can make the rapid immunological detection widely available.</p>", "<p>A promising method to generate recombinant antibodies against human pathogenic viruses like VEEV is the antibody phage display technology. Using antibody phage display, genotype and phenotype of an antibody fragment are linked by fusing the antibody gene fragment to the minor coat protein III gene of the filamentous bacteriophage M13. The resulting antibody fragment::pIII fusion protein is displayed on the surface of the phage particles [##REF##2247164##18##, ####UREF##5##19##, ##REF##1907718##20##, ##REF##1908075##21####1908075##21##]. The most common antibody formats used for this technology are the Fragment antigen binding (Fab) and the single chain Fragment variable (scFv). In comparison to the Fab, that is consisting of the Fragment determining (Fd) of the heavy chain and the light chain linked by a disulphide bond, the scFv simply consists of the variable region of the heavy chain (V_H) and the variable region of the light chain (V_L), connected by a short peptide linker [##REF##3285471##22##,##REF##3045807##23##]. The selection of antibody fragments from antibody gene libraries is performed by an <italic>in vitro </italic>selection process [##REF##16126351##24##,##REF##18314587##25##], that is also referred to as \"panning\".</p>", "<p>In this study, we demonstrated the selection of human antibody fragments from a naïve antibody gene library specific for the detection of VEEV. We describe their immunological properties and discuss their possible application of these antibodies for diagnosis and detection of VEEV after a potential bioterrorism assault or natural outbreak of VEEV.</p>" ]

[ "<title>Methods</title>", "<title>Cell culture and virus production</title>", "<p>Alphaviruses were grown in Vero cells (VERO-B4, African green monkey kidney cells, DSMZ-ACC 33, Deutsche Sammlung von Mikroorganismen und Zellkulturen GmbH, Braunschweig, Germany) in biosafety level 2 and 3 facilities according to standard procedures [##REF##1847149##35##]. Virus titers were determined by the 50 % tissue culture infective dose (TCID₅₀/mL) method [##UREF##6##44##,##UREF##7##45##]. All viruses used in this study represent models for biowarfare agent relevant Alphavirus species and are either part of the strain collection of the Armed Forces Scientific Institute for Protection Technologies – NBC Protection (WIS) or were received from the National Collection of Pathogenic Viruses (NCPV), UK. The viruses used in this study were VEEV strain TC83 (variety 1AB), VEEV strain 12/93, Eastern equine encephalitis virus (EEEV) strain H178/99, Western equine encephalitis virus (WEEV) strain H160/99 and Chikungunya virus (CHIKV) strain S27. The strain TC83 was obtained from the Trinidad donkey strain by serial passages on guinea pig embryo heart cells in 1960 [##UREF##8##46##]. Additionally, VEEV strain 230 antigen (inactivated by β-propiolactone) was purchased from Senova GmbH (Jena, Germany). Strain VEEV 230 is the former USSR vaccine strain. Its history of production is not exactly known but it has been produced by serial passages of a virulent natural strain on chick embryos [##REF##1914828##47##]. If not particularily indicated, active virus material was used throughout the study. Lysates of VEEV infected cells were prepared by incubation with 4 M urea for 15 min at room temperature (RT).</p>", "<title>Purification of Alphaviruses</title>", "<p>Virus containing supernatants from infected Vero cells were either purified by affinity chromatography on Matrex Cellufine Sulfate Medium™ (<italic>Virus Recovery System</italic>, VRS, Chisso America Inc., NY, USA) or by isopycnic density gradient centrifugation. Matrex Cellufine Sulfate Medium™ (VRS) is a cellulose bead medium functionalized with a low concentration of sulfate esters that operates similar to a cation-exchange resin and has a high affinity to enveloped viruses. It selectively adsorbs complete virus particles as well as viral coats according to their charge. Briefly, 50 mL resin was equilibrated with adsorption buffer (0.01 M phosphate buffer, pH 7.5). Up to 200 mL of virus containing prefiltered cell culture supernatant was loaded onto the column which then was washed twice with 0.01 M phosphate buffer, pH 7.5. Elution of virus particles was performed with 1 M NaCl.</p>", "<p>Virus particles were pre-purified using ultracentrifugation through the sucrose cushion method (20% sucrose cushion), which causes low mechanical stress and allows the concentration and collection of morphologically intact particles after centrifugation at 112,000 × g for 2 to 3 hours. The pellet was resuspended in 0.5 to 1 mL phosphate buffered saline (PBS; [##UREF##9##48##]) and further purified by isopycnic density gradient centrifugation (20 to 60 % sucrose) for 18 hours at 217,500 × g. The virus containing fraction was removed, stored at -80°C until subjected to further analysis.</p>", "<title>Selection of recombinant antibodies</title>", "<p>The panning procedure based on protocols by Hust et al. [##REF##18314587##25##] with numerous modifications in 96 well microtitre plates (Maxisorb, Nunc, Wiesbaden, Germany). The mAb 8747 (Chemicon, Temecula, USA; [##REF##11572636##42##]) and mAb VEE-WIS1 (WIS, Munster, Germany) were incubated in concentrations of 1,5 μg/mL each overnight at 4°C in microtitre wells, followed by blocking with 1% (w/v) BSA in PBST (phosphate buffered saline + 1% Tween 20; [##UREF##9##48##]) for 1 h at RT. For every panning round one well was coated for the selection and one well was coated for a preselection step. For the preselection step 50 μL VRS concentrated supernatant from non-infected Vero cells + 50 μL 1% BSA in PBST was incubated. Afterwards, the wells were blocked with 2% skim milk powder in PBST. After 2 h at RT the wells were washed three times with PBST. In parallel, for selection, 50 μL VRS purified VEEV (2.7 mg/mL) + 50 μL 1% BSA in PBST was captured for 1 h by gently shaking at RT, followed by overnight incubation at 4°C.</p>", "<p>The human naïve HAL4/7 antibody gene library [##UREF##10##49##] consisting of in 5 × 10⁹independent clones in total based on the phagemid vector pHAL14 [##UREF##10##49##,##UREF##11##50##] was used for panning. The library was packaged using Hyperphage [##REF##11135557##51##, ####REF##16989094##52##, ##REF##16996161##53####16996161##53##]. Prior to panning 5 × 10¹¹scFv phage particles of HAL4 (kappa V_Lrepertoire) and 5 × 10¹¹scFv phage particles of HAL7 (lambda V_Lrepertoire) were mixed with 150 μL „panningblock\" solution (1% (w/v) BSA +1% (w/v) skim milk in PBST). In the preselection step, the library phage suspension was incubated at RT for 2.5 h in the well with captured VRS concentrated supernatant from non-infected Vero cells to remove non-specific binders. The supernatant containing the depleted library was mixed with 1/10 volume of VRS concentrated supernatant from non-infected Vero cells and 5 μg of a non VEEV-specific murine IgG for competition. For the selection step, the library solution was incubated in the wells with the immobilised VEEV at RT for 2 h followed by 30 times washing with PBST. Afterwards the bound scFv phage particles were eluted with 200 μL trypsin solution (10 μg/mL trypsin in PBS) at 37°C for 30 min. The supernatant containing the eluted scFv phage was transferred into a new tube. For the inactivation of the VEEV particles, 100 μL of 0.1 M glycin buffer pH 2.2 were added and incubated at RT for 15 min. The solution was neutralized with 100 μL 0.1 M phosphate buffer pH 7.6. 10 μL of eluted scFv phage were used for titration as described by Hust et al. [##REF##18314587##25##]. The remaining scFv phage were amplified as described by Hust et al. [##REF##18314587##25##] and used for the next panning round. The second panning round using the amplified phage was performed with the following modifications: the amount of antigen was reduced by 50% and washing cycles during panning were increased to 60. Additionally, in the third panning round the amount of antibody phage was reduced to 1 × 10⁹scFv phage. Furthermore, cell culture supernatant from non-infected Vero cells, 1/10 volume, was used for competition.</p>", "<title>Antigen ELISA using scFv phage, scFvs and scFv-Fc fusions</title>", "<p>All ELISAs were performed in 96 microtitre well plates (Maxisorb™, Nunc) that were coated with VRS purified viral antigen overnight at 4°C. Afterwards the wells were washed three times with PBST and blocked with 2% (w/v) skim milk powder in PBST (M-PBST) or with 1% fetal calf serum (FCS) in PBST for 1.5 h at RT, followed by three washes with PBST. ScFv phage, soluble antibody fragments or scFv-Fc fusion proteins were diluted in 100 μL M-PBST and incubated with the antigen for 1.5 h, followed by five washes with PBST. Bound scFv phage were detected by using the mAb anti-M13 conjugated with horseradish peroxidase (HRP) (GE Healthcare, München, Germany; 1:5.000). Bound soluble antibody fragments were detected by using the murine mAb 9E10 which recognizes the c-terminal c-myc tag. Staining was performed with a goat anti-mouse Ab conjugated to HRP (Sigma; 1:10.000). The specific binding of scFv-Fc fusion proteins to viral antigen was assessed with a goat anti-human Fc specific mAb conjugated to HRP (Sigma; 1:20.000), biotinylated mAbs were detected by using a Streptavidin HRP conjugate (GE Healthcare; 1:4.000). The visualization was performed with TMB (3,3',5,5'-tetramethylbenzidine) as substrate and the staining reaction was stopped by adding 100 μl 1 M sulphuric acid. Absorbance at 450 nm was measured by using a SUNRISE™ microtiter plate reader (Tecan, Crailsheim, Germany).</p>", "<title>Production of soluble antibody fragments in microtitre plate wells</title>", "<p>Microtitre plate wells containing 200 μL 2xTY + 100 mM glucose + 100 μg/mL ampicillin (2xTY-GA) were inoculated with single <italic>E. coli </italic>colonies from the phage titration of the panning and incubated overnight at 37°C and with constant shaking at 1200 rpm. 200 μL 2xTY-GA was inoculated with 10 μL of the overnight culture and grown at 37°C and 1200 rpm for 2 h. Bacteria were harvested by centrifugation for 10 min at 3220 × g. The pellets were resuspended in 200 μl 2xTY + 100 μg/mL Ampicillin + 50 μM isopropyl-beta-D-thiogalacto-pyranoside (IPTG), a substance that induces the prokaryotic <italic>lacZ </italic>promotor in <italic>E. coli</italic>, and incubated at 30°C and 1200 rpm overnight. Cells were removed from the scFv containing supernatant by centrifugation for 10 min at 3220 × g and 4°C.</p>", "<title>Sequencing</title>", "<p>Sequencing was performed using ABI Prism 310 Genetic Analyzer according to the manufacturers instructions using oligonucleotide primer MKpelB_f (5' GCCTACGGCAGCCGCTGG 3') or MKmyc_r (5' GATCCTCTTCTGAGATGAG 3'). The antibody gene fragments were analyzed by using VBASE2 <ext-link ext-link-type=\"uri\" xlink:href=\"http://www.vbase2.org\"/>[##REF##15608286##54##,##UREF##12##55##].</p>", "<title>SDS-PAGE and immunoblot analysis</title>", "<p>In order to analyse the scFv presentation on phage, the SDS-PAGE (sodium dodecyl sulphate-polyacrylamid gel electrophoresis) of scFv phage followed by Western blot and immunostaining of pIII were performed as described by Kirsch et al. [##REF##15992816##56##].</p>", "<p>ScFv-Fc fusion proteins and scFv phage were used to detect VEEV proteins. VEEV particles were separated by SDS-PAGE and blotted onto PVDF membrane. The membrane was blocked with M-PBST for 1 h at RT. ScFv phage or the scFv-Fc fusion protein were incubated for 1.5 h at RT, followed by two times washing with PBST. For the detection of bound scFv phage mAb mouse anti-M13 conjugated with HRP (GE Healthcare, 1:4000) was used for detection and visualized by DAB (diaminobenzidine). For the detection of the scFv-Fc fusion protein goat anti-human (Fc specific) conjugated with alcaline phosphatase (AP) (Dianova, Hamburg, Germany, 1:5000) was used. Murine IgGs were detected using goat anti-mouse (Fc specific) conjugated with AP (Sigma, 1:10000) and visualised by NBT/BCIP.</p>", "<title>Cloning and production of scFv-Fc fusion proteins</title>", "<p>VEEV-specific scFv gene fragments were subcloned from the library vector pHAL14 into the mammalian expression vector pCMV-hIgG1-Fc-XP (Schirrmann, manuscript in preparation) between the murine IgG signal peptide and the human IgG1 gene fragments by using the restriction sites <italic>Nco</italic>I and <italic>Not</italic>I. For the transient production of VEEV-specific scFv-Fc fusion proteins, the human embryonic kidney (HEK) cell line, 293T (American Type Culture Collection, ATCC, Rockwell, MD, No. CRL-11268) was transiently transfected by using the cell line specific lipid transfection reagent HEKfectin (Bio-Rad, München, Germany). 5 × 10⁵HEK293T cells were seeded and cultivated overnight into six well tissue culture plates (Sarstedt, Nürnbrecht, Germany) with 2 mL Dulbecco's Modified Eagle Medium (DMEM), supplemented with 2 mM L-glutamine, 1.5 g/L sodium bicarbonate and 4.5 g/L glucose, 8% (v/v) fetal calf serum (FCS) and 1% (w/v) penicillin/streptomycin (PAA, Parsing, Austria) at 37°C in 7% CO₂atmosphere and at 95% humidity. For the transfection 2.5 μg of plasmid DNA, encoding the scFv-Fc gene construct and 10 μL HEKfectin were preincubated in DMEM before the DNA-liposome complexes were then added to the HEK cells. The cells were incubated overnight with the transfection mixture before the medium was exchanged to fresh one on the following day. After 72 hours culture supernatants containing the scFv-Fc fusion proteins were harvested. The scFv-Fc fusion protein content of the collected supernatants was analyzed using a human IgG capture ELISA as previously described [##REF##17161420##57##].</p>" ]

[ "<title>Results</title>", "<title>Selection of recombinant antibodies against VEEV from a human naïve antibody library</title>", "<p>In order to generate antibody fragments reactive to members of the VEE virus serocomplex the human naive scFv antibody gene library HAL4/7 was used. All pannings were performed in a biosafety level 3 laboratory and the vaccine strain, TC83, as a medically important and epizootic Alphavirus species was used as antigen.</p>", "<p>The phage library was subjected to 3 rounds of panning and representative phage clones were assessed for their ability to bind VEEV TC83 immobilized onto 96 microwell plates. In order to exclude the enrichment of false-positive phage, the binding to supernatant of non-infected Vero cells (VRS concentrated and unconcentrated) was determined. Furthermore, the non-specific bindings of scFv phage to the VEEV-specific capture antibodies mAb 8747 and mAb VEE-WIS without and with an non-specific antigen, like lysozyme, was examined. As shown in figure ##FIG##0##1## a significant enrichment of VEEV-specific polyclonal antibody phage occurred after the third panning round. However, besides the specific accumulation of binders also a severe co-enrichment of antibodies to the capture antibodies was observed.</p>", "<p>Single clones were isolated from the third panning round. Soluble scFvs were produced in microtitre plates and analyzed by antigen ELISA on immobilized inactivated VEEV particles. The ELISA analysis using soluble scFvs instead of scFv phage minimized the occurance of false positives, because some antibody fragments bind only as antibody phage particles. Inactivated VEEV particles were used to ensure that the antibodies selected on active VEEV particles bound inactivated virus, too. In initial tests, we observed the enrichment of antibodies binding to Vero cell culture components. This effect was enhanced if VEEV particles were directly coated onto the wells and not captured by antibodies. Therefore, VRS purified Vero cell culture supernatant from non virus infected cells was used, because proteins from the cell culture supernatant were also enriched by the VRS system. In total, 230 antibody clones were analyzed by antigen ELISA (data not shown). Due to the signal to noise ratio, 26 VEEV binding scFv clones were further subjected to BstNI fingerprinting to sort out clones with identical restriction pattern (data not shown). After DNA sequencing, 10 different scFvs were finally found from this panning (designated with CHN24-x and MK269-x). One additional scFv clone (MK271-G2) was isolated by a slightly different panning strategy. Interestingly, only antibodies with lambda light chains were obtained. According to the integrative database of germ-line variable genes from the immunoglobulin loci of human (VBASE2) the isolated scFv fragments contained the antigen-binding variable domains of the light chains LV1, 2, 3 and 6. The heavy chains of the isolated scFvs belonged to the subfamily HV1, 3 and 4, while HV1 predominated (table ##TAB##0##1##).</p>", "<title>Production and characterization of VEEV-specific scFv phage</title>", "<p>In order to prove the presentation of functional scFvs on the selected monoclonal scFv phage clones, the clones were subjected to immunoblot analysis. ScFv phage were separated by SDS-PAGE under reducing conditions and the corresponding immunoblot was stained using an anti-pIII mAb. All scFv phage preparations showed a nearly equal and efficient display of scFv antibodies on their surface. This straightly allows to compare the cognate ELISA, Western blot and immunohistochemistry results. The anti-pIII immunoblot of a selection of anti-VEEV scFv phage is shown in figure ##FIG##1##2##.</p>", "<title>Verification of the VEEV-specific immunoreaction with scFv phage and scFv-Fc fusions</title>", "<p>In order to evaluate whether the antibody format or design influences the specific binding capacity, ELISA results obtained with selected scFv phage (figure ##FIG##2##3A##) and their corresponding scFv-Fc fusions (figure ##FIG##2##3B##) were compared. Purified formalin inactivated VEEV TC83 antigen was immobilized onto microwells and serial dilutions of either anti-VEEV scFv phage or serial dilution of scFv-Fc fusion proteins were used for detection. All selected scFv phage clones and the corresponding scFv-Fc fusions were able to bind directly immobilized VEEV particles (figure ##FIG##2##3##). The background binding of the control antibody IIB6 scFv phage increased when using very high scFv phage particle concentrations. A scFv phage concentration of about 1 × 10⁹– 5 × 10⁹scFv phage particles, respectively 10–100 ng/mL scFv-Fc fusion proteins are well suited for the detection of immobilized VEEV particles. Additionally, it was also possible to detect direct immobilised active VEEV TC83 particles by ELISA using scFv phage (figure ##FIG##3##4A##), respectively scFv-Fc fusion proteins (figure ##FIG##3##4B##).</p>", "<p>SDS-PAGE and Western blotting are valuable approaches to examine which VEEV structural proteins are recognized by the selected anti-VEEV antibody fragments. Since the viral glycoproteins E1 and E2 can be separated from each other under non-reducing conditions, virus samples were first disintegrated by incubation for 20 minutes at 56°C in Laemmli sample buffer containing no 2-mercaptoethanol. The samples were separated by 10% SDS-PAGE, blotted onto a PVDF membrane and stained as described. In general 5 × 10¹⁰anti-VEEV scFv phage/mL were used for the specific detection of structural proteins (Fig. ##FIG##2##3A##). The E2 protein specific antibodies mAb 8747 (Chemicon, CA, USA) and mAb 8/6 (Greiser-Wilke et al., 1989) served as positive control and displayed the expected electrophoretic profile typical for the Alphavirus E2 protein (46,9 kDa) and the cognate viral E1/E2 heterodimer (94,8 kDa). Interestingly, under non reducing conditions most of the anti-VEEV scFv phage displayed a nearly similar binding pattern and were able to bind either the E1 or E2 glycoprotein and the corresponding heterodimer (figure ##FIG##4##5A##). However, several of the specific scFv phage also caused an undefined smear if used in immunoblot analysis. This might be explained by prolonged staining. If the corresponding scFv-Fc fusions were used for binding, clear and defined bands, representing either E1 or E2 protein, were detectable similar to the E2 glycoprotein positive controls in figure ##FIG##4##5B##. In contrast, if virus samples were either prepared under reducing conditions or boiled prior to SDS-PAGE, no specific binding was observed. Therefore, the epitopes recognized by the scFv phage and scFv-Fc fusions are likely to be conformation dependent and the secondary structure seems to be critical for the binding of viral structural proteins. None of the isolated scFv fragments identified any linear epitopes.</p>", "<title>Evaluation of the cross-reactivity with different Alphavirus species and subspecies</title>", "<p>In order to test the cross-reactivity of the selected antibody clones with other strains of the VEEV as well as with other antigenic complexes, their binding was evaluated in a sandwich antigen catch ELISA by using an Alphavirus specific mAb mixture for capturing and the selected scFv phage for detection.</p>", "<p>An established VEEV-specific (figure ##FIG##5##6A##) and Alphavirus genus-specific sandwich ELISA (figure ##FIG##5##6B##) served as positive control. As negative control, cell culture of non-infected Vero cells was used. As marker antibody the biotinylated anti-VEEV mAb 8/6 was used for the detection of all VEEV strains (figure ##FIG##5##6A##) and a biotinylated mixture of antibodies consisting of mAb 8/6, mAb VEE-WIS1, mAb 12/2 and mAb 42/2 was used for the group specific detection of Alphaviruses (figure ##FIG##3##4B##). All viral antigens were captured by either the VEEV-specific mAb VEEV-WIS1 or a mAb mixture of anti-Alphavirus antibodies, consisting of mAb 3/4, mAb 12/2 and mAb VEE-WIS1 (WIS, Munster, Germany). Some virus strains (VEE-230) were captured better than others (VEE-H12/93).</p>", "<p>All VEEV antigens were employed with a nearly similar TCID₅₀/mL of 3 × 10⁸to 1 × 10⁹. In addition to the VEEV vaccine strain TC83 of subtype IAB, the USSR (Russian) vaccine strain VEEV 230 and the British NCPV strain VEEV 12/93 were applied. Furthermore, the selected scFv clones were tested for the specific detection of Eastern equine encephalitis virus (EEEV) strain H178/99, Western equine encephalitis virus (WEEV) strain H160/99 and Chikungunya virus (CHIKV) strain S27 Petersfield.</p>", "<p>Positive ELISA signals were obtained for the different VEEV strains with all tested scFv phage clones (figure ##FIG##5##6C##). In contrast, the scFv CHN-24-2A1 showed a comparable low antigen binding. Maximum binding in the antigen sandwich ELISA was found for the Russian strain VEEV 230. This might be explained by the fact that this virus sample was chemically inactivated prior to use. We suppose that dependent on the inactivation the critical epitopes are more accessible for antibody detection.</p>", "<p>However, none of the selected anti-VEEV scFv phage showed any cross-reactivity with other Alphaviruses, when used as detection antibody. All positive controls exhibited the expected binding pattern and were captured and detected by their specific mAbs in the ELISA.</p>", "<title>Detection of VEEV antigen in lysates of infected Vero cells</title>", "<p>In order to examine the broad immunological applicability of the selected scFvs, we also tested the recombinant antibody fragments for the specific detection of VEEV TC83 in lysates of infected Vero cells. These cell lysates were prepared by disrupting infected cells with 4 M urea while coupled to microwells. Detection was performed with scFv phage followed by an incubation with mAb anti-M13 conjugated to HRP. Lysates of non-infected Vero cells and VEEV antigen incubated with the mAb II-B6 served as negative control.</p>", "<p>Specific binding could be confirmed for nearly all selected antibody fragments except for the clones CHN24-2-A1 and MK269-E11. The most stringent binding results were obtained with the scFv clones CHN24-2-A2, CHN24-2-C3, CHN24-2-F11 and MK271-G2 (Fig. ##FIG##6##7##).</p>", "<p>In addition, detection of VEEV-specific antigen by immunohistochemistry in TC83 infected and formaldehyde fixed Vero cells was possible. Similar to the results described above, all scFv clones, except for clone CHN24-2-A1, showed a specific cytoplasmic immunostaining of VEEV infected Vero cells (data not shown).</p>" ]

[ "<title>Discussion</title>", "<p>For the detection of VEEV after an potential bioterrorims assault, e.g. by use of a VEEV aerosol, or a natural outbreak of VEEV, a fast diagnosis of these pathogen is necessary. The present work describes for the first time the screening and isolation of anti-VEEV antibody fragments from a human naïve antibody gene library by phage display. Ten out of eleven scFv clones were selected by the panning strategy using a mAb mixture for virus capturing as described. One further clone, MK271-G2, was isolated by an alternative panning, that was performed on directly immobilized viral antigen. The use of antibody captured virus particles was the preferred panning strategy because preliminary tests revealed, that panning on directly immobilized VEEV antigen enhanced especially the enrichment of non-specific binders.</p>", "<p>To our knowledge, there are no studies demonstrating the successful <italic>in vitro </italic>antibody selection against human pathogen complete virus particles using naïve antibody gene libraries. A successful panning against severe acute respiratory syndrome coronavirus using a human semisynthetic library is described by van den Brink et al. [##REF##15650189##26##]. In most other studies recombinant or purified virus proteins were used for panning if using a naïve antibody gene library [##REF##15381361##27##, ####REF##17266749##28##, ##REF##17336997##29####17336997##29##]. The pannings using complete particles are mostly performed using immune libraries, e.g de Carvalho et al. [##REF##11739690##30##], Koch et al. [##REF##12706090##31##] or Duan et al. [##REF##16518967##32##]. The panning procedure described here might be also useful for the <italic>in vitro </italic>antibody selection of scFvs against other viral targets from human naïve antibody gene libraries, in particular when either immunized patients are not available or immunisation is not ethically feasible.</p>", "<p>Nearly all scFvs were able to detect active as well as inactive VEEV TC83 viral antigen. Comparable indirect ELISA data were obtained with scFv phage and their corresponding scFv-Fc fusions. The specific immunoreaction could be verified by Western blot analysis, immunohistochemistry and by immunostaining of urea disrupted cell lysates. The selected antibody clones were reactive with all tested members of the VEE virus serocomplex but showed no significant cross-reactivity with closely related Alphavirus species like WEEV, EEEV and CHIKV, if used as detection molecules.</p>", "<p>All Alphaviruses share a number of structural, sequential and functional similarities. Immunological typing approaches categorize the nearly 30 species into seven serocomplexes or species. The nucleotide and amino acid identity among these antigenic complexes, subtypes and varieties varies from 45 to 96 % [##REF##11581380##2##,##REF##7968923##3##,##REF##2833129##33##,##REF##16847131##34##]. In general, the sequences of structural proteins are more divergent than the sequences of non-structural proteins. In immunoblot analysis the anti-VEEV scFv phage displayed a similar binding pattern like the E2 protein specific antibodies mAb 8747 (Chemicon, CA, USA) and mAb 8/6 [##REF##1847149##35##] and identified probably the E2 glycoprotein and the cognate viral E1/E2 heterodimer. Interestingly, all obtained scFv clones identified structural epitopes that are still folded after denaturation at 56°C under non-reducing conditions. In contrast, if the viral antigens were either prepared under reducing conditions or boiled prior to SDS-PAGE, no specific binding was observed.</p>", "<p>A possible neutralisation activity of the selected scFvs has to be assessed in further studies. Normally, the protective immunity to Alphaviruses is associated with an antibody reactivity to the virion glycoproteins E2 and so far, six conformationally stable epitopes were identified as critical for virus neutralization [##REF##6183343##36##, ####REF##2455383##37##, ##REF##7543231##38##, ##REF##16894184##39##, ##REF##2414905##40####2414905##40##]. Furthermore our antibodies are fully human and therefore better suited for applications like as passive vaccination than murine antibodies.</p>", "<p>To date, VEEV diagnosis is performed using monoclonal and polyclonal antibodies [##REF##15964967##41##] and also scFv fragments have been analysed [##REF##11572636##42##]. This study showed that scFv phage are applicable for a broad range of anti-VEEV diagnosis assays: antigen ELISA on purified virus particles, ELISA on cell lysate and immunoblot. Furthermore, the recombinant fragments offer the possibility to develop a VEEV-specific diagnosis assay since the specific scFv phage can be easily produced and purified in high amounts. This could be an alternative to fullsize IgGs for an ELISA assay. At least, they might be applied for immuno-PCR [##REF##16682441##43##] in order to increase the sensitivity of detection. These methods can be used for the diagnosis of VEEV in the enviroment and for the detection of human or equine VEEV infections.</p>" ]

[ "<title>Conclusion</title>", "<p>For the first time, this study describes the selection of antibodies against a human pathogenic virus from a human naïve scFv antibody gene library using complete, active virus particles as antigen. The described antibody selection procedure may also be useful for the <italic>in vitro </italic>antibody selection of antibody fragments against other viral targets from human naïve antibody gene libraries, in particular when immunized patients are not available or immunisation is not ethically feasible. The broad and sensitive applicability of anti-VEEV scFv-presenting phage for the immunological detection and diagnosis of Alphavirus species was demonstrated. The selected recombinant antibody fragments will improve the rapid and specific detection of VEEV infections after human and equine outbreaks of encephalitis, where an early and definite identification is of critical importance.</p>" ]

[ "<p>This is an Open Access article distributed under the terms of the Creative Commons Attribution License (<ext-link ext-link-type=\"uri\" xlink:href=\"http://creativecommons.org/licenses/by/2.0\"/>), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.</p>", "<title>Background</title>", "<p>Venezuelan equine encephalitis virus (VEEV) belongs to the Alphavirus group. Several species of this family are also pathogenic to humans and are recognized as potential agents of biological warfare and terrorism. The objective of this work was the generation of recombinant antibodies for the detection of VEEV after a potential bioterrorism assault or an natural outbreak of VEEV.</p>", "<title>Results</title>", "<p>In this work, human anti-VEEV single chain Fragments variable (scFv) were isolated for the first time from a human naïve antibody gene library using optimized selection processes. In total eleven different scFvs were identified and their immunological specificity was assessed. The specific detection of the VEEV strains TC83, H12/93 and 230 by the selected antibody fragments was proved. Active as well as formalin inactivated virus particles were recognized by the selected antibody fragments which could be also used for Western blot analysis of VEEV proteins and immunohistochemistry of VEEV infected cells. The anti-VEEV scFv phage clones did not show any cross-reactivity with Alphavirus species of the Western equine encephalitis virus (WEEV) and Eastern equine encephalitis virus (EEEV) antigenic complex, nor did they react with Chikungunya virus (CHIKV), if they were used as detection reagent.</p>", "<title>Conclusion</title>", "<p>For the first time, this study describes the selection of antibodies against a human pathogenic virus from a human naïve scFv antibody gene library using complete, active virus particles as antigen. The broad and sensitive applicability of scFv-presenting phage for the immunological detection and diagnosis of Alphavirus species was demonstrated. The selected antibody fragments will improve the fast identification of VEEV in case of a biological warfare or terroristic attack or a natural outbreak.</p>" ]

[ "<title>Authors' contributions</title>", "<p>MIK, BH performed the most experiments and helped to draft the manuscript. CN and TR performed some of the experiments. In addition, BH participated in the design and coordination of the study. TS and HJM particpated in the design and coordination of the study and helped to draft the manuscript. MH drafted the manuscript, participated in the design and coordination of the study and performed some of the experiments. SD conceived the project and wrote the grant application, particpated in the design and coordination of the study and helped to draft the manuscript. All authors read and approved the final manuscript.</p>" ]

[ "<title>Acknowledgements</title>", "<p>We would like to thank Lars Toleikis for providing the mucin specific control antibody IIB6. We gratefully acknowledge the kind help of Svetlana Mollova, Ida Retter and Werner Müller for adapting VBASE2 to analysis of antibody V-gene sequences directly derived from antibody selection projects. We would like to thank Luzie Voss and Saskia Helmsing for technical assistance. We would like to thank Steven R. Talbot for corrections and carefully reading the manuscript. We gratefully acknowledge the financial support by the German ministry of defense (BMVg) and the financial support by the German ministry of education and research (BMBF, SMP \"Antibody Factory\" in the NGFN2 program).</p>" ]

[ "<fig position=\"float\" id=\"F1\"><label>Figure 1</label><caption><p>ELISA with 10¹⁰(cfu) polyclonal scFv phage of each panning round. Antigen (directly immobilized): 1 μg VRS purified VEEV particles (VEEV VRS), 1 μg VRS concentrated supernatant from non-infected Vero cells (Vero VRS), 1:2 diluted supernatant of non infected VERO cells (Vero SN), 0.5 μg of each anti-VEEV capture mAb 8747 and VEE-WIS1, 1 μg lysozyme. The bound scFv phage were detected using mAb anti-M13 conjugated with HRP (1:5000).</p></caption></fig>", "<fig position=\"float\" id=\"F2\"><label>Figure 2</label><caption><p>Immunoblot of anti-VEEV scFv phage. 1 × 10¹¹(cfu) scFv phage per lane were separated on a reducing 10% SDS-PAGE, followed by Western blot and detection of wildtype pIII or scFv::pII fusion using mouse mAb anti-pIII (1:2000) and goat anti-mouse HRP (1:5000). A selection of seven anti-VEEV scFv phage clones is shown.</p></caption></fig>", "<fig position=\"float\" id=\"F3\"><label>Figure 3</label><caption><p>ELISA on directly immobilized inactive VEEV particles. Antigen: 1 μg VRS purified VEEV particles. <bold>A</bold>. A dilutions series of scFv phage particles were used for VEEV detection. The scFv phage were detected using mAb anti-M13 conjugated with HRP (1:5000). The mean values of two ELISAs from two independent scFv phage productions are shown. <bold>B</bold>. A series of scFv-Fc fusion protein dilutions were used for VEEV detection. The scFv-Fc were detected using goat anti-human IgG Fc specific antibody conjugated with HRP (1:20000). IIB6 is a non VEEV-specific control scFv antibody.</p></caption></fig>", "<fig position=\"float\" id=\"F4\"><label>Figure 4</label><caption><p>ELISA on directly immobilized active VEEV particles. Antigen: 1 μg VRS purified VEEV particles or 1 μg VRS concentrated supernatant from non-infected Vero cells as control. <bold>A</bold>. 1 × 10⁹scFv phage per well were used for VEEV detection. The scFv phage were detected as described in figure 3. The mean values of two ELISAs from two independent scFv phage productions are shown. <bold>B</bold>. 10 ng per well (100 ng/mL) scFv-Fc were used for VEEV detection. The scFv-Fc were detected as described in figure 3.</p></caption></fig>", "<fig position=\"float\" id=\"F5\"><label>Figure 5</label><caption><p>Immunoblot analysis of VEEV antigen detected by scFv phage or scFv-Fc fusions. VEEV VRS preparations were prepared at 56°C under non-reducing conditions and separated by 10% SDS-PAGE. After Western blot membranes were cut in stripes corresponding to 5–6 μg VEEV proteins. <bold>A</bold>. Immunostain was performed with 5 × 10¹⁰(cfu) anti-VEEV scFv phage particles/mL, murine anti-VEEV mAbs 8/6 and 8747 (1:1000) and detected with mAb mouse anti-M13 HRP (1:4000), respectively goat anti-mouse IgG Fc specific HRP (1:10000). IIB6 is a non VEEV-specific control scFv phage. <bold>B</bold>. Additionally, VEEV VRS samples were prepared at 56°C under non-reducing or at 99°C under reducing conditions, respectively. Western blots were stained with 1 μg/mL scFv-Fc and murine anti-VEEV IgG (1:1000) and detected with goat anti-human (gamma chain specific) AP (1:5000), or goat-anti mouse (Fc specific) AP (1:10000), respectively. The marker bands were marked with a pencil.</p></caption></fig>", "<fig position=\"float\" id=\"F6\"><label>Figure 6</label><caption><p>Cross-reactivity of the anti-VEEV scFv clones and different anti-Alphavirus specific mAbs analyzed by ELISA. Antigens: VEEV strains TC83, 230 and H12/93 were captured by using anti-VEEV mAb VEE-WIS1 (3 μg/mL); Eastern equine encephalitis virus (EEE), Western equine encephalitis virus (WEE) and Chikungunya (CHIK) were captured by using an anti-Alphavirus mAb mix consisting of mAb 3/4, mAb 12/2 and mAb VEE-WIS1 (3 μg/mL); Culture supernatant of non-infected Vero cells was captured once by anti-VEEV mAb VEE-WIS1 (VERO VEEWIS1) or by a mAb mix consisting of mAb 3/4, mAb 12/2 and mAb VEE-WIS1 (VERO mAb mix). <bold>A</bold>. Staining with biotinylated anti-VEEV mAb 8/6 (1:10000) and streptavidin conjugated with HRP (1:4000). <bold>B</bold>. Staining with a biotinylated mixture of antibodies consisting of mAb 8/6 (1:10000), mAb VEE-WIS1 (1:10000), mAb 12/2 (1:5000) and mAb 42/2 (1:2000) followed by a streptavidin-HRP (1:4000) incubation. <bold>C</bold>. Staining with 1 × 10⁹(cfu) scFv phage per well was followed by an incubation with mAb anti-M13 conjugated with HRP (1:5000). The IIB6 scFv phage was used as negative control. The mean values of two ELISAs from two independent scFv phage productions are shown.</p></caption></fig>", "<fig position=\"float\" id=\"F7\"><label>Figure 7</label><caption><p>ELISA on VEEV infected cell lysate. Antigen: cell lysate from VEEV infected/non-infected Vero cells. VEEV was detected by using 1 × 10⁹(cfu) scFv phage per well followed by an incubation with mAb anti-M13 conjugated with HRP (1:5000). The IIB6 scFv phage was used as negative control. The mean values of two ELISAs from two independent scFv phage productions are shown.</p></caption></fig>" ]

[ "<table-wrap position=\"float\" id=\"T1\"><label>Table 1</label><caption><p>Anti-VEEV scFvs</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td align=\"center\">scFv clone</td><td align=\"center\" colspan=\"2\">VH</td><td colspan=\"1\"/><td align=\"center\" colspan=\"2\">VL</td></tr><tr><td/><td colspan=\"5\"><hr/></td></tr><tr><td/><td align=\"left\">HV</td><td align=\"left\">D</td><td align=\"left\">HJ</td><td align=\"left\">LV</td><td align=\"left\">LJ</td></tr></thead><tbody><tr><td align=\"left\">CHN24-2-A1</td><td align=\"left\">IGHV1-69*01</td><td align=\"left\">IGHD2-8*02</td><td align=\"left\">IGHJ5*02</td><td align=\"left\">IGLV3-1*01</td><td align=\"left\">IGLJ3*01</td></tr><tr><td align=\"left\">CHN24-2-A2</td><td align=\"left\">IGHV3-9*01</td><td align=\"left\">IGHD6-19*01</td><td align=\"left\">IGHJ3*02</td><td align=\"left\">IGLV3-1*01</td><td align=\"left\">IGLJ1*01</td></tr><tr><td align=\"left\">CHN24-2-B7</td><td align=\"left\">IGHV1-18*01</td><td align=\"left\">IGHD2-21*02</td><td align=\"left\">IGHJ3*02</td><td align=\"left\">IGLV2-14*04</td><td align=\"left\">IGLJ3*01</td></tr><tr><td align=\"left\">CHN24-2-C2</td><td align=\"left\">IGHV1-69*01</td><td align=\"left\">IGHD6-13*01</td><td align=\"left\">IGHJ3*02</td><td align=\"left\">IGLV3-21*01</td><td align=\"left\">IGLJ3*01</td></tr><tr><td align=\"left\">CHN24-2-C3</td><td align=\"left\">IGHV3-23*01</td><td align=\"left\">IGHD6-13*01</td><td align=\"left\">IGHJ6*03</td><td align=\"left\">IGLV3-1*01</td><td align=\"left\">IGLJ1*01</td></tr><tr><td align=\"left\">CHN24-2-D5</td><td align=\"left\">IGHV1-8*01</td><td align=\"left\">IGHD6-6*01inv</td><td align=\"left\">IGHJ6*02</td><td align=\"left\">IGLV2-14*04</td><td align=\"left\">IGLJ3*02</td></tr><tr><td align=\"left\">CHN24-2-F11</td><td align=\"left\">IGHV1-18*01</td><td align=\"left\">IGHD6-6*01</td><td align=\"left\">IGHJ4*02</td><td align=\"left\">IGLV6</td><td align=\"left\">IGLJ3*02</td></tr><tr><td align=\"left\">MK269-C10</td><td align=\"left\">IGHV3-30*04</td><td align=\"left\">IGHD5-5*01</td><td align=\"left\">IGHJ6*02</td><td align=\"left\">IGLV2-14*02</td><td align=\"left\">IGLJ1*01</td></tr><tr><td align=\"left\">MK269-E11</td><td align=\"left\">IGHV4-34*01</td><td align=\"left\">IGHD3-3*01</td><td align=\"left\">IGHJ4*02</td><td align=\"left\">IGLV1-51*02</td><td align=\"left\">IGLJ3*01</td></tr><tr><td align=\"left\">MK269-E12</td><td align=\"left\">IGHV4-4*02</td><td align=\"left\">IGHD2-21*01inv</td><td align=\"left\">IGHJ5*02</td><td align=\"left\">IGLV3-21*02</td><td align=\"left\">IGLJ3*01</td></tr><tr><td align=\"left\">MK271-G2</td><td align=\"left\">IGHV1-69*01</td><td align=\"left\">IGHD3-16*01</td><td align=\"left\">IGHJ6*02</td><td align=\"left\">IGLV3-21*02</td><td align=\"left\">IGLJ3*02</td></tr></tbody></table></table-wrap>" ]

[]

[]

[]

[]

[]

[]

[ "<table-wrap-foot><p>Tab. 1. Given are the names of the gene segments according to VBASE2. Abbreviations: HV: V (variable) gene segments of the heavy chain; D: D (diversity) gene segment; HJ: J (joining) gene segment of the heavy chain; LV: V gene segment of the light chain; LJ: J gene segment of the light chain.</p></table-wrap-foot>" ]

[ "<graphic xlink:href=\"1472-6750-8-66-1\"/>", "<graphic xlink:href=\"1472-6750-8-66-2\"/>", "<graphic xlink:href=\"1472-6750-8-66-3\"/>", "<graphic xlink:href=\"1472-6750-8-66-4\"/>", "<graphic xlink:href=\"1472-6750-8-66-5\"/>", "<graphic xlink:href=\"1472-6750-8-66-6\"/>", "<graphic xlink:href=\"1472-6750-8-66-7\"/>" ]

[]

{ "acronym": [], "definition": [] }

[ "<title>Background</title>", "<p>Asthma is chronic inflammatory disorder of the airways that induces a range of sub-clinical and clinical effects including but not limited to hyperresponsiveness, airflow limitation, and respiratory symptoms. Approximately 6.7% of adults and 8.5% of children in the United States are reported to suffer from asthma with the greatest prevalence among non-Hispanic black and Hispanic children under 18 years of age [##UREF##0##1##]. Triggers of asthma exacerbation are varied and include viral infections, certain animal allergens and criteria air pollutants, mites, environmental tobacco smoke (ETS), mold, chemical irritants, and exercise in cold air [##UREF##1##2##]. Reducing exposure to environmental agents shown to increase asthma symptoms or lead to asthma exacerbations is an important component of a strategy to manage asthma for individuals [##REF##17589609##3##].</p>", "<p>Numerous investigations have demonstrated that indoor air cleaning devices can reduce concentrations of asthma triggers in indoor air [##REF##16201657##4##, ####REF##10380780##5##, ##UREF##2##6##, ##REF##9577519##7##, ##REF##10195274##8##, ##REF##16268832##9##, ##REF##2048821##10####2048821##10##]. Some studies have reported associations between use of air cleaners and improvements in respiratory symptoms and breathing problems for children and adults with asthma or persistent allergic rhinitis [##REF##9192919##11##, ####REF##10452769##12##, ##REF##12534557##13##, ##UREF##3##14####3##14##]. However, the benefits of air cleaners for breathing problems have not always been reproducible [##UREF##3##14##, ####REF##3041876##15##, ##REF##9655716##16##, ##REF##7846339##17##, ##REF##12426250##18##, ##REF##3171006##19####3171006##19##]. An expert panel recently determined that the evidence offered by health studies is not sufficient to conclude that operation of indoor air cleaning devices alleviates asthma symptoms or improves pulmonary function [##UREF##3##14##,##REF##12426250##18##, ####REF##3171006##19##, ##UREF##4##20####4##20##].</p>", "<p>The heterogeneity in results of air cleaner intervention studies for asthma symptoms may reflect in part the limited efficacy of the portable air cleaners used to mitigate exposure to airborne asthma triggers. Portable air cleaners typically have flow rates of 170 – 340 cubic meters per hour (m³/hr) and removal efficiency for fine particle mass (PM_2.5) of only about 70% because of bypass around their high efficiency particle arrestance (HEPA) filters [##UREF##5##21##]. For a typical U.S. home size of 450 m³, a 180 m³/hr portable device has a theoretical particle removal rate of approximately 0.4 per hour (hr^-1), about the same as the air exchange rate for a closed home. Air flow rates through room filters must be equivalent to several air changes per hour in order to achieve substantial control of airborne particulate matter [##REF##16201657##4##]. In contrast, whole house, high efficiency air cleaning systems that can provide clean air delivery rates up to 10 times greater than a portable air cleaner and particle removal rates of approximately 7 per hour are now available for residences [##UREF##6##22##]. The mitigation of asthma triggers in indoor air by these systems and potential health benefits for sensitive subpopulations have yet to be evaluated.</p>", "<p>To address this knowledge gap, we used an indoor air quality modeling system to examine peak and time-integrated concentrations of fungal spores, environmental tobacco smoke, respiratory viruses, and cat allergen in indoor air associated with natural ventilation, portable air cleaners, and forced air ventilation equipped with conventional and high efficiency filtration systems. As part of the modeling, we simulated several conditions that correspond to asthma management guidance published by the American Lung Association and the National Institutes of Health.</p>" ]

[ "<title>Methods</title>", "<p>We used the CONTAM multi-zone indoor air quality model developed by the National Institute of Standards and Technology (NIST) to estimate indoor concentrations of indoor allergens and irritants associated with asthma [##UREF##7##23##]. Airflow among indoor and outdoor zones of the building (i.e. rooms and ambient air) in CONTAM occurs via flow paths such as doors, windows, and cracks. Inter-zonal flow is based on the empirical power law relationship between airflow and the pressure difference across a flow path. Simulation of a mechanical ventilation system in CONTAM also induces circulation of air in CONTAM. After airflow among zones is established, mass balance equations are used to calculate pollutant concentrations based on the sources and sinks in each zone. Each zone (i.e. rooms, hallways) is treated as a single node wherein the air has uniform, well-mixed conditions throughout. Performance evaluations of CONTAM have demonstrated that the model simulations of inter-zonal flow and air exchange rate are within 15% on average of corresponding values measured in a single-family home and test home, respectively [##UREF##8##24##, ####UREF##9##25##, ##UREF##10##26##, ##UREF##11##27##, ##REF##15116834##28##, ##UREF##12##29####12##29##].</p>", "<p>Our analysis included two residential building templates developed by NIST, a two story detached home and a single story detached home. Single family detached homes represent over 60% of the total housing stock in the U.S. [##UREF##13##30##]. The floor areas for the single story and two story-building templates are 180 square meters (m²) and 276 m²respectively. See Additional files ##SUPPL##0##1## and ##SUPPL##1##2## for floor plans of the templates. The templates were based on the U.S. Census Bureau American Housing Survey [##UREF##14##31##] and the U.S. Department of Energy Residential Energy Consumption Survey [##UREF##15##32##] and were intended to represent typical U.S. residential building stock [##UREF##16##33##]. We modified the NIST templates to allow for natural ventilation and leakage through and around windows sized to 11.5% of the area of each wall [##UREF##17##34##].</p>", "<p>Both residential templates were modeled with six different ventilation and filtration configurations (See Table ##TAB##0##1##). The first configuration was a home with natural ventilation (N) and no capacity for indoor air cleaning. The remaining configurations each employ a central forced air heating and cooling system with differing degrees of filtration including: a standard 1 inch media filter (C), a standard 5-inch media filter (C5), the 1-inch filter with one portable HEPA unit in a bedroom (C+1P), the 1-inch media filter with a portable HEPA unit in the bedroom and one in the living/family room (C+2P), and a high efficiency electrostatic air cleaner with HEPA-like removal efficiency for aerosols (HE).</p>", "<p>Homes with central systems were assumed to have air-handling units (AHU) balanced to provide 0.18 m³/min/m²(0.6 cfm/ft²) of air to each room in the house. The duty schedule during heating and cooling periods was simulated with 1-hour resolution based on output from representative runs of the EnergyPlus Energy Simulation Software [##UREF##18##35##]. In general, the fraction of each hour devoted to forced air heating or cooling was proportional to the difference between ambient temperature and a set point of 22°C (72°F). Hourly duty schedules ranged from 4 minutes per hour during temperate periods to 38 minutes per hour during extreme summer periods and 52 min during extreme winter periods. In simulations with the C1 and C5 filters, a conventional AHU that operated only during periods of heating or cooling demand was used. In the simulations with the high efficiency electrostatic air cleaner, we modeled a modern AHU equipped with a variable speed fan that operates at full speed during periods of heating and cooling demand and at half-speed during all other times. Portable air cleaners were modeled as operating at 118 m³/hr for 24 hours per day. For the single story home, the return air duct AHU was located in the living room, for the two story home, there was a return in the hallway of both the first and second story. An air supply diffuser was located in each room of both housing templates.</p>", "<p>For simulations of central forced air systems, removal efficiencies for in-duct air cleaners were based on particle size-specific results observed in our prior assessment of in-duct air cleaning technologies conducted in a fully instrumented test home [##UREF##6##22##]. In that work, we found that the removal efficiency of a polydisperse test dust achieved by in-duct devices (specifically, 1-inch, 5-inch, and high efficiency electrostatic) was approximately 10% lower than the rated efficiencies determined according to ASHRAE Method 52.2, an industry standard performance metric [##UREF##19##36##]. Through diagnostic testing, we determined that the difference between the rated and in-use performance was the result of bypass where 10% of the airflow through the AHU fan entered the AHU cabinet downstream of the filter bay.</p>", "<p>For the portable air cleaners, removal efficiencies were based on studies conducted for the National Center of Energy Management and Building Technologies [##UREF##5##21##]. Similar to the whole house testing, Chen et al. found approximately 30% leakage in portable units and that none of the portable air cleaners reached HEPA-like filtration.</p>", "<p>Meteorological information is used by CONTAM to simulate force convection, radiant leakage, and corresponding air exchange rates. We used year 2005 meteorological data, including hourly wind direction and speed, dry and wet bulb temperature, relative humidity, and cloud cover data, obtained from the National Weather Service for the Cincinnati, Ohio area (Cincinnati/Northern Kentucky International Airport). We chose this area because Cincinnati has four distinct seasons and differences in ventilation are expected to vary by climatic conditions.</p>", "<p>Using a temperature-based probabilistic approach based on data from an EPA analysis [##UREF##20##37##], window and door opening schedules were generated that produced total ventilation rates for centrally and naturally ventilated periods consistent with corresponding air exchange rates determined from field campaigns reported elsewhere [##REF##16675415##38##, ####UREF##21##39##, ##UREF##22##40####22##40##]. During periods in which the windows were open, 40% of the total window area was assumed to be open. The AHU duty schedule and the window schedules were linked so that the AHU was never running when the windows were open. The front door was set to a schedule of opening for 15 minutes five times each day. Particle-size specific deposition rates to indoor surfaces were based on research by Thatcher and colleagues [##UREF##23##41##]. Due to limitations of the model, deposition rates were assumed independent of air exchange rate and the AHU duty schedule.</p>", "<p>A set of indoor allergens and irritants that can play a significant role in triggering asthma attacks was the focus of our analysis. Generation rates and particle size distributions of the contaminants were based on experimental data available in the published literature. Details regarding inputs to the model for the allergens and irritants are presented in Table ##TAB##1##2##.</p>", "<title>Cat Allergen</title>", "<p>Emission rates for cat allergen were based on studies that characterized the occurrence, suspension, and removal of cat allergen, <italic>Fel d 1</italic>, inside homes [##REF##9577519##7##,##REF##10435476##42##,##REF##9574885##43##]. Based on findings from those studies, we chose to model generation of cat allergen with a constant and intermittent source. The constant source was used to represent <italic>Fel d 1 </italic>levels in air during quiescent periods. The intermittent source represented resuspension of cat allergen caused by certain activities such as vacuuming or sitting on a couch [##REF##2048821##10##,##UREF##24##44##]. The intermittent source released a burst of allergen once an hour during typical waking hours, 7:00 AM – 10:00 PM. The constant generation source was located in all rooms of the house other than the bedrooms, while the burst source was released only in the main living space (i.e. living room for template 72 and family room for template 28). We omitted release of cat allergen in bedrooms in order to evaluate the extent to which allergen avoidance achieved by restricting cats from bedrooms as recommended by the NIH (2007), may be influenced by the use and efficacy of indoor air cleaning systems.</p>", "<p>Aerosols that contain cat allergen range in aerodynamic diameter from less than 0.4 micrometers (μm) to greater than 9 μm [##REF##9577519##7##,##REF##10380780##45##]. Previous research has demonstrated removal of airborne cat allergen by portable air cleaners with HEPA filters [##REF##9577519##7##]. For the electrostatic air cleaner, we assumed that the removal efficiency of cat allergen was equivalent to the particle-size specific performance observed for standard test dust and described elsewhere [##UREF##6##22##].</p>", "<title>Environmental Tobacco Smoke</title>", "<p>Particle size information and emission rates for ETS were based on information reported from studies of cigarette smoke in experimental chambers [##UREF##25##46##]. The total particle mass released for each cigarette was 8.3 mg with a release rate of 1.3 mg/min. A recent national survey indicates that the average adult smoker in the United States consumes 15 cigarettes per day [##UREF##26##47##]. Taking into account waking hours spent at home [##REF##11477521##48##], we modeled ETS emissions as cigarette consumption within the home twice in the morning hours and six times in the evening hours. All cigarettes were assumed to be smoked in the main living space (i.e. living room for template 72 and family room for template 28). Particle size for ETS has been reported to range from 0.05 μm to 0.71 μm [##UREF##27##49##]. Removal efficiency for ETS is one component of the industry standard method for determining and rating the performance of indoor air cleaning technologies [##UREF##28##50##].</p>", "<title>Outdoor Fungi</title>", "<p>In contrast to the other asthma triggers that were modeled as indoor sources, we modeled indoor air concentrations of airborne fungi that result from penetration of mold spores in ambient air. To coincide with the meteorological data noted earlier, daily mold spore counts for February 14 to November 23, 2005 measured at the Hamilton County Environmental Services Office in Cincinnati were obtained from the Hamilton County Air Quality Management Division. The daily observations from Cincinnati are short-term samples collected with a Rotorod Sampler (Sampling Technology, Inc., Minnetonka, MN) and therefore do not reflect the temporal variability of spore concentrations that may occur over the course of each day. In the absence of more complete data, we assumed that concentrations within the day were constant for purposes of this analysis. The outdoor level of total fungal spores reported in the data for Cincinnati ranged between 32 and 7935 spores per cubic meter (spores/m³) with a geometric mean of 881 spores/m³. As expected, outdoor spore concentrations were highest in the summer and early fall months. The aerodynamic diameter size distribution for total spores is large, ranging from 1 to 40 μm. While the dominant fungal genera, <italic>Cladosporium</italic>, has a aerodynamic diameter slightly less than 2 μm [##UREF##27##49##], the other dominant types, basidiospores and ascospores have aerodynamic diameters on the order of 5 μm [##UREF##29##51##]. Fungal allergens are borne on spores larger than 2.5 μm as well as hyphael fragments and fragmented spores smaller than 2.5 μm. Because of the absence of information on fungal fragment levels in outdoor spore data for Cincinnati and the paucity of large spore types in the data, we established 2.5 μm as a reasonable central estimate of the aerodynamic diameter for fungi in this analysis.</p>", "<title>Respiratory Viruses</title>", "<p>We modeled the release of two respiratory viruses, influenza virus and rhinovirus, because they have been implicated as triggers of asthma exacerbations and essential information is available on their transmission [##REF##3039011##52##] and aerosol properties. While respiratory syncytial virus and other viruses have also been associated with asthma, a lack of key information on these organisms precluded their inclusion in this analysis. For our respiratory virus modeling we utilized the concept of infectious dose, referred to as quanta, as first described in 1955 [##UREF##30##53##] Estimates of quanta generation rates from an infectious person are based on analyses of outbreaks of infectious diseases as described elsewhere [##REF##12950586##54##,##REF##2653151##55##]. The greater the quanta generation rate the more infectious the organism. Estimates for influenza, a virus that can spread rapidly, are on the order of 15 to 128 quanta per hour [##REF##12950586##54##,##REF##16297217##56##]. Organisms with slower spreading infections, like rhinovirus and tuberculosis have generation rates on the order of 1 to 10 quanta per hour [##REF##12950586##54##]. For this analysis, we assumed the approximate mid-point of published quanta generation rates for influenza and rhinovirus, 67 q/hr and 5 q/hr, respectively.</p>", "<p>We also assumed the quanta were evenly distributed among the particles released during a sneeze. The removal processes are based on the particle sizes of the quanta released. We based the particle size distribution on experimental studies of particles emitted during sneezes and coughs conducted in the 1940s and 1960s [##REF##6018703##57##,##UREF##31##58##] and recently re-analyzed [##REF##15764538##59##]. To establish removal efficiency for respiratory virus achieved by the in-duct media filters, we relied upon size-specific results observed in our test home [##UREF##6##22##]. For the in-duct electrostatic air cleaner, the removal efficiency was based on laboratory studies in which a suspension of live influenza A virus, PR-8 strain (Advanced Biotechnology, Inc., MD) in phosphate buffered saline was aerosolized within a ventilation duct using a 6-jet Collison nebulizer. Aerosol samples were obtained on Teflo filters (Millipore Corporation, Bedford, MA) in triplicate upstream and downstream of the electrostatic air cleaner on three days. The samples were extracted and assayed for influenza by quantitative polymerase chain reaction (qPCR) following procedures described by Van Elden et al. [##REF##11136770##60##]. The average removal efficiency from the tests was greater than 99% with more precise quantitation limited by the sensitivity of the assay. The removal efficiencies obtained from the laboratory studies were coupled with AHU bypass information for use in the model. Details of this novel application of qPCR will be published elsewhere.</p>", "<p>Output from the IAQ model for respiratory virus was expressed as quanta per cubic meter (q/m³) of indoor air. We used a modified Wells-Riley equation [##REF##665658##61##] to estimate the risk of infection based on the concentration of quanta in the room from the model output coupled with conventional central estimates of exposure duration and a breathing rate of 0.48 m³/hr published in a widely used compilation of exposure factors [##UREF##32##62##]. We used the results to analyze the risk of infection for an individual when (1) spending time in the same room as an infectious individual, (2) spending time in an adjacent room, and (3) occupying other rooms in the house when an infected individual is either in a bedroom or in the living room of the home.</p>" ]

[ "<title>Results</title>", "<p>Air exchange rate (AER) is an influential determinant of indoor air quality and hence is a primary output from the CONTAM model. The distributions of 24-hour average AER across the year for the two templates with both natural and forced air ventilation systems are summarized in Table ##TAB##2##3##. The mean and median AER for the natural ventilation configuration were approximately twice those in the forced air configuration due to the increased use of windows during warm weather. AER was lower in the newer home (DH28) than the older home (DH72) which reflects differences in leakage rates between the two homes. With the exception of differences in AER, the modeling results were similar for the two home templates. Therefore, we chose to report only the results from the newer two-story home (DH28).</p>", "<title>Cat Allergen</title>", "<p>The distribution of hourly average concentrations for airborne cat allergen throughout the home for each of the six ventilation configurations is summarized in Figure ##FIG##0##1A##. When operating a high efficiency device, the median allergen concentration (4.0 ng/m³) was 46% lower when compared to conventional filtration (6.4 ng/m³). The next best performance was achieved by two systems – the in-duct 5-inch media filter (C5) and a portable air cleaner in the same room as the intermittent release of allergen (i.e. C+2P). Nominally, peak concentrations were best mitigated by the high efficiency in-duct device (86 ng/m³), although the difference in comparison to peaks associated with the other air cleaning approaches (approximately 100 ng/m³) may not be substantive relative to uncertainties in the modeling analysis. To evaluate the effectiveness of the ventilation configurations at limiting transfer of allergen to bedrooms, all airborne releases of cat allergen in our model occurred outside of the bedrooms. In the bedroom, allergen levels were lower than the whole house average for all configurations, with the high efficiency in-duct filtration performing best at minimizing the transfer of allergen into the bedroom (See Figure ##FIG##0##1B##).</p>", "<title>Environmental Tobacco Smoke</title>", "<p>Modeled whole house concentrations of ETS were strongly influenced by use of air cleaners as illustrated by the distribution of hourly average concentrations estimated across the year (Figure ##FIG##1##2##). The greatest mitigation of ETS was achieved by the high efficiency in-duct device (median &lt;0.01 μg/m³), followed by use of a portable air cleaner in the same room as the smoker (median 3.2 μg/m³), the pleated in-duct media filter (median 9.8 μg/m³), one portable air cleaner in a bedroom (median 17.8 μg/m³), and a conventional in-duct filter (median 29.9 μg/m³). Simulation of a home with natural ventilation yielded hourly average ETS concentrations that were similar to the C5 simulation, probably because of the higher AER throughout the year for a home without forced air conditioning.</p>", "<p>The effect of high efficiency in-duct filtration on peak and short-term time-weighted averaged levels of ETS is depicted in Figure ##FIG##2##3A## and ##FIG##2##3B## for a typical 24-hour period (February 1) that had eight smoking events in the living room. For a home with conventional in-duct filtration, each cigarette smoked is associated with a peak concentration of approximately 80 μg/m³and a subsequent exposure period of at least 8 hours when windows are closed. In contrast, peak ETS concentrations per cigarette during model runs with the high efficiency in-duct device were about 40 μg/m³. First-order removal rates for ETS calculated for the conventional and high efficiency in-duct filtration conditions were 0.008 min^-1and 0.049 min^-1. Use of the high efficiency in-duct device also substantially limited the distribution of the contaminant into other rooms of the home such as the bedroom.</p>", "<title>Outdoor Fungi</title>", "<p>The highest indoor/outdoor ratios for spore concentrations were in the summer and fall months, probably due to the higher AER associated with open windows during those seasons. When averaged over the period for which fungal spore data were available, the indoor/outdoor ratio was highest for the natural ventilation configuration and lowest for the in-duct high efficiency configuration (Table ##TAB##3##4##). Whole house indoor spore concentrations for the in-duct high efficiency configuration were less than one-half the levels in the conventional configuration and more than eight times lower than the mean outdoor level. Even in the bedroom where the portable air cleaner was located, the in-duct high efficiency achieved lower spore levels.</p>", "<title>Respiratory Viruses</title>", "<p>For the virus results, we limited the analysis to December through March to reflect the cold and flu season in the United States. The median AER for this period was 25% lower for the naturally ventilated configuration and essentially unchanged for the mechanically ventilated homes in comparison to the remainder of the year. For this period, we examined the extent to which the risk of infection by either influenza or rhinovirus is modified by the use of an air cleaner for three common scenarios where a healthy individual and infectious individual cohabitate.</p>", "<p>In the first scenario, a healthy individual, perhaps a caregiver, spends one hour in the bedroom of an individual infected with influenza. In this case, the use of a portable air cleaner in the room with the infectious individual limits the average risk of infection to less than one-half the risk when conventional filtration is used (Table ##TAB##4##5##). The high efficiency in-duct system provides the next lowest average risk of infection, followed by the conventional and pleated filter in-duct systems. The risk of infection is lowered for each of the in-duct and portable air cleaner configurations in comparison to natural ventilation</p>", "<p>In the second scenario, we evaluated the risk of infection for a person who spends 12 hours in a bedroom adjacent to a second bedroom occupied by an individual infected with influenza. This scenario is representative of many residential configurations including children who normally sleep in separate bedrooms or two children who normally share a bedroom but are separated temporarily when one of them has a chest cold. In this scenario, the risk of influenza infection for a 12-hour exposure for an occupant in the adjacent bedroom was approximately 16% with conventional filtration, 5% for the configurations with a portable air cleaner in the bedroom and 0.6% with the high efficiency filtration (See Table ##TAB##4##5##).</p>", "<p>For the third scenario, we estimated the risk of infection from an individual who remains in the home over the course of a five-day infectious period. We assumed that the infectious individual spent one-half of their time in the bedroom and the other half in the family room, while a healthy individual spent 69% of the corresponding time indoors at home [##REF##11477521##48##] during which they were exposed to the house-wide average concentration of quanta in air. For this scenario, the risk of infection by influenza was greater than 30% in the ventilation configuration with a portable air cleaner in both of the two rooms frequented by the infectious individual (Table ##TAB##4##5##). In comparison, the risk of infection was 17% for the natural ventilation configuration and less than 4% for the high efficiency in-duct system. The former probably reflects a relatively slow rate of inter-zonal transfer and the latter reflects the comparatively high flow rate and removal efficiency of the in-duct system.</p>" ]

[ "<title>Discussion</title>", "<p>Several studies have assessed the use of air cleaners for reducing indoor air concentrations of chemical and biological materials that exacerbate asthma. In these studies, the air cleaning intervention was typically a portable air cleaner sized for a single room of typical size in a residence. Although based on modeling rather than measurements, our analysis indicates that certain air cleaning configurations can mitigate indoor air concentrations of some common asthma triggers more effectively on average than air cleaning achieved by the type of portable filtration devices evaluated previously as well as by conventional in-duct filtration.</p>", "<p>Prior performance evaluations of CONTAM demonstrate that the model provides a reasonable degree of accuracy for the types of indoor air quality simulations upon which our analyses rely. Inter-zonal airflow predictions from CONTAM simulations of a single story home were within 15% of corresponding measured values [##UREF##12##29##]. Similarly, air exchange rates for a single room building predicted with CONTAM were within 5% of measured levels [##UREF##8##24##]. In a related analysis, the correlation between predicted and observed concentrations of a conservative gas ranged from 0.95 to 0.998 during six tests within a single room test home [##UREF##9##25##]. In a tracer gas study conducted in a multi-room occupied townhouse, gas concentrations predicted by the model were within 25% of measured concentrations [##UREF##10##26##]. Finally, measured and predicted 24-hour average concentrations of 0.3 to 5 μm particles in a single room building were within 30% of each other [##UREF##8##24##].</p>", "<p>Particle removal efficiencies for air cleaning systems considered in this analysis were derived from empirical data obtained from test homes or test chambers [##UREF##5##21##,##UREF##6##22##]. Removal efficiencies for the portable air cleaners were based on chamber studies of four different devices that all claimed to have HEPA filters but whose efficacy under controlled conditions was low compared to HEPA standards [##UREF##5##21##]. If we had assumed that the portable air cleaners had removal efficiencies approaching those of HEPA filters, those systems would have compared more favorably to the other devices for the rooms of the homes in which they were located. Whole house comparisons of portable and in-duct systems are unlikely to have been changed substantially if we had assumed a higher aerosol removal efficiency for the portable devices.</p>", "<p>In terms of controlling residence-wide concentrations of cat allergen, ETS, respiratory viruses, and mold spores in indoor air, use of a high efficiency in-duct air cleaner as part of a forced air ventilation system yielded the greatest benefit, followed by multiple portable air cleaners in conjunction with conventional in-duct filtration. The greatest benefit of air cleaning systems over conventional in-duct filtration was observed for ETS, probably because of its sub-micron size distribution and the correspondingly low rate of deposition to surfaces. The extent to which these findings can be generalized to other constituents of indoor air depends upon their similarity in terms of emission profiles and aerodynamic characteristics. Other important indoor allergens such as dust mite and cockroach that have been shown to be associated with relatively particles are unlikely to be represented accurately by our results for cat allergen, ETS, viruses, and fungal spores.</p>", "<p>Consistent with results from our evaluation of air cleaners in a test home [##UREF##6##22##], the whole house performance of each system was directly related to its clean air delivery rate (CADR), the product of air flow rate and removal efficiency. This analysis focused on single family detached homes, however we anticipate that the findings are applicable to multi-family and attached homes as well. Various types of housing stock may differ systematically in terms of air exchange rate because of variation in construction practices, exterior surface area-to-volume ratios, and other factors. Particle deposition has been reported to be positively associated with air exchange rate due to increased turbulence of indoor air [##REF##11393992##63##,##UREF##33##64##]. Because of modeling constraints, we assumed that particle deposition rates were independent of air exchange rate. This simplifying assumption is unlikely to be a substantial contributor to uncertainty in our results because the range of turbulence-induced deposition rates reported for respirable-sized aerosols is small in comparison to differences in performance among air cleaning devices indicated by our analysis.</p>", "<p>In terms of controlling the contaminant concentrations in a single room, the location of the contaminant source is important. If the contaminant source was in the family room of the home and therefore near a central return, as was the case for the allergen and ETS modeling, the high efficiency in-duct filtration was superior to all configurations including those with a portable air cleaner in the room. Similarly, if the source is outdoors, as was the case with the fungal spore modeling, the high efficiency in-duct filtration was superior. Conversely, when the source was in a location away from a central return, like a bedroom, as was the case for the one-hour influenza scenario, operation of a portable air cleaner in the room was the most effective air cleaning configuration. We anticipate that these results for cat allergen, ETS, and virus are reasonably representative of emissions of other respirable-sized aerosols from indoor sources including fungal spores that may be released from surfaces as a result of mechanical forces.</p>", "<p>The utility of the modeling results presented is related primarily to relative differences between the air cleaning systems included in this assessment. If reasonable however, the absolute levels are also of interest for consideration of potential air quality and exposure benefits afforded by indoor air cleaning systems. To assess the accuracy of the model results, we compared the predicted concentrations to measurements from studies that quantified residential airborne levels of animal allergens [##REF##10380780##5##,##REF##9577519##7##,##REF##2048821##10##], ETS [##REF##16201657##4##,##UREF##34##65##], or fungal spores [##UREF##2##6##,##REF##18458748##66##]. Several of the studies evaluated the effectiveness of portable air cleaners with HEPA filters which allows us to compare our modeled results to not only the reported levels, but also to the changes in contaminant concentrations associated with use of portable air cleaners. Other studies were designed to measure typical residential contaminant concentrations, both with and without a source present. Data from those investigations provide a reasonable benchmark for our modeled results under typical ventilation configurations.</p>", "<p>The relative differences among the ventilation configurations that we considered are similar to the reductions observed in intervention studies designed to evaluate the effectiveness of portable air cleaners. In a study of dog allergen, airborne levels in two rooms with portable air cleaners were reduced to 25% of the baseline allergen level [##REF##10380780##5##]; similar to the difference in modeled cat allergen concentrations for the bedroom when the portable air cleaner was introduced. In a study of portable air cleaner efficacy in four homes with smokers, PM concentrations in the living room were reduced by 30–70% with the use of a portable air cleaner [##REF##16201657##4##]. Our modeling yielded similar reductions in ETS when comparing the conventional filtration to the ventilation configurations with portable air cleaners. In a study designed to evaluate the utility of portable air cleaners for controlling fungal spore concentrations, the intervention effectively reduced spore levels in a bedroom of a residence, however the air cleaner worked best when the bedroom door was closed [##UREF##2##6##].</p>", "<p>When comparing absolute levels of contaminants in the home, the modeled results for the conventional and natural ventilation configurations compare well with values reported in the literature. Our modeled cat allergen concentrations with conventional filtration are similar to concentrations reported in a study of 75 homes with cats in Britain [##REF##9577519##7##] and the levels during our intermittent release of allergen is similar to measured values during periods of disruptions such as vacuuming [##REF##2048821##10##,##UREF##35##67##]. In a study of homes in six U.S. cities, the authors calculated that smoking one pack of cigarettes daily contributed 20 μg/m³to the 24 average hour indoor particle concentration [##UREF##34##65##], which is similar to our modeled ETS concentrations for the conventional and natural ventilation configurations. For fungal spores, our modeled indoor/outdoor ratios for the conventional and natural ventilation configurations are similar to the ratio of 0.32 for total spores reported in a study conducted in six homes in the Cincinnati area [##REF##18458748##66##].</p>", "<p>Results from a controlled study of rhinovirus transmission provide a reasonable comparison for evaluating the accuracy of our modeled likelihood of infection. In the experimental study, groups of eight students with active rhinovirus infection spent 12-hours in a room with 12 susceptible students and followed a protocol designed to allow transmission of an infectious dose only by inhalation [##REF##3039011##52##]. The resulting risk of infection from this study was 61%. While AER or filtration characteristics were not reported for this study, we assumed that the room was either naturally ventilated or had conventional filtration. Our modeled scenario with one infectious individual in a room of approximately one-half the size of the experimental room resulted in an average risk of infection with influenza of 33.6% and 16.5% with natural and conventional filtration, respectively. If the modeling were conducted with four infectious individuals in the smaller bedroom to more closely mimic the conditions of the experimental study, the risk of modeled infection would rise to a level similar to that observed in the experimental study.</p>", "<p>We relied upon the concept of quanta generation to estimate the probability of acquiring an infection through the airborne route, using the Wells-Riley equation [##REF##665658##61##]. The Wells-Riley equation and modifications of the equation have been used by researchers to estimate the risk of airborne transmission of an infection for a variety of organisms including measles [##REF##665658##61##], influenza [##REF##12950586##54##], rhinovirus [##REF##12950586##54##,##REF##14754759##68##], severe acute respiratory syndrome (SARS) [##REF##17100668##69##], and tuberculosis [##REF##2653151##55##]. The Wells-Riley equation only estimates the risk of transmission for the inhalation route of exposure. Organisms like rhinovirus and influenza can be transmitted by other routes of exposure such as direct contact, although the relative importance of the respective routes of exposure is not well understood. The ability of various indoor air cleaning configurations to influence virus transmission through surface-mediated pathways remains to be determined. Consideration of virus transmission via surfaces and other pathways is unlikely to influence our findings for modification of the risk of infection through inhalation because of different ventilation and air cleaning configurations.</p>", "<p>While the Wells-Riley equation accounts for the ventilation rate of the indoor space of interest to calculate the quanta concentration, it does not, as discussed recently [##REF##18211478##70##], explicitly account for other removal processes such as deposition to surfaces, filtration, and loss of infectiousness in the air. However, quanta generation rates are typically based on disease outbreak data, and therefore inherently account for these processes. Our modeling accounted for deposition and filtration, but not loss of infectiousness. Some data suggest that virus die-off is a slow process that can occur over several days at temperature and humidity levels typical of indoor environments [##REF##2999318##71##]. Therefore, not explicitly controlling die-off is unlikely to influence our results substantively. Including removal mechanisms in our model along with the estimates of virus emissions in units of quanta may have resulted in double counting for removal by filtration and deposition. Therefore, our results may underestimate the actual risk of infection. To evaluate the impact of potential double counting for deposition, we conducted model runs without a deposition rate for virus. In these models, the risk of infection increased approximately 30 to 50% depending on the filtration type. Regardless, our analysis was designed to primarily evaluate the differences in ventilation configurations and the differences between these configurations would not be changed by under or over estimating the risk of infection.</p>", "<p>While a number of intervention studies clearly demonstrate exposure reductions attributable to the use of portable air cleaners, associated improvements in health have been more difficult to demonstrate. Some air cleaning interventions have yielded improvements in respiratory symptoms and breathing problems for children and adults with asthma or persistent allergic rhinitis [##REF##9192919##11##, ####REF##10452769##12##, ##REF##12534557##13####12534557##13##,##REF##2191991##72##]; however, the results of these studies have not always been reproducible [##UREF##3##14##, ####REF##3041876##15##, ##REF##9655716##16##, ##REF##7846339##17##, ##REF##12426250##18####12426250##18##,##REF##3171006##73##]. One explanation for the lack of reproducible results could be that portable air cleaners used in these studies have not effectively reduced personal exposure. Our modeling demonstrates that while the use of a portable air cleaner will provide exposure benefits in the room it is located, concentrations of common asthma triggers throughout the residence, and corresponding personal exposures, are not likely to be mitigated. Our modeling analysis indicate that high efficiency in-duct air cleaning systems would yield a more substantial reduction in personal exposure that the portable air cleaners used in intervention studies published to date. Potential benefits of these systems for personal exposure could be evaluated following methods employed in a study of personal exposure to cat allergen [##REF##12801310##74##].</p>", "<p>An Expert Panel convened by the NIH recommended asthmatics with pet allergies that are not willing to part with their pets keep the pet out of the asthmatic's bedroom as one part of an asthma management strategy. Additionally, the National Environmental Education &amp; Training Foundation (NEETF) recommends that the use of portable air cleaners in bedrooms of asthmatics [##UREF##36##75##]. While the use of portable air cleaners in the bedroom prove to be beneficial in our modeling, the results indicate that the use of high efficiency in-duct air cleaners provide an more effective means of controlling allergen levels not only in a single room, but the whole house.</p>" ]

[ "<title>Conclusion</title>", "<p>The modeling results from this study demonstrate that properly maintained forced air systems with a high efficiency aerosol removal system are expected to provide the best control of the indoor exposure to common asthma triggers such as cat allergen, ETS, fungal spores and respiratory viruses. The modeling results also showed that the potential efficacy of avoidance strategies recommended for asthmatics by the American Lung Association and the National Institutes of Health may be enhanced by the use of certain indoor air cleaning systems.</p>" ]

[ "<p>This is an Open Access article distributed under the terms of the Creative Commons Attribution License (<ext-link ext-link-type=\"uri\" xlink:href=\"http://creativecommons.org/licenses/by/2.0\"/>), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.</p>", "<title>Background</title>", "<p>Reducing exposure to environmental agents indoors shown to increase asthma symptoms or lead to asthma exacerbations is an important component of a strategy to manage asthma for individuals. Numerous investigations have demonstrated that portable air cleaning devices can reduce concentrations of asthma triggers in indoor air; however, their benefits for breathing problems have not always been reproducible. The potential exposure benefits of whole house high efficiency in-duct air cleaners for sensitive subpopulations have yet to be evaluated.</p>", "<title>Methods</title>", "<p>We used an indoor air quality modeling system (CONTAM) developed by NIST to examine peak and time-integrated concentrations of common asthma triggers present in indoor air over a year as a function of natural ventilation, portable air cleaners, and forced air ventilation equipped with conventional and high efficiency filtration systems. Emission rates for asthma triggers were based on experimental studies published in the scientific literature.</p>", "<title>Results</title>", "<p>Forced air systems with high efficiency filtration were found to provide the best control of asthma triggers: 30–55% lower cat allergen levels, 90–99% lower risk of respiratory infection through the inhalation route of exposure, 90–98% lower environmental tobacco smoke (ETS) levels, and 50–75% lower fungal spore levels than the other ventilation/filtration systems considered. These results indicate that the use of high efficiency in-duct air cleaners provide an effective means of controlling allergen levels not only in a single room, like a portable air cleaner, but the whole house.</p>", "<title>Conclusion</title>", "<p>These findings are useful for evaluating potential benefits of high efficiency in-duct filtration systems for controlling exposure to asthma triggers indoors and for the design of trials of environmental interventions intended to evaluate their utility in practice.</p>" ]

[ "<title>List of Abbreviations</title>", "<p>ETS: Environmental tobacco smoke; PM_2.5:Particles less than 2.5 microns; HEPA: High efficiency particle arrestance; NIST: National Institute of Standards and Technology; AHU: Air-handling units; AER: Air exchange rates; SARS: Severe acute respiratory syndrome; CADR: Clean air delivery rate; qPCR: Quantitative polymerase chain reaction. </p>", "<title>Competing interests</title>", "<p>Funding for this research was provided by Trane Residential Systems, Inc., Tyler, TX and Environmental Health &amp; Engineering, Inc., Needham, MA.</p>", "<title>Authors' contributions</title>", "<p>TAM conceived of the study, and participated in its design and coordination and helped to draft the manuscript. TM carried out the modeling efforts. JA carried out the data analysis. DLM participated in the design of the study and drafted the manuscript. All authors read and approved the final manuscript.</p>", "<title>Supplementary Material</title>" ]

[]

[ "<fig position=\"float\" id=\"F1\"><label>Figure 1</label><caption><p><bold>Comparison of Hourly Fel d 1 allergen concentrations by filtration configuration for (1A) the whole house average and (1B) bedroom 2</bold>.</p></caption></fig>", "<fig position=\"float\" id=\"F2\"><label>Figure 2</label><caption><p>Comparison of Hourly ETS concentrations by filtration configuration.</p></caption></fig>", "<fig position=\"float\" id=\"F3\"><label>Figure 3</label><caption><p>Comparison of 24-hour environmental tobacco smoke (ETS) concentrations in the living room and bedroom between the conventional filter (3A) and the high-efficiency filter (3B) for February 1.</p></caption></fig>" ]

[ "<table-wrap position=\"float\" id=\"T1\"><label>Table 1</label><caption><p>Ventilation/Filtration Configuration Information</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td align=\"left\">Abbreviation</td><td align=\"left\">Description</td></tr></thead><tbody><tr><td align=\"left\">N</td><td align=\"left\">Natural ventilation with no air cleaning capacity</td></tr><tr><td colspan=\"2\"><hr/></td></tr><tr><td align=\"center\" colspan=\"2\">Forced Air Systems</td></tr><tr><td colspan=\"2\"><hr/></td></tr><tr><td align=\"left\">C</td><td align=\"left\">Conventional 1-inch media filter (MERV 2)</td></tr><tr><td align=\"left\">C5</td><td align=\"left\">Standard 5 inch media filter. Based on Perfect Fit 5 inch media filter, Model BAYFTAH26M, Trane Residential Systems, Tyler, TX, USA (MERV 8)</td></tr><tr><td align=\"left\">HE</td><td align=\"left\">High Efficiency System – CleanEffects™ Model TFD235ALAH000AA, Trane Residential Systems, Tyler, TX, USA</td></tr><tr><td colspan=\"2\"><hr/></td></tr><tr><td align=\"center\" colspan=\"2\">Forced Air Systems plus Portable Air Cleaners</td></tr><tr><td colspan=\"2\"><hr/></td></tr><tr><td align=\"left\">C+1P</td><td align=\"left\">Conventional 1-inch filter plus portable HEPA filter devices. Flow characteristics based on Quiet Flo HEPA Air Purifier Model 20316, Hunter Fan Company, Memphis, TN, USA. Filtration capacity based on Chen et al. (2006).</td></tr><tr><td align=\"left\">C+2P</td><td align=\"left\">Conventional 1-inch filter plus 2 portable HEPA filters devices (See above)</td></tr></tbody></table></table-wrap>", "<table-wrap position=\"float\" id=\"T2\"><label>Table 2</label><caption><p>Model Inputs for Contaminant Emission Rates and Filtration Removal Efficiency Rates</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td align=\"center\">Contaminant<break/>/Particle size</td><td align=\"center\">Emission<break/> Rate</td><td align=\"center\">Deposition<break/> Rate (hr^-1)</td><td align=\"center\">1-inch<break/> (%)</td><td align=\"center\">5-inch<break/> (%)</td><td align=\"center\">High<break/>Efficiency<break/> (%)</td><td align=\"center\">Portable<break/> (%)</td></tr></thead><tbody><tr><td align=\"center\" colspan=\"7\">Cat Allergen^a</td></tr><tr><td colspan=\"7\"><hr/></td></tr><tr><td align=\"center\">0.54</td><td align=\"center\">0.0688 μg/hr</td><td align=\"center\">0.052</td><td align=\"center\">2.5</td><td align=\"center\">29.2</td><td align=\"center\">90.7</td><td align=\"center\">71</td></tr><tr><td align=\"center\">0.875</td><td align=\"center\">0.0688 μg/hr</td><td align=\"center\">0.15</td><td align=\"center\">2.5</td><td align=\"center\">29.2</td><td align=\"center\">90.7</td><td align=\"center\">71</td></tr><tr><td align=\"center\">1.6</td><td align=\"center\">0.1376 μg/hr</td><td align=\"center\">0.35</td><td align=\"center\">20.7</td><td align=\"center\">47</td><td align=\"center\">91.8</td><td align=\"center\">71</td></tr><tr><td align=\"center\">2.7</td><td align=\"center\">0.5502 μg/hr</td><td align=\"center\">1</td><td align=\"center\">20.7</td><td align=\"center\">47</td><td align=\"center\">91.8</td><td align=\"center\">71</td></tr><tr><td align=\"center\">4</td><td align=\"center\">1.8895 μg/hr</td><td align=\"center\">2.2</td><td align=\"center\">55.3</td><td align=\"center\">77.8</td><td align=\"center\">96.5</td><td align=\"center\">72</td></tr><tr><td align=\"center\">5.25</td><td align=\"center\">2.0953 μg/hr</td><td align=\"center\">3.5</td><td align=\"center\">55.3</td><td align=\"center\">77.8</td><td align=\"center\">96.5</td><td align=\"center\">80</td></tr><tr><td align=\"center\">7.4</td><td align=\"center\">5.5885 μg/hr</td><td align=\"center\">6.5</td><td align=\"center\">74.3</td><td align=\"center\">86.9</td><td align=\"center\">98.4</td><td align=\"center\">80</td></tr><tr><td align=\"center\">9</td><td align=\"center\">10.899 μg/hr</td><td align=\"center\">10</td><td align=\"center\">74.3</td><td align=\"center\">86.9</td><td align=\"center\">98.4</td><td align=\"center\">80</td></tr><tr><td colspan=\"7\"><hr/></td></tr><tr><td align=\"center\" colspan=\"7\">ETS^b</td></tr><tr><td colspan=\"7\"><hr/></td></tr><tr><td align=\"center\">0.0575</td><td align=\"center\">1.31 mg/cig</td><td align=\"center\">0.02</td><td align=\"center\">0</td><td align=\"center\">14.6</td><td align=\"center\">90.1</td><td align=\"center\">70</td></tr><tr><td align=\"center\">0.1475</td><td align=\"center\">2.84 mg/cig</td><td align=\"center\">0.005</td><td align=\"center\">0</td><td align=\"center\">14.6</td><td align=\"center\">90.1</td><td align=\"center\">70</td></tr><tr><td align=\"center\">0.31</td><td align=\"center\">2.84 mg/cig</td><td align=\"center\">0.018</td><td align=\"center\">0</td><td align=\"center\">14.6</td><td align=\"center\">90.1</td><td align=\"center\">70</td></tr><tr><td align=\"center\">0.71</td><td align=\"center\">1.31 mg/cig</td><td align=\"center\">0.08</td><td align=\"center\">2.5</td><td align=\"center\">29.2</td><td align=\"center\">90.7</td><td align=\"center\">71</td></tr><tr><td colspan=\"7\"><hr/></td></tr><tr><td align=\"center\" colspan=\"7\">Outdoor Fungal Spores</td></tr><tr><td colspan=\"7\"><hr/></td></tr><tr><td align=\"center\">2.5</td><td align=\"center\">NA</td><td align=\"center\">0.9</td><td align=\"center\">14</td><td align=\"center\">47</td><td align=\"center\">91.8</td><td align=\"center\">71</td></tr><tr><td colspan=\"7\"><hr/></td></tr><tr><td align=\"center\" colspan=\"7\">Virus^c</td></tr><tr><td colspan=\"7\"><hr/></td></tr><tr><td align=\"center\">2.1</td><td align=\"center\">35.3 q/hr</td><td align=\"center\">0.6</td><td align=\"center\">14</td><td align=\"center\">47</td><td align=\"center\">91.8</td><td align=\"center\">71</td></tr><tr><td align=\"center\">4.5</td><td align=\"center\">29.4 q/hr</td><td align=\"center\">2.8</td><td align=\"center\">55</td><td align=\"center\">77.8</td><td align=\"center\">96.5</td><td align=\"center\">72</td></tr><tr><td align=\"center\">7.3</td><td align=\"center\">1.8 q/hr</td><td align=\"center\">6.5</td><td align=\"center\">73</td><td align=\"center\">86.9</td><td align=\"center\">98.4</td><td align=\"center\">80</td></tr><tr><td align=\"center\">9.4</td><td align=\"center\">0.5 q/hr</td><td align=\"center\">10</td><td align=\"center\">74</td><td align=\"center\">86.9</td><td align=\"center\">98.4</td><td align=\"center\">80</td></tr></tbody></table></table-wrap>", "<table-wrap position=\"float\" id=\"T3\"><label>Table 3</label><caption><p>Distribution of simulated 24-hour average air exchange rates for homes with and without forced air ventilation systems.</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td align=\"left\">Ventilation/Filtration</td><td align=\"center\">House<break/> Template</td><td align=\"center\">Mean</td><td align=\"center\">Std<break/> Dev</td><td align=\"center\" colspan=\"5\">Percentiles</td></tr><tr><td/><td/><td/><td/><td colspan=\"5\"><hr/></td></tr><tr><td/><td/><td/><td/><td align=\"center\">5%</td><td align=\"center\">25%</td><td align=\"center\">50%</td><td align=\"center\">75%</td><td align=\"center\">95%</td></tr></thead><tbody><tr><td align=\"left\">Natural</td><td align=\"center\">DH28</td><td align=\"center\">3.7</td><td align=\"center\">5.0</td><td align=\"center\">0.1</td><td align=\"center\">0.2</td><td align=\"center\">0.2</td><td align=\"center\">6.8</td><td align=\"center\">13.0</td></tr><tr><td align=\"left\">Forced Air</td><td/><td align=\"center\">1.8</td><td align=\"center\">3.6</td><td align=\"center\">0.1</td><td align=\"center\">0.1</td><td align=\"center\">0.2</td><td align=\"center\">0.9</td><td align=\"center\">10.9</td></tr><tr><td align=\"left\">Natural</td><td align=\"center\">DH72</td><td align=\"center\">3.0</td><td align=\"center\">3.9</td><td align=\"center\">0.2</td><td align=\"center\">0.4</td><td align=\"center\">0.5</td><td align=\"center\">5.1</td><td align=\"center\">10.6</td></tr><tr><td align=\"left\">Forced Air</td><td/><td align=\"center\">1.6</td><td align=\"center\">2.9</td><td align=\"center\">0.1</td><td align=\"center\">0.3</td><td align=\"center\">0.4</td><td align=\"center\">0.7</td><td align=\"center\">8.7</td></tr></tbody></table></table-wrap>", "<table-wrap position=\"float\" id=\"T4\"><label>Table 4</label><caption><p>Geometric Mean (Geometric Standard Deviation) of Indoor/Outdoor Ratios and Indoor Spore Concentrations by Ventilation/Filtration Type</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td align=\"center\">Ventilation<break/> on/Filtration</td><td align=\"center\">I/O Ratio</td><td align=\"center\">Whole House<break/> (spores/m³)</td><td align=\"center\">Bedroom 2^a<break/>(spores/m³)</td></tr></thead><tbody><tr><td align=\"center\">N</td><td align=\"center\">0.34 (2.6)</td><td align=\"center\">303 (7.0)</td><td align=\"center\">238 (9.4)</td></tr><tr><td align=\"center\">C</td><td align=\"center\">0.16 (2.7)</td><td align=\"center\">141 (5.8)</td><td align=\"center\">131 (6.3)</td></tr><tr><td align=\"center\">C5</td><td align=\"center\">0.13 (3.1)</td><td align=\"center\">111 (6.7)</td><td align=\"center\">97 (8.0)</td></tr><tr><td align=\"center\">C+1P</td><td align=\"center\">0.14 (2.9)</td><td align=\"center\">128 (6.0)</td><td align=\"center\">54 (8.5)</td></tr><tr><td align=\"center\">C+2P</td><td align=\"center\">0.14 (2.9)</td><td align=\"center\">119 (6.2)</td><td align=\"center\">52 (8.8)</td></tr><tr><td align=\"center\">HE</td><td align=\"center\">0.07 (4.1)</td><td align=\"center\">57 (8.3)</td><td align=\"center\">41 (13.0)</td></tr></tbody></table></table-wrap>", "<table-wrap position=\"float\" id=\"T5\"><label>Table 5</label><caption><p>Mean (Standard deviation) percent risk of infection during three exposure scenarios</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td align=\"center\">Scenario</td><td align=\"center\">1</td><td align=\"center\">2</td><td align=\"center\">3</td><td/></tr></thead><tbody><tr><td align=\"left\">Ventilation<break/>/Filtration</td><td align=\"center\">Risk of influenza infection for a one hour exposure in the bedroom with individual infected with influenza</td><td align=\"center\">Risk of influenza infection from 12 hour exposure in adjacent bedroom</td><td align=\"center\">Risk of infection during 5 day infectious period while infected individual in bedroom for 1/2 the day and the family room for the 1/2 the day^a</td><td/></tr><tr><td/><td/><td/><td colspan=\"2\"><hr/></td></tr><tr><td/><td/><td/><td align=\"center\">Influenza</td><td align=\"center\">Rhinovirus</td></tr><tr><td colspan=\"5\"><hr/></td></tr><tr><td align=\"left\">N</td><td align=\"center\">36 (7.9)</td><td align=\"center\">0.6 (1.3)</td><td align=\"center\">17.1 (2.4)</td><td align=\"center\">1.4 (0.2)</td></tr><tr><td align=\"left\">C</td><td align=\"center\">18 (3.4)</td><td align=\"center\">16.1 (1.7)</td><td align=\"center\">70.0 (1.6)</td><td align=\"center\">8.6 (0.4)</td></tr><tr><td align=\"left\">C5</td><td align=\"center\">17 (3.4)</td><td align=\"center\">6.7 (1.0)</td><td align=\"center\">36.6 (1.8)</td><td align=\"center\">3.4 (0.2)</td></tr><tr><td align=\"left\">C+1P</td><td align=\"center\">7 (0.8)</td><td align=\"center\">5.9 (0.7)</td><td align=\"center\">51.9 (1.7)</td><td align=\"center\">5.3 (0.2)</td></tr><tr><td align=\"left\">C+2P</td><td align=\"center\">7 (0.8)</td><td align=\"center\">5.4 (0.6)</td><td align=\"center\">33.7 (2.2)</td><td align=\"center\">3.0 (0.2)</td></tr><tr><td align=\"left\">HE</td><td align=\"center\">13 (1.5)</td><td align=\"center\">0.6 (0.1)</td><td align=\"center\">3.9 (0.2)</td><td align=\"center\">0.3 (0.01)</td></tr></tbody></table></table-wrap>" ]

[]

[]

[]

[]

[]

[ "<supplementary-material content-type=\"local-data\" id=\"S1\"><caption><title>Additional file 1</title><p>Two-Story Home Floorplan (DH28).</p></caption></supplementary-material>", "<supplementary-material content-type=\"local-data\" id=\"S2\"><caption><title>Additional file 2</title><p>One-Story Home Floorplan (DH72).</p></caption></supplementary-material>" ]

[ "<table-wrap-foot><p>^aBetween the hours of 7 am – 10 pm, the cat allergen concentration increases for 33% from the intermittent allergen release. Emission rates based on Custovic et al. [##REF##9537780##76##].</p><p>^bA total of 8 cigarettes per day. Per cigarette emission rates (mg/cigarette) based on Klepeis et al. [##UREF##25##46##].</p><p>^cEmission rate of infectious doses (or quanta) per hour (q/hr) based on Liao et al. [##REF##16297217##56##].</p></table-wrap-foot>", "<table-wrap-foot><p>DH28 Two story detached home</p><p>DH72 Single story detached home</p></table-wrap-foot>", "<table-wrap-foot><p>^aBedroom 2 contains a portable air cleaner</p></table-wrap-foot>", "<table-wrap-foot><p>^aAssumes that that occupant is in the home 68.7% of the time based on Klepeis et al. [##REF##11477521##48##]</p></table-wrap-foot>" ]

[ "<graphic xlink:href=\"1476-069X-7-43-1\"/>", "<graphic xlink:href=\"1476-069X-7-43-2\"/>", "<graphic xlink:href=\"1476-069X-7-43-3\"/>" ]

[ "<media xlink:href=\"1476-069X-7-43-S1.pdf\" mimetype=\"application\" mime-subtype=\"pdf\"><caption><p>Click here for file</p></caption></media>", "<media xlink:href=\"1476-069X-7-43-S2.pdf\" mimetype=\"application\" mime-subtype=\"pdf\"><caption><p>Click here for file</p></caption></media>" ]

{ "acronym": [], "definition": [] }

[ "<title>Background</title>", "<p>Coronary artery disease is the main cause of death in middle-aged men [##REF##3202078##1##] and atherosclerosis is responsible for 50% of all mortality in the USA, Europe and Japan [##REF##3511384##2##]. There are several theories about the pathogenesis of atherosclerosis but the ‚response-to-injury‘-hypothesis of R. Ross [##REF##3511384##2##,##REF##8479518##3##] is widely accepted. Several different sources of injury to the endothelium lead to endothelial cell dysfunction. The initial stages of atherosclerosis are characterized by adhesion of circulating leukocytes to the endothelial cells and subsequent transendothelial migration. This process is mediated in part by cellular adhesion molecules (CAMs) like vascular cell adhesion molecule-1 (VCAM-1), intercellular adhesion molecule-1 (ICAM-1), and E-selectin, expressed on the endothelial membrane, in response to several inflammatory cytokines, including interleukin-1, tumor necrosis factor and interferon [##REF##7507411##4##]. Cellular adhesion molecules are later on detached from the membrane and are found in the circulation in their soluble form. CAM expression is increased in atherosclerotic plaques [##REF##7506307##5##,##REF##7688768##6##]. Soluble forms of these adhesion molecules (sVCAM-1, sICAM-1, sE-selectin) can be detected in the serum and are increased in conditions with an inflammatory component [##REF##7506035##7##,##REF##1676471##8##] and in atherosclerosis resulting in coronary artery disease [##REF##9439492##9##, ####REF##9416885##10##, ##UREF##0##11####0##11##], carotid sclerosis [##REF##9416885##10##] and peripheral vascular disease [##REF##9409238##12##]. Furthermore increased levels of sCAMs are associated with components of the metabolic syndrome like diabetes mellitus [##REF##1280239##13##], hypertension [##REF##9443768##14##] and dislipidemia [##REF##8641021##15##,##REF##9315520##16##]. A slight but significant independent correlation of sCAMs and fasting triglycerides has been observed [##REF##8641021##15##,##REF##9598830##17##]. The impact of postprandial metabolism on sCAMs has been shown during an oral glucose tolerance test, when postprandial insulin levels correlated directly with sICAM-1 [##REF##9568701##18##]. Another study compared 2 h and 4 h postprandial levels of sICAM-1 and sVCAM-1 after a high-fat versus a high-carbohydrate diet in diabetic and normal subjects [##REF##11923038##19##]. In this study, an increase of ICAM-1 and VCAM-1 occurred after high-fat meal in non-diabetic subjects and was prevented by addition of vitamine E and C. Two other studies showed that soluble adhesion molecules were increased after a high fat meal [##REF##14988255##20##,##REF##14632967##21##]. There are no previous reports on the relation between postprandial triglycerides, glucose, insulin and soluble adhesion molecules after a standardized lipid load in healthy subjects.</p>", "<p>The aim of the study was to evaluate soluble adhesion molecules sICAM-1, sVCAM-1 and sE-selectin after a mixed meal in young healthy subjects with normal or high triglyceride response to the meal.</p>" ]

[ "<title>Methods</title>", "<title>Subjects</title>", "<p>Thirty healthy male subjects with normal fasting triglycerides (&lt;150 mg/dl) were recruited according to their postprandial response to a standardized lipid-rich meal. In a previous study, a bimodal distribution of triglyceride maxima following an oral lipid load was observed. In total 15% of the subjects did respond with postprandial triglyceride maxima above 260 mg/dl. The cut-off point to identify the high responders (HR) was chosen according to a bimodal distribution of triglyceride maxima [##REF##8352452##26##,##REF##9329767##27##]. Fifteen normal responders (NR) and fifteen HR were selected out of 182 subjects who underwent the standardized lipid-rich meal in the current study. Inclusion criteria were normal BMI (&lt;25 kg/m²) and fasting blood glucose levels &lt; 110 mg/dl. Subjects had no history of present or past hypertension, hyperlipidemia, diabetes, or cardiovascular disease. All subjects were following previously ad libidum diets, had no recent change in body weight or intercurrent illness and were taking no medication. The study complies with the Declaration of Helsinki. The protocol of the study was approved by the ethics commitee of the University of Kiel, the subjects gave informed written consent before being tested. The characteristics of the study population are reported in table ##TAB##0##1##.</p>", "<title>Standardized lipid-rich meal</title>", "<p>Studies began at 8 AM after a 12-h-overnight fast. After fasting blood was drawn, the subjects consumed 500 ml of a standardized mixed liquid meal (oral metabolic tolerance test, oMTT) containining the following ingredients: 30 g of protein (11,9 energy%), 75 g of carbohydrates (29,6 energy%)(93% saccharose, 7% lactose), 58 g of fat (51,6 energy%)(65% saturated, 35% unsaturated fatty acids), 10 g of alcohol (6,9 energy%), 600 mg cholesterol and 30.000 IU retinylpalmitate (Nutrichem, Roth, Germany). The total energy content was 1017 kcal (4255 kJ). The test meal was drunk within 15 minutes. Blood withdrawal was repeated at 1, 2, 3, 4, 5 and 6 h after ingestion of oMTT. Subjects were allowed to walk or sit, as they wished, but not to exercise during the test. Ad libitum drinking of water or fruit tea without sugar was permitted. Blood was collected through a venous in-dwelling catheter placed in a cubital vein. For assessing glucose tubes containing 1 mg/ml fluoride and 1.2 mg/ml EDTA, for determing triglycerides, insulin and cellular adhesion molecules tubes containing 1.6 mg/ml Potassium EDTA were used. Plasma was separated by centrifugation at 6.48 e+7/min, 4°C and stored at -20°C until analysis.</p>", "<title>Laboratory analyses</title>", "<p>Fasting and postprandial triglyceride and glucose were determined automatically and in duplicate with the kone lab 20i analyzer (Kone, Espoo, Finland) using enzymatic test kits (glucose: Roche, Mannheim, Germany; triglycerides: Boehringer, Mannheim, Germany). Insulin was measured with a radioimmunassay (Biochem Immunosystems, Freiburg, Germany). Plasma concentrations of sICAM-1, sVCAM-1 and sE-selectin were measured in duplicate using a quantitative sandwich ELISA (Boehringer, Mannheim, Germany). The inter-assay and intra-assay coefficients of variation were &lt; 10%.</p>", "<title>Statistical analyses</title>", "<p>The Kolmogorov-Smirnov test was used to determine whether each considered variable had a normal distribution. Comparisons of baseline data among the groups were performed using the unpaired Student's <italic>t </italic>test for normally distributed parameters and with the Mann-Whitney-U test for parameters not following a normal distribution. The paired Student's <italic>t </italic>test was used for comparison of CAMs before and after ingestion of the test meal. If differences reached statistical significance, post-hoc analysis with a two-sided paired <italic>t </italic>test was used to assess differences at individual time periods in the study, using Bonferroni correction to adjust for for multiple comparisons. The statistical significance of postprandial change of sCAMs was determined by comparing the summarized postprandial values (area under the curve, AUC) with the fasting values by a t test or a Mann-Whitney-U test, depending on the presence of a normal distribution. Spearman's coefficient was used to describe correlations of pooled data from NR and HR. In further exploratory analyses, multiple linear regression analysis was used to examine the relative contributions and overlap of metabolic factors possibly contributing to sCAM levels. Results were given as mean ± SEM. The 0–9 h AUC was calculated by the trapezoidal method [##REF##2106931##28##]. Statistical significance was defined as p &lt; 0.05.</p>" ]

[ "<title>Results</title>", "<p>The characteristics of the study population are shown in Table ##TAB##0##1##. High responders showed significantly higher fasting triglycerides and insulin concentrations (p &lt; 0.01, p = 0.039, rsp.), as well as higher fasting sICAM-1, sVCAM-1 and sELAM-1 (p = 0.046, p = 0.034, p = 0.05, rsp.) concentrations.</p>", "<title>Glucose</title>", "<p>None of the subjects showed impaired glucose tolerance. Fasting and postprandial (AUC) glucose levels were not different between groups. In NR, the glucose concentration falls below fasting 60 min after the oMTT, in HR an increase of glucose after 60 min resulted in a significant difference at this time point (p = 0.01) (Figure ##FIG##0##1##).</p>", "<title>Insulin</title>", "<p>Fasting insulin levels were significantly different in normal and high responders (p = 0.04) with higher concentrations for normal responders. After ingestion of the test meal, there was a tendency towards higher insulin levels in HR (p = 0.10 for AUC) (Figure ##FIG##1##2##).</p>", "<title>Triglycerides</title>", "<p>Fasting and following the ingestion of the oral metabolic tolerance test, mean plasma triglyceride levels were significantly higher in high responders as compared to normal responders at any time point (p &lt; 0.001) (Figure ##FIG##2##3##).</p>", "<title>Soluble adhesion molecules</title>", "<p>Soluble adhesion molecules (sICAM-1, sVCAM-1, sE-selectin) did not increase in the postprandial state (Figure ##FIG##3##4##, ##FIG##4##5##, ##FIG##5##6##). Neither the levels at single postprandial time points nor the means over observation time differed from fasting concentrations for sICAM-1, sVCAM-1 and sE-selectin. Therefore, the fasting values were used to analyze differences between NR and HR. sICAM-1 levels were significantly different between NR (210.2 ± 5.94 ng/ml) and HR (304.3 ± 5.69 ng/ml, p = 0.046). Plasma levels of sVCAM-1 were also significantly higher in HR (p = 0.047). Soluble E-selectin levels were not significantly different but tended to be higher in HR (18.4 ± 9.6 vs. 13.2 ± 7.4, p = 0.05).</p>", "<title>Relationship of sCAM and metabolic parameters</title>", "<p>In univariate linear regression analysis, fasting soluble ICAM-1 correlated with with postprandial insulin AUC and the maximum of insulin after oMTT (r = 0.39, p &lt; 0.04, r = 0.7, p &lt; 0.001 rsp.). There was no significant correlation with fasting insulin concentration. Correlation between fasting sVCAM-1 and postprandial triglycerides again showed a stronger correlation (r = 0.37) than with fasting triglycerides (r = 0.34, p = 0.047). There was no significant correlation between fasting sE-selectin and fasting or postprandial glucose, insulin or triglycerides (Table ##TAB##1##2##).</p>", "<p>In multivariate linear regression analysis, including fasting, maximal and AUC-values of insulin, glucose, and triglycerides, we found a strong prediction of fasting sICAM-1 through triglycerides (AUC), insulin maxima, and postprandial glucose (AUC) (r²0.67, p &lt; 0.001). 67% of the variance of fasting sICAM-1 concentration was predicted by these three parameters, although the latter contributed to the model goodness but was not significant as an independent predictor for fasting sICAM-1. For fasting sE-selectin and sVCAM-1, the metabolic parameters were much weaker predictors. Only 8% at most of the variance of fasting sE-selectin and sVCAM-1 where predicted when metabolic parameters were included. sE-selectin was predicted significantly through postprandial triglycerides which remained in the model with a borderline significance for maximal triglycerides (p = 0.08), while for sVCAM-1 none of the included metabolic parameter was an significant predictor.</p>" ]

[ "<title>Discussion</title>", "<p>In this study we showed, that healthy young men with normal fasting triglycerides but high postprandial triglycerides (HR) showed higher levels of sCAMs after a lipid-rich meal compared to subjects with lower postprandial triglycerides (NR). Furthermore, HR are characterized by higher postprandial glucose concentrations, as important early characteristics of the metabolic syndrome, whereas fasting glucose was low in these subjects. The NR showed an initial decrease of glucose concentrations after oMTT which is a sign of intact glucose-insulin regulation and was significantly different from the 60 min increase of glucose in HR.</p>", "<p>Ferri et al. [##REF##9568701##18##] found an increase of sICAM-1 after an oral glucose tolerance test and a correlation with postprandial insulin in subjects with impaired glucose tolerance and hyperlipidemia. Furthermore subjects with isolated hyperlipidemia had normal levels of sICAM in their study. The authors concluded that glucose metabolism and insulin resistance, not hyperlipidemia, has direct impact on levels of sCAM. We showed that in fact postprandial insulin correlates with sICAM. The strongest correlation was seen between both insulin maxima and the insulin area-under-the-curve and sCAMs, especially for sICAM-1. But as there was no postprandial augmentation of sCAM in our study, the conclusion of Ferri et al. that insulin has direct impact on CAMs is not supported by our results. Instead we conclude that there is a general metabolic abnormality seen in high responders which may be an abnormality which eventually may lead to the inflammatory abnormalities accompanying the metabolic syndrome.</p>", "<p>High responders showed higher postprandial lipid peroxidation and reduced intracellular levels of antioxidant vitamin C in previous studies [##UREF##1##22##]. The association between high postprandial triglyceride and insulin levels after a mixed meal with high CAMs is presumably indicating an early stage of endothelial dysfunction.</p>", "<p>In this study strong correlations between postprandial insulin levels and sICAM were seen whereas the correlation between fasting insulin levels and sICAM was less pronounced. Moreover, the linear relationship between sICAM, postprandial triglycerides, and insulin maxima was independent of fasting metabolic parameters in multiple regression analysis. Again, postprandial triglyceride levels showed a significant correlation to sVCAM, while this correlation was not found with fasting triglycerides. In contrast, Ridker et al. [##REF##9439492##9##] found a slight but significant correlation between sICAM-1 and fasting triglycerides in patients with risk for future myocardial infarction. In this study, the postprandial metabolism was not examined. The important role of the postprandial state for induction of early stages of metabolic syndrome and atherosclerosis is evident because of stronger and independent postprandial correlations.</p>", "<p>Intervention studies showed that metabolic disorder as well as diet influence CAM levels and, thus, the endothelial activation. In poorly controlled NIDDM diabetics, increased sE-selectin levels returned to normal after short-term improvement of glycemic control [##REF##9614623##23##]. In other studies, lipid lowering therapy decreased levels of sE-selectin in hypercholesteremic patients [##REF##8641021##15##] and n3-fatty acid treatment decreased sE-selectin and sICAM-1 [##REF##9598830##17##] in hypertriglyceridemic patients.</p>", "<p>According to sICAM-1, we also found higher levels of fasting sVCAM-1 in high responders compared to normal responders. sE-selectin concentrations were also higher in HR, although the difference was statistically not significant. In unstimulated endothelium, adhesion molecule expression is low, with the exception of ICAM-1, which is constitutively expressed to a higher degree. Because of the \"premetabolic\" syndrome in high responders, we assume that ICAM-1 is expressed constantly on endothelial cells in these persons in a higher degree than E-selectin and VCAM, which can explain the most pronounced difference in HR and NR for sICAM-1.</p>", "<p>Correlation analysis of any CAMs with postprandial glucose in our collective was negative, implicating the importance of insulin and triglycerides in early stages of metabolic syndrome when glucose levels are still normal.</p>", "<p>In our subjects there was no evidence of clinical manifested atherosclerosis disease, although early stages of atherosclerosis can not be excluded. HR had a tendency for both higher sCAM levels and elevated postprandial triglyceride and insulin levels. We showed for the first time a correlation between postprandial triglycerides, insulin and sCAMs in healthy subjects. Thus, it may be suggested that postprandial triglycerides and/or insulin and sCAMs are reasonable markers for early metabolic abnormalities and endothelial activation leading to the metabolic syndrome and atherosclerosis, however further studies are needed to confirm this. A limitation of this study is the limited sample size which could explain the tendency for elevated sE-selectin in HR, but without reaching statistical significance. Further studies with alterations of oxidative agens in oral lipid loads are needed to examine postprandial levels of adhesion molecules.</p>", "<p>In this study soluble adhesion molecules did not increase within the 6 h observation period after an oral lipid load. Other investigators found increased levels of CAM as early as 2 hours after a glucose load [##REF##11923038##19##]. An increase of soluble adhesion molecules after high fat meals was shown in other studies [##REF##14988255##20##,##REF##14632967##21##]. Incubation of HUVEC endothelial cells with chylomicrons can induce E-selectin and VCAM-1 expression [##UREF##2##24##,##REF##9288541##25##].</p>", "<p>We assume that the retinol content of our test meal attenuated the postprandial rise of soluble adhesion molecules. We did indeed observe that consumption of the lipid load without retinol increases postprandial levels of sICAM and sE-selectin, and that this effect is prevented by retinol (Pfeuffer et al., personal communication). This fits with the results of Nappo et al. [##REF##11923038##19##]. When these authors administered a high-fat diet meal with or without supplementation of vitamin E and C as antioxidative agents to diabetic and healthy subjects, the high-fat meal without antioxidants significantly increased postpranadial levels of sICAM-1 and sVCAM-1. This effect was prevented by addition of vitamin E and C in both groups.</p>" ]

[ "<title>Conclusion</title>", "<p>In conclusion, soluble adhesion molecules are strongly associated with postprandial but not fasting triglycerides and insulin in young, healthy men which could have implications for future atherogenic risk assessment in healthy subjects, but further studies are needed to confirm this.</p>" ]

[ "<p>This is an Open Access article distributed under the terms of the Creative Commons Attribution License (<ext-link ext-link-type=\"uri\" xlink:href=\"http://creativecommons.org/licenses/by/2.0\"/>), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.</p>", "<title>Background</title>", "<p>The importance of the postprandial state for the early stages of atherogenesis is increasingly acknowledged. We conducted assessment of association between postprandial triglycerides, insulin and glucose after ingestion of a standardized lipid-rich test meal, and soluble cellular adhesion molecules (sCAM) in young healthy subjects.</p>", "<title>Methods</title>", "<p>Metabolic parameters and sICAM-1, sVCAM-1 and E-selectin were measured before and hourly until 6 hours after ingestion of a lipid-rich meal in 30 healthy young men with fasting triglycerides &lt;150 mg/dl and normal fasting glucose levels. Subjects were classified as either normal responders (NR) (postprandial triglyceride maxima &lt; 260 mg/dl) or high responders (HR) (postprandial triglyceride maxima &gt; 260 mg/dl). Levels of CAM were compared in HR and NR, and correlation with postprandial triglyceride, insulin and glucose response was assessed.</p>", "<title>Results</title>", "<p>Fasting sICAM-1 and sVCAM-1 levels were significantly higher in HR as compared to NR (p = 0.046, p = 0.03). For sE-selectin there was such a trend (p = 0.05). There was a strong positive and independent correlation between sICAM-1 and postprandial insulin maxima (r = 0.70, p &lt; 0.001). sVCAM-1 showed significant correlation with postprandial triglycerides (AUC) (r = 0.37, p = 0.047). We found no correlation between sCAMs and fasting insulin or triglyceride concentrations.</p>", "<title>Conclusion</title>", "<p>This independent association of postprandial triglycerides with sICAM-1 may indicate a particular impact of postprandial lipid metabolism on endothelial reaction.</p>" ]

[ "<title>Competing interests</title>", "<p>The authors declare that they have no competing interests.</p>", "<title>Authors' contributions</title>", "<p>DR, MP, and JS were responsible for the study design; DR, and SR were responsible for data collection; DR, and MN were responsible for data analysis; and DR, MP, MN, and JS were responsible for writing the manuscript. All authors read and approved the final manuscript.</p>" ]

[ "<title>Acknowledgements</title>", "<p>We thank C. Laue, A. Thoss and M. Gerull for data and sample collection and technical assistance. This work was financially supported by the BMBF-Project \"Fat and metabolism – gene variation, gene regulation and gene function\" MN (0312823A/B).</p>" ]

[ "<fig position=\"float\" id=\"F1\"><label>Figure 1</label><caption><p><bold>Postprandial concentrations of glucose</bold>. Plasma glucose concentrations following ingestion of an oral metabolic tolerance test in 15 normal (●, NR) and 15 high (△, HR) triglyceride responders (mean ± SEM).</p></caption></fig>", "<fig position=\"float\" id=\"F2\"><label>Figure 2</label><caption><p><bold>Postprandial concentrations of insulin</bold>. Plasma insulin concentrations following ingestion of an oral metabolic tolerance test in 15 normal (●, NR) and 15 high (△, HR) triglyceride responders (mean ± SEM).</p></caption></fig>", "<fig position=\"float\" id=\"F3\"><label>Figure 3</label><caption><p><bold>Postprandial concentrations of triglycerides</bold>. Plasma triglyceride concentrations following ingestion of an oral metabolic tolerance test in 15 normal (●, NR) and 15 high (△, HR) triglyceride responders (mean ± SEM).</p></caption></fig>", "<fig position=\"float\" id=\"F4\"><label>Figure 4</label><caption><p><bold>Postprandial concentrations of sICAM-1</bold>. Plasma sICAM-1 concentrations following ingestion of an oral metabolic tolerance test in 15 normal (●, NR) and 15 high (△, HR) triglyceride responders (mean ± SEM).</p></caption></fig>", "<fig position=\"float\" id=\"F5\"><label>Figure 5</label><caption><p><bold>Postprandial concentrations of sVCAM-1</bold>. Plasma sVCAM-1 concentrations following ingestion of an oral metabolic tolerance test in 15 normal (●, NR) and 15 high (△, HR) triglyceride responders (mean ± SEM).</p></caption></fig>", "<fig position=\"float\" id=\"F6\"><label>Figure 6</label><caption><p><bold>Postprandial concentrations of E-Selectin</bold>. Plasma E-Selectin concentrations following ingestion of an oral metabolic tolerance test in 15 normal (●, NR) and 15 high (△, HR) triglyceride responders (mean ± SEM).</p></caption></fig>" ]

[ "<table-wrap position=\"float\" id=\"T1\"><label>Table 1</label><caption><p>Characteristics of the study population</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td/><td align=\"left\">Normal Responder (n = 15)</td><td align=\"left\">High Responder (n = 15)</td><td align=\"left\">P-value</td></tr></thead><tbody><tr><td align=\"left\">Age (years)</td><td align=\"left\">24.27 ± 2.84</td><td align=\"left\">25.73 ± 2.79</td><td align=\"left\">0.23</td></tr><tr><td align=\"left\">BMI (kg/m²)</td><td align=\"left\">22.43 ± 1.82</td><td align=\"left\">24.15 ± 2.86</td><td align=\"left\">0.12</td></tr><tr><td align=\"left\">WHR</td><td align=\"left\">0.81 ± 0.05</td><td align=\"left\">0.85 ± 0.06</td><td align=\"left\">0.08</td></tr><tr><td align=\"left\">Fasting triglycerides (mg/dl)</td><td align=\"left\">79.4 26.1</td><td align=\"left\">129.1 ± 31.3</td><td align=\"left\">&lt;0.001</td></tr><tr><td align=\"left\">Fasting glucose (mg/dl)</td><td align=\"left\">94.1 ± 16.5</td><td align=\"left\">88.3 ± 9.9</td><td align=\"left\">0.26</td></tr><tr><td align=\"left\">Fasting insulin (mU/l)</td><td align=\"left\">15.3 ± 6.5</td><td align=\"left\">10.9 ± 4.4</td><td align=\"left\">0.039</td></tr><tr><td align=\"left\">Fasting sICAM-1 (ng/ml)</td><td align=\"left\">217.2 ± 40.8</td><td align=\"left\">307.3 ± 79.3</td><td align=\"left\">0.046</td></tr><tr><td align=\"left\">Fasting sVCAM-1 (ng/ml)</td><td align=\"left\">320.6 ± 177.4</td><td align=\"left\">418.7 ± 121.0</td><td align=\"left\">0.034</td></tr><tr><td align=\"left\">Fasting sELAM-1 (ng/ml)</td><td align=\"left\">13.2 ± 7.4</td><td align=\"left\">18.4 ± 9.6</td><td align=\"left\">0.050</td></tr></tbody></table></table-wrap>", "<table-wrap position=\"float\" id=\"T2\"><label>Table 2</label><caption><p>Correlation coefficients of soluble adhesion molecules with fasting and postprandial metabolic parameters</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td/><td align=\"left\">sICAM-1</td><td align=\"left\">sVCAM-1</td><td align=\"left\">sELAM-1</td></tr></thead><tbody><tr><td align=\"left\">Fasting glucose</td><td align=\"left\">-0.070</td><td align=\"left\">0.042</td><td align=\"left\">-0.088</td></tr><tr><td align=\"left\">Glucose max</td><td align=\"left\">0.073</td><td align=\"left\">0.147</td><td align=\"left\">0.241</td></tr><tr><td align=\"left\">Glucose AUC</td><td align=\"left\">0.089</td><td align=\"left\">0.189</td><td align=\"left\">0.057</td></tr><tr><td align=\"left\">Fasting insulin</td><td align=\"left\">-0.038</td><td align=\"left\">0.047</td><td align=\"left\">-0.185</td></tr><tr><td align=\"left\">Insulin max</td><td align=\"left\">0.699*</td><td align=\"left\">0.013</td><td align=\"left\">0.071</td></tr><tr><td align=\"left\">Insulin AUC</td><td align=\"left\">0.385*</td><td align=\"left\">0.249</td><td align=\"left\">-0.004</td></tr><tr><td align=\"left\">Fasting triglycerides</td><td align=\"left\">0.212</td><td align=\"left\">0.337</td><td align=\"left\">0.087</td></tr><tr><td align=\"left\">Triglycerides max</td><td align=\"left\">0.290</td><td align=\"left\">0.316</td><td align=\"left\">0.310</td></tr><tr><td align=\"left\">Triglycerides AUC</td><td align=\"left\">0.288</td><td align=\"left\">0.366*</td><td align=\"left\">0.164</td></tr></tbody></table></table-wrap>" ]

[]

[]

[]

[]

[]

[]

[ "<table-wrap-foot><p>Anthropometric variables, fasting metabolic parameters, and soluble adhesion molecules in 15 normal and 15 high triglyceride responders (mean ± SD).</p></table-wrap-foot>", "<table-wrap-foot><p>Correlation coefficients of fasting and postprandial glucose, insulin, and triglycerides assessed by univariate linear regression analysis following ingestion of an oral metabolic tolerance test in 30 healthy men. The numeric values are standardized regression coefficients. *significant correlation in univariate analysis</p></table-wrap-foot>" ]

[ "<graphic xlink:href=\"1476-511X-7-32-1\"/>", "<graphic xlink:href=\"1476-511X-7-32-2\"/>", "<graphic xlink:href=\"1476-511X-7-32-3\"/>", "<graphic xlink:href=\"1476-511X-7-32-4\"/>", "<graphic xlink:href=\"1476-511X-7-32-5\"/>", "<graphic xlink:href=\"1476-511X-7-32-6\"/>" ]

[]

{ "acronym": [], "definition": [] }

[ "<title>Background</title>", "<p><bold>I</bold>n recent years, several studies have found that point mutation of some tumors was relevant to that of mtDNA, but it is unclear for causal relation, which could not rule out the possibility of mtDNA integration to the nuclear genome and inducing carcinogenesis. Actually there were objective conditions for the intranuclear transfusion and integration of mtDNA and its fragments. Physical, chemical and certain biological factors may cause mtDNA mutations, the collapse of mitochondrial membrane, and give rise to mtDNA and its fragments dissociation into the cytoplasm. When the free mtDNA and its fragments in the cytoplasm generate excessivelly and the activity of DNAase DNAase-like materials is degraded, the free mtDNA or its fragments probably has the similar effect of tumorgenic virus, passing through nucleopore and randomly integrating into genome DNA. The roles of mtDNA intranuclear integration could be as follows: (1) The integration fragments or integration sites do not influence the normal function of genome and have little impact on the biological characteristics of the host cells; (2) activation of a \"healthy gene\" enhances the body's disease resistance and promotes biological evolution; (3) oncogene activation or anti-oncogene inhibition causes cell proliferation and differentiation out of control, which finally leads to cancerization; (4) apoptosis gene activation or anti-apoptosis gene inhibition induces cells apoptosis rapidly. More and more data indicated that mtDNA integration existed in the nuclear genome of tumor cells. Liang etc. [##REF##8692203##1##] has also found the phenomenon of mtDNA fragments intranuclear integration in early glioma cells by fluorescent in situ hybridization of chromosomes. Kamimura etc. [##REF##2614844##2##] detected a section of mtDNA sequence homology in the nDNA of tumor cells, which is composted of three unconsecutive sections of mtDNA: 12S rRNA, cytochrome oxidase I (COX-I) and a part of ND4L/ND4 DNA. Later Shay [##REF##6098801##3##] has got the similar findings in the nuclear genome research on Hela TG cervical cancer cells. mtDNA intranuclear integration may lead to the instability of chromosome DNA and oncogene activation and/or anti-oncogene deactivation, which lead to abnormal cell proliferation and differentiation and finally result in cancerization.</p>", "<p>Carcinoma of the uterine cervix is the second commonest malignancy in women only next to breast cancer. Activation of oncogene and inactivation of anti-oncogene are molecular basis of cancerization of cells. Some scholars [##REF##9407318##4##] suggested that mtDNA, the unique genetic materal outside of chromosome, may be randomly integrated into genome DNA and activate oncogene or inactivate anti-oncogene, and finally induce the development of tumor. Previous study of our lab [##UREF##0##5##] has found that higher frequency of mtDNA mutation existed in cervical cancer. The purpose of this study was to enrich the study of molecular mechanism of cervical cancer by detecting intra-nucleus integration of mtDNA segment in cervical mucosa cells and exploring its correlation with c-myc (an important oncogene).</p>" ]

[ "<title>Materials and methods</title>", "<title>Cases</title>", "<p>40 patients with cervical cancer were collected from 2000 to 2004 for biopsy samples, including 34 cases of squamous cell carcinoma and 6 cases of adenocarcinoma. According to FIGO clinical staging standard, 13 cases were in stage I and 27 cases were in stage II; These cases were classified as histological grading standard: 9 cases in grade I, 21 cases in grade II and 11 cases in grade III. radical hysterectomy plus pelvic curettage of lymph node was performed for all these patients whose age ranged from 36 to 71 years old, and median age was 59.5 years old. 30 cases of CIN and 30 cases of normal cervical epithelia were taken as control. Patients were informed of the nature, goals, potential benefits, and risks of participating in the study and signed a written consent form approved by the Institutional Ethics Committee.</p>", "<title>Pathological section and staining</title>", "<p>Tissues of cervical cancer were taken, fixed with10% formaldehyde and embedded with paraffin, HE staining or IHC were used for sections. Some tissues were taken for 5 μm frozen sections, fixed with 4% paraformaldehyde for hybridization in-situ.</p>", "<title>DNA hybridization in-situ</title>", "<p>mtDNA probe sequence refered to the relative literature [##UREF##1##6##]. Probe marks adopted Roche random primer digoxin marks and reagent kits, according to the manufacture's instruction. 2 μg of restriction endonuclease HaeIII (GG ↓ CC) and HpaII (C ↓ CGG) were respectively added to 1 μg of mtDNA probe in 37°C water bath for 2 h for enzymatic digestion which was prepared for hybridization.) Dilution of the probe was 1.5 ng/μl. The hybridization solution contained: 50% deionized formamide, 0.1%N Lauroylsarcosine,0.02% SDS, 2% blocking reagent and 5 × SSC. Hybridization solution without probe was used as negative control. The steps of the hybridization in situ were as follows: (1) each section was initially treated with 0.01 mol/l HCL and proteinase K at 37°C for 30 min, then washed with 0.1 mol/L glycine for 5 min and fixed with 4% paraformaldehyde for 5 min; (2) Prehybridization: 30 μl prehybridization solution was added to each section, at 37°C for 30 min; (3) Hybridization: 30 μl hybridization solution with probe was added to each section at 37°C for 16 h; (4) 30 μl digoxin antibody labeled with alkaline phosphoric enzyme was added at 42°C for 0.5 h; (5) coloration: the section was colored by NBT/BCIP for 30~60 min for microscopic examination and photograph. If the nucleus had hyacinthine staining but the intercellular substance and control hadn't such staining, it was regarded as positive expression.</p>" ]

[ "<title>Results</title>", "<title>Intranucleus integration of mtDNA sequence</title>", "<p>mtDNA sequence was detected in 15 cases of cervical mucosa nucleus by hybridization in-situ (see Fig ##FIG##0##1##). Integration rates in normal cervix, CIN and cervical cancer were 0%, 13.3% and 27.5% respectively. Difference among the three groups was significant (χ²= 9.054, <italic>P </italic>&lt; 0.05, see Table ##TAB##0##1##).</p>", "<title>C-myc oncogene expression</title>", "<p>Positive expression was diffused in cervical squamous cell cancer but mesenchymal was negative (see Fig ##FIG##1##2##). In CIN, positive expression was seen in atypical proliferation of epithelium whereas normal epithelia and mesenchyma were negative. The expression rates of C-myc in normal cervix, CIN and cervical cancer were 7%, 33% and 73% respectively. Difference among the three groups was significant (χ²= 10.658, χ²= 27.503, <italic>P </italic>&lt; 0.05, Table ##TAB##1##2##).</p>", "<title>Relationship between intranuleus integration of mtDNA sequence and C-myc oncogene expression</title>", "<p>In 100 samples, 41 were c-myc positive, of which 10 cases were positive of mtDNA hybridization in-situ nucleus staining; 59 cases were c-myc negative, of which 5 cases were positive of mtDNA hybridization in-situ nucleus staining. Difference was significant (χ²= 4.81, <italic>P </italic>&lt; 0.05) (Table ##TAB##2##3##)</p>" ]

[ "<title>Discussion</title>", "<p>Proper insertion of mitochondria gene into nucleus genome is important in biological evolution, but improper insertion may be one of the main causes of certain genetic diseases, malformation or tumors. Hu Yide et al [##REF##11866891##7##] adopted gene transfer technology to transfect mtDNA fragment to mouse NIH3T3, which induced cells malignant transformation. It suggested that mtDNA integration in nucleus was an important factor to promote cells cancerization. Ling Xianlong et al [##UREF##2##8##] also found intranucleus integration of mtDNA fragment in cell nucleus of gastric cancer. Shay et al [##REF##1383764##9##] found CoIII of mtDNA in Hela TG cells arranged in c-myc gene. The resulting mRNA contained not only genetic information from c-myc, but also from CoIII. Our lab had similar findings in FISH of Hela cells of cervical cancer cultured in vitro.</p>", "<p>C-myc oncogene was located in chromosome 8q24, total length 6~7 kb. It had 3 exon, coding protein consists of 49 amino acid residues, molecular weight reaches 64/67 kDa. C-myc over-expression and proliferation was often found in cervical cancer tissue. Ngan [##REF##10673985##10##] adopted immunohistochemical technology to study 45 cases of normal cervical tissue, 38 cases of stage I CIN, 37 cases of stage II CIN and 43 cases of stage III CIN: The results showed that c-myc expression was active in poorly developed cells. It suggested that in CIN evolution, c-myc was an important proto-oncogene. Aoyama [##REF##10207671##11##] tested various pathological cervical tissues with PCR. He found that c-myc oncogene was easily proliferated and (or) over-expressed. Present study first used IHC technology to explore c-myc expression in cervical tissues. Data showed that c-myc expression decreases gradually in cervical cancer, CIN and normal cervical tissue. With the increase of malignancy, positive expression became stronger, low differentiation squamous cell cancer was stronger than high differentiation one and grade III CIN was stronger than grade I CIN, which basically conformed to the reports in literature [##REF##8692203##1##].</p>", "<p>In order to exploring the correlation between c-myc oncogene expression and mtDNA integration into genome, DNA hybridization in-situ on frozen sections was also used in present study. The results showed that in the course of chronic cervical inflammation →CIN →cervical cancer, detection rate of mtDNA sequence arranged in DNA genome increased in turn, which suggested its relation with the development of cervical cancer. C-myc gene expression rates of cervical mucosa cells were 67% in mtDNA detection group and 36% in non-detection group respectively. Thus, We infered that acted by certain physical, chemical and biological factors, mtDNA mutation took place. Mutation in D-LOOP may change affinity of trans factors related to mitochondria DNA and replication, copy number of mtDNA therefore increases obviously, normal metabolic balance was damaged, free mtDNA and its fragments were excessive and activity of nucleic acid catabolic enzymes in cells decreases meantime, mtDNA and its fragments were free outside mitochondria, it had similar effect of cancerous virus, arranged randomly into nucleus genome and activated oncogene or inhibited anti-oncogene, which influenced cell proliferation and differentiation and developed tumor.</p>", "<p>Research already found that nucleus DNA and mtDNA might wander in cells [##REF##1942048##12##]. Mitochondria RNA can be reversely transcripted into mtDNA in cytoplasm, there is nucleopore on membrane and DNA ligase in nucleus. So integration of mtDNA in nucleus gene may induce canceration of cells, change of mitochondria structure and quantity increase. Advanced study was needed to confirm the hypothesis above.</p>", "<p>Present study detected intranucleus integration of mtDNA fragment in cervical mucosa cells, and analyzed the correlation between mtDNA fragment integration and c-myc expression. In conclusion, We tentatively assumed that Integration of mtDNA into nuclei of cervical epithelium cells may be involved in the carcinogenesis of cervical epithelium cells and expression of c-myc gene might be related to integration) of mtDNA sequence into nuclei of cervical epithelium cells. It provided new clues to reveal molecular mechanism of cervical cancer.</p>" ]

[]

[ "<p>This is an Open Access article distributed under the terms of the Creative Commons Attribution License (<ext-link ext-link-type=\"uri\" xlink:href=\"http://creativecommons.org/licenses/by/2.0\"/>), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.</p>", "<title>Objective</title>", "<p>To explore the relationship between the integration of mitochondrial DNA(mtDNA) in the nuclei of cervical epithelium cells and the expression of c-myc.</p>", "<title>Methods</title>", "<p>The expression of c-myc protein was measured by immunohistochemical test in 40 cases of the uterine cervix cancer, 30 cases of cervical intraepithelial neoplasia (CIN) and 30 cases of normal cervical epithelium; the sequence of mtDNA in the nuclei was detected by in situ hybridization technique.</p>", "<title>Results</title>", "<p>The detection rates of mtDNA in the nuclei of cervical epithelium cells were 27.5%, 13.3% and 0% in cervical carcinoma, CIN, and normal cervical epithelium respectively. The expression rate of c-myc in cervical mucoma cells was 67% in the mtDNA sequence positive group and was significantly higher than that in the negative group (36%).</p>", "<title>Conclusion</title>", "<p>The integration of mtDNA into the nuclei of cervical epithelium cells may be involved in the carcinogenesis of cervical epithelium cells and the expression of c-myc might be related to the integration of mtDNA sequence into nuclei of cervical epithelium cells.</p>" ]

[ "<title>Detection of the c-myc expression by immunohistochemical test</title>", "<p>Immunohistochemical reagent c-myc monoclonal antibody produced by U.S. Symed (purchased from Beijing Zhongshan Biological Technology Company, China). Operating steps: (1) the section was dewaxed and then put into water. (2) repaired for 20 min by hot platform. (3) The normal horse serum was added to section for 20 min at 1:50 dulitions for blocking) (4) the section was incubated with monoclonal antibody at 1:60 dulitions at 4°C overnight; (5) 1:120 double (secondary) antibody was added at room temperature for 60 min (6) SAHRP 1:150 for 60 min at room temperature. (7) DAB staining. (8) Hematein double staining, dehydration, transparent mount. The standard of c-myc positive reaction: the reaction product of c-myc positive reaction was brown particles, distributed in cell nucleus, it was regarded as positive expression if the percentage of possive cells was more than 30% under 10 high power fields.</p>", "<title>Statistical methods</title>", "<p>The statistical analysis was performed by using SAS (6.12_version)) statistical software. The statistical comparisons between groups were performed by χ²test, <italic>P </italic>&lt; 0.05 was considered statistically significant.</p>", "<title>Competing interests</title>", "<p>The authors declare that they have no competing interests.</p>", "<title>Authors' contributions</title>", "<p>DC conceived of the study, and participated in its design and coordination. WX carried out immunohistochemical test, drafted the manuscript and collected all of cervical cancer samples. JX carried out the DNA hybridization in-situ, participated in the design of the study and performed the statistical analysis. All authors read and approved the final manuscript.</p>" ]

[ "<title>Acknowledgements</title>", "<p>This work was supported by a grant BK2007023 from Jiangsu province Natural Science Foundation of China.</p>" ]

[ "<fig position=\"float\" id=\"F1\"><label>Figure 1</label><caption><p>Integration of mtDNA in the nuclei of cervical epithelium cells detected by in situ hybridization The nucleus had hyacinthine staining but the intercellular substance and control hadn't such staining, it was regarded as positive expression. (A: normal cervix tissure; B: CIN tissure; C: cervical cancer tissure)</p></caption></fig>", "<fig position=\"float\" id=\"F2\"><label>Figure 2</label><caption><p><bold>The expression of c-myc in cervical epithelium cells detected by immunohistochemical test</bold>. The standard of c-myc positive reaction: the reaction product of c-myc positive reaction was brown particles, distributed in cell nucleus, it was regarded as positive expression if the percentage of possive cells was more than 30% under 10 high power fields. (A: normal cervix tissure; B: CIN tissure; C: cervical cancer tissure).</p></caption></fig>" ]

[ "<table-wrap position=\"float\" id=\"T1\"><label>Table 1</label><caption><p>Integration of mtDNA in nuclei of cervical epithelium cells</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td align=\"left\">Groups</td><td align=\"center\">n</td><td align=\"center\">+</td><td align=\"center\">-</td><td align=\"center\">integration rate (%)</td><td align=\"center\">p value</td></tr></thead><tbody><tr><td align=\"left\">Cervical carcinoma</td><td align=\"center\">40</td><td align=\"center\">17</td><td align=\"center\">23</td><td align=\"center\">42.50%</td><td align=\"center\">χ²= 9.054(P &lt; 0.05)</td></tr><tr><td align=\"left\">CIN</td><td align=\"center\">30</td><td align=\"center\">4</td><td align=\"center\">26</td><td align=\"center\">13.30%</td><td align=\"center\">χ²= 9.054(P &lt; 0.05)</td></tr><tr><td align=\"left\">Normal cervix</td><td align=\"center\">30</td><td align=\"center\">0</td><td align=\"center\">30</td><td align=\"center\">0%</td><td/></tr></tbody></table></table-wrap>", "<table-wrap position=\"float\" id=\"T2\"><label>Table 2</label><caption><p>The expression of c-myc in cervical epithelium cells</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td align=\"left\">Groups</td><td align=\"center\">n</td><td align=\"center\">positive</td><td align=\"center\">negative</td><td align=\"center\">positive rate (%)</td><td align=\"center\">p value</td></tr></thead><tbody><tr><td align=\"left\">Cervical carcinoma</td><td align=\"center\">40</td><td align=\"center\">29</td><td align=\"center\">11</td><td align=\"center\">73%</td><td align=\"center\">χ²= 10.658(P &lt; 0.05)<break/>χ²= 27.503(P &lt; 0.05)</td></tr><tr><td align=\"left\">CIN</td><td align=\"center\">30</td><td align=\"center\">10</td><td align=\"center\">20</td><td align=\"center\">33%</td><td align=\"center\">χ²= 10.658(P &lt; 0.05)</td></tr><tr><td align=\"left\">Normal cervix</td><td align=\"center\">30</td><td align=\"center\">2</td><td align=\"center\">28</td><td align=\"center\">7%</td><td align=\"center\">χ²= 27.503(P &lt; 0.05)</td></tr></tbody></table></table-wrap>", "<table-wrap position=\"float\" id=\"T3\"><label>Table 3</label><caption><p>Relationship between integration of mtDNA in nucleus and the expression of c-myc gene</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td/><td align=\"center\" colspan=\"2\">Sequence intranuleus integration (n)</td><td/></tr><tr><td/><td colspan=\"2\"><hr/></td><td/></tr><tr><td align=\"left\">c-myc oncogene expression (n)</td><td/><td/><td align=\"center\">p value</td></tr></thead><tbody><tr><td/><td align=\"center\">negative</td><td align=\"center\">positive</td><td/></tr><tr><td align=\"left\">negative</td><td align=\"center\">54</td><td align=\"center\">5</td><td align=\"center\">χ²= 4.81(P &lt; 0.05)</td></tr><tr><td align=\"left\">positive</td><td align=\"center\">31</td><td align=\"center\">10</td><td align=\"center\">χ²= 4.81(P &lt; 0.05)</td></tr></tbody></table></table-wrap>" ]

[]

[]

[]

[]

[]

[]

[]

[ "<graphic xlink:href=\"1756-9966-27-36-1\"/>", "<graphic xlink:href=\"1756-9966-27-36-2\"/>" ]

[]

{ "acronym": [], "definition": [] }

[]

[]

[]

[]

[ "<title>Conclusion</title>", "<p>Now, the study of KLT-induced apoptosis is focused on the above-mentioned gene. In short, KLT induce apoptosis of cancer cell by way of up regulating the expression of p53 gene, Fas gene, Caspase-3, PCNA, p21^WAFI/CIPIand down regulating the expression of cyclin A, cyclin E1, cyclin F gene. But its effect on the expression of bcl-2 gene and c-myc gene is not yet clear.</p>", "<p>The occurrence of malignant tumor may be caused by the abnormal proliferation of cell or the inhibition of cellular apoptosis pathway. The proliferation of tumor cell and the apoptosis of tumor cell are not only affected by many factors and pathways such as drugs, radioactive ray, etc, but also regulated by some tumor genes or tumor-suppressing gene. It has been proved that the Chinese crude drug-induced apoptosis is one important anticancer mechanism of Chinese crude drug, and it is also relevant to the concentration of the medicine[##REF##12970920##43##]. Just as the result of experimental treatment on hepatoma of mice that we have done [##REF##15567749##44##], we found that KLT can made the cancer cells stop in the G2+M phase of cell life circle, and prevent them form entering the G0 and G1 phase. so it can induce the apoptosis and suppress growth of cells without any effect on the surrounding normal tissues or causing Inflammatory reaction which are unique characteristics. based on these characteristics, KLT has been used in the treatment on many kinds of malignant tumor, and the clinical effect show that KLT can not only repress tumor directly, improve the quality of life obviously, enhance the patient's immunity, but also enhance the effectiveness of chemotherapy and reduce side effects [##UREF##13##45##, ####UREF##14##46##, ##UREF##15##47####15##47##], which is consistent with its pharmacology above-mentioned However, the relationship between the apoptosis of hepatoma carcinoma cell and gene is further studied, although these research can not fully reveal the mechanism of KLT-induced apoptosis, and its experimental result is still preliminary, the trend of using KLT to prevent and cure liver cancer is already formed, and with the development of modern medical technologies, its mechanism will undoubtedly be revealed in the near future. and KLT will bring new hope for the treatment of cancer and the protection of normal tissues with the more sufficient evidence of its effect on liver cancer and other tumor in clinical trials.</p>" ]

[ "<p>This is an Open Access article distributed under the terms of the Creative Commons Attribution License (<ext-link ext-link-type=\"uri\" xlink:href=\"http://creativecommons.org/licenses/by/2.0\"/>), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.</p>", "<p>Many kinds of Chinese herb had been confirmed to have the character of anti-tumor, clinical reports about anti-tumor effects of Chinese herb had also been found in recent years, but most of the reports were focused on the clinical treatment of effectiveness for Chinese herb, on the other hand, review about Chinese herbal related with molecules on cancer-cell-apoptosis was seldom, many scientists could not believe such kinds of clinical describes about anti-tumor effects for Chinese herb, because these describes were lack of molecular biology evidence. Kanglaite(KLT) injection is an anti-tumor new drug which extracts from Chinese medicine-<italic>coix seed </italic>with modern advanced pharmaceutical technology, it is also a new biphase extended-spectrum anticancer medicine, the food and drug administration(FDA) of United States also approved a phase II trial of KLT to test its efficacy in treating non-small-cell lung cancer. Some studies show it could inhibit some anti-apoptotic gene and activate some pro-apoptotic gene, its injection solution is one of the new anticancer medicine that can significantly inhibit a various kinds of tumor cells, so it has become the core of research that how to further explore KLT injection to promote tumor cell apoptosis by impacting on related genes. In this review, the relationship between KLT and some tumor cell apoptosis molecules had been discussed and reviewed generally.</p>" ]

[ "<title>Review</title>", "<p>In recent years, with the lucubrate on tumor cell biology and molecular biology, it has been recognized that the occurrence and development of tumor is not only the result of cell proliferation disorders and disdifferentiation, but also closely correlated with the abnormal apoptosis [##REF##17612632##1##,##REF##17607368##2##]. Although abnormal apoptosis can promote the occurrence and development of tumor, we can also treat tumor by promoting apoptosis of cancer cell [##REF##17698633##3##,##REF##17559085##4##]. Thus it has become the new target in oncotherapy by way of inducing apoptosis of cancer cell. Kanglaite(KLT) injectionis a diphasic broad-spectrum anti-tumor new drug which has depressant effect on many kinds of tumor cells, it is extracted from the chinese crude drug-coixenolide [##UREF##0##5##,##UREF##1##6##] and made use of the latest and most complex modern high technologies in process of preparation [##UREF##2##7##]. Animal experiments show that KLT mainly block G2+M phase of cell circle, thereby reducing the mitotic division of cells, so the proliferation of tumor cell was inhibited, at the same time it can also activate some pro-apoptotic factor, and further lead to apoptosis [##UREF##3##8##]. Clinical application also shows that combined with chemotherapy, KLT has a good effect on the treatment of advanced cancer, particularly in digestive tract cancer, for example, the patient's life span and quality of life improve significantly. The finding show that this preparation has significantly depressant effect and preonunced curative effect on a variety of cancer cells [##REF##18283616##9##]. Although the therapeutic measure of liver cancer contain surgery, radiotherapy, chemotherapy, interventional therapy and so on, their effect is not satisfactory so far [##REF##12970869##10##,##REF##18278788##11##]. However, with the development of cell biology, the theory of apoptosis presents a new hope and path for the treatment of liver cancer, The present study has found that KLT plays an important role in promoting apoptosis of hepatoma cells [##UREF##4##12##]. It has been known that the apoptosis of hepatoma cells is triggered by a variety of receptor-mediated cell signaling, and a variety of protease take part in the apoptotic signal transduction. In addition many kinds of genes are also involved with apoptotic regulation of hepatoma cells. In this paper, the relationship between KLT and the hepatoma cell apoptosis molecules is going to be discussed and reviewed generally.</p>", "<title>The influence of KLT on p53</title>", "<p>There are two types of p53 genes: the wild type p53 gene and the mutant p53 gene. The wild type p53 gene, which is also known as the guardian of gene, is indispensable to regulate the normal cells circle [##REF##18323654##13##,##REF##16912290##14##]. On the one hand, as the important regulating factor in the process of apoptosis, the wild type p53 monitor the integrity of genes all the time, on the other hand, as the nuclear transcripton, it can respectively combines with DNA and RNA polymerase to regulate expression of gene, it can also inhibit the synthesis of DNA, take part in the repair of DNA, induce cell growth stop at the phase of G0. In addition, the wild type p53 gene can induce the Fas-mediated cell procedural death after the damage of DNA, so that the regular growth of cell is maintained. Wang JJ [##UREF##3##8##] had found, while discussing the anticancer mechanism of KLT injection, that the labeling index of wild type p53 in treat group which received KLT injection is 16.8%, while the control group had no expression at all. Furthermore, in the experiment about KLT-induced apoptosis, Bao Y [##UREF##5##15##] discovered that compared with the control group, the mRNA level of the wild type p53 gene significantly Increased in 20 μl/ml KLT experimental group after 48 hours. And in the experiment about apoptosis of multidrug resistance phenotypic human breast cancer cell line MCF7^adrand its cell cycle arrest that induced by KLT. According to the immunohistochemical detection, Guo JW [##UREF##6##16##] found that the wild type p53 gene of MCF7^adrin the control group is negative, but the wild type p53 gene of MCF7^adrin the KLT experimental group is midrange positive, it can be assumed that KLT could up-regulate the expression of p53 and extend half life of p53 protein. In addition, Wei CY [##UREF##4##12##] observed 34 cases' hepatoma carcinoma cell that cultivated in vitro and their changes after treated with KLT, the results indicated that, compared with the control group, the apoptosis of hepatoma carcinoma cell in KLT experimental group is very significant, there was significant difference between them, at the same time, the labeling index of wild type p53 in the treat group is (8.39 ± 1.42)%, but the labeling index of wild type p53 in the control group (2.11 ± 0.97)%, there was significant difference between these two groups. the phenomenon recorded here is the same as the studies of Wang JJ, Guo JW and Bao Y which had indicated [##UREF##3##8##,##UREF##5##15##,##UREF##6##16##]. In conclusion, KLT injection's may induce the apoptosis of tumor cell by way of up-regulate the expression of p53 genes.</p>", "<title>The effect of KLT on the bcl-2 genes</title>", "<p>Proto-oncogenes bcl-2 is the most definite apoptotic antagonist gene so far. In 1984, It was first cloned in t(14; 18) (q32; q21) chromosome translocation of follicular lymphoma cell line. But many studies had also confirmed that the high expression of this gene might inhibit the apoptosis of a variety of cells as well[##REF##18309928##17##,##REF##18309106##18##], thereby it can participate in the occurrence of a variety of tumor. When studying the anticancer mechanism of KLT, Wang JJ [##UREF##3##8##] found that the labeling index of bcl-2 gene was (16.80 ± 3.77)% in the control group, which is higher than that (6.6%) in the 10 μl/ml KLT treat group. Moreover, accompanied with the concentration of KLT increase, the gene expression of bcl-2 decreased in KLT group. In the experiment about KLT-induced the apoptosis of pancreatic cancer cells, according to the Western blot analysis, Bao Y [##UREF##5##15##] discovered that the expression of bcl-2 protein decreased after 72 hours when application of KLT at 20 μl/ml, the results of these two experiments above-mentioned may indicate that KLT induced the apoptosis of cells by down-regulating the expression of bcl-2 genes. Nevertheless, when studying the KLT- induced the apoptosis of cancer cell(HL60), Li Y [##UREF##7##19##] make use of RT-PCR to detect the the gene expression of bcl-2, there was no significant change in genetic transcription after 24 hours when using KLT at 10 ul/ml. So, whether KLT induces apoptosis of cancer cell by down-regulating the expression of bcl-2 genes isn't yet clear, and its role in hepatoma is not learned, which requires further study and research.</p>", "<title>The effect of KLT on Fas-genes</title>", "<p>Fas gene is located on the No.10 chromosome q23, with a length about 25 kb, the codogenic Fas protein consists of 325 amino acids and it can be expressed in many tissues. When Fas protein combining with Fas ligand, signal of apoptosis is send to the cell and the apoptosis of cell is induce [##REF##18310897##20##,##REF##18058722##21##] Many anticancer drugs can induce the apoptosis of cancer cell by up-regulating the expression of Fas gene [##REF##12845662##22##]. when studying the KLT- induced the apoptosis of cancer cell(HL60), Li Y [##UREF##7##19##] make use of RT-PCR to detect the the gene expression of Fas, she observed that genetic transcription strengthened after 24 hours when using KLT at 10 ul/ml. Han SX [##UREF##8##23##] proved that KLT injection can induce the apoptosis of human cervical carcinoma cell by raising the level of Fas gene. Similarly, the experiment, which detected the expression of Fas receptor on the surface of osteogenic sarcoma cell under different concentration of KLT [##UREF##9##24##], showed that under the KLT concentration of 0 ul/ml, 1 ul/ml, 5 ul/ml, 10 ul/ml and 20 ul/ml, the amount of Fas mRNA detected by the RT-PCR analisis is (0.12 ± 0.02) ul/ml, (0.27 ± 0.05) ul/ml, (0.35 ± 0.09) ul/ml, (0.46 ± 0.14) ul/ml and (0.51 ± 0.16) ul/ml respectively, so they considered that accompanied with the concentration of KLT increase, the level of Fas gene in the cancer cell significantly increased. Anyway, KLT may induce apoptosis of cancer cell by up-regulating the expression of Fas genes, but its effect on the Fas gene of liver cancer cell should be further studied.</p>", "<title>The effect of KLT on caspase-3</title>", "<p>Caspases is a group of prolease that induce the apoptosis of cell. Under the normal circumstances, the strict substrate specificity and high effectivity of the activated caspases can assures its narrow spectrum of proteolysis during the process of apoptosis [##REF##9170975##25##,##REF##18339080##26##]. The caspases selectively Shear a group of protein in a simpatico way that lead to functional failure or structural changes of the protein. If the activity of caspases is suppressed, the cellular apoptosis could be disturbed, which could lead to the occurrence and development of tumor, because the dynamic balance between cellular apoptosis and proliferation is disturbed. Caspases-3 and Caspases-8 are the most widely studied in caspases family. Caspase-3, as a important member in the caspases family, is a major functional enzyme in the pathway of cellular apoptotic signal[##REF##10200555##27##,##REF##18076824##28##]. Caspases-3 can cause the clearage of its substrate PARP(116 × 103) and transferred into opyeptide 24 × 103 and 89 × 103 and thereby activate the endonuclease to trigger the complete degradation of DNA. Many factors that regulate the cellular apoptosis can react through caspase-3 prolease. The depressor of caspase-3 can restrain the activation of caspase-3 and degrade the activity of it, so that the apoptosis of cancer cell is restrained [##REF##18057239##29##,##REF##18054197##30##]. Bcl-2 and p53 both interact with caspase family [##REF##18415681##31##, ####REF##18470751##32##, ##UREF##10##33####10##33##]. Some report have discovered, while studying KLT-induced the apoptosis of human pancreatic cancer Paru-8988 cell, that the increase of caspase-3 total protein also show obviously time dependence [##UREF##11##34##]. Moreover, through the test on the substrate of caspase-3-PARP, no degradation product (89 × 103) strap was found in the control group, but the degradation product (89 × 103) strap was discovered after 6 hours in KLT group, which indicates that caspase-3 have enzyme activity. However, no study or research about the effect of KLT on the liver cancer cell had done.</p>", "<title>The effect of KLT on other related genes</title>", "<p>PCNA (proliferating cell nuclear antigen) is a subunit of DNA polymerase and cell cycle-dependent protein whose maximum appearance in the S phase. Some research shows that the expression of PCNA is connection with the low grade tissues[##REF##9778886##35##,##REF##11411868##36##]. Wang JJ found that the gene expression of nuclear PCNA increased obviously, and the labeling index of PCNA was 15. 2% after 48 hours when 0.2 mg/ml KLT effected on the renal cancer cells, while hardly any gene expression of nuclear PCNA was observed in the control group [##UREF##3##8##]. They believed that PCNA took part in the reparative process of KLT- induced DNA injury, and when the degree of DNA injury is too serious to be repaired by PCNA, other genes such as P53 and Bcl-2 will send signals to trigger the apoptosis of cell and they were also convinced that the anticancer effect of KLT was a result of multiple gene interaction and restriction [##UREF##3##8##,##UREF##5##15##,##UREF##6##16##,##UREF##7##19##].</p>", "<p>c-myc is one of the core protein – myc family in oncogene, and it not only is a positive controlling gene in cellular growth and cell life circle, but also take part in the progress of apoptosis, that is to say, it has dualism[##REF##17156951##37##]. Under a circumstances of growth inhibiting, improper expression of c-myc could induce regulatory failure of normal cell life circle and apoptosis[##REF##10378694##38##]. when studying the KLT- induced the apoptosis of cancer cell(HL60), Li Y make use of RT-PCR to detect the the gene expression of c-myc, she observed that there was no significant change in genetic transcription after 24 hours when using KLT at 10 ul/ml [##UREF##7##19##]. As bcl-2 gene, we should do further research and study on it.</p>", "<p>P21^WAFI/CIPIis the downstream gene of p53 gene, its activation contain the p53 dependent path and the p53 non-dependent path[##REF##14580260##39##]. The protein product of P21^WAFI/CIPIcan combine with cell circle protein, cyclin-dependent kinase(CDK), and proliferating cell nuclear antigen(PCNA) to form a quaternionic complex that can stop the cell life circle and depress the cell growth. Guo JW discovered that while up-regulate the expression of p53 protein, KLT can raise the expression of p21^WAFI/CIPImRNA and protein, it indicates that KLT can induce apoptosis of cancer cell by way of the p53 dependent path to up-regulate the expression of p21^WAFI/CIPI[##UREF##6##16##].</p>", "<p>Furthermore, some researchers found that KLT can up-regulate the level of ubiquitin C, RAD17 genes and down regulate t the level of cyclin A, cyclin E1, cyclin F gene in studying the influence of KLT on Patu-8988 cell life circle and gene expression [##UREF##12##40##].</p>", "<p>In addition, some study [##REF##18486908##41##] show that the genes regulated apoptosis are not isolated, they can influence and restrict with each other, for example, Bcl-2 family, IAPs family, c-myc, P53, P35, can affect activation of caspase-3 through regulating activation of caspase-8 and caspase-9[##REF##10089877##42##]. so Kanglaite may promote the interaction of those genes to form a network cycle, amplify cascade reaction and further promote apoptosis. but there is no related research at the present stage.</p>", "<title>Competing interests</title>", "<p>The authors declare that they have no competing interests.</p>", "<title>Authors' contributions</title>", "<p>LY wrote the article under the supervision of DQ. LY, LCS and DQ contributed to the collection and evaluation of date. LY conceived of the study, and participated in its design and coordination. All authors read and approved the final manuscript.</p>" ]

[]

[]

[]

[]

[]

[]

[]

[]

[]

[]

[]

[]

{ "acronym": [], "definition": [] }

[ "<title>Background</title>", "<p>Health-related quality of life is now considered an important endpoint in cancer clinical trials. It has been shown that assessing quality of life in cancer patients could contribute to improved treatment and could even be as prognostic as medical factors could be prognostic [##UREF##0##1##, ####UREF##1##2##, ##REF##12639062##3##, ##REF##11165402##4####11165402##4##]. Above all, studies of quality of life can further indicate the directions needed for more efficient treatment of cancer patients. Among the quality of life studies in cancer patients, breast cancer has received most attention for several reasons. First, the number of women with breast cancer is increasing. It has been reported that each year over 1.1 million women worldwide are diagnosed with breast cancer and 410,000 die from the disease [##UREF##2##5##]. Secondly, early detection and treatment of breast cancer have improved and survivors now live longer, so studying quality of life in this context is important. Thirdly, breast cancer affects women's identities and therefore studying quality of life for those who lose their breasts is vital. In addition, it is believed that females play important roles as partners, wives, and mothers within any family. Thus, when a woman develops breast cancer, all members of family might develop some sort of illnesses. In fact, breast cancer is a family disease. Other reasons could be added, but overall it is crucial to recognize that with increasing improvements in medicine and medical practice during recent years studying quality of life for any cancer, for any anatomical site and for either gender is considered highly relevant. A descriptive study of the published papers (230 articles) on non-biomedical outcomes (quality of life, preferences, satisfaction and economics) in breast cancer patients, covering the literature from 1990 to 2000, found that the most frequently reported outcomes were health-related quality of life (54%), followed by economic analyses (38%), and patient satisfaction (14%). Only 9% measured patient preferences [##REF##15504918##6##].</p>", "<p>Over the past 10 years, much clinical effort has been expended in the treatment of breast cancer in order to improve survival. Now the question is: to what extent have studies of quality of life in breast cancer patients added to our information or contributed to improved outcomes in breast cancer care? This is very difficult to answer, but it is possible to try to investigate the contribution of quality of life studies to breast cancer care as a whole. There are several useful review papers on quality of life in breast cancer patients. However, most published papers have either been overviews or systematic literature searches with very focused objectives. The aim of this review is to collect and examine all literature published since the topic first appeared in English language biomedical journals. It is hoped that this extensive review may contribute to existing knowledge, help both researchers and clinicians to have a better profile on the topic, and consequently aid in improving quality of life in breast cancer patients.</p>" ]

[ "<title>Methods</title>", "<p>As part of a study on quality of life in breast cancer patients, an extensive literature search was carried out using MEDLINE, EMBASE, the Science Citation Index (ISI), the Cumulative Index to Nursing and Allied Health Literature (CINAHL), the PsycINFO, the Allied and Complementary Medicine (AMED), and Global Health databases. The intention was to review all full publications that have been appeared in English language biomedical journals between 1974 and 2007. The year 1974 was chosen because the first study on quality of life in breast cancer patients was published then. The search strategy included the combination of key words 'quality of life' and 'breast cancer' or 'breast carcinoma' in titles of publications. It was though that this might help to focus the investigation. It provided the initial database for the review. The initial search was carried out in early 2006 and updated twice in 2006, twice at the end of January and December 2007, and once for a final check in April 2008.</p>" ]

[ "<title>Results</title>", "<title>Statistics</title>", "<p>A total of 971 citations were identified and after exclusion of duplicates, the abstracts of 606 citations were reviewed. Of these, meetings abstracts, editorials, brief commentaries, letters, errata and dissertation abstracts and papers that appeared online and were indexed ahead of publication were also excluded. The remaining 477 papers were examined in this bibliographic review. The statistics are shown in Table ##TAB##0##1## and a chronological list of all papers is available [Additional file ##SUPPL##0##1##]. Here, the major findings are summarized and presented under the following headings.</p>", "<title>Reviews</title>", "<p>There were several review papers. These were divided into two categories: overviews [##REF##2205372##7##, ####REF##2033420##8##, ##REF##8502817##9##, ##UREF##3##10##, ##REF##7827169##11##, ##REF##10886990##12##, ##REF##7841968##13##, ##REF##9556780##14##, ##REF##10370362##15##, ##REF##11128120##16##, ##REF##11255204##17##, ##REF##15687635##18##, ##REF##12207554##19##, ##REF##12117073##20##, ##REF##15590318##21##, ##UREF##4##22##, ##REF##16476831##23##, ##UREF##5##24##, ##REF##17393190##25##, ##REF##17540137##26####17540137##26##], and systematic reviews [##REF##9396988##27##, ####REF##12372724##28##, ##REF##12185329##29##, ##REF##12591983##30##, ##REF##12657232##31##, ##REF##12491494##32##, ##REF##15528966##33##, ##REF##16226458##34##, ##REF##18028021##35####18028021##35##]. Whilst there were quite significant numbers of commentaries, some brief, a few systematic reviews with focused objectives were also identified. These are summarized in Tables ##TAB##1##2## and ##TAB##2##3##. Both overviews and systematic reviews touched interesting topics pointed to helpful comments and findings among published papers. For instance, a paper by Rozenberg et al. [##REF##17540137##26##] highlighted that most women affected by breast cancer will not die from it but from other diseases, owing to recent improvements in treatment. They also pointed out that women with breast cancer and three or more co-morbid conditions have a 20-fold higher rate of mortality from causes other than breast cancer and a 4-fold higher rate of all-cause mortality when compared with patients who have none.</p>", "<p>Health-related quality of life in patients undergoing systemic therapy for advanced breast cancer was reviewed by Bottomley and Therasse, covering the literature from 1995 to 2001. They indicated that there were 19 studies. Among these, there were 12 studies on chemotherapy, 6 hormonal trials and 1 on biological therapy (Trastuzumab). They concluded that quality of life data provided invaluable insights into the treatment and care of patients [##REF##12372724##28##].</p>", "<p>To help the selection of optimal treatment, Goodwin et al. conducted a review of measurements of health-related quality of life in randomized clinical trials in breast cancer patients, covering the literature from 1980 to 2000. They identified a total of 256 randomized trials in breast cancer that included health-related quality of life or psychosocial outcomes. Of these, 66 trials involved randomized of different treatment options, 46 evaluated biomedical interventions and 20 evaluated psychosocial interventions. They concluded that until the results of ongoing trials are available, caution is recommended in initiating new quality of life studies unless treatment equivalence is expected or unless unique or specific issues can be addressed [##REF##12591983##30##]. Similarly, Fossati's critical review of published literature on randomized clinical trials of cytotoxic or hormonal treatments of advanced breast cancer indicated that quality of life assessments added relatively little value to classical clinical endpoints [##REF##15528966##33##].</p>", "<p>Mols et al. reviewed the literature on quality of life among long-term survivors of breast cancer and found that although these patients experienced some specific problems such as a thick and painful arm and problems with sexual functioning, most reported good overall quality of life. The review also indicated that the current medical condition, amount of social support and current income level were strong positive predictors of quality of life, and the use of adjuvant chemotherapy emerged as a negative predictor. The authors concluded that focusing on the long-term effects of breast cancer is important when evaluating the full extent of treatment [##REF##16226458##34##].</p>", "<p>Grimison and Stockler reviewed quality of life in early-stage breast cancer patients receiving adjuvant systemic therapy, review of clinical randomized trials covering the literature from 1996 to 2007, and concluded that the long-term effects of chemotherapy-induced menopause and hormonal therapy on quality of life were poorly recognized. They found that vasomotor symptoms and altered sexual function were common, distressing and inadequately treated [##REF##18028021##35##].</p>", "<title>Two historical papers</title>", "<p>The first paper on quality of life in breast cancer patients was published in 1974. In this historical paper advanced breast cancer patients receiving adrenalectomy with chemotherapy were assessed for objective and subjective response rates, survival and quality of life. The results showed that in 64% of the patients the subjective palliation involved a return to essentially normal living during the period of improvement [##REF##4136510##36##]. The second historical paper on the topic was appeared two years later, in 1976; Priestman and Baum used a linear analogue self-assessment (LASA) to measure the subjective effects of treatment in women with advanced breast cancer [##REF##58161##37##]. The results showed that this technique might be used to monitor the subjective benefit of treatment and to compare the subjective toxicities of different therapeutic regimens. The results also suggested that the subjective toxicity of cytotoxic therapy was not related to the patient's age and diminished with successive courses of drugs. However, not until the late 1980s and early 1990s was the literature gradually supplemented with papers using relatively standard and established instruments to measure quality of life in breast cancer patients.</p>", "<title>Instruments used</title>", "<p>Broadly, quality of life measures can be classified as: general, disease specific, and site-specific. Although the early studies did not use standard measures, several valid instruments for measuring quality of life in breast cancer patients have been developed in recent years. The most commonly-used instruments were: the European Organization for Research and Treatment of Cancer (EORTC) Quality of Life Questionnaire and its Breast Cancer supplement (EORTC QLQ-C30 and QLQ-BR23); the Functional Assessment of Chronic Illness Therapy General Questionnaire and its Breast Cancer Supplement (FACIT-G and FACIT-B formerly FACT questionnaires); the Breast Cancer Chemotherapy Questionnaire (BCQ); the Hospital Anxiety and Depression Scale (HADS); and the Medical Outcomes Study Short Form Survey (SF-36). Table ##TAB##3##4## lists a number of most important instruments used in studies of quality of life in breast cancer patients. Almost all these instruments proved to be valid and were found to be very popular among researchers and clinicians.</p>", "<title>Validation studies</title>", "<p>Development of instruments for measuring quality of life in breast cancer patients, or cultural adaptation and validation studies of the existing instruments, was the major theme in a number of papers. These are presented in Table ##TAB##4##5##[##REF##3058874##38##, ####UREF##6##39##, ##REF##1768628##40##, ##REF##8853525##41##, ##REF##8998495##42##, ##REF##8874337##43##, ##REF##9060536##44##, ##REF##9739441##45##, ##REF##9713301##46##, ##REF##10481946##47##, ##REF##10983481##48##, ##UREF##7##49##, ##REF##11727963##50##, ##UREF##8##51##, ##REF##12018741##52##, ##UREF##9##53##, ##REF##15342666##54##, ##REF##15088137##55##, ##REF##16252544##56##, ##REF##17140438##57##, ##REF##17377841##58##, ##REF##17221159##59####17221159##59##]. A paper by Levine et al. in 1988 was the first validation study in this field. It reported a quality of life measure in breast cancer patients called the Breast Cancer Chemotherapy Questionnaire (BCQ). This is a 30-item questionnaire that focuses on loss of attractiveness, fatigue, physical symptoms, inconvenience, emotional distress and feelings of hope and support from others [##REF##18028021##35##]. A few studies reported translation and validation findings for the instruments used to assess quality of life among breast cancer patients in different cultures (for example see [##REF##10983481##48##,##REF##15342666##54##,##REF##16252544##56##]).</p>", "<title>Measurement issues</title>", "<p>Papers that dealt with issues of quality of life measurement in breast cancer patients encompassed a variety of topics, mainly focusing on methodological and practical concerns in such assessment, especially in clinical settings. Most authors have tried first to convince clinicians to assess quality of life, and secondly to show how quality of life data could contribute to care and management of breast cancer patients. Table ##TAB##5##6## presents a summary of the results [##REF##2194541##60##, ####REF##2370573##61##, ##REF##1385719##62##, ##REF##1588353##63##, ##REF##8119064##64##, ##REF##8265446##65##, ##REF##8423565##66##, ##REF##8433117##67##, ##REF##8172795##68##, ##REF##7546841##69##, ##REF##8619938##70##, ##REF##9358933##71##, ##REF##9744512##72##, ##REF##9549808##73##, ##REF##9469334##74##, ##UREF##10##75##, ##REF##11236851##76##, ##REF##10785587##77##, ##UREF##11##78##, ##REF##11678311##79##, ##REF##11804378##80##, ##UREF##12##81##, ##REF##16117596##82##, ##UREF##13##83##, ##REF##17474993##84####17474993##84##].</p>", "<title>Surgical treatment</title>", "<p>Breast cancer surgery including conservative surgery followed by irradiation, and modified radical mastectomy or radical mastectomy followed by immediate reconstruction is associated with different side-effects including pain, and fatigue and thus affecting quality of life in breast cancer patients. A list of studies on surgery and quality of life in breast cancer patients is given in Table ##TAB##6##7##[##REF##3881629##85##, ####REF##3780987##86##, ##REF##1551058##87##, ##REF##11091520##88##, ##UREF##14##89##, ##UREF##15##90##, ##REF##11091530##91##, ##REF##9226025##92##, ##REF##9469332##93##, ##REF##9640214##94##, ##REF##10379856##95##, ##REF##10517342##96##, ##REF##10513913##97##, ##REF##10424401##98##, ##REF##11297021##99##, ##REF##14731583##100##, ##REF##11283922##101##, ##REF##11456055##102##, ##REF##12783001##103##, ##REF##12890459##104##, ##REF##15125749##105##, ##REF##14996859##106##, ##UREF##16##107##, ##REF##15927829##108##, ##REF##16180226##109##, ##REF##16163445##110##, ##REF##16887839##111##, ##REF##16387467##112##, ##REF##17574501##113####17574501##113##]</p>", "<p>The most important topic in studies of breast cancer surgery and quality of life relates to the type of surgery. Recent findings suggest that partial and total mastectomy appear to be equivalent treatments in terms of patients' long-term quality of life. However, both short-term and long-term distress levels after partial and total mastectomy may depend on patient's age at diagnosis [##REF##9469332##93##]. A study of early breast cancer patients one year after mastectomy or conservative surgery and radiation therapy found that the differences between treatment groups were mainly accounted for by adjuvant therapies. Those treated by breast conservation reported better body image but worse physical functions. The negative impact of breast cancer and its treatment was greater for younger women across a number of dimensions of quality of life measures regardless of treatment type [##REF##14731583##100##].</p>", "<p>In addition, one study found that aspects of quality of life other than body image were no better in women who underwent breast-conserving surgery or mastectomy with reconstruction than in women who had mastectomy alone. Furthermore, mastectomy with reconstruction was associated with greater mood disturbance and poorer health [##REF##11283922##101##]. However, the results of a 5-year prospective study on quality of life following breast-conserving surgery or mastectomy indicated that mastectomy patients had a significantly worse body image; role and sexual functioning, and their lives were more disrupted [##REF##15125749##105##]. A recent Japanese study on the early effects of surgery in patients with breast cancer performing multivariate analysis reported that there were no significant differences in quality of life before and after surgery, but quality of life was significantly better among women undergoing breast conservation than those undergoing mastectomy [##REF##16887839##111##]. A study comparing the short- and long-term effects of mastectomy with reconstruction, mastectomy without reconstruction, and breast conservation therapy on aspects of psychosocial adjustment and quality of life in a sample of 258 women with breast cancer concluded that overall, the general patterns of psychosocial adjustment and quality of life were similar among the three surgery groups. In addition the study results showed that during the long-term follow-up period (6 months to 2 years after surgery), women in all three groups experienced marked improvements in psychosocial adjustment (depressive symptoms, satisfaction with chest appearance, sexual functioning) and quality of life in physical and mental health domains [##REF##17574501##113##].</p>", "<title>Systemic therapies</title>", "<p>In order to reduce the risk of recurrence and death, breast cancer patients usually receive systemic therapies (chemotherapy, hormonal therapy and biological treatments) after surgery. Several studies evaluated quality of life in breast cancer patients receiving systemic therapies. A list of studies reporting on the topic is given in Table ##TAB##7##8##[##REF##4136510##36##,##REF##58161##37##,##REF##7004560##114##, ####REF##3683485##115##, ##REF##2148877##116##, ##REF##2003707##117##, ##REF##1683557##118##, ##REF##1627368##119##, ##REF##1567662##120##, ##REF##1383437##121##, ##REF##8431375##122##, ##REF##8142261##123##, ##REF##7740333##124##, ##REF##8622073##125##, ##REF##8879621##126##, ##REF##8885485##127##, ##REF##8622502##128##, ##UREF##17##129##, ##UREF##18##130##, ##REF##9928575##131##, ##REF##10533471##132##, ##REF##10855346##133##, ##REF##10482198##134##, ##REF##10445420##135##, ##REF##10673516##136##, ##REF##10717247##137##, ##REF##10930796##138##, ##REF##10930797##139##, ##REF##10944595##140##, ##REF##10899655##141##, ##REF##11016752##142##, ##UREF##19##143##, ##REF##12118024##144##, ##REF##12202668##145##, ##REF##12419743##146##, ##REF##11904787##147##, ##REF##14610048##148##, ##REF##14512398##149##, ##REF##15319567##150##, ##REF##15514369##151##, ##REF##15226325##152##, ##REF##15545973##153##, ##REF##16037685##154##, ##UREF##20##155##, ##REF##16184455##156##, ##REF##15937502##157##, ##REF##15863376##158##, ##REF##15994149##159##, ##REF##16118805##160##, ##REF##16619567##161##, ##UREF##21##162##, ##UREF##22##163##, ##REF##16484701##164##, ##REF##16541325##165##, ##REF##16944295##166##, ##REF##16823511##167##, ##UREF##23##168##, ##REF##17236771##169####17236771##169##].</p>", "<p>Chemotherapy has considerable effect on quality of life of breast cancer patients. In a study of postoperative adjuvant chemotherapy in primary node positive breast cancer patients (one or more axillary node), women receiving a single agent or a multi-drug regimen indicated that the treatment was <italic>'unbearable' </italic>[##REF##7004560##114##] or in a study of patients with early breast cancer receiving preoperative chemotherapy almost all patients considered chemotherapy the most <italic>'burdensome' </italic>aspect of the treatment [##REF##2148877##116##].</p>", "<p>The side-effects of chemotherapy on quality of life in breast cancer patients were the topic of many investigations. In these studies, investigators looked at the issue from different perspectives. For instance, using a decision-analytic approach to evaluate tradeoffs between efficacy and quality of life in the choice of three adjuvant treatments (chemotherapy, surgical ovarian suppression, and medical ovarian suppression) in pre-menopausal women with newly-diagnosed, hormone-responsive early breast cancer, Elkin et al. concluded that when different treatments have similar efficacy, there may be a subgroup of women for whom quality of life considerations dominate the choice. However, they stated that small differences in the relative efficacy of these therapies have a substantial impact on treatment choice [##REF##16184455##156##].</p>", "<p>To improve clinical outcomes an international randomized controlled trial compared dose-intensive chemotherapy with standard systemic chemotherapy in patients with locally advanced breast cancer and showed that a dose-intensive regimen only has a temporary effect on health-related quality of life, thus enabling more research on intensive treatment for patients with locally advanced breast cancer, as it might also offer a survival benefit [##REF##15863376##158##].</p>", "<p>However, recent studies focusing on adjuvant hormonal therapies (tamoxifen or aromatase inhibitors such as anastrozole, letrozole, exemestane) and quality of life in postmenopausal early-stage breast cancer patients reported more encouraging results. Most studies found that overall quality of life was improved in patients receiving either anstrozole or tamoxifen but patients reported different side effects [##REF##15514369##151##,##REF##16944295##166##]. A trial comparing tamoxifen with exemestane showed that quality of life did not change significantly in either groups, but there were improvements in endocrine-related symptoms [##REF##16484701##164##].</p>", "<p>In summary, as noted by Grimison and Stockler, for the majority of breast cancer patients most aspects of health-related quality of life recover after adjuvant chemotherapy ends without long-term effects except vasomotor symptoms and sexual dysfunction. However, tamoxifen and aromatase inhibitors cause long-term effects due to vasomotor, gynecological and sexual problems [##REF##18028021##35##].</p>", "<title>Quality of life as predictor of survival</title>", "<p>Until recently, only a few studies had reported a relationship between quality of life and survival in breast cancer patients [##REF##3683485##115##]. A study using the Daily Diary Card to measure quality of life in advanced breast cancer showed that the instrument offered accurate prognostic data regarding subsequent response to treatment and survival duration [##REF##8431363##170##]. Similarly, Seidman et al. evaluated quality of life in two phase II clinical trials of metastatic breast cancer and found that baseline scores of two validated quality of life instruments independently predicted the overall likelihood of tumour responses [##REF##7544834##171##].</p>", "<p>Studies have shown that baseline quality of life predicts survival in advanced breast cancer but not in early stage of disease [##REF##11078489##172##]. Two recently published papers also confirmed that baseline quality of life is not a prognostic factor in non-metastatic breast cancer patients. One of these two studies, using Cox survival analysis, indicated that neither health-related quality of life nor psychological status at diagnosis or 1 year later was associated with medical outcome in women with early-stage breast cancer [##REF##15483029##173##]. The other study with a sample of 448 locally advanced breast cancer patients, reported that baseline health-related quality of life parameters had no prognostic value in a non-metastatic breast cancer population [##REF##15310784##174##]. However, other studies have demonstrated that some aspects of quality of life data including physical health [##REF##1453197##175##], pain [##REF##10930797##139##,##REF##12826039##176##], and loss of appetite [##REF##15093577##177##] were significant prognostic factors for survival in women with advanced breast cancer. In addition, one study demonstrated that baseline physical aspects of quality of life and its changes were related to survival, but psychological and social aspects were not [##REF##10752779##178##].</p>", "<title>Psychological distress</title>", "<p>Women with breast cancer might develop psychological distress including anxiety and depression during diagnosis and treatment and after treatment. The psychological impact of breast cancer has received considerable attention. Since this is a separate topic, the focus here is on psychological distress as it relates to quality of life studies in breast cancer patients. Table ##TAB##8##9## summarizes the papers on the topic [##UREF##24##179##, ####REF##8861837##180##, ##REF##8874336##181##, ##REF##8699200##182##, ##REF##8911127##183##, ##REF##9176972##184##, ##REF##9734576##185##, ##UREF##25##186##, ##REF##10382190##187##, ##UREF##26##188##, ##UREF##27##189##, ##REF##11436544##190##, ##UREF##28##191##, ##REF##14581426##192##, ##REF##12743147##193##, ##REF##14745853##194##, ##REF##14722597##195##, ##UREF##29##196##, ##UREF##30##197##, ##UREF##31##198##, ##REF##15961434##199##, ##REF##15898865##200##, ##REF##15800768##201##, ##UREF##32##202##, ##REF##16155769##203##, ##REF##16889323##204##, ##REF##17154743##205##, ##REF##16783127##206##, ##REF##16594937##207##, ##REF##18000503##208##, ##REF##17205280##209##, ##REF##17878129##210####17878129##210##].</p>", "<p>Psychological distress in breast cancer patients is mostly related to depression, anxiety, and low emotional functioning and almost all studies have shown that psychological distress contributed to impaired quality of life especially emotional functioning, social functioning, mental health and overall quality of life. The diagnosis of the disease, importance of fears and concerns regarding death and disease recurrence, impairment of body image, and alteration of femininity, sexuality and attractiveness are factors that can cause unexpected psychological distress even years after diagnosis and treatment [##REF##9045314##211##, ####REF##16823173##212##, ##REF##17674188##213####17674188##213##].</p>", "<p>Studies have shown that psychological factors predict subsequent quality of life [##REF##15898865##200##] or even overall survival in breast cancer patients [##REF##17203386##214##]. A study showed that patients with lower coping capacity reported higher prevalence of symptoms, experienced higher levels of distress, and experienced worse perceived health, which in turn decreased their quality of life [##REF##17544244##215##]. Furthermore, it has been shown that psychological adjustment such as the ability to cope with the disease, treatment and effects of treatment could improve outcome. The relationship between positive thinking and longer survival and a better quality of life is well documented [##REF##11399287##216##].</p>", "<title>Supportive care</title>", "<p>A variety of topics were covered to address supportive care issues in breast cancer patients. These ranged from papers on controlling emesis to papers that reported issues related to counseling, social support and exercise to improve quality of life [##REF##1675865##217##, ####REF##2067964##218##, ##REF##1387929##219##, ##REF##8229122##220##, ##REF##8459989##221##, ##REF##10026552##222##, ##REF##9341354##223##, ##REF##10522767##224##, ##REF##10612010##225##, ##REF##10920832##226##, ##REF##11250997##227##, ##REF##11409065##228##, ##UREF##33##229##, ##REF##12546524##230##, ##REF##12533272##231##, ##REF##16594274##232##, ##REF##12721239##233##, ##REF##15170652##234##, ##REF##15378098##235##, ##REF##14734954##236##, ##REF##15251160##237##, ##REF##15363874##238##, ##REF##15870721##239##, ##REF##15856335##240##, ##UREF##34##241##, ##REF##16136270##242##, ##REF##15892425##243##, ##REF##15452188##244##, ##REF##16001991##245##, ##REF##15759065##246##, ##UREF##35##247##, ##REF##17135823##248##, ##REF##16012817##249##, ##UREF##36##250##, ##REF##17785709##251##, ##REF##17396040##252##, ##REF##18062617##253####18062617##253##]. The results are summarized in Table ##TAB##9##10##.</p>", "<title>Symptoms</title>", "<p>There were studies on breast cancer symptoms and their relationship to quality of life. Most of these studies were related to fatigue, lymphedema, pain, and menopausal symptoms. The results are summarized in Table ##TAB##10##11##[##UREF##37##254##, ####REF##9576289##255##, ##UREF##38##256##, ##REF##10219851##257##, ##REF##10673515##258##, ##REF##11200778##259##, ##REF##10908824##260##, ##REF##12413312##261##, ##REF##12377968##262##, ##UREF##39##263##, ##REF##12779081##264##, ##REF##12889597##265##, ##REF##14668143##266##, ##REF##15669929##267##, ##REF##15812652##268##, ##REF##16037759##269##, ##REF##15860132##270##, ##REF##15740827##271##, ##UREF##40##272##, ##REF##16754728##273##, ##REF##16518447##274##, ##REF##16428125##275##, ##REF##16470230##276##, ##REF##17048250##277##, ##REF##17319631##278##, ##REF##17363839##279##, ##UREF##41##280####41##280##].</p>", "<p>Fatigue is the least definable symptom experienced by patients with breast cancer and its effect on impaired quality of life cannot be explained precisely. A recent publication studying 1,588 breast cancer patients showed that fatigue (as measured by the EORTC QLQ-C30 fatigue subscale) independently predicted longer recurrence-free survival when biological factors were controlled in the analysis. When combined with the biological model, fatigue still remained a significant predictor of recurrence-free survival [##REF##17203386##214##].</p>", "<title>Sexual functioning</title>", "<p>Breast cancer could be regarded as a disease that relates to women's identities. In this respect, sexual functioning is an important issue, especially in younger breast cancer patients. Among quality of life studies in breast cancer patients only six papers focused especially on sexual functioning [##UREF##42##281##, ####UREF##43##282##, ##REF##9469334##283##, ##UREF##44##284##, ##REF##17160080##285##, ##REF##17613483##286####17613483##286##]. The findings indicated that disrupted sexual functioning or unsatisfactory sexual life was related to poorer quality of life at younger age, treatment with chemotherapy, total mastectomy, emotional distress consequent on an unsatisfactory sexual life, and difficulties with partners because of sexual relationships.</p>" ]

[ "<title>Discussion</title>", "<p>This bibliographic review has provided an extensive list of studies that focused on quality of life in breast cancer patients. The article might be criticized on the grounds that it included every paper on the topic and that it provides more enumeration than insight. However, this was not an in-depth review but rather, as indicated in the title, a bibliographic investigation and descriptive in nature. The benefit of such an approach is that it reveals how much effort has been made in this area and shows the achievements of a journey that was started more than 30 years ago. If quality of life has now become an important part of breast cancer patients' care, it is due to all these efforts. Furthermore, this approach might help potential investigators to formulate new questions or conduct more focused studies on the topic in the future. It should be admitted that investigations of this type have limitations and are inconclusive. Since in this review the search strategy was limited to the key words 'quality of life' and 'breast cancer' in titles, perhaps many other papers also were missed even from enumeration. However, an up coming complementary review by the author will focus on these missing papers.</p>", "<p>A number of studies that covered measurement issues and introduced instruments used to measure quality of life in breast cancer patients. Hopefully there is now sufficient evidence to use these valid instruments and to adopt the practices that are needed to assess quality of life in research or clinical settings. Since 1974, when the first study on quality of life in breast cancer patients was published, there has been quite impressive progress and improvement, indicating that measuring quality of life in breast cancer patients is both crucial and scientific. Now several valid instruments that capture quality of life dimensions in cancer patients in general and in breast cancer patients in particular are available. The EORTC QLQ-C30, EORTC QLQ-BR23, FACIT-G and FACIT-B are among the most acceptable instruments to patients and health professionals. They have been used in many studies, so it is possible to compare results between studies with similar objectives. It seems that it is time to stop developing new instruments, since there are enough valid and comprehensive measures to assess quality of life in breast cancer patients. New instruments might cause confusion and may be regarded as a waste of resources, so any such developments would need robust justification. Depending on the objectives of any single study, one might use other existing valid measures such as the Satisfaction with Life Domains Scale for Breast Cancer (SLDS-BC), which can briefly and rapidly assess quality of life across the breast cancer continuum of care [##REF##14616940##287##]; the Body Image After Breast Cancer Questionnaire (BIBCQ); which is a valid measure for assessing the long-term impact of breast cancer on body image [##REF##16684320##288##]; and the Fallowfield's Sexual Activity Questionnaire (FSAQ), which is a useful tool for measuring sexual activity in women with cancer [##REF##17785035##289##].</p>", "<p>There were some important technical issues that should be addressed. Some believe that if we perform complex analyses of quality of life data or if we use several instruments in a single study then we might achieve more scientific results. There is evidence that this could merely lead to misleading findings and might be a source of suffering for the patients [##REF##17474993##84##]. The recommendation is to analyze data in a simple way and avoid complexity. The presentation of data should be straightforward and easy to follow; otherwise those who are critical of such findings might conclude that these are manipulations of data, or they might ask whether these numbers and statistics reflect what really happens to breast cancer patients or the clinical teams that care for them. Do these figures convey difficulties that exist in treating breast cancer patients or help to manage their symptoms?</p>", "<p>The present review covered several topics and provided tables to indicate areas that need more attention. It appears that the most common and important disease- and treatment-related side-effects and symptoms in breast cancer patients including arm morbidity, pain, fatigue and postmenopausal symptoms, are among neglected topics. As noted by Cella and Fallowfield, recognition and management of treatment-related side-effects for breast cancer patients receiving adjuvant endocrine therapy is an important issue since such side-effects negatively affect health-related quality of life and adherences to therapy. These authors argue that adverse events constitute the main reason for non-adherence to endocrine treatment, and across all adjuvant endocrine trials regardless of the treatment, vasomotor symptoms such as hot flushes are the most common side effects. Other frequently reported side-effects such as vaginal discharge, vaginal dryness, dyspareunia, and arthralgia vary in prevalence between tamoxifen and aromatase inhibitors [##UREF##45##290##]. It has been recommended that currently in assessing quality of life in breast cancer patients priorities should be given to cognitive functioning, menopausal symptoms, body image and long-term effects of new therapies that might cause musculoskeletal and neurological side-effects [##REF##18028021##35##]. In addition, sexual functioning seems important area that needs more attention, especially for younger breast cancer survivors. It is argued that younger survivors may need interventions that specifically target their needs related to menopausal symptoms and problems with relationships, sexual functioning and body image [##REF##15908646##291##].</p>", "<p>There were few qualitative studies. Since these could provide more insight into quality of life in breast cancer patients, we need more such studies to collect data and indicate how breast cancer patients interpret life after diagnosis and during and after treatment. Breast cancer survivors even might rate their quality of life more favorably than outpatients with other common medical conditions and identify many positive aspects from the cancer experience [##REF##8861837##180##]. However, it is not only the study of quality of life in newly diagnosed breast cancer patients that is necessary; studying quality of life in long-term survivors is equally important. As suggested, when assessing quality of life in breast cancer patients, the stage of disease should also be considered. There are differences in quality of life between patients with non-invasive breast cancer, newly diagnosed breast cancer and advanced local breast cancer, and disease-free breast cancer survivors, women with recurrence breast cancer, and women with advanced metastatic breast cancer [##UREF##46##292##].</p>" ]

[ "<title>Conclusion</title>", "<p>There was quite an extensive body of the literature on quality of life in breast cancer patients. These papers have made a considerable contribution to improving breast cancer care, although their exact benefit was hard to define. However, quality of life data provided scientific evidence for clinical decision-making and conveyed helpful information concerning breast cancer patients' experiences during the course of the disease diagnosis, treatment, disease-free survival time, and recurrences; otherwise finding patient-centered solutions for evidence-based selection of optimal treatments, psychosocial interventions, patient-physician communications, allocation of resources, and indicating research priorities were impossible. It seems that more qualitative research is needed for a better understanding of the topic. In addition, issues related to the disease, its treatment side effects and symptoms, and sexual functioning should receive more attention when studying quality of life in breast cancer patients.</p>" ]

[ "<p>This is an Open Access article distributed under the terms of the Creative Commons Attribution License (<ext-link ext-link-type=\"uri\" xlink:href=\"http://creativecommons.org/licenses/by/2.0\"/>), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.</p>", "<title>Background</title>", "<p>Quality of life in patients with breast cancer is an important outcome. This paper presents an extensive overview on the topic ranging from descriptive findings to clinical trials.</p>", "<title>Methods</title>", "<p>This was a bibliographic review of the literature covering all full publications that appeared in English language biomedical journals between 1974 and 2007. The search strategy included a combination of key words 'quality of life' and 'breast cancer' or 'breast carcinoma' in titles. A total of 971 citations were identified and after exclusion of duplicates, the abstracts of 606 citations were reviewed. Of these, meetings abstracts, editorials, brief commentaries, letters, errata and dissertation abstracts and papers that appeared online and were indexed ahead of publication were also excluded. The remaining 477 papers were examined. The major findings are summarized and presented under several headings: instruments used, validation studies, measurement issues, surgical treatment, systemic therapies, quality of life as predictor of survival, psychological distress, supportive care, symptoms and sexual functioning.</p>", "<title>Results</title>", "<p><italic>Instruments</italic>-Several valid instruments were used to measure quality of life in breast cancer patients. The European Organization for Research and Treatment of Cancer Core Cancer Quality of Life Questionnaire (EORTC QLQ-C30) and its breast cancer specific complementary measure (EORTC QLQ-BR23) and the Functional Assessment Chronic Illness Therapy General questionnaire (FACIT-G) and its breast cancer module (FACIT-B) were found to be the most common and well developed instruments to measure quality of life in breast cancer patients. <italic>Surgery</italic>-different surgical procedures led to relatively similar results in terms of quality of life assessments, although mastectomy patients compared to conserving surgery patients usually reported a lower body image and sexual functioning. <italic>Systemic therapies</italic>-almost all studies indicated that breast cancer patients receiving chemotherapy might experience several side-effects and symptoms that negatively affect their quality of life. Adjuvant hormonal therapies also were found to have similar negative impact on quality of life, although in general they were associated with improved survival. <italic>Quality of life as predictor of survival</italic>-similar to known medical factors, quality of life data in metastatic breast cancer patients was found to be prognostic and predictive of survival time. <italic>Psychological distress</italic>-anxiety and depression were found to be common among breast cancer patients even years after the disease diagnosis and treatment. Psychological factors also were found to predict subsequent quality of life or even overall survival in breast cancer patients. <italic>Supportive care</italic>-clinical treatments to control emesis, or interventions such as counseling, providing social support and exercise could improve quality of life. <italic>Symptoms</italic>-Pain, fatigue, arm morbidity and postmenopausal symptoms were among the most common symptoms reported by breast cancer patients. As recommended, recognition and management of these symptoms is an important issue since such symptoms impair health-related quality of life. <italic>Sexual functioning</italic>-breast cancer patients especially younger patients suffer from poor sexual functioning that negatively affect quality of life.</p>", "<title>Conclusion</title>", "<p>There was quite an extensive body of the literature on quality of life in breast cancer patients. These papers have made a considerable contribution to improving breast cancer care, although their exact benefit was hard to define. However, quality of life data provided scientific evidence for clinical decision-making and conveyed helpful information concerning breast cancer patients' experiences during the course of the disease diagnosis, treatment, disease-free survival time, and recurrences; otherwise finding patient-centered solutions for evidence-based selection of optimal treatments, psychosocial interventions, patient-physician communications, allocation of resources, and indicating research priorities were impossible. It seems that more qualitative research is needed for a better understanding of the topic. In addition, issues related to the disease, its treatment side effects and symptoms, and sexual functioning should receive more attention when studying quality of life in breast cancer patients.</p>" ]

[ "<title>Competing interests</title>", "<p>The author declares that they have no competing interests.</p>", "<title>Authors' contributions</title>", "<p>The author carried out this review and wrote the manuscript, and prepared all the tables and the additional file.</p>", "<title>Supplementary Material</title>" ]

[ "<title>Acknowledgements</title>", "<p>The author wishes to thanks Dr. Elena Elkin, Dr. Lonneke van de Poll-Franse, and Dr. Su Wilson for their helpful comments on early version of the manuscript and also Mrs. T. Rostami for her secretarial assistance. This was a piece of pure academic research work and the author did not receive any financial support or grant for the study.</p>" ]

[]

[ "<table-wrap position=\"float\" id=\"T1\"><label>Table 1</label><caption><p>Number of citations by year of publication (1974–2007)</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td align=\"left\"><bold>Year</bold></td><td align=\"left\"><bold>Breast cancer</bold></td><td align=\"left\"><bold>Quality of life</bold></td><td align=\"left\"><bold>BC+QOL*</bold></td><td align=\"left\"><bold>Papers reviewed**</bold></td></tr></thead><tbody><tr><td align=\"left\">1974</td><td align=\"left\">246</td><td align=\"left\">13</td><td align=\"left\">1</td><td align=\"left\">1</td></tr><tr><td align=\"left\">1975</td><td align=\"left\">312</td><td align=\"left\">23</td><td align=\"left\">0</td><td align=\"left\">0</td></tr><tr><td align=\"left\">1976</td><td align=\"left\">358</td><td align=\"left\">34</td><td align=\"left\">1</td><td align=\"left\">1</td></tr><tr><td align=\"left\">1977</td><td align=\"left\">522</td><td align=\"left\">27</td><td align=\"left\">0</td><td align=\"left\">0</td></tr><tr><td align=\"left\">1978</td><td align=\"left\">527</td><td align=\"left\">33</td><td align=\"left\">0</td><td align=\"left\">0</td></tr><tr><td align=\"left\">1979</td><td align=\"left\">489</td><td align=\"left\">34</td><td align=\"left\">0</td><td align=\"left\">0</td></tr><tr><td align=\"left\">1980</td><td align=\"left\">662</td><td align=\"left\">36</td><td align=\"left\">1</td><td align=\"left\">1</td></tr><tr><td align=\"left\">1981</td><td align=\"left\">634</td><td align=\"left\">45</td><td align=\"left\">1</td><td align=\"left\">0</td></tr><tr><td align=\"left\">1982</td><td align=\"left\">647</td><td align=\"left\">71</td><td align=\"left\">1</td><td align=\"left\">1</td></tr><tr><td align=\"left\">1983</td><td align=\"left\">661</td><td align=\"left\">89</td><td align=\"left\">2</td><td align=\"left\">2</td></tr><tr><td align=\"left\">1984</td><td align=\"left\">830</td><td align=\"left\">73</td><td align=\"left\">0</td><td align=\"left\">0</td></tr><tr><td align=\"left\">1985</td><td align=\"left\">844</td><td align=\"left\">97</td><td align=\"left\">2</td><td align=\"left\">2</td></tr><tr><td align=\"left\">1986</td><td align=\"left\">920</td><td align=\"left\">134</td><td align=\"left\">1</td><td align=\"left\">1</td></tr><tr><td align=\"left\">1987</td><td align=\"left\">961</td><td align=\"left\">211</td><td align=\"left\">2</td><td align=\"left\">2</td></tr><tr><td align=\"left\">1988</td><td align=\"left\">1125</td><td align=\"left\">223</td><td align=\"left\">2</td><td align=\"left\">2</td></tr><tr><td align=\"left\">1989</td><td align=\"left\">1333</td><td align=\"left\">294</td><td align=\"left\">2</td><td align=\"left\">2</td></tr><tr><td align=\"left\">1990</td><td align=\"left\">1470</td><td align=\"left\">422</td><td align=\"left\">7</td><td align=\"left\">6</td></tr><tr><td align=\"left\">1991</td><td align=\"left\">1423</td><td align=\"left\">394</td><td align=\"left\">8</td><td align=\"left\">7</td></tr><tr><td align=\"left\">1992</td><td align=\"left\">1805</td><td align=\"left\">603</td><td align=\"left\">8</td><td align=\"left\">8</td></tr><tr><td align=\"left\">1993</td><td align=\"left\">2088</td><td align=\"left\">641</td><td align=\"left\">18</td><td align=\"left\">17</td></tr><tr><td align=\"left\">1994</td><td align=\"left\">2342</td><td align=\"left\">747</td><td align=\"left\">16</td><td align=\"left\">15</td></tr><tr><td align=\"left\">1995</td><td align=\"left\">2444</td><td align=\"left\">948</td><td align=\"left\">11</td><td align=\"left\">10</td></tr><tr><td align=\"left\">1996</td><td align=\"left\">2926</td><td align=\"left\">1422</td><td align=\"left\">16</td><td align=\"left\">15</td></tr><tr><td align=\"left\">1997</td><td align=\"left\">3249</td><td align=\"left\">1756</td><td align=\"left\">19</td><td align=\"left\">16</td></tr><tr><td align=\"left\">1998</td><td align=\"left\">3597</td><td align=\"left\">2049</td><td align=\"left\">29</td><td align=\"left\">25</td></tr><tr><td align=\"left\">1999</td><td align=\"left\">3872</td><td align=\"left\">2457</td><td align=\"left\">39</td><td align=\"left\">30</td></tr><tr><td align=\"left\">2000</td><td align=\"left\">5026</td><td align=\"left\">2639</td><td align=\"left\">37</td><td align=\"left\">30</td></tr><tr><td align=\"left\">2001</td><td align=\"left\">5206</td><td align=\"left\">2985</td><td align=\"left\">34</td><td align=\"left\">27</td></tr><tr><td align=\"left\">2002</td><td align=\"left\">5720</td><td align=\"left\">3233</td><td align=\"left\">42</td><td align=\"left\">26</td></tr><tr><td align=\"left\">2003</td><td align=\"left\">6441</td><td align=\"left\">3900</td><td align=\"left\">38</td><td align=\"left\">31</td></tr><tr><td align=\"left\">2004</td><td align=\"left\">7422</td><td align=\"left\">4811</td><td align=\"left\">74</td><td align=\"left\">47</td></tr><tr><td align=\"left\">2005</td><td align=\"left\">7862</td><td align=\"left\">5276</td><td align=\"left\">73</td><td align=\"left\">53</td></tr><tr><td align=\"left\">2006</td><td align=\"left\">7021</td><td align=\"left\">4592</td><td align=\"left\">63</td><td align=\"left\">48</td></tr><tr><td align=\"left\">2007</td><td align=\"left\">4641</td><td align=\"left\">2207</td><td align=\"left\">58</td><td align=\"left\">51</td></tr><tr><td colspan=\"5\"><hr/></td></tr><tr><td align=\"left\"><bold>Total</bold></td><td align=\"left\">85626</td><td align=\"left\">42519</td><td align=\"left\">606</td><td align=\"left\">477</td></tr></tbody></table></table-wrap>", "<table-wrap position=\"float\" id=\"T2\"><label>Table 2</label><caption><p>A list of some overview papers on quality of life in breast cancer patients (1974–2007)</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td align=\"left\"><bold>Author(s) [Ref.]</bold></td><td align=\"left\"><bold>Year</bold></td><td align=\"left\"><bold>Main focus</bold></td><td align=\"left\"><bold>Conclusion(s)</bold></td></tr></thead><tbody><tr><td align=\"left\">McEvoy and McCorkle [##REF##2205372##7##]</td><td align=\"left\">1990</td><td align=\"left\">QOL in advanced breast cancer</td><td align=\"left\">Efforts to manage advanced breast cancer must include both current medical therapies and attention to the critical factors associated with enhancing their QOL.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Kiebert et al. [##REF##2033420##8##]</td><td align=\"left\">1991</td><td align=\"left\">Impact of breast conserving surgery vs. mastectomy on QOL</td><td align=\"left\">There were no substantial differences between the two treatment modalities except for body image and sexual functioning in favor of breast conserving surgery.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Aarenson [##REF##8502817##9##]</td><td align=\"left\">1993</td><td align=\"left\">Assessments of QOL and benefits from adjuvant therapies</td><td align=\"left\">Adjuvant therapies could improve QOL in breast cancer patients.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Bryson and Plosker [##UREF##3##10##]</td><td align=\"left\">1993</td><td align=\"left\">Tamoxifen as adjuvant therapy</td><td align=\"left\">Tamoxifen has a low cost-utility ratio in postmenopausal women with node-positive, estrogen receptor-positive breast cancer.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Stefanek [##REF##7827169##11##]</td><td align=\"left\">1994</td><td align=\"left\">QOL research, provider-patient communication, and psychological distress of spouses and other relatives of breast cancer patients</td><td align=\"left\">This review summarizes and critiques publications in three identified areas.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Ganz [##REF##10886990##12##]</td><td align=\"left\">1994</td><td align=\"left\">Review of various approaches to the measurement of QOL, the important QOL issues in the treatment of breast cancer, and what is known about QOL of older women with breast cancer</td><td align=\"left\">Ongoing and future research using newer approaches to QOL assessment should provide additional information on this important topic.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Osoba [##REF##7841968##13##]</td><td align=\"left\">1994</td><td align=\"left\">QOL as a treatment endpoint</td><td align=\"left\">Advances in understanding HRQOL in metastatic breast cancer will aid the development of rational treatment policies.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Carlson [##REF##9556780##14##]</td><td align=\"left\">1998</td><td align=\"left\">QOL in metastatic breast cancer</td><td align=\"left\">Clinician must balance anti-tumor activity, performance status, and the usual toxicity measures as surrogates for QOL associated with each specific therapy.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Leedham and Ganz [##REF##10370362##15##]</td><td align=\"left\">1999</td><td align=\"left\">Psychological concerns and mental health</td><td align=\"left\">Psychological concerns and mental health are important issues for breast cancer patients and should be recognized and treated when necessary.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Rustoen and Begnum [##REF##11128120##16##]</td><td align=\"left\">2000</td><td align=\"left\">Nursing practice</td><td align=\"left\">Nurses play an important role in meeting the needs of breast cancer patients.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Shapiro et al. [##REF##11255204##17##]</td><td align=\"left\">2001</td><td align=\"left\">Relationship between psychosocial variables and QOL</td><td align=\"left\">A broader, more integrative framework that includes psychosocial factors is needed to evaluate breast cancer consequences.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Partridge et al. [##REF##15687635##18##]</td><td align=\"left\">2001</td><td align=\"left\">QOL before, during and after high-dose chemotherapy</td><td align=\"left\">Resulting transient impaired overall QOL with subsequent improvement over time.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Kurtz and Dufour [##REF##12207554##19##]</td><td align=\"left\">2002</td><td align=\"left\">QOL in older patients with metastatic disease receiving either standard treatment or new drugs</td><td align=\"left\">Aromatase inhibitors (such as taxanes and orally administered chemotherapy) provide similar or a better QOL as compared to first line endocrine therapy with tamoxifen.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Costantino [##REF##12117073##20##]</td><td align=\"left\">2002</td><td align=\"left\">Hormonal treatments in metastatic breast cancer patients</td><td align=\"left\">QOL data is useful for both clinicians and patients in evaluating treatment options and developing treatment strategies.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Fallowfield [##REF##15590318##21##]</td><td align=\"left\">2004</td><td align=\"left\">Hormonal therapies</td><td align=\"left\">Tolerability profiles of available treatment options are highlighted.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Sammarco [##UREF##4##22##]</td><td align=\"left\">2004</td><td align=\"left\">QOL of older breast cancer patients</td><td align=\"left\">Outpatient and long-term care should become a key setting for implementation of QOL interventions for women with breast cancer.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Knobf [##REF##16476831##23##]</td><td align=\"left\">2006</td><td align=\"left\">Endocrine effects of adjuvant therapy in younger survivors</td><td align=\"left\">Causes premature menopause that is associated with poorer QOL, decreased sexual functioning, menopausal symptom distress, psychosocial distress related to infertility, and infertility.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Kayl and Meyers [##UREF##5##24##]</td><td align=\"left\">2006</td><td align=\"left\">Side effects of chemotherapy</td><td align=\"left\">QOL issues may help to guide patient-care decision.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Diel [##REF##17393190##25##]</td><td align=\"left\">2007</td><td align=\"left\">Effectiveness of bisphosphonates on bone pain and quality of life in breast cancer patients with metastatic bone disease</td><td align=\"left\">Clinical trial data demonstrate that bisphosphonates offer significant and sustained relief from bone pain and can also improve quality of life in patients with metastatic breast cancer. New treatment schedules using high dose bisphosphonates can offer rapid relief of acute, and severe bone pain.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Rozenberg et al. [##REF##17540137##26##]</td><td align=\"left\">2007</td><td align=\"left\">Co-morbid conditions and breast cancer</td><td align=\"left\">Women with breast cancer and three or more co-morbid conditions have a 20-fold higher rate of mortality from causes other than breast cancer and a 4-fold higher rate of all-cause mortality when compared with patients who have none.</td></tr></tbody></table></table-wrap>", "<table-wrap position=\"float\" id=\"T3\"><label>Table 3</label><caption><p>A list of systematic reviews on different aspects of quality of life in breast cancer patients (1974–2006)</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td align=\"left\"><bold>Author(s) [Ref.]</bold></td><td align=\"left\"><bold>Year</bold></td><td align=\"left\"><bold>Main focus</bold></td><td align=\"left\"><bold>Conclusion(s)</bold></td></tr></thead><tbody><tr><td align=\"left\">Irwig and Bennetts [##REF##9396988##27##]</td><td align=\"left\">1997</td><td align=\"left\">A systematic review of quality of life after breast conservation or mastectomy</td><td align=\"left\">Apart body image it is unclear whether breast conservation or mastectomy results in better psychosocial outcomes.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Bottomley and Therasse [##REF##12372724##28##]</td><td align=\"left\">2002</td><td align=\"left\">Systemic therapy (chemotherapy, hormonal therapy, or biological therapy) in advanced breast cancer (1995–2001)</td><td align=\"left\">QOL data provide invaluable insights into the treatment and care of patients.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Shimozuma et al. [##REF##12185329##29##]</td><td align=\"left\">2002</td><td align=\"left\">Systematic overview of the literature (1982–1999)</td><td align=\"left\">To date there have been almost no appropriate systematic overviews or guidelines issued for QOL assessment studies related to breast cancer.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Goodwin et al. [##REF##12591983##30##]</td><td align=\"left\">2003</td><td align=\"left\">Randomized clinical trials of treatment (review of literature from 1980–2001)</td><td align=\"left\">Until results of ongoing trials in breast cancer are available, caution is recommended in initiating new QOL studies unless treatment equivalency is expected or unless unique or specific issues can be addressed.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Rietman et al. [##REF##12657232##31##]</td><td align=\"left\">2003</td><td align=\"left\">Late morbidity of breast cancer (review of literature from 1980 to 2000)</td><td align=\"left\">Significant relationship between late morbidity and restrictions of daily activities and poorer QOL was reported.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Payne et al. [##REF##12491494##32##]</td><td align=\"left\">2003</td><td align=\"left\">Racial disparities in the palliative care for African-American (review of literature from 1985 to 2000)</td><td align=\"left\">Differences in treatment patterns, pain management, and hospice care exist between African-American and other ethnic groups.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Fossati [##REF##15528966##33##]</td><td align=\"left\">2004</td><td align=\"left\">Randomized clinical trials of cytotoxic or hormonal treatments in advanced breast cancer (review of published literature before Dec 2003</td><td align=\"left\">QOL assessments added relatively little value to classical clinical endpoints.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Mols et al. [##REF##16226458##34##]</td><td align=\"left\">2005</td><td align=\"left\">Systematic review among long-term survivors</td><td align=\"left\">Focusing on the long-term effects of breast cancer is important when evaluating the full extent of cancer treatment.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Grimison and Stockler [##REF##18028021##35##]</td><td align=\"left\">2007</td><td align=\"left\">Adjuvant systemic therapy for early-stage breast cancer (review of literature from 1996 to Feb. 2007)</td><td align=\"left\">For the majority of breast cancer patients most aspects of health-related quality of life recover after adjuvant chemotherapy ends without long-term effects except vasomotor symptoms and sexual dysfunction.</td></tr></tbody></table></table-wrap>", "<table-wrap position=\"float\" id=\"T4\"><label>Table 4</label><caption><p>A list of instruments used to measure quality of life in breast cancer patients (1974–2007)</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td align=\"left\"><bold>Types of measures</bold></td><td align=\"left\"><bold>Measures full name</bold></td><td align=\"left\"><bold>Abbreviation</bold></td></tr></thead><tbody><tr><td align=\"left\"><italic>General measures</italic></td><td/><td/></tr><tr><td/><td align=\"left\">Short Form Health Survey</td><td align=\"left\">SF-36</td></tr><tr><td/><td align=\"left\">Spitzer Quality of Life Index</td><td align=\"left\">QLI</td></tr><tr><td/><td align=\"left\">Sickness Impact Profile</td><td align=\"left\">SIP</td></tr><tr><td/><td align=\"left\">Ferrans and Powers Quality of Life Index</td><td align=\"left\">QLI</td></tr><tr><td align=\"left\"><italic>Cancer specific measures</italic></td><td/><td/></tr><tr><td/><td align=\"left\">European Organization for Research and Treatment of Cancer Core quality of Life questionnaire</td><td align=\"left\">EORTC QLQ-C30</td></tr><tr><td/><td align=\"left\">Functional Assessment of Chronic Illness Therapy General Questionnaire</td><td align=\"left\">FACIT-G (formerly FACT)</td></tr><tr><td/><td align=\"left\">Functional Living Index-Cancer</td><td align=\"left\">FLI-C</td></tr><tr><td/><td align=\"left\">Ferrans and Powers Quality of Life Index-Cancer</td><td align=\"left\">QLI-C</td></tr><tr><td align=\"left\"><italic>Breast cancer specific measures</italic></td><td/><td/></tr><tr><td/><td align=\"left\">European Organization for Research and Treatment of Cancer Breast Cancer Quality of Life Questionnaire</td><td align=\"left\">EORTC QLQ-BR23</td></tr><tr><td/><td align=\"left\">Functional Assessment of Chronic Illness Therapy-Breast</td><td align=\"left\">FCIT-B</td></tr><tr><td/><td align=\"left\">Breast Cancer Chemotherapy Questionnaire</td><td align=\"left\">BCQ</td></tr><tr><td/><td align=\"left\">The Satisfaction with Life Domains Scale for Breast Cancer</td><td align=\"left\">SLDS-BC</td></tr><tr><td align=\"left\"><italic>Psychological measures</italic></td><td/><td/></tr><tr><td/><td align=\"left\">General Health Questionnaire-28</td><td align=\"left\">GHQ-28</td></tr><tr><td/><td align=\"left\">Hospital Anxiety and Depression Scale</td><td align=\"left\">HADS</td></tr><tr><td/><td align=\"left\">Beck Depression Inventory</td><td align=\"left\">BDI</td></tr><tr><td/><td align=\"left\">Center for Epidemiologic Studies Depression Scale</td><td align=\"left\">CES-D</td></tr><tr><td/><td align=\"left\">State-Trait Anxiety Inventory</td><td align=\"left\">STAI</td></tr><tr><td/><td align=\"left\">Profile Mood State</td><td align=\"left\">PMS</td></tr><tr><td/><td align=\"left\">Mental Adjustment to Cancer Scale</td><td align=\"left\">MACS</td></tr><tr><td/><td align=\"left\">Psychosocial Adjustment to Illness Scale</td><td align=\"left\">PAIS</td></tr><tr><td align=\"left\"><italic>Symptom measures</italic></td><td/><td/></tr><tr><td/><td align=\"left\">Functional Assessment of Chronic Illness Therapy-Fatigue</td><td align=\"left\">FACIT-F</td></tr><tr><td/><td align=\"left\">Piper Fatigue Scale</td><td align=\"left\">PFS</td></tr><tr><td/><td align=\"left\">Multidimensional Fatigue Inventory</td><td align=\"left\">MFI</td></tr><tr><td/><td align=\"left\">Functional Assessment of Chronic Illness Therapy-B plus Arm Morbidity Subscale</td><td align=\"left\">FACIT-B + 4</td></tr><tr><td/><td align=\"left\">Hot Flash Related Interference Scale</td><td align=\"left\">HFRDIS</td></tr><tr><td/><td align=\"left\">Shoulder Disability Questionnaire</td><td align=\"left\">SDQ</td></tr><tr><td/><td align=\"left\">Brief Pain Inventory</td><td align=\"left\">BPI</td></tr><tr><td/><td align=\"left\">McGill Pain Questionnaire</td><td align=\"left\">MPQ</td></tr><tr><td/><td align=\"left\">Memorial Symptom Assessment Scale</td><td align=\"left\">MSAS</td></tr><tr><td/><td align=\"left\">Rotterdam Symptom Checklist</td><td align=\"left\">RSC</td></tr><tr><td align=\"left\"><italic>Other measures</italic></td><td/><td/></tr><tr><td/><td align=\"left\">Functional Assessment of Chronic Illness Therapy-Spiritual</td><td align=\"left\">FACIT-SP</td></tr><tr><td/><td align=\"left\">Body Image Scale</td><td align=\"left\">BIS</td></tr><tr><td/><td align=\"left\">Body Image After Breast Cancer Questionnaire</td><td align=\"left\">BIBCQ</td></tr><tr><td/><td align=\"left\">Watts Sexual Functioning Questionnaire</td><td align=\"left\">WSFQ</td></tr><tr><td/><td align=\"left\">Social Support Questionnaire</td><td align=\"left\">SSQ</td></tr><tr><td/><td align=\"left\">Life Satisfaction Questionnaire</td><td align=\"left\">LSQ</td></tr><tr><td/><td align=\"left\">Satisfaction With Life Scale</td><td align=\"left\">SWLS</td></tr></tbody></table></table-wrap>", "<table-wrap position=\"float\" id=\"T5\"><label>Table 5</label><caption><p>A summary of validation studies of quality of life instruments in breast cancer patients (1974–2007)</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td align=\"left\"><bold>Author(s) [Ref.]</bold></td><td align=\"left\"><bold>Year</bold></td><td align=\"left\"><bold>Instrument</bold></td><td align=\"left\"><bold>Main focus</bold></td></tr></thead><tbody><tr><td align=\"left\">Levine et al. [##REF##3058874##38##]</td><td align=\"left\">1988</td><td align=\"left\">The Breast Cancer Chemotherapy Questionnaire (BCQ)</td><td align=\"left\">Development an outcome measure in clinical trials of adjuvant chemotherapy</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Ciampi et al. [##UREF##6##39##]</td><td align=\"left\">1988</td><td align=\"left\">A 27 item Linear Analog Self Assessment</td><td align=\"left\">Factor analysis indicating disease and treatment-related, physical, emotional and social health summary scores</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Tamburini et al. [##REF##1768628##40##]</td><td align=\"left\">1991</td><td align=\"left\">Two simple index</td><td align=\"left\">To assess the impact of therapy on QOL in patients receiving chemotherapy for operable breast cancer</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Osoba et al. [##REF##8853525##41##]</td><td align=\"left\">1994</td><td align=\"left\">The European Organization for Research and Treatment of Cancer Core Quality of Life Questionnaire (EORTC QLQ-C30)</td><td align=\"left\">Evaluation of psychometric properties and responsiveness</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Carlsson and Hamrin [##REF##8998495##42##]</td><td align=\"left\">1996</td><td align=\"left\">The Life Satisfaction Questionnaire (LSQ-32)</td><td align=\"left\">Development a tool to measure life satisfaction in breast cancer patients</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Sprangers et al. [##REF##8874337##43##]</td><td align=\"left\">1996</td><td align=\"left\">The European Organization for Research and Treatment of Cancer Breast Cancer Specific Quality of Life Questionnaire (EORTC QLQ-BR23)</td><td align=\"left\">Development of a breast cancer specific QOL measure</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Brady et al. [##REF##9060536##44##]</td><td align=\"left\">1997</td><td align=\"left\">The Functional Assessment of Cancer Therapy Breast Cancer Specific Questionnaire (FACT-B)</td><td align=\"left\">Development of a breast cancer specific QOL measure</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">de Haes and Olschewski [##REF##9739441##45##]</td><td align=\"left\">1998</td><td align=\"left\">The Rotterdam Symptom Checklist (RSC)</td><td align=\"left\">Cross cultural validation</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">McLachlan et al. [##REF##9713301##46##]</td><td align=\"left\">1998</td><td align=\"left\">The EORTC QLQ-C30</td><td align=\"left\">Validation as a measure of psychological function</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Fallowfiled et al [##REF##10481946##47##]</td><td align=\"left\">1999</td><td align=\"left\">An endocrine symptom subscale for the FACT-B (FACT-B plus ES)</td><td align=\"left\">Validation in women undergoing hormonal therapy for breast cancer</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Montazeri et al. [##REF##10983481##48##]</td><td align=\"left\">2000</td><td align=\"left\">The EORTC QLQ-BR23</td><td align=\"left\">Validation of the Iranian version</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Mihailova et al. [##UREF##7##49##]</td><td align=\"left\">2001</td><td align=\"left\">The EORTC QLQ-C30 and the QLQ-BR23</td><td align=\"left\">Validation of the Bulgarian version</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Coster et al. [##REF##11727963##50##]</td><td align=\"left\">2001</td><td align=\"left\">The Impact of Arm Morbidity (FACT-B+4)</td><td align=\"left\">Development a QOL scale to assess the impact of arm morbidity post-operatively</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Carpenter [##UREF##8##51##]</td><td align=\"left\">2001</td><td align=\"left\">The Hot Flash Related Daily Interference Scale</td><td align=\"left\">Development of a tool for measuring the impact of hot flashes on QOL</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Pandey et al. [##REF##12018741##52##]</td><td align=\"left\">2002</td><td align=\"left\">The FACT Breast Cancer Specific Questionnaire (FACT-B)</td><td align=\"left\">Validation of the Malayalam version</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Chie et al. [##UREF##9##53##]</td><td align=\"left\">2003</td><td align=\"left\">The EORTC QLQ-C30 and the EORTC QLQ-BR23</td><td align=\"left\">Validation of the Taiwan Chinese version</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Lee et al. [##REF##15342666##54##]</td><td align=\"left\">2004</td><td align=\"left\">The Functional Assessment of Cancer Therapy-General (FACT-G)</td><td align=\"left\">Validation of the Korean version</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Yun et al. [##REF##15088137##55##]</td><td align=\"left\">2004</td><td align=\"left\">The EORTC QLQ-BR23</td><td align=\"left\">Cross-cultural application in Korea</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Parmar et al. [##REF##16252544##56##]</td><td align=\"left\">2005</td><td align=\"left\">The EORTC QLQ-C30</td><td align=\"left\">Validation of the Indian version</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Avis and Foley [##REF##17140438##57##]</td><td align=\"left\">2006</td><td align=\"left\">The Quality of life in Adult Cancer Survivors (QLACS)</td><td align=\"left\">Evaluation in long term breast cancer survivors</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Wan et al. [##REF##17377841##58##]</td><td align=\"left\">2007</td><td align=\"left\">The FACT-B</td><td align=\"left\">Validation of the simplified Chinese version</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Wan et al. [##REF##17221159##59##]</td><td align=\"left\">2007</td><td align=\"left\">The EORTC QLQ-BR53</td><td align=\"left\">Psychometric properties of the simplified Chinese version</td></tr></tbody></table></table-wrap>", "<table-wrap position=\"float\" id=\"T6\"><label>Table 6</label><caption><p>A list of quality of life studies that covered measurement issues in breast cancer patients (1974–2007)</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td align=\"left\"><bold>Author(s) [Ref.]</bold></td><td align=\"left\"><bold>Year</bold></td><td align=\"left\"><bold>Main focus</bold></td><td align=\"left\"><bold>Conclusion(s)/Recommendation</bold></td></tr></thead><tbody><tr><td align=\"left\">Baum et al. [##REF##2194541##60##]</td><td align=\"left\">1990</td><td align=\"left\">The issue of measuring QOL in advanced breast cancer</td><td align=\"left\">Efforts are being made to find out ways to measure QOL in advanced breast cancer patients.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Sutherland et al. [##REF##2370573##61##]</td><td align=\"left\">1990</td><td align=\"left\">Ratings of the importance of QOL variables</td><td align=\"left\">Breast cancer patients give different weights to different QOL variables.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Gelber et al. [##REF##1385719##62##]</td><td align=\"left\">1992</td><td align=\"left\">Explaining about the QOL adjusted Time Without Symptom and Toxicity</td><td align=\"left\">Integration of two methods (QOL and symptom free duration) could provide a new tool.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Ganz et al. [##REF##1588353##63##]</td><td align=\"left\">1992</td><td align=\"left\">The influence of multiple variables on the relationship of age to QOL</td><td align=\"left\">The casement plot methodology should be employed for simultaneous evaluation of multiple variables.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Gelber et al. [##REF##8119064##64##]</td><td align=\"left\">1993</td><td align=\"left\">Description of survival estimates with applications to QOL evaluation (Quality adjusted Time Without Symptoms of disease and Toxicity of treatment)</td><td align=\"left\">Estimation showed that patients continued to benefit greatly from long-term-duration chemotherapy between 5 and 10 years following treatment.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Hyden et al. [##REF##8265446##65##]</td><td align=\"left\">1993</td><td align=\"left\">Pitfalls in collecting QOL data</td><td align=\"left\">Several recommendations were made: (a) build support for QOL assessment among the group's leadership, (b) involve physicians and oncology nurses in the study design, (c) identify a QOL liaison at each participating institution, and (d) aggressively monitor the quality and timeliness of data submission.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Fallowfield [##REF##8423565##66##]</td><td align=\"left\">1993</td><td align=\"left\">Measurement issues</td><td align=\"left\">Some recommendations for selecting well validated measures.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Gerard et al. [##REF##8433117##67##]</td><td align=\"left\">1993</td><td align=\"left\">Framing and labeling effects in measuring quality adjusted life years</td><td align=\"left\">A significant difference was found in the particular values of descriptions that were written in the third person that differed in terms of whether the word \"cancer\" was used.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Hurny et al. [##REF##8172795##68##]</td><td align=\"left\">1994</td><td align=\"left\">Timing of baseline QOL assessment</td><td align=\"left\">Timing is an important consideration in QOL assessment.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Fallowfield [##REF##7546841##69##]</td><td align=\"left\">1995</td><td align=\"left\">Discussion on some instruments used to measure QOL</td><td align=\"left\">Monitoring QOL in breast cancer should be a mandatory part of follow-up in clinical trials.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Hietanen [##REF##8619938##70##]</td><td align=\"left\">1996</td><td align=\"left\">Measurement and practical aspects of QOL assessment</td><td align=\"left\">Main factors affecting QOL in the treatment of breast cancer.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Bernhard et al. [##REF##9358933##71##]</td><td align=\"left\">1997</td><td align=\"left\">The International Breast Cancer Study Group (IBCSG) approach</td><td align=\"left\">Confirmation of the feasibility, validity and clinical relevance of quality of life assessment.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Bernhard et al. [##REF##9744512##72##]</td><td align=\"left\">1998</td><td align=\"left\">Factors affecting baseline QOL assessment</td><td align=\"left\">Cultural and biomedical factors are influencing baseline QOL data and should be considered when evaluating the impact of treatment.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Bernhard et al. [##REF##9549808##73##]</td><td align=\"left\">1998</td><td align=\"left\">Practical issues and factors associated with missing data</td><td align=\"left\">The factors most highly associated with missing data were institution and chemotherapy compliance.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Ganz et al. [##REF##9469334##74##]</td><td align=\"left\">1998</td><td align=\"left\">Compliance with QOL data collection</td><td align=\"left\">Educational level of a trial participants might contribute to it compliance.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Coates and Gebski [##UREF##10##75##]</td><td align=\"left\">1998</td><td align=\"left\">Approaches to missing data</td><td align=\"left\">Missing data cannot be assumed to be similar to those available. Optimal assessment requires careful prospective attention to complete data collection.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Jansen et al. [##REF##11236851##76##]</td><td align=\"left\">2000</td><td align=\"left\">Response shift</td><td align=\"left\">Significant recalibration effects were observed.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Curran et al. [##REF##10785587##77##]</td><td align=\"left\">2000</td><td align=\"left\">Summary measures and statistics</td><td align=\"left\">Different techniques in analysis might result in different conclusions.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Perez et al. [##UREF##11##78##]</td><td align=\"left\">2001</td><td align=\"left\">The application of a time trade-off utility measure</td><td align=\"left\">The utility measure and a QOL measure showed fair to moderate concordance.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Nagel et al. [##REF##11678311##79##]</td><td align=\"left\">2001</td><td align=\"left\">A cluster analytic approach to analyze quality of life data</td><td align=\"left\">QOL scores could identify clinically meaningful subgroups of patients.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Mosconi et al. [##REF##11804378##80##]</td><td align=\"left\">2001</td><td align=\"left\">A general introduction to the debate on the methodological issues involved in QOL evaluation</td><td align=\"left\">Open questions regarding the use of QOL measures in surgical, adjuvant therapy and metastatic studies.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Efficace et al. [##UREF##12##81##]</td><td align=\"left\">2002</td><td align=\"left\">Evaluating reliability, validity and cultural relevance of QOL measures in clinical trials</td><td align=\"left\">Suggestions for selecting future measures for use in breast cancer population of patients.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Wilson et al. [##REF##16117596##82##]</td><td align=\"left\">2005</td><td align=\"left\">Comparing two QOL measures (the Rand 36-item and the Functional Living Index-Cancer)</td><td align=\"left\">Neither questionnaire can be replaced by each other in studies of QOL in breast cancer patients.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Carver et al. [##UREF##13##83##]</td><td align=\"left\">2006</td><td align=\"left\">Assessment of demographic, medical and psychological variables on outcome</td><td align=\"left\">Different aspects of QOL at long-term follow-up had different antecedents.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Perry et al. [##REF##17474993##84##]</td><td align=\"left\">2007</td><td align=\"left\">Benefits, acceptability and utilization of QOL assessment in women with breast cancer</td><td align=\"left\">Summarized the benefits, challenges, and barriers of QOL measurement for female breast cancer patients.</td></tr></tbody></table></table-wrap>", "<table-wrap position=\"float\" id=\"T7\"><label>Table 7</label><caption><p>A list of studies of surgical treatment and quality o life in breast cancer patients (1974–2007)</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td align=\"left\"><bold>Author (s) [Ref.]</bold></td><td align=\"left\"><bold>Year</bold></td><td align=\"left\"><bold>Treatment (assessment time)</bold></td><td align=\"left\"><bold>Conclusion(s)</bold></td></tr></thead><tbody><tr><td align=\"left\">de Haes et al. [##REF##3881629##85##]</td><td align=\"left\">1985</td><td align=\"left\">MAS vs. tumorectomy (11 months after surgery)</td><td align=\"left\">No differences expect worse body image in MAS patients.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">de Haes et al. [##REF##3780987##86##]</td><td align=\"left\">1986</td><td align=\"left\">MAS vs. tumorectomy (11 and 18 months after surgery)</td><td align=\"left\">Overall QOL improved over time in both groups; poor body image in MAS.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Ganz et al. [##REF##1551058##87##]</td><td align=\"left\">1992</td><td align=\"left\">MAS vs. BCS after one year</td><td align=\"left\">No significant differences in QOL and both groups improved; BCS patients did not experience significantly better QOL but had fewer problems with clothing and body image.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Shimozuma et al. [##REF##11091520##88##]</td><td align=\"left\">1994</td><td align=\"left\">Surgery-any</td><td align=\"left\">Hospitalization had a strong negative relation to overall QOL; type of surgery had no significant association with QOL.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Neises et al. [##UREF##14##89##]</td><td align=\"left\">1994</td><td align=\"left\">MAS or BCS</td><td align=\"left\">Older women suffer as much as younger patients after MAS.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Fallowfield [##UREF##15##90##]</td><td align=\"left\">1994</td><td align=\"left\">Surgery and tamoxifen vs. tamoxifen alone</td><td align=\"left\">At 2 years similar psychological health; no evidence of impaired QOL for elderly women after surgery</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Shimozuma et al. [##REF##11091530##91##]</td><td align=\"left\">1995</td><td align=\"left\">MRM or BCS (before surgery and 3 times up 2 years after)</td><td align=\"left\">No significant differences in overall QOL; patients with BCS need more psychological support.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Hart et al. [##REF##9226025##92##]</td><td align=\"left\">1997</td><td align=\"left\">MAS + prostheses or MAS + reconstruction or MAS alone</td><td align=\"left\">No one technique is necessary for all women to optimize QOL; women should choose and make their own decisions.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Dorval et al. [##REF##9469332##93##]</td><td align=\"left\">1998</td><td align=\"left\">Partial or total MAS (3 and 18 months after)</td><td align=\"left\">Both appeared to be equivalent in long-term QOL. Younger women might benefit more from partial MAS.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Curran et al. [##REF##9640214##94##]</td><td align=\"left\">1998</td><td align=\"left\">MRM vs. BCS</td><td align=\"left\">Significant benefit in body image and satisfaction in BCS group; no difference in fear of recurrence.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Wapnir et al. [##REF##10379856##95##]</td><td align=\"left\">1999</td><td align=\"left\">Lumpectomy with axillary dissection (LAD) or mastectomy</td><td align=\"left\">No major differences except for dressing, comfort with nudity and sexual drive in favor of ALD.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Shimozuma et al. [##REF##10517342##96##]</td><td align=\"left\">1999</td><td align=\"left\">MRM or BCS (1 year after)</td><td align=\"left\">At one year good QOL, with no relationship to the type of surgery.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Pusic et al. [##REF##10513913##97##]</td><td align=\"left\">1999</td><td align=\"left\">Lumpectomy + irradiation or MAS + reconstruction or MAS alone</td><td align=\"left\">Postoperative QOL varied with age; for age less than 55 QOL was lowest for MAS, over 55 was lowest for lumpectomy.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Amichetti et al. [##REF##10424401##98##]</td><td align=\"left\">1999</td><td align=\"left\">BCS + irradiation in non-infiltrating breast cancer</td><td align=\"left\">Good QOL and body image and lack of negative impact on sexuality.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">King et al. [##REF##11297021##99##]</td><td align=\"left\">2000</td><td align=\"left\">MAS or BCS (3 months and 1 year after)</td><td align=\"left\">Most symptoms declined over time but arm and menopausal symptoms persisted; worse QOL in younger patients.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Kenny et al. [##REF##14731583##100##]</td><td align=\"left\">2000</td><td align=\"left\">MAS or BCS + irradiation (1 year after)</td><td align=\"left\">Better body image and physical function in BCS; more impact on younger women regardless of treatment type.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Nissen et al. [##REF##11283922##101##]</td><td align=\"left\">2001</td><td align=\"left\">MAS or MAS + reconstruction or BCS (6 times assessment up to 2 years after)</td><td align=\"left\">QOL other than body image were not better in BCS or MAS + reconstruction than in who had MAS alone; MAS + reconstruction was associated with greater mood disturbance and poorer QOL.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Janni et al. [##REF##11456055##102##]</td><td align=\"left\">2001</td><td align=\"left\">MAS or BCS (median 46 months follow-up)</td><td align=\"left\">Surgical modalities had no long-term impact on overall QOL, but certain body image related problems in MAS was observed.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Girotto et al. [##REF##12783001##103##]</td><td align=\"left\">2003</td><td align=\"left\">MAS + reconstruction in older women</td><td align=\"left\">Improved QOL in older patients especially improved mental health.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Cocquyt et al. [##REF##12890459##104##]</td><td align=\"left\">2003</td><td align=\"left\">Skin-sparing MAS or BCS</td><td align=\"left\">Both yielded comparable QOL, but cosmetic outcome was better after skin-sparing MAS.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Engel et al [##REF##15125749##105##]</td><td align=\"left\">2004</td><td align=\"left\">MAS or BCS (5 years follow-up)</td><td align=\"left\">MAS patients had lower body image, role and sexual functioning; BCS should be encouraged in all ages.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Ganz et al. [##REF##14996859##106##]</td><td align=\"left\">2004</td><td align=\"left\">Lumpectomy + chemotherapy or MAS + chemotherapy or Lumpectomy alone or MAS alone in non-metastatic breast cancer patients</td><td align=\"left\">At the end of primary treatment all treatment groups reported good emotional functioning but decreased physical health especially among women who had MAS or received chemotherapy.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Dubernard et al. [##UREF##16##107##]</td><td align=\"left\">2004</td><td align=\"left\">SLNB</td><td align=\"left\">Axillary procedure affected only QOL related to arm morbidity.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Elder et al. [##REF##15927829##108##]</td><td align=\"left\">2005</td><td align=\"left\">MAS + immediate breast reconstruction (before and 12 months after)</td><td align=\"left\">After 12 months good QOL comparable with aged-matched women from the general population.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Barranger et al. [##REF##16180226##109##]</td><td align=\"left\">2005</td><td align=\"left\">SLNB vs. ALND in breast-sparing treatment</td><td align=\"left\">SLNB was associated with significantly lower mid term morbidity.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Fleissig [##REF##16163445##110##]</td><td align=\"left\">2006</td><td align=\"left\">SLNB vs. ALND</td><td align=\"left\">Regarding arm functioning and QOL the use of SNB was recommended in patients with node negative breast cancer.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Pandey et al. [##REF##16887839##111##]</td><td align=\"left\">2006</td><td align=\"left\">MAS or BCS</td><td align=\"left\">No significant change in overall QOL after surgery; poorer QOL in MAS patients.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Rietman et al. [##REF##16387467##112##]</td><td align=\"left\">2006</td><td align=\"left\">SLNB or ALND (before and after 2 years)</td><td align=\"left\">Less treatment related upper limb morbidity, perceived disability in activities of daily life and worsening of QOL after SNLB compared with ALND.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Parker et al. [##REF##17574501##113##]</td><td align=\"left\">2007</td><td align=\"left\">MAS or MAS+ reconstruction or BCS (short- and long-term effects on aspects of psychosocial adjustment and QOL</td><td align=\"left\">Overall, the general patterns of psychosocial adjustment and QOL were similar among the three surgery groups.</td></tr></tbody></table></table-wrap>", "<table-wrap position=\"float\" id=\"T8\"><label>Table 8</label><caption><p>A list of studies on systemic therapies and quality of life in breast cancer patients (1974–2007)</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td align=\"left\"><bold>Author(s) [Ref.]</bold></td><td align=\"left\"><bold>Year</bold></td><td align=\"left\"><bold>Treatment/patients</bold></td><td align=\"left\"><bold>Conclusion(s)</bold></td></tr></thead><tbody><tr><td align=\"left\">Moore et al. [##REF##4136510##36##]</td><td align=\"left\">1974</td><td align=\"left\">Adrenalectomy + chemotherapy in advanced breast cancer</td><td align=\"left\">In most patients the subjective palliation involved a return to normal living.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Priestman and Baum [##REF##58161##37##]</td><td align=\"left\">1976</td><td align=\"left\">Chemotherapy in advanced breast cancer</td><td align=\"left\">Toxicity is not related to the patients' age and diminished with successive courses of drugs.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Palmer et al. [##REF##7004560##114##]</td><td align=\"left\">1980</td><td align=\"left\">A single agent vs. five drug combination in node positive primary breast cancer</td><td align=\"left\">Better QOL in single agent group.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Coates et al. [##REF##3683485##115##]</td><td align=\"left\">1987</td><td align=\"left\">Intermittent vs. continuous chemotherapy in metastatic breast cancer</td><td align=\"left\">Continuous chemotherapy was better; changes in the QOL were independent prognostic factor of survival.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Kiebert et al. [##REF##2148877##116##]</td><td align=\"left\">1990</td><td align=\"left\">Peri-operative chemotherapy vs. no chemotherapy in early stage breast cancer</td><td align=\"left\">No differences 1 year after; patients considered chemotherapy most burdensome aspect of treatment.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Gelber et al. [##REF##2003707##117##]</td><td align=\"left\">1991</td><td align=\"left\">Single cycle of combination chemotherapy vs. longer duration chemotherapy for pre-menopausal or chemo-endocrine therapy for postmenopausal women</td><td align=\"left\">Better QOL in longer duration chemotherapy or chemo-endocrine therapy.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Berglund et al. [##REF##1683557##118##]</td><td align=\"left\">1991</td><td align=\"left\">Late effects of adjuvant chemotherapy vs. postoperative radiotherapy in pre- and post-menopausal breast cancer</td><td align=\"left\">Chemotherapy patients had higher overall QOL.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Richards et al. [##REF##1627368##119##]</td><td align=\"left\">1992</td><td align=\"left\">A (weekly for 12 courses vs. every three weeks for 4 courses) in advanced breast cancer</td><td align=\"left\">Similar survival but higher psychological distress in the three weeks group.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Hurny et al. [##REF##1567662##120##]</td><td align=\"left\">1992</td><td align=\"left\">CMF (6 cycles vs. 3 cycles) in operable breast cancer</td><td align=\"left\">QOL improved with increasing time from the study entry.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Campora et al. [##REF##1383437##121##]</td><td align=\"left\">1992</td><td align=\"left\">Adjuvant chemotherapy vs. palliative chemotherapy in metastatic breast cancer</td><td align=\"left\">No significant difference between groups.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Fraser et al. [##REF##8431375##122##]</td><td align=\"left\">1993</td><td align=\"left\">CMF vs. E in advanced breast cancer</td><td align=\"left\">Similar survival and no significant difference in overall global QOL.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Twelves et al. [##REF##8142261##123##]</td><td align=\"left\">1994</td><td align=\"left\">Iododoxorubicin in advanced breast cancer</td><td align=\"left\">Little evidence of benefit in terms of physical symptom relief, level of activity, psychological symptoms or global QOL.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Bertsch and Donaldson. [##REF##7740333##124##]</td><td align=\"left\">1995</td><td align=\"left\">Vinorelbine vs. melphalan</td><td align=\"left\">Vinorelbine was better in some aspects of QOL.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Swain et al. [##REF##8622073##125##]</td><td align=\"left\">1996</td><td align=\"left\">AC + G-CSF in node positive breast cancer</td><td align=\"left\">Tolerable physical symptoms and emotional distress.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">McQuellon et al. [##REF##8879621##126##]</td><td align=\"left\">1996</td><td align=\"left\">High-dose chemotherapy + ABMT</td><td align=\"left\">No significant difference between pre- and post-treatment QOL.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Larsen et al. [##REF##8885485##127##]</td><td align=\"left\">1996</td><td align=\"left\">High-dose chemotherapy + ASCT</td><td align=\"left\">Resulting in poor physical and emotional health.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Hurny et al. [##REF##8622502##128##]</td><td align=\"left\">1996</td><td align=\"left\">6 cycles of CMF vs. 3 cycles CMF in node-positive operable breast cancer</td><td align=\"left\">Worse QOL during treatment but not after treatment completion.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Griffiths and Beaver [##UREF##17##129##]</td><td align=\"left\">1997</td><td align=\"left\">High-dose chemotherapy in advanced breast cancer</td><td align=\"left\">No significant deterioration in QOL.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Lindley et al. [##UREF##18##130##]</td><td align=\"left\">1998</td><td align=\"left\">Systemic adjuvant therapy</td><td align=\"left\">2–5 years after treatment good QOL. Small to modest gain was acceptable to women.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Ganz et al. [##REF##9928575##131##]</td><td align=\"left\">1998</td><td align=\"left\">TAM or chemotherapy alone or chemotherapy + TAM, or no adjuvant therapy</td><td align=\"left\">No significant differences in global QOL among treatment groups; those who received chemotherapy had more sexual problems and those who received TAM had more vasomotor symptoms.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Bernhard et al. [##REF##10533471##132##]</td><td align=\"left\">1999</td><td align=\"left\">Formestane vs. megestrol acetate in postmenopausal advanced breast cancer while on TAM</td><td align=\"left\">No significant difference in QOL; baseline QOL was strong predictive for QOL under treatment but not for time to treatment failure.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Fairclough et al. [##REF##10855346##133##]</td><td align=\"left\">1999</td><td align=\"left\">CAF vs. dose intensive a 16-week multi-drug regimen</td><td align=\"left\">Negative impact of the dose intensive 16-week regimen was observed, although Q-TwiST analysis showed a small gain for this regimen.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Osoba and Burchmore [##REF##10482198##134##]</td><td align=\"left\">1999</td><td align=\"left\">Trastuzumab (Hercptin) in metastatic breast cancer who may or may not have had prior chemotherapy</td><td align=\"left\">Trastuzumab was associated with an amelioration of the deleterious effects of chemotherapy alone; the drug was not associated with worsening of QOL.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">McLachlan et al. [##REF##10445420##135##]</td><td align=\"left\">1999</td><td align=\"left\">Chemotherapy in metastatic breast cancer</td><td align=\"left\">QOL maintained or improved; patients did not want to trade quantity for QOL.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Macquart-Moulin et al. [##REF##10673516##136##]</td><td align=\"left\">2000</td><td align=\"left\">High-dose chemotherapy + G-CSF + ASCT in inflammatory breast cancer</td><td align=\"left\">QOL deterioration disappeared after treatment and returned to baseline after one year.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Riccardi et al. [##REF##10717247##137##]</td><td align=\"left\">2000</td><td align=\"left\">Doubling E within FEC vs. FEC in metastatic breast cancer</td><td align=\"left\">No significant difference in response or improvement of baseline QOL.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Kramer et al. [##REF##10930796##138##,##REF##10930797##139##]</td><td align=\"left\">2000</td><td align=\"left\">Paclitaxel vs. A in advanced breast cancer</td><td align=\"left\">QOL appeared to be prognostic for survival and response to treatment.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Joly et al. [##REF##10944595##140##]</td><td align=\"left\">2000</td><td align=\"left\">CMF + irradiation vs. irradiation in pre-menopausal breast cancer</td><td align=\"left\">Similar QOL was observed.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Hakamies-Blomqvist et al. [##REF##10899655##141##]</td><td align=\"left\">2000</td><td align=\"left\">T vs. sequential MF in metastatic breast cancer</td><td align=\"left\">Difference in QOL was minor favoring MF.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Broeckel et al. [##REF##11016752##142##]</td><td align=\"left\">2000</td><td align=\"left\">Adjuvant chemotherapy treated breast cancer (after 3 to 36 months)</td><td align=\"left\">Younger age, unmarried status, time since diagnosis and chemotherapy completion related to greeter depressive symptoms.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Carlson et al. [##UREF##19##143##]</td><td align=\"left\">2001</td><td align=\"left\">High-dose chemotherapy + ASCT in metastatic breast cancer</td><td align=\"left\">Anxiety and depression continued to increase, loss of sexual interest, worrying and joint pain were reported.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Osoba et al. [##REF##12118024##144##]</td><td align=\"left\">2002</td><td align=\"left\">Chemotherapy + Trastuzumab (Hercptin) vs. Chemotherapy alone in metastatic breast cancer</td><td align=\"left\">More improved global QOL with chemotherapy + Herceptin.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Modi et al. [##REF##12202668##145##]</td><td align=\"left\">2002</td><td align=\"left\">Paclitaxel in metastatic breast cancer</td><td align=\"left\">QOL benefit in tumor response patients.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Heidemann et al [##REF##12419743##146##].</td><td align=\"left\">2002</td><td align=\"left\">Mitoxantrone vs. FEC in metastatic breast cancer</td><td align=\"left\">No significant difference in survival or response but a QOL scores favored mitoxantrone.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Genre et al. [##REF##11904787##147##]</td><td align=\"left\">2002</td><td align=\"left\">High-dose-intensity AC (21 vs. 14 days)</td><td align=\"left\">Shortening cycles had a high negative impact on QOL.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">de Haes et al. [##REF##14610048##148##]</td><td align=\"left\">2003</td><td align=\"left\">Goserelin vs. CMF in peri-and pre-menopausal node-positive early breast cancer</td><td align=\"left\">Better QOL in favor of goserelin.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Brandberg et al. [##REF##14512398##149##]</td><td align=\"left\">2003</td><td align=\"left\">Tailored FEC vs. induction FEC followed with high-dose CTCb + peripheral SCT</td><td align=\"left\">No significant overall differences were found between groups.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Land et al. [##REF##15319567##150##]</td><td align=\"left\">2004</td><td align=\"left\">CMF vs. AC in axillary node negative and estrogen receptor negative breast cancer</td><td align=\"left\">Overall QOL was equivalent between two groups.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Fallowfield et al. [##REF##15514369##151##]</td><td align=\"left\">2004</td><td align=\"left\">ANA vs. TAM alone or in combination in postmenopausal early breast cancer</td><td align=\"left\">Similar overall QOL impact but some small differences in side effects profiles.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Bottomely et al. [##REF##15226325##152##]</td><td align=\"left\">2004</td><td align=\"left\">AT vs. AC in metastatic breast cancer</td><td align=\"left\">No significant differences in QOL between two groups.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Bernhard et al. [##REF##15545973##153##]</td><td align=\"left\">2004</td><td align=\"left\">TAM for 5 years or three prior cycles of CMF followed by 57 months TAM in estrogen receptor-negative and estrogen receptor-positive breast cancer</td><td align=\"left\">At completion there were no differences by treatment groups.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Tong et al. [##REF##16037685##154##]</td><td align=\"left\">2005</td><td align=\"left\">Capecitabine, idarubicin and cyclophosphamide (all-oral regimen, XIC) in metastatic breast cancer</td><td align=\"left\">No significant decease in global QOL scores.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Galalae et al. [##UREF##20##155##]</td><td align=\"left\">2005</td><td align=\"left\">Radiotherapy and adjuvant chemotherapy vs. radiotherapy and hormonal therapy vs. radiotherapy alone after conserving surgery</td><td align=\"left\">Adjuvant chemotherapy lowered QOL vs. hormones or radiotherapy alone.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Elkin et al. [##REF##16184455##156##]</td><td align=\"left\">2005</td><td align=\"left\">Ovarian suppression vs. chemotherapy in pre-menopausal hormone-responsive breast cancer</td><td align=\"left\">Assuming equal efficacy ovarian suppression was superior. Efficacy would have impact on treatment choice.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Conner-Spady et al. [##REF##15937502##157##]</td><td align=\"left\">2005</td><td align=\"left\">High-dose chemotherapy + ABST in breast cancer with poor prognosis</td><td align=\"left\">Impaired QOL in short term but improved after 2 years.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Bottomley et al. [##REF##15863376##158##]</td><td align=\"left\">2005</td><td align=\"left\">Dose-intensives chemotherapy (CE + filgrastim) vs. CEF in locally advanced breast cancer</td><td align=\"left\">Groups did not differ in progression free survival; lower QOL in intensified group at short term but no difference at long term.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Ahles et al. [##REF##15994149##159##]</td><td align=\"left\">2005</td><td align=\"left\">Standard-dose systemic chemotherapy vs. local therapy only in long-term breast cancer survivors</td><td align=\"left\">Lower overall QOL in chemotherapy group.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Peppercorn et al. [##REF##16118805##160##]</td><td align=\"left\">2005</td><td align=\"left\">High-dose chemotherapy + ABMT vs. intermediate-dose chemotherapy in patients with stage II and III breast cancer</td><td align=\"left\">Patients who received more intensive therapy experienced transient declines in QOL; by 12 months after, QOL was comparable between the 2 arms, regardless of therapy intensity, and many QOL areas were improved from baseline.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Semiglazov et al. [##REF##16619567##161##]</td><td align=\"left\">2006</td><td align=\"left\">CMF + mistletoe lectin (PS76A2) vs. CMF + placebo</td><td align=\"left\">PS76A2 improved QOL during and after chemotherapy.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Martin et al. [##UREF##21##162##]</td><td align=\"left\">2006</td><td align=\"left\">FAC vs. TAC or TAC + G-CSF in node negative breast cancer</td><td align=\"left\">Lower QOL in patients treated with TAC. Addition of G-CSF improves QOL.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Hurria et al. [##UREF##22##163##]</td><td align=\"left\">2006</td><td align=\"left\">Anthracyclin-based chemotherapy or CMF in older women with breast cancer</td><td align=\"left\">QOL maintained in both group.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\" colspan=\"\">Fallowfield et al. [##REF##16484701##164##]</td><td align=\"left\">2006</td><td align=\"left\">EXE vs. TAM after 2–3 years of TAM in postmenopausal primary breast cancer</td><td align=\"left\">Temporary decrease in overall QOL for EXE but no other differences.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\" colspan=\"\">Groenvold et al. [##REF##16541325##165##]</td><td align=\"left\">2006</td><td align=\"left\">CMF vs. ovarian ablation</td><td align=\"left\">CMF had more negative impact on QOL.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\" colspan=\"\">Cella et al. [##REF##16944295##166##]</td><td align=\"left\">2006</td><td align=\"left\">ANA vs. TAM alone or in combination in postmenopausal breast cancer</td><td align=\"left\">ANA and TAM had similar impact on QOL.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\" colspan=\"\">Liu et al. [##REF##16823511##167##]</td><td align=\"left\">2006</td><td align=\"left\">DPPE + A vs. A in patients with advanced or metastatic breast cancer</td><td align=\"left\">Patients on A alone had fewer disease and treatment adverse events and better QOL.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\" colspan=\"\">Karamouzis et al. [##UREF##23##168##]</td><td align=\"left\">2007</td><td align=\"left\">Chemotherapy vs. supportive care in metastatic patients</td><td align=\"left\">QOL was better in patients receiving chemotherapy than those under supportive care.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\" colspan=\"\">Hopwood et al. [##REF##17236771##169##]</td><td align=\"left\">2007</td><td align=\"left\">Adjuvant radiotherapy</td><td align=\"left\">QOL and mental health were favorable for most patients about to start radiotherapy but younger age and receiving chemotherapy were significant risk factors for poorer QOL.</td></tr></tbody></table></table-wrap>", "<table-wrap position=\"float\" id=\"T9\"><label>Table 9</label><caption><p>A list of studies on psychological distress and quality of life in breast cancer patients (1974–2007)</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td align=\"left\"><bold>Author (s) [Ref.]</bold></td><td align=\"left\"><bold>Years</bold></td><td align=\"left\"><bold>Main focus</bold></td><td align=\"left\"><bold>Results/conclusion(s)</bold></td></tr></thead><tbody><tr><td align=\"left\">Ferrero et al. [##UREF##24##179##]</td><td align=\"left\">1994</td><td align=\"left\">Mental adjustment to cancer in newly-diagnosed non-mtastatic breast cancer(an xploratory study)</td><td align=\"left\">Strong association between mental adjustment to cancer and reported vague physical symptoms; fighting spirit and denial was associated with better QOL and helpless/hopeless and anxious preoccupation and fatalism were negatively correlated with well-being.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\" colspan=\"\">Ganz et al. [##REF##8861837##180##]</td><td align=\"left\">1996</td><td align=\"left\">Psychosocial concerns 2 and 3 years after primary treatment</td><td align=\"left\">Problems associated with physical and recreational activities, body image, and sexual functions were observed, although many positive aspects from cancer experience were reported.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\" colspan=\"\">Maunsell et al. [##REF##8874336##181##]</td><td align=\"left\">1996</td><td align=\"left\">Brief psychological intervention vs. Brief psychological intervention + psychological distress screening</td><td align=\"left\">Distress screening did not improve QOL. Minimal psychological intervention at initial treatment alone was recommended.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\" colspan=\"\">Andrykowski et al. [##REF##8699200##182##]</td><td align=\"left\">1996</td><td align=\"left\">Psychological adjustment in women with breast cancer or benign breast problems</td><td align=\"left\">Breast cancer patients reported poorer physical health but greater positive psychosocial adaptation and improved life outlook, no difference in psychological distress between two groups.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\" colspan=\"\">Marchioro et al. [##REF##8911127##183##]</td><td align=\"left\">1996</td><td align=\"left\">Evaluation of the impact of a psychological intervention vs. standard care in non-metastatic breast cancer patients</td><td align=\"left\">Cognitive psychotherapy and family counseling improved both depression and QOL indexes.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\" colspan=\"\">Weitzner et al. [##REF##9176972##184##]</td><td align=\"left\">1997</td><td align=\"left\">QOL and mood in long-term breast cancer survivors</td><td align=\"left\">Psychological measures were found to be more robust predictors of QOL than the demographic variables; long-term survivors continue to experience significant depression and lower QOL.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\" colspan=\"\">Kissane et al. [##REF##9734576##185##]</td><td align=\"left\">1998</td><td align=\"left\">Psychological morbidity in early-stage breast cancer</td><td align=\"left\">45% (135/303) had psychiatric disorder, 42% had depression, anxiety or both; QOL was substantially affected.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\" colspan=\"\">Bloom et al. [##UREF##25##186##]</td><td align=\"left\">1998</td><td align=\"left\">Intrusiveness of illness in young women with newly-diagnosed breast cancer</td><td align=\"left\">Intrusiveness of illness mediated the effect of disease and treatment factors on QOL; neither time post-diagnosis nor type of treatment affected the psychological component of QOL.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\" colspan=\"\">Longman et al. [##REF##10382190##187##]</td><td align=\"left\">1999</td><td align=\"left\">Psychological adjustment over time</td><td align=\"left\">Over time depression burden and anxiety burden persist and each was negatively associated with overall and present QOL.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\" colspan=\"\">Cotton et al. [##UREF##26##188##]</td><td align=\"left\">1999</td><td align=\"left\">Relationship among spiritual well-being, QOL, and psychological adjustment</td><td align=\"left\">Spiritual well-being was correlated with both QOL and psychological adjustment, but relationship was found to be more complex and indirect than previously considered.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\" colspan=\"\">Ashing-Giwa [##UREF##27##189##]</td><td align=\"left\">1999</td><td align=\"left\">Psychological outcome in long-term survivors of breast cancer (focus on African-American)</td><td align=\"left\">Patients relied on spiritual faith and family support to cope; socio-cultural contexts of the women's lives need to be considered when studying QOL.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\" colspan=\"\">Lewis et al. [##REF##11436544##190##]</td><td align=\"left\">2001</td><td align=\"left\">Cancer-related intrusive thoughts and social support</td><td align=\"left\">In women with social support cancer-related intrusive thoughts had no significant negative impact on QOL, but in women with low social support there was negative effect on QOL.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\" colspan=\"\">Amir and Ramati [##UREF##28##191##]</td><td align=\"left\">2002</td><td align=\"left\">Post-traumatic distress disorder (PTSD), QOL, and emotional distress in long term survivors of breast cancer and a control group</td><td align=\"left\">Higher PSTD, emotional distress and lower QOL in breast cancer mainly due to chemotherapy and disease stage.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\" colspan=\"\">Ganz et al. [##REF##14581426##192##]</td><td align=\"left\">2003</td><td align=\"left\">Psychosocial adjustment 15 months after diagnosis in older women with breast cancer</td><td align=\"left\">Psychosocial adjustment at 15 months was predicted by better mental health, emotional social support and better self-rated interaction with health care providers.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\" colspan=\"\">Bordeleau et al. [##REF##12743147##193##]</td><td align=\"left\">2003</td><td align=\"left\">Randomized trial of group psychological support vs. control in metastatic breast cancer</td><td align=\"left\">Supportive-expressive group therapy did not appear to influence QOL.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\" colspan=\"\">Badger et al. [##REF##14745853##194##]</td><td align=\"left\">2004</td><td align=\"left\">Depression burden and psychological adjustment</td><td align=\"left\">Depression burden had negative effect on psychological adjustment and QOL.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\" colspan=\"\">Schreier and Williams [##REF##14722597##195##]</td><td align=\"left\">2004</td><td align=\"left\">Anxiety in women receiving either radiation or chemotherapy for breast cancer</td><td align=\"left\">No significant differences for total QOL or any subscales by treatment; trait anxiety was higher for chemotherapy patients; state anxiety was high and did not decrease over the course of the treatment for either group.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\" colspan=\"\">Kershaw et al. [##UREF##29##196##]</td><td align=\"left\">2004</td><td align=\"left\">Coping strategies in advanced breast cancer patients and their family caregivers</td><td align=\"left\">Patients use more emotional support, religion and positive reframing strategies while family use more alcohol or drug. In both active coping was associated with higher QOL.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\" colspan=\"\">Lehto et al. [##UREF##30##197##]</td><td align=\"left\">2005</td><td align=\"left\">Psychological stress factors as predictors of QOL in patients receiving surgery alone vs. adjuvant treatment</td><td align=\"left\">Psychosocial factors were strongest predictors of QOL but not cancer type or treatment; non-cancer related stresses showed strongest QOL decreasing influence.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\" colspan=\"\">Roth et al. [##UREF##31##198##]</td><td align=\"left\">2005</td><td align=\"left\">Affective distress in women seeking immediate vs. delayed breast reconstruction after mastectomy</td><td align=\"left\">Women seeking immediate breast reconstruction showed relatively higher psychological impairment and physical disability.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\" colspan=\"\">Okamura et al. [##REF##15961434##199##]</td><td align=\"left\">2005</td><td align=\"left\">Psychiatric disorders and associated factors after first breast cancer recurrence</td><td align=\"left\">Patients' psychiatric disorders were associated with lower QOL.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\" colspan=\"\">Golden-Kreutz et al. [##REF##15898865##200##]</td><td align=\"left\">2005</td><td align=\"left\">Traumatic stress, perceived global stress, and life events</td><td align=\"left\">Initial stress at diagnosis predicted both psychological and physical health at follow-up.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\" colspan=\"\">Deshields et al. [##REF##15800768##201##]</td><td align=\"left\">2005</td><td align=\"left\">Emotional adjustment (at 4 points in time)</td><td align=\"left\">Primary psychological changes occur quickly after treatment conclusion and then it appeared to become stabled.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\" colspan=\"\">Laid law et al. [##UREF##32##202##]</td><td align=\"left\">2005</td><td align=\"left\">Self-hypnosis or Japanese healing or. control</td><td align=\"left\">Positive change in anxiety level, a general increase in mood and a better QOL were observed.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\" colspan=\"\">Schou et al. [##REF##16155769##203##]</td><td align=\"left\">2005</td><td align=\"left\">Dispositional optimism and QOL.</td><td align=\"left\">Optimism was predictive for better emotional and social functioning one year after surgery; at time of diagnosis and throughout post-diagnosis dispositional optimism was associated with better QOL and fewer symptoms.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\" colspan=\"\">Grabsch et al. [##REF##16889323##204##]</td><td align=\"left\">2006</td><td align=\"left\">Psychological morbidity in advanced breast cancer</td><td align=\"left\">42% (97/277) had a psychiatric disorder, 36% depression or anxiety or both. QOL was substantially affected.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\" colspan=\"\">Antoni et al. [##REF##17154743##205##]</td><td align=\"left\">2006</td><td align=\"left\">Stress management after treatment for breast cancer</td><td align=\"left\">Stress management skill taught had beneficial effects on reduced social disruption, and increased emotional well-being, positive states of mind, benefit finding, positive lifestyle change, and positive affect.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\" colspan=\"\">Wonghongkul et al. [##REF##16783127##206##]</td><td align=\"left\">2006</td><td align=\"left\">Uncertainty appraisal coping</td><td align=\"left\">Social support was used most to cope and confront-coping used the least; year of survival, uncertainty in illness and harm appraisal influenced QOL.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\" colspan=\"\">Yen et al. [##REF##16594937##207##]</td><td align=\"left\">2006</td><td align=\"left\">Depression and stress in breast cancer versus benign tumor</td><td align=\"left\">Stress from health problem was the most significant predictor for QOL among malignant group.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\" colspan=\"\">Costanzo et al. [##REF##18000503##208##]</td><td align=\"left\">2007</td><td align=\"left\">Adjustment to life after treatment</td><td align=\"left\">While breast cancer survivors demonstrated good adjustment on general distress following treatment, some women were at risk for sustained distress.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\" colspan=\"\">Wong and Fielding [##REF##17205280##209##]</td><td align=\"left\">2007</td><td align=\"left\">Change in psychological distress and change in QOL</td><td align=\"left\">The magnitude of change in psychological distress significantly impacted physical and functional, but not social QOL in breast cancer patients.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\" colspan=\"\">Meneses et al. [##REF##17878129##210##]</td><td align=\"left\">2007</td><td align=\"left\">Psycho-educational intervention and QOL</td><td align=\"left\">Breast cancer education intervention is an effective intervention in improving QOL during the first year of breast cancer survivorship.</td></tr></tbody></table></table-wrap>", "<table-wrap position=\"float\" id=\"T10\"><label>Table 10</label><caption><p>A list of quality of life studies covering supportive care topics in breast cancer patients (1974–2007)</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td align=\"left\"><bold>Author (s) [Ref.]</bold></td><td align=\"left\"><bold>Year</bold></td><td align=\"left\"><bold>Intervention</bold></td><td align=\"left\"><bold>Results/conclusion(s)</bold></td></tr></thead><tbody><tr><td align=\"left\">van Holten-Verzantvoort et al. [##REF##1675865##217##]</td><td align=\"left\">1991</td><td align=\"left\">Pamidronate vs. control to reduce skeletal morbidity</td><td align=\"left\">Less short-term mobility impairment and bone pain in treatment group but not at long term.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\" colspan=\"\">Young-McCaughan and Sexton [##REF##2067964##218##]</td><td align=\"left\">1991</td><td align=\"left\">Aerobic exercise</td><td align=\"left\">Higher QOL in women who exercised.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\" colspan=\"\">Soukop et al. [##REF##1387929##219##]</td><td align=\"left\">1992</td><td align=\"left\">Ondansetron vs. metoclopramide to control emesis</td><td align=\"left\">Ondansetron was significantly superior.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\" colspan=\"\">Kornblith et al. [##REF##8229122##220##]</td><td align=\"left\">1993</td><td align=\"left\">Megestrol acetate in dose-response trial to prevent appetite loss</td><td align=\"left\">Lower dose was optimal achieving fewest side effects and a better QOL.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\" colspan=\"\">Clavel et al. [##REF##8459989##221##]</td><td align=\"left\">1993</td><td align=\"left\">Ondansetron to control emesis (review of five randomized trials)</td><td align=\"left\">Ondansetron provided significant QOL benefits compared with metoclopramide and alizapride)</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\" colspan=\"\">Ashbury et al. [##REF##10026552##222##]</td><td align=\"left\">1998</td><td align=\"left\">One-on-one peer support (Reach to Recovery programme)</td><td align=\"left\">Patients were satisfied and the programme had incremental benefits to QOL of patients.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\" colspan=\"\">Lee [##REF##9341354##223##]</td><td align=\"left\">1997</td><td align=\"left\">Social support (Reach to Recovery programme)</td><td align=\"left\">Social support plays a vital role in promoting overall QOL.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\" colspan=\"\">Wengstrom et al. [##REF##10522767##224##]</td><td align=\"left\">1999</td><td align=\"left\">Nursing intervention vs. control</td><td align=\"left\">No measurable effect on side effects or QOL but proved to have a positive effect in minimizing stress.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\" colspan=\"\">Lachaine et al. [##REF##10612010##225##]</td><td align=\"left\">1999</td><td align=\"left\">Ondansetron or metoclopramide to control emesis</td><td align=\"left\">Emesis control was significantly better in ondansetron; global QOL decreased more with metoclopramide.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\" colspan=\"\">Ritz et al. [##REF##10920832##226##]</td><td align=\"left\">2000</td><td align=\"left\">Advanced nursing care (APN)+ standard care vs. standard care</td><td align=\"left\">APN improved some QOL indicators.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\" colspan=\"\">Molenaar et al. [##REF##11250997##227##]</td><td align=\"left\">2001</td><td align=\"left\">Decision support to help patients to choose mastectomy or breast conservation</td><td align=\"left\">Decision-making improved as evaluated in terms of satisfaction and QOL.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\" colspan=\"\">Sammarco [##REF##11409065##228##]</td><td align=\"left\">2001</td><td align=\"left\">Perceived social support and uncertainty in younger breast cancer survivors</td><td align=\"left\">Significant positive correlation between perceived social support and QOL, and significant negative correlation between uncertainty, and QOL.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\" colspan=\"\">Michael et al. [##UREF##33##229##]</td><td align=\"left\">2002</td><td align=\"left\">Social networks</td><td align=\"left\">Pre-diagnosis level of social integration was important factor in future QOL, and explains more of the variance than treatment or tumour characteristics.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\" colspan=\"\">Olsson et al. [##REF##12546524##230##]</td><td align=\"left\">2002</td><td align=\"left\">Erythropoietin (randomized to two different doses epoetin-beta) for treatment of anemia</td><td align=\"left\">Global QOL was significantly improved and there was no difference between two study arms.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\" colspan=\"\">O'Shaughnessy [##REF##12533272##231##]</td><td align=\"left\">2002</td><td align=\"left\">Effects of epoetin-alfa to prevent neuronal apoptosis vs. placebo</td><td align=\"left\">Improved cognitive function, mood and QOL in treatment group.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\" colspan=\"\">Graves et al. [##REF##16594274##232##]</td><td align=\"left\">2003</td><td align=\"left\">8-week intervention based on social cognitive theory vs. standard care</td><td align=\"left\">Women in intervention group improved more on QOL, mood, self-efficacy, and outcome expectations.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\" colspan=\"\">Courneya et al. [##REF##12721239##233##]</td><td align=\"left\">2003</td><td align=\"left\">Exercise training (randomized trial)</td><td align=\"left\">Exercise training had beneficial effects on QOL.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\" colspan=\"\">Turner [##REF##15170652##234##]</td><td align=\"left\">2004</td><td align=\"left\">Seated exercise</td><td align=\"left\">Reduced fatigue and improved QOL observed.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\" colspan=\"\">Headley et al. [##REF##15378098##235##]</td><td align=\"left\">2004</td><td align=\"left\">Effect of seated exercise vs. control</td><td align=\"left\">Women with advanced breast cancer randomized to the seated exercise had a slower decline in total physical well-being and less increase in fatigue.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\" colspan=\"\">Weinfurt et al. [##REF##14734954##236##]</td><td align=\"left\">2004</td><td align=\"left\">Zoledronic asid or pamidornate disodium for metastatic bone lesion</td><td align=\"left\">Overall increase in QOL was observed.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\" colspan=\"\">Diel et al. [##REF##15251160##237##]</td><td align=\"left\">2004</td><td align=\"left\">Ibandronate vs. placebo in breast cancer with metastatic bone pain</td><td align=\"left\">A significant improvement in QOL was observed in intervention group; fatigue and pain were also reduced.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\" colspan=\"\">Body et al. [##REF##15363874##238##]</td><td align=\"left\">2004</td><td align=\"left\">Ibandronate vs. placebo in breast cancer with metastatic bone pain</td><td align=\"left\">Oral ibandronate had beneficial effects on bone pain and QOL and was well tolerated.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\" colspan=\"\">Wardley et al. [##REF##15870721##239##]</td><td align=\"left\">2005</td><td align=\"left\">Zoledronic acid in community setting vs. hospital setting in breast cancer patients with bone metastases</td><td align=\"left\">No difference between settings; safety and QOL benefits were observed.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\" colspan=\"\">Yoo et al. [##REF##15856335##240##]</td><td align=\"left\">2005</td><td align=\"left\">Muscle relaxation training and guided imagery vs. control</td><td align=\"left\">Less anticipatory and post-chemotherapy nausea and vomiting and higher QOL in intervention group.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\" colspan=\"\">Manning-Walsh [##UREF##34##241##]</td><td align=\"left\">2005</td><td align=\"left\">Relationships between persona land religious support and symptom distress and QOL</td><td align=\"left\">Personal support was positively related to QOL and had partial mediated effects on symptom distress but religious support was not.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\" colspan=\"\">Gordon et al. [##REF##16136270##242##]</td><td align=\"left\">2005</td><td align=\"left\">Home-based physiotherapy or group-based exercise or no intervention</td><td align=\"left\">Physiotherapy was found beneficial for functioning, physical and overall QOL.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\" colspan=\"\">Kendall et al. [##REF##15892425##243##]</td><td align=\"left\">2005</td><td align=\"left\">Influence of exercise (13.2 years following diagnosis)</td><td align=\"left\">High level of functioning was observed; those whose exercise increased, maintained a better QOL.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\" colspan=\"\">Chang et al. [##REF##15452188##244##]</td><td align=\"left\">2005</td><td align=\"left\">Effect of weekly epoetin alfa on maintaining hemoglobin levels, and reduction of transfusion vs. standard care</td><td align=\"left\">Epoetin alfa improved QOL, maintained hemoglobin levels and reduced of transfusion.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\" colspan=\"\">Hudis et al [##REF##16001991##245##]</td><td align=\"left\">2005</td><td align=\"left\">Effect of weekly epoetin alfa on hemoglobin levels</td><td align=\"left\">Epoetin alfa improved hemoglobin levels, and QOL in mildly anemic patients.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\" colspan=\"\">Badger et al. [##REF##15759065##246##]</td><td align=\"left\">2005</td><td align=\"left\">Telephone interpersonal counseling (TPC) vs. usual care</td><td align=\"left\">TIP-C was partially effective in symptom management and improved QOL.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\" colspan=\"\">Cheema and Gual [##UREF##35##247##]</td><td align=\"left\">2006</td><td align=\"left\">Full-body exercise training (before and after evaluation study)</td><td align=\"left\">Significant improvements were observed in upper- and lower-body strength, endurance, and QOL.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\" colspan=\"\">Sutton and Erlen [##REF##17135823##248##]</td><td align=\"left\">2006</td><td align=\"left\">Mutual dyadic support intervention</td><td align=\"left\">Most dyadic relationships were supportive, some reciprocal and some experienced conflicts.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\" colspan=\"\">Round et al. [##REF##16012817##249##]</td><td align=\"left\">2006</td><td align=\"left\">Recovery advice to prevent treatment problems</td><td align=\"left\">Recovery advice given to women neither was supported nor refuted to be able improve QOL.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\" colspan=\"\">Giese-Davis et al. [##UREF##36##250##]</td><td align=\"left\">2006</td><td align=\"left\">Peer counseling intervention (newly diagnosed and peer counselors)</td><td align=\"left\">Significant improvement in newly diagnosed was observed in trauma symptoms, emotional well-being, and self-efficacy but increased emotional suppression and declined QOL in peer counselors.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\" colspan=\"\">Moadel et al. [##REF##17785709##251##]</td><td align=\"left\">2007</td><td align=\"left\">Effects of yoga on QOL</td><td align=\"left\">Yoga was associated with beneficial effects on social functioning among breast cancer survivors.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\" colspan=\"\">Hartmann et al. [##REF##17396040##252##]</td><td align=\"left\">2007</td><td align=\"left\">Effects of a step-by-step inpatient rehabilitation programme and QOL</td><td align=\"left\">Although not generally superior to conventional inpatient rehabilitation programmes, the step-by-step rehabilitation provided marked benefits for patients with cognitive impairments.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\" colspan=\"\">Kim et al. [##REF##18062617##253##]</td><td align=\"left\">2007</td><td align=\"left\">Effect of complex decongestive therapy (CDT) on edema and QOL in breast cancer patients with unilateral leymphedema</td><td align=\"left\">CDT for upper limb lymphedema resulted in significant improved edema and QOL.</td></tr></tbody></table></table-wrap>", "<table-wrap position=\"float\" id=\"T11\"><label>Table 11</label><caption><p>A list of studies of quality of life and common symptoms in breast cancer patients (1974–2007)</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td align=\"left\"><bold>Author (s) [Ref.]</bold></td><td align=\"left\"><bold>Year</bold></td><td align=\"left\"><bold>Main focus</bold></td><td align=\"left\"><bold>Results/conclusion(s)</bold></td></tr></thead><tbody><tr><td align=\"left\">Hann et al. [##UREF##37##254##]</td><td align=\"left\">1998</td><td align=\"left\">Fatigue following radiotherapy</td><td align=\"left\">Women experienced fatigue but not worse than expected.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\" colspan=\"\">Carpenter et al. [##REF##9576289##255##]</td><td align=\"left\">1998</td><td align=\"left\">Hot flushes</td><td align=\"left\">65% (n = 114) reported ht flushes, with 59% of women with hot flushes rating the symptom as severe; hot flushes were most severe in women with a higher body mass index, those who were younger at diagnosis, and those receiving tamoxifen.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\" colspan=\"\">Hann et al. [##UREF##38##256##]</td><td align=\"left\">1999</td><td align=\"left\">Fatigue after high-dose therapy and autolougous stem cell rescue</td><td align=\"left\">Fatigue was related to medical and psychosocial factors.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\" colspan=\"\">Velanovich and Szymanski [##REF##10219851##257##]</td><td align=\"left\">1999</td><td align=\"left\">Lymphedema</td><td align=\"left\">Lymphedema occurred in a minority of patients and negatively affected QOL.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\" colspan=\"\">Bower et al. [##REF##10673515##258##]</td><td align=\"left\">2000</td><td align=\"left\">Fatigue, occurrence, and correlates</td><td align=\"left\">About one-third (n = 1957) reported more severe fatigue which was associate with higher level of depression, pain, and sleep difficulties.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\" colspan=\"\">Kuehn [##REF##11200778##259##]</td><td align=\"left\">2000</td><td align=\"left\">Surgery related symptoms following ALND</td><td align=\"left\">Shoulder-arm morbidity following ALND was found to be the most important long-term sources of distress.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\" colspan=\"\">Stein et al. [##REF##10908824##260##]</td><td align=\"left\">2000</td><td align=\"left\">Hot flushes</td><td align=\"left\">Hot flushes have a negative impact on QOL that may be due to fatigue and interference with sleep.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\" colspan=\"\">Beaulac et al. [##REF##12413312##261##]</td><td align=\"left\">2002</td><td align=\"left\">Lymphedema in survivors of early-stage breast cancer</td><td align=\"left\">MAS or BCS patients had similar lymphedema rates (28%–42/151) and had negative impact on long-term QOL in survivors.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\" colspan=\"\">Kwan et al. [##REF##12377968##262##]</td><td align=\"left\">2002</td><td align=\"left\">Arm morbidity after curative breast cancer treatment</td><td align=\"left\">Symptomatic patients and patients with lymphedema had impaired QOL compared to patients with no symptoms.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\" colspan=\"\">Fortner et al. [##UREF##39##263##]</td><td align=\"left\">2002</td><td align=\"left\">Sleep difficulties</td><td align=\"left\">Most patients had significant sleep problems that frequently being disturbed by pain, nocturia, feeling too hot, and coughing or snoring loudly; patients having significant sleep problems had greater deficits in QOL.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\" colspan=\"\">Engel et al. [##REF##12779081##264##]</td><td align=\"left\">2003</td><td align=\"left\">Arm morbidity</td><td align=\"left\">Up to 5 years after diagnosis 38% (n = 990) were still experienced arm problems and for these patients QOL was significantly lower than patients without arm morbidity; extent of axilla, younger age, and operating clinic significantly contributed to arm morbidity.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\" colspan=\"\">Caffo et al. [##REF##12889597##265##]</td><td align=\"left\">2003</td><td align=\"left\">Pain after surgery</td><td align=\"left\">Pain distressed 40% of patients (n = 529) regardless of treatment type and had negative effect on patients' QOL.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\" colspan=\"\">Rietman et al. [##REF##14668143##266##]</td><td align=\"left\">2004</td><td align=\"left\">Impairments and disabilities (2.7 years after surgery)</td><td align=\"left\">Pain was the most frequent assessed impairment after breast cancer treatment with strong relationship to perceived disability and QOL.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\" colspan=\"\">Schults et al. [##REF##15669929##267##]</td><td align=\"left\">2005</td><td align=\"left\">Menopausal symptoms</td><td align=\"left\">Menopausal signs and symptoms may not be different or the breast cancer survivors and they should not be confused with the QOL/psychosocial issues of the cancer survivors.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\">Ridner [##REF##15812652##268##]</td><td align=\"left\">2005</td><td align=\"left\">Lymphedema</td><td align=\"left\">Survivors with lymphedema reported poorer QOL; a symptom cluster including limb sensation, loss of confidence in body, decreased physical activity, fatigue and psychological distress was identified.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\" colspan=\"\">Conde et al. [##REF##16037759##269##]</td><td align=\"left\">2005</td><td align=\"left\">Menopausal symptoms</td><td align=\"left\">Prevalence of menopausal symptoms was similar in women with and without breast cancer; sexual activity was less frequent in breast cancer patients.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\" colspan=\"\">Burckhardt et al. [##REF##15860132##270##]</td><td align=\"left\">2005</td><td align=\"left\">Pain</td><td align=\"left\">Widespread pain significantly caused more experience of pain severity, pain impact and lower physical health than regional pain.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\" colspan=\"\">Mills et al. [##REF##15740827##271##]</td><td align=\"left\">2005</td><td align=\"left\">Fatigue</td><td align=\"left\">Pre-chemotherapy and chemotherapy induced inflammation were related to fatigue and QOL.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\" colspan=\"\">Massacesi [##UREF##40##272##]</td><td align=\"left\">2006</td><td align=\"left\">Effects of endocrine related symptoms in breast cancer who had switched from tamoxifen to anastrozole</td><td align=\"left\">Endocrine related symptoms improved but higher rate of mild arthritic and bone pain were reported.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\" colspan=\"\">Land et al. [##REF##16754728##273##]</td><td align=\"left\">2006</td><td align=\"left\">Tamoxifen or raloxifene related symptoms</td><td align=\"left\">No significant differences between groups; tamoxifen group reported better sexual function, more gynecological problems and vasomotor symptoms while raloxifene group reported more musculoskeletal problems and weight gain.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\" colspan=\"\">Heidrich et al. [##REF##16518447##274##]</td><td align=\"left\">2006</td><td align=\"left\">Symptoms, and symptom beliefs in older breast cancer patients vs. older women without breast cancer</td><td align=\"left\">Symptom experience and QOL of older breast cancer survivors were similar to those of older women with other chronic health problems.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\" colspan=\"\">Gupta et al. [##REF##16428125##275##]</td><td align=\"left\">2006</td><td align=\"left\">Menopausal symptoms</td><td align=\"left\">96% reported vasomotor, 83% psychological and 90% somatic symptoms (n = 200) which negatively correlated not only their own but also with their partners' QOL.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\" colspan=\"\">Byar et al. [##REF##16470230##276##]</td><td align=\"left\">2006</td><td align=\"left\">Fatigue</td><td align=\"left\">Fatigue was associated with other physical and psychological symptoms and higher fatigue compromised QOL.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\" colspan=\"\">Arndt et al. [##REF##17048250##277##]</td><td align=\"left\">2006</td><td align=\"left\">Fatigue</td><td align=\"left\">Fatigue emerged as the strongest predictor of QOL.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\" colspan=\"\">Pyszel et al. [##REF##17319631##278##]</td><td align=\"left\">2006</td><td align=\"left\">Disability, and psychological distress in breast cancer survivors with and without lymphedema</td><td align=\"left\">Patients with arm lymphedema were more disabled, experienced a poorer QOL and had increased psychological distress in comparison to those without lymphedema.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\" colspan=\"\">Dagnelie et al. [##REF##17363839##279##]</td><td align=\"left\">2007</td><td align=\"left\">Fatigue</td><td align=\"left\">Of all QOL domains/subscales, fatigue is by far the predominant contributor to patient-perceived overall QOL in breast cancer patients preceding high-dose radiotherapy.</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\" colspan=\"\">Janz et al. [##UREF##41##280##]</td><td align=\"left\">2007</td><td align=\"left\">Relationship between symptoms and post-treatment QOL</td><td align=\"left\">Five most common symptoms were: systemic therapy side effects, fatigue, breast symptoms, sleep difficulties, and arm symptoms. Fatigue had the greatest impact on QOL.</td></tr></tbody></table></table-wrap>" ]

[]

[]

[]

[]

[]

[ "<supplementary-material content-type=\"local-data\" id=\"S1\"><caption><title>Additional file 1</title><p>Quality of life in breast cancer patients. This is a chronological list of all papers that were published since 1974 to the end of year 2007 in the English biomedical journals. The list is organized for each year and only contains papers that used the word quality of life and breast cancer or breast carcinoma in their titles. The papers are sorted alphabetically.</p></caption></supplementary-material>" ]

[ "<table-wrap-foot><p>* Excluding duplicates and papers that appeared online and indexed ahead of publication.</p><p>** Excluding all meetings abstracts, editorials, brief commentaries, letters, replies, erratum, and dissertation abstracts. For all citations see Additional file ##SUPPL##0##1##.</p></table-wrap-foot>", "<table-wrap-foot><p>Abbreviations</p><p>MRM: modified radical mastectomy, MAS: mastectomy, BCS: breast conservation surgery, SNLB: sentinel lymph node biopsy, ALND: axillary lymph node dissection</p></table-wrap-foot>", "<table-wrap-foot><p>Abbreviations</p><p>C: Cyclophosphamide, M: Methotrexate, F: 5-fluorouracil, A: Doxorubcin, E: Epirubcin, T: Docetaxel, TAM: Tamoxifen, ANA: Anastrozole, EXE: Exemestane, QOL: Quality of life, DPPE: Tesmilifene, Granulocyte colony stimulating factor: G-CSF, CTCb: Cyclophosphamide, thiotepa, and carboplatin</p></table-wrap-foot>", "<table-wrap-foot><p>ALND: axillary lymph node dissection, ASCT: autologous stem cell transplantation, SLNB: sentinel lymph node biopsy.</p></table-wrap-foot>" ]

[]

[ "<media xlink:href=\"1756-9966-27-32-S1.doc\" mimetype=\"application\" mime-subtype=\"msword\"><caption><p>Click here for file</p></caption></media>" ]

{ "acronym": [], "definition": [] }

[ "<title>Background</title>", "<p>The death of a spouse can lead to serious psychological problems [##REF##16798624##1##], although the impact of spousal bereavement on mental health diverges among the widowed. Targeted support can have a preventive effect and reduce psychological problems resulting from the death of the spouse [##REF##3378823##2##, ####REF##3740226##3##, ##REF##8288152##4##, ##REF##9051279##5####9051279##5##]. However, not all studies on bereavement interventions demonstrate positive results [##UREF##0##6##, ####UREF##1##7##, ##REF##14577426##8####14577426##8##]. Bereavement interventions are defined as all interventions developed to benefit bereaved persons in terms of alleviating the emotional and practical problems following the loss of a loved one. Types of intervention can vary from self-help groups to psychotherapy. It appears that the effectiveness of a bereavement intervention is largely determined by the population towards it is directed. Most outreaching preventive interventions for all widowed individuals are not very beneficial [##UREF##2##9##], probably because most of the widowed are able to adjust relatively well over time and do not need a specific intervention to regain pre-bereavement levels of functioning. However, this does not apply to all widowed individuals. Several widowed individuals are not able to deal with the loss on their own, and for those widows and widowers a bereavement intervention could be very helpful. Interventions directed towards widows and widowers with a high risk profile, such as having more severe psychological problems or symptoms of complicated grief, do show desirable results [##UREF##2##9##].</p>", "<p>Despite a growing understanding of the effectiveness of bereavement interventions and the groups that benefit most from them, we know little about the cost-effectiveness of bereavement interventions. It is reasonable to assume that some interventions, especially those provided by volunteers, could be cost-effective – even when no superior clinical results could be demonstrated. It is plausible that widows and widowers that have been offered targeted support by volunteers will make less use of health care services, the latter being considerably more expensive than the attention of (trained) volunteers. In order to test this hypothesis, we conducted a cost-utility analysis alongside a randomized clinical trial on a visiting service for older widowed individuals, which failed to demonstrate superior clinical effects on depression of targeted support by trained volunteers over care as usual (Onrust, Willemse, van den Bout &amp; Cuijpers, in press.).</p>" ]

[ "<title>Methods</title>", "<title>Sample and setting</title>", "<p>This cost utility analysis was based on a randomized controlled study on the effects of a visiting service for older widowed individuals. The research has been judged and ethically approved by the METiGG, a medical-ethical committee for research in mental health care settings in the Netherlands. The study was conducted in 18 municipalities in the Netherlands. Making use of the Registry Office, letters were sent to all residents at the age of 55 and older who had lost their spouse 6 – 9 months earlier. This is consistent with the recruitment procedure of most visiting services in the Netherlands. Respondents were contacted only 6 – 9 months after the loss, because in the initial stage of bereavement, social support is usually available in the direct environment of family, friends and neighbours. The letters contained information about the study, an informed consent form, and a short screenings questionnaire. In order to increase study participation, we used local media to inform the population and stimulate application. This media attention resulted in several participants who had not yet received a letter because the death was more recent than 6 months, or who had reconsidered participation several months after they were contacted. This media attention also resulted in 5 participants (2,3%) slightly younger than 55 years of age. Since it is not unconventional that slightly younger individuals make an appeal to the visiting service, and since age did not predict the effects of the visiting service (Onrust &amp; Willemse et al., in press), these respondents were included in the study.</p>", "<p>To be eligible for the study, respondents had to meet the following inclusion criteria: widowed during the past year, moderate or strong feelings of loneliness, and the absence of a full-blown mental disorder. In addition, respondents had to be capable of participating in a 1-hour-long interview. Further explanation of the inclusion criteria is presented elsewhere (Onrust &amp; Willemse et al., in press).</p>", "<p>Figure ##FIG##0##1## presents the several steps of the recruitment of study participants. During the one year recruitment period, a total of 2,708 letters were sent to widowed individuals. In total, 308 widowed individuals (11,4%) returned the informed consent form. In order to determine whether the widowed individuals agreeing to participate were eligible for the study, we carried out a stepwise screening procedure.</p>", "<p>At first, we measured feelings of loneliness with the 'Loneliness Scale' [##UREF##3##10##], which was enclosed in the request for participation. Only respondents with at least moderate feelings of loneliness were included in the study. We excluded respondents without feelings of loneliness because the intervention was designed for widowed individuals with social support deficits and because bereavement interventions directed towards the entire population of widowed individuals are generally not effective. Of the 308 widowed individuals agreeing to participate, 27 widows and widowers (8,8%) were excluded because they did not report feelings of loneliness.</p>", "<p>All other candidates were contacted by phone for a screening interview, in order to ascertain the capability of engaging in a 1-hour-long interview and the absence of a full-blown mental disorder. The presence or absence of full-blown mental disorders was measured with the M.I.N.I. Plus, a short standardized diagnostic interview [##REF##9881538##11##]. A total of 33 widowed individuals (10,7%) were excluded from the study, as they were considered to be not capable of participating, mostly because they were confused and did not understand the objectives of the study. Based on the M.I.N.I., another 30 widows and widowers (9,7%) were excluded as they appeared to meet the DSM-IV diagnostic criteria for either depression or an anxiety disorder. The remaining 218 widowed individuals (71% of those agreeing to participate; 8,1% of all persons approached) met inclusion criteria and were approached for the baseline measurement. This measurement was completed by 216 widowed individuals who were randomly allocated to the visiting service (n = 110) or a brief brochure (n = 106). The randomization was carried out centrally, using blocked randomization stratified for gender and region with the widowed individual as unit of randomization, with blocks of two widowed individuals. Data were collected at baseline, at 6 months, 12 months and 24 months after baseline. Although all follow-up assessments were attempted to be scheduled as close as possible to the intended point in time, we did allow a deviation of (± 2 weeks). This paper focuses on the cost-utility of the visiting service at the 12 months follow-up assessment, as the intervention was administered during the first 12 months of the study and potential shifts in health care utilization were most likely to occur during this period. At the 12 month follow-up assessment 185 (86%) widows and widowers were retained in the trial. The recruitment of participants and the baseline measurements took place in 2003–2004, data collection for the 12 month follow-up assessment was carried out in 2004–2005.</p>", "<title>Intervention</title>", "<p>The experimental intervention was the visiting service, based on the Widow-to-Widow program [##UREF##4##12##]. Respondents who were allocated to the visiting service were offered 10 – 12 home visits by a trained volunteer. During the home visits, one-to-one support was offered by exchanging experiences and emotions. The volunteers provided the respondents with the opportunity to express their feelings and a better understanding of their grieving process. In addition, the volunteers provided information and sometimes practical help. All volunteers were widowed themselves for some years. They had attended a course of 6 meetings, in which both theoretical knowledge (grief phenomena; tasks of grief; loneliness and social support) and practical skills (empathic listening; conversation techniques; setting boundaries) were learned. Mainly based on the way they participated in this course, their eligibility for the program was evaluated. During the period of home visits, all volunteers were supervised by the coordinator of the visiting service. The coordinators of all visiting services had also attended a course of 6 meetings. In this course, which was based on the \"Manual Visiting Services\" [##UREF##5##13##], information was provided on the organization and procedure of the visiting service and the supervision of the volunteers.</p>", "<p>Respondents who were allocated to the visiting service were allowed to use all other types of health services and community resources during the intervention period.</p>", "<p>The comparison (control) intervention consisted of a brief brochure on depressive symptoms. The brochure provided information and several tips to improve well-being. Respondents who were allocated to the comparison group were not offered any type of intervention, but were allowed to use all types of health services and community resources with the exception of the visiting service during the study. Mostly, widowed individuals are supported by their direct environment. This support generally diminishes over time. Although there are several interventions available for widowed individuals to cope with their grief, it depends on the widowed individual whether he or she will actually use the available services. Generally, only a small amount of the widowed population does make use of special services.</p>", "<title>Clinical end terms</title>", "<p>Quality of life was assessed with the EuroQol (EQ-5D) [##REF##10158943##14##]. The EuroQol is made up of five dimensions: Mobility, Self-care, Usual Activities, Pain/Discomfort, and Anxiety/Depression. Respondents were asked to indicate for each dimension whether they experienced 'no problems', 'some problems', or 'extreme problems'. Subsequently, the separate scores were combined into the EQ-5D Index, a health status index. The EQ-5D Index can be linked directly to empirical values for health status of the general public, which allows the conversion to utilities [##UREF##6##15##].</p>", "<title>Resource use</title>", "<p>For this study we adopted a societal perspective, including the cost of all types of health care services (direct medical costs), patient costs such as costs for traveling and parking (direct non-medical costs) and costs deriving from not being able to perform domestic tasks. We did not include costs attributable to productivity loss, since our sample consisted of older widowed individuals of which only a small part (14%) was employed at baseline. The number of widowed individuals that did report absence from work or reduced efficiency at work (3% at baseline and 1% at follow-up) was too small to be taken into account. Information on the use of health care services and the capability of performing domestic tasks was gathered with parts of the Trimbos and institute of Medical Technology Assessment Questionnaire on Costs Associated with Psychiatric Illness (TiC-P) [##UREF##7##16##].</p>", "<p>Direct medical and direct non-medical costs are presented in Table ##TAB##0##1##. Direct medical costs are treatment costs for several formal (e.g. general practitioners, mental health services, social work, home care) and informal caregivers (such as family and friends), which were calculated by multiplying the number of health service units (e.g. consultations, contacts) by their standard cost price [##UREF##8##17##]. Since the TiC-P measures health care utilization during the past 4 weeks, the costs were subsequently converted to annual costs. We also included the costs of antidepressant, anxiolytic and hypnotic medication, calculated as the price per standard daily dose as reported in the Pharmaceutical Compass [##UREF##9##18##], multiplied by the number of prescription days, plus the pharmacist's dispensing costs of €6,45 per prescription. Since most psychiatric drugs are prescribed for a period of three month on average, we added the pharmacist's dispensing costs 4 times in order to estimate annual costs. Medication use was assessed by the interviewer, who asked the respondents what kind of prescription drugs they used. Participants were encouraged to get the box of the medication in order to ascertain the correct name of the drug. At last, costs arising from being too ill to perform domestic tasks were evaluated at the price of domestic help at €8,30.</p>", "<p>Direct non-medical costs are costs patients had to make by traveling to health service providers and parking. These costs were valued at €0,16/km and €2,50/hour parking time. We also added the costs of patients' time spent in travel, waiting and treatment at €8,30 [##UREF##8##17##].</p>", "<p>All costs are estimated for the reference year 2003 and are presented in euros.</p>", "<title>Intervention costs</title>", "<p>Table ##TAB##1##2## represents the cost of the visiting service. Direct medical costs of the intervention included organizing the visiting service, training of volunteers, supervision of the volunteers and the intake by the coordinator of the visiting service (either a paid social worker or a volunteer), the costs of phone calls to both volunteers and participants and overhead costs. In order to estimate these costs we used different sources. First, we calculated the annual costs per participant based on the financial paragraph of the annual report of two participating visiting services. Second, the Manual Visiting Services [##UREF##5##13##], which was used to set up the visiting services, did also include an estimate of annual costs. This estimate was indexed for 2003 in two ways, by means of a Health Care Index and by means of the General Index as reported by Statistics Netherlands [##UREF##10##19##]. The costs of the coordinator could differ depending on whether the coordinator is paid or not. A paid social worker is more expensive than a volunteer. Since both options were possible, we used examples of both options in the calculation of the intervention costs. Subsequently, we averaged the four different estimates and added time costs for the volunteers valued at €12,45 per visit (visit plus travel time). Together these direct medical costs added up to €453 annually per recipient of the visiting service. Direct non-medical costs were time costs of the participant, valued at €8,30/visit.</p>", "<title>Analyses</title>", "<p>Statistical analysis was guided by some characteristics of our data.</p>", "<p>Primarily, our data were not complete. At the 12 month follow-up assessment 14.4% of the data was missing. All analyses were conducted according to the intention-to-treat principle. Therefore, all missing values were imputed. In order to replace the missing values by plausible estimates, we used the regression imputation procedure as implemented in Stata version 9.1 [##UREF##11##20##].</p>", "<p>Secondly, we had to take into account two confounding variables. Despite random allocation to the research conditions, there were two significant differences between the experimental group and the control group at baseline. Participants in the experimental group were on average more lonely and had a worse quality of life at the start of the study. We adjusted for both confounders (by using residualised QALYs), since they were significant predictors of the QALY end-term,.</p>", "<p>In the cost-utility analysis, we calculated the pre-post changes in costs and the pre-post changes in quality of life in each of the conditions. Then we calculated the incremental cost-utility ratio (ICUR) across the experimental and control conditions, which represents the incremental costs (or savings) per QALY gained in the experimental condition relative to the control condition. Uncertainty was assessed by means of non-parametric bootstrapping (2,500 times) of the data of the individual respondents. The comparison of the simulated ICURs is presented in a cost-utility plane (Figure ##FIG##1##2##), with differences in costs on the vertical axis and differences in QALYs on the horizontal axis. If the majority of the estimates appear in the top left-hand quadrant of the plane, the intervention results in a loss of quality of life against additional costs as compared with the control condition, which makes the intervention clearly unacceptable from a cost-effectiveness perspective. If the majority of the bootstrapped ICURs appear in lower right-hand quadrant of the plane, the intervention produces more QALYs for less costs than the control condition, which makes the intervention clearly superior from a cost-effectiveness perspective. In the other two quadrants the additional costs or savings have to be weighted against a loss or gain in QALYs.</p>", "<p>The results of the cost-utility analysis are also presented in a cost-utility acceptability curve (Figure ##FIG##2##3##). The acceptability curve represents the probability that the intervention is cost-effective, given a varying threshold for the willingness to pay for each QALY gained. Finally, we calculated the Net Monetary Benefit (NMB) for two different ceilings of willingness to pay that are usually applied in the Netherlands (20,000 euro and 80,000 euro). The <italic>net monetary benefit </italic>of a participant is calculated as: <italic>net benefit </italic>= [(willingness to pay) * Δ <italic>effects</italic>] – Δ <italic>costs</italic>.</p>", "<title>Sensitivity analyses</title>", "<p>As already mentioned, we conducted our main analyses without the costs attributable to productivity losses since the majority of our sample was not employed. However, since productivity loss is usually the main cost-driver, we repeated all analyses including the costs attributable to productivity losses. Information on the productivity loss was gathered with parts of the Trimbos and institute of Medical Technology Assessment Questionnaire on Costs Associated with Psychiatric Illness (TiC-P) [##UREF##7##16##]. To evaluate a lost day in a paid job we used age and gender specific friction-costs obtained from Oostenbrink et al. (2004). Friction costs represent the monetary counter-value of production losses that occur during absence from work with a limit to five months [##REF##10154656##21##]. Second, production losses also occur when people are ill, try to work, and are then less efficient. We estimated the number of work cutback days as the number of days actually worked when ill, multiplied by a self-reported inefficiency score, which ranges between 0 and 1 (0 = as efficient as when in good health, 1 = totally inefficient). Again, we used friction costs to valuate these production losses.</p>" ]

[ "<title>Results</title>", "<title>Sample</title>", "<p>The sample consisted of 138 widows (63.8%) and 78 widowers (36.2%). The age of the participants ranged from 50 to 92 years (Mean = 68.8; Sd = 9.3) and the participants had received 13 years of education on average. Duration of widowhood varied from 2 to 14 months (Mean = 7.9; Sd = 1,9). As already mentioned, participants in the visiting service group differed significantly from participants in the control group on two variables: participants in the visiting service group reported more feelings of loneliness than in the CAU group (Mean = 7.1; Sd = 3.0 versus mean = 6.0; Sd = 2.9; t = -2.66; p = 0.008) and a worse health-related quality of life at baseline (EQ-5D utility score mean = 0.76; Sd = 0,25 versus mean = 0.83; Sd = 0.18; t = 2.19; p = 0.030). There was no significant difference in loss to follow-up rates between the research conditions. Furthermore, completers did not differ from non-completers on any of the baseline variables, which indicated that loss to follow-up was at random.</p>", "<title>Quality of Life</title>", "<p>Participants in the visiting service group demonstrated a significant improvement in health-related quality of life (EQ-5D utility score at baseline mean = 0.76 (s.d. = 0,25); EQ-5D utility score at 12 months follow-up mean = 0.80 (s.d. = 0.18); Difference mean = 0.04 (s.d = 0.02) QALY gained; t = -2.273; p = 0.025). Participants in the control group did not (baseline mean = 0.83 (s.d. = 0,18); follow-up mean = 0.81 (s.d = 0.21); Difference mean = 0.01 (s.d. = 0.02) QALY lost; t= 0.696; p = 0.488). However, the visiting service group did not significantly differ from the control group in their changes in health-related quality of life over time when we adjusted for both confounding variables (t = 1.29; p = 0.215).</p>", "<title>Costs</title>", "<p>Table ##TAB##2##3## presents the annual capita costs of both the visiting service group and the control group at baseline and at the 12 months follow-up assessment. In both groups costs increased over time, however these changes in costs were not statistically significant (p = 0.166 in the visiting service group and p = 0.430 in the control group). In the visiting service group, the increased costs included the costs of the intervention (€ 553), but these additional costs were partly compensated for by savings elsewhere in the healthcare and welfare sector. The mean difference of the additional costs was € 210 (s.e. = 363) in favour of the control group, but this difference was not statistically significant (t = -0.579; p = 0.563).</p>", "<title>Cost-utility</title>", "<p>The incremental cost-utility ratio was calculated as (ΔCosts_E- ΔCosts_C)/(ΔQALY_E- ΔQALY_C), where ΔCosts represents the average additional per capita costs and ΔQALY represents the number of QALYs gained over time, controlled for both confounding variables, in both the visiting service group (E) and the control group (C). Substitution yields a cost-utility ratio of (390 - 180)/(0.01 - (-0.02)) = 6,827. This means that for each QALY gained by offering the visiting service, the additional costs amount to € 6,827. Bootstrapping of the data of the individual respondents yields a median ICUR of € 4,123 (95% Confidence Interval : – €627,530 – €668,056).</p>", "<p>The incremental cost-utility ratio is surrounded by a certain amount of uncertainty, which is presented in the cost-effectiveness plane (Figure ##FIG##1##2##). Each dot of the cost-effectiveness plane represents a bootstrap replication (n = 2,500) of the incremental cost-utility ratio; 28% of the dots are in the lower right-hand quadrant, indicating a 28% probability that the visiting services generates better health effects against lower costs; there is a 5% probability that the visiting service generates worse outcomes against higher costs and a 1% probability that the visiting service generates worse outcomes against lower costs. However, most dots (59%) are in the upper right-hand quadrant, indicating better outcomes against higher costs.</p>", "<title>Acceptability</title>", "<p>The acceptability curve for the incremental cost-utility ratio is presented in Figure ##FIG##2##3##. The visiting service had a probability of 31% of being more acceptable than the comparator condition from a cost-effectiveness point of view under the conservative scenario that there is no willingness to pay for a gain of one QALY. However, people are generally willing to pay for a QALY gained. When the willingness to pay is raised to € 10,000, the visiting service has a probability of 55% of being cost-effective compared with the informational brochure. Generally, the willingness to pay for a QALY gained by preventive interventions is approximately €20,000, and at this threshold the visiting service has a probability of 70% of being more acceptable than CAU.</p>", "<title>Net Monetary Benefit</title>", "<p>Given a willingness to pay of 20,000 Euro for a QALY gained, the Net Monetary Benefit is: NMB = 20,000 * 0.031 – 210 = 410. Given a willingness to pay of 80,000 Euro for a QALY gained the Net Monetary Benefit is: NMB = 80,000 * 0.031 – 210 = 2270.</p>", "<title>Sensitivity analyses</title>", "<p>When the indirect costs related to the production losses are included, the incremental cost-utility ratio is €11.239. Bootstrapping of the data yields a median ICUR of € 6,151 (95% Confidence Interval : – €205,706 – €222,067). The distribution of the bootstrapped ICURs over the cost-effectiveness plane is as follows: 63% of the ICURs fall in the upper right-hand quadrant indicating that better effects are obtained against higher costs, 5% fall in the upper left-hand quadrant indicating that the visiting service is inferior, 1% fall in the lower left-hand quadrant indicating that the visiting service has worse clinical outcomes against lower costs, and 24% of the bootstrapped ICURs fall in the lower right-hand quadrant, implying that the visiting service is dominant, because it generates better outcomes against lower costs than the control condition. Under these circumstances, the visiting service has a probability of 27% of being acceptable when the willingness to pay equals zero. When the willingness to pay is increased to € 10,000, and € 20,000, the probability of the visiting service being more acceptable than the control condition increases to 49% and 64% respectively.</p>" ]

[ "<title>Discussion</title>", "<p>We conducted a cost-utility analysis with health-related quality of life as clinical end term. Health related quality of life was measured with the EQ-5D which is a generic measure of health status. As only one of the five components of the EQ-5D has a psychological nature, it is sometimes debated whether the use of this measure is justified in the evaluation of psychological interventions. We believe that it is. First of all, if an individual reports 'extreme problems' on the mental health dimension (anxiety/depression) of the EQ-5D, without any other problems, this health status is still evaluated as 0.36954 by the Dutch general population [##UREF##6##15##], which represents poor health. By resolving these 'extreme problems' in mental health, effective psychological interventions are able to demonstrate changes in QALYs. Secondly, the use of a generic measure of health-related quality of life enables us to compare the all kinds of interventions on their cost-effectiveness. And although most effective medical procedures usually demonstrate larger improvements in QALYs than psychological interventions, their costs are usually much higher as well. Psychological interventions therefore do not need to result in large changes in QALYs to be cost-effective.</p>", "<p>In this study we evaluated the cost-effectiveness of a visiting service for older widows and widowers. The visiting service is a selective preventive intervention. Selective bereavement interventions are directed towards bereaved individuals with a high risk profile. Bereaved individuals with a high risk profile are more likely to experience an abnormal form of grief. The visiting service focussed on loneliness as risk factor. Besides selective bereavement interventions, there are also universal bereavement interventions and indicated bereavement interventions, which are respectively directed towards all bereaved persons and persons who already are experiencing abnormal bereavement. Besides some positive effects of universal prevention for bereaved children, there is hardly any evidence for the effectiveness of universal bereavement interventions [##UREF##2##9##]. Screening for high risk seemed to increase the efficacy of bereavement interventions. Some studies on selective bereavement interventions demonstrated modest effects, although there were some indications that this is only temporary [##UREF##2##9##]. Indicated interventions generally seem to lead to favourable results, both for bereaved individuals suffering from complicated grief and bereaved individuals suffering from bereavement-related depression [##UREF##2##9##]. Given that these bereavement interventions differ in nature and clinical effectiveness, results of this study should not be generalised to indicated interventions or treatment for bereavement related disorders, which are usually administered by a therapist instead of a volunteer and clearly differ from selective interventions like the visiting service.</p>", "<title>Main findings</title>", "<p>The experimental group demonstrated a small improvement in health-related quality of life after the intervention. This improvement was absent in the control group. However, since the baseline scores in the control group were significantly higher, there was less possibility for improvement, and when we controlled for this 'false start' the differences in effects on health-related quality of life were no longer significant, which should be read as a warning against overly optimistic interpretations of our data. In both groups, the total costs were higher at the 12 months follow-up assessment than at baseline and the additional costs were somewhat higher in the experimental group than in the control group, although the difference was not significant. Overall, the experimental group demonstrated slightly better results against slightly higher costs. Whether the visiting service is acceptable depends on the willingness to pay: at a willingness to pay equal to zero, the visiting service has a probability of 31% of being acceptable; beyond €20,000, the visiting service has a probability of 70% of being acceptable.</p>", "<title>Limitations</title>", "<p>We have to place these findings in the context of the limitations of our study. There are several factors that limit the generalizability of our findings. First, this study is conducted alongside a randomized clinical trial on a visiting service for older widowed individuals with sufficient power to detect changes in clinical outcomes. However, the study was underpowered to detect changes in costs, which usually have large standard errors. Therefore, we took a probabilistic course indicating the likelihood that the intervention was superior from a health economic point of view. The second limitation is the high initial non-response. Only 11.4% of the approached widows and widowers returned the informed consent form. Although part of the non-response is caused by individuals not eligible for the study, both by individuals suffering from full-blown psychiatric disorders and by individuals without feelings of loneliness, it is unlikely that this applies for the complete non-response. Given the absence of information on those widows and widowers not participating in the study, the representativeness of the sample is not clear. One of the risks of studying a vulnerable population is self-selection of the least vulnerable individuals. The results of this study indicate that this probably also applies to a certain extent to our sample. Although the average utility score of our sample was significantly lower at baseline (mean = 0.79; s.d. = 0.22) than the average utility score of the general population (0.88), the average baseline score of our sample did not correspond to high distress either. Furthermore, our data were not complete. At the 12 month follow-up assessment, 14.4% of the data was missing. Although a loss to follow-up of 14% is not much, considering the period of 1 year, the imputation of missing values could still have distorted the results. However, completers did not differ from non-completers on any of the baseline variables, which suggested that loss to follow-up was completely at random. Another limitation was the difference at baseline between the visiting service group and the control group. Despite random allocation to the research conditions, respondents of the visiting service group were more lonely (mean loneliness score was 7.1 (s.d. = 3.0) in the experimental group compared to 6.0 (s.d. = 2.9) in the control condition; t = -2.66 and p = 0.008) and had a worse quality of life (mean 0.76 QALY (s.d. = 0.25) in the experimental group compared to 0.83 QALY (s.d. = 0.18) in the control condition; t = 2.19 and p = 0.030). Although we controlled for both confounding variables in our analyses, our results could have been biased. In addition, instead of monitoring health care utilisation over the entire year of interest in our study, which would clearly imply a large burden on the respondents, we made the simplifying assumption that the health care utilisation during the 4 week period that was assessed with the TIC-P could be interpolated to 1-year estimates from the baseline and follow-up measures. Although we expect that potential bias in these estimates are of similar magnitude in both trial conditions and therefore cancel each other out, the changes in health care costs over time are presumably more complex than is assumed under our model. Because of these limitations, the results of this study should be considered with caution.</p>" ]

[ "<title>Conclusion</title>", "<p>Despite its limitations, this study still offers new information on the potential benefits of bereavement interventions by (trained) volunteers. This study indicates that even in the absence of clinical effectiveness of a bereavement intervention its cost-effectiveness could still be acceptable. However, the acceptability of the visiting service we evaluated, depended mainly on the willingness to pay. Beyond a willingness to pay of €8,000, the visiting service has a probability of 50% of being more cost-effective than the control condition. At lower levels of willingness to pay it is more likely that the visiting service is not cost-effective than that the visiting service is superior.</p>", "<p>We assumed that bereavement interventions could be cost-effective because widows and widowers that have been offered targeted support by volunteers will make less use of health care services, the latter being considerably more expensive than the attention of (trained) volunteers. Our data suggests that widowed individuals in the experimental group did indeed make less use of health care services. In the experimental group, total costs without the intervention costs decreased while costs in the control group increased. However, these savings were not large enough to compensate for the intervention costs. In this study, the intervention was still more expensive overall than the control condition.</p>", "<p>We already know that bereavement interventions like the visiting service do not produce large benefits in terms of public mental health when targeted towards the entire population of all widowed individuals [[##REF##14577426##8##,##UREF##2##9##], Onrust et al., submitted]. Based on this cost-utility analysis we can now add that bereavement interventions like the visiting service will also not produce large benefits from the health economic point of view, when targeted towards the entire population of all widowed individuals. Presumably, those widowed individuals that are able to adjust relatively well over time and do not need a specific intervention to regain pre-bereavement levels of functioning, do not make frequent use of the health care services related to their bereavement as well. In light of our findings, we recommend that in depth analyses are conducted to identify who benefits most from this kind of interventions, and in what subgroups the incremental cost-utility is best. In the future bereavement interventions are then best directed to these groups.</p>" ]

[ "<p>This is an Open Access article distributed under the terms of the Creative Commons Attribution License (<ext-link ext-link-type=\"uri\" xlink:href=\"http://creativecommons.org/licenses/by/2.0\"/>), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.</p>", "<title>Background</title>", "<p>Despite a growing understanding of the effectiveness of bereavement interventions and the groups that benefit most from them, we know little about the cost-effectiveness of bereavement interventions.</p>", "<title>Methods</title>", "<p>We conducted a cost-utility analysis alongside a randomized clinical trial on a visiting service for older widowed individuals (n = 110) versus care as usual (CAU; n = 106). The visiting service is a selective bereavement intervention that offers social support to lonely widows and widowers by a trained volunteer. Participants were contacted 6–9 months post-loss. Eleven percent of all contacted persons responded and eight percent participated in the trial. The primary outcome measure was quality adjusted life years (QALYs) gained (assessed with the EQ-5D), which is a generic measure of health status. Costs were calculated from a societal perspective excluding costs arising from productivity losses. Using the bootstrap method, we obtained the incremental cost utility ratio (ICUR), projected these on a cost-utility plane and presented as an acceptability curve.</p>", "<title>Results</title>", "<p>Overall, the experimental group demonstrated slightly better results against slightly higher costs. Whether the visiting service is acceptable depends on the willingness to pay: at a willingness to pay equal to zero per QALY gained, the visiting service has a probability of 31% of being acceptable; beyond €20,000, the visiting service has a probability of 70% of being more acceptable than CAU.</p>", "<title>Conclusion</title>", "<p>Selective bereavement interventions like the visiting service will not produce large benefits from the health economic point of view, when targeted towards the entire population of all widowed individuals. We recommend that in depth analyses are conducted to identify who benefits most from this kind of interventions, and in what subgroups the incremental cost-utility is best. In the future bereavement interventions are then best directed to these groups.</p>", "<title>Trial registration</title>", "<p>Controlled trials ISRCTN17508307</p>" ]

[ "<title>Competing interests</title>", "<p>The authors declare that they have no competing interests.</p>", "<title>Authors' contributions</title>", "<p>SO was the trial's principal investigator, coordinated data-collection, conducted the analysis and wrote the manuscript. FS supervised the analysis, was the research team's advisor with regard to statistical and economic aspects and assisted with writing the manuscript. GW supervised the trial, coordinated the data-collection and assisted with writing the manuscript. JvdB was the research team's advisor with regard to clinical aspects and supervised the writing of the manuscript. PC designed the study and supervised the writing of the manuscript. All agree with the contents of the manuscript.</p>", "<title>Pre-publication history</title>", "<p>The pre-publication history for this paper can be accessed here:</p>", "<p><ext-link ext-link-type=\"uri\" xlink:href=\"http://www.biomedcentral.com/1472-6963/8/128/prepub\"/></p>" ]

[ "<title>Acknowledgements</title>", "<p>This study was financially supported trough a grant from the Netherlands Organisation for Health Research and Development (ZonMw); Grant: 2001-2-22102. The trial is registered in the ISRCTN Register as ISRCTN17508307. We would like to thank all participating visiting services and the participants in the trial, both respondents and volunteers, for their valuable help in making this study possible.</p>" ]

[ "<fig position=\"float\" id=\"F1\"><label>Figure 1</label><caption><p>Participants flow through the study.</p></caption></fig>", "<fig position=\"float\" id=\"F2\"><label>Figure 2</label><caption><p>Cost-effectiveness plane: each dot (n = 2,500) represents a bootstrapped cost-utility ratio.</p></caption></fig>", "<fig position=\"float\" id=\"F3\"><label>Figure 3</label><caption><p>ICUR acceptability curve: probability cost-utility ratio is acceptable given varying thresholds for willingness to pay.</p></caption></fig>" ]

[ "<table-wrap position=\"float\" id=\"T1\"><label>Table 1</label><caption><p>Direct medical and direct non-medical costs by health service type</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td/><td align=\"center\" colspan=\"2\"><bold>Direct Medical Costs (in 2003 €)</bold></td><td align=\"center\" colspan=\"2\"><bold>Direct Non-Medical Costs (in 2003 €)</bold></td></tr></thead><tbody><tr><td align=\"left\"><bold>Health service type</bold></td><td align=\"left\"><bold>unit</bold></td><td align=\"right\"><bold>cost price^a</bold></td><td align=\"right\"><bold>km, P, hrs^b</bold></td><td align=\"right\"><bold>cost price^c</bold></td></tr><tr><td colspan=\"5\"><hr/></td></tr><tr><td align=\"left\">Medical doctor</td><td align=\"left\">Consult</td><td align=\"right\">20,20</td><td align=\"right\">1.8 km, 1 h</td><td align=\"right\">11,10</td></tr><tr><td align=\"left\">Medical specialist</td><td align=\"left\">Consult</td><td align=\"right\">98,00</td><td align=\"right\">7 km, 2 h</td><td align=\"right\">20,20</td></tr><tr><td align=\"left\">Regional mental health service</td><td align=\"left\">Contact</td><td align=\"right\">124,00</td><td align=\"right\">10 km, 3 h</td><td align=\"right\">29,00</td></tr><tr><td align=\"left\">Regional addiction service^d</td><td align=\"left\">Contact</td><td align=\"right\">124,00</td><td align=\"right\">10 km, 3 h</td><td align=\"right\">29,00</td></tr><tr><td/><td/><td/><td/><td/></tr><tr><td align=\"left\">Mental Hospital – Outpatient</td><td align=\"left\">Consult</td><td align=\"right\">88,00</td><td align=\"right\">12 km, 4 h</td><td align=\"right\">37,20</td></tr><tr><td align=\"left\">Mental Hospital – Day care</td><td align=\"left\">Contact</td><td align=\"right\">125,00</td><td align=\"right\">12 km, 4 h</td><td align=\"right\">37,20</td></tr><tr><td align=\"left\">Mental Hospital – Inpatient</td><td align=\"left\">Day</td><td align=\"right\">250,00</td><td align=\"right\">8 h</td><td align=\"right\">66,40</td></tr><tr><td/><td/><td/><td/><td/></tr><tr><td align=\"left\">General Hospital – Outpatient</td><td align=\"left\">Consult</td><td align=\"right\">56,00</td><td align=\"right\">7 km, 3 h</td><td align=\"right\">28,50</td></tr><tr><td align=\"left\">General Hospital – Day care</td><td align=\"left\">Contact</td><td align=\"right\">229,00</td><td align=\"right\">7 km, 4 h</td><td align=\"right\">36.80</td></tr><tr><td/><td/><td/><td/><td/></tr><tr><td align=\"left\">Teaching Hospital – Outpatient</td><td align=\"left\">Consult</td><td align=\"right\">100,00</td><td align=\"right\">12 km, 3 h</td><td align=\"right\">29,30</td></tr><tr><td align=\"left\">Academic Hospital – Day care</td><td align=\"left\">Contact</td><td align=\"right\">229,00</td><td align=\"right\">12 km, 4 h</td><td align=\"right\">37,60</td></tr><tr><td/><td/><td/><td/><td/></tr><tr><td align=\"left\">Private practice psychotherapist</td><td align=\"left\">Session</td><td align=\"right\">76,00</td><td align=\"right\">5 km, 2 h</td><td align=\"right\">19,90</td></tr><tr><td align=\"left\">Social worker^e</td><td align=\"left\">Contact</td><td align=\"right\">45,00</td><td align=\"right\">7 km, 3 h</td><td align=\"right\">28,50</td></tr><tr><td align=\"left\">Physiotherapist</td><td align=\"left\">Contact</td><td align=\"right\">22,75</td><td align=\"right\">1,8 km, 2 h</td><td align=\"right\">19,40</td></tr><tr><td align=\"left\">Alternative Healer</td><td align=\"left\">Contact</td><td align=\"right\">8,30</td><td align=\"right\">1,8 km, 2 h</td><td align=\"right\">19,40</td></tr><tr><td align=\"left\">Self-Help</td><td align=\"left\">Session</td><td align=\"right\">0,00</td><td align=\"right\">10 km, 3 h</td><td align=\"right\">29,00</td></tr><tr><td/><td/><td/><td/><td/></tr><tr><td align=\"left\">Home care, nursing</td><td align=\"left\">Hour</td><td align=\"right\">30,70</td><td align=\"right\">0 km, 0 h</td><td align=\"right\">0,00</td></tr><tr><td align=\"left\">Home care, domestic</td><td align=\"left\">Hour</td><td align=\"right\">21,70</td><td align=\"right\">0 km, 0 h</td><td align=\"right\">0,00</td></tr><tr><td align=\"left\">Informal care (family, friends)^f</td><td align=\"left\">Hour</td><td align=\"right\">8,30</td><td align=\"right\">0 km, 0 h</td><td align=\"right\">0,00</td></tr></tbody></table></table-wrap>", "<table-wrap position=\"float\" id=\"T2\"><label>Table 2</label><caption><p>Calculation of intervention costs</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td align=\"left\"><bold>Source</bold></td><td align=\"right\"><bold>Total Costs</bold></td><td align=\"left\"><bold>Period</bold></td><td align=\"center\"><bold>Participants</bold></td><td align=\"center\"><bold>Annual costs per participant</bold></td></tr></thead><tbody><tr><td align=\"left\">Visiting Service 1</td><td align=\"right\">12,808</td><td align=\"left\">36 months</td><td align=\"center\">20</td><td align=\"center\">213</td></tr><tr><td align=\"left\">Visiting Service 2</td><td align=\"right\">54,900</td><td align=\"left\">24 months</td><td align=\"center\">80</td><td align=\"center\">343</td></tr><tr><td align=\"left\">Manual method 1^a</td><td align=\"right\">3,403</td><td align=\"left\">12 months</td><td align=\"center\">10</td><td align=\"center\">340</td></tr><tr><td align=\"left\">Manual method 2^b</td><td align=\"right\">3,200</td><td align=\"left\">12 months</td><td align=\"center\">10</td><td align=\"center\">320</td></tr><tr><td/><td/><td/><td/><td/></tr><tr><td align=\"left\">Mean Intervention</td><td/><td/><td/><td align=\"center\">304</td></tr><tr><td align=\"left\">Time Costs Volunteers</td><td align=\"right\">12.45/visit</td><td align=\"left\">12 months</td><td/><td align=\"center\">149</td></tr><tr><td align=\"left\">Direct Medical Costs</td><td/><td/><td/><td align=\"center\">453</td></tr><tr><td/><td/><td/><td/><td/></tr><tr><td align=\"left\">Time Costs Participant</td><td align=\"right\">8.30/visit</td><td align=\"left\">12 months</td><td/><td align=\"center\">100</td></tr><tr><td colspan=\"5\"><hr/></td></tr><tr><td align=\"left\">Total Costs Intervention</td><td/><td/><td/><td align=\"center\">553</td></tr></tbody></table></table-wrap>", "<table-wrap position=\"float\" id=\"T3\"><label>Table 3</label><caption><p>Annual per capita costs by item and condition</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td/><td align=\"center\" colspan=\"6\">Annual per capita costs (Direct Medical and Direct Non-Medical) in €</td></tr></thead><tbody><tr><td/><td align=\"center\" colspan=\"3\">Experimental Group (n = 110)</td><td align=\"center\" colspan=\"3\">Control Group (n = 106)</td></tr><tr><td colspan=\"7\"><hr/></td></tr><tr><td/><td align=\"left\">t0</td><td align=\"left\">t2</td><td align=\"left\">Diff. t0-t2</td><td align=\"left\">t0</td><td align=\"left\">t2</td><td align=\"left\">Diff. t0-t2</td></tr><tr><td colspan=\"7\"><hr/></td></tr><tr><td align=\"left\">Health service type</td><td align=\"left\">Mean (sd)</td><td align=\"left\">Mean (sd)</td><td align=\"left\">Mean (sd)</td><td align=\"left\">Mean (sd)</td><td align=\"left\">Mean (sd)</td><td align=\"left\">Mean (sd)</td></tr><tr><td/><td/><td/><td/><td/><td/><td/></tr><tr><td align=\"left\">Medical doctor</td><td align=\"left\">255 (328)</td><td align=\"left\">245 (342)</td><td align=\"left\">-11 (420)</td><td align=\"left\">199 (328)</td><td align=\"left\">265 (393)</td><td align=\"left\">66 (426)</td></tr><tr><td align=\"left\">Medical specialist</td><td align=\"left\">615 (1534)</td><td align=\"left\">490 (1091)</td><td align=\"left\">-125 (1710)</td><td align=\"left\">308 (835)</td><td align=\"left\">437 (1087)</td><td align=\"left\">129 (1100)</td></tr><tr><td align=\"left\">Regional mental health service</td><td align=\"left\">169 (860)</td><td align=\"left\">115 (438)</td><td align=\"left\">-54 (978)</td><td align=\"left\">38 (273)</td><td align=\"left\">82 (471)</td><td align=\"left\">44 (548)</td></tr><tr><td align=\"left\">Regional addiction service</td><td align=\"left\">0 (0)</td><td align=\"left\">0 (0)</td><td align=\"left\">0 (0)</td><td align=\"left\">0 (0)</td><td align=\"left\">0 (0)</td><td align=\"left\">0 (0)</td></tr><tr><td/><td/><td/><td/><td/><td/><td/></tr><tr><td align=\"left\">Mental Hospital – Outpatient</td><td align=\"left\">0 (0)</td><td align=\"left\">0 (0)</td><td align=\"left\">0 (0)</td><td align=\"left\">0 (0)</td><td align=\"left\">0 (0)</td><td align=\"left\">0 (0)</td></tr><tr><td align=\"left\">Mental Hospital – Day care</td><td align=\"left\">0 (0)</td><td align=\"left\">0 (0)</td><td align=\"left\">0 (0)</td><td align=\"left\">0 (0)</td><td align=\"left\">0 (0)</td><td align=\"left\">0 (0)</td></tr><tr><td align=\"left\">Mental Hospital – Inpatient</td><td align=\"left\">0 (0)</td><td align=\"left\">0 (0)</td><td align=\"left\">0 (0)</td><td align=\"left\">0 (0)</td><td align=\"left\">0 (0)</td><td align=\"left\">0 (0)</td></tr><tr><td/><td/><td/><td/><td/><td/><td/></tr><tr><td align=\"left\">General Hospital – Outpatient</td><td align=\"left\">10 (105)</td><td align=\"left\">27 (151)</td><td align=\"left\">17 (186)</td><td align=\"left\">0 (1)</td><td align=\"left\">16 (110)</td><td align=\"left\">16 (110)</td></tr><tr><td align=\"left\">General Hospital – Day care</td><td align=\"left\">0 (0)</td><td align=\"left\">0 (0)</td><td align=\"left\">0 (0)</td><td align=\"left\">0 (0)</td><td align=\"left\">0 (0)</td><td align=\"left\">0 (0)</td></tr><tr><td/><td/><td/><td/><td/><td/><td/></tr><tr><td align=\"left\">Teaching Hospital – Outpatient</td><td align=\"left\">0 (0)</td><td align=\"left\">0 (0)</td><td align=\"left\">0 (0)</td><td align=\"left\">0 (0)</td><td align=\"left\">0 (0)</td><td align=\"left\">0 (0)</td></tr><tr><td align=\"left\">Academic Hospital – Day care</td><td align=\"left\">0 (0)</td><td align=\"left\">0 (0)</td><td align=\"left\">0 (0)</td><td align=\"left\">0 (0)</td><td align=\"left\">0 (0)</td><td align=\"left\">0 (0)</td></tr><tr><td/><td/><td/><td/><td/><td/><td/></tr><tr><td align=\"left\">Private practice psychotherapist</td><td align=\"left\">116 (966)</td><td align=\"left\">15 (119)</td><td align=\"left\">-101 (972)</td><td align=\"left\">36 (364)</td><td align=\"left\">26 (171)</td><td align=\"left\">-9 (405)</td></tr><tr><td align=\"left\">Social worker</td><td align=\"left\">62 (268)</td><td align=\"left\">46 (227)</td><td align=\"left\">-16 (334)</td><td align=\"left\">54 (291)</td><td align=\"left\">42 (186)</td><td align=\"left\">-12 (326)</td></tr><tr><td align=\"left\">Physiotherapist</td><td align=\"left\">298 (810)</td><td align=\"left\">322 (931)</td><td align=\"left\">24 (1007)</td><td align=\"left\">370 (1100)</td><td align=\"left\">249 (652)</td><td align=\"left\">-121 (952)</td></tr><tr><td align=\"left\">Alternative Healer</td><td align=\"left\">10 (59)</td><td align=\"left\">18 (139)</td><td align=\"left\">8 (152)</td><td align=\"left\">7 (70)</td><td align=\"left\">27 (100)</td><td align=\"left\">20 (102)</td></tr><tr><td align=\"left\">Self-Help</td><td align=\"left\">0 (0)</td><td align=\"left\">47 (298)</td><td align=\"left\">47 (298)</td><td align=\"left\">10 (82)</td><td align=\"left\">40 (198)</td><td align=\"left\">30 (217)</td></tr><tr><td/><td/><td/><td/><td/><td/><td/></tr><tr><td align=\"left\">Home care, nursing</td><td align=\"left\">1169 (2346)</td><td align=\"left\">1261 (2460)</td><td align=\"left\">92 (1386)</td><td align=\"left\">1068 (2076)</td><td align=\"left\">1151 (2078)</td><td align=\"left\">83 (1383)</td></tr><tr><td align=\"left\">Home care, domestic</td><td align=\"left\">55 (491)</td><td align=\"left\">35 (246)</td><td align=\"left\">-20 (350)</td><td align=\"left\">32 (200)</td><td align=\"left\">9 (82)</td><td align=\"left\">-24 (141)</td></tr><tr><td align=\"left\">Informal care (family, friends)</td><td align=\"left\">37 (146)</td><td align=\"left\">18 (81)</td><td align=\"left\">-19 (141)</td><td align=\"left\">73 (484)</td><td align=\"left\">32 (162)</td><td align=\"left\">-42 (512)</td></tr><tr><td/><td/><td/><td/><td/><td/><td/></tr><tr><td align=\"left\">Antidepressants</td><td align=\"left\">14 (60)</td><td align=\"left\">13 (68)</td><td align=\"left\">-1 (35)</td><td align=\"left\">1 (13)</td><td align=\"left\">1 (13)</td><td align=\"left\">0 (0)</td></tr><tr><td align=\"left\">Anxiolytics</td><td align=\"left\">7 (23)</td><td align=\"left\">6 (18)</td><td align=\"left\">-1 (23)</td><td align=\"left\">6 (20)</td><td align=\"left\">6 (20)</td><td align=\"left\">0 (23)</td></tr><tr><td align=\"left\">Hypnotics</td><td align=\"left\">10 (24)</td><td align=\"left\">9 (18)</td><td align=\"left\">-1 (20)</td><td align=\"left\">4 (14)</td><td align=\"left\">5 (14)</td><td align=\"left\">1 (13)</td></tr><tr><td/><td/><td/><td/><td/><td/><td/></tr><tr><td align=\"left\">Total without intervention</td><td align=\"left\">2829 (3837)^a</td><td align=\"left\">2666 (3333)^b</td><td align=\"left\">-163 (2938)^d</td><td align=\"left\">2209 (2757)^a</td><td align=\"left\">2389 (2988)^b</td><td align=\"left\">180 (2346)^d</td></tr><tr><td/><td/><td/><td/><td/><td/><td/></tr><tr><td align=\"left\"><bold>Intervention: Visiting Service</bold></td><td align=\"left\"><bold>0 (0)</bold></td><td align=\"left\"><bold>553 (0)</bold></td><td align=\"left\"><bold>553 (0)</bold></td><td align=\"left\"><bold>0 (0)</bold></td><td align=\"left\"><bold>0 (0)</bold></td><td align=\"left\"><bold>0 (0)</bold></td></tr><tr><td colspan=\"7\"><hr/></td></tr><tr><td align=\"left\">Total with intervention</td><td align=\"left\">2829 (3837)^a</td><td align=\"left\">3220 (3333)^c</td><td align=\"left\">390 (2938)^e</td><td align=\"left\">2209 (2757)^a</td><td align=\"left\">2389 (2988)^c</td><td align=\"left\">180 (2346)^e</td></tr></tbody></table></table-wrap>" ]

[]

[]

[]

[]

[]

[]

[ "<table-wrap-foot><p>^aIntegral unit cost prices [17].</p><p>^bBased on average distances (in km), parking costs and travel + waiting + treatment</p><p>times (in hrs) for receiving treatment [17]</p><p>^cCosts of 1 km = € 0.16, 1 h parking = €2.50 €, 1 h patient's time = €8.30 [17].</p><p>^dValued as outpatient mental health services.</p><p>^eFrom DFL 77,00 in 1993, converted into Euro, indexed for 2003 (cf. <ext-link ext-link-type=\"uri\" xlink:href=\"http://www.cbs.nl\"/>) and</p><p>rounded.</p><p>^fValued as domestic help [17].</p></table-wrap-foot>", "<table-wrap-foot><p>^aFrom DFL 6,000 in 1997, converted into Euro, indexed for 2003 based on Health Care</p><p>Index (cf. <ext-link ext-link-type=\"uri\" xlink:href=\"http://www.cbs.nl\"/>) and rounded.</p><p>^bFrom DFL 6,000 in 1997, converted into Euro, indexed for 2003 based on General</p><p>Index (cf. <ext-link ext-link-type=\"uri\" xlink:href=\"http://www.cbs.nl\"/>) and rounded.</p></table-wrap-foot>", "<table-wrap-foot><p>^aNo significant difference between total costs at baseline (t0) in the experimental group and total costs at baseline (t0) in</p><p>the control group.</p><p>^bNo significant difference between total costs without the costs of the intervention at 1-year follow-up (t2) in the</p><p>experimental group and total costs at 1-year follow-up (t2) in the control group.</p><p>^cTotal costs including the costs of the intervention at 1-year follow-up (t2) in the experimental group differ significantly</p><p>from the total costs at 1-year follow-up (t2) in the control group at p &lt; 0.10 (p = 0.055).</p><p>^dNo significant difference between the cost difference (t2 - t0) without the costs of the intervention in the experimental</p><p>group and the cost difference (t2 - t0) in the control group.</p><p>^eNo significant difference between the cost difference (t2 - t0) including the costs of the intervention in the experimental</p><p>group and the cost difference (t2 - t0) in the control group.</p></table-wrap-foot>" ]

[ "<graphic xlink:href=\"1472-6963-8-128-1\"/>", "<graphic xlink:href=\"1472-6963-8-128-2\"/>", "<graphic xlink:href=\"1472-6963-8-128-3\"/>" ]

[]

{ "acronym": [], "definition": [] }

[ "<title>1. Background</title>", "<p>DNA microarray technology enables conducting experiments that measure RNA-transcript abundance (so called gene expression or expression degree) on a large scale of genomic sequences. The quality of the measurement systematically depends on experimental factors such as the performance of the measuring \"device\", e.g., on the chosen array-type, the design of the chip-platform and -generation and on the particular probe design, on one hand; and also on the quality of the sample, e.g. on the source of RNA and the used hybridization-pipeline including the protocol of RNA-extraction, -amplification and -labelling, on the other hand. Other essential factors affecting the quality of the expression measures are the quality and up-to-dateness of the genomic information probed on the chip and last but not least, the performance of the calibration algorithm which transfers raw intensity data into suited measures of transcript abundance. This so-called calibration step aims at removing systematic biases from the raw data which, in the ideal case, would allow the determination of the exact number of transcript copies of every probed transcript and thus direct comparison of expression measures independently of the used array type and sample preparation protocol.</p>", "<p>Apparent sources of variance can be, as for each experimental technique, divided into technical and biological ones, as well as, into systematic (see above) and random ones. The quality of the chip measurement and of the subsequent data calibration is characterized by their accuracy (the systematic bias between the measured and true expression value), precision (the uncertainty in replicated measurements), sensitivity (the expression range potentially covered by the measurement) and specificity (the selective power of the measurement to respond only to the specific targets).</p>", "<p>The development of appropriate calibration method requires in the first instance appropriate models and metrics to identify, to assign and to quantify the biases in each measurement. In the accompanying paper we presented the basics of the so-called hook-method, a simple and intuitive approach providing a natural metric system to characterize the hybridization on a particular array. The method divides into two essential constituents: (i) the analysis of the data in terms of the competitive two-species Langmuir hybridization model using the so-called hook-plot and (ii) the correction of the raw intensities for parasitic effects such as the non-specific hybridization, saturation and sequence-specificity to output expression measures in intrinsic units which are defined by the properties of the measuring device. The hook method is a strict single-chip calibration approach which treats each array as an independent measurement. This way the method accounts for chip-specific systematic effects which the calibration step intents to correct.</p>", "<p>In this paper we illustrate the performance of the hook method. We present examples dealing with different issues of array-measurements: the accuracy and precision of expression measures, the comparability of array experiments for different chip-generations, the effect of up-dating the probe assignments using latest genomic information, of RNA-quality and of different options of the preparation protocol such as labelling reagents and the type of the labelled molecule or replacing RNA-targets with DNA. We deliberately select a relatively wide range of different problems to illustrate the power of the method to estimate various systematic effects within a unique framework of chip-characteristic and to demonstrate the potential of developing new correction algorithms.</p>", "<p>In the first part of the paper we summarize the essential chip characteristics provided by the hook-method. In the second part special benchmark experiments are analyzed to estimate transcript related expression measures. The third part deals with hybridization quality control based on the hook analysis.</p>" ]

[]

[]

[]

[ "<title>4. Summary and Conclusion</title>", "<p>We presented a new method of microarray data analysis based on a physical model. This so-called hook method pre-processes the raw intensity data for further downstream analyses on one hand, and, on the other hand, provides chip characteristics with potential applications in hybridization quality control and array normalization.</p>", "<p>In this publication we illustrate the diagnostic potential of the hook-method by means of different chip- and transcript-related characteristics in various situations:</p>", "<p>- Using the data of spiked-in and dilution experiments it was shown that our single-chip approach provides accurate and precise expression measures over three orders of magnitude in units of the specific binding strength of the transcripts. The correction for saturation and probe-specific non-specific background assures linearity between the input (transcript concentration) and output (expression degree) measures. Among the four alternative measures, PMonly and PM-MM-difference measures perform best, but also the measure extracted from the S/N-ratio provides satisfactory results.</p>", "<p>- The \"present/absent\"-concept of detection calls originally introduced by Affymetrix provides straightforward, simple and helpful information which relates the signal of each transcript to the detection limit of the particular hybridization and, in addition characterizes the mean \"presence\" of transcripts in the hybridization solution. The hook-method calculates an analogous measure based on the break-criterion reflecting the onset of specific hybridization. This criterion implicitly takes into account the different correlations between the PM and MM probes upon non-specific and specific hybridization and thus it \"dynamically\" adapts to each particular hybridization. We have shown that this criterion well classifies into present and absent transcripts using data taken from the two-species yeast 2.0 array and from the golden spike experiment with known batches of \"empty\" probes.</p>", "<p>- The hook method performs reasonably well by comparing expression data of the same origin between two chip generations (HG-U133A and HG-U133 plus 2.0). The hook-diagnosis suggests that subtle differences of the hybridization law due to details of chip-manufacturing and/or -handling upon preparation give rise to slightly biased expression data between different array types and/or different batches of chips of the same type.</p>", "<p>- The re-assembly and filtering of probe sets based on improved genomic information increases the amount of probe sets detected as present ones. This result in turn shows that the hook-calling criterion applied to the original probe set definitions partly removes the \"bad\" (because of inconsistent probe assignments) probe sets from further analysis. The mean hybridization characteristics remain virtually unaffected by the redefinition step of the probe sets. The consequences of probe set-updating for the expression measures on transcript level will be studied separately.</p>", "<p>- The effect of 3'-biased RNA amplification gives rise to the progressively decreased specific hybridization of probes with increasing distance of their position relative to the 3'-end of the transcript which can be detected by hook-analysis using appropriate subsets of probes nearer to the 3'- and the 5'-end, respectively. This analysis properly differentiates between specific and non-specific hybridization where the latter one is, per definition, not affected by the 3'-biased intensity effect. Our data show that overall 3'/5'-signal ratios not considering the difference between specific and non-specific binding can lead to misinterpretations of the amplification bias.</p>", "<p>- Hook analysis reveals detailed insights into consequences of tissue-specific RNA-quality differences on hybridization and expression measures. Degradation of RNA increases the fraction of absent probes paralleled by the decrease of the specific binding strength and counterbalanced by the increase of non-specific background hybridization. Improper separation of both opposite effects can pretend expression changes into the wrong direction. We suggest that the chip characteristics provided by the hook method can serve as calibration benchmarks for alternative normalization algorithms which take into account the different behaviour of the specific and non-specific signal in samples of varying RNA-quality.</p>", "<p>- The variation of the labelling protocol and substitution of RNA-targets by DNA modifies the probe/target interactions. Hook analysis shows for example that DNA-targets, and to a smaller degree, the Affy-labeling protocol (no labelling of cytosines) improve the specificity of the method compared with RNA-targets and the previous ENZO-protocol, respectively. For DNA-targets the sequence correction is of much smaller impact because of smaller sequence-induced variability of the raw intensities.</p>", "<p>In summary, sequence correction and especially the quantification of the non-specific background contribution for each probe enable subtle diagnosis of the hybridization on each array. To extract this information the hook method combines the intensities of each PM/MM-probe pair and utilizes the different properties of both probe types. Here the MM behave like \"weak-affine\" PM and serve as intrinsic reference for the PM over the whole potential concentration range of the transcripts. We illustrated that this intrinsic referencing might be extremely useful for dealing with practical issues of expression analysis such as RNA-quality, hybridization control and calibration of expression measures. This publication outlined several potential applications of the method which will be addressed in our future work.</p>" ]

[ "<p>This is an Open Access article distributed under the terms of the Creative Commons Attribution License (<ext-link ext-link-type=\"uri\" xlink:href=\"http://creativecommons.org/licenses/by/2.0\"/>), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.</p>", "<title>Background</title>", "<p>Microarray experiments rely on several critical steps that may introduce biases and uncertainty in downstream analyses. These steps include mRNA sample extraction, amplification and labelling, hybridization, and scanning causing chip-specific systematic variations on the raw intensity level. Also the chosen array-type and the up-to-dateness of the genomic information probed on the chip affect the quality of the expression measures. In the accompanying publication we presented theory and algorithm of the so-called hook method which aims at correcting expression data for systematic biases using a series of new chip characteristics.</p>", "<title>Results</title>", "<p>In this publication we summarize the essential chip characteristics provided by this method, analyze special benchmark experiments to estimate transcript related expression measures and illustrate the potency of the method to detect and to quantify the quality of a particular hybridization. It is shown that our single-chip approach provides expression measures responding linearly on changes of the transcript concentration over three orders of magnitude. In addition, the method calculates a detection call judging the relation between the signal and the detection limit of the particular measurement. The performance of the method in the context of different chip generations and probe set assignments is illustrated. The hook method characterizes the RNA-quality in terms of the 3'/5'-amplification bias and the sample-specific calling rate. We show that the proper judgement of these effects requires the disentanglement of non-specific and specific hybridization which, otherwise, can lead to misinterpretations of expression changes. The consequences of modifying probe/target interactions by either changing the labelling protocol or by substituting RNA by DNA targets are demonstrated.</p>", "<title>Conclusion</title>", "<p>The single-chip based hook-method provides accurate expression estimates and chip-summary characteristics using the natural metrics given by the hybridization reaction with the potency to develop new standards for microarray quality control and calibration.</p>" ]

[ "<title>2. Chip characteristics</title>", "<title>Hook parameters</title>", "<p>Figure ##FIG##0##1## depicts a typical graphical output-summary of the hook-analysis for two hybridizations performed on two different chip-types taken from the Genelogic dilution [##UREF##0##1##] and the GoldenSpike [##REF##15693945##2##] experimental series (see also Figure ##FIG##1##2## with data taken from the HG-U95 Latin square spiked-in series [##UREF##1##3##]). The Δ-vs-Σ plots characterize the hybridization of the particular chip. They are obtained by transforming the probe intensities of one GeneChip microarray into Δ = logI^PM- logI^MMand Σ = 0.5(logI^PM+ logI^MM) coordinates and subsequent smoothing (I^PMand I^MMdenote the spot intensities of the PM and MM probes after optical background correction; the logs are base 10 throughout the paper). The corrected version of the Δ-vs-Σ plot uses intensity values which are corrected for sequence-specific sensitivity effects. These plots are called hook-curves because of their typical shape. Additional characteristics of a particular chip-hybridization are the signal-density distribution and the four positional-dependent sensitivity profiles of the PM and MM probes upon specific and non-specific hybridization, respectively. These profiles are calculated from the intensity data of the chosen chip and used to correct the intensities for sequence-specific affinities.</p>", "<p>The corrected hook-data are well fitted by the Langmuir-absorption model which predicts the theoretical curve shown in Figure ##FIG##0##1##. The fit provides characteristic parameters (see Table ##TAB##0##1##, see the accompanying paper [##REF##18759985##4##] for details) of the particular hybridization judging properties such as the mean non-specific and specific signal, the saturation intensity and the mean PM/MM- gain of the sensitivity caused by the central mismatch of the MM probes (see Table ##TAB##1##2##, data are taken from the hook-analyses of more than 500 GeneChip arrays of different type and origin, see also [##UREF##2##5##] for details). Note that selected characteristics such as the non-specific binding strength (width) and the PM/MM-gain (height) are directly related to the geometrical dimensions of the hook-curve. Hence, the respective characteristics can be roughly and simply estimated by visual inspection of the Δ-vs-Σ plot.</p>", "<p>Different parts of the hook have been assigned to (see Figure ##FIG##1##2## from the left to the right) the N (non-specific)-, mix (mixed)-, S (specific)-, sat (saturation)- and as (asymptotic)- regimes of hybridization. These regimes reflect the fact that the contribution of specific hybridization to the spot intensities progressively increases along the rising part of the hook from tiny amounts in the N-regime to about 100% near the maximum. In contrast, the degree of saturation progressively increases along the decaying part from almost no saturation effects near the maximum to complete saturation in the as-regime. Note the considerable distortion of the N- and mix-regimes between the raw and corrected hooks. These marked differences between both hook-versions emphasize the importance of the correction step.</p>", "<p>The N-range of the hook-curve is characterized by the variance of the underlying probe-level data, σ, which are well described by a normal distribution. The mean specific signal of the particular hybridization, &lt;λ&gt;, is calculated as log-mean of the S/N-ratio of the probe sets beyond a certain threshold (e.g. R &gt; 0.5, see below). Note that the distribution of the specific signal is well approximated by an exponential decay in many cases. Then, the characteristic \"decay\" constant λ defines the Σ-range over which the probability of detecting a signal decays by one order of magnitude.</p>", "<title>Hook curves of different chip generations</title>", "<p>Figure ##FIG##2##3## shows a collection of representative hook-curves taken from four hybridizations of human-genome chips of different generations. Along the chip generations the spot-size of the probes decreases from 20 μm (U95), over 18 μm (U133A) to 11 μm (U133-plus2). The reduction of spot-size has enabled to increase the number of probe sets per chip from 16.000 over 22.000 to 54.000, respectively [##UREF##3##6##,##UREF##4##7##]. In addition, this development is accompanied by modifications of the reagent-kits and the scanning technique [##UREF##4##7##,##UREF##5##8##]. Importantly, also probe design and selection have been improved by applying more sophisticated genomic and thermodynamic criteria especially for chip generations following the U95. Chip data shown in Figure ##FIG##2##3## refer to RNA prepared from tissue samples (thyroid nodules; [##REF##16407496##9##]) and to Universal Human Reference RNA [##REF##16776839##10##].</p>", "<p>The different shapes of the uncorrected hook curves of the U95 and U133 chips, particularly the broader N-range of the former one, can be explained by the partially suboptimal quality of the probe selection for the U95-generation (which also applies to the design of the DG1-chip shown in Figure ##FIG##0##1## and Figure ##FIG##1##2##) containing a relatively high number of weak-affinity probes. For the U133 series the N-range considerably narrows essentially due to better quality of the probes. It is important to note that our affinity correction levels out this difference to a large extent providing corrected hook curves of very similar shape for chips of different generations such as the U95 and U133 arrays.</p>", "<p>We obtained analogous results for hundreds of GeneChip expression arrays of different specifications: chip generations, species (human, mouse, rat, drosophila, rice, arabidopsis etc.) and samples (patient cohorts, cell lines, benchmark experiments) [##UREF##2##5##]. Table ##TAB##1##2## lists typical parameter-ranges obtained in these studies. For example, the PM/MM-affinity gain for specific hybridization shows that the central mismatch of the MM causes on the average the nearly tenfold (s ~ 7–11) increase of sensitivity of the PM-probes compared with that of the MM. On the contrary, for non-specific binding one expects on the average the same sensitivity for the PM- and MM-probes. The respective PM/MM-gain parameter however indicates a small but significantly increased PM-sensitivity, n ~ 1.05 – 1.25. We tentatively attribute this effect to false positive detections in the N-range, i.e. to a certain amount of specific hybridization among the absent probes (see below). The relatively narrow data-range of the obtained hybridization characteristics reflects the common physical-chemical basis of the method which is determined by properties such as the oligonucleotide density and size of the probe spots, the common MM probe-design and hybridization conditions. A particular example which demonstrates apparent inconsistencies between the expression estimates obtained from different chip-generations will be given below.</p>", "<title>Detection call</title>", "<p>The onset and further increase of specific binding gives rise to a characteristic breakpoint of the hook curve which clearly separates the N- and mix- hybridization ranges. The corresponding change of the slope of the hook curve can be rationalized in terms of relatively strongly correlated PM- and MM-intensities in the N-range which progressively \"decouple\" upon increasing amount of specific binding because it much stronger affects the PM than the MM. We use the breakpoint to classify the probe sets into absent and present ones in analogy with the detection call provided by MAS5 [##UREF##6##11##].</p>", "<p>To verify the used break-criterion in a simple illustrative fashion we analysed two special chip hybridizations. The GeneChip Yeast Genome 2.0 Array (YG 2.0) contains probe sets to detect transcripts of both, the two most commonly studied species of yeast, Saccharomyces cerevisiae and Schizosaccharomyces pombe. The YG 2.0 array thus includes 5,744 probe sets for 5,841 of the 5,845 genes present in S. cerevisiae and 5,021 probe sets for all 5,031 genes present in S. pombe. The evolutionary divergence between S. cerevisiae and S. pombe over 500 million years ago caused enough sequence divergence between the two species to require selection of separate probe sets for all genes, even the closest cross-species orthologs [##UREF##7##12##]. Due to this sequence divergence one expects only weak cross-species hybridization.</p>", "<p>Figure ##FIG##3##4## shows the hook plot for a hybridization of the array with RNA from S. cerevisiae [##REF##17043222##13##]. The break criterion provides a total absent rate of 47% which well agrees with the percentage of probe sets for S. pombe printed on the chip (~47%). Species-specific masking indicates that the absent probes originate nearly exclusively from the probe sets designed for S. pombe which indeed accumulate nearly completely in the N-range of the hook whereas the S. cerevisiae-probe sets cover the mix-, S- and sat-ranges as expected. About 5% of each fraction \"overlap\", i.e. they refer to present probe sets of S. pombe and absent sets of S. cerevisiae, respectively.</p>", "<p>The second example was taken from the Golden Spike experiment in which PCR products from a Drosophila Gene Collection referring to 3,860 probes were spiked onto Drosgenome DG1-arrays [##REF##15693945##2##]. On this array 10,131 probe sets out of the total number of 14,116 are called ,empty' because they are not assigned to any of the added cRNA spikes. Again the absent rate of 70% agrees with the fraction of empty probes (~72%). Selective masking of either the spiked or the empty probe sets shows that the latter ones indeed accumulate in the N-region and are called absent whereas the spikes are predominantly flagged as present (see right part in Figure ##FIG##3##4##).</p>", "<p>The selective masking in these both examples shows that the simple break criterion gives rise to false present calls (of potentially absent probes) of less than 5 – 7% even if one neglects cross hybridization. The break-criterion provides a sort of detection limit for the specific expression signals. The detection call thus divides the probe sets into subsets with detectable and essentially not-detectable amounts of transcripts. The false present and false absent rates depend on the degree of cross hybridization and on other factors which will be addressed below.</p>", "<p>In the next section we present other examples showing that the hook method reasonably estimates the detection limit of the particular array in terms of present and absent calls. The alternative calling-algorithm implemented in MAS5 calculates the so-called discrimination score (DS) of each probe pair which is directly related to its Δ-value [##REF##18759985##4##,##UREF##6##11##]. Then, one-sided Wilcoxson's rank test is applied to the DS-values of each probe set together with appropriate threshold-settings to estimate whether the set is present or absent. The used test strongly penalizes negative PM-MM signal differences. More than 40% of all probe pairs amount to such \"bright MM\" (because MM &gt; PM) in the N-range whereas its percentage steeply decreases with increasing Σ and virtually disappears in the S-range of the hook [##REF##15834006##14##]. This trend explains the correlation between the call-rate obtained by both methods (see next section). For the examples presented here MAS5 provides a distinct smaller (36%) and an equal (70%) absent rate for the yeast and golden spike hybridizations, respectively.</p>", "<p>On the other hand, the hook criterion includes both, the PM-MM difference in terms of the Δ coordinate and the mean total signal in terms of Σ. The latter value adds a second threshold which prevents probe sets with relatively strong mean signals to be called absent. Moreover, the break-criterion detects rather the change of the mutual correlation between the PM and MM signals caused by the onset of specific hybridization than a certain fixed signal level. As a result, the hook-criterion \"dynamically\" shifts with varying signal level using the break as a simple and reasonable landmark whereas the MAS5 threshold is statically and less intuitively given in terms of p-values typically predetermined by the default settings of the used analysis program.</p>", "<title>3. RNA-expression</title>", "<title>Benchmark experiments with variable transcript concentration</title>", "<p>Figure ##FIG##4##5## and Figure ##FIG##5##6## show the hook curves, the absent calls and concentration measures of two special benchmark experiments. In the GeneLogic dilution series, cRNA from human liver tissue was hybridized on HG-U95 GeneChips in various amounts [##UREF##0##1##]. The decrease of the degree of non-specific binding upon dilution widens the horizontal dimension of the hook curve (see upper panel in Figure ##FIG##4##5##). Dilution decreases the concentration of specific and non-specific transcripts in a parallel fashion leaving their concentration ratio virtually constant. As expected, the S/N-ratio R of selected probes remains essentially constant whereas the binding strength of specific binding progressively decreases (compare solid symbols and thick lines in the lower panel of Figure ##FIG##4##5##).</p>", "<p>The hook-method provides a virtually constant fraction of absent probes independent of the dilution step (see middle part in Figure ##FIG##4##5##). This result can be rationalized in terms of the condition of R = const, which corresponds to virtually constant ordinate values, Δ ≈ const, in the mix-range of the hook-plot (see dotted horizontal lines in the upper panel in Figure ##FIG##4##5##). The horizontal shift of the hook upon dilution only weakly affects the fraction of probes below and above a certain R-value. Also the fraction of probes below and above the break criterion for classifying the probe sets into present and absent ones remains essentially constant. The virtually constant absent rate properly reflects the invariant composition of the hybridization solution. Contrarily, the fraction of absent calls estimated by MAS5 progressively increases upon dilution.</p>", "<p>In the U133-spiked-in series of Affymetrix, a set of selected RNA-transcripts (the spikes) is added in definite concentrations to the hybridization solution [##UREF##1##3##]. The hybridization cocktail also contains a RNA-extract from HeLa-cells to mimic complex hybridization conditions. Figure ##FIG##5##6## shows the typical hook-curve calculated from the intensity data of one chip of this experiment. The blue curve corresponds to the probe sets which are mainly hybridized with the non-spike RNA of the added background. The Δ-vs-Σ-coordinates of the probe sets detecting the spikes are shown by open circles. Their positions cover the full range of the hook curve and shift to the right with increasing transcript concentration (0 – 512 pM). Note that the distance of the position of a particular probe set relative to the end point is inversely related to the specific binding strength and thus to the specific transcript concentration.</p>", "<p>Spike probe sets without specific transcripts (0 pM) and with transcripts of only tiny concentrations (&lt; 0.5 pM) assemble mainly within the N-range of the hook curve. Figure ##FIG##5##6## compares the absent call rates for the spikes obtained from the hook and MAS5 methods which both show similar results. The probability of flagging a probe absent increases upon decreasing transcript concentration. The absent rate thus reflects the resolution limit of the method for detecting small transcript concentrations. The vertical shift between the MAS5 and hook data can be adjusted by changing the threshold-parameters used in both methods.</p>", "<p>The fit of the hook-equation provides the S/N-ratio R for each set of spiked-in probes which linearly correlates with the spiked in concentration (Figure ##FIG##5##6##, lower panel). The vertical axes in this figure show that the largest spike-concentration (512 pM) corresponds to a S/N-ratio of R≈ 200 (left axis) and to the specific binding strength of X^S≈ 1 (right axis). Comparison of the absent rates with the S/N-ratio indicates that the threshold for present calls refers to R ≈ 0.1 – 2 and to a binding strength for specific hybridization of X^N≈ (0.5 – 5) 10^-3(see dashed arrows in Figure ##FIG##5##6##). Hence, the relevant measuring range of R and X^Ncovers about three orders of magnitude.</p>", "<title>Expression estimates</title>", "<p>The hook-methods provides potentially four alternative expression measures of each probe set: the S/N-ratio R, which is obtained from the direct fit of the transformed two-species Langmuir isotherm to the hook curve; and PMonly, MMonly and PM-MM-difference estimates which are calculated as the mean generalized logarithm of the background- and sensitivity corrected and de-saturated signal values averaged over the background distribution. The corrections for the latter three expression values are estimated from the hook-curve analysis. Figure ##FIG##6##7## compares the performance, accuracy and precision of the different alternative measures in terms of their correlation with the known spiked-in concentration. The precision reflects the scattering of the estimated data about their mean and was therefore estimated as the respective coefficient of variation. The accuracy reflects the systematic deviation of the estimated from the spiked concentration. Hence, it was quantified as the ratio of the estimated concentration and the known concentration of the spikes. For sake of comparison we also show RMA (robust multiarray analysis, [##REF##12582260##15##,##UREF##8##16##]) expression estimates in Figure ##FIG##6##7##.</p>", "<p>It turns out that all considered methods except MMonly are comparably precise at larger transcript concentrations c^sp-in&gt; 2 pM, at which the transcripts are safely called present (see previous paragraph). Note that the direct fit of the hook equation to the data provides the S/N-ratio which represents only a rough measure of the expression degree. The PMonly and PM-MM estimates more precisely correct the signals for the non-specific background contribution. It does therefore not surprise that these measures outperform the S/N-ratio R at smaller c^sp-in-values in terms of precision. The MMonly expression values are by far the most imprecise ones which does not surprise because the specific signal level and thus the sensitivity of the MM-probe intensities are smaller by nearly one order of magnitude compared with the respective PMonly and PM-MM measures at a comparable non-specific background level. The coefficient of variation of the MMonly expression estimates exceeds CV &gt; 2 over the whole concentration range which exceeds the maximum scaling used in Figure ##FIG##6##7##.</p>", "<p>The hook-measures clearly outperform the RMA-values in terms of the accuracy of the expression values. Note that RMA uses a linear intensity approximation which ignores saturation at high transcript concentrations at one hand-side and corrects the intensities for non-specific hybridization using a global background level on the other hand-side. As a consequence, RMA systematically underestimates the change of the expression values especially at high and small transcript concentrations (see also [##UREF##2##5##] for a detailed discussion). Note that RMA represents a multichip- method which processes a series of chips to adjust the probe-specific sensitivities. In contrast, the hook method provides strictly single-chip estimates which are based on the intensity information of only one particular chip. The accuracy of the PM-MM estimates perform best among the methods at small transcript concentrations presumably because the explicit use of the MM intensities well corrects for sequence-specific background effects not considered by the positional dependent sensitivity model used by the hook method.</p>", "<p>In this context we explicitly refer to the so-called effect of \"bright\" MM, i.e. a certain amount of about 40–50% of negative PM-MM intensity differences on each chip [##UREF##9##17##,##UREF##10##18##]. This systematic bias has been explained by the intrinsic purine-pyrimidine asymmetry of base pairings in the non-specific DNA/RNA probe/target duplexes [##REF##15834006##14##,##UREF##11##19##,##REF##16171364##20##]. The sensitivity correction used by the hook method explicitly corrects the raw intensity data for this sequence effect.</p>", "<title>Reproducibility across GeneChip-generations</title>", "<p>Up to now a large number of microarray data has been collected in public repositories such as GEO (Gene expression Omnibus of NCBI) or ArrayExpress (EBI) referring to a wide variety of different conditions, specimen and array-types. One important challenge in microarray analysis is to take full advantage of these previously accumulated data, e.g., for combining different datasets to get a more comprehensive view in comparative analyses. Difficulties related to the heterogeneous character of array platforms, chip types and hybridization protocols in most cases hinder such meta-analyses. Consistencies and inconsistencies between chip platforms and -types have been previously addressed in a number of studies [##REF##16964229##21##, ####REF##17061323##22##, ##REF##16964228##23##, ##REF##12823866##24##, ##REF##18005448##25####18005448##25##].</p>", "<p>A recent study reports that even identically composed probe sets containing identical numbers and sequences of probes on different GeneChip-types can produce significantly different values of gene expression in cross-chip comparisons for samples containing the same target RNA [##REF##16776839##10##]. Particularly, this study compares the newer HG-U133 plus 2.0 (P-chip) with the previous-generation HG-U133A (A-chip) array. The nearly 55.000 probe sets of the former chip integrate the more than 22.000 probe sets of the HG-U133A chip and, in addition, the probe sets of the HG-U133B array. In the study both, the A- and P-arrays were hybridized with the same Universal Human Reference RNA.</p>", "<p>For subsequent comparison of the expression values the authors masked the additional probe sets on the P-chip (\"not A\"-probes) and processed only the common probe sets present on both chips (\"A\"-probes) using MAS5 and a combination of global and invariant-set normalizations (see ref. [##REF##16776839##10##] for details). The analysis revealed a number of differentially expressed genes which is much larger than the number expected by chance despite the identical probes and target RNA.</p>", "<p>Figure ##FIG##7##8## compares the expression values of four probe sets selected by Zhang et al. as representative examples ranging from small to high expression levels to illustrate the bias caused by the chip-types (see also Fig. ##FIG##2##3## in ref. [##REF##16776839##10##]). Note that the difference between the expression values of both chip-types inverses sign upon increasing expression suggesting that simple re-scaling of the data does not solve the problem.</p>", "<p>We re-analyzed these chip-data using the hook-method. The left part of Figure ##FIG##7##8## shows that the systematic difference between the chip-types essentially disappeared at small expression levels and it is clearly reduced compared with the data of Zhang et al. at larger expression levels. Parallel analyses which either consider or not consider the not A-probes provide virtually the same results (data not shown). We tentatively attribute this improvement to the sequence correction of the intensities and to the proper estimation of the non-specific background correction.</p>", "<p>In the next step we compare the hook-curves of the P- and A-chips to identify possible differences of their hybridization characteristics. Examples of raw and corrected hooks taken from this series are shown in Figure ##FIG##2##3## (see the two parts on the right). In Figure ##FIG##8##9## we re-plotted the corrected hooks and the density distributions for direct comparison. The characteristics of the P-chip were calculated using either all probes or the two subsets of probes shared (A-probe sets) and not-shared (not-A-probe sets) with the A-array. All hook versions fit well to the theoretical function. Table ##TAB##2##3## summarizes the extracted parameter values.</p>", "<p>The widths of the hooks and thus the respective level of non-specific binding are virtually the same for the P- and A-arrays. The not-A-probe sets are, on the average, distinctly less expressed than the A-probe sets as indicated by the more than twice as large amount of absent probes (%N = 64% versus 29%) and the smaller decay rate of the respective density distribution (λ = 0.45 versus 0.65). The percentage of absent probe sets on the P-chip (50%) represents the average of the respective contributions of A- and not-A-probes where the not-A-probes obviously add a considerable larger amount than the A-probes. The total density distribution of the P-chip well agrees with the distribution of the not-A-probes in the N-range and with that of the A-probes in the S- and sat-ranges. In summary, the hybridizations on both chips well agree in terms of the general target properties (N-background, decay rate) but differ with respect to the general probe characteristics (%N). The latter effect simply reflects the different probe-selections of the manufacturer for each chip type.</p>", "<p>Besides these essentially common characteristics, the hook-analysis revealed one significant difference between the chip types, namely the significantly increased height parameter α for the A-chips. This parameter characterizes the PM/MM-gain of the specific signals, or, in other words, the mean incremental effect of introducing one central mismatch into specific probe/target duplexes. Here one expects however virtually identical α-values for the A- and P-chips because the mismatch design and the nominal probe length are identical for both array-types. On the other hand, subtle deviations from the nominal probe design owing to deficiencies of fabrication and/or variations of the hybridization conditions in different preparations can however affect the observed maximum PM/MM ratio: For example, the in-situ synthesis of the GeneChip probes usually produces a non-negligible fraction of truncated probe-oligomers not synthesized to full nominal length. This effect gives rise to systematic deviations from the Langmuir isotherm and, more importantly, it will affect the PM/MM-gain because the relative effect of one middle-mismatch is expected to increase with decreasing length of the probe oligomers [##UREF##12##26##,##REF##17939865##27##]. Also the post-hybridization washing step upon chip preparation is expected to affect the apparent PM/MM-ratio and the binding law as well [##UREF##13##28##,##UREF##14##29##]. We suggest that subtle differences of the hybridization law due to details of chip-manufacturing and/or handling of the chips upon preparation as well as evolving instrumentation and instrument protocols give rise to slightly biased expression data between different array types and/or different batches of chips of the same type. The latter conclusion was derived from another chip series for which we observed a reversed relation of the PM/MM-gain, namely a larger value for the P-array compared with the A-array [##UREF##2##5##] (see also the two A-chips in Figure ##FIG##2##3##). Selected hook parameters can serve as indicators of such effects and can provide hints for their origin.</p>", "<title>Updated probe sets</title>", "<p>One possible approach to partially level out chip-type specific differences is the matching of the probe sets of different array types using genomic sequence information updated with respect to the original probe set assignment of the manufacturer. Recent studies show that significant percentages of existing GeneChip probe set definitions are no longer consistent with gene and transcript assignments in actual versions of public databases. The probe identity issue is of critical importance, as it significantly affects the expression values summarized on probe set level and thus their interpretation and understanding [##REF##16284200##30##,##REF##17288599##31##]. Dai et al. [##REF##16284200##30##] performed reanalysis of probe and probe set annotations resulting in publicity available, regularly updated probe set definitions for most of the GeneChip-types. A series of probe selection and grouping criteria utilizing the latest sequence and annotation information taken from databases such as REFSEQ or ENSEMBLE (gene, transcript and exon based) are applied. (i) This filtering removes \"bad\" probes either without or with multiple perfect match hits along the genomic sequence and, (ii) it re-arranges \"redundant\" probe sets addressing the same gene, transcript or exon into one probe set. The resulting updated probe sets contain variable numbers of probes ranging from four to more than thirty. The mean probe set size is increased for gene- and transcript related sets (e.g., for the HG-U133A array: ENSEMBLE(gene)~14.9; ENSEMBLE(transcript)~13.9; Refsequ~14.9) and decreased for exon-related sets (ENSEMBLE(exon)~9.3) compared with the original Affymetrix set definition (NetAffx~11.1).</p>", "<p>In Figure ##FIG##8##9## and Table ##TAB##2##3## we compare the hook characteristics for different probe set definitions. All updated probe set definitions under consideration give rise to very similar hook curves which essentially also agree with that obtained from the original probe set-assignments. This result again shows that the expressed probe sets follow the same hybridization law where changes of their performance will change their position along the hook. Interestingly, also the decay rates of the density distributions and the mean expression index &lt;φ&gt; of about 2.1 are very similar for all considered cases. This result indicates that the expression degrees of present probe sets located in the mix-, S- and sat-ranges of the original hooks remain, on the average, essentially unchanged after updating the probe sets.</p>", "<p>The amelioration of the probe sets masks out a certain amount of \"bad\", i.e. falsely annotated or ambiguous probes and merges redundant probe sets (see above). As a consequence, the fraction of absent probe sets notably decreases from 34% (A-chip) and 50% (P-chip) to about 20% in both cases (see Table ##TAB##2##3##). The percentage of probe utilization inversely correlates with the reduction of the amount of absent probes detected by the hook method between the original and updated probe sets (see Table ##TAB##2##3##). For example, about 70% probes of the A-chip but only 50% of the P-chip are used after updating the gene-annotations. The obtained common percentage of absent probe sets of 20% reflects the consistent filtering criteria applied to both chip types. Indeed, the verification of probe sets based on genomic sequence data comes out with similar percentages of modified and not-modified probe sets sharing the same target in the original and updated probe set definitions.</p>", "<p>In summary, the verification of probe sets increases the amount of the probe sets detected as present ones on one hand. Hence, the hook-calling criterion automatically removes the \"bad\" probe sets from further analysis. On the other hand, the mean expression degree and the hybridization characteristics reported by the ensemble of probes synthesized on the chip remain virtually unaffected by the redefinition step. Comparison of the updated expression measures of the slightly diverging probe sets shown in Figure ##FIG##7##8## after verification leaves the small systematic biases essentially unchanged (data not shown).</p>", "<title>4. Hybridization control</title>", "<p>Assessment of data quality is an important component of the analysis pipeline for gene expression microarray experiments. Essentially all steps of RNA-preparation (extraction, amplification, in-vitro transcription, labelling), hybridization, washing and signal detection can have significant effects on the extracted \"apparent\" expression values seen between different samples with consequences for subsequent downstream applications. There are, for example, \"technical\" factors associated with the correction for background fluorescence owing to bleed over-effects from surrounding probes on the arrays [##UREF##15##32##], or to spatial artefacts [##REF##16430768##33##,##UREF##16##34##]. Another kind of effects are linked with the RNA integrity and the used amplification and labelling protocols [##REF##15038166##35##, ####REF##16945445##36##, ##UREF##17##37##, ##REF##18298816##38##, ##UREF##18##39####18##39##]. In this section we demonstrate the potential of the hook-analysis to detect and to estimate variations of the data owing to RNA-quality, the effect of substitution of cRNA by cDNA and of the labelling protocol.</p>", "<title>RNA-amplification bias</title>", "<p>The amplification step of cRNA-preparation uses reverse transcriptase primers starting from the 3' -end of the original mRNA resulting in a population of 3' -biased, truncated transcript fragments. This 3'-overrepresentation gives rise to the systematic lowering of signal-intensities when the position of the probes shifts towards the 5'-end [##REF##15038166##35##,##UREF##19##40##,##REF##14606961##41##]. Hence, the probes designed for detecting one and the same transcript apparently report a progressively decreasing expression degree with increasing distance from the 3'-end of the transcripts. This is potentially detrimental for the expression value of the probe set summarized from individual probe-level data.</p>", "<p>To illustrate the consequences of the 3'-biased amplification on the hook-data we ranked each probe in each probe set according to its position from the 3'-end, calculated the Δ- and Σ-coordinates as average value over probes no. #1 – #4 (subset more closely to the 5'-end), #8 – #11 (subset more closely to the 3'-end) and #1 – #11 (total probe set) and presented the hook-plots, the density distributions and the total Σ-coordinates as a function of the \"sub-Σ-values\", Σ_subin Figure ##FIG##9##10##. This approach considers the sequential ordering of the probes as a rough measure of their actual position along the respective gene to estimate the mean effect of the 3'-biased transcript populations on the hook-characteristics.</p>", "<p>As an example, the figure compares two biological replicates A and B of total RNA prepared from rat muscle hybridized on rat genome RG-230 GeneChip arrays. Before microarray analysis RNA integrity and concentration was examined on an Agilent 2100 Bioanalyzer (Agilent Technologies, Palo Alto, CA, USA) using the RNA 6.000 LabChip Kit (Agilent Technologies) according to manufacturers instructions. Quantification of 28S and 18S ribosomal RNA before target amplification using the T7-protocol (see [##REF##14647279##42##,##REF##16945445##36##] and [##REF##16469371##43##] and references cited therein) revealed virtually equal RNA quality from both preparations according to the 28S/18S-ratios of 1.45 (sample A) and 1.43 (B).</p>", "<p>Figure ##FIG##9##10## (upper part) correlates the total probe set average of the probe intensities, Σ, with that of the subsets, Σ_sub. In the N-hybridization range both sub-averages agree each with another. This result is plausible because the 3'-bias due to incomplete amplification of full-length transcripts applies per definition only to specific hybridization: Non-specific transcripts are not-specified with respect to their position relative to the 3'-end and thus they on the average hybridize equally to all probes regardless of their relative position. Upon increasing mean intensity-values Σ_suband Σ, the curves however split into two branches starting with the onset of specific binding. The 3'-biased sub-average exceeds that of the 5'-biased one by a factor of 3.7 (sample A) and 1.9 (sample B) in the S-range. This difference indicates the more uniform amplification in sample B providing a higher yield for longer transcripts. Note that the observed onset of the split between the 3'- and 5'-branches well agrees with the position of the break of the respective hook curve (see vertical dotted line). This type of analysis thus once more confirms the chosen break-criterion to estimate the boundary between the N- and mix-hybridization ranges along the hook curve.</p>", "<p>The total hook and the 3'- and 5'-\"subhooks\" of each sample are well described by the same theoretical function using a common set of parameters (see middle panel in Figure ##FIG##9##10##). To a good approximation, all probes obey the same hybridization law irrespective of their position relative to the 3'-end and irrespective of their amplification yield. The intensities of the probes near the 3'-end however cover a larger Σ-range compared with the respective 5'-biased subset. This effect is manifested by the larger decay constant λ of the signal distribution of the 3'-biased probes compared with that of the 5'-biased probes as illustrated in the lower part of Figure ##FIG##9##10##. The larger λ indicates the better (specific) signal to non-specific background (S/N)-ratio: the average specific signal level is larger in units of the detection limit which is roughly given by the non-specific background. The 3'-subset is also characterized by the larger values of the (negative-logarithmic) expression index φ. It reflects the larger average strength of specific binding owing to the larger fraction of specific full-length transcripts.</p>", "<p>The smaller 3'/5'-ratio in the S-range, the smaller expression index and the larger decay constants, λ, of sample B compared with sample A reveal a generally larger fraction of specific transcripts due to more complete amplification and thus a better RNA-quality. The hook-analysis also reveals that the larger fraction of full length transcripts in sample B is accompanied by a slightly smaller width of the hook and a smaller fraction of absent probes (see middle panel of Figure ##FIG##9##10##). The latter trend can be simply attributed to the fact that the mean occupancy of the probes with specific transcripts and thus the specific signal increases if one improves the RNA-quality in terms of longer transcripts. Note that the increase of the decay constant upon RNA-improvement means that the probe sets on the average shift towards the S-range of the hook-curve. This trend is accompanied by a reduction of the percentage of absent probes from 54% to 42%. MAS5-analysis provides a similar difference of the amount of absent probes with 63% and 53% for the A- and B-samples, respectively.</p>", "<p>The narrowing of the hook upon improvement of RNA quality, indicates a larger relative amount of non-specific binding. This trend seems peculiar because one might expect that the larger amount of specific transcripts reduces the amount of non-specific binding. However, the more efficient amplification step in sample B results in a higher total number of full length transcripts and/or in a larger binding constant for non-specific binding and thus in an increased binding strength of non-specific binding which, in turn, gives rise to the increased level of cross-hybridization as indicated by the slightly narrower hook-curve. Note however that the decreased quality of the RNA-amplification only weakly shifts the rising branch of the hook curve, in contrast to the overall dilution effect shown in Figure ##FIG##4##5##.</p>", "<p>Microarrays of the GenChip-design contain special probe sets for estimating the 3'/5'-amplification bias. They refer to relatively long transcripts such as β-actin and GADPH with probe sets targeting the transcription of their 3'-, mid- (m), and 5'-regions. Small 3'/5'-signal ratios are generally thought to indicate small amplification bias and thus good amplification quality.</p>", "<p>Figure ##FIG##10##11## compares the hook-coordinates (Σ and Δ) and two expression measures (MAS5 and hook/PMonly) of the three GADPH-probe sets in both samples. The intensity-related Σ-values of the different probe sets only marginally differ for sample A \"pretending\" this way better RNA-quality than for sample B with markedly larger 3'-values (see the fold changes above the bars: 1.6×-versus-11× for A and B, respectively). Comparison of the Σ-values with the break criterion however clearly indicates that the GADPH-signals of sample A are dominated by non-specific hybridization which was shown to level-out 3'/5'-expression differences (see also the open circles in the middle panel of Figure ##FIG##9##10## which indicate the position of the 3'- and 5'-probe sets of GADPH along the hook curve). Note that both, the 5'- and m-sets of sample A are called absent by the hook method. Contrarily, all GADPH-sets are present in sample B. Their 3'/5'-ratio consequently can be attributed more reliably to the amplification bias whereas that of sample A simply reflects the virtual absence of GADPH-transcripts. Note that the expression values calculated by MAS5 and, to an even larger degree, hook (PMonly) reflect drastically increased 3'/5'-ratios owing to the N-background correction. Note also, that the 3'/5'-ratios of the GADPH-probe sets of sample B exceed that of the sub-hooks in the S-range (11×-versus-1.9×, see Figure ##FIG##9##10##). This difference simply reflects the longer transcript regions interrogated by the entire set of GADPH-probes compared with the mean transcript-length probed by the subhooks. Analysis of the alternative β-actin control set provides analogous results (data not shown).</p>", "<p>In summary, the 3'/5'-ratio of the respective control probe sets are obviously insufficient for judging the amplification bias because non-specific hybridization keeps the signal of the 5' probe set at the same level as that of the 3' probe set which misleadingly pretends good amplification quality. Consideration of the hook-coordinates of these probes and, more reliably, analysis of 3'-biased \"sub-hooks\" enables the separation of the N- and S-hybridization ranges and this way a clear identification of the 3'/5'-amplification bias.</p>", "<title>Tissue specific RNA quality and normalization of microarray data</title>", "<p>Measurement of gene expression is based on the assumption that an analyzed RNA sample closely represents the amount of transcripts <italic>in vivo</italic>. Transcripts show stability differences of up to several orders of magnitude raising the possibility that partial degradation during cell lysis and sample preparation causes a transcript-specific bias in the expression measures in addition to the amplification bias discussed in the previous section [##UREF##17##37##]. Different RNA quality measures, such as the 28S/18S ratio, the RNA integrity number (RIN) or a degradometer-score have been developed, verified (see [##REF##16469371##43##] and references cited therein for an overview) and related to different microarray hybridization characteristics [##REF##16945445##36##,##UREF##18##39##,##REF##14647279##42##]. It was shown that the decrease in RNA integrity is often paralleled by the decrease of the percentage of present calls [##UREF##17##37##,##UREF##18##39##] which implies the reduction of the expression degree for degraded transcripts. Other studies however reveal more puzzling results, either with virtually no effects of degradation on expression or with opposite correlations between RNA-quality and weak and strong signals where the former ones increase and the latter signals decrease the worse the RNA becomes [##REF##18298816##38##].</p>", "<p>The integrity of the RNA extracted from different tissues systematically depends, among other factors, on the type of the tissue possibly and partly because of variations of the content and the activity of ribonucleases [##UREF##17##37##,##UREF##18##39##]. Estimation of RNA-quality and, if possible, appropriate correction for tissue-specific biases are thus essential steps in establishing tissue-specific expression profiles.</p>", "<p>In Figure ##FIG##11##12## we compare the hybridization characteristics of different tissues. The raw array data are taken from the comparative expression study on 79 human tissues [##REF##15075390##44##]. All hybridizations use the same start-amount of 5 μg of total RNA and the same amplification, hybridization and labelling protocols. Part a of Figure ##FIG##11##12## shows the distributions of the amount of absent calls obtained using MAS5 and hook-method for all considered tissues. The possible percentage of absent probe sets widely varies from values greater than 95% (virtually no present genes) to ~40% (hook) and ~10% (MAS5). Except their different spread, both distributions show essentially the same structure which reflects strong correlations between the MAS5 and hook calls in agreement with our previous findings (see above).</p>", "<p>For more detailed analysis we select two samples with relatively large and small percentages of absent probe sets, the RNA of which were extracted from superior cervical ganglion cells (scg) and from periphal blood/dentritic cells (dc) (see arrows in Figure ##FIG##11##12a##), respectively. Comparison of the respective intensity distributions indicates, except the slightly divergent width, no striking differences (see Figure ##FIG##11##12b##). In contrast, the respective hook-plots and underlying signal-distributions shown in part c and d of Figure ##FIG##11##12## reveal completely different hybridization characteristics: Most of the probe sets of the scg-hybridization accumulate within a relatively narrow Σ-range corresponding mainly to the N- and partly to the mix-hybridization regimes whereas the probes sets of the dc-sample cover a much wider range which includes the S- and sat-hybridization regimes as well.</p>", "<p>The different shapes of the hook curves cannot be explained by a smaller amount of RNA (e.g. due to a smaller yield of cRNA synthesis), less-efficient labelling and/or suboptimal calibration of the scanner. In these cases one expects the shift of the \"whole\" hooks without considerable change of their width and decay of the density distribution (compare, e.g. with Figure ##FIG##4##5##, upper part). Instead, the hook of the scg-sample is distinctly reduced in width reflecting the much higher level of non-specific background hybridization paralleled by the reduction of the decay constant.</p>", "<p>The hook-coordinates of selected probe sets are highlighted by symbols in Figure ##FIG##11##12## to illustrate this result: The symbols refer to probe sets selected to cover essentially the N-, mix-and S-hybridization regimes of the dc-hook. In the scg-hook most of these sets shift towards, and partly behind the detection limit given by the break-criterion. The solid symbols refer to amplification (GADPH-3' and -5') and hybridization (BioB-3) controls. The horizontal shift between amplification controls (see the solid triangles, the left one refers to the 5'- and the right one to the 3'-probe set) suggests a slightly smaller amplification bias of the dc-sample. The transcripts for hybridization controls (the solid circle refers to BioB_3) were added to the RNA-extracts in constant amounts before the inverse transcription step to assess its performance. The position of the respective Σ- and Δ-coordinates along the hook-curve remains relatively invariant indicating that the inverse transcription step has been performed in both samples in comparable quality. The drastic differences in the call rates must be therefore attributed to tissue-specific differences of the RNA-quality.</p>", "<p>In parts e and f of Figure ##FIG##11##12## we show the Σ-coordinates and the expression measures of the selected probe sets in both samples. Part g provides the differences (dc – scg) between them. The log-intensity measures (Σ) and the PMonly hook-expression values clearly reveal the larger signal and expression level of the dc-sample. Importantly, the PMonly-expression estimate of the BioB-hybridization control remains invariant between the samples. This result correctly reflects the equal amounts of BioB-transcripts spiked into both samples. The difference of the Σ-coordinates is however negative for BioB-3. This result and the fact that the differential expression of the PMonly estimates exceeds that of the Σ-data can be attributed to the non-specific background contributing to the latter data. The larger N-background in the scg-sample effectively increases the respective signal. Moreover, the data clearly show that the positive difference of the log-binding strength of specific hybridization of most of the transcripts is counterbalanced by the negative change of the binding strength of non-specific binding (see the horizontal dashed line in part e) – g) of Figure ##FIG##11##12##).</p>", "<p>These trends partly explain the puzzling results of a recent correlation analysis between signal intensities and the degree of RNA-degradation [##REF##18298816##38##]: Our data show, that, on one hand, degradation of RNA increases the non-specific background level with the consequence that the intensities of probes with small specific signal contributions effectively increase. On the other hand, the specific binding strength decreases upon RNA-degradation with the consequence that the signals of strongly expressed signals decrease. The former effect mainly affects weak intensities whereas the latter effect is more relevant for stronger total signals. Both opposite effects contribute to the intensity of each probe with specific weights giving rise to increased, decreased or even unchanged total signals.</p>", "<p>In part e – g of Figure ##FIG##11##12## we also show MAS5 expression estimates taken from ref. [##REF##15075390##44##]. The MAS5-expression measures of the dc-sample agree to a good approximation with that of the hook-method (see Figure ##FIG##11##12##, part e). For the scg-sample MAS5 however provides a considerably larger mean expression level. As a result, the expression differences are either much smaller in magnitude, or more critically, even change sign compared with the hook-results (see part g of Figure ##FIG##11##12##: dc-scg). For example, the hybridization control BioB becomes apparently much less expressed in the dc-sample in contrast to the hook-method which detects essentially no change, as expected.</p>", "<p>These qualitative discrepancies between both approaches uncover a fundamental problem of microarray normalization with no satisfactory solution yet (see, e.g., [##REF##16953902##45##]). Note that in their analysis the authors used MAS5 together with global median normalization of the raw intensities [##REF##15075390##44##]. The vertical bars in part b of Figure ##FIG##11##12## indicate the median of the log-intensity distributions. For the two considered samples the change of the non-specific contribution clearly dominates the observed change of the median chip intensity resulting in a stronger median signal of the scg-sample. The relative effect of, e.g. BioB with respect to the median is larger for the scg-sample (see the open circles in the figure) which gives rise to the negative differential expression reported by MAS5. This result exemplifies the problem with normalization methods which rescale the individual chip intensities to global chip characteristics such as their median or average value or use an averaged distribution as by quantile normalization. For the particular example discussed here such methods mask the larger specific signal in the dc-sample. Contrarily, the hook method disentangles the specific and non-specific signal-contributions with the option to scale them separately in subsequent normalization steps.</p>", "<title>Labelling protocol</title>", "<p>In addition to the quality of start-RNA and the amplification bias there are other methodological differences such as the labelling reaction that can introduce systematic biases. Figure ##FIG##12##13## compares the hook characteristics of two replicated samples of the same amount of starting RNA (5 μg) which are labelled using two different in-vitro-transcription (IVT) labelling kits: the Enzo BioArray high-yield RNA transcript labeling kit (Enzo) and the GeneChip expression 3'-amplification kit for IVT labeling (Affy) [##UREF##5##8##,##UREF##20##46##]. Both methods essentially follow the same experimental steps. Major distinction exists in the use of Biotin-UTP and -CTP in the former and Biotin-UTP only in the latter method. Fluorescent labels thus attach either to C- and U-nucleotides as well or to U-nucleotides only.</p>", "<p>The sensitivity profiles of the N-hybridization range are very similar for both labelling protocols with differences of less than 20% of the respective sensitivity value. Similar results were reported previously by using either Biotin-UTP or Biotin-CTP [##REF##17553856##47##]. The sensitivity terms additively decompose into \"binding-\"contributions related to the effective free energy of the respective base pairing; and into a fluorescence contribution taking into account base-specific labelling [##REF##16171364##20##]. Labelling is expected to decrease the binding contribution (because the bulky label disturbs the base-base interactions) and to increase the fluorescence contribution [##UREF##11##19##,##REF##16171364##20##]. The obtained positional dependent sensitivity profiles reveal that, if at all, labelling has only little effect.</p>", "<p>On the other hand, the width of the hook curve and the decay constant of the density distribution for the Affy-protocol slightly exceed the respective values for the Enzo-labelling at identical percentages of absent probes (~33%) and at identical optical background levels in both preparations. The observed differences indicate the slightly smaller amount of non-specific binding and the stronger specific binding of the former preparation. Hence, the Affy-protocol slightly better performs then the previous Enzo-labeling because it reduces the non-specific background level and increases the effective binding strength for specific binding; this way, giving rise to both, a better specificity and sensitivity of the method [##UREF##12##26##] in agreement with the results of special benchmark experiments [##UREF##20##46##].</p>", "<p>The molecular origin of the observed differences is presently not clear and requires further analyses. Note however that the Enzo-protocol introduces a significantly higher fraction of biotinylated nucleotides with potentially deteriorated binding affinities which provides a tentative explanation of the observed trends. The stronger specific binding caused by the Affy-protocol is paralleled by stronger saturation effects at high intensities which, in turn, give rise to systematic differences between the S-sensitivity profiles of both preparations: The profiles of cytosine (C) and guanine (G) shift systematically towards smaller sensitivities whereas the T- and especially the A-profiles shift into the opposite direction. This vertical \"compression\" of the profiles was previously observed [##REF##16171364##20##]. It reflects the fact that stronger base Watson-Crick pairings of the C- and G-nucleotides are, on the average over all probes, more affected by saturation than pairing of the T and especially A which form weaker bonds. Note also that the saturation effect is much smaller for the MM as expected. These results reveal that the hook-algorithm only incompletely corrects the individual probe intensities for saturation effects probably because the intensity asymptote upon complete saturation is not a chip constant but a sequence- and thus probe-specific property owing to washing effects [##UREF##14##29##,##REF##16723429##48##].</p>", "<title>Replacing RNA targets with DNA</title>", "<p>Microarray technology takes advantage of either of two types of chemical entities as the labelled target, RNA or DNA, considered to be virtually equivalent for the purpose of expression analysis. RNA is usually hybridized on \"conventional\" expression arrays whereas especially newer GeneChip generations such as exon- and tiling-expression arrays as well as genomic SNP- and re-sequencing-arrays use DNA-targets. Figure ##FIG##13##14## compares selected hook characteristics of both options to illustrate the effect of the two binding \"chemistries\" using the same start RNA-extract prepared from Jurkat-cells (chip data are taken from [##REF##16964210##49##]). cRNA was prepared by standard one round in-vitro transcription (see above) whereas cDNA was obtained by means of a different protocol (see [##REF##16964210##49##] and references cited therein). Besides the chemical entity of the targets both protocols differ with respect to preparation steps such as fragmentation (chemical versus enzymatic), labelling (\"during isothermal amplification\" versus \"after fragmentation\") and the position of the label (throughout the sequence versus end-labelled).</p>", "<p>Inspection of the hook-curves reveals several effects caused by the substitution of RNA by DNA: Firstly, the sensitivity correction to a much less extent affects the hook-curve of the DNA-hybridization (compare the corrected and raw hooks). For example, the width of the N-range of the raw RNA-hook (ΔΣ(N) ≈ 0.7) considerably exceeds that of the respective DNA-hook (ΔΣ(N) ≈ 0.3) whereas after correction the N-widths shrink to virtually identical values in both cases (ΔΣ(N) ≈ 0.2). Secondly, DNA/DNA hybridisation shifts the whole hook, and especially the background level, to smaller abscissa values indicating a smaller mean intensity level; thirdly, substitution of RNA by DNA slightly increases the width of the hook (β) and the decay constant of the density distribution in the S-range (λ); and fourthly, it slightly reduces the vertical dimension of the hook (α). Moreover, also the sensitivity profiles indicate characteristic differences: Especially, the profiles for Guanine (G) provide the largest contributions for DNA-binding to DNA-probes whereas the Cytosine-profiles are the largest in most cases for RNA-binding.</p>", "<p>The different target-entities give rise to D(NA)/R(NA)- and D/D-base pairings in the target/probe-duplexes and to R/R- and D/D-interactions for bulk duplexing of the targets in solution. The thermodynamic stability of specific 27 meric oligomer-duplexes was found to follow the order D/D &lt; D/R &lt; R/R with free energy ratios (37°C) of ΔG(D/D)/ΔG(D/R) ≈ 0.9 and ΔG(R/R)/ΔG(D/R) ≈ 1.3 [##REF##11011698##50##]. Note that the PM/MM-gain α ≈ log(s) approximately refers to the free energy increment of one Watson-Crick pairing in 25 meric probe/target duplexes if one neglects the specific mismatch contribution. The decreased PM/MM-gain (α) of the DNA-hybridization thus corresponds to the weaker association of D/D -versus – D/R where the ratio α (D/D)/α (D/R) ≈ 0.85 ± 0.05 roughly agrees with the expected free energy ratio.</p>", "<p>The slightly larger width of the DNA-hook indicates the smaller non-specific binding strength of the D/D-duplexes. This difference and the larger variability of the RNA-hybridization were attributed to relatively-stable, mismatched \"G•u-wobble\" base pairings in the non-specific R/D-duplexes (the lower case letter refers to the target, the upper case letter to the probe) which give rise to less specific binding and stronger scattering of the background compared with D/D hybridizations without such relatively-stable mismatched pairings [##REF##16964210##49##]. The latter D/D-hybridization is consequently more specific than the R/D-hybridization as indicated by the larger decay constant (see Figure ##FIG##13##14##) [##UREF##12##26##].</p>", "<p>Also the sensitivity profiles indicate systematic differences of base-pair interactions in both hybridizations. Particularly, the relative values of the G- and A-profiles for the D/D-duplexes are considerably larger than that for the D/R-duplexes. Exactly this trend is expected from the relative interaction strength of canonical Watson-Crick pairings in the respective duplexes: D/D-pairings are symmetrical with respect to \"bond-reversals\" (i.e. C•g≈ G•c &gt; A•t≈ T•a) in contrast to \"unsymmetrical\" D/R-interactions (C•g &gt; G•c≈ T•a &gt; A•u) [##UREF##11##19##,##REF##16171364##20##,##REF##11011698##50##, ####REF##7545436##51##, ##REF##12071944##52####12071944##52##]. Hence, for D/D-duplexes one expects the relative enhancement of the G and A sensitivity terms compared with those in the D/R duplexes in agreement with the observed profiles.</p>", "<p>Note however that the sensitivity profiles refer to effective binding strengths which include surface and bulk interactions as well [##UREF##12##26##,##UREF##21##53##]. Such effects give rise to specific differences between the S- and N-profiles especially of the RNA-preparation which implies relative strong R/R-interactions in the respective bulk duplexes.</p>", "<title>Competing interests</title>", "<p>The authors declare that they have no competing interests.</p>", "<title>Authors' contributions</title>", "<p>HB designed and leads the project, carried out most of the analyses and wrote the paper. SP wrote the computer program for hook analysis and helped to draft the paper. KK added experimental expertise and helped to draft the paper. All authors read and approved the final manuscript.</p>" ]

[ "<title>Acknowledgements</title>", "<p>The work was supported by the Deutsche Forschungsgemeinschaft under grant no. BIZ 6-1/4 and by grants from the Interdisciplinary Centre for Clinical Research at the Faculty of Medicine of the University of Leipzig (project Z03 to K.K.). SP thanks the International Max Planck Research School for Molecular Cell Biology and Bioengineering (IMPRS-MCBB) Dresden for funding.</p>" ]

[ "<fig position=\"float\" id=\"F1\"><label>Figure 1</label><caption><p>Hook-analysis of hybridizations on the human genome HG-U95 (left panel) and Drosophila genome DG-1 (right panel) GeneChips taken from the Genelogic dilution [##UREF##0##1##] and the GoldenSpike [##REF##15693945##2##] experimental series: The upper panel shows the raw and the sensitivity-corrected hook curves, the fitted theoretical curve and the distribution of the Σ-signal values (right axis, only left panel). Each hybridization is characterized by the parameters given in the figure (see also Table 2). These chip-characteristics are obtained from the fit. They are related to the geometrical dimensions of the corrected hook curve (see text). The lower part in each panel shows the four sensitivity profiles: PM-N and MM-N (left) and PM-S and MM-S (right).</p></caption></fig>", "<fig position=\"float\" id=\"F2\"><label>Figure 2</label><caption><p>Hybridization ranges of the raw (lower part) and the corrected (upper part) hook-curves calculated from hybridizations of the HG-U95 (left) and DG-1 (right) Gene Chips (see also Figure 1). The dotted lines indicate the hybridization ranges characterized by predominantly non-specific (N) and specific (S) binding, by a mixture of significant S- and N-contributions (mix), by the progressive saturation of the probe spots with bound transcripts (sat) and by almost completely saturated probes (as). Affinity correction considerably changes the shape of the hook-curve and the extent of the hybridization ranges. The corrected hook-curve and the fit are characterized by their geometrical dimensions; width (β), height (~α), start- (Σ(0), Δ(0)) and end- (Σ(∞)) positions; which in turn characterize the particular hybridization in terms of the mean non-specific background contribution, the PM/MM-gain etc. (see Table 2 for details). Compare also with Figure 1: The HG-U95 data were taken from different experiment series (Affymetrix spiked-in series here [##UREF##1##3##] and Genelogic dilution series [##UREF##0##1##] in Figure 1).</p></caption></fig>", "<fig position=\"float\" id=\"F3\"><label>Figure 3</label><caption><p>Hook-characteristics of GeneChips of different generations (see figure, from left to the right). The chips are hybridized with mRNA extracts from tumour samples (thyroid nodules, two parts on the left; [##REF##16407496##9##] and references cited therein) and from the Universal Human Reference RNA (chips c and d; see [##REF##16776839##10##] for details). The figures show the raw hook (below), the corrected hook (middle), the probability density distribution (middle, right axis) and the theoretical curve fitted of the mix-, S- and sat-ranges of the corrected hook curves (above). The percentage of absent probes (%N) is given within the figures.</p></caption></fig>", "<fig position=\"float\" id=\"F4\"><label>Figure 4</label><caption><p>Present/absent characteristics of two hybridizations. <bold>Left part</bold>: The Yeast Genome 2.0 (YG 2.0) array contains about 50% probe sets designed for S. cerevisiae and S. pombe each. The hook refers to a chip hybridized with RNA taken from S. cerevisiae [##REF##17043222##13##]. The hooks are calculated either for all probes or masking the probes of one of the two yeast species. The lower part shows the respective signal-density distributions. The added transcripts of S. cerevisiae give rise to virtually absent probes of S. pombe in the N-range of the hook curve. The relative amount of S. cerevisiae-probes called absent (red) and of S. pombe-probes called present (blue) are given within the figure. <bold>Right part</bold>: Hook curves for a DG1-chip taken from the Golden Spike series which has been hybridized with a definite collection of \"spiked\"-transcripts. The selective masking of the spikes and of the remaining \"empty\" probes shows that these probes accumulate in the S- and N-region, respectively. The relative amounts of empty probes called present and of spiked probes called absent are given in the figure.</p></caption></fig>", "<fig position=\"float\" id=\"F5\"><label>Figure 5</label><caption><p>Genelogic dilution experiment: Hook curves for different dilution steps (upper panel), the fraction of absent probes (middle panel) and concentration measures (S/N-ratio and specific binding strength, lower panel) as a function of the amount of added RNA. The dilution of the hybridization solution shifts the increasing part of the hooks to the left and increases its width. The width is inversely related to the non-specific binding strength, ~-log X^N, which consequently decreases upon dilution. The horizontal dotted lines in the upper part indicate the levels of different S/N-ratio (R); the dashed parabola-like curves are fits of the Langmuir-hybridization model. The hook method provides a virtually constant fraction of absent probes which corresponds to the essentially invariant S/N-ratio of the probes upon changing dilution. Contrarily, MAS5 provides an increasing fraction of absent probes (see middle panel). The lower part compares the S/N-ratio of selected probes which remain virtually constant upon dilution with the binding strength which progressively decreases (compare lines and solid symbols in the lower part; the diagonal lines refer to the right coordinate axis).</p></caption></fig>", "<fig position=\"float\" id=\"F6\"><label>Figure 6</label><caption><p>Affymetrix spiked-in experiment: The upper panel shows the hook obtained from one chip of this series. The predominant number of probes is hybridized with RNA of a HeLa-cell extract which was added to the chips to mimic a complex hybridization background (thick blue curve). The spike-probe sets are indicated by the open symbols and the respective transcript concentrations (see the numbers, the concentrations are given in units of pM). The horizontal distance between a spike position and the end point is related to the logarithm of the specific binding strength. The turning point between the N- and the mix-ranges defines the threshold for present probes. The dashed line is the fit of the Langmuir hybridization model to the data. The middle and lower parts show present/absent characteristics and the S/N-ratio of the spikes, respectively. The fraction of absent probes and the S/N ratio were calculated as mean values over all 42 chips of the experimental series (see thick lines). The open circles in the lower part show the individual probe-set values and thus the scatter of these points about their mean value. Spiked probes with nominal concentrations larger than 2 pM are \"safely\" called present. The S/N-ratio linearly correlates with the spiked-in concentration. The right axis of the lower part scales the expression estimates in units of the binding strength. The green dashed lines indicated that the threshold for calling probes as present corresponds to S/N-ratios R ≈ 0.1 – 2 and the S-binding strength of X^N≈ (0.5 – 5) 10^-3.</p></caption></fig>", "<fig position=\"float\" id=\"F7\"><label>Figure 7</label><caption><p>Expression estimates (upper panel, see figure for assignment), their coefficient of variation and the ratio of the estimated and the experimental (\"true\") spiked concentration (lower panel) as a function of the spiked concentration. The latter two measures estimate the precision and the accuracy of the expression values, respectively. The expression estimates in the upper panel are scaled to agree with the diagonal (dashed) line which refers to perfect results. The perfect precision and accuracy refer to zero (no scattering, middle part) and unity (lower part), respectively. All values are averaged over all probe sets detecting spiked transcripts. The figure compares the performance of the hook expression estimates (PMonly, MMonly, PM-MM and R) with that of RMA (see text).</p></caption></fig>", "<fig position=\"float\" id=\"F8\"><label>Figure 8</label><caption><p>Cross-chip comparison of the expression estimates of four selected probe sets taken from the HG-U133A and HG-U133plus2 arrays (chip data were taken from [##REF##16776839##10##]). Both chip types were hybridized with human reference RNA in five replicates (solid symbols). The open symbols are the log-means over the replicates. Expression measures taken from ref. [##REF##16776839##10##] were compared with the four alternative measures provided by the hook-method. Note the systematic shift of the expression values between both different chip-types which changes sign upon increasing expression value. The chip-type specific bias considerably reduces for the hook-measures. The MMonly-method performes worst among the hook-methods. (see also Figure 7). The Zhang-measures are given in arbitrary units which were scaled for comparison with the hook data.</p></caption></fig>", "<fig position=\"float\" id=\"F9\"><label>Figure 9</label><caption><p>Signal distributions (below) and corrected hook curves (above) of Universal Human Reference RNA hybridized on HG133A and HG133plus2 GeneChips (raw data were taken from [##REF##16776839##10##]). The probe sets are composed either according to Affymetrics default settings (left part and \"Affy\" in the other parts), or using different customized transcript definitions (see ref. [##REF##16284200##30##]; version 10; middle and right part) based on the annotations of different resources: ENSEMBLE (ENSG, ENSE, ENST), REFSEQ (see text). The probe sets of the HG133plus2 array were split into two subgroups which are either represented on both chip-types (\"A\") or on the P-chip only (\"notA\"). See the legends within the figure. The respective number of probe sets per array is given within the parentheses. The dotted lines in the lower panel serve as guide for the eye to characterize the respective decay constants λ. The hooks in the left part and the \"Affy\" hooks in the other parts are shifted in vertical direction each to another for sake of clarity. The dotted curves in the upper panel are fits of the hook-equation. The essential parameters are given in Table 3.</p></caption></fig>", "<fig position=\"float\" id=\"F10\"><label>Figure 10</label><caption><p>3'/5'-bias of two replicated hybridizations A (left part) and B (right part) of RNA of different quality on the rat genome array RAE-230: The graph above, in the middle and below show the total log-averaged mean of the PM- and MM-intensities, Σ, taken over all 11 probe-pairs of each set, the hook curves and the signal distributions, respectively, as a function of the sub-mean, Σ_sub, averaged over subsets of the first four probes of a probe set (probes no. 1–4) closer to the 5'-end of the transcripts and the last four probes (no. 8–11) nearer to the 3'-end of the transcripts. The 5'- and 3'-biased sub-means virtually agree in the N-hybridization range whereas upon specific hybridization the 3'-biased sub-mean exceeds that of the 5'-biased one owing to 3'-biased amplification of RNA (upper panel, see arrows, the factors indicate the fold changes of the 3'-end relative to the 5'-end). The dimensions of the different hooks calculated using either all probes (1–11) or the biased subsets roughly agree each with another showing that all probes follow virtually the same hybridization law (middle panel). The higher yield for RNA fragments near the 3'-end of the transcripts gives rise to larger decay constants λ if one plots the signal-density as a function of the Σ_sub-coordinate of the respective subsets of probes (see lower part, φ is the respective mean negative logarithmic expression index). The 3'/5'-bias is larger for sample A shown in the left column of the figure. Note that the width of the respective hooks and the fraction of absent probes (42% versus 54%, see figure) increase upon decreasing RNA-quality (compare A with B). The open circles in the middle panel indicate the positions of the GADPH-probe sets used typically for 3'/5'-hybridization control. The left one in each hook refers to the 5'-biased set and the right one to the 3'-biased set.</p></caption></fig>", "<fig position=\"float\" id=\"F11\"><label>Figure 11</label><caption><p>Characterization of the 3'/5'-amplification bias for the two samples shown in Figure 10 using the GADPH-probe sets Affx_rat_GADPH_x_at with x = 3', m and 5'. These three sets probe the GADPH-transcript with increasing distance from the 3'-end. The bars show the log-averages of the PM and MM intensities after correction for the optical background over the respective probe sets (Σ) and the MAS5 and hook (PMonly) expression estimates. The horizontal line indicates the hook-coordinate of the break, Σ^breakwith Σ ≥ Σ^breakcalled present (P) and Σ &lt; Σ^breakcalled absent (A, see also the middle panel of Figure 10). Note that the Σ-signal of GADPH in sample A is dominated by non-specific hybridization at least for the m- and 5'-probes whereas it contains a much larger specific contribution in sample B. The fold-changes of the 3'/5'-signals are given above the 3'-bars. The circles indicate the respective Δ-coordinate of the hook-curve referring to the right axis.</p></caption></fig>", "<fig position=\"float\" id=\"F12\"><label>Figure 12</label><caption><p>Tissue-specific RNA profiling. Part a) Frequency distribution of absent calls of tissue-specific total RNA hybridized on HG-U133A arrays taken from 79 tissues and analyzed with MAS5 and hook (raw array-intensities and MAS5 data were taken from ref. [##REF##15075390##44##]). Part b) – g) Comparison of two hybridizations with small and large absent rates (see arrows in part a): peripheral blood-BDCA4 dentritic cells (dc, GEO-query GSM18873) and superior cervical ganglion (scg, GEO-query GSM19012). Both hybridizations used the same amount of total RNA (5 μg) for synthesis of biotinylated cRNA and the same labelling protocol. Part b) compares the log-intensity-distributions of the PM-probes: Except the shift and widening of the distribution of the dc-sample, one observes essentially no peculiar differences between the specimens. The median and the probe-set related values of BioB-3 are explicitly shown and discussed in the text. Parts c) and d) show the respective hook-plots together with the signal-density distributions. Note the striking differences: The scg-sample hybridizes much weaker with a markedly larger fraction of probe set with absent calls (95%) and a much steeper decay of the distribution in the mix-range of the hook (λ is the decay constant). The dotted curves are fits of the Langmuir model. The open symbols indicate the hook-coordinates of selected probe set in both preparations to illustrate the apparent expression changes from different regions of the hook (#1 to #4). The solid symbols refer to amplification and hybridization control probe sets. In part e) and f) the Σ-coordinates and the expression measures of these selected probes are explicitly shown, part g shows the respective log-differences between both samples. Note that the difference of the non-specific background level is negative (dashed horizontal line), whereas the difference of the specific binding strengths of most of the considered probe sets is positive (PMonly measures). The specific expression of the BioB-control is virtually invariant in both samples, as expected. Contrarily, MAS5 pretends significant expression changes of the BioB-control due to improper normalization (see text).</p></caption></fig>", "<fig position=\"float\" id=\"F13\"><label>Figure 13</label><caption><p>Hook analysis of two replicated hybridization on RAE-230 rat genome arrays which are labelled using either the Affymetrix- (left) or ENZO- (right) protocols. The Affy-protocols labels the cytosines only whereas the ENZO-protocol labels cytosines and uracyls as well.</p></caption></fig>", "<fig position=\"float\" id=\"F14\"><label>Figure 14</label><caption><p>Hook-characteristics of cRNA (left) and cDNA (right) hybridizations prepared from of a Jurkat-cell RNA-extract on HG-U133Av2 chips (raw data are taken from ref. [##REF##16964210##49##]).</p></caption></fig>" ]

[ "<table-wrap position=\"float\" id=\"T1\"><label>Table 1</label><caption><p>Geometrical parameters of the hook curve</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td align=\"left\"><bold>Hook parameter</bold></td><td align=\"left\"><bold>symbol</bold></td><td align=\"left\"><bold>typical range</bold></td><td align=\"left\"><bold>characteristics</bold></td></tr></thead><tbody><tr><td align=\"left\"><bold>Start point</bold></td><td align=\"left\">Σ(0) ≈ Σ^{<italic>start</italic>},</td><td align=\"left\">1.0 – 2.5</td><td align=\"left\">Non-specific signal</td></tr><tr><td/><td align=\"left\">Δ(0) ≈ Δ^{<italic>start</italic>}</td><td align=\"left\">0.0 – 0.15</td><td align=\"left\">PM/MM-gain (N)</td></tr><tr><td align=\"left\"><bold>End point</bold></td><td align=\"left\">Σ(∞),</td><td align=\"left\">3.5 – 4.8</td><td align=\"left\">Saturation signal</td></tr><tr><td/><td align=\"left\">Δ(∞)</td><td align=\"left\">0</td><td align=\"left\">PM/MM-gain (as)</td></tr><tr><td align=\"left\"><bold>Width</bold></td><td align=\"left\"><italic>β </italic>= Σ(∞)-Σ(0)</td><td align=\"left\">2.2 – 3.2</td><td align=\"left\">Measuring range, non- specific binding strength in logarithmic scale</td></tr><tr><td align=\"left\"><bold>asymptotic height</bold></td><td align=\"left\"><italic>α</italic></td><td align=\"left\">0.75 – 1.1</td><td align=\"left\">PM/MM-gain (S)</td></tr><tr><td align=\"left\"><bold>decay constant</bold></td><td align=\"left\"><italic>λ</italic></td><td align=\"left\">0.5 – 1.5</td><td align=\"left\">Decay rate of the density distribution of the Σ-values; this S/N-index characterizes the mean ratio of specific and non- specific binding (S/N- ratio) in the logarithmic scale.</td></tr><tr><td align=\"left\"><bold>Expression index</bold></td><td align=\"left\"><italic>φ </italic>= (<italic>β </italic>- Δ(0)) - <italic>λ </italic>≈ <italic>β </italic>- <italic>λ</italic></td><td align=\"left\">1.5 – 2.5</td><td align=\"left\">Mean specific signal in logarithmic scale</td></tr></tbody></table></table-wrap>", "<table-wrap position=\"float\" id=\"T2\"><label>Table 2</label><caption><p>Overview of the hybridization characteristics extracted from the hook-analysis.</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td align=\"left\"><bold>Characteristics</bold></td><td align=\"center\"><bold>Equation^b)</bold></td><td align=\"left\"><bold>characterizes...^b)</bold></td><td align=\"center\"><bold>Typical range^c)</bold></td></tr></thead><tbody><tr><td align=\"left\" colspan=\"4\"><bold><italic>Chip-level </italic></bold>(index \"c\" is omitted)</td></tr><tr><td align=\"left\">Optical background, O^a)</td><td align=\"center\">log <italic>O </italic>= ⟨log <italic>O</italic>⟩_{<italic>zones</italic>}</td><td align=\"left\">...residual background intensity not related to hybridization; it is obtained using the Affy-zone algorithm performed prior to hook analysis</td><td align=\"center\">1.4 – 2.0</td></tr><tr><td align=\"left\">N-background signal^a)</td><td align=\"center\">log <italic>N </italic>= Σ(0) + Δ(0)</td><td align=\"left\">... mean background PM-intensity due to N-hybridization</td><td align=\"center\">1.0 – 2.5</td></tr><tr><td align=\"left\">PM/MM-gain in the N- range</td><td align=\"center\">log <italic>n </italic>= Δ(0)</td><td align=\"left\">...the PM-over-MM excess of the intensity presumably due to a certain amount of weakly (S-) expressed transcripts in the N-range</td><td align=\"center\">0.0 – 0.15</td></tr><tr><td align=\"left\">Saturation signal^a)</td><td align=\"center\">log <italic>M </italic>= Σ(∞)</td><td align=\"left\">... the maximum possible intensity of the spots</td><td align=\"center\">4.0 – 4.9</td></tr><tr><td align=\"left\">N-binding strength^a)</td><td align=\"center\">log <italic>X</italic>^{<italic>N </italic>}≡ log <italic>X</italic>^{<italic>PM</italic>, <italic>N </italic>}= -<italic>β </italic>+ Δ(0)</td><td align=\"left\">... the (binding) strength of non- specific hybridization; measuring range of the chip</td><td align=\"center\">2.2 – 3.2</td></tr><tr><td align=\"left\">PM/MM-gain (S, the PM- over-MM excess of the intensity in the S-range)</td><td align=\"center\">log <italic>s </italic>= <italic>α </italic>- Δ(0)</td><td align=\"left\">...the effect of the mismatch on specific binding</td><td align=\"center\">0.8 – 1.1</td></tr><tr><td align=\"left\">Mean S/N-ratio^a)</td><td align=\"center\">⟨<italic>λ</italic>⟩ = ⟨log(<italic>R </italic>+ 1)⟩_{<italic>R </italic>&gt; 0.5}</td><td align=\"left\">...mean (log-) S/N-ratio; R-range over which the density of expression values decays by one order of magnitude</td><td align=\"center\">0.2 – 1.5</td></tr><tr><td align=\"left\">Mean expression level^a)</td><td align=\"center\">⟨<italic>φ</italic>⟩ = ⟨<italic>λ</italic>⟩ + log <italic>X</italic>^{<italic>N</italic>}<break/>⟨<italic>S</italic>⟩ = 10^{-⟨<italic>φ</italic>⟩}</td><td align=\"left\">...mean (log-) expression index in units of the specific binding strength</td><td align=\"center\">1.0 – 2.5</td></tr><tr><td align=\"left\">Standard deviation of the N- distribution^a)</td><td align=\"center\"><italic>σ</italic></td><td align=\"left\">...residual scatter of the corrected PM-intensities in the N-range (log- scale)</td><td align=\"center\">0.25 – 0.35</td></tr><tr><td align=\"left\">Percent non-specific, %N; fraction of N-probes</td><td align=\"center\">%N, f^absent= %N/100</td><td align=\"left\">Percentage of probe sets in the N- range;...amount of \"absent\" probes</td><td align=\"center\">20 – 95%</td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"left\" colspan=\"4\"><bold><italic>Probe-set level </italic></bold>(index \"set\" is omitted)</td></tr><tr><td align=\"left\">Hook coordinates</td><td align=\"center\">Σ^hook, Δ^hook</td><td align=\"left\">...log-mean and log difference of the PM and MM intensities after optical background correction</td><td align=\"center\">1 – 4.7 and 0.0 – 1.1</td></tr><tr><td align=\"left\">S/N-ratio</td><td align=\"center\">R</td><td align=\"left\">...ratio of the specific binding strength of the probe set and the mean non-specific binding strength of the chip, signal-to-noise level</td><td align=\"center\">0 – 100, R = 0 indicates \"absent\" probes</td></tr><tr><td align=\"left\">expression level</td><td align=\"center\">L^S≡ L^{PM, S}</td><td align=\"left\">...expression degree in intensity units (PMonly, MMonly and PM-MM estimates)</td><td align=\"center\">10 – 100,000</td></tr><tr><td align=\"left\">S-binding strength</td><td align=\"center\">X^S≡ X^{PM, S}</td><td align=\"left\">...specific binding strength obtained as PMonly, MMonly or PM-MM_difference estimate</td><td align=\"center\">0 – 1</td></tr></tbody></table></table-wrap>", "<table-wrap position=\"float\" id=\"T3\"><label>Table 3</label><caption><p>Hook characteristics of HG-U133A and HG-U133plus2 chips hybridized with the same RNA using different probe set definitions^a)</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td align=\"left\"><bold>probe set definition</bold></td><td align=\"center\"><bold>chip-type</bold></td><td align=\"center\"><bold>Optical BG</bold></td><td align=\"center\"><bold>non- specific BG</bold></td><td align=\"center\"><bold>N-binding strength</bold></td><td align=\"center\"><bold>PM/MM- gain (S)</bold></td><td align=\"center\"><bold>PM/MM-gain (N)</bold></td><td align=\"center\"><bold>mean S/N-index</bold></td><td align=\"center\"><bold>mean expression index</bold></td><td align=\"center\"><bold>percent absent</bold></td><td align=\"center\"><bold>probe utilization</bold>^d)</td></tr><tr><td/><td/><td align=\"center\"><bold>logO</bold></td><td align=\"center\"><bold>logN</bold></td><td align=\"center\"><bold>β</bold></td><td align=\"center\"><bold>α</bold></td><td align=\"center\"><bold>logn</bold></td><td align=\"center\"><bold>&lt;λ&gt;</bold></td><td align=\"center\"><bold>&lt;φ&gt;</bold></td><td align=\"center\"><bold>%N</bold></td><td align=\"center\"><bold>%P</bold><break/><bold># of probe sets</bold></td></tr></thead><tbody><tr><td align=\"left\" colspan=\"11\"><bold>Affymetrix probe sets^b)</bold></td></tr><tr><td align=\"left\"><bold>total</bold></td><td align=\"center\">HG- U133A</td><td align=\"center\">1.89<break/>± 0.04</td><td align=\"center\">1.71<break/>± 0.04</td><td align=\"center\">2.70<break/>± 0.04</td><td align=\"center\">0.99<break/>± 0.03</td><td align=\"center\">0.10<break/>± 0.01</td><td align=\"center\">0.61<break/>± 0.03</td><td align=\"center\">2.09<break/>± 0.05</td><td align=\"center\">34<break/>± 3</td><td align=\"center\">100%<break/>22,193</td></tr><tr><td align=\"left\"><bold>total</bold></td><td align=\"center\">HG- U133plus2</td><td align=\"center\">1.81<break/>± 0.04</td><td align=\"center\">1.65<break/>± 0.08</td><td align=\"center\">2.75<break/>± 0.07</td><td align=\"center\">0.85<break/>± 0.01</td><td align=\"center\">0.07<break/>± 0.01</td><td align=\"center\">0.57<break/>± 0.03</td><td align=\"center\">2.18<break/>± 0.07</td><td align=\"center\">50<break/>± 3</td><td align=\"center\">100%<break/>54,585</td></tr><tr><td align=\"left\"><bold>A</bold></td><td align=\"center\">HG- U133plus2</td><td align=\"center\">1.82<break/>± 0.06</td><td align=\"center\">1.63<break/>± 0.08</td><td align=\"center\">2.76<break/>± 0.07</td><td align=\"center\">0.86<break/>± 0.01</td><td align=\"center\">0.09<break/>± 0.01</td><td align=\"center\">0.65<break/>± 0.05</td><td align=\"center\">2.11<break/>± 0.07</td><td align=\"center\">29<break/>± 3</td><td align=\"center\">41%<break/>22,187</td></tr><tr><td align=\"left\"><bold>notA</bold></td><td align=\"center\">HG- U133plus2</td><td align=\"center\">1.80<break/>± 0.06</td><td align=\"center\">1.67<break/>± 0.09</td><td align=\"center\">∞</td><td align=\"center\">0.83<break/>± 0.01</td><td align=\"center\">0.06<break/>± 0.01</td><td align=\"center\">0.45<break/>± 0.03</td><td align=\"center\">x</td><td align=\"center\">64<break/>± 5</td><td align=\"center\">59%<break/>32,308</td></tr><tr><td colspan=\"11\"><hr/></td></tr><tr><td align=\"left\" colspan=\"11\"><bold>Customized probe sets^c)</bold></td></tr><tr><td align=\"left\"><bold>Ensemble gene</bold></td><td align=\"center\">HG- U133A</td><td align=\"center\">1.89<break/>± 0.05</td><td align=\"center\">1.71<break/>± 0.03</td><td align=\"center\">2.77<break/>± 0.02</td><td align=\"center\">0.97<break/>± 0.03</td><td align=\"center\">0.12<break/>± 0.003</td><td align=\"center\">0.56<break/>± 0.03</td><td align=\"center\">2.21<break/>± 0.04</td><td align=\"center\">23<break/>± 3</td><td align=\"center\">68%<break/>11,834</td></tr><tr><td/><td align=\"center\">HG- U133plus2</td><td align=\"center\">1.82<break/>± 0.06</td><td align=\"center\">1.61<break/>± 0.09</td><td align=\"center\">2.74<break/>± 0.07</td><td align=\"center\">0.84<break/>± 0.01</td><td align=\"center\">0.08<break/>± 0.02</td><td align=\"center\">0.56<break/>± 0.05</td><td align=\"center\">2.18<break/>± 0.07</td><td align=\"center\">21<break/>± 3</td><td align=\"center\">48%<break/>17,215</td></tr><tr><td align=\"left\"><bold>Ensemble transcript</bold></td><td align=\"center\">HG- U133A</td><td align=\"center\">1.88<break/>± 0.05</td><td align=\"center\">1.71<break/>± 0.03</td><td align=\"center\">2.67<break/>± 0.05</td><td align=\"center\">0.95<break/>± 0.03</td><td align=\"center\">0.10<break/>± 0.01</td><td align=\"center\">0.57<break/>± 0.03</td><td align=\"center\">2.10<break/>± 0.05</td><td align=\"center\">19<break/>± 1</td><td align=\"center\">71%<break/>23,740</td></tr><tr><td/><td align=\"center\">HG- U133plus2</td><td align=\"center\">1.81<break/>± 0.06</td><td align=\"center\">1.65<break/>± 0.09</td><td align=\"center\">2.68<break/>± 0.07</td><td align=\"center\">0.79<break/>± 0.11</td><td align=\"center\">0.04<break/>± 0.09</td><td align=\"center\">0.57<break/>± 0.04</td><td align=\"center\">2.08<break/>± 0.07</td><td align=\"center\">18<break/>± 11</td><td align=\"center\">48%<break/>33,977</td></tr><tr><td align=\"left\"><bold>Ensemble Exon</bold></td><td align=\"center\">HG- U133A</td><td align=\"center\">1.88<break/>± 0.05</td><td align=\"center\">1.69<break/>± 0.04</td><td align=\"center\">2.64<break/>± 0.06</td><td align=\"center\">-0.97<break/>± 0.02</td><td align=\"center\">0.11<break/>± 0.01</td><td align=\"center\">0.58<break/>± 0.03</td><td align=\"center\">2.06<break/>± 0.06</td><td align=\"center\">23<break/>± 3</td><td align=\"center\">63%<break/>22,299</td></tr><tr><td/><td align=\"center\">HG- U133plus2</td><td align=\"center\">1.81<break/>± 0.06</td><td align=\"center\">1.62<break/>± 0.08</td><td align=\"center\">2.72<break/>± 0.13</td><td align=\"center\">0.84<break/>± 0.01</td><td align=\"center\">0.09<break/>± 0.01</td><td align=\"center\">0.57<break/>± 0.06</td><td align=\"center\">2.15<break/>± 0.15</td><td align=\"center\">25<break/>± 5</td><td align=\"center\">43%<break/>34,541</td></tr><tr><td align=\"left\"><bold>Refseq</bold></td><td align=\"center\">HG- U133A</td><td align=\"center\">1.88<break/>± 0.05</td><td align=\"center\">1.68<break/>± 0.06</td><td align=\"center\">2.64<break/>± 0.07</td><td align=\"center\">0.96<break/>± 0.02</td><td align=\"center\">0.09<break/>± 0.02</td><td align=\"center\">0.60<break/>± 0.05</td><td align=\"center\">2.04<break/>± 0.07</td><td align=\"center\">17<break/>± 3</td><td align=\"center\">72%<break/>17,531</td></tr><tr><td/><td align=\"center\">HG- U133plus2</td><td align=\"center\">1.83<break/>± 0.06</td><td align=\"center\">1.67<break/>± 0.08</td><td align=\"center\">2.65<break/>± 0.12</td><td align=\"center\">0.83<break/>± 0.01</td><td align=\"center\">0.09<break/>± 0.01</td><td align=\"center\">0.61<break/>± 0.05</td><td align=\"center\">2.04<break/>± 0.12</td><td align=\"center\">27<break/>± 7</td><td align=\"center\">48%<break/>25,004</td></tr></tbody></table></table-wrap>" ]

[ "<inline-formula><mml:math xmlns:mml=\"http://www.w3.org/1998/Math/MathML\" id=\"M1\" name=\"1748-7188-3-11-i1\" overflow=\"scroll\"><mml:semantics><mml:mrow><mml:mfrac><mml:mn>1</mml:mn><mml:mn>2</mml:mn></mml:mfrac></mml:mrow></mml:semantics></mml:math></inline-formula>", "<inline-formula><mml:math xmlns:mml=\"http://www.w3.org/1998/Math/MathML\" id=\"M2\" name=\"1748-7188-3-11-i1\" overflow=\"scroll\"><mml:semantics><mml:mrow><mml:mfrac><mml:mn>1</mml:mn><mml:mn>2</mml:mn></mml:mfrac></mml:mrow></mml:semantics></mml:math></inline-formula>", "<inline-formula><mml:math xmlns:mml=\"http://www.w3.org/1998/Math/MathML\" id=\"M3\" name=\"1748-7188-3-11-i1\" overflow=\"scroll\"><mml:semantics><mml:mrow><mml:mfrac><mml:mn>1</mml:mn><mml:mn>2</mml:mn></mml:mfrac></mml:mrow></mml:semantics></mml:math></inline-formula>" ]

[]

[]

[]

[]

[]

[ "<table-wrap-foot><p>^a)characteristics refer to the PM-probes; for O, M and σ virtually equal values for PM and MM are obtained</p><p>^b)see the accompanying paper [##REF##18759985##4##] for details</p><p>^c)ranges of typical values are taken from the hook-analyses of more than 500 GeneChip arrays of different type and origin (see [##UREF##2##5##])</p></table-wrap-foot>", "<table-wrap-foot><p>^a)Raw intensity data were taken from ref. [##REF##16776839##10##]; human reference RNA has been hybridized onto both chip types in 5 replicates. The data are log-averages/±SE</p><p>^b)Probe set definition of the manufacturer; total...all probe sets; A/notA...probe sets shared/not shared between the P- and A-chips</p><p>^c)Customized probe sets were filtered using genomic information provided by Ensemble (gene, transcript or exon related) and Refsequ (see [##REF##16284200##30##]); probe set definitions were downloaded from <ext-link ext-link-type=\"uri\" xlink:href=\"http://brainarray.mbni.med.umich.edu\"/> (version 10) as CDF and probe-sequence files</p><p>^d)Percent and total number of the probes on the respective chip which are used in the respective analysis. Note that the number of probes per set varies between 4 and more than 30 for the customized sets. The data are taken from <ext-link ext-link-type=\"uri\" xlink:href=\"http://brainarray.mbni.med.umich.edu\"/></p></table-wrap-foot>" ]

[ "<graphic xlink:href=\"1748-7188-3-11-1\"/>", "<graphic xlink:href=\"1748-7188-3-11-2\"/>", "<graphic xlink:href=\"1748-7188-3-11-3\"/>", "<graphic xlink:href=\"1748-7188-3-11-4\"/>", "<graphic xlink:href=\"1748-7188-3-11-5\"/>", "<graphic xlink:href=\"1748-7188-3-11-6\"/>", "<graphic xlink:href=\"1748-7188-3-11-7\"/>", "<graphic xlink:href=\"1748-7188-3-11-8\"/>", "<graphic xlink:href=\"1748-7188-3-11-9\"/>", "<graphic xlink:href=\"1748-7188-3-11-10\"/>", "<graphic xlink:href=\"1748-7188-3-11-11\"/>", "<graphic xlink:href=\"1748-7188-3-11-12\"/>", "<graphic xlink:href=\"1748-7188-3-11-13\"/>", "<graphic xlink:href=\"1748-7188-3-11-14\"/>" ]

[]

{ "acronym": [], "definition": [] }

[ "<title>Background</title>", "<p>Classical swine fever (CSF) is a highly contagious viral disease of swine and wild boars, causing severe economic losses mainly in countries with dense pig populations. The causative agent is classical swine fever virus (CSFV), a small enveloped, positive-stranded RNA virus that belongs to the genus <italic>Pestivirus </italic>in the <italic>Flaviviridae </italic>family [##REF##16716047##1##,##REF##10785323##2##]. The genus also comprises bovine viral diarrhoea virus (BVDV) and border disease virus (BDV) of sheep.</p>", "<p>Although CSF has been known for more than 150 years, the losses to this disease are still extremely high. For example the 1997–98 outbreaks of CSF caused very heavy losses in the Netherlands, when approximately 12 million pigs were lost due to the disease (about 700,000 heads), culling and welfare reasons [##REF##10619154##3##].</p>", "<p>A number of observations show that antigenic variations exist among CSFV strains and the various field isolates can vary considerably in virulence. Highly virulent viruses cause peracute or acute forms of the disease with high morbidity and mortality in pigs, irrespective of age and breed. In contrast, viruses of moderate to low virulence may cause a very mild or inapparent disease. In the last three-four decades, the most common clinical picture of CSF has changed from acute to subacute, chronic or inapparent forms [##REF##3992564##4##,##REF##8903020##5##]. These changes in the clinical manifestation of the disease frequently complicate the early detection and proper diagnosis of the CSF, considering that the very mild clinical symptoms might easily be overlooked. The delayed diagnosis may cause uncontrolled spread of CSF and heavy losses in large swine populations. Considering this situation, there is a high need to perform comparative studies on the tissue distribution of various variants of the virus in order to study virus biology and to assure the diagnosis.</p>", "<p>The diagnosis can be complicated by the uncharacteristic profiles of CSF clinical symptoms, which may lead to delayed identification of new outbreaks (see World Organisation for Animal Health, OIE, <ext-link ext-link-type=\"uri\" xlink:href=\"http://www.oie.int\"/>). A further diagnostic problem is that rather poor information is available concerning the pathogenicity and invasion capacity of various virulence variants of CSFV. The early studies on viral pathogenicity and invasion were restricted to single strains and comparative aspects were not discussed. For example, Ressang [##UREF##0##6##,##UREF##1##7##] described the quantitative distribution of the virulent Brescia strain in various tissues in increasing intervals. Subsequently, the studies were extended to involve comparative analysis of more than one strain. Such work was performed by Kamolsiriprichaiporn et al. [##REF##1445070##8##] who compared the pathogenicity of the virulent Weybridge and of the low virulent New South Wales strains. Japanese researchers performed comparative immunohistochemical studies on organ specimens of pigs infected with the highly virulent ALD strain or with the less virulent Kanagawa 74 strain, respectively [##REF##11055862##9##].</p>", "<p>The aim of this study was: i) to gain further knowledge on the tissue distribution and pathogenicity of CSFV, by directly comparing the <italic>in vivo </italic>effects of three virulence variants of the virus; ii) to investigate the applicability of various diagnostic procedures to detect the various virulence variants in the experimentally infected host animals. For these purposes, the virus distribution was determined by virus isolation (VI), the CSFV antigen was visualised in paraformaldehyde fixed, paraffin-embedded tissue sections by a monoclonal antibody based, two-step immunohistochemical method and the viral RNA was detected by <italic>in situ </italic>hybridisation, using a digoxigenin (DIG)-labelled riboprobe.</p>", "<p>By comparing the spread of three virulence variants of the virus in 64 animals, this study was performed in order to examine the tissue distribution of CSFV in the natural host, to obtain data of comparative pathology and to compare the applicability of virus detection methods. These observations will contribute to a better understanding of the viral pathogenesis and to the introduction of more effective measures to control CSF.</p>" ]

[ "<title>Methods</title>", "<title>Viruses and animals</title>", "<p>The studies involved three virulence variants of CSFV. The highly virulent ISS/60 virus was isolated from an Italian landrace pig, while the moderately virulent Lorraine isolate, alias Wingene'93, originated from a Belgian domestic pig herd [##REF##8903020##5##]. The attenuated vaccine strain Riems [subgroup 1.1., 10] was used as an avirulent representative of CSFV. The isolates were checked to be free from African swine fever virus (ASFV) by using haemabsorption-inhibition test as well as ELISA. The presence of BVDV was excluded by virus isolation and by immunoperoxidase tests (IPX) using monoclonal antibodies. All assays were performed according to OIE guidelines (In: Manual of standards for diagnostic tests and vaccines. Ed 5. Chap 2.1.12. Paris: OIE, 2004; Office International des Epizooties/World Organization for Animal Health).</p>", "<p>To compare the virulence variants, 67 conventional weaner hybrid pigs (20–25 kgs body mass) were used. The animals were clinically healthy on arrival and serologically tested to be free of CSFV, BVDV, porcine reproductive and respiratory syndrome virus (PRRSV), encephalomyocarditis virus (EMCV) and Aujeszky's disease virus (ADV) by using the standard diagnostic procedures of our institutes and our routine serological tests [##REF##9270350##11##].</p>", "<title>Experimental design</title>", "<p>A standardised protocol was used for the animal experiments, carried out by two partners of EU research project FAIR PL 95–707 in Belgium (Experiment/group II) and in Italy (Experiments/groups I and III). The conditions were harmonised within the consortium of the project. Animal experiments were approved by the ethical committees in both countries. Upon arrival, the animals were clinically examined, randomly numbered and housed in completely separated high-security isolation units. Experiments I and II involved 25 pigs each, while Experiment III was composed of 17 animals.</p>", "<p>After 6-days acclimatisation the animals (24-24-16) were intranasally inoculated with 2 ml volumes of the viruses (10³TCID⁵⁰per/ml) as follows: group I with ISS/60, group II with Wingene'93 and group III with Riems. In each experiment one uninfected, separately housed pig was used as negative control.</p>", "<p>The pigs were sequentially killed by electrocution on various post infection days (PIDs) as indicated in Tables ##TAB##0##1## and ##TAB##1##2##.</p>", "<title>Clinical examinations and sample collection</title>", "<p>The pigs were monitored daily for clinical signs. Rectal temperatures were recorded every day throughout the experiments. Blood samples were collected for VI on all sampling days.</p>", "<p>After euthanasia or death, necropsies were performed and gross lesions were recorded. Tissue samples of tonsils, spleen, ileocoecal, mesenteric and submandibular lymph nodes, kidneys, lungs, heart muscle, cerebrum, cerebellum and striated muscle (M. longissimus dorsi and M. quadriceps) were collected from all animals except seven pigs, which died in Experiment I between PIDs 5 and 7.</p>", "<title>Virus isolation (VI)</title>", "<p>VI was performed from tissue and blood samples. About 1 cm³of tissue samples were homogenised in 9 ml MEM culture medium using an Ultraturrax (Junke and Kunkel). The suspension was centrifuged at 4,000 × g for 10 min and 300 μl of the supernatant was inoculated onto a non-confluent monolayer of BVDV-free PK15 cell cultures on multi-dish plates (Falcon 35; 3047). Concerning the blood samples, serum was separated, 100 μl was diluted in 900 μl culture medium and 300 μl amount of the dilution was inoculated onto a non-confluent monolayer of BVDV-free PK15 cell cultures in a multi-dish plate. The plates were incubated for 48 hours, fixed with isopropanol and stained with a polyclonal immunoperoxidase conjugated polyclonal serum with a dilution of 120 (OIE Manuals 2004 and 2008, <ext-link ext-link-type=\"uri\" xlink:href=\"http://www.oie.int\"/>).</p>", "<title>Direct immunofluorescence (DIF) in Experiment II</title>", "<p>Due to practical reasons and the various technical facilities available at our institutes, the DIF studies were restricted to the 25 animals of experimental group II. This group was selected for the DIF investigations, considering that the moderately virulent (or low virulent) variants of CSFV have large epidemiological importance, since due to the lack of typical clinical manifestation these cases may easily be overlooked in the field. Considering that this may lead to a delayed detection of the disease, special attention should be focused on the comparative pathology of such variants of CSFV. The cell cultures or the cryostat sections were fixed with acetone and stained with fluorescent anti-CSF polyclonal serum by following OIE guidelines (In: Manual of standards for diagnostic tests and vaccines. Ed 5. Chap 2.1.12. Paris: OIE, 2004).</p>", "<title>Histopathology, immunohistochemistry (IHC) and <italic>in situ </italic>hybridisation (ISH)</title>", "<p>For histopathological and immunohistochemical examinations, the collected tissue samples were fixed in 4%, freshly prepared, buffered paraformaldehyde, embedded in paraffin according to routine histological procedures and sectioned at the thickness of 5 μm. The sections were stained with haematoxylin-eosin for histopathological evaluation.</p>", "<p>For immunohistochemical examinations, monoclonal antibody \"WH 303\", specific to CSFV glycoprotein E2, was kindly provided by Dr. David Paton, Veterinary Laboratory Agencies, Addlestone, (recent affiliation: Pirbright Laboratory), UK. The antibody was applied by a two-step peroxidase method, using the DakoEnVision +HP mouse Kit (Dakopatts, Glosstrup, Denmark). Briefly, deparaffinised tissue sections were rinsed in 0.5-mol/l Tris-HCl buffer, pH 7.6 containing 0.15 mol/l NaCl (TBS). Endogenous peroxidase was inactivated by incubating the sections with 1% (v/v) hydrogen peroxide in TBS for 20 min. The tissue sections were then rinsed thoroughly in TBS and incubated for a further 10 min with 2% bovine serum albumin (BSA) in TBS at room temperature before incubation with the primary antibody overnight at 4°C. The monoclonal antibody WH 303 was diluted 1: 200 in TBS containing 1% BSA. As negative controls, duplicate sections were incubated with 2% BSA instead of specific primary antibodies. The sections were washed three times for 5 min each time in TBS followed by 30 min incubation with one drop of peroxidase conjugated rabbit anti-mouse secondary antibody. After a washing step in TBS, peroxidase activity was visualised by incubation sections in TBS containing 0.06% (w/v) 3, 3'diaminobenzidine tetrahydrochloride (DAB, Sigma, St. Louis, USA) and 0.034% (v/v) hydrogen peroxide for 8 min. Finally, the sections were rinsed in tap water, counterstained in Mayer's haematoxylin and mounted with Entellan (Merck, Darmstadt, Germany).</p>", "<p><italic>In situ </italic>hybridisation was performed on sections processed as for IHC and mounted onto 3-aminopropyltrietoxysilane-coated slides (Sigma, St. Louis, MO, USA). Prior to deparaffinisation and rehydration in graded ethanol the slides were heated to 75°C for 15 min. In order to improve the probe penetration, the sections were digested with protease VIII 0.25 mg ml^-1at 25°C for 15 min. Finally the slides were washed twice in distilled water, dehydrated in graded ethanol and air-dried. The DIG-labelled riboprobe was synthesised from a HindIII-BamHI fragment of an infectious cDNA clone of CSFV Riems cloned into pBlueScript II SK+ (Stratagene, La Jolla, CA). Negative strand RNA representing nucleotides 6436-5711 of the Riems full-length sequence was in vitro transcribed using the DIG RNA labelling kit (Roche). The hybridisation mixture consisted of 50% formamide, 10% dextran-sulphate, 2 × SSC (1 × SSC = 0.15 M sodium chloride, 0.015 M sodium citrate), 0.1 mM EDTA, 1 mM Tris-HCl pH 7.5, denatured salmon sperm DNA to a final concentration of 4 mg ml^-1and 0.5 ng/μl freshly denatured DIG-labelled riboprobe. Sixty μl of the hybridisation mixture was applied per slide. The tissue sections were covered and sealed by Frame-Seal chambers (MJ Research Inc, Watertown, MA, USA). The slides were then placed into a PTC 200 Peltier Thermal Cycler (MJ Research Inc) equipped with an interchangeable Twin Towers <italic>in situ </italic>block, heated to 65°C for 15 min. The hybridisation was carried out at 55°C for 2 hours. A bovine herpesvirus type 5 (BHV-5) specific DIG-labelled probe [##REF##10391508##12##] was used on duplicate sections as negative controls. After hybridisation the slides were gently washed as follows: twice (5 min each) with 4 × SSC at room temperature, twice (5 min each) with 1 × SSC at room temperature and once with 0.1 × SSC for 15 min at 55°C. The sections were not allowed to dry at any time during or following the washing steps of post-hybridisation. For the immunological detection of the digoxigenin-labelled hybrids, a DIG Nucleic Acid Detection Kit (Boehringer Mannheim, Germany) was used according to the manufacturer's instructions, utilizing an antibody-conjugate (anti-digoxigenin alkaline phosphatase conjugate, anti-DIG-AP) and an enzyme-catalysed colour reaction with 5-bromo-4-chloro-3-indolyl phosphate (BCIP) and nitroblue tetrazolium salt (NBT), providing a blue-coloured precipitate.</p>" ]

[ "<title>Results</title>", "<title>Clinical signs and viraemia</title>", "<p>In Experiment I (highly virulent virus), three pigs developed febrile reaction (40–40.3°C) at post infection day (PID) 1. From PID 2, twelve out of 18 animals showed pyrexia up to 42°C, which persisted throughout the observation period. Some pigs developed inappetence, apathy and mild diarrhoea from PID 1. Starting from PID 3, three animals showed staggering, shivering and incoordination. At PID 5, one piglet developed posterior paresis. At PID 8, the remaining one piglet showed nervous symptoms such as locomotoric ataxia and paresis. Cutaneous lesions were constantly absent. The animals, which were not sacrificed, died from PIDs 5 to 7 (Table ##TAB##0##1##). Viraemia, as recorded by virus detection in the serum samples, started at PID 2 in three animals and at PID 3 all the animals but two became viraemic, as it was shown by the VI assays. From PID 4 all pigs showed viraemia until the end of the experiment.</p>", "<p>In Experiment II (moderately virulent virus) the first febrile reactions were noticed in two pigs at PID 2 and half of the inoculated animals successively developed fever, up to 41.5°C during the observation period. Apathy and inappetence were recorded at PID 11. At PID 12 diarrhoea and a stringent respiration were noticed. Skin haemorrhages and ataxia appeared one day before death, on PIDs 9 and 13. The animals that were not sacrificed died at PIDs 10 and 14 (Table ##TAB##1##2##). Viraemia started at PID 5 in one animal.</p>", "<p>In Experiment III (avirulent vaccine strain), all the animals showed slightly elevated temperature with an average of 0.4 – 0.9°C from PID 1 until the end of the experiment. No other clinical signs were recorded in the group infected with the avirulent CSFV strain. Viraemia was not observed in this group.</p>", "<title>Gross pathology</title>", "<p>In Experiment I, one animal presented a distinct swelling of the submandibular lymph nodes at PID 2. Spleen infarction was seen in one pig at PID 3. From PID 4, all the remaining animals showed evidence of typical CSF lesions characterized by severe enlargement of lymph nodes with haemorrhages in the periphery, spleen infarction and petechial haemorrhages in the renal cortex.</p>", "<p>The macroscopic lesions in Experiment II were swollen lymph nodes with discrete petechial haemorrhages and haemorrhages in the kidneys of the pigs that were killed at PID 8. Only in the pigs sacrificed and died at the terminal phase of the experiment from PID 12 became the signs more pathognomonic.</p>", "<p>At the post mortem examination of the pigs in Experiment III a general swelling of the lymph nodes was observed in one pig at 36 hours after inoculation. Mild haemorrhages were seen in the lymph nodes of the head and neck regions in one animal at PID 2.</p>", "<p>No macroscopic pathological changes were observed in the uninfected control pigs.</p>", "<title>Virus isolation (VI) from tissue samples</title>", "<p>The virus was detected (re-isolated) from the tissue samples in all the three experiments.</p>", "<p>In Experiment I, the virus was isolated from the tonsils, spleen, lymph nodes and heart muscle at PID 2. Subsequently, the VI tests detected the virus from the tonsils and lymph nodes of all infected animals, with the exception of the tonsil samples of one pig (Table ##TAB##0##1##). The virus was also re-isolated from the spleen, kidneys, lungs, heart, brain and striated muscles, as shown in Table ##TAB##0##1##.</p>", "<p>In Experiment II, CSFV was detected at PID 4 in the tonsil and ileocoecal lymph node of one pig. From PID 5, the virus was isolated from the tonsils and lymph nodes of all infected animals. VI detected the virus also in the spleen, kidneys, lungs, heart, brain and in the striated muscles, see Table ##TAB##1##2##.</p>", "<p>In Experiment III, CSFV was isolated only from the tonsils of three animals at PIDs 3, 5 and 7 and from the ileocoecal lymph node of one pig at PID 7 and mesenteric lymph node of one animal at PID 8.</p>", "<p>The results of virus isolation from tissue specimens are summarised in Tables ##TAB##0##1## and ##TAB##1##2##.</p>", "<title>Direct immunofluorescence in Experiment II</title>", "<p>By the means of DIF, the virus was detected in tonsils, in the superficial and crypt epithelial cells, macrophages, lymphoid and endothelial cells from PID 4 and the fluorescence staining remained fairly homogenous until PID 14, at the end of the experiment. In the spleen, immunofluorescence was first observed at PID 7 in lymphoid and endothelial cells. In the lungs, positive staining was found in the bronchiolar mucosal epithelial cells as well as in the alveolar macrophages and in a few endothelial cells from PID 8 until the end of the experiment. In the kidneys, only a small amount of positively stained duct epithelial, endothelial and mononuclear cells were observed in seven animals from PID 6. In the myocardium, immunostaining was seen only in one pig at PID 10. The immunoreactivity was observed in the endothelial cells of the small capillaries. In the brain and muscle specimens positive immunofluorescence staining has not been detected.</p>", "<title>Histopathological, immunohistochemical examinations and in situ hybridisation</title>", "<p>Microscopic lesions were observed in the examined organs of pigs in all the three infected groups. The changes were more frequent and severe in Experiments I and II. The monoclonal antibody, specific to gp E2 of CSFV, gave specific positive cytoplasmic staining reaction in tonsils, spleen, lymph nodes, lungs and kidneys but not in myocardium and striated muscles. Further immunopositivity was detected in nervous tissues in one single animal in Experiment I (Tables ##TAB##0##1## and ##TAB##1##2##).</p>", "<title>Experiment I</title>", "<p>In tonsils the lesions consisted of some cystically enlarged or plugged tonsillar crypts with cellular debris, neutrophil granulocytes and keratin. A mild hypertrophy of the follicles was observed from 36 hours after inoculation. Necrotic changes were also seen from PID 1. Specific immunoreactivity was detected first in a few crypt-epithelial cells and many migrating macrophages, as well as in the lymphoid cells at PID 2. The immunostaining became more disseminated from PID 4 in the crypt-epithelial cells, macrophages lymphoid and endothelial cells and remained fairly homogenous until PID 8, at the end of the experiment. In addition, at PID 8 very strong immunostaining was observed in the superficial-epithelial cells (Figure ##FIG##0##1##). In lymph nodes, a mild depletion/atrophy of the follicles was seen between 12–24 hours after inoculation, followed by a mild follicular and perifollicular hypertrophy from 36 hours after infection until the end of the experiment. Necrotic changes were seen first in the follicles at PID 1 and became then more diffuse. Acute focal haemorrhages were found in two lymph nodes. Specific immunostaining was observed in reticular cells, macrophages, lymphoid and a few endothelial cells from 60 hours after infection. A fairly uniform, lower amount of virus antigen could be detected in all the lymph nodes at PID 3 and a still uniform but higher amount of positively stained cells between PID 4 and 8. In spleen, a mild depletion/atrophy of the follicles/periarterial lymphatic sheaths (PALS) and perifollicular hyperplasia was observed 12 hours after inoculation, followed by a mild hypertrophy. Immunoreactivity was first observed at PID 3 in reticular cells, macrophages, lymphoid and endothelial cells. In kidneys, six pigs had a very mild focal mononuclear interstitial nephritis between 12 hours and 2 days after inoculation. Only a small number of positively stained duct epithelial, endothelial and mononuclear cells were observed in one animal at PID 8. In lungs, very mild non-suppurative bronchointerstitial inflammatory changes were observed in all the pigs. These lesions were considered as non-specific. Specific immunoreactivity was found in the bronchial and bronchiolar mucosal epithelial cells, in the alveolar macrophages and in a few endothelial cells from PID 4 in two animals. In heart muscle, specific histopathological changes were not observed. In cerebrum and cerebellum, the main changes were confined to the vessels in form of vasculitis consisting of infiltration of mononuclear cells into the wall and around the small blood vessels, most frequently in meninges and white matter. In many cases, swelling and degenerative changes of the endothelial cells occurred. In some cases the vascular changes were accompanied by focal gliosis. The lesions developed one day after inoculation. Positive immunostaining was detected in one single animal at PID 8. In muscles, very mild focal acute muscle degeneration to variable degree and oedema were observed in all the three infected groups and control animals, throughout the experiment (results not repeated below).</p>", "<title>Experiment II</title>", "<p>In tonsils, the microscopic lesions consisted of some cystically enlarged or plugged tonsillar crypts with cellular debris, neutrophil granulocytes and keratin. Necrotic changes were seen from PID 6 and became very severe in the two pigs, which died at PID 14. Specific immunoreactivity was first detected exclusively in the crypt-epithelial cells, from PID 4. From PID 7 a higher amount of virus antigen was detected in the crypt-epithelial cells, migrating macrophages and lymphoid cells as well as in endothelial cells (Figure ##FIG##1##2##). From PID 10 the viral antigen was detected even in the superficial-epithelial cells. The immunostaining remained fairly homogenous until PID 14, at the end of the experiment. In lymph nodes, a mild depletion/atrophy of the follicles was observed from PID 7 in six pigs and a mild follicular as well as perifollicular hypertrophy from PID 3, respectively PID 6 in 12 respectively 3 pigs until the end of the experiment. These changes were most evident in the submandibular lymph nodes. Acute focal haemorrhages were seen in the submandibular lymph node of seven animals from PID 5. Follicular necrosis was observed at PID 5 and 6 in the submandibular lymph node. From PID 7 more diffuse necrotic changes were seen occasionally in all the three examined lymph nodes. Specific immunostaining was observed in reticular cells, macrophages, lymphoid and a few endothelial cells from PID 5. Immunoreactive macrophages and lymphoid cells were most evident in the reactive centre of the follicles. In the submandibular lymph node a greater number of positively stained cells were observed than in the ileocoecal and mesenteric lymph nodes between PID 5 and 8 (Figure ##FIG##2##3##). After that, a fairly uniform but lower amount of virus antigen could be detected in all the lymph nodes until PID 14. In spleen, mild follicular/PALS atrophy was recorded from PID 3. Focal haemorrhages were observed from PID 7 as well as necrotic lesions mainly in the white pulp from PID 4. These necrotic changes were very severe and characterized as vascular necrosis in one pig and as an acute-subacute fibrinopurulent-necrotic peritonitis in another one, which died at PID 14. Specific immunoreactivity was first observed at PID 4 in reticular cells, macrophages, lymphoid and endothelial cells (Figure ##FIG##3##4##). In kidneys, a few acute focal haemorrhages were seen, mainly in the medulla, in five animals from PID 6. Furthermore, mild focal mononuclear interstitial nephritis was observed in four animals and a mild acute focal glomerulonephrosis was detected in two animals at PID 14. In one pig, which died at PID 12, acute pyelonephritis was observed. Only a small amount of positively stained duct epithelial, endothelial and mononuclear cells were observed in four animals from PID 10. In the lungs, very mild non-suppurative bronchointerstitial inflammatory changes were observed in five pigs from PID 10. They consisted of vascular lesions with fibrinoid necrosis and tendency to thrombus formation. In one of these animals focal acute fibrinotic pneumonia with necrosis was also seen at PID 14. Immunoreactivity was found in the bronchial and bronchiolar epithelial cells, in the alveolar macrophages and in a few endothelial cells from PID 4 (Figure ##FIG##4##5##) until the end of the experiment. In the hearts, the pathological findings were confined to the smaller vessels of the myocardium in three pigs, from PID 10. In one animal, which died at PID 10, a marked endothelial proliferation was observed. Necrotic vasculitis occurred in two pigs, which died at PID 14. Specific immunostaining was not detected. In the cerebrums and cerebellums, similar vasculitis was observed as in Experiment I, with severe degenerative changes (Figure ##FIG##5##6##) from PID 10 until the end of the experiment in almost all pigs. In some cases the vascular lesions were accompanied by focal gliosis. In two cases, mild endothelial proliferation was observed at PID 10 and 14. In skeletal muscles a necrotic vasculitis was seen in two pigs at PID 14.</p>", "<title>Experiment III</title>", "<p>In tonsils, mild changes were characterized by expanded crypts plugged with cellular debris and keratin. Positive immunostaining was observed from PID 5, in a few crypts/crypt epithelial cells of three pigs (Figure ##FIG##6##7##). In lymph nodes, neither atrophic changes nor haemorrhages were detected, but occasionally slight follicular and perifollicular hyperplasia were seen in most of the pigs from PID 2. From PID 5, very mild necrotic lesions of variable degree were observed in the lymph follicles of four animals. Positive immunostaining in macrophages was observed in one submandibular and one mesenteric lymph node at PID 7 and 8, respectively. No changes were noted in the spleens. In kidneys, a very mild focal interstitial nephritis with mononuclear cells was seen in about the half of the animals throughout the observation period. In addition, a focal mononuclear perivasculitis in the medulla was detected in two pigs at 60 hours, respective four days after infection. Specific immunostaining was not detected. In brain tissue, necrotic lesions were not seen, only swelling of the endothelial cells of the small vessels was observed. Specific immunostaining was not detected.</p>", "<p>A general observation was that the microscopic changes in Experiments I and II became progressively more severe. In contrast, the changes seen in Experiment III remained fairly homogenous throughout the observation period.</p>", "<p>In the uninfected control pig, histopathological changes and positive immunoreactivity was not observed. The sections of infected animals showed negative results when instead of specific antibody, 2% BSA was applied.</p>", "<p>The presence of CSFV nucleic acid was demonstrated by a pilot <italic>in situ </italic>hybridisation in various organs in all the three experiments. In experiment I the tonsils gave rather strong positive signals (4–10 foci/section) as early as 60 hours post infection. On PID 8 the distribution of viral nucleic acids was wide, strong hybridisation signals (&gt; 10 foci/section) were seen in the tonsils, spleen, kidneys, various lymph nodes and in the lungs. In experiment II also the tonsils became first positive, but much later then in Experiment I. The first positive results in the tonsils were seen here 4 days post infection. Subsequently, 5 days post infection the spleen became positive; while on PID 8 the tonsils, spleen, kidneys, lymph nodes and lungs harboured viral nucleic acids. By reading the hybridisation assay, fewer foci were seen then in Experiment I (1–3 foci per section). The positive nucleic acid hybridisation signals in Experiment III were fewer (1–3 foci per section) and restricted to the tonsils and lymph nodes. The signals were observed between PIDs 3 to 8 in this group. The hybridisation signals were observed in the cytoplasm of the epithelial (Figure ##FIG##7##8##), mononuclear and reticular cells. When using the probe on the sections of the uninfected animals or the BHV-5 specific probe on the sections of the infected pigs, no hybridisation signal was observed.</p>" ]

[ "<title>Discussion</title>", "<p>Although CSF is registered as one of the most important Transboundary Animal Diseases (TADs), notifiable to OIE, the regular re-occurrence of the outbreaks in various regions of the world indicates that many questions are still poorly answered concerning the biology of this devastating disease. One of the problems is that CSF has an increasing tendency to appear and re-appear in a clinically very mild or in a completely unapparent form. By being unnoticed for a time, such mild infections may spread to large populations of pigs, causing serious epizootiological and economic consequences.</p>", "<p>Considering the varying clinical manifestations and the observed diagnostic problems, further research has to be conducted on comparative pathology of CSFV, with special regard to the emerging new viral variants of very mild pathogenicity, causing very weak or completely unapparent clinical symptoms, which can easily be overlooked in the field. Regarding these requirements, we were conducting here <italic>in vivo </italic>studies on three groups of experimentally infected pigs, in order to compare the effects of viruses of varying virulence, which may occur in the field either as single or as multiple infections.</p>", "<p>Other research groups reported on comparative <italic>in vivo </italic>analysis of CSFV strains [##REF##1445070##8##,##REF##11055862##9##,##REF##4179763##13##] but none of the previous investigations provided such a comprehensive analysis of various virulence variants as the present study. The comparative analysis, performed on a large number of pigs under harmonised experimental conditions, is providing further data and demonstration material on the pathogenesis of CSF. In addition, the data are useful for the improvement of CSF diagnosis, with special regard to cases when the virus replication results only in the mild or inapparent clinical symptoms. Considering the high number of pigs (67 animals), it was preferable to divide the tasks and to perform the experiments at two partner laboratories in parallel, under harmonised experimental conditions. The same age groups of pigs were infected and sampled by using standardised procedures. The evaluation methods were also harmonised (like gross pathology, virus isolation) and all samples were collected for testing in a single laboratory by the same researcher (like histopathology, IHC and ISH).</p>", "<p>The clinical signs, which are important for the early detection of the new cases of CSF infections in the field and for early warning [##REF##3992564##4##] varied remarkably in the three groups. The febrile reactions were very marked in group I, since high fever was recorded already at PID 1 and it lasted throughout the experiment. Group II showed a later occurring and milder febrile reaction, while group III reacted only with a slightly elevated temperature, which did not show a marked profile. Inappetence also varied strongly; in groups I and II reduced appetite was observed from PID 1, respectively from PID 11, while in group III loss of appetite was not observed. Nervous symptoms also appeared with great variations: group I developed serious signs of the involvement of nervous system, while the other two groups remained symptomless, except for three pigs in group II, which developed slight ataxia one day before death.</p>", "<p>Concerning viraemia, the differences were also very clear among the groups. In groups I and II the viral invasion in the blood circulation was recorded from PID 2, respectively PID 5, while group III developed no measurable viraemia. It is worth to note that the frequency of viraemia showed strong variations: in group I all animals became viraemic (from PID 4), while in group II only one, indicating that the virulence variants had various capacities of <italic>in vivo </italic>viral replication and invasion.</p>", "<p>It is a known fact that the moderate or the low virulence variants of CSFV frequently cause very mild and/or unapparent clinical symptoms, which are accompanied by a restricted <italic>in vivo </italic>viral replication and invasion [##REF##4179763##13##,##REF##1516362##14##]. This phenomenon was clearly demonstrated and confirmed in the present experiments. Concerning epidemiology and early diagnosis of CSF, group II is the most interesting in our present studies. Based on our previous experiments and on the observations of other groups, we supposed that group II requires special attention (see notes above). This is the reason that group II was tested not only by the same methods as the two other groups, but also by DIF, in order to investigate the tissue distribution of the moderately virulent virus by as many means as possible. This selected group showed that a moderately virulent virus is able to cause infection without the development of any apparent clinical response. Simultaneously, the viral replication and invasion showed a restricted tendency in animals infected with the moderately virulent virus. In this group the number of diseased animals was lower than in group I and only three pigs died. Noteworthy, the development of the CSF varied remarkably in group II between the individual animals, ranging from a symptomless infection to typical, fatal cases. One can conclude that the lack of clinical symptoms and of detectable virus in the blood circulation in a number of animals may create serious problems in the early detection of an outbreak, caused by such variants of the virus. Due to the lack of clinical signs of diagnostic importance, special attention has to be paid to detect successfully and immediately such cases of CSFV infection in the field.</p>", "<p>By comparing the gross pathological changes, in group I a distinct swelling of the submandibular lymph nodes was seen as early as PID 2. From PID 4 all the infected animals showed typical CSF lesions. In contrast, group II exhibited gross pathological signs only from PID 8, such as swelling and haemorrhages of lymph nodes as well as haemorrhages of kidneys. The wide range of typical pathological changes occurred in this group only from PID 12. In group III the lack of pathological changes indicated the attenuated character of the virus. One can conclude that the gross pathological findings were in good correlation with the clinical pictures observed in the three groups.</p>", "<p>The histopathological examinations revealed marked differences among the three groups, which agreed with the clinical and gross pathological findings. The microscopic lesions included vascular changes and necrosis of lymphocytes, which were observed in all the three infected groups. The changes were more frequent and severe in the first two groups. Encephalitis, another major histopathological lesion, was seen only in groups I and II. Compared to group I, in group II the above mentioned lesions developed 5–6 days later, and remarkably, they became more severe at the termination of the experiment. These findings indicate that the highly and the moderately virulent viruses have rather similar capacities to induce histopathological changes, but in the case of the latter, these changes develop after a longer incubation period. These days, when the animals are already CSFV infected but neither clinical signs, nor histopathological changes are yet observed, creates an important risk-period in the safe early diagnosis of CSF.</p>", "<p>By evaluating the findings in comparative histopathology, it has to be stated that in groups I and II the microscopic changes became progressively more severe during the development of the disease, in contrast to the lesions seen in group III, which remained fairly unaltered. These are factors, which should be considered in the comparative pathology and diagnosis of CSF.</p>", "<p>Immunohistochemistry, in correlation to histopathology, also revealed marked differences among the groups. For example, group I showed necrotic changes of the lymphoid cells in the tonsils as early as PID 1 and the viral antigens became detectable from PID 2. In contrast, group II showed lymphoid cell necrosis only from PID 6. It is worth to note that the viral antigen appeared in group II prior to necrosis, since IHC became positive already from PID 4. The observed differences indicate the possibility of a very prompt and destructive viral replication in group I, leading to early cellular damage appearing very rapidly, before the detection of the virus by IHC. In contrast, in group II IHC revealed the signs of viral replication before the appearance of necrotic alterations. This indicates characteristic differences in the replication features and in the pathobiology of these two virulence variants. The phenomenon has diagnostic importance, since it illustrates that in the case of moderately virulent viruses, IHC is detecting the viral infection earlier, compared to the histological examinations. This is in accordance with the observations of Kamolsiriprichaiporn et al. [##REF##1445070##8##]. Thus, the viral antigen demonstration is very important in the early detection of CSFV infections, especially in cases caused by moderately virulent viruses. The importance of diagnostic IHC is further emphasized by the observation that this test gave positive results very long time (seven days in our case) before the appearance of the clinical symptoms.</p>", "<p>In contrast to the other two groups, no necrotic changes of the tonsils were detected in group III, confirming the very low virulent or avirulent character of the vaccine virus. However, viral antigen was detected in the tonsils by IHC, indicating viral replication. When discussing the virulence level of the vaccine strain, it is noteworthy that in the lymph nodes even this virus was able to induce necrotic changes from PID 5. These changes were presumably connected to viral replication, since IHC revealed the presence of viral antigens from PID 5 in the lymph nodes.</p>", "<p>The results of virus isolation were in accordance with the tendency of IHC, since in group I the virus was detected in tonsils, spleen and lymph nodes as early as PID 2, while in group II it was isolated first on PID 4 from tonsils and lymph nodes. In group III the results of virus isolation agreed with the findings of IHC, since the vaccine strain was demonstrated in the lymphoid tissues between PID 3 and 7. However, it is remarkable that the replication of the virus was demonstrated by this test in not more than three animals. One can hypothesize that this was either a technical problem, or the amount of vaccine virus was so low that it was under the level of the detection capacity of the VI test in quite a number of animals.</p>", "<p>As further tools of direct virus detection, DIF and IHC proved to be complementary methods to the \"golden standard\" of VI [##REF##11055862##9##]. In our experiments IHC revealed the presence of the virus in the tonsils and lymph nodes in group I as early as PIH 60. It is interesting that in striated muscles and heart the virus was detected by VI, but not by IHC. One can speculate that this might be due to two basic reasons: i) the sensitivity of IHC is lower; ii) the virus is transported to these organs by blood, due to viraemia, which is detected by VI but not localised by the IHC method [##REF##9763128##15##]. In agreement with the previous results, DIF and IHC detected the virus in group II from PID 4. Similarly to VI, the virus was detected by DIF and IHC in the tonsils and in addition, IHC gave positive results also in the spleen and lungs. It is noteworthy, that nervous tissue showed both histopathological lesions and the presence of viral antigen in Experiment I, while in Experiment II in spite of severe histopathological changes, CSFV antigen was not detected. To explain this peculiar phenomenon, one can speculate that: i) early cell damage may occur already at initial stage of viral replication, when the viral load is still low; ii) immune-mediated reactions may play role [##REF##16846995##16##]. Concerning group III, the results of IHC indicated some virus replication between PIDs 5 and 8 in the tonsils and in the lymph nodes, but similarly to VI, only in three animals. This finding confirms that the vaccine strain replicates in the lymphoid tissues; while the amount of replicating virus is presumably low.</p>", "<p>The quantitative tendencies of the virus replication will be investigated by the real-time PCR assays of our laboratories [##REF##11901858##17##] in the forthcoming experiments Since the viral nucleic acid detection and quantification by PCR and by other means of molecular diagnosis has various approaches and variants, the involvement of those assays would turn the present paper extremely long, complicated and multidisciplinary. Thus, herewith we focused on the comparison of morphology-associated descriptions and diagnostic approaches; compared to the golden standard of virus isolation, while the PCR investigations will be reported and discussed in separate articles.</p>", "<p>The present results show that the <italic>in situ </italic>hybridisation technique, developed in this study, is a useful tool for the detection of CSFV in formalin fixed, paraffin embedded tissue samples. Similarly to the observation of Choi and Chae [##REF##12627722##18##]<italic>in situ </italic>hybridisation assays provide sensitive means for studying the pathogenesis of acute and chronic CSFV infections.</p>", "<p>The parallel studies on the three experimental groups allowed not only the comparison of clinical, pathological and virological parameters, but also the estimation of further aspects of disease development. It is clear that group I represented the case of rapidly developing, fatal CSF. However, group II, which developed an initially more subtle, milder disease, revealed many aspects, which might be useful considering the recent epizootiological situations in swine populations. An interesting observation in this group is that the number of cells immunopositive for the viral antigen had a tendency to decline slightly during the course of virus infection. Similar tendency has been reported in case of BVDV [##REF##9672623##19##]. Concerning CSFV, Sánchez-Gordón et al. [##REF##12724565##20##] have observed a similar decline in the tonsils. The intensity of the phenomenon varied in various experiments and the authors hypothesize that the differences might have been due to the timing of virus spread or differences in the local immune responses [##REF##12724565##20##].</p>" ]

[ "<title>Conclusion</title>", "<p>The <italic>in vivo </italic>studies and the accompanied diagnostic approaches provided useful data on the comparative pathology of three virulence variants of CSFV. The experiments confirmed the previous expectations that the three variants represent various levels of pathogenicity. Data have been obtained concerning comparative aspects of clinical manifestations, development of pathological signs and tissue distribution of CSFV variants. These data have practical importance when discussing the pathobiology of classical swine fever in the host species. The observations are useful for the early diagnosis of classical swine fever, with special regard to the detection and identification of the very mild or inapparent clinical manifestations. The present study demonstrates that in the case of the highly and moderately virulent virus variants the virulence does not affect the pattern of the spread in a pig, but influences the onset, intensity, duration and outcome of the disease. As far as diagnostic tools are concerned, IHC provides useful means of early virus detection and it indicates the localisation of the virus spread in tissues, supporting the determination of the pathogenicity levels of newly emerging viruses.</p>" ]

[ "<p>This is an Open Access article distributed under the terms of the Creative Commons Attribution License (<ext-link ext-link-type=\"uri\" xlink:href=\"http://creativecommons.org/licenses/by/2.0\"/>), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.</p>", "<title>Background</title>", "<p>The aim of this study was to compare the tissue distribution and pathogenicity of three virulence variants of classical swine fever virus (CSFV) and to investigate the applicability of various conventional diagnostic procedures.</p>", "<title>Methods</title>", "<p>64 pigs were divided into three groups and infected with the highly virulent isolate ISS/60, the moderately virulent isolate Wingene'93 and the live attenuated vaccine strain Riems, respectively. Clinical signs, gross and histopathological changes were compared in relation to time elapsed post infection. Virus spread in various organs was followed by virus isolation, by immunohistochemistry, applying monoclonal antibodies in a two-step method and by <italic>in situ </italic>hybridisation using a digoxigenin-labelled riboprobe.</p>", "<title>Results</title>", "<p>The tissue distribution data are discussed in details, analyzing the results of the various diagnostic approaches. The comparative studies revealed remarkable differences in the onset of clinical signs as well as in the development of the macro- and microscopical changes, and in the tissue distribution of CSFV in the three experimental groups.</p>", "<title>Conclusion</title>", "<p>The present study demonstrates that in the case of highly and moderately virulent virus variants the virulence does not affect the pattern of the viral spread, however, it influences the outcome, the duration and the intensity of the disease. Immunohistochemistry has the advantage to allow the rapid detection and localisation of the virus, especially in cases of early infection, when clinical signs are still absent. Compared to virus isolation, the advantage of this method is that no cell culture facilities are required. Thus, immunohistochemistry provides simple and sensitive tools for the prompt detection of newly emerging variants of CSFV, including the viruses of very mild virulence.</p>" ]

[ "<title>Competing interests</title>", "<p>The authors declare that they have no competing interests.</p>", "<title>Authors' contributions</title>", "<p>KB performed the histopathological, immunohistochemical and <italic>in situ </italic>hybridisation studies, participated in the evaluation and summarizing of the findings and wrote the draft of the manuscript. FK applied for funding of the project, participated in the design of the study and performed the second animal experiment inclusive virus isolation as well as participated in the evaluation of the findings and had a major impact on the manuscript. HV participated in the design of the study and performed and evaluated the second animal experiment inclusive virus isolation. CM participated in the design and performing of the <italic>in situ </italic>hybridisation study, evaluated the findings and influenced the manuscript. FF participated in performing the second and third animal experiments inclusive virus isolation and evaluating the results. GMDM participated in the design of the study, participated in performing the second and third animal experiments inclusive virus isolation, evaluated the findings and influenced the manuscript. SB applied for funding of the project, participated in the design of the study and had a major impact on the manuscript. All authors read and approved the final manuscript.</p>" ]

[ "<title>Acknowledgements</title>", "<p>This work was supported by the European Commission (Grants: FAIR PL 95-707 and SSP1-501599). The authors thank the laboratory staff for the excellent technical assistance. Ms. Irja Johansson is specially acknowledged for her valuable support and performing of <italic>in situ </italic>hybridisation. Many thanks are due to Professor Carl Hård af Segerstad and Dr. Dolores Gavier-Widén for the critical reading and constructive suggestions.</p>" ]

[ "<fig position=\"float\" id=\"F1\"><label>Figure 1</label><caption><p><bold>Tonsil. Experiment I, PID 8. Positive immunohistochemical staining</bold>. Tonsil. Experiment I, PID 8. Superficial-epithelial cells, macrophages and lymphoid cells staining intensely for CSFV antigen in the cytoplasm with a monoclonal antibody specific for glycoprotein E2 (WH 303). Immunohistochemistry; EnVision™ +HP mouse system. Magnification 540×.</p></caption></fig>", "<fig position=\"float\" id=\"F2\"><label>Figure 2</label><caption><p><bold>Tonsil. Experiment II, PID 7. Positive immunohistochemical staining</bold>. Tonsil. Experiment II, PID 7. Immunoreactivity to WH 303 monoclonal antibody as a cytoplasmic rim in the crypt-epithelial cells. Immunohistochemistry; EnVision™ +HP mouse system. Magnification 540×.</p></caption></fig>", "<fig position=\"float\" id=\"F3\"><label>Figure 3</label><caption><p><bold>Lymph node. Experiment II, PID 6. Positive immunohistochemical staining</bold>. Lymph node. Experiment II, PID 6. Immunoreactivity to WH 303 monoclonal antibody in the cytoplasm of the reticulocytes and macrophages. Immunohistochemistry; EnVision™ +HP mouse system. Magnification 540×.</p></caption></fig>", "<fig position=\"float\" id=\"F4\"><label>Figure 4</label><caption><p><bold>Spleen. Experiment II, PID 7. Positive immunohistochemical staining</bold>. Spleen. Experiment II, PID 7. Immunoreactivity to WH 303 monoclonal antibody in the cytoplasm of reticulocytes and macrophages. Immunohistochemistry; EnVision™ +HP mouse system. Magnification 540×.</p></caption></fig>", "<fig position=\"float\" id=\"F5\"><label>Figure 5</label><caption><p><bold>Lungs. Experiment II, PID 14. Positive immunohistochemical staining</bold>. Lungs. Experiment II, PID 14. Immunoreactivity to WH 303 monoclonal antibody in the cytoplasm of the bronchiolar epithelial cells. Immunohistochemistry; EnVision™ +HP mouse system. Magnification 1080×.</p></caption></fig>", "<fig position=\"float\" id=\"F6\"><label>Figure 6</label><caption><p><bold>Brain, blood vessel. Experiment II, PID 12. Degenerative changes</bold>. Brain, blood vessel. Experiment II, PID 12. Degenerative changes (pyknosis and karyorrhexis) of the endothelial cells. Haematoxylin-eosin staining. Magnification 540×.</p></caption></fig>", "<fig position=\"float\" id=\"F7\"><label>Figure 7</label><caption><p><bold>Tonsil. Experiment III, PID 8. Positive immunohistochemical staining Tonsil</bold>. Experiment III, PID 8. Immunoreactivity to WH 303 monoclonal antibody as a cytoplasmic rim in the crypt-epithelial cells. EnVision™ +HP mouse system. Magnification 540×.</p></caption></fig>", "<fig position=\"float\" id=\"F8\"><label>Figure 8</label><caption><p><bold>Tonsil. Experiment II, PID 4. <italic>In situ </italic>hybridisation</bold>. Tonsil. Experiment II, PID 4. Intense hybridisation signal for CSFV nucleic acid in the cytoplasm of tonsillar crypt epithelial cells. <italic>In situ </italic>hybridisation; DIG-labelled riboprobe. Magnification 675×.</p></caption></fig>" ]

[ "<table-wrap position=\"float\" id=\"T1\"><label>Table 1</label><caption><p>Results of Experiment I; pigs, infected with the highly virulent isolate, ISS/60</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td align=\"center\"><bold>PID(h)</bold></td><td align=\"center\"><bold>Pig No.</bold></td><td align=\"center\"><bold>Tonsils</bold></td><td align=\"center\"><bold>Spleen</bold></td><td align=\"center\"><bold>Kidneys</bold></td><td align=\"center\"><bold>Ln. 1</bold></td><td align=\"center\"><bold>Ln. 2</bold></td><td align=\"center\"><bold>Ln. 3</bold></td><td align=\"center\"><bold>Lungs</bold></td><td align=\"center\"><bold>Heart</bold></td><td align=\"center\"><bold>Cerebr.</bold></td><td align=\"center\"><bold>Cerebel.</bold></td><td align=\"center\"><bold>Musc. 1</bold></td><td align=\"center\"><bold>Musc. 2</bold></td></tr></thead><tbody><tr><td align=\"center\"><bold>-1</bold></td><td align=\"center\"><bold>1</bold></td><td align=\"center\">*-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td></tr><tr><td align=\"center\"><bold>12 h</bold></td><td align=\"center\"><bold>2</bold></td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td></tr><tr><td/><td align=\"center\"><bold>3</bold></td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td></tr><tr><td align=\"center\"><bold>1</bold></td><td align=\"center\"><bold>4</bold></td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td></tr><tr><td/><td align=\"center\"><bold>5</bold></td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/</td><td align=\"center\">-/-</td></tr><tr><td/><td align=\"center\"><bold>6</bold></td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td></tr><tr><td/><td align=\"center\"><bold>7</bold></td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td></tr><tr><td align=\"center\"><bold>2</bold></td><td align=\"center\"><bold>8</bold></td><td align=\"center\">+/-</td><td align=\"center\">+/-</td><td align=\"center\">-/-</td><td align=\"center\">+/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">+/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td></tr><tr><td/><td align=\"center\"><bold>9</bold></td><td align=\"center\">+/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td></tr><tr><td/><td align=\"center\"><bold>10</bold></td><td align=\"center\">+/+</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">+/+</td><td align=\"center\">-/+</td><td align=\"center\">-/+</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td></tr><tr><td/><td align=\"center\"><bold>11</bold></td><td align=\"center\">+/+</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">+/+</td><td align=\"center\">+/+</td><td align=\"center\">++</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td></tr><tr><td align=\"center\"><bold>3</bold></td><td align=\"center\"><bold>12</bold></td><td align=\"center\">+/+</td><td align=\"center\">+/+</td><td align=\"center\">+/+</td><td align=\"center\">+/+</td><td align=\"center\">+/+</td><td align=\"center\">+/+</td><td align=\"center\">+/-</td><td align=\"center\">+/-</td><td align=\"center\">+/-</td><td align=\"center\">+/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td></tr><tr><td/><td align=\"center\"><bold>13</bold></td><td align=\"center\">+/++</td><td align=\"center\">+/-</td><td align=\"center\">-/-</td><td align=\"center\">+/+</td><td align=\"center\">+/+</td><td align=\"center\">+/+</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td></tr><tr><td align=\"center\"><bold>4</bold></td><td align=\"center\"><bold>14</bold></td><td align=\"center\">+/+++</td><td align=\"center\">+/+</td><td align=\"center\">+/+</td><td align=\"center\">+/+++</td><td align=\"center\">+/++</td><td align=\"center\">+/+++</td><td align=\"center\">+/++</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">+/-</td><td align=\"center\">-/-</td></tr><tr><td/><td align=\"center\"><bold>15</bold></td><td align=\"center\">+/+++</td><td align=\"center\">+/+++</td><td align=\"center\">+/+++</td><td align=\"center\">+/+++</td><td align=\"center\">+/+++</td><td align=\"center\">+/+++</td><td align=\"center\">+/-</td><td align=\"center\">+/-</td><td align=\"center\">+/-</td><td align=\"center\">+/-</td><td align=\"center\">+/-</td><td align=\"center\">+/-</td></tr><tr><td align=\"center\"><bold>5</bold></td><td align=\"center\"><bold>16</bold></td><td align=\"center\">+/+++</td><td align=\"center\">+/+++</td><td align=\"center\">+/+++</td><td align=\"center\">+/+++</td><td align=\"center\">+/+++</td><td align=\"center\">+/+++</td><td align=\"center\">+/-</td><td align=\"center\">+/-</td><td align=\"center\">+/-</td><td align=\"center\">+/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td></tr><tr><td/><td align=\"center\"><bold>17</bold></td><td align=\"center\">+/+++</td><td align=\"center\">+/+++</td><td align=\"center\">+/+++</td><td align=\"center\">+/+++</td><td align=\"center\">+/+++</td><td align=\"center\">+/+++</td><td align=\"center\">+/+</td><td align=\"center\">+/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">+/-</td><td align=\"center\">+/-</td></tr><tr><td/><td align=\"center\"><bold>18 !</bold></td><td align=\"center\">+/n</td><td align=\"center\">+/n</td><td align=\"center\">+/n</td><td align=\"center\">+/n</td><td align=\"center\">+/n</td><td align=\"center\">+/n</td><td align=\"center\">+/n</td><td align=\"center\">-/n</td><td align=\"center\">+/n</td><td align=\"center\">+/n</td><td align=\"center\">-/n</td><td align=\"center\">-/n</td></tr><tr><td/><td align=\"center\"><bold>19 !</bold></td><td align=\"center\">+/n</td><td align=\"center\">+/n</td><td align=\"center\">+/n</td><td align=\"center\">+/n</td><td align=\"center\">+/n</td><td align=\"center\">+/n</td><td align=\"center\">+/n</td><td align=\"center\">+/n</td><td align=\"center\">-/n</td><td align=\"center\">-/n</td><td align=\"center\">+/n</td><td align=\"center\">+/n</td></tr><tr><td align=\"center\"><bold>6</bold></td><td align=\"center\"><bold>20 !</bold></td><td align=\"center\">+/n</td><td align=\"center\">+/n</td><td align=\"center\">+/n</td><td align=\"center\">+/n</td><td align=\"center\">+/n</td><td align=\"center\">+/n</td><td align=\"center\">+/n</td><td align=\"center\">+/n</td><td align=\"center\">+/n</td><td align=\"center\">+/n</td><td align=\"center\">-/n</td><td align=\"center\">-/n</td></tr><tr><td/><td align=\"center\"><bold>21 !</bold></td><td align=\"center\">+/n</td><td align=\"center\">+/n</td><td align=\"center\">+/n</td><td align=\"center\">+/n</td><td align=\"center\">+/n</td><td align=\"center\">+/n</td><td align=\"center\">+/n</td><td align=\"center\">-/n</td><td align=\"center\">-/n</td><td align=\"center\">-/n</td><td align=\"center\">-/n</td><td align=\"center\">-/n</td></tr><tr><td/><td align=\"center\"><bold>22 !</bold></td><td align=\"center\">+/n</td><td align=\"center\">+/n</td><td align=\"center\">+/n</td><td align=\"center\">+/n</td><td align=\"center\">+/n</td><td align=\"center\">+/n</td><td align=\"center\">+/n</td><td align=\"center\">+/n</td><td align=\"center\">+/n</td><td align=\"center\">+/n</td><td align=\"center\">+/n</td><td align=\"center\">+/n</td></tr><tr><td/><td align=\"center\"><bold>23 !</bold></td><td align=\"center\">+/n</td><td align=\"center\">+/n</td><td align=\"center\">+/n</td><td align=\"center\">+/n</td><td align=\"center\">+/n</td><td align=\"center\">+/n</td><td align=\"center\">+/n</td><td align=\"center\">+/n</td><td align=\"center\">+/n</td><td align=\"center\">+/n</td><td align=\"center\">+/n</td><td align=\"center\">+/n</td></tr><tr><td align=\"center\"><bold>7</bold></td><td align=\"center\"><bold>24 !</bold></td><td align=\"center\">+/n</td><td align=\"center\">+/n</td><td align=\"center\">+/n</td><td align=\"center\">+/n</td><td align=\"center\">+/n</td><td align=\"center\">+/n</td><td align=\"center\">+/n</td><td align=\"center\">+/n</td><td align=\"center\">+/n</td><td align=\"center\">-/n</td><td align=\"center\">-/n</td><td align=\"center\">-/n</td></tr><tr><td align=\"center\"><bold>8</bold></td><td align=\"center\"><bold>25</bold></td><td align=\"center\">-/+++</td><td align=\"center\">+/+++</td><td align=\"center\">+/+</td><td align=\"center\">+/+++</td><td align=\"center\">+/+++</td><td align=\"center\">+/+++</td><td align=\"center\">+/+++</td><td align=\"center\">+/-</td><td align=\"center\">+/+</td><td align=\"center\">+/++</td><td align=\"center\">-/-</td><td align=\"center\">+/-</td></tr></tbody></table></table-wrap>", "<table-wrap position=\"float\" id=\"T2\"><label>Table 2</label><caption><p>Results of Experiment II; pigs, infected with the moderately virulent isolate, Wingene'93</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td align=\"center\"><bold>PID</bold></td><td align=\"center\"><bold>Pig No.</bold></td><td align=\"center\"><bold>Tonsils</bold></td><td align=\"center\"><bold>Spleen</bold></td><td align=\"center\"><bold>Kidneys</bold></td><td align=\"center\"><bold>Ln. 1</bold></td><td align=\"center\"><bold>Ln. 2</bold></td><td align=\"center\"><bold>Ln. 3</bold></td><td align=\"center\"><bold>Lungs</bold></td><td align=\"center\"><bold>Heart</bold></td><td align=\"center\"><bold>Cerebr.</bold></td><td align=\"center\"><bold>Cerebel.</bold></td><td align=\"center\"><bold>Musc. 1</bold></td><td align=\"center\"><bold>Musc. 2</bold></td></tr></thead><tbody><tr><td align=\"center\"><bold>0</bold></td><td align=\"center\"><bold>1 (25)</bold></td><td align=\"center\">*-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td></tr><tr><td align=\"center\"><bold>1</bold></td><td align=\"center\"><bold>2 (1)</bold></td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td></tr><tr><td/><td align=\"center\"><bold>3 (2)</bold></td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td></tr><tr><td align=\"center\"><bold>2</bold></td><td align=\"center\"><bold>4 (3)</bold></td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td></tr><tr><td/><td align=\"center\"><bold>5 (4)</bold></td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td></tr><tr><td align=\"center\"><bold>3</bold></td><td align=\"center\"><bold>6 (5)</bold></td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td></tr><tr><td/><td align=\"center\"><bold>7 (6)</bold></td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td></tr><tr><td align=\"center\"><bold>4</bold></td><td align=\"center\"><bold>8 (7)</bold></td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td></tr><tr><td/><td align=\"center\"><bold>9 (9)</bold></td><td align=\"center\">+/+</td><td align=\"center\">-/+</td><td align=\"center\">-/-</td><td align=\"center\">+/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/+</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td></tr><tr><td align=\"center\"><bold>5</bold></td><td align=\"center\"><bold>10 (8)</bold></td><td align=\"center\">+/+++</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">+/-</td><td align=\"center\">+/-</td><td align=\"center\">+/++</td><td align=\"center\">-/+</td><td align=\"center\">-/-</td><td align=\"center\">+/-</td><td align=\"center\">+/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td></tr><tr><td/><td align=\"center\"><bold>11 (10)</bold></td><td align=\"center\">+/+-++</td><td align=\"center\">-/+++</td><td align=\"center\">-/-</td><td align=\"center\">+/++</td><td align=\"center\">+/+</td><td align=\"center\">+/+++</td><td align=\"center\">-/+</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td></tr><tr><td align=\"center\"><bold>6</bold></td><td align=\"center\"><bold>12 (11)</bold></td><td align=\"center\">+/+</td><td align=\"center\">-/+++</td><td align=\"center\">-/-</td><td align=\"center\">+/+++</td><td align=\"center\">+/-</td><td align=\"center\">+/+++</td><td align=\"center\">+/+</td><td align=\"center\">+/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td></tr><tr><td/><td align=\"center\"><bold>13 (12)</bold></td><td align=\"center\">+/+++</td><td align=\"center\">+/+</td><td align=\"center\">-/-</td><td align=\"center\">+/-</td><td align=\"center\">+/++</td><td align=\"center\">+/+++</td><td align=\"center\">+/+</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td></tr><tr><td align=\"center\"><bold>7</bold></td><td align=\"center\"><bold>14 (13)</bold></td><td align=\"center\">+/+++</td><td align=\"center\">+/+++</td><td align=\"center\">+/-</td><td align=\"center\">+/+++</td><td align=\"center\">+/+</td><td align=\"center\">+/+++</td><td align=\"center\">+/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td></tr><tr><td/><td align=\"center\"><bold>15 (14)</bold></td><td align=\"center\">+/+++</td><td align=\"center\">+/++</td><td align=\"center\">+/-</td><td align=\"center\">+/+</td><td align=\"center\">+/+</td><td align=\"center\">+/+++</td><td align=\"center\">+/+</td><td align=\"center\">-/-</td><td align=\"center\">-/nc</td><td align=\"center\">-/nc</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td></tr><tr><td align=\"center\"><bold>8</bold></td><td align=\"center\"><bold>16 (15)</bold></td><td align=\"center\">+/++</td><td align=\"center\">+/+++</td><td align=\"center\">+/-</td><td align=\"center\">+/+</td><td align=\"center\">+/+</td><td align=\"center\">-/+++</td><td align=\"center\">+/-</td><td align=\"center\">+/-</td><td align=\"center\">-/nc</td><td align=\"center\">-/nc</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td></tr><tr><td/><td align=\"center\"><bold>17 (16)</bold></td><td align=\"center\">+/+++</td><td align=\"center\">+/+++</td><td align=\"center\">+/-</td><td align=\"center\">+/+++</td><td align=\"center\">+/++</td><td align=\"center\">+/+++</td><td align=\"center\">+/++</td><td align=\"center\">+/-</td><td align=\"center\">-/nc</td><td align=\"center\">+/nc</td><td align=\"center\">+/-</td><td align=\"center\">+/-</td></tr><tr><td/><td align=\"center\"><bold>18 (17)</bold></td><td align=\"center\">+/+++</td><td align=\"center\">+/+++</td><td align=\"center\">+/-</td><td align=\"center\">+/+++</td><td align=\"center\">+/+++</td><td align=\"center\">+/+++</td><td align=\"center\">+/+++</td><td align=\"center\">+/-</td><td align=\"center\">+/-</td><td align=\"center\">+/-</td><td align=\"center\">+/-</td><td align=\"center\">-/-</td></tr><tr><td align=\"center\"><bold>10</bold></td><td align=\"center\"><bold>19 (18)</bold></td><td align=\"center\">+/++</td><td align=\"center\">+/+++</td><td align=\"center\">+/+</td><td align=\"center\">+/+</td><td align=\"center\">+/+++</td><td align=\"center\">+/+</td><td align=\"center\">+/++</td><td align=\"center\">+/-</td><td align=\"center\">+/-</td><td align=\"center\">+/-</td><td align=\"center\">+/-</td><td align=\"center\">-/-</td></tr><tr><td/><td align=\"center\"><bold>20 (20)!</bold></td><td align=\"center\">+/+++</td><td align=\"center\">+/+</td><td align=\"center\">+/+</td><td align=\"center\">+/+</td><td align=\"center\">+/+</td><td align=\"center\">+/+++</td><td align=\"center\">+/+++</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">+/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td></tr><tr><td align=\"center\"><bold>12</bold></td><td align=\"center\"><bold>21 (19)</bold></td><td align=\"center\">+/++</td><td align=\"center\">+/++</td><td align=\"center\">+/-</td><td align=\"center\">+/+</td><td align=\"center\">+/++</td><td align=\"center\">+/+</td><td align=\"center\">+/-</td><td align=\"center\">+/-</td><td align=\"center\">+/-</td><td align=\"center\">+/-</td><td align=\"center\">+/-</td><td align=\"center\">+/-</td></tr><tr><td/><td align=\"center\"><bold>22 (21)</bold></td><td align=\"center\">+/+++</td><td align=\"center\">-/++</td><td align=\"center\">+/-</td><td align=\"center\">+/-</td><td align=\"center\">+/+</td><td align=\"center\">+/+</td><td align=\"center\">+/++</td><td align=\"center\">+/-</td><td align=\"center\">+/-</td><td align=\"center\">-/-</td><td align=\"center\">+/-</td><td align=\"center\">+/-</td></tr><tr><td align=\"center\"><bold>14</bold></td><td align=\"center\"><bold>23 (22)</bold></td><td align=\"center\">+/++</td><td align=\"center\">-/+</td><td align=\"center\">+/-</td><td align=\"center\">+/+</td><td align=\"center\">+/+</td><td align=\"center\">+/+</td><td align=\"center\">+/-</td><td align=\"center\">-/-</td><td align=\"center\">+/-</td><td align=\"center\">-/-</td><td align=\"center\">+/-</td><td align=\"center\">+/-</td></tr><tr><td/><td align=\"center\"><bold>24 (23)!</bold></td><td align=\"center\">+/-</td><td align=\"center\">+/++</td><td align=\"center\">+/+++</td><td align=\"center\">+/+++</td><td align=\"center\">+/+++</td><td align=\"center\">+/+++</td><td align=\"center\">+/+++</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">-/-</td><td align=\"center\">+/-</td><td align=\"center\">-/-</td></tr><tr><td/><td align=\"center\"><bold>25 (24)!</bold></td><td align=\"center\">+/+++</td><td align=\"center\">+/++</td><td align=\"center\">+/+</td><td align=\"center\">+/++</td><td align=\"center\">+/-</td><td align=\"center\">+/++</td><td align=\"center\">+/+</td><td align=\"center\">+/-</td><td align=\"center\">+/-</td><td align=\"center\">-/-</td><td align=\"center\">+/-</td><td align=\"center\">+/-</td></tr></tbody></table></table-wrap>" ]

[]

[]

[]

[]

[]

[]

[ "<table-wrap-foot><p>*Results of virus isolation/immunohistochemistry.</p><p>Severity of reactions: as measured by IHC = - = negative, + = 1–3 foci/section, ++ = 4–10 foci/section, +++ &gt; 10 foci/section.</p><p>Ln. 1 = ileocecal lymph node</p><p>Ln. 2 = mesenteric lymph node</p><p>Ln. 3 = submandibular lymph node</p><p>Musc. 1 = M. longissimus dorsi</p><p>Musc. 2 = M. quadriceps</p><p>PID = post infection day.</p><p>! = found dead.</p><p>n = not available.</p><p>h = hours</p></table-wrap-foot>", "<table-wrap-foot><p>*Results of virus isolation/immunohistochemistry.</p><p>Severity of reactions: as measured by IHC = - = negative, + = 1–3 foci/section, ++ = 4–10 foci/section, +++ &gt; 10 foci/section.</p><p>Ln. 1 = ileocecal lymph node</p><p>Ln. 2 = mesenteric lymph node</p><p>Ln. 3 = submandibular lymph node</p><p>Musc. 1 = M. longissimus dorsi</p><p>Musc. 2 = M. quadriceps</p><p>PID = post infection day.</p><p>! = found dead.</p><p>nc = not conclusive</p></table-wrap-foot>" ]

[ "<graphic xlink:href=\"1751-0147-50-34-1\"/>", "<graphic xlink:href=\"1751-0147-50-34-2\"/>", "<graphic xlink:href=\"1751-0147-50-34-3\"/>", "<graphic xlink:href=\"1751-0147-50-34-4\"/>", "<graphic xlink:href=\"1751-0147-50-34-5\"/>", "<graphic xlink:href=\"1751-0147-50-34-6\"/>", "<graphic xlink:href=\"1751-0147-50-34-7\"/>", "<graphic xlink:href=\"1751-0147-50-34-8\"/>" ]

[]

{ "acronym": [], "definition": [] }

[ "<title>Background</title>", "<p>Adenocarcinoma accounts for the majority of pancreatic malignancies. Adenosquamous carcinoma (ASC) of the pancreas is an unusual variant of pancreatic neoplasm [##REF##10367867##1##, ####REF##11353055##2##, ##REF##4355621##3##, ##REF##16172546##4####16172546##4##], and is characteristic by histological patterns of both ductal adenocarcinoma and squamous carcinoma within the same tumor. The prognosis of this rare tumor appears to be even less favorable than the common invasive ductal tumor with few patients surviving more than 1 year after surgical resection [##REF##16172546##4##]. Most of studies on this disease have been small series or single case reports, and few studies have investigated the clinicopathologic features and outcome of patients with pancreatic ASC following surgical treatment [##REF##10367867##1##,##REF##11353055##2##,##REF##1372374##5##,##REF##2062081##6##]. Therefore, medical records of 12 patients with pancreatic ASC treated surgically at Chang Gung Memorial Hospital (CGMH), Taoyuan in the past 14 years were retrospectively reviewed.</p>" ]

[ "<title>Methods</title>", "<p>A total of 637 patients with pancreatic malignancies underwent surgical treatment at CGMH between January 1993 and December 2006. Adenocarcinoma was diagnosed in 530 patients and ASC in 12. Institutional Review Board approval was obtained and medical records of 12 patients with pancreatic ASC were retrospectively reviewed. Preoperative imaging studies employed abdominal ultrasonography, abdominal computed tomography (CT)/magnetic resonance imaging (MRI), and endoscopic retrograde cholangiopancreatography (ERCP). Serum tumor markers such as carcinoembryonic antigen (CEA) and carbohydrate antigen 19-9 (CA 19-9) were measured preoperatively. One patient had a preoperative fine needle tumor biopsy. Intraoperative radiotherapy and postoperative chemotherapy were performed in 2 patients and 7 patients, respectively. Tumor stage and TMN stage were defined according to the sixth edition of American Joint Committee on Cancer for pancreatic carcinoma [##UREF##0##7##] based on the histopathologic examination of surgical specimens and clinical findings such as imaging studies and intraoperative records. To further elucidate the outcome following surgical resection for pancreatic ASC from more common pancreatic adenocarcinoma, patients with stage IIB pancreatic adenocarcinoma undergoing surgical resection were also extracted from our databank (at the same studying period as pancreatic ASC). Survival data of patients with stage IIB pancreatic adenocarcinoma and ASC undergoing surgical resection were compared. Survival rate was calculated and graphs plotted using Kaplan-Meier method. Differences in survival curves between the groups were compared by the log-rank test. A <italic>p</italic>-value less than 0.05 were defined as statistically significant. All statistical analyses were performed with SPSS for Windows, version 11.5 (Statistical Package for the Social Science, SPSS, Inc., Chicago, Illinois).</p>" ]

[ "<title>Results</title>", "<p>The demographic features of 12 patients with pancreatic ASC including 5 men and 7 women (age range, 32 to 79 years; median, 71 years) are shown in Table ##TAB##0##1##. Symptoms were abdominal pain in 11 patients (91.7%), body weight loss in 10 (83.3%), anorexia in 5 (41.7%), jaundice in 3 (25.0%). Ten patients had 13 comorbidities including hypertension in 5, diabetes mellitus in 4 and peptic ulcer in 3, and heart disease in 1. Laboratory studies revealed anemia in 9 patients (75.0%), elevated total bilirubin levels in 3 (25.0%), and elevated alkaline phosphatase levels in 3 (25.0%). Elevated serum CEA levels and CA 19-9 levels were identified in 10 patients (83.3%), respectively. Three patients underwent ERCP, which identified tumor obstruction of the pancreatic head duct. All patients underwent abdominal CT or MRI, which accurately determined and localized a pancreatic tumor.</p>", "<p>Table ##TAB##1##2## demonstrates the details of tumor characteristics, management and prognosis of 12 patients with pancreatic ASC. The tumors were located at pancreatic head in 5 (41.7%) patients, tail in 5 (41.7%), and body in 4 (33.3%). Tumor size ranged from 3.5 to 8 cm with a median of 6.3 cm. The lesions from the resected specimens were firm with light tan to yellowish colors and had merged imperceptibly with the surrounding pancreatic parenchyma. Histologically, tumors were a mixture component of adenocarcinoma and squamous cell carcinoma (Figure ##FIG##0##1##). The rates of squamous component in the tumor tissue ranged from 40 to 90% in patients undergoing surgical resection. Lymph node metastases were identified in 11 patients (91.7%). Encasement of superior mesenteric artery by the tumor was found during operation in 5 patients, and carcinomatosis in 1 patient. Surgical resection including pancreaticoduodenectomy (PD) and subtotal or distal pacreatectomy along with total gastrectomy or splenectomy was performed in 7 patients. R0 (radical) resection was identified in 5 patients and R1 resection in 2 (cases 9 and 11). Five patients underwent laparotomy followed by intra-operative biopsy of the pancreatic tumor and three received bypass surgery. Intraoperative irradiation and postoperative chemotherapy were carried out in 2 and 7 patients, respectively. Tumor stage was IIB in 7 patients, III in 4 and IV in 1.</p>", "<p>There was no surgical mortality. The time of follow-up ranged from 0.79 to 122.66 months with a median of 6.49 months. Figure ##FIG##1##2## shows the cumulative survival rates of 12 patients with pancreatic ASC with a median of 4.41 months, ranging from 1.12 to 22.42 months. Eleven of 12 patients with pancreatic ASC died in one year after surgery with one-year survival rate of 8.3% (95% confidence interval, 0.0–24). Figure ##FIG##2##3## demonstrates cumulative survival rates of stage IIB pancreatic adenocarcinoma (n = 101) and ASC (n = 7) patients undergoing surgical resection. Patients with pancreatic ASC had shorter median survival compared to those with adenocarcinoma (6.51 months vs. 9.76 months, <italic>p </italic>= 0.018).</p>" ]

[ "<title>Discussion</title>", "<p>The first report of ASC is credited to Herxheimer in 1907 [##UREF##1##8##]. This admixed tumor has been seen more commonly in other organ systems where adenocarcinomas are generally found, such as the stomach [##REF##3556988##9##], intestine [##REF##5472856##10##] and uterus [##REF##14082276##11##]. It has also been identified in the esophagus [##REF##5686635##12##], anus [##REF##16811005##13##] and vagina [##REF##14155003##14##] where squamous cell carcinomas predominate. In the present studies, the incidence of pancreatic ASC was 1.9% (12/637), within the range of 0.9 to 3.8% reported in the literatures [##REF##11353055##2##, ####REF##4355621##3##, ##REF##16172546##4####16172546##4##]. The histogenesis of pancreatic ASC remains unclear. There are numerous possibilities that account for the presence of a squamous element where adenocarcinoma is expected. Four theories regarding the histogenesis of adenosquamous carcinoma may be summarized as follows: adenocarcinoma transforming into squamous cell carcinoma; bipotential undifferentiated cell origin; collision tumor; and squamous metaplasia origin [##REF##1372374##5##].</p>", "<p>Madura et al. [##REF##10367867##1##] reported that most patients with pancreatic ASC are males in their 60s and frequently located at the head of the pancreas. Different from their findings, more females were identified in our patients, and the patient median age was 71 years. Moreover, our results show that the tumor location was evenly distributed at the pancreatic head, body, or tail. Symptoms of our patients with pancreatic ASC were abdominal pain (92%), body weight loss (83%), anorexia (42%) and jaundice (25%) similar to those of pancreatic adenocarcinoma [##REF##9672355##15##].</p>", "<p>Accurate preoperative diagnosis of pancreatic ASC is made with great difficulty since there are no investigations of its defining characteristics in imaging studies that would differentiate it from the more common pancreatic exocrine neoplasm [##REF##10367867##1##]. Nevertheless, studies have indicated that cytological examination of pure pancreatic juice obtained by endoscopic retrograde pancreatic juice aspiration is a useful modality for the preoperative diagnosis [##REF##14714256##16##]. Rahemtullah et al. [##REF##14681946##17##] also reported that cytological features derived from fine-aspiration biopsy are diagnostic of pancreatic ASC. Furthermore, imaging studies by Nabae et al. [##REF##9672355##15##] showed that the presence of central necrosis in a huge infiltrative pancreatic tumor is suggestive of the diagnosis of ASC. Moreover, a tumor might selectively take up gallium 67 and be visualized by nuclear scanning which is useful in detecting this rare pancreatic tumor [##REF##9095322##18##]. In the present studies, no patient had central necrosis at the pancreatic tumor on abdominal imaging studies indicating a diagnosis of pancreatic ASC. Besides, the preoperative fine needle biopsy of the tumor was performed in 1 patient, which revealed adenocarcinoma.</p>", "<p>As shown in table ##TAB##1##2##, 11 patients with pancreatic ASC (92%) died within 12 months despite aggressive surgical management along with intraoperative irradiation or postoperative chemotherapy. The median cumulative survival of 12 patients was 4.92 months (Figure ##FIG##1##2##). Furthermore, median survival of 7 patients undergoing surgical resection was 6.51 months. These results were similar to that obtained by Madura et al. [##REF##10367867##1##], who reported that 72 patients survived with an average age of 5.7 months, regardless of whether or not surgical resection was performed. To our surprise, 1 patient in our series who had no lymph node involvement without undergoing surgical resection and received intraoperative irradiation had a survival of 22.42 months. No lymph node metastasis and the potential benefit of intraoperative radiation therapy might explain his long survival.</p>", "<p>Once pancreatic ASC is identified either preoperatively or intraoperatively, the choice of treatment becomes a complex decision as survival is typically dismal [##REF##10367867##1##]. In this regard, although PD has been shown to be performed with a very low mortality rate (&lt;4%) in specialized high-volume centers, the incidence of postoperative morbidity can be as high as 30% to 40% [##REF##9339931##19##,##REF##18221566##20##]. Furthermore, a significant high mortality rate (25%) has been reported in the patient subgroup with significant preoperative comorbidities [##REF##18221566##20##]. Thus, anesthesia risks and complications following major surgery in pancreatic ASC patients along with severe medical diseases should be considered before operation. Moreover, we observed that median survival of patients (stage IIB) with pancreatic ASC undergoing surgical resection was 6.51 months, significantly shorter (<italic>p </italic>= 0.018) than patients with stage IIB pancreatic adenocarcinoma receiving resection (median survival, 9.76 months; Figure ##FIG##2##3##), suggesting more aggressive biology of pancreatic ASC than adenocarcinoma. Moreover, nodal metastases were identified in 92% (11/12) of our patients, which might reflect the disease entity tending to have lymph node involvement and at least partly explained the poor prognosis of this virulent tumor.</p>", "<p>It should be noted that this study was based on a retrospective review of patients undergoing surgery. Pancreatic malignancy patients who were not diagnosed as pancreatic ASC without tissue proof treated non-surgically were not enrolled in this study. Whether surgical resection or non-surgical management such as chemotherapy, radiotherapy, chemo-radiotherapy or target therapy would provide survival benefits to patients with pancreatic ASC remains unknown. More studies are necessary to confirm this.</p>" ]

[ "<title>Conclusion</title>", "<p>Pancreatic ASC is a rare pancreatic neoplasm subtype. Abdominal pain and body weight loss are the two predominant symptoms. Distribution of ASC is even in the pancreas, and the tumor size is big at the time of diagnosis. Pancreatic ASC tends to have nodal metastases and has a dismal outcome despite surgical resection. In this limited case study, aggressive surgical management does not appear effective in treating pancreatic ASC patients. Strategies involving non-surgical treatment such as chemotherapy, radiotherapy or target agents should be tested.</p>" ]

[ "<p>This is an Open Access article distributed under the terms of the Creative Commons Attribution License (<ext-link ext-link-type=\"uri\" xlink:href=\"http://creativecommons.org/licenses/by/2.0\"/>), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.</p>", "<title>Background</title>", "<p>Pancreatic adenosquamous carcinoma (ASC) is a rare pancreatic malignancy subtype. We investigated the clinicopathological features and outcome of pancreatic ASC patients after surgery.</p>", "<title>Methods</title>", "<p>The medical records of 12 patients with pancreatic ASC undergoing surgical treatment (1993 to 2006) were retrospectively reviewed. Survival data of patients with stage IIB pancreatic adenocarcinoma and ASC undergoing surgical resection were compared.</p>", "<title>Results</title>", "<p>Symptoms included abdominal pain (91.7%), body weight loss (83.3%), anorexia (41.7%) and jaundice (25.0%). Tumors were located at pancreatic head in 5 (41.7%) patients, tail in 5 (41.7%), and body in 4 (33.3%). Median tumor size was 6.3 cm. Surgical resection was performed on 7 patients, bypass surgery on 3, and exploratory laparotomy with biopsy on 2. No surgical mortality was identified. Seven (58.3%) and 11 (91.7%) patients died within 6 and 12 months of operation, respectively. Median survival of 12 patients was 4.41 months. Seven patients receiving surgical resection had median survival of 6.51 months. Patients with stage IIB pancreatic ASC had shorter median survival compared to those with adenocarcinoma.</p>", "<title>Conclusion</title>", "<p>Aggressive surgical management does not appear effective in treating pancreatic ASC patients. Strategies involving non-surgical treatment such as chemotherapy, radiotherapy or target agents should be tested.</p>" ]

[ "<title>Competing interests</title>", "<p>The authors declare that they have no competing interests.</p>", "<title>Authors' contributions</title>", "<p>HJT: planning, study design, data collection and analysis, drafting, and revising the manuscript. CHM: planning, study design and analysis, surgical management of patients. WRC: pathological review of surgical specimens, preparing the pathological figure. YCN: study design and analysis, surgical management of patients, YTS: study design and analysis, surgical management of patients. HTL: study design and analysis, surgical management of patients, revising the manuscript. JYY: study design and analysis, surgical management of patients. CMF: study design and analysis, surgical management of patients. All authors read and approved final manuscript.</p>" ]

[ "<title>Acknowledgements</title>", "<p>We thank Yi-Hua Liu for assistance with data collection and Shu-Fang Huang for superb help with data analysis and preparing the tables and figures.</p>", "<p>The written consent was obtained from the patients' Family for publication of this study and IRB approval was obtained for collecting the data.</p>" ]

[ "<fig position=\"float\" id=\"F1\"><label>Figure 1</label><caption><p>Histopathology in a patient with pancreatic tumor shows glandular adenocarcinoma foci (black arrowheads) and nests of squamous cell carcinoma (upper middle part), consistent with adenosquamous carcinoma (Hematoxylin-Eosin stain, original magnification ×100).</p></caption></fig>", "<fig position=\"float\" id=\"F2\"><label>Figure 2</label><caption><p>Cumulative survival rates of 12 patients with pancreatic adenosquamous carcinoma after surgery.</p></caption></fig>", "<fig position=\"float\" id=\"F3\"><label>Figure 3</label><caption><p>Cumulative survival rates of patients with stage IIB pancreatic adenocarcinoma and adenosquamous carcinoma undergoing surgical resection.</p></caption></fig>" ]

[ "<table-wrap position=\"float\" id=\"T1\"><label>Table 1</label><caption><p>Demographics of 12 patients with pancreatic adenosquamous carcinoma.</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td align=\"center\">Case</td><td align=\"center\">Age/sex</td><td align=\"center\">Symptoms</td><td align=\"center\">Comorbidity</td><td align=\"center\" colspan=\"3\">Laboratory data</td><td align=\"center\" colspan=\"2\">Tumor markers</td></tr><tr><td/><td/><td/><td/><td colspan=\"5\"><hr/></td></tr><tr><td/><td/><td/><td/><td align=\"center\">Hb<break/>(g/dL)</td><td align=\"center\">Bil (T)<break/>(mg/dL)</td><td align=\"center\">Alk-P<break/>(U/L)</td><td align=\"center\">CEA (ng/mL)</td><td align=\"center\">CA 19-9 (U/mL)</td></tr></thead><tbody><tr><td align=\"center\">1</td><td align=\"center\">73/F</td><td align=\"center\">Abd pain, BWL, jaundice</td><td align=\"center\">HTN</td><td align=\"center\">10.4</td><td align=\"center\">15.7</td><td align=\"center\">258</td><td align=\"center\">262</td><td align=\"center\">240</td></tr><tr><td align=\"center\">2</td><td align=\"center\">66/M</td><td align=\"center\">BWL, jaundice, anorexia</td><td align=\"center\">DM</td><td align=\"center\">10.2</td><td align=\"center\">14.8</td><td align=\"center\">438</td><td align=\"center\">13.1</td><td align=\"center\">&gt;240</td></tr><tr><td align=\"center\">3</td><td align=\"center\">65/F</td><td align=\"center\">Abd pain, BWL, diarrhea</td><td align=\"center\">HTN, heart disease</td><td align=\"center\">13.4</td><td align=\"center\">0.4</td><td align=\"center\">89</td><td align=\"center\">25.2</td><td align=\"center\">15.7</td></tr><tr><td align=\"center\">4</td><td align=\"center\">63/M</td><td align=\"center\">Abd pain, BWL</td><td align=\"center\">Peptic ulcer</td><td align=\"center\">13.2</td><td align=\"center\">0.5</td><td align=\"center\">92</td><td align=\"center\">10.2</td><td align=\"center\">135</td></tr><tr><td align=\"center\">5</td><td align=\"center\">78/M</td><td align=\"center\">Abd pain, anorexia</td><td align=\"center\">Nil</td><td align=\"center\">10.6</td><td align=\"center\">0.7</td><td align=\"center\">78</td><td align=\"center\">12.4</td><td align=\"center\">142.4</td></tr><tr><td align=\"center\">6</td><td align=\"center\">79/M</td><td align=\"center\">Abd pain, BWL, anorexia, abd mass</td><td align=\"center\">Nil</td><td align=\"center\">12.6</td><td align=\"center\">0.5</td><td align=\"center\">95</td><td align=\"center\">17.6</td><td align=\"center\">8.5</td></tr><tr><td align=\"center\">7</td><td align=\"center\">38/F</td><td align=\"center\">Abd pain, BWL, jaundice</td><td align=\"center\">DM</td><td align=\"center\">10.7</td><td align=\"center\">23.2</td><td align=\"center\">192</td><td align=\"center\">14.5</td><td align=\"center\">138</td></tr><tr><td align=\"center\">8</td><td align=\"center\">79/F</td><td align=\"center\">Abd pain, dizziness, malaise</td><td align=\"center\">HTN, DM, Peptic ulcer</td><td align=\"center\">7.0</td><td align=\"center\">0.8</td><td align=\"center\">59</td><td align=\"center\">2300</td><td align=\"center\">&gt;240</td></tr><tr><td align=\"center\">9</td><td align=\"center\">76/F</td><td align=\"center\">Abd pain, BWL</td><td align=\"center\">HTN</td><td align=\"center\">11.3</td><td align=\"center\">0.5</td><td align=\"center\">93</td><td align=\"center\">5.34</td><td align=\"center\">129</td></tr><tr><td align=\"center\">10</td><td align=\"center\">32/M</td><td align=\"center\">Abd pain, BWL</td><td align=\"center\">Peptic ulcer</td><td align=\"center\">15.2</td><td align=\"center\">0.5</td><td align=\"center\">84</td><td align=\"center\">1.79</td><td align=\"center\">84</td></tr><tr><td align=\"center\">11</td><td align=\"center\">69/F</td><td align=\"center\">Abd pain, BWL, anorexia</td><td align=\"center\">DM</td><td align=\"center\">9.4</td><td align=\"center\">0.5</td><td align=\"center\">66</td><td align=\"center\">83.74</td><td align=\"center\">&gt;240</td></tr><tr><td align=\"center\">12</td><td align=\"center\">78/F</td><td align=\"center\">Abd pain, BWL, anorexia, malaise</td><td align=\"center\">HTN</td><td align=\"center\">10.6</td><td align=\"center\">0.3</td><td align=\"center\">68</td><td align=\"center\">0.57</td><td align=\"center\">160.9</td></tr></tbody></table></table-wrap>", "<table-wrap position=\"float\" id=\"T2\"><label>Table 2</label><caption><p>Details of tumor characteristics, management, and prognosis of 12 patients with pancreatic adenosquamous carcinoma.</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td align=\"left\">Case</td><td align=\"center\">Tumor location</td><td align=\"center\">Size (cm)</td><td align=\"center\">Operative method</td><td align=\"center\">Intraoperative irradiation</td><td align=\"center\">Postoperative chemotherapy</td><td align=\"center\">Stage* (TNM)</td><td align=\"center\">Survival (months)</td></tr></thead><tbody><tr><td align=\"left\">1</td><td align=\"center\">Head</td><td align=\"center\">6</td><td align=\"center\">Biopsy, bypass</td><td align=\"center\">ND</td><td align=\"center\">ND</td><td align=\"center\">III (T4N1M0)</td><td align=\"center\">4.04</td></tr><tr><td align=\"left\">2</td><td align=\"center\">Head</td><td align=\"center\">3.5</td><td align=\"center\">PD</td><td align=\"center\">ND</td><td align=\"center\">ND</td><td align=\"center\">IIB (T2N1M0)</td><td align=\"center\">2.50</td></tr><tr><td align=\"left\">3</td><td align=\"center\">Body and tail</td><td align=\"center\">8</td><td align=\"center\">Biopsy, bypass</td><td align=\"center\">ND</td><td align=\"center\">ND</td><td align=\"center\">IV (T4N1M1)</td><td align=\"center\">1.12</td></tr><tr><td align=\"left\">4</td><td align=\"center\">Head</td><td align=\"center\">6</td><td align=\"center\">Biospy</td><td align=\"center\">1,800 cGy</td><td align=\"center\">ND</td><td align=\"center\">III (T4N0M0)</td><td align=\"center\">22.42</td></tr><tr><td align=\"left\">5</td><td align=\"center\">Tail</td><td align=\"center\">8</td><td align=\"center\">Biopsy</td><td align=\"center\">2,000 cGy</td><td align=\"center\">Gemcitabine</td><td align=\"center\">III (T4N1M0)</td><td align=\"center\">5.42</td></tr><tr><td align=\"left\">6</td><td align=\"center\">Body</td><td align=\"center\">8</td><td align=\"center\">Biopsy, bypass</td><td align=\"center\">ND</td><td align=\"center\">Tegafur</td><td align=\"center\">III (T4N1M0)</td><td align=\"center\">4.41</td></tr><tr><td align=\"left\">7</td><td align=\"center\">Head</td><td align=\"center\">3.8</td><td align=\"center\">PD</td><td align=\"center\">ND</td><td align=\"center\">Gemcitabine, Fluorouracil</td><td align=\"center\">IIB (T2N1M0)</td><td align=\"center\">6.84</td></tr><tr><td align=\"left\">8</td><td align=\"center\">Head</td><td align=\"center\">5.5</td><td align=\"center\">PD</td><td align=\"center\">ND</td><td align=\"center\">Gemcitabine</td><td align=\"center\">IIB (T3N1M0)</td><td align=\"center\">6.51</td></tr><tr><td align=\"left\">9</td><td align=\"center\">Body</td><td align=\"center\">7</td><td align=\"center\">subtotal P, total G, S</td><td align=\"center\">ND</td><td align=\"center\">Tegafur, Uracil</td><td align=\"center\">IIB (T3N1M0)</td><td align=\"center\">11.84</td></tr><tr><td align=\"left\">10</td><td align=\"center\">Tail</td><td align=\"center\">5</td><td align=\"center\">distal P, S</td><td align=\"center\">ND</td><td align=\"center\">ND</td><td align=\"center\">IIB (T2N1M0)</td><td align=\"center\">10.82</td></tr><tr><td align=\"left\">11</td><td align=\"center\">Tail</td><td align=\"center\">8</td><td align=\"center\">distal P, total G, S</td><td align=\"center\">ND</td><td align=\"center\">Gemcitabine, Cisplatin</td><td align=\"center\">IIB (T3N1M0)</td><td align=\"center\">3.68</td></tr><tr><td align=\"left\">12</td><td align=\"center\">Body and Tail</td><td align=\"center\">6.5</td><td align=\"center\">subtotal P, S</td><td align=\"center\">ND</td><td align=\"center\">Gemcitabine</td><td align=\"center\">IIB (T3N1M0)</td><td align=\"center\">4.08</td></tr></tbody></table></table-wrap>" ]

[]

[]

[]

[]

[]

[]

[ "<table-wrap-foot><p>Abd, abdominal; Alk-P, alkaline phosphatase; Bil (T): total bilirubin; BWL, body weight loss; CA19-9, carbohydrate antigen 19-9 (&lt; 37 U/mL); CEA, carcinoembryonic antigen (&lt; 5 ng/mL); DM, diabetes mellitus; Hb, hemoglobin; HTN, hypertension.</p></table-wrap-foot>", "<table-wrap-foot><p>G, gastrectomy; ND, not done; P, pancreatectomy; PD, pancreaticoduodenectomy; S, splenectomy; *, clinical and pathological.</p></table-wrap-foot>" ]

[ "<graphic xlink:href=\"1477-7819-6-95-1\"/>", "<graphic xlink:href=\"1477-7819-6-95-2\"/>", "<graphic xlink:href=\"1477-7819-6-95-3\"/>" ]

[]

{ "acronym": [], "definition": [] }

[ "<title>Introduction</title>", "<p>Advances in the management of colon cancer over the past decades have resulted in an improvement of the prognosis of the disease. The proportion of stage I and II has increased from 39.6% to 56.6% leading to a raise of five-year relative survival from 33% in 1970s to 55.3% in 1990s [##REF##16552821##1##].</p>", "<p>Nevertheless, the five-year survival rate of colon cancer has not improved dramatically in the last decade, remaining at approximately 60%, and colon cancer is still one of the leading killers in the Western countries [##REF##7577025##2##].</p>", "<p>In truth, despite curative resection, many patients develop recurrence at the primary site or distant organs. These high risk patients could be candidates for more aggressive treatments (neoadjuvant chemotherapy) in order to improve the prognosis [##REF##9501824##3##]. This target requires not only the development of new therapeutic modalities but also a reliable preoperative stratification of high and low risk patients.</p>", "<p>Prognostic factors derived from clinical, laboratory and pathologic data of colorectal cancer patients have been considered important and have been investigated in order to make a proper choice between surgery chemotherapy and radiotherapy, but the results of the previous studies were often intriguing and conflicting [##REF##1730115##4##,##REF##3366023##5##].</p>", "<p>Actually, most studies investigating prognostic factors for large bowel cancers did not distinguish between the subpopulation of colon and rectal cancer, despite the different biological characteristics, treatment modalities, pattern of recurrence and survival rates of the two group of neoplasms [##REF##11289282##6##].</p>", "<p>Further, it was suggested that proximal and distal colon cancer can differ in histopathologic characteristics, molecular pattern, stage of diagnosis and, consequently, clinical outcome. Over the past 20 years, the literature has demonstrated a stage migration of colorectal cancer from distal to proximal sites with a tendency for proximal tumours to present at a more advanced stage than distal tumours [##REF##11227943##7##].</p>", "<p>At the moment, the most accurate prognostic factor remains the extension of the tumour into the bowel wall as expressed in the Dukes classification or TNM classification [##REF##7577025##2##,##REF##1730115##4##].</p>", "<p>The main endpoint of the present study was to evaluate the prognostic implication of many preoperative clinical, laboratory and patho-morphological data by both univariate and multivariate analysis.</p>", "<p>Therefore, we performed two statistical analysis of clinical, laboratory and patho-morphological data in a group of patients with colon cancer, considering survival and pT staging as the independent variables.</p>" ]

[ "<title>Materials and methods</title>", "<title>Patients</title>", "<p>A total of 103 patients with colon cancer, who were surgically treated between January 1999 and December 2001 at the Department of General Surgery, University Hospital Tor Vergata, Rome, were evaluated for eligibility.</p>", "<p>Patients who suffered from rectal cancer, colon carcinoma with locally advanced invasion (pT4) or colon cancer with distant metastasis were excluded. Only elective surgery cases were considered. Thus, out of 103 subjects, 92 patients, who underwent to curative resection and were followed for at least 5 years, were analysed.</p>", "<p>Preoperative staging was performed using colonoscopy, conventional transabdominal ultrasonography, CT scan of abdomen, barium enema, chest X-ray and blood tests that included tumour markers.</p>", "<p>CT scanning was performed using oral and intravenous contrast. Patients were scanned at 5-mm intervals from the diaphragm through the pubic symphysis. We did not use a three-dimensional endoluminal view; we used only transverse CT images. The assessment of extracolonic compartment metastases of the abdomen and pelvis was performed on 5-mm venous phase contrast-enhanced transverse images.</p>", "<p>Bowel wall thickening of more than 0.5 cm was considered to indicate the presence of a neoplasm. Colorectal wall invasion was analyzed according to a modified T classification reported by Filippone et al. Contrast-enhanced CT criteria for T staging were ≤<italic>T2 </italic>= smooth outer border of thickened colorectal wall with a clear surrounding fat plane, <italic>T3 </italic>= tumor with rounded or nodular advancing margin, <italic>T4 </italic>= obliteration of fat planes between colorectal tumor and adjacent organs. This classification was used to address known limitations at CT in distinguishing T1 and T2 lesions.</p>", "<p>The cancer was found in the ileocecal junction of 11 patients (11.9%), in ascending colon of 14 patients (15.2%), in transverse colon of 8 subjects (8.7%), in hepatic flexure of 8 patients (8.7%), in splenic flexure of 7 patients (7.6%), in descending colon of 10 patients (10.9%) and in sigmoid colon of 34 patients (36,9%).</p>", "<title>Operative procedure</title>", "<p>All patients were surgically treated. Right hemicolectomy was performed in 14 patients, ileocecal resection in 11 cases, transverse colon resection in 8 patients, left hemicolectomy in 25 patients, sigmoid resection in 24 patients, Hartmann procedure in 10 cases.</p>", "<p>T and N staging was based on the international TNM classification, as follows: pT1, tumor invading submucosal layer; pT2, tumor invading muscularis propria or subserosa; pT3, tumor penetrating serosa and perivisceral fat; and pT4, tumor invading adjacent organs. Lymph nodes were likewise classified: N0, no regional lymph node metastasis; N1, metastasis in one to three perirectal lymph nodes; N2, metastasis in four or more perirectal lymph nodes; and N3, metastasis in pelvic lymph nodes. The patients were classified as follows: 5 pT1 N0 M0; 2 pT1 N1 M0; 4 pT1 N2 M0; 16 pT2 N0 M0;; 2 pT2 N1 M0; 2 pT2 N2 M0; 37 pT3 N0 M0; 14 pT3 N1 M0; 10 pT3 N2 M0.</p>", "<title>Data analysis</title>", "<p>Medical records of patients with colon cancer were isolated in a computerized database. The database included 53 demographics, clinical, laboratory and patho-morphological parameters: name, sex, age, symptoms and major medical problems of patients; laboratory data and neoplastic markers values; location, size, endoscopic appearance and preoperative staging of the tumour; operation type, degree of differentiation and pTNM of the cancer; postoperative course, recurrence, and condition at follow-up.</p>", "<p>These patients returned for follow-up every 6 months during the first 3 years and then once a year. When necessary, telephone contact was made with the patient to obtain up-to-date information. The dead line of follow-up was up to January 2007. The longest follow-up time was 96 months with an average period of 60 months.</p>", "<p>Statistical analysis was performed using the software program Statgraphics Plus. Analysis of variance (F-ratio) was used for comparison of quantitative parameters, whereas qualitative parameters were analyzed by Chi-square test. To compare the prognostic value of the statistically significant variables, multivariate analysis (Multivariate analysis of variance MANOVA) was performed with 95% CI for the means of each variable.</p>", "<p>All <italic>P </italic>values were two-tailed. <italic>P </italic>values of less than 0.05 were considered statistically significant.</p>" ]

[ "<title>Results</title>", "<p>Between January 1999 and December 2001, 92 patients with colon cancer underwent surgical resection at the Department of General Surgery, University Hospital Tor Vergata, Rome and were included in the study.</p>", "<p>Patients consisted of 48 males and 44 females who ranged in age from 37 to 94 years (average age of 69.2 years). The average interval between symptoms and diagnosis was 5.4 months (range 1–70).</p>", "<p>In our experience 92 patients with colon cancer were recruited between January 1999 and December 2001 and followed up to December 2006. The average follow up period of the 92 patients was 40.6 months (range 3–96). The 5-year survival rate was 39.1% (36/92).</p>", "<p>The results of univariate analyses of clinical, laboratory and pathomorphologic data considering pT staging as the independent variable are summarized in table ##TAB##0##1##; pT stages were considered as 2 categories: pT &lt; 2 and pT &gt; 2.</p>", "<p>Presence of mucorrhea and anismus, hematocrit value ranging between 16,7% and 31%, WBC count between 4500 and 5800/mm³and fibrinogen value &gt; 400 mg/dl were significantly related to pT staging &gt; 2. Further the degree of CT scan T-staging were significantly related to pT staging of the tumour.</p>", "<p>Among clinical parameters, presence of mucorrhea (p &lt; 0.005; F-ratio 8.75) appeared more significantly related to pT staging &gt; 2 than anismus (p &lt; 0.05; F-ratio 4.26). Among laboratory data, fibrinogen value &gt; 400 mg/dl (p &lt; 0.0005; F-ratio 6,64) appeared more significantly related to pT &gt; 2 than WBC count ranging between 4500 and 5800/mm³(p &lt; 0.01; F-ratio 1,90) or hematocrit value of 16,7–31% (p &lt; 0.05; F-ratio 2,54). Furthermore, CT scan-T staging appeared strongly related to the pathologic-T staging (p &lt; 0.01; F-ratio 5,21).</p>", "<p>Only those variables that appeared significant in the univariate analysis were considered for the multivariable analysis. Fibrinogen value appeared the most significant predictor of pathologic-T staging of the tumour (p &lt; 0.001, F-ratio 5.86), as shown in table ##TAB##1##2##.</p>", "<p>On survival analyses, the pathologic-T staging of the tumour (p &lt; 0.01; F-ratio 2.11), the operation type (p &lt; 0.01; F-ratio 3.51) and the CT scanning (p &lt; 0.05; F-ratio 5,21) appeared to be prognostic indicators (table ##TAB##2##3##).</p>", "<p>Finally, tumour site was found significantly related neither to survival not to the degree of tumour differentiation as shown in table ##TAB##3##4##.</p>" ]

[ "<title>Discussion</title>", "<p>Several studies provided data regarding the survival of patients with colorectal cancer. Different clinico-pathological prognostic factors have been proposed: age, location of the cancer, surgical procedure, radical resection, blood transfusion, pathological type, diameter, depth of tumor invasion, lymph node metastasis and distant metastasis [##REF##16552821##1##,##REF##7577025##2##,##REF##8697383##8##].</p>", "<p>The site of the tumor is one of the prognostic factors investigated. Patients with colon cancer are considered having a better survival than those with rectal cancer [##REF##11289282##6##,##REF##11227943##7##,##REF##3021418##9##]. In previous studies distal location and advanced stage of tumor were determined as independent prognostic factors for survival of patients with colorectal cancer [##REF##12705353##10##]. In the present study, we considered only patients with colon cancer and, among these patients, we found relationship neither between tumor location and survival nor between tumor location and degree of cancer differentiation.</p>", "<p>Differently, pathological classification is one of the prognostic factors proposed for patients with colorectal cancer. It was suggested that patients with different papillary adenocarcinoma have the best outcome, patients with moderately-differentiated and mucinous adenocarcinoma have a moderate outcome, patients with signet-ring cell poorly-differentiated adenocarcinoma have a poor prognosis [##REF##15868237##11##].</p>", "<p>Current data indicates also that radical resection and type of operation are important prognostic factors. In a recent study on prognosis of 96 patients with colon cancer, survival time and 3-year survival rate were, respectively, 24 months and 6.5% for patients who received R2 procedure, 98 months and 87.9% for patients who underwent R0 resection [##REF##16552821##1##]. Further, the prognosis of the patients undergone left hemicolectomy (splenic flexure of colon, descending colon and most part of sigma colon) was not different from that of the patients undergone right hemicolectomy (caecum, ascending colon and hepatic flexure of colon)[##REF##16552821##1##]. Differently, in our experience type of operation was an indicator of survival (P &lt; 0.001; F-ratio 3.51).</p>", "<p>Several analyses confirmed the vital importance of tumour stage, as reflected in Dukes or TNM classification, in predicting survival [##REF##7577025##2##,##REF##1730115##4##,##REF##3366023##5##]. The overall 5-year survival rate of patients with colorectal cancer, reported in literature, is at least 60% and raises to 90% for Dukes A tumour and conversely decreases to 10% for Dukes D².</p>", "<p>Accordingly, stage T4 and vascular invasion were reported as markers of poor prognosis [##REF##16622901##12##]. Petersen and colleagues [##REF##12077094##13##] identified these factors in a series of 268 patients with stage II colonic cancer, together with positive surgical margins and perforation. Burdy et al [##REF##11711742##14##] identified stage T4 in a study of 108 colonic cancers, and also reported independent significance for male sex, bowel obstruction and number of nodes examined. Morris et al [##REF##16622901##12##], in a observational study on 1306 patients, reported only T4 and vascular invasion as significant factors in multivariable analysis. Mulcahy et al [##REF##9118749##15##] found a trend for prognostic significance of vascular invasion in rectal but not in colonic cancers. Accordingly, in our experience the pT staging of the cancer was strongly related to survival (P &lt; 0.001; F-ratio 2.11). Therefore, histopathological factors continue to be the most valuable source of information regarding the possible evolution of patients with colon cancer [##REF##12368937##16##,##REF##15011850##17##].</p>", "<p>Although surgical therapy is the basis of treatment for patients with colon cancer, multimodal therapeutical concepts are currently applied not only in metastatic disease and Dukes C patients but also in Dukes B [##REF##10334518##18##, ####UREF##0##19##, ##REF##16422171##20####16422171##20##]. Further, the introduction of new drugs has extended the therapeutic options [##REF##16162019##21##,##REF##16945169##22##] and ongoing studies with novel targeted therapies will show their results in the next years.</p>", "<p>Other open questions include the timing of treatment (neoadjuvant treatment for locally advanced cancer), the duration of therapy and whether there is a real possibility of stratifying patients for neoadjuvant treatment on the basis of preoperative prognostic factors [##REF##16760294##23##, ####REF##12853343##24##, ##REF##12810454##25####12810454##25##].</p>", "<p>Actually, recent studies, that underlined remission rates as high as 40% in advanced colon cancer and the efficacy of adjuvant chemotherapy, are now leading to revaluation of the therapeutic approach to colon cancer. Data from animal models and tumour biologic hypotheses also point to a possible advantage for preoperative therapy to improve disease-free survival and overall survival in colon cancer patients [##REF##8088204##26##].</p>", "<p>Therefore, the main target of this investigation was to identify preoperative clinical, laboratory and patho-morphological parameters that may be indicators of the pT staging of the tumour.</p>", "<p>Considering clinical data, we observed that the presence of mucorrhea and anismus were indicators of pT staging &gt; 2, it was suggested that these signs are also related to the tumour size. On laboratory, we found that hematocrit value between 16,7 and 31% and WBC count ranging between 4500 and 5800/mm³were significantly associated to pT &gt; 2. Further fibrinogen value &gt; 400 mg/dl both on univariate and multivariate analysis was significantly associated to pT staging &gt; 2; however the importance of fibrinogen value to predict tumour invasion and prognosis of these patients remains uncertain, given that this marker is not specific and it is also involved in acute-phase reaction.</p>", "<p>The significance of increase of CEA and CA 19-9 levels to predict the prognosis of the patients remains a problem for debate; in our experience CEA and CA 19-9 levels were found indicators neither of pT staging nor of survival. Differently, in a recent study on 103 patients with colorectal carcinoma, Nozoe et al [##REF##16847905##27##] observed that high preoperative CEA and CA 19-9 were predictors of survival. On the opposite, in a large study of 279 patients with colon cancer and 293 patients with rectal cancer, Tominaga et al [##REF##8697383##8##] found that, although higher preoperative CEA level group tended to have a higher recurrence rate, preoperative CEA level was not statistically related to the prognosis of both groups of patients (Hazard ratio 1.34 and 1.35).</p>", "<p>Besides this, an interesting result of our investigation is the importance of CT scan preoperative staging both as predictors of survival and of pT staging. These data confirm the diagnostic accuracy of this tool not only for the rectum but also for the colon cancer as pointed out by Blomqvist in a recent review on advances in preoperative staging of colorectal cancer patients. The author underlines the technological advancement of CT scan related to hardware, software, development of CT colonography, new contrast enhancement agents [##REF##14693879##28##,##REF##12705549##29##].</p>", "<p>In summary, these data confirm the vital importance of tumour stage in predicting survival and recurrence and provide the grounds for further work in order to assess the prognostic significance of various clinical, laboratory, patho-morphological markers and to define the subgroups of patients at different risk of recurrence who could be treated more or less intensively.</p>" ]

[]

[ "<p>This is an Open Access article distributed under the terms of the Creative Commons Attribution License (<ext-link ext-link-type=\"uri\" xlink:href=\"http://creativecommons.org/licenses/by/2.0\"/>), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.</p>", "<title>Background</title>", "<p>The long-term prognosis of patients with colon cancer is dependent on many factors. To investigate the influence of a series of clinical, laboratory and morphological variables on prognosis of colon carcinoma we conducted a retrospective analysis of our data.</p>", "<title>Methods</title>", "<p>Ninety-two patients with colon cancer, who underwent surgical resection between January 1999 and December 2001, were analyzed. On survival analysis, demographics, clinical, laboratory and pathomorphological parameters were tested for their potential prognostic value. Furthermore, univariate and multivariate analysis of the above mentioned data were performed considering the depth of tumour invasion into the bowel wall as independent variable.</p>", "<title>Results</title>", "<p>On survival analysis we found that depth of tumour invasion (P &lt; 0.001; F-ratio 2.11), type of operation (P &lt; 0.001; F-ratio 3.51) and CT scanning (P &lt; 0.001; F-ratio 5.21) were predictors of survival. Considering the degree of mural invasion as independent variable, on univariate analysis, we observed that mucorrhea, anismus, hematocrit, WBC count, fibrinogen value and CT scanning were significantly related to the degree of mural invasion of the cancer. On the multivariate analysis, fibrinogen value was the most statistically significant variable (P &lt; 0.001) with the highest F-ratio (F-ratio 5.86). Finally, in the present study, the tumour site was significantly related neither to the survival nor to the mural invasion of the tumour.</p>", "<title>Conclusion</title>", "<p>The various clinical, laboratory and patho-morphological parameters showed different prognostic value for colon carcinoma. In the future, preoperative prognostic markers will probably gain relevance in order to make a proper choice between surgery, chemotherapy and radiotherapy. Nevertheless, current data do not provide sufficient evidence for preoperative stratification of high and low risk patients. Further assessments in prospective large studies are warranted.</p>" ]

[ "<title>Competing interests</title>", "<p>The authors declare that they have no competing interests.</p>", "<title>Authors' contributions</title>", "<p>MG: manuscript preparation and critical review. GM: critical review. GMG: data collection and manuscript preparation. FC: literature review and manuscript preparation. MGM: literature review. CN: data collection and literature review. FR: critical review. AMF: critical review. All authors read and approved the final manuscript.</p>" ]

[]

[]

[ "<table-wrap position=\"float\" id=\"T1\"><label>Table 1</label><caption><p>Results of univariate analyses of clinical, laboratory and pathomorphologic data considering pT staging as independent parameter</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td align=\"center\"><bold>Source</bold></td><td align=\"center\"><bold>Sum of squares</bold></td><td align=\"center\"><bold>Df</bold></td><td align=\"center\"><bold>Mean Square</bold></td><td align=\"center\"><bold>R-ratio</bold></td><td align=\"center\"><bold>P-value</bold></td></tr></thead><tbody><tr><td align=\"left\"><bold>pT staging vs mucorrea</bold></td><td/><td/><td/><td/><td/></tr><tr><td align=\"left\">Between groups</td><td align=\"center\">41.2011</td><td align=\"right\">1</td><td align=\"center\">41.2011</td><td align=\"center\">8.75</td><td align=\"center\">0.005</td></tr><tr><td align=\"left\">Within groups</td><td align=\"center\">423.875</td><td align=\"right\">90</td><td align=\"center\">4.70972</td><td/><td/></tr><tr><td align=\"left\"><bold>Total correct</bold></td><td align=\"center\">465.076</td><td align=\"right\">91</td><td/><td/><td/></tr><tr><td colspan=\"6\"><hr/></td></tr><tr><td align=\"left\"><bold>pT staging vs anismus x</bold></td><td/><td/><td/><td/><td/></tr><tr><td align=\"left\">Between groups</td><td align=\"center\">21.0117</td><td align=\"right\">1</td><td align=\"center\">21.0117</td><td align=\"center\">4.26</td><td align=\"center\">0.0519</td></tr><tr><td align=\"left\">Within groups</td><td align=\"center\">444,064</td><td align=\"right\">90</td><td align=\"center\">4,93405</td><td/><td/></tr><tr><td align=\"left\"><bold>Within groups</bold></td><td align=\"center\">465,076</td><td align=\"right\">91</td><td/><td/><td/></tr><tr><td colspan=\"6\"><hr/></td></tr><tr><td align=\"left\"><bold>pT staging vs-fibrinogens</bold></td><td/><td/><td/><td/><td/></tr><tr><td align=\"left\">Between groups</td><td align=\"center\">453.687</td><td align=\"right\">78</td><td align=\"center\">5.8165</td><td align=\"center\">6.64</td><td align=\"center\">0.0003</td></tr><tr><td align=\"left\">Within groups</td><td align=\"center\">11.3889</td><td align=\"right\">13</td><td align=\"center\">0.876068</td><td/><td/></tr><tr><td align=\"left\"><bold>Total (correct)</bold></td><td align=\"center\">465.076</td><td align=\"right\">91</td><td/><td/><td/></tr><tr><td colspan=\"6\"><hr/></td></tr><tr><td align=\"left\"><bold>pT staging vs WBC</bold></td><td/><td/><td/><td/><td/></tr><tr><td align=\"left\">Between groups</td><td align=\"center\">630.923</td><td align=\"right\">50</td><td align=\"center\">12.6185</td><td align=\"center\">1.90</td><td align=\"center\">0.01</td></tr><tr><td align=\"left\">Within groups</td><td align=\"center\">279.55</td><td align=\"right\">42</td><td align=\"center\">6.65595</td><td/><td/></tr><tr><td align=\"left\"><bold>Total (correct)</bold></td><td align=\"center\">910.473</td><td align=\"right\">92</td><td/><td/><td/></tr><tr><td colspan=\"6\"><hr/></td></tr><tr><td align=\"left\"><bold>pT staging vs hematocrit</bold></td><td/><td/><td/><td/><td/></tr><tr><td align=\"left\">Between groups</td><td align=\"center\">852.117</td><td align=\"right\">79</td><td align=\"center\">10.7863</td><td align=\"center\">2.54</td><td align=\"center\">0.05</td></tr><tr><td align=\"left\">Within groups</td><td align=\"center\">59.5</td><td align=\"right\">14</td><td align=\"center\">4.25</td><td/><td/></tr><tr><td align=\"left\"><bold>Total (correct)</bold></td><td align=\"center\">911.617</td><td align=\"right\">93</td><td/><td/><td/></tr><tr><td colspan=\"6\"><hr/></td></tr><tr><td align=\"left\"><bold>pT vs CT staging</bold></td><td/><td/><td/><td/><td/></tr><tr><td align=\"left\">Between groups</td><td align=\"center\">126.287</td><td align=\"right\">5</td><td align=\"center\">25.2575</td><td align=\"center\">5.21</td><td align=\"center\">0.01</td></tr><tr><td align=\"left\">Within groups</td><td align=\"center\">58.1571</td><td align=\"right\">12</td><td align=\"center\">4.84643</td><td/><td/></tr><tr><td align=\"left\"><bold>Total (correct)</bold></td><td align=\"center\">184.444</td><td align=\"right\">17</td><td/><td/><td/></tr></tbody></table></table-wrap>", "<table-wrap position=\"float\" id=\"T2\"><label>Table 2</label><caption><p>Results of multivariate analysis of clinical and laboratory data of patients with colon cancer</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td align=\"center\"><bold>Source</bold></td><td align=\"center\"><bold>Sum of squares</bold></td><td align=\"center\"><bold>Df</bold></td><td align=\"center\"><bold>Mean Square</bold></td><td align=\"center\"><bold>F-ratio</bold></td><td align=\"center\"><bold>P-value</bold></td></tr></thead><tbody><tr><td align=\"left\">Main effects</td><td/><td/><td/><td/><td/></tr><tr><td align=\"left\">A: fibrinogens</td><td align=\"center\">405.233</td><td align=\"right\">78</td><td align=\"center\">5.19529</td><td align=\"center\">5.86</td><td align=\"center\">0.0014</td></tr><tr><td align=\"left\">B: mucorrhea</td><td align=\"center\">1.38889</td><td align=\"right\">1</td><td align=\"center\">1.38889</td><td align=\"center\">1.57</td><td align=\"center\">0.2366</td></tr><tr><td align=\"left\">C: tenesmus</td><td align=\"center\">0.25</td><td align=\"right\">1</td><td align=\"center\">0.25</td><td align=\"center\">0.28</td><td align=\"center\">0.6059</td></tr><tr><td colspan=\"6\"><hr/></td></tr><tr><td align=\"left\"><bold>Residual</bold></td><td align=\"center\">9.75</td><td align=\"right\">11</td><td align=\"center\">0.886364</td><td/><td/></tr><tr><td colspan=\"6\"><hr/></td></tr><tr><td align=\"left\"><bold>Total (correct)</bold></td><td align=\"center\">465.076</td><td align=\"right\">91</td><td/><td/><td/></tr></tbody></table></table-wrap>", "<table-wrap position=\"float\" id=\"T3\"><label>Table 3</label><caption><p>Prognostic indicators of survival in patients surgically treated for colon cancer</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td align=\"center\"><bold>Source</bold></td><td align=\"center\"><bold>Sum of squares</bold></td><td align=\"center\"><bold>Df</bold></td><td align=\"center\"><bold>Mean Square</bold></td><td align=\"center\"><bold>R-ratio</bold></td><td align=\"center\"><bold>P-value</bold></td></tr></thead><tbody><tr><td align=\"left\"><bold>Survival vs pT staging</bold></td><td/><td/><td/><td/><td/></tr><tr><td align=\"left\">Between groups</td><td align=\"center\">136.587</td><td align=\"right\">15</td><td align=\"center\">9.10583</td><td align=\"center\">2.11</td><td align=\"center\">0.01</td></tr><tr><td align=\"left\">Within groups</td><td align=\"center\">328.489</td><td align=\"right\">76</td><td align=\"center\">4.32222</td><td/><td/></tr><tr><td align=\"left\"><bold>Total correct</bold></td><td align=\"center\">465.076</td><td align=\"right\">91</td><td/><td/><td/></tr><tr><td colspan=\"6\"><hr/></td></tr><tr><td align=\"left\"><bold>Survival vs operation</bold></td><td/><td/><td/><td/><td/></tr><tr><td align=\"left\">Between groups</td><td align=\"center\">22292.1</td><td align=\"right\">6</td><td align=\"center\">3715.35</td><td align=\"center\">3.51</td><td align=\"center\">0.01</td></tr><tr><td align=\"left\">Within groups</td><td align=\"center\">89913.6</td><td align=\"right\">85</td><td align=\"center\">1057.81</td><td/><td/></tr><tr><td align=\"left\"><bold>Total correct</bold></td><td align=\"center\">112206.0</td><td align=\"right\">91</td><td/><td/><td/></tr><tr><td colspan=\"6\"><hr/></td></tr><tr><td align=\"left\"><bold>Survival vs CT staging</bold></td><td/><td/><td/><td/><td/></tr><tr><td align=\"left\">Between groups</td><td align=\"center\">7770.25</td><td align=\"right\">5</td><td align=\"center\">1554.05</td><td align=\"center\">3.90</td><td align=\"center\">0.05</td></tr><tr><td align=\"left\">Within groups</td><td align=\"center\">4778.86</td><td align=\"right\">12</td><td align=\"center\">398.238</td><td/><td/></tr><tr><td align=\"left\"><bold>Total correct</bold></td><td align=\"center\">12549.1</td><td align=\"right\">17</td><td/><td/><td/></tr></tbody></table></table-wrap>", "<table-wrap position=\"float\" id=\"T4\"><label>Table 4</label><caption><p>Statistical correlation between tumour site and degree of differentiation on the cancer</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td align=\"center\"><bold>Degree of differentiation</bold></td><td align=\"center\" colspan=\"2\"><bold>Site</bold></td><td align=\"center\"><bold>Total</bold></td><td align=\"center\"><bold>Significance</bold></td></tr><tr><td/><td align=\"center\"><bold>R</bold></td><td align=\"center\"><bold>L</bold></td><td/><td/></tr></thead><tbody><tr><td align=\"center\"><bold>very differentiated</bold></td><td align=\"center\"><bold>5</bold></td><td align=\"center\"><bold>10</bold></td><td align=\"center\"><bold>15</bold></td><td align=\"center\"><bold>p = n.s.</bold></td></tr><tr><td align=\"center\"><bold>moderately differentiated</bold></td><td align=\"center\"><bold>27</bold></td><td align=\"center\"><bold>42</bold></td><td align=\"center\"><bold>69</bold></td><td align=\"center\"><bold>p = n.s.</bold></td></tr><tr><td align=\"center\"><bold>insufficient differentiated</bold></td><td align=\"center\"><bold>5</bold></td><td align=\"center\"><bold>3</bold></td><td align=\"center\"><bold>8</bold></td><td align=\"center\"><bold>p = n.s.</bold></td></tr><tr><td colspan=\"5\"><hr/></td></tr><tr><td align=\"center\"><bold>Total</bold></td><td align=\"center\"><bold>37</bold></td><td align=\"center\"><bold>55</bold></td><td align=\"center\"><bold>92</bold></td><td/></tr></tbody></table></table-wrap>" ]

[]

[]

[]

[]

[]

[]

[]

[]

[]

{ "acronym": [], "definition": [] }

[ "<title>Background</title>", "<p>Natural glycans are known to take part in many key biological processes such as cell adhesion, recognition, receptor activation or signal transduction, and they also exhibit major structural functions in both bacteria and plants. In addition, bacterial glycans act as virulence, osmoprotection and desiccation protection factors [##UREF##0##1##]. The diversity of structures within the mammalian glycome seems limited and is well described in the literature [##REF##16959566##2##]. On the other hand, the bacterial glycome exhibits greater diversity, stemming largely from the distinct cell wall architecture of these organisms.</p>", "<p>The cell envelope of either Gram-positive or Gram-negative bacteria is based on peptidoglycan, a polymer in which polysaccharide chains are cross-linked with short peptide chains [##UREF##1##3##]. Gram-negative bacteria possess an additional outer membrane that is composed of a lipopolysaccharide-protein complex. Gram-positive bacteria have no outer membrane, but the peptidoglycan wall is thicker (&gt; 30 nm vs. 10 nm in Gram-negative bacteria) and contains polysaccharides with teichoic acids attached (a carbohydrate polymer containing alditols and phosphodiester linkages).</p>", "<p>Both Gram-positive and Gram-negative bacteria produce extracellular polysaccharides, present either as a discrete capsule covalently attached to the cell envelope or as a slime weakly bound to the cell surface. These various glycoconjugates and polysaccharides on the surface of the cell often contain the antigenic determinants that initiate an immunogenic response in a host. In addition, these surface carbohydrates provide recognition elements for pathogens such as bacteriophages.</p>", "<p>The lipopolysaccharide of Gram-negative bacteria contains lipid A, a phosphorylated GlcN-GlcN disaccharide moiety, <italic>N</italic>- and <italic>O</italic>-acylated with fatty acids which anchor the molecule in the outer leaflet of the outer membrane. Lipid A is covalently linked to a heteropolysaccharide which interacts with the environment and consists of an inner core (commonly containing Kdo (3-deoxy-<sc>D</sc>-<italic>manno</italic>-oct-2-ulosonic acid) and <italic>manno</italic>-heptoses) and an outer <italic>O</italic>-specific chain, a complex polysaccharide which determines the serological or antigenic properties of the lipopolysaccharide [##REF##17362200##4##,##REF##12045108##5##]. These so-called <italic>O</italic>-antigens are mainly heteropolymers containing a large variety of residues (mainly monosaccharides, but also alditols, amino acids, etc.). These components, together with the capsular polysaccharides (K-antigens [##REF##2695750##6##,##REF##16756484##7##]), can elicit an immune response in higher organisms.</p>", "<p>The structures of the various carbohydrate antigens are unique, often being characterized by repeating units in the polymer structure. Indeed, all types of monosaccharides, including L-rhamnose (6-deoxy-L-mannose) and L-fucose (6-deoxy-L-galactose), are found in bacteria, together with rarer, modified sugars, such as 3,6-dideoxyhexoses and Kdo. Knowledge of the structures of surface carbohydrates and their variations is required for understanding how cellular recognition, adhesion, and the immune response operate at the molecular level. This understanding provides a basis for the design of synthetic carbohydrate-based vaccines, diagnostic agents, and immunostimulators. Certain fragments of bacterial polysaccharides, in the form of appropriate glycoconjugates, are known to act as vaccines [##REF##16630616##8##].</p>", "<p>Carbohydrates represent the most diverse class of biopolymers, and there is growing interest in the study and analysis of this diversity and its biomedical significance. For example, vertebrate glycan variability is assumed to act as a barrier that prevents the spread of an infection within a given population [##REF##10406840##9##]. Although it is widely known that the diversity of carbohydrates is much greater in bacteria than in mammals, no systematic attempt has been undertaken to examine the diversity of bacterial carbohydrates in detail. The structures deposited in glycoscience databases have been only sporadically evaluated. However, statistical structure-oriented investigations using carbohydrate databases were proven to be useful for immunochemical research and serotyping [##REF##16594963##10##]. Systematic analysis of all publicly available data will not only expand our general knowledge and understanding of the complexity of glycans in biological systems but will also offer a framework for the design of more comprehensive high-throughput screening methods or devices.</p>", "<p>Comprehensive data concerning carbohydrate diversity within the entire bacterial world will be useful for the classification of bacteria according to their glycan structures and facilitate the search for the most widespread carbohydrate markers of various bacterial taxonomic groups. These markers are critical for medical applications, and a simple ranking by abundance is a good starting point for the design of synthetic biologically-active carbohydrates and for corresponding immunological studies. In particular, the statistics of <italic>monomer composition </italic>reveal potential taxonomic markers and also simplify the creation of carbohydrate microarrays by providing candidates for spotting [##REF##17460666##11##].</p>", "<p>A one-enzyme-class/one-saccharide-linkage paradigm applies for almost all individual steps of glycan biosynthesis. Accordingly, complete information on the <italic>diversity of disaccharide fragments </italic>allows one to describe the diversity of the glycosyltransferases expressed in individual taxonomic groups, and these enzymes may become potential targets for antimicrobial treatment.</p>", "<p>For this study we performed statistical analyses of the Bacterial Carbohydrate Structure Data Bank (BCSDB), the largest database for bacterial glycans containing nearly all known bacterial glycan structures published up to 2007 [##UREF##2##12##]. For comparison the mammalian glycans documented in the GLYCOSCIENCES.de database [##REF##16239495##13##] (derived mainly from CarbBank [##REF##2623761##14##]) have also been examined. The properties analyzed include glycan size, branching, and charge density, as well as the frequency of occurrence of specific monosaccharide residues, residue pairs and their linkage configurations. Precise definitions for the terminology used in this study can be found in the Methods section.</p>" ]

[ "<title>Methods</title>", "<p>The analyzed sequences were obtained from the meta-database GlycomeDB [##UREF##4##27##], which contains all sequences from the Bacterial Carbohydrate Structure DataBase (BCSDB) [##UREF##2##12##] and the GLYCOSCIENCES.de portal [##REF##16239495##13##] in a harmonized format. With the help of the NCBI taxonomy database [##REF##10592169##15##], subsets of these databases were taken and further analyzed using routines implemented in JAVA. The results of the analytic routines were stored in a PostgreSQL 8.2 database. Additional analytical procedures were implemented in PHP, which finally generated Microsoft Excel tables used for further analysis and graphical visualization.</p>", "<title>Definition of terms</title>", "<p>According to IUPAC nomenclature a <italic>monosaccharide </italic>is a poly(hydroxy) aldehyde or ketone with three or more carbon atoms (triose, tetrose, etc.); the term denotes a single structural component without glycosidic linkages and includes a variety of derivatives such as amino, deoxy or carboxy forms. <italic>Oligosaccharides </italic>are compounds in which monosaccharides and their derivatives are coupled in a precisely defined manner via glycosidic linkages. The term <italic>polysaccharide </italic>generally refers to oligosaccharides with a large or undefined number of monosaccharide residues. The term <italic>carbohydrate </italic>includes all mono-, oligo- or polysaccharides and molecules derived from monosaccharides modified by reduction, oxidation, or substitution. The term <italic>glycan </italic>is frequently used to refer to any saccharide component of a glycoconjugate, such as a glycoprotein or glycolipid, even when the chain length is short. A <italic>glycoconjugate </italic>is formed by a covalent linkage between a glycan and a nonglycan entity. <italic>Polysaccharide </italic>may be used to refer to polymers with glycosidic and/or phosphodiester linkages (such as teichoic acids). The carbohydrate databases used in this study may contain any of the compounds described above but do not contain DNAs or RNAs.</p>", "<p>In this study we used the following definitions:</p>", "<p>1. The term <italic>sequence </italic>will be used here to refer to a specific carbohydrate molecule or a glycan obtained from a larger molecule (glycoconjugate). Sequences may be linear or branched. Each database record may refer to either an individual carbohydrate sequence or to a particular glycoconjugate containing a given glycan. Thus, each unique glycan may have multiple database records, one for each different glycoconjugate.</p>", "<p>2. A <italic>residue </italic>is a specific building block, e.g. a monosaccharide, within a carbohydrate sequence, analogous to the amino acid residues in proteins.</p>", "<p>3. The term <italic>unit </italic>will be used to specify the smallest sequence fragment which describes a given carbohydrate molecule or glycan. For a nonrepeating oligomer sequence the unit will be the entire sequence; for polymers built up from repeating subsequences, the unit will be one such subsequence.</p>", "<p>4. A <italic>branching point </italic>is a particular residue to which <italic>two or more </italic>carbohydrate residues are attached via nonreducing hydroxy functions or other functional groups.</p>", "<p>5. A <italic>monosaccharide </italic>is a unique carbohydrate residue according to the IUPAC definition given above and is specified by the number of carbons, the ring type, the anomeric (α,β) and absolute (<sc>D</sc>, <sc>L</sc>) configurations, and all primary and secondary modifications. For example, α-<sc>D</sc>-Glc<italic>p</italic>N, β-<sc>D</sc>-Glc<italic>p</italic>N and α-<sc>D</sc>-Glc<italic>p</italic>NAc are three different monosaccharides derived from glucose.</p>", "<p>6. <italic>Primary modifications </italic>of monosaccharides are those which alter the stereochemical designation or electronic hybridisation state of at least one carbon atom (e.g. deoxy, carboxy, keto, double-bond modifications). <italic>Secondary modifications </italic>are all modifications which are not primary (substituents such as amino, <italic>O</italic>-methyl, <italic>O</italic>-acetyl, sulfate, phosphate, etc.).</p>", "<p>7. We define the <italic>basetype </italic>of a monosaccharide to include only those characteristics which specify the order and stereochemical designations of its carbon atom skeleton, i.e., the anomeric and absolute configurations, ring type, and primary modifications. <italic>Basetype </italic>is <italic>not </italic>altered by secondary modifications. Thus, α-<sc>D</sc>-Glc<italic>p</italic>, α-<sc>D</sc>-Glc<italic>p</italic>N and α-<sc>D</sc>-Glc<italic>p</italic>6S all have the same basetype (α-<sc>D</sc>-Glc<italic>p</italic>) while β-<sc>D</sc>-Glc<italic>p </italic>and α-<sc>D</sc>-Glc<italic>p</italic>A are different basetypes.</p>", "<p>8. With respect to common historical usage, the term <italic>basic entity </italic>will be used to specify the following characteristics of a monosaccharide: the stereochemical configuration (D, L), all primary modifications, and only those secondary modifications involving amine groups, including substituted amines, at any position other than the anomeric carbon. The basic entity definition does <italic>not </italic>include anomeric configuration, ring type, or any secondary modifications. Thus, α-<sc>D</sc>-Gal<italic>f</italic>N and β-<sc>D</sc>-Gal<italic>p</italic>NAc have the same basic entity (<sc>D</sc>-GalN) while β-<sc>D</sc>-Gal<italic>p</italic>, β-<sc>D</sc>-Gal<italic>p</italic>A and β-<sc>D</sc>-Fuc<italic>p </italic>are all different basic entities.</p>" ]

[ "<title>Results and Discussion</title>", "<title>Distribution of carbohydrate structures among taxonomic groups</title>", "<p>We first examined the number of sequences found in the BCSDB and GLYCOSCIENCES.de for various taxonomic ranks (class, order, family). Where possible, the taxonomic relationships were traced using the NCBI taxonomy database [##REF##10592169##15##]. The GLYCOSCIENCES.de database currently contains a total of 23120 glycan and glycoconjugate records, of which 13704 records for diverse animal, plant, bacteria and fungi classes have some information concerning taxonomy. In the BCSDB there are a total of 8504 records for bacteria only, and 8479 of these contain information concerning taxonomy. These numbers may include multiple records for a given glycan when the same glycan is reported for more than one species. Note that not all taxonomic classes are represented in the databases and that for bacterial glycans there is considerable overlap between the two databases.</p>", "<p>If we now consider the two databases combined, there are a total of 13775 nonredundant carbohydrate records which include taxonomic information. The distribution of these records among various taxonomic classes is shown both numerically and schematically in Fig. ##FIG##0##1##. A more detailed breakdown of the distribution can be found in the additional material [see Additional file ##SUPPL##0##1##].</p>", "<p>The taxonomic class <italic>Mammalia </italic>is found to have 4739 assigned sequence/taxon pairs, of which 2118 are of human origin (family <italic>Hominidae</italic>). All other animal or plant classes in the database have less than 350 pairs. The category \"unresolved\" refers to the 1482 records for which the source is defined but the specific taxonomic class could not be traced automatically using the NCBI. Only about half of the bacterial phyla are represented in the BCSDB with a total of 6098 sequence/taxon pairs, and nine classes have less than 10 records. Note that the number of carbohydrates or glycoconjugates registered for a given taxonomic class does not necessarily reflect its species diversity, but more likely the intensity with which the class has been studied. Thus, the apparent diversity of carbohydrates in the various taxonomic classes reflects to a large part the information bias in the published literature, and this situation must be kept in mind when making conclusions based on the distributions of properties discussed below.</p>", "<p>In the combined databases there are a total of 12659 records in the category \"no taxonomy\" which means that either no information concerning the taxonomy of the source is available or that the carbohydrate is not of purely natural origin. These records were not included in Fig. ##FIG##0##1## and were not used in the following analyses.</p>", "<title>Choice of taxonomic datasets for statistical comparisons</title>", "<p>For the following more detailed statistical comparisons, we defined two sets of taxonomic <italic>groups</italic>, considering both biological and coverage aspects. Taxonomy Set 1 (Table ##TAB##0##1##) was defined to provide an overview of the total content of the two databases used for the general comparison of bacterial and mammalian carbohydrates, taking into account the fact that bacterial carbohydrates frequently contain repeating units while mammalian sequences usually do not. Thus, Set 1 contains three taxonomic groups: all mammalian carbohydrates, all bacterial carbohydrates with nonrepeating sequences (oligomers), and all bacterial sequences with repeating units (polymers).</p>", "<p>For comparisons within the taxonomic domain <italic>Bacteria</italic>, we defined a more detailed taxonomy Set 2 (Table ##TAB##1##2##), which includes two classes of Gram-positive bacteria (<italic>Actinobacteria </italic>and <italic>Bacilli</italic>) and the various classes of the phylum <italic>Proteobacteria</italic>. The largest of these classes, the γ-<italic>Proteobacteria</italic>, has been further subdivided in Set 2 into the major order <italic>Enterobacteriales </italic>and a subset containing all other γ-<italic>Proteobacteria</italic>. The class δ-<italic>Proteobacteria </italic>(with only 2 records) has been combined with the ε-<italic>Proteobacteria</italic>.</p>", "<p>In order to obtain meaningful statistics, only those taxonomic groups are compared for which at least 200 carbohydrate sequences are available. For this reason the classes Chlamydiae, Clostridia, and Bacteroidetes, for example, have not been included in Set 2. Note that in Tables ##TAB##0##1## and ##TAB##1##2## the total number of <italic>unique </italic>carbohydrate <italic>sequences </italic>in each group is listed, and these sets were utilized in all subsequent analyses.</p>", "<title>Carbohydrate size, branching and charge density</title>", "<p>Frequency distributions for general measures of molecular size, topology (branching) and mean charge density have been calculated for the carbohydrate sequences comprising the various taxonomic groupings described by Set 1 and Set 2 (Tables ##TAB##0##1## and ##TAB##1##2##). In each case the distributions are normalized to the <italic>total number of sequences in each taxonomic group </italic>and expressed as percentages within each group. In Fig. ##FIG##1##2A## distributions for the number of monosaccharides per sequence <italic>unit </italic>(either the entire carbohydrate sequence for oligomers or the repeating unit for polymers, see Definition 4 in the methods) are shown for bacteria vs. mammals (taxonomy Set 1). The distribution is relatively broad for mammals with mean and median values, respectively, of 8.17 and 8 monomers per sequence, while for bacteria the distribution shows a narrow peak at 4–5 monomers for both oligomers (mean: 5.94 median: 5) and for the repeating unit of polymers (mean: 4.17, median: 4). However, oligomers show a significant population of sequences with 8–15 monomers while the distribution for polymers essentially ends at 9 monomers per unit. Of course, the total length of a polymeric sequence with multiple units may very well exceed the maximum length of oligomers. Naturally occurring oligomers may also be longer than the sequences reported in the databases since the process of extracting and isolating a glycan may result in partial digestion and loss of residues.</p>", "<p>In Fig. ##FIG##1##2B## the distributions of the size parameter for the bacterial groups defined in taxonomy Set 2 are found to differ considerably. Narrow distributions with essentially a single prominant peak are found for <italic>Actinobacteria </italic>(mean: 4.51, median: 3), <italic>Bacilli </italic>(mean: 5.18, median: 5) and the order <italic>Enterobacteria </italic>(mean: 5.18, median: 6) with peaks at ca. 2.5, 5.5 and 4.5 residues, respectively. The various other classes of <italic>Proteobacteria </italic>have broader distributions with more or less pronounced multiple peaks, e.g. at 3, 8 and 11 residues for the δ,ε-Proteobacteria group.</p>", "<p>The number of branching points per carbohydrate residue can be considered to be a <italic>branching index </italic>which reflects the complexity of carbohydrate topology. Fig. ##FIG##2##3A## demonstrates that 22% of all mammalian and 50% of all bacterial sequences are linear (branching index = 0). However, for the individual bacterial groups of taxonomy Set 2, the percentage of linear structures ranges from 30% to 78% (Fig. ##FIG##2##3B##). A general feature of all branching point graphs in Fig. ##FIG##2##3## is a peak in the distribution at a branching index of 0.2 – 0.3, which corresponds to carbohydrate sequences with one branching point for every three to five monosaccharide residues. This peak in the distribution is weak for <italic>Actinobacteria </italic>and α-<italic>Proteobacteria </italic>but strong for mammals, <italic>Bacilli</italic>, and other <italic>Proteobacteria</italic>.</p>", "<p>Finally, the <italic>mean charge density </italic>parameter (max. electric charge possible for all ionizable groups divided by the number of carbohydrate residues in a sequence unit) is shown in Fig. ##FIG##3##4## for taxonomy Set 1 and Set 2. About 58% of mammalian sequences and 47% of all bacterial carbohydrate sequence units have no net charge (Fig. ##FIG##3##4A##). For the bacterial groups of Set 2, the frequency of sequence units with no net charge ranges from ca. 32% for β-<italic>Proteobacteria </italic>to 77% for <italic>Actinobacteria </italic>(Fig. ##FIG##3##4B##). All other sequences have a net negative charge due to carboxyl, phosphate or sulfate groups, for example. The distributions for mammals and <italic>Bacilli </italic>both have peaks at charge densities of -0.2 and -0.5; the δ,ε-<italic>Proteobacteria </italic>distribution exhibits a single broad peak at ca. -0.3 while the other bacterial distributions have multiple peaks at -0.3, -0.5, -0.7 and -1.0 (Fig. ##FIG##3##4B##).</p>", "<title>Monosaccharide diversity</title>", "<p>For mammals and even more so for bacteria, the diversity of the monosaccharides used as the building blocks of carbohydrate sequences is significantly larger than that for the residues in proteins or nucleic acids. From the GLYCOSCIENCES.de database a total of 35 different monosaccharides were found for mammalian carbohydrates, according to the nomenclature used in the original databases (Table ##TAB##2##3##). This degree of diversity is at first glance puzzling, in view of the common notion that mammalian carbohydrates are built up of 10 \"classical\" monosaccharides (Glc, Gal, GlcNAc, GalNAc, Man, GlcA, Fuc, Neu, IdoA, Xyl) [##UREF##0##1##]. However, the variety of monosaccharides defined in the primary databases is higher due to (a) residues being specified with unknown anomer or ring type definitions, (b) analytical artifacts from the structure elucidation process (alditols, double bonds), or (c) secondary modifications such as sulfation. Furthermore, carbohydrate sequence databases are not error-free and suffer from incorrect structure elucidations and curation mistakes. Since the existing databases generally use free-text identifiers for the monosaccharides, it was helpful to translate all structural database entries into a machine-readable notation called GlycoCT [##REF##18436199##16##]. Using structural filters based on this notation we were able to significantly reduce the fuzziness introduced by the lack of strictness in the definitions of the original sequences (Table ##TAB##2##3##). During the analysis we excluded manually common artifacts caused by analytical procedures and entries with undefined absolute or anomeric configuration or ring type.</p>", "<p>To minimize the influence of errors and artifacts on the statistics of Table ##TAB##2##3##, a threshold for the occurrence of monosaccharides, basetypes and basic entities was set at 10 for mammals and 2 for bacteria. This means that a given residue type was included in the statistics only when its number of occurrences exceeded the defined threshold. A relatively low threshold was chosen for bacteria because, in contrast to mammals, bacteria are known to produce a great variety of unique monosaccharide residues with low occurrence. When the threshold for bacteria was reduced from 2 to 0, the diversity of detected residues increased by about 25%. A complete list of monosaccharide residues and basetypes found for each taxonomic group is available in GlycoCT nomenclature in the additional material [see Additional files ##SUPPL##1##2## (tables a-i) and ##SUPPL##2##3##].</p>", "<p>For mammals the analysis returned 18 occurrences of <sc>D</sc>-Fuc as a basic entity. However, this residue was excluded from Table ##TAB##2##3## because all carbohydrates in GLYCOSCIENCES.de which are specified to contain <sc>D</sc>-Fucose originate from old publications in which the absolute configuration of Fuc was not specified. The occurrence of <sc>D</sc>-Fuc in mammalian carbohydrate records can be regarded as a data translation error since there is no evidence for a mammalian enzyme which synthesizes <sc>D</sc>-Fuc.</p>", "<p>The 35 monosaccharides with the highest occurrence within taxonomy Set 1 are shown schematically in Fig. ##FIG##4##5##. Of them, all 17 monosaccharides that were found in mammalian carbohydrates are also found in the bacterial world. Rhamnose, <sc>L</sc>-<italic>glycero</italic>-α-<sc>D</sc>-<italic>manno</italic>-Heptose, α-<sc>D</sc>-Galacturonic acid and α-Kdo are the most frequent monosaccharides that are unique to bacteria and, except for Rhamnose, are preferably located in the core portions of bacterial saccharides, in accordance with the typical lipopolysaccharide (LPS) structure of Gram-negative bacteria, the classes which dominate in this analysis (Table ##TAB##1##2##).</p>", "<p>A more detailed analysis of monosaccharide residues in the bacterial taxonomy groups of Set 2 is shown in Fig. ##FIG##5##6##. Kdo and <sc>L</sc>-<italic>glycero</italic>-<sc>D</sc>-<italic>manno</italic>-Heptose are confined to Gram-negative bacteria, whereas Gram-positive bacteria seem to have an excess of arabinoses and methylated hexoses.</p>", "<p>Generally, monosaccharides that are unique to bacteria are of special interest as potential immunogenic targets. Existing vaccines frequently take advantage of the unique saccharides in the complex carbohydrates located on the surface of bacteria [##REF##17460666##11##]. Fig. ##FIG##6##7## presents the unique monosaccharides (see definition of <italic>unique </italic>in the Figure legend) found for the bacterial and mammalian groups. Due to the greater diversity of bacterial monosaccharides, many carbohydrates unique to the bacterial world were found (especially for Gram-positive bacteria), whereas only two mammalian monosaccharides [α-<italic>N</italic>-Glycoloylneuraminic acid (α-Neu5Gc) and β-<sc>D</sc>-<italic>N</italic>-acetylglucosamine-6-<italic>O</italic>-sulfate (β-<sc>D</sc>-GlcpNAc-6S)] appear to have no counterpart in the bacterial world. It is known that neuraminic acid derivatives are typically found at the terminal positions of mammalian glycoconjugates, being mediators for cell-cell interaction or receptors for pathogens [##REF##17420276##17##]. The presence of exposed α-Neu5Ac residues in bacteria may be an evolutionary advantage through which bacteria mask themselves to the host immune system.</p>", "<p>Attention should be paid to the distribution of monosaccharides at the terminal positions of oligomers and side chains of polymers. In higher organisms such residues are optimally positioned to mediate recognition by endogenous carbohydrate-binding proteins [##REF##10406840##9##]. According to our findings bacterial carbohydrates often have glucose residues at the nonreducing ends, in contrast to mammalian glycans (data not shown). This may be the result of the evolutionary adaptation of bacteria since exposed terminal glucose residues are important for the adherence of bacteria and entry into host epithelial cells, as demonstrated for <italic>Salmonella </italic>and <italic>Pseudomonas </italic>[##REF##16495526##18##].</p>", "<p>Fig. ##FIG##7##8## demonstrates that more than 70% of the monosaccharides in every taxonomic group are reported to be in the pyranose form, with most groups even reaching 90%. An interesting finding is that more than 50% of all furanose residues found in bacteria are in the 395 glycan sequences of the class <italic>Actinobacteria </italic>(cf. area of bars in Fig. ##FIG##7##8B##). Nearly 20% of all residues in <italic>Actinobacteria </italic>glycans are in the furanose form, compared to &lt; 4% for all other bacterial groups studied. A high proportion of furanose residues has also been found for plants (data not shown). In the majority of cases where linear forms or rings of unknown size are found, they can be explained as artifacts of the structure elucidation process, especially when present at the reducing end. However, linear monosaccharides are known to occur occasionally in bacterial carbohydrate sequences and are most prevalent in <italic>Bacilli </italic>and <italic>Actinobacteria </italic>(Fig. ##FIG##7##8B##).</p>", "<title>Monosaccharide modifications</title>", "<p>Part of the diversity of monosaccharides can be found in their modifications (Fig. ##FIG##8##9##). For bacteria secondary modifications often play a role in the mediation of reactivity and lability to various environmental conditions such as pH. The <italic>N</italic>-acetylamino group is the most common substituent for carbohydrates in mammals (ca. 45% of all residues) and in most bacteria classes (ca. 18–21% of all residues), except for <italic>α-Proteobacteria </italic>(11%) and <italic>Actinobacteria </italic>(4.5%). Acetylation of amino groups plays a key role in regulating the ability of amino sugars to form hydrogen bonds and to bear charge [##REF##12045109##19##].</p>", "<p><italic>O</italic>-methylation is the most frequent modification for <italic>Actinobacteria </italic>(ca. 23% of all residues, mainly at O6 of glucose) but occurs with a frequency of &lt; 5% in other bacteria classes and is essentially absent in mammals.</p>", "<p><italic>O</italic>-acetyl, amino, or phosphate substituents are also much more prevalent for bacteria (4–7%) than for mammals (&lt; 1%), and the <italic>O</italic>-acetylation pattern is often different for different cultures of a single bacteria strain. <italic>O</italic>-acetyl groups mask the protective epitopes for bacteria through steric hindrance or altered conformations, as shown for <italic>Meningococci </italic>[##REF##17376859##20##].</p>", "<p>Amino sugars with free aminogroups are present in about 7% of bacterial carbohydrate residues compared to ca. 1% for mammals and feature a positively charged -NH₃⁺substituent at neutral pH. The occurrence of these residues in the bacterial cell wall affects hydrophobicity and makes bacteria resistant to the lysozyme of the host, as has been demonstrated for glucosamine in Gram-positive bacteria [##REF##6766437##21##]. Several secondary modifications appear to be unique for bacterial carbohydrates (pyruvate, lactate, ethanolamine, <italic>O</italic>-methyl and formyl) while sulfation or <italic>N</italic>-glycolyl substitution occurs primarily in mammals. Finally, about 7% of bacterial residues have modifications listed under the category \"other\" in Fig. ##FIG##8##9A##, with <italic>Actinobacteria </italic>and <italic>Bacilli </italic>showing the highest frequencies (Fig. ##FIG##8##9B##).</p>", "<title>Disaccharide fragment patterns in bacteria and mammals</title>", "<p>The topological characteristics of glycan architecture can be described by statistics which document the frequency distributions for specific neighboring pairs of monosaccharides connected either with <italic>any </italic>type of linkage (monosaccharide pair analysis) or via <italic>specific </italic>linkage positions (disaccharide pattern analysis). The matrix diagram in Fig. ##FIG##9##10## illustrates the statistics of linked monosaccharide pairs for bacteria. Here the frequencies of any type of glycosidic linkage between the 20 most common donor and acceptor residues are shown. The areas of the circles plotted at the coordinates for a given pair represent its relative abundance within a given bacterial taxonomic group. Note that not all possible monosaccharide pairs are actually found in the natural sequences registered in the database (missing circles). Some combinations are exclusive for Gram-positive bacteria, e.g. those involving α-<sc>D</sc>-Ara<italic>f </italic>or α-<sc>D</sc>-Glc<italic>p</italic>6Me in <italic>Actinobacteria </italic>while others may exhibit similar or widely differing abundances across the taxonomic groups. The high abundances found along the diagonal of the matrix stem from homopolymeric subsequences which are frequent in bacteria. Note that the results for \"Kdo\" (without anomeric configuration) originate from analytical artifacts. Detailed results for a total of 676 pairs in bacterial and mammalian carbohydrates are summarized in the additional material section [see Additional file ##SUPPL##3##4##].</p>", "<p>In order to describe carbohydrate sequences at a higher level of complexity, we need to consider not only the identities of the linked monosaccharides but also the linkage configuration. All free hydroxyl groups on each acceptor monosaccharide are potential sites of glycosyltransferase reactions. Therefore, we define the child (donor) to parent (acceptor) connection in terms of the directed glycosylation linkage pattern, analogous to reaction patterns described elsewhere [##REF##16159923##22##]. Thus, the descriptor \"a1–4\", for example, indicates that an alpha anomeric O1 of the donor is linked to C4 of the acceptor. The statistics of linkage patterns provide a direct description of the expression and activity of glycosyltransferases and the carbohydrate structure repetoire in an organism or taxonomic group. Such statistics have been employed successfully for a variety of bioinformatic tasks with the glycome, e.g. matrix generation [##REF##15585530##23##] and pattern detection [##UREF##3##24##]. This kind of information is also valuable for recognizing both unique and common linkages and can serve as a basis for a deeper understanding of the immunogenicity of bacterial carbohydrates and for designing targeted vaccines.</p>", "<p>The analysis summarized in Fig. ##FIG##10##11## demonstrates that the most prevalent linkages in mammals are <sc>D</sc>-Gal, <sc>D</sc>-Man and <sc>D</sc>-GlcNAc as β1–4 donors to <sc>D</sc>-GlcNAc; <sc>D</sc>-GlcNAc as β1–2 donor to <sc>D</sc>-Man; <sc>D</sc>-GlcNAc as β1–3 donor to <sc>D</sc>-Gal; and <sc>D</sc>-Man disaccharides with α1–3 or α1–6 linkages. These monosaccharides are sufficient to build up the common <italic>N</italic>- and <italic>O-</italic>glycan structures which dominate the mammalian database.</p>", "<p>A complete list of mammalian disaccharide fragments found in GLYCOSCIENCES.de is available [see Additional file ##SUPPL##4##5##], where the data are encoded to illustrate the differences between our database findings and the data from Ohtsubo &amp; Marth [##REF##16959566##2##]. Three disaccharides reported in [##REF##16959566##2##], namely <sc>D</sc>-Glc(α1–2) <sc>D</sc>-Gal, <sc>D</sc>-GlcA(β1–4) <sc>D</sc>-Gal and <sc>D</sc>-GlcNAc(α1–6) <sc>D</sc>-GlcNAc, are absent from our databases. On the other hand, many existing mammalian disaccharides have not been mentioned by Ohtsubo &amp; Marth, among them reasonably abundant ones such as <sc>D</sc>-GlcNAc(β1–3) <sc>D</sc>-GalNAc, <sc>D</sc>-GlcNAc(β1–4) <sc>D</sc>-Man, <sc>D</sc>-GlcNAc(β1–6) <sc>D</sc>-Man, and <sc>D</sc>-GlcNS(α1–4) <sc>L</sc>-IdoA. The last disaccharide listed is present in more than 500 records of GLYCOSCIENCES.DE for human carbohydrates and reported in the literature in association with Sandhoff's Disease [##REF##412673##25##].</p>", "<p>The disaccharide ensemble or total count of unique linkages is considerably larger for most of the bacterial classes compared to mammals, as shown in Table ##TAB##3##4##. For this analysis we have taken a total of 24613 bacterial and 23883 mammalian disaccharide fragments into account.</p>", "<p>The abundance data for the more frequently occurring disaccharide linkage patterns found in mammals and/or bacteria are presented in Table ##TAB##4##5## (detected in the combined database using an abundance threshold = 0.1% for either mammals or bacteria). More comprehensive data (with lower threshold) are presented in the additional material [see Additional file ##SUPPL##5##6##]. In Table ##TAB##4##5## there are 26 disaccharide linkages listed which are found only in bacteria. The four most frequent of these, with abundances of 0.40–0.68% (underlined in Table ##TAB##4##5##) are <sc>D</sc>-Glc(α1–4) <sc>D</sc>-Gal, <sc>D</sc>-Gal(α1–2) <sc>D</sc>-Gal, <sc>D</sc>-Glc(b1–4) <sc>D</sc>-Gal, and <sc>D</sc>-Gal(α1–2) <sc>D</sc>-Man. In Table ##TAB##4##5## there are five disaccharide types which occur in mammals only, and the most abundant of these (underlined in Table ##TAB##4##5##) involve (a) α2–6 linkages from a neuraminic acid (Neu5Ac or Neu5Gc) to <sc>D</sc>-GalNAc or (b) α1–4 linkages from a uronic acid (GlcA, IdoA, or ΔGlcA) to <sc>D</sc>-GlcN-sulfate.</p>", "<p>The abundances of the most common bacterial disaccharide fragments are shown schematically in Fig. ##FIG##11##12##, where separate diagrams are presented for oligomers (A) and polymers (B) using the same color-coded linkage scheme as in Fig. ##FIG##10##11##. The residue names are sorted according to their abundance either as donors (children) or acceptors (parents). The highest abundances in bacterial oligomers (Fig. ##FIG##11##12A##) are exhibited by the constituents of the bacterial lipopolysaccharide core region: <sc>L</sc>-gro-<sc>D</sc>-manHep→Kdo and <sc>L</sc>-gro-<sc>D</sc>-manHep→<sc>L</sc>-gro-<sc>D</sc>-manHep. High abundances for <sc>L</sc>-Rha→<sc>L</sc>-Rha, <sc>D</sc>-Man→<sc>D</sc>-Man, <sc>D</sc>-Gal→<sc>D</sc>-Gal and <sc>D</sc>-Glc→<sc>D</sc>-Glc in Fig. ##FIG##11##12B## arise from the homopolymeric regions prevalent in bacterial polymers.</p>", "<p>A potential application of the information and methods outlined here is the design and validation of carbohydrate vaccines against bacterial pathogens. Carbohydrate-based vaccines against <italic>Haemophilus influenzae </italic>Type b, <italic>Neisseria meningitidis </italic>and <italic>Streptococcus pneumoniae </italic>have already been licensed, and many similar products are in various stages of development. For example, the disaccharides <sc>D</sc>-Glc(α1–2)<sc>D</sc>-Gal and <sc>D</sc>-Glc(β1–4)<sc>D</sc>-Gal are not present in mammalian organisms accordingly to our analysis and are both constituents of the capsular polysaccharides of <italic>Salmonella pneumonia</italic>, which were shown to be target candidates for vaccine development [##REF##15895165##26##].</p>" ]

[ "<title>Results and Discussion</title>", "<title>Distribution of carbohydrate structures among taxonomic groups</title>", "<p>We first examined the number of sequences found in the BCSDB and GLYCOSCIENCES.de for various taxonomic ranks (class, order, family). Where possible, the taxonomic relationships were traced using the NCBI taxonomy database [##REF##10592169##15##]. The GLYCOSCIENCES.de database currently contains a total of 23120 glycan and glycoconjugate records, of which 13704 records for diverse animal, plant, bacteria and fungi classes have some information concerning taxonomy. In the BCSDB there are a total of 8504 records for bacteria only, and 8479 of these contain information concerning taxonomy. These numbers may include multiple records for a given glycan when the same glycan is reported for more than one species. Note that not all taxonomic classes are represented in the databases and that for bacterial glycans there is considerable overlap between the two databases.</p>", "<p>If we now consider the two databases combined, there are a total of 13775 nonredundant carbohydrate records which include taxonomic information. The distribution of these records among various taxonomic classes is shown both numerically and schematically in Fig. ##FIG##0##1##. A more detailed breakdown of the distribution can be found in the additional material [see Additional file ##SUPPL##0##1##].</p>", "<p>The taxonomic class <italic>Mammalia </italic>is found to have 4739 assigned sequence/taxon pairs, of which 2118 are of human origin (family <italic>Hominidae</italic>). All other animal or plant classes in the database have less than 350 pairs. The category \"unresolved\" refers to the 1482 records for which the source is defined but the specific taxonomic class could not be traced automatically using the NCBI. Only about half of the bacterial phyla are represented in the BCSDB with a total of 6098 sequence/taxon pairs, and nine classes have less than 10 records. Note that the number of carbohydrates or glycoconjugates registered for a given taxonomic class does not necessarily reflect its species diversity, but more likely the intensity with which the class has been studied. Thus, the apparent diversity of carbohydrates in the various taxonomic classes reflects to a large part the information bias in the published literature, and this situation must be kept in mind when making conclusions based on the distributions of properties discussed below.</p>", "<p>In the combined databases there are a total of 12659 records in the category \"no taxonomy\" which means that either no information concerning the taxonomy of the source is available or that the carbohydrate is not of purely natural origin. These records were not included in Fig. ##FIG##0##1## and were not used in the following analyses.</p>", "<title>Choice of taxonomic datasets for statistical comparisons</title>", "<p>For the following more detailed statistical comparisons, we defined two sets of taxonomic <italic>groups</italic>, considering both biological and coverage aspects. Taxonomy Set 1 (Table ##TAB##0##1##) was defined to provide an overview of the total content of the two databases used for the general comparison of bacterial and mammalian carbohydrates, taking into account the fact that bacterial carbohydrates frequently contain repeating units while mammalian sequences usually do not. Thus, Set 1 contains three taxonomic groups: all mammalian carbohydrates, all bacterial carbohydrates with nonrepeating sequences (oligomers), and all bacterial sequences with repeating units (polymers).</p>", "<p>For comparisons within the taxonomic domain <italic>Bacteria</italic>, we defined a more detailed taxonomy Set 2 (Table ##TAB##1##2##), which includes two classes of Gram-positive bacteria (<italic>Actinobacteria </italic>and <italic>Bacilli</italic>) and the various classes of the phylum <italic>Proteobacteria</italic>. The largest of these classes, the γ-<italic>Proteobacteria</italic>, has been further subdivided in Set 2 into the major order <italic>Enterobacteriales </italic>and a subset containing all other γ-<italic>Proteobacteria</italic>. The class δ-<italic>Proteobacteria </italic>(with only 2 records) has been combined with the ε-<italic>Proteobacteria</italic>.</p>", "<p>In order to obtain meaningful statistics, only those taxonomic groups are compared for which at least 200 carbohydrate sequences are available. For this reason the classes Chlamydiae, Clostridia, and Bacteroidetes, for example, have not been included in Set 2. Note that in Tables ##TAB##0##1## and ##TAB##1##2## the total number of <italic>unique </italic>carbohydrate <italic>sequences </italic>in each group is listed, and these sets were utilized in all subsequent analyses.</p>", "<title>Carbohydrate size, branching and charge density</title>", "<p>Frequency distributions for general measures of molecular size, topology (branching) and mean charge density have been calculated for the carbohydrate sequences comprising the various taxonomic groupings described by Set 1 and Set 2 (Tables ##TAB##0##1## and ##TAB##1##2##). In each case the distributions are normalized to the <italic>total number of sequences in each taxonomic group </italic>and expressed as percentages within each group. In Fig. ##FIG##1##2A## distributions for the number of monosaccharides per sequence <italic>unit </italic>(either the entire carbohydrate sequence for oligomers or the repeating unit for polymers, see Definition 4 in the methods) are shown for bacteria vs. mammals (taxonomy Set 1). The distribution is relatively broad for mammals with mean and median values, respectively, of 8.17 and 8 monomers per sequence, while for bacteria the distribution shows a narrow peak at 4–5 monomers for both oligomers (mean: 5.94 median: 5) and for the repeating unit of polymers (mean: 4.17, median: 4). However, oligomers show a significant population of sequences with 8–15 monomers while the distribution for polymers essentially ends at 9 monomers per unit. Of course, the total length of a polymeric sequence with multiple units may very well exceed the maximum length of oligomers. Naturally occurring oligomers may also be longer than the sequences reported in the databases since the process of extracting and isolating a glycan may result in partial digestion and loss of residues.</p>", "<p>In Fig. ##FIG##1##2B## the distributions of the size parameter for the bacterial groups defined in taxonomy Set 2 are found to differ considerably. Narrow distributions with essentially a single prominant peak are found for <italic>Actinobacteria </italic>(mean: 4.51, median: 3), <italic>Bacilli </italic>(mean: 5.18, median: 5) and the order <italic>Enterobacteria </italic>(mean: 5.18, median: 6) with peaks at ca. 2.5, 5.5 and 4.5 residues, respectively. The various other classes of <italic>Proteobacteria </italic>have broader distributions with more or less pronounced multiple peaks, e.g. at 3, 8 and 11 residues for the δ,ε-Proteobacteria group.</p>", "<p>The number of branching points per carbohydrate residue can be considered to be a <italic>branching index </italic>which reflects the complexity of carbohydrate topology. Fig. ##FIG##2##3A## demonstrates that 22% of all mammalian and 50% of all bacterial sequences are linear (branching index = 0). However, for the individual bacterial groups of taxonomy Set 2, the percentage of linear structures ranges from 30% to 78% (Fig. ##FIG##2##3B##). A general feature of all branching point graphs in Fig. ##FIG##2##3## is a peak in the distribution at a branching index of 0.2 – 0.3, which corresponds to carbohydrate sequences with one branching point for every three to five monosaccharide residues. This peak in the distribution is weak for <italic>Actinobacteria </italic>and α-<italic>Proteobacteria </italic>but strong for mammals, <italic>Bacilli</italic>, and other <italic>Proteobacteria</italic>.</p>", "<p>Finally, the <italic>mean charge density </italic>parameter (max. electric charge possible for all ionizable groups divided by the number of carbohydrate residues in a sequence unit) is shown in Fig. ##FIG##3##4## for taxonomy Set 1 and Set 2. About 58% of mammalian sequences and 47% of all bacterial carbohydrate sequence units have no net charge (Fig. ##FIG##3##4A##). For the bacterial groups of Set 2, the frequency of sequence units with no net charge ranges from ca. 32% for β-<italic>Proteobacteria </italic>to 77% for <italic>Actinobacteria </italic>(Fig. ##FIG##3##4B##). All other sequences have a net negative charge due to carboxyl, phosphate or sulfate groups, for example. The distributions for mammals and <italic>Bacilli </italic>both have peaks at charge densities of -0.2 and -0.5; the δ,ε-<italic>Proteobacteria </italic>distribution exhibits a single broad peak at ca. -0.3 while the other bacterial distributions have multiple peaks at -0.3, -0.5, -0.7 and -1.0 (Fig. ##FIG##3##4B##).</p>", "<title>Monosaccharide diversity</title>", "<p>For mammals and even more so for bacteria, the diversity of the monosaccharides used as the building blocks of carbohydrate sequences is significantly larger than that for the residues in proteins or nucleic acids. From the GLYCOSCIENCES.de database a total of 35 different monosaccharides were found for mammalian carbohydrates, according to the nomenclature used in the original databases (Table ##TAB##2##3##). This degree of diversity is at first glance puzzling, in view of the common notion that mammalian carbohydrates are built up of 10 \"classical\" monosaccharides (Glc, Gal, GlcNAc, GalNAc, Man, GlcA, Fuc, Neu, IdoA, Xyl) [##UREF##0##1##]. However, the variety of monosaccharides defined in the primary databases is higher due to (a) residues being specified with unknown anomer or ring type definitions, (b) analytical artifacts from the structure elucidation process (alditols, double bonds), or (c) secondary modifications such as sulfation. Furthermore, carbohydrate sequence databases are not error-free and suffer from incorrect structure elucidations and curation mistakes. Since the existing databases generally use free-text identifiers for the monosaccharides, it was helpful to translate all structural database entries into a machine-readable notation called GlycoCT [##REF##18436199##16##]. Using structural filters based on this notation we were able to significantly reduce the fuzziness introduced by the lack of strictness in the definitions of the original sequences (Table ##TAB##2##3##). During the analysis we excluded manually common artifacts caused by analytical procedures and entries with undefined absolute or anomeric configuration or ring type.</p>", "<p>To minimize the influence of errors and artifacts on the statistics of Table ##TAB##2##3##, a threshold for the occurrence of monosaccharides, basetypes and basic entities was set at 10 for mammals and 2 for bacteria. This means that a given residue type was included in the statistics only when its number of occurrences exceeded the defined threshold. A relatively low threshold was chosen for bacteria because, in contrast to mammals, bacteria are known to produce a great variety of unique monosaccharide residues with low occurrence. When the threshold for bacteria was reduced from 2 to 0, the diversity of detected residues increased by about 25%. A complete list of monosaccharide residues and basetypes found for each taxonomic group is available in GlycoCT nomenclature in the additional material [see Additional files ##SUPPL##1##2## (tables a-i) and ##SUPPL##2##3##].</p>", "<p>For mammals the analysis returned 18 occurrences of <sc>D</sc>-Fuc as a basic entity. However, this residue was excluded from Table ##TAB##2##3## because all carbohydrates in GLYCOSCIENCES.de which are specified to contain <sc>D</sc>-Fucose originate from old publications in which the absolute configuration of Fuc was not specified. The occurrence of <sc>D</sc>-Fuc in mammalian carbohydrate records can be regarded as a data translation error since there is no evidence for a mammalian enzyme which synthesizes <sc>D</sc>-Fuc.</p>", "<p>The 35 monosaccharides with the highest occurrence within taxonomy Set 1 are shown schematically in Fig. ##FIG##4##5##. Of them, all 17 monosaccharides that were found in mammalian carbohydrates are also found in the bacterial world. Rhamnose, <sc>L</sc>-<italic>glycero</italic>-α-<sc>D</sc>-<italic>manno</italic>-Heptose, α-<sc>D</sc>-Galacturonic acid and α-Kdo are the most frequent monosaccharides that are unique to bacteria and, except for Rhamnose, are preferably located in the core portions of bacterial saccharides, in accordance with the typical lipopolysaccharide (LPS) structure of Gram-negative bacteria, the classes which dominate in this analysis (Table ##TAB##1##2##).</p>", "<p>A more detailed analysis of monosaccharide residues in the bacterial taxonomy groups of Set 2 is shown in Fig. ##FIG##5##6##. Kdo and <sc>L</sc>-<italic>glycero</italic>-<sc>D</sc>-<italic>manno</italic>-Heptose are confined to Gram-negative bacteria, whereas Gram-positive bacteria seem to have an excess of arabinoses and methylated hexoses.</p>", "<p>Generally, monosaccharides that are unique to bacteria are of special interest as potential immunogenic targets. Existing vaccines frequently take advantage of the unique saccharides in the complex carbohydrates located on the surface of bacteria [##REF##17460666##11##]. Fig. ##FIG##6##7## presents the unique monosaccharides (see definition of <italic>unique </italic>in the Figure legend) found for the bacterial and mammalian groups. Due to the greater diversity of bacterial monosaccharides, many carbohydrates unique to the bacterial world were found (especially for Gram-positive bacteria), whereas only two mammalian monosaccharides [α-<italic>N</italic>-Glycoloylneuraminic acid (α-Neu5Gc) and β-<sc>D</sc>-<italic>N</italic>-acetylglucosamine-6-<italic>O</italic>-sulfate (β-<sc>D</sc>-GlcpNAc-6S)] appear to have no counterpart in the bacterial world. It is known that neuraminic acid derivatives are typically found at the terminal positions of mammalian glycoconjugates, being mediators for cell-cell interaction or receptors for pathogens [##REF##17420276##17##]. The presence of exposed α-Neu5Ac residues in bacteria may be an evolutionary advantage through which bacteria mask themselves to the host immune system.</p>", "<p>Attention should be paid to the distribution of monosaccharides at the terminal positions of oligomers and side chains of polymers. In higher organisms such residues are optimally positioned to mediate recognition by endogenous carbohydrate-binding proteins [##REF##10406840##9##]. According to our findings bacterial carbohydrates often have glucose residues at the nonreducing ends, in contrast to mammalian glycans (data not shown). This may be the result of the evolutionary adaptation of bacteria since exposed terminal glucose residues are important for the adherence of bacteria and entry into host epithelial cells, as demonstrated for <italic>Salmonella </italic>and <italic>Pseudomonas </italic>[##REF##16495526##18##].</p>", "<p>Fig. ##FIG##7##8## demonstrates that more than 70% of the monosaccharides in every taxonomic group are reported to be in the pyranose form, with most groups even reaching 90%. An interesting finding is that more than 50% of all furanose residues found in bacteria are in the 395 glycan sequences of the class <italic>Actinobacteria </italic>(cf. area of bars in Fig. ##FIG##7##8B##). Nearly 20% of all residues in <italic>Actinobacteria </italic>glycans are in the furanose form, compared to &lt; 4% for all other bacterial groups studied. A high proportion of furanose residues has also been found for plants (data not shown). In the majority of cases where linear forms or rings of unknown size are found, they can be explained as artifacts of the structure elucidation process, especially when present at the reducing end. However, linear monosaccharides are known to occur occasionally in bacterial carbohydrate sequences and are most prevalent in <italic>Bacilli </italic>and <italic>Actinobacteria </italic>(Fig. ##FIG##7##8B##).</p>", "<title>Monosaccharide modifications</title>", "<p>Part of the diversity of monosaccharides can be found in their modifications (Fig. ##FIG##8##9##). For bacteria secondary modifications often play a role in the mediation of reactivity and lability to various environmental conditions such as pH. The <italic>N</italic>-acetylamino group is the most common substituent for carbohydrates in mammals (ca. 45% of all residues) and in most bacteria classes (ca. 18–21% of all residues), except for <italic>α-Proteobacteria </italic>(11%) and <italic>Actinobacteria </italic>(4.5%). Acetylation of amino groups plays a key role in regulating the ability of amino sugars to form hydrogen bonds and to bear charge [##REF##12045109##19##].</p>", "<p><italic>O</italic>-methylation is the most frequent modification for <italic>Actinobacteria </italic>(ca. 23% of all residues, mainly at O6 of glucose) but occurs with a frequency of &lt; 5% in other bacteria classes and is essentially absent in mammals.</p>", "<p><italic>O</italic>-acetyl, amino, or phosphate substituents are also much more prevalent for bacteria (4–7%) than for mammals (&lt; 1%), and the <italic>O</italic>-acetylation pattern is often different for different cultures of a single bacteria strain. <italic>O</italic>-acetyl groups mask the protective epitopes for bacteria through steric hindrance or altered conformations, as shown for <italic>Meningococci </italic>[##REF##17376859##20##].</p>", "<p>Amino sugars with free aminogroups are present in about 7% of bacterial carbohydrate residues compared to ca. 1% for mammals and feature a positively charged -NH₃⁺substituent at neutral pH. The occurrence of these residues in the bacterial cell wall affects hydrophobicity and makes bacteria resistant to the lysozyme of the host, as has been demonstrated for glucosamine in Gram-positive bacteria [##REF##6766437##21##]. Several secondary modifications appear to be unique for bacterial carbohydrates (pyruvate, lactate, ethanolamine, <italic>O</italic>-methyl and formyl) while sulfation or <italic>N</italic>-glycolyl substitution occurs primarily in mammals. Finally, about 7% of bacterial residues have modifications listed under the category \"other\" in Fig. ##FIG##8##9A##, with <italic>Actinobacteria </italic>and <italic>Bacilli </italic>showing the highest frequencies (Fig. ##FIG##8##9B##).</p>", "<title>Disaccharide fragment patterns in bacteria and mammals</title>", "<p>The topological characteristics of glycan architecture can be described by statistics which document the frequency distributions for specific neighboring pairs of monosaccharides connected either with <italic>any </italic>type of linkage (monosaccharide pair analysis) or via <italic>specific </italic>linkage positions (disaccharide pattern analysis). The matrix diagram in Fig. ##FIG##9##10## illustrates the statistics of linked monosaccharide pairs for bacteria. Here the frequencies of any type of glycosidic linkage between the 20 most common donor and acceptor residues are shown. The areas of the circles plotted at the coordinates for a given pair represent its relative abundance within a given bacterial taxonomic group. Note that not all possible monosaccharide pairs are actually found in the natural sequences registered in the database (missing circles). Some combinations are exclusive for Gram-positive bacteria, e.g. those involving α-<sc>D</sc>-Ara<italic>f </italic>or α-<sc>D</sc>-Glc<italic>p</italic>6Me in <italic>Actinobacteria </italic>while others may exhibit similar or widely differing abundances across the taxonomic groups. The high abundances found along the diagonal of the matrix stem from homopolymeric subsequences which are frequent in bacteria. Note that the results for \"Kdo\" (without anomeric configuration) originate from analytical artifacts. Detailed results for a total of 676 pairs in bacterial and mammalian carbohydrates are summarized in the additional material section [see Additional file ##SUPPL##3##4##].</p>", "<p>In order to describe carbohydrate sequences at a higher level of complexity, we need to consider not only the identities of the linked monosaccharides but also the linkage configuration. All free hydroxyl groups on each acceptor monosaccharide are potential sites of glycosyltransferase reactions. Therefore, we define the child (donor) to parent (acceptor) connection in terms of the directed glycosylation linkage pattern, analogous to reaction patterns described elsewhere [##REF##16159923##22##]. Thus, the descriptor \"a1–4\", for example, indicates that an alpha anomeric O1 of the donor is linked to C4 of the acceptor. The statistics of linkage patterns provide a direct description of the expression and activity of glycosyltransferases and the carbohydrate structure repetoire in an organism or taxonomic group. Such statistics have been employed successfully for a variety of bioinformatic tasks with the glycome, e.g. matrix generation [##REF##15585530##23##] and pattern detection [##UREF##3##24##]. This kind of information is also valuable for recognizing both unique and common linkages and can serve as a basis for a deeper understanding of the immunogenicity of bacterial carbohydrates and for designing targeted vaccines.</p>", "<p>The analysis summarized in Fig. ##FIG##10##11## demonstrates that the most prevalent linkages in mammals are <sc>D</sc>-Gal, <sc>D</sc>-Man and <sc>D</sc>-GlcNAc as β1–4 donors to <sc>D</sc>-GlcNAc; <sc>D</sc>-GlcNAc as β1–2 donor to <sc>D</sc>-Man; <sc>D</sc>-GlcNAc as β1–3 donor to <sc>D</sc>-Gal; and <sc>D</sc>-Man disaccharides with α1–3 or α1–6 linkages. These monosaccharides are sufficient to build up the common <italic>N</italic>- and <italic>O-</italic>glycan structures which dominate the mammalian database.</p>", "<p>A complete list of mammalian disaccharide fragments found in GLYCOSCIENCES.de is available [see Additional file ##SUPPL##4##5##], where the data are encoded to illustrate the differences between our database findings and the data from Ohtsubo &amp; Marth [##REF##16959566##2##]. Three disaccharides reported in [##REF##16959566##2##], namely <sc>D</sc>-Glc(α1–2) <sc>D</sc>-Gal, <sc>D</sc>-GlcA(β1–4) <sc>D</sc>-Gal and <sc>D</sc>-GlcNAc(α1–6) <sc>D</sc>-GlcNAc, are absent from our databases. On the other hand, many existing mammalian disaccharides have not been mentioned by Ohtsubo &amp; Marth, among them reasonably abundant ones such as <sc>D</sc>-GlcNAc(β1–3) <sc>D</sc>-GalNAc, <sc>D</sc>-GlcNAc(β1–4) <sc>D</sc>-Man, <sc>D</sc>-GlcNAc(β1–6) <sc>D</sc>-Man, and <sc>D</sc>-GlcNS(α1–4) <sc>L</sc>-IdoA. The last disaccharide listed is present in more than 500 records of GLYCOSCIENCES.DE for human carbohydrates and reported in the literature in association with Sandhoff's Disease [##REF##412673##25##].</p>", "<p>The disaccharide ensemble or total count of unique linkages is considerably larger for most of the bacterial classes compared to mammals, as shown in Table ##TAB##3##4##. For this analysis we have taken a total of 24613 bacterial and 23883 mammalian disaccharide fragments into account.</p>", "<p>The abundance data for the more frequently occurring disaccharide linkage patterns found in mammals and/or bacteria are presented in Table ##TAB##4##5## (detected in the combined database using an abundance threshold = 0.1% for either mammals or bacteria). More comprehensive data (with lower threshold) are presented in the additional material [see Additional file ##SUPPL##5##6##]. In Table ##TAB##4##5## there are 26 disaccharide linkages listed which are found only in bacteria. The four most frequent of these, with abundances of 0.40–0.68% (underlined in Table ##TAB##4##5##) are <sc>D</sc>-Glc(α1–4) <sc>D</sc>-Gal, <sc>D</sc>-Gal(α1–2) <sc>D</sc>-Gal, <sc>D</sc>-Glc(b1–4) <sc>D</sc>-Gal, and <sc>D</sc>-Gal(α1–2) <sc>D</sc>-Man. In Table ##TAB##4##5## there are five disaccharide types which occur in mammals only, and the most abundant of these (underlined in Table ##TAB##4##5##) involve (a) α2–6 linkages from a neuraminic acid (Neu5Ac or Neu5Gc) to <sc>D</sc>-GalNAc or (b) α1–4 linkages from a uronic acid (GlcA, IdoA, or ΔGlcA) to <sc>D</sc>-GlcN-sulfate.</p>", "<p>The abundances of the most common bacterial disaccharide fragments are shown schematically in Fig. ##FIG##11##12##, where separate diagrams are presented for oligomers (A) and polymers (B) using the same color-coded linkage scheme as in Fig. ##FIG##10##11##. The residue names are sorted according to their abundance either as donors (children) or acceptors (parents). The highest abundances in bacterial oligomers (Fig. ##FIG##11##12A##) are exhibited by the constituents of the bacterial lipopolysaccharide core region: <sc>L</sc>-gro-<sc>D</sc>-manHep→Kdo and <sc>L</sc>-gro-<sc>D</sc>-manHep→<sc>L</sc>-gro-<sc>D</sc>-manHep. High abundances for <sc>L</sc>-Rha→<sc>L</sc>-Rha, <sc>D</sc>-Man→<sc>D</sc>-Man, <sc>D</sc>-Gal→<sc>D</sc>-Gal and <sc>D</sc>-Glc→<sc>D</sc>-Glc in Fig. ##FIG##11##12B## arise from the homopolymeric regions prevalent in bacterial polymers.</p>", "<p>A potential application of the information and methods outlined here is the design and validation of carbohydrate vaccines against bacterial pathogens. Carbohydrate-based vaccines against <italic>Haemophilus influenzae </italic>Type b, <italic>Neisseria meningitidis </italic>and <italic>Streptococcus pneumoniae </italic>have already been licensed, and many similar products are in various stages of development. For example, the disaccharides <sc>D</sc>-Glc(α1–2)<sc>D</sc>-Gal and <sc>D</sc>-Glc(β1–4)<sc>D</sc>-Gal are not present in mammalian organisms accordingly to our analysis and are both constituents of the capsular polysaccharides of <italic>Salmonella pneumonia</italic>, which were shown to be target candidates for vaccine development [##REF##15895165##26##].</p>" ]

[ "<title>Conclusion</title>", "<p>In this study we combined the BCSDB, the largest available bacterial carbohydrate database, with the GLYCOSCIENCES.de database to obtain a set of 13775 nonredundant glycan/taxon pairs (carbohydrate sequences with a defined taxonomy), of which 6098 were assigned to <italic>Bacteria </italic>and 4739 to <italic>Mammalia</italic>. The representative statistical analyses presented here reveal the basic principles of carbohydrate architecture in bacteria vs. mammals. The major monosaccharides which characterize different branches of the tree of life were extracted from the database and are in accordance with the published literature. Several monosaccharides unique to certain subclasses of bacteria were identified and could prove useful as molecular markers for these classes. Similarly, a variety of structural modifications of monosaccharides have been detected, and many of these are characteristic in that they may be either highly abundant or totally absent in individual taxonomic classes.</p>", "<p>A linkage analysis was performed for all disaccharide fragments of bacterial and mammalian glycans and revealed that there are a number of abundant linkages as well as nonexistent linkages which may be useful for characterizing the various taxonomic groups. Through a comparison of the disaccharide linkage ensembles or spaces for bacteria and mammals, one obtains an overview of those glycosyltransferase activities which are common to both classes and those which appear to be unique for mammals or bacteria or even for specific bacteria subclasses. Thus, differential cross-species expression analysis is possible and may ultimately provide a deeper understanding of immunogenic patterns present in pathogenic bacteria.</p>" ]

[ "<p>This is an Open Access article distributed under the terms of the Creative Commons Attribution License (<ext-link ext-link-type=\"uri\" xlink:href=\"http://creativecommons.org/licenses/by/2.0\"/>), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.</p>", "<title>Background</title>", "<p>There are considerable differences between bacterial and mammalian glycans. In contrast to most eukaryotic carbohydrates, bacterial glycans are often composed of repeating units with diverse functions ranging from structural reinforcement to adhesion, colonization and camouflage. Since bacterial glycans are typically displayed at the cell surface, they can interact with the environment and, therefore, have significant biomedical importance.</p>", "<title>Results</title>", "<p>The sequence characteristics of glycans (monosaccharide composition, modifications, and linkage patterns) for the higher bacterial taxonomic classes have been examined and compared with the data for mammals, with both similarities and unique features becoming evident. Compared to mammalian glycans, the bacterial glycans deposited in the current databases have a more than ten-fold greater diversity at the monosaccharide level, and the disaccharide pattern space is approximately nine times larger. Specific bacterial subclasses exhibit characteristic glycans which can be distinguished on the basis of distinctive structural features or sequence properties.</p>", "<title>Conclusion</title>", "<p>For the first time a systematic database analysis of the bacterial glycome has been performed. This study summarizes the current knowledge of bacterial glycan architecture and diversity and reveals putative targets for the rational design and development of therapeutic intervention strategies by comparing bacterial and mammalian glycans.</p>" ]

[ "<title>Authors' contributions</title>", "<p>SH carried out the data generation, programmed the analytical procedures and drafted the manuscript together with PVT, who made the statistical analyses and the figures. RR participated in the data generation and made significant contributions to the programming framework. C–WvdL, WEH and YAK participated in the design of the study and helped to draft the manuscript. All authors read and approved the final manuscript.</p>", "<title>Supplementary Material</title>" ]

[ "<title>Acknowledgements</title>", "<p>PT thanks the DKFZ for stipends supporting his stay in Heidelberg. The development of GLYCOSCIENCES.de at the DKFZ was supported by a Research Grant from the Deutsche Forschungsgemeinschaft (DFG BIB 46 HDdkz 01-01) within the digital library program (SH). The development of the BCSDB (PT) was supported by the International Science and Technology Center (Project 1197p), the Russian Foundation for Basic Research (Project 05-07-90099) and the Russian President Grant Committee (Project MK-2005.1700.4). The EUROCarbDB project, supported by the EU (6th Research Framework Program, RIDS contract number 011952), has also contributed resources to this analysis (RR).</p>" ]

[ "<fig position=\"float\" id=\"F1\"><label>Figure 1</label><caption><p><bold>Distribution of carbohydrate sequences for various taxonomic classes</bold>. For the combined BCSDB and GLYCOSCIENCES.de databases the pie chart sector areas correspond to the distribution (in percent) of the 13775 assigned sequences within the taxonomic classes shown, while the labels give the absolute numbers of sequences. The white pie sector contains all classes which each have &lt; 1% of the total assigned records. This category is expanded in the bar chart at the right, where the bottom block \"all other\" contains all classes which each have &lt; 0.1% of the assigned records. Class names ending with ...<italic>opsida </italic>or ...<italic>mycetes </italic>correspond to plants or fungi, respectively; <italic>Actinopterygii </italic>contains fish while <italic>Chondrichthyes </italic>contains sharks.</p></caption></fig>", "<fig position=\"float\" id=\"F2\"><label>Figure 2</label><caption><p><bold>Size distribution of carbohydrate sequence units</bold>. The normalized frequency distribution for total carbohydrate residue count per sequence unit is shown in percent of total sequences for each taxonomic group. <bold>A</bold>. For taxonomy Set 1 the solid blue curve represents the cumulative values for bacterial oligomers (blue-shaded region) and the repeating units of polymers (cyan region) in comparison with mammals (black curve). <bold>B</bold>. The frequency distribution for carbohydrate residue count per sequence unit is shown for each of the bacteria groups defined in taxonomy Set 2, using the color coding defined in the legend. For comparison the dotted curve shows the distribution for mammals. The curves are smoothed for visual clarity.</p></caption></fig>", "<fig position=\"float\" id=\"F3\"><label>Figure 3</label><caption><p><bold>Branching index distribution</bold>. The normalized frequency distributions for the number of branching points <italic>per residue </italic>are shown for the carbohydrate sequence units of taxonomy Set 1 (<bold>A</bold>) and Set 2 (<bold>B</bold>), analogous to the graphs in Fig. 2.</p></caption></fig>", "<fig position=\"float\" id=\"F4\"><label>Figure 4</label><caption><p><bold>Mean charge density distribution</bold>. Normalized frequency distributions for the maximum possible mean charge density per residue are plotted for the carbohydrate sequence units of taxonomy Set 1 (<bold>A</bold>) and Set 2 (<bold>B</bold>), analogous to the graphs in Fig. 3.</p></caption></fig>", "<fig position=\"float\" id=\"F5\"><label>Figure 5</label><caption><p><bold>Diversity of monosaccharides in bacteria and mammals</bold>. For taxonomy Set 1 the 35 most abundant monosaccharide residues are listed from left to right in decreasing order of their total abundance across all taxonomic groups. The three rows of circles correspond to the three groups: mammals (gray), bacteria (polymers, cyan), bacteria (oligomers, blue), as defined in Table 1. Circle areas reflect the relative abundance of a monosaccharide residue within each group (monosaccharide count/total residue count per taxonomic group). Residues that are the result of analytical artifacts or incomplete structure elucidation (hexosamine alditols, glycerol, D-Glc, etc.) are highlighted with gray bars. In this and subsequent Figures the anomeric designators α and β are written as a and b, the ring type designators <italic>p </italic>and <italic>f </italic>are shown as p and f.</p></caption></fig>", "<fig position=\"float\" id=\"F6\"><label>Figure 6</label><caption><p><bold>The most abundant monosaccharides in bacteria</bold>. Circle areas reflect relative abundances of the 30 most common monosaccharide residues (<bold>A</bold>) or basic entities (<bold>B</bold>) for the carbohydrates found in the BCSDB for each of the bacterial taxonomy groups of Set 2. Within each group the abundances are normalized to the total number of residues per group. The color code (see legend at bottom) is the same as in Figs. 2–4; for comparison, the open circles represent data for mammals. The residues are sorted from top to bottom in order of decreasing total abundance <italic>in bacteria </italic>(order differs from Fig. 5). The residue Kdo (without anomeric configuration) results from analytical artifacts and is highlighted in gray. For the basic entities defined in <bold>B</bold>, no distinction is made between anomeric configurations and ring types.</p></caption></fig>", "<fig position=\"float\" id=\"F7\"><label>Figure 7</label><caption><p><bold>The most abundant unique monosaccharides</bold>. Circle areas reflect relative abundances within a taxonomic group for those unique monosaccharide residues which appear exclusively, or nearly so, in a single taxonomic group of Set 1 (<bold>A</bold>) or Set 2 (<bold>B</bold>). Uniqueness is defined here as: frequency in the selected group &gt; 0.1%, frequency in other groups &lt; 0.1%. Residues that result from analytical artifacts and those that are ambiguous due to incomplete structure elucidation are highlighted in gray. The symbol <italic>X </italic>represents any substituent.</p></caption></fig>", "<fig position=\"float\" id=\"F8\"><label>Figure 8</label><caption><p><bold>Monosaccharide ring type distributions</bold>. The distribution of residue ring type (pyranose, furanose, linear, unknown) is shown schematically for various taxonomic groups. <bold>A</bold>. For taxonomy Set 1 the areas of the colored bars are proportional to the absolute occurrences (numbers shown) of a given ring type in each taxonomic group. The vertical scale is expanded in the inset. <bold>B</bold>. For taxonomy Set 2 two different ways of viewing the data are presented. (1) For a given ring type the <italic>area </italic>or <italic>height </italic>of each colored bar in a stack represents the relative abundance (%) of that ring type for each of the taxonomic groups, normalized to the total occurrence of that ring type across all groups (stack height = 100% for each ring type). Thus, the bar heights within a stack represent the distribution of a single ring type across all taxonomic groups. (2) The number labeled in each bar represents the frequency (in %) of residues with the corresponding ring type within the bar's taxonomic group, normalized to the total number of residues for that group. The numbers sum horizontally to 100% for each taxonomic group (color) and, therefore, represent the distribution of the different ring types within an individual group.</p></caption></fig>", "<fig position=\"float\" id=\"F9\"><label>Figure 9</label><caption><p><bold>Distribution of monosaccharide modifications</bold>. Frequency distributions (in %) for secondary modifications of monosaccharide residues are shown for taxonomy Set 1 (<bold>A</bold>) and Set 2 (<bold>B</bold>), normalized to the total number of carbohydrate residues within each taxonomic group. In <bold>A </bold>the bars for bacterial oligomers (blue) and polymers (cyan) are stacked to give the cumulative values for all bacteria studied.</p></caption></fig>", "<fig position=\"float\" id=\"F10\"><label>Figure 10</label><caption><p><bold>Distribution of monosaccharide pairs in various bacterial groups</bold>. For taxonomy Set 2 the matrix presents relative abundance data for monosaccharide residue pairs of all linkage types, involving the 20 most common residues serving as donor (children) or acceptor (parent). Each circle area reflects the relative abundance of a given donor-acceptor pair (matrix coordinates) within the corresponding taxonomic group, normalized to the total number of pairs within that group.</p></caption></fig>", "<fig position=\"float\" id=\"F11\"><label>Figure 11</label><caption><p><bold>Glycosidic linkages in mammalian carbohydrates</bold>. Frequency distribution of specific disaccharide linkages in mammalian carbohydrates. Plotted circle areas represent the relative frequencies for disaccharides formed from the 9 most common donors (children) and 9 most common acceptors (parents) in a defined glycosidic linkage (color code in legend). The areas of the circles are proportional to the relative abundances of disaccharide pairs, normalized to the total number of specific disaccharide pairs. The linkage codes α1-n and β1-n correspond to a linkage to any exocyclic carbon at the acceptor, e.g. C6 in hexopyranoses. For donor residues in keto form the linkage is at the anomeric carbon C2 instead of C1. For better visualization some of the circles for a given linkage are offset somewhat from the matrix coordinate corresponding to a given linkage type.</p></caption></fig>", "<fig position=\"float\" id=\"F12\"><label>Figure 12</label><caption><p><bold>Glycosidic linkages in bacterial carbohydrates</bold>. Frequency distribution of specific disaccharide linkages in bacterial carbohydrates. Plotted circle areas represent the relative frequencies for disaccharides formed from the 15 most common donors (children) and 15 most common acceptors (parents) in a defined glycosidic linkage (color code in legend) for bacterial oligomers (<bold>A</bold>) or polymers (<bold>B</bold>). The areas of the circles are proportional to the relative abundances of specific disaccharide pairs, normalized to the total number of disaccharide pairs. The linkage codes and plotting offsets are used as in Fig. 12. The lower left corner of each diagram is plotted at the right with rescaling for better visualization.</p></caption></fig>" ]

[ "<table-wrap position=\"float\" id=\"T1\"><label>Table 1</label><caption><p>Definition of taxonomy Set 1.</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td align=\"left\"><bold>Group Name</bold></td><td align=\"center\"><bold>Number of sequences*</bold></td><td align=\"left\"><bold>Explanation</bold></td></tr></thead><tbody><tr><td align=\"left\">Mammalia (class)</td><td align=\"center\">3328</td><td align=\"left\">Total number of different carbohydrate (glycan) sequences for Mammalia, as registered in GLYCOSCIENCES.de.</td></tr><tr><td align=\"left\">Bacteria (polymers)</td><td align=\"center\">2250</td><td align=\"left\">Total number of different repeating units in the polymeric carbohydrate (glycan) sequences for all bacteria registered in the BCSDB.</td></tr><tr><td align=\"left\">Bacteria (oligomers)</td><td align=\"center\">3210</td><td align=\"left\">Total number of different oligosaccharide sequences (nonrepeating units) for all bacteria registered in the BCSDB.</td></tr></tbody></table></table-wrap>", "<table-wrap position=\"float\" id=\"T2\"><label>Table 2</label><caption><p>Definition of taxonomy Set 2.</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td align=\"left\"><bold>Bacterial Class*</bold></td><td align=\"center\"><bold>Number of sequences**</bold></td><td align=\"center\"><bold>Gram reaction</bold></td></tr></thead><tbody><tr><td align=\"left\"><italic>Actinobacteria</italic></td><td align=\"center\">395</td><td align=\"center\">+</td></tr><tr><td align=\"left\"><italic>Bacilli</italic></td><td align=\"center\">640</td><td align=\"center\">+</td></tr><tr><td align=\"left\">α-<italic>Proteobacteria</italic></td><td align=\"center\">324</td><td align=\"center\">-</td></tr><tr><td align=\"left\">β-<italic>Proteobacteria</italic></td><td align=\"center\">365</td><td align=\"center\">-</td></tr><tr><td align=\"left\">(γ-<italic>Proteobacteria</italic>)</td><td align=\"center\">(3305)</td><td align=\"center\">-</td></tr><tr><td align=\"left\"> <italic>Enterobacteriales </italic>[order]</td><td align=\"center\">2087</td><td/></tr><tr><td align=\"left\"> other γ-<italic>Proteobacteria </italic>[orders]</td><td align=\"center\">1218</td><td/></tr><tr><td align=\"left\">δ/ε-<italic>Proteobacteria</italic></td><td align=\"center\">284</td><td align=\"center\">-</td></tr><tr><td align=\"left\"> (δ-<italic>Proteobacteria</italic>)</td><td align=\"center\">(2)</td><td/></tr><tr><td align=\"left\"> (ε-<italic>Proteobacteria</italic>)</td><td align=\"center\">(282)</td><td/></tr></tbody></table></table-wrap>", "<table-wrap position=\"float\" id=\"T3\"><label>Table 3</label><caption><p>Diversity of monosaccharides, basetypes and basic entities for various taxonomic groups.</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td align=\"left\"><bold>Group</bold></td><td align=\"center\"><bold>Monosaccharides</bold></td><td align=\"center\"><bold>Basetypes</bold></td><td align=\"center\"><bold>Basic entities</bold></td></tr></thead><tbody><tr><td align=\"left\"><italic>Mammalia</italic></td><td align=\"center\">35</td><td align=\"center\">14</td><td align=\"center\">10 *</td></tr><tr><td align=\"left\"><italic>Bacteria</italic></td><td align=\"center\">551</td><td align=\"center\">143</td><td align=\"center\">123</td></tr><tr><td align=\"left\"><italic>Actinobacteria</italic></td><td align=\"center\">100</td><td align=\"center\">48</td><td align=\"center\">33</td></tr><tr><td align=\"left\"><italic>Bacilli</italic></td><td align=\"center\">100</td><td align=\"center\">41</td><td align=\"center\">34</td></tr><tr><td align=\"left\">α-<italic>Proteobacteria</italic></td><td align=\"center\">68</td><td align=\"center\">34</td><td align=\"center\">26</td></tr><tr><td align=\"left\">β-<italic>Proteobacteria</italic></td><td align=\"center\">69</td><td align=\"center\">33</td><td align=\"center\">32</td></tr><tr><td align=\"left\"><italic>Enterobacteriales </italic>(γ-<italic>Proteobacteria</italic>)</td><td align=\"center\">243</td><td align=\"center\">76</td><td align=\"center\">65</td></tr><tr><td align=\"left\">other γ-<italic>Proteobacteria</italic></td><td align=\"center\">203</td><td align=\"center\">70</td><td align=\"center\">76 **</td></tr><tr><td align=\"left\">δ/ε-<italic>Proteobacteria</italic></td><td align=\"center\">52</td><td align=\"center\">35</td><td align=\"center\">32</td></tr></tbody></table></table-wrap>", "<table-wrap position=\"float\" id=\"T4\"><label>Table 4</label><caption><p>Diversity of disaccharide linkages found in various taxonomic groups.</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td align=\"left\"><bold>Group</bold></td><td align=\"center\"><bold>Disaccharide Linkages</bold></td></tr></thead><tbody><tr><td align=\"left\"><italic>Actinobacteria</italic></td><td align=\"center\">364</td></tr><tr><td align=\"left\"><italic>Bacilli</italic></td><td align=\"center\">526</td></tr><tr><td align=\"left\">α-<italic>Proteobacteria</italic></td><td align=\"center\">319</td></tr><tr><td align=\"left\">β-<italic>Proteobacteria</italic></td><td align=\"center\">356</td></tr><tr><td align=\"left\"><italic>Enterobacteriales </italic>(γ-<italic>Proteobacteria</italic>)</td><td align=\"center\">1570</td></tr><tr><td align=\"left\">other γ-<italic>Proteobacteria</italic></td><td align=\"center\">1148</td></tr><tr><td align=\"left\">δ/ε-<italic>Proteobacteria</italic></td><td align=\"center\">218</td></tr><tr><td align=\"left\"><italic>Mammalia</italic></td><td align=\"center\">488</td></tr></tbody></table></table-wrap>", "<table-wrap position=\"float\" id=\"T5\"><label>Table 5</label><caption><p>Identity and abundances of mammalian and bacterial disaccharides.</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td/><td align=\"center\" colspan=\"10\"><bold>Acceptors (parents)</bold></td></tr><tr><td/><td colspan=\"10\"><hr/></td></tr><tr><td align=\"center\"><bold>Donors (children)</bold></td><td align=\"center\"><bold>Fuc</bold></td><td align=\"center\"><bold>Gal</bold></td><td align=\"center\"><bold>GalNAc</bold></td><td align=\"center\"><bold>Glc</bold></td><td align=\"center\"><bold>GlcNAc</bold></td><td align=\"center\"><bold>GlcA, IdoA or ΔGlcA</bold></td><td align=\"center\"><bold>Man</bold></td><td align=\"center\"><bold>Neu5Ac or Neu5Gc</bold></td><td align=\"center\"><bold>Xyl</bold></td><td align=\"center\"><bold>GlcNS</bold></td></tr></thead><tbody><tr><td align=\"center\"><bold>Fuc</bold></td><td align=\"center\">a1–3 (0.01) [0.15]</td><td align=\"center\">a1–2 (2.68) [0.18]</td><td/><td/><td align=\"center\">a1–3 (2.00) [0.32]</td><td/><td/><td/><td/><td/></tr><tr><td/><td/><td/><td/><td/><td align=\"center\">a1–4 (0.54) [0.09]</td><td/><td/><td/><td/><td/></tr><tr><td/><td/><td/><td/><td/><td align=\"center\">a1–6 (2.84) [0.01]</td><td/><td/><td/><td/><td/></tr><tr><td colspan=\"11\"><hr/></td></tr><tr><td align=\"center\"><bold>Gal</bold></td><td/><td align=\"center\">a1–3 (0.92) [0.60]</td><td align=\"center\">a1–3 (0.02) [0.12]</td><td align=\"center\">a1–3 (0.00) [0.32]</td><td align=\"center\">a1–3 (0.00) [0.26]</td><td/><td align=\"center\">a1–2 [0.40]</td><td/><td align=\"center\">b1–4 (0.11) [0.00]</td><td/></tr><tr><td/><td/><td align=\"center\">a1–4 (0.22) [0.54]</td><td align=\"center\">b1–3 (2.41) [0.54]</td><td align=\"center\">a1–6 (0.00) [0.49]</td><td align=\"center\">b1–3 (1.86) [0.52]</td><td/><td align=\"center\">a1–3 [0.10]</td><td/><td/><td/></tr><tr><td/><td/><td align=\"center\">a1–6 (0.02) [0.25]</td><td align=\"center\">b1–4 (0.11) [0.11]</td><td align=\"center\">b1–3 (0.04) [0.18]</td><td align=\"center\">b1–4 (18.06) [1.68]</td><td/><td align=\"center\">a1–6 [0.12]</td><td/><td/><td/></tr><tr><td/><td/><td align=\"center\">b1–3 (0.70) [0.80]</td><td/><td align=\"center\">b1–4 (2.51) [2.32]</td><td/><td/><td/><td/><td/><td/></tr><tr><td/><td/><td align=\"center\">b1–4 (0.11) [0.34]</td><td/><td align=\"center\">b1–6 (0.00) [0.18]</td><td/><td/><td/><td/><td/><td/></tr><tr><td/><td/><td align=\"center\">b1–6 (0.05) [0.42]</td><td/><td align=\"center\">a1–2 [0.25]</td><td/><td/><td/><td/><td/><td/></tr><tr><td/><td/><td align=\"center\"><underline>a1–2 [0.52]</underline></td><td/><td/><td/><td/><td/><td/><td/><td/></tr><tr><td colspan=\"11\"><hr/></td></tr><tr><td align=\"center\"><bold>GalNAc</bold></td><td/><td align=\"center\">a1–3 (0.80) [0.05]</td><td align=\"center\">a1–3 (0.07) [0.32]</td><td/><td align=\"center\">b1–4 (0.68) [0.02]</td><td align=\"center\">b1–4 (0.12) [0.08]</td><td/><td/><td/><td/></tr><tr><td/><td/><td align=\"center\">b1–3 (0.30) [0.44]</td><td align=\"center\">b1–3 (0.03) [0.10]</td><td/><td/><td/><td/><td/><td/><td/></tr><tr><td/><td/><td align=\"center\">b1–4 (0.61) [0.36]</td><td align=\"center\">a1–4 [0.14]</td><td/><td/><td/><td/><td/><td/><td/></tr><tr><td/><td/><td/><td align=\"center\">b1–4 [0.20]</td><td/><td/><td/><td/><td/><td/><td/></tr><tr><td align=\"center\"><bold>Glc</bold></td><td/><td align=\"center\">b1–3 (0.02) [0.40]</td><td align=\"center\">a1–6 [0.12]</td><td align=\"center\">a1–2 (0.06) [0.95]</td><td align=\"center\">b1–3 [0.11]</td><td align=\"center\">a1–4 (0.03) [0.27]</td><td align=\"center\">a1–3 (0.19) [0.19]</td><td/><td/><td/></tr><tr><td/><td/><td align=\"center\">b1–6 (0.01) [0.34]</td><td align=\"center\">b1–3 [0.18]</td><td align=\"center\">a1–3 (0.08) [0.83]</td><td align=\"center\">b1–6 [0.15]</td><td align=\"center\">b1–4 (0.01) [0.12]</td><td align=\"center\">b1–4 (0.00) [0.14]</td><td/><td/><td/></tr><tr><td/><td/><td align=\"center\">a1–2 (*) [0.31]</td><td/><td align=\"center\">a1–4 (0.15) [0.39]</td><td/><td align=\"center\">b1–3 [0.10]</td><td/><td/><td/><td/></tr><tr><td/><td/><td align=\"center\">a1–3 [0.13]</td><td/><td align=\"center\">a1–6 (0.10) [0.57]</td><td/><td/><td/><td/><td/><td/></tr><tr><td/><td/><td align=\"center\"><underline>a1–4 [0.68]</underline></td><td/><td align=\"center\">b1–3 (0.05) [0.62]</td><td/><td/><td/><td/><td/><td/></tr><tr><td/><td/><td align=\"center\">a1–6 [0.13]</td><td/><td align=\"center\">b1–4 (0.00) [1.69]</td><td/><td/><td/><td/><td/><td/></tr><tr><td/><td/><td align=\"center\"><underline>b1–4 [0.40]</underline></td><td/><td align=\"center\">b1–6 (0.01) [0.90]</td><td/><td/><td/><td/><td/><td/></tr><tr><td/><td/><td/><td/><td align=\"center\">a1–1 [0.16]</td><td/><td/><td/><td/><td/><td/></tr><tr><td/><td/><td/><td/><td align=\"center\">b1–2 [0.33]</td><td/><td/><td/><td/><td/><td/></tr><tr><td colspan=\"11\"><hr/></td></tr><tr><td align=\"center\"><bold>GlcNAc</bold></td><td/><td align=\"center\">b1–3 (5.38) [1.64]</td><td align=\"center\">b1–3 (0.68) [0.04]</td><td/><td align=\"center\">b1–3 (0.01) [0.34]</td><td align=\"center\">a1–4 (0.10) [0.05]</td><td align=\"center\">b1–2 (9.46) [0.11]</td><td/><td/><td/></tr><tr><td/><td/><td align=\"center\">b1–4 (0.13) [0.02]</td><td align=\"center\">b1–6 (1.45) [0.00]</td><td/><td align=\"center\">b1–4 (5.61) [0.69]</td><td align=\"center\">b1–4 (0.13) [0.13]</td><td align=\"center\">b1–4 (2.83) [0.05]</td><td/><td/><td/></tr><tr><td/><td/><td align=\"center\">b1–6 (1.17) [0.04]</td><td/><td/><td align=\"center\">b1–6 (0.01) [0.22]</td><td/><td align=\"center\">b1–6 (1.64) [0.02]</td><td/><td/><td/></tr><tr><td/><td/><td/><td/><td/><td align=\"center\">b1–2 [0.15]</td><td/><td/><td/><td/><td/></tr><tr><td colspan=\"11\"><hr/></td></tr><tr><td align=\"center\"><bold>GlcA, IdoA or ΔGlcA</bold></td><td/><td align=\"center\">b1–3 (0.13) [0.23]</td><td align=\"center\">a1–3 (0.14) [0.01]</td><td align=\"center\">b1–4 [0.26]</td><td align=\"center\">a1–4 (0.16) [0.01]</td><td align=\"center\">b1–4 (0.01) [0.19]</td><td align=\"center\">a1–3 [0.22]</td><td/><td/><td align=\"center\"><underline>a1–4 (0.44)</underline></td></tr><tr><td/><td/><td align=\"center\">b1–4 (*) [0.14]</td><td/><td/><td align=\"center\">b1–3 (0.13) [0.14]</td><td align=\"center\">b1–3 [0.13]</td><td align=\"center\">b1–2 [0.19]</td><td/><td/><td align=\"center\">b1–4 (0.10)</td></tr><tr><td colspan=\"11\"><hr/></td></tr><tr><td align=\"center\"><bold>Man</bold></td><td/><td align=\"center\">a1–3 [0.22]</td><td/><td align=\"center\">a1–3 (0.00) [0.27]</td><td align=\"center\">b1–4 (6.22) [0.02]</td><td/><td align=\"center\">a1–2 (1.60) [0.92]</td><td/><td/><td/></tr><tr><td/><td/><td/><td/><td/><td/><td/><td align=\"center\">a1–3 (6.39) [0.41]</td><td/><td/><td/></tr><tr><td/><td/><td/><td/><td/><td/><td/><td align=\"center\">a1–4 (0.01) [0.16]</td><td/><td/><td/></tr><tr><td/><td/><td/><td/><td/><td/><td/><td align=\"center\">a1–6 (6.35) [0.53]</td><td/><td/><td/></tr><tr><td/><td/><td/><td/><td/><td/><td/><td align=\"center\">b1–4 [0.17]</td><td/><td/><td/></tr><tr><td colspan=\"11\"><hr/></td></tr><tr><td align=\"center\"><bold>Neu5Ac or Neu5Gc</bold></td><td/><td align=\"center\">a2–3 (4.85) [1.01]</td><td align=\"center\"><underline>a2–6 (0.49)</underline></td><td/><td align=\"center\">a2–6 (0.22)</td><td/><td/><td align=\"center\">a2–8 (0.29) [0.12]</td><td/><td/></tr><tr><td/><td/><td align=\"center\">a2–6 (2.64) [0.02]</td><td/><td/><td/><td/><td/><td/><td/><td/></tr><tr><td colspan=\"11\"><hr/></td></tr><tr><td align=\"center\"><bold>Xyl</bold></td><td/><td/><td/><td/><td/><td/><td/><td/><td align=\"center\">b1–4 (0.00) [0.10]</td><td/></tr><tr><td colspan=\"11\"><hr/></td></tr><tr><td align=\"center\"><bold>GlcNS</bold></td><td/><td/><td/><td/><td/><td align=\"center\">a1–4 (0.34)</td><td/><td/><td/><td/></tr></tbody></table></table-wrap>" ]

[]

[]

[]

[]

[]

[ "<supplementary-material content-type=\"local-data\" id=\"S1\"><caption><title>Additional file 1</title><p>Supplementary Table 1: Abundance of carbohydrate sequences for various taxonomical classes, orders and families.</p></caption></supplementary-material>", "<supplementary-material content-type=\"local-data\" id=\"S2\"><caption><title>Additional file 2</title><p>Supplementary Table 2: Abundance of basetypes for bacterial and mammalian carbohydrates.</p></caption></supplementary-material>", "<supplementary-material content-type=\"local-data\" id=\"S3\"><caption><title>Additional file 3</title><p>Supplementary Table 3: Abundances of monosaccharide residues found in bacterial and mammalian carbohydrates.</p></caption></supplementary-material>", "<supplementary-material content-type=\"local-data\" id=\"S4\"><caption><title>Additional file 4</title><p>Supplementary Table 4: Abundances of all monosaccharide pairs found in bacterial and mammalian carbohydrates.</p></caption></supplementary-material>", "<supplementary-material content-type=\"local-data\" id=\"S5\"><caption><title>Additional file 5</title><p>Supplementary Table 5: Comprehensive list of mammalian disaccharide fragments and their relative abundances.</p></caption></supplementary-material>", "<supplementary-material content-type=\"local-data\" id=\"S6\"><caption><title>Additional file 6</title><p>Supplementary Table 6: Comparison of mammalian and bacterial disaccharide fragments and their relative abundances.</p></caption></supplementary-material>" ]

[ "<table-wrap-foot><p>* For each group each unique carbohydrate sequence was counted only once; a given sequence may occur in more than one taxonomic group.</p></table-wrap-foot>", "<table-wrap-foot><p>* The groups used to define Set 2 are those with more than 200 sequences and are listed without parentheses.</p><p>** For each class each unique carbohydrate sequence was counted only once; a given sequence may occur in more than one taxonomic group.</p></table-wrap-foot>", "<table-wrap-foot><p>The total number of different monosaccharides, basetypes, and basic entities are listed for each group in taxonomy Set 1 and Set 2. When counting in the combined databases, an occurrence threshold of 10 for mammals and 2 for bacteria was employed.</p><p>* The basic entity <sc>D</sc>-Fuc was excluded from <italic>Mammalia </italic>for reasons given in the text.</p><p>** In this case there are more basic entities compared to basetypes because amino derivatives are included in the basic entities.</p></table-wrap-foot>", "<table-wrap-foot><p>Total count of all unique disaccharide linkages per taxonomic group. Unknown anomeric and absolute configurations and fuzzily defined linkages are counted as distinct entities.</p></table-wrap-foot>", "<table-wrap-foot><p>Linkage types and relative abundances are shown as percentages (rounded to two decimals) of the total disaccharide fragment count for mammals (values in parentheses) or bacteria [values in brackets]. Linkages are listed only for disaccharide fragments present at &gt; 0.1% in either taxonomy group. Values of 0.00 indicate the presence of a linkage at &lt; 0.005%; missing values indicate the absence of that linkage type in the corresponding taxonomy group. An entry (*) denotes that the linkage was reported for mammals in [##REF##16959566##2##] but was not found in the databases. The percentages are cumulative for all ring types (including alditols), all absolute configurations and for all anomeric configurations of the acceptor. For linkages that occur only in mammals or only in bacteria, those with abundences of 0.40% or more are underlined.</p></table-wrap-foot>" ]

[ "<graphic xlink:href=\"1472-6807-8-35-1\"/>", "<graphic xlink:href=\"1472-6807-8-35-2\"/>", "<graphic xlink:href=\"1472-6807-8-35-3\"/>", "<graphic xlink:href=\"1472-6807-8-35-4\"/>", "<graphic xlink:href=\"1472-6807-8-35-5\"/>", "<graphic xlink:href=\"1472-6807-8-35-6\"/>", "<graphic xlink:href=\"1472-6807-8-35-7\"/>", "<graphic xlink:href=\"1472-6807-8-35-8\"/>", "<graphic xlink:href=\"1472-6807-8-35-9\"/>", "<graphic xlink:href=\"1472-6807-8-35-10\"/>", "<graphic xlink:href=\"1472-6807-8-35-11\"/>", "<graphic xlink:href=\"1472-6807-8-35-12\"/>" ]

[ "<media xlink:href=\"1472-6807-8-35-S1.xls\" mimetype=\"application\" mime-subtype=\"vnd.ms-excel\"><caption><p>Click here for file</p></caption></media>", "<media xlink:href=\"1472-6807-8-35-S2.xls\" mimetype=\"application\" mime-subtype=\"vnd.ms-excel\"><caption><p>Click here for file</p></caption></media>", "<media xlink:href=\"1472-6807-8-35-S3.xls\" mimetype=\"application\" mime-subtype=\"vnd.ms-excel\"><caption><p>Click here for file</p></caption></media>", "<media xlink:href=\"1472-6807-8-35-S4.xls\" mimetype=\"application\" mime-subtype=\"vnd.ms-excel\"><caption><p>Click here for file</p></caption></media>", "<media xlink:href=\"1472-6807-8-35-S5.rtf\" mimetype=\"application\" mime-subtype=\"rtf\"><caption><p>Click here for file</p></caption></media>", "<media xlink:href=\"1472-6807-8-35-S6.rtf\" mimetype=\"application\" mime-subtype=\"rtf\"><caption><p>Click here for file</p></caption></media>" ]

{ "acronym": [], "definition": [] }

[ "<title>Background</title>", "<p>The main aim of our molecular modelling investigations was to identify natural compounds for their ability to bind to the NF-kappaB p50 as a strategy to identify molecules exhibiting inhibitory activity on the molecular interactions of the transcription factor with its target DNA sequence. p50–p65 heterodimer is the predominant NF-kappaB complex in T-cells regulating HIV-1 infection and recent studies have shown that p50 unit of NF-kappaB is the one that mainly interacts with HIV-1 LTR [##REF##11160127##1##,##REF##10212261##2##]. The specific protein residues involved in DNA binding to the HIV-1 LTR NF-kappaB sites (sequence 5'-GGGACTTTCCC-3') have been identified [##REF##9450761##3##,##REF##7830764##4##]. Structurally different inhibitors of the NF-kappaB/DNA interactions with a rather low binding constant (in the range of 30 μM and 500 μM) are reported in the literature [##REF##10976530##5##, ####REF##15788163##6##, ##REF##15546742##7####15546742##7##]. Recently, some molecular modelling studies have predicted possible binding mode of the inhibitors molecules to the DNA binding region of subunit p50, starting from the crystallographic structure of the NF-kappaB homodimer [##REF##15788163##6##, ####REF##15546742##7##, ##REF##16759101##8##, ##REF##15124931##9####15124931##9##].</p>", "<p>In particular, Sharma et al. [##REF##16759101##8##] in an effort to rationalize the results obtained from EMSA studies on a set of aurintricarboxylic acid analogues, employed docking studies to explain the structure activity relationships observed within this class. To the best of our knowledge, nowadays the identification of new lead compounds for NF-kappaB inhibition through virtual screening of structures libraries is not yet reported in literature. In this paper, we present docking studies on a series of natural compounds previously identified within medicinal plant extracts by us, into NF-kappaB p50 protein target. After evaluation through electrophoretic mobility shift assays (EMSA), we obtained a fairly good agreement between experimental data and molecular modelling identification of bioactive and inactive compounds.</p>" ]

[ "<title>Methods</title>", "<title>Docking studies</title>", "<title>Ligands data and preparation</title>", "<p>The database of 27 natural structures used in our molecular docking studies, were derived from different medicinal plant extracts (Figure ##FIG##0##1##) as prepared in our laboratory. A dataset of 12 active compounds used as references molecules were collected from four publications [##REF##15788163##6##, ####REF##15546742##7##, ##REF##16759101##8##, ##REF##15124931##9####15124931##9##] reported by one laboratory (Figure ##FIG##1##2##). Ten of these inhibitors (<bold>1i-8i</bold>, <bold>11i </bold>and <bold>12i</bold>) were employed in starting docking studies (protocol 1) and in the Standard Similarity Scoring for subsequently docking simulations.</p>", "<p>Two inhibitory molecules (<bold>9i </bold>and <bold>10i</bold>) were used as test set in all docking simulations. The three-dimensional models of all the molecules under investigation were built by assembling fragments from the SYBYL 7.0 software package standard library [##UREF##0##10##]. Resulting geometries were optimized and molecular charges were assigned by a semi empirical molecular orbital calculations using the AM1 Hamiltonian [##UREF##1##11##] (module MOPAC implemented in SYBYL).</p>", "<title>Proteins data and preparation</title>", "<p>The three dimensional structure of the complex NF-kappaB-DNA [##REF##7830764##4##] was retrieved from the Protein Data Bank (PDB code: <ext-link ext-link-type=\"pdb\" xlink:href=\"1NFK\">1NFK</ext-link>). The cocrystallized DNA macromolecule was removed from the structure. p50 dimer and p50 monomers (chains A and B) were selected for the docking simulations and prepared using the graphical interface Maestro [##UREF##2##12##]. All water molecules were removed, the hydrogen atoms were added to the proteins and all atom force field (OPSL-2001) charges and atom types were assigned. Preparation and refinement were done running ProteinPrep job on the structure in a standard procedure. Minimizations were performed until the average root mean square deviation of non-hydrogen atoms reached 0.3 Å.</p>", "<title>Docking Simulations</title>", "<p>All molecules of plant extracts (<bold>1–27</bold>) and the known inhibitors (<bold>1i-12i</bold>) under study were docked in to the binding site of the receptor (PDB ID: <ext-link ext-link-type=\"pdb\" xlink:href=\"1NFK\">1NFK</ext-link>) using Glide (Grid-Based Ligand Docking With Energetics) software from Schrodinger [##REF##15027866##13##,##REF##15027865##14##]. Grids were prepared for each proteins with the exact same center and the size of the bounding box set on 30 Å. The coordinates of the enclosing box (x = -1,1958 Å; y = 9.0149 Å; z = 19,7598 Å) were defined starting from the set of active site residues involved in hydrogen bonds with the NF-kappaB recognition site of DNA (Arg54, Arg56, Tyr57, Cys59, Lys241, Gln306 and Thr143) and optimised including the double strands DNA helices volume by visual inspection. The Glide algorithm is based on a systematic search of positions, orientations, and conformations of the ligand in the receptor binding site using funnel type approach. The search begins with a rough positioning and scoring phase that significantly limits the search space and reduces the number of poses to be selected for minimization on the precomputed OPLS-2001 van der Waals and electrostatic grids for the protein. The 5–10 lowest-energy poses obtained from this stage are subjected to Monte Carlo simulations and the minimized poses accepted are then rescored using the GlideScore function, which is a more sophisticated version of ChemScore [##REF##9385547##15##]. This force field include additional terms accounting for solvation and repulsive interactions. In order to provide a better correlation between good poses and good scores, Glide Extra-Precision (XP) Mode was subsequently used on the conformations selected from Glide Standard Precision (SP) mode. The atom-pair superimposition of p50 chains A and B, prepared as described above, gave a minimum RMSD of 2,303 Å (heavy atoms). Considering the clear dependence of the docking accuracy of ligands on the protein structure, docking simulations were carried out with the same protocol on both A and B, considered as two slightly different conformations of the same structure.</p>", "<p>Unfortunately, complexes of NF-kappaB cocrystallized with inhibitors has not been solved. Therefore, a common self-docking procedure to evaluate the accuracy of the docking protocol adopted was not practicable. In order to overcame this situation, two structurally similar active compounds (<bold>9i </bold>and <bold>10i</bold>) were used as test set and docked into the DNA binding site of the protein. Moreover in following docking jobs, atom pair similarity (AP) scoring (Similscore) facility as implemented in Glide, was incorporated in GlideScore (G-score), based on the assumption that closely related chemical structure should share similar biological activity and physiochemical property [##UREF##3##16##]. Similscore can have a value between 0 and 1 as implemented in Glide. The adjusting G-score value is illustrated here below:</p>", "<p>1. if 0.0 ≤ SimilScore &lt; 0.3 → G-score = G-score+6.0</p>", "<p>2. if 0.3 ≤ SimilScore &lt; 0.7 → G-score = G-score+(0.7-Similscore)/(0.7-0.3)*6.0</p>", "<p>3. if 0.7 ≤ SimilScore &lt; 1.0 → G-score = G-score+0.0</p>", "<p>All inhibitors molecules, except for <bold>9i </bold>and <bold>10i</bold>, were used just as reference structures for AP similarity method.</p>", "<p>Based on the best final GlideScore ranking, the similarity docking procedure for subsequently docking simulations on p50 subunits was chosen.</p>", "<title>Preparation of nuclear extracts</title>", "<p>Nuclear extracts were prepared as described [##REF##12446679##18##]. Cell were washed twice with PBS and detached by trypsinization. After homogenization with Dounce B homogeneizer, nuclear proteins were obtained and protein concentration was determined using Bio-Rad protein assay. Nuclear extracts were brought to a concentration of 0.5 μg/μl for Electrophoretic Mobility Shift Assay (EMSA) experiments [##REF##12446679##18##].</p>", "<title>Electrophoretic Mobility Shift Assay (EMSA)</title>", "<p>EMSA was perfomed as previously described [##REF##12446679##18##, ####REF##12167479##19##, ##REF##17466942##20####17466942##20##]. Briefly, double-stranded synthetic oligodeoxynucleotides mimicking the NF-κB binding site present in the promoter of the IL-8 gene (IL-8 NF-κB, sense: 5'-AAT CGT GGA ATT TCC TCT-3') have been employed. Oligodeoxynucleotides were labeled with γ³²-P-ATP using 10 Units of T4-polynucleotide-kinase (MBI Fermentas) in 500 mM Tris-HCl, pH 7.6, 100 mM MgCl₂, 50 mM DTT, 1 mM spermidine, 1 mM EDTA in the presence of 50 μCi γ³²-P-ATP) in a volume of 20 μl for 45 minutes at 37°C. Reaction was brought to 150 mM NaCl and 150 ng complementary oligodeoxynucleotide was added. Reaction temperature was increased to 100°C for 5 minutes and left diminishing to room temperature overnight. Nuclear extracts from IB3-1 cells or purified NF-κB p50 dimer protein (Promega) were used at the specified concentrations and poly(dI:dC) (1 mg per reaction) was also added to abolish nonspecific binding [##REF##18258920##21##]. After 5 min binding at room temperature, the samples were run at constant voltage (200 V) under low ionic strength conditions (0.25× TBE buffer: 22 mM Tris-borate, 0.4 mM EDTA) on 6% polyacrylamide gels. Gels were dried and subjected to standard autoradiographic procedures.</p>", "<title>Cell cultures and infection with <italic>Pseudomonas aeruginosa</italic></title>", "<p>IB3-1 cells have been obtained from LGC Promochem [##REF##17466942##20##]. Cells have been grown in LHC-8 basal medium (Biofluids), supplemented with 5% FBS in the absence of gentamycin [##REF##18258920##21##]. All culture flasks and plates have been coated with a solution containing 35 g/ml bovine collagen (Becton-Dickinson), 1 g/ml bovine serum albumin (Sigma) and 1 g/ml human fibronectin (Becton-Dickinson) as described. P. aeruginosa, PAO1 strain, was grown in trypticase soy broth (TSB) or agar (TSA) (Difco). Bacteria colonies from o/n cultures on TSA plates were grown in 20 ml TSB at 37°C. IB3-1 cells were infected with ranging doses of PAO1 at 37°C for 4 hours.</p>", "<title>Quantitation of transcripts of inflammatory genes</title>", "<p>Total RNA from IB3-1 cells was isolated using High Pure RNA Isolation Kit (Roche. Mannheim. Germany) [##REF##18258920##21##]. Total RNA (2.5 μg) was reverse-transcribed to cDNA using the High Capacity cDNA Archive Kit and random primers (Applied Biosystems) in a 100-μl reaction. The cDNA (2 μl) was then amplified for 50 PCR cycles using the Platinum^®SYBR^®Green qPCR SuperMix-UDG (Invitrogen) in an ABI Prism 5700 sequence detection system (Applied Biosystems). The real-time PCR reactions were performed in duplicates for both target and normalizer genes. Primer sequences for detection of IL-8 mRNA were GACCACACTGCGCCAACA (IL-8 forward) and GCTCTCTTCCATCAGAAAGTTACATAATTT(IL-8 reverse). Primer sets were purchased from Sigma-Genosys (The Woodlands. TX). Results were collected with Sequence Detection Software (version 1.3; Applied Biosystems). Relative quantification of gene expression was performed using the comparative threshold (C_T) method as described by the manufacturer (Applied Biosystems User Bulletin 2). Changes in mRNA expression level were calculated following normalization to calibrator gene [##REF##12738678##22##]. The ratios obtained following normalization are expressed as -fold change over untreated samples.</p>" ]

[ "<title>Results and discussion</title>", "<title>Docking analysis</title>", "<p>The docking results for all the reference inhibitory compounds and the natural compounds under study are reported in Table ##TAB##0##1## (references compounds) and in Table ##TAB##1##2## (natural compounds). As shown in the 2.3 Å crystal structure, the DNA/p50 complex is formed by one DNA molecule and two p50 proteins each one consisting of two distinct domains connected by a10-residue linker. Both domains and the segments that connects them, form a sequence-specific DNA-binding surface by contributing 5 loops per subunit that fill the entire major groove of the DNA. The specific interactions that stabilized the NF-kappaB/DNA complex, occur over 10-bp forming the kB recognition site. Unlike many dimeric protein-DNA complexes, many residues of both subunits make specific base contacts in a non-contiguous cooperative network. The plasticity of centre region of the interface carry to the lack of symmetry exhibited by the interactions of Lys 241 from the linker segment, and Lys 272 and Arg 305 from the dimerization region with the symmetrical target site [##REF##7830764##4##]. In the subsequent experimental EMSA studies, a recombinant p50 protein that probably forms a monomer-dimer mixture in binding buffer solution will be used. On the base of structural and experimental assumptions as above mentioned, p50 dimer and monomer were employed as protein target in our molecular modelling investigation.</p>", "<p>In order to evaluate the impact of the introduction of the similarity penalty in the docking algorithm on the results, the positions of <bold>9i </bold>and <bold>10i </bold>used as test set in the final GlideScore ranking were compared in the two different procedures (Table ##TAB##1##2##).</p>", "<p>Known active compound <bold>10i </bold>was ranked at the top positions in both procedures, but only the introduction of the similarity parameter in the scoring function significantly increased the efficiency in <bold>9i </bold>ranking (Table ##TAB##1##2##). In fact the difference in glide-score among these two inhibitors with a similar inhibitory potency (500 μM) [##REF##15124931##9##] was very small (ΔGlideScore = 0.10). For each selected ligand the pose with best E-Model score (combination of energy grid score, GlideScore, and the internal strain of the ligand) was used for in-deph interaction analysis. Compound <bold>21 </bold>clearly showed highest score in respect to docked plant extracts (Table ##TAB##1##2##) outranking the known inhibitors at physiological pH in docking simulation to the dimer.</p>", "<p>Docked compounds <bold>1–27</bold>, <bold>9i </bold>and <bold>10i</bold>, occupied a region of the binding surface creates by the spatial relationship between the N-terminal domain of p50 subunit and the 10 residues long linker loop (Figure ##FIG##2##3##). Molecules <bold>21</bold>, <bold>9i </bold>and <bold>10i </bold>(Figure ##FIG##3##4##) were located in a small cleft surrounded by several polar amino acids (i.e. Tyr57, His109, His141, Tyr143 Lys144, Lys145, Ser208, Asp239, Lys241 and Ser208) and the highest score poses were superimposable with minimum RMSD of 1.36 Å for compounds <bold>9i </bold>and <bold>21</bold>. The RMSD was calculated by superimposing the following atoms pairs: heteroatoms involved in hydrogen bonding with the same residues of the protein (<bold>9i</bold>.O8 and <bold>21</bold>.O2'; <bold>9i</bold>.O7 and <bold>21</bold>.O1a) (Figure ##FIG##3##4B##) and the centroid of aromatic system of coumarin structure with the centroid of benzene ring of <bold>21</bold>. These compounds showed slightly different binding modes in p50 (chain A), p50 (chain B) and p50-p50 targets. Here we reported the highest score poses obtained from docking protocol including the similarity function. H-bond interactions between OH groups of coumarin structures (OH of benzene ring in <bold>21</bold>) and both NH of His141 and the carboxylic group of Asp239 showed to be important for ligands binding.</p>", "<p>Moreover, OH groups of coumarin moiety (carboxylate group in <bold>21</bold>) made an additional hydrogen bond with CO of the backbone and protonated NH₃group of Lys241 (Table ##TAB##2##3##). It is important to note that Lys241 could be involved in the stability of the DNA-binding conformation of the protein. In fact, as discussed above, this residue is situated in the flexible linker segment and interacts with Lys 272 and Arg 305 from the dimerization domain. Finally, carbonyl group of <bold>10i </bold>engage another H-bond with NH of the backbone of the Leu207. Compound <bold>21 </bold>showed the same binding mode of active ligands in the monomer configuration of the target, with the only difference of a stronger interaction of carboxylate group with Lys241 (Table ##TAB##2##3## and Figures ##FIG##4##5A##, ##FIG##5##6A##).</p>", "<p>Interestingly, the best pose of compound <bold>21 </bold>occupied a region formed by residues of both p50 units (chain A and chain B) of NF-kappaB dimer: Lys 145 and Thr143 of chain A and Tyr57, Lys144, Lys145, Glu60, Cys59, Thr143, Lys146 of chain B. In particular, the OH group of the ligand engages a hydrogen bond with the sidechain of Thr143 (chain B), and the carboxylate group forms a salt bridge stabilized by two hydrogen bonds with the side chain of Lys 145 (chain B). Moreover the phenyl structure of compound <bold>21 </bold>could be involved in a weak π-π stacking interaction with the aromatic moiety of Tyr57 (chain B) (centroid-centroid distance: 4.93 Å), a residue specific for kB DNA sequence 5'-GGGATTTCC-3', present in different cellular genes including HIV-LTR. Of course, further dynamics simulation on the protein-ligand complex should be necessary to validate this hypothesis. In addition, the amino group of Lys145 of the opposite p50 unit (chain A) could form an additional π-cation interaction with the aromatic group of <bold>21 </bold>(N⁺-H and centroid of benzene ring distance: 3.87 Å) (Figures ##FIG##4##5B##, ##FIG##5##6B##). These bridge structures are likely to reinforce the anchoring of this molecule to the DNA binding region of the dimer, and might account for the slight better G-score of <bold>21 </bold>in respect to the monomer configuration of the receptor. Moreover, all the residues of the protein involved in molecular interactions with molecule <bold>21</bold>, form hydrogen bonds also with DNA.</p>", "<p>All compounds with higher GlideScore and E-Model score clearly showed the ability to make a maximum number of hydrogen bonding, according with the result as previously reported on a flexible docking studies of known inhibitors <bold>9i </bold>and <bold>10i </bold>[##REF##15124931##9##], even if reported residues involved in binding interaction were different. The highest ranking poses of <bold>21</bold>, <bold>9i </bold>and <bold>10i </bold>formed 3–4 hydrogen bonding with the target protein, whereas molecules in medium positions in docking ranking not more than 2. According, structures not involved in hydrogen bonding were ranked in the last positions (Table ##TAB##1##2##). In particular, compound <bold>5 </bold>with a GlideScore &lt; 0 in similarity protocol lost the ability both to occupy the same positions of active ligands and to form hydrogen bonding with the protein (not shown). In house experimental data were in good agreement with the molecular modelling findings. In accordance with docking results, <bold>21 </bold>and <bold>5 </bold>showed to be active and inactive respectively in further EMSA experimental studies.</p>", "<title>Effects of compound 21 on NF-kappaB/DNA interactions</title>", "<p>The effects of compound <bold>21 </bold>on NF-kappaB interactions were first studied by electrophoretic mobility shift assay (EMSA) as described elsewhere [##REF##12446679##18##, ####REF##12167479##19##, ##REF##17466942##20##, ##REF##18258920##21####18258920##21##]. It is indeed well accepted that molecules binding NF-kappaB might retain inhibitory activity on molecular interaction between NF-kappaB and DNA [##REF##18258920##21##]. Accordingly, we performed EMSA in the presence of increasing amounts of compound <bold>21</bold>. In addition, compounds <bold>5 </bold>was used as possible negative control. This compound, indeed, is expected, from the docking analysis (Table ##TAB##1##2##), to be less active. In addition, extracts from <italic>Cupressus pyramidalis </italic>were also used, since this extract does not inhibit NF-kappaB/DNA interactions (data not shown). Finally, the known inhibitory compound <bold>9i </bold>was used as reference molecule. The results of the gel retardation analysis are shown in Figure ##FIG##6##7## and clearly demonstrate that compound <bold>21 </bold>inhibit the molecular interactions between nuclear factors (Figure ##FIG##6##7B##) or isolated NF-kappaB p50 (Figure ##FIG##6##7C## and ##FIG##6##7D##) and a target double stranded oligonucleotide mimicking the NF-kappaB binding sites. This effect was similar to that exhibited by the reference compound <bold>9i</bold>. Interestingly, compound <bold>5 </bold>and extracts from <italic>C. pyramidalis </italic>were found to be inactive (Figure ##FIG##6##7A##), fully in agreement with the docking data summarized in Table ##TAB##1##2##.</p>", "<title>Biological effects of compound 21: inhibition of <italic>Pseudomonas aeruginosa </italic>mediated increase of IL-8 mRNA</title>", "<p>Several experimental model system are available for biological validation of molecules inhibiting NF-kappaB function. In a recent paper we report that decoy oligonucleotides targeting NF-kappaB are powerful inhibitors of <italic>Pseudomoas aeruginosa </italic>mediated induction of IL-8 in cystic fibrosis IB3-1 cells [##REF##18258920##21##]. Besides the importance of these data for the theoretical point of view, our results are of great interest for the practical point of view, suggesting this treatment as a possible strategy for the therapy of inflammation associated with cystic fibrosis.</p>", "<p>When the effects of <italic>P. aeruginosa </italic>infection on the expression of pro-inflammatory genes of IB3-1 cells infected for 4 hours are analysed, the results shown in Figure ##FIG##7##8A## are obtained. In this preliminary experiment the content of RNAs coding for several pro-inflammatory proteins was analysed by RT-PCR. The results obtained indicate that IL-8 mRNA sequences sharply increases following PAO1 infection by 40 folds (range 12–92) in respect to basal level of uninfected cells, assumed to be 1. In addition to IL-8 mRNA, other genes induced by PAO1 are GRO-γ (15 folds, range 9–36), IL-6 (18 folds, range 5–44), IL-1β (5 folds, range 2–8), ICAM-1 (4 folds, range 2–8). On the contrary, very low increase of accumulation of IP-10, RANTES, MIP-1α, TNF-α, IFN-β, TGF-β and IL-10 mRNA was observed under these experimental conditions.</p>", "<p>Since NF-kappaB is one of the most important transcription factors regulating the expression of IL-8 gene [23] and the data reported in Figure ##FIG##6##7## demonstrate that compound <bold>21 </bold>inhibit NF-kappaB/DNA interactions, we tested the activity of this compound in inhibiting the expression of IL-8 gene in IB3-1 cells infected with PAO1. Cells were exposed to different concentrations of compound <bold>21 </bold>and then infected with 150 cfu/cell of PAO1. After 4 hours, cells were harvested, RNA extracted and quantitative RT-PCR analysis performed. The results obtained (Figure ##FIG##7##8B##) demonstrate that compound <bold>21 </bold>is a strong inhibitor of PAO-1 induced accumulation of IL-8 mRNA.</p>" ]

[ "<title>Results and discussion</title>", "<title>Docking analysis</title>", "<p>The docking results for all the reference inhibitory compounds and the natural compounds under study are reported in Table ##TAB##0##1## (references compounds) and in Table ##TAB##1##2## (natural compounds). As shown in the 2.3 Å crystal structure, the DNA/p50 complex is formed by one DNA molecule and two p50 proteins each one consisting of two distinct domains connected by a10-residue linker. Both domains and the segments that connects them, form a sequence-specific DNA-binding surface by contributing 5 loops per subunit that fill the entire major groove of the DNA. The specific interactions that stabilized the NF-kappaB/DNA complex, occur over 10-bp forming the kB recognition site. Unlike many dimeric protein-DNA complexes, many residues of both subunits make specific base contacts in a non-contiguous cooperative network. The plasticity of centre region of the interface carry to the lack of symmetry exhibited by the interactions of Lys 241 from the linker segment, and Lys 272 and Arg 305 from the dimerization region with the symmetrical target site [##REF##7830764##4##]. In the subsequent experimental EMSA studies, a recombinant p50 protein that probably forms a monomer-dimer mixture in binding buffer solution will be used. On the base of structural and experimental assumptions as above mentioned, p50 dimer and monomer were employed as protein target in our molecular modelling investigation.</p>", "<p>In order to evaluate the impact of the introduction of the similarity penalty in the docking algorithm on the results, the positions of <bold>9i </bold>and <bold>10i </bold>used as test set in the final GlideScore ranking were compared in the two different procedures (Table ##TAB##1##2##).</p>", "<p>Known active compound <bold>10i </bold>was ranked at the top positions in both procedures, but only the introduction of the similarity parameter in the scoring function significantly increased the efficiency in <bold>9i </bold>ranking (Table ##TAB##1##2##). In fact the difference in glide-score among these two inhibitors with a similar inhibitory potency (500 μM) [##REF##15124931##9##] was very small (ΔGlideScore = 0.10). For each selected ligand the pose with best E-Model score (combination of energy grid score, GlideScore, and the internal strain of the ligand) was used for in-deph interaction analysis. Compound <bold>21 </bold>clearly showed highest score in respect to docked plant extracts (Table ##TAB##1##2##) outranking the known inhibitors at physiological pH in docking simulation to the dimer.</p>", "<p>Docked compounds <bold>1–27</bold>, <bold>9i </bold>and <bold>10i</bold>, occupied a region of the binding surface creates by the spatial relationship between the N-terminal domain of p50 subunit and the 10 residues long linker loop (Figure ##FIG##2##3##). Molecules <bold>21</bold>, <bold>9i </bold>and <bold>10i </bold>(Figure ##FIG##3##4##) were located in a small cleft surrounded by several polar amino acids (i.e. Tyr57, His109, His141, Tyr143 Lys144, Lys145, Ser208, Asp239, Lys241 and Ser208) and the highest score poses were superimposable with minimum RMSD of 1.36 Å for compounds <bold>9i </bold>and <bold>21</bold>. The RMSD was calculated by superimposing the following atoms pairs: heteroatoms involved in hydrogen bonding with the same residues of the protein (<bold>9i</bold>.O8 and <bold>21</bold>.O2'; <bold>9i</bold>.O7 and <bold>21</bold>.O1a) (Figure ##FIG##3##4B##) and the centroid of aromatic system of coumarin structure with the centroid of benzene ring of <bold>21</bold>. These compounds showed slightly different binding modes in p50 (chain A), p50 (chain B) and p50-p50 targets. Here we reported the highest score poses obtained from docking protocol including the similarity function. H-bond interactions between OH groups of coumarin structures (OH of benzene ring in <bold>21</bold>) and both NH of His141 and the carboxylic group of Asp239 showed to be important for ligands binding.</p>", "<p>Moreover, OH groups of coumarin moiety (carboxylate group in <bold>21</bold>) made an additional hydrogen bond with CO of the backbone and protonated NH₃group of Lys241 (Table ##TAB##2##3##). It is important to note that Lys241 could be involved in the stability of the DNA-binding conformation of the protein. In fact, as discussed above, this residue is situated in the flexible linker segment and interacts with Lys 272 and Arg 305 from the dimerization domain. Finally, carbonyl group of <bold>10i </bold>engage another H-bond with NH of the backbone of the Leu207. Compound <bold>21 </bold>showed the same binding mode of active ligands in the monomer configuration of the target, with the only difference of a stronger interaction of carboxylate group with Lys241 (Table ##TAB##2##3## and Figures ##FIG##4##5A##, ##FIG##5##6A##).</p>", "<p>Interestingly, the best pose of compound <bold>21 </bold>occupied a region formed by residues of both p50 units (chain A and chain B) of NF-kappaB dimer: Lys 145 and Thr143 of chain A and Tyr57, Lys144, Lys145, Glu60, Cys59, Thr143, Lys146 of chain B. In particular, the OH group of the ligand engages a hydrogen bond with the sidechain of Thr143 (chain B), and the carboxylate group forms a salt bridge stabilized by two hydrogen bonds with the side chain of Lys 145 (chain B). Moreover the phenyl structure of compound <bold>21 </bold>could be involved in a weak π-π stacking interaction with the aromatic moiety of Tyr57 (chain B) (centroid-centroid distance: 4.93 Å), a residue specific for kB DNA sequence 5'-GGGATTTCC-3', present in different cellular genes including HIV-LTR. Of course, further dynamics simulation on the protein-ligand complex should be necessary to validate this hypothesis. In addition, the amino group of Lys145 of the opposite p50 unit (chain A) could form an additional π-cation interaction with the aromatic group of <bold>21 </bold>(N⁺-H and centroid of benzene ring distance: 3.87 Å) (Figures ##FIG##4##5B##, ##FIG##5##6B##). These bridge structures are likely to reinforce the anchoring of this molecule to the DNA binding region of the dimer, and might account for the slight better G-score of <bold>21 </bold>in respect to the monomer configuration of the receptor. Moreover, all the residues of the protein involved in molecular interactions with molecule <bold>21</bold>, form hydrogen bonds also with DNA.</p>", "<p>All compounds with higher GlideScore and E-Model score clearly showed the ability to make a maximum number of hydrogen bonding, according with the result as previously reported on a flexible docking studies of known inhibitors <bold>9i </bold>and <bold>10i </bold>[##REF##15124931##9##], even if reported residues involved in binding interaction were different. The highest ranking poses of <bold>21</bold>, <bold>9i </bold>and <bold>10i </bold>formed 3–4 hydrogen bonding with the target protein, whereas molecules in medium positions in docking ranking not more than 2. According, structures not involved in hydrogen bonding were ranked in the last positions (Table ##TAB##1##2##). In particular, compound <bold>5 </bold>with a GlideScore &lt; 0 in similarity protocol lost the ability both to occupy the same positions of active ligands and to form hydrogen bonding with the protein (not shown). In house experimental data were in good agreement with the molecular modelling findings. In accordance with docking results, <bold>21 </bold>and <bold>5 </bold>showed to be active and inactive respectively in further EMSA experimental studies.</p>", "<title>Effects of compound 21 on NF-kappaB/DNA interactions</title>", "<p>The effects of compound <bold>21 </bold>on NF-kappaB interactions were first studied by electrophoretic mobility shift assay (EMSA) as described elsewhere [##REF##12446679##18##, ####REF##12167479##19##, ##REF##17466942##20##, ##REF##18258920##21####18258920##21##]. It is indeed well accepted that molecules binding NF-kappaB might retain inhibitory activity on molecular interaction between NF-kappaB and DNA [##REF##18258920##21##]. Accordingly, we performed EMSA in the presence of increasing amounts of compound <bold>21</bold>. In addition, compounds <bold>5 </bold>was used as possible negative control. This compound, indeed, is expected, from the docking analysis (Table ##TAB##1##2##), to be less active. In addition, extracts from <italic>Cupressus pyramidalis </italic>were also used, since this extract does not inhibit NF-kappaB/DNA interactions (data not shown). Finally, the known inhibitory compound <bold>9i </bold>was used as reference molecule. The results of the gel retardation analysis are shown in Figure ##FIG##6##7## and clearly demonstrate that compound <bold>21 </bold>inhibit the molecular interactions between nuclear factors (Figure ##FIG##6##7B##) or isolated NF-kappaB p50 (Figure ##FIG##6##7C## and ##FIG##6##7D##) and a target double stranded oligonucleotide mimicking the NF-kappaB binding sites. This effect was similar to that exhibited by the reference compound <bold>9i</bold>. Interestingly, compound <bold>5 </bold>and extracts from <italic>C. pyramidalis </italic>were found to be inactive (Figure ##FIG##6##7A##), fully in agreement with the docking data summarized in Table ##TAB##1##2##.</p>", "<title>Biological effects of compound 21: inhibition of <italic>Pseudomonas aeruginosa </italic>mediated increase of IL-8 mRNA</title>", "<p>Several experimental model system are available for biological validation of molecules inhibiting NF-kappaB function. In a recent paper we report that decoy oligonucleotides targeting NF-kappaB are powerful inhibitors of <italic>Pseudomoas aeruginosa </italic>mediated induction of IL-8 in cystic fibrosis IB3-1 cells [##REF##18258920##21##]. Besides the importance of these data for the theoretical point of view, our results are of great interest for the practical point of view, suggesting this treatment as a possible strategy for the therapy of inflammation associated with cystic fibrosis.</p>", "<p>When the effects of <italic>P. aeruginosa </italic>infection on the expression of pro-inflammatory genes of IB3-1 cells infected for 4 hours are analysed, the results shown in Figure ##FIG##7##8A## are obtained. In this preliminary experiment the content of RNAs coding for several pro-inflammatory proteins was analysed by RT-PCR. The results obtained indicate that IL-8 mRNA sequences sharply increases following PAO1 infection by 40 folds (range 12–92) in respect to basal level of uninfected cells, assumed to be 1. In addition to IL-8 mRNA, other genes induced by PAO1 are GRO-γ (15 folds, range 9–36), IL-6 (18 folds, range 5–44), IL-1β (5 folds, range 2–8), ICAM-1 (4 folds, range 2–8). On the contrary, very low increase of accumulation of IP-10, RANTES, MIP-1α, TNF-α, IFN-β, TGF-β and IL-10 mRNA was observed under these experimental conditions.</p>", "<p>Since NF-kappaB is one of the most important transcription factors regulating the expression of IL-8 gene [23] and the data reported in Figure ##FIG##6##7## demonstrate that compound <bold>21 </bold>inhibit NF-kappaB/DNA interactions, we tested the activity of this compound in inhibiting the expression of IL-8 gene in IB3-1 cells infected with PAO1. Cells were exposed to different concentrations of compound <bold>21 </bold>and then infected with 150 cfu/cell of PAO1. After 4 hours, cells were harvested, RNA extracted and quantitative RT-PCR analysis performed. The results obtained (Figure ##FIG##7##8B##) demonstrate that compound <bold>21 </bold>is a strong inhibitor of PAO-1 induced accumulation of IL-8 mRNA.</p>" ]

[ "<title>Conclusion</title>", "<p>In the present work, we carried out docking studies on the dataset of 27 molecules found in different plant extracts to NF-kappaB-p50, with the purpose of developing a docking protocol fit for the target under study, eventually applicable for more time-consuming virtual screening of larger database of compounds.</p>", "<p>Usually, docking to protein structures that do not have a ligand present, as in the case of NF-kappaB, dramatically reduces the expected performance of structure-based methods. Therefore, the use of NF-kappaB as a target for virtual docking of natural compounds is not feasible. To overcome such a limitation, we proposed to enhance the simple docking procedure by means of a sort of combined target- and ligand-based drug design approach. Advantages of this combination strategy, based on a similarity parameter for the identification of weak binding chemical entities, are illustrated in this work with the discovery of a new lead compound for NF-kappaB. In this respect, this paper represents the first example of successfully individuation of a potential lead compound through molecular docking simulations on a NF-kappaB target. At the same time, information derived from this structure and its different binding modes, could carry through further lead optimization to more potent NF-kappaB inhibitors.</p>", "<p>In order to validate the approach, biochemical analyses based on EMSA were performed on compound <bold>21</bold>; the results obtained sustain the concept that the docking performance is predictive of a biochemical activity (Figure ##FIG##6##7##).</p>", "<p>Our results are of interest also from the practical point of view. The transcription factor NF-kappaB is indeed a very interesting target molecules in the design on anti-tumor, anti-inflammatory, pro-apoptotic drugs.</p>", "<p>In order to validate this last hypothesis, we have employed human cystic fibrosis IB3-1 tracheal epithelial cells. We have elsewhere reported that these cells activate, upon exposure to the bacterium <italic>Pseudomonas aeruginosa </italic>(the PAO-1 strain), the expression of several pro-inflammatory genes, including those coding interleukin-6 (IL-6) and interleukin-8 (IL-8). As supported by several groups, the expression of IL-8 is dependent from NF-kappaB activation. Accordingly, decoy molecules targeting NF-kappaB are strong inhibitors of the IL-8 expression. Therefore, PAO-1 infected IB3-1 cells are a very interesting model system to screen for IL-8 inhibitors. The results of our experiments, in agreement with both docking and EMSA data, demonstrate that compound <bold>21 </bold>is a strong inhibitor of IL-8 and should be considered of interest for modulation of the expression of this gene.</p>" ]

[ "<p>This is an Open Access article distributed under the terms of the Creative Commons Attribution License (<ext-link ext-link-type=\"uri\" xlink:href=\"http://creativecommons.org/licenses/by/2.0\"/>), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.</p>", "<title>Background</title>", "<p>The transcription factor NF-kappaB is a very interesting target molecule for the design on anti-tumor, anti-inflammatory and pro-apoptotic drugs. However, the application of the widely-used molecular docking computational method for the virtual screening of chemical libraries on NF-kappaB is not yet reported in literature. Docking studies on a dataset of 27 molecules from extracts of two different medicinal plants to NF-kappaB-p50 were performed with the purpose of developing a docking protocol fit for the target under study.</p>", "<title>Results</title>", "<p>We enhanced the simple docking procedure by means of a sort of combined target- and ligand-based drug design approach. Advantages of this combination strategy, based on a similarity parameter for the identification of weak binding chemical entities, are illustrated in this work with the discovery of a new lead compound for NF-kappaB. Further biochemical analyses based on EMSA were performed and biological effects were tested on the compound exhibiting the best docking score. All experimental analysis were in fairly good agreement with molecular modeling findings.</p>", "<title>Conclusion</title>", "<p>The results obtained sustain the concept that the docking performance is predictive of a biochemical activity. In this respect, this paper represents the first example of successfully individuation through molecular docking simulations of a promising lead compound for the inhibition of NF-kappaB-p50 biological activity and modulation of the expression of the NF-kB regulated IL8 gene.</p>" ]

[ "<title>Authors' contributions</title>", "<p>LP carried out all the bioinformatic procedures and the docking experiments. EF participated to the EMSA assays; MB purified the nuclear factors for EMSA analysis; I.M. performed semi-quantitative RT-PCR analysis; VB performed the treatment of IB3-1 cells with selected compounds; MCD performed infection IB3-1 cells with P. aeruginosa; EN performed quantitative RT-PCR analysis of IB-8 mRNA; GC was the responsible of the conception, design, analysis and interpretation of the data on cystic fibrosis cell lines; RG was the responsible of the coordination of the project and of the drafting of the manuscript. All authors read and approved the final manuscript.</p>" ]

[ "<title>Acknowledgements</title>", "<p>RG is granted by Fondazione CARIPARO, AIRC, Telethon and by MUR COFIN-2005. VB is fellow of the \"Fondazione Cariverona\", FQ and EN are fellows of the \"Azienda Ospedaliera di Verona\". This work was supported also by grants from the Italian Cystic Fibrosis Research Foundation (grants # 15/2004 and # 13/2007 to RG and GC) and Fondazione Cariverona – Bando 2005 – Malattie rare e della povertà (to GC).</p>" ]

[ "<fig position=\"float\" id=\"F1\"><label>Figure 1</label><caption><p>Structures of compounds found in <italic>Cupressus pyramidalis and Aegle marmelos </italic>extracts and used for docking simulations.</p></caption></fig>", "<fig position=\"float\" id=\"F2\"><label>Figure 2</label><caption><p>NF-kappaB/DNA binding inhibitors used for atom-pair similarity scoring in docking.</p></caption></fig>", "<fig position=\"float\" id=\"F3\"><label>Figure 3</label><caption><p>Stereoview of compounds <bold>1–27</bold>, <bold>9i </bold>and <bold>10i </bold>docked in to DNA binding region of the NF-kappaB p50 monomer chain A. The macromolecule is highlighted in green.</p></caption></fig>", "<fig position=\"float\" id=\"F4\"><label>Figure 4</label><caption><p>Superimposition of the docked poses of inhibitors <bold>9i</bold>, <bold>10i </bold>and compound <bold>21</bold>. A. the DNA binding site of NF-kappaB p50 (monomer chain A) is highlighted in green; B. the ligand atoms involved in hydrogen bonding are labeled. Compounds <bold>9i</bold>, (shown in purple), <bold>10i </bold>(shown in blue) and <bold>21 </bold>are illustrated in stick representations.</p></caption></fig>", "<fig position=\"float\" id=\"F5\"><label>Figure 5</label><caption><p>Binding modes of <bold>21 </bold>docked in to the active site of NF-kappaB p50: A. monomer and B. homodimer (chain A, shown in green). The residues involved in the interaction with the ligand are shown; the hydrogen bonding and the relative distances are indicated in purple.</p></caption></fig>", "<fig position=\"float\" id=\"F6\"><label>Figure 6</label><caption><p>Poses of docked compound <bold>21 </bold>in to the DNA binding region of NF-kappaB p50. A. monomer and B. homodimer (chain A, shown in grey). The inhibitory activity of <bold>21 </bold>may be due to its ability to form a stable complex with the active conformation of the dimer and/or blocking the interaction of DNA with the monomer filling the protein binding site. The DNA was obtained from the crystal structure of the homodimer NF-kappaB (pdb code:<ext-link ext-link-type=\"pdb\" xlink:href=\"1NFK\">1NFK</ext-link>). The surface of the protein is represented in wire frame, the ligand and DNA are highlighted in VDW and stick representation respectively.</p></caption></fig>", "<fig position=\"float\" id=\"F7\"><label>Figure 7</label><caption><p>EMSA analysis. NF-kappaB p50 has been incubated for 15 min in the presence of increasing amounts of compounds <bold>5 </bold>(right side of panel A), extracts from <italic>C. pyramidalis </italic>(left side of panel A), compound <bold>21 </bold>(B-D) and reference compound <bold>9i </bold>(right side of panel C). After the incubation of compounds (or <italic>C. pyramidalis </italic>extracts) to NF-kappaB p50, a further 15 min incubation step was carried on with a ³²P-labelled double stranded oligonucleotide carrying the NF-kappaB binding sites. <italic>C. pyramidalis </italic>extracts were employed at a concentration ranging from 15 to 1000 μg/reaction as mentioned. Comparison between compound <bold>21</bold>, compound <bold>5</bold>, and compound <bold>9i </bold>was performed at 0.01, 0.1, 1 an 10 mM concentrations. Concentration-dependency of the effects displayed by compound <bold>21 </bold>was further analysed with intermediate dosages in the experiment depicted in panel D. The NF-kappaB/DNA complex were identified after polyacrylamide gel electrophoresis. Arrows indicate NF-kappaB/DNA complexes. Asterisks indicate free target DNA.</p></caption></fig>", "<fig position=\"float\" id=\"F8\"><label>Figure 8</label><caption><p>A. Effects of PAO-1 infection of cystic fibrosis IB3-1 cells on the expression of the indicated mRNA. Cells were infected with PAO-1 for 4 hours and then the mRNA analysed by RT-PCR. For RT-PCR analysis the PCR primers have been described in Bezzerri et al. [##REF##18258920##21##] B. Effects of compound 21 on the accumulation of IL-8 mRNA. Cells were incubated for 24 hours in the presence of the indicated concentrations of compound 21, then infected with 150 cfu/cell of PAO-1. Accumulation of IL-8 mRNA was determined by RT-PCR analysis [##REF##18258920##21##].</p></caption></fig>" ]

[ "<table-wrap position=\"float\" id=\"T1\"><label>Table 1</label><caption><p>Ranking of the poses of references inhibitory molecules (1i-8i and 11i-12i) in the target NF-kappaB p50 both as dimer (p50-p50) and as monomers (p50 A and B).</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td align=\"center\" colspan=\"6\"><bold>Docking protocol 1</bold></td></tr></thead><tbody><tr><td align=\"center\" colspan=\"2\"><bold><italic>p50-p50</italic></bold></td><td align=\"center\" colspan=\"2\"><bold><italic>p50 (A)</italic></bold></td><td align=\"center\" colspan=\"2\"><bold><italic>p50 (B)</italic></bold></td></tr><tr><td colspan=\"2\"><hr/></td><td colspan=\"2\"><hr/></td><td colspan=\"2\"><hr/></td></tr><tr><td align=\"center\"><bold>Cpd.</bold></td><td align=\"center\"><bold>G-Score</bold></td><td align=\"center\"><bold>Cpd.</bold></td><td align=\"center\"><bold>G-Score</bold></td><td align=\"center\"><bold>Cpd</bold></td><td align=\"center\"><bold>G-Score</bold></td></tr><tr><td colspan=\"6\"><hr/></td></tr><tr><td align=\"center\"><bold>8i</bold></td><td align=\"center\">-5.88</td><td align=\"center\"><bold>8i</bold></td><td align=\"center\">-6.06</td><td align=\"center\"><bold>1i</bold></td><td align=\"center\">-6.08</td></tr><tr><td align=\"center\"><bold>7i</bold></td><td align=\"center\">-5.88</td><td align=\"center\"><bold>7i</bold></td><td align=\"center\">-5.83</td><td align=\"center\"><bold>3i</bold></td><td align=\"center\">-5.35</td></tr><tr><td align=\"center\"><bold>2i</bold></td><td align=\"center\">-5.84</td><td align=\"center\"><bold>11i</bold></td><td align=\"center\">-5.20</td><td align=\"center\"><bold>7i</bold></td><td align=\"center\">-5.06</td></tr><tr><td align=\"center\"><bold>3i</bold></td><td align=\"center\">-5.78</td><td align=\"center\"><bold>1i</bold></td><td align=\"center\">-5.07</td><td align=\"center\"><bold>2i</bold></td><td align=\"center\">-4.58</td></tr><tr><td align=\"center\"><bold>11i</bold></td><td align=\"center\">-3.87</td><td align=\"center\"><bold>2i</bold></td><td align=\"center\">-2.79</td><td align=\"center\"><bold>11i</bold></td><td align=\"center\">-3.59</td></tr><tr><td align=\"center\"><bold>1i</bold></td><td align=\"center\">-3.57</td><td align=\"center\"><bold>3i</bold></td><td align=\"center\">-2.77</td><td align=\"center\"><bold>4i</bold></td><td align=\"center\">-1.82</td></tr><tr><td align=\"center\"><bold>4i</bold></td><td align=\"center\">-0.76</td><td align=\"center\"><bold>4i</bold></td><td align=\"center\">&gt; 0</td><td align=\"center\"><bold>8i</bold></td><td align=\"center\">&gt; 0</td></tr><tr><td align=\"center\"><bold>12i</bold></td><td align=\"center\">&gt; 0</td><td align=\"center\"><bold>12i</bold></td><td align=\"center\">&gt; 0</td><td align=\"center\"><bold>12i</bold></td><td align=\"center\">&gt; 0</td></tr><tr><td align=\"center\"><bold>5i</bold></td><td align=\"center\">-</td><td align=\"center\"><bold>5i</bold></td><td align=\"center\">&gt; 0</td><td align=\"center\"><bold>5i</bold></td><td align=\"center\">-</td></tr><tr><td align=\"center\"><bold>6i</bold></td><td align=\"center\">-</td><td align=\"center\"><bold>6i</bold></td><td align=\"center\">&gt; 0</td><td align=\"center\"><bold>6i</bold></td><td align=\"center\">-</td></tr></tbody></table></table-wrap>", "<table-wrap position=\"float\" id=\"T2\"><label>Table 2</label><caption><p>Ranking of the poses of natural compounds and test set inhibitors (9i and 10i) in the target NF-kappaB p50 both as dimer (p50-p50) and as monomers (p50 A and B). In the docking protocol 1 the similarity scoring algorithm is not used.</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td align=\"center\" colspan=\"2\"><bold>Docking protocol 1</bold></td><td align=\"center\" colspan=\"6\"><bold>Docking protocol 2</bold></td></tr><tr><td colspan=\"2\"><hr/></td><td colspan=\"6\"><hr/></td></tr><tr><td align=\"center\" colspan=\"2\"><bold><italic>p50-p50</italic></bold></td><td align=\"center\" colspan=\"2\"><bold><italic>p50-p50</italic></bold></td><td align=\"center\" colspan=\"2\"><bold><italic>p50 (A)</italic></bold></td><td align=\"center\" colspan=\"2\"><bold><italic>p50 (B)</italic></bold></td></tr><tr><td colspan=\"2\"><hr/></td><td colspan=\"2\"><hr/></td><td colspan=\"2\"><hr/></td><td colspan=\"2\"><hr/></td></tr><tr><td align=\"center\"><bold>Cpd.</bold></td><td align=\"center\"><bold>G-score</bold></td><td/><td align=\"center\"><bold>G-Score</bold></td><td align=\"center\"><bold>Cpd.</bold></td><td align=\"center\"><bold>G-Score</bold></td><td align=\"center\"><bold>Cpd.</bold></td><td align=\"center\"><bold>G-Score</bold></td></tr></thead><tbody><tr><td align=\"center\"><bold><italic>10i</italic></bold></td><td align=\"center\"><italic>-5.61</italic></td><td align=\"center\"><bold><italic>21</italic></bold></td><td align=\"center\"><italic>-2.34</italic></td><td align=\"center\"><bold><italic>9i</italic></bold></td><td align=\"center\"><italic>-3.81</italic></td><td align=\"center\"><bold><italic>9i</italic></bold></td><td align=\"center\"><italic>-2.13</italic></td></tr><tr><td align=\"center\"><bold>15</bold></td><td align=\"center\">-5.50</td><td align=\"center\"><bold><italic>9i</italic></bold></td><td align=\"center\"><italic>-2.29</italic></td><td align=\"center\"><bold><italic>10i</italic></bold></td><td align=\"center\"><italic>-2.27</italic></td><td align=\"center\"><bold><italic>10i</italic></bold></td><td align=\"center\"><italic>-1.43</italic></td></tr><tr><td align=\"center\"><bold>25</bold></td><td align=\"center\">-5.24</td><td align=\"center\"><bold><italic>10i</italic></bold></td><td align=\"center\"><italic>-2.19</italic></td><td align=\"center\"><bold><italic>21</italic></bold></td><td align=\"center\"><italic>-1.40</italic></td><td align=\"center\"><bold><italic>21</italic></bold></td><td align=\"center\"><italic>-0.32</italic></td></tr><tr><td align=\"center\"><bold>23</bold></td><td align=\"center\">-4.94</td><td align=\"center\"><bold>18</bold></td><td align=\"center\">-0.50</td><td align=\"center\"><bold>18</bold></td><td align=\"center\">-0.07</td><td align=\"center\"><bold>18</bold></td><td align=\"center\">&gt; 0</td></tr><tr><td align=\"center\"><bold><italic>21</italic></bold></td><td align=\"center\"><italic>-4.91</italic></td><td align=\"center\"><bold>23</bold></td><td align=\"center\">-0.05</td><td align=\"center\"><bold>10</bold></td><td align=\"center\">&gt; 0</td><td align=\"center\"><bold>15</bold></td><td align=\"center\">&gt; 0</td></tr><tr><td align=\"center\"><bold>24</bold></td><td align=\"center\">-4.74</td><td align=\"center\"><bold>20</bold></td><td align=\"center\">&gt; 0</td><td align=\"center\"><bold>23</bold></td><td align=\"center\">&gt; 0</td><td align=\"center\"><bold>23</bold></td><td align=\"center\">&gt; 0</td></tr><tr><td align=\"center\"><bold><italic>9i</italic></bold></td><td align=\"center\"><italic>-4.44</italic></td><td align=\"center\"><bold>15</bold></td><td align=\"center\">&gt; 0</td><td align=\"center\"><bold>7</bold></td><td align=\"center\">&gt; 0</td><td align=\"center\"><bold>24</bold></td><td align=\"center\">&gt; 0</td></tr><tr><td align=\"center\"><bold>10</bold></td><td align=\"center\">-4.40</td><td align=\"center\"><bold>25</bold></td><td align=\"center\">&gt; 0</td><td align=\"center\"><bold>8</bold></td><td align=\"center\">&gt; 0</td><td align=\"center\"><bold>10</bold></td><td align=\"center\">&gt; 0</td></tr><tr><td align=\"center\"><bold>20</bold></td><td align=\"center\">-4.37</td><td align=\"center\"><bold>24</bold></td><td align=\"center\">&gt; 0</td><td align=\"center\"><bold>20</bold></td><td align=\"center\">&gt; 0</td><td align=\"center\"><bold>23</bold></td><td align=\"center\">&gt; 0</td></tr><tr><td align=\"center\"><bold>7</bold></td><td align=\"center\">-4.28</td><td align=\"center\"><bold>10</bold></td><td align=\"center\">&gt; 0</td><td align=\"center\"><bold>15</bold></td><td align=\"center\">&gt; 0</td><td align=\"center\"><bold>20</bold></td><td align=\"center\">&gt; 0</td></tr><tr><td align=\"center\"><bold>22</bold></td><td align=\"center\">-4.13</td><td align=\"center\"><bold>16</bold></td><td align=\"center\">&gt; 0</td><td align=\"center\"><bold>25</bold></td><td align=\"center\">&gt; 0</td><td align=\"center\"><bold>16</bold></td><td align=\"center\">&gt; 0</td></tr><tr><td align=\"center\"><bold>16</bold></td><td align=\"center\">-3.48</td><td align=\"center\"><bold>7</bold></td><td align=\"center\">&gt; 0</td><td align=\"center\"><bold>24</bold></td><td align=\"center\">&gt; 0</td><td align=\"center\"><bold>7</bold></td><td align=\"center\">&gt; 0</td></tr><tr><td align=\"center\"><bold>19</bold></td><td align=\"center\">-3.71</td><td align=\"center\"><bold>22</bold></td><td align=\"center\">&gt; 0</td><td align=\"center\"><bold>13</bold></td><td align=\"center\">&gt; 0</td><td align=\"center\"><bold>8</bold></td><td align=\"center\">&gt; 0</td></tr><tr><td align=\"center\"><bold>2</bold></td><td align=\"center\">-3.39</td><td align=\"center\"><bold>19</bold></td><td align=\"center\">&gt; 0</td><td align=\"center\"><bold>12</bold></td><td align=\"center\">&gt; 0</td><td align=\"center\"><bold>2</bold></td><td align=\"center\">&gt; 0</td></tr><tr><td align=\"center\"><bold>27</bold></td><td align=\"center\">-3.09</td><td align=\"center\"><bold>4</bold></td><td align=\"center\">&gt; 0</td><td align=\"center\"><bold>16</bold></td><td align=\"center\">&gt; 0</td><td align=\"center\"><bold>22</bold></td><td align=\"center\">&gt; 0</td></tr><tr><td align=\"center\"><bold>5</bold></td><td align=\"center\">-3.05</td><td align=\"center\"><bold>2</bold></td><td align=\"center\">&gt; 0</td><td align=\"center\"><bold>6</bold></td><td align=\"center\">&gt; 0</td><td align=\"center\"><bold>4</bold></td><td align=\"center\">&gt; 0</td></tr><tr><td align=\"center\"><bold>26</bold></td><td align=\"center\">-2.89</td><td align=\"center\"><bold>5</bold></td><td align=\"center\">&gt; 0</td><td align=\"center\"><bold>9</bold></td><td align=\"center\">&gt; 0</td><td align=\"center\"><bold>1</bold></td><td align=\"center\">&gt; 0</td></tr><tr><td align=\"center\"><bold>18</bold></td><td align=\"center\">-2.69</td><td align=\"center\"><bold>27</bold></td><td align=\"center\">&gt; 0</td><td align=\"center\"><bold>11</bold></td><td align=\"center\">&gt; 0</td><td align=\"center\"><bold>19</bold></td><td align=\"center\">&gt; 0</td></tr><tr><td align=\"center\"><bold>14</bold></td><td align=\"center\">-2.47</td><td align=\"center\"><bold>26</bold></td><td align=\"center\">&gt; 0</td><td align=\"center\"><bold>2</bold></td><td align=\"center\">&gt; 0</td><td align=\"center\"><bold>14</bold></td><td align=\"center\">&gt; 0</td></tr><tr><td align=\"center\"><bold>4</bold></td><td align=\"center\">-2.00</td><td align=\"center\"><bold>14</bold></td><td align=\"center\">&gt; 0</td><td align=\"center\"><bold>3</bold></td><td align=\"center\">&gt; 0</td><td align=\"center\"><bold>5</bold></td><td align=\"center\">&gt; 0</td></tr></tbody></table></table-wrap>", "<table-wrap position=\"float\" id=\"T3\"><label>Table 3</label><caption><p>intramolecular hydrogen bonds of the docked poses of 9i, 10i and 21 with the involved residues of the DNA binding region of NF-kappaB (see Figure 5 for ligand atom labels). The interatomic distances in Angstroms are shown.</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td align=\"center\">Residue interaction</td><td align=\"center\">Ligand atom</td><td align=\"center\">Distance (Å)</td></tr></thead><tbody><tr><td align=\"center\"><bold>(His141) </bold>-Nε-H::O</td><td align=\"center\">O5 <bold>(9i)</bold></td><td align=\"center\">2.12</td></tr><tr><td/><td align=\"center\">O8 <bold>(10i)</bold></td><td align=\"center\">1.88</td></tr><tr><td/><td align=\"center\">O2' <bold>(21)</bold></td><td align=\"center\">1.90</td></tr><tr><td colspan=\"3\"><hr/></td></tr><tr><td align=\"center\"><bold>(Asp239) </bold>-COO::H</td><td align=\"center\">H5 <bold>(9i)</bold></td><td align=\"center\">1.91</td></tr><tr><td/><td align=\"center\">H8 <bold>(10i)</bold></td><td align=\"center\">2.09</td></tr><tr><td/><td align=\"center\">H2' <bold>(21)</bold></td><td align=\"center\">2.00</td></tr><tr><td colspan=\"3\"><hr/></td></tr><tr><td align=\"center\"><bold>(Lys241) </bold>-CO::H</td><td align=\"center\">H7 <bold>(10i)</bold></td><td align=\"center\">1.96</td></tr><tr><td colspan=\"3\"><hr/></td></tr><tr><td align=\"center\"><bold>(Lys241) </bold>-N⁺-H::O</td><td align=\"center\">O7 <bold>(9i)</bold></td><td align=\"center\">2.60</td></tr><tr><td/><td align=\"center\">O1a <bold>(21)</bold></td><td align=\"center\">2.17</td></tr><tr><td colspan=\"3\"><hr/></td></tr><tr><td align=\"center\"><bold>(Leu207) </bold>-N-H::O</td><td align=\"center\">O1 <bold>(10i)</bold></td><td align=\"center\">2.18</td></tr></tbody></table></table-wrap>" ]

[]

[]

[]

[]

[]

[]

[]

[ "<graphic xlink:href=\"1472-6807-8-38-1\"/>", "<graphic xlink:href=\"1472-6807-8-38-2\"/>", "<graphic xlink:href=\"1472-6807-8-38-3\"/>", "<graphic xlink:href=\"1472-6807-8-38-4\"/>", "<graphic xlink:href=\"1472-6807-8-38-5\"/>", "<graphic xlink:href=\"1472-6807-8-38-6\"/>", "<graphic xlink:href=\"1472-6807-8-38-7\"/>", "<graphic xlink:href=\"1472-6807-8-38-8\"/>" ]

[]

{ "acronym": [], "definition": [] }

[ "<title>Background</title>", "<p>Animal models are essential to investigate aspects of human placentation [##UREF##0##1##]. Thanks to striking similarities, the guinea pig <italic>Cavia porcellus </italic>is regarded as an appropriate model for haemochorial placentation, the associated fetomaternal exchange and placental growth processes as well as trophoblast invasion by analogy to the cell columns in the human placenta [##UREF##0##1##, ####UREF##1##2##, ##REF##9778122##3##, ##REF##9481785##4##, ##UREF##2##5##, ##REF##17382996##6##, ##REF##17915313##7####17915313##7##]. To expand the pool of potential models for specific topics, several of its relatives were recently studied [##REF##17915313##7##, ####REF##12061858##8##, ##REF##15081638##9##, ##REF##12950036##10##, ##REF##17016808##11##, ##REF##17607703##12##, ##REF##15737234##13##, ##REF##16310042##14##, ##REF##16448832##15####16448832##15##], supplementing previous work [##REF##14432010##16##, ####REF##13907426##17##, ##REF##1246967##18####1246967##18##]. As a result, a diversity of placental characters have now been identified for guinea pig-related rodents. However, some of the closest relatives of the guinea pig, species of the genus <italic>Galea </italic>[##REF##11861886##19##,##REF##16085429##20##], have not yet been studied. Here, we investigated placental development in the prea, <italic>Galea spixii </italic>Wagler, 1831 (Fig. ##FIG##0##1##), analyzing stages from initial pregnancy to term by means of histology, immunohistochemistry, proliferation activity and electron microscopy. With a standard length of 22.5 to 23.5 cm and an average weight of 375 g and 405 g in males and females respectively, this species is much smaller than the guinea pig. In addition,<italic>Galea </italic>exhibits lower litter sizes of only 2 to 4 young and a shorter gestation period of 48 days [##UREF##3##21##, ####UREF##4##22##, ##UREF##5##23####5##23##]. This species is associated with the semidry to arid caatinga vegetation of northeast Brazil [##UREF##3##21##,##UREF##6##24##]. For reproduction, females build nests among rocks and vegetation. <italic>Galea </italic>is a highly social animal that breeds at different times of the year, even in the dry summer season in the absence of reasonable food abundance. Thus, in addition to understanding of placental development carried out here,<italic>Galea </italic>may also provide insights into pregnancy success under deprived nutritional conditions with respect to climate change.</p>" ]

[ "<title>Methods</title>", "<p>Material of determinate age was collected from a breeding group at the Universidade Federal Rural do Semi-Árido, Mossoró, covering 18 individual from initial pregnancy to term (Table ##TAB##0##1##). Controlled breeding was undertaken and successful copulations were regarded as day zero of pregnancy. The experimental protocol was approved by the Bioethics Committee of the School of Veterinary Medicine, University of Sao Paulo. Primary fixation was mostly done by perfusion with 2.5% glutaraldehyde via the maternal system in situ after the animals were euthanized. Tissues for histology of all relevant placental areas were transferred to 10% formalin in 0.1 M phosphate buffer for 48 h, embedded in paraplast and sectioned at 5 μm using an automatic microtome (Leica RM2155, Germany). Sections were stained with haematoxylin and eosin (HE), Masson's trichrome and the periodic acid-Schiff reaction (PAS) and were examined with an Olympus BX40 microscope (Zeiss KS400 image analysis system 3.4). Tissues for transmission electron microscopy, i.e. samples from various regions such as the labyrinth, trophospongium or subplacenta, were maintained in 2.5% glutaraldehyde for 48 h, post-fixated for 2 h in 2% phosphate-buffered osmium tetroxide (pH 7.4) and embedded in Spurr's Resin. Ultrathin sections were made on an automatic ultramicrotome (Ultracut R, Leica Microsystems, Germany), contrasted with 2% uranyl acetate and 0.5% lead citrate and studied in a transmission electron microscope (Morgagni 268D, FEI Company, The Netherlands; Mega View III camera, Soft Imaging System, Germany). Immunohistochemistry was performed for vimentin to identify mesenchymal cells and stromal decidua and for α-smooth muscle actin to identify vessel walls, following methods established as standard for guinea pig-related rodents [##REF##9778122##3##,##REF##17382996##6##, ####REF##17915313##7##, ##REF##12061858##8####12061858##8##,##REF##15737234##13##, ####REF##16310042##14##, ##REF##16448832##15####16448832##15##,##REF##18512684##25##]. Additional immunohistochemistry was carried out for cytokeratin to identify epithelial cells and trophoblast, but this method did not result in fully specific reactions for guinea pig-related rodents [##REF##9778122##3##,##REF##17915313##7##]. As a proliferation marker, a mouse monoclonal antibody to human anti-PCNA (proliferating cell nuclear antigen) was used. Sections were rehydrated in an ethanol series during the course of which they were submitted to endogenous peroxidase blockage in 3% hydrogen peroxide (v/v) in ethanol for 20 minutes. They were then placed in 0.1 M citrate buffer, pH 6.0, and submitted to microwave irradiation at 700 MHz for fifteen minutes. Sections were equilibrated in 0.1 M phosphate-buffered saline (PBS), pH 7.4, and non-specific binding was blocked using Dako Protein Block for 20 minutes. Tissues were incubated with primary antibodies overnight at 4°C in a humid chamber. Mouse monoclonal anti-human primary antibodies were used to detect vimentin; the cytokeratin was detected by a rabbit polyclonal antibody, α-smooth muscle actin and the PCNA were performed by mouse monoclonal anti-human primaries antibodies. The slices were then rinsed in PBS and incubated with the biotinylated secondary antibody for 45 minutes, followed by streptavidin-HRP for 45 minutes). After rinsing in PBS, the binding was visualized using aminoethyl carbazole as the chromagen. Sections were counterstained with haematoxylin and mounted in Faramount^®. Negative controls were performed using goat anti-Mouse IgG as the primary antibody solution. For complementary data see Table ##TAB##1##2##. Readers unfamiliar with placental terminology may refer to Kaufmann and Davidoff [##REF##602862##27##] and Kaufmann [##UREF##2##5##] for background information.</p>" ]

[ "<title>Results</title>", "<title>Macroscopy</title>", "<p>The chorioallantoic placenta is situated at the antimesometrial side of the uterus (Fig. ##FIG##1##2A##). The placenta establishes itself in initial pregnancy, i.e. before day 9 of pregnancy (Fig. ##FIG##1##2B##, see table ##TAB##0##1##). In this phase, it consists of trophospongium without fetal vessels, made by cellular and syncytial trophoblast (Figs. ##FIG##1##2B##, ##FIG##2##3A##). Strict borders to the decidua do not occur. In subsequent stages of early pregnancy, covering days 9 to 12 (see table ##TAB##0##1##), a disk with clearly defined borders to the decidua develops, which consists of trophospongium, labyrinth and subplacenta (Fig. ##FIG##1##2C##). This shape is maintained in mid gestation (represented by stages of 16 to 32 days of pregnancy, see table ##TAB##0##1##) and late pregnancy (days 36 to 48, see table ##TAB##0##1##), but in these more advanced stages the placenta is highly lobulated (Fig. ##FIG##1##2A, D##). Moreover, the labyrinth is the most prominent area and trophospongium and subplacenta are reduced in size (Fig. ##FIG##1##2A, D##).</p>", "<title>Growing zones</title>", "<p>In initial pregnancy layers of cellular trophoblast with noticeable borders between individual cells are situated on fetal mesenchyme as a border to the central excavation (Fig. ##FIG##2##3A##). This cytotrophoblast is followed by syncytial trophoblast, characterized by the absence of cell borders separating the cytoplasm from the nuclei, which faces towards the maternal blood channels (Fig. ##FIG##2##3A##). The trophoblast cells have large intercellular spaces in between and in relation to the syncytiotrophoblast (Fig. ##FIG##2##3B##). In early pregnancy, the trophospongium is mostly syncytial, but cytotrophoblast is still present (Fig. ##FIG##3##4A,B##). The trophoblast cells cover internally-directed lamellae of mesenchyme, reaching from the placental margin to the labyrinth (Fig. ##FIG##3##4A,B##). These cells are active in proliferation (Fig. ##FIG##3##4C,D##). The described pattern is preserved in mid gestation, but near term the amount of proliferating cells at the periphery is low. Finally, a second centre of proliferation is present in the labyrinth, which is supplied by fetal vessels (Fig. ##FIG##4##5A##), represented by cellular trophoblast within the syncytial areas (Fig. ##FIG##4##5B##).</p>", "<title>Periphery</title>", "<p>The outer surface is defined by a multilayered, villous parietal yolk sac that is associated with a well-developed Reichert's membrane (Fig. ##FIG##5##6A##). Even in mid gestation and near term the parietal yolk sac appears to a large extent as a multilayered structure (Fig. ##FIG##5##6B##). The parietal yolk sac cells possess electron-dense inclusions, large intracellular spaces and apical microvilli (Fig. ##FIG##5##6B##).</p>", "<title>Subplacenta</title>", "<p>Initially the subplacenta is confluent with the main placenta and in contact with the maternal blood channels at its outer borders (Figs. ##FIG##1##2B##, ##FIG##6##7A##). Later it is a distinct organ (Fig. ##FIG##1##2A,C##). In early pregnancy and in one of the mid gestation stages, the subplacenta possesses both fetal vessels and maternal blood channels (Fig. ##FIG##6##7B##). In these cases only some, but not all of the maternal blood spaces are occluded by detritus. In all other investigated stages of mid gestation, the subplacenta is supplied by fetal vessels only. Near term, the organ is reduced and only remnants of its tissue can be found (Fig. ##FIG##1##2D##). Similar to the main placenta, the subplacenta starts as layers of cellular and syncytial trophoblast located on fetal mesenchyme. From early pregnancy on, the subplacenta is highly folded (Figs. ##FIG##1##2A,C##, ##FIG##6##7B,C##, ##FIG##7##8A##). The cytotrophoblast is highly proliferative (Figs. ##FIG##6##7D##, ##FIG##7##8A,B##). The syncytiotrophoblast contains high levels of 1–2 glycol (Fig. ##FIG##6##7C##), which indicates the presence of carbohydrates such as glycogen or mucopolysacharides as well as glycoproteins or proteoglycans. Originating from the subplacenta, strands of extraplacental trophoplast and syncytial streamers continue to the decidua (Figs. ##FIG##1##2B,C##, ##FIG##6##7A##) and the maternal blood channels. They are widespread in early pregnancy, but rare in mid gestation and absent near term (Fig. ##FIG##1##2A,D##).</p>", "<title>Labyrinth</title>", "<p>A labyrinth as an interface between the fetal and maternal blood systems arises as a distinct region in early pregnancy (Fig. ##FIG##1##2C##). At this stage, the endothelium of the maternal arteries has been destroyed and removed by trophoblast, as revealed by immunostaining for vimentin to identify mesenchymal cells and smooth α-actin to trace vessel walls (Fig. ##FIG##8##9A,B##). Positive records found by performing cytokeratin reveal this finding (but see methods and [##REF##17915313##7##]). By contrast, fetal capillaries are enclosed by endothelium (Figs. ##FIG##8##9B##, ##FIG##9##10A–C##). The capillaries are associated with cellular trophoblast possessing large intercellular spaces and some syncytial trophoblast towards the maternal blood channels (Fig. ##FIG##9##10A##). From mid gestation on, syncytial trophoblast is more frequent and becomes dominant (Fig. ##FIG##9##10B##). The syncytiotroblast between the fetal and maternal blood system is, partly, very thin (Fig. ##FIG##9##10B,C##). However, even near term some cellular trophoblast is present (Fig. ##FIG##9##10C##).</p>" ]

[ "<title>Discussion</title>", "<p>Placental development in <italic>Galea spixii </italic>is, in principle, equivalent to that of the guinea pig. Similarities include a highly lobulated, labyrinthine placenta that includes labyrinth, trophospongium and subplacenta as distinct areas with a characteristic structure [##UREF##2##5##,##REF##12950036##10##,##REF##17016808##11##], an invasive haemochorial placental type with a thin interhaemal barrier [##UREF##0##1##, ####UREF##1##2##, ##REF##9778122##3####9778122##3##,##UREF##2##5##,##REF##17915313##7##,##REF##12950036##10##], trophoblast invasion related to the subplacenta that is functionally analogous to the cell columns in the human placenta [##UREF##1##2##,##UREF##2##5##,##REF##17915313##7##], trophoblast cell clusters as growing zones at the periphery and in the labyrinth [##REF##17382996##6##] and the presence of the parietal yolk sac and Reichert's membrane [##UREF##2##5##]. These features are typical for placentation in caviomorph rodents and represent the ancestral condition or stem species pattern of the group [##REF##17382996##6##,##REF##12950036##10##,##REF##17607703##12##,##REF##17586042##26##,##REF##18586321##33##]. However, two differences were observed. Primarily, <italic>Cavia </italic>possesses no coexistence between the fetal and maternal blood supply of the subplacenta [##UREF##2##5##,##REF##602862##27##, ####UREF##7##28##, ##UREF##8##29####8##29##]: supply by the maternal arterial system is stopped by coagulation, resulting in detritus-filled spaces and collapse of the lacunes, once the fetal capillaries achieve access to the subplacenta. This condition was regarded as typical for caviomorphs, although mostly only a few stages were studied [##REF##12061858##8##, ####REF##15081638##9##, ##REF##12950036##10##, ##REF##17016808##11##, ##REF##17607703##12##, ##REF##15737234##13##, ##REF##16310042##14##, ##REF##16448832##15##, ##REF##14432010##16##, ##REF##13907426##17##, ##REF##1246967##18####1246967##18##]. In two individuals of <italic>Galea </italic>in early pregnancy and at mid gestation, both fetal and maternal blood systems are present inside the subplacenta with only some, but not all, of the maternal blood spaces occluded by detritus. The data indicate coexistence between both blood systems. Such a condition was recently documented for another caviomorph species, the degu <italic>Octodon degus</italic>, in which fetal vessels first arise while maternal blood lacunae are still present [##REF##17016808##11##]. Moreover, very recent data indicate an overlap for the capybara too [Kanashiro et al., pers. comm.]. In all three species the maternal and fetal blood systems are separated by several layers of fetal mesenchyme, cytotrophoblast and syncytiotrophoblast, suggesting that the labyrinth is more appropriate for fetomaternal exchange [##REF##17016808##11##]. Thus far, no functional or physiological significance for the overlap in the blood supply of the subplacenta is available. However, according to PAS-staining the subplacental syncytiotrophoblast accumulates glycogen or related substances such as mucopolysacharides, glycoproteins or proteoglycans. This may indicate a secretory function, e.g. for growth factors, hormones or cytokines which might be given to the fetal unit to shut down physiological functions that are not needed during fetal life. Otherwise it could have been accumulated after degeneration of the subplacenta and may be involved in the forthcoming birth, as was previously discussed by several authors [##UREF##2##5##,##REF##17016808##11##,##REF##16448832##15##,##UREF##8##29##, ####UREF##9##30##, ##REF##16740154##31##, ##UREF##10##32####10##32##]. Finally, even in late pregnancy the parietal yolk sac is partly multilayered in <italic>Galea </italic>and related species [##REF##17607703##12##, ####REF##15737234##13##, ##REF##16310042##14##, ##REF##16448832##15##, ##REF##14432010##16##, ##REF##13907426##17##, ##REF##1246967##18####1246967##18##], whereas this condition has been lost in <italic>Cavia </italic>[##UREF##2##5##]. In this case, <italic>Galea </italic>exhibits the caviomorph stem species pattern, whereas <italic>Cavia </italic>possesses a derived character condition [##REF##17586042##26##].</p>" ]

[ "<title>Conclusion</title>", "<p>Development of the placenta in <italic>Galea </italic>exhibits major parallels to the guinea pig and other caviomorphs. Due to similarities in invasion and the expanding processes, it may serve as an additional animal model for human placentation. <italic>Galea </italic>is smaller than the guinea pig and may thus be beneficial for some projects. Moreover, it could serve as an alternative where the center of origin for trophoblast invasion temporarily has access to both maternal and fetal blood systems during pregnancy.</p>" ]

[ "<p>This is an Open Access article distributed under the terms of the Creative Commons Attribution License (<ext-link ext-link-type=\"uri\" xlink:href=\"http://creativecommons.org/licenses/by/2.0\"/>), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.</p>", "<title>Background</title>", "<p>Placentas of guinea pig-related rodents are appropriate animal models for human placentation because of their striking similarities to those of humans. To optimize the pool of potential models in this context, it is essential to identify the occurrence of characters in close relatives.</p>", "<title>Methods</title>", "<p>In this study we first analyzed chorioallantoic placentation in the prea, Galea spixii, as one of the guinea pig's closest relatives. Material was collected from a breeding group at the University of Mossoró, Brazil, including 18 individuals covering an ontogenetic sequence from initial pregnancy to term. Placentas were investigated by means of histology, electron microscopy, immunohistochemistry (vimentin, α-smooth muscle actin, cytokeration) and proliferation activity (PCNA).</p>", "<title>Results</title>", "<p>Placentation in Galea is primarily characterized by an apparent regionalization into labyrinth, trophospongium and subplacenta. It also has associated growing processes with clusters of proliferating trophoblast cells at the placental margin, internally directed projections and a second centre of proliferation in the labyrinth. Finally, the subplacenta, which is temporarily supplied in parallel by the maternal and fetal blood systems, served as the center of origin for trophoblast invasion.</p>", "<title>Conclusion</title>", "<p>Placentation in Galea reveals major parallels to the guinea pig and other caviomorphs with respect to the regionalization of the placenta, the associated growing processes, as well as trophoblast invasion. A principal difference compared to the guinea pig occurred in the blood supply of the subplacenta. Characteristics of the invasion and expanding processes indicate that Galea may serve as an additional animal model that is much smaller than the guinea pig and where the subplacenta partly has access to both maternal and fetal blood systems.</p>" ]

[ "<title>List of all abbreviations</title>", "<p>Ce: central excavation; ct: cellular trophoblast; end: endothelium; evt: extraplacenta trophoplast; fc: fetal capillary; fm: fetal mesenchyme; fv: fetal vessel; lab: labyrinth; mbc: maternal blood channel; pys: parietal yolk sac; rm: Reichert's membrane; sp: subplacenta; sys: syncytial streamers; syt: syncytial trophoblast; tr: trophoblast; tsp: trophospongium; uc: umbilical cord; ut: uterine tissue; vys: visceral yolk sac</p>", "<title>Competing interests</title>", "<p>The authors declare that they have no competing interests.</p>", "<title>Authors' contributions</title>", "<p>MO established the breeding group including the determination of the stages, helped by AD. The material was processed by MO, AM and CA, and immunohistochemistry and semi thin preparation was performed by PF. AM wrote major parts of the manuscript, supported by CA and MAM.</p>" ]

[ "<title>Acknowledgements</title>", "<p>We are grateful for technical support to several members of the University Sao Paulo, Brazil, the Universidade Federal Rural do Semi-Árido, Mossoró, Brazil and the Humboldt-University of Berlin, Germany. Finally, we want to thank Jason Dunlop for help with the English revision.</p>" ]

[ "<fig position=\"float\" id=\"F1\"><label>Figure 1</label><caption><p><bold>The prea, <italic>Galea spixii</italic></bold>. An adult individual at the Universidade Federal Rural do Semi-Árido, Mossoró, Brazil. As a close relative of the guinea pig, this species has a standard length of only 22.5 to 23.5 cm, an average weight of just 375 to 405 g as well as a gestation period of 48 days and litter sizes of only 2 to 4 young [##UREF##3##21##, ####UREF##4##22##, ##UREF##5##23##, ##UREF##6##24####6##24##].</p></caption></fig>", "<fig position=\"float\" id=\"F2\"><label>Figure 2</label><caption><p><bold>Schematic view of the fetal membranes and the placenta</bold>. (A) The schematic view demonstrates the arrangement of the mesometrially situated embryo, the chorioallantoic placenta, positioned at the antimesometrial side, and the other fetal membranes inside the uterus, representing an advanced stage of pregnancy. (B) In initial pregnancy, the placenta contains mainly of trophospongium with the developing subplacenta confluent to the main placenta. Strings of extraplacental trophoblast and syncytial streamers can be followed from towards the maternal blood channels. (C) In early pregnancy, a labyrinth is established and the subplacenta represents a distinct organ. (D) Near term, the placenta is highly villous with the labyrinth as the dominant area. The subplacenta has been reduced. Red numbers in white boxes refer to subsequent figures with more detailed documentation of specific regions. Ce = central excavation, lab = labyrinth, mbc = maternal blood channel, pys = parietal yolk sac, sp = subplacenta, sys = syncytial streamers, tsp = trophospongium, uc = umbilical cord, ut = uterine tissue, vys = visceral yolk sac.</p></caption></fig>", "<fig position=\"float\" id=\"F3\"><label>Figure 3</label><caption><p><bold>Placental trophospongium in initial pregnancy</bold>. (A) HE. The placenta mostly consists of trophospongium (tsp) around the maternal blood channels (mbc). Layers of cellular trophoblast (ct) are situated on fetal mesenchyme along the central excavation, covered by syncytial trophoblast (syt). (B) TEM. The trophoblast cells have large intercellular spaces in between and towards the syncytiotrophoblast. Scale bars = 0.1 mm for histology and 2 μm for TEM.</p></caption></fig>", "<fig position=\"float\" id=\"F4\"><label>Figure 4</label><caption><p><bold>The trophospongium in early pregnancy</bold>. (A,B) HE. Cellular trophoblast (ct) at the placental margin (fetal side, opposite the decidua) covers internally-directed lamellae of fetal mesenchyme (arrow), situated near to the maternal blood channels (mbc). The placental surface is defined by a villous parietal yolk sac (pys). (C) TEM. Cells in the trophospongium show mitotic figures, signifying proliferation activity. (D) Immunostaining for PCNA, indicated by red nuclei, likewise reveals high proliferation activity of the trophoblast cells. Scale bars = 0.1 mm for histology and 2 μm for TEM.</p></caption></fig>", "<fig position=\"float\" id=\"F5\"><label>Figure 5</label><caption><p><bold>Placental organisation in mid gestation</bold>. (A) Immunostaining for α-actin. Only the labyrinth (lab) possesses fetal vessels (red staining of the endothelium), in contrast to the trophospongium (tsp). (B) TEM. The labyrinth includes a second centre of proliferation, represented by trophoblast cells (ct) near the fetal capillaries (fc), whereas syncytial trophoblast faces towards the maternal blood channels (mbc). Scale bars = 0.1 mm for histology and 2 μm for TEM.</p></caption></fig>", "<fig position=\"float\" id=\"F6\"><label>Figure 6</label><caption><p><bold>The parietal yolk sac</bold>. (A) Early pregnancy, PAS, counterstained with HE. The cytotrophoblast below the parietal yolk sac (pys) and the Reichert's membrane (rm) contains glycogen or related substances and may be active in secretion. Below is the trophospongium with its zone of proliferating trophoblast cells (ct) and the maternal blood channels (mbc). (B) Near term, TEM. The parietal yolk sac is still multilayered. It possesses apical microvilli, electron-dense inclusions and large intracellular spaces. Scale bars = 0.1 mm for histology and 2 μm for TEM.</p></caption></fig>", "<fig position=\"float\" id=\"F7\"><label>Figure 7</label><caption><p><bold>The subplacenta</bold>. (A) Initial pregnancy, HE. Originating from the subplacenta (sp), extraplacenta or extravillous trophoplast (evt) and syncytial streamers (sys) continue to the decidua. (B) Early pregnancy, HE. The subplacenta is supplied by fetal vessels (fv) and maternal blood channels (mbc). (C) Same stage, PAS, counterstained with HE. Each fold inside the subplacenta possesses a band of fetal mesenchyme (fm), including fetal vessels, lined with highly proliferative cytotrophoblast (ct). On the other side, syncytial trophoblast (syt) occurs that has accumulated 1–2 glycol containing substances, e.g. carbohydrates such as glycogen or mucopolysacharides, or glycoproteins or proteoglycans. (D) Same stage, TEM. Subplacental cytotrophoblast is active in proliferation. Scale bars = 0.1 mm for histology and 2 μm for TEM.</p></caption></fig>", "<fig position=\"float\" id=\"F8\"><label>Figure 8</label><caption><p><bold>Proliferation of subplacenta trophoblast</bold>. (A,B) Early pregnancy, PCNA, negative IgG controls below. The subplacenta (sp), which is highly lobed and situated around the central excavation (ce), shows a high degree of proliferation activity, related to the layer of cellular trophoblast (ct). The syncytial trophoblast (syt) on the other side of the lobes faces towards the maternal blood channels (mbc). Some fetal capillaries (fc) are already present. Scale bars = 0.1 mm.</p></caption></fig>", "<fig position=\"float\" id=\"F9\"><label>Figure 9</label><caption><p><bold>The maternal blood channels</bold>. Early pregnancy, immunostaining for vimentin (A) and α-actin (B) reveals the presence of trophoblast (tr) in the walls of the maternal blood channels (mbc) in the labyrinth (negative results), in contrast to fetal capillaries (fc) and larger fetal vessels (fv) that are enclosed by endothelium (red nuclei). Scale bars = 0.1 mm.</p></caption></fig>", "<fig position=\"float\" id=\"F10\"><label>Figure 10</label><caption><p><bold>The fetomaternal interface inside the labyrinth</bold>. (A) Initial pregnancy, TEM. Intact endothelium (end) of a fetal capillary (fc), associated with cytotrophoblast (ct) with large intercellular spaces (arrow) and syncytial trophoblast (syt) towards the maternal blood channels (mbc). (B) Mid gestation, TEM. The interhaemal barrier is very thin. (C) Near term, TEM. Cytotrophoblast is still present in the labyrinth. Scale bars = 2 μm.</p></caption></fig>" ]

[ "<table-wrap position=\"float\" id=\"T1\"><label>Table 1</label><caption><p>Investigated material of <italic>Galea spixii</italic></p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td align=\"center\"><bold>Phase of pregnancy</bold></td><td align=\"center\"><bold>Number of animals</bold></td><td align=\"center\"><bold>Gestation-al days</bold></td><td align=\"center\"><bold>Crown-rump length (CRL)</bold></td><td align=\"center\"><bold>Diameter of the placenta</bold></td></tr></thead><tbody><tr><td align=\"left\">Initial pregnancy</td><td align=\"center\">3</td><td align=\"center\">&lt; 9</td><td align=\"center\">&lt; 10 mm</td><td align=\"center\">4 – 8 mm</td></tr><tr><td align=\"left\">Early pregnancy</td><td align=\"center\">3</td><td align=\"center\">9 – 12</td><td align=\"center\">11 – 28 mm</td><td align=\"center\">9 – 12 mm</td></tr><tr><td align=\"left\">Mid gestation</td><td align=\"center\">7</td><td align=\"center\">16 – 32</td><td align=\"center\">36 – 73 mm</td><td align=\"center\">13 – 18 mm</td></tr><tr><td align=\"left\">Near term to term</td><td align=\"center\">5</td><td align=\"center\">36 – 48</td><td align=\"center\">76 – 110 mm</td><td align=\"center\">20 – 35 mm</td></tr></tbody></table></table-wrap>", "<table-wrap position=\"float\" id=\"T2\"><label>Table 2</label><caption><p>Immunohistochemistry. Complete description data related to the performed immunohistochemistry reactions</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td align=\"left\"><bold>Data Reaction</bold></td><td align=\"left\"><bold>Company</bold></td><td align=\"left\"><bold>Origin</bold></td></tr></thead><tbody><tr><td align=\"left\">Dako Protein Block</td><td align=\"left\">DakoCytomation</td><td align=\"left\">Carpinteria, CA, U.S.A</td></tr><tr><td align=\"left\">Vimentin (Type: V9, sc-6260), Dilution: 1:200</td><td align=\"left\">Santa Cruz Biotechnology</td><td align=\"left\">Santa Cruz, California, U.S.A.</td></tr><tr><td align=\"left\">Cytokeratin (PU071-UP), Dilution 1:500</td><td align=\"left\">Biogenex</td><td align=\"left\">San Ramon, California, U.S.A.</td></tr><tr><td align=\"left\">A-smooth muscle actin (Clone 1A4), Dilution 1:300</td><td align=\"left\">DakoCytomation</td><td align=\"left\">Carpinteria, CA, U.S.A</td></tr><tr><td align=\"left\">PCNA (PC10, sc-56), Dilution 1:800</td><td align=\"left\">Santa Cruz Biotechnology</td><td align=\"left\">Santa Cruz, California, U.S.A.</td></tr><tr><td align=\"left\">Streptavidin-HRP; LSAB^®+ System-HRP,</td><td align=\"left\">DakoCytomation</td><td align=\"left\">Carpinteria, CA, U.S.A.</td></tr><tr><td align=\"left\">Aminoethyl carbazole, AEC Substrate Kit</td><td align=\"left\">Zymed Laboratories</td><td align=\"left\">South San Francisco, CA, U.S.A.</td></tr><tr><td align=\"left\">Faramount</td><td align=\"left\">DakoCytomation,</td><td align=\"left\">Carpinteria, CA, U.S.A.</td></tr><tr><td align=\"left\">Goat anti-Mouse IgG (AP308F) Dilution 1:500</td><td align=\"left\">Chemicon International</td><td align=\"left\">Temecula, CA, USA</td></tr></tbody></table></table-wrap>" ]

[]

[]

[]

[]

[]

[]

[ "<table-wrap-foot><p>Investigated stages used by the authors in this study. All material was derived from a breeding group at the Universidade Federal Rural do Semi-Árido, Mossoró, Rio Grande do Norte, Brasil and is housed at the University of Sao Paulo, Faculty of Veterinary Medicine, Sao Paulo, Brasil.</p></table-wrap-foot>" ]

[ "<graphic xlink:href=\"1477-7827-6-39-1\"/>", "<graphic xlink:href=\"1477-7827-6-39-2\"/>", "<graphic xlink:href=\"1477-7827-6-39-3\"/>", "<graphic xlink:href=\"1477-7827-6-39-4\"/>", "<graphic xlink:href=\"1477-7827-6-39-5\"/>", "<graphic xlink:href=\"1477-7827-6-39-6\"/>", "<graphic xlink:href=\"1477-7827-6-39-7\"/>", "<graphic xlink:href=\"1477-7827-6-39-8\"/>", "<graphic xlink:href=\"1477-7827-6-39-9\"/>", "<graphic xlink:href=\"1477-7827-6-39-10\"/>" ]

[]

{ "acronym": [], "definition": [] }

[ "<title>Background</title>", "<p>Assisted reproductive technology (ART) has been commonly used in infertility treatment over the last two decades. The high cost and relatively low implantation and pregnancy rates (PRs) in in-vitro fertilization (IVF) and intracytoplasmic sperm injection (ICSI) treatment cycles has led to a need to evaluate the predictors of success in these patients. One of the important factors is the endometrial receptivity. Endometrial thickness has been utilized as an indirect indicator for endometrial receptivity and is measured in the midsaggital plane during transvaginal ultrasound, which is considered as both atraumatic and simple [##REF##9080229##1##]. The effect of endometrial thickness on pregnancy rates in ART patients has been evaluated by many authors [##REF##8566245##2##, ####REF##7650143##3##, ##REF##8671501##4##, ##REF##10221242##5##, ##REF##9935132##6##, ##REF##10632422##7##, ##REF##11555457##8##, ##REF##11172840##9##, ##REF##14585884##10##, ##REF##1743341##11####1743341##11##], with controversial results. Some authors demonstrated a higher pregnancy rate at certain endometrial thickness [##REF##7650143##3##,##REF##8671501##4##,##REF##14585884##10##, ####REF##1743341##11##, ##REF##1587952##12####1587952##12##], while others did not show a significant correlation between endometrial thickness and PRs in IVF/ICSI patients [##REF##10221242##5##,##REF##10632422##7##,##REF##11555457##8##]. Other authors reported a threshold of &lt;7 mm and/or &gt;14 mm with a significant reduction in implantation rate and PR [##REF##8566245##2##,##REF##9935132##6##].</p>", "<p>With these controversies, no conclusive cut-off value of endometrial thickness has been established in order to help clinicians in counseling the couple about the outcome. The reason for such controversy could be probably due to a relatively low number of cycles for patients with both extremes of endometrial thicknessess.</p>", "<p>The aim of this study is to determine if there is any effect of endometrial thickness measured on the day of administration of human Chorionic Gonadotrphin (hCG) on pregnancy rate while analyzing large number of cycles, and if so, to identify a cut off value at which pregnancy rate is too low, hence helping clinicians in counceling the couples.</p>" ]

[ "<title>Methods</title>", "<p>All fresh cycles of IVF or ICSI conducted at King Faisal Specialist Hospital and Research Center IVF unit from January 2003 to December 2005 were identified from our electronic database and the charts were reviewed. The study was approved by the Ethics Committee of our hospital. All fresh IVF or ICSI treatment cycles that reached oocyte pick up and embryo transfer within the study period were included, women with known intrauterine anomalies were excluded from the study. Endometrial thickness was not used as a criteria for cancellation. Endometrial thickness was defined as the maximal distance between the echogenic interfaces of the myometrium and the endometrium and was measured in the midsagittal plane by two dimensional transvaginal ultrasound on the day of hCG administration.</p>", "<p>Two protocols for pituitary down regulation were used, long or short protocol as previously described [##REF##9080229##1##]. The medication for stimulation used in all cases was human menopausal gonadotrophin (hMG, Menegon^®, Ferring, Germany). When at least three follicles were ≥ 18 mm, hCG 10,000 units was administered. The endometrial thickness was measured by the same sonographer and documented in the chart. Oocyte retrieval was performed 36 hours later. Fertilization was achieved by IVF or ICSI according to the indication. Cleavage stage embryos were transferred on day 3. Maximum two embryos were transferred under transabdominal ultrasound guidance with a full bladder. The patients were started on IM progesterone injections (Gestone, Nordic Pharma, UK) on the same day of embryo transfer for luteal phase support and continued till pregnancy test on day 15. Clinical pregnancy was confirmed by ultrasound observation of fetal cardiac activity two weeks after positive hCG test.</p>", "<p>The patients were divided into two groups; those who got pregnant (group A) and those who did not (group B). Both groups were compared for the various parameters including age, body mass index (BMI), diagnosis, number of oocyte retrieved, length of stimulation, dose of hMG, fertilization rate, number of cleaved embryos, number of transferred embryos.</p>", "<title>Statistical analysis</title>", "<p>Data were analyzed using SPSS version 14 software (Chicago, Ilin, USA). All tests were two tailed, and p &lt; 0.05 was considered statistically significant. Continuous variables are presented as mean and SD and were tested by student's t-test. Comparisons of proportions were made by the chi-squared test. The effect of endometrial thickness on the pregnancy outcome was studied using multivariate analysis, where all other factors affecting the pregnancy outcome were controlled for. To determine the correlation between endometrial thickness, patient characteristics and treatment characteristics a stepwise logistic regression analysis was performed including (age of the patient, body mass index (BMI), endometrial thickness on day 3 of the cycle, duration of stimulation, dose of hMG needed, number of oocytes retrieved, number of cleaved embryos, and number of embryos transferred). The Receiver operating characteristic (ROC) analysis was used to evaluate an endometrial thickness that can predict pregnancy outcome.</p>" ]

[ "<title>Results</title>", "<p>A total of 2464 cycles were included in the study. Clinical pregnancy rate (PR) was 35.8%. 79% of the patients had undergone the long protocol. The pregnancy rate was 39.4% in the long protocol group vs 22.4% in the short protocol group. Compared to group B, group A patients were younger, required lower dose of hMG, had more medium sized and mature follicles, higher number of oocytes retrieved, higher number of oocytes fertilized, and higher number of cleaved embryos. Both groups had similar BMI, duration of stimulation, baseline endometrial thickness (measured on day 3 of the cycle before the start of hMG), and number of transferred embryos (Table ##TAB##0##1##). There was no statistical difference between the two groups in the primary infertility diagnosis (Table ##TAB##1##2##). Endometrial thickness measured on the day of hCG administration ranged between 5 – 20 mm, and was higher in cycles where pregnancy was achieved, with statistical significance (mean 11.6 vs. 11.3 mm, respectively, p &lt; 0.0001). Pregnancy rate increased from 29.4% among patients with an endometrial thickness of ≤6 mm, to 44.4% among patients with an endometrial thickness of ≥17 mm (Table ##TAB##2##3##). (Figure ##FIG##0##1##) shows the positive linear correlation (r = 0.864) and ROC with an area under the curve (AUC) = 0.55. From this ROC a cut-off value of ≥11 mm would be suggested. When dividing the patients into two groups, group 1 with endometrial thickness of &lt;11 mm, and group 2 with endometrial thickness ≥11 mm, PRs were 30.9% and 38.7% respectively, p = 0.001, RR = 1.25 (95%CI 1.12–1.41) (Table ##TAB##3##4##). Multiple logistic regression analysis indicated significant independent effects of age (P = 0.01), Type of protocol used (P = 0.0001), endometrial thickness on hCG day (P = 0.001), number of oocytes retrieved (P = 0.0001), number of cleaved embryos (P = 0.0001), and number of embryos transferred (P = 0.0001) on pregnancy rates.</p>" ]

[ "<title>Discussion</title>", "<p>This study is to our knowledge so far the largest in regards to sample size that addresses the effect of endometrial thickness on PR. The day of the stimulation cycle on which the endometrial thickness is measured to document adequate endometrial development has varied between authors. The most often used is the measurement taken on the day of hCG administration, but some authors have used the measurement that was taken on the day of oocyte retrieval or the day of embryo transfer in their studies, which makes it difficult to compare between studies. We have used the measurement taken on the day of hCG administration in our data. The change in endometrial thickness occurring during IVF stimulation has been evaluated by several authors [##REF##11555457##8##,##REF##2673844##13##,##REF##2199588##14##]. Grant et al 2007, demonstrated a trend toward significance in the overall change in endometrial thickness between the baseline and that on hCG day [##REF##17239871##15##]. Our results are with agreement to those that reported a positive correlation [##REF##7650143##3##,##REF##8671501##4##,##REF##14585884##10##, ####REF##1743341##11##, ##REF##1587952##12##, ##REF##2673844##13##, ##REF##2199588##14##, ##REF##17239871##15##, ##REF##2056019##16##, ##REF##4076435##17##, ##REF##15705371##18##, ##REF##8419225##19##, ##REF##17207799##20##, ##REF##17081537##21####17081537##21##]. Endometrial thickness measured on the day of hCG administration was higher in cycles where pregnancy was achieved (mean 11.6 vs. 11.3 mm, respectively, p &lt; 0.0001), but the difference is not of clinical significance, because results fell within the range of measurement error. When using a multiple logistic regression analysis to control all other confounding variables, we found an independent effect of endometrial thickness on PR. The uniqueness of this study is that it demonstrated a steady and gradual increase in PR as endometrial thickness increases. Many previous studies reported significant differences in PRs above and below a threshold thickness of 8 – 10 mm, but didn't show a continuous relationship such as we found [##REF##7650143##3##,##REF##8671501##4##,##REF##1743341##11##,##REF##1587952##12##,##REF##8419225##19##]. Although we found a clear positive correlation between endometrial thickness and PR, our PR was 29.4% among patients with ≤ 6 mm endometrial thickness in contrast to Gonen et al 1990 who reported poor PR with endometrial thickness &lt; 6 mm [##REF##2199588##14##]. Furthermore, there were several reports of successful pregnancies resulting from cycles with endometrial thickness of ≤ 4 mm [##REF##9688390##22##] indicating that a thin endometrium does not necessarily preclude the possibility of implantation. Hence cancellation of cycle based on a thin endometrium is unwarranted.</p>", "<p>Some authors suggested a detrimental effect of endometrial thickness of ≥ 14 mm on PR [##REF##9935132##6##]. Our results on the contrary, suggest that PRs are highest for patients with the thickest lining, and are consistent with other recent studies finding no reduction in PRs with very thick endometrium [##REF##2056019##16##,##REF##10632437##23##, ####REF##11937134##24##, ##REF##15568328##25####15568328##25##]. In fact there was a case report of a successful twin pregnancy after IVF with an endometrial thickness of 20 mm [##REF##15237015##26##]. Limitations of our study, it is retrospective in nature, but all patients received hMG for stimulation, hence eliminating the bias that can result from different stimulation medications and their different effects on endometrial proliferation. Similarily the number of embryos transferred was limited to two, unless there was only one embryo available for transfer, to control its effect on PR. The poor predictive value of the ROC analysis makes it difficult to accurately determine a cut-off value, never the less, adequate endometrial development is required for pregnancy to occur, and PR were found to be higher when the endometrium reached at least 11 mm thickness.</p>" ]

[ "<title>Conclusion</title>", "<p>The results of the present study identified a positive linear correlation between endometrial thickness measured on hCG day and PR. Therefore, clinicians must pay close attention to endometrial development as well as to follicle growth. But again cancellation of embryo transfer based on a thin endometrial lining is unwarranted.</p>" ]

[ "<p>This is an Open Access article distributed under the terms of the Creative Commons Attribution License (<ext-link ext-link-type=\"uri\" xlink:href=\"http://creativecommons.org/licenses/by/2.0\"/>), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.</p>", "<title>Background</title>", "<p>To evaluate the relationship between endometrial thickness on day of human chorionic gonadotrophin administration (hCG) and pregnancy outcome in a large number of consecutive in vitro fertilization and embryo transfer (IVF-ET) cycles.</p>", "<title>Methods</title>", "<p>A retrospective cohort study including all patients who had IVF-ET from January 2003–December 2005 conducted at a tertiary center.</p>", "<title>Results</title>", "<p>A total of 2464 cycles were analysed. Pregnancy rate (PR) was 35.8%. PR increased linearly (r = 0.864) from 29.4% among patients with a lining of less than or equal to 6 mm, to 44.4% among patients with a lining of greater than or equal to 17 mm. ROC showed that endometrial thickness is not a good predictor of PR, so a definite cut-off value could not be established (AUC = 0.55).</p>", "<title>Conclusion</title>", "<p>There is a positive linear relationship between the endometrial thickness measured on the day of hCG injection and PR, and is independent of other variables. Hence aiming for a thicker endometrium should be considered.</p>" ]

[ "<title>Competing interests</title>", "<p>The authors declare that they have no competing interests.</p>", "<title>Authors' contributions</title>", "<p>All authors made substantial contribution to conception and design of the research. AA acquired the data and wrote the manuscript. KA performed the statistical analysis and interpreted the data, and SC critically revised the manuscript. Again all authors gave the final approval of the manuscript.</p>" ]

[ "<title>Acknowledgements</title>", "<p>We would like to thank Dr Mohamed Gamal El Din Hassan (PhD) Scientist/Biostatistician from Biostatistics, Epidemiology and Scientific Computing (BESC), Research Center KFSHRC) for his help in the statistical analysis.</p>" ]

[ "<fig position=\"float\" id=\"F1\"><label>Figure 1</label><caption><p>The ROC and linear regression curves.</p></caption></fig>" ]

[ "<table-wrap position=\"float\" id=\"T1\"><label>Table 1</label><caption><p>Demographic data</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td align=\"left\">Characteristics</td><td align=\"center\">Group A<break/>mean ± SD</td><td align=\"center\">Group B<break/>mean ± SD</td><td align=\"center\">P value</td></tr></thead><tbody><tr><td align=\"left\">Number of cycles (n)</td><td align=\"center\">882</td><td align=\"center\">1582</td><td/></tr><tr><td align=\"left\">Age (years)</td><td align=\"center\">30.27 ± 5.53</td><td align=\"center\">31.14 ± 5.38</td><td align=\"center\">0.0001</td></tr><tr><td align=\"left\">BMI (weight kg/height m²)</td><td align=\"center\">28.44 ± 4.58</td><td align=\"center\">28.32 ± 4.42</td><td align=\"center\">0.524</td></tr><tr><td align=\"left\">Long protocol # (%)</td><td align=\"center\">765 (39.4%)</td><td align=\"center\">1177 (60.6%)</td><td align=\"center\">&lt; 0.0001</td></tr><tr><td align=\"left\">Short protocol # (%)</td><td align=\"center\">117 (22.4%)</td><td align=\"center\">405 (77.6%)</td><td/></tr><tr><td align=\"left\">Stimulation length (days)</td><td align=\"center\">10.92 ± 2.63</td><td align=\"center\">10.79 ± 2.46</td><td align=\"center\">0.228</td></tr><tr><td align=\"left\">Dose of hMG (ampoules)</td><td align=\"center\">37.67 ± 15.03</td><td align=\"center\">40.73 ± 16.54</td><td align=\"center\">&lt; .0001</td></tr><tr><td align=\"left\">Endometrial thickness cycle day 3 (mm)</td><td align=\"center\">3.23 ± 1.22</td><td align=\"center\">3.21 ± 1.22</td><td align=\"center\">0.696</td></tr><tr><td align=\"left\">Endometrial thickness hCG day (mm)</td><td align=\"center\">11.64 ± 2.13</td><td align=\"center\">11.26 ± 2.17</td><td align=\"center\">&lt; 0.0001</td></tr><tr><td align=\"left\">Number of medium sized follicles</td><td align=\"center\">8.08 ± 5.33</td><td align=\"center\">7.12 ± 5.31</td><td align=\"center\">&lt; 0.0001</td></tr><tr><td align=\"left\">Number of mature follicles</td><td align=\"center\">7.92 ± 3.51</td><td align=\"center\">7.62 ± 3.69</td><td align=\"center\">0.0475</td></tr><tr><td align=\"left\">Number of oocytes retrieved</td><td align=\"center\">10.51 ± 5.43</td><td align=\"center\">9.86 ± 5.73</td><td align=\"center\">0.006</td></tr><tr><td align=\"left\">Number of fertilized oocytes</td><td align=\"center\">5.79 ± 3.23</td><td align=\"center\">4.97 ± 3.35</td><td align=\"center\">&lt; 0.0001</td></tr><tr><td align=\"left\">Number of embryos</td><td align=\"center\">5.3 ± 2.82</td><td align=\"center\">4.44 ± 2.81</td><td align=\"center\">&lt; 0.0001</td></tr><tr><td align=\"left\">Number of embryos transferred</td><td align=\"center\">1.88 ± 0.37</td><td align=\"center\">1.98 ± 0.26</td><td align=\"center\">0.001</td></tr></tbody></table></table-wrap>", "<table-wrap position=\"float\" id=\"T2\"><label>Table 2</label><caption><p>Diagnostic categories</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td align=\"center\">Diagnosis</td><td align=\"center\">Group A<break/>(882)</td><td align=\"center\">Group B<break/>(1582)</td></tr></thead><tbody><tr><td align=\"center\">Male factor 1763 (71.6%)</td><td align=\"center\">623 (70.6%)</td><td align=\"center\">1140 (72.0%)</td></tr><tr><td align=\"center\">Tubal factor 338 (13.7%)</td><td align=\"center\">114 (13.0%)</td><td align=\"center\">224 (14.2%)</td></tr><tr><td align=\"center\">Unexplained 213 (8.6%)</td><td align=\"center\">85 (9.6%)</td><td align=\"center\">128 (8.1%)</td></tr><tr><td align=\"center\">Others 150 (6.1%)</td><td align=\"center\">60 (6.8%)</td><td align=\"center\">90 (5.7%)</td></tr></tbody></table></table-wrap>", "<table-wrap position=\"float\" id=\"T3\"><label>Table 3</label><caption><p>Pregnancy rates at different endometrial thicknesses</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td align=\"center\">Endometrial thickness on day of HCG</td><td align=\"center\">Group A (n)</td><td align=\"center\">Group B (n)</td><td align=\"center\">Pregnancy rate</td></tr></thead><tbody><tr><td align=\"center\">≤6 mm</td><td align=\"center\">5</td><td align=\"center\">12</td><td align=\"center\">29.40%</td></tr><tr><td align=\"center\">7 mm</td><td align=\"center\">11</td><td align=\"center\">34</td><td align=\"center\">24.40%</td></tr><tr><td align=\"center\">8 mm</td><td align=\"center\">35</td><td align=\"center\">96</td><td align=\"center\">26.70%</td></tr><tr><td align=\"center\">9 mm</td><td align=\"center\">70</td><td align=\"center\">171</td><td align=\"center\">29.00%</td></tr><tr><td align=\"center\">10 mm</td><td align=\"center\">162</td><td align=\"center\">321</td><td align=\"center\">33.50%</td></tr><tr><td align=\"center\">11 mm</td><td align=\"center\">140</td><td align=\"center\">240</td><td align=\"center\">36.80%</td></tr><tr><td align=\"center\">12 mm</td><td align=\"center\">174</td><td align=\"center\">275</td><td align=\"center\">38.80%</td></tr><tr><td align=\"center\">13 mm</td><td align=\"center\">130</td><td align=\"center\">202</td><td align=\"center\">39.20%</td></tr><tr><td align=\"center\">14 mm</td><td align=\"center\">82</td><td align=\"center\">122</td><td align=\"center\">40.20%</td></tr><tr><td align=\"center\">15 mm</td><td align=\"center\">38</td><td align=\"center\">62</td><td align=\"center\">38.00%</td></tr><tr><td align=\"center\">16 mm</td><td align=\"center\">19</td><td align=\"center\">27</td><td align=\"center\">41.30%</td></tr><tr><td align=\"center\">≥17 mm</td><td align=\"center\">16</td><td align=\"center\">20</td><td align=\"center\">44%</td></tr><tr><td align=\"center\">Total</td><td align=\"center\">882</td><td align=\"center\">1582</td><td align=\"center\">35.80%</td></tr></tbody></table></table-wrap>", "<table-wrap position=\"float\" id=\"T4\"><label>Table 4</label><caption><p>Pregnancy rates below and above 11 mm endometrial thickness</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td align=\"center\">Endometrial thickness on day of HCG</td><td align=\"center\">Group A (n)</td><td align=\"center\">Group B (n)</td><td align=\"center\">Pregnancy rate</td></tr></thead><tbody><tr><td align=\"center\">&lt; 11 mm</td><td align=\"center\">283</td><td align=\"center\">634</td><td align=\"center\">30.90%</td></tr><tr><td align=\"center\">≥ 11 mm</td><td align=\"center\">599</td><td align=\"center\">948</td><td align=\"center\">38.70%</td></tr><tr><td align=\"center\">Total</td><td align=\"center\">882</td><td align=\"center\">1582</td><td align=\"center\">35.80%</td></tr></tbody></table></table-wrap>" ]

[]

[]

[]

[]

[]

[]

[ "<table-wrap-foot><p>P = 0.319</p></table-wrap-foot>", "<table-wrap-foot><p>P = 0.001</p><p>RR = 1.25, (95% CI 1.12–1.41)</p></table-wrap-foot>" ]

[ "<graphic xlink:href=\"1477-7827-6-37-1\"/>" ]

[]

{ "acronym": [], "definition": [] }

[ "<title>Background</title>", "<p>Toll-like receptors (TLRs) recognize specific pathogen associated molecular patterns (PAMPs) and serve an essential role in the innate immune system by initiating and directing immune response to microbial pathogens. Human TLRs comprise a large family of 10 proteins with member-specific activators and a complex downstream signalling [##REF##12524386##1##]. TLRs are expressed on various immune cells but are also present on mucosal surfaces of the respiratory, gastrointestinal and urinary tract [##REF##12524386##1##]. Applying different adaptor proteins such as toll-like receptor adaptor molecule 1 (TRIF, TICAM1), myeloid differentiation primary response gene 88 (MyD88), myelin and lymphocyte protein Mal, translocation associated membrane protein (TRAM) and sterile alpha and TIR motif containing (SARM), TLRs activate signalling pathways of mitogen-activated protein kinases, nuclear factor kappa-B (NFêB), signal transducers and activators of transcription (STATs) or the activator protein 1 (AP1) [##REF##12524386##1##, ####REF##10623846##2##, ##REF##15379975##3####15379975##3##]. These signalling cascades result in enhanced secretion of various pro- and anti-inflammatory cytokines such as interferons, tumor necrosis factor α (TNFα) and interleukins IL4, IL8, and IL12 [##REF##12524386##1##,##REF##10623846##2##]. Two studies have described the expression of human TLRs in epithelial cells within the female reproductive tract [##REF##15695310##4##,##REF##15196211##5##]. Other than their importance for the interaction between host and pathogen, the receptors might be involved in mucosal homeostasis as described already for the intestine and colon [##REF##18301257##6##]. TLR3 is implicated in the recognition of dsRNA, mRNA and viruses [##REF##12524386##1##,##REF##14729660##7##], whereas TLR4 is a key component of the initial injury response by reacting towards bacterial endotoxin and multiple endogenous ligands [##REF##17047512##8##]. Recent studies have determined the expression pattern of TLR3 [##REF##15695310##4##,##REF##15935884##9##, ####REF##17292969##10##, ##REF##17043100##11##, ##REF##15214945##12####15214945##12##] and TLR4 [##REF##15695310##4##,##REF##17292969##10##, ####REF##17043100##11##, ##REF##15214945##12##, ##REF##15509642##13##, ##REF##15385480##14####15385480##14##] in the human endometrium, but their possible involvement in the pathogenesis of endometrial diseases associated with inflammation remains to be elucidated.</p>", "<p>Endometriosis is a common benign gynaecological condition of reproductive aged women [reviewed in [##REF##15541453##15##]]. The disease is characterised by endometrial tissue fragments outside the uterine cavity and is associated with pelvic pain, dysmenorrhoea, and infertility. Since aetiology and pathogenesis remain uncertain, different theories are discussed including altered immune function. The deregulation of immune response in endometriosis is characterised by increased number of activated macrophages and their secreted products, such as growth factors, cytokines, and angiogenic factors [##REF##14651748##16##,##REF##17028437##17##]. Young et al. reported an increase in interleukin-8 (IL-8) production after stimulating TLR3 and TLR4 in endometrial cell lines with appropriate ligands [##REF##15214945##12##]. IL-8 is a chemotactic activating cytokine for leukocytes and it has been hypothesized to play a role in the growth and maintenance of ectopic endometrial tissue [##REF##11949939##18##]. Recent studies consider endometriosis as a process of sterile inflammation in the pelvis, which is accompanied by elevated levels of inflammatory key regulators such as TNFα [##REF##11834864##19##] or NF-κβ [##REF##17483545##20##]. Both are known downstream targets of TLRs.</p>", "<p>Endometrial carcinoma is the most common gynaecological malignancy in Europe and North America affecting mainly postmenopausal women [##REF##16084259##21##]. In endometrial tumorigenesis, two different types are characterised: the estrogen-related adenocarcinoma (endometrioid type) and the non-endometrioid type such as papillary serous and clear cell carcinoma [##REF##16084259##21##]. Adenocarcinoma accounts for seventy percent of endometrial cancer and is mostly preceded by premalignant changes like endometrial hyperplasia [##REF##16084259##21##]. The majority of adenocarcinoma expresses steroid receptors and occur in women with risk factors associated with an imbalance of estrogen and progesterone. However, inflammation with production of pro-inflammatory cytokines such as TNFα is known to play an important role in cancer development [##REF##12490959##22##]. In endometrial hyperplasia and adenocarcinoma the expression of NFκB and TFNα has been demonstrated [##REF##16364998##23##] indicating that the production of pro-inflammatory cytokines seem to play a role in endometrial tumorigenesis.</p>", "<p>The present study describes the expression pattern of TLR3 and TLR4 mRNA and proteins in healthy endometrium across the menstrual cycle and in postmenopausal tissue. To assess the possible involvement of these toll-like receptors in endometrial pathologies, their expression pattern was also examined in endometriosis and in adenocarcinoma specimens.</p>" ]

[ "<title>Methods</title>", "<title>Endometrial tissues</title>", "<p>Endometrial tissues were obtained from 55 women with regular menstrual cycles (mean 28 ± 2.2 days) who were undergoing gynaecological procedures for benign conditions at the Department of Gynaecology, University Hospital Essen (table ##TAB##0##1##). In this cohort, 20 women have been diagnosed with endometriosis. Menstrual effluents were collected from women without proven endometriosis during first three days of menstrual bleeding as described elsewhere [##REF##18202125##24##].</p>", "<p>The menstrual cycle phase was characterised by morphologic evaluation following the criteria of Noyes et al. [##UREF##0##25##]: proliferative (P, controls: n = 16, endometriotic: n = 13), secretory (S, controls: n = 11, endometriotic: n = 3) and menstrual (M, n = 8) phase. Additionally, four proliferative corresponding ectopic lesions were included, obtained from the above-characterised cohort. In the premenopausal group, patient age ranged from 19 to 52 years (median: 38, detailed data in table ##TAB##0##1##).</p>", "<p>Postmenopausal endometrium was obtained from 34 women including 10 with endometrial hyperplasia and another 16 with endometrial carcinoma (table ##TAB##0##1##). The remaining 8 patients did not have any endometrial abnormalities and were used as the control group. Patients were considered postmenopausal if they have been in menopause for at least one year. Endometrial adenocarcinoma specimens were classified based on the post-operative histopathologic WHO guidelines [##UREF##1##26##] as follow: grade 1 (G1, well differentiated, n = 5), G2 (moderately differentiated, n = 6), G3 (undifferentiated, n = 5). In the postmenopausal group, patient age ranged from 37 to 86 years (median: 66).</p>", "<p>None of the women included in the studies had received any hormonal treatments for at least three months preceding biopsy and routinely analyzed laboratory parameters from blood samples were physiologically analogous to the patient's age. Considering the leukocyte content and level of C-reactive protein, no systemic inflammation was diagnosed at the time of surgery. Volunteers donating menstrual effluents were healthy and without diagnosed infections.</p>", "<p>Institutional ethical approval was granted for all subjects, and all women provided written informed consent.</p>", "<p>All biopsies were transferred into a buffered saline solution directly after surgery and stored in this buffer for maximal two hours until further use. A portion of the biopsy specimen was fixed in 4% formalin and embedded in paraffin for histology and immunohistochemistry, the remainder was flash-frozen in liquid N₂for RNA extraction.</p>", "<title>Quantitative real-time PCR</title>", "<p>Isolation of total RNA from endometrial tissue and reverse transcription into cDNA were carried out applying standard methods as described previously [##REF##18202125##24##]. Following a DNase digest and reverse transcription, quantitative real-time PCR (qPCR) reactions were performed in triplicates using an ABI Prism 7300 Sequence Detector (Applied Biosystems, Weiterstadt, Germany) in a total volume of 20 μl containing 40 ng cDNA, 3.75 pmol gene-specific primers (table ##TAB##1##2##) and SYBR Green reagent (Applied Biosystems) with ROX dye as passive control for signal intensity. The thermal cycle profile was 10 sec at 95°C, followed by 45 cycles of 5 sec at 95°C and 35 sec at 60°C. Melting curve analysis allowed determination of the specificity of the PCR fragments. All melting curves yielded one peak per PCR product.</p>", "<p>To determine the copy number of PCR fragments, serially diluted, gene specific standard cDNAs generated from amplicons of TLR3, TLR4 and β-actin (ACTB) were used. Applying thermal block cyclers and ethidium bromide gel electrophoresis, standard PCRs were conducted. Each gene-specific PCR resulted in one distinct band of the appropriate length. The amplicons were purified by using a Qiagen kit and cDNA concentration was measured photometrically. For each gene, five different dilutions of standard cDNA were used in real time PCR. Threshold cycles for TLR3 signals were between 26 and 38 and for TLR4 between 25 and 36, respectively. Because of the diversity in the RNA quality, each individual sample was normalized to its ACTB mRNA content as an internal standard. These relative values were used for statistics.</p>", "<title>Immunohistochemistry</title>", "<p>Paraffin-embedded specimens were sectioned at 7 μm, rehydrated and microwaved in 0.01 M sodium citrate buffer, pH 6.0, for 10 min for antigen retrieval. Immunostainings were performed on paraffin sections applying the diaminobenzidine staining method with the VECTASTAIN Elite ABC kit (Vector Laboratories, Burlingame, CA) according to the manufacturer's protocol. Endogenous peroxidase activity was quenched with 0.3% H2O2 in methanol for 10 minutes and washed in buffered saline solution (PBS). Unspecific binding of the first antibody was blocked by 30-minute incubation step in PBS containing 0.15% normal horse serum. Slides were incubated in a humidified chamber overnight at 4°C with the monoclonal mouse-anti-human antibodies against TLR3 [##REF##16011525##27##] and TLR4 [HTA125, [##REF##14745724##28##]] at 20 μg/ml and 100 μg/ml, respectively (Acris Antibodies, Hiddenhausen, Germany). Control samples were carried out by omitting the primary antibody. All sections were counterstained with haematoxylin and documented by using a Zeiss Axiophot microscope (Zeiss, Jena, Germany) with a Nikon DS-U1 camera and the LUCIA Image Analysis software (Nikon, Tokyo, Japan).</p>", "<title>Immunofluorescent staining</title>", "<p>Frozen tissues were sectioned at 7 μm and fixed in 70% ethanol. Unspecific binding of the first antibody was blocked by a 30 min incubation step in 5% BSA/PBS. The TLR4 antibody was incubated as described above and was detected using Alexa Fluor 488-conjugated anti-mouse antibody (3.3 μg/ml, MoBiTec, Goettingen, Germany). Sections were fixed in formalin (4%) for two minutes and then washed in PBS. The incubations with CD14 (10 μg/ml, mouse anti-human, BioLegend, San Diego, CA) or CD163 (10 μg/ml, mouse anti-human, HyCult Biotechnology, Uden, The Netherlands) occurred at room temperature for 60 min. CD14 antigen is expressed on on monocytes/macrophages, acting as a dendritic cells precursor [##REF##11642597##29##]. CD163 is a member of the scavenger receptor cystein-rich family class B and is expressed on most subpopulations of mature tissue macrophages [##REF##16164022##30##]. CD163 is highly abundant in human placenta [##REF##10985978##31##] and is present in shed menstrual endometrium [##REF##18202125##24##]. The secondary, goat anti-mouse antibody was Cy3-conjugated (2.5 μg/ml, Dianova, Munich, Germany) and was applied to the specimens for another 60 minutes. Nuclei were identified by 4',6'-diamidino-2-phenylindole staining (DAPI, Sigma, Munich, Germany) using 0.1 μg/ml DAPI in methanol for 15 min at 37°C. Negative controls were performed by omitting the primary antibody and were used to adjust the background fluorescence.</p>", "<p>After mounting with Mowiol (Sigma), confocal microscopy was performed using a Zeiss Axiovert 100 microscope and LSM 510 system (Zeiss, Jena, Germany). TLR4 was detected at 488 nm, CD14 as well as CD163 at 543 nm, and DAPI at 366 nm, respectively.</p>", "<title>Statistical analysis</title>", "<p>Exploratory data analyses, Kruskal-Wallis test for group comparisons, as well as the Mann-Whitney <italic>U </italic>test for nonparametric independent two-group comparisons were performed with the program SPSS 14 for Windows (SPSS Inc., Chicago, IL). Differences with <italic>P </italic>&lt; 0.05 were regarded as statistically significant, <italic>P </italic>&lt; 0.01 as highly statistically significant. Values of mRNA quantification are given as mean ± standard deviation (SD).</p>" ]

[ "<title>Results</title>", "<title>TLR3 and TLR4 expression is deregulated in peritoneal endometriosis</title>", "<p>Both receptors were expressed in all endometrial biopsies with the averaged TLR4 mRNA levels being higher (20-fold) than TLR3 (figure ##FIG##0##1##). This difference was the greatest in the shed menstrual endometrium, where TLR4 transcripts were 564-fold higher than those for TLR3. The relative abundance of both transcripts did not vary throughout the menstrual cycle (figure ##FIG##0##1##).</p>", "<p>TLR3 and TLR4 proteins were expressed mainly in the luminal and glandular epithelium (figure ##FIG##1##2##). Interestingly, the glands presented a heterogeneous immune staining for TLR3 (figure ##FIG##1##2C, I##). Indeed, the TLR3 receptor was found to be locally expressed in a subset of epithelial cells within one gland. In addition, we report the expression of TLR4 protein on immune cells such as monocytes and macrophages, in menstrual phase samples (figure ##FIG##1##2J##). Co-immunostainings on menstrual effluents confirmed that CD14 positive dendritic cells and monocytes (figure ##FIG##1##2K##) as well as CD163 positive resident macrophages (figure ##FIG##1##2L##) expressed TLR4 protein.</p>", "<p>In endometriosis, we can observe a significant decrease in TLR3 and TLR4 mRNA levels in eutopic tissues collected during the proliferative phase, when compared to controls (<italic>P </italic>&lt; 0.05; figure ##FIG##2##3A–B##). Interestingly, endometriotic lesions in proliferative phase showed a significant increase of TLR3 mRNA expression (<italic>P </italic>&lt; 0.05) when compared with the corresponding eutopic endometrium (figure ##FIG##2##3A##). For the TLR4 transcript, a 6-fold increase was observed in the endometriotic lesions in comparison with the diseased eutopic endometrium (<italic>P </italic>&lt; 0.01; figure ##FIG##2##3B##). In the endometrial tissues collected during the secretory phase, the TLR4 mRNA level tended to be lower in eutopic endometrium than in controls (<italic>P </italic>= 0.08; figure ##FIG##2##3D##).</p>", "<p>Immunostaining analyses confirmed these findings at the protein level. In endometriosis, eutopic tissues revealed weaker staining for TLR3 and TLR4 proteins (figure ##FIG##1##2B, D, F, H##) when compared to controls (figure ##FIG##1##2A, C, E, G##). Figure ##FIG##3##4## exemplary presents the expression of TLR3 and TLR4 protein in eutopic compared to ectopic endometrium from the same patient. The TLR3 (Fig. ##FIG##3##4A##) and TLR4 (Fig. ##FIG##3##4C##) immunostaining in diseased eutopic endometrium was barely detectable, whereas corresponding lesion from the uterosacral ligament showed an intense staining in the glandular epithelium for both proteins (Fig. ##FIG##3##4B## and ##FIG##3##4D##). Concerning protein localization, we found TLR3 and TLR4 proteins in glandular and epithelial cells of endometriosis patients.</p>", "<title>TLR3 and TLR4 are expressed in postmenopausal endometrium and regulated endometrial adenocarcinoma</title>", "<p>TLR3 and TLR4 mRNA abundance in healthy postmenopausal tissues is similar to those found during the menstrual cycle. In postmenopausal controls, TLR4 mRNA levels were higher than those for TLR3 (<italic>P </italic>&lt; 0.05, figure ##FIG##4##5##). TLR3 and TLR4 mRNA expression varied significantly between control, hyperplasia and endometrial adenocarcinoma samples (Kruskal-Wallis test, <italic>P </italic>&lt; 0.01). For both receptors, we observe a significant decrease in mRNA abundance in endometrial hyperplasia and adenocarcinoma samples, when compared to postmenopausal endometrium (<italic>P </italic>&lt; 0.05; figure ##FIG##4##5A–B##). In undifferentiated G3 carcinoma, TLR3 and TLR4 mRNA levels were significantly lower than in postmenopausal controls (<italic>P </italic>&lt; 0.01) and in hyperplasic endometrial tissues (<italic>P </italic>&lt; 0.05, figure ##FIG##4##5C–D##).</p>", "<p>TLR3 and TLR4 proteins in hyperplasia and endometrial carcinoma were mostly localized to the luminal and glandular epithelium (figure ##FIG##5##6##). Additionally, we demonstrate a discontinuous staining for TLR3 protein within epithelial glands of G1 carcinoma (figure ##FIG##5##6C##), comparable to the findings in secretory and menstrual phase of premenopausal women (figure ##FIG##1##2C, I##). In undifferentiated G3 carcinoma, staining for TLR3 (figure ##FIG##5##6E##) and TLR4 (figure ##FIG##5##6J##) was not detectable, strengthening our findings of low TLR3 and TLR4 mRNA abundance in G3 carcinoma (figure ##FIG##4##5B##). In accordance with staining patterns obtained during the menstrual phase (figure ##FIG##1##2J##), we were able to find TLR4 protein localized on immune cells (figure ##FIG##5##6F, G, H, I##). We performed co- immunostainings on controls and on malignant endometrial tissues (G2 carcinoma) with antibodies for CD14 and CD163 (figure ##FIG##6##7##). TLR4 protein was expressed on CD14 positive dendritic cells, and monocytes (figure ##FIG##6##7A, C##), as well as on CD163 positive macrophages (figure ##FIG##6##7B, D##).</p>" ]

[ "<title>Discussion</title>", "<p>In the current study, we report that toll-like receptor 3 and 4 expression is modulated in pathogenic alterations of the endometrium. We also found higher TLR4 expression levels in endometrial samples throughout the menstrual cycle and in postmenopausal biopsies, when compared to those for TLR3. In most tissues including gut, gonads and placenta, TLR3 is greater expressed than TLR4 mRNA [##REF##11777946##32##]. TLR3 recognizes RNA and viruses, whereas TLR4 mediates the response to bacterial endotoxins and is activated due to sterile inflammation [##REF##15585605##33##,##REF##16894359##34##]. Thus, the predominant expression of TLR4, observed in uterine tissues, might reflect the occurrence of sterile inflammation during the menstrual cycle. Moreover, ascending bacterial pathogens could contribute to the TLR4 dominance in the uterus.</p>", "<p>In agreement with earlier reports [##REF##15695310##4##,##REF##15214945##12##], both investigated TLRs were mainly localized in the endometrial epithelium, the site of primary immune response in the uterus. In addition, we were able to detect TLR4 protein on endometrial CD14 and CD163 positive immune cells. We found CD14 mainly expressed within the epithelial layer, only a sporadic number of CD14 positive cells was detected in stromal compartment, probably representing the population of monocytes. A recent study performed on bovine endometrial cells, co-localised TLR4 transcripts with CD14 mRNA and protein to stromal and epithelial cells [##REF##16223858##35##]. CD14 is a known accessory molecule for TLR4 and conducts a downstream signalling cascade via MyD88 [##REF##15379975##3##]. In agreement with Pioli et al., who detected TLR4, CD14 and MyD88 transcripts in human endometrium [##REF##15385480##14##], we were able to co-localize TLR4 with CD14 proteins suggesting the presence of both interacting receptors CD14 and TLR4 in the endometrial cells.</p>", "<p>For the first time, we present endometrial effluents expressing high levels of TLR3 and TLR4 proteins. Since the period of menstruation is accompanied by an increased risk of infections due to ascending microorganisms [##REF##815068##36##], we believe that the increased expression of toll-like receptors may be one of the defense mechanisms used by the uterus. Previous studies reported that toll-like receptors are also implicated in epithelial repair as described for intestinal [##REF##18301257##6##] and alveolar epithelial cells [##REF##16799081##37##]. In damaged tissue, necrosis induced inflammation is thought to trigger danger signals, leading to tissue repair response through TLRs [##REF##18301257##6##]. Since repair processes occur every month in the uterus of premenopausal women, we believe that the interaction between hyaluronan and TLR4 might promote the endometrial repair. Hyaluronan is released by necrotic cells, interacts with TLR4 and activates CD44 mediated signalling [##REF##14764599##38##]. In the endometrium, deposition of hyaluronan has been described in stromal compartment [##REF##11702245##39##]. Moreover, hyaluronan has been reported to be involved in attachment of endometrial cells to the mesothelium as a very early step of endometriosis [##REF##11704126##40##]. Further investigations of hyaluronan-TLR4 signalling in healthy and diseased endometrium would be of interest to gain insight into the functional role of TLR4 expression in the uterus.</p>", "<p>Endometriosis causes chronic inflammatory conditions in the pelvic cavity and in the uterus. This disorder is discussed to be accompanied by an activation of the Th2 type of immune response and a shift from Th1 towards Th2 cytokine production [##REF##17234676##41##]. Interestingly, Th2 cytokines were shown to play an important role in balancing TLR signalling in human intestinal epithelial cells by mediating downregulation of TLR3 and TLR4 expression and function [##REF##16670286##42##]. This could also be the case in the diseased eutopic endometrium, where decreased TLR levels were found. It remains to be fully elucidated, if deregulation of TLR expression is involved in the pathogenesis of endometriosis or if altered TLR expression patterns are a consequence resulting from the presence of endometriotic lesions. We could recently show that uterine gene expression patterns are altered due to the existence of ectopic lesions in a non-human primate model for endometriosis [##REF##16481591##43##,##REF##16672717##44##]. Since implantation is a process accompanied by an inflammatory event, an impaired fertility observed in endometriotic women could be one consequence [##REF##16129944##45##]. Continued studies are needed to determine the role of TLR function in diseased endometrium, which could be a promising path towards a better understanding of the pathogenesis of this disease.</p>", "<p>Interestingly, we found a local upregulation of both TLRs in peritoneal endometriotic lesions when compared to eutopic endometriosis from the same patients. A recent study presented a local upregulation of CD14 and CD163 in ovarian endometriotic lesions [##REF##17640886##46##]. However, we observe locally gained TLR4 expression in epithelial cells as demonstrated in immunohistochemical stainings. We propose that the sterile inflammation process, which occurs in the pelvic cavity upon endometriosis, is able to enhance the epithelial TLR4 expression and thus activate the known downstream signalling cascade. One of the potentially activated TLR-downstream molecules is NF-κB, which was recently found as constitutively elevated in endometriotic lesions [##REF##17483545##20##]. It is established that the activation of NF-κB is linked to proliferation, angiogenesis and enhanced production of inflammatory cytokines on ectopic sites [##REF##17028437##17##]. Hence, the TLR-NF-κB cascade might contribute to the chronic persistence of endometriotic lesions.</p>", "<p>In endometrial adenocarcinoma, expression levels of the downstream molecules TNFα and NF-κB were decreased in G2 and G3 but not in the well-differentiated grade 1 carcinoma [##REF##12237915##47##]. Although the evidence is lacking, the almost negligible levels of both toll-like receptors in G3 endometrial adenocarcinoma may reflect lowered differentiation and possibly indicates poor prognosis.</p>", "<p>Both endometriosis and endometrial adenocarcinoma are estrogen dependent diseases. In both conditions, TLR3 and TLR4 were significantly decreased in diseased endometrium when compared to age matched controls. However, it is known that estrogen did not influence the expression of either TLR3 [##REF##16384532##48##] nor TLR4 [##REF##15261308##49##,##REF##17532054##50##] in epithelial cells of endometrium [##REF##16384532##48##], retina [##REF##17532054##50##] and in macrophages [##REF##15261308##49##]. Thus, additional factors are required to decrease TLR-expression in endometriotic endometrium and in endometrial carcinoma.</p>", "<p>Besides excessive estrogen, genetic predisposition presents one of the risk factors associated with the development of endometrial adenocarcinoma. In our study, we observed high inter-individual differences in TLR expression as mirrored by high standard deviations. The results implicate a possible impact of polymorphisms on mRNA expression in physiologic and pathologic endometrium. Recently, a functional polymorphism of the TLR4 gene, associated with impaired TLR signalling, was considered as a significant risk factor for gastric carcinoma [##REF##17324405##51##]. Another single nucleotide polymorphism in 3'-untranslated region of the same gene has been associated with increased risk for prostate carcinoma [##REF##15087412##52##]. It remains to be fully elucidated, if genetic polymorphisms in genes encoding for toll-like receptors might promote endometrial carcinogenesis.</p>" ]

[ "<title>Conclusion</title>", "<p>Our data suggest an involvement of TLR3 and TLR4 in endometrial diseases as we demonstrated altered expression levels for both receptors in endometriosis and endometrial adenocarcinoma. Healthy and differentiated endometrium seems to require an adequate TLR3 and TLR4 expression. Further studies are necessary to investigate the potential function of both receptors in endometrial diseases.</p>" ]

[ "<p>This is an Open Access article distributed under the terms of the Creative Commons Attribution License (<ext-link ext-link-type=\"uri\" xlink:href=\"http://creativecommons.org/licenses/by/2.0\"/>), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.</p>", "<title>Background</title>", "<p>Toll-like receptors (TLRs) play an essential role in the innate immune system by initiating and directing immune response to pathogens. TLRs are expressed in the human endometrium and their regulation might be crucial for the pathogenesis of endometrial diseases.</p>", "<title>Methods</title>", "<p>TLR3 and TLR4 expression was investigated during the menstrual cycle and in postmenopausal endometrium considering peritoneal endometriosis, hyperplasia, and endometrial adenocarcinoma specimens (grade 1 to 3). The expression studies applied quantitative RT-PCR and immunolabelling of both proteins.</p>", "<title>Results</title>", "<p>TLR3 and TLR4 proteins were mostly localised to the glandular and luminal epithelium. In addition, TLR4 was present on endometrial dendritic cells, monocytes and macrophages. TLR3 and TLR4 mRNA levels did not show significant changes during the menstrual cycle. In patients with peritoneal endometriosis, TLR3 and TLR4 mRNA expression decreased significantly in proliferative diseased endometrium compared to controls. Interestingly, ectopic endometriotic lesions showed a significant increase of TLR3 und TLR4 mRNA expression compared to corresponding eutopic tissues, indicating a local gain of TLR expression. Endometrial hyperplasia and adenocarcinoma revealed significantly reduced receptor levels when compared with postmenopausal controls. The lowest TLR expression levels were determined in poor differentiated carcinoma (grade 3).</p>", "<title>Conclusion</title>", "<p>Our data suggest an involvement of TLR3 and TLR4 in endometrial diseases as demonstrated by altered expression levels in endometriosis and endometrial cancer.</p>" ]

[ "<title>Competing interests</title>", "<p>The authors declare that they have no competing interests.</p>", "<title>Authors' contributions</title>", "<p>SA processed tissue samples, established the TLR-assays, carried out the expression analyses, analyzed data, and drafted the manuscript. CB participated in the design of the study, collected patients' tissues, and was involved in the analyses of data. AAK was involved in tissue processing and expression analyses. RK participated in the design and interpretation of the study. IG conceived the study, participated in its design, coordination, and analysis, and helped to draft the manuscript. All authors have read and approved the final manuscript.</p>" ]

[ "<title>Acknowledgements</title>", "<p>We appreciate the support of Prof. Dr. Elke Winterhager, Institute of Anatomy II, University of Duisburg-Essen. We thank Georgia Rauter for her excellent technical assistance and Claudia Jacobs for her support by managing of patients' data.</p>" ]

[ "<fig position=\"float\" id=\"F1\"><label>Figure 1</label><caption><p><bold>TLR3 and TLR4 transcript are expressed in endometrium during the menstrual cycle</bold>. Columns indicate mean TLR3 and TLR4 mRNA quantities from endometrium in proliferative (n = 16), secretory (n = 11) and menstrual phase (n = 8) run in triplicates. The y-axis is scaled logarithmically; error bars represent the standard deviation of the mean.</p></caption></fig>", "<fig position=\"float\" id=\"F2\"><label>Figure 2</label><caption><p><bold>TLR3 and TLR4 protein is localised to endometrial cells during the menstrual cycle</bold>. TLR3 protein staining in healthy late proliferative (LP) tissue was high in luminal and glandular tissue <bold>(A</bold>, brown precipitate) and lower in LP endometriotic tissue <bold>(B)</bold>. Late secretory (LS) endometrium showed highly expressed TLR3 in the epithelium <bold>(C)</bold>, but weakly in endometriosis <bold>(D)</bold>. Intense staining of TLR4 proteins was shown in mid proliferative (MP) tissue <bold>(E)</bold>. In late proliferative phase of endometriosis, TLR4 proteins were comparably lower <bold>(F)</bold>. TLR4 protein was high in mid secretory (MS) normal endometrium <bold>(G)</bold>, whereas it was decreased in endometriotic MS tissue <bold>(H)</bold>. During the menstrual phase, both TLR3 <bold>(I) </bold>and TLR4 <bold>(J) </bold>were highly expressed. Co-immunostaining for TLR4 (green), CD14 (<bold>K</bold>, red) and CD163 (<bold>L</bold>, red) demonstrated that TLR4 proteins were expressed by CD14 positive dendritic cells and monocytes (<bold>K</bold>, yellow) and by CD163 positive macrophages (<bold>L</bold>, yellow). Localisation of TLR4 to immune cells is marked by a black arrow <bold>(J) </bold>and by white arrows <bold>(K, L)</bold>.</p></caption></fig>", "<fig position=\"float\" id=\"F3\"><label>Figure 3</label><caption><p><bold>TLR3 and TLR4 mRNA expression is regulated in endometriosis</bold>. The expression of TLR3 <bold>(A, C) </bold>and TLR4 mRNA <bold>(B, D) </bold>in endometrium during proliferative (n = 13, run in triplicates, <bold>A, B) </bold>and secretory phase (n = 3, <bold>C, D) </bold>was decreased in eutopic endometriotic endometrium when compared to controls. In addition, four proliferative corresponding lesions were evaluated <bold>(A, B) </bold>showing a local upregulation of both receptors on ectopic sites. Columns represent the mean ratio of TLR copy number to ACTB copy number. Error bars represent the standard deviation of the mean. * P &lt; 0.05; ** P &lt; 0.01.</p></caption></fig>", "<fig position=\"float\" id=\"F4\"><label>Figure 4</label><caption><p><bold>TLR3 and TLR4 protein is locally induced in endometriotic lesions</bold>. No specific TLR3 protein staining was seen in eutopic endometriotic tissue <bold>(A) </bold>whereas a high glandular localisation of the protein was detected in a gland of an ectopic endometriotic lesion from the same patient <bold>(B)</bold>. Similarly, TLR4 was not detectable in eutopic endometrium <bold>(C) </bold>but present in glandular epithelium of ectopic endometrium from the same women <bold>(D)</bold>.</p></caption></fig>", "<fig position=\"float\" id=\"F5\"><label>Figure 5</label><caption><p><bold>TLR3 and TLR4 mRNA expression is decreased in endometrial adenocarcinoma</bold>. <bold>(A-B) </bold>Columns indicate mean TLR3 (A) and TLR4 (B) mRNA levels from postmenopausal patients (PMP, n = 8), and those diagnosed with endometrial hyperplasia (HP, n = 10) and endometrial carcinoma (EnCa, n = 16). <bold>(C-D) </bold>TLR3 (C) and TLR4 (D) mRNA expression in different carcinoma grades compared to postmenopausal controls and hyperplastic endometrium: G1 (n = 5), G2 (n = 6) and G3 (n = 5). Error bars represent the standard deviation of the mean. * P &lt; 0.05; ** P &lt; 0.01.</p></caption></fig>", "<fig position=\"float\" id=\"F6\"><label>Figure 6</label><caption><p><bold>TLR3 and TLR4 proteins are present in postmenopausal endometrium</bold>. Localisation of TLR3 in normal postmenopausal endometrium <bold>(A)</bold>, endometrial hyperplasia <bold>(B)</bold>, endometrial adenocarcinoma grade G1 <bold>(C)</bold>, G2 <bold>(D) </bold>and G3 <bold>(E</bold>). Localisation of TLR4 protein in normal postmenopausal endometrium <bold>(F)</bold>, endometrial hyperplasia <bold>(G)</bold>, endometrial adenocarcinoma grade G1 <bold>(H)</bold>, G2 <bold>(I) </bold>and G3(<bold>J</bold>). All stained sections indicated epithelium as the preferred localisation of TLR3 and TLR4 proteins. TLR4 protein was additionally present in immune cells (arrows).</p></caption></fig>", "<fig position=\"float\" id=\"F7\"><label>Figure 7</label><caption><p><bold>TLR4 is localised to immune cells of postmenopausal endometrium</bold>. Co-Immunostaining of TLR4 with CD14 and CD163 in healthy endometrium <bold>(A, B) </bold>and in adenocarcinoma <bold>(C, D)</bold>. TLR4 proteins were expressed by CD14 positive dendritic cells and monocytes <bold>(A, C) </bold>and by CD163 positive macrophages <bold>(B, D)</bold>. Arrows indicate the co-localisation of TLR4 with the immune cells. Scale bar = 20 μm.</p></caption></fig>" ]

[ "<table-wrap position=\"float\" id=\"T1\"><label>Table 1</label><caption><p>Patients' characteristics according to diagnosis at time of surgery</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td/><td align=\"left\"><bold>n</bold></td><td align=\"left\"><bold>Age, mean</bold></td><td align=\"left\"><bold>Age, SD</bold></td><td align=\"left\"><bold>Indications for surgery</bold></td></tr></thead><tbody><tr><td align=\"left\">Premenopausal, controls (proliferative &amp; secretory)</td><td align=\"left\">27</td><td align=\"left\">37</td><td align=\"left\">8.6</td><td align=\"left\">fibroids (n = 7), non endometriotic ovarian cyst (n = 2), infertility (n = 5), dysmenorrhoe (n = 11), pelvic pain (n = 1), uterine prolapse (n = 1)</td></tr><tr><td align=\"left\">Premenopausal, non-endometriotic, menstrual</td><td align=\"left\">8</td><td align=\"left\">39</td><td align=\"left\">10.7</td><td align=\"left\">no surgery</td></tr><tr><td align=\"left\">Premenopausal, endometriotic</td><td align=\"left\">20</td><td align=\"left\">34</td><td align=\"left\">6.8</td><td align=\"left\">endometriosis (n = 11), ovarian cyst (n = 2), infertility (n = 1), dysmenorrhoe (n = 6*)</td></tr><tr><td align=\"left\">Postmenopausal, controls</td><td align=\"left\">8</td><td align=\"left\">68</td><td align=\"left\">10.2</td><td align=\"left\">fibroids (n = 6), uterine prolapse (n = 2)</td></tr><tr><td align=\"left\">Postmenopausal, hyperplasia</td><td align=\"left\">10</td><td align=\"left\">64</td><td align=\"left\">13</td><td align=\"left\">abnormal endometrial thickness, supposed carcinoma</td></tr><tr><td align=\"left\">Postmenopausal, endometrial carcinoma</td><td align=\"left\">16</td><td align=\"left\">67</td><td align=\"left\">12.8</td><td align=\"left\">abnormal endometrial thickness, endometrial carcinoma</td></tr></tbody></table></table-wrap>", "<table-wrap position=\"float\" id=\"T2\"><label>Table 2</label><caption><p>Oligonucleotide primers used for the quantitative real time PCR.</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td align=\"left\"><bold>Gene (GenBank No.)</bold></td><td align=\"left\"><bold>Forward primers (position)</bold></td><td align=\"left\"><bold>Reverse primers (position)</bold></td></tr></thead><tbody><tr><td align=\"left\">TLR3 (<ext-link ext-link-type=\"gen\" xlink:href=\"NM_003265\">NM_003265</ext-link>)</td><td align=\"left\">5'-GTATTGCCTGGTTTGTTAATTGG (2059–2082)</td><td align=\"left\">5'-AAGAGTTCAAAGGGGGCACT (2215–2194)</td></tr><tr><td align=\"left\">TLR4 (<ext-link ext-link-type=\"gen\" xlink:href=\"NM_138557\">NM_138557</ext-link>)</td><td align=\"left\">5'-AAGCCGAAAGGTGATTGTTG (2187–2206)</td><td align=\"left\">5'-CTGAGCAGGGTCTTCTCCAC (2339–2320)</td></tr><tr><td align=\"left\">ACTB (<ext-link ext-link-type=\"gen\" xlink:href=\"NM_001101\">NM_001101</ext-link>)</td><td align=\"left\">5'-ACCAACTGGGACGACATGGA (302–322)</td><td align=\"left\">5'-CCAGAGGCGTACAGGGATAG (510–491)</td></tr></tbody></table></table-wrap>" ]

[]

[]

[]

[]

[]

[]

[ "<table-wrap-foot><p>n = number of patients, * not including patients with known endometriosis before surgery</p></table-wrap-foot>", "<table-wrap-foot><p>All primers were designed using the Primer3 software.</p></table-wrap-foot>" ]

[ "<graphic xlink:href=\"1477-7827-6-40-1\"/>", "<graphic xlink:href=\"1477-7827-6-40-2\"/>", "<graphic xlink:href=\"1477-7827-6-40-3\"/>", "<graphic xlink:href=\"1477-7827-6-40-4\"/>", "<graphic xlink:href=\"1477-7827-6-40-5\"/>", "<graphic xlink:href=\"1477-7827-6-40-6\"/>", "<graphic xlink:href=\"1477-7827-6-40-7\"/>" ]

[]

{ "acronym": [], "definition": [] }

[ "<title>Background</title>", "<p>Higher plant genomes contain considerable amounts of satellite repeats which make up to 20% of nuclear DNA in some species [##REF##8492802##1##]. Satellite repeats occur in a genome as continuous arrays of tandemly arranged basic repeated units (monomers). Although the monomers are usually only tens to hundreds of nucleotides long, they can accumulate into millions of copies, forming megabase-sized clusters distinguishable as heterochromatic regions on mitotic chromosomes or in interphase nuclei. Satellite repeats undergo rapid evolutionary changes of their sequences and abundance, leading to the frequent occurrence of genus- or species-specific families of satellite DNA [##REF##12675302##2##, ####REF##10905342##3##, ##UREF##0##4##, ##REF##15911575##5####15911575##5##]. Contrary to this diversification observed between various taxa, the repeat monomers are usually well-homogenized within a species. This process of intra-specific sequence homogenization, generally referred to as 'concerted evolution', is supposed to arise from a concurrent action of various molecular mechanisms including unequal crossing-over and gene conversion [##REF##12414190##6##,##REF##3744043##7##]. Although some of these mechanisms have been characterized in detail, their overall contribution to satDNA evolution remains elusive. Moreover, theoretical models and computer simulations suggest that these mechanisms alone cannot account for efficient amplification and long-term persistence of satellites within genomes [##REF##8078581##8##, ####REF##8138169##9##, ##REF##3569882##10####3569882##10##]. Therefore, it is supposed that other processes capable of efficient sequence amplification probably act on satDNA. It has been proposed that they involve extrachromosomal circular DNA (eccDNA) molecules, arising from intra-strand recombination between monomers within satellite arrays and subsequently serving as a template for rolling-circle replication. This process would result in the synthesis of linear DNA fragments composed of multiple copies of the circular template molecules and their reintegration into the genome, thus providing an efficient mechanism for amplification and eventual sequence homogenization of satDNA.</p>", "<p>Although eccDNA has been reported from a wide range of eukaryotic organisms including yeast, <italic>Drosophila</italic>, <italic>Xenopus</italic>, mouse and human [##REF##2079966##11##, ####REF##9050997##12##, ##REF##10490607##13##, ##REF##12799349##14####12799349##14##], there are only a few studies focusing on its formation from satellite repeats [##REF##10490607##13##, ####REF##12799349##14##, ##REF##16547499##15####16547499##15##]. In plants, eccDNA derived from centromeric repeats in <italic>Arabidopsis </italic>[##REF##17612498##16##] and repetitive element Bdm29 in <italic>Brachycome dichromosomatica </italic>[##REF##18088310##17##] have been detected. In spite of this progress in eccDNA research, its formation from a wider range of plant satellite repeats and different species has not been studied so far. Consequently, there is only a little known about the structure of circular DNA molecules in plant genomes and mechanisms of their formation. There is also an interesting question concerning the role of eccDNA in the evolution of monomer size of satellite repeats. Similar to other groups of eukaryotes, plant satellites show a clear preference for monomer sizes in ranges between 135 – 195 bp and their multiples [##REF##11836208##18##]. Although the correspondence of this length with the length of DNA wrapped around nucleosome particles has been pointed out [##UREF##1##19##,##REF##11498581##20##], there is no mechanism known to explain this phenomenon. It has been demonstrated that nucleosomes constrain accessibility of enzymatic apparatus to certain regions of associated DNA [##REF##16506092##21##]. Thus, recombination-based sequence homogenization or excision of eccDNA may be more frequent in more accessible regions (e.g. nucleosome linkers), leading to the emergence of the nucleosome-sized repeated units.</p>", "<p>In this study, we addressed some of the questions raised above by investigating the occurrence and properties of eccDNA molecules derived from satellite repeats in a range of species from three genera of higher plants (Fabaceae, Poaceae, Brassicaceae). The repeats to be studied were selected based on their various monomer lengths and eventual presence of higher-order repeats, in order to follow the importance of these properties for formation and size of eccDNA molecules. Our results demonstrated that formation of eccDNA from satellite repeats is a common phenomenon in higher plants, and that it is strongly dependent on sequence similarity.</p>" ]

[ "<title>Methods</title>", "<title>Plant material and genomic DNA isolation</title>", "<p>Seeds of plants used in this study were obtained from Osiva Boršov, Czech Republic (<italic>V. faba </italic>cv. Merkur, <italic>V. pannonica </italic>cv. Detenická panonská), IPK Gatersleben, Germany (<italic>V. grandiflora </italic>Scop. var. grandiflora, <italic>V. narbonensis </italic>L.), NASC, Loughborough, UK (<italic>Arabidopsis thaliana </italic>Columbia), the Breeding Station at Slapy u Tábora, Czech Republic (<italic>Pisum sativum </italic>cv. Carrera), the Agriculture Research Institute at Kromeříž, Czech Republic (<italic>Vicia sativa </italic>cv. Ebena), and the Crop Research Institute, Prague, Czech Republic (<italic>Secale cereale </italic>cv. Dankovské, <italic>Triticum aestivum </italic>cv. Saxana). Seeds of rice (<italic>Oryza sativa </italic>ssp. <italic>japonica </italic>var. Nipponbare) were kindly provided by Prof. J. Jiang (University of Wisconsin, Madison, USA). Total genomic DNA was extracted from leaves pooled from several plants as described by Dellaporta et al. [##UREF##3##45##]. In <italic>V. faba</italic>, DNA was isolated from young (developing) leaves, mature (one-month-old) leaves or 0.5 cm long root tips in order to compare eccDNA levels in various tissues. The DNA concentration measurements were performed using PicoGreen dye (Invitrogen, USA) according to the manufacturer's recommendations.</p>", "<title>Preparation of circular DNA size markers</title>", "<p>Plasmid-based open circle markers were prepared by cloning <italic>Pst</italic>I-digested lambda DNA (Fermentas International Inc., Canada) into plasmid vector pBluescript II SK+ (Stratagene, USA). Selected plasmid clones of different sizes were isolated and converted from a supercoiled to a open form by nicking activity of DNaseI (Boehringer Mannheim, Germany). The reaction was performed in a mixture consisting of 25 pg of DNaseI, 50 mM Tris-HCl pH 7.5, and 10 mM MgCl₂(total volume of 15 μl), for 15 min at 37°C. Open circle markers of small sizes (558, 930, 1,302, 1,674 bp) were designed using LoxP-directed cloning [##REF##11680722##46##]. Complementary oligonucleotides (5'-GAT CTA TAA CTT CGT ATA ATG TAT GCT ATA CGA AGT TAT G-3', 5'-AAT TCA TAA CTT CGT ATA GCA TAC ATT ATA CGA AGT TAT A-3') were annealed to form a linear double-stranded fragment harboring the LoxP site and single-stranded overhangs compatible with <italic>Bam</italic>HI and <italic>Eco</italic>RI restriction sites. The fragment was cloned into <italic>Bam</italic>HI/<italic>Eco</italic>RI-digested plasmid vector and this construct was further modified by incubation with Cre recombinase (New England BioLabs, USA) and the linear LoxP fragment (in a 20 μl reaction mixture containing 2 μg LoxP vector, 0.5 μM annealed LoxP oligonucleotides, 1 × ligase buffer (Fermentas International Inc.), 10 U Cre recombinase, at 37°C for 30 min and heat-inactivated at 70°C for 10 min), producing linearized plasmid carrying one LoxP sequence terminated with a <italic>Bam</italic>HI and <italic>Eco</italic>RI overhang at each end, respectively. This vector was used for cloning <italic>Bam</italic>HI/<italic>Eco</italic>RI-digested PCR fragments of various lengths amplified from the lambda DNA template using a forward primer (5'-TTG CTG AGG ATC CTG TAC CGG CTG TCT GGT ATG TAT G-3') in combination with one of the following reverse primers (5'-TTG CTG AGA ATT CTC CTC CTG CGA TCC CTT C-3', 5'-TTG CTG AGA ATT CAT CGG CAG GGT GAT CGC-3', 5'-TTG CTG AGA ATT CTG GAA CTG GCG AGC CAT C-3', 5'-TTG CTG AGA ATT CGC GGC TTC AAG CGC AAG-3'). The final constructs thus contained LoxP-PCR fragment-LoxP cassettes which were subsequently released by Cre-mediated recombination in the form of covalently closed circular molecules. The length of these circular DNA markers was 558, 930, 1,302 or 1,674 bp. They were treated with a nicking endonuclease Nt.AlwI (New England BioLabs) to convert them into open circles, and purified by agarose-gel electrophoresis. 50 pg of each of these markers were added into genomic DNA samples prior to 2-D electrophoresis and their positions on the gel were determined using Southern hybridization with the lambda DNA probe.</p>", "<title>EccDNA analysis on two-dimensional agarose gel electrophoresis</title>", "<p>Neutral-neutral 2-D agarose gel electrophoresis was performed as described by Cohen and Lavi [##REF##8628266##22##] with the following modifications. Samples of up to 20 μg of genomic DNA were analyzed (in the case of comparative analysis of eccDNA content in different tissues or in stressed plants, the same DNA amouts were always loaded). The DNA was separated on 0.4% agarose in 1 × TBE buffer at 0.7 V/cm for 18 hr and the lanes with samples were excised and stained in 1 × TBE buffer containing 0.3 μg/ml of ethidium bromide for 2 hr. Stained lanes were placed on a gel support at 90° orientation to the direction of electrophoresis and embedded by 1% agarose supplemented with 0.3 μg/ml of ethidium bromide. The second dimension was run in 1 × TBE buffer, 0.3 μg/ml of ethidium bromide at 4 V/cm for 4 hr. Alternatively, the electrophoresis was run on 0.7% and 2% agarose gels (in some cases, 1.5% and 2% was used) for 21 and 8 hours, respectively, in order to improve resolution of small DNA molecules. In addition, the sensitivity of the assay was increased by treating the samples with Plasmid-safe ATP-dependent DNAse (Epicentre Biotechnologies), which selectively degraded linear DNA fragments, thus allowing equivalents of 80 – 240 μg of undigested genomic DNA to be loaded. The Plasmid-safe ATP-dependent DNAse treatment was preceded by passing high molecular weight genomic DNA through a hypodermic needle (Omnican 100, 0.3 mm in diameter, B. Braun Petzold GmbH, Germany), resulting in its slight shearing which promoted linear DNA degradation. The treatment was performed using 160 – 480 U of Plasmid-safe ATP-dependent DNase in 800 – 2400 μl of reaction buffer overnight at 37°C and stopped by incubation at 70°C for 30 min. Short fragments of degraded nucleic acids were removed using molecular weight cut-off columns (Microcon-30 and Microcon-100, Millipore, USA) and the digestion and purification was repeated once more.</p>", "<p>The presence of single-stranded DNA was tested by dividing purified samples into halves and incubating them with or without 5 – 50 U of mung bean nuclease (MBN, Takara Bio Inc., Japan) in 1 × MBN buffer for 10 min at 37°C. Alternatively, RNaseH (Ambion, USA) directed degradation or a combination of RNaseH followed by MBN treatment was performed to check for the presence of DNA-RNA hybrid molecules. RNaseH treatment was carried out in 100 μl of 1 × RNaseH reaction mixture and 25 U of RNaseH for 1 hr at 37°C. Finally, to remove any contamination that could affect subsequent electrophoretic separation, the samples were purified using Wizard SV Gel and PCR Clean-Up System columns (Promega, USA).</p>", "<title>Southern hybridization</title>", "<p>Following electrophoresis, DNA was transferred onto Hybond-N+ membranes (Amersham Biosciences, USA) by capillary transfer. Hybridization probes for satellite repeats were derived from fragments amplified by PCR using specific primers and genomic DNA as a template (AT_180: 5'-ACC TTC TTC TTG CTT CTC AAA G-3', 5'-GTT GGT TAG TGT TTT GGA GTC G-3'; CentO: 5'-AAA ACA TGA TTT TTG GAC ATA TTG G-3', 5'-TGA CAA AAG TTC GCC GCC-3'; PisTR-B: 5'-ACC CAT GAA ATT TGA TTG-3', 5'-CAA CAT TTT CAT CAT TCA CAC-3'; Afa: 5'-GCA TTT CAA ATG AAC TCT GA-3', 5'-GAT GAT GTG GCT TTG AAT GG-3', Sc119: 5'-CCA GAA TCG GCC AAA AC-3', 5'-CCC GTT TCG TGG ACT ATT AC-3'; FokI: 5'-CAT TAT GGA AGG TAG TCT GTT GTC GAG-3', 5'-CAA GGC TAC CAT CCA TTG GAG-3'; VicTR-A: 5'-TAC ATA AAA GTC AYG AAG TT-3', 5'-TAS TAT AAC AYA AGA YA ATC-3'; VicTR-B: 5'-ATA TAA GTC TTC ARA AAA T-3', 5'-GAA GAC TTA TAT TCA CTT-3'). The probe for IGS-like satellite was prepared by insert amplification from clone S12 [##REF##11289513##47##] using T3/T7 primers and removing surrounding polylinker sequences by restriction digestion and gel purification. The same procedure was used for probe preparation for the VG-V subfamily of VicTR-B repeats (using clone c609 [GenBank:<ext-link ext-link-type=\"gen\" xlink:href=\"DQ139394\">DQ139394</ext-link>]) and for VicTR-A clones c653 [GenBank:<ext-link ext-link-type=\"gen\" xlink:href=\"EU568805\">EU568805</ext-link>] and c666 [GenBank:<ext-link ext-link-type=\"gen\" xlink:href=\"EU568818\">EU568818</ext-link>]. Fragment labeling and hybridization were done using the AlkPhos Direct Kit (Amersham Biosciences), according to the manufacturer's recommendations (hybridization and washing temperatures varied between 50°C and 61°C according to probe AT/GC content). Hybridization specificity for the VG-V subfamily was verified in hybridization using clone c609 as a probe and membranes with clones c605 [GenBank:<ext-link ext-link-type=\"gen\" xlink:href=\"DQ139381\">DQ139381</ext-link>], c788 [GenBank:<ext-link ext-link-type=\"gen\" xlink:href=\"DQ139383\">DQ139383</ext-link>], c606 [GenBank:<ext-link ext-link-type=\"gen\" xlink:href=\"DQ139368\">DQ139368</ext-link>] and c610 [GenBank:<ext-link ext-link-type=\"gen\" xlink:href=\"DQ139388\">DQ139388</ext-link>] [##REF##16788823##26##] representing other VicTR-B subfamilies. The specificity of VicTR-A clones c653 and c666 was confirmed in hybridization using sequences of clones c651 [GenBank:<ext-link ext-link-type=\"gen\" xlink:href=\"EU568803\">EU568803</ext-link>], c652 [GenBank:<ext-link ext-link-type=\"gen\" xlink:href=\"EU568804\">EU568804</ext-link>], c654 [GenBank:<ext-link ext-link-type=\"gen\" xlink:href=\"EU568806\">EU568806</ext-link>], c763 [GenBank:<ext-link ext-link-type=\"gen\" xlink:href=\"EU568830\">EU568830</ext-link>] and c768 [GenBank:<ext-link ext-link-type=\"gen\" xlink:href=\"EU568835\">EU568835</ext-link>] as negative controls. To detect signals, blots were incubated with chemiluminescent substrate (CDP-Star, Amersham Biosciences) and exposed to X-ray film for up to 60 hr.</p>", "<title>Cloning and sequence analysis of VicTR-A repeats</title>", "<p>The preparation of <italic>V. pannonica </italic>and <italic>V. narbonensis </italic>shotgun genomic libraries and screening and sequencing of their clones was performed as described by [##REF##16788823##26##] except that VicTR-A clone P5 [##REF##10905342##3##] was used as a probe to screen the libraries by colony hybridization. The sequences were deposited in GenBank under the accession nos. <ext-link ext-link-type=\"gen\" xlink:href=\"EU568802\">EU568802</ext-link> – <ext-link ext-link-type=\"gen\" xlink:href=\"EU568868\">EU568868</ext-link>. Sequence periodicity analysis based on the concept of nucleotide autocorrelation functions [##REF##10222405##27##] was performed as described previously [##REF##16788823##26##]. Dot-plot sequence comparisons were done using a dotter program [##REF##8566757##48##].</p>" ]

[ "<title>Results</title>", "<p>We employed two-dimensional (2-D) agarose gel electrophoresis [##REF##8628266##22##] followed by Southern blot hybridization to examine the presence of eccDNA in extracts of total genomic DNA from selected plant species. In addition to separating DNA molecules based on their size, 2-D electrophoresis also allows resolution based on their structure, resulting in formation of separated arcs on the gel representing linear and various conformations of circular DNA molecules (Fig. ##FIG##0##1A##). In our initial experiments we performed 2-D electrophoresis in 0.4% agarose in the first dimension, and in 1% agarose supplemented with ethidium bromide in the second dimension, which provided efficient resolution of linear and circular molecules from over 9 kb down to 3 kb as assessed from migration of linear and supercoiled or open circular marker molecules (not shown). These conditions were then used to analyze DNA samples of <italic>Vicia faba</italic>, a species containing highly abundant satellite repeat FokI [##REF##6089113##23##,##REF##8654919##24##] (Table ##TAB##0##1##). A strong arc of linear DNA and a weaker one corresponding to open circular molecules were revealed after hybridization with FokI probe. We used this species further for investigating FokI eccDNA levels in different tissues including young/developing or old leaves and root meristems (Fig. ##FIG##0##1B–D##), and also in leaf tissues stressed by wounding (Fig. ##FIG##0##1E–F##). However, we did not observe significant differences in eccDNA signals between these samples, suggesting that there are comparable levels of circular molecules derived from FokI satellite in all investigated <italic>V. faba </italic>tissues and that these levels are not significantly affected by stress conditions. As only open circular molecules were detected and there was no hybridization signal corresponding to covalently closed circular DNA, we tested whether this could be a result of damaging closed circles during DNA isolation. However, including supercoiled control plasmid DNA into tissue samples at the beginning of the procedure revealed that the isolation does not lead to significant nicking or other degradation of this DNA, and thus that open circles are the predominant form of satellite eccDNA in the examined tissues.</p>", "<p>Following these initial experiments, we performed a detailed survey of the presence of eccDNA derived from nine satellite repeats and three subfamilies in ten plant species (Table ##TAB##0##1##). However, as eccDNA was found to occur at relatively low levels, we improved our isolation protocol by including treatment with Plasmid-safe ATP-dependent DNAse, which specifically degrades linear DNA while leaving covalently closed and nicked circular molecules intact. Removing the majority of linear fragments by this treatment allowed loading of equivalents of up to 240 μg of original undigested genomic DNA onto the gel. We also increased the discriminatory power of the technique towards efficient resolution of short circular molecules in order to exactly determine size differences of circular molecules derived from satellites with shorter monomers. This was achieved by increasing agarose concentrations to 0.7% and 2% in the first and second dimension, respectively, which led to improved resolution of circular DNA from linear fragments (down to 500 bp) and to resolving circular molecules differing by at least 90 bp. Size estimations of eccDNA molecules were done using a mixture of specifically designed markers including open circular molecules produced by Cre/lox recombination and nickase treatment of larger plasmid templates (see Methods for details). This is demonstrated in Figure ##FIG##1##2##, showing detection of VicTR-A satellite in <italic>V. narbonensis</italic>. Although the size of VicTR-A monomers is 69 bp, its eccDNA was found to occur in size steps of about 140 bp, thus corresponding to multiples of 138 bp. Moreover, there was an alternating pattern of stronger and weaker spots, suggesting more frequent excision of circular molecules differing in multiples of 276 bp.</p>", "<p>Extrachromosomal circular molecules were detected for all nine satellite families and three subfamilies tested, and in all cases they occurred as open circles (Fig. ##FIG##2##3##). The size of the eccDNA molecules ranged from over 8 kb down to 500 bp and in most cases their hybridization signals formed discrete spots corresponding to multiples of monomer or higher-order unit lengths (Table ##TAB##0##1##). However, there was often a continuous smear underlying these spots, as shown on Fig. ##FIG##2##3A## for Sc119 satellite of <italic>S. cereale</italic>. Treating the samples with mung bean nuclease, which specifically degrades single-stranded but not double-stranded or nicked DNA almost completely removed the smear, resulting in more round and discrete spots (Fig. ##FIG##2##3B##), thus suggesting that the smear was composed of partially single-stranded circular molecules. In addition to the arcs representing circular and linear DNA molecules, in some cases we encountered hybridization signals forming a relatively fuzzy arc that could not be assigned to linear, covalently closed supercoiled or open circular molecules by co-migration with the respective markers. The signal was reproducibly detected for satellites from <italic>Secale cereale, Vicia sativa </italic>and <italic>Oryza sativa </italic>(Fig. ##FIG##2##3B, D, E##). As it was present even in mung bean nuclease-treated samples, its composition from single-stranded DNA or RNA could be ruled out. Thus, we tested whether it could be affected by treatment with RNaseH or RNaseH followed by mung bean nuclease. However, these treatments had no effect, implying that the extra arc did not include hybrid DNA:RNA molecules (data not shown).</p>", "<p>The series of discrete spots representing open circles differing in size by the monomer length of the respective satellites were detected for all repeats with monomers ranging from 340 bp (Afa of <italic>Triticum aestivum</italic>, Fig. ##FIG##2##3C##) down to 118 bp (Sc119 of <italic>S. cereale</italic>, Fig. ##FIG##2##3A, B##). <italic>V. faba </italic>FokI repeats (monomer size of 59 bp) and all other satellites with shorter monomers produced continuous arcs of hybridization signal (Fig. ##FIG##2##3F, G, H##), which could also be made up of the monomer-spaced spots that, due to the limited resolution of agarose gel electrophoresis, were fused into a continuous smear, even when 1.5% agarose was used in the first dimension instead of 0.7% (Fig. ##FIG##2##3G##). However, there were also several repeats showing more complex patterns of eccDNA signals. In the case of the <italic>O. sativa </italic>CentO repeats, the signals formed short smears instead of focused spots (Fig. ##FIG##2##3E##), which could be explained by the presence of monomer variants differing in sequence length (145, 165 bp) interspersed within the arrays of predominant 155 bp monomers, as reported by Lee et al. [##REF##16987952##25##]. VicTR-B repeats of <italic>V. grandiflora </italic>produced a continuous smear consistent with their short monomer length (38 bp); however, there were also faint spots detected on the smear differing by 186 bp (not shown). Further investigation using specifically selected probe and stringent hybridization conditions revealed that the discrete spots represented a minor sequence subfamily VG-V reported by Macas et al. [##REF##16788823##26##] which is homogenized as a 186 bp higher-order repeat derived from five monomers (Fig. ##FIG##2##3I##).</p>", "<p>The VicTR-A satellite of <italic>V. narbonensis </italic>was found to produce eccDNA in size steps corresponding to dimers (138 bp) and more frequently tetramers (276 bp) of the basic repeated unit of 69 bp (Fig. ##FIG##1##2##). In the related species <italic>V. pannonica</italic>, initial investigation of the VicTR-A repeats, using probe amplified in a PCR reaction with genomic DNA as a template, revealed a different hybridization pattern for their eccDNA, consisting of the smear with spots spaced by about 180 bp which could not be derived from multiples of the monomer length (Fig. ##FIG##2##3J##). As there was only limited information available about sequence variability of VicTR-A satellites in <italic>V. narbonensis </italic>and <italic>V. pannonica</italic>, we constructed whole genome shotgun libraries from these species and screened them for VicTR-A clones which were subsequently sequenced and analyzed. The analysis of 29 <italic>V. narbonensis </italic>VicTR-A clones (19,006 bp in total) using nucleotide autocorrelation functions [##REF##16788823##26##,##REF##10222405##27##] revealed that their sequences are well-homogenized with a basic periodicity of 69 bp. However, some preference for homogenization of the tetramer-based units was also revealed by the increased height of the 276 bp peak (Fig. ##FIG##3##4A##). This is consistent with the observed preferential formation of the tetramer-based eccDNA, although it is not clear why similar correlation was not observed for the dimer periodicity, which was detected in eccDNA sizes but not by the sequence analysis.</p>", "<p>In contrast to <italic>V. narbonensis</italic>, analysis of 26 <italic>V. pannonica </italic>clones (16,234 bp in total) showed that they are considerably less homogenized, as revealed by much lower peak heights (Fig. ##FIG##3##4B##). Their basic periodicity was also 69 bp, however, there was one <italic>V. pannonica </italic>clone (c653) with periodicity of 180 bp and high sequence similarity (&gt; 97%) between the 180 bp repeated units. Comparative analysis with other VicTR-A sequences (Fig. ##FIG##4##5A,B##) revealed that this repeat most probably originated by recombination between AT-rich regions resembling the CAAAA motif, which is supposed to be involved in breakage-reunion of repeated sequences [##UREF##2##28##,##REF##9796101##29##]. Using this clone as a probe confirmed that this VicTR-A subfamily gives rise to the 180 bp-spaced signals on the blots of extrachromosomal circular molecules (Fig. ##FIG##2##3K##), while the other subfamily with 69/138 bp periodicity (clone c666) produces weak signals of continuous smear (Fig. ##FIG##2##3L##). Southern blots of restriction enzyme-digested genomic DNA hybridized with the probes differentiating the two subfamilies confirmed that the 180 bp subfamily occurs in the <italic>V. pannonica </italic>genome along with the previously reported VicTR-A repeats with 69/138 bp repeated units [##REF##10905342##3##] (Fig. ##FIG##4##5C##).</p>" ]

[ "<title>Discussion</title>", "<p>Our results show that all the plant satellite repeats that we investigated are prone to the formation of eccDNA. These results complement similar findings described for insects and animals [##REF##10490607##13##, ####REF##12799349##14##, ##REF##16547499##15####16547499##15##,##REF##8559653##30##], and significantly broaden our knowledge about plant satellite repeats [##REF##17612498##16##,##REF##18088310##17##] by detection of satellite-derived eccDNA in a total of ten species. Investigated satellites differed in their monomer length, proportion in the genome and chromosomal localization (Table ##TAB##0##1##). Since eccDNA was detected for all of them, it can be concluded that these features do not have a crucial impact on the formation of circular molecules. A common feature of all investigated satellites was that their eccDNA occurred in the form of open circles, the double-stranded circular molecules relaxed due to the presence of singe strand nicks. The possibility that these findings resulted from DNA damage during sample preparation was excluded by detection of intact supercoiled control plasmid added to the samples at the beginning of the isolation procedure. Similar control was used for eccDNA isolation from <italic>Xenopus </italic>embryos which was also found to occur as open circles [##REF##10490607##13##].</p>", "<p>In yeast, eccDNA formation requires chromosomal replication as it originates from stalled replication forks [##REF##12783853##31##]. On the other hand, eccDNA production in <italic>Xenopus </italic>is supposed to be uncoupled from DNA replication, although some synthesis requiring replicative polymerases was detected on the newly formed eccDNA [##REF##10490607##13##,##REF##11410662##32##]. The experiments that focused on FokI repeats in <italic>V. faba </italic>revealed comparable levels of eccDNA in mature leaves and young (growing) leaves or root meristems, thus indicating that FokI eccDNA formation is also not tightly linked to DNA replication. We did not observe a significant increase of eccDNA concentration in mechanically damaged leaves, suggesting that its formation is not induced by this sort of stress or DNA degradation processes. It should be noted, however, that due to generally low levels of eccDNA in the investigated tissues and the only semi-quantitative nature of the assay, we could not detect subtle changes in eccDNA concentration.</p>", "<p>Our experimental results, together with previous reports [##REF##12799349##14##,##REF##18088310##17##,##REF##11410662##32##], support the hypothesis that eccDNA is produced by homologous intra-strand recombination between satellite repeat units [##REF##2079966##11##,##REF##5070865##33##]. This process is supposed to result in eccDNA molecule sizes corresponding to the multiples of monomer length, which is consistent with our observations. Moreover, the eccDNA patterns detected for different VicTR-A and VicTR-B subfamilies or higher-order repeats suggest that relatively long regions (tens to hundreds of nucleotides) of high sequence similarity are required for efficient recombination. It has been reported that efficiency of homologous recombination depends on similarity of involved sequences [##REF##8196655##34##, ####REF##8811176##35##, ##REF##15611187##36####15611187##36##] and is proportional to the length of the similarity [##REF##6096689##37##]. If the length or degree of sequence similarity is decreased, the rate of recombination is reduced rapidly. In yeast, a divergence of only 1% between 350 bp substrates caused a 5–23-fold reduction of mitotic and meiotic recombination, and divergence of 15% led to a 700-fold reduction [##REF##10101158##38##]. In plants, 1.6% and 1.9% sequence heterogeneity was found to decrease the frequency of intrachromosomal recombination by 3.6 [##REF##15611187##36##] and 9.6-fold [##REF##15584964##39##], respectively, and the recombination between 585 bp inverted repeat substrates was reduced by about 4–20 fold when the level of divergence increased from 0.5% to 9% [##REF##16507082##40##]. A similar adverse effect of decreasing similarity on the efficiency of eccDNA formation was also evident in the case of HOR units of the subfamily VG-V of <italic>V. grandiflora </italic>VicTR-B repeats. The 186 bp HORs have significantly higher average similarity (89%) than the individual monomers from which they are composed (78%) [##REF##16788823##26##]. Consequently, the eccDNA was detected only in sizes corresponding to multiples of the HOR lengths but not in multiples of the 38 bp monomers (Fig. ##FIG##2##3I##). Other VicTR-B subfamilies from the same species that are homogenized at the level of the monomers produced eccDNA with corresponding size distributions (Fig. ##FIG##2##3H##). Thus, homologous recombination seems to be the major mechanism of eccDNA origin from plant satellite repeats. Other possible recombination mechanisms such as the nonhomologous end joining (NHEJ) repair pathway probably do not contribute significantly to eccDNA production [##REF##17612498##16##]. This conclusion is also consistent with previous studies on yeast, which demonstrated the requirement for RAD52-dependent homologous recombination in the formation of eccDNA from rDNA repeats [##REF##10207108##41##].</p>", "<p>As the eccDNA size distributions of satellites corresponded to multiples of monomer or higher-order repeat lengths we did not find any direct evidence of eccDNA production reflecting the periodicity of nucleosomal structure of chromatin. Nevertheless, it cannot be ruled out that a minor subpopulations of circles arise via this recombination pathway and were not detected under our experimental conditions. Such a mechanism could explain the origin of the VG-V subfamily, which represents a clear example of evolutionary shift from 38 bp monomers to the pentamer-based 186 bp HORs. Moreover, the recombination-based elimination of specific sequence regions was likely involved in the formation of the 180 bp VicTR-A subfamily in <italic>V. pannonica </italic>(Fig. ##FIG##4##5A,B##) which could represent another case of evolution towards the nucleosome-sized monomers. The constraints imposed on the formation of eccDNA by chromatin could also explain some discrepancies observed when comparing sequence periodicity of <italic>V. narbonensis </italic>VicTR-A satellite (Fig. ##FIG##3##4A##) to the size distribution of its eccDNA spots detected on 2-D blots (Fig. ##FIG##1##2##). While the periodicity of the sequenced VicTR-A clones was found to be based on a monomer-sized repeated units, the eccDNA was found to occur in multiples of dimer and preferentially tetramer units. Interestingly, the tetramer peak (276 bp) on the periodicity plot is higher that the one corresponding to the trimer (207 bp), suggesting emergent tetramer-based periodicity. Thus, this satellite might be in a transition stage towards HOR periodicity driven by preferential formation of dimer/tetramer-based eccDNA. On the other hand, we can not exclude that the observed eccDNA pattern could arise from some unknown, less abundant subfamily of VicTR-A repeats with already developed HOR periodicity which we did not detect among the cloned sequences.</p>", "<p>The observed common occurrence of eccDNA is important for proving its role in satellite repeat evolution, but it remains to be investigated to what extent it participates in the processes of satellite repeat amplification and sequence homogenization. In the simplest case, recombination-based excision of eccDNA may only represent a deletion mechanism reducing copy numbers of satellite repeats in the genome. Alternatively, open circular molecules can be further utilized as a replication template, leading to production of long linear stretches on newly synthesized DNA fragments composed from multiple copies of the original circular sequence. Thus, this mechanism would provide both amplification as well as sequence homogenization of satellite DNA. In <italic>Drosophila</italic>, specific circle-with-tail structures of tandemly arranged genes corresponding to <italic>Stellate</italic>, <italic>Suppressor </italic>of <italic>Stellate </italic>and histone genes were observed on 2-D gels, suggesting the occurrence of rolling circle replication of these eccDNA. Such rolling circle intermediates (RCIs) of satellite eccDNA molecules were not, however, observed due to the methodological constraints of 2-D electrophoresis [##REF##16091629##42##]. Nor did we find RCIs of satellite eccDNA in our experimental system. This obstacle could be overcome in future experiments by visualizing the content of samples under the electron microscope, as already successfully done for RCIs of mitochondrial plasmid mp1 in <italic>Chenopodium album </italic>[##REF##12065425##43##] and rDNA rolling circles in <italic>Xenopus </italic>[##REF##4517945##44##].</p>" ]

[ "<title>Conclusion</title>", "<p>This work demonstrated the existence of eccDNA molecules derived from various plant satellite repeats, providing strong support for theoretical models predicting eccDNA as an intermediate in satellite DNA evolution. However, it is yet to be seen to what extent and how the eccDNA is utilized in these processes. Future detailed examination of the molecular basis for recombination events and analysis of replication intermediates should provide a better understanding of the biological principles and constraints involved in these processes.</p>" ]

[ "<p>This is an Open Access article distributed under the terms of the Creative Commons Attribution License (<ext-link ext-link-type=\"uri\" xlink:href=\"http://creativecommons.org/licenses/by/2.0\"/>), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.</p>", "<title>Background</title>", "<p>Satellite repeats represent one of the most dynamic components of higher plant genomes, undergoing rapid evolutionary changes of their nucleotide sequences and abundance in a genome. However, the exact molecular mechanisms driving these changes and their eventual regulation are mostly unknown. It has been proposed that amplification and homogenization of satellite DNA could be facilitated by extrachromosomal circular DNA (eccDNA) molecules originated by recombination-based excision from satellite repeat arrays. While the models including eccDNA are attractive for their potential to explain rapid turnover of satellite DNA, the existence of satellite repeat-derived eccDNA has not yet been systematically studied in a wider range of plant genomes.</p>", "<title>Results</title>", "<p>We performed a survey of eccDNA corresponding to nine different families and three subfamilies of satellite repeats in ten species from various genera of higher plants (<italic>Arabidopsis</italic>, <italic>Oryza</italic>, <italic>Pisum</italic>, <italic>Secale</italic>, <italic>Triticum </italic>and <italic>Vicia</italic>). The repeats selected for this study differed in their monomer length, abundance, and chromosomal localization in individual species. Using two-dimensional agarose gel electrophoresis followed by Southern blotting, eccDNA molecules corresponding to all examined satellites were detected. EccDNA occurred in the form of nicked circles ranging from hundreds to over eight thousand nucleotides in size. Within this range the circular molecules occurred preferentially in discrete size intervals corresponding to multiples of monomer or higher-order repeat lengths.</p>", "<title>Conclusion</title>", "<p>This work demonstrated that satellite repeat-derived eccDNA is common in plant genomes and thus it can be seriously considered as a potential intermediate in processes driving satellite repeat evolution. The observed size distribution of circular molecules suggests that they are most likely generated by molecular mechanisms based on homologous recombination requiring long stretches of sequence similarity.</p>" ]

[ "<title>Authors' contributions</title>", "<p>AN and JM designed the study, and AN carried out most of the experimental work. AK constructed and screened genomic DNA libraries and participated in eccDNA detection experiments. JM performed bioinformatic analysis of the newly sequenced satellite repeats. All authors contributed to the manuscript preparation and approved its final version.</p>" ]

[ "<title>Acknowledgements</title>", "<p>We thank Ms. J. Látalová and Ms. H. Štepancíková for their excellent technical assistance. This work was supported by grants GA204/06/P360 from the Czech Science Foundation, AVOZ50510513 from the Academy of Sciences of the Czech Republic, and LC06004 from the Ministry of Education, Youth and Sports of the Czech Republic.</p>" ]

[ "<fig position=\"float\" id=\"F1\"><label>Figure 1</label><caption><p><bold>Detection of eccDNA molecules derived from <italic>Vicia faba </italic>FokI repeats in different tissues and under stress conditions</bold>. (A) Schematic outline of the migration patterns of linear and circular DNA forms on 2-D gel electrophoresis. (B-F) Detection of eccDNA in genomic DNA samples of <italic>V. faba </italic>separated on 2-D electrophoresis (0.4% and 1% agarose), Southern blotted and hybridized with FokI-derived probe. The DNA was isolated from root meristems (B), young (C) or old leaves (D), and from wounded leaves one (E) and two (F) days after the treatment.</p></caption></fig>", "<fig position=\"float\" id=\"F2\"><label>Figure 2</label><caption><p><bold>Improved detection and size determination of VicTR-A repeats in <italic>Vicia narbonensis</italic></bold>. EccDNA detection in <italic>V. narbonensis </italic>genomic DNA sample pre-treated with Plasmid-safe DNase and mung bean nuclease and resolved on 2-D electrophoresis using 0.7% and 2% agarose gels. Prior to electrophoresis, the sample was mixed with a set of circular DNA markers. The Southern blot of the gel was hybridized with VicTR-A probe (A), and then reprobed with the lambda DNA used to visualize DNA markers (B). Panel C shows superposition of signals from VicTR-A nad marker probes and gives the length of the open circle (OC), supercoiled (SC) and linear (L) markers (the lengths are in base pairs). VicTR-A signals on panel C are enhanced by longer exposure compared to panel A in order to reveal spots of the shortest circular molecules.</p></caption></fig>", "<fig position=\"float\" id=\"F3\"><label>Figure 3</label><caption><p><bold>Survey of eccDNA derived from various plant satellites</bold>. Samples of genomic DNA enriched for circular DNA were separated on 0.7% and 2% 2-D gels, blotted and hybridized with satellite-specific probes. The arcs corresponding to linear (L) and open circle (OC) DNA are indicated with black arrowheads and positions of the longest (1,674 bp) and the shortest (558 bp) OC markers are shown as blue spots. The extra arc of unknown origin is indicated with black arrow on panels A, B, D and E. (A-B) Detection of Sc119 repeats in <italic>Secale cereale</italic>, demonstrating the effect of single-stranded DNA degradation by mung bean nuclease (B) compared to untreated control (A). Figure insets contain magnified regions marked by the rectangles showing the effect of the nuclease treatment on shape of the spots. All other samples shown on panels C-L were also treated with mung bean nuclease. (C) Afa repeats in <italic>Triticum aestivum</italic>, (D) IGS-like in <italic>Vicia sativa</italic>, (E) CentO in <italic>Oryza sativa</italic>, (F) FokI in <italic>V. faba</italic>, (G) PisTR-B in <italic>Pisum sativum </italic>(this sample was resolved on 1.5% and 2% 2-D gels), (H) VicTR-B in <italic>V. grandiflora</italic>, (I) VG-V subfamily of VicTR-B in <italic>V. grandiflora</italic>, (J) VicTR-A in <italic>V. pannonica </italic>detected using PCR probe including a mixture of genomic VicTR-A sequences, (K-L) VicTR-A in <italic>V. pannonica </italic>detected using specific clones c653 (K) and c666 (L).</p></caption></fig>", "<fig position=\"float\" id=\"F4\"><label>Figure 4</label><caption><p><bold>Periodicity analysis of VicTR-A repeats in <italic>Vicia narbonensis </italic>and <italic>Vicia pannonica</italic></bold>. The nucleotide autocorrelation analysis of VicTR-A sequences from <italic>V. narbonensis </italic>(A) and <italic>V. pannonica </italic>(B), measuring the excess of pairs of identical nucleotides at various distances (2–500 base pairs). Sequence periodicity was calculated for a distance of <italic>k </italic>base pairs and nucleotide X as a difference C_XX(<italic>k</italic>) = p_XX(<italic>k</italic>)-p_X.p_X, where p_XXis the observed frequency of identical nucleotides X and p_Xis the proportion of nucleotide X in the sequence [##REF##16788823##26##,##REF##10222405##27##]. The periodicity is revealed as regularly spaced peaks. The results combining periodicities of all four nucleotides are shown.</p></caption></fig>", "<fig position=\"float\" id=\"F5\"><label>Figure 5</label><caption><p><bold>Sequence similarity and genomic organization of VicTR-A subfamilies in <italic>Vicia pannonica</italic></bold>. (A) Alignment of the clones c666 and c653, representing the subfamilies with 69/138 and 180 bp periodicities. The A-tracts involved in recombination between the 69/138 repeats which presumably gave rise to the 180 bp subfamily are indicated with orange lines, while other A-tracts are marked with black lines. Only a part of the alignment is shown. (B) Dot-plot comparison of the clones c666 and c653, displaying similarities between the sequences as diagonal lines [##REF##8566757##48##]. (C) Southern blot of <italic>V. pannonica </italic>genomic DNA digested with various enzymes and hybridized with probes derived from VicTR-A clones c666 and c653, respectively. Enzymes used for DNA digestion were: S, <italic>Sau</italic>3A; M, <italic>Mbo</italic>I; T, <italic>Taq</italic>I; B, <italic>BsmF</italic>I; N, <italic>Nla</italic>III; Bp, Np, partial restriction digestion of <italic>BsmF</italic>I and <italic>Nla</italic>III, respectively.</p></caption></fig>" ]

[ "<table-wrap position=\"float\" id=\"T1\"><label>Table 1</label><caption><p>Satellite repeats used in this study</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td align=\"left\"><bold>Satellite repeat family</bold></td><td align=\"left\"><bold>Species</bold></td><td align=\"left\"><bold>Monomer length (bp)</bold></td><td align=\"left\"><bold>Location^(a)</bold></td><td align=\"left\"><bold>Copies/haploid genome (1C)</bold></td><td align=\"left\"><bold>References</bold></td><td align=\"left\"><bold>EccDNA size</bold></td><td align=\"left\"><bold>Figure</bold></td></tr></thead><tbody><tr><td align=\"left\">Afa</td><td align=\"left\"><italic>Triticum aestivum</italic></td><td align=\"left\">340</td><td align=\"left\">P, I, T</td><td align=\"left\">30,000</td><td align=\"left\">[##REF##8355649##49##, ####UREF##4##50##, ##REF##10810139##51####10810139##51##]</td><td align=\"left\">(340)n</td><td align=\"left\">3C</td></tr><tr><td align=\"left\">AT-180</td><td align=\"left\"><italic>Arabidopsis thaliana</italic></td><td align=\"left\">179</td><td align=\"left\">C, P</td><td align=\"left\">&gt; 55,000</td><td align=\"left\">[##REF##7545957##52##,##REF##11853315##53##]</td><td align=\"left\">(179)n</td><td/></tr><tr><td align=\"left\">IGS-like</td><td align=\"left\"><italic>Vicia sativa</italic></td><td align=\"left\">173</td><td align=\"left\">I</td><td align=\"left\">10,000 – 100,000</td><td align=\"left\">[##REF##11289513##47##]</td><td align=\"left\">(173)n</td><td align=\"left\">3D</td></tr><tr><td align=\"left\">CentO</td><td align=\"left\"><italic>Oryza sativa</italic></td><td align=\"left\">155</td><td align=\"left\">C, P</td><td align=\"left\">44,400</td><td align=\"left\">[##REF##9653153##54##,##REF##12172016##55##]</td><td align=\"left\">(155)n</td><td align=\"left\">3E</td></tr><tr><td align=\"left\">Sc119</td><td align=\"left\"><italic>Secale cereale</italic></td><td align=\"left\">118</td><td align=\"left\">I, T</td><td align=\"left\">1,500,000</td><td align=\"left\">[##REF##6244112##56##]</td><td align=\"left\">(118)n</td><td align=\"left\">3A, B</td></tr><tr><td align=\"left\">VicTR-A</td><td align=\"left\"><italic>Vicia narbonensis</italic></td><td align=\"left\">69</td><td align=\"left\">I, T</td><td align=\"left\">100,000 – 1,000,000</td><td align=\"left\">[##REF##10905342##3##,##REF##12770847##57##]</td><td align=\"left\">(138)n</td><td align=\"left\">2</td></tr><tr><td align=\"left\">VicTR-A_c666</td><td align=\"left\"><italic>Vicia pannonica</italic></td><td align=\"left\">69/138</td><td align=\"left\">n.d.^(b)</td><td align=\"left\">n.d.^(b)</td><td align=\"left\">this publication</td><td align=\"left\">continuous arc</td><td align=\"left\">3L</td></tr><tr><td align=\"left\">VicTR-A_c653</td><td align=\"left\"><italic>Vicia pannonica</italic></td><td align=\"left\">180</td><td align=\"left\">n.d.^(b)</td><td align=\"left\">n.d.^(b)</td><td align=\"left\">this publication</td><td align=\"left\">(180)n</td><td align=\"left\">3K</td></tr><tr><td align=\"left\">FokI</td><td align=\"left\"><italic>Vicia faba</italic></td><td align=\"left\">59</td><td align=\"left\">I</td><td align=\"left\">5,400,000 – 21,000,000</td><td align=\"left\">[##REF##8654919##24##]</td><td align=\"left\">continuous arc</td><td align=\"left\">3F</td></tr><tr><td align=\"left\">PisTR-B</td><td align=\"left\"><italic>Pisum sativum</italic></td><td align=\"left\">50</td><td align=\"left\">P, T</td><td align=\"left\">380,000</td><td align=\"left\">[##REF##11550909##58##,##REF##18031571##59##]</td><td align=\"left\">continuous arc</td><td align=\"left\">3G</td></tr><tr><td align=\"left\">VicTR-B</td><td align=\"left\"><italic>Vicia grandiflora</italic></td><td align=\"left\">38</td><td align=\"left\">I, T</td><td align=\"left\">1,000,000 – 5,000,000</td><td align=\"left\">[##REF##10905342##3##]</td><td align=\"left\">continuous arc</td><td align=\"left\">3H</td></tr><tr><td align=\"left\">VicTR-B_VG-V</td><td align=\"left\"><italic>Vicia grandiflora</italic></td><td align=\"left\">186</td><td align=\"left\">T</td><td align=\"left\">n.d.</td><td align=\"left\">[##REF##16788823##26##]</td><td align=\"left\">(186)n</td><td align=\"left\">3I</td></tr><tr><td align=\"left\">VicTR-B</td><td align=\"left\"><italic>Vicia sativa</italic></td><td align=\"left\">38</td><td align=\"left\">I</td><td align=\"left\">1,000,000 – 5,000,000</td><td align=\"left\">[##REF##10905342##3##]</td><td align=\"left\">continuous arc</td><td/></tr></tbody></table></table-wrap>" ]

[]

[]

[]

[]

[]

[]

[ "<table-wrap-foot><p>^(a)Positions of signals on chromosomes. C, centromeric; P, pericentromeric; I, intercalary; T, (sub-)telomeric.</p><p>^(b)VicTR-A family in V. <italic>pannonica </italic>is located in (sub-)telomeric chromosome regions [##REF##10905342##3##], but chromosome localization and copy numbers of its specific sequence variants represented by the clones c666 and c653 were not determined.</p></table-wrap-foot>" ]

[ "<graphic xlink:href=\"1471-2229-8-90-1\"/>", "<graphic xlink:href=\"1471-2229-8-90-2\"/>", "<graphic xlink:href=\"1471-2229-8-90-3\"/>", "<graphic xlink:href=\"1471-2229-8-90-4\"/>", "<graphic xlink:href=\"1471-2229-8-90-5\"/>" ]

[]

{ "acronym": [], "definition": [] }

[ "<title>Background</title>", "<p>Psycholinguistic models of speech production distinguish between four major processing stages to transfer an idea into a meaningful utterance [##UREF##0##1##, ####REF##3749399##2##, ##UREF##1##3##, ##UREF##2##4##, ##UREF##3##5####3##5##], involving conceptual, syntactic, phonological encoding and articulation.</p>", "<p>Evidence for these different levels has been obtained from speech errors and picture naming studies (for a review see [##REF##10354575##6##]). In addition to the question as to what stages can be distinguished during language production, it has also been of interest when these different types of information become available. Electrophysiological measures have been particularly helpful in addressing this question [##REF##9445890##7##, ####REF##15037128##8##, ##UREF##4##9##, ##REF##10934906##10##, ##REF##11388923##11##, ##UREF##5##12####5##12##].</p>", "<p>Existing psycholinguistic theories are mainly based on single word production. Hence, many aspects are discussed related to conceptual representation and selection of single words and as a consequence, little empirical evidence is available for the conceptual planning in complex utterances [##REF##16752086##13##, ####REF##3691025##14##, ##UREF##6##15##, ##REF##1424498##16####1424498##16##]. The present study aims to investigate the neural aspects of conceptualization during speech planning, i.e. the streamlining of ideas into meaningful utterances.</p>", "<p>From a psycholinguistic perspective, conceptualization involves two steps, usually referred to as macro – and micro-planning [##UREF##7##17##]. During macro-planning, an intention or goal is chosen, divided into sub-goals that are planned, ordered and specified for intended mood and content. Its output is an ordered sequence of speech-act intentions (often shortened to speech acts, SA). These speech acts are further specified during the micro-planning phase, where they are assigned particular informational structures (e.g. what should be expressed as topical, focussed, or new information) and perspective. Note that these two processes do not necessarily have to be as separated and serial as is described here. It is very well possible that micro-planning already starts before macro-planning processes are completed.</p>", "<p>The goal of the present study was to gain more insight in macro-planning processing and, to be more specific, in the processes underlying the ordering of events ('the linearization problem').</p>", "<p>The speaker's decision to verbally order events in a particular way is influenced by the speech context in many ways. One such ordering principle is chronological order. Usually, speakers prefer a chronological order of event sequence, for example, 'I woke up this morning and ate breakfast'. But there are situations in which a speaker might choose against a chronological order, for example when the event is the most salient one and is therefore mentioned first ('I got fired when I arrived at work'), a process referred to as 'topicalization' [##UREF##7##17##].</p>", "<p>The speaker can order the events by using temporal connectives available in the language, such as 'before' and 'after'. These are linguistics signals informing the comprehender about the order of the upcoming events. 'After' (on a sentence-initial position) indicates that events will be described in the actual order of occurrence, whereas 'before' signals a reversal (e.g., 'After I ate dinner, I did the dishes' or 'Before I did the dishes, I ate dinner').</p>", "<p>Evidence that a chronological order is preferred to non-chronological order comes from language acquisition and language disorder studies. When asked to act out 'Before/After' instructions, children usually have more difficulty acquiring 'Before' than 'After' [##UREF##8##18##, ####REF##7365706##19##, ##UREF##9##20####9##20##]. For example, Stevenson and Pollitt [##REF##3693459##21##] tested the understanding of temporal terms of English children aged 2 to 5 by letting them act out situations described by sentences containing the words 'Before' and 'After'. Children showed a tendency to act out only the first clause of 'Before' sentences. This suggests that they do not understand the reversed temporal relation between events that is depicted by 'Before' sentences and the authors concluded that the children had greater difficulty understanding 'Before' constructions in comparison to 'After' sentences [##UREF##10##22##].</p>", "<p>Parkinson patients have also been shown to make more errors for sentences starting with 'Before', since they tend to understand 'Before' sentences as if they had started with 'After' [##REF##1878778##23##]. In a related study [##REF##8232851##24##], Parkinson patients also failed to understand so-called object-relative clauses (in which the subject of the main clause serves as the object of the relative clause). These sentences were interpreted as subject-relative clauses, i.e. the subject of the main clause was also assumed to be the subject of the relative clause. Because of fewer filler/gap positions in subject relative clauses, these sentences are less demanding on working memory. Taken together, it appears that the understanding of non-chronological order sentences is more difficult for Parkinson patients because these sentence constructions require more working memory processing.</p>", "<p>The role of working memory in the processing of temporal connectives has been investigated by event-related brain potentials (ERPs) [##REF##9738499##25##]: Participants read sentences that started with the temporal connectives 'Before' and 'After'. The sentences appeared one word at a time whilst EEG was recorded (e.g.; 'Before/After the psychologist submitted the article, the journal changed its policy'). 'Before' sentences differed from 'After' sentences by a ramp-like negativity which started around 300 ms after onset of the sentence's initial word and lasted for the entire sentence. This 'Before/After'-difference was greater for those participants with better individual working memory capacity, indicating an immediate interaction (already at 300 ms after presenting 'Before/After' words) between working memory and linearization of conceptual events. The authors concluded that sentences containing a non-chronological order of events are more demanding on working memory than chronological order sentences, possibly leading to different discourse representations for the two types of sentences.</p>", "<p>In sum, these studies show that non-chronological order constructions (sentences starting with 'Before') seem to be more difficult to understand than chronological order constructions (beginning with 'After'). Furthermore, it seems that the difficulty with 'Before' sentences is due to higher verbal working memory load. The aim of the present paper is to investigate whether this difference in linearization complexity is also reflected in language production. Is it plausible to expect a similar complexity effect in production, or would the mere fact of producing, hence choosing a temporal relation between events extinguish possible differences in difficulty?</p>", "<p>Speech production is achieved with amazing speed, going from the initial planning stage to articulation in just a few hundred milliseconds. If one intends to capture the neural events involved in speaking as they unfold in time, electrophysiological measures are the method of choice. A number of ERP studies have used surrogate tasks (i.e. mapping of specific semantic, syntactic, or phonological features of the word corresponding to a picture to one or two button-press decisions) to study information availability of linguistic information [##REF##9445890##7##,##REF##10934906##10##,##REF##11388923##11##,##REF##11900728##26##, ####REF##11672841##27##, ##REF##9554845##28##, ##REF##10510863##29####10510863##29##]. Some have used delayed vocalization in order to avoid possible speech artefacts. Most recently, overt vocalization tasks in ERP studies [##REF##11068246##30##,##REF##10689046##31##] have shown that reliable and artifact-free ERPs can be generated in the interval between a stimulus onset and the respective vocalization of an overt utterance.</p>", "<p>The present study uses this method to investigate the conceptual planning of chronological and non-chronological order constructions. Moreover, the ERP analysis in the present study focus on a relatively smaller time window (until 600 ms after stimulus onset) to minimize possible interference of speech and other artefacts, but also to reduce the influence of speech preparation. Subjects saw a sequence of two pictures, followed by a cue. They were instructed by the colour of that cue to utter, in German, a sentence describing the typical actions starting with the temporal connectives 'Before' or 'After'. An example for a chronological order would be 'Nachdem ich fahre, lese ich', ['After I drive (a car), I read (a book)']. A non-chronological order would be: 'Bevor ich lese, fahre ich', ['Before I read (a book), I drive (a car)']). The proportion of each utterance format was 50% and it was randomized across blocks.</p>", "<p>As we supposed that the two conditions differ in working memory load, we expected to find ERP differences related to this fact (paralleling the results of Münte et al. [##REF##9738499##25##] in the comprehension domain).</p>" ]

[ "<title>Methods</title>", "<p>All procedures were approved prior to the study by the ethics committee of the University of Magdeburg, which ensured compliance with the Helsinki Declaration.</p>", "<title>Participants</title>", "<p>Thirty-two right-handed, neurologically healthy students aged between 20 and 32 (mean age 23.3, 23 women, 9 men) with normal or corrected to normal vision and German as their native language gave informed consent and were paid for their participation. Subjects with more than 25% loss of trials caused by blinking or movement artefacts were excluded from further analysis. In this study, artefacts were mainly caused by subjects not being able to sit still during talking, or most importantly, by movements prior to vocalization, i.e. in the time-window of interest related to speech planning prior to articulation. It was not uncommon for subjects to move their head or to blink immediately before they started producing an utterance. Some participants could not control these movements and to ensure clean recordings during the conceptualisation window, we excluded 15 subjects, leaving 17 subjects for the final analyses.</p>", "<title>Stimuli and procedure</title>", "<p>A total of 75 pairs of black and white line-drawings (picture data base of the Max-Planck Institute for Psycholinguistics, Nijmegen; [##REF##7373248##47##]) were used. Pictures were edited with Corel Draw version 11.0 to have the same resolution (300 × 300 dpi), size (33 × 33 mm) and colour combination (black on white background). Pictures were combined into pairs such that no semantic and phonological overlap between the words denoting the objects on the two pictures occurred.</p>", "<p>Each picture pair was presented twice in each condition ('Before'/'After') with the position of the pictures switched. This resulted in a total of 150 picture pairs per one condition and 300 picture pairs for the entire experiment.</p>", "<p>Each trial comprised the following sequence: The first object picture was presented for 500 ms, followed by a blank screen for 200 ms. This was replaced by a second object picture, presented also for 500 ms, followed by a coloured fixation cross with a duration of 5000 ms to allow for the overt response. At the end of each trial, a blank screen was shown for 500 ms which prepared the subjects for the next trial. Instructions were to assume the action 1 required for object 1 as having occurred first, while the action 2 associated with object 2 happened subsequent to action 1. Further, subjects were told to start their utterances with 'Before' or 'After'. The color of the fixation cross (red or yellow) specified whether participants generated the event description by means of a (chronological) 'After' or an (non-chronological) 'Before' sentence. For example, subjects saw the object 'book' and then the object 'couch', followed by a red fixation cross. The instructed, correct German utterance would then be 'Nachdem ich lese, sitze ich' [in English; 'After' I read (a book), I sit (down on the couch)']. The same objects followed by a yellow fixation cross, would require the utterance 'Bevor ich sitze, lese ich' [in English; 'Before' I sit (down on the couch), I read (a book)'; information in parentheses for clarification only]. In order to optimally match the two types of utterances, subjects were instructed to use an identical structure for both sentences (except for the initial word). They had to always produce both sentence parts in the present tense. The aim was to minimize variability among the answers, to keep overt production time as short as possible, and to avoid possible differences between the conditions due to lexical effects. Utterances were recorded to check for all of the above mentioned points. The presentation of the fixation cross color was randomized and counterbalanced over the two utterances. Subjects were instructed to utter the required sentence as soon and as correct as possible after the appearance of the cue. After the application of the EEG electrodes, subjects were seated in a sound-proof cubicle and received detailed explanations about their task. They received three practice runs before the actual experiment started. During the first practice run, subjects saw the pictures of the objects with the target verb written below the picture. They were asked to learn the verb-object association (For example, the object picture was 'book', accompanied by the verb 'reading'). The second practice run then showed the same pictures without presentation of the verbs and subjects had to name the pictures out loud. For these two practice sessions, subjects could take as much time as required and they had to perform errorless before continuing with the last practice session. The last session entailed 20 example trials of the experiment itself, so that subjects could familiarize themselves with the timing of the stimuli. Subjects were told to sit as still as possible, and to blink only while they were speaking.</p>", "<title>Recording and analysis</title>", "<p>EEG was recorded with tin electrodes mounted in an electrode cap FP1, FP2, F3, F4, C3, C4, P3, P4, O1, O2, F7, F8, T7, T8, P7, P8, Fpz, Fz, Cz, Pz, Fc1, Fc2, Cp1, Cp2, Po3, Po4, Fc5, Fc6, Cp5, Cp6 positions of the 10/20 system. Two additional electrodes were placed at the left and right mastoid for referencing. The electrode placed on the left mastoid was used for online referencing. Data were re-referenced off-line to the mean of the activity at the two mastoid electrode sites. Vertical eye movements were measured with a bipolar montage comprising electrodes placed above the left eyebrow and below the left orbital ridge. Horizontal eye movements were measured with two electrodes placed at the left and right external canthi. EEG-data were recorded continuously (time-constant 10 seconds, filter settings 0.05 to 30 Hz) with a sampling rate of 250 Hz. Electrode impedances were kept below 5 kΩ. EEG was averaged time-locked to the onset of the colour cue for epochs of 700 ms including a 100 ms pre-stimulus baseline. For eye-blink rejection the maximum difference was set on 150 μV (8 subjects) or 200 μV (9 subjects) for the vertical EOG channel. Shifts were corrected with linear regression. The threshold for shift correction was set to 300 μV/s. Rejections were 25% on average, and there was no difference between the conditions. Only single trials free of blink and movement artefacts were included in the averages. To quantify the ERP effects, mean amplitudes were measured at midline (Fz, Cz, Pz), parasagittal (Fp1/2, F3/4, C3/4, P3/4, O1/2), and temporal (F7/8, T7/8, P7/8) sites in an early (180–230 ms) and a later (350–400 ms) time-interval. These were subjected to repeated measures ANOVA. Factors were Order (After versus Before), Anterior/Posterior (3 levels for midline and temporal, 5 levels for parasagittal) and Hemisphere (left versus right, the factor was not used for midline analyses). The Huynh-Feldt correction for inhomogeneities of covariance was used when appropriate. We report the corrected p-value in conjunction with the original degrees of freedom.</p>" ]

[ "<title>Results</title>", "<title>Behavior</title>", "<p>Voice onset latencies (VOL) were collected by means of a voice key (Presentation, version 9.10). The responses of seventeen subjects were averaged and included in the analysis. The overall amount of errors for each subject was less than 10 percent. There was no significant difference in the error proportions of the two conditions (mean chronological order = 4.9, mean non-chronological order = 5.8, (t(16) = -.74, p = .47) Also, no significant difference in onset latency were observed (mean VOL = 1362 ms for both conditions).</p>", "<title>ERPs</title>", "<p>The grand average ERPs time-locked to the fixation cross are shown in Figure ##FIG##0##1##.</p>", "<p>The two conditions started to diverge around 180 ms, with the 'Before' condition being more positive than the 'After' condition in the entire time window. This difference appeared to have two different portions as is illustrated by the isovoltage maps in Figure ##FIG##1##2##: The earlier portion around 200 ms had a fronto-central distribution, whereas the later portion around 350 ms had a more posterior distribution.</p>", "<p>For the early time-window (180–230 ms), reliable differences between 'Before' and 'After' sentences were revealed by main effects of the factor Order for midline (F(1,16) = 4.83, p &lt; 0.05), parasagittal (F(1,16) = 6.22, p &lt; 0.025) and temporal (F(1,16) = 9.09, p &lt; 0.01) electrode sites. No reliable interactions were obtained between Order and the topographical factors (see Table ##TAB##0##1##).</p>", "<p>For the later time-window (350–400 ms), a significant Order by Anterior/Posterior interaction for the parasagittal (F(4,64) = 7.12, p &lt; 0.003) and temporal sites (F(2,32) = 9.96, p &lt; 0.005) reflected the fact that during this time window a posteriorly distributed difference between 'Before' and 'After' sentences was present (see Figure ##FIG##1##2##). Order did not reach significance (F(1,16) = 1.24/2.02/1.88; all p &gt; .05) for midline/parasagittal/temporal electrode sites, see Table ##TAB##0##1##).</p>", "<p>To test whether indeed the early and late portions of the 'Before/After' differences had different distributions, we determined the mean amplitude of the 'After' minus 'Before' difference waves in the 180–230 ms and 350–400 ms time-windows for all 29 scalp electrodes. These values were subjected to the vector normalization procedure described by McCarthy and Wood [##REF##2581760##32##] and then entered into an ANOVA with time-window (early vs. late) and electrode site (29 levels) as factors. A significant interaction between time-window and site (F(28,448) = 3.06, original p &lt; 0.0001, Huynh-Feldt corrected: p = 0.038) indicated that the difference between condition had indeed a different distribution in the two time-windows.</p>" ]

[ "<title>Discussion</title>", "<p>The present study investigated conceptualization processes while subjects generated a verbal description of two events in a chronological and non-chronological order sequence. The ERP results revealed differences between 'Before/After' sentences in terms of an early fronto-central negativity for 'After' versus 'Before' sentences. This effect is followed by a later parietal positivity for 'Before' in comparison to 'After' sentences. Further, no difference in voice-onset latency (VOL) was found between the two order constructions.</p>", "<p>The observed lack of effects for VOL might seem counterintuitive at a first glance, as one would assume longer latencies for the presumably more difficult planning in the 'Before' case. Note that VOL measures the end of an entire information processing sequence. Strategic effects and execution are included next to the process of interest, i.e. the conceptual planning. However, this result may also be related to a ceiling effect or a strategy to delay naming until the entire utterance plan is available.</p>", "<p>Whereas the VOL are not informative in this case, ERP results show that the brain clearly distinguishes between sentences in which a sequence of events is uttered in chronological order or non-chronological order. An early fronto-central difference (180–230 ms after cue onset) differentiated between 'Before' and 'After' sentences. This result is in line with previous fMRI studies showing that the prefrontal cortex plays a critical role in temporal sequencing [##REF##10617267##33##, ####REF##9872379##34##, ##REF##15589094##35####15589094##35##]. Moreover, patients with prefrontal lesions are known to be impaired in generating and evaluating order of series and actions [##REF##8595032##36##,##REF##7555008##37##]. More specifically, a fronto-temporal network involved in verbal semantic memory decision or categorization processes has been found in several studies before [##REF##12676061##38##,##REF##10769304##39##].</p>", "<p>In parallel to an earlier comprehension study [##REF##9738499##25##] using sentences with temporal connectives which had shown an interaction between working memory and linearization processes reflected in a frontal negativity, it seems plausible to interpret the early negativity in the context of working memory. Indeed, several studies looking at the comprehension of various syntactic sentence structures (SO and SS relative clauses) revealed an anterior negativity for the more complex sentence structures (SO sentences) [##UREF##11##40##,##REF##9088556##41##], which was interpreted in terms of more working memory demands for the understanding of more complex sentence structures.</p>", "<p>The present study reveals a higher negativity for chronological order sentences. This might seem counterintuitive, as it implies that working memory demands are higher during the conceptualization of a chronological order sentence. The direction of this effect can be explained by taking a closer look at the setup of the experiment. The object pictures were presented sequentially during this study, meaning that the first picture needed to be kept in mind longer when subjects had to create a chronological order sentence. This is contrary to the production of the non-chronological order that started with the utterance of the last presented object. Taken together, in order to utter a chronological order sentence, subjects had to go back two steps to start with the first presented object and this process naturally appeared to demand more working memory processing.</p>", "<p>Further, we found a parietal positivity between 350 and 400 ms for the before condition. Its distribution and polarity suggest that this may be an instance of the P300, or P3b. The P300 is a well-known component often found in tasks investigating attention devoted to a stimulus, stimulus salience, task relevance, objective and subjective probability among a stimulus sequence, or the amount of resources needed to process a stimulus [##UREF##5##12##,##REF##7280146##42##, ####REF##3774922##43##, ##REF##887923##44##, ##UREF##12##45####12##45##].</p>", "<p>In this light, the greater P3b for 'Before' constructions in the present study may reflect greater attentional processing in terms of 'context maintenance' processes. A recent study from our laboratory tapping into process-related strategies during conceptualization, found a similar parietal positivity. As described in the introduction, process-related strategies are used when content-related information (e.g. differences in time between events, as used during content-related strategies) is not available. In this experiment, process-related strategies were investigated by manipulating the complexity of utterances describing the direction of an arrow in a network of geometrical forms (easy: downwards, medium: downwards to the triangle, complex: downwards to the grey triangle) [##UREF##13##46##]. In this case, medium and complex utterances were associated with a parietal positivity when compared with the \"easy\" condition.</p>", "<p>To sum up, the present study showed that conceptualization of 'Before' and 'After' order sentences leads to more conceptualization processing when non-chronological order constructions are being built. This finding is in line with language comprehension studies that have shown that these types of sentences are more difficult to understand, and with a second study in our laboratory that investigated complexity differences in process related strategies. Moreover, both the frontal and the parietal effect are in line with a fronto-temporal network found in fMRI studies looking at verbal semantic categorization processes [##REF##12676061##38##,##REF##10769304##39##].</p>", "<p>The result is explained in terms of the P3b component reflecting attentional processing. In addition, this component might not reflect attention related processes per se, but rather a more general demand of resources required for stimulus processing. Seeing that 'before' sentences are the more complex sentences in our study, one could argue by extension that conceptualization in this case takes up more resources. A remaining question, at this point, is what the exact nature of the conceptualization difference is that we found in the present study. Whereas, in daily life, topicalization occurs when one event bears more significance in relation to another event, the current study's use of neutral items made events less salient and therefore, might have affected the conceptualization process of non-chronological orders. While we cannot exclude this line of reasoning entirely at this stage, the lack of difference in voice onset latency and error results for both conditions seem to speak against that possibility.</p>", "<p>Although it is reasonable to suspect that this finding reflects the inverted narration relation between events that are expressed in non-chronological order constructions, further experiments are needed to address this question in more detail.</p>" ]

[ "<title>Conclusion</title>", "<p>In this ERP study addressing conceptualization processes during language production, a frontal negativity likely associated to greater working memory demands for chronological order constructions was found, which can be explained by the fact that the first event needs to be retrieved from working memory. Importantly, as in Marek et al [##UREF##13##46##], a parietal positivity was found for the more difficult (Before) condition, which appears to reflect effects of conceptualization complexity.</p>" ]

[ "<p>This is an Open Access article distributed under the terms of the Creative Commons Attribution License (<ext-link ext-link-type=\"uri\" xlink:href=\"http://creativecommons.org/licenses/by/2.0\"/>), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.</p>", "<title>Background</title>", "<p>During speech production the planning of a description of several events requires, among other things, a verbal sequencing of these events. During this process, referred to as linearization during conceptualization, the speaker can choose between different types of temporal connectives such as 'Before' X did A, Y did B' or 'After' Y did B, X did A'. To capture the neural events of such linearization processes, event-related potentials (ERP) were measured in native speakers of German. Utterances were elicited by presenting a sequence of two pictures on a video screen. Each picture consists of an object that is associated with a particular action (e.g. book = reading). A coloured vocalization cue indicated to describe the sequence of two actions associated with the objects in chronological (e.g. red cue: 'After' I drove the car, I read a book) or reversed order (yellow cue).</p>", "<title>Results</title>", "<p>Brain potentials showed reliable differences between the two conditions from 180 ms after the onset of the vocalization prompt, with ERPs from the 'After' condition being more negative. This 'Before/After' difference showed a fronto-central distribution between 180 and 230 ms. From 300 ms onwards, a parietal distribution was observed. The latter effect is interpreted as an instance of the P300 response, which is known to be modulated by task difficulty.</p>", "<title>Conclusion</title>", "<p>ERPs preceding overt sentence production are sensitive to conceptual linearization. The observed early, more fronto-centrally distributed variation could be interpreted as involvement of working memory needed to order the events according to the instruction. The later parietal distributed variation relates to the complexity in linearization, with the non-chronological order being more demanding during the updating of the concepts in working memory.</p>" ]

[ "<title>Abbreviations</title>", "<p>ERP: event-related potential.</p>", "<title>Authors' contributions</title>", "<p>BH co-designed the study, analysed and performed the experiments and the statistical analyses and co-wrote the manuscript. BMJ and TFM co-designed the study and co-wrote the manuscript. All authors read and approved the final manuscript.</p>" ]

[ "<title>Acknowledgements</title>", "<p>Supported by DFG grant MU1311/13-1 to TFM and NWO ASPASIA grant 015.001.053 to BMJ. Address correspondence and requests for reprints to: Thomas F. Münte, Department of Neuropsychology, P.O. Box 4120, 39016 Magdeburg, Germany, email: <email>[email protected]</email>.</p>" ]

[ "<fig position=\"float\" id=\"F1\"><label>Figure 1</label><caption><p><bold>Grand average ERPs at selected scalp sites time locked to the onset of the coloured fixation cross which prompted the utterance.</bold> The Before condition gave rise to a more positive waveform starting at about 180 ms.</p></caption></fig>", "<fig position=\"float\" id=\"F2\"><label>Figure 2</label><caption><p><bold>Spline interpolated isovoltage maps of the difference between the After and Before condition.</bold> During the first phase a fronto-central distribution is evident, while beyond 350 ms a clear parieto-occipital maximum emerges (min/max scaling: -0.24 to 0.24 μV at 200 ms; -0.98 to 0.98 μV at 360 ms).</p></caption></fig>" ]

[ "<table-wrap position=\"float\" id=\"T1\"><label>Table 1</label><caption><p>F-values from ANOVAs comparing the different conditions at temporal, parasagittal and midline electrode locations.</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td/><td align=\"left\"><bold><italic>Df</italic></bold></td><td align=\"left\"><bold>F</bold></td><td align=\"left\"><bold>180–230 ms</bold></td><td align=\"left\"><bold>350–400 ms</bold></td></tr></thead><tbody><tr><td align=\"left\"><italic>Temporal</italic></td><td/><td/><td/><td/></tr><tr><td align=\"left\">Type of Order (O)</td><td align=\"left\">1,16</td><td align=\"left\">9.09</td><td align=\"left\">.01</td><td/></tr><tr><td align=\"left\">O × Ant</td><td align=\"left\">2,32</td><td align=\"left\">9.96</td><td/><td align=\"left\">.005</td></tr><tr><td colspan=\"5\"><hr/></td></tr><tr><td align=\"left\"><italic>Parasagittal</italic></td><td/><td/><td/><td/></tr><tr><td align=\"left\">Type of Order (O)</td><td align=\"left\">1,16</td><td align=\"left\">6.22</td><td align=\"left\">.025</td><td/></tr><tr><td align=\"left\">O × Ant</td><td align=\"left\">4.64</td><td align=\"left\">7.12</td><td/><td align=\"left\">.003</td></tr><tr><td colspan=\"5\"><hr/></td></tr><tr><td align=\"left\"><italic>Midline</italic></td><td/><td/><td/><td/></tr><tr><td align=\"left\">Type of Order (O)</td><td align=\"left\">1,16</td><td align=\"left\">4.83</td><td align=\"left\">.05</td><td/></tr></tbody></table></table-wrap>" ]

[]

[]

[]

[]

[]

[]

[]

[ "<graphic xlink:href=\"1471-2202-9-77-1\"/>", "<graphic xlink:href=\"1471-2202-9-77-2\"/>" ]

[]

{ "acronym": [], "definition": [] }

[ "<title>Background</title>", "<p>Today about 33 million people are living with HIV. Of them 15 million are women and 2.5 million children under the age of 15 [##UREF##0##1##].</p>", "<p>Vertical transmission of HIV from mother-to-child accounts for the vast majority of the infections among the children. Mother-to-child HIV transmission occurs intrauterine, intra-partum and during breastfeeding. Without antiretroviral treatment, the risk of an infected woman transmitting the virus to her child is between 16 and 40%. Breastfeeding contributes at least a 10% risk of transmission [##REF##10703780##2##]. Timely administration of antiretroviral drugs to the HIV infected pregnant woman and her newborn significantly reduces the risk of mother-to-child transmission [##UREF##1##3##]. With antiretroviral treatment resulting in low viral load, no breastfeeding and elective Caesarean the HIV transmission rate to the child can be reduced to 0–2% [##REF##17253490##4##]. Elective Caesareans are most important for women with high HIV viral loads at the time of delivery.</p>", "<p>Pregnancy has not been proven to have any negative effects on women with asymptomatic HIV infection [##REF##7501333##5##]. It seems that advanced HIV infection, however, can increase the risk for spontaneous abortion or premature birth [##REF##10703780##2##].</p>", "<p>In Kazakhstan, 16 500 people are estimated to live with HIV/AIDS which means a prevalence of 0.1–0.2% [##UREF##2##6##]. The HIV epidemic is strongly concentrated to vulnerable populations, like injecting drug users, sex workers and prisoners. In addition to these groups, the epidemic is spreading among other vulnerable groups like youth, migrants and truck drivers. About three-quarters of the new HIV diagnoses is among unemployed people. Currently the two main routes of HIV transmission in Kazakhstan are injecting drug use and sexual transmission.</p>", "<p>Because of the increased involvement of women and children in drug use and trafficking, it's likely that the HIV prevalence in these groups will rise. There are about 20 000–50 000 female sex workers, of whom 30% are intravenous drug users [##UREF##3##7##]. According to WHO/UNAIDS report in March 2006, there were 1500 people in need of antiretroviral drugs in Kazakhstan at the end of 2005. Just 15%, of those in need, received treatment [##UREF##4##8##].</p>", "<p>In recent years efforts has been made to cover all pregnant women with full prevention services and the mother-to-child prevention is a big part of the Kazakhstan HIV programme [##UREF##2##6##]. In average the number of births in Kazakhstan are 237 000 per year. In 2005 the antenatal care coverage were estimated to 91% and the number of women counselled on prevention of mother-to-child transmission (PMTCT) services were 129 706. The number of HIV infected women was estimated to fewer than 500 and of those 47 (9%) received antiretroviral therapy for PMTCT [##UREF##5##9##].</p>", "<p>The aim of the study was to evaluate current knowledge, risk behaviour and attitudes to voluntary counselling and testing concerning HIV/AIDS among pregnant women in Semey, Kazakhstan. This is important, considering that the gateway for prevention of mother-to-child transmission is voluntary counselling and testing for HIV.</p>" ]

[ "<title>Methods</title>", "<p>The study was conducted between June 14 and July 25, 2007 in Semey, Kazakhstan. We collected 226 questionnaires from a consecutive sample of pregnant women attending four different antenatal clinics at different parts of the city. During a year there are around 1000 pregnancies in Semey followed at 23 antenatal clinics. At the included clinics 520 pregnant women were registered at the time of the study.</p>", "<p>A questionnaire was designed to obtain three areas of interest: sociodemographic characteristics, general awareness of HIV/AIDS and attitudes and risk behaviour. The questionnaire starts out with the sociodemographic questions, like age, educational and employment level for the woman and her husband, parity, socio-economic status, religion and ethnical group. In the next part, general awareness of HIV/AIDS follows questions about if the woman has heard about HIV/AIDS before, and in case of, how she first heard about it, knowledge of route of transmission, knowledge of HIV prevention and symptoms. The third part includes questions concerning if they wanted to be tested and if they wanted more children or would be prepared not to breastfeed their baby, to take pills and deliver through Caesarean to prevent HIV transmission to their baby, given they were HIV positive. In addition, the third part includes questions about whom they would inform if they were tested positive, from whom they would get support and finally questions concerning use of condoms, sexual contacts and intravenous drug use.</p>", "<p>The questionnaire contained three questions about condom use. The first and second questions about use of condom in steady relationship respectively with casual partners were designed to assess the women's actual practice of condom. The third question concerning responsibility for condom use is a question about the women's attitude.</p>", "<p>The questionnaires were translated to Kazakh and Russian and the women were given questionnaires in the language they preferred.</p>", "<p>The women were informed that their participation and the completion of the questionnaire were entirely voluntarily and that they were free not to answer the questions they found too private. The information given would be stored confidentially and no names or identifying information would appear in publications. A verbal consent was obtained for each participant. In addition to the questionnaires, we interviewed 21 of the study participants. In all interviews an interpreter was necessary; because of this a local student was always present to help us to ask questions and to take notes.</p>", "<p>Open-ended questions were used and the women were encouraged to give as much information as they could. We were careful not to suggest answers or to ask leading questions. The questions were clustered around the areas: knowledge of differences between HIV and AIDS, living with HIV/AIDS, attitudes to people living with HIV/AIDS and in particular knowledge about mother-to-child transmission and attitudes to voluntary counselling and testing.</p>", "<title>Ethical considerations</title>", "<p>The participation in the study was voluntary with informed consent. The questionnaire was anonymous with no registration of names, medical data or other personal information. In our manuscript it is not possible to identify individual patients. We have a written decision from the Head of the Ethical Committee at Semey, Kazakhstan that no further ethical review is required. With the anonymous design and no registration of personal sensitive data, the study does not need to be reviewed by the Swedish Ethical Committee according to their rules. No information was collected from the women who refused to participate.</p>" ]

[ "<title>Results</title>", "<title>Study population</title>", "<p>A total of 226 pregnant women participated in our study. Their ages ranged from 18 to 47 years, with a mean age of 26.8 and a median age of 25. Thirteen women did not answer the question about their age; therefore these women were excluded when we referred our results to age. Parity ranged from 0 to 4 children, with a mean of 0.68 and a median of 1 child. Eighteen women did not answer the question about parity.</p>", "<p>Of the 226 women, 76.2% (170/223) were Kazakh, 18.4% (41/223) Russian, 2.2% (5/223) Tatar, 1.8% (4/223) German and 1.3% (3/223) others. The three women who did not answer the question about their ethnicity were excluded when we referred our results to ethnicity. Religion is closely linked with ethnicity. Of the Kazakh women who answered the question about religion, 100% were Muslim. Among the Russian women who answered the question about religion, 93% were Christian Orthodox. Of the whole group 79.7% (173/217) were Muslim, 18.9% (41/217) Christian Orthodox and 1.4% (3/217) other. No significant association was found between occupation and ethnic origin. When it comes to education among the women we refer to additional file ##SUPPL##0##1##.</p>", "<p>Russian women significantly more often attended Special College (66%; 95% CI 51–80%) compared to 35% (95% CI 28–41%) for Kazakh women. Kazakh women had a tendency to more often attend University/Institute/Academy.</p>", "<title>General knowledge</title>", "<p>Ninety-six percent (215/225) of the women had heard of HIV/AIDS.</p>", "<p>The majority, 52.1% (111/213), of the whole group had first heard about it from the media, 40% (85/213) knew it from school, 2.8% (6/213) from their parents, 1.9% (4/213) from friends and 3.3% (7/213) from other sources. Two women did not answer this question. The older women (&gt; 30 years old) had significantly more often, 80% (95% CI 68–91), heard about HIV/AIDS from the media than the younger women at age 20–30 on 47% (95% CI 39–55). In contrast the younger women (age &lt; 20 and 20–30), had significantly more often heard about the disease at school, 64% (95% CI 35–92) for women age &lt; 20 and 46% (95% CI 37–54) for women age 20–30, as compared with women with age over 30, for whom the corresponding figure was 16% (95% CI 6–27). Only one out of the 56 women who correctly answered yes to the statement that there are differences between HIV and AIDS was able to specify those differences, but the 21 women we interviewed with open questions were better at specifying the differences between HIV and AIDS. Among those who specified, many answers were unclear and difficult to interpret. However, we did obtain some more or less correct answers like: \"HIV is the causative agent of the disease AIDS\", \"HIV means you are a carrier, while AIDS means you are sick\" and \"with HIV you live longer than with AIDS\".</p>", "<title>Knowledge about transmission and symptoms</title>", "<p>To the open question regarding the main way HIV/AIDS is spread from one person to another many women gave several answers, the most common being \"sexual contact\" given by 76% (122/160). The second most common answer was blood transfusion, with 30% (48/160). Nineteen percent (30/160) wrote intravenous drug use/needle sharing and 3% (5/160) wrote they didn't know.</p>", "<p>To discover misconceptions about transmission about HIV/AIDS the women were asked to include/exclude ways that HIV/AIDS can or not can be transmitted. These answers, with reference to the women's level of education [see Additional file ##SUPPL##1##2##]</p>", "<p>In general women, with higher levels of education were better than women with low education at correctly excluding and including transmission routes of HIV/AIDS. The differences were significant for \"shaking hands/hugging/living in the same house\", \"changing clothes with someone who has HIV/AIDS\", \"sexual intercourses with condom\", and \"sharing needles while injecting drugs\". For details, see Additional file ##SUPPL##1##2##.</p>", "<p>As many as 76% (157/206) answered correctly, with no, to the statement you can not tell, by looking at a person, whether he/she is infected with HIV/AIDS. Younger women (age &lt; 20) significantly more often, 100%, answered no to the statement that by looking at a person you can see if he/she is infected with HIV/AIDS, as compared with women aged 20–30 years, 76% (95% CI 69–83) and women aged over 30 years, 70% (95% CI 57–84).</p>", "<p>The open question \"Do you know any symptoms of HIV/AIDS\" was left blank by 111 women. Of the responders, 66% (69/104) answered no. Among the women who could mention any symptoms the most common answers were fever/sub febrile 37% (13/35), loss of immunity 34% (12/35), weakness 29% (10/35) and cahexia/weight loss 17% (6/35). When the women's ability to mention symptoms of HIV/AIDS were refered to educational level, the women with low education levels significantly more often, 95% (95% CI 85–100), answered that did not know any symptoms, as compared with women with higher levels of education, 60% (95% CI 50–70).</p>", "<title>Knowledge about treatment</title>", "<p>A total of 14% (28/201) believed that there is a medicine to cure HIV/AIDS, and nearly the same number, 15% (30/202) stated that there are other ways to cure HIV/AIDS. Many more, almost one out of three, stated that there are possible ways to slow the progression of the disease.</p>", "<title>Having HIV/AIDS</title>", "<p>Of the whole group 41% (88/215) of the women wanted to be tested for HIV/AIDS. Fifty-nine percent (124/211) of the women said they would tell the healthcare personnel if they were found to be HIV positive, 35% (74/211) that they would tell their mothers and 32% (68/211) their partners. Fifteen women left the question blank. Statistically significantly more Russian women, 51% (95% CI 36–66) said they would tell their partner as compared with the Kazakh women, 26% (95% CI 19–33), but on the other hand there was a tendency among the Kazakh women to say more often that they would tell their friends than among the Russian women. There was also a tendency for the Russian women to say more often that they would tell their mothers, but the difference was within the margins of error.</p>", "<p>Regarding from whom the women thought they would get support if they were found to be HIV positive, the far most common answer was mother, with 51%. Only 45% thought that they would get support from healthcare personnel, and even fewer, 21%, said they would get support from their fathers, partners 33% or friends 23%. Sixteen women (7%) did not answer the question. A comparison between Kazakh and Russian women showed that more Russian women expected to get support from mothers and partners, while Kazakh women more often mentioned friends and healthcare personnel. Note that the women could tick several alternatives.</p>", "<title>Attitudes to PMTCT, prevention of mother-to-child transmission</title>", "<p>Seven women said they would want to have more children even if they were found to be HIV positive. Among the women who said no to have more children 82% (176/216), the main reason given was the risk of infecting the baby. Significantly more highly educated women said no, 87% (95% CI 81–92), to having more children as compared with the less educated women, 66% (95% CI 53–79). Of the responders 83%, were prepared not to breastfeed their baby if they were found to be HIV positive. Somewhat more, 86%, were prepared to take medicine, but fewer women, 68%, were prepared to accept Caesarean section to prevent mother-to-child transmission.</p>", "<title>Risk behaviour and protection</title>", "<p>Ninety-six percent (179/186) had had one sexual partner during the last sixth months. The three questions about condom use were as follow: \"Use of condom in steady relationship?\", \"Use of condom with casual partners?\" and \"Who has the responsibility for condom use?\" Of the women who answered the first question 21% (42/201) answered always, 40% (80/201) never and 39% (79/201) sometimes. Of the responders to the second question, as few as 57% (78/137) answered always, 34% (47/137) never and 9% (12/137) answered sometimes.</p>", "<p>About two thirds, 68% (141/206), of the women stated that men and women have equal responsibility to make sure a condom is used during sexual intercourses. Of our responders 65% (134/206) answered that there are ways to protect oneself against sexually transmitted HIV/AIDS. Of the women who answered yes to this question, 55% (74/134) specified ways to protect oneself. Note that the women were allowed to give more than one way of protection.</p>", "<p>The far most common answer was to use condoms, 82% (61/74). Other answers were \"avoiding casual sexual contacts\" and \"use of clean syringes\". When the women were asked if they had sufficient information to protect themselves against HIV/AIDS, 42% (89/211) answered no and nearly the same number, 42% (88/211) answered yes. Russian women answered to a significantly higher degree 63% (95% CI 49–78) that they had sufficient information compared to Kazakh women 36% (95% CI 28–43).</p>", "<title>HIV epidemic</title>", "<p>Of our responders 47% (93/198) answered that certain groups of people are more often infected with HIV than others. The far most common group was drug addicts/intravenous drug users, mentioned by 88% (49/56) of the responders. Other groups, given by less than 10% each, were people with many sexual partners, prostitutes, homeless people, homosexual people, medical people, people with low social status, and people without knowledge about the disease and blood donors/recipients. One misconception was that 11% (6/56) of the responders answered that people with poor immune defence systems/organisms are more often infected.</p>" ]

[ "<title>Discussion</title>", "<p>The HIV epidemic in Central Asia is continuing to spread and has reached Kazakhstan. The number of reported HIV cases is still relatively low and mainly concentrated to vulnerable populations such as intravenous drug users and sexual workers. To prevent future spread of the HIV epidemic and to prevent stigmatization in Kazakhstan and other countries in Central Asia it is important to evaluate HIV knowledge and to educate the population with correct information.</p>", "<title>Knowledge</title>", "<p>It is reassuring to see that only ten women out of 226 did not know there was a disease called HIV or AIDS. This corresponds to the findings in several other studies concerning awareness about HIV/AIDS among pregnant women from India [##REF##17134354##10##], China [##UREF##6##11##], Papua New Guinea [##REF##16454397##12##] and Ghana [##REF##17299543##13##]. Compared with the women in Aksu, northwest China, the women in Semey had more often heard of about HIV/AIDS. 95% (95% CI 92–98) compared to 85% (95% CI 80–89) in Aksu. It is positive that the media are a major source of information, exactly as in the Aksu study and the Papua New Guinea study [##REF##16454397##12##]. That a majority of the women are aware of a disease called HIV/AIDS and that media are a major source of information shows that the HIV epidemic is discussed in public and that the media are an efficient way of spreading information. In addition illiteracy is rare in Kazakhstan, which facilitate for the women to take share of the given information.</p>", "<p>However, it is also important that the information is correct, otherwise misconceptions may lead to further stigmatization. Because the women had difficulties distinguish HIV from AIDS and only 16% could mention symptoms of HIV/AIDS, we conclude that the women's knowledge in general was superficial with little understanding of the details and the nature of the disease. We were pleased to find that as many as 76% answered that you can not tell by looking at a person if she/he is infected with HIV/AIDS. The younger women, as compared with the older women, significantly more often answered no to this question. This may be explained by the fact that nowadays students got a lot more information at school.</p>", "<p>The women in Semey were, as compared with the women in Aksu, significantly better at pointing out the high risk behaviour \"sexual intercourse without condom\", 89% (95% CI 85–93), as compared with 73% (95% CI 67–78) in Aksu and \"sharing needles while injecting drugs\", 86% (95% CI 81–90) compared to 55% (95% CI 49–61) in Aksu as routes of transmission. This is positive because these two routes are the most important routes for the general population to be aware of.</p>", "<p>The finding that pregnant women in general are aware of the two main routes: sexual intercourse without a condom and sharing needles while injecting drugs, of HIV transmission is in agreement with previous studies [##UREF##6##11##, ####REF##16454397##12##, ##REF##17299543##13####17299543##13##], and might not be so surprising since media focus on these two main routes in their information to the public. There were more difficulties in excluding incorrect routes of transmission. Almost one fifth though that kissing could be a route of HIV transmission and only 44% answered correctly that HIV can not be spread by mosquitoes. This is slightly lower than in the Hong Kong study where 57% answered to mosquitoes as vectors of HIV [##UREF##7##14##]. The misconception about mosquitoes as a transmission route has also been seen in other similar studies [##REF##17134354##10##, ####UREF##6##11##, ##REF##16454397##12####16454397##12##] and might not be very surprising, since HIV is a blood-borne disease. The pregnant women in our study, like the women in a similar study conducted in the province Yunnan, China [##UREF##6##11##] excluded, to a higher, but not satisfactory extent daily domestic contacts such as eating from the same plates and cups, shaking hands, hugging, living in the same house and changing clothes with someone who has HIV/AIDS as possible routes of transmission. The fact that a high proportion of the women responded \"don't know\" in addition to the women with incorrect answers further illustrates their limited knowledge.</p>", "<p>Our conclusion is, as in the study conducted in Aksu, that most of the women know that HIV is a sexually transmitted disease, but there are still people who have poor knowledge of HIV/AIDS and people with misconceptions about how it is spread. These misconceptions may influence the dissemination patterns of HIV and increase the stigmatization of the HIV positive, as well as giving rise to misguided fears. It is essential to continue to raise the level of knowledge of HIV/AIDS through campaigns in the media and at schools. One thing that needs to be communicated is that there is no medicine to cure a HIV infection, only 40% (95% CI 33–47) of the women answered correctly that this was the case. This may reflect that the information about HIV/AIDS in the media focus on risk behaviour and not on possible treatment. In addition the prevalence of HIV is low in Kazakhstan and few of the women participating in our study had met or knew anyone with the infection i.e. they have never come across persons under treatment for HIV/AIDS. This is significantly lower than the corresponding figure for the women in the Hong Kong study, where 79% (95% CI 73–84) answered no to the statement that there are medicines available to cure HIV/AIDS. The pregnant women in Ghana also significantly more often, 90% (95% CI 86–93), answered no to the statement that there are medicines/treatment available to cure HIV/AIDS. That significantly more women in Ghana know that there are no available treatment to cure HIV/AIDS can be explained by that the prevalence is higher (3.1% in 2003, according to CIA homepage) in Ghana and therefore more women have come across friends or relatives under treatment.</p>", "<title>Having HIV/AIDS</title>", "<p>There were fewer women in our study, who would take a HIV test if it was provided, 41% (CI 95% 34–47) as compared with 77% (95% CI 71–83) in the Hong Kong study. Greater willingness among pregnant women to take an HIV test is also reported from several other studies [##REF##17134354##10##, ####UREF##6##11##, ##REF##16454397##12##, ##REF##17299543##13####17299543##13##,##REF##17461720##15##,##REF##17385672##16##]. The low prevalence of pregnant women willing to take an HIV test in Kazakhstan is not positive and can be explained of that Kazakhstan has a high coverage, 91% in 2005 of antenatal care [##UREF##5##9##]. Therefore, many of the women participating in our study already had been tested for HIV/AIDS in the beginning of their current pregnancy. Considering this the high rate of women answering no to have an HIV test may not indicate unwillingness, instead it may indicate the fact that they already have been tested. When it comes to informing those close to themselves a majority of the women said they would inform the healthcare personnel if they were found to be HIV positive. This indicates that many of the women have put their confidence in the health care system, which increases the system's chance to stop further spread of HIV and to give the infected women treatment. Only one out of three women felt confident enough to inform and get support from their husband, which is a problem, since informing one's partner is an important issue in the prevention of further spread. The study conducted in Aksu showed similar results.</p>", "<title>Knowledge and attitudes toward PMTCT, prevention of mother-to-child transmission</title>", "<p>The women in Semey had limited knowledge about mother-to-child HIV transmission, considering that prevention of mother-to-child transmission of HIV is a large part of the Kazakhstan HIV program. Both the questionnaires and the interviews confirmed this lack of awareness. Although 68% (95% CI 62–75) knew that HIV could be transmitted to the foetus by an infected mother during pregnancy and delivery, only 46% knew that breastfeeding can be a route of transmission. An explanation can be that transmission from mother to child not is a main route in Kazakhstan and therefore the women have not been reached on prevention strategies and information about this transmission way. A comparison with the women in the Hong Kong study shows that the Hong Kong women were significantly better at pointing out pregnancy 97% (95% CI 94–99) and delivery 91% (95% CI 86–94) as possible routes of mother-to-child transmission, but when it came to breastfeeding around the same percentage of women in both our study and the Hong Kong study identified this as a route of HIV transmission. The uncertainty about mother-to-child transmission as a route of HIV transmission has also been identified in other studies [##UREF##6##11##,##REF##17299543##13##,##REF##17461720##15##,##REF##17385672##16##]. In contrast to the studies mentioned above, two similar studies conducted in India [##REF##17134354##10##] and Papua New Guinea [##REF##16454397##12##] reports a higher percentage, 80% and 69% respectively, of women who knew about breastfeeding as a route of transmission of HIV from mother to child. When it comes to knowledge about pregnancy and delivery as routes of transmission, around the same proportion of pregnant women in India and Papua New Guinea, as in the Hong Kong study [##UREF##7##14##], knew about these routes.</p>", "<p>Even if the women in Semey seem to have poor knowledge about routes of mother-to-child HIV transmission, it is positive to see that only seven out of 226 women said they would want to give birth to more children if they were found to be HIV positive. However, in comparison to the women in the Aksu study this result is not satisfactory. The women in the Aksu study stated significantly more often, 97% (95% CI 95–99), that they would not have more children, as compared with 82% (95% CI 77–87) of the Semey women. This may be a result of that a family is central for the women in Kazakhstan and the women may have difficulties to accept not to have any more children, in contrast the women in China are not allowed to give birth to more than one child according to their law. It is also positive that the vast majority of the women were prepared not to breastfeed (the same was seen in the Aksu study) and to take medication to prevent mother-to-child transmission. It is therefore of great importance that HIV positive mothers get help both financially and with information so that they actually can take the necessary medication and not breastfeed.</p>", "<p>The majority of the women were prepared to accept Caesarean section if being HIV positive and being recommended this operation to reduced mother to child transmission. One out of four women answered \"don't know\" to this question. A possible explanation for why just 68% of the women answered they were prepared to have Caesarean section could be that the women don't know what the term \"Caesarean section\" means. Another explanation may be that some women consider it too expensive.</p>", "<title>HIV epidemic</title>", "<p>The women in Semey are not a group with generally high risk behaviour. Many women are aware that intravenous drug use is a problem in Kazakhstan, but few of our responders had been in contact with drug use personally. Slightly less than half our responders wrote that certain groups are more often infected with HIV/AIDS than others. Of those who specified their answers, 11% answered that people with poor immune defence systems/organisms were more often infected, a misconception probably attributable to confusion with the information that HIV causes poor immune defence system. The high level of condom use in steady relations may be reflecting that the majority of the participating women were well educated and living in a city. An average family in the cities has lower number of children than an average family in rural villages.</p>" ]

[ "<title>Conclusion</title>", "<p>Almost all the women in our study had heard of a disease called HIV/AIDS. The women's knowledge, however, was somewhat superficial and many could not specify their answers. Still, most of the women managed to identify sexual contacts and intravenous drug/needle sharing as main routs of transmission. The knowledge among pregnant women in Semey was similar to the knowledge among pregnant women in Aksu, northwest China.</p>", "<p>The media were a main source of information, showing that the HIV epidemic is discussed in public and that the media are an efficient way of disseminating information. However, it is important that the information is correct, including that social contacts are not a risk, otherwise misconceptions may lead to further stigmatization.</p>", "<p>One conclusion is that pregnant women in Semey have poorer knowledge than the pregnant women in Hong Kong about specific mother-to-child HIV transmission and do not know about the means of reducing mother-to-child HIV infection. However, most of the women in Semey were positive to prevention strategies for mother-to-child transmission. It is therefore important that testing and counselling for both men and women are available free of charge. Individuals who test positive need to be supported to prevent further risk behaviour, and HIV positive pregnant women should be offered antiretroviral treatment and help to shorten the breastfeeding period, to prevent mother-to-child transmission.</p>", "<p>It is noteworthy that only a minority of the women specified condom use as a means of protection. There is room for improvement in this respect, and the importance of the use of condoms cannot be underestimated, since condom use is one of the most important means of preventing extension of the HIV epidemic.</p>" ]

[ "<p>This is an Open Access article distributed under the terms of the Creative Commons Attribution License (<ext-link ext-link-type=\"uri\" xlink:href=\"http://creativecommons.org/licenses/by/2.0\"/>), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.</p>", "<title>Background</title>", "<p>Central Asia has one of the most rapidly increasing HIV prevalence in the world. The aim of this study was to evaluate current knowledge, risk behaviour and attitudes to voluntary counselling and testing concerning HIV/AIDS among pregnant women in Semey, Kazakhstan.</p>", "<title>Methods</title>", "<p>We collected 226 questionnaires in a consecutive sample from a population on 520 pregnant women. The results were related to ethnicity, age and education level.</p>", "<title>Results</title>", "<p>Ninety-six percent had heard about HIV.</p>", "<p>Positive findings were that 89% and 86% of the women were aware of the two main routes of transmission: sexual intercourses without a condom and sharing needles while injecting drugs. The women had first heard about HIV/AIDS through the media with, 52%, and at school with 40%. Only 46% and 68% of the women pointed out breastfeeding and mother-to-child transmission during pregnancy or delivery as routes of transmission. Eighty-three percent were prepared not to breastfeed their baby if they were found to be HIV positive. Slightly more, 86%, accepted the need to take medicine, but fewer women, 68%, were positive to Caesarean section. Negative findings were that only 28% answered that there are ways to protect oneself against sexually transmitted HIV/AIDS and specified that this was condom use.</p>", "<title>Conclusion</title>", "<p>The pregnant women in Semey have poor knowledge about specific mother-to-child HIV transmission and do not know about the means of reducing mother-to-child HIV infection. The information in the public health program needs to be improved. However, most of the women in Semey were positive to prevention strategies for mother-to-child transmission after hearing about it.</p>" ]

[ "<title>Competing interests</title>", "<p>The authors declare that they have no competing interests.</p>", "<title>Authors' contributions</title>", "<p>All the authors have together participated in the design of the study. ES and SS have made the interview, distributed the questionnaires and registered the data with the supervision by MU. ES, SS have made the analysis of the data together with RA and the statistician Salmir Nasić. RA, MU, ES, SS have participated in interpretation of data and preparation of the manuscript. All the authors have read and approved the final manuscript.</p>", "<title>Pre-publication history</title>", "<p>The pre-publication history for this paper can be accessed here:</p>", "<p><ext-link ext-link-type=\"uri\" xlink:href=\"http://www.biomedcentral.com/1471-2458/8/295/prepub\"/></p>", "<title>Supplementary Material</title>" ]

[ "<title>Acknowledgements</title>", "<p>We thank all the women who took part in the study.</p>", "<p>We also thank statistician Salmir Nasić for help with the calculations and Secretary Lisbeth Jinnestål Fernow for help with the manuscript.</p>", "<p>The study has been supported by a grant from the Sahlgrenska Academy at Gothenburg University</p>" ]

[]

[]

[]

[]

[]

[]

[]

[ "<supplementary-material content-type=\"local-data\" id=\"S1\"><caption><title>Additional file 1</title><p>Education among the pregnant women in Semey related to ethnic group. The data provided present the level of education among the ethnic groups with 95% confidence intervals.</p></caption></supplementary-material>", "<supplementary-material content-type=\"local-data\" id=\"S2\"><caption><title>Additional file 2</title><p>Knowledge of HIV transmission routes according to educational level and total. The data provided present the level of HIV knowledge related to level of education and ethnic group. The results are presented in a table with 95% confidence intervals.</p></caption></supplementary-material>" ]

[]

[]

[ "<media xlink:href=\"1471-2458-8-295-S1.pdf\" mimetype=\"application\" mime-subtype=\"pdf\"><caption><p>Click here for file</p></caption></media>", "<media xlink:href=\"1471-2458-8-295-S2.pdf\" mimetype=\"application\" mime-subtype=\"pdf\"><caption><p>Click here for file</p></caption></media>" ]

{ "acronym": [], "definition": [] }

[ "<title>Background</title>", "<p>It has long been acknowledged that adult risk factors for cardiovascular disease (CVD), such as obesity and smoking, do not adequately explain the incidence of CVD in later life [##REF##16011473##1##,##REF##9021310##2##]. Indeed, as early as 1977, Anders Forsdahl postulated that there might be an important link between living conditions during childhood and adolescence and heart disease in later life [##REF##884401##3##]. This link was subsequently explored by David Barker and colleagues at the University of Southampton, whose research increasingly focused on the relationship between birth weight and health in later life, using low birth weight as a marker of undernutrition and adverse circumstances <italic>in utero </italic>[##UREF##0##4##]. This research suggested that fetal life might be a particularly sensitive period of development during which physiological mechanisms might be 'programmed' (i.e. permanently altered) in such a way that they lead to an increased risk of CVD in later life [##REF##11312225##5##,##REF##15857857##6##]. However, other researchers have speculated that postnatal development may contain even more sensitive periods than prenatal development [##REF##15145640##7##], with puberty being a particularly sensitive period [##REF##17436055##8##]. Indeed, a growing number of researchers now believe that health in later life is the result of exposures which accumulate and/or interact with one another across the <underline>entire</underline> life course, rather than those that 'programme' future health during a limited number of 'critical periods' [##UREF##1##9##].</p>", "<p>Since it is difficult, expensive and time consuming to undertake prospective studies examining the relationship between living conditions in early life and health in later life, retrospective analyses of cohorts exposed to food deprivation during two 'natural experiments' – the 1944–45 Dutch famine and the 1941–44 Leningrad siege – have proved invaluable for clarifying this relationship. Although these analyses have primarily focused on the relationship between prenatal exposure to food deprivation and health in later life, they include some that have examined the potential impact of postnatal undernutrition during childhood, adolescence or early adulthood. These analyses have found a higher incidence of breast cancer [##UREF##2##10##,##REF##15069116##11##] – though not other cancers [##UREF##2##10##,##UREF##3##12##, ####REF##11549558##13##, ##REF##12491504##14####12491504##14##] – amongst cohorts exposed to the Dutch famine during childhood and adolescence, higher mortality from ischaemic heart disease and stroke amongst men exposed to the Leningrad siege at 9–15 years of age [##REF##17436055##8##,##REF##14660443##15##] and higher systolic blood pressure amongst those exposed to the Leningrad siege at 6–15 years of age [##REF##17436055##8##].</p>", "<p>The present study aimed to clarify the nature of the relationship between undernutrition during pre- and postnatal development and CVD in later life. This involved examining a cohort of Guernsey islanders (born between 1923 and 1937) for whom data on birth weight were available (as a marker of prenatal nutrition), some of whom had been exposed as children, adolescents or young adults to food deprivation (i.e. postnatal undernutrition) during the 1940–45 German occupation of the Channel Islands [##UREF##4##16##]. The occupation culminated in a 9 month siege (from 1944 to 1945) when the islands were cut off from mainland France by the Allied liberation of Normandy following the D-Day landings. During this period, the official ration for islanders fell to around 1000 kcal per day – below the threshold identified as critical by researchers studying cohorts exposed to the Dutch famine [##UREF##5##17##]. By this stage of the occupation very little off-ration food was available [##UREF##6##18##], although the poor and those living in urban areas are likely to have suffered most, since these groups had the least funds to purchase whatever food was still available through the black market, and would have had least access to off-ration agricultural produce [##UREF##7##19##]. However, the paucity of fuel, water and soap, together with a more general decline in both the quantity and quality of food available, meant that the 1944–45 siege affected all but the most privileged islanders.</p>", "<p>Previous studies examining the short-term consequences of the 1940–45 occupation on the health of Channel Islanders have found little evidence of any impact on self-reported birth weight [##UREF##8##20##] or on infant and under-5 mortality rates [##UREF##9##21##]. However, children resident in the islands during the occupation displayed slower rates of growth in both height and weight [##REF##16715836##22##], as well as delayed age at menarche [##REF##17504356##23##]. Preliminary analyses exploring the impact of exposure to the occupation on health in later life found that middle-aged men born in the Channel Islands in 1939–1940, before the occupation began, had significantly higher systolic blood pressure and blood glucose levels than those born in 1945–1946, after the occupation ended [##UREF##10##24##,##UREF##11##25##]. It therefore appears that postnatal exposure to the occupation during childhood and adolescence may have had some short- <underline>and</underline> long-term effects on health. The analyses that follow therefore aim to compare the impact of birth weight and postnatal exposure to the occupation on CVD in later life.</p>" ]

[ "<title>Methods</title>", "<p>The cohort was based on 1673 live births delivered by a community midwife in the 10 parishes of Guernsey between February 1923 and August 1937. The midwife's records included data on birth weight, date and place of birth, sex of the baby and gestational age in weeks. These data were confidentially matched by hand to the Guernsey birth registers, which provided additional information on the occupation of the father and marital status of the mother, as well as confirmation of the parish of residence at birth. Paternal occupation at birth was classified as manual vs. non-manual and used as a marker of social class in the analyses which follow. It was also reclassified as 'agricultural' vs. 'non-agricultural' to explore the potential impact of paternal occupations with access to off-ration foodstuffs during the 1940–45 occupation. Likewise, parish of residence at birth was used as a marker of parish of residence during the occupation and was classified as 'urban' or 'rural' to reflect differences in the availability of off-ration foodstuffs in different parishes during the occupation [##UREF##12##26##,##UREF##13##27##].</p>", "<p>The cohort was then confidentially matched, again by hand, to the 'occupation register' compiled by the German Army during the 1940–45 occupation, which comprised a complete list of all resident islanders. In this way it was possible to identify those cohort members who were resident in Guernsey during the occupation, and those who were not (i.e. those who had left the island or had been evacuated before the 1940–45 occupation began). Finally, cohort members were confidentially matched to the hospital episode statistics (HES) database at the island's only hospital. HES for 1997–2005 were then searched for any admissions for CVD events – defined as myocardial infarction, stroke or unstable angina using the ICD-10 coding system.</p>", "<p>Cox regression models were used to assess whether birth weight and/or exposure to the 1944–45 occupation might be associated with CVD in later life. Birth weight, which had been measured to the nearest 1/4 lb, was converted to kilograms (kg) and used as a continuous variable. The 'occupation register' compiled by the German army was used to identify those cohort members who had been exposed to the 1940–45 occupation (i.e. those resident in Guernsey) and those who had not (i.e. those who had left or had been evacuated from the island). The analyses that follow therefore provide hazard ratios for CVD-related hospital admissions corresponding to a 1 kg increase in birth weight, and to exposure to the 1940–45 occupation.</p>", "<p>To inform the design of the Cox regression models, an analytical framework was drawn up in the form of a causal path diagram [##REF##12435780##28##] which summarised the causal relationships considered likely to exist between all of the variables available for analysis (see Figure ##FIG##0##1##). Many of these relationships are reasonably straightforward and have been replicated in a number of previous studies. However, some of the relationships were particular to the present study and therefore had to be inferred. The latter included the relationships between paternal occupation at birth and parish of residence at birth, and between parish of residence at birth and CVD. For these relationships it was assumed that parish of residence at birth was associated with paternal occupation at birth [##UREF##14##29##] and would therefore be related to CVD in a similar way to paternal occupation at birth.</p>", "<p>From Figure ##FIG##0##1## it is evident that all but one of the covariates appeared to act as potential confounders because they preceded both of the exposures. The exception was exposure to the occupation, which would act as a competing exposure in analyses exploring the relationship between birth weight and CVD in later life, since it did not precede birth weight in this cohort and was not considered to be directly related to birth weight in the causal path diagram [##UREF##8##20##]. For this reason the analytical models exploring each of the exposure variables (birth weight and exposure to the occupation) included slightly different covariates acting as potential confounders and/or competing exposures. Three different sets of models were therefore developed to examine the relationships between each of these exposures and CVD in later life: one for birth weight, one for exposure to the occupation, and one exploring whether there was any interaction between birth weight and exposure to the occupation.</p>", "<p>The first set of models (A: birth weight) included sex, gestational age at birth and paternal occupation at birth (as potential confounders) and exposure to the 1940–45 occupation (as a competing exposure). The first set of models also tested for any interaction between birth weight and sex. The second set of models (B: exposure to the occupation) included gestational age at birth, birth weight, paternal occupation at birth and parish of residence at birth (as potential confounders). The second set of models also tested for any interaction between exposure to the occupation and parish of residence at birth. Finally, the third model (C: birth weight × exposure to the occupation interaction) included sex, gestational age at birth, paternal occupation at birth and parish of residence at birth (as potential confounders) and tested for any interaction between birth weight and exposure to the occupation. All of the temporal variables available for inclusion in the analyses (such as date of birth and date of admission to hospital) were found to be correlated with one another. For this reason, the temporal variable most strongly related to CVD in later life (age at first admission to hospital) was included as the core temporal variable in the Cox regression analyses.</p>", "<p>Sensitivity analyses were conducted to assess the possibility of selection bias occurring due to selective loss to follow-up in the HES database. These analyses assumed that those lost to follow-up did not appear in the HES database because they had <italic>not </italic>experienced a CVD event. Meanwhile, although the sample size used in the analyses that follow was predetermined by the numbers of cohort members who could be matched to the HES database, it was possible to conduct power calculations based on the cumulative incidence of CVD over a 7-year period estimated from the literature, and these were then compared with the actual effect sizes found in the Cox regression models.</p>", "<p>Formal ethical clearance for the study was granted by the States of Guernsey Board of Health's research ethics committee.</p>" ]

[ "<title>Results</title>", "<title>Loss to follow-up and selection bias</title>", "<p>Of the original 1673 live births included in the cohort, 113 (6.8%) had died before the age of 18, and were therefore excluded from these analyses of CVD in adults. Data on hospital admissions in the HES database were identified for 931 of the remaining 1560 cohort members (see Figure ##FIG##1##2##). Of the 629 cohort members not found in the HES database, 172 were recorded in the Guernsey death registers as having died prior to the introduction of the HES database in 1997, leaving 457 cohort members unaccounted for and lost to follow-up. Some of these cohort members may have left Guernsey permanently before or after the 1940–45 occupation, while others may be residents who have yet to be admitted to the island's hospital. Of the 931 cohort members found in the HES database, 873 also had complete birth data and were included in the analyses that follow. This represents 52% follow-up of the original cohort and 63% follow-up of those who were alive when the HES database was first introduced. The matching process identified 225 individuals who were registered as being resident in Guernsey during the 1940–45 occupation, and these were therefore classified as exposed. Since the 'occupation register' used for this purpose provided a complete list of all islanders who were resident at the time of the 1940–45 occupation, none of the remaining cohort members (N = 648) would have been exposed to the occupation, and these were therefore classified as unexposed.</p>", "<title>A: Birth weight and CVD</title>", "<p>The first set of models (A: birth weight) showed that although there was some evidence of a positive relationship between birth weight and CVD, this was not statistically significant either before or after adjustment for sex, paternal occupation at birth, gestational age at birth and exposure to the 1940–45 occupation (see Table ##TAB##0##1##). To test for any interaction between birth weight and sex the data were analysed separately by sex. This comparison found that, although there was a modest inverse relationship between birth weight and CVD in men and a modest positive relationship between birth weight and CVD in women, a formal test for interaction between birth weight and sex was not statistically significant (p = 0.60). Meanwhile, paternal occupation at birth also approached statistical significance in its relationship with CVD, such that those cohort members whose fathers had manual occupations were more likely to have experienced CVD. However, exposure to the 1940–45 occupation was the only variable that displayed a statistically significant relationship with CVD, and this is examined in greater detail in the second set of models (B; see below).</p>", "<title>B: Exposure to the occupation and CVD</title>", "<p>The second set of models (B: exposure to the occupation) confirmed that there was a significant relationship between exposure to the occupation and CVD, such that those who had remained on the island during the occupation had more than twice the HR of hospital admission for CVD in later life (see Table ##TAB##1##2##). After adjustment for potential confounders, this relationship remained statistically significant. Recoding paternal occupation at birth to test for any difference in CVD between cohort members whose fathers had agricultural vs. non-agricultural occupations (and would therefore have had more vs. less access to off-ration foodstuffs, respectively [##UREF##6##18##]), did not alter the relationship between this variable and CVD, nor did it affect the significant relationship between exposure to the occupation and CVD. However, parish of birth approached significance as a predictor of CVD in later life, and this is likely to reflect residual differences in social class related to the higher proportion of non-manual paternal occupations in urban parishes [##REF##12435780##28##]. There was also evidence of a statistically significant interaction between exposure to the occupation and parish of residence at birth (p = 0.01), the size of the coefficient being higher amongst those resident in urban parishes (where the availability of off-ration food was lowest, but where paternal occupational class was highest [##REF##12435780##28##]).</p>", "<title>C: Interaction between birth weight and exposure to the occupation</title>", "<p>The third model (C: interaction between birth weight and exposure to the occupation) divided the cohort into those who had been exposed to the occupation and those who had not and examined the relationship between birth weight and CVD in each of these groups to assess whether there was any evidence of an interaction (see Table ##TAB##2##3##). This found no evidence of an interaction between birth weight and exposure to the occupation, and a formal test for interaction was not statistically significant (p = 0.43).</p>", "<title>Sensitivity analyses and power calculations</title>", "<p>The sensitivity analyses, which assumed that all those lost to follow-up had not been admitted to hospital for CVD, produced results that were no different for any of the three models in Tables ##TAB##0##1##, ##TAB##1##2##, ##TAB##2##3##. Meanwhile, the power calculations suggested that the number of cohort members with lower than mean birth weight in each of the analyses summarised in Table ##TAB##0##1## (n = 425) was sufficient to detect just under a doubling of the hazard for CVD (or a HR of 1.92) at 90% power and a significance level of 0.05. If the Cox regression models in Table ##TAB##0##1## had been calculated with 'lower than mean birth weight' as an exposure (rather than per kg increase in birth weight), the hazard ratio observed in the unadjusted model would have been 1.47. This is substantially smaller than that powered by the study, suggesting that the study was underpowered to examine a relationship of smaller magnitude between birth weight and CVD in later life and is therefore at risk of a type II error. For the models exploring the relationship between exposure to the occupation and CVD (Table ##TAB##1##2##), similar power calculations suggested that the number of cohort members who had been exposed to the occupation in the analyses summarised in Table ##TAB##1##2## (n = 225) was sufficient to detect a doubling of the hazard for CVD (i.e. an HR of 1.99) at 90% power with a significance level of 0.05 and without adjustment for potential confounders. This is less than the actual effect size detected by either of the analyses presented in the second set of models (Table ##TAB##1##2##), and it therefore seems likely that the study was adequately powered for these models.</p>" ]

[ "<title>Discussion</title>", "<title>Potential limitations</title>", "<p>This study had a number of potential limitations which should be taken into account when interpreting its results. First, the cohort comprises only home births delivered by a single community midwife, and as such is unlikely to be representative of all births in Guernsey at that time. Indeed, when these births were compared to a 10% representative sample of births during the same period drawn from the Guernsey birth registers, it was found that the births delivered by the community midwife were significantly more likely to have taken place in urban parishes [##UREF##14##29##]. Sociodemographic differences were found in separate analyses comparing those who had been resident in Guernsey during the occupation and those who had been evacuated and subsequently returned [##UREF##15##30##], as well as analyses comparing cohort members of above and below average birth weight. However, because all the models adjusted for age (as this was the temporal variable included in the Cox regression analyses), parish of residence at birth (for models B), and sex (for models A), any potential selection bias related to these three sociodemographic variables should have been ameliorated in these analyses.</p>", "<p>Nonetheless, the study suffers from substantial loss to follow-up, since the main analyses only include those cohort members who could be linked to the HES database. Those lost to follow-up would include any cohort members who had died prior to the introduction of the HES database in 1997, any who were evacuated from Guernsey and did not return, and any still living in Guernsey who had yet to attend the island's hospital and thereby enter the HES database (whether for CVD or non-CVD complaints). However, a comparison of cohort members included in the HES database with those who were not found few sociodemographic or clinical differences, although those included in the HES database were less likely to have died in Guernsey and more likely to have been exposed to the 1940–45 occupation. While such differences will affect the generalisability of the results, they may well be the result of population movement rather than pre-existing sociodemographic differences between those included in the analyses and those lost to follow-up. In this regard, the results of the sensitivity analyses were encouraging since these were not substantially different from the main analyses when all those lost to follow-up were assumed <italic>not </italic>to have experienced a CVD event.</p>", "<p>Finally, the power calculations indicated that the analyses in the first set of models (A: birth weight) are likely to have been underpowered. If so, it is possible that a significant relationship may have been found between birth weight and CVD had a larger sample been available for analysis. However, because the modest relationship between birth weight and CVD amongst men and women combined was positive rather than inverse, and was only inverse amongst men when the analysis was stratified by sex, it seems unlikely that a larger sample size would have generated the statistically significant inverse relationship between birth weight and CVD that has commonly been observed in previous studies [##UREF##0##4##].</p>", "<title>Interpretation of findings</title>", "<p>Notwithstanding these potential limitations, the lack of a significant inverse relationship between birth weight and CVD in this cohort runs contrary to the findings of numerous previous studies [##REF##15857857##6##,##REF##2495113##31##, ####REF##15554948##32##, ##REF##14729890##33####14729890##33##]. There was also no apparent interaction between birth weight and exposure to the occupation, suggesting that this had not masked any underlying association between birth weight and CVD. However, in many of the previous studies examining the relationships between birth weight, CVD and associated risk factors, these relationships were only apparent after adjusting for one or more measures of current body size – an analytical approach which can invoke a statistical artefact known as the 'reversal paradox' [##REF##12241871##34##,##REF##16691183##35##]. Although it was not possible to adjust for current body size in the present study, it would not have been appropriate to do so had data on current body size been available. Indeed, the absence of a significant negative relationship between birth weight and CVD in the present study adds to concerns that previous reports of such a relationship may be partly artefacts of the 'reversal paradox'.</p>", "<p>Meanwhile, the results of both the first and second sets of models (see Tables ##TAB##0##1## and ##TAB##1##2##) seem to provide clear evidence that exposure to the 1940–45 Channel Islands occupation was associated with an increased likelihood of hospital admission for CVD in later life. Differential exposure to undernutrition and related deprivation in early life between those who remained in Guernsey during the 1940–45 occupation and those who left or were evacuated before the occupation began [##UREF##4##16##,##UREF##13##27##] is one plausible explanation for these findings. Likewise, it is possible that the stress associated with the occupation may have had a number of permanent psychosocial effects that led to a differential risk of CVD in later life. However, those cohort members who were evacuated from the islands before the occupation began often left their families behind and were sent in large groups to unfamiliar towns in the UK, many of which experienced sustained bombing during the war. It therefore seems likely that <underline>both</underline> resident <underline>and</underline> evacuated islanders would have experienced some form of stress, whilst only those who remained on the island would have experienced the undernutrition and related deprivation associated with the 1940–45 occupation.</p>", "<p>It is nonetheless possible that these findings reflect residual confounding for pre-existing sociodemographic differences between islanders resident in Guernsey throughout the 1940–45 occupation and those who left the island or were evacuated before the occupation began [##UREF##15##30##]. Given that the analyses presented here had limited access to sociodemographic variables (such as parish of residence and paternal occupation at birth), it seems likely that these analyses will have suffered from residual confounding. However, it is reassuring that the increase in CVD amongst cohort members resident on the island during the 1940–45 occupation remained statistically significant when paternal occupation was recoded as 'agricultural' vs. 'non-agricultural' and in the sensitivity analyses which assumed that cohort members missing from the HES database had <italic>not </italic>experienced CVD.</p>", "<p>We are therefore confident that the results of these analyses strengthen the evidence provided by two previous studies of cohorts exposed in childhood to nutritional deprivation during the Leningrad siege, which found that cardiovascular mortality and systolic blood pressure were significantly elevated among those exposed at the age of 9–15 years [##REF##14660443##15##], while men exposed at the age of 6–8 years also had increased mortality from ischaemic heart disease [##REF##17436055##8##]. While these are similar findings to those observed in the present study, it is nonetheless important to note that conditions during the 1941–44 Leningrad siege were much more severe than those in Guernsey during the 1940–45 occupation and siege, with widespread starvation of the civilian population [##UREF##16##36##]. Nonetheless, one explanation for the Leningrad studies' findings was that children over the age of 12 appeared to have received disproportionately fewer rations than younger children, and a similar scenario has been described in accounts of the 1940–44 German occupation of Belgium [##UREF##17##37##] and in accounts of rationing from the 1940–45 occupation of the Channel Islands, which suggest that the young and the elderly suffered most [##UREF##12##26##,##UREF##13##27##,##UREF##18##38##]. Indeed, as in Leningrad, some accounts from the Channel Islands describe a scenario in which those in their early teenage years bore the brunt of inadequate food supplies during the occupation because those over 12 years old were often expected to work yet did not receive proportionately higher rations [##UREF##19##39##].</p>", "<p>Moreover, there is further evidence that children and adolescents might have experienced longer-term developmental delays as a result of the Channel Islands siege, in the form of delayed age at menarche [##REF##17504356##23##] and slower growth rates [##REF##16715836##22##]. In particular, a study of Jersey schoolchildren during the occupation found that children who remained on the island displayed significantly slower growth compared to children in the UK [##REF##16715836##22##]. While the lower rates of weight gain were temporary, the lower rates of height gain appeared longer lasting, indicating that exposure to the occupation may have had some longer-lasting developmental effects. Similar patterns of impaired growth were observed following the 1940–44 German occupation of Belgium, where girls who reached puberty during this time grew up to be shorter and lighter adults [##UREF##17##37##]. Likewise, in a study of women exposed to the Dutch famine, those exposed at the ages of 0–9 and 12–16 were found to have shorter stature as adults [##UREF##20##40##]. It is therefore possible that these effects on the stature of children and adolescents may reflect the developmental mechanism(s) through which exposure to the 1940–45 occupation of the Channel Islands might have led to an increased risk of CVD in later life.</p>", "<p>This interpretation is supported by the significant interaction between exposure to the occupation and parish of residence at birth (used as a marker of parish of residence during the occupation), which indicated that exposure to the occupation was more strongly associated with CVD amongst those born/living in urban parishes where there was less access to off-ration foodstuffs [##UREF##12##26##,##UREF##13##27##]. Authors studying the immediate health effects of the Dutch famine have found similar differences between those living in rural and urban areas, and also attributed these to disparities in food availability therein [##UREF##21##41##]. As such, the statistically significant interaction between exposure to the occupation and parish of residence at birth observed in the present study supports the conclusion that the differences in hospital admission for CVD were due to undernutrition and related deprivation during the 1940–45 occupation. The interaction between exposure to the occupation and parish of residence at birth also raises the possibility of either a linear relationship or threshold effect involving the severity of food deprivation during the occupation and subsequent CVD.</p>" ]

[ "<title>Conclusion</title>", "<p>The present study aimed to explore the impact of undernutrition at various stages of early development on the subsequent risk of CVD in later life. In contrast to many previous studies, there was little evidence of an inverse relationship between birth weight (as a marker of nutrition <italic>in utero</italic>) and CVD in later life, although these analyses were substantially underpowered. In better powered analyses, there was clear evidence that exposure to nutritional deprivation during the 1940–45 occupation of the Channel Islands in childhood, adolescence and young adulthood was associated with an increased risk of CVD in later life. These findings suggest that, in this cohort, postnatal undernutrition associated with exposure to the 1940–45 occupation may be a more important determinant of CVD in later life than the levels of undernutrition experienced <italic>in utero </italic>prior to the occupation.</p>" ]

[ "<p>This is an Open Access article distributed under the terms of the Creative Commons Attribution License (<ext-link ext-link-type=\"uri\" xlink:href=\"http://creativecommons.org/licenses/by/2.0\"/>), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.</p>", "<title>Background</title>", "<p>To clarify the nature of the relationship between food deprivation/undernutrition during pre- and postnatal development and cardiovascular disease (CVD) in later life, this study examined the relationship between birth weight (as a marker of prenatal nutrition) and the incidence of hospital admissions for CVD from 1997–2005 amongst 873 Guernsey islanders (born in 1923–1937), 225 of whom had been exposed to food deprivation as children, adolescents or young adults (i.e. postnatal undernutrition) during the 1940–45 German occupation of the Channel Islands, and 648 of whom had left or been evacuated from the islands before the occupation began.</p>", "<title>Methods</title>", "<p>Three sets of Cox regression models were used to investigate (A) the relationship between birth weight and CVD, (B) the relationship between postnatal exposure to the occupation and CVD and (C) any interaction between birth weight, postnatal exposure to the occupation and CVD. These models also tested for any interactions between birth weight and sex, and postnatal exposure to the occupation and parish of residence at birth (as a marker of parish residence during the occupation and related variation in the severity of food deprivation).</p>", "<title>Results</title>", "<p>The first set of models (A) found no relationship between birth weight and CVD even after adjustment for potential confounders (hazard ratio (HR) per kg increase in birth weight: 1.12; 95% confidence intervals (CI): 0.70 – 1.78), and there was no significant interaction between birth weight and sex (p = 0.60). The second set of models (B) found a significant relationship between postnatal exposure to the occupation and CVD after adjustment for potential confounders (HR for exposed vs. unexposed group: 2.52; 95% CI: 1.54 – 4.13), as well as a significant interaction between postnatal exposure to the occupation and parish of residence at birth (p = 0.01), such that those born in urban parishes (where food deprivation was worst) had a greater HR for CVD than those born in rural parishes. The third model (C) found no interaction between birth weight and exposure to the occupation (p = 0.43).</p>", "<title>Conclusion</title>", "<p>These findings suggest that the levels of postnatal undernutrition experienced by children, adolescents and young adults exposed to food deprivation during the 1940–45 occupation of the Channel Islands were a more important determinant of CVD in later life than the levels of prenatal undernutrition experienced <italic>in utero </italic>prior to the occupation.</p>" ]

[ "<title>Competing interests</title>", "<p>The authors declare that they have no competing interests.</p>", "<title>Authors' contributions</title>", "<p>RFH contributed to the design of the study, collection of data, analysis and interpretation, and also drafted and approved the manuscript. MSG contributed to the design of the study, analysis and interpretation, and also assisted with drafting the manuscript and approved the final version. AB contributed to the collection of data and approved the final manuscript. GTHE contributed to the conception and design of the study, collection of data, analysis and interpretation, and also assisted with drafting the manuscript and approved the final version.</p>", "<title>Pre-publication history</title>", "<p>The pre-publication history for this paper can be accessed here:</p>", "<p><ext-link ext-link-type=\"uri\" xlink:href=\"http://www.biomedcentral.com/1471-2458/8/303/prepub\"/></p>" ]

[ "<title>Acknowledgements</title>", "<p>This study would not have been possible without the assistance of staff at Guernsey's Island Archive Service (Dr Darryl Ogier and Don Le Tissier) and Greffe (Keith Robilliard). The study forms part of the Channel Islands Occupation Birth Cohort Study which is funded by the Lloyds TSB Foundation for the Channel Islands. All researchers are independent of the funding body. RFH is funded by a VIP Wellcome Trust Fellowship, MSG is funded by HEFCE, AB is funded by the Guernsey Board of Health, and GTHE is funded by HEFCE.</p>" ]

[ "<fig position=\"float\" id=\"F1\"><label>Figure 1</label><caption><p><bold>Causal path diagram showing relationships between variables available for inclusion in the analyses</bold>. Solid lines depict relationships presumed to be causal on the basis of previous studies or circumstances on Guernsey before, during and after the 1940–45 occupation. Dotted lines depict relationships examined by the analyses presented in Tables 1–3.</p></caption></fig>", "<fig position=\"float\" id=\"F2\"><label>Figure 2</label><caption><p><bold>Structure of the Guernsey midwife cohort</bold>. Boxes in bold are those remaining in the cohort at each stage, and unbolded boxes are those removed from the cohort at each stage.</p></caption></fig>" ]

[ "<table-wrap position=\"float\" id=\"T1\"><label>Table 1</label><caption><p>Cox regression analysis for models exploring the relationship between birth weight and CVD in later life (A).</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td/><td align=\"center\"><bold>Covariate </bold><break/>(referent category)<break/></td><td align=\"center\"><bold>Hazard </bold><break/><bold>Ratio</bold><break/></td><td align=\"center\"><bold>95% </bold><break/><bold>confidence </bold><break/><bold>intervals</bold></td><td align=\"center\"><bold>p value</bold><break/><break/></td></tr></thead><tbody><tr><td align=\"center\"><bold>Unadjusted</bold></td><td align=\"center\">Birth weight (per kg increase)</td><td align=\"center\">1.30</td><td align=\"center\">0.82 – 2.07</td><td align=\"center\">0.26</td></tr><tr><td align=\"center\"><bold>Unstratified, adjusted for potential confounders¹and competing exposures²</bold></td><td align=\"center\">Birth weight (per kg increase)</td><td align=\"center\">1.12</td><td align=\"center\">0.70 – 1.78</td><td align=\"center\">0.65</td></tr><tr><td/><td align=\"center\">Female sex (male)</td><td align=\"center\">0.76</td><td align=\"center\">0.47 – 1.24</td><td align=\"center\">0.27</td></tr><tr><td/><td align=\"center\">Preterm gestational age (term)</td><td align=\"center\">0.26</td><td align=\"center\">0.04 – 1.91</td><td align=\"center\">0.19</td></tr><tr><td/><td align=\"center\">Manual paternal occupation (non-manual)</td><td align=\"center\">1.42</td><td align=\"center\">0.88 – 2.30</td><td align=\"center\">0.15</td></tr><tr><td/><td align=\"center\">Exposure to occupation (unexposed)</td><td align=\"center\">2.65</td><td align=\"center\">1.62 – 4.34</td><td align=\"center\">0.01</td></tr><tr><td align=\"center\"><bold>Stratified by sex and adjusted for potential confounders¹and competing exposures²</bold></td><td align=\"center\">Birth weight in men (per kg increase)</td><td align=\"center\">0.94</td><td align=\"center\">0.52 – 1.72</td><td align=\"center\">0.60³</td></tr><tr><td/><td align=\"center\">Birth weight in women (per kg increase)</td><td align=\"center\">1.60</td><td align=\"center\">0.73 – 3.50</td><td/></tr></tbody></table></table-wrap>", "<table-wrap position=\"float\" id=\"T2\"><label>Table 2</label><caption><p>Cox regression analysis for models exploring the relationship between exposure to the occupation and CVD in later life (B).</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td/><td align=\"center\"><bold>Covariate </bold><break/>(referent category)<break/></td><td align=\"center\"><bold>Hazard </bold><break/><bold>Ratio</bold><break/></td><td align=\"center\"><bold>95% </bold><break/><bold>confidence</bold><break/><bold>intervals</bold></td><td align=\"center\"><bold>p value</bold><break/><break/></td></tr></thead><tbody><tr><td align=\"center\"><bold>Unadjusted</bold></td><td align=\"center\">Exposure to occupation (unexposed)</td><td align=\"center\">2.88</td><td align=\"center\">1.78 – 4.66</td><td align=\"center\">0.01</td></tr><tr><td align=\"center\"><bold>Unstratified, adjusted for potential confounders</bold>¹</td><td align=\"center\">Exposure to occupation (unexposed)</td><td align=\"center\">2.52</td><td align=\"center\">1.54 – 4.13</td><td align=\"center\">0.01</td></tr><tr><td/><td align=\"center\">Preterm gestational age (term)</td><td align=\"center\">0.28</td><td align=\"center\">0.04 – 2.05</td><td align=\"center\">0.21</td></tr><tr><td/><td align=\"center\">Birth weight (per kg increase)</td><td align=\"center\">1.15</td><td align=\"center\">0.73 – 1.83</td><td align=\"center\">0.55</td></tr><tr><td/><td align=\"center\">Manual paternal occupation (non-manual)</td><td align=\"center\">1.41</td><td align=\"center\">0.87 – 2.29</td><td align=\"center\">0.16</td></tr><tr><td/><td align=\"center\">Rural parish of birth (urban)</td><td align=\"center\">1.61</td><td align=\"center\">0.97 – 2.69</td><td align=\"center\">0.07</td></tr><tr><td align=\"center\"><bold>Stratified by parish, and adjusted for potential confounders</bold>¹</td><td align=\"center\">Exposure to occupation in urban parishes (unexposed)</td><td align=\"center\">2.75</td><td align=\"center\">1.41 – 5.35</td><td align=\"center\">0.01²</td></tr><tr><td/><td align=\"center\">Exposure to occupation in rural parishes (unexposed)</td><td align=\"center\">2.28</td><td align=\"center\">1.08 – 4.81</td><td/></tr></tbody></table></table-wrap>", "<table-wrap position=\"float\" id=\"T3\"><label>Table 3</label><caption><p>Cox regression analysis for model exploring interaction between birth weight and exposure to the occupation (C).</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td/><td align=\"center\"><bold>Covariate </bold><break/>(referent category)<break/></td><td align=\"center\"><bold>Hazard </bold><break/><bold>Ratio</bold><break/></td><td align=\"center\"><bold>95% </bold><break/><bold>confidence </bold><break/><bold>intervals</bold></td><td align=\"center\"><bold>p value</bold><break/><break/></td></tr></thead><tbody><tr><td align=\"center\"><bold>Stratified by exposure to the occupation, and adjusted for potential confounders</bold>¹</td><td align=\"center\" colspan=\"4\"><bold>Exposed individuals</bold></td></tr><tr><td align=\"center\"/><td align=\"center\">Birth weight amongst those exposed to the occupation (per kg increase)</td><td align=\"center\">1.22</td><td align=\"center\">0.65 – 2.28</td><td align=\"center\">0.54</td></tr><tr><td/><td align=\"center\">Female sex (male)</td><td align=\"center\">0.63</td><td align=\"center\">0.33 – 1.17</td><td align=\"center\">0.14</td></tr><tr><td/><td align=\"center\">Preterm gestational age (term)</td><td align=\"center\">0.32</td><td align=\"center\">0.04 – 2.31</td><td align=\"center\">0.26</td></tr><tr><td/><td align=\"center\">Manual paternal occupation (non-manual)</td><td align=\"center\">0.93</td><td align=\"center\">0.49 – 1.77</td><td align=\"center\">0.82</td></tr><tr><td/><td align=\"center\">Rural parish of birth (urban)</td><td align=\"center\">1.93</td><td align=\"center\">0.99 – 3.75</td><td align=\"center\">0.05</td></tr><tr><td colspan=\"5\"><hr/></td></tr><tr><td/><td align=\"center\" colspan=\"4\"><bold>Unexposed individuals</bold></td></tr><tr><td/><td align=\"center\">Birth weight amongst those unexposed to the occupation (per kg increase)</td><td align=\"center\">0.92</td><td align=\"center\">0.44 – 1.90</td><td align=\"center\">0.81</td></tr><tr><td/><td align=\"center\">Female sex (male)</td><td align=\"center\">0.93</td><td align=\"center\">0.42 – 2.04</td><td align=\"center\">0.85</td></tr><tr><td/><td align=\"center\">Preterm gestational age (term)</td><td align=\"center\">-²</td><td align=\"center\">-²</td><td align=\"center\">-²</td></tr><tr><td/><td align=\"center\">Manual paternal occupation (non-manual)</td><td align=\"center\">2.53</td><td align=\"center\">1.15 – 5.58</td><td align=\"center\">0.02</td></tr><tr><td/><td align=\"center\">Rural parish of birth (urban)</td><td align=\"center\">1.15</td><td align=\"center\">0.51 – 2.58</td><td align=\"center\">0.73</td></tr></tbody></table></table-wrap>" ]

[]

[]

[]

[]

[]

[]

[ "<table-wrap-foot><p>¹Sex, gestational age at birth and paternal occupation at birth were all considered potential confounders. It is important to note that the interpretation of the estimated hazard ratios for potential confounders is not straightforward since the main exposure variable (birth weight) lies on the pathway between these potential confounders and the outcome variable (CVD in later life).</p><p>²Exposure to the 1940–45 occupation was considered a competing exposure.</p><p>³p value for interaction between birth weight and sex.</p></table-wrap-foot>", "<table-wrap-foot><p>¹Gestational age at birth, birth weight, paternal occupation at birth and parish of residence at birth were all considered potential confounders. It is important to note that the interpretation of the estimated hazard ratios for potential confounders is not straightforward since the main exposure variable (exposure to the occupation) lies on the pathway between these potential confounders and the outcome variable (CVD in later life).</p><p>²p value for interaction between birth weight and sex.</p></table-wrap-foot>", "<table-wrap-foot><p>¹Sex, gestational age at birth, paternal occupation at birth and parish of residence at birth were all considered potential confounders. It is important to note that the interpretation of the estimated hazard ratios for potential confounders is not straightforward since the main exposure variable (birth weight) lies on the pathway between these potential confounders and the outcome variable (CVD in later life).</p><p>²There were no cohort members in this category who were preterm and who had also experienced a CVD event</p></table-wrap-foot>" ]

[ "<graphic xlink:href=\"1471-2458-8-303-1\"/>", "<graphic xlink:href=\"1471-2458-8-303-2\"/>" ]

[]

{ "acronym": [], "definition": [] }

[ "<title>Background</title>", "<p><italic>Chlamydia trachomatis </italic>is the most common sexually transmitted bacterial pathogen in the United States; infection results in devastating sequelae, including pelvic inflammatory disease and infertility. Animal models and clinical studies of infected patients have indicated that protective chlamydial immunity is primarily mediated by Th1 responses [##REF##11908999##1##, ####REF##12010958##2##, ##REF##11834375##3####11834375##3##]. The induction of such immunity involves rapid recruitment and activation of certain effector immune cells, specifically Th1 cells and dendritic cells (DCs) into the local genital mucosa to clear the infection, arrest ascending disease, and prevent major complications [##REF##11908999##1##,##REF##7729886##4##, ####REF##9802562##5##, ##REF##9673231##6##, ##REF##10678969##7####10678969##7##]. In addition, certain complimentary B cell functions – principally Ab-mediated enhancement of Ag presentation – lead to activation of Ag-specific Th1 cells dependent in part by Fc-receptor-mediated events [##REF##11908999##1##,##UREF##0##8##]. Thus, both cell-mediated and humoral immune responses are required for long-term protection against <italic>Chlamydia</italic>.</p>", "<p>With the major elements of protective anti-<italic>Chlamydia </italic>immunity defined, a number of candidate vaccines have been described [##REF##11908999##1##]. Chemokines have emerged as important factors and possible mucosal adjuvants that function in lymphocyte activation and recruitment [##REF##9973465##9##, ####REF##11123289##10##, ##REF##12393512##11####12393512##11##]. Indeed, a qualitative relationship exists between the class of chemokines secreted following infection, the type of immune response (cellular or humoral immunity) elicited, and the fate of the host following infection [##REF##9864948##12##, ####REF##11175801##13##, ##REF##11705916##14##, ##REF##8871660##15####8871660##15##]. The profile of chemokine expression serves as a reliable indicator of immune response type (i.e., Th1 <italic>vs</italic>. Th2). In this respect, the CCL5-CCR5 axis has been demonstrated to be preferentially involved in the activation and function of Thl cells [##REF##11123289##10##,##REF##9419219##16##,##REF##9302298##17##].</p>", "<p>CCL5 is secreted by epithelial cells, macrophages, fibroblasts, platelets, and activated T cells [##REF##11286708##18##]. This CC chemokine is known to regulate T cell differentiation and polarize Th1 &gt;&gt; Th2 subtypes [##REF##11123289##10##,##REF##11175801##13##,##REF##11286708##18##,##REF##11890717##19##] as well as numerous physiological functions of leukocytes including migration. Polymorphisms in CCR5 and CCL5 modulate immune responses as well as susceptibility and progression to HIV-1 and AIDS, respectively [##REF##10200305##20##,##REF##11125885##21##]. We also showed that many of the deleterious complications of genital chlamydial infection, due to Th1-mediated inflammation, are not present in individuals with the <italic>ccr5Δ32 </italic>mutation or in CCR5-deficient mice [##REF##16118671##22##]. CCR5 expression following genital chlamydial infection is followed by an early Th1-like response that precedes activation and mucosal recruitment of Ag-specific Th1 cells necessary for clearance of <italic>Chlamydia </italic>[##REF##11796619##23##]. These findings indicate that CCL5 might be important for inducing protective immunity against <italic>Chlamydia</italic>. However, it is not certain what affect CCL5 deficiency would have on chlamydial disease. We tested the hypothesis that CCL5 is essential for inducing adaptive mucosal immunity against <italic>Chlamydia </italic>by Ab inhibition using a reliable mouse model of genital chlamydial infection. Results revealed CCL5 supports the induction of Th1 cytokine and immunoglobulin IgG2a as well as IgA responses against <italic>Chlamydia</italic>.</p>" ]

[ "<title>Methods</title>", "<title><italic>Chlamydia </italic>stocks</title>", "<p>Stocks of the <italic>Chlamydia muridarum </italic>(trachomatis agent for mice) were prepared by propagating elementary bodies (EBs) in McCoy cells and used to infect mice as previously described [##REF##3258586##49##]. The stocks were tittered by infecting McCoy cells with varying dilutions of EBs. The infectious titer was expressed as inclusion-forming units per milliliter (IFU/ml).</p>", "<title>Anti-CCL5 antibody production</title>", "<p>Murine anti-CCL5 was produced as previously described [##REF##16455992##34##]. Briefly, rabbits were immunized with murine CCL5 (Pepro-Tech) and the hyper immune CCL5 antisera obtained was yielded titers of ~1:10⁶such that 10 μl of rabbit CCL5 antiserum neutralized 20 ng of CCL5. This antiserum was titrated by direct ELISA, and no cross-reactivity was detected when tested against other CCR5 ligands, chemokines and cytokines. Subsequently, antisera were heat-inactivated and purified using an IgG isotype-specific protein A column (Pierce). Anti-CCL5 Ab titers were adjusted to 1 : 4 × 10⁵(i.e., 50× dilution) in PBS (Ab solution) for CCL5 blocking experiments. Similarly prepared normal or preimmune rabbit serum was used to generate the control Ab solution.</p>", "<title>Animal infection, Ab treatment, and analysis of disease course</title>", "<p>Female BALB/c mice, ages 6 to 8 weeks, were purchased from Jackson Laboratory and used to establish a colony at the Morehouse School of Medicine animal facility. The animals were housed and maintained in isolator cages under specific-pathogen-free housing conditions. The guidelines proposed by the Committee for the Care of Laboratory Animal Resources Commission of Life Sciences-National Research Council were followed to minimize animal pain and distress. These studies were approved by the Morehouse School of Medicine Institutional Animal Care and Use Committee (IACUC). Seven days prior to infection, mice received 2.5 mg of medroxy-progesterone acetate (Depo-Provera; The Upjohn Co.) by subcutaneous route in 100 μl of PBS [##REF##12898452##50##]. Groups were intravaginally infected with 5 × 10⁴IFU of <italic>C. muridarum </italic>under phenobarbital anesthesia, whereas uninfected control mice received only phosphate buffered saline (PBS). Mice (15 mice per group) received 100 μl of either control or anti-CCL5 Abs by intraperitoneal injection, 24 hrs before chlamydial infection and every 72 hrs thereafter. The extent of unresolved chlamydial infection was monitored by taking cervico-vaginal swabs from individual mice every week until 42 days after the challenge [##REF##16118671##22##]. Briefly,<italic>Chlamydia </italic>present in swabs were detected by infecting McCoy cells vaginal/swab rinses and staining infected monolayer with fluorescein isothiocyanate (FITC)-labeled, genus-specific, anti-<italic>Chlamydia </italic>Abs (Kallestad Diagnostics) to verify inclusion bodies (&gt; 10 IFUs/ml) by direct immuno-fluorescence [##REF##3258586##49##].</p>", "<title>Sample, tissue, and cell collection</title>", "<p>Vaginal cavities were rinsed three times with 50 μl of PBS to obtain mucosal secretions. Blood collected by retro-orbital bleeding and serum was separated following centrifugation. Serum and mucosal secretions were collected on 0, 7, 14, and 42 days post challenge and analyzed by ELISA. Following sacrifice by CO₂inhalation, spleen, ILN, fallopian tube(s), uterus, and cervix tissues were aseptically removed and single-cell suspensions were prepared by passing tissues through a sterile wire screen to quantify mRNA expression and T helper responses. Reproductive tract tissues were further disrupted to generate single cell suspensions by stirring in collagenase type IV (Sigma) in RPMI 1640 (collagenase solution) at 37°C for 30 min. Lymphocytes were further purified using a discontinuous Percoll (Pharmacia) gradient, collected at the 40/75% interface. CD4⁺T cells were enriched using Mouse CD4 Cellect^®plus columns according to manufacturer's protocol (Biotex Laboratories). Cell suspensions were washed twice in RPMI 1640 and lymphocytes were maintained in medium supplemented with 10 ml/L of nonessential amino acids (Mediatech), 1 mM sodium pyruvate (Sigma), 10 mM HEPES (Mediatech), 100 U/ml penicillin, 100 μg/ml streptomycin, 40 μg/ml gentamycin (Elkins-Sinn), 50 μM mercaptoethanol (Sigma), and 10% fetal calf serum (FCS) (Atlanta Biologicals).</p>", "<title>Chlamydia-specific T helper cell responses</title>", "<p>Purified CD4⁺T cells were cultured at a density of 5 × 10⁶cells/ml with 10⁶cells/ml γ-irradiated (3,000 rads) naïve splenic feeder cells in complete medium in the presence or absence of 10 μg/ml of UV-inactivated <italic>C. muridarum </italic>inclusion bodies (IBs) as Ag at 37°C in 5% CO₂. After 3 days of culture, cells were pulsed with BrdU labeling solution and incorporation was detected by ELISA (Roche Molecular Biochemical). Similarly, cell culture supernatants were collected 3 days after culture and assayed for cytokine secretion by Luminex assay [##REF##16455992##34##].</p>", "<title>Anti-Chlamydia Ab detection in serum and vaginal washes</title>", "<p><italic>Chlamydia</italic>-specific serum IgG subclass antibodies and vaginal wash IgA levels were quantified 6 weeks after challenge by ELISA [##REF##11123289##10##]. Briefly, 96-well Falcon ELISA plates (Fisher Scientific) were coated with 100 μl of anti-IgG or IgA Ab (BD-PharMingen) or 10 μg/ml of UV-inactivated <italic>C. muridarum </italic>IBs in PBS O/N at 4°C and blocked with 10% FCS in PBS for 2 hrs at RT. IgM, IgG subclasses or IgA standards, and experimental samples were serially added after diluted with PBS. After O/N incubation at 4°C and three washes using PBS containing 0.05% Tween 20 (PBS-T), Ag-specific titers of IgM, IgG, IgA, or IgG subclass Abs were determined following the addition of biotinylated detection Abs (BD-PharMingen). After incubation and wash steps, anti-biotin HRP Ab (Vector Laboratories Inc., Burlingame, CA) in PBS-T was added to detection wells and incubated for 1 hr at RT. Following incubation, all plates were washed 6 times and the color reaction was developed by adding 100 μl of 1.1 mM 2,2'-azino-bis(3)-ethylbenzthiazoline-6-sulfonic acid (Sigma) in 0.1 M citrate-phosphate buffer (pH 4.2) containing 0.01% H₂O₂(ABTS solution).</p>", "<title>Cytokine detection</title>", "<p>The presence of the following T helper cell-derived cytokines in culture supernatants was determined by Beadlyte™ mouse multi-cytokine detection system kit (BioRad): interleukin (IL)-2, IL-4, IL-6, IL-10, granulocyte monocyte cell stimulating factor (GM-CSF), and interferon (IFN)-γ. Filter-bottom ELISA plates (BioRad) were rinsed with 100 μl of Bioplex assay buffer and the buffer was removed using a Millipore™ multiscreen separation vacuum manifold system set at 5 mm Hg. Analyte beads in assay buffer were added into wells, followed by 50 μl of serum or standard solution and incubated for 30 mins at RT with continuous shaking (at setting #3) using a Lab-Line™ Instrument Titer Plate Shaker (Melrose, IL). The filter-bottom plates were washed as before and the buffer was removed using a Millipore™ multiscreen separation vacuum manifold system. Subsequently, 50 μl of anti-mouse IL-2, IL-4, IL-6, IL-10, GM-CSF, or IFN-γ Ab-biotin reporter solution was added in each well and the plates were incubated with continuous shaking for 30 mins followed by centrifugation and washing. Next, 50 μl of streptavidin-phycoerytherin (PE) solution was added and the plates were incubated with continuous shaking for 10 mins at RT. 125 μl of Bio-plex assay buffer was added and Beadlyte™ readings were measured using a Luminex™ System and calculated using Bio-plex™ software (Bio-Rad). The cytokine Beadlyte™ assays were capable of detecting &gt; 5 pg/ml for each analyte.</p>", "<title>RNA isolation and gene expression analysis</title>", "<p>Total RNA from the spleen, ILNs, fallopian tube, uterus, and cervix leukocytes was isolated from mouse treated with anti-CCL5 or control Ab using Tri-reagent™ (Molecular Research Center, Cincinnati, OH). Potential genomic DNA contamination was removed from these samples by treatment with RNase-free DNase (Invitrogen) for 15 mins at 37°C. RNA was then precipitated and re-suspended in RNA secure (Ambion). cDNA was generated by reverse transcribing approximately 1.5 μg of total RNA using Taqman™ reverse transcription reagent (Applied Biosystems).</p>", "<p>Mouse mRNA sequences of CCL5, CCR5, IFN-γ, and 18S rRNA were obtained from the NIH-NCBI gene bank database accession numbers NM03653, D83648, K00083, and X00686.1, respectively. These sequences were then used to design primers for real-time polymerase chain reaction (RT-PCR) analysis, which generated amplicons of 97, 100, 98, and 149 base pairs size, respectively, for CCL5 (sense-TCG TGT TTG TCA CTC GAA GG and antisense- GCT GAT GGC CTG ATT GTC TT), CCR5 (sense- CGA AAA CAC ATG GTC AAA CG and antisense- GGG AAG CGT ATA CAG GGT CA, IFN-γ (sense- ACT GGC AAA AGG ATG GTG AC and antisense- GTT CTC CTG TGG ATC GGG TA), and 18S rRNA (sense-GTA ACC CGT TGA ACC CAA TT and antisense- CAA TCC AAT CGG TAG TAG CG). Primers were designed using the primer 3 software program from Whitehead Institute at the Massachusetts Institute of Technology (MIT). Thermodynamic analysis of primers was conducted using the following computer programs: Primer Premier™ (Integrated DNA Technologies) and MIT Primer III (Boston, MA). The resulting primer sets were compared against the entire murine genome using the National Center for Biotechnology Information (NCBI) to confirm specificity and ensure that the primers flanked mRNA splicing regions. cDNA was generated as before and amplified with specific cDNA primers using SYBR^®Green PCR master mix reagents (Applied Biosystems). The copy number (&gt; 10) of mRNA relative to 18S rRNA copies was evaluated by RT-PCR analysis using the BioRad Icycler and software (Hercules, CA).</p>", "<title>Statistics</title>", "<p>Data were expressed as the mean ± standard error of mean (SEM), compared using a two-tailed student's <italic>t</italic>-test or an unpaired Mann Whitney <italic>U</italic>-test, and considered statistically significant if <italic>p </italic>&lt; 0.01. When cytokine levels were below detection (BD) limit, they were recorded as one-half the lower detection limit (e.g., 5 pg/ml for IL-10) for statistical analysis.</p>" ]

[ "<title>Results</title>", "<title>Expression of chemokines after genital Chlamydia infection</title>", "<p>CCL5, CCR5, and IFN-γ mRNAs were measured by quantitative RT-PCR analysis after genital chlamydial infection. A significant increase in CCR5, CCL5, and IFNγ gene expression in the spleen and ILN was observed 7 days after genital infection when compared with levels before infection (Figure ##FIG##0##1##). These mRNA levels modestly declined at inductive sites 14 days after infection. CCR5, but not CCL5 or IFN-γ mRNA expression by fallopian tube-, uterus-, and cervix- derived lymphocytes were considerably higher than levels before infection. Indeed, CCR5 mRNA expression by fallopian tube lymphocytes was significantly higher 7 and 14 days post infection. These data suggest that increases in CCL5, CCR5, and IFN-γ mRNA expression during early stages of infection at inductive sites (i.e., spleen and ILN) preceded CCR5 expression at effectors sites (i.e., fallopian tube(s), uterus, and cervix). The pattern of CCL5 and CCR5 as well as IFN-γ illustrates how this chemokine axis coincides with both the innate (recognition phase, 0 to 7 days) and adaptive (activation, effector, and decline/homeostasis phases) immune responses to this pathogen and associated inflammation.</p>", "<title>Proliferative responses of <italic>C. muridarum </italic>genital infection modulated by CCL5 inhibition</title>", "<p>We next characterized <italic>C. muridarum</italic>-specific proliferative responses of T helper cells isolated from the spleen, ILNs, fallopian tube(s), uterus, and cervix, 42 days after challenge. CD4⁺T cells isolated from these mucosal and systemic immune compartments of <italic>C. muridarum</italic>-infected and control Ab-treated mice exhibited marked increases in Ag-specific proliferative responses compared to disease-free (uninfected) or <italic>C. muridarum</italic>-infected and anti-CCL5 Ab-treated mice (Figure ##FIG##1##2##). Notably, systemic and mucosal inductive sites of the spleen and ILN, respectively, contained Ag-specific CD4⁺T cells that significantly proliferated after <italic>C. muridarum </italic>re-stimulation. These results suggest that CCL5 is required for optimal generation of <italic>Chlamydia</italic>-specific CD4⁺T cells that proliferate after Ag recognition.</p>", "<title>CCL5 modulation of Chlamydia-specific humoral responses</title>", "<p>To test the role of CCL5 in Th1-biased humoral responses to genital infection, we measured <italic>Chlamydia</italic>-specific IgG1, IgG2a, IgG2b, IgG3 and IgM Abs in sera as well as Ag-specific IgA and IgG Abs in vaginal washes. Infected mice that received control Ab displayed significantly higher levels of <italic>Chlamydia</italic>-specific serum IgG2a, followed by IgG2b responses than compared to similar mice that received anti-CCL5 Ab or mock-infected (i.e., naïve) mice (Figure ##FIG##2##3##). Analysis of vaginal secretions revealed a significant increase in Ag-specific IgA Ab responses in <italic>C. muridarum</italic>-challenged mice that received control Ab when compared with similar mice that received anti-CCL5 Ab or uninfected animals. These results indicate <italic>C. muridarum </italic>infection enhanced Th1-biased humoral responses and CCL5 blockade attenuated <italic>Chlamydia</italic>-specific IgG2a serum responses as well as vaginal IgA responses.</p>", "<title>T helper cytokine responses and <italic>C. muridarum </italic>shedding</title>", "<p>Genital <italic>Chlamydia </italic>infection up-regulated Ab responses as well as splenic and ILN CD4⁺T cell proliferative responses. We next examined whether these effects were mediated in part through T helper cytokine responses. IL-2 secreted by Ag-stimulated splenic CD4⁺T cells from infected mice that received control Ab was significantly higher than levels from cells isolated from similar mice that received anti-CCL5 Ab or uninfected mice (Figure ##FIG##3##4##). The secretion of IFN-γ from splenic, but not ILN, CD4⁺T cells from <italic>C. muridarum</italic>-infected and control Ab-treated were considerably higher than T helper cells isolated from similar mice treated with anti-CCL5 Ab or uninfected animals. Chlamydial infection of mice lead to the development of CD4⁺T cells that significantly secreted IL-6, IL-10, and GM-CSF, but not IL-4, in response to <italic>C. muridarum </italic>re-stimulation. Anti-CCL5 Ab-treated mice infected with <italic>C. muridarum </italic>resulted in T helper cells with reduced IL-6, IL-10, and GM-CSF secretion following Ag re-stimulation when compared to CD4⁺T cells from similar mice treated with control Ab; however, the secretion pattern of Th2 cytokine responses by CD4⁺T cells isolated from ILNs did not significantly change. These results indicate that <italic>C. muridarum </italic>infection invoked splenic, but not ILNs, <italic>Chlamydia-</italic>specific CD4⁺T cells that secreted Th1 cytokines as well as IL-6, IL-10, and GM-CSF, which were reduced by CCL5 blockade. The extent of chlamydial infection, determined by C. muridarum detected in cervico-vaginal swabs, was significantly higher in mice receiving anti-CCL5 Ab treatment than compared to control mice (Figure ##FIG##4##5##). However, all mice resolved chlamydial infection in &lt; 90 days.</p>" ]

[ "<title>Discussion</title>", "<p>Th1-mediated immune responses are essential to control the chlamydial infection [##REF##16118671##22##]. It has been shown in previous studies that CCL3, CCL4 and CCL5 enhance adaptive immunity through Th1 cytokine and co-stimulatory molecule modulation [##REF##11123289##10##,##REF##12393512##11##,##REF##8871660##15##,##REF##9419219##16##]. In this study, we demonstrate some of the cellular and molecular mechanisms of CCL5-mediated chlamydial immunity. Importantly, IFN-γ mRNA was not significantly elevated 14 days after <italic>C. muridarum </italic>infection. IL-12p40 mRNA expression coincided with CCL5, but typically preceded IFN-γ mRNA expression (data not shown). Indeed, early genital clearance of <italic>Chlamydia </italic>has been shown to occur in an IL-12-dependent and IFN-γ-independent fashion [##REF##9120292##24##]; so, further studies will be required to dissect the roles of IL-12 and CCL5 in <italic>Chlamydia </italic>clearance and immunity. The present study shows that 7 days after genital <italic>Chlamydia </italic>infection, CCL5, CCR5, and IFNγ mRNA levels were elevated in inductive sites, while CCR5 mRNA expression was higher in the fallopian tubes than in the uterus and cervix. To this end, similar yet differential chemokine expression patterns have been reported within anatomically distinct regions [##REF##11854242##25##].</p>", "<p>CD4⁺Th1 cells (and associated cytokines, chemokines, etc.) are critical elements in the immune response, stimulated by an ascending <italic>C. trachomatis </italic>infection in the female genital tract [##REF##9420616##26##]. Indeed, studies of chlamydial infection in knockout mice support the importance of class II MHC, CD4, IL-12, IFN-γ, and IFN-γ receptor for chlamydial immunity [##REF##18292563##27##, ####REF##17015458##28##, ##REF##9199462##29##, ##REF##9038313##30####9038313##30##]. Our data suggest that CCL5 interactions are comparably important and mediate the temporal recruitment and activation of T cells to mitigate chlamydial infection through protective mucosal adaptive immunity by enhancing Th1 <bold>&gt;&gt; </bold>Th2 humoral and cellular immune responses.</p>", "<p>CCL3 and CCL4 are also CCR5 ligands, but these chemokines were not remarkably elevated compared to CCL5, which was the most abundant CCR5 ligand expressed during the early stages (i.e., &lt; 7 days) of chlamydial infection. CCL5 is an important factor in the homing of lymphocytes that express CCR5, CCR4, CCR3 and CCR1 [##REF##1699135##31##, ####REF##1380064##32##, ##REF##7584498##33####7584498##33##]. It was also demonstrated from our earlier studies that CCL5 can increase the proliferation and activation of Ag stimulated T lymphocytes [##REF##11123289##10##]. These findings coincide with marked increases in CCL5 and IFN-γ production by CD4⁺T cells during chlamydial infection. It is also plausible that CCL5 blockade might reduce the ability of innate immune cells (e.g., NK cells, macrophages, etc.) to respond to <italic>C. muriduram </italic>challenge. Previously, we demonstrated that CCL5 inhibition decreased the numbers of NK1.1⁺and CD11b⁺leukocytes at mucosal effector sites following pneumococcal challenge [##REF##16455992##34##]. However, T cells were the primary lymphocyte-subtype that were reduced following CCL5 blockade during the hyper cellular response to <italic>Streptococcus pneumoniae </italic>in mice.</p>", "<p>Genital <italic>Chlamydia </italic>infection of mice enhanced Ag-specific Abs in serum and vaginal secretions as well as proliferative cytokine responses by CD4⁺T cells isolated from systemic and mucosal compartments. High levels of IgA in cervical secretions in infected women correlate with low numbers of <italic>C. trachomatis </italic>shedding and B cells are required to eliminate <italic>Chlamydia </italic>in a secondary infection [##UREF##1##35##]<bold>. </bold>We have previously shown that CCL5 induces Ag-specific titers of IgG2a, followed by IgG2b, IgG3, and IgG1 [##REF##11123289##10##], CCL3 and CCL4 enhance Ag-specific IgG1 and IgG2b responses [##REF##12393512##11##]. In this study, <italic>Chlamydia </italic>infection enhanced IgG2a Ab responses, which were reduced after CCL5 inhibition. This highlights the importance of this chemokine in Th1-associated Ab responses against <italic>Chlamydia</italic>.</p>", "<p>The precise cytokine signals required for S-IgA production are not completely understood and studies show both Th1-and Th2-type cell-derived cytokines are important for S-IgA production [##REF##3262647##36##, ####REF##2786548##37##, ##REF##8568254##38####8568254##38##]. We have previously shown that chemokines like XCL1 and CCL5 induce IgA production [##REF##9973465##9##,##REF##11123289##10##]. The sharp IgA Ab response generated by <italic>Chlamydia-</italic>infected mice also correlated with the predominant Th1 &gt;&gt; Th2 cytokine response induced by this infection, which were reduced by CCL5 blockade. These results suggest that CCL5 is required in part for optimal <italic>Chlamydia</italic>-specific IgA Ab responses during <italic>C. muridarum </italic>infection.</p>", "<p>Mucosal, but not serum, IgA Abs were selectively elevated after <italic>C. muridarum </italic>challenge, presumbably due to the compartmentalized common mucosal immunity system. In particular, IgA Ab-secreting cells predominantly populate the lamina propria of the mucosa. In contrast, <italic>Chlamydia</italic>-specific IgG Abs were not detected in vaginal secretions. This confirms other studies that show vaginal <italic>Chlamydia</italic>-specific IgG Abs are not present at high levels (relative to IgA) in vaginal washes after <italic>C. muridarum </italic>challenge [##REF##17639158##39##,##REF##16178770##40##]. However, intranasal, subcutaneous or transcutaneous immunization using mucosal adjuvants can induce vaginal wash Ag-specific IgG Ab titers that are comparable to those of IgA [##REF##9973465##9##,##REF##15494499##41##, ####REF##15474723##42##, ##REF##14742549##43##, ##REF##12615442##44##, ##REF##8525685##45####8525685##45##].</p>", "<p>IFN-γ production is often associated with IgG2a production [##UREF##2##46##] and may account for the systemic humoral responses against <italic>Chlamydia </italic>following vaginal challenge. In the present study, <italic>Chlamydia </italic>infection induced a profound IgG2a Ab response compared to other IgG subclass Abs. The analysis of mucosal and systemic responses revealed a bias toward Th1 &gt;&gt; Th2 type of responses. While little is known regarding Ag-specific CD4⁺T cell (IL-4 and GM-CSF) secretory responses during <italic>Chlamydia </italic>infection, we show that these T helper cytokines were increased during chlamydial infection. CCL5 blockade diminished these cellular responses along with <italic>C. muridarum</italic>-specific CD4⁺T cell secretion of IFN-γ and IL-2.</p>", "<p>Contradictory studies demonstrated the ability of CCL5 to promote both Th1-and Th2-type responses. It was reported that anti-CCL5 Ab treatment in mice decreased mycobacterial-inducible Th2-type lesions while increasing schistosomal-inducible Th2-type granulomas [##REF##10384113##47##]. However, studies from our laboratory suggest CCL5 enhances mucosal and systemic humoral responses through help provided by Th1-type cytokines and select Th2-type cytokines, with CCL5 promoting Th1 and Th2 responses [##REF##11123289##10##]. The results in the present study suggest that CCL5 is required for optimal Th1 cellular responses against and clearance of <italic>Chlamydia</italic>. These results corroborate our previous findings that CCR5-dependent mucosal immune responses are required for the efficient clearance of genital chlamydial infection, while functional CCR5 expression reduces infertility as a pathologic consequence of Th1-mediated inflammation associated with infection [##REF##16118671##22##].</p>", "<p>The <italic>ccl5 </italic>gene has a number of physiologically relevant single nucleotide polymorphisms that affect its function and expression. In particular, the In1.1T/C haplotype of <italic>ccl5 </italic>is highly prevalent (~37%) in African Americans, than compared to Americans of European origin (~0.3%), and results in significantly lower CCL5 expression [##REF##12114533##48##]. This has been attributed to some of the health disparities between these ethic/racial groups; specifically, higher human immunodeficiency virus (HIV) susceptibility and faster progression to acquired immune deficiency syndrome (AIDS). Similar to the <italic>ccr5Δ</italic>32 polymorphism, the data in this study and others suggest diminished CCL5-CCR5 interactions could yield reduced fallopian tube scarring, but could also result in higher transmission/shedding of <italic>Chlamydia </italic>via decreased chlamydial clearance. It is plausible that a host with a modestly compromised innate immune system would on one hand – avoid infertility by mounting reduced Th1 responses and on the other – be unable to optimally clear a chlamydial infection. This would provide an evolutionary rationale for not only <italic>ccr5Δ</italic>32 and <italic>In1.1T/C ccl5 </italic>polymorphisms, but also the propagation and transmission of <italic>Chlamydia</italic>. No doubt, demonstration of this postulate will require additional and extensive studies.</p>" ]

[ "<title>Conclusion</title>", "<p>Understanding the cellular and molecular mechanisms that CCL5 uses to modulate mucosal immunity is essential to better understanding the pathogenesis of chlamydial infection. We tested the hypothesis that CCL5 is essential for inducing adaptive mucosal immunity against <italic>Chlamydia</italic>. We conclude that CCL5 supports the induction of Th1 cytokine and IgG2a Ab as well as IgA Ab responses against <italic>Chlamydia</italic>. The suppression of CCL5 correlated with delayed clearance of <italic>C. muriduram </italic>infection, which suggests chlamydial immunity is mediated by Th1 immune responses driven in part by CCL5.</p>" ]

[ "<p>This is an Open Access article distributed under the terms of the Creative Commons Attribution License (<ext-link ext-link-type=\"uri\" xlink:href=\"http://creativecommons.org/licenses/by/2.0\"/>), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.</p>", "<title>Background</title>", "<p>Following genital chlamydial infection, an early T helper type 1 (Th1)-associated immune response precedes the activation and recruitment of specific Th1 cells bearing distinct chemokine receptors, subsequently leading to the clearance of <italic>Chlamydia</italic>. We have shown that CCR5, a receptor for CCL5, is crucial for protective chlamydial immunity. Our laboratory and others have also demonstrated that CCL5 deficiencies found in man and animals can increase the susceptibility and progression of infectious diseases by modulating mucosal immunity. These findings suggest the CCR5-CCL5 axis is necessary for optimal chlamydial immunity. We hypothesized CCL5 is required for protective humoral and cellular immunity against <italic>Chlamydia</italic>.</p>", "<title>Results</title>", "<p>The present study revealed that CCR5 and CCL5 mRNAs are elevated in the spleen, iliac lymph nodes (ILNs), and genital mucosa following <italic>Chlamydia muriduram </italic>challenge. Antibody (Ab)-mediated inhibition of CCL5 during genital chlamydial infection suppressed humoral and Th1 &gt; Th2 cellular responses by splenic-, ILN-, and genital mucosa-derived lymphocytes. Antigen (Ag)-specific proliferative responses of CD4⁺T cells from spleen, ILNs, and genital organs also declined after CCL5 inhibition.</p>", "<title>Conclusion</title>", "<p>The suppression of these responses correlated with delayed clearance of <italic>C. muriduram</italic>, which indicate chlamydial immunity is mediated by Th1 immune responses driven in part by CCL5. Taken together with other studies, the data show that CCL5 mediates the temporal recruitment and activation of leukocytes to mitigate chlamydial infection through enhancing adaptive mucosal humoral and cellular immunity.</p>" ]

[ "<title>Abbreviations</title>", "<p>Ab: antibody; ABTS: 2,2'-azino-bis(3)-ethylbenzthiazoline-6-sulfonic acid; Ag: antigen; AIDS: acquired immunodeficiency syndrome; BD: below detection; BrdU: bromodeoxyuridine; DAB: diaminobenzedine tetrahydrochloride; DC: dendritic cell; EB: elementary body; FCS: fetal calf serum; FITC: fluorescein isothiocyanate; GMCSF: granulocyte monocyte cell stimulating factor; HIV: human immunodeficiency virus; HRP: horse radish peroxidase; IB: inclusion body; IFN: interferon; IFU/ml: inclusion-forming units per milliliter; Ig: immunoglobulin; IL: interleukin; ILN: iliac lymph node; MHC: major histocompatibility complex; MIT: Massachusetts Institute of Technology; NCBI: National Center for Biotechnology Information; OD: optical density; O/N: over night; PBS: phosphate buffered saline; PBS-T: PBS Tween; PE: phycoerytherin; RPMI: Roswell Park Memorial Institute; RT: room temperature; RT-PCR: real-time polymerase chain reaction; SEM: Standard Error of Mean; TBS: tris-buffered saline; Th1: T helper type 1; TNF: tumor necrosis factor; UV: ultra violet.</p>", "<title>Authors' contributions</title>", "<p>SKS and UPS carried-out all animal studies. SKS quantified serum and vaginal wash Ab levels as well as <italic>C. muriduram </italic>shedding. UPS isolated and measured mRNA levels as well as T cell cytokine secretion. DDT provided CCL5 and anti-CCL5 Ab as well as helped to draft the manuscript. JUI provided <italic>C. muriduram </italic>EBs and determined the corresponding titer. JWL as well as JUI conceived the study, participated in its design with all authors, coordinated and helped to draft the manuscript with the assistance of all authors. All authors read and approved the final manuscript. The authors declare that they have no competing interests.</p>" ]

[ "<title>Acknowledgements</title>", "<p>This work was supported in part by National Institute of Health grants RR03034, GM08248, MD000525, AI41231 and AI57808 and the Smith &amp; Lucille Gibson Endowment in Medicine. This research was also supported in part by the Intramural Research Program of the National Institute on Aging, National Institutes of Health. The content of this manuscript benefited from many fruitful conversations with colleagues at Morehouse School of Medicine, Centers for Disease Control &amp; Prevention, and University of Louisville as well as editing by Andrew Marsh.</p>" ]

[ "<fig position=\"float\" id=\"F1\"><label>Figure 1</label><caption><p><bold>CCL5, CCR5, and IFN-γ mRNA expression during <italic>Chlamydia </italic>infections</bold>. Groups of female BALB/c mice challenged with <italic>C. muridarum </italic>and total RNA was isolated from lymphocytes isolated from the spleen, ILNs, fallopian tubes, uterus, and cervix of each mouse, under sterile conditions before or 7 and 14, 21 and 42 days after challenge. The levels of CCL5, CCR5 and IFN-γ mRNA expression were ascertained after RT-PCR analysis. Log₁₀copies of transcripts were expressed relative to the actual copies of 18S rRNA ± SEM. Experiments were repeated 3 times to yield 15 mice per group. Asterisks (*) indicate significant differences (<italic>p </italic>&lt; 0.01) between tissues from naïve mice.</p></caption></fig>", "<fig position=\"float\" id=\"F2\"><label>Figure 2</label><caption><p><bold>Chlamydia-specific CD4⁺T cell proliferation</bold>. Groups of naïve or female BALB/c mice challenged with <italic>C. muridarum </italic>and received 100 μl of control Ab or anti-CCL5 Ab solution every 3 days. Following sacrifice 42 days after challenge, spleen-, ILN-, fallopian tubes-, uterus-, and cervix-derived CD4⁺T cells were purified and cultured at a density of 5 × 10⁶cells/ml with 10⁶cells/ml of γ-irradiated feeder splenocytes for 3 days. Proliferation of CD4⁺T cells was measured by BrdU incorporation. The data presented are the mean OD₄₅₀for proliferative ± SEM of quadruplicate cultures. Experiments were repeated 3 times to yield 15 mice per group. Asterisks (*) indicate statistically significant differences (<italic>p </italic>&lt; 0.01) between untreated, control Ab-treated mice and mice treated with anti-CCL5 Ab.</p></caption></fig>", "<fig position=\"float\" id=\"F3\"><label>Figure 3</label><caption><p><bold>Serum and vaginal Ab responses of mice infected with genital <italic>Chlamydia</italic></bold>. Groups of naïve or female BALB/c mice challenged with <italic>C. muridarum </italic>and received 100 μl of control Ab or anti-CCL5 Ab solution every 3 days. Following sacrifice 42 days after challenge, <italic>Chlamydia</italic>-specific serum and vaginal Ab responses of mice were determined by ELISA that was capable of detecting &gt; 20 pg/ml of IgM, IgG, IgA, and IgG subclass Abs. The data presented are the mean concentration of IgG1, IgG2a, IgG2b, IgG3, IgM, IgG, or IgA ± SEM of three separate experiments. Experiments were repeated 3 times to yield 15 mice per group. Asterisks (*) indicate statistically significant differences (<italic>p </italic>&lt; 0.01) between untreated and control Ab- or anti-CCL5 Ab-treated mice.</p></caption></fig>", "<fig position=\"float\" id=\"F4\"><label>Figure 4</label><caption><p><bold>T helper cytokine secretion by CD4⁺T cells from <italic>Chlamydia</italic>-infected mice</bold>. Groups of naïve or female BALB/c mice challenged with <italic>C. muridarum </italic>and received 100 μl of control Ab or anti-CCL5 Ab solution every 3 days. Following sacrifice 42 days after challenge, spleen- and ILN-derived CD4⁺T cells from these mice were purified and cultured at a density of 5 × 10⁶cells/ml with 10⁶cells/ml of γ-irradiated feeder splenocytes for 3 days. Cytokines present in cultured supernatants were determined by ELISA that was capable of detecting &lt; 10 pg of IL-2, IFN-γ, IL-4, IL-6, IL-10, or GM-CSF. The data presented are the mean cytokine (pg/ml) ± SEM of quadruplicate cultures. Asterisks (*) indicate statistically significant differences (<italic>p </italic>&lt; 0.01) between untreated and control Ab- or anti-CCL5 Ab-treated mice.</p></caption></fig>", "<fig position=\"float\" id=\"F5\"><label>Figure 5</label><caption><p><bold>Changes in <italic>Chlamydia</italic>-shedding</bold>. Mice treated with control Ab or anti-CCL5 Ab were infected with <italic>C. muridarum</italic>. The status of infection was monitored for live organism shedding by culturing periodic cervico-vaginal swabs from individual mice every week for 42 days after challenge. <italic>Chlamydia </italic>present in swabs were detected by infecting McCoy cells with vaginal swab rinses and staining infected monolayer with FITC-labeled, genus-specific, anti-<italic>Chlamydia </italic>Abs to count inclusion bodies (&gt; 10 IFUs/ml) by direct immuno-fluorescence. BD indicates the presence of <italic>Chlamydiae </italic>was below this level of detection. Experiments were repeated 3 times to yield 15 mice per group. Asterisks (*) indicate statistically significant differences (<italic>p </italic>&lt; 0.01) between untreated and control Ab- or anti-CCL5 Ab-treated mice.</p></caption></fig>" ]

[]

[]

[]

[]

[]

[]

[]

[]

[ "<graphic xlink:href=\"1471-2180-8-136-1\"/>", "<graphic xlink:href=\"1471-2180-8-136-2\"/>", "<graphic xlink:href=\"1471-2180-8-136-3\"/>", "<graphic xlink:href=\"1471-2180-8-136-4\"/>", "<graphic xlink:href=\"1471-2180-8-136-5\"/>" ]

[]

{ "acronym": [], "definition": [] }

[ "<title>Background</title>", "<p>Malaria parasites, particularly <italic>Plasmodium falciparum</italic>, impose heavy economic and health burdens on human population worldwide [##REF##15759000##1##]. Hundreds of millions of people are infected by the parasite each year, leading to 1–2 million deaths annually. Lack of effective vaccines and emergence of drug-resistant parasites and insecticide-resistant mosquito vectors are the main reasons for the failure in controlling the parasites and the associated disease. A better understanding of the molecular mechanisms of drug resistance, the molecular basis of the host immune response, and the strategies the parasite employs to evade host immunity is critical for vaccine and drug development.</p>", "<p>Genetic variation in parasites can contribute to drug resistance, immune evasion, and disease manifestation. Genetic mapping is one of the powerful approaches for the identification of mutations that cause drug resistance and changes in other phenotypes [##REF##17572690##2##]. For efficient mapping of a target gene, it is often necessary to genotype a large number of polymorphic markers. In addition to length polymorphisms such as microsatellites and minisatellites and large-scale sequencing, genome-wide single nucleotide polymorphisms (SNP) have been identified from many organisms, including <italic>P. falciparum</italic>, for genotyping and mapping genes associated with different phenotypes [##REF##17159978##3##, ####REF##17159981##4##, ##REF##17159979##5####17159979##5##]. High-throughput SNP typing methods have also been developed [##REF##10742102##6##, ####REF##12960966##7##, ##REF##15838508##8##, ##REF##15687290##9##, ##REF##17225249##10##, ##REF##17335059##11####17335059##11##], leading to recent successful identification of candidate genes (loci) associated with various human diseases [##REF##17632545##12##, ####REF##17293876##13##, ##REF##17554300##14##, ##REF##17463248##15##, ##REF##17463246##16##, ##REF##17637780##17##, ##REF##17632509##18##, ##REF##17618284##19##, ##REF##17618283##20####17618283##20##].</p>", "<p>One of the high-throughput typing methods is array-based hybridization. In this method, labeled genomic DNA is hybridized to microarrays comprising high-density short oligonucleotides designed based on known SNP or systematically tiled along all chromosomes to detect potential polymorphisms. High-density arrays have been successfully used to detect variation in copy number [##REF##14762065##21##, ####REF##16504045##22##, ##REF##17122084##23####17122084##23##] and SNP [##REF##17062589##24##,##REF##17189563##25##]. The human malaria parasite <italic>P. falciparum </italic>has a genome with extremely high AT content (&gt; 80%) as well as numerous repetitive sequences [##REF##12368864##26##], making array design and data analysis challenging. Hybridizations of <italic>P. falciparum </italic>genomic DNA to both Affymetrix GeneChips^®and slides printed with 70 mer oligonucleotides have been reported previously [##REF##16174539##27##, ####REF##16789840##28##, ##REF##17599553##29####17599553##29##]. Kidgell <italic>et al</italic>. recently used an array with 327,782 probes to identify 23,653 single feature polymorphisms (SFP) among 14 isolates. The results from this study suggest that high-density array could be a promising tool for high-throughput detection of genome variations including SNP and copy number variations (CNV). However, calling SNP based on hybridization signals is a complex process, and many factors can affect SNP calling, including array design, GC content of a probe, the position of the SNP in a probe, hybridization conditions, and algorithms used to analyze array signals. Additionally, methods were developed to call SFP in many previous studies, but the accuracy of SFP calls were not verified with known SNP or through DNA sequencing. To investigate the influences of these factors on calling SFP in a highly AT-rich genome and to develop a reliable method for calling SFP from the <italic>P. falciparum </italic>genome using commercially available array platforms, we have analyzed data from a high-density 'tiling' array with ~2.5 million 25 mer probes designed at The Sanger Institute (PFSANGER GeneChips^®) to detect genomic variations in five <italic>P. falciparum </italic>field isolates. Genomic DNA samples from the five parasite isolates were hybridized to the array, and signals from the parasites were compared with known SNP [##REF##17159981##4##] to evaluate SNP calling accuracy under different conditions. Based on the comparison, we identified factors that could affect probe/DNA hybridization dynamics and established a set of conditions that allowed us to call SFP/SNP with ≥ 94% accuracy. We also sequenced 52 SFP calls that did not agree with known SNP and found that ~64% of the 'wrong' calls were actually due to errors in the genome sequences. Parameters that provided best SNP calling accuracy were used to identify 121,087 potential SNP, including ~18,000 new SFP that have not been reported previously.</p>" ]

[ "<title>Methods</title>", "<title>Parasites and parasite culture</title>", "<p><italic>P. falciparum </italic>parasite isolates used in this study have been described [##REF##17159981##4##,##REF##12890022##43##]. The parasites were cultured <italic>in vitro </italic>according to the methods of Trager and Jensen [##REF##781840##44##]. Briefly, parasites were maintained in RPMI 1640 medium containing 5% human O⁺erythrocytes (5% hematocrit), 0.5% Albumax (GIBCO, Life Technologies, Grand Island, NY), 24 mM sodium bicarbonate, and 10 μg/ml gentamycin at 37°C with 5% CO₂, 5% O₂, and 90% N₂.</p>", "<title>DNA extraction and probe labeling</title>", "<p>Parasites were cultured to a parasitemia of 5% or higher; and the cultures were centrifuged at 5000g to collect red blood cells that were lyzed with addition of 10 vol of 0.1% saponin in PBS. The parasites were centrifuged again; and genomic DNA was extracted from the parasite pellet using Wizard Genomic DNA Purification kit (Promega, Madison, WI). Genomic DNA (10 μg) from each parasite was used as probes in the hybridizations. Briefly, genomic DNA was fragmented to an average size of 50–150 bp with DNase I and the quality of the digested DNA evaluated in 2% agarose gels. Subsequently, fragmented DNA was end-labeled using terminal deoxynucleotidyl transferase and a biotin labeling kit (Affymetrix mapping 250 K reagent kit; Affymetrix, Inc., Santa Clara, CA).</p>", "<title>Microarray hybridization</title>", "<p>The PFSANGER Genechip^®was purchased from Affymetrix, Inc. Array hybridization was performed at the microarray facility of the Laboratory of Immunopathogenesis and Bioinformatics, SAIC-Frederick, Inc (Frederick, MD). Briefly, biotin-labeled DNA were hybridized to array chips at 45°C for 16 h with constant rotation at 60 rpm. Affymetrix 20× hybridization control was used to make the hybridization cocktail. Hybridized chips were washed and stained following the company's EukGE-WS2v5 protocol. The chips were then scanned at 570 nm emission wavelength using an Affymetrix scanner 3000. All the parasites have two or more biological replicates (Additional file ##SUPPL##0##1##).</p>", "<title>Microarray chip design and data analysis</title>", "<p>The probes were designed based on <italic>P. falciparum </italic>genome (3D7) sequence v2.1.1 [##UREF##1##45##] covering genomic regions where unique probes with a reasonably broad 'thermal' range could be designed. A brief description of the array design has been reported recently [##REF##18096748##46##]. Because of recent updates of genome databases, all probe sequences were reassigned with new coordinates along each chromosome and their relative positions in a predicted gene (exon, intron, across exon and intron, and intergenic regions) according to the 3D7 genome sequence in PlasmoDB V5.2. The scanned image CEL files were processed and analyzed using the R/Bioconductor package and the robust multichip analysis method [##REF##11017085##47##]. Basically, the programs retrieved probe information (perfect match only), performed background subtraction, quantile-normalized signals from the chips, and transformed the data into a final normalized data matrix of log2 values. Partek Genomics Suite 6.3 (Partek Inc., St. Louis, MO) and in-house programs are also used in SFP calling and copy number analyses.</p>", "<title>Mapping known SNP to array probes</title>", "<p>After determining the correct genomic coordinates for each SNP and each array probe, known SNP from our previous study [##REF##17159981##4##] and those in PlasmoDB [##REF##17159978##3##,##REF##17159979##5##,##REF##16789840##28##,##UREF##1##45##] were mapped to probes that covered known SNP positions. Ambiguous SNP (mapped to multiple positions) were removed, and the remaining SNP were uploaded to a genome browser [##UREF##0##32##] with allele information from different parasites.</p>", "<title>SFP calling</title>", "<p>Because the signals from the probes do not allow for accurate mapping of the position of a SNP within a probe at the given probe density, we can only assert that somewhere within a probe there is likely a polymorphism. Therefore, we simply assigned the polymorphism to a feature (probe) and called it a single feature polymorphism (SFP) as described [##REF##16789840##28##]. Because a polymorphic site was often covered by multiple probes (average ~4 probes), we treated calls from probes within 25 bp as one SFP (called mSFP). To establish optimal parameters for SFP calling, we investigated SFP calling rates and calling accuracies using various conditions. We first identified all of the probes that covered each SNP identified in our previous study [##REF##17159981##4##]. Then we extracted their hybridization signals from a normalized data file. The average probe intensity (average of antilogs of the raw data) from the normalized data for all replicates of each parasite isolate was calculated. This value was compared with the average signal for 3D7 obtained in the same way. A ratio was obtained after comparison with the signal from that of 3D7. We evaluated the influences of SNP position in a probe, GC content of a probe, cutoff ratios of hybridization signal, and numbers of probes on SFP calling accuracy. Probes with GC content &lt; 16% and &gt; 50%, and probes with multiple hits in the genome were excluded for the analyses. The last two nucleotides at each ends of a probe were also discarded, because substitutions at these positions had minimal influences on hybridization signals.</p>", "<p>Once optimal parameters were identified for calling SNP using the NIAID SNP as an input set to test the method, we applied a similar procedure to a whole genome scan for probe-based SFP and mSFP (Additional file ##SUPPL##8##9##). Probe ratios were computed for each parasite for each probe, and raw alleles were generated by applying the cutoff ratio of 5.0 – it was an SFP if a ratio was above the cutoff value and it was not if below the ratio. Next, going through one parasite at a time, all probes were considered where there was more than one positive probe in a row within 25 bp of one another. Once this filtered set of probes was extracted from the full set, the ratios of intensity for each of the isolates compared with 3D7 was computed and tabulated. From this table, a vector was constructed for each parasite isolate where either a '1' or a '0' was added to each position determined by the value of the ratio. This vector was then scanned for stretches of '1's where the distance between the probes was less than 25 bp. In cases where longer stretches were identified, they were output as an additional feature type called long multiprobe polymorphism. Because some probes represent different strands of the exact same sequence region, we also discarded those stretches of '1's where the probes on either strand had a distance of 0 bp from the neighboring probe but did not exceed the threshold ratio value. All of the multiprobe polymorphisms corresponding to the mSFP were then output, and both classes of polymorphisms (single probe SFP and multi-probe mSFP) were then loaded into the genome browser. The procedure also tracks the 'alleles' by parasite isolate to determine the counts of mSFP shared by each possible combination of parasite isolates. Additional parameters that added confidence to a particular mSFP call, such as multiple parasite isolates having the same SFP and matches to known SNP in PlasmoDB, were also indicated.</p>", "<title>Estimating SFP calling rates using ROC curve and Z-score</title>", "<p>Hybridization measurements from Affymetrix CEL files were pre-processed in the R programming environment [##UREF##2##48##] using the read.affybatch function from the affy BioConductor package [##REF##14960456##49##]. Background adjustment was performed using the method developed for the RMA algorithm, and normalization was done using the quantile method. Differential hybridization between parasite isolates was expressed as Z-scores calculated by the LPE package [##REF##14555628##30##,##UREF##3##50##].</p>", "<title>DNA sequencing</title>", "<p>To verify selected mSNP (Table ##TAB##1##2##) that might be called incorrectly or calls that had contradictory signals, we amplified DNA fragments of 200–500 bp containing the probes and sequenced the PCR products directly according to methods described [##REF##12890022##43##]. Primer sequences used in PCR and DNA sequencing are listed in Table ##TAB##1##2##.</p>", "<title>Detection of CNV</title>", "<p>To detect CNV, we imported the filtered probe data into Partek Genomics Suite v6.3 and normalized individual probe signal from the 3D7 reference genome to 1.0 (haploid genome). Basically, the genomic segmentation algorithm finds a segment according to three criteria: 1) neighboring regions have statistically significantly different average intensities (<italic>P </italic>≥ 0.00001); 2) breakpoints (region boundaries) were chosen to give optimal statistical significance (smallest <italic>P</italic>-value); and 3) detected regions must contain a minimum of 15 probes. After determining the segments that had average signals higher or lower than 1.5 fold of those of the 3D7 reference, we filtered out regions that were less than 300 bp long. Detected segments, representing potential deletions or highly polymorphic regions, were plotted along chromosomes to produce CN genome view (Figure ##FIG##4##5##); and the segments were mapped to predicted genes in PlasmoDB to generate additional file ##SUPPL##8##9##. To screen for those highly polymorphic genes from potentially deleted segments, we flagged segments containing <italic>var/rif/stevor </italic>and other multigene families.</p>" ]

[ "<title>Results</title>", "<title>Basic probe statistics and quality control</title>", "<p>The array has 2.56 million perfect-matched probes (25 mer) with 2,206,371 <italic>P. falciparum</italic>-specific probes (the rest of the probes were for rodent malaria parasites). Of the <italic>P. falciparum </italic>probes, 2,107,319 mapped uniquely to the genome and 99,052 mapped to more than one location or were not assigned to any chromosomes. Among the unique probes, 1,446,824 were in the predicted coding regions (CDS); 1,304,180 probes were within exons; 727,200 probes were intergenic; 84,622 were within introns; 58,022 probes spanned exon/intron junctions, and 32,347 probes spanned the predicted translation start sites or stop codons.</p>", "<p>Genomic DNA from five different parasites (Additional file ##SUPPL##0##1##) were labeled and hybridized (2–4 replicates) to the PFSANGER GeneChip^®. After normalization of the hybridization signals across all array chips, an average signal intensity for each probe was calculated from replicates of each parasite. The qualities of the hybridizations were evaluated using various methods including MA plots, scatter plots (data not shown), and coefficient of variance (CV) tests (Additional file ##SUPPL##0##1##). Good reproducibility was obtained among replicates with the majority of the probes (&gt; 90%) having CV less than 25% (Additional file ##SUPPL##0##1##). Histograms of signal ratios relative to 3D7, the reference genome, showed similar data distribution among different parasite samples (Additional file ##SUPPL##1##2##).</p>", "<title>Probe coverage of known SNP</title>", "<p>Accurate SNP calling and detection of insertions/deletions requires optimization of calling parameters. Here we evaluated potential factors that might affect SFP calling accuracy by comparing known SNP between 3D7 and four other parasites (Dd2, HB3, 7G8, and FCR3) identified in our previous study (<italic>i.e</italic>., NIAID SNP) [##REF##17159981##4##] and hybridization signal ratios. Among the 3,836 NIAID SNP (excluding 82 that were mapped to multiple sites) identified previously, 2,651 (69%) were covered by 10,841 probes, including 1,787 covered by 5,600 probes in the predicted exons. The majority of the SNP were covered by 1–5 probes (average 4.4 probes/SNP), with a maximum coverage of 45 probes/SNP (Additional file ##SUPPL##2##3##). Overall, the SNP were distributed evenly across the 25 mer positions in the probe, with ~94% of probes having one SNP (Additional file ##SUPPL##3##4##).</p>", "<title>Probe GC content and hybridization intensity</title>", "<p>Because GC content in a probe is known to affect probe/DNA hybridization dynamics, we investigated the influence of probe GC content on hybridization signal intensity. The GC effect is likely exaggerated even more for the AT-rich genome of <italic>P. falciparum </italic>genome. The majority of the probes in the array have GC contents of 15% to 40% (Figure ##FIG##0##1A##). Signal intensity was similarly low for probes with GC content &lt;16%, but for probes with GC content of 16% or higher, signal intensity increased with the increase of GC content until ~40%, when signal intensity began to plateau (Figure ##FIG##0##1B##). Signal intensity did not change much from 40% to 80% GC in 3D7; however, the intensity began to decrease and fluctuate dramatically after reaching 50% GC content in non-3D7 parasites (Figure ##FIG##0##1C##). Reduction in signal intensity in non-3D7 parasites suggested high levels of polymorphism in these probes. In the parasite genome, the first exons of the <italic>var </italic>gene family have a relatively high GC content and are highly variable in DNA sequence. These high-GC-content probes are therefore likely from the <italic>var </italic>genes. Comparison of the high-GC probes with <italic>var </italic>gene sequences showed that ~44% of the 5,491 probes with 50% or higher GC content were from the <italic>var </italic>genes. These probes likely contributed to the dramatic variation in signal ratio between parasites (Figure ##FIG##0##1D##). These results suggest that probes with GC content &lt;16% and the <italic>var </italic>probes with &gt;50% might not be reliable for the detection of SFP for genetic mapping of the <italic>P. falciparum </italic>traits.</p>", "<title>Substitution positions in a probe and hybridization dynamics</title>", "<p>The position of a nucleotide substitution in a probe can also influence probe hybridization intensity. A substitution in the middle of a probe is expected to affect hybridization stability more dramatically than a change at the end positions of a probe. Comparison of average signal ratios between 3D7 and the other four parasites and SNP at known probe positions showed that substitutions at the two end positions (1 and 25) of a probe did not affect probe-target hybridization; and substitutions at position 2 and 24 had minimal effect on signal intensity (Figure ##FIG##1##2##). Signal ratios (3D7/7G8) of probes with SNP from position 3 to position 7 increased from both ends, averaging more than 10 times of the probes without polymorphism. For all positions in a probe, the average signal ratios were approximately the same (&lt; 1.5) if there was no known polymorphism in a probe. For probes that had known SNP, the signal ratio was generally 5 or higher if two positions at each end of a probe were excluded (Figure ##FIG##1##2##). Our data showed that substitutions located at probe position 3–23 (25 mer probes) had a strong effect on hybridization intensity and should be considered for SFP detection (Figure ##FIG##1##2##).</p>", "<title>Estimates of correct SFP call rates</title>", "<p>We next evaluated different signal cutoff ratios to obtain a value that produced the best SFP calling accuracy realizing that this ratio would balance false positive and false negative calling rates. We found that a signal cutoff ratio of 1.5 produced the highest overall correct call rates (≥ 90%) for Dd2, HB3, and 7G8 (Table ##TAB##0##1##). Correct call rates increased slightly after removing probes with high and low GC contents and increased further after excluding calls from single probes and calls with probe vote ratio &lt; 75%. In contrast, correct call rates decreased with the increase of signal ratio cutoff values, likely because of the exclusion of some real SFP with relatively lower signal ratios. Even using a signal cutoff ratio of 5.0, we obtained correct call rates ≥ 85%. After correcting for wrong calls due to sequence errors (see below), we obtained correct call rates ≥ 94% (Table ##TAB##0##1##). The call rate for FCR3 could not be estimated accurately without known SNP information.</p>", "<title>Sequencing verification of SFP calls</title>", "<p>Both false positive (Fp) and false negative (Fn) calls could be caused by SFP calling errors, sequencing mistakes, or problems in sequence alignment in the databases. To investigate whether the discrepancies between our SFP calls and the known SNP were from array SFP calling or sequencing/alignment errors, we sequenced 52 Fp or Fn SFP calls (positions 3–23, 1.5 cutoff ratio between 3D7 and 7G8) with different probe coverage and probe vote ratios to verify the calls. Our results showed that 33 of the 52 (63.5%) initial wrong calls were due to sequence errors in the databases, including four Fp calls that did not have polymorphism at the expected sites but had new polymorphic sites nearby, leading to the incorrect Fp calls (MAL14.5217, MAL12.3146, MAL11.3013, and PFC0210c in Table ##TAB##1##2##). Among the 19 true wrong-calls verified by sequencing, 9 were called by a single probe, 6 had mixed probes calls, 3 had two one-sided probe calls, and 1 had three one-sided probe calls. If we excluded calls from single probes and mixed probe calls having a probe vote ratio &lt;75% (for example, one probe suggested a SFP, but three others suggested no SFP), we would have had only four calls that were incorrect (7.7% of the 52). In other words, 92% (48/52) of the calls would have been correct if we had excluded single probe calls and calls with a probe call vote ratio of &lt;75%. If we apply these corrections, we obtain a corrected overall SFP call rate of ≥ 94% even using a conservative cutoff value of 5.0 (Table ##TAB##0##1##).</p>", "<title>Use of receiver operating characteristic (ROC) curves to estimate call rates</title>", "<p>To further test the reliability of our method in calling SFP, we also used a ROC curve to evaluate SFP calling accuracy and applied local pooled error (LPE) analysis to obtain Z-scores for calling SFP [##REF##14555628##30##]. LPE generates corrected Z-scores that reduce Fp, which might result when sample variance happens to be low, by using a 'pooled' variance for all the probes that show similar intensities. The ROC curve is a graphic plot of sensitivity <italic>vs</italic>. (1-specificty) or fraction of true positive <italic>vs</italic>. the fraction of Fp [##REF##8843379##31##]. As shown in Figure ##FIG##2##3##, if we allowed a Fp rate of approximately 2% (1-specificity), and at a Z-score of ~1.5, we could obtain a sensitivity of call rate ~81% genome-wide for data from 7G8, Dd2, and HB3.</p>", "<p>SFP were called using Z-scores of 1.5, 2.0, 3.0 and 4.0 and compared with SFP called using signal ratio cutoffs of 1.5, 2.0, 3.0, and 5.0. Results from cutoffs of Z-score of 3.0 and signal ratio of 3.0 had the best overall matches (~99%) and the best positive SFP call matches (~82%) for all 14 chromosomes. To minimize Fp calls (low Fp rate is important for genetic mapping) from unknown parasites that might have higher background, however, we decided to use a conservative signal ratio cutoff value of 5.0. Using this cutoff value, almost all (~98%) of the positive calls matched a positive call from a Z-score cutoff 3.0.</p>", "<title>Detection of genome-wide substitutions among field isolates</title>", "<p>We used a conservative signal cutoff ratio of 5.0 and all the parameters discussed above (Additional file ##SUPPL##4##5##) to call SFP and obtained 121,087 mSFP genome-wide among the five parasites, including 41,700 unique mSFP from 3D7, 8,856 from 7G8, 10,068 from Dd2, 10,449 from HB3, and 5,121 from FCR3 (Table ##TAB##2##3##). Inspection of the calls revealed that the large number of 3D7 unique calls was largely from multigene families such as var, rif, and stevor. We therefore flagged mSFP from multigene families (PFB0935w, PFD0090c, MAL7P1.6, MAL7P1.58, PFI1780w, PFA0655w, PFB0105c, MAL7P1.7, MAL7P1.59, PF10_0380, PFE1600w, PF10_0012, PF10_0005) and their paralogs. Excluding mSFP from these genes removed approximately 67% of the SFP and reduced the total number of mSFP to 40,354, including 6,618 unique mSFP for 3D7, 6,855 for HB3, 2,854 for FCR3, 7,173 for Dd2, and 6,342 for 7G8 (Additional file ##SUPPL##5##6##). A list of SFP and mSFP in each predicted gene and genes that are highly polymorphic (genes encoding potential antigens) can be found in Additional file ##SUPPL##6##7##.</p>", "<p>Some chromosomes appeared to have unusually large numbers of mSFP calls from some parasites. For example, Dd2 had 1636 unique mSFP from chromosome 2, whereas the other four parasites had fewer than 400 mSFP (Table ##TAB##2##3##). Close inspection of the calls revealed that the majority of the extra mSFP was from a deletion at one end of chromosome 2 in Dd2 (Additional files ##SUPPL##7##8## and ##SUPPL##8##9##). Similarly, the higher numbers of mSFP from chromosome 12, 13, and 14 of HB3 were from specific regions either deleted or having highly polymorphic genes in a specific parasite (Additional file ##SUPPL##7##8## and ##SUPPL##8##9##).</p>", "<title>Genome-wide mSFP distribution</title>", "<p>SFP and mSFP were uploaded into the GBrowse genome browser at the ABCC website [##UREF##0##32##] for genome-wide display of the polymorphic site. Probe sequences and locations in predicted exons, introns, and intergenic regions were mapped to chromosomes. SNP in the PlasmoDB and our SFP/mSFP calls were also displayed in the browser with allele information from each parasite. As shown in the browser, the majority of our mSFP (89%) matched well with the PlasmoDB SNP (estimated for 7G8 only), including SNP in the <italic>pfcrt </italic>(Figure ##FIG##3##4A##). This comparison identified ~18,000 new unique mSFP (excluding those from multi-gene families) from the five parasite genomes.</p>", "<p>We noticed that many of the PlasmoDB SNP (51.1%) were located on chromosomal regions that did not have probe coverage (Figure ##FIG##3##4##). Because the majority of the regions without probe coverage were likely in areas of AT-rich repetitive and/or noncoding sequences, the observation suggested that relatively larger numbers of SNP in the PlasmoDB could be from repetitive sequences.</p>", "<p>We next counted mSFP in a window of 10-kb segments and plotted mSFP from each segment along the chromosomes to investigate mSFP distribution on the chromosomes from each parasite (Additional file ##SUPPL##7##8##). Again, these plots showed clusters of some highly polymorphic regions, mostly at chromosome ends, corresponding to <italic>var/rif/stevor </italic>clusters. The plots also identified some unique peaks for individual parasite, for example, a unique peak on chromosome 2 for Dd2 and HB3, respectively. These unique peaks were likely due to deleted DNA segments or reflected the unique selection and evolutionary histories in an individual parasite (Additional file ##SUPPL##7##8##).</p>", "<title>Genome-wide CNV</title>", "<p>Genome-wide segmentation analyses showed that there were relatively few large-scale amplifications or deletions among the parasites (Figure ##FIG##4##5##). The 5 largest amplified regions were a ~28 kb on chromosome 4 of FCR3, a ~80–96 kb on chromosome 5 of Dd2 and FCR3, a ~30 kb on chromosome 9 of FCR3, a ~82.5 kb on chromosome 11 for HB3, and various sizes (~3–180 kb) in the middle of chromosome 12 for different parasites. The chromosome 5 amplified region contained a total of 20 unique genes, including 19 genes (PFE1065w-PFE1155c) amplified ~2–3 copies in FCR3 and 14 genes (PFE1095w-PFE1160w) amplified ~4–5 copies in Dd2 (Additional file ##SUPPL##8##9##) with a total of 13 genes shared by the two parasites. Eight of the shared genes were predicted to encode proteins related to ribosomal subunits, ATP-dependant helicase, nucleotide binding, s-adenosylmethionine-dependent methyltransferase, mitochondrial processing peptidase, G10, and multidrug resistance homolog protein, PfPgh-1. Similarly, segments of different sizes located at the middle of chromosome 12 were amplified ~7–8 copies in 7G8 (PFL1085w, PFL1125c-PFL1160c, ~67 kb), ~5 copies in Dd2 (PFL1085w, PFL1145w-PFL1150c, ~3 kb), ~3–4 copies in FCR3 (PFL1135c-PFL1160c, ~20kb), and ~2–3 copies in HB3 (PFL1085w, PFL1125w-PFL1310c, ~184kb). Only two genes (PFL1145w and PFL1150c) were amplified in all of the four parasites, one of which was a gene encoding putative ribosomal protein L24. A large region on chromosome 11 from HB3 containing 26 genes (PF11_0489 to PF11_0513) was amplified 2-3X, four of the genes were predicted to encode ring-infected erythrocyte surface antigen, antigen 332, and Ser/Thr protein kinase. The amplified region on chromosome 4 of FCR3 (~25 kb) contained genes encoding a putative reticulocyte-binding protein 1 and four hypothetical proteins (PFD0095c-PFD0115c) and was amplified at least five times. This amplified segment may play a role in the higher growth rate for this parasite, because the reticulocyte-binding protein may facilitate parasite invasion.</p>", "<p>The majority of the regions with reduced signals (blue) were located on chromosomes ends or regions containing the <italic>var/rif/stevor </italic>gene clusters, reflecting the highly variable nature of these DNA regions (Figure ##FIG##4##5##). Although it is difficult to distinguish highly polymorphic regions from deletions in this haploid genome, we considered several additional restrictions to exclude potential polymorphic loci. A segment was considered not truly deleted if it contained known highly polymorphic genes such as <italic>var/rif/stevor </italic>[##REF##17599553##29##] or if a segment had reduced signals in all four parasites (suggesting highly polymorphic genes such as genes encoding surface proteins). For segments with reduced signal ratios occurring only in one or two parasites, they were more likely to be true deletions, which could also be detected in mSFP distribution plots (Additional file ##SUPPL##7##8##). For example, a deletion of ~42-kb segment (PFB0070w-PFB0100c) on chromosome 2 of Dd2 and FCR3 was found to contain a gene encoding knob-associated histidine-rich protein (KAHRP). Deletion of KAHRP in Dd2 was reported previously [##REF##16789840##28##,##REF##17599553##29##,##REF##1922044##33##]. Another likely deleted segment was a ~98-kb region on chromosome 9 of HB3 containing 19 genes (PFI1710w-PFI1800w) including the gene encoding cytoadherence linked asexual protein (CLAG) and lysophospholipase. Again, deletion of this region had been reported [##REF##8367496##34##]. A list of chromosome segments and mapped genes potentially amplified or deleted/highly polymorphic, including those reported previously, can be found in Additional file ##SUPPL##8##9##.</p>" ]

[ "<title>Discussion</title>", "<p>The PFSANGER array, despite having ~2.2 million <italic>P. falciparum </italic>probes, was not designed specifically for SNP detection, and whether it was suitable for SNP detection was not certain. This study was initiated to investigate the possibility of using the PFSANGER array for genetic mapping and population studies. The large number of probes on the chip and their high AT content (some &gt; 80%) require critical evaluation of factors that may affect hybridization dynamics before SFP can be reliably called. Based on comparison of mSFP calls with known SNP identified previously [##REF##17159981##4##], we showed that the last two end positions in a probe had limited influence on hybridization signal and that probes with GC contents lower than 16% should be excluded for SFP calling in this genome. We also found that mSFP calls based on a single probe were not reliable after resequencing. For a potential mSFP call, a conservative signal cutoff ratio of 3–5.0 and a vote among several adjacent probes (within 25 bp) with a majority of the probes (at least 75%) should be applied. We demonstrated that this particular microarray could be successfully employed to detect mSFP with high mSFP calling accuracy (≥ 94%). This work provides important information for calling mSFP in the <italic>P. falicparum </italic>genome using microarrays.</p>", "<p>We used a 5.0 cutoff ratio in calling SFP because for genetic mapping, a high Fp rate may lead to misleading results that should be minimized. A higher cutoff value may result in a higher Fn rate or missing some calls too. Missing some calls will not be a big issue as the array can detect a large number of SFP. The 5.0 cutoff therefore represents a conservative value for minimizing Fp calls, considering potential higher backgrounds that may exist in some field isolates such as FCR3 in this study. Higher background in FCR3 requires further investigations, although signal intensity and distribution from this parasite appeared to be similar to those from other parasites (Additional file ##SUPPL##0##1## and ##SUPPL##1##2##). A sample mixed with a smaller percentage of DNA from a different genotype (strain) may increase the hybridization background signal. Indeed, typing DNA from the FCR3 parasite with microsatellites showed that the DNA sample appeared to contain a secondary peak in some markers (data not shown). If this is true, a sample with high background may have to be discarded.</p>", "<p>Using an array with a much higher density of probes than those published previously [##REF##16174539##27##, ####REF##16789840##28##, ##REF##17599553##29####17599553##29##], we identified 121,087 mSFP from five isolates, including ~18,000 new mSFP after excluding mSFP from multigene families. Among the 121,087 mSFP, ~67% were in clusters of highly polymorphic genes such as <italic>var/rif</italic>/<italic>stevor</italic>. Approximately 89% of our mSFP calls that also had probes spanning known SNP in PlasmoDB matched the SNP, reflecting relative high accuracy of our mSFP calls, although our stringent cutoff values may lead to higher Fn rates or \"no-calls\" (such as excluding single probe calls). Our mSFP also provided additional evidence confirming the SNP reported previously, which is important because the majority of SNP in PlasmoDB were generated from shotgun sequences and sequence alignments have not been visually inspected or adjusted. For a genome with large number of repetitive sequences, sequence alignment errors can be generated if sequence alignment is totally relied on computer software [##REF##17159981##4##].</p>", "<p>Distributions of mSFP across the chromosomes among the parasites were very similar except for a few unique peaks that may reflect deletion or amplification in each individual parasite. If we exclude the mSFP from the multigene families, we obtained 40,354 mSFP or approximately 570 bp per SFP in the genome, a frequency that is within the range (519–976 bp per SNP) of our previous estimates [##REF##12124624##35##] and similar to an estimate of 446 bp per SNP by another group [##REF##17159979##5##]. If we consider 45% of the 40,354 mSFP from five isolates as common mSFP, as estimated previously [##REF##17159981##4##], we can expect ~18,000 common mSFP in the five parasite genomes that will be useful for genetic mapping.</p>", "<p>The highly AT-rich <italic>P. falciparum </italic>genome has a large number of repetitive sequences and low complexity regions in protein coding sequences [##REF##12124624##35##, ####REF##11157785##36##, ##REF##14697197##37####14697197##37##]. The non-coding regions consist of more than 40% of the genome and generally have AT content &gt;90% with large numbers of polymorphic AT repeats and polyA/T tracts [##REF##12368864##26##,##REF##8661002##38##]. These high-AT regions not only present a problem for genome sequencing and DNA sequence alignment but also make it difficult to design sequence-specific probes with reliable hybridization dynamics. SNP in these regions may not be very useful for mapping purposes because of difficulty in designing oligonucleotide probes or PCR primers for genotyping. Indeed, analyses of signal intensity from probes with different GC contents showed that probes with GC contents &lt;16% produced similar low signals, suggesting that these probes might not be practical for calling mSFP. Of interest, probes with GC content &gt;50% also produced highly variable signals. The majority of high-GC probes from the variable <italic>var </italic>genes can partly contribute to this variation. We excluded probes with GC content &gt;50% for several reasons: 1) Approximately 44% of the probes with GC content &gt;50% were <italic>var </italic>probes that should be discarded; 2) probes with high GC content would have higher 'affinity' than those with lower GC content during hybridization. A substitution in a probe with high GC content may not reduce the hybridization signal as much as a probe with low GC content; 3) there were only ~3000 probes with GC contents &gt;50%. Exclusion of these probes should not have significant impact on our SFP calls.</p>", "<p>The <italic>P. falciparum </italic>chromosomes have been shown to be highly variable in size in pulse-field gel electrophoresis (PFG) [##REF##3889657##39##]. Genomic segmentation analysis to detect chromosome deletion and amplification showed relatively few amplification/deletion events with segment size &gt; 0.3 kb. The variation in chromosome sizes seen in PFG gels could be mainly due to chromosome translocation, which is difficult if not impossible to detect using microarrays. One of the amplified regions was a segment on chromosome 5 containing the <italic>pfmdr1 </italic>gene in the Dd2 and FCR3 parasites. Amplification of the <italic>pfmdr1 </italic>locus has been reported [##REF##16789840##28##,##REF##17599553##29##,##REF##1922044##33##], which could be due to drug selection pressure [##REF##15288742##40##]. Similarly, there were few deletions larger than 10 kb; many of the deleted/amplified regions detected in our study matched well with those reported previously [##REF##16789840##28##,##REF##17599553##29##]. Two well-known deleted regions on chromosome 2 and 9, respectively, were detected in our analyses [##REF##8367496##34##,##REF##2648403##41##]. Detection of previously reported deletions suggested that our methods for detecting deletion/amplification were working properly. However, using an array with higher probe density than previous studies, we also discovered many deletions/amplifications that have not been described previously (Additional file ##SUPPL##8##9##). We identified 181 amplified and 536 highly variable or deleted genes or fragments, 74 (40.9%) and 30 (5.6%) of which, respectively, were reported previously [##REF##16789840##28##,##REF##17599553##29##,##REF##1922044##33##]. Some of the discrepancies were likely due to different filtering criteria used (e.g. cutoff ratios, minimum number of probes, length cutoff of segment). Because of our small parasite sample size, it is difficult to make any functional inferences from the amplifications and deletions found in this study, although amplification at the <italic>pfmdr1 </italic>locus may be associated with responses to some anti-malarial drugs [##REF##15288742##40##,##REF##8302844##42##], and amplification of chromosome 4 in FCR3 may contribute to its adaptation to higher growth rates.</p>" ]

[ "<title>Conclusion</title>", "<p>This study developed methods for accurate detection of mSFP and CNV in the <italic>P. falciparum </italic>genome after evaluating factors that can influence DNA hybridization dynamics. More than 120,000 mSFP, including ~18,000 new and unique mSFP, and various chromosomal amplification/deletions were identified from the <italic>P. falciparum </italic>genome. Nearly 70% of the polymorphic sites are in clusters of <italic>var/rif/stevor </italic>gene families. Use of this array to analyze DNA samples from large numbers of parasites will facilitate our understanding of parasite diversity and evolution and genetic mapping of important parasite traits.</p>" ]

[ "<p>This is an Open Access article distributed under the terms of the Creative Commons Attribution License (<ext-link ext-link-type=\"uri\" xlink:href=\"http://creativecommons.org/licenses/by/2.0\"/>), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.</p>", "<title>Background</title>", "<p>Genetic mapping is a powerful method to identify mutations that cause drug resistance and other phenotypic changes in the human malaria parasite <italic>Plasmodium falciparum</italic>. For efficient mapping of a target gene, it is often necessary to genotype a large number of polymorphic markers. Currently, a community effort is underway to collect single nucleotide polymorphisms (SNP) from the parasite genome. Here we evaluate polymorphism detection accuracy of a high-density 'tiling' microarray with 2.56 million probes by comparing single feature polymorphisms (SFP) calls from the microarray with known SNP among parasite isolates.</p>", "<title>Results</title>", "<p>We found that probe GC content, SNP position in a probe, probe coverage, and signal ratio cutoff values were important factors for accurate detection of SFP in the parasite genome. We established a set of SFP calling parameters that could predict mSFP (SFP called by multiple overlapping probes) with high accuracy (≥ 94%) and identified 121,087 mSFP genome-wide from five parasite isolates including 40,354 unique mSFP (excluding those from multi-gene families) and ~18,000 new mSFP, producing a genetic map with an average of one unique mSFP per 570 bp. Genomic copy number variation (CNV) among the parasites was also cataloged and compared.</p>", "<title>Conclusion</title>", "<p>A large number of mSFP were discovered from the <italic>P. falciparum </italic>genome using a high-density microarray, most of which were in clusters of highly polymorphic genes at chromosome ends. Our method for accurate mSFP detection and the mSFP identified will greatly facilitate large-scale studies of genome variation in the <italic>P. falciparum </italic>parasite and provide useful resources for mapping important parasite traits.</p>" ]

[ "<title>Abbreviations</title>", "<p>CNV: copy number variation; CV: coefficient of variance; Fn: false negative; Fp: false positive; LPE: local pooled error; MS: microsatellites; ROC: receiver operating characteristic; SFP: single feature polymorphism; mSFP: SFP called by two or more overlapping probes; SNP: single nucleotide polymorphism.</p>", "<title>Competing interests</title>", "<p>The authors declare that they have no competing interests.</p>", "<title>Authors' contributions</title>", "<p>HJ prepared parasite culture, performed DNA extraction, labeling, DNA sequencing, array hybridization, and data analysis, and took an active part in writing the manuscript; MY performed data analysis; JM and LZ performed parasite culture and MS typing as well as DNA extraction; AI and YH performed data analysis; LJK, performed ROC and z-score analyses; RMS performed data analysis and took an active part in writing the manuscript; X-zS designed the project, performed data analysis, and took an active part in writing the manuscript. All authors read and approved the manuscript.</p>", "<title>Supplementary Material</title>" ]

[ "<title>Acknowledgements</title>", "<p>This work was supported by the Division of Intramural Research, National Institute of Allergy and Infectious Diseases, National Institutes of Health; the Intramural Research Program of the Center for Cancer Research, National Cancer Institute, National Institutes of Health; and in part was funded by NCI contract N01-CO-12400. The content of this publication does not necessarily reflect the views or policies of the Department of Health and Human Services, nor does mention of trade names, commercial products, or organizations imply endorsement by the U. S. Government. AI was supported by the Wellcome Trust. We thank NIAID intramural editor Brenda Rae Marshall for assistance and Jun Yang and Brandie Fullmer at the Laboratory of Immunopathogenesis and Bioinformatics, SAIC-Frederick, Inc. for microarray hybridizations, and David Bennett at Partek Inc. for advanced data analysis help.</p>" ]

[ "<fig position=\"float\" id=\"F1\"><label>Figure 1</label><caption><p><bold>Distribution of probes with different GC contents and the influence of GC content on signal intensity.</bold><bold> A</bold>. Number of probes with different GC contents. <bold>B</bold>. Hybridization signals from probes with different GC contents using 3D7 DNA. <bold>C</bold>. Hybridization signals from probes with different GC contents using DNA from 7G8. <bold>D</bold>. Signal ratios of 3D7 over 7G8 from probes with different GC contents. The box plots (<bold>B-D</bold>) showed the lowest intensity, lower quartile, median, upper quartile, and the highest intensity. Note large variations in probes with GC contents higher than 50%.</p></caption></fig>", "<fig position=\"float\" id=\"F2\"><label>Figure 2</label><caption><p><bold>Relationship between probe signal ratios and SNP positions.</bold> 7G8-same indicates signals from probes with no known NIAID SNP within the probes between 3D7 and 7G8 parasites (3D7/7G8); 7G8-diff indicates probes with known differences between 3D7 and 7G8 parasites. The definitions for the rest of the parasites (FCR3, Dd2, and HB3) are the same as those for 7G8. The dashed line indicates signal cutoff ratio value of 5.0.</p></caption></fig>", "<fig position=\"float\" id=\"F3\"><label>Figure 3</label><caption><p><bold>Relationship of receiver operating characteristic (ROC) curve and Z-score values and estimates of SFP call rates.</bold> The black line is the ROC curve, and the red line is the Z-score curve. The vertical dash line indicates false positive rate (1-specificity) of 5%, and horizon lines point to a Z-score value of 1.5 and sensitivity level (call rate) of approximately 81%, respectively. The curves were generated using data from all replicates of hybridization. SFP calls were compared with known NIAID SNP described previously (see text).</p></caption></fig>", "<fig position=\"float\" id=\"F4\"><label>Figure 4</label><caption><p><bold>Genome browser displays (drawn in Canvas) showing SFP, mSFP and SNP from two genomic loci on chromosome 7.</bold><bold>A</bold>. A genome browser window (~3 kb) showing expanded chromosome region covering <italic>pfcrt </italic>gene (top line) and predicted exons/introns of the p<italic>fcrt </italic>gene, SNP in PlasmoDB (blue circle), NIAID SNP (red diamonds), SFP from individual probe (light blue squares), mSFP (black squares) and all genomic probes covering the <italic>pfcrt </italic>gene. Color codes for the genomic probes are: green, probes in coding regions; purple, probes in noncoding regions; and yellow, probes spanning protein coding and noncoding regions. Note the mSFP matched well with those known SNP. <bold>B</bold>. An expanded region (500-bp window) from PF07_0028 showing distributions of PladmoDB SNP and array probe locations. Five of the seven PlasmoDB SNP (blue circle) in the intron were not covered by any probes. One SNP matched a mSFP call (black bars in multiple parasites), and another was covered by one probe and but was not called (filtered out because of single probe). The color codes for the genomic probes are the same as those in <bold>A</bold>; the labels are either SNP ID (blue circles) or probe ID (black and light blue bars).</p></caption></fig>", "<fig position=\"float\" id=\"F5\"><label>Figure 5</label><caption><p><bold>Copy number/segmentation analyses showing amplified and highly variable or deleted regions on 14 chromosomes.</bold> Amplified/deleted regions were displayed as a signal heat map (red, amplified; blue, deleted or highly polymorphic) from each parasite. The 14 chromosome diagrams showed amplified (red, &gt; 1.5) or deleted/highly variable regions (blue, &lt; 0.67) after filtering for regions 0.3 kb or larger. The dashed lines separate the four parasites in each chromosome in the order of 7G8, Dd2, FCR3, and HB3. The arrow indicates the chromosome 5 regions amplified in Dd2 and FCR3.</p></caption></fig>" ]

[ "<table-wrap position=\"float\" id=\"T1\"><label>Table 1</label><caption><p>Comparison of correct mSFP calling rates using different cut off values</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td/><td align=\"center\" colspan=\"3\">Overall rate</td><td align=\"center\" colspan=\"3\">GC filtered</td><td align=\"center\" colspan=\"3\">Probe filtered</td><td align=\"center\" colspan=\"3\">Corrected rate</td></tr><tr><td/><td colspan=\"3\"><hr/></td><td colspan=\"3\"><hr/></td><td colspan=\"3\"><hr/></td><td colspan=\"3\"><hr/></td></tr><tr><td align=\"center\">Cutoff value</td><td align=\"center\">7G8</td><td align=\"center\">Dd2</td><td align=\"center\">HB3</td><td align=\"center\">7G8</td><td align=\"center\">Dd2</td><td align=\"center\">HB3</td><td align=\"center\">7G8</td><td align=\"center\">Dd2</td><td align=\"center\">HB3</td><td align=\"center\">7G8</td><td align=\"center\">Dd2</td><td align=\"center\">HB3</td></tr></thead><tbody><tr><td align=\"center\">1.5</td><td align=\"center\">92.5</td><td align=\"center\">92.4</td><td align=\"center\">90.0</td><td align=\"center\">92.6</td><td align=\"center\">92.5</td><td align=\"center\">90.2</td><td align=\"center\">93.7</td><td align=\"center\">93.4</td><td align=\"center\">91.2</td><td align=\"center\">97.7</td><td align=\"center\">97.6</td><td align=\"center\">96.7</td></tr><tr><td align=\"center\">2.0</td><td align=\"center\">91.5</td><td align=\"center\">90.4</td><td align=\"center\">89.2</td><td align=\"center\">91.3</td><td align=\"center\">90.4</td><td align=\"center\">89.4</td><td align=\"center\">92.8</td><td align=\"center\">92.6</td><td align=\"center\">90.7</td><td align=\"center\">97.4</td><td align=\"center\">97.3</td><td align=\"center\">96.6</td></tr><tr><td align=\"center\">5.0</td><td align=\"center\">82.9</td><td align=\"center\">82.0</td><td align=\"center\">82.5</td><td align=\"center\">82.9</td><td align=\"center\">82.1</td><td align=\"center\">82.8</td><td align=\"center\">86.4</td><td align=\"center\">85.9</td><td align=\"center\">84.5</td><td align=\"center\">95.0</td><td align=\"center\">94.8</td><td align=\"center\">94.3</td></tr></tbody></table></table-wrap>", "<table-wrap position=\"float\" id=\"T2\"><label>Table 2</label><caption><p>DNA sequencing verification of false negative (Fn) and false positive (Fp) calls</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td align=\"left\">Gene ID</td><td align=\"left\">Chr position</td><td align=\"center\">Mism alle</td><td align=\"center\">SFPn</td><td align=\"center\">3D7</td><td align=\"center\">7G8</td><td align=\"left\">Forward (5'-3')</td><td align=\"left\">Reversed (5'-3')</td></tr></thead><tbody><tr><td align=\"left\">MAL2.808</td><td align=\"left\">chr2: 306218</td><td align=\"center\">T/A</td><td align=\"center\">Fn(0/5)</td><td align=\"center\">A</td><td align=\"center\">A</td><td align=\"left\">tcagtagtatcttttgtttc</td><td align=\"left\">atgtaaaactaccatcaaatg</td></tr><tr><td align=\"left\"><bold>PFC0210c</bold></td><td align=\"left\">chr3: 218122</td><td align=\"center\">G/G</td><td align=\"center\">Fp(4/0)</td><td align=\"center\">C</td><td align=\"center\">C</td><td align=\"left\">agatgtgttctttatctaatt</td><td align=\"left\">aaccaagtgataagcacata</td></tr><tr><td align=\"left\">PFC0235w</td><td align=\"left\">chr3: 248155</td><td align=\"center\">A/G</td><td align=\"center\">Fn(0/2)</td><td align=\"center\">A</td><td align=\"center\">A</td><td align=\"left\">ggaaatgtatttgagaaaaac</td><td align=\"left\">caatgtttactatccgaatt</td></tr><tr><td align=\"left\">PFC0770c</td><td align=\"left\">chr3: 718081</td><td align=\"center\">T/T</td><td align=\"center\">Fp(4/0)</td><td align=\"center\">T</td><td align=\"center\">A</td><td align=\"left\">atggggagcaaagaatttc</td><td align=\"left\">tattccatgatgtattatgat</td></tr><tr><td align=\"left\">PFC1065w</td><td align=\"left\">chr3: 995530</td><td align=\"center\">C/G</td><td align=\"center\">Fn(1/8)</td><td align=\"center\">G</td><td align=\"center\">G</td><td align=\"left\">ggaaaaagaagaagatttaa</td><td align=\"left\">aatatatcttccgaatcatc</td></tr><tr><td align=\"left\">PFC1065w</td><td align=\"left\">chr3: 995640</td><td align=\"center\">A/G</td><td align=\"center\">Fn(0/8)</td><td align=\"center\">A</td><td align=\"center\">A</td><td align=\"left\">atagatgtatcgtgtgataa</td><td align=\"left\">attattacttctgtctctag</td></tr><tr><td align=\"left\"><italic>PFE1390w</italic></td><td align=\"left\">chr5: 1154254</td><td align=\"center\">A/A</td><td align=\"center\">Fp(3/0)</td><td align=\"center\">T</td><td align=\"center\">T</td><td align=\"left\">cgaaaaagagaagaaaaact</td><td align=\"left\">tgtgttggcttcttaatatt</td></tr><tr><td align=\"left\">MAL6P1.232</td><td align=\"left\">chr6: 817214</td><td align=\"center\">T/A</td><td align=\"center\">Fn(0/4)</td><td align=\"center\">T</td><td align=\"center\">T</td><td align=\"left\">tccaaatcttctcaaagct</td><td align=\"left\">ggtttattcaaaacattagg</td></tr><tr><td align=\"left\">MAL7.743</td><td align=\"left\">chr7: 181822</td><td align=\"center\">C/C</td><td align=\"center\">Fp(5/0)</td><td align=\"center\">C</td><td align=\"center\">G</td><td align=\"left\">tttaatgcttccctttgctt</td><td align=\"left\">ataattgtgatgaagtgatg</td></tr><tr><td align=\"left\"><italic>MAL7P1.30</italic></td><td align=\"left\">chr7: 512599</td><td align=\"center\">T/T</td><td align=\"center\">Fp(1/0)</td><td align=\"center\">A</td><td align=\"center\">A</td><td align=\"left\">atggtagaataattcatatgt</td><td align=\"left\">ttatcacacatggtttcaac</td></tr><tr><td align=\"left\"><italic>MAL7P1.65</italic></td><td align=\"left\">chr7: 519234</td><td align=\"center\">T/C</td><td align=\"center\">Fn(0/2)</td><td align=\"center\">T</td><td align=\"center\">C</td><td align=\"left\">aaaacaaccgtctgatataa</td><td align=\"left\">taaacaataaatccaactgt</td></tr><tr><td align=\"left\">MAL7.2803</td><td align=\"left\">chr7: 621749</td><td align=\"center\">G/A</td><td align=\"center\">Fn(0/6)</td><td align=\"center\">G</td><td align=\"center\">G</td><td align=\"left\">ttttcgctcggattattaaa</td><td align=\"left\">gcaacatgatttttttttttc</td></tr><tr><td align=\"left\"><italic>MAL7P1.67</italic></td><td align=\"left\">chr7: 677205</td><td align=\"center\">C/A</td><td align=\"center\">Fn(0/2)</td><td align=\"center\">G</td><td align=\"center\">T</td><td align=\"left\">atttaacttactggattggt</td><td align=\"left\">aatggacaaccaggttaaaa</td></tr><tr><td align=\"left\">MAL7P1.82</td><td align=\"left\">chr7: 794419</td><td align=\"center\">A/C</td><td align=\"center\">Fn(0/7)</td><td align=\"center\">A</td><td align=\"center\">A</td><td align=\"left\">gtgtacttcattttgtagtta</td><td align=\"left\">atatctacaaaaggggaatt</td></tr><tr><td align=\"left\">MAL7P1.82</td><td align=\"left\">chr7: 794421</td><td align=\"center\">C/A</td><td align=\"center\">Fn(0/7)</td><td align=\"center\">A</td><td align=\"center\">A</td><td align=\"left\">ccatgtgctttcatatatat</td><td align=\"left\">ccatgtaccagctcatac</td></tr><tr><td align=\"left\">PF07_0102</td><td align=\"left\">chr7: 922368</td><td align=\"center\">C/A</td><td align=\"center\">Fn(0/4)</td><td align=\"center\">C</td><td align=\"center\">C</td><td align=\"left\">aagagtattaataattccgtc</td><td align=\"left\">gaacagaggatgaattattt</td></tr><tr><td align=\"left\"><italic>MAL8P1.42</italic></td><td align=\"left\">chr8: 1017925</td><td align=\"center\">T/A</td><td align=\"center\">Fn(0/1)</td><td align=\"center\">T</td><td align=\"center\">A</td><td align=\"left\">tccatgatatattcccaag</td><td align=\"left\">tattcctcatttcagggtat</td></tr><tr><td align=\"left\">MAL8.3159</td><td align=\"left\">chr8: 1057901</td><td align=\"center\">C/A</td><td align=\"center\">Fn(0/3)</td><td align=\"center\">C</td><td align=\"center\">C</td><td align=\"left\">gtacagctagttgtagtg</td><td align=\"left\">gagctttcttactaaagtat</td></tr><tr><td align=\"left\"><italic>PF08_0017</italic></td><td align=\"left\">chr8: 1179041</td><td align=\"center\">C/T</td><td align=\"center\">Fn(0/1)</td><td align=\"center\">C</td><td align=\"center\">T</td><td align=\"left\">cggtgataataataaatacg</td><td align=\"left\">gaatttatagaactttccgc</td></tr><tr><td align=\"left\">PF08_0017</td><td align=\"left\">chr8: 169329</td><td align=\"center\">T/C</td><td align=\"center\">Fn(0/1)</td><td align=\"center\">T</td><td align=\"center\">T</td><td align=\"left\">ccgtctacacaataattcttt</td><td align=\"left\">gggtagtaaatatgaggaaa</td></tr><tr><td align=\"left\">MAL8.2086</td><td align=\"left\">chr8: 582828</td><td align=\"center\">T/C</td><td align=\"center\">Fn(0/4)</td><td align=\"center\">T</td><td align=\"center\">T</td><td align=\"left\">tgggataaacctatgtataa</td><td align=\"left\">tcattcaaatttacaggtcg</td></tr><tr><td align=\"left\"><italic>PFI1300c</italic></td><td align=\"left\">chr9: 1080645</td><td align=\"center\">T/T</td><td align=\"center\">Fp(1/0)</td><td align=\"center\">A</td><td align=\"center\">A</td><td align=\"left\">tatgatgacaatcatattcc</td><td align=\"left\">ccttctatgaatagagatac</td></tr><tr><td align=\"left\"><italic>PFI1300c</italic></td><td align=\"left\">chr9: 1080729</td><td align=\"center\">G/A</td><td align=\"center\">Fn(0/2)</td><td align=\"center\">T</td><td align=\"center\">C</td><td align=\"left\">tacccatatcttgatttacg</td><td align=\"left\">ctttggagatttgtttagat</td></tr><tr><td align=\"left\">PFI0495w</td><td align=\"left\">chr9: 464714</td><td align=\"center\">G/G</td><td align=\"center\">Fp(5/0)</td><td align=\"center\">G</td><td align=\"center\">A</td><td align=\"left\">attctcccaaaactgaaata</td><td align=\"left\">atatcttcgttagttatgtg</td></tr><tr><td align=\"left\"><italic>MAL9.1104</italic></td><td align=\"left\">chr9: 548591</td><td align=\"center\">A/G</td><td align=\"center\">Fn(0/1)</td><td align=\"center\">A</td><td align=\"center\">G</td><td align=\"left\">tcttcttttcctttctacat</td><td align=\"left\">ttaaggttccttctgaatta</td></tr><tr><td align=\"left\">PFI0690c</td><td align=\"left\">chr9: 603205</td><td align=\"center\">T/A</td><td align=\"center\">Fn(0/2)</td><td align=\"center\">T</td><td align=\"center\">T</td><td align=\"left\">cgaaaaaatcctttacctt</td><td align=\"left\">aaagatttccccctactaaa</td></tr><tr><td align=\"left\">MAL5.878</td><td align=\"left\">chr9: 926274</td><td align=\"center\">G/A</td><td align=\"center\">Fn(0/1)</td><td align=\"center\">C</td><td align=\"center\">C</td><td align=\"left\">gttcgtcttttttttcatatg</td><td align=\"left\">Gaatataagacagatgttcc</td></tr><tr><td align=\"left\">PF10_0314</td><td align=\"left\">chr10: 1294935</td><td align=\"center\">C/G</td><td align=\"center\">Fn(3/17)</td><td align=\"center\">T</td><td align=\"center\">T</td><td align=\"left\">caatgtgaggaatatttatag</td><td align=\"left\">ggcctcattgtggttatta</td></tr><tr><td align=\"left\"><italic>MAL10.3336</italic></td><td align=\"left\">chr10: 1334877</td><td align=\"center\">A/A</td><td align=\"center\">Fp(1/0)</td><td align=\"center\">A</td><td align=\"center\">A</td><td align=\"left\">tttaaacacccctcaaaaaa</td><td align=\"left\">aaatatcaaaaccggaaatg</td></tr><tr><td align=\"left\">MAL10.4084</td><td align=\"left\">chr10: 1433239</td><td align=\"center\">G/A</td><td align=\"center\">Fn(0/1)</td><td align=\"center\">C</td><td align=\"center\">C</td><td align=\"left\">aagaaataattggttgggct</td><td align=\"left\">ttctgtccaccatttttttg</td></tr><tr><td align=\"left\"><italic>PF10_0377</italic></td><td align=\"left\">chr10: 1554669</td><td align=\"center\">T/A</td><td align=\"center\">Fn(9/11)</td><td align=\"center\">T</td><td align=\"center\">A</td><td align=\"left\">taaaacctgtataaccaaata</td><td align=\"left\">tatacaaactttacaaaactc</td></tr><tr><td align=\"left\">PF10_0094</td><td align=\"left\">chr10: 389999</td><td align=\"center\">A/C</td><td align=\"center\">Fn(0/2)</td><td align=\"center\">T</td><td align=\"center\">T</td><td align=\"left\">aaggtataccaatagatttg</td><td align=\"left\">gtaaatcattcaccctcat</td></tr><tr><td align=\"left\"><italic>PF10_0138</italic></td><td align=\"left\">chr10: 556132</td><td align=\"center\">C/C</td><td align=\"center\">Fp(2/1)</td><td align=\"center\">C</td><td align=\"center\">C</td><td align=\"left\">taatgtgtatgtatcagcta</td><td align=\"left\">ggattgtaataagtatatgg</td></tr><tr><td align=\"left\"><italic>MAL10.1222</italic></td><td align=\"left\">chr10: 564556</td><td align=\"center\">T/T</td><td align=\"center\">Fp(1/0)</td><td align=\"center\">T</td><td align=\"center\">T</td><td align=\"left\">gttttatgcttaggcttata</td><td align=\"left\">tgggaaaatataaatgaagg</td></tr><tr><td align=\"left\"><italic>PF11_0338</italic></td><td align=\"left\">chr11: 1272493</td><td align=\"center\">A/A</td><td align=\"center\">Fp(1/0)</td><td align=\"center\">A</td><td align=\"center\">A</td><td align=\"left\">gaatgttaacatacaaatgta</td><td align=\"left\">cttcagggagaatatttattc</td></tr><tr><td align=\"left\"><bold>MAL11.3013</bold></td><td align=\"left\">chr11: 1294419</td><td align=\"center\">T/T</td><td align=\"center\">Fp(3/0)</td><td align=\"center\">T</td><td align=\"center\">T</td><td align=\"left\">tcatggttcaggtataaga</td><td align=\"left\">ccattattttcttgagctgc</td></tr><tr><td align=\"left\">PF11_0353</td><td align=\"left\">chr11: 1327608</td><td align=\"center\">G/A</td><td align=\"center\">Fn(0/5)</td><td align=\"center\">G</td><td align=\"center\">G</td><td align=\"left\">ttataccatatgtgtacaaag</td><td align=\"left\">gaaatatcaaaatttcctaac</td></tr><tr><td align=\"left\">PF11_0360</td><td align=\"left\">chr11: 1369690</td><td align=\"center\">A/G</td><td align=\"center\">Fn(0/3)</td><td align=\"center\">A</td><td align=\"center\">A</td><td align=\"left\">cctattctattcaatactgt</td><td align=\"left\">ctgtatacatttgtttggat</td></tr><tr><td align=\"left\">PF11_0046</td><td align=\"left\">chr11: 151916</td><td align=\"center\">A/G</td><td align=\"center\">Fn(0/2)</td><td align=\"center\">A</td><td align=\"center\">A</td><td align=\"left\">acaagcatagatatcatagc</td><td align=\"left\">ataacatgtcctaaaggtga</td></tr><tr><td align=\"left\"><italic>PF11_0441</italic></td><td align=\"left\">chr11: 1717528</td><td align=\"center\">T/A</td><td align=\"center\">Fn(5/15)</td><td align=\"center\">A</td><td align=\"center\">T</td><td align=\"left\">cagttatatacctttatcag</td><td align=\"left\">ataagaaaaaatatccacac</td></tr><tr><td align=\"left\"><italic>MAL12.4052</italic></td><td align=\"left\">chr12: 1192527</td><td align=\"center\">T/G</td><td align=\"center\">Fn(1/2)</td><td align=\"center\">T</td><td align=\"center\">G</td><td align=\"left\">ggatattcacaatggatttt</td><td align=\"left\">catgtgtatcatttatacatg</td></tr><tr><td align=\"left\"><italic>MAL12.2128</italic></td><td align=\"left\">chr12: 577914</td><td align=\"center\">T/T</td><td align=\"center\">Fp(1/0)</td><td align=\"center\">T</td><td align=\"center\">T</td><td align=\"left\">ctgatgaaagaatacatattg</td><td align=\"left\">tgaacaatatattcggaaac</td></tr><tr><td align=\"left\"><bold>MAL12.3146</bold></td><td align=\"left\">chr12: 817466</td><td align=\"center\">T/T</td><td align=\"center\">Fp(1/0)</td><td align=\"center\">T</td><td align=\"center\">TAT</td><td align=\"left\">aatctaaaaaatccaagtatg</td><td align=\"left\">cataatgattgtatatccttt</td></tr><tr><td align=\"left\">PF13_0184</td><td align=\"left\">chr13: 1376386</td><td align=\"center\">T/C</td><td align=\"center\">Fn(1/9)</td><td align=\"center\">T</td><td align=\"center\">T</td><td align=\"left\">tattcttgaattttcgctac</td><td align=\"left\">tatattttatggatcatctc</td></tr><tr><td align=\"left\">MAL13.4760</td><td align=\"left\">chr13: 2159993</td><td align=\"center\">C/T</td><td align=\"center\">Fn(0/2)</td><td align=\"center\">C</td><td align=\"center\">C</td><td align=\"left\">cacaaaagtatacgtctat</td><td align=\"left\">ttaacagtttaggacacata</td></tr><tr><td align=\"left\">MAL13.670</td><td align=\"left\">chr13: 304167</td><td align=\"center\">C/A</td><td align=\"center\">Fn(0/1)</td><td align=\"center\">A</td><td align=\"center\">A</td><td align=\"left\">attaaataattcttcttccag</td><td align=\"left\">catgtcttgtatttcgtttt</td></tr><tr><td align=\"left\">MAL13P1.67</td><td align=\"left\">chr13: 557320</td><td align=\"center\">A/T</td><td align=\"center\">Fn(0/3)</td><td align=\"center\">A</td><td align=\"center\">A</td><td align=\"left\">gttcttctaacacaaataaa</td><td align=\"left\">tctacaggtaatatgttatc</td></tr><tr><td align=\"left\">PF13_0088</td><td align=\"left\">chr13: 650502</td><td align=\"center\">T/C</td><td align=\"center\">Fn(0/4)</td><td align=\"center\">G</td><td align=\"center\">G</td><td align=\"left\">cggcatgctcctgaagtaaa</td><td align=\"left\">ttatgttagagatgggtata</td></tr><tr><td align=\"left\"><italic>PF13_0125</italic></td><td align=\"left\">chr13: 912350</td><td align=\"center\">T/G</td><td align=\"center\">Fn(2/3)</td><td align=\"center\">A</td><td align=\"center\">G</td><td align=\"left\">catagtactatcacctgaa</td><td align=\"left\">ctatggttataaccaagaaat</td></tr><tr><td align=\"left\">MAL13P1.127</td><td align=\"left\">chr13: 958583</td><td align=\"center\">A/A</td><td align=\"center\">Fp(2/1)</td><td align=\"center\">T</td><td align=\"center\">C</td><td align=\"left\">gatgaatttgttgtaacgttt</td><td align=\"left\">acgttaataacaatcatgtga</td></tr><tr><td align=\"left\"><bold>MAL14.5217</bold></td><td align=\"left\">chr14: 2364467</td><td align=\"center\">A/A</td><td align=\"center\">Fp(3/0)</td><td align=\"center\">A</td><td align=\"center\">A</td><td align=\"left\">ggtatatcctttctacatat</td><td align=\"left\">aattcttttcatagggagtt</td></tr><tr><td align=\"left\"><italic>PF14_0565</italic></td><td align=\"left\">chr14: 2428920</td><td align=\"center\">A/T</td><td align=\"center\">Fn(16/23)</td><td align=\"center\">T</td><td align=\"center\">A</td><td align=\"left\">atcgtcaataccttcctcg</td><td align=\"left\">taaacaaaatatgagcactg</td></tr></tbody></table></table-wrap>", "<table-wrap position=\"float\" id=\"T3\"><label>Table 3</label><caption><p>Summary of mSFP calls for the 14 chromosomes among five parasite isolates</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td align=\"left\">Isolate</td><td align=\"left\">Ch1</td><td align=\"left\">Ch2</td><td align=\"left\">Ch3</td><td align=\"left\">Ch4</td><td align=\"left\">Ch5</td><td align=\"left\">Ch6</td><td align=\"left\">Ch7</td><td align=\"left\">Ch8</td><td align=\"left\">Ch9</td><td align=\"left\">Ch10</td><td align=\"left\">Ch11</td><td align=\"left\">Ch12</td><td align=\"left\">Ch13</td><td align=\"left\">Ch14</td><td align=\"left\">Total</td></tr></thead><tbody><tr><td align=\"left\">0000*</td><td align=\"left\">35727</td><td align=\"left\">54790</td><td align=\"left\">67238</td><td align=\"left\">67463</td><td align=\"left\">90602</td><td align=\"left\">85998</td><td align=\"left\">85713</td><td align=\"left\">86826</td><td align=\"left\">98064</td><td align=\"left\">190050</td><td align=\"left\">223105</td><td align=\"left\">212037</td><td align=\"left\">340227</td><td align=\"left\">231400</td><td align=\"left\">1869240</td></tr><tr><td align=\"left\">0001</td><td align=\"left\">247</td><td align=\"left\">615</td><td align=\"left\">656</td><td align=\"left\">896</td><td align=\"left\">429</td><td align=\"left\">506</td><td align=\"left\">655</td><td align=\"left\">524</td><td align=\"left\">635</td><td align=\"left\">765</td><td align=\"left\">606</td><td align=\"left\">1092</td><td align=\"left\">1486</td><td align=\"left\">1337</td><td align=\"left\">10449</td></tr><tr><td align=\"left\">0010</td><td align=\"left\">209</td><td align=\"left\">116</td><td align=\"left\">130</td><td align=\"left\">331</td><td align=\"left\">304</td><td align=\"left\">211</td><td align=\"left\">252</td><td align=\"left\">313</td><td align=\"left\">1529</td><td align=\"left\">54</td><td align=\"left\">390</td><td align=\"left\">245</td><td align=\"left\">393</td><td align=\"left\">244</td><td align=\"left\">5121</td></tr><tr><td align=\"left\">0011</td><td align=\"left\">180</td><td align=\"left\">41</td><td align=\"left\">49</td><td align=\"left\">244</td><td align=\"left\">55</td><td align=\"left\">137</td><td align=\"left\">143</td><td align=\"left\">269</td><td align=\"left\">121</td><td align=\"left\">173</td><td align=\"left\">98</td><td align=\"left\">165</td><td align=\"left\">156</td><td align=\"left\">76</td><td align=\"left\">1907</td></tr><tr><td align=\"left\">0100</td><td align=\"left\">349</td><td align=\"left\">1636</td><td align=\"left\">348</td><td align=\"left\">688</td><td align=\"left\">537</td><td align=\"left\">344</td><td align=\"left\">683</td><td align=\"left\">527</td><td align=\"left\">497</td><td align=\"left\">761</td><td align=\"left\">829</td><td align=\"left\">1224</td><td align=\"left\">890</td><td align=\"left\">755</td><td align=\"left\">10068</td></tr><tr><td align=\"left\">0101</td><td align=\"left\">248</td><td align=\"left\">556</td><td align=\"left\">73</td><td align=\"left\">479</td><td align=\"left\">279</td><td align=\"left\">203</td><td align=\"left\">336</td><td align=\"left\">268</td><td align=\"left\">193</td><td align=\"left\">191</td><td align=\"left\">178</td><td align=\"left\">315</td><td align=\"left\">164</td><td align=\"left\">137</td><td align=\"left\">3620</td></tr><tr><td align=\"left\">0110</td><td align=\"left\">175</td><td align=\"left\">165</td><td align=\"left\">193</td><td align=\"left\">491</td><td align=\"left\">173</td><td align=\"left\">267</td><td align=\"left\">384</td><td align=\"left\">199</td><td align=\"left\">227</td><td align=\"left\">506</td><td align=\"left\">223</td><td align=\"left\">488</td><td align=\"left\">478</td><td align=\"left\">222</td><td align=\"left\">4191</td></tr><tr><td align=\"left\">0111</td><td align=\"left\">296</td><td align=\"left\">213</td><td align=\"left\">235</td><td align=\"left\">551</td><td align=\"left\">941</td><td align=\"left\">501</td><td align=\"left\">406</td><td align=\"left\">513</td><td align=\"left\">389</td><td align=\"left\">543</td><td align=\"left\">378</td><td align=\"left\">866</td><td align=\"left\">597</td><td align=\"left\">397</td><td align=\"left\">6826</td></tr><tr><td align=\"left\">1000</td><td align=\"left\">300</td><td align=\"left\">278</td><td align=\"left\">359</td><td align=\"left\">563</td><td align=\"left\">345</td><td align=\"left\">467</td><td align=\"left\">580</td><td align=\"left\">598</td><td align=\"left\">511</td><td align=\"left\">1125</td><td align=\"left\">687</td><td align=\"left\">762</td><td align=\"left\">1287</td><td align=\"left\">994</td><td align=\"left\">8856</td></tr><tr><td align=\"left\">1001</td><td align=\"left\">138</td><td align=\"left\">241</td><td align=\"left\">123</td><td align=\"left\">427</td><td align=\"left\">138</td><td align=\"left\">401</td><td align=\"left\">625</td><td align=\"left\">315</td><td align=\"left\">315</td><td align=\"left\">621</td><td align=\"left\">484</td><td align=\"left\">342</td><td align=\"left\">490</td><td align=\"left\">327</td><td align=\"left\">4987</td></tr><tr><td align=\"left\">1010</td><td align=\"left\">179</td><td align=\"left\">75</td><td align=\"left\">70</td><td align=\"left\">219</td><td align=\"left\">55</td><td align=\"left\">81</td><td align=\"left\">117</td><td align=\"left\">136</td><td align=\"left\">49</td><td align=\"left\">182</td><td align=\"left\">275</td><td align=\"left\">106</td><td align=\"left\">209</td><td align=\"left\">162</td><td align=\"left\">1915</td></tr><tr><td align=\"left\">1011</td><td align=\"left\">243</td><td align=\"left\">288</td><td align=\"left\">144</td><td align=\"left\">533</td><td align=\"left\">47</td><td align=\"left\">593</td><td align=\"left\">447</td><td align=\"left\">438</td><td align=\"left\">369</td><td align=\"left\">363</td><td align=\"left\">282</td><td align=\"left\">348</td><td align=\"left\">684</td><td align=\"left\">316</td><td align=\"left\">5095</td></tr><tr><td align=\"left\">1100</td><td align=\"left\">123</td><td align=\"left\">179</td><td align=\"left\">166</td><td align=\"left\">309</td><td align=\"left\">87</td><td align=\"left\">141</td><td align=\"left\">343</td><td align=\"left\">143</td><td align=\"left\">78</td><td align=\"left\">204</td><td align=\"left\">177</td><td align=\"left\">272</td><td align=\"left\">424</td><td align=\"left\">142</td><td align=\"left\">2788</td></tr><tr><td align=\"left\">1101</td><td align=\"left\">477</td><td align=\"left\">271</td><td align=\"left\">356</td><td align=\"left\">860</td><td align=\"left\">240</td><td align=\"left\">1180</td><td align=\"left\">134</td><td align=\"left\">495</td><td align=\"left\">356</td><td align=\"left\">496</td><td align=\"left\">476</td><td align=\"left\">880</td><td align=\"left\">547</td><td align=\"left\">207</td><td align=\"left\">7975</td></tr><tr><td align=\"left\">1110</td><td align=\"left\">363</td><td align=\"left\">197</td><td align=\"left\">340</td><td align=\"left\">731</td><td align=\"left\">222</td><td align=\"left\">306</td><td align=\"left\">628</td><td align=\"left\">433</td><td align=\"left\">133</td><td align=\"left\">408</td><td align=\"left\">402</td><td align=\"left\">560</td><td align=\"left\">644</td><td align=\"left\">222</td><td align=\"left\">5589</td></tr><tr><td align=\"left\">1111</td><td align=\"left\">2507</td><td align=\"left\">1924</td><td align=\"left\">1607</td><td align=\"left\">3829</td><td align=\"left\">968</td><td align=\"left\">3720</td><td align=\"left\">4215</td><td align=\"left\">3149</td><td align=\"left\">2816</td><td align=\"left\">3984</td><td align=\"left\">3072</td><td align=\"left\">3948</td><td align=\"left\">3809</td><td align=\"left\">2152</td><td align=\"left\">41700</td></tr><tr><td align=\"left\">Total</td><td align=\"left\">6034</td><td align=\"left\">6795</td><td align=\"left\">4849</td><td align=\"left\">11151</td><td align=\"left\">4820</td><td align=\"left\">9058</td><td align=\"left\">10948</td><td align=\"left\">8320</td><td align=\"left\">8218</td><td align=\"left\">10776</td><td align=\"left\">8557</td><td align=\"left\">11613</td><td align=\"left\">12258</td><td align=\"left\">7690</td><td align=\"left\">121087</td></tr></tbody></table></table-wrap>" ]

[]

[]

[]

[]

[]

[ "<supplementary-material content-type=\"local-data\" id=\"S1\"><caption><title>Additional file 1</title><p>Parasite sample replicates and basic hybridization statistics after normalization.</p></caption></supplementary-material>", "<supplementary-material content-type=\"local-data\" id=\"S2\"><caption><title>Additional file 2</title><p>Plots of normalized signal ratios averaged from parasite replicates, showing distribution of probe signal ratios from each parasite (over 3D7).</p></caption></supplementary-material>", "<supplementary-material content-type=\"local-data\" id=\"S3\"><caption><title>Additional file 3</title><p>Number of NIAID SNP that are covered by different numbers of probes.</p></caption></supplementary-material>", "<supplementary-material content-type=\"local-data\" id=\"S4\"><caption><title>Additional file 4</title><p>The numbers of probes with NIAID SNP at positions 1–25. Probes with a single SNP are in light blue, two SNP are in red, and more than two SNP are in dark blue.</p></caption></supplementary-material>", "<supplementary-material content-type=\"local-data\" id=\"S5\"><caption><title>Additional file 5</title><p>Summary of procedures for calling genome-wide SFP and copy number variation.</p></caption></supplementary-material>", "<supplementary-material content-type=\"local-data\" id=\"S6\"><caption><title>Additional file 6</title><p>mSFP calls for the 14 chromosomes among five parasite isolates after excluding calls from multigene families.</p></caption></supplementary-material>", "<supplementary-material content-type=\"local-data\" id=\"S7\"><caption><title>Additional file 7</title><p>Numbers of SFP, mSFP, and known SNP in predicted <italic>P. falciparum </italic>genes.</p></caption></supplementary-material>", "<supplementary-material content-type=\"local-data\" id=\"S8\"><caption><title>Additional file 8</title><p>SFP counts per 10-kb bins across the 14 chromosomes from 7G8, Dd2, FCR3, and HB3.</p></caption></supplementary-material>", "<supplementary-material content-type=\"local-data\" id=\"S9\"><caption><title>Additional file 9</title><p>Amplified and deleted chromosomal segments or genes 300 bp or larger.</p></caption></supplementary-material>" ]

[ "<table-wrap-foot><p>To obtain the best correct call rates, we compared mSFP calls using three cutoff values (1.5, 2.0, and 5.0). First we called mSFP using unique probes and probe position 3–23 (Overall rate). We repeated the calls after removing probes with GC contents &lt; 16% and &gt; 50% (GC filtered). We then obtained call rates after removing probes with GC content &lt; 16% and &gt; 50% and excluding calls with single probes and multiple probe calls with less than 75% probe votes (Probe filtered). Corrected rates were obtained after adjusting for 63.5% error rate in the wrong calls due to sequence errors, which were calculated using formula</p><p>[(100-probe filtered rate) × 0.635 + probe filtered rate].</p><p>A correct call was defined as correct calls over the sum of correct, wrong, and tie calls.</p></table-wrap-foot>", "<table-wrap-foot><p>Gene ID, gene ID or SNP ID in PlasmoDB; Chr position, chromosomal position of the polymorphic site; Mism Alle, mismatched alleles of our array calls and known NIAID SNP between 3D7 and 7G8; SFPn, calls not matching known SNP, either false positive (Fp) or false negative (Fn). The numbers in the parentheses are numbers of probes calling for SFP or no SFP. For example, Fp(3/0) indicates three probes called for a SFP and no probe called for no SFP, but there was no known SNP in the databases; and Fn(0/3) indicates three probes called for no SFP, but a known SNP existed (false negative); 3D7, alleles obtained from sequencing 3D7 DNA; and 7G8, alleles obtained from sequencing 7G8 DNA sequences. The gene ID in italic indicates SNP not confirmed by sequencing (true wrong calls) using 1.5 cutoff ratio and 3–23 positions in a probe; and those in <bold>bold </bold>had additional polymorphisms supporting the array calls. TAT in MAL12.3146 is a trinucleotide missing in 7G8. Forward and reverse are primers used in amplification and sequencing of the PCR products.</p></table-wrap-foot>", "<table-wrap-foot><p>*Parasite isolate order is 7G8, Dd2, FCR3, and HB3. For example, '1000' indicates the numbers of unique alleles for 7G8. A '0' indicates that a parasite has the same allele as that of 3D7 (0), and '1' indicates a different allele (a mSFP). The numbers in the first row were positions with probes but no SFP were called (no polymorphism). These numbers were not counted in the total calculation. The counts were based on a signal cutoff value of 5.0. Note these calls were mSFP and were different from those defined previously, where each probe was defined as an independent SFP[##REF##16789840##28##].</p></table-wrap-foot>" ]

[ "<graphic xlink:href=\"1471-2164-9-398-1\"/>", "<graphic xlink:href=\"1471-2164-9-398-2\"/>", "<graphic xlink:href=\"1471-2164-9-398-3\"/>", "<graphic xlink:href=\"1471-2164-9-398-4\"/>", "<graphic xlink:href=\"1471-2164-9-398-5\"/>" ]

[ "<media xlink:href=\"1471-2164-9-398-S1.doc\" mimetype=\"application\" mime-subtype=\"msword\"><caption><p>Click here for file</p></caption></media>", "<media xlink:href=\"1471-2164-9-398-S2.pdf\" mimetype=\"application\" mime-subtype=\"pdf\"><caption><p>Click here for file</p></caption></media>", "<media xlink:href=\"1471-2164-9-398-S3.pdf\" mimetype=\"application\" mime-subtype=\"pdf\"><caption><p>Click here for file</p></caption></media>", "<media xlink:href=\"1471-2164-9-398-S4.pdf\" mimetype=\"application\" mime-subtype=\"pdf\"><caption><p>Click here for file</p></caption></media>", "<media xlink:href=\"1471-2164-9-398-S5.pdf\" mimetype=\"application\" mime-subtype=\"pdf\"><caption><p>Click here for file</p></caption></media>", "<media xlink:href=\"1471-2164-9-398-S6.doc\" mimetype=\"application\" mime-subtype=\"msword\"><caption><p>Click here for file</p></caption></media>", "<media xlink:href=\"1471-2164-9-398-S7.xls\" mimetype=\"application\" mime-subtype=\"vnd.ms-excel\"><caption><p>Click here for file</p></caption></media>", "<media xlink:href=\"1471-2164-9-398-S8.xls\" mimetype=\"application\" mime-subtype=\"vnd.ms-excel\"><caption><p>Click here for file</p></caption></media>", "<media xlink:href=\"1471-2164-9-398-S9.xls\" mimetype=\"application\" mime-subtype=\"vnd.ms-excel\"><caption><p>Click here for file</p></caption></media>" ]

{ "acronym": [], "definition": [] }

[ "<title>Background</title>", "<p>A striking feature of the human genome is its plasticity, which is illustrated by the many occurrences of structural variations such as indels and copy number variations [##REF##17803354##1##, ####REF##16809666##2##, ##REF##15895083##3##, ##REF##16327808##4##, ##REF##17122850##5##, ##REF##17115057##6####17115057##6##]. Exonic structural variation may have a direct influence on gene products, and hence of interest for e.g. resequencing studies. In an earlier paper we studied STRs (originally called periodic DNA; see Figure ##FIG##0##1## and Methods for a definition) and demonstrated that STRs, in contrast to longer tandem repeats, are common in exonic regions, and that SNPs are more frequent in STRs compared with non-STRs [##REF##17673700##7##]. Long intergenic tandem repeats are well known targets for structural variation, and in this study we investigate whether exonic STRs share this property, and hence may serve as a probable target for exonic disease causing mutations.</p>", "<p>The tandem repeat content of mammalian genomes has been investigated in several papers, generally confining the analysis to intergenic regions and/or assuming the repeat element is repeated many times [##REF##15479712##8##, ####REF##12045144##9##, ##REF##16086851##10##, ##REF##15716087##11##, ##REF##15059995##12##, ##REF##16567018##13##, ##REF##18032720##14####18032720##14##]. Reports on tandem repeat sequences in human exons have found that almost all repeats have a period (unit size) that follows the codon size (i.e. a period of 3, 6 or 9 bp) [##REF##15479712##8##,##REF##16567018##13##,##REF##17581576##15##]. In concordance, most known repeat-related diseases are caused by expansion of 3-repeat elements (Trinucleotide Disease) in relatively long tandem repeats [##REF##17581576##15##], but other types of length variations may likewise contribute to disease risk. Here we focus on a class of very short tandem repeats and their contribution to disease risk.</p>", "<p>We found a strong excess of validated indels in STR regions and demonstrated that exonic STRs are likely targets for disease causing mutations by showing that disease-related genes have a significantly higher STR content than non-disease-related genes.</p>" ]

[ "<title>Materials and methods</title>", "<title>Reference genome sequences</title>", "<p>The human reference genome [##REF##11237011##24##] (hg18, NCBI build 36) and the mouse reference genome [##REF##12466850##25##] (mm8, NCBI 36) were used in the analyses. Base pairs assigned 'N' (i.e. gaps) in the reference sequences were omitted in the analysis and the pruned genomes were referred to as the \"entire genomes\" (human genome length: 2,858,013,089 bp; mouse genome length: 2,550,169,439 bp).</p>", "<title>Short tandem repeat identification</title>", "<p>We identified STRs by scanning the genomes for DNA segments that fulfil four criteria. A sequence is defined as STR with period <italic>p</italic>, if it fulfils the following: (1) the length of the sequence is at least 9 bp, (2) a motif (e.g., AT in ATATATATAT) of length <italic>p</italic>(≥1) is repeated at least three times with (3) at most one bp not matching a perfect repetition of the motif in sliding windows of max(12, 3·<italic>p</italic>) bp, and (4) the two flanking bp of the sequence must match the motif. Known polymorphic single nuclear substitutions are used to allow mismatches in the reference genome, consequently all possible alleles are analyzed (Figure ##FIG##0##1##). We used all polymorphic single nuclear substitutions from ENSEMBL 46 (containing dbSNP build 127 [##REF##11125122##26##] for humans and dbSNP build 126 for mouse [##REF##11125122##26##]). The data were downloaded as the \"ENSEMBL 46 VARIATION\" track from the BioMart Browser [##REF##16082012##27##]. If a STR sequence is assigned more than one period, we used the smallest.</p>", "<p>Only 0.8% of all STRs with periods 1–25 have period &gt; 9 (Additional file ##SUPPL##0##1##: Figure S3), hence we only used periods &lt;10 bp when analyzing the entire genomes. The entire human genome has 114,996,351 bp tagged as STRs (4.02% of the entire genome), and the entire mouse genome contains 137,927,765 bp tagged as STRs (5.41% of the entire genome).</p>", "<title>Indels</title>", "<p>We used all insertions and deletions (indels) from ENSEMBL 46 (containing dbSNP build 127 [##REF##11125122##26##]). The data were downloaded as the \"ENSEMBL 46 VARIATION\" track from the BioMart Browser [##REF##16082012##27##]. To obtain validated indels only, the data were filtered to contain only observations with validation \"freq\" and/or \"doublehit\" (the minor allele is seen at least twice) and Mapweight 1 (the highest quality alignments), resulting in 4,351 validated insertions and 16,899 validated deletions. To differentiate between insertions and deletions, we used the state given by dbSNP, which is defined according to the reference sequence.</p>", "<title>Disease-related gene sets</title>", "<p>Human and mouse genes were downloaded using BioMart (ENSEMBL 46) [##REF##16082012##27##] only including \"KNOWN\" genes with \"KNOWN\" transcripts. This resulted in 21,658 human genes with 39,684 transcripts and 21,946 mouse genes with 28,576 transcripts. If a gene had multiple transcripts we clustered all exons from all transcripts into one super-transcript.</p>", "<p>The OMIM Morbid Map (August 30, 2007) which contains the cytogenetic map locations of all disease genes described in the OMIM database [##REF##15608251##28##] was used to assign disease status of human genes. We created four sets of human disease genes: The general set (<italic>all diseases</italic>, 2095 genes) consists of all Morbid Map genes, except genes annotated with terms related to homosexuality and protections against diseases. Three subsets were defined using disease terms: A <italic>leukaemia </italic>set (70 genes, term: 'leukaemia'), a <italic>cancer set </italic>excluding leukaemia (151 genes, terms: 'carcinom', 'cancer', 'tumour', 'burkitt lymphoma', 'malignant melanoma', 'multiple endocrine neoplasia', 'neurofibromatosis', 'polycystic kidney disease', 'harvey ras oncogene', 'retinoblastoma', 'tuberous sclerosis' and 'von hippel-lindau syndrome') and an <italic>immune system disease </italic>set, excluding cancer and leukaemia (52 genes, terms: 'asthma', 'ataxia telangiectasia', 'autoimmune', 'digeorge syndrome' and 'immunodeficiency').</p>", "<p>We defined two non-overlapping sets of mouse disease genes. The first set of 294 mouse cancer genes is the result of querying the Mouse Genome Database (MGD) [##REF##14681461##29##] for \"increased tumour incidence\" in the mammalian phenotype ontology [##REF##14681461##29##]. The second set consists of 764 mouse genes associated with \"postnatal lethality\" after removal of genes overlapping the cancer set.</p>", "<title>Reference gene sets</title>", "<p>The reference set of \"non-disease-related\" human genes was defined as the 11,210 known genes not found in the OMIM database, whereas the mouse reference set was defined as the 17,171 known mouse genes not mapped to the mammalian phenotype ontology [##REF##14681461##29##].</p>", "<title>Known tandem repeats</title>", "<p>The \"Simple Repeats\" track in the UCSC Genome Browser [##REF##14681465##30##] act as a <italic>de facto </italic>definition of tandem repeats (possibly imperfect), identified by Tandem Repeats Finder [##REF##9862982##31##]. The track was created using the following parameter settings for TRF; match = 2, mismatch = 7, indels = 7, matching probability = 0.80, indel probability = 0.10, maximum period = 50, and minimum alignment score = 2000.</p>", "<title>STRs outside known tandem repeats</title>", "<p>STRs outside known tandem repeats are defined by applying the following two filters: (A) STRs inside known tandem repeats are omitted from the analysis; (B) all contiguous segments of STRs are clustered, and all such clusters which are more than 25 bp long are omitted from the analysis.</p>", "<title>Statistical methods</title>", "<p>To test for excess of insertions/deletions in STRs, we used a binomial test. The observed number of insertions/deletions inside STRs was compared to the binomial distribution b(<italic>n, p</italic>) where <italic>n </italic>is the total number of validated insertions/deletions and <italic>p </italic>= 0.0402 is the proportion of STRs in the human genome. We define an indel to be inside a STR segment if the midpoint of the indel is within the segment. The midpoint is defined as (<italic>s</italic>+<italic>e</italic>)/2, where <italic>s </italic>is the start coordinate and <italic>e </italic>is the end coordinate of the indel.</p>", "<p>The distribution is relative STR amount is non-Gaussian (Additional file ##SUPPL##0##1##: Figure S4) and a standard t-test cannot be applied. Instead, we used the Wilcoxon rank-sum test [##UREF##0##32##] to compare the relative STR amount in disease-related genes to the relative amount in reference genes, because the test does not require assumptions about the underlying distribution of relative STR amount. The STR overrepresentation for each disease-related gene set is found by comparing the estimated median relative STR content in the disease-related gene set to the estimated median relative STR content in the reference gene set. Confidence intervals of the estimated overrepresentation are obtained by Gaussian approximation of the distribution of rank sums from the Wilcoxon rank-sum test.</p>", "<p>All data were analyzed using Python <ext-link ext-link-type=\"uri\" xlink:href=\"http://www.python.org\"/> and R <ext-link ext-link-type=\"uri\" xlink:href=\"http://www.R-project.org\"/>[##UREF##1##33##]. All scripts are available upon request.</p>" ]

[ "<title>Results</title>", "<title>STR content</title>", "<p>Initially, we annotated the human and mouse genomes with STRs (see Figure ##FIG##0##1## for an example and Methods for how STRs are identified). The identified STRs make up 4.02% of the human genome, and the majority of the identified STR segments (62.1%) are imperfect repeats regions, i.e. they contain polymorphic base-pairs or base-pairs that do not match the periodic pattern. It appeared that 92.23% of all known human genes have STRs in their exons and 99% of STR regions are shorter than 33 bp. The short length differentiates exonic STRs from known exonic tandem repeats because the two groups share little overlap; 96.42 % of the identified STRs are shorter than 25 bp and outside known tandem repeats as defined by the UCSC Simple Repeats track (see Methods).</p>", "<title>Indels and STRs</title>", "<p>When looking at known indels, we found that STRs are related to hypermutability. Insertions as well as deletions are much more common inside than outside STR regions, both in the entire genome and in exons only (Figure ##FIG##1##2##, Additional file ##SUPPL##0##1##: Table S1). Furthermore, the lengths of indels in STRs are in agreement with the STR period (Figure ##FIG##2##3##, Additional file ##SUPPL##0##1##: Table S2), indicating that indels may be generated by slipped-strand mispairing [##REF##3328815##16##].</p>", "<p>The majority of indels found in exons have lengths different from the codon size (3, 6 or 9 bp), both inside and outside STRs (Figure ##FIG##3##4##). To test whether the increased frequency of indels is confined to long tandem repeat-like regions, we limited our analysis to STRs (≤25 bp) not overlapping with known tandem repeats. We found that the higher frequency of indels in STRs is preserved (Additional file ##SUPPL##0##1##: Figure S1) in the set of STRs with length ≤25 bp.</p>", "<title>STRs are overrepresented in disease genes</title>", "<p>First, we found that exons of disease-related genes generally are longer than those of reference genes, and also that the amount of STRs in exons is larger in disease genes than in reference genes (Figure ##FIG##4##5## and Additional file ##SUPPL##0##1##: Figure S2). To compare the amount of STRs in different subsets of genes we therefore used the relative amount of STRs in a gene, i.e. the length of STRs in the gene relative to the length of the gene. We found that all four subsets of disease-related genes had significantly higher relative amounts of STR regions in exons than non-disease-related genes, and that almost all disease-related genes have STRs in their exons (Table ##TAB##0##1##, Additional file ##SUPPL##0##1##: Figure S2). In contrast, this is not true if we consider introns instead of exons (Additional file ##SUPPL##0##1##: Table S4). To validate the findings, we replicated the analysis in mouse using data from the Mouse Genome Database (MGD) [##REF##17135206##17##] and obtained similar results (Table ##TAB##0##1##, Additional file ##SUPPL##0##1##: Figures S2).</p>", "<p>Unsurprisingly, we found more STRs of periods 3, 6, 9 in exons than in the entire genome (Figure ##FIG##5##6A, C##), but there appears to be no obvious correlation between disease status and STR period (Figure ##FIG##5##6B, D##). Furthermore, the observed excess of STRs in disease genes is preserved when using only STRs of periods different from 3, 6 or 9 bp (Additional file ##SUPPL##0##1##: Table S3)</p>", "<p>Our definition of STR focuses specifically on small, periodic regions with few repeats. To see if the pattern mainly originates from known tandem repeats, we excluded both long STRs (&gt;25 bp) as well as all STRs overlapping the UCSC simple repeats track (see Methods) and found that the results were not affected (Additional file ##SUPPL##0##1##: Table S5). We also allow for imperfections in the definition of STR (see Methods) which potentially biases our results: If more polymorphic sites are known in disease-related genes than in other genes, then this could lead to an (artificial) excess of STRs in disease-related genes. To test whether this is the case, we defined STR from the reference sequence alone (ignoring known polymorphic sites), redid the analysis and found that the results were not affected (Additional file ##SUPPL##0##1##: Table S6).</p>" ]

[ "<title>Discussion</title>", "<p>Our definition of STR identify short, possibly imperfect tandem repeats as short as 9 bp with the motif repeated at least 3 times; it takes both known genetic variation and rare pattern deviations into account. As there is no consensus in the literature about cut-off values for identification of STRs [##REF##12949124##20##], we chose cut-off values of 3 repeats and 9 bp minimum length. Interestingly, two other studies point to these cut-off values as reasonable: Ref [##REF##17926066##21##] identifies orthologous, alignable STRs in the human and chimpanzee reference genomes and estimate that polymerase slippage is negligible below 10 bp. Ref [##REF##17442102##22##] compiles a microsatellite data set with perfect repeats from the reference genome and estimates that polymerase slippage mutations do rarely occur unless the STR length is &gt;8–9 bp and the number of repeats &gt;3. Both studies use only the reference genome(s) and mathematical models to estimate the slippage threshold.</p>", "<p>We have used a specialized algorithm to detect STRs, but other tandem repeats detecting software could potentially identify similar sequences, if the parameters of the software are tuned to look for shorter tandem repeats than those found using the default/standard parameters. However the existing software for tandem repeat detection differ significantly in what is identified as STRs (using default/standard settings) and hence the resulting STR content depends on the software [##REF##17442102##22##]. One benefit of our definition is that it is straightforward to verify whether a sequence is STR.</p>", "<p>Indels involved in diseases are well known [##REF##17581576##15##,##REF##17823326##18##,##REF##16941003##19##] which suggests that the observed excess of STRs in the exons of disease-related genes is linked to the excess of indels in STRs. As described earlier, long tandem repeats in human exons contain almost no repeats with periods different from 3, 6 or 9 bp [##REF##15479712##8##,##REF##12045144##9##], and most diseases related to known tandem repeats are caused by expansion of 3-repeat elements (Trinucleotide Disease). This is in marked contrast to our findings showing that more than 50% of exonic STRs have periods different from 3, 6 and 9 bp and that the observed excess of STRs in disease-related genes is preserved when using only STRs with periods different from 3, 6 and 9 bp. The basic explanation is that STRs generally are shorter and/or less perfect than tandem repeats, and not detected by the commonly used tandem repeats software.</p>", "<p>The main difference between the reference genes and the four sets of disease-related genes is a difference in number of zero- or low STR content (Figure ##FIG##4##5##). The reference set contains 10–12 % genes with no STRs at all, whereas this fraction is substantially lower in the four disease-related sets. A possible explanation could be failure to detect STRs in really short genes, but when we removed short genes (&lt;1000 bases) from the analysis, the results were not affected (data not shown). We conclude that the most likely explanation is a link between STR content, hypermutability and disease-status.</p>", "<p>The observed association with disease genes, hypermutability and wide distribution of STRs suggests that STRs in exons may be good candidates to screen for rare disease causing mutations. This is supported by the observation that exons of disease related genes virtually always harbour STRs (Figure ##FIG##4##5##). Today, primarily genome-wide SNP association studies are used to identify genetic variants or regions implicated in disease, but next-generation sequencing technologies possibly will enable comprehensive whole-genome sequencing of individuals. However, sequencing the entire genome of a large group of affected individuals may still be prohibitively expensive for years to come, and alternative strategies are welcomed. Since there are only 1,973,844 bp of STR segments in human exons and 99% of them are shorter than 33 bp, it may be feasible to screen for rare mutations using selective resequencing of STR regions at a reasonable price and effort [##REF##17982454##23##].</p>" ]

[ "<title>Conclusion</title>", "<p>In summary, 92% of all human genes have STRs in their exons according to the definition used in this paper. Despite their short lengths and simple definition STRs capture a large amount of the known exonic indels and are significantly overrepresented in disease-related genes. These findings constitute STRs as an obvious target when screening for rare disease causing mutations, because of the relatively low amount of STRs in exons (1,973,844 bp in human; 1,544,242 bp in mouse) and the limited length of STR regions (99% are shorter than 33 bp).</p>" ]

[ "<p>This is an Open Access article distributed under the terms of the Creative Commons Attribution License (<ext-link ext-link-type=\"uri\" xlink:href=\"http://creativecommons.org/licenses/by/2.0\"/>), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.</p>", "<title>Background</title>", "<p>In recent years it has been demonstrated that structural variations, such as indels (insertions and deletions), are common throughout the genome, but the implications of structural variations are still not clearly understood. Long tandem repeats (e.g. microsatellites or simple repeats) are known to be hypermutable (indel-rich), but are rare in exons and only occasionally associated with diseases. Here we focus on short (imperfect) tandem repeats (STRs) which fall below the radar of conventional tandem repeat detection, and investigate whether STRs are targets for disease-related mutations in human exons. In particular, we test whether they share the hypermutability of the longer tandem repeats and whether disease-related genes have a higher STR content than non-disease-related genes.</p>", "<title>Results</title>", "<p>We show that validated human indels are extremely common in STR regions compared to non-STR regions. In contrast to longer tandem repeats, our definition of STRs found them to be present in exons of most known human genes (92%), 99% of all STR sequences in exons are shorter than 33 base pairs and 62% of all STR sequences are imperfect repeats. We also demonstrate that STRs are significantly overrepresented in disease-related genes in both human and mouse. These results are preserved when we limit the analysis to STRs outside known longer tandem repeats.</p>", "<title>Conclusion</title>", "<p>Based on our findings we conclude that STRs represent hypermutable regions in the human genome that are linked to human disease. In addition, STRs constitute an obvious target when screening for rare mutations, because of the relatively low amount of STRs in exons (1,973,844 bp) and the limited length of STR regions.</p>" ]

[ "<title>Abbreviations</title>", "<p>STR: Short Tandem Repeats.</p>", "<title>Authors' contributions</title>", "<p>BM and PV came up with the idea for the study, and did the bioinformatics analysis. BM made the statistically analysis and drafted the first version of the paper. CW, PV and BM contributed to the design of the study, interpretation of results and writing the paper.</p>", "<title>Supplementary Material</title>" ]

[ "<title>Acknowledgements</title>", "<p>We thank Roald Forsberg and Sharon Browning for helpful discussions, Enette Berndt Knudsen for technical assistance and David R. Shaw, the Jackson Laboratory, for helping to compile the list of mouse disease genes. Part of this study was done while BEM visited the Department of Statistics at The University of Auckland. CW is supported by the Danish Cancer Society and PV was supported by the Lundbeck Foundation, Denmark.</p>" ]

[ "<fig position=\"float\" id=\"F1\"><label>Figure 1</label><caption><p><bold>Example of a STR.</bold> A STR is a segment of DNA with a strong periodic pattern. A segment of DNA is defined as STR if (1) the minimum length is 9 bp, (2) a sequence motif (e.g., AT in ATATATATAT) is repeated at least three times, (3) there are only few base pairs that do not match the periodic motif (see Methods). The example shows an exonic STR region located in the nestin (<italic>NES</italic>) gene on chromosome 1.</p></caption></fig>", "<fig position=\"float\" id=\"F2\"><label>Figure 2</label><caption><p><bold>Distribution of indels inside versus outside STRs</bold>. Both insertions and deletions are more frequent inside (red bars) than outside (black bars) STRs (P-values shown above columns), in the entire genome as well as in exons only.</p></caption></fig>", "<fig position=\"float\" id=\"F3\"><label>Figure 3</label><caption><p><bold>Indel length versus STR period</bold>. Left: correlation between STR period and insertion lengths. Circle area is proportional to the number of observations. Right: correlation between STR period and deletion length. 25 indels longer than 9 bp are omitted. See Additional file ##SUPPL##0##1##: Table S2 for counts.</p></caption></fig>", "<fig position=\"float\" id=\"F4\"><label>Figure 4</label><caption><p><bold>Distribution of indel lengths in exons</bold>. Left: insertions of different lengths inside STRs (red bars) and outside STRs (black bars). Right: deletions of different lengths inside and outside STRs. A majority of all insertions have a length different from 3, 6 or 9 bp; both inside and outside STRs.</p></caption></fig>", "<fig position=\"float\" id=\"F5\"><label>Figure 5</label><caption><p><bold>STR content in the exons of human disease genes</bold>. Absolute STR amount for human reference genes and the four sets of disease genes with number of genes shown in parentheses. Genes are ranked by absolute STR content, with the STR poorest genes to the left. Note the virtually all disease genes harbour STRs in their exons.</p></caption></fig>", "<fig position=\"float\" id=\"F6\"><label>Figure 6</label><caption><p><bold>Distribution of STR periods</bold>. The proportion of STR of each period in the human (A) and mouse (C) genome; a segment of STR is always assigned to the lowest possible period. The proportion of STR of each period, for each of the disease sets in human (B) and mouse (D).</p></caption></fig>" ]

[ "<table-wrap position=\"float\" id=\"T1\"><label>Table 1</label><caption><p>Estimated STR overrepresentation in disease-related genes, relative to the proportion of STR in reference genes.</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td align=\"center\">Gene set</td><td align=\"left\">  #Genes</td><td align=\"center\">STR overrepresentation ^a</td><td align=\"center\">P-value</td></tr></thead><tbody><tr><td align=\"center\" colspan=\"4\"><bold>Human genes</bold></td></tr><tr><td align=\"left\">Reference set</td><td align=\"center\">11210</td><td align=\"center\">-</td><td align=\"center\">-</td></tr><tr><td align=\"left\">All diseases</td><td align=\"center\">2095</td><td align=\"center\">7.0 % [4.6, inf]</td><td align=\"center\">2.1 × 10^-7</td></tr><tr><td align=\"left\">Leukaemia</td><td align=\"center\">70</td><td align=\"center\">28.3 % [15.2, inf]</td><td align=\"center\">1.7 × 10^-4</td></tr><tr><td align=\"left\">Cancers</td><td align=\"center\">151</td><td align=\"center\">17.5 % [8.8, inf]</td><td align=\"center\">3.3 × 10^-4</td></tr><tr><td align=\"left\">Immune system diseases</td><td align=\"center\">52</td><td align=\"center\">16.5 % [3.0, inf]</td><td align=\"center\">2.1 × 10^-2</td></tr><tr><td align=\"center\" colspan=\"4\"><bold>Mouse genes</bold></td></tr><tr><td align=\"left\">Reference set</td><td align=\"center\">17077</td><td align=\"center\">-</td><td align=\"center\">-</td></tr><tr><td align=\"left\">Cancers</td><td align=\"center\">294</td><td align=\"center\">12.1 % [5.3, inf]</td><td align=\"center\">7.6 × 10^-4</td></tr><tr><td align=\"left\">Postnatal lethality</td><td align=\"center\">764</td><td align=\"center\">25.7 % [21.6, inf]</td><td align=\"center\">&lt;10^-12</td></tr></tbody></table></table-wrap>" ]

[]

[]

[]

[]

[]

[ "<supplementary-material content-type=\"local-data\" id=\"S1\"><caption><title>Additional file 1</title><p><bold>Supporting material</bold>. Supporting Figures S1-S4 and Supporting Tables S1-S5.</p></caption></supplementary-material>" ]

[ "<table-wrap-foot><p>^a95% confidence intervals are in brackets.</p></table-wrap-foot>" ]

[ "<graphic xlink:href=\"1471-2164-9-410-1\"/>", "<graphic xlink:href=\"1471-2164-9-410-2\"/>", "<graphic xlink:href=\"1471-2164-9-410-3\"/>", "<graphic xlink:href=\"1471-2164-9-410-4\"/>", "<graphic xlink:href=\"1471-2164-9-410-5\"/>", "<graphic xlink:href=\"1471-2164-9-410-6\"/>" ]

[ "<media xlink:href=\"1471-2164-9-410-S1.pdf\" mimetype=\"application\" mime-subtype=\"pdf\"><caption><p>Click here for file</p></caption></media>" ]

{ "acronym": [], "definition": [] }

[ "<title>Background</title>", "<p><italic>Conserved synteny </italic>is the (local) maintenance of gene content and order in certain chromosomal regions of related species. Several studies on chromosome evolution [##REF##12529304##1##, ####REF##15153998##2##, ##REF##16983148##3##, ##REF##17009864##4##, ##REF##17387144##5####17387144##5##] demonstrated that conserved synteny exists not only between closely-related species but also over very long evolutionary timescales. Long-range conserved synteny is a particularly frequent feature around developmentally important genes [##REF##17387144##5##], demonstrating that rearrangements are not an unbiased random process in genome evolution.</p>", "<p>Conserved synteny is, however, not only of interest as a phenomenon in genome evolution, but provides valuable practical information for the analysis of families of homologous genes. It is a long-standing problem in comparative genomics to identify orthologs, i.e. pairs of genes from two organisms that are separated from each other by a speciation event. In general, the task to distinguish true orthologs from paralogs cannot be solved based on pairwise comparisons. Gene loss, differences in evolutionary rates [##REF##15983874##6##], and convergent evolution often distort the sequence similarities to an extent that makes it impossible to determine orthology from the gene tree. Genomic linkage with genes whose orthology relationships are clearer (i.e. which are more conserved across species and have fewer in-species paralogs) than others can be exploited because linked genes likely share their duplication history. Local/tandem duplications place new copies into a new genomic context, large-scale duplications coordinately duplicate the genomic context and gene loss becomes obvious if it leaves large parts of the genomic context intact while erasing the gene of interest. Therefore, conserved synteny information may demonstrate the loss of a particular copy of a gene and hence put a restriction on which extant gene copies are potential orthologs. If the genomic context of duplicated genes has sufficiently diverged prior to a speciation event, synteny can even provide direct evidence for or against orthology.</p>", "<p>There are three basic approaches towards automated orthology identification.</p>", "<p>1. Similarity-based clustering methods. This group includes the popular reciprocal pairwise best hit approach and refinements (such as <monospace>Inparanoid</monospace>[##REF##11743721##7##, ####REF##15608241##8##, ##REF##18055500##9####18055500##9##]), as well as more sophisticated methods that initially represent homology as many-to-many relations. In [##UREF##0##10##], for instance, the \"homology graph\" graph of initial <monospace>blast</monospace> hits is refined by iteratively removing sub-optimal edges.</p>", "<p>2. Phylogenomics-based methods (such as the tree-based <monospace>Ensembl Compara</monospace>[##REF##18000006##11##] pipeline). These approaches first cluster homologous genes, then construct a gene phylogeny, attempt to reconcile it with a prescribed species tree and use the resulting mapping between gene-tree and species tree to assign orthology and paralogy relations. An alternative use of phylogenetic information is made by <monospace>PhyOP</monospace>, synonymous rate estimates to distinguish between orthologous and paralogous segments in closely related genomes [##REF##17009864##12##].</p>", "<p>3. Methods utilizing conserved synteny to infer true orthology between relatively recently diverged species. Methods range from whole genome alignments to combinations with similarity- and phylogenomics-based approaches. Examples are the commercial \"syntenic-anchor\" approach from Celera [##REF##15458983##13##], the former <monospace>Ensembl Compara</monospace> pipeline (prior the June 2006 release). Algorithms that are primarily designed to determine syntenic regions and break points between them also fall into this category [##REF##12421767##14##, ####UREF##1##15##, ##REF##15247098##16##, ##REF##16951135##17##, ##UREF##2##18####2##18##].</p>", "<p>Despite substantial improvements in this area, the automatically generated results are still far from being perfect, and the annotation provided by databases such as <monospace>Ensembl Compara</monospace>[##REF##18000006##11##] or <monospace>OrthoDB</monospace>[##REF##17947323##19##] are neither sufficiently complete nor sufficiently accurate for many applications. For instance, an in-depth study of lineage-specific differences in a family of transcription factors requires not only the complete complement of family members in each species, but also a flawless gene phylogeny (which implies a correct orthology assignment).</p>", "<p>The <monospace>SynBlast</monospace> tool is designed to assist the manual curation of such data and to focus on individual loci of interest. In contrast to most approaches to genome-wide orthology annotation, it does not operate on pre-determined gene (proteome) sets but it searches the nucleotide sequence of the entire target genome. Hence it does not exclude possible homologs only because they are missing from annotation tracks. Instead of attempting to automatically extract an assignment of orthologs and paralogs, <monospace>SynBlast</monospace> instead provides the user with detailed information on all plausible homologs and their genomic context. To this end, web-based graphical overviews that can easily be compared with one another are generated. While heuristic rules are employed by the software to propose a plausible rank-ordering of the homologs with the aim of determining the correct orthologs, its primary purpose is to present conflicting information to the researcher in such way that it facilitates the decision of a human curator.</p>" ]

[ "<title>Methods</title>", "<p>The vertebrate genomes were taken from <monospace>Ensembl</monospace> (release 42, Dec 2006). In case of the <italic>ParaHox </italic>application and also the <italic>Danio Hox </italic>example the new assembly version for zebrafish (Zv7, Apr 2007 from <monospace>Ensembl</monospace> release 46, Aug 2007) was used. The new <italic>Danio </italic>assembly was scanned with local <monospace>WU-BLAST</monospace> (<monospace>tblastn</monospace>, version 2.0 MP-WashU, 04-May-2006). All other <monospace>blast</monospace> searches were performed with local <monospace>tblastn</monospace> (<monospace>blastall</monospace> version 2.2.15 of the NCBI <monospace>BLAST</monospace> suite). Genome databases were used in repeat-masked form, and the minimum <italic>E</italic>-value was set to <italic>E </italic>= 10^-5or <italic>E </italic>= 10^-4. The maximal size of the target cluster was restricted to twice the size of the reference cluster for all applications (parameter <italic>L</italic>). The number of different proteins to be contained in a valid synteny region at minimum (parameter <italic>N</italic>) was set to 1 (Parahox application) or 4 (Hox application). The cutoff for the HSP chain score was set to 100. The cutoff for the maximal overlap (w.r.t. query coordinates) of neighboring consistent HSPs was set to 40 amino acid positions. The fraction of the score used for matches of loci with different orientation was set to 90 percent while the gap penalties were set to 10 (gap in reference sequence) and 2 (gap in target sequence). All scripts were written in <monospace>Perl</monospace> (v5.8.8) and executed on PC hardware running Linux.</p>", "<p>The <italic>intra-score </italic>is calculated once for each query protein <italic>s</italic>, and describes the relative difference of the best and the second-best hit onto the reference genome (i.e. for the closest-related paralogs). This is approximated by their bit-score differences, i.e.</p>", "<p></p>", "<p>where <italic>q</italic>_{<italic>s</italic>1}and <italic>q</italic>_{<italic>s</italic>2}are the two top-scoring target loci (i.e., HSP chains) within the reference genome. The more distant the closest paralogs in the reference, the more reliable is the assignment of orthologs from the target species.</p>", "<p>The <italic>inter-score </italic>is calculated for each assigned target locus <italic>t</italic>_{<italic>s </italic>}within a target genome, defined as its relative bit-score difference to the best reference hit locus <italic>q</italic>_{<italic>s</italic>1}:</p>", "<p></p>", "<p>Hence, the inter-score expresses how \"bad\" a putative ortholog hit to the target genome is w.r.t. the maximally expected score <italic>b</italic>(<italic>s</italic>, <italic>q</italic>_{<italic>s</italic>1}).</p>", "<p>The ratio of intra-score and inter-score, <italic>S</italic>_intra/<italic>S</italic>_inter, quantifies the quality of an inter-species (potentially orthologous) hit in relation to the similarity between paralogs in the reference genome. Therefore, it serves as a measure for the confidence in the orthology of the query and target locus.</p>", "<p>The <italic>(log)RatioSum </italic>score is defined as the sum of the (logarithms of the) intra-inter-score ratios of all target loci assigned within the gene order alignment.</p>" ]

[ "<title>Results</title>", "<title>Algorithms and Implementation</title>", "<title>Overview of the SynBlast pipeline</title>", "<p><monospace>SynBlast</monospace> is a \"semi-automatic\" pipeline that is implemented as a suite of <monospace>Perl</monospace> scripts (<monospace>SynBlast</monospace> package). In order to allow automatic retrieval of proteins from syntenic regions and comparison of assignments with existing annotations the <monospace>Ensembl</monospace> system and databases [##REF##18000006##11##] were chosen as standard reference sources. Therefore, the pipeline scripts make use of the <monospace>Ensembl Perl</monospace> API to retrieve reference annotation and sequences from the <monospace>Ensembl Core</monospace> databases as well as homology annotations (for comparison only) from the <monospace>Ensembl Compara</monospace> database.</p>", "<p>The workflow is summarized in Figure ##FIG##0##1##. It starts from a focal protein coding gene of interest whose homologs are to be detected.</p>", "<p>Step 1, in order to include synteny information, the adjacent protein coding genes within a certain genomic distance (flanking size, e.g. 1 Mb up- and downstream) are added to the reference set. The system requires both the sequences and positional information (orientation and (relative) start and end positions) of the reference genes. This information can be provided manually by the user in form of a text file containing tab-separated entries of gene identifiers and their genomic coordinates. Details are given in the <monospace>SynBlast</monospace> Tutorial, which is included in the Online Supplemental Material [##UREF##3##20##]. Alternatively, the corresponding files can be generated using the <monospace>getEnsemblProteins.pl</monospace> script which retrieves sequences and annotation information from <monospace>Ensembl</monospace> databases.</p>", "<p>Step 2 consists of translated-<monospace>blast</monospace> searches using all reference proteins as queries on a selection of genome databases as targets. Resulting <monospace>blast</monospace> hits are expected to be in tabular (<monospace>NCBI BLAST</monospace>) format for further processing. Again this step can be performed manually using any program that creates <monospace>blast</monospace>-like tabular output, including <monospace>NCBI BLAST</monospace>[##REF##9254694##21##], <monospace>WU-BLAST</monospace> (<ext-link ext-link-type=\"uri\" xlink:href=\"http://blast.wustl.edu\"/>), or <monospace>BLAT</monospace>[##REF##11932250##22##]. When the genome data is available locally in <monospace>NCBI BLAST</monospace> format, the script <monospace>doBlastJobs.pl</monospace> automatizes this step using local <monospace>NCBI BLAST</monospace>. It is needed to include the reference species as target genome as well in order to enable the subsequent normalization of <monospace>blast</monospace> scores and to detect possible paralogous clusters that the user should be aware of when interpreting final pipeline results. Those paralogous regions of the reference set should be used as reference in a subsequent pipeline run as well to avoid false positive orthology assignments.</p>", "<p>In step 3, we search for potential regions of conserved synteny (syntenic target blocks). To this end we collect <monospace>blast</monospace> hits that are located within regions of limited size on the target genome. The purpose of this filtering step is to extract candidate subsets of <monospace>blast</monospace> hits (or HSPs, high scoring pairs) that can be treated separately in the following. At this stage we do not consider gene order, but gene content information, i.e. a user-specified number of query-specific hits must be contained at minimum in each candidate subset. The procedure is implemented in the script <monospace>doSyntenyFilter.pl</monospace>, and is described in detail in the following section.</p>", "<p>In step 4, all detected candidate regions of conserved synteny from step 3 are evaluated w.r.t. their agreement with the reference gene order. The technical details of the <monospace>doSyntenyAlignment.pl</monospace> procedure are described below. The candidate syntenic regions are sorted according to a scoring scheme that combines sequence similarity, synteny information, and orthology versus paralogy information. The results are presented as HTML files in a web browser together with graphical representations of gene order alignment matrices and paths. Graphics such as those shown in Figure ##FIG##1##2## allow the user to readily identify small-scale rearrangements such as translocations or inversions.</p>", "<title>Extraction of syntenic target blocks – <monospace>doSyntenyFilter.pl</monospace></title>", "<p>A region of one of the target genomes is considered as a candidate for a syntenic target block if it contains blast hits from at least <italic>N </italic>different proteins of the query set within an interval of length at most <italic>L</italic>. The parameters <italic>N </italic>and <italic>L </italic>are set by the user, both either directly or indirectly (relative to the reference set). They reflect the expected rate of gene loss and the expected structural similarity. The maximal regions of contiguous <monospace>blast</monospace> hits fulfilling these criteria are extracted separately for each target sequence. Depending on the status of the genome assembly, the sequence can be on a chromosome, a scaffold, or a contig. Small scaffolds or contigs pose a problem for this step as a target block syntenic to the query region may be mapped to several different scaffolds. In this case, <monospace>SynBlast</monospace> reports two or more separate syntenic regions and/or misses parts of the regions if less than <italic>N </italic>query proteins map to some of the scaffold regions. Note that for some genomes allelic variants are assembled into different scaffolds. <monospace>SynBlast</monospace> then reports all these scaffolds and it is left to the user to recognize this.</p>", "<p>In its current implementation, the candidate subsets of <monospace>blast</monospace> hits are found by a sliding window approach. In addition to the number of query proteins ≥ <italic>N </italic>that have <monospace>blast</monospace> hits within a sequence window of size ≤ <italic>L </italic>also the sum of all maximal HSP bit-scores for these proteins is recorded. This yields a convenient measure to prefer the higher-scoring subsets when there are overlapping intervals (in particular if <italic>L </italic>and <italic>N </italic>are too small). We currently use a greedy approach that selects a specified number of target block intervals in decreasing order of the score sum and skips all intervals overlapping a previously selected interval by more than a specified threshold. For a detailed description of the various effects following changes to the parameter settings of <italic>L </italic>and <italic>N </italic>we refer to the <monospace>SynBlast</monospace> tutorial, which is provided as part of the Supplemental Material [##UREF##3##20##].</p>", "<title>Evaluation of syntenic target blocks via gene order alignment – <monospace>doSyntenyAlignment.pl</monospace></title>", "<p>In the fourth step of the pipeline, each of the syntenic target blocks (subsets of <monospace>blast</monospace> hits), resulting from step 3, is analyzed separately in comparison to the query region. This part of the pipeline consists of several sub-components, which we discuss separately, see also Figure ##FIG##2##3##.</p>", "<p>(i) The set of reference proteins is linearly ordered (by start/end or mean positions) into so-called \"query loci\". If the query region contains overlapping proteins, these are combined into a single query locus. Thus, a query locus may comprise more than one query protein (an example is the second query locus in Figure ##FIG##2##3A##).</p>", "<p>(ii) For each query protein we chain all corresponding HSPs into models of target loci. Each HSP consists of an interval <italic>α </italic>= [<italic>a</italic>', <italic>a</italic>\"] on the query sequence and a corresponding interval <italic>β </italic>= [<italic>b</italic>', <italic>b</italic>\"] on the target sequence. These intervals are linearly ordered for both query and target based on their coordinates.</p>", "<p>Intervals on the two strands of a target are treated separately. We furthermore take into account that HSPs must not be too far away from each other, i.e. the maximal genomic extension is restricted to either an absolute or query gene dependent length (locus size limit; see also Figure ##FIG##2##3-B1##). Again, groups of intervals that are too far apart from each other are treated separately.</p>", "<p>For each group of HSPs with common orientation we compute an optimal \"alignment\" of these lists of intervals using a variant of the Needleman-Wunsch algorithm, similar to Ref. [##REF##15032508##23##]. The query and target intervals of the HSPs, respectively, are considered as characters within the alignment in which a match occurs if both intervals belong to the same HSP. The score of the match is its bit-score. Pairs of (query/target) intervals that do not correspond to the same HSP are considered as mismatches with score -∞. Small negative scores are given to insertions and deletions, endgaps are treated as free. The resulting alignment then defines a consistent chaining of collinear HSPs, here called an \"HSP chain\" for short. It represents an overall hit for the respective query protein (approximate gene model) with a group score equal to the sum of bit-scores of all HSPs of the chain. A score threshold can be specified to eliminate spurious hits, which otherwise might lead to incorrect groupings of adjacent loci in the next sub step, see also Figure ##FIG##2##3-B5##. Each HSP chain is furthermore characterized by a start, end, and mean position.</p>", "<p>(iii) As in sub-step (i), we now define a linear order of HSP chains by grouping them according to start/end or mean positions into so-called target loci, see Figure ##FIG##2##3##. If there are HSP chains that reside within the same target locus, (i.e. have overlapping intervals with some HSP chain) while representing hits for the same query protein, only the top-scoring chain is kept, see Figure ##FIG##2##3-B6##. Thus, we finally end up with linearly ordered blocks of target loci each containing one or more (overlapping) query-specific HSP chains.</p>", "<p>(iv) As in sub-step (ii), we use a variant of the Needleman-Wunsch algorithm to obtain a maximum weight sequence of collinear pairs of query and target loci. As before, mismatches are prohibited. Matches are scored as the arithmetic mean of the scores of all matching individual query proteins that belong to the same query locus. Formally, for each pair (<italic>q</italic>, <italic>t</italic>) of query and target locus let <italic>ν</italic>_{<italic>qt </italic>}be the number of matching query proteins between the pair of loci. The corresponding similarity score is</p>", "<p></p>", "<p>where <italic>b</italic>(<italic>s</italic>, <italic>t</italic>_{<italic>s</italic>}) is the bit-score of query protein s with its match <italic>t</italic>_{<italic>s </italic>}on the target genome as determined in sub-step (ii). In contrast to step (ii), we do not exclude matches between items of different orientation. Instead, we use only a fraction of the score <italic>b</italic>(<italic>s</italic>, <italic>t</italic>_{<italic>s</italic>}) (adjustable parameter) to penalize those matches. Thus, swapped orientation of a target locus w.r.t. the orientation of its matching query locus within the reference set is generally allowed, but the match score of such a locus is reduced to a user-defined fraction (e.g. 90%). This parameter can also be set to 0, in which case matches with reversed direction are considered as not informative at all. The gene order alignment score is consequently calculated for both orientations of the target block relative to the reference set, and only the higher-scoring alignment is retained in subsequent steps.</p>", "<p>In addition to this absolute scoring we also compute relative weights <italic>b</italic>(<italic>s</italic>, <italic>t</italic>_{<italic>s</italic>})/<italic>b</italic>(<italic>s</italic>, <italic>q</italic>_{<italic>s</italic>}) where the absolute bit-score is scaled by the score obtained by matching the query protein <italic>s </italic>back to its genomic locus <italic>q</italic>_{<italic>s </italic>}within the reference genome. The value <italic>b</italic>(<italic>s</italic>, <italic>q</italic>_{<italic>s</italic>}) is a good approximation for the maximal <monospace>tblastn</monospace> score of a given query protein. The relative score is then used as match score during the gene order alignment. This ensures that the matches to each reference locus are scored relative to the information content of the locus.</p>", "<p>Since match scores are defined directly between loci we can conveniently combine the visualization of the alignment path and the scoring matrix, see Figure ##FIG##1##2##.</p>", "<p>(v) Finally, all evaluated target regions are compiled in a ranked list in browsable HTML format including graphical overviews of loci scoring matrices (dotplot) and alignments as well as an alignment table displaying additional information for assigned loci. Swapped orientation of a single target locus is indicated in the dotplot by a shaded dot. In the alignment graphics, an arrow with reversed orientation w.r.t. the arrow of its assigned query locus is used.</p>", "<p>The ranking can be based on the gene order alignment score (roughly the sum of (weighted) bit-scores for assigned target loci) or the <italic>(log)RatioSum </italic>score, which is calculated as the sum of (the logarithms of) intra-inter-score ratios of assigned target loci. This score ratio, which is described in detail in the \"Methods\" section, measures the ambiguity of orthology between two loci based on the existence of close paralogs within the reference. It has proven useful in the process of identifying true orthologs. In combination with the gene order alignment score this score makes it easier to distinguish between putative orthologous and paralogous hits or clusters.</p>", "<p>If reference and genome data are taken from <monospace>Ensembl</monospace> databases, <monospace>SynBlast</monospace> optionally retrieves the <monospace>Ensembl</monospace><monospace>Compara</monospace> homology annotations and the <monospace>Ensembl Core</monospace> protein coding gene annotations overlapping the target locus interval of the matching HSP chain identified by <monospace>SynBlast</monospace>, for comparison.</p>", "<title>Applications</title>", "<p>As a real-life test of <monospace>SynBlast</monospace>, we consider here the genomic clusters of vertebrate <italic>Hox </italic>and <italic>ParaHox </italic>genes. These genes code for homeodomain transcription factors that regulate the anterior/posterior patterning in most bilaterian animals [##REF##2569969##24##,##REF##1346368##25##]. <italic>Hox </italic>and <italic>ParaHox </italic>genes arose early in metazoan history from a single ancestral \"<italic>UrHox </italic>gene\" [##REF##11253066##26##]. After a few tandem duplication events, a large scale duplication lead to ancestral <italic>Hox </italic>and <italic>ParaHox </italic>clusters. While the ancestral <italic>ParaHox </italic>cluster remained largely unchanged, the evolution of its <italic>Hox </italic>counterpart was dominated by a series of tandem duplications. As a consequence, most bilaterians share at least eight distinct paralogous groups (8 in arthropods, and 13 or 14 in chordates) which retained high sequence similarity at the homeobox. Both <italic>Hox </italic>and <italic>ParaHox </italic>genes are usually organized in tightly linked clusters [##UREF##4##27##], with syntenic conservation extending even beyond the core clusters themselves. For instance, an additional homeobox gene, <italic>Evx</italic>, located at the 5' end of the <italic>Hox </italic>cluster can be seen as part of an extended <italic>Hox </italic>cluster, see [##REF##16341069##28##] for more details.</p>", "<p>The modern vertebrate genome arose from an ancestral chordate by means of two rounds of whole genome duplication [##REF##9254922##29##,##REF##16162861##30##]. Teleost fishes have undergone an additional round of genome duplication [##REF##10597637##31##,##REF##12618368##32##].</p>", "<p>Substantial loss of duplicated genes was the consequence of these duplication events. In the case of <italic>Hox </italic>clusters there is little doubt about the orthology relationships among the <italic>Hox </italic>genes of tetrapoda. In teleost fishes, however, the relationships of the duplicated <italic>Hox </italic>clusters between zebrafish and fugu have long been controversial, see [##REF##18202881##33##] for a discussion, and have only recently been resolved using a dense taxon sampling [##REF##16162861##30##]. It is well known that the relative order and orientation of <italic>Hox </italic>genes in their clusters have been highly conserved in vertebrate evolution, albeit there is substantial gene loss. The <italic>Hox </italic>clusters thus are an excellent test case to demonstrate the gene order alignment functionality of <monospace>SynBlast</monospace>.</p>", "<title>Vertebrate <italic>Hox </italic>clusters</title>", "<p>We used the four human <italic>Hox </italic>clusters as reference and searched the vertebrate target species with <monospace>SynBlast</monospace>. We consider here a diverse set of vertebrate genomes which contains both tetrapods (with 4 paralogous <italic>Hox </italic>clusters) and teleosts (with 8 paralogons). The cluster locations, gene inventories, and <monospace>SynBlast</monospace> scores are listed in Figure ##FIG##3##4##. In case of genomes with complete assemblies, the correct assignment of cluster orthology and the correct assignment of <italic>Hox </italic>gene identity is straightforward by visual inspection of the <monospace>SynBlast</monospace> cluster alignments, see Table ##TAB##0##1## for an example. Here, both the gene order alignment score and the <italic>logRatioSum </italic>score is suitable to assign cluster identity to the target loci in the zebrafish genome. However, the <italic>logRatioSum </italic>score clearly out-performs the gene order alignment score in case of the <italic>Danio Bb </italic>cluster. In combination, the two scores provide the best means to rank orthologous loci at the top. The zebrafish Zv7 assembly contains two inparalog copies <italic>DrCa1 </italic>and <italic>DrCa2 </italic>of the zebrafish <italic>HoxCa </italic>cluster. This is, however, certainly an assembly artifact and contradicts all of the existing literature, see e.g. [##REF##14707165##34##] and the references therein. <monospace>SynBlast</monospace> correctly retrieves both copies with comparable scores.</p>", "<p>In the case of incomplete assemblies, only partial clusters can be obtained. For instance, individual <italic>Hox </italic>loci of oppossum and chicken are located on small separate scaffolds. For the duplicated genomes of the five teleosts in our data set, we obtained all 7 <italic>Hox </italic>genes-containing paralogous clusters in agreement with the literature, see [##REF##18202881##33##,##REF##15967537##35##, ####REF##16162861##36##, ##REF##17845724##37##, ##REF##17553908##38####17553908##38##]. Since our query consisted of the <italic>Hox </italic>cluster only, we could of course not retrieve the 8th zebrafish paralogon, which is completely devoid of homeobox genes [##REF##16736008##39##].</p>", "<p>Several artifacts of preliminary genome assemblies further complicate the analysis: In the fugu <italic>HoxAa </italic>cluster we readily detected the artifactual breakage of the cluster into two fragments on scaf.12 and scaf.346. In the older Zv6 assembly of the zebrafish genome, some <italic>Hox </italic>clusters contained local rearrangements and obviously duplicated gene loci, in particular the <italic>HoxBb </italic>and <italic>HoxCb </italic>clusters. Most of these problems have been resolved in the most recent Zv7 assembly, while the <italic>Ca </italic>artifact has been newly introduced. Table ##TAB##1##2## summarizes the discrepancies of the <monospace>Ensembl Compara</monospace> annotation for the orthology assignments obtained using <monospace>SynBlast</monospace>, the latter conforming to the recent literature [##REF##18202881##33##].</p>", "<p>The very well-understood example of the <italic>Hox </italic>gene cluster demonstrates that the true orthology and paralogy relationships can be determined rather quickly and easily by means of a manual analysis with the assistance of <monospace>SynBlast</monospace>. Automatic orthology annotation pipelines, on the other hand, still produce unsatisfactory results despite recent progress.</p>", "<title>Teleost ParaHox clusters</title>", "<p>The <italic>ParaHox </italic>clusters of teleost fishes have long been used to contradict the whole genome duplication scenario because of a mainly unduplicated repertoire of <italic>ParaHox </italic>genes compared to other vertebrates. Even after the teleost-specific genome duplication had been broadly accepted, the small number of <italic>ParaHox </italic>genes in each cluster and the large amount of gene loss at this locus complicated attempts to decipher their duplication history. Knowledge about the number of paralogous <italic>Cdx </italic>genes and their assignment to paralogous groups is a good starting point for such a reconstruction. Two studies based on publicly available genome sequences arrived at different scenarios for the history of this particular <italic>ParaHox </italic>gene in teleost fishes [##REF##16619246##40##,##REF##16801555##41##]. While Prohaska et al. [##REF##16619246##40##] proposed the existence of a <italic>Cdx2 </italic>gene copy (at least for fugu and tetraodon), Mulley et al. [##REF##16801555##41##] concluded that both copies of <italic>Cdx2 </italic>were lost and suggested that this loss was compensated by two copies of <italic>Cdx1</italic>. A more recent analysis that uses additional sequence data [##REF##17822543##42##] settles the discrepancy in favor of [##REF##16801555##41##], supporting the retention of two <italic>Cdx1 </italic>genes in cichlids. The analysis of [##REF##17822543##42##] in part excludes zebrafish because of problems with the available genome assemblies. Here we demonstrate how <monospace>SynBlast</monospace> can be used to facilitate retrieval of candidate <italic>Cdx </italic>loci and cluster assignments in the zebrafish genome.</p>", "<p>In an intact <italic>ParaHox </italic>cluster, the <italic>Cdx </italic>gene is flanked by two <italic>ParaHox </italic>genes, i.e. <italic>Gsh </italic>and <italic>Pdx</italic>, and a number of genes of other gene families. According to [##REF##17822543##42##], the ancestral gnathostome <italic>ParaHox </italic>genes are organized in four clusters, designated A, B, C, and D in analogy to the <italic>Hox </italic>clusters (see Figure ##FIG##4##5##). The <italic>Cdx </italic>gene of the C cluster has been lost soon after the 1R/2R duplications [##REF##16801555##41##]. No organism with a fourth <italic>Cdx </italic>paralog resulting from this duplication event has yet been found. Therefore, only three of the four ancestral clusters each retained a <italic>Cdx </italic>gene: <italic>Cdx2 </italic>(cluster A), <italic>Cdx4 </italic>(cluster B), and <italic>Cdx1 </italic>(cluster D). As a consequence of the teleost genome duplication we would expect to find 8 <italic>ParaHox </italic>clusters, two A clusters (A1, A2), two B clusters (B1, B2) etc. and up to 6 <italic>Cdx </italic>genes in the 6 clusters A1/2, B1/2, and D1/2. We start our <monospace>SynBlast</monospace> search in the zebrafish genome with the four human <italic>ParaHox </italic>cluster regions.</p>", "<p>Table ##TAB##2##3## shows the scores for all pairs of the four human query loci with the 11 high-scoring target loci of zebrafish. The assignment of true orthologs is more challenging than in the case of <italic>Hox </italic>clusters but still revealing.</p>", "<p>One copy of <italic>ParaHoxA </italic>retained 13 of 24 genes flanking the <italic>Cdx2 </italic>locus even though <italic>Cdx2 </italic>itself was obviously lost. This is a case where gene loss can reliably be distinguished from missing data based on well-conserved synteny information (see Figure ##FIG##5##6##). The second copy retained only 5 of the 24 flanking genes. In line with the analysis of [##REF##16801555##41##,##REF##17822543##42##], we observe that the <italic>Cdx2 </italic>gene has been lost from both copies. We also observe that one of the two <italic>ParaHoxA </italic>contains the only copy of <italic>Gsh1 </italic>which is located at <bold>DrA2 </bold>(Chr.5), while the only copy of <italic>Pdx1 </italic>is located at <bold>DrA1 </bold>(Chr.24). Note that this information independently confirms the assignment of the two zebrafish <italic>ParaHoxA </italic>paralogs to the ancestral A cluster. <monospace>SynBlast</monospace> reports additional syntenic regions in the zebrafish genome that contain homologs of some of the genes of the <italic>HsA </italic>query. These are located on chromosomes 7, 14, 20, and 21, and can be assumed to be orthologs of the <italic>ParaHox </italic>B, C, and D clusters. In order to confirm this assumption, we also consider the remaining three human <italic>ParaHox </italic>regions as queries.</p>", "<p>The query with human <italic>ParaHoxB </italic>yields only poorly conserved synteny information. This can be due to the reorganization of this locus when it got translocated to the mammalian sex chromosome X, see [##REF##16619246##40##,##REF##16778056##43##] for details. Nevertheless, we obtain sufficient information from the linkage of the <italic>Cdx </italic>loci with <italic>chic1 </italic>to see that zebrafish has one <italic>Cdx4 </italic>locus, <bold>DrB1 </bold>(Chr.14). With the human <italic>ParaHoxC </italic>cluster, which contains the <italic>Gsh-2 </italic>gene as a query, two paralogous regions in the zebrafish genome can be identified. <bold>DrC1 </bold>on Chr.20 containing the only surviving copy of <italic>Gsh2</italic>, while a putative <bold>DrC2 </bold>locus on Chr.1 contains three high-scoring reference loci (<italic>fip1l1</italic>, <italic>chic2</italic>, and <italic>clock</italic>) and both neighbors of <italic>Gsh2</italic>, i.e. <italic>chic2 </italic>and <italic>pdgfrA</italic>, but is devoid of homeobox genes.</p>", "<p>Note that \"empty\" parahox clusters are not exceptional. Teleost fishes have also lost all homeobox genes in one of the <italic>HoxD </italic>paralogs (zebrafish, [##REF##16736008##39##]) or one of the <italic>HoxC </italic>paralogs (pufferfish, [##REF##16619246##40##]), respectively. Loss of a gene of interest can nevertheless be identified due to the retention of neighboring genes given sufficient conserved synteny.</p>", "<p>The assignment of orthologs to cluster <italic>HsParaHoxD </italic>is difficult. Conserved synteny information is relatively rare and only locally given, i.e. the orthologous hits for those query regions are scattered more or less across the target chromosome or even genome, which is probably due to extensive rearrangements. Nevertheless, one <italic>Cdx </italic>locus is linked to <italic>pdgfrB</italic>, a D cluster gene. <monospace>SynBlast</monospace> detects multiple fragments that map to two distinct zebrafish chromosomes. A plausible hypothesis is to interpret the three hits of Chr.14 at 22 M, 25 M, and 53 M as remnants of one dissolving cluster <bold>DrD1</bold>, while the two fragments of Chr.21 at 36 M and 43 M constitute the other <bold>DrD2 </bold>paralog.</p>", "<p>In summary, we have located the three retained <italic>Cdx </italic>genes in the highly fragmented zebrafish genome assembly, and we conclude that three <italic>Cdx </italic>genes were lost in the aftermath of the fish-specific genome duplication. Due to synteny information, the three <italic>Cdx </italic>genes can unambiguously be assigned to the paralog groups <italic>Cdx4 </italic>(one copy, B cluster) and <italic>Cdx1 </italic>(two copies, D clusters).</p>", "<title>Further Examples</title>", "<p>In addition to the difficult <italic>Hox </italic>and <italic>Parahox </italic>loci we have investigated several examples of human loci with extensive synteny in other vertebrates, some of which are included in the Online Supplemental Material for comparison. In these rather straightforward cases we encountered a rather common annotation problem. Homology-based protein annotation sometimes produces two (or even more) disconnected annotated fragments, in particular when evidence from different sources is used. Since these fragments are not recognized as parts of the same protein, they are subsequently interpreted as distinct homologs of the same protein, resulting in an erroneous \"within-species-paralog\" assignment. We observed that <monospace>SynBlast</monospace> correctly recognizes such disconnected fragments as belonging to the same query item in the HSP chaining step, and hence avoids these spurious \"paralogs\".</p>" ]

[ "<title>Discussion and Conclusion</title>", "<p>The <monospace>SynBlast</monospace> tool was developed to assist in the interactive preparation of high-quality orthology annotations. It uses synteny in addition to sequence similarity. A major difference to most other tools is that it does not operate on a \"proteome set\". Instead, it uses <monospace>tblastn</monospace> and a two-level alignment procedure to retrieve the homologs of a set of reference proteins. As a consequence, it is independent of gene predictions and annotations of the target genomes. Known or predicted protein sequences are required only for the query genome. This avoids in particular many of the problems with misannotations in the target genome that may confuse automatic pipelines.</p>", "<p>A major advantage of the synteny-based approach is that we also find fairly diverged homologs in a conserved context that would otherwise be discarded due to insufficient sequence similarity, see also [##REF##17888146##44##]. This allows the user to find supporting information for highly diverged genes or gene loss and to distinguish it from the failure to detect sequence similarity. As a consequence, we find that <monospace>SynBlast</monospace> is particularly useful to retrieve homologous regions in the presence of high rates of gene loss, such as after the teleost-specific genome duplication. Syntenic regions are found and gene losses can be identified even when the focal genes are lost from one or more paralogons. As demonstrated in the <italic>ParaHox </italic>cluster example, information on such loci is readily accessible using <monospace>SynBlast</monospace> and can be instrumental in deciphering complex duplication/loss scenarios. This is the case in particular when homologous genes that arose through several distinct duplication events are of interest, as in the case of homeobox clusters. Of course, in cases where synteny is not preserved, <monospace>SynBlast</monospace> cannot do better than a simple <monospace>blast</monospace> search. In such a case, the output of the program at least makes it easy for the user to identify cases of disintegrated synteny. To distinguish orthologous and paralogous regions, <monospace>SynBlast</monospace> provides two scoring schemes: one that attempts to evaluate the overall similarity of two syntenic regions (gene order alignment score), and alternatively the relative similarity in comparison to the most similar within-reference paralog (<italic>(log)RatioSum </italic>score). However, the <monospace>SynBlast</monospace> system was designed to aid a careful manual evaluation rather than to provide an automatic pipeline. Hence, it produces extensive graphical and tabular output of all regions in the target genomes that are potentially syntenic to the query region in the form of HTML pages, which also integrates the existing <monospace>Ensembl Compara</monospace> homology annotation for comparison. This renders the tool most useful when orthology annotation is not obvious and expert knowledge is required to reach a definitive conclusion.</p>" ]

[ "<title>Discussion and Conclusion</title>", "<p>The <monospace>SynBlast</monospace> tool was developed to assist in the interactive preparation of high-quality orthology annotations. It uses synteny in addition to sequence similarity. A major difference to most other tools is that it does not operate on a \"proteome set\". Instead, it uses <monospace>tblastn</monospace> and a two-level alignment procedure to retrieve the homologs of a set of reference proteins. As a consequence, it is independent of gene predictions and annotations of the target genomes. Known or predicted protein sequences are required only for the query genome. This avoids in particular many of the problems with misannotations in the target genome that may confuse automatic pipelines.</p>", "<p>A major advantage of the synteny-based approach is that we also find fairly diverged homologs in a conserved context that would otherwise be discarded due to insufficient sequence similarity, see also [##REF##17888146##44##]. This allows the user to find supporting information for highly diverged genes or gene loss and to distinguish it from the failure to detect sequence similarity. As a consequence, we find that <monospace>SynBlast</monospace> is particularly useful to retrieve homologous regions in the presence of high rates of gene loss, such as after the teleost-specific genome duplication. Syntenic regions are found and gene losses can be identified even when the focal genes are lost from one or more paralogons. As demonstrated in the <italic>ParaHox </italic>cluster example, information on such loci is readily accessible using <monospace>SynBlast</monospace> and can be instrumental in deciphering complex duplication/loss scenarios. This is the case in particular when homologous genes that arose through several distinct duplication events are of interest, as in the case of homeobox clusters. Of course, in cases where synteny is not preserved, <monospace>SynBlast</monospace> cannot do better than a simple <monospace>blast</monospace> search. In such a case, the output of the program at least makes it easy for the user to identify cases of disintegrated synteny. To distinguish orthologous and paralogous regions, <monospace>SynBlast</monospace> provides two scoring schemes: one that attempts to evaluate the overall similarity of two syntenic regions (gene order alignment score), and alternatively the relative similarity in comparison to the most similar within-reference paralog (<italic>(log)RatioSum </italic>score). However, the <monospace>SynBlast</monospace> system was designed to aid a careful manual evaluation rather than to provide an automatic pipeline. Hence, it produces extensive graphical and tabular output of all regions in the target genomes that are potentially syntenic to the query region in the form of HTML pages, which also integrates the existing <monospace>Ensembl Compara</monospace> homology annotation for comparison. This renders the tool most useful when orthology annotation is not obvious and expert knowledge is required to reach a definitive conclusion.</p>" ]

[ "<p>This is an Open Access article distributed under the terms of the Creative Commons Attribution License (<ext-link ext-link-type=\"uri\" xlink:href=\"http://creativecommons.org/licenses/by/2.0\"/>), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.</p>", "<title>Motivation</title>", "<p>In the last years more than 20 vertebrate genomes have been sequenced, and the rate at which genomic DNA information becomes available is rapidly accelerating. Gene duplication and gene loss events inherently limit the accuracy of orthology detection based on sequence similarity alone. Fully automated methods for orthology annotation do exist but often fail to identify individual members in cases of large gene families, or to distinguish missing data from traceable gene losses. This situation can be improved in many cases by including conserved synteny information.</p>", "<title>Results</title>", "<p>Here we present the <monospace>SynBlast</monospace> pipeline that is designed to construct and evaluate local synteny information. <monospace>SynBlast</monospace> uses the genomic region around a focal reference gene to retrieve candidates for homologous regions from a collection of target genomes and ranks them in accord with the available evidence for homology. The pipeline is intended as a tool to aid high quality manual annotation in particular in those cases where automatic procedures fail. We demonstrate how <monospace>SynBlast</monospace> is applied to retrieving orthologous and paralogous clusters using the vertebrate <italic>Hox </italic>and <italic>ParaHox </italic>clusters as examples.</p>", "<title>Software</title>", "<p>The <monospace>SynBlast</monospace> package written in <monospace>Perl</monospace> is available under the GNU General Public License at <ext-link ext-link-type=\"uri\" xlink:href=\"http://www.bioinf.uni-leipzig.de/Software/SynBlast/\"/>.</p>" ]

[ "<title>Availability and requirements</title>", "<p>The <monospace>SynBlast</monospace> package written in <monospace>Perl</monospace> is available under the GNU General Public License at <ext-link ext-link-type=\"uri\" xlink:href=\"http://www.bioinf.uni-leipzig.de/Software/SynBlast/\"/>. It requires a Unix-like environment and several add-on perl modules (<monospace>DBI, GD</monospace>) installed, as well as an installation of the <monospace>Ensembl Core</monospace> and <monospace>Ensembl Compara</monospace> APIs of the appropriate release version, see also the <monospace>SynBlast</monospace> tutorial [##UREF##3##20##] for installation issues. A local version of the NCBI BLAST suite, as well as the genome sequence databases of selected target species is needed to generate the genome-wide similarity search results as part of the pipeline.</p>", "<p><bold>Project name: </bold><monospace>SynBlast</monospace></p>", "<p><bold>Project home page: </bold><ext-link ext-link-type=\"uri\" xlink:href=\"http://www.bioinf.uni-leipzig.de/Software/SynBlast/\"/></p>", "<p><bold>Operating System: </bold>Unix/GNU Linux</p>", "<p><bold>Programming languages: </bold><monospace>Perl</monospace>, <monospace>bash</monospace></p>", "<p><bold>Other requirements: </bold>several add-on <monospace>perl</monospace> modules (<monospace>DBI, GD</monospace>), <monospace>Ensembl Core/Compara</monospace> API, <monospace>NCBI BLAST</monospace> (or similar).</p>", "<p><bold>License: </bold>GNU GPL version 2 or any later version</p>", "<title>Authors' contributions</title>", "<p>SJP and PFS designed the study, JL developed the <monospace>SynBlast</monospace> pipeline and tool, all authors closely collaborated in preparing the manuscript.</p>" ]

[ "<title>Acknowledgements</title>", "<p>We thank Jutta J. Roth, Sven Findeiß, and Stephan Bernhart for valuable comments and discussions. This work was supported in part by the <italic>Deutsche Forschungsgemeinschaft </italic>(DFG), Proj. No. STA850/6-1 and by the European Community FP-6 project \"EMBIO\" <ext-link ext-link-type=\"uri\" xlink:href=\"http://www-embio.ch.cam.ac.uk/\"/>.</p>" ]

[ "<fig position=\"float\" id=\"F1\"><label>Figure 1</label><caption><p><bold>SynBlast pipeline steps</bold>. (1) A focal protein coding gene of interest and its surrounding genes (within a certain flanking size) are selected as reference set. Protein sequences and genomic positional information are either compiled manually or retrieved from Ensembl using the tool getEnsemblProteins.pl. (2) tblastn searches of all reference proteins are performed against selected target genome databases. (3) The resulting blast hit tables are scanned for regions of possibly conserved gene content. These regions are stored in separate blast hit tables. (4) The resulting sets of possibly syntenically conserved blast hits are evaluated based on their sum of blast bit-scores obtained by means of a gene loci order alignment to the reference set. The final decision on orthology or paralogy of ranked syntenic blocks is left to the user.</p></caption></fig>", "<fig position=\"float\" id=\"F2\"><label>Figure 2</label><caption><p><bold>Dotplot graphic and detailed alignment graphic</bold>. The dotplot (top) visualizes the content of the alignment scoring matrix used to calculate the global alignment with free endgaps where the reference loci (p) and target loci (q) are arranged in rows and columns, respectively. Brightness of green color indicates the relative score value (rscore) of sequence similarity for a particular locus match. The dotted squares correspond to the optimal alignment path, which is shown in the alignment drawing (black box). Swapped orientation of a single target locus (w.r.t. a matching reference locus and the cluster orientation) is indicated by shaded boxes and opposite arrow orientation in the dotplot and alignment drawing, respectively. (Note that in this example the alignment path contains only aligned loci which share the same orientation.) The focal reference gene is highlighted by a light frame and orange color in the dotplot and alignment drawing, respectively.</p></caption></fig>", "<fig position=\"float\" id=\"F3\"><label>Figure 3</label><caption><p><bold>Evaluation of syntenic blocks</bold>. Panel <bold>(A) </bold>summarizes the mappings from query to target, panel <bold>(B) </bold>elaborates on particular cases. The query region (panel <bold>(A)</bold>, middle) contains a sequence of syntenic query loci (green), each representing one or more possibly overlapping query proteins <underline>(i)</underline>. Each candidate target region in the genomes of interest (panel <bold>(A) </bold>above and below the query locus), is identified by a set of blast hits, HSPs, (yellow). For each region, the following steps are performed: First, the set of query-specific HSPs is chained <underline>(ii)</underline>, resulting in one or more HSP chains that represent approximate protein models (small boxes). Filtering rules are applied that exclude individual HSPs from a chain for one of the following four reasons: <italic>(1) </italic>if the resulting chain exceeds the prescribed size limit for a locus (<bold>B1</bold>) [default: twice the length of the query locus]; <italic>(2) </italic>if it is inconsistent with a co-linear ordering of other HSPs in the chain (<bold>B2</bold>); <italic>(3) </italic>if it overlaps with another query interval by more than a specified threshold (<bold>B3</bold>) [default: 30aa]; and <italic>(4) </italic>if it lies on the opposite strand (<bold>B4</bold>). Chains of HSPs are excluded if they score below a threshold bit-score [default: 50] after filtering (<bold>B5</bold>). The retained HSP chains are grouped <underline>(iii)</underline> into target loci (big open boxes) that contain all HSP chains (irrespective of their orientation) with overlapping target intervals. For each target locus, only the highest scoring chain for each query protein is kept (<bold>B6</bold>). This results in a sequence of non-overlapping target loci (recall that one locus might represent one or more proteins) that can be aligned <underline>(iv)</underline> with the sequence of query loci in a gene order alignment (gray shading, optimal assignments are shown by darker shading). The score of this alignment is then used to rank the region relative to other syntenic target regions.</p></caption></fig>", "<fig position=\"float\" id=\"F4\"><label>Figure 4</label><caption><p><bold>Overview on pipeline results for vertebrate Hox clusters</bold>. SynBlast results and manually extracted orthologous cluster positions and identities for selected vertebrate species are listed. Unless otherwise indicated, positions correspond to assigned blast hits' intervals from <italic>Hox1 </italic>to <italic>Hox13</italic>/<italic>Evx </italic>hits in gene order alignment. Cluster orientation is w.r.t. the human reference clusters, which are HOXA9_ENSG00000078399_5e5; HOXB9_ENSG00000170689_2e5; HOXC9_ENSG00000180806_3e5; HOXD9_ENSG00000128709_5e5. Unassigned loci from the reference may be due to overlaps of chained HSPs. A '*' indicates loci that are absent in agreement with the literature [##REF##15972462##45##]. Data for Ensembl release 42 (Dec 2006).</p></caption></fig>", "<fig position=\"float\" id=\"F5\"><label>Figure 5</label><caption><p><bold>Schematic representation of genes flanking the Cdx gene locus in the human ParaHox clusters</bold>. Only linked genes relevant for the interpretation of the <italic>SynBlast </italic>output are shown.</p></caption></fig>", "<fig position=\"float\" id=\"F6\"><label>Figure 6</label><caption><p><bold>ParaHox example application</bold>. SynBlast was used to determine the four pairs of paralogous regions generated by the fish-specific genome duplication from the four gnathostome <italic>ParaHox </italic>regions. We show alignment dot-plots for the high-ranking hits (according to the gene order alignment score and <italic>logRatioSum </italic>score (in brackets)) of the four query regions against the zebrafish genome (Zv7, Ensembl release 46, Aug 2007). Parameters for the synteny filtering step were <italic>N </italic>= 1, <italic>L </italic>= 2. See text for more details.</p></caption></fig>" ]

[ "<table-wrap position=\"float\" id=\"T1\"><label>Table 1</label><caption><p><monospace>SynBlast</monospace> results for <italic>Danio rerio</italic>, <monospace>Ensembl</monospace> release 46 (Zv7), with the human <italic>Hox </italic>clusters as query.</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td align=\"right\">Reference</td><td align=\"right\"><italic>DrAa </italic><break/>chr.19 <break/>10.5 M</td><td align=\"right\"><italic>DrAb </italic><break/>chr.16 <break/>16 M</td><td align=\"right\"><italic>DrBa </italic><break/>chr.3 <break/>23 M</td><td align=\"right\"><italic>DrBb </italic><break/>chr.12 <break/>26.5 M</td><td align=\"right\"><italic>DrCa1 </italic><break/>chr.23 <break/>33.7 M</td><td align=\"right\"><italic>DrCa2 </italic><break/>chr.23 <break/>35.2 M</td><td align=\"right\"><italic>DrCb </italic><break/>chr.11 <break/>0.6 M</td><td align=\"right\"><italic>DrDa </italic><break/>chr.9 <break/>2 M</td></tr></thead><tbody><tr><td align=\"right\"><italic>HsA</italic></td><td align=\"right\"><bold>2.58</bold></td><td align=\"right\"><bold>3.29</bold></td><td align=\"right\">-0.76</td><td align=\"right\">-0.39</td><td align=\"right\">-1.98</td><td align=\"right\">-1.98</td><td align=\"right\">-0.4</td><td align=\"right\">-0.74</td></tr><tr><td align=\"right\"><italic>HOXA9</italic></td><td align=\"right\"><bold>5581</bold></td><td align=\"right\"><bold>4073</bold></td><td align=\"right\">4690</td><td align=\"right\">2119</td><td align=\"right\">3322</td><td align=\"right\">3318</td><td align=\"right\">1490</td><td align=\"right\">3269</td></tr><tr><td align=\"right\">chr.7</td><td align=\"right\"><bold>2/1</bold></td><td align=\"right\"><bold>1/3</bold></td><td align=\"right\">7/2</td><td align=\"right\">3/7</td><td align=\"right\">9/4</td><td align=\"right\">10/5</td><td align=\"right\">4/8</td><td align=\"right\">6/6</td></tr><tr><td/><td/><td/><td/><td/><td/><td/><td/><td/></tr><tr><td align=\"right\"><italic>HsB</italic></td><td align=\"right\">-1.14</td><td align=\"right\">-0.01</td><td align=\"right\"><bold>3.13</bold></td><td align=\"right\"><bold>0.42</bold></td><td align=\"right\">-1.66</td><td align=\"right\">-1.69</td><td align=\"right\">-0.64</td><td align=\"right\">-0.48</td></tr><tr><td align=\"right\"><italic>HOXB9</italic></td><td align=\"right\">2702</td><td align=\"right\">1342</td><td align=\"right\"><bold>6201</bold></td><td align=\"right\"><bold>1647</bold></td><td align=\"right\">3008</td><td align=\"right\">2982</td><td align=\"right\">1003</td><td align=\"right\">1678</td></tr><tr><td align=\"right\">chr.17</td><td align=\"right\">6/4</td><td align=\"right\">3/7</td><td align=\"right\"><bold>1/1</bold></td><td align=\"right\"><bold>2/6</bold></td><td align=\"right\">7/2</td><td align=\"right\">8/3</td><td align=\"right\">5/8</td><td align=\"right\">4/5</td></tr><tr><td/><td/><td/><td/><td/><td/><td/><td/><td/></tr><tr><td align=\"right\"><italic>HsC</italic></td><td align=\"right\">-0.89</td><td align=\"right\">-1.51</td><td align=\"right\">-0.26</td><td align=\"right\">-0.34</td><td align=\"right\"><bold>6.44</bold></td><td align=\"right\"><bold>7.58</bold></td><td align=\"right\"><bold>0.22</bold></td><td align=\"right\">-2.51</td></tr><tr><td align=\"right\"><italic>HOXC9</italic></td><td align=\"right\">2361</td><td align=\"right\">2323</td><td align=\"right\">3108</td><td align=\"right\">1346</td><td align=\"right\"><bold>8687</bold></td><td align=\"right\"><bold>6537</bold></td><td align=\"right\"><bold>4150</bold></td><td align=\"right\">3798</td></tr><tr><td align=\"right\">chr.12</td><td align=\"right\">7/6</td><td align=\"right\">8/7</td><td align=\"right\">5/5</td><td align=\"right\">6/8</td><td align=\"right\"><bold>2/1</bold></td><td align=\"right\"><bold>1/2</bold></td><td align=\"right\"><bold>3/3</bold></td><td align=\"right\">9/4</td></tr><tr><td/><td/><td/><td/><td/><td/><td/><td/><td/></tr><tr><td align=\"right\"><italic>HsD</italic></td><td align=\"right\">-0.92</td><td align=\"right\">-0.63</td><td align=\"right\">-0.85</td><td align=\"right\">-0.6</td><td align=\"right\">-0.41</td><td align=\"right\">-0.41</td><td align=\"right\">-0.71</td><td align=\"right\"><bold>1.76</bold></td></tr><tr><td align=\"right\"><italic>HOXD9</italic></td><td align=\"right\">2799</td><td align=\"right\">1811</td><td align=\"right\">2660</td><td align=\"right\">871</td><td align=\"right\">3017</td><td align=\"right\">3013</td><td align=\"right\">1303</td><td align=\"right\"><bold>4326</bold></td></tr><tr><td align=\"right\">chr.2</td><td align=\"right\">8/4</td><td align=\"right\">5/6</td><td align=\"right\">7/5</td><td align=\"right\">4/8</td><td align=\"right\">3/2</td><td align=\"right\">2/3</td><td align=\"right\">6/7</td><td align=\"right\"><bold>1/1</bold></td></tr></tbody></table></table-wrap>", "<table-wrap position=\"float\" id=\"T2\"><label>Table 2</label><caption><p>Incomplete and erroneous <monospace>Ensembl Compara</monospace> orthology annotations for vertebrate <italic>Hox </italic>cluster loci.</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td align=\"left\">Ref. cluster</td><td align=\"center\" colspan=\"3\">Genes</td><td align=\"center\" colspan=\"2\">Clusters</td></tr><tr><td colspan=\"1\"><hr/></td><td colspan=\"3\"><hr/></td><td colspan=\"2\"><hr/></td></tr><tr><td/><td align=\"right\">Δ</td><td align=\"right\"><italic>β</italic></td><td align=\"right\"><italic>n</italic></td><td align=\"right\"><italic>M</italic></td><td align=\"right\"><italic>K</italic></td></tr></thead><tbody><tr><td align=\"left\"><italic>HoxA</italic></td><td align=\"right\">11</td><td align=\"right\">2</td><td align=\"right\">148</td><td align=\"right\">5</td><td align=\"right\">16</td></tr><tr><td align=\"left\"><italic>HoxB</italic></td><td align=\"right\">25</td><td align=\"right\">2</td><td align=\"right\">118</td><td align=\"right\">12</td><td align=\"right\">16</td></tr><tr><td align=\"left\"><italic>HoxC</italic></td><td align=\"right\">11</td><td align=\"right\">2</td><td align=\"right\">88</td><td align=\"right\">6</td><td align=\"right\">10</td></tr><tr><td align=\"left\"><italic>HoxD</italic></td><td align=\"right\">20</td><td align=\"right\">7</td><td align=\"right\">113</td><td align=\"right\">13</td><td align=\"right\">16</td></tr></tbody></table></table-wrap>", "<table-wrap position=\"float\" id=\"T3\"><label>Table 3</label><caption><p><monospace>SynBlast</monospace> results for <italic>Danio rerio</italic>, <monospace>Ensembl</monospace> release 46 (Aug 2007), Zv7 assembly with the human <italic>ParaHox </italic>clusters as query.</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td align=\"right\">Reference</td><td align=\"right\"><italic>DrA1 </italic><break/>chr.24 <break/>20 M <break/><italic>pdx1</italic></td><td align=\"right\"><italic>DrA2 </italic><break/>chr.5 <break/>60 M <break/><italic>gsh1</italic></td><td align=\"right\"><italic>DrB </italic><break/>chr.7 <break/>50 M</td><td align=\"right\"><italic>DrB1 </italic><break/>chr.14 <break/>37 M <break/><italic>cdx4</italic></td><td align=\"right\"><italic>DrC1 </italic><break/>chr.20 <break/>20 M <break/><italic>gsh2</italic></td><td align=\"right\"><italic>DrC2 </italic><break/>chr.1 <break/>10 M</td><td align=\"right\"><italic>DrD1</italic>? <break/>chr.14 <break/>53 M</td><td align=\"right\"><italic>DrD1</italic>? <break/>chr.14 <break/>22 M</td><td align=\"right\"><italic>DrD1 </italic><break/>chr.14 <break/>25 M <break/><italic>cdx1a</italic></td><td align=\"right\"><italic>DrD2 </italic><break/>chr.21 <break/>43 M <break/><italic>CDX1</italic></td><td align=\"right\"><italic>DrD2</italic>? <break/>chr.21 <break/>36 M</td></tr></thead><tbody><tr><td align=\"right\"><italic>HsA/C1</italic></td><td align=\"right\"><bold>2.49</bold></td><td align=\"right\"><bold>0.25</bold></td><td align=\"right\">0.52</td><td align=\"right\">0.05</td><td align=\"right\">-0.65</td><td align=\"right\">-</td><td align=\"right\">-</td><td align=\"right\">-</td><td align=\"right\">1.3</td><td align=\"right\">-</td><td align=\"right\">(0.13)</td></tr><tr><td align=\"right\"><italic>CDX2</italic></td><td align=\"right\"><bold>8343</bold></td><td align=\"right\"><bold>2291</bold></td><td align=\"right\">1605</td><td align=\"right\">1537</td><td align=\"right\">1922</td><td align=\"right\">-</td><td align=\"right\">-</td><td align=\"right\">-</td><td align=\"right\">1757</td><td align=\"right\">-</td><td align=\"right\">(1112)</td></tr><tr><td align=\"right\">chr.13</td><td align=\"right\"><bold>2/1</bold></td><td align=\"right\"><bold>7/2</bold></td><td align=\"right\">5/5</td><td align=\"right\">9/6</td><td align=\"right\">40/3</td><td align=\"right\">-</td><td align=\"right\">-</td><td align=\"right\">-</td><td align=\"right\">3/4</td><td align=\"right\">-</td><td align=\"right\">(8/8)</td></tr><tr><td/><td/><td/><td/><td/><td/><td/><td/><td/><td/><td/><td/></tr><tr><td align=\"right\"><italic>HsB/C2</italic></td><td align=\"right\">(-0.41)</td><td align=\"right\">(-0.7)</td><td align=\"right\"><bold>2.53</bold></td><td align=\"right\"><bold>0.47</bold></td><td align=\"right\">(-0.04)</td><td align=\"right\">-0.02</td><td align=\"right\">-</td><td align=\"right\">-</td><td align=\"right\">-0.75</td><td align=\"right\">-0.14</td><td align=\"right\">-</td></tr><tr><td align=\"right\"><italic>CDX4</italic></td><td align=\"right\">(237)</td><td align=\"right\">(139)</td><td align=\"right\"><bold>2934</bold></td><td align=\"right\"><bold>1192</bold></td><td align=\"right\">(114)</td><td align=\"right\">669</td><td align=\"right\">-</td><td align=\"right\">-</td><td align=\"right\">543</td><td align=\"right\">183</td><td align=\"right\">-</td></tr><tr><td align=\"right\">chr.X</td><td align=\"right\">(19/25)</td><td align=\"right\">(31/38)</td><td align=\"right\"><bold>1/1</bold></td><td align=\"right\"><bold>3/2</bold></td><td align=\"right\">(7/41)</td><td align=\"right\">5/7</td><td align=\"right\">-</td><td align=\"right\">-</td><td align=\"right\">34/10</td><td align=\"right\">12/32</td><td align=\"right\">-</td></tr><tr><td/><td/><td/><td/><td/><td/><td/><td/><td/><td/><td/><td/></tr><tr><td align=\"right\"><italic>HsC/C3</italic></td><td align=\"right\">-1.47</td><td align=\"right\">-1.56</td><td align=\"right\">-0.33</td><td align=\"right\">-2.52</td><td align=\"right\"><bold>3.23</bold></td><td align=\"right\"><bold>0.91</bold></td><td align=\"right\">-</td><td align=\"right\">-</td><td align=\"right\">-0.46</td><td align=\"right\">-</td><td align=\"right\">(-0.37)</td></tr><tr><td align=\"right\"><italic>GSH2</italic></td><td align=\"right\">1478</td><td align=\"right\">507</td><td align=\"right\">1424</td><td align=\"right\">1595</td><td align=\"right\"><bold>4756</bold></td><td align=\"right\"><bold>1880</bold></td><td align=\"right\">-</td><td align=\"right\">-</td><td align=\"right\">469</td><td align=\"right\">-</td><td align=\"right\">(488)</td></tr><tr><td align=\"right\">chr.4</td><td align=\"right\">54/4</td><td align=\"right\">55/7</td><td align=\"right\">19/5</td><td align=\"right\">56/3</td><td align=\"right\"><bold>1/1</bold></td><td align=\"right\"><bold>3/2</bold></td><td align=\"right\">-</td><td align=\"right\">-</td><td align=\"right\">26/9</td><td align=\"right\">-</td><td align=\"right\">(23/8)</td></tr><tr><td/><td/><td/><td/><td/><td/><td/><td/><td/><td/><td/><td/></tr><tr><td align=\"right\"><italic>HsD/C4</italic></td><td align=\"right\">-0.6</td><td align=\"right\">(-2.77)</td><td align=\"right\">(-3.57)</td><td align=\"right\">-3.44</td><td align=\"right\">-3.31</td><td align=\"right\">(-0.06)</td><td align=\"right\"><bold>0.79</bold></td><td align=\"right\"><bold>0.03</bold></td><td align=\"right\"><bold>-0.41</bold></td><td align=\"right\"><bold>0.18</bold></td><td align=\"right\"><bold>5.27</bold></td></tr><tr><td align=\"right\"><italic>CDX1</italic></td><td align=\"right\">439</td><td align=\"right\">(254)</td><td align=\"right\">(2238)</td><td align=\"right\">953</td><td align=\"right\">940</td><td align=\"right\">(991)</td><td align=\"right\"><bold>1796</bold></td><td align=\"right\"><bold>1514</bold></td><td align=\"right\"><bold>846</bold></td><td align=\"right\"><bold>1107</bold></td><td align=\"right\"><bold>2550</bold></td></tr><tr><td align=\"right\">chr.5</td><td align=\"right\">13/41</td><td align=\"right\">(24/57)</td><td align=\"right\">(42/2)</td><td align=\"right\">41/12</td><td align=\"right\">37/13</td><td align=\"right\">(7/11)</td><td align=\"right\"><bold>2/4</bold></td><td align=\"right\"><bold>6/5</bold></td><td align=\"right\"><bold>11/16</bold></td><td align=\"right\"><bold>4/9</bold></td><td align=\"right\"><bold>1/1</bold></td></tr></tbody></table></table-wrap>" ]

[ "<disp-formula id=\"bmcM1\"><label>(1)</label><mml:math xmlns:mml=\"http://www.w3.org/1998/Math/MathML\" id=\"M1\" name=\"1471-2105-9-351-i1\" overflow=\"scroll\"><mml:semantics><mml:mrow><mml:mi>S</mml:mi><mml:mo stretchy=\"false\">(</mml:mo><mml:mi>q</mml:mi><mml:mo>,</mml:mo><mml:mi>t</mml:mi><mml:mo stretchy=\"false\">)</mml:mo><mml:mo>=</mml:mo><mml:mrow><mml:mo>{</mml:mo><mml:mrow><mml:mtable columnalign=\"left\"><mml:mtr columnalign=\"left\"><mml:mtd columnalign=\"left\"><mml:mrow><mml:mfrac><mml:mn>1</mml:mn><mml:mrow><mml:msub><mml:mi>ν</mml:mi><mml:mrow><mml:mi>q</mml:mi><mml:mi>t</mml:mi></mml:mrow></mml:msub></mml:mrow></mml:mfrac><mml:mstyle displaystyle=\"true\"><mml:munder><mml:mo>∑</mml:mo><mml:mrow><mml:mi>s</mml:mi><mml:mo>∈</mml:mo><mml:mi>q</mml:mi></mml:mrow></mml:munder><mml:mrow><mml:mi>b</mml:mi><mml:mo stretchy=\"false\">(</mml:mo><mml:mi>s</mml:mi><mml:mo>,</mml:mo><mml:msub><mml:mi>t</mml:mi><mml:mi>s</mml:mi></mml:msub><mml:mo stretchy=\"false\">)</mml:mo><mml:mo>,</mml:mo></mml:mrow></mml:mstyle></mml:mrow></mml:mtd><mml:mtd columnalign=\"left\"><mml:mrow><mml:mtext>if </mml:mtext><mml:msub><mml:mi>ν</mml:mi><mml:mrow><mml:mi>q</mml:mi><mml:mi>t</mml:mi></mml:mrow></mml:msub><mml:mo>≥</mml:mo><mml:mn>1</mml:mn></mml:mrow></mml:mtd></mml:mtr><mml:mtr columnalign=\"left\"><mml:mtd columnalign=\"left\"><mml:mrow><mml:mo>−</mml:mo><mml:mi>∞</mml:mi></mml:mrow></mml:mtd><mml:mtd columnalign=\"left\"><mml:mrow><mml:mtext>otherwise</mml:mtext></mml:mrow></mml:mtd></mml:mtr></mml:mtable></mml:mrow></mml:mrow></mml:mrow></mml:semantics></mml:math></disp-formula>", "<disp-formula id=\"bmcM2\"><label>(2)</label><mml:math xmlns:mml=\"http://www.w3.org/1998/Math/MathML\" id=\"M2\" name=\"1471-2105-9-351-i2\" overflow=\"scroll\"><mml:semantics><mml:mrow><mml:msub><mml:mi>S</mml:mi><mml:mrow><mml:mtext>intra</mml:mtext></mml:mrow></mml:msub><mml:mo stretchy=\"false\">(</mml:mo><mml:mi>s</mml:mi><mml:mo stretchy=\"false\">)</mml:mo><mml:mo>=</mml:mo><mml:mfrac><mml:mrow><mml:mi>b</mml:mi><mml:mo stretchy=\"false\">(</mml:mo><mml:mi>s</mml:mi><mml:mo>,</mml:mo><mml:msub><mml:mi>q</mml:mi><mml:mrow><mml:mi>s</mml:mi><mml:mn>1</mml:mn></mml:mrow></mml:msub><mml:mo stretchy=\"false\">)</mml:mo><mml:mo>−</mml:mo><mml:mi>b</mml:mi><mml:mo stretchy=\"false\">(</mml:mo><mml:mi>s</mml:mi><mml:mo>,</mml:mo><mml:msub><mml:mi>q</mml:mi><mml:mrow><mml:mi>s</mml:mi><mml:mn>2</mml:mn></mml:mrow></mml:msub><mml:mo stretchy=\"false\">)</mml:mo></mml:mrow><mml:mrow><mml:mi>b</mml:mi><mml:mo stretchy=\"false\">(</mml:mo><mml:mi>s</mml:mi><mml:mo>,</mml:mo><mml:msub><mml:mi>q</mml:mi><mml:mrow><mml:mi>s</mml:mi><mml:mn>1</mml:mn></mml:mrow></mml:msub><mml:mo stretchy=\"false\">)</mml:mo></mml:mrow></mml:mfrac></mml:mrow></mml:semantics></mml:math></disp-formula>", "<disp-formula id=\"bmcM3\"><label>(3)</label><mml:math xmlns:mml=\"http://www.w3.org/1998/Math/MathML\" id=\"M3\" name=\"1471-2105-9-351-i3\" overflow=\"scroll\"><mml:semantics><mml:mrow><mml:msub><mml:mi>S</mml:mi><mml:mrow><mml:mtext>inter</mml:mtext></mml:mrow></mml:msub><mml:mo stretchy=\"false\">(</mml:mo><mml:msub><mml:mi>t</mml:mi><mml:mi>s</mml:mi></mml:msub><mml:mo stretchy=\"false\">)</mml:mo><mml:mo>=</mml:mo><mml:mfrac><mml:mrow><mml:mi>b</mml:mi><mml:mo stretchy=\"false\">(</mml:mo><mml:mi>s</mml:mi><mml:mo>,</mml:mo><mml:msub><mml:mi>q</mml:mi><mml:mrow><mml:mi>s</mml:mi><mml:mn>1</mml:mn></mml:mrow></mml:msub><mml:mo stretchy=\"false\">)</mml:mo><mml:mo>−</mml:mo><mml:mi>b</mml:mi><mml:mo stretchy=\"false\">(</mml:mo><mml:mi>s</mml:mi><mml:mo>,</mml:mo><mml:msub><mml:mi>t</mml:mi><mml:mi>s</mml:mi></mml:msub><mml:mo stretchy=\"false\">)</mml:mo></mml:mrow><mml:mrow><mml:mi>b</mml:mi><mml:mo stretchy=\"false\">(</mml:mo><mml:mi>s</mml:mi><mml:mo>,</mml:mo><mml:msub><mml:mi>q</mml:mi><mml:mrow><mml:mi>s</mml:mi><mml:mn>1</mml:mn></mml:mrow></mml:msub><mml:mo stretchy=\"false\">)</mml:mo></mml:mrow></mml:mfrac></mml:mrow></mml:semantics></mml:math></disp-formula>" ]

[]

[]

[]

[]

[]

[ "<table-wrap-foot><p>The <italic>logRatioSum </italic>score, the gene order alignment score, and the corresponding ranks are given. Putative orthologs are depicted in bold. In combination, the two scores provide the best means to rank orthologous loci at the top. In most cases, identification of orthologous regions is unambiguous. For completeness we list both copies of the artifactually duplicated DrCa cluster of chr.23.</p></table-wrap-foot>", "<table-wrap-foot><p>We list the total number <italic>n </italic>of <italic>Hox </italic>and <italic>Evx </italic>genes in the dataset; the number Δ of <italic>Hox </italic>and <italic>Evx </italic>gene orthology assignments (see Figure 4 that are well-supported (query coverage &gt; 30%) by <monospace>SynBlast</monospace> but that are missing in the <monospace>Compara</monospace> annotation; the number <italic>β </italic>of well-supported assignments with incorrect annotations in <monospace>Compara</monospace> (paralog). We furthermore list the number <italic>M </italic>of the <italic>K Hox </italic>clusters in the dataset which contain apparently missing and/or erroneous <monospace>Compara</monospace> annotations. All data refer to <monospace>Ensembl</monospace> release 42 (Dec 2006).</p></table-wrap-foot>", "<table-wrap-foot><p>The <italic>logRatioSum </italic>score, the gene order alignment score, and the corresponding ranks are given. Putative orthologs are depicted in bold. In combination, the two scores provide the best means to rank orthologous loci at the top. Numbers in parentheses indicate that the target region (column) is only approximately matched, and/or that only a single query gene was found. See text for more detail.</p></table-wrap-foot>" ]

[ "<graphic xlink:href=\"1471-2105-9-351-1\"/>", "<graphic xlink:href=\"1471-2105-9-351-2\"/>", "<graphic xlink:href=\"1471-2105-9-351-3\"/>", "<graphic xlink:href=\"1471-2105-9-351-4\"/>", "<graphic xlink:href=\"1471-2105-9-351-5\"/>", "<graphic xlink:href=\"1471-2105-9-351-6\"/>" ]

[]

{ "acronym": [], "definition": [] }

[ "<title>Background</title>", "<p>In recent decades, extraordinary advances in biochemistry and molecular biology have led to an unprecedented level of understanding biological systems at the molecular level. The complexity of cellular pathways and networks often makes it difficult or impossible to reliably predict the behavior of a system from knowledge of its components, and thus there is considerable interest in formulation of quantitative, predictive mathematical models of cellular functions. Such efforts, collectively described by such terms as systems biology and <italic>in silico </italic>biology [##REF##15724281##1##, ####REF##14744094##2##, ##REF##16971329##3##, ##REF##11701654##4##, ##REF##11872829##5##, ##REF##12432404##6##, ##REF##15113093##7##, ##REF##11283699##8##, ##REF##16714341##9####16714341##9##], aim in the long term toward goals such as predicting the effects of drugs or other interventions on the state of diseased cells, and enhancing our fundamental understanding of how cells respond to stimuli and regulate their internal environments.</p>", "<p>The internal dynamics of cells are driven by the kinetics of a complex set of biochemical reactions: the state of the cell may be viewed as the numbers and binding states of all species of interest, and the time evolution of that state is defined by how those species react with one another. A central challenge in cellular modelling is to formulate correct biochemical reaction schemes to represent a process of interest, and then to populate the reaction system with appropriate rate constants [##REF##11872829##5##, ####REF##12432404##6##, ##REF##15113093##7##, ##REF##11283699##8##, ##REF##16714341##9####16714341##9##]. Within this effort, two persistent difficulties arise: populating mathematical models based on incomplete experimental information [##UREF##0##10##,##REF##12145321##11##]; and the computational demands of simulating the resulting systems, which can grow large for even moderately complex processes.</p>", "<p>We have previously carried out a study aimed at the first of these problems, in which we used bulk expression data from <italic>Escherichia coli </italic>to deduce the numbers of free RNA polymerases available to transcribe a target gene of interest [##UREF##0##10##]; this information is not currently experimentally available, with bulk studies [##UREF##1##12##] able to provide the average numbers of each enzyme type but not to determine how many are \"tied up\" in the cell, transcribing other genes, at any given time. When simulating the expression of a gene or network of genes, whether an engineered or \"synthetic\" system [##REF##11849948##13##, ####REF##15995697##14##, ##REF##12432407##15##, ##REF##14527313##16##, ##REF##16978856##17##, ##UREF##2##18####2##18##], or a natural one [##REF##14744094##2##,##REF##12432404##6##,##REF##11283699##8##,##REF##15562315##19##, ####REF##15883588##20##, ##REF##12447976##21####12447976##21##], the total number of RNA polymerases is less relevant than the number that are not currently occupied expressing genes outside the target system of interest. Our method for deducing this number involved using bulk measurements (collected as a function of growth rate [##UREF##1##12##]) to create an average (or \"mean field\") behaviour for the set of genes in the bacterial genome; we then tested how many expression enzymes our target gene had available to transcribe it, and generate free enzyme levels as a function of growth rate [##UREF##0##10##].</p>", "<p>We turn now to the second of the challenges mentioned above, that of computational time. Having the rest of the genome present in the system, even in our bulk-averaged way, added significantly to the computational demands of the simulations. Further investigation shows, however, that there are regimes in which the target system is not significantly affected by the presence of the remainder of the genome, and may thus well approximated by excluding the genome portion and simulating only the target system. The key quantity is the \"on rate\" of binding between RNA polymerase and the promoter of the target gene: for certain ranges of this parameter, the perturbation introduced by the presence of other genes (the rest of the genome in the cell) is small enough to be neglected, saving significant amounts of computational time. We explore the details of these ranges, as a function of other system parameters, below. We view this work as complementary to the various ongoing large-scale cellular simulation projects [##REF##14744094##2##,##REF##15113093##7##,##REF##15562315##19##,##UREF##3##22##, ####UREF##4##23##, ##REF##14683613##24##, ##REF##14681416##25####14681416##25##], offering a method of simplifying the system in cases where including genes outside the immediate system of interest does not alter the overall behaviour significantly. Although our results are obtained for our particular gene expression model, we anticipate that our promoter on-rates will apply, at least approximately, to other studies of transcription in bacteria, and thus offer guidance to others wishing to simplify their system by omitting the genomic influence.</p>" ]

[ "<title>Methods</title>", "<title><italic>E. coli </italic>gene expression model</title>", "<p>Our technique relies on the existence of experimental results [##UREF##1##12##] reporting bulk average assays of the amounts of each species present in the biological system of interest, as a function of growth rate; quantities such as average RNA polymerase per cell, average transcript content per cell, and so on, are much more readily obtained than specific rate constants for individual reactions. Using the bacterium <italic>Escherichia coli </italic>as a model organism, we have formulated a picture of the biochemical reactions underlying gene expression from an inserted plasmid carrying a promoter controlling the transcription of our target gene. We implemented a \"mean-field\" modelling approach, generating genome-wide averages for the mean transcript length, mean elongation time, and so on, adjusting the model parameters so that it generated numbers matching the bulk averages that had previously been reported experimentally [##UREF##0##10##,##UREF##1##12##]. A full list of the reactions included in the model and the nomenclature used for the species is provided in Tables ##TAB##0##1##, ##TAB##1##2##, and ##TAB##2##3##. The following sections provide an overview of the processes represented in the model, with further details provided in the Appendix and in our previously published work [##UREF##0##10##].</p>", "<title>Cell growth and division</title>", "<p>The cellular volume grows exponentially until a threshold is reached, at which point it is approximately halved (a binomial distribution is used) and exponential growth restarts. A counter species, <italic>v</italic>, is used to represent volume: <italic>v </italic>→ 2<italic>v</italic>, with rate constant <italic>k </italic>= ln(2)/τ, where τ is the doubling time of the cells. At cell division, all species present are divided between two hypothetical daughter cells, and the simulation follows one of these daughters. We treat the system as ergodic, and average over long times for a single cell to obtain ensemble averages across the cellular population.</p>", "<title>Enzyme binding and isomerization</title>", "<p>RNA polymerases (Rpoly) are responsible for initiating and catalyzing the transcription of messenger RNA (mRNA) strands. As the model assumes all mRNA transcripts reside in operons, Rpoly binds to promoter sequences in the DNA (operon) and forms a closed complex (Rpoly+operon→closed_Rpoly_prom). This closed complex then must isomerize into an open complex (closed_Rpoly_prom→open_Rpoly_prom) before transcription can begin.</p>", "<title>Enzyme clearance</title>", "<p>RNA polymerases clear the promoters, leaving those sites free to bind additional enzymes while transcription proceeds further down the DNA strand. We model this by regenerating the promoter after clearance occurs, forming an enzyme-template complex plus the original site: open_Rpoly_prom→Rpoly_operon+incom_mRNA+operon. We create a nascent transcript (mRNA_incom) at this step to allow subsequent translation to proceed; this feature will prove very helpful in studying future simulated studies of protein synthesis. Conservation of the number of promoters is maintained: when the enzyme-template complex finishes elongation, only the enzyme and the polymerized product are released.</p>", "<title>Elongation</title>", "<p>To avoid the complexity of accounting for each enzyme at different stages of elongation, a single reaction is used to represent the process of completing the mRNA chain: Rpoly_operon→Rpoly+mRNA. Compliment to this reaction is the disappearance of the nascent transcript made available during transcription: incom_mRNA→(), where () is a null placeholder. Both the reactions have the same elongation rate constant that can be summarized as k_elongation= ρ/λ, where ρ and λ are the polymerization rate and length of template, respectively.</p>", "<title>Enzyme production</title>", "<p>Since the kinetics of RNA polymerase assembly are not fully known, the model is simplified by treating enzyme production as a zero-order process in which enzymes appear from outside the model at a constant rate: ()→Rpoly. The enzymes are partitioned at cell division like all other species. The rate constant for production can be summarized as k_rep= (ν/1.5)/τ, where ν and τ are the average number per cell and cellular doubling time, respectively.</p>", "<title>DNA replication</title>", "<p>DNA replication in bacteria is a complex process involving multiple replication forks. We represent the coding portion of the genomic DNA by the number of operons present (operon), and simplify the replication process as a zero-order process: ()→operon. Rate constants for this process are chosen to match the number of genomes per cell at different growth rates.</p>", "<title>mRNA degradation</title>", "<p>RNases act to destroy mRNA in <italic>E. coli</italic>, and we represent the degradation of mRNA by these enzymes with first-order reactions: mRNA→(), and incom_mRNA→(); the latter is an additional RNase-driven degradation, beyond the above-mentioned rate of disappearance of incomplete mRNA through conversion to complete mRNA strands.</p>", "<title>RNA production from operons</title>", "<p>We assume that all genes in the genome are clustered into operons: groups of genes transcribed from a single promoter, as in the <italic>lac </italic>operon. The model keeps track of which gene on the mRNA operon Rpoly is currently transcribing and makes available completed transcripts of the nascent operon (this latter point will prove relevant in future protein synthesis models): Rpoly_operon1→Rpoly_operon2+mRNA. In response to the genome-wide average of 6.9 genes per operon [##UREF##0##10##,##UREF##1##12##] the model tracks the 7 transcripts representing the average mRNA operon (six genes of equal size, one 90% the length of the average size).</p>", "<p>In addition to messenger RNA, other forms of RNA collectively known as stable RNA (sRNA) are produced within the cell. Since sRNA is transcribed but not translated the model does not consider nascent sRNA production.</p>", "<title>With-genome and no-genome models</title>", "<p>We have constructed two versions of the model, one containing a representation of the host cell genome and the reporter gene, the other neglecting the cellular genome and representing only the reporter gene on the plasmid. The with-genome model incorporates 26 reactions involving 27 species, while the no-genome version has 10 reactions involving 10 species; the two versions are shown schematically in Figures ##FIG##0##1A## and ##FIG##0##1B##. The genome affects a plasmid-borne gene of interest by competing for RNA polymerase binding with the plasmid-borne promoter, while in the no-genome version of the model we omit the genomic promoter sites and thus this competition does not occur. The goal, then, is to determine the parameter regimes in which this omission has an acceptably small influence on the behaviour of the system, and to determine how much more quickly the computational simulations will run as a result of the simplification.</p>", "<title>Computational simulation method</title>", "<p>The chemical kinetics of this system were initially simulated using the Gillespie Monte Carlo algorithm [##REF##15113411##26##, ####UREF##5##27##, ##UREF##6##28####6##28##], and these results were used to validate a deterministic, ordinary differential equation (ODE) version of the system, which was shown to yield identical average transcript numbers, allowing us to use the significantly faster ODE model to generate larger numbers of points in parameter space. Comparing the two models allowed us to determine the point at which the on-rate constant between the target promoter and RNA polymerase, k_on, crossed a threshold where the two models (with and without the host genome included) generated average transcript numbers differing by more than a certain percentage; here, we have chosen a five percent difference as an admittedly arbitrary significance threshold.</p>", "<p>The original experimental measurements in the literature were carried out over a range of cellular growth rates, each of which yielded different average quantities of biomolecules per cell. Stochastic simulations of our system were carried out at each experimentally-examined growth rate (doubling times of 24, 30, 40, 60, and 100 minutes [##UREF##1##12##]) and sampled at discrete points in parameter space, as follows: plasmid copy numbers of 10, 100, and 1000; mRNA half-lives of 2, 6, 10, and 14 min; and gene lengths of 10, 100, 1000, and 10000 bp. The relationship between these independent variables and the point at which the promoter-RNAP on-rate begins to yield a significant difference between the genome and no-genome models is complex and highly nonlinear, and not amenable to reduction to a single equation. We have instead produced a MATLAB script (The MathWorks, Natick, MA) that generates an on-rate threshold given a user's input of plasmid copy number, mRNA degradation rate, gene length and cellular doubling time: any promoter on-rate constant larger than this predicted value can exclude the computationally expensive genome from the simulations without creating more than a five-percent error, while any constant smaller than this should include the genome.</p>", "<title>Stochastic modelling approach and software</title>", "<p>Deterministic chemical kinetics apply in the regime of large numbers of randomly interacting molecules. Inside cells, molecule numbers are often small enough to produce significant fluctuations [##REF##11283699##8##,##REF##15883588##20##,##UREF##6##28##, ####REF##16736013##29##, ##REF##16715097##30##, ##REF##12687005##31##, ##REF##16883354##32##, ##REF##12183631##33##, ##REF##15124029##34##, ##REF##11720979##35##, ##REF##14749823##36##, ##REF##15790857##37##, ##REF##16179466##38##, ##REF##958399##39##, ##UREF##7##40##, ##REF##16822033##41##, ##REF##11438714##42##, ##REF##15261147##43##, ##REF##16372021##44####16372021##44##], thus requiring a stochastic simulation of the reaction kinetics. The Gillespie algorithm [##REF##14744094##2##] treats chemical reactions as Poisson processes, with event (reaction) rates given by microscopic rate constants and the current state of the system. For an elementary reaction of the form A+B→C with rate constant <italic>k</italic>, the Poisson rate of the forward reaction is <italic>kab</italic>/<italic>V</italic>, where <italic>a </italic>and <italic>b </italic>represent the numbers of molecules of species A and B present, and <italic>V </italic>is the reaction volume (note that this volume is a changing parameter in a living bacterial cell). We use the unit \"n\" to represent the number of molecules present in the system, rather than concentration units such as molarity. To advance the simulation, the timing of the next reaction event is randomly selected using the exponential distribution of inter-event times for the set of Poisson processes representing the reactions, and the probability of each reaction being the one that occurs at that instant is given by its fraction of the sum of all reaction rates [##REF##15113411##26##, ####UREF##5##27##, ##UREF##6##28####6##28##].</p>", "<p>Bacterial cells have often been approximated as well-stirred reactors: based on their small size, it is assumed that diffusion is sufficiently fast to yield a well-mixed system. Early experimental results showed protein mobility <italic>in vivo </italic>consistent with normal diffusion [##REF##9864330##45##], and though the diffusion coefficients were substantially lower than for the same proteins in water, the diffusion was fast enough to spread the proteins over the volume of a bacterium on a time scale of seconds. Recent theoretical treatments [##REF##15261147##43##,##REF##16204833##46##, ####UREF##8##47##, ##REF##11590012##48##, ##REF##15142746##49####15142746##49##] have questioned the picture of bacterial cells as well-mixed systems, and recent experimental results [##REF##16606319##50##] have reported subdiffusive behavior in the motion of individual RNA molecules, where each RNA is rendered visible through binding to multiple fluorescent protein labels. In this paper, we use the well-stirred reactor picture as a first approximation to gain insight, but it should be noted that this is a significant simplification, and that future refinements and extensions are possible. Approaches proposed to deal with crowded cellular environments include rate laws obeying fractal-like kinetics [##REF##15142746##49##,##UREF##9##51##,##REF##17820893##52##], and Monte Carlo simulations wherein two- or three-dimensional spatial information is retained for each molecule [##REF##15261147##43##,##REF##16204833##46##,##REF##15142746##49##,##REF##12324410##53##].</p>", "<p>The gene expression model was initially implemented using BioNetS (Biochemical Network Stochastic Simulator) [##REF##15113411##26##], which provides a convenient interface for specifying reactants, products and kinetic data. The software generates C++ source code implementing the system using the Gillespie stochastic simulation algorithm (or an approximation, if desired), and this code is then compiled and executed with user-tunable parameters as inputs. Some species in the model exist in small numbers while others exist in large numbers; although continuum approximations and hybrid schemes are available through BioNetS [##REF##15113411##26##], the Gillespie algorithm with no approximations yielded the best simulation speed. The data from the BioNetS-generated code was processed using DataTask (Visual Data Tools, Inc) and its run manager DataTask, which automated the process of sweeping parameter values and analyzing the results. The complete gene expression models used are available as BioNetS scripts and are provided along with this paper (see Additional File ##SUPPL##0##1##).</p>", "<title>Derivation of <italic>E. coli </italic>gene expression parameters</title>", "<p>To derive the on-rate constant between RNA polymerase and the reporter promoter where there is 5% difference in transcript average between models, we employ bulk cellular averages obtained by Bremer and Dennis for several different cellular growth rates [##UREF##1##12##]. We implement a \"mean-field\" approach [##UREF##0##10##] by considering the production of generic transcripts with properties derived from genome-wide averages: we compute mean transcript lengths, mean elongation rates, and so on. With these quantities in hand, the unknown between models is reduced to the RNA polymerase on-rate constant for binding to the reporter promoter, and we find its value by sweeping until the difference in transcript average between models is 5%.</p>", "<p>The model has been constructed to be as detailed as possible, using all available information about the biochemical processes underlying gene expression. This leads to a large number of species and reactions, the full details of which are provided in the Appendix. For a derivation of average genome parameters, please see Iafolla and McMillen [##UREF##0##10##].</p>", "<title>Stochastic model parameter sweeping</title>", "<p>The first step in deriving the on-rate constant that determines a 5% difference in transcript averages between models is to obtain steady-state values of all species in the simulations. Figure ##FIG##1##2## shows the time series for one species in the model, the reporter mRNA. An initial run of 10 cell divisions in length is generated for each simulation, and the final state of this run is used as the initial state for the long-duration run in which statistics are accumulated to determine average species levels; this prevents the initial transient approach to steady state from distorting the averages.</p>", "<p>Parameter sweeping begins by using on-rates that vary by a factor of 10 (Figure ##FIG##2##3A##). When the desired percent difference between models lies between two on-rate constants, another sweep is performed between these new limits incrementing the on-rate by a unit multiple of the smaller limit. The third parameter sweep uses a unit increment of the next significant digit between the new limits; this change in on-rate is small enough to approximate linearity (Figure ##FIG##2##3B##). Only <italic>R</italic>²≥ 0.90 were accepted for interpolation; the range was narrowed until this level of linearity was achieved.</p>", "<p>The duration of the stochastic simulations was varied to obtain linearity with R²≥ 0.90; this is achieved by using a minimum of 1000 cell divisions, although some simulations use more cell divisions to obtain the desired linearity. Since the doubling time of the cells is varied, the total duration in real time varies among the simulations; the number of cell divisions explored appears to be the key factor in obtaining well-converged statistics, rather than the absolute duration.</p>", "<p>The minimum 1000 cell division duration was deduced by qualitative analysis of multiple simulations with the same seed but different durations; we examined the effect of duration on the mean values obtained from the reporter mRNA histograms. The on-rate constants used in the duration analysis was determined by comparing the histograms between models over a range of on-rate constants (10^-7n^-1s^-1to 1 n^-1s^-1); the range of on-rate constants that bound the percent difference in the above statistical parameters by 5% was investigated for duration analysis (this range was from 10^-5n^-1s^-1to 10^-2n^-1s^-1). Ultimately, longer-duration runs produced averages that were not statistically different from those obtained after 1000 divisions (see the Appendix for additional explanation), implying that longer durations only increase computational expense.</p>", "<p>After interpolation, the validity of the on-rate was tested: using a different seed for 30 simulations – all employing steady-state initial conditions and the same duration, kinetics and interpolated on-rate – the sample mean difference between models of the 30 simulations was statistically compared to the population mean of 5%. The on-rate was accepted if the two means were not proven statistically different using a level of significance α = 0.95. All simulations, either in parameter sweeping or verification, employ different nucleating random number generator seeds.</p>", "<title>Additional deterministic simulations</title>", "<p>The stochastic simulations are very computationally intensive, and thus we investigated methods of speeding up the calculations. The ordinary differential equations corresponding to the full reaction system for each model (genome and no genome) were derived using standard chemical kinetics and solved numerically using the solvers provided by MATLAB. To take cell growth and division into account, the ODEs were solved one cell cycle at a time, with the numbers of molecules at the end of the cycle cut in half to simulate division, then used as the initial state for the next cell cycle. Within each set of parameter values, each ODE was run for ten cell cycles to allow the system to reach a steady state, then for more ten more cell cycles, during which state values were averaged to obtain the average mRNA numbers for the reporter gene. As shown in Figure ##FIG##3##4##, the average mRNA numbers from the stochastic simulations matched nearly perfectly with those generated by the ODEs, and on this basis we used the deterministic ODEs to increase the number of points in the parameter space that could be feasibly sampled. (This reduction to the deterministic model is possible because here we are considering only the mean values from the stochastic simulation; in cases where the fluctuations were the point of interest, fully stochastic simulations would of course be required.) Full-scale stochastic simulations were carried out for the experimentally available doubling times (24, 30, 40, 60, and 100 minutes [##UREF##1##12##]), varying the other parameters as follows: gene lengths of 10, 100, 1000, and 10000 base pairs (bp); mRNA half-lives of 2, 6, 10, and 14 minutes; and plasmid copy numbers of 10, 100, and 1000 per cell. These were supplemented by deterministic simulations for the same doubling times, at the following parameter values: gene lengths from 10 to 100 in steps of 10 bp, from 100 to 1000 in steps of 100 bp, and from 1000 to 10000 in steps of 1000 bp; mRNA half-lives from 2 to 14 minutes in steps of 1 minute; and plasmid copy numbers from 1 to 9 in steps of 1, from 10 to 100 in steps of 10, and from 100 to 1000 in steps of 100 copies per cell.</p>", "<p>Similar to the parameter sweeping carried out for the stochastic simulations, we used the deterministic simulation results for each parameter set to calculate the RNA polymerase-promoter binding on rate, k_on, at which there will be a five percent difference between the models with and without a representation of the host cell genome; for the deterministic results, the 5% threshold was determined using the fzero function in MATLAB, which searches for a zero-crossing between two given points.</p>", "<title>Interpolation of on-rate thresholds</title>", "<p>The on-rate (k_on) threshold above which a 5% deviation between the genome and no-genome models occurred has been calculated explicitly only at the set of parameter values listed above (based on stochastic simulations supplemented by cross-validated deterministic simulations to increase the density of the sampling of parameter space). To allow the k_on threshold to be calculated at values other than those explicitly simulated, we created a MATLAB script to carry out the necessary interpolation using a local minimization method. In local linear fitting, to find the unknown point at a desired parameter value, one draws a straight line connecting the known points on either side of the desired value, and takes the point on that straight line as the interpolated result at the desired parameter value. Note that this process minimizes the total distance between the interpolated point and the two known points, and we use this idea to perform our interpolation in our 5-dimensional space (k_on as a function of four parameters: growth rate, gene length, mRNA half life, and plasmid copy number). For any single given 4-dimensional parameter set, the nearest available set of parameter values is determined by finding the two nearest parameter values in each direction on this 4-dimensional mesh; combining all four dimensions yields the 16 nearest points on the mesh. Since these 16 data points do not generally fit well to a linear function, we obtain the interpolated on-rate value for a given parameter set by searching for the k_on value that minimizes the total distance in 5-dimensional space to those nearest 16 points, using the MATLAB fminsearch function to carry out the minimization operation.</p>", "<p>The above interpolation has been implemented in MATLAB script that presents a simple user interface allowing the user to enter the desired parameter values (within the ranges spanned by the simulations), after which the script will carry out the interpolation for the given point and return the k_on value above which a 5% difference arises between the genome and no-genome models: any promoter on-rate constant larger than this predicted value can exclude the computationally expensive genome from the simulations without creating more than a five-percent error, while any constant smaller than this should include the genome. The user interface is shown in Figure ##FIG##4##5##, and the MATLAB files required to implement it are provided along with this paper (see Additional File ##SUPPL##1##2##).</p>" ]

[ "<title>Results and discussion</title>", "<title>Percent difference of reporter transcript averages between models</title>", "<p>As shown in Figure ##FIG##2##3##, using the stated parameters as a representative example, the percent difference of reporter transcripts between models changes as a function of binding constant between RNA polymerase and the target promoter (k_on). An excessively small binding constant (≈ 10^-10n^-1s^-1to 10^-7n^-1s^-1) prevents the RNA polymerase from binding to the promoter, thereby producing an insignificant number of transcripts, usually less than one per cell division, as shown in Figure ##FIG##5##6##. The constant can be so small that noise dominates the system, leading to essentially random results, including some in which more reporter transcripts are produced in simulations that use the genome, relative to the simulations that only use the plasmid-borne reporter genes. Eventually the binding constant becomes large enough to produce a considerable quantity of transcripts; at this point the genome's presence competes with the reporter gene for access to RNA polymerase and reduces the transcription of the reporter gene, producing a significant percent difference between models. As the binding constant to the reporter promoter further increases, the RNA polymerase binding saturates and the promoter generates nearly the same number of transcripts with or without the presence of the competing genome; the difference between models trends towards zero as the binding constant approaches infinity.</p>", "<p>Figure ##FIG##2##3## shows there are two binding constant ranges for each set of parameters where there is less than a 5% difference in transcript production. We have not considered the lower range, here, because of the insignificant number of transcripts produced, usually an average of much less than one per cell division. In this regime, the two versions of the model are both matching simply because they are both yielding a result of \"nearly zero.\" For the case we wish to consider, that of observing the output of a target gene through the expression of a reporter, such low levels of transcription would be invisible to current detection techniques, requiring single-molecule resolution against the noisy background of the cytoplasm, and thus for the moment we consider it justified to exclude this near-zero range in our simulations. The higher k_on rate constant limit corresponds to transcript numbers on the order of 10²to 10⁴, a magnitude that is much more amenable to experimental access and thus potentially more significant for use in other studies.</p>", "<title>Accuracy of the interpolated on-rates</title>", "<p>To test the accuracy of the interpolated on-rates, the on-rates were entered back into both versions of the model and run for 30 different simulations seeds for a duration of 30 cell divisions, after creating steady state values for all species within the model. The percent differences were assembled and statistically compared to the population mean of 5% using a level of significance α = 0.95. This process was repeated for all 240 different kinetic situations generated using the stochastic simulations. There was no statistical difference between the population mean and the sample mean obtained from the simulations (data not shown), thereby ensuring that the interpolated values are the correct ones for producing a percent difference of 5%.</p>", "<title>Time reduction via genome exclusion</title>", "<p>Excluding the genome from simulation studies does reduce CPU simulation time in the computationally intensive fully stochastic simulations. To illustrate this, the verification runs were used for comparison between models; these simulations employ the same kinetic parameters and duration, and offer a large population size (since each run was repeated multiple times with varying random seeds).</p>", "<p>Dividing the average run time of the genome by those models excluding it produces a direct measure of the benefit of excluding the host cell genome in the simulations. As Figure ##FIG##6##7## shows, computational time can be reduced by a factor ranging from two to 24-fold. Accurate analysis of the time saved between models requires standard CPU power. The verification simulations in this study have been spread out over many computers, most of which have different CPUs. To normalize the results, 10 replicates of a standard run with the same kinetic parameters, duration and random number seed was run (with minimal other processor load) on each type of CPU, for each version of the model. The simulation duration was set to take approximately 30 minutes of CPU time, to average away any aberrations caused by minor fluctuations in CPU availability over time. The run durations for these standard runs were then used to create a scaling factor for each CPU type, and the simulation times reported in Figure ##FIG##6##7## were corrected by these factors.</p>", "<p>The simulation spends most of its time on the RNA polymerase binding/binding reactions: the reactions operon_ns+Rpoly, operon_s+Rpoly, and plas+Rpoly in the with-genome model, and simply plas+Rpoly in the no-genome model. Figures ##FIG##7##8A## and ##FIG##7##8B## show the number of reaction steps simulated in the with-genome and no-genome versions of the model (keeping plasmid copy number, mRNA half-life, and gene length fixed, while varying cell doubling time). As Figure ##FIG##7##8A## shows, the number of reaction steps dedicated to simulating the genomic RNA polymerase binding operations falls off more rapidly with growth rate than does the number of steps required to simulate the plasmid-to-RNA polymerase binding. Figure ##FIG##7##8B## shows that the number of reaction steps simulated in the no-genome version of the model falls off as a function of growth rate, but less rapidly than in the with-genome case; this is the cause of the reduction in the relative advantage of the no-genome version as the growth rate increases, seen in Figure ##FIG##6##7##. For large plasmid copy numbers, the RNA polymerase binding steps are more time-consuming in the no-genome version of the model, and the computational advantage of excluding the genome is correspondingly smaller; again, this is seen in Figure ##FIG##6##7##.</p>", "<title>Relationship between the parameters</title>", "<p>Figures ##FIG##8##9##, ##FIG##9##10##, and ##FIG##10##11## show the dependence of the k-on value on gene length, plasmid number and mRNA half-life, while the doubling time is fixed at 30 minutes. These plots are 3D slices through the full 5D space of results (where the five dimensions are the four input parameters, mRNA half life, gene length, plasmid number, and doubling time, and the output promoter on-rate, k_on). The plots show some of the nonlinearity inherent in the relationship of k_on to the parameters, and help to indicate why it has not proven to be possible to reduce the parameter relationships to a single regression equation.</p>", "<title>Potential extensions</title>", "<p>Simulating the translation of mRNA to protein, downstream of the transcriptional events discussed here, requires a significantly more elaborate model [##UREF##0##10##] with correspondingly greater computational demands. One extension of this study would be to investigate the binding on-rates for ribosomes binding to the ribosome-binding-sites (RBS) of the mRNA binding sites, and once again compare the results when the presence of the genome is modelled to those when it is excluded; presumably there would be a similar possibility of excluding the representation of the genome under some parameter ranges (where the main parameters would remain the same: doubling time, gene length, mRNA half-life, and plasmid copy number). Since translation follows transcription in the gene expression process, the range of parameter values in which one can exclude the genome from studies of the translational output of a target gene should be smaller than the regions found in the current study of transcriptional output: the system will be subject to the constraints imposed by matching the transcriptional results, as well as additional constraints required to match the translational results.</p>", "<p>The ability of RNA polymerase to produce an approximately equal amount of transcripts at large enough binding constants for both models raises an important question: are there enough RNA polymerases left when a large rate law exists for the reporter promoter to transcribe the necessary genomic genes for cell division? The presence of a large rate for the reporter transcript will produce metabolic strain on the cell [##REF##18352063##54##, ####REF##18600670##55##, ##REF##11388089##56####11388089##56##], possibly leading to an increase in doubling time that is not captured within the current model. Further studies on modelling the effect of metabolic strain and its feedback with cellular doubling time will help to clarify this issue.</p>" ]

[ "<title>Results and discussion</title>", "<title>Percent difference of reporter transcript averages between models</title>", "<p>As shown in Figure ##FIG##2##3##, using the stated parameters as a representative example, the percent difference of reporter transcripts between models changes as a function of binding constant between RNA polymerase and the target promoter (k_on). An excessively small binding constant (≈ 10^-10n^-1s^-1to 10^-7n^-1s^-1) prevents the RNA polymerase from binding to the promoter, thereby producing an insignificant number of transcripts, usually less than one per cell division, as shown in Figure ##FIG##5##6##. The constant can be so small that noise dominates the system, leading to essentially random results, including some in which more reporter transcripts are produced in simulations that use the genome, relative to the simulations that only use the plasmid-borne reporter genes. Eventually the binding constant becomes large enough to produce a considerable quantity of transcripts; at this point the genome's presence competes with the reporter gene for access to RNA polymerase and reduces the transcription of the reporter gene, producing a significant percent difference between models. As the binding constant to the reporter promoter further increases, the RNA polymerase binding saturates and the promoter generates nearly the same number of transcripts with or without the presence of the competing genome; the difference between models trends towards zero as the binding constant approaches infinity.</p>", "<p>Figure ##FIG##2##3## shows there are two binding constant ranges for each set of parameters where there is less than a 5% difference in transcript production. We have not considered the lower range, here, because of the insignificant number of transcripts produced, usually an average of much less than one per cell division. In this regime, the two versions of the model are both matching simply because they are both yielding a result of \"nearly zero.\" For the case we wish to consider, that of observing the output of a target gene through the expression of a reporter, such low levels of transcription would be invisible to current detection techniques, requiring single-molecule resolution against the noisy background of the cytoplasm, and thus for the moment we consider it justified to exclude this near-zero range in our simulations. The higher k_on rate constant limit corresponds to transcript numbers on the order of 10²to 10⁴, a magnitude that is much more amenable to experimental access and thus potentially more significant for use in other studies.</p>", "<title>Accuracy of the interpolated on-rates</title>", "<p>To test the accuracy of the interpolated on-rates, the on-rates were entered back into both versions of the model and run for 30 different simulations seeds for a duration of 30 cell divisions, after creating steady state values for all species within the model. The percent differences were assembled and statistically compared to the population mean of 5% using a level of significance α = 0.95. This process was repeated for all 240 different kinetic situations generated using the stochastic simulations. There was no statistical difference between the population mean and the sample mean obtained from the simulations (data not shown), thereby ensuring that the interpolated values are the correct ones for producing a percent difference of 5%.</p>", "<title>Time reduction via genome exclusion</title>", "<p>Excluding the genome from simulation studies does reduce CPU simulation time in the computationally intensive fully stochastic simulations. To illustrate this, the verification runs were used for comparison between models; these simulations employ the same kinetic parameters and duration, and offer a large population size (since each run was repeated multiple times with varying random seeds).</p>", "<p>Dividing the average run time of the genome by those models excluding it produces a direct measure of the benefit of excluding the host cell genome in the simulations. As Figure ##FIG##6##7## shows, computational time can be reduced by a factor ranging from two to 24-fold. Accurate analysis of the time saved between models requires standard CPU power. The verification simulations in this study have been spread out over many computers, most of which have different CPUs. To normalize the results, 10 replicates of a standard run with the same kinetic parameters, duration and random number seed was run (with minimal other processor load) on each type of CPU, for each version of the model. The simulation duration was set to take approximately 30 minutes of CPU time, to average away any aberrations caused by minor fluctuations in CPU availability over time. The run durations for these standard runs were then used to create a scaling factor for each CPU type, and the simulation times reported in Figure ##FIG##6##7## were corrected by these factors.</p>", "<p>The simulation spends most of its time on the RNA polymerase binding/binding reactions: the reactions operon_ns+Rpoly, operon_s+Rpoly, and plas+Rpoly in the with-genome model, and simply plas+Rpoly in the no-genome model. Figures ##FIG##7##8A## and ##FIG##7##8B## show the number of reaction steps simulated in the with-genome and no-genome versions of the model (keeping plasmid copy number, mRNA half-life, and gene length fixed, while varying cell doubling time). As Figure ##FIG##7##8A## shows, the number of reaction steps dedicated to simulating the genomic RNA polymerase binding operations falls off more rapidly with growth rate than does the number of steps required to simulate the plasmid-to-RNA polymerase binding. Figure ##FIG##7##8B## shows that the number of reaction steps simulated in the no-genome version of the model falls off as a function of growth rate, but less rapidly than in the with-genome case; this is the cause of the reduction in the relative advantage of the no-genome version as the growth rate increases, seen in Figure ##FIG##6##7##. For large plasmid copy numbers, the RNA polymerase binding steps are more time-consuming in the no-genome version of the model, and the computational advantage of excluding the genome is correspondingly smaller; again, this is seen in Figure ##FIG##6##7##.</p>", "<title>Relationship between the parameters</title>", "<p>Figures ##FIG##8##9##, ##FIG##9##10##, and ##FIG##10##11## show the dependence of the k-on value on gene length, plasmid number and mRNA half-life, while the doubling time is fixed at 30 minutes. These plots are 3D slices through the full 5D space of results (where the five dimensions are the four input parameters, mRNA half life, gene length, plasmid number, and doubling time, and the output promoter on-rate, k_on). The plots show some of the nonlinearity inherent in the relationship of k_on to the parameters, and help to indicate why it has not proven to be possible to reduce the parameter relationships to a single regression equation.</p>", "<title>Potential extensions</title>", "<p>Simulating the translation of mRNA to protein, downstream of the transcriptional events discussed here, requires a significantly more elaborate model [##UREF##0##10##] with correspondingly greater computational demands. One extension of this study would be to investigate the binding on-rates for ribosomes binding to the ribosome-binding-sites (RBS) of the mRNA binding sites, and once again compare the results when the presence of the genome is modelled to those when it is excluded; presumably there would be a similar possibility of excluding the representation of the genome under some parameter ranges (where the main parameters would remain the same: doubling time, gene length, mRNA half-life, and plasmid copy number). Since translation follows transcription in the gene expression process, the range of parameter values in which one can exclude the genome from studies of the translational output of a target gene should be smaller than the regions found in the current study of transcriptional output: the system will be subject to the constraints imposed by matching the transcriptional results, as well as additional constraints required to match the translational results.</p>", "<p>The ability of RNA polymerase to produce an approximately equal amount of transcripts at large enough binding constants for both models raises an important question: are there enough RNA polymerases left when a large rate law exists for the reporter promoter to transcribe the necessary genomic genes for cell division? The presence of a large rate for the reporter transcript will produce metabolic strain on the cell [##REF##18352063##54##, ####REF##18600670##55##, ##REF##11388089##56####11388089##56##], possibly leading to an increase in doubling time that is not captured within the current model. Further studies on modelling the effect of metabolic strain and its feedback with cellular doubling time will help to clarify this issue.</p>" ]

[ "<title>Conclusion</title>", "<p>Efforts to create accurate, quantitative models of <italic>Escherichia coli </italic>genomic networks using chemical equations results in large reaction schemes, with reactions potentially proceeding at a wide range of rates. The large computational time required to simulate these reactions is a persistent problem for large-scale cellular simulation. To help address one aspect of this problem, we have investigated the necessity of simulating the presence of the <italic>E. coli </italic>genome when studying a target gene inserted on a plasmid. The presence of the genome, introduced using our \"mean-field\" approach, is felt by the target gene through the competition for free RNA polymerases available to bind to the target gene's promoter and generate transcripts. However, there are ranges of the parameter space in which the presence of the genome yields a negligible difference in the number of reporter transcripts produced from the target gene, and in these cases is it possible to exclude any explicit representation of the genome and save the computations required to simulate the associated additional reactions. Stochastic simulations show speed increases of from two to 24 times, when the genome is excluded from our models. We have generated a set of fully stochastic simulations and found the promoter on-rate values for which the genome and no-genome models differ by less than 5%, and augmented these stochastic simulations with cross-validated deterministic runs to increase the number of sampled points in parameter space. Within the ranges of our four independent parameters (growth rate, gene length, mRNA degradation half-life, and plasmid copy number), we have produced a MATLAB user interface that will allow the user to input any set of parameters and obtain the promoter on-rate value (k_on) above which the effect of the genome will fall below our 5%-difference threshold. Given the increasing computational demands of cellular simulations, we hope that this approach will aid in the efficiency of other studies, and suggest other methods in which portions of the full cellular system may be excluded without significantly affecting the final results.</p>" ]

[ "<p>This is an Open Access article distributed under the terms of the Creative Commons Attribution License (<ext-link ext-link-type=\"uri\" xlink:href=\"http://creativecommons.org/licenses/by/2.0\"/>), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.</p>", "<title>Background</title>", "<p>Simulating the major molecular events inside an <italic>Escherichia coli </italic>cell can lead to a very large number of reactions that compose its overall behaviour. Not only should the model be accurate, but it is imperative for the experimenter to create an efficient model to obtain the results in a timely fashion. Here, we show that for many parameter regimes, the effect of the host cell genome on the transcription of a gene from a plasmid-borne promoter is negligible, allowing one to simulate the system more efficiently by removing the computational load associated with representing the presence of the rest of the genome. The key parameter is the on-rate of RNAP binding to the promoter (k_on), and we compare the total number of transcripts produced from a plasmid vector generated as a function of this rate constant, for two versions of our gene expression model, one incorporating the host cell genome and one excluding it. By sweeping parameters, we identify the k_on range for which the difference between the genome and no-genome models drops below 5%, over a wide range of doubling times, mRNA degradation rates, plasmid copy numbers, and gene lengths.</p>", "<title>Results</title>", "<p>We assess the effect of the simulating the presence of the genome over a four-dimensional parameter space, considering: 24 min &lt;= bacterial doubling time &lt;= 100 min; 10 &lt;= plasmid copy number &lt;= 1000; 2 min &lt;= mRNA half-life &lt;= 14 min; and 10 bp &lt;= gene length &lt;= 10000 bp. A simple MATLAB user interface generates an interpolated k_on threshold for any point in this range; this rate can be compared to the ones used in other transcription studies to assess the need for including the genome.</p>", "<title>Conclusion</title>", "<p>Exclusion of the genome is shown to yield less than 5% difference in transcript numbers over wide ranges of values, and computational speed is improved by two to 24 times by excluding explicit representation of the genome.</p>" ]

[ "<title>Authors' contributions</title>", "<p>MAJI conceived of the study, designed molecular simulations, implemented stochastic simulations, compiled data, analyzed results and drafted the manuscript. GQD carried out deterministic simulations, compiled data, analyzed results, implemented networked runs of stochastic simulations, and produced the MATLAB parameter interpolation routines and user interface. DRM participated in design of the study, and helped draft and finalize the manuscript and revisions. All authors participated in the writing and approved of the final form of the manuscript.</p>", "<title>Appendix</title>", "<p>Below is a detailed explanation of the gene expression model, expanding on the information presented in the Methods section. A full list of kinetic parameters for each reaction is provided in Iafolla and McMillen [##UREF##0##10##].</p>", "<title>Nomenclature</title>", "<p>The following is a complete list of species names used in the model:</p>", "<p>The cellular processes represented in the model are discussed individually, below:</p>", "<title>Cellular division</title>", "<p>To reflect the exponential growth of bacterial cells in a nutrient-rich liquid culture, we include cell growth and division, incorporated as a process that grows to a threshold volume and is then halved. At division, all species have their numbers cut approximately in half: for large numbers, a binomial distribution is used to calculate the new number, while small numbers (less than 100) have each molecule explicitly checked and randomly assigned to a daughter cell with equal probability [##REF##15113411##26##]. The model follows only one cell as a representative of the full population, so the second daughter effectively vanishes after division. Tracking such a representative cell over long times yields the same statistics as tracking an ensemble of many cells over shorter times, if we make the reasonable assumption that the system is ergodic.</p>", "<p>Cell volume is represented by a \"counter\" species, <italic>v</italic>, whose exponential growth is governed by the following reaction, with rate constants adjusted to produce various doubling times to match the experimental conditions being examined:</p>", "<p></p>", "<p>For a doubling time τ, the rate constant is set to k = ln(2)/τ. The reaction is initialized at <italic>v </italic>= V₀, and cellular division occurs when <italic>v </italic>reaches 2V₀. Our model treats all processes as stochastic, but the resulting degree of variability depends strongly on the number of molecules participating in the reaction. The range of cell division times can thus be tuned by the choice of V₀; here we set V₀= 1000, which yields a very slight degree of variability in the cell division times. This variability arises from two sources: the stochastic rate of reaction R1, and the random assortment of the counter <italic>v </italic>between daughter cells at division: <italic>v </italic>is cut only approximately in half at cell division, like all other species, and thus the initial volume after cell division lies in a small range around V₀.</p>", "<title>Enzyme binding, unbinding, isomerization and clearance</title>", "<p>Since the only enzymes used in this model are RNA polymerases only binding to promoters need consideration. The bimolecular reactions for RNA polymerase (Rpoly) binding to a promoter on a gene (plas or operon) are shown below:</p>", "<p></p>", "<p></p>", "<p>RNA polymerase initially forms a closed-complex with the promoter region, which then undergoes isomerization (R3) into an open-complex. The rate constants for R2 and R3 are adjacent to the reactions; that for R2 is scaled to mimic dilution of cell cycle progression: as the cell grows, the increase in volume decreases the probability of the two species coming into contact and reacting, effecting reducing the rate constant [##REF##15113411##26##]; this effect is incorporated by dividing the rate constants by <italic>v</italic>/V₀.</p>", "<p>Following binding, the enzyme clears the promoter at a particular rate. The elementary reactions for this process are shown below:</p>", "<p></p>", "<p>We create a nascent transcript (mRNA_incom) at this step to allow subsequent translation to proceed; this feature will prove very helpful in studying future simulated studies of protein synthesis. Reaction R4 also shows an important assumption: the regeneration of a binding site after clearance allows another enzyme to bind to the same gene, creating the multiple simultaneous elongation processes observed in actual bacterial cells.</p>", "<title>Elongation</title>", "<p>To avoid the computational complexity of accounting for all elongating intermediates (growing mRNA and peptides of every possible length), the following approximation has been employed: a single intermediate is converted to the final product at a rate corresponding to the average time taken by the complete polymerization process. Using average elongation rates for specific cell growth rates as specified by Bremer and Dennis [##UREF##1##12##], the elongating species produce only the enzyme and the polymerized product, not the template that is read. This is shown below in Reaction R5:</p>", "<p></p>", "<p>Compliment to this reaction is the disappearance of the nascent transcript made available during transcription: incom_mRNA→(), where () is a null placeholder. The elongation rate constant can be summarized as k_elongation= ρ/λ, where ρ and λ are the polymerization rate and length of template, respectively.</p>", "<title>Enzyme and genome production</title>", "<p>Many processes involved in molecular biology are either too complex to model or not characterized at present. In our model, we use simplified zeroth-order production rates for complicated species involved: although the assembly details of some species are not fully available, there is considerable information on population size of these species. In <italic>E. coli</italic>, the average number of RNA polymerases and genome equivalents per cell are known at several cellular growth rates [##UREF##1##12##], and their production is represented by the elementary reactions below:</p>", "<p></p>", "<p></p>", "<p></p>", "<p>The operon species in R8 is representative of the genome, since our model employs RNA polymerase binding directly to the promoter sequence of the average operon. The rate constant for production can be summarized as k_rep= (ν/1.5)/τ, where ν and τ are the average number per cell and cellular doubling time, respectively.</p>", "<title>mRNA degradation</title>", "<p>The presence of RNases in <italic>E. coli </italic>implies that mRNA possess a finite life-span. The following reactions are used to represent mRNA degradation:</p>", "<p></p>", "<p></p>", "<p>For a half-life <italic>h</italic>, the rate constant for R9 and R10 is set to k = ln(2)/<italic>h</italic>.</p>", "<p>We assume that RNases can degrade nascent transcripts. To account for degrading a transcript while it is being created we propose the following elementary reaction and rate constant:</p>", "<p></p>", "<p></p>", "<p>The reaction indicates that an RNA polymerase currently producing a transcript becomes an unscathed RNA polymerase and a degraded mRNA. Although this reaction implies that all RNA polymerases producing a transcript are subject to degradation, the proportionality to incomplete transcripts is specified in the rate constant. The Rpoly_mRNA species present in the denominator of the rate constant makes the reaction rate independent of the number of elongating RNA polymerases.</p>", "<title>Modelling RNA production from operons</title>", "<p>We assume that all genes in our relevant genome are clustered into operons. Our model creates a single transcript for the entire operon, mimicking the <italic>lac </italic>operon [##REF##5360547##57##]. To make the elementary reactions simple and accurate for mRNA and subsequent peptide production, RNA polymerase binds once to the promoter and produces a transcript of average length under corresponding kinetics; the ejection of the mRNA occurs simultaneously with RNA polymerase transcribing the adjacent gene on the operon, or in the case of the last gene on the operon, being released. This is shown in the following reactions for a hypothetical three gene operon, where the binding (R2), isomerization (R3) and clearance steps (R4) have been omitted:</p>", "<p></p>", "<p></p>", "<p></p>", "<p>The numeric suffix on the Rpoly_operon species represents the gene number adjacent to the promoter. Notice that the rate constants for the above reactions are all equivalent. The release of the mRNA while the RNA polymerase is still elongating the operon allows ribosomes to bind and perform translation without requiring additional species; the act of transcription is conserved since RNA polymerase only binds once to the promoter. Evidently, the total time to transcribe all three genes is equivalent to the time for transcribing the whole operon.</p>", "<p>Contrast to mRNA production, stable RNA is easily produced. Since this RNA is not translated there is no need to include ribosomes translating complete transcripts before the operon is finished elongation. Hence, the length of stable RNA in the model is equivalent to the average stable RNA operon length.</p>", "<title>Supplementary Material</title>" ]

[ "<title>Acknowledgements</title>", "<p>This work has been funded by the Natural Sciences and Engineering Research Council of Canada (NSERC), the Ontario Research Fund (ORF), the Canada Foundation for Innovation (CFI), the Ontario Photonics Consortium (OPC), and the Canadian Institutes for Health Research (CIHR).</p>" ]

[ "<fig position=\"float\" id=\"F1\"><label>Figure 1</label><caption><p><bold>Schematic of the two versions of the model system</bold>. Our simulations compare two versions of a bacterial gene expression model. (A) In the first version, the genome is represented as an \"average\" open generating generic transcripts, rather than as the full set of individual genes. Bulk experimental measurements are used to generate the correct average number of transcripts from this generic operon in the genome. In this version of the model, the promoter residing on a plasmid of interest (plas) competes with the genomic operons for access to RNA polymerase (Rpoly) enzymes. (B) In this version, all references to the host cell genome are excluded from the model, leaving only the plasmid-borne promoter (plas) to be transcribed by RNA polymerase (Rpoly). Full lists of the reactions that constitute the models are given in Tables 1 and 2, with a list of species names given in Table 3.</p></caption></fig>", "<fig position=\"float\" id=\"F2\"><label>Figure 2</label><caption><p><bold>Typical time series generated by the model</bold>. Typical time series generated by the model. Plot A (left) shows a run with all intermediates and products initially set to zero, illustrating the initial transient. Plot B (right) shows a run initialized with the state obtained after 10 cell divisions in the left-hand run, thus removing the initial transient. Simulations were performed at a variety of cellular growth rates with different kinetic parameters. Parameters for this example: doubling time = 24 min; on-rate constant = 10^-3n^-1s^-1; plasmid copy number = 10; gene length = 10⁴bp; and mRNA half-life = 14 min.</p></caption></fig>", "<fig position=\"float\" id=\"F3\"><label>Figure 3</label><caption><p><bold>Parameter sweeping</bold>. Parameter sweeping. Here, we compare two versions of the gene expression model, one incorporating the host cell genome and one excluding it. The RNA polymerase on-rate constant for binding to the promoter that produces the reporter mRNA is varied until the percentage difference between these models exceeds 5% (the value we have selected as our threshold for a significant difference between the two models, marked by a horizontal dashed line on each plot). The on-rate is first varied by a factor of 10 to determine the general location of the desired value (plot A, left), followed by a sweep on a finer scale to narrow in on an approximately linear region near the threshold crossing (plot B, right). The solid vertical line in Graph B shows the interpolated on-rate constant when the percent difference in transcript production between models crosses the 5% threshold. The parameters for this example are: doubling time = 24 min; plasmid copy number = 10; gene length = 10 bp; and mRNA half-life = 6 min.</p></caption></fig>", "<fig position=\"float\" id=\"F4\"><label>Figure 4</label><caption><p><bold>Comparison of stochastic and deterministic simulation outputs</bold>. Comparison of stochastic and deterministic simulation outputs. The stochastic simulations required too much computational time for it to be practical to sample the parameter space very densely. Since we have used only the mean values from the stochastic simulations, we explored the possibility of using deterministic simulations, which require a tiny fraction of the stochastic simulation time, to increase our sampling of the parameter space. The plot shows the average number of mRNA transcripts generated by the two methods, stochastic and deterministic. The straight diagonal line indicates a good match, and in fact the two methods differ by less than one percent in most cases. Parameter values are the same as those used in Figure 2A.</p></caption></fig>", "<fig position=\"float\" id=\"F5\"><label>Figure 5</label><caption><p><bold>MATLAB graphical user interface</bold>. MATLAB graphical user interface. The on-rate at which excluding the genome yields less than a 5% difference between the genome and no-genome models is a complex function of the parameters: population doubling time, gene length, mRNA degradation half-life, and plasmid copy number. This space is sampled only at discrete points, but the MATLAB user interface (provided in the additional files accompanying this paper) allows the user to enter any value within the ranges sampled by our simulations (the allowable range is specified above each parameter's input window). A threshold on-rate (above which the genome and no-genome models differ by less than 5%) is calculated by a minimum-distance interpolation between the nearest available points (see text for more detail).</p></caption></fig>", "<fig position=\"float\" id=\"F6\"><label>Figure 6</label><caption><p><bold>The effect of the genome on the reporter transcript output</bold>. The effect of the genome on the reporter transcript output. At small enough binding constants neither model is able to produce a significant number of transcripts: the average time between transcriptions is much larger than the doubling time, leading to an average of much less than one transcript per cell division. As the binding constant increases, the reporter promoter starts to compete with the genomic promoters for RNA polymerase, ultimately producing a difference in the number of transcripts between models. The above example has been arbitrarily chosen; it uses the same parameters as in Figure 2 (doubling time = 24 min; plasmid copy number = 10; gene length = 10 bp; and mRNA half-life = 6 min). The error bars are a single standard deviation in the transcript number distributions generated by the stochastic simulations.</p></caption></fig>", "<fig position=\"float\" id=\"F7\"><label>Figure 7</label><caption><p><bold>Ratios of simulation times in models with and without the genome included</bold>. The ratio of simulation time between models with the genome to those excluding it, as a function of doubling time and plasmid copy number. Removing the genome from simulation studies can be 2 to 24 times more efficient compared to those that include it. The data was constructed by averaging the simulation times for all verification runs that employed the set doubling times and plasmid copy numbers, regardless of mRNA half-life and gene length. All computer simulation times were normalized with respect to the computer's CPU strength. The trends suggest that the ratio will approach 1 for sufficiently long doubling times.</p></caption></fig>", "<fig position=\"float\" id=\"F8\"><label>Figure 8</label><caption><p><bold>Number of reaction steps simulated</bold>. Fixing three of the input parameters (plasmid copy number is 10, mRNA half-life is 6 min, and gene length is 1000 bp), we plot the number of reaction steps taken in a stochastic run simulating 9000 seconds of time. (A) With-genome model. The total number of reaction steps, and the number of reactions dedicated to RNA polymerase binding/unbinding to the genomic operons, and to the plasmid carrying our gene of interest. (B) No-genome model. The total number of reaction steps, and the number of reactions dedicated to RNA polymerase binding/unbinding to the plasmid carrying our gene of interest.</p></caption></fig>", "<fig position=\"float\" id=\"F9\"><label>Figure 9</label><caption><p><bold>Dependence of promoter strength on gene length and plasmid number</bold>. The full set of simulations yields promoter strengths, k_on, as a function of four input parameters (gene length, plasmid number, mRNA half-life, and cell doubling time). Here, we fix the doubling time at 30 minutes and the mRNA half-life at 8 minutes, and plot k_on as a function of the two remaining parameters: gene length and plasmid copy number.</p></caption></fig>", "<fig position=\"float\" id=\"F10\"><label>Figure 10</label><caption><p><bold>Dependence of promoter strength on plasmid number and mRNA half-life</bold>. The full set of simulations yields promoter strengths, k_on, as a function of four input parameters (gene length, plasmid number, mRNA half-life, and cell doubling time). Here, we fix the doubling time at 30 minutes and the gene length at 4000 base pairs, and plot k_on as a function of the two remaining parameters: plasmid copy number and mRNA half-life.</p></caption></fig>", "<fig position=\"float\" id=\"F11\"><label>Figure 11</label><caption><p><bold>Dependence of promoter strength on mRNA half-life and gene length</bold>. The full set of simulations yields promoter strengths, k_on, as a function of four input parameters (gene length, plasmid number, mRNA half-life, and cell doubling time). Here, we fix the doubling time at 30 minutes and the plasmid copy number at 200, and plot k_on as a function of the two remaining parameters: mRNA half-life and gene length.</p></caption></fig>" ]

[ "<table-wrap position=\"float\" id=\"T1\"><label>Table 1</label><caption><p>Biochemical reactions that make up our bacterial gene expression model (version incorporating the host's genome). </p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td align=\"left\">Left</td><td align=\"left\">Right</td><td align=\"left\">Forward Rate constant</td><td align=\"left\">Backward rate constant</td></tr></thead><tbody><tr><td align=\"left\">operon_ns + Rpoly</td><td align=\"left\">closed_Rpoly_prom_ns</td><td align=\"left\">k_on_Rpoly_prom_ns/(v/1000)</td><td align=\"left\">k_off_Rpoly</td></tr><tr><td align=\"left\">closed_Rpoly_prom_ns</td><td align=\"left\">open_Rpoly_prom_ns</td><td align=\"left\">k_isomerization</td><td/></tr><tr><td align=\"left\">open_Rpoly_prom_ns</td><td align=\"left\">operon_ns + Rpoly_operon_ns_1</td><td align=\"left\">k_prom_clearance</td><td/></tr><tr><td align=\"left\">Rpoly_operon_ns_1</td><td align=\"left\">Rpoly_operon_ns_2 + mRNA</td><td align=\"left\">k_transcription_ns</td><td/></tr><tr><td align=\"left\">Rpoly_operon_ns_2</td><td align=\"left\">Rpoly_operon_ns_3 + mRNA</td><td align=\"left\">k_transcription_ns</td><td/></tr><tr><td align=\"left\">Rpoly_operon_ns_3</td><td align=\"left\">Rpoly_operon_ns_4 + mRNA</td><td align=\"left\">k_transcription_ns</td><td/></tr><tr><td align=\"left\">Rpoly_operon_ns_4</td><td align=\"left\">Rpoly_operon_ns_5 + mRNA</td><td align=\"left\">k_transcription_ns</td><td/></tr><tr><td align=\"left\">Rpoly_operon_ns_5</td><td align=\"left\">Rpoly_operon_ns_6 + mRNA</td><td align=\"left\">k_transcription_ns</td><td/></tr><tr><td align=\"left\">Rpoly_operon_ns_6</td><td align=\"left\">Rpoly_operon_ns_7 + mRNA</td><td align=\"left\">k_transcription_ns</td><td/></tr><tr><td align=\"left\">Rpoly_operon_ns_7</td><td align=\"left\">Rpoly + mRNA_small</td><td align=\"left\">k_transcription_ns/0. 871794871794871</td><td/></tr><tr><td align=\"left\">operon_s + Rpoly</td><td align=\"left\">closed_Rpoly_prom_s</td><td align=\"left\">k_on_Rpoly_prom_s/(v/1000)</td><td align=\"left\">k_off_Rpoly</td></tr><tr><td align=\"left\">closed_Rpoly_prom_s</td><td align=\"left\">open_Rpoly_prom_s</td><td align=\"left\">k_isomerization</td><td/></tr><tr><td align=\"left\">open_Rpoly_prom_s</td><td align=\"left\">operon_s + Rpoly_operon_s</td><td align=\"left\">k_prom_clearance</td><td/></tr><tr><td align=\"left\">Rpoly_operon_s</td><td align=\"left\">Rpoly + stable_RNA</td><td align=\"left\">k_transcription_s</td><td/></tr><tr><td align=\"left\">V</td><td align=\"left\">2v</td><td align=\"left\">k_cell_div</td><td/></tr><tr><td/><td align=\"left\">operon_ns</td><td align=\"left\">k_rep_operon_ns</td><td/></tr><tr><td/><td align=\"left\">operon_s</td><td align=\"left\">k_rep_operon_s</td><td/></tr><tr><td/><td align=\"left\">Rpoly</td><td align=\"left\">k_rep_Rpoly</td><td/></tr><tr><td align=\"left\">plas + Rpoly</td><td align=\"left\">closed_Rpoly_prom_reporter</td><td align=\"left\">k_on_Rpoly_prom_reporter/(v/1000)</td><td align=\"left\">k_off_Rpoly</td></tr><tr><td align=\"left\">closed_Rpoly_prom_reporter</td><td align=\"left\">open_Rpoly_prom_reporter</td><td align=\"left\">k_isomerization</td><td/></tr><tr><td align=\"left\">open_Rpoly_prom_reporter</td><td align=\"left\">plas + Rpoly_reporter + incom_mRNA_reporter</td><td align=\"left\">k_prom_clearance</td><td/></tr><tr><td align=\"left\">Rpoly_reporter</td><td align=\"left\">Rpoly + mRNA_reporter</td><td align=\"left\">k_transcription_reporter</td><td/></tr><tr><td align=\"left\">incom_mRNA_reporter</td><td/><td align=\"left\">k_transcription_reporter</td><td/></tr><tr><td align=\"left\">mRNA_reporter</td><td/><td align=\"left\">k_deg_mRNA_reporter</td><td/></tr><tr><td align=\"left\">incom_mRNA_reporter</td><td/><td align=\"left\">k_deg_mRNA_reporter</td><td/></tr><tr><td align=\"left\">Rpoly_reporter</td><td align=\"left\">Rpoly</td><td align=\"left\">incom_mRNA_reporter*k_deg_mRNA_reporter/Rpoly_reporter</td><td/></tr></tbody></table></table-wrap>", "<table-wrap position=\"float\" id=\"T2\"><label>Table 2</label><caption><p>List of the biochemical reactions that make up our bacterial gene expression model (version excluding the host's genome). Gene expression model excluding the host's genome</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td align=\"left\">Left</td><td align=\"left\">Right</td><td align=\"left\">Forward rate constant</td><td align=\"left\">Backward rate constant</td></tr></thead><tbody><tr><td align=\"left\">V</td><td align=\"left\">2v</td><td align=\"left\">k_cell_div</td><td/></tr><tr><td/><td align=\"left\">Rpoly</td><td align=\"left\">k_rep_Rpoly</td><td/></tr><tr><td align=\"left\">plas + Rpoly</td><td align=\"left\">closed_Rpoly_prom_reporter</td><td align=\"left\">k_on_Rpoly_prom_reporter/(v/1000)</td><td align=\"left\">k_off_Rpoly</td></tr><tr><td align=\"left\">closed_Rpoly_prom_reporter</td><td align=\"left\">open_Rpoly_prom_reporter</td><td align=\"left\">k_isomerization</td><td/></tr><tr><td align=\"left\">open_Rpoly_prom_reporter</td><td align=\"left\">plas + Rpoly_reporter + incom_mRNA_reporter</td><td align=\"left\">k_prom_clearance</td><td/></tr><tr><td align=\"left\">Rpoly_reporter</td><td align=\"left\">Rpoly + mRNA_reporter</td><td align=\"left\">k_transcription_reporter</td><td/></tr><tr><td align=\"left\">incom_mRNA_reporter</td><td/><td align=\"left\">k_transcription_reporter</td><td/></tr><tr><td align=\"left\">mRNA_reporter</td><td/><td align=\"left\">k_deg_mRNA_reporter</td><td/></tr><tr><td align=\"left\">incom_mRNA_reporter</td><td/><td align=\"left\">k_deg_mRNA_reporter</td><td/></tr><tr><td align=\"left\">Rpoly_reporter</td><td align=\"left\">Rpoly</td><td align=\"left\">incom_mRNA_reporter*k_deg_mRNA_reporter/Rpoly_reporter</td><td/></tr></tbody></table></table-wrap>", "<table-wrap position=\"float\" id=\"T3\"><label>Table 3</label><caption><p>Species nomenclature used in biochemical models</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td align=\"center\">Species Label</td><td align=\"center\">Species</td></tr></thead><tbody><tr><td align=\"right\">Rpoly</td><td align=\"left\">RNA polymerase</td></tr><tr><td align=\"right\">closed_Rpoly_prom_ns*</td><td align=\"left\">RNA polymerase in a closed-complex with the average genomic mRNA operon promoter</td></tr><tr><td align=\"right\">closed_Rpoly_prom_s*</td><td align=\"left\">RNA polymerase in a closed-complex with the average genomic sRNA operon promoter</td></tr><tr><td align=\"right\">closed_Rpoly_prom_reporter</td><td align=\"left\">RNA polymerase in a closed-complex with the reporter mRNA promoter on the plasmid</td></tr><tr><td align=\"right\">open_Rpoly_prom_ns*</td><td align=\"left\">RNA polymerase in an open-complex with the average genomic mRNA operon promoter</td></tr><tr><td align=\"right\">open_Rpoly_prom_s*</td><td align=\"left\">RNA polymerase in an open-complex with the average genomic sRNA operon promoter</td></tr><tr><td align=\"right\">open_Rpoly_prom_reporter</td><td align=\"left\">RNA polymerase in an open-complex with the reporter mRNA promoter on the plasmid</td></tr><tr><td align=\"right\">Rpoly_operon_ns*</td><td align=\"left\">RNA polymerase elongating the average genomic mRNA transcript from the average genomic mRNA operon</td></tr><tr><td align=\"right\">Rpoly_operon_s*</td><td align=\"left\">RNA polymerase elongating the average genomic sRNA transcript from the average genomic sRNA operon</td></tr><tr><td align=\"right\">Rpoly_reporter</td><td align=\"left\">RNA polymerase elongating the reporter mRNA transcript from the plasmid</td></tr><tr><td align=\"right\">incom_mRNA*</td><td align=\"left\">Nascent average genomic mRNA</td></tr><tr><td align=\"right\">incom_mRNA_small*</td><td align=\"left\">Nascent genomic mRNA where its final length is approximately 90% of the average genomic mRNA</td></tr><tr><td align=\"right\">incom_mRNA_reporter</td><td align=\"left\">Nascent reporter mRNA</td></tr><tr><td align=\"right\">mRNA*</td><td align=\"left\">Average genomic mRNA (represented as the length of the average genomic mRNA gene)</td></tr><tr><td align=\"right\">mRNA_small*</td><td align=\"left\">Genomic mRNA that is approximately 90% of the average genomic mRNA</td></tr><tr><td align=\"right\">stable_RNA*</td><td align=\"left\">Average genomic sRNA (represented as the length of the average genomic sRNA operon)</td></tr><tr><td align=\"right\">mRNA_reporter</td><td align=\"left\">Reporter mRNA – the mRNA of interest</td></tr><tr><td align=\"right\">operon_ns*</td><td align=\"left\">Average genomic mRNA operon</td></tr><tr><td align=\"right\">operon_s*</td><td align=\"left\">Average genomic sRNA operon</td></tr><tr><td align=\"right\">plas</td><td align=\"left\">Reporter promoter on the plasmid</td></tr><tr><td align=\"right\">v</td><td align=\"left\">Cell volume</td></tr><tr><td align=\"right\">*</td><td align=\"left\">Refers to species used exclusively in the gene expression model that incorporates the host's genome</td></tr></tbody></table></table-wrap>", "<table-wrap position=\"float\" id=\"T4\"><label>Table 4</label><caption><p>Species used in both versions of the model</p></caption><table frame=\"hsides\" rules=\"groups\"><tbody><tr><td align=\"left\"><italic>Species Name</italic></td><td align=\"left\"><italic>Species</italic></td></tr><tr><td align=\"left\">Rpoly</td><td align=\"left\">RNA polymerase</td></tr><tr><td align=\"left\">Rpoly_reporter</td><td align=\"left\">RNA polymerase elongating the reporter mRNA transcript from the reporter gene</td></tr><tr><td align=\"left\">closed_Rpoly_prom_reporter</td><td align=\"left\">RNA polymerase in a closed-complex with the repoter promoter</td></tr><tr><td align=\"left\">deg_mRNA_incom_reporter</td><td align=\"left\">Nascent reporter mRNA degradation product</td></tr><tr><td align=\"left\">deg_mRNA_reporter</td><td align=\"left\">Reporter mRNA degradation product</td></tr><tr><td align=\"left\">incom_mRNA</td><td align=\"left\">Nascent reporter mRNA</td></tr><tr><td align=\"left\">mRNA_reporter</td><td align=\"left\">Reporter mRNA</td></tr><tr><td align=\"left\">open_Rpoly_prom_reporter</td><td align=\"left\">RNA polymerase in an open-complex with an mRNA reporter promoter</td></tr><tr><td align=\"left\">plas</td><td align=\"left\">Promoter on the plasmid</td></tr><tr><td align=\"left\">v</td><td align=\"left\">Counter (representing cell volume)</td></tr><tr><td align=\"left\" colspan=\"2\"><bold>Species used exclusively in the model containing the genome</bold></td></tr><tr><td align=\"left\">Rpoly_operon_ns</td><td align=\"left\">RNA polymerase elongating an average mRNA transcript from a template operon</td></tr><tr><td align=\"left\">Rpoly_operon_s</td><td align=\"left\">RNA polymerase elongating an average RNA transcript from a template operon</td></tr><tr><td align=\"left\">closed_Rpoly_prom_ns</td><td align=\"left\">RNA polymerase in a closed-complex with an mRNA operon promoter</td></tr><tr><td align=\"left\">closed_Rpoly_prom_s</td><td align=\"left\">RNA polymerase in a closed-complex with an mRNA operon promoter</td></tr><tr><td align=\"left\">mRNA</td><td align=\"left\">Average mRNA (gene length)</td></tr><tr><td align=\"left\">mRNA_small</td><td align=\"left\">Approximately 90% of the average mRNA; all species names that include \"small\" refer to this shorter species and its products/complexes</td></tr><tr><td align=\"left\">open_Rpoly_prom_ns</td><td align=\"left\">RNA polymerase in an open-complex with an mRNA operon promoter</td></tr><tr><td align=\"left\">open_Rpoly_prom_s</td><td align=\"left\">RNA polymerase in an open-complex with a stable RNA operon promoter</td></tr><tr><td align=\"left\">operon_ns</td><td align=\"left\">Average mRNA operon</td></tr><tr><td align=\"left\">operon_s</td><td align=\"left\">Average stable RNA operon</td></tr><tr><td align=\"left\">stable_RNA</td><td align=\"left\">Average stable RNA (full operon length)</td></tr></tbody></table></table-wrap>" ]

[ "<disp-formula id=\"bmcMR1\"><label>(R1)</label><italic>v </italic>→ 2<italic>v </italic></disp-formula>", "<disp-formula id=\"bmcMR2\"><label>(R2)</label>Rpoly + plas ⇔ closed_Rpoly_prom   K_R2= [k_on_Rpoly/(<italic>v</italic>/V₀)]/k_off_Rpoly</disp-formula>", "<disp-formula id=\"bmcMR3\"><label>(R3)</label>closed_Rpoly_prom → open_Rpoly_prom   k_R3</disp-formula>", "<disp-formula id=\"bmcMR4\"><label>(R4)</label>open_Rpoly_prom → Rpoly_operon + operon + incom_mRNA</disp-formula>", "<disp-formula id=\"bmcMR5\"><label>(R5)</label>Rpoly_operon → Rpoly + mRNA</disp-formula>", "<disp-formula id=\"bmcMR6\"><label>(R6)</label>() → Rpoly</disp-formula>", "<disp-formula id=\"bmcMR7\"><label>(R7)</label>() → plas</disp-formula>", "<disp-formula id=\"bmcMR8\"><label>(R8)</label>() → operon</disp-formula>", "<disp-formula id=\"bmcMR9\"><label>(R9)</label>mRNA_reporter → ()</disp-formula>", "<disp-formula id=\"bmcMR10\"><label>(R10)</label>incom_mRNA_reporter → ()</disp-formula>", "<disp-formula id=\"bmcMR11\"><label>(R11)</label>Rpoly_mRNA_reporter → Rpoly</disp-formula>", "<disp-formula><italic>k</italic>_R11= incom_mRNA_reporter·<italic>k</italic>_{mRNA_degradation}/Rpoly_mRNA_reporter</disp-formula>", "<disp-formula>Rpoly_operon1 → Rpoly_operon2 + mRNA   k = k_{transcription}</disp-formula>", "<disp-formula>Rpoly_operon2 → Rpoly_operon3 + mRNA   k = k_{transcription}</disp-formula>", "<disp-formula>Rpoly_operon3 → Rpoly + mRNA   k = k_{transcription}</disp-formula>" ]

[]

[]

[]

[]

[ "<supplementary-material content-type=\"local-data\" id=\"S1\"><caption><title>Additional file 1</title><p>Bionets files for the models. Files used to generate the stochastic simulations, using the Bionets stochastic simulation tool (required to read the files, and freely available from <ext-link ext-link-type=\"uri\" xlink:href=\"http://x.amath.unc.edu/BioNetS/\"/>). The ZIP file extracts to a directory containing files corresponding to the with-genome and no-genome versions of the model.</p></caption></supplementary-material>", "<supplementary-material content-type=\"local-data\" id=\"S2\"><caption><title>Additional file 2</title><p>MATLAB user interface. Files used to create the MATLAB user interface, allowing the user to enter four parameters (plasmid copy number, gene length, mRNA half-life, and bacterial cell doubling time), and get back the k_on rate above which excluding the genome will make less than a five percent difference in the simulated transcription levels of the plasmid-borne gene of interest. The ZIP file extracts to a directory containing three files that should be placed in the directory where the user interface will be used; the interface may be executed by opening MATLAB and running the script kon_gui.m.</p></caption></supplementary-material>" ]

[ "<table-wrap-foot><p>Gene expression model incorporating the host's genome. Lists of the biochemical reactions that make up our bacterial gene expression model, for the version of the model that includes an \"averaged\" version of the host cell's genome. Columns Left and Right represent the left and right sides of chemical reactions, and the Forward and Backward rate constants are associated with the forward and reverse reactions.</p></table-wrap-foot>", "<table-wrap-foot><p>Lists of the biochemical reactions that make up our bacterial gene expression model, for the version of the model that excludes the host cell's genome. Columns Left and Right represent the left and right sides of chemical reactions, and the Forward and Backward rate constants are associated with the forward and reverse reactions.</p></table-wrap-foot>", "<table-wrap-foot><p>List of species names used in the two versions of the model.</p></table-wrap-foot>" ]

[ "<graphic xlink:href=\"1471-2105-9-373-1\"/>", "<graphic xlink:href=\"1471-2105-9-373-2\"/>", "<graphic xlink:href=\"1471-2105-9-373-3\"/>", "<graphic xlink:href=\"1471-2105-9-373-4\"/>", "<graphic xlink:href=\"1471-2105-9-373-5\"/>", "<graphic xlink:href=\"1471-2105-9-373-6\"/>", "<graphic xlink:href=\"1471-2105-9-373-7\"/>", "<graphic xlink:href=\"1471-2105-9-373-8\"/>", "<graphic xlink:href=\"1471-2105-9-373-9\"/>", "<graphic xlink:href=\"1471-2105-9-373-10\"/>", "<graphic xlink:href=\"1471-2105-9-373-11\"/>" ]

[ "<media xlink:href=\"1471-2105-9-373-S1.zip\" mimetype=\"application\" mime-subtype=\"x-zip-compressed\"><caption><p>Click here for file</p></caption></media>", "<media xlink:href=\"1471-2105-9-373-S2.zip\" mimetype=\"application\" mime-subtype=\"x-zip-compressed\"><caption><p>Click here for file</p></caption></media>" ]

{ "acronym": [], "definition": [] }

[ "<title>Background</title>", "<p>Genome-wide gene expression profiling using microarray technologies has been ubiquitously used in biological research. An important problem is to identify gene sets that are significantly changed under a certain treatment (for example, two different cell lines or tissues or the same cell line under different conditions). A gene set is basically a group of genes with related functions, e.g., genes in a biological process or in the same complex. There are a variety of ways by which genes, and, ultimately, gene sets may be defined. For example, gene sets can be defined according to the information provided by several databases, such as GeneOntology [##REF##10802651##1##], KEGG [##REF##9847135##2##], Biocarta <ext-link ext-link-type=\"uri\" xlink:href=\"http://www.biocarta.com\"/>, and Pfam [##REF##9223186##3##]. Gene sets may also be defined by cytogenetic bands, by region of genomic sequence or by establishing the functional relationships among them. Importantly, by using a gene set-based approach, a high power can potentially be achieved for detecting differentially expressed gene sets by integrating expression changes of genes inside the same gene set, even when the expression changes of individual genes are modest.</p>", "<p>Moreover, because the gene sets have already been annotated by their common functions in the databases, the biological interpretation for a given list of significant gene sets will also be clear. At least one study [##REF##15941488##4##] showed that using such gene set-based approaches did increase the congruence of the identified gene sets between different data sets addressing the same biological problem. To detect differentially expressed gene sets, several methods have been proposed, which can be roughly categorized into three groups.</p>", "<p>The first group identifies a list of significant differentially expressed genes (DEGs) using individual gene analysis methods, and then examines the enrichment of gene sets within this gene list using different statistical tests, such as the binomial test, Fisher's exact test, or the hypergeometric test [##REF##14668247##5##, ####REF##14519205##6##, ##REF##12540299##7##, ##UREF##0##8##, ##UREF##1##9##, ##REF##14990455##10##, ##REF##14962934##11####14962934##11##]. Khatri and Draghici [##REF##15994189##12##] compared fourteen different methods within this group. Each of these methods is easy to implement, but flawed by 1) sensitivity to the cutoff value for defining the list of significant DEGs, 2) non-consideration of the relative position of genes inside the significant DEG list, and 3) assumption of independence between the genes, which may make the resulting p-value misleading.</p>", "<p>The second group of methods does not depend on the predefined DEG list. Instead, these methods calculate a gene-specific statistic, known as the \"local\" statistic, which measures the strength of association between the gene expression and the phenotype for each gene. A \"global\" statistic for a gene set is then constructed as a function of the local statistic for each gene in it. The significance of the global statistic is assessed by permutation test, and different methods arrive at this assessment in different ways [##REF##12808457##13##, ####REF##16199517##14##, ##REF##16174746##15##, ##REF##17303618##16##, ##UREF##2##17##, ##REF##15176478##18##, ##REF##17127676##19##, ##UREF##3##20##, ##REF##15647293##21####15647293##21##]. In contrast to calculating a gene-specific statistic, the third group of methods directly combines the expression levels of all genes in the gene sets and they are represented as gene set-specific features. These features are then compared between the treatment and the control groups to identify significantly affected gene sets [##REF##16156896##22##, ####REF##14693814##23##, ##REF##15657105##24##, ##UREF##4##25##, ##REF##16772399##26##, ##REF##17044931##27####17044931##27##]. Some methods also integrated the interaction information between genes in the gene sets [##REF##17483504##28##, ####REF##17571924##29##, ##UREF##5##30##, ##REF##17267429##31####17267429##31##].</p>", "<p>The available methods generally tested the association of all the genes in a gene set with the phenotype. In reality, however, it is more likely that only genes in a subset of the gene set of interest are associated with the phenotype. Three possible factors may explain this. First, since the function annotation defined in the available databases, such as GO, KEGG, and Biocarta, are incomplete, or even erroneous, some of the genes in a gene set of interest may not truly belong to the set. Second, the gene sets are sometimes defined according to the genomic regions of the genes. Thus, the expression of the genes in the same set may not coordinate with each other. For example, although a group of gene sets has been defined by the cytogenetic bands on human chromosomes [##REF##16199517##14##], the expression of genes on the same cytogenetic bands do not necessarily correlate with each other. The result is that only a subset of genes in a given cytogenetic band will then be correlated with a phenotype. Third, even if all the genes in the gene set have the same function, or belong to the same complex, it is possible that only genes in one branch of the pathway are associated with the phenotype. The cumulative effect of these considerations strongly suggests that the currently available methods for gene set enrichment analysis may not be powerful enough to detect the association of a given gene set with a phenotype, particularly in the case where only a subset of the genes is associated with the phenotype.</p>", "<p>In this paper, therefore, we extend a set-association strategy for genetic polymorphism association studies developed by [##REF##11731502##32##] to a set-enrichment analysis. In so doing, we want to test the null hypothesis that no subsets of genes in the gene set are associated with the phenotype. We refer to the resulting method as Gene Set Enrichment by testing Subset association, or Sub-GSE. Using two simulated data sets, we first show that Sub-GSE has higher sensitivity in identifying gene sets associated with a phenotype compared to GSEA, SigPath, and GSA, when only a fraction of the genes are associated with the phenotype. Next, we apply Sub-GSE to two real data sets. One involves gene expression data related to gender and the other identified functional gene sets related to p53 mutation status. For the first dataset, Sub-GSE identified cytogenetic bands Xp22 as significantly associated with gender, while GSEA failed to detect this association. For the second dataset, Sub-GSE identified several novel functional gene sets, including DNA damage genes, cell cycle checkpoints genes and programmed cell death genes that are associated with p53 mutation status and that were also not detected by GSEA. Overall, this method provides a complementary approach for identifying gene sets associated with a phenotype, especially when only a subset of genes in a gene set is associated with the phenotype.</p>" ]

[ "<title>Methods</title>", "<title>Association strength measures for individual genes</title>", "<p>For a given gene, suppose the gene expression levels measured in the experiment are (<italic>e</italic>₁, <italic>e</italic>₂, ⋯, <italic>e</italic>_{<italic>m</italic>}), where <italic>m </italic>is the number of samples. The corresponding phenotypic data for the <italic>m </italic>samples are denoted as <italic>C </italic>= (<italic>c</italic>₁, <italic>c</italic>₂,⋯, <italic>c</italic>_{<italic>m</italic>}). Depending on the measurement levels of <italic>C</italic>, we measure the association strength between the gene expression and phenotype by the absolute value of t-statistic, Kruskal-Wallis statistic or Pearson correlation coefficient for binary, discrete or continuous phenotypic data, respectively.</p>", "<title>Local association statistic for each strict subset and the global statistic for the gene set</title>", "<p>Suppose the given gene set has a total of <italic>s </italic>genes and the sorted association strength measures are <italic>a</italic>₁≥ <italic>a</italic>₂≥ ⋯ ≥ <italic>a</italic>_{<italic>s</italic>}. According to the definition of the strict subsets, the <italic>i</italic>-th strict subset include genes that correspond to (<italic>a</italic>₁, <italic>a</italic>₂,⋯, <italic>a</italic>_{<italic>i</italic>}). The local association statistic for the strict subset <italic>i </italic>is defined as</p>", "<p></p>", "<p>We use the permutation test described below to define a p-value, <italic>p</italic>_{<italic>i </italic>}for the <italic>i</italic>-th strict set. The statistic for the given gene set is the minimum p-value over all the strict sets, i.e.</p>", "<p></p>", "<title>Permutation test</title>", "<p>We employ the algorithm in [##REF##15290652##41##] to assess the significance of the given gene set. The algorithm permutes the phenotypic data <italic>C </italic>= (<italic>c</italic>₁, <italic>c</italic>₂,⋯, <italic>c</italic>_{<italic>m</italic>}) for <italic>N </italic>times and keep the gene expression data intact. After each permutation, the association strength measures are re-calculated using the permuted phenotypic data and the \"strict subsets\" are re-defined. Suppose that the statistics for the <italic>n</italic>-th permutation and the <italic>i</italic>-th strict subset are organized as</p>", "<p></p>", "<p>where <italic>c </italic>is the minimum number of genes in the strict sets. Here we take the observed data as the permutation 0.</p>", "<p>Based on the data matrix, we calculate the raw p-value for the <italic>i</italic>-th strict set of the observed data as</p>", "<p></p>", "<p>The testing statistic for the gene set is .</p>", "<p>To estimate the distribution of <italic>P</italic>_{<italic>min </italic>}under the null hypothesis, we replace the observed \"strict subset\" statistics with the permuted ones. The permuted raw p-values for the <italic>n</italic>-th permutation and <italic>i</italic>-th strict subset are calculated by</p>", "<p></p>", "<p>The minimum p-value is taken as one of the random sample from the distribution of <italic>P</italic>_{<italic>min </italic>}under the null hypothesis. Finally the significance of the gene set is calculated as</p>", "<p></p>", "<title>Multiple testing correction for multiple gene sets</title>", "<p>In a typical situation, there will be multiple gene sets to be analyzed. After we assess the significance level for each of them according to the procedure described above, q-values of all the given gene sets are calculated using the QVALUE R package [##UREF##7##34##] for multiple testing correction.</p>" ]

[ "<title>Results and discussion</title>", "<p>In this section, we first give a brief overview of our method. Second, we apply Sub-GSE, GSEA, SigPathway, and GSA to two simulated data sets and compare their performance. Third, we apply Sub-GSE to two real datasets: one related to gender and the other related to p53 mutation status. We also show some new biological findings related to the two real data sets using Sub-GSE.</p>", "<title>Outline of the method</title>", "<p>To assess the enrichment of a given gene set, we construct a statistical hypothesis testing model. The null hypothesis is that no subsets of genes in the given gene set are associated with the phenotype.</p>", "<title>Defining the \"strict subsets\"</title>", "<p>Given the fact that the number of all subsets of a gene set increases exponentially with the number of genes in the gene set, it is impractical and less powerful to test every subset of the gene set. Therefore, we define the \"strict subsets\" to only those subsets that are more likely to be related with the phenotype. To define the \"strict subsets\" of a gene set, we first calculate the association strength between the gene expression and the phenotype for each gene in the gene set. Depending on the measurement levels of the phenotypic data, we calculate the absolute t-statistics for comparing the mean gene expression levels for binary phenotypic data, Kruskal-Wallis statistics for comparing the mean expression levels of different groups for discrete phenotypic data, and the absolute Pearson correlation coefficient between the gene expression levels and the phenotype for continuous data. All genes are sorted in decreasing order of the association strength measures. The \"strict subsets\" are defined to include genes up to each position from the top to the bottom in the ranked gene list, which are most strongly associated with the phenotype. That is to say, for each position in the ranked gene list, we define a strict subset that includes all the genes that are ranked higher than this position. Thus, if there are <italic>n </italic>genes in the gene set, there will be <italic>n </italic>\"strict subsets\" among which the <italic>i</italic>-th subset contains the top <italic>i </italic>genes in the ranked gene list according to the association strengths. In this way, the number of subsets to be tested increases linearly with the number of genes in the gene set. Since the strict subsets includes the genes that are most associated with the phenotype, we expect them to be more probable to be related with the phenotype. The strict subsets are defined to be contiguous to include as many as possible subsets that are expected to be more likely related with the phenotype. However, the method we propose here cannot detect the gene sets in which individual genes are not associated with the phenotype but they can interact with each other to affect the phenotype. To overcome the problem that the \"strict subsets\" contain too few genes, we add a tuning parameter to control the sizes of the \"strict subsets\". Throughout the paper, we set this tuning parameter to be 5 which means the \"strict subsets\" are required to contain at least 5 genes. The cutoff for the set size of the strict subsets is set to be 5 so that the method is not too sensitive to detect gene set which has only one gene strongly correlated with the phenotype. There are other ways of deciding the cutoff of the set size as discussed in the Conclusions section.</p>", "<title>Testing Statistic</title>", "<p>The hypothesis testing statistic is calculated in three steps. First, for each \"strict subset\", we calculate the average association strength across all member genes, which is also called the local set association statistic <italic>T</italic>. Second, the statistical significance (raw p-value) of the local set association statistic <italic>T </italic>for each \"strict subset\" is calculated by permuting the phenotypes of the individuals. Finally, the minimum raw p-value among all the \"strict subsets\" is evaluated and taken as the hypothesis testing statistic. If there is any strict subset related with the phenotype, the minimum p-value will be significantly small.</p>", "<title>Significance Assessment</title>", "<p>To assess the significance of the minimum p value, nested permutation is needed since we do not know the distribution of <italic>T </italic>under the null hypothesis However, nested permutation is computation intensive. Fortunately, previous work [##UREF##6##33##] has shown that a single set of permutation is sufficient to accomplish the significance assessment. For the permutation, we decide to permute the phenotypic data and keep the gene expression data intact due to the criticism on gene-based permutation which assumes the independence between genes [##REF##17303618##16##].</p>", "<p>The phenotypic data is permuted for <italic>N </italic>times. After each permutation, the \"strict subsets\" are re-defined according to the newly calculated association strengths using the permuted phenotypic data. The strict subsets are defined in the same way as we did with the observed data including the threshold of set size.</p>", "<p>The only difference is that the phenotypic data is changed. By comparing the set association statistic <italic>T </italic>from the observed data and those from the permuted data, raw p values can be calculated for all the observed \"strict subsets\" and thus the observed <italic>P</italic>_{<italic>min </italic>}is obtained. To estimate the distribution of <italic>P</italic>_{<italic>min </italic>}under the null hypothesis, as classic permutation does, we replace the observed phenotypic data with every permuted phenotypic data and compare the set association statistic with those from all the other <italic>N </italic>- 1 permutations. In other words, we repeat exactly the same procedure to obtain the minimum raw p values for every permuted data. Finally, the significance of the gene set will be the percentage of permutations that result in minimum raw p values smaller than the observed <italic>P</italic>_{<italic>min</italic>}. If there are more than one given gene set, multiple testing correction can be done using any multiple testing correction method. In this paper, we use the QVALUE R package [##UREF##7##34##] to calculate the q-values for the two biological data sets so that the results by Sub-GSE are comparable with other gene set enrichment analysis methods.</p>", "<title>Simulation Studies</title>", "<p>We first elucidate that the p-values do not depend on the size of the gene set and the p-value has a uniform distribution in [0,1] under the null hypothesis. To achieve these objectives, we simulate a data set where all the gene sets have different set sizes and no gene sets are related to the phenotype. The simulation is implemented in the following steps:</p>", "<p>1. Generate 100 gene sets whose sizes are 5,6,7,8,...,104. The total number of genes is 5450;</p>", "<p>2. The gene expression levels in 100 samples for each gene are generated from a standard normal distribution. Different genes are independent of each other. Different samples are also independent of each other;</p>", "<p>3. Generate the phenotypic data from another independent standard normal distribution in 100 samples;</p>", "<p>4. Repeat steps 1–3 for 100 times;</p>", "<p>In total, the simulation generates 100 data sets that have gene expression data and a corresponding phenotypic data. We apply Sub-GSE to the 100 data sets separately.</p>", "<p>First, since the gene sets have different sizes, we plot the average p-values of all the gene sets across the 100 different data sets against their set sizes to see whether the gene set size affects the significance level. Figure ##FIG##0##1## shows that the set size does not affect the p-values.</p>", "<p>Second, the phenotypic data is independent of the expression levels of all the genes. Therefore, Sub-GSE should not detect any significant gene sets. In Figure ##FIG##1##2##, the histogram of all the p-values of the 100 gene sets from the 100 data sets is shown. The histogram illustrates that the p-values from the Sub-GSE have a uniform distribution for gene sets that are not related to the phenotype, which is consistent with the theoretical uniform distribution under the null hypothesis.</p>", "<title>Simulation 1</title>", "<p>We first evaluate the performance of Sub-GSE using simulated data in which gene expression profiles with different correlations within the gene set are generated. The expression profiles for 1000 genes in 100 samples are simulated. The genes are divided into 50 non-overlapping gene sets with 20 genes in each. The gene expression profiles for the 100 samples represent 100 independent vectors of random variables generated from a multivariate normal distribution. The multivariate normal distribution has 1000 dimensions corresponding to the 1000 genes. The mean is a vector of 1000 zeroes, and the variance of the expression levels of each gene is 1. To simulate the dependence between genes, we randomly select a certain percentage of correlated genes (PCG) = (0%, 10%, ⋯, 90%) in each gene set and let the correlation coefficient between any two of them be <italic>ρ </italic>= 0, 0.1, 0.2, ⋯, 0.9. The remaining genes are independent of each other and those that are chosen. We use this simulation strategy based on the following considerations. The chosen genes in the gene set correspond to those in the same complex or pathway; thus, their expression profiles are correlated. Also, since the remaining genes represent those not belonging to the group, they are more likely to be independent of the chosen genes and each other.</p>", "<p>If a given gene is among those that are chosen, we use its expression levels as the phenotype. The rationale of this step is to determine if our Sub-GSE method can identify the gene set to which this particular gene belongs. We repeat this process for all the chosen genes. Thus, we have a total of 1000 × <italic>PCG </italic>different phenotypic data. To avoid the problem where a gene has exactly the same expression profile as the phenotype, we eliminate the gene's expression profile from the expression data if it is used as the phenotype.</p>", "<p>We use the following approach to study the robustness of Sub-GSE. For each given correlation coefficient and PCG, we randomly choose one of the simulated phenotypic data and the corresponding gene expression data. Sub-GSE is applied to the chosen data set for 100 times. The standard deviations of the p-values across the 100 different runs are plotted against the average p-values for all the gene sets in Figure ##FIG##2##3##. The figure shows that the standard deviation of the p-value for the same gene set is smaller than 0.006 and even smaller when the p-value is close to either 1 or 0. The closer the p-value is to 0 or 1, the smaller the standard deviation is. The maximum standard deviation is achieved when the average p-value is around 0.5.</p>", "<p>For each given pair of percentage of correlated genes (<italic>PCG</italic>) and correlation coefficient, we apply all four methods, Sub-GSE, GSEA, GSA, and SigPath, to the corresponding data. All the gene sets are ranked in an increasing order of their q-values so more significant gene sets have smaller rank. The rank of the gene set, some of whose member genes are correlated with the phenotypic data, is extracted to evaluate the performance of the methods. Figure ##FIG##3##4## shows the average rank of the gene set related to the phenotype for different combinations of <italic>PCG </italic>and correlation coefficient.</p>", "<p>First, as seen in the left panel in Figure ##FIG##3##4##, for small <italic>PCG </italic>= 10%, 20% and 30%, the average rank of the gene set related to the phenotype based on Sub-GSE is always the lowest, irrespective of the coefficient value. On the other hand, for large <italic>PCG</italic>, the performance of Sub-GSE is similar or slightly worse than GSEA and GSA for small correlation. The right panel in Figure ##FIG##3##4## confirms this because the average rank of the gene set related to the phenotype based on Sub-GSE decreases much faster than those for the other methods when <italic>PCG </italic>is small. The results of Figure ##FIG##3##4## can be explained as follows. When <italic>PCG </italic>is low, only a small fraction of the genes in the target gene set are correlated with the phenotype. GSEA, GSA, and SigPath cannot distinguish the target gene set from the other gene sets since these methods consider all the genes in the gene set of interest in their statistics. In contrast, Sub-GSE incrementally tests each strict set and chooses the smallest p-value across all the strict sets as a test statistic, thus making the test more powerful.</p>", "<p>Second, across different combinations of <italic>PCG </italic>and correlation coefficient, we find that GSA and GSEA achieve similar results. Both GSA and GSEA use t-statistics to obtain the ranking list of genes. For applications in this article, the only diference between them is that GSA restandardizes the statistics before the permutation to reduce the effect of correlation between genes. However, in both panels of Figure ##FIG##3##4##, the average ranks of the target gene set by GSEA and GSA are quite similar, especially when the <italic>PCG </italic>is high. Consequently, restandardization in GSA does not seem to be very efficient in this simulation study, especially when there are many correlated genes.</p>", "<p>Third, to show the sensitivity and specificity of Sub-GSE, we need a group of gene sets that are related with the phenotype. Therefore, we do another set of simulations similar as simulation 1. The detailed descriptions of the simulation and the resulting ROC curves can be found in the supplementary materials [see Additional file ##SUPPL##0##1##]. The results show that the higher the PCG and the correlation coefficient are, the higher the AUC score is. Once the correlation coefficient is higher than 0.4, the AUC score is higher than 0.85 no matter what the PCG is. When PCG is higher than 0.5, the AUC score can be higher than 0.75 regardless of the correlation coefficient.</p>", "<title>Simulation 2</title>", "<p>In reality, most phenotypes are the joint effect of multiple genes, probably from multiple pathways. Therefore, we also simulate a more realistic case where the phenotypes are assumed to be a complex function of expression levels from two gene sets. As in simulation 1, we again consider 1000 genes divided equally into 50 non-overlapping gene sets of 20 genes in each. For fixed <italic>PCG </italic>and <italic>ρ</italic>,</p>", "<p>1. Simulate the expression profiles of the 1000 genes as in the first simulation for 100 individuals;</p>", "<p>2. Choose two gene sets <italic>K</italic>₁and <italic>K</italic>₂from the 50 gene sets. Let <italic>SK</italic>₁and <italic>SK</italic>₂be the correlated genes in <italic>K</italic>₁and <italic>K</italic>₂, respectively. Define the phenotype for the <italic>j</italic>-th individual as</p>", "<p></p>", "<p>where <italic>ϵ</italic>_{<italic>j </italic>}has a normal distribution with mean 0 and variance 0.25.</p>", "<p>3. Analyze the data using GSEA, SigPath, GSA, and Sub-GSE to rank the gene sets. Rank all the gene sets in increasing order of their q-values.</p>", "<p>4. Repeat steps 1–3 100 times to assess the performance of the different analytic methods by the effects of the different gene expression data.</p>", "<p>We study the robustness of Sub-GSE as follows. Similar to the process in simulation 1, for each given correlation coefficient and PCG, we randomly choose a phenotypic data and the corresponding gene expression data. Sub-GSE is applied to the chosen data set for 100 times. The standard deviation of the p-values across the 100 runs for all the gene sets is plotted against the average p-values in Figure ##FIG##4##5##. Again, the standard deviation of the p-values across different runs for each gene set is smaller than 0.006. The closer the average p-value is to either 0 or 1, the smaller the standard deviation is. The maximum standard deviation is achieved when the average p-value is around 0.5.</p>", "<p>For this simulation study, we again apply the four different methods to prioritize the gene sets as in the first simulation study and calculate the average rank of the two target gene sets. The results can be found in Figure ##FIG##5##6##.</p>", "<p>As shown in Figure ##FIG##5##6##, the average ranks of the target gene sets based on all the methods are relatively high. This could result from the involvement of two different gene sets when simulating the phenotypic data and the fact that the phenotypic data are the sum of the squared expression levels of correlated genes. Another potential complicating factor is that the phenotype includes a noise in addition to the function of the gene expression levels of the component genes. All these facts can weaken the correlation between the phenotypic data and the gene expression profile of individual genes inside the true gene sets. Despite these problems, Sub-GSE performs relatively well compared to the other three methods, especially when <italic>PCG </italic>is low. When <italic>PCG </italic>is high, the performances of Sub-GSE are close to those of GSEA and GSA since both GSEA and GSA consider all the genes inside the gene sets. Again, the performances of GSEA and GSA are similar.</p>", "<title>Male Vs. Female Lymphoblastoid Cells</title>", "<p>We also apply Sub-GSE to two real data sets from [##REF##16199517##14##]. The first data set measured the mRNA expression profiles from lymphoblastoid cells derived from 15 males and 17 females using Affymetrix U133A chip. The gender of the individuals represents the corresponding phenotypic data. The gene sets are chosen as the cytogenetic sets (C1, 319 gene sets) and the functional gene sets (C2, 522 gene sets) defined in [##REF##16199517##14##]. The cytogenetic sets contain 24 gene sets, one for each of the 24 human chromosomes, and 295 gene sets corresponding to cytogenetic bands along the chromosomes. The functional sets include 472 gene sets containing genes whose products are involved in specific metabolic and signaling pathways, as reported in eight publicly available, manually curated databases, and 50 gene sets containing genes co-expressed in response to genetic and chemical perturbations, as reported in various experimental studies (see supporting text in [##REF##16199517##14##] for details). We apply Sub-GSE, GSEA, and SigPath to these two types of gene sets independently with the objective of identifying the cytogenetic regions that are differentially expressed between males and females and the functional gene sets related to sex distinction, respectively.</p>", "<p>First, we apply Sub-GSE to investigate the enrichment of cytogenetic gene sets (C1). As expected, the three most significant cytogenetic bands are chrY, chrYp11 and chrYq11 which all have a q-value of 0 and are the only three cytogenetic bands from chromosome Y in gene sets C1. They are also the only three significant gene sets in C1 by GSEA (FDR &lt; 0.2) and SigPath (max q-value &lt; 0.2). Besides these expected bands on chromosome Y, other bands that are ranked as the top 7 among all the gene sets by Sub-GSE, GSEA and SigPath are listed in Table ##TAB##0##1##. As seen from the lists, Sub-GSE is sensitive enough to identify cytogenetic bands on both chromosomes X and Y at the q-value threshold of 0.20. On the contrary, neither GSEA nor SigPath is able to detect any bands on chromosome X at the FDR threshold of 0.20. Again, this result shows the sensitivity of Sub-GSE.</p>", "<p>Second, we apply Sub-GSE to investigate the enrichment of functional gene sets (C2). Both Sub-GSE and GSEA detect three significant gene sets whose significance levels are listed in Table ##TAB##1##2##: testis-related genes, genes that escape X inactivation, and female reproductive tissue-expressed genes. The q-values of all the other gene sets are larger than 0.9 by Sub-GSE. Hence, in this dataset, the results by Sub-GSE are roughly the same as those achieved by GSEA.</p>", "<title>P53 Status in Cancer Cell Lines</title>", "<p>The second real data set corresponds to the gene expression data and phenotypic data related to p53 mutation status from [##REF##16199517##14##]. The objective of this study is to identify novel targets of the transcription factor p53. The p53 mutation status gene expression data examined the gene expression patterns from the NCI-60 collection of cancer cell lines. The expression profiles were measured using Affymetrix U95Av2 chips. The mutational status of the p53 gene had been reported for 50 of the NCI-60 cell lines, with 17 being classified as normal and 33 as carrying mutations in the gene. We take the gene expression profiles of these 50 cell lines as the expression data and the vector of binary indicators of the mutational patterns (normal or mutated) as the corresponding phenotypic data. For the gene set data, we only use the functional sets (C2, 522 gene sets) in [##REF##16199517##14##], which was already described in the application noted above. Functional sets that have a q-value smaller than or equal to 0.03 by Sub-GSE are extracted and listed in Table ##TAB##2##3## together with their q-values. Comparing the list of significant gene sets by Sub-GSE and GSEA [##REF##16199517##14##], we can see that Sub-GSE obtains more significant gene sets than GSEA does, which again shows the sensitivity of Sub-GSE. The relationship between the identified gene sets by Sub-GSE (Table ##TAB##2##3##) and p53 is illustrated in Figure ##FIG##6##7##. Basically, all these gene sets are significantly enriched in genes that are differentially expressed in p53 mutants versus those without p53 mutations. Therefore, the identified gene sets by Sub-GSE are potentially those regulated by p53. According to the definitions of these gene sets, as shown in Figure ##FIG##6##7##, we can roughly divide them into three groups.</p>", "<p>The first group includes gene sets that are directly regulated or affected by p53, including the \"p53 signaling pathway\", \"p53 signaling pathway genes\", and \"p53 upregulated genes\". This group of gene sets was detected by both Sub-GSE and GSEA [##REF##16199517##14##] (FDR &lt; 0.015).</p>", "<p>The second group contains gene sets that are \"downstream\" of p53. These gene sets can either be induced or inhibited by p53 [##REF##11099028##35##, ####REF##16741928##36##, ##REF##9664074##37##, ##REF##12670900##38####12670900##38##]. For example, it is well known that p53 induces cell cycle arrest during the G1/S phase and the G2/M phase checkpoint [##REF##16741928##36##]. By itself, p53 can activate an important death receptor, Fas, which triggers the \"Fas Signaling Pathway\" and thus leads to apoptosis [##REF##12670900##38##]. It is also well known that p53 functions \"upstream\" of ceramide in response to genotoxic stress [##REF##9664074##37##].</p>", "<p>The third group includes gene sets related to the \"upstream\" biological processes or genes for p53. These \"upstream\" biological processes, such as DNA damage caused by radiation or chemical carcinogens, for example, pass the DNA damage signal down to p53 and further induce some of the \"downstream\" pathways. Two genes, TrkA and Pitx2, are known to affect apoptosis through regulation of p53 [##REF##9852160##39##,##REF##16129685##40##]. The gene sets related to these \"upstream\" biological processes actually include genes related to those \"downstream\" biological processes in the second group.</p>", "<p>In this dataset, Sub-GSE not only detects the gene sets identified by GSEA [##REF##16199517##14##], but also detects more novel gene sets related to p53. Previous studies from the literature support the findings in that all the significant gene sets identified by Sub-GSE are related to p53, as shown in Figure ##FIG##6##7##.</p>" ]

[ "<title>Results and discussion</title>", "<p>In this section, we first give a brief overview of our method. Second, we apply Sub-GSE, GSEA, SigPathway, and GSA to two simulated data sets and compare their performance. Third, we apply Sub-GSE to two real datasets: one related to gender and the other related to p53 mutation status. We also show some new biological findings related to the two real data sets using Sub-GSE.</p>", "<title>Outline of the method</title>", "<p>To assess the enrichment of a given gene set, we construct a statistical hypothesis testing model. The null hypothesis is that no subsets of genes in the given gene set are associated with the phenotype.</p>", "<title>Defining the \"strict subsets\"</title>", "<p>Given the fact that the number of all subsets of a gene set increases exponentially with the number of genes in the gene set, it is impractical and less powerful to test every subset of the gene set. Therefore, we define the \"strict subsets\" to only those subsets that are more likely to be related with the phenotype. To define the \"strict subsets\" of a gene set, we first calculate the association strength between the gene expression and the phenotype for each gene in the gene set. Depending on the measurement levels of the phenotypic data, we calculate the absolute t-statistics for comparing the mean gene expression levels for binary phenotypic data, Kruskal-Wallis statistics for comparing the mean expression levels of different groups for discrete phenotypic data, and the absolute Pearson correlation coefficient between the gene expression levels and the phenotype for continuous data. All genes are sorted in decreasing order of the association strength measures. The \"strict subsets\" are defined to include genes up to each position from the top to the bottom in the ranked gene list, which are most strongly associated with the phenotype. That is to say, for each position in the ranked gene list, we define a strict subset that includes all the genes that are ranked higher than this position. Thus, if there are <italic>n </italic>genes in the gene set, there will be <italic>n </italic>\"strict subsets\" among which the <italic>i</italic>-th subset contains the top <italic>i </italic>genes in the ranked gene list according to the association strengths. In this way, the number of subsets to be tested increases linearly with the number of genes in the gene set. Since the strict subsets includes the genes that are most associated with the phenotype, we expect them to be more probable to be related with the phenotype. The strict subsets are defined to be contiguous to include as many as possible subsets that are expected to be more likely related with the phenotype. However, the method we propose here cannot detect the gene sets in which individual genes are not associated with the phenotype but they can interact with each other to affect the phenotype. To overcome the problem that the \"strict subsets\" contain too few genes, we add a tuning parameter to control the sizes of the \"strict subsets\". Throughout the paper, we set this tuning parameter to be 5 which means the \"strict subsets\" are required to contain at least 5 genes. The cutoff for the set size of the strict subsets is set to be 5 so that the method is not too sensitive to detect gene set which has only one gene strongly correlated with the phenotype. There are other ways of deciding the cutoff of the set size as discussed in the Conclusions section.</p>", "<title>Testing Statistic</title>", "<p>The hypothesis testing statistic is calculated in three steps. First, for each \"strict subset\", we calculate the average association strength across all member genes, which is also called the local set association statistic <italic>T</italic>. Second, the statistical significance (raw p-value) of the local set association statistic <italic>T </italic>for each \"strict subset\" is calculated by permuting the phenotypes of the individuals. Finally, the minimum raw p-value among all the \"strict subsets\" is evaluated and taken as the hypothesis testing statistic. If there is any strict subset related with the phenotype, the minimum p-value will be significantly small.</p>", "<title>Significance Assessment</title>", "<p>To assess the significance of the minimum p value, nested permutation is needed since we do not know the distribution of <italic>T </italic>under the null hypothesis However, nested permutation is computation intensive. Fortunately, previous work [##UREF##6##33##] has shown that a single set of permutation is sufficient to accomplish the significance assessment. For the permutation, we decide to permute the phenotypic data and keep the gene expression data intact due to the criticism on gene-based permutation which assumes the independence between genes [##REF##17303618##16##].</p>", "<p>The phenotypic data is permuted for <italic>N </italic>times. After each permutation, the \"strict subsets\" are re-defined according to the newly calculated association strengths using the permuted phenotypic data. The strict subsets are defined in the same way as we did with the observed data including the threshold of set size.</p>", "<p>The only difference is that the phenotypic data is changed. By comparing the set association statistic <italic>T </italic>from the observed data and those from the permuted data, raw p values can be calculated for all the observed \"strict subsets\" and thus the observed <italic>P</italic>_{<italic>min </italic>}is obtained. To estimate the distribution of <italic>P</italic>_{<italic>min </italic>}under the null hypothesis, as classic permutation does, we replace the observed phenotypic data with every permuted phenotypic data and compare the set association statistic with those from all the other <italic>N </italic>- 1 permutations. In other words, we repeat exactly the same procedure to obtain the minimum raw p values for every permuted data. Finally, the significance of the gene set will be the percentage of permutations that result in minimum raw p values smaller than the observed <italic>P</italic>_{<italic>min</italic>}. If there are more than one given gene set, multiple testing correction can be done using any multiple testing correction method. In this paper, we use the QVALUE R package [##UREF##7##34##] to calculate the q-values for the two biological data sets so that the results by Sub-GSE are comparable with other gene set enrichment analysis methods.</p>", "<title>Simulation Studies</title>", "<p>We first elucidate that the p-values do not depend on the size of the gene set and the p-value has a uniform distribution in [0,1] under the null hypothesis. To achieve these objectives, we simulate a data set where all the gene sets have different set sizes and no gene sets are related to the phenotype. The simulation is implemented in the following steps:</p>", "<p>1. Generate 100 gene sets whose sizes are 5,6,7,8,...,104. The total number of genes is 5450;</p>", "<p>2. The gene expression levels in 100 samples for each gene are generated from a standard normal distribution. Different genes are independent of each other. Different samples are also independent of each other;</p>", "<p>3. Generate the phenotypic data from another independent standard normal distribution in 100 samples;</p>", "<p>4. Repeat steps 1–3 for 100 times;</p>", "<p>In total, the simulation generates 100 data sets that have gene expression data and a corresponding phenotypic data. We apply Sub-GSE to the 100 data sets separately.</p>", "<p>First, since the gene sets have different sizes, we plot the average p-values of all the gene sets across the 100 different data sets against their set sizes to see whether the gene set size affects the significance level. Figure ##FIG##0##1## shows that the set size does not affect the p-values.</p>", "<p>Second, the phenotypic data is independent of the expression levels of all the genes. Therefore, Sub-GSE should not detect any significant gene sets. In Figure ##FIG##1##2##, the histogram of all the p-values of the 100 gene sets from the 100 data sets is shown. The histogram illustrates that the p-values from the Sub-GSE have a uniform distribution for gene sets that are not related to the phenotype, which is consistent with the theoretical uniform distribution under the null hypothesis.</p>", "<title>Simulation 1</title>", "<p>We first evaluate the performance of Sub-GSE using simulated data in which gene expression profiles with different correlations within the gene set are generated. The expression profiles for 1000 genes in 100 samples are simulated. The genes are divided into 50 non-overlapping gene sets with 20 genes in each. The gene expression profiles for the 100 samples represent 100 independent vectors of random variables generated from a multivariate normal distribution. The multivariate normal distribution has 1000 dimensions corresponding to the 1000 genes. The mean is a vector of 1000 zeroes, and the variance of the expression levels of each gene is 1. To simulate the dependence between genes, we randomly select a certain percentage of correlated genes (PCG) = (0%, 10%, ⋯, 90%) in each gene set and let the correlation coefficient between any two of them be <italic>ρ </italic>= 0, 0.1, 0.2, ⋯, 0.9. The remaining genes are independent of each other and those that are chosen. We use this simulation strategy based on the following considerations. The chosen genes in the gene set correspond to those in the same complex or pathway; thus, their expression profiles are correlated. Also, since the remaining genes represent those not belonging to the group, they are more likely to be independent of the chosen genes and each other.</p>", "<p>If a given gene is among those that are chosen, we use its expression levels as the phenotype. The rationale of this step is to determine if our Sub-GSE method can identify the gene set to which this particular gene belongs. We repeat this process for all the chosen genes. Thus, we have a total of 1000 × <italic>PCG </italic>different phenotypic data. To avoid the problem where a gene has exactly the same expression profile as the phenotype, we eliminate the gene's expression profile from the expression data if it is used as the phenotype.</p>", "<p>We use the following approach to study the robustness of Sub-GSE. For each given correlation coefficient and PCG, we randomly choose one of the simulated phenotypic data and the corresponding gene expression data. Sub-GSE is applied to the chosen data set for 100 times. The standard deviations of the p-values across the 100 different runs are plotted against the average p-values for all the gene sets in Figure ##FIG##2##3##. The figure shows that the standard deviation of the p-value for the same gene set is smaller than 0.006 and even smaller when the p-value is close to either 1 or 0. The closer the p-value is to 0 or 1, the smaller the standard deviation is. The maximum standard deviation is achieved when the average p-value is around 0.5.</p>", "<p>For each given pair of percentage of correlated genes (<italic>PCG</italic>) and correlation coefficient, we apply all four methods, Sub-GSE, GSEA, GSA, and SigPath, to the corresponding data. All the gene sets are ranked in an increasing order of their q-values so more significant gene sets have smaller rank. The rank of the gene set, some of whose member genes are correlated with the phenotypic data, is extracted to evaluate the performance of the methods. Figure ##FIG##3##4## shows the average rank of the gene set related to the phenotype for different combinations of <italic>PCG </italic>and correlation coefficient.</p>", "<p>First, as seen in the left panel in Figure ##FIG##3##4##, for small <italic>PCG </italic>= 10%, 20% and 30%, the average rank of the gene set related to the phenotype based on Sub-GSE is always the lowest, irrespective of the coefficient value. On the other hand, for large <italic>PCG</italic>, the performance of Sub-GSE is similar or slightly worse than GSEA and GSA for small correlation. The right panel in Figure ##FIG##3##4## confirms this because the average rank of the gene set related to the phenotype based on Sub-GSE decreases much faster than those for the other methods when <italic>PCG </italic>is small. The results of Figure ##FIG##3##4## can be explained as follows. When <italic>PCG </italic>is low, only a small fraction of the genes in the target gene set are correlated with the phenotype. GSEA, GSA, and SigPath cannot distinguish the target gene set from the other gene sets since these methods consider all the genes in the gene set of interest in their statistics. In contrast, Sub-GSE incrementally tests each strict set and chooses the smallest p-value across all the strict sets as a test statistic, thus making the test more powerful.</p>", "<p>Second, across different combinations of <italic>PCG </italic>and correlation coefficient, we find that GSA and GSEA achieve similar results. Both GSA and GSEA use t-statistics to obtain the ranking list of genes. For applications in this article, the only diference between them is that GSA restandardizes the statistics before the permutation to reduce the effect of correlation between genes. However, in both panels of Figure ##FIG##3##4##, the average ranks of the target gene set by GSEA and GSA are quite similar, especially when the <italic>PCG </italic>is high. Consequently, restandardization in GSA does not seem to be very efficient in this simulation study, especially when there are many correlated genes.</p>", "<p>Third, to show the sensitivity and specificity of Sub-GSE, we need a group of gene sets that are related with the phenotype. Therefore, we do another set of simulations similar as simulation 1. The detailed descriptions of the simulation and the resulting ROC curves can be found in the supplementary materials [see Additional file ##SUPPL##0##1##]. The results show that the higher the PCG and the correlation coefficient are, the higher the AUC score is. Once the correlation coefficient is higher than 0.4, the AUC score is higher than 0.85 no matter what the PCG is. When PCG is higher than 0.5, the AUC score can be higher than 0.75 regardless of the correlation coefficient.</p>", "<title>Simulation 2</title>", "<p>In reality, most phenotypes are the joint effect of multiple genes, probably from multiple pathways. Therefore, we also simulate a more realistic case where the phenotypes are assumed to be a complex function of expression levels from two gene sets. As in simulation 1, we again consider 1000 genes divided equally into 50 non-overlapping gene sets of 20 genes in each. For fixed <italic>PCG </italic>and <italic>ρ</italic>,</p>", "<p>1. Simulate the expression profiles of the 1000 genes as in the first simulation for 100 individuals;</p>", "<p>2. Choose two gene sets <italic>K</italic>₁and <italic>K</italic>₂from the 50 gene sets. Let <italic>SK</italic>₁and <italic>SK</italic>₂be the correlated genes in <italic>K</italic>₁and <italic>K</italic>₂, respectively. Define the phenotype for the <italic>j</italic>-th individual as</p>", "<p></p>", "<p>where <italic>ϵ</italic>_{<italic>j </italic>}has a normal distribution with mean 0 and variance 0.25.</p>", "<p>3. Analyze the data using GSEA, SigPath, GSA, and Sub-GSE to rank the gene sets. Rank all the gene sets in increasing order of their q-values.</p>", "<p>4. Repeat steps 1–3 100 times to assess the performance of the different analytic methods by the effects of the different gene expression data.</p>", "<p>We study the robustness of Sub-GSE as follows. Similar to the process in simulation 1, for each given correlation coefficient and PCG, we randomly choose a phenotypic data and the corresponding gene expression data. Sub-GSE is applied to the chosen data set for 100 times. The standard deviation of the p-values across the 100 runs for all the gene sets is plotted against the average p-values in Figure ##FIG##4##5##. Again, the standard deviation of the p-values across different runs for each gene set is smaller than 0.006. The closer the average p-value is to either 0 or 1, the smaller the standard deviation is. The maximum standard deviation is achieved when the average p-value is around 0.5.</p>", "<p>For this simulation study, we again apply the four different methods to prioritize the gene sets as in the first simulation study and calculate the average rank of the two target gene sets. The results can be found in Figure ##FIG##5##6##.</p>", "<p>As shown in Figure ##FIG##5##6##, the average ranks of the target gene sets based on all the methods are relatively high. This could result from the involvement of two different gene sets when simulating the phenotypic data and the fact that the phenotypic data are the sum of the squared expression levels of correlated genes. Another potential complicating factor is that the phenotype includes a noise in addition to the function of the gene expression levels of the component genes. All these facts can weaken the correlation between the phenotypic data and the gene expression profile of individual genes inside the true gene sets. Despite these problems, Sub-GSE performs relatively well compared to the other three methods, especially when <italic>PCG </italic>is low. When <italic>PCG </italic>is high, the performances of Sub-GSE are close to those of GSEA and GSA since both GSEA and GSA consider all the genes inside the gene sets. Again, the performances of GSEA and GSA are similar.</p>", "<title>Male Vs. Female Lymphoblastoid Cells</title>", "<p>We also apply Sub-GSE to two real data sets from [##REF##16199517##14##]. The first data set measured the mRNA expression profiles from lymphoblastoid cells derived from 15 males and 17 females using Affymetrix U133A chip. The gender of the individuals represents the corresponding phenotypic data. The gene sets are chosen as the cytogenetic sets (C1, 319 gene sets) and the functional gene sets (C2, 522 gene sets) defined in [##REF##16199517##14##]. The cytogenetic sets contain 24 gene sets, one for each of the 24 human chromosomes, and 295 gene sets corresponding to cytogenetic bands along the chromosomes. The functional sets include 472 gene sets containing genes whose products are involved in specific metabolic and signaling pathways, as reported in eight publicly available, manually curated databases, and 50 gene sets containing genes co-expressed in response to genetic and chemical perturbations, as reported in various experimental studies (see supporting text in [##REF##16199517##14##] for details). We apply Sub-GSE, GSEA, and SigPath to these two types of gene sets independently with the objective of identifying the cytogenetic regions that are differentially expressed between males and females and the functional gene sets related to sex distinction, respectively.</p>", "<p>First, we apply Sub-GSE to investigate the enrichment of cytogenetic gene sets (C1). As expected, the three most significant cytogenetic bands are chrY, chrYp11 and chrYq11 which all have a q-value of 0 and are the only three cytogenetic bands from chromosome Y in gene sets C1. They are also the only three significant gene sets in C1 by GSEA (FDR &lt; 0.2) and SigPath (max q-value &lt; 0.2). Besides these expected bands on chromosome Y, other bands that are ranked as the top 7 among all the gene sets by Sub-GSE, GSEA and SigPath are listed in Table ##TAB##0##1##. As seen from the lists, Sub-GSE is sensitive enough to identify cytogenetic bands on both chromosomes X and Y at the q-value threshold of 0.20. On the contrary, neither GSEA nor SigPath is able to detect any bands on chromosome X at the FDR threshold of 0.20. Again, this result shows the sensitivity of Sub-GSE.</p>", "<p>Second, we apply Sub-GSE to investigate the enrichment of functional gene sets (C2). Both Sub-GSE and GSEA detect three significant gene sets whose significance levels are listed in Table ##TAB##1##2##: testis-related genes, genes that escape X inactivation, and female reproductive tissue-expressed genes. The q-values of all the other gene sets are larger than 0.9 by Sub-GSE. Hence, in this dataset, the results by Sub-GSE are roughly the same as those achieved by GSEA.</p>", "<title>P53 Status in Cancer Cell Lines</title>", "<p>The second real data set corresponds to the gene expression data and phenotypic data related to p53 mutation status from [##REF##16199517##14##]. The objective of this study is to identify novel targets of the transcription factor p53. The p53 mutation status gene expression data examined the gene expression patterns from the NCI-60 collection of cancer cell lines. The expression profiles were measured using Affymetrix U95Av2 chips. The mutational status of the p53 gene had been reported for 50 of the NCI-60 cell lines, with 17 being classified as normal and 33 as carrying mutations in the gene. We take the gene expression profiles of these 50 cell lines as the expression data and the vector of binary indicators of the mutational patterns (normal or mutated) as the corresponding phenotypic data. For the gene set data, we only use the functional sets (C2, 522 gene sets) in [##REF##16199517##14##], which was already described in the application noted above. Functional sets that have a q-value smaller than or equal to 0.03 by Sub-GSE are extracted and listed in Table ##TAB##2##3## together with their q-values. Comparing the list of significant gene sets by Sub-GSE and GSEA [##REF##16199517##14##], we can see that Sub-GSE obtains more significant gene sets than GSEA does, which again shows the sensitivity of Sub-GSE. The relationship between the identified gene sets by Sub-GSE (Table ##TAB##2##3##) and p53 is illustrated in Figure ##FIG##6##7##. Basically, all these gene sets are significantly enriched in genes that are differentially expressed in p53 mutants versus those without p53 mutations. Therefore, the identified gene sets by Sub-GSE are potentially those regulated by p53. According to the definitions of these gene sets, as shown in Figure ##FIG##6##7##, we can roughly divide them into three groups.</p>", "<p>The first group includes gene sets that are directly regulated or affected by p53, including the \"p53 signaling pathway\", \"p53 signaling pathway genes\", and \"p53 upregulated genes\". This group of gene sets was detected by both Sub-GSE and GSEA [##REF##16199517##14##] (FDR &lt; 0.015).</p>", "<p>The second group contains gene sets that are \"downstream\" of p53. These gene sets can either be induced or inhibited by p53 [##REF##11099028##35##, ####REF##16741928##36##, ##REF##9664074##37##, ##REF##12670900##38####12670900##38##]. For example, it is well known that p53 induces cell cycle arrest during the G1/S phase and the G2/M phase checkpoint [##REF##16741928##36##]. By itself, p53 can activate an important death receptor, Fas, which triggers the \"Fas Signaling Pathway\" and thus leads to apoptosis [##REF##12670900##38##]. It is also well known that p53 functions \"upstream\" of ceramide in response to genotoxic stress [##REF##9664074##37##].</p>", "<p>The third group includes gene sets related to the \"upstream\" biological processes or genes for p53. These \"upstream\" biological processes, such as DNA damage caused by radiation or chemical carcinogens, for example, pass the DNA damage signal down to p53 and further induce some of the \"downstream\" pathways. Two genes, TrkA and Pitx2, are known to affect apoptosis through regulation of p53 [##REF##9852160##39##,##REF##16129685##40##]. The gene sets related to these \"upstream\" biological processes actually include genes related to those \"downstream\" biological processes in the second group.</p>", "<p>In this dataset, Sub-GSE not only detects the gene sets identified by GSEA [##REF##16199517##14##], but also detects more novel gene sets related to p53. Previous studies from the literature support the findings in that all the significant gene sets identified by Sub-GSE are related to p53, as shown in Figure ##FIG##6##7##.</p>" ]

[ "<title>Conclusion</title>", "<p>To summarize, we have developed a method, called Sub-GSE, to identify gene sets that are associated with a phenotype by testing the association between the strict subsets of genes and the phenotype. In many applications, it is very likely that only a subset of genes in a gene set of interest is associated with the phenotype. However, since currently available methods for gene set enrichment analysis usually test the association of all the genes in a gene set with the phenotype, the power of these methods is correspondingly reduced. In contrast, Sub-GSE is based on the idea of set-association approach first proposed by [##REF##11731502##32##] and it incrementally tests the association of \"strict subsets\" with the phenotype. The strict subsets contain the genes having the top association strength of individual genes with the phenotype. We first study the performance of Sub-GSE and compare it with three widely used methods for gene set enrichment analysis: GSEA, GSA, and SigPath. Our simulations show that Sub-GSE outperforms GSEA, GSA, and SigPath in prioritizing gene sets associated with a phenotype when the fraction of genes associated with the phenotype is relatively small. On the other hand, these four methods all achieve similar results when the fraction of associated genes is large. When applied to two real data sets, Sub-GSE is shown to detect more biologically meaningful gene sets than GSEA. For example, Sub-GSE identified cytogenetic band Xp22 as significantly associated with gender (q-value &lt; 0.20), while neither GSEA nor SigPath identified them as significant at a FDR &lt; 0.20. Similarly, Sub-GSE identified many gene sets including, for instance, DNA damage genes, cell cycle checkpoints genes and programmed cell death genes, as significantly associated with p53 mutation status. These were not identified by GSEA. This evidence supports the high sensitivity of Sub-GSE. Most of the detected gene sets have supports from previous studies for the association between them and the p53 mutations.</p>", "<p>Usually a large number of sets will be detected as significant for most tests of gene enrichment analysis. Since Sub-GSE is more sensitive in detecting significant gene sets than other tests for gene set enrichment analysis, we expect that many more gene sets will be identified as significant. This may reflect biological reality instead of statistical artifacts. For example, cancer can affect a large number of genes and gene categories. By studying the GO relationship among the significant gene sets, the more specific significant GO categories may represent the real underlying affected function categories.</p>", "<p>The advantages of Sub-GSE over other approaches for testing gene set enrichment are most evident when only a fraction of the genes in the gene set of interest are associated with the phenotype. If we believe that most genes in a gene set of interest are associated with the phenotype, other approaches, including GSEA, GSA, and SigPath, may perform better than Sub-GSE. Under this scenario, the use of the minimum p-value across all the strict subsets as a test statistic, which is done in Sub-GSE, would result in the introduction of more noise. It is possible that the minimal p-value may be achieved for some subsets of the gene set of interest, making Sub-GSE less powerful. The results of our simulations are consistent with this observation. On the other hand, our simulations also showed that the performance of Sub-GSE is only marginally worse than the other approaches under the conditions noted above. We do not claim that Sub-GSE is always better than GSEA, GSA, or SigPath. Instead, Sub-GSE complements other approaches for gene set enrichment analysis when the fraction of associated genes is relatively small.</p>", "<p>The speed of Sub-GSE is determined by the number of gene sets and the number of genes inside each gene set. To give an example of the running time, we download the gene expression data with accession number GSE5081 from NCBI <ext-link ext-link-type=\"uri\" xlink:href=\"http://www.ncbi.nlm.nih.gov/\"/> which hybridized total RNAs from gastric biopsy specimens of patients with Helicobacter pylori positive (HP+) and Helicobacter pylori negative (HP-) antrum erosions (ER+), and the corresponding, adjacent normal mucosae (ER-). The gene expression data includes 54675 probes and 32 samples. HP+ and HP- are treated as the phenotype. Mappings between the probes and GO categories are from the R package <ext-link ext-link-type=\"uri\" xlink:href=\"http://www.r-project.org/\"/> named \"hgu133plus2\". All the probes are mapped to 8310 GO categories in total. We run Sub-GSE on this data set using a computer with Pentium 4 CPU 3.60 GHz/3.59 GHz, 1.00 GB of RAM. It took 12.7 hours when 1000 permutations are required and the strict set size threshold is 1.</p>", "<p>Usually we need to simultaneously test the association of a large number of gene sets with the phenotype. For each gene set, we can use Sub-GSE to test the association of the gene set with the phenotype to obtain a p-value. We have shown in the \"Results\" section that the p-value is uniformly distributed under the null hypothesis that no subset is associated with the phenotype. When we test for a large number of gene sets, the issue of multiple testing is of concern. To solve this problem, conventional methods such as Bonferroni correction can be used. However, Bonferroni correction is too conservative in most situations. Another currently widely used method dealing with multiple testing is to control false discovery rate (FDR) as implemented in the software package QVALUE [##UREF##7##34##]. For the QVALUE package to work well, the p-values for all the gene sets need to be weakly dependent. When the sizes of the gene sets are relatively small compared to the total number of genes, we expect that the p-values to be weakly dependent since the genes usually form modules and genes from different modules are more likely to be independent. When these assumptions are in doubt, we can use the p-values obtained from Sub-GSE to indicate the statistical significance of the gene sets.</p>", "<p>There are two options in Sub-GSE: the minimal size for the strict sets and the statistic to measure the association strength between gene expression profiles and the phenotype. We set the tuning parameter <italic>c </italic>to be the minimal size of the strict subsets on which to test. Parameter <italic>c </italic>can control the sensitivity and the specificity of Sub-GSE, thus having a significant effect on its performance. Generally, the sensitivity of Sub-GSE decreases and the specificity increases as <italic>c </italic>increases. Therefore, the choice of <italic>c </italic>should depend on the balance between sensitivity and specificity. Although we set <italic>c </italic>= 5 in this paper, which restricts the minimal size for the strict sets, we can, instead, require that the minimal size of the strict sets depend on the size of the gene set of interest. For example, one could consider the subsets of genes that cover at least 10% of the given genes inside each gene set. Since the minimal set size of the subset may be different for different gene sets, the effects of this type of restriction need to be further studied. The other Sub-GSE option involves the statistic used to measure the association strength between gene expression profiles and the phenotype. In this paper, we use t-statistics, Kruskal-Wallis statistics, and Pearson's correlation to evaluate the association strength between the gene expression profiles and discrete, categorical, and quantitative phenotypes, respectively. Other statistics can also be applied. The power of Sub-GSE to detect enriched gene sets for different types of statistics also needs to be further studied.</p>", "<p>It is well known that genes in the same pathway or complex tend to be correlated. A natural question is whether it is better first to do principal component analysis (PCA) and then apply Sub-GSE to the principal components. We implemented this idea and found the approach less powerful than the method implemented in this paper. A potential explanation is that the expression profiles of the genes in the gene sets among the cases and controls do not satisfy the normality assumption making the PCA approach less powerful. More studies are needed to see under what conditions the combination of PCA and Sub-GSE is more powerful than Sub-GSE alone.</p>" ]

[ "<p>This is an Open Access article distributed under the terms of the Creative Commons Attribution License (<ext-link ext-link-type=\"uri\" xlink:href=\"http://creativecommons.org/licenses/by/2.0\"/>), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.</p>", "<title>Background</title>", "<p>Many methods have been developed to test the enrichment of genes related to certain phenotypes or cell states in gene sets. These approaches usually combine gene expression data with functionally related gene sets as defined in databases such as GeneOntology (GO), KEGG, or BioCarta. The results based on gene set analysis are generally more biologically interpretable, accurate and robust than the results based on individual gene analysis. However, while most available methods for gene set enrichment analysis test the enrichment of the entire gene set, it is more likely that only a subset of the genes in the gene set may be related to the phenotypes of interest.</p>", "<title>Results</title>", "<p>In this paper, we develop a novel method, termed Sub-GSE, which measures the enrichment of a predefined gene set, or pathway, by testing its subsets. The application of Sub-GSE to two simulated and two real datasets shows Sub-GSE to be more sensitive than previous methods, such as GSEA, GSA, and SigPath, in detecting gene sets assiated with a phenotype of interest. This is particularly true for cases in which only a fraction of the genes in the gene set are associated with the phenotypes. Furthermore, the application of Sub-GSE to two real data sets demonstrates that it can detect more biologically meaningful gene sets than GSEA.</p>", "<title>Conclusion</title>", "<p>We developed a new method to measure the gene set enrichment. Applications to two simulated datasets and two real datasets show that this method is sensitive to the associations between gene sets and phenotype. The program Sub-GSE can be downloaded from <ext-link ext-link-type=\"uri\" xlink:href=\"http://www-rcf.usc.edu/~fsun\"/>.</p>" ]

[ "<title>Authors' contributions</title>", "<p>XY developed and implemented the algorithm and FS provided the original idea. Both XY and FS contributed to the writing of the manuscript. XY and FS read and approved the final manuscript.</p>", "<title>Supplementary Material</title>" ]

[ "<title>Acknowledgements</title>", "<p>We thank Dr. Chao Cheng for substantially useful discussion of the method and the results. This work was partly supported by National Institutes of Health (NIH)/National Science Foundation Joint Mathematical Biology Initiative grant DMS-0241102 and NIH grants P50 HG 002790.</p>" ]

[ "<fig position=\"float\" id=\"F1\"><label>Figure 1</label><caption><p><bold>Effect of set size on the p-values</bold>. The average p-values of the 100 gene sets across the 100 different data sets are plotted against their corresponding set sizes.</p></caption></fig>", "<fig position=\"float\" id=\"F2\"><label>Figure 2</label><caption><p><bold>The distribution of p-values under the null hypothesis of no association</bold>. The histogram of the p-values under the null hypothesis of no association between the gene sets and the phenotype.</p></caption></fig>", "<fig position=\"float\" id=\"F3\"><label>Figure 3</label><caption><p><bold>The robustness of Sub-GSE based on simulation 1</bold>. The standard deviation of the p-values across the 100 runs of Sub-GSE on simulation 1 is plotted against the average p-values for all the gene sets.</p></caption></fig>", "<fig position=\"float\" id=\"F4\"><label>Figure 4</label><caption><p><bold>Comparison results of the different tests based on simulation 1</bold>. The average ranks of the gene set related to the phenotype for Sub-GSE, GSEA, GSA, and SigPath for different percentages of correlated genes (<italic>PCG</italic>) and correlation coefficients within the chosen genes. The left panel compares the average ranks in 2-D plots in which each subplot corresponds to one value of <italic>PCG</italic>. For a given <italic>PCG</italic>, the average ranks from the four methods are plotted against the correlation coefficient between the correlated genes. The right panel shows the average rank versus the percentage of correlated genes and correlation coefficient in a 3-D plot.</p></caption></fig>", "<fig position=\"float\" id=\"F5\"><label>Figure 5</label><caption><p><bold>The robustness of Sub-GSE based on simulation 2</bold>. The standard deviation of the p-values across the 100 runs of Sub-GSE on simulation 2 is plotted against the average p-values for all the gene sets.</p></caption></fig>", "<fig position=\"float\" id=\"F6\"><label>Figure 6</label><caption><p><bold>Comparison results of the different tests based on simulation 2</bold>. The average ranks of the target gene sets by Sub-GSE, GSEA, GSA and SigPath for different percentages of correlated genes and correlation coefficients in Simulation II. The left panel compares the average ranks in 2-D plots in which each subplot corresponds to one value of <italic>PCG</italic>. For a given <italic>PCG</italic>, the average ranks of the target sets from the four methods are plotted against the correlation coefficient between the correlated genes. The right panel shows the average ranks of the target sets versus the <italic>PCG </italic>and correlation coefficient in a 3-D plot. The four cubes correspond to Sub-GSE, GSEA, GSA and SigPath.</p></caption></fig>", "<fig position=\"float\" id=\"F7\"><label>Figure 7</label><caption><p><bold>Relationship between the significant gene sets and p53</bold>. Relationship between significant gene sets and p53. Solid arrows describe the interactions between p53 and different biological processes or genes. Dashed arrows show the definition of those gene sets in the corresponding dashed rectangle. The solid arrows are constructed based on previous works in [##REF##11099028##35##, ####REF##16741928##36##, ##REF##9664074##37##, ##REF##12670900##38##, ##REF##9852160##39##, ##REF##16129685##40####16129685##40##].</p></caption></fig>" ]

[ "<table-wrap position=\"float\" id=\"T1\"><label>Table 1</label><caption><p>Comparison of the top 7 cytogenetic bands related to gender detected by different methods.</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td align=\"center\" colspan=\"2\">Sub-GSE</td><td align=\"center\" colspan=\"2\">GSEA</td><td align=\"center\" colspan=\"2\">sigPath</td></tr></thead><tbody><tr><td align=\"left\">set names</td><td align=\"left\">q-value</td><td align=\"left\">set names</td><td align=\"left\">FDR</td><td align=\"left\">set names</td><td align=\"left\">max q-value</td></tr><tr><td colspan=\"6\"><hr/></td></tr><tr><td align=\"left\">chrY</td><td align=\"left\">0</td><td align=\"left\">chrY</td><td align=\"left\">0</td><td align=\"left\">chrY</td><td align=\"left\">0</td></tr><tr><td align=\"left\">chrYq11</td><td align=\"left\">0</td><td align=\"left\">chrYq11</td><td align=\"left\">0.000801147</td><td align=\"left\">chrYq11</td><td align=\"left\">0</td></tr><tr><td align=\"left\">chrYp11</td><td align=\"left\">0</td><td align=\"left\">chrYp11</td><td align=\"left\">0.000811309</td><td align=\"left\">chrYp11</td><td align=\"left\">0</td></tr><tr><td align=\"left\">chrXp22</td><td align=\"left\">0.147</td><td align=\"left\">chrYp11_Xp22</td><td align=\"left\">0.25508726</td><td align=\"left\">chrYp11_Xp22</td><td align=\"left\">0.315700257</td></tr><tr><td align=\"left\">chrX</td><td align=\"left\">0.294</td><td align=\"left\">chr11p12</td><td align=\"left\">0.31999215</td><td align=\"left\">chr1q22</td><td align=\"left\">0.990782566</td></tr><tr><td align=\"left\">chrXp11</td><td align=\"left\">0.686</td><td align=\"left\">chr6q24</td><td align=\"left\">0.6108403</td><td align=\"left\">chr20p11</td><td align=\"left\">0.991916264</td></tr><tr><td align=\"left\">chr15q11</td><td align=\"left\">0.924</td><td align=\"left\">chr5p14</td><td align=\"left\">0.6131663</td><td align=\"left\">chr15q21</td><td align=\"left\">0.991916264</td></tr></tbody></table></table-wrap>", "<table-wrap position=\"float\" id=\"T2\"><label>Table 2</label><caption><p>Comparison of the significant functional gene sets related to gender by Sub-GSE and GSEA.</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td align=\"left\">Set Name</td><td align=\"left\">Sub-GSE (q-value)</td><td align=\"left\">GSEA (FDR)</td></tr></thead><tbody><tr><td align=\"left\">Testis related genes</td><td align=\"left\">&lt;0.001</td><td align=\"left\">0.012</td></tr><tr><td align=\"left\">Genes that escape X inactivation</td><td align=\"left\">0.165</td><td align=\"left\">&lt;0.001</td></tr><tr><td align=\"left\">Female reproductive tissue expressed genes</td><td align=\"left\">0.165</td><td align=\"left\">0.045</td></tr></tbody></table></table-wrap>", "<table-wrap position=\"float\" id=\"T3\"><label>Table 3</label><caption><p>Significant functional gene sets related to p53 mutational status by Sub-GSE.</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td align=\"left\">Set Name</td><td align=\"left\">q-value</td></tr></thead><tbody><tr><td align=\"left\">Hypoxia and p53 in the Cardiovascular system</td><td align=\"left\">&lt;0.001</td></tr><tr><td align=\"left\">G1 and S Phases of the Cell Cycle</td><td align=\"left\">&lt;0.001</td></tr><tr><td align=\"left\">p53 Signaling Pathway (p53 Pathway)</td><td align=\"left\">0.022</td></tr><tr><td align=\"left\">TrkA Signaling Pathway</td><td align=\"left\">0.022</td></tr><tr><td align=\"left\">P53 Upregulated Genes</td><td align=\"left\">0.022</td></tr><tr><td align=\"left\">p53 Signaling Pathway Genes(p53_signalling)</td><td align=\"left\">0.029</td></tr><tr><td align=\"left\">Cell Cycle: G2/M Checkpoint</td><td align=\"left\">0.029</td></tr><tr><td align=\"left\">G2 and M Phases of the Cell Cycle</td><td align=\"left\">0.029</td></tr><tr><td align=\"left\">Programmed Cell Death</td><td align=\"left\">0.029</td></tr><tr><td align=\"left\">DNA Damage Signaling Pathway</td><td align=\"left\">0.029</td></tr><tr><td align=\"left\">Radiation Sensitivity Genes</td><td align=\"left\">0.029</td></tr><tr><td align=\"left\">Cell Cycle Regulator Genes</td><td align=\"left\">0.029</td></tr><tr><td align=\"left\">ATM Signaling Pathway</td><td align=\"left\">0.029</td></tr><tr><td align=\"left\">Ceramide Signaling Pathway</td><td align=\"left\">0.029</td></tr><tr><td align=\"left\">Drug Resistance and Metabolism Genes</td><td align=\"left\">0.029</td></tr><tr><td align=\"left\">Stress Induction of HSP Regulation</td><td align=\"left\">0.029</td></tr><tr><td align=\"left\">Multi-step Regulation of Transcription by Pitx2</td><td align=\"left\">0.029</td></tr><tr><td align=\"left\">Fas Signaling Pathway</td><td align=\"left\">0.029</td></tr></tbody></table></table-wrap>" ]

[ "<disp-formula><mml:math xmlns:mml=\"http://www.w3.org/1998/Math/MathML\" id=\"M1\" name=\"1471-2105-9-362-i1\" overflow=\"scroll\"><mml:semantics><mml:mrow><mml:msub><mml:mi>y</mml:mi><mml:mi>j</mml:mi></mml:msub><mml:mo>=</mml:mo><mml:mn>1</mml:mn><mml:mo>+</mml:mo><mml:mstyle displaystyle=\"true\"><mml:munder><mml:mo>∑</mml:mo><mml:mrow><mml:mi>i</mml:mi><mml:mo>∈</mml:mo><mml:mi>S</mml:mi><mml:msub><mml:mi>K</mml:mi><mml:mn>1</mml:mn></mml:msub></mml:mrow></mml:munder><mml:mrow><mml:msubsup><mml:mi>e</mml:mi><mml:mrow><mml:mi>i</mml:mi><mml:mi>j</mml:mi></mml:mrow><mml:mn>2</mml:mn></mml:msubsup></mml:mrow></mml:mstyle><mml:mo>+</mml:mo><mml:mstyle displaystyle=\"true\"><mml:munder><mml:mo>∑</mml:mo><mml:mrow><mml:mi>i</mml:mi><mml:mo>∈</mml:mo><mml:mi>S</mml:mi><mml:msub><mml:mi>K</mml:mi><mml:mn>2</mml:mn></mml:msub></mml:mrow></mml:munder><mml:mrow><mml:msubsup><mml:mi>e</mml:mi><mml:mrow><mml:mi>i</mml:mi><mml:mi>j</mml:mi></mml:mrow><mml:mn>2</mml:mn></mml:msubsup><mml:mo>+</mml:mo><mml:msub><mml:mi>ε</mml:mi><mml:mi>j</mml:mi></mml:msub></mml:mrow></mml:mstyle><mml:mo>,</mml:mo></mml:mrow></mml:semantics></mml:math></disp-formula>", "<disp-formula><mml:math xmlns:mml=\"http://www.w3.org/1998/Math/MathML\" id=\"M2\" name=\"1471-2105-9-362-i2\" overflow=\"scroll\"><mml:semantics><mml:mrow><mml:msub><mml:mi>T</mml:mi><mml:mi>i</mml:mi></mml:msub><mml:mo>=</mml:mo><mml:mfrac><mml:mrow><mml:mstyle displaystyle=\"true\"><mml:msubsup><mml:mo>∑</mml:mo><mml:mrow><mml:mi>j</mml:mi><mml:mo>=</mml:mo><mml:mn>1</mml:mn></mml:mrow><mml:mi>i</mml:mi></mml:msubsup><mml:mrow><mml:msub><mml:mi>a</mml:mi><mml:mi>j</mml:mi></mml:msub></mml:mrow></mml:mstyle></mml:mrow><mml:mi>i</mml:mi></mml:mfrac><mml:mo>.</mml:mo></mml:mrow></mml:semantics></mml:math></disp-formula>", "<disp-formula><mml:math xmlns:mml=\"http://www.w3.org/1998/Math/MathML\" id=\"M3\" name=\"1471-2105-9-362-i3\" overflow=\"scroll\"><mml:semantics><mml:mrow><mml:msub><mml:mi>P</mml:mi><mml:mrow><mml:mi>m</mml:mi><mml:mi>i</mml:mi><mml:mi>n</mml:mi></mml:mrow></mml:msub><mml:mo>=</mml:mo><mml:munderover><mml:mrow><mml:mi>min</mml:mi><mml:mo>⁡</mml:mo></mml:mrow><mml:mrow><mml:mi>i</mml:mi><mml:mo>=</mml:mo><mml:mi>c</mml:mi></mml:mrow><mml:mi>s</mml:mi></mml:munderover><mml:msub><mml:mi>p</mml:mi><mml:mi>i</mml:mi></mml:msub><mml:mo>.</mml:mo></mml:mrow></mml:semantics></mml:math></disp-formula>", "<inline-formula><mml:math xmlns:mml=\"http://www.w3.org/1998/Math/MathML\" id=\"M4\" name=\"1471-2105-9-362-i4\" overflow=\"scroll\"><mml:semantics><mml:mrow><mml:msubsup><mml:mi>T</mml:mi><mml:mi>i</mml:mi><mml:mrow><mml:mo stretchy=\"false\">(</mml:mo><mml:mi>n</mml:mi><mml:mo stretchy=\"false\">)</mml:mo></mml:mrow></mml:msubsup></mml:mrow></mml:semantics></mml:math></inline-formula>", "<disp-formula><mml:math xmlns:mml=\"http://www.w3.org/1998/Math/MathML\" id=\"M5\" name=\"1471-2105-9-362-i5\" overflow=\"scroll\"><mml:semantics><mml:mrow><mml:mrow><mml:mo>(</mml:mo><mml:mrow><mml:mtable><mml:mtr><mml:mtd><mml:mrow><mml:msubsup><mml:mi>T</mml:mi><mml:mi>c</mml:mi><mml:mrow><mml:mo stretchy=\"false\">(</mml:mo><mml:mn>0</mml:mn><mml:mo stretchy=\"false\">)</mml:mo></mml:mrow></mml:msubsup></mml:mrow></mml:mtd><mml:mtd><mml:mrow><mml:msubsup><mml:mi>T</mml:mi><mml:mrow><mml:mi>c</mml:mi><mml:mo>+</mml:mo><mml:mn>1</mml:mn></mml:mrow><mml:mrow><mml:mo stretchy=\"false\">(</mml:mo><mml:mn>0</mml:mn><mml:mo stretchy=\"false\">)</mml:mo></mml:mrow></mml:msubsup></mml:mrow></mml:mtd><mml:mtd><mml:mo>⋯</mml:mo></mml:mtd><mml:mtd><mml:mrow><mml:msubsup><mml:mi>T</mml:mi><mml:mi>s</mml:mi><mml:mrow><mml:mo stretchy=\"false\">(</mml:mo><mml:mn>0</mml:mn><mml:mo stretchy=\"false\">)</mml:mo></mml:mrow></mml:msubsup></mml:mrow></mml:mtd></mml:mtr><mml:mtr><mml:mtd><mml:mrow><mml:msubsup><mml:mi>T</mml:mi><mml:mi>c</mml:mi><mml:mrow><mml:mo stretchy=\"false\">(</mml:mo><mml:mn>1</mml:mn><mml:mo stretchy=\"false\">)</mml:mo></mml:mrow></mml:msubsup></mml:mrow></mml:mtd><mml:mtd><mml:mrow><mml:msubsup><mml:mi>T</mml:mi><mml:mrow><mml:mi>c</mml:mi><mml:mo>+</mml:mo><mml:mn>1</mml:mn></mml:mrow><mml:mrow><mml:mo stretchy=\"false\">(</mml:mo><mml:mn>1</mml:mn><mml:mo stretchy=\"false\">)</mml:mo></mml:mrow></mml:msubsup></mml:mrow></mml:mtd><mml:mtd><mml:mo>⋯</mml:mo></mml:mtd><mml:mtd><mml:mrow><mml:msubsup><mml:mi>T</mml:mi><mml:mi>s</mml:mi><mml:mrow><mml:mo stretchy=\"false\">(</mml:mo><mml:mn>1</mml:mn><mml:mo stretchy=\"false\">)</mml:mo></mml:mrow></mml:msubsup></mml:mrow></mml:mtd></mml:mtr><mml:mtr><mml:mtd><mml:mo>⋯</mml:mo></mml:mtd><mml:mtd><mml:mo>⋯</mml:mo></mml:mtd><mml:mtd><mml:mo>⋯</mml:mo></mml:mtd><mml:mtd><mml:mo>⋯</mml:mo></mml:mtd></mml:mtr><mml:mtr><mml:mtd><mml:mrow><mml:msubsup><mml:mi>T</mml:mi><mml:mi>c</mml:mi><mml:mrow><mml:mo stretchy=\"false\">(</mml:mo><mml:mi>N</mml:mi><mml:mo stretchy=\"false\">)</mml:mo></mml:mrow></mml:msubsup></mml:mrow></mml:mtd><mml:mtd><mml:mrow><mml:msubsup><mml:mi>T</mml:mi><mml:mrow><mml:mi>c</mml:mi><mml:mo>+</mml:mo><mml:mn>1</mml:mn></mml:mrow><mml:mrow><mml:mo stretchy=\"false\">(</mml:mo><mml:mi>N</mml:mi><mml:mo stretchy=\"false\">)</mml:mo></mml:mrow></mml:msubsup></mml:mrow></mml:mtd><mml:mtd><mml:mo>⋯</mml:mo></mml:mtd><mml:mtd><mml:mrow><mml:msubsup><mml:mi>T</mml:mi><mml:mi>s</mml:mi><mml:mrow><mml:mo stretchy=\"false\">(</mml:mo><mml:mi>N</mml:mi><mml:mo stretchy=\"false\">)</mml:mo></mml:mrow></mml:msubsup></mml:mrow></mml:mtd></mml:mtr></mml:mtable></mml:mrow><mml:mo>)</mml:mo></mml:mrow><mml:mo>,</mml:mo></mml:mrow></mml:semantics></mml:math></disp-formula>", "<disp-formula><mml:math xmlns:mml=\"http://www.w3.org/1998/Math/MathML\" id=\"M6\" name=\"1471-2105-9-362-i6\" overflow=\"scroll\"><mml:semantics><mml:mrow><mml:msub><mml:mi>p</mml:mi><mml:mi>i</mml:mi></mml:msub><mml:mo>=</mml:mo><mml:msubsup><mml:mi>p</mml:mi><mml:mi>i</mml:mi><mml:mrow><mml:mo stretchy=\"false\">(</mml:mo><mml:mn>0</mml:mn><mml:mo stretchy=\"false\">)</mml:mo></mml:mrow></mml:msubsup><mml:mo>=</mml:mo><mml:mfrac><mml:mrow><mml:mo>#</mml:mo><mml:mo>{</mml:mo><mml:mi>n</mml:mi><mml:mo>&gt;</mml:mo><mml:mn>0</mml:mn><mml:mo>:</mml:mo><mml:msubsup><mml:mi>T</mml:mi><mml:mi>i</mml:mi><mml:mrow><mml:mo stretchy=\"false\">(</mml:mo><mml:mi>n</mml:mi><mml:mo stretchy=\"false\">)</mml:mo></mml:mrow></mml:msubsup><mml:mo>≥</mml:mo><mml:msubsup><mml:mi>T</mml:mi><mml:mi>i</mml:mi><mml:mrow><mml:mo stretchy=\"false\">(</mml:mo><mml:mn>0</mml:mn><mml:mo stretchy=\"false\">)</mml:mo></mml:mrow></mml:msubsup><mml:mo>}</mml:mo></mml:mrow><mml:mi>N</mml:mi></mml:mfrac><mml:mo>.</mml:mo></mml:mrow></mml:semantics></mml:math></disp-formula>", "<inline-formula><mml:math xmlns:mml=\"http://www.w3.org/1998/Math/MathML\" id=\"M7\" name=\"1471-2105-9-362-i7\" overflow=\"scroll\"><mml:semantics><mml:mrow><mml:msub><mml:mi>P</mml:mi><mml:mrow><mml:mi>m</mml:mi><mml:mi>i</mml:mi><mml:mi>n</mml:mi></mml:mrow></mml:msub><mml:mo>=</mml:mo><mml:msubsup><mml:mi>P</mml:mi><mml:mrow><mml:mi>m</mml:mi><mml:mi>i</mml:mi><mml:mi>n</mml:mi></mml:mrow><mml:mrow><mml:mo stretchy=\"false\">(</mml:mo><mml:mn>0</mml:mn><mml:mo stretchy=\"false\">)</mml:mo></mml:mrow></mml:msubsup><mml:mo>=</mml:mo><mml:msubsup><mml:mrow><mml:mi>min</mml:mi><mml:mo>⁡</mml:mo></mml:mrow><mml:mrow><mml:mi>i</mml:mi><mml:mo>=</mml:mo><mml:mi>c</mml:mi></mml:mrow><mml:mi>s</mml:mi></mml:msubsup><mml:msubsup><mml:mi>p</mml:mi><mml:mi>i</mml:mi><mml:mrow><mml:mo stretchy=\"false\">(</mml:mo><mml:mn>0</mml:mn><mml:mo stretchy=\"false\">)</mml:mo></mml:mrow></mml:msubsup></mml:mrow></mml:semantics></mml:math></inline-formula>", "<disp-formula><mml:math xmlns:mml=\"http://www.w3.org/1998/Math/MathML\" id=\"M8\" name=\"1471-2105-9-362-i8\" overflow=\"scroll\"><mml:semantics><mml:mrow><mml:msubsup><mml:mi>p</mml:mi><mml:mi>i</mml:mi><mml:mrow><mml:mo stretchy=\"false\">(</mml:mo><mml:mi>n</mml:mi><mml:mo stretchy=\"false\">)</mml:mo></mml:mrow></mml:msubsup><mml:mo>=</mml:mo><mml:mfrac><mml:mrow><mml:mo>#</mml:mo><mml:mo>{</mml:mo><mml:mi>l</mml:mi><mml:mo>≥</mml:mo><mml:mn>0</mml:mn><mml:mo>,</mml:mo><mml:mi>l</mml:mi><mml:mo>≠</mml:mo><mml:mi>n</mml:mi><mml:mo>:</mml:mo><mml:msubsup><mml:mi>T</mml:mi><mml:mi>i</mml:mi><mml:mrow><mml:mo stretchy=\"false\">(</mml:mo><mml:mi>l</mml:mi><mml:mo stretchy=\"false\">)</mml:mo></mml:mrow></mml:msubsup><mml:mo>≥</mml:mo><mml:msubsup><mml:mi>T</mml:mi><mml:mi>i</mml:mi><mml:mrow><mml:mo stretchy=\"false\">(</mml:mo><mml:mi>n</mml:mi><mml:mo stretchy=\"false\">)</mml:mo></mml:mrow></mml:msubsup><mml:mo>}</mml:mo></mml:mrow><mml:mi>N</mml:mi></mml:mfrac><mml:mo>.</mml:mo></mml:mrow></mml:semantics></mml:math></disp-formula>", "<inline-formula><mml:math xmlns:mml=\"http://www.w3.org/1998/Math/MathML\" id=\"M9\" name=\"1471-2105-9-362-i9\" overflow=\"scroll\"><mml:semantics><mml:mrow><mml:msubsup><mml:mi>P</mml:mi><mml:mrow><mml:mi>m</mml:mi><mml:mi>i</mml:mi><mml:mi>n</mml:mi></mml:mrow><mml:mrow><mml:mo stretchy=\"false\">(</mml:mo><mml:mi>n</mml:mi><mml:mo stretchy=\"false\">)</mml:mo></mml:mrow></mml:msubsup><mml:mo>=</mml:mo><mml:msubsup><mml:mrow><mml:mi>min</mml:mi><mml:mo>⁡</mml:mo></mml:mrow><mml:mrow><mml:mi>i</mml:mi><mml:mo>=</mml:mo><mml:mi>c</mml:mi></mml:mrow><mml:mi>s</mml:mi></mml:msubsup><mml:msubsup><mml:mi>p</mml:mi><mml:mi>i</mml:mi><mml:mrow><mml:mo stretchy=\"false\">(</mml:mo><mml:mi>n</mml:mi><mml:mo stretchy=\"false\">)</mml:mo></mml:mrow></mml:msubsup></mml:mrow></mml:semantics></mml:math></inline-formula>", "<disp-formula><mml:math xmlns:mml=\"http://www.w3.org/1998/Math/MathML\" id=\"M10\" name=\"1471-2105-9-362-i10\" overflow=\"scroll\"><mml:semantics><mml:mrow><mml:mi>p</mml:mi><mml:mo>−</mml:mo><mml:mi>v</mml:mi><mml:mi>a</mml:mi><mml:mi>l</mml:mi><mml:mi>u</mml:mi><mml:mi>e</mml:mi><mml:mo>=</mml:mo><mml:mfrac><mml:mrow><mml:mo>#</mml:mo><mml:mo>{</mml:mo><mml:mi>n</mml:mi><mml:mo>≥</mml:mo><mml:mn>1</mml:mn><mml:mo>:</mml:mo><mml:msubsup><mml:mi>P</mml:mi><mml:mrow><mml:mi>m</mml:mi><mml:mi>i</mml:mi><mml:mi>n</mml:mi></mml:mrow><mml:mrow><mml:mo stretchy=\"false\">(</mml:mo><mml:mi>n</mml:mi><mml:mo stretchy=\"false\">)</mml:mo></mml:mrow></mml:msubsup><mml:mo>≤</mml:mo><mml:msubsup><mml:mi>P</mml:mi><mml:mrow><mml:mi>m</mml:mi><mml:mi>i</mml:mi><mml:mi>n</mml:mi></mml:mrow><mml:mrow><mml:mo stretchy=\"false\">(</mml:mo><mml:mn>0</mml:mn><mml:mo stretchy=\"false\">)</mml:mo></mml:mrow></mml:msubsup><mml:mo>}</mml:mo></mml:mrow><mml:mi>N</mml:mi></mml:mfrac></mml:mrow></mml:semantics></mml:math></disp-formula>" ]

[]

[]

[]

[]

[ "<supplementary-material content-type=\"local-data\" id=\"S1\"><caption><title>Additional file 1</title><p><bold>Supplementary Materials</bold>. This file contains the descriptions of the simulations that show the ROC curves and the robustness of Sub-GSE.</p></caption></supplementary-material>" ]

[ "<table-wrap-foot><p>The top 7 gene sets by Sub-GSE, GSEA and sigPath and their corresponding p-values, FDR and max q-values. The cytogenetic bands chrYp11_Xp22 was named chrYp22 in the original gene set data which include 8 genes that are on both chrYp11 and chrXp22.</p></table-wrap-foot>", "<table-wrap-foot><p>The top 3 functional gene sets and their corresponding q-values and FDR rates by Sub-GSE and GSEA.</p></table-wrap-foot>", "<table-wrap-foot><p>Significant functional gene sets (q-value ≤ 0.03) detected in 50 of the NCI-60 cell lines. The q-values are calculated by Sub-GSE.</p></table-wrap-foot>" ]

[ "<graphic xlink:href=\"1471-2105-9-362-1\"/>", "<graphic xlink:href=\"1471-2105-9-362-2\"/>", "<graphic xlink:href=\"1471-2105-9-362-3\"/>", "<graphic xlink:href=\"1471-2105-9-362-4\"/>", "<graphic xlink:href=\"1471-2105-9-362-5\"/>", "<graphic xlink:href=\"1471-2105-9-362-6\"/>", "<graphic xlink:href=\"1471-2105-9-362-7\"/>" ]

[ "<media xlink:href=\"1471-2105-9-362-S1.pdf\" mimetype=\"application\" mime-subtype=\"pdf\"><caption><p>Click here for file</p></caption></media>" ]

{ "acronym": [], "definition": [] }

[ "<title>Background</title>", "<p>High-throughput experiments such as microarrays compare the expression levels of thousands of genes at once. Individual gene readings, when compared to a control, measure the degree to which the gene is up- or down-regulated. Microarray experiments contain significant noise, and typically only a few genes are found whose expression is significantly changed. Recently, several groups have begun to examine microarray experiments from the perspective of biologically related gene sets. There are a large number of such methods based on thresholding the initial microarray results by fold-change or p-value and then using a Fisher exact test to determine significance (e.g. [##UREF##0##1##]). We use the term <italic>gene set database query </italic>to describe the comparison of a microarray experiment against a database of sets of genes (often with associated biology). The results from a gene set database query consist of biologically meaningful groups, such as the set of all genes annotated with a GO term or a pathway, rather than individual genes. Non-parametric algorithms such as GSEA [##REF##16199517##2##] and PAGE [##REF##15941488##3##] are also available for this type of search. General tools to assist biologists with the analysis are being developed [##UREF##1##4##,##REF##18202032##5##], and in this context, quantifying the success of particular algorithms becomes even more important.</p>", "<p>In addition to queries against a database of biologically meaningful sets, we may wish to query a database consisting of other experiments. We use the term <italic>gene vector database query </italic>to describe a query against a database whose entries are themselves vectors of gene readings. Such a query may aim to find related experiments – for example, queries of signatures against the Connectivity Map corpus were able to identify compounds with similar effects [##REF##17008526##6##].</p>", "<p>The naive null model for a gene set database query (in both GSEA and PAGE) is that the genes in the set are drawn independently from the overall distribution. However, many gene sets of biological interest consist of co-regulated genes. The expression responses of these genes will typically be highly correlated. This tight correlation may cause us to reject the naive null model with high confidence, even in cases where the genes are not differentially regulated. One way to compensate for this interdependence of genes within a set is through permutation of class labels [##UREF##2##7##]. However, a disadvantage of permutation testing is that it requires a large number of replicates. (Note that for experiments involving fewer than 13 microarrays, fewer than 1,000 distinct permutations exist, thus permutation based p-values may be limited to &gt; 0.001). We show that it is more effective to calibrate p-values for each set (or vector) in the database using a large corpus of experiments. Once this calibration is performed, queries can be performed with higher accuracy than permutation tests, and with less computational cost.</p>", "<p>Statistical methods which apply to set queries, such as Fisher Exact, may not apply to vector queries. If different statistical methods are used, the p-values from gene set database queries and gene vector database queries may not be comparable. For this reason, we developed a method that can query both sets and vectors, using a common statistical framework. With this method, one can query microarray readings against a database of sets (as in GSEA), or query gene signatures against a database of vectors (as used in the Connectivity Map). In addition, one can use our method to query microarray readings against a database of previous microarray experiments (e.g. to find drugs which offset the transcriptional changes associated with a disease). The source code implementing our query tool, Geneva, is available <ext-link ext-link-type=\"uri\" xlink:href=\"http://bioinfo2.ucsd.edu\"/>.</p>", "<p>We report the results of evaluation experiments using publicly available microarray experiments from the GEO data repository. Related microarray experiments are those that differ only by (for example) severity of disease, dose of compound, or sampling of subjects. Gene set enrichment methods should identify very similar enriched sets for related experiments. We formalize this idea to identify 5 pairs of related experiments from GEO as an evaluation set, thus, extending the data sets from PAGE [##REF##15941488##3##]. We also use 28 pairs of mismatched (or unrelated) experiments as a negative set. This provides an objective framework to evaluate multiple methods as to their accuracy. The value of standard evaluation data sets is well proven – for example, see the influence of the Burset and Guigo data set in gene-finding [##REF##8786136##8##]. Using this evaluation data enables us to compare statistical measures within Geneva.</p>", "<p>We divide the generalized gene vector analysis problem into three steps. The first step is the acquisition of a <italic>reading </italic>for each gene to be used for querying, and compiling a database of gene sets and/or gene vectors. The second step is the calculation of an <italic>enrichment score </italic>for each gene set (or gene vector) in the database. The third step is the conversion of these enrichment scores into <italic>p-values</italic>, using the distribution of enrichment scores on a corpus of data. The computation of these values is described in the Methods.</p>" ]

[ "<title>Methods</title>", "<title>Gene Readings</title>", "<p>In order to perform a query, we need a single reading quantifying the degree of up- or down-regulation of each gene. The gene readings for the genes included in the microarray will be represented as a query vector of length N, whose n^thvalue represents the change in transcription of the n^thgene. The levels of transcription of each gene in the public data-sets we used were initially quantified using MAS5 [##REF##12204100##10##] or related methods. To quantify the up- or down-regulation of each gene, we employed the Cyber-T algorithm [##REF##11259426##11##]. The Cyber-T statistic itself is retained as the reading for a gene. The Cyber-T statistic has the advantage that it reflects both direction (up-versus down-regulation) and confidence. A variety of methods are available to quantify up- and down-regulation [##REF##17233564##12##], which can be incorporated similarly. In addition, we tried applying log fold change (which reflects direction) and the Cyber-T p-value (which reflects only confidence). In addition, p-values from SAM [##REF##11309499##13##] were computed and tested, and found to give similar results to Cyber-T p-values (data not shown).</p>", "<p>The Connectivity Map (CMAP) corpus consists of a total of 463 microarray experiments involving the exposure of human cell cultures to various perturbagens [##REF##17008526##6##]. As a second corpus, we obtained all GEO data-sets available for the Affymetrix HG-U133A chip (GPL96) as of February 1^st, 2007. The SOFT-format files for each data-set were parsed, and expression differences were measured using Cyber-T for each pair of sample sets which (a) contained three or more entries per set, and (b) were disjoint. To avoid over-representing particular treatments in our corpus, we selected at most three such comparisons per data-set. The resulting corpus contains a total of 285 gene vectors. For each comparison (A vs. B), we also added the reverse comparison (B vs. A) which increased the number of corpus to 570 vectors. This was done for technical and expository reasons as it made the distributions of scores symmetrical.</p>", "<p>A database of gene sets was constructed from several sources: GOA [##REF##10802651##14##], GenMAPP [##REF##11984561##15##], HumanCyc [##REF##15642094##16##], BioCarta <ext-link ext-link-type=\"uri\" xlink:href=\"http://biocarta.com/genes/allpathways.asp\"/>, and TRANSFAC [##REF##12520026##17##]. Gene identifiers from the source databases, along with Affymetrix microarrays, are mapped to a collection of common identifiers. Because small gene sets do not lead to statistically significant results, we ignored any set containing fewer than five genes. A total of 4,256 gene sets of sufficient size were used.</p>", "<title>Enrichment Scores</title>", "<p>Given a query vector of gene readings (as described above) and a gene set, we considered several statistical models for computing an enrichment score for the gene set:</p>", "<p>• Pearson correlation. We construct a binary membership vector for the set. This membership vector's n^thentry is 1 if the n^thgene is a member of the set, and 0 otherwise. We then compute the Pearson correlation between the membership vector and the query vector. The enrichment score is the Pearson correlation coefficient, <italic>r</italic>.</p>", "<p>• Spearman correlation. As with Pearson correlation, we first construct a binary membership vector for the set. We then compute a Spearman (rank-based) correlation, <italic>ρ</italic>, between the membership vector and the query vector. The enrichment score is the variable <italic>t</italic>, defined as:</p>", "<p></p>", "<p>• PAGE. We implemented the Parametric Analysis of Gene Set Enrichment (PAGE) method as described by Kim &amp; Volsky [##REF##18202032##5##]. PAGE is based on using the normal distribution for statistical inference, and is possibly more sensitive than GSEA.</p>", "<p>The two correlation-based methods have the advantage that they apply equally well to queries against a database of vectors. The accuracy of these methods was compared on an evaluation data set.</p>", "<p>In the past, researchers have compiled a set of genes of interest from a microarray experiment (e.g. two-fold or greater change in expression), then compared the set against a database of biologically related genes using Fisher's Exact Test. Geneva can be used in essentially the same way if Cyber-T is replaced by a binary-valued vector set to one for precisely those genes of interest. However, querying based upon the readings themselves is more informative than applying an arbitrary cutoff and then querying upon gene sets.</p>", "<title>Calibration of p-values</title>", "<p>For the GEO corpus and for each gene set, we fitted a normal distribution to the empirical distribution of the enrichment scores. (See Figure ##FIG##0##1## for an empirical cumulative distribution for two different gene sets.) The inferred mean (<italic>μ</italic>) and standard deviation (<italic>σ</italic>) parameters of the normal distribution were then used to compute p-values for that gene set for all queries. This was done independently for each of the three enrichment score methods: Pearson, Spearman, and PAGE, and also done for the CMAP corpus in addition to the GEO corpus.</p>", "<p>Under reasonable assumptions, the theoretical distribution of Pearson correlation scores follows a normal distribution whose variance is inversely proportional to the number of genes [##UREF##3##18##]. In practice, the distribution of Pearson correlation scores for gene sets in our database across the corpus is indeed fit well by a normal distribution, but with a standard deviation that varies between gene sets. The variance of the enrichment score distribution correlates with size (r = 0.41), but is also affected by the degree of co-regulation.</p>", "<p>We expect the distribution of the enrichment scores across the corpus to follow a normal distribution. We evaluated the quality of the fit to the normal distribution using the Kolmogorov-Smirnov statistic. (This KS test was used to test for normality, and should not be confused with the use of KS test in GSEA.) The median p-values for the GEO and CMAP corpora were 0.87 and 0.60 respectively. Thus, we could not reject the hypothesis that the p-values are normally distributed. Similarly, Spearman correlation p-values follow a normal distribution (median p-values 0.61 and 0.53), as do Z-scores (p-values 0.88 and 0.68). The standard deviations of enrichment scores for gene sets across the two corpora are tightly correlated (r = 0.87). This suggests that any sufficiently large and diverse corpus provides a reasonable measurement of the degree to which genes in a set are co-regulated.</p>", "<title>Evaluation of query algorithm</title>", "<p>We obtained several publicly-accessible microarray data-sets from the GEO repository [##REF##16939800##19##,##UREF##4##20##]. Five pairs of related experiments were used in our evaluation experiment, as follows:</p>", "<p>• Muscle: Muscle tissue from old males (67–75 years) vs. young (21–27 years) males (GDS287) and old females (65–71 years) vs. young (20–29 years) females (GDS472) [##REF##12783983##21##].</p>", "<p>• Malaria: Whole blood from children with mild malaria vs. healthy children and severe malaria vs. healthy children (GDS1971)</p>", "<p>• AD: Brain tissue from subjects with moderate AD (Alzheimer's disease) vs. normal and with severe AD vs. normal (GDS810) [##REF##14769913##9##].</p>", "<p>• Glioma: Grade III gliomas vs. control (non-tumor) cells and grade IV gliomas vs. control cells (GDS1962) [##REF##16616334##22##].</p>", "<p>• Obesity: Skeletal muscle tissue samples from obese vs. non-obese and morbidly obese vs. non-obese subjects (GDS268) [##REF##16849634##23##].</p>", "<p>The pairs of experiments described above are considered <italic>related</italic>, as they involve similar biological changes and should affect transcription in similar ways. In order to quantify the performance of our queries, we tabulate the gene sets that are considered as significant for both the related experiments. If a gene set is found to be enriched in both experiments, we have increased confidence that the gene set is indeed undergoing a biologically relevant change in expression. By contrast, we expect to see few (if any) shared gene sets between two experiments chosen from different biological conditions. In practice, some overlap was seen between some of those pairs, for instance, we saw some overlap between the effects on muscle tissues of obesity and aging. So we selected an even cleaner negative control set by picking a collection of seven <italic>unrelated </italic>experiment pairs (as close to \"biologically disjoint\" as possible), which should share few up- or down-regulated genes (Table ##TAB##1##2##).</p>", "<p>We listed the top N gene sets reported as enriched for any of the ten evaluation experiments, for N ranging from 1 to 1000. We then checked if a gene set is reported as enriched for two experiments. We count the number of such shared gene sets for related experiments (e.g. obesity in male and in female), denoting the count as <italic>V</italic>. Similarly, let <italic>I </italic>denote the number of shared gene sets for unrelated experiments; these are (to a first approximation) all invalid. These gene sets shared between unrelated experiments serve as an estimate of the number of spurious gene sets shared between related experiments. For any given N, the false discovery rate (FDR) [##UREF##5##24##] for gene sets shared between related experiments can be readily computed (for V &gt; I) as <italic>cI/V</italic>. Here the scaling factor, <italic>c </italic>is the number of related experiment pairs divided by the number of unrelated experiment pairs. We can also define Precision to be (<italic>1-FDR</italic>) = <italic>(1-cI/V</italic>). This Precision is plotted against N in figures ##FIG##2##3## &amp;##FIG##3##4##. Method A is considered better than method B for threshold N if it has higher Precision. For each method after evaluation on the above data set, we selected an FDR cutoff of 10% for our comparisons (see Table ##TAB##0##1## for Pearson hits at FDR = 10%).</p>", "<title>Queries against a gene vector database</title>", "<p>The Pearson and Spearman correlation enrichment models can be applied equally well to queries against a database of vectors. As a test of this procedure, we measured differential expression using Cyber-T for all data sets in the GEO corpus (described above), then performed an all-against-all vector query. We modeled the distribution of correlation values R for a given data-set X with a normal distribution. This enables us to compute the p-value, P_X(R), for a given value of R. When comparing vectors X and Y, a p-value for the association of X and Y is computed as the geometric mean of P_X(R) and P_Y(R). This score reflects the significance of a particular correlation R relative to the correlation values observed for X and Y across the entire corpus. In the absence of a training set of query results, we examined the query results for several GEO data-sets to determine whether they were biologically reasonable.</p>" ]

[ "<title>Results</title>", "<p>The distribution of enrichment scores for a given gene set across the corpus reflects the co-regulation (formally: the correlation in transcription changes) of the genes across various treatments. Figure ##FIG##0##1## compares these distributions for two gene sets: A set of 118 genes related to oxidative phosphorylation, and a large set of 1,222 genes related to mRNA processing. For an example of how p-value calibration provides improved query results, let us consider a data-set (GDS287) comparing muscle tissue from young and aged males. Using a naive query that performs no p-value calibration, we obtained a p-value of 1.2 × 10^-34for the mRNA processing set, much lower than the value for oxidative phosphorylation (6.4 × 10^-11). Similar results were observed using PAGE [##REF##15941488##3##]. However, after calibration against the CMAP corpus (see Methods), this ordering is reversed, and the p-value for mRNA processing is no longer significant after correcting for multiple hypothesis testing. Naive queries frequently detect the mRNA processing set as enriched – indeed; it receives an uncorrected p-value below 0.05 in the <italic>majority </italic>of the 463 CMAP experiments. This demonstrates that the naive null hypothesis does not suffice to filter out false positives based on gene sets comprised of highly co-regulated genes from gene set queries.</p>", "<p>As described in the Methods, we computed the false discovery rate for queries across pairs of related and unrelated experiments (Figure ##FIG##1##2##). If p-value calibration is performed, many more gene sets are observed at a 10% false discovery rate. The results also demonstrate that either large corpus provides a reasonable training set for p-value calibration, as queries calibrated with either GEO or CMAP perform significantly better than those without calibration. The GEO corpus has the advantage that it includes a wide array of treatments and tissue types, and that it uses t-scores (available only for the GEO corpus) rather than fold changes (available for the CMAP corpus). On the other hand, the CMAP corpus is somewhat larger, and has the advantage that it was generated by one lab with high reproducibility. The GEO corpus was arguably more effective, as measured by the slower decrease in precision. However, when we list the top 10 gene sets for these experiments (as measured by product of p-values), the lists reported using calibration against the CMAP corpus appeared to be most biologically reasonable (in our subjective opinion).</p>", "<p>Calibrating p-values using a corpus of experiments is less expensive computationally than using a permutation of class labels, particularly if many queries will be run against the same database of gene sets. The initial corpus calibration is time-consuming (requiring approximately 1 day of running time on a typical desktop PC), but need only be done once for each gene set. Perhaps surprisingly, our results show that calibrating p-values across a corpus of experiments yielded higher accuracy than generating p-values by permuting the class labels. However, we note that permutation of class labels is clearly more effective than no p-value calibration at all.</p>", "<p>In a related experiment, we compared the query precision obtained when using the Cyber-T statistic, Cyber-T p-values, or log fold change as our gene readings (Figure ##FIG##2##3##). Queries using the Cyber-T statistic or p-value are noticeably more accurate than those driven solely by log fold change. This reflects the large amount of noise in fold-change measurements for genes expressed at a low level. Not surprisingly, precision declines as the significance threshold drops (i.e. N increases).</p>", "<p>A final evaluation experiment compared the accuracy obtained using several different enrichment models (Figure ##FIG##3##4##). Pearson correlation is more accurate than Spearman correlation, as might be expected when comparing parametric and non-parametric models. However, Pearson and PAGE were almost identical. Noticeably, both the FDR q-value and FWER p-values from GSEA performed worse, possibly because we did not calibrate those p-values.</p>", "<title>Gene set results</title>", "<p>Table ##TAB##0##1## lists several of the top gene sets returned for the five pairs of experiments described in Methods. (A full table of gene sets, together with the corresponding Affymetrix probe IDs, is available in Additional file ##SUPPL##0##1##). Some gene sets of clear biological interest arise. For example, the set of genes annotated with the biological process \"Long-term memory\" was down-regulated in the Alzheimer's disease samples. Gene sets related to immune response were differentially regulated in response to malaria infection. As reported previously [##REF##15941488##3##], gene sets related to glycolysis and the TCA cycle are differentially regulated in young and aged muscle.</p>", "<p>The original study of the Alzheimer's disease samples [##REF##14769913##9##] identified several differentially expressed gene sets using a modified Fisher's exact test. Several biological processes were identified again by our study, including downregulation of ATP biosynthesis and GPCR signaling, and upregulation of apoptosis. We believe that the reliability of our results is improved by the use of a parametric statistic, as well as a more reasonable null model, which accounts for co-regulated gene sets.</p>", "<title>Vector query results</title>", "<p>As described in Methods, we computed the correlation of the Cyber-T vectors for all pairs of experiments in the GEO corpus. Given these values, we performed vector queries, to identify all experiments significantly related to a microarray experiment. Such experiments may affect the cell similarly (e.g. exposure to related compounds), or may perturb similar pathways with opposite effects (e.g. disease response versus exposure to a treatment). This query involved approximately 800,000 pairwise comparisons, and required 3 CPU days of running time on a compute-cluster. We examined the query results up to a p-value of 0.05.</p>", "<p>We examined the query results for pairs of related experiments (see Methods). As expected, the two muscle data-sets (GDS287 and GDS472) are related to each other (p-value 3.55 × 10^-6). Two other data-sets were significantly similar – a study of sarcopenia (GDS749) and a study of the effects of exercise on muscle in elderly males (GDS1340). These results show the effectiveness of exercise in offsetting age-associated muscle loss at the transcriptional level. The Alzheimer's disease experiment (GDS810) was similar to an experiment on bipolar disorder (GDS2190), suggesting these disorders might be similar.</p>", "<p>Our full GEO against GEO results are reported [see Additional file ##SUPPL##1##2##]. Relationships between compounds can be discovered by this kind of undirected data-mining. For instance, a close relationship was observed between experiments exposing a prostate cancer cell line to two different androgens: DHT (GDS2057) and methyltrienolone/R1881 (GDS536). Hits were also seen for experiments with the transcriptional changes induced by the estrogen hormone estradiol (GDS1549) and by the estrogen receptor agonist tamoxifen (GDS2367).</p>", "<p>Other hits come from experiments with related treatments – comparisons of transcription in blood versus liver (GDS1023) and kidney versus liver (GDS1663) were closely related, presumably due to transcription of liver-specific genes. Some of the confident query hits come from experiments from different labs which applied essentially the same treatments – for instance GDS1549 and GDS2367 both measure the effects of estradiol on breast cancer cell lines. In the future, programmatically examining meta-data (e.g. from MIAME) may permit highlighting the most interesting search hits by filtering out closely related experiments.</p>" ]

[ "<title>Discussion and conclusion</title>", "<p>The high-throughput gene collection database query problem can be formulated in several ways, focusing on either gene set database queries or gene vector database queries. The use of a parametric statistical framework that accepts either gene names or gene values as input is important, particularly when combining heterogeneous data-sources (e.g. microarrays and literature-curated gene lists). Queries against both sets and vectors work well with simple metrics such as Pearson correlation, provided that p-values are calibrated properly.</p>", "<p>Calibration of gene set database queries against a corpus of experiments provides much more accurate results than using a naive null model. Calibration against a training corpus is certainly not ideal for all situations. In cases where a suitable corpus is not available (e.g. if one is investigating an organism that has not yet been extensively studied), class label permutation is the only practical approach. If a set of genes is not significantly expressed in the training corpus, then the training corpus will not adequately measure their degree of correlation. Therefore, it is desirable to use a training corpus containing as wide a variety of tissues and conditions as possible.</p>", "<p>Identifying standard data sets that can be used to compare different algorithms and different metric is beneficial. Our proposed standard data set of 5 related pairs of experiments and 28 unrelated pairs of experiments is an advance over evaluations based on one or two anecdotal comparisons. However, there is much room for improvement by extending the numbers of both related and unrelated pairs and further reducing the bias caused by any single experiment. We also recognize that the figures do not provide statistical tests to determine if in fact a method is superior to another. Again this was impossible given the small size of the data set. But it also cautions us to not over interpret the findings in figures ##FIG##1##2## and ##FIG##3##4##.</p>", "<p>The emergence of large microarray repositories, such as GEO, provide researchers with the ability to search for experiments with similar (or opposite) gene changes. Such searches provide an ideal approach to find compounds which offset the gene expression changes associated with disease states. Calibration of p-values using a corpus of experiments significantly improves the accuracy of such queries by providing a reasonable null model without the need for large numbers of controls.</p>" ]

[ "<title>Discussion and conclusion</title>", "<p>The high-throughput gene collection database query problem can be formulated in several ways, focusing on either gene set database queries or gene vector database queries. The use of a parametric statistical framework that accepts either gene names or gene values as input is important, particularly when combining heterogeneous data-sources (e.g. microarrays and literature-curated gene lists). Queries against both sets and vectors work well with simple metrics such as Pearson correlation, provided that p-values are calibrated properly.</p>", "<p>Calibration of gene set database queries against a corpus of experiments provides much more accurate results than using a naive null model. Calibration against a training corpus is certainly not ideal for all situations. In cases where a suitable corpus is not available (e.g. if one is investigating an organism that has not yet been extensively studied), class label permutation is the only practical approach. If a set of genes is not significantly expressed in the training corpus, then the training corpus will not adequately measure their degree of correlation. Therefore, it is desirable to use a training corpus containing as wide a variety of tissues and conditions as possible.</p>", "<p>Identifying standard data sets that can be used to compare different algorithms and different metric is beneficial. Our proposed standard data set of 5 related pairs of experiments and 28 unrelated pairs of experiments is an advance over evaluations based on one or two anecdotal comparisons. However, there is much room for improvement by extending the numbers of both related and unrelated pairs and further reducing the bias caused by any single experiment. We also recognize that the figures do not provide statistical tests to determine if in fact a method is superior to another. Again this was impossible given the small size of the data set. But it also cautions us to not over interpret the findings in figures ##FIG##1##2## and ##FIG##3##4##.</p>", "<p>The emergence of large microarray repositories, such as GEO, provide researchers with the ability to search for experiments with similar (or opposite) gene changes. Such searches provide an ideal approach to find compounds which offset the gene expression changes associated with disease states. Calibration of p-values using a corpus of experiments significantly improves the accuracy of such queries by providing a reasonable null model without the need for large numbers of controls.</p>" ]

[ "<p>This is an Open Access article distributed under the terms of the Creative Commons Attribution License (<ext-link ext-link-type=\"uri\" xlink:href=\"http://creativecommons.org/licenses/by/2.0\"/>), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.</p>", "<title>Background</title>", "<p>Microarray experiments measure changes in the expression of thousands of genes. The resulting lists of genes with changes in expression are then searched for biologically related sets using several divergent methods such as the Fisher Exact Test (as used in multiple GO enrichment tools), Parametric Analysis of Gene Expression (PAGE), Gene Set Enrichment Analysis (GSEA), and the connectivity map.</p>", "<title>Results</title>", "<p>We describe an analytical method (Geneva: Gene Vector Analysis) to relate genes to biological properties and to other similar experiments in a uniform way. This new method works on both gene sets and on gene lists/vectors as input queries, and can effectively query databases consisting of sets of biologically related sets, or of results from other microarray experiments. We also present an improvement to the null model estimate by using the empirical background distribution drawn from previous experiments. We validated Geneva by rediscovering a number of previous findings, and by finding significant relationships within microarrays in the GEO repository.</p>", "<title>Conclusion</title>", "<p>Provided a reasonable corpus of previous experiments is available, this method is more accurate than the class label permutation model, especially for data sets with limited number of replicates. Geneva is, moreover, computationally faster because the background distributions can be precomputed. We also provide a standard evaluation data set based on 5 pairs of related experiments that should share similar functional relationships and 28 pairs of unrelated experiments from GEO. Discovering relationships amongst GEO data sets has implications for drug repositioning, and understanding relationships between diseases and drugs.</p>" ]

[ "<title>Authors' contributions</title>", "<p>SWT implemented the algorithm and built the test data-set. PA formulated the problem and directed the comparison of methods. Both authors prepared the final manuscript.</p>", "<title>Supplementary Material</title>" ]

[ "<title>Acknowledgements</title>", "<p>The authors gratefully acknowledge the assistance and ideas of Liwen Liu and William Reisdorf. ST was supported by NSF IGERT training grant DGE0504645.</p>" ]

[ "<fig position=\"float\" id=\"F1\"><label>Figure 1</label><caption><p><bold>Cumulative distribution functions of Pearson enrichment scores for two gene sets across the CMAP corpus.</bold> There are clear differences in the variance of these two distributions of two gene sets. However, the empirical distribution of scores across the corpus fits a normal distribution well for most gene sets. For each gene set, we calculated a p-value based on the specific normal distribution associated with that gene set.</p></caption></fig>", "<fig position=\"float\" id=\"F2\"><label>Figure 2</label><caption><p><bold>Precision (1-FDR) of gene set queries methods with Pearson based p-values calibrated on GEO and CMAP compared to permutation based p-values and no permutation.</bold> Also included, for comparison are GSEA q-value (based on FDR) and GSEA p-value (based on FWER). We compared Precision across the various queries for threshold of gene sets (N) plotted on the x-axis (as described in Methods).</p></caption></fig>", "<fig position=\"float\" id=\"F3\"><label>Figure 3</label><caption><p><bold>Comparison of query accuracy, on the evaluation set, with p-values calibrated against the GEO corpus using Pearson correlation.</bold> Queries based upon signed p-value were more effective than just p-value. Cyber-T was also extremely effective especially for N &gt; 60. Using log fold changes as gene values was least effective consistently, perhaps due to the noise in the log fold change for genes with low expression.</p></caption></fig>", "<fig position=\"float\" id=\"F4\"><label>Figure 4</label><caption><p><bold>Comparison of query accuracy, for the evaluation set, using various enrichment models.</bold> This is based on the GEO corpus using the Cyber-T. PAGE produces the best results. Pearson is a very close second; an additional advantage of Pearson correlation is that it is effective for queries against vectors and gene sets.</p></caption></fig>" ]

[ "<table-wrap position=\"float\" id=\"T1\"><label>Table 1</label><caption><p>Top-scoring differentially expressed gene sets found for pairs of related microarray experiments (from the category in column 1 above) using Geneva.</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td align=\"left\"><bold>Experiment</bold></td><td align=\"left\"><bold>Rank</bold></td><td align=\"left\"><bold>p-value</bold></td><td align=\"left\"><bold>Name</bold></td><td align=\"left\"><bold>Source</bold></td></tr></thead><tbody><tr><td align=\"left\">Muscle</td><td align=\"left\">1</td><td align=\"left\">7.89E-10</td><td align=\"left\">Glycolysis_and_Gluconeogenesis</td><td align=\"left\">GenMAPP</td></tr><tr><td align=\"left\">Muscle</td><td align=\"left\">2</td><td align=\"left\">6.93E-09</td><td align=\"left\">Costamere: CC</td><td align=\"left\">GOA</td></tr><tr><td align=\"left\">Muscle</td><td align=\"left\">3</td><td align=\"left\">4.37E-07</td><td align=\"left\">superpathway of glycolysis, pyruvate dehydrogenase, TCA, and glyoxylate bypass</td><td align=\"left\">HumanCyc</td></tr><tr><td align=\"left\">Muscle</td><td align=\"left\">4</td><td align=\"left\">4.86E-07</td><td align=\"left\">Contractile Fiber Part: CC</td><td align=\"left\">GOA</td></tr><tr><td align=\"left\">Muscle</td><td align=\"left\">5</td><td align=\"left\">6.54E-07</td><td align=\"left\">Z Disc: CC</td><td align=\"left\">GOA</td></tr><tr><td align=\"left\">Muscle</td><td align=\"left\">6</td><td align=\"left\">9.20E-07</td><td align=\"left\">Small Leucine-Rich Proteoglycan (SLRP) Molecules</td><td align=\"left\">BioCarta</td></tr><tr><td align=\"left\">Muscle</td><td align=\"left\">7</td><td align=\"left\">1.69E-06</td><td align=\"left\">aspartate degradation II</td><td align=\"left\">HumanCyc</td></tr><tr><td align=\"left\">Muscle</td><td align=\"left\">8</td><td align=\"left\">4.80E-06</td><td align=\"left\">Myofibril: CC</td><td align=\"left\">GOA</td></tr><tr><td align=\"left\">Muscle</td><td align=\"left\">9</td><td align=\"left\">4.87E-06</td><td align=\"left\">gluconeogenesis</td><td align=\"left\">HumanCyc</td></tr><tr><td align=\"left\">Muscle</td><td align=\"left\">10</td><td align=\"left\">5.38E-06</td><td align=\"left\">Contractile Fiber: CC</td><td align=\"left\">GOA</td></tr><tr><td align=\"left\">Malaria</td><td align=\"left\">1</td><td align=\"left\">1.60E-08</td><td align=\"left\">Immune Response-Regulating Signal Transduction: BP</td><td align=\"left\">GOA</td></tr><tr><td align=\"left\">Malaria</td><td align=\"left\">2</td><td align=\"left\">1.60E-08</td><td align=\"left\">Immune Response-Regulating Cell Surface Receptor Signaling Pathway: BP</td><td align=\"left\">GOA</td></tr><tr><td align=\"left\">Malaria</td><td align=\"left\">3</td><td align=\"left\">1.60E-08</td><td align=\"left\">Immune Response-Activating Signal Transduction: BP</td><td align=\"left\">GOA</td></tr><tr><td align=\"left\">Malaria</td><td align=\"left\">4</td><td align=\"left\">1.60E-08</td><td align=\"left\">Immune Response-Activating Cell Surface Receptor Signaling Pathway: BP</td><td align=\"left\">GOA</td></tr><tr><td align=\"left\">Malaria</td><td align=\"left\">5</td><td align=\"left\">1.60E-08</td><td align=\"left\">Antigen Receptor-Mediated Signaling Pathway: BP</td><td align=\"left\">GOA</td></tr><tr><td align=\"left\">Malaria</td><td align=\"left\">6</td><td align=\"left\">1.67E-08</td><td align=\"left\">T Cell Receptor Signaling Pathway: BP</td><td align=\"left\">GOA</td></tr><tr><td align=\"left\">Malaria</td><td align=\"left\">7</td><td align=\"left\">1.76E-08</td><td align=\"left\">Regulation Of T Cell Receptor Signaling Pathway: BP</td><td align=\"left\">GOA</td></tr><tr><td align=\"left\">Malaria</td><td align=\"left\">8</td><td align=\"left\">2.69E-08</td><td align=\"left\">Regulation Of Antigen Receptor-Mediated Signaling Pathway: BP</td><td align=\"left\">GOA</td></tr><tr><td align=\"left\">Malaria</td><td align=\"left\">9</td><td align=\"left\">2.64E-07</td><td align=\"left\">Activation Of Csk By cAMP-Dependent Protein Kinase Inhibits Signaling Through The T Cell Receptor</td><td align=\"left\">BioCarta</td></tr><tr><td align=\"left\">Malaria</td><td align=\"left\">10</td><td align=\"left\">5.02E-07</td><td align=\"left\">Locomotion: BP</td><td align=\"left\">GOA</td></tr><tr><td align=\"left\">AD</td><td align=\"left\">1</td><td align=\"left\">1.13E-11</td><td align=\"left\">Proton-Transporting Two-Sector ATPase Complex: CC</td><td align=\"left\">GOA</td></tr><tr><td align=\"left\">AD</td><td align=\"left\">2</td><td align=\"left\">1.13E-11</td><td align=\"left\">Hydrogen-Translocating V-Type ATPase Complex: CC</td><td align=\"left\">GOA</td></tr><tr><td align=\"left\">AD</td><td align=\"left\">3</td><td align=\"left\">9.31E-11</td><td align=\"left\">Long-Term Memory: BP</td><td align=\"left\">GOA</td></tr><tr><td align=\"left\">AD</td><td align=\"left\">4</td><td align=\"left\">4.91E-10</td><td align=\"left\">aspartate degradation II</td><td align=\"left\">HumanCyc</td></tr><tr><td align=\"left\">AD</td><td align=\"left\">5</td><td align=\"left\">1.78E-09</td><td align=\"left\">Proton-Transporting ATP Synthase Complex: CC</td><td align=\"left\">GOA</td></tr><tr><td align=\"left\">AD</td><td align=\"left\">6</td><td align=\"left\">1.78E-09</td><td align=\"left\">Proton-Transporting ATP Synthase Complex (sensu Eukaryota): CC</td><td align=\"left\">GOA</td></tr><tr><td align=\"left\">AD</td><td align=\"left\">7</td><td align=\"left\">1.78E-09</td><td align=\"left\">Hydrogen-Translocating F-Type ATPase Complex: CC</td><td align=\"left\">GOA</td></tr><tr><td align=\"left\">AD</td><td align=\"left\">8</td><td align=\"left\">1.95E-09</td><td align=\"left\">Hydrogen Ion Transporter Activity: MF</td><td align=\"left\">GOA</td></tr><tr><td align=\"left\">AD</td><td align=\"left\">9</td><td align=\"left\">6.44E-09</td><td align=\"left\">Monovalent Inorganic Cation Transporter Activity: MF</td><td align=\"left\">GOA</td></tr><tr><td align=\"left\">AD</td><td align=\"left\">10</td><td align=\"left\">7.75E-09</td><td align=\"left\">Ubiquinol-Cytochrome-C Reductase Activity: MF</td><td align=\"left\">GOA</td></tr></tbody></table></table-wrap>", "<table-wrap position=\"float\" id=\"T2\"><label>Table 2</label><caption><p>Treatments considered <italic>unrelated </italic>for the purpose of evaluation experiments.</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td align=\"left\"><bold>Pair A</bold></td><td align=\"left\"><bold>Pair B</bold></td></tr></thead><tbody><tr><td align=\"left\">Muscle</td><td align=\"left\">Malaria</td></tr><tr><td align=\"left\">Muscle</td><td align=\"left\">Glioma</td></tr><tr><td align=\"left\">Malaria</td><td align=\"left\">Glioma</td></tr><tr><td align=\"left\">Malaria</td><td align=\"left\">Obesity</td></tr><tr><td align=\"left\">AD</td><td align=\"left\">Malaria</td></tr><tr><td align=\"left\">AD</td><td align=\"left\">Obesity</td></tr><tr><td align=\"left\">Glioma</td><td align=\"left\">Obesity</td></tr></tbody></table></table-wrap>" ]

[ "<disp-formula><mml:math xmlns:mml=\"http://www.w3.org/1998/Math/MathML\" id=\"M1\" name=\"1471-2105-9-348-i1\" overflow=\"scroll\"><mml:semantics><mml:mrow><mml:mi>t</mml:mi><mml:mo>=</mml:mo><mml:mfrac><mml:mrow><mml:mi>ρ</mml:mi><mml:msqrt><mml:mrow><mml:mi>N</mml:mi><mml:mo>−</mml:mo><mml:mn>2</mml:mn></mml:mrow></mml:msqrt></mml:mrow><mml:mrow><mml:msqrt><mml:mrow><mml:mn>1</mml:mn><mml:mo>−</mml:mo><mml:msup><mml:mi>ρ</mml:mi><mml:mn>2</mml:mn></mml:msup></mml:mrow></mml:msqrt></mml:mrow></mml:mfrac></mml:mrow></mml:semantics></mml:math></disp-formula>" ]

[]

[]

[]

[]

[ "<supplementary-material content-type=\"local-data\" id=\"S1\"><caption><title>Additional file 1</title><p>Enriched gene sets.</p></caption></supplementary-material>", "<supplementary-material content-type=\"local-data\" id=\"S2\"><caption><title>Additional file 2</title><p>GEO vs. GEO query.</p></caption></supplementary-material>" ]

[ "<table-wrap-foot><p>Pearson correlation was used, and was calibrated against a corpus of experiments from GEO (see Methods).</p><p>The p-value reported is the product of the p-values for the two related experiments.</p></table-wrap-foot>", "<table-wrap-foot><p>Each treatment has two experiments, for a total of 28 unrelated experiment pairs.</p></table-wrap-foot>" ]

[ "<graphic xlink:href=\"1471-2105-9-348-1\"/>", "<graphic xlink:href=\"1471-2105-9-348-2\"/>", "<graphic xlink:href=\"1471-2105-9-348-3\"/>", "<graphic xlink:href=\"1471-2105-9-348-4\"/>" ]

[ "<media xlink:href=\"1471-2105-9-348-S1.txt\" mimetype=\"text\" mime-subtype=\"plain\"><caption><p>Click here for file</p></caption></media>", "<media xlink:href=\"1471-2105-9-348-S2.zip\" mimetype=\"application\" mime-subtype=\"x-zip-compressed\"><caption><p>Click here for file</p></caption></media>" ]

{ "acronym": [], "definition": [] }

[]

[]

[]

[]

[]

[ "<p>This is an Open Access article distributed under the terms of the Creative Commons Attribution License (<ext-link ext-link-type=\"uri\" xlink:href=\"http://creativecommons.org/licenses/by/2.0\"/>), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.</p>", "<p>Identifying the complete repertoire of genes and genetic variants that regulate the pathogenesis and progression of human disease is a central goal of post-genomic biomedical research. In cancer, recent studies have shown that genome-wide association studies can be successfully used to identify germline polymorphisms associated with an individual's susceptibility to malignancy. In parallel to these reports, substantial work has also shown that patterns of somatic alterations in human tumors can be successfully employed to predict disease prognosis and treatment response. A paper by Van Ness et al. published this month in <italic>BMC Medicine </italic>reports the initial results of a multi-institutional consortium for multiple myeloma designed to evaluate the role of germline polymorphisms in influencing multiple myeloma clinical outcome. Applying a custom-designed single nucleotide polymorphism microarray to two separate patient cohorts, the investigators successfully identified specific combinations of germline polymorphisms significantly associated with early clinical relapse. These results raise the exciting possibility that besides somatically acquired alterations, germline genetic background may also exert an important influence on cancer patient prognosis and outcome. Future 'personalized medicine' strategies for cancer may thus require incorporating genomic information from both tumor cells and the non-malignant patient genome.</p>" ]

[ "<title>Commentary</title>", "<title>Germline variations and human health</title>", "<p>A major advancement of genetic research in recent years has been the explosion of genome-wide association studies (GWAS) in the literature from different investigators and laboratories [##REF##18398418##1##]. The completion of the reference human genome sequence, and its subsequent comparison across different human sub-populations, has identified millions of genetic polymorphisms that differ between different individuals, families, and ethnic groups [##REF##16255080##2##]. With the availability of increasingly affordable chip technologies for interrogating these polymorphisms <italic>en masse </italic>in individual genomes, it is now possible to consider identifying, on a comprehensive genome-wide scale, all genes and genetic variants associated with human disease. In the area of cancer, GWAS studies have been performed for multiple different tumor types including breast, lung, and stomach cancers [##REF##17529967##3##, ####REF##18385738##4##, ##REF##18488030##5####18488030##5##]. These studies have both reconfirmed previously known disease genes (eg <italic>FGFR2 </italic>in breast cancer) [##REF##17529967##3##], and also identified novel genetic loci, such as <italic>TNRC9</italic>, <italic>MAP3K1</italic>, and <italic>LSP1 </italic>for breast cancer [##REF##17529967##3##] and the nicotinic acetylcholine receptor subunits in lung cancer [##REF##18385738##4##]. To date, the majority of reported GWAS studies have employed a case-control design, where affected individuals with a disease are compared against a matched population of non-affected normal controls. The genetic variants identified using such case-control designs thus represent 'disease susceptibility' loci that can either increase or decrease an individual's risk to developing disease. A report published this month in <italic>BMC Medicine </italic>by Van Ness et al. [##REF##18778477##6##] seeks to extend this theme, by asking whether germline polymorphisms can influence not simply the onset of disease, but the actual course of disease prognosis in cancer.</p>", "<title>Somatic alterations as dominant drivers of cancer progression</title>", "<p>The focus of Van Ness et al. on the cancer patient germline is particularly notable when one considers how the concept of cancer as an acquired somatic disease has dominated the field. In this model, tumor cells are believed to arise as a consequence of accumulated genetic lesions, which cause the pathologic activation of oncogenes and inactivation of key tumor suppressor pathways [##REF##10647931##7##]. Furthermore, multiple studies have already described various somatically-derived genomic 'signatures' in tumors that can predict both disease outcome and response to therapy, such as a 70 gene expression signature in breast tumors that can identify patients with particularly good prognoses [##REF##12490681##8##], and <italic>EGFR </italic>mutations in lung cancers that can predict tumor response to <italic>EGFR</italic>-targeted therapies [##REF##15118073##9##,##REF##15118125##10##]. In contrast to the altered cancer genome, studies analyzing the germline analysis of cancer patients, for the most part, have been largely confined to the pharmacodynamic/pharmacokinetic (PK/PD) arena, where patients are genotyped for polymorphisms in various drug-metabolizing genes to identify individuals at greatest risk of incurring severe drug toxicities (eg <italic>UGT1A1 </italic>in irinotecan treatment) [##REF##15007088##11##].</p>", "<p>More recently, however, emerging evidence suggests that in addition to somatic alterations, germline variations may also play an important role in influencing cancer prognosis and disease outcome. This might occur if particular germline variants increase the risk of developing a particular cancer subtype intrinsically associated with poor prognosis. For example, in stomach cancer, polymorphisms in the <italic>PSCA </italic>gene have been shown to be associated with the development of diffuse-type gastric adenocarcinoma, a histologic variant traditionally associated with poor clinical outcome [##REF##18488030##5##]. The influence of host genetic background on the development of cancers with differing metastatic traits has also been observed in mouse models of cancer [##REF##12721549##12##]. Germline polymorphisms could also influence cancer prognosis by affecting the regulatory circuitry of cancer cells, by altering promoter-binding sites for important cancer-related genes such as <italic>mdm2</italic>, a negative regulator of <italic>p53 </italic>[##REF##16983111##13##]. Although such studies are still relatively few in number, they do suggest that it may be time to initiate more systematic efforts to understand the specific role of germline genetic background in determining the course of cancer progression.</p>", "<title>Germline variants may affect outcome in multiple myeloma</title>", "<p>The report by Van Ness et al. provides promising initial data that this idea may indeed have scientific merit. This group has focused on multiple myeloma (MM), a hematopoetic malignancy of plasma B-cells. Although considered a uniformly fatal disease, individual MM patients are known to exhibit significant clinical heterogeneity in terms of disease morphology, time to progression and response to treatment [##REF##11841419##14##]. To ask whether germline polymorphisms might underlie some aspect of this clinical heterogeneity, the investigators designed a customized single nucleotide polymorphism (SNP) array to measure genetic polymorphisms across ~1000 genes in biological pathways relevant to MM or MM therapy, including immunity and inflammatory pathways, and genes related to drug metabolism and transport. Although not a genome-wide approach, the use of a targeted array is not without its advantages. First, by making use of prior literature knowledge, the investigators were able to incorporate many SNPs in pathways and genes relevant for MM not represented on standard genome-wide SNP arrays. Second, the use of a smaller SNP set (3500 SNPs) allowed the study to be performed with relatively smaller numbers of patient samples while still preserving statistical power, compared with a typical GWAS study. Third, because the choice of treatment regimen is an important contributor to clinical outcome, the use of smaller numbers of patient samples also facilitates standardization of treatment therapies across independent patient cohorts.</p>", "<p>The investigators applied their customized array to two separate cohorts of MM patients treated with comparable chemotherapeutic regimens. The recruitment of these patients was orchestrated through 'Bank on a Cure' (BOAC), a centralized collection agency for MM patient material from different corporative groups and institutional trials, established by expert researchers and clinicians in the MM field [##UREF##0##15##]. Using a series of computational training algorithms, the investigators showed that they could classify the patients on the basis of the germline SNP profiles into two distinct groups of 'good prognosis' (&gt;3 year progression-free survival, PFS) vs 'bad prognosis' (&lt;1 year PFS) groups above random chance. Although this initial result will undoubtedly require further validation to assess its ultimate accuracy, several intriguing trends have already emerged from their data. Among these, the authors found that accurate classification was highly dependent on using a multiplex panel of SNPs rather than any single SNP in isolation, strongly suggesting that the factors driving disease outcome in MM are likely to be complex and multifactorial. Another interesting finding was that patient classification accuracy also increased when the analysis was restricted to non-synomymous SNPs, ie those SNPs causing amino acid differences in proteins. This may imply that stronger effects on clinical outcome are likely to be modulated through alterations in protein function, rather than by alterations in the regulatory pathways controlling the transcription of these genes.</p>", "<title>Challenges for the future</title>", "<p>Although obviously exploratory in nature, the promising results of this initial study have paved the way for more ambitious and rigorous experimental designs and projects. Some potential issues with the current study include the somewhat arbitrary threshold for defining the prognosis categories, which focuses on extreme cases (&lt;1 year vs &gt;3 years). It would have been interesting, for example, to ask if similar SNP associations were also observed if the analysis was repeated treating patient survival as a continuous variable. It could also be argued that the lack of using an unbiased genome-wide approach prohibited the investigators from discovering potentially novel gene-disease associations, which might have shed further light on the regulatory pathways underlying MM. Ultimately, because this approach is relatively new, the study data will have to be further scrutinized in several other independent cohorts to assess the true prognostic power of these SNPs.</p>", "<p>Finally, it will be fascinating to determine whether the stratification power provided by these germline SNPs reflects an enhanced propensity of certain patients to develop a particular poor or good prognosis 'molecular subtype' of MM, since there is already a significant body of work describing various somatic alterations in MM that are predictive of clinical outcome [##REF##16616336##16##,##REF##16728703##17##]. Alternatively, it is also possible that these survival-associated SNPs may provide further stratification power beyond that observed by studying the somatic MM genome alone. In conclusion, this study by Van Ness is quite exciting and likely represents a model for similar studies in other cancers. It recognizes that the factors determining disease outcome in cancer are complex and multifactorial, ranging from the propensity of a cancer to proliferate and metastasize to how that cancer might respond to different types of therapy. One possibility for the future might be to test how such germline information can best be further integrated with somatic genomics to derive a holistic model for predicting outcome (Figure ##FIG##0##1##). Such information might directly translate to the development of new pharmaceutics, and diagnostic panels for personalized and predictive medicine.</p>", "<title>Pre-publication history</title>", "<p>The pre-publication history for this paper can be accessed here:</p>", "<p><ext-link ext-link-type=\"uri\" xlink:href=\"http://www.biomedcentral.com/1741-7015/6/27/prepub\"/></p>" ]

[]

[ "<fig position=\"float\" id=\"F1\"><label>Figure 1</label><caption><p><bold>Determinants of clinical outcome in cancer</bold>. The schematic outlines the interaction between germline genetic background and somatic alterations in influencing cancer outcome. Germline variations may result in differences between individuals in drug metabolizing activities, cancer pathways, and development of distinct molecular subtypes of cancer (top boxes). Alternatively, somatic alterations can cause differences in histopathology, gene expression, and gene amplifications and deletions (bottom boxes). Overlaid upon this germline/somatic interaction is the specific choice of treatment regimen. All these factors interplay to ultimately determine patient outcome (Kaplan-Meier survival curve on right). Pictures of human figures were adapted from <ext-link ext-link-type=\"uri\" xlink:href=\"http://commons.wikimedia.org/wiki/Image:Wikicouple.svg\"/>. Pictures of FISH images and Kaplan-Meier survival curves were generated in the author's laboratory and previously used in [##REF##18223220##18##].</p></caption></fig>" ]

[]

[]

[]

[]

[]

[]

[]

[]

[ "<graphic xlink:href=\"1741-7015-6-27-1\"/>" ]

[]

{ "acronym": [], "definition": [] }

[ "<title>Background</title>", "<p>Prostate specific antigen (PSA) is accepted as screening test for prostate cancer (PCa) detection but has its limitations especially in test specificity [##REF##10411025##1##]. To improve PSA specificity, many methods have been introduced, e.g. measurements of molecular forms of PSA like free PSA [##REF##1716536##2##,##REF##1703033##3##], PSA in relation to prostate volume (PSA density, PSAD) [##REF##1371554##4##], age related reference PSA values [##REF##7688054##5##] or PSA changes over time which is known as PSA velocity (PSAV) [##REF##7505974##6##]. To date only the use of percent free PSA (%fPSA) has been clinically accepted to improve specificity [##REF##9605898##7##].</p>", "<p>Recently the clinical usefulness of PSAV has debated intensively. Some authors argue for a lower cutoff for PSAV of 0.4 ng/mL per year instead of the former 0.75 ng/mL per year cutoff especially in younger men [##REF##17296371##8##]. Others introduced age adjusted PSAV and PSA cutoffs for biopsy indication [##REF##17222618##9##]. For younger patients (age 50–59) the PSAV cutoff should be lowered to 0.4 ng/mL per year to improve specificity [##REF##17222618##9##]. Ito et al. [##REF##12110095##10##] described a yearly threshold of 0.3 ng/mL as the optimal cutoff value of PSAV if the initial PSA level is 1–1.9 ng/mL and 0.75 ng/mL if the initial PSA is 2–4 ng/mL. Berger et al. [##REF##17270635##11##] found significant differences in PSAV between PCa and patients with no evidence of malignancy.</p>", "<p>In contrast, two recent studies did not prove the additional value of PSAV over PSA alone [##REF##16442212##12##,##REF##16622122##13##]. When considering the large biological variability of PSA of up to 20% [##REF##15961552##14##] or differences in PSA values regarding the used assay [##REF##10971300##15##] this may lead to misinterpretation. Different values are even more obvious when considering %fPSA [##REF##11678748##16##, ####REF##16762996##17##, ##REF##16391327##18####16391327##18##].</p>", "<p>An improved PCa detection rate was shown when using multivariate models like logistic regression [##REF##10388474##19##] or artificial neural networks [##REF##10962306##20##] which include %fPSA, PSA, and partially patient age, prostate volume and other clinical factors as input variables. However, until now the parameter PSAV has not been included in such a multivariate model.</p>", "<p>The aim of this study was to combine both methods, the ANN and PSAV, and to validate the diagnostic usefulness of this new model with regard to the differentiation between PCa and benign prostate hyperplasia (BPH). For that purpose, we compared the diagnostic usefulness of the conventional PSAV and other parameters with the so-called ANN velocity (ANNV) that included the PSA and free PSA velocity data into an ANN model.</p>" ]

[ "<title>Methods</title>", "<p>From a cohort of 2959 patients visiting the Department of Urology (Charité Hospital Berlin) with total PSA (tPSA) and free PSA (fPSA) measurements from 1996–2006, a total of 199 patients were included. The selection criteria for this PSAV and ANNV study were at least three PSA and fPSA measurements with a minimum of three months interval between two measurements before treatment. All serum samples were drawn before any prostate manipulation (or at least 3–4 weeks after an earlier manipulation) and centrifuged within 2–3 hours after sampling. The samples were analyzed immediately or stored at -20°C for no longer than 48 hours before assay. The study was carried out in accordance with the standards of the local ethics board and the Helsinki Declaration of 1975 as revised in 1996.</p>", "<p>All 199 patients (44–85 years) had a histological proven diagnosis of PCa (n = 49) or BPH (n = 150) based on examination of tissue samples obtained by transrectal ultrasound (TRUS)-guided sextant (until 1999) or octant prostate biopsies. Additionally, the status of digital rectal examination (DRE), age, and prostate volume (measured by TRUS) were also available.</p>", "<p>Total and free PSA were measured with the IMMULITE PSA and Free PSA kits (Diagnostic Products, Los Angeles, CA, USA). The analytical performance and comparisons to other PSA tests have been described earlier [##REF##16391327##18##,##REF##8674185##21##]. Prostate volume was determined by TRUS using the prolate ellipse formula. A DRE finding non-suspicious for cancer was defined as negative and a finding suspicious for cancer as positive. All patients had a complete data set on tPSA, %fPSA, age, prostate volume, and DRE status at the time of the last PSA and fPSA measurement. In 14 of the 150 BPH patients (9.3%) the ANNV was not calculated using the first but the first available complete data set (fPSA and tPSA at 2^ndor 3^rdmeasurement) since partially the fPSA was not measured when tPSA was less than 2 ng/mL (1996–1999) or less than 1 ng/mL (1999–2006).</p>", "<p>PSA values were included in the velocity calculation using the formula: (last PSA – first PSA)/time interval in ng/mL/year). Based on this formula, a one year short-term-PSAV, which describes the PSAV within the last 12 months before diagnosis, was also calculated.</p>", "<p>The ANNV was calculated analogous with the ANN output values instead of the PSA values by using the same formula (last ANN output – first ANN output/time interval). The ANN was constructed with the SPSS-module Neural connection 2.0 (SPSS) as described earlier [##REF##12142385##22##]. The back-propagation network consists of one input layer with the five neurons tPSA, %fPSA, patient age, prostate volume, and DRE status. Each ANN contains one hidden layer with three neurons. Each ANN finally contains one output neuron representing the output value as the probability of PCa. The activation function for the hidden neurons was the tanh while the activation function for the output neuron was linear in the range 0 to 1 to get a value for the probability of PCa. Training of the ANN took place in 4 steps with 100 sweeps each of them. Stopping criteria were a RMS error less than 0.001 or a rate of 95% correct classified samples. The initial weights were set randomly to values between -1 and +1. Before training all variables were normalized to mean value 0 and standard deviation 1 and ordered randomly. To avoid over-fitting we used 10-fold cross-validation. During training always 10% of the data were used for internal validation.</p>", "<p>The respective PSA- and ANN-follow ups were divided into three groups. Group 1 consisted of increasing values (PSA &gt;0.75 ng/mL/year; ANN &gt;4/year), group 2 of stable values (PSA -0.75 to 0.75 ng/mL/year; ANN -4 to 4/year), and group 3 of the decreasing follow up values (PSA &lt;-0.75 ng/mL/year; ANN &lt;-4/year). When analyzing the follow up of %fPSA only, it has been shown that due to the large variability between the measurements there is no usefulness at all for the parameter %fPSA velocity (data not shown).</p>", "<p>Statistical calculations were performed with SPSS 14.0 for Windows (SPSS, Chicago, USA). We used the non-parametric Mann-Whitney U test and the Kruskal-Wallis test. The diagnostic validity of all parameters was evaluated by Receiver-operating characteristic (ROC) curve analysis. The areas under the ROC curves (AUCs) and the specificities at 90% and 95% sensitivity were compared by a nonparametric method using the software GraphROC 2.1 for Windows. Significance was defined as <italic>P </italic>&lt; 0.05.</p>" ]

[ "<title>Results</title>", "<p>The median tPSA for the PCa patients at the time of diagnosis (last tPSA value) was 8.3 ng/mL (range 3.2–107 ng/mL). The BPH patients had a significantly lower (<italic>P </italic>&lt; 0.0001) median tPSA value of 5.3 ng/mL (0.4–37.1 ng/mL). Descriptive data for all analyzed tPSA ranges 0–4, 4.1–10, 10.1–20 and &gt;20 ng/mL for all PCa and BPH including the median tPSA, %fPSA and prostate volume are shown in Table ##TAB##0##1##. The median age for all patients was 68 years and the age distribution revealed no differences between the 4 tPSA groups for the PCa where the median age was 66 years (45–80 years). The median age for the BPH patients was somewhat but not significantly higher (68 years, range: 44–85, <italic>P </italic>= 0.1) but did also not differ between the 4 groups.</p>", "<p>Regarding the follow up of the PCa patients, the number of PSA and fPSA measurements ranged from 3 to 13 (median: 4.5) whereas the BPH patients had on average more PSA and fPSA measurements (range 3 to 22, median: 7). The distribution of the follow up related to the years before diagnosis of PCa or total follow up time for the BPH patients is shown in Table ##TAB##1##2##. The median follow up time for all patients was 3.4 years while PCa patients had a shorter median follow up (1.8 years) compared with BPH patients (4.2 years).</p>", "<p>Table ##TAB##2##3## shows the ROC analysis for all 199 patients by comparing the AUC for tPSA, %fPSA, PSAD, PSAV, ANN output and the ANNV. PSAD was the best parameter to differentiate between PCa and BPH and neither ANN nor ANNV could improve this. At 95% sensitivity, PSAD performed better than all other parameters. On the other hand, at 95% specificity, the ANNV was the best available parameter with a sensitivity of 32.7% and significantly better performance compared with all others except %fPSA (<italic>P </italic>= 0.44). A similar behavior is seen for the 4–10 ng/mL tPSA range in Table ##TAB##3##4##. Again, regarding the AUC comparison and the specificities at 95% sensitivity, PSAD performed best, but did only reach significance levels to all others at 95% sensitivity but not for the AUC comparison. At 95% specificity, the ANNV (sensitivity 37.5%) demonstrated also within the tPSA range 4–10 ng/mL the ability to perform significantly better than all other parameters except the ANN output (<italic>P </italic>= 0.07). Figure ##FIG##0##1## shows for the tPSA range 4–10 ng/mL that the ANNV has the steepest increase of the ROC curve with the highest sensitivities at 95% and 90% specificity, respectively. This may be more important for repeat biopsies, where biopsies in general should be avoided.</p>", "<p>In Table ##TAB##4##5## the respective three groups for PSAV and ANNV (increasing, stable and decreasing values) are given. More than two third of all PCa patients have the typical increasing PSAV. In comparison, the ANNV is only indicated at 45% of all PCa patients' increasing values. The differences between the PSAV and ANNV are also given in the Table ##TAB##4##5##. It can be seen that only BPH benefit from the additional ANNV since the stable values are significantly more (+17.4%). Also, there is a reduction of increasing values (-10.6%). This avoids repeated prostate biopsies in at least 11% of all BPH patients. Another observation is that more than half of all patients (52%) show an atypical PSAV with regard to their diagnosis.</p>", "<p>When using the traditional PSAV cutoff of 0.75 ng/mL/year the regular median PSAV for the PCa patients was 1.24 ng/mL/year whereas the PSAV for the BPH patients was 0.16 ng/mL/year (<italic>P </italic>&lt; 0.0001). Further descriptive data are given for the patients with increasing, stable and decreasing PSAV (Table ##TAB##5##6##) and ANNV (Table ##TAB##6##7##).</p>", "<p>When analyzing the short-term-PSAV over 12 months, the median PSAV for the PCa patients is higher with 1.61 ng/mL/year but the median PSAV for the BPH patients is almost zero with 0.04 ng/mL/year (<italic>P </italic>= 0.0001). There are only slight differences between PCa patients when looking at the PSAV for a 12 months period (data not shown). However, around one third of all BPH patients change the status of stable values which were visible over a long time observation to increasing or decreasing values when only calculating the PSAV over 12 months. Instead of 65% by using the regular PSAV, only 32% of all BPH patients had stable values when using the short-term-PSAV.</p>" ]

[ "<title>Discussion</title>", "<p>The discussion regarding the use of PSAV to improve the low specificity of PSA has gained increasing attention. Initially, Carter et al. [##REF##7505974##6##] presented at a PSAV cut off of 0.75 ng/mL/year a specificity of 90% – significantly higher than the 60% specificity of a single PSA cutoff of 4 ng/mL. The results, though, were based on analyzing only 18 cancer cases [##REF##7505974##6##]. In a current analysis on a large cohort of patients the authors found that the PSAV cutoff of 0.75 ng/mL/year underestimated the risk of PCa especially in younger men and recommended age-adjusted cutoffs [##REF##17222618##9##]. However, this analysis excluded 45% of men (5,381 from 11,861 men) with a PSAV of 0 or less and 30% (504 from 1654) PCa patients without increasing PSAV. The remaining 70% PCa patients had an increasing PSAV, which is the same percentage of PCa patients as we found in our study with increasing PSA values (Table ##TAB##4##5##). However, in our analysis, all patients regardless of the PSAV were considered. On the other hand, 30% of all PCa patients do not present with the typical increasing PSA values. 10% of all PCa patients even have decreasing PSA values. This shows that PSAV can detect approximately 2/3 of all PCa patients but stable or decreasing PSA values do not really reduce the risk of having PCa if the PSA alone is elevated.</p>", "<p>It was assumed that additional clinical and laboratory data would improve the PCa detection rate when using ANN models where %fPSA, age, prostate volume and the DRE status are considered. However, this study could not demonstrate a positive effect by using the ANNV to detect more cancer patients since only 45% of all PCa patient had an increasing ANNV. This is a reduction of 26.5% compared with the PSAV. The number of stable or decreasing ANNV output values increased compared with the PSAV indicating that PSAV alone is the better indicator for a PCa risk. More than half of all PCa patients had a stable or even decreasing ANNV. It should be noted that this poor performance of the ANN is only related to the follow up but not to the ANN use at all for PCa detection. Here it has been demonstrated that ANN models with clinical and laboratory values can significantly improve the PCa detection rate compared with PSA and %fPSA [##REF##10962306##20##,##REF##12142385##22##, ####REF##16388506##23##, ##REF##16697520##24##, ##REF##17333205##25####17333205##25##]. However, the relatively small number of patients with only one third PCa is a limitation of this study compared with other ANN studies where no follow up was analyzed.</p>", "<p>Importantly, the inclusion of the ANNV can substantially save repeated biopsies in BPH patients. Whereas the PSAV shows only for 65.3% of the BPH patients the typical continuous follow up, this number increases to 82.7% when using the ANNV. Thus, at the best case approximately 17% of all BPH patients may benefit if taking into account not only the PSA but also the ANN follow up. When only looking at the difference between an increasing PSAV and increasing ANNV, the ANNV could save approximately 11% of all biopsies compared with the PSAV. As seen in Figure ##FIG##0##1##, the ANNV has the highest sensitivities at 95% and 90% specificity, respectively. Thus, relatively good sensitivity values at high specificity cutoffs argue for a usability of the ANNV especially for repeat biopsies, where biopsies in general should be avoided. This is an important result of the study. A further possibility is to look at partial ROC areas, which has been published before [##REF##15643193##26##]. When only including the AUC between 80% and 100% specificity, the ANNV has clearly the largest AUC compared with all others. Hence, for a ROC comparison one should not only consider the AUC but also the ROC curve shape for a better interpretation.</p>", "<p>Another problem is the relatively poor performance of the PSAV alone, which can be partially explained by the biological variation of PSA of up to 20% [##REF##15961552##14##]. Differences in PSA values regarding the used assays may be also responsible [##REF##10971300##15##]. However, this could be excluded in our study since only the IMMULITE assays were used for the tPSA and fPSA measurements over the whole time period from 1996 until 2006. The use of %fPSA revealed large differences between commonly used assays [##REF##11678748##16##, ####REF##16762996##17##, ##REF##16391327##18####16391327##18##]. In a recent study on 4,480 men in 5 different populations with 5 different PSA and fPSA assays and the application of different assay-adapted ANNs it has been demonstrated that our recently multicentric evaluated ANN \"ProstataClass\" [##REF##12142385##22##] should not be used without consideration of the PSA assay [##REF##17333205##25##]. In another study, Okamura and colleagues [##REF##15948720##27##] reported an acceptable comparability between two PSA assays by using a %fPSA-based logistic regression model.</p>", "<p>To calculate the PSAV we subdivided the PSA follow ups into three categories with increasing (&gt; 0.75 ng/mL/year), stable (-0.75 to 0.75 ng/mL/year) and decreasing (&lt; -0.75 ng/mL/year) values. The same procedure was performed with the ANNV, where 4/year was taken as cutoff. Contrary to others [##UREF##0##28##], we found it difficult to further subdivide also the category of inconsistent values. We did not find it useful to determine a definitive cutoff for the ANNV as Carter et al. [##REF##7505974##6##] did for the PSAV but the cutoff 4/year for the ANNV was taken for this preliminary study which equals to 90% specificity to have the possibility to discriminate between increasing, stable and decreasing values. However, the number of patients is relatively small and the usefulness of a cutoff especially for the PCa detection has not been shown.</p>", "<p>Recently, a study has reported that different methods to calculate the PSAV either with simple arithmetic or linear regression does not change the outcome [##REF##17085120##29##]. Data from this study were also not different when using the linear regression calculated compared with the arithmetic method (not shown).</p>", "<p>Whereas the advantage of ANN models compared with PSA or %fPSA has been proven in many studies [##REF##10962306##20##,##REF##12142385##22##], the situation for PSAV compared with PSA is unclear [##REF##17296371##8##,##REF##17270635##11##, ####REF##16442212##12##, ##REF##16622122##13####16622122##13##,##REF##14972481##30##,##REF##17382716##31##]. Studies on large populations have shown a clear advantage for PSAV compared with PSA alone [##REF##17296371##8##,##REF##17270635##11##,##REF##17382716##31##]. Apart from PSAV, the age and the tPSA range should be also considered [##REF##17296371##8##,##REF##17382716##31##]. Another study in screened men proved a significant difference for the PSAV between PCa (median: 0.62) and men with a negative biopsy (median: 0.46) but could not confirm a clinical advantage [##REF##14972481##30##]. Furthermore, the studies by Thompson et al. [##REF##16622122##13##] and Schroeder et al. [##REF##16442212##12##] also on large populations, did not show any advantage of using PSAV instead of PSA. In a recent review on studies of PSAV it was explained why the association between PSAV and disease-specific survival, which has been shown in other studies, does not necessarily imply that PSAV will be a useful screening tool [##REF##17925534##32##]. Moreover, other results show that patients who have a PSA which returns to normal levels still have a significant risk of PCa which led the authors to the conclusion that prostate biopsy might be most appropriate even after a single abnormal PSA [##UREF##0##28##]. Here we only partially agree because a simple repeated measurement of the PSA can avoid a significant number of biopsies and inclusion of ANN models give further certainty for a correct biopsy indication [##REF##15180645##33##].</p>" ]

[ "<title>Conclusion</title>", "<p>To conclude, this study demonstrates limited usefulness of PSAV to detect PCa with only 71% of increasing PSA values while approximately 30% of all PCa do not have the typical PSA follow up. The ANNV cannot improve the PCa detection rate but may save 11–17% of unnecessary prostate biopsies in BPH patients. Further studies on screening and larger populations are needed to determine the usefulness of the ANNV.</p>" ]

[ "<p>This is an Open Access article distributed under the terms of the Creative Commons Attribution License (<ext-link ext-link-type=\"uri\" xlink:href=\"http://creativecommons.org/licenses/by/2.0\"/>), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.</p>", "<title>Background</title>", "<p>To validate an artificial neural network (ANN) based on the combination of PSA velocity (PSAV) with a %free PSA-based ANN to enhance the discrimination between prostate cancer (PCa) and benign prostate hyperplasia (BPH).</p>", "<title>Methods</title>", "<p>The study comprised 199 patients with PCa (n = 49) or BPH (n = 150) with at least three PSA estimations and a minimum of three months intervals between the measurements. Patients were classified into three categories according to PSAV and ANN velocity (ANNV) calculated with the %free based ANN \"ProstataClass\". Group 1 includes the increasing PSA and ANN values, Group 2 the stable values, and Group 3 the decreasing values.</p>", "<title>Results</title>", "<p>71% of PCa patients typically have an increasing PSAV. In comparison, the ANNV only shows this in 45% of all PCa patients. However, BPH patients benefit from ANNV since the stable values are significantly more (83% vs. 65%) and increasing values are less frequently (11% vs. 21%) if the ANNV is used instead of the PSAV.</p>", "<title>Conclusion</title>", "<p>PSAV has only limited usefulness for the detection of PCa with only 71% increasing PSA values, while 29% of all PCa do not have the typical PSAV. The ANNV cannot improve the PCa detection rate but may save 11–17% of unnecessary prostate biopsies in known BPH patients.</p>" ]

[ "<title>Abbreviations</title>", "<p>ANN: artificial neural network; ANNV: ANN velocity; BPH: benign prostatic hyperplasia; DRE: digital rectal examination; fPSA: free PSA; PCa: prostate cancer; PSA: prostate-specific antigen; PSAD: PSA density; PSAV: PSA velocity; percent free PSA; tPSA: total PSA; TRUS: transrectal ultrasound</p>", "<title>Competing interests</title>", "<p>The authors declare that they have no competing interests.</p>", "<title>Authors' contributions</title>", "<p>All authors read and approved the final manuscript.</p>", "<title>Pre-publication history</title>", "<p>The pre-publication history for this paper can be accessed here:</p>", "<p><ext-link ext-link-type=\"uri\" xlink:href=\"http://www.biomedcentral.com/1471-2490/8/10/prepub\"/></p>" ]

[ "<title>Acknowledgements</title>", "<p>This work was supported by Diagnostics Products Corp. and by the Berliner Sparkassenstiftung Medizin. We gratefully thank Paul EC Sibley for his helpful corrections.</p>" ]

[ "<fig position=\"float\" id=\"F1\"><label>Figure 1</label><caption><p>ROC curves for tPSA (green, AUC 0.5), %fPSA (blue, AUC 0.64), PSAD (black, AUC 0.69) and ANNV (red, AUC 0.57) to show the different behavior of the curve regardless of the AUC at tPSA 4–10 ng/mL (PSAV and ANN not shown, given in table 4).</p></caption></fig>" ]

[ "<table-wrap position=\"float\" id=\"T1\"><label>Table 1</label><caption><p>Distribution of patients within the different PSA ranges and median values for tPSA, %fPSA and prostate volume for all patients</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td/><td align=\"center\" colspan=\"4\">All patients</td><td align=\"center\" colspan=\"4\">PCa</td><td align=\"center\" colspan=\"4\">BPH</td></tr><tr><td/><td colspan=\"12\"><hr/></td></tr><tr><td align=\"center\">tPSA range (ng/mL)</td><td align=\"center\">number</td><td align=\"center\">tPSA (ng/mL)</td><td align=\"center\">%fPSA (%)</td><td align=\"center\">volume (ml)</td><td align=\"center\">number</td><td align=\"center\">tPSA (ng/mL)</td><td align=\"center\">%fPSA (%)</td><td align=\"center\">volume (ml)</td><td align=\"center\">number</td><td align=\"center\">tPSA (ng/mL)</td><td align=\"center\">%fPSA (%)</td><td align=\"center\">volume (ml)</td></tr></thead><tbody><tr><td align=\"center\">0–4</td><td align=\"center\">47</td><td align=\"center\">2.1</td><td align=\"center\">19.4</td><td align=\"center\">40</td><td align=\"center\">2</td><td align=\"center\">3.5</td><td align=\"center\">17.2</td><td align=\"center\">36.5</td><td align=\"center\">45</td><td align=\"center\">2.1</td><td align=\"center\">19.5</td><td align=\"center\">40</td></tr><tr><td align=\"center\">4.1–10</td><td align=\"center\">94</td><td align=\"center\">5.6</td><td align=\"center\">15.9</td><td align=\"center\">44.5</td><td align=\"center\">24</td><td align=\"center\">5.8</td><td align=\"center\">14.3*</td><td align=\"center\">33.5*</td><td align=\"center\">70</td><td align=\"center\">5.6</td><td align=\"center\">17.4</td><td align=\"center\">50</td></tr><tr><td align=\"center\">10.1–20</td><td align=\"center\">50</td><td align=\"center\">13.6</td><td align=\"center\">11</td><td align=\"center\">49.1</td><td align=\"center\">20</td><td align=\"center\">13.5</td><td align=\"center\">9.7*</td><td align=\"center\">44.5*</td><td align=\"center\">30</td><td align=\"center\">13.9</td><td align=\"center\">13.7</td><td align=\"center\">59</td></tr><tr><td align=\"center\">&gt;20</td><td align=\"center\">8</td><td align=\"center\">25.7</td><td align=\"center\">9</td><td align=\"center\">35.5</td><td align=\"center\">3</td><td align=\"center\">29.3</td><td align=\"center\">6.8*</td><td align=\"center\">35*</td><td align=\"center\">5</td><td align=\"center\">25.6</td><td align=\"center\">11</td><td align=\"center\">80</td></tr><tr><td align=\"center\">all</td><td align=\"center\">199</td><td align=\"center\">6.2</td><td align=\"center\">15.1</td><td align=\"center\">45</td><td align=\"center\">49</td><td align=\"center\">8.3*</td><td align=\"center\">10.9*</td><td align=\"center\">35*</td><td align=\"center\">150</td><td align=\"center\">5.3</td><td align=\"center\">16.7</td><td align=\"center\">47</td></tr></tbody></table></table-wrap>", "<table-wrap position=\"float\" id=\"T2\"><label>Table 2</label><caption><p>PSA follow up for all patients for all patient groups</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td/><td align=\"center\" colspan=\"2\">all patients</td><td align=\"center\" colspan=\"2\">PCa</td><td align=\"center\" colspan=\"2\">BPH</td></tr><tr><td/><td colspan=\"6\"><hr/></td></tr><tr><td align=\"left\">Follow up (years)</td><td align=\"center\">number</td><td align=\"center\">percentage in %</td><td align=\"center\">number</td><td align=\"center\">percentage in %</td><td align=\"center\">number</td><td align=\"center\">percentage in %</td></tr></thead><tbody><tr><td align=\"left\">0.5 to 1</td><td align=\"center\">12</td><td align=\"center\">6</td><td align=\"center\">6</td><td align=\"center\">12</td><td align=\"center\">6</td><td align=\"center\">4</td></tr><tr><td align=\"left\">1 to 2</td><td align=\"center\">47</td><td align=\"center\">24</td><td align=\"center\">21</td><td align=\"center\">43</td><td align=\"center\">26</td><td align=\"center\">18</td></tr><tr><td align=\"left\">2 to 4</td><td align=\"center\">56</td><td align=\"center\">28</td><td align=\"center\">15</td><td align=\"center\">31</td><td align=\"center\">41</td><td align=\"center\">27</td></tr><tr><td align=\"left\">4 to 6</td><td align=\"center\">41</td><td align=\"center\">20.5</td><td align=\"center\">5</td><td align=\"center\">10</td><td align=\"center\">36</td><td align=\"center\">24</td></tr><tr><td align=\"left\">6 to 9</td><td align=\"center\">43</td><td align=\"center\">21.5</td><td align=\"center\">2</td><td align=\"center\">4</td><td align=\"center\">41</td><td align=\"center\">27</td></tr><tr><td colspan=\"7\"><hr/></td></tr><tr><td align=\"left\">all</td><td align=\"center\">199</td><td align=\"center\">100</td><td align=\"center\">49</td><td align=\"center\">100</td><td align=\"center\">150</td><td align=\"center\">100</td></tr></tbody></table></table-wrap>", "<table-wrap position=\"float\" id=\"T3\"><label>Table 3</label><caption><p>Areas under the curves (AUC), specificities at 95% sensitivity and sensitivities at 95% specificity with the respective confidence intervals (in parenthesis) for the parameters tPSA, %fPSA, PSAD^§, PSAV, ANN and ANNV^$for all patients (n = 199)</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td align=\"left\">Parameter</td><td align=\"center\">AUC<break/>(Confidence Intervals)</td><td align=\"center\">P-values and significance levels^§</td><td align=\"center\">Specificity at 95% Sensitivity</td><td align=\"center\">P-values and significance levels^§</td><td align=\"center\">Sensitivity at 95% Specificity</td><td align=\"center\">P-values and significance levels^$</td></tr></thead><tbody><tr><td align=\"left\">PSA<break/></td><td align=\"center\">0.69<break/>(0.61–0.77)</td><td align=\"center\">0.0001**<break/></td><td align=\"center\">27.3<break/>(21.4–34)</td><td align=\"center\">&lt;0.0001***<break/></td><td align=\"center\">14.3<break/>(7–25.4)</td><td align=\"center\">0.008**<break/></td></tr><tr><td align=\"left\">%fPSA<break/></td><td align=\"center\">0.70<break/>(0.71–0.78)</td><td align=\"center\">0.007**<break/></td><td align=\"center\">17.3<break/>(12.5–23.3)</td><td align=\"center\">&lt;0.0001***<break/></td><td align=\"center\">20.4<break/>(11.6–32.3)</td><td align=\"center\">0.44<break/></td></tr><tr><td align=\"left\">PSAD<break/></td><td align=\"center\">0.76<break/>(0.69–0.83)</td><td align=\"center\">-<break/></td><td align=\"center\">44<break/>(37.1–51.7)</td><td align=\"center\">-<break/></td><td align=\"center\">16.3<break/>(8.5–27.7)</td><td align=\"center\">0.013*<break/></td></tr><tr><td align=\"left\">PSAV<break/></td><td align=\"center\">0.76<break/>(0.67–0.84)</td><td align=\"center\">0.835<break/></td><td align=\"center\">4.7<break/>(2.2–8.7)</td><td align=\"center\">&lt;0.0001***<break/></td><td align=\"center\">16.3<break/>(8.5–27.7)</td><td align=\"center\">0.023*<break/></td></tr><tr><td align=\"left\">ANN<break/></td><td align=\"center\">0.66<break/>(0.57–0.75)</td><td align=\"center\">0.001**<break/></td><td align=\"center\">10<break/>(6.3–15.1)</td><td align=\"center\">&lt;0.0001***<break/></td><td align=\"center\">18.4<break/>(10–30)</td><td align=\"center\">0.023*<break/></td></tr><tr><td align=\"left\">ANNV<break/></td><td align=\"center\">0.56<break/>(0.44–0.68)</td><td align=\"center\">&lt;0.0001***<break/></td><td align=\"center\">1.3<break/>(0.2–4.3)</td><td align=\"center\">&lt;0.0001***<break/></td><td align=\"center\">32.7<break/>(21.7–45.4)</td><td align=\"center\">-<break/></td></tr></tbody></table></table-wrap>", "<table-wrap position=\"float\" id=\"T4\"><label>Table 4</label><caption><p>Areas under the curves (AUC), specificities at 95% sensitivity and sensitivities at 95% specificity with the respective confidence intervals (in parenthesis) for the parameters tPSA, %fPSA, PSAD^§, PSAV, ANN and ANNV^$for the tPSA range 4–10 ng/mL (n = 94)</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td align=\"left\">Parameter</td><td align=\"center\">AUC<break/>(Confidence Intervals)</td><td align=\"center\">P-values and significance levels^§</td><td align=\"center\">Specificity at 95% Sensitivity</td><td align=\"center\">P-values and significance levels^§</td><td align=\"center\">Sensitivity at 95% Specificity</td><td align=\"center\">P-values and significance levels^$</td></tr></thead><tbody><tr><td align=\"left\">PSA<break/></td><td align=\"center\">0.50<break/>(0.36–0.63)</td><td align=\"center\">&lt;0.0001***<break/></td><td align=\"center\">11.4<break/>(5.9–19.8)</td><td align=\"center\">0.002**<break/></td><td align=\"center\">4.2<break/>(1.3–19.1)</td><td align=\"center\">0.037*<break/></td></tr><tr><td align=\"left\">%fPSA<break/></td><td align=\"center\">0.64<break/>(0.51–0.77)</td><td align=\"center\">0.09<break/></td><td align=\"center\">20<break/>(12.6–29.6)</td><td align=\"center\">0.045*<break/></td><td align=\"center\">12.5<break/>(3.5–29.7)</td><td align=\"center\">0.023*<break/></td></tr><tr><td align=\"left\">PSAD<break/></td><td align=\"center\">0.69<break/>(0.58–0.81)</td><td align=\"center\">-<break/></td><td align=\"center\">35.7<break/>(26.2–46.2)</td><td align=\"center\">-<break/></td><td align=\"center\">4.2<break/>(1.3–19.1)</td><td align=\"center\">0.013*<break/></td></tr><tr><td align=\"left\">PSAV<break/></td><td align=\"center\">0.66<break/>(0.53–0.80)</td><td align=\"center\">0.529<break/></td><td align=\"center\">4.3<break/>(1.2–11.0)</td><td align=\"center\">&lt;0.0001***<break/></td><td align=\"center\">12.5<break/>(3.5–29.7)</td><td align=\"center\">0.023*<break/></td></tr><tr><td align=\"left\">ANN<break/></td><td align=\"center\">0.66<break/>(0.53–0.79)</td><td align=\"center\">0.264<break/></td><td align=\"center\">8.6<break/>(3.8–16.4)</td><td align=\"center\">&lt;0.0001***<break/></td><td align=\"center\">16.7<break/>(5.9–34.6)</td><td align=\"center\">0.074<break/></td></tr><tr><td align=\"left\">ANNV<break/></td><td align=\"center\">0.57<break/>(0.40–0.75)</td><td align=\"center\">0.008**<break/></td><td align=\"center\">0<break/></td><td align=\"center\">&lt;0.0001***<break/></td><td align=\"center\">37.5<break/>(21.2–56.4)</td><td align=\"center\">-<break/></td></tr></tbody></table></table-wrap>", "<table-wrap position=\"float\" id=\"T5\"><label>Table 5</label><caption><p>Comparison of the PSA and ANN velocity in PCa and BPH patients</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td/><td align=\"center\" colspan=\"2\">PSA velocity (PSAV)</td><td align=\"center\" colspan=\"4\">ANN velocity (ANNV)</td></tr></thead><tbody><tr><td align=\"left\">follow up group</td><td align=\"center\">PCa in %<break/>(n = 49)</td><td align=\"center\">BPH in %<break/>(n = 150)</td><td align=\"center\">PCa in %<break/>(n = 49)</td><td align=\"center\">Difference to PSAV</td><td align=\"center\">BPH in %<break/>(n = 150)</td><td align=\"center\">Difference to PSAV</td></tr><tr><td colspan=\"7\"><hr/></td></tr><tr><td align=\"left\">increasing</td><td align=\"center\">71.4%</td><td align=\"center\">21.3%</td><td align=\"center\">44.9%</td><td align=\"center\">- 26.5%</td><td align=\"center\">10.7%</td><td align=\"center\">- 10.6%</td></tr><tr><td align=\"left\">stable*</td><td align=\"center\">18.4%</td><td align=\"center\">65.3%</td><td align=\"center\">34.7%</td><td align=\"center\">+ 16.3%</td><td align=\"center\">82.7%</td><td align=\"center\">+ 17.4%</td></tr><tr><td align=\"left\">decreasing</td><td align=\"center\">10.2%</td><td align=\"center\">13.3%</td><td align=\"center\">20.4%</td><td align=\"center\">+ 10.2%</td><td align=\"center\">6.7%</td><td align=\"center\">- 6.6%</td></tr></tbody></table></table-wrap>", "<table-wrap position=\"float\" id=\"T6\"><label>Table 6</label><caption><p>Median values and p-values between the 3 groups of increasing, stable or decreasing PSAV values</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td/><td align=\"left\" colspan=\"3\">Increasing PSAV<break/>(&gt;0.75 ng/mL/year)</td><td align=\"left\" colspan=\"3\">Stable PSAV<break/>(-0.75 to 0.75 ng/mL/year)</td><td align=\"left\" colspan=\"3\">Decreasing PSAV<break/>(&lt; -0.75 ng/mL/year)</td></tr></thead><tbody><tr><td align=\"left\">Parameter</td><td align=\"left\">PCa (n = 35)</td><td align=\"left\">BPH (n = 32)</td><td align=\"left\">p-value</td><td align=\"left\">PCa (n = 9)</td><td align=\"left\">BPH (n = 98)</td><td align=\"left\">p-value</td><td align=\"left\">PCa (n = 5)</td><td align=\"left\">BPH (n = 20)</td><td align=\"left\">p-value</td></tr><tr><td colspan=\"10\"><hr/></td></tr><tr><td align=\"left\">Age (years)</td><td align=\"left\">65</td><td align=\"left\">68</td><td align=\"left\">0.15</td><td align=\"left\">68</td><td align=\"left\">68</td><td align=\"left\">0.86</td><td align=\"left\">72</td><td align=\"left\">68.5</td><td align=\"left\">0.92</td></tr><tr><td align=\"left\">tPSA (ng/mL)</td><td align=\"left\">11.7*</td><td align=\"left\">11.5^*$</td><td align=\"left\">0.72</td><td align=\"left\">4.8^§</td><td align=\"left\">4.96</td><td align=\"left\">0.96</td><td align=\"left\">6.2*</td><td align=\"left\">4.39</td><td align=\"left\">0.067</td></tr><tr><td align=\"left\">%fPSA (%)</td><td align=\"left\">9.6^*§</td><td align=\"left\">14.1</td><td align=\"left\">0.001</td><td align=\"left\">15</td><td align=\"left\">17.4</td><td align=\"left\">0.49</td><td align=\"left\">21</td><td align=\"left\">16</td><td align=\"left\">0.13</td></tr><tr><td align=\"left\">Volume (mL)</td><td align=\"left\">35</td><td align=\"left\">46.5</td><td align=\"left\">0.066</td><td align=\"left\">33</td><td align=\"left\">45</td><td align=\"left\">0.1</td><td align=\"left\">55</td><td align=\"left\">57.5</td><td align=\"left\">0.89</td></tr></tbody></table></table-wrap>", "<table-wrap position=\"float\" id=\"T7\"><label>Table 7</label><caption><p>Median values and p-values between the 3 groups of increasing, stable or decreasing ANNV values</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td/><td align=\"left\" colspan=\"3\">Increasing ANNV (&gt; 4)</td><td align=\"left\" colspan=\"3\">Stable ANNV (-4 to 4)</td><td align=\"left\" colspan=\"3\">Decreasing ANNV (&lt; -4)</td></tr></thead><tbody><tr><td align=\"left\">Parameter</td><td align=\"left\">PCa (n = 22)</td><td align=\"left\">BPH (n = 16)</td><td align=\"left\">p-value</td><td align=\"left\">PCa (n = 17)</td><td align=\"left\">BPH (n = 124)</td><td align=\"left\">p-value</td><td align=\"left\">PCa (n = 10)</td><td align=\"left\">BPH (n = 10)</td><td align=\"left\">p-value</td></tr><tr><td colspan=\"10\"><hr/></td></tr><tr><td align=\"left\">Age (years)</td><td align=\"left\">65*</td><td align=\"left\">68.5</td><td align=\"left\">0.61</td><td align=\"left\">70</td><td align=\"left\">68</td><td align=\"left\">0.7</td><td align=\"left\">62</td><td align=\"left\">64</td><td align=\"left\">0.29</td></tr><tr><td align=\"left\">tPSA (ng/mL)</td><td align=\"left\">8.3</td><td align=\"left\">6.8</td><td align=\"left\">0.17</td><td align=\"left\">10.2</td><td align=\"left\">5.3</td><td align=\"left\">0.002</td><td align=\"left\">5.5</td><td align=\"left\">4.35</td><td align=\"left\">0.5</td></tr><tr><td align=\"left\">%fPSA (%)</td><td align=\"left\">8.0^*$</td><td align=\"left\">13.15</td><td align=\"left\">0.017</td><td align=\"left\">15.15</td><td align=\"left\">17.4</td><td align=\"left\">0.85</td><td align=\"left\">11.2</td><td align=\"left\">13.8</td><td align=\"left\">0.1</td></tr><tr><td align=\"left\">Volume (mL)</td><td align=\"left\">33.5*</td><td align=\"left\">36.5*</td><td align=\"left\">0.34</td><td align=\"left\">50</td><td align=\"left\">53.5^$</td><td align=\"left\">0.87</td><td align=\"left\">30.5*</td><td align=\"left\">34*</td><td align=\"left\">0.26</td></tr></tbody></table></table-wrap>" ]

[]

[]

[]

[]

[]

[]

[ "<table-wrap-foot><p>*significantly different from BPH patients with <italic>P </italic>&lt; 0.0001</p></table-wrap-foot>", "<table-wrap-foot><p>^§PSAD with largest AUC and highest specificity at 95% sensitivity, all others compared to PSAD</p><p>^$ANNV with highest sensitivity at 95% specificity, all others compared to ANNV</p><p>*<italic>P </italic>&lt; 0.05</p><p>**<italic>P </italic>&lt; 0.01</p><p>***<italic>P </italic>&lt; 0.0001</p></table-wrap-foot>", "<table-wrap-foot><p>^§PSAD with largest AUC and highest specificity at 95% sensitivity, all others compared to PSAD</p><p>^$ANNV with highest sensitivity at 95% specificity, all others compared to ANNV</p><p>*<italic>P </italic>&lt; 0.05</p><p>**<italic>P </italic>&lt; 0.01</p><p>***<italic>P </italic>&lt; 0.0001</p></table-wrap-foot>", "<table-wrap-foot><p>*by using the cutoff of 0.75 ng/mL/year for tPSA (range -0.75 to 0.75) and by using the cutoff of 4 per year for ANN output (range -4 to 4)</p></table-wrap-foot>", "<table-wrap-foot><p>The p-value is given for the comparison between the respective PCa and BPH patients (Mann-Whitney U Test)</p><p>*significantly different to the respective patients in the stable group (<italic>P </italic>&lt; 0.05; Mann-Whitney U Test)</p><p>^§significantly different to the respective patients in the decreasing group (<italic>P </italic>&lt; 0.05; Mann-Whitney U Test)</p><p>The Kruskal-Wallis Test for the PCa patients between all 3 groups showed for tPSA (p = 0.0003) and %fPSA (p = 0.0008) significant differences but not for age (p = 0.26) or volume (p = 0.17).</p><p>The Kruskal-Wallis Test for the BPH patients between all 3 groups showed for tPSA (p &lt; 0.0001) significant differences but not for %fPSA (p = 0.4), age (p = 0.96) or volume (p = 0.44).</p></table-wrap-foot>", "<table-wrap-foot><p>The p-value is given for the comparison between the respective PCa and BPH patients (Mann-Whitney U Test)</p><p>*significantly different to the respective patients in the stable group (<italic>P </italic>&lt; 0.05; Mann-Whitney U Test)</p><p>^§significantly different to the respective patients in the decreasing group (<italic>P </italic>&lt; 0.05; Mann-Whitney U Test)</p><p>The Kruskal-Wallis Test for the PCa patients between all 3 groups showed only for %fPSA (p = 0.014) significant differences but not for age (p = 0.32), tPSA (p = 0.16) or volume (p = 0.98).</p><p>The Kruskal-Wallis Test for the BPH patients between all 3 groups showed for volume (p = 0.022) significant differences but not for age (p = 0.31), tPSA (p = 0.61) or %fPSA (p = 0.17).</p></table-wrap-foot>" ]

[ "<graphic xlink:href=\"1471-2490-8-10-1\"/>" ]

[]

{ "acronym": [], "definition": [] }

[ "<title>Background</title>", "<p>The prevalence of childhood obesity increased dramatically during the last decades in industrialized countries [##REF##12365956##1##,##REF##15917848##2##]. This increase in prevalence seems rather to be due to a shift of the upper part of the body mass index (BMI) distribution than to a shift of the entire BMI distribution as for example observed in the NHANESIII survey from 1988 to 1994 [##REF##10918526##3##]. This increased positive skewness could be due to exposure to obesogenic environmental determinants among a subpopulation with a high degree of susceptibility. TV watching, formula feeding, smoking in pregnancy, maternal obesity or parental social class are well known environmental, constitutional or sociodemographic risk factors [##REF##16076998##4##,##REF##16339125##5##]. However, it remains unknown if these factors affect the entire BMI distribution or only parts of it. A recent descriptive study reported an effect of several risk factors for childhood obesity on upper BMI percentiles, while the middle part of the BMI distribution was virtually unaffected. However, this study did not adjust for potential confounders [##REF##18402677##6##].</p>", "<p>In the literature most authors used linear or logistic regression to model effects on body mass index (BMI) measures. However, BMI data are usually positively skewed, and therefore a transformation of the response variable and/or other regression methods might be more appropriate. Possible approaches include lognormal or Box Cox power transformations of the BMI prior to linear regression modeling, gamma regression, quantile regression or GAMLSS models.</p>", "<p>Quantile regression has been applied in various BMI-related studies [##REF##17470452##7##, ####REF##16140349##8##, ##REF##12548387##9####12548387##9##]. Several risk factors for increased adult body size had different effects on specific quantiles. Comparisons between different regression models were discussed, but not quantified by model fit criteria such as Akaike Information Criterion (AIC) [##UREF##0##10##].</p>", "<p>The aim of our study was to compare generalized linear models, GAMLSS models and quantile regression models among BMI data on 4967 preschoolers in order to identify the best approach for obesity risk factor analysis. Additionally, we aimed to assess the effect of different risk factors on the BMI distribution (change of mean, variance, skewness or kurtosis) that might have implications for preventive measures (population based approach vs. targeted approach).</p>" ]

[ "<title>Methods</title>", "<title>Data</title>", "<p>Data on 7026 children participating in the school entry health examination in Bavaria, Southern Germany, were collected between September 2001 and August 2002. Children's age ranged from 54 to 88 months. Parental questionnaires on sociodemographic, lifestyle and other risk factors for obesity were distributed together with the invitation to the compulsory school entry examination. Children's weight and height were measured in light clothing and with calibrated balances and fixed stadiometers during the examination. The study has been described in detail elsewhere [##REF##16076998##4##].</p>", "<p>Sex and age were considered as confounders, while explanatory variables with previously reported associations to childhood body composition were <italic>a priori </italic>considered as exposures (abbreviations in brackets). These exposure variables included maternal smoking in pregnancy (PS), amount of watching TV (TV), breast feeding (BF), daily meal frequency (MF), highest graduation of either parent (elementary/secondary/at least A-level) (PG), maternal BMI (MB) and child's weight gain from birth to 2 years of life (WG) [##REF##16076998##4##,##REF##16339125##5##,##UREF##1##11##]. The sample was confined to cases with complete information on these variables leaving data of 4967 children for the analyses.</p>", "<title>Statistical methods</title>", "<p>Simple linear regression uses an identity link and models the relationship between a dependent variable <bold><italic>Y</italic></bold>_{<italic>i</italic>}, independent variables (<italic>z</italic>₁, ..., <italic>z</italic>_{<italic>m</italic>}) with <italic>m </italic>as total number of covariates included, and residuals (<italic>ε</italic>₁, ..., <italic>ε</italic>_{<italic>n</italic>}) for the individual <italic>i</italic>, <italic>i </italic>= 1, ..., <italic>n</italic>. The model can be denoted as</p>", "<p></p>", "<p><italic>Generalized linear models </italic>(GLM) allow a more flexible modeling [##UREF##2##12##] of the linear predictor <italic>η</italic>_{<italic>i </italic>}= <italic>g</italic>(<italic>μ</italic>_{<italic>i</italic>}) which can be denoted as</p>", "<p></p>", "<p>The link function <italic>g</italic>(.) can be specified e.g. by</p>", "<p>• the identity link <italic>g</italic>(<italic>μ</italic>) = <italic>μ</italic>, resulting in the simple linear regression model,</p>", "<p>• the log link <italic>g</italic>(<italic>μ</italic>) = log(<italic>μ</italic>) yielding loglinear regression,</p>", "<p>• the Box Cox power link [##UREF##3##13##]</p>", "<p></p>", "<p>• or the inverse link <italic>g</italic>(<italic>μ</italic>) = <italic>μ</italic>^-1.</p>", "<p>The inverse link function is the natural link function for the normal gamma distribution and was used in this study to perform gamma regression.</p>", "<p>One approach for model selection is the Generalized Akaike Information Criterion (GAIC)</p>", "<p></p>", "<p>with <italic>c </italic>= 2 for the 'classical' Akaike Information Criterion (AIC) [##UREF##0##10##], and <italic>c </italic>= log(<italic>n</italic>) for the Bayes Information Criterion (BIC) [##UREF##4##14##]. The GAIC includes the log likelihood</p>", "<p></p>", "<p>containing the relevant parameter vector (e. g. <italic>μ</italic>) and a penalty term <italic>c </italic>× <italic>p </italic>for the number of parameters and <italic>p </italic>= <italic>m </italic>+ <italic>f </italic>with <italic>f </italic>for the extra degrees of freedom needed for special model fitting techniques (e. g. splines). A statistical model is considered as better fitting if its GAIC is smaller than the GAIC of another statistical model.</p>", "<p><italic>Generalized Additive Models for Location, Scale and Shape </italic>(GAMLSS) offer an approach to model data with consideration of <italic>μ </italic>as location parameter as well as <italic>σ </italic>as scale parameter, and the skewness parameter <italic>ν </italic>and the kurtosis parameter <italic>ζ </italic>as shape parameters. A GAMLSS model is based on independent observations <italic>y</italic>_{<italic>i </italic>}for <italic>i </italic>= 1, ..., <italic>n </italic>and monotone link functions <italic>g</italic>_{<italic>k</italic>}(.), relating the parameters <italic>μ</italic>, <italic>σ</italic>, <italic>ν </italic>and <italic>ζ </italic>to the <italic>J</italic>_{<italic>k </italic>}explanatory variables [##UREF##5##15##,##UREF##6##16##] through semiparametric predictors. The common choice of the link functions is:</p>", "<p></p>", "<p>A multiplicative rather than an additive model for <italic>μ </italic>can be obtained by setting <italic>g</italic>₁(<italic>μ</italic>) = log(<italic>μ</italic>). Calculations with GAMLSS in this study use the Box Cox t (BCT) distribution, which is defined as</p>", "<p></p>", "<p>with <italic>z </italic>assumed to follow a t distribution with <italic>ζ </italic>degrees of freedom (<italic>ζ </italic>&gt; 0). Under this assumption it is possible to perform likelihood calculations.</p>", "<p>Additionally, cubic and penalized splines were considered to model continuous covariates [##UREF##7##17##,##UREF##8##18##]. The model selection can also be performed by GAIC because GAMLSS represents a general framework of regression models, including the class of GLMs [##UREF##9##19##]. The authors of GAMLSS used values for <italic>c </italic>in the range of 2 to 3 to calculate the GAIC [##UREF##9##19##].</p>", "<p>In contrast to the above mentioned distribution based methods, quantile regression estimates conditional quantile functions. It can be used to obtain information about specific quantiles of the underlying distribution.</p>", "<p><italic>Quantile regression </italic>for the sample quantile <italic>τ </italic>works by minimizing</p>", "<p></p>", "<p>with the so-called check function [##UREF##10##20##]</p>", "<p></p>", "<p>In (3), the predictor in equation (1) is taken as <italic>η </italic>= <italic>Q</italic>_{<italic>τ </italic>}with <italic>Q</italic>_{<italic>τ </italic>}being the modeled <italic>τ </italic>quantile.</p>", "<p>The comparison of quantile regression and generalized linear models is a major challenge due to the inapplicability of the GAIC in quantile regression. To compare GAMLSS and quantile regression, we plotted estimated values of the 90^thand 97^thBMI percentiles for weight gain in the first two years, while the other covariates were considered at their mean values (if continuous) or their modes (if categorical). We similarly calculated the estimated percentiles for each category of meal frequency, holding the other variables fixed accordingly.</p>", "<p>All calculations were carried out with R 2.5.1 <ext-link ext-link-type=\"uri\" xlink:href=\"http://cran.r-project.org\"/>.</p>" ]

[ "<title>Results</title>", "<p>The overall mean of the BMI of the 4967 children was 15.34 kg/m²with a median of 15.08 kg/m². The data included 2585 males (vs. 2382 females), 417 (vs. 4550) children whose mother had smoked in pregnancy, 384 children with more than 2 TV hours per day (vs. 4583 in 3 lower categories), 1197 (vs. 3770) children who had never been breastfed, 816 children with 3 daily meals at maximum (vs. 4151 with 4 or more meals), and 1466 children whose parents had only an elementary school degree or less (vs. 3501 in other categories). In addition to these categorical covariates, we considered the metric variables children's age in months with a mean of 72.86 (SD 4.77), the maternal BMI (in kg/m²) which ranged from 15.9 to 49.5 (mean 23.44, SD 3.99), and the children's weight gain (in kg) in the first 2 years of life, ranging from 5.5 to 15.3 (mean 9.45, SD 1.40).</p>", "<p>Figure ##FIG##0##1## shows univariate non-parametric kernel density estimates of the children's BMI distributions with regard to underlying risk factors. Maternal BMI and weight gain in the first 2 years were categorized by common cut points (Maternal BMI &gt; 25 kg/m², weight gain ≥ 10 kg [##REF##16076998##4##]). When present, most risk factors seemed to increase BMI values of upper BMI regions: For example, there was a higher proportion of children with a BMI &gt; 18 in non-breastfed compared to breastfed children, although the distribution curves of both strata were of almost identical shape for BMI values of &lt; 18.</p>", "<p>Simple linear models assessing the impact of certain risk factors might be limited under such varying key characteristics of the density distributions with and without underlying risk factors due to their intense assumptions.</p>", "<p>In the multivariable regression analyses, we considered the following <italic>a priori </italic>defined interaction terms with reported or assumed interrelations: a) sex as confounder with every covariate except age, b) weight gain in the first 2 years with parental education [##REF##16076998##4##], c) weight gain in the first 2 years with breast feeding [##REF##15737952##21##] and d) maternal smoking in pregnancy with breastfeeding [##REF##17226091##22##].</p>", "<p>Full multivariable linear, loglinear, gamma and linear regression models with Box Cox power transformed BMI values included all covariates and all <italic>a priori </italic>defined interaction terms. The backward elimination procedure yielded models without any interaction term and without parents' graduate, maternal smoking in pregnancy or breastfeeding for all 4 GLM models,</p>", "<p></p>", "<p>with <italic>η </italic>= <italic>μ </italic>for LR, for example.</p>", "<p>We chose <italic>c </italic>= 3 in equation (2) for the GAIC because this factor yielded stable and plausible results in a univariate preanalysis (data not shown). We decided not to fit the multivariable GAMLSS model by considering all covariates from the beginning and starting the fitting process due to the high computational demand of this approach. Instead, we calculated separate univariate GAMLSS models for all covariates and thereafter combined the resulting models to a multivariable model in terms of a pre-selecting forward selection procedure. During the fitting process of univariate models, we considered the strict parameter hierarchy for GAMLSS models in four steps, according to the suggestion of the GAMLSS authors [##UREF##11##23##]: first a model for <italic>μ </italic>should be fitted, after that for <italic>σ</italic>, followed by <italic>ν </italic>and <italic>ζ</italic>. If a parameter term did not reduce the GAIC(3), it was not considered for the univariate model of the respective covariate. For example, <italic>ν </italic>and <italic>ζ </italic>did not enhance the fit of the univariate model for the variable watching TV, yielding (table ##TAB##0##1##):</p>", "<p></p>", "<p></p>", "<p></p>", "<p></p>", "<p>Cubic and penalized splines up to three degrees of freedom were considered in models of the continuous covariates age, maternal BMI and weight gain in the first 2 years. Parameters that were not significant anymore in the combined multivariable model were excluded from the final multivariable model. Apart from age, increase (or decrease) in the location parameter <italic>μ </italic>for covariates was always associated with significant increase (or decrease) in the scale parameter <italic>σ</italic>.</p>", "<p>The final multivariable GAMLSS model yielded the same significant covariates as the GLM methods using backward selection, with exception of breastfeeding for which the scale parameter <italic>σ </italic>was significant in the GAMLSS (tables ##TAB##0##1## and ##TAB##1##2##). The <italic>a priori </italic>defined interaction terms were not significant in any considered model.</p>", "<p>The fit of the multivariable GAMLSS was far better than the fit of the multivariable GLM models. The GAIC(3) of GAMLSS was 17 470, while linear regression with Box Cox Power transformation, gamma regression, loglinear regression and the simple linear regression model yielded increased GAICs with 17 955, 18 120, 18 219 and 18 616, respectively.</p>", "<p>Apart from parental education, all considered covariates were significant in quantile regression considering the quantile <italic>τ </italic>= 0.9 (equals 90^thpercentile). In quantile regression (QR) models with <italic>τ </italic>= 0.97 (equals 97^thpercentile), however, only TV watching, breastfeeding, meal frequency, maternal BMI and weight gain in first two years of life were significantly associated with child's BMI. For example, the model for QR, <italic>τ </italic>= 0.9, was (table ##TAB##2##3##):</p>", "<p></p>", "<p>An overview on significant variables in respective models and differences across models is shown in table ##TAB##1##2##. The covariates TV watching, meal frequency, maternal BMI and weight gain in the first two years of life were significantly associated with child's BMI regardless of the method or chosen link. In contrast, parental education was not significant in any multivariable model. Its influence on offspring's BMI might sufficiently be explained by effects of the other considered covariates. An effect of breastfeeding on the BMI distribution was only detected by GAMLSS and quantile regression. Pregnancy smoking, however, was only significant in the quantile regression model of the <italic>τ </italic>= 0.9 quantile.</p>", "<p>In figure ##FIG##1##2##, estimated values of the 90^thand 97^thBMI percentiles from GAMLSS and quantile regression were compared for weight gain with fixed values of the other covariates. Similarly, table ##TAB##3##4## shows percentile values estimated with both methods for different values of meal frequency. Both figure ##FIG##1##2## and table ##TAB##3##4## indicate that estimated values for the 90^thpercentile obtained by GAMLSS and quantile regression were similar, while the 97^thpercentile was slightly higher in quantile regression models. While percentile curves estimated by quantile regression were linear, those obtained by GAMLSS showed a shaped curve due to the combinations of the additional parameters <italic>σ</italic>, <italic>ν </italic>and <italic>ζ</italic>.</p>" ]

[ "<title>Discussion and conclusion</title>", "<p>In our study, GAMLSS showed a much better fit examining obesity risk factors compared to GLM models by GAIC. The same explanatory variables had significant associations to body composition across all GLM models, although models contained either additive (linear regression) or multiplicative components (loglinear regression, Box Cox regression and gamma regression).</p>", "<p>In general, GAMLSS offers a flexible approach due to the large number of implemented distribution families. With GAMLSS, it is possible to assess the effect of specific parameters on the outcome variable distribution. For example, we observed that some variables did not only affect the mean, but additionally the scale of the BMI distribution. Additionally, interdependencies of considered parameters can be examined by GAMLSS. We observed that an increase (decrease) of the mean (<italic>μ</italic>) was mostly associated with an increase (decrease) of the scale (<italic>σ</italic>). The scale parameter <italic>σ </italic>in the distribution used (BCT) in GAMLSS is an approximative centile based coefficient of variation measure [##UREF##6##16##]. Therefore risk factors of overweight seem to affect both, the BMI itself and its variation. For example, children with a high weight gain in the first 2 years of life had higher BMI values as well as a higher coefficient of variation in BMI compared to those with a low infant weight gain. Thus, low infant weight gain might be a better predictor for underweight than is high infant weight gain for overweight. A change of the skewness term <italic>ν</italic>, however, did not improve the goodness of fit for modeling the skewed BMI distribution. This might be due to a sufficient consideration of skewness by a change of both parameters <italic>μ </italic>and <italic>σ</italic>.</p>", "<p>Quantile regression allows additional interpretation, e.g. of risk factors affecting only parts of the distribution [##REF##17470452##7##]. While GAMLSS models consider the entire BMI distribution, quantile regression directly examines possible associations between explanatory variables and certain predefined percentiles. Logistic regression is in principal based on a similar idea, but in case of overweight, for example, it has to deal with a big loss of information due to transformation of the continuous BMI to a binary variable. Quantile regression, in contrast, uses the whole information of the data. Furthermore, the interpretations of logistic and quantile regression differ. For example, logistic regression assesses the odds ratio for overweight in relation to certain risk factors, whereas quantile regression quantifies the linear impact of risk factors on overweight children.</p>", "<p>In our study, the variables TV watching, maternal BMI and weight gain in the first 2 years of life were directly and meal frequency was inversely significantly associated with body composition in every examined model type. However, the strength of the associations was of different magnitude across model types (table ##TAB##3##4##).</p>", "<p>In our study breastfeeding seemed to have a protective effect on the upper percentiles of the BMI estimated by quantile regression (e.g. -0.41 for the 90^thpercentile, s. table ##TAB##2##3##), although generalized regression models and GAMLSS did not assess breastfeeding as being significantly associated with the mean BMI (although it was a significant predictor of <italic>σ</italic>). The latter is in accordance with a recent study on mean BMI and DXA derived fat mass measures [##REF##17556696##24##]. Additionally, different aspects might be detected by modeling different quantiles, for example quantiles referring to underweight.</p>", "<p>We confined our sample to cases with complete information in all variables. Since underreporting with respect to pregnancy smoking and high values of maternal BMI is well-known, this might have led to underestimation of the effects of the corresponding covariates on childhood BMI. However, such an underestimation is likely to similarly affect all examined statistical approaches and therefore be of minor relevance for assessment of the appropriate approach. It might be of interest, however, to compare how sensitive the statistical models are to several methods of missing data imputation such as multiple imputation. However, this question leads deeply into other statistical methodology and is therefore beyond the scope of our study.</p>", "<p>GAMLSS and quantile regression have recently been compared, along with many other methods, in a WHO study to identify standard reference values for child growth [##REF##16143968##25##]. Four out of five construction methods taken under further examination were GAMLSS methods with different distribution functions: Box Cox t (like in this study), Box Cox power exponential [##REF##15351960##26##], Box Cox normal [##REF##1518992##27##] and Johnson's SU (sinh^-1normal) [##REF##18132090##28##]. The other considered method used modulus-exponential-normal distribution [##UREF##12##29##]. The authors finally calculated reference values by GAMLSS with Box Cox power exponential distribution, using AIC and GAIC(3) in parallel for model selection [##UREF##13##30##]. This indicates that GAMLSS is a very appropriate method for constructing reference curves which are based on estimated percentile curves.</p>", "<p>In our study, a comparison of GAMLSS and quantile regression by estimated values of the 90^thand 97^thpercentiles with respect to certain covariates (weight gain and meal frequency) showed similar results for both methods at the 90^thpercentile, while the estimated 97^thpercentile was slightly higher in the quantile regression model. Since implementation of percentile curves is existent only for univariate models in the <italic>gamlss </italic>package, some computational effort was necessary to gain the respective GAMLSS curves with fixed effects of other covariates. Furthermore, it might be worthwhile to consider nonlinear quantile regression (20) in future studies.</p>", "<p>The statistical model that should be used, largely depends on the observed data and on the aim of the study. GAMLSS models provide exact modeling of continuous outcomes, e.g. for the calculation of standard reference values. While GLMs provide helpful information on mean response changes, GAMLSS additionally provides information on distribution parameters like scale or skewness. On the other hand, quantile regression can be used to model specific parts of the BMI distribution such as the 90^thor 97^thpercentile and should be preferred to logistic regression if the original scale of the outcome variable was continuous and a GLM or GAMLSS cannot answer the research question.</p>" ]

[ "<title>Discussion and conclusion</title>", "<p>In our study, GAMLSS showed a much better fit examining obesity risk factors compared to GLM models by GAIC. The same explanatory variables had significant associations to body composition across all GLM models, although models contained either additive (linear regression) or multiplicative components (loglinear regression, Box Cox regression and gamma regression).</p>", "<p>In general, GAMLSS offers a flexible approach due to the large number of implemented distribution families. With GAMLSS, it is possible to assess the effect of specific parameters on the outcome variable distribution. For example, we observed that some variables did not only affect the mean, but additionally the scale of the BMI distribution. Additionally, interdependencies of considered parameters can be examined by GAMLSS. We observed that an increase (decrease) of the mean (<italic>μ</italic>) was mostly associated with an increase (decrease) of the scale (<italic>σ</italic>). The scale parameter <italic>σ </italic>in the distribution used (BCT) in GAMLSS is an approximative centile based coefficient of variation measure [##UREF##6##16##]. Therefore risk factors of overweight seem to affect both, the BMI itself and its variation. For example, children with a high weight gain in the first 2 years of life had higher BMI values as well as a higher coefficient of variation in BMI compared to those with a low infant weight gain. Thus, low infant weight gain might be a better predictor for underweight than is high infant weight gain for overweight. A change of the skewness term <italic>ν</italic>, however, did not improve the goodness of fit for modeling the skewed BMI distribution. This might be due to a sufficient consideration of skewness by a change of both parameters <italic>μ </italic>and <italic>σ</italic>.</p>", "<p>Quantile regression allows additional interpretation, e.g. of risk factors affecting only parts of the distribution [##REF##17470452##7##]. While GAMLSS models consider the entire BMI distribution, quantile regression directly examines possible associations between explanatory variables and certain predefined percentiles. Logistic regression is in principal based on a similar idea, but in case of overweight, for example, it has to deal with a big loss of information due to transformation of the continuous BMI to a binary variable. Quantile regression, in contrast, uses the whole information of the data. Furthermore, the interpretations of logistic and quantile regression differ. For example, logistic regression assesses the odds ratio for overweight in relation to certain risk factors, whereas quantile regression quantifies the linear impact of risk factors on overweight children.</p>", "<p>In our study, the variables TV watching, maternal BMI and weight gain in the first 2 years of life were directly and meal frequency was inversely significantly associated with body composition in every examined model type. However, the strength of the associations was of different magnitude across model types (table ##TAB##3##4##).</p>", "<p>In our study breastfeeding seemed to have a protective effect on the upper percentiles of the BMI estimated by quantile regression (e.g. -0.41 for the 90^thpercentile, s. table ##TAB##2##3##), although generalized regression models and GAMLSS did not assess breastfeeding as being significantly associated with the mean BMI (although it was a significant predictor of <italic>σ</italic>). The latter is in accordance with a recent study on mean BMI and DXA derived fat mass measures [##REF##17556696##24##]. Additionally, different aspects might be detected by modeling different quantiles, for example quantiles referring to underweight.</p>", "<p>We confined our sample to cases with complete information in all variables. Since underreporting with respect to pregnancy smoking and high values of maternal BMI is well-known, this might have led to underestimation of the effects of the corresponding covariates on childhood BMI. However, such an underestimation is likely to similarly affect all examined statistical approaches and therefore be of minor relevance for assessment of the appropriate approach. It might be of interest, however, to compare how sensitive the statistical models are to several methods of missing data imputation such as multiple imputation. However, this question leads deeply into other statistical methodology and is therefore beyond the scope of our study.</p>", "<p>GAMLSS and quantile regression have recently been compared, along with many other methods, in a WHO study to identify standard reference values for child growth [##REF##16143968##25##]. Four out of five construction methods taken under further examination were GAMLSS methods with different distribution functions: Box Cox t (like in this study), Box Cox power exponential [##REF##15351960##26##], Box Cox normal [##REF##1518992##27##] and Johnson's SU (sinh^-1normal) [##REF##18132090##28##]. The other considered method used modulus-exponential-normal distribution [##UREF##12##29##]. The authors finally calculated reference values by GAMLSS with Box Cox power exponential distribution, using AIC and GAIC(3) in parallel for model selection [##UREF##13##30##]. This indicates that GAMLSS is a very appropriate method for constructing reference curves which are based on estimated percentile curves.</p>", "<p>In our study, a comparison of GAMLSS and quantile regression by estimated values of the 90^thand 97^thpercentiles with respect to certain covariates (weight gain and meal frequency) showed similar results for both methods at the 90^thpercentile, while the estimated 97^thpercentile was slightly higher in the quantile regression model. Since implementation of percentile curves is existent only for univariate models in the <italic>gamlss </italic>package, some computational effort was necessary to gain the respective GAMLSS curves with fixed effects of other covariates. Furthermore, it might be worthwhile to consider nonlinear quantile regression (20) in future studies.</p>", "<p>The statistical model that should be used, largely depends on the observed data and on the aim of the study. GAMLSS models provide exact modeling of continuous outcomes, e.g. for the calculation of standard reference values. While GLMs provide helpful information on mean response changes, GAMLSS additionally provides information on distribution parameters like scale or skewness. On the other hand, quantile regression can be used to model specific parts of the BMI distribution such as the 90^thor 97^thpercentile and should be preferred to logistic regression if the original scale of the outcome variable was continuous and a GLM or GAMLSS cannot answer the research question.</p>" ]

[ "<p>This is an Open Access article distributed under the terms of the Creative Commons Attribution License (<ext-link ext-link-type=\"uri\" xlink:href=\"http://creativecommons.org/licenses/by/2.0\"/>), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.</p>", "<title>Background</title>", "<p>Body mass index (BMI) data usually have skewed distributions, for which common statistical modeling approaches such as simple linear or logistic regression have limitations.</p>", "<title>Methods</title>", "<p>Different regression approaches to predict childhood BMI by goodness-of-fit measures and means of interpretation were compared including generalized linear models (GLMs), quantile regression and Generalized Additive Models for Location, Scale and Shape (GAMLSS). We analyzed data of 4967 children participating in the school entry health examination in Bavaria, Germany, from 2001 to 2002. TV watching, meal frequency, breastfeeding, smoking in pregnancy, maternal obesity, parental social class and weight gain in the first 2 years of life were considered as risk factors for obesity.</p>", "<title>Results</title>", "<p>GAMLSS showed a much better fit regarding the estimation of risk factors effects on transformed and untransformed BMI data than common GLMs with respect to the generalized Akaike information criterion. In comparison with GAMLSS, quantile regression allowed for additional interpretation of prespecified distribution quantiles, such as quantiles referring to overweight or obesity. The variables TV watching, maternal BMI and weight gain in the first 2 years were directly, and meal frequency was inversely significantly associated with body composition in any model type examined. In contrast, smoking in pregnancy was not directly, and breastfeeding and parental social class were not inversely significantly associated with body composition in GLM models, but in GAMLSS and partly in quantile regression models. Risk factor specific BMI percentile curves could be estimated from GAMLSS and quantile regression models.</p>", "<title>Conclusion</title>", "<p>GAMLSS and quantile regression seem to be more appropriate than common GLMs for risk factor modeling of BMI data.</p>" ]

[ "<title>Competing interests</title>", "<p>The authors declare that they have no competing interests.</p>", "<title>Authors' contributions</title>", "<p>The authors' responsibilities were as follows: AB (guarantor) did the statistical analysis with help by LF and wrote the first draft of the manuscript. AMT, LF and UM reviewed and critiqued the manuscript and made substantial intellectual contributions to subsequent drafts. AB and AMT had the idea for the study and wrote the final draft together.</p>", "<title>Pre-publication history</title>", "<p>The pre-publication history for this paper can be accessed here:</p>", "<p><ext-link ext-link-type=\"uri\" xlink:href=\"http://www.biomedcentral.com/1471-2288/8/59/prepub\"/></p>" ]

[ "<title>Acknowledgements</title>", "<p>This study was supported by the innovative research priority project Munich Center of Health Sciences (sub-project II) of the Ludwig Maximilians University Munich and by grants of the Bundesministerium für Bildung und Forschung (Obesity network: LARGE).</p>", "<p>We thank Nora Fenske for her help in computing the comparison between GAMLSS and quantile regression.</p>" ]

[ "<fig position=\"float\" id=\"F1\"><label>Figure 1</label><caption><p>Univariate density distributions of children's BMI with regard to underlying risk factors. Maternal BMI and weight gain in the first two years were divided up into two categories. The risk factors seem to produce a slightly right-skewed distribution for exposed in comparison to non-exposed children, whereas the confounder variable sex does not.</p></caption></fig>", "<fig position=\"float\" id=\"F2\"><label>Figure 2</label><caption><p>Values for the 90^thand 97^thBMI percentiles in respect to weight gain in the first two years (in kg), estimated by GAMLSS (dark lines) and quantile regression (grey lines), with fixed values for all other covariates. The dashed lines denote the estimated values for the 97^thpercentiles for GAMLSS and quantile regression (QR), respectively. The dots represent observed values in the dataset.</p></caption></fig>" ]

[ "<table-wrap position=\"float\" id=\"T1\"><label>Table 1</label><caption><p>Estimators (EST) and 95% confidence intervals (CI) of the multivariable GAMLSS model in the School Entry Health Examination Study in Bavaria, 2001–2002.</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td align=\"left\">Variable</td><td align=\"left\" colspan=\"2\"></td><td align=\"left\" colspan=\"2\">log </td><td align=\"left\" colspan=\"2\"></td><td align=\"left\" colspan=\"2\">log </td></tr></thead><tbody><tr><td/><td align=\"left\">EST</td><td align=\"left\">95% CI</td><td align=\"left\">EST</td><td align=\"left\">95% CI</td><td align=\"left\">EST</td><td align=\"left\">95% CI</td><td align=\"left\">EST</td><td align=\"left\">95% CI</td></tr><tr><td colspan=\"9\"><hr/></td></tr><tr><td align=\"left\">Intercept</td><td align=\"left\">7.74</td><td align=\"left\">7.10, 8.38</td><td align=\"left\">-3.37</td><td align=\"left\">-3.49, -3.15</td><td align=\"left\">-1.41</td><td align=\"left\">-1.66, -1.19</td><td align=\"left\">1.72</td><td align=\"left\">-0.18, 3.62</td></tr><tr><td align=\"left\">Sex (SEX)</td><td align=\"left\">-0.10</td><td align=\"left\">-0.17, -0.03</td><td align=\"left\">-0.06</td><td align=\"left\">-0.11, -0.01</td><td align=\"left\">---‡</td><td align=\"left\">---</td><td align=\"left\">---</td><td align=\"left\">---</td></tr><tr><td align=\"left\">Watching TV (TV) *</td><td/><td/><td/><td/><td/><td/><td/><td/></tr><tr><td align=\"left\"> Up to 1 h</td><td align=\"left\">0.00</td><td align=\"left\">-0.09, 0.09</td><td align=\"left\">-0.03</td><td align=\"left\">-0.09, 0.03</td><td align=\"left\">---</td><td align=\"left\">---</td><td align=\"left\">---</td><td align=\"left\">---</td></tr><tr><td align=\"left\"> 1–2 h</td><td align=\"left\">0.08</td><td align=\"left\">-0.02, 0.18</td><td align=\"left\">0.05</td><td align=\"left\">-0.01, 0.11</td><td align=\"left\">---</td><td align=\"left\">---</td><td align=\"left\">---</td><td align=\"left\">---</td></tr><tr><td align=\"left\"> More than 2 h</td><td align=\"left\">0.39</td><td align=\"left\">0.20, 0.58</td><td align=\"left\">0.21</td><td align=\"left\">0.12, 0.30</td><td align=\"left\">---</td><td align=\"left\">---</td><td align=\"left\">---</td><td align=\"left\">---</td></tr><tr><td align=\"left\">Breastfeeding (BF)</td><td align=\"left\">---‡</td><td align=\"left\">---</td><td align=\"left\">-0.08</td><td align=\"left\">-0.13, -0.03</td><td align=\"left\">---‡</td><td align=\"left\">---</td><td align=\"left\">---</td><td align=\"left\">---</td></tr><tr><td align=\"left\">Meal frequency (MF) †</td><td/><td/><td/><td/><td/><td/><td/><td/></tr><tr><td align=\"left\"> 4/day</td><td align=\"left\">-0.01</td><td align=\"left\">-0.13, 0.11</td><td align=\"left\">-0.20</td><td align=\"left\">-0.26, -0.14</td><td align=\"left\">---</td><td align=\"left\">---</td><td align=\"left\">-1.48</td><td align=\"left\">-3.09, 0.13</td></tr><tr><td align=\"left\"> 5 or more/day</td><td align=\"left\">-0.16</td><td align=\"left\">-0.28, -0.04</td><td align=\"left\">-0.26</td><td align=\"left\">-0.32, -0.20</td><td align=\"left\">---</td><td align=\"left\">---</td><td align=\"left\">-1.94</td><td align=\"left\">-3.55, -0.33</td></tr><tr><td align=\"left\">Age (AGE)</td><td align=\"left\">0.02</td><td align=\"left\">0.01, 0.02</td><td align=\"left\">---</td><td align=\"left\">---</td><td align=\"left\">---</td><td align=\"left\">---</td><td align=\"left\">---</td><td align=\"left\">---</td></tr><tr><td align=\"left\">Maternal BMI (MB)</td><td align=\"left\">0.07 §</td><td align=\"left\">0.06, 0.08</td><td align=\"left\">0.02 §</td><td align=\"left\">0.02, 0.02</td><td align=\"left\">---</td><td align=\"left\">---</td><td align=\"left\">---</td><td align=\"left\">---</td></tr><tr><td align=\"left\">Weight gain in first 2 y (WG)</td><td align=\"left\">0.50</td><td align=\"left\">0.47, 0.53</td><td align=\"left\">0.07 §</td><td align=\"left\">0.06, 0.09</td><td align=\"left\">---</td><td align=\"left\">---</td><td align=\"left\">0.22</td><td align=\"left\">0.10, 0.34</td></tr></tbody></table></table-wrap>", "<table-wrap position=\"float\" id=\"T2\"><label>Table 2</label><caption><p>Variables in the models with GLM (linear regression, lognormal regression, gamma regression, regression with Box Cox power transformation), GAMLSS, quantile regression for <italic>τ </italic>= 0.9 (QR 0.9) and for <italic>τ </italic>= 0.97 (QR 0.97) for the School Entry Health Examination Study data in Bavaria, 2001–2002.</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td/><td align=\"left\">GLM</td><td align=\"left\">GAMLSS</td><td align=\"left\">QR 0.9</td><td align=\"left\">QR 0.97</td></tr></thead><tbody><tr><td align=\"left\">Sex (SEX)</td><td align=\"left\">+</td><td align=\"left\">+</td><td align=\"left\">+</td><td align=\"left\">0</td></tr><tr><td align=\"left\">Pregnancy smoking (PS)</td><td align=\"left\">0</td><td align=\"left\">[0]</td><td align=\"left\">+</td><td align=\"left\">0</td></tr><tr><td align=\"left\">Watching TV (TV)</td><td align=\"left\">+</td><td align=\"left\">+</td><td align=\"left\">+</td><td align=\"left\">+</td></tr><tr><td align=\"left\">Breastfeeding (BF)</td><td align=\"left\">0</td><td align=\"left\">(+)</td><td align=\"left\">+</td><td align=\"left\">+</td></tr><tr><td align=\"left\">Meal frequency (MF)</td><td align=\"left\">+</td><td align=\"left\">+</td><td align=\"left\">+</td><td align=\"left\">+</td></tr><tr><td align=\"left\">Parents' graduate (PG)</td><td align=\"left\">0</td><td align=\"left\">[0]</td><td align=\"left\">0</td><td align=\"left\">0</td></tr><tr><td align=\"left\">Age (AGE)</td><td align=\"left\">+</td><td align=\"left\">+</td><td align=\"left\">+</td><td align=\"left\">0</td></tr><tr><td align=\"left\">Maternal BMI (MB)</td><td align=\"left\">+</td><td align=\"left\">+</td><td align=\"left\">+</td><td align=\"left\">+</td></tr><tr><td align=\"left\">Weight gain in first 2 y (WG)</td><td align=\"left\">+</td><td align=\"left\">+</td><td align=\"left\">+</td><td align=\"left\">+</td></tr></tbody></table></table-wrap>", "<table-wrap position=\"float\" id=\"T3\"><label>Table 3</label><caption><p>Estimators and 95% confidence intervals (CI) of the quantile regression models with <italic>τ </italic>= 0.9 (QR 0.9) and <italic>τ </italic>= 0.97 (QR 0.97).</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td/><td align=\"center\" colspan=\"2\">QR 0.9</td><td align=\"center\" colspan=\"2\">QR 0.97</td></tr></thead><tbody><tr><td/><td align=\"center\">Estimator</td><td align=\"center\">95% CI</td><td align=\"center\">Estimator</td><td align=\"center\">95% CI</td></tr><tr><td colspan=\"5\"><hr/></td></tr><tr><td align=\"left\">Intercept</td><td align=\"center\">5.16</td><td align=\"center\">3.26, 7.06</td><td align=\"center\">6.33</td><td align=\"center\">4.61, 8.05</td></tr><tr><td align=\"left\">Sex (SEX)</td><td align=\"center\">-0.25</td><td align=\"center\">-0.47, -0.03</td><td align=\"center\">---</td><td align=\"center\">---</td></tr><tr><td align=\"left\">Pregnancy smoking (PS)</td><td align=\"center\">0.54</td><td align=\"center\">0.11, 0.97</td><td align=\"center\">---</td><td align=\"center\">---</td></tr><tr><td align=\"left\">Watching TV (TV) *</td><td/><td/><td/><td/></tr><tr><td align=\"left\"> Up to 1 h</td><td align=\"center\">-0.03</td><td align=\"center\">-0.27, 0.21</td><td align=\"center\">0.33</td><td align=\"center\">-0.20, 0.86</td></tr><tr><td align=\"left\"> 1–2 h</td><td align=\"center\">0.30</td><td align=\"center\">0.05, 0.55</td><td align=\"center\">0.68</td><td align=\"center\">0.23, 1.13</td></tr><tr><td align=\"left\"> More than 2 h</td><td align=\"center\">1.31</td><td align=\"center\">0.80, 1.82</td><td align=\"center\">2.11</td><td align=\"center\">1.01, 3.21</td></tr><tr><td align=\"left\">Breastfeeding (BF)</td><td align=\"center\">-0.41</td><td align=\"center\">-0.72, -0.10</td><td align=\"center\">-0.63</td><td align=\"center\">-1.00, 0.26</td></tr><tr><td align=\"left\">Meal frequency (MF) †</td><td/><td/><td/><td/></tr><tr><td align=\"left\"> 4/day</td><td align=\"center\">-0.19</td><td align=\"center\">-0.62, 0.24</td><td align=\"center\">-0.88</td><td align=\"center\">-1.53, -0.23</td></tr><tr><td align=\"left\"> 5 or more/day</td><td align=\"center\">-0.44</td><td align=\"center\">-0.01, -0.87</td><td align=\"center\">-1.13</td><td align=\"center\">-1.76, 0.50</td></tr><tr><td align=\"left\">Age (AGE)</td><td align=\"center\">0.03</td><td align=\"center\">0.01, 0.05</td><td align=\"center\">---</td><td align=\"center\">---</td></tr><tr><td align=\"left\">Maternal BMI (MB)</td><td align=\"center\">0.16</td><td align=\"center\">0.13, 0.19</td><td align=\"center\">0.22</td><td align=\"center\">0.16, 0.28</td></tr><tr><td align=\"left\">Weight gain in first 2 y (WG)</td><td align=\"center\">0.73</td><td align=\"center\">0.65, 0.81</td><td align=\"center\">0.87</td><td align=\"center\">0.75, 0.99</td></tr></tbody></table></table-wrap>", "<table-wrap position=\"float\" id=\"T4\"><label>Table 4</label><caption><p>Values for the 90^thand 97^thBMI percentiles (τ) estimated by GAMLSS and quantile regression (QR) in respect to meal frequency (MF), with fixed values for all other covariates.</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td/><td align=\"center\">MF ≤ 3</td><td align=\"center\">MF = 4</td><td align=\"center\">MF ≥ 5</td></tr></thead><tbody><tr><td align=\"left\">GAMLSS, <italic>τ </italic>= 0.9</td><td align=\"center\">17.15</td><td align=\"center\">16.82</td><td align=\"center\">16.62</td></tr><tr><td align=\"left\">QR, <italic>τ </italic>= 0.9</td><td align=\"center\">17.08</td><td align=\"center\">16.89</td><td align=\"center\">16.64</td></tr><tr><td align=\"left\">GAMLSS, <italic>τ </italic>= 0.97</td><td align=\"center\">18.39</td><td align=\"center\">17.96</td><td align=\"center\">17.83</td></tr><tr><td align=\"left\">QR, <italic>τ </italic>= 0.97</td><td align=\"center\">19.35</td><td align=\"center\">18.46</td><td align=\"center\">18.22</td></tr></tbody></table></table-wrap>" ]

[ "<disp-formula><italic>y</italic>_{<italic>i </italic>}= <italic>β</italic>₀+ <italic>β</italic>₁<italic>z</italic>_{<italic>i</italic>1}+ ... + <italic>β</italic>_{<italic>m</italic>}<italic>z</italic>_{<italic>im </italic>}+ <italic>ε</italic>_{<italic>i</italic>}, <italic>ε</italic>_{<italic>i </italic>}~ <italic>N</italic>(0, <italic>σ</italic>²).</disp-formula>", "<disp-formula id=\"bmcM1\"><label>(1)</label><italic>η</italic>_{<italic>i </italic>}= <italic>β</italic>₀+ <italic>β</italic>₁<italic>z</italic>_{<italic>i</italic>1}+ ... + <italic>β</italic>_{<italic>m</italic>}<italic>z</italic>_{<italic>im</italic>}.</disp-formula>", "<disp-formula><mml:math xmlns:mml=\"http://www.w3.org/1998/Math/MathML\" id=\"M1\" name=\"1471-2288-8-59-i1\" overflow=\"scroll\"><mml:semantics><mml:mrow><mml:mi>g</mml:mi><mml:mo stretchy=\"false\">(</mml:mo><mml:mi>λ</mml:mi><mml:mo>,</mml:mo><mml:mi>μ</mml:mi><mml:mo stretchy=\"false\">)</mml:mo><mml:mo>=</mml:mo><mml:mrow><mml:mo>{</mml:mo><mml:mrow><mml:mtable><mml:mtr><mml:mtd><mml:mrow><mml:mo stretchy=\"false\">(</mml:mo><mml:msup><mml:mi>μ</mml:mi><mml:mi>λ</mml:mi></mml:msup><mml:mo>−</mml:mo><mml:mn>1</mml:mn><mml:mo stretchy=\"false\">)</mml:mo><mml:mo>/</mml:mo><mml:mi>λ</mml:mi><mml:mo>,</mml:mo><mml:mtext> if </mml:mtext><mml:mi>λ</mml:mi><mml:mo>≠</mml:mo><mml:mtext>0</mml:mtext></mml:mrow></mml:mtd></mml:mtr><mml:mtr><mml:mtd><mml:mrow><mml:mi>log</mml:mi><mml:mo>⁡</mml:mo><mml:mo stretchy=\"false\">(</mml:mo><mml:mi>μ</mml:mi><mml:mo stretchy=\"false\">)</mml:mo><mml:mo>,</mml:mo><mml:mtext> if </mml:mtext><mml:mi>λ</mml:mi><mml:mo>=</mml:mo><mml:mtext>0</mml:mtext></mml:mrow></mml:mtd></mml:mtr></mml:mtable></mml:mrow></mml:mrow></mml:mrow></mml:semantics></mml:math></disp-formula>", "<disp-formula id=\"bmcM2\"><label>(2)</label><mml:math xmlns:mml=\"http://www.w3.org/1998/Math/MathML\" id=\"M2\" name=\"1471-2288-8-59-i2\" overflow=\"scroll\"><mml:semantics><mml:mrow><mml:mi>G</mml:mi><mml:mi>A</mml:mi><mml:mi>I</mml:mi><mml:mi>C</mml:mi><mml:mo stretchy=\"false\">(</mml:mo><mml:mi>c</mml:mi><mml:mo stretchy=\"false\">)</mml:mo><mml:mo>=</mml:mo><mml:mo>−</mml:mo><mml:mn>2</mml:mn><mml:mi>L</mml:mi><mml:mi>L</mml:mi><mml:mo stretchy=\"false\">(</mml:mo><mml:mover accent=\"true\"><mml:mi>θ</mml:mi><mml:mo>^</mml:mo></mml:mover><mml:mo stretchy=\"false\">)</mml:mo><mml:mo>+</mml:mo><mml:mo stretchy=\"false\">(</mml:mo><mml:mi>c</mml:mi><mml:mo>×</mml:mo><mml:mi>p</mml:mi><mml:mo stretchy=\"false\">)</mml:mo></mml:mrow></mml:semantics></mml:math></disp-formula>", "<disp-formula><mml:math xmlns:mml=\"http://www.w3.org/1998/Math/MathML\" id=\"M3\" name=\"1471-2288-8-59-i3\" overflow=\"scroll\"><mml:semantics><mml:mrow><mml:mi>L</mml:mi><mml:mi>L</mml:mi><mml:mo stretchy=\"false\">(</mml:mo><mml:mover accent=\"true\"><mml:mi>θ</mml:mi><mml:mo>^</mml:mo></mml:mover><mml:mo stretchy=\"false\">)</mml:mo><mml:mo>=</mml:mo><mml:mi>log</mml:mi><mml:mo>⁡</mml:mo><mml:mrow><mml:mo>(</mml:mo><mml:mrow><mml:mi>L</mml:mi><mml:mo stretchy=\"false\">(</mml:mo><mml:msub><mml:mi>y</mml:mi><mml:mn>1</mml:mn></mml:msub><mml:mo>,</mml:mo><mml:mn>...</mml:mn><mml:mo>,</mml:mo><mml:msub><mml:mi>y</mml:mi><mml:mi>n</mml:mi></mml:msub><mml:mo>|</mml:mo><mml:mover accent=\"true\"><mml:mi>θ</mml:mi><mml:mo>^</mml:mo></mml:mover><mml:mo stretchy=\"false\">)</mml:mo></mml:mrow><mml:mo>)</mml:mo></mml:mrow><mml:mo>=</mml:mo><mml:mstyle displaystyle=\"true\"><mml:munderover><mml:mo>∑</mml:mo><mml:mrow><mml:mi>i</mml:mi><mml:mo>=</mml:mo><mml:mn>1</mml:mn></mml:mrow><mml:mi>n</mml:mi></mml:munderover><mml:mrow><mml:mi>log</mml:mi><mml:mo>⁡</mml:mo><mml:mi>f</mml:mi><mml:mo stretchy=\"false\">(</mml:mo><mml:msub><mml:mi>y</mml:mi><mml:mi>i</mml:mi></mml:msub><mml:mo>|</mml:mo><mml:mover accent=\"true\"><mml:mi>θ</mml:mi><mml:mo>^</mml:mo></mml:mover><mml:mo stretchy=\"false\">)</mml:mo></mml:mrow></mml:mstyle></mml:mrow></mml:semantics></mml:math></disp-formula>", "<inline-formula><mml:math xmlns:mml=\"http://www.w3.org/1998/Math/MathML\" id=\"M4\" name=\"1471-2288-8-59-i4\" overflow=\"scroll\"><mml:semantics><mml:mover accent=\"true\"><mml:mi>θ</mml:mi><mml:mo>^</mml:mo></mml:mover></mml:semantics></mml:math></inline-formula>", "<disp-formula><mml:math xmlns:mml=\"http://www.w3.org/1998/Math/MathML\" id=\"M5\" name=\"1471-2288-8-59-i5\" overflow=\"scroll\"><mml:semantics><mml:mrow><mml:mtable columnalign=\"left\"><mml:mtr columnalign=\"left\"><mml:mtd columnalign=\"left\"><mml:mrow><mml:msub><mml:mi>g</mml:mi><mml:mn>1</mml:mn></mml:msub><mml:mo stretchy=\"false\">(</mml:mo><mml:mi>μ</mml:mi><mml:mo stretchy=\"false\">)</mml:mo><mml:mo>=</mml:mo><mml:mi>μ</mml:mi><mml:mo>=</mml:mo><mml:msub><mml:mi>η</mml:mi><mml:mn>1</mml:mn></mml:msub><mml:mo>=</mml:mo><mml:mstyle displaystyle=\"true\"><mml:munderover><mml:mo>∑</mml:mo><mml:mrow><mml:mi>j</mml:mi><mml:mo>=</mml:mo><mml:mn>1</mml:mn></mml:mrow><mml:mrow><mml:msub><mml:mi>J</mml:mi><mml:mn>1</mml:mn></mml:msub></mml:mrow></mml:munderover><mml:mrow><mml:msub><mml:mi>f</mml:mi><mml:mrow><mml:mi>j</mml:mi><mml:mn>1</mml:mn></mml:mrow></mml:msub><mml:mo stretchy=\"false\">(</mml:mo><mml:msub><mml:mi>x</mml:mi><mml:mrow><mml:mi>j</mml:mi><mml:mn>1</mml:mn></mml:mrow></mml:msub><mml:mo stretchy=\"false\">)</mml:mo><mml:mo>+</mml:mo><mml:msub><mml:mi>z</mml:mi><mml:mn>1</mml:mn></mml:msub><mml:mo>'</mml:mo><mml:msub><mml:mi>β</mml:mi><mml:mn>1</mml:mn></mml:msub></mml:mrow></mml:mstyle></mml:mrow></mml:mtd></mml:mtr><mml:mtr columnalign=\"left\"><mml:mtd columnalign=\"left\"><mml:mrow><mml:msub><mml:mi>g</mml:mi><mml:mn>2</mml:mn></mml:msub><mml:mo stretchy=\"false\">(</mml:mo><mml:mi>σ</mml:mi><mml:mo stretchy=\"false\">)</mml:mo><mml:mo>=</mml:mo><mml:mi>log</mml:mi><mml:mo>⁡</mml:mo><mml:mo stretchy=\"false\">(</mml:mo><mml:mi>σ</mml:mi><mml:mo stretchy=\"false\">)</mml:mo><mml:mo>=</mml:mo><mml:msub><mml:mi>η</mml:mi><mml:mn>2</mml:mn></mml:msub><mml:mo>=</mml:mo><mml:mstyle displaystyle=\"true\"><mml:munderover><mml:mo>∑</mml:mo><mml:mrow><mml:mi>j</mml:mi><mml:mo>=</mml:mo><mml:mn>1</mml:mn></mml:mrow><mml:mrow><mml:msub><mml:mi>J</mml:mi><mml:mn>2</mml:mn></mml:msub></mml:mrow></mml:munderover><mml:mrow><mml:msub><mml:mi>f</mml:mi><mml:mrow><mml:mi>j</mml:mi><mml:mn>2</mml:mn></mml:mrow></mml:msub><mml:mo stretchy=\"false\">(</mml:mo><mml:msub><mml:mi>x</mml:mi><mml:mrow><mml:mi>j</mml:mi><mml:mn>2</mml:mn></mml:mrow></mml:msub><mml:mo stretchy=\"false\">)</mml:mo><mml:mo>+</mml:mo><mml:msub><mml:mi>z</mml:mi><mml:mn>2</mml:mn></mml:msub><mml:mo>'</mml:mo><mml:msub><mml:mi>β</mml:mi><mml:mn>2</mml:mn></mml:msub></mml:mrow></mml:mstyle></mml:mrow></mml:mtd></mml:mtr><mml:mtr columnalign=\"left\"><mml:mtd columnalign=\"left\"><mml:mrow><mml:msub><mml:mi>g</mml:mi><mml:mn>3</mml:mn></mml:msub><mml:mo stretchy=\"false\">(</mml:mo><mml:mi>ν</mml:mi><mml:mo stretchy=\"false\">)</mml:mo><mml:mo>=</mml:mo><mml:mi>ν</mml:mi><mml:mo>=</mml:mo><mml:msub><mml:mi>η</mml:mi><mml:mn>3</mml:mn></mml:msub><mml:mo>=</mml:mo><mml:mstyle displaystyle=\"true\"><mml:munderover><mml:mo>∑</mml:mo><mml:mrow><mml:mi>j</mml:mi><mml:mo>=</mml:mo><mml:mn>1</mml:mn></mml:mrow><mml:mrow><mml:msub><mml:mi>J</mml:mi><mml:mn>3</mml:mn></mml:msub></mml:mrow></mml:munderover><mml:mrow><mml:msub><mml:mi>f</mml:mi><mml:mrow><mml:mi>j</mml:mi><mml:mn>3</mml:mn></mml:mrow></mml:msub><mml:mo stretchy=\"false\">(</mml:mo><mml:msub><mml:mi>x</mml:mi><mml:mrow><mml:mi>j</mml:mi><mml:mn>3</mml:mn></mml:mrow></mml:msub><mml:mo stretchy=\"false\">)</mml:mo><mml:mo>+</mml:mo><mml:msub><mml:mi>z</mml:mi><mml:mn>3</mml:mn></mml:msub><mml:mo>'</mml:mo><mml:msub><mml:mi>β</mml:mi><mml:mn>3</mml:mn></mml:msub></mml:mrow></mml:mstyle></mml:mrow></mml:mtd></mml:mtr><mml:mtr columnalign=\"left\"><mml:mtd columnalign=\"left\"><mml:mrow><mml:msub><mml:mi>g</mml:mi><mml:mn>4</mml:mn></mml:msub><mml:mo stretchy=\"false\">(</mml:mo><mml:mi>ς</mml:mi><mml:mo stretchy=\"false\">)</mml:mo><mml:mo>=</mml:mo><mml:mi>log</mml:mi><mml:mo>⁡</mml:mo><mml:mo stretchy=\"false\">(</mml:mo><mml:mi>ς</mml:mi><mml:mo stretchy=\"false\">)</mml:mo><mml:mo>=</mml:mo><mml:msub><mml:mi>η</mml:mi><mml:mn>4</mml:mn></mml:msub><mml:mo>=</mml:mo><mml:mstyle displaystyle=\"true\"><mml:munderover><mml:mo>∑</mml:mo><mml:mrow><mml:mi>j</mml:mi><mml:mo>=</mml:mo><mml:mn>1</mml:mn></mml:mrow><mml:mrow><mml:msub><mml:mi>J</mml:mi><mml:mn>4</mml:mn></mml:msub></mml:mrow></mml:munderover><mml:mrow><mml:msub><mml:mi>f</mml:mi><mml:mrow><mml:mi>j</mml:mi><mml:mn>4</mml:mn></mml:mrow></mml:msub><mml:mo stretchy=\"false\">(</mml:mo><mml:msub><mml:mi>x</mml:mi><mml:mrow><mml:mi>j</mml:mi><mml:mn>4</mml:mn></mml:mrow></mml:msub><mml:mo stretchy=\"false\">)</mml:mo><mml:mo>+</mml:mo><mml:msub><mml:mi>z</mml:mi><mml:mn>4</mml:mn></mml:msub><mml:mo>'</mml:mo><mml:msub><mml:mi>β</mml:mi><mml:mn>4</mml:mn></mml:msub></mml:mrow></mml:mstyle></mml:mrow></mml:mtd></mml:mtr></mml:mtable></mml:mrow></mml:semantics></mml:math></disp-formula>", "<disp-formula><mml:math xmlns:mml=\"http://www.w3.org/1998/Math/MathML\" id=\"M6\" name=\"1471-2288-8-59-i6\" overflow=\"scroll\"><mml:semantics><mml:mrow><mml:mi>z</mml:mi><mml:mo>=</mml:mo><mml:mrow><mml:mo>{</mml:mo><mml:mrow><mml:mtable><mml:mtr><mml:mtd><mml:mrow><mml:mn>1</mml:mn><mml:mo>/</mml:mo><mml:mi>σ</mml:mi><mml:mi>ν</mml:mi><mml:mrow><mml:mo>(</mml:mo><mml:mrow><mml:msup><mml:mrow><mml:mo stretchy=\"false\">(</mml:mo><mml:mi>y</mml:mi><mml:mo>/</mml:mo><mml:mi>μ</mml:mi><mml:mo stretchy=\"false\">)</mml:mo></mml:mrow><mml:mi>ν</mml:mi></mml:msup><mml:mo>−</mml:mo><mml:mn>1</mml:mn></mml:mrow><mml:mo>)</mml:mo></mml:mrow><mml:mo>,</mml:mo><mml:mtext> </mml:mtext><mml:mi>ν</mml:mi><mml:mo>≠</mml:mo><mml:mn>0</mml:mn></mml:mrow></mml:mtd></mml:mtr><mml:mtr><mml:mtd><mml:mrow><mml:msup><mml:mi>σ</mml:mi><mml:mrow><mml:mo>−</mml:mo><mml:mn>1</mml:mn></mml:mrow></mml:msup><mml:mi>log</mml:mi><mml:mo>⁡</mml:mo><mml:mo stretchy=\"false\">(</mml:mo><mml:mi>y</mml:mi><mml:mo>/</mml:mo><mml:mi>μ</mml:mi><mml:mo stretchy=\"false\">)</mml:mo><mml:mo>,</mml:mo><mml:mtext> </mml:mtext><mml:mi>ν</mml:mi><mml:mo>=</mml:mo><mml:mtext>0</mml:mtext></mml:mrow></mml:mtd></mml:mtr></mml:mtable></mml:mrow></mml:mrow><mml:mo>,</mml:mo></mml:mrow></mml:semantics></mml:math></disp-formula>", "<disp-formula id=\"bmcM3\"><label>(3)</label><mml:math xmlns:mml=\"http://www.w3.org/1998/Math/MathML\" id=\"M7\" name=\"1471-2288-8-59-i7\" overflow=\"scroll\"><mml:semantics><mml:mrow><mml:munder><mml:mrow><mml:mi>min</mml:mi><mml:mo>⁡</mml:mo></mml:mrow><mml:mi>θ</mml:mi></mml:munder><mml:mstyle displaystyle=\"true\"><mml:munderover><mml:mo>∑</mml:mo><mml:mrow><mml:mi>i</mml:mi><mml:mo>=</mml:mo><mml:mn>1</mml:mn></mml:mrow><mml:mi>n</mml:mi></mml:munderover><mml:mrow><mml:msub><mml:mi>ρ</mml:mi><mml:mi>τ</mml:mi></mml:msub><mml:mo stretchy=\"false\">(</mml:mo><mml:msub><mml:mi>y</mml:mi><mml:mi>i</mml:mi></mml:msub><mml:mo>−</mml:mo><mml:mi>η</mml:mi><mml:mo stretchy=\"false\">)</mml:mo></mml:mrow></mml:mstyle></mml:mrow></mml:semantics></mml:math></disp-formula>", "<disp-formula><mml:math xmlns:mml=\"http://www.w3.org/1998/Math/MathML\" id=\"M8\" name=\"1471-2288-8-59-i8\" overflow=\"scroll\"><mml:semantics><mml:mrow><mml:msub><mml:mi>ρ</mml:mi><mml:mi>τ</mml:mi></mml:msub><mml:mo stretchy=\"false\">(</mml:mo><mml:mi>u</mml:mi><mml:mo stretchy=\"false\">)</mml:mo><mml:mo>=</mml:mo><mml:mi>u</mml:mi><mml:mrow><mml:mo>(</mml:mo><mml:mrow><mml:mi>τ</mml:mi><mml:mo>−</mml:mo><mml:mi>I</mml:mi><mml:mo stretchy=\"false\">(</mml:mo><mml:mi>u</mml:mi><mml:mo>&lt;</mml:mo><mml:mn>0</mml:mn><mml:mo stretchy=\"false\">)</mml:mo></mml:mrow><mml:mo>)</mml:mo></mml:mrow><mml:mo>=</mml:mo><mml:mrow><mml:mo>{</mml:mo><mml:mrow><mml:mtable><mml:mtr><mml:mtd><mml:mrow><mml:mi>τ</mml:mi><mml:mo>×</mml:mo><mml:mi>u</mml:mi><mml:mo>,</mml:mo><mml:mtext> </mml:mtext><mml:mi>u</mml:mi><mml:mo>≥</mml:mo><mml:mn>0</mml:mn></mml:mrow></mml:mtd></mml:mtr><mml:mtr><mml:mtd><mml:mrow><mml:mo stretchy=\"false\">(</mml:mo><mml:mi>τ</mml:mi><mml:mo>−</mml:mo><mml:mn>1</mml:mn><mml:mo stretchy=\"false\">)</mml:mo><mml:mo>×</mml:mo><mml:mi>u</mml:mi><mml:mo>,</mml:mo><mml:mtext> </mml:mtext><mml:mi>u</mml:mi><mml:mo>&lt;</mml:mo><mml:mn>0</mml:mn></mml:mrow></mml:mtd></mml:mtr></mml:mtable></mml:mrow></mml:mrow></mml:mrow></mml:semantics></mml:math></disp-formula>", "<disp-formula><italic>η </italic>= <italic>β</italic>₀+ <italic>β</italic>₁<italic>SEX </italic>+ <italic>β</italic>₂<italic>TV </italic>+ <italic>β</italic>₃<italic>MF </italic>+ <italic>β</italic>₄<italic>MF </italic>+ <italic>β</italic>₄<italic>AGE </italic>+ <italic>β</italic>₅<italic>MB </italic>+ <italic>β</italic>₆<italic>WG</italic></disp-formula>", "<inline-formula><mml:math xmlns:mml=\"http://www.w3.org/1998/Math/MathML\" id=\"M9\" name=\"1471-2288-8-59-i9\" overflow=\"scroll\"><mml:semantics><mml:mover accent=\"true\"><mml:mi>μ</mml:mi><mml:mo>^</mml:mo></mml:mover></mml:semantics></mml:math></inline-formula>", "<inline-formula><mml:math xmlns:mml=\"http://www.w3.org/1998/Math/MathML\" id=\"M10\" name=\"1471-2288-8-59-i10\" overflow=\"scroll\"><mml:semantics><mml:mover accent=\"true\"><mml:mi>σ</mml:mi><mml:mo>^</mml:mo></mml:mover></mml:semantics></mml:math></inline-formula>", "<inline-formula><mml:math xmlns:mml=\"http://www.w3.org/1998/Math/MathML\" id=\"M11\" name=\"1471-2288-8-59-i11\" overflow=\"scroll\"><mml:semantics><mml:mover accent=\"true\"><mml:mi>ν</mml:mi><mml:mo>^</mml:mo></mml:mover></mml:semantics></mml:math></inline-formula>", "<inline-formula><mml:math xmlns:mml=\"http://www.w3.org/1998/Math/MathML\" id=\"M12\" name=\"1471-2288-8-59-i12\" overflow=\"scroll\"><mml:semantics><mml:mover accent=\"true\"><mml:mi>ς</mml:mi><mml:mo>^</mml:mo></mml:mover></mml:semantics></mml:math></inline-formula>", "<disp-formula><italic>η</italic>₁= <italic>μ </italic>= <italic>β</italic>₀₁+ <italic>β</italic>₁₁<italic>TV</italic></disp-formula>", "<disp-formula><italic>η</italic>₂= log(<italic>σ</italic>) = <italic>β</italic>₀₂+ <italic>β</italic>₁₂<italic>TV</italic></disp-formula>", "<disp-formula><italic>η</italic>₃= <italic>ν </italic>= <italic>β</italic>₀₃</disp-formula>", "<disp-formula><italic>η</italic>₄= log(<italic>ζ</italic>) = <italic>β</italic>₀₄</disp-formula>", "<disp-formula><italic>η </italic>= <italic>β</italic>₀+ <italic>β</italic>₁<italic>SEX </italic>+ <italic>β</italic>₂<italic>PS </italic>+ <italic>β</italic>₃<italic>TV </italic>+ <italic>β</italic>₄<italic>BF </italic>+ <italic>β</italic>₅<italic>MF </italic>+ <italic>β</italic>₆<italic>AGE </italic>+ <italic>β</italic>₇<italic>MB </italic>+ <italic>β</italic>₈<italic>WG</italic></disp-formula>" ]

[]

[]

[]

[]

[]

[ "<table-wrap-foot><p>* \"never\" as reference</p><p>† \"1–3/day\" as reference</p><p>‡ Parameter only significant in the respective univariate model</p><p>§ Splines used for parameter estimation</p></table-wrap-foot>", "<table-wrap-foot><p>\"+\" denoting significant variables, \"0\" non-significant variables and, in case of GAMLSS, \"(+)\" variables only significant for the <italic>σ </italic>term and \" [0]\" variables only significant in the univariate models.5</p></table-wrap-foot>", "<table-wrap-foot><p>* \"never\" as reference</p><p>† \"1–3/day\" as reference</p></table-wrap-foot>" ]

[ "<graphic xlink:href=\"1471-2288-8-59-1\"/>", "<graphic xlink:href=\"1471-2288-8-59-2\"/>" ]

[]

{ "acronym": [], "definition": [] }

[ "<title>Introduction</title>", "<p>In previous research, we validated a subset of items from the ACTG adherence battery as prognostic of virologic suppression at 6 months and moderately correlated with adherence estimates from the Medication Event Monitoring System (MEMS) [##REF##12015870##1##]. The objective of the current study was to validate the longitudinal use of the Owen Clinic adherence index in analyses of time to initial virologic suppression and maintenance of suppression.</p>" ]

[ "<title>Methods</title>", "<p>A retrospective observational cohort study was conducted including all HIV-infected adults under care at the UCSD Owen Clinic between January 2003 and June 2006. Patients were included in the analyses reported here if they: (1) had at least one self report medication adherence score recorded; (2) either initiated antiretroviral therapy for the first time or began a new regimen during the study period; (3) had a plasma viral load ≥ 400 copies/ml prior to initiation of the index regimen; (4) had at least one post baseline plasma viral load; and (5) remained on the index regimen for at least 30 days. Only time on the first regimen during the study period (index regimen) is included in reported analyses. During the study period, patients on antiretroviral therapy were asked to complete, prior to meeting with their medical provider, a computer-assisted four item antiretroviral medication adherence survey [##UREF##4##20##] (Figure ##FIG##3##1##) at every primary care visit. The adherence assessment takes 2–3 minutes to complete and is overseen by the medical assistant who is also recording vital signs. Clinicians review adherence scores and are expected to document adherence counseling in the clinic electronic medical record if scores indicate adherence problems. The adherence items are a subset of the AIDS Clinical Trials Group (ACTG) adherence battery [##REF##10928201##2##]. Items 1 and 2 query the number of missed doses of each antiretroviral medication over each of the preceding four days. The number of missed doses for each drug is summed across all four antecedent days. The sum scores of the two drugs with the highest number of missed doses (designated items 1 and 2) are included in the index score. Item 3 asks \"During the past 4 days, on how many days have you missed <underline>all</underline> your pills?\" (response options (numeric code): no days (0), one day (1), two days (2), three days (3), all four days (4)). Item 4 inquires \"How closely did you follow your specific schedule over the last 4 days?\" (response options (numeric code): never (4), some of the time (3), about half the time (2), most of the time (1), all of the time (0)). Item 5 deals with weekend adherence behavior asking \"Did you skip any of the HIV medications last weekend – last Saturday <underline>or</underline> Sunday?\" (response options (numeric code): no (0), yes (1)). The index score is the sum of responses to the four items with a possible range of 0 (best adherence) to 25 (poorest adherence) if each component of the regimen was dosed twice daily.</p>", "<p>Two outcome measures were operationally defined as: (1) time to first virologic suppression defined as HIV plasma viral load (pVL) ≤ 400 copies/ml after regimen initiation; and (2) maintenance of virologic suppression (pVL ≤ 400 copies/ml). Follow up time for each patient began with the date of initiation of the index antiretroviral regimen and ended with the earliest of the following events: (1) change or discontinuation of the index regimen; (2) last clinic visit date; or (3) end of the study period. Time to first virologic suppression on the index regimen was examined using extended Cox models incorporating time-updated adherence scores. It was confirmed that the proportional hazards assumption was met for all covariates included in the Cox models using log(t) by covariates interactions [##UREF##5##21##]. Maintenance of virologic suppression was evaluated in logit models using population-averaged generalized estimating equations (GEE) with time varying covariates [##UREF##6##22##,##UREF##7##23##]. GEE are a family of methods suitable for the analysis of the longitudinal relationship between a continuous or dichotomous outcome variable and both time-dependent and time independent covariates. The within subject dependency of observations is handled by assuming a working correlation structure for the repeated measurements of the outcome variable [##REF##15469034##24##]. The analysis for the maintenance of virologic suppression analysis included only those patients who achieved an initial pVL ≤ 400 copies/ml and their follow up began on the date of initial virologic suppression. The primary independent variable was time-updated adherence score. Because adherence scores were highly skewed toward higher scores (reflecting poorer adherence), adherence scores were first fit using univariate regression splines to examine the functional relationship between adherence score and the outcome measures [##UREF##2##16##,##UREF##3##17##]. Spline techniques are a family of methods for determining the functional form of the relationship between a continuous predictor variable (e.g. adherence score) and an outcome variable [##REF##18058845##25##]. After determining that the functional relationships were approximately monotonic, eight binary cut points on adherence score were examined in ascending order (e.g. ≥ 1/&lt; 1, ≥ 2/&lt; 2, ≥ 3/&lt; 3) until a threshold demonstrating statistical significance in adjusted models was found (lowest detectable cut point). Because multiple ascending potential cut points were examined, tests of significance for adherence score were adjusted using the Bonferroni method to maintain a overall type I error rate of 0.05 [##UREF##2##16##,##REF##10623917##26##]. Thus the critical p-value for each cut point was 0.05/8 = 0.00625. Examined covariates included: age, sex, race/ethnicity, HIV transmission risk factor, treatment experience (naïve or experienced at time of index regimen initiation), regimen type (number and type of antiretroviral drug classes in the regimen), and both CD4 and pVL measured at the closest time prior to initiation of the index regimen.</p>", "<p>Because HIV plasma viral load and adherence score were not always measured on the same dates, records with missing values for adherence score after the first adherence measurement date were imputed using the last observation carried forward (LOCF) principle. Because the first adherence measurement date usually occurred after the regimen start date, records with missing early adherence scores were backfilled to the regimen start date using the score of the first adherence measurement. Adherence scores were carried forward and backfilled no more than 90 days from the temporally closest adherence measurement date. Viral load data were not carried forward.</p>", "<p>Statistical analyses were performed using Stata 10.0 (Stata Corporation, College Station, TX). This research was approved by the University of California San Diego Human Subjects Committee (Project No. 040394)</p>" ]

[ "<title>Results</title>", "<p>Study eligibility criteria were met by 278 patients whose baseline characteristics are presented in Table ##TAB##0##1##. Participants were predominantly male (88%), middle aged (median 39 years), men having sex with men (MSM) (64%), white (47%), and antiretroviral therapy treatment naive (60%). The median absolute CD4+ lymphocyte count and log₁₀transformed HIV plasma viral load were 173 and 5.0, respectively. Index antiretroviral regimens were distributed as follows: ≥ 2 nucleoside reverse transcriptase inhibitors (NRTIs) + 1 boosted protease inhibitor (PI/r) 73%, ≥ 2 NRTIs + 1 non-nucleoside reverse transcriptase inhibitor (NNRTI) 23%, and other regimens 4%. Enfuvirtide was included as part of the index regimen in only two patients. Median [IQR] days on the index regimen was 286 [115–566] overall. According to prior antiretroviral experience, the median [IQR] days on therapy was 285 [116–566] for treatment naïve patients and 286 [93–562] for treatment experienced patients. 217 patients (78%) achieved an undetectable pVL at median 63 days. 8.3% (18/217) of patients experienced viral rebound (pVL &gt; 400) after initial suppression. The median number of per-patient administrations of the adherence instrument was 4, varying from 1 to 27 administrations. Adherence scores varied from 0 – 25 (mean 1.06, median 0).</p>", "<p>Of the 1155 records in the final analysis dataset representing the longitudinal histories of 278 patients, HIV viral load and adherence were measured on the same date in 556 (48%) records. Of the 1155 records, 599 (52%) represented missing adherence scores at dates of viral load measurement. Of the 599 missing adherence scores, 426 were imputed using the last observation carried forward approach (LOCF) and 173 were imputed by backfilling values. Even though these missing adherence scores technically represent missing values at the time the viral load measures were taken, they conceptually represent values that were obtained at a different time point than the viral load measures. These instances typically represent patients for whom blood is drawn either before of after a clinic visit at which adherence assessment was conducted. The median (IQR) time between the regimen start date and date of the first recorded adherence score was 21 (13–60) days.</p>", "<title>Time to First Viral Suppression Analysis</title>", "<p>Because the distribution of adherence scores was highly skewed (Figure ##FIG##0##2##) we modeled adherence scores using binary indicator variables. In addition to adherence categories, the following potential covariates were examined in separate unadjusted Cox regression models: sex, race/ethnicity, HIV transmission risk factor, age, baseline CD4+ lymphocyte category (0–49, 50–199, ≥ 200), baseline log₁₀HIV plasma viral load, prior antiretroviral treatment experience (naïve, experienced), index regimen type. Of these potential covariates, baseline HIV viral load and race were significantly (p &lt; 0.05) associated with time to viral suppression. Table ##TAB##1##2## presents unadjusted and adjusted analyses of the effect of time updated adherence scores on time to viral suppression. Adjusted hazard ratios (HR) less than 1 are interpretable as indicating longer time to achieving viral suppression relative to the reference category. As anticipated, treatment experienced patients and those with higher baseline viral loads had longer times until achieving viral suppression. Race was not independently associated with the outcome in a model controlling for these two factors and adherence, and was therefore omitted from the final model. Controlling for the remaining two covariates (prior treatment status and baseline HIV viral load), having a time-updated adherence score of five or more (the lowest detectable cut point after Bonferroni correction of overall Type I error rate) was significantly predictive of longer time to achieve viral suppression. There were no 2-way statistical interactions (p &gt; 0.10) between adherence score and either baseline viral load or prior treatment experience. The functional relationship between covariate-adjusted adherence sum score modeled as a regression spline and the log (HR)+residual is presented in Figure ##FIG##1##3##.</p>", "<title>Maintenance of Viral Suppression Analysis</title>", "<p>Table ##TAB##2##3## presents the results of unadjusted and adjusted effects of time-updated adherence scores on maintenance of viral suppression in population averaged GEE logit regression models. The table reports crude and adjusted odds ratios of final models. The same potential covariates were examined as those reported above for the time to initial suppression analysis. With the exception of the time-updated adherence scores, none of the examined covariates were significantly associated with maintenance of viral suppression in unadjusted analysis. Prior treatment experience and baseline plasma viral load were included in the adjusted model to maintain comparability with the time to initial viral suppression analysis (Table ##TAB##1##2##). In both unadjusted and adjusted models, the lowest detectable cut point on adherence score was the same as that observed in the time to initial viral suppression analysis (≥ 5/&lt; 5). The functional relationship between covariate-adjusted adherence sum score modeled as a regression spline and the partial predictor of viral suppression is presented in Figure ##FIG##2##4##.</p>" ]

[ "<title>Discussion</title>", "<p>In the developmental phase of adherence measurement in our clinic, we constructed a 5-item instrument whose individual items were selected from the 51-item ACTG adherence battery [##REF##10928201##2##] on the basis of factor structure and internal consistency reliability. In the manuscript presenting this developmental work, we showed that responses on the 5-item adherence index, administered on one occasion 30 days after initiating a new antiretroviral regimen, were moderately correlated (Spearman rho 0.40 – 0.48) with measures of electronic drug monitoring (EDM) and were predictive of HIV viral load responses at 3 and 6 months after start of treatment in models controlling for baseline viral load and prior antiretroviral experience. We also showed that a cut point of 5 or more on the index distinguished those with viral load suppression (≤ 400 copies/ml) at 3 and 6 months from those failing to suppress at the same time points [##REF##12015870##1##]. The currently reported analyses were conducted to evaluate whether the same 5-item index, when administered repetitively under longitudinal follow up, predicted initial viral suppression and maintenance of suppression while patients continued the index regimen. We found, conditional upon the study eligibility criteria and analytic methods, that the self-report adherence index scores were predictive of both outcomes in models controlling for prior antiretroviral treatment experience and baseline plasma viral load. For the time to initial viral suppression outcome, adherence scores ≥ 5 were associated with an approximately 60% reduced hazard of achieving a plasma viral load ≤ 400 copies/ml. For the maintenance of viral suppression outcome, adherence scores ≥ 5 predicted an approximately 80% lower chance of maintaining viral suppression relative to scores less than 5.</p>", "<p>These findings are not directly comparable to the effects demonstrated in our earlier study for several reasons including: (1) period effects (1998 – 1999 vs. 2003 – 2006) associated with changes in potency and simplicity of antiretroviral regimens; (2) differences in prior treatment experience (22% vs. 60% antiretroviral naïve comparing the earlier to the current study); (3) conditions of adherence measurement (written completion [earlier study] vs. computer assisted [current study]); and (4) differences in analytic approach (outcomes analyzed cross sectionally at fixed time points [earlier study] vs. longitudinally in continuous time [current study]). Nonetheless, the current results contribute to the predictive validation of the instrument as it has been used in routine clinical care of patients on antiretroviral therapy.</p>", "<p>In a recent review of the status of HIV adherence measurement, Chesney presented a conceptual model of adherence assessment and intervention, distinguishing research from clinical applications, and resource-rich from resource-poor settings. In discussing the \"elusive gold standard\" of adherence measurement, she emphasized that \"efforts should continue to develop a portfolio of different valid and reliable self-report measures with varying strengths and weaknesses that can be optimally applied, depending on the situation [##REF##17133199##3##].\" In that spirit, we discuss a number of challenges that emerged in exploring the relationship between routine longitudinal adherence measurement using the Owen Clinic instrument and viral suppression.</p>", "<p>First, adherence score distributions in the current (Figure ##FIG##0##2##) and previous study were highly skewed, with most observations clustered in a range reflecting good adherence and the remainder of observations distributed in the long tail of the distribution reflecting poorer adherence. The clustering of observations toward the excellent adherence end of the distribution creates <italic>ceiling effects </italic>[##UREF##0##4##]. Others have noted the same phenomenon for other self report measures [##REF##17577653##5##, ####REF##16804749##6##, ##REF##17133207##7##, ##REF##16783537##8####16783537##8##]. The clustering of scores toward excellent adherence likely represents a mixture of responses from truly adherent patients and from others exhibiting <italic>social desirability bias </italic>[##UREF##1##9##]. Simoni et al have commented on approaches to minimize both ceiling effects and social desirability bias in adherence assessment [##REF##16783535##10##]. Comparison of self report scores to independent and hopefully more objective measures of adherence (e.g. pharmacy refill data, pill counts, EDM) offer an opportunity to assess the effect of social desirability bias. In other contexts, the use of measures designed to measure social desirability as a construct have been used as covariates to explain self reported health behaviors subject to such response bias [##REF##13813058##11##,##REF##12413181##12##]. With regard to ceiling effects not contaminated by social desirability bias, designing items to capture more challenging aspects of adherence behavior, such as timing of doses or dose taking at inconvenient times (e.g. at work, on weekends, or in the presence of persons not knowing the patient's diagnosis), has been recommended to mitigate the strict ceiling commonly observed in self reported adherence. It should be noted, however, that our instrument included three items (Figure ##FIG##3##1##: items 2–4) dealing with such recommended approaches.</p>", "<p>Second, the modeling of adherence score is not straightforward. As constructed its scale of measurement is discrete numerical with a possible range of 0 – 25 with skewness not amenable to a normalizing transformation. Although cut point selection for an underlying numerical measure may introduce bias in effect measurement [##REF##16828672##13##] and may reduce power to detect effects in comparison with use of the numerical measure [##REF##16217841##14##], cut point models are often preferred because of simplicity of data summarization and interpretation. Post hoc cut point selection, as pointed out by the authors of the STARD initiative [##REF##12513067##15##] (Item 9), may not be replicable with other datasets. In our modeling of the effect of adherence score, we employed an approach adapted from Williams et al [##UREF##2##16##], first exploring the functional form of the relationship between adherence score as a numerical measure using smoothing regression splines as implemented by Royston and Sauerbrei in STATA followed by cut point examination adjusted for multiple comparisons [##UREF##3##17##]. Cut points alternative to what we have described as the lowest detectable cut points could be recommended if alternate methodologies of correction for multiple comparisons were employed (e.g. cross validation or split sample approaches, or examination in independent data sets). It is of interest that in our earlier study, a similar cut point on the same instrument (≥ 5/&lt; 5) was felt to be the most discriminating cut point [##REF##12015870##1##]. After examining the regression spline plots for both outcome metrics (Figures ##FIG##1##3## and ##FIG##2##4##) in the current study, we felt that a cut point around 5 identified a region above which a monotonic relationship between adherence score and functions of the outcome metrics was suggested. In clinical care settings, we believe, based on these data, that our clinicians should be alert to clinically significant problems with adherence for scores at or above 5.</p>", "<p>Third, because of the observational nature of the data, measurements of adherence and HIV plasma viral load were not scheduled to occur simultaneously. Typically clinicians order viral loads every 3 – 6 months depending on clinical factors. Adherence in contrast is measured in our clinic at all routine visits. Conceptually, adherence is a construct representing a daily health behavior for which various self-report indicators have been developed and mapped to estimates of percentage adherence over a defined period or, as in the case of the Owen Clinic instrument, given interpretability primarily through demonstrated association with viral suppression. Because of the staggered nature of data accrual in the clinic, decisions must be made regarding how to line up sequential viral load and adherence measures. At a conceptual level, it is a non-trivial question to decide over how long a period an adherence measure based on a limited recall period (4 days in the case of our instrument) can be extrapolated with regard to preceding and future adherence behaviors for which the self-report data represents an imperfect indicator. In our primary analysis, we made the assumption that a given adherence assessment carried forward no longer than 90 days from the antecedent adherence measurement. Whether the observations that are not temporally matched represent truly missing observations is debatable since the very nature of the data accrual process in clinical care did not require temporal matching of adherence and viral load measurement. Because the LOCF principle has been criticized in recent years [##REF##17230434##18##], we explored alternate analyses to evaluate the robustness of our findings. First, to determine if the frequency of adherence measurement was related to adherence scores such that longer intervals between measurements were associated with better or poorer adherence, we calculated rates of adherence measurement per 100 days of follow up. We then divided the adherence measurement rate distribution into quartiles and used analysis of variance to test for equality of mean adherence scores across the quartiles, finding no significant difference (p = 0.89). This provided limited evidence that, in our data set, adherence scores were not systematically related to frequency of measurement, although others have found that missing adherence values were associated with nonadherence [##REF##11352698##19##]. Second, we restructured the data set by grouping follow up time in 6 month intervals, taking the median adherence score for the interval as representative, the last viral load in the interval as the outcome, and repeating the panel regression for longitudinal viral suppression. In a model comparable to that shown in Table ##TAB##2##3## controlling for prior treatment experience and baseline log₁₀-HIV viral load, the adjusted odds ratio for viral suppression was 0.14 (95% CI: 0.06 – 0.33, p &lt; 0.0001) for a 6-month median adherence score greater than 5. Finally, in a third analysis of maintenance of longitudinal viral suppression, mean adherence scores were calculated for the period immediately prior to each viral load measurement, creating a score for each interval between viral load measurements. This operationalization of adherence was then fit in a GEE logit model for maintenance of viral suppression, again controlling for prior treatment experience and baseline log₁₀-HIV viral load. The adjusted adherence odds ratio for maintaining viral suppression for a mean interval adherence score greater than 5 was 0.28 (95% CI: 0.14 – 0.57, p &lt; 0.0001) Therefore, although the adherence effect estimates were model dependent, the direction of effect was consistent and significant across models.</p>" ]

[ "<title>Conclusion</title>", "<p>Despite the limitations of self-report adherence measures, they are likely to remain the most frequent modality of adherence assessment in clinical settings. The brief self-report instrument examined in this study and in an earlier developmental study has been demonstrated to correlate with electronic drug monitoring and to be predictive of viral load responses both when administered at baseline and also when administered in longitudinal follow up of unselected patients in clinical care for HIV infection.</p>" ]

[ "<p>This is an Open Access article distributed under the terms of the Creative Commons Attribution License (<ext-link ext-link-type=\"uri\" xlink:href=\"http://creativecommons.org/licenses/by/2.0\"/>), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.</p>", "<title>Background</title>", "<p>Newer antiretroviral (ARV) agents have improved pharmacokinetics, potency, and tolerability and have enabled the design of regimens with improved virologic outcomes. Successful antiretroviral therapy is dependent on patient adherence. In previous research, we validated a subset of items from the ACTG adherence battery as prognostic of virologic suppression at 6 months and correlated with adherence estimates from the Medication Event Monitoring System (MEMS). The objective of the current study was to validate the longitudinal use of the Owen Clinic adherence index in analyses of time to initial virologic suppression and maintenance of suppression.</p>", "<title>Results</title>", "<p>278 patients (naïve n = 168, experienced n = 110) met inclusion criteria. Median [range] time on the first regimen during the study period was 286 (30 – 1221) days. 217 patients (78%) achieved an undetectable plasma viral load (pVL) at median 63 days. 8.3% (18/217) of patients experienced viral rebound (pVL &gt; 400) after initial suppression. Adherence scores varied from 0 – 25 (mean 1.06, median 0). The lowest detectable adherence score cut point using this instrument was ≥ 5 for both initial suppression and maintenance of suppression. In the final Cox model of time to first undetectable pVL, controlling for prior treatment experience and baseline viral load, the adjusted hazard ratio for time updated adherence score was 0.36_{score ≥ 5}(95% CI: 0.19–0.69) [reference: &lt;5]. In the final generalized estimating equations (GEE) logistic regression model the adjusted odds ratio for time-updated adherence score was 0.17_{score ≥ 5}(0.05–0.66) [reference: &lt;5].</p>", "<title>Conclusion</title>", "<p>A brief, longitudinally administered self report adherence instrument predicted both initial virologic suppression and maintenance of suppression in patients using contemporary ARV regimens. The survey can be used for identification of sub-optimal adherence with subsequent appropriate intervention.</p>" ]

[ "<title>Competing interests</title>", "<p>The authors declare that they have no competing interests.</p>", "<title>Authors' contributions</title>", "<p>WCM designed the study, conducted the final analysis, and prepared the manuscript; EB and EW conducted extensive medical record review and prepared preliminary analysis of the data; CB and BC contributed to design of the study and manuscript preparation; SM advised on statistical analysis and contributed to the manuscript preparation. All authors read and approved the final manuscript.</p>" ]

[ "<title>Acknowledgements</title>", "<p>This work was supported in part by the UCSD Center for AIDS Research (AI 36214) and by the CFAR-Network of Integrated Clinical Systems (AI067039). The funding agencies had no role in the study design; collection, analysis, or interpretation of the data; manuscript preparation; or decision to submit the work for publication.</p>" ]

[ "<fig position=\"float\" id=\"F2\"><label>Figure 2</label><caption><p>Distribution of first adherence scores during the study period (n = 278 patients).</p></caption></fig>", "<fig position=\"float\" id=\"F3\"><label>Figure 3</label><caption><p>Regression spline (95% confidence interval) of adherence score in Cox model of time to viral suppression, adjusted for treatment experience and baseline viral load.</p></caption></fig>", "<fig position=\"float\" id=\"F4\"><label>Figure 4</label><caption><p>Regression spline (95% confidence interval) of adherence score in GEE logit model of maintained viral suppression, adjusted for treatment experience and baseline viral load.</p></caption></fig>", "<fig position=\"float\" id=\"F1\"><label>Figure 1</label><caption><p>Screen shot of adherence instrument.</p></caption></fig>" ]

[ "<table-wrap position=\"float\" id=\"T1\"><label>Table 1</label><caption><p>Patient Characteristics at Study Entry (n = 278)</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td align=\"left\"><bold>Characteristic</bold></td><td/></tr></thead><tbody><tr><td align=\"left\">Sex [n (%)]</td><td/></tr><tr><td align=\"left\">   Female</td><td align=\"center\">33 (12)</td></tr><tr><td align=\"left\">   Male</td><td align=\"center\">245 (88)</td></tr><tr><td align=\"left\">HIV Transmission Risk Factor [n (%)]</td><td/></tr><tr><td align=\"left\">   MSM₁, not IDU₂</td><td align=\"center\">179 (64)</td></tr><tr><td align=\"left\">   Heterosexual contact</td><td align=\"center\">52 (19)</td></tr><tr><td align=\"left\">   IDU</td><td align=\"center\">23 (8)</td></tr><tr><td align=\"left\">   Other/Unknown</td><td align=\"center\">24 (9)</td></tr><tr><td align=\"left\">Race/Ethnicity [n (%)]</td><td/></tr><tr><td align=\"left\">   White</td><td align=\"center\">130 (47)</td></tr><tr><td align=\"left\">   Black</td><td align=\"center\">30 (11)</td></tr><tr><td align=\"left\">   Hispanic</td><td align=\"center\">87 (31)</td></tr><tr><td align=\"left\">   Other/Unknown</td><td align=\"center\">31 (11)</td></tr><tr><td align=\"left\">Age (years)</td><td/></tr><tr><td align=\"left\">   [mean (sd)]</td><td align=\"center\">39.5 (9.2)</td></tr><tr><td align=\"left\">   [median (range)]</td><td align=\"center\">39 (19–77)</td></tr><tr><td align=\"left\">ART₃Treatment Experience [n (%)]</td><td/></tr><tr><td align=\"left\">   Naive</td><td align=\"center\">168 (60)</td></tr><tr><td align=\"left\">   Experienced</td><td align=\"center\">110 (40)</td></tr><tr><td align=\"left\">Baseline absolute CD4</td><td/></tr><tr><td align=\"left\">   [mean (sd)]</td><td align=\"center\">201 (163)</td></tr><tr><td align=\"left\">   [median (range)]</td><td align=\"center\">173 (0–883)</td></tr><tr><td align=\"left\">Baseline log₁₀HIV-1 Plasma Viral Load</td><td/></tr><tr><td align=\"left\">   [mean (sd)]</td><td align=\"center\">4.9 (0.7)</td></tr><tr><td align=\"left\">   [median (range)]</td><td align=\"center\">5.0 (2.7–6.3)</td></tr><tr><td align=\"left\">Days on new regimen</td><td/></tr><tr><td align=\"left\">   [median (range)]</td><td align=\"center\">286 (30–1221)</td></tr><tr><td align=\"left\">Year of study entry [n (%)]</td><td/></tr><tr><td align=\"left\">   2003</td><td align=\"center\">51 (18)</td></tr><tr><td align=\"left\">   2004</td><td align=\"center\">103 (37)</td></tr><tr><td align=\"left\">   2005</td><td align=\"center\">81 (29)</td></tr><tr><td align=\"left\">   2006</td><td align=\"center\">43 (16)</td></tr><tr><td align=\"left\">New Regimen Type₄[n(%)]</td><td/></tr><tr><td align=\"left\">   NNRTI &amp; ≥ 2 NRTIs</td><td align=\"center\">63 (23)</td></tr><tr><td align=\"left\">   PI_b&amp; ≥ 2 NRTIs</td><td align=\"center\">204 (73)</td></tr><tr><td align=\"left\">   NNRTI &amp; PI_b&amp; ≥ 1 NRTI</td><td align=\"center\">8 (3)</td></tr><tr><td align=\"left\">   ≥ 2 NRTI</td><td align=\"center\">3 (1)</td></tr><tr><td align=\"left\"># Adherence Scores per patient</td><td/></tr><tr><td align=\"left\">   [median (range)]</td><td align=\"center\">4 (1–27)</td></tr></tbody></table></table-wrap>", "<table-wrap position=\"float\" id=\"T2\"><label>Table 2</label><caption><p>Unadjusted and adjusted effects of time-updated adherence scores on time to first HIV viral load ≤ 400 copies/ml in Cox regression models (n = 278 patients)</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td/><td align=\"center\" colspan=\"3\">Unadjusted</td><td align=\"center\" colspan=\"3\">Adjusted</td></tr></thead><tbody><tr><td align=\"left\">Predictor</td><td align=\"center\">HR₁</td><td align=\"center\">95% CI</td><td align=\"center\">p-value</td><td align=\"center\">HR</td><td align=\"center\">95% CI</td><td align=\"center\">p-value</td></tr><tr><td colspan=\"7\"><hr/></td></tr><tr><td align=\"left\">Adherence Score</td><td/><td/><td/><td/><td/><td/></tr><tr><td align=\"left\">   &lt; 5</td><td align=\"center\">1.0</td><td/><td/><td align=\"center\">1.0</td><td/><td/></tr><tr><td align=\"left\">   ≥ 5</td><td align=\"center\">0.42</td><td align=\"center\">0.22–0.79</td><td align=\"center\">0.007</td><td align=\"center\">0.36</td><td align=\"center\">0.19–0.69</td><td align=\"center\">0.002</td></tr><tr><td align=\"left\">Antiretroviral Experience</td><td/><td/><td/><td/><td/><td/></tr><tr><td align=\"left\">   Naïve</td><td align=\"center\">1.0</td><td/><td/><td align=\"center\">1.0</td><td/><td/></tr><tr><td align=\"left\">   Experienced</td><td align=\"center\">0.79</td><td align=\"center\">0.60–0.1.05</td><td align=\"center\">0.10</td><td align=\"center\">0.68</td><td align=\"center\">0.50–0.91</td><td align=\"center\">0.01</td></tr><tr><td align=\"left\">Baseline log₁₀HIV viral load</td><td/><td/><td/><td/><td/><td/></tr><tr><td/><td align=\"center\">0.82</td><td align=\"center\">0.68–0.99</td><td align=\"center\">0.04</td><td align=\"center\">0.71</td><td align=\"center\">0.58–0.87</td><td align=\"center\">0.001</td></tr><tr><td align=\"left\">Race</td><td/><td/><td align=\"center\">0.047</td><td/><td/><td/></tr><tr><td align=\"left\">   White</td><td align=\"center\">1.0</td><td/><td/><td/><td/><td/></tr><tr><td align=\"left\">   Black</td><td align=\"center\">1.42</td><td align=\"center\">0.92–2.19</td><td align=\"center\">0.11</td><td align=\"center\">---</td><td align=\"center\">---</td><td align=\"center\">---</td></tr><tr><td align=\"left\">   Hispanic</td><td align=\"center\">1.51</td><td align=\"center\">1.10–2.06</td><td align=\"center\">0.01</td><td/><td/><td/></tr><tr><td align=\"left\">   Unknown/Other</td><td align=\"center\">1.01</td><td align=\"center\">0.63–1.62</td><td align=\"center\">0.98</td><td/><td/><td/></tr></tbody></table></table-wrap>", "<table-wrap position=\"float\" id=\"T3\"><label>Table 3</label><caption><p>Unadjusted and adjusted effects of time-updated adherence scores on maintenance of HIV viral load ≤ 400 copies/ml in generalized estimating equation logit regression models (n = 217 patients achieving initial viral suppression)</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td/><td align=\"center\" colspan=\"3\">Unadjusted</td><td align=\"center\" colspan=\"3\">Adjusted</td></tr></thead><tbody><tr><td align=\"left\">Predictor</td><td align=\"center\">OR₁</td><td align=\"center\">95% CI</td><td align=\"center\">p-value</td><td align=\"center\">OR</td><td align=\"center\">95% CI</td><td align=\"center\">p-value</td></tr><tr><td colspan=\"7\"><hr/></td></tr><tr><td align=\"left\">Adherence Score</td><td/><td/><td/><td/><td/><td/></tr><tr><td align=\"left\">   &lt; 5</td><td align=\"center\">1.0</td><td/><td/><td align=\"center\">1.0</td><td/><td/></tr><tr><td align=\"left\">   ≥ 5</td><td align=\"center\">0.20</td><td align=\"center\">0.05–0.79</td><td align=\"center\">0.02</td><td align=\"center\">0.17</td><td align=\"center\">0.05–0.66</td><td align=\"center\">0.01</td></tr><tr><td align=\"left\">Antiretroviral Experience</td><td/><td/><td/><td/><td/><td/></tr><tr><td align=\"left\">   Naïve</td><td align=\"center\">1.0</td><td/><td/><td align=\"center\">1.0</td><td/><td/></tr><tr><td align=\"left\">   Experienced</td><td align=\"center\">0.78</td><td align=\"center\">0.28–2.24</td><td align=\"center\">0.65</td><td align=\"center\">0.60</td><td align=\"center\">0.21–1.70</td><td align=\"center\">0.34</td></tr><tr><td align=\"left\">Baseline log₁₀HIV viral load</td><td/><td/><td/><td/><td/><td/></tr><tr><td/><td align=\"center\">0.56</td><td align=\"center\">0.26–1.23</td><td align=\"center\">0.15</td><td align=\"center\">0.49</td><td align=\"center\">0.22–1.11</td><td align=\"center\">0.09</td></tr></tbody></table></table-wrap>" ]

[]

[]

[]

[]

[]

[]

[ "<table-wrap-foot><p>1. MSM: men having sex with men.</p><p>2. IDU: injection drug use.</p><p>3. ART: antiretroviral therapy</p><p>4. NNRTI: non-nucleoside reverse transcriptase inhibitor; NRTI: nucleoside/nucleotide reverse transcriptase inhibitor; PI_b: ritonavir boosted protease inhibitor.</p></table-wrap-foot>", "<table-wrap-foot><p>HR: hazard ratio</p></table-wrap-foot>", "<table-wrap-foot><p>OR: odds ratio</p></table-wrap-foot>" ]

[ "<graphic xlink:href=\"1742-6405-5-20-2\"/>", "<graphic xlink:href=\"1742-6405-5-20-3\"/>", "<graphic xlink:href=\"1742-6405-5-20-4\"/>", "<graphic xlink:href=\"1742-6405-5-20-1\"/>" ]

[]

{ "acronym": [], "definition": [] }

[ "<title>Introduction</title>", "<p>Familial history is a risk factor of several chronic diseases of public health significance, including obesity. Therefore, it has been proposed that familial history information could be useful in clinical practice and construction of family pedigrees could provide important data for use in genetic studies [##REF##12568818##1##, ####REF##9756239##2##, ##REF##6519911##3####6519911##3##]. However, little is known about the accuracy of self-reported familial history, particularly self-reported familial history of obesity (FHO).</p>", "<p>At present, the literature mainly provides information on self-reported height and weight. More than two decades ago, Stunkard and Albaum [##REF##7270483##4##] reported that self-reported weights were remarkably accurate across different ages and sexes. Subsequent studies confirmed that self-reported values are valid and provide reliable indicators of measured weight and height [##REF##9756239##2##,##REF##10490794##5##, ####REF##11209581##6##, ##REF##11255492##7##, ##REF##12186665##8##, ##REF##11033979##9##, ##REF##7061685##10##, ##REF##17228047##11####17228047##11##]. Recently, Gorber and colleagues [##REF##17578381##12##] have published a review of the literature determining what empirical evidence exists regarding the agreement between objective (measured) and subjective (reported) measures in assessing height, weight and body mass index (BMI) in observational and experimental studies of adult populations. Overall, this review, including 64 studies, reported an evident trend for an overestimation of height and an underestimation of weight and BMI, in both men and women [##REF##17578381##12##]. Height and weight estimations from a family member (father, mother and siblings) follow the same pattern. Indeed, Reed and Price [##REF##9756239##2##] conducted a study using a cohort of 374 first-degree relatives from 94 Caucasian families, to assess the value of family informant estimates of height and weight. It was shown that respondents systematically overestimated heights (mean = 1.4 cm) and underestimated weights (mean = 4.1 kg) of their family members [##REF##9756239##2##]. Although the accuracy of self-reported height and weight and the accuracy of family member estimates have been studied, there is still no available data on the validity of self-reported measure of FHO. This information could be of great importance for use in genetics studies and other studies such as those designed to understand differences between subjects with and without FHO. Indeed, these studies should benefit from information about the accuracy of a self-reported measure of FHO which has the advantages of practicality and being a low cost method applicable to a large number of individuals. Thus, the aim of this study was to examine the validity of a self-reported measure of FHO. Before doing so, we first compared self-reported and measured weight and height in a cohort of 245 men and 372 women from the greater Quebec City area (study 1). Secondly, we assessed the correlation between weight and height estimations reported by a participant and the values provided by each family member (mother, father and siblings) in a cohort of 78 subjects including 199 family members (study 2). Finally, we examined the validity of a self-reported measure of FHO in the same cohort (study 2).</p>" ]

[ "<title>Methods</title>", "<title>Study 1</title>", "<title>Study population and data collection</title>", "<p>Participants were adults aged between 18 to 55 years. Subjects were recruited in the Quebec City metropolitan area through public advertisements in local newspapers and by electronic messages sent to university and hospital employees. A trained research assistant conducted a telephone interview with people who responded to the advertisement messages. The assistant asked the participants to report their body weight and height. Subjects had to answer to the following questions: <italic>What is your current weight?, What is your current height? </italic>Following the interview, eligible participants were given an appointment within the next 2–3 weeks to come to the laboratory to meet trained research assistants for anthropometric measurements. The beam Scale with height rod graduated in centimetres was used (Detecto, Webb City, USA) to obtain a measure of weight and height of each participant. Weight was measured to the nearest 0.1 kg and height was measured to the nearest 0.5 cm. The scale was calibrated before each examination. BMI was computed as weight in kilograms divided by height in meters squared. Enrolment of the subjects took place between 2004 and 2006. The final study sample consisted of 245 men and 372 women. All subjects gave their written consent to participate into this study which has been approved by the Ethics Committee of the local university Hospital Research Center.</p>", "<title>Statistical analysis</title>", "<p>Paired Student's <italic>t</italic>-test was used to compare the means of self-reported and measured height and weight. Pearson's correlation coefficients were used to examine the association between the self-reported and measured height and weight. All statistical analyses were performed using SAS statistical software, version 8.2 (SAS Institute Inc, Cary, NC) and statistical significance was defined as p &lt; 0.05.</p>", "<title>Study 2</title>", "<title>Study population and data collection</title>", "<p>Participants were adults aged between 18 to 55 years. Subjects were recruited as described in study 1. Enrolment of the subjects took place in 2006. The sample included 78 respondents (52 women and 26 men) and their family members (mother, father, and siblings) (n = 199). Each of the 78 volunteers was given an appointment to come to the laboratory within the next 5–6 days after the initial contact. During their visit, they were first asked to report on a self-administrated questionnaire their own weight and height and to estimate weight and height of each of their family members (mother, father and siblings). Subjects had to answer the following questions: <italic>What is your current weight?, What is your current height?, What is the current weight of your mother, father, and siblings?</italic>, and <italic>What is the current height of your mother, father, and siblings?</italic>. Second, volunteers had to identify whether any of their family members (mother, father and siblings) were obese. If the participant identified at least one obese first-degree relative, a subjective FHO was determined as positive. The subjective FHO was considered negative if no obese first-degree relative was identified. Third, the beam Scale with height rod graduated in centimetres was used (Detecto, Webb City, USA) to obtain a measure of weight and height of each participant. Weight was measured to the nearest 0.1 kg and height was measured to the nearest 0.5 cm. The scale was calibrated before each examination. BMI was computed as weight in kilograms divided by height in meters squared. Each participant was asked to transmit (if needed by mail) an informed consent and a self-administrated questionnaire to each of his/her family members, and to not discuss the measurements or the study with their family members. These family members were asked to complete the informed consent, to report their own weight and height, and to send this information back to the research team. All family members (mother, father, and siblings) were asked the following questions: <italic>What is your current weight?</italic>, and <italic>What is your current height?</italic>. This allowed the calculation of BMI for each family member (mother, father, and siblings) and to define an objective measure of FHO. Thus, objective FHO+ was defined as having at least one obese (BMI ≥ 30 kg/m²) family member and objective FHO- as having no family member with a BMI ≥ 30 kg/m². All subjects gave their written consent to participate into this study which was approved by the Ethics Committee of the local university Hospital Research Center. Complete informations about FHO were obtained for 44 families including 199 family members.</p>", "<title>Statistical analysis</title>", "<p>The student <italic>t</italic>-test was used 1) to compare the means of self-reported and measured height and weight and 2) to compare the means of weight and height estimations reported by the participants to values reported by each family member. Pearson's correlation coefficients were calculated to assess the linear associations between self-reported and measured height and weight. Pearson's correlation coefficients were also used to examine the relation between height and weight estimations reported by the participant and the values provided by each family member. Sensitivity and specificity of FHO were calculated. The sensitivity represents the capacity of the participant to correctly report positive FHO (correctly reported positive FHO/all subjective positive FHO) and the specificity represents the capacity of the participant to correctly report negative FHO (correctly reported negative FHO/all subjective negative FHO). Positive predictive value estimating the proportions of participants who correctly reported positive FHO (correctly reported positive FHO/all objective positive FHO) was also computed together with the negative predictive value which estimates the proportions of participants who correctly reported FHO (correctly reported negative FHO/all objective negative FHO). The likelihood ratio was calculated to estimate the ratio of the true positive rate versus the false positive rate. The ROC-Curve analysis was computed to test the capacity of the self-reported FHO to classify correctly participants. Finally, kappa coefficients were calculated to measure congruence between the subjective FHO and the objective FHO. The following classification suggested by Landis and Koch [##REF##843571##13##] was used: poor-to-fair (kappa &lt; 0.40), moderate (kappa of 0.41 to 0.60), substantial (kappa of 0.61 to 0.80) and excellent (kappa of 0.81 to 1.00). All statistical analyses were performed in SAS statistical software, version 8.2 (SAS Institute Inc, Cary, NC) and statistical significance was defined as p &lt; 0.05.</p>" ]

[ "<title>Results</title>", "<title>Study 1</title>", "<p>The mean (± SD) age of the 617 participants was 37.9 ± 11.3 years. Mean values of self-reported and measured values of weight, height and BMI are presented in Table ##TAB##0##1##. In men, self-reported and measured values were not significantly different (p &gt; 0.05). In women, the mean difference between self-reported and measured weight and BMI were not significantly different. However, women overestimated their height by a mean of 1.2 cm (p &lt; 0.05). Differences observed between self-reported and measured values were similar in men and women (p &gt; 0.05). Strong correlations between self-reported and measured weight, height and BMI were observed (r = 0.99, p &lt; 0.0001; 0.97, p &lt; 0.0001; r = 0.97, p &lt; 0.0001, respectively) (data not shown).</p>", "<title>Study 2</title>", "<p>The mean (± SD) age of the 78 participants was 31.7 ± 10.4 years whereas the mean age of family members was 45.0 ± 15.7 years. On average, each family had 4.5 ± 1.1 members with a minimum of three and a maximum of eight family members. Participants of study 2 were comparable to those of study 1. Despite that participants of study 2 were slightly younger, they had similar weight and BMI (after adjustment for age and sex, data not shown). The mean values of self-reported and measured weight, height, BMI as well as their differences are presented in Table ##TAB##0##1##. No significant differences were observed between self-reported and measured weight and BMI. Correlations between self-reported and measured weight, height and BMI were observed (r = 0.98, p &lt; 0.0001; 0.98, p &lt; 0.0001; r = 0.95, p &lt; 0.0001, respectively) (data not shown).</p>", "<p>To determine the accuracy of estimations provided by the participant about the weight, height, and BMI of their family members, values reported by the study participants were compared to weight, height, and BMI reported by each family member. Mean values of weight, height and BMI are presented in Table ##TAB##1##2##. No significant differences were observed between informant estimates and family member self-reported values. In addition, highly significant correlations were observed for weight, height, and BMI reported by the participants and self-reported values of family members (weight: r = 0.96, height: r = 0.95, BMI: r = 0.91 p &lt; 0.0001) (data not shown).</p>", "<p>For intrinsic and predictive validity, the distribution of individuals according to the objective and subjective measure of FHO are presented in Table ##TAB##2##3##. Sensitivity was 90.5%, whereas specificity was 82.6%. The positive and negative values were respectively 82.6% and 90.5% and the likelihood ratio was 5.2. The area under the ROC-curve was 0.87 indicating that the classification into FHO categories was excellent (not shown). Finally, the validity of the self-reported measure of FHO was verified using the kappa statistic. Considering the 44 participants, kappa analysis indicated that the degree of agreement between the subjective measure of FHO reported by the participants and the objective measure derived from the self-reported BMI of family members was substantial as indicated by the value of 0.72. The degree of agreement was also verified separately for men and women, age groups, and body weight categories. Kappa values of 0.68 and 0.85 were observed respectively for men and women. For age (median split), a substantial degree of agreement was observed in younger individuals (&lt;28 years) (kappa = 0.63) and an excellent degree of agreement (kappa = 0.81) was observed in older individuals (≥ 28 years). With respect to body weight, there was a moderate degree of agreement in heavier individuals (median split, BMI ≥ 23.8 kg/m², kappa = 0.58) and a substantial degree of agreement in individuals having a BMI &lt; 23.8 kg/m²(kappa = 0.71).</p>" ]

[ "<title>Discussion</title>", "<p>Although familial history information is collected in clinical and research studies, little information is available on the validity of a proband's reported familial history. To our knowledge, this is the first study in the field of obesity which investigates the validity of a self-reported measure of FHO. Before doing so, we first verified that individuals were able to adequately report their own weight and height. Thus, the accuracy of self-reported weight and height has been first examined in study 1. Results obtained are concordant with the literature which has demonstrated that self-reported height and weight are highly predictive of measured values [##REF##9756239##2##,##REF##10490794##5##, ####REF##11209581##6##, ##REF##11255492##7##, ##REF##12186665##8##, ##REF##11033979##9####11033979##9##]. In contrast to some previous studies [##REF##9756239##2##,##REF##10490794##5##,##REF##12186665##8##,##REF##11033979##9##,##REF##10715748##14##,##REF##16800200##15##] and a recent review [##REF##17578381##12##], similar means of self-reported and measured weight, height (except for women), and BMI were observed in men and women of study 1. In women, self-reported values of height were overestimated. The overestimation was minor (1.2 cm) and did not significantly affect the mean BMI values (self-reported versus measured values). Similar analyses were performed in a second study sample (study 2) and no significant differences were observed between self-reported and measured values in both men and women.</p>", "<p>Regarding weight and height estimations of family members provided by the participants in study 2, high correlations were observed between estimations reported by the participants and the values reported by each family member. Moreover, no significant difference was observed between the values reported by the participants and the one reported by each family member. The use of self-reported current weight and height for each family member instead of the use of measured values could be considered as a limitation of the present study. However, no major difference in self-reported and measured weight, height with (except among the women of study 1) and BMI were reported in participants from two different samples. Thus, we assume that family members had also truthfully reported their weight and height.</p>", "<p>Although results of the present study suggest that individuals were able to adequately report their own weight and height and those of their family members, it is important to mention that the major aim of the present study was to examine the validity of a self-reported measure of FHO. By comparing, subjective FHO (self-reported by the participant) with objective FHO (defined by the self-reported BMI of family members), we found that objective FHO are close to perfectly reliable. Very good sensitivity values were observed indicating that individuals with positive FHO correctly reported the presence of FHO. The high level of specificity observed indicates that individuals without FHO correctly reported the absence of FHO. The positive and negative predictive values were high, suggesting that the method use in this study to determine the presence or the absence of FHO is reliable. In other words, individuals are able to report the presence or the absence of obesity in their first-degree family members. It is important to note that substantial or excellent agreement was also observed when kappa statistic was calculated according to gender and age. Moreover, substantial and moderate degree of agreement was observed depending of weight status.</p>", "<p>The present study has several limitations. First, since demographic characteristics can have an influence on the degree of reporting error [##REF##17578381##12##] and that bias of estimated heights and weights are influenced by informant characteristics [##REF##9756239##2##], it would be relevant to perform statistical analyses after stratification on the basis of gender, age, education, weight, etc. However, the sample size of the study 2 is relatively small for such a stratification. Other studies with a larger number of families are needed to examine this point. Second, the determination of the objective measure of FHO was available only for families in which all members have completed the questionnaire. It is possible that these families may be more willing than non-respondent families to provide precise weight and height. Third, most of the participants may have been aware that stature and weight would be measured following the self-reported, which could have mitigated misreporting and could minimise errors between self-reported and measured values which usually tend to deviate towards a 'preferred' body size [##REF##8762365##16##]. This bias could explain why self-reported and measured weight, height (except for women in study 1), and BMI were similar in the present study in contrast to other studies [##REF##9756239##2##,##REF##10490794##5##,##REF##12186665##8##,##REF##11033979##9##,##REF##17578381##12##,##REF##10715748##14##,##REF##16800200##15##].</p>", "<p>In conclusion, although the results of the present study need to be replicated in other cohorts or populations with larger number of families, the present study reports important findings. Subjects can accurately self-report their weight and height and those of their family members. More importantly, the results of this study indicate that a self-reported measure of FHO is valid, suggesting that individuals are able to detect the presence or the absence of obesity in their first-degree family members. This finding is important for future research. Indeed, since the presence of a positive FHO has been highlighted as a predictive risk factor for the development of weight excess [##REF##10342802##17##, ####REF##10757206##18##, ##REF##9487953##19##, ##REF##11357211##20##, ##REF##9302300##21##, ##REF##17822567##22####17822567##22##], and because genetic counselling is done in clinical settings, obtaining an accurate FHO is essential. Studies such as those designed to understand difference between subjects with and without FHO should benefit from information about the accuracy of self-reported FHO.</p>" ]

[]

[ "<p>This is an Open Access article distributed under the terms of the Creative Commons Attribution License (<ext-link ext-link-type=\"uri\" xlink:href=\"http://creativecommons.org/licenses/by/2.0\"/>), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.</p>", "<title>Background</title>", "<p>Familial history information could be useful in clinical practice. However, little is known about the accuracy of self-reported familial history, particularly self-reported familial history of obesity (FHO).</p>", "<title>Methods</title>", "<p>Two cross-sectional studies were conducted. The aims of study 1 was to compare self-reported and objectively measured weight and height whereas the aims of study 2 were to examine the relationship between the weight and height estimations reported by the study participants and the values provided by their family members as well as the validity of a self-reported measure of FHO. Study 1 was conducted between 2004 and 2006 among 617 subjects and study 2 was conducted in 2006 among 78 participants.</p>", "<title>Results</title>", "<p>In both studies, weight and height reported by the participants were significantly correlated with their measured values (study 1: r = 0.98 and 0.98; study 2: r = 0.99 and 0.97 respectively; p &lt; 0.0001). Estimates of weight and height for family members provided by the study participants were strongly correlated with values reported by each family member (r = 0.96 and 0.95, respectively; p &lt; 0.0001). Substantial agreement between the FHO reported by the participants and the one obtained by calculating the BMI of each family members was observed (kappa = 0.72; p &lt; 0.0001). Sensitivity (90.5%), specificity (82.6%), positive (82.6%) and negative (90.5%) predictive values of FHO were very good.</p>", "<title>Conclusion</title>", "<p>A self-reported measure of FHO is valid, suggesting that individuals are able to detect the presence or the absence of obesity in their first-degree family members.</p>" ]

[ "<title>Competing interests</title>", "<p>The authors declare that they have no competing interests.</p>", "<title>Authors' contributions</title>", "<p>AMP performed the statistical analysis of the data and took the primary role in drafting the manuscript. LP, GG and MCV guided the strategy of the data analysis, assisted with the interpretation of the results, and provided critical review of the manuscript. LP, GG and MCV conceived the study. All authors read and approved the final manuscript.</p>" ]

[ "<title>Acknowledgements</title>", "<p>This research would not have been possible without the cooperation of the participating families. We would like to thank Marie-Eve Bouchard, Steve Amireault, Diane Drolet, and Dominique Beaulieu for their collaboration to the recruitment of the subjects, the study coordination and the data collection.</p>", "<p>Ann-Marie Paradis is supported by a doctoral research award from the Canadian Institutes of Health Research (CIHR) and the Fonds de la Recherche en Santé du Québec (FRSQ). Gaston Godin is Tier 1 Canada Research Chair in Health Related Behaviour, Laval University. This work was supported by a grant from CIHR – New Emerging Teams Programs (NET) (# OHN 63276).</p>" ]

[]

[ "<table-wrap position=\"float\" id=\"T1\"><label>Table 1</label><caption><p>Mean (± SD) self-reported and measured anthropometric measurements and their differences in men and women of the study 1 and study 2.</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td/><td align=\"left\">Self-reported</td><td align=\"left\">Measured</td><td align=\"left\">Difference</td></tr></thead><tbody><tr><td align=\"left\"><bold>Study 1</bold></td><td/><td/><td/></tr><tr><td align=\"left\"> Men (N = 245)</td><td/><td/><td/></tr><tr><td align=\"left\">  Weight (kg)</td><td align=\"left\">86.85 ± 17.73</td><td align=\"left\">87.00 ± 18.86</td><td align=\"left\">0.15 ± 3.27</td></tr><tr><td align=\"left\">  Height (cm)</td><td align=\"left\">176.77 ± 6.72</td><td align=\"left\">175.60 ± 6.90</td><td align=\"left\">-1.17 ± 2.3</td></tr><tr><td align=\"left\">  BMI (kg/m²)</td><td align=\"left\">27.73 ± 5.02</td><td align=\"left\">28.17 ± 5.54</td><td align=\"left\">0.43 ± 1.37</td></tr><tr><td align=\"left\"> Women (N = 372)</td><td/><td/><td/></tr><tr><td align=\"left\">  Weight (kg)</td><td align=\"left\">69.98 ± 14.99</td><td align=\"left\">70.61 ± 16.62</td><td align=\"left\">0.62 ± 2.79</td></tr><tr><td align=\"left\">  Height (cm)</td><td align=\"left\">163.38 ± 6.59</td><td align=\"left\">162.18 ± 6.46</td><td align=\"left\">-1.20 ± 2.21*</td></tr><tr><td align=\"left\">  BMI (kg/m²)</td><td align=\"left\">26.23 ± 5.43</td><td align=\"left\">26.88 ± 5.87</td><td align=\"left\">0.65 ± 1.36</td></tr><tr><td align=\"left\"><bold>Study 2</bold></td><td/><td/><td/></tr><tr><td align=\"left\"> Men (N = 26)</td><td/><td/><td/></tr><tr><td align=\"left\">  Weight (kg)</td><td align=\"left\">80.35 ± 10.83</td><td align=\"left\">80.86 ± 11.36</td><td align=\"left\">0.51 ± 1.28</td></tr><tr><td align=\"left\">  Height (cm)</td><td align=\"left\">175.17 ± 6.84</td><td align=\"left\">176.88 ± 6.72</td><td align=\"left\">1.71 ± 1.91</td></tr><tr><td align=\"left\">  BMI (kg/m²)</td><td align=\"left\">26.21 ± 3.39</td><td align=\"left\">25.88 ± 3.63</td><td align=\"left\">-0.32 ± 0.57</td></tr><tr><td align=\"left\"> Women (N = 52)</td><td/><td/><td/></tr><tr><td align=\"left\">  Weight (kg)</td><td align=\"left\">64.96 ± 14.34</td><td align=\"left\">65.93 ± 16.17</td><td align=\"left\">0.96 ± 4.44</td></tr><tr><td align=\"left\">  Height (cm)</td><td align=\"left\">161.42 ± 6.51</td><td align=\"left\">163.12 ± 6.78</td><td align=\"left\">1.70 ± 1.69</td></tr><tr><td align=\"left\">  BMI (kg/m²)</td><td align=\"left\">24.97 ± 5.53</td><td align=\"left\">24.86 ± 6.42</td><td align=\"left\">-0.11 ± 2.17</td></tr></tbody></table></table-wrap>", "<table-wrap position=\"float\" id=\"T2\"><label>Table 2</label><caption><p>Self-reported and family informant estimates of height, weight and body mass index (BMI) (study 2).</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td/><td align=\"left\">Self-reported values</td><td align=\"left\">Informant estimates</td><td align=\"left\">Difference</td></tr></thead><tbody><tr><td align=\"left\">Weight (kg)</td><td align=\"left\">72.72 ± 16.55</td><td align=\"left\">71.90 ± 16.00</td><td align=\"left\">0.81 ± 4.87</td></tr><tr><td align=\"left\">Height (cm)</td><td align=\"left\">166.91 ± 9.32</td><td align=\"left\">166.35 ± 8.92</td><td align=\"left\">0.56 ± 3.01</td></tr><tr><td align=\"left\">BMI (kg/m²)</td><td align=\"left\">26.04 ± 5.23</td><td align=\"left\">25.92 ± 5.05</td><td align=\"left\">0.12 ± 2.15</td></tr></tbody></table></table-wrap>", "<table-wrap position=\"float\" id=\"T3\"><label>Table 3</label><caption><p>Distribution of individuals according to the measure of the subjective and objective FHO.</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td/><td align=\"center\" colspan=\"2\">Objective measure of FHO</td><td/></tr><tr><td/><td colspan=\"2\"><hr/></td><td/></tr><tr><td/><td align=\"center\">Positive</td><td align=\"center\">Negative</td><td align=\"center\">Total</td></tr></thead><tbody><tr><td align=\"left\">Subjective measure of FHO</td><td/><td/><td/></tr><tr><td align=\"left\"> Positive</td><td align=\"center\">19</td><td align=\"center\">4</td><td align=\"center\">23</td></tr><tr><td align=\"left\"> Negative</td><td align=\"center\">2</td><td align=\"center\">19</td><td align=\"center\">21</td></tr><tr><td align=\"left\">Total</td><td align=\"center\">21</td><td align=\"center\">23</td><td align=\"center\">44</td></tr></tbody></table></table-wrap>" ]

[]

[]

[]

[]

[]

[]

[ "<table-wrap-foot><p>*Significantly different from the self-reported value, p &lt; 0.05.</p></table-wrap-foot>" ]

[]

[]

{ "acronym": [], "definition": [] }

[ "<title>Background</title>", "<p>It is increasingly recognized that individual variation is an important consideration in dietary interventions for weight loss, cardiovascular disease and other conditions. Position papers of scientific organizations make the point that it is \"unlikely that a single strategy is best for all\" (e.g. [##REF##15277443##1##]). It seems, however, that the principle is more widely applied in the breach than in the observance and most dietary trials continue to report means and standard deviations or other measures of group statistics. The use of group statistics in a dietary intervention implies not only that the population is uniform but, in addition, that there is a best diet for all individuals. In fact, given the difficulty in effecting weight loss, the test of a dietary strategy might better be whether it can be made to work for anybody. In this sense, the outliers become the key parameters in diet and the probability of reaching some criterion, rather than average behavior is intuitively a useful measure of efficacy of a diet. Further, weight loss involves inconvenience and dedication and a prospective dieter or physician is sensibly asking about relative payoffs of different dietary approaches – in essence, which diet to bet on. Thus, what one wants from a dietary comparison is to be able to compare individual performances on different diets rather than group values. Non-parametric statistics based on ranking such as the Mann-Whitney tests are applicable but because the rank becomes the primary variable, information is lost that a potential dieter would like to know.</p>", "<p>Here we describe a simple way to present individual data in comparative dietary trials that preserves the maximum amount of information and allows judgments of efficacy to be made. The idea is based on assessment of all possible paired comparisons and deduction of benefit from evaluation of a pay-off matrix, analogous to a matrix comparing different outcomes in games of strategy [##UREF##0##2##].</p>", "<p>The experimental ideal for taking account of individual variation is the cross-over protocol, that is, paired variables, in which subjects alternate between the two different diets. In this case, an odds-ratio for specified differential performance or some similar parameter will give the information that a dieter might want. Because it is difficult to perform any dietary intervention, however, a cross-over experiment is not always feasible and one may have to compare two groups in parallel who cannot generally be assumed to have uniform responses to diet.</p>" ]

[ "<title>Methods</title>", "<p>Individual results for the two diets to be compared are ranked according to outcome, for example, weight loss. All possible differences are calculated and these differences constitute the elements of a payoff matrix. The matrix elements can be color coded to indicate particular levels of relative payoff, e.g. &gt; 2 kg difference in weight loss. Probabilities of particular outcomes can also be calculated. The underlying rationale is that, since we don't know which subject in intervention A should be matched with which subject in intervention B, we consider all such possibilities and consider all the outcomes. The method has the advantage that it does not assume any particular distribution. It has the disadvantage that it assumes that the sample distribution of responses is somehow representative of the population distribution, and it therefore tends to over-emphasize differences.</p>", "<p>The matrix presentation brings out the qualitative features of individual responses and allows a graphic representation of the comparative effects. It emphasizes differences and therefore is exaggerated in comparison with sample means which tend to emphasize consistency.</p>", "<p>We illustrate the method with data from a paper by Volek, et al. [##REF##15533250##3##]. In this cross-over study, 15 men and 13 men were assigned to a low fat (LF) diet or a very low carbohydrate ketogenic diet (VLCKD). After a certain period subjects were switched to the other diet. We first treat the data as if this were a parallel experiment, that is, as if the LF and the VLCKD group were separate individuals. We then ask how the results compare to the actual outcomes where the pairings are known.</p>", "<p>The matrix shown in Figure ##FIG##0##1## represents the differences in all responses for the two diets (regardless of order). The horizontal row across the top of the matrix shows the individual values for weight loss on the VLCKD, while the vertical column on the left shows individual weight losses on LF. The matrix elements are the differences between the two diets, that is, the column value minus the row value. Positive values indicate that the VLCKD did better than the low fat. Examination of the color coding of the matrix shows that, consistent with the mean responses, there is a clear choice of the VLCKD. The actual probability predicted by the theoretical pairing shown in Table ##TAB##0##1## are calculated from the total number of matrix elements for each condition divided by the total.</p>", "<p>The matrix display shows that, beyond average performance, those subjects who benefited more from carbohydrate restriction compared to reduced fat generally did so in a big way (&gt; 5 kg differential weight loss). Table ##TAB##0##1## summarizes the probabilities of the various theoretical outcomes. An individual dieter is not guaranteed better outcome on either diet but the table demonstrates that they have a far better chance of showing dramatic results if they go with the VLCKD.</p>", "<p>Volek's experiment [##REF##15533250##3##] had a second phase in which subjects switched diets allowing a test of the matrix method, that is, one can ask how the actual within-subject comparisons compare to the anticipated outcomes of the matrix presentation. Table ##TAB##0##1## compares the probabilities from the matrix analysis and from the data in the cross-over experiment. The table shows that there is generally good agreement – fortuitously good for the overall comparison of the two groups – and it is clear that the matrix predicts, and only slightly exaggerates, the value of the VLCKD compared to LF.</p>" ]

[]

[ "<title>Discussion</title>", "<p>Group statistics tend to destroy information in order to gain reliability. The case might be made that not every experiment should have the same standards for the relative importance of these two parameters. In experiments where people can reasonably be expected to have similar responses, such as drug trials, an expectation value based on average outcome is the most relevant. On the other hand, given the difficulty of staying on any diet, the high non-Gaussian prevalence of obesity in the general populations [##REF##17498505##4##] and the variability due to hidden variables (like non-exercise activity thermogenesis (NEAT, [##REF##16026422##5##]), the prospective dieter is really asking which diet to bet on, that is, what is the best possible payoff for each diet and what are the odds, analogous to picking a strategy in the theory of games [##UREF##0##2##]. In the end, nobody loses an average amount of weight and the frequently quoted conclusion that low-carbohydrate and LF diets are the same at one year (e.g., [##REF##12761365##6##]) might be further enlightened by individual analysis along the lines suggested from a matrix presentation.</p>", "<p>Group statistics are integrative methods and decrease the signal to noise ratio and increase reliability but, because they obscure differences can repress future experiments. The matrix method is a derivative method and therefore increases the noise. It has reduced reliability but because it highlights potential differences, can provide a guide to future experiment and is hypothesis generating. Outlier data can be compared against prospective trials using Chi-square or Fischer exact testing to improve the reliability of the original observation. Use of both methods would provide maximum information about the nature of the comparison. Finally, while the example given is a weight loss experiment, the method is generalizable to all experimental parameters, such as lipid profile, that may be under the control of diet.</p>", "<p>Comparison of matrix data from the results of multiple experiments will be required to make the most reliable conclusions. Along these lines, a comparison of the numerical differences between LF and low carbohydrate diets, regardless of statistical significance showed the likelihood that the carbohydrate-restricted diet was better [##UREF##1##7##], that is, treated as separate Bernoulli trials, the results were not likely to be random. The matrix method would demonstrate whether the numerical benefit of the low carbohydrate diet was due to a sub-group with unusually good performance.</p>" ]

[]

[ "<p>This is an Open Access article distributed under the terms of the Creative Commons Attribution License (<ext-link ext-link-type=\"uri\" xlink:href=\"http://creativecommons.org/licenses/by/2.0\"/>), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.</p>", "<title>Objective</title>", "<p>To provide a simple method for presentation of data in comparative dietary trials.</p>", "<title>Methods</title>", "<p>Individual data from each diet are ranked and all possible paired comparisons are made and displayed in a pay-off matrix which can be color-coded according to the magnitude of the differences between the two diets. Probability of outcome can be calculated from the fraction of matrix elements corresponding to specified conditions. The method has the advantage of emphasizing differences and providing the maximum amount of information.</p>", "<title>Results</title>", "<p>The method was tested with values from the literature and allows intuitive sense of the comparative effectiveness of the two diets. In a test case in which a cross-over study had been performed the matrix derived from theoretical paired comparisons (treating the data as two parallel studies) was consistent with the results from the actual pairing in the cross-over.</p>", "<title>Conclusion</title>", "<p>The matrix method is a simple way of providing access to the differences between dietary trials. It exaggerates differences but can be used in combination with group statistics that, conversely, provide reliability at the expense of detailed information.</p>" ]

[ "<title>Competing interests</title>", "<p>The authors declare that they have no competing interests.</p>", "<title>Authors' contributions</title>", "<p>Authors contributed equally to the manuscript.</p>" ]

[]

[ "<fig position=\"float\" id=\"F1\"><label>Figure 1</label><caption><p><bold>Payoff matrix for dietary comparisons</bold>. Matrices show (theoretical) paired comparisons: Weight loss (in kg) for each individual in the VLCKD is shown in rank order across the top of the matrix (X-axis). Weight loss for the LF is shown down the side of the matrix (X-axis). Each matrix element shows the difference between the value for the VLCKD (column) and the value for the LF (row):VLCKD-LF. Positive values indicate more weight loss for the VLCKD value than the LF, negative values indicate the reverse. Data are from reference [##REF##15533250##3##] in which subjects were assigned to two diets with roughly similar caloric levels (VLCKD: 1855 kcal/d; LF: 1550 kcal/d) differing in nutrient composition: VLCKD = %carbohydrate:fat:protein = ~9:63:28, LF, ~58:22:20. After a fixed period (50 days for men; 30 days for women) subjects switched to the other diet. Data in the matrices are for performance in each phase. In the cross-over data, weight loss for each subject from the LF phase is subtracted from weight loss in the VLCKD phase (regardless of which came first in the experiment) and displayed in rank order. Color-coding as indicated in the figure.</p></caption></fig>" ]

[ "<table-wrap position=\"float\" id=\"T1\"><label>Table 1</label><caption><p>Probability of differences in outcomes in diet comparisons</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td/><td/><td align=\"center\" colspan=\"2\"><bold><italic>PROBABILITY</italic></bold></td></tr></thead><tbody><tr><td/><td/><td align=\"left\">MATRIX</td><td align=\"left\">X-OVER</td></tr><tr><td/><td/><td/><td/></tr><tr><td align=\"left\"><bold><italic>MEN</italic></bold></td><td align=\"left\"><bold>VLCKD &gt; LF</bold></td><td align=\"left\"><bold>0.73</bold></td><td align=\"left\"><bold>0.73</bold></td></tr><tr><td/><td align=\"left\">LF &gt; VLCKD</td><td align=\"left\">0.27</td><td align=\"left\">0.27</td></tr><tr><td/><td align=\"left\"><bold>VLCKD – LF &gt; 2</bold></td><td align=\"left\"><bold>0.59</bold></td><td align=\"left\"><bold>0.53</bold></td></tr><tr><td/><td align=\"left\">LF – VLCKD &gt; 2</td><td align=\"left\">0.17</td><td align=\"left\">0.27</td></tr><tr><td/><td align=\"left\"><bold>VLCKD – LF &gt; 5</bold></td><td align=\"left\"><bold>0.31</bold></td><td align=\"left\"><bold>0.33</bold></td></tr><tr><td/><td align=\"left\">LF – VLCKD &gt; 5</td><td align=\"left\">0.02</td><td align=\"left\">0</td></tr><tr><td/><td/><td/><td/></tr><tr><td align=\"left\"><bold><italic>WOMEN</italic></bold></td><td align=\"left\"><bold>VLCKD &gt; LF</bold></td><td align=\"left\"><bold>0.78</bold></td><td align=\"left\"><bold>0.62</bold></td></tr><tr><td/><td align=\"left\">rLF &gt; VLCKD</td><td align=\"left\">0.21</td><td align=\"left\">0.31</td></tr><tr><td/><td align=\"left\"><bold>VLCKD – LF &gt; 2</bold></td><td align=\"left\"><bold>0.69</bold></td><td align=\"left\"><bold>0.54</bold></td></tr><tr><td/><td align=\"left\">LF – VLCKD &gt; 2</td><td align=\"left\">0.09</td><td align=\"left\">0.31</td></tr><tr><td/><td align=\"left\"><bold>VLCKD – LF &gt; 5</bold></td><td align=\"left\"><bold>0.09</bold></td><td align=\"left\"><bold>0.31</bold></td></tr><tr><td/><td align=\"left\">LF – VLCD &gt; 5</td><td align=\"left\">0</td><td align=\"left\">0</td></tr></tbody></table></table-wrap>" ]

[]

[]

[]

[]

[]

[]

[ "<table-wrap-foot><p>The number of matrix elements corresponding to each condition are divided by the total number of matrix elements (paired differences) in the matrices in Figure 1. For cross-over, difference for results (VLCKD – LF) for each subject are divided by the number of subjects.</p></table-wrap-foot>" ]

[ "<graphic xlink:href=\"1475-2891-7-24-1\"/>" ]

[]

{ "acronym": [], "definition": [] }

[ "<title>Introduction</title>", "<p>In this review we discuss the issues associated with the application of bioelectrical impedance analysis (BIA) to measure body composition in large epidemiologic studies with multiethnic populations. The review is limited to healthy adults and does not include children, adolescents, elderly, and unhealthy individuals. The most recent system such as foot (or hand) to foot system is the main focus of this review and the early tetra-polar electrode system will not be discussed. These recent models are readily available and easy to use.</p>", "<p>Percent body fat is strongly associated with the risk of chronic diseases such as hypertension, dyslipidemia, diabetes mellitus, and coronary heart disease [##REF##16157073##1##, ####REF##15867303##2##, ##REF##15868117##3##, ##REF##15364185##4####15364185##4##]. In epidemiological studies, surrogate measures of body fatness such as body mass index (BMI), waist circumference, waist-hip ratio and skin fold thickness have been used extensively. However, these techniques do not precisely characterize persons by body composition (percentage of body fat or muscle mass), and there is substantial variation across age, sex and ethnic groups [##REF##15660034##5##, ####REF##10865763##6##, ##REF##911746##7####911746##7##]. Several techniques have been used to assess percent body fat in controlled laboratory conditions. These include underwater weighing (densitometry), dual energy x-ray absorptiometry (DEXA), bioelectrical impedance analysis (BIA) and magnetic resonance imaging (MRI). However, densitometry, DEXA, and MRI are expensive, inconvenient for the participant, and not feasible to conduct in the field because they require large specialized equipment. For these reasons, their use in large epidemiological studies is limited.</p>", "<p>BIA, by contrast, is relatively simple, quick (takes only a few minutes), and non-invasive which gives reliable measurements of body composition with minimal intra- and inter-observer variability [##REF##2707216##8##]; the results are available immediately and reproducible with &lt;1% error on repeated measurements [##REF##2058583##9##]. This technique became commercially available for the first time in the mid- 1980s [##REF##16215137##10##], and requires inexpensive, portable equipment, making it an appealing alternative to assess body composition in epidemiological studies [##REF##14618504##11##].</p>", "<title>Principles of bioelectrical impedance technique</title>", "<p>BIA analysis is based on the principle that electric current flows at different rates through the body depending upon its composition. The body is composed mostly of water with ions, through which an electric current can flow. The water in the body is localized in two compartments: extra-cellular water (ECW, approximately 45%) and intracellular water (ICW, approximately 55%) [##REF##15809537##12##]. On the other hand, the body also contains non-conducting materials (body fat) that provide resistance to the flow of electric current. Adipose tissue is significantly less conductive than muscle or bone [##REF##11319654##13##]. The principal of BIA is that electric current passes through the body at a differential rate depending on body composition. Hence, there is a direct relationship between the concentrations of ions and the electrical <italic>conductivity </italic>and an indirect relationship exists between the ion concentration and the <italic>resistance </italic>of the solution.</p>", "<p><bold>Body impedance (Z) </bold>is defined as the opposition of a conductor to the flow of an alternating current, and consists of two components: resistance (R) and reactance (Xc). <bold>Resistance (R) </bold>is the major opposition of the conductor and at usual low frequency (50 kHz), the extra-cellular part of non-adipose tissue works as a resistor [##REF##8780369##14##]. <bold>Reactance </bold>is an additional opposition or the storage of an electrical charge by a condenser for a short period of time; the lipid component of the membranes of the Body Cell Mass (BCM) behave as capacitors and reduce the flow of intracellular ions. In practice, impedance is the amount of dropped voltage when a small constant current (800 uA) with a fixed frequency (50 kHz) passes between electrodes spanning the body. However, lean tissue, which is rich in water and electrolytes, has minimal impedance and increases to a maximum when all lean tissue is replaced by fat/adipose tissue. Hence, lean body mass and Fat Mass (FM) can be calculated from the difference in conductivity [##REF##12508951##15##].</p>", "<p>The other assumptions for BIA measurement are that the body is a cylindrical-shaped ionic conductor with homogeneous composition, a fixed cross-sectional area and a uniform distribution of current density [##REF##8780360##16##,##REF##15556267##17##]; BIA measures the impedance to the flow of an electric current through the total body fluid. Therefore, the conductive volume (V) which represents total body water (TBW) or FFM is directly related to the square length of conductor (S) and inversely correlated to resistance of the cross-section area (R), while <italic>p </italic>is the specific receptivity of the conductor, yielding the equation: V = <italic>p </italic>× S²/R. Based on this assumption, the same arms and legs respectively contribute to almost 47% and 50% of whole body resistance despite contributing to 4% and 17% of body weight respectively. In contrast, the trunk, which contains 50% of the body mass, contributes only 5–12% of whole body resistance [##REF##15809537##12##].</p>", "<title>Predictive equations</title>", "<p>Many empirical equations have been developed for estimation of TBW, FFM and body cell mass (BCM), by using sex, age, weight, height and race as explanatory variables. However, predictive equations are generally population-specific and can be useful only for those populations with characteristics similar to those of the reference populations [##REF##11960296##18##,##REF##12806211##19##]. When these equations have been used to predict body composition in different populations, the results have been inconsistent. The developed predictive equations cannot be generalized to diverse populations. Heyward and Wagner reviewed the reliability and validity of different equations for African Americans, Asians and Indian Americans. They found that the majority of studies indicated that the BIA method is not accurate when a generalized equation is applied for different ethnic groups [##UREF##0##20##].</p>", "<title>Summary of bio-impedance technique</title>", "<p>▪ Based on the principle that body fat impedes electric current more than body protein</p>", "<p>▪ Impedance is a drop in voltage when a small constant current with a fixed frequency passes between electrodes spanning the body</p>", "<p>▪ Predictive equations estimate TBW, FFM and body cell mass (BCM) using sex, age, weight, height and race</p>", "<title>Validity of BIA measurements</title>", "<p>The human body is not uniform either in length, cross-sectional area, or ionic composition and this affects the accuracy of BIA measurements [##REF##12508951##15##]. In addition, body impedance varies among different ethnic groups and influences the accuracy of BIA [##REF##10865738##21##]. Validity of hand to hand (Omron BF306 BIA) with a 4-C model was tested among Chinese and Japanese participants which showed different levels of biases in predicted levels of body fat (SEE = 4.5% BF) which may have resulted from different levels of body fat, age and relative arm span [##REF##11890632##22##]. Demura et al. in a sample of 50 Japanese men aged 18 to 27 y. validated foot-to-foot (Tanita, TBF-102), and hand-to-hand (Omron, HBF-300) and hand-to-foot (Selco, SIF-891) BIA analyzers against hydro-densitometry (HD) [##REF##11811572##23##]. They found higher correlation between hand to foot (r = 0.96) than foot to foot (r = 0.71) against HD as a reference method and there was 2.2% to 3.3% overestimation when they used the manufacturer's equations, therefore, they developed new equations for their sample. Jebb et al. tested the validity of foot-to-foot (Tanita -350) among 104 men and 101 women recruited from Dunn Nutrition Centre using DEXA as a reference method. The observed limit of agreement for fat mass was ± 7.9 kg [##REF##10743490##24##]. A number of other factors that influence BIA results are described in this section.</p>", "<title>Consumption of food or beverages</title>", "<p>Although food or fluid intake before BIA measurement affects TBW and ECW, a general agreement on the ideal amount of time between food and fluid intake and BIA measurements has yet to be consolidated. It has been suggested that due to the large cross-sectional surface of the trunk, even fluid intake of up to 2 L is shown to be \"electrically silent\" during the first hour after consumption [##REF##9569522##25##,##REF##8780358##26##]. Kaminsky and Whaley (1993) compared body fat percentage measurements after 3 hours and 12 hours of fasting and found no significant difference between these values [##REF##8412053##27##]. Lukaski et al., (1986) emphasizes that dehydration increases resistance by nearly 40 Ω, which results in a 5.0 kg underestimation of FFM [##REF##3700310##28##] and Evans et al., (1998) showed increased impedance one hour after eating a heavy meal [##REF##9569522##25##]. In contrast, investigators have reported that food intake, its absorption and the resulting increase in movement of fluid into the bloodstream from 2–4 hours before BIA measurement, decreases the impedance value from 4 to 15 Ω, or &lt;3% and results in overestimation of FFM by almost 1.5 kg [##REF##3234328##29##]. Slinde and Rossander-Hulthen, after giving standard food to 18 healthy subjects, measured BIA 18 times during 24 hr. Their results showed that percentage of body fat varied by 8.8% and 9.9% from the highest to the lowest measurement in women and men respectively [##REF##11566645##30##]. In contrast, Chumlea et al., (1987) found no effect of food consumption before BIA measurement on impedance measurements [##REF##3596566##31##]. For these reasons undertaking an overnight fast is recommended as a routine standardization technique before impedance measurements [##REF##15556267##17##,##REF##8110141##32##].</p>", "<title>Exercise</title>", "<p>Although exercise of mild intensity may not affect BIA measurements, moderate and intensive exercise before measurements may change the measured impedance by different mechanisms [##REF##2401285##33##]. For example, exercise increases cardiac output and vascular perfusion and subsequently increases blood flow to skeletal muscle, which warms the muscle and decreases muscle resistance which results in reduced impedance [##REF##8780358##26##]. In addition, intensive activity causes vasodilatation, an increase in skin temperature, which also reduces measured impedance [##REF##3193865##34##]. Jogging or cycling at moderate intensities for 90–120 min decreases measured impedance by 50 to 70 Ω, which results in nearly a 12 kg overestimation of FFM [##REF##3364394##35##]. Therefore, to reduce measurement error, BIA should not be performed within several hours of moderate to intensive exercise. In addition, the chosen mode for each individual may affect the accuracy of measurement. Swartz et al, in a well designed study, compared the % BF measured among high or moderately active and inactive individuals by hydrostatic weight and BIA using different athletic and adult modes in a foot-to-foot BIA (Tanita TBF-305). Their results showed that although the electrical impedance was not significantly different, the chosen adult mode for highly and moderately active individuals significantly overestimated the percent of body fat [##REF##12144724##36##].</p>", "<title>Medical conditions</title>", "<p>Although some investigators have applied BIA method in various patients and clinical settings, it should be noted that there are some medical conditions which change serum electrolytes, hematocrit and blood flow, affecting Z and <italic>p</italic>, independent of body fluid volume [##REF##8780358##26##]. Conversely, there are some other medical conditions, which via a change in fluid distribution alter Z measurements. Significant alteration in body hydration, fluid distribution and differences in the ratio of ECW to ICW caused by a medical condition will affect impedance measurements [##REF##8849209##37##,##REF##8676831##38##]. Among those conditions, the most significant confounding variable is edema of the distal extremities, which is mainly caused by peripheral venous insufficiency. This insufficiency may result from congestive heart failure, cirrhosis, nephrotic syndrome, hypoalbuminemia, and lympheodema [##REF##10372157##39##]. Other medical conditions, which affect BIA validity, include cutaneous disease that may alter electrode-skin electrical transmission in patients with amputations, poliomyelitis and muscular dystrophies. These conditions will have significant effects on the application of BIA in the clinical population [##REF##15556267##17##,##REF##15380917##40##].</p>", "<title>Environmental factors</title>", "<p>Although environmental changes do not significantly affect actual whole body volume, they appear to alter the Z measurements by changing skin temperature. The result of several studies showed an inverse relation between skin temperature and impedance which means impedance increases with a lowering in temperature and decrease with a rise in skin temperature. Gudivaka et al observed 8% change in resistance at 50 kHz with 8.4°C change in skin temperature [##REF##8872654##41##]. Thus, changes in cutaneous and muscle blood flow may have a large impact on BIA measurements in both clinical and field settings.</p>", "<title>Within-subject variability</title>", "<p>Due to increased progesterone plasma levels after ovulation and the change in hydration status, within-subject variability of impedance may be higher in women. The effect of this variability has been examined by several studies and various results have been reported. Gualdi-Russo et al., did not find significant differences in TBW estimated at different points in time during the follicular and premenstrual stages [##REF##12365353##42##]. On the other hand, Gleichauf et al., suggested that the average of several measurements during a menstrual cycle could be considered as an estimation of body composition [##REF##2816797##43##]. However, it has been recommended that BIA measurement not be taken at a time while the participant is experiencing large weight gain related to the menstrual cycle [##REF##9738136##44##]. Menopause changes body composition and fat distribution and women experience a loss in lean mass and an increase in weight, fat mass and central fat deposition [##REF##12833110##45##, ####REF##8002510##46##, ##REF##7985622##47##, ##REF##1985614##48##, ##REF##10591232##49####10591232##49##]. The ratio of fat/lean mass, especially in the lower part of the body increases [##REF##16155280##50##,##REF##7589642##51##], which may affect the estimated impedance as the current passes through the legs. Therefore, the accuracy of BIA measurements increases by applying specific prediction equations for postmenopausal women [##REF##10557029##52##].</p>", "<title>Ethnicity</title>", "<p>In recent years, BIA has been extensively applied among different age groups of both sexes, including mostly Caucasian populations of USA and Europe, and several prediction equations have been developed for these samples [##REF##2043597##53##, ####REF##11568498##54##, ##REF##11312069##55####11312069##55##]. Also, a few prediction equations have been developed based on samples from African Americans, Hispanics and Native Americans [##REF##9209163##56##]. Stolarczky et al., (1997) showed that by applying population-specific equations for estimation of lean body mass among Native American women, the standard error for estimating (SEE) decreased from 8.1 kg to 2.6 kg [##REF##9209163##56##]. However, it has been suggested that biological and physiological assumptions for estimation of body composition, which are mainly based on Caucasian samples, may not be accurate for other ethnic groups. <bold>Hence, the validity of these equations must be tested in the population under study</bold>. There are several factors responsible for ethnic differences, which may affect the extent and direction of the error while measuring body composition by BIA such as:</p>", "<p>• <bold>Fat distribution </bold>Ethnicity affects fat patterning and consequently influences the validity of equations. It has been shown that the proportion of fat deposition on trunk varies by 5.7% between different ethnic groups of Asians, Mexican Americans, Caucasians and African Americans [##REF##8589780##57##].</p>", "<p>• <bold>Body density </bold>Body density may have a significant impact on the accuracy of estimated lean body mass and fat-free mass. Several studies showed that African Americans have greater body density and greater body mass cell compared to Caucasian Americans [##REF##9539193##58##,##REF##13411180##59##]. Swinburn et al., (1999) found that Polynesians in New Zealand have higher levels of fat-free mass and less body fat than Europeans at any given body mass index [##REF##10578208##60##]. In contrast, Kyle et al., (2001) indicated that Japanese men and women had 10–12% higher body fat than Swiss men and women [##REF##11312069##55##]. It has also been reported that Asian populations (Chinese, Malay, Singaporean Indians) have higher body fat percentages at a given BMI and Wang et al. reported a lower hydration of the FFM in Asians [##REF##10865763##6##,##REF##10951540##61##].</p>", "<p>In prediction equation calculations, it has been assumed that the fat free mass density does not vary among different ethnic groups. Because the density of FFM differs between different ethnic groups, this assumption may be a major source of error.</p>", "<p>• <bold>Differences in proportional limb lengths </bold>as mentioned before, impedance demonstrates a direct relationship between conductive volume (V) and the square length of a conductor (S). Since whole body impedance is mainly based on the impedance of limbs [##REF##2627926##62##], the differences among different racial groups may mostly relate to differences in proportion of limb lengths [##REF##10837277##63##]. This hypothesis is supported by several studies, for example, whole-body impedance of Nigerians was significantly greater than that of matched Caucasian individuals, but was not different among different tribes of Nigeria [##REF##14618504##11##]. Also, several other studies showed that black populations have longer limbs than white populations and increased lumbar lordosis [##REF##721084##64##, ####REF##5091624##65##, ##REF##14618332##66####14618332##66##].</p>", "<p>Generally speaking, based on the preceding hypothesis, regarding age, race, level of activity etc. it has been suggested that the general prediction equation across different age and ethnic groups should not be applied without cross validating the study population [##REF##10951540##61##,##REF##8329342##67##].</p>", "<title>Summary of factors impacting BIA results</title>", "<p>▪ Contact between limbs and trunk</p>", "<p>▪ Inaccurate body weight</p>", "<p>▪ Consumption of food and drink (overnight fast suggested)</p>", "<p>▪ Moderate to intense level physical activity 2–3 hours before measurement</p>", "<p>▪ Medical conditions impacting fluid and electrolyte balance</p>", "<p>▪ Ambient temperature (cold increases impedance)</p>", "<p>▪ Individual characteristics (abdominal obesity, muscle mass, weight loss, menstrual cycle, menopause)</p>", "<p>▪ Ethnic variation, possibly mediated by body density and proportional limb length</p>" ]

[]

[ "<title>Summary of factors impacting BIA results</title>", "<p>▪ Contact between limbs and trunk</p>", "<p>▪ Inaccurate body weight</p>", "<p>▪ Consumption of food and drink (overnight fast suggested)</p>", "<p>▪ Moderate to intense level physical activity 2–3 hours before measurement</p>", "<p>▪ Medical conditions impacting fluid and electrolyte balance</p>", "<p>▪ Ambient temperature (cold increases impedance)</p>", "<p>▪ Individual characteristics (abdominal obesity, muscle mass, weight loss, menstrual cycle, menopause)</p>", "<p>▪ Ethnic variation, possibly mediated by body density and proportional limb length</p>" ]

[]

[ "<title>Conclusion</title>", "<p>BIA has become a popular method for estimation of body composition during the last two decades. Since 1990, more than 1600 published articles have been reported using BIA as a tool of body composition measurement [##REF##15556267##17##,##REF##15380917##40##,##REF##15809535##68##] and our search with the key words of body composition and bioelectrical impedance showed that 235 articles were published in English between 2004 and 2006 and we found different levels of agreements between different BIA models and reference methods. Also, there are many different equations for BIA calibration thus results of studies should be compared with more caution. BIA seems to reasonably estimate body composition in controlled conditions for healthy and euvolemic adults by applying a population specific predictive equation and it is not recommended to generalize a few equations for international epidemiologic studies, which involve participants from diverse populations. As far as we know, for some ethnic groups such as South Asians or Middle Easterners, or African residing in Africa predictive equations have not yet been developed. Hence, it is necessary to develop new predictive equations or cross validate existing equations on new populations to be studied.</p>", "<p>If the BIA equation is not appropriately chosen based on age, gender, level of physical activity, level of body fat and ethnicity, the results of the study will not be reliable.</p>", "<p>Overall BIA is a useful tool for clinical studies, but for large epidemiological studies with diverse population, particularly in developing nations, BIA has limited use unless valuation studies are conducted specifically for the populations under study.</p>" ]

[ "<p>This is an Open Access article distributed under the terms of the Creative Commons Attribution License (<ext-link ext-link-type=\"uri\" xlink:href=\"http://creativecommons.org/licenses/by/2.0\"/>), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.</p>", "<p>Percentage of body fat is strongly associated with the risk of several chronic diseases but its accurate measurement is difficult. Bioelectrical impedance analysis (BIA) is a relatively simple, quick and non-invasive technique, to measure body composition. It measures body fat accurately in controlled clinical conditions but its performance in the field is inconsistent. In large epidemiologic studies simpler surrogate techniques such as body mass index (BMI), waist circumference, and waist-hip ratio are frequently used instead of BIA to measure body fatness. We reviewed the rationale, theory, and technique of recently developed systems such as foot (or hand)-to-foot BIA measurement, and the elements that could influence its results in large epidemiologic studies. BIA results are influenced by factors such as the environment, ethnicity, phase of menstrual cycle, and underlying medical conditions. We concluded that BIA measurements validated for specific ethnic groups, populations and conditions can accurately measure body fat in those populations, but not others and suggest that for large epdiemiological studies with diverse populations BIA may not be the appropriate choice for body composition measurement unless specific calibration equations are developed for different groups participating in the study.</p>" ]

[ "<title>Competing interests</title>", "<p>The authors declare that they have no competing interests.</p>", "<title>Authors' contributions</title>", "<p>MD ran the electronic searches, reviewed all abstracts and articles, coordinated and drafted the manuscript. ATM participated in reviewing the articles and helped to draft the manuscripts.</p>" ]

[ "<title>Acknowledgements</title>", "<p>We wish to thank Dr. Yusuf (Director of Population Health Research Institute) for all his supports and guidance.</p>" ]

[]

[]

[]

[]

[]

[]

[]

[]

[]

[]

[]

{ "acronym": [], "definition": [] }

[ "<title>Background</title>", "<p>Recent estimates suggest that up to 1.7 billion people worldwide are overweight or obese, making it one of the biggest health threats facing world's population. Obesity lies at the other end of malnutrition scale and is becoming a public health problem in developing countries as well. Over 115 million people suffer from obesity related problems in developing countries [##REF##11010948##1##]. In Brazil and Colombia for example, 36 and 41% of the population respectively is overweight. Prevalence of obesity in Mexico was unknown until recently [##UREF##0##2##,##UREF##1##3##]: about 26% of children between 5 and 12 years of age and 35% of the adult women are obese. The high prevalence of obesity in the Mexican population must be contributing to the increment in chronic diseases that has been observed in recent years [##UREF##2##4##]. Health officials and academia have recognized the need for urgent preventive measures to stop this accelerating trend.</p>", "<p>Several studies have identified an excessive intake of dietary fat as a major mechanism for increasing the amount of body fat in humans and experimental animals. Diets with a high fat content are energy dense [##UREF##3##5##]. Thus, reduction of dietary fat as a treatment for obesity has been a widely used approach. A number of trials with low-fat diets have demonstrated the effectiveness of such recommendation [##UREF##4##6##, ####REF##10889789##7##, ##UREF##5##8####5##8##]. In addition to weight loss, low fat diets help maintain low cholesterol and triglyceride levels in blood, reduce leptin concentration, increase adiponectin and reduce insulin resistance, and decrease cardiovascular and diabetes risk [##REF##16256015##9##,##UREF##6##10##].</p>", "<p>An increase in the carbohydrate to fat ratio is associated with the reduction in energy density of the diet [##REF##10721885##11##]. A dietary recommendation to increase cereal consumption is a possible approach to improve the carbohydrate to fat ratio. Studies in adult men and women have demonstrated that an increase in dietary carbohydrates from ready-to-eat cereals (RTEC) or other foods, even in the lack of an advice to reduce fat, is a potentially effective approach for weight reduction [##UREF##3##5##,##REF##9234028##12##,##UREF##7##13##].</p>", "<p>The objectives of the study were: 1) To determine if an increase in cereal intake by consuming RTEC, among overweight or at risk of overweight children is an effective treatment to reduce excess body fat, 2) To determine if the inclusion of a nutrition education program in addition to an increase in carbohydrate intake has an effect on body weight and body fat, and 3) To determine if an increase in RTEC intake alone or with a nutrition education program has an effect on plasma lipid profile.</p>" ]

[ "<title>Methods</title>", "<title>Subjects and place of study</title>", "<p>Children were eligible if they had a BMI for age &gt; 85% and were attending elementary school with an age range from 6 to 12 years. In order to detect children as being overweight or at risk of overweight, 6 elementary schools of the city of Queretaro were randomly selected and invited to participate; 5 schools accepted participation. Parents of all children from 1^stto 6^thgrade were invited to a session where details of the study were explained, including benefits and potential risks of child participation. Parents of 905 children accepted voluntarily to participate in an initial screening to detect overweight or at risk of overweight children. Weight and height were determined in all children at their schools. Children were weighed without sweater or jacket and without shoes using an electronic scale (SECA, Erecta 844, Hamburg, Germany) to the nearest 1 g. Height was measured using portable stadimeters (SECA, Bodymeter 208, Germany). Children with a BMI-for-age above the 85^thpercentile were enrolled in the experimental study. According to the Center for Disease Control and Prevention (CDC) references, a child at risk of overweight is defined as having a BMI-for-age between the 85^thand 95^thpercentile of the CDC growth charts [##UREF##8##14##]. Overweight is defined as a BMI-for age at or above the 95^thpercentile (14).</p>", "<p>Of the 905 children initially screened, 17% had a BMI-for-age percentile between 85% and 95%, and 18% had a BMI-for-age percentile equal or above 95%. Of these overweight and at risk of overweight children, 256 accepted to participate in a longitudinal controlled study, from which 178 children completed the study. Lost to follow-up was mainly due to the children's lack of compliance to the study protocol. The sample size of 178 subjects that completed the study accomplishes the expected sample size with an alpha error of 0.05 and a beta error of 0.2, to detect a BMI expected difference of 1 kg/m², with an expected standard deviation of BMI change of 1 kg/m². Blood samples were taken from children if parents agreed to the procedure. Of the 178 subjects that completed the study, a blood sample was obtained from 129 children. Children included in the study were healthy volunteers with no apparent disease apart from being overweight.</p>", "<title>Experimental groups and treatments</title>", "<p>Children were randomly assigned to one of four different treatments. They were stratified into 4 groups with similar age, height and BMI percentile and same gender, in order to create groups with similar baseline characteristics. A randomization of treatments was done to each group with a computer generating random number list. The randomization was done at a central office by someone who did not have contact with the children or their parents. Children in group 1 consumed one serving of 33 ± 7 g of RTEC (Kellogg's de Mexico, Querétaro, Mexico) at breakfast. Children in group 2 consumed two servings of 33 ± 7 g of RTEC, one at breakfast and another serving at dinner. Children in group 3 consumed one serving of 33 ± 7 grams of RTEC and in addition, both children and mothers received a nutrition education guide that contained recommendations for healthy eating. Children in group 4 were involved in the study and had no treatment. Follow up in all groups was for 12 weeks.</p>", "<p>To allow for variety in the diet, children consumed from 4 different types of RTEC: a corn based RTEC (Corn Flakes^®, Kellogg Company Mexico), a pre-sweetened corn based RTEC (Zucaritas^®, Kellogg Company Mexico), a pre-sweetened corn based, chocolate flavored RTEC (Choco Zucaritas^®, Kellogg Company Mexico), and a pre-sweetened rice based, chocolate flavored RTEC (ChocoKrispis^®, Kellogg Company Mexico). These RTEC were chosen because of the high consumption among children. The children were allowed to choose from the 3 pre-sweetened RTEC only for 3 days in one week and were not allowed to repeat. The remaining four days children consumed from corn-based cereal only. The mean nutrient composition of RTEC per serving was as follows: 165 Kcal (712 KJ), 5.8 g of protein, 0.5 g total fat, and 35 g of carbohydrates. The RTEC was consumed with 250 mL of cold skimmed milk in a bowl with a spoon. Compliance was recorded by weekly interviews to the parents.</p>", "<p>A nutrition education guide was prepared by one of us (RA) based on general recommendations for obese individual developed by Perez-Lizaur and Marvan [##UREF##9##15##] which included recommendations for the whole family and recommendations for the child. The nutrition education program included 12 sessions (one per week) that were given at school to the children's parents (usually the mother), both in oral and written form. The dietary recommendations were given by a nutritionist. Practice of the recommendations mentioned above was monitored weekly during RTEC delivery at the school by asking the parents if they had any difficulty following the nutrition education guide. Table ##TAB##0##1## shows a summary of the major aspects included in the nutrition education guide. As part of the education guide, a sample menu was provided so that parents could use it to plan their meals at home and for school.</p>", "<p>Children in all four groups were evaluated for anthropometry, body composition, physical activity, and blood lipids at the beginning of the study before treatments and after 12 weeks with each respective treatment.</p>", "<title>Anthropometry, body composition and blood lipids</title>", "<p>Anthropometric measures included weight and height and were done as described above. Standardization in height and weight measures was done following standard procedures recommended by the World Health Organization [##UREF##10##16##]. Each child was evaluated by the same observer at basal and post-treatment. Body composition analysis was carried out by bioelectrical impedance using a conductance measurement apparatus (BIA 101, RJL Systems, Clinton TWP, MI). The apparatus was calibrated everyday before measures were carried out. Children were laid down in a bed placed in a quiet room inside the school, apart from where the rest of the measurements took place. Electrodes were placed on the left foot and right hand, after cleaning the area with alcohol. Children were asked to remain calm and not to move for the duration of evaluation.</p>", "<p>A fasting blood sample was drawn from every child at basal and after 12 weeks of treatment. Children in all schools were asked to attend at 8 in the morning. Mother and child were instructed that the child should not have any food after 9 p.m. on the night before. Both mother and child were asked before the blood sample was taken if the child had fasted. Blood samples were centrifuged at 1800–2000 rpm during 15 minutes and plasma was stored at -20°C until analysis. Biochemical analysis in plasma samples included triglycerides, total cholesterol and HDL cholesterol and were done using a commercial kit (Sera-Pak Kit Bayer Diagnostics, France).</p>", "<title>Physical activity evaluation</title>", "<p>Physical activity of all children was evaluated by asking the child's mother to fill out a questionnaire at the beginning of the study and 12 weeks after treatment began. The questionnaire asked to recall different physical activities normally carried out by children throughout the week as well as their duration. This questionnaire has been validated and applied in previous studies [##UREF##11##17##]. The outcome of the questionnaires showed the time spent performing different activities during the week. Time of each type of activity was transformed into Metabolic Equivalent units (Mets/hr), which is the ratio of the metabolic rate during the physical activity to the resting metabolic rate, according to the compendium of physical activities from the Prevention Research Center of the University of South Carolina [##UREF##12##18##]. For data analysis, physical activities were grouped into intense, moderate and low depending on Mets/hr spent as follows: Low = 0 to 3 Mets/hr, Moderate = 3 to 6 Mets/hr and Intense = 6 or more Mets/hr.</p>", "<title>Data analysis</title>", "<p>Percent fat and fat free mass were calculated from the reactance and resistance values obtained in the bioimpedance analysis using the equation suggested by Kottler et al. (1996) [##REF##8780369##19##]. LDL and VLDL were calculated from total cholesterol, HDL and triglycerides concentrations [##UREF##13##20##]. BMI and BMI percentile were calculated in Epi-Info v.3.3.2. Treatment effect was measured as the change on anthropometric and biochemical determinations within initial and final measures and mean change among groups comparison. Partial measurements were analyzed to confirm validity of initial and final measurements. Within effects were carried out with a paired T-Test. Between groups effect in lipids and anthropometry changes was observed with a one-way ANOVA to compare unadjusted changes and with a univariate general linear model adjusted for baseline value, gender and interactions in case they resulted significant and the school random effect. Physical activity analysis was evaluated as the final evaluation controlled for baseline value, gender and the school random effect. Treatments' pairwise comparisons were tested with the least significant difference test [##REF##3440385##21##]. Additionally, an analysis of variance and a chi square test was carried out to compare baseline age, anthropometry and gender of subjects included in the analysis versus children that had missing data and were not included in the analysis. Statistical analyses were performed using the software SPSS, V.9.0 (Chicago, IL).</p>" ]

[ "<title>Results</title>", "<p>Children were recruited from October to December 2002 and the fieldwork was from January to June 2003. The statistical analyses considered all children that had initial and final measurements in an intention to treat basis. Only one child that had an extreme weight final value was excluded from analysis. The participants' flow chart is shown in figure ##FIG##0##1##. Age, gender and height were not different between children included and children excluded from the analysis. Characteristics of subjects in the experimental groups at the beginning of the study are shown in table ##TAB##1##2##. Changes in weight, BMI and body composition are shown in Table ##TAB##2##3##. After 12 weeks of intervention there was a significant increase in body weight in the two RTEC groups and in the control group, only the group that had RTEC plus nutrition education had no increment in body weight. In analysis of variance, children that consumed one serving of RTEC and had nutrition education had a difference in unadjusted weight change of 2.03 kg compared with children in the control group (p &lt; 0.01). Body weight change in the RTEC and nutrition education group adjusted for gender, school and baseline body weight was also significantly different from the control (p &lt; 0.001) and the other two treatment groups (p &lt; 0.01). Unadjusted and adjusted changes in body weight with both treatments with RTEC alone were not statistically different from the control group. BMI reduced significantly only in the group of children that received RTEC and nutrition education (p &lt; 0.01); children in this group had an unadjusted change in mean BMI of 0.64 kg//m²higher than the control group (p &lt; 0.01). This group's adjusted change in BMI was also statistically greater than control (p &lt; 0.01) and the other two treatments with RTEC only (p &lt; 0.05). Children in the RTEC and nutrition education group showed an unadjusted decrease in total body fat of 1.15 kg compared to the control group (p &lt; 0.05) and the change adjusted for sex, school and baseline body fat was different from the control group and from the group with 1 dose or two of RTEC. Boys reduced total body fat 1.3% more than girls did (p &lt; 0.05) (Data not shown). Unadjusted and adjusted changes in indicators of body composition in the two RTEC groups that did not receive any nutrition education were not different compared with the control group.</p>", "<p>The effect of different treatments on blood lipids is shown in table ##TAB##3##4##. Only children that had RTEC and nutrition education showed a significant reduction in triglycerides (p &lt; 0.05), an increase in HDL (p &lt; 0.01) and a small reduction in VLDL (p &lt; 0.05). Changes in the other groups were not statistically significant. Comparison of unadjusted changes among groups showed that only HDL increased significantly in the group with RTEC and nutrition education compared to the control group. Treatment changes adjusted for baseline value and school were not different from the control group.</p>", "<p>Changes ± standard deviation (SD) in intense, moderate and low physical activities in Mets/week were the following: for with 1 dose of RTEC, 13.4 ± 41.3, 4.3 ± 12.7, 3 ± 18, for group with 2 doses of RTEC 2.4 ± 61.5, -2.2 ± 17.1, 2.6 ± 31.9, for with 1 dose of RTEC + education guide -3.6 ± 52.4, -0.2 ± 16, -5 ± 19 and for the control group 4.6 ± 31.6, -1.1 ± 9.9, 6.7 ± 18.5. Changes were not statistically different neither between basal and final evaluations nor among experimental groups. When adjusting model for gender and school, boys increased their intense physical activity while girls decreased it resulting in a significant difference between boys and girls (8.8 ± 60.2 vs 12.6 ± 85.7).</p>" ]

[ "<title>Discussion</title>", "<p>Although there are many environmental factors promoting excess energy intake, consumption of high fat diets increases the likelihood of obesity and the risk of obesity is lower in individuals consuming low fat diets. A review of clinical trials and animal studies [##REF##10721889##22##] suggests that the Public Health recommendation to lower dietary fat intake continues to be an appropriate measure to reduce energy intake and obesity. Fat compared with carbohydrates and proteins, increases the energy density of foods and diets.. Thus, a logical suggestion has been to replace fat with carbohydrate and therefore, decrease the energy density of the diet [##UREF##3##5##].</p>", "<p>The present study showed that the increase in RTEC consumption as a source of carbohydrate in children was effective in reducing body weight and body fat only when a nutrition education guide was included as part of the treatment. The inclusion of either one or two servings of RTEC in the diet without nutrition education was not effective in reducing body fat and did not cause any significant change in body weight compared with the control no-treatment group. Kirk et al [##UREF##3##5##] found a significant reduction of 2 kg body weight in 29 adults that replaced one meal with a serving of RTEC everyday during 4 weeks as a high carbohydrate regime. Differences between our study and this study include the difference in the population studied but more important is that the study by Kirk et al [##UREF##3##5##] did not include a control group. This makes its conclusion about the effectiveness of increasing carbohydrate consumption as an effective approach to treat obesity weak. Rodearmel et al. [##REF##16988082##23##] studied the impact of increasing 2 serving of RTEC/day and increasing daily steps in a 13-week intervention study as a family-based approach to prevent obesity and found significant differences in children's BMI-for-age and body fat between the experimental and the control groups. The control group in this study did not receive any intervention, therefore, the effect of the RTEC seen cannot be separated from the increase in physical activity in the children.</p>", "<p>Our study agrees with other studies [##UREF##3##5##,##REF##16988082##23##] in the fact that an increase in RTEC consumption as a source of carbohydrates was shown to be an effective strategy to lose weight in obese children but our study suggests that only when it is given with nutrition education. The change in body weight in the group receiving education and RTEC was accompanied with a reduction in total body fat. These changes did not occur in the groups that received one or two servings of RTEC and that the mother and child did not receive any nutrition education guide. These findings suggest that in our population a nutrition education guideline might be necessary for the beneficial effects of increasing carbohydrate consumption.</p>", "<p>The importance of education programs in the treatment of obesity has been known for a number of years, but only recently it has been suggested that nutrition education should be part of any successful strategy to reduce obesity in children [##REF##11494647##24##, ####REF##12094001##25##, ##UREF##14##26##, ##REF##11274525##27##, ##REF##11494642##28##, ##REF##11707548##29####11707548##29##], adolescents [##REF##11274525##27##, ####REF##11494642##28##, ##REF##11707548##29##, ##UREF##15##30####15##30##] and in adults [##REF##14993861##31##,##REF##12468634##32##]. Also, nutrition education has proven to be effective in improving nutritional status of individuals in different populations at risk. Studies of nutrition education programs that are continuous, specific, focused and targeted to vulnerable populations have been successful in improving nutritional status [##REF##15333726##33##, ####REF##16549488##34##, ##REF##10721928##35##, ##REF##10801920##36##, ##REF##1274890##37####1274890##37##]. Our study suggests that providing a nutrition education guide makes dietary changes, such as increase in carbohydrate consumption, more effective, and that a lack of an adequate nutritional education can cause nutritional strategies to fail.</p>", "<p>It is important to consider that since we did not include a group receiving nutrition education alone, we are unable to conclude that the group receiving RTEC in addition to a nutrition education program works any better than nutrition education alone to increase carbohydrate intake. The study was not designed to test the effect of a nutrition education program without the increase in RTEC consumption.</p>", "<p>The reduction in body fat and body weight in the RTEC and nutrition education group of the children was accompanied by a significant reduction in plasma triglycerides and by an elevation in HDL. Changes in these two variables are clearly associated with a reduction in body fat and are beneficial to reduce health complications associated with excess of body fat.</p>" ]

[ "<title>Conclusion</title>", "<p>We found that a strategy to increase carbohydrate consumption to reduce obesity in children is effective only when accompanied with a nutrition education program. An increase in RTEC intake as a source of carbohydrates with a simple nutrition education guideline produced a significant loss of body weight, a decrease in body fat and in plasma triglycerides, and an increase in high density lipoproteins. The importance of nutrition education could be extrapolated to other nutritional manipulations intended for treatment of obesity.</p>" ]

[ "<p>This is an Open Access article distributed under the terms of the Creative Commons Attribution License (<ext-link ext-link-type=\"uri\" xlink:href=\"http://creativecommons.org/licenses/by/2.0\"/>), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.</p>", "<title>Background</title>", "<p>The main emphasis of dietary advice for control of obesity has been on reducing dietary fat. Increasing ready to eat cereal (RTEC) consumption could be a strategy to reduce fat intake and increase carbohydrate intake resulting in a diet with lower energy density.</p>", "<title>Objectives</title>", "<p>1. To determine if an increase in RTEC intake is an effective strategy to reduce excess body weight and blood lipids in overweight or at risk of overweight children. 2. To determine if a nutrition education program would make a difference on the response to an increase in cereal intake. 3) To determine if increase in RTEC intake alone or with a nutrition education program has an effect on plasma lipid profile.</p>", "<title>Experimental design</title>", "<p>One hundred and forty seven overweight or at risk of overweight children (6–12 y of age) were assigned to one of four different treatments: a. One serving of 33 ± 7 g of RTEC for breakfast; b. one serving of 33 ± 7 g of RTEC for breakfast and another one for dinner; c. one serving of 33 ± 7 g of RTEC for breakfast and a nutrition education program. d. Non intervention, control group. Anthropometry, body composition, physical activity and blood lipids were measured at baseline, before treatments, and 12 weeks after treatments.</p>", "<title>Results</title>", "<p>After 12 weeks of intervention only the children that received 33 ± 7 g of RTEC and nutrition education had significantly lower body weight [-1.01 (-1.69, -0.34) ], p &lt; 0.01], lower BMI [-0.95 (-1.71, -0.20), p &lt; 0.01] and lower total body fat [-0.71 (-1.71, 0.28), p &lt; 0.05] compared with the control group [1.19 (0.39, 1.98), 0.01 (-0.38, 0.41), 0.44 (-0.46, 1.35) respectively]. Plasma triglycerides and VLDL were significantly reduced [-20.74 (-36.44, -5.05), -3.78 (-6.91, -0.64) respectively, p &lt; 0.05] and HDL increased significantly [6.61 (2.15, 11.08), p &lt; 0.01] only in this treatment group. The groups that received 1 or 2 doses of RTEC alone were not significantly different to the control group.</p>", "<title>Conclusion</title>", "<p>A strategy to increase RTEC consumption, as a source of carbohydrate, to reduce obesity is effective only when accompanied by nutrition education. The need for education could be extrapolated to other strategies intended for treatment of obesity.</p>", "<title>Trial Registration</title>", "<p>Australian New Zealand Clincial Trial Registry. Request no: ACTRN12608000025336</p>" ]

[ "<title>Consent</title>", "<p>Written informed consent was obtained from the parents for participation and publication of results. A copy of the written consent is available for review by the Editor-in-Chief of this journal. This study was approved by the Internal Committee of Human Research of the University of Queretaro.</p>", "<title>Competing interests</title>", "<p>The authors declare that they have no competing interests.</p>", "<title>Authors' contributions</title>", "<p>JLR developed the study design, supervised the study, and made a substantial contribution with interpretation of data, drafting the manuscript and revising it critically for important intellectual content. MRA participated in the study design and coordinated the field research. KM coordinated the field research and supervised the quality of data collection. OPG contributed with the study conception and design, interpretation of data and writing the publication. And MCC participated in managing the data, carried out the statistical analysis and contributed to revising the manuscript. All authors read and approved the final manuscript.</p>" ]

[ "<title>Acknowledgements</title>", "<p>We would like to thank Juana Ramirez Anguiano, Paola García Juarez and Abigail Dominguez Chavero for their dedicated work, and to Elba Suaste for her special touch when taking blood from children and her assistance with laboratory work. We are grateful to Kellogg Company México S.A de C.V. that contributed with a scientific grant for partial funding of the study.</p>" ]

[ "<fig position=\"float\" id=\"F1\"><label>Figure 1</label><caption><p>Flow-chart.</p></caption></fig>" ]

[ "<table-wrap position=\"float\" id=\"T1\"><label>Table 1</label><caption><p>Summary of the nutrition education guide used in one treatment group</p></caption><table frame=\"hsides\" rules=\"groups\"><tbody><tr><td align=\"left\"><underline>Recommendations concerning the family:</underline></td></tr><tr><td align=\"left\"> • Parents are responsible for teaching their children healthy food choices in and outside their home.</td></tr><tr><td align=\"left\"> • Lunch must be a simple, appetizing, easy to carry, economical and nutritious meal.</td></tr><tr><td align=\"left\"> • When eating together, the rest of the family should eat the same type of foods, but the serving size may vary individually.</td></tr><tr><td align=\"left\"> • Mealtimes at home should be calmed and trouble-free; this is not a good time for arguments about the child's diet.</td></tr><tr><td align=\"left\"> • The amount of foods eaten at home is influenced by family preferences. Watch out what is bought and stored at home. Avoid storing foods that the child may crave for such as deserts, soft drinks, candies, potato chips and other calorie dense foods.</td></tr><tr><td align=\"left\"><underline>Recommendations concerning the child</underline></td></tr><tr><td align=\"left\"> • The child must continue with his/her usual physical activity.</td></tr><tr><td align=\"left\"> • Food preferences should be considered when planning the child's meals.</td></tr><tr><td align=\"left\"> • A child must always have breakfast before school, or during the weekend.</td></tr><tr><td align=\"left\"> • Consume only skimmed milk, low-fat cheese and low-fat yogurt.</td></tr><tr><td align=\"left\"> • Eat the regular foods prepared at home, following the general recommendations given in the sample menu.</td></tr><tr><td align=\"left\"> • If there is a mealtime out of home, child may be allowed to eat the foods available, but the amount of food consumed should be less than usual.</td></tr><tr><td align=\"left\"> • Avoid beverages with a high content of sugar; instead drink natural water.</td></tr><tr><td align=\"left\"> • The child should replace snacks with low-sugar beverages or water and may have a calorie dense snack of his/her choice occasionally (once a week).</td></tr><tr><td align=\"left\"><underline>Foods to be included in the child's diet:</underline></td></tr><tr><td align=\"left\"> • Whole grain breads, pasta, rice, cereal.</td></tr><tr><td align=\"left\"> • Turkey ham, turkey sausages, chicken, tuna, eggs and beans</td></tr><tr><td align=\"left\"> • Low-fat milk, cheese, yogurt</td></tr><tr><td align=\"left\"> • Lettuce, tomato, carrots</td></tr><tr><td align=\"left\"> • Any kind of fruit</td></tr><tr><td align=\"left\"><underline>Foods to be avoided in the child's diet:</underline></td></tr><tr><td align=\"left\"> • Foods with a high sugar content such as soft drinks, candy, commercial fruit juices, and chocolate.</td></tr><tr><td align=\"left\"> • Foods with a high fat content, such as cream, desserts made from whole milk, peanut butter, fried food, pork and lamb meat and their products such as bacon and pork sausage.</td></tr><tr><td align=\"left\"><underline>Example of a lunch:</underline></td></tr><tr><td align=\"left\"> • A sandwich with one item (low fat cheese, turkey ham, or tuna fish in water) + 1 fruit or vegetable + natural water or one glass of a low-sugar beverage and the rest of beverages as natural water.</td></tr></tbody></table></table-wrap>", "<table-wrap position=\"float\" id=\"T2\"><label>Table 2</label><caption><p>Characteristics of subjects in experimental groups at baseline *†</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td/><td align=\"center\">1 dose of RTEC</td><td align=\"center\">2 doses of RTEC</td><td align=\"center\">1 dose of RTEC + Nutrition education guide</td><td align=\"center\">Control</td></tr></thead><tbody><tr><td align=\"left\">N</td><td align=\"center\">46</td><td align=\"center\">48</td><td align=\"center\">45</td><td align=\"center\">39</td></tr><tr><td align=\"left\">Males %</td><td align=\"center\">56.4</td><td align=\"center\">40.5</td><td align=\"center\">47.5</td><td align=\"center\">51.6</td></tr><tr><td align=\"left\">Females %</td><td align=\"center\">43.6</td><td align=\"center\">59.5</td><td align=\"center\">52.5</td><td align=\"center\">48.4</td></tr><tr><td align=\"left\">Age (m)</td><td align=\"center\">110.3 ± 19.7</td><td align=\"center\">109.3 ± 18.9</td><td align=\"center\">107.8 ± 18.8</td><td align=\"center\">110.1 ± 18.9</td></tr><tr><td align=\"left\">Anthropometry:</td><td/><td/><td/><td/></tr><tr><td align=\"left\"> Weight (Kg)</td><td align=\"center\">47.0 ± 12.9</td><td align=\"center\">47.7 ± 12.7</td><td align=\"center\">46.4 ± 12.2</td><td align=\"center\">48.2 ± 11.7</td></tr><tr><td align=\"left\"> Height (Cm)</td><td align=\"center\">139.2 ± 12.1</td><td align=\"center\">139.01 ± 9.4</td><td align=\"center\">138.2 ± 10.8</td><td align=\"center\">139.8 ± 11.4</td></tr><tr><td align=\"left\"> BMI (Kg/M²)</td><td align=\"center\">23.7 ± 3.3</td><td align=\"center\">24.3 ± 3.7</td><td align=\"center\">23.8 ± 3.1</td><td align=\"center\">24.3 ± 3.1</td></tr><tr><td align=\"left\"> Height for age (Z-score)</td><td align=\"center\">0.4 ± 0.8</td><td align=\"center\">0.5 ± 0.8</td><td align=\"center\">0.5 ± 0.9</td><td align=\"center\">0.6 ± 1.1</td></tr><tr><td align=\"left\"> Weight for age (Z-score)</td><td align=\"center\">2.1 ± 1.0</td><td align=\"center\">2.4 ± 1.4</td><td align=\"center\">2.2 ± 1.0</td><td align=\"center\">2.4 ± 1.0</td></tr><tr><td align=\"left\"> Weight for height (Z-score)</td><td align=\"center\">2.9 ± 1.1</td><td align=\"center\">3.0 ± 1.0</td><td align=\"center\">2.8 ± 0.7</td><td align=\"center\">3.0 ± 0.9</td></tr><tr><td align=\"left\">Blood lipids:</td><td/><td/><td/><td/></tr><tr><td align=\"left\"> N</td><td align=\"center\">27</td><td align=\"center\">36</td><td align=\"center\">34</td><td align=\"center\">32</td></tr><tr><td align=\"left\"> Total Cholesterol (mg/dL)</td><td align=\"center\">141.3 ± 31.3</td><td align=\"center\">140.6 ± 32.9</td><td align=\"center\">127.4 ± 23.3</td><td align=\"center\">138.8 ± 32.9</td></tr><tr><td align=\"left\"> Triglycerides (mg/dL)</td><td align=\"center\">108.6 ± 45.2</td><td align=\"center\">132.2 ± 46.4</td><td align=\"center\">130.2 ± 47.7</td><td align=\"center\">125.1 ± 45.1</td></tr></tbody></table></table-wrap>", "<table-wrap position=\"float\" id=\"T3\"><label>Table 3</label><caption><p>Effect of treatments on anthropometry and body composition in the different groups *</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td/><td align=\"center\">1 dose of RTEC</td><td align=\"center\">2 doses of RTEC</td><td align=\"center\">1 dose of RTEC + Nutrition education guide</td><td align=\"center\">Control</td></tr></thead><tbody><tr><td align=\"left\">N</td><td align=\"center\">46</td><td align=\"center\">48</td><td align=\"center\">45</td><td align=\"center\">39</td></tr><tr><td align=\"left\">Weight (Kg)</td><td/><td/><td/><td/></tr><tr><td align=\"left\">Initial</td><td align=\"center\">47.0 (43.0, 51.1)</td><td align=\"center\">47.7 (43.6, 51.8)</td><td align=\"center\">47.0 (43.2, 50.8)</td><td align=\"center\">48.2 (44.04, 52.3)</td></tr><tr><td align=\"left\">Final</td><td align=\"center\">47.92 (43.9, 52.0)</td><td align=\"center\">48.6 (44.6, 52.7)</td><td align=\"center\">46.08 (42.5, 49.7)</td><td align=\"center\">49.30 (45.2, 53.4)</td></tr><tr><td align=\"left\">Unadjusted change</td><td align=\"center\">0.9 (0.4, 1.4) ‡</td><td align=\"center\">0.9 (0.3, 1.5) ‡</td><td align=\"center\">-0.9 (-2.2, 0.5) §</td><td align=\"center\">1.2 (0.8, 1.5) ‡</td></tr><tr><td align=\"left\">Adjusted change †</td><td align=\"center\">1.03 (0.3, 1.7)</td><td align=\"center\">0.6 (-0.1, 1.3)</td><td align=\"center\">-1.01 (-1.7, -0.3) **</td><td align=\"center\">1.2 (0.4, 2.0)</td></tr><tr><td align=\"left\">BMI(Kg/M²)</td><td/><td/><td/><td/></tr><tr><td align=\"left\">Initial</td><td align=\"center\">23.7 (22.7, 24.8)</td><td align=\"center\">24.3 (23.1, 25.5)</td><td align=\"center\">24.1 (23.1, 25.2)</td><td align=\"center\">24.3 (23.2, 25.4)</td></tr><tr><td align=\"left\">Final</td><td align=\"center\">23.8 (22.6, 24.9)</td><td align=\"center\">24.1 (22.8, 25.3)</td><td align=\"center\">23.2 (22.3, 24.1)</td><td align=\"center\">24.3 (23.2, 25.4)</td></tr><tr><td align=\"left\">Unadjusted change</td><td align=\"center\">0.04 (-0.3, 0.4)</td><td align=\"center\">-0.2 (-0.5, 0.1)</td><td align=\"center\">-1.0 (-1.7, -0.2) ‡, §</td><td align=\"center\">0.02 (-0.1, 0.2)</td></tr><tr><td align=\"left\">Adjusted change †</td><td align=\"center\">0.1 (-0.3, 0.4)</td><td align=\"center\">-0.3 (-0.7, 0.1)</td><td align=\"center\">-0.9 (-1.2, -0.5) **</td><td align=\"center\">0.01 (-0.4, 0.4)</td></tr><tr><td align=\"left\">Total Body Fat (%)</td><td/><td/><td/><td/></tr><tr><td align=\"left\">Initial</td><td align=\"center\">23.6 (20.6, 26.6)</td><td align=\"center\">25.9 (22.8, 28.9)</td><td align=\"center\">24.4 (21.6, 27.3)</td><td align=\"center\">27.1 (23.9, 30.2)</td></tr><tr><td align=\"left\">Final</td><td align=\"center\">24.1 (20.9, 27.2)</td><td align=\"center\">25.5 (22.5, 28.5)</td><td align=\"center\">23.7 (20.7, 26.7)</td><td align=\"center\">27.5 (24.5, 30.5)</td></tr><tr><td align=\"left\">Unadjusted change</td><td align=\"center\">0.5 (-0.1, 1.1)</td><td align=\"center\">-0.4 (-1.0, 0.3)</td><td align=\"center\">-0.7 (-1.7, 0.3) §</td><td align=\"center\">0.4 (-0.4, 1.2)</td></tr><tr><td align=\"left\">Adjusted change †</td><td align=\"center\">0.4 (-0.4, 1.1)</td><td align=\"center\">-0.5 (-1.3, 0.3)</td><td align=\"center\">-0.8 (-1.6, -0.04) ††</td><td align=\"center\">0.4 (-0.5, 1.4)</td></tr></tbody></table></table-wrap>", "<table-wrap position=\"float\" id=\"T4\"><label>Table 4</label><caption><p>Effect of treatments on plasma lipids in the different groups *</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td/><td align=\"center\">1 dose of RTEC</td><td align=\"center\">2 doses of RTEC</td><td align=\"center\">1 dose of RTEC + Nutrition education guide</td><td align=\"center\">Control</td></tr></thead><tbody><tr><td align=\"left\">N</td><td align=\"center\">32</td><td align=\"center\">34</td><td align=\"center\">36</td><td align=\"center\">27</td></tr><tr><td align=\"left\">Total Cholesterol</td><td/><td/><td/><td/></tr><tr><td align=\"left\">Initial</td><td align=\"center\">143.3 (132.6, 154.1)</td><td align=\"center\">141.3 (130.0, 152.7)</td><td align=\"center\">128.6 (121.1, 136.1)</td><td align=\"center\">134.6 (123.3, 145.8)</td></tr><tr><td align=\"left\">Final</td><td align=\"center\">149.6 (138.8, 160.3)</td><td align=\"center\">147.5 (135.7, 159.4)</td><td align=\"center\">136.8 (128.0, 145.6)</td><td align=\"center\">141.3 (132.0, 150.7)</td></tr><tr><td align=\"left\">Unadjusted change</td><td align=\"center\">6.2 (-7.3, 19.7)</td><td align=\"center\">6.2 (-8.0, 20.4)</td><td align=\"center\">8.2 (-3.4, 19.8)</td><td align=\"center\">6.7 (-5.8, 19.3)</td></tr><tr><td align=\"left\">Adjusted change †</td><td align=\"center\">14.7 (4.5, 24.9)</td><td align=\"center\">14.2 (4.3, 24.1)</td><td align=\"center\">9.5 (-0.7, 19.6)</td><td align=\"center\">6.2 (-4.7, 17.1)</td></tr><tr><td align=\"left\">Triglycerides</td><td/><td/><td/><td/></tr><tr><td align=\"left\">Initial</td><td align=\"center\">109.5 (92.9, 126.1)</td><td align=\"center\">134.2 (118.1, 150.2)</td><td align=\"center\">129.5 (113.4, 145.6)</td><td align=\"center\">121.9 (106.6, 137.1)</td></tr><tr><td align=\"left\">Final</td><td align=\"center\">134.5 (109.7, 159.2)</td><td align=\"center\">119.4 (102.9, 135.8)</td><td align=\"center\">108.7 (92.8, 124.6)</td><td align=\"center\">121.6 (102.2, 141.0)</td></tr><tr><td align=\"left\">Unadjusted change</td><td align=\"center\">25.0 (-3.6, 53.6)</td><td align=\"center\">-14.8 (-31.8, 2.2)</td><td align=\"center\">-20.7 (-36.4, -5.1) ‡</td><td align=\"center\">-0.2 (-19.3, 18.8)</td></tr><tr><td align=\"left\">Adjusted change †</td><td align=\"center\">13.5 (-6.5, 33.4)</td><td align=\"center\">-10.3 (-29.0, 8.4)</td><td align=\"center\">-18.1 (-36.7, 0.6)</td><td align=\"center\">-4.3 (-24.6, 16.0)</td></tr><tr><td align=\"left\">HDL cholesterol</td><td/><td/><td/><td/></tr><tr><td align=\"left\">Initial</td><td align=\"center\">48.4 (43.9, 52.8)</td><td align=\"center\">48.1 (44.3, 51.8)</td><td align=\"center\">43.1 (39.3, 47.0)</td><td align=\"center\">47.5 (42.4, 52.6)</td></tr><tr><td align=\"left\">Final</td><td align=\"center\">47.0 (42.4, 51.6)</td><td align=\"center\">48.5 (44.8, 52.2)</td><td align=\"center\">49.7 (46.5, 53.0)</td><td align=\"center\">44.8 (40.5, 49.1)</td></tr><tr><td align=\"left\">Unadjusted change</td><td align=\"center\">-1.4 (-7.3, 4.6)</td><td align=\"center\">0.4 (-4.6, 5.5)</td><td align=\"center\">6.6 (2.2, 11.1) ‡, §</td><td align=\"center\">-2.7 (-6.5, 1.1)</td></tr><tr><td align=\"left\">Adjusted change †</td><td align=\"center\">-2.2 (-5.7, 1.4)</td><td align=\"center\">1.0 (-2.4, 4.3)</td><td align=\"center\">1.7 (-1.7, 5.1)</td><td align=\"center\">-3.0 (-6.7, 0.7)</td></tr><tr><td align=\"left\">LDL cholesterol</td><td/><td/><td/><td/></tr><tr><td align=\"left\">Initial</td><td align=\"center\">122.6 (111.0, 134.3)</td><td align=\"center\">123.6 (112.2, 135.1)</td><td align=\"center\">114.1 (104.9, 123.4)</td><td align=\"center\">116.7 (106.7, 126.8)</td></tr><tr><td align=\"left\">Final</td><td align=\"center\">137.1 (123.5, 150.6)</td><td align=\"center\">125.2 (112.9, 137.5)</td><td align=\"center\">112.3 (101.3, 123.2)</td><td align=\"center\">125.4 (114.4, 136.4)</td></tr><tr><td align=\"left\">Unadjusted change</td><td align=\"center\">14.4 (1.0, 27.8) ‡</td><td align=\"center\">1.6 (-11.8, 14.9)</td><td align=\"center\">-1.9 (-15.7, 12.0)</td><td align=\"center\">8.7 (-5.0, 22.3)</td></tr><tr><td align=\"left\">Adjusted change †</td><td align=\"center\">19.3 (7.3, 31.3)</td><td align=\"center\">9.2 (-2.3, 20.7)</td><td align=\"center\">1.8 (-10.0, 13.5)</td><td align=\"center\">8.0 (-4.8, 20.8)</td></tr><tr><td align=\"left\">VLDL cholesterol</td><td/><td/><td/><td/></tr><tr><td align=\"left\">Initial</td><td align=\"center\">21.9 (18.6, 25.2)</td><td align=\"center\">26.8 (23.6, 30.0)</td><td align=\"center\">25.7 (22.6, 28.9)</td><td align=\"center\">24.4 (21.3, 27.4)</td></tr><tr><td align=\"left\">Final</td><td align=\"center\">26.9 (22.0, 31.9)</td><td align=\"center\">23.9 (20.6, 27.2)</td><td align=\"center\">21.9 (18.8, 25.1)</td><td align=\"center\">24.3 (20.5, 28.2)</td></tr><tr><td align=\"left\">Unadjusted change</td><td align=\"center\">5.0 (-0.7, 10.7)</td><td align=\"center\">-3.0 (-6.4, 0.5)</td><td align=\"center\">-3.8 (-6.9, -0.6) ‡</td><td align=\"center\">-0.04 (-3.9, 3.8)</td></tr><tr><td align=\"left\">Adjusted change †</td><td align=\"center\">2.6 (-1.4, 6.6)</td><td align=\"center\">-2.0 (-5.8, 1.7)</td><td align=\"center\">-3.3 (-7.0, 0.4)</td><td align=\"center\">-0.8 (-4.9, 3.2)</td></tr></tbody></table></table-wrap>" ]

[]

[]

[]

[]

[]

[]

[ "<table-wrap-foot><p>* Values are means ± standard deviation unless otherwise is mentioned.</p><p>† No statistical significant difference among groups was found</p></table-wrap-foot>", "<table-wrap-foot><p>* Values are means (95% Confidence Interval).</p><p>† Estimated mean change adjusted for initial value, gender, school random effect and significant interactions.</p><p>‡Difference between initial and final is significant at p &lt; 0.05 in paired T-Test.</p><p>§ Change is different from the control group at p &lt; 0.05 in ANOVA</p><p>** Estimated change from group with 1 dose of RTEC + Nutrition education guide is different to all other groups, at p &lt; 0.05 in Adjusted ANOVA</p><p>†† Estimated change from group with 1 dose of RTEC + Nutrition education guide is different to group with 1 dose of RTEC and to control group, at p &lt; 0.05 in Adjusted ANOVA</p></table-wrap-foot>", "<table-wrap-foot><p>* Values are means (95% CI) in mg/dL.</p><p>† Estimated mean change adjusted for initial value, gender, school random effect and significant interactions.</p><p>‡ Difference between initial and final significant at p &lt; 0.05 in paired T-test.</p><p>§ Change in group with 1 dose of RTEC + Nutrition education guide is different to change in group with 1 dose of RTEC and to the control group at p &lt; 0.01 in ANOVA.</p></table-wrap-foot>" ]

[ "<graphic xlink:href=\"1475-2891-7-28-1\"/>" ]

[]

{ "acronym": [], "definition": [] }

[ "<title>Background</title>", "<p>Malaria represents about 1.4% of the global burden of disease [##UREF##0##1##] and in Africa, it is the primary cause of disease burden as measured by Disability Adjusted Life Years (DALY) lost of 10.8% [##UREF##1##2##,##REF##15331814##3##]. The continent bears over 90% of the global burden of about 2.7 million deaths attributable to malaria; and houses over 300 million people who suffer from this disease yearly, the worst hit being young children and pregnant women [##UREF##1##2##, ####REF##15331814##3##, ##REF##11425184##4####11425184##4##].</p>", "<p>More than three quarter of global malaria deaths occur in under-five children living in malarious countries in sub-Saharan Africa (SSA) [##UREF##2##5##], where 25% of all childhood mortality below the age of five (about 800,000 young children [##UREF##3##6##]) is attributable to malaria [##UREF##1##2##]. Of those children who survive cerebral malaria, a severe form of the disease, more than 15% suffer neurological deficits [##REF##11425184##4##,##REF##11425178##7##], which include weakness, spasticity, blindness, speech problems and epilepsy. Where such children are poorly managed and do not have access to specialized educational facilities, these deficits may interfere with future learning and development [##UREF##2##5##].</p>", "<p>About 30–40% of all fevers seen in health centres in Africa are due to malaria with huge seasonal variability between rainy and dry seasons. At the end of the dry season, it is less than 10% and more than 80% as the rainy season winds up [##UREF##4##8##].</p>", "<p>In Nigeria, malaria is the leading cause of under-five mortality contributing 33% of all childhood deaths and 25% infant mortality. As a child will typically be sick of malaria between 3–4 times in one year, the disease is a major cause of absenteeism in school-aged children, thus impeding their educational and social development [##UREF##2##5##] and subsequently robbing the country of its future human resources.</p>", "<p>Several global and regional attempts have been made at controlling the disease in the past with little success as a result of ineffective strategies used and insufficient resources. However, the most recent launching of the Roll Back Malaria initiative has generated a lot of resources for the control of the disease with simple and cost-effective interventions, with a special focus on the most at risk. At the historic Malaria Summit hosted by Nigeria in 2000, African Heads of States made a declaration to halve the burden of malaria by the year 2010. One of the targets set for the first five years was to ensure that the vulnerable groups, children under five years of age and pregnant women, have access to and sleep under insecticide-treated nets (ITNs) [##UREF##5##9##].</p>", "<p>Several studies have demonstrated the efficacy of ITNs [##REF##9684633##10##, ####REF##1675368##11##, ##REF##11157527##12##, ##REF##11418148##13##, ##REF##12749481##14##, ##REF##8944251##15##, ##REF##12749495##16##, ##REF##1680284##17##, ##REF##15106149##18####15106149##18##]. Between 1986 and 2002, at least 81 trials and over 30 descriptive studies carried out in every type of malaria setting worldwide have documented the positive impact of ITNs on child and adult morbidity and mortality [##REF##11418148##13##]. Most of these studies were summarized by Lengeler in 2004. He found that in five randomized controlled trials (RCTs) an overall reduction in child mortality of 17% could be demonstrated, with six lives saved per year for every 1,000 children protected. The use of ITNs in areas with stable malaria reduced the incidence of uncomplicated episodes by 50% compared to areas where nets were not used, and 39% compared to areas were the nets were untreated. ITNs also impacted on severe malaria, parasite prevalence, high parasitaemia, splenomegaly and improvement in haemoglobin levels of children [##REF##15106149##18##]. A reduction of 27% in child mortality was also demonstrated in an ITN social marketing programme in Tanzania [##REF##11157527##12##,##REF##11418148##13##]. This overwhelming evidence of the efficacy of utilization of ITN was the basis of its adoption as one of the four global Roll Back Malaria (RBM) strategies for malaria control [##UREF##1##2##].</p>", "<title>Ownership versus utilization of ITN</title>", "<p>Two important RBM indicators for monitoring progress towards the set target are the proportion of households which own one or more nets and the proportion of under-five children who sleep under a net [##UREF##6##19##]. Net ownership is important to assess the effectiveness of the distribution channels of the RBM programme and suggest programme modifications where there are lapses. However, utilization is the crucial indicator that will generate the desired epidemiological impact [##REF##16772004##20##].</p>", "<p>Few studies have examined the difference between the two indicators. A meta-analysis of household surveys on net utilization and ownership found a wide gap between net possession and use. ITN ownership was found to be between 0.1% and 28.5%, while use among children less than five years of age ranged between 0% and 16% [##REF##12869090##21##].</p>", "<p>Equality is a major issue in ITN ownership. Net ownership has been found to be lowest among the poorest households [##UREF##7##22##]; thus possibly linking possession to the cost of the net [##REF##12358619##23##]. Authors of a study conducted on the effect of lowering tariffs on nets and netting materials predict that reducing tariffs on insecticides and ITNs from 42% to 0% and the tariff on netting materials from 40% to 5% would increase demand for ITNs by 9–27% [##REF##12481212##24##]. Wiseman <italic>et al </italic>reported a significant association between good access roads to the community and net ownership [##REF##17488900##25##]. Perceived risk of malaria and benefits of the nets by the population also drive demand. Onwujekwe <italic>et al</italic>, in a Nigerian study, found that households with a recent attack of malaria and those with higher willingness to pay were more likely to purchase a net than their counterparts [##REF##12917269##26##]</p>", "<p>Utilization has, however, been found to vary with seasons of the year and acceptability of the nets in terms of size, colour and shape. Binka <italic>et al </italic>showed that the time of the year during which the nets are delivered affects use. 99% of the net recipients were found to use the nets during rainy season, while only 20% used it during the dry season [##REF##9217706##27##]. Demographic characteristics like age, education, size of household, and ethnicity also influence use of bed nets. Some studies show that children are less likely to use nets [##REF##8212106##28##,##UREF##8##29##], particularly in rural areas [##UREF##7##22##], while others found no significant association between age and net use [##REF##10475665##30##].</p>", "<p>Few studies have documented household net coverage and utilization in Nigeria. The 2003 National Demographic Health Survey reported a 12% household ownership of any net and 2% of ITN. Under-five utilization of ITN was 1.2% while 5.9% of them used any net [##UREF##9##31##]. NetMark, a United States Agency for International Development (USAID)-funded project, conducted a study in 2004 and compared findings with an older one done by the same organization in 2000 in five states in the country. Overall household ownership of any net was found to have more than doubled, from 12% in 2000 to 27% in 2004, while ITN ownership increased from 0 to 9%. The study also documented an increase in utilization of ITN by under-five children compared to previous years to 3.3% [##UREF##10##32##]. In both studies, rural households were more likely to have a net than urban households.</p>", "<p>The President of Nigeria launched an ITN Massive Promotion and Awareness Campaign (IMPAC) and made provision of free ITNs for the vulnerable groups a priority on commencement of implementation of the RBM programme. With support from international donor agencies, ITNs (either re-treatable nets bundled with re-treatment kits or long-lasting insecticidal nets) were procured and distributed.</p>", "<p>ITNs were given at immunization posts to children completing their immunization schedule; during stand-alone ITN campaigns in specific rural local government areas; and by means of the Expanded Programme on Immunization, EPI-plus (giving measles vaccine plus ITN) in tandem with supplemental immunization campaigns to saturate the population with the nets in line with WHO's recommendation [##UREF##11##33##].</p>", "<p>The huge cost implication of the Abuja Declaration and the limitedness of available resources [##REF##12787469##34##, ####UREF##12##35##, ##UREF##13##36####13##36##] coupled with the enormous pressure to meet set targets within limited-time spans puts a prerogative on sound monitoring and periodic evaluation. This is crucial for identifying gaps in programme implementation or areas where modifications in specific technical strategies may be needed, where resources should be focused, assessing progress or otherwise and providing feedback to inform future planning [##UREF##14##37##].</p>", "<p>In addition, the United Nations Millennium Development Goals (MDGs) 4, 5 and 6 are directly linked to malaria control, while 1 and 2 are indirectly related. The achievement of the MDGs depends, therefore, on the success of the RBM initiative. An assessment of the country's performance over the first five years of implementation would also give an indication of where the country is with regard to achieving the MDGs [##UREF##15##38##].</p>", "<p>Since the declaration was made, studies have been done to evaluate coverage, acceptability, willingness to pay for and effectiveness of all the interventions in Nigeria [##UREF##9##31##,##UREF##10##32##,##REF##16704735##39##, ####REF##15023234##40##, ##REF##17255221##41####17255221##41##]. However, only one of those which studied utilization of ITN analysed nationally representative data [##UREF##9##31##]. The sampling methodology used for the selection of the states included in the NetMark study was purposive, selective and not random; the sample was drawn only from areas where the agency was active with intensive distribution of nets and provision of targeted subsidies for the vulnerable groups and only households with under-five aged children were included in the survey. This selection bias affects the external validity of the study and makes it difficult to generalize the results to the whole country as positive results are likely to be overestimated.</p>", "<p>Although the NDHS survey was a larger study, it was carried out two years before the due time for the mid-term evaluation of the Abuja Targets. The results are, therefore, not likely to reflect the current coverage of ITN among children under-five.</p>", "<p>To inform policy and re-engineering of programme delivery to meet the set targets, evaluation should not just be limited to percentage coverage of ownership and use. It is important to investigate the predictors of utilization of ITN in a national survey. The two studies described above also failed to examine these associations being limited by the study design, which was descriptive rather than analytical.</p>", "<p>This study aims at bridging the current information gap on the status of implementation of the Abuja targets in Nigeria as regards the ownership and utilization of ITNs among the children under the age of five years and determining the factors that predict utilization. This will help identify gaps in programme implementation and provide a scientific basis for policy decisions on scaling up interventions where necessary.</p>" ]

[ "<title>Methods</title>", "<title>Study area and population</title>", "<p>The study was conducted in Nigeria, where malaria is endemic and transmission is stable all year round, with peaks during the rainy seasons which varies by regions; March-November in the south and July-September in the north. The population is huge (about 140 million) and culturally diverse with about 250 ethnic groups and 521 languages. However, the dominant tribes are Yoruba in the south-west, Ibo in the south-east and Hausa and Fulani in the north. The country is divided into six geopolitical regions namely; south-west zone (SWZ), south-east zone (SEZ), south-south zone (SSZ), north-west zone (NWZ), north-east (NEZ) and north-central zone (NCZ). About 70% of its population live in rural areas, where the predominant occupation is peasant farming and standard of living relatively poor. About 20% of the population are aged 0–5 years.</p>", "<title>Data collection</title>", "<p>Primary data collected in October 2005, during a national household survey on mid-term evaluation of the Abuja targets in Nigeria, was utilized for the analysis. A multistage stratified cluster sampling technique was employed at national, zonal, state and Local Government Areas (LGA) levels.</p>", "<p>The sampling frame, which consisted of all the 36 States + Federal Capital Territory, was divided into six strata (geopolitical zones) and two states from each of the country's six geo-political zones were randomly selected. In each selected state, three senatorial districts were chosen, from which one LGA each was randomly selected (one urban and two rural areas). At the LGA level, all wards were listed and grouped into communities with and without health facilities, and one community each was chosen at random from the communities with and without health facility. A fixed number of households, 100 per community (total of 7,200), was imposed on the design.</p>", "<p>Households were randomly selected by spinning a bottle in front of the market/community leader's house and starting from the direction in which the bottle pointed. Interviews were conducted by trained interviewers in the local languages using questionnaires adapted from the WHO Malaria Indicator Survey (MIS) tools.</p>", "<p>The household head (usually male) was interviewed for all information about the household, including net ownership and utilization, while the wife, or any eligible woman (aged 15–49 years) in the household was asked about her reproductive history and fever episodes among her under-five children if any. Where there was more than one eligible woman in the house, the one with under-five children was interviewed.</p>", "<p>There was no information on the required sample sizes for the women and children in the sampling methodology available from those who conducted this survey. However, using a design effect of 1.25 [##UREF##16##42##], requiring a 95% confidence interval (CI) length of 1% and a known ITN coverage of 1.2% for children under five and 1.4% for women aged 15–49 years, WINPEPI [##REF##15606913##43##] gave a required sample size of 2,277 and 2,651 for children under five and women aged 15–49 years respectively. The sample sizes for the two groups-6,390 eligible women and 3,585 eligible children-were, therefore, found to be adequate. To detect an acceptable difference of 0.5% in household ownership of nets, with a known current coverage of 2%, the required household sample size is 3,764; however, 5,588 households participated in the survey.</p>", "<p>Eligible children were defined as those who are under five and who stayed in the household the night before the interviewer's visit (<italic>de facto children</italic>), while eligible women were those who were in the specified age group (15–49 years) and who stayed in the household the night before the interviewer's visit (<italic>de facto women</italic>). A mosquito net was defined as an ITN if it was pre-factory-treated or has been dipped in insecticide within the last 6–12 months.</p>", "<title>Analysis</title>", "<p>Principal Component Analysis (PCA) [##UREF##17##44##] was used to develop wealth indices for the households based on ownership of durable assets including radio, television, telephone, refrigerator, bicycle, motorcycle/scooter and car/truck. Ownership was coded as 0 or 1 and missing cases were excluded. The households were then divided into socioeconomic quartiles based on their scores. In order to capture wealth differences between urban and rural residences, PCA scores were generated for the two areas separately (urban and rural household wealth indices) and for the combined data as well. The first dimension of the PCA was taken as the score for the household. Cronbach's alpha was calculated to test consistency-reliability. Infrastructural variables were not used to avoid 'urban bias' that could prevent comparism between rural and urban wealth indices; since urban areas are more likely to have higher quality of building and amenities. Internal coherence was tested comparing asset ownership by socioeconomic groups (quartiles) and this gave a clear trend of increasing asset ownership from lowest to highest quartiles. The distributions of scores for the three groups showed little evidence of 'ceiling' or 'floor' effects as they followed normal curves; suggesting appropriate and sufficient choices of variables.</p>", "<p>Univariate and multivariate analyses addressed ownership and utilization of any net and ITN. On the univariate level, frequencies and proportions were calculated for household ownership and under-five utilization of any net and ITN, and cross-tabulated with background characteristics of the households and demographic characteristics of the children. Pearson's Chi squared test was used to determine association with a P-value of &lt; 0.05 accepted as significant. Fisher's exact test was calculated for borderline significance and for cells with counts less than five. Logistic regression models were used to determine the predictors of household ITN and any net ownership separately. A model was developed for the combined data in which variables that were significantly associated with net ownership at the univariate level were used as covariates. Data was split into two regions, north and south, and separate models were developed to investigate the predictors of net ownership in the two regions.</p>", "<p>Regression models were also used to determine predictors of utilization. A model was developed for the combined data and then split by residence to generate output for urban and rural areas separately. Interactions between region and residence, wealth index and household heads gender, and religion and region were also explored for net ownership and utilization. Odds ratio and 95% confidence interval, CI, were generated for the final models and p-value of &lt; 0.05 for Wald's statistic accepted as significant. The logarithmic transform of the variable 'family size' was used to normalize the distribution before inclusion in the model and all missing data were excluded from the analysis.</p>", "<p>The prevalence of fever episodes in under-five children was analyzed on a univariate level by background characteristics and presence of health facility in the area of residence. Fisher's 95% CI was then calculated for the rate ratios and an overall p-value generated.</p>" ]

[ "<title>Results</title>", "<title>Background characteristics of study population (Additional File ##SUPPL##0##1##)</title>", "<p>Of the 5,588 household surveyed, 21% were from the SWZ, 14.4% from SEZ, 13.5% from SSZ, 17% from NWZ, 13.2% from NEZ and 20.9% from NCZ. 36% were urban and 64% rural (ratio 1:2). About two thirds of the households were sampled from communities with health facilities. Overall, distribution of wealth was unequal with slightly higher proportions in the upper two quartiles of the combined data, and the lowest two quartiles of the urban and rural data respectively.</p>", "<p>In the combined data, wealth index also varied within and between the zones. In the southern zones (SWZ, SEZ and SSZ), a majority of the population fell within the upper two quartiles (&gt;50%) while the northern zones have more households in the lower two quartiles (54% in NWZ and 76% in NEZ), except NCZ where more than 57% are also in 3^rdand highest quartiles, the largest proportion (33.5%) being in the 3^rdquartile.</p>", "<p>Of the 4,625 households with valid data on household religion, 3,044 (65.8%) were Christian, 1,529 (33.1%) Muslim and 52 (1.1%) practiced other forms of religion. A majority of the households in the southern zones (SWZ, SEZ and SSZ) were Christians; the NEZ was largely Muslim (73%), while the ratio of Muslim to Christian households in NWZ and NCZ was about 1.2:1.</p>", "<p>Males headed 93% of the households while females headed 7% of them. Generally, female family heads are commoner in the south (ranging from about 9% to 16%) than the north (1% – 3%). Overall mean family size was 5.4 (SD 3.0) ranging from 4.7 (SD 2.4) in SWZ to 6.5 (SD 3.4) in NWZ (Additional File ##SUPPL##0##1##).</p>", "<p>15.3% of household members were children under-five and 14.5% were 5–14 years old. Age group 50 and above constituted 5.6%, typical for a developing country.</p>", "<title>Univariate analysis</title>", "<title>Household net ownership</title>", "<p>Overall household ownership of any net was 23.9% (95% CI, 22.8%–25.1%) and 10.1% (95% CI, 9.2%–10.9%) for ITN (Additional file ##SUPPL##1##2##). A significantly (p &lt; 0.0001) larger percentage of rural dwellers (22.8%) owned any nets compared to urban dwellers (18.3%), and were more likely (p &lt; 0.0001) to own more than one net (11%) than urban dwellers (0.6%). Net ownership varied very significantly (&lt;0.001) by region. Households in SSZ consistently own the least nets in all categories while NEZ households own more nets than other regions, except for more than one net, which is commoner in the NWZ (35.4%). However, information about net ownership was only available from 58% of the households in the NEZ, which could have resulted in an overestimate if the valid cases were more likely to own nets than the missing cases. Also, the SEZ has the 2^ndlargest proportion of households with any net and the 3^rdlargest for more than one net. Since the population in the sample from SEZ was mostly rural (98%, possibly attributable to an error in data collection) and rural households were more likely to own nets than urban households, this could have caused an overestimate of the true proportion for the zone.</p>", "<p>When household net ownership by region (dichotomized N/S) was stratified by residence (urban or rural), residence was found to modify the association between region and ownership of any net. Rural households in the north were one-and-a-half times more likely to own nets than urban households in the same region (59.5% vs. 40.5%, RR, 1.47, 95% CI, 1.25–1.74), and in the south, about four times more likely than urban households (79.8% vs. 20.2%, RR, 3.94, 95% CI, 3.17–4.91) and this was statistically significant (p &gt; 0.0001). There was no synergism according to the additive model.</p>", "<p>These findings were similar for ITN ownership, with the rural households in the north almost twice as likely to own ITNs as the urban households (62.7% vs. 37.3%, RR, 1.67, 95% CI, 1.29–2.19) while rural households in the south were twice more likely to own ITN than their urban counterparts (67% vs. 33%, RR, 2.01, 95% CI, 1.52–2.73). No evidence for synergism was found on the additive model.</p>", "<p>Using the wealth index generated for all households irrespective of their urban or rural status (combined), households that fell in the highest quartile were hardly more likely to own any net (26.3%) compared to the lowest quartile (24.1%) with rate ratio, RR of 0.91 (CI, 0.79–1.07); while the poorest households were more likely to have more than one net (13.3%) compared to the richest (10.0%) with a significant RR of 1.3. There was no significant difference in ownership of at least one ITN between the rich and the poor (p = 0.05) and though poorest households were more likely to own more than one ITN than those in the other quartiles, this was significant only for the 3^rdquartile (RR 1.5, CI, 1.03–2.21).</p>", "<p>Using wealth index generated for urban and rural areas separately, the scenario was similar for urban households and that of the combined data. More households in the lowest group owned any compared to the 2^ndand 3^rd, but those in the highest owned more nets than any other group (p = 0.003). When urban wealth index was dichotomised, the rich households (upper two quartiles) owned significantly (p = 0.045) more nets (22%) than the poorer households (lower two quartiles; 18%).</p>", "<p>By rural wealth index, the 2^ndquartile households had the largest proportion of nets in all categories. In the category for any net, this was followed by the highest quartile and the lowest had the least proportion of households with nets (&lt;0.0001). When dichotomized, the richer households (upper two quartiles) were more likely to have nets (28.2%) than the poorer (lower two quartiles) households (23.2%).</p>", "<title>Multivariable analysis</title>", "<title>Ownership of any net</title>", "<p>Variables that were significantly associated with net ownership at the univariate level; region, residence, and household wealth index; and other covariates; presence of at least one under-five child in household, religion, gender and age of household head, family size and presence of health facility in the community; were entered in an unconditional logistic regression model (combined Model A) to determine predictors of household net ownership using the stepwise forward likelihood ratio.</p>", "<p>In the initial model, combined wealth index, residence, region, religion, household head's gender, presence of at least one under-five child in the household were significantly associated with net ownership; however, when 'presence of an educated eligible woman in the household' as a proxy for household education status, and interaction terms for region by residence, religion by region, and combined wealth index by household head's gender, were entered into the model, the final model did not include religion, however, education and family size became significant.</p>", "<p>The odds of a household owning a net if there was an educated eligible woman in the household were 30% higher than a household without an educated eligible woman (p = 0.005). A unit increase in family size increased odds of ownership of any net more than twice (p &lt; 0.0001) while controlling for all other variables; and if a household had at least one under-five child the odds of owning any net was about 60% higher than households with no under-five child (adjusted OR, 1.60, CI, 1.40–1.90). Living in a rural area raised ownership odds by 26% compared to urban residence (p = 0.021), and there was interaction between residence and region with a more than 50% increase in the odds of owning a net if the household was in the north and rural compared to urban households in the south (p = 0.024). Every unit rise in wealth index was found to increase ownership odds 1.24 times (CI, 1.15–1.34).</p>", "<p>Data was split by residence and separate models were developed for urban and rural areas (Models B and C respectively) and outputs were generated for the north and south regions separately to identify predictors for the regions. Additional File ##SUPPL##2##3## shows the adjusted odds ratios for the final variables in the different models.</p>", "<p>In the final model for the urban region (Model B), the presence of an educated woman in the household raised the odds of owning a net by 42% in the north compared to those without (p &lt; 0.0001), while this was not predictive in the south controlling for other variables in the model. The odds of owning a net were almost three times higher for households with a health facility in their community in the south than those without (OR, 2.88, CI, 1.54–5.37); while they were more than one-and-a-half times lower in the north (OR, 0.59, CI, 0.41–0.81), controlling for other variables in the model. Also, in the north, a unit increase in urban wealth index independently increased the odds of net ownership by 28% (OR, 1.28, CI, 1.09–1.50), a difference not found in the South. An under-five child in the household was very significantly predictive of net ownership in the south with an OR of 3.24 (CI, 1.99–5.27) but not in the north.</p>", "<p>For rural areas (Model C), presence of under-five child in the household and health in the community, family size, presence of health facility in the community, wealth index, education and religion were predictors of ownership of any net in the south while only family size and religion were important for net ownership in the north.</p>", "<p>Households with at least one under-five child was 1.7 times (p &lt; 0.0001) as likely to own any net as those without, and those with a health facility in their community were 1.6 times (p = 0.001) more likely to have nets than those without health facilities, controlling for other variables in the model. The odds of net ownership increased by more than three-and-a-half in the south, and more than two-and-a-half in the north, for every additional family member (p &gt; 0.0001 and 0.002 respectively).</p>", "<p>Christians were more than twice more likely to own nets in the south (OR, 2.35, CI, 1.32–4.20) and twice less likely to in the north than Muslims (OR, 0.23, CI, 0.03–0.91). A unit rise in wealth index and having an educated woman in the house independently raised the odds of net ownership in the south by 33% and 55% respectively, with no effects in the north.</p>", "<title>Ownership of ITN</title>", "<p>Similar logistic regression models were developed in which the dependent variable was ownership of at least one ITN by the household. Variables analysed on the univariate level were adjusted for potential confounders in a combined model (Model I) and the data was then split by residence (urban, Model II/rural, Model III) and outputs generated for the north and the south (Additional File ##SUPPL##3##4##).</p>", "<p>In the combined model, although region and combined wealth index were significant on the univariate level, when entered into the model with other covariates, including presence of an under-five child in the house, presence of an educated eligible respondent in the household (as a proxy for education status of the household), health facility status, gender of household head, residence and interaction terms for region by residence, education by combined wealth index and combined wealth index by household head's gender; these two variables were not significantly associated with household ITN ownership. The single variables in the final model were under-five child in the household; health facility status, household head's gender, education status and family size and these were found to predict ITN ownership. Additional File ##SUPPL##4##5## shows the adjusted odds ratio, 95% confidence intervals and p-values for these predictors.</p>", "<p>A household with an under-five child was more than on-and-a-half times as likely to own an ITN as those without (OR, 1.55, CI, 1.53–1.94), while presence of health facility in the community and male gender of household head independently increased the odds of ITN ownership by about 100% (p &lt; 0.0001 and 0.023 respectively). A unit increase in family size raised the odds of ownership by about 90% (OR, 1.88, CI, 1.13–3.13) and the presence of an educated woman in the household increased the odds by 36% (P = 0.012).</p>", "<p>Interestingly, there was evidence for interaction between region and residence in relation to ITN ownership; being a rural household in the north significantly raised the odds of owning ITN about twice (p &lt; 0.0001), even though residence in itself was not predictive. There was also evidence for synergism on a multiplicative scale between education and household wealth index on ITN ownership. A unit rise in wealth index when there was an educated woman in the household led to a 15% rise in the odds for owning a net (p = 0.025). There was, however, no evidence for interaction between combined wealth index and gender of household head.</p>", "<p>For urban households, under-five child in the household and household religion were significant predictors of ITN ownership in the south after controlling for family size, education (using presence of an educated woman in the household as proxy), urban wealth index, and urban wealth index by household head's gender, household head's gender and health facility status. Only education was significantly predictive in the north and strongly so (OR, 2.05, CI, 1.18–3.55); however when an interaction term for education by urban wealth index was introduced into the model, while it did not change the picture in the south, it was significant in the north (OR, 1.33, CI, 1.02–1.75), such that in the final model, the effect of education alone on ownership of nets was slightly reduced, but was still significant (OR, 1.93, CI, 1.01–3.38).</p>", "<p>Christianity increased the odds of ITN ownership about seven fold compared to Islam in the south (OR, 6.78, CI, 1.64–29.12) while the presence of an under-five child in the household increased the odds 3.4 times (OR, 3.42, CI, 1.92–6.40).</p>", "<p>Similarly, ITN ownership in the south was also dependent on under-five child in the household and household religion, and rural wealth index in addition. Model III showed that ownership odds significantly increased 1.5 times with under-five in the household (p &lt; 0.0001) and six-fold when household religion was Christianity compared to Islam, each variable adjusted for others in the model.</p>", "<p>Religion was predictive of ITN ownership in the rural north and south. Just like it was for urban households in the south, in the rural south, the odds of ownership was significantly increased if the household was Christian (OR, 5.95, CI, 1.45–24.4) compared to if they were Muslim. However in the north, the relationship was in the opposite direction; Christian households had twice lowered odds of ITN ownership in this region compared to Muslim households, even after controlling for possible interaction between religion and wealth index. There was evidence for interaction between wealth index and household head's gender in the north; the odds of net ownership reduced by 30% with a unit rise in wealth index when the household head was female. Other religions were unimportant in determining net ownership in both regions.</p>", "<title>Number of nets in the household</title>", "<p>Overall, household ownership of any net was 23.9%, with 10.9% households owning more than one net. ITN ownership was however 10.1%, with 4.7% reporting ownership of more than one ITN. This means that of all households with nets, 42% had ITN while 58% had untreated bed nets; and of those who have more than one net, 43% were treated nets while 57% were untreated (Additional File ##SUPPL##4##5##). The mean number of nets per household reporting any net ownership was 1.82 (SD, 1.11) and 1.17 (SD, 1.25) for the total sample population.</p>", "<title>Utilization of mosquito nets by under-five children</title>", "<p>Overall, 11.5% (95% CI, 10.4%–12.6%) and 1.7% (95% CI, 1.3%–2.2%) of all eligible children slept under any net or ITN respectively, the night before the survey. Younger children (&lt;2 years old) were more likely to be put under any net than older children although this was not significant for ITN. There was no association between the gender of the child and the use of nets (p = 0.36) however, region was significantly associated with utilization (p = &lt; 0.0001); southerners were more likely to keep their children under nets than northerners, who were more likely to own nets. Utilization was commoner among rural children than urban children for any net but did not differ for ITNs (Additional File ##SUPPL##5##6##).</p>", "<p>Education was very significantly associated with net utilization among under-five children (p = &lt; 0.0001). The rate of utilization increased monotonously with level of education; those with higher education than secondary were about thrice as likely to put their under-five children under a net as the uneducated, and twice as much secondary school leavers.</p>", "<p>In the combined data, household wealth index was significantly associated with utilization of any net and ITN by children under-5 (p = 0.004 and 0.003 respectively). Children who fell in the upper 2 quartiles are more likely to have slept under a net the night before the survey than the lower two quartiles. Those in the highest quartile are 1.5 times and 1.3 times as likely to use any net or ITN than those in the lowest quartiles respectively. However, when rural and urban dwellers were separated according to their wealth indices, utilization of nets was independent of the household wealth, although utilization still varied with residence by caregiver's level of education.</p>", "<p>Additional File ##SUPPL##6##7## shows utilization rates of ITN and any net for rural and urban under-five children by caregiver's level of education. Reported use of any net was significantly (p &lt; 0.0001) higher among rural children of the less educated and uneducated caregivers (56%) than those whose caregivers were more educated (44%). However, the proportion of under-five children who used ITN among them was significantly (p &lt; 0.0001) higher for caregivers with higher levels of education (secondary and higher, 56.7%).</p>", "<p>In urban households, caregivers with secondary and higher levels of education were significantly (p &lt; 0.0001) more likely to put their children under any net and were more likely to protect them with ITN, although this was not statistically significant (p = 0.146).</p>", "<title>Utilization of any net</title>", "<p>Multivariate analysis of utilization of any net by under-five children using the combined data showed that fever/convulsion in the previous two weeks, availability of health facility, residence and caregiver's level of education were predictors of utilization of any net, controlling for other variables in the model (Model 1, Additional File ##SUPPL##7##8##). Combined wealth index (CWI), child's age, family size, religion, region and region by residence, were not predictive of utilization of any net by under-five children.</p>", "<p>If a child had fever/convulsion in the last two weeks, the odds of using a net the night before the survey was about 1.3 times higher than if fever/convulsion was not reported (p &lt; 0.0001). This finding was similar for presence of a health facility in the community (OR, 1.29, CI, 1.01–1.63). In addition, the odds of an under-five child sleeping under any net the night before the survey was 40% higher if the caregiver was educated compared to uneducated caregivers (p = 0.016). There was also evidence for interaction between CWI and caregiver's level of education; for an educated caregiver, a unit increase in wealth index resulted in a 30% rise in the odds of utilization of any net (OR, 1.29, CI, 1.14–1.45) independent of other variables.</p>", "<p>Under-five children who live in rural areas were about one-and-a-half times as likely to use any net as their counterparts in urban areas. When the data was split by urban or rural status of residence, using the combined wealth index as a measure of household wealth in both residences, (Models 2 and 3 respectively), fever/convulsion episode in the child and wealth index were found to be important predictors in the rural areas. The odds of utilization of any net was 1.5 times higher in children who had fever two weeks before the survey than under-five children who did not have fever (OR, 1.49, CI, 1.11–1.99); while wealth index independently increased the odds by 17% for every unit rise (OR, 1.17, CI, 1.01–1.37).</p>", "<p>The presence of health facility in the community, caregiver's level of education and age of the child were found to predict utilization of any net in urban children. The odds of an under-five child using a net the previous night was 2.3 times higher for children who live in communities with health facilities than those without heath facilities (p = 0.001). An educated caregiver had a 2.16 higher odds of putting her child under a net than the uneducated, controlling for other factors (p = 0.008).</p>", "<p>Caregiver's education was also found to interact with wealth index; the odds of net use increased by 30% when caregiver was educated (OR, 1.3, CI, 1.05–1.57) per unit rise in wealth index compared to an uneducated caregiver. Children less than two years of age were almost twice as likely to be put under a net as children between two and five years (OR, 0.56, CI, 0.36–0.85). When the effect of urban and rural wealth indices were examined with separate models developed for urban and rural children, the predictors were overall similar to those found using the combined wealth index. However, the age of the child was not predictive in urban region and fever was not significant for rural children.</p>", "<title>Utilization of ITN</title>", "<p>Additional File ##SUPPL##8##9## shows the variables in the final logistic regression models developed to predict the use of ITN by under-five children. Only valid cases (2009) were included in the analysis. In the combined data (Model X), presence of health facility in the community where a child lived strongly predicted the use of ITN the night before the survey, with an odds three times higher than where health facilities were absent (OR, 2.93, CI, 1.29–6.78). Among Christian caregivers, the odds of an under-five child sleeping under an ITN the night before the survey was more than three times higher than for Muslim caregivers. When a caregiver was educated, a unit rise in wealth index increased the odds of utilization of ITN by 43%, other factors controlled for.</p>", "<p>Age, family size, combined wealth index, residence, region, fever/convulsion episodes and caregiver's level of education were not found to independently predict utilization of ITN by under-five children in the combined data. However, when spilt by residence and output for the combined model generated for urban and rural communities, region was predictive in urban areas (Model Y) while there was no variable in the final model for rural communities. Children living in the north had five-fold lower odds of sleeping under an ITN than children living in the south (OR, 0.18, CI, 0.06–0.54).</p>", "<p>Using rural and urban wealth indices, separate models were used to check for predictors. These models showed the same result for urban communities with region predicting use of ITN; while an interaction was found between rural wealth index and caregiver's level of education. When a caregiver in a rural area was educated, a unit increase in rural wealth index raised the odds of an under-five child sleeping under an ITN by 57% compared to an uneducated caregiver (OR, 1.57, CI, 1.06–2.32).</p>", "<title>Utilization of nets and fever and/or convulsion prevalence</title>", "<p>Overall fever/convulsion prevalence among under-five-children who slept under any net was 1.2 times greater than those who did not use any net; and more than one-and-a-half times greater in those who used ITN the night before the survey (Additional File ##SUPPL##9##10##).</p>", "<p>There was a positive association between fever episodes in the last two weeks and the use of nets in under-five-children. The odds of using any net the night before the survey were about one-and-a-half times higher in children who had fever and/or convulsion in the last two weeks than children who had no fever and/or convulsion; and this was significant (p = 0.013). The difference was explained by the use of ITN; children with history of fever were twice as likely to have been put under an ITN the previous night as those with no such history, p = 0.006, while there is no significant difference among children who used other nets.</p>", "<p>Among all children who used any net, those with history of fever and/or convulsion were 1.7 times more likely to use an ITN than an ordinary net; this was however not statistically significant. Usage of other nets was independent of fever history. Users and non-users had a prevalence of roughly 30%.</p>" ]

[ "<title>Discussion</title>", "<p>With data collected during a national household survey in October 2005, the set year for the achievement of the Abuja targets and mid-term assessment of the Abuja declaration, this study has demonstrated that Nigeria, though it has made some progress, is still very far from achieving the Abuja targets with regard to ITN coverage of the vulnerable groups.</p>", "<title>Household net ownership</title>", "<p>Overall household ownership of any net was 23.9% (95% CI, 22.8%–25.1%) and ITN was 10.1% (95% CI, 9.2%–10.9%). Given the fact that ITNs were barely existent in the country before the launching of the RBM programme in Nigeria (coverage was 0% as reported by NetMark), these figures represent non-negligible progress. These results show about 100% improvement over the figures reported for any net and five-fold increase for ITN from the National Demographic Health Survey (NDHS) conducted in 2003 in the country; which reported 12% of the population owned any net and 2% owned ITN [##UREF##9##31##].</p>", "<p>In 2004, NetMark reported 27% for ownership of any net and 9% for ITN, even though the study was done one year later than the NDHS. The exclusion of households without under-five children, which are less likely to have nets, could have caused an overestimation of the proportions and probably explains the higher proportion of ownership of any net compared to the current study. However, the higher ITN coverage in this study could reflect the massive distribution of free ITNs embarked upon by the National Malaria Control Programme (NMCP) in the last two years before the survey [##UREF##18##45##]. Although household ownership of ITN has increased in the Nigerian population, it is still low compared to other African countries, like Senegal, Malawi and Eritrea, which is the only country that has achieved the Abuja Targets [##UREF##1##2##].</p>", "<p>The relatively higher proportion of net ownership in the rural area shows an existing demand in these areas and forms a good entry point for introduction of ITN since possession is the first step to utilization.</p>", "<p>Although ITN ownership was not found to differ significantly by wealth index, the fact that the possession of any net varied by household wealth, with the rich more likely to own any net than the poor, may be a better indicator of the relationship between wealth and net ownership. There was no information on how the nets in the study were acquired, whether they were received free from the National Malaria Control Programme (NMCP) or purchased. However since the nets distributed free of charge to the vulnerable groups by NMCP (through health facilities and campaigns) were ITNs, this could account for the equity in ownership of ITN since those with ordinary nets must have paid for them. The correlation between presence of a health facility in the community and under-five child in households with ITN, also raises the probability of having received them from the programme, as this was one of the distribution channels.</p>", "<p>This finding further highlights poverty as a potential barrier to scaling up of ITN in Nigeria as documented by other studies [##REF##12917269##26##] and raises questions about funding large-scale distribution and the sustainability of such measures.</p>", "<p>Two main strategies have been suggested to improve ITN coverage, forming two poles. Some authors [##REF##12901886##46##,##UREF##19##47##] argue that to maintain a balance between equity and sustainability, the public sector should target subsidies at the vulnerable groups while the private sector is given room to grow. On the other hand, some believe ITN should be treated as a public good and should therefore be given out free of charge to everyone, especially the poor, who are most vulnerable to the disease; they argued that the international community can afford the cost for SSA [##REF##12726981##48##].</p>", "<p>The two positions are however valid. A large proportion of the population still live under a dollar per day and the vulnerable groups may not be able to pay for ITNs, even when subsidized. On the other hand, the limitedness of resources, considering the huge population of the country, makes free distribution unsustainable without guaranteed continuous support from donor agencies.</p>", "<p>A recent study showed that an average Nigerian household was willing to pay Naira 7,324 (USD 61) per month for the control of malaria, an amount that was 37% higher than the current expenditure on malaria in form of protection, treatment and indirect costs [##REF##17517146##49##] Mokuolu <italic>et al</italic>, in another study on patterns of consumption of the new and more expensive artemisinin-based combination treatment (ACT), reported a rise in the demand for the more expensive drug as a result of increased awareness about the efficacy of the medicine and the ineffectiveness of the cheaper alternative [##REF##17255221##41##].</p>", "<p>While for some, willingness to pay may match actual ability to pay, many may not be able to afford payment even with a high perception of risk. A window of opportunity for social marketing and commercial sales exists on one hand, while, on the other hand, it is also expedient for the most vulnerable groups to be provided for, so that ITN coverage can be scaled up and the RBM goal achieved [##REF##17713981##50##].</p>", "<p>Wealth index interacted significantly with presence of an educated eligible woman in the household (which was also an independent factor in the total population), underscoring the importance of education in ITN possession. Education as a determinant of net ownership has also been documented by a study in The Gambia that modeled the determinants of bed net ownership and the factors that influence the number of nets purchased [##REF##17488900##25##].</p>", "<p>This shows a cross-linkage between achieving the Abuja targets and other millennium development goals and it underscores the integration of the programmes targeted at achieving these goals. The relationship between religion and household net ownership points to the role of religious leaders in the propaganda for net ownership.</p>", "<title>Utilization of nets</title>", "<p>Proportion of under-five children using ITN is still very low, although the rate found in this study – 1.7% (95% CI, 1.3%–2.2%) – was higher than the 2000 baseline of 0% and the 2003 NDHS figure of 1.2%, considering the 60% set as the Abuja target. There is an urgent need for the NMCP to intensify its efforts.</p>", "<p>This study demonstrated higher utilization rates for any net among rural children than urban children. This was contrary to what was found in a meta-analysis of 13 surveys, which included five Demographic Health Surveys for African countries, where children in urban households were found to be more likely to use nets than children in rural households [##REF##12869090##21##]. A similar meta-analysis carried out on studies conducted between 1999–2002 also reported the same findings [##REF##15331842##51##]</p>", "<p>However, these studies analysed surveys that were done as baseline assessment before the full-scale implementation of the RBM programme. There findings are, therefore, not likely to reflect the state of the art with utilization of nets.</p>", "<p>Although use of ITN is low, the higher rates of use of ordinary nets show an existing culture of net use and represent an entry point for scaling up ITN. A massive net (re)treatment campaign should be embarked upon to turn these nets into ITN. Utilization was disproportionate to household net ownership. The north-south divide of this scenario particularly raises questions about the factors that could be responsible for this. Koreromp <italic>et al </italic>also found this gap between ownership and possession, and noted that nets were less likely to be used during hot and dry seasons [##REF##12869090##21##]. However, this study was conducted during the rainy season when malaria transmission peaks and net utilization were supposedly high. Binka <italic>et al </italic>in a study of acceptability of nets in Ghana, found that 99% used their nets during rainy season [##REF##9217706##27##]. The low rate of use by this vulnerable group therefore raises concern. Could it be that they are not given priority in the household for using the nets as there is scarcity of ITNs in the home? This is possible considering that the ratio of mean family size to mean number of ITNs per household was 4.5 ± 3.0:1 among families with ITN; however, since it is possible for more than one person to share a net and information was unavailable on this, it is difficult to ascertain this. Could it also be that caregivers are unaware of the need for consistent use or could the perception of risk be lower in the north than the south?</p>", "<p>Educated caregivers with an increasing wealth index had higher likelihood of putting their children under ITNs than uneducated ones, even with increase in wealth index. This pre-supposes that, given equity in ownership of nets, there still remains a gradient between those children whose caregivers are educated and those who are not. This is an important finding that calls for a paradigm shift in the malaria control efforts since actual protection of these children depends on the use of the nets rather than ownership. The fight against malaria cannot be left only in the hands of the NMCP; it calls for a multi-sectoral approach involving, for instance, the Ministry of Education and the Ministry of Women Development.</p>", "<p>The correlation between the presence of health facilities in the community and utilization of ITN, as has been highlighted above, could be due to the fact that these are ITN distribution outlets for the NMCP and could point to the effectiveness of this mechanism in reaching the vulnerable groups. However, the fact that this did not predict use when data was split by urban and rural residence calls for further analysis in this area. The north-south divide in under-five utilization of ITN also calls for further research into the possible barriers to use. Could it reflect cultural or religious reasons? This arises especially as Christians were more than three times more likely to put their children under ITNs than Muslims in the total population.</p>", "<title>Limitations of the study</title>", "<p>In interpreting the results of this study, certain issues must be borne in mind. The sampling methodology used in this study (cluster design) is fraught with the problem of high intra-class correlation and taking outcomes to the exact percentage point may be misleading. However, since a large number of clusters were studied in this survey, this effect will be minimal.</p>", "<p>Also, the survey questionnaire was not translated into local languages, the interviewers interpreted them to the respondents who did not understand English. It is possible for interviewers to misinterpret questions or introduce personal preferences. However, because they were all trained and the hypothesis of this study was generated after data collection, this should probably not have biased the study in a significant way.</p>", "<p>The household participation rate in the study was 78%. Two zones (NCZ and SWZ) were overly represented in the sample, contributing over 20%, while two other zones contributed less than 15% to the sample. Differences between these zones could affect the parameters measured in this study, the direction of which cannot be ascertained as no information was available about those who did not participate in the study. Nevertheless, since most states in the south are similar and the same in the north, and the north to south ratio of the sample was about 1:1, this limitation may not adversely affect the study.</p>", "<p>Although the PCA method used in developing the household wealth index only measures the long-term household wealth and does not account for short-term wealth or shocks to the household as other methods like income and expenditure do; this study was not particularly focused on the current resources available to households but on the class to which they belong based on their durable possessions and in relation to other households. This is therefore not likely to affect the socio-economic classification. Moreover, data on validation of the reported assets was not available in this study, thus this could be a potential limitation, however, some confidence in the result can be drawn from the fact that the findings are similar to what other studies have documented with regards to net ownership.</p>", "<p>The use of an educated eligible woman in the household as a proxy for educational status of the household could introduce bias in the result. If there was more than one eligible woman in the house and the educated one happens not to have any under-five child, then she would not have been interviewed according to the study protocol, but since this would have been a non-differential misclassification, it is likely to have caused an underestimation of the odds of net ownership.</p>", "<p>The use of fever as proxy for malaria also calls for caution in interpreting these results. Being a symptom for many other childhood illnesses with shared risk factors for malaria, like age, it is probable that children with diseases other than malaria could have been counted as malaria cases since no specific diagnosis was done to ascertain the cases. However, fever is the commonest symptom of malaria and 60% of every fever episode among under-five children in childhood, especially in the rainy season, is likely to be due to malaria. Moreover, it has been documented that in malaria endemic regions like Nigeria, malaria increases the risk of other childhood diseases e.g., pneumonia, and frequently co-exist with them in a co-morbid state [##REF##17651488##52##,##REF##15764695##53##]</p>", "<p>Since this study was done during the rainy season, during which malaria transmission peaks, children are more likely to have more episodes of malaria. Therefore, it is unlikely that this proxy will adversely affect the findings, and even if it does, the misclassification will be non-differential as the question about fever and net use were asked independently, and as such would underestimate the prevalence throughout.</p>", "<p>Finally, since this study, for the most part was a retrospective evaluation of self-reported behaviour patterns, it was subject to recall bias, however as the time-period was short (two weeks) this bias is minimal.</p>" ]

[ "<title>Conclusion</title>", "<p>Over a five year period, Nigeria has succeeded in achieving only 2.8% of the 60% expected coverage for under-five children with insecticide-treated nets. Although this is a non-negligible achievement considering the baseline situation of 0%, this progress is much too slow, if the target set for 2010 are to be achieved and, therefore, puts a prerogative on a more concerted effort by the malaria control body.</p>", "<p>Achieving the set goal requires focused and well-informed strategies that are based on scientific evidence generated from local circumstances. This study has identified key issues constituting impediments in the cogwheels of progress; while poverty militates against ownership of net, lack of education curtail its use. This presents the policy makers with the challenge of addressing these issues decisively, if the RBM goal, or indeed the MDGs will be met.</p>", "<p>Considering the time constraints, the aim must be a rapid scale-up among the target groups, while not neglecting the other members of the population. A pluralistic approach to scaling up, in which several distribution and financing mechanisms are combined has been recommended; [##UREF##19##47##] this includes commercial sales of nets, social marketing, community-based distribution and targeted subsidies. However, the choice of strategy must be based on evidence of what works in the context of the country to ensure adaptability. These strategies must target not only increasing household ownership but also utilization, which is most important for epidemiologic impact.</p>", "<p>In addressing poverty as a hindrance to ownership, the government must decide what is workable for it within the limits of its resources; whether to continue giving out the nets free of charge, if it can sustain it, or to highly subsidize the nets while continuing to use the MCH clinics as distribution outlets, and continuing the mass ITN distribution campaign. The most important consideration should be that cost must not be a barrier to access to nets for the vulnerable groups.</p>", "<p>Other measures of reducing the cost of ITN should be explored to ensure that the rest of the population not covered by subsidy or free distribution can have access to the nets at an affordable cost. Seeking transfer of net manufacturing technology from a sister African country like Tanzania, and removing taxes and tariffs on netting materials and insecticides are some of these measures. These will create an enabling environment for, and encourage local manufacturing of nets, which not only has the potential to substantially reduce the market price of the nets, but also to create employment opportunities and, as such, alleviate poverty.</p>", "<p>Evidence from this study shows that NMCP needs to reassess the communication strategies it has been using for net delivery and promotion in the last five years. If nets have been given out free-of-charge to under-five children for five years and the coverages are still low, then it raises questions about the kind of messages being delivered to the population and how well they understand them. While recognizing that NMCP cannot embark on adult education of the vast majority in the rural communities who are illiterate, it could liaise with the Ministries of Communication, Education and Women Development in developing and disseminating information on ITN use and benefits in the language that the audience best understand.</p>", "<p>With the foregoing, it is paramount for the federal government to increase its health expenditure to 15% of GNP as agreed during the RBM summit by the African Heads of States. The currently poor spending on health (less than 5% of GNP) is too small if the country is to achieve a quick win over malaria and achieve the MDGs, to which malaria control is so intricately linked.</p>" ]

[ "<p>This is an Open Access article distributed under the terms of the Creative Commons Attribution License (<ext-link ext-link-type=\"uri\" xlink:href=\"http://creativecommons.org/licenses/by/2.0\"/>), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.</p>", "<title>Background</title>", "<p>The Abuja target of increasing the proportion of people sleeping under insecticide-treated nets (ITNs) to 60% by the year 2005, as one of the measures for malaria control in Africa, has generated an influx of resources for malaria control in several countries in the region. A national household survey conducted in 2005 by the Malaria Control Programme in Nigeria assessed the progress made with respect to ITN ownership and use among pregnant women and children under five years of age since 2000. The survey was the first nationally representative study of ITN use assessing progress towards the Abuja target amongst vulnerable groups.</p>", "<title>Population and Method</title>", "<p>A cross-sectional survey of a sample of 7,200 households, selected by a multistage stratified sampling technique from 12 randomly selected states from the six geopolitical zones of the country. Data collection was done during the malarious rainy season (October 2005) using a modified WHO Malaria Indicator Survey structured questionnaire about household ownership and utilization of mosquito nets (treated or untreated) from household heads.</p>", "<title>Results</title>", "<p>Household ownership of any net was 23.9% (95% CI, 22.8%–25.1%) and 10.1% for ITNs (95% CI, 9.2%–10.9%). Education, wealth index, presence of an under-five child in the household, family size, residence, and region by residence were predictive of ownership of any net. The presence of an under-five child in the household, family size, education, presence of health facility in the community, gender of household head, region by residence and wealth index by education predicted ITN ownership.</p>", "<p>Utilization of any net by children under-five was 11.5% (95% CI, 10.4%–12.6%) and 1.7% (95% CI, 1.3%–2.2%) for ITN. Predictors of use of any net among under-five children were fever in the previous two weeks, presence of health facility in the community, caregiver's education, residence, and wealth index by caregiver's education; while religion, presence of health facility and wealth index by caregiver's education predicted the use of ITN among this group.</p>", "<title>Conclusion</title>", "<p>This study demonstrated that the substantial increase in ITN utilization among children under five years of age in Nigeria is still far from the Abuja targets.</p>" ]

[ "<title>Competing interests</title>", "<p>The authors declare that they have no competing interests.</p>", "<title>Authors' contributions</title>", "<p>OBO conceived the design, acquired the data, analysed and interpreted the results. MH contributed to the analysis and revised the manuscript while OS contributed to acquiring the data and proof read the manuscript.</p>", "<title>Supplementary Material</title>" ]

[ "<title>Acknowledgements</title>", "<p>The primary data collection for this study was carried out by the Federal Ministry of Health with the support of the Global Fund to Fight AIDS, Tuberculosis and Malaria. We are grateful to Dr. T. O. Sofola, National Coordinator for the Malaria Control Programme in Nigeria for granting permission to use the data. We also acknowledge Dr. Bayo Fatunmbi of World Health Organization and Babatunde Oresanya, who facilitated the dispatch of the data from Nigeria to Israel where the analysis was done.</p>" ]

[]

[]

[]

[]

[]

[]

[]

[ "<supplementary-material content-type=\"local-data\" id=\"S1\"><caption><title>Additional file 1</title><p>Socio-demographic characteristics of study households by region.</p></caption></supplementary-material>", "<supplementary-material content-type=\"local-data\" id=\"S2\"><caption><title>Additional file 2</title><p>Household ownership of mosquito nets by background characteristic.</p></caption></supplementary-material>", "<supplementary-material content-type=\"local-data\" id=\"S3\"><caption><title>Additional file 3</title><p>Predictors of household ownership of any net.</p></caption></supplementary-material>", "<supplementary-material content-type=\"local-data\" id=\"S4\"><caption><title>Additional file 4</title><p>Predictors of household ownership of ITN.</p></caption></supplementary-material>", "<supplementary-material content-type=\"local-data\" id=\"S5\"><caption><title>Additional file 5</title><p>Net Ownership and Type of net by Number of nets in households.</p></caption></supplementary-material>", "<supplementary-material content-type=\"local-data\" id=\"S6\"><caption><title>Additional file 6</title><p>Utilization of mosquito nets by children.</p></caption></supplementary-material>", "<supplementary-material content-type=\"local-data\" id=\"S7\"><caption><title>Additional file 7</title><p>Proportion of rural and urban under-five children who used nets the night before the survey by caregiver's education level and type of net used.</p></caption></supplementary-material>", "<supplementary-material content-type=\"local-data\" id=\"S8\"><caption><title>Additional file 8</title><p>Logistic regression models for prediction of use of any net by under-five children.</p></caption></supplementary-material>", "<supplementary-material content-type=\"local-data\" id=\"S9\"><caption><title>Additional file 9</title><p>Logistic regression models for prediction of utilization of ITN by under-five children.</p></caption></supplementary-material>", "<supplementary-material content-type=\"local-data\" id=\"S10\"><caption><title>Additional file 10</title><p>Prevalence of fever episodes among under-five children by utilization of net.</p></caption></supplementary-material>" ]

[]

[]

[ "<media xlink:href=\"1475-2875-7-145-S1.pdf\" mimetype=\"application\" mime-subtype=\"pdf\"><caption><p>Click here for file</p></caption></media>", "<media xlink:href=\"1475-2875-7-145-S2.pdf\" mimetype=\"application\" mime-subtype=\"pdf\"><caption><p>Click here for file</p></caption></media>", "<media xlink:href=\"1475-2875-7-145-S3.pdf\" mimetype=\"application\" mime-subtype=\"pdf\"><caption><p>Click here for file</p></caption></media>", "<media xlink:href=\"1475-2875-7-145-S4.pdf\" mimetype=\"application\" mime-subtype=\"pdf\"><caption><p>Click here for file</p></caption></media>", "<media xlink:href=\"1475-2875-7-145-S5.pdf\" mimetype=\"application\" mime-subtype=\"pdf\"><caption><p>Click here for file</p></caption></media>", "<media xlink:href=\"1475-2875-7-145-S6.pdf\" mimetype=\"application\" mime-subtype=\"pdf\"><caption><p>Click here for file</p></caption></media>", "<media xlink:href=\"1475-2875-7-145-S7.pdf\" mimetype=\"application\" mime-subtype=\"pdf\"><caption><p>Click here for file</p></caption></media>", "<media xlink:href=\"1475-2875-7-145-S8.pdf\" mimetype=\"application\" mime-subtype=\"pdf\"><caption><p>Click here for file</p></caption></media>", "<media xlink:href=\"1475-2875-7-145-S9.pdf\" mimetype=\"application\" mime-subtype=\"pdf\"><caption><p>Click here for file</p></caption></media>", "<media xlink:href=\"1475-2875-7-145-S10.pdf\" mimetype=\"application\" mime-subtype=\"pdf\"><caption><p>Click here for file</p></caption></media>" ]

{ "acronym": [], "definition": [] }

[ "<title>Background</title>", "<p>Recent advances in genomics and bioinformatics are allowing science to test the boundaries of reductionism as never before: how well can biological processes observed at the level of individual molecules, genes or cells predict the behaviour of more complex systems such as whole organs, individuals, or populations? These exciting technological developments have generated renewed interest within existing, more long-established biological disciplines to seek out empirical tools for quantifying and testing the relationship between phenomena occurring at different levels of biological organization in order to generate better predictions. One such field is medical entomology. Typically research in medical entomology is conducted at two very different scales: the first being laboratory-based studies of arthropod disease vectors under controlled insectary conditions, and the second being large-scale epidemiological surveys of their abundance and distribution in nature. While the former approach is clearly advantageous for identifying potential biological mechanisms, and the latter for generating hypotheses from correlations, it is increasingly recognized that the ability to make inferences across these scales is hindered by a conceptual and methodological gap in between that limits our ability to conduct hypothesis-driven empirical research on relatively large and environmentally realistic scales [##REF##17243075##1##].</p>", "<p>Arguably for the first time in the history of their discipline, medical entomologists now find themselves in the unique position of having both the need for such experimental initiatives recognized [##REF##16583503##2##], and the financial support to create them becoming available through new funding streams in global health. The need to fill the research gap between laboratory and field is also stimulated by awareness that although current approaches to important vector-borne diseases such as malaria based on artemisinin-based combination therapies and insecticide treated bed nets are proving successful [##REF##11418148##3##, ####REF##12749495##4##, ##UREF##0##5##, ##REF##15655009##6##, ##REF##16515516##7####16515516##7##], their long-term effectiveness may be undermined by the emergence of drug and insecticide resistance [##REF##16265886##8##, ####UREF##1##9##, ##REF##17579229##10####17579229##10##]. Consequently, the need for new strategies that exploit novel aspects of vector genetics, physiology, behaviour and ecology are increasingly needed. These innovations must be drawn from an understanding of vector biology within natural transmission settings if they are to yield rapid, locally appropriate strategies for disease control.</p>", "<p>While almost all new approaches to vector control could benefit from a closer integration of laboratory and field perspectives [##UREF##2##11##], the most prominent candidate is the development of transgenic parasite resistant and/or sterile vectors whose release into the wild could reduce disease transmission by reducing parasite and/or vector populations [##REF##12901936##12##, ####UREF##3##13##, ##REF##17373671##14##, ##REF##18165498##15####18165498##15##]. In the past decade, the genetic transformation of a number of important disease-transmitting mosquito species has become possible [##REF##17406416##16##, ####REF##17160283##17##, ##REF##11580043##18##, ##REF##10879538##19##, ##REF##11005862##20####11005862##20##]. Transgenes have been identified that deliver effector molecules that substantially reduce the development of rodent malaria parasites [##REF##15185949##21##, ####REF##12024215##22##, ##REF##12167627##23####12167627##23##] and human dengue virus within mosquitoes [##REF##16537508##24##], which has fuelled optimism that mass-release of laboratory-reared genetically-modified individuals could reduce disease transmission. The greatest unknown with respect to the feasibility of this approach is whether genetically-modified mosquitoes would be able to survive and successfully compete for mates against their wild counterparts outside of the confines of the laboratory. Initial laboratory studies indicated that transgenes impose fitness costs which reduce the reproductive success of the mosquito bearer [##REF##12595691##25##, ####REF##14711992##26##, ##REF##15082552##27####15082552##27##]. A recent study suggests this disadvantage can be reduced by use of out-crossed mosquito lines [##REF##17372227##28##], although so far only under conditions where exposure to parasites is substantially greater than mosquitoes encounter in the wild. While this improvement is encouraging, it does not address the problem that all laboratory-reared mosquitoes, regardless of their genotype, may have poor competitive ability in the wild. For example, recent comparative analysis of <italic>Anopheles gambiae s.s</italic>. in captivity and in nature in southern Tanzania suggest free-living males are larger and have greater lipid reserves than those reared under apparently optimal laboratory conditions [##REF##17690243##29##]. Regardless of whether this reduction in energetic reserves was due to selection for smaller individuals during the colonization process and/or sub-optimal conditions of insectary environments, it suggests laboratory-reared mosquitoes could be at a sizeable disadvantage to their wild counterparts.</p>", "<p>Additional studies have shown that the mating success of male mosquitoes depends on subtle variation in environmental conditions experienced during larval development [##UREF##4##30##,##REF##16197541##31##], which may not be fully captured in mass-rearing facilities. These limitations are thought to have been responsible for the failure of many genetic control trials during the 1970's and 80's, which found that laboratory-reared male mosquitoes were unable to compete in the wild [##REF##16701369##32##]. Clearly, to avoid repeating these failures with the new generation of transgenic mosquitoes, intermediary testing grounds between the laboratory and field within disease-endemic countries are needed.</p>", "<p>Semi-Field Systems (SFS) have been proposed as the best mechanism for bridging this gap. A semi-field system is here defined as an enclosed environment, ideally situated within the natural ecosystem of the target disease vector and exposed to ambient environmental conditions, within which all features necessary for its lifecycle completion are present [##REF##12537599##33##]. In the case of mosquito vectors of human disease, this typically involves a large outdoor cage in which the movement of the disease vector of interest either in or out of the unit is restricted by netting, and within which features such as aquatic larval habitats, blood hosts for adult females, sugar sources (plants) for adults, appropriate resting sites (houses, cattle sheds, etc.) and environmental features (e.g swarm markers to stimulate mating), are present. There are no general guidelines for the appropriate size of such a unit, but ideally it should be large enough to sustain a population of similar density to that encountered in the target environment for numerous generations.</p>", "<p>This definition of a SFS differs from others that apply 'semi-field' to studies that actually involve observation of vectors in a non-contained setting or habitat, where only one part of its life cycle is present [##REF##16771214##34##]. A major goal of SFS is to establish multiple generations of a vector population within a contained setting, without outside intervention [##REF##1085660##35##] in addition to facilitating short-term behavioural or ecological studies based on a single cohort. The main advantage of this approach is that because the abundance and composition of vectors within the SFS can be known, and if desired experimentally manipulated (either at the time of introduction, or through removal of some target individuals), much more precise estimates of the value and variability of demographic and life-history parameters can be obtained than would be from the field. Additionally, they allow researchers to conduct high throughput assays of control tools and ecological phenomena year round without risk of exposure to infection, as all mosquitoes used within the SFS will be free of parasites.</p>", "<p>The concept of simulating the natural environment within contained settings in order to experimentally test ecological hypotheses does not originate in medical entomology. This approach has a long history in aquatic ecology, where hundreds of studies have successfully employed pond meso- and microcosms to examine the impact of biotic and abiotic factors on population and community dynamics [##UREF##5##36##]. Furthermore, neither is this approach new within medical entomology. Almost 70 years ago, Hackett and Bates [##UREF##6##37##] commented on this need for ecological experimentation within natural disease transmission settings: \"The study of behavior under natural, semi-natural and laboratory conditions necessitates locating the laboratory at the source of material. Self evident as this may seem, there are very few laboratories of this kind functioning at present in malarial regions\". Since that time, only a handful of attempts have been made to create large-scale research facilities within semi-natural conditions in disease endemic settings, with the majority being initiated only in the last decade (Table ##TAB##0##1##). Early work in Albania and India used outdoor cages (&lt; 75 m²) to conduct basic ecological observation of anopheline species [##UREF##6##37##,##UREF##7##38##]. Thirty years later this approach was revived for comparative evaluation of different genetically-based population suppression methods for the Indian vectors <italic>Aedes aegypti </italic>and <italic>Culex fatigans </italic>[##REF##1085660##35##,##REF##56331##39##,##UREF##8##40##] but was discontinued after the abandonment of the Sterile Male Release programme that motivated this research [##UREF##9##41##]. Within the last decade, several research programmes in Africa, Asia, Europe and Australia have revitalized SFS for examination of mosquito vector ecology and control (Table ##TAB##0##1##). This approach has been used particularly productively in western Kenya [##REF##12537599##33##], where SFS studies of the malaria vector <italic>An. gambiae s.s. </italic>within 85 m²modified greenhouses have yielded valuable insights into basic ecology and vector-parasite interactions [##REF##12296972##42##, ####REF##15350861##43##, ##REF##15189235##44####15189235##44##] and novel control and monitoring methods [##REF##11963983##45##, ####REF##12389946##46##, ##REF##12174767##47##, ##REF##16700902##48####16700902##48##]. Here the establishment of what is currently the largest SFS in the world for the purpose of experimental study of the ecology and control of African anopheline malaria vectors is described. This facility was built over a two-year period at the Ifakara Health Institute (2004–2006) and is the site of several new studies on vector behaviour, ecology and control.</p>" ]

[ "<title>Materials and methods</title>", "<title>Study site</title>", "<p>The SFS was estabished at the Ifakara Health Institute (IHI) located in the Kilombero district of southern Tanzania. Malaria transmission intensities within this area are amongst the highest described for sub-saharan Africa [##REF##8103627##49##,##REF##12950662##50##]; with annual entomological inoculation rates exceeding three hundred infectious bites a year in some locations [##REF##8103627##49##,##REF##8765455##51##,##UREF##10##52##]. The major malaria vectors in this region are <italic>Anopheles arabiensis, An. gambiae s.s</italic>. and <italic>An. funestus </italic>[##UREF##10##52##, ####REF##7785522##53##, ##REF##11304064##54####11304064##54##].</p>", "<title>SFS site selection</title>", "<p>The crucial first step in establishing a SFS is identifying an appropriate site that adequately captures the environmental conditions experienced by local mosquito species. Additional logistic criteria include ease of access by research personnel and electricity/water supply, being situated where potential hazards to surrounding residents arising from accidental vector release are negligible, and continual monitoring by security staff is possible. Trade-offs may arise in attempting to maximize all these criteria at particular locations which will require careful case-by-case consideration. For example, it has been suggested that the best way to limit hazards posed by unintentional release of mosquitoes into the environment would be to build containment units as far away from communities as possible [##UREF##11##55##]. However, the majority of SFS currently in existence and being planned are located within disease-endemic settings in the developing world. In many of these settings, access to roads, water, an electrical supply, and reliable 24-hour surveillance is possible only near towns or cities. In balancing these components of potential risk, it was decided to select a site for the SFS that is within the campus of the IHI, which is located in Ifakara town. By building within the fenced-off perimeter of the research centre, it was possible to ensure constant surveillance and containment, and strictly control those who had access to the SFS.</p>", "<p>Another key factor in the site selection process for SFS is the availability of background data on the dynamics of local vector populations and their disease transmission ability [##UREF##11##55##]. This information is essential to examine how closely the behaviour, life-history and population dynamics of contained vectors represent those of the wild. As mosquitoes in the SFS will be exposed to many of the same environmental conditions as those of neighbouring populations (e.g temperature, humidity, vegetation), it is anticipated they will be subject to similar selective forces. However, one deviation from complete 'naturalness' was made in the IHI SFS by covering its roof with polyethylene plastic; a decision taken on the basis that this compromise would permit experimental manipulation of rainfall in future experiments. How this modification influences the environmental suitability of the SFS relative to ambient conditions can be assessed by comparison of mosquito population dynamics in the SFS with those of the surrounding area. An advantage of selecting a site in Ifakara was that substantial baseline epidemiological and entomological information on the dynamics of malaria and <italic>Anopheles </italic>populations in the area is already available [##REF##12950662##50##,##REF##11304064##54##,##REF##14728609##56##,##UREF##12##57##]. Additionally, detailed knowledge of mosquito ecology exists for the Kilombero valley, and new studies specifically addressing the mating biology [##REF##17690243##29##, ####UREF##4##30##, ##REF##16197541##31####16197541##31##,##REF##16573828##58##] and population genetics (Ng'habi et al., in prep.) of <italic>An. gambiae </italic>and <italic>An. arabiensis </italic>within this region were initiated concurrently with the establishment of the SFS.</p>", "<title>Planning and design</title>", "<p>Given that Ifakara town is occasionally subject to flooding during the rainy reason, it was decided that the entire SFS structure should be raised 1.6 m above ground level to ensure floodwaters would not breach the structure even during heavy precipitation. The SFS was thus mounted on top of a 22 × 30 m steel-reinforced concrete platform of 0.16 m thickness. This platform was supported by 56 steel-reinforced concrete posts (1.1 m × 1.1 m) equidistantly spaced along the length and width which would allow for natural water flow to continue unimpeded under the structure during times of heavy floods.</p>", "<p>The SFS outer was built from a pre-fabricated greenhouse frame (Shelter 9600, Filclair, Venelles, France). This structure originally consisted of 3 connected compartments of 9.6 × 21 m, but was modified by subdividing the first section into two units of 9.6 × 9 m and 9.6 × 12 m respectively (Figure ##FIG##0##1##). Rather than leaving the roof exposed to natural climatic conditions, it was covered with thick opaque white polyethylene plastic to guarantee protection from intense seasonal rains. The walls of the SFS were covered by PVC coated polyester netting of 346 holes per inch²(Polytex UK), which generates a mesh width approximately two times smaller than the standard recommended for bed nets [156 holes per inch2, [##UREF##13##59##]]. This product was selected on the basis that its filaments were woven together which prevents the mesh being stretched, its high degree of porosity (81%) which facilitates air movement, a shade factor of 56.5% to help reduce temperatures, and its UV-stabilization. After installation of the netting, data loggers (Tinytag TV-1500, Gemini Data Logger, UK) were placed in several areas of the SFS and surrounding outside environment to record temperature variation (taking readings approximately once every 10 minutes).</p>", "<title>Ethical considerations and community awareness</title>", "<p>A potential risk of using SFS in disease-endemic settings is the accidental release of vectors into an environment where they could become infected with a human pathogen, increase the size of the local vector population or introduce a novel phenotype with enhanced transmission capacity. Consequently great care and vigilance is required to ensure the physical integrity of the structure and containment protocols. Access to the IHI SFS is restricted to a small number of research personnel. Research technicians conduct weekly intensive inspections of all areas of the inner and outer structure for physical damage that could allow mosquitoes to escape or enter from outside.</p>", "<p>In addition to making sure that mosquitoes do not escape from a SFS, it is also imperative to ensure that malaria parasites are not accidentally introduced through mosquito contact with an infected person. A protocol for weekly malaria screening for all those working within the SFS was developed. Individuals found to be infected during this screening would be immediately treated with appropriate first line anti-malarial medication and excluded from the screen-houses for one month. Should it be found that a staff member has had malaria parasites while working within the SFS, the experimental chamber in which they worked can be shut down and all mosquitoes within it killed (by depriving them of water, blood and breeding sites for at least two weeks) to ensure no potentially infectious mosquitoes remain within it. Additional methods to reduce the risk of the unintentional introduction of parasites include the use of non-amplifying animal hosts such as livestock as the main blood source for captive vector populations. This procedure is being adopted in the IHI SFS where cows are used as the blood source for free-living <italic>Anopheles </italic>populations.</p>", "<p>In addition to the precautions described above, a key ethical requirement of working with SFS is the creation and maintenance of strong support and awareness within the local community for these research activities. A series of public meetings with IHI staff, workers involved with the construction of the SFS, district health and government officials, and local residents were held in which information on the function and purpose of the SFS was disseminated. Ethical clearance from both the IHI Institutional Review Board (IHDRC/EC4/CL.N96/2004) and Tanzanian National Institute of Medical Research (NIMR/HQ/R.8a/Vol.IX/345) for SFS studies was obtained before the start of this study.</p>" ]

[ "<title>Results</title>", "<title>Constructing the SFS</title>", "<p>Construction of the SFS began in July 2005. Work began by clearing all vegetation from the site, leveling the ground, and digging 56 holes (1 m depth) in the soil for the foundation platform posts (Figure ##FIG##1##2a##). Due to limited access to cement mixers and a constant supply of electricity, all cement required for the construction (approximately 250 m³) was mixed and poured by hand (Figure ##FIG##1##2b##). Approximately 20 full-time labourers were engaged in constructing the foundation over a 3-month period. Once the foundation had been completed, the pre-fabricated greenhouse frame with netting fitted was assembled over a period of 2 weeks (Figure ##FIG##1##2c##). Angled gutters were installed along the outside edge of each compartment to prevent the accumulation of rainwater on the roof (Figure ##FIG##1##2d##). Two electricity points were fitted into each compartment. Drainage and water pipes were fitted into each of the 4 compartments. Soil to a depth of 30 cm was added to sections 3 and 4 of the structure (Figure ##FIG##0##1##). Prior to adding soil to these compartments, sand and rocks were used to construct a drainage system to draw runoff from the soil towards outflow pipes (Figure ##FIG##1##2e##). Two main entrances were built at either end of the SFS, the front being accessible by a 6 m concrete ramp that permits livestock movement, and the posterior by stairs. Double-entry doors were constructed at both main entrances, and between section 1 and 3 (Figure ##FIG##1##2f##). The entire outer structure, including electricity and mains water supply was completed by October of 2006 (Figure ##FIG##2##3a##).</p>", "<title>Establishing research activities</title>", "<p>Different research activities were allocated to each of the four SFS sections on the basis of maximizing logistical efficiency and minimizing the risk of mosquito escape or entry from outside. The first section behind the main front access point (section 1, Figure ##FIG##2##3b##) was designated for use as an <italic>An. arabiensis </italic>insectary. All mosquitoes in this section are thus additionally contained either in adult cages, or larval trays covered with netting. Initial consultation with the greenhouse manufacturer suggested that the netted outer walls would allow temperatures inside the SFS to equilibrate with ambient conditions outside. Ambient temperatures during the hot rainy season in Ifakara can exceed 40°C for several hours each day which are sufficiently high to kill adult and larval mosquitoes. To buffer the insectary from extreme temperatures that could knock out the colony, a traditional thatch roof was built within the insectary area to provide additional shading and cooling (Figure ##FIG##2##3b##). Under these conditions, an average of 638 pupae per day (± 114.8) were obtained from the F1 generation of <italic>An. arabiensis </italic>collected from a nearby village in March 2008.</p>", "<p>The insectary connects directly onto two experimental spaces; the first being a 9.6 m × 10 m chamber within which an experimental hut (3.5 m × 4 × 2.5 m) was constructed for studies of mosquito host seeking and house entry behaviour (Figure ##FIG##2##3c##). This hut was fitted with 6 window traps that can be used to capture mosquitoes leaving and/or attempting to enter the house while a live host is within it [##REF##12109716##60##]. This experimental hut section is designated for short-term behavioural studies in which no more than 300 mosquitoes at a time are released (at dusk), and subsequently recaptured the next day and removed in a cage. A further screen door separates this section from the insectary area meaning that three security doors must be passed through before reaching the outside, and minimizing the risk of mosquito escape during exit or entry.</p>", "<p>The insectary also connects directly to a 9.1 × 21 m chamber designated for establishment of a free-living, self-replicating <italic>An. arabiensis </italic>population within a realistic ecosystem (Figure ##FIG##2##3d##). This section is intended for study of <italic>Anopheles </italic>behaviour, ecology and gene flow within an environment that mimics the natural surroundings as closely as possible. The exact number of free-living mosquitoes that will be held within this unit is uncertain, and will depend upon the carrying capacity of the established population at equilibrium. This section is linked to the main insectary by another double entry door system, requiring four doors to be passed through before reaching outside.</p>", "<p>The fourth experimental section (9.1 × 21 m) is set up as an stand-alone experimental unit isolated from all other areas of the SFS, within which studies of olfaction and chemical ecology are ongoing (Figure ##FIG##2##3e##). This section is physically separated from the adjoining central section by thick polyethylene plastic which minimizes the direct flow of air and odours between them. Entry into this section is possible only from the rear SFS double entry door, and not through any other adjoining section. Studies using odour-baited traps to compare the attraction and repellency of different compounds to <italic>Anopheles gambiae </italic>s.s. are being conducted in this section.</p>", "<title>Replicating the natural environment</title>", "<p>As described above, one section of the SFS was set aside for establishment of a free-living population of <italic>An. arabiensis </italic>within conditions that mimic those of the natural environment. To achieve this, a domestic compound consisting of a mud-walled, thatched-roof house (2.6 m × 3 m × 2.5 m, Figure ##FIG##3##4a##), a typical outdoor toilet (1.4 m × 1.7 m × 2 m), and traditional chicken coop (1.8 m × 1.9 m × 2 m, Figure ##FIG##3##4b##) were constructed within this section by local builders. Grasses and other plants that emerged from the soil brought in from the local environment were allowed to grow. Additional plants common to the surrounding environment such as banana (Figure ##FIG##3##4c##), potatoes, rice, and castor bean (<italic>Ricinus communis </italic>L.) were introduced. A sprinkler system was installed so that varying levels of rainfall could be simulated. Five breeding sites were created by burying plastic buckets into the soil (50 cm diameter), adding 5 cm of soil, and filling them with water to a depth of 25 cm (Figure ##FIG##3##4c##). As <italic>An. arabiensis </italic>is somewhat zoophilic [##REF##4420769##61##, ####REF##11706651##62##, ##REF##7709869##63##, ##REF##14710739##64####14710739##64##], regular blood meals can be provided to free-living mosquitoes within this section by introducing a cow or calf for a few nights each week (Figure ##FIG##3##4d##).</p>", "<title>Climatic conditions</title>", "<p>A primary aim was to create climatic conditions within the SFS representative of the natural environment within the Kilombero region. Initial consultations with the greenhouse manufacturers indicated that the netting walls would allow temperatures inside the SFS to equilibrate with those outside. However, hot temperatures substantially higher than what is generally deemed acceptable for <italic>An. gambiae </italic>and <italic>An. arabiensis </italic>survival and reproduction (e.g. &gt; 30°C for several hours each day) were soon observed within the SFS. These periods of high temperature, however, were similar to those of ambient conditions nearby but outside of the SFS where temperatures at ground level exceed 40°C up to 8 hours each day during the hot rainy season (Figure ##FIG##4##5##). Thus, although temperatures within the SFS were above the threshold for adult mosquito survival for periods of the day, they did not in general differ in mean or variability from those experienced in the nearby environment (e.g May 9–14^th2008: mean temperature inside SFS: 34.24°C ± 10.64°C SD, mean temperature outside the SFS: 34.33°C ± 11.20 SD). For mosquitoes to survive periods of excessively high temperatures both in nature and within the SFS, environmental refugia of substantially lower and less variable temperatures such as houses must be available [##UREF##14##65##]. The simple shaded refugia that were constructed within several areas of the SFS successfully reduced temperatures to within the acceptable range for adult and larval survival (Figure ##FIG##5##6##). The average temperature within the mud-walled house in the central SFS section was 3.5°C lower than within exposed areas of the SFS, and was substantially less variable (Table ##TAB##1##2##). Notably, temperatures inside the mud house did not exceed 35°C which is a critical threshold above which <italic>An. arabiensis </italic>in the laboratory begin to exhibit avoidance behaviour [##REF##15385063##66##]. At 29.20°C (± 3.29°C), the average temperature in our artificial larval habitat was also within the natural range observed in <italic>An. gambiae s.l. </italic>aquatic habitats in east Africa, and did not exceed the upper tolerable limit of 40°C [##REF##17038186##67##] (Table ##TAB##1##2##, Figure ##FIG##5##6##). The construction of a simple thatched roof over the insectary section of the SFS reduced temperatures by approximately 4°C in comparison to exposed areas of the SFS (Table ##TAB##1##2##, Figure ##FIG##5##6##), and considerably reduced the maximum temperature from 51.91°C to 34.69°C. Thus the climatic conditions within the SFS successfully represented the range of temperature extremes experienced in nearby field conditions, while providing realistic environmental refugia with temperatures appropriate for mosquito growth, survival and reproduction.</p>" ]

[ "<title>Discussion</title>", "<p>In just under two years a 625 m²state-of-the-art SFS for large-scale experimentation on anopheline mosquito ecology and control was established within a remote area of southern Tanzania where malaria transmission intensities are amongst the highest ever recorded [##REF##8103627##49##, ####REF##12950662##50##, ##REF##8765455##51##, ##UREF##10##52####10##52##,##REF##9715940##68##]. This unique facility is more than 4 times larger than any SFS previously or currently in existence, and has capacity for a wide variety of research activities including mass-rearing of African malaria vectors under natural conditions, high throughput evaluation of novel control and trapping techniques, short-term assays of host-seeking behaviour and olfaction, and long-term experimental investigation of anopheline population dynamics and gene flow within a contained environment that simulates a local village domestic compound. This was accomplished through a multidisciplinary collaboration between entomologists, senior public health scientists of the IHI, architects, engineers, site managers and a dedicated team of labourers who built this structure largely in the absence of electricity or any other mechanized construction aids.</p>", "<p>Experimental activities have only recently been initiated within the SFS, and the ultimate value of this facility as a research tool will be realized as the studies now underway reach conclusion. Preliminary results from short-term behavioural assays of <italic>An. gambiae </italic>host-seeking behaviour using odour-baited traps and live animal baits suggest that realistic and repeatable results can be obtained within the SFS in a relatively short period of time (I. Lyimo &amp; S. Moore, pers. comm.). The longer-term task of establishing a self-replicating, free-living population of <italic>An. arabiensis </italic>within simulated village conditions is currently underway. Although too early to forecast the outcome of this objective, the fact that an amenable spectrum of climatic conditions can be generated within the SFS is encouraging. Although daily temperatures within exposed areas of the SFS routinely exceeded the optimum temperature of <italic>An. gambiae </italic>[26.5 C under insectary conditions, [##REF##13684306##69##]] for some periods of the day, they were not significantly higher than those of the natural environment immediately outside of the SFS. Had the aim been to create an insectary facility for efficient mass production of <italic>An. gambiae</italic>, the regular daily periods of excessive ambient temperatures within the SFS (&gt;35°C) would be a cause for concern. However the goal was instead to simulate the ambient climatic conditions within the Kilombero region and this was accomplished. Furthermore, the features that were built within the SFS provided microclimatic refuges in which mean temperature and variability was substantially reduced and stayed within the acceptable limits for adult survival and reproduction. For example, the mean temperature within the mud house inside the SFS was 3.5°C lower than exposed areas of the SFS. This observation matches reports from South Africa of air temperature inside mud and thatch houses being 3–6°C cooler than ambient conditions [##UREF##15##70##]. Temperatures within the mud house in our SFS were significantly higher than those reported in a similar structure within the SFS at Mbita, western Kenya [##REF##12537599##33##], which may be more reflective of the different climatic conditions between study sites than structural differences in SFS design. Water temperatures within the artificial larval habitats in our SFS were higher than the reported optimal value gauged from insectary studies [24–26 C, [##REF##14641976##71##]] and those reported for the Mbita SFS [##REF##12537599##33##], but remained within the natural range observed in <italic>An. gambiae s.l. </italic>larval habitats in east Africa [##UREF##16##72##,##UREF##17##73##]. Given that free-living <italic>An. gambiae s.s. </italic>were able to complete their life cycle within the slightly cooler and much smaller confines of the Mbita SFS [##REF##12537599##33##] there is optimism that the same can be achieved in the IHI SFS with <italic>An. arabiensis</italic>, a species known to have greater tolerance of hot and arid environments [##REF##15385063##66##,##REF##9633110##74##].</p>", "<p>While early observations are promising, much still remains to be known about how representative conditions inside the SFS will be of mosquito ecology in the wild. Open questions include whether a self-replicating population can be maintained over numerous generations on this spatial scale, what carrying capacity this population will reach under ambient climatic and host (bovine) conditions, whether additional climatic refugia or controls will be needed, and if existing plant and nectar sources within the SFS will be sufficient to maintain the adult male population. Importantly, the identification of limitations in the ability of our SFS to replicate natural mosquito dynamics as experimental work progresses will in itself provide valuable knowledge of the crucial determinants of anopheline population growth and stability that would not be possible under natural field conditions.</p>", "<p>As research begins at the SFS in Ifakara and similar facilities around the world (Table ##TAB##0##1##), it is useful to consider the major challenges to the successful use of this research tool. These challenges are varied and range from the purely scientific to those of logistics and ethics. Five key areas merit discussion. The first is the possibility that although biological inferences made from SFS may be much more realistic than those from cage studies, they may still misrepresent some areas of mosquito ecology and population processes in nature. For example, although full life cycle completion of <italic>An. gambiae s.s. </italic>was achieved within the SFS in western Kenya, it was noted that the artificial breeding sites within it gave rise to considerably fewer larvae than expected [##REF##12537599##33##]. Whether this reduced efficiency was due to a problem with the environmental conditions inside the SFS, or maladaptation of the laboratory population used in these experiments to ambient conditions is unknown, but suggests there could be unique constraints or bottlenecks acting on population growth within these systems. Conversely, absence of the full range of environmental risks within the SFS such as stochasticity in host encounter rates, insecticide treated bed nets, predation by small vertebrates, pathogens and extreme environmental conditions such as flooding may result in an overestimation of life-history and demographic rates. For example, Knols <italic>et al </italic>estimated the daily survival of <italic>An. gambiae s.s </italic>within their SFS to be 90% which is higher than reported in many field studies [##REF##11289661##75##]. In order to reduce the risk of accidental parasite introduction, non-amplifying animal hosts will be used as the main source of blood in many SFS. While numerous disease vectors include the blood of non-human animals in their diet, many of the species that are most problematic exhibit a pronounced preference for humans [##UREF##18##76##]. Several studies have shown that the fitness haematophagous insects derive from blood varies with host species [##REF##5448824##77##, ####REF##1238572##78##, ##REF##2280395##79##, ##REF##11372967##80##, ##UREF##19##81##, ##REF##12518856##82####12518856##82##]. It is also known that selection for divergent preference for human or cow hosts in <italic>An. gambiae </italic>mosquitoes can be generated in as little as 5–6 generations of selection [##REF##14204067##83##]. Thus constantly exposing vector populations within SFS to non-human hosts could result in the generation of individuals with different phenotypes, genotypes and population dynamics than those who feed on and transmit disease to humans. Continued monitoring and comparison of SFS results to those observed in the field will be useful to identify which, if any, of these issues pose serious obstacles to interpretation and reinforce the point that SFS studies are intended to complement but not replace field studies.</p>", "<p>A second scientific concern is that vector populations established within SFS will likely be considerably smaller than those in the wild and thus experience inbreeding and a resultant reduction in genetic diversity which could impede fitness. It is well known that genetic diversity within insect vectors can be considerably reduced during laboratory colonization [##REF##9288819##84##, ####REF##11296845##85##, ##UREF##20##86####20##86##]. Free-living populations established within SFS may be considerably larger than typical laboratory colonies and thus avoid a similar intensity of inbreeding, however it is unlikely they will escape some bottle-necking and an associated loss of diversity from founder populations. Within only a few generations of laboratory colonization, mosquitoes can develop significant behavioural divergence from wild populations which restricts mating between them [##UREF##21##87##]. This phenomenon may also occur within SFS, although perhaps at a slower rate than in small laboratory cages. As genetic and phenotypic divergence between contained SFS and wild populations may be unavoidable, the need for repeated comparative sampling of individuals in both settings is advocated to track if and how genetic diversity is reduced in captivity, and provide guidelines for how frequently captive populations should be enriched by fresh genetic material to maintain representative levels of diversity.</p>", "<p>Should self-replicating vectors be successfully established in SFS, a logistical obstacle to the estimation of precise demographic rates from them will be the problem of disentangling overlapping generations. While much more precise estimates of mosquito population size will be possible within the contained environment of an SFS than in nature, it will remain difficult to accurately monitor individual-level activities such as mating behaviour and resource acquisition, and its resultant impact on fitness. The development of novel marking schemes using stable isotopes [##REF##17371925##88##,##UREF##22##89##] or distinct genetic traits may permit more precise monitoring of the behaviour and reproductive success of specific subsets of individuals, or individuals themselves.</p>", "<p>For greatest public health relevance, SFS should be situated within or as near as possible to natural disease transmission environments as possible. Placing a large contained population of competent disease vectors within an appropriate transmission setting will always raise biosecurity concerns. Any breach of containment could result in increasing the disease transmission within the local area, and the accidental introduction of parasites into contained populations from asymptomatic carriers working within the facility could also generate the potential for infection. Awareness and discussion of how to prevent those risks are absent from early accounts of SFS use, but are justifiably coming to the forefront as plans for large-scale studies with genetically-modified disease vectors come under development. Recently an international committee of scientists formalized guidelines on recommended biosecurity measures and precautions for contained SFS trials with genetically-modified mosquitoes [##UREF##11##55##]. The publication of these guidelines represents a significant step forward in thinking regarding the ethical responsibility for good practice within these facilities.</p>", "<p>A final, crucial issue for the expansion of SFS-based research programmes throughout the world is the need to engage and promote awareness within the communities that host these facilities. The communities surrounding SFS research facilities should be the primary beneficiaries of research conducted within them, and their particular needs as end-users must be kept in mind when using these facilities to trial new vector control strategies. While researchers working in SFS may have this goal clearly in mind, it will be of little value unless clearly and regularly communicated to local communities in an open and discursive manner. Understandably local residents may be apprehensive about the placement of an SFS containing live insect vectors near their home, and misinformation about the purpose of this work and risks associated with it could cause considerable friction. This lesson was painfully learnt by scientists working at the Indian Council for Medical Research in the 1970's who had their research unit on <italic>Aedes </italic>mosquitoes shut down when journalists falsely alledged that the actual purpose of the work was biological warfare and human population control [##UREF##9##41##]. That such a debacle could occur in one of the most successful disease vector control research programmes of all time [with 104 papers published in 6 years and numerous insights into population suppression gained, [##UREF##9##41##]] is a sobering thought for all those involved in this new generation of SFS research. Community awareness activities have begun in Ifakara, and must be sustained and scaled up if both local community members and researchers working at the SFS are to obtain maximum benefits from this research facility.</p>", "<p>Engagement must extend beyond local communities to include scientists and students working within the disease endemic countries that host SFS. These facilities can provide substantial indirect benefits by acting as state-of-the-art training tools for young vector biologists in which methodological skills can be honed, and independent research hypotheses experimentally tested in a disease-free setting. Currently, at the IHI, there are three east African postgraduate students pursuing their PhD studies on research based within the SFS and plans to recruit several more underway. Thus this research tool will contribute to the IHI's goal of substantially increasing Ph.D-level capacity in malaria vector research within Tanzania and east Africa. Much of the recent motivation for initiating SFS programmes has been driven by laboratory-based research on genetic modification of disease vectors that has occurred almost exclusively in developing countries. For both the transgenic approach and other emerging vector control strategies to fulfill their potential, it is absolutely imperative that endemic country scientists are actively involved in driving SFS-based research and taking forward innovative techniques developed within it.</p>" ]

[]

[ "<p>This is an Open Access article distributed under the terms of the Creative Commons Attribution License (<ext-link ext-link-type=\"uri\" xlink:href=\"http://creativecommons.org/licenses/by/2.0\"/>), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.</p>", "<title>Background</title>", "<p>Medical entomologists increasingly recognize that the ability to make inferences between laboratory experiments of vector biology and epidemiological trends observed in the field is hindered by a conceptual and methodological gap occurring between these approaches which prevents hypothesis-driven empirical research from being conducted on relatively large and environmentally realistic scales. The development of Semi-Field Systems (SFS) has been proposed as the best mechanism for bridging this gap. Semi-field systems are defined as enclosed environments, ideally situated within the natural ecosystem of a target disease vector and exposed to ambient environmental conditions, in which all features necessary for its life cycle completion are present. Although the value of SFS as a research tool for malaria vector biology is gaining recognition, only a few such facilities exist worldwide and are relatively small in size (&lt; 100 m²).</p>", "<title>Methods</title>", "<p>The establishment of a 625 m²state-of-the-art SFS for large-scale experimentation on anopheline mosquito ecology and control within a rural area of southern Tanzania, where malaria transmission intensities are amongst the highest ever recorded, is described.</p>", "<title>Results</title>", "<p>A greenhouse frame with walls of mosquito netting and a polyethylene roof was mounted on a raised concrete platform at the Ifakara Health Institute. The interior of the SFS was divided into four separate work areas that have been set up for a variety of research activities including mass-rearing for African malaria vectors under natural conditions, high throughput evaluation of novel mosquito control and trapping techniques, short-term assays of host-seeking behaviour and olfaction, and longer-term experimental investigation of anopheline population dynamics and gene flow within a contained environment that simulates a local village domestic setting.</p>", "<title>Conclusion</title>", "<p>The SFS at Ifakara was completed and ready for use in under two years. Preliminary observations indicate that realistic and repeatable observations of anopheline behaviour are obtainable within the SFS, and that habitat and climatic features representative of field conditions can be simulated within it. As work begins in the SFS in Ifakara and others around the world, the major opportunities and challenges to the successful application of this tool for malaria vector research and control are discussed.</p>" ]

[ "<title>Competing interests</title>", "<p>The authors declare that they have no competing interests.</p>", "<title>Authors' contributions</title>", "<p>HMF was the primary project coordinator of SFS establishment at the IHI and drafted the manuscript. KRN designed and set up the areas of the SFS where a free-living <italic>An. arabiensis </italic>population will be established in simulated village conditions, and assisted with collection of temperature data. TW was the lead architect and DK the head of the maintenance unit that carried out construction. SJM oversaw parts of the construction and set up the olfaction study chamber. IL designed and set up the experimental hut and insectary area of the SFS. TR oversaw parts of the construction and helped supervise research activities within it. HU and HM provided institutional support and guidance in logistics, ethics, and community sensitization. GFK provided support with project planning and coordination. BGJ initiated this project, obtained financial support for it, and provided scientific and logistical guidance.</p>" ]

[ "<title>Acknowledgements</title>", "<p>This work is dedicated to the people of Kilombero and Ulanga districts, Tanzania in the hope that the facilities described here will help bring practical solutions to reduce the unacceptable burden that malaria places upon their communities. The authors express their sincere gratitude to the TTCIH maintenance team whose commitment made this project feasible. We thank the IHI administration for their excellent guidance in project management, and all members of the Public Health Entomology team for their support. This work was supported by a VIDI grant (no. 864.03.004) awarded by the Dutch Scientific Organization (NWO) to BGJK and BBSRC David Phillips Fellowship to HMF, and the Bill and Melinda Gates Foundation (grant # GCGH121). We thank the Filclair corporation for provision of the SFS at reduced cost, and their technical support.</p>" ]

[ "<fig position=\"float\" id=\"F1\"><label>Figure 1</label><caption><p><bold>Schematic diagram of the IHI Semi-field system (SFS) for research on African <italic>Anopheles </italic>ecology and control</bold>.</p></caption></fig>", "<fig position=\"float\" id=\"F2\"><label>Figure 2</label><caption><p><bold>Key steps in the construction of the IHI SFS.</bold> (a) digging holes for foundation posts, (b) pouring the concrete foundation platform, (c) installing the netting, (d) roof gutters draining precipitation during peak rainfall, (e) French drain system installed under soil to divert surface water run off, (f) double entry door system.</p></caption></fig>", "<fig position=\"float\" id=\"F3\"><label>Figure 3</label><caption><p>IHI SFS on completion (a) outer structure, (b) insectary section with thatched roof, (c) experimental hut trial area, (d) section for establishment of free-living, self-replicating <italic>An. arabiensis </italic>population, (e) section for olfaction and chemical ecology research.</p></caption></fig>", "<fig position=\"float\" id=\"F4\"><label>Figure 4</label><caption><p>Habitat features within the SFS section designated for a free-living <italic>An. arabiensis </italic>population: (a) traditional mud-walled house, (b) chicken coop with clay pot refugia, (c) artificial breeding site and banana plant, (d) Cattle shed containing calf host.</p></caption></fig>", "<fig position=\"float\" id=\"F5\"><label>Figure 5</label><caption><p>Average hourly temperatures at ground-level within the central section of the SFS and a nearby site outside of the SFS (3 m away) from May 9 – 14^th2008.</p></caption></fig>", "<fig position=\"float\" id=\"F6\"><label>Figure 6</label><caption><p>Average hourly temperatures at different locations within the SFS in the period from Feb 29^th– March 9^th2008.</p></caption></fig>" ]

[ "<table-wrap position=\"float\" id=\"T1\"><label>Table 1</label><caption><p>Previous and current location, size, target species and research aims of Semi-Field Systems (SFS) established for mosquito vector research.</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td align=\"left\">Country</td><td align=\"left\">Year*</td><td align=\"center\">Dimensions<break/> (m)</td><td align=\"center\">Number of units</td><td align=\"left\">Mosquito Species</td><td align=\"left\">Purpose</td><td align=\"left\">Refs</td></tr></thead><tbody><tr><td align=\"left\">Albania</td><td align=\"left\">1939</td><td align=\"center\">10 × 5 × 6</td><td align=\"center\">1</td><td align=\"left\">Various European anophelines</td><td align=\"left\">Basic ecological studies</td><td align=\"left\">[##UREF##6##37##]</td></tr><tr><td align=\"left\">India – Madras</td><td align=\"left\">1942</td><td align=\"center\">12.2 6.1 × 3.05</td><td align=\"center\">3</td><td align=\"left\"><italic>An. culifacies</italic></td><td align=\"left\">Basic ecological studies, evaluation of genetic control strategies for population suppression</td><td align=\"left\">[##UREF##7##38##]</td></tr><tr><td align=\"left\">India – Delhi</td><td align=\"left\">1976</td><td align=\"center\">5.6 × 3.3 × 2.1</td><td align=\"center\">1</td><td align=\"left\"><italic>Ae. aegypti Cx. Fatigans</italic></td><td/><td align=\"left\">[##REF##1085660##35##,##REF##56331##39##,##UREF##8##40##]</td></tr><tr><td align=\"left\">Kenya</td><td align=\"left\">2002</td><td align=\"center\">11.4 × 7.1 × 4.4</td><td align=\"center\">7</td><td align=\"left\"><italic>An. gambiae s.s.</italic></td><td align=\"left\">Basic ecological studies, vector-malaria parasite interactions, evaluation of novel trap designs and repellents</td><td align=\"left\">[##REF##12537599##33##,##REF##12296972##42##, ####REF##15350861##43##, ##REF##15189235##44##, ##REF##11963983##45##, ##REF##12389946##46##, ##REF##12174767##47##, ##REF##16700902##48####16700902##48##]</td></tr><tr><td align=\"left\">Thailand</td><td align=\"left\">2003</td><td align=\"center\">10 × 10 × 4</td><td align=\"center\">1</td><td align=\"left\"><italic>Ae. aegypti</italic></td><td align=\"left\">Basic ecological studies</td><td align=\"left\">[##UREF##23##90##]</td></tr><tr><td align=\"left\">Tanzania – Muheza</td><td align=\"left\">2003</td><td align=\"center\">12.2 × 8.2 × 4.6</td><td align=\"center\">3</td><td align=\"left\"><italic>An. gambiae s.s Cx. quinquefasciatus</italic></td><td align=\"left\">Evaluation of trapping methods, training and basic ecological studies</td><td align=\"left\">No publ.</td></tr><tr><td align=\"left\">Sudan</td><td align=\"left\">2006</td><td align=\"center\">18 × 8 × 2.75</td><td align=\"center\">3</td><td align=\"left\"><italic>An. arabiensis</italic></td><td align=\"left\">Fitness of sterilized males, basic ecological studies</td><td align=\"left\">[##UREF##14##65##]</td></tr><tr><td align=\"left\">Tanzania – Ifakara</td><td align=\"left\">2007</td><td align=\"center\">29.8 × 21 × 7.1</td><td align=\"center\">4</td><td align=\"left\"><italic>An. gambiae s.s An. arabiensis</italic></td><td align=\"left\">Basic ecological studies, evaluation of trapping methods and repellents</td><td align=\"left\">This paper</td></tr><tr><td align=\"left\">Australia</td><td align=\"left\">2008</td><td align=\"center\">17 × 9 × 4.3</td><td align=\"center\">2</td><td align=\"left\"><italic>Ae. aegypti</italic></td><td align=\"left\">Assessment of biocontrol strategy using Wolbachia, basic ecological studies</td><td align=\"left\">No publ.</td></tr><tr><td align=\"left\">Austria</td><td align=\"left\">TBC</td><td align=\"center\">25 × 10 × 3</td><td align=\"center\">TBD</td><td align=\"left\"><italic>An. arabiensis</italic></td><td align=\"left\">Research on Sterile Insect Technique</td><td align=\"left\">No publ.</td></tr></tbody></table></table-wrap>", "<table-wrap position=\"float\" id=\"T2\"><label>Table 2</label><caption><p>Average temperatures at different locations within the SFS from February 29^th– May 9^th2008.</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td align=\"left\">Location</td><td align=\"center\">Average<break/> Temperature (°C)</td><td align=\"center\">Standard<break/> Deviation (°C)</td><td align=\"center\">Range (°C)</td></tr></thead><tbody><tr><td align=\"left\">Ground-level in SFS</td><td align=\"center\">31.24</td><td align=\"center\">9.62</td><td align=\"center\">21.77–51.91</td></tr><tr><td align=\"left\">Inside mud house</td><td align=\"center\">27.84</td><td align=\"center\">2.66</td><td align=\"center\">23.86 – 34.69</td></tr><tr><td align=\"left\">Artificial breeding site</td><td align=\"center\">29.20</td><td align=\"center\">3.29</td><td align=\"center\">25.19 – 36.67</td></tr><tr><td align=\"left\">Thatched-roof insectary</td><td align=\"center\">26.72</td><td align=\"center\">3.58</td><td align=\"center\">22.60 – 34.43</td></tr></tbody></table></table-wrap>" ]

[]

[]

[]

[]

[]

[]

[ "<table-wrap-foot><p>Year refers to the time when the first research publication from these facilities was published, or year of establishment in cases where no published references to these facilities are yet available ('TBC' = to be constructed).</p></table-wrap-foot>" ]

[ "<graphic xlink:href=\"1475-2875-7-158-1\"/>", "<graphic xlink:href=\"1475-2875-7-158-2\"/>", "<graphic xlink:href=\"1475-2875-7-158-3\"/>", "<graphic xlink:href=\"1475-2875-7-158-4\"/>", "<graphic xlink:href=\"1475-2875-7-158-5\"/>", "<graphic xlink:href=\"1475-2875-7-158-6\"/>" ]

[]

{ "acronym": [], "definition": [] }

[ "<title>Background</title>", "<p>Indigenous <italic>vivax </italic>malaria was a common cause of death in the 18^thand 19^thcentury in Finland [##REF##15847704##1##]. The decline of malaria commenced late in the 18^thcentury and malaria gradually decreased without countermeasures during the 19^thand 20^thcentury [##UREF##0##2##]. Medication was sparingly used. At the beginning of the 1930's there were fewer than 10 cases annually and the last cases of indigenous malaria were documented in 1953 and 1954 [##UREF##1##3##].</p>", "<p>During World War II Finland fought two wars with the Soviet Union. The Winter War lasted from November 1939 to March 1940. Allied with Germany, Finland started the Continuation War in July 1941. Finnish troops invaded large parts of Soviet Karelia. In January 1942 the front line became more immobile and the positional warfare lasted to June 1944 [##UREF##2##4##,##UREF##3##5##]. The war was lost in September 1944. During the war recurrent malaria epidemic became an increasing problem among Finnish soldiers. The malaria situation among the Soviet and German soldiers is not known. The Finnish malaria epidemic in 1941–1945 has been studied in detail by Hernberg [##UREF##4##6##, ####UREF##5##7##, ##UREF##6##8####6##8##]. He proposed the widely accepted explanation that infective mosquitoes spread malaria over the front line from Soviet soldiers. The main military events in relation to malaria in southern Finland are presented in Figures ##FIG##0##1##, ##FIG##1##2##, ##FIG##2##3## and ##FIG##3##4##.</p>", "<p>A corresponding positional warfare malaria epidemic commenced in the 1990's in the western part of the demilitarized zone (DMZ) between South and North Korea [##UREF##7##9##,##REF##17170570##10##]. At the beginning of the 20^thcentury <italic>vivax </italic>malaria was common in Korea. During the Korean War (1950–53) there was a peak in malaria among both civilians and soldiers. Malaria declined rapidly in the 1960's and 1970's [##UREF##8##11##]. Both countries were declared officially free of malaria in 1979 [##UREF##9##12##]. Malaria re-emerged in South Korea in 1993 when one soldier got malaria near the DMZ. Later malaria increased rapidly peaking in 2000 when thousands of people, mostly soldiers, got malaria close to the DMZ. It also spread southwards in South Korea. There was a simultaneous epidemic of <italic>vivax </italic>malaria starting in 1995 in North Korea close to the DMZ [##UREF##7##9##]. The number of malaria cases in North Korea reached nearly 300,000 in 2001. The prevalent opinion is that malaria re-emerged in South Korea because of a continuous influx of infective mosquitoes from North Korea where malaria is expected to have persisted [##REF##17170570##10##,##REF##10507220##13##]. Detailed maps of the evolution of the recent epidemic in South Korea have been published [##REF##17170570##10##,##UREF##8##11##].</p>", "<p>In both Finland and Korea the malaria epidemics started in the military forces and only later spread among civilians. The data from Finland and Korea suggest that the key to understanding these epidemics is strongly connected with some special features of military activities. This study aims at resolving the origin of re-emerging malaria among soldiers in positional warfare conditions.</p>" ]

[ "<title>Methods</title>", "<p>Statistics on monthly war malaria in Finland in 1941–1945 was taken from Hernberg's studies [##UREF##4##6##, ####UREF##5##7##, ##UREF##6##8####6##8##]. Annual malaria cases in Finland in 1860–1954 were collected from several publications [##UREF##0##2##,##UREF##1##3##,##UREF##10##14##]. The military preparations in 1941 in Finland are quantitatively quite well known and they are used to explain the dynamics of the malaria epidemics during 1941–1945 [##UREF##2##4##,##UREF##3##5##]. Statistics on Korean malaria was taken from various sources [##UREF##7##9##, ####REF##17170570##10##, ##UREF##8##11##, ##UREF##9##12####9##12##]. The Korean military statistics, however, are less known because of the international tensions. United States sources were used to interpret recent changes in North Korean military strategy [##UREF##11##15##]. The estimation of the troop strengths near DMZ in South Korea and North Korea during the epidemics was indirectly interpreted according to the statistics on malaria prophylaxis [##REF##17170570##10##]. Statistics on urbanization trends in the three countries was collected from scientific or official sources [##UREF##12##16##, ####UREF##13##17##, ##UREF##14##18##, ##REF##18430203##19####18430203##19##]. The criteria for urban and rural settlements varies between countries, but for the present study only the varying steepness of the urbanization rate is relevant independently of particular values.</p>" ]

[ "<title>Results and discussion</title>", "<title>Finland 1941–1945</title>", "<p>Finland mobilized 610,000 people in the armed forces in the Continuation War in 1941 and concentrated some 310,000 men along the front line with Soviet Union in Karelia in the southeast and on Hanko peninsula by the Gulf of Finland (ceded to Soviet Union after the Winter War in 1939–1940). The allied German troops (about 200,000 soldiers) were concentrated in the northern part of Finland [##UREF##0##2##,##UREF##1##3##].</p>", "<p>Annual civil malaria cases in Finland and the Karelian Isthmus from 1923 until the end of Continuation War in 1944 is presented in Figure ##FIG##4##5##. It is notable that malaria had almost disappeared from the Karelian Isthmus during the 1930's but suddenly increased in 1937 before the Winter War. In 1940 and 1941 there are no documented civil cases in that area. Already in July 1941 several malaria cases appeared among Finnish soldiers along both southern frontlines. Monthly military cases of malaria in 1941–1944 and civil cases after the war in 1945 are presented in Figure ##FIG##5##6##. Hernberg [##UREF##3##5##] studied the situation from 1941 to 1945 and used his statistics as a proof of existence of a dormant stage in humans. He concluded that Finnish soldiers at the beginning of the 1940's must have got the infection from infective mosquitoes flying across the front line from the Soviet positions because indigenous malaria was very rare in Finland at that time.</p>", "<p>According to recent studies on historical malaria in Finland, primary infections occur only indoors from mid winter to spring and infective <italic>Anopheles </italic>mosquitoes do not exist in summer [##REF##15847704##1##,##REF##18430203##19##]. The <italic>Anopheles </italic>mosquitoes reach the adult stage in the latter half of July and August. In 1941 the mean temperature of June was cold (about 11°C) and July was warm (about 21°C). As a consequence, <italic>Anopheles </italic>mosquitoes had certainly hatched already in July, but they cannot then have been infective with sporozoites. Because of the low temperature in August the sporogony was slow.</p>", "<p>Because the primary infections of malaria during the summer of 1941 are excluded the early malaria cases can only have been relapses from activated hypnozoites. Relapses are principally expected to be triggered by the bites of mosquitoes in natural conditions [##REF##18430203##19##]. Hypnozoites can remain viable in the human liver at least nine years but probably a much longer time [##REF##18430203##19##]. The main issue is to explain the source of hypnozoite carriers along the front line.</p>", "<p>The basic cause of decline of malaria in Finland is now expected to have been gradual decrease of household size since the latter half of 18^thcentury from about ten to four members [##UREF##15##20##]. The household size of four members seems to be a threshold value for prevalence of malaria. The declining household size has a very high correlation with declining malaria frequency (p &lt; 0.001). According to the Finnish statistics malaria disappeared when the household size decreased below four members in the 1950's. Likewise the European malaria declined in step with declining household size until it disappeared when household size passed below four members in all countries [##UREF##16##21##,##UREF##17##22##]. The mechanism of the decline, however, will be discussed in a separate paper.</p>", "<p>The increasing level of urbanization is another factor that must be considered when interpreting the occurrence of relapses. If we accept that <italic>Anopheles </italic>mosquitoes are the principal triggers of relapses, then relapses should occur in situations where mosquitoes and humans meet. This happened in the rural Finland each year when the new generation of <italic>Anopheles </italic>hatched in July. The rural population decreased gradually when people moved into cities to look for better job opportunities. During the last phase of malaria in the 1920's and 1930's many people were still hypnozoite carriers when they moved into cities. People living in urban conditions in multi-storey buildings are rarely exposed to mosquitoes compared with people in rural conditions. A conclusion is that a certain fraction of the Finnish urban population in the late 1930's still had dormant hypnozoites in their liver. A steep decline of malaria frequency occurred in urban settlements about 20 years later than in rural settlements. Although the number of malaria cases was declining, the number of people carrying hypnozoites was still high enough for a new epidemic during suitable conditions. A suitable condition develops when a large random sample of the population congregates and is stationary for a considerable time. An occasional relapse may be triggered immediately but an epidemic requires a longer time for the new cycles of malaria to develop. Typical situations are positional warfare, refugee camps and large scale public work with provisional lodging during times of unemployment and depression (corresponding to \"frontier malaria\" in the 19^thcentury United States).</p>", "<p>Most of the Finnish malaria cases since the late 1920's were probably relapses. The peak in malaria cases in 1931 is explained by depression and public work. The workers lived in temporary quarters in primitive conditions and were exposed to nocturnal mosquitoes. A major part of the malaria cases in 1937–1939 was clustered within the Karelian Isthmus during intensified fortification work before the war in November 1939. The military actions during the Winter War 1939–1940, however, did not contribute to the malaria cases because of the highly mobile front lines during the cold months. The soldiers slept in a tent or in a lean-to with temperatures down to -40°C. In such conditions mosquitoes cannot be active.</p>", "<p>During the Continuation War, the situation changed. Positional warfare conditions prevailed from January 1940 to June 1944. The soldiers slept in primitive dugouts and barracks (6–24 people in the same space) and remained stationary and continuously exposed to mosquitoes which over wintered together with humans. Because the sleeping places were heated the over wintering mosquitoes were forced to take blood during winter. Any malaria carrier would then have got a relapse triggered by the mosquitoes. Exactly in these conditions the malaria epidemic commenced. The first outbreak revealed in the summer in 1941 that hypnozoite carriers among the soldiers were able to initiate an epidemic. It is worth noting that there were very few civil cases on the Karelian Isthmus during the Continuation War although 280,000 civilians returned to their homes [Figure ##FIG##6##7##].</p>", "<p>According to Hernberg [##UREF##4##6##] there were about 3000 cases of malaria among the Finns during 1941–1945. Because he did not know the true dynamics of relapses his judgement about separating primary infections from relapses cannot be relied on. A crude estimate of 1500 primary infections and 1500 relapses is accepted in this study. About 310,000 men were concentrated along the front line and exposed to mosquitoes. It can therefore be estimated that about 0.5% of the Finnish population still had hidden dormant hypnozoites in their liver in the 1940's. In 1945 after the war 1551 cases of malaria were reported all over Finland from February to October peaking in May and June. 299 patients had malaria two or several times so the remaining 1252 cases represent the true number of malaria infected persons. A detailed study on 868 patients revealed only 12 women. Practically all of them represented repatriated soldiers. Probably nearly all of the cases were relapses. They originated from primary infections that must have taken place during the war in the spring 1944 when many soldiers lived together in primitive conditions. The mosquitoes were then able to transmit malaria between infected and uninfected soldiers during the transmission season from December to May inside the dugouts and barracks. When the infected soldiers were repatriated after the war only those soldiers returning to conditions where they were locally exposed to mosquitoes got relapses. Those soldiers that returned to housing conditions without mosquitoes could still have hidden dormant stages of malaria in their liver a long time after the war. The schema in Figure ##FIG##7##8## presents the principal mechanism of the outbreak of malaria in positional warfare conditions.</p>", "<p>The key factor for the re-emergence of malaria is the number of people sleeping in the same space. The average number of soldiers sleeping in the same space is usually much higher than the empirically found threshold value of four humans per room. The success of <italic>Plasmodium </italic>is linked to interactions between groups of humans. If the groups are too small with two or three humans, the groups become more isolated and the <italic>Plasmodium </italic>faces a higher extinction rate than the re-colonization rate. If the group size is more than four humans, the <italic>Plasmodium </italic>can gradually spread among humans. Principally malaria epidemics among military forces could be avoided if each soldier sleeps alone in a separate space like a room, a tent or a bed net. In such situations occasional malaria cases will always be isolated and more easily controlled than if a large number of people get malaria simultaneously.</p>", "<p>Of special interest in the outbreak of malaria among Finnish troops is that it did not happen in fight situations, it was always behind the immediate front line. Because most <italic>Anopheles </italic>species are night biters (like <italic>A. messeae </italic>in Finland) they usually bite when people are sleeping. The behaviour of an infective mosquito is altered by the <italic>Plasmodium</italic>. Sporozoites in the salivary glands decrease the production of anti-coagulants and vasodilators by the mosquito which in effect forces the mosquito to bite repeatedly to get blood [##REF##18430203##19##]. Another effect is that the chance of the bite getting unnoticed by the victim increases as human skin does not react on the decreased load of alien proteins. That means that all people sleeping in the same space could quickly get malaria by only one infective mosquito. All these circumstances points to a basically indoor transmission pattern of malaria in general.</p>", "<p>Malaria has sometimes been associated with famine and hardship. During 1749 – 1849 there were 27 years associated with famine because of extreme weather conditions (floods, frost or drought with thousands of deaths) in Finland [##UREF##18##23##]. The mean annual number of deaths from malaria was 95.24 for the whole period and 103.07 for the 27 famine years. The difference is statistically insignificant. As a consequence the effect of famine or other hardship on malaria prevalence is expected to be irrelevant.</p>", "<title>Korea 1993–2006</title>", "<p>There are several similarities in the evolution of malaria in Korea and Finland. <italic>Plasmodium vivax </italic>with long incubation time is common to both regions. During the Korean War in 1950–54, there was a high rate of malaria among all military forces. Later it declined in step with decreasing household size and increased urbanization. The average household size in South Korea reached four members in the 1980's. In North Korea, the household size was still higher in the 1990's, varying from 4.2 to 4.7 depending on sources [##REF##10440306##24##].</p>", "<p>Malaria was officially eradicated in both North Korea and South Korea in 1979 [##UREF##9##12##]. But at least in South Korea occasional cases, which were considered relapses after a long incubation, were still reported until 1984 [##REF##16172490##25##]. Re-emergence of malaria commenced in 1993 in South Korea [##REF##17170570##10##] and in 1995 in North Korea [##UREF##7##9##]. The number of cases in North Korea has been nearly 100 times higher than in South Korea. The malaria trend is now reversed in both countries by means of large scale national and international (especially in North Korea) counter measures.</p>", "<p>The re-emergence of malaria has received much attention in South Korea. It is generally believed that it was a result of influx of infective <italic>Anopheles sinensis </italic>mosquitoes (the main vector) from North Korea [##UREF##8##11##,##UREF##18##23##]. This explanation is inconsistent with the fact that during the main transmission season in the summer, the prevailing wind direction comes from the south [##UREF##19##26##,##UREF##20##27##]. Unfed females of the vector species, <italic>Anopheles sinensis</italic>, can fly quite long distances, up to 9–12 km [##REF##12325443##28##], but after a blood feed it remains 2–3 days in a safe place until egg-laying [##UREF##21##29##,##REF##16192749##30##]. After the first egg batch, the female returns within 24 hours to a host to the next blood meal [##UREF##22##31##]. Infective mosquitoes have been collected at very low frequency in Korea, 0.01–0.06% of studied samples [##UREF##18##23##]. It has been shown that <italic>A. sinensis </italic>females are highly zoophilic and prefer cattle [##UREF##21##29##,##REF##16192749##30##]. They take, however, readily blood from humans if cattle are not available [##UREF##21##29##]. It is also striking that malaria foci in South Korea in the 1960's were relatively stationary and malaria did not spread in neighbouring regions by means of flying mosquitoes [##UREF##8##11##]. The South Korean troops near DMZ had about a 6–7 time higher malaria incidence rate than the United States troops stationed in the same region in 1993–2005 [##REF##17170570##10##]. That large difference in the incidence rate would not be expected if infective mosquitoes at random would bite soldiers close to DMZ. Considering the sum of all these factors, large scale spreading of malaria by mosquitoes from North Korea seems unlikely.</p>", "<p>The behaviour of an infective <italic>A. sinensis </italic>is probably altered in the same way as <italic>A. messeae </italic>in Finland. The production of anticoagulants and vasodilators are reduced because of sporozoites in the salivary glands [##REF##18430203##19##]. An infective mosquito female is forced to take repeated blood meals and will not fly far away from the blood source. A plausible explanation is that infected humans themselves are spreading malaria and the role of the mosquito is to locally transfer malaria between human individuals. In this way, <italic>A. sinensis </italic>can be regarded a good vector although the proportion of infective individuals is very low.</p>", "<p>Re-emergence of malaria in Korea in the 1990's can be explained in a similar way as the epidemic during positional warfare malaria in Finland in the 1940's: hidden dormant stages in the liver of a small fraction of the Korean population. In the case of North Korea the strength of military forces increased from 288,000 to 665,000 within 100 km of the DMZ from 1980 to 1994 [##UREF##11##15##]. The number of soldiers close to the DMZ during the recent malaria epidemic has been at least 350,000 based on statistics on malaria prophylaxis [##REF##17170570##10##]. In retrospect it is very likely that occasional relapses of <italic>vivax </italic>malaria still occurred in the 1980's and beginning of 1990's in North Korea corresponding to the South Korean cases in the 1980's. The increased military forces close to the DMZ created a situation resembling that in Finland in 1941. An increasing part of the urban population was mobilized in conditions with higher exposure to mosquitoes. The occasional carriers of hypnozoites developed relapses leading to a regional epidemic among military forces. This is in short the origin of the re-emergence of malaria in North Korea. To maintain parity in the military balance South Korea probably intensified the guarding of DMZ although public reports are not readily available. It is, however, known that more than 200,000 South Korean soldiers in the risk zone near DMZ got malaria prophylaxis in 2005 [##REF##17170570##10##]. Increased number of soldiers close to the DMZ created a suitable situation for the re-emergence of malaria because viable hypnozoites presumably still occurred in occasional soldiers along DMZ in South Korea.</p>", "<p>The higher magnitude of malaria cases in North Korea compared with South Korea is explained by a combination of several factors. The household size in North Korea is obviously higher than in South Korea as mentioned above. There was a higher proportion of urban hypnozoite carriers in North Korea than in South Korea because of a more rapid urbanization rate in North Korea which caused an apparent rapid decline of manifested malaria in 1970's. Rapid urbanization enabled a higher proportion of the human population to escape exposure to <italic>Anopheles </italic>mosquitoes. During the mobilization of more troops close to DMZ, a higher proportion of the civil population became again exposed to mosquitoes and caused a rapid epidemic (although not immediately reported). North Korea was badly prepared to withstand a malaria epidemic compared with South Korea because of many political and economic reasons.</p>", "<p>The effect of the difference in urbanization rates on malaria decline is illustrated when Finland and Korea are compared (Figure ##FIG##8##9##, data on urbanization from [##UREF##21##29##, ####REF##16192749##30##, ##UREF##22##31####22##31##]). Urbanization level in Finland was still very low during the decline of malaria in the 1920's and 1930's. Malaria declined more slowly in Finland and the positional warfare malaria produced a broader peak than in Korea. In Finland malaria disappeared within ten years after World War II.</p>", "<title>The benefit of induced relapses for <italic>Plasmodium</italic></title>", "<p>The significance of the mosquito induced relapses is obvious. The timing of activation of a dormant stage in malaria is closely correlated with the emergence of the new generation of adult <italic>Anopheles </italic>mosquitoes as demonstrated for <italic>vivax </italic>malaria in the case of Finland [##REF##18430203##19##]. The <italic>Plasmodium </italic>can optimize the production of gametocytes during the peak of mosquito activity. This leads to an efficient production of sporozoites for transmissions to new human hosts. If there are big clusters of people present, as in positional warfare conditions, there will be a rapid spreading of <italic>Plasmodium </italic>among humans.</p>", "<p>The basic function of a mosquito induced relapse probably is that it enables meiotic recombination in mosquitoes. A human can have several primary infections without the activation of all hypnozoites. Each primary infection will increase the number of hypnozoites in the liver. When the hypnozoites are activated during the next mosquito season, there will be a synchronous development of gametocytes of different strains. This optimizes the possibility for a meiotic recombination within mosquitoes. Regular occurrences of meiotic recombination among a large human population will increase the adaptability of the <italic>Plasmodium </italic>to various <italic>Anopheles </italic>species and to variable seasonal and environmental conditions like those in Finland and Korea.</p>" ]

[ "<title>Results and discussion</title>", "<title>Finland 1941–1945</title>", "<p>Finland mobilized 610,000 people in the armed forces in the Continuation War in 1941 and concentrated some 310,000 men along the front line with Soviet Union in Karelia in the southeast and on Hanko peninsula by the Gulf of Finland (ceded to Soviet Union after the Winter War in 1939–1940). The allied German troops (about 200,000 soldiers) were concentrated in the northern part of Finland [##UREF##0##2##,##UREF##1##3##].</p>", "<p>Annual civil malaria cases in Finland and the Karelian Isthmus from 1923 until the end of Continuation War in 1944 is presented in Figure ##FIG##4##5##. It is notable that malaria had almost disappeared from the Karelian Isthmus during the 1930's but suddenly increased in 1937 before the Winter War. In 1940 and 1941 there are no documented civil cases in that area. Already in July 1941 several malaria cases appeared among Finnish soldiers along both southern frontlines. Monthly military cases of malaria in 1941–1944 and civil cases after the war in 1945 are presented in Figure ##FIG##5##6##. Hernberg [##UREF##3##5##] studied the situation from 1941 to 1945 and used his statistics as a proof of existence of a dormant stage in humans. He concluded that Finnish soldiers at the beginning of the 1940's must have got the infection from infective mosquitoes flying across the front line from the Soviet positions because indigenous malaria was very rare in Finland at that time.</p>", "<p>According to recent studies on historical malaria in Finland, primary infections occur only indoors from mid winter to spring and infective <italic>Anopheles </italic>mosquitoes do not exist in summer [##REF##15847704##1##,##REF##18430203##19##]. The <italic>Anopheles </italic>mosquitoes reach the adult stage in the latter half of July and August. In 1941 the mean temperature of June was cold (about 11°C) and July was warm (about 21°C). As a consequence, <italic>Anopheles </italic>mosquitoes had certainly hatched already in July, but they cannot then have been infective with sporozoites. Because of the low temperature in August the sporogony was slow.</p>", "<p>Because the primary infections of malaria during the summer of 1941 are excluded the early malaria cases can only have been relapses from activated hypnozoites. Relapses are principally expected to be triggered by the bites of mosquitoes in natural conditions [##REF##18430203##19##]. Hypnozoites can remain viable in the human liver at least nine years but probably a much longer time [##REF##18430203##19##]. The main issue is to explain the source of hypnozoite carriers along the front line.</p>", "<p>The basic cause of decline of malaria in Finland is now expected to have been gradual decrease of household size since the latter half of 18^thcentury from about ten to four members [##UREF##15##20##]. The household size of four members seems to be a threshold value for prevalence of malaria. The declining household size has a very high correlation with declining malaria frequency (p &lt; 0.001). According to the Finnish statistics malaria disappeared when the household size decreased below four members in the 1950's. Likewise the European malaria declined in step with declining household size until it disappeared when household size passed below four members in all countries [##UREF##16##21##,##UREF##17##22##]. The mechanism of the decline, however, will be discussed in a separate paper.</p>", "<p>The increasing level of urbanization is another factor that must be considered when interpreting the occurrence of relapses. If we accept that <italic>Anopheles </italic>mosquitoes are the principal triggers of relapses, then relapses should occur in situations where mosquitoes and humans meet. This happened in the rural Finland each year when the new generation of <italic>Anopheles </italic>hatched in July. The rural population decreased gradually when people moved into cities to look for better job opportunities. During the last phase of malaria in the 1920's and 1930's many people were still hypnozoite carriers when they moved into cities. People living in urban conditions in multi-storey buildings are rarely exposed to mosquitoes compared with people in rural conditions. A conclusion is that a certain fraction of the Finnish urban population in the late 1930's still had dormant hypnozoites in their liver. A steep decline of malaria frequency occurred in urban settlements about 20 years later than in rural settlements. Although the number of malaria cases was declining, the number of people carrying hypnozoites was still high enough for a new epidemic during suitable conditions. A suitable condition develops when a large random sample of the population congregates and is stationary for a considerable time. An occasional relapse may be triggered immediately but an epidemic requires a longer time for the new cycles of malaria to develop. Typical situations are positional warfare, refugee camps and large scale public work with provisional lodging during times of unemployment and depression (corresponding to \"frontier malaria\" in the 19^thcentury United States).</p>", "<p>Most of the Finnish malaria cases since the late 1920's were probably relapses. The peak in malaria cases in 1931 is explained by depression and public work. The workers lived in temporary quarters in primitive conditions and were exposed to nocturnal mosquitoes. A major part of the malaria cases in 1937–1939 was clustered within the Karelian Isthmus during intensified fortification work before the war in November 1939. The military actions during the Winter War 1939–1940, however, did not contribute to the malaria cases because of the highly mobile front lines during the cold months. The soldiers slept in a tent or in a lean-to with temperatures down to -40°C. In such conditions mosquitoes cannot be active.</p>", "<p>During the Continuation War, the situation changed. Positional warfare conditions prevailed from January 1940 to June 1944. The soldiers slept in primitive dugouts and barracks (6–24 people in the same space) and remained stationary and continuously exposed to mosquitoes which over wintered together with humans. Because the sleeping places were heated the over wintering mosquitoes were forced to take blood during winter. Any malaria carrier would then have got a relapse triggered by the mosquitoes. Exactly in these conditions the malaria epidemic commenced. The first outbreak revealed in the summer in 1941 that hypnozoite carriers among the soldiers were able to initiate an epidemic. It is worth noting that there were very few civil cases on the Karelian Isthmus during the Continuation War although 280,000 civilians returned to their homes [Figure ##FIG##6##7##].</p>", "<p>According to Hernberg [##UREF##4##6##] there were about 3000 cases of malaria among the Finns during 1941–1945. Because he did not know the true dynamics of relapses his judgement about separating primary infections from relapses cannot be relied on. A crude estimate of 1500 primary infections and 1500 relapses is accepted in this study. About 310,000 men were concentrated along the front line and exposed to mosquitoes. It can therefore be estimated that about 0.5% of the Finnish population still had hidden dormant hypnozoites in their liver in the 1940's. In 1945 after the war 1551 cases of malaria were reported all over Finland from February to October peaking in May and June. 299 patients had malaria two or several times so the remaining 1252 cases represent the true number of malaria infected persons. A detailed study on 868 patients revealed only 12 women. Practically all of them represented repatriated soldiers. Probably nearly all of the cases were relapses. They originated from primary infections that must have taken place during the war in the spring 1944 when many soldiers lived together in primitive conditions. The mosquitoes were then able to transmit malaria between infected and uninfected soldiers during the transmission season from December to May inside the dugouts and barracks. When the infected soldiers were repatriated after the war only those soldiers returning to conditions where they were locally exposed to mosquitoes got relapses. Those soldiers that returned to housing conditions without mosquitoes could still have hidden dormant stages of malaria in their liver a long time after the war. The schema in Figure ##FIG##7##8## presents the principal mechanism of the outbreak of malaria in positional warfare conditions.</p>", "<p>The key factor for the re-emergence of malaria is the number of people sleeping in the same space. The average number of soldiers sleeping in the same space is usually much higher than the empirically found threshold value of four humans per room. The success of <italic>Plasmodium </italic>is linked to interactions between groups of humans. If the groups are too small with two or three humans, the groups become more isolated and the <italic>Plasmodium </italic>faces a higher extinction rate than the re-colonization rate. If the group size is more than four humans, the <italic>Plasmodium </italic>can gradually spread among humans. Principally malaria epidemics among military forces could be avoided if each soldier sleeps alone in a separate space like a room, a tent or a bed net. In such situations occasional malaria cases will always be isolated and more easily controlled than if a large number of people get malaria simultaneously.</p>", "<p>Of special interest in the outbreak of malaria among Finnish troops is that it did not happen in fight situations, it was always behind the immediate front line. Because most <italic>Anopheles </italic>species are night biters (like <italic>A. messeae </italic>in Finland) they usually bite when people are sleeping. The behaviour of an infective mosquito is altered by the <italic>Plasmodium</italic>. Sporozoites in the salivary glands decrease the production of anti-coagulants and vasodilators by the mosquito which in effect forces the mosquito to bite repeatedly to get blood [##REF##18430203##19##]. Another effect is that the chance of the bite getting unnoticed by the victim increases as human skin does not react on the decreased load of alien proteins. That means that all people sleeping in the same space could quickly get malaria by only one infective mosquito. All these circumstances points to a basically indoor transmission pattern of malaria in general.</p>", "<p>Malaria has sometimes been associated with famine and hardship. During 1749 – 1849 there were 27 years associated with famine because of extreme weather conditions (floods, frost or drought with thousands of deaths) in Finland [##UREF##18##23##]. The mean annual number of deaths from malaria was 95.24 for the whole period and 103.07 for the 27 famine years. The difference is statistically insignificant. As a consequence the effect of famine or other hardship on malaria prevalence is expected to be irrelevant.</p>", "<title>Korea 1993–2006</title>", "<p>There are several similarities in the evolution of malaria in Korea and Finland. <italic>Plasmodium vivax </italic>with long incubation time is common to both regions. During the Korean War in 1950–54, there was a high rate of malaria among all military forces. Later it declined in step with decreasing household size and increased urbanization. The average household size in South Korea reached four members in the 1980's. In North Korea, the household size was still higher in the 1990's, varying from 4.2 to 4.7 depending on sources [##REF##10440306##24##].</p>", "<p>Malaria was officially eradicated in both North Korea and South Korea in 1979 [##UREF##9##12##]. But at least in South Korea occasional cases, which were considered relapses after a long incubation, were still reported until 1984 [##REF##16172490##25##]. Re-emergence of malaria commenced in 1993 in South Korea [##REF##17170570##10##] and in 1995 in North Korea [##UREF##7##9##]. The number of cases in North Korea has been nearly 100 times higher than in South Korea. The malaria trend is now reversed in both countries by means of large scale national and international (especially in North Korea) counter measures.</p>", "<p>The re-emergence of malaria has received much attention in South Korea. It is generally believed that it was a result of influx of infective <italic>Anopheles sinensis </italic>mosquitoes (the main vector) from North Korea [##UREF##8##11##,##UREF##18##23##]. This explanation is inconsistent with the fact that during the main transmission season in the summer, the prevailing wind direction comes from the south [##UREF##19##26##,##UREF##20##27##]. Unfed females of the vector species, <italic>Anopheles sinensis</italic>, can fly quite long distances, up to 9–12 km [##REF##12325443##28##], but after a blood feed it remains 2–3 days in a safe place until egg-laying [##UREF##21##29##,##REF##16192749##30##]. After the first egg batch, the female returns within 24 hours to a host to the next blood meal [##UREF##22##31##]. Infective mosquitoes have been collected at very low frequency in Korea, 0.01–0.06% of studied samples [##UREF##18##23##]. It has been shown that <italic>A. sinensis </italic>females are highly zoophilic and prefer cattle [##UREF##21##29##,##REF##16192749##30##]. They take, however, readily blood from humans if cattle are not available [##UREF##21##29##]. It is also striking that malaria foci in South Korea in the 1960's were relatively stationary and malaria did not spread in neighbouring regions by means of flying mosquitoes [##UREF##8##11##]. The South Korean troops near DMZ had about a 6–7 time higher malaria incidence rate than the United States troops stationed in the same region in 1993–2005 [##REF##17170570##10##]. That large difference in the incidence rate would not be expected if infective mosquitoes at random would bite soldiers close to DMZ. Considering the sum of all these factors, large scale spreading of malaria by mosquitoes from North Korea seems unlikely.</p>", "<p>The behaviour of an infective <italic>A. sinensis </italic>is probably altered in the same way as <italic>A. messeae </italic>in Finland. The production of anticoagulants and vasodilators are reduced because of sporozoites in the salivary glands [##REF##18430203##19##]. An infective mosquito female is forced to take repeated blood meals and will not fly far away from the blood source. A plausible explanation is that infected humans themselves are spreading malaria and the role of the mosquito is to locally transfer malaria between human individuals. In this way, <italic>A. sinensis </italic>can be regarded a good vector although the proportion of infective individuals is very low.</p>", "<p>Re-emergence of malaria in Korea in the 1990's can be explained in a similar way as the epidemic during positional warfare malaria in Finland in the 1940's: hidden dormant stages in the liver of a small fraction of the Korean population. In the case of North Korea the strength of military forces increased from 288,000 to 665,000 within 100 km of the DMZ from 1980 to 1994 [##UREF##11##15##]. The number of soldiers close to the DMZ during the recent malaria epidemic has been at least 350,000 based on statistics on malaria prophylaxis [##REF##17170570##10##]. In retrospect it is very likely that occasional relapses of <italic>vivax </italic>malaria still occurred in the 1980's and beginning of 1990's in North Korea corresponding to the South Korean cases in the 1980's. The increased military forces close to the DMZ created a situation resembling that in Finland in 1941. An increasing part of the urban population was mobilized in conditions with higher exposure to mosquitoes. The occasional carriers of hypnozoites developed relapses leading to a regional epidemic among military forces. This is in short the origin of the re-emergence of malaria in North Korea. To maintain parity in the military balance South Korea probably intensified the guarding of DMZ although public reports are not readily available. It is, however, known that more than 200,000 South Korean soldiers in the risk zone near DMZ got malaria prophylaxis in 2005 [##REF##17170570##10##]. Increased number of soldiers close to the DMZ created a suitable situation for the re-emergence of malaria because viable hypnozoites presumably still occurred in occasional soldiers along DMZ in South Korea.</p>", "<p>The higher magnitude of malaria cases in North Korea compared with South Korea is explained by a combination of several factors. The household size in North Korea is obviously higher than in South Korea as mentioned above. There was a higher proportion of urban hypnozoite carriers in North Korea than in South Korea because of a more rapid urbanization rate in North Korea which caused an apparent rapid decline of manifested malaria in 1970's. Rapid urbanization enabled a higher proportion of the human population to escape exposure to <italic>Anopheles </italic>mosquitoes. During the mobilization of more troops close to DMZ, a higher proportion of the civil population became again exposed to mosquitoes and caused a rapid epidemic (although not immediately reported). North Korea was badly prepared to withstand a malaria epidemic compared with South Korea because of many political and economic reasons.</p>", "<p>The effect of the difference in urbanization rates on malaria decline is illustrated when Finland and Korea are compared (Figure ##FIG##8##9##, data on urbanization from [##UREF##21##29##, ####REF##16192749##30##, ##UREF##22##31####22##31##]). Urbanization level in Finland was still very low during the decline of malaria in the 1920's and 1930's. Malaria declined more slowly in Finland and the positional warfare malaria produced a broader peak than in Korea. In Finland malaria disappeared within ten years after World War II.</p>", "<title>The benefit of induced relapses for <italic>Plasmodium</italic></title>", "<p>The significance of the mosquito induced relapses is obvious. The timing of activation of a dormant stage in malaria is closely correlated with the emergence of the new generation of adult <italic>Anopheles </italic>mosquitoes as demonstrated for <italic>vivax </italic>malaria in the case of Finland [##REF##18430203##19##]. The <italic>Plasmodium </italic>can optimize the production of gametocytes during the peak of mosquito activity. This leads to an efficient production of sporozoites for transmissions to new human hosts. If there are big clusters of people present, as in positional warfare conditions, there will be a rapid spreading of <italic>Plasmodium </italic>among humans.</p>", "<p>The basic function of a mosquito induced relapse probably is that it enables meiotic recombination in mosquitoes. A human can have several primary infections without the activation of all hypnozoites. Each primary infection will increase the number of hypnozoites in the liver. When the hypnozoites are activated during the next mosquito season, there will be a synchronous development of gametocytes of different strains. This optimizes the possibility for a meiotic recombination within mosquitoes. Regular occurrences of meiotic recombination among a large human population will increase the adaptability of the <italic>Plasmodium </italic>to various <italic>Anopheles </italic>species and to variable seasonal and environmental conditions like those in Finland and Korea.</p>" ]

[ "<title>Conclusion</title>", "<p>The new interpretation of re-emerging <italic>vivax </italic>malaria in positional warfare conditions is based on two independent discoveries on Finnish malaria. Firstly, malaria transmission is mainly an indoor phenomenon. High correlation of a threshold value of four persons in household size and malaria prevalence points to the indoor transmission of malaria. The overall decline of household size below four persons and disappearance of malaria in Europe in the 20^thcentury strongly corroborates a universal indoor transmission independently of vector species and <italic>Plasmodium </italic>species. When any vector species becomes infective the <italic>Plasmodium </italic>alters the vector in the same direction. The vector is forced to repeatedly probe for blood and to remain close to humans. Outdoor transmissions of malaria are not excluded, but they have small impact on the long term statistics.</p>", "<p>Secondly, the mosquito triggered relapse explains the exact timing of the activation of hypnozoites and presence of mosquitoes. Hypnozoites of <italic>P</italic>. <italic>vivax </italic>malaria may be viable for decades in human liver. This fact explains the re-emergence of malaria when it seemed to have been eradicated in recent decades. <italic>Vivax </italic>malaria is often described as unstable malaria. Actually, it is stable malaria but well adapted to fluctuating transmission opportunities.</p>", "<p>Some principal conclusions from the study can be drawn. The primary means of the spreading of <italic>Plasmodium vivax </italic>is by humans and only secondary by the mosquitoes. The <italic>Plasmodium </italic>spends only a short time of its lifetime in the mosquito based on the relative length of sporogony. The time in a human can be several years. Statistically humans have a much bigger chance to spread <italic>Plasmodium </italic>than the mosquitoes. A seemingly healthy human carrier of hypnozoites can easily take the parasite into previously non-malarious area. In suitable conditions in the presence of mosquitoes and other humans a new malaria epidemic could then start.</p>", "<p>The role of the mosquito in malaria is two fold, to be a tool for the sexual reproduction of the <italic>Plasmodium </italic>and to transfer sporozoites into as many members of humans as possible. It is risky for the mosquito female to leave the place where it succeeded both to take a blood meal and to lay eggs. Any other behaviour by the female would not only decrease its reproduction rate but also decrease the reproduction rate by the <italic>Plasmodium</italic>. From an evolutionary point of view the <italic>Plasmodium </italic>would not adapt to a mosquito host which could not be altered to serve the basic needs of the <italic>Plasmodium</italic>.</p>", "<p>Eradication of <italic>vivax </italic>malaria is a much longer process than has been realized. Hypnozoites of <italic>vivax </italic>malaria survive a long time in the humans and might possibly remain viable throughout the life time of humans after the primary infection. From an epidemiological point of view the crucial issue is to isolate humans from each other indoors in such a way that mosquitoes cannot fly between them during the night. A human crisis with people living in crowded conditions for a longer time always creates a situation where malaria epidemics can break out. In this case military personnel were studied. The same situation may emerge in refugee camps.</p>" ]

[ "<p>This is an Open Access article distributed under the terms of the Creative Commons Attribution License (<ext-link ext-link-type=\"uri\" xlink:href=\"http://creativecommons.org/licenses/by/2.0\"/>), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.</p>", "<title>Background</title>", "<p>A sudden outbreak of <italic>vivax </italic>malaria among Finnish troops in SE-Finland and along the front line in Hanko peninsula in the southwest occurred in 1941 during World War II. The common explanation has been an invasion of infective <italic>Anopheles </italic>mosquitoes from the Russian troops crossing the front line between Finland and Soviet Union. A revised explanation is presented based on recent studies of Finnish malaria.</p>", "<title>Methods</title>", "<p>The exact start of the epidemic and the phenology of malaria cases among the Finnish soldiers were reanalyzed. The results were compared with the declining malaria in Finland. A comparison with a corresponding situation starting in the 1990's in Korea was performed.</p>", "<title>Results and discussion</title>", "<p>The malaria cases occurred in July in 1941 when it was by far too early for infective mosquitoes to be present. The first <italic>Anopheles </italic>mosquitoes hatched at about the same time as the first malaria cases were observed among the Finnish soldiers. It takes about 3 – 6 weeks for the completion of the sporogony in Finland. The new explanation is that soldiers in war conditions were suddenly exposed to uninfected mosquitoes and those who still were carriers of hypnozoites developed relapses triggered by these mosquitoes. It is estimated that about 0.5% of the Finnish population still were carriers of hypnozoites in the 1940's. A corresponding outbreak of <italic>vivax </italic>malaria in Korea in the 1990's is similarly interpreted as relapses from activated hypnozoites among Korean soldiers.</p>", "<p>The significance of the mosquito induced relapses is emphasized by two benefits for the <italic>Plasmodium</italic>. There is a synchronous increase of gametocytes when new mosquitoes emerge. It also enables meiotic recombination between different strains of the <italic>Plasmodium</italic>.</p>", "<title>Conclusion</title>", "<p>The malaria peak during the positional warfare in the 1940's was a short outbreak during the last phase of declining indigenous malaria in Finland. The activation of hypnozoites among a large number of soldiers and subsequent medication contributed to diminishing the reservoir of malaria and speeded up the eradication of the Finnish malaria. A corresponding evolution of Korean malaria is anticipated with relaxed tensions and decreasing troop concentrations along the border between South and North Korea.</p>" ]

[ "<title>Competing interests</title>", "<p>The authors declare that they have no competing interests.</p>", "<title>Authors' contributions</title>", "<p>LeH drafted the manuscript and collected the historical data. LaH added and/or removed various sections and participated in the design of the study. Both authors read and approved the final manuscript.</p>" ]

[ "<title>Acknowledgements</title>", "<p>This study was supported by grants from Svenska Litteratursällskapet i Finland and Kommerserådet Otto A. Malms Donationsfond.</p>" ]

[ "<fig position=\"float\" id=\"F1\"><label>Figure 1</label><caption><p>Military situation in Finland before the Winter War 1939–40.</p></caption></fig>", "<fig position=\"float\" id=\"F2\"><label>Figure 2</label><caption><p><bold>Military situation after the Winter War in 1940.</bold> Over 400,000 karelians were evacuated from the occupied area. There were no cases of malaria among military troops.</p></caption></fig>", "<fig position=\"float\" id=\"F3\"><label>Figure 3</label><caption><p><bold>Finnish and German maximum advances during Continuation War in 1941–44 in the surroundings of Finland.</bold> Malaria occurred among Finnish troops along three frontlines: Hanko peninsula, the Karelian Isthmus and East Karelia east of Lake Ladoga. During the war 280,000 Karelians returned to their homes but there were very few cases of malaria among civilians.</p></caption></fig>", "<fig position=\"float\" id=\"F4\"><label>Figure 4</label><caption><p><bold>Post war Finland in 1944.</bold> The Karelians were again evacuated from Karelia. There were very few cases of malaria among civilians.</p></caption></fig>", "<fig position=\"float\" id=\"F5\"><label>Figure 5</label><caption><p><bold>Civil cases of malaria in 1923–44 in Finland.</bold> The Karelian Isthmus and the rest of Finland are compared. Malaria cases had disappeared in the Karelian Isthmus in the beginning of 1930's. The cases in 1937–39 were mosquito triggered relapses. Most of the civil cases in 1942–44 are expected to be soldiers on leave [##UREF##4##6##].</p></caption></fig>", "<fig position=\"float\" id=\"F6\"><label>Figure 6</label><caption><p><bold>Monthly military cases of malaria form 1941 to 1944 and civil cases in 1945.</bold> The red horizontal line indicate the threshold temperature for sporogony of <italic>Plasmodium vivax </italic>in South Finland. Summer peaks of malaria cannot be explained by primary infections because <italic>Anopheles messeae </italic>hatch on average only in the latter half of July [##REF##15847704##1##,##REF##18430203##19##].</p></caption></fig>", "<fig position=\"float\" id=\"F7\"><label>Figure 7</label><caption><p><bold>Proportion of military and civil cases of malaria in the Karelian Isthmus. </bold>The absolute numbers are given inside the bars. Most of the civil cases are expected to represent soldiers on leave [##UREF##4##6##].</p></caption></fig>", "<fig position=\"float\" id=\"F8\"><label>Figure 8</label><caption><p><bold>Scheme of re-emerging malaria during the Continuation War.</bold> In crowded conditions with mosquitoes the malaria epidemic starts with mosquito triggered relapses in hypnozoite carriers. In the next step mosquitoes become infective and gradually spread sporozoites into healthy humans. After the war hypnozoite carriers got relapses if <italic>Anopheles </italic>mosquitoes were present.</p></caption></fig>", "<fig position=\"float\" id=\"F9\"><label>Figure 9</label><caption><p><bold>Trench warfare malaria in Korea and Finland. </bold>Time scale for Finland has been moved 55 years forwards. Note: The population of South Korea is about 10 times that of Finland and the lowest absolute values for Finland were about the same as those for Korea. On the logarithmic scale 1 for Finland and 0.1 for Korea both represent 4 cases of malaria. The malaria situation in Finland was only marginally different from that in Korea when a new epidemic commenced. The pronounced decline of malaria in Korea should be compared with the urbanization rate.</p></caption></fig>" ]

[]

[]

[]

[]

[]

[]

[]

[]

[ "<graphic xlink:href=\"1475-2875-7-171-1\"/>", "<graphic xlink:href=\"1475-2875-7-171-2\"/>", "<graphic xlink:href=\"1475-2875-7-171-3\"/>", "<graphic xlink:href=\"1475-2875-7-171-4\"/>", "<graphic xlink:href=\"1475-2875-7-171-5\"/>", "<graphic xlink:href=\"1475-2875-7-171-6\"/>", "<graphic xlink:href=\"1475-2875-7-171-7\"/>", "<graphic xlink:href=\"1475-2875-7-171-8\"/>", "<graphic xlink:href=\"1475-2875-7-171-9\"/>" ]

[]

{ "acronym": [], "definition": [] }

[ "<title>Background</title>", "<p>Pregnancy-associated malaria (PAM) has serious adverse outcomes such as low birth weight neonates, increased perinatal and maternal mortality, anaemia and increased risk of hypertension in first-time pregnant mothers [##REF##15081631##1##,##REF##17105340##2##]. PAM is coupled with massive accumulation of parasitized erythrocytes (PEs) and monocytes in the placental intervillous blood spaces [##REF##8495954##3##,##REF##6758604##4##]. The basis for this accumulation in the placenta results from the capacity of placental PEs to bind to chondroitin sulfate A (CSA) but not to CD36, a common receptor for PEs sequestration in the microvasculature [##REF##8633247##5##]. In endemic areas, women acquire antibodies against placental parasites over successive pregnancies, as they become resistant to PAM [##REF##9804416##6##]. Women who have acquired antibodies against placental PEs have higher haemoglobin levels, deliver heavier babies and are much less susceptible to PAM than primigravid and HIV-infected women lacking these antibodies [##REF##15183624##7##, ####REF##14573685##8##, ##REF##14751701##9####14751701##9##]. Furthermore, naturally acquired antibodies from multigravid women react against placental PEs or CSA-binding parasites collected around the world, indicating that target epitopes are globally conserved [##REF##9804416##6##,##REF##10395863##10##, ####REF##10496918##11##, ##REF##10975848##12####10975848##12##].</p>", "<p>Recent evidences suggest that var2CSA, a member of the <italic>Plasmodium falciparum </italic>Erythrocyte Membrane Protein 1 (PfEMP1) family, may have an important role in PAM disease and immunity [##REF##17224156##13##]. PfEMP1 proteins are clonally variant parasite adhesion ligands expressed on the surface of infected erythrocytes [##REF##7606788##14##,##REF##7541722##15##]. Var2CSA is a very large protein with an estimated molecular weight of 350 kDa, and can be divided into six Duffy binding-like domains (DBL1-6). Among them DBL2-X, DBL3-X and DBL6-ε specifically bind to CSA [##REF##15717280##16##]. <italic>Var2csa </italic>gene orthologs are present in all parasite isolates [##REF##17669514##17##] and are transcriptionally upregulated in both placental isolates and laboratory parasites selected to bind CSA [##REF##12823820##18##, ####REF##15962229##19##, ##REF##16861676##20####16861676##20##]. Importantly, <italic>var2csa </italic>knock-out parasites revealed that no other parasite ligand can promote massive adhesion in the placenta [##REF##16025132##21##, ####REF##17878945##22##, ##REF##16631964##23####16631964##23##]. Furthermore, the var2CSA protein is the target of naturally acquired maternal antibodies and the presence of var2CSA specific IgG has been correlated with higher birth weight babies [##REF##16790811##24##, ####REF##15520249##25##, ##REF##16453268##26####16453268##26##].</p>", "<p>All these data point to var2CSA as the key target for the development of a PAM vaccine, but a number of obstacles need to be overcome, such as the identification of regions in the large polymorphic molecule (350 kDa) able to induce broadly transcendent neutralizing antibodies that would de-sequester and/or mediate parasite phagocytosis. Given the var2CSA protein size, strategies to express the entire protein are not envisaged, but expression of correctly folded and biologically active cysteine-rich DBL domains is the most promising strategy.</p>", "<p>In this study, an expression system was developed to produce recombinant DBL domains in the culture supernatant of transiently transfected Human Embryonic Kidney 293 cell line (HEK293) grown in serum-free medium in suspension. Using the HEK293 cell expression system, five DBL domains out of six were successfully secreted in the growth medium. As DBL2-X and DBL3-X were not found or found in low amount in the HEK293 growth medium, their expression in <italic>Escherichia coli </italic>was evaluated. Previously described CSA binding domains DBL3-X and DBL6-ε were then chosen to raise antisera in mice in order to evaluate their capacity to induce antibodies recognizing the PEs surface of different CSA-binding strains and for their capacity to inhibit FCR3-CSA PEs binding to Sc1D cells in static adhesion experiments.</p>" ]

[ "<title>Methods</title>", "<title>HEK293 DBL expression and purification</title>", "<p><italic>Var2csa </italic>synthetic genes were designed with optimized codons for human expression. Synthetic genes encoding FCR3 var2CSA DBL domains (accession AY372123); DBL1-X (residues 58–383), DBL2-X (residues 530–863), DBL3-X (residues 1221–1548), DBL4-ε (residues 1594–1888), DBL5-ε (residues 2003–2270) and DBL6-ε (residues 2322–2590) were cloned into the pTT3 vector between the EcoRI and BamHI restriction sites. Synthetic genes contained an N-terminal murine Ig κ-chain leader sequence to allow secretion of the proteins [##REF##1640112##27##] and a His6-tag on the C-terminus. Potential N-glycosylation sites were removed from synthetic genes by converting asparagine to glutamine or by replacing asparagine with an amino acid from another <italic>var2csa </italic>allele (Table ##TAB##0##1##).</p>", "<p>FreeStyle 293-F cells (Invitrogen) were grown in Freestyle 293 serum free expression medium and transfected with the pTT3 vector containing the synthetic genes following Invitrogen's recommendations. 72 hours post-transfection, cells were centrifuged and the culture medium was harvested. After filtration on a 0.22 μm filter, supernatants were concentrated five times using a 10 kDa cut-off Vivaflow 200 System (Vivasciences). Samples were then diafiltrated against PBS pH 7.2 (GIBCO) supplemented to 500 mM NaCl final concentration and charged on a HisTrap FF Ni-affinity column (GE Healthcare) previously equilibrated with the same buffer, and connected to an FPLC Akta System (Amersham Pharmacia Biotech). After washing, proteins were eluted with an imidazole gradient (0 to 0.5 M). Aliquots containing purified DBL domains were pooled and dialyzed against 0.9% NaCl. Purified proteins were subsequently concentrated using Macro- and Micro-sep concentrators (Pall/Gellman). Protein concentrations were determined using the Bio-Rad protein assay. Purity of the samples was checked by SDS-PAGE and Western blot.</p>", "<title>Prokaryotic expression, refolding and purification</title>", "<p>Synthetic genes encoding for var2CSA DBL2-X and DBL3-X were designed with optimized codons for <italic>E. coli </italic>expression. Synthetic genes encoding FCR3 var2CSA (accession AY372123) DBL2-X (residues 530–863) and DBL3-X (residues 1221–1548) domains fused to a His6-tag on the C-terminus were cloned into the pET24a expression vector between the NdeI and EcoRI restriction sites. Transformed <italic>E. coli </italic>BL21 (DE3) cells were grown at 37°C in LB medium with 30 μg/ml of kanamycin to an absorbance of 0.5 at 600 nm, and were then induced with 1.0 mM isopropyl-β-D-thiogalactopyranoside (IPTG) for 3 h at 37°C under good aeration. Cells were harvested by centrifugation at 6,000 g. The pellets from 800 ml of culture were resuspended in 30 ml of 50 mM Tris.HCl, pH 8.5, containing 150 mM NaCl, in the presence of protease inhibitors. The cells were disrupted by sonication on ice and the suspensions were centrifuged for 20 min at 5,000 g. Pellets were resuspended and washed twice with 30 ml of 50 mM Tris.HCl, pH 8.5, containing 150 mM NaCl, in the presence of protease inhibitors, and finally centrifuged for 20 min at 10,000 g. The pellets containing DBL2-X or DBL3-X inclusion bodies were then denatured overnight at 25°C under agitation in 50 mM Tris.HCl pH 8.5 buffer, containing 150 mM NaCl, 8 M Urea and 5 mM Dithiothreitol (DTT). The suspensions were centrifuged for 30 min at 15,000 g and the pellet was discarded. Refolding was assayed with the AthenaES™ kit (Athena Environmental Sciences, Inc.) according to the manufacturer's instructions. Once the best condition was established, denatured proteins were immobilized on a HisTrap FF Ni-affinity column previously equilibrated with the same buffer, and connected to an FPLC Akta System. The proteins were refolded on the column by adding quickly the refolding buffer (50 mM Tris.HCl pH 8.5, 20 mM NaCl, 0.4 mM KCl, 0.5% Triton-X-100 and 0.5 mM DTT). After extensive washing with the refolding buffer without Triton-X-100, the protein was eluted with an imidazole gradient (0 to 0.5 M) and aliquots containing purified DBL domains were pooled. After dialysis against 0.9% NaCl and concentration by means of Macro- and Micro-sep concentrators, a further stage of gel filtration (Superdex 75, Amersham Pharmacia Biotech) was required to separate the monomeric proteins from the aggregated material and other impurities. Purified DBLs were subsequently concentrated by means of Macro- and Micro-sep concentrators. Protein concentration was determined using the Bio-Rad protein assay. Purity of the samples was checked by SDS-PAGE and Western blot.</p>", "<p>Free thiol content was estimated using the Ellman's Reagent (Pierce) by comparison to a cysteine standard curve composed of known concentrations of Cysteine Hydrochloride Monohydrate following Pierce's recommendations.</p>", "<title>Parasite and cell culture</title>", "<p>The <italic>P. falciparum </italic>FCR3, HB3 and 7G8 strains were maintained in culture according to standard conditions in O+ human erythrocytes in RPMI 1640 containing L-glutamine (Invitrogen) supplemented with 5% Albumax I, 1× hypoxanthine and 20 μg/ml gentamicin [##REF##781840##28##]. To maintain knob-positive parasites, cultures were routinely selected by gelatin flotation using Plasmion (Fresenius Kabi) [##REF##15713452##29##]. Parasites were tested <italic>Mycoplasma </italic>negative by PCR. Laboratory isolates, FCR3, 7G8 and HB3 were initially selected on bovine CSA or on the human choriocarcinoma placenta BeWo cell line and subsequently maintained by panning to the BeWo cell line as previously described [##REF##16274691##30##]. CD36 or CSA binding phenotypes of PEs were verified on receptors immobilized on plastic Petri dishes as previously described [##REF##16274691##30##].</p>", "<title>Balb/c immunization</title>", "<p>Two groups of three Balb/c mice (Charles River) received a primary subcutaneous injection of 20 μg refolded recombinant DBL3-X or secreted DBL6-ε protein dissolved in 100 μl of NaCl 0.9% and emulsified at 1:1 in 100 μl of complete Freund adjuvant (Pierce). Two additional injections were performed at 4 week intervals by using the same amount and antigen batches emulsified at 1:1 in incomplete Freund adjuvant. Mice were bled by orbital sinus puncture one day before the primary antigen injection and each time preceding an antigen boost, and 2, 4 and 6 weeks after the 3rd and last boost. Serum samples were collected after centrifugation of the blood, de-complemented for 30 min at 56°C and stored at -20°C until use.</p>", "<title>Flow cytometry</title>", "<p>Synchronous PEs cultures (3–12% parasitaemia) at mid/late trophozoite stages were purified using the VarioMACS and CS columns (Miltenyi) as previously described [##REF##10213198##31##] and resuspended to 10⁷PEs/ml in PBS 0.2% BSA. PEs were added to 96-well U-bottom plates and the mouse serum (heat inactivated and pre-adsorbed on uninfected erythocytes from mock culture) was added into the wells to get a 1:20 dilution of serum to cell suspension (using a final volume of 50 to 100 μl). After incubation at room temperature for 30 min, 100 μl of PBS containing 0.2% BSA were added, and cells were centrifuged at 2000 rpm for 2 min. Cells were washed twice with 160 μl of PBS/0.2% BSA and resuspended in 50 μl PBS/0.2% BSA containing the secondary antibody (anti-mouse IgG Alexa 488) diluted 1:100. After incubation on ice in the dark for 30 min, cells were centrifuged and washed as before, and resuspended in 2–4% paraformaldehyde in PBS for fixation overnight at 4°C. Fixed cells were centrifuged at 2000 rpm for 2 min and resuspended in 100 μl PBS before transfer to flow cytometry tubes. Analysis was carried out on a FACScan using CellQuest software (BD Biosciences). The flow cytometer gated the PEs' population based on forward scatter and side scatter, and cells were plotted as a histogram by fluorescence in channel 1 (Alexa fluor^®488). At least 10000 cells were counted for each sample. The level of fluorescence was stated as the Geometric Mean Fluorescence Index (MFI) of all the gated cells as determined in at least two independent experiments. Data were then converted to normalized MFI by subtracting the preimmune MFI value from the immune MFI value.</p>", "<title>Inhibition experiments</title>", "<p>Cytoadhesion microassays were performed on confluent <italic>Saimiri </italic>brain endothelial cell (SBEC) Sc1D cells as previously described [##REF##9307979##32##]. Briefly, endothelial cells were rinsed twice with cytoadhesion medium at pH 7.2 before the addition to the cells of 20 μl of serum (diluted 1:5) preincubated with 20 μl of PEs (at 10⁷cells/ml). The slides were incubated for two hours at 37°C in a CO₂incubator. Non-adherent PEs were removed by washing with cytoadhesion medium and cytoadherent PEs were counted under a light microscope (Nikon TMS, magnification of ×300), in four randomly selected fields (area of 0.2827 mm²) for each spot.</p>" ]

[ "<title>Results</title>", "<title>Expression and Purification of recombinant DBL domains</title>", "<p>Var2CSA is a key target for the development of a PAM vaccine. Due to the large protein size, the most promising strategy is a vaccine based on var2CSA protein subunits. Therefore, the six var2CSA DBL domains from the FCR3 strain were tested for expression as secreted His-tagged recombinant proteins in the human embryonic cell line HEK293. For that purpose synthetic genes were designed with optimized codons for human expression and with an N-terminal murine Ig κ-chain leader sequence to allow secretion of the proteins and a His6-tag on the C-terminus. As <italic>P. falciparum </italic>proteins are not N-glycosylated, potential N-glycosylation sites were removed from synthetic genes by converting asparagine to glutamine or by replacing asparagine with an amino acid from another <italic>var2csa </italic>allele (Table ##TAB##0##1##).</p>", "<p>Using this expression system, DBL1-X, DBL4-ε and DBL6-ε were produced at relatively high levels in the culture supernatant, while DBL3-X and DBL5-ε were produced at much lower levels (Figure ##FIG##0##1a##). No expression of DBL2-X could be detected in the culture supernatants. However, after cell lysis, all the recombinant proteins were found in the insoluble fraction, indicating that some of the proteins are likely to be retained in the cells due to their incorrect conformation.</p>", "<p>After purification on a HisTrap FF Ni-affinity column, highly pure recombinant (over 95%), DBL1-X, DBL4-ε and DBL6-ε were obtained at a yield ranging from 0.6 to 5 mg/l of cell culture supernatant, while only a minimal amount of DBL3-X was obtained (Figure ##FIG##0##1b##). Under reducing conditions in SDS-PAGE, all four proteins migrated at the expected molecular weight (Figure ##FIG##0##1b##). DBL4-ε and DBL6-ε migrated as a double-band as a consequence of the presence of two isoforms of the protein, which could result from post-transductional modifications such as O-glycosylation. N-terminal sequencing and SDS-PAGE indicated that proteins remained intact, without proteolytic degradation occurring during the purification procedures.</p>", "<p>As the chosen strategy was to express CSA-binding domains rather than non CSA binding domains, DBL2-X and DBL3-X domains were tested for expression in <italic>E. coli</italic>. For that purpose, synthetic genes encoding for var2CSA DBL2-X and DBL3-X fused to a His6-tag on the C-terminus were designed with optimized codons for <italic>E. coli </italic>expression.</p>", "<p>Recombinant DBL2-X and DBL3-X expressed in <italic>E. coli </italic>accumulate in inclusion bodies as insoluble, misfolded aggregates. Misfolded DBL domains were solubilized in 8 M urea, purified under denaturing conditions by metal-affinity chromatography and refolding tests were carried out by the method of rapid dilution. Using that strategy, a buffer containing 0.5% Triton-X-100 was identified as the best solution for refolding both insoluble recombinant DBL domains. However, due to the incapacity to completely remove Triton X-100 through classical methods, such as dialysis or filtration, the refolding process was done after immobilization of the denatured proteins on a metal-affinity column. In this procedure, DBL domains solubilized in 8 M urea were immobilized onto the metal-affinity column and the denaturation solution changed rapidly for the previously identified refolding solution containing Triton-X-100 (Figure ##FIG##1##2a##). Although due to sudden buffer change some DBL detached from the column, an important fraction of the protein was retained on the column. After extensive washing with the refolding buffer without the detergent, the proteins were eluted using an imidazole gradient (Figure ##FIG##1##2a##). After dialysis against 0.9% NaCl, gel filtration chromatography using Superdex 75 was performed to purify highly pure (over 95%) recombinant DBL monomers to homogeneity (Figure ##FIG##1##2a##). Using that strategy, the protein yield after refolding and all purification steps was 4 mg/l of cell culture for DBL3-X (Figure ##FIG##1##2b##) and 0.3 mg/l of cell culture for DBL2-X (Figure ##FIG##1##2c##). Due to the low yield recovered, DBL2-X was not further characterized. However, refolded and purified DBL3-X was characterized using a variety of biochemical and biophysical methods. N-terminal sequencing of recombinant DBL3-X yields the expected sequence, namely, MNATN. No other sequence was detected. Refolded DBL3-X migrated slower on SDS-PAGE gels after reduction with DTT indicating the presence of disulfide bonds (Figure ##FIG##1##2b##). Free thiol content was estimated by the method of Ellman to further examine the oxidation state of refolded DBL3-X. Free thiols can be clearly detected up to 30 μM thiol concentrations in this assay. No free thiols are detected in refolded DBL3-X at a protein concentration of 100 μM. Given that DBL3-X contains 12 cysteines, greater than 96% of cysteines are thus disulfide linked.</p>", "<title>Antibodies to var2CSA DBL6-ε partially cross-react with different CSA binding strains</title>", "<p>To investigate whether the recombinant var2CSA DBL domains can induce biologically active antibodies recognizing the native var2CSA, mice were immunized with the refolded DBL3 or the HEK293 secreted DBL6-ε. Final bleed antisera were evaluated for their capacity to recognize the surface of FCR3^CSAbinding PEs by liquid immunofluorescence and flow cytometry (Figure ##FIG##2##3a##). Two out of the three mice immunized with DBL6-ε (DBL6-1 and DBL6-3) had high level of antibodies recognizing the surface of FCR3^CSAPEs. Very low reactivity was detected for mouse DBL6-2. However, in contrast to DBL6-ε, the DBL3-X antisera did not react or reacted very weakly with the surface of FCR3^CSAPEs (Figure ##FIG##2##3a##).</p>", "<p>To investigate whether mice antisera could also recognize other var2CSA variants, flow cytometry analysis was performed on two other CSA-binding parasite lines from South America (7G8^CSA) and Central America (HB3^CSA) (Figure ##FIG##2##3b##). Two out of the three mice immunized with DBL6-ε (DBL6-2 and DBL6-3) had a significant level of antibodies recognizing the surface of 7G8^CSAPEs and a low amount of antibodies reacting against HB3^CSA(Figure ##FIG##2##3a##). Curiously all three DBL3-X immunized antisera reacted against 7G8^CSAPEs and recognized weakly the HB3^CSAstrain, even if they did not recognize the surface of the homologous strain FCR3. No significant recognition of all the antisera was observed against the CD36 binding parasites.</p>", "<title>Var2CSA DBL6-ε antibodies partially inhibit CSA adhesion</title>", "<p>In order to evaluate the inhibitory capacities of the DBL3-X and DBL6-ε antisera, the binding of FCR3^CSAPEs to the CSA expressing endothelial cell line Sc1D was assessed in the presence of non-immune or immune antisera at a final 1:10 dilution. One month after the third injection (Day 85), the three DBL6-ε immunized mice recognized the surface of FCR3^CSAbinding PEs by liquid immunofluorescence and inhibited PEs cytoadhesion to Sc1D cells from 32 to 64% (Figure ##FIG##3##4a##). However, no inhibition was observed using the DBL3-X antisera. After an additional boost, DBL6-ε immunized mice were sacrificed and the final bleeds were tested for inhibition. Although the final DBL6-ε antisera bleeds recognize the surface of CSA-binding PEs (Figure ##FIG##2##3a##), almost no inhibitory activity could be detected (Figure ##FIG##3##4b##), indicating a change in the immune response. From these data we can conclude that the DBL6-ε domain can induce an immune response that block PEs adhesion, indicating that it is a critical target domain.</p>" ]

[ "<title>Discussion</title>", "<p>Several lines of evidence point to var2CSA as the leading vaccine candidate to prevent PAM. However, the major limiting steps for validating this molecule as a protective immunogen reside in its high molecular weight and the production of sufficient functional recombinant protein. Indeed, the var2CSA protein size (around 350 kDa) and the hydrophobic and cysteine-rich sequence contribute to aggregation as insoluble material upon expression in the available prokaryotic systems. Therefore we assessed the human embryonic kidney cell line HEK293 as a new system for expressing var2CSA DBL domains from the FCR3 strain as secreted His-tagged recombinant proteins. Engineering the expression vector with a sequence signal allowed us to obtain high quality soluble affinity purified material in a quick and easy way. Although the HEK293 expression system was useful for secreting in relatively high quantity (up to 5 mg/l) three out of six different domains in the culture supernatant, the three others were either absent or present in very low quantities in the culture supernatant. We also showed here that antisera from mice immunized with DBL6-ε recognized the surface of the homologous parasite FCR3^CSAand a variable degree of cross-reactivity was observed for 7G8^CSAand HB3^CSAparasites in individual mice. As 7G8 var2CSA is closer to FCR3 var2CSA than both var2CSA variants present in HB3 (HB3.1 and HB3.2) (Figure ##FIG##2##3b##), this could explain the higher cross-reactivity observed against the 7G8^CSAPEs surface. Considering that DBL6-ε domain is the least conserved var2CSA DBL domain, this limited cross-reactivity is somehow expected. It also suggests that the immunodominant epitopes are in the polymorphic region, whereas the conserved block may not be exposed in the native var2CSA molecule (Figure ##FIG##2##3b##).</p>", "<p>As two of the previously described CSA binding domains present in var2CSA (DBL2-X and DBL3-X) were not expressed or expressed in low amount in the growth supernatant, the <italic>E. coli </italic>expression system was evaluated. The proteins were recovered from inclusion bodies under denaturing conditions and a column refolding based process allowed us to obtain DBL3-X refolded material. Surprisingly, antisera from mice immunized with DBL3-X did not react with the surface of the homologous CSA binding PEs, but recognized weakly the surface of 7G8^CSAPEs. Although DBL3-X is one of the most conserved var2CSA DBL domains, it is hard to understand why the antisera react in a better way with a heterologous antigen than with the homologous one. One explanation could be due to a better accessibility to the antigen on the surface of 7G8^CSAPEs compared to FCR3^CSAor HB3^CSAparasites, or that the DBL3-X domain was not correctly refolded as it is in the context of the whole PfEMP1.</p>", "<p>The DBL3-X and DBL6-ε immunized mice immunological responses were also followed by testing the antisera for their ability in blocking PEs adhesion to Sc1D cells. We showed here that all the antisera raised against DBL6-ε inhibited PEs adhesion with values ranging from 32% to 64% inhibition at D85 after immunization. However, after an additional boost and even with a good surface reactivity, the inhibitory activity was strongly reduced in the final bleed, indicating that other epitopes are targeted by these antisera. It is therefore conceivable that most anti-DBL6-ε antibodies are against the surface of the CSA binding PEs but do not inhibit adhesion. Another explanation could be that some of these non-inhibitory antibodies could compete with the inhibitory antibodies preventing the latter from reaching the CSA binding pocket. In terms of vaccine development, as it seems that the balance between inhibitory and non-inhibitory antibodies may change during the course of immunization, it may be important to specifically target the critical regions involved in adhesion in order to overcome this potential problem.</p>" ]

[ "<title>Conclusion</title>", "<p>This is the first report showing adhesion inhibitory antibodies obtained through a var2CSA recombinant DBL domain immunization protocol. These results support the current strategies using var2CSA as immunogen in the aim of blocking placental sequestration of malaria parasites and set up the basis for developing a vaccine against pregnancy-associated malaria based on this protein.</p>" ]

[ "<p>This is an Open Access article distributed under the terms of the Creative Commons Attribution License (<ext-link ext-link-type=\"uri\" xlink:href=\"http://creativecommons.org/licenses/by/2.0\"/>), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.</p>", "<title>Background</title>", "<p>Pregnancy-associated malaria (PAM) is a serious consequence of <italic>Plasmodium falciparum</italic>-infected erythrocytes sequestration in the placenta through the adhesion to the placental receptor chondroitin sulfate A (CSA). Although women become resistant to PAM as they acquire transcending inhibitory immunity against CSA-binding parasites, hundreds of thousands of lives could be saved if a prophylactic vaccine targeting the surface proteins of placental parasites could be designed. Recent works point to the variant protein var2CSA as the key target for the development of a pregnancy-associated malaria vaccine. However, designing such a prophylactic vaccine has been hindered by the difficulty in identifying regions of var2CSA that could elicit broadly neutralizing and adhesion-blocking antibodies.</p>", "<title>Methods</title>", "<p>Var2CSA is a very large protein with an estimated molecular weight of 350 kDa, and can be divided into six cysteine rich Duffy binding-like domains (DBL). The human embryonic kidney 293 cell line (HEK293) was used to produce secreted soluble recombinant forms of var2CSA DBL domains. The <italic>Escherichia coli </italic>expression system was also assessed for the domains not expressed or expressed in low amount in the HEK293 system. To investigate whether var2CSA binding DBL domains can induce biologically active antibodies recognizing the native var2CSA and blocking the interaction, mice were immunized with the refolded DBL3-X or the HEK293 secreted DBL6-ε domains.</p>", "<title>Results</title>", "<p>Using the HEK293 expression system, DBL1-X, DBL4-ε and DBL6-ε were produced at relatively high levels in the culture supernatant, while DBL3-X and DBL5-ε were produced at much lower levels. DBL2-X and DBL3-X domains were obtained after refolding of the inclusion bodies produced in <italic>E. coli</italic>. Importantly, mice antisera raised against the recombinant DBL6-ε domain, specifically reacted against the surface of CSA-binding parasites and revealed adhesion blocking activity.</p>", "<title>Conclusion</title>", "<p>This is the first report showing inhibitory binding antibodies obtained through a var2CSA recombinant DBL domain immunization protocol. These results support the current strategies using var2CSA as immunogen in the aim of blocking placental sequestration of malaria parasites. This work is a step towards the development of a var2CSA based vaccine that will prevent pregnancy-associated malaria and improve pregnancy outcomes.</p>" ]

[ "<title>List of abbreviations</title>", "<p>PEs: Parasitized Erythrocytes; CSA: Chondroitin Sulfate A; PAM: Pregnancy Associated Malaria; PfEMP1: <italic>P. falciparum </italic>Erythrocyte Membrane Protein 1; DBL: Duffy binding-like; HEK293: Human Embryonic Kidney 293; SBEC: Saimiri brain endothelial cell.</p>", "<title>Competing interests</title>", "<p>The authors declare that they have no competing interests.</p>", "<title>Authors' contributions</title>", "<p>PF participated in the design of the study, carried out the expression and purification of the recombinant DBL domains and wrote the manuscript. NV participated in the design of the study, performed the FACS experiments and wrote the manuscript. SD, CL and JG evaluated the inhibitory capacities and the surface reactivity by liquid IFA of the antisera. AS participated in the design of the study and wrote the manuscript. BG expressed DBL domains, conceived the study, participated in its design and coordination and wrote the manuscript. All authors read and approved the final manuscript.</p>" ]

[ "<title>Acknowledgements</title>", "<p>We would like to thank Dr Y. Durocher for providing the pTT3 vector. This work had financial support from the European Malaria Vaccine Initiative (grant n° 01/2005), the BIOMALPAR network of excellence (LSHP-CT-2004-503578) and from the \"Fonds dédié: Combattre les Maladies parasitaires\" Sanofi Aventis – Ministère de la Recherche.</p>" ]

[ "<fig position=\"float\" id=\"F1\"><label>Figure 1</label><caption><p><bold>Protein purification of FCR3 var2CSA DBL domains expressed in HEK293</bold>. a. SDS PAGE and Western blot analysis. NuPAGE Novex 4–12% Bis-Tris gels under reduced conditions were loaded with HEK293 extracts obtained after transfection with the different FCR3 var2CSA DBL domains. Gels were either stained with Coomasie blue or transferred to a PVDF membrane to detect recombinant products using an anti-His Tag antibodies. Lanes 1, 2, 3, 4, 5, and 6 correspond respectively to DBL1-X, DBL2-X, DBL3-X, DBL4-ε, DBL5-ε, and DBL6-ε. Culture supernatant, soluble and insoluble fractions after cell lysis are indicated. b. Electrophoresis of purified HEK293 expressed DBL domains. NuPAGE Novex 4–12% Bis-Tris gels under reduced (+DTT) or non reduced (-DTT) conditions were loaded with DBL1-X, DBL3-X, DBL4-ε and DBL6-ε obtained after His-Tag purification and stained with Coomasie blue. Lanes 1, 3, 4, and 6 correspond respectively to DBL1-X, DBL3-X, DBL4-ε and DBL6-ε domains. Protein yields were 0.6 mg/L for DBL1-X, 0.9 mg/l for DBL4-ε and 5 mg/l for DBL6-ε.</p></caption></fig>", "<fig position=\"float\" id=\"F2\"><label>Figure 2</label><caption><p><bold>DBL2-X and DBL3-X expression in <italic>E. coli </italic>and purification</bold>. a. Electrophoresis in an SDS 12% polyacrylamide gel of aliquots from different steps in expression, refolding and purification of DBL3-X. Lane 1, non-induced BL21 (DE3) (containing pET24a-DBL3-X plasmid); lane 2, induced BL21 (DE3) containing the pET24a-DBL3-X plasmid; lane 3, inclusion bodies suspension; lane 4, inclusion bodies after washing and denaturation in Urea 8 M; lane 5, affinity column flow through; lane 6, portion of DBL3-X lost during refolding onto the affinity column; lane 7, affinity column elution pool; lane 8, molecular size exclusion column purified DBL3-X. b. Electrophoresis of refolded DBL3-X in a SDS 12% polyacrylamide gel before (-) and after (+) reduction with dithiothreitol (DTT). c. Electrophoresis of refolded DBL2-X in a SDS 12% polyacrylamide gel before (-) and after (+) reduction with dithiothreitol (DTT). Higher mobility before reduction indicates the presence of disulfide linkages.</p></caption></fig>", "<fig position=\"float\" id=\"F3\"><label>Figure 3</label><caption><p><bold>PEs surface recognition</bold>. a. Anti-var2CSA DBL mouse sera were tested for surface recognition of erythrocytes infected with parasites of the FCR3, 7G8 and HB3 strains with CSA- and CD36 binding phenotypes, respectively. Analysis was performed using flow cytometry and fluorescence microscopy. A representative example of a specific surface immunolabeling using the DBL6-1 antisera and the corresponding differential interferential contrast microscopy field are shown. Flow cytometry data shown are the normalized geometric mean fluorescence index (MFI) (± SD) obtained by subtracting the preimmune MFI value from the immune MFI value. b. Alignment of the DBL6-ε var2CSA sequences from the 3 parasite lines used to assess the antisera cross-reactivity by FACS. Conserved residues between FCR3 sequence and the other sequences are boxed with a yellow background.</p></caption></fig>", "<fig position=\"float\" id=\"F4\"><label>Figure 4</label><caption><p><bold>Inhibition of PEs adhesion</bold>. Mouse antisera (1:10 dilution) were tested for inhibition of binding of PEs to Sc1D cells in static binding assays at D85 after immunization (a) and for final bleeds (b). The number of bound PEs in the presence of preimmune serum (PI) was used as a reference (100%). The numbers of bound PEs for the three anti-DBL6-ε sera were then expressed as a percentage relative to the PI reference value. Values represent the means of two individual experiments performed in duplicates.</p></caption></fig>" ]

[ "<table-wrap position=\"float\" id=\"T1\"><label>Table 1</label><caption><p>N-Glycosylation sites mutated in HEK293 expressed DBL domains</p></caption><table frame=\"hsides\" rules=\"groups\"><thead><tr><td align=\"center\">Domains</td><td align=\"center\">Boundaries</td><td align=\"center\" colspan=\"2\">N-Glycosylation sites mutated*</td></tr></thead><tbody><tr><td align=\"center\">DBL1-x</td><td align=\"center\">58 – 383</td><td align=\"center\">5</td><td align=\"center\"><bold>N</bold>85<bold>I</bold></td></tr><tr><td/><td/><td/><td align=\"center\"><bold>N</bold>147<bold>Q</bold></td></tr><tr><td/><td/><td/><td align=\"center\"><bold>N</bold>356<bold>K</bold></td></tr><tr><td/><td/><td/><td align=\"center\"><bold>N</bold>362<bold>H</bold></td></tr><tr><td/><td/><td/><td align=\"center\"><bold>N</bold>380<bold>Q</bold></td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"center\">DBL2-x</td><td align=\"center\">530 – 863</td><td align=\"center\">2</td><td align=\"center\"><bold>N</bold>692<bold>Q</bold></td></tr><tr><td/><td/><td/><td align=\"center\"><bold>N</bold>730<bold>Q</bold></td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"center\">DBL3-x</td><td align=\"center\">1221 – 1548</td><td align=\"center\">3</td><td align=\"center\"><bold>N</bold>1222<bold>S</bold></td></tr><tr><td/><td/><td/><td align=\"center\"><bold>N</bold>1290<bold>Q</bold></td></tr><tr><td/><td/><td/><td align=\"center\"><bold>N</bold>1428<bold>K</bold></td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"center\">DBL4-ε</td><td align=\"center\">1594 – 1888</td><td align=\"center\">4</td><td align=\"center\"><bold>N</bold>1674<bold>H</bold></td></tr><tr><td/><td/><td/><td align=\"center\"><bold>N</bold>1744<bold>Q</bold></td></tr><tr><td/><td/><td/><td align=\"center\"><bold>N</bold>1749<bold>Q</bold></td></tr><tr><td/><td/><td/><td align=\"center\"><bold>N</bold>1844<bold>Q</bold></td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"center\">DBL5-ε</td><td align=\"center\">2003 – 2270</td><td align=\"center\">3</td><td align=\"center\"><bold>N</bold>2134<bold>Q</bold></td></tr><tr><td/><td/><td/><td align=\"center\"><bold>N</bold>2210<bold>Q</bold></td></tr><tr><td/><td/><td/><td align=\"center\"><bold>N</bold>2223<bold>Q</bold></td></tr><tr><td colspan=\"4\"><hr/></td></tr><tr><td align=\"center\">DBL6-ε</td><td align=\"center\">2322 – 2590</td><td align=\"center\">2</td><td align=\"center\"><bold>N</bold>2442<bold>Q</bold></td></tr><tr><td/><td/><td/><td align=\"center\"><bold>N</bold>2537<bold>Q</bold></td></tr></tbody></table></table-wrap>" ]

[]