accession_id
stringlengths 9
11
| pmid
stringlengths 1
8
| introduction
stringlengths 0
134k
| methods
stringlengths 0
208k
| results
stringlengths 0
357k
| discussion
stringlengths 0
357k
| conclusion
stringlengths 0
58.3k
| front
stringlengths 0
30.9k
| body
stringlengths 0
573k
| back
stringlengths 0
126k
| license
stringclasses 4
values | retracted
stringclasses 2
values | last_updated
stringlengths 19
19
| citation
stringlengths 14
94
| package_file
stringlengths 0
35
|
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
PMC3014918 | 21118541 | Background
Vibrio cholerae is the etiological agent of the severe diarrheal disease cholera. Similar to many Gram-negative enteric pathogens, horizontal gene transfer and recombination plays a significant role in the evolution and emergence of new pathogenic strain of this species [ 1 - 12 ]. The main cause of the explosive rice water diarrhea characteristic of cholera is the cholera toxin (CT), an AB type enterotoxin, which is encoded within the ssDNA filamentous phage CTXɸ [ 13 , 14 ]. The B subunit of CT binds to the GM 1 gangliosides, which are exposed when higher order gangliosides found in the intestinal mucus are cleaved by sialidase/neuraminidase (NanH). This protein is encoded within a 57 kb region named Vibrio Pathogenicity Island-2 (VPI-2) [ 15 , 16 ]. In addition to encoding sialidase, VPI-2 also encodes the sialic acid catabolism (SAC) gene cluster (Figure 1A ) [ 16 - 19 ]. The SAC cluster was shown to be present only in pathogenic isolates of V. cholerae and enables the bacterium to grow on sialic acid as a sole carbon source [ 18 , 20 ]. Recently, we demonstrated that the ability to catabolize sialic acid gives V. cholerae a competitive advantage in vivo [ 19 ]. In non-O1/O139 pathogenic isolates, in addition to the SAC cluster are the genes required for a type 3 secretion system which is important for virulence [ 21 - 25 ]. The toxin co-regulated pilus (TCP), an essential intestinal colonization factor for V. cholerae , is encoded within the 40 kb Vibrio Pathogenicity Island-1 (VPI-1 or TCP Island) region [ 26 , 27 ].
VPI-1 and VPI-2 do not share any genes in common but do share some functional characteristics such as the ability to integrate into the chromosome, specifically at a tRNA site using an integrase belonging to the tyrosine recombinase family [ 16 , 18 , 23 , 26 , 28 ]. VPI-2 integrates into chromosome 1 at a tRNA-serine locus, whereas VPI-1 is located at the tmRNA locus. Both regions are flanked by direct repeats (DRs) named attL and attR [ 16 , 18 , 23 , 26 , 28 ]. These integrases, IntV1 (VC0847) and IntV2 (VC1758), are believed to mediate insertion into the host chromosome through site specific recombination between an attachment site attP , present in the pathogenicity island, and attB , present in the bacterial chromosome.
Pathogenicity islands have been shown to excise from their host genome in pathogenic Escherichia coli and Yersinia species [ 29 - 36 ]. In E. coli strain 536, a uropathogenic isolate, Hacker and colleagues have identified six PAIs, all of which encode a tyrosine recombinase integrase and are flanked by DRs [ 31 , 33 , 36 - 39 ]. They demonstrated that PAI-I, II, III and V can excise from the chromosome by site-specific recombination involving their respective DRs ( attL and attR ) [ 31 , 33 ]. The PAIs were shown to excise at different frequencies depending on the growth conditions [ 31 , 33 ]. Likewise, both VPI-1 and VPI-2 have been shown to excise from their host chromosome [ 23 , 28 ]. Rajanna and colleagues demonstrated that VPI-1 can excise from V. cholerae N16961 at very low rates [ 28 ]. They determined that the integrase IntV1 (VC0847) was not essential for excision since a transposase within the region appeared to compensate for an IntV1 knockout [ 28 ]. Recently, Murphy and Boyd demonstrated that VPI-2 from V. cholerae N16961 can excise from chromosome 1, which also occurred at very low frequency under optimal growth conditions [ 23 ]. Their study showed that IntV2 (VC1758) was essential for excision and the formation of a circular intermediate (CI) [ 23 ]. Pathogenicity islands from both E. coli and V. cholerae are non-self mobilizable, they do not encode any proteins such as those for phage structural proteins or conjugation systems needed for cell to cell mobility [ 23 , 28 , 31 , 33 , 36 - 39 ]. The mechanism of transfer for most pathogenicity islands remains to be elucidated but likely involves hitchhiking with plasmids, conjugative transposons, Integrative and Conjugative Elements (ICEs), or generalized transducing phages or uptake by transformation.
It is known that for some mobile and integrative genetic elements (MIGEs) the presence of a recombination directionality factor (RDF)/excisionase is required for excision [ 40 , 41 ]. For instance, Xis is required for the excision of the ICE SXT from V. cholerae [ 41 ], Hef from the High Pathogenicity Island of Yersinia pestis [ 32 ], and Rox from the Shigella Resistance Locus (SRL) of Shigella flexneri [ 42 ]. RDFs are small basic proteins that bind and bend DNA on the recombination sites attL and attR triggering excision by coordinating the assembly of the excisive intasome [ 43 - 45 ]. In addition, some RDFs have been found to inhibit reintegration of the CI by converting attP into a catalytically inactive structure and are thought to stabilize the appropriate positioning of the integrase within the excisive intasome [ 46 - 48 ]. To date, no RDFs have been identified in E. coli or V. cholerae pathogenicity islands.
Here, we report the environmental conditions that induce excision of VPI-2. We examined the VPI-2-encoded factors that are required for VPI-2 excision, determining that V. cholerae cells subjected to stress conditions showed an increase in the excision levels of VPI-2 compared to cell grown at optimal conditions. Bioinformatic analysis of the VPI-2 region identified two open reading frames (ORFs) VC1785 and VC1809 that show homology to previously described RDFs, which we named VefA and VefB. We examined the role of these genes in VPI-2 excision. | Methods
Bacterial strains and growth conditions
The strains and plasmids used in this study are listed in table 1 . Bacteria were grown in lysogeny broth more commonly known as Luria-Bertani broth (LB), LB agar, or LB agar 10% sucrose without NaCl (LB-Suc) [ 49 ]. Strains harboring the pBAD33 expression vector were grown on LB supplemented with 0.02% W/V of L-Arabinose (LB-Ara). Bacteria were incubated overnight at 37°C with aeration unless otherwise indicated. When required, ampicillin (Amp, 100 μg/ml), streptomycin (Sm, 200 μg/ml), or chloramphenicol (Cm, 25 μg/ml) were added to the media.
Determination of VPI-2 excision rate
Excised circular VPI-2 DNA containing attP is expected to be a very rare event given the predicted low excision rate under normal conditions and the inability of VPI-2 to replicate after excision [ 23 ]. Therefore, we quantified the excision rates of VPI-2 by measuring the presence of attB , the locus present on the V. cholerae chromosome after VPI-2 excises (Figure 1B ), in different strains under different conditions, and comparing it with the presence of attB in cultures of V. cholerae N16961 grown under standard optimal conditions: 12 hours in LB at 37°C with aeration. Using the O.D. values of 1 mL of a culture of V. cholerae N16961 grown for 12 hours in LB at 37°C with aeration as a reference, 750 μL to 4 mL were pelleted by centrifugation and genomic DNA was extracted using ABI PrepMan Ultra reagent from the test cultures. We took 50 μL from each DNA extraction and diluted each with 200 μL of sterile ddH 2 O. A 5 μL aliquot of DNA after dilution was used as template for Real-Time quantitative PCR (QPCR) reactions. The QPCR assay calculated the percentage of cells in a culture that contained an unoccupied VPI-2 attB site. We quantified attB sites present in cell grown under different growth conditions and normalized to the amount of attB present in N16961 grown for 12 hours at 37°C. The gene-specific primers were designed using Primer3 software according to the real-time PCR guidelines, and are listed in Table 2 . The Applied Biosystems 7000 system was used for RT fluorescence detection of PCR products that resulted from binding of the dye SYBR Green to double stranded DNA and the results were examined with Applied Biosystems SDS software V 1.3. The reference gene mdh was assayed both separately and in the same reaction. To confirm that primer pairs only amplified target genes to assure accurate quantification of the results, non-template controls were included in each replicate. The attB and mdh PCR products were visually checked on agarose gels. The melting curves of PCR products were used to ensure the absence of primer dimers, contamination with genomic DNA and non-specific homologous sequences. PCR reactions were performed in 10 uL volumes containing 5 uL of 2X SYBR Green PCR Master Mix (Applied Biosystems), 900 nm of each primer, and 1 uL of DNA template. PCR cycling conditions were 30 sec at 95°C followed by 40 cycles of 15 sec at 95°C and 30 sec at 60°C. Serial doubling dilutions were used as templates for QPCR to generate standard curves for each PCR reaction by plotting relative DNA concentrations versus log (C t ) value (C t is the PCR cycle at which fluorescence rises beyond background). The C t value for mdh was 15 cycles and for attB 30 cycles. Every sample was assayed in triplicate and each experiment was performed using a minimum of three different samples. Differences in the attB ratio were extrapolated using the delta-delta Ct method as developed by Pfaffl [ 50 ].
Bioinformatic analysis
BLAST search was performed using Xis (ABA87014), an RDF from V.cholerae SXT ICE element required for excision, AlpA, a well known RDF from E. coli (AAA18418) and the Hef protein (NP_405464) from Y. pestis pathogenicity island, as seeds on the genome sequence of V. cholerae N16961 [ 51 ]. DNA sequences from putative RDFs were downloaded from GenBank and the sequences were aligned using ClustalW [ 52 ]. Next, the protein sequences of characterized RDFs were used as seeds to perform BLASTN and BLASTP searches against Vibrio genomes sequences in the database [ 51 ]. The retrieved sequence must give an e-value below 10 -3 , relative to known RDFs.
RNA extraction and Real-Time quantitative PCR (QPCR)
Total RNA from V. cholerae N16961 was extracted 12 hours post-inoculation in LB broth from one group treated with sub-lethal UV-irradiation and one group untreated as follows. The cells from 5 mLs 11 hours growth cultures were pelleted and resuspended in 5 mLs of PBS. A 100 uL aliquot was taken from each sample prior to treatment to calculate colony forming units (CFUs). Each 5 mL sample was placed in a plastic Petri dishes without a cover and one set of samples was irradiated with a sublethal dose of 25 J/m 2 of UV irradiation in a Fisher Scientific UV cross linker (FB-UVXL-1000) and the other 5 mL set of samples was left untreated as previously described by others [ 53 ]. The cells from both UV treated and untreated samples were recovered, pelleted, resuspended in 5 mLs of LB broth and grown for 1 hour at 37°C. A 100 uL aliquot was taken from each sample to calculate CFUs post treatment from both sets of samples. The CFU counts pre and post treatment were identical at ~9.75 × 10 9 /ml as expected. Every experiment was performed in triplicate. Total RNA was extracted from each culture using RNAprotect Bacteria reagent (Qiagen, Valencia, CA) and an RNeasy mini kit (Qiagen) according to the manufacturer's protocols. RNA purity and the presence of genomic DNA were assessed using an ND-1000 NanoDrop UV-Vis spectrophotometer (NanoDrop Technologies) giving values of A 260 / A 280 > 2.0 and A 260 / A 230 > 2.0 indicating of no protein and solvent contamination, respectively. In addition, 1 μg of each sample of RNA was run on a 1% agarose gel in 1× TBE buffer to examine quality of the samples. RNA was measured to calculate the volume of sample to be added to perform a reverse transcriptase (RT) reaction using SuperScript II Reverse Transcriptase and random hexamers following manufacturer's instructions (Invitrogen). The purity and quantity of cDNA was examined using an ND-1000 NanoDrop UV-Vis spectrophotometer as above. QPCR was performed using standard protocol using primer pairs for vc1758, vc1785, vc1809 and vc0432 ( intV2, vefA, vefB and mdh , respectively) listed in Table 2 using SYBR green PCR Master Mix (Invitrogen) on an Applied Biosystems 7000 Real Time PCR System (Foster City, CA). To confirm that primer pairs only amplified target genes to assure accurate quantification of the results, non-template controls were included in each replicate. The intV2, vefA, vefB and mdh PCR products were visually checked on agarose gels. The melting curves of PCR products were used to ensure the absence of primer dimers, contamination with genomic DNA and non-specific homologous sequences. The data was analyzed using ABI PRISM 7000 SDS software (Applied Biosystems). Differences in the gene ratios were extrapolated using the delta-delta Ct method [ 50 ]. Every sample was assayed in triplicate and each experiment was performed using a minimum of three different samples.
Construction of mutant strains
To construct the mutant strains, primers were designed to conduct Splice Overlap Extension (SOE) PCR followed by allelic exchange [ 54 ]. SOE PCR primers were designed to produce non-functioning constructs of the 204-bp vefA and the 228-bp vefB genes. The size of the regions removed from vefA and vefB is 169-bp and 191-bp, respectively and were constructed in V . cholerae strain N16961 to create mutant strains V . cholerae SAM-3 and SAM-4, respectively (Table 1 ). Primer pairs SOEVC1785A/SOEVC1785B and SOEVC1785C/SOEVC1785 D were used to amplify PCR products from VC1785 from V. cholerae strain N16961 (Table 2 ). The ligated product was amplified with primer pair SOEVC1785A and SOEVC1785 D, which was restricted with enzymes, XbaI and SacI and ligated with pDS132 (New England Biolabs) resulting in pΔ1785. pΔ1785 was transformed into E. coli strain DH5αλpir, plasmid purified and then transformed into E. coli β2155 cells. E. coli β2155 transformants were conjugated with N16961. V. cholerae cells were passaged in LB-suc to cure them of the integrated pΔ1785. PCR was used to screen for V. cholerae strains in which the wild type gene was replaced by the mutant gene, which was confirmed by sequencing. The Δ1785 strain was designated V. cholerae strain SAM-3. A knockout mutant of VC1809 was constructed in N16961 as described above using primer pairs listed in Table 2 . Complementation of RAM-1 and SAM-3 mutant strains were generated by creating pIntV2 and pVefA, by cloning intV2 (VC1758) and vefA (VC1785), respectively into the SacI/XbaI sites of the expression plasmid pBAD33 (New England Biolabs) using standard cloning protocol (Table 1 and 2 ). | Results and Discussion
VPI-2 excision rates under different growth conditions
It was previously shown that the four pathogenicity islands identified in V. cholerae N16961 can excise from chromosome 1 and form circular intermediates (CI) [ 23 , 28 ]. The excision of VPI-1 and VPI-2 occurs at very low levels suggesting that excision is tightly controlled, although it may also suggest that the excision event is inefficient, possibly due to poor expression of the regulatory genes, an altered regulatory circuit, or mutations that might occur in these sequences as the region become evolutionarily integrated into the host chromosome [ 23 , 28 ]. First, we quantified the excision levels of VPI-2 in cultures of V. cholerae N16961 grown for 12 hours in LB at 37°C (standard conditions) by measuring the presence of attB , the locus present on the chromosome after VPI-2 excises (Figure 1 ), and comparing it with the housekeeping gene mdh using QPCR. We used attB as a surrogate for VPI-2 excision measurements since the copy number of attP in the CI is minuscule compared to attB , which replicates along with the rest of the chromosome unlike excised VPI-2. We compared the presence of attB with mdh since all cells encode one functional copy of the latter. PCR products of attB and mdh were visually checked on an agarose gel and their melting temperature analyzed to ensure we had the correct PCR products. The reference gene was assayed both separately and in the same reaction. Both primer pairs used were tested by comparing the results obtained using previously quantified cloned copies of mdh and attB and gave comparable results. We found that attB was present in 1 in every 1.6 (±0.2) × 10 6 V. cholerae cells under optimal growth conditions.
Next, we measured the presence of attB from V. cholerae cells cultured under different conditions compared with the presence of attB under our standard condition, growth for 12 hours at 37°C. We determined that incubation time does not affect the excision levels of VPI-2 indicating that excision does not occur in a growth phase dependent manner (Figure 2 ). However, V . cholerae cultures grown at 25°C showed a 2-fold increase in the presence of the attB site compared to cells grown at the optimum temperature 37°C (Figure 2 ). In addition, we found that nutrient limitation affected the excision level showing over a 5-fold decrease in the presence of attB when compared to the growth on LB at the same temperature (Figure 2 ). Furthermore, we found that sub-lethal UV-light irradiation of cell cultures compared to untreated cells, resulted in a significant increase in the level of excision of VPI-2, over 4-fold compared to untreated cells grown under the same conditions (Figure 2 ). Taken together, these data indicate that environmental factors can affect the induction of excision and circularization of VPI-2, which is probably the first step required for the horizontal transfer of the region. These results are consistent with what was previously shown for other mobile and integrative genetic elements as well as PAIs from E. coli , where excision occurs upon exposure to stress conditions such as sub-lethal UV-light irradiation [ 53 , 55 , 56 ].
VPI-2 encodes two novel recombination directionality factors
Both the high pathogenicity island HPI from Y. pestis and ICE SXT from V. cholerae encode small accessory proteins called recombination directionality factors (RDFs) or excisionases (Xis) that are required for efficient excision of these elements [ 29 , 41 ]. In order to identify candidate RDFs within VPI-2 from V. cholerae N16961, we performed BLAST and PSI-BLAST searches on the V. cholerae N16961 genome using RDFs, the V. cholerae Xis protein (ABA87014) from SXT, the Y. pestis Hef protein (NP_405464) from HPI and E. coli K12 AlpA protein (AAA18418) from λ phage as queries [ 57 ]. The most significant BLAST result in these searches was ORF VC0497, which is annotated as a transcriptional regulator, and is encoded within Vibrio Seventh Pandemic island-II (VSP-II). VSP-II also encodes a tyrosine recombinase integrase at ORF VC0516 (IntV3) [ 58 ]. ORFs VC1785 and VC1809 encoded within VPI-2 were the second and third most significant hits retrieved from these BLAST searches, which we termed VefA (for Vibrio excision factor A) and VefB, respectively (Figure 3 ). The VefA and VefB proteins share 46% amino acid identity/72% similarity. VefA shares 37% amino acid identities with AlpA, 46% identity with Hef and 29% with Xis from the V. cholerae SXT element as was previously shown [ 53 ] (Figure 3 ). The vefB gene is located at the 3' end of VPI-2 at ORF VC1809 marking the end of the island, and vefA (VC1785) is adjacent to neuraminidase gene, nanH (VC1784) in the middle of the island (Figure 1A ).
As shown in Figure 2 , UV-light irradiation increased excision of VPI-2 over 4-fold. In order to investigate this further, we determined the effect of UV-light irradiation on the expression of intV2, vefA and vefB in V. cholerae N16961 (Figure 4 ). We examined transcript levels of intV2, vefA and vefB in cells grown for 12 h in LB and in cells grown for 12 h in LB followed UV-light irradiation treatment. We found that all three genes showed negligible levels of transcription under standard optimum growth conditions but after UV-light treatment both intV2 and vefA show a 10-fold and vefB a 5-fold increase in expression levels (Figure 4 ). These results indicate that UV-light induces expression of factors potentially involved in VPI-2 excision.
IntV2 and VefA are essential for the excision of VPI-2
To determine in more detail the role of intV2, vefA and vefB in VPI-2 excision, we created deletion mutations in each gene and measured excision levels of VPI-2 by determining attB levels in cells. In V. cholerae RAM-1, an intV2 mutant, we did not detect any VPI-2 attB products, demonstrating that intV2 is essential for excision as was previously shown (Figure 5 ) [ 23 ]. We complemented RAM-1 with a functional copy of intV2 by transforming V. cholerae RAM-1 with pIntV2 creating strain SAM-1. In our SAM-1 strain, we found that excision of VPI-2 was restored in addition, attB levels were approximately four-fold higher than wild-type levels which is represented by the dotted broken horizontal line in Figure 5 . These data demonstrate that over expressing intV2 ectopically induces excision of VPI-2. In our control experiments, transformation of either wild-type N16961 or RAM-1 with pBAD33 alone (strains SAM-11 and SAM-12 respectively) did not affect attB levels (data not shown).
Next, we determined whether one, both, or neither of the putative RDFs uncovered by our bioinformatic analysis are required for VPI-2 excision. To do this, we constructed in-frame deletion mutations in each gene to create mutant strain SAM-3 (Δ vefA ) and SAM-4 (Δ vefB) . The two mutant strains and the wild-type N16961 were each inoculated into LB and all three strains grew similarly indicating that the mutant constructs did not have any general growth defect (data not shown). We determined the attB levels using QPCR in strain SAM-3 compared to the wild-type strain grown under the same conditions. We found that no VPI-2 excision occurs in SAM-3 cells when compared with the wild type, indicating that a functional copy of vefA is essential for efficient excision of VPI-2 (Figure 5 ). We complemented SAM-3 with a functional copy of vefA (SAM-5) and measured attB levels in these cells with the wild type levels both under standard conditions, to find that some excision occurred, but it was less than in wild-type cells (Figure 5 ). In our vefB mutant strain (SAM-4), we found no difference in VPI-2 excision levels compared to wild-type grown under the same conditions, which demonstrates that vefB is not essential for excision (Figure 5 ). From these data it appears that vefA is the cognate RDF for VPI-2 excision. In our control experiments, transformation of SAM-3 with pBAD33 alone (resulting in strain SAM-13) did not affect attB levels (data not shown).
Vibrio species island-encoded integrases with corresponding RDFs
Given that our initial search for RDFs within one V. cholerae genome (strain N16961) yielded three putative RDFs (VC0497, VC1785, and VC1809), we decided to investigate further the occurrence of RDFs among Vibrio species whose genome sequence is available in the database. We performed BLAST searches against the 20 Vibrio species in the genome database, and we uncovered a total of 27 putative RDFs (Table 3 ). Next, we identified putative integrases within the genomes of the RDF homologues using BLAST search analysis by using IntV2 as a seed. For each of the RDFs identified among the 27 genomes encompassing 10 different Vibrio species (V. cholerae, V. coralliilyticus, V. furnissii, V. harveyi, V. parahaemolyticus, V. splendidus, V. vulnificus, Vibrio sp. Ex25, RC341, and MED222), we identified a corresponding integrase with greater than 40% amino acid identities to IntV2 (VC1758) (Table 3 ). We examined the gene context of each RDF and integrase within each of the 20 strains to determine whether the RDF and integrase were located on the same region within a strain. From these analyses, we found that each of the 27 RDFs has a corresponding integrase within approximately 100 kb of each other (Table 3 ). It should be noted that from table 3 , only three of the strains have been annotated completely and for many of the strains examined their ORF annotation numbering is not consecutive. Within V. cholerae , integrases and RDFs located in the same region of the genome in different strains had the same gene content indicating the same island is present in different strains. Among the different species, however, integrases and RDFs associated with the same insertion site did not have the same gene content indicating a novel island region in the different species (data not shown).
From our analysis, no RDF was identified within the VPI-1 or the VSP-I regions in N16961 or within homologous regions in the other 27 sequenced strains of V. cholerae in the database. Both the VPI-1 and VSP-I regions have been shown to excise from their chromosome location, and VPI-1 encodes a tyrosine recombinase with homology to IntV2, thus they may therefore use an alternative mechanism of excision or perhaps co-opt an RDF from another region on the genome. Overall our data indicates that the presence of both an integrase and a cognate RDF pairing is a relatively conserved feature but not an essential one. | Results and Discussion
VPI-2 excision rates under different growth conditions
It was previously shown that the four pathogenicity islands identified in V. cholerae N16961 can excise from chromosome 1 and form circular intermediates (CI) [ 23 , 28 ]. The excision of VPI-1 and VPI-2 occurs at very low levels suggesting that excision is tightly controlled, although it may also suggest that the excision event is inefficient, possibly due to poor expression of the regulatory genes, an altered regulatory circuit, or mutations that might occur in these sequences as the region become evolutionarily integrated into the host chromosome [ 23 , 28 ]. First, we quantified the excision levels of VPI-2 in cultures of V. cholerae N16961 grown for 12 hours in LB at 37°C (standard conditions) by measuring the presence of attB , the locus present on the chromosome after VPI-2 excises (Figure 1 ), and comparing it with the housekeeping gene mdh using QPCR. We used attB as a surrogate for VPI-2 excision measurements since the copy number of attP in the CI is minuscule compared to attB , which replicates along with the rest of the chromosome unlike excised VPI-2. We compared the presence of attB with mdh since all cells encode one functional copy of the latter. PCR products of attB and mdh were visually checked on an agarose gel and their melting temperature analyzed to ensure we had the correct PCR products. The reference gene was assayed both separately and in the same reaction. Both primer pairs used were tested by comparing the results obtained using previously quantified cloned copies of mdh and attB and gave comparable results. We found that attB was present in 1 in every 1.6 (±0.2) × 10 6 V. cholerae cells under optimal growth conditions.
Next, we measured the presence of attB from V. cholerae cells cultured under different conditions compared with the presence of attB under our standard condition, growth for 12 hours at 37°C. We determined that incubation time does not affect the excision levels of VPI-2 indicating that excision does not occur in a growth phase dependent manner (Figure 2 ). However, V . cholerae cultures grown at 25°C showed a 2-fold increase in the presence of the attB site compared to cells grown at the optimum temperature 37°C (Figure 2 ). In addition, we found that nutrient limitation affected the excision level showing over a 5-fold decrease in the presence of attB when compared to the growth on LB at the same temperature (Figure 2 ). Furthermore, we found that sub-lethal UV-light irradiation of cell cultures compared to untreated cells, resulted in a significant increase in the level of excision of VPI-2, over 4-fold compared to untreated cells grown under the same conditions (Figure 2 ). Taken together, these data indicate that environmental factors can affect the induction of excision and circularization of VPI-2, which is probably the first step required for the horizontal transfer of the region. These results are consistent with what was previously shown for other mobile and integrative genetic elements as well as PAIs from E. coli , where excision occurs upon exposure to stress conditions such as sub-lethal UV-light irradiation [ 53 , 55 , 56 ].
VPI-2 encodes two novel recombination directionality factors
Both the high pathogenicity island HPI from Y. pestis and ICE SXT from V. cholerae encode small accessory proteins called recombination directionality factors (RDFs) or excisionases (Xis) that are required for efficient excision of these elements [ 29 , 41 ]. In order to identify candidate RDFs within VPI-2 from V. cholerae N16961, we performed BLAST and PSI-BLAST searches on the V. cholerae N16961 genome using RDFs, the V. cholerae Xis protein (ABA87014) from SXT, the Y. pestis Hef protein (NP_405464) from HPI and E. coli K12 AlpA protein (AAA18418) from λ phage as queries [ 57 ]. The most significant BLAST result in these searches was ORF VC0497, which is annotated as a transcriptional regulator, and is encoded within Vibrio Seventh Pandemic island-II (VSP-II). VSP-II also encodes a tyrosine recombinase integrase at ORF VC0516 (IntV3) [ 58 ]. ORFs VC1785 and VC1809 encoded within VPI-2 were the second and third most significant hits retrieved from these BLAST searches, which we termed VefA (for Vibrio excision factor A) and VefB, respectively (Figure 3 ). The VefA and VefB proteins share 46% amino acid identity/72% similarity. VefA shares 37% amino acid identities with AlpA, 46% identity with Hef and 29% with Xis from the V. cholerae SXT element as was previously shown [ 53 ] (Figure 3 ). The vefB gene is located at the 3' end of VPI-2 at ORF VC1809 marking the end of the island, and vefA (VC1785) is adjacent to neuraminidase gene, nanH (VC1784) in the middle of the island (Figure 1A ).
As shown in Figure 2 , UV-light irradiation increased excision of VPI-2 over 4-fold. In order to investigate this further, we determined the effect of UV-light irradiation on the expression of intV2, vefA and vefB in V. cholerae N16961 (Figure 4 ). We examined transcript levels of intV2, vefA and vefB in cells grown for 12 h in LB and in cells grown for 12 h in LB followed UV-light irradiation treatment. We found that all three genes showed negligible levels of transcription under standard optimum growth conditions but after UV-light treatment both intV2 and vefA show a 10-fold and vefB a 5-fold increase in expression levels (Figure 4 ). These results indicate that UV-light induces expression of factors potentially involved in VPI-2 excision.
IntV2 and VefA are essential for the excision of VPI-2
To determine in more detail the role of intV2, vefA and vefB in VPI-2 excision, we created deletion mutations in each gene and measured excision levels of VPI-2 by determining attB levels in cells. In V. cholerae RAM-1, an intV2 mutant, we did not detect any VPI-2 attB products, demonstrating that intV2 is essential for excision as was previously shown (Figure 5 ) [ 23 ]. We complemented RAM-1 with a functional copy of intV2 by transforming V. cholerae RAM-1 with pIntV2 creating strain SAM-1. In our SAM-1 strain, we found that excision of VPI-2 was restored in addition, attB levels were approximately four-fold higher than wild-type levels which is represented by the dotted broken horizontal line in Figure 5 . These data demonstrate that over expressing intV2 ectopically induces excision of VPI-2. In our control experiments, transformation of either wild-type N16961 or RAM-1 with pBAD33 alone (strains SAM-11 and SAM-12 respectively) did not affect attB levels (data not shown).
Next, we determined whether one, both, or neither of the putative RDFs uncovered by our bioinformatic analysis are required for VPI-2 excision. To do this, we constructed in-frame deletion mutations in each gene to create mutant strain SAM-3 (Δ vefA ) and SAM-4 (Δ vefB) . The two mutant strains and the wild-type N16961 were each inoculated into LB and all three strains grew similarly indicating that the mutant constructs did not have any general growth defect (data not shown). We determined the attB levels using QPCR in strain SAM-3 compared to the wild-type strain grown under the same conditions. We found that no VPI-2 excision occurs in SAM-3 cells when compared with the wild type, indicating that a functional copy of vefA is essential for efficient excision of VPI-2 (Figure 5 ). We complemented SAM-3 with a functional copy of vefA (SAM-5) and measured attB levels in these cells with the wild type levels both under standard conditions, to find that some excision occurred, but it was less than in wild-type cells (Figure 5 ). In our vefB mutant strain (SAM-4), we found no difference in VPI-2 excision levels compared to wild-type grown under the same conditions, which demonstrates that vefB is not essential for excision (Figure 5 ). From these data it appears that vefA is the cognate RDF for VPI-2 excision. In our control experiments, transformation of SAM-3 with pBAD33 alone (resulting in strain SAM-13) did not affect attB levels (data not shown).
Vibrio species island-encoded integrases with corresponding RDFs
Given that our initial search for RDFs within one V. cholerae genome (strain N16961) yielded three putative RDFs (VC0497, VC1785, and VC1809), we decided to investigate further the occurrence of RDFs among Vibrio species whose genome sequence is available in the database. We performed BLAST searches against the 20 Vibrio species in the genome database, and we uncovered a total of 27 putative RDFs (Table 3 ). Next, we identified putative integrases within the genomes of the RDF homologues using BLAST search analysis by using IntV2 as a seed. For each of the RDFs identified among the 27 genomes encompassing 10 different Vibrio species (V. cholerae, V. coralliilyticus, V. furnissii, V. harveyi, V. parahaemolyticus, V. splendidus, V. vulnificus, Vibrio sp. Ex25, RC341, and MED222), we identified a corresponding integrase with greater than 40% amino acid identities to IntV2 (VC1758) (Table 3 ). We examined the gene context of each RDF and integrase within each of the 20 strains to determine whether the RDF and integrase were located on the same region within a strain. From these analyses, we found that each of the 27 RDFs has a corresponding integrase within approximately 100 kb of each other (Table 3 ). It should be noted that from table 3 , only three of the strains have been annotated completely and for many of the strains examined their ORF annotation numbering is not consecutive. Within V. cholerae , integrases and RDFs located in the same region of the genome in different strains had the same gene content indicating the same island is present in different strains. Among the different species, however, integrases and RDFs associated with the same insertion site did not have the same gene content indicating a novel island region in the different species (data not shown).
From our analysis, no RDF was identified within the VPI-1 or the VSP-I regions in N16961 or within homologous regions in the other 27 sequenced strains of V. cholerae in the database. Both the VPI-1 and VSP-I regions have been shown to excise from their chromosome location, and VPI-1 encodes a tyrosine recombinase with homology to IntV2, thus they may therefore use an alternative mechanism of excision or perhaps co-opt an RDF from another region on the genome. Overall our data indicates that the presence of both an integrase and a cognate RDF pairing is a relatively conserved feature but not an essential one. | Conclusions
In this study, we analyzed the excision dynamics of VPI-2 encoded within V. cholerae N16961. Our results indicate that excision is controlled by at least two conserved factors within the island, an integrase encoded by intV2 and an RDF encoded by vefA , whose expression is induced by environmental stimuli similar to other MIGEs such as prophages, ICEs and integrons. We identified two putative RDFs and found that of the two we identified, only one VefA is essential for the efficient excision of VPI-2. We determined the occurrence of RDFs among the genomes of sequenced Vibrio species and found 27 putative RDFs that also had a homologue of IntV2 associated with it, which suggests that requirement for both an RDF and a corresponding integrase is a relatively common feature. | Background
Vibrio Pathogenicity Island-2 (VPI-2) is a 57 kb region present in choleragenic V. cholerae isolates that is required for growth on sialic acid as a sole carbon source. V. cholerae non-O1/O139 pathogenic strains also contain VPI-2, which in addition to sialic acid catabolism genes also encodes a type 3 secretion system in these strains. VPI-2 integrates into chromosome 1 at a tRNA-serine site and encodes an integrase intV2 (VC1758) that belongs to the tyrosine recombinase family. IntV2 is required for VPI-2 excision from chromosome 1, which occurs at very low levels, and formation of a non-replicative circular intermediate.
Results
We determined the conditions and the factors that affect excision of VPI-2 in V. cholerae N16961. We demonstrate that excision from chromosome 1 is induced at low temperature and after sublethal UV-light irradiation treatment. In addition, after UV-light irradiation compared to untreated cells, cells showed increased expression of three genes, intV2 (VC1758), and two putative recombination directionality factors (RDFs), vefA (VC1785) and vefB (VC1809) encoded within VPI-2. We demonstrate that along with IntV2, the RDF VefA is essential for excision. We constructed a knockout mutant of vefA in V. cholerae N16961, and found that no excision of VPI-2 occurred, indicating that a functional vefA gene is required for excision. Deletion of the second RDF encoded by vefB did not result in a loss of excision. Among Vibrio species in the genome database, we identified 27 putative RDFs within regions that also encoded IntV2 homologues. Within each species the RDFs and their cognate IntV2 proteins were associated with different island regions suggesting that this pairing is widespread.
Conclusions
We demonstrate that excision of VPI-2 is induced under some environmental stress conditions and we show for the first time that an RDF encoded within a pathogenicity island in V. cholerae is required for excision of the region. | Authors' contributions
EFB designed the research; SA-M and MGN performed the research; SA-M, MGN and EFB analyzed data; SA-M, MGN and EFB wrote the paper. | Acknowledgements
This research was supported by National Science Foundation CAREER award DEB-0844409 to E.F.B.
The authors declare no conflicts of interest. | CC BY | no | 2022-01-12 16:12:49 | BMC Microbiol. 2010 Nov 30; 10:306 | oa_package/31/95/PMC3014918.tar.gz |
PMC3014919 | 21122131 | Background
Streptococcus iniae ( S. iniae ) is a hemolytic Gram-positive coccus that is a major pathogen of culture fish. It has been associated with disease outbreak in several species of freshwater and marine fish cultured worldwide, including tilapia [ 1 , 2 ], barramundi [ 3 ], channel catfish [ 4 ], hybrid striped bass [ 1 , 5 ], Japanese flounder [ 6 , 7 ], olive flounder [ 8 ], rabbitfish [ 9 ], and rainbow trout [ 9 , 10 ]. Streptococcal infection can lead to serious symptoms such as meningoencephalitis and generalized septicaemia with high mortality rates of up to 50% [ 9 , 11 ]. S. iniae is also known to be an opportunistic pathogen that can cause fulminant soft tissue infection in humans, such as bacteremic cellulitis, septicarthritis, and endocarditis [ 12 ]. Identifying potential virulence determinants of streptococcal infection will eventually help to the control and eradication of the disease.
Iron plays a significant role in many biological processes and is vital for several metabolic processes. Moreover, many proteins such as cytochromes and tricarboxylic acid metalloenzymes use iron as a cofactor [ 13 ]. Iron is also required for important cellular functions such as the transport and storage of oxygen and as a catalyst in electron transport processes [ 14 ]. The levels of several virulence determinants produced by bacterial pathogens, such as toxins and hemolysins, are depressed under iron-restricted conditions [ 15 ]. Despite its abundance in the natural environment, iron has low solubility under physiological conditions. Moreover, it may be associated with heme or hemo-proteins such as transferrin, lactoferrin, haptoglobin, hemoglobin, and ferritin and such forms do not readily support the growth of microorganisms. Many microorganisms circumvent this nutritional limitation by forming direct contacts with iron-containing proteins through ATP-binding cassette (ABC) transporters.
The ABC transporter superfamilies constitute many different systems that are widespread among living organisms and show different functions, such as ligands translocation, mRNA translation, and DNA repair. The general principle of ABC transport systems involves the ligands translocation through a pore formed by two integral membrane protein domains. This is accompanied by ATP hydrolysis through two nucleotide-binding domains associated with the cytoplasmic side of the pore. In bacteria, ligand translocation is preceded by interaction with an accessory component, i.e., the periplasmic-binding protein [ 16 ].
In this study, an ABC transporter member, named as mtsABC (metal ABC transport system) was cloned from S. iniae HD-1 which is cotranscribed by three genes and was shown to share amino acid sequence homology with the metal ABC transport proteins of other Gram-positive and Gram-negative bacteria. BLAST-mediated sequences similarity searches of the derived amino acid sequences of the mtsABC operon indicated that mtsA encodes a metal solute-binding lipoprotein, mtsB encodes an ATP-binding protein (ATPase), and mtsC encodes a transmembrane permease protein. Our data showed that MtsA is a lipoprotein, and associated with heme. Moreover, this protein is expressed in vivo during Kunming mice infection by S. iniae HD-1. These results provide information on the role of MtsA in heme utilization and the possibility of using MtsA as an effective S. iniae vaccine candidate. | Methods
Bacterial strains and growth conditions
Streptococcus iniae HD-1 was isolated from Threeband sweetlips ( Plectorhynchus cinctus ) from Guangdong province, PRC. The microorganism was stored in our lab and cultured according to the methods described by Zhou et al [ 45 ]. Briefly, S. iniae isolate HD-1 cells were grown in brain heart infusion broth (BHI, Oxoid Ltd.), and BHI broth with 1.5% agar (Guangdong Huankai Microbial Sci. & Tech, Co., Ltd.) was used as the solid medium. Escherichia coli DH5α and BL21 (DE3) strains (Beijing Newprobe Biotechnology Co., Ltd.) were used for gene cloning and protein expression, respectively.
Cloning and reverse transcription analysis of mtsABC
Genomic DNA was extracted from the S. iniae HD-1 strain using the Wizard genomic DNA purification kit (Promega Co., Ltd.), as recommended by the manufacturer, and the material was quantified by measuring the absorbance at 260 nm. PCR was carried out with 1 μg of DNA using the primers listed in Additional file 1 , Table S6. The primers were designed based on the conserved regions of the published amino acid sequence of metal ABC transporter (Additional file 1 , Table S6-1), and the full-length product was obtained by SiteFinding-PCR (Additional file 1 , Table S6-2, 6-3), as described by Tai et al [ 46 ]. The PCR products were sequenced to rule out spurious mutations (Invitrogen Co., Ltd.).
S. iniae HD-1 cells grow to the logarithmic phase were harvested by centrifugation, and total RNA was extracted by the Pure YieldTM RNA midiprep system (Promega, USA, Co., Ltd.). Total RNA was then incubated with RNase I at 37°C for 30 min to remove the contaminating genomic DNA. The material was quantified spectrophotometrically by ultraviolet absorption spectrometry (CE2302, Gene Quest), and its integrity was verified on a 0.8% agarose gel. First-strand cDNA was synthesized from 1 μg total RNA using the first-strand cDNA synthesis kit with ReverTra Ace-α-reverse transcriptase (Toyobo Co., Ltd.). The cDNA synthesized above was used as the template to amplify genes using the ORF-specific primers listed in Additional file 1 , Table S7, and the PCR products were sequenced at Invitrogen Corporation to confirm their specificity.
Expression of recombinant MtsA
The genomic DNA of S. iniae HD-1 was used as the template to amplify the gene encoding amino acids 27 to 310 of mtsA together with primers 5'-GCG GGATCC GCCTCTAAAGATAAG-3' (underlined nucleotides refer to the Bam HI restriction site-initiation codon) and 5'-GCG CTCGAG TTATTTTGCTAAGCCTTCTGAA-3' (underlined nucleotides refer to the Xho I restriction site-stop codon), which were designed from the sequenced mtsA gene. The protein from this cloned amino acid sequence lacks the presumed signal sequence (amino acids 1 to 26). The cloned amino acid fragment was sequenced by Invitrogen Corporation to rule out the possibility of spurious mutations. Recombinant MtsA was purified from E. coli BL21 (DE3) under native conditions using nickel-nitrilotriacetic acid (Ni-NTA) columns (Qiangen, USA) as recommended by the manufacturer. The protein purified by this protocol was free of contaminating proteins, as assessed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). It was quantified by the Bradford assay (CE2302, Gene Quest) using BSA (0.5 mg ml -1 ) as the standard. Specific fractions were then pooled.
Preparation of anti-MtsA antibodies
Anti-sera against histidine-tagged MtsA were prepared in male New Zealand white rabbits (2.2 kg), and approval from the Animal Ethics Committee of Life Sciences Institute was obtained prior to using the animals for research. The experiments were performed as stipulated by the China State Science and Technology Commission [ 47 ]. Rabbits were purchased from Guangdong Laboratory Animals Research Center and acclimatized for 2 weeks in the laboratory of the Life Science Institute prior to use. The rabbits were maintained at the SPF animal center and fed twice daily. They were immunized with 850 μg purified MtsA in 100 μl complete Freund adjuvant (Sigma-Aldrich, Inc.) and then boosted with 170 μg MtsA in 100 μl incomplete Freund adjuvant (Sigma-Aldrich, Inc.) three times at an interval of 15 days. The sera were collected 1 day before the first immunization and 7 days after each booster dose. Purified MtsA and collected sera were used to determine the rabbit anti-MtsA antibody titer by the dot blotting assay.
Extraction of the S. iniae HD-1 lipoprotein
TritonX-114 was used to extract the S. iniae HD-1 lipoprotein, according to the method modified by Cockayne et al [ 48 , 49 ]. Briefly, S. iniae HD-1 cells were cultured, harvested, suspended, and sonicated. Next, 100 μl of 10% TritonX-114 in PBS was added to 2 ml of HD-1 cells lysate and incubated at 4°C for 2 h. After centrifugation at 13,000 × g for 10 min, the supernatant was transferred to a fresh tube and incubated at 37°C for 30 min to allow phase separation. The detergent layer was retained after centrifugation at 13,000 × g for 10 min at room temperature, washed with 1 ml PBS at 4°C for 1 h, and separated from the aqueous phase after incubation at 37°C [ 50 ]. The detergent layer was diluted 1:1 with water, and analyzed by western blotting using the rabbits anti-MtsA antibodies.
Preparation of MtsA cellular fractions
To determine the subcellular localization of MtsA in S. iniae HD-1, the cellular fractions were subjected to two different treatments. In the first treatment procedure, S. iniae HD-1 cells were cultured overnight in 50 ml BHI, harvested, and resuspended in one-tenth volume of Tris buffer (1 M, pH 7.4), and disrupted by sonication (300 W, 5 min). After removing unbroken cells by centrifugation at 10,000 × g , the crude cell lysate was further centrifuged at 248,000 × g for 1 h (OptimaTML-100XP ultracentrifuge, Beckman Coulter). The supernatant and pellet were used as the soluble and particulate fractions of S. iniae cells, respectively [ 51 ]. In the second treatment procedure, the cellular fractions were obtained from S. iniae HD-1 by centrifugation using the protocol of Homonylo-McGavin & Lee [ 52 , 53 ]. Briefly, S. iniae HD-1 cells were grown overnight in 30 ml BHI and then washed by centrifugation at 4°C in a buffer composed of ice-cold 20 mM Tris and 1 mM MgCl 2 (pH 7.0). The cell pellets were resuspended and incubated for 90 min in 0.3 ml of protoplast buffer (150 μl 60% raffinose (Beijing Newprobe Biotechnology Co., Ltd.), 15 μl 1 M Tris (pH 7.4), 6 μl 100 mM phenyl-methyl sulfonyl fluoride (MBchem, Inc.), 3 μl 1 M MgCl 2 , 15 μl 25,000 U ml -1 mutanolysin (Sigma-Aldrich, Inc.), 15 μl 270,000 U ml -1 lysozyme, and 96 μl ddH 2 O). The cell wall extracts were separated from the spheroplasts by centrifugation at 10,000 × g for 10 min. The pelleted protoplasts were washed, suspended in 2 ml PBS-sucrose buffer, and disrupted by sonication, as described above. The supernatant and pellet obtained after centrifugation at 248,000 × g for 1 h were used as the soluble and particulate fractions of the protoplasts, respectively. All cellular fractions were analyzed by western blotting using the rabbit anti-MtsA antibodies.
Detection of the heme-binding activity of MtsA
The pyridine hemochrome assay [ 28 ] was used to analyze heme binding to MtsA. Purified MtsA in 750 μl of 10 mM Tris-HCl (pH 8.0) was mixed with 170 μl of pyridine (Sigma-Aldrich, Inc.), 75 μl of 1 N NaOH, and 2 mg of sodium hydrosulfite (Beijing Newprobe Biotechnology Co., Ltd.), and heme content was determined by measuring the absorbance (■, black square) at 418 nm with a UV-visible spectrophotometer (Uvmini-1240, Shimadzu). Purified catalase-peroxidase (KatG, Beijing Newprobe Biotechnology Co., Ltd.), a known heme-containing protein, was used as the positive control (Δ, white triangle) [ 54 ].
Measurement of iron in MtsA by ICP-AES
The levels of Fe, Zn, Ca, Mg, and Mn in purified MtsA were determined by inductively coupled plasma-atomic emission spectrometry (ICP-AES) using an IRIS (HR) ICP-AES instrument [ 55 ]. Briefly, 0.1 g purified MtsA was immersed in 15 ml nitric acid in an electric cooker. After 3 h nitrification, 1 ml perchloric acid was added and treated for 1 h. The liquid was filter sterilized and analyzed by ICP-AES. A sample lacking purified MtsA was used as the negative control. To achieve contamination-deprivation conditions, all utensils were soaked overnight in nitric acid, and rinsed 6 times with ddH 2 O.
Detection of anti-MtsA antibodies in sera from Kunming mice that were experimentally infected with S. iniae HD-1
To detect the presence of specific anti-MtsA antibodies in the sera from Kunming mice, 10 male Kunming mice (20 ± 2 g) were purchased from Guangdong Laboratory Animals Research Center, and approval from the Animal Ethics Committee of Life Sciences Institute was obtained prior to using the animals for research. The experiments were performed as stipulated by the China State Science and Technology Commission [ 47 ]. Mice were acclimatized at the SPF animal center and fed twice daily for 2 weeks in the laboratory of the Life Science Institute prior to use. Each mouse was injected with 100 μl of 6.2 × 10 8 CFU ml -1 S. iniae HD-1 cells, and the infected sera were collected 10 days post infection. The infected sera and purified MtsA were used in dot-blot and western-blot assays. The sera from 10 Kunming mice injected with PBS were used as the negative control.
Statistical analysis
The nucleotide and deduced amino acid homology analysis of mtsABC was carried out by ClustalX 1.83 and NCBI BLAST http://blast.ncbi.nlm.nih.gov/Blast.cgi . The presumed signal sequence was predicted by the signalP 3.0 Server http://www.cbs.dtu.dk/services/SignalP/ . The theoretical pI/MW was analyzed by the ExPASy Compute pI/MW tool http://www.expasy.org/tools/pi_tool.html . The main domains of mtsABC were detected by the SMART software http://smart.embl-heidelberg.de/ . The amino acid sequences were aligned using the SECentral Align Multi 4 program. To determine whether mtsABC is a Lipoprotein, its sequence was assessed by the ScanProsite analysis software http://www.expasy.ch/tools/scanprosite/ . All statistical analyses were performed using the SPSS 16.0 software (SPSS Inc., USA). | Results
Cloning and reverse transcriptase-PCR analysis of mtsABC
To clone mtsABC from S. iniae HD-1, primers designed based on the conserved regions of the published amino acid sequence of metal ABC transporter were used. The PCR products from genomic DNA template were subsequently sequenced by Invitrogen Corporation. The results showed that the ORFs of mtsA [GeneBank: HQ170628 ], mtsB [GeneBank: HQ170629 ], mtsC [GeneBank: HQ170630 ], mtsAB , and mtsBC had 930, 729, 852, 1724, and 1574 bp respectively. Reverse transcriptase-PCR analysis confirmed that the mtsA gene is the first of three contiguous ORFs that are preceded by a potential promoter region. These three genes are transcribed in the same direction and have very short intergenic sequences, with transcription terminating between mtsA and mtsB , suggesting that these genes constitute a single transcriptional unit (Figure 1 ). No corresponding PCR products were obtained with the same mRNA sample as the template, indicating that the RNA sample was not contaminated with DNA.
Sequence analysis of mtsABC
ABC systems are widespread among living organisms and have been detected in all genera of the three kingdoms of life. These systems show remarkable conservation in the primary sequence of the cassette and in the organization of constitutive domains or subunits [ 17 , 18 ]. All ABC systems share a highly conserved ATP-hydrolyzing domain (nucleotide-binding domain [NBD]) that is unequivocally characterized by three short sequence motifs, i.e., Walker A, Walker B, and a signature motif that is unique to ABC proteins and is located upstream of the Walker B motif [ 19 - 24 ]. BLAST of the derived amino acid sequences of the mtsABC operon indicated that mtsA encodes a metal solute-binding lipoprotein (MtsA, 309 residues), mtsB encodes an ATP-binding protein (MtsB, 242 residues), and mtsC encodes a transmembrane permease protein (MtsC, 283 residues). The closest homologs for these proteins are putative metal ABC transporter proteins encoded by the mtu locus of Streptococcus uberis 0140J and the mts locus of Streptococcus equi subsp. zooepidemicus MGCS10565 (Additional file 1 , Table S1, and Figure 2 ). mtsA contains a helical backbone metal receptor (TroA-like domain) that functions in the ABC transport of ferric siderophores and metal ions such as Fe 3+ , Mn 2+ , Cu 2+ , and/or Zn 2+ (Additional file 1 , Table S2). mtsB contains Walker site A, Walker site B, a signature sequence, and the 4 th motif as defined by Linton & Higgins [ 25 ]. mtsC contains eight transmembrane subunits (TMs) of the periplasmic-binding protein (PBP)-dependent ABC transporters that are possibly involved in the uptake of siderophores, heme, vitamin B 12 , or divalent cations (Additional file 1 , Table S2). Based on these observations, we concluded that mtsABC is a member of the ABC transporter systems.
Three patterns of signal peptide (Additional file 1 , Table S3) were used to identify bacterial lipoproteins from bioinformatics data [ 26 ]. To characterize the MtsA protein, the ScanProsite analysis was performed. The results indicated that the sequence of MtsA showed similarity to the following 2 patterns: residues D1 to D24 (MFKKISLAFAMLLSIFCITACSSQ) hit the G+LPPv2 pattern, and residues D17 to D21 (CITAC) hit the PS51257 pattern (Figure 3A ) [ 27 ], which suggested that MtsA is a lipoprotein. mtsA contains an lipoprotein peptidase cleavage site signal sequence as defined by Linton & Higgins [ 25 ]. To confirm that MtsA is a lipoprotein, the crude cell lysate of S. iniae HD-1 was mixed with Triton X-114, and the detergent phase was analyzed by western blotting using rabbit anti-MtsA antibodies (Figure 3B ). The results showed that MtsA protein was extracted by Triton X-114. Together, the results indicated that MtsA protein is a lipoprotein.
Purification of recombinant MtsA
To be able to further characterize MtsA, we first expressed recombinant MtsA consisting of amino acid residues D27 to D310 that lacked the putative signal sequence. Briefly, mtsA gene was cloned and the PCR product was isolated from the plasmid after a double digestion with restriction enzymes Bam HI and Xho I, and ligated into the compatible site of pET-32a-c (+) Vector to yield recombinant protein MtsA. The expressed MtsA had a molecular mass of 49.5-kDa (Figure 4 ) with a tag from Trx·Tag to EcoR V of pET-32a-c (+), which has a molecular weight of 17.7-kDa. The expression level of MtsA peaked after induction with 1 mM IPTG at 37°C for 4 h. The MtsA protein was purified from E. coli BL21 (DE3) under native condition n the soluble form and immunized the New Zealand white rabbits. The results showed that the rabbit anti-MtsA antibody titers increased from essentially zero to 1:50,000 after four rounds of immunization (Additional file 1 , Table S4). The western blotting analysis was performed to show the specificity of immunized sera against purified MtsA (Figure 4 , and Additional file 2 , Figure S3-4).
Subcellular localization of MtsA
To determine the subcellular localization of MtsA, the western blotting was performed with the cellular fractions of S. iniae HD-1 using rabbit anti-MtsA antibodies (Figure 5A ). MtsA was detected in the particulate fraction of the cells when the cellular fractions were prepared by centrifugation of the crude cell lysate (the first treatment). MtsA was found to be associated with the protoplast and cell wall extracts when the cellular fractions were prepared by protoplast formation. After separation of the protoplasts, MtsA was detected in the particulate fraction (the second treatment).
To detect surface exposure of MtsA, cells of S. iniae HD-1 cells were harvested, washed, centrifuged, and resuspended in PBS. The cells were subjected to proteinase K (5 μg ml -1 ) treatment with gentle agitation at room temperature for 1 h, and the cells were collected. Western blotting showed that peptide fragments in the cells can be detected after 1 h incubation with proteinase K. However, when the cell wall extracts and protoplasts were used in the experiment, it were completely hydrolyzed and no peptide fragments were detected (Figure 5B ). Together, this result indicated that MtsA is not exposure on surface, but is on the outside of the cytoplasmic membrane and is buried inside the cell wall.
MtsA had heme-binding activity
To examine whether heme is the chromophore associated with MtsA, the pyridine hemochrome assay was performed [ 28 ]. The UV-visible absorption spectrum of purified MtsA exhibited peaks at 275, 420, 525, and 560 nm, which were identical to those obtained from purified KatG, a well-known heme-containing protein with spectral peaks at 418, 524, and 556 nm. The molar ratio of associated heme to purified MtsA was 0.806 (Figure 6 ), this value is consistent with the hypothesis that one protein molecule is associated with one heme molecule.
To determine whether iron is present and its amount in purified MtsA, the ICP-AES analysis were performed. The results showed that Fe was present (Additional file 1 , Table S5) in purified MtsA; however, four other bivalent metallic elements Ca, Mg, Zn and Mn were not detected. The amount of iron present in purified MtsA (20 μM) was 1.43, 1.38, and 1.33 mg L -1 , in three independent purification experiments respectively.
In vivo production of MtsA during S. iniae HD-1 infection
To determine whether MtsA is produced in vivo during S. iniae infection, we infected Kunming mice with S. iniae HD-1 and performed western blotting analysis with purified MtsA to determine the presence of anti-MtsA antibodies in infected sera (Figure 7 ). The results indicated that MtsA is produced in vivo during experimental S. iniae HD-1 infection. | Discussion
Heme is an important nutrient for several bacteria and can serves as a source of essential iron. The most abundant source of iron in the body is heme, so it is not surprising to find that pathogenic bacteria can use heme as an iron source [ 29 ]. The presence of the central iron atom in heme allows it to undergo reversible oxidative change and act as a virulence-regulated determinant [ 30 - 36 ]. It is necessary for bacterial pathogens to acquire sufficient iron from their surroundings, and scavenging heme from the environment requires much less effort than synthesizing it de novo [ 30 , 34 ].
Acquiring iron from the micro-environment is important for the growth of bacterial pathogens. Pathogens often use low environmental iron levels as a signal to induce virulence genes [ 14 ]. Many pathogenic bacteria secrete exotoxins, proteases, and siderophores to rapidly increase the local concentration of free heme [ 37 ], and it is common for pathogens to directly acquire iron from host iron-binding proteins by using receptor-mediated transport systems specific for host-iron complexes [ 38 ].
To define the role of MtsA in heme utilization, the binding activity and subcellular localization of purified MtsA were investigated. The coding sequence of mtsA was cloned into the expression vector pet-32a-c (+). The major induced protein in E. coli (BL21) migrated as a 49.5-kDa band in SDS-PAGE gels; this size is consistent with the predicted molecular mass of MtsA. The 49.5-kDa MtsA was purified by Ni 2+ affinity chromatography and reacted with anti-MtsA antibodies from infected mice to confirm the in vivo production of MtsA. The UV-visible absorbance spectrum of KatG is a typical heme-containing protein, and the results of the pyridine hemochrome assay indicated that MtsA is associated with heme. Moreover, measurements of the iron level by ICP-AES indicated that purified MtsA is a holo-protein that is associated with iron.
In general, there are four major types of cell surface display proteins in Gram-positive bacteria, which are as follows: proteins anchored to the cytoplasmic membrane by hydrophobic transmembrane domains; lipoproteins that are covalently attached to membrane lipids after cleavage by signal peptides II; proteins that contain the C-terminal LPXTG-like motif and are covalently attached to peptidoglycan by sortase; and proteins that recognize some cell wall components by specific domains [ 39 ]. ABC transporters are integral membrane proteins that transport diverse substrates across lipid bilayers [ 40 ]. In bacteria, ABC transporters catalyze the uptake of essential nutrients or the extrusion of toxic substances [ 41 ]. ABC importers, present only in prokaryotes, require a binding protein that delivers the captured substrate to the external face of the transporter [ 42 ]. As MtsA is a solute-binding protein of the ABC transporter, its major function is presumed to be the capture and transfer of iron compounds to the downstream gene of the iron transport system of S. iniae HD-1. The signal peptide pattern analysis and Triton X-114 extraction results confirmed that MtsA is a lipoprotein. This result is reliable because the original G+LPP pattern was present in the analysis of the signal peptide features of 33 experimentally verified lipoproteins. Lipoproteins in Gram-positive bacteria are cell envelope proteins anchored to the outer leaflet of the plasma membrane. Lipid modification is achieved through covalent addition of a diacylglyceride to an indispensable cysteine residue in the lipoprotein signal peptide that provides a common anchoring mechanism for what is now recognized as an abundant and functionally diverse class of peripheral membrane proteins [ 40 ]. In Gram-positive bacteria, substrate-binding proteins of ABC transporters are typically lipoproteins [ 41 , 42 ], and the western blotting results is consistent with the notion that MtsA is an ABC transporter lipoprotein [ 43 , 44 ].
The results of this study indicated that mtsABC is a member of the ABC transporter family. MtsA protein is a solute-binding protein that can bind to heme and facilitate the latter's use as a substrate by the S. iniae . Western blotting indicated that MtsA is produced in vivo during experimental S. iniae HD-1 infection, and MtsA may be a potentially useful S. iniae protein vaccine candidate. Further studies are needed to clarify the role of MtsA in the utilization of other iron compounds and other important cations, and to establish its candidacy as a useful S. iniae vaccine component. | Conclusions
In summary, this study presents MtsA as a novel solute-binding protein that can contribute to iron transport. This is the first ABC transporter member to be identified from S. iniae . We have shown that MtsA is a lipoprotein which can bind to heme, and is expressed in vivo during Kunming mice infection by S. iniae HD-1. More importantly, this is the first report on the cloning of ABC transporter lipoprotein from S. iniae genomic DNA, and its immunogenicity is indicative of its possible use as an S. iniae subunit vaccine. | Background
Streptococcus iniae ( S. iniae ) is a major pathogen that causes considerable morbidity and mortality in cultured fish worldwide. The pathogen's ability to adapt to the host affects the extent of infection, hence understanding the mechanisms by which S. iniae overcomes physiological stresses during infection will help to identify potential virulence determinants of streptococcal infection. Grow S. iniae under iron-restricted conditions is one approach for identifying host-specific protein expression. Iron plays an important role in many biological processes but it has low solubility under physiological condition. Many microorganisms have been shown to be able to circumvent this nutritional limitation by forming direct contacts with iron-containing proteins through ATP-binding cassette (ABC) transporters. The ABC transporter superfamilies constitute many different systems that are widespread among living organisms with different functions, such as ligands translocation, mRNA translation, and DNA repair.
Results
An ABC transporter system, named as mtsABC (metal transport system) was cloned from S. iniae HD-1, and was found to be involved in heme utilization. mtsABC is cotranscribed by three downstream genes, i.e., mtsA , mtsB , and mtsC . In this study, we cloned the first gene of the mtsABC transporter system ( mtsA ), and purified the corresponding recombinant protein MtsA. The analysis indicated that MtsA is a putative lipoprotein which binds to heme that can serve as an iron source for the microorganism, and is expressed in vivo during Kunming mice infection by S. iniae HD-1.
Conclusions
This is believed to be the first report on the cloning the ABC transporter lipoprotein from S. iniae genomic DNA. Together, our data suggested that MtsA is associated with heme, and is expressed in vivo during Kunming mice infection by S. iniae HD-1 which indicated that it can be a potential candidate for S. iniae subunit vaccine. | Authors' contributions
LLZ carried out the molecular genetic studies, participated in the sequence alignment studies, performed the statistical analysis, and drafted the manuscript. JW carried out the function studies and participated in the sequence alignment studies. HBF carried out the infection assay. MQX conceived of the study and participated in its design and coordination. AXL participated in the conceived of the study and helped to draft the manuscript. All authors read and approved the final manuscript.
Supplementary Material | Acknowledgements
Project support was provided in parts by grants from Key Projects in the National Science & Technology Pillar Program in the Eleventh Five-year Plan Period (2007BAD29B05) to Dr. An-Xing Li. Project support was provided in parts by grants from Chongqing Engineering Technology Research Centre of Veterinary Drug (CSTC, 2009CB1010) to Dr. Lili Zou.
We thank Prof. Shaoping Weng and Drs. Lichao Huang, Xiangyun Wu, Yangsheng Wu, Jianfeng Yuan, and Suming Zhou for their helpful technical advice. We also thank Dr. Shenquan Liao for providing plasmid pet-32a-c (+) used in this study, and the professional copyediting service from the International Science Editing. | CC BY | no | 2022-01-12 15:21:37 | BMC Microbiol. 2010 Dec 1; 10:309 | oa_package/8f/9f/PMC3014919.tar.gz |
PMC3014920 | 21126367 | Background
The ciliated planktotrophic pilidium larva of nemertean worms (phylum Nemertea) is one of the most characteristic and unmistakable representatives of marine plankton. It is shaped like a helmet and equipped with an apical tuft of long cilia, and, typically, four lobes, spanned by the larval ciliated band [ 1 ]. The thin-walled and funnel-like pilidial esophagus (technically a vestibule) is separated from the thick-walled globose stomach by a sphincter (which could be considered as the pilidial mouth). The larval gut ends blindly. The larva swims with the apical organ pointing forward, and the larval antero-posterior axis corresponds to the animal-vegetal axis of the zygote. The pilidium larva, first described by Müller in 1847 [ 2 ], looks nothing like the juvenile nemertean worm, which Metschnikoff discovered in 1869 [ 3 ] develops inside the larva from a series of isolated rudiments called the imaginal discs. In a typical pilidium the antero-posterior axis of the juvenile is at a 90° angle with respect to the antero-posterior axis of the larva. The unusual development of the pilidium culminates in a catastrophic metamorphosis during which the juvenile worm emerges from and devours the larval body [ 4 , 5 ]. Although the term "metamorphosis" has been applied, particularly in the older literature, to the entire process of development of the juvenile worm inside the pilidium larva, I am using it in a more restrictive sense, referring to the end-point of pilidial development - i.e. the rapid transition between the pilidium and the juvenile.
The pilidium larva is found in a single clade called the Pilidiophora [ 6 ], which comprises roughly a third of all nemertean species [ 7 ], and represents a derived mode of development for the phylum [ 8 ]. Other nemerteans possess uniformly ciliated planula-like planktotrophic or lecithotrophic larvae, or direct development [ 1 , 9 , 10 ]. Pilidia belonging to different species vary in shape, size and color, but few have been matched to their respective adults [ 5 , 10 - 12 ]. Modifications of pilidial development include Desor's larva in Lineus viridis [ 13 ] and Schmidt's larva in Lineus ruber [ 14 ], Iwata's larva in Micrura akkeshiensis [ 15 ] and a few recently discovered others [[ 10 , 16 ], Megan Schwartz (University of Puget Sound) pers. comm.]. Development of Desor's and Schmidt's larva is encapsulated, while the others have planktonic lecithotrophic development. In all of these cases, the pilidial lobes are lacking, the blastocoel is reduced, the larva does not feed, but, similar to the canonical pilidium the juvenile develops via imaginal discs.
Although the development of the juvenile from imaginal discs is surely the most remarkable feature of the pilidium, this process remains poorly understood. The larval epidermis is very thin and transparent, making it easy to observe juvenile rudiments in live larvae. However, because the larvae are planktotrophic and the development of the juvenile takes weeks to months, the few studies of pilidial development are limited to descriptions of early stages, i.e. before onset of rudiment formation [ 12 , 17 - 20 ], or observations based on unidentified pilidium larvae opportunistically collected from plankton [ 21 - 23 ]. The only published report of pilidium larvae raised from fertilization to the point of formation of a nearly complete set of imaginal discs is that by Schmidt [ 24 ] for Cerebratulus marginatus (see [ 25 ] for German translation). Most of what we know about the development of the juvenile inside the pilidium comes from the 1912 study by Salensky [ 23 ].
On the other hand, the development of the juvenile inside the encapsulated Desor's "larva" has been studied in detail by many authors [ 13 , 26 - 29 ]. These embryos are opaque, so the observations are based on studies of preserved and sectioned material. Because this mode of development is derived for Pilidiophorans, it is not clear how much of what is known about the juvenile development inside Desor's larva is peculiar to it, and how much pertains to the pilidial development. For example, Desor's larva is described to have a total of eight imaginal discs, including a separate proboscis rudiment, whereas the typical pilidium is considered to have a total of seven discs, with proboscis developing as a derivative of the cephalic discs [[ 23 ] and references therein, and modern textbooks, e.g. [ 1 , 9 , 30 ]], although see Bürger [ 22 ] and Schmidt [ 24 , 25 ].
This paper describes the development of a typical pilidium larva in M. alaskensis from fertilization to metamorphosis, focussing particularly on the relative timing of developmental events, and the number and origin of the various juvenile rudiments. A staging scheme is proposed based on characteristic developmental milestones. Larval and juvenile serotonergic nervous systems are briefly described. The results are discussed in the light of previously published accounts of pilidial development. | Methods
Collecting adults
Micrura alaskensis (Heteronemertea; Lineidae) is a common intertidal pilidiophoran nemertean which inhabits sandy mudflats of the Northwest Pacific coast from Alaska to California (Roe et al 2007). The adults are typically about 5-10 cm long and 2-3 mm wide, light pinkish to darker salmon pink with lighter colored head (Figure 15 ). M. alaskensis lacks ocelli and possesses lateral longitudinal cephalic slits and a caudal cirrus, both easily observed under the dissecting microscope in specimens with intact anterior and posterior end. As in other heteronemerteans, the mouth opens ventrally some distance from the terminal proboscis pore. Adults are usually found in the top 10-20 cm of silty sand and co-occur with nemerteans Carinoma mutabilis and Cerebratulus spp., polychaetes (Fams. Maldanidae, Onuphidae, Oweniidae, Nereidae, Nephtidae, Lumbrinereidae etc.), phoronids ( Phoronopsis harmeri ) and a variety of sand-burrowing bivalves, such as Macoma nasuta , Clinocardium nuttallii and Tresus capax .
Reproductive adults of M. alaskensis were collected during negative low tides in False Bay (San Juan Island, Washington, USA) in July-August 2005, June-August 2006 and July 2007, and from several different mudflats in vicinity of Charleston, Oregon during the summer months of 2008-2010. Occasionally, ripe individuals were encountered outside the usual summer reproductive period, e.g. in February 2007 in False Bay and in March 2009 and March 2010 in Charleston, OR. Ripe males and females can be distinguished under the dissecting microscope, as testis and ovaries are visible through the body wall. Gametes were obtained by dissection. Several hundred to thousands of eggs can be obtained from a single female. Primary oocytes with distinct germinal vesicles dissected from the females or spawned naturally, undergo germinal vesicle breakdown soon upon contact with sea water, and can be fertilized by a dilute suspension of sperm obtained by dissection.
Larval culture
Developing embryos were cultured in 150 ml glass custard dishes in 0.45 μm-filtered seawater (FSW) until they started swimming. Swimming stages were transferred to larger volumes of FSW and cultured with continuous stirring using plexiglass paddles. Initial concentrations were approximately 1 larva per ml, and they were progressively thinned to about 1 larva per 4-10 ml over the course of several weeks of development. Water was changed by reverse filtration every three to four days, and after each water change larvae were fed either Rhodomonas lens at a concentration of 10 4 cells/ml, or a mixed diet of 5 × 10 3 cells/ml each of Rhodomonas lens and Isochrysis galbana .
Light microscopy
Live pilidium larvae were gently trapped between the slide and the coverglass supported by clay feet, and photographed using Nikon Coolpix 4500 digital camera mounted on a Nikon Eclipse 600 equipped with DIC optics or a Leica DCF 400 digital camera mounted on an Olympus BX51 microscope equipped with DIC. The timelapse video of metamorphosis was captured with side illumination through the Nikon Eclipse 600 microscope equipped with a Sony 3CCD video camera.
Fluorescent labeling and confocal microscopy
Larvae were relaxed in a 1:1 mixture of 0.33 M MgCl 2 and filtered seawater for 10-15 minutes at room temperature (RT) and fixed for 30 min at RT in 4% paraformaldehyde freshly prepared from 16% ultrapure paraformaldehyde (Electron Microscopy Sciences) and filtered seawater. Following fixation, larvae were either rinsed in three 10 min changes of phosphate buffer saline (PBS) and stored in PBS at 4°C, or immediately permeabilized and stained. For confocal microscopy larvae were permeabilized in three 10 min changes of phosphate buffered saline (PBS) with 0.1% Triton X-100 (PBT), stained with Bodipy FL Phallacidin (1 U per 100 μl of PBT) and a nuclear dye Hoechst 33342 (1 μM) for 40 min at RT, rinsed in three 10 min changes of PBS and mounted in Vectashield (Vector Laboratories) on Poly-L-lysine coated coverslips supported by foil tape over microscope slides. These semi-permanent preps were sealed with nail polish and viewed immediately or stored at 4 degrees C.
To visualize the serotonergic nervous system, larvae were fixed and permeabilized as described above, then blocked in 5% Normal Donkey Serum (Jackson Immunoresearch) in PBT for 2 hours at RT. After three 10 min washes in PBT, larvae were incubated for 2 hours at RT or overnight at 4 degrees C; with the rabbit-anti-5HT primary antibody (ImmunoStar Cat.# 20080) diluted 1:500 in PBT, washed in three 10 min changes of PBT and incubated for 2 hours at RT with Alexa Fluor 488 donkey-anti-rabbit secondary antibody (Molecular Probes) diluted 1:600 in PBT, adding 10 μg/ml of propidium iodide for the last 40 minutes of incubation to stain the nuclei. Stained larvae were washed in three 10 min changes of PBS and mounted as described above.
Stained samples were examined with a BioRad Radiance 2000 laser scanning confocal mounted on a Nikon Eclipse E800 microscope with a 40× 1.3 N.A. oil lens or an Olympus Fluoview 1000 laser scanning confocal mounted on an Olympus IX81 inverted microscope with a UPlanFLN 40× 1.3 NA oil lens or a PlanApoN 60× N.A. 1.42 oil lens. Stacks of 0.5-1 μm optical sections were imported into ImageJ v. 1.42i (Wayne Rasband, National Institutes of Health, Bethesda, MD, USA) for image processing. | Results
Developmental timeline
Using the method described here, I successfully reared to metamorphosis dozens of cohorts of pilidium larvae of Micrura alaskensis .
Dissected oocytes are frisbee-shaped and possess conspicuous germinal vesicles. They round up and undergo germinal vesicle breakdown within 30-40 min of contact with sea water. The round eggs are relatively opaque, 76 μm in diameter on average (n = 13), without a distinct egg envelope (Figure 1A ). The transparent layer of jelly surrounding the eggs is easily removed by pipetting or repeatedly passing the eggs through a nitex mesh with mesh size slightly larger than the egg diameter. As in other nemerteans, spiral cleavage is equal and holoblastic (Figure 1A-I ). At 11°C the first polar body is formed within the first 20-30 minutes (Figure 1B ), and the second polar body forms at about 1 hour after fertilization (Figure 1C ). The first cleavage occurs at about 2 hours 15 min after fertilization (Figure 1D-F ), and the subsequent divisions occur about once an hour. The initial cleavage furrow is broad (Figure 1D ) and the blastomeres separate almost completely (Figure 1E ), then re-compact (Figure 1F ). The third division is dextral (Figure 1H ) and, as in other nemerteans, the first-quartet micromeres are slightly larger than the macromeres in the 8-celled embryos (Figure 1I ).
The blastula of Micrura alaskensis is flattened along the animal-vegetal axis, and square-shaped, hence I refer to it as a "blastosquare" (Figure 2A-B, D ). The blastosquare has a small blastocoel (Figure 2B ). Blastosquares become ciliated and start swimming as early as 16 hours after fertilization at ambient sea temperature (11-14°C). By 24 hours of development swimming gastrulae are radially symmetrical, possess an apical tuft at the animal pole, and a prominent invagination at the vegetal pole (Figure 2C ). By the end of the second day of development the young pilidium larva exhibits bilateral symmetry, as the developing gut is inclined toward the dorsal side (Figure 2E-F ). The young helmet-like pilidium larva is equipped with a long apical tuft, two lateral lappets, and a band of longer cilia spanning the lappets. The gut of the pilidium larva is a blind sack differentiated into the funnel-like thin-walled esophagus (or vestibule) and the thick-walled stomach, with a muscle sphincter in between (Figure 2G-H ). The posterior wall of the esophagus (with respect to the axis of the future juvenile worm) has a pair of ciliated ridges (Figure 3A ). The pilidia are capable of feeding on unicellular algae as early as 66 hours after fertilization. The stomachs of pilidia feeding on Rhodomonas lens gradually accumulate brown-magenta pigment (Figure 3 ). As the pilidia grow the lateral lappets become more and more pronounced, and two additional lobes develop, which are usually referred to as the anterior and the posterior (with respect to the juvenile-to-be) (Figure 3C ).
The rate of development depends on the temperature, feeding regime, and, possibly, other factors. Development of M. alaskensis in culture is somewhat asynchronous, even among full siblings raised in the same container. However, the juvenile rudiments always appear in the same sequence: first a pair of cephalic discs (Figure 3B ), then a pair of trunk discs (Figure 3C ), followed by the unpaired proboscis rudiment (Figure 3D ), a pair of cerebral organ discs (Figure 3E ), and the unpaired dorsal rudiment (Figure 3F ). I observed two pairs of imaginal discs, the cephalic and the trunk discs, as early as 9 days after fertilization, although, more typically, at 14-15 days. The proboscis rudiment appears at about the same time as the cerebral organ discs, as early as 14 days of development, but more typically around 20 days.
The proboscis rudiment fuses with the cephalic discs, forming the head rudiment, and the cerebral organ discs fuse with the trunk discs (Figure 3G ). About the same time, a small dorsal rudiment appears just below the stomach and dorsal to the trunk discs (Figure 3F ). It is initially clearly separate, and later fuses with the trunk discs contributing to the trunk rudiment. Subsequently, the head rudiment fuses with the trunk rudiment on both sides forming a toroid juvenile rudiment (Figure 3H ). This "torus" stage was observed as early at 28 days after fertilization, although development to this point may take 5 weeks or longer. The proboscis rudiment elongates and reaches beyond the margin of the juvenile head, stretching out along the dorsal side of larval esophagus toward the stomach (Figures 3I , 4A ). The dorso-posterior wall of the juvenile rudiment gradually grows over the larval stomach and overlaps the proboscis. At this stage the juvenile rudiment most resembles a slipper shoe with a hole in the middle of its "sole" (Figure 4B ). The edge of the dorsal wall continues to grow anteriorly, covering up the proboscis, and eventually it fuses with the margin of the head rudiment, thus closing the dorsal gap and completing the juvenile (Figure 4C ).
Metamorphosis of cultured pilidia of M. alaskensis occurred as early as 35 days after fertilization, but more typically at 6-7 weeks. Pilidium larvae with complete juveniles readily underwent metamorphosis in culturing containers, and even on microscope slides, trapped under coverglass while being filmed (Figure 5 , Additional file 1 -- Movie 1). In the vast majority of cases, the emerging juvenile devoured the entire larval body (Figure 5C , Additional file 1 -- Movie 1).
Because development of pilidia in culture is typically asynchronous, it makes little sense to refer to the larval stages by absolute age. To facilitate future studies of pilidial development in this and other nemertean species, I propose a staging scheme based on the presence of certain developmental milestones (Table 1 ).
Development of the juvenile rudiments
Larvae of Micrura alaskensis have three pairs of imaginal discs. The cephalic discs develop first as invaginations of pilidial epidermis just apical of the ciliated band at the junction of pilidial lateral lappets and the anterior lobe - one on each side (Figure 6 , Additional file 2 -- Movie 2). The trunk discs develop next as invaginations of the pilidial subumbrellar epidermis (i.e. below the ciliated band), also in the immediate proximity to the larval ciliated band, near the junction of pilidial lateral lappets and the posterior lobe (Figure 7 , Additional file 3 -- Movie 3). The last pair, the cerebral organ discs, appear as invaginations of the subumbrellar epidermis on the inner side of lateral lappets in close proximity to the esophageal ciliated ridges and the trunk discs (Figure 8 , Additional files 4 and 5 -- Movies 4 and 5).
Once the invaginated rudiments of cephalic and trunk discs separate from the larval epidermis, they consist of two layers separated by a narrow cavity, like a flattened balloon. The thick inner layer gives rise to the juvenile tissues, including the definitive epidermis, and the thin outer layer composed of a single layer of squamous cells gives rise to the amniotic sack, or amnion. The amnion produced by the fused thin outer walls of cephalic and trunk imaginal discs encloses the juvenile inside the pilidium larva and separates it from the blastocoel. In Micrura alaskensis the amnion is characteristically decorated with black and brown pigment granules (Figures 3E, G-I , 4A, C, F-H , 5 ), which become apparent even before the imaginal discs fuse, and are evident in metamorphosing larvae. The black pigment granules are concentrated along the ventral side of the amnion, while the brown pigment granules form a polka-dot pattern on the lateral sides.
In addition to the six imaginal discs described above larvae of Micrura alaskensis possess two unpaired rudiments -- the proboscis rudiment and the dorsal rudiment. These do not form as epidermal invaginations, but appear to be mesenchymal in origin. The proboscis rudiment is first evident as a distinct small cluster of few cells in the plane of bilateral symmetry of the larva -- between the pilidial epidermis of the anterior lobe and the two cephalic discs (Figures 8A', D , 9A, C ). All three pairs of imaginal discs are present at this stage (Figures 8 , 9B ). Subsequently, the proboscis rudiment fuses with the two cephalic discs (Figure 9D ), and forms a distinctly bilayered proboscis bud inside the head rudiment at the "head and trunk" stage (Figure 9E ). The proboscis bud elongates (Figure 9F ) and protrudes beyond the margin of the head rudiment at "extended proboscis" stage (Figure 9G ). The proboscis follows along the larval esophagus in the larval mid-plane and reaches the larval stomach at "complete proboscis" stage (Figure 9H ).
The dorsal rudiment, like the proboscis rudiment, is first evident as a cluster of what seems to be few mesenchymal cells in the plane of bilateral symmetry, at the diametrically opposite end from the proboscis. It is sandwiched between the epidermis of the posterior lobe and the stomach (Figure 3F ). It subsequently fuses with the trunk discs to form the trunk rudiment (Figure 3G ).
Larval and juvenile serotonergic nervous system
Serotonergic neurons were observed in the young pilidium larva as early as 40 hours after fertilization, even before onset of feeding. Two groups of neurons can be distinguished -- several apical neurons, and several associated with the developing ciliated band (Figure 10A ). The serotonergic nervous system of a fully developed pilidium larva includes numerous serotonergic neurons connected by the marginal ciliary nerve (terminology after Lacalli and West [ 31 ]) associated with the ciliated band, as well as numerous neurons that form an extensive subepidermal nerve net (Figure 10B ). Two monociliated serotonergic neurons are always found associated with the apical plate, one on each side (Figure 10B-C ). In addition, one can distinguish the circular oral nerve [ 31 ] in the sphincter between the larval esophagus and the stomach (Figure 11 ).
The juvenile inside the pilidium possesses a separate serotonergic nervous system, which becomes evident after torus stage. As in adult nemerteans, the most prominent component of the juvenile nervous system includes the two lateral nerve cords that originate from the cerebral ganglia and are connected anteriorly via dorsal and ventral brain commissures and posteriorly in the juvenile caudal cirrus (Figure 11 ). In addition, one can distinguish the juvenile subepidermal nerve network, and a number of longitudinal nerves in the proboscis (Figure 11 ). The only component of the larval serotonergic nervous system that appears to be incorporated into the juvenile is the oral nerve.
In addition to serotonergic neurons and fibers, I was able to visualize the fibrous core (axons) or the major components of the juvenile nervous system, such as the cerebral ganglia, the lateral nerve cords, dorsal and ventral commissures of the brain, and the two cerebral organ nerves, using phalloidin labeling (Figure 12A-C , Additional files 6 and 7 -- Movies 6 and 7). From this it is clear that the serotonergic neurons comprise only a small fraction of the juvenile nervous system. | Discussion
Development of the paired imaginal discs
The results of this study confirm classical reports that the typical pilidium larva possesses three pairs of imaginal discs which develop as ectodermal invaginations -- the cephalic discs, the trunk discs, and the cerebral organ discs. This study disagrees with the only other detailed report on the order of appearance and the origin of these discs in pilidium larvae [ 23 ], and confirms Schmidt's report for Cerebratulus marginatus [ 24 , 25 ]. Salensky, who relied solely on planktonic samples reported that all three pairs of imaginal discs appear simultaneously [[ 23 ]: p. 21]. Schmidt [ 24 , 25 ] reared pilidia of C. marginatus from fertilization to the point of formation of almost all of the juvenile rudiments and reported that the discs appear in the following sequence: cephalic discs first, followed by the trunk discs, and finally, the cerebral organ discs. The present study shows that in M. alaskensis , the appearance of the paired imaginal discs follows the same order: the cephalic discs appear after about one week of development, followed by the trunk discs at two weeks, and the cerebral organ discs at about three weeks of development at ambient sea temperature.
My unpublished observations of development in another Northwest Pacific coast species Cerebratulus cf. marginatus (which is likely not the same species as Schmidt's C. marginatus from the Gulf of Naples, Italy), as well as in numerous pilidia belonging to different species captured in plankton in coastal waters of Washington and Oregon suggest that they all follow the same sequence of appearance of imaginal discs. Although the development of the juvenile was not the focus of his study, Cantell's observations of pilidium larvae of Lineus albocinctus, Lineus bilineatus, Micrura purpurea , and several unidentified pilidial morphotypes collected from plankton [ 12 ] suggest the same order of imaginal disc formation as described by Schmidt [ 24 , 25 ] and in this study. It is easy to see, however, that if one is relying solely on the larvae obtained from plankton, as opposed to rearing them in the lab, one might be misled into thinking that all three pairs of imaginal discs appear simultaneously, if intermediate stages of development did not present themselves to the collector.
According to Salensky [ 23 ], all three pairs of imaginal discs develop as invaginations of subumbrellar pilidial epidermis, i.e. below the larval ciliated band, or in other words -- from the hyposphere. I document here for M. alaskensis that while the trunk discs and the cerebral organ discs indeed invaginate from the subumbrellar epidermis, the cephalic discs are derived from the umbrellar epidermis, i.e. above the larval ciliated band, or from the episphere (Figure 13 ). This particular observation is in agreement with the known cell lineage of the cephalic discs, which are derived from the first-quartet micromeres in pilidium larvae of Cerebratulus lacteus [[ 32 ], reviewed in [ 33 ]]. The cell lineage of the trunk discs and the cerebral organ discs remains to be determined using long-term lineage markers, but their position suggests that they are produced by the progeny of the second-quartet or the third-quartet micromeres.
The proboscis rudiment in pilidial development
According to textbooks, the juvenile inside the pilidium larva develops from seven imaginal discs: the paired cephalic discs, paired trunk discs, paired cerebral organ discs, and an unpaired dorsal disc [ 1 , 9 , 30 ]. This is largely based on the most detailed published study of development of the juvenile inside the pilidium larva by Salensky [ 23 ]. Eight imaginal discs, including a separate proboscis rudiment are described in development of Desor's larva [ 26 - 29 ]. According to Salensky [ 23 ], the proboscis of the pilidium larva develops from the fused cephalic discs, and a separate proboscis rudiment is lacking. Interestingly, Bürger [ 22 ] and Schmidt [ 24 , 25 ] described a separate proboscis rudiment in development of the pilidium larva, but this observation was not confirmed by Salensky [ 23 ] and is ignored in the more recent literature.
The results of the present study clearly show that in the pilidium larva of Micrura alaskensis there are eight separate juvenile rudiments, including an unpaired proboscis rudiment, which confirms earlier reports by Bürger for an unidentified species [ 22 ] and Schmidt for Cerebratulus marginatus [ 24 , 25 ]. Because the proboscis rudiment is relatively small and is evident as a separate entity from the cephalic discs only for a short period of time, it is possible that Salensky and other investigators relying as they did on planktonic samples instead of culturing larvae, did not have the opportunity to observe and confirm its presence.
The proboscis rudiment in M. alaskensis does not develop as an invagination, contrary to Bürger's report [ 22 ]. It is not clear whether it arises by delamination from the pilidial epidermis, or from a cluster of mesodermal cells associated with the pilidial epidermis. In the first case the proboscis rudiment would be derived from the first-quartet micromeres. In the latter it would most likely originate from the 4d cell, the spiralian mesentoblast. The 4d cell gives rise to a population of scattered mesenchyme cells in the pilidium of Cerebratulus lacteus [ 32 ], including a small cluster of cells in vicinity of the cephalic discs, i.e. where proboscis rudiment would be developing later (Figure 7d in [ 32 ]). It appears very likely that the bilayered proboscis bud described here (Figure 6E ) has a dual origin, so that its inner part ("the arm") is derived from the fused cephalic discs, as proposed by Salensky [ 23 ], while the outer part ("the sleeve") from the separate proboscis rudiment described here. Conceivably, the separate rudiment is derived from the 4d cell and gives rise to the muscle layers of the proboscis and the rhynchocoel, while the portion derived from the cephalic discs gives rise to the glandular epidermis of the proboscis. This would make sense in the context of a now widely accepted hypothesis that the nemertean rhynchocoel is homologous to the annelid and mollusk coeloms [ 34 ], which are also derived from the 4d cell. A study utilizing long-lasting cell lineage markers is necessary to determine which is the case.
Another piece of evidence that the proboscis is not derived solely from the cephalic discs, as suggested by Salensky [ 23 ], comes from an incidental observation of development of one particular cohort of pilidium larvae reared in the lab, which exhibited numerous abnormalities in juvenile, but not larval, development [Maslakova unpublished]. In many of these pilidia one or several juvenile rudiments were missing, while others appeared to be unaffected. Some larvae had all of the discs, including both of the cephalic discs, but were missing the proboscis. This suggests that even if cephalic discs normally contribute to the proboscis, it is possible that they may need some sort of inductive signal from the unpaired proboscis rudiment in order to do so. It would be possible to test whether the anterior unpaired rudiment observed in this study is essential to the formation of the proboscis by experimentally destroying it (e.g. by laser ablation) in otherwise normally developing pilidium larvae.
Origin of the dorsal rudiment
Although textbooks often describe all of the juvenile rudiments as invaginations of larval epidermis, the results of this study suggest that only the paired imaginal discs develop as distinct invaginations, whereas the proboscis rudiment and the dorsal rudiment do not. Salensky [ 23 ] reported that the unpaired dorsal disc develops via delamination from the pilidial dorsal epidermis. The present study is inconclusive as to whether the dorsal disc is of mesodermal origin (i.e. likely derived from the 4d cell) or is in fact an epidermal derivative (i.e. likely derived from one of the first-quartet micromeres).
Nephridial rudiments and the juvenile foregut
Several earlier studies describe a pair of esophageal pouches during juvenile development inside Desor's larva [ 26 ] and pilidium larvae [ 22 , 23 ] as rudiments of the nephridia (although Nusbaum and Oxner [ 27 ] disagree with this interpretation). Bürger [ 22 ] reported that the two pouches completely separate from the esophageal cavity of the pilidium larva, fuse with the trunk discs, then branch like fingers of a glove, and that the nephridiopores develop after metamorphosis. Salensky [ 23 ] reported two pouches in a similar position (between the cerebral organ discs and the larval stomach), but interpreted them as being derived from the subumbrellar epidermis near esophagus, rather than the esophageal wall. He insisted on this particular distinction, because he believed that the pilidial esophagus is of endodermal origin, and the nephridia must be derived from the ectoderm. Histologically, however, there is no difference between the subumbrellar epidermis and esophageal epidermis; one gradually transitions into the other. Both are composed of squamous cells, and both are likely ingested by the juvenile during pilidial metamorphosis. Moreover, cell lineage analysis in the pilidium larva of Cerebratulus lacteus [ 32 ] shows that the larval esophagus is derived from the second-quartet and third-quartet micromeres, i.e. is of ectodermal origin. More importantly, however, Salensky [ 23 ] did not observe these pouches closing off from the esophagus, even in the latest stages of juvenile formation inside the pilidium larva, nor did he observe any branching.
I observed a pair of thick-walled rounded esophageal pouches sandwiched between the cerebral organs and the larval stomach in advanced developmental stages (e.g. "complete proboscis" and "hood") of M. alaskensis (Figure 12 , Additional files 6 and 7 -- Movies 6 and 7). Similar to Salensky [ 23 ], I did not observe these pouches closing off from the esophagus, or their distal ends branching. Judging from their position and morphology, it seems unlikely that they represent rudiments of nephridia (see also [ 27 ]). A more likely explanation is that they form the juvenile foregut, while the larval stomach gives rise to the juvenile midgut. In adult pilidiophorans, these two regions of the digestive tract are well differentiated histologically, and the transition between the two is very distinct. The adult foregut comprises a muscular, thick-walled tube of densely ciliated, glandular, and typically deeply folded epithelium. It appears unlikely that it could be derived from the larval esophagus, or the larval stomach, as previously suggested ([ 23 ] and references therein). I observed a distinctly bipartite digestive tract in newly metamorphosed juveniles of M. alaskensis with a thick-walled foregut positioned between the cerebral organs and the midgut, where the paired pouches used to be (data not shown). Upon request from one of the reviewers I am including a diagram, which shows relative position of the juvenile foregut rudiments to the various parts of the larval digestive system (Figure 14 ).
At the same time, I have also observed what appears to be a pair of protonephridia -- one on each side of the developing juvenile, in the immediate vicinity of the paired esophageal pouches described above, but not physically connected to them. These organs, provisionally referred to as the nephridia, are sandwiched between the cerebral organs and the stomach on two sides, and the esophageal pouches and the juvenile body wall on two other sides. Each nephridium has a branched distal portion, and at least one efferent duct leading through the body wall to the nephridiopore on the lateral side of the developing juvenile in advanced stages of M. alaskensis (Figure 12 , Additional file 7 ). The position of these rudiments corresponds to that of protonephridia in adult nemerteans (restricted to the foregut region in most species). Because the nephridial rudiments are located in close proximity to the foregut rudiments, and because Bürger's study [ 22 ] was based on regular histological sections which are typically 7-8 μm thick (compared to 1-μm-thick confocal sections in this study), he might have been mislead into thinking that a) the nephridial rudiments are connected to the esophageal pouches, b) these pouches are closed off from the esophagus, and c) the nephridial openings develop only after metamorphosis. Future studies utilizing TEM of ultra-thin sections of advanced developmental stages of pilidium larvae should be able to confirm the nature of these provisional nephridia.
Larval and juvenile nervous systems
The results of this study largely confirm previous reports on the structure of the pilidial larval and juvenile nervous systems based on classical histological methods [ 22 , 23 ], TEM [ 31 ] and immunohistochemistry and fluorescent microscopy [ 35 ].
Salensky [ 23 ] believed that the central nervous system of the juvenile (cerebral ganglia and lateral nerve cords) is derived from the cephalic imaginal discs, that there are no rudiments of the nervous system in the trunk discs, and that the lateral nerve cords invade the tissue of the trunk discs after fusion with the cephalic discs. Bürger [ 22 ] believed that the central nervous system is derived from both the cephalic imaginal discs and the trunk discs and that the juvenile nervous system develops at about the time when the cephalic and the trunk discs fuse with each other. According to Bürger [ 22 ] the cephalic discs give rise to the dorsal cerebral ganglia, while the trunk discs give rise to the ventral cerebral ganglia and the lateral nerve cords. The results of this study cannot confirm or disconfirm either hypothesis. It does appear very likely that at least some portion of the cerebral ganglia originates from the cephalic discs. However, a long-term lineage tracing is necessary to determine with confidence which imaginal discs give rise to various parts of the juvenile nervous system. | Conclusions
This study represents the first report documenting pilidial larval development from fertilization to metamorphosis. It is based on observations of laboratory cultures of the common Northwest Pacific coast species Micrura alaskensis . Larvae typically reach metamorphosis after 5-8 weeks of development at ambient sea temperature. During metamorphosis the juvenile worm escapes from the larval enclosure and routinely devours the entire pilidial body. The larval and juvenile serotonergic nervous system is for the first time illustrated using confocal microscopy.
Larval development is asynchronous in culture, so a staging scheme is proposed based on certain developmental milestones to facilitate comparison with other species. Characteristic developmental stages include, in that order: blastosquare, gastrula, young pilidium, feeding pilidium, cephalic discs, trunk discs, cerebral organ discs, head and trunk, torus, extended proboscis, complete proboscis, hood, metamorphosis. The order of appearance of juvenile rudiments is always the same and appears to be conserved between different pilidiophoran species.
The paired cephalic discs, paired trunk discs, and paired cerebral organ discs develop as invaginations of the larval epidermis. The cephalic discs invaginate from the larval episphere, while the trunk discs and the cerebral organ discs invaginate from the hyposphere. A separate unpaired dorsal rudiment previously described for pilidium larvae is also found in M. alaskensis , and does not develop as an invagination. The rudiments of juvenile nephridia and foregut, which have been described but confused with each other in the earlier literature and ignored by the modern literature are described and illustrated using confocal microscopy. One of the most significant findings of this study is the separate proboscis rudiment, which brings the total number of juvenile rudiments inside the pilidium larva from seven, as described in textbooks, to eight. The unpaired proboscis rudiment does not develop as an invagination, and may be of mesodermal, rather than ectodermal origin. | Background
The nemertean pilidium is one of the most notable planktotrophic larval types among marine invertebrates. The juvenile forms inside the larva from a series of isolated rudiments, called the imaginal discs. The development culminates in catastrophic metamorphosis, in which the larval body is consumed by the juvenile worm. Although the pilidium was first described in 1847, and is commonly found among marine plankton, there is not a single complete description of its development. The few published studies of pilidial development are based on observations of typically unidentified larvae opportunistically collected from plankton at various developmental stages.
Results
The development of Micrura alaskensis , a common Northwest Pacific coast intertidal nemertean, is described from fertilization to metamorphosis. A staging scheme is proposed based on characteristic developmental milestones. Three pairs of imaginal discs develop as invaginations of larval epidermis. The cephalic discs invaginate from the larval epidermis above the ciliated band, while the cerebral organ discs and the trunk discs invaginate below the ciliated band. All paired imaginal disc invaginations are closely associated with different portions of the larval ciliated band. In addition, two unpaired rudiments contribute to the juvenile - the proboscis rudiment and the dorsal rudiment, which do not develop as invaginations. A pair of thick-walled esophageal pouches previously thought to represent nephridial rudiments give rise to the juvenile foregut. Branched rudiments of protonephridia, and their efferent ducts are also described. Larval and juvenile serotonergic nervous systems are briefly described. Development of the juvenile is completed by 5-8 weeks at 11-15 degrees C. During the rapid metamorphosis the juvenile emerges from and devours the larva.
Conclusions
This study is the first description of pilidial development from fertilization to metamorphosis in a single species. It is illustrated with photomicrographs of live larvae, diagrams, confocal images, and videos. The findings are discussed in the context of previously published accounts of pilidial development, with which they disagree on several accounts. The results described here indicate a different number, origin and fate of various juvenile rudiments. The proposed staging scheme will be useful in subsequent studies of pilidial development. | Competing interests
The author declares that they have no competing interests.
Supplementary Material | Acknowledgements
A large part of this study was carried out at the University of Washington's Friday Harbor Laboratories, while I was supported by the FHL Postdoctoral Fellowship. I thank the FHL researchers and staff, and particularly Dr. Richard Strathmann, for pointing me to M. alaskensis , as a suitable species for long-term laboratory culture, and Dr. George von Dassow for help with videomicroscopy and image processing and comments on the manuscript. Ms. Laurel Hiebert assisted in rearing pilidium larvae at the Oregon Institute of Marine Biology (University of Oregon). | CC BY | no | 2022-01-12 15:21:37 | Front Zool. 2010 Dec 2; 7:30 | oa_package/f2/38/PMC3014920.tar.gz |
PMC3014921 | 21106063 | Background
Endometrial carcinoma (EC) is the most common gynecological malignancy observed in the western society with an incidence rate of approximately 15-20 per 100.000 women per year. Endometrioid adenocarcinoma (EAC) arises from cells that form the glands in the endometrium. It is the most prevalent subtype constituting approximately 80% of all endometrial cancers. Endometroid andenocarcinoma can be histologically graded according to the FIGO system (International Federation of Gynecology and Obstretics) or classified as low grade or high grade by an alternative architectural binary grading system. The disease may either be an estrogen-dependent low-grade endometroid variant (type I) or a non-estrogen-dependent high grade variant (type II). As expected, estrogen-dependent endometroid carcinomas preferentially affect women in the pre- or peri-menopausal phase, whereas type II EACs usually develops in older post-menopausal women. Type II tumors are typically of high-grade endometroid adenocarcinomas, papillary serous or clear cell types, and generally carry a poor prognosis [ 1 - 5 ].
Due to the complexity of cancer etiology caused by the genetic heterogeneity present in the human population and the influences of environmental factors, it can be advantageous to turn to inbred animal models. In the present and previous works, we have used tumor material from crosses including the BDII/Han rat model, where more than 90% of the female virgins of the BDII inbred rat strain spontaneously develop endometrial cancer during their life time. These tumors are hormone dependent ECs and thus represent spontaneous hormonal carcinogenesis [ 6 - 8 ].
In a recent comprehensive microarray study of endometrial cancer cell lines, we found that the expressions of 354 genes were significantly altered relative to normal/pre-malignant endometrium. When applying traditional statistical analyses and gene classification analysis on the microarray data (Waikato environment of knowledge analysis, Weka), we could identify a three-gene signature ( Gpx3 , Bgn and Tgfb3 ), that might have important implications in EAC carcinogenesis [ 9 ]. It was also revealed that Gpx3 displayed the most significantly altered gene expression in comparisons among endometrial tumors and normal/pre-malignant endometrium.
GPX3 constitutes the backbone of the cellular antioxidant defense system, together with Superoxide Dismutase (SOD) and Catalase (CAT)[ 10 ]. GPX3 catalyses the reduction of peroxides and protects cells against oxidative damage. The decreased mRNA expression of Gpx3 in rat endometrial tumors might result in an impaired defense against endogenous and exogenous genotoxic compounds, which could potentially lead to an increased mutation rate including genes involved in carcinogenesis. The gene expression of GPX3 in human has previously been shown to be silenced in prostate cancer, ovarian clear cell adenocarcinoma, gastric carcinoma and in Barret's disease by epigenetic mechanisms, such as hyper-methylation [ 11 - 16 ]. Furthermore, Yu, et al. [ 17 - 23 ], suggested that GPX3 also contains tumor suppressor activity by, directly or in-directly, regulating cell growth and proliferation through unknown mechanisms. GPX3 may influence the expression of MET (mesenchymal-epithelial transition factor) that encodes a tyrosine kinase receptor for hepatocyte growth factor (HGF). Abnormal expression/activation of the MET receptor has been reported in numerous human cancer diseases [ 17 - 23 ].
The aim of this study was to investigate the mechanisms underlying the down-regulation of Gpx3 and potential implications in EAC carcinogenesis. The production of hydrogen peroxide in the endometrial tumors displaying loss of Gpx3 expression as well as in the endometrial samples with normal/high expression was measured. Since GPX3 has been suggested to exhibit tumor suppressor activity, that could regulate the transcription of the oncogene, Met , we estimated the correlation of the expression between Met and Gpx3 by employing a real time RT-PCR expression study. In order to verify the results from the BDII endometrial cancer rat model in human, we also examined the mRNA expression of GPX3 and MET in human endometroid tumor samples of FIGO grade I, II and III and benign samples. | Materials and methods
Animal crosses and tumor material
The animal material was derived from crosses between BDII/Han females and males from two non-susceptible rat strains, BN/Han and SPRD Cu3 /Han, where at first an F1 progeny was produced. Subsequently, an F2 offspring by brother/sister mating of the F1 progeny, and a backcross progeny (N1), by crossing the F1 males to BDII females, were produced. The female progeny was palpated twice each week for identification of uterine tumors. Animals suspected to have tumors were euthanized and the tumor tissue surgically removed, subjected to pathological characterization and subsequently used to establishment of cell cultures [ 29 ]. In this study we have investigated cell lines established from tissues pathologically classified as endometrial adenocarcinomas (EAC) and from tissues of normal/pre-malignant endometrium (NME). RUT cell lines originate from tumors developed in the F1 and F2 progeny and NUT cell lines originate from the tumors in the backcross progeny (Table 1 ).
In vitro cell culture conditions
Primary cell cultures established from the EAC tumors were propagated in Dulbecco's modified Eagle medium, supplemented with 100 IU/100 μg/ml penicillin/streptomycin, L-glutamine, MEM amino acids, MEM Non Essential Amino acids, MEM Vitamins solution and 10% heat-inactivated fetal bovine serum, for 3-5 passages in order to obtain the required amount of cells. The NME cell lines were cultured under the same conditions, but in medium containing 20% fetal bovine serum. The cells were grown at 37°C in an atmosphere of 95% humidity and 5% CO 2 and harvested by trypsinizination at a confluence of 80-90% (approx 1x10 6 cells).
RNA extraction
Total RNA was extracted from the harvested cells of the different endometrial rat cell lines with a KingFisher mL Instrument (Thermo Electron Corporation, USA) according to the manufacturer's protocol (MagAttract Tissue Mini M48 Kit, Qiagen). RNA was spectrophotometrically quantified (NanoDrop technologies, USA).
Quantitative PCR (qPCR) of Gpx3 and Met in cell lines from adenocarcinomas (EAC) and normal/premalignant tissue (NME)
A total of 14 EAC and 2 NME cell lines were used in qPCR analysis with GAPDH as an endogenous control and Universal Rat Reference RNA, Agilent Technologies, Inc as a calibrator (Table 1 ). RT-PCR was performed using High Capacity cDNA Reverse Transcription Kit according to the manufacturer's protocol (Applied Bio-systems). Template cDNA was added to TaqMan Universal Master Mix (AB; Applied Biosystems, Foster City, CA, USA) in a 12.5 μl reaction with specific pre-designed probes for the Gpx3 and Met (Applied Biosystems). Reactions were performed in duplicates and threshold cycle number was averaged. Relative gene expression quantification was calculated according to the comparative Ct method using GAPDH as an endogenous control and Universal Rat reference RNA (Stratagene) as calibrator. The relative quantitative gene expression, were determined as follows: 2-(Ct sample-Ct calibrator), where Ct values of the calibrator and sample are determined by subtracting the Ct value of the target gene from the value of the GAPDH gene.
Bisulfite treatment and methylation-specific PCR
One μg DNA from the susceptible rat strain BDII, 15 EAC and 2 NME cell lines were denatured, sodium bisulfite treated and purified using Epitect Bisulfite Kit according to the manufacturer's protocol (Qiagen) (Table 1 ). As a positive control, DNA from an endometrial cell line (NUT43) with a normal expression of Gpx3 , and thus unmethylated was treated with methylase and subsequently treated with sodium bisulfite. The modified tumor/control DNA was used as template for methylation-specific PCR. Methylation specific primers (MSP) were designed using the publicly available MethPrimer program http://www.urogene.org/methprimer/ [ 30 ]. The bisulphite modification of DNA converts unmethylated cytosines to uracils, whereas methylated cytosines will remain unchanged. Bisulphite treated DNA was amplified with either methylation specific or un-methylation specific primer sets. PCR was carried out in a final 25 μl volume containing 50 ng of template DNA The mixture was heated at 94°C for 1 min and then subjected to 35 cycles of 94°C, 55°C and 72°C and a final extension at 72°C for 7 min. The PCR product was analyzed on a 2% agarose gel with appropriate size marker and the absence or presence of PCR product were detected.
5-aza-2 ́-deoxycytidine and trichostatin A (TSA) treatment in rat endometrial tumor cell lines
Two EAC cell lines (NUT12 and NUT81) with confirmed biallelic methylated promoter status of Gpx3 were treated with 5-aza-2 ́-deoxycytidine and trichostatin A (Sigma). Cells were grown in a medium containing 2.5 μM 5-aza-cytidine for 96 hours, with the medium and drug being replaced every 24 hours and the addition of 300 nM TSA was added for the last 16 hours. After 96 hours, the drugs were removed and total RNA for Gpx3 RT-PCR expression analysis was extracted using AllPrep RNA/DNA Mini Kit according to the manufacturer's protocol (Qiagen).
Hydrogen peroxide measurements
The intra- and extra-cellular amount of hydrogen peroxide in the endometrial cell lines were measured using the Amplex ® Red Peroxide/Peroxidase Hydrogen assay kit according to the manufacturer's protocol (Molecular Probes, Invitrogen). The three cell lines investigated (NUT12, NUT43, NUT56) were seeded to a 96 wells plate (Corning) with an initial number of 5000 cells/well (Table 1 ). In each well, the amount of hydrogen peroxide was measured post 72 hours of incubation, as described above. Each cell line was replicated 20 times. The cells were lysated by adding RIPA buffer (25 mM TRIS-HCl pH 7.6, 150 mM NaCl, 1% deoxycholate, 0, 1% SDS and 1% NP40) followed by incubation on ice for 30 minutes. Intra- and extra-cellular H 2 O 2 concentrations were assessed by pooling 50 μl of cell lysate with 50 μl of used cell culture media.
Development of FISH probes for Gpx3
DNA from the pre-malignant cell line, NUT43, with a normal expression of Gpx3 was used to generate a probe that represented only the Gpx3 gene. Six primer pairs, specific for the Gpx3 gene, with a product size of approximately 600 base pair each (Table 5 ), were designed by using the Primer 3 program available on the internet: http://fokker.wi.mit.edu/primer3/input.htm . Amplification was performed by PCR and carried out in a final 25 μl volume containing 100 ng of template DNA. The mixture was heated at 94°C for 1 min and then subjected to 35 cycles of 94°C, 58°C and 72°C and a final extension at 72°C for 7 min. Sizes of the PCR products were determined on a 2% agarose gel with an appropriate size marker. The PCR products were then purified by Mini Elute PCR purification kit (QIAGEN) and the concentration of DNA was measured by NanoDrop (NanoDrop Technologies, USA) amplified product. The amplified sequences were then pooled and fluorescently labeled by dNTP in DOP-PCR and subsequently used as a Gpx3 specific probe in FISH.
One probe, which was used as positive control, was developed from a RNO2 BAC clone (CH230-397A17 from BACPAC Resources Center, Oakland, California). The BAC DNA was amplified by DOP-PCR as follows. The reactions were performed in a final volume of 25 μl and with a BAC DNA concentration of 20 ng/μl. The mixture was heated at 94°C for 1 min and then subjected to 35 cycles of 94°C, 55°C and 72°C and a final extension at 72°C for 7 min. The product was verified on a 2% agarose gel with an appropriate size marker. For FISH, the Nick Translation kit from Abbott molecular was used according to the manufactures protocol.
Human material
A total of 30 EACs in FIGO grade I-III (10 tumors from each grade embedded in archival formalin fixed paraffin (FFPE) were used in the study. Apart from the endometrial tumors, 21 benign endometrial tissues were collected, and reference material from lung was used in the normalization process (Table 2 ). All samples were anonymous. A pathologist marked the tumor area in samples in the hematoxylin and eosin slide. Using a Tissue Micro Array-equipment (Pathology Devices), 3-4 cores (∅0.6 mm) of tumor tissue was punched out from the paraffin block. After standard tissue sample deparaffinization using xylene and alcohols, samples were lyzed in a Tris-chloride, EDTA, sodium dodecyl sulfate (SDS) and proteinase K containing buffer. RNA was then extracted and used for the real time qPCR.
Quantitative PCR (qPCR) of GPX3 and MET in FIGO grade I-III human EACs
Total RNA was extracted and used for qPCR according to the same procedure as for the rat samples.
Statistical analysis
For statistical evaluations of Ct values for differences among replicates we applied paired samples t-test and for comparisons of normal and malignant tissues independent sample t-test was applied. (PASW Statistics 18, SPSS Inc, Chicago, USA). In both tests the null hypotheses were assuming no differences between replicates, and no differences between tissue types respectively. The Pearson correlation test was performed to check for correlation between the expression of Gpx3 and Met . The significance levels were set to P < 0.5 in all statistical tests. | Results
QPCR expression analysis of Gpx3 and Met
The statistical analyses applied for comparing replicates, revealed no significant differences among the replicates in either of the data sets. An average of the replicates Ct value were used in the following calculations of the relative quantitative gene expression, the delta-delta Ct value.
The majority of the rat tumors displayed an almost total loss of expression of Gpx3 whereas the non-/pre-malignant endometrial samples displayed a normal/high expression, as also shown by the highly significant difference among malignant and premalignant/normal samples, (p < 0.001, Figure 1 , Table 1 ). The mRNA expression of GPX3 in the human material assessed in 30 human EACs in FIGO grade I-III (10 tumors from each grade), and 21 benign endometrial samples differed significantly between normal and malignant tissues (P < 0.001), but no differences among the different groups of malignant tumors were seen (P > 0.5) (Table 2 ).
In the rat cell lines, the expression of the Met gene was slightly higher among the malignant cell lines (P = 0.054). In the human material the expression of MET was significantly lower among the malign tumors (P < 0.001), but there were no differences between FIGO grades (P > 0.5) (Figure 1 ).
Epigenetic inactivation of the Gpx3 gene
The methylation status of the Gpx3 promoter region on bisulfite treated DNA was examined with methylation specific primers (MSP) in methylation-specific PCR of the inbred rat strain BDII, one NME cell line and 15 EAC cell lines (Table 3 ). The NME cell line was not methylated, while in the BDII rat strain and in 14 out of 15 EAC cell lines (93%), the Gpx3 promoter region was hypermethylated of (Table 4 ).
Demethylation of the Gpx3 gene
Two of the tumor cell lines (NUT12 and NUT81) that displayed Gpx3 promoter biallelic hypermethylation were randomly selected for treatment with the demethylating agent 5-aza-2 ́-deoxy-cy-tidine (5Aza-dC) in combination with the histone deacetylace inhibitor, trichostatin A (TSA) (Figure 2 ). As is shown, the Gpx3 mRNA expression was fully restored after the demethylation treatment.
Reactive oxidative species (ROS) in cancer cells
The production of hydrogen peroxide in one rat endometrial tumor cell line (NUT12) with loss of expression of Gpx3 and in two non-malignant endometrial cell lines (NUT43 and NUT56) with Gpx3 expression was measured. A higher generation of hydrogen peroxide was produced in the tumor cell line than in the premalignant cell lines (Figure 3 ).
Deletion of the gene Gpx3 in the rat EAC tumor NUT84
FISH was performed on the cell line where Gpx3 was unmethylated, but with a decreased gene expression of Gpx3 . The Gpx3 probe was generated from an NME cell line (NUT43). Six primer pairs representing all exons of Gpx3 were designed and used in PCR (Table 5 ). The PCR products were subsequently pooled, amplified and fluorescently labeled in DOP-PCR. As a positive control, we used a probe generated from a RNO2 BAC. The result of the FISH experiment showed that Gpx3 was deleted in the NUT84 cell line. | Discussion
Gpx3 was identified as a potential molecular biomarker for rat EAC in a previous microarray study [ 9 ]. In this study, we sought to confirm the previously demonstrated down-regulation of Gpx3 expression from the microarray study, of a selection of previously used rat endometrial tumors (Table 1 ).
As indicated in the previous study, the rat tumors as well as the human tumors indeed displayed an almost total loss of expression of Gpx3 , whereas the non-/pre-malignant endometrial samples displayed a normal/high expression, p < 0.001 (Figure 1 ). The mRNA expression of GPX3 in the human material was measured in 30 human EACs in FIGO grade I-III (10 tumors from each grade), and 21 benign endometrial samples using qPCR (Table 2 ). Regardless of tumor grade, the expression of GPX3 was low in all tumors, whereas the benign endometrial tissues exhibited a relatively high expression. Thus, in comparisons between tumor FIGO grades we found no significant difference (p > 0.5), while between malignant tumors and benign tissues, we found significant differences, (p < 0.001) (Figure 1 ). These observations are in agreement with the down-regulated expression of Gpx3 in the rat EACs described above. Accordingly, the dramatic loss of Gpx3 expression in almost all rat EAC samples and the consistently low expression of GPX3 in human EAC, suggest that the down-regulation of Gpx3 expression is an early event in EAC carcinogenesis.
In a study performed by Yu et al . [ 17 - 23 ], it was suggested that GPX3 has tumor suppressor activity as they could show that induced over-expression of GPX3 in prostate cancer cell lines decreased invasiveness, anchorage independent growth and colony formation. Moreover, xenografted prostate cancer cells expressing GPX3 , showed reduction of tumor size, elimination of metastasis and reduction of animal death. Their findings also suggest that the GPX3 tumor suppressor activity seems to involve transcriptional regulation of the tumor oncogene, MET [ 12 ]. MET activation by HGF ligand binding, initiates a whole spectrum of biological activities, including growth promotion, motility and invasiveness [ 24 ]. Multiple signaling transduction pathways are induced by MET engagement, including; the MAPK pathway, the RAS pathway, the P13K pathway, the STAT pathway, the beta catenin pathway and the Notch pathway [ 25 ]. To investigate whether Gpx3/GPX3 regulates the expression of Met / MET in the rat EAC and human tumor material, we performed qPCR on Met as well. The expression of Met in the rat tumors was slightly higher in the endometrial tumors with a loss of Gpx3 expression. Applying Pearsson ́s correlation test, no significant correlation between the Gpx3 and Met gene expression was detected in the rat tumor material. Thus, in the tumor materials included in this study, we could not confirm any tumor suppressor activity of Gpx3 . In the human material highly significant differences in the expression of MET between normal and tumor tissues were found (P < 0.001), but not in the way as would be expected from previous studies, where it was suggested that down regulation of GPX3 causes up-regulation of MET in some cancer types [ 17 - 23 ]. Here, when GPX3 was down regulated, also the MET gene was down regulated, which was also confirmed by the positive correlation displayed by Pearsson ́s correlation test (0.67, P < 0.001). Thus, in this study we could not detect any tumor suppressor activity of GPX3 in MET/HGF mediated pathways as suggested by Yu, et al. [ 17 - 23 ].
The GPX3 gene was shown to display epigenetic inactivation in prostate cancer, ovarian clear cell adenocarcinoma, gastric carcinoma and Barret's disease [ 11 - 16 ]. We therefore investigated the methylation status in the BDII inbred rat strain, the EAC cell lines with confirmed decreased mRNA expression (n = 14) in the NME cell lines with a confirmed normal expression of Gpx3 (n = 2) (Figure 1 ). Hypermethylation of the Gpx3 promoter region was found in 14 EAC cell lines. Thirteen of the EAC samples the Gpx3 promoter were biallelically methylated (Table 4 ), One tumor, NUT39, showed monoallelic methylation and Gpx3 was therefore not as much down-regulated in this tumor as in the tumors with biallelic hypermethylation, which was also confirmed in the expression analysis (Table 4 ). The premalignant samples (NME), with an up-regulated expression of Gpx3 , were not methylated.
Two of the tumor cell lines, NUT12 and NUT81, that displayed biallelic hypermethylation of the Gpx3 promoter, were selected for treatment with the demethylating agent 5-aza-2 ́-deoxy-cy-tidine (5Aza-dC) in combination with the histone deacetylace inhibitor, trichostatin A (TSA). The Gpx3 mRNA expression was fully restored after the demethylation treatment (Figure 2 ). Hence, these results confirm that promoter methylation plays an essential role in silencing of the Gpx3 expression.
Cancer cells are constantly exposed to oxidative stress, and it has been shown that human tumor cell lines generate reactive oxidative species (ROS) to a much higher extent than do non-transformed cell lines [ 26 ]. When we measured the production of hydrogen peroxide in two rat endometrial tumor cell lines with loss of expression of Gpx3 and in one premalignant/normal endometrial cell line with normal Gpx3 expression, we found that a higher generation of hydrogen peroxide was produced in the tumor cell lines than in the premalignant cell line (Figure 3 ). The loss of the protective properties of Gpx3 most likely makes the endometrial cells more vulnerable to ROS damage and genome instability. These findings suggest that the GPX3 function is impaired in endometroid adenocarcinoma, and a likely consequence is an increased amount of hydrogen peroxide and other reactive oxidative species (ROS). Clearly, more tumor cell lines have to be investigated and further functional analyses are required to elucidate the role of ROS in EAC.
In our study Gpx3 was strongly down regulated in NUT84. From the result of the methylation studies of this cell line it became clear to us that the loss of expression did not depend on epigenetic inactivation of the gene (Table 4 ). However, the down-regulation may be explained by other mechanisms, such as structural aberrations at the site of the Gpx3 locus. In fact in a previous study a chromosomal deletion has been observed in the region of RNO10 where Gpx3 is located, (RNO10q22, 40.3 Mb) which might explain the low expression of the gene [ 27 ]. Consequently, we decided to perform FISH on that cell line with a Gpx3 probe that was generated from an NME cell line (NUT43) with normal expression of Gpx3 . From the results of the FISH experiment, we could determine that Gpx3 was included in the deletion of that was previously detected [ 27 ], which explains the decreased expression of Gpx3 in NUT84. As the BDII rat strain did not exhibit a deletion in the region including Gpx3 most probably the deletion is an event that has occurred during the tumor development. | Conclusion
We have found a consistent down-regulated expression of Gpx3 in both rat and human EACs. The limited expression of Gpx3 in the rat cell lines was correlated to biallelic hyper-methylation of the Gpx3 promoter region. Demethylation of the genomes resulted in a restored expression, suggesting that the hypermethylation is responsible for the down-regulation of Gpx3 . From the FISH images we could confirm that the Gpx3 gene was deleted in one of the tumors that was down regulated, but not methylated (NUT84). In previous studies in human prostate cancer, it was suggested that GPX3 exhibits tumor suppressor activity by transcriptional regulation of the oncogene MET [ 11 - 16 ]. We could not confirm any such tumor suppressor activity of GPX3/ Gpx 3 either in human or rat endometrial tumors.
It has been proposed that ROS overproduction is required for hypoxic activation of HIF-1 [ 28 ] and the results from the preliminary experiments in this study indicate that the GPX3 function is impaired also in endometroid adenocarcinoma. To conclude, the results presented here propose that there are important clinical implications of GPX3 expression in EAC, both as an important molecular biomarker for EAC and as a potential target for therapeutic interventions. | Glutathione peroxidase 3 (GPX3) is one of the key enzymes in the cellular defense against oxidative stress and the hepatocyte growth factor receptor, (MET) has been suggested to be influenced by the GPX3 gene expression. In a previous microarray study performed by our group, Gpx3 was identified as a potential biomarker for rat endometrial adenocarcinoma (EAC), since the expression was highly downregulated in rat EAC tumors. Herein, we have investigated the mRNA expression and Gpx3 and Met in rat EAC by real time quantitative PCR (qPCR), and the methylation status of Gpx3 . In addition we have examined the expression of GPX3 and MET in 30 human EACs of different FIGO grades and 20 benign endometrial tissues. We found that the expression of GPX3 was uniformly down regulated in both rat and human EAC, regardless of tumor grade or histopathological subtype, implying that the down-regulation is an early event in EAC. The rate of Gpx3 promoter methylation reaches 91%, where biallelic methylation was present in 90% of the methylated tumors. The expression of the Met oncogene was slightly upregulated in EACs that showed loss of expression of Gpx3 , but no tumor suppressor activity of Gpx3/GPX3 was detected. Preliminary results also suggest that the production of H 2 O 2 is higher in rat endometrial tumors with down-regulated Gpx3 expression. A likely consequence of loss of GPX3 protein function would be a higher amount of ROS in the cancer cell environment. Thus, the results suggest important clinical implications of the GPX3 expression in EAC, both as a molecular biomarker for EAC and as a potential target for therapeutic interventions. | List of abbreviations
EAC: endometrial adenocarcinoma; NME: normal/pre-malignant endometrium; NUT: backcross, rat uterine tumor; RUT: rat uterine tumor
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
SK and EF contributed with all original ideas, designed all studies, performed the data analysis and drafted the manuscript. SK and EF carried out the methylation specific PCR and Q-PCR. SK and JC performed the hydrogen-peroxide assays and the de-methylation study. EF, GH and MK were responsible for the human endometrial study. KKL supervised the project and contributed with ideas and took part in the preparation of the manuscript. All authors have read and approved the final version of the manuscript. | Acknowledgements
This work was supported by the Swedish National Research School in Bioinformatics and Genomics, Erik Philip-Sörensen Foundation, and the Nilsson-Ehle foundation. We are grateful to Karin Lilja for excellent technical assistance. | CC BY | no | 2022-01-12 15:21:37 | Cancer Cell Int. 2010 Nov 24; 10:46 | oa_package/b5/8b/PMC3014921.tar.gz |
PMC3014922 | 21539753 | Introduction
In the last decade the etiologic background of several monogenic autoinflammatory syndromes has been identified including FMF, MKD, TRAPS, CAPS - CINCA, FCAS and MWS, Blau syndrome and PAPA. Specific mutations of the genes encoding proteins that modulate inflammasome activity have been discovered for these syndromes. Periodic fever syndrome in which no mutation has been found so far is PFAPA syndrome [ 1 , 2 ]. The spectrum of autoinflammatory diseases is still expanding with continuous reports of novel genetic syndromes [ 3 , 4 ]. With better understanding and genetic testing the number of patients diagnosed with periodic fever syndromes is increasing but there are probably many unrecognized cases especially in countries where these diseases are rare.
To our knowledge, the frequency of periodic fever syndromes in Eastern and Central European (ECE) countries has not been reported so far. In this study we aimed to identify the prevalence of diagnosed and suspected periodic fever syndromes and other autoinflammatory diseases in ECE countries. | 2. Methods
We performed a cross-sectional survey on periodic fever syndromes in ECE countries. Two different strategies were used to collect data on patients with periodic fever syndromes: i) the Eurofever survey- a secured web-based questionnaire administered to 54 pediatric rheumatology centers in ECE countries, members of PRINTO, ii) collection of data with the structured questionnaire distributed by e-mail to the pediatric immunologists included in the "J project" which is a physicians' education campaign promoted by ESID in order to sensitize pediatricians in ECE countries to the recognition of immune-mediated diseases. Questionnaire was also sent to other pediatricians with an interest in rheumatology and immunology, not included in PRINTO or ESID networks, mainly working in the tertiary care centers of pediatric rheumatology and immunology in ECE countries. Two rounds of e-mails with structured questionnaires were sent for both initiatives.
In total, the questionnaire was sent to 69 physicians from 16 ECE countries, including Albania, Bulgaria, Bosnia and Herzegovina, Croatia, Czech Republic, Estonia, Hungary, Latvia, Lithuania, Macedonia, Montenegro, Poland, Romania, Serbia, Slovakia and Slovenia. According to the 2007 Eurostat data for each included country, it is estimated that 123.6 million people live in these countries [ 5 ].
The structured questionnaire was mainly focused on monogenic inflammatory diseases. In particular we asked each center about the number of patients with a genetically confirmed diagnosis of monogenic autoinflammatory disease, where the genetic testing was performed and the possibility of genetic testing within the country. A suspected case was defined on the basis of the positivity of diagnostic criteria, whenever available (such as for FMF [ 6 ]) and/or on the intention to test for a given gene associated with autoinflammatory disease. We were also interested in the numbers of patients with suspected syndromes. In its third part the questionnaire included queries on the number of patients with other diseases grouped under the term autoinflammatory. | Results
All together we received data from 37 physicians from 35 centers (54% combined response rate for both surveys) from 14 ECE countries including Albania, Bosnia and Herzegovina, Bulgaria, Croatia, Czech Republic, Hungary, Latvia, Lithuania, Macedonia, Poland, Romania, Serbia, Slovakia and Slovenia. The estimated resident population in these countries is 121.5 million and the number of children 0-19 years is approximately 27.0 million. This data was obtained from the latest available statistical report for each participating country.
Among the physicians who responded to the survey the distribution of specialists was as follows: 14 pediatric rheumatologists, 8 pediatric immunologists and 11 pediatricians with interest in this field.
Genetically confirmed patients with monogenic periodic fevers
The overall numbers of patients reported with FMF, MKD, TRAPS and CINCA in ECE countries is shown in Table 1 . Five adult patients were included- 3 with MKD and 2 with TRAPS. No country reported more than 2 genetically confirmed FMF patients. The majority of MKD patients were from the Czech Republic (8). Patients carrying TNFRSF1A mutations were reported from the Czech Republic (4), Hungary (3), Poland (2), Latvia (1) and Slovenia (1). CINCA patients were reported from Hungary (2), Czech Republic (1) and Slovakia (1).
One genetically confirmed case of Blau syndrome was reported from Croatia. Genetic testing was done in Portland, Oregon, USA.
Genetic testing for suspected FMF patients from Slovenia was performed at the University Children's Hospital Ljubljana and for patients from Macedonia and Serbia at the Medical faculty Skopje. At the time of our survey, no genetic testing for other autoinflammatory diseases was available in ECE countries. Six centers from ECE countries sent DNA for genetic testing for periodic fever syndromes to Genova (Italy), two centers to Germany, including one to University of Freiburg, two centers to Turkey, one center to Montpellier (France), London (UK), Roma (Italy) and Barcelona (Spain).
Suspected cases of monogenic periodic fever syndromes and reported cases of other autoinflammatory diseases
The majority of patients with suspected FMF were from Romania (13), followed by Poland (8), Bulgaria (5), Hungary (4) and Croatia (4). Two adult patients were included. Other countries reported 3 or less patients with suspected FMF.
Among suspected MKD patients the highest number was reported from Romania (5), followed by Serbia (4), Czech Republic (3) and Slovakia (3).
Among suspected TRAPS patients the highest number was reported from Poland (6) followed by the Czech Republic (3). One suspected TRAPS patient was reported from Albania, Croatia, Hungary, Latvia, Romania, Serbia and Slovakia.
On the basis of their clinical presentation some patients were included in the study as suspected CAPS. Three suspected CINCA cases were referred from Poland and one from the Czech Republic, Hungary, Latvia and Slovakia. Two suspected FCAS patients were reported from Croatia and one from Serbia. One suspected MWS case was reported from Latvia and Poland. One suspected PAPA case was reported from the Czech Republic and one suspected case with Blau syndrome from Hungary.
The data about other reported autoinflammatory diseases are presented in Table 2 . Two adult patients were included- 1 PFAPA patient and patient with Schnitzler syndrome. | Discussion
Published studies investigating periodic fever syndromes were mainly conducted in Mediterranean countries, western European countries and United States. To our knowledge, there has been no systematic evaluation of patients with periodic fevers in ECE countries.
The aim of this study has been to find out if periodic fever syndromes and other autoinflammatory diseases are recognized in ECE countries. We contacted all physicians from eastern and central Europe who are members of PRINTO and ESID (J-project for eastern and central European countries) and pediatricians with an interest in rheumatology and immunology working in the tertiary care centers of pediatric rheumatology and immunology in ECE countries. The response rate for surveys was 54% which is a possible limitation to the collected data. We believe that the main reason for this is linked to the low awareness of these diseases in ECE countries.
The estimated frequency of FMF patient in ECE countries appears exceedingly low comparing to the data from Western European countries and it is likely that many FMF patients remain unrecognized. It is possible that clinical presentation of FMF is milder or different in ECE countries because of environmental modifiers. The estimated frequency of FMF in Mediterranean population ranges from 1/256 among North African Jews to 1/500 among Armenians and Israelis [ 7 ]. Lower but significant frequencies have been reported in other countries around the Mediterranean Sea [ 8 , 9 ]. In some patients with clinical features of FMF no mutation of the MEFV gene could be identified. In the Western European countries, only a minority of patients' compatible with FMF present MEFV mutations [ 8 ].
Genetically confirmed MKD was in our survey reported in 14 patients, which is the highest number among all genetically confirmed periodic fever syndromes in ECE countries. Moreover, an MKD was suspected in further 24 cases. According to these data, it seems reasonable that MKD may play a relevant role as a possible cause of inherited periodic fever syndromes in ECE countries, as already observed in populations other than Dutch or north-European [ 10 - 12 ].
The number of patients with recurrent fever carrying one mutation of TNFRSF1A gene in ECE countries was equal to that observed for genetically confirmed FMF cases. TRAPS was originally considered an exceedingly rare disease with only few families described in the literature. During the past decades more than 50 disease associated mutations have been identified in hundreds of patients with different ethnic backgrounds [ 13 ]. Since the survey was not aimed to collect information on the type of mutations found for each patient we cannot exclude that part of the mutations found were hypomorphic variants of the gene, such as R92Q or P46L.
CINCA is a very rare congenital inflammatory disease [ 14 ] and we were able to identify only four patients from ECE countries.
In the present study, a relevant number of patients were classified as PFAPA. It is likely that at least part of the 282 patients with a PFAPA phenotype identified in ECE countries might also be positive at genetic testing [ 15 ].
In the majority of patients included in the survey genetic analysis was not done. With the effort of the Eurofever project http://www.printo.it/eurofever the network of laboratories that offer genetic testing for periodic fever syndromes was established. The next important step is to further spread the knowledge about these rare diseases among physicians of different specialties in ECE countries. A low number of trained pediatric rheumatologists and immunologists in these countries is likely one of the contributing factors for low recognition of these diseases. With the purpose to improve the knowledge about autoinflammatory diseases in the region of ECE countries and southern Europe, the first educational meeting was organized in October 2009 in Ljubljana, Slovenia. The meeting was officially supported by the PReS and the J Project, and was attended by 121 participants from 22 countries.
In summary, this study presents the first data on the prevalence of diagnosed and suspected autoinflammatory diseases in ECE countries. It demonstrates that all recognized periodic fever syndromes do affect also patients from ECE countries, with a frequency which appears to be lower than in the Mediterranean and Western European countries. In order to identify more patients in the future, it is important to continue with educational activities, that is one of the main aims of the Eurofever project. | Objective
To analyze the prevalence of diagnosed and suspected autoinflammatory diseases in Eastern and Central European (ECE) countries, with a particular interest on the diagnostic facilities in these countries.
Methods
Two different strategies were used to collect data on patients with periodic fever syndromes from ECE countries- the Eurofever survey and collection of data with the structured questionnaire.
Results
Data from 35 centers in 14 ECE countries were collected. All together there were 11 patients reported with genetically confirmed familial Mediterranean fever (FMF), 14 with mevalonate-kinase deficiency (MKD), 11 with tumor necrosis factor receptor associated periodic syndrome (TRAPS) and 4 with chronic infantile neurological cutaneous and articular syndrome (CINCA). Significantly higher numbers were reported for suspected cases which were not genetically tested. All together there were 49 suspected FMF patients reported, 24 MKD, 16 TRAPS, 7 CINCA and 2 suspected Muckle-Wells syndrome (MWS) patients.
Conclusions
The number of genetically confirmed patients with periodic fever syndromes in ECE countries is very low. In order to identify more patients in the future, it is important to organize educational programs for increasing the knowledge on these diseases and to establish a network for genetic testing of periodic fever syndromes in ECE countries. | Abbreviations
FMF: familial Mediterranean fever; MKD: mevalonate-kinase deficiency (MKD); TRAPS: tumor necrosis factor receptor associated periodic syndrome; CAPS: diseases from the spectrum of cryopyrin associated periodic fever syndromes; CINCA: chronic infantile neurological cutaneous and articular syndrome; FACS: familial cold autoinflammatory syndrome; MWS: Muckle-Wells syndrome; PAPA: pyogenic sterile arthritis, pyoderma gangrenosum and acne syndrome; PFAPA: periodic fever syndrome with aphthous stomatitis, pharyngitis and cervical adenopathy; CRMO: chronic recurrent multifocal osteomyelitis; PRINTO: Paediatric Rheumatology International Trials Organization; ESID: European Society for Primary Immunodeficiencies
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
NT, MG and TA participated in the design of the study. All authors participated in the acquisition of data from their countries. NT, NR, MG and TA participated in the statistical analysis and interpretation of data. NT, PD, MG and TA drafted the manuscript. All authors read and approved the final manuscript. | Acknowledgements
This work was partially supported by the Eurofever Project (EAHC, Project No2007332) and by the Slovenian Research Agency (Grant NoL3-0624).
Members of Eastern and Central European autoinflammatory collaborating group in alphabetic order : Mihaela Bataneant, University of Medicine and Pharmacy "Victor Babes" Timisoara, IIIrd Pediatric Clinic, Children's Hospital "Louis Turcanu", Romania; Iva Bilič, University Hospital Centre Rijeka, Croatia; Beata Derfalvi, 2 nd Department of Pediatrics, Semmelweis University Budapest, Hungary; Nicolae Iagaru, Institute for Mother and Child care Bucharest, Romania; Krzysztof Kalwak, Medical University Department, Ped. Hemat/Onc/BMT, Wroclaw, Poland; Anuela Kondi, UHC Mother Theresa, Tirana, Albania; Elena Koškova, National Institute of Rheumatic Diseases, Piestany, Slovakia; Petra Krol, Rheumatology Unit, Department of Paediatrics and Adolescent medicine, Charles University in Prague, 1 st medical School and General University Hospital in Prague, Czech Republic; Dafina Kuzmanovska, University Pediatric Clinic, Skopje, Republic of Macedonia, Jiri Litzman, Department of Clinical Immunology, Allergy, St. Anne University Hospital, Brno, Czech republic; Dragana Malčić Zanič, Children's Hospital Banja Luka, Bosnia and Herzegovina; Henryka Mazur Zielinska, Department of Pediatrics, Medical University of Silesia, Poland; Dimitrina Mihaylova, University Children's Hopsital Bulgaria; Lenka Minxova, University Hospital, Department of Pediatrics, Hradec Králové, Czech Republic; Velma Mulaosmanovič, Children's Hospital University Clinical Center Sarajevo, Bosnia and Herzegovina; Andrea Ponyi, Semmelweis University, 2 nd Department of Pediatrics, Unit of Pediatrics Rheumatology, Budapest, Hungary; Tatjana Prokofjeva, Children's Clinical University Hospital, Riga, Latvia, Vojko Rožmanič, University Hospital Centre Rijeka, Croatia; Ingrida Rumba, University of Latvia, Riga; Skirmante Rusoniene, Vilnius University Children's Hospital, Lithuania; Lidia Rutkowska Sak, Institute of Rheumatology, Warsaw, Poland; Ruta Santere, University Children's Hospital, Latvia; Valda Stanevicha, Riga Stradins University, Children University Hospital, Latvia; Gordana Susič, Institute of Rheumatology, Belgrade, Serbia; Anna Szaflarska, Department of Clinical Immunology, Polish-American Institute of Pediatrics, Jagiellonian University Medical College, Cracow, Poland; Lana Tambič Bukovac, Children's Hospital Srebrnjak, Zagreb, Croatia; Veronika Vargova, Children Faculty Hospital Košice, Slovakia; Richard Vesely, Slovakia; Gordana Vijatov Djurić, Institute for Child and Youth Health Care of Vojvodina, Novi Sad, Serbia; Jelena Vojinović, Department of Pediatric Rheumatology, Clinic of Pediatrics, University Clinical Center Niš, Serbia; Zuber Zbigniew, Wojewódzki Specjalistyczny Szpital, Krakow, Poland; Ellenes J Zoltan, City Hospital Oradea, Romania; Balogh Zoltos, National Institute of Rheumatology and Physiotherapy (ORFI), Budapest, Hungary;
Eurofever steering committee: Joost Frenkel (The Netherland), Seza Ozen (Turkey). | CC BY | no | 2022-01-12 15:21:37 | Pediatr Rheumatol Online J. 2010 Dec 2; 8:29 | oa_package/71/fc/PMC3014922.tar.gz |
|
PMC3014923 | 21134287 | Background
Children with juvenile idiopathic arthritis (JIA) are at risk of bone destructions and reduced bone mass. The pathogenesis for the bone loss is complex and is influenced by inflammation, physical inactivity, nutrition and medication. The immune and skeletal systems share a number of regulatory molecules, and there is accumulating evidence indicating interactions between the two systems [ 1 ].
Bone remodelling is a lifelong continuous process conducted by osteoblasts, synthesizing bone matrix and its resorption by osteoclasts. Important regulators of osteoclast recruitment and function are the three key molecules Osteoprotegerin (OPG), Receptor Activator of Nuclear factor -κB (RANK) and its ligand (RANKL). RANKL stimulates osteoclast production and survival via the membrane -bound receptor RANK, [ 2 - 5 ] while OPG inhibits osteoclast differentiation and activation due to its function as a non-signalling decoy receptor for RANKL [ 6 ]. The physiological balance between RANKL and OPG is regulated by various calcitropic cytokines and hormones and alterations in their ratio are critical in the pathogenesis of bone diseases [ 7 ]. Osteoblasts and T cells are important producer cells of RANKL. An inflammatory environment with T-cell activation may tilt the balance between OPG and RANKL and increase osteoclast activation and bone resorption.
In patients with early rheumatoid arthritis (RA), baseline serum OPG/RANKL ratio and inflammation have independently predicted radiographic progression of joint damage [ 8 ]. RANKL-expressing cells and RANK-expressing osteoclast precursor cells, and more limited OPG, have been demonstrated at sites of subchondral bone erosions [ 9 ]. In children with JIA, over expression of RANKL has been detected in synovial fluid mononuclear cells from joints [ 10 ]. Higher serum RANKL and lower serum OPG/RANKL ratio has been found in two studies of patients with JIA compared to controls [ 11 , 12 ], while higher serum levels of OPG [ 13 , 14 ] or an increased OPG/RANKL ratio [ 14 ] has been shown in other studies. Publications concerning this topic in patients with JIA are still few, the studies cross-sectional, and the clinical relevance of the observations is not clear.
To further elucidate the potential role of OPG and RANKL in JIA we prospectively explored serum levels of OPG and RANKL in an observational cohort study of children with early disease, compared to individually matched children, and in relation to radiographic score, bone and lean mass, disease activity, and medication. | Methods
Study participants
All Caucasian children with JIA between the ages of 6 and 18 years who were attending the Department of Rheumatology, Oslo University Hospital, Rikshospitalet, for the first time from May 1995 to February 1999 were invited to participate in a two-year prospective study of bone mass and bone turnover. Of 127 eligible patients, 108 (85%), living in 16 different counties (latitudes from 58°N to 68°N) were included. Each patient was individually matched to a healthy child with the same sex, age, race, and county of residence who was randomly selected from the National Population Register. The characteristics of the total patient group, including nutritional status, level of physical activity, markers of bone formation and bone resorption, and bone mass, have previously been described [ 15 ]. In the present study we included only patients who met the JIA criteria [ 16 ] for oligoarthritis (n = 59) or polyarthritis (n = 31) and their controls (n = 90). Patients' mean disease duration was 19.4 months (SD 12.3). The participants were examined at baseline and at follow-up, a mean of 24.2 months (SD 1.4) later. The Regional Ethics Committee for Medical Research approved the study. Written informed consent was obtained from the parents and from children older than 16 years.
Clinical examination
Clinical information was obtained by interviews, physical examination and questionnaires [ 15 , 17 - 22 ].
Radiographic examination
Radiographs of the non-dominant hand and wrist were taken for assessment of skeletal maturity of all study subjects at baseline and follow-up [ 23 ]. Radiographs of the patients' knees and ankles were obtained at the time of admission to hospital; other joints when clinically indicated, and were scored by one of 2 trained radiologists (KD and VJ) according to a radiographic classification system for juvenile rheumatoid arthritis: grade 0 (normal joints), grade 1 (juxtaarticular osteoporosis and/or periarticular soft tissue swelling), grade 2 (growth abnormality, bony erosion not present), grade 3 (growth abnormality and marginal bony erosions), grade 4 (deformation and severe erosions), and grade 5 (gross destruction and deformity) [ 24 , 25 ]. Radiographic progression was defined as an increase in the radiographic grade during the study. An increase from grade 0 to 1 was not considered to be radiographic progression.
Laboratory measures
Venous blood samples were obtained before noon at baseline and follow-up. Bone markers were measured in the second void urine sample of the morning. For analysis of OPG, RANKL and C-reactive protein (CRP) serum were stored at -70°C until analyzed. OPG and CRP (detection limit 0.16 mg/L) was quantified by enzyme immunoassays (EIA) using commercially available matched antibodies from R&D systems (Minneapolis, Minnesota) [ 26 ] and Dako cytomation (Glostrup, Denmark), [ 27 ] respectively. RANKL was quantified by EIA (Bender MedSystems, Vienna, Austria). Bone formation was assessed by serum levels of bone-specific alkaline phosphatase and osteocalcin, bone resorption by serum levels of C-telopeptide of type 1 collagen and urinary concentration of deoxypyridinoline, [ 28 - 30 ] and vitamin D stores by serum concentration of 25-hydroxyvitamin D and the active hormonal metabolite by serum 1,25-dihydroxyvitamin D3, all samples analyzed immediately by routine laboratory methods.
Bone mass measurements
Measurements of the total body, distal radius, femoral neck and L2-L4 of the lumbar spine were evaluated with the same dual x-ray absorptiometry (DXA) equipment (Lunar Expert-XL; GE Lunar, Madison, WI). All analyses were performed by one investigator using Expert-XL software version 1.72 and 1.91 and were read by one investigator (GL). The bone mineral content (BMC) was calculated from all the regions. Fat and lean composition of soft tissue was calculated from the total body scan. The long-term precision of the scanner was tested daily with minimal drift in the measurements and a coefficient of variation (CV) of 0.5%. The in vivo precision of the operator technique (patients and healthy subjects) was 1.6% for lumbar spine and 2.0% for femoral neck [ 31 ]. The total body BMC findings were calculated as Z-scores in terms of the number of standard deviations (SD) above or below the age-specific mean for healthy individuals (Z-score = [subject's measurement - mean measurement of the reference population]/SD of the reference population). By definition, 16% of healthy children will have a Z-score less than -1.0 SD, and 2.3% will have a Z-score less than -2.0, but no evidence-based guidelines exist to define osteopenia and osteoporosis in children [ 32 ]. We defined low BMC as Z-score between -1 SD and -2 SD and very low BMC as a Z-score > 2 SD below the mean.
Statistical analysis
Differences between patients and matched controls for clinical, radiographic, bone mass and laboratory measures were tested by paired samples t-test for continuous variables and McNemar's test for categorical variables. Within the patient cohort, differences were tested by the independent samples t-test or by one-way analysis of variance using the Bonferroni correction for multiple comparisons for continuous variables, and the chi-square test for categorical variables.
Multiple regression analyses were performed to identify predictors of the 2-year changes in body lean mass, BMC and numbers of active or mobility restricted joints. Explanatory variables were included in the model if the p value was less than 0.2 in unadjusted linear regression analyses or if a variable was known to be associated with the outcome variable [ 33 ]. Highly intercorrelated independent variables (r > 0.7) in a multiple model were avoided. To reduce the possibility of body size-related artefacts in the analyses of bone mass, bone area, weight, and height were included in the multiple regression models for BMC [ 34 ]. Forward stepwise regression methods were used.
For all analyses, p values less than or equal to 0.05 (2-tailed tests) were considered significant. The statistical analysis was performed using SPSS software version 14.0 (SPSS, Chicago, IL). | Results
Demographic and clinical data
A total of 90 JIA patients and 90 controls were included. Clinical features at baseline and 2-year follow-up are shown in table 1 .
Serum levels in JIA patients and controls
Figure 1 shows the serum levels and changes from baseline to 2-year follow-up in patients and controls. The OPG level was significantly lower in the patients than in the controls at baseline (p = 0.003). The levels of RANKL were higher in the patients at follow-up (p = 0.073), but the findings did not reach statistical significance. The OPG/RANKL ratio tended to be lower in the patients than controls at baseline and follow-up (p = 0.061 and p = 0.200). The level of CRP was significantly higher in the patients at baseline (p = 0.036). There were no significant differences in serum levels of OPG, RANKL, OPG/RANKL ratio or CRP between males and females (data not shown).
Changes in serum levels, radiographic score and bone mass
Changes in the OPG/RANKL ratio and serum levels of CRP from baseline to 2-year follow-up, in relation to different radiographic scores and total body BMC Z-scores, are shown in Figure 2 . No patient had higher radiographic score than grade 3 with growth abnormality and bony erosions. The increments from baseline to follow-up of the OPG/RANKL ratio and CRP were significantly higher in patients with very low BMC and patients with erosions, compared to patients with less serious findings (p values ranged from <0.001 to 0.007). A greater increase in OPG/RANKL ratio was also seen in patients with radiographic progression compared to the patients without radiographic progression (difference of increase 33.7, 95%CI = 10.7, 56.7, p = 0.004).
Changes in serum levels and disease-onset type
The levels of RANKL increased significantly more from baseline to follow-up in the polyarthritis-group than in the oligoarthritis-group (p = 0.015), while the changes in OPG, OPG/RANKL ratio and CRP were not significantly different (Figure 3 ).
Serum levels, antirheumatic and corticosteroid treatment
Figure 4 shows serum OPG, RANKL, OPG/RANKL ratio and CRP at 2-year follow-up in relation to current DMARD and corticosteroid treatment. The levels of the OPG/RANKL ratio and CRP was significantly higher in the patients who were currently treated with DMARDs compared to the patients who were not (p = 0.013 and p = 0.037). Analyses comparing DMARD subgroups did not show significant differences. The serum OPG/RANKL ratio and CRP were significantly higher as well in patients who were current users of corticosteroids compared to those who were not (p = 0.012 and p < 0.001).
Predective value of OPG, RANKL, OPG/RANKL ratio and changes in lean mass, bone and affected joints
Laboratory measurements, patient and disease characteristics, physical activity and medication assessed at baseline were explored as independent predictors of the 2-year changes in lean mass, BMC and no. of active or mobility restricted joints by unadjusted linear regression analysis. Predictors chosen for the multiple regression model of the change in total body lean mass were baseline weight, height, bone age, weight-bearing physical activity, serum ionized calcium, 25-hydroxyvitamin D, parathyroid hormone, bone-specific alkaline phosphatase, C-telopeptide type 1, urinary deoxypyridinoline, baseline OPG, RANKL and CRP. In the final regression model baseline RANKL was a significant independent negative predictor of increased lean mass at follow-up (regression coefficient -1.5, 95% CI = -2.88,-0.12, p = 0.034) together with the independent positive predictors weight-bearing physical activity (regression coefficient 216, 95% CI = 51, 381, p = 0.011), bone-specific alkaline phosphatase (regression coefficient 31.7, 95% CI = 10.7, 52.6, p = 0.004) and 25-hydroxyvitamin D (regression coefficient 38.9, 95% CI = 0.48, 77.3, p = 0.047). Serum levels of OPG, RANKL or the OPG/RANKL ratio were not identified as independent predictors of the changes in BMC of total body, distal radius, femur neck or lumbar spine, or the changes of no. of joints with active disease or restricted mobility (data not shown). | Discussion
In the present prospective, observational cohort study of 90 JIA patients, early in the disease course, the level of OPG was significant lower than in 90 matched healthy children. Furthermore, during 2-years follow-up, RANKL increased more in patients with polyarthritis than in patients with oligoarthritis and RANKL was a significant negative predictor of lean mass. The OPG/RANKL ratio at follow-up was higher in patients on DMARD or CS treatment. To our knowledge, this is the first prospective controlled study of OPG and RANKL in children with JIA.
The findings of lower serum levels of OPG in our JIA patients with early disease are consistent with findings in children with untreated juvenile dermatomyositis [ 35 ], and so are the trends towards higher levels of RANKL and a lower OPG/RANKL ratio. Other studies exploring serum OPG and RANKL in JIA, differs from ours, but consist of patients with longer disease duration [ 11 , 12 , 14 ], more severe disease [ 12 , 14 ] or different subgroups [ 11 ], complicating comparisons. A recent study of polyarticular JIA patients showed higher serum levels of RANKL, a lower OPG/RANKL ratio, and comparable OPG levels compared to controls [ 12 ]. Another recent study, with a mixture of children and adults, found similar results in oligoarticular and polyarticular subtypes of JIA compared to controls [ 11 ]. A previous study has shown higher serum OPG levels and a higher OPG/RANKL ratio in oligoarticular and polyarticular JIA patients compared to controls, in contrast to the other studies [ 14 ]. These differences may have several explanations. As mentioned, the composition, age range and disease severity is not necessarily comparable between the studies. Importantly, we measured free RANKL and immunoassays measuring both OPG bound and free RANKL may give different results. An increase in circulating OPG has often been viewed as a compensatory response [ 14 ]. However, members of the TNF ligand superfamily often circulate at low levels with a short half time. Thus, as seen for other soluble TNF receptors [ 36 ], OPG may represent a reliable marker of the overall activity of the OPG/RANK/RANKL axis as well as a stable marker of inflammation. The correlation between CRP and the OPG/RANKL ratio in our JIA cohort supports this.
We found that baseline RANKL was a significant and independent negative predictor and weight-bearing physical activity a positive predictor of the gain in total body lean mass in the JIA patients. There are few data elucidating the impact of RANKL on the variation in lean mass in children but our results are most likely related to the chronic inflammation. In adult RA patients cachexia with muscle wasting and fat gain is common, and the mechanisms probably include cytokine-driven hypermetabolism during active disease [ 37 , 38 ].
The molecular pathways and specific effects of conventional DMARDs on bone and cartilage are not clearly defined [ 39 ]. However, DMARDs may have an effect on the osteoclast formation. In a study of cultures of fibroblast-like synoviocytes from patients with RA, the DMARDs methotrexate and sulfasalazine, have been shown to decrease the ratio of RANKL/OPG in a dose-dependent manner [ 40 ]. Another study of synovial tissue from patients with RA, treated with DMARDs, has shown increased OPG expression and decreased RANKL expression [ 41 ]. Our results with an increased OPG/RANKL ratio in DMARDs treated and corticosteroid treated patients are in accordance with these findings. Although corticosteroids are generally not considered conventional DMARDs there is evidence that corticosteroids have structure-sparing effects and can reduce the rate of erosion progression in RA [ 39 , 42 ].
There have been limited knowledge of the disease course during the first years of JIA and this study was part of a larger prospective comprehensive study [ 43 ]. Our patients were comparable to JIA patients in epidemiologic studies [ 44 ]. They were explored early in the disease course, the mean disease activity was low, the numbers of patients with radiographic erosions (6%) were low, and the numbers of patients with very low bone mass (4%) were low. The overall low disease activity in our patients is however a limitation for the interpretation of the results and the present study seems somewhat underpowered in places with strong trends but statistical significance not being attained. If we had chosen a patient cohort with higher disease activity and longer disease duration, the numbers of patients with joint erosions and very low bone mass would more likely have been higher. In addition, if more sensitive imaging methods as ultrasonography and magnetic resonance imaging (MRI), had been available as supplements to conventional radiographs, we might have detected a higher number of patients with structural damage [ 45 , 46 ]. | Conclusions
In summary, the JIA patients with oligo- or polyarthritis had significant lower levels of OPG early in the disease course compared to controls. The patients tended to have higher levels of RANKL and a lower OPG/RANKL ratio consistent with earlier findings. Baseline RANKL was a significant negative predictor of total body lean mass. To better understand bone loss and the clinical significance of the balance between OPG and RANKL in children with JIA more prospective data are warranted. Inclusion of children early in the disease course and before treatment with oral corticosteroids, DMARDs or biologic therapy will add information to our knowledge. | Background
The clinical relevance of observations of serum levels of osteoprotegerin (OPG) and receptor activator of nuclear factor -κB ligand (RANKL) in juvenile idiopathic arthritis (JIA) is not clear. To elucidate the potential role of OPG and RANKL in JIA we determined serum levels of OPG and RANKL in patients with early JIA compared to healthy children, and prospectively explored changes in relation to radiographic score, bone and lean mass, severity of the disease, and treatment.
Methods
Ninety children with early oligoarticular or polyarticular JIA (ages 6-18 years; mean disease duration 19.4 months) and 90 healthy children individually matched for age, sex, race, and county of residence, were examined at baseline and 2-year follow-up. OPG and RANKL were quantified by enzyme-immunoassay. Data were analyzed with the use of t-tests, ANOVA, and multiple regression analyses.
Results
Serum OPG was significantly lower in patients than controls at baseline, and there was a trend towards higher RANKL and a lower OPG/RANKL ratio. Patients with polyarthritis had significantly higher increments in RANKL from baseline to follow-up, compared to patients with oligoarthritis. RANKL was a significant negative predictor for increments in total body lean mass. Patients who were receiving corticosteroids (CS) or disease-modifying antirheumatic drugs (DMARDs) at follow-up had higher OPG/RANKL ratio compared with patients who did not receive this medication.
Conclusions
The data supports that levels of OPG are lower in patients with JIA compared to healthy children, and higher levels of RANKL is associated with more serious disease. RANKL was a significant negative predictor of lean mass in patients with JIA. The OPG/RANKL ratio was higher in patients on DMARDs or CS treatment. | Abbreviations
BMC: Bone mineral content; CS: Corticosteroids; CRP: C-reactive protein; DMARDs: disease-modifying antirheumatic drugs; DXA: dual x-ray absorptiometry; JIA: juvenile idiopathic arthritis; OPG: Osteoprotegerin; RANKL: receptor activator of nuclear factor -κB ligand; TNF: tumor necrosis factor
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
GL contributed to conception and design, acquisition, analysis and interpretation of data, drafting of manuscript and final approval of manuscript. TU contributed to conception and design, acquisition, analysis and interpretation of data, revision of manuscript and final approval of manuscript. KG, AMS and BF contributed to conception and design, acquisition and analysis of data, revision of manuscript and final approval of manuscript. OF contributed to conception and design, revision of manuscript and final approval of manuscript. | Acknowledgements
We thank the participants and their parents for their patience, Gunn J.Hovland for analyzing the DXA scans, Knut Dale and Virginia Johnston for grading the radiographs, and Berit Halmrast and Helga V. Bruaseth for assistance with the data collection. This study was supported by the Oslo Women's Public Health Association, the Jan A. Pahles Research Legacy, the Solveig Amalie Husbys Memorial Foundation and the Eimar Munthe Legacy. | CC BY | no | 2022-01-12 15:21:37 | Pediatr Rheumatol Online J. 2010 Dec 6; 8:30 | oa_package/55/26/PMC3014923.tar.gz |
PMC3014924 | 21245904 | Introduction
Modification of chromatin structure regulates many aspects of cell and developmental biology. Epigenetic regulators are known to affect complex neuronal processes such as learning and memory [1] – [3] and contribute significantly to the occurrence of cognitive disorders, such as schizophrenia and intellectual disability (previously referred to as mental retardation) [4] , [5] . However, little is known about the “writers” of the neuronal epigenome that lay down the basis for proper cognition. Are these chromatin writers required to safeguard neuronal homeostasis/fitness by influencing the expression of a large and heterogeneous group of genes, such as house-keeping and non-neuronal genes? Or do they lay down specific epigenetic programs to regulate neuronal genes that are directly involved in determination of connectivity, plasticity, learning, and memory?
The euchromatin histone methyltransferases (EHMTs) are a family of evolutionarily conserved proteins that write part of the epigenetic code through methylation of histone 3 at lysine 9 (H3K9) [6] – [10] . In mammals, two EHMT paralogs exist, EHMT1/GLP and EHMT2/G9a. Heterozygous mutations or deletions of the human EHMT1 gene cause Kleefstra Syndrome (OMIM #610253), a neurodevelopmental disorder that is characterized by autistic-like features and severe intellectual disability [11] – [14] . Studies in mice have shown that Ehmt1/GLP and Ehmt2/G9a form a heterodimeric complex [8] , and that loss of either protein resulted in nearly identical phenotypes, such as early embryonic lethality, reduced H3K9 dimethylation (H3K9me2), and inappropriate gene transcription [7] , [8] . Furthermore, mice with neuronal ablation of Ehmt1/GLP and Ehmt2/G9a in adulthood show defects in fear conditioned learning. A gene expression study in different brain areas of these mice has led to the suggestion that EHMT proteins act as repressors of non-neuronal genes in neuronal tissues [15] . Here, by using classic Drosophila genetics, extensive neurodevelopmental and behavioral phenotyping, expression profiling, and genome-wide mapping of EHMT target loci by H3K9me2 profiling, we uncover EHMT as a key epigenetic regulator of neuronal genes and processes. | Materials and Methods
Fly Stocks and Maintenance
Flies were reared on standard medium (cornmeal/sugar/yeast) at 25 degrees and 45%–60% humidity with a 12-h light/dark cycle. All fly stocks were obtained from the Bloomington Drosophila stock center (Indiana University) (see Text S1 for stock descriptions) except for EHMT deletion strains and UAS-EHMT strains, which were generated according to standard procedures (see Text S1 and Figure 2 ).
Immunohistochemisty and Stainings
Tissues were dissected and fixed using standard methods. Rabbit-anti-EHMT antibodies were a gift from Dr. A. Lambertson [10] and were used at a 1/100. Rat-anti-elav (1/500), mouse-anti-repo (1/500), mouse-anti-DLG (1/100), and mouse anti-dac (1/100) antibodies were obtained from the Developmental Studies Hybridoma Bank (University of Iowa). Nuclei were visualized using the fluorescent nuclear dye DAPI. For imaging of type IV md neurons expressing a membrane targeted mCD8-GFP fusion protein (memGFP) we used a rat-anti-mCD8 antibody (Invitrogen) at 1/100. Secondary antibodies were conjugated to either alexa-fluor-568 or alexa-fluor-488 (Invitrogen). Images were acquired using either a Leica DM-IRE2 confocal microscope (Leica Microsystems) or a Zeiss Axioimager Z1 fluorescent microscope equipped with an ApoTome (Carl Zeiss B.V.). Where possible, colocalization was shown in a color-blind-friendly manner using photoshop to copy red or blue signals into both the red and blue channel to produce magenta.
Western Blotting
Proteins were extracted from 0–3 h embryo collections as previously described [64] and subjected to Western blot analysis according to standard procedures using the Bio-Rad electrophoresis system (Bio-Rad) (see Text S1 for details).
Analysis of md Neurons
We have analyzed the morphology of the solitary type 4 md neuron in the ventral cluster called vdaB (ventral dendritic arborization neuron B) [18] . For the visualization of dendritic arbors we used the type 4 md neuron-specific driver 477-Gal4 [20] to drive expression of memGFP. Details of crosses, confocal microscopy, and quantification of dendrite ends are provided in Text S1 .
Analysis of Larval Locomotory Behavior
Larval crawling was assayed as described previously (refer to Text S1 for details) [25] . Approximately 30 individuals per strain were tested per day over a 5-d period, resulting in a total of approximately 150 larvae per genotype and experiment. Experiments were performed at least twice. Quantification of path lengths was performed using Adobe Photoshop and Image J.
Light-Off Jump Reflex Habituation
Male flies of the genotypes EHMT + /Y , EHMT DD1 /Y , EHMT DD2 /Y , and female flies of the genotypes EHMT + /EHMT + and EHMT DD1 /EHMT DD2 were tested for light-off jump reflex habituation in a modified assay that was previously described by Engel and Wu [27] . Details of this high throughput assay are described Text S1 .
Courtship Conditioning
Flies were tested for learning and/or memory at 4 d of age using the courtship conditioning assay as previously described [53] . For induced EHMT expression via hs-Gal4 , flies were incubated at 37 degrees for 45 min on days 1–3, with the final heat shock treatment taking place 24 h before training and testing on day 4.
ChIP-seq
Chromatin immunoprecipitation was performed using standard methods with anti-H3K9me2 antibodies (07-441, Upstate) and Prot A/G beads (Santa Cruz) to capture antibody bound chromatin (for details see Text S1 ). Massive-parallel sequencing was performed using the Illumina Genome Analyzer IIx according to standard protocols of the manufacturer (Illumina) (for details see Text S1 ). All sequence analyses were conducted using the BDGP Release 5 genome assembly (DM3) and the release 5.12 annotations provided by FlyBase. To compensate for differences in sequencing depth and mapping efficiency among the two ChIP-seq samples, the total number of unique tags of each sample was uniformly equalized relative to the sample with the lowest number of tags (7,043,913 tags), allowing for quantitative comparisons. For association of individual bins with genes, we determined the distances from the middle of the bin to the nearest tss or polyA site using the Pinkthing tool ( http://pinkthing.cmbi.ru.nl/cgi-bin/index50.pl ). The ChIP-Seq data from this study are available at the NCBI Gene Expression Omnibus ( http://www.ncbi.nlm.nih.gov/geo/ ) under series accession no. GSE22447.
Mircroarray Expression Analysis
Total RNA was isolated in triplicate from third instar larvae using the RNeasy Lipid Tissue Midi kit (Qiagen). RNA quality was evaluated using spectrophotometry and integrity was confirmed using gel electrophoresis of glyoxal-denatured samples. Total RNA samples were labeled using the ‘indirect’ method [65] . Superscript II reverse transcriptase (Invitrogen) was used to produce cDNA incorporated with aminoallyl-dUTP (Fermentas). Reactive fluorescent dyes (Alexa647 or Alexa555; Invitrogen) were conjugated to the individual samples. Two differently labeled samples, whole larvae from EHMT mutant versus EMHT + wild-type, were pooled and co-hybridized to the 14K long oligo array from the Canadian Drosophila Microarray Centre ( www.flyarrays.com ) according to previously described methods [66] . Images of the hybridized microarrays were obtained using a ScanArray 4000 scanner (Perkin-Elmer) and were quantified using QuantArray 3.0 software (Perkin-Elmer). Data were normalized using lowess sub-grid normalization using Genetraffic Duo (Stratagene) analysis software. Normalized data were exported and analyzed using the one-class test available in the Statistical Analysis of Microarrays (SAM) software package. The false discovery rate of the one-class test was adjusted such that the expected number of false positive results was less than one. Gene lists generated in SAM were filtered to include only those genes that displayed at least a 2.5-fold increase or decrease in abundance with respect to the wild-type sample and whose coefficient of variance was less than 100%.
Statistical Analysis
For all data, normal versus non-normal distribution was assessed using the Shapiro-Wilk test and by visual examination of histograms. For comparison of more than two variants with a normal distribution, one-way ANOVA analysis was used to determine the probability that there were differences between the variants. In the cases that ANOVA indicated that there was a significant differences between variants ( p <0.05) we performed post hoc pair-wise comparisons using the Bonferroni correction, which takes into account that multiple comparisons are being made and therefore increases the stringency of the test. This method was applied for normally distributed data in Figures 3c , 5c , S2e, S2f , S3 , and S4b . For comparison of more than two variants with a non-normal distribution, the Kruskal-Wallis test was used to determine if there were significant differences between any of the means. For data sets in which there was a significant difference ( p <0.05), we subsequently performed pair-wise comparisons using the Mann-Whitney test, a post hoc test that can be used to compare two means with non-normal distributions. The combination of Kruskal-Wallice and Mann-Whitney was used for data sets which were not normally distributed ( Figures 4b–4c and 5d–5f ). All of the above tests were performed using SPSS software (SPSS Inc.).
To test whether the number of LOMBs in a given genomic position (e.g. gene body) was significantly enriched compared to the distribution of all bins in the genome, we applied a hypergeometric test using an online tool ( http://stattrek.com/Tables/Hypergeometric.aspx ). This test was performed for all genomic regions defined in Figure 6c —Upstream >1 kb, Upstream <1 kb, Gene Body, Downstream <1 kb, Downstream >1 kb, and Distant—to obtain individual p values. Since we performed this test for six genomic regions, p values were corrected using the Bonferroni method to account for multiple comparisons using the same data set.
Enrichment of gene ontolology terms was analyzed using GOToolBox [36] to perform a hypergeometric test with Benjamini & Hochberg correction. | Results
EHMT/G9a Is the Solitary Drosophila Ortholog of EHMT1 and EHMT2/G9a
We set out to study EHMT / G9a in Drosophila . In contrast to mammals, which have two EHMT genes, flies possess a single ortholog [6] , [10] that we will subsequently refer to as EHMT throughout our manuscript. Phylogenetic analysis of the EHMT protein family in Drosophila , human, and mouse shows that Drosophila EHMT is equally similar to both EHMT1 and EHMT2/G9a ( Figure 1a ).
EHMT Is Expressed in the Fly Nervous System
We first examined expression and subcellular localization of the EHMT protein in the fly nervous system using an anti-EHMT antibody [10] . In the adult brain EHMT staining is widely abundant, in a pattern resembling nuclear DAPI staining ( Figure 1b ). Analysis at single cell resolution in the ventral nerve cord of third instar larvae demonstrates that EHMT is localized in the nuclei of neurons as revealed by colocalization with the neuronal nuclear marker elav ( Figure 1c ) [16] . Weaker EHMT staining colocalized with repo ( Figure 1d ), a nuclear glial marker [17] . EHMT staining was also observed in the nuclei of the larval multiple dendrite (md) sensory neurons of the peripheral nervous system labeled using the 109(2)80-Gal4 driver [18] to express memGFP ( Figure 1e ). In addition, EHMT staining was observed at low levels in non-neuronal tissues such as the muscle and epidermis ( Figure S1c ). The anti-EHMT immunolabeling is specific, as it is absent in EHMT mutant embryos, md neurons, adult brains, and larval body-walls ( Figures 2c and S1 ). These data reveal that EHMT is widely expressed in the Drosophila nervous system.
Generation of EHMT Mutant Flies
In order to uncover the functions of EHMT we generated deletions in the EHMT gene by excision of a P-element, KG01242 , located in the 5′ UTR. We screened 80 independent excision lines and identified two downstream deletions (DD) resulting from imperfect excisions of KG01242 . Both deletion strains are viable to adulthood, which is consistent with a viable EHMT knock-out allele generated by homologous recombination in Drosophila [19] . EHMT DD1 and EHMT DD2 lack 870 and 1473 base pairs of DNA downstream of the original P-element insertion site, respectively, including the EHMT translational start site ( Figure 2a ). We also isolated a precise transposon excision line that represents the same genetic background as our deletion lines and served as a control in all subsequent experiments (referred to as EHMT + ). Western blot analysis revealed a band of the expected size (180 kDa) in EHMT + embryonic protein extracts, which was absent in extracts from both deletion lines ( Figure 2b ). No extra bands were detected by the C-terminally directed EHMT antibody [10] that would point to expression of an N-terminally truncated protein. EHMT protein was also undetectable by immunohistochemistry in EHMT mutant embryos, md neurons, and adult brains, while showing a nuclear staining pattern in EHMT + animals ( Figures 1 , 2c , and S1 ). Expression of the neighbor gene, CG3038 , was not affected by the deletions ( Figure S2 ). These data show that EHMT DD1 and EHMT DD2 are strong and specific loss of function mutants, most likely complete null alleles.
EHMT Regulates Dendrite Branching in Type 4 Multiple Dendrite Neurons
Next, we examined several aspects of neuronal development in EHMT mutant flies. Analysis of adult mushroom body architecture, synaptic morphology of the larval neuromuscular junction, and adult photoreceptor function (assessed by electroretinography) ( Figure S3 ) as well as analysis of embryonic nervous system integrity (unpublished data) did not reveal significant differences in mutant versus control conditions, indicating that general nervous system development and neuronal function is not affected.
Loss of EHMT did however result in altered dendrite development in multidendrite (md) neurons, which are sensory neurons that tile the larval body wall. We specifically examined dendritic arbors of type 4 md neurons, which were highlighted using the 477-Gal4 driver [20] and the UAS-Gal4 system [21] ( Figure 3a and 3b ). These neurons are highly stereotyped in their number, position, and morphology, thus allowing for quantitative analysis of dendritic arbors of identical neurons in different animals and genotypes. While the basic organization of these arbors is maintained in EHMT mutants (primary branches labeled I, II, III, and IV in Figure 3a and 3b ), reduction of higher order branching resulted in dendritic fields of appreciably reduced complexity ( Figure 3a, 3a' versus 3b, 3b' ). We quantitatively assessed this defect by counting the number of dendrite ends per standardized field of view in stacked confocal images. This analysis confirmed that EHMT DD1 and EHMT DD2 had a consistent and statistically significant decrease in the total number of dendrite ends, showing 16 and 17.5 percent reduction, respectively, when compared to EHMT + ( Figure 3c ).
To address whether this phenotype results cell-autonomously from EHMT deficiency in neurons, we generated UAS-EHMT transgenic flies and performed cell-specific rescue experiments ( Figure 3c ). Re-expression of EHMT in mutant type 4 md neurons (using 477-Gal4 ) did indeed rescue dendrite branching towards wild-type levels ( Figure 3c , red and orange bars). This reversal is specific, since expression of EHMT in the EHMT + genetic background did not increase branching ( Figure 3c , black bars). Rather, EHMT overexpression appeared to reduce dendrite branching as compared to controls expressing Gal4 and GFP in the absence of UAS-EHMT , although this reduction was not statistically significant ( Figure 3c , black bars). These data show that EHMT is cell-autonomously required in type 4 md neurons to establish normal dendrite complexity.
EHMT Affects Larval Locomotory Behavior
Drosophila md neurons are important in the regulation of larval locomotion behavior [22] – [24] . We therefore examined larval locomotory patterns during the early third instar using an established larval foraging assay ( Figure 4 ) [25] . Larval crawling paths were analyzed for total path length over a 5 min period and for specific crawling patterns, such as branched versus straight paths. The total path length covered by foraging larvae was not different between mutants and EHMT + controls ( Figure 4b ), indicating that crawling ability is not hindered in these larvae. However, striking differences in larval locomotory patterns were observed between mutant and wild-type. Foraging EHMT mutant larvae often stop, retract, and turn, causing increased branching in their crawling paths ( Figure 4a ). Quantitative analysis of the length of the resulting side branches revealed an increase of approximately 4-fold and 2-fold, respectively, in EHMT DD1 and EHMT DD2 ( Figure 4c ). Thus, the dendrite phenotype of EHMT mutant larvae is associated with an altered crawling behavior. In contrast, other innate behaviors, such as adult phototaxis and negative geotaxis, were normal in EHMT mutants ( Figure S4 ). To address whether the dendrite phenotype of type 4 md neurons alone is sufficient to cause the abnormal crawling pattern, we attempted to rescue this phenotype by re-expression of UAS-EHMT in type 4 md neurons using 477-Gal4 . This was not sufficient to restore normal larval locomotor behavior, indicating that dendritic defects in type 4 md neurons and abnormal locomotory behavior might arise independently.
EHMT Regulates Habituation, a Form of Non-Associative Learning
Next, we analyzed the role of Drosophila EHMT in learning. Habituation is a form of non-associative learning where an initial response to a repeated stimulus gradually wanes [26] . In the light-off jump reflex habituation assay [27] flies were exposed to a sudden light-off pulse and measured for a jump response over the course of 100 trials with a 1 s inter-trial interval. Figure 5a and 5b show the proportion of flies that do show a jump response over the course of 100 trials. Hemizygous EHMT mutant males (genotypes: EHMT DD1 /Y and EHMT DD2 /Y ) and transheterozygous EHMT mutant females (genotype: EHMT DD1 /EHMT DD2 ) both displayed a drastically slower response decrement during the habituation procedure as compared to wild-type EHMT + flies ( Figure 5a and 5b ). Individual flies were deemed to have habituated when they failed to jump in five consecutive trials (no-jump criterion). Habituation was scored as the number of trials required to reach the no-jump criterion (trials to criterion). The mean number of trials to criterion for mutants, EHMT DD1 /Y , EHMT DD2 /Y , and EHMT DD1 /EHMT DD2 , was significantly higher (12-, 8-, and 6-fold, respectively) than for EHMT + wild-type flies ( p <0.001) ( Figure 5c ). These experiments establish a role for EHMT in regulating non-associative learning.
EHMT Is Required for Courtship Memory
Having established a role for EHMT in habituation, a simple learning process, we asked whether EHMT is also involved in more complex forms of learning and/or memory using the courtship conditioning paradigm. This assay is based on the conditioning of male courtship behavior by exposure to a non-receptive female, which in presence of normal learning and memory capacities results in suppression of courtship [28] . Male flies were paired with a non-receptive pre-mated female for appropriate time intervals (see Experimental Procedures) and tested for courtship suppression immediately following the training period, after 30 min or after 24 h, to assess learning, short-, and long-term memory, respectively. The mean Courtship Index (CI, the percentage of time spent on courtship during a 10 min interval) of trained males and of socially naïve males was assessed to calculate a Learning Index (LI), which is defined as the percent reduction in mean courtship activity in trained males compared to naïve males; LI = (CI naive − CI trained )/CI naive . We found that EHMT mutant flies are perfectly capable of this form of learning, as they efficiently suppressed courtship immediately following the training period ( Figure 5d ). Strikingly, the Learning Index of EHMT DD1 males was reduced by 50% at 30 min after training (STM-short term memory), and even more dramatically, to 17% of the wild-type value at 24 h after training (LTM-long term memory) ( Figure 5d ). These results indicate that EHMT is dispensable for courtship learning but necessary for both short- and long-term courtship memory.
The Requirement for EHMT-Based Courtship Memory Maps to 7B-Gal4 Positive Neurons
To provide evidence for the specificity of the courtship conditioning phenotype and to roughly map where EHMT is required to control learning and memory in this paradigm, we performed rescue experiments in the EHMT DD2 background using tissue specific expression of UAS-EHMT and short-term memory (30 min after training) as a read-out. The elav-Gal4 driver was used to express EHMT in all neurons, and the 7B-Gal4 promoter for more selective expression. Indeed, pan-neuronal expression of EHMT in the mutant background restored the Learning Index to normal levels ( Figure 5e , orange bars, pan neuronal versus EHMT mutants), providing evidence that EHMT is required cell-autonomously in neurons to achieve normal memory. Elav-driven expression of EHMT in the EHMT + genetic background had no significant effect on Learning Index ( Figure 5e , black bars, pan neuronal versus EHMT mutant). 7B-Gal4 is predominantly expressed in the mushroom bodies of adult brains but is also expressed and at lower levels in some other brain regions, including the antennal lobe ( Figure S5 ) [29] . Expression of EHMT with this driver in the EHMT mutant background was able to rescue the Learning Index ( Figure 5e , orange bars, 7B-Gal4 versus EHMT mutant), revealing that EHMT function in 7B-Gal4 neurons is sufficient for normal memory. We also observed that overexpression of EHMT using 7B-Gal4 in the EHMT + background significantly reduced the Learning Index ( Figure 5e , black bars, 7B-Gal4 versus EHMT mutants). Since the Learning Index was normal in EHMT mutants containing both 7B-Gal4 and UAS-EHMT , we conclude that there is no deleterious effect due to the expression of Gal4 or the 7B-Gal4 P-element insertion itself. We therefore asked whether the presence of endogenous EHMT could make a significant difference to the absolute protein levels in the mushroom body upon 7B-Gal4 -mediated overexpression. We observe a very high and uniform EHMT staining in all mushroom body Kenyon cells upon UAS-EHMT expression with 7B-Gal4 in the EHMT + background ( Figure S6 ). A similar staining pattern was observed using this driver in the EHMT mutant background, although staining intensity was noticeably lower, likely due to the absence of endogenous EHMT ( Figure S6 ). In contrast, the elav-Gal4 driver resulted in a non-uniform staining pattern, with high EHMT levels in only a small proportion of Kenyon cells ( Figure S6 ). Thus, overexpression of EHMT in 7B-Gal4 neurons appears to be deleterious when above a certain threshold. These results suggest that appropriate levels of EHMT in the Drosophila nervous system are critical for courtship memory and indicate that the requirement for EHMT in this process is confined to 7B-Gal4 positive neurons.
Taken together with the defect in jump reflex habituation, these data reveal an important role for EHMT not only in a simple form of learning but also in a more complex cognitive process such as courtship memory.
EHMT-Mediated Memory Can Be Rescued in Adulthood
Recently, it has been reported that postnatal loss of Ehmt1 and G9a in mice causes cognitive defects in the absence of obvious developmental abnormalities [15] . We therefore asked whether the memory defects of EHMT mutants in the courtship conditioning paradigm can be rescued by expression of EHMT in adulthood. Indeed, induced expression of EHMT using hs-Gal4 after eclosion (see Experimental Procedures) completely restored memory defects shown by siblings of the same genotype that had not undergone the heat-shock procedure ( Figure 5f ). This demonstrates that EHMT is required for memory in adult flies and highlights that cognitive defects are reversible in EHMT mutant flies.
Generation of Genome Wide H3K9me2 Methylation Profiles
The reversible memory defects in EHMT mutant flies suggest a critical role for EHMT in neuronal function in addition to its role in dendrite development. We therefore wanted to determine the molecular mechanisms through which EHMT regulates neuronal processes. EHMT proteins mediate histone 3 lysine 9 dimethylation (H3K9me2) in euchromatic regions of the mammalian and fly genomes [6] , [7] , [10] . Therefore, we investigated EHMT target sites by generating genome-wide H3K9me2 profiles for EHMT mutant and wild-type larvae using chromatin immunoprecipitation (ChIP) with an H3K9me2 antibody followed by massive parallel sequencing of the co-precipitated DNA (ChIP-seq technology). Mapping of the sequenced tags to the Drosophila genome revealed a genome-wide profile that is consistent with known H3K9me2 patterns [30] , [31] . High H3K9me2 is a known characteristic of heterochromatin [31] . Accordingly, we find high H3K9me2 levels in both wild-type and EHMT mutant strains in annotated heterochromatic regions that are contiguous with the assembled euchromatic chromosome arms (Chr2Lh, Chr2Rh, Chr3Lh, Chr3Rh, and ChrXh) ( Figure S7 ) [32] . This was expected, since EHMT is known to have no effect on heterochromatin formation and heterochromatic H3K9me2 levels are known to be unaffected by loss of EHMT/G9a in fly and mouse [7] , [15] , [19] , [33] . The generated H3K9me2 profiles also follow expected patterns in euchromatin. H3K9me2 is known to dip immediately before the transcriptional start site (tss) and near the polyadenylation site (polyA) of genes [30] , [34] either due to nucleosome depletion or decreased H3K9me2 in these regions. We indeed observe a dip in H3K9 dimethylation in these regions ( Figure 6d and 6e, left panel s, black lines), thus demonstrating the accuracy and reliability of our H3K9me2 profiles.
EHMT Affects H3K9me2 Levels in Discrete Regions of the Euchromatic Genome
Since the global pattern of H3K9me2 appeared to be normal in EHMT mutants, we reasoned that EHMT must affect discrete regions within the genome. To identify these regions we divided the euchromatic genome into 300 bp bins and compared the number of sequenced tags per bin in wild-type versus mutant samples. For each of the 384,944 bins in the euchromatic genome we calculated a methylation ratio by dividing the number of tags in wild-type by the number of tags in the mutant. Thus, a ratio greater than 1 identifies regions where methylation is decreased in EHMT mutant flies. We have plotted the log of these ratios (log(2)wt/mt) in a histogram, in which Loss of Methylation Bins (LOMBs) are found in the area of positive log values. The histogram roughly follows a normal distribution but is asymmetric, with 19,258 bins outside two-times the standard deviation of the mean on the positive side, while only 50 bins outside two-times the standard deviation on the negative side ( Figure 6a ). The 19,258 LOMBs constitute about 5% of the euchromatic genome and provide an unbiased confirmation for the role of EHMT in H3K9 dimethylation. Loss of methylation (LOM) can also be visualized in the USCS genome browser as areas where H3K9me2 levels are depleted in the mutant but remain high in wild-type (two examples given in Figure 6b ). Interestingly, we find that LOMBs are not randomly distributed in the genome but are enriched in the areas 1 kb upstream of the tss and 1 kb downstream of the polyA site by 1.6-fold and 3.3-fold, respectively ( Figure 6c ). As mentioned above, we observe a local depletion of H3K9me2 in these regions in wild-type animals ( Figures 6d and 6e, left panel , black line). In EHMT mutants, this local depletion is strongly augmented both upstream of the tss and near the polyA site ( Figures 6d and 6e, left panel , orange line), providing further evidence that EHMT deposits H3K9me2 marks in discrete euchromatic loci, with a bias towards the 5′ and 3′ ends of genes.
Loss of H3K9me2 in EHMT Mutants Can Affect Gene Transcription
H3K9me2 is a marker for condensed, transcriptionally repressive chromatin [35] , however the modification itself does not strongly correlate with transcription levels on a genome wide scale as is seen for some other histone modifications, like H3K4me3 and H3K27me3 [30] . To determine whether H3K9me2 can contribute to transcriptional repression in Drosophila , we performed microarray expression analysis to compare mRNA levels in EHMT wild-type and mutant larvae. We then analyzed H3K9me2 levels upstream of the tss and downstream of the polyA site for genes that were up- and downregulated in EHMT mutants. Genes that are activated by EHMT (i.e. greater that 2.5-fold downregulated in mutants, Table S1 ) showed no difference in H3K9me2 profiles upstream of the tss or downstream of the polyA site when comparing EHMT wild-type and mutant strains ( Figure 6d and 6e , middle panels). In contrast, genes that are repressed by EHMT (i.e. greater that 2.5-fold upregulated in mutants, Table S2 ) have a clearly augmented dip in methylation both at the tss and polyA sites ( Figure 6d and 6e, right panel s) when compared to the wild-type profile and to the average methylation profiles of all genes. These data indicate that EHMT-mediated H3K9 dimethylation immediately up and downstream of genes can affect transcriptional repression in Drosophila.
EHMT Target Loci
Next, we investigated which genes were affected by loss of methylation in EHMT mutants by associating each LOMB with its closest gene. In total, LOMBs were found in or near 5,136 genes; 1,229 genes had LOMBs upstream of the tss (upstream LOMB) and 1,712 genes had LOMBs downstream of the polyA site (downstream LOMB) ( Table S3 ). The two groups overlap by 255 genes ( Figure 7a ).
To assess the function of LOMB-associated genes, we analyzed their gene ontology for enrichment of specific biological processes using GOToolBox [36] . Genes associated with LOMBS are highly enriched in terms related to the nervous system ( Figure 7b , for lists of genes associated with selected terms see Table S4 ). The broad term nervous system development, associated to more than 350 LOMB genes, reaches the highly significant p value of 2.3×10 −28 and is the most enriched tissue-specific term. Consequently, the term is highly depleted from the pool of genes with unaltered H3K9me2 in EHMT mutants; i.e. in genes that are not associated with LOMBs ( Figure 7b , no LOMBs). Strikingly, all GO terms that describe EHMT mutant phenotypes (e.g. short- and long-term memory, non-associative learning, dendrite morphogenesis, and larval locomotory behavior) show significant enrichment when considering all LOMB-associated genes and genes associated with downstream LOMBs ( Figure 7b , observed phenotypes). Other neuronal terms that show high enrichment are also shown ( Figure 7b , neuronal terms). Signal transduction is also amongst the most strongly enriched terms, with a p value of 6.2×10 −48 . Many specific signaling pathway terms are highly overrepresented amongst LOMB-associated genes, with G-protein coupled receptor protein signaling pathway and small GTPase mediated signal transduction being the top terms ( Figure 7b , signaling pathways). We also note significant enrichment of pathway terms that directly relate to EHMT mutant phenotypes, such as cAMP signaling, a major pathway involved in learning and memory.
Notably, there is a stark contrast in enriched terms when comparing genes associated with either upstream or downstream LOMBs ( Figure 7b , Downstream LOMBs and Upstream LOMBs). Downstream LOMBs are associated with genes that are enriched for neuronal terms, signaling pathways, and terms describing observed EHMT mutant phenotypes, while upstream LOMBs are associated with genes involved in biological processes requiring a high transcriptional plasticity, such as stress response and actin cytoskeleton remodeling ( Figure 7b , enriched in upstream LOMBs). The contrast between these two groups in their gene ontology illustrates the importance of H3K9me2 position at target gene loci and provides further support as to the biological relevance of these data.
Finally, genes involved in regulatory processes such as translation, chromatin assembly/disassembly, and chromosome organization are highly depleted from LOMB-associated genes ( Figure 7b , depleted), which contrasts the striking enrichment of nervous system and phenotype-relevant terms amongst LOMB-associated genes. | Discussion
Here, we demonstrate that Drosophila EHMT, a histone methyltransferase, regulates sensory dendrite development, larval locomotory behavior, simple learning (habituation), and complex memory (courtship conditioning). Notably, EHMT mutants are viable, appear healthy, and many other aspects of neuronal development and function are normal, highlighting the selectivity with which EHMT regulates specific aspects of neuronal development and function. Genome-wide molecular analysis of EHMT mutant flies supports this idea, revealing altered histone methylation at target loci encompassing a selection of neuronal genes that control learning, memory, and other phenotype-relevant processes.
EHMT Affects H3K9me2 at Discrete Regions in the Euchromatic Genome
The EHMTs are an evolutionarily conserved family of proteins that regulate H3K9 methylation at euchromatic DNA [6] , [7] , [10] , [15] . Previous studies have shown that EHMTs affect transcription through H3K9 dimethylation in the promoters of certain genes [8] , [37] – [39] . Our study provides the first genome-wide overview of EHMT function with respect to its role in post-translational histone modifications. We provide evidence that Drosophila EHMT induces H3K9 dimethylation at a proportion (about 5%) of the euchromatic genome, with a preference for discrete regions at the 5′ and 3′ ends of genes ( Figure 6 ). Genes with differential H3K9me2 levels at the 5′ end (within 1 kb upstream of the transcriptional start site) are predominantly involved in biological processes related to stress response (e.g. heat shock response and actin cytoskeleton remodeling), which require rapid and frequent changes in transcription. This observation is consistent with studies in yeast and humans, which show that chromatin structure immediately upstream of transcriptional start sites directly correlates with transcriptional plasticity [40] . In contrast, genes that are differentially methylated at the 3′ end are highly enriched in genes that control neuronal processes that are disrupted in EHMT mutants ( Figure 7 ). The general view is that gene expression is regulated through interactions at the promoter, or 5′ end. However, recent studies have revealed that 3′ gene ends also play an important and complex role in the regulation of transcription by: (1) mediating gene looping [41] – [45] , which is necessary for transcriptional memory, i.e. the altered transcriptional responsiveness of genes after a previous cycle of activation and repression [43] , [46] , [47] ; (2) serving as an initiation site for antisense transcripts [41] ; and (3) regulating transcript termination, a process that also affects transcript levels [48] . Currently there is no evidence linking H3K9me2 to any of these processes, however it is conceivable that differential histone methylation at the 3′ end of neuronal genes may act as a mechanism to control their expression. In line with this idea a recent study has reported that the DNA methyltransferase, Dnmt3a, also targets neuronal genes in “non-promoter” regions, including 3′ ends [49] . Thus, it appears that epigenetic alterations to non-promoter regions is emerging as a general theme for the regulation of neuronal gene expression.
Regulation of Dendrite Development and Locomotory Behavior by EHMT
EHMT mutants show a decrease in dendrite branching in sensory neurons of the Drosophila peripheral nervous system ( Figure 3 ). Type 4 md neurons are known to provide the sensory input that they receive via their dendrites as an essential functional component to the neuronal circuitry governing larval movement [23] . Our analysis of larval locomotion in EHMT mutants revealed a behavioral phenotype characterized by an increased performance of stops, retractions, and turns ( Figure 4 ). It has been reported that such a phenotype can directly arise from dysfunction of type 4 md neurons [22] , [24] , which raised the possibility that decreased dendrite branching and altered locomotory behavior are connected traits. Re-expression of EHMT in type-4 md neurons did, however, not rescue larval locomotion defects, suggesting that larval locomotion and type 4 md neuron development are controlled independently by EHMT. Thus, this lack of rescue may be due to requirements for EHMT in additional peripheral or central neurons relevant to the crawling pattern. We can also not exclude unspecific secondary effects or that precise levels of re-expressed EHMT may be crucial for turning behavior. Ultimately, the relevance of EHMT in both dendrite development and crawling is illustrated by the observation that EHMT mutants show loss of H3K9me2 at 65 of 147 genes annotated to be involved in dendrite development and 15 of 16 genes involved in larval locomotory behavior (see Table S4 for gene IDs).
Regulation of Learning and Memory by EHMT
We have shown that EHMT is required for light-off jump reflex habituation ( Figure 5a–5c ), a simple form of non-associative learning that is known to require classic learning and memory genes such as rutabaga [27] . In this paradigm a sequence of leg extension and flight initiation is induced by sudden darkness. This behavioral response is mediated by the giant fiber interneurons, which receive sensory input from the visual system in the brain and relay this information through the thoracic ganglion where efferent neurons descending from the giant fiber to thoracic muscles are stimulated [50] , [51] . Only a few genes are known to control jump reflex habituation and most of these are ion channels, or are involved in cAMP and cGMP second messenger signaling pathways [50] . EHMT is the first histone modifying enzyme to be implicated in this simple form of learning. Jump-reflex habituation is not an official gene ontology term, but significantly, seven of the eight genes known to be involved in jump-reflex habituation [50] show loss of H3K9 dimethylation in EHMT mutants ( Table S4 ).
We have also identified a role for EHMT in courtship memory ( Figure 5d–5f ). This is a complex form of memory that allows male flies to discriminate between receptive and non-receptive females, presumably to optimize the energy that they spend on courtship. We demonstrate that loss of EHMT leads to impaired short- and long-term memory while the learning capacity of the EHMT mutants was unaffected ( Figure 5d ). Moreover, we show that normal courtship memory is restored upon re-expression of EHMT in the whole nervous system and in a subset of neurons labeled by 7B-Gal4 , which is predominantly expressed in the mushroom body neurons of the adult brain ( Figure 5e ). Although further work is required to map the specific circuits required for EHMT-dependent courtship memory, the mushroom body is known to be crucial for courtship memory, but not learning [52] , pointing towards a deficit in this area of the brain. Significantly, EHMT affects histone methylation in 22 of 36 genes that were annotated at the time of our analysis to be involved in memory ( Table S4 ). Other relevant memory genes, such as Orb2 [53] ( Figure 6b ), nemy [54] , and ben [55] , that were not yet included in gene ontology databases are also affected by loss of EHMT . Together, these data suggest that EHMT targets two-thirds of all currently known memory genes.
Importantly, we were able to fully restore memory deficits by re-expression of EHMT during adulthood ( Figure 5f ). Thus, although EHMT can affect neuronal hardwiring (dendrite development in the peripheral nervous system; Figure 3 ), it appears that adult cognitive defects do not arise from neurodevelopmental defects occurring prior to eclosion. This is consistent with a recently reported impairment in fear conditioning that has been observed in mice with postnatal loss of Ehmt1 in the brain [15] and with our observation that mushroom body morphology appears unaffected in EHMT mutant flies. Thus, EHMT-mediated H3K9 dimethylation of specific loci is required in adult post-mitotic neurons to consolidate or retrieve consolidated memories. Interestingly, other epigenetic regulators, such as the DNA methyltransferases Dnmt1 and Dnmt3a, are also required in post-mitotic neurons for normal memory [56] . These studies support the idea that the process of learning induces reprogramming of the neuronal epigenome, which crucially underlies memory [1] – [3] . Such “stable” chromatin modifications, including DNA and histone methylation, appear to be good candidates for “writing” long-term memory, however these marks must also remain dynamic allowing for memories to be modified. Our understanding of this stable versus dynamic state of epigenetics in neurons and its consequences are highly limited. It will thus be important to dissect the extent of epigenetic plasticity during the different phases of learning, memory consolidation, and memory retrieval, and to determine how these alterations to the epigenetic landscape translate into transcriptional changes required for information processing and storage.
A recent study of mRNA levels in mice with brain region-specific loss of Ehmt1 has identified 56 genes that are consistently misregulated in the mutant mouse brain [15] . Of these 56 genes, 18 are non-neuronal, which led to the interpretation that EHMT proteins control cognition through repression of non-neuronal genes in neuronal tissues. In contrast to this view, our data show that Drosophila EHMT mediates H3K9 dimethylation at more than 350 neuronal gene loci with proven critical roles in nervous system development and function. Does this apparent discrepancy reflect evolutionary differences? Of the 56 differentially expressed genes identified by Schaefer et al. [15] , 30 are conserved in flies and 20 show loss of H3K9me2 in EHMT mutants ( Table S4 ). This correlation is very unlikely to occur by chance ( p <2.9×10 −4 ; hypergeometric test), suggesting that EHMT target genes are, at least in part, evolutionarily conserved. The great number of highly enriched neuronal genes amongst Drosophila EHMT targets, their striking match with EHMT mutant phenotypes, and the reversibility of cognitive defects argue that EHMT orchestrates an epigenetic program that directly regulates a battery of neuronal players underlying the molecular basis of cognition. It is also noteworthy that EHMT targets include fly orthologs of NF1 , FMR1 , FMR2 , CNTNAP2 , GDI , DLG3 , and of many more genes underlying syndromic and non-syndromic forms of intellectual disability. Also, the major signaling pathways known to underlie intellectual disability, Rho and Ras GTPase pathways [57] , [58] , are highly enriched in our ontology analysis (GO term: small GTPase mediated signal transduction).
Our study complements a number of reports on post-embryonic rescue of cognitive phenotypes in disease models of intensively studied disorders such as Fragile X syndrome, Neurofibromatosis I, Tuberous sclerosis, Rubinstein-Taybi, Angelman, and Rett syndrome [59] . The growing number of such examples provides an argument for reappraisal of the traditional view that genetic forms of intellectual disability are largely due to irreversible neurodevelopmental defects, findings which open prospects for therapeutic intervention. Currently, clinical trials are underway to treat Fragile X patients with compounds that have initially been identified to rescue phenotypes in fly models of Fragile X syndrome [60] – [63] . The EHMT mutant fly has provided novel insights into the epigenetic regulation of cognition and will be a valuable tool to work further towards such translational approaches. Furthermore, a better understanding of the epigenetic mechanisms regulating cognitive processes is relevant to the wider medical community, considering the increased awareness of the epigenetic contributions to neurodevelopmental and psychiatric disorders in general [4] , [5] . | The author(s) have made the following declarations about their contributions: Conceived and designed the experiments: JMK HvB AS. Performed the experiments: JMK KK MAWO DK ZA JTW. Analyzed the data: JMK HM EKDJ ZA JTW HZ AS. Contributed reagents/materials/analysis tools: ZA JTW HGS MBS KK HZ. Wrote the paper: JMK HvB AS.
Behavioral phenotyping and genome-wide profiling of the histone modifier EHMT in Drosophila reveals a mechanism through which an epigenetic writer may control cognition.
The epigenetic modification of chromatin structure and its effect on complex neuronal processes like learning and memory is an emerging field in neuroscience. However, little is known about the “writers” of the neuronal epigenome and how they lay down the basis for proper cognition. Here, we have dissected the neuronal function of the Drosophila euchromatin histone methyltransferase (EHMT), a member of a conserved protein family that methylates histone 3 at lysine 9 (H3K9). EHMT is widely expressed in the nervous system and other tissues, yet EHMT mutant flies are viable. Neurodevelopmental and behavioral analyses identified EHMT as a regulator of peripheral dendrite development, larval locomotor behavior, non-associative learning, and courtship memory. The requirement for EHMT in memory was mapped to 7B-Gal4 positive cells, which are, in adult brains, predominantly mushroom body neurons. Moreover, memory was restored by EHMT re-expression during adulthood, indicating that cognitive defects are reversible in EHMT mutants. To uncover the underlying molecular mechanisms, we generated genome-wide H3K9 dimethylation profiles by ChIP-seq. Loss of H3K9 dimethylation in EHMT mutants occurs at 5% of the euchromatic genome and is enriched at the 5′ and 3′ ends of distinct classes of genes that control neuronal and behavioral processes that are corrupted in EHMT mutants. Our study identifies Drosophila EHMT as a key regulator of cognition that orchestrates an epigenetic program featuring classic learning and memory genes. Our findings are relevant to the pathophysiological mechanisms underlying Kleefstra Syndrome, a severe form of intellectual disability caused by mutations in human EHMT1 , and have potential therapeutic implications. Our work thus provides novel insights into the epigenetic control of cognition in health and disease.
Author Summary
Epigenetic regulators can affect gene transcription through modification of DNA and histones, which together form chromatin. The importance of such regulators for cognition is increasingly appreciated, but only few key factors have been identified so far. Excellent candidates are histone modifiers that are involved in intellectual disability, such as EHMT1, implicated in Kleefstra Syndrome. Here, we characterized the neuronal function of EHMT in Drosophila . Flies that lack EHMT are viable but show highly selective defects in specific aspects of neuronal development and function, including learning and memory. Genome-wide analysis of EHMT-mediated histone methylation revealed that EHMT targets the majority of all currently known Drosophila learning and memory genes. It also targets genes known to be involved in the other aspects of behavior and neuronal development that are compromised in EHMT mutants. Remarkably, EHMT mutant memory deficits can be reversed in adulthood, suggesting that epigenetic influences on cognition are not always permanent. Our results provide novel insights into the epigenetic control of cognition in health and disease. | Supporting Information | We thank the Drosophila Genomic Resource Center, the Bloomington Drosophila stock center, A. Lambertson, and the Developmental Studies Hybridoma Bank for providing constructs, fly strains, and antibodies. We are grateful to R. Habets and P. Verstreken for help with electroretinography; to K. J. Françoijs, S. van Heeringen, E. Janssen-Megens, and Y. Tan for help with sequencing and sequence analysis; and to D. Kasimer and Z. Razak for help with microarrays and microarray analysis.
Abbreviations
Courtship Index
downstream deletion
euchromatin histone methyltransferase
histone 3 at lysine 9
Learning Index
loss of methylation bin
long term memory
short term memory | CC BY | no | 2022-01-13 00:00:18 | PLoS Biol. 2011 Jan 4; 9(1):e1000569 | oa_package/05/5e/PMC3014924.tar.gz |
|
PMC3014925 | 21245905 | Introduction
In-vitro fertilisation (IVF) is now widely used for the treatment of infertility, and validated age-stratified national success rates and outcomes are published annually [1] , [2] , [3] . To facilitate patient counselling, clinical decision-making, and access to health care provision, prediction models for live birth after IVF have been constructed [4] . However, these studies have been limited by their sample size, development before the introduction of intracytoplasmic sperm injection (ICSI), or lack of validation in external populations [5] , [6] , [7] , [8] , [9] . Established multivariable prediction models may therefore not be applicable to contemporary couples seeking treatment. Consequently, clinicians and regulatory bodies have not adopted prediction models and predominantly quote age-related success rates [1] , [2] , [3] .
Given the known complications with multiple gestations and prematurity, the focus has moved to defining the most appropriate IVF outcome variable as a singleton term live birth [10] , [11] , [12] . Low birth weight and macrosomia are also known to be associated with immediate and long-term risk to offspring heath [13] , and IVF singletons are at increased risk of these complications [14] , [15] . It is now recognised that factors leading to infertility may be responsible for adverse perinatal outcome rather than the process itself [16] , [17] , [18] , [19] ; however, which parental characteristics of infertile couples contribute to adverse perinatal outcomes in IVF singletons and can thereby be targeted for intervention remain unknown.
In this prospective cohort study of 144,018 treatment cycles we assessed the extent to which baseline characteristics can be used to predict live birth after IVF-assisted conception, and for those cycles in which a singleton pregnancy was achieved we identified which factors were associated with preterm delivery, low birth weight, and macrosomia. | Methods
Source of Data
The UK Human Fertilisation and Embryology Authority (HFEA), which is responsible for the regulation of assisted conception treatment in the UK, has had a Parliamentary statutory obligation to prospectively collect baseline information and birth outcomes on all licensed fertility treatment cycles performed in the UK since 1991 [7] . All treatment cycles and outcomes registered on the HFEA database between January 2003 and December 2007 were used in our study, with the final analysis cohort details and exclusion criteria provided in Figure 1 [20] . Treatment cycles that were for storage or donation of gametes, were not IVF, or were frozen embryo transfers were excluded. Although there is a move to greater use of frozen embryo cycles we excluded these from our analyses to be consistent with previous publications, including that by Templeton et al. [7] in which the established model was developed. Furthermore, during the time studied very few elective single embryo transfers were performed (<0.05% of all cycles). HFEA data relating to treatments between April 1999 and March 2002 were not verified by licensed treatment centres and are therefore deemed less accurate. Furthermore, few treatment cycles had treatment with ICSI before and during this period. Whilst data have been collected beyond 2007, validation checks on the computerised data undertaken by HFEA are currently complete only to December 2007. Anonymised data were provided by the HFEA per cycle of treatment rather than for individual women, so our outcomes are expressed as rates and/or odds per cycle of treatment (rather than per individual woman). Ethical approval of the study was provided by the HFEA.
Measurements
Maternal age, duration and cause of infertility, previous number of IVF attempts, number of previous spontaneous and IVF live births, source of gametes, and cycle number were recorded at the time of treatment. Duration of fertility, number of previous IVF attempts, number of previous pregnancies, number of previous IVF pregnancies, and total number of previous live births were all categorised in accordance with the previous analysis by Templeton [7] . Cycle number was collapsed, with more than three cycles as one category, because of small numbers. Live birth was defined as a baby born alive after 24 wk gestation. Our main outcome was at least one live birth, which was defined as any birth event in which at least one baby was born alive and survived for more than 1 mo. This outcome is consistent with previous publications, including that by Templeton et al. [7] used to define the established prediction model. In a sensitivity analysis we repeated associations with this main outcome after exclusion of multiple pregnancies, defined as those in which two or more fetal heartbeats were noted at 8 wk gestation.
For assessment of perinatal outcomes in cycles with a singleton live birth, gestational age at birth was defined as completed weeks of gestation. For the main outcome of preterm we examined multivariable associations with preterm birth, defined as ≤36 completed weeks; we also examined associations with extreme preterm (<33 wk). Birth-weight outcomes were supplied in 500 g increments and categorised as low birth weight (LBW <2.5 kg), normal (≥2.5 to <4.0 kg) or macrosomic (≥4 kg). In these analyses we included only cycles in which there was one heartbeat at 8 wk gestation and one live birth (i.e. these were singleton live births).
Statistical Methods
We performed univariable and multivariable logistic regression to assess associations with at least one live birth. Given that the historical multivariable Templeton model [7] had been externally validated we first tested its predictive ability. We used the reported characteristics that were associated with live birth and their respective regression coefficients from that model to generate the probability of live birth in our cohort (see Text S1 for full details of these calculations) [7] . The predictive ability of the model was assessed by determining the discrimination, using the area under the curve of receiver operator characteristics (AUROC), and its calibration. Calibration was assessed by ranking participants into tenths based on their predicted risk for the Templeton prediction model, and then within each tenth comparing the predicted mean rate to the observed rate of live birth.
The multivariable logistic regression model formed the basis of our novel prediction model of live birth. In the novel prediction model we used the same characteristics as those used in the Templeton model but included all causes of infertility (Templeton includes only tubal versus all other causes), and allowed the coefficients for this and all other variables to be newly derived, and included four additional characteristics—the source of the egg (donor or patient's own), type of hormonal preparation used (antioestrogen, gonadotrophin, or hormone replacement therapy), whether or not ICSI was used, and the number of the treatment cycle (1, 2 or ≥3). We tested all two-way interactions between pairs of predictors included in our multivariable analyses and used a Bonferroni-corrected (for multiple testing) p -value threshold of 0.05 to define statistical evidence of an interaction. The discrimination and calibration of this novel model was assessed as described above. The AUROC between the Templeton model and our model was compared using the ROCCOMP command in Stata [21] . When we repeated the multivariable analyses using 1,000 bootstrap replications, the estimates and their standard errors were essentially the same and results are presented here without bootstrapping.
Lastly, we examined model reclassification by determining the integrated discrimination improvement (IDI) of the novel prediction model compared to the original Templeton model [7] . The IDI is a summary measure of the extent to which a new prediction model increases risk prediction in individuals who ultimately have the outcome of interest [22] , and reduces risk prediction in those who remain healthy in comparison to the established risk prediction model (in this case the Templeton model [7] ).
To explore risk factors for adverse perinatal outcomes (preterm, extreme preterm, low birth weight, and macrosomia) we used logistic regression to examine the univariable and independent multivariable associations of all risk factors assessed in the multivariable analyses of at least one live birth, as described above. The selection of these potential risk factors for adverse perinatal outcomes was based on previous studies and the plausibility that risk factors that influence odds of live birth are also likely to affect gestational age and birth weight. For associations with preterm as the outcome we additionally adjusted for mean birth weight, and for outcomes with low birth weight and macrosomia we adjusted for mean gestational age. These analyses were conducted only for cycles in which there was only one heartbeat at 8 wk gestation and at least one live birth.
All statistical analyses were performed using Stata version 11 (StataCorp LP).
Dealing with Missing Data
For the vast majority of variables there was no missing data; 3.9% of cycles had missing data on method of hormonal preparation used and 8.4% had missing data on duration of infertility; overall 12% of the eligible cohort had some missing data ( Figure 1 and Table S1 ). Univariable associations were very similar when maximum numbers for each variable were used ( Table S2 ) and when only those with complete data were used ( Table 1 ), suggesting that missing data did not result in bias. | Results
Figure 1 shows how we established the eligible cohort of IVF treatment cycles ( N = 163,425) and the sample used for the main multivariable analyses (i.e. without any missing data N = 144,018; 88% of eligible). Table S1 shows the study characteristics. Amongst the 163,425 eligible cohort, the overall rate of at least one live birth was 23.4 per 100 cycles (95% CI 23.2–23.7). Rates of successful live birth increased linearly over time from 22.7 per 100 cycles in 2003 to 24.9 per 100 cycles in 2007 ( p <0.001 for linear trend) ( Figure S1 ).
Table 1 shows univariable and multivariable associations of live birth. The odds of successful live birth decreased with increasing maternal age, increasing duration of infertility, greater number of previously unsuccessful IVF treatments, when the woman's own egg (as opposed to donor) was used, and when this was the second or third (as opposed to first) treatment cycle. Odds of successful live birth were lower when the cause of infertility was tubal, anovulatory, or cervical disease or when it was due to a male cause. Women who had at least one previous live birth (either natural or with IVF) had increased odds of a successful live birth with this cycle, as did those in whom gonadotrophin or hormone replacement (as opposed to antioestrogens) were used and ICSI was used with IVF. A previous IVF live birth increased the odds of future success (OR 1.58, 95% CI 1.46–1.71) more than previous spontaneous live birth (OR 1.19, 95% CI 0.99–1.24); p -value for difference in estimate <0.001 (estimated using 1,000 bootstrap replications to estimate standard errors of differences between the log odds between the two regression coefficients).
There was statistical evidence for four interactions, and stratified analyses reflecting these interactions are shown in Table S3 (for interactions with age) and Table S4 (for interactions with ICSI). The increased odds of success in cycles in which the duration of infertility was less than one year increased with increasing maternal age, though only a very small proportion of all cycles were in the category of less than one year duration of infertility. The reduced odds of successful outcome amongst own versus donor oocytes strengthened with increasing age. In couples who had not used ICSI all three causes—male infertility, infertility due to cervical disorders, and infertility due to a combination of causes—were associated with reduced odds of live birth, whereas there were no such associations in those using ICSI. Requiring three or more treatment cycles was associated with reduced odds of live birth in those in which ICSI was used, but not where it was not used. These four interactions were included in our novel prediction model, which is described in Text S2 .
Table 2 shows the AUROC for each of the Templeton and our new prediction models, with statistically significant improvement in discrimination for our novel model. Figure 2 and Table 3 show the observed to predicted rate of successful live birth by tenths of the distribution of the linear prediction models for each of the models. Calibration was poor with the Templeton model, which markedly underestimated the likelihood of successful live birth across the entire distribution of risk, particularly in those at lowest risk. By contrast the novel model had excellent calibration and reclassified cycle probability of a live birth in a way that improved upon the original Templeton model (IDI = 2.1%, p <0.001 comparing the novel model to the Templeton [7] ).
Of the 144,018 cycles 9931 (7%) were multiple pregnancies (i.e. had two or more fetal heart beats noted at 8 wk gestation). Of these, 1,264 (13%) resulted in one live birth, 7,925 (80%) in two live births, and 109 (1%) in three live births; 633 (6%) did not result in a live birth. When we removed these 9,931 cycles from our analyses results were essentially unchanged from those presented here. For example, Table S5 shows the univariable and multivariable associations of potential predictors with live birth after these exclusions (i.e., the equivalent of Table 1 in this paper). The AUROC, observed to predicted ratios and IDI were the same as those presented in Tables 2 and 3 with these exclusions.
Table 4 provides examples of how our novel prediction model could be used in clinical practice to give an estimate of a couple's probability of achieving a live birth in a given cycle of treatment. This illustrates not only the clinical use of this model (which we have developed into a freely available computer programme, http://www.IVFpredict.com , and iPhone/Android application, IVFpredict) but also how both couple characteristics and treatment choice influence prognosis.
Of the 144,018 cycles included in our main analyses for prediction of successful live birth, there were 24,226 live singleton births; 24,096 (99.5%) of these had gestational age data and 24,050 (99.3%) had birth weight data. Mean (SD) gestational age was 38.98 (2.12) completed weeks, with 472 (2.0%) being less than 33 wk, 1,598 (6.6%) between 33–36 weeks and 22,026 (91.4%) 37 or greater weeks. Mean (SD) birth weight was 3.277 (0.629) kg, with 2,100 (8.7%) low birth weight (<2.5 kg), 19,704 (81.9%) a healthy birth weight, and 2,246 (9.3%) macrosomic (≥4.0 kg). Table S6 shows the univariable associations of risk factors preterm (<37 completed weeks), low birth weight (<2.5 kg) and macrosomia (≥4.0 kg). Table 5 shows the multivariable associations of these risk factors with preterm birth, low birth weight, and macrosomia. In multivariable analyses the odds of both preterm birth and low birth weight in singleton IVF live births were reduced when the woman's (rather than donor) egg was used and when ICSI was used. The odds of low birth weight were also reduced with increasing maternal age and with a history of previous pregnancy (either spontaneous or following IVF). Odds of macrosomia increased with increasing maternal age and in cycles in which there was history of a previous pregnancy (either spontaneous or IVF). Odds of all three factors—preterm birth, low birth weight, and macrosomia—were increased when infertility was due to a cervical disorder. Tubal causes of infertility were associated with increased odds of preterm birth, anovulatory causes with low birth weight, and male causes of infertility with macrosomia. Table S7 shows univariable and multivariable associations with extreme preterm birth (<33 wk; n = 472). Characteristics that were associated with preterm birth in general were also associated with extreme preterm birth. In addition, extreme preterm birth was lower in 38- to 39-year-olds compared to all other ages and was increased in those who had a previous history of IVF. | Discussion
In this study we identify precise estimates of the strength and independence of the factors affecting the odds of IVF success and their association with adverse perinatal outcome. To date, successful prediction of live birth after assisted conception has been limited, with a recent systematic review [4] finding that models were limited by their sample size, incorporating fewer than 3,100 cycles or couples and their lack of external validation. The notable exception was the model of Templeton et al., which analysed 36,961 treatment cycles undertaken in the UK between 1991 and 1994 and was validated in a population of 1,253 couples receiving IVF treatment in The Netherlands between 1991 and 1999 [7] , [23] . Since then, ICSI for male factor infertility has been widely adopted, and consequently we demonstrate that this previously validated model, although showing reasonable discrimination, is poorly calibrated and of limited use in contemporary populations. We have developed a new model, which encompasses a series of new measures including use of donor oocytes, ICSI, cycle number, and whether there had been a previous spontaneous or IVF-related live birth or fetal loss. Using this novel model we can statistically significantly improve the overall prediction of live birth as assessed by area under the curve and attain excellent calibration with accurate identification of couples with a poor, moderate, or good prognosis. We also find that maternal characteristics, in particular maternal age, source of the oocyte and cervical causes of infertility are strongly associated with the risk of low birth weight and preterm delivery in singleton live births resulting from IVF. Notably, some of these associations were in the opposite direction to those seen for successful live birth. Thus, in women who successfully have a singleton live birth with IVF, the risk of low birth weight is reduced in older compared with younger women and both low birth weight and preterm are reduced when the woman's own embryo has been used.
The use of assisted conception has increased dramatically over recent years, with concomitant increases in success rates, in part driven by the widespread uptake of ICSI for male factor infertility [24] . The importance of ICSI in general and in particular causes of infertility is demonstrated in our study by its association with increased odds of successful live birth and by the fact that couples with male causes of infertility, cervical causes, or combined causes have reduced odds of success if ICSI has not been used, but are unrelated to success if ICSI has been used. Recent technical advances have, however, failed to overcome the reduction in success rates associated with increasing duration of infertility, necessity for repeated IVF attempts, or increasing maternal age, all of which are independently associated with reduced odds of live birth. The detrimental impact of prolonged infertility suggests that early recourse to treatment is appropriate and that extended treatment waiting times, for example whilst trying lifestyle interventions, might militate against eventual success. The marked reduction in the success of the second cycle but then a relative plateau is in contrast to previous reports, which suggested a subtle decline with increasing cycle number [7] . This difference may indicate that previous declines in success rates with increasing cycle number principally reflected increasing maternal age, which we have adjusted for.
In keeping with all previous reports, live birth rates decline with increasing maternal age [2] , [4] , [7] , [23] . By contrast, ours is the first study that we are aware of to find that, in women with successful IVF delivery of a singleton live birth, younger maternal age is associated with increased risk of low birth weight. This latter finding is however, in keeping with the recent observation that maternal age is positively associated with first trimester growth [25] , which if impaired is an important determinant of later adverse perinatal outcome [26] , [27] . For older women, the use of donor oocytes is a successful strategy for the attainment of a live birth, however, we identify that donor oocyte recipients have a marked increase in the risk of delivering a preterm or low birth weight infant. This may reflect the primary relationship between ovarian senescence and vascular function. Premature and natural menopause have both been associated with widespread vascular dysfunction, dyslipidaemia, a proinflammatory phenotype, and an increased risk of cardiovascular events [28] , [29] , [30] . These same factors have been implicated in the aetiology of fetal growth restriction and preeclampsia, the major determinants of preterm birth [31] . Furthermore, increased incidence of these complications have been reported in young and old donated oocyte recipients [32] , [33] . With respect to macrosomia the associations with older maternal age may reflect higher maternal socioeconomic class due to deferred child bearing or increased maternal obesity, both of which would be contribute to improved fetal nutrition. Similarly, a previous successful pregnancy would be associated with potential maternal weight retention and thereby increased fetal weight in subsequent pregnancies [34] .
We examined the associates of preterm birth (<37 weeks the established definition of preterm). Although it is possible that obstetricians may consider IVF pregnancies as precious and have a lower threshold for iatrogenic preterm birth, we think this is unlikely because of the established associations of prematurity with neonatal respiratory complications [35] , [36] , and our finding that similar associations were also found for extreme preterm birth support this assumption. In the UK since 2005 only two embryos are allowed to be replaced under the age of 40 to reduce the risks of preterm birth and low birth weight, which are associated with multiple pregnancies. We have restricted the analysis of perinatal outcomes to delivery of a singleton pregnancy only because of the relevance of understanding risk factors associated with these outcomes in couples requiring IVF even when there is a singleton pregnancy. Few previous studies have examined the relationship of couple and treatment characteristics with gestational age and birth weight after live singleton IVF birth; our findings highlight important areas for further research aimed at maximising the success of IVF in terms of a healthy-weight, term live birth.
We demonstrate that a previous live birth as a consequence of IVF has an even greater effect on the prospect of successful assisted conception therapy than does previous spontaneous conception. Although many couples undergoing assisted conception feel encouraged by achieving a pregnancy, even if it subsequently results in fetal loss, we found no beneficial or negative effect of a history of a nonviable pregnancy on live birth. This suggests that embryonic chromosomal errors, rather than a defective maternal environment, may be primarily responsible.
Our work has a number of strengths. We have considered a range of anamnestic couple characteristics simultaneously with respect to validated live birth and perinatal outcomes rather than one or two in isolation. As a result, our data give a better overall reflection of predictive abilities, or lack thereof, for many factors. The size of our study was extremely large compared to other such studies in the literature. Finally, we considered a relevant multivariable historical model for consistency of findings before developing and assessing a novel prognostic model.
We acknowledge, however, a number of limitations. Data were not complete on 12% of the eligible cohort; however, univariable analysis was similar in the whole cohort, and multivariable multiple imputation did not alter the overall conclusions (results available from authors on request). Treatment cycles rather than individual patients were identified because of concerns regarding confidentiality and breach of the terms of the HFEA Act, and therefore it was not possible to examine the effect of multiple cycles within one patient or to use robust standard errors that take account of clustering of women. However, the previous HFEA analysis could account for clustering and did not show a significant effect as compared with per treatment cycle [7] . Maternal age was supplied in categories because of recent concerns over confidentiality; however, our findings of a decline in live birth rates with increasing age are in keeping with the previous analysis of the HFEA database and population reports [1] , [7] , [37] . We accept that the cause of infertility may have been underinvestigated or misreported [38] , although for male factors this was cross-validated with the use of ICSI, and for tubal disease our observed decrease in success rate is consistent with the control arm of randomised controlled trials of salpingectomy prior to IVF [39] , suggesting that these data are accurate.
For the main analyses with successful birth as the outcome we included both single and multiple pregnancy, i.e. our outcome was at least one live birth irrespective of whether there was one or more heart beats at 8 wk. Our reasons for doing this were, first, that this is a relevant outcome for infertile couples and, second, this was the outcome used in the study that developed the established prediction model, and therefore we wanted to test this model with the same outcome. Note that restriction of the data to pregnancies in which only one fetal sac was evident at 8 wk gestation produced similar results.
Finally, we acknowledge lack of external validation of our model. Nonetheless, we believe that this model will improve the ability to stratify contemporary couples seeking IVF on the basis of low, moderate, or high likelihood of success. This is of particular relevance to couples willing to consider all therapeutic options, including use of donor oocytes, as there is a 5-fold difference in live birth between the lowest and highest decile of our prediction model. To facilitate validation of the model we are currently generating a free web-based prediction tool ( http://www.IVFpredict.com ) and iPhone/Android application (IVFpredict) for widespread use of our new prediction tool. These will acknowledge the current lack of external validation and will request provision of anonymised data (all variables included in the prediction model, country of treatment, and outcome) that in the coming years we will use as a means of external validation of this model. We have included full model details in Text S2 thereby facilitating model validation by other research groups.
In conclusion, we show that baseline couple and treatment characteristics can provide a basis for counselling and informing couples of their likely prognosis in terms of low, moderate, or high odds of success (see Table 4 ). | ICMJE criteria for authorship read and met: SMN DAL. Agree with the manuscript's results and conclusions: SMN DAL. Designed the experiments/the study: SMN DAL. Analyzed the data: SMN DAL. Collected data/did experiments for the study: SMN. Wrote the first draft of the paper: SMN DAL. Contributed to the writing of the paper: SMN DAL.
Using the HFEA database of all 144,018 live births in all IVF cycles in the UK between 2003 and 2007, Scott Nelson and Debbie Lawlor show that couple- and treatment-specific factors can be used to help predict successful outcome following IVF.
Background
The extent to which baseline couple characteristics affect the probability of live birth and adverse perinatal outcomes after assisted conception is unknown.
Methods and Findings
We utilised the Human Fertilisation and Embryology Authority database to examine the predictors of live birth in all in vitro fertilisation (IVF) cycles undertaken in the UK between 2003 and 2007 ( n = 144,018). We examined the potential clinical utility of a validated model that pre-dated the introduction of intracytoplasmic sperm injection (ICSI) as compared to a novel model. For those treatment cycles that resulted in a live singleton birth ( n = 24,226), we determined the associates of potential risk factors with preterm birth, low birth weight, and macrosomia. The overall rate of at least one live birth was 23.4 per 100 cycles (95% confidence interval [CI] 23.2–23.7). In multivariable models the odds of at least one live birth decreased with increasing maternal age, increasing duration of infertility, a greater number of previously unsuccessful IVF treatments, use of own oocytes, necessity for a second or third treatment cycle, or if it was not unexplained infertility. The association of own versus donor oocyte with reduced odds of live birth strengthened with increasing age of the mother. A previous IVF live birth increased the odds of future success (OR 1.58, 95% CI 1.46–1.71) more than that of a previous spontaneous live birth (OR 1.19, 95% CI 0.99–1.24); p -value for difference in estimate <0.001. Use of ICSI increased the odds of live birth, and male causes of infertility were associated with reduced odds of live birth only in couples who had not received ICSI. Prediction of live birth was feasible with moderate discrimination and excellent calibration; calibration was markedly improved in the novel compared to the established model. Preterm birth and low birth weight were increased if oocyte donation was required and ICSI was not used. Risk of macrosomia increased with advancing maternal age and a history of previous live births. Infertility due to cervical problems was associated with increased odds of all three outcomes—preterm birth, low birth weight, and macrosomia.
Conclusions
Pending external validation, our results show that couple- and treatment-specific factors can be used to provide infertile couples with an accurate assessment of whether they have low or high risk of a successful outcome following IVF.
Please see later in the article for the Editors' Summary
Editors' Summary
Background
Worldwide, more than 10% of couples are infertile. Sometimes there is no obvious reason for a couple's inability to have children but, for many couples, problems with their eggs or sperm prevent “fertilization”—the union of an egg and a sperm that leads, eventually, to the birth of a baby. Until recently, little could be done to help infertile couples. Then, on the 25 July 1978, the world's first “test-tube baby” was born. Since then, 4 million babies have been born through in vitro fertilization (IVF). In IVF, mature eggs are collected from the woman (or from an egg donor if the woman cannot make her own eggs) after a course of special hormones, and they are mixed in a dish with her partner's sperm. If her partner has a low sperm count or abnormal sperm, a single sperm can be injected directly into the egg in a procedure called intracytoplasmic sperm injection (ICSI), which became widely available in the mid 1990s, or sperm from a donor can be used. Finally, a number (depending on the country) of embryos (eggs that have begun to divide and develop) are put back into the woman where, hopefully, they will establish a successful pregnancy.
Why Was This Study Done?
Not every attempt at IVF is successful. In the US and the UK, IVF is successful in about a third of women under 35 years old but in only 5%–10% of women over the age of 40. It would be useful to have a way to predict the likelihood of a live birth after IVF for individual couples. Such a “prediction model” would facilitate patient counseling, clinical decision making, and the allocation of IVF resources. In this study, the researchers use information on IVF cycles collected by the Human Fertilisation and Embryology Authority (HFEA), which regulates IVF in the UK, to assess the extent to which the characteristics of infertile couples and the treatment they receive can be used to predict live birth after IVF. They also use these data to identify which factors are associated with preterm delivery, low birthweight, and macrosomia (the birth of an unusually large baby), three undesirable birth characteristics.
What Did the Researchers Do and Find?
Between 2003 and 2007, 163,425 IVF cycles were completed in the UK, 23.4% of which resulted in at least one live birth. The researchers used the data collected by the HFEA on 144,018 of these cycles (the other cycles had missing data) to develop a multivariable logistic regression prediction model (a type of statistical model) for the outcome of IVF. According to this model, a decreased chance of at least one live birth was associated with several factors including increasing maternal age, increasing duration of infertility, and the use of the woman's own oocytes. By contrast, a previous IVF live birth and the use of ICSI were associated with increased chances of success. Importantly, compared with an established multivariable prediction model, which was developed before the introduction of ICSI, the researchers' new prediction model predicted the chance of a live birth following IVF with greater accuracy. Finally, the researchers report that the chances of preterm and low birthweight after IVF were increased if donor eggs were required and ICSI was not used, that an increased risk of macrosomia was associated with increasing maternal age and with a history of previous live births, and that all three undesirable birth characteristics were associated with infertility due to cervical problems.
What Do These Findings Mean?
These findings indicate that couple- and treatment-specific factors can be used to provide infertile couples with an accurate assessment of whether they have a low or high chance of a successful outcome following IVF. The prediction model developed here provides a more accurate assessment of likely outcomes after IVF than a previously established model. Furthermore, because the new model considers the effect of ICSI on outcomes, it should be more useful in contemporary populations than the established model, which does not consider ICSI. However, before this new prediction model is used to guide clinical decisions and to counsel patients, it needs to be validated using independent IVF data. To facilitate the external validation of their model, the researchers are currently generating a free web-based prediction tool and iPhone application (IVFpredict).
Additional Information
Please access these websites via the online version of this summary at http://dx.doi.org/10.1371/journal.pmed.1000386 .
The Human Fertilisation and Embryology Authority provides information on IVF and IVF statistics for the UK
The UK National Health Service Choices website provides information for patients on infertility and on IVF
The American Pregnancy Association has information for patients on infertility and on IVF
MedlinePlus has links to further resources on infertility and IVF (in English and Spanish)
The history of the development of IVF is described on the Nobel Prize website
The prediction tool that was used in this study is at http://www.IVFpredict.com | Supporting Information | We thank the staff of the HFEA and the contributing UK clinics for the data.
Abbreviations
area under the curve of receiver operator characteristics
confidence interval
UK Human Fertilisation and Embryology Authority
intracytoplasmic sperm injection
integrated discrimination improvement
in vitro fertilisation | CC BY | no | 2022-01-13 03:10:32 | PLoS Med. 2011 Jan 4; 8(1):e1000386 | oa_package/f6/96/PMC3014925.tar.gz |
|
PMC3014926 | 21245906 | Introduction
Sensation in the rodent vibrissal system relies on active whisking for interactions with the environment [1] , [2] . Motor circuits control whisker movement, while sensory afferents collect information about contact with objects. Interactions between motor and sensory systems are necessary for object localization and identification [3] – [5] .
Ascending sensory and descending motor pathways interact at multiple levels including the brainstem [6] , thalamus [7] , and cortex [8] . Three areas in the cerebral cortex are activated by whisker stimulation. Primary somatosensory cortex (vS1) responds with short latencies [9] , whereas secondary somatosensory cortex (S2) and vibrissal motor cortex (vM1) respond 10–20 ms later [10] . These areas are also strongly interconnected in a bidirectional manner [8] , [11] .
In rodents, some of the cytoarchitectonic features of vM1, vS1, and S2 are area-specific, such as the presence of “barrels” in layer (L) 4 of vS1, and others are not, such as the presence of most cortical layers, including L1, L2/3, L5A, L5B, and L6 [12] . Here, to explore the synaptic organization of cortical circuits in these three areas, we used glutamate uncaging and laser scanning photostimulation (LSPS) to map the local sources of excitatory synaptic input to individual excitatory neurons in vM1, vS1, and S2. We recorded from postsynaptic neurons distributed across L2–6 (i.e., all the cortical layers that contain excitatory neurons) and, for each one, stimulated presynaptic neurons also distributed across L2–6. The collection of synaptic input maps for each area was analyzed to extract a laminar connectivity matrix representing the local pathways between excitatory neurons in each area [13] , [14] . These connectivity matrices provide a quantitative survey of the interlaminar organization of local excitatory networks in each of these three cortical areas. | Materials and Methods
Terminology for Cortical Axes
We use the term radial to refer to the axis defined by the apical dendrites of pyramidal neurons; this axis is approximately normal to the cortical surface. Normalized radial distance is along the radial axis, bounded by the pia and the white matter, where pia = 0 and the L6/white matter border = 1. Vertical is synonymous with radial . Horizontal , or lateral , refers to planes normal to the radial axis, approximately parallel to layers , or laminae ( Figure 6 ). Oblique refers to off-axis interlaminar connections.
Slice Preparation
Mice were decapitated at postnatal day 20–25 under isofluorane anesthesia, and the brain rapidly placed in ice cold choline solution (in mM: 110 choline chloride, 25 NaHCO 3 , 25 D-glucose, 11.6 sodium ascorbate, 7 MgCl 2 , 3.1 sodium pyruvate, 2.5 KCl, 1.25 NaH 2 PO 4 , 0.5 CaCl 2 ). Coronal brain slices (300 μm) were cut (Microm HM 650V), incubated 30 min at 37°C in oxygenated ACSF (in mM: 127 NaCl, 25 NaHCO 3 , 25 D-Glucose, 2.5 KCl, 2 CaCl 2 , 1.25 NaH 2 PO 4 , 1 MgCl 2 ), and maintained in a holding chamber at 22°C for up to 5 h during recording. For vM1 slices, the brain was pitched upward ∼10° to optimize alignment with the radial axis of vM1, and slices ∼0.7–1.3 mm anterior to bregma were used; for vS1 and S2 slices, the brain was cut coronally, and slices ∼1–2 mm posterior to bregma were used ( Figure 1A,B ). To determine the optimal slice angle for each area, we used the appearance of the intact apical trunk at high magnification to select slices for recording and avoided those sections where the apical dendritic trunk was at an angle with respect to the slice plane. Thus, only one or two sections per animal could be used for recording. Separate experiments in our laboratory using the photostimulation methods in vS1 [42] and vM1 (unpublished data) measure input to L1 dendrites of L5 pyramidal neurons, confirming the apical trunk is intact using this slice angle. We added biocytin to visualize a subset of dendritic arbors, some of which are reconstructed in Figure 3 and Figure 5 . These neurons appeared radially symmetric, with arbors ranging from 300–500 μm in diameter. Since the neurons were 50–100 μm deep in the slice, a portion of the apical and basal dendrites are truncated by slicing and the deep half of the arbor is intact.
Electrophysiology
Recordings were performed at room temperature (22°C) in ACSF. Neuronal excitability was reduced by increased divalent ions (4 mM CaCl 2 and 4 mM MgCl 2 ), and NMDA receptor blockade with 5 μM 3-((R)-2-carboxypiperazin-4-yl)-propyl-1-phosphonic acid (CPP; Tocris). Patch pipettes were fabricated from borosilicate glass with filament (4–6 MΩ). Intracellular solution contained (in mM): 128 K-gluconate, 10 HEPES, 10 sodium phosphocreatine, 4 MgCl 2 , 4 Na 2 ATP, 3 sodium L-ascorbate, 1 EGTA, and 0.4 Na 2 GTP (pH 7.25; 290 mOsm). To visualize dendritic arbors, 20 μM Alexa 594 or 488 (Molecular Probes) was added to the internal solution. In some cases, 2–3 mg/mL biocytin was included. Electrophysiological signals were amplified with an Axopatch 700B amplifier (Molecular Devices), filtered at 4 KHz, and digitized at 10 KHz. Data I/O were controlled by Ephus , a suite of custom Matlab-based (Mathworks) software tools available online ( https://openwiki.janelia.org [54] ).
Neurons were selected based on pyramidal appearance, or in the case of L4 recordings in vS1, either pyramidal or stellate appearance. In vS1, recordings were generally made in the middle of the barrel field and not a specific whisker barrel. Following patching, a family of current steps was presented to determine firing properties. Neurons with narrow APs and high firing rates were rejected for analysis as presumed interneurons.
Laser Scanning Photostimulation (LSPS)
Methods followed published procedures [13] , [26] . MNI-glutamate (0.2 mM; Tocris, MO [55] ) was added to a recirculating bath. Photolysis was performed by shuttering (1.0 ms pulse) the beam of an ultraviolet (355 nm) laser (DPSS Lasers, San Jose, CA), ∼20 mW in the specimen plane, set by a combination of a gradient neutral density filter wheel and Pockels cell (electro-optical modulator; Conoptics). A 16×16 standard stimulus grid for input maps had row and column spacing of 110 μm for vM1 and 90 μm for vS1 and S2 recordings. Maps were recorded in voltage clamp at −70 mV. Inhibitory input amplitude was minimized by recording near the chloride reversal potential. The 256 grid sites were visited in a sequence that optimized the spatiotemporal separation between sites [25] . The sequence was repeated 2–4 times per neuron. In vS1 and S2, the map was aligned to the top of the pia and centered on the soma. In vM1, the medial edge of the map was aligned to the interhemispheric fissure, and the top to the dorsal-most edge of pia.
To convert each map's set of traces into an array of pixels that represent response amplitudes, we calculated the average current over a 50 ms post-stimulus window. Direct dendritic responses were excluded on the basis of temporal windowing [56] , rejecting traces with events (detected as >3 SD above baseline) with onset latencies of <7 ms. At locations where some maps had direct responses at a given pixel while others did not, the average of the non-direct responses was used; pixels were excluded from the average raw input map for a given neuron if all traces had direct responses.
We measured excitation profiles using loose-seal recordings with the amplifier in voltage-follower mode, to gauge the efficacy of photostimulation for neurons in the different layers in the three areas. Excitation profiles were recorded and analyzed following previous methods [13] , [25] , [26] , [30] . To characterize the size of the excitation profile, we calculated the mean weighted distance from the soma of AP generating sites as: Σ(APs × distance from soma)/Σ(APs).
Connectivity Matrix Analysis
Procedures build on [13] . A transformation step was added, to account for cortical curvature, which was especially strong in vM1. As described in Results, we assigned each point in the stimulus grid a normalized radial distance and horizontal offset ( Figure 6 ). Individual recorded neurons were also assigned a postsynaptic radial distance based on the same criteria. Individual input maps for a given neuron could then be averaged together based on postsynaptic radial distance. Furthermore, when computing the input to a given neuron for the purpose of determining the connectivity matrix, a presynaptic point would be averaged into a bin appropriate to its location. Most aspects of local connectivity were robust to changes in binning. Subsequent corrections to the connectivity matrices were made to account for variations in excitability between layers, and the number of neurons in pre- and post-synaptic layers; these were then presented as neuron→neuron connectivity matrices and layer→layer connectivity matrices (see Results and Text S1 ).
Quantitative Comparisons of Connectivity
To make quantitative comparisons between the strength of pathways in different areas, we determined both the average strength of pathways and their variability using a bootstrap-based analysis ( Figure 9 ). After selecting the pre- and postsynaptic neuron populations by relative laminar depth, the strength of corresponding pixels in the input map (limited to maps from neurons in the postsynaptic layer, and pixels in the presynaptic layer within 300 μm horizontal distance of the soma) were averaged for each map. We resampled the individual map averages 10,000 times with replacement and resampled other factors contributing to the individual neuron→neuron strength (number of APs from cortical area's excitation profiles and neuron density). Pathways were presented with the average strength and SD from the bootstrap analysis.
Analysis of Cortical Lamination
In vS1 and S2, we performed morphometric measurements of cortical landmarks in video images of brain slices. Along a radial line, we marked the locations of the soma, pia, white matter, and major laminar boundaries and calculated the absolute and normalized radial distances of these locations. The bottom extent of cortex was marked at the border between L6 (including the subplate zone) and white matter [57] . The distances to lower borders of layers (±SD) are given in Table 1 . The division between L2 and L3 in vS1 was drawn between groups of neurons that did not receive appreciable L4 input (L2 [58] ) and those that did. Since this functional division was not clear in S2, L2/3 was divided in half. These values are bracketed in the table. vM1 appearance was similar to somatic motor cortex, with a prominent clear zone in the upper middle part of the cortex, corresponding to L5A [13] . Thus, landmarks indicating the border between L1, a compressed L2/3, and the bottom of L5A were apparent in video images and used to measure laminar boundaries in vM1. The division between L5B and L6 was estimated as the radial distance where cell density increased ( Figure S4 ; Table 1 ), as a clear border was not apparent based on image contrast. Alternative methods of determining cortical layers in motor and sensory cortex were performed on images of gene expression patterns from the Allen Brain Atlas ( Figure S8 ). | Results
Identification of Cortical Areas
We identified vibrissal motor cortex (vM1), primary somatosensory (barrel) cortex (vS1), and secondary somatosensory cortex (S2) based on anatomical coordinates, cytoarchitectonic features, anatomical labeling experiments, and in the case of vM1, optical microstimulation mapping. vM1 ( Figure 1A ) was located in the posteromedial part of frontal agranular cortex, anteromedial to the barrel cortex [10] , [15] – [17] . When anterograde tracers were injected into vM1, fluorescently labeled axons were observed in brainstem nuclei involved in whisker motor control ( Figure S1 ) [18] . Furthermore, microstimulation mapping using channelrhodopsin-2 (ChR2) revealed that vM1 had the lowest thresholds for whisker movements ( 14 ) [19] , [20] . vS1 ( Figure 1B ) was identified by the presence of cytoarchitectonic “barrels” in L4 [21] . S2 ( Figure 1C ) was located in dysgranular cortex, lateral to the barrel cortex [8] , [10] , [22] . Axons projected from vS1 to S2, and from S2 to vM1 and vS1 ( Figure S1 ). These experiments enabled us to target our mapping experiments to specific cortical locations corresponding to vM1, vS1, and S2.
Mapping Local Excitatory Pathways with Laser Scanning Photostimulation (LSPS)
We prepared coronal brain slices containing vM1, vS1, or S2 ( Figure 1A–C ) and used LSPS with glutamate uncaging [23] – [25] to map excitatory inputs to excitatory neurons ( Figure 1D–G ). We excited small clusters of neurons at each site in an array of locations while recording from individual excitatory neurons ( Figure 1D,E ), obtaining maps of local intracortical sources of excitatory input ( Figure 1F,G ).
To calibrate LSPS, we recorded in cell-attached mode from excitatory neurons, while uncaging glutamate on a grid around the cell ( Figure 2A,B ). The spatial distribution of action potentials (APs) evoked by uncaging (the “excitation profile”) provides a measure of the effective spatial resolution of photostimulation ( Figure 2A,B ). These data were used to estimate neuronal photoexcitability ( Figure 2C ) and the spatial resolution of LSPS ( Figure 2D ) for photostimulating neurons in different cortical layers and areas. Photostimulation-evoked APs always occurred in perisomatic regions ( Figure 2B , Figure S2 ) with short latencies, and almost always as singlets. Stimulation of strong synaptic pathways, such as L4→L3 in vS1, did not cause APs in the target location ( Figure S2 ), indicating that synaptic activity did not cause APs in neurons that were not directly photostimulated. Ultraviolet (UV) attenuation in scattering tissue causes photoexcitation to decline as a function of depth in the slice; consistent with this, excitation was not observed for neurons deeper than 100 μm ( Figure S3 ) [26] . The total number of neurons excited per stimulus, estimated from the excitation profiles and measured densities of neurons ( Figure 2 and Figure S4 ), was in the range 50–200, consistent with previous results [13] , [26] . Only a small fraction of these neurons were synaptically connected to the recorded postsynaptic neuron [27] .
An input map represents the aggregate functional synaptic connectivity between small clusters of presynaptic excitatory neurons at the stimulus locations and individual postsynaptic neurons. Pixels in input maps do not represent the strengths of unitary connections; rather, they measure average monosynaptic excitatory responses to a single uncaging event (see Text S1 , Equations 1–4) [26] : where ρ cell is the neuronal density at the point of uncaging (neurons/μm 3 ), V exc is the volume of excited neurons (μm 3 ), and S AP is number of APs fired per presynaptic neuron (AP/neuron). The average strength of a synaptic connection ( q con ) is calculated from equation (1). The collection of q con for different neuronal populations defined by laminar location is the basis of connectivity matrices. We first present the mapping data for each area in the more familiar form of average input maps. In subsequent sections we summarize connectivity in laminar connectivity matrices, which take into account the parameters in equation (1).
vM1 Maps
Unlike vS1, vM1 lacks a distinct granular L4. The superficial layers L2/3 and L5A are compressed, and deeper layers L5B and L6 are expanded, consistent with vM1's location at the crest of a cortical convexity [28] . In addition, L1 was thicker than in the other areas ( Table 1 ). Both superficial and deep L5 neurons had dense basal dendrites and a single apical dendrite extending to L1, and L6 neurons had apical dendrites that did not extend to L1; in some cases, these were inverted pyramids ( Figure 3A ).
We recorded from 95 excitatory neurons located in all layers (i.e., from upper L2 to lower L6) and mapped the local sources of excitatory synaptic input with LSPS using a stimulus grid that spanned vM1 ( Figure 3B ; Figure S5 ). We pooled neurons into groups by dividing the cortex into 10 equal distance bins; the top-most bin was empty, because L1 lacks excitatory neurons. We averaged the maps in each bin ( Figure 3C ). The strongest pathway was a descending projection, L2/3→ upper L5. Weaker ascending projections, within L5 and L5A→L2/3, were also found ( Figure 3C ). On average, neurons in the lower one-third (0.7–1.0) of vM1 showed weak inputs. However, individual neurons in this deeper range received strong inputs, but these tended to be spatially dispersed and sparse ( Figure S5 ).
vS1 Maps
We recorded from 80 excitatory neurons in vS1, using a different stimulus grid matched to the cortical thickness ( Figure 4 ; Figure S6 ). In vS1, laminar boundaries were distinct, allowing pooling of cytoarchitectonically defined groups for binning ( Table 1 ; Figure S8 ). The ascending L4→L3 pathway and the descending L2/3→L5 pathway were both prominent ( Figure 4 ; Figure S6 ). Similar to vM1, L6 neurons had relatively weak inputs (mainly from L4). L4 neurons also showed little intracortical interlaminar input [29] . In addition, we further distinguished sub-layers within L2/3 and L5B based on patterns of connectivity observed in the input maps. For example, L2 constituted a narrow superficial layer of neurons lacking strong input from L4, but with input from L5A [30] . Binning with a simple three-layer scheme (‘supragranular-granular-infragranular’; Figure 4 ) conveyed the main feedforward local excitatory connections in vS1.
S2 Maps
S2 abuts the lateral edge of vS1, where the barrel pattern terminates ( Figure S1 ). The cytoarchitectonic layers appeared similar in S2 and vS1, except that the cortex was thinner and L5A thicker. L4 included neurons with a sparse apical dendrite, and neurons lacking an apical dendrite ( Figure 5A ). L5 neurons had many basal dendrites and an apical dendrite that ramified in L1; L6 neurons' apical dendrites did not extend above L4.
We recorded input maps for 100 excitatory neurons in S2 ( Figure 5B,C ; Figure S7 ). Similar to vS1, an ascending pathway to more superficial layers (L4→L3) was present but was not the strongest projection. Instead, the descending projection L2/3→L5 was predominant. L5 also received substantial ascending input from L6.
Derivation of Connectivity Matrices
Connectivity matrices represent local circuits in a compact manner [13] , [14] , [27] , [31] , [32] . Each element ( i, j ) in the matrix ( W i,j ) corresponds to the strength of a connection ( q con ; Equation 1) from the j th presynaptic location (along the rows) to the i th postsynaptic location (along the column). Distance is measured in normalized units along the radial directions (pia, 0; white matter, 1). Because of the curvature of vM1 at the cortical flexure ( Figure 6A,B ), we converted map data from the coordinates of the slice image ( x, y ) to coordinates corresponding to an unfolded cortex ( h, r ), where h is the horizontal distance along the laminar contour and r is the distance along the radial axis. Figure 6 provides a graphical illustration of the process of converting the pixels in an input map from x-y coordinates ( Figure 6A ), using a spatial transform defined on the basis of the radial structure of the cortex ( Figure 6B ), into r-h coordinates ( Figure 6C,D ).
This approach allowed us furthermore to convert input maps into vectors, by averaging input across the horizontal dimension ( h ) at a given presynaptic radial distance ( r ) into bins ( Figure 6E–G ; a similar analysis in the horizontal dimension is given in Figure S9 ). This is identical to averaging along the rows of input maps, except that it takes into account the curvature of the cortex. One neuron's input vector ( Figure 6G ) thus represents the inputs to one neuron from different laminar locations; i.e., the horizontal dimension has been collapsed. Each neuron was also assigned a postsynaptic radial distance. This allowed us to group all the input vectors and then sort them by the postsynaptic neuron's depth in the cortex ( Figure 6H ). Stacking the vectors on top of each other, sorted by depth, provided a raw connectivity matrix, W raw ( r post , r pre ), describing connectivity between neurons at different locations along the radial axis ( Figure 6H , Figure 7A ). The rows in such a connectivity matrix represent synaptic input to a particular laminar location, and the columns represent synaptic output from that laminar location. Intralaminar connections lie along the main diagonal. We note that intralaminar connectivity was undersampled because of direct excitation of the postsynaptic neurons' dendrites.
In addition to deriving matrices based on the collections of input vectors ( Figure 7A,D,G ), we further analyzed the data in terms of the excitation parameters given in equation (1). To compute the average connectivity matrix at the level of individual neurons ( W neuron ), we binned the data and applied correction factors to derive the strength of input per presynaptic neuron per AP. We divided the connection strength in the raw connectivity matrix by the mean number of APs per uncaging event at the presynaptic region ( Figure 7B,E,H ; Figure S10 ; Text S1 ) and the number of presynaptic neurons stimulated. The number of stimulated neurons was obtained from measurements of ρ cell ( Figure S4 ) and V exc . To compute the connectivity matrix at the level of cortical layers ( W layer ) we multiplied the neuron→neuron connections by the number of presynaptic and postsynaptic neurons per layer ( Figure 7C,F,I ; a detailed calculation is illustrated in Figures S11 and S12 ). Values for all connectivity matrices are provided in Table S1 and Dataset S1 . | Discussion
We used glutamate uncaging and LSPS to map local synaptic connections among excitatory neurons in mouse vM1, vS1, and S2, three cortical areas centrally involved in vibrissa-based somatosensation. From single cell input maps recorded at different cortical depths, we derived connectivity matrices that compactly describe the local network. Our main findings were that vM1 contains a strong pathway from L2/3 to upper L5; that vS1 and S2 contain two strong pathways, corresponding to L4→L3 and L2/3→L5; and that S2 contains these plus pathways between L6 and L5B.
Connectivity Matrix Descriptions of Cortical Circuits: Neuron→Neuron and Layer→Layer
The connectivity matrix description allows us to directly contrast local circuits in different cortical regions. The elements (pixels) in the neuron→neuron connectivity matrices, W neuron ( Figure 7B,E,H ), represent the mean strength of postsynaptic response in a single neuron extrapolated to a single presynaptic AP in a single cell of the indicated layer ( q con ). Pixel values were 10–100 times lower than typical unitary EPSCs, reflecting both the generally low probability of connections between excitatory neurons in cortical circuits (typically 0.1–0.2) [27] , [33] – [35] , and the fact that the current amplitude in the maps represents a mean over 50 ms rather than the peak of the EPSC.
In contrast, the elements in the layer→layer connectivity matrices, W layer ( Figure 7C,F,I ), represent the average strength of connections extrapolated to the entire projection from one layer to another. The W layer matrices differ from the W neuron matrices in that they enhance thicker and more neuron-dense layers and diminish thinner and less neuron-dense layers. For example, because in vS1 the L5A is thin ( Table 1 ) and both L5A and L5B are low in neuronal density ( Figure S4 ), the projections to and from L5, such as L5A→L2/3 and L2/3→L5B, are relatively strong at the level of neuron→neuron connectivity ( Figure 7E ) but relatively weak at the level of layer→layer connectivity ( Figure 7F ). Interestingly, in rat vS1 the L4→L2/3 projection is functionally weak compared to the structural density of L4 axons and L2/3 dendrites, while the converse holds for the L5A→L2/3 projection [36] . Our results here show how weak neuron→neuron connections may be strong in aggregate at the layer→layer level. Further structure-function analyses will be required to determine whether it is generally the case that larger and more neuron-dense layers have weaker neuron→neuron but stronger layer→layer projections.
Major Features of Connectivity Matrices in the Three Areas
The connectivity matrix representations of vM1 show strong descending projections from L2/3→upper L5 ( Figure 7A–C ), similar to the forelimb area of mouse M1 [13] , [14] , [37] . This input straddled the L5A/B border. L5B received an additional hotspot from itself, which appeared strong when considered as an entire layer ( Figure 7C ). The deepest one-third of vM1 (consisting mostly of L6) had weak inputs and outputs.
The vS1 excitatory circuits were more complex ( Figure 7D–F ). The major ascending pathway from L4→L3 was paralleled by an ascending component from L5A. The high cell density in L4 made the L4→L3 connection prominent in the laminar analysis ( Figure 7F ). Another prominent projection was from L2 and L3 to L5A and L5B; inputs originating in more superficial regions of L2/3 targeted relatively more superficial regions of L5A/L5B (note the diagonal shape of the L2/3→L5 hotspot in Figure 7E ). On a neuron→neuron basis, the L3→L5B connection was stronger than L4→L3, although the layer→layer analysis showed a reduction in cell density relative to L4. L2 received input from L3, and weaker input from L5A. However, L2 was thin and thus contributed little to W layer . As in vM1, deep layers had weak inputs and outputs.
In S2 ( Figure 7G–I ), an ascending L4→L2/3 pathway and descending L3→L5 pathway were present. Neurons on the L5A/L5B border also showed strong intralaminar connections. The L6 output evident in the input maps ( Figure 5 ; Figure S7 ) also supplied potent input to L5B. Although not as strong at the single cell level, the entire L6 excited L5B as much as L3 ( Figure 7I ). L6 was enhanced in S2 relative to other regions as both a source of synaptic output and a recipient of synaptic input, due to the relatively high density of neurons ( Figure S4 ) and their relatively low photoexcitability ( Figure 2C–E ). The functional connectivity in the local excitatory circuits of all three regions is simplified into quantitative laminar wiring diagrams ( Figure 8 ).
Limitations in the Derivation of Connectivity Matrices
LSPS with glutamate uncaging simultaneously excites a group of presynaptic neurons, while the postsynaptic response is measured. To derive average connection strength per neuron ( q con ), the number of excited neurons needs to be estimated, based on the excitability ( S AP ), neuron density ( ρ cell ), and excitation volume ( V exc ) at the uncaging location (Equation 1). The accuracy of the estimate of q con is limited by our measurement of ρ cell and neuronal excitation ( S AP , V exc ; Text S1 Equations 3–4): Measurements of neuronal density vary by a factor of two [27] , [38] , [39] . Although excitation profiles give a direct measure of evoked APs in brain slices under the relevant recording conditions ( Figure 2 ), excitation varies across neurons and somewhat across cortical areas, and decreases with depth in the slice; these effects together introduce uncertainty roughly on the order of a factor of two ( Figure S3 ). Despite these uncertainties, our estimates of q con are broadly consistent with those derived from pair recordings ( Figure S13 ).
Because LSPS excites many neurons, this strong stimulus allows weak pathways to be detected. However, the average connection strength, q con , reflects both the connection probability and unitary connection strength: It is therefore not possible to separate connection probability and unitary connection strength directly. Furthermore, p con is inversely related to the horizontal separation between cell pairs [34] . LSPS averages inputs from a range of presynaptic locations with varied horizontal offset. For each cell class, a broad distribution of p con values contributes to LSPS maps.
In addition, by computing the average connection strength, we average out the underlying distribution of unitary connection strength, which is a skewed distribution of numerous weak and a few strong connections [27] , [33] . This inherent averaging also makes LSPS insensitive to certain non-laminar aspects of cell-type specificity in cortical connectivity [14] , [33] , [35] , [40] – [43] .
Comparisons with Previous Studies of vS1 Connectivity
Comparison of our neuron→neuron connectivity matrix with a pair-recording study [27] reveals qualitative similarities ( Figure S13 ). After both methods are corrected to similar units (peak amplitude in pA/AP), the general shape of the connectivity matrix and values for neuron→neuron connectivity are similar. The major interlaminar pathways are L2/3→L5 and L4/5A→L2/3. However, local intralaminar connections are underestimated in our data set due to direct responses to uncaging. Furthermore, descending projections from L4→L5A and from L5A→L5B may be underestimated in LSPS relative to pair recording due to exclusion of direct responses along the apical dendrite of the postsynaptic neuron (see L5A and L5B maps in Figure S6 ). Under-sampling of connected pairs in low- p con pathways, such as L4→L6, may account for differences from LSPS, where many L4 neurons are excited during each L6 recording. Lastly, L2 connectivity differs in part because of differences in the definition of this layer.
Inter-Areal Comparisons: Ascending Pathways to Supragranular Layers
We compared the matrices for the four areas so far studied, vM1 (present study), the forelimb region of somatic M1 [13] , vS1 (also the present study) [27] , and S2 (present study). Overall, the main differences are attributable to the presence of a distinct granular layer in somatosensory cortex. Specifically in vS1, L4 outflow contributed strongly to the connectivity matrix. L4→L2/3 is also a major pathway in rodent V1 [44] . In S2, the local excitatory circuit differs from vS1 most prominently in that the L4→L3 pathway is reduced. LSPS analyses of auditory cortex circuits have found L4→L2/3 inputs [45] , [46] , which is adjacent to S2. However, ascending pathways were not unique to vS1, as a similar but weaker L3/5A→L2/3 pathway was prominent in forelimb M1, and present but weaker still in vM1 ( Figure 3C and Figure S5C , leftmost panels). The upward compression of layers in vM1, typical of cortical convexities [28] , may be why L3/5A→L2 was less distinct in vM1 than in forelimb M1 (e.g., it was more prone to masking by dendritic responses of L2 neurons). However, inspection of individual maps and traces ( Figure S5C ) showed that these ascending pathways were present for some L2 neurons.
Inter-Areal Comparisons: Descending Pathways to Deep Layers
A second main interlaminar hotspot in vS1 was the descending pathway(s) L2/3→L5, which was the predominant hotspot in the two motor areas. We noted that this pathway was present in all three cortical regions studied here and was similarly prominent in somatic M1 [13] . Indeed, it was the predominant pathway in S2. Thus, a strong supragranular to infragranular descending connection emerged as a common element of local cortical circuits examined here. Superficial L5B neurons and deep L5A neurons at the laminar border were most strongly activated, suggesting that the cytoarchitectonic boundaries identified do not correspond well with functional gradient within L5. Perhaps an alternative molecular marker, such as Etv1 ( Figure S8 ), better denotes this functional division.
Inter-Areal Comparisons: Involvement of Deep Neurons in Local Circuit Function
In three of the four areas, L6 neither received nor sent strong projections (but vS1 neurons in L6 received a weak projection from L4). L6 output is provided by an ascending connection to L4 in cat visual cortex [31] , [32] , [47] but was absent or reduced in all vibrissal areas we studied. L6→L4 projections studied in mouse somatosensory and auditory cortical areas have “modulator” rather than “driver” properties, including paired pulse facilitation [48] . Although deeper neurons tend to have relatively small dendritic arbors [49] , which may account for a reduction (but not absence) of inputs, this difference in arbor size is not of sufficient magnitude to account for the paucity of inputs. Similarly, the paucity of outputs was not due to lack of photoexcitability of these neurons. Channelrhodopsin-assisted circuit mapping experiments [50] have shown that the supragranular layers indeed connect preferentially to upper rather than lower infragranular neurons. Thus, the lack of inputs was not due simply to slice-related artifacts such as severing of pathways. Consistent with weak local inputs, in vivo recordings in cat motor cortex suggest that a large number of L6 neurons are virtually silent, even during motor activity [51] . Thus, the sources and modes of excitation for L6 neurons remain to be determined [49] , [52] . However, L6 was more engaged in local circuits in S2, supplying a measurable output to L5A and L5B and to other L6 neurons. In addition to input from L5B, L6 neurons in S2 collected inputs from a wide horizontal distance, sometimes >300 μm ( Figure S7B ,C at right). Thus, S2 may be better suited for studying L6 function.
Quantitative Comparison of Cortical Microcircuits
One major difficulty in making a comparison of connectivity between two cortical areas is selecting the laminar position of pre- and postsynaptic neurons for the comparison. Is it better to compare identical relative laminar depths between cortical areas, not accounting for the decreased thickness of superficial layers, and increased thickness of deep layers, in motor areas? How shall we treat the presence or apparent absence of a distinct layer 4? We present a direct quantitative comparison of three major areas identified in our study, based on cytoarchitectonic laminar divisions ( Table 1 , Figure S8 ) ( Figure 9 ). In vS1 [53] and vM1 (Tianyi Mao, BMH, GMGS, KS, unpublished observations) these layers correspond to distinct cell types with different projection patterns. The descending projection from L2/3→L5A/B was prominent in all areas, but the strength of the pathways at a neuron→neuron level varied by a factor of four between the areas. Ascending projections from middle layers to superficial ones (L4→L2/3 in vS1 and S2; L5A→L2/3 in vM1 for comparison) were also present in all regions but were the least prominent in agranular vM1. Lastly, the L6→L5 projection identified in S2 was more than twice as strong at the neuron→neuron level than in vS1 (and the difference was greater with vM1). Our approach provides a defined framework for measuring similarities and differences between cortical microcircuits in a quantitative manner. | The author(s) have made the following declarations about their contributions: Conceived and designed the experiments: BMH DH LP KS GMS. Performed the experiments: BMH SAH YXZ DH. Analyzed the data: BMH DH KS GMS. Contributed reagents/materials/analysis tools: GMS. Wrote the paper: BMH KS GMS.
Optical and electrophysiological tools were used to map out the neural circuits within and between cortical layers in three different brain regions, and the results suggest regional specializations for sensory versus motor information processing.
Rodents move their whiskers to locate and identify objects. Cortical areas involved in vibrissal somatosensation and sensorimotor integration include the vibrissal area of the primary motor cortex (vM1), primary somatosensory cortex (vS1; barrel cortex), and secondary somatosensory cortex (S2). We mapped local excitatory pathways in each area across all cortical layers using glutamate uncaging and laser scanning photostimulation. We analyzed these maps to derive laminar connectivity matrices describing the average strengths of pathways between individual neurons in different layers and between entire cortical layers. In vM1, the strongest projection was L2/3→L5. In vS1, strong projections were L2/3→L5 and L4→L3. L6 input and output were weak in both areas. In S2, L2/3→L5 exceeded the strength of the ascending L4→L3 projection, and local input to L6 was prominent. The most conserved pathways were L2/3→L5, and the most variable were L4→L2/3 and pathways involving L6. Local excitatory circuits in different cortical areas are organized around a prominent descending pathway from L2/3→L5, suggesting that sensory cortices are elaborations on a basic motor cortex-like plan.
Author Summary
The neocortex of the mammalian brain is divided into different regions that serve specific functions. These include sensory areas for vision, hearing, and touch, and motor areas for directing aspects of movement. However, the similarities and differences in local circuit organization between these areas are not well understood. The cortex is a layered structure numbered in an outside-in fashion, such that layer 1 is closest to the cortical surface and layer 6 is deepest. Each layer harbors distinct cell types. The precise circuit wiring within and between these layers allows for specific functions performed by particular cortical regions. To directly compare circuits from distinct cortical areas, we combined optical and electrophysiological tools to map connections between layers in different brain regions. We examined three regions of mouse neocortex that are involved in active whisker sensation: vibrissal motor cortex (vM1), primary somatosensory cortex (vS1), and secondary somatosensory cortex (S2). Our results demonstrate that excitatory connections from layer 2/3 to layer 5 are prominent in all three regions. In contrast, strong ascending pathways from middle layers (layer 4) to superficial ones (layer 3) and local inputs to layer 6 were prominent only in the two sensory cortical areas. These results indicate that cortical circuits employ regional specializations when processing motor versus sensory information. Moreover, our data suggest that sensory cortices are elaborations on a basic motor cortical plan involving layer 2/3 to layer 5 pathways. | Supporting Information | We thank Tianyi Mao for valuable discussion of the data and Yeka Aponte for comments on the manuscript.
Abbreviations
action potentials
channelrhodopsin-2
laser scanning photostimulation | CC BY | no | 2022-01-13 00:00:18 | PLoS Biol. 2011 Jan 4; 9(1):e1000572 | oa_package/cf/c3/PMC3014926.tar.gz |
|
PMC3014927 | 21171987 | Background
Gastrointestinal stromal tumors (GIST) are the most common mesenchymal tumors of the gastrointestinal tract and are characterized by the expression of the KIT receptor (stem cell factor receptor, CD117) and to a lesser extent of PDGFRA (platelet derived growth factor receptor alpha), representing two closely related receptor tyrosine kinases (RTK) [ 1 , 2 ]. The majority of GIST shows oncogenic mutations either in KIT or PDGFRA [ 3 , 4 ]. Mainly, mutations in exon 9 or 11 of the KIT gene or in exon 18 of PDGFRA lead to ligand independent, constitutive activation of the kinase function [ 5 ]. About 60% of all GIST carry an exon 11 mutation of KIT which encodes the juxtamembrane domain of the receptor possessing an autoinhibitory function [ 6 , 7 ]. Less common mutations in PDGFRA (~ 10%) are detected in GIST that often display gastric location and epithelioid morphology [ 2 ].
In a minority of cases (10-15%) no mutations in the known KIT or PDGFRA hot spots are detected although these tumors express the KIT protein. This subgroup is called wild type GIST (wt-GIST) and comprises tumors in pediatric patients, in patients affected by the Carney triad, neurofibromatosis type 1 (NF1) associated GIST and a subset of sporadic adult GIST [ 8 - 11 ]. The pathogenetic mechanisms underlying wt-GIST are poorly understood and there is limited benefit of imatinib therapy in these patients [ 12 ]. Therefore the identification of additional genetic factors contributing to the pathogenesis of GIST may help to find new concepts of individualized therapy.
Recently, the BRAF mutation p.V600E was found in 4-13% of wt-GIST [ 13 - 15 ]. For another subgroup of wt-GIST including pediatric tumors, a strong IGF1R expression combined partly with gene amplification was described [ 16 - 18 ]. Two other alternative RTK probably involved in the pathogenesis of GIST are AXL and MET. Both kinases have been shown to be upregulated in GIST resistant to treatment [ 19 ]. AXL is a member of the Ufo/AXL subfamily and activates the same signaling pathway as KIT. The tyrosine kinase domain of MET is mutated in sporadic papillary renal carcinomas. Some mutations in the MET gene are located in codons homologous to those in KIT and it is suggested that these missense mutations lead to constitutive activation of the MET protein [ 20 ]. To develop additional therapy approaches it would be of interest to know whether this RTK also plays a role in wt-GIST.
Besides KIT and PDGFRA, CSF1R (colony stimulating factor 1 receptor), FLT3 (fms like tyrosine kinase 3) and PDGFRB (platelet derived growth factor receptor β) belong to the same family of type III RTK. These five tyrosine kinases show a homologous structure and a comparable function in activation, proliferation and suppressing apoptosis [ 21 - 23 ].
Aberrant expression and mutations in either CSF1R , FLT3 and PDGFRB or their ligands have been described in several malignant diseases. Overexpression of CSF1R is found in epithelial tumors such as breast and ovarian cancer [ 24 ]. The translocation t(1;2) of its ligand CSF1 leads to the development of the tenosynovial giant cell tumor [ 25 ].
Aberrantly expressed FLT3 is observed at high levels in a spectrum of hematologic malignancies [ 26 ]. Additionally, in AML an internal tandem duplication in the transmembrane domain of FLT3 was identified which leads to constitutive activation of its kinase domain. It seems that this mutation is not present or very rare in GIST [ 27 ].
PDGFRB is overexpressed in malignant peripheral nerve sheath tumors (MPNST) and chordomas [ 28 , 29 ].
To analyse expression profiles of certain tumors for research and diagnostic purposes, qPCR (quantitative PCR) is frequently applied because of its reproducibility and high sensitivity [ 30 , 31 ]. This method is based on the normalization of the target gene expression on stably expressed internal reference genes. A major challenge is the application of suitable reference genes which have to be tested and verified under defined experimental conditions [ 32 , 33 ]. Ideal reference genes have to be non-regulated, stable and not affected by biological or experimental conditions. The target gene is amplified together with the reference gene in order to minimize experimental variability concerning reverse transcription enzymatic efficiencies, PCR efficiency, amount of starting material and differences between human tissues. The reference gene and the target gene should have very robust and stable expression profiles to ensure accurate normalization and interpretation of results. The most stable expressed gene from a set of genes can be identified by geNorm [ 34 ], a software program which additionally provides the number of genes required to calculate a robust normalization factor based on the geometric mean of these genes.
Typical reference genes regulate basic and ubiquitous cellular functions and are responsible for the cellular maintenance, e.g. GAPDH or ß-actin. However, these commonly used reference genes vary considerably in different tissue types or under different experimental conditions [ 35 , 36 ]. There is no standard reference gene for all kinds of tissue types. To our best knowledge only few studies investigated the alteration of stability of reference genes in different mesenchymal tumour entities.
In the present study we attempted to identify suitable reference genes in gastrointestinal stromal tumors by using a set of sixteen reference genes which are currently applied in qPCR procedures. Furthermore we evaluated the gene expression of KIT as well as the alternative RTK FLT3 , CSF1-R , PDGFRB , AXL and MET in mutated and non-mutated gastrointestinal stromal tumors by qPCR using the identified reference genes. The study was focussed on wt-GIST compared to samples with mutations in KIT exon 9 and 11 and PDGFRA exon 18 in order to evaluate whether overexpression of these alternative RTK might contribute to the pathogenesis of wt-GIST. | Methods
Samples
A total of 107 samples were included into this study. All specimens were obtained in the years 2005 and 2006 under approved ethical protocols and with informed consent from each patient. All samples were fixed in neutral-buffered formalin prior to paraffin embedding. 20 samples from normal tissue (i.e., muscularis propria of stomach and gut) as control group and 87 GIST representing different mutational subgroups were evalutated. 20 samples of wt-GIST, 7 samples of wt-GIST associated with neurofibromatosis type 1, 20 samples with exon 9 mutation in KIT , 20 samples with exon 11 mutation in KIT and 20 samples with exon 18 mutation in PDGFRA ). Sequence analysis of KIT (exons 8, 9, 11, 13, 14, 15 and 17) and PDGFRA (exons 12, 14 and 18) was carried out as described earlier [ 1 , 37 , 38 ]. Additionally, in all samples the wild type status of the BRAF gene was ascertained. All GIST samples were stained immunohistochemically for CD117. Four normal tissues from the gastrointestinal region were available as fresh frozen and formalin-fixed, paraffin-embedded samples.
RNA extraction and cDNA synthesis
Prior to RNA extraction, paraffin-embedded tissues were cut into 10 μm sections and mounted on glass-slides. Six slides of each specimen were used for RNA extraction. The sections were deparaffinized by extracting twice in xylene for 10 min at room temperature. Rehydration was done in 100% ethanol, 90% ethanol, 80% ethanol and 70% ethanol made with DEPC-water for 10 min each. Tumor areas previously marked on a H&E slide were scraped from the sections with a sterile scalpel. Sections were transferred into a sterile 1.5 ml tube. Extraction and purification of RNA was done using the RNeasy FFPE KIT (Qiagen, Hilden, Germany) according to the manufacturer's recommendations. After tumor localization by H&E staining fresh frozen material was rasped into 10 μm thick pieces and RNA purification was carried out using the RNeasy Kit (Qiagen, Hilden, Germany). Both RNeasy Kits contain a step of DNase treatment. Finally RNA from fresh and from fixed tissue was eluted in water. The quantification was done spectrophotometrically (NanoDrop, PeqLab Technology, Erlangen, Germany). 500 ng of RNA from each sample was reverse-transcribed using a random-hexamer primer and Avian Myoblastosis Virus reverse transcriptase (AMV-RT) according to the manufacturer's protocol (Qiagen). cDNA from those four fresh frozen and corresponding formalin-fixed, paraffin-embedded (FFPE) control samples was generated in only one reverse transcription and loaded on the reference low density arrays (see below). Further cDNA samples were generated from the four FFPE control tissues, pooled within one patient and loaded onto a second independent reference low density array.
Identification of reference genes
The reference genes used here were preselected because of their constitutive, non-regulated stable expression over a wide spectrum of tissues. But nevertheless the preselected reference genes are not suitable for every kind of tissue and therefore need to be analyzed prior to use in a certain study. The detection of suitable reference genes was carried out using TaqMan Low Density Arrays (TLDAs, Microfluidic Cards, Applied Biosystems, Darmstadt, Germany). These arrays are prefabricated 384-well cards where gene-specific primer and probe sets are spotted in small reaction chambers during manufacturing. The cards have 8 separate loading ports leading into 48 wells each. In this study, arrays with triplicates of 16 putative reference genes were used, so 8 different samples could be analyzed. The list of assays is given in table 1 . For each sample, 500 ng cDNA was mixed with 2× TaqMan Universal PCR Master Mix (Applied Biosystems). 100 μl of this mixture was loaded into each port and distributed into the reaction chambers by centrifugation. The card was sealed and the quantitative PCR (qPCR) was performed on an ABI PRISM HT 7900 (Applied Biosystems) sequence detection system. After pre-incubation for 2 min at 50°C and 10 min at 95°C, the PCR reaction was performed (15 s at 94°C followed by 60 s at 60°C, 40 cycles). The fluorescent signal was measured in each cycle.
Analyses of gene expression by qPCR
qPCR analysis was performed using the assays-on-demand products (Applied Biosystems) listed in table 2 . These gene-specific qPCR assays consist of a pair of unlabeled PCR primers and a FAM labeled specific probe. According to the manufacturer of these assays, probe and primer sets that would amplify pseudogenes are excluded in the process of development. Reactions were carried out in a reaction volume containing 5 μl PCR Master Mix (Applied Biosystems), 0.5 μl forward and reverse primer mix, 500 ng cDNA ad 10 μl A.dest. Triplicate reactions were carried out for each transcript. Control reactions were performed using a minus RT preparation and a sample with A.dest instead of RNA. PCR conditions were the same as for the TaqMan Low Density Arrays.
Statistical analyses
The geNorm applet for Microsoft Excel was used to determine the most stable genes among the sixteen candidate reference genes. Raw Cq values were converted into relative quantities for analysis with geNorm, where the highest relative quantity for each gene is set to 1. The program selects from a panel of candidate reference genes the two most stable genes or a combination of multiple stable genes for normalization. The gene expression stability (M) value is based on the combined estimate of intra- and intergroup expression variations of the genes studied and takes the PCR efficiency into account. The limited M-value is 1.5. The most stable genes are stepwise selected from the investigated gene panel to estimate how many reference genes should be used. The normalization factors define the optimal number of reference genes required for a precised normalization design. The analysis of the expression data of tumor samples compared to control samples was performed with the REST software (Relative Expression Software Tool) [ 39 ]. The software normalizes the measured Cq-values of the target genes with those of the reference genes and compares the expression data of tumor and control samples by considering the PCR efficiency and the mean crossing point deviation. | Results
RNA quality
RNA quality of the samples was inspected on a 1% agarose gel. The concentration and purity of the RNA was characterized by the mean A 260/280 ratio and was on average 1.99 for fresh frozen as well as for formalin fixed, paraffin-embedded samples and reflected pure and protein-free RNA.
Expression variability of cDNA synthesis
Four normal control FFPE tissue samples from the gastric and bowel wall were used to determine suitable reference genes. One TLDA was done with cDNA from only one reverse transcription for each sample. As it is assumed that cDNA synthesis from FFPE tissue varies strongly in efficiency, pooled cDNA synthesis samples from the same FFPE tissue were used for a second TLDA. The pooled cDNA preparation showed an essentially lower variance in contrast to the cDNA samples from only one reverse transcription (Figure 1 ). The median of the M-values ranges from 0.1 to 0.2, thus the variability of the technical replicates is low. Within a single specimen the range is also very small. The median of the cDNA samples from a single reverse transcription ranges from 0.1 to 0.4. However, each cDNA sample from one reverse non-pooled transcription showed a high variance of the values (median of M-values 0.1 to 0.8, Figure 1 ). Therefore the variance of the samples can be lowered by cDNA pooling. The median of the cDNA samples from one reverse transcription from fresh frozen tissue ranged from 0.1 to 0.2 and can be compared to the median of the pooled cDNA from the FFPE tissue. Within a single specimen the range is higher than in the pooled cDNA samples, but smaller than in the cDNA from only one reverse transcription from the FFPE samples.
Identification of suitable reference genes for normalization
All 16 putative reference genes had a high expression stability and the 'M'-value (0.02-0.06) was clearly below the 'M'-cutoff-value of 1.5. By stepwise exclusion of genes, the expression stability value 'M' is calculated (data not shown) and the expression stability rises with the exclusion of further genes. Samples generated from only one reverse transcription, showed POLR2A , TFRC , RPLPO and GAPDH as the most stable genes (Figure 2a ). The average stability value was between 0.05 and 0.07. The range of the values was 0.01-0.15. The most stable genes for pooled cDNA samples were PGK1 , PPIA , RPLPO and IPO8 with a median of 0.04-0.05 (Figure 2b ). The range of the values was lower and varied between 0.01 and 0.09. All in all the gene expression variability of the pooled cDNA samples is much lower compared to the single reverse transcription cDNA synthesis. The gene UBC is the most stable one in fresh frozen tissue with the smallest value range (0.01-0.8), but RPLPO and PPIA belong also to the most stable genes in fresh frozen tissue with a median of 0.03-0.045 and the range of the values was 0.1-0.12 (Figure 2c ). TBP was not expressed in the tissue. By combining the lowest variability values of fixed and fresh tissue, the genes POLR2A , PPIA , RPLPO and TFRC were detected (Figure 3 ). Hence these four genes were used for further analysis with the GIST samples.
Expression profiles of KIT, FLT3, CSF1-R, PDGFRB, AXL and MET
Six patient groups (NT = normal tissue -control, WT = wild type, Ex9 = KIT Exon 9 mutated, Ex11 = KIT Exon 11 mutated, NF1 = wild type and neurofibromatosis type 1, Ex18 = PDGFRA Exon 18 mutated) were used to analyze the gene expression of KIT , FLT3 , CSF1-R , PDGFRB , AXL and MET . Here, the REST analysis with POLR2A is shown exemplarily. The REST analysis of the target genes KIT, CSF1R , FLT3 , PDGFRB, AXL and MET with the reference genes RPLPO , PPIA and TRFC showed the same significant results for the differential expression or at least the same trend in mutated and non-mutated GIST (data not shown). As suspected, we found a significant overexpression of KIT in exon 9 and exon 11 mutated GIST in comparison with normal tissue. Also a significantly lowered expression of PDGFRB in both groups compared to normal tissue was shown. The same effect was observed in PDGFRA exon 18 mutated GIST compared to normal tissue (table 3 ).
On closer inspection of normal tissue and wt-GIST a clear overexpression of KIT in the tissue of wild type tumors was shown. The same effect, but even stronger, could be detected by comparing NF1-associated wt-GIST with normal tissue. Concerning KIT expression, the tumors without mutation behave like KIT mutated GIST (table 3 ). This was in accordance with immunohistochemical staining (Figure 4 ).
The other target genes, FLT3 , CSF1R , PDGFRB , AXL and MET showed lower expression levels in the tumor tissue compared to normal tissue except AXL in NF1 associated GIST. MET was even significantly reduced in wt-GIST, KIT exon 9, 11 and PDGFRA exon 18 mutated GIST. No changes in expression levels were detected comparing normal tissue and wt-GIST with KIT exon 11 mutated GIST, indicating that both groups possess similar expression levels for the analyzed receptors. None of the groups showed significant expression alterations for CSF1R and FLT3 . An exception is the group of KIT exon 9 mutated GIST. A lower expression of CSF1R , FLT3 , PDGFRB and MET compared to normal tissue was identified in contrast to KIT , which showed a significant overexpression. In summary, overexpression of KIT was detected in each GIST subgroup compared to the corresponding normal tissue by using the preassigned reference gene POLR2A (table 3 ). GIST with a mutation in exon 18 of PDGFRA did not show a significant upregulation of KIT compared to normal tissue. GIST with a mutation in KIT exon 9, 11 and in the PDGFRA exon 18 showed significant PDGFRB downregulation. | Discussion
Most gastrointestinal stromal tumors exhibit mutations in exon 11 of the KIT gene. This exon encodes for the juxtamembrane domain of the receptor which possesses an autoinhibitory function. In wt-GIST, activating mutations are found neither in the KIT nor in the PDGFRA gene. Clinical treatment with the tyrosine kinase inhibitor imatinib targets the ATP binding site in the kinase domain of the KIT receptor. In wt-GIST, response to treatment is often poor [ 12 ] although most of them overexpress the KIT protein. A screen for activating mutations in the juxtamembrane domain of the alternative receptor tyrosine kinases CSF1R, FLT3, PDGFRB in 30 wt-GIST samples yielded only wild type sequences (data not shown) in agreement with previously published results [ 40 ]. Therefore we investigated in this study whether the expression of alternative receptor tyrosine kinases may contribute to the pathogenesis of wt-GIST and thus may help to identify wt-GIST subgroups with different response to imatinib and elucidate novel therapeutic targets.
The pathogenesis of several malignant tumors is associated with overexpression of CSF1R and PDGFRB [ 24 , 28 ]. Imatinib inhibits cell invasion in malignant peripheral nerve sheath tumors by blocking PDGFRB [ 41 ] and it has been found to have antitumor activity in patients with chordoma [ 42 ]. In our study, the gene-expression level of CSF1R , FLT3 and PDGFRB was determined in a cohort of 87 GIST samples. Furthermore, we assessed the expression of AXL and MET , two receptor tyrosine kinases which were found to be alternatively activated in therapy resistant GIST [ 19 ].
Gene expression analysis by qPCR requires suitable reference genes. The expression of reference genes like GAPDH or BETA-ACTIN is regulated differentially depending on the tissue type. Therefore they are not suited as univocal reference genes [ 35 , 43 ]. The determination of reference genes with stable expression in the experimental system used is essential to ensure accurate normalization and interpretation of results.
Whereas fresh tissue is frequently not available for genetic analysis, FFPE material is the standard. After formalin-fixation and paraffin-embedding of tissue, the isolated RNA is often heavily fragmented. In our study we used fresh as well as fixed material from the gastrointestinal tract to determine suitable reference genes and to analyze whether their expression levels are comparable. We then validated the reference genes by qPCR in our GIST cohort.
As reported also by others [ 44 , 45 ] the variability of expression levels for the reference genes was very diverse comparing fresh frozen and formalin-fixed tissue. Therefore, we decided to validate separate reference genes for each tissue type.
The genes TRFC , POLR2A, PPIA and RPLPO were validated as appropriate reference genes for FFPE tissue. For fresh frozen tissue, PPIA and RPLPO were also found to be suitable. Additionally, UBC is suited as a reference gene in fresh frozen tissue. The variability of the reference genes for fresh frozen tissue was lower than for fixed tissue. To overcome the problem of high variability in FFPE tissue, we pooled two independent cDNA syntheses from one sample as recommended in the MIQE guidelines [ 30 , 46 ]. The MIQE guidelines give considerations for a consistent application of the qPCR technology including experimental details, data analysis and reporting principles. Unequal efficiency of cDNA synthesis might be a reason for deviation. Additionally, we tried to select consistent patient material for our cohort by choosing paraffin blocks having the same age. It was shown by Bibikova [ 45 ], that Cq-values in qPCR experiments depend on the age and condition of the tissue blocks. Because the expression value depends also on amplicon length [ 30 , 44 ], only primer sets generating amplicons of about 100 bp were chosen for qPCR. This corresponds to the fragment length of degraded RNA between 100 and 200 bp. The application of the MIQE guidelines results in a minimum variability for reference genes. In summary, considering the MIQE guidelines FFPE material can be used reliably for expression analysis in GIST, but the use of separate reference genes for FFPE tissue is indispensable.
Our qPCR analysis included GIST with wild type sequences in the hot-spot regions of KIT and PDGFRA (wt-GIST), KIT exon 9 and exon 11 mutated samples, PDGFRA exon 18 mutated samples, NF-1-associated wt-GIST and normal tissue controls. The wt-GIST and the samples with the two different KIT mutations showed a significantly increased expression of KIT in contrast to the normal tissue. The results correspond to the immunohistochemical stainings of our samples and were in agreement with published data [ 47 ]. It was shown immunohistochemically that PDGFRA mutated GIST have only a slightly increased protein expression of KIT [ 13 ]. Our data revealed the same trend on RNA level compared to the normal tissue. Additionally, we could show in all groups of mutated GIST a significantly reduced expression of PDGFRB compared to the normal tissue. This could be due to the concomitant increase of KIT expression. When comparing NF-1-associated GIST without KIT mutation with wt-GIST without NF-1 association the latter showed a lower expression of KIT . Thus, the results of the two wt-GIST groups give a heterogenous profile, which suggests that different genomic events may be responsible for the development of these tumors.
Comparing our sample groups with each other, no significant difference in the gene expression levels of FLT3 , CSF1R and AXL were determined. An exception was seen in the sample group with KIT exon 9 mutation. Here, a significantly reduced expression of CSF1R , FLT3 and PDGFRB compared to the normal tissue was detected. The results lead us to the assumption that KIT exon 9 mutated GIST play a special role compared to GIST carrying other mutations. Interestingly, KIT exon 9 mutated GIST need a double daily dose of the tyrosine kinase inhibitor imatinib to be effectively treated [ 48 ]. Furthermore, they develop preferentially in the small intestine but only rarely in the stomach where the majority of GIST are detected [ 49 ].
All qPCR data were calculated four times with TRFC , POLR2A, PPIA and RPLPO as reference genes using the REST software. Concordant results with the four reference genes are based on the extensive and complex preselection of our cohort and the preparation of cDNA synthesis according to the MIQE guidelines. | Conclusions
In summary, we conclude that none of the alternative receptor tyrosine kinases analyzed here are associated with the pathogenesis of wild type or mutated GIST. It remains to be clarified whether an increased expression of receptor tyrosine kinase ligands is responsible for tumorigenesis of wt-GIST as it is described for dermatofibrosarcoma protuberans (DFSP) and tenosynovial giant cell tumor (TGCT) [ 25 , 50 ]. Further studies are needed to elucidate the role of ligand-driven pathogenesis in wt-GIST. | Background
Gastrointestinal stromal tumors (GIST) represent the most common mesenchymal tumors of the gastrointestinal tract. About 85% carry an activating mutation in the KIT or PDGFRA gene. Approximately 10% of GIST are so-called wild type GIST (wt-GIST) without mutations in the hot spots. In the present study we evaluated appropriate reference genes for the expression analysis of formalin-fixed, paraffin-embedded and fresh frozen samples from gastrointestinal stromal tumors. We evaluated the gene expression of KIT as well as of the alternative receptor tyrosine kinase genes FLT3 , CSF1-R , PDGFRB , AXL and MET by qPCR. wt-GIST were compared to samples with mutations in KIT exon 9 and 11 and PDGFRA exon 18 in order to evaluate whether overexpression of these alternative RTK might contribute to the pathogenesis of wt-GIST.
Results
Gene expression variability of the pooled cDNA samples is much lower than the single reverse transcription cDNA synthesis. By combining the lowest variability values of fixed and fresh tissue, the genes POLR2A , PPIA , RPLPO and TFRC were chosen for further analysis of the GIST samples. Overexpression of KIT compared to the corresponding normal tissue was detected in each GIST subgroup except in GIST with PDGFRA exon 18 mutation. Comparing our sample groups, no significant differences in the gene expression levels of FLT3 , CSF1R and AXL were determined. An exception was the sample group with KIT exon 9 mutation. A significantly reduced expression of CSF1R , FLT3 and PDGFRB compared to the normal tissue was detected. GIST with mutations in KIT exon 9 and 11 and in PDGFRA exon 18 showed a significant PDGFRB downregulation.
Conclusions
As the variability of expression levels for the reference genes is very high comparing fresh frozen and formalin-fixed tissue there is a strong need for validation in each tissue type. None of the alternative receptor tyrosine kinases analyzed is associated with the pathogenesis of wild-type or mutated GIST. It remains to be clarified whether an autocrine or paracrine mechanism by overexpression of receptor tyrosine kinase ligands is responsible for the tumorigenesis of wt-GIST. | Authors' contributions
MCB, JF and HK carried out the molecular genetic studies and JF wrote and drafted the manuscript. SMB and HUS developed the design of the study. SMB coordinated the study together with EW and HUS. BB and MZ performed the statistical analysis of the expression data. EW and HUS participated in the diagnosis and selection of tumor material. RB revised the manuscript for important intellectual content and approved the final manuscript. All authors read and approved the manuscript. | Acknowledgements and Funding
This study was supported by a grant of the BONFOR programme of the Bonn Medical centre to HUS. | CC BY | no | 2022-01-12 15:21:37 | BMC Mol Biol. 2010 Dec 20; 11:100 | oa_package/56/d0/PMC3014927.tar.gz |
PMC3014928 | 21176138 | Background
The CTCF protein, formerly known as NeP1, is an eleven zinc finger protein that is highly conserved from fruit flies to man. The protein was first identified in the chicken as a negative regulator of the c -myc oncogene [ 1 ] and the lysozyme gene [ 2 ]. The CTCF protein has a central zinc finger domain that shows 100% amino acid conservation between the chicken form and the human form of the protein. This central zinc finger domain is flanked by an NH 2 -terminal domain and a carboxy-terminal domain, both of which have an unknown structure. The CTCF protein has the ability to bind to different CTCF consensus sites by using different combinations of its eleven zinc fingers and is therefore frequently described as a multivalent protein [ 3 ]. These binding studies were performed by deleting different CTCF zinc fingers and observing the effects the deletions had on the ability of the protein to bind to different consensus sites [ 4 - 8 ]. A more recent study has determined that CTCF uses 4 to 5 core zinc fingers to bind to CTCF consensus sites [ 9 ]. Recent whole genome analyses of the CTCF binding sites in Drosophila and human cell lines support the idea that CTCF protein binds to a single consensus sequence [ 10 , 11 ].
The molecular mechanisms regulating the many diverse functions of CTCF are in part governed by the posttranslational modification of the protein. Phosphorylation of CTCF has been shown to relieve its repressive activity at the c-myc P2 promoter [ 12 , 13 ] and poly-(ADP)-ribosylation has been implicated in its role as an insulator protein [ 14 ]. In addition, we have recently shown that the posttranslational modification of CTCF by the small ubiquitin-like modifier proteins (SUMOs) contributes to its role as a transcriptional repressor at the c-myc P2 promoter [ 15 ].
CTCF is implicated in a diverse number of biological roles including gene repression, gene activation, chromatin insulator function, X-chromosome inactivation and the maintenance of genomic imprinting [ 3 , 16 ]. Recently, CTCF has been found to play a role in the organization of the mammalian genome and has been implicated in the genomic organization of the β-globin locus [ 17 , 18 ], the H19/Igf2 imprinting control region [ 19 - 23 ], the major histocompatibility complex class II genes [ 24 ] and the cystic fibrosis transmembrane conductance regulator gene locus [ 25 ]. The CTCF protein binds to approximately 15 000 sites in the human genome [ 11 , 26 - 29 ] and is hypothesized to organize the genome by forming DNA loops [ 30 , 31 ]. The evidence of CTCF's ability to loop DNA is the result of in vivo chromatin conformation capture assays (3C), and chromatin immunoprecipitation. The CTCF DNA binding site has been shown to be necessary for long-range chromatin interactions at the H19/IGF2 imprinting control region [ 23 ] and the knockdown of CTCF protein in chicken cells disrupts long-range chromatin interactions at the β-globin locus [ 17 ]. It has also been suggested that CTCF forms loops in DNA by tethering DNA to the nucleolus through its interaction with the protein nucleophosmin [ 31 ].
Since CTCF has previously been shown to bend DNA [ 32 ], we asked whether its SUMOylation altered its bending ability. In the course of answering this question, we obtained some unexpected results. Although SUMOylation had no effect on CTCF's ability to bend DNA, we found that the CTCF protein does not act as a typical DNA bending protein. The CTCF protein forms an unusual structure in DNA that we believe to be a DNA loop. This unusual DNA structure forms at all three CTCF binding sites tested: the chicken β-globin FII insulator, the chicken lysozyme gene F1 silencer element and the human c-myc P2 promoter. We find that the SUMOylation of CTCF does not affect its ability to form this unusual DNA structure. We discuss the possible mechanisms of DNA looping by CTCF and their roles in genome organization. | Methods
Plasmids and Cloning
Oligonucleotides containing CTCF binding sequences in the chicken β-globin FII HS4 insulator element (pr2214 and pr2215) and chicken lysozyme F1 gene silencer (pr2236 and pr2237) are described in Table 1 . The oligonucleotides were annealed, digested with the restriction enzyme XbaI and cloned into the XbaI site in the DNA bending vector pBEND2 [ 33 ]. Plasmids containing the FII site in both orientations were isolated. We arbitrarily name the orientation in which the FII site is pointing "towards" the HindIII site in the pBEND2 vector as the "forward" orientation. Conversely, when the FII site is pointing towards the EcoRI site in the pBEND2 vector the orientation is defined as the "reverse" orientation (See Figure 1A and 2A ). A fragment containing the c-myc -P2 promoter was amplified by PCR using the primers pr2216 and pr2217 (Table 1 ) as described previously [ 15 ]. The PCR product was digested with the restriction enzyme XbaI and was cloned into the XbaI site in the vector pBEND2 resulting in the vector that was called pBEND2- c-myc -P2. Only plasmids containing the forward orientation were isolated.
To construct the shuffled clones, we randomized the original 115 nt DNA sequence from the SalI restriction enzyme site to the HindIII restriction enzyme site in the pBEND2 vector using the online Bioinformatics tool, the Sequence Manipulation Suite [ 41 ]. The shuffled oligonucleotides pr2232 and pr2233 (Table 1 ) were annealed and phosphorylated using polynucleotide kinase (New England Biolabs). The primers were engineered to contain SalI and HindIII restriction enzyme sites for ligation into the SalI and HindIII sites in the pBEND2-FII-Forward or Reverse vectors. The result was the formation of the plasmids pBEND2-FII-Forward or Reverse-Shuffled in which the tandemly repeated restriction enzyme sites to the right of the XbaI site were replaced with the shuffled sequence. The random clones were constructed in a similar manner except that a random 115 nt DNA sequence of 50% A+T and 50% G+C base composition was generated using the Sequence Manipulation Suite [ 41 ]. The random oligonucleotides pr2228 and pr2229 were cloned into pBEND2-FII-Forward or Reverse to form the plasmids pBEND2-FII-Forward or Reverse-Random as described above (see Figure 6A ).
The phasing clones were constructed by excising the 156 bp BamHI-BamHI DNA fragment from the plasmids pK10, pK12, pK14, pK16, pK18 and pK20 and replacing it with annealed oligonucleotides that contained the chicken β-globin FII HS4 insulator element cut with BamHI (see Table 1 and Figure 7A ). Clones in both orientations were isolated.
Purification of the zinc finger domain of CTCF
The zinc finger domain of CTCF was amplified by PCR from a mouse CTCF cDNA clone using the primers pr1964 (BamHI) and pr1965 (XhoI) (Table 1 ) and was cloned into the BamHI and XhoI restriction sites in the vector pGEX4T-1 (GE Healthcare) to create an NH 2 -terminal glutathione S -transferase (GST) fusion protein. The GST-fusion of the zinc finger domain was purified as previously described [ 15 ]. The purity of the GST-ZnF-CTCF domain was determined to be about 37% pure by scanning a Coomassie stained SDS-PAGE using the program ImageJ.
In vitro transcription/translation of CTCF and in vitro SUMOylation
Full length CTCF was prepared by in vitro transcription/translation as described previously. The full length CTCF was posttranslationally modified by SUMO 1 in vitro using a SUMOylation control kit purchased from LAE Biotech International (catalogue no. K007) as described with the following modifications [ 15 ]. Briefly, 2 μL of in vitro translated CTCF were incubated with the recommended amounts of E1 and E2 enzymes and 20 mM ATP with or without SUMO 1 protein in a 12 μL reaction mixture for one hour at 37°C. The SUMOylation of CTCF was quantitative as shown previously [ 15 ].
Circular Permutation Assay
Probes were prepared from the pBEND2 clones by digesting plasmid DNA with the following restriction enzymes: MluI, BglII, NheI, SpeI, EcoRV, SmaI, SspI, KpnI and BamHI. The digested probes were purified from an agarose gel using the MinElute Gel Purification Kit (Qiagen) and 5' labeled with γ- 32 P-ATP (Perkin Elmer) and T4 Polynucleotide Kinase (New England Biolabs). Electrophoretic mobility shift assays were carried out as described previously with the following modifications. Briefly, 2 μL of SUMOylated or unmodified CTCF from the SUMOylation reactions were used in each gel shift reaction. In experiments where the effects of SUMOylation were not being assayed, 0.5 μL to 1 μL of the in vitro translate was used in each gel shift reaction. In experiments using the GST-fusion of the zinc finger domain of CTCF, 250 ng of the purified protein domain was used in each gel shift reaction. The reactions were run on 4% native polyacrylamide gels in 0.25X TBE at 9 V/cm. The gel dimensions for the FII and F1 mobility shifts with full-length CTCF were 10 cm × 7 cm, while gels 19.5 cm × 16 cm were used to analyse the c-myc P2 promoter and the GST-zinc finger domain mobility shifts. The dried gels were exposed to a phosphor screen and imaged using a Phosphorimager. The electrophoretic mobilities of the CTCF-FII forward and reverse probe complexes and free probes were measured using ImageQuant software. The relative electrophoretic mobility (μ) of a CTCF-probe complex was calculated as the mobility of the complex divided by the mobility of the free probe. The relative mobilities of the complexes were plotted as a function of the position (bp) from the middle of the left EcoRV site to the middle of the restriction enzyme used to generate the probe and the graphs were fitted with the best fit polynomial curve using Microsoft Excel.
Phasing Analysis
The fragments for the phasing experiments were prepared by digesting the pK10, pK12, pK14, pK16, pK18 and pK20 plasmids containing the FII insulator element in either the forward or reverse orientation, with the restriction enzymes RsaI, PvuII and NheI. Since the probes were not gel purified, we included the NheI restriction enzyme to digest a 464 bp RsaI vector fragment that would otherwise co-migrate with the FII insulator probes. The 391 bp to 401 bp RsaI to PvuII fragment contains the FII insulator element. Fragments were 5' labeled and electrophoretic mobility shifts were performed as described above, using 19.5 cm × 16 cm 4% native PAGE gels. The electrophoretic mobilities of the CTCF-FII forward and reverse phasing probe complexes and free probes were measured using ImageQuant software. The relative electrophoretic mobility (μ) of a CTCF-probe complex was calculated as the mobility of the complex divided by the mobility of the free probe. The relative mobilities of the complexes were plotted as a function of the linker length (bp). The graphs were fitted with the best fit polynomial curve using Microsoft Excel. | Results
The CTCF insulator protein forms an unusual directional DNA structure
CTCF has been found previously to bend DNA [ 32 ]. We initially wished to determine whether SUMOylation affected CTCF's ability to bend DNA. Therefore, we cloned the well characterized chicken β-globin FII insulator site into the XbaI site of the pBEND2 vector [ 33 ]. In this plasmid, the XbaI site is flanked by a set of tandemly repeated restriction enzyme sites (See Figure 1A ). When the plasmid is digested with each of these restriction enzymes, a set of probes of equal length is generated; the DNA binding site is permuted along the length of the probe. The probes were radiolabeled with 32 P and were incubated with CTCF that we synthesized in vitro [ 15 ]. The DNA-protein complexes were analysed on 4% native acrylamide gels.
In a typical DNA bending experiment, the DNA-protein complex has the slowest mobility when the DNA binding site is centrally located. Conversely, the fastest mobility occurs when the DNA binding site is located near either end of the probe. Therefore, we expected that the complexes formed by both the MluI probe and the BamHI probe would migrate more quickly than the EcoRV fragment containing the CTCF-binding site in the middle of the fragment (See Figure 1C ). We tested the ability of CTCF to bend DNA using the well-characterized chicken HS4 β-globin FII insulator element. As expected, when the FII site was located near the right end of the probe (digestion with MluI), the complex migrated more quickly in the native gel than when the site was in the middle of the fragment (see Figure 1B , left). As the FII site was permuted to a more central location (digestion with BglII, NheI, SpeI and EcoRV), the mobility of the CTCF-probe complex decreased, as predicted. However, we expected that the probes generated by digestion with SmaI, SspI, KpnI and BamHI would yield complexes that would migrate progressively more quickly through the gel, mirroring the permuted probes on the opposite side of the EcoRV site as is usual in DNA bending assays. Instead, we saw even more slowly migrating complexes. We wondered if these results could be explained by CTCF binding to a second site in the pBEND2 probe itself. Therefore, we generated a probe that does not contain the FII CTCF binding site by digesting the parent pBEND2 plasmid with the enzyme BamHI. We observed no CTCF-DNA complexes with this substrate, indicating that CTCF does not bind DNA from the parent pBEND2 plasmid in the absence of the FII insulator site (data not shown).
Our original interest was to determine whether the SUMOylation of CTCF affected its ability to bend DNA; therefore, we SUMOylated CTCF quantitatively in vitro with SUMO1 [ 15 ]. When SUMOylated CTCF was used in the EMSA assay, the DNA bending pattern was identical to that seen using unmodified CTCF (see Figure 1B , right). Hence, the modification of the CTCF protein by SUMO did not affect its ability to deform DNA containing a CTCF binding site. The relative electrophoretic mobility (μ) of a CTCF-probe complex was calculated as the mobility of the complex divided by the mobility of the free probe. The relative mobilities of the complexes were plotted as a function of the position (bp) from the middle of the left EcoRV site to the middle of the restriction enzyme used to generate the probe (Figure 1E ). The graphs were fitted with the best fit polynomial curve using Microsoft Excel. Two CTCF DNA complexes were observed in these experiments. The top complex on the electrophoretic mobility gels corresponds to the probe bound by full length CTCF, whereas the bottom complex on the gels corresponds to the probe bound by a C-terminally truncated form of CTCF caused by premature termination during in vitro translation. The major SUMOylation site in CTCF is found in the C-terminal domain of the protein [ 15 ]. The polynomial curves fitting the relative mobility of the top complexes of SUMOylated and unmodified CTCF are distinct, thus indicating that CTCF was efficiently SUMOylated due to the slower migration of the SUMOylated complexes (Figure 1E , bottom two curves). Conversely, when the protein is C-terminally truncated the bottom complexes exhibit similar relative electrophoretic mobilities since the C-terminally truncated CTCF is not being efficiently SUMOylated (Figure 1E , top two curves). The abnormal electrophoretic mobility exhibited by CTCF during circular permutation creates an unusual polynomial curve. Since we cannot extrapolate to the position at which the curve reaches a maximum relative mobility it was difficult to determine the centre of the CTCF-induced bend (see Figure 1C bottom). Likewise, we were unable to determine the possible differences in the bend angles induced by unmodified and SUMOylated forms of CTCF. The CTCF-induced bend is not a typical asymmetrical DNA bend (see Figure 1D ). We refer to the DNA structure formed upon CTCF binding to the β-globin FII insulator element as an unusual DNA structure.
We then asked whether the behaviour of the CTCF-DNA complexes would persist if we inverted the FII CTCF binding site in the pBEND2 vector. We therefore generated a new plasmid called pBEND2-FII-reverse and repeated the circular permutation experiments using the new set of probes (see Figure 2A ). The experiments yielded results that mirrored those of the FII forward oriented probes (see Figure 2B , left). The probe generated with BamHI formed a CTCF-DNA complex that migrated most rapidly, whereas probes generated with EcoRV, SpeI, NheI, BglII and MluI formed CTCF-DNA complexes that migrated more slowly through the gel. These results show that the altered DNA structure formed by CTCF is dependent on the orientation of the CTCF FII binding site in the pBEND2 vector. Furthermore, they show that the mobility of the CTCF-probe complex is inversely correlated with the length of the DNA downstream of the FII site. This suggests that in addition to its primary site of binding in the FII sequence, CTCF uses DNA sequence outside the FII consensus sequence to form an altered DNA structure. Once again, the SUMOylation of CTCF had no influence on the formation of the altered DNA structure (see Figure 2B , right).
Although there are several explanations for the altered DNA structure, we believe that the unusual DNA structure formed by CTCF in our bending experiments is a small DNA loop. Some of the alternative models are dealt with in the Discussion.
The CTCF insulator protein forms a directional unusual DNA structure at two other CTCF binding sites
The CTCF protein binds to a rather loose consensus sequence that occurs some 15 000 times in the human genome. We were therefore interested to know whether other well-characterized CTCF binding sites would exhibit this same behaviour when placed in the pBEND2 vector. When we cloned the c-myc P2 promoter-binding site and the chicken lysozyme gene F1 silencer site into pBEND2, the permuted fragments also showed a similar unusual conformational behaviour in DNA bending permutation assays (Figure 3 ). The chicken F1 silencer site was the one previously used to characterize DNA bending by CTCF [ 32 ]. We conclude that the unusual conformation of the protein-DNA complexes occurs independently of the sequence of the CTCF DNA binding site. Incidentally, we again showed that this altered conformation occurs independently of the SUMOylation of CTCF (Figure 3A , lanes 10-17). Note that the differences in the mobilities of the DNA-CTCF complexes at the c-myc P2 promoter are not as striking as those seen at the F1 and FII sites since the c-myc P2 probes are larger than those of the F1 and FII elements. Since the mobility of the DNA-protein complexes depends upon the molecular mass of both the DNA probe and the bound protein as well as the extent of the bending, it is not unusual to see a smaller effect using a larger DNA probe.
CTCF-DNA complexes do not involve intermolecular interactions between DNA molecules
The slowly migrating complexes seen in the bending assays might have arisen through the multimerization of two DNA fragments mediated by CTCF. To address this question, we performed a mixing experiment using separate probes containing either the c-myc P2 promoter region or the FII insulator. These were prepared by radiolabeling EcoRV digested pBEND2- c-myc P2 and pBEND2-FII-reverse plasmids, respectively. The FII probe is smaller than the c-myc P2 fragment and the two probes are easily resolved on a 4% native polyacrylamide gel (see Figure 4 , lanes 1 and 3). Upon the addition of CTCF, the CTCF- c-myc P2 complex and the CTCF-FII complexes are also easily resolved on the native gel (Figure 4 , lanes 2 and 4). If CTCF were acting as a DNA bridging protein between the two probes, then the incubation of CTCF with both probes should cause the appearance of an additional higher molecular weight complex migrating behind the c-myc P2-CTCF complex (see schematic below the gel). As seen in lane 6 of Figure 4 , no such complex is detected. Therefore, we conclude that CTCF is not forming an intermolecular bridge between two DNA molecules in our experiments.
The zinc finger domain of CTCF is sufficient for the formation of the unusual DNA structure
The CTCF protein is thought to use multiple permutations of its zinc fingers to bind to its highly diverse DNA binding sequences [ 3 , 4 , 6 - 8 ]. Therefore, it was of interest to learn whether the ability to form the altered DNA structure on the pBEND2 vectors also resided in the zinc finger domain. We used permuted probes from the pBEND2-FII-reverse plasmid in electrophoretic mobility shift assays with the CTCF zinc finger domain (Figure 5A ). When the zinc finger domain is incubated with the pBEND2-FII-reverse BamHI fragment (the FII binding site at the end of the probe), the DNA-protein complex migrates rapidly through the acrylamide gel. When the probe was prepared by digesting with MluI, (FII binding site is at the opposite end of the probe), a slower migrating DNA-protein complex was obtained. Therefore, the zinc finger domain of CTCF is sufficient for the formation of the unusual DNA structure at the FII insulator element. We obtained the same circular permutation results using permuted probes containing the chicken F1 lysozyme gene silencer (Figure 5B ). Because of its reduced molecular mass, the zinc finger domain produces a smaller effect on electrophoretic mobility than that of the full-length protein during circular permutation assays. When the results of the circular permutation assay using the CTCF zinc finger domain are plotted, the curve is not indicative of an asymmetric bend but is once again unique (data not shown). As can be seen, three of these probes (MluI, BglII and NheI) show the increased mobility at the F1 element. We conclude that CTCF's ability to form the unusual DNA structure resides in its zinc finger domain.
The formation of the unusual DNA structure by CTCF is not dependent on a specific DNA sequence downstream of the CTCF binding site
We have observed that when the length of the DNA "downstream" of the FII site increased, the mobility of the CTCF-DNA complex in the native gel decreased. This suggests that CTCF binds to its consensus-binding site in a unique orientation and uses DNA downstream of the FII consensus sequence to form the unusual DNA structure with reduced electrophoretic mobility. To determine whether the actual DNA sequence downstream of the FII consensus affected the ability of CTCF to form the unusual DNA structure, we replaced the DNA sequence downstream of the FII site with two new DNA sequences. In the plasmid pBEND2-FII-Forward-Shuffled we replaced the original sequence with a shuffled version of the original downstream sequence. We also constructed the plasmid pBEND2-FII-Forward-Random by replacing the original downstream sequence with a random DNA sequence composed of a 50% A+T and 50% G+C base composition (Figure 6A ). When probes were prepared by digesting the new plasmids with BamHI and HindIII we observed that CTCF is still able to form the unusual DNA structure regardless of the DNA sequence downstream of the FII binding site (Figure 6B , lanes 4, 5 versus lane 6). As controls, we also cloned both the shuffled DNA sequence and the random DNA sequence into the pBEND2-FII-Reverse plasmid. When the same probes were prepared with the control clones, CTCF did not form the DNA structure showing that the reduced mobility of the experimental probes was not due to the new DNA sequences cloned downstream of the FII site (Figure 6B , lanes 2, 3 versus lane 1).
CTCF phases DNA in a manner dependent upon the orientation of the CTCF binding site
The DNA bending assays using the pBEND2 vector cannot distinguish between a directional bend in the DNA and DNA flexure. DNA phasing assays indicate whether the bend has directionality as opposed to a random DNA flexure in which the DNA may be bent in any direction. We used the phasing vectors of Zinkel et al . [ 34 ] to determine the behaviour of the CTCF-induced bend. These vectors contain a BamHI cloning site for the DNA-binding site of interest (in this case the FII site) joined to a kinetoplast DNA sequence. The kinetoplast DNA consists of sequences of A-tract repeats that cause a sequence-directed bend toward the minor groove of the helix. The length of the linkers between the two sites is varied over 1 turn of the DNA helix: 10, 12, 14, 16, 18 and 20 bp. As can be seen in Figure 7A , the FII insulator element was cloned into the BamHI site of the phasing vectors in two orientations. The forward orientation occurs when the FII site is pointing towards the kinetoplast DNA (κDNA) and the reverse orientation occurs when the FII site is pointing away from the κDNA as indicated by an arrow. When the FII site is cloned "facing" the kinetoplast DNA (forward orientation), CTCF induces a typical DNA bend rather than a DNA flexure. The migration of the DNA-probe complexes changes depending on the orientation of the sequence-induced bend of the kinetoplast DNA and the CTCF induced bend in the FII insulator element. The minimal migration of the CTCF-DNA complexes corresponds to the "cis" isomer, the DNA conformation where the protein-induced bend and the sequence induced bends are in the same direction as shown by the schematic in Figure 7B . When these bends are in the opposite direction, the CTCF-DNA complexes show maximal migration in the polyacrylamide gel; this corresponds to the "trans" isomer (Figure 7B ). In our phasing experiments, when the FII site is in the forward orientation the CTCF-DNA complexes exhibit minimum mobility at a linker length of between 10 bp and 12 bp (Figure 7C , "cis" configuration) whereas maximum mobility of the CTCF-probe complex is seen when the linker length is approximately 18 bp ("trans" configuration). The relative mobilities of the top CTCF-DNA complexes (full length CTCF) normalized to the mobility of each free probe are plotted below the gel in Figure 7C . The best fit polynomial equation was fitted to the data using Microsoft Excel to determine the linker length corresponding with both maximal and minimal CTCF-DNA complex mobilities. The CTCF protein is phasing DNA over a length of 10 bp, which corresponds to one helical repeat as seen by the overlap between the mobilities of the CTCF-probe complexes at linker lengths of 10 bp and 20 bp in the graph. This experiment confirms that when the FII site is cloned "facing" the kinetoplast DNA (forward orientation), CTCF induces a typical DNA bend rather than a DNA flexure (Figure 7C ). However, when the orientation of the FII site was reversed we observed an unusual electrophoretic behaviour of the CTCF-DNA complex. The mobility pattern is multiphasic, seemingly varying with a periodicity of 2 bp (Figure 7D ). Since our circular permutation assays yielded these unusual results we were unable to calculate the centre of the DNA bend and are therefore unable to determine the direction of the bend using phasing analysis.
We believe that the complex multiphasic mobility pattern is due to a CTCF-induced directional bend towards the kinetoplast DNA in addition to an unusual DNA structure formed towards the PvuII restriction enzyme site. These complex DNA conformations may not phase because the unusual DNA structure is not a fixed structure. When the phasing analysis was done in the forward orientation, the unusual DNA structure may have been unable to form due to the topological constraints posed by the bent kinetoplast DNA. Therefore, the CTCF-induced bend alone is responsible for the DNA phasing. | Discussion
In this paper we have presented biochemical evidence that shows CTCF is able to form an unusual DNA structure. We believe that this DNA structure is a CTCF-induced DNA loop and that CTCF acts as a DNA looping protein. We show that this unusual DNA structure is formed at the chicken β-globin FII insulator element and that its formation depends upon the orientation of the FII site. The unusual DNA structure also forms at two other well-characterized CTCF binding sites: the c-myc P2 promoter and the chicken lysozyme gene F1 silencer element. These results were unexpected since Arnold et al . [ 32 ] had previously characterized CTCF (formerly named NeP1) as a DNA bending protein. However, there are several differences between our circular permutation assay and those of these authors. We examined the effect of CTCF protein only on a circular permutation substrate that contained a single CTCF binding consensus sequence. Arnold et al . [ 32 ] examined a complex protein mixture of CTCF (NeP1), TR, and RXR in their circular permutation assays. The large number of protein-DNA complexes formed in their mobility shift assays may have obscured the aberrant electrophoretic mobilities of the CTCF-probe complexes alone. Careful re-examination of their data shows it to be consistent with our experiments.
We believe that the unusual DNA structure formed by CTCF in the circular permutation assays is a small DNA loop and we have depicted the results of our circular permutation experiments and the CTCF loop topologies in a speculative schematic diagram in Figure 8 . As indicated in our model, as the length of the probe downstream of the CTCF binding site increases a DNA loop is formed (Figure 8 , lanes 1 and 2) and the electrophoretic mobility of the CTCF-DNA complex is decreased. As the length of the probe upstream of the FII insulator element decreases and the length of the probe downstream of the binding site increases, the size of the loop is also increased (Figure 8 , lane 3). In lane 4, the electrophoretic mobility of the CTCF-DNA complex is slightly greater than the mobility exhibited by the complex in lane 3. We believe that this is due to the reduction in "drag" caused by the DNA upstream of the FII site as the loop reaches its maximal size.
There are several alternative models that could account for the unusual DNA topology caused by CTCF. One possible explanation might be that CTCF induces formation of single-stranded regions of DNA downstream of the FII binding site. However, when we probed the altered DNA structure with potassium permanganate or diethyl pyrocarbonate, the presence of single stranded DNA was not detected by DNA footprinting (not shown). Therefore, we conclude that the altered DNA structure is therefore not caused by melting of the DNA.
A second explanation is that the FII binding site nucleates the binding of multiple CTCF molecules downstream of the FII binding site. For example, multiple RecA protein molecules nucleate onto single stranded DNA in a unidirectional manner [ 35 ]. Interestingly, a C-terminally truncated form of CTCF is present in the in vitro translation reaction. This complex was identified using "supershift" experiments in the presence of a C-terminal-specific antibody (data not shown). If the complexes formed at the FII binding site are caused by CTCF nucleation, we might have expected to detect a mixed protein-DNA complex containing both the truncated and full-length CTCF. Since we do not detect this mixed complex, we conclude that the CTCF-DNA complexes contain a single molecule of CTCF or its C-terminal truncation.
It is possible that the nucleation of other proteins present in the reticulocyte lysate accounts for the aberrant mobility of the CTCF-probe complexes. However, we note that the purified GST-zinc finger domain of CTCF exhibits the same overall electrophoretic mobility pattern as the in vitro translated CTCF lysates. The presence of such proteins in the GST-zinc finger protein purified from E. coli is unlikely.
A further possible explanation for the aberrant DNA structure is the wrapping of DNA around CTCF. Circular permutation experiments using the mitochondrial protein mtTF1, a DNA wrapping protein, show a protein-DNA mobility pattern similar to those observed in classical circular permutation experiments [ 35 ]. While these other models may provide an explanation for the abnormal DNA structure formed by CTCF, the DNA looping model is the strongest explanation and fits well with the in vivo data describing CTCF's role in mediating long-range chromatin interactions.
Our bending results provide insight into the molecular mechanisms of the CTCF protein. It is of interest to note that the unusual DNA structure formed by CTCF is directional. CTCF uses DNA downstream of the binding site to form the structure while it does not appear to use DNA upstream of the binding site. Interestingly, the Drosophila insulator protein Su (Hw), which is also a zinc finger protein that binds to the gypsy insulator element, has also been studied in circular permutation experiments. Like CTCF, Su(Hw) protein is believed to contain twelve zinc fingers in a central protein domain and like CTCF, the Su(Hw) protein bends DNA in an unusual manner [ 36 ]. These authors attributed the aberrant behaviour to either an increase in flexibility or a melting of the DNA by Su(Hw).
In our experiments, the zinc finger domain of CTCF is sufficient for both the bending and looping activities. Arnold et al . [ 31 ] used extracts from mammalian cells transfected with a CTCF zinc finger domain construct to conclude that CTCF's DNA bending ability resided outside the zinc finger domain of the protein. It is conceivable that our results differ from those of Arnold et al . since we used a partially purified zinc finger domain instead of a complex cell lysate. The Su(Hw) protein needs its zinc fingers to bind DNA but requires the presence of its acidic C-terminal domain for DNA bending. This suggests that the mechanism by which Su(Hw) forms an unusual DNA structure may differ from that used by CTCF.
We determined that the nucleotide sequence of the DNA outside the FII binding site is not critical for the formation of the DNA loop. We speculate that these results could provide mechanistic insight into the ability of CTCF to organize the human genome. If CTCF were to bind specifically to a core consensus sequence it might then make a DNA loop using any nucleotide sequence. The protein might conceivably be able to make a DNA loop at any of the 15 000 sites in the genome to which it has been shown to bind.
Previous studies have attributed CTCF's ability to bind to different CTCF consensus sites to its use of different combinations of its eleven zinc fingers [ 3 ]. These studies were performed by deleting different CTCF zinc fingers and observing the effects the deletions had on the ability of the protein to bind to different consensus sites [ 4 - 8 ]. A more recent study has determined that CTCF uses 4 to 5 core zinc fingers (zinc fingers 4-8) that are critical to providing high affinity binding to a 12bp core sequence in CTCF consensus sites [ 9 ]. Recent whole genome analyses of CTCF binding sites in Drosophila and human cell lines support the idea that CTCF protein binds to a single consensus sequence [ 10 , 11 ].
If CTCF needs only 4 to 5 zinc fingers to bind to consensus sites (Figure 8 ), then the remaining six or seven zinc fingers might be free to bind DNA non-specifically to form a loop. A crystal structure of CTCF in the presence of DNA would add significant insight into its ability to form a DNA loop and its combinatorial use of zinc finger during DNA binding. The analysis of CTCF's DNA binding ability by surface plasmon resonance has shown that the binding of CTCF to DNA is a two-stage reaction [ 7 ]. It is possible that the first stage of the binding reaction is due to the initial binding of CTCF to the consensus site followed by a second binding step whereby CTCF forms a DNA loop. Alternately, the initial binding may non-specific followed by binding to the consensus site.
The CTCF protein has been previously shown by other workers to act as a DNA bridging protein [ 36 ]. In electrophoretic mobility shift assays, CTCF was able to form an intermolecular complex between two probes that contained target site 3 or 4, respectively, from the mouse IGF2/H19 imprinting control region. We were unable to detect bridged complexes between probes containing the FII insulator element and the c-myc P2 promoter. It is possible that CTCF acts in a different manner when binding to the IGF2/H19 ICR than at other CTCF consensus sites. It is also possible that our EMSA experiments did not contain sufficient amounts of CTCF protein to overcome the necessary entropic costs of forming bridged, cross-joined or sandwiched structures. Several proteins are known to form these DNA structures; however, sufficiently high binding energy is needed to achieve these DNA topologies. The lac repressor protein is known to form such structures only when the protein is present at sufficiently high concentrations [ 37 ]. It is difficult to compare the amount of CTCF protein generated in our in vitro translation reaction to those obtained by Pant et al . for use in their gel shift experiments [ 37 ].
The experiments presented in our paper add biochemical evidence to CTCF's loop-forming ability. DNA looping is a central phenomenon in gene regulation in both prokaryotes and eukaryotes. There are several well-characterized models of DNA looping in the control of prokaryotic gene expression including the E. coli lac and deo operons as well as the integrase protein of the bacteriophage lambda [ 38 , 39 ]. In all of these models protein:protein interactions either through the homodimerization of looping proteins or the heterodimerization of several looping proteins are required for DNA loop formation. The mechanism of CTCF DNA loop formation that we observed seems to be somewhat different. We propose a model of CTCF insulator function whereby a single CTCF protein molecule could form a DNA loop in an orientation dependent manner instead of the dimerization of two CTCF protein molecules or a bridging of two DNA molecules mediated by CTCF. By extension, a CTCF-dependent insulator element is a cis regulatory element that is the site of a DNA loop and the enhancer blocking ability of a DNA insulator function could be the result of the formation of the loop. It is interesting to note that Kyrchanova et al . [ 40 ] recently demonstrated that functional pairing interactions between Drosophila insulators was an orientation dependent interaction. Their findings fit nicely with our discovery of the orientation dependence of the unusual DNA structure formed by CTCF.
The results we have presented in this paper are a starting point towards a greater understanding of CTCF's DNA looping ability and the molecular mechanisms regulating this ability. | Conclusions
We conclude that the CTCF insulator protein is able to form an unusual DNA structure in vitro that we believe is a DNA loop. This unusual DNA structure is formed at several CTCF binding sites and is formed in a directional manner. The CTCF zinc finger domain is sufficient for the formation of the unusual DNA structure. CTCF uses DNA downstream of the CTCF binding site to form the unusual DNA structure but the sequence of this downstream DNA does not affect the formation of the structure. When the DNA sequence downstream of CTCF is topologically constrained, the unusual DNA structure is unable to form and CTCF acts as a DNA bending protein. The results of this study could provide mechanistic insights into CTCF's ability to mediate long-range chromatin interactions and form DNA loops. | Background
The CTCF insulator protein is a highly conserved zinc finger protein that has been implicated in many aspects of gene regulation and nuclear organization. The protein has been hypothesized to organize the human genome by forming DNA loops.
Results
In this paper, we report biochemical evidence to support the role for CTCF in forming DNA loops. We have measured DNA bending by CTCF at the chicken HS4 β-globin FII insulator element in vitro and have observed a unique DNA structure with aberrant electrophoretic mobility which we believe to be a DNA loop. CTCF is able to form this unusual DNA structure at two other binding sites: the c-myc P2 promoter and the chicken F1 lysozyme gene silencer. We also demonstrate that the length though not the sequence of the DNA downstream of the binding site is important for the ability of CTCF to form this unusual DNA structure. We hypothesize that a single CTCF protein molecule is able to act as a "looper" possibly through the use of several of its zinc fingers.
Conclusions
CTCF is able to form an unusual DNA structure through the zinc finger domain of the protein. This unusual DNA structure is formed in a directional manner by the CTCF protein. The findings described in this paper suggest mechanisms by which CTCF is able to form DNA loops, organize the mammalian genome and function as an insulator protein. | Authors' contributions
MJM designed and performed the experiments described in this paper. Both MJM and PDS conceived the study and wrote the manuscript. Both authors read and approved the final manuscript. | Acknowledgements
This work was supported in part by the Canadian Institutes of Health Research and the Department of Molecular Genetics at the University of Toronto. MJM is the recipient of a University of Toronto Open Fellowship. We thank Lori Frappier, Barbara Funnell and Linda Lee for their incisive comments on the manuscript. We thank Howard Lipshitz and Marc Shulman for their generous support of this work. We thank Dan Sadowski for drawing Figure 8. | CC BY | no | 2022-01-12 15:21:37 | BMC Mol Biol. 2010 Dec 21; 11:101 | oa_package/9e/b4/PMC3014928.tar.gz |
PMC3014929 | 21245907 | Introduction
It is estimated that 3.2 million stillbirths occur each year globally, 1 million of which happen during birth [1] . In addition, complications from preterm birth (before 37 completed weeks of gestation) are the leading cause of death for newborns, contributing an additional 1 million or 12% of child deaths [2] , [3] . In 2009, more than 200 stakeholders attended the International Conference on Prematurity and Stillbirth convened by the Global Alliance to Prevent Prematurity and Stillbirth (GAPPS, http://www.gapps.org/ ). The community expert group at the conference included 15 members drawn from technical and funding organizations in addition to program implementers and researchers from around the world (see Acknowledgments section for specific names). In their discussions, the group framed efforts to address preterm and stillbirths within the broader context of maternal–newborn interventions. As most of the evidence supporting these interventions emanates from research projects in controlled settings in specific contexts, the group identified the main challenge being implementing interventions at scale in different contexts. Based on these discussions, the group began a research prioritization exercise for implementation research on community-based maternal-newborn interventions that address prematurity and stillbirths at scale in different contexts. In this paper, we present the results of this exercise. | Methods
A number of research prioritization efforts have recently been applied to various health topics and health system themes [4] – [7] . The GAPPS community expert group chose the methodology proposed by the Child Health and Nutrition Initiative (CHNRI) to systematically list and score research questions. The CHNRI methodology was selected because its conceptual framework [8] – [10] has been used in numerous areas by different national and international organizations [11] – [16] (further information on CHNRI methodology, validity, and potential limitations are discussed in Table S1 ). The group followed three main stages to derive research priorities (detailed in Box 1 ). Briefly, guided by the CHNRI methodology the group evaluated 55 research questions against five main criteria:
Is the research question answerable in an ethical way?
Does the research question have the potential to reduce the disease burden (due to prematurity and stillbirths)?
Is it likely that the proposed research would address obstacles to scaling up?
Would the proposed research attract funding support and national policy attention?
Would the research results be owned by local actors, including political authorities and elected representatives, health workers, district managers, and communities?
Respondents were 39% women and diverse in terms of regional representation (26% sub-Saharan Africa, 16% Asia, 16% Latin America, 10% Europe, 32% North America). While a substantial number of respondents were based in North America, they all work full-time in developing country contexts. Half of the respondents were based in research institutions, whereas the other half were in charge of implementing programs whether through nongovernmental organizations, UNICEF country offices, or USAID headquarters. Nonrespondents were not significantly different from respondents ( Table S2 ). | Results
The research question that was highlighted as the most important out of all 55 reviewed was “Evaluate ways to reduce the financial barriers to facility births at the community level—e.g., user fee exemptions, emergency loans, conditional cash transfers, transportation vouchers, etc.” Other research questions among the top five prioritized also addressed equity issues (reaching the poor and marginalized), but also behavioral practices and skills (engaging with social norms, identifying prematurity) and service delivery (measuring and maintaining quality of care provided by community health workers). The remaining top ten research questions ( Table 1 ) include other behavioral skills and practices (thermal care and feeding for preterm babies, birth planning), concerns about how to best motivate and compensate community health workers and their supervisors, and different dimensions of making referrals more effective. Congruent with the priority need to measure and maintain quality of care by community health workers, rational drug use by community health workers and community engagement with regard to audits was also listed among the top 25 research questions that received an overall research priority score (RPS) of 0.75 or greater ( Table 2 ).
Table 3 shows the ten research questions that were assigned the lowest RPSs. Several broad policy questions (human resource planning, gender profiles, budget flows, accountability, and monitoring systems) are listed here, along with some questions related to the sequencing of community interventions and one specific question regarding private provider practice (delayed cord clamping). Questions from almost all research avenues were found among the bottom ten research questions, suggesting that no one area was completely discriminated against by the scoring. Furthermore, even these lower-ranked research questions received relatively high RPSs compared to those arising from other CHNRI exercises. The RPS for all 55 questions ranged from 0.86 to 0.56, in contrast to other CHNRI exercises, which have generated RPS ranges from 0.90 to 0.25 [12] – [16] . This suggests that respondents collectively considered all implementation research questions as fairly important.
Research questions did vary in specificity. For example, broad questions such as “evaluate community-based strategies to reach the poor and marginalized” were scored alongside very specific questions like “evaluate ways to provide thermal care and feeding the preterm baby.” Both broad and specific questions were ranked in the top and bottom ten implementation research questions, suggesting that no bias existed against the kind of question asked.
The CHNRI methodology evaluates certain dimensions of each research question according to defined criteria. For example, “Evaluate methods and levels of accountability that can be ensured” was not considered to affect disease burden and “Evaluate ways to ensure delayed cord clamping in deliveries assisted by private providers” was not scored as likely to attract funding support or national policy attention. Among the five criteria, the most discriminative was the one related to disease burden reduction, while the least discriminative was the one regarding ethical answerability.
As mentioned, the relatively high mean scores assigned to questions across all criteria (apart from disease burden reduction) indicate that most of the respondents were fairly optimistic about the value of implementation research questions. Average expert agreement (AEA) ranged from 0.82 to 0.49. Similar to other CHNRI exercises, AEA showed a direct positive association with RPSs, indicating that there was more agreement among experts about what were the priority research questions. This is a property that is inherent to the way AEA is measured: very high or very low RPS scores require high levels of expert agreement, while substantial disagreement among experts will lead to RPS moving closer to a mean value [12] - [16] .
To determine whether any systematic bias existed against certain questions due to the profile of the respondent, we analyzed scores for researchers and implementers separately. We found at least a 10% difference in the scoring assigned for 20% of the research questions ( Table 4 ). The 11 questions for which there was a significant difference between researchers and implementers are spread across each research avenue, suggesting no one particular research area was affected by this difference of opinion. In ten out of these 11 questions, implementers ranked the implementation research question as being of higher value than researchers. | Discussion
The top 25 research questions that have been prioritized span a broad range of issues ( Table 2 ). These implementation research priorities address fostering and sustaining specific behavioral skills and practices at the community level, engaging communities in monitoring service delivery through audits, and improving referral. With regard to service delivery, a host of implementation research questions about the management of community health workers, along with the health system supports they require to function, were stressed. Finally, issues of equity, financing, and referral were highlighted, reflecting how community-based approaches cannot be dealt with in isolation from broader health system concerns.
While many of the implementation research priorities identified can be generalized across community-based maternal, newborn, and child health areas, a few distinctions may be particular to this specific exercise. Issues related to referral were present three times within the top 25 research questions. There is little implementation research on linking families from homes to facilities, or referral more broadly, in low-income countries [17] – [19] . While important gains have been made with task-shifting, effective and equitable referral remains vital, because the most serious cases of prematurity and other birth complications cannot be handled at the community level.
Implementation research questions related to community engagement and some other broader policy concerns central to managing health systems, such as human resource planning and monitoring systems, were overall not given high priority by respondents. Nonetheless, even the bottom ten research questions received high RPSs relative to other CHNRI exercises. This could be because the other exercises had more discriminatory criteria or because previous exercises compared different kinds of research (basic science versus implementation research). It may be easier for experts to discern between very different research areas (basic science versus implementation research) than to discern between areas of implementation research, which they may consider to be of relatively similar importance.
In addition, many of the implementation research questions do not by themselves contribute to improved maternal newborn outcomes. Their value comes forth when combined with other implementation issues that together make a more comprehensive and coherent community-based response with linkages to primary health care service delivery. It might therefore be difficult for respondents to think about specific implementation research questions in isolation from their broader social and health systems contexts.
The partiality toward some areas of implementation research could reflect the profile of respondents. A comparison of scoring by implementers and researchers did show some differences—not across any particular kind of research question, but in the direction of the bias, with implementers ranking implementation research questions higher than did researchers. The reasons for this difference among 20% of the questions are not known, but seem to indicate that implementers perceive the results of implementation research to be more powerful if effectively implemented.
While the CHNRI methodology provides a systematic and transparent way to rank research questions that purposefully avoids biases introduced by group dynamics dominated by powerful individuals, it still is a very lengthy process to undertake. Respondents must score 55 research questions according to five criteria that have three subcomponents each, resulting in 825 dimensions to respond to in the spreadsheet. This makes it a complex spreadsheet and likely does not help response rates. Eliciting participation via e-mail alone was not successful—only 42 out of 85 experts responded to the preliminary e-mail. The 42 experts that did express interest reflected a group that was more familiar with the GAPPS conference and had current working relationships with the lead authors who managed the exercise.
Despite these drawbacks, this exercise represents an important collaboration between researchers and program implementers to jointly identify the key implementation research questions vital to improving community-based maternal and newborn interventions that address preterm and stillbirths. The exercise also developed new criteria deemed more appropriate to implementation research, which require further testing and refinement to improve their discriminatory power.
Success in reducing stillbirth and prematurity rates, and in increasing the survival of preterm infants in low-income countries, is strongly dependent on achieving high and equitable coverage with existing cost-effective interventions [20] , [21] . Yet coverage of such interventions remains unacceptably low in most countries. For example, across 68 countries with the highest mortality, only 54% of women deliver with a skilled birth attendant and 38% receive a postnatal visit [22] . Furthermore, coverage levels are particularly low among poor and rural families in these countries. Community-based interventions are therefore essential to reach population subgroups whose current access to health facilities is severely limited. The effect of expanding coverage of family and community care to 90% can by itself lead to a 15%–32% reduction in neonatal mortality [22] . Nonetheless, the knowledge gaps around how to sustain these programs at scale in different contexts remain significant. | Conclusion
While important reviews [23] – [28] have helped to spur attention to community-based maternal newborn issues, with intriguing results for specific interventions [29] , [30] , the implementation research priorities identified in this article will, we hope, help to secure further research attention and financing for this important area. Priority research areas identified include equity concerns (such as removal of financial barriers and responsiveness to the poor and marginalized), specific behavioral skills and practices, and the management of community health workers including referral care. The challenge is now raised; will communities, governments, donors, research institutions, and international organizations respond? | ICMJE criteria for authorship read and met: AG MY AB KYC IR CGV MC CR. Agree with the manuscript's results and conclusions: AG MY AB KYC IR CGV MC CR. Designed the experiments/the study: AG MY IR CGV MC. Analyzed the data: AG MY KYC IR MC. Collected data/did experiments for the study: AG CGV. Enrolled patients: AG. Wrote the first draft of the paper: AG MY. Contributed to the writing of the paper: AG MY KYC IR CGV MC CR. Participated in designing the study and contributed to writing the manuscript by way of comments: AB. Led the development of the CHNRI methodology with assistance of the large multidisciplinary team of experts: IR. Suggested the original idea: CGV. Reviewed the data analysis: CR.
¶ Membership of the GAPPS expert group on community based strategies and constraints is provided in the Acknowledgments.
Asha George and colleagues from the GAPPS group report the implementation research priorities to address prematurity and stillbirths at the community level that resulted from their recent expert consensus exercise. | Supporting Information | The authors thank the CHNRI team who helped to develop these methods, particularly Robert Black, Shams El Arifeen, and Harry Campbell. We thank Paola Canahuati, the UN translation service, Mariame Sylla, and Jose Rolando Figueroa for translating the CHNRI scoring spreadsheet from English into French and Spanish. We are grateful to Zulfiqar Bhutta, Neal Brandes, and Luc de Bernis for their expert review of an earlier draft of this manuscript.
The GAPPS expert group on community-based strategies and constraints includes those who attended the Seattle GAPPS meeting and those who participated in the research prioritization exercise: Kaosar Afsana (Bangladesh Rural Advancement Committee, Dhaka, Bangladesh), Fernando Althabe (Institute for Clinical Effectiveness and Health Policy, Buenos Aires, Argentina), Shams Arifeen (International Centre for Diarrheal Disease Research, Bangladesh, Dhaka, Bangladesh), Abhay Bang (Society for Education, Action and Research in Community Health, Gadchiroli, India), Maria Bardolet (UNICEF, Niamey, Niger), Fernando Barros, Cesar G. Victora (University of Pelotas, Pelotas, Brazil), Neal Brandes, Troy Jacobs (USAID Washington, DC, USA), Guy Clarysse (UNICEF, Dakar, Senegal), Joseph de Graft-Johnson (Saving Newborn Lives, Washington, DC, USA), Alfredo Fort (Macro International, Calverton, MD, USA), Hernando Gaitan (Institute of Clinical Research, National University of Columbia, Bogota, Columbia), Asha George, Mark Young (UNICEF New York, USA), Kate Gilroy, Vishwajeet Kumar, Henry Perry (John Hopkins Bloomberg School of Public Health, Baltimore, USA), Zelee Hill (University College London, London, UK), Hakeem Jokhio (Aga Khan University, Karachi, Pakistan), Nilda Lambo (UNICEF, Lusaka, Zambia), Tippawan Liabsuetrakul (Faculty of Medicine, Prince of Songkla University, Hat Yai, Songkla, Thailand), Stephanie Martin (Program for Appropriate Technology in Health, Seattle, WA, USA), Grace Mlava (UNICEF, Lilongwe, Malawi), David Marsh, Salim Saddrudin (Save the Children, Washington, DC, USA), Indira Narayan (BASICS/USAID Washington, DC, USA), Pius Okong (St. Francis Nsambya Hospital, Kampala, Uganda ), Adama Ouedrago (UNICEF, Niamey, Niger), Luwei Pearson (UNICEF, Addis Ababa, Ethiopia), Stefan Peterson (Karolinska Institute, Stockholm, Sweden), Shamim Qazi (WHO, Geneva, Switzerland), M. G. Sall (University of Dakar, Dakar, Senegal), Sharad Sharma (Department of Health Services, Kathmandu, Nepal), Judith Standley (Independent Program Consultant, Israel), Aboubacry Thiam (BASICS/USAID, Dakar, Senegal), Celestin Traore (UNICEF Bujumbura, Burundi).
Abbreviations
average expert agreement
Child Health Nutrition Research Initiative
Global Alliance to Prevent Prematurity and Stillbirth
Millennium Development Goal
research priority score
World Health Organization | CC BY | no | 2022-01-13 03:10:32 | PLoS Med. 2011 Jan 4; 8(1):e1000380 | oa_package/04/97/PMC3014929.tar.gz |
PMC3014930 | 21143852 | Discussion
Pneumothorax is present in about one fifth of the blunt chest trauma cases. Insertion of an intercostal tube drainage is one effective treatment and significant morbidity can be avoided by prompt pleural decompression using proper techniques [ 1 ]. Both ventral and lateral approaches are equally preferred by the clinicians and no statistically significant difference between the two approaches for functional malposition is observed [ 2 ]. Inadvertent abdominal insertion of the intercostal tube is not rare but it is diagnosed immediately by absent air column movement in tube as well as with development of pneumoperitoneum and abdominal symptoms. Injury to the stomach or bowel may bring ingested or digested food particles into the chest tube [ 3 ]. In present case, the inadvertent entry of chest tube into jejunal loop was concealed, may be, because of snug fitting of tube into jejunum which prevented leak of intestinal air and fluid into peritoneum. The air column movement was present in the tube as the proximal holes in the tube were in left chest. The drainage of bile was not apparent initially as it was mixed with more quantity of blood in chest. It was revealed only when an Ascaris worm made its way out through the tube. | Conclusion
Close observation of the chest tube drainage bag contents should be the routine practice. | Inadvertent insertion of the intercostal tube into abdomen is not rare. It can present by different ways. In the present case an Ascaris worm crept into the intercostal drainage bag to reveal the false passage of the tube. | Case report
A middle age man presented in the emergency department late night with the history of recent blunt trauma over left chest complaining of breathlessness and chest pain. Air entry was absent on the left side of chest and x-ray chest showed left pneumothorax with collapsed lung. Emergency intercostal tube drainage was planned. One and half centimeter skin was incised at fifth intercostal space in anterior axillary line. An artery forceps was inserted through the incision making its way through intercostal muscles till parietal pleura gave way. The forceps was removed and the index finger was inserted into the wound to confirm its entry into pleural cavity. The 32 French intercostal tube was held into the artery forceps and thrust through the incision into the left pleural cavity. Approximately half liter of blood was drained through the tube. Tube was fixed after confirming the air fluid column movement in the tube. Another half liter of dark blood was drained overnight. Next morning, chest x-ray showed the tube in the left chest directing downward into the costophrenic angle above the diaphragm. The left lung was well expanded and there was no air under diaphragm. In the afternoon, an Ascaris worm was noticed in the intercostals drainage bag along with fifty milliliters of blood mixed with bile (See Figure 1 ). The patient had no abdominal complaints, no air was noticed under diaphragm on erect abdominal x-ray and there was no free fluid in peritoneal cavity on ultrasonography of abdomen. Emergency exploratory laparotomy was planned suspecting bowel injury following breach of diaphragm by intercostal tube. In the laparotomy, intercostal tube was found perforating the left dome of diaphragm with tip entering into the loop of jejunum. The tube was repositioned inside the left chest and diaphragmatic rent was repaired with 2-0 polypropelene. Jejunal perforation was closed in two layers using Polyglactin (Vicryl) suture. Chest tube was removed on second day of operation and the patient made swift recovery.
Consent
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.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
PNM: Manuscript preparation, design; JHM: Manuscript review; HM: Concept; BSP: Literature search. The manuscript has been read and approved by all the authors and the requirements for authorship have been met, and each author believes that the manuscript represents honest work. | CC BY | no | 2022-01-12 15:21:37 | J Cardiothorac Surg. 2010 Dec 8; 5:125 | oa_package/50/07/PMC3014930.tar.gz |
||||
PMC3014931 | 21114825 | Background
Multiple rounds of WGD events (both allopolyploidy and autopolyploidy) cyclically increase the genetic diversity in vascular plants, and subsequently this is steadily eroded by genomic fractionation toward diploid-like genomes. Lineage-specific WGD events followed by genome repatterning and descending dysploidy toward diploid-like genomes have been revealed in several angiosperm groups and are probably best characterized in grasses [ 1 - 5 ] and Brassicales [ 6 - 11 ]. Despite still scant knowledge of the number and genealogical context of ancient WGD episodes [ 12 ], c. 15% of speciation events among extant angiosperms are associated with polyploidy [ 13 ].
Polyploidy has also played a significant role in colonization and species radiation on islands. Multiple examples of long-distance dispersals of diploid progenitors or polyploid founders followed by adaptive radiation are documented on well-studied archipelagos (Canary Islands, New Zealand, Hawaiian Islands) [ 14 - 17 ]. Remarkably, species radiation on islands is usually characterized by chromosomal stasis, i.e. uniformity of chromosome numbers/ploidy levels [ 15 - 17 ]. This means that adaptive or species radiations proceed through homoploid divergence, rather than by changing the number of linkage groups by dysploidy and/or polyploidy. The reasons for insular chromosomal stasis are most likely complex and lineage-specific, albeit the young age of radiating polyploid lines and the adaptive advantage of successful polyploid founders and their descendants are suspected as crucial factors. Although chromosomal stasis does not necessarily imply karyotypic stasis [ 17 ], only a handful of reports deal with the evolution of whole chromosome complements in island endemics. With the exception of the Hawaiian silverswords (Asteraceae), analyzed through inter-species crossing experiments and meiotic chromosome pairing configurations [ 18 ], none of the homoploid species complexes on islands has been analyzed for whole-genome collinearity.
The genus Pachycladon (Brassicaceae) comprises nine morphologically and ecologically diverse species in mainly alpine habitats of the South Island of New Zealand, and a single species occurs in alpine habitats in Tasmania [ 19 , 20 ]. The morphology of Pachycladon is so diverse that prior to the genus being recircumscribed by [ 21 ], species were also placed in Cheesemania and Ischnocarpus . Pachycladon is monophyletic [ 21 - 23 ], characterized by little genetic variation amongst species at a variety of genetic loci [ 24 ], and the species are interfertile [ 25 , 26 ]. Furthermore, six Pachycladon species analyzed karyologically all have the same chromosome number of 2n = 20 [ 27 , 28 ] and comparable genome sizes (430 to 550 Mb [ 26 , 28 ]). Pachycladon is related to Arabidopsis [ 21 , 23 ], with both these genera belonging to the polyphyletic tribe Camelineae [ 29 , 30 ]. The close relationship between these genera is underlined by the generation of a sexually derived intergeneric hybrid between A. thaliana and P. cheesemanii [ 31 ].
Based on chromosome counts and preliminary cytogenetic data, Pachycladon species were thought to have a polyploid origin (M. Lysak and P. Heenan, unpublished results). Indeed, an allopolyploid origin of the genus during the Pleistocene between ~0.8 and 1.6 mya (million years ago) has been confirmed through identification of two paralogous copies of five single copy nuclear genes [ 23 ]. Phylogenetic data of Joly et al. [ 23 ] suggested that one of the purported parents comes from the polyphyletic Camelineae or the genus Boechera (i.e. from crucifer lineage I [ 29 ]), whereas the Brassica copy from the crucifer lineage II. Our recent comparative phylogenomic study of some allopolyploid Australian Camelineae species ( Ballantinia antipoda , 2n = 12; Stenopetalum nutans , 2n = 8 and S. lineare , 2n = 10) revealed their close phylogenetic affinity to Pachycladon and other Camelineae taxa [ 10 ]. The ~6 to 9 million old allopolyploid event in the ancestry of Australian genera was found to be obscured by extensive chromosome repaterrning leading to the extant diploid-like karyotypes (n = 4-6). Such concealed WGD episodes still detectable by comparative genetic and cytogenetic analysis were classified as mesopolyploid [ 10 ]. Although both recent studies [ 10 , 23 ] argued for an allopolyploid origin of the New Zealand and Australian Camelineae genera, the unknown genome structure of Pachycladon species did not yet allow to elucidate the relationship between the two polyploid Camelineae groups.
In the present paper comparative chromosome painting (CCP) has been applied to four Pachycladon species ( Pachycladon cheesemanii , P. enysii , P. exile , and P. novae-zelandiae ) that represent the morphological, ecological and phylogenetic diversity of the genus (Figure 1 and [ 21 ]), and for which genetic maps are not available. Pachycladon enysii is a monocarpic, lanceolate and serrate-leaved, stout terminal inflorescence species of high altitude (975-2492 m) alpine greywacke rock; P. novae-zelandiae is a polycarpic, lobed-leaved, lateral inflorescence species of mid-altitude (1080-2031 m) alpine schist rock; and P. chessemanii and P. exile are polycarpic, heterophyllous, slender terminal inflorescence and generalist species of high fertility rock such as limestone, schist, and volcanics and occur from near sea-level to the alpine zone (10-1600 m altitude) [ 19 ]. We used CCP to study the extent of chromosome collinearity between the ten chromosomes of Pachycladon species and the eight chromosomes of the theoretical Ancestral Crucifer Karyotype (ACK [ 32 , 33 ]). Combining comparative cytogenetic data with already published accounts on phylogenetics, biogeography, and ecology of the genus we addressed (i) genome structure and evolution of Pachycladon species, (ii) the genome relationship to other crucifer species, particularly to the endemic Australian Camelineae taxa, and (iii) the role of major karyotypic reshuffling in the species radiation in the island setting. | Methods
Plant material
The four species of Pachycladon included in this study represent the morphological and ecological diversity of the genus [ 19 ]. Plants were cultivated in a glasshouse at Landcare Research, Lincoln, New Zealand. All species have known wild origins: P. cheesemanii (Bobs Cove, Queenstown, Otago; 168°37'E, 47°08'S), P. enysii (Mount Potts, Canterbury; 170°55'E, 43°30'S), P. exile (Awahokomo, Otago; 170°23E, 44°42'S), and P. novae-zelandiae (Old Man Range, Otago; 169°12'E, 45°20'S).
Chromosome preparation
Entire inflorescences were fixed in ethanol:acetic acid (3:1) fixative overnight and stored in 70% ethanol at -20°C until use. Selected flower buds were rinsed in distilled water and citrate buffer (10 mM sodium citrate, pH 4.8) and incubated in an enzyme mix (0.3% cellulase, cytohelicase, and pectolyase; all Sigma) in citrate buffer at 37°C for 3 h. Individual flower buds were disintegrated on a microscopic slide by a needle in a drop of citrate buffer and the suspension softened by adding 20 μL of 60% acetic acid. The suspension was spread on a hot plate at 50°C for ~0.5 min. Chromosomes were fixed by adding of ethanol:acetic acid (3:1, 100 μL) and dried with a hair dryer. Suitable slides were postfixed in 4% formaldehyde in distilled water for 10 min and air-dried.
DNA probes for fluorescence in situ hybridization (FISH)
For CCP in P. exile , on average each third Arabidopsis thaliana BAC clone was used to establish contigs corresponding to the 24 genomic blocks of the ACK [ 33 ]. For the detail composition of the BAC contigs see [ 49 ]. After initial CCP experiments in P. exile , some BAC contigs were split into smaller subcontigs to pinpoint rearrangement of ancestral blocks. (Sub)conting characterizing chromosome rearrangements in P. exile were used as CCP probes to reconstruct karyotypes of P. cheesemanii , P. enysii and P. novae-zelandiae . The A. thaliana BAC clone T15P10 (AF167571) containing 45 S rRNA genes was used for in situ localization of NORs, and A. thaliana clone pCT4.2 (M65137), corresponding to a 500-bp 5 S rRNA repeat, was used for localization of 5 S rDNA loci. The Arabidopsis -type telomere repeat (TTTAGGG)n was prepared according to [ 76 ]. All DNA probes were labeled with biotin-dUTP, digoxigenin-dUTP, or Cy3- dUTP by nick translation as described by [ 49 ].
FISH
To remove cytoplasm prior to FISH, the slides were treated with pepsin (0.1 mg/mL; Sigma) in 0.01 M HCl for 10 min, postfixed in 4% formaldehyde in 2× SSC (1× SSC is 0.15 M NaCl and 0.015 M sodium citrate) for 10 min, and dehydrated in an ethanol series (70, 80, and 96%). Selected BAC clones were pooled and ethanol precipitated. The pellet was resuspended in 20 μL of hybridization mix (50% formamide and 10% dextran sulfate in 2× SSC) per slide. The probe and chromosomes were denatured together on a hot plate at 80°C for 2 min and incubated in a moist chamber at 37°C overnight. Posthybridization washing was performed in 20% formamide in 2× SSC at 42°C. Detection of was as described by [ 49 ]. Chromosomes were counterstained with 4',6-diamidino-2-phenylindole (2 μg/mL) in Vectashield (Vector Laboratories). Fluorescence signals were analyzed with an Olympus BX-61 epifluorescence microscope and AxioCam CCD camera (Zeiss). Images were acquired separately for all four fluorochromes using appropriate excitation and emission filters (AHF Analysentechnik). The four monochromatic images were pseudocolored and merged using the Adobe Photoshop CS2 software (Adobe Systems). Pachytene chromosomes in Figure 2 were straightened using the plugin 'Straighten Curved Objects' [ 77 ] in ImageJ program (National Institutes of Health). | Results
Comparative structure of Pachycladon karyotypes
The karyotype structure of four Pachycladon species ( P. cheesemanii , P. enysii , P. exile , and P. novae-zelandiae ) has been reconstructed by comparative chromosome painting (CCP) (Figure 2 ). Considering the close phylogenetic relationship between Pachycladon and Arabidopsis [ 21 , 23 ], we assumed that both genera descended from the Ancestral Crucifer Karyotype (ACK) with eight ancestral chromosomes AK1 to AK8 [ 32 , 33 ]. Hence A. thaliana BAC clones and contigs corresponding to the 24 conserved genomic blocks (GBs) of the ACK were used as painting probes to identify collinear chromosome regions in Pachycladon species. The four reconstructed karyotypes showed overall similarity, comprising seven (sub)metacentric (PC1, PC3, PC4, PC6 - PC8, and PC10) and three acrocentric (PC2, PC5, and PC9) chromosomes with the identical arrangement of ancestral GBs (Figure 3 ). The structural uniformity of all reconstructed karyotypes suggests that this structure is the ancestral Pachycladon karyotype.
All 24 GBs were found to be duplicated within the analyzed pachytene complements displaying regular meiotic pairing (Figure 2 and 3 ). The Pachycladon karyotype comprises one AK chromosome (PC7), seven AK-like chromosomes discernible within the composite Pachycladon chromosomes (four chromosomes modified by inversions), and 14 AK-like chromosome arms (Figure 3 and Table 1 ). Thus, in total forty-three ancestral GBs (90%) remained intact and only five blocks were split within one chromosome arm (block L on PC5), between two arms of the same chromosome (W on PC6), or between two different chromosomes (D to PC1 and PC2, J to PC9 and PC10, R to PC4 and PC5). Except chromosome PC7 resembling chromosome AK7, all Pachycladon chromosomes originated through "fusion" of two or three AK chromosomes (Figure 3 ).
Evolution of the ten Pachycladon chromosomes
We have reconstructed the origin of the nine "fusion" chromosomes of the ancestral Pachycladon karyotype using the minimal number of rearrangements and assuming that the ten PC chromosomes originated from the duplicated ACK (i.e. from 16 AK chromosomes).
PC1 and PC2 chromosomes
(Figure 4A ). PC1 originated via a reciprocal translocation between chromosomes AK1 and AK2 with breakpoints in the (peri)centromeric region of AK1 (close to block B) and in the block D of AK2. The second translocation product harbouring the AK1 centromere has been involved in a subsequent reciprocal end-to-end translocation with AK5, resulting in chromosome PC2. As the four GBs (K-N) of AK5 have the ancestral position within PC2 chromosome, we infer an inactivation and/or loss of the AK5 centromere.
PC3 chromosome
(Figure 4B ). The origin of PC3 can be reconstructed as a paracentric inversion of the block D on AK2 followed by nested "fusion" of this chromosome into the (peri)centromere of AK3. The nested "fusion" required three or four breakpoints: two at the chromosome termini of AK2 and one or two at the centromere of AK3. One breakpoint would presumably disrupt the AK3 centromere, whereas two breaks at pericentromeric regions of the opposite arms would yield a dispensable minichromosome as a second translocation product.
PC4 and PC5 chromosomes
(Figure 4C ). PC4 and PC5 were generated through the reshuffling of ancestral chromosomes AK4 and AK6, and AK4, AK5 and AK6, respectively. A pericentric inversion (GBs O and P) transforming AK6 into a telocentric chromosome was followed by a reciprocal translocation between this chromosome and AK4. This translocation joined the long arm of AK4 (block J) with the AK6 telocentric (= PC4). The AK4-derived telocentric chromosome comprising only the centromere and block I has undergone a reciprocal end-to-end translocation with AK5. As the GB collinearity around the AK5 centromere between blocks L and M remained conserved, we inferred an inactivation and/or loss of this centromere on PC5. A small reciprocal translocation between the bottom arm of PC4 (block R) and the upper arm of PC5 occurred after the major reshuffling steps. A paracentric inversion between GBs K and L on PC5 could have occured before the origin of both PC chromosomes or it is a later event.
PC6 chromosome
(Figure 4D ). This chromosome most likely originated via a reciprocal end-to-end translocation between AK6 and AK8 and was probably followed by a concurrent or subsequent inactivation and/or loss of the AK8 centromere, reflected by the ancestral position of blocks V and Wa on PC6. This event was followed by a pericentric inversion with breakpoints in the (peri)centromeric region (close to block Q) and within block W.
PC8 chromosome
(Figure 4E ). PC8 originated via a reciprocal translocation between AK7 and AK8, yielding the fusion PC8 chromosome and a meiotically unstable minichromosome containing the centromere of AK8. The translocation was preceded by a paracentric inversion on AK7 (block S) and pericentric inversion on AK8 (block V).
PC9 and PC10 chromosomes
(Figure 4F ). Chromosome PC9 originated through a reciprocal translocation between AK1 and AK4 with breakpoints in the AK1 pericentromere (close to block B) and the proximal part of the bottom arm of AK4 (block J). The second translocation product (GBs C and Jb, and the AK1 centromere) participated in a reciprocal end-to-end translocation with AK3 which resulted in the origin of PC10 and small acentric fragment. Ancestral arrangement of AK3-derived GBs suggests that the AK3 centromere has been lost or inactivated.
The reconstructed chromosome origins are congruent with the reduction of 16 ancestral chromosomes (centromeres) to only 10 in Pachycladon . Centromeres of both homeologues of AK1, AK2, AK4, AK6 and AK7 remained functional, whereas six centromeres were lost (Figure 3 ). The centromere of one AK3 homeologue was eliminated due to the nested chromosome fusion, one AK8 centromere was eliminated via symmetric translocation, and both AK5 centromeres and centromeres of second homeologues of AK3 and AK8 were inactivated/lost (Figure 4 ).
Out of the 33 breakpoints inferred for the origin of ten Pachycladon chromosomes, 12 (36%) map to pericentromeric regions, 16 (49%) to telomeric regions, whereas only five (15%) occurred within GBs (Figure 4A to 4F ).
Chromosome landmarks (heterochromatin, telomeres and rDNA)
We have analyzed mitotic and pachytene chromosome complements of the four Pachycladon species for the distribution of heterochromatin domains, localization of ribosomal RNA genes (rDNA) and the Arabidopsis -type telomere repeat (Figure 5 ). Except for prominent heterochromatin of pericentromeres and terminal nucleolus organizing regions (NORs) (Figure 2 ), a single heterochromatic knob occurs in P. enysii and two knobs were found in P. exile . Whereas two of the three knobs reside within genomic blocks (B on PC1 in P. enysii , U on PC7 in P. exile ), the knob on the bottom arm of PC10 in P. exile is localized between blocks G and H, i.e. at the site of presumably inactivated centromere of AK3 (Figure 4F ). No heterochromatic domains were observed at the sites of other presumably inactivated and/or lost paleocentromeres. The telomere (TTTAGGG)n repeat hybridized only to chromosome ends and no interstitial telomeric signals were observed (data not shown). Whereas the four species have a single 5S rDNA locus at the same position, the number of terminal 45S rDNA loci varies. P. novae-zelandiae has one, P. cheesemanii and P. exile possess two, and P. enysii has three 45S rDNA loci, with 45S locus on the upper arm of PC2 being common to all species (Figure 5 ). Thus, the cross-species karyotypic stasis does not apply to the number of terminal 45S rDNA loci. | Discussion
We have used comparative chromosome painting to reconstruct karyotype structure and evolution in the genus Pachycladon . Interestingly, our analysis showed that the four analyzed species representing the phylogenetic, ecological and morphological diversity of the genus possess an identical karyotype, which is also most likely to be the ancestral karyotype of the genus Pachycladon .
Chromosomal and karyotypic stasis in Pachycladon
The present study of four Pachycladon species is the first whole-genome analysis of an island species radiation. Pachycladon species have uniformly ten chromosomes [ 27 , 28 ] and this infrageneric chromosomal stasis has been now extended for karyotypic stasis. Overall similar genome structures supported the monophyletic origin of the genus [ 21 - 23 ] and allowed inference of the ancestral Pachycladon karyotype whose structure remained conserved in the extant species. Karyotypic stasis revealed in Pachycladon clearly indicates that the Pleistocene species radiation on the South Island of New Zealand [ 19 ] was not associated with major chromosome rearrangements. The four karyotypes differ only by the number of heterochromatic knobs and NORs, without an apparent link to infrageneric phylogenetic relationships (Figure 1 ). Hence, the speciation proceeded through homoploid divergence from the ancestral allopolyploid genome.
Perhaps with the exception of meiotic studies in the Hawaiian silversword alliance [ 18 ] there is virtually no data on karyotype evolution during island angiosperm speciation. Hence, only the variation in chromosome number/ploidy level can be discussed more extensively. Several surveys of angiosperm chromosome numbers showed the trend of chromosomal stasis during species radiation on islands (see reviews by [ 15 - 17 ]). This tendency might appear paradoxical considering geographical isolation and a wealth of diverse insular environments potentially promoting the origin of novel chromosomal races and karyotypes. However, genomes diverging on islands are under multiple constraints determining chromosomal stasis or chromosomal variation. As self-evident factors influencing the insular species radiation and genomic stability are the age of islands and their distance from the mainland, the number of colonization events, the incidence of polyploidy and phylogenetic constraints. Colonizations followed by adaptive radiation on (volcanic) islands represent often relatively young evolutionary events and therefore many island endemics represent monophyletic lineages comprising closely related species with uniform chromosome numbers. Furthermore, it was concluded that chromosomal stasis vs. lability is under a strong phylogenetic constraint as some lineages (e.g. Asteraceae, Sideritis ) seem to be more prone to genome reshuffling than others [ 15 , 16 , 34 ].
Generally the low incidence of polyploidy has been claimed for island floras [ 16 ]. These estimates collated prior to the era of indepth whole-genome analyses revealing multiple whole-genome duplications of a different age (e.g., [ 7 , 12 , 10 ]) had to be, by definition, too conservative. Recent studies suggest that colonization of islands has been frequently associated with hybridization and allopolyploidy (see [ 35 , 36 ] for examples). Allopolyploid ancestors originated either on continents and spread to islands (e.g. the allopolyploid ancestor of the Hawaiian mints [ 37 ]) or diploid ancestors hybridize in situ after long-distance dispersal (e.g. the New Zealand and Australian Lepidium species [ 38 ]). The allopolyploidy-driven speciation on islands is frequently associated with chromosomal stasis as shown for the Hawaiian flora with the high incidence of polyploidy (> 80% [ 15 ]).
Polyploidy is also a pronounced feature of the New Zealand flora, with 72% of the species being polyploid in families with 25 or more species [ 39 ]. Chromosomal features of New Zealand plants indicative of polyploidy are the high number of species with even haploid numbers and/or haploid numbers n = > 10-14 [ 40 ]. Many of the polyploid genera that are like Pachycladon exhibiting chromosomal stasis are species-rich and generally considered to be recent species radiations, often into mainly alpine-montane habitats. They include, for example, Aciphylla (42 species, 2n = 22), Brachyglottis (30 species, 2n = 60), Chionochloa (22 species, 2n = 42), Gentianella (40 species, 2n = 36), Epilobium (38 species, 2n = 36), and Ourisia (20 species, 2n = 48) (data from [ 27 ]).
Chromosomal stasis is also observed in the few crucifer genera that have species radiations on islands. All seven Parolinia species endemic to the Canary Islands probably have 2n = 22 (4 species counted [ 41 ]), and seven shrubby species of Descurainia endemic to the Canary Islands share 2n = 14 [ 42 ]. Similarly, of nine Diplotaxis species in the Cape Verde Islands, five have 2n = 26 [ 43 ]. Unfortunately, insufficient chromosomal data are available for c. 40 Cardamine species endemic to New Zealand (P. Heenan, unpublished data) as well as for most crucifer genera endemic to Australia [ 41 , 44 ].
Pachycladon karyotype is derived from the duplicated Ancestral Crucifer Karyotype
Our data suggest that the ancestral Pachycladon karyotype (n = 10) was derived from the duplicated Ancestral Crucifer Karyotype (n = 8) through allopolyploidy. The ACK was expected to be inferred as an ancestral genome of Pachycladon , as all of the Camelineae genomes analyzed thus far have descended from the ACK (for instance, Arabidopsis , Capsella , Turritis , and Neslia [ 32 , 45 , 46 ], including the analyzed Australian Camelineae species [ 10 ]). Furthermore, the karyotypes of Crucihimalaya and Transberingia , two genera often found as being the closest relatives of Pachycladon [ 21 , 23 ], resemble the ACK structure [ 10 ]. Similarly, the ACK was proposed as an ancestral karyotype for tribes Boechereae and Cardamineae [[ 47 ], Mandáková and Lysak, unpublished data]. It is likely, therefore, that the ancestral Pachycladon genome has been derived from the hybridization between two ACK-like genomes. The primary allopolyploid had either the structure of duplicated ACK with n = 16 or the participating genome(s) were reduced (n = 8 → n = 7-5?) prior to the hybridization event and the allopolyploid had less than 16 chromosome pairs. The fact that paralogous genomic blocks do not lay on the same chromosome suggests that the modern Pachycladon chromosomes were reshuffled prior to the hybridization event, rather than due to homeologous recombination between two ACK-like genomes within the allopolyploid ancestor.
The ten composite Pachycladon chromosomes originated through inversions, reciprocal translocations and centromere inactivation/loss events within the duplicated ACK complement (Figure 4 ). Chromosome "fusions" were mediated by reciprocal translocations with or without preceding para- and pericentric inversions. These translocations yielded a "fusion" chromosome and (a)centric fragment as the second translocation product. Small acentric fragments and the minichromosome harbouring one AK8 centromere were meiotically unstable and eliminated. Whereas Robertsonian-like translocations eliminating one AK centromere together with the minichromosome is a common mechanism of the karyotype evolution in Brassicaceae [ 32 , 48 , 49 ], asymmetric translocation events yielding miniature acentric fragments and dicentric chromosomes with one AK centromere apparently inactivated or removed by recombination were proposed for the origin of composite chromosomes in the Australian Camelineae species [ 10 ]. Centromere inactivation and/or loss has been inferred on bottom (long) arms of four Pachycladon chromosomes (PC2, PC5, PC6, and PC10) based on the absence of ancestral centromeres and conserved organization of adjacent genomic blocks (Figure 3 and 4 ). The incidence of centromere inactivation in Australian and New Zealand Camelineae species might be tentatively related to the common ancestry of both lineages and/or to the duplicated character of the allopolyploid ancestral genomes. Centromere inactivation of AK4 can also be suggested for the origin of chromosome At2 in A. thaliana [ 32 ], and centromere inactivation of AK5 for the origin of Bst5 in Boechera stricta [ 47 ] and chromosome AK4/5 in Neslia paniculata [ 32 ]. Nevertheless, an alternative mechanism of centromere removal through subsequent paracentric and pericentric inversions followed by a symmetric translocation (Figure 2C in [ 32 ]) is also plausible, though more breakpoints have to be considered. A dicentric chromosome could also be stabilized by intrachromosomal translocation, with breakpoints in pericentromeric region of one of the ancestral centromeres, followed by a loss of the resulting centric fragment.
Chromosome PC3 originated probably through a nested "fusion" of chromosome AK2 between chromosome arms of AK3. As both AK chromosomes within PC3 possess the ancestral structure of genomic blocks (except inverted block D) translocation events with breakpoints at chromosome termini of AK2 and centromere of AK3 seems to be the parsimonious scenario. In grasses (Poaceae), insertional chromosome "fusion" has been proposed as a general mechanism of descending dysploidy [ 4 , 50 ], whereas in crucifers it can be suggested only for the origin of chromosome AK2/5 in Hornungia alpina [ 32 ]. Thus, Pachycladon chromosome PC3 is most likely the first instance of reconstructed insertional dysploidy in Brassicaceae. An alternative mechanism of the PC3 origin via end-to-end reciprocal translocation coupled with the elimination of the AK3 centromere requires two more breakpoints.
Common origin of Pachycladon and Australian Camelineae species?
Based on the phylogenetic analysis of Australian Camelineae taxa and Pachycladon species, [ 10 ] concluded that both groups might originate from a very similar allopolyploid ancestor. Although the authors could not reject a single origin of both lineages, they considered two successive allopolyploidization events as more likely, i.e. mesopolyploid Australian Camelineae species originated and radiated in continental arid habitats before the mesopolyploid ancestor of Pachycladon . The present data corroborate this conclusion. Specifically, the two species groups do not share any cytogenetic signature, i.e. a taxon/lineage-specific chromosome rearrangement, such as the rearranged AK8 homeologue shared by five Australian species analyzed [ 10 ]. In the Australian species, any two paralogous GBs differ by the length and fluorescence intensity as revealed by CCP [ 10 ]. This difference was either present already in the hybridizing progenitors or was caused by preferential fractionation of paralogous regions belonging to only one subgenome [ 51 ]. In Pachycladon , two paralogous copies of all GBs cannot be distinguished upon CCP analysis. Furthermore, higher chromosome number in Pachycladon species (n = 10) than in the Australian species (n = 4-7) implies a more recent origin and less extensive diploidization in Pachycladon . Indeed, the significantly lower number of non-ancestral junctions of genomic blocks in Pachycladon compared to Ballantinia antipoda and the two Stenopetalum species (Table 1 and [ 10 ]) underlines the less extensive genome reshuffling in Pachycladon . Also the number of split GBs in Pachycladon (10%) is lower than in the Australian species (13% to 19%; [ 10 ]). Interestingly, both groups do not differ substantially by the number of preserved AK chromosomes, chromosome arms and GBs (Table 1 ). This comparison suggests that the most recent steps of chromosome number reduction in the Australian Camelineae species have been mediated by tandem end-to-end translocations followed by centromere inactivation/loss, not disrupting the structure of AK-like chromosomes and chromosome arms.
Altogether, the differences in genome structure between the mesopolyploid Australian and New Zealand lineage indicate two successive WGD events involving the same pool of parental species. The existence of the progenitor species in Australia for a long period of time is a credible assumption considering the remarkable stasis of the ACK and AK chromosomes across crucifer lineages I and II [ 32 , 33 , 49 ]. Further research is needed to elucidate if the ancient ACK-like karyotype could be found in some not yet analyzed Australian crucifer species. Recurrent formation of allopolyploids from the same or closely related parents has been documented, e.g. in the North American allopolyploid species of Tragopogon [ 52 ], in Persicaria [ 53 ] or Arabidopsis kamchatica [ 54 ], and also proven by the generating synthetic allopolyploids as Arabidopsis suecica [ 55 ], tobacco [ 56 ] or Tragopogon mirus and T. miscellus [ 57 ].
Although less likely, we cannot rule out that karyotypic change in the Australian Camelineae species and in Pachycladon had significantly different dynamics. The Australian Brassicaceae species exhibit a predominantly annual growth habit [ 44 ] in comparison to the perennial Pachycladon [ 21 ], and a more rapid rate of genome evolution could therefore be brought about with faster nucleotide substitution rates that occur in many annuals [ 58 , 59 ]. Perennials are thought to have greater chromosomal stasis than annuals [ 60 , 61 ]. Certainly the annuality could have accelerated genome reshuffling in the Australian lineage. However, for Brassicaceae we have insufficient data on large-scale genome evolution in relation to the life forms, reproduction systems and ecological factors, and as noted by [ 15 ] and [ 62 ] chromosomal evolution is often stochastic and does not obey the models.
Phylogeographic scenario of the origin of Pachycladon
Pachycladon is the only New Zealand genus from the polyphyletic tribe Camelineae (other endemic crucifer species belong to Cardamineae, Lepidieae, and Notothlaspideae), and therefore an in situ origin seems unlikely. The closest Camelineae relatives of Pachycladon occur in Australia (e.g., Arabidella , Ballantinia , and Stenopetalum ) and Eurasia/Beringia (e.g., Arabidopsis , Crucihimalaya , Transberingia ) [ 10 , 23 ]. It seems more plausible that the hybridization event giving rise to Pachycladon has taken place on the Australian continent.
There are strong taxonomic and biogeographic links between Australia and New Zealand and dispersal across the Tasman Sea can occur in both directions. Tasmania and New Zealand have about 200 species in common [ 63 ], and there are many genera in continental Australia and New Zealand that have species that are closely related (e.g., Aciphylla , Celmisia , Gentianella , Melicytus , and Ranunculus ). For these shared genera, species diversity is often highest in New Zealand and the Australian species are considered to be the result of westward dispersal from New Zealand and subsequent speciation (e.g. [ 64 , 65 ]). Indeed, P. radicatum occurs in the Tasmanian mountains and is considered to have dispersed there and diverged contemporaneously with the radiation of Pachycladon in New Zealand [ 21 , 22 ]. Other taxa are also shared between the two countries, but these are considered to have dispersed eastward from Australia to New Zealand and include, for example, Craspedia [ 66 ], Montigena [ 67 ], Poranthera [ 68 ], Scleranthus [ 69 ], and Stylidiaceae [ 70 ]. This pattern of eastward dispersal means it is plausible that Pachycladon could have originated in Australia and then subsequently dispersed to New Zealand.
An alternative scenario of the origin of the Pachycladon allopolyploid ancestor in (eastern) Eurasia followed by a later dispersal to New Zealand is unlikely and incongruent with the close phylogenetic ties of Pachycladon to Australian Camelineae. Also, the origin of Pachycladon and Australian crucifer species in New Zealand is very unlikely, considering the diversity of endemic Australian Brassicaceae taxa (15 genera and 65 species [ 44 ]).
Many of the Australian Camelineae are distributed in the arid Eremaean Zone and/or the southeastern temperate biome [ 44 ], whereas in New Zealand Pachycladon mainly occupies montane-alpine habitats. These three environments have expanded in both Australia and New Zealand during the Pliocene and Pleistocene and are generally considered important drivers of species radiations (e.g. [ 71 , 72 ]). For the Australian Camelineae their origin and diversification ~6 to 9 mya [ 10 ] is consistent with other dated molecular phylogenies of a diverse range of arid-adapted taxa [ 73 ]. These dated phylogenies show the deepest divergences of taxa are consistent with the beginning of the formation of the arid zone in the mid-Miocene and that most arid-zone species lineages date to the Pliocene or earlier. The molecular clock date of 0.8 to 1.6 mya for the origin of Pachycladon [ 23 ] is also consistent with its alpine distribution and habitats in the Southern Alps in the South Island of New Zealand [ 19 ]. Uplift of the Southern Alps occurred over the last 8 million years, but only reached a suitable height to permanently support alpine plants during the Pleistocene.
Reconstructed genome evolution corroborates the close relationship of Pachycladon to Arabidopsis and other Camelineae species
The phylogenetic position of Pachycladon has been investigated repeatedly using various nuclear, chloroplast, and mitochondrial genes [ 10 , 21 , 23 ]. All studies are congruent in placing the genus into the crucifer lineage I, within the polyphyletic tribe Camelineae [ 21 , 29 , 30 , 74 ]. Although the phylogenetic relationships within Camelineae are unclear, these studies have shown Transberingia and Crucihimalaya (Camelineae), Sphaerocardamum (Halimolobeae), Physaria (Physarieae), and Boechera (Boechereae) to be among the closest relatives of Pachycladon . Based on the analysis of five single-copy nuclear genes, [ 23 ] showed that Pachycladon has an allopolyploid origin and that the two genomes were associated with two divergent Brassicaceae lineages (lineage I and II [ 29 , 75 ]). One putative parental genome was associated with Camelineae sensu lato (and Boechereae) and the second genome being related to Brassiceae, Sisymbrium , Eutremeae, Thlaspideae, and remarkably also to Cardamineae on the chalcone synthase gene tree. This pattern has been interpreted as the evidence of an inter-tribal allopolyploidization event at the origin of Pachycladon .
A recent study using nuclear, mitochondrial and chloroplast genes, as well as significantly increasing the sampling of Australian Camelineae (in comparison to that of [ 23 ]), has confirmed the allopolyploid origin of Pachycladon and provides confidence that the two gene paralogues that constitute Pachycladon are derived from within lineage I [ 10 ]. Most importantly, this study has disclosed the close relationship of Pachycladon to the Australian genera Arabidella , Ballantinia , and Stenopetalum , and the maternal gene paralogues of Pachycladon and these three genera clustered with Eurasian Camelineae ( Arabidopsis , Capsella, Crucihimalaya , Olimarabidopsis , Transberingia ) and North American Boechereae. The position of the paternal gene copy was less evident, but it was always embedded within lineage I, and therefore different from the study by [ 23 ]. Mandakova et al. [ 10 ] and the present study convincingly show that the Pachycladon ancestor orginated from hybridization between a Camelineae species and either another species of that tribe or a very closely related tribe of lineage I. Future phylogenomic analyses of other Australian crucifer genera are likely to further resolve the parentage and phylogenetic relationships of Pachycladon . | Conclusion
We have shown that the remarkable infrageneric morphological and ecological differentiation in Pachycladon is characterized by the genome stability manifested as chromosomal and karyotypic stasis. The monophyletic Pachycladon species descended from a common allopolyploid ancestor (n = 10) through a whole-genome duplication of the Ancestral Crucifer Karyotype (n = 8) and subsequent diploidization by descending dysploidy. Furthermore, the present study and the phylogenetic data of [ 10 ] clearly demonstrate the close relationship between the allopolyploid Pachycladon and the allopolyploid Australian Camelineae taxa. CCP data demonstrate that both mesopolyploid groups most likely originated from two different WGD events that involved identical or very similar diploid parents. We argue that the Pachycladon ancestor has its origin in Australia and later dispersed to the South Island of New Zealand. The endemic Australian and New Zealand Camelineae provide an excellent framework to examine the nature and consequences of differently-aged WGD events within a complex of closely related species. | Background
Pachycladon (Brassicaceae, tribe Camelineae) is a monophyletic genus of ten morphologically and ecogeographically differentiated, and presumably allopolyploid species occurring in the South Island of New Zealand and in Tasmania. All Pachycladon species possess ten chromosome pairs (2n = 20). The feasibility of comparative chromosome painting (CCP) in crucifer species allows the origin and genome evolution in this genus to be elucidated. We focus on the origin and genome evolution of Pachycladon as well as on its genomic relationship to other crucifer species, particularly to the allopolyploid Australian Camelineae taxa. As species radiation on islands is usually characterized by chromosomal stasis, i.e. uniformity of chromosome numbers/ploidy levels, the role of major karyotypic reshuffling during the island adaptive and species radiation in Pachycladon is investigated through whole-genome CCP analysis.
Results
The four analyzed Pachycladon species possess an identical karyotype structure. The consensual ancestral karyotype is most likely common to all Pachycladon species and corroborates the monophyletic origin of the genus evidenced by previous phylogenetic analyses. The ancestral Pachycladon karyotype (n = 10) originated through an allopolyploidization event between two genomes structurally resembling the Ancestral Crucifer Karyotype (ACK, n = 8). The primary allopolyploid (apparently with n = 16) has undergone genome reshuffling by descending dysploidy toward n = 10. Chromosome "fusions" were mediated by inversions, translocations and centromere inactivation/loss. Pachycladon chromosome 3 (PC3) resulted from insertional fusion, described in grasses. The allopolyploid ancestor originated in Australia, from the same or closely related ACK-like parental species as the Australian Camelineae allopolyploids. However, the two whole-genome duplication (WGD) events were independent, with the Pachycladon WGD being significantly younger. The long-distance dispersal of the diploidized Pachycladon ancestor to New Zealand was followed by the Pleistocene species radiation in alpine habitats and characterized by karyotypic stasis.
Conclusions
Karyotypic stasis in Pachycladon suggests that the insular species radiation in this genus proceeded through homoploid divergence rather than through species-specific gross chromosomal repatterning. The ancestral Pachycladon genome originated in Australia through an allopolyploidization event involving two closely related parental genomes, and spread to New Zealand by a long-distance dispersal. We argue that the chromosome number decrease mediated by inter-genomic reshuffling (diploidization) could provide the Pachycladon allopolyploid founder with an adaptive advantage to colonize montane/alpine habitats. The ancestral Pachycladon karyotype remained stable during the Pleistocene adaptive radiation into ten different species. | Authors' contributions
MAL and PBH conceived the study. TM carried out the research. TM, PBH and MAL analyzed the data and wrote the manuscript. All authors read and approved the final paper. | Acknowledgements
We acknowledge excellent technical assistance of Kateřina Toufarová. This work was supported by research grants from the Grant Agency of the Czech Academy of Science (IAA601630902) and the Czech Ministry of Education (MSM0021622415), MAL was supported by a Humboldt Fellowship. PBH was supported by the New Zealand Foundation for Research, Science and Technology through the Defining New Zealand's Land Biota OBI and the Marsden Fund. | CC BY | no | 2022-01-12 15:21:37 | BMC Evol Biol. 2010 Nov 29; 10:367 | oa_package/ed/ed/PMC3014931.tar.gz |
PMC3014932 | 21126350 | Background
Rattus is the most speciose genus among mammals with some 66 described species [ 1 ]. The genus is thought to have originated in the general region of the Indonesian Islands with subsequent dispersal into continental Asia, New Guinea and Australia [ 2 ]. Over the last five million years (My) a combination of tectonic and climatic change meant that this geographic region experienced major fluctuations in sea level and in the extent of emergent land mass [ 3 ]. These changes provided opportunities for dispersal and speciation within Rattus .
Previous studies [ 4 , 5 ] based on mitochondrial DNA (mtDNA) identified two major groups within Rattus . These are the Asian group, comprising rats from Southeast Asia and Island Southeast Asia, and the Australo-Papuan group comprising rats from Australia and New Guinea. The Asian group includes the three most widely distributed rat species, Rattus rattus and Rattus norvegicus which were dispersed around the world via European sailing ships [ 6 ] and Rattus exulans which was transported throughout the Pacific by prehistoric peoples [ 7 ]. The Australo-Papuan group of rats are largely restricted to Australia and New Guinea with the exception of Rattus praetor that also occurs in the Solomon Islands in addition to New Guinea. In addition subfossil remains of this rat have been found at archaeological sites as far east as Fiji indicating that prehistoric peoples in all likelihood transported it into the Pacific [ 8 ]. While the Robins et al. [ 4 ] study shows good species resolution among the rats of Australia the deeper relationships were not well resolved. The focus of this investigation is to use mtDNA, including whole mitochondrial genomes, to help evaluate these evolutionary relationships and to date the divergences among the Rattus species of Australia.
Rattus has been included in other recent molecular phylogenetic analyses. Jansa et al. [ 9 ] investigated relationships among endemic rodents in the Philippine Islands, using single nuclear (exon 1 of IRBP) and mitochondrial (cytochrome b (cyt b )) markers, but they included only one species from the Australo-Papuan group. Rowe et al. [ 10 ] sampled broadly among the genera of murine rodents with the aim of inferring relationships among the "Old Endemics" of the Australo-Papuan, or Sahul, region. Their use of six nuclear (exon 10 of GHR, exon 11 of BRCA1, the single large exon of RAG1, intron 3 and flanking regions of BDR, exon 1 of IRBP and intron 2 and flanking regions of ATP5) and six mitochondrial markers (the genes cytochrome c oxidase subunit I (COI), cytochrome c oxidase subunit II (COII), ATPase 8 and cyt b , plus the two tRNA's between COI and ATPase 8) enabled the resolution of deep divergences but their inclusion of only three Rattus species means that there is little overlap with our study. Most recently Pagès et al. [ 11 ] developed a large dataset of Rattus and related species based on one nuclear (exon 1 of IRBP) and two mitochondrial markers (cyt b and COI). That study, however, focussed on the species found in Southeast Asia and did not include any of the species endemic to the Australo-Papuan region.
There has been considerable debate regarding the number of rat species in Australia and New Guinea and the relationships among them. Taylor and Horner [ 12 ] revised the systematics of the native Australian rats and recognized five species R. fuscipes , R. leucopus , R. lutreolus, R. sordidus and R. tunneyi . Based on a multivariate analysis of craniometric data Taylor et al. [ 13 ] included R. fuscipes and R. leucopus within a clade containing New Guinean native rat species whereas the other Australian native rat species ( R. sordidus , R. tunneyi and R. lutreolus ) formed a separate but deeper clade. A phylogeny based on isozyme electrophoresis of 55 loci of the Australian Rattus species, together with R. norvegicus and R. rattus as outgroups, [ 14 ] showed similarities with the phenogram of Taylor et al. [ 13 ] with R. lutreolus diverging early, followed by R. tunneyi and lastly a more recent divergence of a clade containing R. sordidus , Rattus colletti and R. villosissimus . This study was unable to resolve the deeper relationships among R. fuscipes, R. lutreolus and R. leucopus and since it did not include rats from New Guinea it could not address the relationship between R. fuscipes and the New Guinean rats.
In the recent taxonomic study by Musser and Carleton [ 1 ] the rats of Australia and New Guinea are divided into two groups. The Rattus fuscipes species group comprises four of the five Australian native rats described by Taylor and Horner [ 12 ] while the fifth, R. leucopus , is placed among the New Guinean species in the Rattus leucopus species group. Musser and Carleton [ 1 ] and Baverstock et al. [ 14 ] give R. colletti and R. villosissimus full species status whereas Taylor and Horner had classified them both as sub-species of R. sordidus . The Rattus leucopus species group of Musser and Carleton [ 1 ] comprises the 14 species; Rattus arfakiensis, Rattus arrogans, Rattus giluwensis, Rattus jobiensis, R. leucopus, Rattus mordax, Rattus niobe, Rattus novaeguineae, Rattus omichlodes, R. praetor, Rattus richardsoni, Rattus steini, Rattus vandeuseni and Rattus verecundus . Of the Australian and New Guinean species, only two are known to occur in both regions and the results of the craniometric study of Taylor et al. [ 13 ] strongly supported the hypothesis of Tate [ 15 , 16 ] that R. sordidus colonised southern and south-eastern New Guinea from Australia, and R. leucopus colonised Cape York in north-eastern Australia from New Guinea. The presence of ephemeral land bridges between Australia and New Guinea, especially in the area of the Torres Strait, is believed to be associated with glaciation cycles during the last 2.6 - 3 million years and would have provided opportunities for such faunal interchange [ 17 , 18 ]. Aplin (quoted by Musser and Carleton [ 1 ] pg 1462), suggested on the basis of preliminary cyt b data that "members of the R. fuscipes and the R. leucopus groups each may represent discrete radiations, albeit closely related and with some exchange across the Torres strait." Aplin's [ 17 ] schematic phylogeny, although it included some New Guinean Rattus species, was similar to that of Baverstock et al. [ 14 ]. Aplin placed R. lutreolus basal in the clade containing R. tunneyi , R. sordidus , R. colletti and R. villosissimus and did not resolve the position of R. fuscipes . He stated that although both positions were uncertain R. fuscipes was probably more closely related to the New Guinean rat cluster and R. lutreolus to the Australian rat cluster. It is thus fair to say that there is a need for further clarification of the relationships within this speciose group.
In an earlier study Robins et al. [ 4 ] used three mitochondrial regions (D-loop, cyt b and COI) to identify Rattus species. Several well-defined clades were recovered. In many cases these corresponded to named species, that is all members of the clade had the same species label assigned at source by either the collector or the museum. In some cases, clades contained multiple species labels, possibly resulting from misidentification, faulty taxonomy or biological processes such as introgression. All 16 clades were assigned a name, as shown in Figure two of Robins et al. [ 4 ], effectively making them operational taxonomic units (OTUs) although 10 of these names corresponded well to a species. These OTUs are also used in this paper.
From the sample base of the OTUs in Robins et al. [ 4 ] six rats were chosen for whole mitochondrial genome analysis and together with data from GenBank divergence times and relationships among five Rattus species were inferred [ 5 ]. That 2008 study, however, included only one rat species from the Australo-Papuan region. To increase our understanding of the relationships among the rats of this region we sequenced seven whole mitochondrial genomes from six Rattus species from the Australo-Papuan clade; R. fuscipes , R. lutreolus , R. sordidus , R. tunneyi , R. villosissimus and two samples of R. leucopus (one originating from Australia and one from New Guinea). We have again chosen rats from the sample base of Robins et al. [ 4 ] except for the more recently acquired samples of R. lutreolus and R. villosissimus . We have sequenced whole mitochondrial genomes from four species unique to Australia and two occurring in both Australia and New Guinea. We used these data together with data from GenBank to estimate divergence times and to examine the evolutionary relationships among the Australian rats. We have compared the phylogenies inferred from whole mitochondrial genomes (n = 16) with those obtained by analysing the hypervariable region of the D-Loop together with two mitochondrial gene regions (cyt b and COI) from the same sample set and also from a much larger sample set (n = 126). Here we show that the use of this subset of the mitochondrial genome, while generally reliable for species identification is inadequate for dating or resolving deeper relationships. Our analyses of 16 whole mitochondrial genomes, on the other hand, have made a substantial contribution to our understanding of the relationships among the Australo-Papuan Rattus species and the timing of their divergences. | Methods
Whole mitochondrial genomes were obtained from seven rats representing six different species (see sample details in Table 1 ). The species were R. fuscipes , R. leucopus, R. lutreolus, R. sordidus, R. tunneyi , and R. villosissimus . Two specimens of R. leucopus (one from New Guinea and the other from Australia) were processed because this species occurs in both New Guinea and in the Cape York region of Australia. These seven novel genomes were compared with other available whole mitochondrial genomes from wild caught rats. There are a number of mitochondrial genomes available from highly inbred strains of R. norvegicus (e.g., 10 of the 12 genomes from Schlick et al. [ 19 ]) which we did not include in the analysis. DNA was extracted from an additional two samples of R. lutreolus and three samples of R. villosissimus (see sample details in Table 2 ). Three mitochondrial regions were sequenced from these samples (750 bp of COI, 762 bp of cyt b and 585 bp of D-loop) and analysed together with the dataset from Robins et al. [ 4 ].
DNA was extracted from liver or muscle tissue using the High Pure PCR Template Preparation Kit (Roche). Mitochondrial DNA was amplified in four overlapping long range pieces from 4 to 6 kb in length. See Table 3 for amplicon length and primer details. Long range PCR was performed using the DNA polymerase Taq and protocols of the Expand Long Template PCR System (Roche). For short-range PCR the amplification reactions contained: TrisHCl pH 8.3 10 mM; KCl 50 mM; forward and reverse primers at 0.5 μM each; dNTPs at 0.15 mM each; 0.5 U of Taq polymerase; 1 μL of DNA template. We used a standard 3-step amplification protocol: 94°C, 3 min, 35 cycles of: 94°C, 30 s, 60°C, 30 s, 72°C, 1 min followed by 1 cycle of 72°C, 5 min and a hold at 15°C. PCR products were visualized and quantified on ethidium bromide stained 1% agarose gels using a low mass ladder from Invitrogen for comparison and purified using either ExoSAP-ITTM from USB Corporation or band cut and column cleaned using the QIAquick Gel Extraction KitTM from QIAGEN.
All sequencing was done by The Allan Wilson Centre Genome Service at Massey University, Palmerston North, New Zealand. The long range products from each sample except R. fuscipes which was sequenced as part of a separate experiment, described in McComish et al. [ 20 ] were pooled and processed for sequencing using the multiplexed sequencing kit from Illumina. A 75-bp single read run was performed on an Illumina Genome Analyser GAII (Illumina, Inc.) according to the manufacturer's instructions. After sequencing, the resultant images were analysed with the proprietary Illumina pipeline (version 1.4). This resulted in approximately 964 Mb of sequence, with 90% of clusters passing the initial filtering step. Reads for each sample were trimmed at the 3' end by 5, 10, 15 and 20 bases. Assemblies were made using Velvet 0.7 [ 21 ] with a range of hash lengths from 33 to 63 and a minimum k-mer coverage of 5×. Maximum contig lengths and N50 values were tabulated, and the best assembly for each sample was selected for further analysis. Assembled contigs were aligned to the R. praetor mitochondrial genome using Geneious 4.7 [ 22 ].
Sanger sequencing was used to sequence both ends of each long range fragment to check consistency and to complete the coverage of the mitochondrial genomes. It was also used to sequence the three mitochondrial regions of the additional samples of R. lutreolus and R. villosissimus . Sanger sequencing was carried out using the BigDye Terminator version 3 sequencing kit, the GeneAmp PCR System 9700 and a capillary ABI3730 DNA Analyser, all from Applied Biosystems. The final assembly of each mitochondrial genome was carried out in SequencherTM (Gene Codes Corporation).
Sequence Alignments
Three main datasets were assembled. The whole genome (WG) dataset (for the alignment see Additional File 1 ) contained mitochondrial DNA sequences from the genomes of 13 species and included the vole (Cricetidae) and mouse (Muridae) as outgroups (Table 1 ). The dataset includes 12 mitochondrial genome proteins, rRNA and tRNA coding sequences, totalling 14,471 nucleotides after manual alignment. The D-loop was excluded because of difficulty in alignment. The sequence for protein coding gene ND6 was also excluded because it is encoded on the mitochondrial L-strand. The sequence alignment was partitioned into the three codon positions, RNA stem regions and RNA loop regions. As in the previous work of Robins et al. [ 5 ], the third codon positions were coded as purines or pyrimidines (RY-coded) because it was demonstrated there that these positions had a higher substitution rate than the other partitions and substitution models did not sufficiently account for saturation at these sites at deeper levels in the tree. RY-coding greatly reduced saturation in 3 rd codon transitions, removed highly significant composition bias and provided better phylogenetic resolution, even within Rattus [ 5 ]. Compositional heterogeneity is a recognised problem in mitochondrial genome analysis and for a more detailed discussion of the issues see [ 23 ] and other references therein.
A second alignment was assembled from three genomic regions (3G), comprising 126 sequences, spanning 672 bp of cyt b , 702 bp of the COI and the hypervariable region of the D-loop, for a total length of 1952 bp. This dataset which includes Mus as the outgroup extends that of Robins et al [ 4 ] by the inclusion of sequences from seven new samples, three of R. lutreolus and four of R. villosissimus (for sample details see Tables 1 and 2 ). The vole outgroup was not included because of the difficulty in aligning the D-loop. This alignment was analysed in three ways: with three partitions (codon positions 1+2, codon position 3 as RY, and all other positions see Additional File 2 ), with each genomic region as a separate but unencoded partition (Additional File 3 ), and as a concatenated sequence with no partitions. The coding of sequence labels in the nexus files to species names and GenBank accessions is given in Additional File 4 .
A third alignment (3G-WG) was created (Additional File 5 ) by extracting the 3G regions from the specimens in the WG dataset. This enabled us to assess whether the differences in the results obtained using the WG and 3G datasets were due to the specimens or to the genomic regions used.
Of the gene regions that we have sequenced, cytochrome b is the most widely represented among rodent species in GenBank. To test whether the Australo-Papuan clade was monophyletic we assembled a rodent alignment of cytochrome b sequences from 245 representatives of Rattus sensu lato [ 24 ] (see Additional File 6 ). This included our data as well as additional data deposited in GenBank by Pagès et al. [ 11 ], Jansa et al. [ 9 ], and Rowe et al. [ 10 ]. The sequences from the various studies tended to be either ~700 or ~1200 bp in length. Further we wished to use both unencoded and RY-encoded data. Consequently we created four alignments. The first used the complete alignment from Additional File 6 (1-1182 bases), and the second used a short alignment (bases 1-713). In the third and fourth alignments the data from both the complete and the short alignment were partitioned into codon positions 1+2 (unencoded) and codon position 3 (coded as RY). These codon-partitioned alignments were trimmed by 40 bases at the 3' end to remove the non-coding tRNA region from the sequences leaving bases 41-1182 and 41 - 713 respectively.
Phylogenetic methods
The optimal substitution model was determined in previous work [ 5 ] to be GTR + I + G 4 . This model was used in partitioned maximum likelihood (ML) and Bayesian inference analyses. ML analyses were performed on the WG dataset with PAUP* [ 25 ] and RAxML [ 26 ]. Other exploratory ML analyses were performed using PHYML [ 27 ]. Phylogenetic trees were estimated from the WG, 3G, 3G-WG and the four rodent cytochrome b alignments under Bayesian inference with MrBayes [ 28 ]. For the WG dataset, three Markov chain Monte Carlo (MCMC) chains for each of two independent runs proceeded for 8,500,000 generations with trees being sampled every 5000 generations. The burnin length 1,500,000 was determined by examination of -ln L versus MCMC generation plots and sampling efficiency within Tracer 1.5 [ 29 ]. For the 3G and 3G-WG datasets, three MCMC chains for each of two independent runs were sampled for 10,000,000 generations with trees sampled every 1000 generations. The burnin length of 100,000 generations was chosen by the same method.
In addition to the three phylogenetic trees estimated from each of the WG, 3G and 3G-WG datasets, estimations were made using ML for alignments based on single gene regions from both the WG and 3G datasets. In all, 18 tree topologies were obtained for the species represented in the WG dataset. Support in the WG dataset among these alternative topologies was examined using the Kishino-Hasegawa (KH) [ 30 ] and approximately unbiased (AU) [ 31 ] tests within the CONSEL program [ 32 ]. Sitewise log likelihoods were combined from the GTR + I + G 4 (CF87 + I + G 4 for RY-coded data) maximum likelihood (ML) analyses run in PAUP* for each of the separately run partitions.
We assessed variation along the genome in the strength of support for the two main topologies, that with the clade comprising R. lutreolus , R. sordidus , R. tunneyi and R. villosissimus (Figure 1 ) or that with the clade comprising R. fuscipes , R. leucopus and R. praetor (Figure 2B ) as a more recent divergence. The whole genome alignment was subdivided into 1000 bp sections, and an ML bootstrap analysis performed on each non-partitioned section using PHYML and the GTR+G+I model of evolution. We estimated the frequency of each clade for each region among the bootstrap trees.
BEAST [ 33 ] was used to apply Bayesian inference to the estimation of divergence dates within the phylogenies of the WG, 3G and 3G-WG species. For the WG dataset, as in previous work [ 5 ], a partitioned analysis was performed in which codon positions 1 and 2 were combined, RNA stem and loop partitions were combined and codon position 3 was coded as RY, to counteract the effects of rate heterogeneity and substitution saturation at hypervariable sites. The MCMC chain was run for 10,000,000 generations, with both trees and ages sampled every 5,000 generations. The "relaxed" clock method was used in which rates along branches are distributed according to a lognormal distribution [ 34 ]. For the 3G and 3G-WG datasets, a partitioned analysis was performed in which codon positions 1 and 2 were combined, codon position 3 was coded as RY and all other sites formed the third partition. A birth-death model of speciation was used with the WG and 3G-WG datasets, in which each species is represented by a single sequence, whereas a coalescent model was used with the 3G dataset, in which each species has multiple specimens. The sampled trees were summarised as maximum clade credibility trees with node heights set to the median of sampled values using TreeAnnotator [ 33 ].
We applied age constraints on two nodes to calibrate our BEAST analysis of the WG dataset: the split between Muridae (mouse plus rats) and Cricetidae (vole), that is the root node, and the split between Mus and Rattus within Muridae. As in previous work [ 5 ] we used the recommendations of Benton and Donoghue [ 35 ] for the minimum and maximum age constraints of the mouse-rat divergence. They state that current research suggests that the split between Mus and Rattus is early in the evolution of Murinae but not basal in the divergence of the clade. They recommend a lower bound of 12.3 Mya since this is the oldest record of Progonomys , the genus assumed to contain the common ancestor of Mus and Rattus and an upper bound of 11 Mya which is based on records of the extinct genus Karnimata which is believed to be on the lineage leading to Rattus . These bounds were used respectively for the middle 95% of the normally distributed prior. The Mus plus Rattus clade was enforced relative to the vole which is in agreement with all recent molecular and morpological interpretations. A uniform prior from 11 to 34 Mya for Muridae (mouse and rats) versus Cricetidae (vole) provided a calibration on the root that is very conservative at both the upper and lower bounds (see [ 36 ]).
The chronogram obtained (as described above) from BEAST for the 3G dataset contained unrealistically old ages for the nodes in the tree and a topology that differed markedly from that estimated for the WG dataset. Consequently, to test the effect of substitutional saturation, a further BEAST analysis was performed using a coalescent prior and constant population size. Here, the Mus outgroup was removed because of the long branch lengths between Mus and Rattus and potential saturation in the hypervariable region of the D-loop. The Asian and Australo-Papuan clades were each constrained to be monophyletic and the age of the root, the age of the most recent common ancestor (tMRCA) for Rattus , was assigned a lognormal prior based on the empirical results from the WG analysis. | Results
Topology
Three datasets were used for phylogenetic reconstructions. These differed in their extent of taxon sampling and in their sequence length (whole mitochondrial genomes (WG): 16 taxa 16293-16309 characters; three genomic regions (3G): 126 taxa 1952 characters; whole mitochondrial genomes trimmed to three regions (3G-WG): 16 taxa 1952 characters). As described below, similar but different topologies were inferred for relationships among the Asian clade when the larger taxon dataset was used. Differences in the relationships among Australian species were also inferred when New Guinean sequences were included in the analysis of Australian species. The greatest uncertainty in relationships among the Australian species was due to different root placements being optimal for the Australo-Papuan clade in some analyses of the different datasets. As we discuss, these observations are consistent with expectations for rapid speciation [ 37 , 38 ].
The whole genome (WG) dataset
The trees estimated by ML and Bayesian inference from the WG dataset have the same topology (Figure 1 ) with the minor exception of differences in the branching order among the R. norvegicus sequences. The topological relationships within the clade of Asian species ( R. exulans , R. norvegicus , R. rattus and R. tanezumi ) are the same as previously recovered [ 5 ]. The Australo-Papuan species ( R. fuscipes , R. leucopus , R. lutreolus , R. praetor , R. sordidus , R. tunneyi and R. villosissimus ) form a sister clade to the Asian clade. Overall eighteen hypotheses regarding the relationships among the Australo-Papuan clade of Rattus were tested using the CONSEL program. The log-likelihood differences between the optimal topology and the other hypotheses of relationship for the WG data are given in Table 4 . The relationships represented in Figure 1 had the highest likelihood, but four other topologies were clearly not significantly poorer and a fifth was only marginally not significantly poorer. Common features of the five best trees include a) the early divergence of R. praetor and R. leucopus from the other lineages, and b) the late divergence of R. sordidus and R. villosissimus . Trees in which R. praetor diverge late (e.g. 12, 16, 18 ) are much poorer than the best trees.
The Australo-Papuan clade has strong support from both ML and Bayesian inference (Figure 1 ). Within the Australo-Papuan clade, the analyses of the WG dataset indicate support for R. sordidus and R. villosissimus as sister taxa, with the inclusion of R. lutreolus , R. tunneyi and then R. fuscipes as one moves back in time. The order of divergence of R. praetor and R. leucopus is somewhat uncertain; the time interval between the dates of divergence is small and ML bootstrap support is weak.
Three mitochondrial gene (3G) region (Cyt b, COI, D-Loop) datasets
Phylogenetic analysis of the 3G dataset obtained from ML and MrBayes analyses, when the codon positions are partitioned and RY encoding used, returned the same topology (Figure 2B ) corresponding to topology 12 in Table 4 where, as a member of the PNG I clade, R. praetor diverges recently and R. lutreolus , R. sordidus , R. tunneyi and R. villosissimus have more basal positions. This topology is not supported by the WG dataset (see Figure 1 ).
A slightly different topology again was found when the 3G dataset was analysed with MrBayes without RY encoding, either with the alignment partitioned by region or simply concatenated. Here, the order of divergence of R. leucopus and R. praetor is reversed, and R. lutreolus and R. tunneyi form a clade. This corresponds to topology 4 of Table 4 .
The major differences in tree topology inferred from the 3G dataset, as compared with the WG dataset, may derive from the impact of the many additional sequences in the 3G dataset or from the reduction in the number of genomic regions and partitions, and the inclusion of the hypervariable D-loop. To test this possibility, the 3G dataset was pruned of all sequences not in the WG dataset. For the tree returned by a BEAST analysis of this 3G-WG dataset (Figure 3A ; topology 16 in Table 4 ), the unrooted topology of the Australo-Papuan clade is highly similar to that shown in Figure 1 . It differs only in the position of R. praetor , however, the root of the Australo-Papuan clade is placed differently. Both ML and MrBayes estimations of the phylogenetic tree produced a topology (Figure 3B ) with notable differences in the positions of the root of the Australo-Papuan clade and of R. tunneyi (topology 18 in Table 4 ).
The frequency along the genome for the two topologies, that with the clade comprising R. lutreolus , R. sordidus , R. tunneyi and R. villosissimus (Figure 1 ) or that with the clade comprising R. fuscipes , R. leucopus and R. praetor (Figure 2B ) as a more recent divergence is summarized in Figure 4 . The support for these clades is highly variable along the genomes. The fragment containing the D-loop provided the strongest support for the ( R. fuscipes , R. leucopus and R. praetor ) clade. Further, it contained strong support for the ( R. lutreolus , R. sordidus , R. tunneyi and R. villosissimus ) clade. The sections containing the COI and Cyt b genomic regions provided only weak support for either of these clades.
Another approach to testing the effects of the gene regions used was to remove the outgroup and constrain the height of the tree. This will reduce the possibility of misplacing the root of the Australo-Papuan clade as a result of long branch attraction relative to the outgroup and stochastic artefacts. It will also diminish the influence of hypervariable regions of the sequence. The topology returned from this analysis (Figure 2A ) was very similar to, and not significantly different from, that obtained in the WG analysis (Figure 1 ), corresponding to topology 2, or marginally topology 1, in Table 4 .
In summary, both the 3G and 3G-WG datasets gave topologies significantly different from that obtained using the WG dataset when Bayesian inference (both BEAST and MrBayes) or maximum likelihood was used to analyse partitioned and RY-coded datasets. Reducing the effects of the outgroup or the hypervariable regions enabled us to recover the WG topology. Consequently we can infer that it was the use of only these three limited genomic regions, rather than the inclusion of many additional sequences, that led to the inference of a different topology from the 3G dataset than from those of the pruned 3G-WG and the WG datasets. Nevertheless our analyses of the 3G dataset provided generally well supported species clades which are useful for species identification.
The four cytochrome b alignments returned very similar results. We provide the tree with species clades collapsed for greater legibility in Additional File 7 and as a full tree file in Additional File 8 . In each analysis the order of divergence was Maxomys deepest followed by Micromys , Niviventer and Leopoldamys . The Australo-Papuan rats always formed a well-supported monophyletic clade. The clades representing Bandicota , Berylmys , Tarsomys , Limnomys , R. norvegicus and its sister species R. nitidus , and the remaining clade of Asian rats were each observed consistently with high support. However, the relationships among them were unstable and were not resolved with high posterior probability or bootstrap support (results not shown).
Dates of Divergence
The addition of seven genomes has made more recent the estimated MRCAs within the Rattus genus when compared with previous work [ 5 ]. The median age of the split between the Asian and Australo-Papuan clades is estimated to be 2.7 Mya (95% credible interval 1.7 - 3.9 Mya). Our previous estimate of this date, based on eight Asian rat genomes but only a single genome from the Australo-Papuan clade ( R. praetor ), was near the upper bound at 3.8 Mya. Similarly, the previous estimate of tMRCA of the Asian clade, at 3.1 Mya, falls near the upper bound of the estimate obtained here with median 2.3 and 95% credible range 1.4 - 3.3 Mya. The MRCA for the Australo-Papuan clade is 1.66 Mya, and the most recent divergence within that clade, between R. sordidus and R. villosissimus , is estimated to be approximately 500,000 ya (Table 5 ).
When the dates of divergence were estimated for the 3G dataset, nodes in the tree were more than twice as old as when estimated using the WG dataset. However, when the MRCA for Rattus was constrained, the ages of the divergences in the Australo-Papuan clade (Figure 2A ) more closely resembled those obtained from the WG dataset.
The estimates of the node ages obtained using the 3G-WG dataset (Figure 3 ) are comparable but slightly older than those obtained from the whole genome. The Asian - Australo-Papuan split is at 3.2 Mya as compared with 2.7 Mya, and the Australo-Papuan clade has a MRCA of 1.87 Mya as compared with 1.66 Mya from the whole genome. In contrast, the deeper splits within the Asian clade have nearly the same estimated age. The divergence of R. norvegicus is estimated to be 2.28 Mya, as compared with 2.25 Mya, and the divergence of R. exulans from the other Rattus at 1.81 versus 1.77 Mya when the whole genome was used. | Discussion
Whole mitochondrial genomes have been used to infer phylogenetic relationships in a wide range of organisms e.g., mammals [ 39 - 43 ], birds [ 44 ], fish [ 45 ] and worms [ 46 ]. Phylogenies based on mitochondrial genomes are effectively gene trees and so may not recover the true species tree [ 47 ]. While it may be useful also to develop phylogenies based on nuclear markers, there are several reasons why the use of whole mitochondrial genomes provide good indicators of species history. The mitochondrial genome, being haploid and lacking recombination, effectively has four-fold lower coalescence times relative to diploid nuclear genes. This gives much greater phylogenetic resolution than slower evolving nuclear genes. Hence mitochondrial genomes are expected to be good indicators of species history.
Our phylogeny estimated from the WG dataset (Figure 1 ) is well resolved except for the relative positions of R. leucopus and R. praetor . The top six observed tree topologies (Table 4 ) have one or the other of these two species diverging earliest in the Australo-Papuan clade. Both species diverge earlier than any of the Australian native rats and there is marginally more support for R. praetor diverging earliest. This phylogeny clarifies the relationships among the rats of the Rattus fuscipes species group and is consistent with the taxonomy of Musser and Carleton [ 1 ]. The occurrence of R. villosissimus and R. sordidus as sister species in a shallow divergence is concordant with the phenetic results of Taylor et al. [ 13 ], and the phylogenies of Baverstock [ 14 ] and Aplin [ 17 ]. The more recent divergence of R. lutreolus compared with R. tunneyi differs from these earlier phylogenies where R. lutreolus is either basal to R. tunneyi [ 13 , 17 ] or in an unresolved polytomy that includes R. fuscipes and R. leucopus [ 14 ]. In our phylogeny the position of R. lutreolus within the Australian native rat clade supports Aplin's [ 17 ] view that this rat is more closely related to Australian than to New Guinean rats. The position of R. fuscipes basal in our Australian native rat clade but a more recent divergence than that of R. leucopus differs from the results of Taylor et al. [ 13 ] whose phenogram shows R. fuscipes within a clade of New Guinean rats and diverging earlier than the New Guinean species R. leucopus .
Our choice of samples for whole mitochondrial genome analysis was influenced by our confidence in the correct assignment of species to those samples. Where there was congruence between the named species and the phylogenetic clades we were confident in the identifications. In Robins et al. [ 4 ] we discussed the difficulties that arose when OTUs contained heterogeneous collections of named species. These names were assigned by either the museum or the collector. The source of such heterogeneity could be misidentification, faulty taxonomy due to the plethora of synonyms used for Rattus species, hybridisation, incomplete lineage sorting or some combination of the above. Pagès et al. [ 11 ], although they explicitly avoided rats from the Australo-Papuan clade, reported a similar problem. They identified a heterogeneous clade (R3) that was similar to, and where they incorporated data from Robins et al. [ 4 ], essentially the same as the diardii clade (Figure 2 ) therein. In this paper we again resolved the diardii clade. When Pagès et al. [ 11 ] considered the same species as we, their genetic based species identifications are largely congruent with ours. This finding supports the usefulness of shorter mitochondrial sequences for species identification. Of the 14 Rattus specimens we used for the whole mitochondrial genomes, 11 were from homogeneous clades. The exceptions were the exulans clade which from a total of 21 samples contained two almost certainly misidentified samples and two New Guinean clades PNG I and verecundus . Our New Guinean species clades are all problematic and to resolve the conflict DNA is needed from specimens that have accurate morphological identifications. All six of the specimens of R. praetor that we have processed fell into the PNG I clade and we used one of these samples to represent the clade. At this time we have insufficient information to account for the heterogeneity of this clade although some possibilities are discussed in Robins et al. [ 4 ]. The Australian Rattus specimens on the other hand all fell into well supported homogeneous clades thus the phylogenetic assignment of species was in agreement with that of the museum or collector.
Our cytochrome b analyses generally gave robust species clades. Further, the Australo-Papuan rats were in a well-supported and monophyletic clade in all analyses. However, the relationships within Rattus sensu stricto [ 24 ], especially among Bandicota , Berylmys , the Philippine endemics ( Tarsomys , Limnomys and Rattus everetti ) and the main Asian and Australo-Papuan clades of Rattus , were poorly resolved. On the other hand, the deeper relationships within Rattus sensu lato were well-supported. Pagès et al. [ 11 ] stated that they did not obtain a robust phylogeny when they analysed cytochrome b alone. Interestingly, even when they analysed three concatenated genes (one nuclear and two mitochondrial) they were unable to place Bandicota conclusively within or sister to Rattus . Thus the currently available data is insufficient to address the question of monophyly within Rattus as a whole.
The ancestral node in the Australo-Papuan clade differed with the genomic regions used. The inclusion of the hypervariable D-loop (Figures 2B and 3 ) moved the root deeper into the clade, relative to its placement when the whole genome was used (Figure 1 ) or when the MRCA of Rattus was constrained (Figure 2A ). It is likely that the differences arise from substitution rate differences between D-loop and the rest of the genome. The hypervariable nature of the D-loop, with the potential for substitution saturation, also contributed to the inflation (two- to three-fold increase) of dates of divergence in the unconstrained BEAST tree based on the 3G dataset compared with that of the WG dataset.
Although the deeper branching orders are uncertain within the Australo-Papuan clade for the 3G and the 3G-WG phylogenies (Figures 2 and 3 ) the Australian native Rattus species, including the newly added R. lutreolus and R. villosissimus samples fall into well differentiated clades reinforcing the usefulness of shorter mitochondrial DNA sequences for species identification [ 48 ]. The position of two samples from New Guinea, misidentified as R. rattus and labelled as OTU rattus II in Figures 2A and 2B , as sister to R. villosissimus which is in turn sister to R. sordidus suggests that these samples are in fact R. villosissimus . The usefulness of these shorter sequences, however, to elucidate deeper evolutionary relationships or to date divergences is expected to be limited given the difficulties in inferring species relationships for rapid radiations [ 37 , 38 ]. The deeper branches of the Australo-Papuan clade in Figures 2 and 3 are all quite short, indicative of a period of rapid diversification, but the branching pattern varies with the methods used and no one pattern is well supported. It is possible that the inclusion of one or more nuclear markers could help resolve this pattern and is an approach to be pursued in the future, however, given their greater effective population size and longer coalescence times, nuclear genes are likely to be less powerful in resolving the order of divergences. They will, however, provide evidence regarding whether mitochondrial gene trees provide accurate estimates of the deeper species relationships [e.g., [ 49 ]].
The dates inferred from the WG dataset analysis are congruent with earlier studies. Jansa et al. [ 9 ], using just the IRBP nuclear gene, estimated a date of 3-4 Mya for the origin of a clade containing R. exulans , R. tanezumi and, R. preator (sic). This is the equivalent of our 2.7 Mya split between the Asian and the Australo-Papuan clade (See Figure 1 ). When Jansa et al. [ 9 ] included both IRBP and cyt b data in their analysis the age of the nodes increased three fold and when cyt b alone was used the ages increased four fold. We had a similar problem with the 3G dataset which when analysed in BEAST, without the constraints described in the methods, returned dates that were almost three fold older than those returned by the WG dataset. It is worth noting also that we observed slightly older dates ~4.6 Mya (unpublished data) for the Asian/Australo-Papuan split on standard (ACGT) coding, compared with ~3.51 Mya obtained when we RY coded our 2008 [ 5 ] whole mitochondrial dataset. Rowe et al. [ 10 ] included three Rattus sp. in their analyses of the old endemic murines of the Sahul. They recovered evidence for the Asian/Australo-Papuan split with R. leucopus and R. villosissimus forming a sister pair relative to R. norvegicus . Although they did not report a chronogram, we can obtain a rough estimate of the age of the MRCA for Rattus from their Figure 4 . The depth of the MRCA of Rattus was 26% of the depth of the closest dated node (B) in its ancestry (their Table 1 ), giving an estimated age of 2.52 (2.26 - 2.81) Mya. This date is concordant with our estimate of 2.7 Mya.
Our dates are also consistent with the earlier study of Robins et al. [ 5 ]. Whilst the node dates obtained from our WG BEAST analysis are all slightly younger than the equivalent ones from the previous study they are within the 95% credible intervals of the 2008 study. Since we have increased the sample size of the Australo-Papuan clade representation from one to eight the dates we report here are more reliable. These younger dates align more closely with those from the L1 (LINE-1, long interspersed repeated) retrotransposable elements study of Verneau et al. [ 24 ] who estimated the timing of several major speciation events within Rattus sensu stricto of which the two most recent are relevant to our study. The first at ~2.7 Mya gave rise to five lineages one of which led to R. fuscipes (the only representative of the Australian and New Guinean rats in their study) and another led to Asian Rattus species. This divergence is the equivalent of the split between our Asian and Australo-Papuan clades also dated at ~2.7 Mya.
The five lineage speciation events described by Verneau et al. [ 24 ] and the origin of the Australo-Papuan and Asian clades all estimated at ~2.7 Mya coincided with Pleistocene Ice Age events. The onset of the Pleistocene glaciation cycles was at about 3 Mya [ 50 ] and by 2.6 Mya the Pleistocene Ice Age was well established [ 51 ]. Major changes occurred in climate and available land in the region encompassing Island Southeast Asia, New Guinea and Australia. Such changes, for example at the highest glacial maxima New Guinea, Australia, and the islands of Misool, Waigeo, the Aru archipelago and Tasmania formed a single land mass[ 18 ], would have provided routes for dispersal and subsequently opportunities for speciation in Rattus . Also during the same time period, there was considerable dispersal of marsupials (bandicoots, dasyurids, phalangerid possums, and kangaroos) between New Guinea and Australia but diversification was more limited [ 52 ]. We can speculate that there was greater niche overlap between the newly migrated and endemic marsupials while Rattus was not so ecologically constrained. Glacial cycling with some 20 cycles in the last two million years continued to contribute to ongoing climate and sea level changes [ 17 , 50 ] and during this time radiation also continued in Rattus .
The more recent speciation events described by Verneau et al. [ 24 ] were a radiation in the Asian Rattus clade. Verneau et al's [ 24 ] study included 12 Asian Rattus species and four of these are represented in our study; R. norvegicus, R. exulans R. rattus and R. tanezumi . Their divergence estimates were in the time period 1.2 Mya to about 0.5 Mya. Our results are reasonably concordant with but probably more reliable than theirs because their analysis assumed a fixed rate molecular clock. In their L1 study the timing of the divergence of the R. norvegicus lineage at ~1.8 Mya compares with ~2.3 Mya in our study; R. exulans at ~1 Mya compares with our ~1.7 Mya and the divergence between R. rattus and R. tanezumi at 0.5 Mya compares with ours at ~0.3 Mya. The divergences in our Australo-Papuan clade occur over a similar time period (1.7 Mya - 0.5 Mya) as that in the Asian clade. R. praetor diverged at ~1.7 Mya, R. leucopus at ~1.6 Mya, R. fuscipes at ~1.4 Mya, R. tunneyi at ~1 Mya, R. lutreolus at ~0.9 Mya and the sister species R. sordidus and R. villosissimus at ~0.5 Mya. Despite the uncertainty in assigning the order of the deepest divergences in our Australo-Papuan clade, which is exacerbated by the relatively short time frame of ~0.2 My in which they occurred, the overall pattern from the WG and the 3G datasets suggests that the founding Rattus lineages reached New Guinea by at least 1.7 Mya and Australia by 1.4 Mya with further diversification occurring in both areas between ~1 - 0.5 Mya.
When sea levels were as little as 10 m below present land bridges would have connected New Guinea and Australia at the Torres Strait [ 53 ]. These land bridges which have occurred intermittently have also provided opportunities in more recent times for the reinvasion and subsequent isolation of rats between New Guinea and Australia in the Torres Strait area. In our phylogeny R. sordidus and R. villosissimus diverged most recently in the Australian Rattus clade. This position is consistent with an Australian origin of R. sordidus . In contrast the R. leucopus lineage diverged earlier than any of the native Australian rats thus supporting a New Guinean origin. The recent split between the Australian and New Guinean sister lineages at ~0.6 Mya suggests an invasion of Australia by R. leucopus followed by isolation. Our findings, therefore, support the Torres Strait faunal interchange hypothesis first put forward by Tate [ 15 , 16 ] with later support from Taylor and Horner [ 12 ], Taylor et al. [ 13 ] and Aplin [ 17 ]. | Conclusions
Although Rattus is a problematic genus and further investigations of the Australo-Papuan group are needed to resolve the relationships of the Rattus species of New Guinea we have clarified the pattern and timing of divergences among the Australian rats. Our whole mitochondrial genome analyses are concordant with Musser and Carleton's [ 1 ] Rattus fuscipes species group and resolve the positions of R. fuscipes and R. lutreolus within it. Our findings suggest that Rattus invaded Australia from New Guinea about 1.4 Mya and that this was followed by a period of rapid speciation. Our results also support and suggest a date of ~0.5 Mya for the hypothesized [ 15 , 16 ] more recent invasions across Torres Strait of R. sordidus into New Guinea and R. leucopus into Australia. | Background
The genus Rattus is highly speciose and has a complex taxonomy that is not fully resolved. As shown previously there are two major groups within the genus, an Asian and an Australo-Papuan group. This study focuses on the Australo-Papuan group and particularly on the Australian rats. There are uncertainties regarding the number of species within the group and the relationships among them. We analysed 16 mitochondrial genomes, including seven novel genomes from six species, to help elucidate the evolutionary history of the Australian rats. We also demonstrate, from a larger dataset, the usefulness of short regions of the mitochondrial genome in identifying these rats at the species level.
Results
Analyses of 16 mitochondrial genomes representing species sampled from Australo-Papuan and Asian clades of Rattus indicate divergence of these two groups ~2.7 million years ago (Mya). Subsequent diversification of at least 4 lineages within the Australo-Papuan clade was rapid and occurred over the period from ~ 0.9-1.7 Mya, a finding that explains the difficulty in resolving some relationships within this clade. Phylogenetic analyses of our 126 taxon, but shorter sequence (1952 nucleotides long), Rattus database generally give well supported species clades.
Conclusions
Our whole mitochondrial genome analyses are concordant with a taxonomic division that places the native Australian rats into the Rattus fuscipes species group. We suggest the following order of divergence of the Australian species. R. fuscipes is the oldest lineage among the Australian rats and is not part of a New Guinean radiation. R. lutreolus is also within this Australian clade and shallower than R. tunneyi while the R. sordidus group is the shallowest lineage in the clade. The divergences within the R. sordidus and R. leucopus lineages occurring about half a million years ago support the hypotheses of more recent interchanges of rats between Australia and New Guinea. While problematic for inference of deeper divergences, we report that the analysis of shorter mitochondrial sequences is very useful for species identification in rats. | Authors' contributions
JR and PM designed the project, planned and conducted lab work. All authors analysed data and contributed to the writing and editing of the manuscript. All authors approved the final manuscript.
Supplementary Material | Acknowledgements
We thank Steve Donnellan and the South Australian Museum for tissue samples. We thank David Penny and Peter Lockhart for discussions and support throughout the project and for advice on the manuscript. We thank two anonymous reviewers for helpful comments. This research was funded by the Allan Wilson Centre for Molecular Ecology and Evolution and the Marsden Fund of New Zealand. | CC BY | no | 2022-01-12 15:21:37 | BMC Evol Biol. 2010 Dec 2; 10:375 | oa_package/21/17/PMC3014932.tar.gz |
PMC3014933 | 21126360 | Background
Discernible transposed elements (TEs) occupy about half of the human genome [ 1 ]. They integrate into host DNA in waves of activity. In the face of increasing density, they frequently insert into each other. Nested insertions encrypt valuable historical information about the relative age of the elements, comparable to fossils in distinct layers of earth. As old fossils are absent in young layers, older inactive TEs are not inserted into younger elements. In contrast, young TEs are able to occupy all strata of older elements as well as those active at the same time. Hence TEs active at different historical periods display characteristic insertion profiles. Comprised as they are of a substantial fraction of TEs, mammalian genomes are ideally suited for such analyses. Moreover, even low genomic accumulations (e.g., about 3% genomic coverage of CR1 elements in chicken; [ 2 ]), are sufficient for distinct profiles of retroposon activity [ 3 ].
Over more than one hundred and sixty million years, mammals have accumulated elements from four major classes of transposons, L ong IN terspersed E lements (LINEs), S hort IN terspersed E lements (SINEs), retrovirus-like L ong T erminal R epeats (LTRs), and DNA transposons [ 4 ]. While members of the last group move via a cut and paste mechanism, the other three elements transpose by a copy and paste mechanism via an RNA intermediate reverse transcribed into cDNA. In humans such RNA transposons represent more than 90% of all transposed elements [ 1 ]. Active LINE and LTR elements encode the enzymatic machinery that is necessary for their own propagation, and in the case of LINEs also the co-propagation of SINEs or any other RNA. For LINE1-mediated retroposition, there is a slight preference for A-rich integration sites known as kinkable sites [ 5 ]. Such regions contain a TTAAAA consensus motif and are frequently found in the junction of dimeric retroposons such as Alu elements in primates. Alu elements are primate-specific, 7SL RNA-derived SINEs that arose from Fossil Left and Right Alu monomers [ 6 ].
A retrospective, sequence-based insight into deep evolutionary periods is feasible via inferences from sequence divergence, but is accompanied by uncertainties due to changing regional and temporal substitution rates, mutation saturation, and the occurrence of highly mutated CpG sites. Especially older, highly diverged, and short elements lead to unreliable estimations. Counting and comparing nested insertions, however, is less sensitive to such considerations.
There are currently two different approaches for calculating the relative activity periods of subtypes of transposed elements, both of which draw on RepeatMasker annotations. The T ransposon C luster F inder (TCF) estimates how often certain elements have been fragmented by the insertions of other elements over evolutionary time [ 7 ]. A compilation of representative subsets of interacting transposed elements is then presented in an adjacency matrix displaying frequencies of interruptions optimized for their potential chronological order. This I nterruptive M atrix A nalysis (IMA) starts from a random chronological order of elements and systematically repositions them so as to minimize the number of nonzero entries in the part of the matrix defined by the artificial transposition of old elements into new ones.
At about the same time as the TCF application was developed, we developed the T ransposition in T ransposition (TinT) algorithm [ 3 ], which also uses RepeatMasker coordinates to compile interrupted and nested retroposons. The frequencies of fragmented versus nested elements are counted, assembled in a data matrix, and sorted by pre-selected retroposon types. This matrix applies a specific probabilistic likelihood model (Additional file 1 ) to calculate the relative integration period for each retroposon subtype in relation to all other subtypes.
Due to the high frequency and multiple interactions of different elements, both the TCF and the TinT methods exhibit high intrinsic complexities and are neither easy nor self-explanatorily applicable for the scientific community. To compensate for these shortcomings, we have now developed an easy to use, web-based interface for the TinT application. TinTs can be directly screened for in model organisms or in any allocated RepeatMasker report data. To demonstrate and test the web-based TinT method, we investigated the representative primate genomes of Homo sapiens , Macaca mulatta (rhesus), Callithrix jacchus (New World marmoset), Tarsius syrichta (Tarsius), and Microcebus murinus (grey mouse lemur) and their well-characterized, primate-specific Alu dimeric elements. Because of the well-known evolutionary histories of both the species and their retroposons [ 8 , 9 ], primates represent an ideal test group for the TinT application. | Methods
Required RepeatMasker Input Data
The RepeatMasker source file can be a critical source of errors due to miss-annotations of elements or their fragments. To overcome this potential problem, we developed a quality check of the RepeatMasker TinT coordinates and automatically selected only unambiguously nested insertions for our analyses. The stringent selection works well for genome data and frequently occurring elements, and provides a reliable TinT pattern. For the analysis of lesser quantities of data or genomes with low copy number elements, we have provided the option of applying less stringent parameters (relaxed conditions; Figure 1 ).
However, for the human genome, the minimal amount of data that is necessary under stringent conditions to retain a full TinT resolution is about 10% of the genome, for instance about 300,000 traces are sufficient to receive the representative full TinT pattern. This means, the pattern is stable and reproducible after adding additional portions of data. It is noteworthy that this calculation varies from species to species and depends on the frequency of available elements. The precision of the TinT approach increases with the amount and quality of the input data. To derive the most reliable TinT pattern, all available sequences of selected species should be downloaded from genome ( ftp://ftp.ncbi.nih.gov/genomes/ or trace databases ftp://ftp.ncbi.nih.gov/pub/TraceDB/ ). The most time-consuming step of the TinT analysis is the upstream RepeatMasker screening. Depending on the amount of genomic data, the size of the RepeatMasker library used, and the available computational power, this process might run for several days. To reduce this screening time, it is advisable to restrict the RepeatMasker library to specific element groups (e.g., SINEs or LINEs). The local RepeatMasker library can be assembled with specific elements or element groups. Similar retroposon types should always be included in one run to avoid artificial annotation of the masked repeats. The report file can be directly applied to downstream processes.
Primate Test Sets of Data
Genomes of Homo sapiens (hg19), Macaca mulatta (rheMac2), and Callithrix jacchus (calJac1) were downloaded from the UCSC Genome Bioinformatics site; http://hgdownload.cse.ucsc.edu/downloads.html ; Tarsius syrichta (Tarsyr1.0) and Microcebus murinus (MicMur1.0) genomes were downloaded from the Broad Institute; http://www.broadinstitute.org/ftp/pub/assemblies/mammals . | Results
Principle of the TinT
The first step in generating a TinT profile is to detect nested retroposons. The local version of RepeatMasker http://www.repeatmasker.org/RMDownload.html produces report files containing all necessary information about element types and coordinates of nested and interrupted elements (Figure 1 ). We considered an element to be unambiguously nested if (1) it is located at the same genomic region as the interrupted element parts, (2) its element index is higher than the identical indices of the interrupted element parts, (3) the starting and end-coordinates of all elements span ≥20 nt each (minimal query length), (4) the interrupted host sequences show the same orientation, and (5) the separated parts of the host element's consensus sequence are preferably ≥50 nt (minimal repeat extension), but at least ≥18 nt (minimal repeat extension overlay) and include an overlap of ≤35 nt (maximal repeat overlay; overlapping host sequence regions are the result of target site duplications or low complexity regions). In cases where the separated host parts have been incorrectly assigned to different subfamilies (as evidenced by detailed retroposon inspection), we adopted the name of the largest part. Single elements (that did not insert or were not fragmented by other elements) were excluded from analyses. Nested integrations of identical elements were used only to tune the parameters of the model. All parameters shown in Figure 1 were optimized by empirical data and can be changed individually. To relax the conditions, the element indexes can be ignored (see Figure 1 ; element index and Additional file 2 : item 4). With this setting, the TinT application only considers whether the interrupted host fragments refer to the same class of elements. Furthermore, the stringency can be altered if the minimal query length, the minimal repeat extension, and minimal repeat extension overlay and/or the maximal repeat overlay is changed. Relaxed conditions are only recommended if the amount of data is reduced or rare elements are involved. If elements are considered that integrate without recognizable target site duplications, such as CR1 elements in birds, the maximal repeat overlay parameter can be reduced and the minimal repeat extension overlay proportionally increased.
After identifying nested retroposons, they were counted, sorted by element subtypes, and compiled in a data matrix. Finally, we developed a symmetric probabilistic likelihood model based on a normal distribution of element activity that transforms the information of the TinT matrix into a pattern of chronological integration periods indicating the probability of activity for each analyzed element type. The underlying mathematical model considers a simple scenario with only one period of activity for each element type and similar probabilities of insertions based on the following assumptions:
1. Elements of type i inserted at time points ( k = 1, 2, ..., n i ; where n i is the number of all elements of type i ).
2. In each of such points in time, inserted element of type i may fragment some elements of type j with a certain probability (including the case: j = i ).
3. Considering an identical probability of insertion into any preexisting element, denoted by α, then probability can be represented as , where η j ( t ) is the number of elements of type j preexisting at time point t .
4. Function η j ( t ) is approximated using the normal (Gaussian) distribution with mean t j and standard deviation n j , e.g., , where σ j = n j , and its derivative has the maximum at time point t = t j . All details of the model are presented as Additional file 1 .
Web-based version of TinT
The web-based version of TinT is located at http://www.bioinformatics.uni-muenster.de/tools/tint and requires Java version 1.5. The application is written as a java applet and was developed using the multi-language software development environment Eclipse, which is an integrated development environment with a repository system (CVS) in the background that keeps software changes disposable. There are two input options. First, any RepeatMasker report file can be uploaded and variable subsets and combinations of elements can be selected for TinT calculations. Furthermore, pre-analyzed model organisms and specific elements can be selected for a TinT analysis. Currently, 19 pre-computed genomes are available for the TinT analysis and the data (RepeatMasker output files) can be downloaded from http://www.bioinformatics.uni-muenster.de/tools/tint/download/RepeatMasker/.DIR . The TinT activity pattern is then graphically displayed. It should be mentioned that the application is executed locally on the computer where it is accessed. Optional parameters for reading RepeatMasker data may be entered into a special dialog box. Transpositions can be grouped and this information can be loaded from a file to provide flexibility for further analysis. Printing or exporting the generated graphs is a basic part of the software, so the results can be used in other applications. Exporting depends on the standard printer dialog of the computer system - if available the print is directed to a postscript file.
An example data set from primates
After selectively screening the human genome with RepeatMasker, we detected 1,004,931 dimeric Alu elements, 2,268 of which were considered to be unambiguous nested insertions. The frequencies of insertions extracted from the retroposon matrix (Figure 2A ) were used to calculate their activity probabilities (Additional file 2 ). Because of the multidimensional insertion pattern, the probable relative activity of each given element subtype is directly interrelated to those of the other subtypes. Alu Jo appears as the first active Alu dimer, followed by Alu Jb. Alu Sx shows the most expanded activity with the 75% interval of probable activity overlapping that of both the Alu J elements and the other Alu S subfamilies. The Alu Y elements are clearly separated from the older elements and contain those that are currently still active representatives of Alu dimers (Additional file 2 ).
From the rhesus monkey ( Macaca mulatta ) genome sequences we detected 950,960 dimeric elements including 2,008 nested, TinTs. The pattern of these Alu dimers (Figure 2B ) is similar to that of human; but includes, in addition, rhesus monkey-specific, Alu Y-related Alu R elements [ 10 ]. Because the resolution of individual Alu YR elements was too low, the related Alu YRb-d elements (TinT-option: merging elements) were combined.
The 2,132 nested elements from the 1,057,994 Alu elements detected in the New World marmoset (Figure 2C ) also show a distribution comparable to those of human and rhesus monkey. In addition, there are three New World monkey-specific Alu Ta elements with the most recent activities [ 11 ].
The 995 nested elements from the 917,008 dimeric Alu elements detected in the Tarsius genomic sequences (Figure 2D ) comprise only the two known Alu J elements. The TinT analysis showed that Alu Jo was older than the Alu Jb. Tarsius-specific Alu elements have not yet been detected.
The RepeatMasker screen of the gray mouse lemur ( Microcebus murinus ) sequences (Figure 2E ) revealed a total of 438,443 Alu elements, of which 1,905 were unambiguously nested. The mouse lemur lacks the Alu Jb elements that are present in all other primate groups. In addition to other potential but as yet uncharacterized Alu dimers, there are several recently described, lemur-specific Alu elements with very dominant and recent distributions ( Alu L, Alu La, and Alu Mim; Repbase; [ 12 ]).
Comparison to the TCF defragmentation pattern
Giordano et al. [ 7 ] presented their fragmentation analysis based on a Transposon Cluster Finder (TCF) software package. The transposon defragmentation analysis included most known mammalian TE classes and families but only the three main Alu groups Alu J, Alu S, and Alu Y. Therefore a direct comparison to our TinT activity pattern of Alu elements is limited. Furthermore, the TCF software is not freely available to derive a comparable set of data. The TCF pattern for Alu elements roughly confirms the TinT-derived succession of these elements, but indicates an artificial activity overlay of Alu J with Alu S and Alu Y elements. | Discussion
In light of the many ongoing genome sequence projects, the TinT method should prove to be quite valuable for characterizing the retroposon-influenced architecture and evolutionary history of genomes and provides a basic aid in conducting efficient retroposon-based phylogenetic reconstructions. To test and demonstrate the advanced efficiency of the TinT algorithm and to present a user-friendly web-based application, we performed a comparative analysis of nested primate specific dimeric Alu SINEs, a group of elements with an established evolutionary history [ 8 , 9 ]. Using standard consensus sequences of Alu repeats [ 13 ] to screen all available genomic sources of primates, represented by human and macaque (both Catarrhini), marmoset (Platyrrhini), Philippine Tarsius (Tarsiiformes), and gray mouse lemur (Strepsirrhini), we extracted and analyzed more than 9,300 nested from 4.5 million detected Alu SINEs. The relative activity periods of Alu elements revealed by the TinT analyses coincide with our current knowledge of these elements in primates [ 14 ].
It should be mentioned that a substantial proportion of the nested elements are ancestral insertion events and consequently are shared among different primate groups. Such common TinTs lead to similar activity patterns of species, especially for older elements (see also for example [ 15 ], Figure 1 ).
Nine diagnostic mutations distinguish Alu Jo from Alu Jb [ 14 ]. The TinT profiles support the activity of Alu Jo having preceded that of Alu Jb and, with the probable absence of Alu Jb in strepsirrhines, indicate an origin of latter elements in the common ancestor of Tarsius and higher primates. The phylogenetic affiliation of these two groups in a clade Haplorrhini was previously significantly supported by four orthologous insertions of retroposed elements [ 16 , 17 ]. This relationship is now overwhelmingly supported with quantitative and chronological evidence from 414,037 Alu Jb elements specific for Haplorrhini that are clearly absent in strepsirrhines. Beside some few specific elements humans and macaques have nearly identical profiles of Alu SINE activity. Similar activity profiles for older Alu SINEs ( Alu Jo, Alu Jb and Alu Sx) were also detected in New World monkey (marmoset). In contrast, the overlapping activity patterns of the younger Alu S and Alu Y SINEs vary among primate groups. The TinT patterns of element activities (Figure 2 ) fit well to the sequence-based reconstruction of the evolution of Alu elements (Additional file 3 ) and to the commonly accepted phylogenetic tree of primates (Figure 3 ). Three implications can be drawn from the TinT patterns of Alu SINEs: (1) several subtypes of Alu elements were active during overlapping periods, (2) a significant change in Alu activity took place after Tarsius separated from a common ancestor with anthropoids, and (3) the TinT activity profiles correlate perfectly with the well known activity patterns of Alu elements [ 14 ]. Comparing TinT profiles of dimeric Alu elements to the phylogenetic relationships of different primate species documents the correlation between the activity of retroelements and species evolution (Additional file 3 ).
However, comparing the TinT analysis of dimeric Alu elements to the TCF defragmentation pattern [ 7 ] demonstrates that the TinT analysis provides a more accurate activity pattern and implements information about the number of elements in the genome. The TCF defragmentation analysis shows an Alu J activity time span that overlaps with that of both Alu S and Alu Y elements. The TinT profiles clearly indicate that Alu J elements were already silent before the divergence of Anthropoidea and before the appearance of most Alu S subtypes and Alu Y (Additional file 3 ). Furthermore, the TCF analysis requires that any given element type interact with at least 29% of all other analyzed element types. The TinT model requires that a given element type interact with only two additional types. Especially for small amounts of genomic data, this raises the sensitivity drastically.
For TinT analyses it is important to carefully preselect and compile the elements of interest. Large elements (e.g., the 6,000 bp, full length, primate-specific L1P type of LINEs) have a higher chance of being occupied by other elements than do smaller ones (e.g., the 300 bp Alu SINE elements.) Therefore, we advise users to analyze such groups separately (see for example [ 18 ]). TinT analyses of both monomeric and dimeric elements together should be interpreted with care. Alu elements for example are composed of two monomers connected via an oligo(A) sequence. Such A-rich regions are preferred targets for insertions [ 5 , 19 ] and can bias the insertion profile. | Conclusions
The insertion patterns of retroposed elements provide a homoplasy-free character set for tracing the evolutionary history of species [ 20 ]. The insertion of a given element at the same genomic location in two species and its absence in a reference species indicates a close relationship between the two sharing species [ 21 ]. However, randomly choosing retroposed elements for laborious phylogenetic analysis is highly inefficient, while preselecting specific informative element types (e.g., for deep phylogenetic splits) significantly raises the efficiency of downstream experimental analyses. The TinT application provides a priori information about the relative activity periods of given elements (e.g., to investigate old splits by selectively analyzing old elements that were active in the potential common ancestor of a specific group). The phylogenetic application of TinT-derived element activities significantly aided in resolving the evolutionary histories of galliformes [ 3 ], marsupials [ 15 ], and lagomorphs [ 22 ].
It is well known that retroposed elements significantly influence genome evolution, architecture, and gene function; hence, a clear understanding of their insertion events is a key to understanding the genomic architecture of present-day genomes. Therefore, in addition to a statistical compilation of such elements, TinT affords an invaluable tool for analyzing the chronological activity of retroposed elements. Because SINE elements depend on the LINE retroposition machinery for their insertion, their activity is closely connected to LINE activity. As an example, mammalian-wide interspersed elements (MIRs [ 23 ]) coincide with the activity of LINE2 elements and presumably the propagation of MIRs ended after the inactivation of such elements. To understand the dependence of SINE-LINE associations, overlapping activity periods are a first indication of potential interactivity. A potential non-autonomous and autonomous element affiliation was shown for a novel SINE-like snoRTE element and BovB_Plat autonomous retroposons in platypus [ 24 ].
Genome-wide chronological analyses of transposed elements using TinT build on the RepeatMasker detection of elements or fragments thereof. The detection is based on sequence similarity to a predefined compilation of transposons. Although TinT performs a subsequent stringent quality-check of detected fragmented elements, miss-annotations, especially if old and thus highly diverged elements are involved, cannot be completely excluded. Therefore, more sequence data leads to an increase in precision.
In future TinT updates, we plan to implement two additional levels of complexity. By a genome-wide pre-screening of element-specific insertions, we intend to add empirical retroposon information of type-specific (monomer-dimer, short-long elements) insertion probabilities; thus, freeing the algorithm from the assumption that all elements have similar insertion probabilities. Associated with this, we intend to improve the accuracy of TinT analyses by introducing an asymmetric model of element activity, whereby elements will not necessary reach their highest probability of activity at the center of their activity range. Furthermore, we plan to incorporate an absolute time scale of activity by incorporating divergence data of elements. | Background
DNA sequences afford access to the evolutionary pathways of life. Particularly mobile elements that constantly co-evolve in genomes encrypt recent and ancient information of their host's history. In mammals there is an extraordinarily abundant activity of mobile elements that occurs in a dynamic succession of active families, subfamilies, types, and subtypes of retroposed elements. The high frequency of retroposons in mammals implies that, by chance, such elements also insert into each other. While inactive elements are no longer able to retropose, active elements retropose by chance into other active and inactive elements. Thousands of such directional, element-in-element insertions are found in present-day genomes. To help analyze these events, we developed a computational algorithm ( T ranspositions in T ranspositions, or TinT) that examines the different frequencies of nested transpositions and reconstructs the chronological order of retroposon activities.
Results
By examining the different frequencies of such nested transpositions, the TinT application reconstructs the chronological order of retroposon activities. We use such activity patterns as a comparative tool to (1) delineate the historical rise and fall of retroposons and their relations to each other, (2) understand the retroposon-induced complexity of recent genomes, and (3) find selective informative homoplasy-free markers of phylogeny. The efficiency of the new application is demonstrated by applying it to dimeric Alu S hort IN terspersed E lements (SINE) to derive a complete chronology of such elements in primates.
Conclusion
The user-friendly, web-based TinT interface presented here affords an easy, automated screening for nested transpositions from genome assemblies or trace data, assembles them in a frequency-matrix, and schematically displays their chronological activity history. | Implementation
The TinT application is implemented in a Java environment (version 1.5 or higher) and executed from a bioinformatics web page that runs as an applet on the client computer. TinT reads and optimizes RepeatMasker information of nested transposons and transfers this information into a data matrix of transpositions in transpositions (TinT). The data matrix is than included in a probability calculation to derive a graphical framework of relative activity periods of transposed elements. The probabilistic model considers a simplified assumption with just one period of activity of elements and no specific target site preference. The applet calculates the relative activity periods of elements, but in the current version no time calibration is implemented. The usage of the web-based application is illustrated in Additional file 2 .
Availability and Requirements
Project name: TinT
Project home page : http://www.bioinformatics.uni-muenster.de/tools/tint
Operating system : Platform independent (Requires a Java Virtual Machine (JVM) on the target system)
Programming language : Java
Requirements : Java Runtime Environment
License : GPL for academic users
Authors' contributions
GC, JS, WM, and NG conceived of the user-friendly TinT web-interface. AK and GC adjusted the probabilistic model of TinT, NG and WM implemented TinT into the Java-based web-interface. JB contributed computational equipment. JS wrote the paper. All authors read, edited, and approved the final manuscript.
Supplementary Material | Acknowledgements
We thank Marsha Bundman for editing the manuscript. The primate paintings were provided by Jón Baldur Hlíðberg. This work was supported by the Deutsche Forschungsgemeinschaft (SCHM 1469/3-1). | CC BY | no | 2022-01-12 15:21:37 | BMC Evol Biol. 2010 Dec 2; 10:376 | oa_package/e9/6f/PMC3014933.tar.gz |
PMC3014934 | 21126361 | Background
Over the last few years, the phylogenomic approach was successful in untangling several aspects of the early evolution of eukaryotes. Most eukaryotic diversity has been assigned to one of few supergroups [ 1 ] and new relationships between these large assemblages have emerged. For example, the unexpected close evolutionary affinity of Rhizaria to two of the "chromalveolate" groups, stramenopiles and alveolates (the SAR group, or Harosa in [ 2 ]), was recovered in several phylogenomic analyses [ 3 - 6 ]. Even orphan lineages that have been very challenging to place within the eukaryotic tree, such as the telonemids and centrohelids, or the breviate amoebae, have recently been shown to be related to haptophytes and centrohelids or to Amoebozoa, respectively [ 7 - 9 ]. However, several question marks remain, notably concerning the placement of the root [ 10 ], the monophyly of some supergroups [ 11 ], and the relationships within and between the supergroups, especially where uncultivated protists dominate.
The supergroup Rhizaria, composed of several phyla that are difficult to maintain in laboratory cultures, is a good example of the persisting uncertainties for the relationships between the major members of this assemblage. Although a few rhizarians can be isolated and cultivated [ 12 ], the majority is known only from environmental sequences [ 13 ] or single-specimens extractions [ 14 , 15 ]. Consequently, rhizarians are represented in sequence databases almost entirely by ribosomal rDNA [ 16 ]. A few protein sequences of actin, α-tubulin, β-tubulin, RNA polymerase II, and polyubiquitin are available for selected taxonomic groups [ 17 - 20 ] but for other lineages, such as radiolarians, only the actin-coding gene has been sequenced, which is in sharp contrast to the great diversity of the group and its ecological importance. Recently, five small rhizarian cDNA libraries have been sequenced (3 Cercozoa and 2 Foraminifera), partially filling the gap in comparison to other supergroups, and one genome project ( Bigelowiella natans ) is in progress ( http://www.jgi.doe.gov/sequencing/why/50026.html ).
According to the current consensus, Rhizaria are composed of three highly diverse and possibly monophyletic phyla, Cercozoa, Foraminifera, and Radiolaria (including Acantharea, Polycystinea and Taxopodida, but excluding Phaeodarea that were shown to branch among Cercozoa [ 21 ]). The Rhizaria comprise also the parasitic Phytomyxea and Haplosporidia, as well as various marine filose and reticulose protists, including Gromiida and Filoreta , sometimes considered members of Cercozoa [ 22 , 23 ]. The relationships between these groups are uncertain, due to the lack of resolution observed in the SSU and LSU rDNA as well as the few available protein trees. The most controversial is the position of Foraminifera, whose fast evolving SSU rDNA sequences branch either close to Haplosporidia and Gromiida [ 19 , 24 ] or as sister group to Radiolaria [ 13 , 25 , 26 ]. The weakly supported grouping of Foraminifera and Radiolaria observed in some SSU and LSU rDNA trees led to the creation of the infrakingdom Retaria [ 26 , 27 ]. Another source of information came from the insertions of one or two amino acids at the monomer junctions in the highly conserved protein polyubiquitin. These insertions have been found in Cercozoa and Foraminifera but not in all other eukaryotes studied to date, including radiolarians [ 17 , 23 ]. It has been argued that the ancestor of polycystine and acantharean Radiolaria could have lost the insertion, but the lack of insertion could also be explained by contamination of DNA samples by non-rhizarian protists [ 23 ].
To test the Retaria hypothesis and to shed light on the relationships between most of the deeply branching rhizarian groups, a protocol was developed to prepare cDNA libraries suitable for 454 sequencing from a handful of cells collected from environmental samples. We obtained and analyzed more than 670,000 ESTs from 2 marine acantharean Radiolaria ( Astrolonche sp. and Phyllostaurus sp.), 2 parasitic Phytomyxea ( Plasmodiophora brassicae and Spongospora subterranea ) and Gromia sphaerica , a giant marine testate protist that is capable of producing macroscopic bilaterian-like traces [ 28 ]. Phylogenetic analyses of 167 genes support the Retaria hypothesis and suggest that this group may be most closely related to Phytomyxea and Gromia . Moreover, our study confirms the presence of polyubiquitin insertion in some Acantharea and reveals another possible rhizarian-specific signature in one of the ribosomal proteins. | Methods
Collecting and isolation of specimens
G. sphaerica was collected near Little San Salvador Island in the Bahamas at about 720 m depth (24°34.5'N; 076°00.1'W) and total RNA prepared as described in [ 28 ].
Acanthareans were collected during May-June 2008 at the outlet of the Villefranche Bay, Mediterranean sea (43°41'N; 7°18'48E). Plankton samples were taken using a plankton net (mesh diameter 20 μm) drawn vertically from the depth of 200 to 0 m. Concentrated samples were immediately brought to the lab and processed. Living acanthareans were picked from the plankton with needles, washed with filtered seawater and placed in RNAlater solution (Ambion). The solution was allowed to penetrate into the cells for 24 hours at 4°C, after which the samples were kept frozen until further processing. In total about 300 cells of Phyllostaurus and 50 cells of Astrolonche were collected and used for library preparation.
The 2 phytomyxean samples were prepared from in vitro grown callus consisting of S. subterranea infected Solanum tuberosum cells and P. brassicae infected Brassica rapa cells, respectively. Full details of the generation, growth and characterization of these callus lines will be detailed elsewhere (Bulman et al. submitted). Briefly, sections of S. subterranea or P. brassicae root galls were surface sterilized and placed on MS media. Segments of white/green multiplying cells were transferred to new media as they proliferated. Fresh green callus cells were harvested and transferred into RNAlater (Qiagen).
Preparation of cDNA libraries for 454 sequencing
The G. sphaerica cDNA library was prepared from approximately half of an individual, as described in [ 39 ], with minor modifications (described below) that ensured enrichment of the data with protein-coding sequences. Briefly, the methodology involves cDNA synthesis and amplification using SMARTer cDNA synthesis kit (TaKaRa BIO/Clontech, Mountain View, CA), normalization using Trimmer kit (Evrogen, Moscow, Russian Federation), fragmentation by sonication, end-polishing, ligation of adaptors, and amplification of the 454-ready sample. The design of the adaptors ensures that the sequencing proceeds only from the cDNA breaks introduced by sonication, rather than from original termini, which helps to reduce the amount of adaptor-derived sequences in the resulting data.
For acanthareans, we originally prepared the libraries from 100 Phyllostaurus and 20 Astrolonche cells using the same protocol without normalization, but after Sanger-sequencing 24 randomly picked clones per species we found that the libraries consisted predominantly of non-coding sequences, most likely representing 3'-UTRs of the original transcripts. To enrich our libraries with the coding regions, we amplified the SMARTer kit-synthesied cDNA with a long primer (5'-AGTGGACTATCCATGAACGCAAAGCAGTGGTATCAACGCAGAGT-3') at a concentration of 0.1 μM, instead of using the one included in the kit, which resulted in preferential amplification of the longest cDNA fragments due to mild PCR-suppression effect [ 40 ]. Moreover, after fragmentation and adaptor ligation, only the 5'-ends of the original cDNAs were amplified and subjected to sequencing, by using the primer annealing to the ligated sequencing adaptor [ 39 ] and the primer matching to the unique sequence of the template-switch oligonucleotide used during the cDNA synthesis (5'-GCCTCCCTCGCGCCATCAGCCGCGCAGGTACGTATCAACGCAGAGTACGCGG-3'). The libraries were sequenced using 454 GS-FLX. The latest version of the cDNA preparation protocol, adapted for the latest 454 version (Titanium), is available on Matzlab website [ 41 ].
For the 2 phytomyxean species, cDNA libraries were constructed by Vertis Biotechnology AG (Germany) according to their Random-Primed (RPD) cDNA protocol. Frozen cells were ground under liquid nitrogen and total RNA isolated from the cell powder using the mirVana RNA isolation kit (Ambion). Poly(A)+ RNA was prepared from total RNA. First-strand cDNA synthesis was primed with an N6 randomized primer and second-strand cDNA was synthesized according to the classical Gubler-Hoffman protocol [ 42 ]. Double stranded DNA (dsDNA) was blunted and 454 adapters A and B ligated at the 5' and 3' ends. dsDNA carrying both adapter A and adapter B attached to its ends was selected and amplified with PCR using a proof reading enzyme (24 cycles). For 454 sequencing the cDNA in the size range of 250 - 600 bp was eluted from a preparative agarose gel.
Contig assembly and sequence alignment
All newly generated reads were assembled into contigs using the Newbler assembler with default parameters, generating the following number of contigs larger than 100 bp: G. sphaerica , 24,433; Astrolonche sp., 6426; Phyllostaurus sp., 5056; P. brassicae , 27,333; S. subterranea , 14,531. To filter plant sequences out of the phytomyxean datasets, the phytomyxid-callus contigs were compared to plant cDNA sequences using the BLAT tool [ 43 ] and those with very high similarity were discarded (e-value threshold < 1e-50). The 27,333 P. brassicae contigs (containing B. rapa ) were compared against 2,529,141 Brassicaceae ESTs (NCBI as of 25 March 2009, http://www.ncbi.nlm.nih.gov /) and resulted in 24,166 contigs showing nearly identical hits. The 14,531 S. subterranea contigs (containing S. tuberosum ) were compared to 1,000,784 Solanaceae ESTs (NCBI as of 25 March 2009, http://www.ncbi.nlm.nih.gov /), giving 8,282 contigs with a nearly identical hit. Manual inspection (using blastn) of a subset of the filtered sequences confirmed that these were indeed likely to originate from the plant host cells. The remaining 3,167 P. brassicae contigs with low or no hit against the Brassicaceae ESTs and the 6,249 S. subterranea contigs with low or no hit against the Solanaceae ESTs were used in subsequent screenings.
Blast searches against databases containing the 5 species above and taxa downloaded from GenBank ( http://www.ncbi.nlm.nih.gov/ ) and JGI ( http://genome.jgi-psf.org/ ) were performed to retrieve the genes of interest (see Table S1 for the list of species). These genes corresponded in part to genes that we used in previous phylogenomic studies [ 3 , 4 , 7 ], but also to 55 genes representing additional members of large protein families that were not previously included in our alignment, such as more minichromosome maintenance proteins (MCM) or more proteasome subunits, but also several new ribosomal proteins and proteins for which a broad sampling was available (see Additional file 3 for a list of the newly added genes). In total, 202 single-gene alignments were constructed, automatically aligned with Mafft [ 44 ], using Gblocks [ 45 ] to remove ambiguously aligned positions (with half of the gapped positions allowed, the minimum number of sequences for a conserved and a flank position set to 50% of the number of taxa plus one, the maximum of contiguous non-conserved positions set to 12, and the minimum length of a block set to 5) and followed by manual adjustment when needed with BioEdit [ 46 ]. The orthology and possible contamination in each gene was tested by ML reconstructions with 100 bootstrap replicates using RAxML 7.2.2 (LG substitution matrix) [ 47 ], and visual check of the resulting individual trees. Out of the complete set of 202 genes, 23 showed evidence for deep paralogy and were therefore discarded. In addition, we also excluded 12 extra genes as they did not contain any species of Rhizaria. Our final dataset contained 167 genes, which represented 36,735 amino acid positions after concatenation, and 39 species belonging to Rhizaria, stramenopiles, alveolates and green plants (outgroup). The single-gene and concatenated alignments are available at http://www.fabienburki.com . A table listing all genes and a detailed view of the missing data repartition for each taxa can be found in Additional file 1 . The 167 trees constructed from the final selection of genes are available in Additional file 4 . Since many of the Phytomyxea contigs were short due to the relatively small number of contigs that were obtained from the phytomyxid parasites, 35 genes were chosen as targets for further sequence acquisition. PCR primers were designed based on the short contigs and longer DNA sequences were obtained by RT-PCR or 3'RACE using the callus RNA as template (as in [ 48 ]). Scafos 1.2.5 [ 49 ] was used for performing the concatenation process of the single-genes.
The actin and beta-tubulin alignments were built by retrieving from GenBank sequences belonging to all rhizarian (actin) or eukaryotic (beta-tubulin) groups, and adding to them the sequences identified in the datasets generated in this study. 72 rhizarian and 6 stramenopiles (outgroup) sequences were included in the actin tree; 119 eukaryotes were analyzed for the beta-tubulin tree. We also searched our datasets for the 1 or 2 amino acids insertion at the monomer-monomer junctions carried in most rhizarian species. Finally, the construction of the single-gene alignments revealed a rhizarian specific insertion of 2 amino acids in the 60S ribosomal protain L10a.
Phylogenetic analyses
ML analyses were performed using RAxML 7.2.2 [ 47 ] in combination with the LG amino acid replacement matrix [ 50 ]. The best ML tree was determined with the PROTGAMMA + F implementation in multiple inferences using 10 randomized parsimony starting trees. Statistical support was evaluated with 100 bootstrap replicates. Bayesian analysis using the WAG + G + F model (4 gamma categories) was done with the parallel version of MrBayes 3.1.2 [ 51 ]. The inference consisted of 1,000,000 generations with sampling every 100 generations, starting from a random starting tree and using 4 Metropolis-coupled Markov Chain Monte Carlo (MCMCMC). 2 separate runs were performed to confirm the convergence of the chains. The average standard deviation of split frequencies was used to assess the convergence of the 2 runs. Bayesian posterior probabilities were calculated from the majority rule consensus of the tree sampled after the initial burnin period as determined by checking the convergence of likelihood values across MCMCMC generations (corresponding to 50'000 generations). PhyloBayes 3.1 [ 52 ] was run using the site-heterogeneous mixture CAT model with the rates-across-sites heterogeneity handled by a Dirichlet process (ratecat). 2 independent Markov chains with a total length of 19'000 cycles were performed, discarding the first 2'000 points as burnin, and calculating the posterior consensus on the remaining trees. Convergence between the 2 chains was ascertained by examining the difference in frequency for all their bipartitions (< 0.1 in all analyses). Bootstrap CAT proportions were obtained after 5000 cycles with a conservative burnin of 1000 on 100 pseudo-replicates generated with Seqboot (Phylip package [ 53 ]). Manual verification of 10 replicates showed that the burnin is generally between 500-700 cycles. For each replicate, trees were collected after the initial burnin period and a consensus tree was computed by readpb (PhyloBayes package). Consense (Phylip package [ 53 ]) was then used to calculate the bootstrap support based on these 100 consensus trees. Due to limited size of single-genes for parameter estimation under non-parametric models such as CAT, the PhyloBayes-based actin and beta-tubulin phylogenies were ran under the LG model.
The site-removal analysis was performed using PAML [ 54 ] to identify the fast-evolving sites, as implemented in the AIR package [ 55 ]. Because the topology chosen to estimate the site-wise rates strongly influences the results [ 32 ], rates were calculated for the 105 possible different topologies describing the evolutionary relationships of the 5 rhizarian lineages, and sorted according to the mean of the rates estimated on all topologies. 5% to 90% (10% intervals between 10% and 50%, 5% intervals between 55% and 90%) of the fastest evolving sites were then removed (percentage of the total rate distribution), and bootstrapped ML analyses were run with each of these 14 shorter alignments. PhyloSort [ 56 ] was used to search the pseudo-replicate trees for the relationships of interest. Phylobayes analyses with the CAT model were also done on each reduced alignment.
The statistical model comparison was done using the cross-validation (CV) method available in PhyloBayes 3.1 [ 52 ]. A learning and a test sets were generated by randomly splitting (no replacement) the original alignment into 10 replicates made of 90% and 10% of the original sites, respectively. Each of the learning and test alignments amounted to 33,062 and 3673 positions, respectively. A MCMC run was performed for each replicate under a fixed topology (either the "CAT" or "LG" tree) for a total of 5000 cycles ("CAT") or 1500 cycles ("LG"). Due to technical reasons associated with this test in PhyloBayes, a discrete gamma distribution with 4 categories was used for modelling the rate heterogeneity across site (i.e. dgam instead of ratecat). This slightly different model should not affect the conclusions given the big difference between CAT and LG. The lower number of cycles under "LG" was due to a much greater computational time per cycle as compared to when the "CAT" model was used. The first 500 and 150 points were discarded as burnin for the "CAT" and "LG" runs, respectively, and the remaining points used to compute the cross-validation log-likelihood.
Topology comparisons were conducted using the approximately unbiased (AU) test [ 30 ]. For each tested tree, site likelihoods were calculated using RAxML 7.2.2 with the LG model and the AU test was performed using CONSEL [57]. | Results
Dataset construction
The phytomyxean P. brassicae and S. subterranea are parasites of the plant genera Brassica and Solanum , respectively, and the in vitro callus samples were prepared according to an unpublished protocol (Bulman et al. submitted). Consequently, an unknown amount of plant contamination was expected in the phytomyxean ESTs. An initial blast examination showed that many of the phytomyxid-callus contigs had high similarity to plant sequences and were thus possibly derived from the host cells. We took advantage of the large amount of data available for Brassica and Solanum to filter out these plant sequences and simplify data searching for constructing the single-gene alignments (see Methods).
A total of 167 gene alignments with at least one rhizarian species represented in each were constructed for phylogenetic analyses. Based on recently published results suggesting close evolutionary affinities between Rhizaria, stramenopiles and alveolates, forming the so-called SAR group [ 3 , 5 , 6 ], representatives for these 3 groups were included. The full dataset comprised 10 rhizarians, 9 stramenopiles, and 9 alveolates. In order to reduce the risks of artifacts, 11 green plant taxa were chosen to root our phylogenies because 1) of the availability of complete genomes for many lineages, thus considerably reducing the amount of missing data in the outgroup; 2) they have evolved more slowly comparatively to most of the SAR species; and 3) their relative evolutionary proximity to the SAR group in the tree of eukaryotes [ 4 ]. However, an alternative outgroup, haptophytes, was also tested as it was proposed to be more closely related to the SAR group (data not shown) [ 7 ]. We did not select it for the final analyses because only medium-sized EST datasets are available for a limited number of species, except for one complete genome ( Emiliana huxleyi ), and the intra-Rhizaria relationships remained identical to the trees rooted using the green plants (see below). Each single-gene dataset was thoroughly tested by bootstrapped maximum likelihood (ML) analyses for deep paralogy or suspicious relationships possibly indicative of lateral gene transfer (LGT) or contamination. The acanthareans are known to harbor zooxanthellae symbionts and polycystine radiolarians are hosts of prasinophytes, dinoflagellates and other alveolates [ 29 ]. Accordingly, non-acantharean sequences were expected to be found. Out of the 167 selected genes, we could identify 1 sequence related to haptophytes in Astrolonche and Phyllostaurus , 2 sequences of dinoflagellate origin in Astrolonche , 5 and 2 sequences of general plant affinity in Astrolonche and Phyllostaurus , respectively, and, surprisingly, 25 sequences in Astrolonche and 21 sequences in Phyllostaurus clearly belonging to streptophytes (angiosperms). It is not clear to us why streptophyte sequences were present in our acantharean dataset, but one possible explanation could be that the samples were contaminated with a small amount of pollen. All these contaminant sequences were removed from our alignments. The curated protein alignments were concatenated into a supermatrix amounting to 36,735 unambiguously aligned amino acid positions (global percentage of missing data: 40%; see Additional file 1 for details) that was subjected to phylogenetic analyses.
Phylogenetic analyses of the supermatrix
We analyzed our concatenated alignment using probabilistic methods of tree reconstruction with (i) empirical site-homogeneous models of sequence evolution in ML (LG model) and Bayesian (WAG model) frameworks and (ii) site-heterogeneous mixture model in a Bayesian framework (CAT model). Figures 1 and 2 depict the relationships inferred from these analyses. The "LG" and "WAG" trees received maximal bootstrap support proportions (BP) and posterior probabilities (PP), respectively, for nearly all nodes (Figure 1 ). As expected, the 3 major groups included in this study, i.e. rhizarians, stramenopiles and alveolates, were strongly recovered, and the relationships between and within them corresponded to previously published trees [ 4 , 6 , 7 ]. All 3 analyses robustly supported the monophyly of the four rhizarian phyla: Foraminifera, Acantharea, Phytomyxea and Cercozoa. Notably, Foraminifera were placed as a sister group to Acantharea with 100% BP ("LG" and "CAT") and 1.0 PP ("WAG" and "CAT"). The branching order within Rhizaria was identical and highly supported in the "LG" and WAG" trees, with the Foraminifera + Acantharea clade being sister to Cercozoa and a group composed of Phytomyxea and Gromia (Figure 1 ). On the other hand, the site-heterogeneous CAT model inferred a different topology that received low to high PP and BP, recovering the Foraminifera + Acantharea group in an internal position, sister to Gromia (0.93 PP; 90% BP) and more closely related to Phytomyxea (0.51 PP; 50% BP) to the exclusion of Cercozoa (Figure 2 ). Because the discrepancies between the "LG" and "CAT" topologies are an indication that some relationships may be artifactual, we estimated the fit of these 2 models based on a cross-validation test (see Methods). The "CAT" model was found to much better fit the data than the "LG" model with a score averaged over 10 replicates of 1547 ± 71 (all replicates favoured the "CAT" model), indicating that the topology in Figure 1 is likely the results of biases not correctly handled by the site-homogeneous models (LG and WAG).
To better evaluate these differences, a topology comparison analysis using the approximately unbiased (AU) test was performed [ 30 ]. Both trees in figure 1 ( P = 0.916) and in figure 2 ( P = 0.084) were not rejected at the 5% significance level. This test was based on the comparison of trees obtained with 2 non-nested models, "LG" (Figure 1 ) and "CAT" (Figure 2 ), using the "LG" empirical matrix. Hence, if the topology in Figure 2 had been rejected, it would not have been very informative because the "CAT" model could still have inferred the true tree. In the present case, however, the LG-based AU test kept the "CAT" tree among the trees possibly correctly describing the relationships within Rhizaria, thus strengthening the branching pattern showed in Figure 2 . In addition, a topology with Acantharea alone in a sister position to the rest of Rhizaria was also tested in order to estimate the likelihood of the basal branching of Radiolaria seen in some SSU trees (see [ 16 ] for a discussion). This topology was strongly rejected ( P = 7e-09), further supporting the association of Foraminifera and Radiolaria.
Evaluating the branching order within Rhizaria
In our trees, both foraminiferans and acanthareans appeared as fast-evolving taxa. This raised a concern about their potentially erroneous grouping due to the long branch attraction (LBA) artifact [ 31 ] that would affect not only the position of these diverging lineages but also the relationships among all rhizarian groups. To evaluate for the possibility of LBA, we first conducted a fast-evolving taxa removal experience in which, in turn, the most diverging foraminiferan representative Reticulomyxa filosa (Figure 3 ), both foraminiferan species (Figure 4 ), and the acanthareans (Figure 5 ) were discarded. The removal of R. filosa had no impact on the sister relationship of foraminiferans and acanthareans: both groups remained monophyletic with maximum support. However, this slightly different taxon sampling largely affected the branching order among the rhizarian groups. The "LG" model robustly placed Gromia as the most closely related lineage to the Foraminifera + Acantharea group (93% BP), and Phytomyxea were recovered as sister to this assemblage with 87% BP (Figure 3A ). The "CAT" model inferred the same topology (Figure 3B ), which also corresponded to the full tree inferred with this model (Figure 2 ) but, interestingly, the support values increased from 0.92 to 1.0 PP and from 0.51 to 0.92 PP for the node joining Foraminifera + Acantharea + Gromia and the node uniting Phytomyxea to this group, respectively. Similarly, when Foraminifera were removed altogether, both models again recovered the "CAT" topology (Figure 2 ) with high BP and PP values, exactly as in absence of Reticulomyxa only (Figure 3 ). Finally, discarding Acantharea led in both "LG" and "CAT" analyses to the basal position of Foraminifera (98% BP; 0.73 PP) and the sister position of Gromia to Phytomyxea (100% BP; 0.9 PP), as observed in the complete "LG" tree (Figure 5 and Figure 1 ).
To assess the robustness of the Foraminifera-Acantharea clade and to further investigate the two competing topologies for intra-Rhizaria relationships (Figure 1 and 2 ), we then conducted a site removal analysis in which the fastest-evolving sites were progressively removed from the original alignment. The rationale behind this analysis is that fast-evolving sites are more likely to be saturated and not correctly interpreted as convergence by tree reconstruction methods, thus strongly influencing the potential artifactual grouping of highly diverging lineages [ 32 ]. Specifically, we tested 14 shorter alignments ranging from 35,230 aa to 14,281 aa and reconstructed phylogenetic trees with LG and CAT models at each step to determine the support value for several nodes of interest (Figure 6 ). First, the highly supported association between Foraminifera and Acantharea was not affected by the removal of fast-evolving sites, with almost no decrease in bootstrap values even for the smallest number of positions remaining in the alignment. This result provides additional evidence that the grouping of Foraminifera and Acantharea is not caused by artifacts of tree reconstruction. Second, we monitored the bootstrap supports for the sister position of Gromia with respect to Phytomyxea, the basal position of the Foraminifera-Acantharea clade (as observed in the "LG" tree, Figure 1 ), as well as the alternatives: the sister grouping of Gromia to the Foraminifera-Acantharea group, and the basal position of Cercozoa (as observed in the "CAT" tree, Figure 2 ). Interestingly, as the fast-evolving sites were removed, the bootstrap values for the phylogenetic relationships obtained in the LG-based analysis of the complete dataset decreased (Figure 6 , blue line) and, at the same time, the branching order supported by the CAT-based reconstruction gained statistical significance (Figure 6 , red line). When 13'379 fast-evolving positions were removed, the LG-based analysis converged with high support (94% BP; 0.99 PP) towards the topology that was weakly suggested by the CAT-based analysis of the complete dataset for the association of Phytomyxea, Gromia , Foraminifera and Acantharea, before diverging likely due to lack of phylogenetic signal in the shortest alignments. The position of Gromia remained more ambiguous throughout the removal process and, although the support for the association with Phytomyxea rapidly decreased to below 50% BP, its sisterhood to the Foraminifera-Acantharea clade suggested by the "CAT" model did not gain significance.
Actin phylogeny
Although our multigene analysis represents the broadest rhizarian sampling to date, three important rhizarian groups, Haplosporidia, Filoreta , and Polycystinea, are still missing. Therefore, we performed a separate phylogenetic analysis based on actin, the only protein-coding gene sequenced in all rhizarian groups. ML and Bayesian analyses of our alignment (317 amino acid positions), containing 73 rhizarians and 6 stramenopiles as outgroup, indicated that the acantharean Astrolonche possesses 2 actin paralogues branching as sister groups to 2 of the actin paralogues present in Foraminifera (Figure 7 ). The only actin sequence found in Phyllostaurus grouped with Astrolonche as sister to the foraminiferal paralogue 2. Sister to this clade were two previously obtained actin sequences of the polycystinean radiolarians Thalassicolla pellucida and Collozoum inerme , and their grouping with Acantharea and Foraminifera was strongly supported in Bayesian inferences (0.99 and 1.0 with PhyloBayes and MrBayes, respectively) but not supported in ML (31% BP). However, the relationships between these 3 groups remained unresolved, leaving open the question of a possible radiolarian monophyly. For both paralogues, Haplosporidia appeared as sister to the Foraminifera + Radiolaria clade, albeit without much support.
Rhizarian signatures
In addition to the multigene and actin analyses, we screened our newly generated data for the presence of molecular signatures characteristic of Rhizaria. First, polyubiquitin sequences were searched for the 1 or 2 amino acid insertion previously described at the monomer-monomer junction in all Rhizaria except in Radiolaria [ 17 , 22 ]. We found threonine (T) in 4 sequences of Astrolonche and one sequence of Phyllostaurus and alanine (A) in 6 sequences of Phyllostaurus (Figure 8A ). The presence of 2 different amino acids in Phyllostaurus was surprising, but this is not exceptional as it has already been observed in Lotharella amoeboformis (AY099125) where both A and S insertions have been found [ 17 ]. A serine (S) was also found in the Gromia sphaerica sequence, which was identical to the available polyubiquitin of Gromia oviformis (AY571670). In addition, a new polyubiquitin sequence amplified from the phagomyxid Maullinia ectocarpii was included.
Interestingly, we identified a new insertion of 2 and 4 amino acids in the 60S ribosomal protein L10a, a characteristic also apparently unique to Rhizaria. A phenylalanine (F), an asparagine (N), and a serine (S) followed by a lysine (K) were inserted at position 104 in G. sphaerica , R. filosa , B. natans , and Paracercomonas sp., respectively (Figure 8B ). In G. sphaerica , the sequence contained 2 additional inserted amino acids, i.e. a valine (V) and a glycine (G). Unfortunately, this gene was not present in the acantharean dataset and several attempts to amplify it by PCR failed. Blast searches against GenBank-nr and dbEST revealed no other known rpl10a gene containing this insertion. | Discussion
This study provides the first robust evidence for a relationship between Foraminifera and Acantharea, a member of Radiolaria. This result is rather surprising, taking into account the considerable differences in morphology, composition of the skeleton, and lifestyle between these groups. Radiolarians have intracellular celestite (SrSO 4 ) (in Acantharea) or siliceous (in Polycystinea) skeleton consisting of strontium sulphate and are holoplanktonic. In contrast, the foraminiferal skeleton (when present) is extracellular, agglutinated or calcareous, and the majority of foraminiferans are benthic. Pseudopodia morphology is also markedly different: radiolarians possess stiff, ray-like pseudopodia called axopodia, while foraminifers are defined by the presence of fine, anastomosing granuloreticulopodia. Still, there are also common cell characteristics shared between these 2 groups, the importance of which must be re-evaluated in view of our data. For example, the network of fine reticulopodia observed in some radiolarians exhibits bidirectional streaming, and is used for capturing prey and locomotion in a similar manner as the foraminiferal granuloreticulopodia [ 29 ]. Further studies of proteins involved in pseudopodial formation in both groups are needed to examine these properties at the molecular level. In that respect, it is interesting to note that our acantharean ESTs contained an unusual beta-tubulin strongly resembling the highly diverging type 2 sequences reported in Foraminifera by [ 33 ], as well as a less diverging isoform weakly grouping with a new foraminiferan beta-tubulin type (here named "type 3") found in R. filosa cDNA library (Additional file 2 ).
The clustering of Foraminifera and Acantharea observed in our analyses partially confirms the Retaria hypothesis [ 26 ]. Although multigene data for the 2 other main groups of Radiolaria, Polycystinea and Taxopodida, are still unavailable, we predict that they will also group with Foraminifera. This relationship is suggested by the phylogenetic position of three fast-evolving sequences of polycystinean actin as sister to foraminiferan actin paralogue 2 [ 19 ] as well as by the grouping of Foraminifera with environmental clones assigned to Polycystinea and Sticholonche in an analysis of combined SSU and LSU rDNA [ 25 , 34 ]. However, the branching order of these groups was uncertain and Foraminifera may in fact branch within the radiolarian clade, suggesting that Radiolaria (Radiozoa) could be paraphyletic [ 25 , 34 , 35 ]. The next challenge will be testing whether this surprising pattern arises due to an artifact of LBA, and testing the monophyly of radiolarians.
Further effort is also required to resolve the relationships among the rhizarian groups. Thus far, all phylogenetic studies of this supergroup have recovered Radiolaria alone, or together with Foraminifera as the most basal clade [ 16 ]. The latter topology was supported by our LG and WAG-based tree reconstructions (Figure 1 ), but not by the Bayesian inference with the CAT model (Figure 2 ). Instead, this method suggested that Retaria are closely related to Gromia and Phytomyxea, to the exclusion of Cercozoa. Although this association received only low support with the full dataset, it was strengthen by the experiments with the foraminiferans or the fast-evolving sites removed, as well as by the AU and cross-validation tests. The removal of R. filosa was particularly informative in indicating that this species alone could have attracted Retaria at the base of Rhizaria in the "LG" tree, due to its high rate of evolution. Indeed, when it was not included, the topology suggested by the "CAT" model was robustly recovered. Acanthareans, on the other hand, seemed to be less prone to LBA as they consistently branched in a derived position when both foraminiferans were removed. The fast-evolving sites removal analysis also convincingly supported the grouping of Retaria, Gromia , and Phytomyxea, and was in agreement with the properties of the CAT model; it has been shown that this model infers homoplasies better than empirical models (such as the LG model used in the RAxML analyses) [ 36 ]. Therefore, it might be interpreted that for our complete dataset, the CAT model detected and correctly interpreted more of the saturated positions that were misleading in the RAxML analysis. Within this group, the position of Gromia could not be inferred with precision as it branched either as sister to Retaria or Phytomyxea. However, the grouping of Gromia with Phytomyxea was recovered only when Acantharea were absent or R. filosa was included in the analyses with the "LG" model. Moreover, the better fitted CAT model robustly placed Gromia in a sister position to Retaria. Interestingly, the association of Foraminifera, Acantharea, Gromia and Phytomyxea has never been described, although SSU and actin trees showed generally unsupported relationships between some but not all lineages [ 37 , 38 ]. In addition, other lineages, such as Haplosporidia or Filoreta also belong to this group and will likely be crucial for resolving the internal branching order.
Finally, our study clarifies the question whether acanthareans and polycystines truly lack the rhizarian-specific polyubiquitin insertion, as previously reported [ 22 ]. To explain the apparent absence of the insertion in these two groups, it has been proposed that it was lost in radiolarians, or was acquired after their divergence [ 22 ]. In our EST data, both acantharean species feature polyubiquitin sequences with the insertion, suggesting that the sequences presented in [ 22 ] were not of acantharean origin, but perhaps originated from unidentified symbionts. | Conclusions
Our multigene analysis elucidates the relationship between two important rhizarian phyla, Foraminifera and Radiolaria (Acantharea), which has been a matter of recent controversy. Because Acantharea do not fully represent the radiolarian diversity and genomic data for other important groups (Polycystinea and Taxopodida) is still missing, we cannot rule out the possibility that Radiolaria are paraphyletic. Nevertheless, our study strongly indicates that a basal position of Radiolaria with respect to the rest of Rhizaria is highly unlikely. Instead, our analysis suggests a novel grouping including Foraminifera, Radiolaria, Gromia and Phytomyxea. Within this group, Gromia might be most closely related to Foraminifera and Radiolaria, but its specific phylogenetic position will depend on other important lineages such as Haplosporidia or Filoreta . | Background
Recent phylogenomic analyses have revolutionized our view of eukaryote evolution by revealing unexpected relationships between and within the eukaryotic supergroups. However, for several groups of uncultivable protists, only the ribosomal RNA genes and a handful of proteins are available, often leading to unresolved evolutionary relationships. A striking example concerns the supergroup Rhizaria, which comprises several groups of uncultivable free-living protists such as radiolarians, foraminiferans and gromiids, as well as the parasitic plasmodiophorids and haplosporids. Thus far, the relationships within this supergroup have been inferred almost exclusively from rRNA, actin, and polyubiquitin genes, and remain poorly resolved. To address this, we have generated large Expressed Sequence Tag (EST) datasets for 5 species of Rhizaria belonging to 3 important groups: Acantharea ( Astrolonche sp., Phyllostaurus sp. ), Phytomyxea ( Spongospora subterranea, Plasmodiophora brassicae ) and Gromiida ( Gromia sphaerica ).
Results
167 genes were selected for phylogenetic analyses based on the representation of at least one rhizarian species for each gene. Concatenation of these genes produced a supermatrix composed of 36,735 amino acid positions, including 10 rhizarians, 9 stramenopiles, and 9 alveolates. Phylogenomic analyses of this large dataset revealed a strongly supported clade grouping Foraminifera and Acantharea. The position of this clade within Rhizaria was sensitive to the method employed and the taxon sampling: Maximum Likelihood (ML) and Bayesian analyses using empirical model of evolution favoured an early divergence, whereas the CAT model and ML analyses with fast-evolving sites or the foraminiferan species Reticulomyxa filosa removed suggested a derived position, closely related to Gromia and Phytomyxea. In contrast to what has been previously reported, our analyses also uncovered the presence of the rhizarian-specific polyubiquitin insertion in Acantharea. Finally, this work reveals another possible rhizarian signature in the 60S ribosomal protein L10a.
Conclusions
Our study provides new insights into the evolution of Rhizaria based on phylogenomic analyses of ESTs from three groups of previously under-sampled protists. It was enabled through the application of a recently developed method of transcriptome analysis, requiring very small amount of starting material. Our study illustrates the potential of this method to elucidate the early evolution of eukaryotes by providing large amount of data for uncultivable free-living and parasitic protists. | Authors' contributions
FB and JP conceived the project. AK collected the acanthareans. MVM and GVA collected G. sphaerica and prepared the cDNA libraries for G. sphaerica and the acanthareans. SB isolated the plasmodiophorids. MF analyzed the plasmodiophorid ESTs. FB prepared the alignments and performed the phylogenetic analyses. FB, AK, MVM, GVA, SB, PJK and JP contributed to writing the manuscript and approved its final version.
Accession numbers
454 reads generated in this study for G. sphaerica , Astrolonche sp. and Phylostaurus sp. were deposited in GenBank under the study accession SRP004044.1 , and for P. brassicae and S. subterranea under the study accession SRP003604.2 . The RACE products for P. brassicae and S. subterranea were deposited in GenBank under the accession HO772678 - HO772709 . The new Maullinia ectocarpi polyubiquitin sequences were deposited in Genbank under the accession HQ366774 - HQ366778 .
Supplementary Material | Acknowledgements
We thank John Dolan and Collette Febvre from Biological Station in Villefranche Sur Mer for help in collecting and identifying acanthareans, as well as 2 anonymous reviewers for valuable comments and suggestions The project was supported by the Swiss NSF grant 31003A-125372, National Oceanic and Atmospheric Administration Office of Ocean Exploration Grant # NA07OAR46000289 (''Operation Deep Scope 2007'') to MVM, and a G. & A. Claraz Donation. FB is currently supported by a prospective researcher postdoctoral fellowship from the Swiss National Science Foundation and by a grant to the Centre for Microbial Diversity and Evolution from the Tula Foundation. The phylogenetic analyses were performed on the freely available Bioportal at the University of Oslo ( http://www.bioportal.uio.no ) and the Vital-IT ( http://www.vital-it.ch ) Center for high-performance computing of the Swiss Institute of Bioinformatics. | CC BY | no | 2022-01-12 15:21:37 | BMC Evol Biol. 2010 Dec 2; 10:377 | oa_package/1a/2b/PMC3014934.tar.gz |
PMC3014935 | 21143917 | When species shift their ranges to track climate change, they are almost certain to experience novel environments to which they are poorly adapted. Otaki and co-workers document an explosion of wing pattern variation accompanying range expansion in the pale grass blue butterfly. This pattern can be replicated in the laboratory using artificial selection on cold shocked pupae, at temperature extremes typical of recently colonized environments. We discuss how this phenotypic plasticity may be associated with successful colonization and how significant local adaptation is likely to re-establish developmental control. Integrating knowledge of trait plasticity into current genetic models of adaptation is central to our understanding of when and where a colonising population will be able to persist and adapt in novel surroundings. | Commentary
Species ranges are never fixed, but remain in continual flux in response to demographic, genetic, ecological and environmental variation. Colonization occurs at the range margin when populations spill over into new sites, typically followed by population extinction as environmental and other forces prevent persistence in these new habitats. As the climate has warmed, this turn-over has resulted in expansions that appear more permanent. Consistent northward range shifts have been documented for several vertebrate and invertebrate species in the northern hemisphere, but particularly Lepidoptera, whose historical distributions are usually well-known [ 1 , 2 ]. But what do colonizers look like in these new populations - are they just a sample of the main population or do they differ phenotypically or genotypically? How do these new populations adapt to the novel environmental conditions they encounter and is change likely to persist over longer time scales?
Joji Otaki and co-workers [ 1 ] have charted recent range expansion in the beautifully named pale grass blue butterfly ( Zizeeria maha ). From 1990-2000, this species progressively marched more than 100 km up the west coast of Japan, towards the top of Honshu island - known in precise detail thanks to assiduous collecting of amateur Japanese lepidopterists. Similar northward movements are known in several other Japanese butterflies [ 1 ]. Temperature records suggest increasing summer temperatures facilitated the northward range shift.
Accompanying this colonisation, the pale grass blue has undergone an outbreak of wing pattern diversification. Three novel ventral wing patterns are recognizable with either inner spot elongation, outer spot elongation or reduction of both inner and outer spots (Figure 1 ). Novel phenotypes occurred at high frequencies in the new environments (10-15%), but were hardly ever seen across the main species range. These individuals showed no other obvious phenotypic changes, and the response was not similar to general stress responses seen in other butterflies [ 3 , 4 ]. A possible explanation of this diversity is the breakdown of developmental canalisation due to cooler winter temperatures in more northerly habitats. This is plausible as the pale grass blue butterfly typically does not experience temperatures less than 12°C across its largely tropical, southerly range. To investigate this, Otaki et al. [ 1 ] exposed pupae derived from the eggs of females caught in a southern population to cold-shock (6°C) and found this generated all three novel spot patterns. They also found a dose-response, with more novel forms developing after longer exposure to cold shock (10, 15 or 20 days).
Phenotypic plasticity has been proposed to be important in colonization and persistence in new environments [ 5 - 7 ]. Plasticity in a trait can reduce the difference between the mean phenotype of a recently-colonized population and the new environmental optimum, enabling the population to persist and so have sufficient time to adapt to the change in environment. Conversely, it could drive the mean phenotype away from the environmental optimum. In either case, plasticity increases the standing genetic load, by increasing phenotypic variance around the optimum. In the pale grass blue butterfly, the most obvious effect of colder temperatures in the northern environment is to vastly increase variation in wing spot patterns [ 1 ]. Whether there is a net benefit or loss in fitness associated with this is unclear. Similar responses to cold shock were seen with pupae derived from females caught in the northern range. However the eclosion rate was higher and the induction of wing pattern variation was less. This hints at phenotypic plasticity enabling individuals to gain a foothold outside the normal range, followed by local adaptation to more extreme temperatures and then the re-emergence of canalization to control variation in wing patterns.
Another feature of the northern population is that the novel wing spot patterns appear in summer butterflies. These could not have been exposed to lower temperatures as pupae, so temperature can't be the sole explanation. To investigate this further, Otaki et al. [ 1 ] undertook artificial selection for outer spot elongation under laboratory conditions. To simulate northern conditions, pupae were subjected to temperature shock during development. Otaki et al. observed a quick evolutionary response to selection, with over 80% of individuals developing modified wing patterns by the fifth generation. Elongated wing spots then started to appear without cold shock, with more than 40% of individuals showing novel wing spots without cold treatment by the tenth generation. This appears to be an example of "genetic assimilation" as proposed by C H Waddington in the 1940s, which proposes that exposure to a novel environment causes the expression of previously hidden genetic variation, which is then "assimilated" and expressed even in the absence of the novel environmental stimulus [ 8 , 9 ]. Such a process could underlie the appearance of butterflies with elongated spots in summer populations across the northern distribution of the pale grass blue which would not have been exposed to low temperatures.
For this explanation to make sense, selection needs to have favoured novel spot patterns in colonizing butterflies. Little is known about natural or sexual selection on ventral spot patterns in the pale grass blue, so this remains somewhat hypothetical. However, recent theoretical work has shown that sexual selection can enhance the effect of random genetic drift and act to diversify mate preferences and sexual colour patterns [ 10 ]. A change in mate preference could change the adaptive value of elongated spots from maladaptive in the main range, to adaptive in the new environment.
Another, perhaps more plausible, explanation is that selection favours traits in the novel environment that are genetically correlated to developmental control of wing patterning. It seems likely that there has been local adaptation to the novel colder environmental conditions in the colonizing population. Could such traits be genetically correlated to wing pattern traits? Unfortunately it is not clear from Otaki et al.'s study whether artificial selection on wing spots caused correlated genetic change in adaptation to cold shock, which would be indicative of an underlying genetic correlation. Pupal survival rate under temperature shock increased through time in the selection line, as predicted under this hypothesis. But there was no appropriate control. We need to know the degree of adaptation (i.e. pupal survival rate) of a population held under the same conditions of cold shock but without artificial selection on wing spots. Moreover, there was only a single replicate of artificial selection in Otaki et al.'s study; several would be needed to exclude random effects.
Certain types of phenotypic plasticity enable an organism to tolerate environmental variation without experiencing a major reduction in fitness in different environments. However environmental variation outside of the range usually experienced by an organism can generate novel phenotypic responses, which have not been previously shaped by selection [ 5 ], and may also reflect the expression of previously hidden genetic variation. Poleward range shifts provide an interesting natural system in which to study this phenomenon. Selection for canalization of developmental traits also makes these interesting candidates for responding in a novel manner to extreme environmental cues [ 8 ]. It would be interesting to see whether increased phenotypic variance is a common pattern in other species' range expansions, and whether there are underlying correlated genetic effects similar to those seen in wing patterning in the pale grass blue butterfly [ 1 ]. The unpredictability of the effects on fitness of such phenotypic change makes it hard to predict how such novel variability impacts on population persistence. Additional empirical examples, such as that presented by Otaki et al. [ 1 ], are necessary to explore the importance and generality of this interesting phenomenon, as is a more detailed understanding of the fundamental links between traits that lie within gene and developmental networks.
Authors' contributions
JB, JRB and AP wrote the commentary. All authors read and approved the final manuscript. | CC BY | no | 2022-01-12 15:21:37 | BMC Evol Biol. 2010 Dec 9; 10:382 | oa_package/11/8e/PMC3014935.tar.gz |
||||||
PMC3014936 | 21129225 | Background
The population of children with spina bifida surviving into adulthood in rural Kenya is growing because of improved health education, care as well as an increasingly supportive environment [ 1 ]. Improved survival and integration into such social structures as schooling, work, marriage and child-bearing places significant demands on this population: the need for a lifestyle that is protective/preventive against the development of such life-threatening complications as renal failure and pressure ulcers, amongst others. Prevention requires active bladder and bowel care, as well as regular shifting of position to avoid prolonged pressure leading to the development of pressure ulcers. Failure to adhere to this 'protective lifestyle' almost invariably leads to the development of pressure ulcers; these ulcers may heal with appropriate care. Others may suffer either frequent ulcer relapses or chronic non-healing ulcers that may degenerate into Marjolin's ulcers. A number of hypotheses have been proposed to explain malignant degeneration of chronic wounds and scar tissue (Table 1 ) [ 2 - 16 ].
Four clinical signs have been proposed as characteristic for malignant pressure ulcer degeneration: the appearance of a mass, new onset of pain, a change in drainage odor and change in volume, character or appearance of drainage [ 17 ]. Unfortunately, most spina bifida patients lack sensation, and they and their caretakers may not recognize any significant changes in their ulcers. Health education, with an emphasis on ulcer prevention and care, should be taught to healthcare workers and parent(s)/guardian(s); it is ulcers that develop in childhood that may later degenerate into malignancy [ 18 ].
Our understanding of the process of pressure ulcer development amongst spina bifida patients, and their subsequent degeneration into malignant ulcers is limited. The purpose of this study was to collect and review the various theories on Marjolin's ulcers, the different prognostic factors, with a view to applying these to spina bifida patients. This understanding would aid the healthcare worker in developing programs suited to a growing population of spina bifida patients, especially in the low income countries. The author also sought to describe atypical clinical presentation of Marjolin's ulcers in these patients. | Patients and methods
A chart review of two young adults with spina bifida who had presented to the author's hospital between 2004 and August 2010 with chronic pressure ulcers found to be squamous cell carcinomas on histopathological examination was performed.
An internet/Medline/PubMed search of English literature for pressure ulcer theories as well as on the prognostic features of Marjolin's ulcers was performed. The terms 'pressure ulcer', 'pressure sore', 'decubitus ulcer' independently and with the term 'theory' or 'theories' were used, as were the terms, 'Marjolin's ulcers', 'malignant pressure ulcers', 'prognosis', 'prognostic features', in various combinations. | Results
The two patients, both females, were aged 20 and 26 years. While one of the patients was ambulant with bilateral below-knee prostheses [ 1 ], the other was wheelchair-bound. Both had chronic pressure ulcers; one had lasted 16 years, while the second patient had had the ulcer for five years, with a previous history of ulcers from the same site that had recurred a number of times in the past, with none having lasted for more than a year. The ulcer of one patient was deep, while the other was a shallow flat ulcer: both had a foul smelling purulent discharge and multiple sinuses that communicated with the ulcer. The areas with the ulcers and the sinuses were indurated, and on digital pressure exuded discharge both from the ulcer and sinuses. The margins of the ulcers were of normal appearance, (not elevated), and would thus not suggest malignancy to the casual observer (Figure 1 and 2 ). The excised surgical margins on both patients were clear of tumor. There was no evidence of underlying chronic osteomyelitis.
The internet/Medline/PubMed search on pressure ulcer theories revealed a total of nine different hypotheses (Table 1 ) [ 2 - 16 ], while a search for prognostic features of Marjolin's ulcers revealed seven clinical and four histological features (Table 2 ) [ 19 - 24 ]. | Discussion
A review of theories on Marjolin's ulcer evolution reveals that no single theory explains their evolution fully. These postulates include the toxin, the chronic irritation, the traumatic epithelial elements implantation, the co-carcinogen and the initiation and promotion theory; these theories include trauma as an integral part of the process of the evolution of Marjolin's ulcers [ 2 - 9 ]. The immunologically privileged site theory, which has a large number of proponents, attempts to explain the poor prognosis of Marjolin's ulcers [ 10 , 11 ]. The hereditary and ultraviolet rays' theories were proposed after genetic changes were found in patients with Marjolin's ulcers [ 12 - 15 ]. The environmental and genetic interaction theory seeks to explain the evolution of acute Marjolin's ulcers [ 16 ]. A combination of theories better explains the process: for example, the chronic irritation, the initiation and promotion, the toxin and the co-carcinogen theories when combined together, explain the evolution of pressure ulcer carcinomas, under which spina bifida pressure ulcers fall. The current author proposes the multifactorial theory, a combination of any of the current theories (Table 1 ) [ 2 - 16 ], as the one that best explains this process. It is to be noted that some of these theories may overlap.
Marjolin's ulcers complicating pressure ulcers in spina bifida patients are rarely reported: there are less than ten reported cases in English literature [ 1 ]. Marjolin's ulcers in general, develop in younger patients amongst sub-Saharan patients than those reported from other regions [ 18 ]; therefore, patients presenting with pressure ulcers should be investigated during the initial evaluation for this possibility. Additionally, at surgery, all the excised tissue should be submitted for histopathological investigation. Unfortunately, surgical margins clear of malignancy do not necessarily improve the prognosis of pressure ulcer carcinomas [ 1 , 18 ], which have a much poorer prognosis than Marjolin's ulcers arising from other sources [ 4 ]. Table 2 highlights prognostic features of Marjolin's ulcers in general - it is notable that a pressure ulcer carcinoma is a poor prognostic indicator. Further, Marjolin's ulcers located on the lower limbs or trunk, those with diameters above two centimeters, and latency of five years or more, all common features in the two spina bifida patients presented here, made their prognosis even poorer, especially in an environment with limited resources and options [ 1 , 3 , 11 , 19 - 24 ].
Marjolin's ulcers are characteristically either grossly flat, indurated, infiltrative shallow ulcers with well-defined, elevated margins, or exophytic proliferative ulcers [ 1 ]. The two ulcers in this report had a benign appearance of both the ulcer edges and the bases, and except for a foul smell, none of the other four hallmark signs of pressure ulcer carcinoma [ 17 ] were found. The other common features in these two ulcers were: induration and multiple sinuses communicating with the ulcers, two signs that have not been previously noted in pressure ulcer carcinomas. Pressure ulcer malignancy in spina bifida patients may thus not present with the classical descriptions, and whereas the current rarity of Marjolin's ulcers in spina bifida patients may be partially explained by the fact that not many spina bifida patients have survived long enough to develop this complication in the past, these peculiar presentations of the Marjolin's ulcers is more difficult to explain. The extent to which the congenital immobility, incontinence and lack of sensation, (factors that predispose to pressure ulcer development in both spinal cord injured patients and those with spina bifida), differs from the same factors when these develop secondary to trauma or tumors, is difficult to determine, but may be another variable that could explain the low incidence of pressure ulcer malignancy in spina bifida patients.
It is conceivable that our environment will see more such survivors, and lack of preparedness for prevention of pressure ulcers may lead to increased numbers with Marjolin's ulcers. Prevention is better that cure, more so when the cure is not possible, especially in an environment such as rural Kenya. All chronic ulcers should undergo multiple biopsies, to help define their therapy, and to avoid missing malignant ulcers [ 1 , 18 ]. | Conclusion
The multifactorial theory best explains the malignant degeneration of pressure ulcers, independent of the cause. Appropriate Marjolin's ulcer patient prognostication should aid in clinical decision making, especially the utilization of resources in poor income countries.
There is need for spina bifida patients and their guardians/caretakers to receive a close follow-up throughout life; health education focused on pressure ulcer prevention as well as early treatment of pressure ulcers when they occur, will avert the development of Marjolin's ulcers, and save lives. | Background
Due to improved care, more and more children born with spina bifida in rural Kenya are surviving into adulthood. This improved survival has led to significant challenges in their lifestyles, especially the need to ensure pressure ulcer prevention and treatment. Malignant degeneration of pressure ulcers in spina bifida patients is very rare. The author describes the clinical presentation of two pressure ulcer carcinomas that are at variance from classical descriptions.
Materials and methods
An internet/Medline/PubMed search of English literature for theories on Marjolin's ulcer evolution and prognostic features of Marjolin's ulcers was performed.
A chart review of two young adults with spina bifida who had presented to the author's hospital between 2004 and August 2010 with chronic pressure ulcers found to be Marjolin's ulcers on histo-pathological examination was performed, and the clinical features are reported.
Results
The two ulcers appeared clinically benign: one was a deep ulcer, while the other was shallow; both had normal, benign-appearing edges, and a foul smelling discharge. The two ulcers were surrounded by induration and multiple communicating sinuses, with no evidence of chronic osteomyelitis. The internet search revealed a total of nine theories on Marjolin's ulcer development, as well as seven clinical and four histological prognostic features.
Discussion
The multifactorial theory, a coalescence of a number of proposed theories, best explains the evolution of Marjolin's ulcers. Poor prognostic features include pressure ulcer carcinomas, lesions and location in the lower limbs/trunks, all present in the two patients making their prognosis dim: this is despite the surgical margins being clear of tumor. Benign appearance, induration and presence of multiple communicating sinuses are features that have not been previously described as presenting features of pressure ulcers carcinomas.
Conclusion
There is need for spina bifida patients and their guardians/caretakers to receive a close follow-up throughout life; health education focused on pressure ulcer prevention as well as early treatment of pressure ulcers when they occur, will avert the development of Marjolin's ulcers, and save lives. | Competing interests
The author declares he has no competing interests. No grants were given for this work, and no financial benefits are expected from this work. This paper has not been presented in any form, in any forum. There is no association between the author with any commercial firm, and no grants were granted for this article. There are no competing interests in the publication of this article.
Consent statement
Publication of these cases without patients consent was exempted by the AIC Kijabe hospital ethics committee as the patients consent for publication could not be obtained. | CC BY | no | 2022-01-12 15:21:37 | World J Surg Oncol. 2010 Dec 5; 8:108 | oa_package/b5/44/PMC3014936.tar.gz |
|
PMC3014937 | 21167074 | Background
Although survival of patients with gastric cancer after surgery has been improved by early detection and curative surgery, the prognosis of patients with highly advanced gastric cancer, especially with distant metastasis such as peritoneal dissemination or hematogenous metastasis, is usually very poor. Chemotherapy is the treatment of choice for metastatic advanced gastric cancer; however, a standard treatment regimen has not been established. Neoadjuvant chemotherapy for advanced gastric cancer with or without distant metastasis has been reported [ 1 - 4 ]. For patients with peritoneal dissemination, intraperitoneal chemotherapy is an additional treatment option [ 1 , 5 ]; however, only a few long-term survivors have been reported after intraperitoneal chemotherapy.
Gastric outlet obstruction (GOO) caused by unresectable antral gastric cancer is another challenging aspect of patient care. Treatment of GOO is important for the patient with unresectable gastric cancer who needs chemotherapy. There have been no reports involving patients with gastric cancer and GOO and several incurable factors curatively treated by multimodal treatments.
In this report, we describe our experience of a patient with unresectable gastric cancer with GOO and multiple metastases who achieved a histologic complete response (CR) to intraperitoneal and FOLFOX chemotherapy after palliative laparoscopic gastrojejunostomy. | Discussion
The prognosis of patients with highly advanced gastric cancer with distant metastasis, such as peritoneal dissemination or hematogenous metastasis, is usually very poor. When peritoneal dissemination is present, curative surgery cannot be achieved. However, aggressive treatment, including cytoreductive surgery with peritonectomy and intraperitoneal chemotherapy, have been reported for such patients [ 8 ]. Recently, Yonemura et al.[ 5 ] developed a new multimodal treatment referred to as bidirectional chemotherapy (neoadjuvant intraperitoneal-systemic chemotherapy protocol) for the treatment of peritoneal carcinomatosis [ 5 ]. They reported a 50% complete cytoreduction rate by cytoreductive surgery after neoadjuvant intraperitoneal and systemic chemotherapy. The median survival time of all 51 patients in that study was 14.4 months.
For bidirectional chemotherapy, various chemotherapeutic agents have been used for intraperitoneal and systemic chemotherapy [ 1 , 5 ]. For intraperitoneal chemotherapy in the present report, 40 mg of docetaxel and 150 mg of carboplatin were introduced, as described by Yonemura et al.[ 1 ].
For metastatic advanced gastric cancer, S-1 plus cisplatin was introduced as a standard treatment in Japan based on the randomized controlled trial [ 9 ]. Oxaliplatin has powerful anti-neoplasm activity, an intriguing alternative to cisplatin with at least comparable activity, a synergistic effect with 5-FU, and a satisfactory safety profile. The combination of oxaliplatin with 5-FU and leucovorin (FOLFOX regimen) was selected for systemic chemotherapy because it has a favorable activity as first-line therapy with locally advanced and metastatic gastric cancer or second-line treatment in advanced or metastatic gastric cancer patients, and may be considered a viable treatment alternative. This regimen was suggested to be active with a 40%-55% objective response rate in patients with gastric cancer [ 10 - 12 ]. However, this regimen has rarely been reported as a neoadjuvant chemotherapy agent in the treatment of advanced gastric cancer [ 13 ].
Neoadjuvant chemotherapeutic agents in the previous studies included S-1 [ 4 , 5 ], S-1 plus cisplatin [ 3 ], methotrexate plus 5-fluorouracil [ 1 ], paclitaxel plus doxifluridine [ 2 ], EEP (etoposide, epirubicin, and cisplatin) [ 14 ], and cisplatin plus 5-fluorouracil with leucovorin [ 15 ]. The FOLFOX regimen has usually been used for palliative chemotherapy for patients with metastatic gastric cancer. We used the FOLFOX regimen as neoadjuvant chemotherapy, which is widely used for metastatic advanced gastric cancer after curative or palliative resection or for unresectable gastric cancer in our institute.
In the present case, the patient had several incurable factors, including peritoneal carcinomatosis, hepatic metastasis, and locally advanced tumor that induced GOO. Therefore, the first aim of treatment was palliation of the GOO symptoms. We performed laparoscopic bypass surgery first, and at that time, an intraperitoneal port was implanted for the following intraperitoneal chemotherapy.
GOO is a challenging problem in patients with advanced gastric cancer in the distal part of the stomach. The presence of GOO is an independent prognostic factor, even after radical surgery [ 16 ]. Additionally, for systemic chemotherapy, GOO is a problem that should be treated first because adequate oral intake is essential for systemic chemotherapy. For GOO, various treatment options, such as endoscopic stenting, palliative bypass surgery (open or laparoscopic), or palliative resection, can be chosen. Resection is theoretically the most effective treatment option for GOO by achieving intestinal continuity and a reductive therapeutic effect. In the present case, however, other incurable factors existed, such as hepatic metastasis and peritoneal carcinomatosis. Furthermore, palliative bypass surgery (gastrojejunostomy) is thought to provide better long-term results compared to endoscopic stent placement, and is therefore the treatment of choice in patients with a life expectancy of > 2 months [ 17 ]. Thus, we performed laparoscopic palliative gastric bypass surgery rather than palliative resection or stent placement.
If curative resection is not possible or not effective, and response to chemotherapy is expected, laparoscopic palliative procedure is an effective and safe procedure option for the patient with GOO because this palliative procedure is minimally invasive and enable the patient to receive early post-operative chemotherapy. In addition, the port for the intraperitoneal chemotherapy can be placed during this procedure. Furthermore, laparoscopic bypass surgery can prevent post-operative adhesions and enable easy, definitive surgery for the patient who has a response to the chemotherapy.
For the second operation after CR was suspected by image study and endoscopic biopsy, we performed radical subtotal gastrectomy with extended lymph node dissection. The aim of the surgical resection was to confirm the CR by pathologic examination and to provide the therapeutic effect of resection if the residual cancer cells were present. However, hepatic resection was not performed because the initial hepatic metastasis was multiple and bilobar, and the location of hepatic metastasis could not be identified at the time of the second operation.
According to previous reports regarding neoadjuvant chemotherapy for advanced gastric cancer, aggressive treatment for peritoneal dissemination has been limited to patients with peritoneal carcinomatosis alone, rather than hematogenous metastasis, such as hepatic metastasis or distant lymph node metastasis [ 1 , 5 , 8 , 13 ]. D'Ugo et al. [ 14 ] reported neoadjuvant chemotherapy in patients with resectable gastric cancer. However, in the present case, a CR was observed in spite of the fact that he had three individual incurable factors; peritoneal dissemination, hepatic metastasis, and GOO caused by locally advanced primary tumor. The reason for the good response of our case may be related to the following: 1) the size of the multiple hepatic metastases was small; 2) peritoneal dissemination was limited to the omentum and not disseminated to the distant peritoneum; and 3) the tumor size was relatively small in spite of GOO, and there was no adjacent organ invasion (pancreas or colon). Thus, to evaluate the effectiveness of chemotherapy following bypass surgery for unresectable gastric cancer with GOO, a large randomized controlled trial is needed. | Conclusion
Combination chemotherapy with systemic and intraperitoneal chemotherapy following laparoscopic bypass surgery showed marked efficacy in the treatment for unresectable advanced gastric cancer with GOO. | Background
Gastric outlet obstruction (GOO) caused by unresectable gastric cancer is a challenging aspect of patient care. There have been no reports involving patients with obstructing gastric cancer and several incurable factors curatively treated by multimodal treatments.
Case presentation
We report a case of 55-year-old man who was diagnosed with a poorly differentiated adenocarcinoma in the pre-pyloric antrum with GOO by gastroscopy. An abdominal computed tomography (CT) scan revealed thickening of the gastric wall and adjacent fat infiltration, and a large amount of food in the stomach suggesting a passage disturbance, enlarged lymph nodes along the common hepatic and left gastric arteries, and multiple hepatic metastases. The serum carcinoembryonic antigen (CEA) level was 343 ng/ml and the carbohydrate antigen (CA) 19-9 level was within normal limits. The patient underwent a laparoscopic gastrojejunostomy for palliation of the GOO. On the 3 rd and 12 th days after surgery, he received intraperitoneal chemotherapy with 40 mg of docetaxel and 150 mg of carboplatin. Simultaneously, combined chemotherapy with 85 mg/m 2 of oxaliplatin for the 1 st day and 600 mg/m 2 of 5-FU for 2 days (FOLFOX regimen) was administered from the 8 th post-operative day. After completion of nine courses of FOLFOX, the patient achieved a complete response (CR) with complete disappearance of the primary tumor and the metastatic foci. He underwent a radical subtotal gastrectomy with D3 lymph node dissection 4 months after the initial palliative surgery. The pathologic results revealed no residual primary tumor and no lymph node metastasis in 43 dissected lymph nodes. He has maintained a CR for 18 months since the last operation.
Conclusion
Combination chemotherapy with systemic and intraperitoneal chemotherapy following laparoscopic bypass surgery showed marked efficacy in the treatment for unresectable advanced gastric cancer with GOO. | Case report
A 55-year old man visited our hospital for evaluation of epigastric pain and poor oral intake of 2 months duration. A gastroscopy demonstrated a deep ulcerative lesion in the lesser curvature side of the pre-pyloric antrum and a large amount of food retained in the stomach because of GOO (Figure 1A ). The biopsy revealed poorly differentiated adenocarcinoma.
An abdominal computed tomography (CT) scan revealed thickening of the gastric wall, adjacent fat infiltration, and a large amount of food in the stomach, suggesting passage disturbance. The CT scan also showed enlarged lymph nodes along the common hepatic and left gastric arteries, multiple enhancing omental masses, nodular peritoneal thickening suggestive of peritoneal carcinomatosis, and multiple hepatic metastases (Figure 2A ).
The serum carcinoembryonic antigen (CEA) level was 343 ng/ml and the carbohydrate antigen (CA) 19-9 level was within normal limits. The patient was diagnosed with unresectable gastric cancer with GOO and several incurable factors. The clinical stage was stage IV (cT4N2M1). The Eastern Cooperative Oncology Group (ECOG) performance status was grade 1 [ 6 ].
The patient underwent a laparoscopic gastrojejunostomy for palliation of the GOO. The peritoneal drainage catheter was placed during the laparoscopic procedure. During laparoscopic inspection, there were several metastatic nodules on the omentum. One of the nodules was biopsied and confirmed to be metastatic adenocarcinoma. On the 3 rd and 12 th days after surgery, for intraperitoneal chemotherapy, 40 mg of docetaxel and 150 mg of carboplatin were introduced over 2 hours in 1000ml of saline. Hyperthermia was not used in the intraperitoneal chemotherapy. Simultaneously, he received 85 mg/m 2 of oxaliplatin on the 1 st day as a 2-hour infusion followed by 600 mg/m 2 of 5-fluorouracil (FU) as a 22-hour infusion for 2 days (FOLFOX regimen) from the 8 th post-operative day repeated every 2 weeks.
After six courses, a CT scan revealed marked reduction in the size of the hepatic metastases, and the tumor markers returned to normal levels. After the completion of nine courses of chemotherapy, the gastric wall thickening, metastatic foci in the liver, omental infiltration of the metastatic nodules, and enlarged lymph nodes noted on CT scan resolved (Figure 2B ). Positron emission tomography (PET) scan showed no metabolic evidence of malignancy. A gastroscopy showed an ulcerated scar in the antrum and the GOO and passage disturbance had also resolved (Figure 1B ).
No malignant cells were detected on the endoscopic biopsy specimen, and a CR was determined according to the response evaluation criteria in solid tumors (RECIST) guideline[ 7 ].
He underwent a radical subtotal gastrectomy with a D3 lymph node dissection 4 months after the initial palliative surgery. The pathologic results showed no residual primary tumor and no lymph node metastasis of 43 dissected lymph nodes. There was no peritoneal metastatic focus in the surgical specimen including omentum and visceral peritoneum. Postoperatively, he received oral chemotherapy with daily 600 mg of doxifluridine for 12 months. He maintained a CR for 18 months after the last operation.
Consent
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.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
JP and KC equally contributed to this study and were responsible for treatment of the case and writing the manuscript. All authors read and approved the final manuscript. | CC BY | no | 2022-01-12 15:21:37 | World J Surg Oncol. 2010 Dec 20; 8:109 | oa_package/d5/4c/PMC3014937.tar.gz |
|||
PMC3014938 | 21176192 | Background
Goblet cell carcinoid (GCC) of the large intestine is a rare neoplasm, usually located in ascending colon and rectum. Histologically, it is similar to goblet cell carcinoid of the appendix [ 1 ]. GCC has both endocrine and glandular differentiation. Dual differentiation probably arises from a pluripotent intestinal stem cell instead of two different mature cells. The mean age for diagnosing GCC of the appendix is 58.89 years with equal representation in both genders. Regional and systemic metastasis is common at initial diagnosis. These tumors perform aggressive behavior with tendency for metastasis and wide local dissemination [ 2 ]. Lesions are treated according to the same conventional oncologic approach to adenocarcinoma [ 3 ]. We present here, a 60 year-old male patient, who diagnose as mixed carcinoid-adenocarcinoma located in transverse colon and at gastroenterostomy site. | Discussion
Since more than 30 years ago, a new variant type of epithelial tumor of appendix has been recognized. This tumor is reported under different names including goblet cell carcinoid (GCC), adenocarcinoid, mucinous carcinoid, intermediate type of carcinoid, crypt cell carcinoma, amphicrine (endo-exocrine) neoplasia, composite tumor and microglandular carcinoma. All names except GCC have been omitted from the current World Health Organization (WHO) classification [ 2 ]. Subbuswamy et al described the first report of GCC in 1974 [ 4 ]. The histology and biology appears to be intermediate between carcinoid tumors and adenocarcinomas. This tumor appears to combine features of epithelial and carcinoid neoplasms and in addition the surface mucosal epithelium is not neoplastic. Histopathological features such as increased number of Paneth cells, increased amount of mucin secretion and presence of pancreatic polypeptide may predict a more aggressive behavior [ 2 ].
Mucocarcinoids also called mucinous or adenocarcinoids, show a quite different histological appearance from carcinoids and endocrine cell carcinomas. The tumor is composed predominantly of small clumps, strands, or glandular collections of mucin-producing cells looking like goblet cells or signet-ring cells, and intermingled with endocrine cells in a variable number and occasionally with Paneth's cells. The admixed endocrine cells comprise a variety of cell types, such as somatostatin-containing D cells, serotonin-containing endocrine cells and enterochromaffin-like cells containing histamine. They are often sparse and, in about 10% of cases, difficult to find. The tumor was originally considered to be a variant of a carcinoid. The frequent paucity of endocrine cells and more aggressive clinical nature are not consistent with such speculation. Mucocarcinoids are a variant of adenocarcinomas showing differentiation to both mucin-producing cells and endocrine cells. They occur most frequently in the appendix, but rarely in stomach. Ito et al reported only one case treated in 10 years period [ 5 ].
Even if there are many questions about histogenesis of tumors with mixed differentiation, it is hypothesized that these neoplastic lesions may probably arise from a single pluripotent stem cell as well as different mature cells [ 6 ]. Histologically, these tumors are divided into three subtypes: mixed (composite) tumors, collision tumors and amphicrine tumors. In mixed tumors, the two elements typically merge and intermingle, and in some areas transitions can be seen, such that the two components can be difficult to distinguish. A carcinoid component should compose at least one third of the tumor cell population in a composite tumor. In collision tumors, the two elements should be in intimate contact without intermixture of individual cell types. Amphicrine tumors differ from the above tumor types in that endocrine and nonendocrine epithelial cell constituents are present within the same cell [ 7 ]. These tumors behave more like adenocarcinomas than carcinoids. Two cases with mixed carcinoid-adenocarcinoma, for the first time, were reported by Moyana et al in 1988 [ 8 ]. | Conclusions
This case with the characteristics of mixed carcinoid-adenocarcinoma with carcinoid component, predominantly composed of goblet cells, is reported because of its rarity and points out that more data should be collected to develop our knowledge about diagnosis, histopathological and clinical features, prognosis, and conventional treatment of this neoplasm. | Goblet cell carcinoid of the large intestine is a rare neoplasm, usually located in ascending colon and rectum. A 60-year-old male patient underwent surgery after the diagnosis of acute abdomen. Exploratory laparotomy revealed perforation with a diameter of 1 cm at the site of the previously performed gastroenterostomy and dilatation of the right colic flexure, secondary to a solid obstructive mass located in the mid-portion of transverse colon. Histopathological investigation of the biopsies, taken from the gastroenterostomy site and the tumor, revealed mixed carcinoid-adenocarcinoma with carcinoid component, predominantly composed of goblet cells. Three cycles of FOLFOX-4 protocol was administered. Following respiratory distress secondary to pulmonary metastasis, the patient's condition deteriorated and subsequently died in the fourth postoperative month. Our aim with this paper is to point out that more cases should be reported for more effective diagnosis, histopathological study, clinical investigation, treatment and prognosis of this specific neoplasm. | Case
A 60 year-old male patient presented with complains of nausea, vomiting, abdominal distension, and no discharge for three days. He also had intermittent cramping abdominal pain, mainly located in the upper left abdominal quadrant. He had a history of prior gastric surgery, performed 26 years ago, for peptic ulcer disease. His vital signs included temperature of 36.4°C, blood pressure of 100/80 mmHg, pulse rate of 60 beats/min, respiratory rate of 22 breaths/min. On physical examination, the scar of the midline incision was inspected and the abdomen was distended and tender to palpation with guarding. Routine hematological and biochemical investigations were within normal limits except for raised total leucocytes count (32,000/mm3). Serum carcinoembryonic antigen (CEA) and cancer antigen (CA) 19-9 levels were not elevated on the postoperative 3rd day of the follow-up. Plain X-ray of abdomen revealed few fluid levels and free gas in subphrenic spaces whereas the abdominal ultrasonography showed no finding but diffuse intestinal gas. The patient underwent surgery after the diagnosis of acute abdomen was made. Exploratory laparotomy revealed perforation with a diameter of 1 cm at the site of the previously performed gastroenterostomy and dilatation of the right colic flexure, secondary to a solid obstructive mass located in the mid-portion of transverse colon. There were no metastatic liver lesions whereas metastatic lymph nodes were detected in mesocolon. The gastroenterostomy was reconstructed after anastomosis and the mid segment of the transverse colon with approximately 5-6 cm margins on either side of the tumor was resected.
Histopathological investigation of the biopsies, taken from the gastroenterostomy site and the tumor, revealed mixed carcinoid-adenocarcinoma with carcinoid component, predominantly composed of goblet cells. Ulcero-vegetative mass in the transverse colon with the size of 5 × 5 × 1.5 cm, infiltrating the intestinal serosa, and three tissue samples, each measuring approximately 2.5 × 1.5 × 0.3 cm, taken from the gastroenterostomy site were microscopically similar and had the characteristics of mixed carcinoid-adenocarcinoma with carcinoid component, predominantly composed of goblet cells (Figure 1 ). Tumor invasion in all layers of the transverse colon and the gastroenterostomy site are accompanied by perforation. Immunohistochemical stains showed that neoplastic cells were positive for neuron-specific enolase (NSE), synaptophysin and E-cadherin and negative for chromogranin. Ten metastatic lymph nodes were detected in mesocolon. At three months postoperatively the needle biopsy specimen of the liver revealed metastasis.
The 24 hours urine vanillylmandelic acid (VMA) level was within normal range on the postoperative 5th week of the follow-up. In-111 octreotide scintigraphy detected increased uptake in the region of the para-aortic lymph node, compatible with a lesion which had the expression of somatostatin receptors. A bone scan was performed using 20 mCi of Tc-99 m MDP, and uncovered no evidence of abnormality. Three cycles of FOLFOX-4 protocol was administered by the medical oncology department. He was hospitalized three months after the operation because of poor health status. Ultrasonography of the liver showed an inhomogeneous echo texture, hyperechoic nodules with peripheral hypoechoic halos and the largest lesion with size of 3 × 2.8 × 2.4 cm was localized in the anterosuperior portion of the right lobe. A needle biopsy of the liver was positive for metastasis of the carcinoma. Following respiratory distress secondary to pulmonary metastasis, his health situation got worse and subsequently died in the fourth postoperative month. In additionally, the patient was questioned about any symptoms of the carcinoid syndrome, which includes flushing, diarrhea, wheezing etc. pre-operatively once the post op diagnosis was made and post-operatively. He did not encounter any symptoms of the carcinoid syndrome.
Consent
Written informed consent was obtained from the family of the deceased 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.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
Eİ performed the operation and helped manuscript preparation particularly in describe the findings and the follow-up, revised and edited most of the manuscript. ÖP performed the histopathological and immunohistochemical analyses of all surgical specimens, provided the figure of the microscopic appearance of the tumor, helped literature search, and corrected the final draft. SE helped with the editing of the manuscript and literature search. BRK and NFY were involved in preparation of initial draft and literature search. All authors read and approved the final manuscript for publication. | CC BY | no | 2022-01-12 15:21:37 | World J Surg Oncol. 2010 Dec 22; 8:110 | oa_package/4d/fd/PMC3014938.tar.gz |
|||
PMC3014939 | 21126333 | Background
From 1995 until 2007, the number of reported Legionnaires' disease (LD) cases in Finland has varied from 10 up to 31 per year, with an annual incidence of 2-6 cases per million [ 1 ]. The reporting of LD and Legionella -positive laboratory results has been mandatory for both clinicians and clinical laboratories since 1995. The incidence rate in Finland is much lower compared to the mean incidence rate of LD in European countries, which has been 11 cases per million in 2007 and 2008 [ 2 ].
Previous Finnish Legionella survey studies revealed that 30% of the hot water systems and 47% of the cooling water systems were contaminated with legionellae [ 3 , 4 ]. In addition, the few Finnish case studies where both clinical and environmental Legionella strains were obtained for molecular typing have indicated that hot water systems were the source of infection [ 5 , 6 ].
Following Swedish reports of extremely high concentrations of legionellae in biological waste water treatment plants and LD in an employee working near a plant [ 7 ], an environmental study was initiated focusing on waste water systems used by the Finnish paper and pulp industries. In the first part of the study, culturable legionellae were detected in 73% (11/15) of industrial active sludge basins containing waste water, with the highest concentration being 1.9 × 10 9 cfu/l (Unpublished data, Kusnetsov J, Torvinen E, Lehtola M and Miettinen IT). In addition, the microscopic PNA-FISH method [ 8 ] revealed Legionella pneumophila ( L. pneumophila ) cells to be present in all waste water basins (up to 1.7 × 10 10 cells/l). After these environmental findings, two cases of LD were diagnosed via the occupational health services of the participating paper and pulp mills. These cases are reported here. | Methods
Legionella antigens were detected by urinary antigen immunochromatography (Binax-now, Inc. Portland). Serum antibodies against legionellae were first detected by an in-house EIA-method (TYKSLAB, Turku) and later with an in-house IFA-method (HUSLAB, Helsinki). The EIA-method detects antibodies against L. pneumophila serogroups 1 to 4 and L. micdadei [ 9 ] and the IFA-method detects IgG-, IgA-and IgM-antibodies against L. pneumophila serogroups 1 to 8 and L. gormanii, L. longbeachae, L. dumoffii, L. bozemanii and L. micdadei [ 10 ]. The definitions given by the European Working Group for Legionella Infections were followed to determine whether these cases were confirmed or presumptive LD cases [ 11 ].
At the workplaces, the water samples were analysed before and, in more detail, after these cases were diagnosed. For culture of waste waters, samples were diluted 3-fold before processing according to ISO 11731 [ 12 ]. Clean water samples were also diluted in the same way and also concentrated by filtration. Portions of diluted, undiluted and concentrated samples were inoculated directly, acid-washed (pH 2.2, 4 min) or heat-treated (50°C, 30 min) before inoculation onto GVPC medium plates (buffered charcoal yeast extract medium containing glycine, vancomycin, polymyxin B and cycloheximide, Oxoid Ltd, Cambridge, UK). Water samples from the hot and cold water systems of the cases' homes were analysed with the standard method [ 12 ], without dilution. Media plates were incubated for 10 days at 36 ± 1°C and colonies resembling legionellae were further confirmed by growth tests according to the standard method [ 12 ].
Serotyping of Legionella strains was first performed with the Oxoid Legionella Latex Test (DR0800 M, Oxoid). The L. pneumophila strains were further serogrouped with the Denka Seiken antisera set (Denka Seiken Co. Ltd, Tokyo, Japan) or the Dresden panel of monoclonal antibodies (MAb) [ 13 ]. L. pneumophila serogroup 1 strains were further subgrouped using the Dresden monoclonal panel. Non -pneumophila Legionella strains were identified to species level by growth and biochemical tests and partial 16 S rRNA sequencing by commercial service at FIMM (FIMM, Helsinki, Finland) using primers fD1 Mod and 533r [ 14 ] and GenBank database [ 14 ].
In order to identify if Case B had been exposed via aerosols in the wind blowing from the direction of the waste water treatment plant, meteorological data for the period of his working hours were obtained from a local weather station situated 1200 meters from the waste water treatment plant. | Results of environmental study
In plant A, in February 2006, 2.0 × 10 7 cfu/l of L. pneumophila serogroup 13 was isolated from the active sludge basin, and 1.0 × 10 4 cfu/l of L. pneumophila serogroup 3 from water circulating around the plant (Table 1 ). After Case A had been diagnosed, the plant was sampled again, and high concentrations of L. pneumophila serogroup 1 were isolated from water in the active sludge basin, and from water and sludge in the post-clarification basin. Three strains of L. pneumophila serogroup 1 isolated from the post-clarification basin water and sludge were subtyped and the strains belonged to the monoclonal subgroup Bellingham.
The domestic water systems of Case A were studied on August 29, 2006. Four samples were taken from the shower (mixture of hot and cold water, 35.5°C), kitchen tap (hot water, 55.0°C), toilet tap (mixture of hot and cold water, 35.5°C) and hose used outside the building (cold water, 10.2°C). All these domestic water samples were Legionella -negative.
In plant B, the active sludge basin was first sampled in August 2005, and 3.0 × 10 7 cfu/l of L. pneumophila serogroup 5 was detected (Table 2 ). After Case B had been diagnosed, new samples from the active sludge basin contained high concentrations of Legionella rubrilucens (8.0 × 10 9 cfu/l) and repeatedly L. pneumophila serogroup 5 (4.3 × 10 7 cfu/l). Legionellae were also isolated from a well where the rejected waste water was directed. In addition, a cooling tower at the water treatment plant, used for cooling of waste water, was sampled and yielded low concentration of L. pneumophila serogroup 2 (1.7 × 10 3 cfu/l). A few samples were also taken in November inside the mill, from paper machines and the shower used by Case B, but they did not contain culturable legionellae (Table 2 ).
The domestic water systems of Case B were studied on November 23, 2006. The samples taken from the shower (mixture of hot and cold water, 37.7°C), kitchen tap (hot water, 53.4°C) and toilet tap (cold water, 11.3°C) were Legionella -negative.
Meteorological data reported that the wind had blown from south-west, west-southwest, south, south-south-west and west during the period of the latest working days and hours of Case B before the onset of symptoms. The wind speed was only from 1.1 to 4.2. m/s. This data from the local weather station, which was situated close to the waste water treatment plant B, was used to draw the inverse wind roses in Figure 1 . Case B had been working in a field situated east from the waste water treatment plant B, and was exposed to the wind blowing exactly from west to east for at least four hours, 13 days before the onset of his symptoms. | Discussion and Conclusion
This is the first report of LD cases associated with industrial waste water systems in Finland. In Case A, the positive urinary antigen test, and the antibody respose indicated that the infection was caused by L. pneumophila serogroup 1. The exposure to aerosols generated from waste water of the post-clarification basin was apparent and because L. pneumophila serogroup 1 was detected only in waste water samples, the most likely source of this Legionella infection was the post-clarification basin.
The antibody response of Case B suggested that LD was caused by L. pneumophila serogroup 2. This serogroup was isolated only from the cooling tower at the waste water treatment plant. In September 2006, the cooling towers were used occasionally, which was usual at that time of the year, meaning that a direct route via the cooling towers was possible. Water from the cooling towers flowed into the active sludge basin. Thus L. pneumophila serogroup 2 was very likely to be present in the active sludge basin, at least in low concentrations. However, it was not possible to detect by culture among the abundant growth of other Legionella strains and other microbes.
In 2007 and 2008, a total of 11897 LD cases were detected in Europe [ 2 ]. Most of these cases were diagnosed with UA test (81%) and only a few of the cases with the isolation of legionellae (8.8%). Thus it is very common that Legionella infections are diagnosed without Legionella isolates, which are needed for confirming the source. We were also unable to obtain clinical isolates in this study but were able to exclude some of the sources and focus on the most likely sources of transmission.
As Case B was working in a field about 200 m east of the active sludge basin and the cooling towers, meteorological data indicated that he was likely to have been exposed to aerosols originating from the waste water treatment plant. The incubation period of LD is generally between two and ten days. However, in the Dutch Flower Show outbreak in 1999, 16% of cases occurred after 10 days, up to 19 days [ 15 ]. In addition, in the 1976 Philadelphia outbreak incubation periods as long as 26 days were reported [ 16 ]. In Case A, the incubation time was five days. In Case B, the exact incubation period remained uncertain, as he had been working during the previous two weeks (days 4, 5, 6, 11, 12, 13) before the onset of symptoms under similar conditions. However, particularly on day 13, before the onset of symptoms, the wind was blowing for four hours from the waste water treatment plant exactly in the direction in which he was working. It is therefore possible that the incubation period exceeded 10 days, being up to 13 days in his case. After the domestic water systems and other water systems in the mill were excluded as possible sources, the waste water treatment plant remained the most likely source of the Legionella infection in Case B.
It is notable that even though very high concentrations of L. rubrilucens and L. pneumophila serogroup 5 were found in the waste water system of plant B, the antibody response indicated that he was suffering a L. pneumophila serogroup 2 infection; this serogroup was detected in much smaller concentrations at the plant. According to EWGLI data of LD cases from 1995 to 2006, 0.8% (33/4390) of the clinical Legionella isolates were L. pneumophila serogroup 2 [personal communication, Carol Joseph, 2009, EWGLI data, Health Protection Agency, London, England]. Equally rare were isolations of serogroup 5 of L. pneumophila (1.0%), in comparison to the isolations of serogroup 1 (73.8%), serogroup 3 (3.8%) and serogroup 6 (2.2%). In addition, L. rubrilucens has been known to be a cause of LD only once [ 17 ], showing that the species probably is much less virulent than all serogroups of L. pneumophila . The antibody response to L. pneumophila serogroup 2 does not rule out the possibility that the infection could have been caused by other legionellae, especially by those L. pneumophila serogroups prevailing in the waste water plant. Previously, L. pneumophila serogroup 5 has been associated with an outbreak of nosocomial legionellosis in Finland [ 6 ].
The dose of Legionella cells causing LD is not known precisely, nor how long one has to breath air with aerosols contaminated by legionellae before becoming infected. In the large outbreak in Pas-de-Calais, spending over 100 minutes outdoors daily increased the risk of LD significantly [ 18 ]. Prior to the onset of symptoms, Case B worked for at least 240 minutes (on day 13) downwind of the waste water treatment plant.
The laboratory results and environmental evidence of these two cases indicated that the LD infections were acquired at work, at or very close to the waste water treatment plants. No other previous or simultaneous Legionella infections have been known to occur among employees working at plants A or B, even though equally high Legionella concentrations most probably have existed in these waste water systems for years. The employees have worn respirators as protection against legionellae since 2005, and this may have prevented infections. Another explanation for the low number of cases could be underdiagnosis, since industrial waste water systems have been associated with abundant Legionella growth only since the Pas-de-Calais outbreak in 2003-2004 [ 18 ]. Therefore these waste water systems have not been investigated as possible sources of Legionella infections. Furthermore, the monoclonal subtype Bellingham of L. pneumophila serogroup 1 found in the plant A, does not belong to MAb 3/1 subgroup, which is the most virulent subtype of serogroup 1 [ 19 ], and L. pneumophila serogroups 2 and 5 rarely cause LD.
The two cases reported here have similarities to the previous Swedish case, where an employee of a paper and pulp mill most likely acquired infection while working 100 meters from a waste water treatment plant [ 7 ]. In that case, the clinical Legionella strain ( L. pneumphila serogroup 1 subtype Benidorm, MAb 3/1 positive) and the environmental strain from the waste water basin were identical by molecular typing. In addition, an outbreak of five cases of Pontiac fever occurred after exposure to aerosols from sludge in a sewage treatment plant of the Danish food industry [ 20 ]. The strain isolated from sludge in concentrations of 1.5 × 10 7 cfu/g was L. pneumophila serogroup 1, subgroup OLDA/Oxford (MAb 3/1 negative). In this outbreak, the workers used respirators, but the filters were effective only against chemical substances.
Previously, a large community outbreak with 86 cases of LD was associated with cooling towers and waste water basins of a petrochemical plant occurred in France in 2003-2004 [ 18 ]. The strain causing the outbreak was L. pneumophila serogroup 1, strain Lens. The aerosols with legionellae, spread by a cooling tower, were infectious at a distance of at least six kilometers. The plant was later closed. In Norway, an air-scrubber was spreading Legionella aerosols over a distance of ten kilometers, resulting in 103 LD cases and ten deaths [ 21 , 22 ]. It is assumed that aerosols containing Legionella from the waste water aeration ponds originally contaminated the air scrubber. The scrubber was cleaned of legionellae and the aeration pond changed to an anaerobe pond. Thus, the sources of these larger outbreaks have been limited.
In contrast, Swedish and Finnish studies have indicated that heavy contamination of active sludge basins with legionellae is very common [ 7 ], (Unpublished data, Kusnetsov J, Torvinen E, Lehtola M and Miettinen IT). Further, air samplings in Norway and France in the vicinity of active sludge basins revealed that viable Legionella cells can be isolated up to 180-270 meters downwind [ 23 , 24 ]. It seems therefore likely that any waste water treatment plant with an active sludge basin under aeration can contain higher concentrations of Legionella bacteria and also produce aerosols with legionellae. Evidence of exposure to legionellae can be established if an increased frequency of elevated Legionella antibodies in serum samples can be detected [ 22 ].
In addition, some of these waste water treatment plants use cooling towers to lower the waste water temperature. It would be very useful to know if these waste water cooling towers are clean and maintained in accordance with the European and WHO Legionella guidelines [ 11 , 25 ]. In the European guidelines, a concentration of 1000 cfu/l of legionellae is recommended as the highest acceptable concentration which can be present in cooling water (technical guidelines).
Water treatment plants with active sludge basins should be considered as a possible source of community acquired Legionella infections, directly or indirectly via cooling towers. In addition, the employees should protect themselves by using respirators at or in the vicinity of water treatment plants. The Finnish Work safety act (738/2002) [ 26 ] has stated that employees should be protected against biological factors, including Legionella bacteria, at waste water treatment plants. Especially in industrial waste water treatment plants with high water temperatures, Legionella concentrations may be very high. In 2005, Finnish forest industry employees were instructed to use respirators while working in the vicinity of the water treatment plants. Developing ways to lower Legionella concentrations in these water systems would be the next step in diminishing the risk of Legionella infection.
These LD cases might have remained undiagnosed if our environmental study had not increased awareness about potential Legionella exposure in waste water treatment plants. These findings suggest that the clinicians should consider LD when treating patients with pneumonia from these industrial settings.
Consent
Written informed consents were obtained from the Cases for publication of this report. | Discussion and Conclusion
This is the first report of LD cases associated with industrial waste water systems in Finland. In Case A, the positive urinary antigen test, and the antibody respose indicated that the infection was caused by L. pneumophila serogroup 1. The exposure to aerosols generated from waste water of the post-clarification basin was apparent and because L. pneumophila serogroup 1 was detected only in waste water samples, the most likely source of this Legionella infection was the post-clarification basin.
The antibody response of Case B suggested that LD was caused by L. pneumophila serogroup 2. This serogroup was isolated only from the cooling tower at the waste water treatment plant. In September 2006, the cooling towers were used occasionally, which was usual at that time of the year, meaning that a direct route via the cooling towers was possible. Water from the cooling towers flowed into the active sludge basin. Thus L. pneumophila serogroup 2 was very likely to be present in the active sludge basin, at least in low concentrations. However, it was not possible to detect by culture among the abundant growth of other Legionella strains and other microbes.
In 2007 and 2008, a total of 11897 LD cases were detected in Europe [ 2 ]. Most of these cases were diagnosed with UA test (81%) and only a few of the cases with the isolation of legionellae (8.8%). Thus it is very common that Legionella infections are diagnosed without Legionella isolates, which are needed for confirming the source. We were also unable to obtain clinical isolates in this study but were able to exclude some of the sources and focus on the most likely sources of transmission.
As Case B was working in a field about 200 m east of the active sludge basin and the cooling towers, meteorological data indicated that he was likely to have been exposed to aerosols originating from the waste water treatment plant. The incubation period of LD is generally between two and ten days. However, in the Dutch Flower Show outbreak in 1999, 16% of cases occurred after 10 days, up to 19 days [ 15 ]. In addition, in the 1976 Philadelphia outbreak incubation periods as long as 26 days were reported [ 16 ]. In Case A, the incubation time was five days. In Case B, the exact incubation period remained uncertain, as he had been working during the previous two weeks (days 4, 5, 6, 11, 12, 13) before the onset of symptoms under similar conditions. However, particularly on day 13, before the onset of symptoms, the wind was blowing for four hours from the waste water treatment plant exactly in the direction in which he was working. It is therefore possible that the incubation period exceeded 10 days, being up to 13 days in his case. After the domestic water systems and other water systems in the mill were excluded as possible sources, the waste water treatment plant remained the most likely source of the Legionella infection in Case B.
It is notable that even though very high concentrations of L. rubrilucens and L. pneumophila serogroup 5 were found in the waste water system of plant B, the antibody response indicated that he was suffering a L. pneumophila serogroup 2 infection; this serogroup was detected in much smaller concentrations at the plant. According to EWGLI data of LD cases from 1995 to 2006, 0.8% (33/4390) of the clinical Legionella isolates were L. pneumophila serogroup 2 [personal communication, Carol Joseph, 2009, EWGLI data, Health Protection Agency, London, England]. Equally rare were isolations of serogroup 5 of L. pneumophila (1.0%), in comparison to the isolations of serogroup 1 (73.8%), serogroup 3 (3.8%) and serogroup 6 (2.2%). In addition, L. rubrilucens has been known to be a cause of LD only once [ 17 ], showing that the species probably is much less virulent than all serogroups of L. pneumophila . The antibody response to L. pneumophila serogroup 2 does not rule out the possibility that the infection could have been caused by other legionellae, especially by those L. pneumophila serogroups prevailing in the waste water plant. Previously, L. pneumophila serogroup 5 has been associated with an outbreak of nosocomial legionellosis in Finland [ 6 ].
The dose of Legionella cells causing LD is not known precisely, nor how long one has to breath air with aerosols contaminated by legionellae before becoming infected. In the large outbreak in Pas-de-Calais, spending over 100 minutes outdoors daily increased the risk of LD significantly [ 18 ]. Prior to the onset of symptoms, Case B worked for at least 240 minutes (on day 13) downwind of the waste water treatment plant.
The laboratory results and environmental evidence of these two cases indicated that the LD infections were acquired at work, at or very close to the waste water treatment plants. No other previous or simultaneous Legionella infections have been known to occur among employees working at plants A or B, even though equally high Legionella concentrations most probably have existed in these waste water systems for years. The employees have worn respirators as protection against legionellae since 2005, and this may have prevented infections. Another explanation for the low number of cases could be underdiagnosis, since industrial waste water systems have been associated with abundant Legionella growth only since the Pas-de-Calais outbreak in 2003-2004 [ 18 ]. Therefore these waste water systems have not been investigated as possible sources of Legionella infections. Furthermore, the monoclonal subtype Bellingham of L. pneumophila serogroup 1 found in the plant A, does not belong to MAb 3/1 subgroup, which is the most virulent subtype of serogroup 1 [ 19 ], and L. pneumophila serogroups 2 and 5 rarely cause LD.
The two cases reported here have similarities to the previous Swedish case, where an employee of a paper and pulp mill most likely acquired infection while working 100 meters from a waste water treatment plant [ 7 ]. In that case, the clinical Legionella strain ( L. pneumphila serogroup 1 subtype Benidorm, MAb 3/1 positive) and the environmental strain from the waste water basin were identical by molecular typing. In addition, an outbreak of five cases of Pontiac fever occurred after exposure to aerosols from sludge in a sewage treatment plant of the Danish food industry [ 20 ]. The strain isolated from sludge in concentrations of 1.5 × 10 7 cfu/g was L. pneumophila serogroup 1, subgroup OLDA/Oxford (MAb 3/1 negative). In this outbreak, the workers used respirators, but the filters were effective only against chemical substances.
Previously, a large community outbreak with 86 cases of LD was associated with cooling towers and waste water basins of a petrochemical plant occurred in France in 2003-2004 [ 18 ]. The strain causing the outbreak was L. pneumophila serogroup 1, strain Lens. The aerosols with legionellae, spread by a cooling tower, were infectious at a distance of at least six kilometers. The plant was later closed. In Norway, an air-scrubber was spreading Legionella aerosols over a distance of ten kilometers, resulting in 103 LD cases and ten deaths [ 21 , 22 ]. It is assumed that aerosols containing Legionella from the waste water aeration ponds originally contaminated the air scrubber. The scrubber was cleaned of legionellae and the aeration pond changed to an anaerobe pond. Thus, the sources of these larger outbreaks have been limited.
In contrast, Swedish and Finnish studies have indicated that heavy contamination of active sludge basins with legionellae is very common [ 7 ], (Unpublished data, Kusnetsov J, Torvinen E, Lehtola M and Miettinen IT). Further, air samplings in Norway and France in the vicinity of active sludge basins revealed that viable Legionella cells can be isolated up to 180-270 meters downwind [ 23 , 24 ]. It seems therefore likely that any waste water treatment plant with an active sludge basin under aeration can contain higher concentrations of Legionella bacteria and also produce aerosols with legionellae. Evidence of exposure to legionellae can be established if an increased frequency of elevated Legionella antibodies in serum samples can be detected [ 22 ].
In addition, some of these waste water treatment plants use cooling towers to lower the waste water temperature. It would be very useful to know if these waste water cooling towers are clean and maintained in accordance with the European and WHO Legionella guidelines [ 11 , 25 ]. In the European guidelines, a concentration of 1000 cfu/l of legionellae is recommended as the highest acceptable concentration which can be present in cooling water (technical guidelines).
Water treatment plants with active sludge basins should be considered as a possible source of community acquired Legionella infections, directly or indirectly via cooling towers. In addition, the employees should protect themselves by using respirators at or in the vicinity of water treatment plants. The Finnish Work safety act (738/2002) [ 26 ] has stated that employees should be protected against biological factors, including Legionella bacteria, at waste water treatment plants. Especially in industrial waste water treatment plants with high water temperatures, Legionella concentrations may be very high. In 2005, Finnish forest industry employees were instructed to use respirators while working in the vicinity of the water treatment plants. Developing ways to lower Legionella concentrations in these water systems would be the next step in diminishing the risk of Legionella infection.
These LD cases might have remained undiagnosed if our environmental study had not increased awareness about potential Legionella exposure in waste water treatment plants. These findings suggest that the clinicians should consider LD when treating patients with pneumonia from these industrial settings.
Consent
Written informed consents were obtained from the Cases for publication of this report. | Background
Finnish and Swedish waste water systems used by the forest industry were found to be exceptionally heavily contaminated with legionellae in 2005.
Case presentation
We report two cases of severe pneumonia in employees working at two separate mills in Finland in 2006. Legionella serological and urinary antigen tests were used to diagnose Legionnaires' disease in the symptomatic employees, who had worked at, or close to, waste water treatment plants. Since the findings indicated a Legionella infection, the waste water and home water systems were studied in more detail. The antibody response and Legionella urinary antigen finding of Case A indicated that the infection had been caused by Legionella pneumophila serogroup 1. Case A had been exposed to legionellae while installing a pump into a post-clarification basin at the waste water treatment plant of mill A. Both the water and sludge in the basin contained high concentrations of Legionella pneumophila serogroup 1, in addition to serogroups 3 and 13. Case B was working 200 meters downwind from a waste water treatment plant, which had an active sludge basin and cooling towers. The antibody response indicated that his disease was due to Legionella pneumophila serogroup 2. The cooling tower was the only site at the waste water treatment plant yielding that serogroup, though water in the active sludge basin yielded abundant growth of Legionella pneumophila serogroup 5 and Legionella rubrilucens . Both workers recovered from the disease.
Conclusion
These are the first reported cases of Legionnaires' disease in Finland associated with industrial waste water systems. | Case presentation
General awareness of potential Legionella exposure has increased recently in the paper and pulp industries. As a consequence, these two severe respiratory infections suffered by employees working in proximity of waste water treatment plants of two different paper and pulp mills were studied in more detail for suspected Legionella infection.
Case A
Case A (male, 51 years, previously healthy, smoker) fell ill with pneumonia on August 9, 2006. Five days earlier, he had been accidentally exposed to aerosols of waste water, when water splashed during the installation of a new pump to the post-clarification basin. This basin separates water and sludge after the active sludge basin. The employees had been instructed to use respirators while working in the vicinity of the water treatment plant. Compliance with these instructions, however, was not good, and during the aerosol exposure, Case A had not been wearing a respirator.
Case A was diagnosed with legionellosis by urinary antigen immunoassay, this giving a positive response on the 7 th day. The first serological tests displayed an antibody response to L. pneumophila (the 6 th day, IgM++ and IgA+; the 23 rd day, IgM+++, IgA+++, IgG+++) and L. micdadei (the 6 th day, IgM+; the 23 rd day IgM++, IgG+, in-house EIA). The last test was conducted ten weeks after the onset of illness with the in-house IFA, giving values of 1:256 for L. pneumophila serogroup 1 and 1:1024 for L. dumoffii , while other L. pneumophila serogroups and Legionella species had a titer 1:128 at their highest. Despite repeated cultures no clinical isolate was ever obtained. Medical treatment was started two days after the onset of symptoms and was initially based on cefuroxime, roxithromycin, meropenem, but subsequently on piperacillin, and moxifloxacin was administered after the diagnosis of LD had been established. The patient spent four days in intensive care, and he was discharged after two weeks in hospital. This radiography-confirmed pneumonia case was classified as a confirmed case of LD according to EWGLI case definitions.
Case B
Case B (male, 61 years, had stopped smoking five years previously) was diagnosed with pneumonia on September 24, 2006. He had been working about 200 m eastward from the active sludge basin and the cooling towers of the waste water treatment plant on days 4, 5, 6, 11, 12 and 13 before the onset of the respiratory symptoms. In the intervening days, he was not at work. This mill also instructed employees working in the vicinity of the water treatment plant to use respirators. However, Case B was not working at the waste water treatment plant and did not wear a respirator.
The diagnosis of LD for this radiography-confirmed case of pneumonia was based on serological tests. Nine days after the diagnosis, the patient tested seropositive for L. pneumophila serogroups 1-4 (IgM++, IgA+++, IgG+) and L. micdadei (IgG+, EIA). A subsequent serum sample analysed with IFA in another laboratory revealed a high titre for L. pneumophila serogroup 2 (1:4096, the 26 th day) and it was still high even though declining (1:2048) 12 weeks after the onset of the illness. For L. micdadei , the titers were 1:256 and 1:256, and for other L. pneumophila serogroups and Legionella species the titer values were only up to 1:64 and 1:128 (the 26 th day and 12 week samples). The difference between the causative serogroup and the other serogroups and species was at least two titer steps, showing the immune response specific for L. pneumophila serogroup 2.
His urinary antigen test was negative when tested on the 36 th day. No culturing for legionellae from clinical samples was performed. He was treated with roxithromycin and he recovered at home. This pneumonia was classified as a presumptive LD case.
Neither of the two cases had travelled abroad during the incubation period.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
JK arranged environmental samplings, interviewed the patients, wrote the first version of the manuscript, and had final responsibility for the decision to submit for publication. L-KN and TK were the physicians who cared for the cases, SAU and SM did typing of Legionella strains, and TP, TMNN and K-PM collected data. All the authors revised the article and approved the final version.
Pre-publication history
The pre-publication history for this paper can be accessed here:
http://www.biomedcentral.com/1471-2334/10/343/prepub | Acknowledgements
We would like to acknowledge the case employees for their co-operation and for providing consent for publication of this report. The awareness of occupational health nurse Ms. Anneli Junes is gratefully acknowledged. Health inspectors involved in this study are warmly acknowledged. We also wish to thank Ms. Riitta Lahtinen for providing expert help with the environmental study. The skillful laboratory assistance of Ms. Marjo Tiittanen is gratefully acknowledged. We warmly thank Dr. Ewen Macdonald and Dr. Diane Lindsay for their support with English language. A part of this study was funded by Finnish Work Environment Fund (FEEL-study, number 105304). | CC BY | no | 2022-01-12 15:21:37 | BMC Infect Dis. 2010 Dec 2; 10:343 | oa_package/85/eb/PMC3014939.tar.gz |
PMC3014940 | 21245908 | Introduction
Over five million people in the United States alone suffer with heart failure [1] because unlike some organs, the heart is unable to repair itself after injury [2] . Human embryonic stem cells (hESCs) grow and divide indefinitely while maintaining the potential to develop into many tissues of the body. As such, they provide an unprecedented opportunity to treat a variety of human diseases characterized by tissue loss or insufficiency. Animal studies have shown that pluripotent hESCs have a high risk of tumor formation [3] , while fully differentiated hESC-derived cardiomyocytes (CMs) confer only modest functional benefit [4] . This suggests that from a developmental standpoint, mature CMs may be beyond the ability to fully incorporate into existing muscle. Therefore, the identification of hESC-derived myocardial precursors that are committed to the cardiac lineage, but retain the plasticity to facilitate complete engraftment has been an important goal [5] .
Work over the past decade has shown that hESCs differentiate into a heterogeneous population of CMs in culture, with gene expression patterns and electrophysiological properties reminiscent of embryonic atrium, ventricle and specialized conduction tissue [6] , [7] , [8] , [9] , [10] . The mechanisms that drive CM subtype specification, however, are not well understood.
To approach both the need for a model system with which to elucidate the process of human CM subtype specification, and provide a source of human myocardial precursors for cell therapy studies, we engineered a hESC line that identifies multipotent myocardial precursor (hMP) cells. These cells give rise to multiple CM subtypes, and is therefore uniquely suited to address both of these needs. | Materials and Methods
hESC culture and differentiation
All work with hESCs was done with the approval of the UCSF Stem Cell Research Oversight Committee. The parent H9 hESC line (WA09; WiCell) was maintained on irradiated mouse embryonic fibroblast (MEF) feeder cells (Millipore) in a medium comprised of Knockout DMEM (Invitrogen) supplemented with 20% Knockout Serum Replacement (Invitrogen), 2 mM glutamine, 0.1 mM nonessential amino acids, 0.1 mM β-mercaptoethanol and 15 ng/ml recombinant human FGF-basic (R&D Systems). Differentiation was initiated by human embryoid body (hEB) formation in suspension as previously described [28] . Briefly, colonies of hESCs were dissociated into small clusters by exposure to Collagenase IV (Sigma-Aldrich), then allowed to differentiate in a medium comprised of Knockout DMEM (Invitrogen) supplemented with 20% Defined Fetal Bovine Serum (Hyclone), 2 mM glutamine, 0.1 mM non-essential amino acids, and 0.1 mM β-mercaptoethanol. After 4–7 days in suspension, hEBs were attached to gelatin-coated 12-well culture plates and allowed to differentiate for an additional 14–21 days. For re-culture and expression profiling experiments, hEBs were dissociated with TrypLE Express (Invitrogen) to generate single cell suspensions, stained with propidium iodide to distinguish between live and dead cells, and sorted on the basis of GFP expression using a FACSAria (Becton Dickinson) with standard filter sets using previously described methods [29] , [30] .
Plasmid and cell line construction
The 2K7 bsd lentivector (kindly provided by David Suter [14] ) was used to assemble a lentiviral plasmid capable of driving the expression of enhanced green fluorescent protein (GFP) by a ubiquitin-C or α-myosin heavy chain (αMHC) promoter. The human ubiquitin-C promoter (position −1225 to −6 upstream from the translation start site; [31] ), amplified by PCR, or a 1.7 kb Eco RI- Sal I fragment from the mouse αMHC promoter (generously provided by Jeffrey Robbins [13] ) was inserted into the pENTR 5′-TOPO entry vector using the pENTR 5′-TOPO TA Cloning Kit (Invitrogen). A cDNA encoding GFP was inserted into the pENTR/D-TOPO entry vector using the pENTR/D-TOPO Cloning Kit (Invitrogen). The promoters and reporter gene were inserted into the double recombination site of the 2K7 bsd plasmid using LR Clonase Plus Enzyme Mix (Invitrogen) according to the manufacturer's instructions.
Immunocytochemistry
The lineage fate of differentiated hEBs was determined by staining 14-day-old, adherent differentiating hEBs attached to cover slips in 12-well culture plates. Cover slips were fixed with 4% paraformaldehyde, then permeabilized with 50% methanol/50% PBS, then 100% methanol, then 50% methanol/50% PBS/0.1% Triton X-100, and finally PBS/0.1% Triton X-100. hEBs were incubated with blocking buffer (PBS/10% horse serum/1% BSA/0.1% Triton X-100), then with primary antibody (1–5 μg/ml) in blocking buffer. Primary antibodies used were mouse anti-human cardiac troponin T (LabVision/Neomarkers MS-295-P1; clone 13–11), mouse anti-human cardiac α-actinin (Sigma A7732; clone EA-53), mouse anti-human alpha-fetoprotein (Sigma A8452; clone C3), mouse anti-human nestin (R&D Systems MAB1259; clone 196908) or mouse anti-human smooth muscle actin (R&D Systems MAB1420; clone 1A4). Cover slips were washed with blocking buffer, incubated with 1∶500 dilution of goat anti-mouse Alexa-Fluor 594 (Invitrogen A20185), washed with PBS/1% Triton X-100, then mounted with PBS and DAPI and analyzed by confocal microscopy using a Zeiss LSM510 META system. Alternatively, 8-day-old differentiating hEBs were sorted for GFP expression and re-cultured in differentiation medium in the presence of 10 μM p160 Rho-associated coiled-coil kinase inhibitor (Calbiochem 688000; Y-27632). At 14 days, differentiating hEBs were stained in situ in culture plates as described above using the same antibodies, and analyzed by immunofluorescence microscopy using a Nikon Microphot-FXA fluorescence/phase microscope and QImaging Retiga 2000R digital camera (Diagnostic Instruments) with MetaMorph software (Molecular Devices).
Teratomas
All experiments involving animals were done with the approval of the UCSF Institutional Animal Care and Use Committee. To form teratomas, 5×10 5 hESCs were mixed with an equal volume of 1 mg/ml Phaseolus vulgaris lectin (PHA-P L1668; Sigma), pelleted, and incubated in growth medium overnight at 37°C, 5% CO 2 in a 0.4 μm MILLICELL (Millipore). At least 2 cell pellets were grafted under each kidney capsule of 8-week-old female CB17 SCID-Beige mice (n = 7) using published techniques [32] . Transplanted cells formed teratomas in the recipients and were analyzed 10 weeks after grafting. Teratomas were fixed in 10% buffered formalin, embedded in paraffin, and 5 μm sections were stained with purified polyclonal rabbit anti-GFP (Molecular Probes A11120) at 1∶1500 and biotinylated goat anti-rabbit IgG (Vector BA-1000). Slides were developed using the VECTASTAIN Elite ABC kit (Vector) and counterstained with hematoxylin and eosin to identify tissue structures.
Quantitative real-time PCR
For analysis of transcript expression, GFP + hESCs were sorted by FACS at indicated time points, or beating areas from approximately 20 hEBs were visualized with a Leica MZ6 microscope and manually excised using an 18 g needle. The collected tissue samples were treated with 0.05% Trypsin-EDTA to generate a single cell suspension prior to RNA isolation. RNA was isolated and cDNA synthesized from ∼50,000 hEB-derived cells or proliferating hESCs using the Taqman Gene Expression Cells-to-CT kit (Ambion). cDNA was quantitated using a Nanodrop ND-1000 Spectrophotometer (Nanodrop Technologies, ND Software version 3.3.0). Linear pre-amplification of target sequences was accomplished using the Applied Biosystems PreAmp system. Relative expression was determined using the TaqMan Assay (Applied Biosystems) on an ABI 7300 Real-Time PCR system with the following primer pairs (ABI): GATA4 (Hs00171403_m1), NKX2-5 (Hs00231763_m1), Is~l1Hs 01099687_m1;Hs00158126_m1sarcolipin (Hs00161903_m1;Hs01888464_s1), HCN4 (Hs00975492_m1;Hs00175760_m1), MLC2a (Hs00221909_m1), MLC2v (Hs00166405_m1;Hs01125721_m1), αMHC (Hs00411908_m1), cTnT (Hs00165960_m1), smooth muscle actin (Hs00242273_m1), α-fetoprotein (Hs00173490_m1), nestin (Hs00707120_s1), α1 integrin (Hs00235030_m1), α2 integrin (Hs00158148_m1), α4 integrin (Hs00168433_m1), α5 integrin (Hs00233732_m1), α6 integrin (Hs01041011_m1), α7 integrin (Hs00174397_m1), αv integrin (Hs00233790_m1), β1 integrin (Hs00559595_m1), β5 integrin (Hs00609896_m1), and GAPDH (4326317E). Cycle times to detection were normalized against a reference gene, GAPDH, and relative changes were calculated using ABI Version 1.4 Sequence Detection Software.
Microelectrode array analysis
Individual GFP-expressing and wild type hEBs demonstrating contractile activity were mechanically dissected and plated on fibronectin-coated microelectrode arrays (MEA; Multi Channel Systems, Reutlingen, Germany). Mouse atrial HL-1 cells (generously provided by William Claycomb, LSU Health Sciences Center) were cultured on MEAs as control. The MEA system consisted of a 50×50 mm glass substrate with an embedded 1.4×1.4 mm matrix of 60 titanium nitride-gold contact electrodes with interelectrode distance of 200 μm. This allowed for simultaneous recording of extracellular field potentials (FP) from all electrodes over extended periods of time. Spontaneous electrical activity was recorded at 10 kHz. The temperature was kept at 37°C. Analysis of recordings was done using MC_Rack (Multi-Channel Systems) and a customized toolbox for MATLAB. The following parameters were determined: size of the largest negative peak (FP min ), last postive peak of the cycle (FP max ), the time interval between FP min and FP max (FP dur ), and decay of extracellular potential (time from the onset of the FP to FP min = FP rise ) [19] , [20] .
Array comparative genomic hybridization
Genomic DNA (300–500 ng) from the αMHC-GFP reporter hESC line was isolated using the Qiaeasy DNA kit for tissue (Qiagen). DNA was labeled with Cy3 and Cy5 using the BioPrime DNA labeling system (Invitrogen), hybridized to HumArray 3.2 Human chromosome arrays, and analyzed for chromosomal composition. Differentially labeled human male reference genomic DNA was run as a control. The HumArray 3.2 array contains 2,464 bacterial artificial chromosome clones spotted in triplicate and distributed uniformly across the genome. Each clone contains at least one STS and is mapped to the human genome sequence. Clones containing unique sequences near the telomeres and genes known to be significant in cancer and medical genetics are included on these arrays.
mRNA expression profiling
Sample preparation, labeling, and array hybridizations were performed according to standard protocols from the UCSF Shared Microarray Core Facilities and Agilent Technologies ( http://www.arrays.ucsf.edu and http://www.agilent.com ). Total RNA quality was assessed using a Pico Chip on an Agilent 2100 Bioanalyzer (Agilent Technologies). RNA was amplified using the Sigma whole transcriptome amplification kit following the manufacturer's protocol (Sigma-Aldrich), and subsequent Cy3-CTP labeling was performed using the NimbleGen one-color labeling kit (Roche-NimbleGen). The size distribution and quantity of the amplified product was assessed using an Agilent 2100 Bioanalyzer and a Nanodrop ND-8000 (Nanodrop Technologies); the labeled DNA was assessed using the Nandrop 8000, and equal amounts of Cy3 labeled target were hybridized to Agilent human whole genome 4×44 K Ink-jet arrays. Hybridizations were performed for 14 hrs, according to the manufacturers protocol. Arrays were scanned using an Agilent microarray scanner and raw signal intensities were extracted with Feature Extraction v.10.1 software (Agilent Technologies). Data were further analyzed using Ingenuity Pathways Analysis (Ingenuity® Systems) to identify biological pathways involved in CM differentiation. The false discovery rate (FDR) from the data set was used for canonical pathways analysis.
Statistics
Student's t-test was used to compare the means of the frequency count distributions of FP dur and FP rise . A value of p<0.05 was considered significant. These analyses were performed using SPSS v.16 (SPSS, Inc.) for Macintosh.
The mRNA expression array dataset was normalized using the quantile normalization method [33] . No background subtraction was performed, and the median feature pixel intensity was used as the raw signal before normalization. A one-way ANOVA linear model was fit to the comparison to estimate the mean M values and calculated FDR for each gene for the comparison of interest. All procedures were carried out using functions in the R package limma in Bioconductor [34] , [35] . | Results
Construction of an α-myosin heavy chain human embryonic stem cell reporter line
The α-myosin heavy chain (αMHC) gene has been shown to be expressed both early and late during murine cardiac development [11] , [12] . Taking advantage of this developmentally broad expression pattern, we used a subsequence of the murine αMHC promoter consisting of nucleotides −1679 through +1 relative to the translation start site of the αMHC mRNA to construct a myocardial-specific enhanced GFP reporter in the H9 hESC line. This contains MEF1 and MEF2 binding sites, two thyroid hormone response elements, the cardiac troponin T-responsive MCAT element, and SRF binding motif [13] . We used the HIV-derived, self-inactivating lentiviral vector, 2K7 bsd [14] , in which the HIV gag , pol , and env genes are deleted and the HIV-1 flap sequence and woodchuck hepatitis virus post-transcriptional regulatory element are included to improve infectious titer and gene expression. To mitigate against the effects of clonal variation, we isolated stable transfectants by population selection. Presumably, cells in which viral integration disrupted essential genes did not survive selection. Upon differentiation of the resulting αMHC-GFP hESC line by human embryoid body (hEB) formation, GFP expression was detected solely in 14 day hEBs co-expressing cardiac troponin T (cTnT) ( Fig. 1 ).
The αMHC-GFP reporter is activated solely in hESC-derived cardiomyocytes
To specifically localize cardiac versus non-cardiac proteins within differentiating αMHC-GFP + hEBs, and confirm that GFP did not inhibit the expression of non-cardiac markers, we determined co-localization of alpha-fetoprotein, a marker of primitive endoderm, nestin, a marker of neuroectoderm, and smooth muscle actin, a marker of non-cardiac mesoderm, with GFP in the reporter line, compared to the co-localization of these proteins with ubiquitin C-driven, constitutive expression of GFP in a ubiC-GFP hESC line at day 14 of hEB differerentiation. This demonstrated that while CM-specific cTnT, but not smooth muscle actin, alpha-fetoprotein or nestin, co-localized with αMHC-driven GFP expression in the αMHC-GFP reporter line, all of these markers co-localized with ubiquitin C-driven GFP expression in the ubiC-GFP line ( Fig. 1 ). This also confirmed that the absence of smooth muscle actin, alpha-fetoprotein or nestin co-expression with αMHC-GFP was not due to interference by GFP expression.
To determine the fate of selected αMHC-GFP + hEBs, we sorted hEBs for GFP expression at day 8 of differentiation, then re-cultured αMHC-GFP + cells under differentiation conditions for an additional 6 days (day 14 of hEB differentiation) and assessed CM-specific cTnT and cardiac α-actinin expression by in situ immunocytochemistry. This showed that cells activating the reporter on day 8 generated a lawn of cTnT + cardiac α-actinin + CMs in culture ( Fig. 2 ), and that these cells demonstrate spontaneous contractile activity ( Movie S1 and S2 ).
αMHC-GFP-derived myocardial precursors give rise to muscle in vivo
To establish that hESCs expressing the αMHC-GFP reporter also give rise to muscle in vivo , we evaluated undifferentiated αMHC-GFP hESCs by teratoma assay. We grafted 10 6 hESCs beneath the kidney capsule of 8-week-old SCID mice, and analyzed for teratoma formation after 10 weeks. The αMHC-GFP hESCs consisted of tissues arising from all three embryonic germ layers, however, indirect immunohistochemical analysis showed that GFP expression was restricted to striated muscle, specifically mononucleated myofibers characteristic of cardiac muscle ( Fig. 3 ). Restricted GFP localization to cardiac muscle in all teratomas analyzed also suggested that hESCs expressing the αMHC-GFP reporter do not form tumors comprised of other tissues.
αMHC-GFP hESCs remain genetically stable
It is well-established that hESCs maintained in culture are at risk for developing aneuploidy [15] , [16] . Therefore, it becomes especially important to confirm genomic stability in any hESC-derived line to be used for either developmental investigation or therapy. To assess the genomic integrity of the αMHC-GFP hESC line, we analyzed cells for single chromosomal region gains and losses by array comparative genomic hybridization [17] . This showed that αMHC-GFP hESCs maintain a normal karyotype at the level of individual gene regions ( Fig. 4 ).
The αMHC-GFP reporter identifies an early myocardial precursor that gives rise to multiple CM subtypes
Others have shown that CMs derived from hESCs in culture can display the molecular and electrical properties of embryonic ventricular, atrial and nodal tissue [6] . This heterogeneity occurs with differentiation in culture using a variety of conditions [7] , [8] , [9] , [10] , and suggests that hESCs are capable of giving rise to myocardial precusors that precede the developmental branchpoint between first and second heart field specification. We counted the number of GFP + (αMHC + ) hESCs over the course of hEB differentiation, and found that these cells started to appear in culture at day 5, commensurate with the onset of GATA4 and NKX2-5 expression, and several days before the expression of cTnT, which is associated with spontaneous contractile activity [18] ( Fig. 5 ).
Since the reporter appeared to identify early human myocardial precursors (hMPs), we wanted to determine whether these hMPs were capable of generating a heterogeneous population of human CMs based on molecular and electrical characteristics. We differentiated αMHC-GFP hESCs for 21 days, and manually dissected GFP + hEBs for analysis by qPCR and microelectrode array. Evaluation of individual GFP + hEBs by qPCR demonstrated that molecular signatures characteristic of embryonic atrial (Islet-1 (ISL-1) + sarcolipin (SLN) + Na + /K + hyperpolarization-activated cyclic nucleotide-gated channel 4 (HCN4) − myosin light chain-2 atrial (MLC2a) + MLC2v − αMHC + cTnT + ), left (ISL-1 − SLN − HCN4 − MLC2a − MLC2v + αMHC + cTnT + ) and right (ISL-1 + SLN − HCN4 − MLC2a − MLC2v + αMHC + cTnT + ) ventricular, and nodal (ISL-1 + SLN − HCN4 + MLC2a − MLC2v − αMHC + cTnT − ) tissue were present in distinct hEBs ( Fig. 6A ) at an approximate ratio of 40% ISL-1 + SLN + HCN4 − MLC2a + MLC2v − αMHC + cTnT + , 10% ISL-1 + SLN − HCN4 + MLC2a − MLC2v − αMHC + cTnT − , 30% ISL-1 + SLN − HCN4 − MLC2a − MLC2v + αMHC + cTnT + and 20% ISL-1 − SLN − HCN4 − MLC2a − MLC2v + αMHC + cTnT + among >30 hEBs analyzed.
To determine whether the electrical properties of GFP + hEBs were similarly heterogeneous in nature, we allowed dissected hEBs to adhere to microelectrode arrays (MEAs) ( Fig. 6B ) and performed simultaneous recording of extracellular field potentials (FP) over time ( Fig. 6C ). This allowed us to measure the size of the largest negative peak (FP min ), the last postive peak of the cycle (FP max ), the time interval between FP min and FP max (FP dur ), and decay of the extracellular potential (time from the onset of the FP to FP min = FP rise ). Others have shown that FP rise directly correlates to action potential rise time, and FP dur directly correlates to action potential duration [19] , [20] . Compared to HL-1 cells, which represent a homogenous culture of mouse atrial CMs [21] with FP dur ≤75 ms and FP rise ≤50 ms, both αMHC-GFP + and wild type, beating hEBs demonstrated similarly heterogeneous combinations of FP dur (25–250 ms versus 10–475 ms; p = 0.86) and FP rise (10–175 ms versus 10–275 ms; p = 0.83) ( Fig. 6D ).
Expression profiling of human CM differentiation
Since we had established that this αMHC-GFP hESC line could identify myocardial precusors capable of giving rise to all embryonic CM subtypes, we used it to examine potential pathways involved in CM differentiation. We isolated hMPs at day 8 of differentiation, after the onset of GATA4, NKX2-5 and αMHC expression, but before cTnT expression or spontaneous contractions ( Fig. 5 ), and at day 14 of differentiation, after the onset of cTnT expression and spontaneous beating ( Fig. 1 ), and compared the gene expression profiles of these two differentiation time points to each other and to undifferentiated hESCs using an array of >41,000 unique sequences representing the whole human genome. Analysis of the 830 most highly differentially expressed genes from biological replicates of each time point resulted in the clustering shown in Fig. 7 . This indicated that the reproducibility of hMP differentiation was reasonably robust. Overall, there were 3,279 probes targeting human genes that showed differential expression across the three time points with a false discovery rate <0.05. The complete data set has been deposited in Gene Expression Omnibus ( http://www.ncbi.nlm.nih.gov/geo/ ; Accession Number GPL6480).
Analysis of cardiac-specific gene expression demonstrated that hMP differentiation proceeded in a manner commensurate with cardiac morphological development ( Table 1 ). Differentiating hMPs expressed markers of cardiac tube formation (NKX2-5, GATA4) and looping (MEF2c, αMHC) by day 8, and markers of spontaneous contractile activity (cTnT, cTnI) by day 14. In addition, expression levels of atrial (ISL-1, ANF) and right ventricular (ISL-1, HAND2) genes were observed by day 8 and continued to rise through day 14, while expression of genes associated with development of the atrioventricular septum (TBX2) and left ventricle (HAND1, TBX5) were observed by day 14.
The role of Wnt signaling during myocardial specification and early cardiac morphogenesis has been described [22] . Wnt ligands and Frizzled receptors specifically have been implicated during early heart development. This is supported by our findings that the expression of Wnt1, 2, 5B and 11, as well as Frizzled 4 and 5 was upregulated by day 8 of hMP differentiation ( Table 2 ). In addition, secreted Frizzled-related proteins (SFRPs) that are known to inhibit Wnt signaling and regulate CM proliferation, were downregulated at day 8, but then upregulated in the case of SFRP3 at day 14 ( Table 2 ). CM migration during heart development has been shown to rely on transforming growth factor β (TGFβ) signaling [23] . Consistent with its role in cardiogenesis, several effectors and inhibitors of TGFβ pathways were upregulated and downregulated, respectively, over the course of hMP differentiation ( Table 3 ). | Discussion
Although β-myosin heavy chain (βMHC) is the predominant isoform expressed during murine embryonic and fetal development, the transition from βMHC to αMHC begins as early as day 7.5 p.c. in mouse embryos with the appearance of somites and the onset of cardiac tube formation [11] , [12] . Expression of both isoforms approaches equivalence as the cardiac tube begins to contract by day 9.5 p.c., and the β/α ratio begins to reverse with αMHC becoming the predominant isoform during postnatal life [11] , [12] . These observations during mouse development in vivo parallel mouse ESC (mESC) differentiation in vitro . αMHC is expressed by day 8 of mESC differentiation and is expressed exclusively in beating mouse EBs [12] . While many studies have used the murine αMHC promoter to track mature CMs, our data suggest that because of its expression across early and mature stages of CM development even between species, the murine αMHC promoter can be used to track CM differentiation from a multipotent myocardial precursor in real time. This has afforded us an unprecedented opportunity to study human CM differentiation from undifferentiated ESCs through EB formation and ultimately embryonic CM subtypes.
Other laboratories have developed transgenic/reporter hESC lines to derive differentiated CMs. As distinct from the work described here, however, these lines have allowed either for the selection of mature CMs only, or of progenitors that give rise to non-muscle cardiac cells in addition to myocardial cells. Huber et al. used lentiviral vectors to produce stable hESC lines in which enhanced GFP was expressed under control of the MLC2v promoter [24] . While these lines were able to generate electrically active CMs of unspecified subtype, MLC2v is expressed later in CM differentiation than αMHC, and does not afford the same insight into early CM differentiation. Xu et al. generated stable hESC lines using a plasmid containing the αMHC promoter driving expression of the neomycin resistance gene [25] . While this afforded an effective strategy for enriching CMs for cell transplant experiments, it does not accommodate the study of developmental stages over time. Kita-Matsuo et al. recently reported the design of a set of lentiviral vectors to generate multiple stable hESC lines with eGFP and mCherry reporters or with puromycin resistance using the αMHC promoter [26] . The focus of these studies was to create tools to enhance CM production for large-scale clinical application. While these investigators demonstrated that hESC-derived CMs have gene expression profiles similar to those found in adult hearts and electrophysiological properties of embryonic CMs, analysis of CM differentiation using these lines was not reported. To track the fate of human Isl1 + cells and their progeny during hESC differentiation, Bu et al. used Isl1: cre hESCs transfected with a pCAG-flox-DsRed reporter plasmid to achieve irreversible DsRed expression in Isl1 + cells. In clonal assays of day 8 hEBs, about half of the DsRed + clones expressed markers of the three major cardiac lineages, cTnT (CMs), PECAM1/CD31 (endothelial cells), and smooth muscle troponin (smooth muscle cells) [27] , suggesting the identification of a multipotent precursor that is not restricted to the cardiac muscle lineage.
It has long been appreciated that hESCs differentiate into a heterogeneous population of atrial, ventricular and specialized conduction CMs in culture [6] , [7] , [8] , [9] , [10] . Whether this reflects a stochastic process in vitro , and is driven by a combination of genetic programming and the extracellular milieu in vivo , has not been established. Understanding the mechanisms that drive CM subtype specification, however, will be essential to both understanding cardiac development and developing cell-based reagents for myocardial therapy.
In summary, we have identified multipotent human myocardial precursors (hMPs) using an αMHC-GFP reporter hESC line. We have demonstrated that reporter activation is restricted to hESC-derived CMs differentiated in vitro and in vivo , and that the reporter does not interfere with hESC genomic stability. Importantly, we show that hMPs give rise to multiple CM subtypes and can be used to explore CM differentiation on the molecular level by expression profiling. These precursors will provide important insight into the pathways regulating human myocardial development, and provide a novel therapeutic approach to stem cell therapy for cardiac disease. | Conceived and designed the experiments: CR SSYW FWK SSM WL DJE RJL HSB. Performed the experiments: CR SSYW FWK SSM WL HSB. Analyzed the data: CR SSYW SSM WL DJE RJL HSB. Contributed reagents/materials/analysis tools: DJE RJL HSB. Wrote the paper: CR SSYW HSB. Obtained permission for use of cell line: RJL.
Unlike some organs, the heart is unable to repair itself after injury. Human embryonic stem cells (hESCs) grow and divide indefinitely while maintaining the potential to develop into many tissues of the body. As such, they provide an unprecedented opportunity to treat human diseases characterized by tissue loss. We have identified early myocardial precursors derived from hESCs (hMPs) using an α-myosin heavy chain (αMHC)-GFP reporter line. We have demonstrated by immunocytochemistry and quantitative real-time PCR (qPCR) that reporter activation is restricted to hESC-derived cardiomyocytes (CMs) differentiated in vitro , and that hMPs give rise exclusively to muscle in an in vivo teratoma formation assay. We also demonstrate that the reporter does not interfere with hESC genomic stability. Importantly, we show that hMPs give rise to atrial, ventricular and specialized conduction CM subtypes by qPCR and microelectrode array analysis. Expression profiling of hMPs over the course of differentiation implicate Wnt and transforming growth factor-β signaling pathways in CM development. The identification of hMPs using this αMHC-GFP reporter line will provide important insight into the pathways regulating human myocardial development, and may provide a novel therapeutic reagent for the treatment of cardiac disease. | Supporting Information | The authors acknowledge technical support from A. Barczak, R. Barbeau and C. Eisley of the UCSF Sandler Asthma Basic Research Center Functional Genomics Core Facility, and members of the Bernstein Laboratory for helpful discussion. | CC BY | no | 2022-01-13 08:14:21 | PLoS One. 2011 Jan 4; 6(1):e16004 | oa_package/0f/08/PMC3014940.tar.gz |
|
PMC3014941 | 21245909 | Introduction
CD8 + T cells are critical to protection against infection by intracellular pathogens, including liver stage malaria parasites. CD8 + T cells induced by immunization with radiation-attenuated Plasmodium sporozoites (γ-spz) or sub-unit vaccines are capable of inhibiting the development of liver stage parasites [1] , [2] , [3] , [4] . T cell priming by γ-spz occurs primarily in the skin-draining lymph node after parasite inoculation in the skin by either needle or the bite of an infected mosquito, followed by dissemination of effector T cells throughout the body, including the spleen and liver [4] . This priming in the lymph node is closely dependent on CD4 + cells and the absence thereof results in a reduced effector population [5] . This dependence on helper T cells at such an early time point is unique among models of CD8 + T cell priming, which often demonstrate unaltered primary CD8 + T cell responses to pathogens in the absence of helper T cells [6] , [7] , [8] , [9] , [10] , with defects only apparent in functional recall of resting memory cells [7] , [9] , [11] . These studies have demonstrated an uncoupling of CD8 + T cell clonal expansion, survival, and acquisition of effector function. In view of the clear and early dependence of γ-spz-induced CD8 + T cells on CD4 + T cells, we sought to characterize the effect of helper T cells on the functional development of anti-parasite CD8 + T cells.
In the current study, we evaluated the role of CD4 + T cell help on the development of functional anti-malaria effector and memory CD8 + T cells by using Thy-1 allelic mismatched T cells so that survival could be measured independently of effector function. We found that while effector and memory CD8 + T cells from CD4-depleted mice (“helpless” T cells) were severely reduced in magnitude compared to those primed in the presence of CD4 + T cells, helpless effector and memory CD8 + T cells were fully competent to produce cytokines and degranulate upon restimulation ex vivo . Interestingly, however, helpless CD8 + T cells failed to confer any level of protection against live sporozoite infection, indicating that large numbers of anti-parasite CD8 + T cells are critical to protection. Our studies indicate that the role of CD4 + T cells in modulating the CD8 + T cell response to the circumsporozoite protein of irradiated P. yoelii sporozoites appears to be restricted to ensuring the survival of activated T cells, without a discernible effect on the development of their functional properties. | Methods
Ethics Statement
All animal procedures were approved by the Institutional Animal Care and Use Committee of the Johns Hopkins University (Protocol Number MO09H41) following the National Institutes of Health guidelines for animal housing and care.
Mice
Female BALB/c mice 5–8 weeks of age were purchased from Taconic Farms and housed in microisolater cages. Transgenic mice expressing a TCR specific for a H-2K d -restricted epitope of the Plasmodium yoelii circumsporozoite protein (SYVPSAEQI) have been previously described [41] . For adoptive transfer, TCR-transgenic CD8 + T cells from whole splenocytes were injected intravenously into naïve BALB/c mice. Jα18 −/− mice were kindly provided by Mitchell Kronenberg (La Jolla Institute for Allergy and Immunology) with permission of Masaru Taniguchi (RIKEN). CD1d −/− mice were purchased from Jackson Laboratories.
Depletion of CD4 + and Thy-1.2 + cells
For in vivo depletion of CD4 + cells, 200 μg of anti-CD4 monoclonal antibody (clone GK1.5) was injected intraperitoneally on two consecutive days prior to immunization with irradiated sporozoites on the fourth day. Mice were then treated with 200 μg of GK1.5 antibody weekly until sacrifice. For depletion of Thy-1.2 + cells, 200 μg of anti-Thy-1.2 monoclonal antibody (clone 30-H12) was administered every other day beginning three days before immunization for ten days. For reconstitution of Thy-1.2-depleted mice with CD4 + Thy-1.1 + T cells, spleen and lymph node cells were pooled from BALB/c Thy-1.1 + mice and CD4 + cells were enriched by depletion of CD8 + and B220 + cells according to the manufacturer's instructions (Miltenyi Biotec). 3×10 6 enriched CD4 + T cells were then transferred into anti-Thy-1.2-treated mice one day before immunization.
Immunization
Plasmodium yoelii 17XNL sporozoites were harvested from the salivary glands of infected female Anopheles stephensi mosquitoes and irradiated as previously described [41] . For immunization, irradiated sporozoites were suspended in HBSS containing 1% heat-inactivated mouse serum and mice were then immunized with 5×10 4 sporozoites in the skin at the base of the tail.
Quantification of parasite development the liver
For challenge experiments, live sporozoites were injected intravenously to ensure uniform trafficking of parasites to the liver and the development of liver stage parasites was determined. Forty hours after challenge, livers were excised and parasite load was determined by quantitative PCR for P. yoelii 18S rRNA using SYBR Green (Applied Biosystems) as previously described [42] .
Ex vivo stimulation
For analysis of T cell functionality based on cytokine production, lymphocytes were incubated with peptide-coated target cells in the presence of protein transport inhibitors and then stained for intracellular cytokines. Briefly, A20 target cells were pulsed with SYVPSAEQI peptide (2 μg/mL) and control A20 cells were incubated without peptide. Peptide-coated or control target cells were added to effector cells from the spleen or lymph nodes of immunized mice. To measure cytokine production, cells were incubated with brefeldin A (GolgiPlug, BD Biosciences) to block protein transport. To measure degranulation, cells were incubated with anti-CD107a-FITC (clone 1D4B, BD Bioscience) with monensin (GolgiStop, BD Biosciences). Cells were incubated for 4 hours at 37°C and then washed twice in cold media.
Flow Cytometry
All antibodies were purchased from eBioscience unless otherwise noted. The frequency of parasite-specific CD8 + TCR-Tg T cells was determined by staining of single cell suspensions with anti-CD8-APC (clone 53–6.7) and anti-Thy-1.1-PE (clone His51, BD Biosciences), followed by analysis on a BD FACSCalibur (BD Biosciences). For intracellular cytokine staining, cells were stimulated as described above and surface stained with anti-CD8-APC-AlexaFlour750 (clone 53–6.7) and anti-Thy-1.1-PE. Cells were then permeabilized and fixed using a Cytofix/Cytoperm kit (BD Biosciences) according to the manufacturer's instructions and stained for intracellular cytokines using anti-IFN-γ-PE-Cy7 (clone XMG1.2), anti-TNF-α-Pacific Blue (clone MP6-XT22), or anti-IL-2-APC (clone JES6-5H4) at pre-determined concentrations. Cells were then washed and analyzed on a LSR II flow cytometer (BD Bioscience).
Polyfunctional Analysis
Analysis of all FACS data was done using FlowJo software (TreeStar). Boolean combination gates were created and data was exported to PESTLE and SPICE for analysis (both kindly provided by Mario Roederer). For analysis of cytokine production, approximately 2×10 3 CD8 + Thy-1.1 + cells were collected in each sample (individual samples range from 4×10 2 –2×10 4 ).
Statistical Analysis
Mann-Whitney tests were done using Prism 4 software (GraphPad Software). Permutation tests of significance of polyfunctional distributions were done using SPICE software. | Results
CD4 + helper T cells are necessary for CD8 + T cell responses to γ-spz
We have previously shown that CD4 + cells are critical for optimal priming of both endogenous polyclonal CD8 + T cells and antigen-specific TCR-Tg CD8 + T cells following immunization with irradiated P. yoelii sporozoites [5] . In the current study, we further characterized the effects of helper T cells on the development of effector and memory CD8 + T cells and their ability to protect from infection. While depletion of CD4 + cells by antibody treatment clearly limits the development of the CD8 + T cell response to γ-spz ( Figure 1A ), the possible role of CD4 + cells of the non-T-cell lineage, as well that of CD4 + NKT cells required further elucidation. To determine if CD4 depletion was modulating the CD8 + T cell response to γ-spz independently of depletion of helper T cells, we treated mice with anti-Thy-1.2 antibodies to deplete endogenous T cells while leaving the transferred Thy-1.1 + TCR-Tg cells intact. Depletion of Thy-1.2 + cells prior to γ-spz immunization resulted in a markedly reduced CD8 + T cell response ( Figure 1B ), similar to the effects of depletion of CD4 + cells. Moreover, we found that transfer of CD4 + Thy-1.1 + cells into Thy-1.2-depleted mice rescued the CD8 + T cell response to γ-spz immunization ( Figure 1B ). Taken together, these data strongly suggest that CD4 + T cells provide critical support to the development of effector CD8 + T cells in response to γ-spz. It is important, however, that our results cannot rule out that other CD4 + Thy-1 − cells may also have role in modulating the CD8 + T cells responses.
Given that expression of CD4 among T cells is not exclusive to MHC II-restricted T cells, we next tested if CD4 + CD1d-restricted NKT cells played a role in supporting the CD8 + T cell response. We found that CD1d −/− mice, which lack NKT cells, were fully competent to support CD8 + T cell priming to γ-spz immunization ( Figure 1C ). Further, Jα18 −/− mice, which lack NKT cells expressing the invariant TCR, also showed no defect in the CD8 + T cell response. These results extend our previous observations indicating that a subset of NK1.1 + NKT cells are dispensable for CD8 + T cell priming to γ-spz immunization [5] . Together, these data strongly support the notion that CD4 help to CD8 + T cells following immunization with γ-spz is provided by MHC II-restricted CD4 + T cells.
CD4 + T cell help is needed for peak clonal expansion of anti-parasite CD8 + T cells
To gain insight into the nature of the CD4 help, we analyzed the early events of the T cell response to parasite immunization. As early as four days post-immunization, a major reduction in the number of effector CD8 + T cells in CD4-depleted mice was observed in the spleen ( Figure 2A ). In the lymph nodes, a defect was not observed until day 5. During the first week after immunization, we observed similar kinetics of expression of most surface activation markers independent of CD4 help ( Figure S1 ), consistent with normal initiation of proliferative responses and early expansion ( Figure 2A,B ). Strikingly, however, we found that helpless CD8 + T cells failed to express CD25 at any time during the first week after immunization, in contrast to control CD8 + T cells that transiently expressed high levels of CD25 at three days post-immunization ( Figure 2B,C ). Since day 3 post-immunization was the last day that the CD8 + T cell response was indistinguishable between control and CD4-depleted mice, it is possible that lack of CD25 expression may be related to the early contraction of the helpless T cell population. Thus, we next evaluated if helper T cells were dispensable after the early clonal burst of CD8 + T cells and expression of CD25. To test this, we depleted CD4 + T cells beginning on day 4 after immunization and evaluated the number of antigen-specific CD8 + T cells in the spleen at day 14 post-immunization ( Figure 2D ). We found that delayed depletion of CD4 + T cells did not alter the number of CD8 + T cells recovered two weeks later compared to control mice. Together, these data suggest that helper T cells were critical for supporting the size of the CD8 + T cell compartment through mechanisms acting early in the T cell response, potentially by inducing CD25 expression on the anti-parasite CD8 + T cells.
The size of the memory CD8 + T cell pool induced by γ-spz is dependent on helper T cells in a precursor frequency-dependent manner
We next evaluated the role of CD4 + T cells in the memory development of γ-spz-induced CD8 + T cells. At day 35 post-immunization, the number of anti-parasite CD8 + T cells found in CD4-depleted mice was reduced 76% compared to control mice ( Figure 3A ). This relative reduction was similar to the 78% reduction observed at day 10 ( Figure 2A ), suggesting that helper T cells do not affect the size of the CD8 + T cell pool that transitions from effector cells into memory cells and appear to only support the size of the early effector pool. When the number of naïve TCR-Tg precursors was reduced from 2×10 5 to 2×10 4 or 2×10 3 , no differences were observed in the number of memory CD8 + T cells recovered thirty days after immunization from intact mice and only a slight decrease was observed when 2×10 2 naïve T cells were transferred (39% reduced compared to 2×10 5 , p = 0.047; Figure 3A ). However, in the absence of CD4 + cells, the relative recovery of memory CD8 + T cells decreased dramatically with reduced precursor frequency. With ten-fold serial reductions in precursor frequency from 2×10 5 , the size of the helpless memory CD8 + T cell pool was reduced 76%, 79%, 97%, and 99.8% compared to control, respectively ( Figure 3A ). Importantly, the TCR-Tg cells found in CD4-depleted mice at day 30 using higher precursor frequencies (2×10 5 ) did not represent undifferentiated naïve cells, as they uniformly expressed surface markers consistent with effector and memory T cells: CD44 hi , CD62L lo , CD122 + , and CD11a hi ( Figure 3B ), demonstrating both normal T cell differentiation in the absence of CD4 + T cell help and a genuine memory population not composed of residual naïve T cells left unprimed after immunization. Taken together with the results that depletion of helper T cells after day 4 did not affect the size of the effector CD8 + T cell pool ( Figure 2D ), these data suggest that reduced memory pools observed one month after immunization can be attributed to the early effects of helper T cells on CD8 + T cell clonal expansion.
Helpless CD8 + T cells fail to protect from live malaria challenge
While helper T cells clearly supported the size of the effector CD8 + T cell pool, the CD4-dependence for survival was not absolute, as helpless T cells were readily detected at ten and thirty days post-immunization ( Figure 2A , 3A ). Thus, we next evaluated the ability of the helpless CD8 + T cells to protect from malaria infection by challenging immunized mice with live sporozoites. Protection was quantified by measuring parasite load in the liver forty hours after challenge. In control mice that received TCR-Tg cells and γ-spz immunization, the parasite load in the liver was reduced by greater than 99% compared to control mice that received TCR-Tg cells without immunization ( Figure 4A ). Depletion of CD8 + cells just prior to challenge demonstrated that protection was predominantly mediated by CD8 + T cells. Remarkably, the “helpless” CD8 + T cells in mice that were immunized in the absence of CD4 + T cells conferred no protection against subsequent live sporozoite challenge. This effect was not due to loss of protective immunity mediated directly by CD4 + T cells, as depletion of CD4 + cells from immunized mice just prior to challenge had no effect on protection ( Figure 4B ). Thus, CD8 + T cell-mediated control of liver stage malaria parasites is critically dependent on CD4 help.
Development of cytotoxic effector function by effector and memory CD8 + T cells occurs independently of helper T cells
Given that CD8 + T cells primed in the absence of helper T cells conferred no protection against live malaria sporozoite challenge, we next directly evaluated the functionality of these T cells ex vivo . Following adoptive transfer of TCR-Tg cells and γ-spz immunization, lymphocytes from the spleen and draining lymph nodes of control and CD4-depleted mice were harvested at various time points, stimulated ex vivo with peptide-pulsed target cells and stained for intracellular cytokines (IFN-γ, TNF-α, or IL-2) and surface mobilization of CD107a as an indicator of degranulation and cytotoxic capacity [12] , [13] , [14] . Three days post-immunization (prior to the peak of the response) primed T cells had rapidly acquired effector functions, with cells from both normal and CD4-depleted mice displaying robust capacity to degranulate (CD107a + ), produce IFN-γ, TNF-α, or IL-2 in the lymph nodes draining the site of immunization ( Figure 5A ). Combinatorial analysis of polyfunctional effector function demonstrated indistinguishable profiles between CD8 + T cells from control and CD4-depleted mice ( Figure 5B ). Distinct combinations of effector molecule expression were observed (primarily CD107a + IFN-γ + TNF-α + IL-2 + ), with co-expression of 3-4 effector molecules accounting for more than half of the T cells in both groups at three days post-immunization. Similar polyfunctional effector profiles were observed at the peak of the response at day 5 ( Figure 5C ), despite the fact that the helpless CD8 + T cells had already begun a premature contraction. The patterns of cytokine co-expression and degranulation observed here are consistent with those reported in human studies [15] , [16] , [17] , [18] and demonstrate the combination of effector functions exhibited by CD8 + T cells is not stochastic, but appears to follow discrete programs.
Previous reports on CD4-CD8 T cell interactions have demonstrated the manifestations of CD4 help are evident only in recall responses to resting memory CD8 + T cells, when helpless CD8 + T cells display defective cytokine production, killing, and re-expansion [7] , [9] , [11] . Thus, we evaluated the recall response of the memory anti-parasite CD8 + T cells recovered from the spleen and lymph nodes thirty days after γ-spz immunization. Remarkably, we found no differences between control and CD4-depleted mice ( Figure 6 ). Overall, the percentage of T cells producing effector cytokines and degranulating (CD107a + ) ex vivo from both groups was lower than at the peak of the response, but the distribution of co-expression patterns of effector molecules was the same. Similarly, we found no role for NKT cells in supporting CD8 + T cell effector function, as donor TCR-Tg CD8 + T cells from immunized CD1d −/− recipient mice were indistinguishable from control recipients (not shown). In all, these data demonstrate that CD8 + T cells primed in the absence of helper T cells fail to expand robustly, but develop into fully functional effector and memory cells. | Discussion
Memory CD8 + T cells induced by irradiated malaria sporozoites are critical for elimination of liver stage parasites and the optimal development of primary effector CD8 + T cells to sporozoite antigen is dependent on the presence of CD4 + helper T cells [5] . In this report, we have shown that memory CD8 + T cells developed in the absence of CD4 help failed to protect from live malaria challenge. Nevertheless, CD8 + T cells primed without CD4 + T cell help were fully functional by all indices measured, including production of IFN-γ, TNF-α, and IL-2, as well cytotoxic degranulation. This disconnect between protection and T cell functionality suggests that frequency or total numbers of antigen specific CD8 + T cells are likely to be critical factors in determining the ability of CD8 + T cells to control liver stage malaria parasites. Indeed, recent reports on anti- Plasmodial CD8 + T cell immunity have confirmed that large numbers of circulating T cells are necessary to achieve sterile immunity [19] , [20] .
A striking feature of this model of CD8 + T cell priming is the very early time at which CD4 + T cell help is manifested (four days post-immunization), which contrasts much literature that has documented unaltered primary CD8 + T cell responses to pathogens [6] , [7] , [8] , [9] , [10] . The defects in priming to γ-spz without CD4 + T cell help resemble models of priming under non-inflammatory conditions [21] , [22] , [23] , [24] , [25] , [26] , as well as priming to select pathogens [27] , [28] , [29] . Thus, it is unclear what factors govern the CD4-dependency of a given CD8 + T cell response and how different aspects of CD8 + T cell biology are impacted separately. Our observations that without CD4 + T cell help clonal expansion is diminished but effector function is unaltered upon recall contrasts models that demonstrate intact clonal expansion but defective functional recall [7] , [9] , [11] . A variety of mechanisms of CD4 + T cell help have been proposed, including: CD40-CD40L interactions [21] , [22] , [24] , DC “licensing” [29] , cross-presentation [30] , or chemokine-related help [31] , [32] ; though these mechanisms are certainly not mutually-exclusive. In response to irradiated sporozoites, the lack of expression of CD25 (IL-2Rα) by helpless CD8 + T cells is intriguing and suggests that induction of signaling through the high affinity IL-2 receptor is a mechanism of CD4 help, as has been suggested by previous studies [28] , [29] . Indeed, IL-2 secreting CD4 + T cells are associated with slower HIV disease progression [33] and studies evaluating the functionality of anti-HIV CD8 + T cell responses point to a role of IL-2 in supporting recall responses [34] , [35] .
Interestingly, the magnitude of the defect in clonal population size in the absence of CD4 help was enhanced at low CD8 + T cell precursor frequency. These results are consistent with the observation that helpless CD8 + T cells proliferate and expand for only 3–4 days before contraction. Under conditions of progressively lower precursor frequency, the peak of the proliferative expansion is progressively delayed [36] , [37] , thus reducing the total number of effector T cells generated in the first few days after immunization. This reduction may ultimately diminish the size of the surviving memory population, as the size of the memory pool is closely related to the size of the initial clonal burst [38] , [39] . Given the correlation between precursor frequency and relative expression of CD62L on memory CD8 + T cells [36] , [40] , we examined CD62L expression and found no differences in CD62L expression on memory anti-parasite T cells from intact and CD4-depleted hosts, though the relative expression of CD62L on the memory cells was reduced with decreased precursor frequency (not shown), consistent with previous reports [36] , [40] . These data suggest there are not selective differences in the need for CD4 + T cell help of effector and central memory T cells for survival that could account for our observations. Importantly, while changes in precursor frequency has subtle changes in the percent of CD62L + memory T cells, the polyfunctional profile of CD8 + T cells was consistent, regardless of precursor frequency (data not shown). Finally, we found no defect in helpless CD8 + T cell function in the absence of helper T cells when the initial CD8 + T cell precursor frequency was reduced 10-fold to 2×10 4 cells per mouse ( Figure S2 ).
Taken together, our data provide further evidence that CD8 + T cell expansion and functionality are separate facets of T cell programming that can be uncoupled and revealed in the absence of helper T cells. However, contrary to current models that attribute CD4 + T cell help to CD8 + T cell functionality and not clonal size, our data demonstrate that CD8 + T cells specific for the circumsporozoite protein of P. yoelii that are induced by γ-spz require CD4 help to promote optimal clonal expansion, but not effector function. Resolution of the signals that regulate these separate facets are critical to our understanding of T cell biology. | Conceived and designed the experiments: MGO FZ. Performed the experiments: MGO IAC YCC SWT. Analyzed the data: MGO YCC. Wrote the paper: MGO FZ.
Current address: David H. Smith Center for Vaccine Biology and Immunology, University of Rochester Medical Center, Rochester, New York, United States of America
CD4 + helper T cells are critical orchestrators of immune responses to infection and vaccination. During primary responses, naïve CD8 + T cells may need “CD4 help” for optimal development of memory populations. The immunological factors attributed to CD4 help depend on the context of immunization and vary depending on the priming system. In response to immunization with radiation-attenuated Plasmodium yoelii sporozoites, CD8 + T cells in BALB/c mice fail to generate large numbers of effector cells without help from CD4 + T cells – a defect not observed in most systems. Given this unique early dependence on CD4 help, we evaluated the effects of CD4 + cells on the development of functional properties of CD8 + T cells and on their ability to abolish infection. First, we determined that this effect was not mediated by CD4 + non-T cells and did not involve CD1d-restricted NKT cells. We found that CD8 + T cells induced by sporozoites without CD4 help formed memory populations severely reduced in magnitude that could not limit parasite development in the liver. The inability of these “helpless” memory T cells to protect is not a result of defects in effector function, as their capacity to produce cytokines and undergo cytotoxic degranulation was indistinguishable from control memory T cells. These data indicate that CD4 + T help may not be necessary to develop the functional attributes of CD8 + T cells; however they are crucial to ensure the survival of effector and memory cells induced in primary responses. | Supporting Information | CC BY | no | 2022-01-13 08:14:21 | PLoS One. 2011 Jan 4; 6(1):e15948 | oa_package/8e/ac/PMC3014941.tar.gz |
||
PMC3014942 | 21283511 | Introduction
Biologic phenotypes emerge as a consequence of genes interacting through complex networks. Oncogenesis has been shown to be dependent on biologic networks that control processes such as apoptosis, senescence, proliferation, and angiogenesis [1] , [2] . However, it is clear that current knowledge of which processes influence diverse cancer phenotypes is incomplete. This is especially true when it comes to understanding processes associated with disease outcome.
A complex collection of genomic alterations occur during tumor cell evolution, including mutations, translocations, and copy number alterations. For example, genome-wide analysis of breast tumors by numerous techniques have reproducibly demonstrated recurrent patterns of copy number alteration (CNA) [3] , [4] , [5] , [6] , [7] , [8] , [9] , [10] , [11] . The expression of genes within these altered segments has been demonstrated to be correlated with the copy number state of the region [3] , [9] , [12] , [13] , [14] , [15] , [16] , [17] , [18] , [19] . However, it is unclear whether these recurrent patterns represent the most important set of CNAs or represent only a subset of key regions.
Patterns of copy number alteration have proven valuable in classification of cancer subtypes and can serve as predictors of patient outcome [19] . These alterations target genes that influence networks that provide the tumors with a selective advantage over cells of normal composition. Given their association with outcome, it is likely they also influence processes that drive clinical phenotypes and response to interventions.
Identifying the processes targeted by the regions identified through system-wide analysis is complex. For example, copy number-altered regions contain large numbers of genes. There is also a tremendous degree of between-individual heterogeneity in the inventory of regions found to be altered.
Work by others to identify processes underpinning complex traits has combined inherited variants and network analysis to map multifactorial, heterogeneous disease phenotypes [20] . In this work, the authors extend traditional gene mapping approaches by including putative gene interactions to address heterogeneity. Others have examined multidimensional data sets that include different genome-scale measurements simultaneously in the context of pathways [21] , [22] , [23] .. They apply statistical method to measure pathway enrichment and use gene-expression data to assess variation of pathway activity. Through such analyses they hypothesize new cell functions.
In the work presented here, we compliment and extend these approaches to systematically analyze somatic CNAs to identify biologic networks underpinning cancer phenotypes. We demonstrate the method using the breast cancer data set of Chin et al [24] . We identify altered pathways differentially targeted by copy number aberrations.
Similar to previous approaches, we addresse the heterogeneity of patterns by recognizing that differing patterns of CNA may represent alternative routes that cancer cells may take to alter the same core set of common biologic processes. The apparent heterogeneity in map location associated with CNAs may simply reflect the fact that the genes comprising a given network are distributed throughout the genome. We therefore test whether individual canonical pathways are non-randomly targeted across copy number change regions. In contrast to previous approaches, we leverage existing network structure as opposed to de novo creating networks. The network interaction structure for these canonical networks is then leveraged for mapping phenotypes. We utilize previously described methods [25] to determine whether altered state of non-randomly altered processes can predict patient outcome. | Materials and Methods
Mapping Entrez Gene to Golden Path
NCBI's Entrez Gene database contains 36470 human records, 25441 of them annotated as protein-coding. For each gene in this set we used a variety of methods to find its location Golden Path genome sequence. Version (hg18) of the genome database contains extensive annotations which we used wherever possible. In some cases we used BLAT to find genomic locations.
The positions of approximately 18,342 (∼54%) genes were annotated directly in Golden Path's refLink and refGene tables. While this is the most straightforward reference, it leaves 18,128 genes unmapped, 6,757 (∼18.5%) of them protein-coding.
In cases where a direct gene annotation was not available, we searched Golden Path's annotations for the locations of associated sequences from a variety of sources, listed below in order of preference:
mRNA accessions from Entrez Gene's “gene2accession” table
cross-referenced accessions from the HUGO database
cross-referenced accessions from the uniSTS database
primary representative sequence from associated UniGene cluster
mRNA sequences from associated UniGene cluster
EST sequences from associated UniGene cluster
Accessions were gathered from each of these sources in turn, and then looked up in various Golden Path annotation tables (all_mrna, stsMap, clonePos, and all_est). A locally-built database of mRNA and refseq BLAT results (assembled by Robert Clifford) was also searched, providing some additional matches. The resulting genomic locations of the search sequences were aggregated, and accepted as the gene's position if the locations fell within a 3 mb region (3 mb being a somewhat arbitrary cutoff based on the largest observed refLink-based gene mapping of approximately 2.3 mb). If a chromosome annotation was available from Entrez Gene, HUGO, or uniSTS, genomic positions were only included if they were on the same chromosome. A known chromosome annotation was required in the case of UniGene mRNA and EST sequence lookups.
In cases where accession annotations were available but the positions were not found, we performed our own BLAT searches. This was necessary for certain classes of accessions which do not appear in the Golden Path database (e.g. the “XM_” series of predicted refseqs). If a chromosome annotation was available for the gene, a BLAT search was run only against that chromosome, otherwise all chromosomes were searched. Results were aggregated and accepted as the gene's position if they fell within a 10 mb or smaller region. This is a less strict requirement than used in the accession-based mapping system, yet it can provide at least a general position, much more specific than a cytogenetic-based coordinate (the only mapping information available for some Entrez Gene entries). If plausible matches were found on multiple chromosomes, the gene mapping was rejected as ambiguous.
BLAT results are annotated with one of four categories of match types, so the annotations may be excluded later if they are considered too broad. The four categories are:
A single perfect match for the query sequence was found. The ideal mapping result.
More than one perfect match for the query sequence was found.
A single near-perfect match (at least 95% but less than 100% identity) was found.
Multiple near-perfect matches were found.
Preferential treatment was given to perfect refseq matches in the results – i.e. a perfect BLAT match to a refseq was considered the gene's genomic position, regardless of the presence of other near-perfect matches in the results.
If mapping failed by any of the above methods a few crude methods of last resort were attempted:
if a gene was positioned on an NCBI genomic contig sequence(NC_* series accession, via EG's “gene2refseq” table), and a neighboring gene on the same chromosome, arm,
and band could be found in Golden Path, the relative
distance between the two genes in the NCBI sequence
was applied to the Golden Path coordinates to approximate
its position.
If a gene had only a cytogenetic location available, coordinates of Golden Path-mapped genes with the same cytogenetic location were aggregated and a union of their position generated. The resulting mappings are extremely broad but at least point to a general molecular region which may still be useful in some circumstances.
Mapping BACs to Golden Path
The second dataset to be mapped to Golden Path consisted of the set of BACs used in the CGH arrays from Chin et al [24] . As with the Entrez Gene mapping process, the Golden Path annotation database contains an ideal table for our purposes, “bacEndPairs”, holding the genomic positions of BACs whose end sequences have both been mapped. However, only approximately 39% of the BACs in our set contain an entry in this table. The “fishClones” table provided mappings for an additional 6% of the BACs. For the remainder we used BAC-related annotations as a basis for mapping.
The NCBI clone registry provided a major source of BAC annotations. From it, we extracted BAC-related accession, end sequence, STS and chromosome information. The registry also provided cross-connections to uniSTS, from which we gathered additional related accessions. We searched for the resulting sequences in Golden Path's all_mrna, clonePos, stsMap, and all_ests tables. We also took special note of any matches for BAC end sequences. In addition to the clone registry, we also used annotations from the UCSF 2.0 arrays (data from http://cancer.ucsf.edu/array/analysis/ ), as well as GenBank records referencing BAC names in the title block. Genome mappings were accepted for the BACs if they were no longer than 500 kb in length, and mappings to ambiguous chromosomes were rejected.
For BACs which could not be found using NCBI clone registry or UCSF array annotations, we attempted a surrogate-based mapping approach. Chin et als [1] CGH array annotations provided rough genomic positions (in megabases) whose coordinates aligned most closely with an older genome build, hg16. For each BAC, we extracted sequence IDs from hg16 which were annotated as being near this position. Sets of sequences were extracted from each of the all_mrna, stsMap, and all_est annotation tables. For mRNAs and STSs, we used sequences located within plus or minus 5 kb of the target location. For ESTs, we took sequences within plus or minus 1 kb of the target position. These extracted sequences were used as surrogates for the BACs, and looked up in hg18, searching (in order of preference) mRNAs, STSs, and ESTs. This approach was used to generate hg18 positions for approximately 8.7% of the BACs.
For BACs that could not be mapped to hg18 using any of the above methods, a second pass was performed to find generate approximate positions based on interpolated neighboring BAC locations. For each BAC, we tried to find flanking BACs with hg18 mappings. We then applied relative offsets to the hg18 positions based on the spacings in the hg16 positions. This was only required for approximately 1.4% of the BACs.
BAC preprocessing
Two sets of modified genomic positions are generated for each BAC, which we refer to as expanded and extended coordinates.
Expanded coordinates are an attempt to compensate for the many cases where BAC mapping and end-sequence information is incomplete. They are intended to ensure that all BACs cover a minimum amount of the genome, and that fully-mapped BACs do not crowd out BACs having less complete mapping annotations. This involves expanding mapped BAC coordinates up to approximately 165kb, which is our observation of the median size of BACs where both end sequences have been mapped. Coordinates are not expanded in cases where both end sequences have been mapped, or if existing mapping information spans 100kb or more. If a single end sequence mapping is known, the expansion is made away from the anchored end, otherwise the coordinates are expanded equally in either direction. Collisions during expansion between closely-mapped BACs are detected and resolved by a multi-pass process where the available intervening space is assigned equally between BACs. If expansion in one direction causes a collision with a neighboring BAC, appropriate compensatory expansion is attempted in the other direction, unless that end is fixed by the presence of a known end sequence.
Extended coordinates build upon the expanded mappings by dividing unassigned regions of the genome between neighboring BACs. This provides pseudo-tiling coverage of the genome, allowing any given region to be associated with the most appropriate BAC in the set. Generating extended coordinates requires expanded coordinates to be calculated first, to allow the most equitable assignment of intervening regions.
Expanded and extended coordinates are computed dynamically based on the BAC membership of the CGH array being worked with. While the hg16-based CGH arrays were intended to sample the genome at regular intervals, their computed positions in hg18 are not as neatly spaced. For these purposes the BACs were arranged as we observed them in hg18.
There are cases where BAC coordinates overlap. In cases where a BAC is computed to lie entirely within a larger BAC, the smaller BAC receives the same final coordinates as the larger BAC (it is essentially considered a duplicate). In cases where a BAC partially overlaps with another, the coordinates in the overlap region are left unchanged, and no expansion or extension is performed on the end with the overlap.
Associating BACs with genes
There are three basic types of intersections between gene and BAC coordinates:
The gene's mapping falls entirely within the BAC's mapping.
The gene's mapping lies partly within the BAC's mapping and partly outside.
The gene's mapping is larger than the BAC's mapping. This can happen for genes with very broad cytogenetically-derived gene mappings.
Gene-to-BAC associations of the first type are trivial to calculate. The latter two cases require some additional steps to determine whether a gene should be associated with a BAC or not. Associations are generally rejected if the length of the BAC mapping is less than one-third the length of the gene mapping. This prevents associations from being formed based on insubstantial overlaps. If the extended set of BAC coordinates is being used, an association is rejected unless at least 50% of the gene's coordinates lie within the BAC's coordinates. Since in extended mode BACs tile the genome completely, this step ensures that genes in border regions will be assigned to one BAC exclusively. Specific associations of BACs and their genes has been previously described in Chin et al. [24] .
Identifying Genes in Copy Number Altered Regions. In order to identify the genes in the copy number altered regions it was necessary to translating BACs coordinate used in the comparative genomic hybridization (CGH) assays into genome coordinates. This involved mapping the Entrez Gene database and the CGH BACs to a common coordinate space (Golden Path human genome build hg18), and then overlaying the results. These processes are described in the supplemental material [19] .
Mapping Genes to Pathways
We determined the list of genes used in each pathway in by query of the Pathway Interaction Database [49] .
p-value for a pathway's genomic alterations in a specific sample
Each pathway network has been taken as a set of genes. That is, for each pathway, and according to (2.4), we listed the genes which are members of the pathway.
To determine the probability that a pathway is to be hit by exactly k hits, we first calculate the probability that the pathway is randomly hit times. With G genes quantified in a given platform (for example, a platform that covers the entire genome will cover roughly G = 24,000), and N i genes in a pathway i (N i is usually between 10–70 genes) we get: The probability of randomly hitting zero to k i , j genes, given that M j genes are altered in sample j is the hypergeometric cumulative distribution function: The associated p-value is therefore defined as:
p-value for a global pathway targeting across a population
To be able to statistically quantify genomic targeting of a pathway across a population of subjects we need to iterate across the p-values defined in (2.5). This is in effect a combination of one sided binomial tests. This has been solved by different techniques, including Fisher's Omnibus [50] , which we are using here. This test statistics for pathway i is expressed here as: and the corresponding p-value is: where is the Chi-square cumulative distribution function and d are the number of degrees of freedom (number of samples). | Results
Chin et al. have previously reported genome-wide copy number and gene expression analysis of 145 primary breast cancer tumors [19] . These alterations were determined using genome BAC array CGH [26] , [27] , [28] , [29] comprised of 2464 BACs selected at approximately mega base intervals along the genome as described previously [26] , [28] . Utilizing this data set and the process described in Materials and Methods , the gene content of each segment described in Chin et al. was identified.
Canonical biologic network structure information and gene content was obtained from public sources [30] , [31] , [32] .A total of 565 canonical pathways were examined. These pathways represent collections of interactions that are subsets of larger biologic networks curated to capture specific functions. Therefore, their gene content is not unique. The gene content of these pathways ranges dramatically. For example, as the pathway “degradation of the RAR and RXR by the proteasome [33] ) contains only 2 genes while IL12 Signaling Pathway” [34] , [35] , [36] )contains 80.
To account for heterogeneity of gene involvement when analysis is performed using a network model we define a new statistical metric (described in equations (2.5) and (2.6) in Materials and Methods ). Significance for each pathway across samples was assessed using the Fisher's Omnibus [49] and adjusted for multiple comparisons using the Bonferoni method.
Applying the methods to the data provided by Chin et al., we identify pathways in which the genes altered by CNAs are highly significantly over-represented when compared to random expectations ( Table S1 ).
To illustrate the diverse over-representation patterns for a given network we present the CNA events associated with the pathway “CDC25 and CHK1” [37] ( Figure 1 ). In the figure, gene amplification is denoted through a purple square and gene deletion through black squares.
As Figure 1 demonstrates, no single gene within the pathway appears to be the differential target of CNA across the 18 breast cancer samples shown... or when examined across the remaining 127 individuals in the study.
On the other hand, we can see that the pathway, as a unit, is targeted in almost every subject in the panel (the entire panel of subjects for this pathway is included in Table S2 ). Note, the metric (see Materials and Methods ) compensates for pathway size. As such, to obtain a significant p-value, larger pathways need to accumulate a larger number of gene amplifications or deletions.
We next assessed whether the networks identified by over-representation of CNA are associated with disease outcome. Using pathway activity and pathway consistency scores [26] , we clustered the individuals according to their pathway metrics and performed survival analysis. When we stratify the patients to two groups, we can draw the survival curves and check to see if they separate the population in a significant manner ( Figure 2 ).
Iterating over the collection of hundreds of pathways, we find 29 pathways that meet significance criteria of p<0.05 ( Table S3 ). However when adjusting for multiple testing using the Bonferroni method only two pathways significantly targeted by genomic alterations are also highly associated with survival;”“Hypoxic and oxygen homeostasis regulation of HIF-1-alpha” [38] , [39] , [40] , and Glycosaminoglycan degradation [refs].
An alternative approach to adjusting for multiple comparisons for assessing significance is to validate findings those pathways that show marginal significance across data sets. Two public data sets with expression data and disease outcome were selected from the Gene Expression Omnibus database ( http://www.ncbi.nlm.nih.gov/geo ) [41] The first data set (GSE2990) [42] contained 189 individuals. The second (GSE3494) [43] contained 251 individuals. Gene expression in both datasets utilized the Affymetrix platform for determining gene expression state. Of the original 29 pathways observed to be significantly associated with survival in Chin et al. [19] , 8 were observed to be significant in GSE2990 and 8 were observed to be significant in GSE3494. A total of 4 pathways were observed to be significant in all three data sets. Concordance among the datasets is more than would be expected by chance alone. | Discussion
The above results suggest that genes in CNA non-randomly target processes important for oncogenic state. In the work presented here, we provide a means for objectively identifying the biologic processes that may be the target of these alterations. Moreover, the pathways over-represented in these segments show differences in activity and consistency that is related to cancer outcome.
The total number of pathways identified as non-randomly targeted is striking. One possible explanation is the lack of independence of the gene content associated with each pathway. Hierarchical clustering of the pathways utilizing the p-value associated with the non-random targeting ( Table S4 ) confirms that pathways with related names commonly cluster with high correlation (r>0.5, data not shown). Inspection of the pathway p-values across individuals shows tremendous variability ( Table S4 ). This suggests diverse underlying molecular mechanisms driving oncogenesis. Unfortunately, no obvious pattern of clustering of individuals emerges from analysis of pathway-specific variability.
CNA have been previously demonstrated to show association with patient outcome [44] , [45] , [46] , [47] ). In the Chin et al. [19] individual copy number altered segments showed association with survival and disease recurrence, but performed unevenly. When taken as a set, they found that alteration of any of what they identified as “recurrent amplicons” was associated with reduced survival duration (p<0.04) and distant recurrence (p<0.01).
The results obtained from pathway-based analysis of the same data set produce a striking improvement and suggest that pathways may represent a better way to evaluate recurrent alterations. Two pathways show a highly significant association within Chin et al. alone and 4 pathways show significance across multiple data expression datasets. Because of the high dimensionality of systems-wide data, there is always a danger of over fitting. As such, results from an individual study should be viewed skeptically. However, the significant concordance across multiple provides independent validation.
The increased reproducibility and magnitude of the effect associated with pathway state compared with that observed in the direct examination of “recurrent” regions may be attributable to several factors. At a mechanical level, examination of data at the pathway level permits the information from different regions to be integrated across the network. The fact that any given recurrent region is amplified is no longer the critical predictor. What emerges instead is the importance of sets of altered regions whose individual members hit different parts of a targeted pathway. Pathways pre-aggregate the effects of multiple genes. As such, it is possible to detect multigene interactions that influence cancer phenotypes but which, if not aggregated in a pathway, might fail to meet the test of statistical significance in a small dataset.
CNA is only one factor that could be driving pathway involvement in phenotypes. Many other genomic mechanisms (e.g. individual gene mutations, epigenetic activation/silencing) can influence the state of the pathway. As such, the pathways identified here represent a subset of those likely involved.
Conceptually, it is likely that because the pathway is the underlying unit of the phenotype, focusing on pathways increases signal and reduces noise. Genomic alterations that accumulate during oncogenesis and disease progression occur at random. The observed coherence likely arises because certain processes must be altered to arrive at the given phenotype. Apparent genomic heterogeneity, “noise”, arises because there are multiple ways a pathway can be changed. All of these ways are “signal” from the perspective of a pathway.
It is possible to speculate that analysis similar to those performed for copy number alteration to pathway (above) may prove useful for other genome analyses such as genome-wide mutational screens or association studies. For example, the complex mutational patterns seen in the 1672 genes characterized in human and breast cancer [48] are all observed to mutate genes in one or more of 6 canonical pathways state identified from gene expression data which universally differentiates tumor from normal [25] . Similarly, complex, low odd-ratios haplotype associations patterns may reflect heterogeneous routes to alter common pathways. The above observations have several practical implications in considering next-generation intervention strategies. First, the networks provide a basis for designing combinatorial therapies. Examination of the networks, and their activity states, provides a rational means of determining which combination of genes need to be targeted in order to alter the state of critical nodes. It is also interesting that not all alterations in pathways states influence outcome. This observed difference in effect on outcome, which may reflect the result of natural experiments by the tumor, may also prove important in prioritizing which genes and interactions might be most productively targeted to improve outcome. | Conceived and designed the experiments: SE CS KHB. Performed the experiments: SE KHB. Analyzed the data: SE RBH ME SG CS KHB. Contributed reagents/materials/analysis tools: KHB. Wrote the paper: SE CS KHB.
High resolution, system-wide characterizations have demonstrated the capacity to identify genomic regions that undergo genomic aberrations. Such research efforts often aim at associating these regions with disease etiology and outcome. Identifying the corresponding biologic processes that are responsible for disease and its outcome remains challenging. Using novel analytic methods that utilize the structure of biologic networks, we are able to identify the specific networks that are highly significantly, nonrandomly altered by regions of copy number amplification observed in a systems-wide analysis. We demonstrate this method in breast cancer, where the state of a subset of the pathways identified through these regions is shown to be highly associated with disease survival and recurrence. | Supporting Information | The authors wish to thank Dr. Liran Carmel for his help with the manuscript. | CC0 | no | 2022-01-13 08:14:21 | PLoS One. 2011 Jan 4; 6(1):e14437 | oa_package/8b/20/PMC3014942.tar.gz |
|
PMC3014943 | 21245910 | Introduction
Throughout the last decade several large-scale preventive chemotherapy campaigns, waged against neglected tropical diseases, have progressively scaled up operations to reach nationwide coverage levels in Uganda [1] , [2] . For control of intestinal schistosomiasis, as caused by Schistosoma mansoni infection, an active monitoring and surveillance programme, set within the national control programme (NCP), has provided important disease-specific information, assessing the impact of treatment upon the recipient population, as well as, re-alignment of original control objectives first set forth in 2003 [3] , [4] .
Following WHO guidelines, mass-drug administration of praziquantel (PZQ) is typically focused towards treatment of school-aged children (≥6 years) and adults who reside within disease endemic regions [5] , [6] . PZQ is provided free of charge by the NCP and analysis of school and(or) community treatment registers has shown that several million people have received at least one annual treatment of PZQ within the last five years [1] , [7] . Although this represents a considerable achievement, targeted epidemiological surveys have revealed that coverage is incomplete as in certain areas, e.g. shoreline environments of Lakes Victoria and Albert, large numbers of preschool-aged children (≤5 years) and infants (≤1 years) are infected with S. mansoni and have been largely overlooked by the treatment campaign [8] , [9] , [10] .
To ensure that this unfortunate health inequality does not persist the treatment needs of younger children are being assessed and we have recently called for formal inclusion of these young children within the Ugandan NCP [11] . It can be safely assumed, for example, that mass-treatment initiatives are vital in most in shoreline villages where infections can be common. Given the geographical focality of schistosomiasis and itinerancy of lakeshore communities, however, an important future challenge for the NCP is collection of sufficient disease-specific information to better tailor local drug needs and set parameters for subsequent programme monitoring [12] , [13] . Attention will therefore focus upon those sections of villages where young children are frequently bathed in freshly drawn lake water or are within range of regular ambulation to the lake margins.
Owing to the unique natural history and developmental biology of schistosomes within the mammalian host [14] , accurate identification of infected cases is challenging [15] , even more so in the younger child where the founding worm population has only recently established and begun to mature. Before female worms develop their full egg-laying capacity, sporadic deposition of eggs may take place with a proportion of these being voided into the bowel lumen and ejected in faeces whilst the remainder become trapped within the host's tissues [16] . Interacting with this are also the beginnings of the child's innate and adaptive immune responses to excretory-secretory products of the worms themselves, as well as these responses being primed or modulated by maternally induced effects, for example, during pregnancy and(or) breastfeeding [17] , [18] , [19] , [20] . It is also of particular note that the child's immune system is in a maturing flux of recognition between self- and non-self epitopes [21] and the efficacy of PZQ, which is poor against immature worms of S. mansoni [22] , is only starting to be explored in this ageclass [11] . From a general diagnostic perspective as existing tools are sub-optimal, improvement of methods and techniques for detection of intestinal schistosomiasis continues [15] but in the context of the younger child, it is not yet clear which of the present methods, or combinations thereof, is either most appropriate or applicable for routine use within the NCP.
We therefore report on a field-based study which attempted to determine the age of first infection in very young children with available techniques and also estimate, as accurately as possible, the general prevalence of intestinal schistosomiasis within this ageclass from a typical lakeshore community. The performance of methods that detect schistosome - antigens in urine, antibodies to egg antigens in serum and eggs in stool - was compared. For ease of comparison, our methods are subsequently referred to as: an antigen detection method (ADM), an indirect egg detection method (IEDM) and a direct egg detection method (DEDM), respectively. | Materials and Methods
This field study was carried out in April 2009 in Bugoigo on Lake Albert (GPS co-ordinates, N 01° 54′.481′′, E 31° 24′.597′′), a fishing village impoverished both in terms of sanitation and hygiene that has been the location of several previous research/control studies on intestinal schistosomiasis [23] , [24] , [25] , [26] . Prevalence of infection within local school-aged children has been continuously high (>50%) despite annual chemotherapy [27] and infections in infants and preschool children were first formally recorded in July 2007 [11] .
Study location and participants
Owing to itinerancy, the exact number of inhabitants in Bugoigo is not precisely known but is likely in the region of several thousand. The village contains up to three thousand traditional hut dwellings which stretch 3–4 km along the lakeshore and up to 1–2 km inland. Sanitation and hygiene in this village is minimal with few potable water sources and insufficient pit latrines. Household water is typically drawn directly from the lake at specific collection points and then taken back to each homestead in plastic jerry cans for subsequent domestic use. These lakeshore margins, like elsewhere on Lake Albert, provide conducive aquatic habitats for Biomphalaria spp., the intermediate snail hosts of S. mansoni , and can be found throughout the year, although infected snails vary in numbers seasonally [28] , [29] .
The immediate and longer-term objectives of this study were explained to the local community mobiliser who identified a total of 134 mothers that were willing to participate, bringing up to two of their infants/preschool children (≤5 years of age), and attend the two-day clinic commencing on the following day. After obtaining written informed consent from each mother on her own behalf and on behalf of her child(ren), urine, stool and fingerprick blood samples were obtained from all participants on the first day of the clinic. Mothers were then asked a suite of detailed questions recording their demography and water contact behaviours (the questionnaire is available upon request to the corresponding author).
After receipt of the second-day stool (and urine sample), all participants, regardless of their infection status, were treated for schistosomiasis and soil-transmitted helminthiasis with PZQ (40 mg/kg) (CIPLA, Mumbai, UK) and 400 mg albendazole (GSK, Uxbridge, UK) under medical supervision in conditions typical of mass-drug administration [30] . For smaller children, a chewable albendazole half-tablet (200 mg) was given and PZQ tablets were first crushed in orange juice before being administrated by spoon-feeding by their mother under supervision. The diagnostic findings for schistosomiasis here are reported for the children only.
Schistosome antigens in urine (ADM)
Each child's urine sample was visually inspected for macro-haematuria/turbidity and a random sample was tested for micro-haematuria with Hemastix (Bayer, UK) to exclude the possibility of urinary schistosomiasis or other active urinary tract infections. A 50 μl aliquot was then tested for the presence of schistosome circulating cathodic antigen (CCA) using a commercially available lateral flow immuno-chromatographic urine dipstick (Rapid Medical Diagnostics, Pretoria, RSA) originally developed in Holland [31] . On a subset of 90 children, urine-CCA tests were performed in duplicate to assess variation between dipsticks.
To facilitate better recording of the visual intensity of the CCA reaction band within the test zone, results were visually graded against a reference chart for: trace, single (+), double (++) and triple (+++) positive reactions [32] . When creating binomial variables to depict infection status according to CCA, two variations were taken into account: the first considering trace results as negative infection status and the second considering trace results as positive infection status. The urine CCA reagent strip is referred to as an ADM (antigen detection method) from now on.
Antibodies to soluble egg antigens (IEDM)
A commercially available ELISA kit (IVD Inc.; Carlsbad, USA) was used to test for host antibodies (IgG/M) to soluble egg antigens (SEA) according to manufacturer's instructions. Approximately 75 μl of finger-prick blood was taken from each child and serum was harvested, then diluted 1∶40 with specimen dilution buffer before loading a total of 100 μl into each ELISA microwell [11] . Positive and negative control sera were included on each batch of testing. Upon completion, each ELISA plate was placed on a white card and the colour within each microwell (ranging from colourless to yellow) was recorded by visual inspection. Positive reactions were classified either as trace (faint yellow), single (+, light yellow), double (++, yellow) or triple (+++, dark yellow) upon visual comparison with the control sera. The SEA-ELISA is referred to as an IEDM (indirect egg detection method) from now on.
Direct egg-detection methods in stool (DEDM)
Three parasitological methods Kato-Katz, percoll and FLOTAC, henceforth referred to as direct egg detection methods (DEDMs), were attempted on each stool specimen to visualise eggs. However, owing to the differing amounts of stool required for each technique, it was not always possible to assemble a complete data set for every child with each of these three methods.
Duplicate Kato-Katz (K-K) thick smears (41.7mg) were made from first and second day stool samples ( N = 242 children) [33] . The four faecal smears were each examined under the microscope at x100, schistosome eggs were counted and later expressed as eggs per gram (epg) of faeces. Infection intensity was classified as light (1–100 epg), medium (101–400 epg) and heavy (>400 epg) infections according to WHO guidelines [5] . The methodology of Eberl [34] using sedimentation of schistosome eggs by centrifugation through a solution of percoll (Percoll 77237 (1.130 g/ml), Fluka, Sigma-Aldrich Chemie GmbH, Switzerland) was also implemented on-site to visualize eggs ( N = 96 children on first day stool). The egg-floatation procedure known as FLOTAC [35] was performed off-site back in Kampala on a formalin-fixed stool specimen archive ( N = 191 children taken from the second day stool) whereby schistosome eggs are collected by floatation centrifugation through a solution of zinc sulphate at specific gravity of 1.35.
Data handling and statistical analyses
Data were collected from each individual using pro-forma data sheets, which were then transferred into electronic format using Microsoft Excel. The data thus collated were analysed using MS Excel and R statistical package version 2.8.0 [36] . For prevalence data and diagnostic parameters, 95% confidence intervals (CI 95 ) were estimated using the exact method [37] . Prevalence comparisons were performed using (one-tailed) Fisher's exact modification of the 2×2 chi-squared test [38] . For infection intensity values, the arithmetic mean of positive cases was chosen as the measure of central tendency. Data from the FLOTAC and percoll methods were analysed by combining with K-K results and revising the diagnostic criterion so individuals were considered positive if an egg was detected by at least one DEDM.
The diagnostic performances of the ADM (including and excluding trace reactions as a positive diagnosis) and IEDM were tested qualitatively as a rapid diagnostic for intestinal schistosomiasis, considering DEDMs as the ‘gold-standard’. Additionally, a third ‘gold standard’ was created using data from the ADM (including and excluding trace reactions as positive diagnoses) against which to test IEDM data ( N = 242). Diagnostic sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) were calculated according to the different ‘gold standards’ [38] . The diagnostic powers of ADM and IEDM were calculated using all individuals, and then segregated by sex or age (≤3 years of age versus >3 years of age). P -values <0.05 were considered indicative of statistical significance [38] .
Ethical approvals
Approvals for this study were granted by the Ugandan Council for Science and Technology and the London School of Hygiene and Tropical Medicine (application numbers 06.45 and 5538.09). After sensitisation of the local community to the study objectives, verbal assent was first requested from each mother which was then formalised upon written informed consent (for her and behalf of her child), as either a thumbprint or signature on data recording sheet. This was witnessed by a Vector Control Division Officer. PZQ treatment (40 mg/kg) was offered to all study participants irrespective of their infection status. | Results
A complete data set for the ADM, IEDM and Kato-Katz examinations was obtained from a total of 242 children (134 boys: 108 girls, mean age 2.9 years, minimum 5 months and maximum 5 years). However, owing to insufficient amounts of stool available, the FLOTAC and percoll methods could only be performed on 191 and 96 children, with the former and latter finding 4 and 2 additional egg-positive cases, respectively.
Estimating infection prevalence
The prevalence of intestinal schistosomiasis estimated by each diagnostic method, and combinations thereof, is shown in Table 1 and Fig. 1 . Prevalence inferred by DEDM, ADM (including trace reactions as positive diagnoses) and IEDM (considering traces as negatives) were: 24.4%, 42.6% and 45.9%. Of the children who were egg-positive by K-K, three quarters had ‘light’ intensity infections. Girls were equally as likely as boys to be diagnosed positively for intestinal schistosomiasis by ADM (Odds Ratio (OR) = 1.06, p = 0.90) and DEDM examinations (OR = 0.72, p = 0.29). Children under the age of three, however, were less likely to be positive by ADM (OR = 0.51, p = 0.016) or by DEDM (OR = 0.26, p<0.0001) than their older counterparts. The prevalence of positives by IEDM was 45.9%. General prevalence inferred by pooling DEDM and IEDM was 47.7%, with no further change in prevalence when ADM was then added, see Fig. 1 . There was no discordance between duplicate CCA testing for negative or positive classifications (data not shown).
Ages of becoming first positive
The age of first positive (AFP) for each method is presented in Table 1 . For DEDM, the youngest child with eggs in stool was 9 months old, with medium and heavy infections found at 3 and 5 years of age, respectively. For ADM, trace reactions, single, double and triple positives were found in an ascending series of 6 months, 9 months, 11 months and 2 years of age, respectively. For IEDM, trace reactions began at 5 months of age while single, double and triple positive reactions were found in children as young as 6 months, 1 year and 9 months old, respectively. All tests concur on a mean age of first infection within the third year of life. ADM detected infections slightly ahead of I/DEDMs (3.2 years v. 3.4 years v. 3.7 years, respectively). The order of this temporal series is largely concordant with an absolute minimum age of becoming first positive.
Cross-tabulations of diagnostic scores
In the absence of a genuine ‘gold standard’ where the infection status of each child is precisely known, it is necessary to explore relationships between diagnostic scores and infection intensities empirically, and to cross-tabulate diagnostic permutations by investigation. There was negligible variation in diagnostic performance of all protocols tested when classifying the data according to sex and age (data not shown) and general trends were reported from now on.
Plotting the relationship between ADM and DEDM revealed some immediate trends, Fig. 2 . Whilst there were children positive for ADM who were egg-negative, as the epg increases there was a corresponding increase in the proportion of positive ADM tests and once medium/heavy intensity infections were reached, all ADM tests were clearly positives, see Fig. 2A . Plotting the faecal epg of each child against the intensity of the corresponding ADM test further revealed this positive association, see Fig. 2B .
Considering the relationship between IEDM and DEDM revealed similar trends, see Fig. 3 . Despite some children being positive for IEDM while being egg-negative, as the faecal epg increases there was a corresponding increase in the IEDM reaction strength, with all medium/heavy intensity infections diagnosed as clear strong positives ( Fig. 3A & B ).
The relationship between ADM and IDEM was less clear-cut. Children who were ADM negative or trace had a median negative (or trace) IEDM reaction, but the proportionate increase of ADM positives with rising IEDM designations of positive (+) or strong positives (++/+++) was not as great as that seen with DEDM. For example, nearly 40% of children who were IEDM strong positive elicited a negative ADM reaction, Fig. 4A . As the intensity of the ADM result stepped up towards double and triple positive reactions, this typically corresponded better with increasing IEDM classifications, Fig. 4B .
Using available data it was possible to conduct an exploration of diagnostic performances of the ADM and IEDM versus DEDM and against each other ( Table 2 ). First, considering an ADM trace reaction to be an infection negative and comparing with positive diagnosis by at least one of the DEDMs, the ADM had a sensitivity of 59.3%, specificity of 95.6%, PPV of 81.4% and NPV of 87.9%. When considering an ADM trace reaction to be a positive infection, and comparing to diagnosis by at least one of the DEDMs, ADM had a sensitivity of 81.4%, specificity of 69.9%, PPV of 46.6% and NPV of 92.1%.
The IEDM when compared to diagnosis by DEDM, demonstrated a sensitivity of 93.2%, specificity of 69.4%, PPV of 49.5% and NPV of 96.9%. For details on the performance of the ADM or IEDM compared to diagnosis by all DEDMs (using a subset of the data), and for CI 95 around each value, see Table 2 . | Discussion
With limited access to safe water sources, and high levels of local transmission of S. mansoni, conditions in Bugoigo are particularly conducive for young children to acquire S. mansoni infections, and from a very early age. Approximately half of our children had intestinal schistosomiasis. As might be expected, regardless of techniques used, there was an obvious positive association between increasing diagnostic patency of infection with increasing age of the child. Presumably this was resultant from a progressive temporal accumulation of antigens, eggs and antibodies. Congruence between diagnostic methods became most apparent in children between 31⁄4–33⁄4 years of age, broadly consistent with the overall mean age of infected children within our sample. Prevalence of intestinal schistosomiasis in children under 3 years of age, however, was 35.5% (CI 95 27.9–43.8%) and other studies have also revealed that schistosomiasis in very young children can be common [9] , [11] , [39] .
While egg excretions of these children were of ‘light’ intensity, such infected children will not normally receive praziquantel treatment until they have either entered primary school or if the NCP now formally extends its treatment remit to include this ageclass. Thus an infected child could therefore wait up to 3–4 years before receiving first treatment, and with this may have already entered a more ‘chronic’ stage of disease [40] , [41] , [42] , [43] . For example, earlier clinical and ultrasound studies in Uganda in children aged 6 and above, have shown significant hepatosplenomegaly (i.e. putative morbidity from intestinal schistosomiasis), and while they have not yet developed pipe-stem liver fibrosis, up to 15% can have diffusely echogenic livers with pocketed foci, typical of image pattern B (‘the starry sky’ classification) [3] , [27] . Without medication, it is likely that these preschool children will progress towards ‘moderate’ infection intensities before they become of school age. This might better explain the observations of Balen et al . that many adolescent Ugandan children have surprisingly severe intestinal schistosomiasis [44] . Although it is not yet proven that infection in very early childhood leads to heightened morbidity in later childhood and adolescence, this scenario appears plausible.
From animal models, it is known that only a fraction of penetrating cercariae successfully migrate to, and later mature in, the hepatic portal system. After adult worms reach full fecundity, schistosome eggs can be found in stool around 6 weeks after cercarial exposure and it is commonly held that females of S. mansoni produce up to 100–300 eggs per day, although many fail to be voided into the faeces [16] . Given the insensitivity of DEDMs in stool [34] , it is not surprising that false negatives are inferred and the low egg-detection threshold(s) likely contribute to the longer apparent lag of 7–8 months between infection and egg-patency apparent between experimental schistosomiasis and the situation in the field. Moreover, it should be noted that the relationship between excreted eggs in stool and worm burdens is not always straightforward [45] and that infected laboratory animals are typically exposed with a single substantive dosing of cercariae. By contrast, and in this natural setting, exposure and infection is likely a more gradual process, i.e. the so-called trickle infection dynamic [46] , and our children are at least two orders of magnitude greater in body size than most animal models.
An age of first infection?
While some children were patently infected during the first year of life, others were not. Thus a sub-set of children exists with increased infection risk factors which we explain by the following synopsis. As children are born throughout the year, in a largely asynchronous fashion, whilst their initial age of first exposure to unsafe water might be broadly similar (i.e. within first few months of life as mothers begin to bathe them in jerry-can collected water or in the lake directly) their accumulated infection risk will not be equivalent owing the seasonality of local transmission factors and their particular timeframe of exposure within it contingent upon their mother's infant bathing and domestic water drawing practices [11] .
Estimating this accumulated risk of infection reliably over the seasonal time frame of potential exposure is problematic as day-to-day variations within water collection times, its storage and actual domestic use (within each household) introduce many stochastic processes. Estimating cumulative infection risk is therefore easily confounded but an ad hoc investigation of infection risk associated with jerry-can collected water in June 2009, however, has confirmed that sentinel laboratory-bred mice could become infected to freshly drawn water [28] . Seasonal patterns, which operate in umbrella fashion over and above these specific-exposure patterns, no doubt effect this asynchronous age of first infection. Thus there will be no ‘absolute age’ of first infection but rather a ‘range of ages’ depending upon these intricate covariates of exposure. Only after a child has passed through sufficient ‘windows of exposure’, their probability of infection rises to an eventual certainty, after which, it is incumbent on the diagnostic tools to capture their parasitological status as accurately as possible.
Comparison of diagnostic scores
From first appearances the ADM looks to best capture and identify infections in early stage, especially when we consider trace results as putative infection positives. A contentious issue in the use of the CCA reagent strip has been the interpretation of the exact diagnosis of this ‘trace’ result which can be confounded by non-specific inflammatory factors or breast-feeding [32] , [47] . Interpretation of ‘trace’ is more contentious when surveying children under three years of age, where worm burdens are presumably lower than what might be expected in their school-aged counterparts. Interestingly, the percentage change in prevalence estimated according to the CCA reagent strip when excluding and including trace results as a positive diagnosis is significantly larger in the very young children (≤3 years of age) –10.1% v. 36.2% (+358%) – than in those aged four and five years of age –30.1% v. 52.7% (+175%) which is fitting with our understanding of increasing worm burdens through time. Thus we postulate that using ‘trace’ as positive firmly points towards a future use of the urine CCA-dipstick as an early indicator of infections which are as yet to become egg- or antibody-patent. It is particularly notable that the prevalence based on the ADM, when considering trace as positive, is very close to that of IEDM ( Fig. 1 ), yet the diagnostic performance with it was not particularly congruent (see Fig. 3 and Table 2 ) so we still have an incomplete understanding of this infection progression. The dynamics of other ADM have been explored elsewhere in the context of recently acquired infection but not in very young children [48] .
The ADM showed very promising diagnostic performance and robust field performance with high sensitivity and NPV scores (83.9% and 85.3%, respectively) when we considered trace results as a positive diagnoses and very high specificity and PPV scores (95.5% and 90.5%, respectively) when we considered trace results as negative diagnoses. This bimodal use of the test criteria could be advantageous from a control perspective. For instance, if a confident estimate of the suspected occurrence of infections within a population is needed, one should consider trace results as positives. On the other hand, to monitor the prevalence of ‘actual’ infection, or rather more easily identify those who do not, one should consider trace results as negatives. The former would be important if treatments were to be given out en masse as triggered by exceeding an aggregated local prevalence threshold while the latter would be important if treatment were to be withheld in an individual patient setting on the basis of test and treat.
Towards promotion of safe water
With the gradual rise of infection prevalence in older children (over and above our asynchronous infection hypothesis), this trend must represent the spread of several risk factors, rather incipiently, across our cohort. Aggregation of infections in schistosomiasis is well-known [49] but it would be interesting to establish why approximately half of our study cohort had no evidence of infection despite living within the same village. As we were insufficiently aware of the exact locations of sampled households within the village, this could simply represent a cryptic spatial micro-patterning (i.e. these children who live slightly further away from the lake have less contact with viable cercariae) so we are now undertaking fine scale mapping of these individual households with GPS units. If, however, other causal factors could be identified and, perhaps more importantly, were these amenable to manipulation, it could lead to future infection mitigation measures.
Presently, within the NCP there are no health education materials targeted towards these mothers and their young children. More importantly and in terms of policy realignment of the NCP, a useful formal recommendation would be to initiate cross-sectorial activities with water and sanitation NGOs to improve immediately the domestic water quality at Bugoigo and elsewhere along the Lake Albert shoreline. Rather than focusing upon expensive infrastructure development, it could be achieved by introduction of simple water storage or modification measures. For example, as schistosome cercariae are an ephemeral larval stage, freshly drawn water can be rendered harmless for schistosomiasis by simple resting for 24 hrs, by crude filtration or by introduction of mild disinfectants [16] . Thus without initiating a better dialogue with these women of children bearing age through better public health education, mothers will remain sadly ignorant of the risks that making use of this unsafe water has for themselves and that of the future health of their child [11] . In this dialogue, the NCP should be receptive to explore which infection mitigation measures are best feasible and, by this token, help to provide safe water for domestic use which is well-received, implementable and effective.
Epidemiological indicators and treatment needs
It is evident that infants and preschool children in Bugoigo, and other similar lakeshore villages of Uganda [12] , are living in need of treatment. However, addressing how these children could be best identified is not yet clear, as are epidemiological parameters which should be collected for estimating treatment needs and also impact assessment. For example, should mass-treatment of all infants/preschool children take place when a sub-sample of an equivalent age range has been proven to be infected, or should treatment be allocated based at an individual level using the result of a diagnostic test in a ‘test and treat’ setting?
It is outside the remit of this present paper to make a cost-effectiveness calculation but a clear drawback of the IEDM method is that, whilst initially useful to establish if a child is infected (and there is no evidence in these data to suggest a passive maternal transfer of antibodies has been confounding), monitoring this parameter after treatment will be largely uninformative owing to residual antibody titres remaining after infection has putatively cleared. Thus IEDM is only useful at baseline but as an initial estimate of infection prevalence could be powerfully applied in identification and selection of villages, or sentinel locations, to first define the extent of the problem at intervention baseline. This of course assumes the majority of examined children can mount an antibody response and is not confounded by high levels of immune-suppression, by HIV for example, which is likely high in these fishing villages.
In contrast, both ADM and DEDMs have the potential ability to better track the dynamics of worm populations after treatment [22] , [34] but the insensitivity of DEDMs is of particular concern. Put simply, numerous adult worms may reside within the host but are yet not depositing sufficient eggs to be visualised in stool on the day of sampling. Thus through lack of alternatives a pragmatic way forward would be to focus upon more widespread application of the ADM. The advantages of the ADM have been discussed elsewhere in the context of programmatic monitoring [14] but the future challenge will be for the NCP to meet the financial costs of using these rapid diagnostic tests in scale-up of operations. This is particularly true if these are to be used in a ‘test and treat’ setting when large numbers of tests would be utilized [14] . Given the low price of PZQ treatment, to maintain an affordable diagnosis versus treatment differential, a rational strategy would be to examine a sub-set of children and if local prevalence exceeded a given threshold, mass-treatment is advised. Such a strategy is presently within the resources available to the NCP but best sample sizes and prevalence thresholds remain to be determined. | Conceived and designed the experiments: J. Stothard, J. Sousa-Figuereido, N. Kabatereine. Performed the experiments: J. Stothard, J. Sousa-Figuereido, M. Betson, M. Adriko, M. Arinaitwe, C. Rowell, F. Besiyge. Analyzed the data: J. Stothard, J. Sousa-Figuereido, M. Betson. Contributed reagents/materials/analysis tools: J. Stothard, N. Kabatereine. Wrote the paper: J. Stothard, J. Sousa-Figuereido, M. Betson, M. Adriko, M. Arinaitwe, C. Rowell, F. Besiyge, N. Kabatereine.
Background
In Uganda, control of intestinal schistosomiasis with preventive chemotherapy is typically focused towards treatment of school-aged children; the needs of younger children are presently being investigated as in lakeshore communities very young children can be infected. In the context of future epidemiological monitoring, we sought to compare the detection thresholds of available diagnostic tools for Schistosoma mansoni and estimate a likely age of first infection for these children.
Methods and Findings
A total of 242 infants and preschool children (134 boys and 108 girls, mean age 2.9 years, minimum 5 months and maximum 5 years) were examined from Bugoigo, a well-known disease endemic village on Lake Albert. Schistosome antigens in urine, eggs in stool and host antibodies to eggs were inspected to reveal a general prevalence of 47.5% (CI 95 41.1–54.0%), as ascertained by a positive criterion from at least one diagnostic method. Although children as young as 6 months old could be found infected, the average age of infected children was between 31⁄4–33⁄4 years, when diagnostic techniques became broadly congruent.
Conclusion
Whilst different assays have particular (dis)advantages, direct detection of eggs in stool was least sensitive having a temporal lag behind antigen and antibody methods. Setting precisely a general age of first infection is problematic but if present Ugandan policies continue, a large proportion of infected children could wait up to 3–4 years before receiving first medication. To better tailor treatment needs for this younger ageclass, we suggest that the circulating cathodic antigen urine dipstick method to be used as an epidemiological indicator.
Author Summary
In sub-Saharan Africa, intestinal schistosomiasis is a debilitating disease caused by a worm infection. To arrest disease progression, de-worming medications are given out, often en masse , to school-aged children. In Uganda, however, much younger children can be infected, and in lakeshore communities both infants and pre-school children can already show signs and symptoms of intestinal schistosomiasis. To change de-worming practices, further information on the occurrence of infections in these younger is needed for evidence-based decision making. Our study applied current methods of disease diagnosis to better define the ‘age of first infection’ and estimate general infection prevalence within a disease-endemic village. Up to 50% of young children were clearly shown to have schistosomiasis and could likely wait up to 3–4 years before obtaining first treatment if present de-worming policies are not changed. In the context of identifying future treatment needs, we propose that antigen detection methods are most suitable. | Supporting Information | We would like to thank several other members of the VCD field team for their help and assistance during the survey: Aaron, Andrina, Annet, Aida, David and Fiddi, as well as, local VCD assistants at Bugoigo camp: Chris, Perez, Caesar and Ashuman. We are grateful for the time and effort our mothers took to attend our clinic. JRS and MB are grateful to the staff of Giuseppe Cringoli and Laura Rinaldi's laboratory for their assistance and hospitality in Naples for training in use of FLOTAC. We thank David Rollinson for his helpful comments which improved this manuscript as well as those from the anonymous referees. | CC BY | no | 2022-01-13 05:08:15 | PLoS Negl Trop Dis. 2011 Jan 4; 5(1):e938 | oa_package/73/84/PMC3014943.tar.gz |
|
PMC3014944 | 21245911 | Introduction
The trematode liver fluke, Fasciola hepatica , along with Fasciola gigantica are the causative agent of fasciolosis, a foodborne zoonotic disease affecting grazing animals and humans worldwide. The infective metacercariae are ingested by the definitive host where they subsequently excyst in the duodenum. The juvenile fluke migrate to the liver to mature before entering the host bile ducts [1] . Fascioliasis, liver fluke disease, causes annual losses of more than US$3000 million to livestock production worldwide through livestock mortality and by decreased productivity via reduction of milk, wool and meat yields [2] . F. gigantica is one of the most important helminth infections of ruminants in Asia and Africa and is most prominent in poorer regions impacting on individual and small farming communities; it inflicts significant losses in cattle, buffaloes, goats and sheep and in India, infection levels can reach 55% in isolated regions [2] . Fasciolosis is a particularly heavy burden in the agricultural based economy of the developing world including India.
F. hepatica is also a re-emerging worldwide zoonosis, with estimates of between 2.4 and 17 million people infected worldwide and a further 180 million at risk [3] , [4] , [5] , [6] . Climate changes, altered land use, socio-economic factors and livestock movements provide the opportunity for the increased spread and introduction of pathogenic isolates to humans. The World Health Organisation (WHO) have added fasciolosis to their preventative chemotherapy concept [7] supported by Novartis Pharma AG, with the ultimate aim to implement large scale drug distributions where fasciolosis is a public health concern [8] .
Thus, in the absence of commercial vaccines, control of fascioliosis in livestock is based on the use of anthelmintic drugs. The current drug of choice for treatment of fasciolosis is triclabendazole, a benzimidazole-derivative, which shows activity against both juvenile and mature flukes. However, recent reports of triclabendazole resistance have emerged suggesting control of this infection in livestock may become compromised [9] , [10] , [11] , [12] . In addition, consumers worldwide are concerned about drug residues in the environment and food leading to an increased demand for non-chemical based treatments [13] . Research which is directed towards robustly identifying, characterising and validating vaccine candidates is therefore timely.
The pathology associated with fasciolosis is related to the release of proteins from F. hepatica directly into the host via specific secretory and non-specific passive processes [14] . The predominant excretory-secretory (ES) products from in vitro studies are the cathepsin L (Cat L) proteases [14] , [15] . Furthermore, the secreted Cat L proteases from F. hepatica appear to have pivotal roles in parasite survival, including immune evasion, nutrition and migration [16] . Additionally, the Cat L proteases are expressed in both juvenile [17] , [18] and adult fluke, although the degree of expression, and the isoform expressed, may vary with ontogenic stage [16] .
Cat L proteases are found in all liver fluke life stages with many forms expressed and secreted, likely reflecting different biological functions. Unfortunately, singleton sequence deposits in the public domain have caused confusion. F. hepatica Cat L public database entries have been deposited with relatively high degrees of sequence similarity between them. For example, according to Tort [19] , the Cat L sequences described by Wijffels et al. [20] and Roche et al. [21] share sequence similarities with Fcp6, described by Heussler and Dobbelaere [22] , of greater than 94%. Additionally, an isolated Cat L2 (CL2) clone [21] shows 97% sequence similarity to Fcp1, which was also isolated by Heussler and Dobbelaere [22] . However, phylogenetics can delineate between members of the Cat L protease family, and shows that this is a large family which has expanded within Fasciola via gene duplications, leaving a monophyletic group with distinct clade structures [23] , [24] .
Variable protection rates have been reported using Cat L protease isoforms in vaccine formulations in both field and laboratory trials. A limited understanding of the Cat L protease sub-proteome may be hindering development of this vaccine candidate protein family [25] . A key consideration for vaccine development is to target functional components that are required for the survival of the parasite [26] . To this end, Cat L proteases from F. hepatica have been validated as targets. However, an effective broad spectrum commercial vaccine must also overcome the problems of antigenic diversity [27] . Challenge based vaccination trials with a Cat L protease variant derived from a limited liver fluke population analysis may produce variable protection rates as a partial consequence of altered antibody responses [28] . It is clear there will be more Cat L antigen variability within natural liver fluke populations, highlighting that robust vaccine development requires robust population level vaccine discovery with sensitive assay tools.
Therefore, an unbiased global assay of the Cat L proteases, produced in vivo , will untangle the complexity of a problematic analysis of individual Cat L proteases in different laboratories. Detailed proteomic experimentation into the Cat L protease family has been performed in vitro [24] . Robinson et al. [24] identified members of the Cat L protease family from clades 1A (CL1A) and B (CL1B), 2 (CL2) and 5 (CL5), but not from those originating from the newly excysted juvenile (clades 3 and 4) or F. gigantica (clades 1C, 3 and 4). However, there are clear discrepancies between ex-host and in vivo based studies [14] , it is vital to confirm what complement of Cat L proteases are actually expressed in the host environment if vaccines are to be developed on this target. An in vivo analysis that avoids the additional non-biologically relevant consequences of ex-host studies will produce more robust datasets for vaccine discovery. Therefore, we incorporate sub-proteomics to delineate the Cat L protease family that are secreted by adult F. hepatica , comparing for the first time in vitro and in vivo Cat L profiles using 2-DE, mass spectrometry, bioinformatics and phylogenetics.
We demonstrate that population level variations in a key parasite vaccine candidate can be revealed by sensitive proteomic level assays. This Fasciola case study provides a general strategy to accelerate the pace of vaccine discovery and subsequently vaccine development. | Materials and Methods
ES Product Collection and Preparation for 2-DE
Live adult F. hepatica were cultured for 4 h and prepared as previously described [14] in order to collect in vitro ES products. Gall bladders from naturally infected sheep livers were collected immediately post-slaughter, from a local abattoir, and bile extracted and prepared as previously described in order to obtain in vivo ES protein products [14] . Samples prepared for 2-DE SDS-PAGE were re-solubilised in buffer containing 8 M urea, 2% CHAPS w/v, 33 mM DTT, 0.5% carrier ampholytes (pH 3–10, 4.9–5.7 or 5.5–6.7) v/v and protease inhibitors (CompleteMini, Roche, U.K.) for in vitro ES products or buffer containing 6 M urea, 1.5 M thiourea, 3% w/v CHAPS, 66 mM DTT, 0.5% v/v carrier ampholytes (pH 3–10, 4.9–5.7 or 5.5–6.7) and protease inhibitors (MiniComplete, Roche, U.K.) for in vivo ES products.
2-D Electrophoresis
A total of 300 μl of ES product samples were used to actively rehydrate and focus 17 cm linear pH 4–7, 4.9–5.7 or 5.5–6.7 IPG strips (Biorad, U.K.) at 20°C for separation in the first dimension. All IPG strips were focussed between 40,000 and 60,000 Vh using the Ettan IPGphor system (Amersham Biosciences, U.K.). Each IPG strip was equilibrated for 15 minutes in 5 ml of equilibration buffer (containing 50 mM Tris-HCl pH 8.8, 6 M Urea, 30% v/v Glycerol and 2% w/v SDS [29] ) with the addition of DTT (Melford, U.K.) at 10 mg/ml. The equilibration buffer containing DTT was removed and replaced with equilibration buffer containing IAA (Sigma, U.K.) at 25 mg/ml again for 15 mins. The IPG strips were separated in the second dimension on the Protean II system (Biorad, U.K.) using 11% polyacrylamide gels and run at 40 mA for approximately 1 h until through the stacking gel followed by 60 mA through the resolving gel until completion.
Gels were Coomassie blue stained (PhastGel Blue R, Amersham Biosciences, U.K.) over night in 10% v/v acetic acid and 30% v/v methanol. The background of coomassie stained gels was removed using 10% v/v acetic acid and 30% v/v methanol leaving visibly stained protein spots. All coomassie stained gels were imaged with a GS-800 calibrated densitometer (Biorad, U.K.) set for coomassie stained gels at 400 dpi. Imaged 2-DE gels were analysed using Progenesis PG220 v.2006. Analysis was performed using the Progenesis ‘Mode of non-spot’ background subtraction method on average gels created from a minimum of three biological replicates. Normalised spot volumes were calculated using the Progenesis ‘Total spot volume multiplied by total area’ method and were used to determine the degree of up and/or down regulation between in vitro / in vivo comparisons (with significance set at +/−2 fold change). Significance of fold changes was confirmed by a one way ANOVA using LOG 10 transformation, where appropriate, following a Kolmogorov-Smirnov test for normally distributed spot volumes. Unmatched protein spots were also detected between gel comparisons.
Key protein spots of interest were excised and tryptically digested (Modified trypsin sequencing grade, Roche, U.K.). Briefly, protein spots were destained in 50% v/v acetonitrile and 50% v/v 50 mM ammonium bicarbonate at 37°C until clear. Destained spots were dehydrated in 100% acetonitrile at 37°C for 30 mins followed by rehydration with 50 mM ammonium bicarbonate containing trypsin at 10 ng/μl at 4°C for 45 mins. This was followed by overnight incubation at 37°C. Protein tryptic fragments were then eluted according to Shevchenko et al. [30] . Samples were re-suspended in 10 μl of 1% v/v formic acid and 0.5% v/v acetonitrile for tandem mass spectrometry (MSMS).
Mass Spectrometric Analysis
Samples for MSMS were loaded into gold coated nanovials (Waters, U.K.) and sprayed at 800–900 V at atmospheric pressure using a QToF 1.5 ESI MS (Waters, U.K.). Selected peptides were isolated and fragmented by collision induced dissociation using Argon as the collision gas. Fragmentation spectra were interpreted directly using the Peptide Sequencing programme (MassLynx v 3.5, Waters. U.K.) following spectrum smoothing (2×smooths, Savitzky Golay+/−5 channels), background subtraction (polynomial order 15, 10% below the curve) and processing with Maximum Entropy (MaxEnt) 3 deconvolution software (All MassLynx v 3.5, Waters. U.K.). Sequence interpretation using the Peptide Sequencing programme was conducted automatically with an intensity threshold set at 1 and a fragment ion tolerance set at 0.1 Da. Carbamidomethylation of cysteines, acrylamide modified cysteines and oxidised methionines were taken into account and trypsin specified as the enzyme used to generate peptides. A minimum mass standard deviation was set at 0.025 and the sequence display threshold (% Prob) set at 1. Samples that did not show significant scores and probability when using automated sequence prediction were also interpreted manually to generate sequence tags rather than full peptide sequence information. In these circumstances, the MassLynx program Peptide sequencing was again used with the parameters described above.
Database Searches and Analysis
Peptide sequences and sequence tags from MSMS were used separately to search the Genbank protein database ( www.ncbi.nlm.nih.gov/ ) using BLAST adjusted for short nearly exact matches [31] . Consequently, all protein accession numbers reported here relate to Genbank. Only peptides with E values of less than 0.1 were used to assign an identity or a clade to a protein (see Table S3 for all E values). In some cases peptides produce E values greater than 0.1 despite 100% sequence matching. As stated, these peptides were not included for Cat L clade assignment but added confidence to the identifications. To identify any novel Cat L isoforms, all sequences that did not show 100% sequence identity to Genbank entries were subjected to a local BLAST analysis using BioEdit Version 7.0.5.3 (10/28/05) [32] searching an in house translated database of F. hepatica ESTs (available by anonymous FTP from the Wellcome Trust Sanger Institute ftp://ftp.sanger.ac.uk/pub/pathogens/Fasciola/ ). Again, only matches with E values less than 0.1 were used to assign a Cat L clade.
EST Analysis
Peptides sequenced during MSMS analysis in conjunction with the F. hepatica EST database yielded a novel Cat L protease sequence. Several of the EST sequences were found to be only partial sequences. Of these matching ESTs the two longest (Fhep22e06 and Fhep21e10) were selected for further sequence confirmation to obtain a more complete sequence. Stratagene BlueScript SK(+) plasmids containing F. hepatica inserts (Fhep22e06 and Fhep21e10) were provided by Dr Elizabeth Hoey (Queens University Belfast) were transformed into competent Escherichia coli cells (strain DH5α). Fresh plasmid was prepared from 24-hour cultures of transformed single colonies using a Promega (U.K.) Wizard SV Plus MiniPrep DNA purification system according to the manufactures instructions. Forward and reverse DNA sequencing using standard T7 and T3 primers ( 5′ TAATACGACTCACTATAGGG 3′ T7 primer; 5′ ATTAACCCTCACTAAAGGGA 3′ T3 primer) was performed at the commercial DNA sequencing service of Lark Technologies, Inc. (Essex, U.K.). Nucleotide sequences corresponding to the correct reading frame and similarity were aligned, using BioEdit Version 7.0.5.3 (10/28/05) [32] to establish overlapping regions and facilitate construction of two full-length sequences. For signal peptide prediction the SignalP 3.0 Server [33] , available at http://www.cbs.dtu.dk/services/SignalP/ , was used. SignalP was set for eukaryotes using both neural networks and hidden Markov models. For epitope prediction, a Kolaskar and Tongaonkar Antigenicity prediction method [34] , available at http://tools.immuneepitope.org/tools/bcell/iedb_input , was used.
Phylogenetics
Alignments of Fasciola ( F. hepatica and F. gigantica ) Cat L protease nucleotide and amino acid sequences were constructed using ClustalX [35] . The Cat L protease sequences used were taken from the Genbank database ( www.ncbi.nlm.nih.gov/ ) and also included Cat L proteases from this study (EU835857 and EU835858), identified from BLAST analysis searching with novel peptides. Both C-terminal and N-terminal nucleotide sequences were removed where sequence information was limited for many sequences. To construct the nucleotide phylogenetic tree the alignment was exported into Molecular Evolutionary Genetics Analysis (MEGA) software version 4.0 [36] . The phylogenetic tree was generated using a bootstrapped, 1000-replicate, neighbour-joining method. The data were codon based modified using Nei-Gojobori/Jukes-Cantor calculation as a distance based method. Following alignment of amino acid data, amino acid phylogenetic trees were again constructed using MEGA v 4.0. Analysis was performed using a neighbour-joining method, 1000-replicate, bootstrapped tree. The amino acid data was corrected for a gamma distribution (level set at 1.0) and with a Poisson correction. | Results
2-DE Mapping of F. hepatica ES
The F. hepatica ES proteins prepared from in vitro culture and in vivo from host bile were analysed by 2D electrophoresis using methods well developed in our lab. ES protein arrays produced from in vitro ES products were highly reproducible, with the average percentage matching between replicates at 91%, and good matching for in vivo bile analysis at 75.7%. These 2D arrays of the ES protein revealed a group of protein spots migrating to just below the 30 kDa protein marker and ranging in p I from 4.6–6.6 ( in vitro : p I 4.60–6.62. in vivo : p I 4.65–6.52). This group consisted of 32 protein spots from in vitro samples and 20 protein spots from in vivo samples ( Figure 1 ). ES samples were also analysed using micro range IPG strips in order to check for potential overlapping or co-migrating protein spots. When using a micro range from pH 4.9 to 5.7, spot 18, previously resolved as one spot ( Figure 2A ), migrated to produce three distinct spots ( Figure 2B ). When using IPG strips ranging from pH 5.5 to 6.7, no overlapping protein spots were seen ( Figure 2C ). All of the fore mentioned protein spots were excised for MSMS analysis to identify the ES Cat L proteases, in total 34 from in vitro samples and 22 from in vivo samples.
Following MSMS analysis, 30 protein spots from in vitro samples and 19 from in vivo samples were identified as F. hepatica Cat L proteases ( Table 1 ). To assign a Cat L protease to a clade, according to the classification of Robinson et al. [24] , clade specific peptides were identified (Clade 1: VTGYYTVHSGSEVELK and NSWGLSWGER ; Clade 2: VTGYYTVHSGDEIELK , LTHAVLAVGYGSQDGTDYWIVK and HNGLETESYYPYQAVEGPCQYDGR ; Clade 5: PDRIDWR and FGLETESSYPYR together with TSISFSEQQLVDCSR ). From identifying clade specific peptides, 14 different isoforms of F. hepatica Cat L protease isoforms were identified representing the adult CL1, CL2 and CL5 Cat L protease clades. No matches were made to the juvenile clades CL3 and CL4 or from the F. gigantica CL1C sub-clade.
Despite classifying all the identified Cat L proteases to a Cat L protease clade, the specific isoform or sub-clade could not be assigned, suggesting the specific number of isoforms identified may be under represented. Where possible, MSMS was used to provide sequence information for at least three peptides per spot to definitively identify Cat L proteases. However, in our hands, seven identifications based on single peptides were still able to place a Cat L protease to a single Clade (CL1, CL2 or CL5) although not to a single sub-clade such as CL1A or CL1B. These Cat L protease isoforms were often labelled as ‘NFD’ (not fully designated) to a specific Cat L protease clade. The inability to identify a sub-clade was highlighted from members of the CL1A and 1B clades.
Only two Cat L protease isoforms were identified as single protein spots in the in vitro ES and in vivo bile 2DE arrays. These were identified as a CL1B isoform (Accession number CAC12806) and a CL2 isoform (Accession number AAC47721). Both of these enzymes account for a vast proportion of the secreted Cat L proteases (CAC12806: 9–15%; AAC47721: 8–9%).
Several peptides from spots 22, 27, 30 (both in vitro and in vivo ) and 31 ( in vitro only) did not appear in any entries in the public domain. Consequently, these peptides were used to locally BLAST a F. hepatica EST database ( ftp://ftp.sanger.ac.uk/pub/pathogens/Fasciola/ ). As a result from local searches, identical matches were made with twenty three EST sequences, prior to EST assembly. One specific sequence identified, corresponded to a novel amino acid chain; VTGYYTLHSGNEAGLK ( Figure 3A ) and led to the identification of a new Cat L protease isoform representing a resembling a CL1 clade member (See section 3.4).
ES Product Comparison; In Vitro versus In Vivo
Progenesis PG220 v.2006 gel analysis software was used to identify Cat L proteases with altered expression levels from in vitro culture and from in vivo samples and Cat L proteases absent or present in either sample. The two samples, in vivo ES Cat L proteases (twenty spots) and in vitro ES Cat L proteases (thirty two spots), were matched to one another using Progenesis PG220 v. 2006 to give a percentage matching between both of 51.4% matching comparing in vitro to in vivo and 90% comparing in vivo to in vitro .
Having matched in vitro and in vivo preparations to one another, an assessment of the relative quantification of the Cat L proteases could be made. This analysis was conducted using ‘normalised’ spot volumes, which facilitates a relative quantitative assessment despite different protein quantities loaded onto each array (100 μg of in vitro ES Cat L proteases versus 250 μg of in vivo ES Cat L proteases). From this analysis, five Cat L proteases from in vivo preparations show altered expression levels when compared to in vitro Cat L proteases ( Table 1 ). Three of which show an increase in relative expression and the remaining two showing decreases (reversed for in vitro compared to in vivo Cat L proteases). However, following ANOVA with LOG 10 transformed data, only one change was confirmed as significant; Spot 12 increased in vivo (F 1,5 = 25.90 P = 0.015*) and identified as a CL1 protease. An additional spot approached significance; Spot 1 increased in vivo (F 1,5 = 5.77 P = 0.074) but remaining unidentified. The remaining protein fold changes, Spots 21, 28 and 30, all identified as CL1 members, proved not significant due to a high variance exhibited between replicates. To fully confirm these differential levels of expression (>2 fold) statistically, further sampling would be required.
Four Cat L protease spots were clearly identified in vitro and not in vivo . These protein spots, 5, 7, 19 and 20, were highly abundant in the in vitro samples and absent in bile preparations. All four of these protein spots (5, 7, 19 and 20) had also been identified elsewhere in the ES Cat L profile (in both in vitro and in vivo samples), spots 5 and 7 (a CL2 member) were previously identified in spot 2, spot 19 (a CL1A member) was previously identified in spot 16, and spot 20, a CL1A/1B member potentially identified in spot 17. A further 7 protein spots identified only in vitro samples (10, 13, 24, 25, 29, 31 and 32) were of lower abundance and so cannot been completely regarded as absent from in vivo samples. In all cases these 7 proteins were identified as CL1 members with spot 10 also containing a CL5 member.
Identification of Single Nucleotide Polymorphisms
The MSMS tryptic fragments strategy was designed to delineate the Cat L protease superfamily but also identified two non-synonymous single-nucleotide polymorphisms (nsSNPs) that ultimately gave rise to single amino acid polymorphisms (SAAPs) ( Figure 4 ). Firstly, the peptide sequence characteristic of CL1 Cat L proteases, VTGYYTVHSGSE V ELK ( Figure 4 sequence A), was observed in spot 18A, identified as CL1B (CAC12806), as well as the alternate sequence VTGYYTVHSGSE A ELK ( Figure 4 sequence B) within the same spot (replicated in spot 29, see Table S3 ). This nsSNP shows a nucleotide switch from a thymine to a cytosine, creating a conservative amino acid switch, from a valine residue to an alanine. Sequence A was identified on eleven occasions (including twice in spot 19 and three times in spot 11) where as sequence B was only located in three spots. Analysis of ESTs within our in house translated database of F. hepatica ESTs (available at ftp://ftp.sanger.ac.uk/pub/pathogens/Fasciola/ ) revealed a total of 125 matching sequences equivalent to sequence A. In contrast, for sequence B, only eight matching EST sequences were identified. As a result, sequence B has an estimated minor allele frequency of 6.4%, based on the entries currently within the EST database.
A second potential nsSNP was identified showing an amino acid switch from a polar threonine residue to a basically charged arginine residue creating a tryptic cleavage site ( Figures 3B and 4 ). Underlying this amino acid switch was a nucleotide substitution from a cytosine to a guanine. Analysing the in house database, a total of 118 ESTs were found to match QFGLETESSYPY T AVEGQCR ( Figure 4 sequence C) and 30 ESTs that match to the sequence QFGLETESSYPY R ( Figure 4 sequence D). Therefore, this amino acid change, producing sequence D, has an estimated minor allele frequency of 25.4%. Both sequences, C and D, were found in spots 13, 17 and 20. Interestingly, wherever the novel sequence VTGYYTLHSGNEAGLK was located, the peptide QFGLETESSYPYR resulting from a nsSNP creating the arginine residue, was always present. However, the reverse was not seen.
EST Cathepsin L Protease
Two, full length, Cat L protease contigs were constructed from the matching ESTs to further characterise the novel peptides identified during MSMS analysis ( Figure 5 ). This gave rise to two sequences 99.0% similar in nucleotide sequence (see Figure S2 ) and 98.5% similar in amino acid sequence. This corresponded to 10 nucleotide changes with only 5 of these resulting in amino acid substitutions. One of these substitutions occurred within the signal peptide but was still confirmed as a predicted signal peptide. The other four changes found, occurred within the pro-peptide segment, with the highly conserved auto-activation motif GXNXFXD unaltered. Both sequences, when compared at the amino acid level, showed 78.2% sequence identity (87% comparing nucleotides) to the previously described secreted cathepsin L2 (AAC47721). However, minor variations were observed when compared to secreted cathepsin L1 (AAB41670, clade CL1A), showing 92.9% (EU835857, 95% using nucleotides) and 94.5% (EU835858, 96% using nucleotides) sequence similarity.
The new sequences share 100% identity with CL1B members from analysis of the S2 subsite residues determined by Turk et al. [37] , namely amino acid residues 67, 68, 133, 157, 158, 160 and 205 (Papain numbering). However, there is variation in 3 of the 5 mutation hotspots defined by Irving et al. [23] of which the majority are in hotspots I and II. Residues 156 (263) , 158 (265) and 159 (266) in hotspot I, residues 66 (173) , 79 (186) and 91 (198) in hotspot II and residue 173 (280) in hotspot III ( Fasciola numbering [Numbers in superscript correspond to their position in Figure 5 ]) share no homology with CL1B members. However, many show homology to CL1A members including all three residues from mutation hotspot II and the solitary residue in hotspot III.
The novel sequences were passed through a Kolaskar and Tongaonkar Antigenicity prediction method [34] to identify potential antibody epitopes. Interestingly, an epitope identified in CL1A and CL1B members ( Figure 5 : residues 199 to 236) was now split into two smaller epitopes spanning residues 199–222 and 229–236 as a result of a maximum of 3 sequence variations ( Figure 5 : residues 225, 227 and 228).
Phylogenetic Analysis
The objective of a phylogenetic analysis was to identify the origins of our newly identified Cat L protease sequences and to assess the overall clade structure of the Fasciola Cat L protease sub-family. Phylogenetic trees were constructed using nucleotide and amino acid data separately in order to delineate the phylogenetic relationship of the Fasciola Cat L proteases. This strategy produced trees of high similarity ( Figures 6A , S3 and S4 ) providing high levels of confidence when assessing the overall Fasciola Cat L protease relationship.
As with previous studies our Trees divided the Cat L proteases into 5 distinct clades [23] , [38] . The NEJ specific CL3 proteases and the adult CL2 and CL5 proteases formed three of these five clades with strong bootstrap support. The public entry AY428949 from F. gigantica juveniles, did not cluster with any other Cat L protease in our analyses and forms a second juvenile clade termed CL4 [24] . The fifth clade encompasses the CL1 proteases where extensive sub-division has previously been identified; in our study, CL1 being split into 4 sub-clades. The F. hepatica Cat L1 proteases have undergone much expansion, as would also be expected in F. gigantica , and as a result of the greater number of deposited cDNAs in the public domain three of the 4 sub-clades consist entirely of F. hepatica entries; namely CL1A, CL1B and a new sub-clade CL1D. The novel polymorphs identified in the present study clustered together along with AY573569, previously classified as CL1A, forming the new CL1D clade. The remaining CL1 clade, CL1C, contains F. gigantica entries only, although one Japanese entry, classified as F. hepatica , is also included; however this is most likely to be a hybrid species [38] . There was much variation with the positioning of AB010924, which may indicate a possible sixth clade. However, further entries may be required or genomic sequencing to confirm this finding.
Analysis of the Cat L protease amino acid sequences produced trees that were closely similar to one another and to those produced using nucleotides, providing a further level of confidence in the trees produced ( Figures 6B , S5 and S6 ). As with nucleotide data, the Cat L proteases could be divided into 5 distinct clades. The NEJ CL3 and CL4 clades and adult CL2 and CL5 clades were resolved as previous, providing a high degree of certainty. The CL1 clades clustered together, as prior analyses, yielding CL1A, CL1B, CL1C and the novel CL1D. CL1D contains the new isoforms outlined in the present study but without the CL1 protease ATT76664 (Nucleotide AY573569) supported with strong bootstrap support (77% Minimum Evolution, 76% Neighbour Joining and 78% Maximum Parsimony).
Both our phylogenetic and proteomic studies supported one another regarding the clade structure of the Cat L protease family in Fasciola . As with Robinson et al. [38] only members of adult F. hepatica Cat L protease clades could be identified during 2-DE analysis, namely CL1A, CL1B, CL2, CL5 and the new CL1D. Supporting the in vitro study of Robinson et al. [38] , the ES Cat L proteases were predominantly made up of CL1 proteases (71.68% in vitro and 72.78% in vivo from this clade). ES Cat L proteases from the novel CL1D sub-clade appeared to be expressed at levels similar to the CL1B sub-clade constituting 10.34% and 9.11% of the overall protease content in vitro and in vivo respectively ( Table S1 ). Previously, the CL1B clade constituted 32.09% in vitro [24] , which approximates to the combined CL1B and CL1D clades along with those NFD (29.47% in vitro and 36.80% in vivo ). The remaining Cat L proteases originate from the CL2 clade (21.51% in vitro and 15.07% in vivo ) and the CL5 clade (5.52% in vitro and 5.61% in vivo ). | Discussion
The evolutionary success of Fasciola hepatica is, in part, due to its adaptability to successfully invade and establish in different mammalian hosts [23] . The invasion of multiple host species is supported by the secretion of the multifunctional and multi-family member Cat L proteases within the host environment. Cat L proteases, stored as inactive zymogens [39] , are released in relatively large quantities [0.5–1 μg/adult/hour: 40] in order to facilitate obligate blood feeding [41] ; often degrading 1.5×10 8 red blood cells h −1 worm −1 [41] .
The present study demonstrates that comprehensive high resolution 2-DE mapping of these ES Cat L proteases using narrow and micro range IPG strips and large format SDS-PAGE resolves many issues derived from reductionist based experiments. In support of previous studies [24] we have 1) identified 3 of the 5 Cat L clades (CL1, CL2 and CL5) in Fasciola species from adult liver fluke, of which the CL1 and CL2 clades are the major constituents 2) failed to identify F. _igantic CL1 representatives and 3) failed to identify CL3 and CL4 representatives, juvenile specific and enhancing the belief that they are important in gut invasion [42] . No other proteases were identified during this study highlighting the sole reliance of F. Hepatica on these proteases to provide nutrition.
In addition, the current study has also revealed differences in the Cat L protease complement from the artificial in vitro biology platform of liver fluke ex-host, and liver fluke within the natural host, in vivo . A novel Cat L CL1 clade isoform has also been identified along with the first report of single amino acid polymorphism (SAAP) identified via experimental non-model organism proteomic investigations.
Previous proteomic studies have encountered difficulties in Cat L protease identification in the non-genome sequenced F. Hepatica using PMF [14] , [15] . MSMS peptide sequencing allowed for a more robust analysis of the Cat L proteases [the present study and 38] . However, although all sequences were confirmed as Cat L proteases and assigned to the appropriate clade, not all sequences could be firmly linked to a specific sub-clade or database entry. This is directly related to the large degree of allelic diversity observed between Cat L proteases produced by F. Hepatica [43] , especially with the onset of triploidy in Fasciolids [44] . The full extent of Cat L protease diversity will require a significant high-throughput sequencing effort of natural populations of liver fluke.
Thus, with new confidence in assigning Cat L proteases to a specific clade (CL1, CL2 or CL5) the Cat L protease expression profiles from both in vitro and in vivo preparations could be robustly assessed. The assays revealed there were differences between the two preparations with respect to the presence or absence of protein spots and the regulation of protein spots. Key changes between in vitro and in vivo Cat L proteases included the addition of four abundant protein spots (Spots 5 and 7 – CL2, Spots 19 and 20 – CL1A or B) within the in vitro profile (and a further 7 CL1 lower abundant proteins). All four of these proteins were identified in other locations within the in vitro ES proteome profile, which suggests they are post translationally modified (PTM).
The most likely PTM seen on Cat L proteases will be related to mannose 6-phosphate phosphorylation, signalling transport to the lysosomes. The residues important for lysosomal targeting, via phosphorylation, in human Cat L proteases are partially conserved in Fasciola sp. [45] suggesting PTM of Fasciola Cat L proteases via phosphorylation is a likely candidate. As these parasites are cultured ex-host, post translational phosphorylation may indicate an increase in Cat L protease secretion in response to a poor nutritional environment or may represent modification related to the chemical environment. While the overall biochemical activity of the cat L proteases secreted in vitro and in vivo may be similar with model and calculated natural substrates (see Table S1 ), the potential finding of additional PTM in the ex-host preparations warrants further investigation to reveal their influence on both masking enzymatic activity and future Cat L/protein interactions. However, PTM identification might be a reflection of plasticity of the host-parasite in changing host environments.
A significant increase in the abundance of Cat L protease from the CL1 clade was observed in preparations derived from in vivo treatments (Protein Spot 12). All the remaining changes in expression between in vitro and in vivo samples, although not confirmed significant (further sampling required), were also identified as CL1 members. Within the host, F. Hepatica is involved in interactions between both host and parasite which are clearly absent ex-host in vitro culture. Therefore, it is possible that the variability in expression of these CL1 clade Cat L proteases may represent a specific response of the parasite to the host environment, such as immune evasion and nutrient acquisition [41] , [46] . The selection pressure exerted by the host on the CL1 Cat L proteases and the plasticity of CL1 expression in F. Hepatica has most likely led to the divergence seen in the CL1 clade producing the repertoire of sub-clades seen in the present study. The main challenge faced by F. Hepatica in vitro culture and within the host bile ducts is nutrient acquisition. As the CL1 clade is most likely responsible for the degradation of host haemoglobin for nutritional requirements [47] , the variation in CL1 expression seen in the present study will be in response to nutritional acquisition. With lower risks of host immune attack in bile [48] , [49] and the reduced need to migrate through the interstitial matrices the requirement to vary the regulation of CL2 and CL5 clade Cat L proteases in these 2 systems ( in vitro and in vivo ) may have become less important.
The selection pressure exerted on the CL1 clade by the host is highlighted by two further aspects of this study, namely the discovery of a new sub-clade CL1D and the discovery of SAAPs in the CL1 clade. Using contig sequences derived from the novel Cat L EST sequences, a 92.9–94.5% amino acid sequence identity to CL1A was shown (78.2% to CL2 84–85% to CL5), suggesting these novel Cat L sequences are CL1 members. In addition, they appear not to be the potentially novel Cat L protease sequences described by Robinson et al. [42] . Following phylogenetics it appears the new Cat L protease sequences constitute a novel CL1 sub-clade, CL1D. This clade appears to have initially diverged with the rest of the CL1 members following the division of the CL5 clade. The separation of this clade also seems to be an early divergence post-division of F. Hepatica and F. Gigantic CL1 members. Additionally, it appears this clade has branched from the CL1B clade, highlighted by the clustering of a previously classified CL1B (AY573569) with the two CL1D sequences at the nucleotide level but not at the protein level, suggesting they are now functionally different.
A comparative analysis of the S2 active site residues in the substrate binding region [50] predicts that the new CL1D would be biochemically identical to CL1B, and thus not support the phylogenetic analysis of a new sub-clade. However, due to significant variations in 2 of the 5 mutation hotspots identified by Irving et al. [23] (mutation hotspots I and II, one on either side of the active site cleft) relative biochemical activity would need to be confirmed. Three amino acids under positive selection pressure in each of the two mutation hotspots vary between CL1B members and CL1D members. Both hotspots are suggested to be involved in interactions with substrates or other proteins outside of the normal binding regions [23] . In addition, amino acids 156 and 159 ( Fasciola numbering) in hotspot I are suggested to be involved in S2 sub-site interactions [37] and may well influence biochemical activity.
An additional variation between CL1B and CL1D members can be found in mutation hotspot III, the hotspot found on the edge of the R-domain of the cat L proteases [23] . The exact function of hotspot III, and IV and V, are unclear but suggested to be involved in proteolytic interactions with globular proteins [23] . It appears that CL1D members have an intermediary biochemical activity between CL1A and CL1B members as they share similarities with both clades. Therefore, until confirmation of the biochemical activity of CL1D members, and additionally CL1B members, has been confirmed in relation to CL1A members, it seems warranted to keep the CL1D members in a separate sub-clade of the CL1 proteases.
The second aspect of variation seen in the CL1 clade relates to the discovery of two nsSNPs, including one variant in the newly identified CL1D sub-clade. This is the first report of single amino acid polymorphism (SAAP) identified via experimental non-model organism proteomic investigation as population genetics and genomic discovery approaches are the usual methods for SAAP identification [51] . This finding highlights the power of gel based proteomics to reveal differences at the amino acid substitution level.
The first SAAP, position 120 ( Fasciola numbering), involved a conservative amino acid switch, from a valine residue to a alanine residue, both non-polar. This SAAP had a low estimated minor allele frequency of 6.4% and therefore, was only located in three Cat L proteases spots (18A, 28 and 29, all CL1 members). However, the second SAAP was a switch from a small, polar threonine residue to a large polar, positively charged, arginine residue. This second switch was shown by Irving et al. [23] to be an amino acid residue (position 91 Fasciola numbering) under positive selection pressure increasing the frequency of this substitution and, accordingly, had an approximate minor allele frequency of 25%. This particular SAAP was located in eight protein spots, all CL1 members, and demonstrates the increased frequency when compared to SAAPs not subjected to selection pressures. Only one variant of this T91R SAAP was located to the novel cat L1D protease identified in this study, further suggesting divergence from the CL1B clade.
Although outside of the extended Cat L protease active site the T91R substitution could have potential effects on the function of these enzymes. This particular SAAP site (site 91 Fasciola numbering) is located in a mutation hot spot (II) and forms one side of the active site cleft. As mentioned, this region may be involved in interactions with substrates outside of the normal binding regions [23] and may therefore have implications in the specificity of higher order interactions in the enzymes that carry this SAAP. However, this will need to be confirmed.
Both SAAPs observed in expressed Cat L proteases in the present study may confer structural alterations that could affect recognisable epitopes and, if used as a chemotherapeutic target, drug binding sites [52] . Therefore, these SAAPs provide an opportunity to study antigenic variants which may be useful for future development of control measures. Additionally, they lend themselves to modelling based studies to ascertain any conformational alterations such as solvent exposure and interactions.
The plasticity revealed in the CL1 clade of Cat L proteases may impact on the future development of vaccines based on this target. Will a vaccine targeted towards a CL1 member effectively overcome the antigenic diversity seen in this clade? It has been postulated that significant economic benefit would arise from vaccination resulting in a reduction of worm burden of >50% [53] , [54] . Other vaccine trials would favour formulating vaccines based on CL2, CL3 or CL5 members. Although direct comparisons between vaccine trials are difficult to perform there appears strong evidence that trials with Cat L protease from CL2, CL3 and CL5 clades, individually and in combination with other antigens, are more promising in relation to worm burdens than CL1 trials. Recent trials with the NEJ clade CL3 have produced early success reaching reductions of 52% after vaccination with only this Cat L protease [55] . Successful early trials with the CL5 clade have also been performed providing a 51.4% reduction using the Cat L protease alone and a substantial 83% reduction when used in combination with Cat B [56] . CL2 trials have been more extensive than with CL3, CL4 or CL5 members and reported protection ranges from 33–60% using CL2 alone [57] but in conjunction with fluke haemoglobin have reached 72.4% [58] . Trials involving CL1 have been by far the most studied with protection beginning at 0% and reaching a maximum of 69.5% [58] , with the majority consistently below the recommended 50% reduction.
The plasticity within the CL1 clade may be underpinning the observed variability in previous vaccine trials using CL1 proteases. Others have shown that SAAPs can have a profound effect on the antigenicity of pathogenic organisms [59] . Patterns of excessive polymorphisms in parasitic antigens are consistent with high selection pressure and are suggested to function in immune evasion [28] . Furthermore, Irving et al. [23] identified excessive polymorphism in 5 mutation hotspots (previously discussed) which may be affecting interactions with immune effector molecules [23] , [59] . The expansion of the F. Hepatica CL1 clade into 3 sub-clades (1A, 1B and 1D) could be a direct effect of immune selection. Indeed, the novel CL1D isoforms identified in this study show altered predicted epitopes from CL1A and CL1B members as a result of three SAAPs splitting a large epitope into 2 of smaller size. However, it has been shown that a single SAAP can be responsible for altering the immune recognition of parasitic antigens [28] . The evidence presented in the current study, showing further expansion of the CL1 clade, in addition to SAAPs only identified in CL1 members enhances the possibility that this plasticity underpins the variability of protection seen in vaccine trials. CL1 trials have already shown potent anti-embryonation effects and significant reductions in fecundity [58] . This raises the possibility of formulating improved combinations of Cat L proteases to significantly reduce worm burden and fecundity/embryonation in tandem by robust population proteomics assays [60] .
This study has also raised the possibility of a sixth Cat L protease clade. The Cat L protease (Accession number AB010924) was always placed singularly and did not cluster with any other members. In prior analyses this entry has been classified as a CL2 member. In the present study, AB010924 clustered in varying positions between CL5 and CL1, off the CL1 clade or near CL2 and CL5. Further investigation or enhanced genomic information will be needed to confirm this finding.
To summarise, comparison of the ES product sub-proteomes has highlighted variations in the Cat L protease profile between ex-host artificial platforms and direct in vivo assays most likely related to PTM. Therefore, in vitro studies on the Cat L proteases from Fasciolids may increase the understanding of host-parasite relationships by revealing potential plasticity of an important vaccine target superfamily. In this case study, the plasticity of Cat L protease expression has been shown to be limited to the CL1 clade, leading to the discovery of a new CL1 sub-clade revealed through proteomic-EST sequencing-phylogenetic studies. For the first time, this study has identified experimentally single amino acid polymorphisms (SAAP) in a key immunotherapeutic parasite target. Gel based proteomics of pooled samples from populations should be considered for SAAP based biomarker discovery. To effectively formulate a vaccine based on the Cat L proteases we suggest that discovery programmes focus on an alternate Cat L protease clade, such as CL5 where promising early results have been shown [56] . | Conceived and designed the experiments: RMM DJW PMB. Performed the experiments: RMM EJL. Analyzed the data: RMM HAW JP. Contributed reagents/materials/analysis tools: JB DJW. Wrote the paper: RMM PMB.
Background
Fasciola hepatica , along with Fasciola gigantica , is the causative agent of fasciolosis, a foodborne zoonotic disease affecting grazing animals and humans worldwide. Pathology is directly related to the release of parasite proteins that facilitate establishment within the host. The dominant components of these excretory-secretory (ES) products are also the most promising vaccine candidates, the cathepsin L (Cat L) protease family.
Methodology/Principal Findings
The sub-proteome of Cat L proteases from adult F. hepatica ES products derived from in vitro culture and in vivo from ovine host bile were compared by 2-DE. The individual Cat L proteases were identified by tandem mass spectrometry with the support of an in-house translated liver fluke EST database. The study reveals plasticity within the CL1 clade of Cat L proteases; highlighted by the identification of a novel isoform and CL1 sub-clade, resulting in a new Cat L phylogenetic analysis including representatives from other adult Cat L phylogenetic clades. Additionally, for the first time, mass spectrometry was shown to be sufficiently sensitive to reveal single amino acid polymorphisms in a resolved 2-DE protein spot derived from pooled population samples.
Conclusions/Significance
We have investigated the sub-proteome at the population level of a vaccine target family using the Cat L proteases from F. hepatica as a case study. We have confirmed that F. hepatica exhibits more plasticity in the expression of the secreted CL1 clade of Cat L proteases at the protein level than previously realised. We recommend that superfamily based vaccine discovery programmes should screen parasite populations from different host populations and, if required, different host species via sub-proteomic assay in order to confirm the relative expression at the protein level prior to the vaccine development phase.
Author Summary
Vaccines for neglected parasitic diseases are of paramount importance. An understanding of the basic biology underpinning target expression within parasite populations is one of the pre-requisites for vaccine discovery and development. Fasciola hepatica causes global disease in humans and their livestock. The pathology of the disease is associated with the release of cathepsin L (Cat L) proteases from the parasite into the host. The Cat L proteases are the leading vaccine candidates and are split into 5 clades with different functions. The CL1 clade has undergone significant divergence resulting in the formation of sub-clades. We have studied this vaccine candidate family at the population level with proteomic based assays using F. hepatica as a case study. We have identified differences in Cat L protein expression profiles between in vitro culture compared to in vivo host bile, with CL1 members showing greater expression plasticity. Selection pressure exerted by the host driving the divergence of the CL1 clade is revealed by single amino acid polymorphisms. This case study highlights that high resolution population based proteomic assays at the vaccine discovery stage will support the successful development of broad population based commercial vaccines based on defined antigens and their families. | Supporting Information | The authors would like to thank Mr Jim Heald and Dr Charly Morgan for their MS skills, the staff at Dunbia (Llanybydder), in particular Gerwyn Probert, for their assistance, Dr Elizabeth Hoey (Queen's University, Belfast, U.K.) for kindly donating plasmids containing the EST sequences and Dr Joe Moxon for technical assistance. | CC BY | no | 2022-01-13 05:08:15 | PLoS Negl Trop Dis. 2011 Jan 4; 5(1):e937 | oa_package/cd/d1/PMC3014944.tar.gz |
|
PMC3014945 | 21245912 | Introduction
Pathogen associated molecular patterns (PAMP) trigger innate immunity against pathogens and this response represents the first line of defense against various microorganisms [1] . Double strand RNA (dsRNA), a viral replication intermediate, is sensed by cytoplasmic RNA helicases retinoic acid-inducible gene I (RIG-I) and melanoma differentiation-associated gene 5 (MDA5) as well as by toll-like receptors-3 (TLR3) [2] . TLR3 and RNA helicases interact with different PAMP during the proximal signaling events triggered by the dsRNA. However, these two parallel viral recognition pathways converge at the level of IFN regulatory factor-3 (IRF3). Phosphorylation of IRF3 initiates antiviral responses, including the activation of type I interferon (IFN), interferon stimulating genes (ISGs) and proinflammatory cytokines [3] , [4] .
While TLR3 is primarily responsible for recognizing viral components such as viral nucleic acid and envelope glycoproteins in the extracellular and endosomal compartments [5] , DExD/H box–containing RNA helicases - RIG-I, MDA5 - recognize intracellular dsRNA and they constitute the TLR-independent IFN induction pathway. Although both RIG-I and MDA-5 share high degree of functional and structural homology, they were observed to respond to different dsRNA moieties and RNA viruses. They contain caspase-recruiting domains (CARD) that allow them to interact with Interferon Promoter Stimulated 1 (IPS-1) (otherwise known as Virus-induced Signaling adapter (VISA); mitochondrial antiviral signaling protein (MAVS) or Cardif) [6] . Similar to TLR3, IPS-1 mediates activation of TBK1 and IKKε which in turn activates/phosphorylates IRF3. Phosphorylated IRF3 then homodimerises and translocates to the nucleus [7] to stimulate the expression of type I interferons – IFN-α and IFNβ. IFN-α/β, together with an array of other interferon stimulated genes (ISGs) and cytokines, lead to the establishment of an antiviral state which restricts virus spread in the host cells.
Dengue virus was reported to induce type I IFN even in RIG-I or MDA5 null cells [8] . The same is observed with West Nile virus [9] , another Flavivirius. Japanese encephalitis virus [10] and Hepatitis C virus [11] , also belonging to the Flavivirdae family, on the other hand, are recognized only by RIG-I. These results suggest that Flaviviruses, despite their common genomic features and replication strategies, are differentially recognized by the host. Despite having the IFN pathway activated in response to viral infection, pathogenic viruses have evolved ways to manipulate the IFN system to favor their survival in the host cell. Reports have shown that DV can antagonize the IFN pathway via their non-structural proteins such as NS2A, NS2B, NS4B and NS5 [12] , [13] .
A recent microarray analysis of dengue virus type 1 (DV1)-infected lung carcinoma cell line, H1299, showed upregulation of a number of innate immune response genes. In particular, RIG-I, MDA5 and TLR3 were up-regulated more than 8-, 5- and 2-fold respectively [14] . Furthermore, Sumpter and colleagues [11] reported that inactivation of RIG-I in HUH-7 cells resulted in permissiveness of hepatitis C virus (HCV; a flavivirus) RNA replication. Since HUH-7 cells have low basal level of Toll-like receptor 3 (TLR 3) gene expression, this cell line would be a good in vitro model system to investigate RIG-I-dependent signaling in DV1 infection [15] .
In this study, we present evidence that DV1 infection results in the upregulation of RIG-I, MDA5 and TLR3 expression in HUH-7 cells. This is the first study that shows the role of all three viral RNA sensors – RIG-I, MDA5, TLR3 – in the same cellular system. Previous studies have shown the role of these sensors in different cell lines which may not take into consideration the differences in the genetic make-up of the cell lines. We show, in this study, how RIG-I, MDA5 and TLR3 signaling pathways play a role in DV1-infected cells. | Materials and Methods
Cell lines, virus infection and total cellular RNA isolation
Wild type and TLR3 knockout macrophages were kind gifts from Dr. Xu Shengli, Bioprocessing Technology Institute, Singapore. Macrophages, HUH-7 and shRIG-I cells were cultured in DME medium containing 5% fetal bovine serum (FBS) and 1% penicillin and streptomycin antibiotics (PSA). One set of uninfected macrophages, HUH-7 and shRIG-I cells served as a control, while another set was infected with the Singapore strain of dengue type 1 virus at a multiplicity of infection (MOI) of 1 and incubated at 37°C for 2 h. The supernatant was replaced with fresh DME containing 1% FBS, and infected and uninfected cells were harvested after 3 days. Mock represents cells incubated/transfected with cell lysates/vector for a period of time similar to the infected cells. Total RNA was extracted using the Trizol reagent (Invitrogen, USA) and RNA concentrations quantified via UV spectrophotometry at 260 and 280 nm. RNAs with an OD 260 nm ∶OD 280 nm absorbance ratio of at least 1.9 with intact ribosomal 28S and 18S RNA bands were used in this study.
UV inactivation of DV1
DV1 was inactivated by exposing the virus to a UV lamp (wavelength, 254 nm) at a distance of 5 cm for 1 h. UV-inactivation was confirmed by the inoculation of C6/36 cells before use and, in individual experiments, by monitoring the exposed cells for synthesis of viral non-structural protein, NS3, at 72 h. The supernatant fluids from the inoculated cells were also monitored for the presence of infectious virus.
Mice and TLR3−/− knockout macrophages
TLR3−/− mice have been described previously [16] and maintained in the C57BL/6 background. Wild type C57BL/6 mice were from the Jackson Laboratory (Bar Harbor, Me). Bone marrow-derived macrophages (BMDMs) were generated by culture of bone marrow cells in DMEM containing 20% FCS and 30% L929 conditioned medium (DMEM-C) for 6 days. BMDMs were harvested and tested for purity by flow cytometry with antibodies specific for F4/80 and Mac-1 (BD Pharmingen). The purity of BMDMs was typically 90–95%
Real-time and semi-quantitative RT-PCR
Total RNA (5 μg) from DV1 infected and uninfected cells were reverse-transcribed. Real-time RT-PCR was carried out for the selected genes using gene-specific primers and the LightCycler-FastStart DNA Master PLUS SYBR Green 1 reaction mix (Roche Molecular Biochemicals, Mannheim, Germany). The LightCycler system was used to monitor the SYBR Green signal at the end of each extension period for 40 cycles. The threshold cycle (C T ) for each gene of interest and for the GAPDH housekeeping gene, and the difference between their C T values (ΔC T ) were determined. The relative expression values (2 −ΔΔCT ) between uninfected and infected samples for the selected genes were determined by using the uninfected sample as the reference with its ΔC T value subtracted from the ΔC T value of the infected sample (i.e. ΔΔC T ). Relative fold difference values shown in figures are average of at least two independent experimental results.
A two-step semi-quantitative RT-PCR method was used to measure gene expression in the DV1 infected and uninfected samples. Random hexamers (Qiagen Inc.) was used as primer in the first step of cDNA synthesis. Total RNA (5 μg) was combined with 2 μl of random hexamers, 200 μM dNTPs and H 2 0 and preheated at 65°C for 10 min to denature secondary structures. The mixture was then cooled rapidly in ice and then 5 μl 5×RT Buffer, 10 mM DTT, 0.5 μl RNAse inhibitor (Roche), 1.0 μl (10 mM) dNTP and 200 U reverse transcriptase (Roche) were added for a total volume of 20 μl. After pulse spinning, the RT mix was incubated at 43°C for 90 min and then stopped by heating at 95°C for 5 min. The cDNA stock was stored at −20°C. The yield of cDNA was measured according to the PCR signal generated from the internal standard house-keeping gene GAPDH amplified 30 cycles with 1 μl of the cDNA solution. Gene-specific PCR amplifications were carried out by adding 5 μl of 10 x PCR buffer, 1 μl (5 U/μl) Taq Polymerase, 1.0 μl (10 mM) dNTP, 1 μl of each primer, 1.5 μl (50 mM) MgCl 2 , 3 μl of the first strand cDNA and double-distilled water to 50 μl. The PCR products were loaded onto ethidium bromide-stained 1% agarose gels. A 100 bp DNA ladder molecular weight marker (Fermentas) was run on every gel to confirm expected molecular weight of the amplification product.
The Primer pairs used were: OAS2: sense, TGAGAGCAATGGGAAATGGG , anti-sense, AGGTATTCCTGGATAAACCAACCC ; RIG-1: sense, TGTGGGCAATGTCATCAAAA , anti-sense, GAAGCACTTGCTACCTCTTGC ; MDA5: sense, GGCACCATGGGAAGTGATT , anti-sense, ATTTGGTAAGGCCTGAGCTG ; IFNβ: sense, CTCTCCTGTTGTGCTTCTCC , anti-sense, GTCAAAGTTCATCCTGTCCTTG ; ISG15: sense, TGGTGGACAAATGCGACGAA , anti-sense, CAGGCGCAGATTCATGAAC ; ISG56: sense, TCTCAGAGGAGCCTGGCTAAG , anti-sense, CCACACTGTATTTGGTGTCTAGG ; XBP1: sense, CTGGAAAGCAAGTGGTAGA , anti-sense, CTGGGTCCTTCTGGGTAGAC ; GAPDH: sense, GACAACTTTGGTATCGTGGAA , anti-sense, CCAGGAAATGAGCTTGACA .
Flow cytometry and statistical analysis
TUNEL assay. TdT-mediated dUTP-biotin nick-end labeling (TUNEL) was performed using ApoAlert DNA Fragmentation Assay kit (Clontech) according to the manufacturer's instructions. Briefly, the cells were fixed with 4% formaldehyde/PBS and resuspended in 0.2% Triton X–100 and incubated on ice for 5 min. The cells were labeled by adding 50 μl TUNEL mix. The samples were then resuspended in PBS prior to flow cytometry (FACS Calibur; Becton-Dickinson, San Jose, CA) and results displayed using WinMDI 2.8 software program.
Subgenomic content. To measure subgenomic content, cells were fixed with 70% ice-cold ethanol and stained with 50 μg/ml propidium iodide (PI) containing RNase and subgenomic content was evaluated by a flow cytometer.
Statistical analysis. Error bars in figures represent data expressed as the mean ± S.D. of at least three independent experiments. Z-test for two simple means was used to calculate P-value.
Immunoblotting and antibodies
Cells were harvested up to 72 hpi, and lysed in RIPA buffer containing protease inhibitors. The lysates were subjected to immunoblotting using primary and HRP-conjugated secondary antibodies followed by the ECL-Plus chemiluminescence substrates (Amersham). Mouse monoclonal DV1 E and NS3 antibodies used were prepared in-house. Anti RIG-I and anti-MDA5 (Axxora); anti-actin (Santa Cruz); anti-calreticulin (BD Transduction Laboratories); anti-IRF3 (Santa Cruz) and anti-myc (Sigma Aldrich).
Generation of RIG-I knock-down cells, plasmid constructs and transfection of polyI:C
As shRNA target in the RIG-I sequence, 5′-AATTCATCAGAGATAGTCA -3′ was chosen HUH-7 cells were transfected with shRNA constructs using Lipofectamine 2000 reagents (Invitrogen). Clones were selected in the presence of G418 and screened for reduced RIG-I expression. For over-expression studies, RIG-I and MDA5 coding regions were cloned into a mammalian expression vector. The identities of the clones were confirmed by DNA sequencing. Synthetic dsRNA polyI:C was purchased from Sigma Aldrich. Cells were transfected with 1 μg/ml polyI:C for various times using Lipofectamine 2000 (Invitrogen) according to the manufacturer's instructions.
Small interfering RNA (siRNA) assay
For gene silencing of RIG-I: sense, CAGAAGAUCUUGAGGAUAAUU was used. siTLR3 was purchased from Sigma Aldrich. HUH-7 cells (1×10 5 per well) were plated in six-well plates. At 24 h after incubation, cells were washed, replenished with medium without serum and transfected, at 40–60% of confluency, with specific siRNA or control siRNA by using a siRNA transfection reagent (Lipofectamine RNAi Max reagents, Invitrogen) according to the manufacturer's instructions. After 8 h incubation at 37°C, the liposome suspension was removed and complete culture medium was added. After 24–48 h, cells were infected with DV1 and then harvested at 48 h for analysis. Cells that were untreated or treated with control siRNA served as controls.
IRF3-dimerisation assay
Cells were lysed in buffer containing 50 mM Tris HCl (pH 7.4), 150 mM NaCl, 1 mM EDTA, 1% NP-40, protease inhibitors and phosphatase inhibitors for 30 min at 4°C. Proteins were separated by electrophoresis in 8% non-denaturing polyacrylamide gels containing 1% sodium deoxycolate in the cathode buffer. IRF3 monomers and dimers were detected by Western blot using polyclonal antibodies against the full-length IRF3 (Santa Cruz).
Immunofluorescence
Cells were grown on 24-well plates. After infection with DV1 for the indicated times, cells were fixed in cold methanol for 10 min, washed in PBS and blocked in normal goat serum for 1 h. DV1 E antibody was then applied for 1 h followed by FITC-conjugated secondary for another 1 h. The cells were then washed 4 times with PBS and examined under a fluorescence microscope or quantified using a flow cytometer.
Enzyme-linked immunosorbent assay (ELISA)
Cells transiently transfected or cells infected with virus were cultured for up to 48 h. IFN-β in the supernatants was measured using VeriKine Human IFN-β ELISA kit (PBL Biomedical Laboratories) according to the manufacturer's instructions. | Results
RIG-I and MDA5 knockdown in HUH-7 cells results in enhancement of DV1 replication
To study if RIG-I influences DV1 infection and understand the relationship between innate antiviral responses and virus replication, we used short-hairpin RNA interference (shRNA) technology to knockdown RIG-I in HUH-7 cells. A clonal population of stably transfected cells that showed decreased expression for RIG-I was selected. RIG-I knock-down cells (shRIG-I) were infected with DV1 for up to 72 hours and subjected to immunoblotting analysis. RIG-I expression increased over time in DV1-infected wild type HUH-7 cells and there was minimal RIG-I up-regulation in the shRIG-I cells ( Fig. 1A ). Since Loo and colleagues [8] demonstrated that dengue virus type 2 triggered both RIG-I and MDA5, we probed for MDA5 expression in DV1-infected cells. Minimal MDA5 expression was observed in DV1-infected shRIG-I cells while increased MDA5 expression was observed in HUH-7 cells ( Fig. 1A ). Although MDA5 seems to be upregulated faster than RIG-I in HUH-7 cells, the basal level (see mock infection) of the two proteins are not the same. It appears that HUH-7 cells have a higher basal level of MDA5 compared with RIG-I.
Mock-infected HUH-7 cells showed expression of MDA5, whereas shRIG-I cells did not. To investigate if this MDA5 down regulation also occurred at the transcriptional level, RT-PCR was performed on RNA extracted from DV1-infected cells. MDA5 expression at the mRNA level was visibly low in shRIG-I cells as compared to HUH-7 cells ( Fig. 1B ). To define if this MDA5 down regulation was due to clonal specificity, another clone of shRIG-I cells were infected with DV1 and probed for MDA5 expression. Minimal MDA5 expression was observed in the second clone as well (data not shown). It was important to ensure that this phenomenon is not due to DV1 infection, which may not be activating or could be inhibiting MDA5 expression in these cells. A synthetic analog of viral dsRNA, polyinosine-polycytidylic acid (polyI:C), that can bind and activate MDA5 expression was used to investigate the level of MDA5 expression in shRIG-I and HUH-7 cells. MDA5 expression in shRIG-I cells was reduced even with polyI:C transfection ( Fig. 1C ). Quantitative RT-PCR showed that the MDA5 expression was decreased 8-fold in shRIG-I cells as compared to HUH-7 cells upon DV1 infection ( Fig. 1D ). To understand if MDA5 expression was related to RIG-I expression in mammalian cells, we transfected RIG-I siRNA (which had a different target sequence to that of RIG-I shRNA) in A549 and HUH-7 cells. We observed that transfection of RIG-I siRNA significantly knocked down both RIG-I and MDA5 in HUH-7 cells but only RIG-I in A549 cells ( Fig. 1E ). Knock down of RIG-I in both A549 and HUH-7 cells showed an increase in DV1 propagation. Since the sequences of RIG-I shRNA and RIG-I siRNA are different, it is thus unlikely that these siRNAs knocked down a common off-target gene such as MDA5. To show that this response is not cell-type specific, A549 cells were transfected with siRIG-I or/and siMDA5. We observed that knockdown of both RIG-I and MDA5 showed an increase in DV1 propagation as noticed in shRIG-I cells ( Fig. 1F ). This result confirms that the observations in HUH-7 and shRIG-I cells are not cell-type specific. To rule out the possibility that contamination of cellular nucleic acids in the virus preparations that could activate RIG-I and MDA5, the same virus preparations were UV-treated and used to infect HUH-7 and shRIG-I cells. Infection of cells with UV-treated DV1 did not result in any significant increase in RIG-I or MDA5 expression ( Fig.1G ).
DV1 infection of HUH-7 and shRIG-I cells over a period of 72 hours showed that the latter was highly permissive for DV1 propagation as noted by high levels of viral proteins ( Fig. 1A ). Antibodies against DV1 E (structural) and NS3 (non-structural) proteins were used to show effective and efficient DV1 propagation in shRIG-I cells. Furthermore, knockdown of RIG-I, using siRNA duplexes, in A549 and HUH-7 cells also showed an increase in DV1 propagation ( Fig. 1E ). HUH-7 and shRIG-I cells were infected with DV1 and fixed at 48 and 72 h for immunofluorescence assay using a monoclonal antibody directed against DV1 E protein. Enhanced fluorescence intensity was observed in DV1-infected shRIG-I cells as compared to infected HUH-7 cells (data not shown). This fluorescence intensity was quantified by flow cytometric analysis using monoclonal antibody directed against DV1 E protein. A significant increase in DV1-infected cells in shRIG-I cells as compared to infection in HUH-7 cells was observed ( Fig. 2A ). To define if this enhancement effect was due to increase in viral replication, quantitative RT-PCR (real-time RT-PCR) for negative strand DV1 RNA level, which is indicative of viral replication, was performed. Real-time RT-PCR showed more than 3-fold increase in negative strand DV1 RNA level in shRIG-I cells compared to HUH-7 cells (data not shown). An immunofluorescent-based TCID50 assay was used to titrate the amount of infectious particle in HUH-7 and shRIG-I cells ( Fig. 2B ). DV1 propagated significantly higher in shRIG-I cells than in HUH-7 cells.
DV1 activates innate immune-responsive genes in RIG-I- and MDA5-knockdown cells
Recently, Loo and colleagues [8] demonstrated that RIG-I and MDA5 are equally essential for IFN-β production induced in response to infection with dengue virus. To determine whether DV1 induces IFN-β promoter activity in HUH-7 and shRIG-I cells through a RIG-I and MDA5-dependent pathway, we infected these cells with DV1 for up to 72 h. Total RNA extracted from DV1-infected cells were subjected to qualitative ( Fig. 2C ) and quantitative ( Fig. 2D ) RT-PCR to detect mRNA levels of 2′,5′-oligoadenylate synthetase 2 (OAS2), interferon stimulated gene (ISG) 15, ISG56 and IFN-β genes. These genes were chosen for detection because a recent microarray analysis of DV1-infected H1299 cells (human non-small lung cancer cells) showed increased expression of OAS2, ISG15 and ISG56 [14] . RT-PCR assays showed a significant increase in mRNA levels of IFN-β and IFN stimulated genes (OAS2, ISG15 and ISG56) in DV1-infected shRIG-I cells as compared to infected HUH-7 cells. Qualitative RT-PCR for negative strand viral RNA was performed to show level of virus replication in the infected cell lines ( Fig. 2C ). Virus replication was much pronounced in shRIG-I cells. Quantitative RT-PCR showed at least 200-, 50-, 700- and 4-fold difference for OAS2, ISG15, ISG56 and IFN-β gene expression respectively between DV1- infected HUH-7 and shRIG-I cells ( Fig. 2D ). IFN-β production was assayed using ELISA. DV1-infected shRIG-I cells produced more significant levels of IFN-β compared with DV1-infected HUH-7 cells ( Fig. 2D ; lower panel).
When dsRNA is sensed in the cytoplasm, RIG-I and/or MDA5 recruit the adaptor protein IPS-1 [17] , [18] resulting in the downstream activation (dimerization) and nuclear localization of IRF3 which acts as a transcriptional factor for IFN-α/β expression. IRF3 dimerization was assessed by native PAGE with anti-IRF3 antibody as a probe. Figure 2D shows stronger IRF3 dimer formation in DV1-infected shRIG-I cells compared to infected HUH-7 cells. Ratio of IRF3 dimer/monomer shows greater dimer formation in shRIG-I cells (0.92) as compared to HUH-7 cells (0.59) at 72 h. Increased permissiveness for DV1 infection in shRIG-I cells could have lead to induction of other stimulators of IRF3, such as TLR3. These results, collectively, show that factors other than RIG-I/MDA5 may play a role in DV1-induced immunity.
Cellular response to higher dengue virus replication in shRIG-I cells
Since shRIG-I cells are highly permissive for DV1 replication, this might lead to accumulation of viral RNA and perhaps an overload of the cellular protein synthesis machinery triggering host response to the infection [19] . Viral protein overload in the endoplasmic reticulum could result in unfolded protein response (also known as ER stress) and activation of apoptotic cell death in DV1-infected cells. We investigated if this proliferation of viral antigens in shRIG-I cell would lead to increased ER stress and eventually apoptosis/cell death. ER stress leading to apoptosis in dengue virus-infected cells has been reported previously [20] . As a signal of ER stress, X box binding protein 1 (XBP1) gene expression was detected. XBP-1 is up-regulated as part of the endoplasmic reticulum (ER) stress response, the unfolded protein response (UPR) [21] . During ER stress, IRE-I/XBP1 pathway is activated by cleavage of a 26-nucleotide intron from unspliced XBP1 (uXBP1) mRNA resulting in XBP1 in its mature form (sXBP1) [20] . Infection of shRIG-I and HUH-7 cells by DV1 resulted in splicing of XBP1 ( Fig. 3A ). Splicing of XBP1 in DV1-infected shRIG-I cells occurs as early as 24 hpi as compared to 48 hpi in infected HUH-7 cells. Our results conform to that of Umareddy et al., [22] whereby the authors showed that XBP1 is spliced upon dengue virus replication.
Since most DV1 proteins are localized on the luminal side of the ER membrane prior to cleavage and processing into mature forms, we investigated ER stress by also detecting the expression of calreticulin, a multifunctional, multi-compartmental protein most abundant in the ER lumen. Calreticulin interacts and assists in the folding of various glycoproteins, including viral proteins [23] . The DV1 envelope protein which is produced in large quantities during infection is an N-linked glycoprotein [24] . More pronounced increase in calreticulin protein expression was observed in DV1-infected shRIG-I cells as compared to HUH-7 cells over a period of 72 hpi ( Fig. 3A ). Taken together, these results show that ER stress is more profound in shRIG-I cells due to increased permissiveness to DV1 replication.
Apoptosis/cell death is the final outcome of dengue virus infections [25] . Although UPR is necessary for cell survival and viral replication, prolonged UPR can lead to cell death [22] . Furthermore, activation of RIG-I upon virus infection had been reported to activate the apoptotic cascade [26] . Interferon-β promoter stimulator-1 and IRF-3 were shown to be required for efficient apoptosis following reovirus infection, suggesting a common mechanism of antiviral cytokine induction and activation of the cell death response [26] . Since DV1 infection of HUH-7 and shRIG-I cells showed activation of UPR, IRF-3 dimer formation and IFN- β activation, cell death induced by DV1 infection in HUH-7 and shRIG-I cells was investigated. DNA fragmentation, a hallmark of apoptosis, was observed in both DV1-infected HUH-7 and shRIG-I cells in varying degrees ( Fig. 3B ). In situ DNA fragmentation was investigated using TUNEL assay, which relies on the specific binding of terminal deoxynucleotidyl transferase (TdT) to exposed 3′-OH ends of the fragmented DNA. The signal is then amplified by avidin-peroxidase, enabling detection of in situ DNA fragmentation by flow cytometry. The results of this assay on DV1-infected cells showed an approximately 10% increase in cell death in shRIG-I cells as compared with HUH-7 cells after 72 hpi ( Fig. 3B , upper panel). Cell death in DV1-infected cells was further investigated by analysis of sub-G1/DNA content, indicative of cell death, in HUH-7 and shRIG-I cells. Analysis of the changes in the DNA content distribution showed 20% more sub-G1 DNA in DV1-infected shRIG-I cells as compared to DV1-infected HUH-7 cells ( Fig. 3B , lower panel).
RIG-I and MDA5 synergistically mediate type 1 IFN response
To investigate if RIG-I, MDA5 or both helicases influence the initiation of IFN response and to understand their antiviral potencies, we overexpressed these genes in HUH-7 cells. Upon infection with DV1 for up to 48 h, less DV1 propagation was observed in cells transfected with both helicases ( Fig. 4A ). Real-time RT-PCR showed more than 5-fold decrease in negative strand DV1 RNA level in cells transfected with both RIG-I and MDA5 (0.13) compared to infected HUH-7 cells (1.0). Significant increase in IFN-β gene expression was also noted in infected cells overexpressed with both RIG-I and MDA5 compared to infected HUH-7 cells ( Fig. 4B ). When transfected individually, RIG-I and MDA5 were not able to efficiently suppress DV1 replication but induced elevated amounts of IFN-β expression. These results clearly show that RIG-I and MDA5 synergistically mediate an antiviral response during DV1 infection.
Activation of TLR3 in DV1-infected cell lines
Inspite of knocking down RIG-I and MDA5, the level of IFN-β produced seems to increase upon DV1 infection in shRIG-I cells. Furthermore, an increase in cytokine production was noted ( Fig. 2C ) and an increase in IRF3 dimerization was also observed in native PAGE analysis ( Fig. 2E ). This phenomenon could be due to an increase in virus replication in the knock down cell line or due to other intracellular receptors such as Toll-like receptor 3 (TLR3) that recognize double-stranded RNA synthesized during DV1 replication. TLR3 has been shown to recognize double-stranded RNA [27] and RIG-I/MDA5 and TLR3 signaling pathways converge to phosphorylate IRF3 [28] . We investigated if TLR3 could recognize DV1 infection. Since the basal level of TLR3 is low in HUH-7 cells, we used wild type (WT) and TLR3-knockout (TLR3ko) macrophages. Furthermore, TLR3 is abundant on macrophages [29] . While this work was in progress, Tsai et al., [30] reported that TLR3 in HEK293 cells recognizes dengue virus type 2 and induces cytokine production. In our study, WT and TLR3ko macrophages were infected with DV1 for 48 h and cell lysate and RNA were collected for analysis. Immunoblotting analysis and both semi-quantitative and real-time RT-PCR showed that DV1 virus propagation was more efficient in TLR3ko macrophages ( Fig. 5A ). Negative strand viral RNA, indicative of DV1 replication, was quantitated using real-time RT-PCR. Approximately 2.5 fold increase in viral RNA was noted in DV1-infected shRIG-I cells as compared to infected HUH-7 cells. These experiments show that TLR3 recognizes DV1 double stranded RNA intermediate.
DV1 infected HUH-7 and shRIG-I cells showed increase in TLR3 expression over 72 h ( Fig. 5B ). This result was confirmed by real time RT-PCR analysis (data not shown). To determine if TLR3 could modulate DV1 replication, we overexpressed TLR3 in HUH-7 cells and infected with DV1 for up to 48 h. Semi-quantitative RT-PCR and immunoblotting show that over-expression of TLR3 plasmid greatly inhibited DV1 replication in cells ( Fig. 5C ). IFN-β gene expression was quantified for the same set of experiments. Cells over-expressing TLR3 showed significant increase in IFN-β production ( Fig. 5C ).
To gain deeper understanding of the role of TLR3 in DV1 infection, siRNA silencing of TLR3 in HUH-7 and shRIG-I cells was carried out. HUH-7 and shRIG-I cells were transfected with siTLR3 and infected with DV1 for 48 h. Since the basal level of HUH-7 is low, real-time RT-PCR analysis was used to determine TLR3 knock down efficiency in cells ( Fig. 6 ). Control siRNA transfected cells showed increase in TLR3 expression upon DV1 infection as compared to siTLR3 transfected cells. Semi-quantitative and real-time RT-PCR analysis for DV1 negative strand RNA showed increase in DV1 replication in siTLR3 transfected cells with reference to GAPDH expression ( Fig. 6 ). IFN-β production was quantified by ELISA for the same set of experiments, showing significant increase in IFN-β expression in shRIG-I cells transfected with siTLR3 and infected with DV1, as compared to control siRNA transfected and infected cells for the same ( Fig. 6 ). Quantitative real-time RT-PCR for negative strand viral RNA shows a higher fold increase in DV1 replication in shRIG-I cells (2.14) as compared to HUH-7 cells (1.39). These data are evidence that DV1 RNA is recognized by TLR3 and such recognition modulates DV1 replication by eliciting IFN-β production. | Discussion
Innate immune defenses are the first line of host anti-pathogen mechanisms. Among them, type 1 IFNs play a critical role in host defense against virus infection. IFN pathway is triggered through host recognition of viral macromolecular motifs also known as pathogen-associated molecular patterns (PAMP). In this study, we show that IFN-β production in response to DV1 was enhanced in RIG-I and MDA5 down-regulated cells. Knockdown of RIG-I and MDA5 enhanced cellular permissiveness to DV1 RNA replication and also increased virus propagation. Consequently, higher expression of IFN-related and stimulated genes, such as OAS2, ISG15, ISG56, as well as strong activation of IRF3 was observed in the cell line. Analysis of ER stress responses, such as splicing of XBP1 and activation of calreticulin, also showed earlier splicing of XBP1 and accumulation of more calreticulin in RIG-I- and MDA5-knockdown cells. Prolonged ER stress, probably due to increase in viral load, resulted in stronger induction of apoptosis in these cells. Further analysis of the functional roles of RIG-I and MDA5 demonstrated that RIG-I and MDA5 synergistically suppressed DV1 infection with induction of IFN-β production in wild type HUH-7 cells over-expressing the two helicases. Meanwhile, using siRNA against TLR3 showed that down-regulation of TLR3 resulted in higher DV1 infection, and over-expression of TLR3 inhibited DV1 infection significantly by inducing high levels of IFN-β production. When siTLR3 was transfected in shRIG-I cells (down regulated for RIG-I and MDA5), increased permissiveness to DV1 infection was observed. TLR3 knockout macrophages also showed increased permissiveness to DV1 infection than wild type cells. Collectively, our results showed that RIG-I, MDA5 and TLR3 are able to recognize DV1 infection and establish a strong antiviral state in these cells.
Available evidence reveals that RIG-I and MDA5 play differential roles in host antiviral defence against viral infections. RIG-I was shown to respond to RNA bearing a triphosphate at their 5′ end [31] and a wide variety of RNA viruses including Japanese encephalitis viruses, influenza viruses and paramyxoviruses [8] , [32] , while MDA5 responds primarily to poly (I:C), a dsRNA mimetic, and piconaviruses [33] . Saito et al. [34] and Uzri and Gehrke [35] identified the Hepatitis C virus polyuridine motif of the 3′ untranslated region (UTR) and its replication intermediate as the PAMP substrate of RIG-I. Uzri and Gehrke [35] also showed that the dengue virus full length 5′ UTR and 3′ UTR activated IFN-β to moderate levels. On the other hand, despite the innate immune system's surveillance and detection of PAMP, pathogens like DV1 still establish infection in the presence of a robust innate immune system, as viruses have evolved different mechanisms to counteract the host anti-viral response. For example, DV1 proteins such as NS2A, NS2B, NS4A, NS4B and NS5 could inhibit IFN production [36] , [13] , thus establishing infection in the presence of a functioning innate immune system. Other DNA and RNA viruses have been shown to comprise of IFN inhibitors as well. For example, paramyxovirus V proteins induce STAT degradation or block STAT phosphorylation, inhibiting IFN signaling [37] . Adenovirus E1A proteins and large T antigen proteins of murine polyoma virus also inhibit IFN signaling [38] , [39] . NS1 protein of influenza A virus inhibits IFN-β production [40] .
RIG-I-mediated IFN response was shown to play a critical role in restriction of virus infection in cultured cells. Hepatitis C virus [11] , herpes simplex virus-1 and adenovirus [41] were shown to replicate to much higher titers in RIG-I mutant human hepatoma cells HUH-7.5.1 compared to wild type HUH-7 cells. Modulation of dengue virus infection by IFN has been clearly demonstrated in other studies [42] , [43] . However, despite experimental evidence, the roles played by individual pattern recognition receptors in restriction of dengue virus infection are still not clearly understood. Using a microarray in our earlier study, we demonstrated IFN-related gene induction in DV1-infected cells [14] . A majority of genes, including RIG-I, MDA5 and TLR3, that were strongly up-regulated were IFN-related genes. Previous studies have shown that RIG-I and MDA5 are necessary for establishing an antiviral state against dengue virus [8] . Evidence presented in this study lends further support that both RIG-I and MDA5 are involved in induction of IFN response in dengue virus-infected cells. Interestingly, knockdown of RIG-I gene expression in HUH-7 cells in isolated stable clones showed down-regulation of both RIG-I and MDA5 expression. RIG-I and MDA5 share a limited homology in their overall primary structure [6] . The shRNA used in this study was against a target sequence of RIG-I, as described by Seth and colleagues [44] and no homology to MDA5 sequences was noted during sequence alignment. Thus, it is unclear why silencing RIG-I in HUH-7 cells would also suppress the expression of MDA5. Further investigations are required to investigate if MDA5 expression in HUH-7 cells is RIG-I-dependent or the shRNA sequence is silencing MDA5 expression in unknown ways.
The observation that RNA silencing of RIG-I and MDA5 in HUH-7 resulted in a significant increase in DV1 replication and IFN-β production raised a possibility that the third RNA sensor, TL3, may also play an important anti-DV1 role in this cell system. As the basal level of TLR3 in HUH-7 cells is low [15] , its role in induction of IFN and in restriction of virus infection in this type of cells is not fully appreciated. The fact that knockdown of TLR3 by siRNA in RIG-I- and MDA5-knockdown cells (shRIG-I cells) further enhances dengue virus infection and reduces IFN-β response demonstrates that all the three pattern recognition receptors are implicated in host innate immune response to the same virus in a same infected cell, and may play a synergistic role. Although, as discussed above, DV1 produces IFN antagonists, no viral protein has been shown to interfere in the upstream sensing of viral double stranded RNA by RIG-I, MDA5 or TLR3. As these receptors are inducible in DV1-infected cells [14] , the enhancement effect on viral replication by RIG-I- and MDA5-knockdown, in turn, enhances the TL3 induction and IFN-β response. This may explain why higher levels of IFN-β response were observed in RIG-I- and MDA5-knockdown cells. It also suggests that the presence of minute amounts of a specific receptor in a particular cell type would play an important restriction role. On the other hand, as these three receptors show distinct substrate specificities, it would be reasonable to assume that they may recognize different parts of the viral components. A clear advantage for the host is that the invading pathogens can be effectively captured even though mutations occurred in a certain part of the viral genome. If the virus evades one pathway of detection, another could detect and trigger innate immunity against the virus.
In summary, our data show that RIG-I, MDA5 and TLR3 synergistically activate innate immune response against DV1 infection in HUH-7 cells. Our study shows the involvement of RIG-I, MDA5 and TLR3 in innate immune response against dengue virus in the same cellular system (HUH-7). The results also show that over-expression of RIG-I or MDA5 individually induces weak IFN-β expression as compared to when both genes are over-expressed. However, over-expression of TLR3 alone could recognize DV1 and initiate strong IFN-β response. This study, thus, contributes to ongoing characterization of the innate antiviral response to dengue virus infection in cells. Understanding and identifying the molecular patterns that trigger innate immune signaling may lead to targeted and specific antiviral strategies against dengue virus infection. | Conceived and designed the experiments: AMAN DXL. Performed the experiments: AMAN HHW PT. Analyzed the data: AMAN HHW DXL. Contributed reagents/materials/analysis tools: AMAN HHW PT SX KPL. Wrote the paper: AMAN DXL.
Dengue virus (DV) infection is one of the most common mosquito-borne viral diseases in the world. The innate immune system is important for the early detection of virus and for mounting a cascade of defense measures which include the production of type 1 interferon (IFN). Hence, a thorough understanding of the innate immune response during DV infection would be essential for our understanding of the DV pathogenesis. A recent application of the microarray to dengue virus type 1 (DV1) infected lung carcinoma cells revealed the increased expression of both extracellular and cytoplasmic pattern recognition receptors; retinoic acid inducible gene-I (RIG-I), melanoma differentiation associated gene-5 (MDA-5) and Toll-like receptor-3 (TLR3). These intracellular RNA sensors were previously reported to sense DV infection in different cells. In this study, we show that they are collectively involved in initiating an effective IFN production against DV. Cells silenced for these genes were highly susceptible to DV infection. RIG-I and MDA5 knockdown HUH-7 cells and TLR3 knockout macrophages were highly susceptible to DV infection. When cells were silenced for only RIG-I and MDA5 (but not TLR3), substantial production of IFN-β was observed upon virus infection and vice versa. High susceptibility to virus infection led to ER-stress induced apoptosis in HUH-7 cells. Collectively, our studies demonstrate that the intracellular RNA virus sensors (RIG-I, MDA5 and TLR3) are activated upon DV infection and are essential for host defense against the virus.
Author Summary
Dengue fever, dengue haemmorhagic fever and dengue shock syndrome, which are caused by dengue virus infection, are a major public health problem in many parts of the world, especially South East Asia. The investigation of host cell transcriptional changes in response to virus infection using DNA microarray technology has been an area of great interest. In our previous study, we used microarray technology to study expression of individual human genes in relation to dengue virus infection. Most of the genes that were upregulated were type 1 interferon related genes. To gain a better understanding of the innate immune response to dengue virus, we knocked down RIG-I, MDA5 and TLR3 genes in HUH-7 cells. Silencing these genes using siRNA technology resulted in significant increase in viral replication. This increase in viral load induced ER stress leading to apoptosis. This study demonstrates a synergistic role for RIG-I, MDA5 and TLR3 in restricting dengue virus infection. | CC BY | no | 2022-01-13 05:08:15 | PLoS Negl Trop Dis. 2011 Jan 4; 5(1):e926 | oa_package/04/7b/PMC3014945.tar.gz |
|||
PMC3014946 | 21245913 | Introduction
Burkina Faso is one of the poorest countries in the world ranking 177 th out of 182 according to the Human Development Index (HDI) [1] . Its economy relies predominantly on agriculture (40% of GDP), with cotton production being the most important, followed by livestock production, accounting for 12% of the GDP [2] . The main livestock species are cattle, small ruminants, poultry and pigs. The relative importance of each species varies across the thirteen regions of the country. The country's pig population is estimated at 2 million, and is more concentrated in the regions of Centre West, South West, and Boucle du Mouhoun (with 45% of the pig population) [3] . Pigs are kept mainly in a traditional way, with most being either tethered or allowed to roam freely for some time during the year ( Figure 1 ).
Porcine cysticercosis is a parasitic zoonosis caused by the larval stage of Taenia solium . While pigs are the intermediate host, man is the only natural definitive host. Human tapeworm carriers shed thousands of eggs daily through their feces. Pigs usually get infected by eating infected human feces or by consuming feed or water contaminated with human feces. Humans can also become accidental intermediate hosts upon ingestion of T. solium eggs. In both humans and pigs, the larval stage of T. solium can establish in the muscles and/or in the brain, the latter resulting in neurocysticercosis (NCC). Human NCC may lead to acute seizures, epilepsy and other neurological manifestations [4] . Ingestion of larvae (cysticerci) present in raw or under-cooked pork may result in human tapeworm infection. Porcine cysticercosis is common in developing countries where pigs are raised. Porcine cysticercosis is associated with poor sanitation and hygiene (absence of latrines, defecation in pigpens, poor handwashing practices), poor methods of pig husbandry (free-roaming), lack of meat inspection, and poor knowledge of the disease [5] – [11] , all of which are associated with poverty. In order to avoid economic losses due to condemnation of infected pig carcasses in places where meat inspection is performed [10] , farmers may sell contaminated pigs, either alive or for clandestine slaughtering.
Information on porcine cysticercosis in Burkina Faso is very limited. Coulibaly and Yameogo (2000) [12] reported a prevalence of cysticercosis of 0.6% among pigs inspected at slaughter. Based on data from the main abattoir in Ouagadougou for 2005 to 2007, of 10,505, 12,651, and 11,887 pigs slaughtered during each of these three years, respectively, only 8 (0.08%), 13 (0.10%), and 10 (0.08%) were condemned because of cysticercosis; whereas 472, 186, and 133 carcasses, respectively, were reportedly condemned because of illegal slaughtering. These data clearly show that inspection of pork is very poor in Burkina, and reliable prevalence of porcine cysticercosis is lacking in the country. The aims of this study were therefore: 1) to estimate the prevalence of T. solium cysticercosis among pigs in three villages with distinct pigs raising techniques, and 2) to measure the association between potential risk factors and the prevalence of infection in pigs. | Methods
We conducted a cross-sectional study in three villages including serological detection of active cysticercosis in pigs and a questionnaire survey for the identification of potential risk factors. In two villages, Batondo and Pabré, located 140 km west and 20 km north of Ouagadougou (the capital of Burkina Faso), respectively, pig breeding and pork consumption are very common. In a third village, Nyonyogo, located 30 km north of Ouagadougou, both pig breeding and pork consumption are rare. The sampling of pigs took place between 26 May and 29 October 2007, which corresponds to the rainy season and the very start of the dry season. The sampling was implemented by a research team composed of one physician, one veterinarian, two interviewers, and one translator (for Batondo only).
Sampling and measurement of potential risk factors
The first step in the sampling process was to determine the sampling frame. This was done by using information from the national population census conducted in 2006 during which all villages were divided into enumeration areas (EA), a geographical unit which is intended to include approximately 1,000 persons. In each of the three villages, the sampling process started by identifying the geographic limits of each EA. In each EA, each concession (a grouping of several households, usually members of the same family) was numbered. In Batondo (4 EAs) and Nyonyogo (3 EAs), all the concessions were included because of the small number of concessions present. In Pabré (5 EAs), 50% of all concessions were randomly selected.
In each concession, all the households were invited to participate. The head of each household was asked to list all the members of the household and the mother or the oldest woman (in the case of polygamy) was asked questions regarding selected characteristics of the household, such as the source of drinking water, the presence of latrines, breeding livestock, and cooking pork. The person in charge of pig farming was asked about how pigs were managed and slaughtered. Knowledge of porcine cysticercosis was assessed through questions about seeing lesions in the meat on dead animals or cysts under the tongue of live animals. Before the data collection started, the questionnaires were validated during a pre-test conducted in Tenado, a village located eight km from Batondo. One pig was randomly sampled per household for blood sample collection. Blood samples were also collected in humans (one person randomly selected per household) (see [13] for more details on the human component of the study). All questionnaires are available on request.
Blood samples were left to decant and the sera were transported to the Institut de Recherche en Sciences de la Santé (IRSS) in Bobo-Dioulasso where they were kept at −20°C until analysis.
Serological test
The serum samples were tested in duplicate with the enzyme-linked immunosorbent assay (ELISA) for the detection of circulating antigens of the metacestode of T. solium (Ag-ELISA) [14] . Sera showing a coefficient of variation of more than 50% between the duplicate results were considered as missing values (n = 10). The Ag-ELISA has been reported to have a sensitivity of 76.3% (95%CI: 60.9%–88.6%), 86.7% (95%CI: 62.0%–98.0%) and 85.8% (95%CI: 71.9%–99.7%) and a specificity of 84.1% (95%CI: 74.4%–93.3%), 94.7% (95%CI: 90.0%–99.7%) and 98.9% (95%CI: 97.3%–100%) in pigs in South Africa, Zambia and West Cameroon, respectively [6] , [14] , [15] .
Statistical analysis
First, the relative frequencies of categorical variables and the mean and extreme values of quantitative variables were calculated. Wealth quintiles for households were derived from a score calculated based on the asset ownership of each household (e.g., source of drinking water, presence of toilet, radio, bicycle, presence, type and number of livestock, types of roof and walls) using factor analysis as described by Gwatkin et al. (2000) [16] . Secondly, bivariate analyses were conducted to assess the associations between pigs' serological status and location (village), human socio-demographic and health factors (source of drinking water, presence of latrine, presence of antigens to the larval stages of T. solium , wealth quintile), and pig management factors. Finally, a multiple logistic regression was used to estimate the adjusted associations between the factors mentioned above and the presence of antigens to T. solium in pigs.
Data from Nyonyogo were excluded from the logistic regression because of the small number of pigs in the village (only 6 pigs were sampled there). All the analyses were performed using STATA9 software.
The prevalence of infection and 95% Bayesian credible intervals (BCI), adjusted for measurement error in serology, were also estimated in Batondo and Pabré using a Bayesian approach suggested by Joseph et al. (1995) [17] . We used two sets of priors for the sensitivity and specificity of the AgELISA test based on two studies conducted in infected pigs, one in Zambia and one in South Africa [6] , [14] .
Ethical issues
The animal component of this protocol was approved by the Institutional Animal Care and Use Committee (IACUC) of the University of Oklahoma Health Sciences Center. No IACUC exists in Burkina Faso. The human component of this protocol was reviewed and approved by the ethical committee of the Center MURAZ (Ref. 02-2006/CE-CM) and by the Institutional Review Board of the University of Oklahoma Health Sciences Center (IRB# 12694) for both human and porcine participants. Informed consents for the interviews of participants and the provision of blood samples were obtained separately. The consent process was done orally because a very large proportion of the population had never been to school (62.5%). Oral consent was documented on the individual consent forms by the research staff (fingerprints were collected in place of signatures). Both IRBs approved the use of oral consents. | Results
Socio-demographic characteristics of the three villages are described in Table 1 . The majority of people in Batondo and Pabré drank water from traditional or open wells, whereas in Nyonyogo, most used water from bore wells. There were a few households in Batondo (1.2%) and Pabré (0.3%) in which people drank water from the river, marsh or spring. More than one-third of households in Pabré had latrines in contrast to Batonto and Nyonyogo where latrines were very rare. Wealth quintiles classified approximately one-half (50.9%) of households in Batondo among the poorest compared to around 1% in Pabré and Nyonyogo. In contrast, nearly 30% of households in Pabré and Nyonyogo were classified among the wealthiest compared to only 4.4% in Batondo.
Characteristics related to pigs and pig management are reported in Table 2 . Pigs were raised in two-thirds (66.4%) of the surveyed households in Batondo and in more than half (52.8%) of the surveyed households in Pabré. In Nyonyogo, most surveyed households owned ruminants and only 3% owned pigs. In Batondo, pigs were almost exclusively cared for by women (98.8%) whereas in Pabré, pigs were taken care of by men in nearly two-thirds of cases (63.1%). Similarly, the mother or eldest woman of the family was in charge of pigs in 71.7% of surveyed households in Batondo, but in only 25.0% of surveyed households in Pabré. Children were more often in charge of pigs in Pabré (10.5%) than in Batondo (1.8%). More than one-half of the surveyed households owning pigs in Pabré kept them penned during the rainy season. This was in contrast to Batondo where only 10.4% were kept penned, although over 90% were tethered. During the dry season, pigs were left to roam in all three villages to scavenge around cropped fields.
A total of 336 pigs were sampled: 173 in Batondo, 157 in Pabré, and 6 in Nyonyogo. Four sera in Batondo, 3 in Pabré, and 1 in Nyonyogo were not tested because of insufficient volume. Using prior information on the sensitivity and specificity of the AgELISA from South Africa, the adjusted estimates of the seroprevalence of infection in Batondo and Pabré were 32.7% (95% BCI: 12.1%–54.3%) and 48.3% (95% BCI: 27.9%–74.8%), respectively. Using the priors from Zambia, the adjusted estimates of the seroprevalence of infection in Batondo and Pabré were 32.7% (95% BCI: 25.4%–68.3%) and 48.2% (95% BCI: 35.4%–82.6%), respectively. A total of 1.9% and 3.4% were found to be borderline positive in Batondo and Pabré, respectively. Neither serologically positive nor borderline animals were found in Nyonyogo.
The results of the multivariable logistic regression are reported in Table 3 . Allowing pigs to roam during the rainy season was associated with an increased prevalence odds of infection when compared to keeping pigs penned (POR = 6.48, 95%CI: 1.23–34.17). Being aware of pig cysticercosis was also associated with an increased prevalence odds of infection (POR = 4.70, 95%CI: 1.76–12.52). The village, while not statistically significant, was retained in the model because it was a confounding factor for most of the other variables explored, including those retained in the final model. Once the model is adjusted for pig management during the rainy season and awareness of the infection, the prevalence odds of infection tended to be higher in Pabré than in Batondo (OR = 1.63, 95%CI: 0.94–2.83). | Discussion
This is the first community-based study conducted in Burkina Faso to report seroprevalence of porcine cysticercosis. Our results show that cysticercosis is highly prevalent in the two villages where pigs are commonly raised. These estimates are nearly 100 times higher than the previous estimate of 0.57% based on reports from meat inspection services [12] . This finding supports the opinion of Zoli et al. (2003) [10] that data from meat inspection are not representative of the real situation, mainly because the method is insensitive [14] and depends on the technical skills and the motivation of the examiner. Pigs infected with cysticercosis may also never be sold in the official market.
One-third (32.5%) of the pigs in Batondo and more than one-third (39.6%) in Pabré were infected, but none of the five sampled pigs in Nyonyogo was found to be positive. The latter is not surprising since pig breeding and pork consumption are very rare in this village. Similarly, Carabin et al. (2009) [13] reported very few cysticercosis seropositive inhabitants in Nyonyogo, and the few positives had all borderline reactions to the Ag ELISA.
Using priors from either South Africa or Zambia to adjust for measurement error resulted in the same median prevalence, but 95% BCI were shifted towards higher values when using the Zambia priors. This is because both the prior values for sensitivity and specificity from the Zambian study were higher than the South Africa values. The B158/B60 AgELISA test used here was recently shown to have the highest sensitivity and specificity for detection of cysticercosis-infected pigs when compared to tongue examination, EITB, and HP10 Ag-ELISA [6] , [14] . The seroprevalences in Batondo and Pabré were comparable to that reported in pigs in Mozambique [5] but were higher than that reported in several other African countries [9] , [15] , [18] , [19] and lower than the seroprevalence reported in pigs slaughtered on the clandestine market in Lusaka, Zambia [14] and in the Eastern Cape province in South Africa [6] .
Pigs allowed to roam some of the time during the rainy season were more likely to be seropositive than pigs penned during all of the rainy season. This association was statistically significant even though the proportion of pigs allowed to roam during the rainy season was small because the association was so strong (POR = 6.48, 95% CI 1.23–34.17). Given the imprecision of this estimate, however, the association will need to be further explored. This observation agrees with the findings of Pondja et al . (2010) [5] who identified the free-range pig husbandry system as an important risk factor for porcine cysticercosis.
Conducting this pilot study lead to an appreciation that the pig management systems in Batondo and Pabré were different from what had been reported to us when the study was being planned. Indeed, in both of the two villages with pigs, animals were left roaming during dry seasons but restrained to one degree or another during crop production period (rainy season). The main difference in restraints was that pigs were confined in pens in Pabré and tethered to pillars in Batondo.
The difference in pig management, mainly during the rainy season, between Batondo and Pabré might reflect the difference in who owns and is responsible for raising the pigs. In Batondo, almost all pigs were owned by women whereas they were kept by men in two-third of the cases in Pabré. This result suggests that the target for intervention when planning future cysticercosis control strategies depends on who is responsible for pig raising in any given village.
Pigs owned by people who had heard of porcine cysticercosis were more likely to be seropositive than owners who were unaware of this infection. If pigs raised by the same farmers tend to become re-infected, this would suggest that people become aware of porcine cysticercosis when their pig is found to be infected.
Despite the high frequency of households with latrines in Pabré and the low prevalence of active infections in humans [13] , the seroprevalence in pigs in this village tended to be higher than that in Batondo. This result is counter-intuitive but could be explained by the following. First, the presence of latrines does not certify their use because of foul odors, flies, flooding, etc. Second, latrines may be inappropriately built (closing door, septic tank) allowing access of pigs to human feces [20] . More research is needed to explain this observation.
It was anticipated that there would be an association of pig seropositivity and variables such as presence of a toilet in the household, household source of drinking water, and human serological status, but this was not confirmed by the statistical analysis. Indeed, the structure of community life in these villages means that any risk associated with lifestyle in a family is shared by other families nearby. For example, if a given family does not have access to latrines and thus defecates in nature, this will also affect the members of families equipped with latrines through environmental contamination. People with taeniasis may also infect a large number of pigs if they do not use latrines. Lescano et al. (2007) [21] demonstrated that a T. solium carrier can infect pigs within an area of 500 meters or more.
In conclusion, a high prevalence of porcine cysticercosis was observed in pig populations in Burkina Faso in villages with traditional pig-rearing practices. The co-occurrence of knowledge about cysticercosis and owning seropositive pigs suggests that information about this disease may not be available to livestock managers until the disease is already present. If so, this emphasizes the clear need for improving education in order to control this zoonosis. | Conceived and designed the experiments: RG NP HC AM PD SH LDC. Performed the experiments: RG AM ZT SH AS PN. Analyzed the data: RG HC. Contributed reagents/materials/analysis tools: NP ZT PD AS PN LDC. Wrote the paper: RG NP HC AM ZT PD SH AS PN LDC. PI for the study: HC AM.
Background
Little is known about porcine cysticercosis in Burkina Faso. We conducted a pilot study to estimate the prevalence of antigens of Taenia solium cysticercosis and to identify associated factors in pigs of three villages in Burkina Faso, selected to represent different pig management practices: one village where pigs are allowed to roam freely (Batondo), one village where pigs are penned part of the time (Pabré) and one village with limited pig farming (Nyonyogo).
Methods/Principal Findings
A clustered random sampling design was used. Data on socio-demographic characteristics (source of drinking water, presence of latrines in the household, type and number of breeding animals) and pig management practices were collected using a standardized questionnaire. Blood samples were collected from one pig per household to determine the presence of antigens of the larval stages of T. solium by the B158/B60 Ag-ELISA. The associations between seropositivity and socio-demographic and pig management practices were estimated using logistic regression. Proportions of 32.5% (95% CI 25.4–40.3), 39.6% (31.9–47.8), and 0% of pigs, were found positive for the presence of circulating antigens of T. solium in Batondo, Pabré, and Nyonyogo, respectively. The results of the logistic regression analyses suggested that people acquire knowledge on porcine cysticercosis following the contamination of their animals. The presence of antigens in the pigs' sera was not associated with the absence of latrines in the household, the source of drinking water or the status of infection in humans but was associated with pig rearing practices during the rainy season.
Conclusions/Significance
The results suggest that education of pig farmers is urgently needed to reduce the prevalence of this infection.
Author Summary
Taenia solium cysticercosis is a neglected tropical infection transmitted between humans and pigs. This infection is particularly common in areas where sanitation, hygiene and pig management practices are poor, and can sometimes lead to epilepsy in humans. There is very little information about the importance of this infection in Burkina Faso, even though pork meat is widely consumed in many villages. We conducted a pilot study in three villages: two villages where pig rearing and pork consumption are common (Batondo and Pabré) but with different pig management practices, and one village with limited pig farming and pork consumption (Nyonyogo). Blood tests were done on pigs and information on pig raising was collected from farmers. Our study demonstrated that at least one third of pigs are infected with cysticercosis in villages where they are raised, and, particularly when pigs are left to roam some or all of the time. It also demonstrated that farmers may not be aware of this disease until one of their animals is found to be infected. Thus, the study concluded that there is an urgent need for improving education in order to control this tropical disease. | We would like to thank Dr. Adèle Kam from IRSS for her substantial assistance in the analyses of the blood samples. We are also thankful to Henri Somé for all his help in the database structuring and data entry, and to Dr Christian Kompaoré, Alida Da, Chantal Millogo and Leopold Bado for collecting the data. Finally, we are grateful to Dr Raphaël Nébié and Elie Sawadogo and to all the inhabitants from the three villages for their wonderful welcome and participation. | CC BY | no | 2022-01-13 05:08:15 | PLoS Negl Trop Dis. 2011 Jan 4; 5(1):e927 | oa_package/d0/60/PMC3014946.tar.gz |
||
PMC3014947 | 21245914 | Introduction
Chromatin in higher eukaryotes is compacted and folded on several scales. While the folding mechanisms on the scale of the chromatin fiber are rather well understood [1] , [2] , the higher-order arrangement of whole chromosomes inside the nucleus remains an open question [3] – [5] . Chromosomes, in comparison to other polymeric systems, display a vast amount of unexpected types of behaviour. Most importantly, chromosomes are highly compartmentalized objects being well separated during interphase [6] . Such a separation is not only observed regarding complete chromosomes, rather Mb-sized stretches of chromatin, when labelled fluorescently with different colors, also display little intermingling [7] . A lot of speculation has been going on about the mechanisms driving such kind of compartmentalization. It has been proposed that compartmentalization results from non-equilibrium effects: During metaphase chromosomes are condensed and well separated. The entanglement time disregarding topoisomerase-II activity is supposed to be much larger than the lifetime of the cell [8] . Based on the assumption of a linear polymer model, calculations show that this holds also true given the activity of topoisomerase-II [9] .
Nowadays however, there is an ever growing body of evidence that chromatin is not organized as a simple linear polymer. Rather functional loops seem to play a pivotal role in transcriptional regulation of higher eukaryotes. Several experimental techniques have allowed the determination of specific intra- as well as interchromosomal contacts [10] – [12] . 4C experiments revealed that abundant short as well as long-range chromosomal contacts are established ranging from a few kb to several Mb, these contacts being cell-type specific. The formation of chromatin loops has been shown to be important for transcriptional regulation in the -globin locus [13] . Experimental evidence further suggests that loop formation involves specific proteins like CTCF [14] .
Recently, several computational studies have been put forward investigating the entropic effects of looping. It was shown that ring polymers in proximity show less intermingling than corresponding linear systems [15] . Cook et al. [16] simulated linear chains as well as rosette-like structures in a dense system and found that the probability of inter-chain contacts decreases by the transition from a linear to a looping polymer. In another study, the potential of mean force arising when two ring polymers are brought in proximity has been analyzed [17] . A three-fold increase in the repulsion compared to linear chains (self-avoiding walks) at short separations was found. Topological constraints arising from the non-catenation of rings results in a further increase of repulsion. However, it remains an open question, whether similar predictions can be derived for a system of chromosomes.
The scope of this paper is to extend the study of effective interactions between ring polymers to the more complex system of looping chromosomes. Chromsomes are simulated using the Dynamic Loop (DL) model [18] , which has been shown to be consistent with a variety of experimental observation. Among others, the model correctly displays the observed folding into a confined sub-space of the nucleus [19] , the power-law decay of the contact probabilities determined in a recent study [12] and the formation of chromosome territories (CTs). This DL model assumes that loops form on the basis of diffusional motion, two chromatin segments having a certain probability to stick together for a certain time when being in proximity. Amongst others, we want to investigate how the strength of the repulsive interaction changes when introducing more and more loops into the system. While a coarse-grained model of the chromosome is employed here, a mapping to physical units can be conducted using results from fluorescence in situ hybridization (FISH) experiments [19] . | Materials and Methods
Computer simulations of chromosomes
Model chromosomes are simulated using the Dynamic Loop model introduced in Ref. [18] . The Dynamic Loop model is implemented employing lattice Monte Carlo simulations [27] in order to simplify the handling of excluded volume. Calculation of excluded volume interactions thereby is reduced to checking whether one lattice site is already occupied or not. To allow for more flexibility than a simple local-move algorithm on a cubic lattice, we employ the bond-fluctuation method introduced by Carmesin [28] allowing 108 different bond vectors; the length of a bond can take the values [29] . The bond-fluctuation model is especially suited for dense and compact systems where a lattice algorithm would no longer be feasible due to high rejection rates during the Monte Carlo process. The simulation method applied fulfills the following important features: (i) it produces unbiased results, i.e. each possible conformation out of the ensemble is sampled with equal probability, (ii) it takes into account excluded volume interactions, i.e. two monomers are not allowed to occupy the same region in space and (iii) using some restrictions on the moves and bond vectors it ensures that no bond crossings can occur during a Monte Carlo step, i.e. it preserves the topological state of the conformation. The algorithm conducts only local moves in order to resemble the dynamics of real polymers [28] . Using a coarse-grained lattice approach is reasonable as we are only interested in features of looping chromatin independent on local structure. Coarse-graining allows us to abstract from the complex environment and highlight the main driving forces and effects of chromatin folding.
We perform simulations of isolated chromosomes with coarse-grained lengths ranging from to . Chromosomes are initially equilibrated as self-avoiding walks using local moves of a monomer to one of the nearest neighbors on the lattice. After the initial equilibration steps, the Monte Carlo algorithm allows for the formation of loops. After each Monte Carlo trial move, one monomer is selected at random. It is then checked whether another monomer on the same chain is in the neighborhood, i.e. co-localized. The co-localization condition is fulfilled whenever the distance between the monomers is less than 3 lattice units. If the two monomers are co-localized, then a loop is formed with a certain probability . If the loop is created, it is assigned a certain lifetime which is drawn from a Poissonian distribution where the parameter determines the average lifetime of the loops. The lifetime of the loops is chosen here to be As it has been shown [17] that the chosen lifetime does not strongly influence equilibrium properties, parts of the analysis presented here are restricted to the choice .
Since subsequent conformations in the Markov chain created by the Monte Carlo algorithm are highly correlated, one has to perform a certain number of Monte Carlo steps to obtain two independent conformations. For each set of parameters (chain length , looping probability and lifetime of loops ) we determine the autocorrelation function (see e.g. Ref. [30] ) of the squared radius of gyration . The integrated autocorrelation time is then determined using the windowing method introduced by Sokal [31] .
Calculation of the effective potential
To analyze the strength of the repulsive interactions, the potential acting between the chromosomes' centers of mass is determined using the method introduced by Dautenhahn and Hall [24] . In short, two equilibrated isolated chromosome conformations are selected and shifted such that the distance between their centers of mass equals . If the excluded volume condition is satisfied, i.e. no lattice site is occupied by more than one bond, the conformation is accepted, otherwise it is rejected. The fraction of accepted conformations to the total number of trial conformations defines the effective potential at distance , From the set of accepted two-chain conformations, the conformational properties can be calculated. | Results
Simulation of chromosomes
To obtain information about the repulsive interactions between chromosomes, we use the recently developed Dynamic Loop (DL) model. This model initally assumes chromatin to consist of a coarse-grained linear polymer chain. Loop formation is achieved using diffusional motion of the monomers in the following way: Whenever two segments co-localize by diffusional motion, a chromatin loop is formed with a certain probability between these two sites. The loop is assigned a certain lifetime, thus loop attachment points dissolve again during the course of time.
The stochastic nature of loop formation provides a method to effectively incorporate protein-chromatin and chromatin-chromatin interactions. Probabilistic looping, which is often thought to be mediated by DNA-binding factors [14] , [20] , [21] or by transcription factories [22] , mimics the effect of protein concentration (there being either proteins binding DNA sites or not) and binding affinity. In the following we denote by “loop” a functional interaction between two parts of a chromatin fiber existing for a certain time as created by the method. In contrast, a “contact” denotes two parts of the chromatin fiber close together by thermal fluctuations without necessarily being an interaction.
A typical human chromosome has a length of about 100 mega basepairs (Mb), rendering a detailed description on the molecular level computationally impossible. Coarse-graining approaches, where a long stretch of chromatin is modeled as an effective monomer, are well justified on a scale above the persistence length nm [23] . Thus, it is reasonable to conduct computer simulations on a coarse-grained scale where it can be securely assumed that the fiber is flexible.
The effective repulsion between chromosomes increases strongly with loop number
What happens, when two polymeric coils are brought close together? Clearly, in the absence of other interactions than excluded volume forces, polymers repell each other due to the constrained conformational space available. Such a behaviour has been found both for linear self-avoiding walks [24] and ring polymers [17] . Here, we investigate the potential of mean force between the centers of mass of two chromosomes modelled by the Dynamic Loop model. Results are shown for chains of length in Figure 1 . In principle, dependent on the coarse-graining used, such chains could represent small chromosomal regions up to whole chromosomes. To allow comparison for different sets of parameters (chain length , looping probability , lifetime of loops ), the center-of-mass distance is scaled with the mean radius of gyration of the corresponding isolated chains. The radius of gyration is a measure of the typical size of a chromosome, i.e. the chromosome territory.
Evidently, the effective potential increases when approaching the two chromosomes, i.e. lowering the center-of-mass distance . This result is expected, as the accessible conformational space becomes smaller the more the monomer clouds are in proximity. More importantly, we find that the effective potential is pronouncedly stronger for chromosomes with a large average number of loops compared to linear chains. To assess whether these results are dependent on the chain length, and therefore the level of coarse-graining used, we have plotted the same results for chain lengths and 512 in Figure S1 . Obviously, the dependence on chain length using the scaled center-of-mass distance is rather subtle, indicating that the level of coarse-graining used in simulations does not affect the results on the large scale. This is well-known to be true for self-avoiding walks and ring polymers, where the effective potential at full overlap adopts a constant value in the order of in the limit of infinite chain length. For the Dynamic Loop model, a comparison is more difficult, since the effective potential is also a non-trivial function of the looping probability . Importantly, the repulsive potential increases most strongly in the range where , i.e. around the size of the chromosome territory, indicating a huge energy cost for a high degree of intermingling of chromosome territories.
To demonstrate how the data can be mapped onto physical units, we use model polymers of chain length . We set one coarse-grained bead to 400 kb in order to model a sufficiently long stretch in the size range of a typical chromosome. To determine the spatial extend of this stretch of chromatin, we employ long distance experimental data from human chromosome 11 [19] . In principal, such a mapping is always connected with a lot of uncertainy: The detailed Kuhn length is not known, disallowing for a precise mapping on the short scale; as chromatin is organized much more complex than a linear chain, other parameters (looping, binding, heterogeneity) enter the calculations. To obtain a simple mapping, we adjust the plateau level of the model polymers to that of experimental data. Figure 2A shows the results of the mapping using 177 nm for one lattice unit. Clearly, given such a coarse-graining approach, each polymer segment has to be considered as a bunch of chromatin, making it impossible to resolve the detailed interactions leading to the loop on a molecular basis. Importantly, the model displays well the leveling-off observed in experiments for intermediate looping probabilities (the cyan symbol ▴ corresponds to 131 loops on average). The effective potential in units of is displayed in Figure 2B . While the effective potential profile is rather flat for self-avoiding walks, the existence of loops leads to a strong increase in the potential at distances of about 2–3 m, the region where the experimental data displays a leveling-off. Importantly, these results are independent of the looping lifetime ( Figure S2 ).
The quantitative increase in the effective potential of looping polymers over linear chains (self-avoiding walks) is shown in Figure 3 . The factor by which the effective potential of the model chromosomes is larger than that of the linear chain is plotted on the -axis. Standard errors are in the size range of the symbols and therefore not displayed. Likewise, the abscissa shows the center-of-mass distance scaled by the radius of gyration . We find that the repulsive potential is stronger by more than one order of magnitude for chains in the parameter range where leveling off occurs.
Looping polymers become aspherically elongated
Chromosomes, when brought in close proximity, not only reveal a strong repulsion between their centers of mass; besides this, their structural properties undergo significant changes. Here, we investigate how size and shape of a model chromosome changes in presence of a second one. Such effects play an important role inside the cell nucleus, as chromosomes are located in a complex environment being typically separated by only a few m; comparison to linear chains allows us to learn something about the effect of looping. The change in dimensionality is measured by the swelling factor , given by Here, denotes the root mean squared radius of gyration for a chromosome being in a distance to a second one. denotes the corresponding quantity for isolated chromosomes. In a recent study on topological effects between ring polymers [17] it has been shown that both linear chains as well as ring polymers swell when being brought together, the swelling factor being about 10% for rings and slightly smaller for linear chains.
While linear chains and ring polymers only show a mild swelling in the presence of a second chain, we find that model chromosomes swell enormously. Figure 4A displays for different looping probabilities , i.e. different values of the average number of loops, in relation to the scaled center-of-mass distance . To demonstrate similar behaviour independent of the level of coarse-graining, results are shown for different chain length . Swelling factors are strongly dependent on the average number of loops, increasing by a factor in the order of 2–10 for the range of looping probabilities where a leveling-off is observed (cf. also Figure 2A ). In fact, diverges for large loop numbers, indicating that the chains can not be approached closer than approx. .
The swelling of the chromosomes might suggest that they open up to create space for the monomers of the other chromosome, i.e. allow for intermingling. In the following, we will show that this is not the case, rather the contrary is observed. To achieve this, we investigate how the shape of the chromosomes changes when being close together. The asphericity of the gyration ellipsoid has been established [15] , [25] as a measure of shape, being zero for a spherically shaped polymer and unity for a rod-like polymer, Here, , and denote the eigenvalues of the gyration tensor. To highlight the changes in asphericity when approaching two chromosomes, we show the ratio in Figure 4B , being the asphericity of an isolated model chromosome with the same parameters. While the change in asphericity for linear chains (self-avoiding walks, no loops) is rather small even at full overlap (about 20%), we find a pronounced aspherical deformation on our model chromosomes in the regime of looping probabilities that force a leveling-off in the mean square distance. Asphericity values increase by about 200–400% at genomic separations of , i.e. the typical size of the chromosome.
The changes in shape and dimension are visualized in Figure 4C . Shown are the average gyration ellispoids of three different model polymers: (i) linear chains (self-avoiding walks, 0 loops), (ii) chromosomes with 86 loops on average (purple symbol ▴ in Figure 2B ) and chromosomes with 131 loops on average (cyan symbol ▴ in Figure 2B ). For each set of model parameters, the ellipsoids are displayed for three different center-of-mass distances: isolated chains (infinite CM distance), and . We find that isolated linear chains require a huge amount of space, while looping polymers are pronouncedly smaller and more spherical. When being in contact with a second chromosome, the shape of self-avoiding walks changes only slightly, while chromosomes with loops become markedly aspherical compared to their isolated shape.
These findings indicate that chromosomes, similar to ring polymers, do not swell in order to create space for the second chromosome. On the contrary, they swell to avoid each other by elongating in such a way that the overlapping is minimized. This is achieved by aligning the highly aspherical gyration tensors in a parallel way when close together.
Looping polymers avoid intermingling
To answer the question whether chromosomes swell to create space for each other or if they rather try to avoid each other, the mutual alignment of the polymers is studied. An established measure for the mutual alignment is given by the average angle between the gyration tensors largest principal axes [15] , [17] . In case of the chromosomes being adjusted independently of each other, the average angle would adopt the value of [15] . This value arises by averaging over two randomly oriented orientations in three-dimensional space. Deviations from this value indicate a tendency of the polymers to align in a certain non-random way with respect to each other. Figure 5 displays results for linear chains (black symbols) and chromosomes with loops in relation to the center-of-mass distance for chains of length . The symbol and color code used is the same as in Figure 2 . We find that chromosomes display a pronouncedly stronger tendency to align perpendicular at short center-of-mass separations than linear chains or ring polymers (cf. 5A). Similarly, a slightly parallel alignment can be found at intermediate distances.
These findings might be explained by a tendency of the chromosomes to minimize the overlap area. When the distance between chromosomes is lowered to values in the order of the size of the chromosome, they start to feel the presence of the second chromosome. Thus, the space of accessible conformations is reduced and the chromosome stretches in a direction perpendicular to the center-of-mass axis. However, when the chromosomes are forced even closer together, nearly overlapping completely, the observed perpendicular alignment together with the strong elongation minimizes the volume shared by both chains ( Figure 5B ).
To quantitatively assess the amount of intermingling between the model chromosomes in dependency of the average number of loops, we project the monomer positions to the line connecting the centers of mass of both chromosomes. Thus, a density distribution can be obtained as shown in Figure 6A . Here, chromosomes with a coarse-grained length of have been simulated. Mapping is done according to the procedure described above. Results are shown for different looping probabilities and a fixed center-of-mass distance m. The average number of loops being 0 (linear case, black symbol), 86 (purble symbol ▴) and 131 (purple symbol ▴). Clearly, introducing loops in the chromatin structure results in more compact polymers, the monomers being distributed closer around the centers of mass. We determine the overlap fraction by integrating the overlap area between the distributions of both chromosomes. The results are shown in Figure 6B using two different center-of-mass separations m and m. For a center-of-mass distance of 2 m, being comparable to the size of chromosomal regions [19] , we find that the overlap fraction decreases strongly from about 0.7 down to less than 0.1 for large loop numbers. Interestingly, in the range where leveling-off occurs, overlap fractions are in the range of 20–30%. These values are, however, an overestimate, as the projection procedure does not capture segregation in the direction perpendicular to the line connecting the centers of mass. Although not directly comparable to experimental results, these values are in the size range of experimental data from FISH cryo sections, where an overlap volume of 20% has been observed [26] . | Discussion
In this paper, we have analyzed the effect of loops on the repulsive interactions between polymers. As a measure for these interactions we applied the theory of effective potentials, where monomeric degrees of freedom in the partition sum are traced out. The resulting effective potential gives the interaction between both polymers in dependence of their center-of-mass distance . has been determined for self-avoiding walks in recent decades [24] and for ring polymers [17] . Both linear polymers and rings display a repulsive interaction at full overlap ( ), asymptotically converging to a finite value in the order of in the limit of large chain lengths. Here, we applied the concept of effective interactions to the Dynamic Loop model, which has been proposed as a model for chromatin organization in Ref. [18] .
The major finding of this study is that introducing dynamic loops in the structure of chromatin results in a strong increase of the repulsive interactions by about one order of magnitude ( Figure 1 ). Using a mapping to physical units based on recent FISH experiments, we found that the repulsive forces are strongest at center-of-mass separations of m, i.e. the size of the chromosome territory ( Figure 2 ). These observations indicate that chromatin looping plays a dominant role in the entropy-driven segregation of chromosomes.
Moreover, we found that the existence of loops introduces strong changes in the size and shape properties of the chromosomes. Indeed, when being brought close together, looping polymers swell and become pronouncedly aspherical, the observed effect being multi-fold larger than for linear chains or ring polymers. Chromosomes in proximity display a highly non-random orientation of their gyration ellipsoids. Increasing the number of loops leads to a significant decrease in the overlap fraction.
Clearly, the modelling approach pursued here is simplified with regard to the cellular system. In reality, we have a dense system of chromosomes with a chromatin content of about 10%. It remains to be studied in further publications, whether bringing multiple chromosomes together leads to even stronger effects. Here, for ease of simplicity, we focussed on a two-chromosome-system because it allows best to highlight pairwise interactions between chromosomes and the effect of loops while keeping simulational effort reasonable. Studying the impact of multiple chromosome in a dynamic system the future might also shed light on the question whether the proximity between chromosomes reversly might affect looping probabilty. This, however, is not possible in the framework of the Dynamic Loop Model as presented here because it takes the looping probability as an input parameter.
Our findings indicate that chromatin loops not only play an important role in transcriptional regulation. Rather, they help to impose a certain state of order and segregation. Thus, loops seem to constitute a highly efficient regulatory mechanism concerning gene regulation as well as chromatin compartmentalization. The Dynamic Loop model used in this study refrains from assuming active driving mechanisms for loop formation, rather loops form by diffusional motion, minimizing the energy cost. | Conceived and designed the experiments: MB DH. Performed the experiments: MB. Analyzed the data: MB DH. Wrote the paper: MB DH.
One striking feature of chromatin organization is that chromosomes are compartmentalized into distinct territories during interphase, the degree of intermingling being much smaller than expected for linear chains. A growing body of evidence indicates that the formation of loops plays a dominant role in transcriptional regulation as well as the entropic organization of interphase chromosomes. Using a recently proposed model, we quantitatively determine the entropic forces between chromosomes. This Dynamic Loop Model assumes that loops form solely on the basis of diffusional motion without invoking other long-range interactions. We find that introducing loops into the structure of chromatin results in a multi-fold higher repulsion between chromosomes compared to linear chains. Strong effects are observed for the tendency of a non-random alignment; the overlap volume between chromosomes decays fast with increasing loop number. Our results suggest that the formation of chromatin loops imposes both compartmentalization as well as order on the system without requiring additional energy-consuming processes. | Supporting Information | CC BY | no | 2022-01-13 08:14:21 | PLoS One. 2011 Jan 4; 6(1):e14428 | oa_package/6f/21/PMC3014947.tar.gz |
||
PMC3014948 | 21245915 | Introduction
New world cutaneous leishmaniasis resulting from infection with Leishmania mexicana is under different genetic and immunoregulatory controls to those controlling L. major infection [1] . Also, unlike L. major , the majority of mouse strains are susceptible to L. mexicana infection [2] – [3] . As with the other Leishmania species, protective immunity against L. mexicana is the result of a STAT-4 dependent type-1 immune response, although this can be generated independently of IL-12 [4] . While the immunological pathways resulting in non-healing L. major infections in susceptible BALB/c mice remain somewhat controversial, IL-4 plays the major role in promoting non-healing L. mexicana infections in this mouse strain [5] – [7] . Thus, mice lacking IL-4 develop small lesions that heal while those lacking IL-4Rα fail to develop lesions [6] . This also indicates some input from IL-13 in the non-healing response to L. mexicana infection as IL-4 and IL-13 receptors share the IL-4Rα sub unit [6] . However, IL-4 and IL-13 are pleiotropic cytokines and numerous cell types of both the innate and adaptive immune responses produce these cytokines as well as express their receptors.
In order to better differentiate both the cellular sources and targets of IL-4/IL-13 initiating lesion growth and facilitating progressive non-healing disease, we have previously examined parasite growth in SCID mice reconstituted with IL-4 −/− , IL-4Rα −/− , or wild type splenocytes [5] – [6] . These studies indicated that non-lymphocyte sources of IL-4/IL-13 may contribute to early lesion growth during L. mexicana infection. However, the non-healing disease phenotype was dependent on a lymphocyte source of IL-4 and, in its absence, IL-4-deficient splenocyte-reconstituted SCID mice generated a healing response [5] . In addition, SCID mice reconstituted with IL-4Rα −/− splenocytes demonstrated that initial lesion development was also dependent upon cells from this source responding to IL-4/IL-13 [6] .
In order to better differentiate the specific role of IL-4/IL-13 responding cells from global effects in vivo , tissue specific IL-4Rα −/− mice have been produced. So far macrophage/neutrophil specific (LysM cre IL-4Rα −/lox ) [8] and CD4 + T cell specific (Lck cre IL-4Rα −/lox ) [9] IL-4Rα −/− mice have been generated and the consequences for L. major infection studied. In contrast to susceptible BALB/c mice, BALB/c LysM cre IL-4Rα −/lox mice showed a significantly delayed disease progression after infection with L. major , concomitant with normal Th2 and type 2 antibody immune responses but with improved macrophage leishmanicidal activities [8] . These results suggest that alternatively activated macrophages were contributing to the susceptible phenotype in non-healer BALB/c mice. Furthermore T cell-specific Lck cre IL-4Rα −/lox BALB/c mice infected with L. major were significantly more resistant than global IL-4Rα −/− mice and developed a disease phenotype and clinical immunity similar to genetically resistant C57BL/6 mice [9] ; not only showing the importance of IL-4Rα signaling via CD4 + T cells in the non-healing BALB/c phenotype but paradoxically indicating a protective role for IL-4Rα signaling in a non-CD4 + T cell population.
In the present study we demonstrate that in contrast to L. major infection [8] , macrophage/neutrophil signaling via IL-4Rα has minimal effect on the outcome of L. mexicana infection in BALB/c mice. In addition, unlike global IL-4Rα −/− mice infected with L. mexicana that display no lesion growth, infected CD4 + T cell specific (Lck cre IL-4Rα −/lox ) IL-4Rα −/− mice initially develop lesions indicating that early susceptibility to L. mexicana is dependent on an IL-4 responsive non-CD4 + T cell population. However, subsequent lesion growth is significantly curtailed in infected CD4 + T cell specific (Lck cre IL-4Rα −/lox ) IL-4Rα −/− mice compared with IL-4Rα intact mice, and a strong Th1 response generated in the presence of significant elements of Th2 activity. Despite reduced susceptibility in all CD4 + T cell specific (Lck cre IL-4Rα −/lox ) IL-4Rα −/− mice, a dichotomy between the sexes was identified during L. mexicana infection and while lesions in female CD4 + T cell specific (Lck cre IL-4Rα −/lox ) IL-4Rα −/− mice healed they persisted in male mice associated with elevated IL-4 production in this sex compared with females. Together, our results suggest that initial development of the L. mexicana lesion is dependent on an IL-4/13-responsive non-T cell population, whilst progressive infection is dependent on CD4 + T cells responsive to IL-4. | Methods
Mouse model
LysM cre IL-4Rα −/− , Lck cre IL-4Rα −/− , IL-4Rα lox/lox mice were generated and maintained as previously described [9] – [10] . Cell-specific gene disruption in macrophages/neutrophils or T cells was achieved through an intercross between either LysM cre IL-4Rα −/− or Lck cre IL-4Rα −/− and IL-4Rα lox/lox mice. Transgene-bearing LysM Cre IL-4Rα −/lox and Lck cre IL-4Rα −/lox , were identified by PCR genotyping as described [9] – [10] . The mice were maintained under specific pathogen free conditions. Animal experiments were performed in strict accordance with the UK Home Office Animal [Scientific Procedures] Act 1986 (licence number 60/3929) with approval by the University of Strathclyde Ethical Review Panel.
L. mexicana parasites and infection
L. mexicana (MYNC/BZ/62/M379) was maintained by serial passage of amastigotes inoculated into the shaven rumps of BALB/c mice. Amastigotes for use in infections were isolated from lesions and enumerated as described below. Two sites of infection were examined and either 5×10 6 L. mexicana amastigotes in a final volume of 50μl were inoculated subcutaneously into the shaven base of the tail, or 2×10 5 L. mexicana amastigotes in a final volume of 25μl were inoculated subcutaneously into the hind footpad. 6–8 week old male or female LysM Cre IL-4Rα −/lox and Lck cre IL-4Rα −/lox mice were used in each experiment, with age and sex matched cre negative IL-4Rα −/lox littermates used as controls. The lesion diameter was measured using a sliding gauge micrometer at weekly intervals.
Enumeration of parasites
Lesions were excised from L. mexicana infected LysM Cre IL-4Rα −/lox , Lck cre IL-4Rα −/lox and IL-4Rα −/lox mice and disrupted through a metal mesh with 5mL of RPMI 1640 (Cambrex Bio Science Verviers, Belgium). The parasites were washed twice at 350g in RPMI and then enumerated using an improved Naubauer haemocytometer. Alternatively parasite numbers were quantified by limiting dilution, as previously described [11] .
Splenocyte stimulation and cytokine detection
Splenocytes were isolated from infected mice and cultured for 72 hours in 96-well plates (Corning-Costar, NY, USA) in the presence or absence of L. mexicana antigenic lysate, as previously described [6] . IFN-γ and IL-4 levels were detected in the supernatants by capture ELISA. Briefly the wells of Immulon 1B flat-bottomed microtitre plates (ThermoLabsystems, MA, USA) were coated with 50μL of 1μg ml −1 purified anti-mouse IFN-γ capture antibody R4-6A2 (BD Biosciences, Oxford, UK) or 500ng ml −1 IL-4 capture antibody 11B11 (BD Biosciences) in PBS (pH 9.0) overnight at 4°C. Supernatants were then added to the individual wells and either 30μL recombinant mouse IFN-γ (R&D Systems, Abingdon, UK) or recombinant mouse IL-4 (Genzyme, Cambridge, UK) added to individual wells in duplicate in a doubling dilution with a solution of pH 7.4 PBS supplemented with 10% v/v FCS (Harlan Sera-Lab Ltd., Crawley, UK), ranging form 20ng mL −1 to 0.01ng mL −1 (IFN-γ) or 2ng mL −1 to 0.977pg mL −1 (IL-4). The plates were then incubated for 2 hours at 37°C. The bound cytokines were incubated with either biotinylated rat anti-mouse IFN-γ monoclonal antibody XMG1.2 or biotinylated rat anti-mouse IL-4 antibody BVD6-24G2 (both BD Biosciences) and detected with either conjugated streptavidin-alkaline phosphatase or conjugated streptavidin-horseradish peroxidase (BD Biosciences). The appropriate substrate was then added to the wells, p-nitrophenyl-phosphate (Sigma-Aldrich, Poole, UK) or tetramethylbenzidine in pH 5.5 sodium acetate buffer, containing 0.0003% hydrogen peroxide (BDH, Poole, UK). Finally the plates were read at an absorbance of 405nm for IFN-γ or at 450nm for IL-4.
Detection of Leishmania mexicana specific -IgG1, -IgG2a and total IgE
L. mexicana specific-IgG1 and -IgG2a were detected in the plasma of infected mice by ELISA, as previously described [12] . Briefly, Immulon 1B flat-bottomed microtitre plates were coated with 100μL of 10μg ml −1 Leishmania mexicana lysate (lysate preparation previously described [13] in PBS (pH 9.0) overnight at 4°C. Plasma samples were serially diluted in duplicate, followed by a 1 hour incubation at 37°C. Bound Leishmania specific antibodies were detected with a 1 hour incubation with horseradish peroxidase conjugated goat anti-mouse IgG1 or goat anti-mouse IgG2a (Southern Biotechnology Associates Inc., AL, USA). The substrate tetramethylbenzidine in pH 5.5 sodium acetate buffer, containing 0.0003% hydrogen peroxide, was then added to the wells and, following colour development, the reaction stopped by the addition of 10% sulphuric acid, absorbance measured at 450nm using a SOFTmax PRO (Molecular Devices, CA, USA) and the endpoint dilution was determined. Total IgE was detected in the plasma of infected mice by capture ELISA as previously described [12] , using R35–72 capture IgE mAb (BD Biosciences) and biotinylated rat anti-mouse IgE (Southern Biotechnology Associates Inc.).
Flow cytometry
Draining lymph node cells were activated for 4 hours with 50 ng ml −1 PMA and 500 ng ml −1 Ionomycin (both Sigma-Aldrich) along with GolgiPlug (BD Biosciences). Following stimulation, cells were harvested and washed, resuspended in FACS Buffer containing Fc Block (2.4G2 hybridoma supernatant) together with the appropriate combinations of the following antibodies: CD4-APC, CD8-PerCP or B220-FITC (all from BD Biosciences). Intracellular cytokine staining was carried out using PE-conjugated anti-mouse IL-4 or IFN-γ with Cytofix/Cytoperm solution (all from BD Biosciences). Data was obtained using FACSCanto (BD Bioscience) and analysed using FlowJO (Tree Star Inc., CA, USA).
Statistical analysis
Antibody analysis was performed using the Mann-Whitney U test and all other analysis used an unpaired Student's t test. | Results
IL-4Rα signaling via macrophages/neutrophils plays little role in the susceptibility of BALB/c mice to L. mexicana
To compare the progression of L. mexicana lesion growth in LysM Cre IL-4Rα −/lox with IL-4Rα −/lox littermate control and global IL-4Rα −/− mice, animals were infected with 5×10 6 amastigotes into the shaven base of the tail. While no discernible lesions were identified in infected global IL-4Rα −/− mice, as previously demonstrated [6] , rapidly growing non-healing lesions were observed in both LysM Cre IL-4Rα −/lox and IL-4Rα −/lox mice ( Figure 1A ). Parasite burdens were also similar in LysM Cre IL-4Rα −/lox , and IL-4Rα −/lox mice and significantly higher (p<0.0001) than those recorded from global IL-4Rα −/− animals ( Figure 1B ). In line with the non-healing progressive disease phenotype displayed by macrophage/neutrophil IL-4Rα −/− mice, parasite antigen induced spleen cell IFN-γ production was similar to IL-4Rα −/lox mice and significantly less (p<0.01) than that of global IL-4Rα −/− mice ( Figure 1C ). Antigen induced splenocyte IL-4 production was similar in all 3 strains ( Figure 1D ), demonstrating once again that IL-4 induction can be independent of IL-4Rα signaling [6] , [14] – [16] . Further studies demonstrated that the close similarities in the disease phenotypes of LysM Cre IL-4Rα −/lox and IL-4Rα −/lox mice were independent of site, dose of inoculum, and life cycle stage initiating infection (data not shown). These data suggest that the expression of IL-4Rα by a macrophage/neutrophil population is not important in determining susceptibility to L. mexicana infection.
IL-4Rα signaling via CD4 + T cells is essential for progressive non-healing disease following infection with L. mexicana
In subsequent studies, infection of CD4 + T cell specific (Lck cre IL-4Rα −/lox ) IL-4Rα −/− mice with 5×10 6 L. mexicana amastigotes into the shaven base of the tail resulted in control of the lesion growth observed in control IL-4Rα −/lox mice (wild-type equivalent). Interestingly, in one experiment using male mice, lesions in Lck cre IL-4Rα −/lox mice did not heal completely, while experiments utilizing female mice fully resolved ( Figure S1 ). Consequently, to confirm and further investigate this apparent gender-dependent difference in control of L. mexicana infection male and female Lck cre IL-4Rα −/lox , control IL-4Rα −/lox , and global IL-4Rα −/− mice were infected in parallel with 5×10 6 L. mexicana amastigotes into the shaven base of the tail ( Figure 2A–D ). While infected global IL-4Rα −/− mice displayed a non-lesion growth phenotype, lesion growth was progressive in control IL-4Rα −/lox mice ( Figure 2A and B ). These disease phenotypes were independent of gender. By comparison L. mexicana infected female CD4 + T cell specific (Lck cre IL-4Rα −/lox ) IL-4Rα −/− mice developed lesions which completely healed after 4–5 weeks ( Figure 2A ), while male CD4 + T cell specific (Lck cre IL-4Rα −/lox ) IL-4Rα −/− developed lesions which failed to fully heal ( Figure 2B ). Indeed in agreement with lesion size parasite burdens up until week 6 were similar in both male Lck cre IL-4Rα −/lox and IL-4Rα −/lox mice ( Figure S2 ).
Parasite numbers at the termination of the study at week 12 were of a similar order of magnitude in both female and male control IL-4Rα −/lox mice while global IL-4Rα −/− mice of both sexes were equally able to control infection with L. mexicana ( Figure 2C and D ). By contrast while infected CD4 + T cell specific (Lck cre IL-4Rα −/lox ) IL-4Rα −/− mice of both sexes were able to significantly control parasite growth (p<0.0001 and p<0.01 respectively for female and male mice), male mice ( Figure 2D ) were significantly limited in this ability and had significantly higher parasite burdens (p<0.001) than female mice ( Figure 2C ).
IL-4Rα signaling via CD4 + T cells inhibits the production of a specific Th1 response following infection with L. mexicana
L. mexicana infection of CD4 + T cell specific (Lck cre IL-4Rα −/lox ) IL-4Rα −/− mice resulted in an enhanced Th1 response in both male and female Lck cre IL-4Rα −/lox mice compared with control IL-4Rα −/lox mice, as demonstrated by significantly enhanced antigen induced splenocyte IFN-γ production ( Figure 3A and B ; p<0.025 for females; p<0.01 for males). As previously described, antigen stimulated splenocyte IFN-γ production from infected global IL-4Rα −/− mice was significantly higher than for wild-type equivalent (IL-4Rα −/lox ) mice and this was true whether examining female or male mice (p<0.02). Antigen stimulated splenocytes from infected female but not male Lck cre IL-4Rα −/lox mice produced significantly more IFN-γ than antigen stimulated splenocytes from infected global IL-4Rα −/− animals. An expanded Th1 response was also indicated by enhanced antigen specific IgG2a production compared with control IL-4Rα −/lox mice, and in magnitude similar to that generated by global IL-4Rα −/− mice ( Figure 3C and D ). However, while specific IgG2a production in the absence of IL-4Rα signaling via CD4 + T cells in female Lck cre IL-4Rα −/lox mice was significantly greater (p<0.025) than in infected control IL-4Rα −/lox mice as early as week 6 post-infection ( Figure 3C ) it was not until week 12 post-infection that male Lck cre IL-4Rα −/lox mice were producing significantly more IgG2a (p<0.025) than their infected control IL-4Rα −/lox counterparts ( Figure 3D ). Similar results were recorded in 3 separate experiments.
IL-4Rα signaling via CD4 + T cells promotes the induction of a Th2 response following infection with L. mexicana differentially in male and female mice
A clear dichotomy in antigen-induced splenocyte IL-4 production was identified between infected female and male Lck cre IL-4Rα −/lox mice both compared with each other and compared with their wild-type equivalent counterparts ( Figure 4A and B ). Splenocyte IL-4 production was barely detectable in female Lck cre IL-4Rα −/lox mice, and significantly less than IL-4 production by male Lck cre IL-4Rα −/lox mice (p<0.01). On the other hand, antigen induced splenocyte IL-4 production was similar in all infected male mice, independent of IL-4Rα expression ( Figure 4B ). Similar results were obtained with another Th2 cytokine, IL-5 ( Figure 4C and D ), as well as IL-10 ( Figure 4E and F ), with production of both cytokines significantly lower in female but not male Lck cre IL-4Rα −/lox compared with similarly infected sex matched control IL-4Rα −/lox mice. Similarly, Th2 associated antigen specific IgG1 production was significantly less in infected female (p<0.05 week 12) but not infected male Lck cre IL-4Rα −/lox mice compared with their respective control IL-4Rα −/lox counterparts ( Figure S2 ). Minimal IgG1 production was detected in the serum of infected global IL-4Rα −/− mice. Total IgE levels in infected female and male Lck cre IL-4Rα −/lox mice were similar to each other and intermediate between those induced in control IL-4Rα −/lox and the absence of IgE in infected IL-4Rα −/− mice ( Figure S3 ).
Anti-CD3 and ConA stimulation ( Figure 5A and B ) of spleen cells from 12 week infected mice demonstrated quite clearly that not only was IL-4 production from female Lck cre IL-4Rα −/lox mice significantly less than that of female control IL-4Rα −/lox mice (p<0.003 for ConA and p<0.01 for anti-CD3 respectively), but also significantly less than similarly treated male Lck cre IL-4Rα −/lox mice (p<0.02 for ConA and p<0.05 for anti-CD3 respectively). Conversely male Lck cre IL-4Rα −/lox mice splenocytes produced similar quantities of IL-4 to control male IL-4Rα −/lox mice with either treatment. Examination of draining inguinal lymph node cells indicated a significantly lower (p<0.05) percentage of IL-4 and greater (p<0.05) percentage of IFN-γ secreting CD4 + T cells in infected female but not male Lck cre IL-4Rα −/lox mice compared with control sex-matched IL-4Rα −/lox mice ( Figure 6 A–D ). | Discussion
We have previously demonstrated that signaling via IL-4Rα plays the major role in the non-healing response of BALB/c mice following infection with L. mexicana [6] and that IL-4Rα −/− mice, unlike their wild-type counterparts that produce progressively growing non-healing lesions, display a non-lesion growth disease phenotype associated with an enhanced type-1 response. In the course of the present study using macrophage/neutrophil specific IL-4Rα −/− mice (LysM cre IL-4Rα −/lox ) we failed to identify any significant role for IL-4Rα signaling via macrophages/neutrophils in the normal non-healing response of BALB/c mice infected with L. mexicana . In contrast, following early lesion growth, CD4 + T cell specific (Lck cre IL-4Rα −/lox ) IL-4Rα −/− mice were able to inhibit disease progression. However, while lesions in female CD4 + T cell specific (Lck cre IL-4Rα −/lox ) IL-4Rα −/− mice healed those in male CD4 + T cell specific (Lck cre IL-4Rα −/lox ) IL-4Rα −/− mice persisted. Furthermore, although both female and male CD4 + T cell specific (Lck cre IL-4Rα −/lox ) IL-4Rα −/− mice had significantly enhanced type-1 responses compared with IL-4Rα intact (IL-4Rα −/lox ) mice, male CD4 + T cell specific (Lck cre IL-4Rα −/lox ) IL-4Rα −/− mice maintained strong type-2 responses compared with their female counterparts.
Although signaling via IL-4Rα plays a significant role in the outcome of infection with L. mexicana as well as L. major [6] , [8] – [9] , [15] the cell targets for IL-4/IL-13 activity and whether they promote or inhibit the disease process differ significantly between species. Thus, while IL-4Rα signaling via macrophages/neutrophils promotes early lesion growth in L. major infected BALB/c mice and macrophage/neutrophil specific (LysM cre IL-4Rα −/lox ) IL-4Rα −/− mice display delayed lesion growth [8] , we have failed to identify any contributory role for IL-4Rα signaling via macrophages/neutrophils during L. mexicana infection. The control of L. major early in infection in LysM cre IL-4Rα −/lox mice has been identified as being due to enhanced macrophage microbicidal NO mediated activity in the absence of alternative macrophage activation. What may be critical in this regard is that L. amazonensis parasites, which belong to the “ mexicana ” complex of parasites, have been shown to be more resistant to macrophage-mediated control than L. major requiring higher levels of NO to induce killing [16] – [17] . Furthermore, recent evidence indicates that, unlike L. major , there is in fact enhanced replication of the amastigote stage of L. amazonensis in IFN-γ-stimulated murine macrophages [18] , reportedly due to the induction of a novel L-arginine pathway independent of iNOS or host arginase [19] . In addition it has been demonstrated that arginase null-mutant L. mexicana have attenuated virulence in vitro and in vivo with the indication that the parasite arginase has a potential role in depleting host L-arginine available for iNOS activity [20] – [21] . Furthermore, the authors suggest that there could be different roles of arginase between L. mexicana and L. major as the Th2 response is blunted in animals infected with arginase null mutant L. mexicana parasites while pharmacological inhibition of arginase during L. major infection did not inhibit the Th2 immune response [22] .
CD4 + T cell specific (Lck cre IL-4Rα −/lox ) IL-4Rα −/− mice are more resistant than global IL-4Rα −/− mice to infection with L. major , indicating that in the absence of a polarized Th2 response, there is a role for an IL-4/IL-13 responsive non-CD4 + T cell in early resistance to infection [9] . Conversely CD4 + T cell specific (Lck cre IL-4Rα −/lox ) IL-4Rα −/− mice are more susceptible than global IL-4Rα −/− mice to infection with L. mexicana , indicating a role for an IL-4/IL-13 responsive non CD4 + T cell population in early susceptibility. We have now studied the course of L. mexicana infection in newly generated iLck cre IL-4Rα −/lox female and male mice that have IL-4Rα deleted on all T cell populations [23] . These produce the same disease and immunological phenotypes as CD4 + T cell specific (Lck cre IL-4Rα −/lox ) IL-4Rα −/− mice (data not shown). Consequently IL-4 responsive CD8 + T cells do not play a role in early susceptibility or the non-healing response following infection with L. mexicana . Studies utilizing macrophage specific BALB/c IL-4Rα −/− mice have demonstrated that IL-4/IL-13 operates through this population to enhance L. major parasite growth via alternative macrophage activation [8] and consequently these are unlikely to be the population driving a Th1 response in CD4 + T cell specific (Lck cre IL-4Rα −/lox ) IL-4Rα −/− mice. However, IL-4 treatment of BALB/c mice prior to T cell priming has previously been demonstrated to instruct dendritic cells to produce IL-12 and facilitate a protective Th1 response against L. major [24] and is required for protective type-1 responses to Candida [25] . In addition IL-13 is able to prime monocytes for IL-12 production [26] , which is also observed in listeriosis [27] . Furthermore both IL-4 and IL-13 promote CD40L-induced IL-12 production by macrophages and dendritic cells [28] . This would indicate that dendritic cells may be the IL-4/IL-13 responsive cells facilitating protection against L. major in the absence of IL-4Rα responsive CD4 + T cells in BALB/c mice. As no distinct disease phenotype could be discerned in macrophage/neutrophil specific (LysM cre IL-4Rα −/lox ) IL-4Rα −/− mice compared with IL-4Rα intact animals infected with L. mexicana , no IL-4/IL13 responsive non-T cell population can easily be ruled out in promoting early infection against this parasite. While a role for B cells and antibody production in the non-healing response to L. mexicana is well established (as reviewed in [29] – [30] ), the fact that IL-4Rα −/− BALB/c mice are more resistant to this parasite than IL-4 −/− mice suggests that IL-13 responsive cells and consequently non-lymphoid cells via IL-4Rα signaling play a role in disease susceptibility during L. mexicana infection [6] .
Unlike the epidemiological and experimental reports on L. major and L. tropica which identify females as more susceptible, females are more resistant than males to cutaneous infection with L. mexicana (humans and mice) [5] , [12] , [31] – [33] and visceral leishmaniasis caused by L. donovani (humans) [34] – [35] or L. infantum (dogs) [36] . Female DBA/2 mice infected with L. mexicana develop much stronger Th1 responses, as measured by IFN-γ production, delayed–type hypersensitivity and IgG2a antibody levels, than similarly infected male mice [12] , [32] . Similarly, in humans infected with L. mexicana , females generally have increased Th1 responses as measured by DTH reactions and decreased Th2 responses as measured by IgE production than males [33] . The present studies using CD4 + T cell specific IL-4Rα −/− BALB/c mice have revealed a previously undetected, underlying male susceptibility to L. mexicana involving T cells. Thus, unlike female mice, male mice were unable to resolve infection and overall had a less polarized Th1 response and more polarized Th2 than their female counterparts. This was associated with IL-4 production independently of IL-4Rα signaling in male but not female CD4 + T cell specific (Lck cre IL-4Rα −/lox ) IL-4Rα −/− BALB/c mice. The L. mexicana induced IL-4 producing Th2 phenotype in male Lck cre IL-4Rα −/lox BALB/c mice was not the result of differential Cre-mediated deletion efficiency of IL-4Rα in male mice as compared with female mice, as no CD4 + T cell population expressing IL-4Rα was detected from either gender (data not shown). While IL-4 production independently of IL-4Rα signaling has been observed in a number of immunological models previously [6] , [14] – [15] this is the first time a sex associated influence on this ability has been identified. How significant this could be with regard to the numerous gender related differences observed in inflammatory and infectious diseases is an intriguing question but outside the scope of the present study.
To conclude in this study utilising tissue specific IL-4Rα −/− mice we demonstrate that upon infection with L. mexicana initial lesion growth is dependent on a non-CD4 + T cell population responsive to IL-4/IL-13, while progressive infection is dependent on CD4 + T cells responsive to IL-4. Furthermore whether lesions heal or not is gender determined, suggesting a subtle but significant effect of sex hormones on CD4 + T cell function whereby infected male but not female CD4 + T cell specific IL-4Rα −/− mice can drive IL-4 production independently of IL-4Rα signaling. | Conceived and designed the experiments: FB JA. Performed the experiments: KJB ORM TM HAM JA. Analyzed the data: KJB ORM HAM JA. Contributed reagents/materials/analysis tools: FB JA. Wrote the paper: ORM FB JA.
Immunologically intact BALB/c mice infected with Leishmania mexicana develop non-healing progressively growing lesions associated with a biased Th2 response while similarly infected IL-4Rα-deficient mice fail to develop lesions and develop a robust Th1 response. In order to determine the functional target(s) for IL-4/IL-13 inducing non-healing disease, the course of L. mexicana infection was monitored in mice lacking IL-4Rα expression in specific cellular compartments. A deficiency of IL-4Rα expression on macrophages/neutrophils (in LysM cre IL-4Rα −/lox animals) had minimal effect on the outcome of L. mexicana infection compared with control (IL-4Rα −/flox ) mice. In contrast, CD4 + T cell specific (Lck cre IL-4Rα −/lox ) IL-4Rα −/− mice infected with L. mexicana developed small lesions, which subsequently healed in female mice, but persisted in adult male mice. While a strong Th1 response was manifest in both male and female CD4 + T cell specific IL-4Rα −/− mice infected with L. mexicana , induction of IL-4 was manifest in males but not females, independently of CD4 + T cell IL-4 responsiveness. Similar results were obtained using pan-T cell specific (iLck cre IL-4Rα −/lox ) IL-4Rα −/− mice. Collectively these data demonstrate that upon infection with L. mexicana , initial lesion growth in BALB/c mice is dependent on non-T cell population(s) responsive to IL-4/IL-13 while progressive infection is dependent on CD4 + T cells responsive to IL-4.
Author Summary
Leishmania species are parasites, transmitted by sandflies which are of extensive public health importance in the tropical and subtropical regions of the world. A large number of distinct Leishmania species cause cutaneous disease and the vast majority of studies utilize the caustive agent of Old World cutaneous leishmaniasis, L. major . Other species, for example, L. mexicana , are associated with quite different patterns of disease following infection of mice when compared with L. major . Thus, while susceptible BALB/c mice deficient in the ability to respond to the cytokines IL-4/IL-13 are not protected against development of cutaneous leishmaniasis caused by L. major they are totally resistant to infection with L. mexicana . Here we describe the outcome of L. mexicana infection in BALB/c mice with cell-specific deletion of the receptor for IL-4/IL-13 on macrophages/neutrophils or T helper cells. Infections develop in both mutants but lesion growth is controlled only in T cell specific knockouts and female but not male mice heal. Male but not female T cell specific knockouts maintain a strong IL-4/IL-13 response. This highlights the role of IL4/IL-13 in driving a non-healing response and may in part explain why human males are more susceptible to this infection than females. | Supporting Information | CC BY | no | 2022-01-13 05:08:15 | PLoS Negl Trop Dis. 2011 Jan 4; 5(1):e930 | oa_package/35/22/PMC3014948.tar.gz |
||
PMC3014949 | 21245916 | Introduction
The Mekong river delta is located in the south of Vietnam ( Figure 1 ) in an area of 40,000 square kilometres (12% of Vietnam's land mass) and is home to over 20% of Vietnam's population. It is the area where the Mekong river divides into multiple channels and drains into the South China sea. The low-lying nature of the land and the seasonal fluctuation in water level make the region particularly vulnerable to flooding. The human-restricted disease typhoid fever is endemic to the Mekong delta region [1] , [2] , with a mean incidence of ∼80 cases per 100,000 people per year [1] , [2] , [3] , [4] . Salmonella Typhi ( S . Typhi), the bacterium causing typhoid fever, is transmitted human-to-human in areas with poor sanitation.
The first multidrug resistant (MDR; defined as resistance to chloramphenicol, ampicillin and co-trimoxazole) typhoid outbreak in Vietnam occurred in Kien Giang in the Mekong river delta in 1993 [5] , and since then the fluoroquinolones have become the first choice for the treatment of typhoid fever. MDR S . Typhi is usually associated with self-transferrable IncHI1 plasmids carrying multiple resistance genes encoded within mobile genetic elements [6] , [7] , [8] , [9] , [10] . Between 1994 to 1998, over 80% of S. Typhi strains isolated in the Mekong delta region were reported to be MDR [11] , and declined to approximately 50% between 2002 and 2004 [5] , [11] , [12] . This decline may have been catalysed by the change in treatment policy and the widespread use of fluoroquinolones (such as ciprofloxacin and ofloxacin), which are effective against MDR strains [13] , [14] .
While high-level resistance to fluoroquinolones remains uncommon in Vietnam and other endemic typhoid regions, there has been a sharp increase in the proportion of S . Typhi isolates that are resistant to nalidixic acid [11] . Nalidixic acid (Nal) is a quinolone antimicrobial (the precursor of fluoroquinolones) and the main mechanism for Nal resistance in S. Typhi is mutation of the DNA gyrase gene, gyrA [11] , [15] . S. Typhi strains with Nal resistance-conferring mutations in the gyrA gene usually have elevated minimum inhibitory concentrations (MIC) to fluoroquinolone antibiotics such as ciprofloxacin (MIC ≥0.125 μg/ml) [16] . However, these organisms are not resistant according to CLSI guidelines, which are currently defined by MIC ≥4 μg/ml to ciprofloxacin [17] . Even though these strains are not classified as resistant, they are of clinical importance since typhoid patients infected with such strains respond less well to fluoroquinolone therapy [14] , [15] , [18] , [19] . Such patients frequently have a protracted fever and an increased rate of relapse, compared to those infected with strains that do not have an elevated MIC to fluoroquinolones (MIC <0.125 μg/ml to ciprofloxacin and <0.25 μg/ml to ofloxacin) [15] , [18] , [19] . Resistance to Nal is therefore often used as a marker to predict how well a patient will respond to therapy with fluoroquinolones.
The incidence of typhoid fever has declined in Vietnam. Between 1991 and 2001 approximately 17,000 cases of typhoid fever (blood culture confirmed and syndromic cases) were reported annually through the Vietnamese national surveillance system [1] , [2] , while only 4,323 and 5,030 annual typhoid fever cases were reported in 2004 and 2005, respectively (Source: National Institute of Health and Epidemiology, Ministry of Health, Vietnam). However, 75% of these cases occurred in the Mekong delta [1] , [2] , likely associated with high population density and the propensity of the land to become saturated with floodwaters. In this region, the occurrence of S . Typhi isolates that are MDR and Nal resistant severely limits treatment options. More than 95% of S. Typhi isolated in the Mekong delta are now resistant to Nal, placing a considerable pressure on the effective use of fluoroquinolones [11] , [12] . To compare alternative therapies for typhoid fever patients infected with strains that are MDR and Nal resistant, a randomized controlled trial comparing gatifloxacin (a newer 8-methoxy fluoroquinolone) and azithromycin (a macrolide) was conducted during 2004–2005 in the Mekong delta region [20] . Typhoid patients (adults and children) were recruited into the study at three hospitals in the south of Vietnam (details in Materials and Methods , locations are highlighted in Figure 1B ). Here, we used a high-throughput single nucleotide polymorphism (SNP) typing assay to investigate the population structure of S . Typhi collected during the study [20] , and to determine the genetic mechanisms of drug resistance in this S . Typhi population. | Materials and Methods
Ethics statement
The study was conducted according to the principles expressed in the Declaration of Helsinki and approved by the Institutional Review Board of the Hospital for Tropical Diseases and the Oxford Tropical Research Ethics Committee (OXTREC). All patients provided written informed consent for the collection of samples and subsequent analysis (written informed consent was provided by the parents or guardian of children under 18 years of age).
Patient recruitment
S . Typhi isolates were collected during a multicenter clinical trial [20] conducted between January 2004 and December 2005 at (a) the Hospital for Tropical Diseases in Ho Chi Minh City (n = 10), (b) Dong Thap Provincial Hospital, Cao Lanh, Dong Thap province (n = 25) and (c) An Giang Provincial Hospital, Long Xuyen, An Giang province (n = 232). Locations of (b) and (c) are shown in Figure 1B . Adults and children over 6 months of age were eligible to be included in the study if they had clinically suspected or culture-confirmed uncomplicated typhoid fever and if fully informed written consent had been obtained. Patients were tested for typhoid carriage (via stool culture) during follow-up appointments at 1, 3 and 6 months after discharge from hospital. The 267 isolates presented in this study constitute nearly the full complement of 287 S. Typhi isolated from culture-confirmed typhoid patients enrolled in the trial; the recruitment flow for which is described in detail in [20] .
Antimicrobial susceptibility testing
Antimicrobial susceptibility testing was performed at the time of initial isolation by disc diffusion according to Clinical Laboratory Standards Institute (CLSI) guidelines [17] . Antimicrobial agents tested were: ampicillin, chloramphenicol, trimethoprim-sulfamethoxazole (co-trimoxazole), nalidixic acid, ofloxacin, ciprofloxacin and ceftriaxone (Oxoid, Basingstoke, UK). Minimum Inhibitory Concentrations (MICs) for amoxicillin, chloramphenicol, nalidixic acid, ofloxacin, ciprofloxacin, gatifloxacin, ceftriaxone and azithromycin were determined by E-test (AB Biodisk, Solna, Sweden). Multidrug resistance (MDR) of isolates was defined as resistance to chloramphenicol (MIC ≥32 μg/mL), ampicillin (MIC ≥32 μg/mL) and trimethoprim-sulfamethoxazole (MIC ≥8/152 μg/mL). Nalidixic acid resistance was defined by an MIC ≥32 μg/mL.
Bacterial isolation and DNA preparation
After initial isolation, S . Typhi was stored at −70°C in a 20% glycerol solution until required for further analysis and DNA extraction. To revive frozen organisms, MacConkey and Xylose Lysine Decarboxylase (XLD) agar plates were inoculated from the glycerol solution and incubated at 37°C overnight. To ensure correct identification, colonies were checked using slide agglutination with serotype specific antisera (Vi, O9) and an irrelevant antisera as a negative control (O4) (Murex, Dartford, United Kingdom). Two mL of nutrient broth were inoculated with single S . Typhi colonies and incubated overnight. Overnight cultures were centrifuged and S . Typhi DNA was extracted using Wizard Genomic DNA Purification kit (Promega, USA) as recommended by the manufacturer's guidelines. DNA was stored at −20°C. DNA was quantified using the Quant-IT PicoGreen dsDNA Reagent and Kit (Invitrogen, UK). S . Typhi DNA concentrations were adjusted to 50 ng/mL and 250 ng of DNA were pipetted into 96-well plates. Each 96-well plate contained two isolates in duplicate and the sequenced S . Typhi isolate CT18 as control for assay reproducibility.
Determination of chromosomal and plasmid haplotypes
The chromosomal haplotype of S . Typhi isolates was determined based on alleles present at 1,485 chromosomal SNP loci identified previously from genome-wide surveys [12] , [21] and listed in [22] , [23] . IncHI1 plasmid haplotypes were determined based on eight SNPs identified previously [22] , [24] and resistance gene sequences were interrogated using additional oligonucleotide probes (listed in Table S1 ). All loci were interrogated using a GoldenGate custom oligonucleotide array according to the manufacturer's standard protocols (Illumina), as described previously [22] , [23] . A maximum-likelihood phylogenetic tree based on chromosomal SNPs was constructed using the RAxML software [25] .
Statistical analysis
Clinical data were entered into an electronic database (Epi Info 2003, CDC, Atlanta, USA). For comparison of patient characteristics according to infecting S . Typhi haplotypes, Kruskal-Wallis tests were used for analysis of continuous variables (age, length of stay in hospital, fever clearance time) and logistic regression was used for categorical variables (presence of symptoms). Odds ratios were adjusted for duration of fever prior to admission and use of antibiotics prior to admission by including these variables in the logistic regression model. Where data was missing for a particular patient and variable, that patient was excluded from analysis of that variable (N≤35 patients). Two-tailed p-values are reported; statistical analysis was performed using the R package ( http://www.r-project.org/ ).
PCR amplification and sequencing of gyrA gene in S . Typhi
Oligonucleotide primers for the amplification of the quinolone resistance determining regions in the S . Typhi gyrA gene were as follows [11] : GYRA/P1 5′-TGTCCGAG ATGGCCTGAAGC-3′ and GYRA/P2 5′-TACCGTCATAAGTTATCCACG-3′ . Predicted PCR amplicon size was 347 bp. PCR was performed under the following conditions; 30 cycles of: 92°C for 45 seconds, 45–62°C for 45 seconds and extension at 74°C for 1 minute, followed by a final extension step at 74°C for 2 minutes. PCR products were purified and directly sequenced using the CEQ DTCS - Quick Start Kit (Beckman Coulter, USA) and the CEQ 8000 capillary sequencer. The resulting DNA sequence was analyzed using CEQuence Investigator CEQ2000XL (Beckman Coulter, USA). All sequences were verified, aligned and manipulated using Bioedit software ( http://www.mbio.ncsu.edu/BioEdit/bioedit.html ) and compared to other gyrA sequences by BLASTn at NCBI.
Spatial data collection and analysis
Patient addresses were recorded at the time of hospital admission. The latitude and longitude of the residences of typhoid fever patients (to the hamlet/village level) was assigned from the collected address data using 1/50,000 scale maps (Source: Cartographic Publishing House and VinaREN, Ministry of Natural Resources and Environment, Vietnam) and cross-checked using the websites http://www.basao.com.vn and http://ciren.vn . Location data was analysed using Quantum GIS version 1.4.0 ( http://www.qgis.org/ ). Locations were colour-coded according to S. Typhi haplotype and clustering of specific haplotypes was calculated using the nearest-neighbour analysis function. Nearest-neighbour analysis examines the distances between each point and the closest point to it, and then compares these to expected values for a random sample of points from a CSR (complete spatial randomness) pattern. Significant clustering was inferred by Z-score value (standard normal variable) of less than 0; a positive score was interpreted as dispersion of locations. | Results
S . Typhi population structure
A recently developed typing system, based on the simultaneous interrogation of 1,485 S . Typhi chromosomal single nucleotide polymorphisms (SNPs) using a custom Illumina GoldenGate array [22] , [23] , was used to analyse each of the S . Typhi isolates. This approach facilitates the unequivocal assignment of isolates to haplotypes, allowing closely related strains to be distinguished phylogenetically based on single nucleotide changes. From 287 patients with culture confirmed typhoid fever recruited between January 2004 and December 2005 [20] , 267 S . Typhi were available for SNP typing. These included 264 S . Typhi isolated from blood culture at admission [20] one relapse isolate and two faecal carriage isolates. A total of 24 S . Typhi (23 isolated from An Giang and one from Dong Thap, randomly distributed throughout the study period) were not available for SNP typing.
A total of 261 S . Typhi isolates (98%) were of the common H58 haplogroup. The remaining isolates were of haplotypes H1 (isolates BJ105, BJ63, BJ64), H45 (isolate BJ264), H50 (isolate BJ9) and H52 (isolate BJ3; see Figure 2 and Table 1 ). The H58 S . Typhi isolates displayed variation at 10 SNP loci (detailed in [23] ), which differentiated seven distinct sub-H58 haplotypes, shown in Figure 2 . However, 242 (93%) of these isolates belonged to just three closely related H58 haplotypes, designated C, E1 and E2 in Figure 2 (numbers given in Table 1 ). The genome of S . Typhi strain AG3, isolated during the study (March 2004) from a typhoid fever patient living in An Giang province, was sequenced previously [21] . AG3 belongs to the H58-E2 haplotype, and the SNPs separating E2 from haplotypes E1 and C were originally identified by analysis of the AG3 genome. Therefore, the ability to differentiate within the cluster of 242 S . Typhi isolates was dependant on the inclusion of strain AG3 in the initial genome sequencing study used to identify SNP loci [21] .
All but one S . Typhi isolated from the blood culture of patients admitted to An Giang Provincial hospital (231/232), as well as the two faecal S . Typhi strains isolated from chronic carriers in An Giang, belonged to the S . Typhi H58 haplogroup. The remaining S . Typhi isolate BJ264 (see Figure 2 ) was of the H45 haplotype and was isolated from a typhoid fever patient who was resident in neighbouring Can Tho province. One patient at An Giang Provincial hospital relapsed with symptoms of typhoid fever and had S . Typhi isolated from blood culture 11 days after the initial treatment (gatifloxacin) had been completed. The mother of the patient was found to be a chronic S . Typhi carrier. All three S . Typhi strains - the patient's admission and relapse blood culture isolates and the mother's faecal isolate - belonged to the S . Typhi H58-E2 subtype. The patient's isolates were both MDR and carried the IncHI1 ST6 plasmid (see below), whereas the mother's S . Typhi isolate was plasmid-free and susceptible to all first line antimicrobials at the time of isolation. All three isolates were Nal resistant but sensitive to gatifloxacin (MIC 0.19 mg/ml). Stool cultures were taken at 1 month (96% of patients), 3 months (93%) and 6 months (44% of follow-up. Chronic faecal carriage of S . Typhi was detected in only one trial patient. This was a MDR H58-C strain isolated from stool 6 months after treatment (with gatifloxacin), which was indistinguishable at all assayed loci from the patient's original blood culture isolate. Both isolates were Nal resistant but sensititve to gatifloxacin (MIC 0.19 mg/ml).
At Dong Thap Provincial Hospital, only 3 of the 25 S . Typhi isolates did not belong to the H58 haplogroup. Two H1 isolates (BJ63 and BJ64; Figure 2 ) were identical at all assayed loci and were isolated on consecutive days from two patients resident in Dong Thap. A third H1 strain (BJ105; Figure 2 ) differed from BJ63 and BJ64 at 16 chromosomal SNP loci and was isolated in Dong Thap 14 months after these isolates. Two siblings from Dong Thap province were admitted on consecutive days in 2004 and were both infected with MDR S . Typhi of the haplotype H58-C.
Of the ten S . Typhi strains isolated at the Hospital for Tropical Diseases in Ho Chi Minh City, eight were members of the H58 haplogroup, with patients resident in Ho Chi Minh City (n = 4), Long An (n = 1), Kien Giang (n = 2) and An Giang (n = 1) provinces, reflecting the larger catchment area of the hospital. The remaining two S . Typhi were of haplotypes H52 (BJ3) and H50 (BJ9) and were isolated from patients living in Binh Hoa province and Ho Chi Minh City, respectively.
There was no simple association between S . Typhi haplotype and patient age, length of stay in hospital, fever clearance time, vomiting, abdominal pain, hepatomegaly or relapse ( Table 2 ). However, upon admission, patients infected with S . Typhi haplotype H58-E2 tended to report lower frequencies of diarrhoea and headache and higher frequencies of constipation compared to patients infected with other haplotypes, including H58-C (see Table 2 ).
Plasmids and antimicrobial resistance
The GoldenGate assay incorporated probes targeting IncHI1 plasmid sequences, allowing for detection of the presence of IncHI1 plasmid within the genomic DNA extracted from each S . Typhi isolate. The assay indicated that a total of 139 S . Typhi isolates harboured an IncHI1 plasmid. All plasmids were of the IncHI1 ST6 sequence type [24] and all plasmid-bearing isolates belonged to the S . Typhi H58 haplogroup (see Table 1 ). The MDR IncHI1 plasmid was more common among H58-C isolates than H58-E2 isolates (86% vs 19%, see Table 1 ). Of the 139 S . Typhi isolates giving positive signals for IncHI1 SNP loci, 137 (99%) were classified as MDR by antimicrobial susceptibility testing conducted at the time of isolation. One other IncHI1-positive isolate tested positive by GoldenGate assay for the genes sul1 , sul2 , dfrA7 , tetACDR , strAB , bla and cat (resistance genes; functions outlined in Table S1 ) like the MDR isolates, yet had low MICs for chloramphenicol, ampicillin and trimethoprim-sulfamethoxazole. An additional S . Typhi isolate, BJ5, was resistant to ampicillin and trimethoprim-sulfamethoxazole but sensitive to chloramphenicol. This was consistent with GoldenGate assay results, which gave positive signals for the repC replication initiation gene of IncHI1, resistance genes strAB , bla , sul1 , sul2 , dfrA7 , but no signal for sequences from the cat gene encoding chloramphenicol resistance. A further 17 S . Typhi isolates were recorded as MDR according to their antimicrobial susceptibility pattern at the time of isolation, but did not test positive for IncHI1 plasmid loci. This likely reflects loss of the IncHI1 plasmid in culture or storage between the time of isolation and DNA extraction. The MDR status of the infecting S . Typhi isolate was not associated with fever clearance time (p = 0.3, two-sided T-test) or treatment failure (p = 0.18, Chi 2 test).
A total of 257 S . Typhi isolates were resistant to nalidixic acid (Nal). All of these isolates belonged to the H58 haplogroup ( Table 1 ) and all were susceptible to gatifloxacin, ciprofloxacin and ofloxacin according to current CLSI guidelines [17] . S . Typhi haplotypes H58-C, H58-E1 and H58-E2 were uniformly resistant to Nal, with the exception of a single H58-C isolate which had an intermediate MIC of 28 μg/mL (resistance defined as MIC ≥32 μg/mL). The sequenced H58-E2 isolate AG3 harbours a mutation changing serine (TCC) to phenylalanine (TTC) at codon 83 in the gyrA gene (GyrA-Ser83Phe) [21] , which is known to confer resistance to Nal [26] . In the present study we sequenced the gyrA gene in 223 of the Nal resistant isolates (87%) and found the same GyrA-Ser83Phe amino acid substitution in all isolates tested.
Spatial and temporal distribution
Figure 3 shows the spatial distribution of the residences of 160 typhoid patients (this information was not available for the remaining patients). Of the patients admitted at An Giang Provincial Hospital and Dong Thap Provincial Hospital, sufficient address detail to allow for assignment of latitude and longitude was provided in 61% and 73% of cases, respectively. This represents 50% and 20% of all blood culture confirmed typhoid fever patients at An Giang Provincial Hospital and Dong Thap Provincial Hospital, respectively, during 2004–2005. In An Giang, patients' homes clustered around the An Giang Provincial Hospital, but also around the Sông H u branch of the Mekong river (see Figure 3 ). Most S . Typhi isolated from patients living near this point in An Giang province were of the H58-E2 haplotype (orange in Figure 3 ), and this group demonstrated significant clustering using nearest-neighbour analysis (n = 57, Z-score = −14.145). In contrast, S . Typhi of the H58-C haplotype were isolated relatively frequently in neighbouring provinces and had a more sporadic clustering pattern (red in Figure 3 ). While isolates from An Giang Provincial Hospital are overrepresented in this spatial analysis, the apparent increase in typhoid density in An Giang is consistent with total Typhi isolation rates at the two hospitals during the study period (284 at An Giang Provincial Hospital and 90 at Dong Thap Provincial Hospital).
The temporal distribution of S . Typhi haplotypes over 2004 and 2005 is shown in Figure 4 . Typhoid fever cases peaked just prior to the onset of the wet season in each year, as has been observed previously in this region [1] , [3] (see monthly rainfall, solid line in Figure 4 ). In 2004, H58-E2 and H58-C were both prevalent (62 C, red in Figure 4 ; 103 E2, orange in Figure 4 ), whereas few isolates of H58-E2 Typhi were observed during 2005 (55 C, 4 E2; see Figure 4 ). The decline of H58-E2 may be associated with selection for the IncHI1 MDR plasmid, which was much more common in H58-C ( Table 1 ). As Figure 4 highlights, the majority of isolates collected during the second season were MDR and carried the IncHI1 plasmid ST6. | Discussion
Our data show the vast majority of S . Typhi isolates (n = 261, 98%) isolated from the Mekong delta during the two-year study period belonged to the H58 haplogroup. Furthermore, 91% of isolates (n = 242) belonged to a single clonal complex of S . Typhi H58 (nodes C, E1, E2 shown in Figure 2 ), demonstrating remarkable homogeneity in the S. Typhi population in this location during the study period. The observed level of clonal dominance is greater than that observed in previous haplotyping studies of local S . Typhi populations. Among 54 S. Typhi isolates from Jakarta, Indonesia between 1975 and 2005, a total of nine haplotypes were detected, with the dominant H59 haplotype accounting for 53% of isolates; the next most frequent haplotype was genetically distant from H59 and comprised 24% of isolates [27] . In Kathmandu, Nepal, a collection of S . Typhi isolated from children hospitalised with typhoid fever in 2005–2006 was dominated by the H58-G haplotype (66%) but the distant H42 haplotype was also present at high frequency (19%) [23] . Among S . Typhi isolated between 2001 and 2008 in Nairobi, Kenya, 87% were H58, although two distinct subtypes (nodes B and J, see Figure 2 ) were co-circulating at equally high frequencies (>40% each) between 2004 and 2008 [22] . S . Typhi H58 nodes B and J (co-circulating in Nairobi) represent distinct lineages, each acquiring unique SNPs since the last common ancestor of H58 ( Figure 2 ). However H58 nodes C, E1 and E2, which account for 91% of isolates in this study in Vietnam, are closely related and formed a tight clonal complex differentiated by just two SNPs ( Figure 2 ). Thus the overall level of clonality of the S. Typhi population analysed in this study was unexpectedly high. The clonal complex comprising H58-C, -E1 and -E2 was not detected in study populations in Nepal and Kenya where the same SNP typing method was used [22] , [23] , suggesting it may have arisen locally in Vietnam.
Despite the genetic homogeneity we observed, the availability of whole-genome sequence data for S. Typhi H58-E2 isolate AG3 [21] , collected during the study, allowed us to differentiate closely related organisms within the H58 group. Just two SNPs identified in strain AG3 subdivided the homogeneous group into three nodes C, E1 and E2 ( Figure 2 ), of which two (C and E2) were dominant (>40% each). Isolates belonging to the H58-C node were present at a constant rate during the two years of the study (62 isolates in 2004 and 55 in 2005). However, isolates belonging to the H58-E2 node were common during 2004 (103 isolates), yet were virtually undetected in 2005 (3 isolates). This change in both haplotype distribution and total number of typhoid cases from 2004 to 2005 is striking, and suggests an outbreak caused by S . Typhi H58-E2 during 2004. We additionally found that H58-C strains had a much stronger association with the ST6 IncHI1 MDR plasmid than H58-E2 ( Table 1 ). We speculate that the persistence of H58-C strains and the corresponding disappearance of H58-E2 may be associated with a competitive phenotypic advantage conferred by the IncHI1 MDR plasmid. However, it is important to remember that node C is a precursor of node E2 and we can only differentiate E2 from C because we had whole genome sequence data for an H58-E2 strain from which to identify SNPs [21] . Thus the population of S . Typhi isolates assigned to node C by our SNP typing assay may be more diverse than that assigned to H58-E2. It is also important to note that since our data covers just two years, it is possible that any competitive advantage of H58-C strains may be short-lived and there is no evidence for long-term replacement of H58-E2.
We identified two cases of chronic faecal carriage of S . Typhi during the course of the study, one in a patient's relative and one in a patient after 6 months of follow-up. This underlines the importance of screening procedures to identify carriers and effective treatment to eliminate carriage and reduce transmission. The faecal S . Typhi isolates were of the dominant H58-E2 and H58-C haplotypes, respectively. In a previous case-control study performed in the Mekong delta, close contact with a patient with typhoid fever was significantly associated with developing the disease compared to hospital controls (adjusted odds ratio (OR) = 5.2, 95% confidence interval (95% CI) [1.7, 15.9]) or community controls (adjusted OR = 11.9, 95% CI [2.3, 60.7]) [28] .
We were able to collect residential location data from 160 typhoid patients (61%). While typhoid patients reporting to An Giang Provincial Hospital are overrepresented in this data set (50% of all confirmed cases vs. 20% of all confirmed cases at Dong Thap Hospital), the apparent clustering in An Giang ( Figure 3 ) is consistent with the overall isolation rates at the two hospitals, which during the study period was more than three times higher at An Giang Provincial Hospital than Dong Thap Hospital. The data set provides roughly equal representation of patients infected with S . Typhi H58-C (65%) and H58-E2 (62%), thus any differences in spatial distribution between patients presenting at the different hospitals should not affect the differences between spatial distribution of these haplotypes. Spatial clustering of S . Typhi H58-E2 was evident particularly around the Sông H u branch of the Mekong river, while other S . Typhi haplotypes were more broadly distributed ( Figure 4 ). The spatial clustering of H58-E2 S. Typhi further supports a localised outbreak in 2004 caused by these isolates. In contrast, the broader spatial and temporal distribution of S . Typhi H58-C during the study suggests it may be well established in the community and can persist over longer distances and time periods.
We also observed that some symptoms reported by patients infected with H58-E2 S . Typhi differed from those infected with other S . Typhi haplotypes ( Table 2 ). After adjusting for antibiotic use and duration of fever prior to admission, patients infected with H58-E2 S . Typhi were more likely to report diarrhoea and headache compared with other S . Typhi haplotypes (OR = 0.56, 95% CI [0.34, 0.93] and OR = 0.66, 95% CI [0.40, 1.09], respectively), but were more commonly associated with constipation (OR = 2.6, 95% CI [1.1, 6.0]). This suggests there may be some phenotypic differences between H58-E2 and other S . Typhi with respect to disease, however these were post-hoc analyses and no adjustments were made for multiple comparisons, hence these associations should be interpreted with caution. However if confirmed in subsequent prospective studies, it would be of interest to know whether these phenotypic characteristic were associated with specific mutations in H58-E2 S . Typhi. The two SNPs differentiating the E2 node from E1 and C are both synonymous mutations (C->T in melA (nt 315); G->A in rbsA (nt 576)) and our earlier analysis of the AG3 sequence data detected no phage insertions and no large deletions that were not also detected in other sequenced H58 isolates [21] . However, we were unable to verify if other single-base insertions or deletions were present, which may result in gene inactivation with corresponding phenotypic effects.
Patterns of antimicrobial resistance of S. Typhi tend to vary markedly between different typhoid-endemic regions. In this present work, as in the recent study of Kenyan isolates [22] , there were high rates of MDR associated with IncHI1 ST6 plasmids among strains of the S . Typhi H58 haplogroup. This suggests that the presence of the plasmid may contribute to the success of the dominant S . Typhi haplotypes, and the results of our study corroborate this hypothesis. The S . Typhi H58-E2 subtype (which was generally not associated with a plasmid) was only transient, while the H58-C subtype (which was more commonly associated with the IncHI1 MDR plasmid) was present in 2004 and 2005 in southern Vietnam. In Kenya, almost all isolates of the dominant H58 haplotypes carried the MDR plasmid, while the plasmid-free H58-G subtype was only detected twice [22] . All the H58 isolates analysed in the present study were resistant to Nal, conferred by an identical mutation in gyrA . This is consistent with previous studies reporting strong associations between the Nal resistance phenotype and the H58 haplogroup of S. Typhi [12] , [22] , [23] . The presence of the same mutation conferring Nal resistance in all isolates of H58-C, -E1 and -E2 suggests this mutation may have arisen in the common ancestor of this clonal complex, perhaps in situ in the Mekong delta region, and its continued presence is likely maintained by selective pressure exerted by the use of fluoroquinolones.
Conclusions
During 2004–2005, typhoid in the Mekong river delta region of Vietnam was almost exclusively caused by a single Nal-resistant clonal complex of S . Typhi. This reflects a higher level of clonality than observed in other localised S . Typhi populations studied to date, which may be indicative of higher transmission rates in this location. The high level of Nal resistance and multidrug resistance, frequently in the same strains, is concerning and continues to pose problems for the successful treatment of typhoid fever. | Conclusions
During 2004–2005, typhoid in the Mekong river delta region of Vietnam was almost exclusively caused by a single Nal-resistant clonal complex of S . Typhi. This reflects a higher level of clonality than observed in other localised S . Typhi populations studied to date, which may be indicative of higher transmission rates in this location. The high level of Nal resistance and multidrug resistance, frequently in the same strains, is concerning and continues to pose problems for the successful treatment of typhoid fever. | Conceived and designed the experiments: KEH CD NVVC TTH JF GD SB. Performed the experiments: TTC PTD TTPL NVMH TVTN JIC BHM NVVC. Analyzed the data: KEH CD BHM SB. Contributed reagents/materials/analysis tools: TTPL GD. Wrote the paper: KEH CD NVVC TTH JF GD SB.
Background
Typhoid fever remains a public health problem in Vietnam, with a significant burden in the Mekong River delta region. Typhoid fever is caused by the bacterial pathogen Salmonella enterica serovar Typhi ( S . Typhi), which is frequently multidrug resistant with reduced susceptibility to fluoroquinolone-based drugs, the first choice for the treatment of typhoid fever. We used a GoldenGate (Illumina) assay to type 1,500 single nucleotide polymorphisms (SNPs) and analyse the genetic variation of S . Typhi isolated from 267 typhoid fever patients in the Mekong delta region participating in a randomized trial conducted between 2004 and 2005.
Principal Findings
The population of S . Typhi circulating during the study was highly clonal, with 91% of isolates belonging to a single clonal complex of the S . Typhi H58 haplogroup. The patterns of disease were consistent with the presence of an endemic haplotype H58-C and a localised outbreak of S . Typhi haplotype H58-E2 in 2004. H58-E2-associated typhoid fever cases exhibited evidence of significant geo-spatial clustering along the Sông H u branch of the Mekong River. Multidrug resistance was common in the established clone H58-C but not in the outbreak clone H58-E2, however all H58 S . Typhi were nalidixic acid resistant and carried a Ser83Phe amino acid substitution in the gyrA gene.
Significance
The H58 haplogroup dominates S . Typhi populations in other endemic areas, but the population described here was more homogeneous than previously examined populations, and the dominant clonal complex (H58-C, -E1, -E2) observed in this study has not been detected outside Vietnam. IncHI1 plasmid-bearing S . Typhi H58-C was endemic during the study period whilst H58-E2, which rarely carried the plasmid, was only transient, suggesting a selective advantage for the plasmid. These data add insight into the outbreak dynamics and local molecular epidemiology of S . Typhi in southern Vietnam.
Author Summary
Typhoid fever remains a serious public health issue in some parts of Vietnam, including the Mekong delta region. Typhoid is caused by the bacterium Salmonella Typhi, which is frequently multidrug resistant and shows reduced susceptibility to fluoroquinolone-based drugs. We assayed single nucleotide variation in the genomes of S . Typhi organisms isolated from 267 patients with typhoid fever in the Mekong delta between 2004 and 2005, and identified genetically distinct S . Typhi strains. We also detected the presence of genes or mutations that confer drug resistance in those strains. We found that the vast majority of typhoid cases were caused by one of two subgroups of H58 S . Typhi, referred to as H58-C and H58-E2. The H58-E2 group appeared to cause an outbreak in 2004, affecting patients living in a small zone near the Mekong River. The other group, H58-C, was present throughout the study period and affected patients living in a broader area of the Mekong River delta. Most of the H58-C strains were resistant to multiple drugs and carried a plasmid encoding multiple resistance genes. However very few H58-E2 strains were multidrug resistant, which may explain why the strain did not persist after the initial outbreak. | Supporting Information | We thank the Directors and all the doctors, nurses and microbiology staff involved in the study at An Giang Provincial Hospital, Long Xuyen, An Giang Province, Dong Thap Provincial Hospital, Cao Lan, Dong Thap Province and at the Hospital for Tropical Diseases, Ho Chi Minh City, Vietnam. GoldenGate assays were performed by the genotyping team at the Wellcome Trust Sanger Institute. | CC BY | no | 2022-01-13 05:08:15 | PLoS Negl Trop Dis. 2011 Jan 4; 5(1):e929 | oa_package/2e/e3/PMC3014949.tar.gz |
PMC3014950 | 21245917 | Introduction
Sequence alignment achieves many purposes and comes in several different varieties [1] : Local versus global (and even “glocal”: [2] ), pairwise versus multiple, and DNA/RNA versus proteins. Rather than listing all applications, we cite just two numbers: According to Google Scholar the two original papers on the BLAST algorithm for local alignment by [3] and on one of its improvements [4] have been cited more than 30,000 times each, and the number of daily file uploads to the NCBI server providing BLAST is [5] . A partial list of alignment tools in the public domain can be found in http://pbil.univ-lyon1.fr/alignment.html .
In global alignment, which we focus on here, two sequences of comparable length are placed one below the other. The algorithm inserts blanks in each of the sequences such that the number of positions at which the two sequences agree is maximized. More precisely, a scoring scheme is used. Each position at which the two sequences agree is rewarded by a positive score, while each disagreement (“mutation”) and each insertion of a blank (“gap”) is punished by a negative one. The best alignment is that with the highest total score. In local alignment, one aligns only subsequences against each other and looks for the highest scores between any pairs of subsequences. Regions that cannot be well-aligned are simply ignored. Existing algorithms use either heuristic scoring schemes or scores derived from explicit probabilistic models [6] .
Similarities between DNA sequences, e.g. for distance-based phylogenetic tree construction, are typically not based on alignment scores. Instead they use explicit evolutionary assumptions (e.g. the Kimura two-parameter model [7] ) or are simply obtained by counting the number of nucleotide substitutions (like the p-distance or the Poisson corrected p-distance [7] ). An important property of a similarity measure, from the point of view of phylogeny, is that distances should grow linearly with evolution time. This results in a measure satisfying the so-called four point condition [8] , which in turn makes the measure useful for neighbor joining , the most popular distance based algorithm for inferring phylogenetic trees [9] . The most important metrics developed from this view point are the closely related paralinear [10] and log-det [11] distances. In this paper we refer to both as “log-det”, for simplicity's sake.
In the above mentioned distances, distinct rates of different substitution types are either taken into account using a model, or are not taken into account at all. This fact stands in stark contrast with mutual information (MI), which takes the amount of information shared between two objects as a measure of their similarity [12] . For instance, more frequent substitutions can be encoded more efficiently, and should thus be a weaker indicator for dissimilarity than rare, and thus “surprising”, substitutions. The crucial point to note is that the frequency of substitutions and indels and their correlations can be counted directly from the alignment, and no model is required. As a consequence, MI is, in principle, a model-free, universal, and objective similarity measure, in stark contrast to all metrics discussed above.
Indeed, there are two variants of information theory: The more traditional Shannon theory, based on a probabilistic interpretation of the sequences, and the less well known Kolmogorov (or algorithmic ) “complexity” theory [12] . In this paper we use Kolmogorov information as our main vehicle, but we also show that Shannon theory gives comparable results.
Roughly, in algorithmic information theory the complexity of a sequence is the minimal amount of information (measured in bits) needed to specify uniquely, on a given computer, with a given operating system. Numerical results depend on the latter, but this dependence will, in general, be weak and is ignored in this paper. For two sequences and , the conditional complexity (or conditional information) is the information needed to specify , if is already known (i.e., either it or its specification was already input before). If and are similar, this information might consist of a short list of changes needed to go from to , and is small. If, on the other hand, and have nothing in common, then knowing is useless and . Finally, the mutual information (MI) is defined as the difference It is the amount of information which is common to and , and is also equal to the amount of information in which is useful for describing , and vice versa . Indeed, it can be shown that, up to correction terms that become negligible for long sequences (see [12] ): (a) ; (b) if and only if and are completely independent; (c) ; and (d) . Moreover, the probability (in the Bayesian sense) that is smaller than , if and are independent (see Theorem 2 of [13] ). Hence, the similarity is significant and not by chance when is large.
The fundamental difference between Shannon theory and Algorithmic Information theory is that Shannon theory makes no attempt to quantify the minimum information required to specify a any particular sequence. Instead Shannon theory assumes that a sequence can be treated as though it was generated as a “typical” case of a probabilistic process. As a result of this assumption, Shannon information has no dependence on hardware. However, the main drawback is that it cannot, strictly speaking, deal with individual sequences and it needs an assumption on the probability distribution. Any numerical result obtained from individual sequences implies the assumption that the specific sequences are ‘typical’ of the underlying probabilistic process. As a result it involves statistical inference, even if the result does not strongly depend on this inference. In particular, the assumption of independence of letters in a sequence (used also below) will lead to over-estimation of Shannon entropy, and thus implies no risk of overfitting.
The fact that alignment and information theory are closely related has been realized repeatedly. However, most work in this direction has focused on aligning images rather than sequences [14] . Conceptually, these two problems are closely related, but technically, they are not. The effects of sequence randomness on the significance of alignments has also been studied in [15] . Finally, attempts to extend the notion of edit distance [1] to more general editing operations have been made. In this case the similarity of two sequences is quantified by the complexity of the edit string, see [16] . Indeed, the aims of [16] are similar to ours, but their approach differs in several key respects and leads to markedly different results. | Methods
Translation Strings
At the heart of our approach is the concept of a translation string . The translation string contains the minimal information necessary to recover the sequence from another sequence . Similarly, contains the information needed to obtain from . Here we focus on DNA sequences, consisting of the letters A,C,G and T, and corresponding to complete mitochondrial genomes. But the approach is more general and can be applied to protein sequences without further effort. We refer to the element of sequence as , and denote the length of as . Any global alignment algorithm, when applied to and , outputs a pair of sequences of equal length . The sequences and are obtained from and by inserting hyphens (“gaps”) such that the total score is maximized. The strings and also have length , and are composed from an alphabet of nine characters. For each , the letter is a function of and only. An example of this process is found in Figure 1 ; the rules to create are as follows:
if , then ;
if is a hyphen (gap), then has to specify explicitly what is in ; hence A,C,G,T ;
if is a hyphen (gap), then has to indicate that something is deleted from , but there is no need to specify what. Hence ;
if is a transition , i.e. a substitution A G or C T, then ;
if is a transversion A C or T G, then ;
if is a transversion A T or G C, then .
is defined such that (and thus also ) is obtained uniquely from . But can be obtained from using . Thus does exactly what it is intended to do: it allows one to recover from . It does not, however, allow one to recover from . Due to the second and third bullet points above, is not the same as . This distinguishes our approach from typical edit string methods.
Algorithmic Information Theory: Mutual Information
An estimate of the conditional complexity is obtained by compressing using any general purpose compression algorithm such as zip, gzip, bzip2, etc. In the results shown here we use lpaq1 [17] (see also this reference for a survey of public domain lossless compression algorithms). Denoting by the compressed version of and by the length of in bits, gives an exact upper bound If there were no correlations between sequence and the translation string, , this would also be the best possible upper bound. However, in general we must expect that such correlations exist, although we find them to be weak (see the second figure in Material S1 ). Thus Eq. (1) is still a good estimate, but the best one is obtained by compressing conditionally on , More precisely, one can show that .
In order to obtain an estimate of MI, we have to subtract from , which is also estimated via compression. Unlike , is a DNA string. Since general purpose compression algorithms are known to be inferior for DNA [18] , [19] we could use an efficient DNA compressor like, e.g., ‘GeneCompress” or “XM” [18] , [19] (as we shall do in Eq. (4) below). To avoid any question of consistency, we shall not do this. Instead, the compression is carried out using a general purpose compression algorithm, to get
This is to be compared to the general definition of algorithmic MI, based entirely on concatenation and compression [12] , [20] without using any alignment. This estimate is obtained by comparing the size of the compressed concatenation to the sum of the sizes of the compressed individual files,
At first sight it might seem paradoxical that can even be positive. Not only does involve a larger alphabet than , but, in general, it is also a longer string. Thus one could expect that would not typically compress to a shorter size than . The reason why this first impression is wrong is clear from Figure 1: If and are similar, then consists mostly of zeroes and compresses readily. In practical alignment schemes, the scores for mismatches are carefully chosen such that more frequent substitutions are punished less than unlikely substitutions. In contrast, coding each mismatch simply by a letter in seems to ignore this issue. However, more frequent mismatches will give letters occurring with higher frequency, and general purpose compression algorithms utilize frequency differences to achieve higher compression.
Conceptually our approach is similar to encoding generalized edit strings in [16] . However, there are several pivotal differences between that work and ours. First, the authors in [16] did not compress their edit strings and as a result the conclusions they were able to draw from a quantitative analysis were much weaker than ours. Second, our approach utilizes an alignment algorithm to achieve an efficient encoding of . In addition to producing a better estimate of , this allows us to make quantitative evaluations of the alignment algorithm itself. An additional difference between our approach and the traditional edit methods used in approximate string matching [21] is that our translation strings do not give both translations and from the same string. This asymmetry is crucial to establish the relations to conditional and mutual information.
For long strings, should be symmetric in its arguments. In general, the estimates satisfy (see the third figure in Material S1 ). Indeed, the translation strings and can differ substantially, resulting in different estimates for and via Eq. (2). This difference is mostly canceled by differences between and . Take, for instance, the case where is much shorter than . Then consists mostly of hyphens and is highly compressible. On the other hand, is similar to , since most letters have to be inserted when translating to . Thus both and are small compared to , but for different reasons. Further details are given in Material S1 .
Shannon Theory
Compared to algorithmic information, Shannon theory is the more widely known version of information theory [12] . The basic concept of Shannon theory is that of a block or word probability . It gives the probability that the ‘word’ of consecutive letters (such as A,C,G or T for DNA) appears at any random position in the string . Here we assume stationarity, but we do not assume absence of correlations. The entropy (analogous to the complexity in algorithmic information theory) of a string comprised of letters from an alphabet is defined as with From this, MI is defined as in algorithmic theory: [12] . If entropy is measured in bits, then the logarithm is to base 2. In practice, the limit is rarely feasible, and one usually approximates by the single-letter entropy or, at most, by the pair approximation based on the probabilities for words of length two.
Eq. (6) is valid under the assumption that correlations between consecutive letters in the string can be neglected. Similarly, for two sequences of equal length is estimated by assuming that consecutive letter pairs with and are independent. If we make this assumption, there are still two ways to estimate the MI of two strings. In the first we use the fact that carries the same information as to employ a five-letter alphabet . This has the drawback that indels are usually correlated. In the second we thus neglect all indels and reduce the alphabet to . In the following we shall mostly use the latter to compare with other pairwise distance metrics, but we stress that we do this only for simplicity and convenience (and since it is sufficient to make our point). However, the more interesting MI estimate remains the one obtained from algorithmic theory, due to the fact it takes into account both indels and all possible correlations within each string and between them.
Distances, Trees and Quartets
The value of the MI itself is useful for many purposes: Estimating similarities between different pairs (and thus of finding closest neighbors of a given sequence in a large data set); comparing the qualities of alignments obtained by different algorithms; or assessing the significance of an alignment (i.e., verifying that it is better than an alignment between two unrelated sequences). But in the case of phylogeny, one wants more. Ideally, one wants an additive metric distance , i.e. a non-negative symmetric pairwise function for which and which satisfies both the triangle inequality for any triple, and the four-point condition [22] for any quartet. The latter is a necessary and sufficient condition for all pairwise distances between sequences to be representable as distance sums over links in a tree [8] with the sequences represented by the leaves. Thus distances satisfying Eq. (8) are also called ‘tree metrics’.
Several potential metrics can be derived from MI [20] , [23] , [24] . According to [20] , [24] , the preferred one is the normalized compression distance where can be either or , depending on the way it is estimated. For Shannon theory we can use the same construct with replaced by [25] . Since it would be confusing to use the word “compression” for this metric, we have to use another name. We call it the normalized Shannon distance
Although has been used to produce meaningful phylogenetic trees [20] , [23] – [25] , it has one important drawback for phylogenetic applications: It is not additive. Indeed, for two completely unrelated sequences (corresponding to infinite evolutionary distance), both and do not go to infinity, but rather to 1. They are not linear but convex functions of evolutionary distance. Such metrics are well known to lead to long branch attraction (or the ‘Felsenstein phenomenon’ [26] ).
If evolution is assumed to be a Markov process, then the data processing inequality [12] guarantees that MI decreases with evolutionary distance. A natural assumption – following from the dominance of a single maximal eigenvalue of the Markov matrix – is that it decreases exponentially to zero. In this case the log-MI “distance” would increase linearly with evolution and would be thus additive. Unfortunately, is not a proper metric, as it does not even satisfy the triangle inequality. This can be seen from the following example: Take three sequences over an alphabet of four letters (like DNA) where each letter is represented by two bits (purine/pyrimidine, double/triple hydrogen bonds). Sequence is random, sequence is obtained from by replacing randomly the first bit but conserving the second, and is obtained by replacing the second but conserving the first. Then and are non-zero, while . At the same time, all single sequence complexities (and entropies) are the same, thus while and are finite, clearly violating Eq. (7).
Fortunately, real evolution is most likely not as extreme as this counter example, and the triangle inequality is not really required for distance based phylogeny. In particular, the relationship between trees and metric additivity is not restricted to metrics satisfying the triangle inequality, as seen from the proof in [8] . Also, the neighbor joining algorithm [9] does not require the triangle inequality. Thus we claim that is an a priori better distance measure for phylogeny than or , although a final evaluation can only be made through detailed tests on real biological sequences.
Such tests are presented in the results section, with the log-det (or, more precisely, the paralinear) distance [10] , [11] and two distances based on Kimura's model [7] (see the supplementary information) as other competitors. In the latter, one assumes different rates for transitions (A G, C T) and for transversions (all others).
Assume that for two aligned sequences, and , one first eliminates all positions with indels. Thus, at each site one sees one of the 16 possibilities with . Denote the measured frequencies for these possibilities . The single-sequence (‘marginal’) frequencies are and . We introduce matrices with matrix elements , with , and (here, is the Kronecker delta, i.e. and are diagonal matrices). The log-det distance is then defined as In [10] , this is called paralinear distance; in [11] the name log-det is used either for this or for simplified versions where the matrices and are omitted. This difference is irrelevant for additivity and for use in the neighbor joining algorithm. It can be shown that is additive under rather general evolutionary models, although not when evolutionary speed is site dependent.
Before moving on, we should point out that the data required to compute the log-det distance are precisely the same as those needed to compute the two MI-based distances and , provided one uses for the latter the single-letter Shannon formulas with indels deleted. In that case, and where , and is the number of sites in the alignment. This is the main reason we will later compare these three distances in detail.
Tools
We utilized the MAVID [27] and Kalign [28] global sequence alignment programs available for download at [29] and [30] . We also experimented with STRETCHER [31] , lagan [32] and CLUSTALW 2 [33] , and observed similar results. We made no effort to optimize the scoring parameters of the algorithms and only used the default values.
To evaluate we utilized the expert model (XM) DNA compression algorithm [19] . To evaluate we used lpaq1 [17] . Using lpaq1 was not crucial, with the standard LINUX tools gzip and bzip2 producing similar results. For DNA we also explored GenCompress [23] and bzip2. Both showed markedly inferior results to XM (see supplementary information), although their ability to compress single sequences is not so much inferior to XM [19] . Presumably this is due to the fact that XM is more efficient in finding and exploiting approximate repeats, which is crucial in compressing concatenated strings.
The complete mtDNA sequences used in our analysis were downloaded from [34] . We paid special attention to eliminate incomplete sequences and sequences with too many wild cards. We also took care to circularly shift the sequences (mtDNA forms in most cases a closed ring) in order to improve the alignments. We used different subsets of sequences for different plots. In a few cases we also flipped the strands, if this led to much better alignments. Overall, we used nearly 1800 sequences. | Results
Alignment based mutual informations versus compression based mutual informations
Our first results concern the agreement between the two estimates and . In Figure 2 we compare estimates obtained with XM to estimates obtained with the MAVID alignment tool [29] and with subsequent compression using lpaq1. It is well known that DNA and amino acid sequences are hard to compress [18] , [19] , thus one might expect that depends strongly on the compression algorithm used. This is indeed the case, as seen from the first figure in Material S1 , where we compare values of obtained with three different compression algorithms: The general purpose compressor lpaq1 [17] and the two special DNA compressors GeneCompress [18] and XM [19] . From this figure it is clear that XM is far better the other two. Note that it is very likely that an imperfect compression algorithm underestimates rather than overestimates MI – although we do not know a rigorous theorem to this effect.
In view of this, it is not obvious that the estimates produced by XM are realistic either. It is thus highly significant that the two estimates shown in Figure 2 are approximately equal, despite the fact that alignment algorithms and compression algorithms follow drastically different routes. The slight downward shift from the diagonal, particularly visible for large MI pairs, is due to an off-set of bytes in the XM algorithm. Points above the diagonal indicate that concatenation and compression – using the XM algorithm – produce a better estimate of MI, while points below indicate that MAVID alignment followed by compression of its translation string produced a better estimate. The invertebrate-invertebrate pairs far above the diagonal in Figure 2 correspond to pairs of species where the individual genes are similar, but their ordering is changed (this refers in particular to all pairs with and Kbyte). In that case a compression algorithm is superior to a global alignment algorithm, since it is not affected by shuffling open reading frames (ORFs). Most negative estimates for MI seen in Figure 2 represent cases where shuffling the ORFs prevented reasonable global alignments. Particularly interesting are pairs of mammals with . We checked that all of them involve a subspecies of sikka deer (Cervus nippon taiouanus, GenBank accession number DQ985076), in which a single gene (NADH6) is supposedly on the opposite strand compared to all other mammals.
Agreement between and could have been improved presumably in many cases by masking part of the genome, but we have not tried this. In any case, the occasional disagreements are of particular interest, since they indicate where one of the two approaches encountered particular difficulty. Generally speaking Figure 2 suggests that DNA compression can still be improved slightly, as seen from pairs with between 1 and 2 Kbyte (corresponding roughly to species in different families but the same orders). On the other hand, purely compression based MI estimates give non-trivial (at least positive) results even across different classes.
Comparison between different alignment algorithms
MI estimates obtained using other global alignment algorithms are similar to those obtained with MAVID; an example is shown in Figure 3 . In this figure we see that MAVID produced slightly, but systematically better alignments. However, because neither algorithm's scoring scheme was optimized, we do not consider this figure to indicate which of the two alignment algorithms is better. Rather, it represents a proof of principle that our method can be used to identify strengths and weakness of different alignment algorithms and evaluate objectively the sequence similarity in any given alignment.
Correlations within single translation strings: Shannon informations
In Figure 4 we show compression based conditional complexity estimates for animal mtDNA translation strings plotted against the corresponding single letter Shannon entropies . In the latter, we have not eliminated indels, i.e. they are based on the nine letter alphabet . Thus the difference between and is entirely based on correlations, detected by the compression algorithm (in this case lpaq1).
As goes to zero, the two estimates agree, up to a small initialization cost for lpaq1 of bytes. The estimates agree because the translation string is mostly composed of zeros, with the few substitutions being far apart and weakly correlated. For increasing , however, the compression algorithm often gives significantly lower estimates, thus indicating strong correlations within the translation string. More detailed analysis of pairwise correlations (unpublished) suggests that these are mostly correlations between letters (i.e., inserts and gaps rather than substitutions). The fact that indels occur strongly correlated is well known [1] , and is also assumed in most alignment scoring schemes.
Therefore, if the information encoded in gaps is to be taken into account, it is necessary to go beyond the single letter approximation when estimating realistic and absolute sequence similarities. Furthermore, taking into account only pairwise letter correlations would not be sufficient either. This, of course, is not completely new, and the most common way to deal with this problem is to simply ignore indels [7] . Indeed there seems to exist a wide spread opinion that indels are not very informative and useful. Whether this is true or whether it just reflects an inability to deal with this information efficiently is an open question. In any case, the most straightforward way to deal with it would be based on algorithms using data compression.
Comparison with p-distances: The effect of indels
A very simple but popular distance measure between sequences (both DNA and amino acids) is p-distance. It is defined by first removing all indel positions and then counting the number of positions where the two sequences disagree [7] , where is the number of observed substitutions and is the total number of (non-indel) sites. Since this quantity saturates with increasing evolutionary distance, a slightly more sophisticated version is the Poisson corrected (PC) p-distance [7] , . We note that neither nor take into account the type of substitutions, any information contained in indels, or any information contained in internal correlations within the translation strings.
Our main interest here is to see which of these three neglected aspects (type of substitution, indels, correlations) has the biggest effect. In Figure 5 we show a scatter plot of the normalized compression distance , estimated via , against , for pairs taken from all over the animal kingdom. In order to avoid meaningless alignments, we took in each pair only members in the same (sub-) phylum (hexapoda, mollusca, crustacea, chelicerata, cnidaria, porifera, platyhelminthes, echinodermata) or in the same (super-) class (mammals, sauropsida, amphibia, actinopterygii). We also eliminated pairs with , as we would have otherwise too many biologically meaningless alignments. Here, are the sequence lengths; this criterion guarantees that there are not too many insertions into the longer sequence, and not too many deletions from the shorter. We found that there is a roughly monotonic relationship between and , with occasional, strong, deviations. By far the strongest factor leading to these deviations is the difference in length of the paired raw (i.e. unaligned) sequences. Nearly all gross outliers in Figure 5 correspond to pairs in which one member has a very long mitochondrial genome, leading to a large number of indels.
As we had pointed out in the previous subsection, it is widely believed that indels are not very informative. We plan to check this more carefully in a future publication, using a methodology based on a large number of quartets for sequences similar to the one described in the next section.
Comparison with log-det distances: The effect of substitution types
Finally, we want to compare our distance metrics and to the log-det distance given in Eq. (12). In order to simplify the discussion and to use exactly the same input for all three metrics, we use the same alignment algorithm (MAVID) for each pair and delete all indels. As mentioned above, does not, in general, satisfy the triangle inequality. But this does not preclude it from being satisfied in all “typical cases”. To test this we first check whether the triangle inequality is actually violated or not in randomly chosen triplets, drawn from the entire animal kingdom, with the same selection criteria as in the previous subsection. Note that due to the omission of indels none of these “distances” actually have to satisfy the triangle inequality. Indeed, we found 11 violations for the log-det distance, and none for either of the MI-based distances.
Next we tried to check whether is at least approximately additive. Since we do not have the true evolutionary distances, we take as a proxy. In Figure 6 we plot against for random pairs. We see that:
Roughly, the dependence is linear. Thus, to the extent that is linear, is too. Thus it should not be affected by long branch attraction. This is in contrast to which – when plotted against – is strongly non-linear (data not shown).
On a finer scale, one sees several deviations. The most conspicuous, perhaps, is that insects (hexapods) are systematically above the main curve. This is due to the strong compositional bias in most insects, where C/G is underrepresented compared to A/T. This reduces the entropies of individual sequences. At the same time, however, substitution rates involving C and G are not as suppressed. As a consequence, the ratio is enhanced compared to other phyla, and is increased. This is a desirable effect. It is well known [35] that similar compositional bias can make two sequences look more closely related, even if they are not closely related evolutionarily. While this applies fully to , the effect is at least smaller for .
For intermediate distances ( ), many mammals are below the main line. In particular, consider the two pairs well below it at . Both involve the spectacled bear (Tremarctos ornatus) and another Ursinae species. For whatever reasons, these two translation strings contain an unusually large ratio of transitions to transversions that would otherwise only be typical for much more closely related species. This reduces the information content when compared to unbiased substitutions with the same total frequencies. At the same time, the individual sequences are not very strongly biased. Thus is reduced, but is not – since it is only weakly dependent on the detailed substitution rates. Again we claim that this favors over .
A clear decision whether this is indeed true can only be made by detailed comparison of phylogenies predicted on the basis of these metrics with the true phylogenies. Since the latter are of course unknown, we take inferences made in the literature as proxies. Our detailed strategy is the following:
We first choose random quadruples from all over the animal kingdom. We use the same taxonometric restrictions, to avoid too many pairs which cannot be meaningfully aligned. Thus each quadruple (or “quartet”) contains only species from the same (sub-)phylum or the same (super-)class, respectively. We also used the same cut on the number of indels, in order to eliminate false alignments.
For each quartet, we find the topologies suggested by each of the three metrics, and count the number of cases where two metrics disagree. This gave 185543 quartets (1.9%) where and disagree, 429386 quartets (4.3%) where and disagree, and 380487 quartets (3.8%) where and disagree.
For each quartet we compute a significance with which the suggested topology is actually preferred. This significance is explained in detail in the supplementary information. It involves both the amount by which the four-point condition is violated, and the relative length of the central edge, if the data are approximated by an additive tree. For each pair of metrics we then pick the quartets for which the metrics disagree most significantly (as measured by the sum of the two significances). Actually, we do not strictly choose the worst disagreements, as they would cluster within a few taxa and we want our results to represent as much of the entire animal kingdom as possible. As such, we take relatively more quartets in taxa which are underrepresented in GenBank, and we reject quartets (not entirely systematically) if three of the four species had already appeared in many selected quartets.
In this way we selected 129 “worst” disagreements between and , and 129 “worst” disagreements between and . For reasons that will become clear later, we did not select worst disagreements between and , except for a few cases. For each of these worst cases we searched the literature and established the “correct” topology. Details are again given in the tables found in the Material S1 .
The final results of this are summarized in Table 1 . They clearly indicate that the log-MI metric is vastly superior the log-det distance, in spite of the latter's superior theoretical foundations. This is at odds with the fact that the log-MI metric is not a proper distance, and does not, in any reasonable model, satisfy the four-point condition (Eq. (8)). The reason obviously is that takes into account, in an optimal model-independent manner, compositional details that does not. The comparison between and is much less clear. One might have expected that the strong non-additivity of makes it unsuitable for this sort of phylogenetic application. But this is not so clear; is only marginally better. This seems surprising, but a possible reason for it will be given in the discussion.
Before moving on we highlight a few notable observations about our quartet analysis. Previously, we pointed out that the spectacled bear (T. ornatus) is anomalous either in or in . Indeed, it appears twice in the first table of the supplementary, and both times gives the correct grouping. A similar anomaly is seen in Figure 6 for fish (actinopterygii) at . Most of these correspond to Albula glossodonta (GenBank AP002973) paired with other fish. The second table of Material S1 shows that for most of these pairs the log-MI distance gives a better estimate.
We find that discrepancies between and are very unevenly distributed over the taxa. While we found no disagreements in the chaetognatha, there are a large number in the nematods, most favoring . Indeed, it seems that the nematod phylogenetic tree constructed using would be systematically different from the tree constructed using and other analyses.
It is well known [10] , [11] that the log-det distance is additive only when the evolutionary rate is constant over all sites. One can argue that an analysis that does not distinguish sites with different evolutionary speeds is not fair to . In response we put forth the following three points: (i) The main problem with does not seem to be a lack of additivity, but rather insufficient attention to the specific types of substitution; (ii) Inhomogeneities in the evolutionary speed should affect not only the log-det distance, but most other distance measures as well. Specifically we cannot see why it should not negatively affect too; (iii) Similarly, analyzing sites with different speeds separately should improve the results for any distance measure – as long as it can be done unambiguously, without too much effort, and without reducing the amount of usable data excessively. In view of the last three caveats we believe that “naive” analyses, such as the one presented above, have and will continue to have their merits.
The full picture: Comparison of several distance metrics
So far we have only compared in detail quartet classifications based on log-det distances and on single letter Shannon MI. We have used Shannon MI because its estimation is less ambiguous than compression based MI estimation, and because it uses exactly the same input — the base substitution frequency matrix after removing indels — as the log-det distance. But our tenet is, of course, that compression based estimates should be superior as long as they use the information about indels efficiently. In addition to the log-det distance, there are several measures that are often used. In this subsection we make several pairwise comparisons similar to the one made in the previous subsection. But we restrict ourselves to mammals, as these have the best understood phylogeny, and we expect the least numbers of errors in the literature classification.
In this subsection we compare MI based distances with the log-det and with both versions of the Kimura distance (Eqs. (S8,S9)) discussed in Material S1 . We do not present all possible combinations, as this would produce a huge matrix. Instead, we focus on a subset of the distance measure pairs, but we claim that this subset is large enough to present a clear overall picture.
Results are shown in Table 2 . As mentioned above, we analyzed only mammals for this, but we looked at all possible quartets. Our criteria for identifying the “worst” disagreements is the same as in the previous subsection. Each comparison is based on at least 60 disagreeing quartets. In this table, “Kimura ” and “Kimura ” refer to Eqs. (S8) and (S9) in Material S1 , respectively; “Shannon nolog-MI” stands for (Eq. (10)), “Shannon log-MI” stands for the logarithmic version of the Shannon distance (Eq. (11), right hand side), “transl. string, nolog” stands for (Eq.(9) with the MI estimated via alignment), “transl. string, log” stands for its logarithmic version (Eq.(11), left hand side), “XM compression, nolog” stands for with the MI estimated via concatenation and compression with XM, and “XM compression, log” stands for its logarithmic version.
In the present paper we have not presented any detailed application to a specific open phylogenetic problem. We also have not considered larger phylogenetic trees, in view of the imperfections of all existing distance based tree reconstruction algorithms. Instead, we have concentrated on quartets, since there we can obtain high statistics and the inference of the tree from a given distance matrix is trivial. Also, for the most detailed numerical comparison we have concentrated on Shannon information based methods, rather than on compression based methods for estimating MI. The reason is simply that we desired a comparison with other methods (mainly the log-det distance) which is as straightforward and unambiguous as possible. Indeed, it is trivial to replace Eq. (12) by Eqs. (10), (13), (14). In this way we hope to have the best chance to convince even skeptical readers that mutual information based distance measures are useful in sequence analysis.
We have also presented similar – but less complete – analyses based on large numbers of random quartets for (at least partially) compression based algorithms and have demonstrated that distances based on data compression give even better phylogenies. Indeed, from Table 2 we can draw a number of conclusions:
All versions based on MI are better than any version not based on MI.
Kimura (based directly on the log-likelihood of the data with respect to the Kimura model) seems better than the conventional Kimura , which just estimates the total number of substitutions. This supports our suspicion that counting transitions and transversions with the same weight is not a good strategy.
Nevertheless, does worse than , as expected: As we point out in Material S1 , the log-likelihood for the Kimura model is essentially a coarse grained MI, where different substitutions are lumped together (resp., the probabilities predicted by the model replace the true observed probabilities). It would be hard to see why this should give superior results, given the ease and robustness with which single letter Shannon entropies can be estimated.
Within the class of MI based distances, those which do not neglect indels seem systematically better.
Among the latter, distances based on do better than those based on . This is surprising, as we saw that is for mammals systematically larger (and thus supposedly better) than .
Logarithmic transformation of MI based distances seems to give mixed results. It improves the distances slightly for Shannon MI and for , but it has very negative effects when used with based on XM. We conjecture that this reflects two sides of the logarithmic transformation for distantly related pairs. On the one hand, it largely eliminates systematic errors due to deviations from metric additivity (the Felsenstein phenomenon). On the other hand it amplifies noise. To illustrate this, we discuss in Material S1 a quartet where both the original Shannon MI based distances and their log-transformed versions give wrong results, but for opposite reasons. We speculate that the detrimental effect dominates for , because MI estimation by compression is more noisy (due to the less systematic way that present state-of-the-art compression algorithms work) than .
Thus, contrary to wide spread opinion, information about indels can be directly used for phylogenetics, even without any detailed model for how they were generated. A more detailed presentation of these data and their implications will be given elsewhere.
We believe that so far we have only scratched the true potential of (algorithmic) information theory for sequence analysis. Several generalizations and improvements are feasible and are listed below:
Use more efficient encodings of the translation string. For instance, we only used the letters and to reconstruct , but one could also use in addition , and/or .
Use local alignments instead of global ones. In a local alignment between sequences and , large parts of are not aligned with at all and are encoded without reference to . Only the aligned parts give information from that can be used to recover . Before making the jump from global to local alignments, an intermediate step would be a “glocal” alignment tool such as shuffle-lagan (“slagan”) of [2] .
Construct objective measures based on information theory for the quality of multiple alignments. A straightforward measure is the information about sequence obtained from aligning it simultaneously with and . Assume e.g. that the sequences and are much more similar to each other than either and or and (as for human, chimpanzee, and chicken). In order to measure the MI between chicken and the primates, one could first align and and then align, in a second step, to the fixed alignment .
Conclusions
At present, biological sequence analysis is heavily based on the concept of alignment. There exist proposals for alignment-free approaches, and it has been suggested that they will become more and more important as more sequence data become available [36] . To us it seems an open question whether alignment-free algorithms for sequence comparison will become widely used, whether they will eventually displace alignment-based algorithms, or whether both approaches will merge into a unified approach. We hope that we have shown with the present work that an amalgamation of both methods (alignment-based and alignment-free) is possible. More precisely, by showing that mutual informations between two sequences can be easily estimated from global alignments, we have established a direct link between sequence alignment, Shannon information theory, and methods based on data compression and Kolmogorov information theory. Technically, we have dealt only with pairwise global alignment, but at least the basic concepts should have much wider applicability.
From another point of view, the present paper deals with the basic notion of parsimony . In bioinformatics (and in phylogeny in particular) maximal parsimony in dealing with several objects is often taken as synonymous to minimal number of changes needed to go from the description of one object to the description of another. This is most clearly formulated in the so-called “maximum parsimony method” of distance-free phylogenetic tree construction [7] , but it also underlies the concepts of p- and log-det distances. However, the invention of the Morse alphabet in the nineteenth century, and the theoretical works by Shannon, Kolmogorov, and others in the middle of the last century might cast some doubt on it. It is Rissanen's minimum description length principle [37] , [38] , however, that makes this view obsolete today. Instead of paying attention to the number of changes, one should pay attention to the information needed to encode these changes. We call this “true parsimony”. In this sense, the maximum parsimony method does not really aim for maximal true parsimony. On the other hand, likelihood based and Bayesian methods do aim for true parsimony, but at the cost of depending on explicit models. One goal of the present paper is to show how true parsimony can be measured in less model dependent ways and how maximum true parsimony can be achieved to various degrees of approximation. Moreover, even the crudest approximation – based on MI obtained via single-letter Shannon entropies, with all information about indels discarded – can lead to important practical improvements. | Conclusions
At present, biological sequence analysis is heavily based on the concept of alignment. There exist proposals for alignment-free approaches, and it has been suggested that they will become more and more important as more sequence data become available [36] . To us it seems an open question whether alignment-free algorithms for sequence comparison will become widely used, whether they will eventually displace alignment-based algorithms, or whether both approaches will merge into a unified approach. We hope that we have shown with the present work that an amalgamation of both methods (alignment-based and alignment-free) is possible. More precisely, by showing that mutual informations between two sequences can be easily estimated from global alignments, we have established a direct link between sequence alignment, Shannon information theory, and methods based on data compression and Kolmogorov information theory. Technically, we have dealt only with pairwise global alignment, but at least the basic concepts should have much wider applicability.
From another point of view, the present paper deals with the basic notion of parsimony . In bioinformatics (and in phylogeny in particular) maximal parsimony in dealing with several objects is often taken as synonymous to minimal number of changes needed to go from the description of one object to the description of another. This is most clearly formulated in the so-called “maximum parsimony method” of distance-free phylogenetic tree construction [7] , but it also underlies the concepts of p- and log-det distances. However, the invention of the Morse alphabet in the nineteenth century, and the theoretical works by Shannon, Kolmogorov, and others in the middle of the last century might cast some doubt on it. It is Rissanen's minimum description length principle [37] , [38] , however, that makes this view obsolete today. Instead of paying attention to the number of changes, one should pay attention to the information needed to encode these changes. We call this “true parsimony”. In this sense, the maximum parsimony method does not really aim for maximal true parsimony. On the other hand, likelihood based and Bayesian methods do aim for true parsimony, but at the cost of depending on explicit models. One goal of the present paper is to show how true parsimony can be measured in less model dependent ways and how maximum true parsimony can be achieved to various degrees of approximation. Moreover, even the crudest approximation – based on MI obtained via single-letter Shannon entropies, with all information about indels discarded – can lead to important practical improvements. | Conceived and designed the experiments: OP PG MP. Performed the experiments: OP PG MP. Analyzed the data: OP PG MP. Contributed reagents/materials/analysis tools: OP PG MP. Wrote the paper: OP PG MP.
Background
Existing sequence alignment algorithms use heuristic scoring schemes based on biological expertise, which cannot be used as objective distance metrics. As a result one relies on crude measures, like the p- or log-det distances, or makes explicit, and often too simplistic, a priori assumptions about sequence evolution. Information theory provides an alternative, in the form of mutual information (MI). MI is, in principle, an objective and model independent similarity measure, but it is not widely used in this context and no algorithm for extracting MI from a given alignment (without assuming an evolutionary model) is known. MI can be estimated without alignments, by concatenating and zipping sequences, but so far this has only produced estimates with uncontrolled errors, despite the fact that the normalized compression distance based on it has shown promising results.
Results
We describe a simple approach to get robust estimates of MI from global pairwise alignments. Our main result uses algorithmic (Kolmogorov) information theory, but we show that similar results can also be obtained from Shannon theory. For animal mitochondrial DNA our approach uses the alignments made by popular global alignment algorithms to produce MI estimates that are strikingly close to estimates obtained from the alignment free methods mentioned above. We point out that, due to the fact that it is not additive, normalized compression distance is not an optimal metric for phylogenetics but we propose a simple modification that overcomes the issue of additivity. We test several versions of our MI based distance measures on a large number of randomly chosen quartets and demonstrate that they all perform better than traditional measures like the Kimura or log-det (resp. paralinear) distances.
Conclusions
Several versions of MI based distances outperform conventional distances in distance-based phylogeny. Even a simplified version based on single letter Shannon entropies, which can be easily incorporated in existing software packages, gave superior results throughout the entire animal kingdom. But we see the main virtue of our approach in a more general way. For example, it can also help to judge the relative merits of different alignment algorithms, by estimating the significance of specific alignments. It strongly suggests that information theory concepts can be exploited further in sequence analysis. | Supporting Information | PG would like to thank Olav Zimmermann for numerous discussions. | CC BY | no | 2022-01-13 08:14:21 | PLoS One. 2011 Jan 4; 6(1):e14373 | oa_package/ed/aa/PMC3014950.tar.gz |
|
PMC3014951 | 21134293 | Background
Being overweight is also a major worldwide public health problem. In Japanese adults, it was demonstrated that 29.3% of men and 19.5% of women are overweight (body mass index (BMI); weight (kg)/height (m) 2 , ≥25 kg/m 2 ), with approximately 3.3% of men and 3.2% of women designated as obese (BMI ≥ 30 kg/m 2 ) [ 1 ]. Several large-scale studies have shown that overweight people, defined on the basis of a high BMI, increases the risk of hypertension, dyslipidemia, diabetes and developing cardiovascular disease (CVD) such as stroke and myocardial infarction than subjects with normal BMI [ 2 - 4 ]. Moreover, it has been suggested that individuals exist who are not overweight, but who, like people with overt obesity, also have hyperinsulinemia, are insulin resistant, and are predisposed to hypertriglyceridemia, type 2 diabetes mellitus, and premature cardiovascular disease [ 5 , 6 ]. Much investigation has been focused on insulin resistance as the primary determinant of atherosclerotic risk factors, independent of BMI [ 7 ].
On the other hand, insulin resistant persons also have a characteristic dyslipidemia [ 8 ] and measuring these variables might help identify insulin resistance. Plasma triglycerides, high-density lipoprotein cholesterol (HDL-C) [ 9 , 10 ], and total cholesterol (T-C)/HDL-C are independently associated with insulin resistance and risk factors of CVD [ 11 ]. However, in Japanese community-dwelling persons, there are few studies to demonstrate a relationship between lipid profiles and insulin resistance, categorized by BMI.
We took advantage of the large representative sample of Japanese adults who participated at the time of their annual health examination. We investigated how lipid profiles were associated with insulin resistance in healthy Japanese adults. For this, we used cross-sectional data from community-dwelling participants without clinical diabetes. | Methods
Subjects
Participants were recruited at the time of their annual health examination in a rural town with a total population of 11,136 (as of April 2002) and located in Ehime prefecture, Japan, in 2002. Among the 9,133 adults (4,395 of them male) aged 19 to 90 years in this population, a random sample of 3,164 subjects (34.6%) was recruited. Information on medical history, present conditions, and drugs was obtained by interview. Other characteristics, e.g., smoking and alcohol habits, and medication, were investigated by individual interviews using a structured questionnaire. Subjects taking medications for hypertension, diabetes, or dyslipidemia were excluded. The final study sample included 614 men and 779 women. This study was approved by the ethics committee of Ehime University School of Medicine, and written informed consent was obtained from each subject.
Evaluation of confounding factors
Information on demographic characteristics and risk factors was collected using clinical files. Body mass index was calculated by dividing weight (in kilograms) by the square of the height (in meters). We measured blood pressure in the right upper arm of participants in a sedentary position using an automatic oscillometric blood pressure recorder (BP-103i; Colin, Aichi, Japan) while the subjects were seated after having rested for at least 5 min. Smoking status was defined as the number of cigarette packs per day multiplied by the number of years smoked (pack·year), and the participants were classified into never smokers, past smokers, light smokers (<30 pack·year) and heavy smokers (≥30 pack·year). The daily alcohol consumption was measured using the Japanese liquor unit in which a unit corresponds to 22.9 g of ethanol, and the participants were classified into never drinkers, occasional drinkers (<1 unit/day), light drinkers (1-1.9 unit/day), and heavy drinkers (≥2 unit/day). T-C, TG, HDL-C, fasting blood glucose (FBG), creatinine (enzymatic method), uric acid, and immunoreactive insulin (IRI), and high molecular weight (HMW) adiponectin (FUJIREBIO, Tokyo, Japan) were measured during fasting low-density lipoprotein cholesterol (LDL-C) levels were calculated using the Friedewald formula [ 12 ]. Participants with TG levels ≥400 mg/dl were excluded. Homeostasis of minimal assessment of insulin resistance (HOMA-IR) was calculated from FBG and IRI levels using the following formula; {FBG (mg/dL) × IRI (mU/mL)}/405 [ 13 ], and a borderline level of Insulin resistance was defined as HOMA-IR ≥ 1.6 [ 13 ], and a definite level, HOMA-IR ≥ 2.5 [ 14 ].
Statistical analysis
Statistical analysis was performed using PASW Statistics 17.0 (Statistical Package for Social Science Japan, Inc., Tokyo, Japan). All values are expressed as mean ± standard deviation (SD), unless otherwise specified. Data for TG, FBG, HOMA-IR, and serum HMW adiponectin were skewed, and log-transformed for analysis. Subjects were divided into two groups based on BMI (non-obese, <25.0 kg/m 2 ; overweight, ≥25.0 kg/m 2 ), and differences between the two groups were determined by Student's t-test andχ 2 test. In addition, areas under the receiver operating characteristic (ROC) curves were determined for each variable to identify the predictors of insulin resistance. Areas under the ROC curves are provided with standard errors. An ROC curve is a plot of the sensitivity (true positive) versus 1-specificity (false positive) for each potential marker tested. The area under the ROC curve is a summary of the overall diagnostic accuracy of the test. The best markers have ROC curves that are shifted to the left with areas under the curve near unity. Nondiagnostic markers are represented by diagonals with areas under the ROC curves close to 0.5. Likelihood ratios were calculated as the ratios of sensitivity - (1 - specificity) (positive likelihood ratio) and (1 - sensitivity) - specificity (negative likelihood ratio). A value of P < 0.05 was considered significant. | Results
Background factors of subjects categorized by BMI
Table 1 shows the value of each background factor categorized by BMI. The subjects comprised 614 men aged 58 ± 14 (mean ± standard deviation; range, 20-89) years and 779 women aged 60 ± 12 (range, 21-88) years. The mean BMI in the study sample was 23.1 ± 3.1 kg/m 2 , with 1,042 non-obese (BMI < 25.0 kg/m 2 ) (74.8%) and 351 overweight (BMI ≥ 25 kg/m 2 ) (25.2%). Alcohol consumption, systolic blood pressure (SBP), diastolic blood pressure (DBP), T-C, TG, LDL-C, TG/HDL-C ratio, non-HDL-C, LDL-C/HDL-C ratio, and uric acid were significantly higher in subjects with a BMI ≥ 25.0 kg/m 2 , but age, HDL-C and serum HMW adiponectin were significantly lower in that group. There were no inter-group differences in sex, smoking status, eGFR, and prevalence of CVD.
Insulin resistance of subjects categorized by BMI
Fasting blood glucose, IRI, and HOMA-IR were significantly higher in overweight subjects (Table 2 ), and prevalence of insulin resistance (HOMA-IR > 1.6 or >2.5) was significantly higher in overweight subjects than in non-obese subjects.
Comparison of areas under ROC curves (95% CI) for potential markers of insulin resistance of subjects categorized by BMI
In non-obese subjects, the ROC curve analyses showed that the best marker of insulin resistance was LDL-C/HDL-C ratio, with an area under the ROC curve of 0.74 (0.66-0.80) (Table 3 ; Figure 1 ). The TG/HDL-C ratio, HDL-C, and non-HDL-C also discriminated insulin resistance, as they had areas under the ROC curve of 0.31 (0.24-0.38), 0.69 (0.62-0.75) and 0.69 (0.62-0.75), respectively. In overweight subjects, TG and TG/HDL-C ratio were more effective.
Optimal cut-off point of TG/HDL-C ratio and LDL-C/HDL-C ratio for predicting insulin resistance of subjects categorized by BMI
Table 4 shows the cut-off points of TG/HDL-C ratio and LDL-C/HDL-C ratio for identifying insulin resistance. The optimal cut-off point to identifying insulin resistance for these markers yielded the following values: TG/HDL-C ratio of ≥1.50 and LDL-C/HDL-C ratio of ≥2.14 in non-obese subjects, and ≥2.20, ≥2.25 in overweight. In non-obese subjects, the positive likelihood ratio value indicates that the odds of insulin resistance increased by 2.30-fold if the LDL-C/HDL-C ratio was positive (the value ≥2.14). This ratio was greater for LDL-C/HDL-C ratio than TG/HDL-C ratio. The negative likelihood ratios indicate the extent to which the odds of insulin resistance decrease if the test is negative. These odds also decreased more so for LDL-C/HDL-C ratio. In overweight subjects, these values were similar. | Discussion
In the present study, we examined whether lipid profiles (i.e., TG, HDL-C, LDL-C, TG/HDL-C ratio, LDL-C/HDL-C ratio and non-HDL-C) were associated with insulin resistance in Japanese adults, categorized by body mass index. Most fundamental is the fact that not all overweight or obese persons are insulin resistant. In non-obese subjects, 6.0% of them were insulin resistant, and the best marker of insulin resistance was LDL-C/HDL-C ratio, but HDL-C, TG/HDL-C ratio and non-HDL-C also discriminated insulin resistance. In overweight subjects, the areas of TG and TG/HDL-C ratio were greater than those of the other parameters. The optimal cut-off point to identifying insulin resistance for these markers yielded the following values: TG/HDL-C ratio of ≥1.50 and LDL-C/HDL-C ratio of ≥2.14 in non-obese subjects, and ≥2.20 and ≥2.25, respectively in overweight. The positive likelihood ratio was greatest for LDL-C/HDL-C ratio in non-obese subjects. Lipid ratio of LDL-C/HDL-C might be used as an integrated and simple lipid measure to evaluate insulin resistance in non-obese subjects.
Resistance to insulin-mediated glucose disposal is distributed continuously through the general population [ 15 ], and we have no criterion with which to identify a participant as being insulin resistance or insulin sensitive. However, we classified a participant as insulin resistant if he or she was in HOMA-IR > 2.5 [ 14 ]. Previous studies have shown that HOMAIR-based insulin resistance scores strongly correlate with glucose clamp-assessed insulin resistance [ 13 , 15 ]. However, the validation was carried out in only a few subjects, and HOMA-IR is less accurate and precise than the glucose clamp in measuring insulin resistance, but this limitation is mitigated when the number of subjects examined is large, as in our study [ 16 ].
Hypertriglyceridemia and low HDL-C almost never occurred as isolated disorders, and were nearly always associated with insulin resistance because insulin affects very low-density lipoprotein and HDL-C metabolism [ 17 ]. In previous studies, several lipid ratios have been proposed as simple and useful clinical indicators of insulin resistance. The TG/HDL-C, the T-C/HDL-C, and the LDL-C/HDL-C ratio have shown similar potential for insulin resistance, though the reports are not entirely consistent. In 50 white Americans, both TG and TG/HDL-C ratio were acceptable markers for insulin resistance, with area under the ROC curve of 0.763 and 0.770, respectively, but poor predictors in 99 African Americans, with the values at 0.625 and 0.639, respectively [ 18 ]. It was demonstrated that the relationship between TG and TG/HDL-C with insulin resistance differs by ethnicity. In 3,014 patients (mean age 54 years; 55% women), TG/HDL-C ratio and T-C/HDL-C ratio were related to insulin resistance assessed by the top quartile of the HOMA-IR, and the area under the ROC curves for predicting insulin resistance with TG/HDL-C ratio and T-C/HDL-C ratio were 0.745 and 0.707, respectively [ 19 ]. LDL-C/HDL-C ratio [ 20 ] as well as TG/HDL-C ratio [ 21 ] have advantages for standardization and identification of patients with an atherogenic lipoprotein profile as a surrogate maker of insulin resistance. Also in our study, both LDL-C/HDL-C and TG/HDL-C ratio were useful makers of insulin resistance, especially in all subjects or non-obese subjects. However, these makers were weaker in overweight subjects. Kimm et al. [ 21 ] demonstrated that the lipid ratios of TC/HDL-C, LDL-C/HDL-C and TG/HDL-C, as well as TG and HDL-C, were consistently associated with metabolic syndrome and insulin resistance in participants without metabolic syndrome, though these relations were weaker in participants with metabolic syndrome. Lipid ratios that include information on at least two measures might have a more integrated explanation than single lipid measures such as TG or HDL-C [ 21 ].
Some limitations of this study must be considered. First, the response rate was as low as 35% that is usually the casein other conventional community studies in Japan. However, the relatively large sample size enabled the assessment of an extensive array of insulin resistances in relation to lipid profiles. Second, the cross-sectional study design is limited in its ability to eliminate causal relationships between lipid profiles and HOMA-IR. Third, our definition of HOMA-IR is based on a single assessment of FBS and IRI, which may introduce misclassification bias. Therefore the demographics and referral source may limit generalizability.
In conclusion, the present study demonstrated that special lipid profiles are associated with insulin resistance according to BMI in a general population. The ability to identify who is non-obese or overweight and who are insulin resistant could help health care professionals in bringing about lifestyle interventions. In that context, use of the cutoff-points of LDL-C/HDL-C ratio or TG/HDL-C ratio described in this report is simple and useful. The present data documented that insulin resistance was present even in subjects with BMI within the normal range. Further prospective population-based studies are needed to investigate the changes in lipid metabolism by lifestyle interventions. | Background
The aim of the present study was to examine how lipid profiles are associated with insulin resistance in Japanese community-dwelling adults.
Methods
This cross-sectional study included 614 men aged 58 ± 14 (mean ± standard deviation; range, 20-89) years and 779 women aged 60 ± 12 (range, 21-88) years. The study sample were 1,042 (74.8%) non-obese (BMI < 25.0 kg/m 2 ) and 351 (25.2%) overweight (BMI ≥ 25 kg/m 2 ) subjects. Insulin resistance was defined by homeostasis model assessment of insulin resistance (HOMA-IR) of at least 2.5. The areas under the curve (AUC) of the receiver operating characteristic curves (ROC) were used to compare the power of these serum markers.
Results
In non-obese subjects, the best marker of insulin resistance was low-density lipoprotein cholesterol (LDL-C)/high-density lipoprotein cholesterol (HDL-C) ratio of 0.74 (95% confidence interval (CI), 0.66-0.80). The HDL-C, triglyceride (TG)/HDL-C ratio, and non-HDL-C also discriminated insulin resistance, as the values for AUC were 0.31 (95% CI, 0.24-0.38), 0.69 (95% CI, 0.62-0.75) and 0.69 (95% CI, 0.62-0.75), respectively. In overweight subjects, the AUC for TG and TG/HDL-C ratio were 0.64 (0.58-0.71) and 0.64 (0.57-0.70), respectively. The optimal cut-off point to identifying insulin resistance for these markers yielded the following values: TG/HDL-C ratio of ≥1.50 and LDL-C/HDL-C ratio of ≥2.14 in non-obese subjects, and ≥2.20, ≥2.25 in overweight subjects. In non-obese subjects, the positive likelihood ratio was greatest for LDL-C/HDL-C ratio.
Conclusion
In non-obese Japanese adults, LDL-C/HDL-C ratio may be the best reliable marker of insulin resistance. | Competing interests
The authors declare that they have no competing interests.
Authors' contributions
RK, YT, and KK participated in the design of the study, performed the statistical analysis and drafted the manuscript. NO, TaK, and ToK contributed to acquisition of data and its interpretation. ST and MA contributed to conception and design of the statistical analysis. TM conceived of the study, participated in its design, coordination and helped to draft the manuscript. All authors read and approved the manuscript. | Acknowledgements
This work was supported in part by a grant-in-aid for Scientific Research from the Foundation for Development of Community (2009). | CC BY | no | 2022-01-12 15:21:37 | Lipids Health Dis. 2010 Dec 7; 9:138 | oa_package/bc/9b/PMC3014951.tar.gz |
|
PMC3014952 | 21143884 | Background
Ageing, characterized by an increase of lipid consumption and sedentary lifestyle is known to be linked with metabolic pathologies. The excess of lipid provided by food is stored in localized sites (abdomen, liver, hips) and becomes pathological, disturbing a great number of metabolic functions and contributing to the rise in metabolic syndrome [ 1 ]. Abdominal obesity is also strongly linked with diabetes mellitus and an increase of cardiovascular disease, resulting in a prevalence of adverse cardiovascular events [ 2 , 3 ].
The mechanisms leading to the local accumulation of adipose tissue are not fully elucidated in both humans and animals. Therefore, the main issue for clinicians or researchers is to have techniques for assessing the abdominal fat deposition and its accumulation or change over time. Moreover, the sacrifice of experimental animals permits the determination of the exact value of visceral fat tissue but prohibits follow up of the change to this tissue in vivo.
Studies on the accumulation of abdominal fat mass in animals could help to better understand the mechanisms of metabolic syndrome development but the techniques of investigation are lacking.
In rodents, the total body weight measure does not give relevant information about body composition, which is an essential parameter in this kind of investigation. On the other hand, waist circumference measures have been shown to significantly correlate with cardiovascular disease in humans [ 4 ]. Among several anthropometric measurements, waist circumference was also found to be the best predictor of intra-abdominal fat thickness in normal subjects and therefore of central obesity [ 5 ]. While the ability to predict visceral fat from circumferences and diameters in obese humans has been validated [ 6 ], no anthropometric technique has been developed in animals to date. As aforementioned, the accumulation of intra-abdominal fat can be measured by direct weighing of fat after dissection. Some studies weighed intra-abdominal visceral fat including retroperitoneal, epididymal, and mesenteric fat tissues [ 7 - 11 ]. Lac et al. [ 12 ] measured visceral fat mass by weighing the left perirenal fat pad alone, considering that it was representative of whole abdominal fat tissue. However, because of the sacrifice involved by such techniques, longitudinal studies are excluded. The use of computed tomography in assessing body composition has been shown to be an useful technique to quantify fat tissue deposits [ 7 ]. Magnetic resonance imaging has also been used to measure abdominal fat, and discriminate between locations of adipose tissue [ 13 ]. Unfortunately, these two techniques are hampered by availability and cost; resulting in infrequent use in longitudinal studies. Thus, no technique is currently available to estimate in vivo abdominal fat mass in animals.
Due to its excellent reproducibility [ 14 ], Dual-energy X-ray absorptiometry (DXA) has become one of the best technique for assessing body composition in humans [ 15 ] and rodents [ 16 ]. The DXA analysis software allows specific regions to be identified making it an expedient and cost-effective method to estimate abdominal adipose tissue in humans [ 17 , 18 ]. The validity and reliability of DXA to measure adiposity in the abdominal region has been determined in humans [ 19 ]. Kamel et al. [ 20 ] showed that central fat mass determined by this technique was strongly associated with the value of abdominal visceral fat in both men and women, but no study has investigated abdominal fat mass in rats. Similarly, abdominal circumference measures are among established criteria for human metabolic syndrome diagnosis, yet no anthropometric technique has been developed in animals.
Therefore, the purpose of the present study was to validate a DXA technique and abdominal circumference measurement as predictors of visceral adipose tissue in lean and fat male rats. | Methods
Animal Care
All experimental designs and procedures were made in accordance to the current legislation on animal experience in France and were approved by the ethical committee for animal experimentation (CREEA Auvergne, CE1-09). Considering a minimum of 10% difference of visceral fat mass between fat and lean rats, the statistical power calculation resulted in a number of 12 rats per groups for p < 0.05 and a type II error of 0.01%. Therefore, we included groups of 14 rats.
Forty-three Wistar male rats (CERJ Janvier ® , Le Genest Saint-Isle, France) were included in the study. In order to measure a large range of subjects, we chose a group of 29 rats aged 11 months and 14 rats aged 7 months with a range of body weights. Among the group of 29 rats, 14 rats were previously fed with a normal diet (lean group) and 15 rats were given a high fat/high sucrose diet (fat group) for four months to increase fat mass storage. The 7-month old rats were fed with a normal diet. Rats were individually housed in a temperature-controlled room (20-22°C) and a reversed light-dark cycle (light on 20h00-08h00) was maintained. The rats had free access to water. All measures were performed in the early morning to avoid diurnal complications.
Total Body Composition by DXA
A Hologic QDR 4500 device was used with an internal adapted collimator for small animal measurements (Hologic QDR Software for Windows XP version, Copyright © 1986-2002 Hologic Inc.). Rat whole body measurement required 240 seconds and provided global and regional body composition results. The scan field was adjustable to a maximum of 36 cm (L) * 18 cm (W). Spatial resolution was approximately 1 mm. A special designed small animal step phantom was scanned daily to calibrate the body composition results. Rats were anaesthetized before measurements. Anesthesia consisted on an intra-peritoneal injection of a solution of Acepromazine Vetranquil ® (0.5 ml.kg -1 of body weight) and Ketamine Imalgène ® (0.75 ml.kg -1 of body weight). After anesthesia, rats were positioned ventrally on a reference film to reproduce the position illustrated in figure 1 .
One week before sacrifice, body mass, fat mass and lean mass of all animals were assessed by DXA using specific small animal body composition software (line spacing: 0.254 mm, resolution: 0.127 mm). The coefficients of variation (CV's) were determined for these parameters from six repeated measurements with repositioning on eight animals. The CV's were 4.79%, and 0.19 % for fat mass, and body mass, respectively.
Dissection of rats
Rats were fasted for 12 hours before sacrifice. They were euthanized by decapitation under isoflurane anesthesia. Visceral fat mass was assessed by weighing the total perirenal and peri-epididymal adipose tissues. The weights of these two tissues were combined to form the ex-vivo Fat Mass .
Central fat mass (CFM) Area Determination
The area determination was performed in a group of 28 rats including 14 rats aged of 7 months and the fat group (n = 14, 11 months old). In order to select the more accurate method for assessing abdominal fat mass by DXA, four regions of interest (ROI) extracted from the whole body scan were tested (figures 2A, B, C and 2D ). These ROIs were defined after enlarging the total spine. They consisted of rectangular boxes extending vertically from one vertebral space to another and the lateral borders extending to the edge of the abdominal soft tissue. All trunk tissues, including intra abdominal fat and peripheral subcutaneous fat within these standardized height regions were selected for analysis. From these ROIs, we selected the region displaying the best correlation with ex-vivo Fat Mass. We called the fat mass measured within the chosen region "Central Fat Mass" (CFM) in order to distinguish it from the anatomic visceral fat. All investigations were performed by the same technician in order to avoid any bias linked to the inter-operator error.
Abdominal Circumference (AC)
The abdominal circumference was performed in a group of 29 rats including the group of lean rats (n = 15) and the group of fat rats (n = 14) Before each DXA measures, abdominal circumference (AC) was assessed on the largest zone of the rat abdomen using a plastic non extensible measuring tape (Rollfix, Hoechstmass ® , Germany) with an accuracy of 0.1 cm. Rats were placed in ventral position.
Statistical Analysis
The Gaussian distribution for each parameter was assessed by a Shapiro-Wilk test. In case of non-normal distribution, the data were log-transformed for analyses. In order to assess the potential relationships between the variables, we performed a correlation matrix between fat mass from each DXA ROI and ex-vivo Fat Mass. The CV for AC measures was determined following three analyses on 13 rats. The same operator repositioned the measuring tape three times. For assessing the CV of CFM, the ROI was repositioned two times on twelve scans by the same operator. Correlation matrix was generated for AC, ex-vivo Fat Mass and CFM. In case of significant correlation, linear regressions were used to analyze the relationships between parameters. Characteristics differences between lean and fat rats were tested using unpaired t-tests. A correlation matrix between AC, ex-vivo Fat Mass and CFM distinguishing fat and lean groups was compiled. A paired Student t-test was used to compare ex-vivo Fat Mass and CFM for the whole population. A 2 × 2 ANOVA allowed comparing the differences between lean and fat groups, and the ex-vivo Fat Mass and CFM values. Analysis was carried out using SPSS Advanced Statistics software version 17 (SPSS Inc., Chicago, IL) and data are presented as mean ± SD. | Results
The characteristics of the rats involved in the CFM determination are presented table 1 . Table 2 shows the correlation matrix between the four different ROIs and ex-vivo Fat Mass. All ROIs were significantly correlated with ex-vivo Fat Mass (p < 0.001) The ROI with the strongest correlation (r = 0.94, p < 0.001) was the ROI extending from L2 to L5 vertebrae (FM_L2_L5). This ROI was subsequently selected to represent the CFM. Using this ROI, the coefficient of variation calculated for CFM was 1.2%. Figure 3 displays the linear regression between CFM and ex-vivo Fat Mass. The explained variance for this relationship (R2 = 0.88, p < 0.0001), highlights the very strong association between these data.
The characteristics of lean, fat and the whole group rats are presented in table 3 . As expected, the characteristics of fat rats were significantly larger than those of lean rats. The CFM displayed higher values than ex-vivo Fat Mass in the whole group as well as in fat and lean groups (p < 0.005). The CV for AC was 2.6 %. The correlation matrix between AC, CFM and ex-vivo Fat Mass for the total population (fat and lean rats) is displayed table 4 . Within the whole population, AC correlated positively with both CFM (r = 0.90, p < 0.001) and ex-vivo Fat Mass (r = 0.82, p < 0.001). The slope of regression between AC and CFM (Figure 4A ) displayed a strong association (R2 = 0.82, p < 0.0001). Figure 4B shows a strong relationship between AC and ex-vivo Fat Mass (R2 = 0.68, p < 0.0001). The correlation matrix between AC, CFM, and ex-vivo Fat Mass for fat and lean rats is displayed table 5 . Correlation between CFM and ex-vivo Fat Mass remained significant in the two subgroups of rats (r = 0.60, p < 0.05 for lean rats and r = 0.734, p < 0.01 for fat rats) When considering the lean rats alone, the correlation between AC and both CFM (r = 0.51) and ex-vivo Fat Mass (r = 0.43) failed to reach significance. Nevertheless, for fat rats, AC correlated strongly with CFM (r = 0.93, p < 0.001) and the coefficient of correlation between AC and ex-vivo Fat Mass was weaker than for the whole population but remained significant (r = 0.583, p = 0.029). | Discussion
The main result of the present study demonstrated the ability to perform an indirect determination of the visceral fat mass in rats, using the technique of DXA. In addition, this study showed that fat mass could be accurately assessed by an anthropometric technique in fat rats.
The first objective was to establish the DXA region of interest (ROI) from the whole body scan in order to determine the fat mass demonstrating the best correlation with ex-vivo Fat Mass obtained by weighing. In human studies, landmarks based upon the intervertebral spaces are used to determine abdominal fat [ 17 , 18 ]. The inter-vertebral spaces are also easily identifiable landmarks in rats. Therefore, we decided to form ROIs as rectangular boxes extending vertically from one inter-vertebral space to another. The ROI determined in our study (FM_L2_L5) was quite similar to the ROI used in humans, extending from L2 to L4 vertebrae [ 18 ].
The ROI chosen includes only a small part of peri-epydidimal adipose tissue. Indeed, the localization of this adipose tissue is lower in the abdomen. Nevertheless, correlations between the different ROIs and only perirenal or peri-epididymal adipose tissues were tested but coefficients of correlation were lower than if the two adipose tissues were combined.
Different studies [ 7 - 12 ] have weighed visceral fat after dissection, but no real consensus has been established to identify which fat deposit among retroperitoneal, epididymal, peri-renal and mesenteric fat tissues is more representative of visceral fat mass. In rats, dissection of mesenteric and subcutaneous fat is difficult and some errors due to dissection might occur. In contrast, removing perirenal and peri-epididymal adipose appears to constitute a reliable technique. Therefore we combined these two adipose tissues to form ex-vivo Fat Mass.
The strong correlation observed between CFM assessed by DXA and ex-vivo Fat Mass demonstrates that CFM could be a useful predictor of visceral fat mass in rats. In humans, similar results have already been observed in obese or non obese women when comparing Intra-Abdominal Fat by Magnetic Resonance Imaging and Total Fat or CFM by DXA [ 18 , 21 ]. In the present study, the specific analyses of these relationships were conducted in subgroups of lean and fat rats. The correlations between CFM and ex-vivo Fat Mass remained positive in each group, although blunted in the lean rats.
However, as DXA measures of CFM, includes abdominal fat plus peripheral subcutaneous fat, DXA values were higher than ex-vivo Fat Mass because it includes only the total perirenal and peri-epididymal adipose tissues. Thus, the higher values found with the DXA method can be explained by subcutaneous fat which is excluded in ex-vivo Fat Mass. Indeed, DXA cannot distinguish between intra-abdominal and subcutaneous fat in humans and animals [ 22 ]. Despite a small over-estimation of DXA, the strong correlation observed between CFM and ex-vivo Fat Mass showed that DXA could be a useful predictor of ex-vivo visceral fat mass in rats.
The second objective was to validate the AC measure as an anthropometric method for assessing visceral fat mass. Within the whole population of rats, a strong significant correlation was observed between AC values and both CFM by DXA and ex-vivo Fat Mass. These results showed that AC could be a useful tool to predict visceral fat mass. When dividing the whole population into fat and lean rats, we observed the relationships between AC, CFM and ex-vivo Fat Mass were maintained for the fat rats but correlations with the lean rats disappeared. The small sample of subjects resulting from the division of the whole group may affect the statistical power of relationships between variables. Despite the sample effect, these results underline the fact that when animals are not characterized by obesity, AC is no longer relevant to assess abdominal fat. These results revealed that AC is a good indicator for predicting visceral fat mass in case of obesity. To date, no other study has addressed abdominal circumference in animals. Nevertheless, different studies conducted in human subjects concluded that waist circumference is the best anthropometric predictor of intra abdominal fat thickness [ 5 , 17 , 18 ]. Similarly, the present study has shown that the AC measure could be a useful anthropometric technique for assessing in-vivo abdominal fat mass storage in fat rats. | Conclusions
Both AC and CFM by DXA were good indicators for measuring abdominal fat mass storage. Coefficients of variation were found to be as low as 1.2% and 2.6% for CFM and AC, respectively. These values highlight the fact that these two measures can be used with precision and reliability. DXA and AC measures can be considered as non-invasive alternative techniques in determining visceral fat mass in rodents. These techniques have the potential to reduce the cost and number of rats required when using traditional sacrificial protocols. Moreover, abdominal circumference like other anthropometric measurements is inexpensive, quick and easy to perform. The most important outcome of the present study was that these two techniques can be used for in-vivo investigations, thereby allowing follow-up and longitudinal studies. | Background
Because abdominal obesity is predisposed to various metabolic disorders, it is of major importance to assess and track the changes with time of this specific fat mass. The main issue for clinicians or researchers is to use techniques for assessing abdominal fat deposition and its accumulation or changes over time, without sacrificing of experimental subjects. In the rat, techniques to investigate in-vivo visceral fat mass are lacking. The purpose of the study was to validate indirect Dual-energy X-ray Absorptiometry technique and abdominal circumference measurement as tools to predict visceral adipose tissue in rats.
Forty-three Wistar male rats from different body weight, fat mass and ages were included in the study. Visceral fat mass was assessed by weighing the total perirenal and peri-epididymal adipose tissues after dissection. Statistical methods were used to discriminate the best region of interest allowing the in-vivo measure of Central Fat Mass by DXA. Abdominal circumference was measured at the same time as the DXA scan.
Results
A region of interest including Central Fat Mass from the whole body DXA scan (extending from L2 to L5 vertebrae), correlated strongly with ex-vivo Fat Mass (r = 0.94, p < 0.001). Abdominal circumference correlated significantly with ex-vivo Fat Mass (r = 0.82, p < 0.001) and Central Fat Mass (0.90, p < 0.001) in the whole group of rats. When dividing the whole group into lean and fat rats, correlations remained significant between Central Fat Mass and ex-vivo Fat Mass but disappeared for the lean group between abdominal circumference and ex-vivo Fat Mass.
Conclusions
This study validates the Central Fat Mass determined by DXA as a non-sacrificial technique to assess visceral fat for in-vivo investigations in rats. The abdominal circumference measure appears useful in studying overweight or obese rats. These two techniques could be convenient tools in follow-up and longitudinal studies. | Competing interests
The authors have no conflict of interest.
Authors' contributions
MG has participated in the investigation as PhD student responsible of the protocol; intervention on animals, analysis and interpretation of data, drafting the manuscript.
LM has contributed to conception and design, intervention on animals, acquisition of data. ER has taken responsibilities in the housing protocol and measurements.
DC had responsibility of the design, analysis and interpretation of data and participated in drafting the manuscript, revising it critically for intellectual content and has given final approval of the version to be published.
All authors read and approved the final manuscript. | Acknowledgements
The authors wish to thank the "Heart and Arteries Foundation" (Fondation Coeur et Artères") for granting this project and its financial support. Our thanks to Professor Geraldine Naughton and Dr Douglas Whyte from the Australian Catholic University for help with manuscript preparation. | CC BY | no | 2022-01-12 15:21:37 | Lipids Health Dis. 2010 Dec 9; 9:140 | oa_package/be/41/PMC3014952.tar.gz |
PMC3014953 | 21122154 | Introduction
Neurorehabilitation based on the simulation hypothesis
Over the last decades, promising strategies in neurorehabilitation, e.g. following cerebral stroke [ 1 - 3 ], have been introduced based on the so-called simulation hypothesis [ 4 , 5 ]. The hypothesis suggests that the neural networks of a action-observation system located in the primary motor cortex (M1) and secondary motor areas, such as premotor cortex (PMC), supplementary motor area (SMA) and the parietal cortices, are not only activated during overt motor execution, but also during observation or imagery of the same motor action [ 6 ]. These networks are activated when individuals learn motor actions via execution (as in traditional motor learning), imitation, observation (as in observational learning) and motor imagery. Activation of these brain areas following observation or motor imagery may thereby facilitate subsequent movement execution by directly matching the observed or imagined action onto the internal simulation of that action [ 7 ]. It is therefore believed that this multi-sensory action-observation system enables individuals to (re)learn impaired motor functions through the activation of these internal action-related representations [ 8 ].
We have integrated this knowledge in a novel neurorehabilitative treatment system, based on motor and imagery performance in a virtual reality (VR) environment [ 9 ]: the system consists of a VR environment containing virtual representations of the patient's own arms and hands, which are displayed on a large screen and controlled by the patient wearing arm position trackers and data gloves. To activate the action-observation system, patients can train impaired upper limb function by playing interactive games in which they have to perform or imagine specific upper limb movements to interact with the VR environment. By adjustably mapping the movements of both the paretic and healthy limb onto the virtual limbs, the system offers individual training of upper limb motor function even in patients with little arm or hand movement ability.
Functional near-infrared spectroscopy
To monitor the VR system's effects on brain activation, we chose functional near-infrared spectroscopy (fNIRS). fNIRS is a non-invasive technique based on neurovascular coupling, which exploits the effect of metabolic activity due to neural processing on the oxygenation of cerebral tissue. Utilizing this tight coupling between neuronal activity and localized cerebral blood flow, fNIRS measures hemodynamic changes associated with cortical activation [ 10 ]. Optical NIR technology has been shown to be a reliable tool for functional neuroimaging of the human brain [ 11 ]. Although NIR technologies feature lower spatial resolution and are only able to image cortical tissue while not providing deeper tissue interrogation as compared to traditional neuroimaging methods such as functional magnetic resonance imaging (fMRI), they offer the advantage of portable systems and, in theory, insensitivity to electromagnetic fields and ferromagnetic materials. In this study a novel miniaturized wireless fNIRS instrument was used [ 12 ]. This wireless and portable NIRS technology does not require the subject's body or head to be restrained, and therefore represents an optimal brain monitoring tool for our purpose to record from subjects performing movements in a VR environment. It is thought that this wireless fNIRS technology could overcome some of the limitations inherent to traditional neuroimaging methods.
While the action-observation system described above has been widely investigated using traditional neuroimaging methods [ 13 - 15 ], so far there are only a few studies using NIRS based techniques [ 16 - 19 ]. Further studies have shown fNIRS to be a reliable tool to measure brain oxygenation related to motor imagery performance [ 20 - 27 ], confirming the well-known cortical areas located in primary and secondary motor areas.
The focus of the present study was to obtain evidence of the VR system's efficacy in neurorehabilitation by evaluating its effects on brain activation. In particular, we aimed 1) to provide evidence, that our VR system is able to elicit the action-observation system and 2) to draw conclusions for the system's further application in neurorehabilitative treatment. We hypothesized that the observation, imagery and imitation of a hand motor task in an interactive VR environment enhances the related cortical oxygenation changes of the action-observation system as measured by fNIRS. The long-term aim is to implement the data obtained in the development of a VR-fNIRS based brain computer interfaces (BCIs). Such a VR-fNIRS based BCI could enhance patients' motivation by providing real-time neurofeedback thereby allowing therapists to record pre-post treatment progress assessing training-induced oxygenation changes. | Materials and methods
Subjects
Right-handed subjects were recruited via advertisements at the University of Zurich and ETH Zurich. Exclusion criteria were any history of visual, neurological or psychiatric disorders or any current medication. All subjects gave informed consent. All subjects had normal or corrected-to-normal vision. The study was approved by the ethics committee of the Canton of Zurich and was in accordance with the latest version of the Helsinki declaration.
Experimental protocol
Prior to recording, subjects completed the Edinburgh Handedness Inventory (EHI) [ 28 ] assessing hand dominance to exclude left-handed subjects. The right-handed subjects were assigned to one of two groups: either to the 'unilateral' group (N = 15) or to the 'bilateral' group (N = 8). Each subject in either group participated in one experimental session. However, bilateral wireless NIRS measurements are more demanding with respect to the instrumentation: two devices are needed instead of one and they must be time-synchronized.
All experiments were conducted in a quiet room. Subjects sat in front of a custom made VR table-system with a computer screen (94 cm diagonal) to display the VR environment [ 9 ]. The subjects were asked to place their hands on the table with the palms facing downwards, and faced the monitor at a distance of approximately 70 cm. The image on the monitor showed a virtual arm in the same orientation and relative position as the real arms, resting on a flat surface representing the table. The close correspondence between the virtual and real arms in terms of position and relative (first-person) orientation was designed to optimally stimulate the patient to imagine the virtual arms as their own during the experimental session.
Unilateral group
In the subject group 'unilateral', fNIRS was recorded over the left hemisphere while the subject performed the VR tasks under four conditions:
▪ 'Observation (O)': subjects passively watched a VR video which displayed a right upper limb with the hand repeatedly grasping an incoming ball (13 actions, approx. 0.86 Hz) (Figure 1 ).
▪ 'Observation & motor imagery (O&MI)': same as condition O, except that subjects were asked to imagine that the virtual arm was their own.
▪ 'Motor imagery (MI)': same as condition O&MI, but without visual input - subjects had to imagine performing the action.
▪ 'Imitation (IM)': subjects imitated the hand movements performed in the VR task by the virtual arm while watching the VR video.
The session began with a practice trial (approx. 5 min) to allow subjects to become familiar with the tasks. After the practice trial, all subjects first performed condition O followed by a randomized presentation of conditions O&MI, MI and IM (Easy Randomizer, Version 4.1. [ 29 ]). Subjects were reminded to perform the executed or imagined movements with the same frequency as shown in the video (approx. 0.86 Hz). Each condition lasted 530 s (8 min 50 s) consisting of 10 trials each comprising an initial rest period (30 s), followed by 10 stimulation periods (20 s) alternated with rest periods (30 s) (Figure 2 ). The total number of trials per subject was 40; the total duration of the experiment was approx. 35 min per subject. We chose these irregular periodic alternations of 20 s stimulation and 30 s rest periods to avoid the induction of synchronization of the sequence of the motor stimulation/rest periods in the motor stimulation protocol with systemic rhythms such as heartbeat, respiration and heart rate fluctuations.
Bilateral group
The subject group 'bilateral' had the same VR task as the group 'unilateral', but was recorded bilaterally. This group was included to test for a lateralized distribution of oxygenation patterns for the ipsi- and contralateral side, as seen in related studies [ 30 - 33 ]. We hypothesized that, on the one side, the hemisphere contralateral to the hand performing the task would show larger [O 2 Hb] changes as compared to the ipsilateral hemisphere. The detection of larger [O 2 Hb] changes over the hemisphere contralateral would provide evidence that we were indeed recording from the correct position, i.e. covering motor-related cortical areas. Conditions O and MI were chosen as we assumed that these conditions would elicit the smallest oxygenation changes, both unilaterally and bilaterally. Therefore conditions O&MI and MI were dropped as we assumed that these conditions would follow a similar pattern to the other conditions.
▪ 'Observation right (O_R)': Same as condition O in the unilateral group.
▪ 'Observation left (O_L)': Same as condition O_R, except that a left hand was shown in the VR video.
▪ 'Imitation right (IM_R)': Same as condition IM in the unilateral group.
▪ 'Imitation left (IM_L)': Same as condition IM_R, except that a left hand was shown in the VR video and subjects were asked to imitate the movement with their left hand.
After the practice trial, all subjects performed condition O_R or O_L first, which was randomly assigned, followed by condition IM_R or IM_L, which was also randomized (Easy Randomizer, Version 4.1. by [ 29 ]). The procedure and timing were the same for both the 'unilateral' and the 'bilateral' groups.
NIRS instrumentation
The novel miniaturized continuous wave wireless fNIRS sensor has been previously described in detail [ 12 ]. The optical and electronic components are mounted onto a four-layer rigid-flexible printed circuit board (PCB) which, in combination with a highly flexible casing made of medical grade silicone, enables the sensor to be aligned to curved body surfaces such as the head. The size of the device is 92 × 40 × 22 mm and weighs 40 g. The optical system comprises four light sources at two different wavelengths (760 nm and 870 nm) and four detectors (PIN silicon photodiodes). The distance between light sources and detectors is 25 mm, four light source-detector pairs are linearly arranged every 12.5 mm and thus cover an area of 37.5 × 25 mm (Figure 3 ). Each light source consists of two pairs of serially connected light emitting diodes (LED) is driven using current control and is time multiplexed with an on-time of 120 μs per sample and a forward voltage of 4 V per diode. Although LEDs have a broader emission spectrum than lasers, they have several advantages: they can be applied directly on the body surface without need for lenses or fibers and they are inexpensive. Furthermore, they are harmless for the eye, which is an important advantage in a clinical environment. The power is provided by a rechargeable battery, which allows continuous data acquisition for 180 minutes at full light emission power. The light intensity is sampled at 100 Hz and the resulting data are transmitted wirelessly to the host computer by Bluetooth. The operating range of the sensor is about 5 m. The wireless sensor has been found to be capable of detecting both localized changes [O 2 Hb] and [HHb] in the adult brain and oxygenation changes of muscular tissue [ 12 , 34 ].
For fNIRS recording, the sensor(s) was(were) placed either contralateral (unilateral group) or bilaterally (bilateral group) on the subject's head presumably covering F3 according to the international 10-20 system [ 35 ]. With the compact sensor of 37.5 mm length and 25 mm width, we assumed that we covered secondary motor areas [ 36 ]. Hairs under the sensor(s) were carefully brushed away before fixation; shaving was not required. The sensor was fixed on the subject's head using medical-grade, disposable, self-adhesive bandages (Derma Plast CoFix 40 mm, IVF Hartmann, Neuhausen, Switzerland).
For final data processing, by measuring intensity of NIR light after its transmission trough tissue, it is possible to determine changes over time in the concentration of oxy-hemoglobin (O 2 Hb) and deoxy-hemoglobin (HHb), which represent the dominant light absorbers for living tissue in the NIR spectral band. By applying the modified Beer-Lambert law (MBBL), the concentration for O 2 Hb and HHb ([O 2 Hb], [HHb]) were computed from the measured absorption changes [ 37 , 38 ].
A MATLAB ® (Version 2008a) program was applied to pre-process the raw light intensity values and to compute [O 2 Hb] and [HHb] changes. The measurement files that were acquired during the fNIRS experiment contain the intensity signals of the NIR light, sampled at 100 Hz for all combinations of light-sources, wavelengths and detectors, as well as the intensity of the ambient light. The program subtracts the ambient light intensities from the NIRS measurement values before low-pass filtering (7 th order Chebyshew with 20 dB attenuation at 5 Hz) and decimating the signals to a sampling rate of 10 Hz. Consecutively, the MBLL is used to compute the changes of [O 2 Hb] and [HHb] applying differential path lengths factors (DPF) of 6.75 for the 760 nm and 6.50 for the 870 nm light-sources [ 39 ]. The [O 2 Hb] and [HHb] signals acquired with the wireless NIRS signal characteristically drift slightly over time, which can mostly be attributed to thermal effects. Therefore, data was recorded only two minutes after starting the fNIRS sensor, allowing the setup to reach thermal equilibrium. The remaining signal drift [ 12 ] was highly linear as assessed by visual inspection and thus their linear least squares approximation was subtracted from [O 2 Hb] and [HHb] for drift elimination. | Results
Behavioral data
23 healthy subjects were included in the analysis (15 unilateral group, 8 bilateral group, 9 males, mean age 26 years, range 22 - 33 years). Five subjects (2 in unilateral group; 3 in bilateral group) were excluded from analysis due to a missing signal in the IM condition. All subjects were right-handed according to the EHI with a mean LQ of 81.9 (range 73 - 100) and a mean deciles level of 6.1 (range 3 - 10).
fNIRS measurements
Unilateral group
The mean Δ[O 2 Hb] (Table 1 ) was largest in the IM condition, followed by MI, O, and O&MI. Mean Δ[HHb] was largest in condition MI, followed by IM, O&MI, and O. The data showed a higher degree of inter-subject variability observed for Δ[O 2 Hb] compared to Δ[HHb] as calculated by the standard deviation (SD) of the oxygenation changes.
Intra-condition analysis of the median changes between [O 2 Hb] rest and [O 2 Hb] stim using a paired t-test (Table 1 ) revealed statistical significance in the MI (p = 0.049) and IM (p < 0.001) conditions. No significant differences were detected between [HHb] rest and [HHb] stim . Figure 4 shows an example of a sample subject of the oxygenation changes from rest to stimulation period in each of the four conditions.
Inter-condition analysis of the mean amplitude changes of Δ[O 2 Hb] and Δ[HHb] between rest and stimulation periods between the four conditions using one-way ANOVA (Table 1 , Figure 5 ) revealed neither a main effect of condition, nor statistical significant between the four conditions.
Bilateral group
In this group a smaller number of subjects was included, although sufficient to reach statistical significance.
In the left hemisphere, the mean Δ[O 2 Hb] (Table 2 ) were largest in condition IM_L, followed by IM_right, O_R, and O_L. Mean Δ[HHb] were largest in condition IM_L, followed by IM_R, O_L and O_R.
On the right hemisphere, mean Δ[O 2 Hb] were largest in condition IM_L, followed by IM_R, O_R, and O_L. Mean Δ[HHb] were largest in condition IM_L, followed by IM_R, O_L and O_R. As also seen in the unilateral group a relatively high inter-subject variability was observed, as documented by the standard deviation (SD).
Intra-condition analysis (left hemisphere (LH), right hemisphere (RH)) of the median change between [O 2 Hb] rest and [O 2 Hb] stim using the paired t-test (Table 2 ) revealed statistical significant differences in conditions O_R (LH p = 0.016, RH p = 0.006), O_L (LH p = 0.046, RH p = 0.025), IM_R (LH p = 0.003, RH p < 0.001) and IM_L (LH p < 0.001, RH p = 0.001). Between [HHb] rest and [HHb] stim statistical significance was observed in condition IM_L (LH p = 0.040, RH p < 0.001).
Inter-condition analysis of the mean amplitude changes of Δ[O 2 Hb] and Δ[HHb] between the four conditions using one-way ANOVA (Table 2 , Figure 6 ) revealed a main effect of condition for [O 2 Hb] (LH p = 0.028, RH p < 0.001) and for [HHb] (RH p < 0.001). Statistical significance was found for Δ[O 2 Hb] between-conditions O_R and IM_L (RH p < 0.001), O_L and IM_L (RH p = < 0.001) and IM_R and IM_L (RH p < 0.001); analog for Δ[HHb] between-conditions O_R and IM_L (RH p < 0.001), O_L and IM_L (RH p = < 0.001) and IM_R and IM_L (RH p < 0.001).
In the following discussion we concentrate on the observed [O 2 Hb] changes, since this parameter shows the relevant significant oxygenation changes, whereas [HHb] did show overall significant levels. This is supported by previous fNIRS work suggesting that interpretations about task-relevant activation increases are usually attributed to the prominent increases in [O 2 Hb] [ 40 ], whereas [HHb] is often not reported. | Discussion
Virtual reality based neurorehabilitation
Recent experimental evidence suggests that rapid advancement of VR technologies has great potential for the development of novel strategies for sensory-motor training in neurorehabilitation [ 41 ]. The combination with our wireless and portable fNIRS brain monitoring technique [ 12 ] is particularly appealing from a rehabilitation perspective as it allows therapists and patients unconstraint monitoring while testing and training motor performance [ 21 , 42 ].
In this study we provide evidence for the efficacy of our new VR neurorehabilitation system [ 9 ] by evaluating its effects on brain activation. In particular, we show that our VR system is able to elicit the action-observation system as described by the simulation hypothesis. Based on these results we aim in the long-term to develop a VR-fNIRS based BCI that providing the possibility of real-time neurofeedback combined with an assessment of training-induced cortical oxygenation changes.
Observation, imagery and imitation
From the comparisons between stimulation and rest periods, our results confirm the simulation hypothesis in accordance with well-known findings in fMRI and EEG [ 3 , 14 , 15 , 43 , 44 ] and previous fNIRS studies [ 21 - 25 , 45 ] that have shown that oxygenation changes can be found within the same secondary motor areas during observation, motor imagery and overt motor execution (unilateral and bilateral group, Figure 5 and 6 ). Although not all of the observed changes reached statistical significance, our results revealed that averaged Δ[O 2 Hb] during observation and motor imagery were approximately one-third lower compared to the imitation task. This result is in line with the previous studies mentioned above where both imagery and observation have been reported to elicit consistently lower oxygenation changes.
Inter-subject variability
We observed a high inter-subject variability in Δ[O 2 Hb] in both our samples. General reasons for variability between individuals may be effects of anatomical variance such as thickness of the skull or cerebrospinal fluid layers [ 46 , 47 ]. Another contributing factor might be that our subjects had no prior specific experience in the tasks presented. They were not specifically trained to perform the tasks prior to the experiment (but only received a short practice trial), yet this has been done in a previous fNIRS controlled BCI [ 24 ]. Therefore, in our untrained subjects, inter-subject variability in the hemodynamic response patterns might have been higher than it would have been after substantial pre-experimental training. The question of the extent to which a person is able to generate a mental representation of movements is even more relevant in the assessment of individuals following brain injury. Lesions involving specific cortical areas may impair certain imagery abilities [ 48 ], such as overall slowing of imagery processes resulting in modified temporal characteristics of motor imagery [ 49 , 50 ].
Bilateral oxygenation
As observed in previous studies, brain activation in response to executed or imagined actions can differ depending on the hemisphere recorded [ 51 - 53 ]. In general, unimanual tasks show hemispheric asymmetry with predominant activation of the contralateral hemisphere controlling the moving hand, as assessed by fMRI and PET [ 30 - 33 ]. Additionally, ipsilateral activation is both found in M1 and shifted laterally, ventrally, and anteriorly towards PMC for unimanual tasks with respect to that observed during contralateral hand movements [ 54 - 60 ]. Accordingly, we observed ipsi- and contralateral oxygenation changes, both during observation and imitation.
The difference observed between the unilateral and the bilateral group is concerned about the aspect of handedness. Interestingly, we found that performance during the condition IM_L (imitation with the subject's left non-dominant hand) revealed larger Δ[O 2 Hb] in both hemispheres as compared to IM_R (imitation with the subject's right dominant hand) (Figure 6 ). Further, the Δ[O 2 Hb] in the right hemisphere during movement of the subjects' left hand (i.e. the non-dominant, contralateral hand) is considerably larger than that in the left hemisphere during ipsilateral movement. Additionally, in the left hemisphere during ipsilateral movement (non-dominant hand) the Δ[O 2 Hb] was larger than that observed during contralateral movement (dominant hand; according to the unilateral group). Figure 5 and 6 reflect these findings showing the observed inter-condition differences in the right hemisphere including lower level Δ[O 2 Hb] amplitude during observation as compared to imitation (Figure 6 ). These findings might be explained by the hand dominance of our right-handed sample. Previous fMRI studies described that non-dominant hand movements appear to require more cortical activity and therefore may result in greater recruitment of ipsi- and contralateral cortical motor areas [ 61 ], perhaps because they are less 'automatic'. It has been further observed that this ipsilateral activation was most pronounced in pre-central areas (presumably corresponding to secondary motor areas) during both dominant and non-dominant performance [ 62 ]. However, further fNIRS studies are needed to confirm whether or not our findings of larger Δ[O 2 Hb] during non-dominant performance are in fact caused by the right-handedness of our sample.
Neurorehabilitative potential of combined VR NIRS applications
Taken together the findings of the uni- and bilateral groups, the results show that our VR system can activate the action-observation system as described by the simulation hypothesis. In particular, 1) the study provides evidence that fNIRS recording does not impede interaction with the VR environment This point is an important precondition for further development of combined VR-fNIRS based applications for use in neurorehabilitation. It increases usability in that it requires a short time to fit fNIRS sensor important for therapy. Further, the results revealed two factors that need to be taken into account when dealing with fNIRS signals aimed to provide a basis for neural interfaces: 2) The inter-subject variability is obvious at the group level and will be even more prominent at he single subject level. The reasons for inter-subject variability, i.e. individual experience in motor imagery performance, physiological and anatomical differences, require further assessment. 3) The combined factors of recording side, i.e. uni- or bilateral hemispheres, as well as hand side, i.e. left or right hand used during motor or imagery tasks, need to be taken into account. Our findings may reflect an aspect of handedness in right-handed subjects who may require more cortical activity when using the non-dominant hand. Future studies may include both left-handers and right-handers. Considering these factors may contribute to differentiation of individual oxygenation pattern and permit classification of activation tasks used for neurofeedback or BCI applications.
Study limitations
Although the present study revealed interesting results concerning the potential of the new wireless NIRS system, it was subject to some known limitations. We did not record an electromyogram (EMG) in order to exclude the presence of muscular activation during observation and motor imagery. It could be claimed that weak motor activity might have been present during the imagery tasks. However, previous neuroimaging studies suggested that brain signals during imagery of hand motor tasks are not correlated with EMG activity [ 63 ]. Another possible limitation is that we referenced the positioning of the NIRS device according to the 10-20 system [ 35 ]. However, this positioning may be inaccurate due to inter-subject variability in anatomical head size and shape, and the location on underlying (pre-)motor areas. The location of NIRS recording can therefore generally only be assumed to have correctly covered the preferred areas, i.e. in our case secondary motor areas. | Conclusion
This study shows that our combined VR-fNIRS based neurorehabilitation system is able to activate the action-observation system as described by the simulation hypothesis during performance of observation, motor imagery and imitation of hand actions elicited by a VR environment. Further, in accordance with previous studies, the findings of this study revealed that both inter-subject variability as well as handedness needs to be taken into account when recording in untrained subjects. In the long term, these findings are of relevance for the VR-fNIRS instrument's potential in neurofeedback applications.
LH conceived of the study, conducted the fNIRS recordings, carried out the statistical analysis, and drafted the manuscript. TM and FS carried out the MATLAB ® pre-processing. KE and DK participated in the design of the study. MW participated in the design and coordination of the study. All authors read and approved the final manuscript. | Background
Several neurorehabilitation strategies have been introduced over the last decade based on the so-called simulation hypothesis. This hypothesis states that a neural network located in primary and secondary motor areas is activated not only during overt motor execution, but also during observation or imagery of the same motor action. Based on this hypothesis, we investigated the combination of a virtual reality (VR) based neurorehabilitation system together with a wireless functional near infrared spectroscopy (fNIRS) instrument. This combination is particularly appealing from a rehabilitation perspective as it may allow minimally constrained monitoring during neurorehabilitative training.
Methods
fNIRS was applied over F3 of healthy subjects during task performance in a virtual reality (VR) environment: 1) 'unilateral' group (N = 15), contralateral recording during observation , motor imagery , observation & motor imagery , and imitation of a grasping task performed by a virtual limb (first-person perspective view) using the right hand; 2) 'bilateral' group (N = 8), bilateral recording during observation and imitation of the same task using the right and left hand alternately.
Results
In the unilateral group, significant within-condition oxy-hemoglobin concentration Δ[O 2 Hb] changes (mean ± SD μmol/l) were found for motor imagery (0.0868 ± 0.5201 μmol/l) and imitation (0.1715 ± 0.4567 μmol/l). In addition, the bilateral group showed a significant within-condition Δ[O 2 Hb] change for observation (0.0924 ± 0.3369 μmol/l) as well as between-conditions with lower Δ[O 2 Hb] amplitudes during observation compared to imitation , especially in the ipsilateral hemisphere (p < 0.001). Further, in the bilateral group, imitation using the non-dominant (left) hand resulted in larger Δ[O 2 Hb] changes in both the ipsi- and contralateral hemispheres as compared to using the dominant (right) hand.
Conclusions
This study shows that our combined VR-fNIRS based neurorehabilitation system can activate the action-observation system as described by the simulation hypothesis during performance of observation, motor imagery and imitation of hand actions elicited by a VR environment. Further, in accordance with previous studies, the findings of this study revealed that both inter-subject variability and handedness need to be taken into account when recording in untrained subjects. These findings are of relevance for demonstrating the potential of the VR-fNIRS instrument in neurofeedback applications. | Data Analysis
Descriptive analysis was calculated for all median signal amplitudes (μmol/l ± SD). Each source-detector combination (channel) and each condition was averaged to attempt to provide a detectable signal. The criterion for a detectable signal was the relative value between stimulation and baseline, i.e. increase in [O 2 Hb] and decrease in [HHb]. At this point those channels that did not show task related oxygenation changes were excluded from further analysis, since it was assumed that those channels did not cover the activated cerebral region at all. For the same reason, subjects that did not display statistically significant changes of the [O 2 Hb] median for the condition IM (control condition) were excluded as well.
All data were positively tested for Gaussian distribution using the Kolmogorov-Smirnov test. Consecutively, dependant variables for further statistical analysis were derived from the non-excluded [O 2 Hb] and [HHb] datasets. Specifically, the median of the last 10 s of the stimulation periods ([HHb] stim , [O 2 Hb] stim , stimulation amplitudes) and the median of the last 10 s of the rest periods ([HHb] rest , [O 2 Hb] rest , baselines) were tested in the analysis. The median was chosen instead of the mean as it is more robust to outliners that may have statistically unbalanced the analysis in our relatively small subject sample. The statistical significance of the intra-condition differences between ([HHb] rest , [O 2 Hb] rest ) and ([HHb] stim , [O 2 Hb] stim ), later referred to as Δ[HHb] and Δ[O 2 Hb], was analyzed using the paired t-test.
The statistical significance of inter-conditional differences of [O 2 Hb] stim and [HHb] stim as well as for [HHb] rest and [O 2 Hb] rest were first assessed across all conditions. Then, if a significant difference was found, it was followed by a pair wise comparisons for all possible condition pairs using one-way ANOVA; the alpha-value for significance was set to ≤ 0.05 and the Bonferroni correction was applied to eliminate the problem of multiple comparisons.
Declaration of competing interests
The authors declare that they have no competing interests. | Acknowledgements
The authors thank all participants for their assistance in carrying out this research and the Swiss Society for Neuroscience (SSN), the International Brain Research Organization (IBRO), the Swiss National Research Foundation and the Stiftung für wissenschaftliche Forschung, University of Zurich, for providing the funding. | CC BY | no | 2022-01-12 15:21:37 | J Neuroeng Rehabil. 2010 Dec 2; 7:57 | oa_package/7e/37/PMC3014953.tar.gz |
PMC3014954 | 21159182 | Introduction
The use of complementary and alternative medicine (CAM) by the pediatric population is increasing [ 1 ]. A recent study by Vlieger et al found that perceived adverse effects of allopathic medication, low effect of conventional treatment, school absenteeism and age less than 11 years were predictors of use of CAM care for children in the Netherlands [ 2 ]. It is estimated that 11.8% of children in the USA use CAM therapies [ 3 ]. The Center for Disease Control in the USA reported that manual therapy was the most common type of practitioner-based CAM therapy chosen for children and that musculoskeletal conditions were the most common types of conditions for which treatment was sought [ 4 ]. A 2007 Canadian study corroborated these findings, stating that musculoskeletal care was the most common type of CAM treatment chosen by parents for their children. Personal experience, lack of appropriate treatments available from conventional medicine or referral from a physician were the reasons given for seeking alternative care [ 5 ].
The aim of this study was to investigate the pediatric patients who attended a university-affiliated chiropractic teaching clinic on the south coast of England between 2006 and 2010. The goals were to determine the frequency of presentation in each age group, reasons for attendance, referral patterns and usage of other types of health care prior to presentation and demographic features. | Methods
The data presented in this report were obtained from a computerised system maintained by clinicians overseeing the care of pediatric patients up to 16 years of age between January 2006 and January 2010. This was a cross-sectional study of baseline demographic data of pediatric patients presenting to the Anglo-European College of Chiropractic (AECC) outpatient clinic.
Descriptive statistical analysis was performed using Microsoft Excel. All data were held confidentially. Parents consented that the data could be used for research purposes. Ethical approval was granted by the AECC Projects Panel.
Complaints were categorised as musculoskeletal if they pertained to the axial or appendicular skeleton or resulted in a dysfunction or discomfort of movement or posture. | Results
This data system included 2,645 pediatric patients (0-15 years of age). Of these, 2,303 (87%) were under the age of five and 342 (13%) were between the ages of 5 and 15 (Figure 1 ). Fifty-seven percent were male and 43% were female. These patients were categorised according to complaint on presentation (Figure 2 ). The complaints of all children over five years were categorized as musculoskeletal. In all age groups, just over a third (34.7%) presented with musculoskeletal problems, 29.6% presented with excess crying (previously known as infant colic) and 15.7% with feeding disorders. All children had previously presented to at least one medical practitioner for the same condition and some had presented to multiple healthcare practitioners (Figure 3 ). The younger the child, the more common the referral with 83% of infants under 12 weeks of age being sent by a medical practitioner, 39% between 3 and 12 months of age and a 4-5% rate of referral in age groups over one year. Figure 4 shows referral rates relative to age group. Over-all, 20.5% of the clinic patients were aged between two days and 16 years. | Discussion
Boys were more commonly presented than girls. This may be due to the prevalence of musculoskeletal health problems which have previously been shown to be more common in boys [ 6 ]. The patient proportions in that study (57.4% male) versus girls (42.6%) [ 6 ] were virtually the same as in our clinic (57% male and 43% female). At birth, boys are often larger than girls and intra-uterine constraint may result in biomechanical imbalance or asymmetries in their cranium, spine or extremities [ 7 ].
It is not surprising that musculoskeletal problems were the most common presentation of the pediatric patient to our clinic. First, chiropractors are known to specialize in the musculoskeletal system and second, musculoskeletal pain affects a significant number of children [ 6 ]. Further, these conditions carry a significant economic burden due to time lost at school, lost time from work for parents and diagnostic procedures and referrals and consultation with multiple practitioners [ 6 ]. What may be surprising is that parents have heightened awareness to recognize musculoskeletal pain in the very youngest children, particularly neonates. A common complaint of early infancy is that the baby "refuses" to lie on his/her back (shows pain behaviours when lying supine) [ 8 ]. The "back to sleep" program is required for cot death prevention [ 9 ]. Manual therapy might be considered useful to treat the infant to help the infant to sleep comfortably [ 10 ]. Manual therapy was the most commonly chosen therapy by parents for their child in a USA study [ 4 ]. In a recent UK survey, clinicians (pediatricians, orthopedists, primary care (both new trainees and experienced) and emergency medicine doctors were asked how confident they felt dealing with pediatric musculoskeletal (pMSK) problems. Seventy-four percent had "no" or "some confidence" [ 11 ]. It is possible that clinicians with little confidence to treat pMSK problems may refer these cases to manual therapists. In a London study of general practitioners, 83% had referred for CAM therapies or influenced such referral [ 12 ], although this study was not specific to pMSK.
Referrals to this chiropractic teaching clinic from medical professionals were common. Children under three months of age had the highest (83%) referral rates. It is not surprising that medical professionals referred pediatric patients to this clinic for musculoskeletal conditions such as torticollis and other postural preferences that cause difficulty and perhaps even pain when the infant is moved out of their antalgic posture. However, crying and feeding problems were also commonly referred. These early "quality of life" problems such as excess crying (previously known as infant colic) and feeding problems as well as sleep dysomnias may be considered to be amenable to biomechanical attention [ 13 ]. However, the efficacy of chiropractic care for these conditions has not yet been proven or disproven [ 14 ]. There are some suggestions that feeding problems in the neonate may be biomechanical in nature [ 15 ] and one study suggests that multidisciplinary care which included chiropractic may be helpful [ 16 ]. There may also be benefit to ruling out a simple musculoskeletal lesion which could be corrected quickly with little risk before the child undergoes more invasive testing or procedures.
The population most commonly presented by their parents for care in this study were young, under 12 weeks of age. These results are similar to a Danish study that found the most common pediatric patients to present to chiropractors were under four months of age [ 17 ]. | Conclusion
In this chiropractic clinic, pediatric patients most commonly presented for excessive crying in the early months and for musculoskeletal complaints at all ages. Parents often brought their child to this clinic on the recommendation of medical professionals, particularly in the infant population. All children had consulted a medical practitioner prior to their presentation to this clinic. Further research is required to ascertain therapeutic benefit, cost/benefit and rates of satisfaction for this type of treatment. | Background
Considering the increasing use of alternative therapies for children, it is appropriate to determine the demographic profile of pediatric patients entering a chiropractic clinic.
Methods
Collection of demographic data including age, gender, condition at presentation, previous clinicians consulted and medical referral rates of pediatric patients presenting to a chiropractic teaching clinic between 2006 and 2010.
Results
Over-all, 20.5% of patients were aged between two days and 15 years and classified as pediatric patients. The most common presenting complaint was musculoskeletal (35%). Excess crying (30%) was the most common complaint in the largest presenting age group which was under 12 weeks of age (62.3%). All children had previously presented for medical care for the same condition. Most (83%) of the infant patients under 12 weeks of age were referred for care by a medical practitioner.
Conclusion
Parents commonly presented their child for care at this chiropractic clinic with a recommendation from a medical practitioner. The most common complaints were musculoskeletal and excessive crying conditions and the most prevalent age group was under 12 weeks of age. | Competing interests
The authors declare that they have no competing interests. | Acknowledgements
I wish to thank Professor Jennifer Bolton and Steve Goode and Gary Fitzgerald for assistance in data collection and Dave Mitchell for data preparation. | CC BY | no | 2022-01-12 15:21:37 | Chiropr Osteopat. 2010 Dec 15; 18:33 | oa_package/e5/b5/PMC3014954.tar.gz |
PMC3014955 | 21126347 | Background
Clinical and research electives abroad offer medical students many unique experiences. Shouldering responsibility in a different health care system and working with underserved patients broadens the personal and medical horizon. This may even influence future career choice as international medical experience is associated with an increase in the choice for a primary care specialty [ 1 ]. A number of studies have surveyed the health risks facing students during an elective abroad and the pre-travel advice [ 2 - 9 ]. Particular regard has been given to the risk of bloodborne viral infection. For example, it is worrying that 75% of students fail to report exposures to potentially infectious body fluids [ 4 ].
Each year approximately 300 students enroll in the medical program at Leiden University Medical Center (LUMC) in The Netherlands. Approximately half of them perform one or more electives abroad. Unlike other medical schools, ours allows students to go on electives in countries where infection with Human Immunodeficiency Virus (HIV) is endemic and does not restrict senior students who have completed the fourth college year from performing surgical or obstetric practice in such countries. To receive study credits it is mandatory that the students obtain permission from the student registrar before departure. If study credits are obtained, it is also mandatory for students to seek a Dutch supervisor who assesses the quality of the planned elective and who judges the students' written report at the end of the elective. The registrar's office provides general information on preparation for an elective abroad and advises students to obtain pre-travel counseling and immunization. Although the university occupational health department provides such counseling and immunizations, the students are free to visit any other travel clinic including the LUMC in-hospital travel clinic or their general practitioner. As part of the travel advice, and depending on the destination and intended elective, the health department or travel clinic may refer the student to an infectious disease consultant for counseling on the need of carrying post-exposure prophylaxis for HIV (PEP) with them and on its use. Upon return home, no standard post-travel counseling is offered.
To improve pre-and post-travel care, we performed a questionnaire study of students returning from an elective abroad. We assessed the health risks and the quality and comprehensiveness of pre-and post-travel care. This led to improvements that are described in the discussion. | Methods
All medical students who had performed an elective abroad between July 2006 and December 2008, who had visited countries where hepatitis A is endemic, and who had notified the student registrar to obtain study credits, were sent an informative email asking them to complete a web-based questionnaire. This study was designed in 2008. Students who had returned home prior to December 2007 were sent an email in May 2008. Students who returned between December 2007 and November 2008, which is during the conduct of this study, were sent an email in November 2008. Non-responders were sent a reminder two weeks after the first email. The questionnaire was designed to seek information on pre-travel preparation including vaccinations, on characteristics of the elective, on health risks (in particular the exposure to and protection against bloodborne viruses), on adherence to advice regarding anti-malarial measures and on illness while abroad and upon returning home. In addition the rate of routine screening for tuberculosis using one pre-and one post travel Tuberculin Skin Test (TST) was surveyed. We also surveyed the rate and result of screening for methicillin-resistant Staphylococcus aureus (MRSA) as students visiting foreign hospitals may import MRSA to Dutch hospitals. Finally we surveyed the rate and result of screening for schistosomiasis. The questionnaire was piloted among acquaintances and among staff of the department of Clinical Epidemiology at the Leiden University Medical Center. The protocol of this study (protocol 08/37B) was studied by the Medical Ethics Committee of Leiden University Medical Center in The Netherlands. The Medical Ethics Committee did not object to the conduct of this study. | Results
The mean number of days between having completed the elective and completing the questionnaire was 235 days (interquartile range 121 to 325 days, range 2 to 638 days). The characteristics of the responders and of the electives are described in Table 1 ; 242 students were sent a questionnaire. Of the 180 (74%) who completed it the majority (78%) was female; 77% had planned a holiday before or after the elective, and the mean duration of the time spent abroad was 74 days (median 69 days, range 10 to 224 days). The majority went for the purpose of a clinical (47%) or pre-clinical elective (16%) as opposed to research or volunteer work (37%). Surinam was visited by 31%, making it the most popular destination. Obstetrics and gynecology (42%) was the most popular rotation. Before departure 90% consulted a center specialized in travel medicine; 4% sought advice from their general practitioner and 6% did not obtain advice from a qualified source.
Risk of infection with bloodborne viruses
All 180 students had been vaccinated against hepatitis B. The vaccine response is checked by the university occupational health department. For privacy reasons we did not have access to the response data; 120 students (67%) performed at least one type of procedure that is associated with an increased risk of exposure to bloodborne viral infection (i.e. surgical or obstetric practice, suturing, phlebotomy) (Table 2 ). In general, before completing the fourth college year, students have not yet been trained to perform many of these procedures. Therefore it is surprising that of the 58 junior students, 18 (32%) did take part in such activities. Procedures associated with an increased risk of exposure to bloodborne viral infection were also performed in countries with high HIV prevalence rates (Table 3 ). Some students received medical care while on elective which increases the risk of exposure to bloodborne viruses. Two students received dental care and ten received an intramuscular or intravenous injection.
Depending on type of elective, the destination and the on-site availability of antiretroviral drugs students were advised to take post-exposure prophylaxis with them; 31 students (17%) carried their own supply of PEP but 12 of these students need not have done so as they did not perform procedures that put them at risk of exposure to HIV. Of the 120 students who did perform such procedures, 66 (55%) either had onsite access to PEP or carried a personal supply; 51 (43%) did not know whether the hospital where they performed their elective had PEP and three students (2%) knew that they did not have onsite access to PEP.
Four students experienced mucosal or percutaneous exposure to potentially infectious body fluids while on elective (two in Surinam, one in South Africa and one in Malawi). Five students were unsure whether the event they had experienced qualified as such. None of the students had reported the exposure at the time it occurred and none had used PEP even though all except one either had onsite access to PEP or carried a personal supply. As a result of their response to the questionnaire these nine students were offered screening for HIV and hepatitis C. For reasons of confidentiality we could not find out whether these students opted to be screened.
Other health risks
Nearly all students (98%) filled out the optional questions regarding sexual contact during the time abroad. Eight female students (6%) and three male students (8%) reported having had sex with a new partner; in seven instances with a partner native to the country where the elective was performed. We did not ask whether a condom was used.
Schistosomiasis may be acquired through fresh water contact; 76 students had swum or waded in fresh water in countries where schistosomiasis is prevalent. Of these students 22 had swum in highly endemic countries in sub-Saharan Africa. Eleven of these 22 students had consulted a physician upon return and had mentioned the fresh water contact, 10 were screened of which two showed seroconversion for antischistosomal antibodies.
One student reported a bite by an unidentified animal in the forest in Surinam. He was not vaccinated for rabies. Overall 28 students had been vaccinated against rabies prior to departure.
Malaria chemoprophylaxis
The majority of students (83%) who visited areas that are endemic for malaria used a bed net. Of the 129 students who visited such areas nearly all were prescribed an adequate chemoprophylaxis (75 atovaquone-proguanil, 43 mefloquine, two proguanil, one primaquine and one doxycycline). One student had been prescribed chloroquine by a relative and six students did not remember which prophylaxis had been prescribed. Many students visited countries where malaria prophylaxis is only indicated for selective parts of the country. Of this group 17 did not start prophylaxis. In total 112 students started malaria chemoprophylaxis.
Of the 40 students who used mefloquine 18 (33%) reported an adverse effect: mainly sleep or mood disorder. One student returned prematurely due to neuropsychological adverse effects. Of the 62 students on atovaquone-proguanil 12 (19%) experienced an adverse effect: mainly gastro-intestinal complaints. Eight students who used mefloquine (20%) stopped the drug prematurely as did ten students on atovaquone-proguanil (16%) and the student on doxycycline. Only two of these students switched to another prophylaxis. One did so after having had malaria. All students who stopped using mefloquine did so due to adverse effects. Shortage of tablets or simply forgetting to take the prophylaxis constituted the main reasons for stopping the use of atovaquone/proguanil. Premature stopping of prophylaxis left eight students (15%) unprotected during part of their elective in hyperendemic regions in sub-Saharan Africa.
One student in Benin and one in Kenia were diagnosed with malaria. Both had used mefloquine, but the latter was one of those who had stopped the use due to side effects.
Health problems
Diarrhea was the most common illness and was reported by 117 of 180 students (65%). The incidence was even higher (93%) among 40 students who did not have running water at their lodgings. Most cases were self-limiting and did not last beyond a week. However, 25 of 117 students (21%) had diarrhea accompanied by either bloody stools or fever, and in 29 of 117 students (25%) diarrheal illness caused a temporary interruption of the elective for a mean duration of 2.5 days (median 2 days, range 1 to 7 days). Thirteen of 117 students (11%) consulted a physician for diarrheal illness, three were admitted to hospital, and five received intramuscular or intravenous treatment.
Other common health problems were: constipation (33%), skin infections and wounds (29%) and upper respiratory tract infection (11%). Two students were involved in a traffic accident.
Twenty eight students used an antimicrobial agent; thirteen for enteritis, seven for a urinary tract infection and four each to treat a skin infection and respiratory tract infection.
Post-travel
Seven students (4%) reported having had a fever shortly after returning home. Two of these students consulted a physician and one was diagnosed with Dengue. Travel-related illness after having returned home caused five of 180 students to interrupt their medical course for a period of 7 to 28 days; one due to Dengue, one due to neuropsychological problems attributed to the use of mefloquine, one due to an upper respiratory tract infection and two because they were identified as carriers of MRSA. Dutch hospitals have a low MRSA infection rate and adopt a strict policy to prevent spread of this bacterium [ 10 ]. Screening for MRSA using pharyngeal and nasal swabs is mandatory for hospital employees with recent employment abroad. Upon return, 79 of 180 students (44%) were screened of which two were found to be MRSA carriers (3%). The main focus of screening should be aimed at senior year students involved in clinical work; 70 of 121 senior year students (58%) had been screened for MRSA. Screening was mainly done at the instigation of hospital occupational health departments.
Depending on the destination and the duration of the elective, students are advised to have themselves tested for tuberculosis before departure and 8 weeks after returning home; 84 of 173 students (49%) had a TST performed after returning home. Two students (2%) had a positive reaction which had been negative before the elective abroad. Both had been on a clinical elective; one in Benin and one in Nepal. Both were referred to the municipal health service for counseling. | Discussion
We assessed the health risks that face medical students on an elective abroad to improve the quality and comprehensiveness of pre-and post-travel care. A number of results are related to the risk of bloodborne viral infection. Firstly, we found that regardless of the study year the students were in, none took action following mucosal or percutaneous exposure to potentially infectious body fluids. This result is similar to that of a survey among British medical students [ 4 ]. Secondly, junior students on pre-clinical electives often took part in procedures that pose a risk for bloodborne viral infection. We were not informed about the individual capabilities of the students. Junior students may have had extra-curricular training to perform certain procedures before starting the elective, or they may have been supervised adequately during the elective while learning new procedures. Nevertheless, junior students have not yet received the standard curricular training and in general have limited clinical experience. This puts them at a greater risk of mucosal or percutaneous exposure to potentially infectious body fluids while performing procedures. They may also be less well informed how to act in case of such exposure. Thirdly, we found that allocation of PEP starter kits was inadequate. Kits were commonly handed out to students who turned out not to be at risk of coming in contact with potentially infected body fluids and were not handed out to a sizeable group of students who may have been at risk of such exposure. Due to the difficulty in predicting what students will do while on the elective, improving the pre-travel assessment of who should carry a PEP starter kit is not straightforward. Lastly, systematic education on safe sex should be stressed, as 6% of the students reported that they had sex with a new partner while abroad.
This survey also detected other health risks. One in five students stopped using mefloquine due to adverse effects, which means that a sizeable proportion was left unprotected against malaria. Diarrheal illness was very common as is to be expected. Importantly, a small proportion needed to be hospitalized or required intramuscular or intravenous treatment for diarrheal illness. We also found that medical care following return from the elective can be improved upon. Screening for schistosomiasis, tuberculosis and MRSA did not encompass all who should have been screened.
This study has a number of strengths and limitations. It was restricted to students who had applied for study credits, and we expect this group to constitute the majority of students who perform an elective abroad. For a web-based questionnaire, the response rate was relatively high and none of the questionnaires was incomplete. There are two limitations. The survey was not completely anonymous as we asked the age, gender, study year and e-mail address of the participants. This may have prompted socially desirable answers. The time between having completed the elective and filling out the questionnaire was not standardized and was sometimes quite long which may have reinforced recall bias. To reduce the chance of such bias, we mainly surveyed events that are unlikely to be forgotten, such as needle-stick injury, malaria and diarrheal illness.
Measures that are intended to limit the health risks associated with an elective abroad
Based on the results of this study a number of measures have been adopted to reduce the health risks associated with an elective abroad. Firstly, it has been made mandatory that all medical students planning an elective abroad follow a module on Global Health prior to departure [ 11 - 13 ]. The aim of this module is to enhance student safety and student learning, and to highlight the ethical dimension of an elective abroad. Secondly, at the visit to the administrative department all students are now strongly advised to visit the university occupational health department instead of opting to visit another travel clinic or the general practitioner. By centralizing pre-travel advice, as has been suggested by Tilzey and Banatvala [ 14 ] we expect to achieve a number of improvements. The risk of bloodborne viral infection and the on-site availability of PEP are systematically assessed. This assessment has been standardized. We now ask students to fill out a form describing which procedures they plan to perform. This form is signed by the Dutch supervisor, who judges whether the student is competent to perform the planned procedures and who judges whether the student will be adequately supervised during the elective in case he/she is to learn a new procedure. Based on this signed form, an assessment can be made during the pre-travel consult whether PEP needs to be provided. Whereas students first had to pay for their PEP kit, it is now provided at no cost by the university. To reduce the threshold for reporting and acting on an exposure to potentially infectious body fluids, the written information has been adapted. It now contains a checklist that specifies which steps to take in case of exposure.
If a traveler is to experience adverse effects when using mefloquine, such effects often manifest in the first few weeks of usage. Therefore it is common policy to prescribe mefloquine on trial prior to departure. By centralizing pre-travel advice we aim to increase the proportion of students that receive mefloquine on trial. We have also adapted the written information. In case of adverse effects which seem attributable to mefloquine, students are advised to use half the dosage twice weekly instead of the standard full dosage once a week in order to lower the peak plasma concentration [ 15 ]. Furthermore, students are urged to contact the on call infectious disease consultant in our hospital if they are considering stopping chemoprophylaxis.
To improve post-travel care, upon return all students must now fill out a standard short web-based checklist which assesses certain health risks (exposure to potentially infected body fluids, the risk of schistosomiasis and the need for screening for tuberculosis and MRSA). This results in a computer generated recommendation which states whether the student needs to contact the occupational health department or another care provider for a post-travel consult. | Conclusion
Many of the health risks that were detected in this survey are probably not unique to Dutch medical students. We believe that adopting a standardized pre-and post-travel consult will reduce these health risks by reinforcing knowledge regarding prevention of bloodborne viral infection, by maintaining a clear-cut policy on provision of PEP, by addressing the problem of treatment limiting adverse events with regard to malaria prophylaxis, by reducing the chance of (latent) tuberculosis and chronic schistosomiasis and by preventing spread of MRSA. In a future survey we intend to see whether the new policy is indeed effective in protecting our medical students by limiting health risks. | Background
Clinical and research electives abroad offer medical students many unique experiences. However, participating in an unfamiliar health-care setting combined with limited medical experience may place students at risk of illness. To improve pre-and post-travel care, we assessed the health risks and the quality and comprehensiveness of pre-and post-travel care in a cohort of Dutch medical students returning form an elective abroad.
Methods
All medical students who had performed an elective in the tropics between July 2006 and December 2008 were sent an informative email asking them to complete a web-based questionnaire.
Results
180 of 242 (74%) students completed the questionnaire. Regarding the risk of bloodborne viral infection: 67% of all students and 32% of junior students engaged in procedures that constitute a risk of exposure to bloodborne viral infection, often in countries with high HIV prevalence rates. None of nine students who experienced possible or certain mucosal or percutaneous exposure to potentially infectious body fluids reported the exposure at the time it occurred and none used PEP. Regarding other health risks: 8 of 40 (20%) students stopped using mefloquine due to adverse effects. This left a sizeable proportion unprotected in countries that are hyperendemic for malaria. Post-travel screening for schistosomiasis, tuberculosis (tuberculin skin test) and carriage of methicillin-resistant Staphylococcus aureus (MRSA) encompassed approximately half of all students who should have been screened.
Conclusions
Based on the results of this study we have adopted an integral set of measures to reduce the health risks associated with an elective abroad. The pre and post-travel consult has been centralized and standardized as well as the distribution of PEP. In addition we have developed a mandatory module on Global Health for all medical students planning an elective abroad. | List of abbreviations
(LUMC): Leiden university Medical Center; (HIV): Human Immunodeficiency Virus; (PEP): post-exposure prophylaxis for HIV; (TST): Tuberculin Skin Test; (MRSA): methicillin-resistant Staphylococcus aureus
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
ES participated in the preparation of the protocol and in data acquisition and in revising the manuscript. DS participated in the data analysis and interpretation of the data and in writing the manuscript. JPV participated in the preparation of the protocol and in interpretation of the data and in revising the manuscript. EH participated in the preparation of the protocol and in data acquisition and in revising the manuscript. LGV conceived the study, participated in the preparation of the protocol, interpretation of the data and in revising the manuscript. All authors read and approved the final manuscript."
Funding
There was no dedicated funding for this project.
Pre-publication history
The pre-publication history for this paper can be accessed here:
http://www.biomedcentral.com/1472-6920/10/89/prepub | CC BY | no | 2022-01-12 15:21:37 | BMC Med Educ. 2010 Dec 2; 10:89 | oa_package/03/f1/PMC3014955.tar.gz |
|
PMC3014956 | 21129220 | Background
Virtual Patient Simulation (VPS) is a growing research field, as these software applications are considered to have entered "the mainstream of medical education" [ 1 ]. The overwhelming majority of articles published are quantitative in nature and in most cases address the use of virtual patients for learning, while only a fraction are dedicated to assessment and even fewer to other topics of interest. In spite of promising research results in the recent years and of some attempts to integrate VPS in different curricula, the "effective use requires evidence to guide design and integration" [ 2 ]. One important factor in terms of success or failure of curricular implementation of VPS is the opinion of users, both students and faculty staff. For example, if the users' expectations do not match the curricular use envisaged by the faculty or the options offered by the application itself, the possibility of a successful implementation of VPS in a medical curriculum is limited.
One way to generate the much needed information on how to best implement VPS is by means of qualitative research studies, where the different stakeholders-students, clinical teachers, course directors, faculty board-can convey their views on the role of VPS in healthcare education. Qualitative methods are extensively used in health sciences education research, especially in nursing and psychology; however, to our knowledge, qualitative methodology is not so frequently used as quantitative methodology for research on virtual patient simulation systems, with few exceptions [ 3 - 5 ]. Amongst qualitative methods, different interviewing techniques help researchers to detail the stakeholders' personal views [ 6 , 7 ].
This study aimed to report on the experiences of medical students with a virtual patient simulation system integrated in the Internal Medicine curriculum at Universidad el Bosque (Colombia). We explored the participants' opinions on the educational use of the application by using focus group interviews, as we believe that the learners' perspectives on such systems are quintessential to a successful VPS development and implementation. | Methods
We conducted focus group interviews to explore medical students' opinions on the educational use of the system Web-based Simulation of Patients (Web-SP). The interviews were performed at the Faculty of Medicine of Universidad el Bosque (Bogota), in January 2008.
Study design
The Virtual Patient application
Web-SP is an explorative linear-interactive virtual patient simulation developed at the Department of Learning, Informatics, Management and Ethics, Karolinska Institutet, Sweden [ 8 ]. A Spanish variant of the software has been developed, localized [ 9 ] and used from mid 2005 for learning and performance-based assessment in the Internal Medicine course at the Faculty of Medicine, Universidad el Bosque [ 10 , 11 ]. The virtual cases used at Universidad el Bosque were created from actual clinical records of patients from university hospitals in Bogota. As a consequence, the virtual cases contained patient photographs and diagnostic media pertaining to the clinical records. All patients had previously signed an informed consent form, enabling the clinical teachers to use all the information in the records as a base for case creation. In order to solve a virtual patient case, the students were required to gather information from patient interviews, physical exam and ancillary tests, in order to formulate the diagnosis and the treatment course. After submitting the treatment, the students gained access to the feedback module; in the case of the Spanish version, the feedback consisted of a detailed case discussion provided by a senior clinical expert and the actual patient follow-up.
Participant selection
The participants were medical students who had used Web-SP cases for learning and/or for assessment as part of their Internal Medicine curriculum in 2007. They belonged to a cohort of 216 students participating in a larger study on virtual patient assessment results [ 10 , 11 ]. At the time of the interviews, 49 students had their clinical rotations in Bogota and were geographically available (convenience sampling); out of the 49, 16 students were selected by simple randomization to participate in the focus-group interviews. Randomization was performed to ensure representativeness, so that students with "good experiences" with Web-SP in terms of assessment results (who could have been more positive to the educational use of the application) were not purposively selected. We deliberately recruited more students than needed, in order to allow students to decline participation or drop out at a later time. Equal numbers were recruited from the ̈study ̈ and ̈control ̈ arms of the study mentioned earlier, to avoid any biased opinions about the educational tool. The randomly chosen students were directly contacted by the vice dean's office and invited to participate in the study. None of the students declined participation, nor dropped out of the study. When all students confirmed their participation, we had to arrange two focus groups, as 16 participants would have been too many to accommodate in one session.
The main researcher had been the course director for Internal Medicine, implemented Web-SP in the study plan and conducted two other previous studies in the same population, so a trustful relationship between the principal investigator and the students had already been established prior to study commencement; however, the existence of a relative position of power is hereby acknowledged. Ethical permit to perform the interviews was given by the Faculty Board.
Setting
The interviews were conducted in facilities of the Faculty of Medicine. MB led the focus group interviews, assisted by one of the local clinical teachers. The discussion began by recapitulating their experience with the system and by stating the scope of the interviews, i.e. the exploration of their perception of the system. Then several topics were introduced, as the students had used the system for learning and/or for assessment and had also collected clinical cases in the hospitals. During the interviews the students themselves raised issues the researchers did not plan to bring up, such as communication skills development or their motivation in using the system.
Data collection
The interviews were conducted in Spanish, by the main investigator (MB). An interview guide had previously been developed and peer validated on site. Both audio and video recordings were made during the interviews, which were 60 and 62 minutes long, respectively. No field notes were taken during or after the focus group interviews. Each topic was discussed till estimated saturation.
Data analysis
The audio tapes were transcribed by the main author. A non-verbatim approach was used, meaning that the transcriptions were edited for pauses, interjections and other audio utterances whose omission would not alter in any way the message conveyed by the participants. The video recordings were used to identify the speakers on the audiotapes, which were assigned consecutive numbers. The transcripts were translated into English by an authorised translator not affiliated to Karolinska Institutet or to Universidad el Bosque. The theoretical framework used in this study was content analysis [ 6 ]. MB and UF reviewed the transcripts independently, compared notes and reconciled the differences. Emerging coding was used to obtain categories and themes, whose final form was reached after another consensus discussion between the two coders. Given the manageable length of the two focus group discussions, no software was used to analyze the data. | Results
Emerging coding identified eighteen categories in the transcripts, which were further clustered into five themes. The presentation of the results includes quotations, which are given in brackets and are followed by the student identification number. The most articulate quotation was chosen in each case, in order to avoid redundancy. However, this does not mean that other students did not express the same idea in other words and did not merely agree. When several participants concurred, the identification numbers were omitted.
For a synopsis of the results, please consult Table 1 .
Theme 1: Learning
-Category 1: Clinical reasoning
The main educational scope of virtual patients found in the interviews is the use for training clinical reasoning-" I believe that clinical reasoning abilities are reinforced with virtual patients " (student 1)-, for which the constant use of VPS is deemed essential. Training with VPS is also thought to enhance analytic and synthetic reasoning-" Analysis and synthesis...are basically the aspects that are strengthened through virtual cases...and may be reinforced by constant practice using the program " (student 3). Clinical reasoning development is also linked to a stepwise approach to case solving, fostered by VPS-" How I see it is that it does help the mental process of how to approach a patient, from the reason why a patient needs a consultation until the final diagnosis. When a real patient comes, we already know what to ask, what we have to do, in what order to do it, so we do not miss anything, from the reason for the consultation till diagnosis and treatment in an adequate manner without forgetting any important aspect " (student 4). Even though clinical reasoning development is seen as the primary scope of learning with VPS, input on factual and core knowledge is welcome as part of the "package", especially as feedback-" When receiving feedback, we learn more about pathophysiology, molecular biology etc. and thus complement the clinical part with a more theoretical, abstract part " (student 1); the VPS scope-" would be to reinforce clinical reasoning abilities, but also give feedback with regards to the theoretical aspects of the disease in question " (student 8). The participants greatly appreciate that VPS enable them to follow up a patient from the beginning to the end; the cases give students a sense of closure, which enhances motivation-" In the hospital you meet the patient in an initial phase, during the investigation or treatment, but you never know what happened later" (student 2). The students understand VPS as a "preparation for the real life as doctors".
-Category 2: Transferable skills
The real cases, seen later on as a young doctor, could be solved by association and comparison with virtual patient cases with similar characteristics-" It could happen that we remember a virtual case in order to use that experience or some of it in reality " (student 2). The knowledge acquired with VPS is considered transferable to other types of exams. Moreover, most students consider that knowledge transfers directly to the real patients, especially when the cases used in teaching were created from real life clinical records-" If we meet a patient with cardiac failure on Web-SP... and we meet a patient with similar characteristics in real life, then one can say: the patient has cardiac failure or possible cardiac failure and then the second part would be like choosing the indicated options through Web-SP for ordering lab tests in order to reach a diagnosis and treat the patient " (student 6).
-Category 3: Retention enhancement
Students appreciate that VPS help retain information:-" I remember more Mr. X's case than the page in the textbook " (student 10);-" We associate a disease more to a patient than to the textbook. If I saw the patient, saw the photo and questioned the patient in the program, I will remember more easily, I'll have my flashback of that pathology, more than if I only studied my class notes or a book " (student 4).
-Category 4: Making mistakes and learning from them
Recognizing and correcting mistakes, either in the clinical reasoning path or in previously acquired knowledge, is considered crucial for successful learning-" Web-SP lets us correct or reinforce certain knowledge so that, later, when we meet with reality or real people or cases, we won't make, or at least make less mistakes, in order to reach a correct diagnosis or adequate treatment for the patients " (student 2). Moreover, students even think that they are less prone to repeat the mistakes made with VPS in their clinical practice-"... so that you can go on to the clinical practice and maybe not make those same mistakes "-as in Web-SP (student 10). Making mistakes with VPS is also not as stressful as in the real life context-" It seems to me that it is good in part because one learns from mistakes and one is not too stressed about making mistakes, making mistakes doesn't have severe consequences " (student 11);-" Better to make virtual mistakes than real ones " (student 13).
Communication skills
The participants consider that the development of communication skills falls out of the scope of VPS. On one hand, VPS are ill-suited for such purpose-" Web-SP is what connects theory and practice, because it really applies to what we experience in real life, but without communication skills " (student 5)-and on the other, communication skills are acquired by direct contact with the patient-" I think that those skills are directly acquired in another way " (student 5).
Theme 2: Teaching
Category 1: Clinical specialties
-The participants believe that VPS should enjoy a broad use across clinical specialties. The majority of students favour the curricular use of VPS across the main clinical specialties, such as internal medicine, surgery, obstetrics-gynecology and paediatrics.
-Within a given clinical specialty, the virtual patient cases should present frequent diseases and their complications; topics not included in the study plan and in the clinical rotations should also benefit from VPS, which could close the gap-" If we do not have rounds in a specialty, it would be good to see virtual patients, we gain by it " (student 4). Diseases that might easily be missed during a short clinical rotation (due to seasonality or to being endemic in a different geographic area), but are relevant for the future clinical practice, should also be presented as VPS-" Knowing about a particular disease in a specific geographic area is very important " and " Here in Bogota there's hardly any malaria or Leishmaniasis or yellow fever.. ." (student 11). The students point out that 5-6 cases of tropical diseases would make an acceptable minimum in the study plan.
Category 2: Regulatory effect of VPS
A possible regulatory effect of VPS is envisaged by the participants. Different rotation locations in Colombia offer a varying infectious disease spectrum, as well as patients with different socio-economic status. VPS are seen by most students as a chance to make instruction uniform among rotations sites.
The majority of students in our study seem to require external regulation, by means of limiting the availability of the system-" Without a time limit we can say: I'll check the cases later, and then nothing happens; but if there's a time limit, well, this week I see cardiac failure patients etc. It's more practical for us and also for the teachers, I think " (student 5). Time restrictions help students to plan and organize their workload. The use of a system permanently left open is estimated by the participants at a surprisingly low 3%.
Theme 3: Assessment
-Category 1: VPS assessment-qualitatively different from regular exams
According to the interviewees, VPS evaluate knowledge in a different manner, emphasizing the clinical reasoning process. The students consider VPS to be a more didactic form of evaluation and an intrinsically better evaluation tool than traditional exams-" VPS evaluation lets you see your strengths and weaknesses, where you are failing and what you need to improve, while on a paper exam one can, many times, get it right just by chance " (student 15);-" VPS are a more didactic evaluation method because...little by little we move forward to the big question, the diagnosis and treatment of this patient, after looking at all the patient's data, including images, audio and video " (student 1). However, students agree that VPS should not be the only assessment form used in a course. As pointed out earlier, VPS also allow an increased retrieval of information in comparison with regular examinations and therefore grades tend to be higher with virtual patient evaluation-"... one has better chances to get information about the patient, study the case " (student 10).
-Category 2: Increased motivation
VPS evaluation-" does not feel as exam in the long run " (student 1). For example, Web-SP features open questions, which are "harder" than closed/multiple-choice questions; despite that, they "feel more natural" and the evaluation is deemed by most students as less stressful than with other assessment methods.
-Category 3: Professional focus in assessment
Most students want to learn from summative assessment and from the feedback on assessment. Participants say they " get more " from VPS assessment-"... the reasoning process is evaluated more, not just the answer, not just the diagnosis, but the explanation you submitted for the diagnosis, the tests you ordered, the questions you asked the patient " (student 12);-"... the analysis you made " (student 14). However, assessment should always be relevant for the future clinical practice as a young doctor. VPS assessment, as well as evaluation in general terms, should be directed to-" what a general practitioner should know on a subject " (student 13).
-Category 4: Production assessment
The students are positive to open questions, which-" make you think and analyze the patient " (student 2). They seem to favour open questions, even if the grades might be lower than with other types of evaluation.
Theme 4: Authenticity
-Category 1: Design and content
In the opinion of participants, VP design should reflect the real clinical practice. The option menus should offer realistic and localized choices in terms of history taking, physical exam and diagnostic tools, as well as feedback on treatment alternatives-" Simulated cases should teach us how to act in real life, they have to be similar with real life practice, and if the computer gives me all those choices and I won't have access to them in the hospital, well, it doesn't give me much " (student 5). Realism could encompass even the actual costs for the healthcare system-" It would be great if the program showed how much a certain test costs and how much the patient and the health insurance company would have to pay " (student 11). Too many unstructured options in a VPS may be misguiding; the abundance is not necessary to meet the learning objective-"... there are options that might confound what you already know...you can't make a diagnosis because you're lost " (student 15).
-Category 2: Localization of the socio-cultural context
In the case of applications "born" in one country and "adopted" in another, it is crucial to adapt the patient interview section to the socio-cultural context of the patient-" If I ask the patient a question literally drawn from Web-SP he/she might say " I don't understand the question " (student 10).
-Category 3: Realism and virtuality
The participants think that real life records make better cases than fictitious scenarios; moreover, such cases are thought to provide transferable skills. Media presence is essential to authenticity, and realism starts with the patient photo-" The fact alone of being able to see a photo of the patient...gives us more than a paper case; it is different to see for myself that the patient is sad, then to read: the patient looks sad" (student 16). A video recording is a must in certain circumstances:-" In a paper case I can read that the patient is seizing, but in Web-SP I can see him seizing " (student 10). The participants no longer perceive the cases created from real life patient records as "virtual"-" the case was not virtual anymore because my patient was a real one " (student12).
-Category 4: Feedback
Feedback on the actual evolution of the patient adds to the realism of the application. The feedback section should show the actual clinical evolution and the effect of treatment-" It is very different to imagine the patient got better with antibiotic administration, than having proof-in the feedback-that the antibiotic actually worked " (student 10).
Theme 5: Implementation
As students seemed unanimous with regard to implementation strategies, no quotes are given in this section. A number of important issues were brought up by the students, including:
-Category 1: Number of cases
More cases were linked to more knowledge. More than one case per topic can be necessary for common diseases which are often complicated/have co-morbidities at presentation.
-Category 2: Access/availability
Broad access preferred in relation to site-campus, hospital, home -, but not in relation to timeframe, which should be restricted, as previously shown.
-Category 3: Virtual Patient exchange
The tropical diseases cases developed at Universidad el Bosque should be subject to exchange with other Higher Education Institutions. | Discussion
Focus group interviews were the methodology chosen for this study. From the array of qualitative methods, both questionnaires and individual in-depth interviews would have been reasonable alternatives. After discussing the possibility of using in-depth interviews and questionnaires as opposed to focus groups, we chose the latter. The reason was that we intended to collect as much data as possible while still keeping the project feasible. We felt that while allowing the collection of sufficient, good quality data, it was also a practical way-financially and time-wise-to investigate the participants' opinions. The students have also come up with novel information, i.e. descriptions, explorations and ideas that brought up issues new to the researchers, which in our understanding might have not surfaced with individual interviews or questionnaires.
Learning
Clinical reasoning development is already recognized by the literature as the main scope of VPS in education [ 2 , 4 , 10 ]. Our study adds the "factual and core knowledge" input via the virtual patient application, which was considered by the participants as a highly desirable feature of VPS.
Instead of re-experiencing the daily frustration of not knowing what happened with their patients (rotation ended, patient was moved to another floor etc.), the students conveyed the importance of getting closure by means of patient feedback.
The students indicate that the transfer of knowledge to the real patient is the ultimate goal of simulation technology, in agreement with the literature [ 12 , 4 , 11 ].
Medical students feel they remember more with VPS. Retention enhancement with VPS has previously been demonstrated in the same cohort of students, where the effect of VPS on early and late retention was studied [ 11 ].
Making mistakes was not expected to rise to the level of category. Free from a stressful context (patient, family, hospital staff), the students still perceive errors as serious events, but at the same time as meaningful learning opportunities, to the point that they consider their repetition as unlikely in their future clinical practice.
Communication . In our experience, VPS are not a proper tool for communication skills development. This opinion is shared by the VPS community at large and by the participants in our study. The latter found their motivation in the reduced interactivity of the system, which does not reflect the richness of direct interaction with the real patient.
Teaching and assessment
The students in our study had used VPS both for learning and for assessment in their Internal Medicine course, where Web-SP was a curricular component. Not surprisingly, they do not envisage a use for VPS outside of the curricular context, e.g. as an add-on. They would like other major clinical specialties to offer the opportunity to work with VPS; especially so if certain topics are not comprised in the study plan, but are likely to surface in the first level of attention, finding in alignment with the recent accreditation requirements of the Liaison Committee on Medical Education (LCME) [ 1 , 13 ]. Other than that, they also express the need for solving cases of common, frequent diseases, particularly cases already complicated at debut, patients with several co-morbidities, as well as cases of drug interaction. To our surprise, the participants evoked a regulatory effect of VPS at individual level (to help program their learning) and at institutional level (to even out the differences in disease range among rotation sites).
As for summative assessment, we were not surprised to see that participants consider VPS a good evaluation tool [ 10 , 11 ]. To come across the conviction that VPS assessment is qualitatively different in comparison with other evaluation methods was however quite unexpected. A further benefit of a virtual patient application is that assessment may not feel as an exam, leading to an increased motivation for learning itself. The students are also aware that they can and should learn from assessment; feedback is deemed crucial for such learning through assessment to occur. They do not support, however, a VPS assessment directed to anything else than knowledge and skills essential for the clinical practice as a general practitioner, or the use of VPS outside regular course evaluation. Caution should be exerted when generalizing a benefit for learning across different systems, as the assessment formats are different (open-ended questions in Web-SP, multiple choice in other systems) and the feedback, if provided, also varies in layout.
Authenticity
The design of the application should reflect the reality of clinical practice [ 4 , 14 ] and offer localized menus and content choices [ 9 ]. Authenticity might be increased by considering the actual costs involved in diagnosing and treating the virtual patient, but we believe that such a development should be subject to a localization demand. Layouts that feel artificial in terms of menus or content do not meet their educational goal and are considered by participants as downright misleading [ 14 ]. In contrast, knowledge derived from virtual cases created from real life clinical records is thought to be directly transferred to actual patients, which is a novel finding. An additional possible benefit of using real cases to create virtual patients is a further enhancement of authenticity to such level as to consider the application as a surrogate for reality (a desirable feature in the case of rare diseases, topics not seen during a clinical rotation, diseases unavailable geographically or seasonally etc.). Feedback derived from real cases greatly adds to the realism of the application.
In our experience, media use should serve an educational goal and not become a goal. The literature [ 2 , 4 ] supports our results, that authenticity can be conveyed by means as simple as the face photo, together with shots of the main findings of the physical exam. We were somewhat puzzled by the students' emphasis on the importance of the face photo; nevertheless, the Spanish version of Web-SP features photos of genuinely ill individuals, who may well wait in line for a consult at the clinic. The students responded empathetically to faces they considered familiar. We believe that more media resources are rarely needed. In our experience at Karolinska Institutet, the creation of such cases is unnecessarily expensive and their added value is minimal, if not null. Informed consent for any media use is an obvious must for cases created from actual clinical records, as well as for the content of the medical history itself.
Implementation
More cases equate more knowledge in the students' opinion. However, they seem aware of the practical difficulties of achieving even a modest goal, such as one case per topic; here, as previously discussed, the options are wide: common diseases, pathologies not seen during the clinical rotations and/or not included in the study plan etc. Such difficulties are more apparent if the cases are to be created from real life records (mainly patient informed consent issues and hospital approval for retrieving information from clinical records).
Most participants recommend that the availability of the application be restricted to certain term timeslots, as a means of external regulation. Such a position is conflicting with the scope of a web-based application-namely round-the-clock access from any site-and might reflect special characteristics of the student population (whose identification was beyond the scope and the methodology of this study).
The exchange of virtual patient cases is not a priority for many institutions [ 14 ] and time will tell to what extent current inter-operability efforts are worth undertaking (e.g. as number of user sessions per case exchanged).
Limitations
Based on our experience at LIME with VPS of a linear design, we believe our findings could be generalized to such systems. However, our setting was somewhat unusual, in the sense that the virtual patient cases were created from real life clinical records and included the actual clinical follow-up in the feedback section. Furthermore, the sample population-undergraduate medical students in Latin America-may well have different socio-cultural characteristics than participants in other studies (which makes appealing the option of designing VPS adapted to specific categories of learners). | Conclusions
Our study found five main themes to be associated with successful VPS use in medical curriculum: Learning, Teaching, Assessment, Authenticity and Implementation. Medical students perceive Virtual Patients as important learning and assessment tools, fostering clinical reasoning, in preparation for the future clinical practice as young doctors. However, an application must comply with certain design, authenticity and implementation requirements, in order to reach its educational goal. | Background
The learners' perspectives on Virtual Patient Simulation systems (VPS) are quintessential to their successful development and implementation. Focus group interviews were conducted in order to explore the opinions of medical students on the educational use of a VPS, the Web-based Simulation of Patients application (Web-SP).
Methods
Two focus group interviews-each with 8 undergraduate students who had used Web-SP cases for learning and/or assessment as part of their Internal Medicine curriculum in 2007-were performed at the Faculty of Medicine of Universidad el Bosque (Bogota), in January 2008. The interviews were conducted in Spanish, transcribed by the main researcher and translated into English. The resulting transcripts were independently coded by two authors, who also performed the content analysis. Each coder analyzed the data separately, arriving to categories and themes, whose final form was reached after a consensus discussion.
Results
Eighteen categories were identified and clustered into five main themes: learning, teaching, assessment, authenticity and implementation. In agreement with the literature, clinical reasoning development is envisaged by students to be the main scope of VPS use; transferable skills, retention enhancement and the importance of making mistakes are other categories circumscribed to this theme. VPS should enjoy a broad use across clinical specialties and support learning of topics not seen during clinical rotations; they are thought to have a regulatory effect at individual level, helping the students to plan their learning. The participants believe that assessment with VPS should be relevant for their future clinical practice; it is deemed to be qualitatively different from regular exams and to increase student motivation. The VPS design and content, the localization of the socio-cultural context, the realism of the cases, as well as the presence and quality of feedback are intrinsic features contributing to VPS authenticity.
Conclusions
Five main themes were found to be associated with successful VPS use in medical curriculum: Learning, Teaching, Assessment, Authenticity and Implementation. Medical students perceive Virtual Patients as important learning and assessment tools, fostering clinical reasoning, in preparation for the future clinical practice as young doctors. However, a number of issues regarding VPS design, authenticity and implementation need to be fulfilled, in order to reach the potential educational goals of such applications. | Competing interests
The authors declare that they have no competing interests.
Authors' contributions
MB conducted the focus group interviews. MB and UF independently coded the data. MB and UF performed the content analysis. All authors contributed substantially to the conception and design of the study, as well as to the critical revision of the paper. All authors approved the final manuscript.
Authors' information
MB is a physician specialized in hematology and a doctoral candidate at the Department of Learning, Informatics, Management and Ethics (LIME), Karolinska Institutet. MB has several years' experience as a clinical teacher and is fluent in Spanish. HH is visiting professor in Medical Education at LIME. UF, professor of Medical Educational Simulation, is the head of LIME.
Pre-publication history
The pre-publication history for this paper can be accessed here:
http://www.biomedcentral.com/1472-6920/10/91/prepub | Acknowledgements
To Mesfin Kassaye Tessma (LIME), for the critical revision of the paper. And to our (ever patient) students. | CC BY | no | 2022-01-12 15:21:37 | BMC Med Educ. 2010 Dec 4; 10:91 | oa_package/49/d2/PMC3014956.tar.gz |
PMC3014957 | 21118498 | Background
Behavioural and psychological symptoms are common in dementia. These symptoms include depression and anxiety, psychotic symptoms, wandering, agitated behaviour and sleep disorders, and are collectively known as the behavioural and psychological symptoms of dementia (BPSD). These symptoms confer a large proportion of the social burden of dementia [ 1 , 2 ], and are important targets for intervention in dementia patients [ 3 ]. Yet these symptoms are not restricted to those with dementia. Many behavioural and psychological problems are also present in a significant proportion of the non-demented older population [ 4 , 5 ]. Behavioural and psychological symptoms have been associated with cognitive impairments not sufficient for a diagnosis of dementia, although their prevalence and relationship to cognitive impairments are not clear. Two recent reviews have highlighted the variation in estimates of the prevalence of behavioural and psychological symptoms in those with mild cognitive impairment (MCI) and suggested that differences in the settings of studies, the characteristics of participant groups, and the definitions of symptoms and of MCI contribute to this variation [ 6 , 7 ]. We have previously compared the prevalence of BPSD in the older population with and without dementia [ 4 ]. In this article we use data from the population-representative Medical Research Council Cognitive Function and Ageing Study (MRC CFAS) to explore the relationship between behavioural and psychological symptoms with the socio-demographic, health and cognitive factors that are used to define different classifications of MCI. | Methods
Study participants
The Medical Research Council Cognitive Function and Ageing Study (MRC CFAS), is a multi-centre longitudinal study that is fully described elsewhere [ 8 ]. Between 1990 and 1991 a random sample of individuals aged 65 years and older living in five centres representative of rural and urban areas in England and Wales (Cambridgeshire, Gwynedd, Newcastle, Nottingham and Oxford) were contacted, with a response rate of 82%. At baseline, 13,004 participants completed the screening interview. A subgroup of those screened including all participants with severe cognitive impairments and a stratified subsample of the remainder (n = 2640) also completed a more detailed assessment. This included selected items from the Geriatric Mental State Automated Geriatric Examination for Computer Assisted Taxonomy (GMS-AGECAT), Mini Mental State Examination (MMSE) [ 9 ] and the multi-domain Cambridge Cognitive Examination (CAMCOG) [ 10 ].
Persons diagnosed with dementia (defined as an AGECAT organicity rating case level of O3 or above, n = 587 [ 11 ]), an unknown dementia status (n = 3), a MMSE score that is missing or below 18 (n = 243), or Parkinson's disease (n = 26) were excluded. Therefore of the 2640 who undertook the assessment interview 1781 persons were included in this study. The aim of this study was to analyse the psychiatric correlates of cognitive impairment, and so those with underlying psychiatric conditions were not excluded from the study.
Informed Consent and Confidentiality
MRC CFAS has Multi-centre Research Ethics Committee's approval and ethical approval from the relevant local research ethics committees. All participants gave informed consent and patient confidentiality was not breached.
Assessment of behavioural and psychological symptoms
Symptoms assessed during the screen, assessment and informant interviews include depression, apathy, anxiety, feelings of persecution, hallucination, agitated behaviour, elation, irritability, sleep problems, wandering, confabulation and misidentification. Symptoms were assessed using interviews conducted with both study participants and their informants, combined with observations made by the interviewer during the course of the interviews. Depression was measured with the Cambridge Mental Disorders of the Elderly Examination (CAMDEX) [ 12 ]. Details of the questions used to assess each symptom have been described previously [ 4 ].
Assessment of cognitive function and definition for 'mild cognitive impairment' criteria
Cognitive function was assessed using the MMSE [ 9 ] and by classification based on definitions of mild cognitive impairment (MCI) and its subtypes [ 13 ]. Using the MMSE, cognitive function was classified as low (18-21), intermediate (22-25) or high (26-30). Classifications of MCI aim to identify individuals with cognitive impairments likely to indicate incipient dementia. Three subtypes are considered, amnestic (A-MCI) including deficits in memory, non-amnestic (N-MCI) in which cognitive impairment is restricted to non-memory domains and multiple (M-MCI) in which both memory and non-memory domains of cognition are impaired [ 13 ].
A clinical diagnosis of MCI requires a subjective complaint of memory impairment by either the respondent themselves or an informant. Performance on cognitive tests must fall below a specific threshold, typically below 1 or 1.5 standard deviations of the mean age-specific performance, or if test scores are highly skewed below the 16 th or 7 th percentiles. To exclude those with possible dementia, general cognitive function must not be severely impaired. Exclusion criteria can include any health or psychiatric problems that might be the alternative cause of cognitive impairment, or a severe impairment in activities of daily living. The operationalisation of these criteria, including the level of subjective memory complaint, the specific cognitive tests and exact exclusion criteria, are not specified.
For the present study, 'subjective memory complaint' was defined as a report of memory problems by either the respondent or an informant at either the screening or the assessment interview. 'Objective memory impairment' was defined using as a score below the 16 th age-specific percentile on one or more of the remote memory, recent memory or learning memory domains of the CAMCOG. 'Non-memory impairment' was defined using as a score of below the 16 th age-specific percentile on any of the other cognitive domains measured by CAMCOG which includes orientation, language, attention/calculation, praxis, abstract thinking or perception. CAMCOG cut-off scores were age adjusted using five 5-year age groups including 65-69, 70-74, 75-79, 80-84 and 85+ years. General cognitive function was deemed to be severely impaired when participants scored less than 22 on MMSE. No health-related or functional exclusions were applied, since the relationships between these and behavioural and psychological symptoms were of interest in the present study.
MCI subtypes were assigned as follows: All MCI subtypes required subjective memory complaint and no severe impairment of general cognitive function (ie MMSE > = 22). Those with memory impairment but no non-memory impairment were classified as A-MCI. Those with non-memory impairment but no memory impairment were classified as N-MCI. Those with impairments in memory and non-memory domains were classified as M-MCI. Participants were not excluded on the basis of medical or psychiatric problems. Full details of the mapping of MCI in MRC CFAS have been published elsewhere [ 14 ].
A 'not impaired' group was defined using the following criteria: normal general cognitive function, no severe functional impairment, normal memory and non-memory test performance. Participants who could not be classified as "not impaired" or MCI using any definition were classified as "other cognitive impairment, no dementia" (OCIND). This heterogeneous group includes both respondents with an objective memory impairment but no subjective memory complaint, and those with impaired general cognitive function but not diagnosed with dementia.
Covariates
As well as those factors that contribute to the definition of MCI, other factors considered in our analysis included sex, age, institutionalization, vascular co-morbidity, education, and social class. Age was reported continuously as well as being dichotomized into 65-74 years versus 75 years and older. Vascular co-morbidity was defined using self or informant reported history of a heart attack, stroke or diabetes. Institutionalisation was divided into two groups including, independent (living alone or in a warden controlled flat) or dependent (living in a residential home, nursing home or hospital). Level of education was measured in number of years of education and was divided into two groups including less than 10 years and 10 or more years of education. Functional impairment was defined as requiring help at least several times per week with activities of daily living (ADLs) such as washing, cooking for themselves, dressing, or if the respondent was housebound. Data were missing in fewer than 2% of individuals for each covariate and was assumed to be missing at random.
Statistical analysis
All analyses were performed using Stata 10.0. The prevalence of each behavioural and psychiatric symptom was estimated in cognitive groups. Participants were grouped both by MMSE scores, and by MCI classification. Three MMSE categories were defined including low (18-21), mild (22-25) and high (26-30). MCI classifications were 'not cognitively impaired', 'MCI' and 'OCIND'. Prevalences were also compared across MCI subgroups A-MCI, N-MCI and M-MCI as previously described. Differences in prevalence across groups were tested using likelihood ratio tests. Post-hoc pairwise tests comparing subgroups were not conducted.
Participants were back-weighted in prevalence estimation to adjust for over-sampling in the study population of individuals aged 75 years or older and the subsequent stratified sampling for the assessment interview. The relation between symptoms, subjective memory complaint, objective memory impairment, vascular disease, disability and socio-demographic factors was investigated using both univariate and multivariate logistic regression.
Co-occurrence between symptoms was measured using odds ratios estimated using logistic regression. Univariate analyses and multivariate analyses adjusting for the presence of all other symptoms, MMSE, age and sex were conducted. Factor analysis was used to determine the structure of the co-occurrence of symptoms. The Spearman's correlation matrix relating the symptoms was estimated and this was subjected to the principle factor analysis algorithm. The resulting factor solution was rotated for ease of interpretation using the Varimax rotation. This factor analysis was repeated within each of the MMSE and MCI subgroups described above. | Results
Characteristics of study participants
Table 1 shows the characteristics of the study participants stratified by MCI classification. The age of all three groups was similar, but women were more likely to be cognitively impaired. Few individuals were living in institutions. Those without cognitive impairments were substantially more likely to be better educated and in a higher social class, and to have better self rated health. The proportion of participants with reported vascular disease was similar across groups. The OCIND group had lower cognitive function than the MCI group or the NCI group, and a lower prevalence of impairments in ADLs. The OCIND group reported most frequent use of anti-psychotic medication, while the MCI group had greater reported use of anti-depressants.
Prevalence of symptoms across MMSE cognitive classifications
The prevalence of each behavioural and psychological symptom by MMSE group and MCI classification is shown in table 2 . The previously reported prevalence in those with dementia is also shown for comparison. The prevalence of many symptoms was higher in those with worse cognition.
There was a statistically significant trend to increased depression, apathy, psychosis, agitation, and irritability in those with worse cognition. Inappropriate elation, wandering and confabulation were rare in this sample making differences between groups difficult to detect, nevertheless there is some evidence that each of these is more common in those with worse cognition. Sleep problems were common and reported with equal frequency across MMSE groups. There was no significant difference in the prevalence of anxiety across cognitive groups.
Prevalence across MCI classifications
Significant differences in the prevalence of depression, apathy, persecution and irritability were seen across MCI groups, with the prevalence of each of these lowest in those without cognitive impairment. Sleep problems were reported least often in the OCIND group (38.6%), and most often in the MCI group (49%). Wandering and confabulation were rarely reported in this sample with little evidence of a difference across groups. Hallucinations, agitated behaviour and anxiety were seen least often in those with no impairment and most often in MCI, but these differences were not statistically significant. There was little difference in the prevalence of misidentification across groups.
Cognitive function was higher in the A-MCI group (mean MMSE = 26.3) than in the N-MCI (mean MMSE = 25.3) and M-MCI (mean MMSE = 24.5) groups (F(2,387) = 10.91:p < 0.001). The only statistically significant differences in symptom prevalence observed across MCI subgroups were a high prevalence of elation in A-MCI and a higher prevalence of hallucinations in those with memory impairments (ie M-MCI and A-MCI). There is some evidence that depression and apathy are more common in those with non-memory impairments (N-MCI and M-MCI) with anxiety more common in those with memory impairments. Feelings of persecution and sleep problems were equally common across all groups.
Effects of health-related, sociodemographic and cognitive characteristics on presence of symptoms
In univariate analyses (additional file 1 : table S1) memory impairment, subjective memory complaint, ADL impairment, and lower MMSE score were associated with higher prevalence of many symptoms. Vascular comorbidity is associated with an increased risk of symptoms of mood including depression, anxiety, apathy and irritability.
After adjustment for all other risk factors (additional file 2 : table S2) the effect of objective memory impairment and of MMSE score is largely attenuated, while the presence of subjective memory complaint remains associated with the presence of many symptoms. The association between vascular disease with depression, apathy, anxiety and irritability remains although strength is reduced.
Many symptoms appear less common in older people in multivariate analyses, suggesting younger people with lower cognition were more at risk of BPSD than older people. Depression and sleep problems are more frequently reported in women, while apathy is more common in men. There is a non-significant increase in other symptoms in women, which remains in multivariate analyses.
Co-occurrence of symptoms
Table S3 in additional file 3 shows univariate and multivariate pairwise associations between symptoms. There are many significant associations between pairs of symptoms which remain after adjusting for the presence of all other symptoms, MMSE, age and sex. Factor analysis (table 3 ) suggests three factors arising from this observed co-occurrence. The three factor solution included: (a) symptoms of hallucination and misidentification, (b) symptoms of anxiety, apathy and depression and (c) irritability, persecution and wandering. The fourth strongest factor does not include any symptoms with loading greater than 0.3. However the uniqueness of each symptom was high, suggesting that much of the variability of each symptom cannot be explained by this factor model, and there are many statistically significant pairwise associations between symptoms that the factor model does not include (supp table 3). The factor structure is largely consistent across cognitive and MCI groups (table 4 ) and is similar although with weaker associations to that seen in the population with dementia [ 4 ]. | Discussion
Summary of findings
In this large, representative study of the population without dementia or severe cognitive impairment we have reported the prevalence of behavioural and psychological symptoms, and the relationship between such symptoms and aspects of cognitive impairment, adjusting for a wide range of sociodemographic and health related risk factors. The prevalence of many symptoms increases with worsening cognition, the presence of a subjective memory complaint and impaired functional ability. Depression, anxiety and apathy were more common in those with a history of vascular disease whereas agitated behaviour is more common in those without.
Many symptoms were reported more frequently by those with MCI than by those who were cognitively impaired without meeting the definition of MCI (the OCIND group), the majority of who failed to meet the criteria for MCI owing to the absence of a subjective memory complaint. Within the MCI group, those without memory impairments (N-MCI) typically suffered fewer symptoms than those with only memory (A-MCI) or with memory and non-memory cognitive impairment (M-MCI). However, in multivariate analysis adjusting for subjective memory complaint and cognitive function, those with an objective memory disorder were not found to be at higher risk. Age and sex also affected the prevalence of psychiatric conditions, with depression more common in women and in those under 75 years of age, while apathy was more frequently reported in men, echoing our earlier findings in the population with dementia [ 4 ]
Strengths and limitations of the study
We assessed the prevalence of behavioural and psychological symptoms with respect to a wide variety of socio-demographic, cognitive and health-related factors, allowing a detailed investigation of the associations between symptoms and the factors that comprise the definitions of MCI. Our assessment of symptoms has been previously used to report the prevalence and correlates of behavioural and psychological symptoms of dementia in demented and non demented groups, however it is difficult to compare prevalences of symptoms here with those reported using the informant based Neuropsychiatric Inventory (NPI) [ 15 ]. Our study was population-based and our estimates were therefore not biased by referral filters. Prevalences were backweighted to account for the study design being representative of the older population of England and Wales without dementia. Symptoms were assessed using interviews conducted with both study participants and their carers, combined with observations made by the interviewer during the course of the interviews. However, neither the severity of symptoms nor the extent to which they were problematic to participants was recorded.
Wandering and confabulation were rare in the non-demented sample and very few individuals lived in institutions making associations between these and other factors difficult to detect. Although the response rate of the study is high at 82%, variables were missing for some subjects. The primary cause of a missing MMSE score in our study is severe cognitive impairment and we excluded individuals with missing MMSE from this investigation. Symptom presence was determined on the basis of any evidence being present from screen or assessment interviews. Where no evidence was present, symptoms were assumed to be absent. Covariate data were missing in only a small proportion of cases and where missing were assumed to be missing at random.
Depression
Depression has a complex relationship with dementia and cognitive impairment [ 16 ]. It is known that depression is a risk factor for dementia [ 17 ], but also that depression is a direct cause of cognitive impairment that may not indicate incipient dementia [ 18 ]. Depression is further acknowledged as a symptom of dementia with a high incidence in the early stages of dementia [ 4 ]. Depression is the most commonly studied non-cognitive correlate of MCI, with estimates of the prevalence of depressive symptoms ranging from 16-40% in population based studies [ 6 ]. The varying instruments and threshold for assessment of depression make cross study comparison difficult. The population based Three-City Study found that 16% of 2879 individuals with MCI had some depressive symptoms, 9.2% sub-threshold depression and 2.4% had DSM-IV clinical depression [ 19 ]. We found depression to be most common in those with the lowest MMSE scores (17.6%) and more common in MCI (14.5%) than in those with no cognitive impairment (6%) or other cognitive impairments (8.5%).
Relationship between behavioural and psychological symptoms with cognitive impairment
A subjective memory complaint is the strongest independent predictor of reported symptoms in multivariate analyses. This might be explained by behavioural and psychological symptoms causing memory impairments to be noticed, or a generally high awareness of both cognitive and non-cognitive symptoms. Subjective memory complaints are commonly reported by older people, although previous studies have shown depression to be a stronger predictor of subjective memory complaint than objective cognitive impairments [ 20 ]. Subjective memory complaints are required for a clinical diagnosis of MCI, and so those diagnosed with MCI are more likely to report behavioural and psychological symptoms than those who are similarly cognitively impaired in the population.
We found objective memory impairment to be associated with many symptoms in univariate analysis, and some evidence for differing symptom profiles across MCI subtypes, although the power to detect these differences was limited. Hallucinations were significantly more common in A-MCI (7.6%) and M-MCI (8.8%) than N-MCI (2.4%) which contradicts the findings of a clinical study of 120 subjects [ 21 ] in which hallucinations were more prevalent in those with N-MCI (6 of 26) than in A-MCI/M-MCI (4 of 94). In the current population-based study the numbers with amnestic and non-amnestic subtypes are more equal with 204 A-MCI/M-MCI and 185 N-MCI which suggests that N-MCI may be under-represented in the clinic and that individuals with N-MCI presenting to clinical services may not be representative and are those with more behavioural and psychological problems.
Lower general cognitive function has been consistently associated with a higher prevalence of behavioural and psychological symptoms in population based studies of mild to moderate dementia [ 4 , 22 ] and in MCI [ 23 ]. In the present study most symptoms were more prevalent in those with low MMSE scores. In multivariate analysis the risk of apathy, persecution and misidentification was significantly increased in persons with a lower MMSE score after adjustment for subjective memory complaints and objective memory impairment. The risk of anxiety was associated with a higher MMSE score.
Correlates of behavioural and psychological symptoms
In univariate analyses, being over 75 years of age had little effect on symptom prevalence. However, in multivariate analyses after adjusting for cognitive function younger participants had a significantly increased risk of depression, feelings of persecution and agitated behaviour, echoing previous findings in dementia [ 4 ]. Female subjects were significantly more likely to report depression in common with previous findings [ 24 ] while contrary to a previous study of 214 MCI patients we found that men were more likely to report apathy [ 25 ], this difference possibly owing to cross study differences in the assessment of apathy.
In those with vascular disease the onset of cognitive impairments may signify early vascular or mixed dementia, with potentially a different profile of behavioural and psychological symptoms. A review of behavioural symptoms in different dementia subtypes found an increase in the risk of depression, emotional labiality, anxiety and apathy in vascular dementia compared to AD, while delusions, delusional misidentification, wandering and restlessness were less frequent in vascular dementia compared with Alzheimer's disease [ 26 ]. In our study of the population without dementia the presence of vascular disease significantly increased the risk of depression, apathy, anxiety, irritability and sleep problems, with these associations moderately attenuated after adjusting for cognitive impairment. After adjustment for cognitive impairment agitated behaviour was negatively associated with a history of vascular disease, suggesting an association with non-vascular sources of early cognitive change.
Consistent with previous studies [ 23 , 27 - 29 ] we observed a positive, significant relationship between ADL impairment and depression, apathy, anxiety, persecution, hallucination, irritability, sleep problems and misidentification. While the first definitions of MCI excluded those with ADL impairments [ 30 ], more recent clinical definitions are more inclusive allowing impairment in complex or instrumental ADLs [ 31 ] although those with severe ADL impairments may be excluded from a diagnosis.
Co-occurrence of symptoms
The relationships between some behavioural and psychological symptoms in dementia are well known [ 32 , 33 ], but there are few reports of their co-occurrence in early cognitive impairment [ 34 ]. We report a moderate statistically significant association between many pairs of symptoms, which remained after adjusting for cognitive and socio-demographic factors. Factor analysis identified three factors corresponding to misidentification and hallucination; apathy, anxiety and depression; and feelings of persecution and irritability. However, the high uniqueness of each symptom and the significant co-occurrence that was not represented in the factor structure indicates that most of the variance cannot be explained by these factors. In a previous study of the population with dementia in MRC CFAS, similar factors but with stronger patterns of co-occurrence were observed [ 4 ]. | Conclusions
Recent reviews have demonstrated large variation in estimates of the prevalence of behavioural and psychological symptoms in MCI [ 6 , 7 ]. This variation can in part be explained by different diagnostic criteria for MCI and use of different study settings. Most symptoms are less prevalent in the population than in clinical samples [ 6 ]. We have shown that behavioural and psychological symptoms are related to many of the criteria used in inclusion and exclusion criteria for MCI diagnosis. In particular, those with subjective memory complaints were more likely to report symptoms independently of objective measures of cognitive function and memory impairment. Behavioural or psychological problems may precipitate contact with medical services, which may contribute higher prevalence in tertiary referral centres compared with the general population.
Behavioural and psychological problems are common in dementia [ 4 ] and have become accepted as central characteristics of the disorder. Behavioural and psychological symptoms in dementia are at least as problematic for patients and caregivers as cognitive impairments, significantly affecting quality of life and cost of care of people with dementia [ 1 , 2 ]. Furthermore, they currently offer greater opportunities for intervention and management than does cognitive impairment [ 3 ]. Current definitions of MCI focus entirely on cognition and may exclude those with psychiatric symptoms on the basis that psychiatric disorders might underlie cognitive impairment which should then not be considered an indicator of incipient Alzheimer's disease. Yet we have shown that many behavioural and psychological symptoms are present in those with mild cognitive impairments with a similar pattern of occurrence to that seen in individuals with dementia. Behavioural and psychological symptoms should be assessed as possible targets for management in cognitively impaired older people. Several studies in patients with MCI have shown that those with behavioural and psychological symptoms have an increased risk of dementia incidence and suggest that non-cognitive symptoms should be a consideration when identifying those in the earliest stages of dementia [ 6 , 7 ]. It remains difficult to differentiate patients with psychological symptoms as a consequence of early dementia from those in whom cognitive impairment is secondary to other psychological conditions. Further population-based longitudinal studies are needed to establish whether behavioural and psychological symptoms can be used alongside memory and other cognitive impairment to improve the identification of those at highest risk of dementia incidence. | Background
Behavioural and psychological symptoms are associated with dementia, but are also present in a significant number of the older population without dementia. Here we explore the distribution of behavioural and psychological symptoms in the population without dementia, and their relationship with domains and severity of health and cognitive impairment.
Methods
The Medical Research Council Cognitive Function and Ageing Study is a two-phase longitudinal study of ageing representative of the population aged 65 and over of England and Wales. A subsample of 1781 participants without a study diagnosis of dementia was included in this study. Information on symptoms including depression, apathy, anxiety, feelings of persecution, hallucination, agitated behaviour, elation, irritability, sleep problems, wandering, confabulation and misidentification, cognitive function, health related factors and socio-demographic information was extracted from interviews with participants and knowledgeable informants. Participants were classified according to the Mini-Mental State Examination and by criteria for subtypes of mild cognitive impairment (MCI). The prevalence of behavioural and psychological symptoms and associations with cognitive function, health and socio-demographics was examined. Co-occurrence of symptoms was tested using factor analysis.
Results
Most symptoms were reported more frequently in those with more severe cognitive impairment. Subjective memory complaints were the strongest independent predictor of reported symptoms, and most were reported more often in those classified as having MCI than in those with cognitive impairments that did not meet the MCI criteria. The pattern of co-occurrence of symptoms is similar to that seen in dementia.
Conclusions
Our results highlight that behavioural and psychological symptoms are prevalent in the cognitively impaired older population, and partly explain the variation observed in previous cohorts of individuals with MCI. Behavioural and psychological symptoms offer a target for intervention and so are an important consideration in the assessment of cognitively impaired older people. | Competing interests
The authors declare that they have no competing interests.
Authors' contributions
RL, GS and BS designed and conducted the analysis and drafted the manuscript. All authors provided scientific input into the design of the analysis, the interpretation of findings and made scientific revisions to the manuscript.
Pre-publication history
The pre-publication history for this paper can be accessed here:
http://www.biomedcentral.com/1471-2318/10/87/prepub
Supplementary Material | Acknowledgements
We would like to thank all participants and investigators of the MRC Cognitive Function and Ageing Study. MRC CFAS is funded by the Medical Research Council (grant number G9901400). BCMS is funded by the Joint European Post-Doctoral Programme: The European Research Area in Ageing (ERA-AGE) Network FLARE Programme. RL received a stipend from Alzheimer Nederland. None of the sponsors influenced the design or conduct of the study or the analysis or interpretation of the findings. | CC BY | no | 2022-01-12 15:21:37 | BMC Geriatr. 2010 Nov 30; 10:87 | oa_package/ee/e4/PMC3014957.tar.gz |
PMC3014958 | 21167038 | Background
Increased medical co-morbidity is commonly associated with the aging process [ 1 , 2 ]. There are many reasons for this. For example, age is a risk factor for many common diseases such as diabetes and chronic obstructive lung disease. According to previous research, medical co-morbidity rates tend to be higher among patients living in large urban centres, often due to low socioeconomic status and poorer access to health services [ 3 ]. High rates of medical co-morbidity have also been found to be associated with a number of adverse outcomes, including impaired function in activities of daily living, increased utilization of acute care services, and increased mortality [ 2 , 4 , 5 ]. The question of the impact of medical co-morbidity on cognitive function is a complex one. It is generally known that having multiple medical disorders may be associated with impaired brain function, partly due to the direct neurotoxic effects of disease on the brain and partly due to the possible neurotoxic effects of medications used to treat these diseases [ 6 ]. Patients with low SES may be particularly at risk [ 7 , 8 ]. In spite of this common assertion, there have been few studies that have systematically evaluated the impact of medical co-morbidity on brain function and no studies that have controlled for SES.
Of the existing studies that have examined the effect of medical co-morbidity on cognitive function, the majority have focused on objective function as opposed to subjective function. Lyketsos et al (2005) [ 9 ] examined a community sample and discovered medical co-morbidity was much higher among individuals with dementia. Cullum et al [ 10 ] examined a sample of older medical inpatients and discovered a strong correlation between medical co-morbidity and depression but surprisingly no significant correlation with cognition. It is possible that the lack of an association found in this study may be attributable to the low sensitivity of the cognitive instrument used (AMTS) in picking up changes in cognitive functioning. Mariani et al [ 11 ] examined a sample of patients with amnestic MCI (aMCI) and discovered cognitive functioning was a strong predictor of IADL performance while co-morbidity was not, possibly due to the fact that traditionally changes in ADL performance are more strongly linked to medical co-morbidity. Salvi et al [ 12 ] examined a cohort of acutely ill medical inpatients and found a strong correlation between medical co-morbidity, cognitive functioning, length of stay, depression and disability. Many studies have looked at the link between medical co-morbidity and depression in older adults, demonstrating a strong association [ 13 - 15 ]. However, no studies to date have explored how SES might impact on this association.
Subjective memory loss may or may not correlate with objective memory function based on previous research [ 16 , 17 ]. What is clear is that depressive symptoms and personality traits may mediate part of this discrepancy. In other words, patients with depressive symptoms may have elevated subjective memory complaints in the face of normal cognition [ 18 , 19 ]. The opposite pattern (normal subjective memory complaints in the face of objective memory function) may be observed in patients with poor frontal lobe function, such as patients with vascular dementia or advanced dementia [ 20 , 21 ]. Patients with elevated medical co-morbidity are at risk on both counts. First, medical co-morbidity has been found to be associated with high levels of depression [ 13 - 15 ]. Second, medical co-morbidity is often associated with dementia and cognitive dysfunction [ 9 ].
The objective of this study was to examine the impact of medical co-morbidity and other SES factors on subjective and objective memory function in a sample of patients with mixed diagnoses (dementia, MCI, normal cognitive functioning) referred to a Memory Disorders Clinic. As a result of it's geographic location the clinic sees patients with both high and low SES, making it an ideal environment to study the effects of SES. Furthermore, patients are referred to the clinic by their family doctor and/or specialist often with a complaint of memory loss but not necessarily objective findings. Thus, the base rates of memory complaints tends to be high and it provides an ideal opportunity for studying how SES and associated factors may modify the relationship between subjective and objective cognitive function in a memory clinic setting. We hypothesized based on the literature that patients referred to our Memory Clinic with low SES, increased medical co-morbidity, low education, and the presence of depression would be more likely to have objective cognitive impairment but not necessarily subjective memory loss. | Methods
Data from 85 patients aged 50 or above attending a Memory Disorders Clinic located in a large urban centre was examined retrospectively. The sample consisted of community dwelling patients with a variety of diagnoses (dementia, MCI, cognitively normal) referred by their family physicians with a complaint of memory loss. While patients were all medically stable many were on a number of medications. The study was approved by the St. Michaels Hospital Research Ethics Board. | Results
In terms of descriptive statistics (see Table 1 ), the mean age of the sample was 69.2(10.0) years, the mean level of education was 14.0(3.5) years and the mean CIRS score was 5.2(2.9), suggesting most patients had more than one chronic medical illness. Men and women were equally represented in the sample, and 45% had symptoms compatible with major depression. SES status based on residential income quintiles was evenly distributed across the subgroups. The mean MMSE score of subjects was 27/30 and the mean BNA score was 90/114. 22% of the sample was cognitively normal, 39% were demented (Alzheimer's disease, mixed dementia, dementia with lewy bodies, frontal-temporal dementia), 30% had mild cognitive impairment (MCI) and 9% had dementia secondary to some other cause (traumatic brain injury, lyme disease, cancer).
In terms of comparative statistics, medical co-morbidity as measured by the CIRS score correlated negatively with the MMSE score (p < .001), BNA (p < .05) and all BNA sub scores (see Table 2 ). Conversely there was no correlation between medical co-morbidity and subjective memory complaints as measured by the PAOF (p > .05). Furthermore, there was no correlation with any of the PAOF sub scores (p > .05). Education correlated, as predicted, with cognition as measured by the MMSE (p < .05), but surprisingly not with the BNA or any of the sub scores. Residential SES correlated negatively with overall subjective memory complaints (p < .001), and with two sub scores, (memory complaints (p < .001) and language complaints (p < .01)) while it correlated positively with executive function (p < .01). Depression did not correlate significantly with any measure of objective cognitive function (p > .05) but positively with overall subjective general complaints (p < .001), memory complaints (p < .001) and language complaints (p < .05) with the exception of complaints in motor function (p > .05). Given that age may be a covariate or suppressor variable in the analyses, we reran the spearman rank correlations analyses with age as a co-variate and the results were similar except for a small change in the magnitude of the observed associations. In addition, we conducted multivariable linear regression to examine whether depression or subjective memory complaints are associated with objective memory/cognitive performance after the effects of medical co-morbidity are removed. We found no significant associations. We felt that it was important to run these additional analyses to ensure that the observed association between medical co-morbidity and objective memory impairment was not being driven by other confounders such as age, depression or subjective memory loss. | Discussion
We had predicted based on prior research that changes in objective cognitive function in our Memory Disorders Clinic sample would correlate strongly with increased medical co-morbidity, depression, low SES and low education. Surprisingly, decreased objective memory function correlated strongly with only medical co-morbidity, partially with education and not at all with residential SES. Subjective memory complaints did correlate with depression but not at all with medical co-morbidity and inversely with residential SES.
Prior research has suggested that SES correlates strongly with cognitive functioning [ 8 ]. It is possible in our study, given that we used residential SES as opposed to income levels as our measure of SES, that our sample size was not sufficiently large to detect differences. The weak link with education, however, is even more surprising given that low levels of education have been shown to be significantly associated with poor cognitive functioning. Education did correlate mildly with impaired objective memory performance as measured by the MMSE but did not correlate with a more specific measure of cognitive function, the BNA. Given that the sample looked at is quite elderly, it is possible that years of education is not a good proxy for level of intellectual function. The inverse relationship between subjective memory complaints and SES does suggest that patients with low SES tend to complain more about their memory function, the opposite trend from what would be observed using the theory of cognitive reserve.
Our study suggested a strong correlation between medical co-morbidity and objective memory function but a poor correlation with subjective memory function in our Memory Clinic sample. Thus, patients evaluated in our clinic with significant medical issues are clearly at risk for cognitive impairment. Subjective memory complaints are a measure of awareness of brain dysfunction. Awareness of such brain dysfunction is important as it may lead to the adoption of compensatory strategies, including use of calendars, aids, etc, which may enhance brain function. It may also make patients more likely to seek help or perhaps go on medications (such as cognitive enhancers) that may preserve their brain function. Previous studies have examined subjective memory complaints and have found a relatively poor correlation between complaints and actual cognitive performance [ 16 , 17 ]. This study contributes further to our understanding of the link between subjective memory complaints and objective memory function by demonstrating that this lack of association persists even in the context of significant medical illness.
It is interesting to speculate as to why subjective memory complaints would be low in patients with elevated medical co-morbidity, especially given the strong correlation with reduced cognitive performance. One potential explanation is that such patients may be so preoccupied with issues regarding their physical health that they have little awareness of cognitive issues. In addition, it is possible that it may have something to do with the theory of cognitive reserve [ 26 ]. According to this theory, it is possible that patients with high medical co-morbidity have low cognitive reserve to begin with and therefore experience cognitive changes in a more gradual way, resulting in poor awareness. Finally, it is possible that the relationship between medical co-morbidity and subjective memory complaints may be mediated by some other correlate, such as low SES or limited education. Our study in fact showed only a partial correlation with education and a negative correlation with SES, suggesting this is not likely the mechanism.
Interestingly approximately half of the patients in the clinic studied had a diagnosis of dementia while half did not. There is some evidence to suggest that patients with more severe cognitive impairment may be more inclined to underreport their cognitive impairment while the opposite is true of patients with less severe cognitive impairment. Thus, it is possible that diagnosis alone may explain to some degree the observed associations. Whatever the mechanism, the implications of these research findings are apparent. Patients evaluated in our Memory Disorders Clinic who are in a sense more impaired medically are less likely to be aware of their cognitive deficits, making them more vulnerable. Such patients may be less likely to seek help or develop compensatory strategies, resulting in elevated risk. At a clinical level, it may mean that such patient with high levels of medical co-morbidity need to be assessed differently, perhaps screened more rigorously for the presence of cognitive deficits, or the physician may need to make a greater effort to contact collateral sources. Medication compliance is an area where physicians need to be especially vigilant, as poor cognition may lead to poor compliance, resulting in a greater burden of medical illness.
This study has a number of limitations which should be discussed. First, all of the measurement scales used have some limitations. Specifically, the PAOF has not been validated in the elderly and the BNA is a compilation of tests that has not been correlated with other neuropsychological tests. However, the PAOF has been used in other cognitively impaired populations and the BNA has shown to be superior to the MMSE in detecting dementia. The CIRS is a good attempt to quantify medical co-morbidity but as with many scales may not capture the full impact of medical illness. Also, in our sample we looked at all patients and did not stratify based on diagnosis (cognitively normal, MCI, dementia) as we were more interested in looking at the effects of medical co-morbidity on overall subjective/objective memory function as opposed to diagnosis. Furthermore, the design of the study was cross sectional as opposed to longitudinal, the sample size was relatively small and patients were recruited from a Memory Disorders clinic as opposed to the community. Thus, the findings are not generalizable to other clinical settings such as the community, family practice setting, etc. Finally, while more sophisticated statistical methodology could be used to analyze the relationship between mood, medical co-morbidity and cognition we feel such analyses goes beyond the current scope of our paper given our sample size. | Conclusions
In spite of these limitations, this study raises a number of issues that warrant further exploration. While it is clear from our preliminary findings that medical co-morbidity leads to low complaints and increased cognitive dysfunction, the precise mechanism of this is less clear. In addition, there are a number of health policy implications of our findings. Patients with high levels of medical co-morbidity should be screened more carefully for the presence of cognitive impairment and offered resources to help cope with this. Future studies should examine what the potential causes for our findings are and also look at how these research findings can be applied. | Background
Medical co-morbidity may be associated with impaired cognitive function based on prior studies. However, no studies to date have determined to what extent this association is linked to medical illness or other factors that may be linked to medical illness (such as education, income levels, depression or subjective memory loss). The present study examined how medical co-morbidity, socioeconomic status (defined as residential SES), education and depression are associated with subjective and objective memory function in a sample of patients recruited from a university affiliated Memory Disorders Clinic located in a large Canadian inner city teaching hospital.
Methods
Data was collected from 85 consecutive referrals to an Inner City Memory Disorders Clinic including socio-demographic characteristics, cognitive status and medical co-morbidity. Descriptive and correlational analyses were conducted.
Results
Impaired objective cognitive function correlated significantly with increased medical co-morbidity and partially with education but not with residential SES or depression. Elevated memory complaints correlated significantly with depression, inversely with residential SES and not at all with medical co-morbidity or education.
Conclusions
Increased medical co-morbidity is significantly associated with impaired cognitive performance but not with subjective memory complaints in an Inner City Memory Clinic sample. | Measures
Residential SES
We chose to define SES based on neighbourhood income quintiles. Prior studies that have used residential SES as opposed to personal SES have typically been done on a much larger scale and have been used to determine the effect of neighbourhood SES on the prevalence of dementia, etc. We elected to use residential SES as opposed to personal SES (ie income levels) in order to maximize our clinical sample. Neighbourhood income quintiles were computed using existing census data and divided the region of Toronto into five quintiles. The quintiles ranged from one (lowest income) to five (highest income) and for the purposes of analysis we pooled the two lowest income groups and the three highest income groups with an income less than $35,000.00 being defined as the cut off for low versus high. For the purposes of our study we have decided to separate out education from SES.
Cumulative Illness Rating Scale (CIRS)[ 22 ]
Medical co-morbidity was measured using the cumulative illness rating scale, geriatric version (CIRS-G). In this scale the body is divided into 13 different systems corresponding to 13 different organ areas. Based on the nature of the illness a severity score is assigned ranging from 0 (no impairment) to 4 (severe impairment). Certain illnesses that affect multiple organ systems may in some circumstances be coded more than once. The total score is then achieved by adding up the sub-scores for each organ system and indicates overall medical burden.
Patient Assessment of Own Function (PAOF)[ 23 ]
Subjective memory complaints were measured using the Patient Assessment of Own Function (PAOF) [ 23 ]. The Patient Assessment of Own Functioning (PAOF) is a 33 item tool designed to measure subjective memory complaints. It includes questions about a variety of cognitive functions including memory, language and communication, sensory motor skills and higher level intellectual functions. Participants are asked to respond while thinking about the last two months and to rate their complaints on the 6 points scale. While it has not been validated in older patients with dementia it has been used in a number of cognitively impaired populations including those with psychiatric problems, HIV patients and those with chronic medical illnesses.
Behavioural Neurology Assessment [ 24 ]
The BNA is a clinician administered cognitive test that measures multiple cognitive domains including language, attention, visual spatial function, naming and executive function. It is scored out of 114, with the cut off for dementia being low 80. It has been validated in patients with dementia and has been demonstrated to be superior to the MMSE in detecting dementia [ 24 ].
Folstein Mini-Mental Status Examination [ 25 ]
The MMSE is a measure of global memory function [ 25 ]. It is scored out of 30 and measures a variety of cognitive functions including recall, orientation, visual spatial function, etc. It has been extensively used in dementia trials, with the cut off for dementia being 26 or less.
Diagnosis of Dementia/Depression
A diagnosis of dementia was made following comprehensive assessment by a behavioural neurologist and geriatric psychiatrist in accordance with DSM-IV criteria. All patients had brain imaging and routine blood work to rule out reversible causes of dementia as well as cognitive testing using the BNA and MMSE. While cut off scores were not routinely used, most patients with dementia scored below 26 out of 30 on the MMSE or below 80 on the BNA. Depression was diagnosed using DSM-IV criteria following comprehensive assessment by a geriatric psychiatrist.
Statistical methods
Continuous variables were summarized in the descriptive analyses using the mean, standard deviation and range. Categorical variables and ordinal variables were summarized using the number (percent) in various categories. Spearman rank correlations were used to examine the impact of education, socioeconomic status, depression and medical co-morbidity on subjective and objective memory function.
To avoid problems with confounders we ran two additional analyses. The first was a repeat of the spearman rank correlation analysis adjusting for the effects of age. The second was a multivariable linear regression analysis to determine if the effects remained when medical co-morbidity was removed from the analyses. Statistical analyses were performed using SAS Version 9.2 (SAS Institute, Cary, NC). All p-values were two-tailed and a p-value less than 0.05 was regarded as being significant.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
CF was responsible for writing the article and assisting in interpretation of the data analyses. DJ was responsible for analyzing the data, defining the statistical approach and writing the statistical section of the manuscript. TS was responsible for critiquing the article, assisting with interpretation of the data analyses and aiding with the scientific design of the study. All of the authors have read and approved the final manuscript.
Authors' information
CF is an adjunct scientist in the Li Ka Shing Knowledge Institute at St. Michael's Hospital in Toronto with an academic appointment of assistant professor at the University of Toronto Department of Psychiatry. DJ is a senior scientist and statistician at the University of Manitoba. TS is a scientist at the Li Ka Shing Knowledge Institute, St. Michael's Hospital and an assistant professor in the Department of Neurosurgery, University of Toronto.
Pre-publication history
The pre-publication history for this paper can be accessed here:
http://www.biomedcentral.com/1471-2318/10/89/prepub | Acknowledgements
Funding for personnel involved in abstracting the data and conducting preliminary analyses was provided by the Alzheimer society of Canada and the Heather and Eric Donnelly endowment, St. Michael's Hospital Foundation, St. Michael's Hospital. | CC BY | no | 2022-01-12 15:21:37 | BMC Geriatr. 2010 Dec 17; 10:89 | oa_package/26/d4/PMC3014958.tar.gz |
PMC3014959 | 21114857 | Background
Environmental change driven by anthropogenic activities is causing unprecedented rates of species extinction, presenting a major threat to global biodiversity [ 1 ]. Among vertebrates, all classes have suffered high extinction rates, but amphibians have been most severely impacted. Based on recent estimates, more than one-third of the worlds amphibians are threatened with extinction [ 2 , 3 ], and almost one-half of the remaining species are in a state of decline [ 2 ]. In response to this crisis, the international Amphibian Conservation Action Plan (ACAP), devised in 2005, urged the establishment of captive assurance colonies for threatened species [ 4 , 5 ]. In accordance with this recommendation, a large number of institutions worldwide have initiated captive breeding programmes for declining and endangered amphibians. However, in almost all cases, breeding attempts have failed due to the inherent difficulties associated with simulating the complex combination of social and environmental factors that trigger amphibians to breed [ 6 , 5 ]. In reaction to this captive breeding crisis, there has been a growing interest in determining whether threatened amphibians can be propagated, and genetically managed, using assisted reproductive technologies (ART)[ 5 , 7 , 8 ].
One component of ART is the artificial manipulation of reproductive events using exogenous hormones. Specifically, males and females are administered hormones to stimulate the production and release of gametes (spermatozoa and oocytes), which are then used to generate embryos via in vitro fertilisation (IVF), also referred to as artificial fertilisation (AF) [ 5 ]. Among anurans (frogs and toads), it has been known for several decades that exogenous gonadotropins, and gonadotropin-releasing hormones, can be used to induce both sperm release (spermiation) and oocyte release (ovulation) [ 9 - 12 ]. Pituitary extracts are an effective source of amphibian gonadotropins and such preparations have been successfully used to induce gamete release in various anuran species, including Bufo arenarum , Rana pipiens , Hyla regilla and Eleutherodactylus coqui [ 9 , 10 , 13 - 16 ]. However, the use of pituitary extracts is now strongly discouraged due to a high risk of pathogen transmission [ 5 , 7 ], and because production of pituitary preparations requires the euthanasia of large numbers of reproductively mature anurans [ 17 ]. An alternative approach is the use of synthetic hormones, in particular, commercially available analogues of luteinizing-hormone releasing hormone (LHRH) and human chorionic gonadotropin (hCG) [ 8 , 17 - 19 ].
Luteinizing-hormone releasing hormone is a hypothalamic hormone that acts by stimulating the anterior pituitary to synthesize and release natural luteinizing hormone (LH), which in turn stimulates gonadal activity. In contrast, hCG acts by mimicking LH, due to identical alpha and shared beta subunits [ 20 ], bypassing the hypothalamic-pituitary-gonadal axis to exert a direct influence on the gonads [ 21 ]. Evidence that synthetic analogues of LHRH and hCG can successfully induce gamete release has been obtained for a broad diversity of anuran species [ 7 , 17 , 18 , 22 - 26 ], but the efficacy of these hormones has been found to vary considerably. For example, LHRHa is significantly more effective than hCG at stimulating ovulation in Eleutherodactylus coqui [ 17 ], but hCG is highly effective at stimulating ovulation in Xenopus laevis [ 26 ].
Although there is still much to learn about the relative efficacy of hCG and LHRHa across species, baseline knowledge concerning their potency has permitted IVF to be attempted in a small, but growing number of anuran families, including the bufonidae [ 8 ], pipidae [ 26 ] and myobatrachidae [ 27 ]. Surprisingly, however, almost all anuran IVF studies conducted to date have refrained from artificially fertilizing oocytes using hormonally induced sperm, instead opting to use sperm obtained from testes macerates [ 13 , 27 - 29 ]. The benefit of conducting IVF using testes macerates is that sperm can be obtained in high concentrations, and can also be acquired at the exact moment when females begin ovulating, eliminating the need for sperm storage [ 19 ]. Given these practical advantages, euthanizing males for the purpose of IVF might be useful in common species, but this approach cannot be justified in endangered species where individual animals are of high genetic value [ 5 , 7 ].
To date, few studies have attempted IVF in anurans using hormonally induced gametes collected from live animals, and outcomes have been highly variable. For example, Waggener and Carrol [ 22 ] achieved 100% fertilisation success in the leptodactylid frogs Lepidobatrachus laevis and L. illanensis [ 22 ], but Browne et al. [ 8 ] reported a mean fertilisation success of 12.7% in the endangered toad Bufo baxteri . This extreme variance in IVF success, which is probably related to species-specific differences in reproductive mode and physiology, suggests that protocols for combining hormonally induced gametes are not readily transferable between species [ 19 ]. Given this inherent level of unpredictability, there is a need to develop techniques for collecting and combining gametes obtained from live individuals. In particular, there is an urgent need to develop these protocols for endangered species [ 7 , 8 ].
The southern corroboree frog Pseudophryne corroboree , is one of Australia' s most critically endangered frog species [ 30 ]. Of the species known to be extant in Australia, P. corroboree was ranked by the Australasian Regional Association of Zoological Parks and Aquaria as the highest priority anuran requiring ex-situ conservation [ 31 ]. The distribution of P. corroboree is highly restricted, with the species confined to a linear distance of 51 km within subalpine regions of Koscuiszko National Park [ 32 ]. According to recent field surveys, there are currently less than 50 individuals remaining at natural breeding sites, and the species is predicted to go extinct in the wild within the next ten years [ 33 ]. Population declines in P. corroboree were first observed in the 1980's, and since 1996 the species has been the focus of an intensive management and recovery programme [ 33 , 34 ]. Early management of the species concentrated on habitat protection, but more recently efforts have turned towards the establishment of captive assurance populations [ 33 , 35 ]. At present, multiple populations of P. corroboree are being maintained in zoos and biological institutions throughout Australia, but successful breeding and recruitment in captivity has been limited. Despite this critical situation, there has been no attempt to bolster captive breeding activity using assisted reproductive technologies.
The aim of this study was twofold. First to test the efficacy of using LHRH and hCG to induce spermiation, and LHRH to induce ovulation, in captively reared P. corroboree , and second, to test whether hormonally induced sperm and oocytes can be used to generate embryo's via in-vitro fertilisation (IVF). | Methods
All research was conducted in compliance with the Monash University Animal Ethics Committee (AEC), permit number BSCI/2009/27.
Study population
Frogs were obtained from a captive colony maintained at the Amphibian Research Centre (ARC) in Melbourne Victoria. The captive colony was established by collecting natural egg clutches from multiple populations throughout the species range during the 2004 and 2005 breeding seasons. The eggs were reared to maturity and housed in indoor terrariums (150 × 43 × 38 cm) exposed to seasonal fluctuations in temperature and photoperiod that mimicked natural conditions. Frogs used in this study were approximately 4-5 years old because sexual maturity in P. corroboree is not reached until 3-4 years post-metamorphosis [ 36 ]. In total, the study involved 24 males and 25 females, but one male was not treated. In P. corroboree phenotypic traits do not provide reliable indicators of sex. Therefore, prior to commencement of the study all frogs were genetically sexed using Amplified Fragment Length Polymorphism (AFLP) analysis, with the presence of a 66 bp DNA marker (representing a Y chromosome sequence) diagnostic for males [ 37 , 38 ].
Frogs were collected from the ARC and transported to Monash University (Clayton campus) where they were held for the duration of the study (January 29-April 9, 2009). On the day of collection, frogs were weighed and snout-vent length measured before being randomly assigned to same sex groups (n = 6 frogs per group, with the exception of one female group that contained 7 frogs). Groups of individuals were housed in plastic enclosures (360 mm × 200 mm × 170 mm), each containing a layer of fine gravel (~15 cm thick) covered with a layer of sphagnum moss (~5 cm thick). Once a week, containers were flushed with approximately 2 L of deioinized water and frogs were fed ten-day old crickets (~200 per container). All containers were kept in a constant temperature room maintained on a 17°C/12°C day/night temperature cycle and a 14.5 h/9.5 h light/dark cycle.
Hormonal induction of spermiation
Males were administered a single dose of either 20 IU per gram bodyweight hCG (chorolon ® ) (n = 6 males) or 5 μg per gram bodyweight LHRHa (Leuprorelin oxo-Pro-His-Trp-Ser-Tyr-[d-leu]-Leu-Arg-Pro-NHEt : Lucrin ® ) (n = 11 males). These doses were selected because they approximate doses previously found to induce gamete release in anurans [ 5 , 22 , 24 , 39 ]. Hormones were diluted in 100 μL of Simplified Amphibian Ringer (113 mM NaCl, 2 mM KCl, 1.35 mM CaCl 2 , 1.2 mM NaHCO 3 ) and administered via subcutaneous injection into the dorsal lymph sac. As a control for the injection and handling procedures, a third group of frogs (n = 6) were administered 100 μL of Simplified Amphibian Ringer (SAR). Following hormone administration, frogs were returned to plastic holding tanks (50 mm × 90 mm) containing moist sponge hydrated with 5 mL of distilled water. Under these conditions, frogs were sufficiently hydrated to permit urine collection at each of the sampling times.
Spermic urine was collected at 3, 7, 12, 24, 36, 48, 60 & 72 h post hormone administration (PA). The collection method involved gently inserting the end of a glass microcapillary tube (fire polished and cooled) into the cloaca to stimulate urination. Immediately post collection, the volume of urine collected was measured in microlitres (μL), and the sample was then homogenized with 5 μL of a 1:50 dilution of SYBR-14 (Invitrogen L-7011), and then incubated in the dark for 7 min. Following this, a 2 μL aliquot of propidium iodide (PI) was then added and the solution was incubated in the dark for a further 7 min. SYBR-14 and PI are membrane permanent DNA stains that are commonly used to quantify sperm number and viability in anurans [ 39 , 40 ] and other vertebrates [ 38 ]. SYBR-14 specifically stains the DNA of live (viable) sperm, while PI specifically stains the DNA of dead (non-viable) sperm. Under UV light, live sperm (stained with SYBR-14) fluoresce bright green, while dead sperm (stained with PI) fluoresce bright red [ 41 ]. Immediately after staining, wet mount slides were prepared and the viability of sperm was evaluated within 30 min using fluorescent microscopy at a wavelength of 490 nm. For each sample, we calculated the total sperm count, sperm concentration (number of sperm/urine volume (μL) × 1000), and sperm viability (the proportion of live/total sperm).
Hormonal induction of ovulation
Females were randomly allocated to one of two treatment groups; a hormone treatment (n = 17 females) administered LHRHa (Lucrin ® ), or a control treatment (n = 8 females) administered Simplified Amphibian Ringer (SAR). For the hormone treatment, a stock solution of Lucrin ® was diluted in Simplified Amphibian Ringer to produce a final concentration of 100 μg mL -1 . Females received an anovulatory dose of 1 μg LHRHa per gram bodyweight diluted in 100 μL of SAR administered via subcutaneous injection into the dorsal lymph sac. This dosage was administered to prime the ovary without inducing ovulation [ 8 , 42 ]. Twenty-six hours after administration of the 'priming dose', each female received an ovulatory dose of 5 μg LHRHa per gram bodyweight. For the control treatment, females were administered 100 μL of SAR in place of the priming and ovulatory doses. Following treatment, individual females were placed into plastic enclosures (200 mm × 120 mm × 90 mm) lined with moist sponge and sphagnum moss. Individual animals were removed from their holding tanks 12 h PA of the ovulatory dose and stripping (expulsion of eggs from the oviducts) was attempted. Stripping was facilitated by holding a frog with its legs extended and gently applying pressure to the abdomen in a craniocaudal direction [ 21 , 43 ]. Stripping was attempted every 12 h ± 0.5 h for a period of five days.
In-vitro fertilisation (IVF)
At each sampling time, any eggs expelled from a female were placed in an individual dry Petri dish and IVF was conducted using available spermic urine samples. An aliquot of approximately 170 μL of pooled spermic urine was activated in approximately 100 μL 1:4 SAR. Sperm concentrations used for IVF ranged between 1.14 × 10 2 and 2.87 × 10 2 . The sperm solution was pipetted directly onto the oocytes and the dish was agitated for one minute. Each dish was enclosed within a petri dish and left to develop in a constant temperature room set to 10°C. Developing embryos were supplied with 100 μL of deionised water at 12 h, and a further 1000 μL at 24 h post fertilisation. Fertilisation success was calculated as the proportion of eggs at Gosner stage 4 to 6 [ 44 ] approximately 12 h post application of spermic urine to the oocytes. Embryonic development was checked every 6-12 h for a period of 7 days, and developmental stage quantified, using a stereo dissecting microscope.
Statistical analyses
The number of males that released sperm was compared between experimental treatments (LHRHa versus hCG), and between each experimental treatment and the control (hCG versus control, LHRHa versus control), using Fisher's exact tests. All comparisons were one-tailed due to the expectation of a positive treatment response. Comparison of mean total sperm count and mean total sperm concentration over the 72 h sampling period was made between treatments using Welch ANOVA's, due to unequal variance. Calculations of total sperm concentration only included sampling times in which sperm were released. Comparisons of sperm count and sperm concentration over time were made between treatments using repeated measures MANOVA's, with the main factors set as hormone treatment, and the within subject factors set as time. The MANOVA's were based on 6 equally spaced time intervals: 12 h, 24 h, 36 h, 48 h, 60 h and 72 h PA. Because the data violated the assumption of sphericity (Mauchly's test: P < 0.05), univariate analyses were corrected, and degrees of freedom adjusted, using the Greenhouse-Geisser method. The mean proportion of live sperm released over 72 h was compared between treatments using a student t-test. For all spermiation analyses, sperm counts were square root transformed, and sperm proportions were arcsine transformed. The number of females ovulating in response to hormone treatment (LHRHa injection) was compared to the control treatment (saline injection) using a one-tail Fisher's exact test. All statistical comparisons were performed using JMP software, with significance levels set at P < 0.05 . | Results
Hormonal induction of spermiation
No frogs (0/6) released sperm following injection of a saline control, but a high proportion of males released sperm following injection of hCG (83.3%, 5/6) and LHRHa (81.8%, 9/11). The number of males releasing sperm was significantly higher in response to hormone treatment compared to the control treatment (control versus hCG, Fisher's exact test: p = 0.007; control versus LHRHa, Fisher's exact test p = 0.002), but the number of responding males did not significantly differ between hormone treatments (hCG versus LHRHa, Fisher's exact test: p = 0.72). Males treated with hCG started releasing sperm within 3 hrs post administration (PA), and ceased sperm release after 48 h PA (Figure 1 ). In contrast, males treated with LHRHa, did not commence sperm release until 7 h PA, and at 72 h PA over 25% (3/11) of males were still releasing sperm (Figure 1 ). The highest proportion of males releasing sperm occurred between 12 and 48 h in response to hCG, and between 36 and 60 h in response to LHRHa (Figure 1 ).
The mean total number of sperm released over 72 h was more than 11 times higher in LHRHa treated males than in hCG treated males (Figure 2 ), and this difference was significant (Welch's ANOVA; F 1, 11.06 = 11.28, P = 0.006). There was no overall time effect on the number of sperm released (MANOVA: F 3.01, 36.15 = 1.42, p = 0.46) and no significant interaction between time and treatment (MANOVA: F 3.01, 36.15 = 1.05, p = 0.38), indicating that the number of sperm released did not differ between treatment groups over time. However, there was a significant treatment effect (MANOVA: F 1,12 = 8.14, p = 0.01), indicating that at each sampling period, LHRHa treated males generally released more sperm than hCG treated males (Table 1 ). Peak sperm release occurred at 12 h PA for hCG treated males, but not until 36 h PA for LHRHa treated males (Table 1 ).
Mean total sperm concentration was significantly higher in LHRHa treated males than in hCG treated males (Welch's ANOVA; F 1, 11.70 = 9.33, p = 0.01, Figure 3 ). Sperm concentration did not significantly differ over time (MANOVA: F 2.19, 26.32 = 2.68, p = 0.082), and there was no significant interaction between time and treatment (MANOVA: F 2.19, 26.32 = 1.79, p = 0.18), indicating that over time sperm concentration did not differ between treatment. However, there was a significant overall treatment effect (MANOVA: F 1, 12 = 18.31, p = 0.001), indicating that at individual sampling times sperm concentration was generally higher for LHRHa treated males than hCG treated males (Table 1 ). Peak sperm concentration occurred at 12 h for hCG treated males and 36 h for LHRHa treated males, corresponding with peaks in total sperm number (Table 1 ).
The mean total proportion of live sperm (sperm viability) released over 72 h, was almost 10% higher in LHRHa treated males than in hCG treated males (Figure 4 ), and this difference was significant (t-test; t = 2.31, df = 11, p = 0.04). Males treated with hCG released live sperm between 3 and 36 h PA, and over this period sperm viability was generally high (>60%), but at 12 h PA, which was the peak time for sperm release (maximum sperm number and concentration), average viability was lower than 35% (Table 1 ). For LHRHa treated males, live sperm were released between 7 and 72 h PA, and the sperm viability remained above 60% up until 48 h PA, after which time viability started to drop (Table 1 ). At the time of peak sperm production and concentration (36 h PA), the proportion of live sperm was higher than 80% (Table 1 ).
Hormonal induction of ovulation
No frogs (0/8) released eggs following injection of a saline control, and almost 30% (5/17) of females released eggs following injection of LHRHa, but this difference was not significant (Fishers exact test p = 0.116). The mean body size (SVL) of females that released eggs (mean ± SE = 28.3 ± 0.75) was not significantly different from females that did not release eggs (mean ± SE = 27.58 ± 0.48) (t = -0.793, df = 15, p = 0.44). There was also no significant relationship between female body size (SVL) and total clutch size (r 2 = 0.473, n = 5, p = 0.199). Of the females that responded positively to hormone treatment, all of them (5/5) released their clutches in discrete batches over 2-4 sampling times (Table 2 ). Females released between one and twelve eggs per batch, and average total clutch size was 15.2 ± 2.67 (Table 2 ). No female released eggs until 24 h PA, and only one female was still releasing eggs at 72 h PA (Figure 5 ). On average, females released the greatest proportion of their clutches between 24 and 48 h PA, with a peak in egg release at 36 h PA (Figure 5 ).
In-vitro fertilisation
On average, fertilisation success across females (n = 5) was moderate (mean ± SE = 54.72 ± 12.80%), but there was considerable variation between females (Table 2 ). Of the eggs that were fertilised, all commenced embryonic development, but no embryos survived beyond gastrulation [ 44 ]. The exact stage of failure varied between a female's egg batches, as well as between females (Table 2 ). | Discussion
Our study showed that hormone treatment induced gamete release in both male and female corroboree frogs, permitting us to test whether IVF can be used to augment captive breeding in this critically endangered species. Spermiation was induced following administration of either hCG or LHRHa, but these hormones were not equally effective. Both hormones led to approximately 80% of males releasing sperm, but there were significant differences in the speed and duration of response, and the number, concentration and proportion of live sperm released. Administration of hCG, led to a more rapid response, but LHRHa induced the release of a significantly higher number and concentration of viable sperm, and over a longer time period. This difference indicates that LHRHa was more effective at inducing spermiation in this species. That hCG and LHRHa elicited different responses was not unexpected because past research in anurans has shown that the relative efficacy of these hormones is highly species specific [ 5 ]. It is important to recognise, that as for the vast majority of anuran ART studies conducted to date, our study only tested each hormone at a single dose [ 5 ]. Therefore, to more thoroughly evaluate the relative efficacy of each hormone, it will be necessary to establish spermiation responses to hCG and LHRHa across a range of doses. Testing dose response relationships was beyond the scope of the current study, but this research will be incorporated into the recovery plan for the species.
Despite our finding that LHRHa was successful at inducing spermiation, hormone treated males did not release exceptionally high concentrations of sperm. On average, LHRHa treated males released approximately 4.5 × 10 3 sperm per millilitre, a value that is several orders of magnitude lower than hormonally induced sperm concentrations (4.0 × 10 5 to 4 × 10 7 sperm per millilitre) previously reported for anurans [ 22 , 24 , 39 ]. There may be several explanations why sperm concentrations were comparatively low in P. corroboree . First, the doses used may have been too low, or too high, to induce an optimal response. Even though we used concentrations approximating those found to be effective in a broad range of frog species [ 18 , 19 ], LHRHa may lack potency in P. corroboree. Another possibility is that males were not in prime physiological condition at the time of hormone treatment. In seasonally reproducing anurans such as P. corroboree , individuals typically coordinate their physiological state with environmental cues, which exert their effect by stimulating gonadotropin-releasing hormone neurons at the apex of the hypothalamus-pituitary-gonad axis [ 45 ]. In nature, spermatogenetic activity in P. corroboree commences several months prior to the onset of breeding [ 46 ], so if environmental changes that normally take place in spring (e.g. increasing photoperiod) were not suitably replicated in the captive environment, male investment in spermatogenic activity may have been limited prior to treatment. Another explanation for low sperm yield is that P. corroboree is not a species that invests heavily in spermatogenesis. Theoretically, anurans should only experience strong selection for high sperm production if male's either experience a high risk of sperm competition [ 47 ], are required to fertilise large egg clutches [ 48 ], or have exceptionally high mating rates [ 49 ], but all these conditions are absent in P. corroboree [ 50 , 51 ]. In fact, testes size relative to body size in P. corroboree is amongst the smallest reported in the family Myobatrachidae [ 46 , 51 ], so assuming that testis size reflects sperm production capacity in anurans [ 27 ], male P. corroboree might actually be incapable of producing high sperm yields.
For females, administration of LHRHa stimulated the release of multiple batches of oocytes over a 24 hour period, indicating that hormone treatment was effective at inducing ovulation, and that final egg maturation was asynchronous [ 52 ]. However, the average percentage of females responding was low (<30%), and the average clutch size (mean = approx 15 eggs) was at the lower end of the range (16-40 eggs) previously recorded for this species [ 50 ]. The suboptimal response might mean that LHRHa lacks potency in P. corroboree females, as was suggested for males (see above). Indeed, the ability of LHRHa to induce ovulation is known to vary considerably between anuran species. For example, doses required to reliably stimulate ovulation in the leptodactylid frog Eleutherodactylus coqui are 25 times higher than needed in the Wyoming toad Bufo baxteri [ 8 , 17 ]. Quantifying dose response relationships for female P. corroboree , as well as for males (see above), would provide valuable insight into the potency of LHRHa in this species. An alternative explanation is that phenotypic differences between test females influenced their responsiveness to hormone treatment. For example, in a recent study on boreal toads ( Bufo boreas boreas ), Roth et al. [ 23 ] showed that female age, body size and condition were all important factors influencing the efficacy of LHRHa treatment. However, in our study, females were of similar age (4-5 yrs) and there was no relationship between female body size and number of eggs released, so it is unlikely that these factors underpinned the variable responses reported. A more plausible explanation is that unresponsive females failed to ovulate because their oocytes were immature at the time of hormone treatment [ 42 ]. This explanation seems likely because post experimental tactile inspection of unresponsive females revealed that most individuals were still carrying small to medium sized oocytes.
In nature, female P. corroboree undergo rapid oocyte growth 4-8 weeks before the onset of breeding [ 46 ], so assuming that similar changes take place in captivity, our treatments may not have perfectly coincided with the time when females were undergoing the final stages of oogenesis. If females were treated too early, oocytes might not have been competent for ovulation [ 53 ], but if they were treated too late, oocytes may have already commenced reabsorption [ 46 ]. Increasing egg yield in P. corroboree may therefore require developing techniques for reliably assessing changes in oocyte growth and development [ 54 ]. Furthermore, it may be necessary to artificially accelerate and synchronise ova maturation prior to any attempt to induce ovulation. Past work with anurans has shown that this might be achieved by repeatedly administering females with low-dose injections of gonadotropins. For example, Browne et al. [ 8 ] reported that two priming injections of hCG, in combination with LHRHa, significantly increased the percentage of spawning females, the number of oocytes released, and the survival of fertilised eggs in the Wyoming toad Bufo baxteri .
Oocyte maturation in frogs might also be enhanced via in vitro or in vivo treatment with steroid stimulants [ 53 , 55 ]. The process of oocyte development in anurans has been extensively studied in Xenopus laevis and Rana pipiens and it is well established that the secretion of progesterone from late-stage follicles plays a fundamental role in germinal vesicle breakdown (GVBD) and oocyte maturation [ 55 - 63 ]. Furthermore, there is also experimental evidence to suggest that in vivo administration of progesterone can accelerate oocyte maturation and increase the effectiveness of hormone treatment. Specifically, Browne et al. [ 42 ] recently reported that administration of progesterone, in combination with LHRHa, significantly improved the number and quality of hormone-induced oocytes released by the toad Bufo fowleri . Based on these results, incorporating steroid treatment into future ART work with P. corroboree might be a valuable next step towards optimising ovulatory responses in this species.
Importantly, the hormone treatment protocols we employed permitted gametes to be collected and IVF to be attempted, indicating that there is real potential for ART to assist with the captive breeding of P. corroboree . However, the IVF trials resulted in variable levels of fertilisation success. Variable fertilisation success may have occurred because sperm concentrations were suboptimal. In previous anuran ART studies, concentrations of sperm resulting in high fertilisation success have been within the range of 5 × 10 5 to 1 × 10 6 sperm per mL [ 5 ], but we were restricted to using concentrations of less than 2.8 × 10 2 sperm per mL, which may have greatly reduced the probability of gamete fusion. Alternatively, variable fertilisation success may have resulted because the osmolality of the fertilisation medium was inappropriate, as has been reported in the Australian myobatractid frog Limnodynastes tasmaniensis [ 27 ]. During the study all males were kept well hydrated, so if their spermic urine was too dilute, this may have significantly reduced the osmolality of the fertilisation medium. If so, sperm may have been activated soon after urine collection and lost viability before IVF was attempted. Furthermore, if osmolality was too low this could have resulted in the large egg capsules swelling too rapidly, making it impossible for individual sperm to penetrate an egg cortex. Clearly, further work will be needed to identify the primary cause of low fertilisation success. It may also be necessary to investigate alternative IVF techniques, such as intra-cytoplasmic sperm injection (ICSI). This more sophisticated IVF approach has been trialled in Bufo arenarum [ 64 ] and Xenopus laevis [ 65 ], and provides a promising solution for achieving high fertilisation success when numbers of spermatozoa are limiting [ 7 ].
Of greater immediate concern than suboptimal fertilisation, is the result that all embryos failed during early development. Early embryo failure may have occurred because hormonal induction impaired gamete viability [ 5 ]. In most studies testing the feasibility of IVF in anurans, sperm has been obtained from testes macerates, and embryo survival has been high [ 5 , 13 , 27 - 29 ]. However, in a recent study in which hormonally induced sperm was used to conduct IVF in the endangered toad Bufo baxteri , levels of embryo survival were also low [ 8 ]. These results suggest that there might be problems associated with using spermic urine to fertilize hormonally induced oocytes. Potential reasons for this are not obvious, so this is an area that may require research attention. An alternative explanation for embryo failure is that the incubation conditions employed were inappropriate for this species. Terrestrial breeding Pseudophryne species have extremely large and gelatinous eggs whose capsule size and surface area is primarily determined by hydration state [ 66 ]. Consequently, the amount of water present during incubation will significantly affect rates of gas exchange, and subsequently, embryo growth and survival [ 67 ]. If eggs were kept too hydrated, or were hydrated too early in development, this may have resulted in oxygen limitation within egg capsules, which in turn could have resulted in embryo mortality [ 67 ]. However, this seems unlikely because our incubation conditions were similar to those previously used to successfully rear Pseudophryne guentheri embryos (Silla unpublished data). As such, we suspect that there may be a more intrinsic explanation for embryo failure.
In a recent study that used a cross classified breeding design to examine genetic compatibility in Pseudophryne bibronii , a sister species to P. corroboree , we discovered that crosses made between individuals from the same population had high embryo survival, but those made between populations experienced complete early embryo failure (Byrne and Silla, unpublished data). These results indicate that terrestrial toadlets may be susceptible to high levels of genetic incompatibility, as has been reported in other Australian anurans [ 68 , 69 ]. Critically, the frogs used in our study were all sourced from mixed populations, so there is a real possibility that embryo failure was indeed linked to developmental problems arising from genetic incompatibility. To address this potential problem, it will be necessary to conduct future IVF trials in P. corroboree using individuals derived from the same source populations. | Conclusion
In conclusion, this study demonstrated that exogenous hormones induced spermiation and ovulation in the southern corroboree frog P. corroboree . The hormone-treatment protocols used did not lead to exceptionally large numbers of gametes being released, but did permit the conduction of IVF, which resulted in moderate fertilisation success. Critically, however, all embryos failed during early stages of development. Embryo developmental failure may have occurred either because hormone treatment compromised gamete viability or because the incubation conditions employed were inappropriate. Alternatively, failure may have been linked to genetic incompatibility resulting from crosses being made between frogs sourced from several different populations. Additional work will be required to increase gamete yield for IVF and identify the causation of embryo failure. The findings are an important first step towards developing artificial reproductive technologies for assisting with the captive breeding of Australia's most critically endangered anuran. | Background
Conservation Breeding Programs (CBP's) are playing an important role in the protection of critically endangered anuran amphibians, but for many species recruitment is not successful enough to maintain captive populations, or provide individuals for release. In response, there has been an increasing focus on the use of Assisted Reproductive Technologies (ART), including the administration of reproductive hormones to induce gamete release followed by in vitro fertilisation. The objective of this study was to test the efficacy of two exogenous hormones to induce gamete release, for the purpose of conducting in vitro fertilisation (IVF), in one of Australia's most critically endangered frog species, Pseudophryne corroboree .
Methods
Male frogs were administered a single dose of either human chorionic gonadotropin (hCG) or luteinizing hormone-releasing hormone (LHRHa), while female frogs received both a priming and ovulatory dose of LHRHa. Spermiation responses were evaluated at 3, 7, 12, 24, 36, 48, 60 and 72 h post hormone administration (PA), and sperm number and viability were quantified using fluorescent microscopy. Ovulation responses were evaluated by stripping females every 12 h PA for 5 days. Once gametes were obtained, IVF was attempted by combining spermic urine with oocytes in a dilute solution of simplified amphibian ringer (SAR).
Results
Administration of both hCG and LHRHa induced approximately 80% of males to release sperm over 72 h. Peak sperm release occurred at 12 h PA for hCG treated males and 36 h PA for LHRHa treated males. On average, LHRHa treated males released a significantly higher total number of live sperm, and a higher concentration of sperm, over a longer period. In female frogs, administration of LHRHa induced approximately 30% of individuals to release eggs. On average, eggs were released between 24 and 48 h PA, with a peak in egg release at 36 h PA. IVF resulted in a moderate percentage (54.72%) of eggs being fertilised, however all resultant embryos failed prior to gastrulation.
Conclusions
Hormone treatment successfully induced spermiation and ovulation in P. corroboree , but refinement of gamete induction and IVF techniques will be required before ART protocols can be used to routinely propagate this species. | Competing interests
The authors declare that they have no competing interests.
Authors' contributions
PGB and AJS were equally responsible for designing the experiments, coordinating and conducting the experiments, analysing the data and writing the manuscript. Both authors read and approved the final manuscript. | Acknowledgements
Thanks to Dale Roberts, Dave Hunter and four anonymous referees for comments on the manuscript. The work was supported by grants from the NSW Department of Environment, Climate Change and Water (DECCW), the Corroboree Frog Conservation Trust, the Murray Catchment Management Authority (Murray CMA) and the Foundation for National Parks and Wildlife (FNPW). The research was also supported by Abbott Australia. The project was part of a multi-pronged recovery program for the Southern Corroboree Frog. The recovery team includes experts from the NSW Department of Environment, Climate Change and Water (Dr David Hunter), the Amphibian Research Cente (Dr Gerry Marintelli and Erika Marintelli), Taronga Zoo (Dr Peter Harlow and Michael McFadden), Tidbinbilla Nature Reserve (Murray Evans), Healesville Sanctuary (Kristy Penrose), Melbourne Zoo (Raelene Hobbs), and the Murray Catchment Management Authority (Elisa Tack). | CC BY | no | 2022-01-12 15:21:37 | Reprod Biol Endocrinol. 2010 Nov 29; 8:144 | oa_package/58/a7/PMC3014959.tar.gz |
PMC3014960 | 21118539 | Introduction
The occurrence of squamous cell carcinomas (SCC) in the colorectum is a rare entity representing a small fraction of colorectal malignancies, since more than 90% of colorectal diseases are adenocarcinoid tumors [ 1 ]. Very little information is available in the literature about the etiology, prognosis and optimal treatment of this malignancy [ 2 ]. Here in this study, we describe a patient with SCC of the rectum who underwent a lower anterior resection (LAR) for the possible treatment of the malignancy. | Discussion
Colorectal cancer (CRC) is the third most common cause of cancer-related death in the world [ 3 ]. Almost 90% of CRC are adenocarcinomas, while the remaining 10% are made up of carcinomas, sarcomas and lymphoid tumors [ 1 ]. The occurrence of SCC in the gastrointestinal tract (GIT) is a rare phenomenon, and its occurrence in the colorectum is extremely unusual [ 4 ]. The incidence of SCC of the colorectum has been reported to be almost 0.1 to 0.25 per 1000 CRC [ 4 , 5 ]. A look into the research work and the reported cases of SCC dates back to 1907, when Herxheimer reported adenosquamous carcinoma of the cecum but it was in 1919 when the first case of pure SCC of the colon was reported by Schmidtmann [ 6 ] in a 65-year-old man [ 7 ]. It was not until 1933 that the first case involving the rectum was subsequently described by Raiford [ 8 ]. In India, Bhat et al . [ 9 ] reported the first case of pure SCC of the colon in 1993 in a 55-year-old female from the southern part of the country. Until now almost 120 cases of SCC have been reported from all over the world (See Table 1 ). Surprisingly, a study from Russia reported 107 cases of SCC from a single center alone [ 10 ] but there has been no such reports of high incidence of SCC in the colorectum from any other part of the world.
Before the diagnosis of primary SCC of colorectum is made, certain criteria must be fulfilled as given by Williams et al . in 1979 [ 11 ]. This criteria includes: (A) absence of evidence of squamous cell carcinoma of any other part of the body, ruling out any chance of possible metastasis from any organ to the colorectal site; (B) exclusion of any proximal extension of anal squamous cell carcinoma; (C) absence of fistulous tract lined by squamous cells; and (D) confirmation of SCC by histological analysis [ 1 , 4 , 12 ]. All of these criteria were fulfilled by our case.
A look at the available literature reveals that squamous cell carcinoma of the colorectum affects individuals with a mean age of 55 to 60 years Women are more frequently predisposed to SCC than men, around 66% of cases occurred in women and 34% in men. Furthermore, SCC occurs in concomitance with an advanced tumor stage (Duke's C) [ 4 , 13 ]. Since SCC of the rectum is a rare tumor, epidemiological data constituting patient demographics, risk factors and natural history are lacking in the literature. The clinical characteristics of the patients with SCC of the colorectum are similar to those with adenocarcinoma: rectal bleeding, abdominal pain, change in bowel habits and weight loss [ 4 ]. Because of the rare nature of this malignancy the prognosis for patients is difficult to establish, Comer et al . suggested a poorer prognosis for patients with colorectal SCC than adenocarcinoma [ 1 , 4 , 14 ].
Almost four different pathophysiological theories regarding the origin of squamous cell carcinoma of the colorectum have been proposed in the literature so far. These can be summarized as: (A) Proliferation of uncommitted basal cells into squamous cells which undergo malignant transformation following mucosal injury [ 15 ]; (B) Ability of pluripotent stem cells to undergo spontaneous squamous differentiation [ 16 ]; (C) Squamous metaplasia of glandular epithelium resulting from chronic inflammation or irritation, secondary to inflammatory bowel disease [ 17 ], infection [ 18 ] or radiation [ 19 ]; (D) Origin from embryonal nests of ectodermal cells; and (E) Arousal of carcinomas from preexisting adenomas or adenocarcinomas [ 7 , 20 ]. | Conclusion
In conclusion, advanced colorectal SCC has a poor prognosis. Since colorectal SCC is a very rare disease, treatment selection is difficult. However, surgical resection and adjuvant chemotherapy [ 21 ] is a better approach to the treatment of colorectal SCC. | Background
Primary squamous cell carcinomas of the colorectum are very uncommon. Until now, to the best of our knowledge, only 114 cases of squamous cell carcinoma in the colorectum exist in the reported literature. Here we report a case of squamous cell carcinoma of the rectum in the ethnic Kashmiri population in northern India.
Case Presentation
The case of a 60-year-old male patient (Asian) with a pure squamous cell carcinoma of the rectum is presented here. The patient underwent a curative surgery with concomitant chemotherapy. Two years after the initial curative resection of the tumor he is still alive.
Conclusion
The prognosis for squamous cell carcinoma of the colorectum is worse than for that of adenocarcinoma, because of the delayed diagnosis. The etiopathogenicity of squamous cell carcinoma of the colorectum is discussed. Surgical resection of the lesion seems to be the treatment of choice. Chemotherapy also helps in improvement of the prognosis. | Case presentation
A 60-year-old male patient from an urban area of Kashmir (Asian) visited the Department of General Medicine of our institute with the chief complaints of severe lower-abdominal pain for the past eight months. The patient also complained of severe constipation, nausea, vomiting, anorexia, loss of appetite, abdominal cramps, incontinence of faeces and weight loss during the past four months. He experienced profuse bleeding from the rectum for the last month. Initial interviews with the patient revealed that the he was a heavy smoker and frequent user of noon-chai (Salt tea), meat and pickles. On examination the patient was found to be anemic. Digital rectal examination revealed an ulcero-infiltrative lesion with restricted mobility about 4 cm from the anal verge on the left lateral wall. A colonoscopy confirmed the rectal examination and biopsies taken at the time of the colonoscopy revealed squamous cell carcinoma (SCC) of basal cell type in the first histopathological examination. The report was re-confirmed by a second independent pathologist. A Contrast-Enhanced Computed Tomography (CECT) of the chest, abdomen and pelvis was also done but no lesions were found in any other site than the rectum. The lesion was without any fat stranding or lymphadenopathy. Furthermore, following the provisional diagnosis, the patient was referred to the Department of General Surgery for radical treatment, where he underwent LAR of the rectum using the standard technique of mesorectal excision (Figure 1 ). The continuity of the gut was restored by a circular stapler for low colorectal anastomosis with formation of a colonic pouch. The colonic pouch takes over the function of rectal reservoir which is lost after excision of the middle and lower rectum. Microscopic examination of the resected lesion demonstrated a 2.5 cm × 3 cm SCC tumor of the rectum infiltrating the serosa. The margins of the excised tissue were found to be free of the tumor. However, four regional lymph nodes were also infiltrated by the metastatic SCC cells. The liver and the rest of the organs were free of any metastasis. The slides were reviewed by a third histopathologist who reported the lesion as poorly differentiated squamous cell carcinoma. The stage of the tumor was found to be T 3 N 2 M o . The post-operative period was uneventful. Post-operatively the patient received four cycles of chemotherapy with cisplastin and 5-fluorouracil for five days. The patient is on two years of follow-up and has not shown any evidence of recurrenceas of the present time.
Consent
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 with the corresponding author of this manuscript and is accessible for review by the Editor-in-Chief of this journal
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
ASS conceived and designed the study and wrote the manuscript. NS suggested the necessary changes and copyedited the manuscript. NAC and FQP procured and provided the tumor samples for the study. MAS coordinated the study and revised the manuscript. All authors read and approved the final manuscript. | Acknowledgements
The authors gratefully acknowledge the Sher-I-Kashmir Institute of Medical Sciences, Kashmir for providing funds for this research work. The authors also gratefully acknowledge the technical staff, especially Miss Roohi and Mr. Reyaz of the Department of General Surgery for helping in the procurement of tumor tissue samples from the Operation Theater. We also thank the anonymous pathologists of the Department of Pathology for the histopathological assessment of the tumor tissues. | CC BY | no | 2022-01-12 15:21:37 | J Med Case Reports. 2010 Nov 30; 4:392 | oa_package/56/1f/PMC3014960.tar.gz |
||
PMC3014961 | 21143833 | Introduction
Aggressive angiomyxoma is a rare mesenchymal tumor of the pelvis and perineum that occurs almost exclusively in adult women [ 1 ]. It preferentially arises from the soft tissue of the pelvic region, perineum, and genital area. Its incidence is approximately sixfold higher in women, and 24 male cases have been reported in the literature [ 1 ]. The tumor is usually locally infiltrative and has a high rate of local recurrence after surgical excision [ 1 ]. The adjective "aggressive" emphasizes the neoplastic character of the blood vessels, its locally infiltrative nature, and the high risk of local recurrence, not indicating a malignant potential of the lesion. Rarely, this tumor appears in men, simulating inguinal hernia, testicular neoplasm, spermatic cord neoplasm, hydrocele, or spermatocele [ 2 , 3 ]. | Discussion
Since 1983, when aggressive angiomyxoma was first described by Steeper et al ., about 100 cases in both sexes (including 24 men) were reported worldwide [ 1 ]. It often occurs in middle-aged patients (mean age, 46 years) [ 1 ]. Occurrence of aggressive angiomyxoma in men is extremely rare, and, in men, aggressive angiomyxoma is usually derived from the pelviperineal interstitial tissue involving the scrotum (38%), spermatic cord (33%), perineal region (13%), and intrapelvic organs (8%) [ 1 ].
Macroscopically, its typical cut-surface appearance is a large, grossly gelatinous, and locally infiltrative tumor. Microscopically, the stroma is rich in collagen fibers with a prominent vascular component, including many thick-walled vessels. The differential diagnosis includes angiomyoblastoma, myxoid neurofibroma, myxoma, spindle cell lipoma, and myxoid liposarcoma [ 1 , 2 ]. Immunohistochemically, the stromal cells of the tumor show strong positivity for vimentin and variable positivity for desmin, α-smooth muscle actin, and CD34 [ 2 , 4 ].
Immunohistochemical studies have revealed that tumor cells are immunoreactive for no specific marker. Male angiomyxoma may be positive for estrogen and progesterone receptor [ 2 ]. The tumor cells in this case, however, were negative for the two markers.
Cytogenic analysis reveals chromosomal translocation involving chromosomes 8 and 12, associated with rearrangement of the HMGIC gene [ 1 ].
Surgery is the principal first-line treatment to date and, because of the high risk of local recurrence, a long-term postoperative follow-up with either ultrasound (US) or computed tomography (CT) is recommended [ 1 ]. The recurrence may be attributed to incomplete tumor resection, because of the infiltrating nature, and the absence of a definite capsule. The earliest recurrence has been reported as appearing nine months after surgery [ 5 ]. No distant metastasis, however, has been reported. | Conclusions
In conclusion, aggressive angiomyxoma is a very rare neoplasm that is more predominant in women. After surgery, close follow-up is needed because of the high risk of local recurrence. | Introduction
Aggressive angiomyxoma is a rare myxoid mesenchymal tumor of the pelvis and perineum, which occurs almost exclusively in adult women. The tumor is especially rare in men.
Case presentation
We report the case of a 68-year-old Japanese man with a slowly growing inguinal swelling. At surgery, a huge mass in the soft tissue of the inguinal region was found, not involving the adjacent organs. The morphologic picture was compatible with aggressive angiomyxoma of the inguinal region.
Conclusions
Aggressive angiomyxoma is a very rare, locally infiltrative neoplasm. Thus, after surgery, close follow-up is needed because of a high risk of local recurrence. | Case presentation
A 68-year-old healthy Japanese man presented with a slowly growing swelling of the soft tissue in the inguinal region (Figure 1 ). The duration of symptoms was about five years. At surgery, a large encapsulated mass (7.5 cm) was found, not involving the adjacent structure. The tumor was easily removed, as it was discrete and without adhesions. The cut surface of the tumor was smooth, homogeneous, and gray-white (Figure 2a ). Histologically, it was a paucicellular (hypocellular) tumor composed of fibrotic and myxoid areas, showing a sparse population of spindle-shaped tumor cells without significant cytologic atypia or mitosis (Figure 2b ). Foci of thick-walled blood vessels of various sizes were identified. The tumor cells were positive for CD34, and negative for α-smooth muscle actin and desmin. The tumor cells were negative for hormone receptors (ER and PgR). Chronic inflammatory cells were found scattered in the stroma. The morphologic picture and the immunostain were compatible with aggressive angiomyxoma in the inguinal region. The operation itself was uneventful and, on follow-up, no signs of recurrence have appeared for about one year.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
TK performed histologic examination, analyzed the case, and wrote the manuscript.
Consent
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. | Acknowledgements
I thank Nao Yoshida, Mika Ohya, and Yuko Nishikawa at Shinko Hospital, and Michiko Tajiri-Mori and Shuichi Matsuda at Kobe University for their excellent technical assistance. I also thank Dr. Takahiro Tokiyoshi at Shinko Hospital for providing clinical images. | CC BY | no | 2022-01-12 15:21:37 | J Med Case Reports. 2010 Dec 8; 4:396 | oa_package/5b/1e/PMC3014961.tar.gz |
||
PMC3014962 | 21143888 | Introduction
"Zygomycosis" refers to infections caused by a class of fungi called Zygomycetes, which includes the genera Rhizopus , Absidia , and Rhizomucor . They had previously been assigned to the genus Mucor and were considered responsible for the disease known as "mucormycosis" [ 1 , 2 ]. These fungi are ubiquitous in nature and are common inhabitants of decomposing matter. They can cause serious and rapidly fatal infections, particularly in individuals with compromised immune systems, such as those with poorly controlled diabetes with ketoacidosis [ 1 - 4 ].
The fungi invade major blood vessels, leading to extensive necrosis, and in extensive pulmonary disease, bleeding into the lungs may occur. In patients with diabetes mellitus, pulmonary mucormycosis may develop, with a less fulminant disease course but with atypical presentation of a solitary nodule [ 5 ].
Biopsy (surgical or transbronchial) of abnormal tissue retrieved by bronchoscopic aspiration or bronchoalveolar lavage (BAL) via a bronchoscope and microbiologic evaluation are the most efficient methods for detecting endobronchial Mucor [ 3 - 5 ]. We report a rare case of diffuse pulmonary mucormycosis in a patient with well-controlled type 2 diabetes who had a fatal pulmonary hemorrhage during a fiberoptic bronchoscopy procedure. | Discussion
Mucormycosis can be radiologically misdiagnosed as active tuberculosis, chronic necrotizing aspergillosis, coccidioidomycosis, or bronchiectasis. These have all been reported in the diabetes population and may be associated with massive or recurrent hemoptysis [ 6 - 8 ].
Therefore, fiberoptic bronchoscopy with biopsy or bronchial aspirate or both are needed to confirm a suspected diagnosis and to start an appropriate therapy.
Previous reports of rhinocerebral or pulmonary mucormycosis in HIV-negative diabetes patients involved subjects with poorly controlled diabetes: acidosis and hyperglycemia provide an excellent environment for the fungus to grow [ 9 ].
In our patient's case, no evidence of severe or persistent hyperglycemia was noted. Her general condition and absence of coagulation alterations indicated that a bronchoscopy examination could be carried out. However, the fatal hemorrhagic event occurred during standard bronchoscopic procedures performed to obtain specimens for histologic and microbiologic assessment. Autopsy findings confirmed that endobronchial mucous and necrotic plugs seen during the bronchoscopic procedure were related to pulmonary vascular invasion by mucoraceous hyphae.
Fiberoptic bronchoscopic examination is a useful procedure for identifying bronchial obstructions and endoluminal lesions, as well as for assessing the tracheobronchial tree beyond stenoses. Moreover, the procedure makes it possible to restore normal airflow in airless areas around the blockage (possible atelectasis or subatelectasis). Al Majed [ 10 ] reported the removal of a mucormycosis lesion through a rigid bronchoscope.
The risks associated with bronchoscopy procedures are well known. However, our case study suggests that, in the absence of a clear or well-defined diagnosis, particular caution should be exercised when conducting an endoscopic examination in diabetes patients with suspected pulmonary mucormycosis. Similar to other cases of this group of fungi, angioinvasion, thrombosis, and necrotic lesions are the hallmark features. Moreover, diabetes patients have endothelial dysfunction, increased arterial stiffness, or decreased arterial distensibility [ 11 ]. Therefore, any sampling procedure such as aspiration may trigger vessel rupture, with massive bronchial hemorrhage.
In these cases, BAL could be advocated as a less invasive technique. Moreover, non-invasive techniques, such as virtual bronchoscopy, have been found to be useful for assessing lesion friability in cavitated disease [ 12 ].
It has been suggested that an air-crescent sign on a chest radiograph is an important sign of potentially fatal hemoptysis [ 13 ]. Moreover, lesion changes and progression should indicate the need to start early antifungal and surgical therapy [ 4 ]. However, in our case, none of these signs was observed.
Starting treatment early seems to be the significant factor in reducing mortality associated with this disseminated pulmonary disease. Most patients with mucormycosis have been treated with lipid preparations of amphotericin (predominantly liposomal) with few reactions or adverse events [ 3 ]. For years, amphotericin B has been the drug of choice for these highly aggressive infections. Recently, patients who were unresponsive to monotherapy with liposomal amphotericin B responded favorably to the addition of echinocandin caspofungin acetate [ 14 ]. Newly introduced, second-generation triazoles include voriconazole, which is not active against the Zygomycetes, and posaconazole, which has been demonstrated to be active in vitro , in animal models, and in case reports [ 15 ].
Other mechanisms to prevent or limit this fatal complication are unclear. | Conclusion
Two interesting findings emerge from this case study. First, that mucormycosis should be considered in all diabetes patients regardless of degree of metabolic control.
Second, that fungal lesions may be more friable in these subjects, who might be at greater risk of complications associated with broncoscopy procedures. | Introduction
During infection, Mucorales fungi invade major blood vessels, leading to extensive necrosis, and in cases of extensive pulmonary disease, bleeding into the lungs may occur.
Case presentation
We report an unexpected event of post-bronchoscopy fatal endobronchial hemorrhage in a 62-year-old HIV-negative Italian woman with well controlled diabetes mellitus who presented with diffuse cavitated pulmonary lesions. Fiberoptic bronchoscopy revealed bilateral obstruction of the segmental bronchi. Fatal massive bleeding occurred after standard biopsy procedures. Histologic examination showed that the hyphae were more deeply colored by hematoxylin-eosin (H&E) than by other stains for fungi. Culture and autopsy confirmed bronchopulmonary mucormycosis.
Conclusion
Infection by Mucorales fungi should be considered in the diabetes population regardless of the degree of metabolic control. In these patients, particular caution should be taken during bronchoscopic procedures because of the greater friability of the fungal lesions. | Case presentation
A 62-year-old woman of Italian origin and nationality with a history of fever and a persistent cough for three weeks was admitted to our hospital for a scheduled fiberoptic bronchoscopy (FB) to assess the nature of a diffuse pulmonary lesion revealed by computed tomography (CT) chest scan (Figure 1 ).
Multiple cavitated lesions in the lungs were diagnosed. Some contained air and had a hyperdense capsule in the bronchi, consistent with bronchiectasis, whereas others resembled heteroplastic cavitary lesions (Figure 1 ). The patient had no significant clinical history except for type 2 diabetes mellitus (DM), controlled with oral anti-diabetic treatment and a suitable diet. In the three years of follow-up, her hemoglobin A 1c remained below 7%.
At the time of admission to hospital, the patient's white blood cell count was 10 × 10 9 /L with 50% neutrophils; hemoglobin and blood glucose were 13.0 g and 260 mg/100 ml, respectively. No coagulation alterations were observed, and the thrombocyte count was 200.000 cells/ml. A human immunodeficiency virus (HIV) test performed on admission was negative.
The bronchoscopy examination revealed a mucous and necrotic plug completely occluding the lingular bronchus and the apical segment of both lower-lobe bronchi (B6) (Figure 2a ). Grasping forceps were used to remove mucous plugs for cytohistologic analysis, and other samples were obtained by bronchial aspiration. At the end of the sampling procedures, performed by a team of bronchologists with more than 20 years of experience in these techniques, the patient had a massive hemorrhage followed by cardiorespiratory arrest and was transferred to the Intensive Care Unit (ICU) in a state of coma.
Histologic examination showed large areas of necrosis in the bronchial mucosa, with peripheral granulomatous reaction and several multinucleated cells (Figure 2b ). In the necrotic areas, numerous broad, very rarely septate, haphazardly branched hyphae were evident. Significantly, the hyphae were more deeply stained by hematoxylin-eosin (H&E) than by other special stains for fungi, such as periodic acid Schiff (PAS) (Figure 2c, d ) and Gomori methenamine silver stain (GMS).
Morphologic aspects and stain affinity suggested mucormycosis. This hypothesis was confirmed by the results of the bronchial aspirate culture (histologic sample), which showed filamentous mycetes belonging to the class Zygomycetes. Consequently, the patient was treated with 5 mg of liposomal amphotericin B per pound of body weight.
No evidence of disseminated infection was found, and analysis of cerebrospinal fluid did not reveal cerebral mycoses. The patient died 20 days after being admitted to the ICU. Autopsy confirmed bronchopulmonary mucormycosis.
Abbreviations
GMS: Gomori methenamine silver stain; H&E: hematoxylin-eosin; PAS: periodic acid Schiff.
Competing interests
The authors declare that they have no competing interests.
Consent
Written informed consent was obtained from the patient's family 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.
Authors' contributions
PD, DC, and DD, participated in the conception of the idea, review of the literature, writing of the manuscript, and interpretation of histologic assays. FL collected and interpreted data. AML and ES wrote the pathologic section and reviewed the manuscript. All authors have read and approved the final manuscript. | CC BY | no | 2022-01-12 15:21:37 | J Med Case Reports. 2010 Dec 9; 4:398 | oa_package/f7/e5/PMC3014962.tar.gz |
|||
PMC3014963 | 21143900 | Introduction
Resorption of the root of the maxillary incisors during ectopic eruption of the maxillary canines is not an uncommon phenomenon, and must be considered in all patients with seriously diverging eruption of the maxillary canines. The root of the lateral incisor is the most commonly affected by resorption, although the central incisors can be affected [ 1 - 3 ]. Incisor resorption has been reported to occur more frequently in women, with the female to male ratio being reported as 2:1 [ 3 ], 4:1 [ 1 ] and 10:1 [ 4 ]. However, no sex differences have been found in the severity or location of root resorption [ 1 ].
Root resorption of the maxillary incisors is often difficult to identify on intra-oral radiographs, mainly due to the overlapping of the incisors by the ectopic canine. Cone-beam computed tomography (CT) is superior to conventional X-ray methods for the assessment of incisor root resorption associated with ectopically positioned maxillary canines, as it eliminates the blurring problem of conventional tomography and increases the perceptibility of root resorption substantially [ 5 ]. It was found that 50% more resorptions are detected with CT compared with conventional radiographic methods [ 6 , 7 ].
Early detection and assessment of the extent of resorption is, therefore, of fundamental importance if preventive and early corrective measures are to be taken in order to reduce later complications and to prevent the resorption from getting worse.
Interventional orthodontic treatment consists of a brief period of orthodontic therapy or the removal of teeth (deciduous and/or permanent), with the attempt to eliminate tooth impaction. In this case report we describe the use of a fixed palatal appliance with a torquing spring to move the root of lateral incisors away from the erupting canines. | Treatment progress and results
Because our patient had a reverse overjet and 100% complete overbite, the conventional labial bracket system could not be used. Therefore, a specially designed upper fixed appliance with a torquing spring was constructed to apply torque to the upper lateral incisors from the palatal side. Tip-edge brackets were bonded to the palatal surface of the upper lateral incisors for tying the palatal arch-wire into the teeth. | Discussion
In this case report, a fixed palatal appliance was used to prevent and eliminate root resorption of the upper lateral incisors caused by the erupting canines. Resorption of adjacent teeth during the eruption of maxillary canines is rare in children. However, if it happens it can require comprehensive orthodontic treatment including extractions. Early detection and assessment of the extent of resorption is very important for treatment modality, and may decrease the extent of later complications [ 8 ].
In our patient, the main objective was to prevent and minimize root resorption of the lateral incisors due to the erupting maxillary canines. The interceptive treatment objective was achieved by applying a palatal torque to the root of the lateral incisors away from the canines and by extraction of the primary first molars, which provided more room for the canines to erupt in a more distal direction away from the roots of the lateral incisors.
Because of the excellent tissue contrast and precise three-dimensional images achievable, the use of CT scans for localization of the impactions and for evaluation of resorption is advocated by many authors. In our patient's case, a CT scan provided us with an accurate evaluation of the position of the erupting canines in relation to the roots of the lateral incisors, and as we found resorption had already started in the upper left lateral incisor and could start any time in the upper right lateral incisor, we decided to intervene immediately to save both teeth. However, the radiation dose of a CT scan is approximately 10 times to that of an orthopantomogram [ 9 ]. Therefore, once we ascertained that the tip of the canine had passed away from the apex of the lateral incisor using conventional radiography, the chance of root resorption of the lateral incisor was small and another cone-beam CT scan to reconfirm the situation was not justified.
One may question why extraction of the primary canines was not selected to provide the increased space needed for the eruption of the permanent canines in this case. Our explanation is that the plan was to allow the upper canines to erupt distal to the root apices of the lateral incisors. The primary canines were intentionally not extracted so that the lateral incisors would not drift distally, which would in turn have prevented the canines from erupting distally.
The reverse overjet was not corrected at this stage for two reasons. The first reason is our patient has a significantly increased mandibular length, therefore the chances of relapse to reverse overjet would be great. The second reason is the root of right central incisor is in contact with the crown of the unerupted left central incisor, as shown in the cone-beam CT. To move the right central incisor may have increased the risk of root resorption. For the upper left central incisor, a later radiograph showed the appearance of odontome-like structures that were not apparent in the original cone-beam CT. The odontomes were subsequently removed surgically and a bracket was bonded onto the upper left central incisor for orthodontic extrusion.
Evaluation of our patient after treatment showed an acceptable outcome. This case illustrates the importance of monitoring canine eruption and taking immediate action if necessary. Further monitoring on the erupting teeth will be carried out. | Conclusions
Close monitoring of erupting canines is very important and appropriate intervention is critical to avoid or minimize root resorption. Early intervention can spare the patient time, expense, more complex treatment and injury to otherwise healthy teeth. | Introduction
Resorption of the root of the maxillary incisors during ectopic eruption of the maxillary canines is not an uncommon phenomenon, and must be considered in all patients with seriously diverging eruption of the maxillary canines.
Case presentation
We report on the diagnosis and treatment of a 10-year-old Chinese boy with severe crowding and risk of root resorptions caused by impacted canines in the upper arch and reverse overjet. With the aid of cone-beam computed tomography, the upper right canine crown of our patient was positioned in close proximity to the right lateral incisor while the left canine crown was hitting the root apex of the left lateral incisor. To avoid any progress of root resorption, use of an upper fixed palatal appliance with torquing spring to move the root of lateral incisors away from the canines, plus extraction of upper primary first molars, was selected as an interceptive treatment.
Conclusions
Careful planning is crucial to avoid any complication through orthodontic treatment and to reduce the treatment time and cost. | Case presentation
Diagnosis and treatment plan
A 10-year-old Chinese boy was referred to our Orthodontic Department at Hong Kong University with the chief complaint of delayed eruption of the upper left central incisor. The child was in good health and had no relevant medical history or previous dental trauma. Intra-oral examination showed that he had a mixed dentition, reverse anterior overjet with complete deep bite and clinically missing upper left central incisor (Figures 1 and 2 ).
Radiography showed all the permanent teeth were present and the possibility of root resorption of the upper lateral incisors (12, 22), caused by mesially erupting position canines could not be ruled out (Figure 3 ).
A cone-beam CT scan was taken to assess the extent of resorption, if any, and to aid in the creation of a suitable treatment plan. The three-dimensional view from the CT scan (Figures 4 and 5 ) revealed that the crown of the upper right canine was in close proximity to the upper right lateral incisors. The crown tip of the upper left canine was touching the root of the upper left lateral incisors, causing some root resorption, which was confirmed in the sequential transaxial views of the CT scan (Figure 6 ).
The treatment objective was to prevent and eliminate any root resorption of the lateral incisors that might be caused by the erupting canines. The treatment plan was extraction of the upper primary first molars and palatal root torque of the lateral incisors by using a specially designed upper fixed lingual appliance (Figure 7 ).
Description of the appliance
A 19 × 25 stainless steel (ss) wire was soldered to the fixed palatal arch and extended anteriorly to the palatal surface of the lateral incisors. Two torquing springs were attached to the wire to rest on the palatal surface of the lateral incisors.
The appliance was cemented onto the upper first molars, and fitted anteriorly into the bracket slot and secured in place by using ligature ties through the vertical slot of the bracket. Prior to cementation the torquing spring was activated to apply a palatal root torque to the upper lateral incisors.
Then, three months after the treatment, radiography showed that the lateral incisors were in a more upright position and the upper canines were erupting in a more favorable distal direction away from the root of lateral incisors, and no further resorption was evident (Figure 8 ). Our patient was further reviewed and a subsequent radiograph (Figure 9 ) showed that the both canines had erupted further down from the roots of the lateral incisors, and there were no signs of root resorption. Later photographs (Figure 10 ) show the successful eruption of both canines.
Consent
Written informed consent was obtained from the parent of 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.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
RWKW designed the fixed appliance, supervised the treatment of our patient and revised the manuscript. BKA conducted the treatment and was a major contributor in writing the manuscript. All authors read and approved the final manuscript. | CC BY | no | 2022-01-12 15:21:37 | J Med Case Reports. 2010 Dec 9; 4:399 | oa_package/8a/a3/PMC3014963.tar.gz |
||
PMC3014964 | 21143957 | Introduction
Insomnia is a subjective experience of poor or unrefreshing sleep that may be apparent from a delayed onset or decreased duration of sleep [ 1 ]. The short acting hypnotic cyclopyrrolone derivative, zopiclone (RP 27,267) was synthesised in the late 1970 s and introduced into clinical practice some ten years later. Zopiclone, like the other Z-drugs, zolpidem and zaleplon, interacts with the same molecular target as the benzodiazepines on the GABA A (γ-aminobutyric acid) receptor [ 2 , 3 ]. Zopiclone is available as the racemic mixture although the S-enantiomer, eszopiclone, has a 50 times higher affinity for the receptor than the R-enantiomer [ 4 ]. In terms of clinical efficacy and adverse reactions, the Z-drugs showed little difference from the short acting hypnotic benzodiazepines [ 5 ] although the fatality toxicity index (deaths/10 6 prescriptions) for zopiclone, 2.1, was much lower than the 9.9 for temazepam [ 6 ].
When the Z-drugs were first introduced, there was little epidemiological evidence of misuse; however, as time went on, more cases began to appear in the literature. In a report published in 1995 [ 7 ], zopiclone misuse was identified in three male teenagers attending their local CDAT (Community Drug and Alcohol Team). The zopiclone had either been prescribed or bought on the street (zim-zims) at a cost of £1 a tablet. In two individuals, zopiclone was being used to intensify and prolong the effects of alcohol while the third, a polysubstance user, was crushing up the tablets for intravenous injection. However, in none of the three was there any evidence of dependence.
In a survey of 100 clients attending a methadone maintenance program at a Liverpool clinic, six admitted to using zopiclone in daily doses up to 50 times the therapeutic dose over periods of between six and 24 months [ 8 ]. All six had previously used temazepam but preferred zopiclone because of its lack of amnesic side effects. Zopiclone was readily obtainable from people who were prescribed the drug or by forging prescriptions.
A more recent study of individuals on a methadone maintenance program in Dublin found that 37 out of 158 (23%) tested positive for a zopiclone metabolite, 2-amino-5-chloropyridine, in the urine [ 9 ]. Benzodiazepines were also present in the urine of just under 70% of this group, none of whom were currently being prescribed these drugs.
In a systematic review of the literature from 1996 to 2002, 22 cases of zopiclone abuse and 36 for zolpidem were identified with doses used up to 51 (zopiclone) and 120 (zolpidem) times higher than those recommended for insomnia [ 10 ]. The authors concluded that the two Z-drugs were relatively safe compared to the hypnotic BZs and both dependence following chronic use and addictive non-medical use were quite rare although they did warn of the possible risk to patients with a history of substance use or mental illness. This warning has been confirmed in later case reports of opioid-dependents injecting zolpidem [ 11 ] and a cocaine user snorting zaleplon obtained from several crushed capsules [ 12 ]. | Discussion
After two decades of clinical use, the Z-drugs are considered to have a relatively low risk of being misused or producing dependence compared to benzodiazepine hypnotics like temazepam. This was the conclusion reached by Hayak et al . [ 10 ] on the basis of the number of global prescriptions and a systematic Medline search covering a seven-year period up to 2002. However, the authors cautioned against their uncritical use in patients with a history of substance misuse and/or psychiatric illness. A similar note of caution was sounded by Cimolai [ 14 ] who identified 20 cases of dependence on zopiclone (dose range 7.5 to 390 mg/d) in Canada between 1991 and 2006.
Dependence on alcohol may co-exist with dependence on prescribed psychotropic drugs, thus complicating clinical treatment. Johansson et al . [ 15 ] collected data from a sample of alcohol dependents in both community (n = 130) and inpatient (n = 23) settings in Sweden. Seven percent of the outpatients and 13% of the inpatients were found to be dependent on zopiclone or zolpidem compared to 12% and 30% respectively for benzodiazepines. However, the severity of dependence as measured relative to the defined daily dose (DDD for zopiclone = 7.5 mg, zolpidem = 10 mg) was between 1.0 and 1.5 (range of mean values for the two Z-drugs in the two treatment settings) whereas for the benzodiazepines, the values were 5.5 for outpatients and 11.5 for inpatients; a value of ≥4.0 being considered high dose dependence (HDD).
Using these values, our patient would be defined as HDD with a value of 12. In other case studies on zopiclone dependence, ranges of 2 to 40 [ 9 ] and 1 to 51 [ 10 ] have been noted. In one case described as ultrahigh dose dependence, a 34-year old woman was taking zolpidem equivalent to 200 and when abruptly discontinuing the drug, suffered withdrawal seizures [ 16 ].
In the Dublin sample, the mean age of onset and duration of zopiclone use was 28 and 4.2 years respectively with the mean age of the cohort being 32 years [ 9 ]. From the review of the 22 cases of zopiclone dependence, where the individual ages were presented (n = 15), the mean age was 39 years [ 10 ]. Our patient was 31 years old and although there are a few case reports of zopiclone being taken at doses higher than her typical daily intake of 60 mg, the 13 years of almost continuous use might be considered particularly unusual. It is also interesting to note that she had every intention of returning to zopiclone as soon as possible after discharge and that her experience with benzodiazepines was considered inferior to zopiclone. Comorbidity of alcohol dependence and anorexia nervosa is common and a serotonergic neuronal dysfunction has been implicated in both disorders [ 17 ]. In our patient, her cravings for zopiclone following withdrawal and interim substitution with diazepam would be indicative of dependence. We would recommend a cautious approach to prescribing Z-drugs to patients with a dual diagnosis and that particular attention be paid to any request for an increase in the daily dose. | Conclusion
We have described an individual with an atypically long history of high dose dependence on zopiclone, alcohol dependence and anorexia nervosa. This case adds further support to the growing body of evidence that prescribing zopiclone to drug users with an underlying psychiatric disorder should be carefully monitored. | Introduction
The Z-drugs, zaleplon, zopiclone and zolpidem, are short-acting hypnotics which act at the same receptor as the benzodiazepines, but seemingly without the potential for misuse and the development of dependence of the older benzodiazepines. However, with increased prescribing of Z-drugs, reports of misuse and possible dependence began to appear in the literature, particularly in people with a history of substance misuse and comorbid psychiatric illness. Here we report the case of a woman with a history of chronic zopiclone use and anorexia nervosa, admitted for alcohol detoxification.
Case presentation
A 31-year old Caucasian British woman with a history of long-term zopiclone use and anorexia nervosa was admitted as an inpatient for a ten-day alcohol detoxification. Her weekly (four days out of seven) intake of alcohol was 180 units and her daily intake of zopiclone, 30 mg. Apart from a short period five years ago, she had been taking zopiclone for 13 years at daily doses of up to 90 mg. She admitted to using 'on top' of her prescribed medication, purchasing extra tablets from friends or receiving them gratis from her partner. After detoxification from alcohol and zopiclone, she was prescribed diazepam which she found ineffectual and voiced her intention of returning to zopiclone on leaving the hospital.
Conclusion
Zopiclone is generally regarded as safer than benzodiazepines, however, this particular individual, who was using high doses of zopiclone over many years, may provide further evidence of a risk of dependency when this drug is prescribed for substance users with a comorbid psychiatric illness. | Case presentation
Our patient was a 31-year-old unemployed Caucasian British woman with a diagnosis of alcohol dependence (ICD-10, F10.2) who was referred to the Acute Assessment Unit (AAU) of the hospital for a ten-day detoxification by her local CDAT. On the 28 days immediately prior to admission she had been drinking 6 L of cider (7.5% ABV) equivalent to 45 units four times a week and had been having blackouts as a result. High levels of both aspartate transaminase (AST = 86 U/L) and γ-glutamyltransferase (γGT = 187 U/L) suggested possible hepatic dysfunction but there was no evidence of cognitive impairment (MMSE score = 29). On admission, she was taking chlorpromazine (50 mg twice a day.) for anxiety, fluoxetine (40 mg once daily) for low mood and zopiclone (7.5 mg four times a day). Her physical health screen did not reveal any abnormalities.
She had a long history of both anorexia nervosa and alcohol dependence. Anorexia was first diagnosed in 1994, and when she was 17 years old she was treated as an inpatient. By the age of 18 years, her problem with alcohol had become evident and over the intervening years she has had six separate detoxifications with varying lengths of abstinence; relapses being due to life events or trauma. She also has a history of self-harm, overdosing, burning and lacerating; her last admission to Accident and Emergency was two years ago. Her father died of alcohol-related problems and her uncles are also alcohol dependent. Her sister had anorexia nervosa and died from cardiac complications, a common consequence of the severe calorific deprivation associated with this eating disorder [ 13 ].
Zopiclone (7.5 mg nocte ) was first prescribed to help her to sleep when she was being treated for anorexia in the rehabilitation unit. She found the calming effect of taking it during the day highly desirable and when she was discharged, asked her doctor to increase the dose, claiming she had become tolerant of its hypnotic effect. She reported a typical daily intake of 60 mg, but sometimes she used up to 90 mg, beginning when she woke up and continuing throughout her waking cycle. Her use of alcohol did not change throughout the time she was taking zopiclone. Apart from prescribed zopiclone, she obtained the drug from friends (paid for) and her partner (donated).
Zopiclone was described as "her best friend" and like alcohol it gave her confidence, relaxed her and enhanced her self esteem. She said that she was ultra-possessive about her supply and would not be separated from it, keeping it with her at all times. During the 13 years of using there had only been one relatively short period of abstinence, which occurred six years ago, when she was in the hospital for an alcohol and zopiclone detoxification. However, this ended with the recurrence of the anorexia and she was prescribed zopiclone to help her sleep.
The current detoxification followed the standard protocol used in the AAU, namely a tapering off of doses of chlordiazepoxide (130 mg to zero over six days) and on each of the first five days, an i.m. injection of Pabrinex ® and thereafter Vitamin B Compound Strong tablets. Her zopiclone was reduced from 7.5 mg nocte via 3.75 mg to zero over the same period as the chlordiazepoxide after which she was started on diazepam 20 mg, with the dose being reduced incrementally by 1 mg each day. She found diazepam an ineffectual substitute and had cravings for zopiclone and said she could not wait to return to taking it as soon as possible. She had no intention of stopping zopiclone in the forseeable future.
Our patient is still off zopiclone (and alcohol) after 17 months although she continues to have strong cravings for zopiclone (more so than for alcohol) which would be easy to satisfy. She is concerned that zopiclone is not considered addictive and that there is no specific protocol for detoxification, other than substitution with diazepam, and help for understanding this addiction and preventing relapse.
Patient's perspective
Gratefully I'm still off zopiclone though I am aware I could get hold of it easily. As yet, there's no detox for coming off zopiclone. No help provided for the mental or physical reasons and withdrawal.
Consent
Written informed consent was obtained from the patient for publication of this case report. A copy of the written consent is available for review by the Editor-in-Chief of this journal.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
FK is the Consultant Psychiatrist on the AAU and AM is a Research Scientist (Pharmacology). AM interviewed the patient and wrote the manuscript. Both authors read and approved the final manuscript. | CC BY | no | 2022-01-12 15:21:37 | J Med Case Reports. 2010 Dec 10; 4:403 | oa_package/24/7a/PMC3014964.tar.gz |
|||
PMC3014965 | 21156058 | Introduction
Spontaneous subarachnoid hemorrhage (SAH) is a significant clinical problem that occurs most commonly as a result of aneurysm rupture. In approximately 15% of cases, however, no aneurysm can be identified by cerebral angiography. Although in a minority of cases occult aneurysms are eventually identified, non-aneurysmal SAH represents an interesting clinical problem that can occur as a result of many different pathologies, including vasculitis, arterial dissection, intra-cranial or cervical arteriovenous malformation or fistula, clotting diatheses, antiplatelet and/or anticoagulant medication, pituitary apoplexy and tumors [ 1 ]. Benign perimesencephalic hemorrhage (BPH) is another described type of non-aneurysmal SAH and is thought to account for approximately one- to two-thirds of non-aneurysmal SAH and 5-10% of SAH as a whole [ 1 , 2 ]. The presenting symptoms of both aneurysmal SAH and BPH overlap and include sudden onset "thunderclap" headache, nausea, emesis and meningismus. The diagnosis of BPH can be made on the basis of the appearance of hemorrhage limited to the prepontine and/or perimesencephalic cisterns on computed tomography (CT) scans in the absence of an aneurysm on cerebral angiography [ 1 ].
Despite the fact that aneurysmal SAH and BPH are the respective leading diagnoses in spontaneous SAH with and without an identifiable point of origin [ 1 ], to the best of the authors' knowledge, no cases of both vascular pathologies occurring in the same individual have been previously reported. Herein we present the case of a patient with aneurysmal SAH followed five years later by BPH. | Discussion
Despite the recent identification of BPH as a distinct vascular pathology [ 2 ], it is now purported to be a primary etiology of non-aneurysmal SAH [ 1 , 3 ]. In contrast to aneurysmal SAH, BPH, for which some authors have proposed the term pre-truncal non-aneurysmal hemorrhage [ 4 ], is thought to arise from multiple possible non-arterial sources [ 2 , 5 ]. Previous studies have reported that patients with BPH have normal life expectancies and are not at risk for re-bleeding [ 1 ]. Other studies have noted some, albeit reduced, post-hemorrhage complications compared to aneurysmal SAH, including vasospasm, post-hemorrhagic hydrocephalus and death [ 6 ].
The patient in this case had BPH that was both temporally and spatially remote from her previous aneurysmal SAH. Other than her general risk factors for SAH, which include female sex, hypertension and previous ruptured aneurysm, the literature offers little insight regarding a probable underlying pathology that could account for both of these hemorrhages. There have been previous reports of BPH occurring in individuals with various vascular pathologies, including ischemic stroke [ 5 ] and venous stenosis or thrombosis [ 7 , 8 ], but none that the authors know of in the setting of a previously ruptured intra-cranial aneurysm. Rebleeding from BPH has been reported only once, although it occurred after early anti-coagulation [ 9 ].
The location of the patient's aneurysmal rupture was fortuitous as it related to her subsequent BPH. It was evident that the location of her perimesencephalic hemorrhage did not extend to the region of her previously coiled Acomm aneurysm and most likely arose from a separate process. This permitted the diagnosis of BPH to be made and a less aggressive treatment course to be pursued. Her treatment would have been significantly more complicated had her aneurysm been in the posterior circulation within the region of her BPH. In this setting, a diagnosis of BPH would have been difficult to justify, and she possibly would have undergone attempts at either recoiling or even open clipping of a suspected unsecured aneurysm.
Multiple previous case series have attested to the relatively benign course of perimesencephalic hemorrhage [ 2 , 3 , 6 , 10 ]. It would not be unreasonable, however, to posit that the currently presented patient may have fared worse than expected, given her previous aneurysmal SAH. Fortunately, this was not the case. There is no evidence that the first bleeding event rendered her more susceptible to a second, less severe event. It is unknown what effect BPH occurring shortly after aneurysmal SAH or in someone with a poorer grade injury would have on neurologic outcome. | Conclusions
This work represents the first report of both aneurysmal SAH and non-aneurysmal BPH occurring in the same individual. The diagnosis of BPH may be complicated by previous aneurysmal SAH. The expected good prognosis associated with BPH does not appear to be altered by a previous episode of aneurysmal SAH. To those in the fields of neurology and neurosurgery, this case serves as an important reminder that in patients with a history of previous aneurysmal SAH, subsequent episodes of SAH need to be fully investigated because they may be attributable to an entirely different pathology. | Introduction
Both aneurysmal subarachnoid hemorrhage and benign perimesencephalic hemorrhage are well-described causes of spontaneous subarachnoid hemorrhage that arise as a result of different pathologic processes. To the best of the authors' knowledge, there have been no reports of both vascular pathologies occurring in the same individual.
Case presentation
A 51-year-old Caucasian woman with a history of aneurysmal subarachnoid hemorrhage presented five years after her initial treatment with ictal headache, meningismus, nausea and emesis similar to her previous bleeding event. Computed tomographic imaging revealed perimesencephalic bleeding remote from her previously coiled anterior communicating artery aneurysm. Both immediate and delayed diagnostic angiography revealed no residual filling of the previously coiled aneurysm and no other vascular anomalies, consistent with benign perimesencephalic hemorrhage. The patient had an uneventful hospital course and was discharged to home in good condition.
Conclusions
This report for the first time identifies benign perimesencephalic hemorrhage occurring in the setting of previous aneurysmal subarachnoid hemorrhage. The presence of a previously treated aneurysm can complicate the process of diagnosing benign perimesencephalic hemorrhage. Fortunately, in this case, the previously treated anterior communicating artery aneurysm was remote from the perimesencephalic hemorrhage and could be ruled out as a source. The patient's prior aneurysmal subarachnoid hemorrhage did not worsen the anticipated good outcome associated with benign perimesencephalic hemorrhage. | Case Presentation
The patient was a Caucasian, non-smoking 51-year-old woman with insulin-dependent diabetes and hypertension who initially presented at the age of 46 with acute-onset ictal headache, meningismus and emesis (Hunt/Hess grade I). A non-contrasted head CT scan revealed SAH in an aneurysmal pattern (Figure 1A ). She underwent cerebral angiography, which revealed a 6 mm anterior communicating artery (Acomm) aneurysm (Figure 2A ) that was treated with endovascular coiling in the same setting. At the end of the procedure, a 0.5 mm residual was noted at the base of the aneurysm that incorporated the anterior cerebral arteries and was not treated (not shown). Follow-up angiography eight months later showed complete obliteration (Figure 2B ). The patient's hospital course was complicated by vasospasm, which was treated with hypervolemia, hypertension and intra-arterial nicardipine, as well as cerebral salt wasting, which was treated with sodium and volume supplementation. She was ultimately discharged to home and returned to work six weeks later with no residual neurologic deficits. After her eight-month posthemorrhage angiogram, she was lost to neurosurgical follow-up. Of note, the patient developed peritoneal dialysis-dependent renal failure three years later. A renal ultrasound did not demonstrate evidence of polycystic kidney disease (not shown).
Prior to the patient's current admission (five years after her aneurysmal SAH), the patient had experienced three days of intermittent nausea and vomiting. On the day of admission, she developed an acute-onset severe headache with worsened nausea and emesis. On admission, the patient's condition was Hunt/Hess grade II with an initial non-contrasted head CT scan demonstrating perimesencephalic SAH in the prepontine, interpeduncular, ambient and crural cisterns (Figure 1B and 1C ). Of note, no hemorrhage was noted adjacent to the previously treated Acomm aneurysm (Figure 1C , arrowhead). Although there was no evidence of thrombocytopenia or other coagulopathy on her admission laboratory testing, she was taking clopidogrel and received a pool of platelets. A toxicology screen revealed no evidence of sympathomimetic use. Diagnostic cerebral angiography did not reveal any new aneurysms or vasculopathy and showed the previously treated aneurysm to be stable with no residual (Figure 2C ). The diagnostic angiogram also demonstrated patent cerebral venous sinuses without evidence of thrombosis or stenosis. Given the lack of vasculitic changes on the angiogram, further workup for vasculitis was not performed. The patient had a follow-up angiogram eight days later that again failed to show any source for the hemorrhage, consistent with BPH (not shown). The patient had an uneventful hospital course and was discharged to home in good condition on post-bleed day 10. Follow-up magnetic resonance imaging and angiography performed six months later demonstrated no vascular abnormalities (not shown).
Consent
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.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
RS was responsible for the initial care of the patient, along with the conception and writing of the manuscript. DK and HK were responsible for patient management and workup for benign perimesencephalic hemorrhage. MH was responsible for the patient's original aneurysmal management and supervised her case on her subsequent admission. All authors read and approved the final manuscript. | CC BY | no | 2022-01-12 15:21:37 | J Med Case Reports. 2010 Dec 14; 4:405 | oa_package/b4/d8/PMC3014965.tar.gz |
|||
PMC3014966 | 21162730 | Background
Social capital can be understood as features of social structures - including norms, inter-personal trust, and mutual support - which act as resources for individuals and also facilitate collective action [ 1 , 2 ]. Although the concept originated in sociology to explain diverse phenomena such as educational success, labor market attachment, and the prevention of crime, an increasing number of studies in the last decade and a half have turned to the exploration of social capital in public health [ 3 ]. To date, studies have suggested that social capital may be a determinant of health based on its association with mortality [ 4 ], health behaviors [ 5 ], mental health [ 6 ] and self-rated health [ 7 , 8 ].
Several mechanisms have been suggested by which social capital can influence health [ 9 ]. It is important to note that the influence of social capital on health can be either health-promoting or health-damaging, depending on the social context. Thus, social capital can enhance the diffusion of deleterious health practices such as the spread of smoking among adolescent social networks via social learning [ 10 ]. On the other hand in other situations, such as school-based, peer-led interventions, social capital can enhance the diffusion of health-promoting behaviors such as smoking cessation. A second mechanism by which social capital can influence health is through the provision and exchange of mutual psychosocial support. Again, the effects can be either health-promoting or health-damaging depending on the social context. Thus, it has been documented that involvement in dense social ties can be a detriment to health in resource-deprived settings stemming from the excess burden of providing social support to others [ 11 ]. By contrast, mutual support and norms of reciprocity have been found to be protective for mental and physical health in settings where basic individual needs are already met. Yet a third mechanism by which social capital can influence health is via the exercise of collective action and informal social control [ 12 ]. Groups that are rich in stocks of social capital have greater capacity (i.e. the "collective efficacy") to bring about desired outcomes such as the prevention of antisocial behaviors like drug abuse among minors in the community. On the other hand, a community that is too cohesive can lead to pressure to conform as well as ostracism of individuals who behave differently from the others [ 13 ]. The challenge in social capital research is thus to identify and isolate the specific instances in which it can be a force for health promotion versus a detrimental factor in the patterning of health outcomes.
By now an established convention in social capital research is to distinguish between the impacts of the so-called cognitive dimension of social capital - e.g. perceived trust of others - versus the behavioral dimension - such as participation in groups [ 14 ]. Increasingly the distinction is also drawn between "bonding" social capital and "bridging" social capital. Bonding capital refers to the resources that accrue to members of a group who are similar to each other with respect to social identity, such as class, race/ethnicity, or religion. By contrast, bridging capital refers to resources that span across social cleavages [ 15 ]. Although very few empirical studies in the health field have addressed this distinction, research suggests that bonding capital can be both promoting and damaging for health (depending on the context), whereas bridging capital seems to be more uniformly protective for health because of the potential to link resource-deprived individuals to material and symbolic resources.
In the present study, based on an adolescent sample in Brazil, we sought to address three questions. First, we sought to focus on the association of social capital with self-rated health among youth. The bulk of empirical studies to date have been conducted among adult samples, and less is known about the links between social capital and health outcomes among adolescents [ 16 , 17 ]. Secondly, we sought to distinguish between the different dimensions of social capital-cognitive versus behavioral, and bonding versus bridging. Lastly, we focused on the Latin American context where few studies have been conducted on social capital so far. In a systematic review of studies conducted throughout the world, Islam and colleagues [ 18 ] noted that the association between social capital and health tended to be stronger, and more consistently observed, in societies characterized by unequal distribution of social resources (e.g. the United States). By contrast, among more egalitarian societies (such as the Nordic welfare states), social capital was less consistently associated with health outcomes. Brazil is a very unequal society and we hypothesized that having access to social capital could account for variation in the health status of adolescents. | Material and methods
We carried out a cross-sectional study in 2009 in a town that belongs to the metropolitan urban area of Belo Horizonte, capital of Minas Gerais state, Brazil.
Our sample of adolescents was drawn from a philanthropic non-governmental organization (NGO) which was established in 1966 with the purpose of connecting working adolescents to local employers. The adolescents self-refer to the services of the NGO, which maintains the following admission criteria: a) the adolescents must be aged between 16 years to 17 years and 11 months, b) they must be students enrolled in a public school, c) they must come from economically vulnerable family backgrounds, and d) they must take and complete the admission course provided by the NGO. Neither race nor religious backgrounds are taken into consideration at admission. During the admission course, the youth have an opportunity to learn about the NGO statute and its bylaws, to attend Portuguese and reading classes, and to be prepared to fit in the companies that will hire them.
By the end of the course, all the adolescents receive a qualification certificate on human relations, work team and professional behaviors. Once accepted by the NGO, the youth have an opportunity to get a part-time job and also to attend free sports classes and extra recreational activities on weekends.
The recruitment of the participants required one day and the adolescents recruited in this present study were all those present during a special weekend meeting and who had signed assent to participate in the survey (and from whom we also obtained consent from either their parents or guardians). For this present study, we used the World Bank Integrated Questionnaire to Measure Social Capital (SC-QI), which is a psychometrical validated instrument [ 19 ]. The World Bank group study distributed some indicators into dimensions based on their previous experience with social capital survey and reading scientific literature. This instrument has 27 items and is divided into six different dimensions of social capital as follow: a) groups and network, b) trust and solidarity, c) collective action and cooperation, d) information and communication, e) social cohesion and inclusion, f) empowerment and political action. The IQ-SC questionnaire was conceptualized for the micro level (individuals) and does not collect data on the level of communities, thus it is an adequate instrument for the objective of our study.
The IQ-SC has no overall scoring algorithm. This explains the possibility of our selection of some social capital indicators among the total 27. One of co-authors (I.K.) is an expert on social capital and health studies and selected 14 of 27 items based on social epidemiological evidence and the peculiarities of adolescence age group.
All the adolescents in the NGO (N = 363) participated in the survey. Ethics approval was granted by Research Ethics Committee at Pontificia Universidade Catolica de Minas Gerais.
Outcome Measure
Our outcome measure was a one-item self-rated overall health status question. Individuals responded to the following question: "How would you describe your overall state of health these days, would you say it is very good, good, fair, poor or very poor?". We combined the categories to have a binary outcome of self-rated health where 0 = very good, good or fair and 1 = poor or very poor.
Social Capital
The cognitive dimensions of social capital were assessed by four questions inquiring about: a) trust in others, b) perceived helpfulness of neighbors and c) perceptions of whether the youth could borrow money from others in case of need. The variable "trust in people" was dichotomized as: 0 = people can be trusted or 1 = you can't be too careful. The variable "helpfulness of neighbors" was assessed by the extent to which youth agreed with the 2 statements that: "most people in this neighborhood are willing to help you in case of need" and "in this neighborhood, one has to be alert or someone is likely to take advantage of you". The answers (agree strongly, agree somewhat, neither agree nor disagree, disagree somewhat, disagree strongly) were dichotomized into 0 = agree versus 1 = disagree or unsure. The variable "borrow money" was determined by a five scale question asking if there was someone beyond family or close relatives willing to help the youth in case of need to borrow a small amount of money (definitely, probably, unsure, probably not or definitely not), which was dichotomized to 0 = yes and 1 = no or unsure.
The behavioral dimensions of social capital were assessed by: a) participation in community activities during the past twelve months; b) time or money contribution to a community project (0 = yes versus 1 = no); c) whether they belonged to a group (0 = at least one or more, 1 = zero); d) whether they had a close friend, and e) whether they got together with people to have food or drink in the past month 0 = at least one or more, versus 1 = no.
Finally, we assessed bonding versus bridging social capital by asking respondents to think about the people with whom they associated. Separate questions inquired whether the youth associated with others from different ethnic backgrounds, different economic or social status or different religious groups. For each of these questions, the responses were coded as 0 = yes (bridging) and 1 = no (bonding).
Confounders
The variables sex (male or female), age (0 for 17 years old vs. 1 for 15-17 years old), skin color (white, black, brown, yellow, Indian - classification of Brazilian Institute of Geographic and Statistic) and educational background (first or second grade of middle school - correspond to the 11 th grade American high school) were entered as possible confounders in regression models.
Statistical Analyses
The SPSS statistical package version 18.0 was used for data analyses. Descriptive statistics were performed to characterize the participants including relative and absolute frequencies of the explanatory variables and confounders as well their associated odds ratio for poor self-rated health with 95% confidence intervals (95% CI).
Next, the Enter method was used for the logistic regression models. Variables with a p -value lower than 0.2 were included in the logistic regression models. Although social demographic variables did not meet the formal inclusion criterion [(sex p = 0.510), (age p = 0.714), (educational background p = 0.725), (skin color p = 0.515)], they were nonetheless retained in the regression models due to their theoretical plausibility. | Results
Of the total respondents who took part in our study (N = 363), 95.9% were male and 4.1% were female. Regarding age, the proportion of 15-16 and 17 years old was similar, 53.3% and 46.7%, respectively. Fifty-seven percent of the adolescents reported their skin color as yellow (Asiatic) and 19.3% as black; 88.7% of the adolescents were enrolled in the second grade of Brazilian high school (Table 1 - Additional file 1 ).
Table 1 (Additional file 1 ) shows the descriptive statistics for individual variables and their crude odds ratios for poor self-rated health. Individuals who agreed that "one has to be alert or someone is likely to take advantage of he/she" were more likely to report poor self-rated health (OR = 2.2, 95% CI = 1.1-4.5) compared to those who did not express this sentiment. Individuals who did not get together to have food or drink with people of different social status (i.e. lacking bridging social capital) were more likely to report poor health (OR = 2.0, 95% CI = 1.1-3.5) compared to who did have a different opportunity. Contrary to expectation, however, adolescents who did not belong to a group had a 60% lower odds ratio (OR = 0.4, 95% CI = 0.2-0.9) of reporting poor self-rated health. The remaining indicators of social capital (e.g. trust, having someone to borrow money from, perception of reciprocity) were each associated with self-rated health in the expected direction (i.e. low perception of social capital = increased odds of poor health), but the associations were not statistically significant.
The first step of multivariable-adjusted logistic regression ( Model I ) was controlled for socioeconomic and demographic variables. Model II was controlled for Model I and all social capital indicators. The variable "got together to have drink/food" was constant for selected cases thus excluded from logistic regression. Among socio-demographic variables, Indian ethnic background was the only variable that was statistically significantly associated with an increased odds ratio of poor health. Four of the fourteen indicators of social capital were significantly associated with self-rated health. Adolescents who reported having no one to borrow money from (OR = 2.1, 95% CI = 1.1-4.3) and who agreed that someone was likely to take advantage of them (OR = 2.9, 95% CI = 1.2-7.2) were at increased risk of poor-self-rated health. Distrust (OR = 2.0, 95% CI = 0.2-19.2) and to believe that people are not willing to help in case of need (OR = 1.9, 95% CI = 0.9-3.9) increased the chance of reporting poor health although not statistically significant (Table 2 - Additional file 2 ).
The behavioral dimension of social capital was also associated with self-rated health - although in opposite directions depending on the indicator. Adolescents who did not contribute time to a community project had increased risk of poor self-rated health (OR = 1.9, 95% CI = 1.1-3.7). But by contrast, adolescents who reported not belonging to a group were at lower risk of poor self-rated health (OR = 0.5, 95% CI = 0.2-1.4). Being involved in communal activities was not associated with adolescents' health. Youth who did not contribute money to a community project (OR = 1.6, 95% CI = 0.8-3.1) and who did not have a close friend were more likely to report poor health when compared with those did (Table 2 - Additional file 2 ).
Of the three different types of bridging social capital that we inquired about (getting together with people of different economic status, social status, and race/ethnicity), only bridging relationships with people of different social status was significantly associated with self-rated health. Adolescents who did not get together with people of different social status were at 2.3 times the risk of poor self-rated health (95% CI: 1.1-5.2) compared with those who did. Lastly, Model III was adjusted for Model I and the four social capital indicators that were remained associated with the outcome (borrow money, advantage, time contribution to a community project and get together with people of different social status). | Discussion
The objective of this study was to analyze a possible association between self-rated overall health and social capital among adolescents using four aspects of social capital (cognitive, behavioral and bonding/bridging). Our findings showed that social support, trust, civic participation and bridging social capital still remained associated with self-rated health after adjustment of all the other social capital indicators and confounders.
Adolescents who said that they had no one to borrow a small amount of money, who believed that someone else in the neighborhood would be likely to take advantage of them, who did not contribute time to a community project, and who did not get together with people from different social status had a greater likelihood of reporting poor or very poor health.
These findings converge with a previous study conducted in a developing country context. Khawaja and colleagues (2006) [ 17 ] investigated social capital and health status among 13-19 years old adolescents in Beirut addressing a similar question as our current study. Their results showed that distrust and social fragmentation were prevalent among individuals living in impoverished suburban communities; additionally, adolescents with low levels of social capital were almost four times more likely to report poor health compared to the others. Similar results were found in affluent countries such as Canada where youths from families of low affluence and low levels of social capital were more likely to report poor health [ 16 ].
In our study, the majority of respondents (95.8%) agreed that someone would likely to take advantage of them and almost 36% reported that there was no one beyond family or close relatives willing to help them in case of need to borrow a small amount of money. Capacity to trust could be seen as a consequence of personal trust which is related to connectedness in family and community as well [ 20 ]. Distrust and absent of reciprocity were associated with poor self-rated health among residents of a low income community nested in a Latin America country [ 11 ]. Our results were somewhat expected if we take into to account the reality of the actual structure of Brazilian society characterized by its inherent socioeconomic inequalities with huge gaps among different social classes and also a general distrust both in people and in government.
Bridging relations with people with different social status showed a positive effect on self-rated health, which may be a marker of the ability of adolescents to access valued resources held by individuals from higher social strata. While bonding social capital may be essential for emotional and social support as a whole, it does not enable the redistribution of resources from the advantaged to the disadvantaged groups [ 21 ]. Thus bridging social capital is important in enabling people from disadvantage groups to access materials and psychosocial resources and there is growing evidence that bridging social capital is positively associated with self-rated health, especially among disadvantaged groups [ 22 , 23 ].
With regard to civic participation, in Brazil there are a limited number of youth-oriented civic engagement and extracurricular activities through school and community-based youth organizations. Duke and colleagues (2009) [ 24 ] found that adolescents with strong familial relationships and community connections may experience healthy development and promote future civic engagement. Thus, connections at a young age contribute to important elements of social capital.
Some efforts have been done to try to increase stocks of social capital among adolescents in developing countries through controlled trials of interventions but these efforts are still in their infancy [ 25 ]. Empowerment, which is considered one of social capital dimensions, was studied by Pattussi and colleagues [ 26 ] among Brazilian adolescents. The multilevel analyses revealed that neighborhood empowerment may explain inequalities in the levels of dental caries. Adolescents from areas with higher levels of empowerment scores had lower levels of dental caries when compared with the adolescents whom lived in areas with low level of empowerment [ 26 ] and especially among boys [ 27 ].
Study limitations
The cross-sectional nature of the data does not allow any conclusion concerning causal intercourse between overall self-rated health and social capital. It is important to highlight the possibility of reverse causation and reporting bias once social capital may influence health and vice-versa. For instance it has been pointed out that the lower stock of social capital lead to poor levels of health of populations. However, the opposite may occur since individuals with poor health could generate lower levels of social capital. Another limitation of this study is the inherent bias in self-reporting both self-rated health and social capital (common method variance). Additionally, given that social capital is often conceptualized as a group resource, the ideal study design would be multi-level, i.e. individuals nested within different contexts (e.g. neighborhoods) with different levels of social capital [ 28 ]. Unfortunately, we did not collect information on the areas in which the adolescents resided.
Our sample was limited in terms of socioeconomic variability, thus reducing the generalizability of our findings. The main pre-requisites for the adolescents' application to the NGO are that they must be from disadvantaged socioeconomic backgrounds and they must be enrolled at high school. These two requirements explain the comparative homogeneity of our sample. The majority of adolescents were males due the type of jobs offered by the employers, i.e. issuing parking tickets on streets.
Lastly, it is important to highlight that only four out of fourteen social capital items are significantly associated to self-rated health. This fact could indicate the absence of a robust social capital effect rather than particular social capital specificities or might indicate a power problem of our sample size.
Study strengths
One of the strengths of our study is that we assessed all three dimensions of social capital - including the cognitive, behavioral and bonding/bridging aspects. Our findings underscore the point that specific indicators of social capital may have different associations with health outcomes.
Our study is also among the few so far that has been carried out in a developing country marked by wide socioeconomic disparities. | Conclusion
Social capital may play an important role explaining better self-rated health and the pattern of association differs according its specific dimensions. The three specific dimensions - cognitive, behavioral and bridging social capital - seem to be protective for the health of working adolescents living in a developing country context such as Brazil. | Background
Social capital may influence health and the patterns of association differ according its dimension such as cognitive, behavioral, bridging or bonding. There is a few numbers of studies in Latin America which comprise these aspects of social capital and health. The aim of this study was to examine the association between social capital and self-rated health among youth, and distinguish between the different forms of social capital - cognitive versus behavioral, and bonding versus bridging.
Findings
A cross-sectional study was conducted in 2009 among working adolescents supported by a Brazilian NGO. The sample comprised 363 individuals and data were collected using a validated structured questionnaire. The outcome, self-rated health, was measured as a dichotomous variable (poor/good health) and fourteen social capital indicators were investigated (cognitive, behavioral and bonding/bridging). Data were analyzed using multivariate logistic regression. Cognitive (social support and trust), behavioral (civic participation) and bridging social capital were associated with good self-rated health after adjustment of all the other social capital indicators and confounders (sex, age, skin color and educational background).
Conclusions
Social capital was associated with self-rated health and the patterns of association differed according its specific dimensions. Cognitive, behavioral and bridging social capitals were protective for adolescents health living in a developing country context.. | Competing interests
The authors declare that they have no competing interests.
Authors' contributions
IK idealized the design of this study, drafted all its versions and did the critical and final review. CB idealized the design of the study, did statistical analysis and wrote the manuscript. AC collected the data and with EF and AV gave important contributions during all stages of the draft. All authors read and approved the final draft.
Supplementary Material | Acknowledgements
We would like to thank the Brazilian Coordination for the Improvement of Higher Level Personnel (CAPES) responsible for the first author's PhD scholarship. | CC BY | no | 2022-01-12 15:21:37 | BMC Res Notes. 2010 Dec 16; 3:338 | oa_package/53/2f/PMC3014966.tar.gz |
PMC3014967 | 21118494 | Background
Wild species usually exhibit large genetic variability, which serves as a resource for adaptability to changing environments. On the contrary, cultivated plants are usually more limited in number and display less genetic variability, as a result of the genetic bottlenecks occurring at domestication, translocation and transition from landraces to modern breeding [ 1 ]. A consequence of such genetic erosion is genetic uniformity, which may result in the loss of relevant traits, such as resistance to biotic and abiotic stresses.
Thus, wild species related to cultivated crops represent interesting sources of genetic variation, through the introgression of new and better performing alleles. The large genetic variability present in the wild cereals is an invaluable resource for cereal crop improvement. Hordeum chilense Roemer et Schultes, a native South American diploid wild barley (2n = 2x = 14), offers a high potential for cereal breeding among the species of the genus Hordeum , because of its high crossability with other members of the Triticeae tribe and its agronomically interesting characteristics. Crosses between wheat and H. chilense lead to fertile amphiploids named tritordeums. They represent the basic genetic material for introducing genetic variability from H. chilense into wheat breeding programs [ 2 ] and for transferring useful genes from H. chilense to wheat. The analysis of the germplasm genetic structure is the basis of management, research and utilization of such germplasm [ 3 ], since it is critical to identify and correctly interpret the associations between functional and molecular diversity [ 4 , 5 ]. H. chilense has been found in a wide range of environments, and shows high genetic as well as phenotypic diversity [ 6 ]. The analysis of the structure of such high variations is important for breeding purposes, especially to identify genes or genomic regions involved in environmental adaptation and showing high diversity. The genetic structure of populations has been widely documented in most of the studies investigating the diversity of elite crop germplasm, especially in self-pollinating cereals [ 7 - 9 ]. Molecular markers and development of statistical techniques to analyze such data have been the subject of recent intensive studies [ 10 - 21 ], allowing the analysis of the genetic structure in several species and eliminating many of the problems linked with spurious associations. However, although significant efforts to increase the availability of genomic tools, such as molecular markers for cereal crops, have been undertaken in the last years, these developments were not made for wild species with scarce direct agronomic interest, like H. chilense . To solve this problem, the transferability of wheat and barley microsatellite markers (or Simple Sequence Repeats; SSR) to wild related species was evaluated [ 22 , 23 ]. Comparative genomic analyses have indicated a good conservation of coding regions across genomes of different grass species, suggesting that this part of the genome can be used to develop transferable molecular markers [ 24 - 27 ]. The development of high throughput sequencing technologies in recent years has allowed the generation of large Expressed Sequence Tag (EST) datasets in a number of plant species, including cereals, which can be systematically searched for SSR [ 28 ]. For example, Yu et al. [ 29 ] tested EST-SSR primers originating from hexaploid wheat and rice ESTs on four cereal crops (wheat, rice, barley and maize) and found that 62% of the primer pairs produced Polymerase Chain Reaction (PCR) amplicons on at least two species. Similarly, Zhang et al. [ 30 ] reported the transferability of 116 wheat EST-SSRs on 168 accessions, representing 18 grass species. The transferability among the Triticeae ranged from 73.7% for Aegilops longissima to 100% for wheat subspecies ( Triticum compactum ), but was also good for less related species such as rye (72.8%) or maize (40.4%). In barley, Varshney et al. [ 31 ] reported that 78.2% of the SSR markers used (165) showed amplification in wheat, followed by 75.2% in rye and 42.4% in rice. Finally, Gupta et al. [ 32 ] reported that 55.12% of wheat EST-SSRs were transferable to barley. Recently, it was shown that the barley EST-SSRs represent a promising source of molecular markers to screen the H. chilense genome [ 33 ].
In addition of their high degree of transferability across species, it was recently demonstrated that EST-SSRs are useful for genetic variability studies. For example, Gupta et al. [ 32 ] assessed the genetic diversity of EST-SSRs in a collection of 52 elite exotic wheat genotypes. Their results indicate that EST-SSRs are more useful for diversity analyses than genomic microsatellites (g-SSRs). Yang et al. [ 34 ] also used EST-SSRs to measure the genetic diversity among three hexaploid wheat populations. They concluded that EST-SSR markers are ideal markers for assessing genetic diversity in wheat. In addition, Balfourier et al. [ 35 ] used 38 g-SSRs or 44 EST-SSRs to analyze the structure and the diversity of a collection of 372 wheat varieties and obtained identical results with both types of markers. Recently, Hübner et al. [ 21 ] studied the genetic analysis of a new collection of the wild barley H. spontaneum with a set of 42 EST-SSRs, revealing that wild barley populations can be divided into seven major genetically differentiated clusters, as well as the evidence of temperature and precipitation as environmental cues that have shaped the genetic makeup of wild barley. Pan et al. [ 36 ] investigated the genetic diversity among 15 wild emmer wheat ( T. dicoccum ) populations using 25 EST-SSRs, detecting a considerable amount of genetic variation, partly related to ecological factors.
The goals of the current study were to determine: (i) the transferability level of wheat EST-SSR markers and their usefulness for H. chilense ; (ii) the genetic diversity of the H. chilense species, using a wide set of available microsatellite markers; (iii) the genetic structure of a natural collection of 94 H. chilense accessions; and (iv) the possible influence of spatial, morphological and environmental factors in the observed structure. | Methods
Plant material and DNA sampling
The DNA collection, isolated by Castillo et al. [ 38 ], consisted of 94 samples of H. chilense collected during several expeditions to Chile [ 67 , 68 ]. This germplasm is maintained at the Germplasm Bank of the Institute for Sustainable Agriculture (Prof. A. Martín, IAS, CSIC, Cordoba, Spain). Eight samples of H. chilense DNA were used for selecting transferable and polymorphic microsatellite markers, using Triticum aestivum cv. 'Chinese Spring' DNA as control, to corroborate the pattern of the microsatellites and the fragment sizes. In addition, a DNA set of T. aestivum - H. chilense addition lines developed by Miller [ 37 ] were used to determine the chromosomal locations of the polymorphic SSR. The addition lines for chromosomes 4H ch , 5H ch , 6H ch and 7H ch were disomic, whereas the addition lines for chromosomes 1H ch S, 2H ch alpha arm, 5H ch , 6H ch S arm and 7H ch alpha and beta arm were ditelosomic.
Amplification and transferability of wheat EST-SSRs
A selection of 98 SSRs derived from wheat ESTs [ 69 , 70 ] uniformly distributed across wheat chromosomes were initially screened for their transferability and polymorphism in H. chilense genome, and 20 were selected for germplasm analysis. A set of 21 barley EST-SSRs and eight wheat and barley g-SSRs, previously identified as useful for the genetic analysis in H. chilense [ 33 , 68 ] was added. In total, 49 polymorphic microsatellite markers were applied for fingerprinting H. chilense accessions. The polymerase chain reaction (PCR) amplification and fragment analysis were as previously described [ 43 , 71 , 72 ], or PCR was carried out using the M13 protocol as described in Nicot et al. [ 69 ], with an annealing temperature of 60°C for 30 cycles (30 s at 94°C, 30 s at 60°C, and 30 s at 72°C) and 56°C annealing for eight cycles. Amplification products were visualized using an ABI PRISM 3700 Genetic Analyzer from Life Technologies (Carlsbad, CA, USA). The fragment sizes were calculated using GeneMapper software from the same manufacturer.
Ecogeographical data
The geographic data (altitude, latitude and longitude) of 76 accessions were available [ 6 ], and thus they were used to project the data using the DIVA-GIS software http://www.diva-gis.org ;. The geographic location of the study area was between 28°15' and 38°42' South latitude and between 70°18' and 73°24' West longitude. The altitude on the sites varied within a wide range, from sea level to high mountains (> 2000 m). Since only one accession is available in some provinces, the accessions were grouped in eight zones along Chile, from North to South, including in some cases various close provinces with similar ecological characteristics. The ecological data like rainfall of wettest and driest month and mean temperature of the warmest and coldest month were obtained for each site using DIVA-GIS. The ecological regions were described following the bioclimatic classification of DiCastri and Hajek [ 41 ].
Statistical analysis
The summary statistics including the number of alleles per locus, polymorphism information content (PIC) values and gene diversity were determined using the application PowerMarker version 3.25 [ 73 ]. The unrooted neighbor-joining (NJ) tree was constructed using the Nei's index distance [ 74 ]. One thousand matrices were obtained by bootstrapping, and the consensus tree was constructed with the program Consense of the Phylip package (version 3.66) [ 75 ]. The dendrogram was visualized using the TreeView 1.6.6 software [ 76 ]. We performed a Mantel test correlation [ 42 ] between Nei's genetic distance and the natural (Napierian) logarithm of the geographic distances, using the library ade4 in the R package (version 2.10.1; R Development Core Team 2008) [ 78 ].
A Bayesian model-based analysis for inference of population structure was performed using the program Structure (version 2.2) [ 78 ] to estimate the number of groups (K) represented by all sampled individuals and the individual admixture proportions. The Structure software assumes a model in which there are K populations (where K may be unknown), each being characterized by a set of allele frequencies at each locus. Individuals in the sample are probabilistically assigned to a particular population, or associated to two or more populations (if their genotypes indicate that they are admixed). The number of clusters was inferred using 20 independent runs with 100,000 burn-ins and 100,000 iterations after burn-ins, following the admixture ancestry model and correlated allele frequencies, with K ranging from 1 to 10. We have followed the procedure by Evanno et al. [ 39 ] to better detect the real number of clusters determined by Structure. Also, the clusteredness index [ 40 ] was calculated, which is based on the Q matrix of Structure, being 1 when individuals are assigned completely to a single cluster and 0 when they are equally assigned to all clusters. The individuals can have membership coefficients summing 1 across clusters.
The Distruct 1.1 software [ 40 ] was used to graphically represent the estimated population structure, according to geographic proximity, ecological region and agronomical data. Each individual was represented by a thick line, which was partitioned into K colored segments, representing the individual's estimated membership fractions in K clusters.
The genetic structure of the population was also inferred by the Geneland package [ 79 ], implemented in the R software. The Geneland software uses geographic coordinates and does not assume admixture, whereas the Structure software does not use geographic coordinates and does assume admixture. We carried out five independent runs using independent allele frequencies with 100,000 iterations, from which each 100th observation was sampled from the Markov chain, with minimum and maximum K being 1 to 10. The run with the highest likelihood was post-processed to obtain the posterior mode of population membership. The genetic differentiation among genetic groups inferred by Structure was estimated by hierarchical analysis of molecular variance (AMOVA), implemented in the Arlequin 3.0 software [ 80 ].
We used the Lositan software [ 81 ] to identify outlier loci that had excessively high or low Fst compared to neutral expectations. The basic rationale is that (i) loci influenced by directional (also called adaptive or positive) selection will show a larger genetic differentiation than neutral loci; and that (ii) loci that have been subject to balancing (also called negative or purifying) selection will show a lower genetic differentiation. Thus, the methods generally consist of identifying loci that present Fst coefficients that are "significantly" different from those expected under neutral theory (they are called outlier loci). To avoid false positives caused by population structure, the Fst was calculated for the inferred structure groups (the significance level chosen was 0.001, which corresponds to a statistical significance level of 0.05), applying a Bonferroni standard correction. The association of alleles of outlier loci with ecogeographical factors was assayed by linear regression analyses, using the SPSS package version 17.0.0 from SPSS (Chicago, IL, USA). Alleles with frequencies below 5% were excluded. Alleles of each locus were introduced as dependent variables in the model and ecogeographical factors were the independent variables. Significance was calculated for the model, which included only one allele, with the significance threshold set at 0.05, using a Bonferroni correction, as already mentioned.
Values of environmental variables were first standardized and the Euclidean distance between the samples was computed using SPSS. The correlation between genetic distance and environmental distance in the collection was calculated by the Mantel test. Also, the Principal Component Analysis (PCA) was computed from environmental values, and the samples were plotted in genetic structure grouping. The Spearman rank correlation was used to assess differences in mean number of alleles and ecogeographic variables among the inferred groups. | Results
Transferability and polymorphism of wheat EST-SSRs
The transferability of the 98 wheat EST-SSRs was evaluated on a set of eight accessions of H. chilense . A SSR was considered as transferable when the PCR amplification of a band of the expected size and SSR pattern was observed. The primers that showed null alleles in some samples were tested at least twice, in order to avoid false negatives (eg., non-amplification due to PCR failure).
Among the 98 wheat EST-SSRs, 53 (54%) showed cross-species transferability. The percentage of transferred markers was about 50% on each chromosome, thereby indicating a uniform distribution across the genome. Among the 53 transferable SSRs, 20 PCR primer pairs (20.4%) showed polymorphism in the accessions studied and were used for further analysis in the H. chilense germplasm. Between two and 10 alleles per primer pair were observed and only 11.7% of the total alleles had the same size as the allele found in Triticum aestivum cv. 'Chinese Spring', confirming the good potential of H . chilense for wheat and barley genetic diversity improvement. These markers were first assigned to H. chilense chromosomes using the available set of wheat- H. chilense addition lines ([ 37 ]). Fifteen of the 20 polymorphic primer pairs were located on the same linkage group as in wheat. One of them was located on chromosome 7 D in wheat (CFE135), while it amplified a product on chromosome 1H ch in H. chilense . Four markers showed the same PCR amplicon sizes in both species, and thus their locations could not be confirmed.
Genetic variability analysis
To perform the genetic variability analysis, 21 barley EST-SSRs and 8 gSSRs previously identified as useful for the genetic analysis in H. chilense ([ 33 , 38 ]) were added to the 20 polymorphic wheat EST-SSRs transferred in this work. A total of 351 alleles were detected over the whole sample of 94 accessions for the 49 SSR loci. Among the 351 alleles, 162 originated from the 21 barley EST-SSRs, 94 from the 20 wheat EST-SSRs and 95 from the eight g-SSRs. The number of alleles per locus ranged from two (for GBM1411, GBM1323, GPW7425, CFE10 and CFE23) to 27 (for GBM1464), with a mean of 7.2 alleles and Polymorphic Information Content (PIC) of 0.5 per locus (Table 1 ). The highest (0.91) and lowest (0.04) PIC values were observed for GBM1464 and GPW7213, respectively. Generally, wheat EST-SSRs exhibited lower PIC values and fewer number of alleles than barley EST-SSRs. Barley EST-SSRs detected almost twice more alleles and higher PIC values than wheat EST-SSRs.
Additionally, several rare or specific alleles were found among the analyzed germplasm. Out of the 351 alleles, 184 were found at a frequency lower than 5%, and were therefore considered as rare. Clustering showed that the germplasm can be separated into two groups (see below) with a total of 66 specific alleles found in group I (39 being rare) and 134 specific alleles found in group II (82 being rare). The gene diversity, polymorphic information content (PIC), and the number of specific alleles was lower in group I than in group II (Table 2 ).
Analysis of genetic structure and differentiation among inferred groups
The genotyping data obtained from the 49 SSRs were used to analyze the genetic structure of the germplasm, using the Bayesian clustering model implemented in the Structure software. The natural logarithm of the probability of the data, proportional to the posterior probability of K, showed no clear peak between 1 and 10 for K, and therefore it was difficult to determine the true number of populations (K) (Figure 1a ). We applied the rate of change in the Napierian logarithm probability relative to standard deviation (ΔK), as described by Evanno et al. [ 39 ]. The results showed the highest peak at 2 (Figure 1b ), which was confirmed by the clusteredness index [ 40 ], showing the highest median level at K = 2 (Figure 1c ). Moreover, using the Geneland software, we observed that the posterior distributions of the estimated number of populations (K) across 10 replicates displayed a clear mode at K = 2 in all of them (additional file 1 ), corroborating again the previous data. Thus, these results suggest that the analyzed H. chilense germplasm can be divided into two genetically distinct groups.
To find the key determinants to the inferred structure of these two groups, we investigated the geographical proximity, as well as the morphological, agronomical and ecological characteristics in the accessions belonging to each group. Figure 2a shows the distribution of each accession into the two populations (identified by the Structure software, and designated as group I and group II, hereafter) and the three clusters (according to morphological and agronomical data as classified by Vaz Patto et al. [ 6 ]). Geographic origins divided in 8 zones (Figure 2b ) and ecological regions (Figure 2c ) according to the classification established by DiCastri and Hajek [ 41 ] at K = 2 populations, are also shown.
The genetic structure analysis, according to the geographical origins of the accessions (Figure 2b ) revealed that, in some regions, the accessions were grouped according to the geographical location. The zones 2, 4, 5 and 7 showed a uniform structure, while the rest of the zones were more or less admixed. Uniform structure was considered when more than 80% of the accessions in one group had more than 80% of membership in this group. The geographical origin of the accessions and their membership to the inferred groups are represented in Figure 3 . We calculated the correlation coefficient ( r ) between the geographic and the genetic distance matrices using the Mantel test [ 42 ]. We observed an r value of 0.21 (Figure 4a ), revealing a low but significant correlation (p < 0.001). The correlation was then analyzed separately for both inferred groups (I and II), and the results showed that group I had a uniform distribution in the Center of the country, yet group II expanded across the North, Center and South of the country. Therefore, group II was analyzed separately for the North, the Center and the South regions. We found a significant correlation between geographical proximity and genetic distance for group I and for group II Central. Thus, our results demonstrate a geographical influence in population structure in the Central Chilean region for both structure groups (additional file 2 ). Such influence was not detected for the group II accessions, either at the North or South regions (the latter is probably due to the scarcity of accessions).
To investigate the impact of ecological characteristics in the inferred structure (K = 2), the accessions were grouped according to bioclimatic parameters. The results revealed admixed populations, except for some provinces that showed a more uniform structure (like humid Mediterranean, very arid Mediterranean, very humid Mediterranean, and sub-humid Mediterranean; see Figure 2c ). Comparative analyses between ecogeographical data (latitude, longitude, altitude, mean temperature of coldest and warmest month, and rainfall of driest and wettest month) and genetic data (Nei's genetic distance matrix) revealed a weak but significant correlation (r = 0.108; p = 0.0017) (Figure 4b ). When analyzing the influence of the same ecogeographical data in both structured groups separately, the group I did not show any ecogeographical influence, whereas the group II Central exhibited a significant ecogeographical influence (additional file 3 ). In addition, the ecogeographical data were used to separate the H. chilense accessions through Principal Component (PC) analysis. The first component (PC1, Figure 5 and Table 3 ) was explained by variation in latitude and rainfall, accounting for 42.9% of the variation. The second component (PC2) accounted for 35.4% of the variation, being explained by variation in temperature and longitude. PC3 accounted for 12.3% of the variation, and was explained by the variation in elevation.
The unrooted Neighbor-Joining (NJ) tree (Figure 6 ) distinguished two groups of accessions, corresponding to the structure grouping. Neither geographical nor ecological evidence was detected in the grouping. Results of distance and Bayesian cluster analyses evidenced the presence of a structured genetic diversity among the groups. The Analysis of MOlecular VAriance (AMOVA) of the two inferred groups by the Structure software revealed a 33.16% of the genetic variation among groups, with the remaining 66.84% due to differences within groups. The genetic variances within and among groups were significant (F ST = 0.331, p < 0.001; being F ST the variance among subpopulations relative to the total variance), supporting the presence of a genetic structure.
Association between markers and ecogeographical factors
We identified 12 outlier loci that detected high or low variability with respect to the expected neutrality. Among those, 11 loci (GBM1350, GBM1064, GBM1008, GBM1060, GBM1464, GWM1047, GBMS14, GWM1302, CFE135, GPW7335, GPW7663 and GPW7577) are candidates for balanced selection, while the locus CFE135 is a candidate for being subjected to positive selection. The 12 markers were assayed for their association with ecogeographical data. The marker GWM1302 exhibited four different alleles (188, 190, 192 and 194 bp) among 10 alleles that were associated with low rainfall (Figure 7 ). The other markers did not show any significant association with any of the ecogeographical traits. | Discussion
Transferability and polymorphism of wheat EST-SSR
The transferability of EST-SSRs across related species has been demonstrated in several species and genera [ 24 , 26 , 29 - 32 , 43 - 45 ]. Recently, we reported on the usefulness of barley EST-SSRs for genetic analysis in H. chilense [ 33 ]. In the present work, we show that more than half (54%) of the assayed wheat EST-SSRs can be transferred to H. chilense , which is lower than the transferability of the barley EST-SSRs (66%). This is likely due to the fact that wheat is evolutively more distant from H. chilense than barley. This result is relatively consistent with the findings of Zhang et al. [ 30 ] and Gupta et al. [ 46 ], who reported higher transferability of wheat EST-SSRs to barley than to more evolutively distant species, such as maize, rice, sorghum, lolium (ryegrass), oats and purple false brome ( Brachypodium ).
Despite of the evolutive distance, the overall conservation of the wheat EST-SSRs linkage groups in H. chilense was high, indicating a good level of synteny between these two species. Only one disagreement was observed, that may be due to chromosomal rearrangements, which may be frequent during speciation [ 47 ]. Thus, the wheat EST-SSRs markers transferred to H. chilense have an added value as intergeneric syntenic markers, in addition to their direct application to analyze gene diversity. Since numerous additional wheat EST-SSRs are available in the public databases (eg., GrainGenes http://wheat.pw.usda.gov/GG2/index.shtml ; [ 48 ]), this number could be further increased.
Among the 53 wheat EST-SSRs showing good transferability to H. chilense , about 40% exhibited polymorphism between at least two accessions, which represents 20% of the initial set of wheat EST-SSRs. This was lower compared to what was previously observed between wheat and barley (60% of the transferable EST-SSRs; [ 30 ]). This is also lower compared to what was found using barley EST-SSRs (36%), because barley is more closely related to H. chilense than wheat. Thus, the markers transferred from wheat to H. chilense are only those that are more conserved, and therefore likely to show lower polymorphism. Similarly, the number of alleles detected in H. chilense with wheat EST-SSRs was lower compared to the number obtained when using barley EST-SSRs (94 vs 162, respectively). On the other hand, due to their wheat origin, they are not suitable for direct in-tube detection methods [ 49 ]. Nevertheless these drawbacks are largely overcome by the fact that they provide valuable anchors for synteny inference [ 50 ]. Therefore, we conclude that both wheat and barley represent a good source of markers for genetic diversity and structure studies of H. chilense germplasm collections. In addition, they represent an invaluable tool for the introgression of H. chilense alleles to other cereal species.
Genetic variability analysis
The microsatellite markers revealed a total of 351 alleles across all the 49 loci. The high level of genetic diversity detected could be an adaptive strategy in response to a heterogeneous environment. According to the marker's origin, the SSRs identified in EST databases detected a lower number of alleles and PIC than those obtained through general genomic libraries, including non-transcribed regions. Such is an expected result, due to the more conserved nature of EST-SSRs. The number of alleles detected and PIC found when using EST-SSRs from wheat was lower than with barley EST-SSR and non-transcribed genomic microsatellites, as expected, due to the evolutionary distance between species. Despite these facts, our results show a sufficient level of variation when using EST-SSRs (both from barley and wheat origins) to carry out genetic structure and future association mapping analysis. Therefore, this is one more case where the EST-SSR markers provide an opportunity to examine the functional diversity of germplasm collections, as shown by Eujayl et al. 2002 [ 51 ].
The group I identified by Structure was fixed to one allele in eight loci. Seven of these loci could be assumed as being non-neutral, due to their origin from EST databases. The higher number of fixed alleles in non-neutral SSRs could be explained by temporal variation of external factors, generating selection pressures that maintain variation within populations [ 52 ]. This may also be due to the fact that not all nucleotide bases on a transcribed DNA are of selective nature. In fact, the third base of the mRNA triplets is less specific (wobble hypothesis, [ 53 ]). Furthermore, some amino acids may share similar chemical properties (eg., nonpolar or polar, including acidic or basic), thus being less prone to generate a phenotypic change. A high number of specific alleles were identified, which could be an indication of the relatively high rate of mutation at SSR loci [ 54 ], or to a germplasm with a rich genetic diversity and a divergent population structure. The gene diversity is significantly lower in group I than in group II (Table 2 ). Therefore, the group II is genetically more diverse, corresponding to accessions present in a wider ecogeographical range.
The Spearman correlation showed that genetic diversity is influenced negatively by altitude and positively by temperature. Accessions in group II showed a higher number of alleles. Besides, they were found mainly in places with low altitude and higher temperature than accessions in group I. Therefore, accessions stressed by cold showed less genetic variation.
In this genetic variability analysis of H. chilense germplasm using wheat and barley gSSRs and EST-SSRs, we have defined two main germplasm groups (group I and group II). A previous analysis based on AFLP markers and a Principal-CoOrdinate analysis ([ 6 ]) divided the same germplasm into three clusters. The group II defined in this work contains two subgroups, corresponding to the clusters II and III described by Vaz Patto et al. ([ 6 ]).
Genetic structure in H. chilense
The analysis of the genetic structure using both Bayesian approaches (Structure and Geneland software) and genetic distance approaches (cluster analysis) of a set of 94 H. chilense accessions, using 49 microsatellite markers, revealed two genetically differentiated groups.
The 'admixture model' implemented by Structure gave a better fit to the species ecophysiological clusters, as defined by Vaz Patto et al. [ 6 ], and it was chosen for further association analysis. Therefore, the morphological and agronomic characteristics, which determined the ecophysiological clusters, were key determinants of the population structure of the H. chilense germplasm. Thus, the two inferred groups are mainly in accordance with the agro-morphological clusters described by Vaz Patto et al. [ 6 ], as the group I corresponds to cluster I, while the group II (with the exception of three lines) includes clusters II and III (see Figure 2a ). According to the geographical origins and the ecological distribution, the inferred genetic structure showed both uniform and admixed populations (Figure 2b and 2c ). The accessions included in the geographical zones 1 and 2 showed a uniform genetic structure. They were found in the driest places of these zones, corresponding with Mediterranean arid environments. The accessions grouped in the geographical regions 7 and 8 revealed also a uniform genetic structure, and were found in the wettest places, corresponding to and Mediterranean humid and very humid environments. This points to an influence of the rainfall in shaping the population structure.
Several studies have been carried out to detect the population structure in barley. In most of them, the key factors affecting the genetic structure were growth habit or spike morphology and geographic origin [ 15 , 16 , 19 , 20 , 55 - 57 ]. In our study, the morphological and agronomic characteristics also have appeared as the main factors to affect the population structure, although we have also shown that geographic locations and ecological patterns of distribution also affect this structure. The analysis of the total germplasm set revealed low but significant associations between geographical and genetic distance, as well as between ecological and genetic distance. By analyzing separately the three main geographical regions of provenance of the species (North, Center and South), a more significant correlation between geographical and genetic distance was detected for the accessions from Central Chile, but no association was found either in Northern or in Southern Chile accessions. This is in agreement with the basal phylogenetic position of H. chilense in South America, established initially in Central Chile from a long-distance continental dispersal from North America, followed by two independent dispersals to the North and to the South [ 58 ]. On the other hand, by analyzing both population structure groups separately, a higher correlation between ecological and genetic distance could be detected for group II accessions, but no correlation was found for group I accessions. In fact, group II accessions have shown a better ability to colonize the North and the South Chilean regions (see Figure 3 ). Thus, the adaptation to geographic and ecological factors may be one of the causes of the genetic structure in the studied germplasm. Thus, the results of our work illustrate the interest to further investigate how morphological characteristics and ecophysiological traits affect the species selection and the population structure. Moreover, the presence of a high level of structure within the H. chilense germplasm should be considered in future association mapping studies. The AMOVA detected higher differences among individuals within-population structure groups than among groups, which is consistent with findings from other studies, indicating that considerable genetic diversity is partitioned within rather than between wild barley populations [ 59 - 63 ]. The proportion of genetic variation within population groups reflected high levels of genetic diversity.
Association of markers with ecogeographical factors
The loci that show unusually low or high levels of genetic differentiation are often assumed to be under natural selection [ 64 ]. The accessions that show association of alleles of the locus GWM1302 with low precipitation belong mainly to the group I, and were collected from dry places, thereby suggesting that this environmental factor is involved in a local adaptation after colonization.
Significant correlations between microsatellite markers and ecogeographical factors have been observed in several studies in wild wheat [ 65 ] and in wild barley [ 66 ], suggesting the impact of natural selection on these markers by creating regional divergence.
Genetic clustering in a principal component analysis revealed that the combination of geographic and ecological data, such as the latitude with rainfall, as the main contributor to the genetic structure of the H. chilense germplasm. The second principal component explained by longitude and temperature significantly contributed to the separation of the two groups. Hübner et al. [ 21 ] studied the population structure in Hordeum spontaneum and found a strong correlation of population structure with temperature and precipitation. In our study, the genetic structure of the analyzed germplasm showed a correlation with morphological and ecophysiological characteristics, influenced also to a minor extent by geographic and ecological factors. | Conclusions
Our study shows the utility of barley EST-SSR for the genetic analysis of H. chilense , with a remarkably high level of polymorphism within this species, despite of the evolutionary distance between the wheat and barley genera. The current set of SSR markers available for Hordeum chilense , which includes wheat and barley gSSRs and EST-SSRs, is useful to analyze the genetic structure and ecogeographical adaptation of H. chilense wild barley populations. Both wheat and barley represent a good source of markers for genetic diversity and structure studies of H. chilense germplasm collections. In addition, they represent an invaluable tool for the introgression of H. chilense alleles into other cereal species, and are useful as anchors for the syntenic maps. The analyzed germplasm can be divided into two groups, with morphological and ecophysiological characteristics being key determinants of the population structure. Geographic and ecological structuring was also revealed in the analyzed germplasm. A significant correlation between geographical and genetic distance was detected in the Central Chilean region for the first time in the species. In addition, significant ecological influence in genetic distance has been detected for one of the population structure groups (group II) in the Central Chilean region. Finally, one marker was found significantly associated with precipitation. These findings have a potential application in cereal breeding. | Background
Multi-allelic microsatellite markers have become the markers of choice for the determination of genetic structure in plants. Synteny across cereals has allowed the cross-species and cross-genera transferability of SSR markers, which constitute a valuable and cost-effective tool for the genetic analysis and marker-assisted introgression of wild related species. Hordeum chilense is one of the wild relatives with a high potential for cereal breeding, due to its high crossability (both interspecies and intergenera) and polymorphism for adaptation traits. In order to analyze the genetic structure and ecogeographical adaptation of this wild species, it is necessary to increase the number of polymorphic markers currently available for the species. In this work, the possibility of using syntenic wheat SSRs as a new source of markers for this purpose has been explored.
Results
From the 98 wheat EST-SSR markers tested for transferability and polymorphism in the wild barley genome, 53 primer pairs (54.0%) gave cross-species transferability and 20 primer pairs (20.4%) showed polymorphism. The latter were used for further analysis in the H. chilense germplasm. The H. chilense - Triticum aestivum addition lines were used to test the chromosomal location of the new polymorphic microsatellite markers. The genetic structure and diversity was investigated in a collection of 94 H. chilense accessions, using a set of 49 SSR markers distributed across the seven chromosomes. Microsatellite markers showed a total of 351 alleles over all loci. The number of alleles per locus ranged from two to 27, with a mean of 7.2 alleles per locus and a mean Polymorphic Information Content (PIC) of 0.5.
Conclusions
According to the results, the germplasm can be divided into two groups, with morphological and ecophysiological characteristics being key determinants of the population structure. Geographic and ecological structuring was also revealed in the analyzed germplasm. A significant correlation between geographical and genetic distance was detected in the Central Chilean region for the first time in the species. In addition, significant ecological influence in genetic distance has been detected for one of the population structure groups (group II) in the Central Chilean region. Finally, the association of the SSR markers with ecogeographical variables was investigated and one marker was found significantly associated with precipitation. These findings have a potential application in cereal breeding. | Authors' contributions
AC carried out the experiments and drafted the manuscript. CF, PS, GD and PH were involved in designing and planning the work, interpreting the results and critically editing the manuscript. PH, CF and PS conceived the study. PH coordinated the study and helped to draft the manuscript. All authors have read and approved the final manuscript.
Supplementary Material | Acknowledgements
Financial support from the Spanish Ministry of Science and Innovation (MICCIN-FEDER grant AGL2010-17316) is acknowledged. We gratefully acknowledge Dr Marion Röder for providing GWM primer sequences. The authors are grateful for the careful sampling and maintenance of the Hordeum chilense germplasm (see materials and methods), and for the data made publicly available by the DIVA-GIS project. This work is dedicated to the memory of Dr. Juan Ballesteros. | CC BY | no | 2022-01-12 15:21:37 | BMC Plant Biol. 2010 Nov 30; 10:266 | oa_package/c5/c2/PMC3014967.tar.gz |
PMC3014968 | 21159189 | Background
Citrus trees produce non-climacteric hesperidium fruits with outstanding agricultural and economic relevance. At the morphological level, citrus fruits consist of an inner edible flesh (endocarp), an intermediate spongy layer (albedo or mesocarp) and an external coloured peel containing pigments and essential oils (flavedo or epicarp). Fruit development in oranges has been divided into three consecutive phases, characterized by a high rate of cell division but slow fruit growth during approximately two months after anthesis (phase I), a second phase of rapid increase in fruit size due to cell enlargement and water accumulation (phase II), and finally a phase of very reduced rate of fruit growth and ripening (phase III) [ 1 ].
Citrus fruit maturation shows specific features in flesh and flavedo tissues. Whereas internal maturation in the flesh is accompanied by an increase in the content of solutes and a decrease in acidity, external maturation is typically characterized by a change in colour from green to orange caused by the concomitant catabolism of chlorophylls and the synthesis of carotenoids [ 2 - 4 ]. Under specific environmental conditions, the changes in colour occurring in flavedo may be reversible and are affected by endogenous factors, such as nutrients (sucrose and nitrogen) and phytohormones (gibberellins and ethylene) [ 5 - 7 ]. The biochemical pathways underlying these transformations of pigments have been partially elucidated. Ethylene-induced chlorophyllase activity and gene expression has been negatively related to chlorophyll content suggesting the involvement of the enzyme in colour breakdown of flavedo [ 8 - 10 ]. The characteristic orange colouration of oranges and mandarins is due to the accumulating carotenoids in chromoplasts, particularly oxygenated derivatives (β,β-xanthophylls) and several specific carotenoid cleavage products (apocarotenoids) [ 11 ]. Citrus genes coding for enzymes involved in the synthesis and modification of carotenoids have been previously isolated and their evolution during natural and ethylene-induced ripening described [ 12 - 16 ].
Despite such extensive analysis of the physiological and biochemical aspects of fruit external maturation, studies describing induced or natural mutants showing an altered pattern or timing of colour acquisition are scarce yet. Among them, the orange ( Citrus sinensis L. Osbeck) mutant Pinalate produced yellow-coloured fruits due to an unusually high accumulation of linear carotenes instead of cyclic and oxygenated carotenoids. The mutant also exhibited reduced synthesis of ABA. However, the specific alteration of the carotenoid biosynthesis pathway in Pinalate is currently unknown [ 17 ]. The nan spontaneous mutation of 'Washington Navel' orange, as formerly characterized in our group, showed an abnormal brown colour in the ripe flavedo caused by a defective catabolism of chlorophylls. Transcript profiling indicated that a SGR -like ( STAY-GREEN ) gene was expressed at lower levels in nan flavedo, suggesting that nan mutation could be associated to a SGR -like upstream regulatory factor [ 18 ]. Recently, the delay in fruit colouration displayed by the slow-ripening clementine mutant 'Tardivo' ( Citrus clementina Hort. Ex Tan.) has been associated with altered expression of carotenoid biosynthetic genes and different sensitivity to the exogenous application of ethylene [ 19 ].
As part of a mutagenic approach to citrus functional genomics, our group established a collection of near 10,000 independent Citrus clementina mutants obtained by fast neutrons bombardment, which were expected to contain genomic deletions in hemizygous dosage. Two of these mutants, called 39B3 and 39E7, were molecularly characterized by array-Comparative Genomic Hybridization for the identification of deleted genes. The structure of 39B3 deletion, determined at the BAC resolution, contained more than 21 identified genes spanning a large genomic region [ 20 ]. Phenotypic evaluation for several consecutive years demonstrated that 39B3 and 39E7 mutants have a significant delay in external fruit colour break. In this work we complete the phenotypic characterization and provide the transcriptomic profiling of flavedo from these mutants. | Methods
Plant material
About 7 years-old clementine trees ( Citrus clementina Hort. Ex Tan. cv. clemenules) were grown at the Instituto Valenciano de Investigaciones Agrarias (IVIA) under standard agricultural practices. The 39B3 and 39E7 mutants were obtained by fast neutrons irradiation of clemenules buds [ 43 ].
GA 3 treatment
Individually labelled fruits were periodically treated on-tree with 60 mg/L gibberellin A 3 (GA 3 ) (Sigma). In each treatment, fruits on four adult trees were sprayed every 3 days from October 7 (189 days after anthesis) to November 18 (231 days after anthesis). After colour index determination (see below), flavedo tissue from treated and untreated trees was collected at three different dates: September 12 (previous to GA 3 treatment), October 25 and November 18 [ 15 ].
Colour index determination
The L, a , and b Hunter lab parameters of the colour system were measured on the flavedo surface with a Minolta CR-200 chromameter. The values presented are the results of the 1000 a /L b transformation that results in negative and positive values for the green and orange colours, respectively, in citrus fruit [ 44 ]. In this transformation, the zero value coincides with the midpoint of the colour break period. Eight and twenty fruits were measured per sample for the colour change curve and the GA experiment respectively.
Extraction and quantification of chlorophylls and carotenoids
Flavedo pigments were extracted as previously described [ 17 ]. Briefly, frozen ground material (500 mg) of flavedo was extracted with a mixture of methanol and 50 mM Tris-HCl buffer (pH 7.5) containing 1 M NaCl and partitioned against chloroform until plant material was uncoloured. The chlorophyll ( a + b ) content was determined by measuring the absorbance of the extracts at 644 nm and 662 nm and calculated according to the Smith and Benitez equations [ 45 ]. After chlorophylls measurement, the pigment ethereal solution was dried and saponified using a KOH methanolic solution. The carotenoids were subsequently re-extracted with diethyl ether. Extracts were dried under N 2 and kept at -20 oC until HPLC analysis. Prior to HPLC analysis, carotenoid extracts were dissolved in acetone and incubated overnight at -20oC to precipitate sterols that could interfere in the carotenoid analysis and subsequently dried under N 2 .
Carotenoid composition of each sample was analyzed by HPLC with a Waters liquid chromatography system equipped with a 600E pump and a model 996 photodiode array detector, and Empower software (Waters). A C30 carotenoid column (250 × 4.6 mm, 5 μm) coupled to a C30 guard column (20 × 4.0 mm, 5 μm) (YMC Europe GMBH) was used. Samples were prepared for HPLC by dissolving the dried carotenoid extracts in CHCl 3 : MeOH: acetone (3:2:1, v:v:v). A ternary gradient elution with MeOH, water and methyl tert -butyl ether (MTBE) was used for carotenoid separation reported in previous works [ 17 , 46 ]. Briefly, the initial solvent composition consisted of 90% MeOH, 5% water and 5% MTBE. The solvent composition changed in a linear fashion to 95% MeOH and 5% MTBE at 12 min. During the next 8 min the solvent composition was changed to 86% MeOH and 14% MTBE. After reaching this concentration the solvent was gradually changed to 75% MeOH and 25% MTBE at 30 min. Final composition was reached at 50 min and consisted of 50% MeOH and 50% MTBE. Initial conditions were re-established in 2 min and re-equilibrated for 15 min before next injection. The flow rate was 1 mL/min, column temperature was set to 25°C and the injection volume was 20 μL. The photodiode array detector was set to scan from 250 to 540 nm, and for each elution a Maxplot chromatogram was obtained, which plots each carotenoid peak at its corresponding maximum absorbance wavelength. Carotenoids were identified by comparison of the spectra and retention time with those of authentic standards, when available, or by matching the observed versus literature spectral data and retention time under identical chromatographic conditions [ 12 , 46 , 47 ]. The carotenoid peaks were integrated at their individual maxima wavelength and their content were calculated using calibration curves of β-apo-8'-carotenal (a gift from Hoffman-LaRoche) for apo-8'-carotenal and β-citraurin, β-cryptoxanthin (Extrasynthese) for α- and β-cryptoxanthin, lutein (Sigma) for lutein, neoxanthin, violaxanthin isomers and mutatoxanthin, zeaxanthin (Extrasynthese) for zeaxanthin and antheraxanthin, and β-carotene (Sigma). Standards of phytoene, phytofluene and ζ-carotene for quantification were obtained from flavedo extracts of Pinalate fruits, which accumulate large amounts of these compounds [ 17 ], and afterward purified by TLC.
Samples were extracted at least twice and each analytical determination was replicated at least once. All operations were carried out on ice under dim light to prevent photodegradation, isomerisations and structural changes of carotenoids.
Expression profiling
Total RNA was isolated from flavedo of clementine and mutant fruits collected in November, using RNeasy Plant Mini Kit (Qiagen). The transcripts present in 1.5 μg of total RNA were reverse-transcribed, amplified and labelled with the Amino Allyl MessageAmpTMII aRNA Amplification kit (Ambion), following the manufacturer's instructions. Cy3 and Cy5 fluorescent dyes coupled to the aRNA were obtained from the CyDyeTMPost-Labeling Reactive Dye Pack (Amersham). Purified Cy5 and Cy3 labelled probes (200 pmol each) were combined, diluted with water to a final volume of 9 μL, and fragmented using the RNA Fragmentation Reagents (Ambion). Fragmented samples were heat-denatured for 2 min at 80 oC, mixed with 50 μL of pre-heated hybridization buffer (5 × SSC, 50% formamide, 0.1% SDS, 0.1 mg/mL salmon sperm DNA) and applied to the microarray slide prehybridized in 5 × SSC, 0.1% SDS, 1% BSA, for at least 1 h at 42 oC. We employed the 20 K Citrus cDNA microarrays containing 21240 EST generated by the Spanish Citrus Functional Genomics Project [ 22 , 48 , 49 ]. Three biological replicates of each mutant were compared to three replicates of control in a dye-swap experiment requiring six slides per mutant.
Hybridization was performed overnight at 42 oC. After hybridization, slides were washed 5 min twice at 42 oC in 2 × SSC, 0.1% SDS followed by two washes at room temperature for 5 min in 0.1 × SSC, 0.1% SDS, then by 5 washes at room temperature for 1 min in 0.1 × SSC and rinsed briefly in 0.01 × SSC before drying by centrifugation at 300 rpm 5 min.
Arrays were scanned at 5 μm. Cy3 and Cy5 fluorescence intensity was recorded by using a ScanArray Gx (Perkin Elmer). The resulting images were overlaid and spots identified by the ScanArray Express program (Perkin Elmer). Spot quality was first measured by the signal-to-background method with parameters lower limit (200) and multiplier (2), and subsequently confirmed by visual test. Data analysis was performed using the Limma package from the R statistical computing software [ 50 ]. A mutant/wild type signal higher than 2 or lower than 0.5, with a P-value not higher than 10 -5 were the cut-off values for positive EST identification. The 39B3 and 39E7 microarray experiments have been loaded into the ArrayExpress database under accessions E-MEXP-2638 and E-MEXP-2641, respectively.
Quantitative RT-PCR
Total RNA was isolated from excised flavedo using RNeasy Plant Mini Kit (Qiagen). RNA concentration was determined by a fluorometric assay with the RiboGreen dye (Molecular Probes) following the manufacturer's instructions. Five μg of total RNA was reverse transcribed with the SuperScript III First-Strand Synthesis System for RT-PCR (Invitrogen) in a total volume of 20 μL. Two μL of a 20 times diluted first-strand cDNA was used for each amplification reaction. Quantitative real-time PCR was performed on a LightCycler 2.0 instrument (Roche), using the LightCycler FastStart DNA MasterPLUS SYBR Green I kit (Roche). Reaction composition and conditions followed manufacturer's instructions. The primers employed were 5'-CCGAGAAGTTGTCTGAGCTAGA-3' and 5'-CCCACAAGACTGCTTTTCTTCT-3', which amplified a fragment of 164 base pairs on a cDNA template. Cycling protocol consisted of 10 min at 95°C for pre-incubation, then 40 cycles of 10 sec at 95°C for denaturation, 10 sec at 60°C for annealing and 10 sec at 72°C for extension. Fluorescent intensity data were acquired during the extension time. Specificity of the PCR reaction was assessed by the presence of a single peak in the dissociation curve after the amplification and through size estimation of the amplified product by gel electrophoresis. For expression measurements, we used the absolute quantification analysis from the LightCycler Software 4.0 package (Roche), and calculated expression levels relative to values of a reference sample. Reference sample was flavedo from the parental clementine in Figure 5 and flavedo tissue before GA application in Figure 6 . Results were the average of 3 independent biological replicates repeated twice.
Phylogenetic analysis
The fused sequences of the GARP and coiled-coil domains of CcGCC1 and other 18 proteins obtained or deduced from databases were aligned with the ClustalW2 program [ 51 ]. Phylogenetic analysis was performed using programs from the PHYLIP.
group, PHYLogeny Inference Package, Version 3.6 [ 52 , 53 ]. A distance matrix was computed according to the Dayhoff PAM model by the program Protdist and then it was used as input for the program Neighbor, where the Neighbor-joining method of clustering was selected. A bootstrap analysis based on 1000 replicates was performed. CrPSR1 from Chlamydomonas reinhardtii was defined as the outgroup species. | Results and discussion
Delay of colour change in 39B3 and 39E7 mutants
Mutants 39B3 and 39E7 showing delayed fruit colour break for several consecutive years were obtained from a population of near 10,000 Citrus clementina plants mutagenized by fast neutrons irradiation. Fruits from 39B3 and 39E7 retained an appreciable greenish colour at the end of November, while fruits from the non-mutagenized parental (for simplification designated as clementine in this work) had already initiated the shift to orange at this time (Figure 1A ). A previous structural analysis of the hemizygous genomic deletions found in these mutants reported large DNA lesions containing a high number of genes, but no evidences of overlapping regions in the 39B3 and 39E7 deletions were observed [ 20 ]. In order to characterize the nature of colour break alterations affecting these mutants, changes in flavedo colour index (CI) were measured throughout fruit development in both mutants. As shown in Figure 1B , flavedo CI in clementine followed a sigmoid curve shifting from negative (green colour) to positive values (orange colour), approximately at mid November. The pattern of colour change in 39B3 fruits showed a similar behaviour but with a delay of three-four weeks. In 39E7 mutants, however, CI increased at a slower rate (Figure 1B ) and reached lower final values than the clementine and 39B3 plants (Figure 1C ). These observations suggested that 39B3 and 39E7 mutations affect fruit external ripening in distinct ways; the 39B3 mutation causes a simple delay in flavedo colour change, while the 39E7 mutant is characterized by a reduced rate of colour acquisition leading to an unusual yellowish external appearance after full ripening.
Chlorophyll and carotenoids accumulation in the mutants
In order to gain a deeper insight into the biochemical alterations affecting 39B3 and 39E7 mutants, total chlorophylls and total and individual carotenoids were determined in flavedo from both mutants and parental fruits at two different developmental stages: in November when colour differences with clementine were more evident (Figure 1A ) and in January when all three genotypes had reached the final colouration.
Total chlorophyll pigments were four to five-fold more abundant in 39B3 and 39E7 than in the parental (Table 1 ), which suggests a slower chlorophyll catabolism or a general delay of maturation. These data were in close agreement with the greenish appearance of mutant fruits in November, while clementine fruits were already changing to orange (Figure 1A ). No chlorophylls were however detected two months later, when the three genotypes showed an intense orange (clementine and 39B3) or yellowish (39E7) colouration. Total carotenoids were in 39B3 lower than in clementine in November as expected from the delayed colour break in this mutant, and also in January. On the contrary, 39E7 mutant showed a higher amount of carotenoids in November, but no significant statistical differences were observed with respect to the parental in January.
The profile of individual carotenoids obtained in the flavedo of clementine essentially coincided with previous reports in this variety, characterized by a reduction in β,ε-carotenoids and neoxanthin and an increase of specific β,β-xanthophylls during ripening [ 15 ]. In January, the 39B3 mutant exhibited a carotenoid profile very similar to that of the clementine. However, the pattern of pigment distribution in 39B3 in November differed significantly from the parental, showing lower percentages of phytoene, phytofluene and β-citraurin, and higher amounts of β-carotene, neoxanthin, α-cryptoxanthin and lutein (Table 1 ), characteristics of chloroplastic tissues, in good agreement with the delayed external colouration in 39B3 fruit. Other β,β-xanthophylls more typical of chromoplastic citrus peel, as β-cryptoxanthin, anteraxanthin and 9- Z -violaxanthin, were found in a percentage higher than expected, however the lower amount of total carotenoids in 39B3 indicated a roughly similar absolute accumulation of them in both 39B3 and clementine.
The carotenoid profile of 39E7 mutant showed common features in November and January. In both samples, the absence of the apocarotenoid β-citraurin (C 30 ) was associated with a higher accumulation of the xanthophylls β-cryptoxanthin and zeaxanthin. This observation is of special significance because despite the relevant contribution of β-citraurin, a red-orange pigment, to the typical peel colour of oranges and mandarines [ 11 , 21 ], the specific cleavage reaction producing this C 30- apocarotenoid has not been yet elucidated. The total absence of β-citraurin in fully ripened flavedo of 39E7 mutant suggests that such cleavage reaction could be impaired in this genotype, leading to its distinctive pale yellowish peel. The concomitant increase of β-cryptoxanthin and zeaxanthin in 39E7 might indicate a substrate-product relationship between them and β-citraurin, reinforcing previous suggestions [ 11 , 12 , 21 ]. Such alteration in the carotenoid biosynthesis pathway corroborates at the biochemical level colour-based observations on the different developmental defects affecting 39B3 and 39E7 mutants. However we cannot rule out the presence of multiple mutations in 39E7 leading to separate effects on colour break delay and carotenoid accumulation. Under this assumption, the observed delay in external colouration could be caused by the same locus in both mutants.
Differential expression profiling in flavedo
The availability of 39B3 and 39E7 mutants has been exploited to identify major factors involved in regulation of fruit maturation through the transcriptomic analysis of flavedo tissue from these mutants. We took advantage of a citrus cDNA microarray previously described [ 22 ] to perform large scale hybridization experiments comparing mRNA isolated from green flavedo of both mutants and clementine flavedo undergoing colour break collected the same day. After microarray hybridization and analysis, cDNAs showing a signal intensity more than double or less than half of control, under a P-value threshold of 10 -5 , were considered as differentially expressed genes. Signal ratios and false discovery rates of selected genes have been included as supplementary material in Additional file 1 . As shown in Figure 2 from the 503 and 165 cDNAs overexpressed in 39B3 and 39E7 mutants, respectively, 73 were common. Similarly, a relatively high percentage of down-regulated cDNAs were shared by 39B3 and 39E7 flavedos (90 from 236 and 273, respectively). The occurrence of common transcripts confirms the alteration of particular transcriptional programs in both mutants, which could be revealed by data mining of these coincident clones. None of the 90 cDNAs that were found to be simultaneously down-regulated in both mutants were coincident with the known deleted genes of 39B3 and 39E7. Therefore, they are not expected to reduce their expression as a consequence of their occurrence in a genomic deletion. However, additional deletions to those reported in the published structural characterization of the 39B3 and 39E7 hemizygous deletions [ 20 ] might occur in the genome of these mutants and consequently we cannot elucidate whether or not a certain down-regulated gene is included in a deleted fragment.
Functional classification of cDNAs differentially expressed in both 39B3 and 39E7 mutants highlighted major biochemical features underlying peel colour progression (Table 2 and 3 ). Thus, "photosynthesis" was one of the pivotal enriched categories in the mutants due to the presence in flavedo of photosynthetically active green chloroplasts. Several cDNAs coding for proteins involved in light harvesting, photosynthetic electron transfer chain, Calvin cycle and chlorophyll biosynthesis were overexpressed in mutant samples (Table 2 ), while no "photosynthesis" category could be created in the list of underexpressed cDNAs (Table 3 ).
Similarly, known biochemical and physiological features of citrus fruit flavedo at an advanced maturation stage, such as substitution and accumulation of secondary metabolites and cell wall degradation properly correlated with the enrichment and large size of the functional category "metabolism" and to a lesser extend with the category of "cell wall modification" (Table 2 and 3 ). For example, a valencene synthase responsible for the accumulation of valencene, an important sesquiterpene in the aroma of ripened citrus fruits [ 23 ], is catalogued as a down-regulated gene (Table 3 ). Conversely, a γ-terpinene synthase, involved in the biosynthesis of the monoterpene γ-terpinene in immature green fruits [ 24 ], is in the list of up-regulated genes (Table 2 ).
Interestingly, the transcriptomic study revealed that colour change appears to be also highly dependent upon a major transport activity. The most striking and novel observation in this regard was the high number of putative transporters of mineral elements and metals included in the functional category of "transport" that were down-regulated in the green flavedo of both mutants (Table 3 ). Thus, several sulfate and nitrate transporters, including a membrane transporter NRT1 .2 implicated in chloride homeostasis [ 25 ], generic metal membrane transporters and specific zinc transporters were common in this category, suggesting that the mobilization of mineral elements such as sulfur, nitrogen, chloride, zinc and other metals may play a relevant role in flavedo ripening. The presence of a gene coding for a ferritin-like protein in the listing of down-regulated cDNAs ([GenBank: CX307912 ]; Table 3 ) may exemplify the relevance of these transporters in the colour-break flavedo. Plant ferritins have been described as chloroplastic and mitochondrial proteins involved in Fe(II) oxidation and Fe(III) storage, protecting the cells from the oxidative damage caused by reactive oxygen species produced by free iron [ 26 , 27 ]. For instance, limited iron availability in Chlamydomonas reinhardtii has been postulated to induce ferritin coding genes in order to buffer iron released by the degradation of photosystem I (PSI), an important sink for this metal [ 28 ]. Similarly, a related ferritin-like gene, up-regulated during leaf senescence in Brassica napus , has been proposed to be involved in mobilization of iron from senescing cells to developing organs, where the metal is highly required [ 29 ]. Thus, ferritin accumulation in clementine flavedo tissue undergoing colour break may apparently contribute to the sequestering and recycling of iron molecules released during the degradation of photosystems and light-harvesting complexes, at the transition from chloroplast to chromoplast. The membrane transporters listed in Table 3 could initiate subsequent mobilization of the sequestered iron and maybe other metals and mineral elements to the cells requiring them.
The category of "transport" was also enriched with cDNAs coding for other several kinds of transporters including ABC transporters, sugar and protein transporters, aquaporin, H + -ATPases and other unidentified membrane transporters associated with the green stage of the flavedo (Table 3 ).
A MYB-related transcription factor down-regulated in 39B3 and 39E7
Whereas nine different cDNAs coding for transcriptional regulators, including three ethylene response factors, were up-regulated in 39B3 and 39E7 (Table 2 ), only one was down-regulated in both mutants ([GenBank: CX287481 ]; Table 3 ). This transcription factor belongs to a subgroup of the GARP ( G OLDEN2, AR R-B and P sr1) subfamily of MYB-related proteins containing a coiled-coil domain, which has been recently designated GCC ( G ARP and c oiled- c oil) [ 30 , 31 ]. Consequently, we named the protein deduced from this cDNA CcGCC1 (for Citrus clementina GCC ). This gene was not found in a previous genomic approach to identify deleted genes in 39B3 and 39E7 mutants [ 20 ], and hence no gene dosage effects are expected to contribute to lower its expression in the mutants.
The partial sequence of CcGCC1 cDNA annotated in clone [GenBank: CX287481 ] was completed by sequencing its 3' end. The resulting nucleotide and amino acid sequences are shown in Figure 3 . Database similarity search by BLASTP analysis [ 32 ] of the 233 residues long protein deduced from the cDNA confirmed a high similarity to other members of the GCC subgroup. We used the SMART [ 33 ] and COILS [ 34 ] applications to localize the GARP DNA-binding and the coiled-coil domains respectively, which are highlighted in Figure 3 .
In order to compare CcGCC1 and other related citrus ESTs with known members of this GCC subgroup, we selected a 90 amino acids long fragment fusing GARP and coiled-coil domains of CcGCC1 and several homologous proteins and translated ESTs [ 35 - 41 ]. The phylogenetic tree of these proteins showed two major groups with CcGCC1 clustered with PHR1 from Arabidopsis thaliana , a protein involved in phosphate starvation signalling (Figure 4 ). The closest homolog to CcGCC1 among those polypeptides was [GenBank: AAT06477 ], coded by At5g06800 gene from Arabidopsis . Interestingly, a search into the AtGenExpress database containing microarray expression data of Arabidopsis genes revealed that At5g06800 is mostly expressed in tissues lacking chloroplasts such as roots and to a lesser extent pollen and flower organs [ 42 ].
CcGCC1 expression correlates with colour change processes
In order to investigate the time-dependent expression of CcGCC1 during fruit external maturation in clementine, flavedo tissues collected before (September), during (November) and after fruit colour break (January), were subject to RNA extraction and quantitative RT-PCR with CcGCC1 specific primers. Figure 5A shows that the expression level of CcGCC1 in clementine increased about 15-fold during flavedo ripening, while 39B3 mutant maintained low expression levels in November and only experienced a slight increase in January. A similar change on CcGCC1 expression was observed in 39E7 mutant when samples harvested in November were assayed (Figure 5B ). These results confirmed that CcGCC1 gene expression was induced during colour break of clementine fruits whereas mutants 39B3 and 39E7 affected in the rate of colour break were unable to properly express the gene.
To determine if the expression of CcGCC1 gene was also responsive to other factors modulating colour change, a further experiment using external applications of gibberellins was performed (Figure 6 ). Gibberellins (GA) operate as colour change retardants during fruit external maturation since GA application on green flavedo causes a significant delay in colour break [ 14 , 15 ]. Forty-two days after the first application, fruits treated periodically with gibberellin A 3 showed a delay of about 10 colour units with respect to untreated fruits (Figure 6A ). Interestingly, the GA-dependent retard in peel colour was accompanied by a parallel delay in CcGCC1 induction (Figure 6B ). These results indicate that CcGCC1 also responds to the GA-dependent pathway regulating flavedo ripening and taken together with the previous observations suggest the participation of CcGCC1 in a regulatory pathway acting in parallel or subsequently to colour break processes. | Results and discussion
Delay of colour change in 39B3 and 39E7 mutants
Mutants 39B3 and 39E7 showing delayed fruit colour break for several consecutive years were obtained from a population of near 10,000 Citrus clementina plants mutagenized by fast neutrons irradiation. Fruits from 39B3 and 39E7 retained an appreciable greenish colour at the end of November, while fruits from the non-mutagenized parental (for simplification designated as clementine in this work) had already initiated the shift to orange at this time (Figure 1A ). A previous structural analysis of the hemizygous genomic deletions found in these mutants reported large DNA lesions containing a high number of genes, but no evidences of overlapping regions in the 39B3 and 39E7 deletions were observed [ 20 ]. In order to characterize the nature of colour break alterations affecting these mutants, changes in flavedo colour index (CI) were measured throughout fruit development in both mutants. As shown in Figure 1B , flavedo CI in clementine followed a sigmoid curve shifting from negative (green colour) to positive values (orange colour), approximately at mid November. The pattern of colour change in 39B3 fruits showed a similar behaviour but with a delay of three-four weeks. In 39E7 mutants, however, CI increased at a slower rate (Figure 1B ) and reached lower final values than the clementine and 39B3 plants (Figure 1C ). These observations suggested that 39B3 and 39E7 mutations affect fruit external ripening in distinct ways; the 39B3 mutation causes a simple delay in flavedo colour change, while the 39E7 mutant is characterized by a reduced rate of colour acquisition leading to an unusual yellowish external appearance after full ripening.
Chlorophyll and carotenoids accumulation in the mutants
In order to gain a deeper insight into the biochemical alterations affecting 39B3 and 39E7 mutants, total chlorophylls and total and individual carotenoids were determined in flavedo from both mutants and parental fruits at two different developmental stages: in November when colour differences with clementine were more evident (Figure 1A ) and in January when all three genotypes had reached the final colouration.
Total chlorophyll pigments were four to five-fold more abundant in 39B3 and 39E7 than in the parental (Table 1 ), which suggests a slower chlorophyll catabolism or a general delay of maturation. These data were in close agreement with the greenish appearance of mutant fruits in November, while clementine fruits were already changing to orange (Figure 1A ). No chlorophylls were however detected two months later, when the three genotypes showed an intense orange (clementine and 39B3) or yellowish (39E7) colouration. Total carotenoids were in 39B3 lower than in clementine in November as expected from the delayed colour break in this mutant, and also in January. On the contrary, 39E7 mutant showed a higher amount of carotenoids in November, but no significant statistical differences were observed with respect to the parental in January.
The profile of individual carotenoids obtained in the flavedo of clementine essentially coincided with previous reports in this variety, characterized by a reduction in β,ε-carotenoids and neoxanthin and an increase of specific β,β-xanthophylls during ripening [ 15 ]. In January, the 39B3 mutant exhibited a carotenoid profile very similar to that of the clementine. However, the pattern of pigment distribution in 39B3 in November differed significantly from the parental, showing lower percentages of phytoene, phytofluene and β-citraurin, and higher amounts of β-carotene, neoxanthin, α-cryptoxanthin and lutein (Table 1 ), characteristics of chloroplastic tissues, in good agreement with the delayed external colouration in 39B3 fruit. Other β,β-xanthophylls more typical of chromoplastic citrus peel, as β-cryptoxanthin, anteraxanthin and 9- Z -violaxanthin, were found in a percentage higher than expected, however the lower amount of total carotenoids in 39B3 indicated a roughly similar absolute accumulation of them in both 39B3 and clementine.
The carotenoid profile of 39E7 mutant showed common features in November and January. In both samples, the absence of the apocarotenoid β-citraurin (C 30 ) was associated with a higher accumulation of the xanthophylls β-cryptoxanthin and zeaxanthin. This observation is of special significance because despite the relevant contribution of β-citraurin, a red-orange pigment, to the typical peel colour of oranges and mandarines [ 11 , 21 ], the specific cleavage reaction producing this C 30- apocarotenoid has not been yet elucidated. The total absence of β-citraurin in fully ripened flavedo of 39E7 mutant suggests that such cleavage reaction could be impaired in this genotype, leading to its distinctive pale yellowish peel. The concomitant increase of β-cryptoxanthin and zeaxanthin in 39E7 might indicate a substrate-product relationship between them and β-citraurin, reinforcing previous suggestions [ 11 , 12 , 21 ]. Such alteration in the carotenoid biosynthesis pathway corroborates at the biochemical level colour-based observations on the different developmental defects affecting 39B3 and 39E7 mutants. However we cannot rule out the presence of multiple mutations in 39E7 leading to separate effects on colour break delay and carotenoid accumulation. Under this assumption, the observed delay in external colouration could be caused by the same locus in both mutants.
Differential expression profiling in flavedo
The availability of 39B3 and 39E7 mutants has been exploited to identify major factors involved in regulation of fruit maturation through the transcriptomic analysis of flavedo tissue from these mutants. We took advantage of a citrus cDNA microarray previously described [ 22 ] to perform large scale hybridization experiments comparing mRNA isolated from green flavedo of both mutants and clementine flavedo undergoing colour break collected the same day. After microarray hybridization and analysis, cDNAs showing a signal intensity more than double or less than half of control, under a P-value threshold of 10 -5 , were considered as differentially expressed genes. Signal ratios and false discovery rates of selected genes have been included as supplementary material in Additional file 1 . As shown in Figure 2 from the 503 and 165 cDNAs overexpressed in 39B3 and 39E7 mutants, respectively, 73 were common. Similarly, a relatively high percentage of down-regulated cDNAs were shared by 39B3 and 39E7 flavedos (90 from 236 and 273, respectively). The occurrence of common transcripts confirms the alteration of particular transcriptional programs in both mutants, which could be revealed by data mining of these coincident clones. None of the 90 cDNAs that were found to be simultaneously down-regulated in both mutants were coincident with the known deleted genes of 39B3 and 39E7. Therefore, they are not expected to reduce their expression as a consequence of their occurrence in a genomic deletion. However, additional deletions to those reported in the published structural characterization of the 39B3 and 39E7 hemizygous deletions [ 20 ] might occur in the genome of these mutants and consequently we cannot elucidate whether or not a certain down-regulated gene is included in a deleted fragment.
Functional classification of cDNAs differentially expressed in both 39B3 and 39E7 mutants highlighted major biochemical features underlying peel colour progression (Table 2 and 3 ). Thus, "photosynthesis" was one of the pivotal enriched categories in the mutants due to the presence in flavedo of photosynthetically active green chloroplasts. Several cDNAs coding for proteins involved in light harvesting, photosynthetic electron transfer chain, Calvin cycle and chlorophyll biosynthesis were overexpressed in mutant samples (Table 2 ), while no "photosynthesis" category could be created in the list of underexpressed cDNAs (Table 3 ).
Similarly, known biochemical and physiological features of citrus fruit flavedo at an advanced maturation stage, such as substitution and accumulation of secondary metabolites and cell wall degradation properly correlated with the enrichment and large size of the functional category "metabolism" and to a lesser extend with the category of "cell wall modification" (Table 2 and 3 ). For example, a valencene synthase responsible for the accumulation of valencene, an important sesquiterpene in the aroma of ripened citrus fruits [ 23 ], is catalogued as a down-regulated gene (Table 3 ). Conversely, a γ-terpinene synthase, involved in the biosynthesis of the monoterpene γ-terpinene in immature green fruits [ 24 ], is in the list of up-regulated genes (Table 2 ).
Interestingly, the transcriptomic study revealed that colour change appears to be also highly dependent upon a major transport activity. The most striking and novel observation in this regard was the high number of putative transporters of mineral elements and metals included in the functional category of "transport" that were down-regulated in the green flavedo of both mutants (Table 3 ). Thus, several sulfate and nitrate transporters, including a membrane transporter NRT1 .2 implicated in chloride homeostasis [ 25 ], generic metal membrane transporters and specific zinc transporters were common in this category, suggesting that the mobilization of mineral elements such as sulfur, nitrogen, chloride, zinc and other metals may play a relevant role in flavedo ripening. The presence of a gene coding for a ferritin-like protein in the listing of down-regulated cDNAs ([GenBank: CX307912 ]; Table 3 ) may exemplify the relevance of these transporters in the colour-break flavedo. Plant ferritins have been described as chloroplastic and mitochondrial proteins involved in Fe(II) oxidation and Fe(III) storage, protecting the cells from the oxidative damage caused by reactive oxygen species produced by free iron [ 26 , 27 ]. For instance, limited iron availability in Chlamydomonas reinhardtii has been postulated to induce ferritin coding genes in order to buffer iron released by the degradation of photosystem I (PSI), an important sink for this metal [ 28 ]. Similarly, a related ferritin-like gene, up-regulated during leaf senescence in Brassica napus , has been proposed to be involved in mobilization of iron from senescing cells to developing organs, where the metal is highly required [ 29 ]. Thus, ferritin accumulation in clementine flavedo tissue undergoing colour break may apparently contribute to the sequestering and recycling of iron molecules released during the degradation of photosystems and light-harvesting complexes, at the transition from chloroplast to chromoplast. The membrane transporters listed in Table 3 could initiate subsequent mobilization of the sequestered iron and maybe other metals and mineral elements to the cells requiring them.
The category of "transport" was also enriched with cDNAs coding for other several kinds of transporters including ABC transporters, sugar and protein transporters, aquaporin, H + -ATPases and other unidentified membrane transporters associated with the green stage of the flavedo (Table 3 ).
A MYB-related transcription factor down-regulated in 39B3 and 39E7
Whereas nine different cDNAs coding for transcriptional regulators, including three ethylene response factors, were up-regulated in 39B3 and 39E7 (Table 2 ), only one was down-regulated in both mutants ([GenBank: CX287481 ]; Table 3 ). This transcription factor belongs to a subgroup of the GARP ( G OLDEN2, AR R-B and P sr1) subfamily of MYB-related proteins containing a coiled-coil domain, which has been recently designated GCC ( G ARP and c oiled- c oil) [ 30 , 31 ]. Consequently, we named the protein deduced from this cDNA CcGCC1 (for Citrus clementina GCC ). This gene was not found in a previous genomic approach to identify deleted genes in 39B3 and 39E7 mutants [ 20 ], and hence no gene dosage effects are expected to contribute to lower its expression in the mutants.
The partial sequence of CcGCC1 cDNA annotated in clone [GenBank: CX287481 ] was completed by sequencing its 3' end. The resulting nucleotide and amino acid sequences are shown in Figure 3 . Database similarity search by BLASTP analysis [ 32 ] of the 233 residues long protein deduced from the cDNA confirmed a high similarity to other members of the GCC subgroup. We used the SMART [ 33 ] and COILS [ 34 ] applications to localize the GARP DNA-binding and the coiled-coil domains respectively, which are highlighted in Figure 3 .
In order to compare CcGCC1 and other related citrus ESTs with known members of this GCC subgroup, we selected a 90 amino acids long fragment fusing GARP and coiled-coil domains of CcGCC1 and several homologous proteins and translated ESTs [ 35 - 41 ]. The phylogenetic tree of these proteins showed two major groups with CcGCC1 clustered with PHR1 from Arabidopsis thaliana , a protein involved in phosphate starvation signalling (Figure 4 ). The closest homolog to CcGCC1 among those polypeptides was [GenBank: AAT06477 ], coded by At5g06800 gene from Arabidopsis . Interestingly, a search into the AtGenExpress database containing microarray expression data of Arabidopsis genes revealed that At5g06800 is mostly expressed in tissues lacking chloroplasts such as roots and to a lesser extent pollen and flower organs [ 42 ].
CcGCC1 expression correlates with colour change processes
In order to investigate the time-dependent expression of CcGCC1 during fruit external maturation in clementine, flavedo tissues collected before (September), during (November) and after fruit colour break (January), were subject to RNA extraction and quantitative RT-PCR with CcGCC1 specific primers. Figure 5A shows that the expression level of CcGCC1 in clementine increased about 15-fold during flavedo ripening, while 39B3 mutant maintained low expression levels in November and only experienced a slight increase in January. A similar change on CcGCC1 expression was observed in 39E7 mutant when samples harvested in November were assayed (Figure 5B ). These results confirmed that CcGCC1 gene expression was induced during colour break of clementine fruits whereas mutants 39B3 and 39E7 affected in the rate of colour break were unable to properly express the gene.
To determine if the expression of CcGCC1 gene was also responsive to other factors modulating colour change, a further experiment using external applications of gibberellins was performed (Figure 6 ). Gibberellins (GA) operate as colour change retardants during fruit external maturation since GA application on green flavedo causes a significant delay in colour break [ 14 , 15 ]. Forty-two days after the first application, fruits treated periodically with gibberellin A 3 showed a delay of about 10 colour units with respect to untreated fruits (Figure 6A ). Interestingly, the GA-dependent retard in peel colour was accompanied by a parallel delay in CcGCC1 induction (Figure 6B ). These results indicate that CcGCC1 also responds to the GA-dependent pathway regulating flavedo ripening and taken together with the previous observations suggest the participation of CcGCC1 in a regulatory pathway acting in parallel or subsequently to colour break processes. | Conclusions
In this work, the citrus mutants 39B3 and 39E7 showing a delay in fruit colour change have been phenotypically characterized and used to identify through the analyses of their transcriptomes about 160 genes that were directly related to peel ripening. The results indicated that the 39B3 genotype exhibits a simple delay in the rate of flavedo colouration, while 39E7 shows an additional altered pattern of carotenoid accumulation. We postulate that the yellowish colour of fully ripe 39E7 flavedo was due to a defective synthesis or accumulation of β-citraurin. Analyses of differentially expressed genes revealed that colour change during peel ripening was strongly associated with a major mobilization of mineral elements and other previously known metabolic and photosynthetic changes. Transcriptomic data also showed that expression of CcGCC1 gene coding for a transcription factor containing GARP and coiled-coil domains, was strongly down-regulated in flavedo tissue of both mutants. Application of gibberellin to green fruits postponed colour break and abolish the induction of CcGCC1 expression. Taken together the results indicated that CcGCC1 down-regulation correlated with a delay in colour break induced by genetic, developmental and hormonal cues. | Background
External ripening in Citrus fruits is morphologically characterized by a colour shift from green to orange due to the degradation of chlorophylls and the accumulation of carotenoid pigments. Although numerous genes coding for enzymes involved in such biochemical pathways have been identified, the molecular control of this process has been scarcely studied. In this work we used the Citrus clementina mutants 39B3 and 39E7, showing delayed colour break, to isolate genes potentially related to the regulation of peel ripening and its physiological or biochemical effects.
Results
Pigment analyses revealed different profiles of carotenoid and chlorophyll modification in 39B3 and 39E7 mutants. Flavedo from 39B3 fruits showed an overall delay in carotenoid accumulation and chlorophyll degradation, while the flavedo of 39E7 was devoid of the apocarotenoid β-citraurin among other carotenoid alterations. A Citrus microarray containing about 20,000 cDNA fragments was used to identify genes that were differentially expressed during colour change in the flavedo of 39B3 and 39E7 mutants respect to the parental variety. The results highlighted 73 and 90 genes that were respectively up- and down-regulated in both mutants. CcGCC1 gene, coding for a GCC type transcriptional factor, was found to be down-regulated. CcGCC1 expression was strongly induced at the onset of colour change in the flavedo of parental clementine fruit. Moreover, treatment of fruits with gibberellins, a retardant of external ripening, delayed both colour break and CcGCC1 overexpression.
Conclusions
In this work, the citrus fruit ripening mutants 39B3 and 39E7 have been characterized at the phenotypic, biochemical and transcriptomic level. A defective synthesis of the apocarotenoid β-citraurin has been proposed to cause the yellowish colour of fully ripe 39E7 flavedo. The analyses of the mutant transcriptomes revealed that colour change during peel ripening was strongly associated with a major mobilization of mineral elements and with other previously known metabolic and photosynthetic changes. The expression of CcGCC1 was associated with peel ripening since CcGCC1 down-regulation correlated with a delay in colour break induced by genetic, developmental and hormonal causes. | Authors' contributions
GR carried out transcriptomic and sequence analysis and drafted the manuscript. MAN carried out real-time PCR analyses. MJR and LZ performed pigments measurements and contributed to draft the manuscript. EA and MC designed and made the GA experiment. MT conceived the study and assisted in the drafting of the manuscript. All the authors read and approved the final manuscript.
Supplementary Material | Acknowledgements
Work was supported by grants AGL2007-65437-C04-01/AGR (Centro de Genómica) and AGL2009-11558 (L. Zacarías and M. J. Rodrigo) from the Ministerio de Educación y Ciencia of Spain. Assistance and technical expertise of A. Almenar, E Blázquez, I. López, I. Sanchís and M. Sancho from IVIA, and A. Beneyto from IATA are gratefully acknowledged. | CC BY | no | 2022-01-12 15:21:37 | BMC Plant Biol. 2010 Dec 15; 10:276 | oa_package/10/2a/PMC3014968.tar.gz |
PMC3014969 | 21118520 | Background
The modular structure of biochemical reaction networks greatly facilitates the systematic investigation of their design principles [ 1 , 2 ]. In this way it is often possible to identify small functional units called network motifs [ 3 ] which convey a particular functionality. A thorough understanding of the relationship between network design and functionality is not only important for a smart modification and regulation of existing networks, but it is also essential to design novel circuits with prescribed functionality [ 4 , 5 ].
Regulatory properties of cellular networks arise from an interplay between positive and/or negative feedback reactions. These feedback reactions can be effective at the transcriptional level, at the posttranslational level or through allosteric interactions. For example, in transcriptional networks feed forward loops can act as a low pass filter [ 6 ] or as a fold-change detector [ 7 ] depending on the sign of the genetic interactions. Signal transduction cascades often utilize post-translational modifications such as phosphorylation/dephosphorylation cycles to generate ultrasensitivity [ 8 ] or bistability [ 9 ]. This behavior is advantageous for cell fate decisions where irreversible switch-like transitions are required, e.g. during maturation [ 10 ] or cell-cycle progression [ 11 ]. Metabolic enzymes are often regulated through allosteric interactions with positive and/or negative effector molecules. A classical example is the allosteric product activation of the glycolytic enzyme phospho-fructokinase (PFK) which may lead to an oscillatory behavior of the glycolytic pathway [ 12 , 13 ].
So far, the relation between particular molecular reaction mechanisms and the resulting macroscopes behavior has been mainly investigated in well-mixed reaction systems where the dynamics is conveniently described by ordinary differential equations [ 14 , 15 ]. However, if the enzymes in reversible modification cycles are located in different cellular compartments diffusive coupling between neighboring enzyme/substrate molecules may generate steep gradients [ 16 ] resulting in front-like wave propagation of phosphoproteins [ 17 ]. In that way spatially distributed signaling pathways may create step-like activation profiles that can affect the downstream response of the system in a threshold based manner [ 18 ]. Hence, spatial coupling can significantly alter the macroscopic behavior of biochemical reaction systems [ 19 ] and bring about new functionality to network motifs [ 20 , 21 ].
In reaction-diffusion systems spiral shaped concentration waves and stationary Turing patterns are among the most fascinating spatiotemporal structures. While spiral waves can occur in systems with excitable and oscillatory reaction dynamics [ 22 , 23 ] Turing patterns typically emerge in activator-inhibitor systems with long range inhibition [ 24 , 25 ]. Here, we investigate the effect of different mechanisms of product activation on the generation of such reaction-diffusion patterns in an enzymatic reaction system centered around the PFK which is a central part of the glycolytic pathway. Under well-stirred conditions this system exhibits oscillatory behavior in both cell free extracts [ 26 , 27 ] and in living cells [ 28 ], and diffusive coupling was shown to generate waves of glycolytic activity in yeast extracts [ 29 - 31 ]. Recently, we have observed a novel type of spiral wave behavior in that system [ 32 ]. By increasing the overall protein concentration of the extract a transition from outward to inward rotating spiral waves (also known as anti-spirals) was induced. While outward propagating waves have been observed in several biological systems [ 29 , 33 , 34 ] inward rotating spiral waves were, so far, only observed in purely chemical systems [ 35 , 36 ].
Although we could reproduce the inward propagating waves in numerical simulations with the Goldbeter model [ 32 ] the underlying molecular mechanism for their generation is still unclear. The simulations have shown that the negative feedback on the PFK activity, as provided by its substrate ATP, is not required to generate anti-waves. Therefore, we focus here on the allosteric activation of the PFK by its product ADP. Specifically, we address the question whether the symmetry model of Monod, Wyman and Changeux (MWC) [ 37 ], as employed in the Goldbeter model, is necessary to generate inward propagating waves or whether a more simple Hill kinetics, as it was used by Sel'kov [ 12 ] to model the PFK activation, is suffcient. Since the regulatory properties of the PFK play a key role for the emergence of oscillatory behavior in glycolysis [ 27 , 38 ], in particular in yeast extracts [ 39 ], these simple models have been remarkably successful in describing general aspects of glycolytic oscillations [ 12 , 40 ].
The spatio-temporal dynamics of the two PFK effectors ADP and ATP is described by a reaction-diffusion system of the type
where D is a (diagonal) diffusion matrix and p denotes the kinetic parameters. The reaction mechanism is encoded in the form of the function f ( u , p ). Due to the diffusive coupling, the analysis of wave patterns in Eq. 1 is more complicated than the investigation of reaction mechanisms under well-mixed conditions. However, near an oscillatory instability due to a supercritical Hopf bifurcation the spatiotemporal dynamics of Eq. 1 is well described by the complex Ginzburg Landau equation (CGLE) [ 41 ]
whose solution types are well known [ 42 ]. Here, the complex amplitude A ( x , t ) describes slow spatio-temporal modulations around the unstable steady state. At the level of the CGLE the details of the molecular reaction mechanism are encoded in the dependence of the two real parameters c 1 = c 1 ( D, p ) and c 3 = c 3 ( p ) on the original system parameters p and D in Eq. 1. To find the mapping between the two sets of parameters is tedious, but straight-forward [ 41 ] (see Methods). The transition between inward and outward propagating waves is marked by the curve c 1 - c 3 = 0 [ 43 - 45 ] where the region c 1 - c 3 > 0 corresponds to inward propagating waves in Eq. 1.
By explicitly calculating the two CGLE coefficients c 1 and c 3 we show that inward propagating waves can arise in the Goldbeter model due to the sequential binding of product molecules to the allosteric enzyme as implied in the MWC mechanism. In contrast, in the limit of an in finitely large binding a finity, as implicitly assumed in the Sel'kov model, the formation of inward propagating waves is sup-pressed by a Turing instability. We also find a relation between enzyme cooperativity and the occurrence of inward propagating waves. However, it is not the absolute magnitude of the cooperativity which is important here (as it is for the occurrence of oscillations [ 40 ]), but the sensitivity of the co-operativity with respect to changes in the activator concentration. Finally, we observe that the sequential activation mechanism has a stabilizing effect on the wave dynamics. Together, this shows that in the presence of diffusive coupling the particular choice of a molecular mechanism can have a significant impact on the type and the stability of spatio-temporal patterns even though the dynamics under well-mixed conditions is qualitatively the same.
Model Definitions
Sel'kov Model
In the Sel'kov model it is assumed that the PFK is an oligomeric enzyme which has n independent binding sites for the product ADP (activator), but only one binding site for the substrate ATP (inhibitor) [ 12 ]. Simultaneous binding of n product molecules activates the enzyme which allows for subsequent substrate binding and conversion into product with specific rate k . Hence, the Sel'kov model distinguishes only three enzyme states: an inactive state ( T 00 ) which has neither substrate nor product molecules bound and two fully activated states which can have either zero ( R 0 n ) or one ( R 1 n ) substrate molecule bound (cf. Figure 1A ). Substrate molecules are sup-plied at rate ν i and product molecules are used by downstream reactions with specific rate k d . The inhibition of the PFK by ATP at high ATP concentrations is neglected in the Sel'kov model.
Under steady-state conditions for the enzyme binding reactions the local dynamics of the PFK effectors ATP and ADP is described by the (dimensionless) set of equations [ 12 ]
where φ S denotes the fractional saturation. Substrate ( α = ATP / K M ) and product concentrations are measured in terms of the Michaelis-Menten constant and the apparent dissociation constant for product binding . The other parameters are given by and . Time is measured in units of 1/ k d and e 0 denotes the total enzyme concentration.
MWC Model
Based on experimental evidence Goldbeter proposed an alternative approach to describe the allosteric regulation of the PFK [ 13 ] which utilizes the Monod-Wyman-Changeux mechanism [ 37 ]. Here, the free form of the oligomeric enzyme performs concerted transitions between a catalytically active ( R 00 ) and a catalytically inactive ( T 00 ) conformation where the allosteric constant L = k + / k - defines the equilibrium between both conformations in the absence of any ligands (Figure 1B ). The enzyme is activated by sequential binding of product molecules with dissociation constant for each binding step. Hence, there are n + 1 active enzyme forms R 0 m to which substrate molecules can bind to form n + 1 enzyme-substrate complexes R 1 m . Each complex can release product molecules at the specific rate k .
Similar as in the Sel'kov model substrate molecules are supplied at rate ν i and product molecules leave the system with specific rate k d such that the local dynamics of ATP and ADP is described by the same set of equations as in Eqs. 3 with φ S (Eq. 4) being replaced by (see Methods )
The parameters have the same meaning as in Eqs. 3 if the apparent dissociation constant is replaced by the true dissociation constant K P .
Compared with the original Goldbeter model we have neglected the cooperativity with respect to substrate binding and the inhibitory effect of ATP on the PFK activity (as suggested by numerical simulations [ 32 ]). With these simplifications we treat the Sel'kov and the Goldbeter model on an equal footing which allows for a direct comparison between their PFK activation mechanisms. Since our model retains the MWC mechanism for PFK activation as an essential part we shall call it the MWC model. We also remark that the PFK actually exhibits sigmoidal behavior with respect to its second substrate fructose-6-phosphate while it does not show any co-operativity with respect to ATP [ 46 , 47 ]. Hence, the simplifying assumption of a hyperbolic dependence of the PFK activity on ATP as a substrate seems to be reasonable.
Diffusion and Unified Description
The simplest way to incorporate diffusive coupling between the PFK effectors ATP and ADP is to add 'diffusion terms' in Eqs. 3 with constant (effective) diffusion coefficients. Thereby, we neglect complications arising from allosteric interactions between the PFK effectors and the enzyme which may lead to cross-diffusion terms (non-diagonal elements in the diffusion matrix D ) and a dependence of the effective diffusion coefficients on the effector concentrations [ 48 , 49 ]. For a recent review of the effects of cross-diffusion on pattern formation see Ref. [ 50 ].
Due to the structural similarity between the two expressions in Eqs. 4 and 5 it is convenient to rewrite the effective PFK reaction rate in a unified form. The resulting reaction-diffusion equations read
where the parameter δ ≡ D ATP / D ADP denotes the ratio between the effective diffusion coefficients of inhibitor and activator. Length scales are measured in units of the activator diffusion length given by ( D ADP / k d ) 1/2 = ( D ATP / δk d ) 1/2 .
Eqs. 6 and 7 will be analyzed near a supercritical Hopf bifurcation where the dynamics is well de-scribed by the CGLE in Eq. 2. We are particularly interested in the type and the stability of the emerging patterns as we change from a sequential activation mechanism ( L M = L > 1, ε M = 1) to a Hill-like activation mechanism ( L S ≡ 1, ε S = 0). Note that the Hill mechanism leads to a factor γ n in Eq. 7 while the sequential mechanism produces a factor (1 + γ ) n . The latter results from the (binomial) summation over the intermediate enzyme states R l0 , ..., R ln ( l = 0, 1) (see Methods ).
Transition from the MWC to the Sel'kov Model
Given the structural similarity between the Sel'kov and the MWC model it will be beneficial to investigate the relation between the two models in more detail. In particular, we expect that the MWC mechanism reduces to that of the Sel'kov model as the affinity for subsequent product binding steps increases (i.e. K P decreases) such that the product activation becomes more and more cooperative.
To show this explicitly we note that under the rescaling K P → εK P with 0 < ε < 1 the normalized activator concentration γ = ADP/K P changes as γ → γ / ε and φ M becomes φ M ( α,γ/ε ): = φ ε ( α , γ ) with
Hence, φ ε interpolates between φ M and φ S since φ 1 ≡ φ M and as ε → 0 (the binding a finity in-creases) φ ε approaches φ S provided that the product L M ε n converges to L S = 1. However, this means that in the MWC model the enzyme cooperativity, as measured by the allosteric constant L M , has to become increasingly large which is in agreement with the idea that the product activation becomes more cooperative as we change from the MWC to the Sel'kov mechanism.
Formally, we can describe this transition by
This relation between the MWC and the Sel'kov model will be helpful when we analyze how the type and the stability of the spatio-temporal patterns changes as we change the PFK activation mechanism from the MWC to the Hill type. | Methods
Derivation of the rate law for the PFK in the MWC model
In general, if all enzyme binding reactions are in quasi-steady state the effective reaction rate can be written as [ 65 ] v = ke 0 φ . Here k is the intrinsic substrate conversion rate of a single enzyme subunit, e 0 denotes the total enzyme concentration and
is the fractional saturation function which measures the number of occupied substrate binding sites relative to the total number of substrate binding sites. In quasi-steady state the active enzyme states R lm can expressed in terms of binding constants and substrate/product concentrations. For example, for the case n = 2 shown in Figure 1B we have R 01 = 2 R 00 ADP/K S and R 02 = R 01 ADP/ 2 K S = R 00 ( ADP/K S ) 2 . Hence, the summation over the intermediate enzyme states in produces a binomial series. Similarly, we obtain such that φ as defined in Eq. 12 reproduces Eq. 5 by taking into account that L = T 00 / R 00 .
Calculation of the Hopf and the Turing Instability
We begin by rewriting Eqs. 6 and 7 in the form
where the functions f α and f γ are given by
The function φ i is defined as
where the Sel'kov model is characterized by ε S = 0 and L S = 1 while the Monod-Wyman-Changeux model is obtained for ε M = 1 and L M = L > 1. In Eqs. 13 the vector p = (ν, σ , q , n , L i , ε i ) collectively denotes the kinetic parameters appearing in the functions f α and fγ .
Steady States
The unique (spatially homogeneous) steady state of Eqs. 13 is given by
For α s > 0 to be positive we require that σ > ν .
Turing and Hopf bifurcation thresholds
The stability of the fixed point ( α s , γ s ) against spatio-temporal perturbations of the form δ α ( x , t ) = α 0 exp ( ikx + λ t ) and δγ ( x , t ) = γ 0 exp ( ikx + λ t ) is determined by the characteristic polynomial
where λ ( k ) characterizes the temporal evolution of a spatial growth mode with wave vector k . The co-efficient functions a 0 and a 1 in Eq. 15 are given by
Where
Instabilities occur if there exists a k c for which Re (λ( k c )) > 0. The type of instability depends on whether this occurs for k c = 0 (Hopf bifurcation) or for k c ≠ 0 (corresponding to a Turing bifurcation if, in addition, Im (λ( k c )) = 0).
Turing bifurcation
The critical wave number k c of the most unstable mode in the Turing bifurcation is determined by da 0 / dk = 0, and the corresponding parameter set is implicitly given by a 0 ( k c ) = 0.
Hence,
and the parameter set for the Turing bifurcation is described by
Hopf bifurcation
The Hopf bifurcation is determined by a 1 ( k = 0) = 0 or 1 + A - qB = 0 which is a quadratic equation for σ . The respective solution is given by
With
Turing-Hopf codimension-2 bifurcation
In general, oscillations are observed for σ ≤ σ H if σ H < σ T while Turing patterns emerge for σ ≤ σ T if σ T < σ H and σ T is the smallest (real) root of Eq. 17. However, if both of these codimension-1 bifurcations occur simultaneously ( σ H = σ T ) a Turing-Hopf codimension two bifurcation takes place. An implicit expression for this bifurcation curve is obtained by using the explicit representation for σ H (Eq. 18) in the expression for T ( δ , p ) = 0 (Eq. 17). Note that near the Turing-Hopf bifurcation curve it can be difficult to predict whether wave or Turing patterns are observed since both can be simultaneously stable. Alternatively, mixed mode patterns can appear near a Turing-Hopf bifurcation [ 66 ].
Calculation of the CGLE coefficients c 1 and c 3
Near the supercritical Hopf bifurcation the spatio-temporal dynamics of the reaction-diffusion system in Eqs. 13 is well described by the complex Ginzburg Landau equation (CGLE) [ 41 ]
Here, A ( x , t ) is a complex amplitude describing slow spatio-temporal modulations around the (spatially homogeneous) unstable steady state of Eqs. 13 while c 1 and c 3 are real coefficients. In general, c 3 = c 3 ( p ) only depends on the reaction mechanism through the kinetic parameters while c 1 = c 1 ( p , δ ) additionally depends on the ratio of the diffusion coefficients δ = D α / D γ .
To determine the borderline between inward and outward propagating waves, given by c 1 - c 3 = 0, we will calculate the two CGLE coefficients c 1 and c 3 as a function of the original system parameters following the approach in Ref. [ 41 ].
Calculation of c 1
The first CGLE coefficient c 1 is given by c 1 = d 2 / d 1 where . Here, i is the imaginary unit, D = diag ( δ , 1) denotes the diagonal diffusion matrix and are the left(right) eigenvectors of the Jacobian matrix
where f α , α ≡ ∂ f α /∂ α , etc. denote the respective partial derivate. Here and in the following all expressions have to be evaluated at the Hopf bifurcation by eliminating from the expression for A using Eqs. 16 and 18. The eigenvectors
are normalized as and their respective eigenvalues are given by and . Hence, the oscillation frequency at the Hopf bifurcaption is given by . For c 1 we find
Calculation of c 3
The calculation of c 3 is more tedious, but straight-forward [ 41 ]. The expression for c 3 is given by c 3 = g 2 / g 1 where
The 'overbar' in denotes complex conjugation and transposition such that is a column vector. In Eq. 20 we have defined the tensor-valued vectors M 0 = ( M 1 , M 2 ) and N 0 = ( N 1 , N 2 ). The components of M 0 are given by the second-rank tensor
while each component of N 0 is a third-rank tensor given by
Here we have set (α, γ) ≡ ( x 1 , x 2 ) and i, j, k = 1, 2. Hence, in component form Eq. 20 reads
where we have used the Einstein summation convention according to which over indices appearing twice in a term has to be summed automatically - from 1 to 2 in our case. The vectors V 0 and V + are given by
where Id ≡ diag (1, 1) denotes the 2 × 2 identity matrix. The matrix L 0 and the eigenvalue λ 0 have been defined in Section.
Explicit expression for c 1 and c 3 in the limit of low glycolytic flux
The calculation of c 3 has been automatized using a computer algebra system which was also used to plot the graphs in Figure 2 and 4 . In contrast to c 1 (Eq. 19) the output for c 3 is too clumsy to be displayed in a comprehensive manner. How-ever, in the case of low glycolytic flux where Eqs. 13 can be approximated as (cf. Eq. 10)
the expression for c 3 considerably simplifies to
where the polynomial functions P i are given by
The expression for c 1 is still given by Eq. 19 with the Hopf frequency
Note, that for the Sel'kov case, where ε i ≡ ε S = 0, the coefficients c 1 and c 3 become independent of the parameter combination νq corresponding to the steady state value of the activator concentration.
Numerical Simulations
The results of numerical simulations shown in Figure 3 were generated by discretizing Eqs. 6 and 7 in space and time where the Laplacian was approximated by finite differences (five-point-scheme). The resulting set of ordinary differential equations was integrated using a 4th order Runge-Kutta scheme with no-flux boundary conditions. The respective parameters are summarized in Table 1 . The Turing patterns in Figure 3C and 3F were generated from random initial conditions chosen as
where r is a random number equally distributed in the interval [0,1] and x , y = 1, ..., N . A similar expression was used for γ .
To generate a (anti-)spiral wave as in Figure 3A, D and 3E one has note that the spiral core corresponds to a phase singularity [ 41 ]. If not created initially such a phase defect will not develop spontaneously in a spatially homogeneous medium. To simulate the spiral waves we have, thus, created a phase defect by initially imposing a spatial gradient of the inhibitor ( α ) along the x -direction and a second spatial gradient of the activator ( γ ) along the y -direction as
To generate Figure 3A and 3B we have chosen k = 1/70 while Figure 3D and 3E were generated with k = 2/3.
We remark that the physical side length of the simulation domain is given by λ D · dx · N where
denotes the diffusion length of the activator, dx is the spatial step size of the discretization and N = 176 is the number of grid points which we kept fixed for all simulations. As a result, the physical dimensions of each panel in Figure 3 are different since the simulations were done for different values of the spatial scale separation δ = D ATP / D ADP and dx (cf. Table 1 ). Once specific values for the ATP diffusion co-efficient D ATP and the product consumption rate k d are provided the physical side length is determined by ( D ATP / k d ) 1/2 l s where l s = dx · N/δ 1/2 denotes the dimensionless side length. | Results
The diffusive coupling of locally oscillatory reactions as in Eqs. 6 and 7 can generate different types of reaction-diffusion wave patterns which can be broadly classified into outward and inward propagating waves [ 45 ]. Near a supercritical Hopf bifurcation the transition between these wave types occurs for c 1 - c 3 = 0 (Eq. 2). Depending on the initial and/or boundary conditions these waves may appear in the form of circular or spiral shaped waves.
More complex dynamic behavior can occur near a Benjamin-Feir instability which is indicated by the condition 1 + c 1 c 3 < 0 [ 42 ]. In this bifurcation plane wave solutions become unstable against long wave length perturbations which may result in the occurrence of spatio-temporal chaos.
Finally, when the spatial scale separation δ becomes sufficiently large the oscillatory instability may be suppressed and stationary Turing patterns can emerge. The transition between wave dynamics and stationary patterns is indicated by the codimension-two Turing-Hopf bifurcation.
To compare the spatio-temporal dynamics of the Sel'kov and the MWC model we have calculated the two CGLE coefficients ( c 1 and c 3 ) and the Turing-Hopf curve for Eqs. 6 and 7 as a function of the systems parameters L i , ε i , ν , q , n and δ (see Methods). Note that σ has been eliminated by the requirement for the system to be near the Hopf bifurcation.
Sequential vs. Hill-like Activation Mechanism
To decouple the two limiting prescriptions in Eq. 9 we study Eqs. 6 and 7 first in the regime of low glycolytic flux as it is typically observed during oscillatory behavior [ 51 ]. This regime is characterized by the conditions α s = O (1) and for the Sel'kov and (1 + γ s ) n ≪ L for the MWC model. Here, α s and γ s are the respective inhibitor and activator concentrations of the (unstable) steady state. In this approximation, the function φ i in Eq. 7 becomes [ 12 ]
The parameter L i can be absorbed into the definition of the new parameter combination in Eqs. 6 such that we can simply change from the MWC to the Hill mechanism by decreasing ε from ε M = 1 to ε S = 0 (Figure 2 ).
As the affinity for the sequential binding of product molecules increases ( ε decreases) the stability region of inward propagating waves (dark shaded area) decreases (Figure 2A and 2B ). At ε = 0.1 the transition curve between outward and inward propagating waves ( c 1 - c 3 = 0, dashed line) has crossed the Turing-Hopf bifurcation line (Figure 2C ). Thus, for ε ≤ 0.1 the transition to inward propagating waves occurs in the non-oscillatory regime where wave behavior (shaded area) is suppressed in favor of stationary Turing patterns. This shows that the inward propagating waves, as predicted by the CGLE, are not necessarily observable at the level of the original reaction-diffusion system (Eqs. 6 and 10). Since the binding of subsequent product molecules becomes more cooperative as ε → 0 the occurrence of inward propagating waves in the MWC model seems to be related to the sequential activation of the PFK (Figure 1B ) which exhibits less cooperativity because the binding a finity (1/ εK p ) is finite for ε = 1.
Wave Stability and Numerical Simulations
For the MWC mechanism outward propagating waves are stable even if the activator diffuses faster than the inhibitor ( δ < 1). However, as ε decreases these waves become unstable as indicated by the appearance of a Benjamin-Feir (BF) instability (Figure 2B, 2C and 2D , dash-dotted line). The BF in-stability (1 + c 1 c 3 < 0) marks the region in parameter space where plane waves with a wave number k = 0 become unstable to long wave length perturbations. Hence, in the region to the left of the BF curve no stable wave patterns are observable. How-ever, since even for 1 + c 1 c 3 > 0 waves with a finite wave number k ≠ 0 also can become unstable [ 41 ] there is typically a whole region of unstable wave behavior (extending to the right of the BF curve) where spatio-temporal chaos emerges (Figure 3A ). Thus, increasing the cooperativity for the product activation steps ( ε → 0) leads to a destabilization of the coherent wave behavior for δ ≤ 1.
Away from the BF instability curve ( δ > 1) outward propagating waves become stable even in the limit ε = 0 (Figure 3B ). As δ increases further these waves turn into stationary Turing patterns (Figure 3C ). In contrast, for the sequential product activation mechanism ( ε = 1) there is no Benjamin-Feir instability for typical numbers of PFK subunits n ≤ 8 [ 52 ] and sufficiently low values of the steady state activator concentration γ s = νq (cf. Figure 4E ). Hence, there is no spatio-temporal chaos in that regime. Instead, outward propagating waves are stable at low values of δ (Figure 2A and Figure 3D ). As δ increases the direction of wave propagation changes from outward to inward such that anti-spirals become observable (Figure 3E ). At sufficiently large values of δ wave behavior is suppressed by a Turing instability (Figure 3F ) similar as for the Sel'kov model.
Inward Propagating Waves and Enzyme Cooperatively
The fact that the occurrence of inward propagating waves depends on the binding a finity for sub-sequent product activation steps suggests that the strength of enzyme cooperativity might play a role in this respect. The amount of cooperativity can be conveniently quantified by an effective Hill coefficient which is defined (with respect to the activator concentration) as [ 40 ] n H = ( γ/M )( dM/dγ ) where M ≡ φ /(1 - φ ). For φ = φ i (Eq. 7) this quantity is explicitly given by
Positive (negative) cooperativity corresponds to values n H > 1 ( n H < 1) while n H = 1 indicates no cooperativity.
To perform the transition from the MWC to the Hill mechanism when L is not necessarily large we introduce in Eq. 8 an effective allosteric constant as L eff ≡ L M ε n = ε ( L - 1) + 1. This definition ensures that L eff has the correct limiting behavior as required by Eq. 9, i.e. L eff = L M = L for ε = 1 and L eff → L S = 1 as ε → 0. Note that the true allosteric constant L M increases as 1/ ε n when ε → 0.
In Figure 4 we relate the occurrence of inward propagating waves to the properties of the effective Hill coefficient n H as described by Eq. 11. For the MWC model n H exhibits a single maximum as a function of the steady state activator concentration γ s = νq (Figure 4A ). While oscillatory behavior (grey shaded area) is observed on both sides of the maximum (dotted line) of the Hill curve inward propagating waves only occur to the left of it where dn H /dγ s > 0 (Figure 4E ). This behavior is independent of the particular choice of the other parameters L and q (Figure 5 ). As ε decreases the cooperativity of product activation increases as indicated by an in-crease in the maximum of the Hill curve (Figure 4B and 4C ). Concomitantly, the stability region of anti-spiral waves rapidly shrinks and subsequently shifts to small activator concentrations where the steep-ness of the Hill curve is sufficiently large (Figure 4F and 4G ). This suggests that it is not the strength of the cooperativity per se, but the sensitivity of the Hill coefficient with respect to changes in the activator concentration which determines whether inward propagating waves can occur or not. This is in agreement with the fact that for the Sel'kov model, where is a monotonically decreasing function ( dn H /dγ s ≤ 0), the formation of inward propagating waves is suppressed (Figure 4D and 4H ). | Discussion
Beginning in the 1960s glycolytic oscillations have become one of the best studied biochemical oscillators both in cell-free extracts [ 26 , 38 , 53 ] and in living cells [ 28 , 54 ]. Later it was found in studies with yeast cell populations that glycolytic oscillations represent a collective phenomenon. The oscillations in individual cells are synchronized through the exchange of metabolic intermediates such as acetaldehyde [ 55 ] or glucose [ 56 ]. At low cell densities the oscillations at the population level disappear (synchronously in all cells) indicating a quorum sensing mechanism [ 57 ]. Synchronized behavior was also observed in cell-free extracts where diffusive coupling of glycolytic enzymes can generate waves of glycolytic activity [ 30 , 31 ]. However, clear experimental evidence for metabolic waves in living cells remains scarce [ 58 ] although mathematical modeling supports the feasability of such waves [ 59 ].
Recently, a novel type of wave dynamics, called inward rotating spiral waves, has been observed in cell-free yeast extracts [ 32 ]. Such wave behavior has, so far, only been observed in purely chemical systems [ 35 , 36 ]. Here, we have investigated the molecular mechanism underlying the generation of such anti-spiral waves in simple glycolytic model systems which focus on the allosteric activation of the glycolytic enzyme phosphofructokinase (PFK). We have shown that in the Goldbeter model inward rotating spiral waves can arise due to the sequential activation of the PFK implied in the Monod-Wyman-Changeux mechanism [ 13 , 37 ]. In the limit of an in finitely large binding a finity where the PFK activation is described by a Hill function, as in Sel'kov model [ 12 ], the capability to generate inward propagating waves is lost. This suggests that the MWC mechanism, as in Figure 1B , can not be further implied. On the other hand, as we have shown earlier [ 32 ] the capability to generate anti-waves is retained by the Goldbeter model where the cooperativity with respect to substrate binding and the allosteric inhibition by ATP are additionally taken into account. Hence, the MWC model can be regarded as a 'core' mechanism for the generation of inward propagating waves for allosteric enzyme systems with product activation.
For well-mixed reaction systems a simple Hill function is often employed to model cooperative behavior in a 'generic' way. Near the onset of oscillations choosing a Hill kinetics instead of a more complex activating function, as in the MWC model, does not lead to a qualitative change in the dynamics under well-stirred conditions. However, as we have shown, the choice of the activating function can significantly change the type and the stability of dynamic patterns in the presence of diffusive coupling. For example, the appearance of a Benjamin-Feir instability in the Sel'kov model indicates the occurrence of spatio-temporal chaos which is mostly absent in the MWC model (Figure 2 , 3 and 4 ). This suggests that the intermediate enzyme forms in the MWC model, which are only partially saturated with product molecules, can stabilize the system dynamics against long wave length perturbations.
Sel'kov and Goldbeter have shown that for the PFK mediated reaction to become oscillatory a sufficiently strong positive enzyme cooperativity is required [ 12 , 60 ]. However, as far as oscillations are concerned the detailed shape of the Hill coefficient curve (Eq. 11) is not important. Consequently, they occur on the ascending branch of the cooperativity curve (where dn H /dγ s > 0) as well as on the descending branch (where dn H / dγ s < 0) as long as n H > 1 (Figure 4A, B, C and 4D ). Interestingly, the occurrence of inward propagating waves does not seem to depend on the magnitude of the enzyme cooperativity, but on its sensitivity with respect to changes in the activator concentration. Our simulations show that the formation of inward propagating waves correlates with a positive sensitivity ( dn H / dγ s > 0) which indicates that for the pattern forming aspects of allosteric enzyme systems more subtle enzyme properties play a role than they do for the occurrence of oscillations.
Since the glycolytic model systems in Eqs. 6 and 7 are of the substrate-depletion type [ 14 ] it is not surprising that both models predict the occurrence of stationary Turing patterns if the spatial scale separation between inhibitor and activator dynamics becomes sufficiently large [ 49 ]. What is surprising is the fact that this transition already occurs for comparably small values of δ = 2, ..., 4 if the number n of enzyme subunits is sufficiently large (Figure 2A and 2D ). This strong dependence on the enzyme cooperativity has been largely neglected in earlier work [ 61 , 62 ] which mostly focused on the case n = 2 (corresponding to muscle PFK). However, in yeast the PFK is an octamer ( n = 8) for which Turing pat-terns are predicted to occur for δ > 4 in the MWC model and for δ > 2 in the Sel'kov model. The necessary spatial scale separation could be generated, for example, through preferential allosteric interactions of the PFK effectors with immobilized enzymes [ 32 , 63 ], including the PFK itself. Although Turing patterns can be systematically generated only in chemical systems yet [ 25 ] our results suggest that high oligomeric enzyme systems are promising candidates to generate such patterns also in properly designed biochemical reaction-diffusion systems. | Conclusions
In well-mixed reaction systems the systematic investigation of molecular reaction mechanisms has led to considerable insights into the design principles for the generation of a specific type of dynamic behavior such as bistability or oscillations [ 14 , 15 , 64 ]. Here, we have expanded this approach to the case of spatially extended systems. Specifically, we have demonstrated that amplitude equations are a valuable tool to investigate how the occurrence of particular spatio-temporal patterns depends on the details of the underlying molecular reaction mechanism in the presence of diffusive coupling. In that way we could provide a molecular explanation for the occurrence of inward rotating spiral waves as they were recently observed in glycolysis in cell-free yeast extracts. Our results support the view that in yeast the allosteric enzyme phosphofructokinase is activated by a Monod-Wyman-Changeux and not by a Hill mechanism. They also highlight the importance of the number of enzyme subunits for a possible experimental generation of Turing patterns in biological systems. | Background
A central question for the understanding of biological reaction networks is how a particular dynamic behavior, such as bistability or oscillations, is realized at the molecular level. So far this question has been mainly addressed in well-mixed reaction systems which are conveniently described by ordinary differential equations. However, much less is known about how molecular details of a reaction mechanism can affect the dynamics in diffusively coupled systems because the resulting partial differential equations are much more difficult to analyze.
Results
Motivated by recent experiments we compare two closely related mechanisms for the product activation of allosteric enzymes with respect to their ability to induce different types of reaction-diffusion waves and stationary Turing patterns. The analysis is facilitated by mapping each model to an associated complex Ginzburg-Landau equation. We show that a sequential activation mechanism, as implemented in the model of Monod, Wyman and Changeux (MWC), can generate inward rotating spiral waves which were recently observed as glycolytic activity waves in yeast extracts. In contrast, in the limiting case of a simple Hill activation, the formation of inward propagating waves is suppressed by a Turing instability. The occurrence of this unusual wave dynamics is not related to the magnitude of the enzyme cooperativity (as it is true for the occurrence of oscillations), but to the sensitivity with respect to changes of the activator concentration. Also, the MWC mechanism generates wave patterns that are more stable against long wave length perturbations.
Conclusions
This analysis demonstrates that amplitude equations, which describe the spatio-temporal dynamics near an instability, represent a valuable tool to investigate the molecular effects of reaction mechanisms on pattern formation in spatially extended systems. Using this approach we have shown that the occurrence of inward rotating spiral waves in glycolysis can be explained in terms of an MWC, but not with a Hill mechanism for the activation of the allosteric enzyme phosphofructokinase. Our results also highlight the importance of enzyme oligomerization for a possible experimental generation of Turing patterns in biological systems. | Authors' contributions
RS conceived the study, performed the analytical calculations, the numerical simulations and drafted the manuscript. EMN helped to interpret the results and contributed to the manuscript. All authors read and approved the final manuscript. | Acknowledgements
RS acknowledges financial support from the Ministry of Education of Saxony-Anhalt within the Research Center 'Dynamic Systems'. | CC BY | no | 2022-01-12 15:21:37 | BMC Syst Biol. 2010 Nov 30; 4:165 | oa_package/29/21/PMC3014969.tar.gz |
PMC3014970 | 21122111 | Background
Escherichia coli has not only been the prime organism for developing new molecular biology methods but also for producing recombinant proteins, low molecular weight compounds etc. in industrial biotechnology for decades due to its low cost manufacturing and end-product purification and its ability to reach high cell densities grown aerobically [ 1 , 2 ]. However, a major problem exists with aerobic E. coli cultivation on glucose at high growth rates-formation and accumulation of considerable amounts of acetic acid i.e . overflow metabolism. In addition to being detrimental for target product synthesis, accumulated acetate inhibits growth and diverts valuable carbon from biomass formation [ 3 , 4 ].
The acetate synthesis and utilization pathways [ 5 ] can be seen in Figure 1 : acetate can be synthesized by phosphotransacetylase (PTA)/acetate kinase (ACKA) and by pyruvate oxidase (POXB). Acetic acid can be metabolized to acetyl-CoA either by the PTA-ACKA pathway or by acetyl-CoA synthetase (ACS) through an intermediate acetyl-AMP. The high affinity ( K m of 200 μM for acetic acid) ACS scavenges acetate at low concentrations whereas the low affinity PTA-ACKA pathway ( K m of 7-10 mM) is activated in the presence of high acetate concentrations [ 6 ].
The phenomenon of overflow metabolism has been studied widely over the years and it is commonly believed to be caused by an imbalance between the fluxes of glucose uptake and those for energy production and biosynthesis [ 7 , 8 ]. Several explanations such as the saturation of catalytic activities in the tricarboxylic acid (TCA) cycle [ 9 , 10 ] and respiratory chain [ 7 , 11 , 12 ], energy generation [ 5 , 13 ] or the necessity for coenzyme A replenishment [ 14 ] have been proposed. In addition to bioprocess level approaches [ 1 , 15 ], various genetic modifications of the acetate synthesis pathways extensively reviewed in De Mey et al . [ 15 ] have been made to minimize acetic acid production. For instance, it has been shown that deleting the main acetate synthesis route PTA-ACKA results in a strong reduction (up to 80%) of acetate excretion, maximum growth rate ( ca 20%) and elevated levels of formate and lactate ( ca 30-fold) [ 4 , 16 - 18 ], whereas poxB disruption causes reduction in biomass yield ( ca 25%) and loss of aerobic growth efficiency of E. coli [ 19 ]. The latter indicates that acetate excretion cannot be simply excluded by disrupting its synthesis routes without encountering other unwanted effects. Unfortunately, no clear conclusions could be drawn from batch experiments with an acs knock-out strain [ 4 ]. It should be noted that studies with E. coli genetically modified strains engineered to diminish acetate production in batch cultures have not fully succeeded in avoiding acetate accumulation together with increasing target product production yields and rates [ 15 ]. Additionally, these studies have not allowed elucidating the mechanism of overflow metabolism unequivocally [ 4 , 20 , 21 ].
Acetate overflow is a growth rate dependent phenomenon, but no study has specifically focused on growth rate dependency of protein and gene expression regulation, intra-and extracellular metabolite levels using also metabolic modeling. Describing the physiology of an organism on several 'omic levels is the basis of systems biology that facilitates better understanding of metabolic regulation [ 22 ]. In this study, E. coli metabolism at proteomic, transcriptomic and metabolomic levels was investigated using continuous cultivation methods prior to and after overflow metabolism was switched on. Usually, chemostat cultures are used for steady state metabolism analysis, however, we applied two changestat cultivation techniques: accelerostat (A-stat) and dilution rate stat (D-stat), see Methods section for details [ 23 , 24 ]. These cultivation methods were used as they provide three advantages over chemostat. Firstly, these changestat cultivation techniques precisely detect metabolically relevant switch points ( e.g . start of overflow metabolism, maximum specific growth rate) and enable to monitor the dynamic patterns of several metabolic physiological responses simultaneously which could be left unnoticed using chemostat. Secondly, it is possible to collect vast amount of steady state comparable samples and by doing so, save time. Thirdly, both A-stat and D-stat enable to quantitatively study specific growth rate dependent co-utilization of growth substrates. Latter advantage was applied for investigating acetic acid consumption capability of E. coli at various dilution rates in this study. Combining changestat cultivation methods enables to study metabolism responses of the same genotype at different physiological states in detail without encountering the possible metabolic artifacts accompanied when using genetically modified strains.
Results obtained by studying specific growth rate dependent changes in E. coli proteome, transcriptome and metabolome in continuous cultures together with metabolic modeling allowed us to propose a new theory for acetate overflow: acetate excretion in E. coli is triggered by carbon catabolite repression mediated down-regulation of Acs resulting in decreased assimilation of acetate produced by Pta, and disruption of the PTA-ACS node. | Methods
Bacterial strain, medium and continuous cultivation conditions
The E. coli K12 MG1655 (λ - F - rph-1Fnr + ; Deutsche Sammlung von Mikroorganismen und Zellkulturen, Germany) strain was used in all experiments. Growth and physiological characteristics in accelerostat (A-stat) cultivations were determined using a defined minimal medium as described before by Nahku et al . [ 51 ], except 4.5 g/L α-(D)-glucose and 100 μl L -1 Antifoam C (Sigma Aldrich, St. Louis, LO) was used. The latter was also used in dilution rate stat (D-stat) experiments as the main cultivation medium. In addition, a second medium was used in D-stat where the main medium was supplemented by acetic acid and prepared as follows: 300 ml medium was withdrawn from the main cultivation medium and supplemented with 3 ml of glacial acetic acid (99.9%). One A-stat experiment (referred to as two-substrate A-stat) was carried out with the same medium as other A-stats, but in addition supplemented with acetic acid (final concentration 5 mM).
The continuous (both A-stat and D-stat) cultivation system consisted of 1.25 L Biobundle bioreactor (Applikon Biotechnology B.V., Schiedam, the Netherlands) controlled by an ADI 1030 biocontroller (Applikon Biotechnology B.V.) and a cultivation control program "BioXpert NT" (Applikon Biotechnology B.V.). The system was equipped with OD, pH, pO 2 , CO 2 and temperature sensors. The bioreactor was set on a balance whose output was used as the control variable to ensure constant culture volume (300 ± 1 mL). Similarly, the inflow was controlled through measuring the mass of the fresh culture medium.
A-stat cultivation system and control algorithms used are described in more detail in our previous works [ 24 , 51 , 56 ]. Dilution rate stat (D-stat) is a continuous cultivation method where dilution rate is constant as in a chemostat while an environmental parameter is smoothly changed [ 24 ]. The D-stat experiments in this study were carried out with a slight modification: instead of changing an environmental parameter, two different media were used to keep dilution rate constant. After achieving steady state conditions in chemostat using minimal medium supplemented with glucose, addition of the second medium complemented with glucose and acetic acid was started. The feeding rate of the initial medium was decreased at the same time, resulting in constant glucose concentration in the feed. The acetic acid concentration in the bioreactor as a result of inflow has to be determined to enable precise acetic acid consumption/production rate calculation for the bacteria. Hence, increase of acetic acid concentration in bioreactor was calculated and validated in duplicate non-inoculated D-stat test experiments producing an average standard deviation of 1.24 mM between calculated and measured acetic acid concentrations.
All continuous cultivation experiments were carried out at 37°C, pH 7 and under aerobic conditions (air flow rate 150 ml min -1 ) with an agitation speed of 800 rpm. Four A-stat cultivations were performed with acceleration rate (a) 0.01 h -2 . Duplicate D-stat experiments were performed at dilution rates 0.10; 0.30; 0.505 ± 0.005 h -1 and single experiments at 0.19; 0.24; 0.40; 0.45 h -1 . The acetic acid addition profile was set to achieve 32 ± 6 mM and 58 ± 5 mM in 7 hours inside the bioreactor for experiments at dilution rates 0.10-0.24 h -1 and 0.30-0.51 h -1 , respectively. The growth characteristics of the bacteria were calculated on the basis of total volume of medium pumped out from bioreactor (L), biomass (g DCW), organic acid concentrations in culture medium (mM) and CO 2 concentration in the outflow gas (mM). Formulas were as described in a previous study [ 24 ]. It should be noted that the absolute CO 2 concentrations could be error-prone due to measurement difficulties. However, this does not influence the dynamic pattern of specific CO 2 production rate ( r CO2 ) during specific growth rate increase.
Analytical methods
The concentrations of organic acids (lactate, acetate and formate), ethanol and glucose in the culture medium were determined by HPLC and cellular dry weight (expressed as DCW) as described by Nahku et al . [ 51 ].
Protein expression analysis
Refer to Text S1 in Additional file 1 for detailed description. Shortly, protein expression ratios for around 1600 proteins (identified for each growth rate at a > 95% confidence interval in average from 89,303 distinct 2 or more high-confidence peptides) were generated from mass spectrometric spectra by firstly calculating the ratios between continuous cultivation samples at specific growth rates 0.10 ± 0.01 h -1 (chemostat point prior to the start of acceleration in A-stat); 0.20 ± 0.01; 0.26; 0.30 ± 0.01; 0.40 ± 0.00; 0.49 ± 0.01 h -1 and batch sample grown on medium containing 15 NH 4 Cl as the only source of ammonia. Secondly, the ratios between the mentioned specific growth rates with chemostat point (μ = 0.10 ± 0.01 h -1 ) for two biological replicates were calculated to yield protein expression levels for respective specific growth rates. Protein (and gene) expression measurement results are shown in Additional file 2 . Proteomic analysis data is also available at the PRIDE database [ 57 ] http://www.ebi.ac.uk/pride under accession numbers 12189-12199 (username: review74613, password: Ge9T48e8). The data was converted using PRIDE Converter http://code.google.com/p/pride-converter [ 58 ].
Gene expression profiling
DNA microarray analysis of 4,321 transcripts was conducted with the Agilent platform using the data from one A-stat cultivation (a = 0.01 h -2 ), and gene expression ratios between specific growth rates 0.21; 0.26; 0.31; 0.36; 0.40; 0.48 h -1 and μ = 0.11 h -1 were calculated. Transcript spot intensities of chemostat sample (sample from D-stat prior to acetic acid addition) from μ = 0.51 h -1 and A-stat μ = 0.48 h -1 were used for the two method's comparison at transcriptome level. Gene (and protein) expression measurement results are shown in Additional file 2 . DNA microarray data is also available at NCBI Gene Expression Omnibus (Reference series: GSE23920). The details of the procedure are provided in Text S1 in Additional file 1 .
Metabolome analysis
Sampling was carried out by the rapid centrifugation method. Acquity UPLC (Waters, Milford, MA) together with end-capped HSS C18 T3 1.8 μm, 2.1 × 100 mm column for compound separation coupled to TOF-MS with an electrospray ionization (ESI) source was used for detection (LCT Premiere, Waters). The details of the procedure are provided in Text S1 in Additional file 1 . | Results
E. coli metabolic switch points characterization
In all accelerostat (A-stat) cultivation experiments, after the culture had been stabilized in chemostat at 0.10 h -1 to achieve steady state conditions, continuous increase in dilution rate with acceleration rate (a) 0.01 h -2 (0.01 h -1 per hour) was started. Continuous change of specific growth rate resulted in detecting several important changes in E. coli metabolism as demonstrated in Figure 2 . Firstly, in A-stat cultivations where glucose was the only carbon source in the medium, acetic acid started to accumulate ( i.e . overflow metabolism switch) at μ = 0.27 ± 0.02 h -1 (average ± standard deviation) and a two-phase acetate accumulation pattern was observed (discussed below; Figure 2 ). Cells reached maximum CO 2 production and O 2 consumption at μ = 0.46 ± 0.02 h -1 and metabolic fluctuations were observed at μ = 0.49 ± 0.03 h -1 followed by washout of culture at μ = 0.54 ± 0.03 h -1 (corresponding to maximum specific growth rate at given conditions). The nature of these fluctuations will be studied further and not covered in the current publication. All A-stat results were reproduced with relative standard deviation less than 10% with the exception of acetate production per biomass (Y OAc- ) (Table 1 and Figure S1 in Additional file 1 ).
Metabolomic responses to rising specific growth rate
A-stat cultivation enabled to study acetic acid accumulation profile in detail with increasing specific growth rate. Interestingly, a two-phase acetate accumulation pattern was observed (Figure 2 ). Slow accumulation of acetic acid started at μ = 0.27 ± 0.02 h -1 with concomitant change in specific CO 2 production rate (Figure 2 ). Faster accumulation of acetate was witnessed after cells had reached maximum CO 2 production at μ = 0.46 ± 0.02 h -1 . Quite surprisingly, production of the important carbon catabolite repression (CCR) signal molecule cAMP (Y cAMP ) rose from steady state chemostat level 2.45 ± 0.26 μmol/g dry cellular weight (DCW) (μ = 0.10 h -1 ) to 3.55 ± 0.32 μmol/g DCW (μ = 0.30 h -1 ) after which it sharply decreased to 1.30 ± 0.44 μmol/g DCW at μ = 0.45 h -1 (Figure S1 in Additional file 1 ). This abrupt decline took place simultaneously with the faster acetate accumulation profile described above (Figure 2 and Figure S1 in Additional file 1 ). In addition, similar two-phase acetate accumulation phenomenon was observed in a two-substrate (glucose + acetic acid) A-stat during the decrease of cAMP around specific growth rate 0.39 h -1 (Figure S2 in Additional file 1 ).
Significant fall in two of the measured pentose phosphate pathway intermediates ribose-5-phosphate (R5P) and erythrose-4-phosphate (E4P) was detected with increasing specific growth rate which could point to possible limitation in RNA biosynthesis during growth (Figure 3A ). PTA-ACS node related compound nonesterified acetyl-CoA (HS-CoA) level declined two-fold simultaneously with cAMP after acetate started to accumulate (Figure 3B ). This indicates the possible increase of other CoA containing compounds e.g . succinyl-CoA. Accumulation of TCA cycle intermediates α-ketoglutarate and isocitrate (Figure 3B ) with increasing dilution rate could be associated with pyrimidine deficiency and decrease of ATP expenditure in the PTA-ACS cycle. Concurrently, intracellular concentrations of fructose-1,6-bisphosphate (FBP) and glyceraldehyde-3-phosphate (GAP) from the upper part of energy generating glycolysis increased 6- and 3-fold, respectively (Figure 3C ).
Functional-genomic responses to rising specific growth rate
The two main known pathways for acetate synthesis phosphotransacetylase-acetate kinase (PTA-ACKA) and pyruvate oxidase (POXB) were down-regulated, both on gene and protein expression levels, from μ = 0.20 h -1 i.e . before acetate overflow was switched on. At the same time, there was a concurrent 10-fold repression of the acetic acid utilization enzyme acetyl-CoA synthetase (Acs). This substantial difference (5-fold) between the acetate synthesis and assimilation pathways expression suggests that the synthesized acetic acid cannot be fully assimilated with increasing growth rates (Figure 1 ).
We observed the beginning of carbon catabolite repression (CCR) induction prior to acetate accumulation in parallel with Acs down-regulation. This was indicated by down-regulation (3-fold on average) of CCR-mediated components: alternative (to glucose) substrate transport and utilization systems like galactose (MglAB), maltose (MalBEFKM), galactitol (GatABC), L-arabinose (AraF), D-ribose (RbsAB), C 4 -dicarboxylates (DctA) and acetate (ActP, YjcH) (Figure 4C and Additional file 2 ). Moreover, expression of transcription activator Crp (cyclic AMP receptor protein which regulates the expression of Acs transcribing acs-yjcH-actP operon) and Cra (catabolite repressor activator; a global transcriptional protein essential for acetic acid uptake [ 25 ]) were reduced 1.5 and 2 times, respectively, in like manner to carbon catabolite repressed proteins mentioned above (Figure 1 ). Simultaneously, components of the gluconeogenesis pathway (Pck, MaeB, Pps) and glyoxylate shunt enzymes AceA, AceB (vital for acetate consumption) were repressed with growth rate increase (Figure 4B and Additional file 2 ). It should be emphasized that most of the TCA cycle gene and protein levels were maintained or even increased up to μ = 0.40 h -1 followed by sudden repression simultaneous to achieving maximum specific CO 2 production rate (μ = 0.46 ± 0.02 h -1 , see above; Figure 1 Figure 2 and Figure 4A ). This may allude to no limitation at the TCA cycle level around the specific growth rate where overflow metabolism was switched on.
Acetic acid consumption capability studied by dilution rate stat (D-stat) and two-substrate A-stat cultivations
The beginning of a strong decrease in acetate assimilation enzyme Acs expression before overflow switch point implies to a possible connection between acetate assimilation capability and overflow metabolism of acetate (Figure 1 ). Therefore, specific growth rate dependent acetic acid consumption capabilities were investigated using D-stat and two-substrate A-stat methods. It was shown by D-stat experiments at various dilution rates that more than a 12-fold reduction in acetate consumption capability took place around overflow switch point, and its total loss was detected between dilution rates 0.45 and 0.505 ± 0.005 h -1 (Figure 5 ). Acetic acid consumption and production was also studied in a single experiment using two substrate (glucose + acetic acid) A-stat cultivation (Figure S2 in Additional file 1 ) which demonstrated that acetic acid consumption started to decrease at μ = 0.25 h -1 and was completely abolished at μ = 0.48 h -1 which fits into the range of dilution rates observed in D-stat.
A-stat comparison with chemostat
As could be seen from Table 1 major growth characteristics such as biomass yield (Y XS ), acetate (Y OAc- ), cyclic AMP (Y cAMP ) and carbon dioxide (Y CO2 ) production per biomass from A-stat and chemostat are all fully quantitatively comparable. The latter results enable to use A-stat data for quantitative modeling calculations. In addition, the two continuous cultivation methods were examined at transcriptome level using DNA microarrays. Transcript spot intensities from quasi steady state A-stat sample at μ = 0.48 h -1 and chemostat sample at μ = 0.51 h -1 showed an excellent Pearson product-moment correlation coefficient R = 0.964 (Figure S3 in Additional file 1 ; Additional file 3 ). This indicates good biological correlation between E. coli transcript profiles at similar specific growth rates in chemostat and A-stat. These results showed that our quasi steady state data from A-stat and D-stat cultures are steady state representative.
Proteome and transcriptome comparison
E. coli protein expression ratios for around 1600 proteins were generated by comparing two biological replicates at specific growth rates 0.20 ± 0.01; 0.26; 0.30 ± 0.01; 0.40 ± 0.00; 0.49 ± 0.01 h -1 with sample at μ = 0.10 ± 0.01 h -1 (chemostat point prior to the start of acceleration in A-stat) which produced Pearson correlation coefficients for two biological replicates in the indicated pairs of comparison in the range of R = 0.788-0.917 (Figure S4 in Additional file 1 ).
DNA microarray analysis of 4,321 transcripts was conducted with the Agilent platform using the samples from one A-stat cultivation. Gene expression ratios between specific growth rates 0.21; 0.26; 0.31; 0.36; 0.40; 0.48 h -1 and μ = 0.11 h -1 (chemostat point prior to the start of acceleration in A-stat) were calculated. Comparison of gene and protein expression changes (between respective specific growth rates) revealed that components of the PTA-ACS node were regulated at transcriptional level as the absolute majority of the studied transcripts and proteins indicated by the good correlation between transcriptome and proteome expression profiles (Figure 1 and Figure S5 in Additional file 1 ).
Most recent studies have either failed to find a significant correlation between protein and mRNA abundances or have observed only a weak correlation (reviewed in [ 22 ]). It has been suggested that the main reasons for uncoupling of mRNA and protein abundances are protein regulation by post-translational modification, post-transcriptional regulation of protein synthesis, differences in the half-lives of mRNA and proteins, or possible functional requirement for protein binding [ 22 ]. As the cells in these studies were mostly cultured in non steady state condition, our steady state data with very good correlation between transcriptome and proteome implies that the physiological state of the culture (steady state vs. non steady state) could be an important factor in terms of mRNA and protein correlation determination. Transcriptome and proteome data are presented in Additional file 2 and at NCBI Gene Expression Omnibus and PRIDE database (see Methods for details), respectively. | Discussion
To gain better insights into the regulation of acetate overflow metabolism in E. coli , we studied specific growth rate dependent proteomic, transcriptomic and metabolomic patterns combined with metabolic modeling using advanced continuous cultivation methods, which has not been carried out before. Continuous monitoring of the specific growth rate effect on E. coli metabolism enabled us to precisely detect important metabolic shift points, the most important being the start of acetate overflow at μ = 0.27 ± 0.02 h -1 (Figure 2 ), and changing patterns of a number of important metabolites e.g . acetate, cAMP. Quite surprising was the down-regulation of the known acetate synthesis pathways, PTA-ACKA and POXB expression before overflow switch with increasing growth rate (Figure 1 ). A similar pattern has been seen before in chemostat cultures but without emphasizing the possible physiological consequences [ 26 - 28 ]. A 10-fold repression of the acetic acid utilization enzyme acetyl-CoA synthetase (Acs) expression was observed concurrently with the down-regulation of the PTA-ACKA pathway indicating that acetic acid synthesis might exceed its assimilation (Figure 1 ). Our two substrate A-stat and D-stat experiments directly proved that acetate consumption capability of E. coli is specific growth rate dependent as acetate consumption started to decrease at μ = 0.25 h -1 (Figure S2 in Additional file 1 ) and acetate consumption capability decreased 12-fold around overflow switch growth rate μ = 0.27 ± 0.02 h -1 , respectively (Figure 5 ). In addition, it was shown that activation of carbon catabolite repression (CCR) and repression of Acs take place simultaneously prior to the start of overflow metabolism (Figure 1 Figure 4 and Figure 5 ). As a result, it is proposed that acetate overflow metabolism in E. coli is triggered by Acs down-regulation resulting in decreased assimilation of acetic acid produced by Pta, and disruption of the PTA-ACS node.
We showed that Acs was concurrently down-regulated five times more compared to the acetate synthesis pathways (Figure 1 ). In addition, the TCA cycle flux decrease as shown by change in CO 2 production at overflow switch growth rate indicates that carbon is not metabolized by the TCA cycle after the start of acetate accumulation with pre overflow switch rates (Figure 2 and Additional file 4 ). The latter is caused because the amount of acetyl-CoA entering the TCA cycle decreases after carbon is lost into excreted acetate. Stronger repression of the acetate consuming Acs in comparison with acetate synthesizing PTA-ACKA together with a decline in TCA cycle flux suggest disruption of acetic acid cycling at the PTA-ACS node (Figure 1 ). While this node may seem as a futile cycle, the fact is that numerous metabolic tasks involving the intermediate molecules of this cycle-acetyl phosphate (acetyl-P) and acetyl-AMP-are essential for proper E. coli growth (Figure 6 ). For instance, these molecules play a crucial role in bacterial chemotaxis regulation in which flagellar rotation is controlled by the activation level of the response regulator CheY [ 29 ] through either phosphotransfer from CheA [ 30 , 31 ] or acetyl-P [ 31 , 32 ], acetylation by acetyl-AMP [ 33 , 34 ] or co-regulation of both mechanisms [ 29 ]. It has been also demonstrated that acetyl-P synthesis is vital for EnvZ-independent regulation of outer membrane porins [ 35 ], pathogenesis [ 36 ] and regulation of several virulence factors [ 5 ]. Furthermore, it has been presented that acetyl-P interacts with phosphate concentration regulators PhoB-PhoR [ 37 ] and NRI protein which is part of a complex nitrogen sensing system [ 38 ]. Acetyl-P is critical for efficient degradation of unfolded or damaged proteins by ATP-dependent proteases [ 39 ]. Altogether, acetyl-P can influence the regulation of almost 100 other genes [ 40 ]. Finally, pta and/or ackA mutations were shown to affect repair-deficient E. coli mutants [ 41 ] and a pta mutant has been reported to be impaired in its ability to survive during glucose starvation, while the ackA mutant survived as well as the parent strain [ 42 ]. It is important to note that the only known pathway in E. coli for acetyl-P synthesis is the PTA-ACKA [ 5 , 31 ]. Taking all the previous into account, we conclude that acetyl-P as well as acetyl-AMP are essential for cellular growth of E. coli , and as acetic acid formation is the result of their dephosphorylation, acetic acid should be synthesized and consumed simultaneously during growth to maintain proper balance between metabolites of the PTA-ACS node. This is in agreement with Shin et al . [ 28 ] who proposed that wild-type E. coli constitutively synthesizes acetate even when growing on non-acetogenic carbon source succinate or at low growth rates in carbon limited cultures. It also has to be mentioned that acetic acid is a by-product in the synthesis of cysteine, methionine and arginine, covering around 0.4 mmol/g DCW (Additional file 4 ). Based on our experimental and literature data, production and re-assimilation of acetate might be over 1 mmol/g DCW at μ = 0.2 h -1 (Text S2 in Additional file 1 ) which further supports the hypothesis of the necessity for constant acetic acid synthesis.
A similar regulation for overflow metabolism of acetate was posed for Saccharomyces cerevisiae by Postma and co-workers: they postulated that acetate accumulation is the result of insufficient acetyl-CoA synthetase activity for the complete functioning of the pyruvate dehydrogenase bypass because of glucose repression of ACS at high growth rates [ 43 ]. The hypothesis proposed here is also consistent with the observation that an acs mutant of E. coli accumulated acetate in chemostat cultures at dilution rate (D) 0.22 h -1 whereas acetate overflow was started in wild-type at a higher D = 0.35 h -1 [ 28 ]. Furthermore, it has been shown that over-expression of acs [ 44 ] and constitutively expressed acs together with glyoxylate shunt repressors iclR and fadR mutant resulted in a significant reduction in acetate accumulation in glucose batch fermentations [ 28 ]. Adams and co-workers showed that as a result of micro-evolution, E. coli increased acetate consumption capability by over-expressing Acs (not AckA) [ 45 , 46 ], further supporting the connection between Acs activity and acetate accumulation.
As Acs down-regulation is responsible for triggering overflow metabolism and the resulting accumulation of acetate is detrimental to cellular growth, it bears questioning why E. coli has not evolved towards maintaining sufficient Acs levels for acetate assimilation in all growth conditions. Growth conditions in E. coli native environments are rough as concentrations of utilizable carbon sources including acetate are in the low mg L -1 range and access to nutrients is troublesome [ 47 ]. These harsh conditions force E. coli to make its metabolism ready for scavenging all possible carbon sources including acetate. However, in nutrient rich laboratory conditions, E. coli focuses on anthropic growth [ 48 ] and biomass production rate, primarily realized by enhancing readily oxidizable substrate (glucose) uptake kinetics which in turn results in Acs repression through CCR and thus, acetate accumulation [ 46 ]. This indicates that an active Acs is not essential for rapid growth for E. coli . It seems that maintaining high expression levels of acetate assimilation components (and also other alternative substrates ones) is energetically not favorable at higher growth rates. Moreover, as the space on cell membrane is limited and as E. coli achieves more rapid growth probably by increasing the number of glucose transport machinery components on the membrane, using area for alternative substrate transport proteins is not beneficial for faster growth. Interestingly, even in one of its natural environments-urinary tract-where a continuous dilution of acetate occurs, it has been shown that metabolizing acetate to acetyl-CoA by Acs is not essential for normal E. coli colonization as PTA-ACKA pathway and maintenance of a proper intracellular acetyl-P concentration are necessary for colonizing murine urinary tract [ 32 ].
Based on all the points discussed above, PTA-ACS might function as a futile cycle to provide rapid regulation of acetyl-P concentration in the cell for an active chemotaxis that is vital in natural nutrient-depleted environments, fighting against other organisms (acetate production), pathogenesis, biofilm formation etc. This hypothesis is consistent with the fact that the flagellar assembly and regulation operon ( tar-tap-cheRBYZ ) was more intensively expressed at lower growth rates (Additional file 2 ) where residual glucose concentration is smaller.
Concerning Acs down-regulation, it is possible that CCR is responsible for its repression as proposed by Treves et al . [ 46 ] showing the link between ACS expression level and acetate accumulation. In our experiments, it was shown that activation of CCR and repression of Acs take place simultaneously prior to the start of overflow metabolism (Figure 1 and Figure 4 ). As it is well known that CCR is initiated by the presence of glucose in the medium [ 49 , 50 ], we propose that increasing residual glucose concentration accompanying smooth rise of dilution rate in A-stat triggers Acs down-regulation by CCR. The cAMP-Crp complex is one of the major players in CCR of E. coli as cAMP binding to Crp drastically increases its affinity towards activating the promotors of catabolic enzymes, including Acs [ 6 , 49 , 50 ]. We measured a 1.5-fold decrease in Crp expression with increasing growth rate (Figure 1 ) that is in agreement with the data in the literature [ 51 ]. In addition, when E. coli mutant defective in the gene crp was cultivated in glucose-limited chemostat at a low D = 0.10 h -1 , it accumulated acetate whereas the wild-type did not [ 52 ]. Furthermore, it exhibited a 34% higher biomass yield relative to the wild-type-this increase might be explained by reduced ATP wasting in the acetate futile cycle, which can be directed to biomass synthesis. Moreover, Khankal et al . [ 53 ] noted that E. coli CRP* mutants that do not require Crp binding to cAMP to activate the expression of catabolic genes showed lowered glucose effect on xylose consumption, 3.6 times higher acs expression levels and secreted substantially less acetate in xylitol producing batch fermentations. The connection between cAMP concentration and acetic acid consumption capability, together with the two-phase acetate accumulation profile observed in A-stat and D-stat cultures (Figure 2 and Figure 5 ) suggests a correlation between increasing residual glucose concentration mediated cAMP-Crp repression and acetate accumulation. Thus, cAMP-Crp dependent regulation of Acs transcribing acs-yjcH-actP operon might be a reason for acetate excretion, as also proposed by Veit et al . [ 10 ]. Our hypothesis of the CCR mediated acetate overflow metabolism is as well in agreement with the fact that rising glucose lowers the intracellular Crp level through the autoregulatory loop of the crp gene [ 54 ]. However, other mechanisms can also be involved in Acs down-regulation, for example by Cra (Figure 1 ). Indeed, Sarkar and colleagues have shown that glucose uptake and acetate production rates increased with a decrease of acetate consumption in an E. coli cra mutant [ 55 ].
What could be the biological relevance of the disruption of the PTA-ACS node? Firstly, decline of the ATP-spending PTA-ACS cycle throughput with increasing growth rate points to possible lower ATP spilling (our model calculations). Secondly, disruption of the PTA-ACS node decreases the energy needed for expression of this cycle's components. As the disruption of PTA-ACS cycle is CCR-mediated, repression of other alternative substrate transport and utilization enzymes by CCR enables to save additional energy. This could all lead to the decrease of ATP production as was indicated by the diminishing TCA cycle fluxes (Figure 2 ). Hence, it is plausible that cells repress (by CCR) the expression levels of alternative substrate utilization components (including Acs) for making space on the cell membrane for more preferred substrate (glucose) utilization and ATP producing components to achieve faster growth (see above).
Finally, it was demonstrated that highly reproducible A-stat data are well comparable to chemostat at the level of major growth characteristics and transcriptome, hence quasi steady state data from A-stat can be considered steady state representative (Table 1 ; Figure S1 and Figure S3 in Additional file 1 ). Furthermore, as shown also by Postma et al . for S. cerevisiae [ 43 ], chemostat is not fully suitable for characterization of dilution rate dependent metabolic transitions, whereas A-stat should be considered an appropriate tool for this. A-stat is especially well suited for the studies of the details of transient metabolism processes. Dynamic behavior of acetate, cAMP etc. with increasing specific growth rate (Figure 2 Figure 3 and Figure S1 in Additional file 1 ) and change in acetic acid consumption capability in the two-substrate A-stat (Figure S2 in Additional file 1 ) could be cited as good examples of the latter. | Conclusion
This study is an excellent example of how a systems biology approach using highly reproducible advanced cultivation methods coupled with multiple 'omics analysis and metabolic modeling allowed to propose a new possible regulation mechanism for overflow metabolism in E. coli : acetate overflow is triggered by carbon catabolite repression mediated Acs down-regulation resulting in decreased assimilation of acetate produced by Pta, and disruption of the PTA-ACS node. The practical implications derived from this could lead to better engineering of E. coli in overcoming several metabolic obstacles, increasing production yields etc. | Background
The biotechnology industry has extensively exploited Escherichia coli for producing recombinant proteins, biofuels etc. However, high growth rate aerobic E. coli cultivations are accompanied by acetate excretion i.e . overflow metabolism which is harmful as it inhibits growth, diverts valuable carbon from biomass formation and is detrimental for target product synthesis. Although overflow metabolism has been studied for decades, its regulation mechanisms still remain unclear.
Results
In the current work, growth rate dependent acetate overflow metabolism of E. coli was continuously monitored using advanced continuous cultivation methods (A-stat and D-stat). The first step in acetate overflow switch (at μ = 0.27 ± 0.02 h -1 ) is the repression of acetyl-CoA synthethase (Acs) activity triggered by carbon catabolite repression resulting in decreased assimilation of acetate produced by phosphotransacetylase (Pta), and disruption of the PTA-ACS node. This was indicated by acetate synthesis pathways PTA-ACKA and POXB component expression down-regulation before the overflow switch at μ = 0.27 ± 0.02 h -1 with concurrent 5-fold stronger repression of acetate-consuming Acs. This in turn suggests insufficient Acs activity for consuming all the acetate produced by Pta, leading to disruption of the acetate cycling process in PTA-ACS node where constant acetyl phosphate or acetate regeneration is essential for E. coli chemotaxis, proteolysis, pathogenesis etc. regulation. In addition, two-substrate A-stat and D-stat experiments showed that acetate consumption capability of E. coli decreased drastically, just as Acs expression, before the start of overflow metabolism. The second step in overflow switch is the sharp decline in cAMP production at μ = 0.45 h -1 leading to total Acs inhibition and fast accumulation of acetate.
Conclusion
This study is an example of how a systems biology approach allowed to propose a new regulation mechanism for overflow metabolism in E. coli shown by proteomic, transcriptomic and metabolomic levels coupled to two-phase acetate accumulation: acetate overflow metabolism in E. coli is triggered by Acs down-regulation resulting in decreased assimilation of acetic acid produced by Pta, and disruption of the PTA-ACS node. | Authors' contributions
KV, KA, and RV drafted the manuscript. RN, PL, and LA helped in preparing the manuscript. KV, RN, and PL designed and performed the experiments. KV analysed the experimental data. RN, PL, and LA carried out the 'omics analysis. KV, KA, and RV guided and coordinated the project. All authors read and approved the manuscript.
Supplementary Material | Acknowledgements
The financial support for this research was provided by the Enterprise Estonia project EU29994, and Ministry of Education, Estonia, through the grant SF0140090s08. The authors would like to thank Lauri Peil and Elina Pelonen for help in carrying out 'omics analysis. | CC BY | no | 2022-01-12 15:21:37 | BMC Syst Biol. 2010 Dec 1; 4:166 | oa_package/a2/d5/PMC3014970.tar.gz |
PMC3014971 | 21122128 | Background
Half of the world's population is at risk of malaria, and among the 243 million clinical cases every year, the majority occurs in the world's poorest countries [ 1 ]. This is not an inevitable burden. In the past decade, there have been considerable advances in the cost-effectiveness of preventing and treating malaria, together with an increasing commitment from donors and affected countries to combat this disease. However, life-saving measures continue to be beyond the reach of the majority of people who need them. Among strategies, early diagnosis and treatment (EDAT) has been recommended [ 2 ].
In the Lao People's Democratic Republic (Lao PDR), with a current estimated population of 6 million, malaria is considered endemic throughout the country. However, intensity of transmission varies between different ecological areas ranging from 0.001 to 32.7 per thousand infected bites and according to the seasons [ 3 ]. Around 60% of the population is estimated at risk [ 3 ].
After reports of increasing resistance to chloroquine and sulphadoxine-pyrimethamine, the national malaria treatment guidelines were changed to artemisinin-based combination therapy (ACT) (artemether/lumefantrine, Coartem ® ) in 2004 [ 4 - 6 ] and malaria rapid diagnostic tests (RDT) were introduced. This early diagnosis and treatment (EDAT) strategy was initially implemented in early 2005 at the village and health centre level in three pilot provinces (PP) (Sekong, Saravan and Attapeu) covering a total of six districts and 60 villages. Two village health volunteers (VHVs) per village were selected and trained. Following the pilot phase, EDAT was adopted as a national strategy and extended to all 17 provinces.
An evaluation of a short training on MRDT procedure of 64 VHVs had shown a good reliability in 2003 [ 7 ]. However, previous studies indicated that RDT accuracy was highly user-dependent despite its apparent simplicity. Basic diagnostic tests risked being performed and interpreted poorly, but by providing simple and clear instructions these errors could be reduced, thus increasing their accuracy from 70% to 80% [ 8 ].
This study was conducted more than one year after the initial expansion of the EDAT strategy in six endemic provinces of Lao PDR. Its objective was to evaluate the effects of training initiated in the three pilot areas in 2005 compared to the training in the expanded phase for non-pilot provinces: performing an RDT test, its interpretation and prescribing behaviour of the malaria volunteers. | Methods
Malaria in the Lao PDR
In 2009, only five deaths among 22 784 confirmed cases of malaria were reported compared with 600 deaths and 70 000 confirmed cases in 1997 [ 9 ]. The mortality rate declined over the past 10 years from 0.12 per thousand in 1996 to 0.01 in 2006. Anopheles dirus , Anopheles minimus and Anopheles maculatus are the main vectors, and Anopheles jeyporiensis and Anopheles nivipes the secondary ones [ 10 ]. The National Malaria Control Programme (NMCP) is based on Roll Back Malaria strategies. The main preventive strategy is the use of insecticide-treated nets (ITN) [ 9 ]. By 2007, operational coverage of ITN was estimated to reach half of the population at risk. RDT and ACT are provided free of charge through the public health facilities [ 9 ].
Malaria village health volunteers
The village health volunteers (VHVs) constitute the most peripheral level of the public health care system in Lao PDR. Volunteers provide primary health care services, including diagnosis and management of basic diseases (respiratory diseases, diarrhoea, and uncomplicated malaria) as well as performing health education, assist in vaccination campaigns and insecticide treatment of bed-nets, and reporting of morbidity and mortality data to health or district health offices [ 3 ]. VHVs (not malaria-specific volunteers but primary health care workers) are selected by the village committee in consultation with villagers, the Community Health and the District Health Office. Selection is based on pre-determined criteria. Often the criteria cannot be met and the best person for the role is selected. Each village is required to have two volunteers. But there is often only one person who has the literacy skill to be a VHV in rural areas.
VHVs play a pivotal role in malaria control:
• Village data collection
• Monthly reports to health centre
• Health education and community mobilization campaigns
• Diagnosis and treatment
• Distribution and re-dipping of ITN [ 11 ].
Training of VHVs on EDAT
The EDAT strategy was first introduced in 20 villages/districts of three districts of the three pilot provinces in 2005. The training covered both health staff at district hospitals, health centres within the catchment areas of the target villages and VHVs (2 per village). The training was done mainly with trainers and facilitators from the central level. In subsequent years, between 2006 and 2008, the national programme gradually scaled up the coverage of RDTs and ACTs to more provinces, districts, health centres and villages. The cascade training involved trainers and facilitators from provincial level who attended training of trainers' sessions organized at the central level. The VHVs training took place over three days, which also included training in ITN distribution and use. In some provinces, pre- and post-tests were done although results of these were not available to the authors.
Cross sectional study
Due to budgetary issues and time constraints, a cross sectional study was performed in six provinces from January to February 2007. The first three pilot provinces (PP) (Sekong, Saravan, Attapeu) were selected. Of the 14 other provinces in the country, three other non-pilot provinces (NPP) (Luangnamtha, Savanakhet and Champassak) were selected because nationally they had the highest retrospective incidence of malaria (Figure 1 ). Then a two-stage random sample procedure was applied. One district per province and six villages in each district were selected to make up a total of 36 villages.
The total population of the six provinces was 2.1 million. Of these, 10 200 (4.8%) and 11 700 (5.5%) people respectively, were suspected of, or confirmed with malaria in 2005 (2005 NMCP, unpublished data). At the village level, VHVs were included in this survey if they had been trained on ACT/RDT with the malaria programme, had more than six months experience and, if ACT/RDT was theoretically available in the village, according to the provincial distribution list. In each village one out of the two VHVs was selected using a random table. A pretested questionnaire was used to assess the characteristics of VHVs, their knowledge of malaria, and to provide basic information on their activities related to patients and malaria.
The quality of practice of VHVs was assessed by observation performed by the investigators. Two checklists were used and scores were calculated as previously described [ 2 , 7 ]. Briefly:
- A standardized 18-item form with four sections (history, patient care, observation of diagnosis and treatment, information and advice) to evaluate the VHVs practice (Table 1 ). Each item was given one point. The score of VHVs practice was defined as: excellent >80% of total score, good: 60-80%, poor <60%.
- A checklist of 15 criteria to evaluate the RDT performance (Table 1 ). The total of good responses scored was 15. Using the checklist, observers noted whether the VHVs performed the steps correctly (= 1), incorrectly or did not perform it (= 0). The score for RDT was defined as bad = <5; insufficient = 5 - 10; good = 10 - 13; excellent = >13.
All villagers with symptoms potentially related to malaria were enrolled, with a minimum of 20 cases per VHVs. Patients were included on entrance to the VHVs home, if they had a history of fever of 2-3 days duration, and no history of anti-malarial therapy during the previous week.
Laboratory testing
A finger-prick blood sample was collected from each case to prepare thick and thin smears, and for RDT. A maximum of 20 tests was recorded per VHV. RDT tests were performed at the volunteer's house. All RDT results were confirmed by one expert reader of the investigation team on site. The Paracheck ® Pf Rapid One Step device for Plasmodium falciparum histidine rich protein (Orchid Biomedical Laboratories, Goa, India) was used according to a previously described procedure [ 8 ].
Thick and thin blood films were made immediately after blood collection and stained with 10% Giemsa for ten minutes. Slides were further read by two expert microscopists at the Centre of Malariology, Parasitology and Entomology (CMPE) in Vientiane. All microscopists were unaware of the results of the RDT tests. A minimum of 200 consecutive fields was counted in the thick blood film before classifying a slide as negative. Parasites in thick blood films were counted against 200-500 white blood cells. The parasite density was estimated for 8,000 white blood cells. If malaria parasites were seen, species were identified, based on their morphological features on the associated thin smear. The microscopic reading at the CMPE was considered the gold standard. RDTs were, as possible, performed according to the VHVs normal practice. In fact, some more patients came when they heard of a medical team doing tests.
Retrospective study
Anti-malarial prescriptions were retrospectively assessed by checking the VHVs' register books for the period June - November 2006.
Data management and analysis
Data was entered in Epidata 3.1 freeware (Odense, Denmark) and cross-checked against original data sheets. Analysis was carried out with Stata, Version 8 (Stata Corporation, College Station, TX, USA). Fisher's exact test was used to assess associations between categorical variables, and Student's t- test for two normally distributed continuous variables. P < 0.05 was considered as significant.
Verbally-informed consent was obtained from all participants. The survey was performed with the permission of Lao national, regional and village authorities and in agreement with the Declaration of Helsinki [ 12 ]. Ethical approval from the National Ethical Review Board of Laos was not sought since the survey did not directly involve patient interventions beyond national MoH standard guidelines (Standard treatment guidelines) and routine national malaria programme activities (i.e. active case detection surveys are frequently carried out by the NMCP). Investigators were medical doctors from the Institut Francophone de Médecine Tropicale (IFMT) attending a master's course with special lectures on epidemiology, field research and public health. Pre-tests were performed before conducting the survey. | Results
VHV assessment
This study was performed in January 2007, during the dry season. Most of the 36 villages had very basic services: seven villages (19%) had a private pharmacy and 27 (75%) had a village drug kit provided by the Ministry of Health (which included only chloroquine, quinine, and traditional herbal medicines for malaria). Since the last distribution of ACT from the CMPE (on average seven months earlier) nearly half of the VHVs had a shortage of ACT (Table 2 ).
A total of 36 VHVs and their 720 patients with fever were assessed. Their characteristics are shown in tables 2 and 3 . Of 720 patients, only 31 (4.3%) had a positive RDT. Except for fever, headache, and chills no other symptoms were reported to perform a RDT.
The global performance of VHVs was better in PP than in NPP (Table 1 , table 4 ). Of the 36 VHVs, 75% of them reported using the village's drug kit for malaria but none of these contained ACT. None knew where ACT could be obtained from, or what its cost was. ACT was prescribed infrequently either for confirmed or suspected uncomplicated malaria. VHVs from PP gave chloroquine more frequently than others (p = 0.004) (Table 2 ).
Only 16 (44%) VHVs could describe the symptoms of severe malaria. Coma and convulsions were more frequently related with severe malaria by VHVs of PP (p = 0.09, p = 0.01). Only 15 (42%) had previously seen patients with severe malaria, three (20%) died, nine (60%) were treated by the VHVs and eight were referred to the hospital after the first treatment. Of these eight, three received artesunate injection, one oral quinine, one quinine with doxycycline, and four dextrose infusion only. Artesunate injections are not included in the village's kit.
Patients' assessment
Of the 720 patients, 31 (4%) tested positive using the Paracheck ® test, versus 35 (5%) positive with microscopy; 31 (89%) Plasmodium identified were Plasmodium falciparum , and four Plasmodium vivax (Table 5 ). The sensitivity and specificity of the rapid tests for detecting Plasmodium falciparum , using the microscopy as a gold standard, were 74.3% and 99.3%, respectively. The positive and negative predictive values were 83.9% and 98.7%, respectively.
RDT assessment
VHVs generally reported that RDT was easy to use. The mean global score rate of the different steps of RDT's procedure (10.7 on 15) was rated as good (Table 1 ). Two procedures related to checking expiry date and reporting of results were performed correctly by 50% of the respondents only. The VHVs from PP had an overall better score (12.2) than the others (9.1). Two NPP provinces (Luangnamtha and Savannakhet) yielded an insufficient score (Table 6 and Figure 2 ).
ACT prescription for confirmed malaria
The overall mean performance was rated as poor (Table 1 ). VHVs of PP had a better performance than the others (score: 60% and 45%, respectively). Only one PP (Savannakhet Province) reached a satisfactory score (> 70%).
Of the 18 items, only eight were correctly answered (Table 1 ). The best performances were obtained regarding information on ACT dosage (97%) and treatment duration (97%). These items were related to the actual treatment given, which can be considered as fairly correct. Low scores were obtained with regard to disease history (29%) and patient information about worsening symptoms (45%) or side effects.
From June to November 2006, 2 188 febrile patients were tested, of which 35.19% were positive (Table 6 ). Patients testing positive for malaria ranged from 5.4% in NPP to 41.9% in PP, respectively (p < 0.001). They had a 10-fold chance of testing positive compared to January ((35.19% vs. 4.86%, OR: 10.6, 95%CI: 7.4-15.5, p < 0.001). Correct prescription of ACT ranged from 33% to 97% of confirmed cases. Two PP reached an excellent level of ACT prescription (Table 6 ). | Discussion
The study evaluates the knowledge and practice of recently trained VHVs regarding the new EDAT malaria national strategy in the Lao PDR. It evaluates both the practice of RDT and the quality of ACT prescription of a sample of VHVs in 36 villages of six provinces. It confirms an overall satisfactory performance of RDT and supports the effectiveness and feasibility of RDT at the village level. It underlines persistent gaps in knowledge, care of patient and mistreatments among VHV as providers that have to be addressed. Overall, pilot provinces yielded better results than non-pilot provinces regarding care of patients, practice of RDT and ACT dispensation.
Using RDT, the survey also illustrates the decreasing rate of malaria morbidity at the village level in provinces considered as endemic. The incidence of confirmed malaria was low in northern, central and south-east provinces during the potential high malaria season. In the Lao PDR, the number of confirmed cases dropped from 53 156 in 1997 to 18 382, the year of the survey [ 1 ]. Transmission was probably very low and probably even stopped in January (dry season) in these regions [ 3 ]. By contrast, the incidence of malaria was persistent all over the year in the southern region. This justified the choice of these regions as pilot provinces. The remote locations of the survey at the village level probably reflect the reality of malaria since most of the patients had no other access to care. The population coverage was not calculated, which does not allow for estimating the incidence of malaria in these regions.
Plasmodium vivax represented 12.9% of confirmed malaria, a rate similar to that previously described, which suggests no relative decrease of this species in the Lao PDR, as experienced in Thailand [ 13 , 14 ]. However, the seemingly uniform lower prevalence of vivax malaria in endemic provinces might be partly because of the tendency of P. vivax to achieve and maintain lower-density parasitaemia, and the inverse relation between diagnostic sensitivity and parasitaemia count [ 15 ]. The detection of low level asymptomatic and mixed species infections with conventional light microscopy is limited and the RDT adopted in the Lao PDR does not diagnose P. vivax infection. Albeit the many challenges of introducing P. vivax diagnosis and treatment in Laos, the CMPE should consider a preliminary introduction of a P. vivax RDT.
VHVs are key persons to improving EDAT at village level in countries with poor health facilities. Until recently, most patients with fever were treated for malaria, based only on clinical suspicion by the VHVs, and most VHVs and pharmacy owners lacked basic knowledge of malaria management [ 2 ]. As a result, malaria tended to be overestimated and many patients were probably wrongly treated (infectious diseases such as rickettsia, typhoid or leptospirosis mimicking malaria are common in Laos) [ 16 , 17 ]. The survey supports the positive impact of training VHVs in PP, which yielded a generally better score of performance than others regarding both RDT assessment and ACT prescription. More training focussing on prevention of disease, worsening symptoms, and information of the patients are suggested for both groups of provinces.
Analysis within the provinces reveals a high discrepancy between them: only one province (Saravan) had an excellent score of achievement, but two out of the six had a very poor level. Comparing RDT at village level with the gold standard microscopy at the provincial level yielded predictive values above 80%. This confirmed the accuracy of RDT reading among the VHVs surveyed at the village level, but results were less optimistic than the previous survey, and warrants regular monitoring and quality control [ 7 ].
A possible limitation of this survey is that it was conducted during the low transmission season. The peak malaria transmission months in Laos are from July to October. However, the survey shows that malaria was still acute in two settings (Attapeu and Sekong) during the non-peak seasons. This study attempted to take this into consideration by including a retrospective survey covering six months of the transmission period. This later survey had the usual limitations of retrospective survey (incomplete files, omissions, inaccuracy).
To decrease the possibility of a bias due to the attractiveness of the medical team performing the survey, and the possibility of the VHVs showing their practice in their best light, VHVs were informed that they should continue their work as usual and that their identity would remain confidential. Nevertheless, the survey may probably have slightly overestimated the performance of VHVs, mostly for the first patients.
In fact, of the 720 febrile patients seen in January, only 4.8% had malaria parasites and 4.3% had a positive RDT. This potentially saves around 689 malarial treatments that would have previously been given in the absence of a confirmed diagnosis, and the use of ineffective chloroquine increasing the drug pressure [ 6 ]. The RDT strategy allowed to dramatically reduce the number of unnecessary cases of malaria treatment and potentially allows focusing better on other diseases. The benefits of RDT strategy were also well supported by the retrospective survey in the high malaria season, which revealed a 35% rate of malaria cases among the 2,188 suspected patients. This allowed the saving of 1 418 malaria treatments enabling the patient to save money, an important issue in a country were two-thirds of the population lives on less that USD 2 a day [ 18 ].
The low use of ACT despite the new policy is of concern. VHVs still prescribed non-ACT and ineffective chloroquine. Only a quarter of them would choose ACT. This signals the need for further information/training of VHVs to ensure compliance with the national treatment strategy.
Paracheck ® and Coartem ® are supplied not through the MoH village drug kits but through a separate vertical distribution system from the CMPE, which will probably improve after the first years of implementation. Timely reporting on consumption and stock levels by VHV may require an incentive program if out-of-stock situations are to be avoided.
Detecting symptoms of severe malaria and starting early treatment are key strategies to decreasing malaria related mortality. Despite recent training, VHVs still lacked knowledge on severe malaria detection and treatment [ 7 ]. Less than 20% of the VHVs identified 3-4 of the criteria of severe malaria. Only 41% ever diagnosed severe malaria. A total of 4 752 cases of malaria in 2006 were hospitalized (a proxy for severe malaria) in Laos, of which 3 910 (82%) were reported in the six surveyed provinces (CMPE, unpublished data). Since only half of the severe malaria cases were referred in the survey, this suggests reconsidering and probably doubling the estimation of severe malaria in Laos.
Additionally, in this study, the rate of mortality among severe malaria patients reached 20% at the village level. Hence, the figure of mortality obtained in only 36 villages suggests a probable underestimation of the official malaria deaths (21 reported deaths in 2006) [ 9 ].
Pre-referral treatment was only performed for half the severe malaria cases. Of them, only five (55%) received an anti-malarial treatment. Interestingly, injectable artesunate was not available in the village kits and the CMPE, as per guidelines, distributed this drug only to district and provincial hospitals. It can be hypothesized that some VHVs were given vials of artesunate from the CMPE programme. In fact, VHVs did not know where to get ACT or what it cost. They may also be vulnerable to ambulant drug vendors who sell fake or substandard artemisinin based anti-malarials, a major threat in Asia [ 19 , 20 ]. More studies should focus on the alternative treatments given at the village level and assess their effectiveness. VHVs need to know that counterfeit and sub-standard anti-malarials, including artesunate monotherapies, are widespread and that a safe source of ACT would be the national malaria programme. These results question the availability of potent anti-malarial treatment at the village level where one cannot expect a high rate of referral to a higher health care facility. | Conclusion
RDT and ACT, the new tools for EDAT in Laos are effective and well understood by VHVs. The survey confirms the ability of VHVs to use RDT, but underlines persistent gaps in knowledge, care of patients and treatment, which warrant further follow up, and training. The simplification brought by EDAT at village level will probably help the control of malaria, together with other strategy programmes. | Background
Early diagnosis and treatment (EDAT) is crucial to reducing the burden of malaria in low-income countries. In the Lao PDR, this strategy was introduced in 2004-2005 and an assessment was performed at the community level in January 2007.
Methods
EDAT with malaria rapid diagnostic test (MRDT) and artemisinin combination therapy (ACT) was prospectively assessed among 36 randomized village health volunteers (VHVs) and 720 patients in six malaria-endemic provinces of Laos (three pilot provinces (PP), and three non-pilots provinces (NPP)). ACT was also retrospectively assessed among 2188 patients within the same areas from June to November 2006. Two checklists were used and scores were calculated.
Results
EDAT performance of the VHVs was rated better in PP than in NPP (16.67% versus 38.89%, respectively, p = 0.004). Nearly all VHVs could diagnose malaria but only 16 (44%) could describe the symptoms of severe malaria. In January 2007, 31/720 (4%) patients tested positive using the Paracheck ® test, 35 (5%) with microscopy (sensibility: 74.3%, specificity 99.3%, positive and negative predictive values: 83.9% and 98.7%, respectively). Patients from June to November were at higher risk of malaria: 35.19% of 2,188 febrile patients were positive (OR: 10.6, 95%CI: 7.4-15.5, p < 0.000). VHVs reported the MRDT easy to use, and yielded a satisfactory performance score. EDAT performance was rated as poor despite satisfactory results regarding ACT treatment, duration and dosages. Pre-referral treatment of severe malaria was infrequent and often inadequate, with 20% of these patients dying. Results suggest a higher mortality from severe malaria than officially reported. Shortage of ACT was frequent.
Discussion and conclusion
MRDT and ACT are useful and efficient and can be used by VHVs. VHVs' global EDAT performance is enhanced through training and monitoring. Persistent gaps in knowledge, care of patients and wrong treatment have to be addressed. | Perspectives
RDT is easy to use, reliable and an appropriate test for use in the field by VHVs in the Lao PDR. Results also suggest more training to improve EDAT. Training and monitoring of user behaviour have to be reinforced especially in the provinces with poor success rates. Appropriate training and instruction have to be adapted to each province. Provinces with stable or decreasing transmission of malaria should be trained in the treatment of non-malaria fevers (dengue, arbovirosis but also infectious diseases such as leptospirosis, melioidosis, typhoid, rickettsia which are common but with scarce prevalence data at the national level) [ 16 - 18 , 21 ]. In endemic provinces the problem may be to scale up prevention and care of patients with appropriate diagnosis and early treatment. Current decrease of malaria and the emergence of dengue fever outbreaks have to be regularly assessed and monitored in the countryside.
Conflict of interests
The authors declare that they have no competing interests.
Authors' contributions
HB wrote the manuscript and is responsible for the overall coordination, design and analysis of the surveys in the Lao PDR. He is also guarantor. HB and GD designed the survey with the participation of PV and KT. PV and KT collected the data and performed the first analysis under HB's monitoring with regular advice from GD. HB performed the final analysis. GD, PV, KT, JPR and HB discussed the results. PV and KT wrote the first report with the participation of GD, JPR and HB. All authors contributed to the writing of the paper and read the final version.
Funding
Global Fund against aids, tuberculosis and malaria through the CMPE of Lao PDR funded the survey | Acknowledgements
We thank CMPE staff, IFMT staff and all the people who participated in collecting data during the surveys and all the participating VHVs, patients and families, and the Lao authorities. We thank Dr Bouasy, Dr Viengsay, Dr Keophouphaphone from the CMPE and Dr Manisak and the staff of National Institute of Public Health for help during the survey preparation and data collection.
We thank Dr. Philaysak Naphayvong and Dr Phengta Vongphrachanh for facilitating the relationship with the Lao Authorities during the preparation of the survey. We also thank Dr B Hongvankhame, V Navaviseth, the provincial and district malaria stations in the survey areas for their helpful contribution and collaboration during the survey. We thank Percy Aaron for editing the manuscript. We thank the Lao National Health Authorities for their support, WHO for valuable advice, the Global Fund for making available the funds to CMPE to fund the surveys and IFMT staff for their administrative and technical support. | CC BY | no | 2022-01-12 15:21:37 | Malar J. 2010 Dec 1; 9:347 | oa_package/82/ef/PMC3014971.tar.gz |
PMC3014972 | 21129192 | Background
Malaria parasites are a major cause of mortality and morbidity, resulting in over a million deaths each year and 350 to 500 million clinically significant malaria infections [ 1 ]. Folate metabolism is the target of a number of anti-malarial drugs, which, though compromised by the occurrence and spread of resistance within parasite populations, remain important in treatment and prophylaxis [ 2 , 3 ]. For almost all organisms, the folate pathway is essential in maintaining a constant supply of cofactors that act as donors or acceptors of one-carbon (C 1 ) units in a variety of biosyntheses. In malaria parasites, the most prominent of these is the synthesis of pyrimidines required for DNA replication [ 4 ]. Unlike mammals, Plasmodium falciparum cannot salvage thymidine and thus relies completely on the folate-dependent production of dTMP. The folate pathway can be conveniently divided into two main sections: the first five enzyme activities effect the de novo biosynthesis of the basic folate moiety, 7,8-dihydrofolate (DHF), with further enzymes interconverting the fully reduced form 5,6,7,8-tetrahydrofolate (THF) to the various derivatives utilized in C 1 transfer reactions. Plants and most micro-organisms, including many protozoa, are able to synthesize folates de novo . In contrast, higher organisms must obtain folate from the diet or commensal microorganisms. It has been shown that P. falciparum has the ability to exploit both de novo synthesis and folate salvage routes for its metabolic needs [ 5 - 7 ].
The later part of the folate pathway directly relevant to DNA replication is termed the thymidylate cycle. In this, dihydrofolate reductase (DHFR; EC 1.5.1.3) catalyses the reduction of DHF to THF. Serine hydroxymethyltransferase (SHMT; EC 2.1.2.1), the subject of this study, reversibly catalyses the conversion of serine to glycine, whereby the hydroxymethyl group of the former is transferred to THF yielding 5,10-methylenetetrahydrofolate (5,10-methylene-THF), which is then used by thymidylate synthase (TS; EC 2.1.1.45) as the C 1 donor to convert dUMP to dTMP. Concomitantly, the folate cofactor is oxidized to the dihydro-form, making a functional cycle that is capable of reducing this back to THF essential for continued DNA synthesis. A further activity, folylpolyglutamate synthase (FPGS; EC 6.3.2.17), part of a bifunctional protein also carrying dihydrofolate synthase (DHFS; EC 6.3.2.12) [ 8 - 10 ] adds a variable length polyglutamate tail to reduced folate cofactors, a phenomenon involved in subcellular storage and the retention of folates within the cell [ 11 - 13 ].
Despite much research detailing the biochemistry of the folate pathway and the genetic basis of resistance to antifolate drugs, there has been very little investigation of the subcellular location of folate pathway enzymes or their metabolites in malaria parasites. In other eukaryotes there is substantial evidence for the compartmentalization of folate metabolism within the cell. In particular, fully reduced substituted folates (such as 5-methyl-THF) appear not to exchange between mitochondrial and cytoplasmic compartments, suggesting that limited transport of intermediates between subcellular compartments may be an important factor in enzyme localization [ 14 ].
Consistent with this, the majority of methyl derivative forms are associated with the vacuole and cytosol, whereas formyl derivatives make up the greatest proportion of folates located within organelles, at least in plants [ 15 ]. In these organisms, the first two enzymes of biosynthesis, GTP cyclohydrolase I (GTPCH; EC 3.5.4.16) and dihydroneopterin aldolase (DHNA; EC 4.1.2.25), are found exclusively in the cytoplasm, whereas most of the remaining enzymes are located exclusively in the lumen of the mitochondrion. SHMT and FPGS are exceptional as they are found in both the cytoplasm and mitochondrion as well as the plant chloroplast. The two forms of SHMT, SHMTc and SHMTm, occur as distinct proteins encoded by different genes [ 16 ], but the chloroplast enzyme appears not yet to have been characterized as a separate isoform. Distinct SHMT isoforms are also found in the cytoplasm and mitochondria of yeast and the trypanosomatid Leishmania major [ 17 ], and in mammals, the distribution of both SHMTc and SHMTm differs between tissues and at different stages of development [ 18 , 19 ]. Glycine formed from SHMT acting on serine can feed into the glycine cleavage complex (GCV) of the mitochondrion, which provides an additional source of C 1 units by transferring the α-carbon of glycine onto THF [ 20 ]. Plasmodium falciparum SHMT, described here as PfSHMTc, is encoded by a single copy gene at locus PFL1720w [ 8 ] and its enzymic properties are well characterized [ 21 - 23 ]. A second open reading frame has also been identified (PF14_0534) that encodes a product with an 18% identity to PfSHMTc and incorporates a putative mitochondrion-specific tag [ 24 ]; this protein is described here as PfSHMTm. However, its sequence displays an almost complete lack of conservation of those amino acids that constitute the active site residues of all other SHMT isoforms, both cytoplasmic and mitochondrial, consistent with a failure to detect SHMT activity in a recombinant form of the protein [ 23 ]. The metabolic or other function of this related gene product thus remains to be identified, and particularly whether it could act in conjunction with a GCV in plasmodial mitochondria. Components of a potential GCV have been identified bioinformatically in P. falciparum [ 24 ], although experimental evidence for their mitochondrial location has thus far only been established for the H-protein [ 25 ].
The multiplicity of environments that the parasite must accommodate in its complex life cycle suggests that adaptability in folate metabolism and its enzymes is highly probable, and that variation in enzyme localization over the life cycle might occur. Additionally the parasite exhibits a number of unusual developmental features that could result in differences in folate metabolism from other eukaryotes. The schizogonic nature of asexual reproduction, with its repeated and apparently asynchronous cryptomitoses, results in an atypical cell cycle [ 26 , 27 ]. Peculiarities in the timing and duration of events associated with DNA replication may result in temporal variation in the demand for pyrimidine synthesis [ 28 ]. Here, patterns of localization of PfSHMTc during the erythrocytic cycle of P. falciparum are investigated because of its key role in dTMP synthesis, the strongly modulated level of transcriptional control of its gene and its relatively higher levels of expression compared to other enzymes in the folate pathway [ 29 - 31 ]. In parallel, the localization of the enigmatic PfSHMTm protein is investigated, which shows similarities in its behaviour, but with distinct and important differences from the PfSHMTc isoform. | Methods
Cloning and heterologous expression of the pfshmt genes
The full-length cytoplasmic pfshmt gene (PFL1720w) was amplified from cDNA previously cloned into pMALc2x (New England Biolabs). The intronless full-length shmt homologue PF14_0534 was amplified from K1 isolate genomic DNA, and both products were cloned into the pET-46 Ek/LIC vector (Novagen). The cytoplasmic pfshmt clone was expressed in the BL21 (DE3) pLysS expression host whilst the PF14_0534 ORF was expressed in Rosetta 2 (DE3) pLysS (Novagen). Cultures of both clones were harvested using the Bugbuster kit (Novagen) and the insoluble phases subjected to SDS-PAGE and subsequently blotted onto nitrocellulose. Fractions were loaded on the PAGE gel to give equal protein quantities in the bands of interest between the two expressed protein products. Western blots were probed with anti-PfSHMTc IgY or anti-PfSHMTm IgY, or anti-polyhistidine IgG primary antibodies (see below) followed by the appropriate AP conjugate secondary antibodies (Promega), and developed using standard methods [ 32 ].
Parasite culture and transfection
Parasites (either K1 or 3D7) were grown in 25 cm 2 tissue culture flasks with 1 ml of blood (type O; 50% haematocrit) and 10 ml of medium as described [ 33 ]. Flasks were harvested at a parasitaemia of 8 - 15%. The use of synchronous cultures was investigated but yielded no significant advantage owing to the inherent asynchrony of the repeated mitoses in individual cells [ 26 , 34 ]. The developmental stage of a particular parasite within asynchronous cultures was ascertained through its size, haemozoin development, number of nuclei and overall morphology. For GFP-labelling studies, 3D7 parasites were transfected with appropriate plasmid constructs encoding SHMT-GFP fusion proteins essentially as described [ 35 ], using the primers pfSHMTm-kpn-s (gcgcggtaccATGCTGAAGGAGTTTGTTAAAAATG) and pfSHMTm100-avr-as (gagacctaggGCAACCCCAATATTTCTTTTGTAA) to clone the truncated pfshmtm gene described in Results into the pARL1a- vector [ 36 ].
Western blotting of parasite extracts
Parasite extracts were also prepared for western blotting by freeze-thawing, in which 1 ml of blood at 50% haematocrit and ca . 10% parasitaemia was saponin lysed and washed in PBS. The resulting parasite pellet was resuspended in 0.1 ml of deionized water and subjected to 5 rounds of freezing and thawing. Following centrifugation the supernatant was recovered and 20 μl (equivalent to ca . 10 8 parasites) used per lane on 12% acrylamide SDS-PAGE gels. Protein was transferred to nitrocellulose using a Biorad Mini Protean II blotter; blots were probed with primary antibody and secondary alkaline phosphatase-conjugated antibodies (Promega).
Antibodies
The PfSHMTc primary polyclonal antibody (IgY) was raised in chickens against the denatured product of a 70-codon DNA segment [369-GIRIG...QWAKN-438] located towards the 3' terminus of the pfshmt gene (PFL1720w) expressed in E. coli as a GST fusion. The PfSHMTm primary polyclonal antibody (IgY) was raised in chickens against the denatured full-length gene product expressed in E. coli as a His-tagged fusion. The same gene product was additionally used to raise antibodies (IgG) in rabbits. All three antibodies were commercially produced by Eurogentec. The donated apicoplast-specific antibody, anti-acyl carrier protein (anti-ACP; IgG) was raised in rabbits [ 37 ], as were the donated antibodies against the cytoplasmic enzymes chorismate synthase and cyclin-dependent protein kinase 5 [ 38 ]. The donated 3D7 parasite transfectant strain carrying pSSPF2/PfACP-DsRED [ 39 ] was used to confirm data obtained with the anti-PfACP antibody and control for possible interactions between this and other primary antibodies used simultaneously. The secondary antibodies, Alexafluor (488, 546 and 594 nm) goat anti-chicken IgY and anti-rabbit IgG, were obtained from Molecular Probes, as were the MitoTracker Orange CMTMRos mitochondrial probe and the DNA stain YOYO-1 (491/509 nm).
Immunofluorescence: parasite fixation, permeabilization and staining
A preparative method was developed to maximize the fluorescence intensity of the target proteins by ensuring a high degree of penetration of both primary and secondary antibodies and sufficient incubation with the primary antibody. This was necessary as soluble enzymes are frequently found in relatively low concentrations and are thus less readily visualized than structural, membrane-associated or exclusively organelle-bound proteins. Moreover, malarial folate pathway enzymes are known from both transcriptional and proteomic measurements to be expressed at low levels [ 29 - 31 ]. Preservation of the erythrocyte membrane proved to be largely impracticable due to lysis caused by a combination of detergent extraction and the mechanical stresses inherent in the mixing and centrifugation steps. Transmission light images were thus relatively poor and often obscured by erythrocyte ghosts; they are included merely to indicate the position of the haemozoin within the parasite pigment vacuole. However, the high degree of preservation of the parasites themselves using this method is evident in the undistorted images of internal fine structure shown. Significantly, the control apicoplast-specific antibody (anti-PfACP; see below) showed no fluorescence within the apicoplast in the absence of detergent permeabilization, demonstrating that any immunofluorescence investigation of the internal distribution of proteins within parasites should always ascertain the necessity of such a step.
Giemsa-stained thin blood smears were taken from all cultures used for immunofluorescence imaging upon harvesting to ensure that parasites showed normal undamaged morphology and healthy growth. Cultures selected for mitochondrial staining were incubated at 37°C for 45 min with MitoTracker Orange CMTMRos freshly dissolved in dimethylsulfoxide to give 100 nM final concentration in the medium. A 1 ml volume of parasitized erythrocytes was harvested by centrifugation (3,000 g, 5 min). Pelleted cells were resuspended and fixed in 5 ml freshly prepared 3.7% (w/v) paraformaldehyde in phosphate buffered saline (PBS) for a minimum of 2 h at 4ËšC. Following fixation, the parasitized blood was centrifuged (as above) and washed twice in blocking/wash solution (1 ml PBS, 0.5% (w/v) BSA, 2% (v/v) bovine serum; Sigma) in parallel-sided, screw-capped microfuge tubes with rotational mixing at room temperature for 5 min followed by centrifugation (8,000 g, 30 s). Cells were then incubated in wash buffer plus 0.25% (v/v) Triton X-100 for 5 min to increase permeability. After a further three washes in wash buffer (also used in all subsequent washes), primary antibodies (diluted 1:100 in wash buffer) were added and incubated overnight at 4°C. The samples were then washed four times and incubated with fluorescent secondary antibodies, diluted to the manufacturer's specification (usually 2.5 μl in 1,000 μl), for 2 - 4 h at room temperature. This was followed by three washes, then the DNA was labelled by the addition to the cells suspended in 1 ml of wash solution of 20 μl YOYO-1 (diluted 1:1,000 in wash buffer) and incubation at room temperature for 5 min. The cells were then centrifuged, washed for 1 min and then immediately centrifuged again (as above). The pelleted cells were resuspended in 100 - 250 μl Mowiol (Harco, UK) mountant and mounted on microscope slides under a coverslip [ 32 ].
Microscopy
Parasites labelled for immunofluorescence were viewed by laser scanning confocal microscopy using a Zeiss Axiovert 200 M microscope with argon (548-514 nm) and helium/neon lasers (543 nm, 632.8 nm) using a 100× oil immersion objective lens. Images were viewed and analysed using a combination of Zeiss LSM image software and Imaris 5.7.1 software (Bitplane Scientific Solutions). The latter allows the qualitative display of combined colours from co-localized probes to be quantitatively analysed by providing measurements of their overlap in three dimensions by analysing z-stack scans taken through the whole span of the organelle [ 40 ]. This enables a much more accurate assessment of coincidence of the labels than is possible with single-plane images. Co-localization is expressed as a percentage of the individual fluorochrome volume and material (the latter derived from volume and fluorescence intensity) that occupies the same 'voxels' (three-dimensional pixels) as the second fluorochrome. For a single cell, similar values between volume % and material % co-localized indicate similar concentrations inside and outside the organelle, whereas a higher organellar material % compared to volume % indicates concentration of the target protein relative to the cytoplasm. Three-dimensional projections were created from scans with a z-axis interval of 0.2 μm. This was the minimum increment possible before the scans became excessively long, resulting in unacceptable levels of photobleaching. Scans were sequential, with each colour wavelength scanned in rotation for each single plane image or within each plane in a z stack. For clarity, orange wavelength fluorescence (Alexafluor 546, DsRED and MitoTracker) is false-coloured green, green wavelength fluorescence (Alexafluor 488 and YOYO-1) is false-coloured blue, but far-red fluorescence (Alexafluor 594) is unchanged in all of the images displayed. Transfected parasites expressing GFP-fusion protein endogenously were imaged as previously described using Hoechst 33342 to visualize the nuclei [ 35 ] and MitoTracker Red CMXRos (0.625 nM) to visualize the mitochondrion. All images presented are representative examples of each feature as seen in multiple samples.
Imaging controls
Antibody extracts from pre-immune yolk showed no bands when applied to western blots of total parasite proteins in parallel with the antigen-specific antibodies. Control slides labelled with secondary antibodies alone, or with combinations of secondary antibodies (used to control for artifactual interactions), showed no visible fluorescence when scanned using identical microscope settings and computer processing parameters as those used in producing the images shown. As the anti-PfSHMTc antibody was raised to a GST fusion polypeptide, the possibility of it recognising a plasmodial GST orthologue was excluded using a commercial polyclonal anti-GST IgG (GE Healthcare) on parasites as described above, which also showed no visible fluorescence. Haemozoin auto-fluoresces at a number of wavelengths, however its crystalline nature makes its fluorescence easily recognized and an appropriate choice of filters avoided interference with any of fluorochromes used.
To provide controls against the possibility that the preparative method used might artifactually produce organellar fluorescence in a non-specific manner, polyclonal antibodies against two unrelated enzymes, chorismate synthase and cyclin-dependent protein kinase 5, were also employed, which had been previously characterized as showing a simple cytoplasmic distribution in P. falciparum [ 38 ]. These control antibodies were employed with 3D7 parasites expressing the DsRED labelled apicoplast-specific protein PfACP [ 39 ]. Parasites were treated with control antibody in parallel procedures alongside parasites treated with anti-PfSHMTc and anti-PfSHMTm antibodies. No level of apicoplast-specific fluorescence was observed with either control antibody, which produced a generalized staining of the parasites with no evidence of fluorescence adopting the shape of apicoplasts (see Additional file 1 Negative control images for organellar staining). Furthermore, to exclude the possibility that artifactual interactions between the apicoplast-specific antibody anti-PfACP and the anti-PfSHMT antibodies were occurring, the latter were also used in conjunction with the above DsRED-transfected parasites, yielding identical patterns as those obtained using two primary antibodies simultaneously (see Results). | Results
Antibody specificity with respect to PfSHMTc and PfSHMTm
The pfshmt gene from P. falciparum (PFL1720w) [ 41 ] encodes a product that has been functionally characterized as a conventional cytoplasmic SHMT [ 21 - 23 ]. However, a predicted SHMT-like gene product (PfSHMTm, encoded on PF14_ 0534) was also identified that carries a putative mitochondrial signal sequence [ 24 ] with 18% amino acid identity and 44% similarity to PfSHMTc, but lacks almost all (16 of 21) of the known, very highly conserved residues [ 42 ] contributing to the active site in SHMT orthologues from other organisms, whether cytoplasmic or organellar (See Additional file 2 Sequence alignments of the PfSHMT isoforms). Despite the relatively low level of identity, it was essential to establish the specificity of the anti-PfSHMTc and anti-PfSHMTm antibodies that had been raised to be certain of the identity of the protein yielding positive signals. Both full-length open reading frames were therefore cloned in Escherichia coli expression systems and equal amounts of protein products processed for western blotting. The anti-PfSHMTc antibody recognized the heterologously expressed cognate protein (Figure 1B ) and blots of total parasite lysates from two lines, K1 and 3D7, showed a single band also at the predicted size (49.8 kDa) for the full length PfSHMTc protein (Figure 1D ). Importantly, there was no evidence for cross-reaction with the PfSHMTm product of PF14_0534 (Figure 1B ), whereas control anti-His-tag antibodies recognized both recombinant products essentially equally (Figure 1A ). This engendered confidence that subsequent immunofluorescence signals using the cognate antibody arose solely from PfSHMTc. In the case of PfSHMTm, this antibody was raised to the whole protein (unlike the anti-PfSHMTc antibody), some cross-reaction with PfSHMTc was not unexpected and was evident on blots against recombinant protein. However, this was approximately fourfold less intense than that seen in recognising the cognate PfSHMTm protein (Figure 1C ). Against parasite extracts, the anti-PfSHMTm antibodies gave a predominant band with the same mobility as the recombinant protein (Figure 1E ). It was noted on all blots that PfSHMTm ran slightly ahead of PfSHMTc, despite its somewhat higher predicted molecular weight (55.2 kDa). These differences in specificity led us to conclude that the differences seen below in immunofluorescence images of parasites probed with anti-PfSHMTc from those produced using anti-PfSHMTm are a reliable indicator of biologically significant variations in the distribution of the respective target proteins.
Cytoplasmic distribution of the PfSHMT isoforms
SHMT subcellular distribution in a number of organisms shows a partition between cytoplasmic SHMT and distinct isoforms of the enzyme located within organelles. As only PfSHMTc has thus far been confirmed as enzymatically active in P. falciparum [ 21 - 23 ], a single cellular location might be predicted. However, initial probing using its cognate antibody showed that PfSHMTc does not follow such a simple distribution pattern during the erythrocytic cycle. All stages showed an expected generalized cytoplasmic staining and this, by visual examination and volumetric analysis by the Imaris software, is where the majority of the PfSHMTc molecules are located for most of the time. However, fluorescence brightness within the cytoplasm was not uniform and constriction of cytoplasm between organelles, especially nuclei, produced a patchy appearance (Figure 2 ). The PfSHMTm protein (Figure 3 ) showed an almost identical cytoplasmic distribution to that described for the PfSHMTc enzyme, as can also be seen in Figure 4 , 5 , 6 , 7 , 8 and 9 , in which the anti-PfSHMTc and anti-PfSHMTm antibodies are used in various combinations with organellar labels. However, images obtained where both antibodies were used in combination did show some minor differentiation in cytoplasmic localization and relative concentration within individual parasites, exemplified by Figure 5C, D and 5F .
Mitochondrial localization of PfSHMTc
The mitochondrion and the apicoplast undergo a similar, though not simultaneous, morphological evolution during the development of erythrocytic stage parasites. The two organelles are found in close physical association and a junction between their respective membranes has been described [ 43 , 44 ]. The organelles increase in size, and in the case of the K1 isolate used here, were often observed to adopt a globular shape in the early schizont stage; thereafter they lengthen and ramify, eventually dividing to allow one of each organelle to associate with each individual developing merozoite [ 45 ]. These organelles thus have a requirement for folate pathway metabolites for the synthesis of DNA precursors needed for the replication of their genomes.
The mitochondria, visualized using MitoTracker, showed some evidence of associated PfSHMTc fluorescence throughout the erythrocytic cycle but predominantly during the stages associated with DNA replication. In many early and late parasites, the mitochondria were physically very small and consequently it could not be concluded with any certainty that PfSHMTc fluorescence was within the organelle lumen or merely in the adjacent cytoplasm. Indeed some early to mid-trophozoites showed no evidence of PfSHMTc fluorescence within their mitochondria. However, some mid-trophozoites showed a more convincing co-localization, e.g. Figure 2A , while the larger mitochondria found in very late trophozoites and early schizonts, such as shown in Figure 2B , clearly showed PfSHMTc fluorescence within the lumen, though at a similar concentration to that in the immediately surrounding cytoplasm. Figure 2C is an example of a post-mitotic schizont where very little co-localization remains.
Calculation of the levels of co-localization of PfSHMTc and MitoTracker reinforces this qualitative conclusion. The percentage of PfSHMTc material co-localizing varied between 2.2% and 5.8% (Figure 2 ) and the percentage volume of PfSHMTc co-localized showed only a similar, or slightly higher, value in comparison, confirming that the mitochondrion does not accumulate a noticeably higher concentration of PfSHMTc than that found in the cytoplasm. The three-dimensional projection within Figure 2C gives a particularly good view of a post-mitotic schizont showing the close spatial connection between the nuclei and mitochondria destined to occupy the same daughter merozoite. However, the mitochondria in this late stage parasite showed little evidence of PfSHMTc staining.
Mitochondrial localization of PfSHMTm
The use of anti-PfSHMTm revealed a different pattern of mitochondrial co-localization. The PfSHMTm protein, in contrast to PfSHMTc, was found strongly associated with the mitochondria throughout the erythrocytic cycle, from early trophozoites to late, post-mitotic, schizonts. The mitochondrion was always found within regions of relatively high intensity PfSHMTm fluorescence (Figure 3A and 3E ) and in many instances the shape of the PfSHMTm fluorescence conformed to the shape of mitochondria (Figure 3B, C and 3D ). Importantly, there were no instances of scanned images where mitochondria were found without associated PfSHMTm fluorescence or where such fluorescence was visibly lower than that of the adjacent cytoplasm. However, the quantitative analysis for Figure 3 gave very similar figures for percentage PfSHMTm material co-localized with the MitoTracker compared with percentage volume co-localized, suggesting that there was no active accumulation of PfSHMTm within the mitochondria above the levels in the cytoplasm. The percentages of PfSHMTm material co-localized with MitoTracker varied between 5.0% and 12.9%, a higher range of values than measured for PfSHMTc (2.2 - 5.8%).
Apicoplast localization of PfSHMTc
In contrast to the relatively weak spatial association between subcellular PfSHMTc distribution and the mitochondrion, the apicoplast exhibited a distinctly more pronounced relationship. The apicoplast was visualized in two ways: using antibodies to acyl carrier protein (anti-ACP), which is apicoplast specific [ 37 , 45 ] and using a transfected 3D7 line constitutively expressing DsRED-tagged PfACP [ 39 ]. The parasite shown in Figure 4A was at the mid-trophozoite stage, and although the apicoplast was still relatively small, PfSHMTc fluorescence was clearly co-localized with anti-ACP, indicating that it was within the lumen of this organelle. The parasites in Figure 4B and 4C are early schizonts, at which stage the apicoplast is considerably larger in absolute volume, as well as relative to overall cell volume. In the K1 isolate used in these images, the apicoplast often assumes first an enlarged globular form, which then elongates before ramifying. All of these parasites showed a bright PfSHMTc fluorescence coincident, or largely coincident, with the anti-ACP fluorescence that defines the position of the apicoplast, with the surrounding general cytoplasmic PfSHMTc fluorescence being perceptibly less bright. The parasite shown in Figure 4B displays the earlier globular apicoplast morphology, the parasite in Figure 4C contains an apicoplast that has started to elongate. The three-dimensional projection of the parasite in Figure 4C also allows a clear visualization of the small punctate concentrations of fluorescence that are suggestive of a vesicle-associated location of PfSHMTc, often seen in close proximity to the haemozoin containing pigment vacuole in trophozoite and early schizont stages. The parasite in Figure 4D shows an apicoplast in the ramifying stage of its development and the correspondence of the anti-PfSHMTc fluorescence to the 'Y' shaped apicoplast is striking. In the 3D7 transfectant expressing PfACP with a DsRED tag, the development of the apicoplast did not exhibit the globular stage often seen in K1 parasites, with narrow ramifying apicoplasts being far more evident (Figure 8B and 8C ). However, the close coincidence of the apicoplast and PfSHMTc fluorescence was equally evident as when using K1 and two primary antibodies (see also below).
Quantitative analysis again supports the visual interpretation of the apicoplast data. The trophozoite shown in Figure 4A had a percentage material co-localization of PfSHMTc with anti-ACP of 5.9% and a percentage volume co-localization of 5.8%, indicating that the PfSHMTc fluorescence in this parasite was not appreciably higher within the apicoplast than without. The early schizont stage parasites in Figure 4B and 4C showed significantly higher percentages of PfSHMTc material co-localization of 10.1% and 22.3%, indicating that a considerable proportion of the PfSHMTc of these particular parasites was located within the comparatively small volume of the apicoplast. Moreover, the percentage material co-localized for PfSHMTc fluorescence in these two parasites was about one-third higher than the respective percentage volumes, reinforcing the visual impression that in these parasites PfSHMTc was at a higher concentration within the apicoplast than in the cytoplasm generally. A slightly later parasite (Figure 4D ), showing a ramifying apicoplast, displayed a somewhat lower level of co-localization of PfSHMTc with anti-ACP of 4.5% at a concentration that is again no higher than that of the surrounding cytoplasm.
A direct comparison between mitochondrial and apicoplast PfSHMTc concentrations was made in triple staining experiments. In this case, the limitations of wavelengths available precluded using a dye to simultaneously stain the DNA so that the precise stage of the parasites viewed was not clearly discernible; however, the size of the organelles and overall size of the parasites suggest that those shown in Figure 5A and 5B are mid-trophozoites. In these experiments, PfSHMTc was stained using Alexafluor anti-chicken IgY 488 nm (false coloured blue), which proved to be especially prone to bleaching and therefore unsuited to the repeated exposure to laser light necessary in building a z-stack scan. Unlike the other images presented here, therefore, those showing both the mitochondrion and the apicoplast are from single plane scans where both organelles were in the same z-axis plane. The parasite in Figure 5A shows apicoplast-specific fluorescence located within a discrete region of bright PfSHMTc fluorescence, whereas in contrast, the mitochondrion appears to have no associated PfSHMTc fluorescence. The parasite in Figure 5B also shows the apicoplast fluorescence within a region of high PfSHMTc fluorescence whilst the mitochondrion occupies a pocket of lower intensity PfSHMTc fluorescence. Quantitative analysis confirmed the much more substantial association of PfSHMTc with the apicoplast than with the mitochondrion. As these figures refer to pixels in a single plane rather than voxels in a three-dimensional projection from a z-stack scan, extrapolation to volumetric values was unsafe in this particular case.
Apicoplast localization of PfSHMTm
Use of anti-PfSHMTm in conjunction with anti-ACP showed that the PfSHMTm protein was also found within the apicoplast. The temporal distribution of PfSHMTm within the apicoplast through the erythrocytic cycle was qualitatively similar to that of PfSHMTc. Thus, there was no discernible co-localization seen in the early trophozoite (Figure 6A ), however, there was a marked presence of PfSHMTm fluorescence within the apicoplasts of both late trophozoites and mitotic schizonts (Figure 6B and 6C , Figure 7 ; see also Figure 10 ). The later, post-mitotic, schizonts showed a similar lowering of apicoplast-associated PfSHMTm fluorescence to that found using the PfSHMTc specific antibody (Figure 6D ). However, the spatial distribution of the PfSHMTm fluorescence within the elongating apicoplasts of early schizonts was, in contrast, dissimilar to that shown by PfSHMTc. Whereas the latter exhibited fluorescence relatively uniformly across the apicoplasts (Figure 4C and 4D ; Figure 8B and 8C ), PfSHMTm was distinctly concentrated in their extremities, and was notably absent, or in very much lower concentration, within the medial sections of these organelles (Figure 6C ). This phenomenon is further illustrated by the sequential z plane views (at 0.2 μm intervals) through the same parasite shown in Figure 7B , especially in the second panel of this sequence, which clearly shows concentration of PfSHMTm fluorescence at the tips, and the fourth and fifth panels, where the lower degree of staining of the medial regions relative to the tips is apparent. The percentage co-localization of anti-PfSHMTm material with anti-ACP fluorescence was indicative of a low level of apicoplast PfSHMTm concentration in the trophozoite, e.g. 1.0% for Figure 6A , followed by much higher apicoplast PfSHMTm concentrations in the mitotically active schizont: e.g. 11.3% for Figure 6B , 11.8% for Figure 7A and 36.3% for Figure 6C . In the later, post-mitotic, schizonts, levels of co-localization fell back to lower values, the parasite shown in Figure 6D having a percentage of PfSHMTm material co-localizing with anti-ACP of only 0.3%.
Imaging using an endogenously expressed apicoplast marker
The simultaneous use of two primary antibodies, even when raised in different species, combined with their respective fluorochrome-conjugated secondary antibodies, raised the formal possibility that any observed co-localization was the result of fortuitous interactions between those antibodies. To eliminate this possibility, 3D7 transfected parasites expressing the apicoplast-specific protein ACP fused to the DsRED reporter were employed [ 39 ]. When these parasites were probed with the single anti-PfSHMTc antibody, the images obtained showed an identical incidence of co-localization of the PfSHMTc fluorescence with the apicoplast (Figure 8A-C ) as was seen using the two antibody approach above, although the relatively low absolute brightness of the DsRED fluorescence made these images unsuited to quantitative evaluation. The conclusion from this result is that the images created using two primary antibodies are a true reflection of the sub-cellular distribution of the proteins investigated and that the same distribution is found in two independent lines of the parasite, K1 and 3D7.
The parasites expressing PfACP-DsRED were also simultaneously probed with antibodies to both PfSHMTc (IgY) and PfSHMTm (IgG), again employing single plane scans without a DNA-specific dye rather than z-stacks for this triple labelling experiment. The parasite in Figure 5C (estimated to be a mid to late trophozoite), and that in Figure 5D (an early schizont) both show overlapping, though not identical, PfSHMTc and PfSHMTm fluorescence distribution in the cytoplasm. Both parasites show PfSHMTc and PfSHMTm coincident with the apicoplast as indicated by white colouration in the relevant merged image. Quantitative image analysis reinforces the visual indication of co-localization of both PfSHMTc and PfSHMTm with each other, and with the apicoplast specific fluorescence. In particular the apicoplast specific fluorescence was almost entirely (between 82.8% and 99.7%,) co-localized with the signals from both isoforms of SHMT.
SHMT distribution in the post-mitotic schizont
In late, post-mitotic, schizonts, PfSHMTc fluorescence was characterized by a concentration in the central portion of the parasite. The peripheral regions of the parasite occupied by the nuclei and other constituents of the developing merozoites contained conspicuously lower levels of fluorescence, as shown in Figure 9A and 9B . The central area of late schizonts is the region that becomes the residual body upon completion of merozoite maturation and lysis of the erythrocyte, a prominent component of which is the pigment vacuole containing the crystalline haemozoin. In the very late schizont when the majority of the haemoglobin has been digested, the pigment vacuole occupies a large volume. Figure 9A and 9B show the central mass of dense haemozoin exhibiting no PfSHMTc staining but with marked PfSHMTc fluorescence in the region immediately surrounding it. To assess the relative frequency of this category of PfSHMTc distribution, 48 scans of post-mitotic schizonts were viewed, of which 15 showed a marked concentration of fluorescence in the centre of the schizont when compared to their periphery, an incidence of 31%. The use of anti-PfSHMTm antibody in conjunction with anti-PfSHMTc showed that PfSHMTm has a very similar concentration within the central region of the very late schizont (Figure 9C ). Additional to this general distribution pattern, the post-mitotic schizont contains numerous small apicoplasts, each associated with a developing merozoite. Despite the diminutive size of these organelles, the persistence of PfSHMTc fluorescence within the 'daughter' apicoplasts was still discernible in some images, e.g. in Figure 9B , where the lower right plastid in the parasite clearly shows its presence. The three-dimensional projection shown in Figure 9B is interesting as it shows a relatively late stage of daughter merozoite biogenesis.
Nuclear localization
In most parasites viewed there was a distinctly lower PfSHMTc fluorescence within nuclei than was found in the cytoplasm. However, PfSHMTc fluorescence was very rarely entirely excluded from the nucleus (see especially Figure 4C and 4D ; Figure 9B and 9C ). The level of nuclear relative to cytoplasmic fluorescence was variable with higher levels of intranuclear PfSHMTc fluorescence seen in some late trophozoites and mitotic schizonts. Nuclear PfSHMTc fluorescence rarely approached the intensity of cytoplasmic fluorescence, however. In contrast, PfSHMTm showed very little evidence of nuclear localization throughout the erythrocytic cycle, with most images showing an essentially complete exclusion of PfSHMTm fluorescence from nuclei (Figure 3D and 3E ; Figure 6A, C and 6D ).
Relative incidence of organellar SHMT fluorescence through the erythrocytic cycle
In view of the initially surprising results that PfSHMTc showed organellar co-localization patterns, a large number of z-axis scans of parasites were analysed in order to ascertain the relative incidence of organellar fluorescence for this isoform over the erythrocytic cycle. Parasites were assigned to one of four broadly defined developmental stages by examination of overall size, haemozoin development and nuclear morphology (Figure 10 ). These results emphasize the stage-specific dependence of organellar PfSHMTc fluorescence, which was undetectable in parasites up to and including the early trophozoite stages, visible from mid-trophozoites onward and peaking at the mitotic schizont stages. The corresponding analysis for PfSHMTm with respect to the mitochondrion is strikingly different in that 100% of parasites showed fluorescence in this organelle, regardless of the cell cycle stage. However, its incidence in the apicoplast was similar to that of PfSHMTc, in that it was not seen in the early trophozoite stage but peaked in the late trophozoite stage, although the percentage of parasites displaying this pattern was significantly higher than was the case for PfSHMTc.
GFP-tagging of SHMT via transfection
To support the immunofluorescence studies in a complementary manner, independent attempts were made in the two collaborating laboratories to produce transfected parasites expressing GFP-tagged, full length PfSHMTc and PfSHMTm endogenously, as well as shorter versions carrying a GFP-tag downstream of the first 100 amino acids of each protein (i.e. about one-quarter of their total length). Despite repeated transfections using several different protocols for these four constructs, viable parasites could only ever be recovered in the case of the truncated version of PfSHMTm + GFP. Fluorescence microscopy clearly located this hybrid protein in the mitochondrion (Figure 11 ), confirming the initial prediction based on sequence analysis that PfSHMTm carries a mitochondrial targeting signal at its N-terminus [ 24 ]. However, in contrast to the studies above using the anti-PfSHMTm antibody, no additional distribution in the cytoplasm or apicoplast was apparent, suggesting that localization to these areas was dependent upon properties of the full-length molecule.
Transcript analysis of PfSHMTc
In the original characterization of the gene encoding PfSHMTc, comparison of cDNA and genomic sequences, together with RACE analyses of the transcript start point in two independent laboratories [ 8 , 21 ], identified only a ~240 base 5' UTR on the mRNA which lacked any AUG motif upstream of the documented start codon. To confirm and extend this result, we carried out RT-PCR experiments using a range of internal primers based on genomic sequence extending up to 1 kb upstream of the start codon. However, no splice variants were detected (data not shown), nor could any putative splicing event using the normal GU and AG intron junction signals within this sequence create an alternative start codon. Thus there was no evidence that PfSHMTc might employ a conventional signal sequence that had previously been overlooked to gain access to the organellar compartments. | Discussion
SHMT is the principal agent by which one-carbon units are introduced onto folate carriers for subsequent essential transfer reactions, including the formation of thymidylate. Of the two SHMT isoforms expressed by P. falciparum , PfSHMTc is an enzymatically active member of the thymidylate cycle [ 21 - 23 ], whereas PfSHMTm is more enigmatic, as it lacks most of the conserved active site residues found in other SHMTs, whether cytoplasmic or organellar [ 24 , 42 ] and has been found to be inactive in studies of the recombinant protein [ 23 ]. With a combination of antibody probes and endogenous expression of tagged molecules, the cellular distribution of these two species across the parasite erythrocytic cycle has been investigated to gain possible insight into their biological function.
Although PfSHMTc lacks any obvious signal sequences and was expected to be confined to the cytoplasm, the most surprising result was that, although it is indeed in the cytoplasm that the majority population is found throughout erythrocytic development, PfSHMTc also localizes within parasite organelles, including the mitochondrion and particularly the apicoplast. Moreover, the distribution varies in a dynamic and developmental stage-dependent manner, consistent with a temporally mediated variation in the targeting of newly expressed enzyme protein between sub-cellular locations from a cytoplasmic pool. PfSHMTm, on the other hand, was identified as having an N-terminal mitochondrial targeting sequence [ 24 ], which was functionally confirmed here using a transfected parasite construct expressing GFP fused to its N-terminal domain. However, full-length PfSHMTm in its native state shows complex localization patterns that are similar to, but subtly different from those of PfSHMTc in the cytoplasm, mitochondrion and apicoplast.
Although the parasite preparations undergo a multi-step protocol for immunofluorescence that necessarily includes a mild detergent permeabilization, there are strong reasons to believe that the results obtained are not artifactual. Thus (i) organellar fine structures of the parasites are preserved, (ii) both primary anti-PfSHMT antibodies show a high degree of specificity, (iii) antibodies against two known cytoplasmically located enzymes show no organellar co-localization, (iv) conversely, the anti-ACP primary antibody locates exclusively to the apicoplast as expected, with no evidence of staining elsewhere that might indicate a loss of organellar integrity, (v) the incidence of PfSHMTc in the organelles shows a strong stage-dependency, being absent in early trophozoites, peaking in late trophozoites and mitotic schizonts, then diminishing in later (post-mitotic) schizonts, (vi) a different stage dependency for PfSHMTm is evident, particularly in the case of the mitochondrion, where co-localization is seen in all parasites throughout the cycle, and (vii) localization patterns of PfSHMTc within the apicoplast are the same, regardless of whether two primary antibodies (anti-PfSHMTc and anti-PfACP) are employed as probes or anti-PfSHMTc alone plus the endogenous apicoplast fluorescence from transfected parasites in which PfACP is tagged with DsRED. The use of z-stack scanning and quantitation of overlapping fluorescence signals considerably increased confidence in assigning position compared to conventional 2 D analysis of images in a single plane, especially in the case of late trophozoites and early schizonts, where organelles were large enough to permit at least 3 and up to 7 or 8 planes within the organelle to be examined, as exemplified in Figure 7 . Stages at which a specifically stained organelle was at its smallest unavoidably gave less clear-cut images that could be more easily compromised by adjacent cytoplasmic staining. Moreover the close apposition of the mitochondrion and apicoplast is also potentially problematic when both are very small or narrow, although this was less of an issue with the larger organelles of K1, primarily used for this study, compared to 3D7. However, attempts to increase resolution further by scanning z-planes every 0.1 μm instead of 0.2 μm resulted in unacceptable levels of photobleaching before such lengthy scans could be completed.
The observed association of PfSHMTc with the mitochondrion and the apicoplast is not of equal degree. The mitochondrion shows relatively low levels of PfSHMTc fluorescence in the earlier and later stages of the cycle, with signal within the organellar lumen only obvious in the late trophozoite and early schizont stages, when the mitochondrion is expanding and then elongating, with significant synthesis of internal constituents [ 45 ]. However, even at maximum visibility, the fluorescence intensity, as confirmed by quantitative image analysis, does not exceed that found in the surrounding cytoplasm. In contrast to the mitochondrion, the apicoplast shows a significantly higher level of PfSHMTc association over a longer period of the erythrocytic cycle. PfSHMTc fluorescence within the apicoplast lumen is first detected in trophozoite stages before the organelle has expanded noticeably and persists into the small daughter plastids of very late schizonts. In both of these developmental stages, the concentration of PfSHMTc within the apicoplast does not exceed that in the cytoplasm. However, between these stages, in the early schizont, when the apicoplast is expanding maximally and subsequently elongating, PfSHMTc fluorescence inside this organelle becomes very marked, such that the concentration of PfSHMTc within it now exceeds that in the surrounding cytoplasm. The percentage of total PfSHMTc fluorescence co-localising with apicoplast-specific fluorescence was measured in some parasites at this stage at >20%, a considerable proportion of the cellular total.
The PfSHMTm protein, despite its N-terminal mitochondrial targeting sequence, shows a similar spatial and temporal distribution to that of PfSHMTc, albeit with some important variation. Thus, the level of localization of PfSHMTm within the mitochondrion is distinctly higher than that of PfSHMTc. Moreover, the occurrence of concentrations of PfSHMTm within the extremities of elongating apicoplasts and a corresponding paucity in the medial sections is a polarization not seen with Pf SHMTc and may suggest a more specific developmental role for PfSHMTm in this organelle. Given also the occurrence of Pf SHMTm in the cytoplasm, there are likely to be further signals downstream of the N-terminal domain yet to be characterized that contribute to the complex and dynamic distribution patterns, as discussed further below.
The data suggest a connection between these apparently rather short-lived associations and internal organellar metabolism. Both organelles must replicate their own genomes, prior to which local demand for folate pathway products would be high. Moreover, there is evidence for the existence of a glycine cleavage complex (GCV) in the mitochondrion [ 24 ], which is dependent on the provision of folate cofactor. Studies in other systems indicate that substituted tetrahydrofolates cannot freely exchange between mitochondrial and cytoplasmic compartments [ 14 ], suggesting that transport of folate enzymes into membrane-bound organelles may be essential. In the very late post-mitotic schizont, PfSHMTc is found concentrated in the centre of the parasite, in the region that forms the residual body on erythrocyte lysis and merozoite release. This association can be rationalized in that, after mitosis, demand for DNA precursors is low, however, the late schizont is very active in protein production for organellogenesis and other aspects of merozoite maturation. As amino acids are released from haemoglobin, the late concentration of PfSHMTc in and around this organelle may be connected with an increased demand for the reversible Ser/Gly interconversion function of the SHMT enzyme and/or methionine metabolism.
Although plant cells also exhibit a partitioning of SHMT across cytoplasm, mitochondrion and plastid, this association is a persistent feature [ 46 ], rather than the more transient phenomenon seen here in the malaria parasite. By contrast, the parasite must undergo rapid asexual reproduction at the blood stages, requiring parasite metabolism to be highly efficient in its production and use of folate pathway components. The locations of greatest demand for such products would thus vary through the processes of growth, repeated mitoses and cytodifferentiation in the erythrocytic cycle. It would, therefore, be advantageous to be able to translocate folate enzymes to varying subcellular locations as the demand for folates changed throughout development. The results of this study support a view of the cellular location of folate pathway enzymes being dynamic and responsive to the changing needs of the parasite over time.
Important questions that now need to be addressed are how PfSHMTm and, more puzzling, PfSHMTc, are targeted to the organelles investigated here and their precise function(s) therein. Dual targeting of proteins to the mitochondrion and plastid is frequently observed in other systems [ 47 ]. However, in organisms as diverse as plants, yeast and mammals, dedicated cytoplasmic and organellar isoforms of SHMT are employed to effect compartmentalized folate metabolism. Only PfSHMTm has a recognizable mitochondrial targeting sequence and neither isoform possesses the conventional bipartite topogenic signal associated with plastid targeting via the endoplasmic reticulum in P. falciparum and the closely related Toxoplasma gondii [ 48 , 49 ]. Moreover, although a proportion of PfSHMTm is located in the mitochondrion throughout the erythrocytic cycle and also in the apicoplast in the middle to late stages, how can this isoform perform the necessary enzymic steps when its sequence (see Additional file 2 Sequence alignments of the PfSHMT isoforms) and the inactivity of the recombinant protein [ 23 ] strongly indicate that it cannot? It could be that the folate metabolism essential for the replication of their genomes is provided by the enzymatically competent PfSHMTc after gaining access to these organelles. The sequence and transcript analyses exclude the possibility of a hitherto unidentified upstream leader sequence that could splice onto the ORF as currently defined, but cryptic motifs further downstream in the encoded protein cannot be excluded, especially as numerous plastid proteins in other systems depend upon ill-defined, sometimes non-contiguous regions of the molecule [ 50 ]. However, another intriguing possibility is suggested by a closer analysis of the primary sequences of PfSHMTc and PfSHMTm, which reveals that each carries one or the other of two internal conserved sequence motifs (residues 138-146 in PfSHMTc and residues 262-277 in PfSHMTm), both of which are present in the SHMTs of higher organisms and are known to mediate intersubunit interactions (Figure 12 ). Thus, mammalian SHMTs are stable homotetramers, whereas bacterial SHMTs, which lack these motifs entirely, are homodimers [ 42 ]. The plasmodial proteins seem to represent a complementary pairing system thus far unique among SHMT types and this leads us to the hypothesis that while PfSHMTm is in itself apparently catalytically inactive, a dimer thereof might be able to form a stable but readily reversible heterotetramer with dimeric PfSHMTc, generating a complex in which the requisite SHMT activity is provided (from PfSHMTc) together with (an) organellar targeting sequence(s) (from PfSHMTm) that could be modulated as necessary. Such an arrangement would have parallels with the S-adenosyl-L-methionine decarboxylase (AdoMetDC) system of trypanosomes, where a catalytically inactive paralogue of AdoMetDC forms a heterodimer with AdoMetDC itself and thereby regulates the activity of the latter allosterically [ 51 ]. This scenario could explain why the localizations of the two PfSHMT isoforms show considerable overlap, and why our GFP construct attached to the first 100 amino acids of PfSHMTm, and thus lacking both of the above motifs, migrates solely to the mitochondrion. Moreover, many nuclear-encoded proteins destined for plastids and mitochondria in other organisms are translated on cytoplasmic ribosomes and imported (often long) after their synthesis [ 47 ], thus providing a credible precedent for encounters between PfSHMTc and PfSHMTm. However, it is certainly unclear at this point precisely how such a complex could be successfully translocated across the requisite membranes, although analyses of organellar protein transport found in other systems serve to emphasize the considerable diversity of mechanisms associated with this phenomenon [ 50 , 52 ]. | Conclusions
The two isoforms of SHMT in P. falciparum , PfSHMTc and PfSHMTm, exhibit complex distribution patterns across the cytoplasm and organelles of the parasite that are similar but differ in their levels of occupancy and cell-cycle stage dependency. PfSHMTm is confirmed as having an N-terminal mitochondrial targeting sequence whereas PfSHMTc lacks any obvious targeting signals. Interactions between the two isoforms suggested by sequence analysis may be involved in the dynamic patterns of localization observed and may be important in overcoming the apparent lack of catalytic competence of the PfSHMTm isoform. Further studies are required to establish whether such an association can occur and play a role in ensuring the provision of essential folate cofactors for replication of the nuclear, mitochondrial and apicoplast genomes. | Background
The folate pathway enzyme serine hydroxymethyltransferase (SHMT) converts serine to glycine and 5,10-methylenetetrahydrofolate and is essential for the acquisition of one-carbon units for subsequent transfer reactions. 5,10-methylenetetrahydrofolate is used by thymidylate synthase to convert dUMP to dTMP for DNA synthesis. In Plasmodium falciparum an enzymatically functional SHMT (PfSHMTc) and a related, apparently inactive isoform (PfSHMTm) are found, encoded by different genes. Here, patterns of localization of the two isoforms during the parasite erythrocytic cycle are investigated.
Methods
Polyclonal antibodies were raised to PfSHMTc and PfSHMTm, and, together with specific markers for the mitochondrion and apicoplast, were employed in quantitative confocal fluorescence microscopy of blood-stage parasites.
Results
As well as the expected cytoplasmic occupancy of PfSHMTc during all stages, localization into the mitochondrion and apicoplast occurred in a stage-specific manner. Although early trophozoites lacked visible organellar PfSHMTc, a significant percentage of parasites showed such fluorescence during the mid-to-late trophozoite and schizont stages. In the case of the mitochondrion, the majority of parasites in these stages at any given time showed no marked PfSHMTc fluorescence, suggesting that its occupancy of this organelle is of limited duration. PfSHMTm showed a distinctly more pronounced mitochondrial location through most of the erythrocytic cycle and GFP-tagging of its N-terminal region confirmed the predicted presence of a mitochondrial signal sequence. Within the apicoplast, a majority of mitotic schizonts showed a marked concentration of PfSHMTc, whose localization in this organelle was less restricted than for the mitochondrion and persisted from the late trophozoite to the post-mitotic stages. PfSHMTm showed a broadly similar distribution across the cycle, but with a distinctive punctate accumulation towards the ends of elongating apicoplasts. In very late post-mitotic schizonts, both PfSHMTc and PfSHMTm were concentrated in the central region of the parasite that becomes the residual body on erythrocyte lysis and merozoite release.
Conclusions
Both PfSHMTc and PfSHMTm show dynamic, stage-dependent localization among the different compartments of the parasite and sequence analysis suggests they may also reversibly associate with each other, a factor that may be critical to folate cofactor function, given the apparent lack of enzymic activity of PfSHMTm. | Competing interests
The authors declare that they have no competing interests.
Authors' contributions
MR developed the immunofluorescence protocols, carried out most of the experimental work in Manchester and drafted the manuscript. IBM constructed GFP transfection plasmids, produced parasite transfectants and carried out the imaging thereof. SLM contributed to the establishment of the immunofluorescence procedures. PFGS and JEH conceived the study, wrote parts of the manuscript and prepared the final version. All authors were involved in data interpretation and analysis and have approved the final manuscript.
Supplementary Material | Acknowledgements
We thank Chi Tang and Apolinar Maya-Mendoza (University of Manchester) for assistance with the microscopy, Geoffrey McFadden (University of Melbourne) for providing the anti-ACP antibody, Teresa Fitzpatrick (University of Zürich) and Barbara Kappes (University of Heidelberg) for the anti-CS and anti-PK5 antibodies, Shigeharu Sato (NIMR, London) for the 3D7 parasite transfectant strain carrying pSSPF2/PfACP-DsRED and Ping Wang (University of Manchester) for contributions to the transfection experiments. We are also grateful to the Wellcome Trust, UK (grant no. 073896) and BBSRC, UK (studentship for S.L.M), for financial support. | CC BY | no | 2022-01-12 15:21:37 | Malar J. 2010 Dec 3; 9:351 | oa_package/4d/01/PMC3014972.tar.gz |
PMC3014973 | 21159170 | Background
In spite of the huge global morbidity and mortality inflicted by malaria, an effective and practical vaccine against this disease has not yet been achieved. Early studies on immunization against sporozoite-induced rodent malaria resulted in close to 100% protection when mice were immunized with Plasmodium berghei sporozoites irradiated to sufficient levels and the immunized mice were subsequently challenged with non-irradiated sporozoites [ 1 , 2 ].
Nevertheless, immunization with attenuated P. berghei sporozoites via routes other than intravenous (IV) was found to be far less protective. Thus, even after five immunizations, mice immunized by intramuscular (IM), intraperitoneal (IP) or intradermal (ID) routes were protected only 32%, 26% and 24%, respectively, in contrast to 95% protection after IV immunization [ 3 ]. In another P. berghei study done with similar protocols, IM immunization resulted in only 11% protection, although addition of albumin to the immunizing inoculum raised this to 42%; in contrast, IV immunization yielded 100% protection [ 4 ].
Because sporozoite suspensions used for immunization were heavily contaminated with microorganisms and mosquito components, it was clear that such immunization trials by IV injection could not be directly extended to humans. An alternate approach, however, allowed irradiated mosquitoes to directly inoculate attenuated sporozoites into hosts, the mosquitoes thereby acting as vehicles of immunization. This approach was first established with rodent malaria [ 5 ] and then extended to the first successful human vaccination trial against P falciparum malaria [ 6 ]. A compendium of subsequent human vaccination trials with this approach showed that when sufficient numbers of mosquitoes were used for immunization, greater than 90% of volunteers were completely protected against challenge by bite of infected mosquitoes [ 7 , 8 ]. Recent progress by this group, under the auspices of the biopharmaceutical company Sanaria, has permitted the raising of large numbers of mosquitoes infected with Plasmodium falciparum sporozoites, the purification of these sporozoites sufficient to render them acceptable for human vaccination, and the successful freeze-preservation of the attenuated sporozoites. Trials are currently underway to attempt to vaccinate humans by syringe injection of these sporozoites [ 9 , 10 ].
A central question for any human trials relates to an appropriate route of immunization. It had long been assumed that most sporozoites injected by mosquitoes rapidly reach the blood, after which they travel to the liver for further development. Thus, there was a supposition that sporozoite inoculation by mosquitoes mimicked IV inoculation of sporozoites by syringe. But studies have shown that most if not all mosquito-injected sporozoites are deposited into avascular portions of the skin and sub-cutaneous tissues and that sporozoites then use gliding motility to reach blood vessels to travel to the liver [ 11 , 12 ], or enter lymph vessels to travel to local draining lymph nodes [ 12 ]. This has led to the possibility that inoculation of isolated sporozoites directly into the skin by syringe might successfully replicate the recognized successful approach of allowing mosquitoes to inoculate attenuated sporozoites into skin.
Accordingly, the Plasmodium yoelii rodent malaria system was used to explore this approach and investigate ways of further enhancing the protective immunogenicity of ID-injected, attenuated sporozoites. The P. yoelii system is far superior to the P. berghei system in the infectivity of sporozoites, appearing to be similar in that respect to the human malarias. Indeed, others have contended that "the P. yoelii system has accurately predicted the success or failure of every approach to malaria vaccination that has been tested in humans" [ 13 ]. A recent publication reported that it took at least four ID immunizations with a total of 6,000 P. yoelii sporozoites to equal the protective immunity that could be accomplished with only three doses and a total of only 2,250 sporozoites administered IV [ 14 ]. Because of the practical value of achieving protective immunity with fewer doses of sporozoites, the current study tested whether a substantially higher number of sporozoites given in only two doses ID might result in protection equivalent to that obtained with IV immunization.
The main goal was to assess the protective immunogenicity of attenuated P. yoelii sporozoites injected by syringe into the skin. ID immunization with P. yoelii sporozoites was found to give far better protection than had been observed in previous attempts using non-IV routes of administration to immunize with P. berghei sporozoites [ 3 , 4 ]. Furthermore, administering larger numbers of P. yoelii sporozoites ID was able to give a degree of protective immunity equivalent to what had been reported for ID immunization with twice the number of immunizing doses [ 14 ].
Finally, because many sporozoites injected by mosquitoes remain in the skin and either deteriorate [ 11 ] or differentiate [ 15 ] and because such sporozoites may be involved in induction of the immune response [ 16 ], two methods were tested to possibly enhance the immune response with adjuvants in conjunction with ID-administration of sporozoites; a) topical application of the toll-like receptor (TLR) agonist Imiquimod, as previously done for immunization studies using sub-unit vaccines against malaria [ 17 ], b) "tape-stripping" (TS), a procedure known to disrupt the stratum corneum and enhance the immunogenicity of ID-injected antigens [ 18 ]. All ID-injection procedures were found to lead to levels of protection not significantly different from that which had been observed after IV immunization. | Methods
Host-parasite system
Plasmodium yoelii sporozoites (strain 17XNL) were produced in Anopheles stephensi mosquitoes. Standard protocols were used for infecting and maintaining mosquitoes [ 19 ], which were infected by feeding upon gametocyte-carrying 6-8 wk-old Swiss-Webster mice (Taconic Farms Inc., Germantown, NY). Protocols for maintenance and use of experimental animals were approved by the Institutional Animal Care and Use Committee at New York University School of Medicine, whose animal facility is accredited by the Association for Assessment and Accreditation of Laboratory Animal Care International (Rockville, MD).
Isolation of sporozoites for immunization
Sporozoites for immunization were isolated from mosquitoes 18 days after the mosquitoes had received an infective blood meal. To obtain sporozoites, mosquitoes were anesthetized on ice, then washed with 70% ethanol, followed by RPMI medium 1640 (Gibco BRL, Grand Island, NY). Salivary glands were dissected out and triturated in RPMI medium supplemented with 2% mouse serum albumin, after which the freed sporozoites were counted in a haemacytometer and diluted to appropriate concentrations for immunization.
Immunization and challenge protocols
Sporozoites were irradiated within a gamma irradiator (MDS Nordion Gammacell® 1000 Elite) to a central dose of ± 12,049 cGy and a minimum dose of ± 10,266 cGy. Mice (BALB/c females from Taconic Farms Inc., Germantown, NY) were 6 wk old at the initiation of each experiment (n = 6/group for each experiment). For immunization, they received an initial injection of 60,000 irradiated sporozoites, with a subsequent booster injection of 30,000 irradiated sporozoites 15 days afterwards. Intravenous (IV) injections into a tail vein were given in a volume of 200 μl per mouse. For intradermal (ID) immunizations, a portion of flank skin was shaved and injections were given with a Nanofil TM Sub-Microliter syringe with a 33G flexifil beveled tip (World Precision Instruments, Sarasota, Florida, USA) in two adjacent sites in a volume of 5 μl per site.
Parallel ID immunizations were given to groups of mice that received topical treatment with an adjuvant or an adjuvant procedure. For topical treatment with Imiquimod (obtained as a 5% cream containing 12.5 mg/ml of Imiquimod [Aldara; 3 M, St. Paul, MN]), anesthetized mice were treated with approximately 25 μl (1.25 mg of imiquimod) [ 17 ]. It was applied by rubbing into the shaved skin area 4 h after ID immunization, then reapplied 1 and 2 days after immunization. This process was repeated for the booster immunization.
For assessment of the adjuvant effect of TS, a small area of the shaved flank skin of anesthetized mice was treated by applying and pulling off a strip of adhesive tape (Scotch Brand 3 M Magic tape) 10 times. A fresh piece of tape was used for each of the 10 strippings. This procedure is known to disrupt the stratum corneum [ 18 ], and resulted in local erythema and inflammation typical of what has been demonstrated in previous studies with tape-stripped mice [ 18 ]. Tape-stripping was done 30 min after ID immunization with sporozoites and was repeated for the booster immunization.
Mice were anesthetized and then challenged by mosquito bite 15 days after the second immunization. Parallel challenges were done on non-immunized control mice from the same cohort. For each mouse, five different infected mosquitoes from the same cohort were allowed to probe and feed for 5 min. Thin blood smears were prepared from drops of tail blood up to 14 days after challenge and stained with Giemsa.
Statistics
The percentages of mice that developed parasitemia and the pre-patent periods of those that developed parasitemia after challenge by mosquito bite were recorded. Comparisons were made between the percentages of mice infected after the various immunization regimens (IV, ID, ID+imiquimod, ID + TS) versus the percentage of non-immunized controls that were infected. To do these comparisons, all data were transformed using the following equation y = arcsin[√(y/100)], where y represents percentage of infection. Gaussian distribution of transformed data was then confirmed using the Kolmogorov-Smirnov normality test. A one-way ANOVA followed by Tukey's post hoc test was then used to compare the differences between groups of mice. The analyses were performed using GraphPad Prism Version 5 software (San Diego, California). | Results
Challenges by mosquito injection of sporozoites
The results of three independent experiments are shown in Table 1 . Mice immunized by two IV injections of sporozoites and challenged by mosquito bite had an overall protection of 94%, compared with non-immunized controls. Mice immunized by ID injections had an overall protection of 78%. When attempts were made to enhance ID immunizations by adjuvant treatments, treatment with Imiquimod resulted in 67% protection, while treatment by TS gave 94% protection. In spite of the fact that the percentage of mice protected after ID immunization with TS was the same as the percentage protected after IV immunization, it was not possible to demonstrate that these were significantly better than protection observed after ID immunization alone (P > 0.05). | Discussion
A proof of concept study done more than 35 years ago showed the possibility of immunizing humans against malaria by irradiating infected mosquitoes and allowing them to inject attenuated P. falciparum sporozoites into volunteers [ 6 ]. In subsequent human trials with sporozoite vaccination via mosquito injection, greater than 90% of volunteers were protected against challenge by bite of irradiated, infected mosquitoes [ 7 , 8 ]. Nevertheless, the impracticality of immunization by mosquito bite was obvious. A more recently proposed approach is based on syringe injection of purified and freeze-preserved, attenuated P. falciparum sporozoites isolated from mosquito salivary glands [ 9 , 10 ].
However, early studies with P. berghei sporozoites suggested that immunization by syringe injection of sporozoite suspensions was not very effective unless the sporozoite suspensions were administered IV. As this is not a feasible mode of administration for large scale immunization of humans, more suitable modes of syringe injection of sporozoites were revisited with another species of rodent malaria, P. yoelii . The rationale for this is that far fewer P. yoelii than P. berghei sporozoites are required, either to infect [ 20 , 21 ] or to protectively immunize [ 22 ] mice.
Results for three experiments in the present study showed an overall 94% protection for mice immunized IV vs. 78% for ID immunization after only two immunizations. An initial pilot experiment using three immunizations for this study had yielded 100% protection after IV immunization, compared to 83% protection with ID immunization (data not shown). It has been reported that two IV doses of irradiated P. yoelii sporozoites rarely provides solid protection, while three IV doses does so consistently [ 22 , 23 ]. Thus, to better discriminate between the effects of various types of immunization protocols with and without adjuvant treatments, subsequent immunization experiments were performed with only two immunizations, as presented in this paper. It was anticipated that such suboptimal immunization might better differentiate the effects of adjuvant treatments. Furthermore, it has been noted that a goal for eventual human vaccination is to achieve high-level protection with the fewest numbers of doses of vaccine [ 10 ].
When adjuvant treatments in conjunction with the ID immunizations were evaluated, it was found that topical application of Imiquimod resulted in an overall protection rate of 67%, whereas TS gave 94% protection. IV immunization is the generally accepted "gold standard" for protection achieved with attenuated sporozoites. Although TS in conjunction with ID immunization yielded the same degree of protection as IV-immunization, it was not possible to show statistically that the protection induced by ID immunization together with TS was significantly better than ID immunization alone. Imiquimod is a TLR agonist that stimulates local immunocytes, including dermal dendritic cells, when topically applied [ 24 ]. It has been successfully used as an adjuvant to increase protective immunity against Leishmania major in BALB/c mice [ 25 ] and to enhance immunogenicity of a peptide-based vaccine against Plasmodium sporozoites [ 17 ]. However, topically applied Imiquimod showed no enhancing effects when used in association with ID-injection of radiation attenuated sporozoites.
Tape stripping together with ID-injection of sporozoites, on the other hand, protected the same percentage of mice as observed after IV immunization. Tape stripping consists of partially removing the stratum corneum, the outermost layer of the epidermis, by use of adhesive strips applied to the skin and then removed [ 18 ]. This procedure had been shown to enhance the immune response to an ID-injected protein antigen in the tape-stripped area [ 18 ] and [ 26 ]. It is known to result in local inflammation, with movement of antigen-presenting cells into affected regions and enhancement of immune surveillance [ 26 ], as well as induction of TLR-9 mRNA [ 27 ]. The rationale for using TS was based on previous observations that mosquitoes inject sporozoites into avascular portions of skin and subcutaneous tissues and that many sporozoites remain in the skin and become fragmented within several hours [ 11 ] and [ 28 ]. Furthermore, it has been shown that an anti-sporozoite response originates early in lymphoid tissues linked to cutaneous infection sites and it was suggested that this originates with internalization of sporozoite antigen by immature dendritic cells in the skin [ 16 ]. The current study hypothesized that the effects of TS might enhance interactions between immunocytes and sporozoites deposited in the skin.
The 78% protection reported here after simple ID administration of radiation-attenuated P. yoelii sporozoites with only two immunizations is substantially higher than any non-IV route of immunization previously reported with P. berghei sporozoites. A prior study by others had also reported good protective results after non-IV immunization with attenuated P. yoelii sporozoites [ 14 ]. Nevertheless, that study had reported that four doses of P. yoelii sporozoites given ID or SC were required to achieve protection similar to what was achieved with three doses of fewer sporozoites administered IV. Those results, as well as the ones reported here, are encouraging because it has been argued that "experiments using attenuated P. yoelii sporozoite vaccines have consistently paralleled or predicted results in humans exposed to P. falciparum sporozoites" [ 22 ]. Because of the practical operational value of achieving protective immunity with as few doses of sporozoites as possible, the present study tested whether only two ID doses given with a relatively high number of sporozoites might result in protection comparable to that obtained with IV immunization. The current study now shows that this is the case with P. yoelii malaria. When two immunizing doses with large numbers of sporozoites were used, there were no significant differences between results obtained with IV-versus ID-immunization. Indeed, ID-immunization accompanied by TS gave the same degree of protection obtained with the IV immunization "gold standard. These results are in accord with a retrospective analysis showing that immunization of humans with P. falciparum sporozoites injected by irradiated mosquitoes was most effective when large numbers of immunizing mosquito bites were used to introduce larger numbers of attenuated sporozoites into the vaccinees. Immunization with greater than 1,000 such bites led to total protection in 33 of 35 challenges, with protection lasting up to 10.5 months [ 9 ]. It remains to be determined whether an adjuvant treatment in conjunction with ID-immunization by syringe can be found to substantially reduce the numbers of attenuated sporozoites required to achieve a strong protective immunity with as few immunizing doses as possible. | Conclusion
ID-immunization with large numbers of radiation-attenuated P. yoelii sporozoites with as few as two immunizing doses leads to levels of protective immunity comparable to those achieved by IV-immunization. | Background
Intravenous injection of mice with attenuated Plasmodium berghei sporozoites induces sterile immunity to challenge with viable sporozoites. Non-intravenous routes have been reported to yield poor immunity. Because intravenous immunization has been considered to be unacceptable for large scale vaccination of humans, assessment was made of the results of intradermal immunization of mice with Plasmodium yoelii , a rodent malaria parasite whose infectivity resembles that of human malaria.
Methods
Mice were immunized with two injections of isolated, radiation-attenuated P. yoelii sporozoites, either by intravenous (IV) or intradermal (ID) inoculation. In an attempt to enhance protective immunogenicity of ID-injections, one group of experimental mice received topical application of an adjuvant, Imiquimod, while another group had their injections accompanied by local "tape-stripping" of the skin, a procedure known to disrupt the stratum corneum and activate local immunocytes. Challenge of immunized and non-immunized control mice was by bite of sporozoite-infected mosquitoes. Degree of protection among the various groups of mice was determined by microscopic examination of stained blood smears. Statistical significance of protection was determined by a one-way ANOVA followed by Tukey's post hoc test.
Results
Two intravenous immunizations produced 94% protection to mosquito bite challenge; intradermal immunization produced 78% protection, while intradermal immunization accompanied by "tape-stripping" produced 94% protection. There were no statistically significant differences in degree of protective immunity between immunizations done by intravenous versus intradermal injection.
Conclusions
The use of a sub-microlitre syringe for intradermal injections yielded excellent protective immunity. ID-immunization with large numbers of radiation-attenuated P. yoelii sporozoites led to levels of protective immunity comparable to those achieved by IV-immunization. It remains to be determined whether an adjuvant treatment can be found to substantially reduce the numbers of attenuated sporozoites required to achieve a strong protective immunity with as few doses as possible for possible extension to immunization of humans. | Competing interests
The authors declare that they have no competing interests.
Authors' contributions
All authors shared in experimental design, conduct of experiments, collection of data and experimental analysis. All authors helped to finalize the manuscript and approved the final manuscript. | Acknowledgements
We thank Adela Nacer for help with statistics. This study was supported by Public Health Service grant # AI63530 from the NIH Institute of Allergy and Infectious Diseases to J.V. | CC BY | no | 2022-01-12 15:21:38 | Malar J. 2010 Dec 15; 9:362 | oa_package/11/95/PMC3014973.tar.gz |
PMC3014974 | 21245918 | Introduction
Despite a declining incidence in developed countries, acute measles infection is still responsible for an estimated 164,000 deaths/year and is therefore a major vaccine-preventable cause of death worldwide [1] , [2] . Subacute sclerosing panencephalitis (SSPE) is a rare but usually fatal late complication which presents 3–10 years after the acute infection. SSPE is a distinctive clinical entity characterized by behavioural changes and myoclonic jerks, followed by motor dysfunction and profound global cognitive impairment, and then death within a few years of presentation in most cases. The diagnosis is made by the presence of characteristic clinical signs and, if available, electroencephalographic (EEG) findings in conjunction with elevated measles-specific antibodies in serum and cerebrospinal fluid (CSF) [3] .
The incidence of SSPE in most countries is <5 per million population <20 years of age, although this figure can be higher in the developing world where vaccination programs are not fully established [4] . The first reports of an unusually high incidence in Papua New Guinea (PNG) were published in the early 1990's [5] , with rates between 1988 and 1999 that varied from 13 [5] to 98 [6] per million population <20 years of age. However, these data need to be interpreted against fluctuations in the incidence of acute measles infection over the preceding decade, and should take into account background vaccination coverage and the possibility that localized clusters may contribute disproportionately to overall incidence rates estimated at provincial or country level. In addition, published PNG data to date have come from highland areas which may not be representative of the country as a whole.
We report a series of children presenting to a coastal PNG provincial referral hospital with clinical and laboratory features typical of SSPE. Using available local demographic data, as well as retrospective vaccination and disease surveillance, we have estimated the annual incidence of SSPE in Madang Province and interpreted this figure in relation to prior national measles vaccination coverage and acute measles incidence, as well as the regional distribution of cases. | Methods
Ethics statement
Approval for the study was provided by the PNG Institute of Medical Research Institutional Review Board and the Medical Research Advisory Committee of the PNG Health Department. Written informed consent for participation was obtained from parent(s)/guardian(s). The risks and benefits of lumbar puncture (LP) were explained to parent(s)/guardian(s) by the attending ward pediatrician who carried out the procedure with regard for conventional indications (suspicion of meningitis, subarachnoid hemorrhage or central nervous system disease) and contraindications (such as increased intracranial pressure or coagulopathy) [7] .
Study site and patients
Madang Province on the North Coast of PNG has an estimated population of approximately 450,000 people, 54% of whom are <20 years old [8] . Modilon Hospital is the provincial referral hospital and the only health care facility in the province that offers diagnostic and treatment facilities for severely ill patients. A longitudinal detailed observational study of severe illness in all children aged 6 months to 10 years was started at Modilon Hospital at the end of 2006. Prior to this initiative, documentation of cases was insufficient to allow epidemiologic analyses of specific diseases. In November 2007, the first child with symptoms and signs of SSPE was admitted to the present study. There was a subsequent increase in the numbers of similar cases before a decline after 12 months. Data collection was continued until July 2009, at which time relatively few such cases were being admitted.
National and local measles epidemiology
Measles immunization was started in PNG in 1982. A modified two-dose schedule at six and nine months of age was used with the aim of providing partial coverage for young infants at high risk of pneumonia and SSPE [9] . However, subsequent available national data indicate that coverage has remained low (see Figure 1 ). In a recent study of 2007 data, for example, 58% of eligible children received the first dose and 47% the second dose [10] . Cyclical measles epidemics have continued to occur, the last in 2002 (see Figure 2 ) [11] , [12] , [13] . Supplementary immunisation activities (SIA) for children aged 6 months to 7 years have been deployed since 2004, with a reported coverage of 79% in 2008 [13] .
The measles vaccine coverage recorded in the health diaries of children in Madang Province is similar to that reported elsewhere in PNG, with 41% of children <10 years of age surveyed at two sites within a 20 km radius of Madang town between September 2007 and June 2008 having received at least one dose [10] . Nevertheless, an increasing seroprevalence with age (60% and 79% for children 1–4 years and 5–9 years old, respectively) may indicate that wild measles virus remains prevalent in the community [10] and that there is under-reporting of cases as found in other epidemiologic settings [14] , [15] .
Clinical assessment
After recruitment, a standardized case report form was completed detailing demographic information, medical history and history of the current illness. Vaccination history was identified from the health record book held by the parent(s)/guardian(s) of each child where this was available. Since there is no local or central vaccination register, it was assumed that children without such documentation were unvaccinated. Standardized physical assessment included nutritional status assessed by calculating a weight-for-height Z-score [16] , with a value <2 considered to indicate malnutrition. We defined severe illness as the presence of one or more of the following features: i) impaired consciousness or coma (Blantyre Coma Score (BCS) <5 [17] ), ii) prostration (inability to sit or stand unaided), iii) multiple seizures, iv) hyperlactatemia (blood lactate >5 mmol/L), v) severe anemia (hemoglobin <50 g/L), vi) dark urine, vii) hypoglycemia (blood glucose <2.2 mmol/L), viii) jaundice, or xi) respiratory distress. These criteria are consistent with the World Health Organisation definition for severe malaria [18] .
Children with clinical evidence of SSPE, including myoclonic jerks, behavioural changes, and/or speech and motor deficits, underwent detailed neurologic examination by study clinicians (LM, ML). Level of consciousness was graded according to Blantyre Coma Score [17] . Upper motor neuron signs were considered to be present if the child had i) extensor plantar responses, ii) increased muscle tone of either upper or lower limbs, iii) sustained clonus, iv) hyperreflexia, and/or v) pyramidal tract muscle weakness of either upper or lower limbs. In children whose parents/guardians provided informed consent and who had no contraindications, LP was performed. All children were examined daily until discharge at which time a basic assessment of performance status was made. Moderate disability was defined as that requiring considerable assistance with self-care and severe disability as that requiring special assistance with all self-care, categories that are consistent with Karnovsky's performance scores of 50% and <50%, respectively [19] .
Laboratory tests
CSF was examined macroscopically for turbidity, blood staining and clots. We used the Neubauer Improved counting chamber (BoeCo, Germany) to obtain total and differential CSF white cell counts (WCC). Semi-quantitative measures of CSF glucose and protein were performed using dipsticks (Acon Laboratories, San Diego, USA). Specific measles IgG in CSF and serum was measured using a standard indirect immunofluorescence antibody assay (IFA). Serial two-fold dilutions of patient samples were added to separate wells of glass slides to which were fixed measles virus-infected Vero cells. After incubation and washing, anti-human IgG fluorescein isothiocyanate conjugate was then added and, following further incubation and washing, slides were examined under an ultra-violet microscope. Fluorescence was scored as 1+ to 4+, with levels of ≥1+ regarded as positive. This test was performed in an accredited laboratory and had been assessed and approved by the Australian National Association of Testing Authorities in accordance with requirements of the Australian National Pathology Accreditation Advisory Council. Details of other laboratory tests including malaria microscopy, plasma biochemistry and bacterial culture have been published elsewhere [20] .
Case definition of SSPE
Confirmed SSPE was defined as clinical features of SSPE and the presence of measles-specific IgG in CSF, regardless of titer. Probable SSPE was defined as clinical features of SSPE and negative measles-specific IgG in CSF or when no LP was performed.
Data analysis
The calculation of SSPE incidence was based on PNG Census data for the year 2000 [8] which includes population structure at provincial, district and local-level government (LLG, sub-district) level. The 2008 population was estimated by applying an annual growth rate of 2.6% (Dr Bryant Allen, Australian National University, Canberra, Australia; personal communication). Using this approach, the total population for Madang Province was estimated to be 448,330 with 241,165 (53.8%) <20 years of age. The Global Positioning System co-ordinates of each child's home village were obtained to facilitate LLG incidence estimates [21] . All SSPE incidence rates were expressed per million population <20 years of age which ranged from 5,545 in Iabu Rural LLG to 28,066 in Amenob Rural LLG with an inter-quartile range of 9,880 to 21,267. Reported annual rates of measles vaccination coverage [11] , [12] and cases of acute measles infection reported to the PNG Department of Health [12] , [13] were obtained from World Health Organization sources. Statistical testing was by means of parametric or non-parametric tests using PASW Statistics (version 17; SPSS Inc. Chicago, Ill) and a level of significance of 0.05. | Results
Presenting features and clinical course
Baseline, clinical and laboratory data relating to cases of SSPE identified during the 19-month surveillance period are summarized in Table 1 . These 22 children (16 confirmed and 6 probable cases; see below) were a subset of 671 admitted with severe illness during the study period. Although the median duration of illness prior to admission reported by the parent(s)/guardian(s) was 60 (range 1 to 1,000) days, the data provided were not sufficient to allow an accurate estimate of the age of each child at symptom onset.
Two children had documentation or parental knowledge of a past history of acute measles infection, one at six months and the other at two years of age. Neither had a documented history of measles vaccination. There were 14 (64%) children in whom the first dose of measles vaccine had been given and all but one of these (59%) had subsequently received the second dose. In a contemporaneous sample of 44 children hospitalized with other severe non-SSPE illness matched 2∶1 by age and sex with the SSPE cases, the equivalent percentages were 67% and 67% respectively ( P >0.55 by Chi-squared test). Two children diagnosed with SSPE within a few months of each other were first cousins.
Sixteen children had characteristic myoclonic jerks on admission and four had a clear prior history of myoclonic jerks obtained from the child's parents. One child presented with a short (two-week) history of severe involuntary muscle spasms and died soon after admission, while another presented with complex involuntary dyskinetic movements of upper and lower limbs. The majority of children had additional neurologic findings such as impaired consciousness, difficulty walking and impairment of speech.
LP was performed in 18 of the 22 children. Sixteen of these (89%), including the two with atypical non-myoclonic features, had high titre measles-specific antibodies in both serum and CSF and were therefore confirmed cases of SSPE. Of the probable cases, four did not undergo LP but each had high serum titres of measles-specific antibodies. The remaining two children presented with clinical features consistent with SSPE (myoclonus, motor and speech deficits) with negative CSF measles IgG titers but elevated serum titres at 1∶2048 and 1∶16, respectively. The latter child had no history of measles vaccination.
In all six probable SSPE cases, no other cause of encephalopathy was identified. Normal plasma electrolytes and hepatorenal function excluded metabolic, renal and hepatic encephalopathy. Giemsa-stained thick blood films were negative for malaria parasites and plasma C-reactive protein, blood lactate, white cell count and blood culture results did not suggest an acute infective aetiology. The absence of a CSF pleocytosis in the two children with probable SSPE in whom LP was performed made tuberculous meningitis or cryptococcal meningitis unlikely. CSF from both children was negative by PCR for enteroviruses, Japanese encephalitis virus, Murray Valley encephalitis virus, West Nile virus (including Kunjin) and dengue virus, and serum was negative for the presence of IgM to flaviviruses. Based on clinical presentation and course, serum measles antibody titres and the exclusion of other causes of an encephalopathy, the six children with probable SSPE were included in estimates of SSPE incidence.
Although only one child died in hospital, the remaining children were discharged in line with usual management of SSPE in PNG. These children had moderate or severe disability requiring assistance with most or all activities of daily living. An examination of post-discharge outcome was beyond the scope of the present study.
Incidence of SSPE in relation to national vaccine coverage and acute measles incidence
Figures 1 and 2 show PNG national vaccine coverage and acute measles cases since 1997 [11] , [12] , [13] , and the year of birth of the present 22 SSPE cases is shown in Figure 3 . Despite relatively stable vaccination coverage between 50% and 65% from 1997 to 2008, there was a substantial increase in the numbers of reported acute measles cases in 2002 with a smaller prior peak in 1999 and 2000. There is a close concordance between the distribution of the years of birth of the SSPE cases and that for acute measles nationally (Spearman r = 0.88, P = 0.002).
Provincial and district incidence of SSPE
The location of the home village for each child with SSPE and the annual incidence of SSPE in the 13 districts in Madang Province are shown in Figure 4 . The majority of the children were from remote rural districts with very limited health care access. The overall estimated annual incidence for Madang province was 29 (95% confidence intervals [18 to 45])/million total population or 54/million population <20 years of age. In Josephstaal, Yawar, Astrolabe Bay and Bundi LLGs, the estimated annual incidence was 296 [96 to 691], 194 [78 to 400], 122 [15 to 442] and 119 [3 to 660]/million, respectively. There were no reported SSPE cases from 4 districts. Three of these have no roads and are only accessible by air, river or foot. | Discussion
The present study conducted in coastal Madang Province confirms the relatively high incidence of SSPE in PNG shown previously in several highland provincial surveys conducted during the 12 years up to 1999 [5] , [6] , [22] . However, our data also show that such incidence rates must be interpreted in the light of prior measles epidemiology. There was a clear association between the year of birth of our SSPE cases and national figures for acute measles infection that included a substantial increase in cases in the year 2002. This relationship suggests that, despite the possibility of under-reporting [10] , [14] , [15] , temporal trends in measles cases in PNG are relatively accurate. Without equivalent antecedent data, it is difficult to interpret prior reports [5] , [6] , [22] in which a high SSPE incidence may have simply reflected peaks in measles cases 3–10 years beforehand. The decline in reported acute measles in PNG since 2002, including very few cases over the last 5 years [13] , should herald a substantial reduction in SSPE incidence in PNG over the next few years. Nevertheless, a rising seroprevalence during childhood which exceeds that associated with vaccination coverage and SIA may mean that continued local measles transmission will sustain future low-level presentation of new cases [10] . Although our study captured the delayed peak in SSPE incidence attributable to the 2002 measles epidemic, there have been two further children admitted to Modilon Hospital with a clinical diagnosis of SSPE in the 12 months since recruitment to the present study finished.
The demographic features and clinical course of our patients were similar to those of published series from PNG and other countries. Consistent with previously-reported studies [23] , we could not always determine age of onset of symptoms accurately, but the median age at the time of admission in our children (7.3 years) and the male∶female ratio (1.4∶1) were similar to those in SSPE cases from the PNG highlands a decade ago (7.9 years and 1.2∶1, respectively) [4] . Although a male excess is usual, there has been a large age range at presentation, from <5 years in one of the first PNG studies [22] to >10 years in Europid populations [23] , [24] . This is likely to reflect population-specific differences in contributing factors such as persistence of maternal antibodies and vaccination policies.
The present study is the first to have had access to LLG population data to facilitate an assessment of SSPE epidemiology at a sub-provincial level in PNG. The incidence of SSPE exceeded 100 per million population <20 years old in four LLGs of Madang Province. Although there were relatively few cases in some of sub-districts, this is the highest rate yet recorded. Only half of PNG children receive both doses of measles vaccine before their first birthday [10] , [13] , but the prior measles vaccination rate documented for the children from these districts did not differ significantly from that of the non-SSPE severely ill control children nor from national coverage at the time of likely measles infection. This suggests that factors other than vaccine delivery were responsible. There are known continuing difficulties with ensuring a reliable vaccine cold chain in PNG [25] , [26] , but it is also possible that post-measles vaccination seroconversion rates were unusually low in these areas or that the acute measles incidence was particularly high. Alternatively, the children in these communities have an increased susceptibility to SSPE.
Vaccine seroconversion is highly age-dependent. Only 36% of Melanesian children will develop protective measles immunity after their first vaccination at 6 months [27] while recent data from Madang also indicate low rates of protective immunity to measles in children who had received one or both doses of measles vaccine before one year of age [10] . This reflects, in part, persistence of low-level interfering passive maternal antibodies for up to 12 months [28] , especially when maternal immunity has been acquired by natural infection rather than vaccination [29] . The weight of epidemiologic evidence suggests that SSPE is more likely to occur when measles infects a child in the first year of life [23] , [30] , [31] . Indeed, the clear relationship between year of birth of our children with SSPE and nationally reported measles incidence implies that our cases were very young when they encountered measles virus for the first time. Given that most young children are vulnerable to measles, even if vaccinated, it is likely that differences in the numbers of SSPE cases between districts reflect similar local differences in acute measles incidence between 1998 and 2003.
To ensure adequate herd immunity to measles, countries must achieve 92–95% vaccination coverage that includes two separate doses of the vaccine [32] . Given that the expanded programme for immunisation started in 1982 in PNG and the fact that measles vaccine coverage in PNG has remained ≤70% for at least the last 10 years [11] , it is likely that herd immunity was very low in the more remote communities of Madang Province early in the millennium and that acute measles cases were correspondingly high. The fact that four very remote districts were not represented in our series suggests that either they were isolated from the increase in acute measles cases at that time or that children with SSPE were not brought to Modilon Hospital because of the logistic issues involved with patient transfer.
Prior to widespread vaccination, the incidence of SSPE was between 1.2–6.7 per million population <20 years of age in countries where valid data were available [33] , [34] , [35] , [36] . However, incidence rates up to 43 per million population <20 years of age have been estimated in some developing countries [6] . Furthermore, even within closely located communities, the incidence of SSPE is not uniform. For example, Ashkenazi Jews in Israel have a lower rate of SSPE than Separdic Jews (0.5 vs 3.4 cases per million population, respectively) [36] . Our children are from a comparatively homogenous Melanesian group but there were two first cousins of similar age diagnosed within a year of one another. Although from a single set of close-living relatives, SSPE has been described previously in sibling and twin pairs implying at least some familial predisposition, but a clear genetic basis for susceptibility has yet to be defined [37] , [38] , [39] , [40] , [41] . Single nucleotide polymorphisms in a number of immunity-related genes have been found to be associated with SSPE in Japanese [42] , [43] and Turkish [44] patients, but not in other ethnic groups [45] . The search for genetic associations is difficult in an uncommon disease like SSPE, even in high incidence settings such as PNG, and they would have to account for socio-cultural factors that might promote measles infection at an early age in relatively non-immune, unvaccinated populations [46] .
SSPE is diagnosed on clinical grounds alone in resource poor, high incidence settings similar to that of the present study where there are no brain imaging or EEG facilities. Serologic testing for measles is only available as a research or epidemiologic tool. In our case series, six children had probable SSPE without confirmatory CSF serology. Four of these children did not have LP performed but had very high serum titres of measles-specific IgG. The remaining two had negative CSF serology but characteristic clinical features. Because relatively comprehensive clinical and laboratory investigations excluded other likely causes of encephalopathy, the fact that measles CSF and serum titres in SSPE cases can overlap those of controls [3] , and given the high pre-test probability of SSPE, we believe that these two latter children represent part of the spectrum of the disease. Our series also included two children with atypical clinical features. One presented with muscle spasms and rapidly progressed to death within two weeks of illness onset. The other had a subacute history and complex dyskinetic movements. Atypical presentations have been described previously and are sometimes accompanied by radiologic evidence of extensive brainstem as well as cortical involvement [47] . It is likely that, had cerebral imaging been available, there would have been similar radiologic findings in these two children.
The PNG Pediatric Guidelines recommend that the first measles vaccine be given at 6 months of age and the second at 9 months of age [7] . The reason for this policy is the increased morbidity and mortality from acute measles in younger infants rather than high rates of SSPE [48] . However, the low seroconversion rates in this age-group argue for a delay in the age of the first measles vaccine to 9 months of age followed by a second dose at 12–15 months. This could be reconsidered if there were to be an outbreak of measles, but this has not happened in PNG for the last 8 years.
The present study extends past published data suggesting that PNG has the highest reported incidence of SSPE globally. However, this high incidence is related to prior measles epidemics. We have also shown that the incidence of SSPE varies between communities. This could reflect localized failure of the vaccine cold chain, community-specific factors that increase measles transmission, variability in public health surveillance and/or differences in genetic susceptibility to SSPE. Young PNG children do not respond well to measles vaccine. Because of this, efforts such as SIA should continue in order to reduce the pool of non-immune older people surrounding the youngest and most vulnerable members of PNG communities. SSPE has a high mortality, but most children with SSPE require prolonged care because of profound disabilities. Such dependence comes at substantial cost for caregivers. | Conceived and designed the experiments: LM IM HAK IH PMS TMED. Performed the experiments: LM ML HE. Analyzed the data: LM ML IM DS TMED. Wrote the paper: LM ML IM HAK DS IH PMS TMED.
Introduction
Subacute sclerosing panencephalitis (SSPE) is a late, rare and usually fatal complication of measles infection. Although a very high incidence of SSPE in Papua New Guinea (PNG) was first recognized 20 years ago, estimated measles vaccine coverage has remained at ≤70% since and a large measles epidemic occurred in 2002. We report a series of 22 SSPE cases presenting between November 2007 and July 2009 in Madang Province, PNG, including localized clusters with the highest ever reported annual incidence.
Methodology/Principal Findings
As part of a prospective observational study of severe childhood illness at Modilon Hospital, the provincial referral center, children presenting with evidence of meningo-encephalitis were assessed in detail including lumbar puncture in most cases. A diagnosis of SSPE was based on clinical features and presence of measles-specific IgG in cerebrospinal fluid and/or plasma. The estimated annual SSPE incidence in Madang province was 54/million population aged <20 years, but four sub-districts had an incidence >100/million/year. The distribution of year of birth of the 22 children with SSPE closely matched the reported annual measles incidence in PNG, including a peak in 2002.
Conclusions/Significance
SSPE follows measles infections in very young PNG children. Because PNG children have known low seroconversion rates to the first measles vaccine given at 6 months of age, efforts such as supplementary measles immunisation programs should continue in order to reduce the pool of non-immune people surrounding the youngest and most vulnerable members of PNG communities.
Author Summary
Subacute sclerosing panencephalitis (SSPE) is a disabling and usually fatal brain disorder that typically occurs 3–10 years after acute measles infection. Papua New Guinea (PNG) has particularly high rates of SSPE. We report 22 cases of PNG children presenting to the provincial referral hospital in Madang Province who probably contracted acute measles when <12 months of age during a national epidemic in 2002 and who developed SSPE 5–7 years later. Based on these cases, the estimated annual SSPE incidence in Madang province in 2007–2009 was 54/million population aged <20 years. Four sub-districts had an annual incidence >100/million population aged <20 years, the highest rates ever reported. Young PNG children do not respond well to measles vaccine. Because of this, efforts such as supplementary measles immunisation programs should continue in order to reduce the pool of non-immune older people surrounding the youngest and most vulnerable members of PNG communities. | Supporting Information | We thank the patients and their families, the medical and nursing staff of Modilon Hospital's Pediatric Ward, Drs David Williams and Meredith Hodge and the staff of the serology section at PathWest Laboratory Medicine for assistance with sample processing and measles serologic testing, and the clinical research nurses, scientific and administrative staff of the Papua New Guinea Institute of Medical Research. | CC BY | no | 2022-01-13 05:08:15 | PLoS Negl Trop Dis. 2011 Jan 4; 5(1):e932 | oa_package/9c/59/PMC3014974.tar.gz |
|
PMC3014975 | 21245919 | Introduction
Systemic Lupus Erythematosus (SLE) is a chronic autoimmune disease characterized by the production of antibodies to double stranded DNA (dsDNA) and ribonucleoproteins. The etiology of SLE is unknown, although genetic and environmental causes have been implicated. Several viruses have been linked to SLE, however, the strongest association has been made with the Epstein-Barr virus (EBV). EBV is a lymphotropic, dsDNA herpes virus that infects 90–95% of adults in the United States [1] . Despite this high incidence of infection, only a small subset of infected individuals will develop SLE [2] . Epidemiological studies have demonstrated a higher incidence of EBV infection and higher titers of antibodies to EBV in both young and adult lupus patients relative to healthy individuals. James et al., observed seroconversion (development of IgG antibodies to EBV viral capsid antigen) in 99% of adolescent SLE patients compared to 70% of healthy adolescents and 72% of adolescents with other rheumatic diseases [3] . In addition, they observed by PCR analysis, the presence of EBV DNA in lymphocytes of 100% of SLE patients tested, compared to 72% of controls. McClain et.al. observed that antibodies to a major EBV nuclear antigen, EBNA-1, which is continuously expressed in latently infected B cells, arose in all pediatric SLE patients examined compared to only 69% of healthy pediatric controls [4] .
EBNA-1 is a DNA binding protein that maintains replication of the EBV genome within infected cells. It is also required for maintaining viral latency. Several studies suggest that exposure to EBNA-1 following EBV infection, can lead to an autoimmune response in some individuals, which may play a role in SLE disease etiology. It has been reported that antibodies to epitopes on EBNA-1 cross-react with epitopes on Sm, a ribonucleoprotein complex consisting of a core of polypeptides (B/B′, D, E, F, G) [5] , [6] . Sabbatini et al. demonstrated that antibodies to Sm D could be generated in mice immunized with a Gly-Arg rich peptide derived from the amino terminal end of EBNA-1 [7] . James et al revealed that antibodies to Sm B/B′ could be elicited in rabbits and mice following immunization with a proline rich peptide in the carboxyl end of EBNA-1 (PPPGRRP) that has homology to a proline rich region (PPPGMRPP) found in Sm [8] . In addition, they observed that some animals subsequently developed antibodies to dsDNA , which they hypothesized arose as a consequence of epitope spreading, although this was not proven. More recently, Poole et al showed that rabbits and mice injected with the proline rich peptide of EBNA-1, subsequently develop antibodies to U1 ribonucleoproteins, RNP A and RNP C as a consequence of epitope spreading [9] .
Our laboratory previously reported, that BALB/c mice immunized with an EBNA-1 expression vector that expressed either the entire EBNA-1 protein or EBNA-1 lacking the Gly-Ala repeat, developed antibodies to dsDNA as well as to Sm [10] . It was assumed that the antibodies to Sm arose because of cross-reactivity with EBNA-1 as previously reported, however, the basis for the anti-dsDNA response was unknown. The present study was undertaken to address this issue. Our results strikingly reveal that many antibodies elicited in response to EBNA-1 actually cross-react with dsDNA. | Materials and Methods
All animals were handled in strict accordance with good animal practice as defined by federal and state policies set forth by The Public Health Service Policy on the Humane Care and Use of Laboratory Animals (PHS 1986), The Guide for the Care and Use of Laboratory Animals (ILAR 1996), and The USDA Animal Welfare Act (CFR 1985). All work done with animals in this study, was approved by The Institutional Animal Care and Use Committee (IACUC) at The City College of New York, (approval numbers 626 and 828).
Extraction, Purification and Characterization of rEBNA-1 lacking the Gly-Ala repeat
The EBNA-1 baculovirus expression vector used in this study was a generous gift from Dr. Lori Frappier (McMaster University, Ontario, Canada). Recombinant EBNA-1 protein (rEBNA-1) was isolated from this baculovirus expression vector according to Lori Frappier (personal communication and modifications of Frappier and O'Donnell) [35] . This vector encodes an EBNA-1 protein that has a deletion of most of the Gly-Ala repeat and has a 6× His tag on the N-terminus, which allows for the protein's isolation on a Ni 2+ metal affinity column. Briefly, SF9 cells were grown in serum-free insect cell culture medium, Sf-900 II SFM (Invitrogen, Carlsbad, CA) at 27°C. Cells were resuspended at a concentration of 1×10 6 cells per ml and 100 ml of cells (100×10 6 cells total) were infected with 500 μl of high titer recombinant EBNA-1 baculovirus and grown in 500 ml Erlenmeyer flasks (Corning, Acton, MA) at 27°C in an air shaker for 60 hours. The cells were then harvested by centrifuging at 2000 rpm at 4°C for 10 minutes. The cell pellets were resuspended in 25 ml of a hypotonic buffer (20 mM HEPES pH 7.8, 1 mM MgCl 2 , 1 mM PMSF and 10 μM leupeptin) and allowed to swell on ice for 10 minutes. Cells were then dounced 20 times on ice and centrifuged at 4°C at 3000 rpm for 10 minutes. Supernatant was discarded. The pellet containing intact nuclei was resuspended in 25 ml of hypotonic buffer containing 2.7 ml of 5 M NaCl. After douncing on ice to open the nuclear envelope, the fraction was centrifuged at 18,000 rpm for 20 minutes and the supernatant containing rEBNA-1 protein was collected. Further purification of rEBNA-1 was performed employing a nickel agarose (Ni 2+ -NTA) (QIAGEN, Valencia, CA) column according to modifications of Ceccarelli and Frappier [12] . Ni 2+ -NTA agarose (1ml) was equilibrated in column buffer (0.2 M Hepes pH 7.8, 0.5 M NaCl, 10% glycerol) containing 5 mM imidazole, at room temperature. The nuclear extract was incubated with pre-equilibrated Ni 2+ -NTA at room temperature for 2 hours, with rocking. After incubation, a column was packed with the nuclear extract/Ni 2+ -NTA slurry. The column was washed slowly with column buffer containing 5 mM imidazole followed by column buffer containing 25 mM imidazole. Next, the EBNA-1 protein was eluted with column buffer containing 300 mM imidazole. The protein was then concentrated and the buffer exchanged with PBS, 250 mM NaCl using an Amicon Centrifugal filter (10,000 molecular weight cut off) (Millipore, Billerica, MA). The protein was then resolved by 12% SDS-PAGE followed by a Western blot and immunostaining with a monoclonal antibody to EBNA-1.
Injection of mice with rEBNA-1 protein
Fifteen, six week old, female BALB/c mice were used for injection studies. Five mice were injected intraperitoneally (ip) with 50 μg of rEBNA-1 protein in complete Freund's adjuvant (CFA) (Sigma, St Louis, MO) in a 1∶1 (v/v) ratio and boosted twice (at weeks 3 and 9) with 25 μg of rEBNA-1 in incomplete Freund's adjuvant (IFA). Five mice were injected with CFA only and boosted with IFA and 5 age-matched control mice remained uninjected throughout the study. The mice were bled immediately before injection and at weeks 1.5, 4, 6, 10, 12, 15 and 18. The sera obtained, from these mice were tested for anti-EBNA-1 and anti-dsDNA antibodies by ELISA.
Construction of plasmids encoding the amino and carboxyl regions of EBNA-1
Truncated EBNA-1 proteins were isolated from plasmid transformed E. coli cells. The pLS8 expression plasmid carries the encoding sequence for the amino terminus of the EBNA-1 antigen, from the initial Met residue to amino acid position 404 and lacks virtually all of the Gly-Ala repeat. It was prepared by PCR amplification of the EBNA-1 gene from pMRC72 [10] which contains the EBNA-1 coding sequence, but lacks the Gly-Ala repeat, using the following primer pair; EBV7, 5 – CATATGTCTGACGAGGGGC CAGGT-3′ (forward primer) and EBV6, 5′-CTCGAGTTATGGCCTTCTACCTGG-3′ (reverse primer). The pLS7 expression plasmid also carries the encoding sequence for the amino terminus of the EBNA-1 protein, from the initial Met residue but it terminates at amino acid position 393. Like pLS8, it lacks most of the Gly-Ala repeat. However, unlike pLS8, it is missing the PPPGRRP epitope (aa 398–404). It was prepared from pMRC72 using the following primer pair; EBV7 (see above) and EBV5, 5′ CTCGAGTTAAGACCCGGAT GATGA 3′ (reverse primer). The pLS9 expression plasmid carries the EBNA-1 encoding sequence for the carboxyl terminus of EBNA-1 from amino acids 410 to 641. It was also prepared by PCR amplification of the EBNA-1 gene from pMRC72 using the following primer pair; EBV3, 5′-CATATGGGGGAA GCCGATTA TTTTGAAT-3′ (forward primer) and EBV 4, 5′-CTCGAGTTACTCCTGCCCTTCCTC-3′ (reverse primer). The PCR amplifications were performed for 30 cycles. The amino and carboxyl PCR fragments were digested with Nde1 and Xho1 and inserted into the pET28A expression vector (Novagen, San Diego, CA) which contains an N-terminal 6× His tag.
Isolation of truncated recombinant EBNA-1 proteins
E. coli colonies transformed with pLS7, pLS8, or pLS9 (see above) were selected on LB ampicillin plates and grown at 37°C in 50 ml of LB media containing 1% glucose. Cultures were diluted in 490 ml LB with 0.1 mM IPTG and grown for several hours at 20°C to a final OD 600 of approximately 0.6. Cultures were harvested and re-suspended in lysis buffer (50 mM Tris-Hcl, pH 7.8, 250 mM NaCl) containing 1.0 mM PMSF. Cells were sonicated for 15 minutes on ice with a 4 second on pulse, 6 seconds off at a 30% amplitude. The cell lysate was cleared by centrifugation at 10,000 rpm for 30 minutes at 4°C and filtered through a 45 μM filter. Five mls of Ni 2+ -NTA beads equilibrated with lysis buffer were added to the cleared supernatant and incubated with gentle rocking at room temperature. The beads (bound to the recombinant protein) were separated from the supernatant by low- speed centrifugation. They were then washed 6 times with wash buffer (50 mM Tris-HCl, ph 7.8, 250 mM Nacl, 60 mM imidazol, and 10% glycerol). Two ml of elution buffer (50 mM Tris-Hcl, pH 7.8, 250 mM NaCl, 250 mM imidazol, 10% glycerol) were added to the beads and beads were rocked for 15 minutes. The beads were removed from the reaction by low-speed centrifugation. Supernatants containing the recombinant protein were concentrated and the buffer was exchanged with PBS, 250 mM NaCl using an Amicon Centrifugal filter. Proteins were analyzed by SDS-PAGE and Western Blot.
Plasmids (vector pET15b) expressing the following EBNA-1 amino acid sequences; EBNA 452–641 , EBNA 459–607 , and EBNA 459–619 were gifts from Dr. Lori Frappier [16] . Soluble truncated EBNA-1 proteins were produced in Escherichia coli strain BL21 (DE3) and isolated from cell-lysates. Proteins were then purified over a Ni-NTA agarose column as described above. Proteins were analyzed by SDS-PAGE and Western Blot.
ELISAs
Detection of antibodies to EBNA-1, dsDNA, Sm, LPS, Proteinase 3 and BSA
Diluted serum samples from EBNA-1 injected mice, hybridoma supernatants, or purified monoclonal antibodies were tested for binding to EBNA-1, dsDNA, Sm, LPS, or PR-3 by ELISA as previously described [10] , [36] . For the detection of antibodies to EBNA-1, LPS, Proteinase 3, and BSA, Costar plates (Corning Incorporated, Corning, NY) were coated in PBS with 2.0–5.0 μg/ml of antigen. Costar plates were coated overnight with 5.0 μg/ml of Sm (Immunovision, Springdale, AR) in 0.1M carbonate buffer for the detection of antibodies to Sm. For the detection of antibodies to dsDNA, Immulon-2 plates (Dynatech Laboratories, Inc., Chantilly, VA) were coated with 100 μg/ml of calf thymus dsDNA.
Detection of antibody binding to truncated amino or carboxyl fragments of EBNA-1
Purified monoclonal antibodies were tested for binding to truncated amino (LS7 and LS8) and carboxyl regions (LS9, EBNA 452–641 , EBNA 459–619 , and EBNA 459–607 ) of EBNA-1. ELISA plates were coated with 2.0 μg/ml of the purified, truncated recombinant proteins isolated in this laboratory. Subsequent steps in the ELISA were performed according to Sundar et al [10] .
Isotype ELISA
ELISA plates were coated with 50 μl of a 1∶1000 dilution of either unlabeled goat anti-mouse IgG1, IgG2a, IgG2b or IgG3 (Southern Biotech, Birmingham, Alabama) and incubated at 37°C for one hour and overnight at 4°C. Monoclonal 3D4 antibody was diluted to 1.5 μg/ml and incubated on the plate for one hour at 37°C. Next, 50 μl of a 1∶1000 dilution of goat anti-mouse IgG1 conjugated to alkaline phosphatase (AP), anti-IgG2a-AP, anti-IgG2b-AP, or anti-IgG3-AP (Southern Biotech) was added to wells coated with unlabeled anti-IgG1, anti-IgG2a, anti-IgG2b, or anti- IgG3 respectively. Color development was measured following the addition of 4-nitrophenyl-phosphate disodium salt as substrate and plates were read at 405 nm on a Titertek Multiscan ELISA plate reader.
Quantitative ELISA
A quantitative ELISA was performed, as previously described, to determine the concentration of purified monoclonal IgG antibodies in hybridoma supernatants [36] . Briefly, ELISA plates were coated overnight with 1.0 μg/well of goat anti-mouse IgG antibody (Southern Biotechnology). A commercial mouse monoclonal IgG antibody (Sigma) was serially diluted, beginning at a concentration of 200 ng/ml and used to generate a standard curve. Serial dilutions of monoclonal antibody purified in this laboratory were applied to the anti-IgG coated wells and the concentration of antibody was extrapolated from the standard curve. Monoclonal antibodies were detected with goat anti-mouse IgG antibody conjugated to AP followed by the addition of 4-nitrophenyl-phosphate disodium salt as substrate.
Crithidia Assay
Ready to use Crithidia slides from the CrithiDNA Anti-nDNA Antibody Test Kit from Antibodies Inc. (Davis, CA), were immunostained either with mouse sera from EBNA-1 injected mice, diluted 1/50 or with purified monoclonal antibody diluted to 10 μg/ml. Slides were incubated in a moist, dark chamber for 30 minutes at room temperature (RT). A positive control anti-dsDNA antibody was provided with the kit. A nonspecific monoclonal mouse IgG1 antibody was used as an isotype control (Sigma). Next, the slides were extensively washed with PBS and immunostained for 30 minutes at RT with a 1∶250 dilution of biotinylated goat anti mouse IgG (Southern Biotech). This was followed by 20 μl of a 1∶500 dilution of Streptavidin-FITC (Southern Biotech) for 30 minutes at RT. Slides were washed again and Prolong Gold Antifade, (Invitrogen, Carlsbad, CA) was added prior to examination by fluorescence microscopy using a Nikon Eclipse microscope, model, TE 2000-S at a magnification of 400×.
Western Blot
Proteins were analyzed by SDS-PAGE on a 12% gel and transferred to a nitrocellulose membrane using a Bio-Rad wet transfer apparatus (BioRad, Hercules, CA). After transfer, the membranes were blocked with 3% Milk-PBS for one hour at RT with shaking. The blot was incubated overnight at 4°C with a MAb generated in our laboratory (3D4), diluted to 1 μg/ml or a commercially prepared MAb, 0211 (Thermo Fisher Scientific/Pierce, Rockford, IL) diluted to 10 μg/ml according to the manufacturers protocol. The membrane was washed 6 times in wash buffer (PBS, 0.05% Tween-20). Bound MAbs antibodies were detected with HRP-conjugated goat anti mouse IgG (Southern Biotech) diluted 1∶20,000, followed by chemiluminescence using the Pierce ECL kit according to the manufacturers protocol (Pierce, Rockford, IL). Molecular weight markers conjugated to strep-tag (Precision plus protein WesternC) (Biorad, Hercules, CA) were detected with a 1∶20,000 dilution of Strep-Tactin-HRP (Biorad).
Somatic Cell Fusion
BALB/c mice were immunized intraperitoneally (ip) with 50 μg/ml of rEBNA-1 in CFA and then boosted at 3, 7, and 12 weeks with 25 μg/ml of rEBNA-1 in IFA. Three to four days following the third boost, splenocytes were fused with NSO cells according to Iliev et al [37] . They were grown in complete HAT media supplemented with 20% FBS, 10% NCTC, 1% Penicillin-Streptomycin, 1% non-essential amino acids and 1% L-glutamine. Supernatants from hybridomas were tested for IgG anti-EBNA-1 and anti-dsDNA antibodies by ELISAs as described above.
Purification of Monoclonal Antibodies
Hybridomas producing a MAb to EBNA-1 were grown in serum free media (Hyclone, Logan, Utah) and 400 ml of supernatant were collected for IgG purification. Antibody was purified from the supernatant by eluting it off a protein G Sepharose column (Gamma BindTM Plus Sepharose TM gel beads, Amersham Pharmacia, Uppsala, Sweden) with 0.1M glycine pH 2.5, according to the manufacturer's protocol. Column eluate was neutralized with 1M Tris-HCl. The purified antibody was dialyzed overnight with PBS and antibody concentration was determined by a quantitative ELISA (above).
Antibody adsorption on dsDNA-cellulose columns
Columns were packed with 0.5ml of calf thymus dsDNA-cellulose or cellulose beads (Sigma, St.Louis, MO) according to the manufacturer's protocol. The columns were washed with 10 mM Tris buffer pH 7.9 containing 1 mM EDTA. Columns were then blocked with 5% FBS-PBS overnight at 4°C. A 1/1,000 dilution of week 4 and a 1/5000 dilution of week 12, rEBNA-1 injected mouse sera or 5 μg/ml of MAbs, 3D4 or 0211 were slowly loaded onto cellulose and dsDNA cellulose columns and allowed to sit for 1 hour at 4°C. The flow through was collected and pre and post adsorbed sera or monoclonal antibody were tested for binding to dsDNA and EBNA-1 by ELISA and Western blot as described above. | Results
Mice injected with purified recombinant EBNA-1 protein develop antibodies to dsDNA
We were interested in determining how anti-dsDNA antibodies could arise in mice that develop anti-EBNA-1 antibodies upon exposure to EBNA-1 protein. In our previous study, we generated an anti-EBNA-1 response in mice by injecting them with an EBNA-1 expression vector. However, not all mice developed anti-EBNA-1 antibodies, presumably because they did not all express an adequate concentration of the EBNA-1 protein [10] . In the present study, in order to examine the EBNA-1 response, we decided to inject mice with purified recombinant EBNA-1 protein (rEBNA-1) rather than the EBNA-1 expression vector. EBNA-1 protein used for injections was prepared in our laboratory from a baculovirus vector obtained from Lori Frappier (McMaster University, Ontario, Canada). The rEBNA-1 protein encoded by this vector lacks most of the Gly-Ala repeat. It has been shown that the Gly-Ala repeat is not required for the replication, transactivation or segregation function of EBNA-1, although, it does enable EBNA-1 to escape detection by cytotoxic CD8 + T cells [11] , [12] . The MW of the rEBNA-1 protein lacking the Gly-Ala repeat is approximately 52Kda.
Five, 6 week old, female, BALB/c mice were injected with 50 μg of rEBNA-1 protein in CFA and were boosted 2 times at weeks 3 and 9 with 25 μg of rEBNA-1 in IFA. Five age and sex matched control BALB/c mice were immunized with CFA alone and boosted with IFA and 5 mice were used as uninjected age matched controls. We observed that all 5 mice injected with rEBNA-1 developed IgG antibodies to EBNA-1 within the first 3 weeks ( Figure 1A ). In addition, mice developed antibodies to dsDNA, although, the kinetics of the anti-dsDNA response lagged behind that of the anti-EBNA-1 response suggesting that anti-dsDNA antibodies may have developed over time as a consequence of epitope spreading or somatic mutation ( Figure 1B ). Some mice immunized with adjuvant only, also developed antibodies to dsDNA but with the exception of one mouse, their levels of anti-dsDNA antibody were never as high as that of mice injected with rEBNA-1 in adjuvant. Intraperitoneal delivery of CFA has been shown by others to elicit the production of autoantibodies in mice, including the production of anti-DNA antibodies [13] . It is extremely unlikely that the anti-dsDNA response in rEBNA-1 injected mice was due primarily to adjuvant, as our previous DNA based inoculation studies using EBNA-1 expression vectors in the absence of adjuvant, also elicited the production of anti-dsDNA antibodies [10] .
Week 12 sera from all 5 rEBNA-1 injected mice were serially diluted and tested for binding to EBNA-1 ( Figure 1C ) and dsDNA ( Figure 1D ) by ELISA. All rEBNA-1 injected mice developed high titers of antibody to EBNA-1 and dsDNA. However, the anti-EBNA-1 titers were higher than the anti-dsDNA titers suggesting that either the concentration of antibodies to EBNA-1 were higher than the concentration of antibodies to dsDNA or the affinities of the antibodies to EBNA-1 were higher than for dsDNA. At a dilution of 1∶6400, the anti-EBNA-1 response in all mice was greater than 3 standard deviations above the mean of similarly diluted uninjected control mice (dotted line). At a dilution of 1∶800, the anti-dsDNA response was greater than 3 standard deviations above the mean of uninjected control mice (dotted line). No anti-dsDNA response was observed at a dilution of 1∶6400.
To confirm the specificity of the anti-dsDNA response, week 12 sera from mice injected with rEBNA-1 were diluted 1∶50 and used to immunostain Crithidia luciliae slides. The presence of antibody to dsDNA was indicated by binding to the kinetoplast ( Figure 1E , left panel). In contrast, sera from adjuvant immunized mice did not reveal kinetoplast binding ( Figure 1E , right panel ) indicating that either the anti-DNA antibodies present in these mice were of lower affinity than the anti-dsDNA antibodies obtained from rEBNA-1 injected mice or they were not specific for dsDNA.
To determine whether any of the antibodies to EBNA-1generated in rEBNA-1 injected mice also cross-reacted with dsDNA, week 12 sera from all 5 EBNA-1 injected mice were adsorbed over dsDNA-cellulose beads to remove dsDNA reactive antibodies and then sera were tested by ELISA to determine if adsorbed sera showed reduced binding to EBNA-1. Loss of antibody in the sera due to non specific sticking to cellulose was determined by adsorbing sera to cellulose only beads. Figure 2 , represents the OD 405 nm of anti-dsDNA (A) and anti-EBNA-1 antibody (B), pre and post adsorption onto dsDNA cellulose beads, after the value for non specific binding to cellulose was subtracted. We observed a significant decrease in anti-dsDNA antibody activity following adsorption on dsDNA cellulose beads, in mouse 1, 3, 4, and 5 (p<0.001) ( Figure 2A ). In a similar trend, we observed a significant decrease in anti-EBNA-1 activity in the sera from mouse 1, 3, 4, and 5, following adsorption on dsDNA cellulose (p<0.005) ( Figure 2B ). Mouse 2 showed a small decrease in anti-dsDNA and anti-EBNA-1 activity following adsorption on dsDNA cellulose although it was not significant. This is likely because mouse 2 developed a negligible response to dsDNA following injection with rEBNA-1 although the anti-EBNA-1 response was significant. The observation that a reduction of anti-dsDNA antibody on a dsDNA cellulose column, led to a parallel reduction of anti-EBNA-1 activity, suggests that anti-EBNA-1 antibody in the sera of some rEBNA-1 injected mice, cross-reacts with dsDNA.
Since week 4 rEBNA-1 injected mice displayed a significant delay in the development of a high titer anti-dsDNA but not anti-EBNA-1 response ( Figure 1B ), we wanted to examine whether there was any evidence of cross-reactivity to dsDNA in week 4 sera. We therefore adsorbed week 4 sera over dsDNA cellulose beads. Interestingly, we also observed some reduction in binding to dsDNA and EBNA-1 following adsorption on dsDNA cellulose (data not shown) suggesting that the cross-reactive response arose early. However, since the anti-dsDNA response was much weaker at week 4 than week 12, epitope spreading may have played a role in refining the cross-reactive response over time.
Generation of monoclonal antibodies to EBNA- 1 that cross-react with dsDNA
Although, adsorption studies suggested that antibodies to EBNA-1 generated in rEBNA-1 injected mice cross-creacted with dsDNA they did not prove this. To confirm these results, we therefore generated monoclonal anti-EBNA-1 antibodies from rEBNA-1 injected mice and tested these antibodies for binding to dsDNA by ELISA. To generate monoclonal antibodies to EBNA-1, splenocytes from rEBNA-1 injected BALB/c mice containing serum IgG antibodies to both EBNA-1 and dsDNA, were fused to NSO cells. Hybridoma supernatants were screened by ELISA for IgG antibodies to EBNA-1 and dsDNA. In an initial screen of one fusion, we observed that the majority of clones that tested positive for antibody reactivity to EBNA-1 were also positive for reactivity to dsDNA (10 out of 14 or 71%). Seven clones that were positive for antibodies to EBNA-1 were subcloned two times to insure clonality. Supernatants from these hybridomas were then tested for antibodies to EBNA-1 and dsDNA. In addition, they were tested for the presence of antibodies to the blocking agent Bovine Serum Albumin (BSA). One of the subclones, 3F3, secreted antibody specific for EBNA-1 only ( Table 1 ). Three subclones, secreted monoclonal antibody that reacted strongly with both EBNA-1 and dsDNA and did not bind BSA. Subclone 3D4 is representative of this group ( Table 1 ). Three other subclones, represented by 9G3, secreted antibody that bound not only to EBNA-1 and dsDNA but BSA as well ( Table 1 ).
In the present study we chose to focus on 3D4 because it is an IgG antibody that binds strongly to dsDNA, which is characteristic of many pathogenic IgG anti-dsDNA antibodies that arise in SLE. In addition, it's strong binding to both EBNA-1 and dsDNA made it a good candidate for studying the basis of this cross-reactive response. The other antibodies such as 9G3 that also reacted with BSA (and casein), were not further characterized in this study because of concern that this would lead to non specific binding in ELISAs and Western blots; assays which require these blocking reagents. In addition these antibodies displayed a much weaker affinity for EBNA-1 and dsDNA making them less desirable to use in our initial attempt to identify epitopes in EBNA-1 that play a role in cross-reactivity to dsDNA.
3D4 was isolated from hybridoma supernatant, on a protein G column. Following purification, 3D4 was shown to bind to EBNA-1 by ELISA even at concentrations as low as 0.125 μg/ml ( Figure 3A ). Specificity of this antibody was demonstrated by its lack of binding to a control viral antigen, cystovirus RNA polymerase, P2, isolated by Gottlieb et al [14] and BSA ( Figure 3A ).
Purified 3D4 was also shown to cross-react with dsDNA by ELISA ( Figure 3B ). 3D4 was shown by ELISA to be of the IgG1 isotype ( Figure 3C ). Reactivity of 3D4 with dsDNA was confirmed by its ability to recognize the dsDNA containing kinetoplasts of Crithidia luciliae ( Figure 3D , left panel). An IgG1 isotype control MAb failed to bind kinetoplasts ( Figure 3D , right panel). Cross-reactivity of 3D4 was further demonstrated by adsorption of the purified MAb over dsDNA cellulose beads and then testing pre and post adsorbed antibody for binding to dsDNA and EBNA-1 by ELISA and Western blot ( Figure 3E and F ). Adsorption over dsDNA cellulose resulted in complete depletion of 3D4 as detected by anti-dsDNA and and anti-EBNA-1 ELISAs ( Figure 3E ). Post dsDNA-cellulose adsorbed 3D4 antibody also showed dramatically reduced binding to EBNA-1 compared to pre-adsorbed 3D4, by Western blot ( Figure 3F , right panel, compare lanes 2 and 4). No binding of pre-adsorbed 3D4 antibody to BSA was observed (lane 3).
We were also interested in determining whether a monoclonal anti-EBNA-1antibody isolated from a completely different source would have similar binding properties to the 3D4 MAb isolated in our laboratory. We therefore examined the ability of a commercially prepared monoclonal IgG1 anti-EBNA-1 antibody, 0211 (Thermo Fisher Scientific/Pierce, Rockford, IL) to cross-react with dsDNA. The only information known about 0211 is that it was generated in response to EBNA-1, however, the exact epitope that it recognizes has not yet been identified. We first confirmed by ELISA, that this antibody binds to EBNA-1 but not to BSA or P2 ( Figure 4A ). We next observed by ELISA that this antibody also cross-reacts with dsDNA ( Figure 4B ). Furthermore, adsorption of 0211on a dsDNA cellulose column, resulted in complete depletion of the antibody as detected by anti- dsDNA and anti-EBNA-1 ELISAs ( Figure 4C ). A reduction in binding of post dsDNA cellulose adsorbed antibody to EBNA-1 was also demonstrated by Western blot ( Figure 4D , right panel, compare lanes 2 and 4). No binding of pre-adsorbed 0211 antibody to BSA was observed (lane 3).
A comparison of MAbs 3D4 and 0211 revealed that although both antibodies bind strongly to EBNA-1, 3D4 has an even higher affinity for EBNA-1 than 0211 ( Figure 5A ). At a concentration of 0.1 μg/ml, 3D4 still bound robustly to EBNA-1 while binding by 0211 was negligible.
MAbs 3D4 and 0211 were examined for binding to Sm and a panel of antigens
MAbs 3D4 and 0211 were also examined for binding to Sm, lipopolysaccharide (LPS), BSA, and proteinase -3 (PR-3) which is the target autoantigen in Wegener's granulomatosis. Antibodies to PR-3 are a subgroup of classic anti-neutrophil cytoplasmic antibodies (cANCA ). At 5 μg/ml, 3D4 displayed negligible binding to Sm relative to BSA ( Figure 5B ) while 0211 bound moderately well to Sm ( Figure 5C ). We also tested the binding of 3D4 and 0211 to LPS because it is negatively charged [15] . Since dsDNA is negatively charged, we wondered whether the MAbs would bind other negatively charged antigens. However, we observed that both 3D4 and 0211 failed to bind LPS.
MAbs 3D4 and 0211 display differences in reactivity to the amino and carboxyl regions of EBNA-1
To begin to understand whether MAbs 3D4 and 0211 recognize the same or different regions of EBNA-1, they were examined by ELISA for binding to three truncated recombinant EBNA-1 proteins, LS7, LS8, and LS9, isolated in this laboratory from E. coli. These truncated recombinant proteins are comprised of the amino or carboxyl regions of EBNA-1. The protein designated LS8, is comprised of the amino region of rEBNA-1, from the initial Met residue to aa 404 ( Figure 6A ). Like rEBNA-1 used in this study, it lacks most of the Gly-Ala repeat. It contains the PPPGRPP region in EBNA-1 (aa 398–404) that was shown by James et al to be homologous to a proline rich epitope in Sm B/B′ [8] . LS7 is identical to LS8 except that it terminates at aa 393 and therefore lacks the PPPGRPP epitope ( Figure 6A ). The rational for generating two amino fragments, one with and one without the proline rich epitope was to determine whether this epitope which is responsible for eliciting cross-reactivity with Sm is also involved in eliciting cross-reactivity with dsDNA. LS9 comprises the carboxyl region of the rEBNA-1 protein from aa 410 to the terminal aa 641 and lacks the proline epitope ( Figure 6A ). MAb 3D4 was observed to bind strongly to LS9 but not at all to LS7 or LS8 ( Figure 6B ). The kinetics of 3D4 binding to LS9 closely paralleled the kinetics of binding to the entire rEBNA-1 protein indicating that this carboxyl region (aa 410–641) is sufficient for optimal recognition by 3D4. Adsorption of 3D4 to dsDNA cellulose was also observed to remove all binding to the carboxyl fragment. Taken together these results suggest that the cross-reactive epitope recognized by 3D4, is configured within the carboxyl region.
MAb 0211 was observed to bind all three truncated proteins indicating that it recognizes epitopes in both the amino and carboxyl regions of EBNA-1, however the binding to the amino proteins, LS7 and LS8 is better than the binding to the carboxyl protein, LS9 ( Figure 6C ). Interestingly, 0211 binds more strongly to LS7 than LS8 and since LS7 does not contain the PPPGRPP epitope, this indicates that the proline epitope is not necessary for the binding of 0211 to EBNA-1. Furthermore, this proline rich region may structurally interfere with binding by 0211. It cannot be determined at this time whether the epitope in the amino or carboxyl region of EBNA-1 is responsible for MAb 0211's cross-reactivity with dsDNA.
Despite the fact that 3D4 and 0211 bind differently to the amino and carboxyl regions of EBNA-1, both antibodies still cross-react with dsDNA. Consequently there could be more than one EBNA-1 epitope that could be linear or conformational, that acts as a mimotope for dsDNA. Alternatively, the epitope (s) in the carboxyl region may be more important for cross-reactivity with dsDNA and since 0211 also binds Sm, the epitope in the amino region may be more important for cross-reactivity with Sm.
3D4 binds to a 148 aa core domain in the carboxyl region of EBNA-1 that lacks the negatively charged C-terminal amino acids
To begin to identify a smaller fragment in the carboxyl region of EBNA-1 that contains the epitope recognized by 3D4, we examined the binding of this MAb to three truncated carboxyl fragments; EBNA 452–641 , EBNA 459–619 , and EBNA 459–607 ( Figure 7A ). These fragments are expressed by plasmids kindly provided to us by Dr. Lori Frappier [16] . We observed that 3D4 bound all 3 fragments equally well and did not show diminished binding to these fragments relative to the entire carboxyl region (EBNA 410–641 ) ( Figure 7B ). In fact 3D4 displayed optimal binding to the smallest fragment, EBNA 459–607 suggesting that the cross-reactive epitope lies within this 148 aa region. The carboxyl region of EBNA-1 has a net negative charge due to the high frequency of negatively charged amino acids at the C-terminus (aa 619–641). Twelve out of 22 of the C terminal amino acids are either glutamic or aspartic acid. Both, EBNA 459–619 , and EBNA 459–607 lack these negatively charged amino acids. Since removal of these negatively charged amino acids did not diminish recognition by 3D4, this suggests that charge interaction is not the basis for 3D4's binding to EBNA-1. MAb, 0211 displays a similar binding pattern to the truncated carboxyl fragments of EBNA-1 with maximal binding to the two smallest fragments EBNA 459–619 and EBNA 459–607 (data not shown). | Discussion
This study demonstrates for the first time, that some antibodies that arise in response to EBNA-1 cross-react with dsDNA. Our laboratory previously demonstrated that EBNA-1 expression could elicit an anti-dsDNA response, however, it was not known at the time whether the antibodies to dsDNA were distinct from the anti-EBNA-1 antibodies or whether the same antibodies that bound EBNA-1 were also able to recognize dsDNA. The demonstration that purified monoclonal IgG antibodies to EBNA-1 also bind dsDNA and that adsorption of these antibodies on a dsDNA cellulose column, removes EBNA-1 reactivity, confirms the cross-reactive nature of these antibodies. However, this does not exclude the potential role of epitope spreading in the development of the cross-reactive response. The delay in the development of a strong anti-dsDNA response relative to the anti-EBNA-1 response suggests that cross-reactivity continues to develop over time. It may be that early in the response to EBNA-1, the epitopes that are targeted, elicit only a weakly cross-reactive response to dsDNA. Later in the response, other epitopes may be targeted as a result of intra-molecular epitope spreading and these latter epitopes may be the ones that are responsible for cross-reactivity with dsDNA. Alternatively, antibodies that cross-react more strongly with dsDNA may arise later in the immune response as a consequence of somatic mutation. A specific mutation in the variable heavy and/or light chain regions of an anti-EBNA-1 antibody may alter its specificity from one that only recognizes EBNA-1 to one that recognizes EBNA-1 as well as dsDNA.
While all mice injected with EBNA-1 developed antibodies to dsDNA, we did not consistently observe features of clinical lupus in these mice. Two out of 5 injected mice had significant levels of protein but no blood in their urine relative to uninjected mice. The kidney of 1 out of 3 mice examined at 3 months post injection, had evidence of some IgG immune complex deposition, however, none of the kidneys examined showed signs of lupus histopathology (data not shown). Future studies will include examining a larger cohort of mice for evidence of glomerulonephritis and investigating whether 3D4 and 0211 can deposit in the kidney.
Antibodies to a variety of self proteins have been reported to cross-react with dsDNA, such as antibodies to extracellular matrix protein, HP8, ribosomal P protein, elongation factor-2 (EF-2), α-actinin, the NMDA receptor and Sm D [17] , [18] , [19] , [20] , [21] , [22] . Antibodies targeting peptide mimotopes of dsDNA such as DWEYSVWLSN and RLTSSLRYNP have also been reported [23] , [24] . In addition, antibodies to microbial antigens such as glycolipid components of the cell wall of Mycobacterium tuberculosis, phosphorylcholine in the cell wall of Streptococcus pneumoniae or proteins in Burkholderia fungorum have been observed to cross-react with dsDNA [25] , [26] , [27] , [28] . It is unclear how these antigens act as molecular mimics to dsDNA, but it may be due to similarities in the 3 dimensional structures of these antigens and dsDNA. Evidence from some studies suggest that conformational epitopes are the targets of antibodies that cross-react with dsDNA and self proteins [29] , [30] .
Most of the monoclonal anti-EBNA-1 antibodies generated in this study were found to cross-react with dsDNA. Very few of them were found to recognize EBNA-1 only. This may be because mice that were selected for fusion had already developed maximal levels of cross-reactive antibodies either due to epitope spreading or somatic mutation. In addition, the rEBNA-1 protein used in our injection studies, lacks the Gly-Ala repeat which has been shown to be a major epitope that elicits anti-EBNA-1 antibodies in normal individuals [31] . It may be that in the absence of the Gly-Ala repeat, the response is biased towards other epitopes some of which happen to elicit antibodies that also cross-react with dsDNA. It was previously demonstrated that patients with lupus tend to mount an immune response to different epitopes on EBNA-1 than healthy individuals [4] , [31] , [32] . While sera from healthy individuals, preferentially react with the Gly-Ala repeat, sera from lupus patients tend to recognize epitopes in the amino and carboxyl terminal regions of EBNA-1 that are more likely to be cross-reactive with nuclear autoantigens. It is not clear whether lupus patients are genetically predisposed to developing cross-reactive antibodies or whether they have a defect in B cell tolerance leading to failed regulation of the autoreactive B cells producing these antibodies.
The two MAbs that were extensively characterized in this study, 3D4 and 0211, bind to EBNA-1 and dsDNA, yet 3D4 recognizes only the carboxyl region in EBNA-1, while 0211 recognizes both the amino and carboxyl regions. A homology search failed to find any region in EBNA-1 that is homologous to two previously described peptide mimotopes for dsDNA; DWEYSVWLSN and RLTSSLRYNP. It is not yet known whether 3D4 and 0211 recognize a homologous epitope in the carboxyl region, however the C-terminal negatively charged amino acids do not appear to be necessary for the binding of either of these MAbs to the carboxyl region.
MAb 0211 binds moderately well to Sm while 3D4 displays negligible binding to Sm. The basis for the cross-reactivity of 0211 with Sm does not seem to be dependent on the proline rich epitope described by James et al, since 0211 binds even stronger to a truncated amino fragment of EBNA-1 lacking this determinant [8] . It is not yet clear whether the epitopes in the amino and carboxyl region recognized by 0211 share homology. However, the observation that 0211 binds to both Sm and the amino fragment, while 3D4 binds to the carboxyl region only, suggests that an epitope in the amino fragment may be more important for cross-reactivity with Sm while an epitope in the carboxyl region may be important for cross-reactivity with dsDNA. In a preliminary study that is consistent with this, we recently identified a monoclonal IgM antibody that reacts strongly with EBNA-1 and dsDNA but not Sm and only recognizes the carboxyl region of EBNA-1. In addition, recent studies reveal that the MAb, 9G3 ( Table 1 ) that recognizes BSA, also cross-reacts with Sm, and binds to both the amino and carboxyl fragments of EBNA-1 (preliminary data). The polyreactive nature of this antibody, which will be examined in more depth in future studies, is potentially important since polyreactive antibodies have been shown to be the precursors of more pathogenic antibodies in lupus [33] .
As previously mentioned, the basis for antibody cross-reactivity with EBNA-1 and dsDNA does not appear to be charge interactions since removal of the negatively charged amino acids from the carboxyl region of EBNA-1 does not diminish binding to this region ( Figure 7 ). Furthermore, neither 3D4 nor 0211 recognize LPS which is negatively charged ( Figure 5 ). It is possible that the epitopes in EBNA-1 that cross-react with dsDNA are structural. The X-ray structure of the crystallized VBS/DNA binding region of EBNA-1 has been determined and reveals two distinct domains; a core domain that mediates protein dimerization and a flanking domain that mediates base contact with dsDNA [34] . These domains possess much secondary structure, which may serve as targets for antibodies that cross-react with dsDNA. The core domain (aa 504–607) contains a ß sheet, an α helix and a proline loop and the flanking domain (aa 470–503) contains α helixes. Potentially, the cross-reactive antibodies may recognize a portion of the α helix that mimics the α helix in dsDNA. The X-ray crystal structure of the N-terminus has not yet been resolved. However, this region appears to be less structured than the carboxyl region and could be flexible enough to fold onto itself or the carboxyl region providing multiple opportunities for antibody interaction.
It will be important in future studies to map the epitopes in EBNA-1 that lead to the cross-reactivity with dsDNA and determine whether or not these epitopes are conformational. Identifying these epitopes and individuals who preferentially produce antibody responses to them, may be useful for determining those who are at risk of developing lupus so that early treatment strategies can be initiated. In addition, knowledge of these epitopes may help in the design of therapeutic strategies that can mask these epitopes thereby preventing the immune system from mounting a cross-reactive response to them. | Conceived and designed the experiments: PY PG LAS. Performed the experiments: PY HT RE HW AMW AK EK. Analyzed the data: PY HT RE LAS. Contributed reagents/materials/analysis tools: PG. Wrote the paper: PY PG LAS. Assisted teaching students how to perform somatic cell fusions: RHL.
Background
Several genetic and environmental factors have been linked to Systemic Lupus Erythematosus (SLE). One environmental trigger that has a strong association with SLE is the Epstein Barr Virus (EBV). Our laboratory previously demonstrated that BALB/c mice expressing the complete EBNA-1 protein can develop antibodies to double stranded DNA (dsDNA). The present study was undertaken to understand why anti-dsDNA antibodies arise during the immune response to EBNA-1.
Methodology/Principal Findings
In this study, we demonstrated that mouse antibodies elicited in response to EBNA-1 cross-react with dsDNA. First, we showed that adsorption of sera reactive with EBNA-1 and dsDNA, on dsDNA cellulose columns, diminished reactivity with EBNA-1. Next, we generated mononclonal antibodies (MAbs) to EBNA-1 and showed, by several methods, that they also reacted with dsDNA. Examination of two cross-reactive MAbs—3D4, generated in this laboratory, and 0211, a commercial MAb—revealed that 3D4 recognizes the carboxyl region of EBNA-1, while 0211 recognizes both the amino and carboxyl regions. In addition, 0211 binds moderately well to the ribonucleoprotein, Sm, which has been reported by others to elicit a cross-reactive response with EBNA-1, while 3D4 binds only weakly to Sm. This suggests that the epitope in the carboxyl region may be more important for cross-reactivity with dsDNA while the epitope in the amino region may be more important for cross-reactivity with Sm.
Conclusions/Significance
In conclusion, our results demonstrate that antibodies to the EBNA-1 protein cross-react with dsDNA. This study is significant because it demonstrates a direct link between the viral antigen and the development of anti-dsDNA antibodies, which are the hallmark of SLE. Furthermore, it illustrates the crucial need to identify the epitopes in EBNA-1 responsible for this cross-reactivity so that therapeutic strategies can be designed to mask these regions from the immune system following EBV exposure. | We wish to thank Dr. Lori Frappier for supplying us with the recombinant EBNA-1 baculovirus. We also wish to thank Dr. David Fox Schechter for his advice and assistance in the isolation of the recombinant EBNA-1 protein. | CC BY | no | 2022-01-13 08:14:21 | PLoS One. 2011 Jan 4; 6(1):e14488 | oa_package/67/ac/PMC3014975.tar.gz |
||
PMC3014976 | 21245920 | Introduction
Mycobacterium ulcerans disease, also known as Buruli ulcer (BU), is the third most prevalent disease caused by mycobacteria [1] . It is characterized by deep and necrotizing skin ulcers with undermined edges resulting from the secretion by M. ulcerans of an immunosuppressive macrolide toxin, termed mycolactone [2] . It is predominantly found in scattered foci in tropical riverine and marshy regions throughout the world. In certain parts of Africa its prevalence may exceed 150/100,000 individuals [3] .
Until 2004, the recommended treatment for BU was surgical excision and skin grafting [1] . However, experimental studies using the mouse footpad model demonstrated that the combination of rifampin (RIF) and an aminoglycoside was bactericidal for M. ulcerans [4] , [5] , [6] , [7] . Based on these findings and subsequent studies in humans [8] , [9] the daily administration of the streptomycin-rifampin (STR+RIF) combination for 2 months was recommended by the World Health Organization (WHO) for the treatment of BU [10] . Depending on the size, severity and location of the ulcer, additional surgical intervention with skin grafting was also recommended. Treatment with STR requires intramuscular injection, which is difficult and expensive to implement in resource-poor countries since it requires use of sterile needles and syringes to avoid infection with blood borne pathogens. Therefore, the development of an entirely oral regimen is desirable [11] .
In vitro , M. ulcerans is susceptible to a limited number of oral antibiotics including fluoroquinolones and macrolides [12] , [13] , [14] , [15] , [16] . However, in the mouse model, the combination of clarithromycin (CLR), a bacteriostatic or weakly bactericidal drug against M. ulcerans , and RIF, which also has limited bactericidal activity [4] , [6] has not consistently shown efficacy similar to the standard STR+RIF regimen. Three mouse studies assessing the bactericidal activity and the relapse rate after treatment completion yielded conflicting results. In the first study [5] , the oral combination was less effective than the standard aminoglycoside plus RIF combination whereas in the second and third [12] , [13] studies, both combinations appeared to be as effective as the STR+RIF controls. There is, therefore, a need to directly address the issue of oral antibiotic treatment of BU, both with RIF and CLR and with new treatment regimens including anti-BU drugs that may have improved activity. Rifapentine (RPT), a rifamycin derivative with a much longer half-life than RIF could be an ideal substitute. In the murine model of tuberculosis when substituted for RIF at 10 mg/kg in a daily regimen in combination with isoniazid and pyrazinamide, it shortened the duration of treatment necessary to achieve cure [17] , [18] . Such regimens are currently under evaluation in at least 3 Phase II trials for tuberculosis treatment. In a murine model of M. ulcerans disease, daily RPT at the lower dose of 5 mg/kg has also been shown to be as active as, or even more active than, daily RIF at 10 mg/kg [13] .
In this study, we hypothesized that the use of daily RPT along with CLR would increase the efficacy of the rifamycin-CLR combination and help in the development of an entirely oral regimen for treatment of BU. We first demonstrated that there were no negative in vitro interactions of CLR and RIF (as a representative rifamycin) and then compared the efficacy of the RIF+CLR regimen to that of the RPT+CLR regimen using the STR+RIF standard regimen as control to determine whether daily RPT is a better substitute for daily RIF in the treatment of M. ulcerans disease in this murine model and whether daily RPT+CLR is also a better substitute for the standard daily STR+RIF combination. | Materials and Methods
Antimicrobials
STR and RIF were purchased from Sigma (St. Louis, MO) and RPT was a gift from sanofi-aventis pharmaceuticals (Paris, France). CLR was a gift from Abbott Laboratories (Abbott Park, U.S.A.). Stock solutions of RIF, RPT and CLR were prepared in sterile 0.05% agarose solution and STR was prepared in sterile normal saline. All stock solutions were prepared weekly and were stored at 4°C. All antimicrobials were administered orally (by gavage) using an esophageal cannula, except STR which was given by subcutaneous injections.
Bacterial strain
A recent isolate of M. ulcerans from a Ghanaian patient, strain Mu 1059 [19] provided by Dr. Pamela Small, was used for the study.
In vitro checkerboard study
To determine whether the RIF and CLR interaction is synergistic, indifferent or antagonistic, serial two-fold concentrations ranging from 0.125 to 2 μg/ml of both drugs alone and in combination were prepared in 7H11 agar+Oleic Acid-Albumin-Dextrose-Catalase (OADC) supplement. Eight-week-old colonies of Mu 1059 from 7H11 agar plates were suspended in phosphate-buffered saline (PBS), briefly vortexed, and kept undisturbed for 30 minutes to allow larger particles to settle. The optical density at 600 nm of this suspension was adjusted to 1, and 500 μl of the appropriate dilutions were plated in duplicate on antibiotic-containing plates and control plates without antibiotic. Plates were incubated at 32°C, and final CFU counts were performed after 12 weeks. The MIC was defined as the lowest drug concentration to inhibit growth of at least 99% of CFU on drug-free control plates. The fractional inhibitory concentration (FIC) value of individual drugs was then calculated using the MIC of the drug alone and MIC of the drug in combination. The sums of the two FIC values were combined to give the ΣFIC value which was then used to determine whether synergism (ΣFIC≤0.5), indifference ΣFIC (>0.5 to ≤4) or antagonism (ΣFIC>4) occurred between the antibacterial agents. All calculations were performed in accordance with current accepted standards [20] , [21] , [22] .
Inoculum preparation for the in vivo study
For each infection, an aliquot of a twice-mouse-passaged Mu 1059 strain stored at −80°C was thawed and inoculated in mouse footpads. Once the footpads were swollen to a lesion index of 2–3 (defined as inflammatory footpad/hind foot swelling) [5] , mice were sacrificed and footpad tissue was harvested, minced and suspended in sterile PBS. The solution was vortexed briefly, allowed to stand for 30 minutes, and the supernatant was used for footpad infection. Prior to infection, the inoculum was checked qualitatively for acid-fast bacilli, serially diluted, and plated for CFU counts on Middlebrook selective 7H11 plates (Becton-Dickinson, Sparks, MD).
Mouse model, infection and treatment
The kinetic method developed by Shepard for assessing the activity of anti-leprosy drugs [4] , [5] , [6] , [23] , was used to assess drug activity. In brief, 320 female BALB/c mice aged 4-to-6 weeks (Charles River, Wilmington, MA) were infected in the right hind footpad with 0.03 ml of the M. ulcerans suspension. After infection, mice were randomized to one of two control groups or one of six test groups. The control groups included untreated negative controls (n = 50), and mice treated with STR+RIF as positive controls (n = 55). The test groups included mice treated with each antibiotic alone, i.e., CLR (n = 30), STR (n = 25), RIF (n = 25), and RPT (n = 25), and the two-drug combinations RIF+CLR (n = 55) and RPT+CLR (n = 55). Ten mice from the untreated group were sacrificed the day after infection (D1) and 11 days later at treatment initiation (D11) to establish baseline CFU counts in the footpads. All mice were treated for 4 weeks, 5 days per week. The drugs were given at the following doses that are equivalent (similar AUC) to the human doses [25] , [26] : RIF 10 mg/kg, RPT 10 mg/kg, STR 150 mg/kg and CLR 100 mg/kg. On treatment completion, 5 mice from each group were sacrificed for quantitative CFU counts in the footpads and all of the remaining mice were kept without treatment to determine the time to footpad swelling.
For quantitative footpad CFU counts, each footpad was harvested after having been thoroughly disinfected with soap and sterile PBS followed by 70% alcohol swabs. The footpad tissue was homogenized by fine mincing and suspended in 2 ml sterile PBS. Appropriate dilutions were plated on selective 7H11 plates and incubated at 32°C for 12 weeks before CFU were enumerated. This study was carried out in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health. All animal procedures were approved by the Johns Hopkins Animal Care and Use Committee (protocol MO08M240) and conducted according to relevant national and international guidelines.
Assessment of treatment efficacy
The activity of each treatment was assessed in terms of CFU counts on treatment completion and median time to footpad swelling in treated mice compared with untreated control mice. CFU counts were performed by harvesting and homogenizing the footpad as described above and suspending each footpad in 2 ml PBS. Serial 10-fold dilutions were prepared and 0.5 ml of appropriate dilutions were plated in duplicate on 7H11 selective plates. The plates were then incubated at 32°C for 12 weeks before the CFU counts were made. Median time to footpad swelling was assessed by checking the footpads of mice every week for 39 weeks after infection. If the median time to footpad swelling in treated mice exceeded that in untreated mice by no more than the duration of the treatment, i.e. 4 weeks, then the treatment was considered to be bacteriostatic. Longer median time to swelling was indicative of bactericidal activity or prolonged post-antibiotic effect. Absence of swelling at the end of the follow-up period was indicative of sterilizing potential.
Pharmacokinetic studies of rifamycin-CLR combinations
Because we observed a negative antimicrobial interaction between both rifamycin derivatives and CLR in vivo , a series of single-dose pharmacokinetic (PK) studies were performed in BALB/c mice. In the first study mice were co-administered 100 mg/kg of CLR and RIF 10 mg/kg. One sample for PK analysis was collected per mouse at 1, 2, 4, 6, 9 or 16 hrs after dosing. From these data, composite concentration-time curves were developed and compared. Because RIF serum concentrations appeared to be diminished when the two drugs were dosed together, we conducted a second study in which mice received 10 mg/kg of RIF alone, 10 mg/kg of RIF followed by 100 mg/kg of CLR one hour later, or 10 mg/kg of RIF co-administered with a lower dose of CLR (10 mg/kg). In a third study we substituted RPT for RIF and assessed RPT serum concentrations in mice receiving 10 mg/kg of RPT alone, 10 mg/kg of RPT co-administered with 100 mg/kg of CLR, or 10 mg/kg of RPT followed 1 hour later by 100 mg/kg of CLR. In a fourth study we evaluated RPT serum concentrations after RPT 10 mg/kg was co-administered with CLR at 10 mg/kg. Serum samples were frozen at −80°C and shipped overnight on dry ice to the Infectious Disease Pharmacokinetics Laboratory, National Jewish Medical and Research Center, Denver, CO. Drug concentrations were determined using validated HPLC methods. PK parameters were calculated using non compartmental methods with Phoenix WinNonlin software, version 6.1.0 (Pharsight, Cary, NC).
Statistical analysis
Survival analysis, with footpad swelling as the measurement, was performed using the Kaplan-Meier method [27] . The log rank test was used to determine the level of statistical significance when comparing survival curves of the different treatment groups with the control group. p values were two-tailed, and a value of p <0.05 was considered statistically significant. CFU counts were log-transformed before analysis. Culture-negative footpads were assigned a log value of 0. Group means for experimental treatment groups were compared with that of the standard treatment control by one-way analysis of variance with Dunnett's post-test. Paired t-tests were also used to compare groups of equal size. All analyses were performed with GraphPad Prism version 4.01 (GraphPad, San Diego, CA). | Results
Checkerboard assay
The results of the checkerboard study are shown in table 1 . They indicated that the interaction (ΣFIC = 0.75) between the two drugs was neither synergistic nor antagonistic and therefore was termed indifferent using the current guidelines [20] , [21] , [22] .
CFU counts
The initial footpad suspension used for the inoculum contained 5.76 log 10 CFU per ml or 4.24 log 10 in the 0.03 ml that was inoculated per footpad. The next day (D1) 10 mice were sacrificed and the mean CFU count per footpad was 3.29±0.41 log 10 . On initiation of treatment, 11 days after infection (D11), the mean CFU count in the 10 mice that were sacrificed was 3.35±0.16 log 10 CFU, indicating that there was no substantial multiplication in the footpads during the first 11 days.
On treatment completion ( Figure 1 ), 4 weeks later, the mean log 10 CFU count was 5.01±0.62 in untreated control mice (W4 UT), demonstrating that M.ulcerans had multiplied well, increasing by about 2 log 10 in the footpads during the 4 weeks following treatment initiation, and suggesting a division time close to 4 days. In the positive control mice treated with STR+RIF, the mean log 10 CFU count was 0.76±0.52, with one footpad out of the 5 harvested footpads culture-negative, underscoring the potent bactericidal activity of the STR+RIF combination against actively multiplying M. ulcerans . Among the test mice, the mean log 10 CFU count was 3.54±0.18 in mice treated with CLR alone, a value similar to the 3.35±0.16 log 10 value on treatment initiation, confirming the bacteriostatic activity of CLR against actively multiplying M. ulcerans . For other antibiotics alone or in combination, the mean (including footpads with negative culture) log 10 CFU counts were significantly reduced ( p <0.01) compared to the baseline value ( Figure 1 ) but were not significantly different from each other except that mice treated with STR+RIF had a lower mean CFU count compared to RIF alone by paired t-test analysis ( p = 0.0335, though not significant after adjustment for multiple comparisons): 0.82±0.58 for STR alone (no CFU was isolated from 1 of the 5 mice); 1.33±0.24 for RIF alone and 1.37±1.15 for RIF+CLR (no CFU was isolated from 1 of the 5 mice); 0.48±0.56 for RPT alone (no CFU was isolated from 2 of the 5 mice) and 0.20±0.31 for RPT+CLR (no CFU was isolated from 3 of the 5 mice).
Time to footpad swelling
After completing 4 weeks of treatment, mice were monitored on a weekly basis for footpad swelling. Time to median swelling in untreated mice was 5 weeks after infection ( Figure 2 ). In accordance with the CFU counts on treatment completion, mice treated with CLR were the first to reach footpad swelling. But the median time to swelling was 16 weeks after infection, well beyond the 9 weeks that would have been expected after 4 weeks of treatment with a purely bacteriostatic drug added to the 5 weeks time to swelling in untreated control mice. CLR treatment is thus accompanied by a prolonged delay in footpad swelling possibly due to a significant post-antibiotic effect. Mice treated by CLR were followed by mice treated by RIF alone and STR alone, with median time to footpad swelling of 23.5 and 34 weeks, respectively. Only 26.3% of mice treated with RPT alone and 11.4% of the positive controls treated with STR+RIF developed footpad swelling at the end of the 8-month follow-up period after treatment completion, emphasizing the potent sterilizing effect of both regimens. The difference between RPT alone and STR+RIF was not statistically significant ( p = 0.33).
Surprisingly, as illustrated in Figure 2B , the time to footpad swelling was much shorter in mice treated with RIF+CLR ( p <0.008) or RPT+CLR ( p = 0.116) than in mice treated with RIF alone or RPT alone, respectively, suggesting antimicrobial or pharmacological antagonism between rifamycins and CLR in the mouse. Despite this antagonism, however, the RPT+CLR regimen caused a significantly greater ( p = 0.0007) delay in footpad swelling than did RIF+CLR. Because the checkerboard study did not reveal antagonism between the two antimicrobials, the antagonism was likely to be due to pharmacokinetic drug-drug interaction as demonstrated below.
Impact of CLR administration on rifamycin pharmacokinetics
In the first PK study, as in the second and third PK studies, the experiments were performed as single and first-dose assessments, RIF (10 mg/kg) was either administered alone or co-administered with CLR (100 mg/kg). The mean AUC 0–16h and Cmax of RIF were 118.63±18 μg*hr/ml and 11.97±1.3 μg/ml, respectively, when RIF was administered alone, and 92.5±27 μg*hr/ml and 8.48±0.54 μg/ml, respectively, when RIF was co-administered with CLR ( Figure 3A ), suggesting that CLR co-administration led to diminished RIF concentrations. In a second PK study, RIF (10 mg/kg) was administered alone, with CLR (100 mg/kg) given 1 hr later, or co-administered together with CLR at a lower dose of 10 mg/kg. The mean AUC 0–21h and Cmax of RIF were 123±21 μg*hr/ml and 15.7±4.2 μg/ml, respectively, when RIF was administered alone, 133±32 μg*hr/ml and 16.7±2.5 μg/ml, respectively, when CLR (100 mg/kg) was administered 1 hr after RIF, and 125±22 μg*hr/ml and 15.5±1.2 μg/ml, respectively when RIF and CLR were co-administered at an equal dose of 10 mg/kg ( Figure 3B ).
Similar observations were made in mice given RPT and CLR. The AUC 0–24h and Cmax of RPT were moderately decreased from 317.24±25 μg*hr/ml and 18.11±1.0 μg/ml, respectively, when RPT was given alone to 241.09±0.37 μg*hr/ml and 13.07±1.6μg/ml, respectively, when RPT was co-administered with CLR (100 mg/kg). Delaying administration of CLR (100mg/kg) by 1 hr resulted in a RPT AUC 0–24h of 279.70±0.47 μg*hr/ml and Cmax of 15.14±1.7 μg/ml ( figure 4A ). Reducing the dose of CLR to 10 mg/kg when co-administered with 10 mg/kg RPT resulted in RPT AUC 0–24h and Cmax values of 302.51±28 μg*hr/ml and 17.52±1.1 μg/ml, respectively, similar to those of 324.23±44 μg*hr/ml and 19.24±1.6 μg/ml, respectively, obtained with RPT alone ( Figure 4B ). | Discussion
The main result of the present work is that RPT alone administered 5 days a week at a dose of 10mg/kg is at least as active in terms of bactericidal effect and as active in terms of relapse prevention as the standard combination of STR+RIF against experimental M.ulcerans disease in the mouse. Such a result is extremely promising for the future of M.ulcerans disease treatment because it suggests that an entirely oral treatment may be as active as the present regimen containing parenteral STR. However, RPT cannot be administered alone because of the risk of drug resistance resulting from monotherapy, and CLR is the oral companion drug of choice to combine with a rifamycin [28] . As CLR alone exhibited clear-cut bacteriostatic activity, an additive effect of the combination RPT+CLR and even RIF+CLR was expected. Unfortunately the co-administration of a rifamycin and CLR, both drugs given orally at doses equivalent to human doses on the basis of serum AUC, was less effective than each rifamycin alone in mice infected with M. ulcerans . As no antagonistic effect between RIF and CLR was exhibited in vitro in the checkerboard assay, the lesser in vivo effectiveness of the combination could not be related to a negative antimicrobial drug-drug interaction. Rather, it appears that the negative drug-drug interaction was pharmacokinetic in nature. Indeed, co-administration of a 10-mg/kg dose of RIF with a 100 mg/kg dose of CLR resulted in a 22% reduction of RIF AUC and a 29% reduction of RIF Cmax compared to administration of RIF alone. Similarly, RPT 10 mg/kg given together with 100 mg/kg CLR resulted in a 24% and 28% reduction of the RPT AUC and Cmax, respectively, compared to RPT administered alone. When administration of 100mg/kg of CLR was delayed by one hour from RIF administration, the pharmacokinetic interaction became insignificant. Similarly, there were also no negative pharmacokinetic drug-drug interactions when mice were co-administered 10 mg/kg of RIF and 10 mg/kg of CLR. These results indicate that the co-administration of CLR and RIF negatively interacts with the blood levels of rifamycins in mice probably by interfering with their absorption or by another mechanism and that this drug interaction is dose-dependent. However, a recent study in humans showed that concomitant CLR did not impact the absorption rate constant, the Cmax, or the Tmax of RIF at steady state, indicating that, at clinically relevant doses (7.5 mg/kg of CLR and 10 mg/kg of RIF), CLR does not negatively affect the levels of rifamycins in humans [29] .
Our findings illustrate the difficulties in designing experiments in the murine model that aim to instruct treatment of a human infectious disease and in interpreting their results.
In order to adequately assess in mice the antimicrobial potential of a given drug, that drug should be given at doses deemed equivalent to human doses. As, drugs are usually metabolized much more rapidly in mice than in humans, the drug doses in mice have to be increased to obtain similar drug exposure in mice as in humans [25] , [30] . That is the case for CLR [30] , [31] . But the dose of 100 mg/kg that is adequate in mice to assess the antimicrobial activity of CLR when the drug is used alone presents a problem when it is co-administered with 10 mg/kg of RIF, most likely by reducing the absorption of RIF. Therefore the fact that combinations of CLR and a rifamycin were less active than the corresponding rifamycin alone should be considered an experimental artifact, and both drugs should be administered separately, with an interval of no less than one hour between them. Interestingly, the same phenomenon is observed in the experimental chemotherapy of tuberculosis for which RIF should be administered at least one hour before isoniazid and pyrazinamide to prevent a negative pharmacokinetic interaction in mice [32] , [33] .
Although the negative pharmacokinetic interactions prevented a reliable assessment of the antimicrobial activity of the RIF+CLR and RPT+CLR combinations against experimental M.ulcerans disease in mice, they did not prevent assessment of each drug alone in reference to the positive controls receiving STR+RIF. Besides the promising potency of RPT, our study also emphasizes the peculiar prolonged delay in footpad swelling resulting from treatment with CLR in mice infected with M. ulcerans . Whatever its antimicrobial or immunomodulatory [34] nature, this delay in footpad swelling is favorable and supports the use of CLR in the treatment of Buruli ulcer.
Finally, it is important to note that, in our experimental model, drug activity was assessed during a 4-week period during which untreated animals had a 2 log 10 increase in CFU counts in their footpads. As expected, CLR exhibited bacteriostatic activity whereas other drugs exhibited bactericidal activity, especially RPT. But, very interestingly, even the most active drugs and drug regimens did not reduce the CFU counts by more than 3 log 10 in 4 weeks. This was much less than the 5–6 log 10 reduction in the CFU counts observed by Ji et al. [12] , [13] , [14] when treatment was initiated at the plateau phase of growth, i.e., when the organisms were no longer actively multiplying likely because of immune containment. In our experimental model because antibiotic treatment was initiated during the incubation phase of the disease, the reduction in the CFU counts and the time to foot pad swelling are measuring only the antimicrobial activity. When treatment is initiated at the plateau phase of growth, its effect is likely a mixture of antimicrobial activity, immune containment, and shutting down the enzymes involved in mycolactone production. It does not facilitate the assessment of the respective antimicrobial value of each drug regimen, even though it might better recapitulate the response of patients to antibiotic therapy. In the chemotherapy of BU, as in the chemotherapy of tuberculosis, the drug activity against actively multiplying organisms, usually termed bactericidal activity, is very different from the drug activity against organisms that are no longer actively multiplying, i.e., sterilizing activity. Consequently, the occurrence, magnitude, and duration of the antimicrobial effect depend on the experimental model used. The information provided by each model is therefore different yet complementary, and not at all contradictory. | Conceived and designed the experiments: DA PJC ELN JHG. Performed the experiments: DA PJC ZA. Analyzed the data: DA PJC KED ELN JHG. Contributed reagents/materials/analysis tools: DA KED ELN JHG. Wrote the paper: DA JHG.
Background
Treatment of Mycobacterium ulcerans disease, or Buruli ulcer (BU), has shifted from surgery to treatment with streptomycin(STR)+rifampin(RIF) since 2004 based on studies in a mouse model and clinical trials. We tested two entirely oral regimens for BU treatment, rifampin(RIF)+clarithromycin(CLR) and rifapentine(RPT)+clarithromycin(CLR) in the mouse model.
Methodology/Principal Findings
BALB/c mice were infected in the right hind footpad with M. ulcerans strain 1059 and treated daily (5 days/week) for 4 weeks, beginning 11 days after infection. Treatment groups included an untreated control, STR+RIF as a positive control, and test regimens of RIF, RPT, STR and CLR given alone and the RIF+CLR and RPT+CLR combinations. The relative efficacy of the drug treatments was compared on the basis of footpad CFU counts and median time to footpad swelling. Except for CLR, which was bacteriostatic, treatment with all other drugs reduced CFU counts by approximately 2 or 3 log 10 . Median time to footpad swelling after infection was 5.5, 16, 17, 23.5 and 36.5 weeks in mice receiving no treatment, CLR alone, RIF+CLR, RIF alone, and STR alone, respectively. At the end of follow-up, 39 weeks after infection, only 48%, 26.4% and 16.3% of mice treated with RPT+CLR, RPT alone and STR+RIF had developed swollen footpads. An in vitro checkerboard assay showed the interaction of CLR and RIF to be indifferent. However, in mice, co-administration with CLR resulted in a roughly 25% decrease in the maximal serum concentration (Cmax) and area under the serum concentration-time curve (AUC) of each rifamycin. Delaying the administration of CLR by one hour restored Cmax and AUC values of RIF to levels obtained with RIF alone.
Conclusions/Significance
These results suggest that an entirely oral daily regimen of RPT+CLR may be at least as effective as the currently recommended combination of injected STR+oral RIF.
Author Summary
Buruli ulcer (BU) is found throughout the world but is particularly prevalent in West Africa. Until 2004, treatment for this disfiguring disease was surgical excision followed by skin grafting, procedures often requiring months of hospitalization. More recently, an 8-week regimen of oral rifampin and streptomycin administered by injection has become the standard of care recommended by the World Health Organization. However, daily injections require sterile needles and syringes to prevent spread of blood borne pathogens and streptomycin has potentially serious side effects, most notably hearing loss. We tested an entirely oral regimen, substituting the long acting rifapentine for rifampin and clarithromycin for streptomycin. We also evaluated each drug separately. We found that rifapentine alone is as good as rifampin plus streptomycin, but the simultaneous addition of effective clarithromycin doses, at least in the mouse, reduces the activity of both rifampin and rifapentine, making it difficult to assess the efficacy of the oral regimens in the model. Studies of serum drug concentrations indicated that separating treatment times by one hour or reducing the clarithromycin dose to one active in humans should overcome this issue in experimental and clinical BU treatment, respectively. | We thank Dr. Pamela Small, University of Tennessee, for providing us with the M. ulcerans 1059 strain, and Dr. Charles Peloquin for determination of rifamycin concentrations in serum. | CC BY | no | 2022-01-13 05:08:15 | PLoS Negl Trop Dis. 2011 Jan 4; 5(1):e933 | oa_package/cf/ac/PMC3014976.tar.gz |
||
PMC3014977 | 21245921 | Introduction
Dengue viruses (DEN) are arthropod-borne flaviviruses that cause dengue fever (DF) with significant morbidity and mortality in tropical and subtropical regions of the world. There are four serotypes of dengue viruses (DEN types 1–4). Classic DF is a self-limited illness characterized by fever, headache, myalgia, arthralgia, and abdominal pain. Since the 1950s, a more severe form of the disease, dengue hemorrhagic fever (DHF), has been recognized worldwide [1] . Patients who develop DHF typically have initial symptoms similar to those in DF patients, but develop cytokine-storm-like plasma leakage manifested by hemoconcentration, thrombocytopenia, ascites, and pleural effusion near the defeverence stage [2] . DHF pathogenesis has been attributed to viral virulence versus immune enhancement; however, that has been the subject of debate for many years [2] , [3] .
The innate immune system is the first line of host defense against pathogens and is involved in early recognition and uptake of microbes by the host's professional phagocytes such as dendritic cells (DCs) and macrophages, through germline-encoded receptors, known as pattern recognition receptors (PRRs) [4] . These PRRs recognize microbial antigens and initiate innate immune responses followed by adaptive immunity [5] . PRRs are involved in phagocytosis, antigen presentation, and they trigger intracellular signaling and cytokine secretion [5] . PRR polymorphisms may therefore affect virus entry, replication, and immunity.
Among the PRRs, the CD209 molecule, also known as DC-SIGN (dendritic cell-specific intercellular adhesion molecule-3 grabbing non-integrin), plays an important role in the early interaction of a pathogen with a dendritic cell [6] – [8] and has a key role in DC-T cell interaction [9] , DC migration [10] , [11] , and pathogen uptake [12] . DC-SIGN is organized into three domains, the N-terminal domain is located in the cytoplasm, the transmembrane domain anchors to the cytoplasmic membrane, and the extracellular domain consists of a neck region formed by seven highly conserved 23 amino acid repeats and a carbohydrate domain for pathogen binding [13] .
The CD209 gene is located on chromosome 19p13.2–3 and is highly polymorphic. Numerous single nucleotide polymorphisms (SNPs) have been reported [14] – [18] . One of these SNPs represents a guanine (G) to adenine (A) transition at position −336 within the CD209 gene promoter (rs4804803). This variant has been associated with an increased risk for parenteral acquisition of human immunodeficiency virus (HIV) infection [15] , severity of dengue infection [16] , and confered high susceptibility to tuberculosis in a South African cohort [17] . Nevertheless, Vannberg et al. found that G variant allele of rs4804803 was associated with protection against tuberculosis in individuals from sub-Saharan Africa [18] . This variant affects CD209 promoter activity with multiple transcription factor binding sites for the Sp1/GATA1/CACCC- and CAC-binding transcription factors in a transfection study [16] . As an in vitro study of promoter activity might not reflect an actual functional association, we aimed to test whether the rs4804803 SNP in the promoter region of CD209 was associated with the susceptibility to DF and/or DHF in Taiwanese, and whether monocyte-derived DCs from humans with various genotypes of rs4804803 would reveal differences in DC-SIGN membrane expression and implicate the viral replication and immune reactions after DEN infection. | Materials and Methods
Ethics statement and subjects studied
This study was approved by the Institution Review Board (IRB) of Chang Gung Memorial Hospital-Kaohsiung Medical Center, Taiwan. The dengue patients were recruited as described previously in the 2002–2003 DEN-2 outbreak in Taiwan [19] – [23] . A larger retrospective cohort was designed and re-approved by an additional IRB review (Document No.: 97-2111B). To validate cell surface expression and immune functions of rs4804803 SNP, we obtained informed consent to collect blood leukocytes from normal volunteers with AA or AG genotypes of rs4804803.
Definition of cases and controls
DEN infection was confirmed by clinical dengue symptoms and signs along with detection of DEN-2 RNA by quantitative RT-PCR in blood, detection of IgM to DEN or at least a 4-fold increase in dengue-specific hemagglutination inhibition titers in convalescent serum compared with that in acute-phase serum [20] , [21] . In those with DEN-2 infection, blood was drawn at least once a day subsequent to admission into the hospital to measure the platelet counts and hematocrit levels. A Chest X-ray and abdominal ultrasonography were performed routinely in individuals without evidence of hemoconcentration or hypoalbuminemia to refine the differential diagnosis of DHF vs. DF based on pleural effusion or ascites. A clinical diagnosis of DHF was assigned according to the DHF criteria of the World Health Organization (WHO); including a reduced platelet count (<100,000/mm 3 ), petechiae, hemorrhagic manifestation, and plasma leakage showing hemoconcentration (peak hematocrit ≥20% above the mean for the population, or an increase in hematocrit of 20% or more), pleural effusion, ascites, or hypoalbuminemia [24] . Patients with DF were defined by detectable DEN-2 RNA by RT-PCR or DEN-specific IgM, but without evidence of DHF. Primary or secondary DEN infections were identified using previously established serologic criteria for IgM/IgG ELISAs [19] .
Patients with other non-dengue febrile illnesses (OFI) were defined by febrile illness with no detectable DEN-specific IgM, no detectable DEN RNA, and no obvious or reported bacterial etiology for their illness during the same study period. Population controls were healthy, unrelated volunteers from the same community, with neither signs nor previous history of dengue infection, with a DEN IgG sero-positive rate of 1.37% (1/73).
Genotyping of CD209 rs4804803 SNP
Genomic DNA was isolated from heparin-anticoagulated blood samples using a standard phenol-chloroform extraction followed by 70% alcohol precipitation. Genotyping for the CD209 variant (−336 A/G; rs4804803) was carried out using Custom TaqMan SNP Genotyping Assays (Applied Biosystems, Foster City, CA, USA). The primer sequences were 5′-GGACAGTGCTTCCAGGAACT-3′ (forward) and 5′-TGTGTTACACCCCCTCCACTAG-3′ (reverse). The TaqMan minor groove binder probe sequences were 5′-TACCTGCCTACCCTT G-3′ , and 5′-CTGCCCACC CTTG-3′ . The probes were labeled with the TaqMan fluorescent dyes VIC and FAM, respectively. The PCR was conducted in total volume of 15 μL using the following amplification protocol: denaturation at 95°C for 10 min, followed by 40 cycles of denaturation at 94°C for 20 s, followed by annealing and extension at 60°C for one minute. After the PCR, the genotype of each sample was determined by measuring the allele-specific fluorescence in the ABI Prism 7500 Sequence Detection System, using SDS 1.1 software for allele discrimination (both Applied Biosystems). To validate the genotyping by real-time PCR analysis, 100 PCR products were subject to restriction fragment length polymorphism (RFLP) analysis with MscI restriction enzyme (New England Biolabs, Beverly, MA, USA) and showed 100% identical result between these two genotyping systems.
Generation of DCs from individuals with AA or AG phenotype of rs4804803
Peripheral blood mononuclear cells were collected from peripheral blood of 20 healthy, DEN-specific IgM or IgG seronegative volunteers with AA or AG genotype. CD14 + monocytes were isolated by positive selection according to the manufacturer's specifications using CD14 microbeads and a magnetic cell separator (MACS) (Miltenyi Biotec, Bergisch Gladbach, Germany). Enriched CD14 + cells (purity>95%) were cultured for 6 days in six-well plates in complete RPMI medium in the presence of 10 ng/mL rhGM-CSF and 5 ng/mL rhIL-4 at 37°C, and 5% CO 2. On day 3, half of the medium was replaced with fresh medium supplemented with rhGM-CSF and rhIL-4. Expression of markers was measured by flow cytometer using specific antibodies and their corresponding isotype controls.
DCs infection with DEN-2 and endogenous IP-10 neutralizing experiments
Unless otherwise stated, monocyte-derived dendritic cells (MDDCs) were infected with DEN-2 at a multiplicity of infection (MOI) of 5 for 2 h at 37°C and 5% CO 2 . Cells were washed twice to remove cell-free virus, and cultured in complete RPMI medium (without cytokines) at a density of 2×10 5 cells/ml in 48-well plates. Cells and supernatants were removed and analyzed at 24, 48, and 72 h post-infection. For the neutralization experiments, cells were incubated in the medium alone or in the medium with the addition of anti-human CXCL10/IP-10 antibody (R&D Systems, Minneapolis, MN, USA) at 10 μg/mL for 30 min. Cells and supernatants were harvested and analyzed 24 h post-infection.
Detection of CD209 mRNA by real time RT-PCR and cell surface expression on MDDCs by flow cytometry
Total RNA extracted from MDDCs was subjected to quantitative RT-PCR to assess levels of mRNA corresponding to CD209 and ß2–microglobulin ( B2MG ) using the ABI PRISM 7500 instrument (Applied Biosystems). The forward primer, reverse primer sequence for detecting CD209 and B2MG were 5′-AACAGCTGAGAGGCCTTGGA-3′ , 5′-GGGACCATGGCCAAGACA-3′ , and 5′-AATTGAAAAAGTGGAGCATTCAGA-3′ , 5′-GGCTGTGACAAAGTCACATGGTT-3′ , respectively. The PCR cycling parameters were 40 cycles of PCR reactions at 94°C for 20 s, and 60°C for one minute. The results were detected in real-time and recorded on a plot showing fluorescence vs. time. RT-PCR products were also visualized on ethidium bromide-stained 1.5% agarose (Pierce Co., Rockford, IL, USA) gel with a 100- bp ladder (Pharmacia Biotech, Piscataway, NJ, USA) as a reference.
To measure the CD209 cell surface expression, MDDCs were stained with FITC-conjugated mAbs specific for DC-SIGN (R&D Systems, Minneapolis, MN, USA). An isotype-matched FITC-labeled control, mouse IgG2b (clone MOPC195, Immunotech, Beckman Coulter, Fullerton, CA, USA) was included in each experiment.
Quantitation of viral replication in MDDCs by real time RT-PCR
Total RNA extracted from MDDCs was subjected to assess DEN-2 RNA viral copies. Fluorescent RT-PCR was performed in an ABI 7500 quantitative PCR machine (Applied Biosystems) for 40 cycles using TaqMan technology as previously described [21] .
Detection of cytokine/chemokine production by ELISA
Cytokine/chemokine production and viral replication were determined at 24, 48, and 72 h post-infection. Cell-free culture supernatants TNFα and macrophage chemoattractant protein 1 (MCP-1) concentrations were measured using ELISA kits from eBioscience Inc. (San Diego, CA, USA); IL-12p40 and IFN-inducible protein 10 (IP-10) concentrations were measured using ELISA kits from R&D Systems as per manufacturer's instructions.
Statistical Analyses
Data are presented as mean ± SEM values. Alleles and genotypes distribution of rs4804803 are presented as numbers (percentages). Conformance of the allele frequencies to Hardy-Weinberg equilibrium proportions was tested to compare the observed and expected frequencies of heterozygotes and homozygotes. Differences among patients with DEN, DF, DHF, OFI, and population controls were determined using two-sided Chi-Square test or Fisher exact test. Odds ratio (OR) values were calculated with 95% confidence intervals (CI). The Student's t-test or Mann-Whitney U test was used for statistical comparisons between continuous variables. The Wilcoxon signed-rank test was used for statistical comparison of the neutralization experiments. All analyses were performed using SPSS 13.0 (SPSS Inc. Chicago, IL, USA). | Results
Demographic characteristics of patients with DF and DHF
During a large DEN-2 outbreak in southern Taiwan between June 2002 and January 2003, a hospital-based case-control study was used to identify the risk immune parameters [19] – [23] . Employing the decoded DNA samples from that same cohort of the population that study has been extended to investigate the association of rs4804803 SNP with DF, DHF, viral replication, and immune response. Based on the previous case-control study design, we have included DNA samples from 135 DHF, 176 DF, and 143 OFI patients in this expanded study. The main characteristics of the study population are summarized in Table 1 . There were no significant differences in sex or total leukocyte counts between patients with DF and those with DHF. However, age (41.7±1.6 years vs. 45.7±1.3 years, p <0.001), serum GOT levels (70.1±8.1 U/mL vs. 313.8±74.6 U/mL, p = 0.002) and GPT levels (67.1±11.3 U/mL vs. 142.7±21.9 U/mL, p = 0.003) were significantly higher in the DHF group ( Table 1 ). A patient manifested with abdominal pain had ascites as evidenced by abdominal ultrasonography was classified as DF because the patient revealed no thrombocytopenia (<100,000/mm 3 ), petechia or hemorrhagic manifestation during the admission period.
CD209 −336 A/G polymorphisms (rs4804803) associated with dengue infection and DHF susceptibility
We investigated the association of rs4804803 SNP in the promoter region of CD209 with protection from dengue infection and the susceptibility of DHF. Genomic DNA obtained from DEN patients (n = 311), OFI patients (n = 143), and population controls (n = 120) was genotyped for rs4804803 SNP. We found that GG/AG genotypes in 16.0% of the DEN patients were significantly higher than OFI patients (5.6%, OR = 3.23, p = 0.002) and population controls (7.5%, OR = 2.36, p = 0.020; Table 2 ). Moreover, the GG/AG genotypes were significantly higher in DHF patients (23.0%) than OFI patients and population controls (OR = 5.03 and 3.68, p = 3×10 −5 and 0.001), and also significantly higher than DF patients (10.8%, OR = 2.46; p = 0.004; Table 2 ).
Analysis of the allele distribution between DEN and OFI patients or population controls showed that the G allele frequency was higher in DEN patients (8.4%), compared with OFI patients (2.8%, OR = 3.17, p = 0.002) or population controls (3.8%, OR = 2.34, p = 0.018; Table 3 ). Moreover, the frequency of G allele of rs4804803 was significantly higher in DHF patients (12.2%) than OFI patients or population controls (OR = 4.84 and 3.57, p = 2×10 −5 and 0.001), and higher than DF patients (5.4%, OR = 2.44, p = 0.002; Table 3 ). Few DHF patients (n = 6) had dengue shock syndrome in this cohort; one of them carrying AG genotype.
No association of CD209 −336 A/G polymorphism (rs4804803) with primary and secondary dengue infection
To investigate whether the rs4804803 SNP in the promoter region of CD209 associated with the primary and secondary DEN infection, we used serological methods to detect DEN antibodies for differentiation into primary and secondary dengue infection. Of the 293 DEN patients, 141 (48%) had secondary DEN infections, based on detectable DEN-2 virus RNA and DEN IgG. As shown in Table 1 , secondary DEN-2 infection was more frequently found in patients with DHF than in those with DF (65% vs. 36%, p <0.001). We found the rs4804803 GG/AG genotypes were found in 12.5% of patients with primary DEN infection and 16.3% of patients with secondary DEN infection, which did not reach significantly different (OR = 1.36; p = 0.352). As shown in Table 4 , there was no association between rs4804803 SNP and primary or secondary dengue infection in DF patients (OR = 0.64; p = 0.409), or in DHF patients (OR = 1.80; p = 0.251). In addition, there was no association between allele distribution and primary or secondary dengue infection in DF and DHF patients (data not shown).
Increase of DC-SIGN expression on MDDCs from individuals with AG genotype of rs4804803
Due to the low frequency of GG genotype in our population (2 cases, 0.6%), we could not recall the patients because the data file had been decoded for identification. We examined DC-SIGN (CD209) expression in both mRNA level of MDDCs and protein level on their cell surface from healthy subjects with AA or AG genotype by quantitative RT-PCR and flow cytometry, respectively. A significant increase in CD209 mRNA expression was detected in the MDDCs from individuals with AG genotype than those from individuals with AA genotype ( p = 0.032, Fig 1B ). Similarly, individuals with AG genotype had a significantly higher cell surface DC-SIGN expression ( p = 0.029; Fig 1D ). However, the surface DC-SIGN expression declined rapidly along with DEN-2 infection on MDDCs from both genotypes' subjects, which showed no difference at 24, 48, and 72 h post-infection ( Fig. 2A ).
Lower viral replication in MDDCs from individuals with AG genotype of rs4804803
To investigate whether the rs4804803 SNP was correlated to viral replication, we measured DEN-2 RNA copies in MDDCs with AA or AG genotype of rs4804803 at 24, 48, and 72 h post-infection. DEN-2 replication was significantly higher in MDDCs from individuals with AA genotype than those with AG genotype at 48 h post-infection (1.07±0.45×10 6 copies/10 5 cells vs. 3.90±0.67×10 6 copies/10 5 cells, p = 0.006) and 72 h post-infection (4.83±0.70×10 5 copies/10 5 cells vs. 2.32±0.68×10 6 copies/10 5 cells, p = 0.003; Fig. 2B ). Viral replication, as measured at 72 h post-infection, increased more remarkably in MDDCs at MOI of 5 and 10 ( p <0.001 and 0.002, respectively; Fig. 2C ).
Elevation of cytokines TNFα, IL-12p40, and chemokine IP-10 production in DEN-infected MDDCs with AG genotype of rs4804803
To investigate whether higher cell surface DC-SIGN expression was correlated with immune response, we investigated kinetic cytokine/chemokine production by MDDCs from individuals with AA or AG genotype of rs4804803. Results showed that MDDCs with AG genotype had significantly higher TNFα production than those with AA genotype at 24 and 48 hr post-infection (303.51±66.75 pg/mL vs. 143.97±68.80 pg/mL and 202.35±19.35 pg/mL vs. 73.00±9.55 pg/mL; p = 0.021 and 0.002, respectively; Fig 3A ). IL-12p40 production significantly increased by MDDCs with AG genotype than those with AA genotype at 24, 48, and 72 h post-infection ( p <0.001, 0.007 and 0.001, respectively; Fig 3B ). We also measured the concentration of two chemokines, MCP-1 and IP-10, which had been implicated in the recruitment and stimulation of monocytes, macrophages, dendritic cells, NK cells, and T lymphocytes [25] . It was found that IP-10, but not MCP-1, production was significantly higher by MDDCs with AG genotype than those with AA genotype at 24, 48, and 72 h post-infection (620.60±175.56 pg/mL vs. 243.02±41.64 pg/mL, 889.92±91.46 pg/ml vs. 168.02±24.02 pg/mL, and 614.44±49.16 pg/mL vs. 322.32±69.62 pg/mL; p = 0.034, 0.009 and 0.010, respectively; Fig 3C ). The MCP-1 levels peaked at 48 hr in subjects with both genotypes', but there was no significant difference between AA genotype and AG genotype (550.72±60.73 pg/mL vs. 463.92±66.80 pg/mL, p = 0.157; Fig. 3D ).
IP-10 production by MDDCs involved in viral replication of DEN infection
IP-10, produced by non-infected bystander DCs in response to DEN infection, is a potent chemoattractant for activated T and NK cells [26] , and the modulation of adaptive immune response [27] . IP-10 has also been known to inhibit the binding ability of DEN in immortalized cells [28] . In our MDDC model, cells from individuals with AG genotype exhibited an augmented innate immune reaction, showing higher IP-10 production, post-infection ( Figure 3C ). Based on these results, we hypothesized that DEN-infected MDDCs with AG genotype produced higher levels of IP-10, which might block viral entry or viral replication in MDDCs. We used an anti-IP-10 neutralizing mAb to block endogenous IP-10 production by MDDCs. With both genotypes, the viral replication 24 h post DEN infection increased significantly more in the presence of neutralizing antibody than in the absence of neutralizing antibody ( p = 0.034 and 0.040, respectively; Fig. 4A ). IP-10 production by MDDCs from individuals with AG genotype significantly decreased (795.3±368.1 pg/mL vs. 273.8±87.8 pg/mL; p = 0.037), but it did not decrease in MDDCs from individuals with AA genotype (329.8±114.2 pg/mL vs. 201.8±87.0 pg/mL; p = 0.091; Fig. 4B ). These results suggest that IP-10 produced by MDDCs is involved in the viral replication of DEN infection. | Discussion
DC-SIGN has been shown to be an important receptor for DEN and a number of viruses, including HIV, Helicobacter pylori , and Mycobacterium tuberculosis and hepatitis C virus (HCV) [29] . Some studies have demonstrated that genetic variations of CD209 (rs4804803) were associated with the susceptibility to HIV [14] , Mycobacterium tuberculosis [17] , HCV [30] , and dengue [16] . Few studies have demonstrated how the rs4804803 SNP is involved in viral replication or immune response. We are the first in the field to demonstrate the relationship among functional cell surface expression, viral replication, and immune responses in DEN-infected MDDCs from subjects with rs4804803 SNP. Here we found that rs4804803 SNP was strongly associated with the risk of DHF vs. DF and controls. Functional studies have determined that MDDCs from individuals with AG genotype have a significantly higher cell surface DC-SIGN expression than from those with AA genotype. MDDCs with AG genotype produced higher TNFα, IL-12p40, and IP-10 levels but lower viral replication in response to dengue infection.
Because the physiopathology of various manifestations of DHF is not fully understood, several studies have supported the supposition that secondary dengue infection [31] , age [32] , a number of preexisting chronic diseases such as diabetes and bronchial asthma [33] , and host genetic factor [16] , [22] increase the risk of progression to DHF. This indicates that multiple factors are involved in the development of DHF. Our findings regarding rs4804806 SNP associated with DHF vs. OFI control ( p = 3×10 −5 ; Table 2 ) in a case-control association study suggests that rs4804806 SNP contributes in part to the development of DHF.
Our study shows that the GG/AG genotypes of rs4804803 were associated with susceptibility to DHF, compared with DF, which is consistent with the observation of Sakuntabhai et al. [16] . In our study, the AA genotype was associated with protection against DHF, compared with OFI and population controls, while G allele was associated with protection against DF in Sakuntabhai's observation. The inconsistency between these studies regarding the protection for DHF or DF may result from two possibilities. First, the frequency of G allele in Chinese population is 3.8%; while in Thailand, it is 9.5–10.4% [16] , [34] . Second, definition of DF and DHF might be also different. We defined DF and DHF according to WHO criteria, while in the study by Sakuntabhai et al. , DF was defined by criteria of severe dengue fever syndrome with hemorrhage but no plasma leakage, excluding patients with flu-like symptoms or those having only fever. Moreover, the rs4804803 SNP was demonstrated to be in linkage disequilibrium with three other intronic polymorphisms in a Thai population, and these might also have contributed to the susceptibility of DHF [16] .
Our results suggest that humans carrying the rs4804803 AG genotype have a higher DC-SIGN expression and lower DEN-2 replication in MDDCs. These results differ from a previous study by Loach et al. who demonstrated that the DC-SIGN expression levels on Raji cells after transfection of various DC-SIGN cDNA constructs were significantly correlated to the infection rate of DEN-1 [35] . DC-SIGN is an endocytic receptor shown to induce endocytosis of several pathogens, including dengue [36] – [38] . The difference between these two studies might be due to different cell types and ex vivo culture systems. In our study, it was found that MDDCs from subjects with rs4804803 AG genotype had higher surface DC-SIGN expression with higher production of chemokines such as IP-10, which could limit DEN-2 replication ( Fig. 4A ); however, the higher surface DC-SIGN expression in subjects with AG genotype decreased remarkably 24 h post-infection ( Fig. 2A ). In the study by Loach et al. , ectopic expression levels of DC-SIGN on Raji cells enhanced DEN-1 replication, which might be related to a higher quantity of receptors or lower production of IP-10 favoring DEN replication. Results from these studies suggest that the correlation of viral replication to higher or lower DC-SIGN expression depends on genetic factors in the host, cell type, and dynamic changes in the receptor following DEN infection.
In functional studies of rs4804803 SNP, we determined that MDDCs with AG genotype had a higher DC-SIGN expression correlated to augmented immune responses with higher TNFα, IL-12p40, and IP-10, than those with AA genotype, but not MCP-1 production. DEN replication was significantly lower in individuals with AG genotype. The addition of anti-IP-10 neutralizing antibody blocked the production of endogenous IP-10 and significantly enhanced the replication of DEN-2 ( Fig. 4A ). This suggests that rs4804803 SNP was involved in the DC-SIGN expression associated with augmented immune response, such as the increase in the production of IP-10 that repressed the replication of DEN. This is supported by the fact that altered immune response, but not viral load, was observed in DHF patients [21] , [39] . CLEC5A-mediated DEN infection in animals that was susceptible to DEN hemorrhagic infection also revealed augmented immune response [40] .
In contrast, it is interesting to note that the viral replication in MDDCs from individuals with rs4804803 AA genotype was significantly higher than in individuals with AG genotype following DEN-2 infection. The mechanism by which rs4804803 SNP influences DEN replication in MDDCs is currently unknown. Chan et al. showed that certain polymorphisms of L-SIGN , a DC-SIGN homologue, mediated more efficient viral degradation of SARS-CoV [41] . The clinical implications of screening genotypes to prevent DEN infection might be supported if different viral loads could be demonstrated among humans with various genotypes of rs4804803 in future outbreaks of DEN.
The outcome of DEN infection is determined by a myriad of interactions among viral, immunological, and human genetic factors, as well as kinetic interactions between innate and adaptive immunity. This study provides new evidence that CD209 rs4804803 SNP, correlated to cell surface expression on dendritic cells, mediates augmented immune responses against DEN-2 infection and is implicated in the susceptibility of DHF. Further studies are warranted, particularly with regard to the genetic variants of CD209 on the DC polarization of adaptive immunity, and how they may promote or protect the development of DHF. | Conceived and designed the experiments: LW KDY. Performed the experiments: LW RFC CPL. Analyzed the data: LW RFC. Contributed reagents/materials/analysis tools: JWL IKL HCK SKH. Wrote the paper: LW RFC KDY.
Background
The C-type lectin DC-SIGN (CD209) is known to be the major dengue receptor on human dendritic cells, and a single nucleotide polymorphism (SNP) in the promoter region of CD209 (−336 A/G; rs4804803) is susceptible to many infectious diseases. We reason that variations in the DC-SIGN gene might have a broad influence on viral replication and host immune responses.
Methods and Findings
We studied whether the rs4804803 SNP was associated with a susceptibility to dengue fever (DF) and/or dengue hemorrhagic fever (DHF) through genotyping analysis in a Taiwanese cohort. We generated monocyte-derived dendritic cells (MDDCs) from individuals with AA or AG genotype of rs4804803 to study the viral replication and immune responses for functional validation. A total of 574 DNA samples were genotyped, including 176 DF, 135 DHF, 143 other non-dengue febrile illnesses (OFI) and 120 population controls. A strong association between GG/AG genotypes of rs4804803 and risk of DHF was found when compared among DF, OFI and controls ( p = 0.004, 3×10 −5 and 0.001, respectively). The AA genotype was associated with protection against dengue infection compared with OFI and controls ( p = 0.002 and 0.020, respectively). Moreover, MDDCs from individuals with AG genotype with a higher cell surface DC-SIGN expression had a significantly higher TNFα, IL-12p40, and IP-10 production than those with AA genotype in response to dengue infection. However, the viral replication in MDDCs with AG genotype was significantly lower than those with AA genotype. With both genotypes, MDDCs revealed an increase in viral replication following the addition of anti-IP-10 neutralizing antibody.
Conclusions/Significance
The rs4804803 SNP in the CD209 promoter contributed to susceptibility to dengue infection and complication of DHF. This SNP with AG genotype affects the cell surface DC-SIGN expression related to immune augmentation and less viral replication.
Author Summary
Dengue fever (DF) is an arthropod-borne disease that is prevalent in tropical and subtropical regions of the world. DC-SIGN [dendritic cell-specific intercellular adhesion molecule 3 (ICAM-3)-grabbing non-integrin] is a major receptor for dengue infection. DC-SIGN, also called CD209, expresses on dendritic cells (DCs) that bind to ICAM-3, which is expressed on T cells to facilitate the initial interaction between DCs and T cells. Variations in the CD209 promoter (−336 A/G; rs4804803) genotype are involved in the pathogenesis of human infectious diseases. Here we found that patients with dengue hemorrhagic fever (DHF) had a higher frequency of the AG or GG genotype of rs4804803 than DF or controls. Functional studies determined that monocyte-derived DCs (MDDCs) from individuals with AG genotype had significantly higher cell surface DC-SIGN expression, associated with higher TNFα, IL-12p40, and IP-10 production, but lower viral replication than those with AA genotype. An increase in DEN-2 replication in MDDCs was observed following the addition of anti-IP-10 neutralizing antibody. These findings highlight the fact that the rs4804803 SNP in the CD209 promoter is associated with DHF and correlated to DC-SIGN expression and immune augmentation. | The authors have benefited from the statistical advice by Dr. Eng-Yen Huang. For technical assistance, we would like to thank Mrs Ya-Ting Lo and Mrs Yu-Ni Su. | CC BY | no | 2022-01-13 05:08:15 | PLoS Negl Trop Dis. 2011 Jan 4; 5(1):e934 | oa_package/f1/3e/PMC3014977.tar.gz |
||
PMC3014978 | 21245922 | Introduction
Dengue infection constitutes a major threat for urban populations of Latin America and Asia [1] , [2] . Important differences in the clinical and epidemiological profile of dengue between the countries of Latin America and Southeast (SE) Asian have been observed. While in SE Asian countries dengue hemorrhagic fever (DHF) is common and morbidity and mortality has traditionally concentrated in children under 15 years of age, in American countries the disease affects mostly adult populations and manifests primarily as dengue fever (DF) [3] , [4] .
Several hypotheses have been proposed to explain these differences. It has been shown that children from Central America, Venezuela, and Colombia may not develop vascular permeability as readily as children from SE Asia after secondary dengue infection, [5] – [6] and that there may be a high prevalence of dengue resistance genes among black populations of Brazil and the Caribbean [7] , [8] . An additional explanation for the low numbers of DHF in American countries may be underreporting of cases that do occur, due to technical difficulties or a limited capacity to perform diagnosis that meet the criteria of the WHO case definition [3] . None of these explanations are fully satisfactory in explaining the differences between the two regions.
Dengue was reintroduced in Brazil in 1986, after an absence of at least 20 years (except for an epidemic in Roraima in 1981 and sporadic cases). Since then, Brazil has become the country that reports the largest number of cases to the WHO, accounting for over 70% of cases reported in the Americas [9] , [10] . Three serotypes currently circulate throughout the country; DENV 1 was reintroduced to Rio de Janeiro in 1986, DENV 2 in 1990, and DENV 3 in 2002, and from Rio they spread to the rest of the country [11] . While prior to 2007 the majority of DHF cases in Brazil occurred among adults aged 20–40 years of age, in 2007 the annual number of DHF cases more than doubled over previous years and a shift in the age distribution was reported [12] . In 2007, 53% of cases occurred in children under 15 years old. The shift was most noticeable in the Northeast region, where children accounted for 65% of the total number of DHF cases, while other regions such as the Central-West and North did not experience a significant shift and most of the DHF cases continued to occur among adults [12] .
Although the cause of this shift is likely to be multifactorial, we propose that the conditions for it were being set gradually since the re-emergence of DENV in 1986 and that the current epidemiological profile represents the transition from re-emergence to hyperendemicity. In a setting where transmission is constant, people who are exposed for a longer time have a greater cumulative probability of infection. In Brazil, circulation of DENV virus for over 20 years has resulted in the accumulation of immunity in older individuals, driving the average age of primary and secondary infection towards younger age groups.
Using data from a serological study performed in Recife, in Northeast Brazil, we estimate the force of infection and basic reproductive number of dengue in three areas of distinct socio-economic status for the period 1986–2006, in order to better understand transmission intensity over this period. We then use these estimates to simulate the accumulation of monotypic and multitypic immunity in a population previously susceptible to dengue virus, and to predict the expected age distribution of DHF cases in the future. | Materials and Methods
Ethics statement
This study was reviewed and approved by the ethics committee of the CPqAM-Fiocruz/Brazilian Ministry of Health (No. 49/04). Written consent to participate in the study was obtained from each person (or their guardian) after a full explanation of the study was provided. All personal identifiers were removed prior to secondary data analysis at Johns Hopkins University.
Data sources
This study was based on a serological sample of households in Recife, Pernambuco, Brazil, conducted between August and September 2006. The first dengue outbreak in the state of Pernambuco occurred in 1987 (DENV 1). No additional autochthonous cases were reported until 1995, when DENV 2 was introduced causing a new epidemic. Since 1995 cases have been reported every year. DENV 3 was first isolated in Pernambuco in 2002 [13] .
The study population and methods have been described in detail by Braga et al. [14] Briefly, Recife has 1.5 million inhabitants. The climate is humid, with an average temperature of 25°C and rainfall of approximately 2000 mm per year. Three neighborhoods were selected to represent low, medium and high socio-economic areas. A systematic age stratified sample was obtained, using the Census 2000 data that provides the total population size, number of households and age distribution in the three areas [15] . Residents aged between 5 and 64 years were eligible for the survey. Serum samples were screened for IgG antibodies against DENV with an enzyme-linked immunoassay commercial kit (Dengue IgG-ELISA, PanBio, Ltd., Brisbane, Australia). Tests were performed in duplicate according to the manufacturer's instructions. This test does not determine the presence of immunity to specific dengue serotypes, but the presence of immunity to any dengue serotype.
Estimating the force of infection
The force of infection (λ) is a measure used to characterize the intensity of transmission in a given setting and estimates the per capita rate of acquisition of infection by susceptible individuals. Age stratified serological surveys can provide information about the force of infection over a period of time, λ(t), as described elsewhere [16] . Assuming that the risk of infection does not vary with age, the difference in seroprevalence between subjects a and a+1 years of age can be attributed to the transmission intensity between a and a+1 years ago. To estimate λ(t), for the period 1986–2006, we used a model based upon one described by Ferguson et al. [17] We estimated constant and time-varying forces of infection. Detailed information regarding the methods used can be found in Text S1 in Supporting Information S1 .
Estimating the basic reproductive number (R 0 )
R 0 is the number of secondary infections generated by a primary case in a completely susceptible population. R 0 gives insight into the level of control that is required to reduce incidence and eventually block transmission. Detailed information regarding the methods used to estimate R 0 can be found in Text S1 in Supporting Information S1 .
Immunologic shift simulations
To estimate the accumulation of monotypic and multitypic immunity in a population previously susceptible to dengue, we performed a discrete-time simulation by applying the forces of infection estimated from the seroprevalence data onto a simulated immunologically naive population structured by-age like the one of Recife. We used independent data on the years in which the different serotypes were introduced into Brazil/Pernambuco to apply the estimated hazards only in those years when particular serotypes were known to have circulated [11] , [13] . The age profile of the population was obtained from the 2000 census data. We conducted simulations until age distributions of immunity reached equilibrium and used both constant and time-varying hazards.
Since we did not have seroprevalence data to estimate the force of infection beyond 2006, we assumed that λ(t) after 2006 has been constant and equal to the average hazard over the period 1986–2006.
All statistical analyses were performed using R statistical package (version 2.10.1). | Results
The dataset contained data on 1427 subjects aged 5 to 20 years, 593 (41.6%) from area 1, 480 (33.6%) from area 2 and 342 (24.0%) from area 3. Figure 1 shows the age-specific seroprevalences of each of the areas (black dots). Area 1, the neighborhood of low socioeconomic status showed a significantly higher seroprevalence when compared to Area 3, the high socioeconomic stratum neighborhood (0.85 (95%CI 0.82–0.88) vs. 0.70 (95%CI 0.65–0.75), p<0.0001). The middle class neighborhood (Area 2) also showed a significantly higher seroprevalence when compared to Area 3 (0.82 vs. 0.70, p = 0.0002). Based on the seroprevalence and that approximately 87000 cases were notified in Recife during this period, it is clear that less then 10% of the infections were reported [13] .
Force of infection
The estimated average time-constant force of infection for the period 1986–2006 was 0.052 (95% CI 0.041–0.063). On average, each serotype infected 5.2% of susceptible individuals each year.
Time constant s for the three areas were 0.068 (95%CI 0.045, 0.091), 0.056 (95%CI 0.035, 0.077) and 0.035 (95%CI 0.019, 0.051). Though the difference between these forces of infection is not statistically significant, a trend is seen towards higher hazards of infection in settings of lower socioeconomic status.
As can be expected, the fit of the model improved significantly when we allowed for time-varying forces of infection (likelihood ratio test, p = 0.006). According to this model ( Figure 2 ), the average yearly force of infection ranged between 0 and 0.057 between 1986 and 1998, and then peaked at 0.26 in 1999. As has been reported elsewhere the correlation between incidence and estimated force of infection is poor (r = 0.21) [18] .
Given that it has been reported that between 1987 and 1995 there were no autochthonous dengue cases in the state of Pernambuco, we also fit a model constraining the force of infection for these years to be 0 [13] . The fit of this 8-parameter model was not significantly different from the fit of the model that did not constrain these hazards to be zero (LR test, p = 0.99) or from the saturated model (LR test, p = 0.34).
Figure 1 shows the fit of 1) constant (red lines) and 2) time-varying models (green lines) to the age-specific seroprevalence data in the three areas and overall areas.
The correlation between the annual hazards estimated in areas 1 and 3 (r = 0.79) is high, while the correlation between 1 and 2 and between 2 and 3 is poor (r = 0.06 and 0.18, respectively).
Basic reproductive number
Using the time constant and time varying λ s we estimated an overall R 0 of dengue in Recife of 2.7 (95%CI 2.45, 3.11). For the three areas the R 0 estimates were 3.3 (95%CI 2.45, 4.18), 2.8 (95%CI 2.09, 3.64) and 2.1 (95%CI 1.56, 2.66), respectively.
Simulations
Figure 3 shows the age distribution of susceptible, monotypically immune and multitypically immune at different time-points after the introduction of DENV 1, 2 and 3 into a previously susceptible population, assuming a constant risk of infection of 0.052/year/serotype. As the number of years of DENV circulation increases, multitypic immunity accumulates among adults, and susceptibles and monotypically immune become increasingly concentrated in younger age groups.
Assuming that cases of DHF occur primarily among people who experience secondary infection, the age distribution of people who are at risk of secondary infection (i.e. of people who have been exposed to a single dengue serotype) should approximate the age distribution of DHF cases [19] . Hence, our results suggest that as years after re-emergence go by, the mean, median and modal ages of cases will decrease. For = 0.052, the model estimates that while 10 years after re-emergence the median, mean and modal age of cases (monotypically immune) would be 24, 29.0 and 14 years respectively, these numbers would decrease to 13, 15.2 and 11 years 50 years after re-emergence. Similarly, while it is expected that only up to 27% of DHF cases would occur in children under 15 years of age 10 years after the re-emergence, 50 years after re-emergence this proportion would increase to 58%. Figure 4 shows the age distribution of hospitalized dengue cases in Pernambuco in 2007, based on official notification records, and the estimated distribution according to our model (20 years after re-emergence) [20] .
The strength and the speed of the shift in the age distribution of immunity depend on the underlying force of infection ( Figure 5 and Table S1 in Supporting Information S1 ). The estimated median and modal ages of monotypically immune for = 0.03 are 19 and 15 years respectively, 50 years after re-emergence, while these ages drop to 11 and 6 years for = 0.07.
Results were similar if time-varying, instead of constant forces of infection were applied, or if for each serotype was weighted taking into account the serotype predominance reported for the different years in the state of Pernambuco [13] . | Discussion
A dramatic increase in the number of DHF cases and a shift in age group predominance of DHF were observed during the 2007 dengue epidemic in Brazil, the first re-emergence of the DENV-2 serotype predominance since 1990. Our results suggest that this shift can be partly explained by the accumulation of multitypic immunity in the adult population over time after the re-emergence of DENV-1 in 1986, DENV-2 in 1990 and DENV-3 in 2002. As the length of time of co-circulation of multiple serotypes of dengue in Brazil increases, adults have a lower probability of remaining susceptible to infection. As a result, cases become on average younger as completely susceptible individuals and monotypically immune individuals are more likely to be from younger age groups.
If the accumulation of multitypic immunity in adult population is in part responsible for the observed shift in age group predominance of severe dengue cases, we would expect similar shifts to have occurred in central and northern South America, where several DENV serotypes have been known to circulate since the 1970's. In Mexico, Venezuela, Nicaragua and Colombia most of the severe cases occur among children <15 years of age and a similar trend is being observed in Honduras. [21] , [22] , [23] . In contrast, such a trend has not been observed in countries where multiple serotypes only started circulating in the 90's. If the central/west regions of Brazil continue to experience high DENV forces of infection and multiple circulating serotypes we expect a similar shift in age group predominance to occur in the coming years.
Since DHF is more likely to occur in children, a decrease in the mean age of secondary infection might also be expected to lead to an increase in the proportion of dengue infections that lead to severe symptoms or DHF cases [24] , [25] . In the 1990's, after DENV-2 was introduced, 0.06% of reported dengue cases in Brazil resulted in DHF/DSS. This percentage increased to 0.21% in 2007 [9] . This observed increase in DHF may also have been a result of changes in virulence of particular dengue viruses that were circulating or due to the fact that the overall force of infection has increased as has been proposed.
According to our model, the speed of the shift is proportional to the magnitude of the average force of infection. Higher average forces of infection lead to a more rapid shift of the age distributions of immunity and to a younger median and modal age of monotypically immune. Thus, the shift can be expected to be slower in regions that have been exposed to weaker forces of infection or where the re-emergence of multiple serotypes was delayed. This may explain why the shift has only been observed in major cities and certain regions of Brazil. The Northeast region of Brazil, where Recife is located, has the highest proportion of children among DHF cases, and it has also been traditionally the region with the highest incidence rates of dengue fever since 1986 [9] , [12] . The Central-West region, where the shift is not yet apparent has shown high incidence rates of dengue fever only during the last 7 years [9] .
Our estimate of the average λ and R 0 in Recife is lower than those estimated for Thailand for the period 1980–2005 (λ = 0.1, R 0 = 5.2) [26] . Our model predicts that average forces of infection of 0.1 would be associated with a mean age of severe or DHF cases of 8 years, and this is consistent with what has been traditionally observed in SE Asian countries. Both Thailand and Singapore have experienced significant decreases in transmission intensity over the last few years that have been accompanied by an increase in the average age of cases [26] , [27] , [28] . If the force of infection in Recife continues to be as high as it has been over the last 20 years, or higher, it is likely that within the next decade the age distribution of DHF in Recife (and other American regions with high forces of infection) will resemble the age distribution observed in SE Asia, with most cases concentrated in the adolescent population. However, our projections are meant to be qualitative rather than quantitative. The actual seroprevalences observed in the future in Recife may differ from our projections depending on secular trends in the transmission intensity of dengue and population demographics.
There are several limitations to this study. Even though our results present an explanation for why DHF may have shifted towards children over the years since introduction, the mechanism that we propose is gradual and does not explain the sudden change observed in 2007–2008. The recirculation of DENV-2 into certain cities in 2007, after almost 7 years of DENV 3 predominance and the resultant increase in secondary cases may have determined the observation of an age shift in 2007 and not before, even though it had been gradually taking place [9] . As reported by the Ministry of Health, during 1998–2006 the percentage of severe dengue cases in children increased from 9.5% (in 1998) to 22.6% (in 2001). Although our results suggest that the major driver of the shift is the accumulation of immunity in older age groups, fluctuations in serotype specific transmission intensity, serotype predominance, characteristics of the virus or serotype predominance may have also played a role in determining the visibility of the shift.
Our model predicts that after 20 years of exposure to a constant force of infection of 0.05 per year, children 15 years old or younger should only account for 31% of DHF cases while the data shows that in 2007, 70% of cases in Recife occurred among children of this age group. This discrepancy may arise due to the fact that the model does not take into account age-dependence of infection or clinical presentation. If children are more likely to develop severe disease, then the observed distribution of cases is likely to be skewed towards lower age groups.
The fact that the available serological study does not contain serotype specific information limits our ability to estimate serotype specific forces of infection, interactions (enhancement/inhibition) and basic reproductive numbers. Similarly, the cross-sectional nature of this dataset does not allow us to control for potential confounding by age dependent transmission intensity. Longitudinal data and data from seroprevalence studies using serotype specific methods such as the PRNT are essential in order to properly reconstruct the transmission intensity over the last 20 years.
This analysis has important public health implications on planning public health responses to dengue for the next decade. Dengue is the most rapidly spreading vector borne viral disease. If the age shift in fact represents the transition from re-emergence to hyperendemicity, similar shifts in age are likely to be observed in the rest of Brazil, the American continent and other regions where dengue has emerged more recently. | Conceived and designed the experiments: CB WVdS ETM DATC. Performed the experiments: MTC CB. Analyzed the data: IRB WVdS. Contributed reagents/materials/analysis tools: IRB ETM DATC. Wrote the paper: IRB MTC CB WVdS ETM DATC. Data/sample collection and processing: MTC CB WVdS.
Background
Dengue virus (DENV) was reintroduced into Brazil in 1986 and by 1995 it had spread throughout the country. In 2007 the number of dengue hemorrhagic fever (DHF) cases more than doubled and a shift in the age distribution was reported. While previously the majority of DHF cases occurred among adults, in 2007 53% of cases occurred in children under 15 years old. The reasons for this shift have not been determined.
Methods and Findings
Age stratified cross-sectional seroepidemiologic survey conducted in Recife, Brazil in 2006. Serostatus was determined by ELISA based detection of Dengue IgG. We estimated time-constant and time-varying forces of infection of DENV between 1986 and 2006. We used discrete-time simulation to estimate the accumulation of monotypic and multitypic immunity over time in a population previously completely susceptible to DENV. We projected the age distribution of population immunity to dengue assuming similar hazards of infection in future years. The overall prevalence of DENV IgG was 0.80 (n = 1427). The time-constant force of infection for the period was estimated to be 0.052 (95% CI 0.041, 0.063), corresponding to 5.2% of susceptible individuals becoming infected each year by each serotype. Simulations show that as time since re-emergence of dengue goes by, multitypic immunity accumulates in adults while an increasing proportion of susceptible individuals and those with monotypic immunity are among young age groups. The median age of those monotypically immune can be expected to shift from 24 years, 10 years after introduction, to 13 years, 50 years after introduction. Of those monotypically immune, the proportion under 15 years old shifts from 27% to 58%. These results are consistent with the dengue notification records from the same region since 1995.
Interpretation
Assuming that persons who have been monotypically exposed are at highest risk for severe dengue, the shift towards younger patient ages observed in Brazil can be partially explained by the accumulation of multitypic immunity against DENV-1, 2, and 3 in older age groups, 22 years after the re-introduction of these viruses. Serotype specific seroepidemiologic studies are necessary to accurately estimate the serotype specific forces of infection.
Author Summary
The spread of dengue virus is a major public health problem. Though the burden of dengue has historically been concentrated in Southeast Asian countries, Brazil has become the country that reports the largest number of cases in the world. While prior to 2007 the disease affected mostly adults, during the 2007 epidemic the number of dengue hemorrhagic fever cases more than doubled, and over 53% of cases were in children under 15 years of age. In this paper, we propose that the conditions for the shift were being set gradually since the re-introduction of dengue in 1986 and that they represent the transition from re-emergence to hyperendemicity. Using data from an age stratified seroprevalence study conducted in Recife, we estimated the force of infection (a measure of transmission intensity) between 1986–2006 and used these estimates to simulate the accumulation of immunity since the re-emergence. As the length of time that dengue has circulated increases, adults have a lower probability of remaining susceptible to primary or secondary infection and thus, cases become on average younger. If in fact the shift represents the transition from re-emergence to hyperendemicity, similar shifts are likely to be observed in the rest of Brazil, the American continent and other regions where transmission emerges. | Supporting Information | CC BY | no | 2022-01-13 05:08:15 | PLoS Negl Trop Dis. 2011 Jan 4; 5(1):e935 | oa_package/e0/84/PMC3014978.tar.gz |
||
PMC3014979 | 21092319 | Background
In humans there are more than 40 relatively rare progressive neurodegenerative, neurological and neuromuscular diseases linked with repeat instability mutations [ 1 - 4 ]. A distinct subset of this group, which includes for example Huntington disease and spinocerebellar ataxias, is caused by polymorphic variation in trinucleotide repeats, typically CAG repeats, often present in the coding sequences of specific genes [ 1 - 3 , 5 - 8 ]. In this group of 11 diseases the repeats usually encode long poly-glutamine (poly-Q) tracts, although other poly-amino acid tracts also occur but at lower frequencies. These human diseases are caused by extreme expansions of the repeats, which adversely impact protein structure, often causing intracellular protein aggregation and altered protein function [ 1 , 3 , 7 , 9 , 10 ]. However, trinucleotide repeat tracts which are impure, but maintain the ability to encode poly-Q tracts, have less severe impacts on function in model systems. This suggests that repeat expansions influence both protein and mRNA structure and function [ 1 , 4 , 8 , 11 ].
The length of the expanded repeat is typically inversely related to the age of disease onset and directly related to the severity of the disease [ 6 , 11 ]. Moreover, there is often a tendency for the expanded repeat to further increase in size in subsequent generations thereby resulting in greater disease severity and earlier onset, a phenomenon called genetic anticipation [ 1 ]. Trinucleotide coding sequence repeats in many genes are often polymorphic in the normal population but these repeats are typically shorter in length compared with disease-causing alleles. In some instances the distinction between normal and disease states can be minor in terms of repeat length [ 1 , 11 ].
In a restricted examination of 7,039 human, rat and mouse orthologs 75, 53 and 58 genes, respectively, containing poly-Q encoding CAG repeats (CAG poly-Q genes) were identified [ 12 ]. Another analysis of the human genome reported a total of 66 CAG poly-Q genes [ 11 ]. These regions were defined as containing five or more CAG repeats encoding poly-Q. Although poly-Q containing proteins are generally more highly conserved in mammals, paradoxically their repeat regions are typically poorly conserved both in sequence and length, and they have been implicated in the rapid functional diversification of these proteins and the continuous morphological evolution of mammals [ 13 - 15 ]. It has also been suggested that normal polymorphic variation in these genes contributes to morphological variation within a species, as many of the affected genes are involved in regulating aspects of development and often encode transcription factors and transcription co-regulators [ 11 , 14 ]. Indeed, polymorphic trinucleotide repeats in two genes, RUNX2 (runt-related transcription factor 2) and ALX4 (aristaless-like homeobox 4), have been linked with morphological variation in canine skulls and limbs [ 15 , 16 ]. It is possible therefore that those genes containing polymorphic coding sequence trinucleotide repeats may contribute to variation in a broad range of phenotypes within a species.
What is not clear is whether the propensity for normal polymorphic variation in the CAG repeats of some of these genes also occurs in non-primates. There is significant positive correlation between the length of a pure CAG repeat and its propensity for polymorphic variation in human genes [ 11 ]. Interruptions of the CAG repeat, even with the alternate codon for glutamine, CAA, decrease the likelihood of polymorphic variation, as do reading frame shifts maintaining the CAG repeat structure but encoding different amino acids [ 11 ]. Moreover, the potential for a gene to cause known human diseases was also associated with greater CAG repeat polymorphic diversity [ 17 ]. Thus, it is probable that long and pure CAG tracts encoding poly-Q are predictors of polymorphic variation that is likely to impact on phenotype.
Identification of poly-Q encoding genes in additional mammalian species may identify species-specific differences in both the repertoire of these genes as well as the diversity of polymorphic variants in a subgroup of these genes. The former genes may contribute to an understanding of morphological and functional differences between mammalian species while the latter may contribute to phenotypic variation within a species. The current investigation has identified all bovine poly-Q encoding genes and within this group a number of genes whose orthologs in human and mouse do not contain these repeats. The bovine poly-Q encoding genes were prioritised using multiple criteria for their likelihood of being polymorphic and then the highest ranking group was experimentally tested for polymorphic variation within a cattle diversity panel. Extensive polymorphic variation was identified. The poly-Q encoding genes were strongly enriched for molecular functions relating to transcriptional regulation and their encoded proteins are involved in large transcriptional regulatory complexes in the cell nucleus. | Methods
Animal resources
Large cattle diversity panel
Bos taurus taurus animals (150) were represented by seven dairy breeds (Australian Red Breed (n = 16), Ayshire (n = 3), Brown Swiss (n = 3), Guernsey (n = 7), Holstein Friesian (n = 24), Illawarra Shorthorn (n = 12) and Jersey (n = 15)) and five beef breeds (Angus (n = 8), Belgium Blue (n = 3), Charolais (n = 4), Poll Hereford (n = 11) and Shorthorn (n = 9)). The Bos taurus indicus animals (35) included four breeds (Africander (n = 7), Boran (n = 8), Brahman (n = 12) and Tuli (n = 8)) (Additional file 2 ). The Australian Dairy Herd Improvement Scheme (ADHIS) database was used to ensure that the most unrelated dairy animals available to us were included in the panel; while other animals were sourced from diverse properties throughout Queensland, Australia.
Small cattle diversity panel
Eight animals were selected from the large diversity panel for initial screening for polymorphic variation. These animals included one of each from the Shorthorn, Poll Hereford, Boran, Tuli, Africander, Brahman, Holstein Friesian and Charolais breeds.
Identification of bovine genes encoding poly-Q tracts
Tracts of five or more glutamines (Q-tracts) encoded by bovine genes were identified using the following procedure. Cattle protein sequences in NCBI's RefSeq database, which were derived from annotation of the Btau 3.1 genome sequence assembly [ 19 , 20 ], were down-loaded ( ftp://ftp.ncbi.nih.gov/genomes/Bos_taurus/protein/protein.fa.gz ). Within this set 178 proteins containing at least one Q-tract were identified using a simple script. This analysis was performed without regard to whether glutamine was encoded by CAG or CAA codons. The genes were also manually re-examined using the Btau 4.0 assembly using both RefSeq and Ensembl gene models to confirm Q-tracts and the nature of their coding sequences. Reciprocal mapping of human and bovine orthologs was also performed (Additional file 1 ). This highlighted human orthologs that map to more than one GLEAN5 model, which may represent local assembly issues in the bovine genome assembly and/or incomplete gene models. Only unique bovine GLEAN5 models were used.
To identify evidence of poly-Q tract length variation in sequence databases due to possible allelic contributions, the 178 protein sequences were used as queries in tBlastn searches run with NCBI's Bos taurus genomic Blast service [ 18 ]. The Expect cut-off value was set to 0.00001 and default values were used for other parameters. Two databases were queried: Bos taurus HTGS (High Throughput Genome Sequence) and Bos taurus expressed sequence tags (ESTs). Results with a maximum of 50 alignments were formatted as 'query anchored' HTML and scrutinised for alignments in which there was variation in the length of a poly-Q tract. After excluding non-allelic variation (splicing in the case of the EST database searches, and gene paralogy in the case of the HTGS database searches), 19 candidate proteins were identified (NFYA, FXC1, FAM48A, CACNA1A, FOXP1, FAM155A, ABCF1, MLL3, NUFIP2, CREBBP, FOXP2, GLG1, MED12, LRRC4, FBXO11, ST6GALNAC5, THAP11, NCOR1, MED15). Throughout the text gene symbols were italicised except when used in the context of their encoded proteins.
Prioritization of bovine genes for experimental analysis of polymorphic variation in regions containing poly-Q tracts
A number of characteristics have been associated with propensity for polymorphic variation in genes containing trinucleotide repeats encoding poly-Q tracts [ 11 ]. The bovine genes encoding poly-Q tracts were hierarchically prioritised according to the following criteria: (i) length of pure CAG repeat encoding a poly-Q tract (≥(CAG) 5 ); (ii) lack of conservation of the poly-Q tract lengths in mammalian orthologs (manual inspection using HomoloGene Release 64; [ 18 ]); (iii) evidence for polymorphic variation in other mammalian species [ 1 , 8 , 11 , 17 ]; (iv) demonstration that polymorphic variations in the human or mouse orthologs caused diseases [ 1 , 8 , 11 , 17 ], and; (v) apparent disparities between bovine cDNA sequences and the bovine reference genome sequence in genes encoding poly-Q repeats. Genes in each category were assigned a value of 1 or 0 and ranked according to their cumulative category score. Genes with identical scores were then ranked according to CAG repeat length. The top 32 ranked genes selected according to criteria (i) to (iv) also included 16 of the 19 genes identified by criterion (v). The 32 genes were screened for the presence of polymorphic variation in their CAG repeat regions using the small cattle diversity panel consisting of eight animals. Some genes were assessed for multiple repeat regions. Genes that showed evidence of polymorphic variation in the small diversity panel were then assessed in the larger cattle diversity panel. Also included in both analyses were the genes AR , ODAM and NFYA , representing genes that were lowly ranked in the prioritisation process and therefore predicted to be monomorphic.
DNA purification
Genomic DNA was extracted from white blood cells using 200 μl of whole blood and the QiAmp DNA Mini kit (Qiagen). DNA quality was assessed as an A 260 /A 280 ratio greater than 1.8. The DNA was quantified by spectrophotometry and diluted to a working concentration of 10-50 ng/μl with 10 mM Tris-HCl, pH 7.5, 1 mM EDTA.
Genotyping
PCR screening
Preliminary PCR analysis of 32 prioritised genes and the 3 low ranking genes containing poly-Q tracts was undertaken using DNA from the small diversity panel. Each gene was subjected to PCR to generate an amplicon traversing the region(s) containing the repeats encoding the poly-Q tract. Some genes contained more than one region and these were all analysed. Primers were designed from bovine genome sequence information (Btau4.0; [ 19 ] (Additional file 6 ). PCR reactions (10 ml) contained 10-50 ng genomic DNA, 0.5 μM of each primer, 200 μM dNTP in 1 × Reaction Buffer (1 × Q solution) and Hot-Star Taq (Qiagen). The PCR conditions were 94 °C for 15 min, followed by a touch-down protocol: (11 initial cycles of 94 °C for 30 s, 65-57°C for 30 s with a reduction of 0.7°C/cycle, and 72°C for 45 sec; 32 cycles of 94°C for 30 s, 57°C for 30 s and 72°C for 45 s; and a final extension step at 72°C for 15 min). The amplified product was visualised on 3% agarose gels stained with ethidium bromide. Controls containing no sample were routinely performed. The amplified DNA products were sequenced to confirm gene identity and to provide size references for the repeat regions identified by multiplex genotyping.
Multiplex genotyping assays
Nineteen of the initial 32 prioritised genes showed evidence of polymorphic variation in the initial PCR screen using the small diversity panel. Multiplex genotyping assays were then developed for these genes so they could be analysed on an ABI 3130 × l Genetic Analyser (Applied Biosystems, Australia). Forward primers were labelled with the fluorophores specified in Additional file 6 and three sets of markers were developed for the efficient genotyping of all 19 genes. The first set included RUNX2 , PCR amplified as a singleplex and ( ST6 , FAM48A , MLLQ3 and FXC1 ) amplified as a multiplex; the second set included two PCR multiplexes ( CREBBP , NCOR1 , ABCF1 and MED12-Q1 ) and ( THAP11 , HTT and MED15-Q1 )); while the third set comprised ( C10ORF26 , EXDL2 , TBP ) and ( FAM55A , MEF2A , CACANA1 and LRCH4 ) amplified as two separate multiplexes. The suffix '-Q1' attached to a gene symbol refers to the first of more than one poly-Q encoding region within a gene. Each PCR product (0.5 μl of a 1:100 dilution) from each marker set was co-loaded onto the ABI 3130 × l Genetic Analyser for capillary electrophoresis according to the manufacturer's instructions with LIZ 500 (Applied Biosystems) as the size standard. Sizing of PCR products was accomplished using GeneMapper software (v4.0, Applied Biosystems). PCR reactions (10 ml) contained 10-50 ng genomic DNA, 0.1 μM of FAM-labelled primer pair, 0.25 μM of VIC-, NED- and PET-labelled primer pairs, 1× Q solution and 1 × reaction buffer (Qiagen Multiplex PCR Kit). Multiplex PCR conditions were 94 °C for 15 min; followed by a touch-down protocol: 11 cycles of 94 °C for 30 s, 65-57°C for 1 min 30 s with a reduction of 0.7°C/cycle, and 72°C for 45 s; 32 cycles of 94°C for 30 s, 57°C for 1 min 30 s, and 72°C for 45 s, and; a final extension step at 72°C for 15 min. PCR conditions were similar for singleplex reactions except that Hot-Star Taq (Qiagen) and 200 μM dNTP were used in the reaction and the annealing time in the touch-down protocol was reduced to 30 s.
cDNA synthesis and reverse transcription PCR (RT-PCR)
Total RNA was extracted from a range of bovine tissues using Trizol reagent (Invitrogen) followed by DNase I (Ambion) treatment [ 57 ]. RNA was quantified, its quality verified and cDNA synthesized using an anchored oligo-T 18 primer combined with random hexamers [ 57 ]. To validate the absence of genomic DNA, control experiments were performed without reverse transcriptase. cDNA derived from each bovine tissue sample was subjected to PCR for specific amplification of NFYA . The FAM-labelled sense and antisense primers were designed to traverse the region encoding the poly-Q sequence (sense primer, 5'-CAAACAGCAACAGTTCAGCAG-3'; antisense primer, 5'-CAGGGTCTGGACTTGCTGG-3'). The conditions used for PCR were as described above. The amplified product was sequenced to confirm its identity. All samples contained the same quantity of first-strand cDNA (0.9 ng). Controls included no template or sample that was not subjected to cDNA synthesis using reverse transciptase. In addition, PCR was undertaken using genomic DNA but with NFYA primers spanning the exon 2 - intron 2 and intron 2 - exon 3 boundaries (sense primer, 5'-CAAACAGCAACAGTTCAGCAG-3'; antisense primer, 5'-AGAAGATCCGATGCCTCTCA-3': and sense primer, 5'-GAAGACCCTTCCTCTCCCAC-3'; antisense primer, 5'-CCTCAGTCTGCAACTGGACA-3', respectively).
Functional annotations
To identify higher level functional themes associated with the bovine poly-Q encoding genes, they were analysed by using the Database for Annotation, Visualization and Integrated Discovery (DAVID) with a background containing all genes represented in the bovine genome sequence including Chromosome Unknown [ 26 , 27 ]. DAVID provides statistical methods for identification of enriched biological terms within data sets. Statistically over-represented Gene Ontology (GO) terms, keywords and pathways were identified by selecting those with a Benjamini-adjusted p-value < 0.05. Similar results were obtained by using the full complement of human genes as background. Functional Annotation Clustering was then performed using the DAVID system. This higher level analysis displayed similar functional annotations together based on overlaps of genes associated with each function term and therefore gives a clearer overview of gene function information associated with large datasets. In this instance, enrichment scores (- log (geometric mean of the P-values for terms in the cluster)) greater than 1.3 were considered significant. The conservative High Stringency option was used in conjunction with default settings for other parameters. Analysis of the subgroup of bovine genes containing poly-Q tracts which were not present in their human and murine orthologs used the same analysis procedure except that the background was all bovine genes containing poly-Q tracts.
Interaction networks
The InnateDB database was used to identify networks of physically interacting poly-Q proteins using their human RefSeq orthologs as input [ 28 , 29 ]. The database contains comprehensive records of protein-protein and protein-gene interactions that have been experimentally defined, as well as links to relevant PubMed records [ 18 ]. The interactions were restricted to only those within the group of submitted proteins and were visualised using Cerebral, which is a Java plugin for the Cytoscape biomolecular interaction viewer [ 58 ]. Proteins that regulated the expression of genes in the list were also included and their links were colored red. | Results
Identification of poly-Q encoding bovine genes
Bovine genes encoding tracts of five or more glutamines (poly-Q tracts) were initially identified using the NCBI RefSeq collection [ 18 ] and GLEAN5 gene models in conjunction with the BTAU 3.1 bovine genome sequence assembly [ 19 , 20 ]. They were subsequently manually confirmed using the current Btau_4.0 assembly. A tract of five or more Q in a protein sequence has been identified as statistically different from the average protein with average composition [ 21 ]. A total of 178 poly-Q encoding genes were identified of which 36 contained multiple tracts and 123 contained pure CAG encoding poly-Q repeats (Additional file 1 ).
The presence of a polymorphic repeat tract occasionally confused automated gene model annotation pipelines which utilised cDNA and genomic sequence information representing different alleles in the repeat region. For example, most gene models for bovine FXC1 (fractured callus 1) are incorrect as they attribute an additional intron to the repeat region, which clearly showed allelic differences between genomic DNA and EST sequences [ 22 ]. This annotation error is prevalent in other unrelated gene models where polymorphic repeats are prevalent (e.g. [ 23 ]). The mean poly-Q length encoded by the 178 genes was 7.26 ± 3.55 (1 sd). The longest tract, containing 33 Q residues, was present in FOXP2 (forkhead box P2) while MLL2 (myeloid/lymphoid or mixed lineage leukemia 2) contained the greatest number of repeat tracts, 19. A similar analysis revealed that the human genome contained 201 poly-Q encoding genes of which 109 are orthologs of bovine poly-Q encoding genes (result not shown). Figure 1 shows that most bovine poly-Q proteins have poly-Q tract sizes approximately similar to their human orthologs. Twenty six bovine poly-Q encoding genes were identified with Q-tracts that were not evident in their human and murine orthologs when examined using HomoloGene [ 18 ] (bolded in Additional file 1 ). The mean size of the bovine poly-Q proteins (1,034 ± 913 amino acids (1 s.d.; n = 155)) deduced from the longest Ensembl gene model information [ 22 ] was significantly larger than for all bovine proteins (mean = 365 ± 276 amino acids; 1 s.d.; n = 14,985; p < 0.001). Poly-Q tracts were not enriched in any one third of the protein sequence lengths (p > 0.05). Therefore, they were not biased toward the ends of the polypeptides where they may be expected to be better tolerated in terms of their impacts on polypeptide functions.
Polymorphic variation of bovine poly-Q encoding genes
Several criteria have been associated with the propensity for polymorphic variation in poly-Q encoding genes [ 11 ]. Bovine poly-Q encoding genes were prioritised in a hierarchical manner according to the following criteria: (i) length and purity of CAG repeat; (ii) lack of conservation of CAG repeat length in mammalian orthologs; (iii) evidence for polymorphic variation in other mammalian species; (iii) evidence that polymorphic variations in the human or mouse orthologs caused diseases, and; (iv) disparities between bovine cDNA sequences and the BTAU4.0 reference genome sequence in regions of bovine genes encoding poly-Q tracts. The top ranked 32 genes were initially screened by PCR for polymorphic variation in amplicon size using a small diversity panel consisting of eight cattle. Five of these genes contained two poly-Q tracts and in each case both were independently assayed. In addition, three genes that were not highly ranked were also tested (i.e. ODAM (odontogenic ameloblast-associated protein), AR (androgen receptor) and NFYA (nuclear transcription factor Y, alpha)). The latter analysis was used as a control for the prioritization process. ODAM is situated within a cluster of casein genes and hence it is also of biological interest due to potential strong selection at this locus. Polymorphic poly-Q variants of human AR have been linked with a number of reproductive traits [ 24 ], while NFYA was particularly interesting as its poly-Q encoding tract traverses a splice site (see below).
Of the total of 35 genes tested, 19 showed suggestive evidence for polymorphic variation in the initial 8 animal screen. Figure 2 shows representative profiles for a number of these genes as well as two ( RUNX2 and FBX011 ) that showed no amplicon size variation in the small cattle diversity panel. FXC1 was the most polymorphic gene in the initial analysis.
A larger cattle diversity panel consisting of 150 animals (Additional file 2 ) was then used for automated genotyping of the polymorphic regions in the 19 genes identified in the initial screen. The analysis also included poly-Q encoding regions within ODAM , NFYA and AR , all of which were found to be monomorphic in this larger cattle diversity panel, as they were in the initial small animal screen. The results of this analysis are presented in Table 1 . A total of 16 of the 19 genes identified in the preliminary screen were confirmed as polymorphic in the large diversity panel, with nearly all (15/16), having more than two alleles. Representatives of each gene were sequenced to confirm the allelic variation. As anticipated from the preliminary screen, FXC1 was the most polymorphic gene with 18 alleles. Also of note were the nine HTT (Huntington) alleles. The latter gene is the prototypical CAG poly-Q gene in humans. Extreme repeat expansions in this gene are associated with the autosomal dominant neurological disorder, Huntington disease [ 1 , 6 ]. CAG repeat expansions or mutations in 4 of 16 human orthologs of the polymorphic bovine poly-Q genes are associated with diseases [ 1 , 3 , 8 , 11 ]. The mean size of the proteins encoded by the 16 bovine polymorphic genes was not different from all bovine poly-Q proteins (p > 0.05). Figure 3 shows that for these 16 genes there is a linear relationship between the largest number of repeats in an allele of a gene and the number of alleles. An independent analysis of these 16 genes in 82 Holstein dairy cattle, representing 28 trios, each consisting of both parents and one offspring, revealed stable inheritance of all 16 polymorphic genes (result not shown).
The allele frequencies in the large cattle diversity panel showed significant differences between Bos taurus taurus and Bos taurus indicus animals for 15 of 16 of the polymorphic poly-Q genes (p < 0.05; Fisher exact test, [ 25 ]) (Table 1 and Additional file 3 ). For some genes there was a clearly predominant allele (e.g. MED12 (mediator complex 12) and THAP11 (THAP domain containing 11)) while others showed broad distribution profiles (e.g. FXC1 and HTT ) (Additional file 3 ). Allele frequencies of four genes, CACNA1A (calcium channel, voltage dependent P/Q type, alpha 1A subunit), FXC1, HTT and TBP (TATA box binding protein), were also different (p < 0.05; Fisher exact test) in Bos taurus taurus cattle breeds specialised for beef or dairy uses. Larger cattle populations will need to be assessed in subsequent studies to confirm these results and to also re-examine some of the highly ranked genes that were apparently monomorphic as the process used for gene selection was likely to exclude low frequency allelic variation.
Enrichment of poly-Q encoding genes for functional terms
To test the hypothesis that particular functional classes of genes are enriched in the full repertoire of bovine poly-Q encoding genes we undertook gene function enrichment analyses. The Database for Annotation, Visualization and Integrated Discovery (DAVID) functional analysis tool was used to identify over-represented (p < 0.05; Benjamini correction for multiple testing) gene ontology (GO) terms, pathways and keywords [ 26 , 27 ]. DAVID was then used to group the significantly enriched terms into Functional Annotation Clusters as many functional terms were associated with overlapping gene contents. This process provided a higher level perspective of the enriched functions associated with these genes. There were 21 functional annotation clusters with enrichment scores greater than 1.3, which is considered a significance threshold. Additional file 4 contains the complete analysis. Cluster 1 was the most significant (enrichment score = 21.36) and it contained the most terms. This cluster was over-represented with many terms relating to control of gene transcription, as well as terms associated with regulation of metabolism and biological processes. Strikingly, Clusters 1, 5-7, 9-17, 19 and 20 all contained terms linked with transcriptional regulation and consistent with this there was corresponding enrichment for nuclear organelle structure (Clusters 2 and 4). As might be expected, there was enrichment for themes relating to trinucleotide repeats (Cluster 21). The analysis also highlighted metabolic processes (Clusters 1 and 3). Only one KEGG pathway was over-represented i.e. The Notch Signalling Pathway (P = 2.3E-3, Benjamini correction for multiple testing). Although not reaching significance after correction for multiple testing, Neurodegenerative Diseases and Huntington's Disease KEGG pathways were close to the significance threshold. In humans both of these pathways are closely associated with coding sequence CAG repeat expansions that generate diseases.
Human orthologs of the bovine poly-Q encoding genes were enriched for gene expression (UP_Tissue database) in Epithelium (p = 7.6E-13, after Benjamini correction for multiple testing), Brain (p = 7.1E-4, corrected for multiple testing) and Testis (p = 3.8E-2, corrected for multiple testing) (DAVID database). Manual examination of the expression of these genes using human and murine gene expression databases (GEO database; GDS182 (mouse) and GDS1402 (human); [ 18 ]) generally revealed ubiquitous expression patterns but with bias towards greater expression in epithelia. The 26 poly-Q genes unique to bovine (i.e. containing poly-Q tracts that were not present in mouse or human orthologs) were not significantly different from all of the bovine poly-Q encoding genes with respect to functional terms as measured by DAVID or by examination of gene expression patterns. Likewise, the 16 polymorphic genes were also not enriched for functional terms relative to all poly-Q encoding genes.
The gene function analysis indicated strong enrichment for transcription factors and transcriptional co-regulators in the bovine poly-Q genes. Many transcription factors and their co-regulators physically interact to form functional units controlling gene expression. We therefore tested the extent of interactions for the proteins encoded by human orthologs of the bovine poly-Q genes by using the InnateDB database [ 28 , 29 ]. Figure 4(a) shows the network of human orthologs to the bovine poly-Q proteins that have been experimentally documented to be involved in physical interactions with other proteins from the same group. The diagram emphasises the nuclear location of many poly-Q proteins and that they have potentially extensive physical interactions within the nucleus. Some poly-Q proteins, such as CREBBP (cAMP response element binding (CREB) protein), EP300 (E1A binding protein p300) and SP1 (SP1 transcription factor) had more extensive heterologous interactions. In addition, the transcription of several genes in this group was also controlled, at least in part, by proteins present in the same group (red connections). Figure 4(b) shows a smaller interaction network involving only those proteins whose bovine orthologs were shown to be polymorphic. CREBBP and NCOR1 (nuclear receptor co-repressor 1) are both directly involved in physical associations with HTT [ 30 , 31 ].
The impact of CAG repeat regions on mRNA splicing
During analysis of the bovine genome sequence we identified discordant ESTs associated with poly-Q encoding tracts of bovine genes, which subsequently led to the identification of a number of polymorphic variants. The same analysis also highlighted discordances potentially arising from alternative mRNA splicing events. NFYA is a ubiquitously expressed highly conserved transcription factor subunit that, in conjunction with NFYB and NFYC, binds to CCAAT motifs in the promoters of many genes that are expressed in a tissue specific manner [ 32 ]. Reverse transcriptase-PCR (RT-PCR) results demonstrated that NFYA was expressed in a broad spectrum of bovine tissues (result not shown). This result was confirmed by manual inspection of murine and human gene expression databases (GEO accessions GDS182 (mouse) and GDS1402 (human) [ 18 ]). Close inspection of the cDNA amplicons, which traversed the poly-Q encoding region, suggested that they were not all equivalent in size. Figure 5(a) shows a diagrammatic representation of the organisation of exons 2 and 3 of the NFYA gene and includes the sequences flanking the exon 2 - intron 2 and intron 2 - exon 3 boundaries. The poly-Q encoding region traversed the mRNA splice site i.e. three CAG codons were present in exon 2 and two in exon 3. Neither of these flanking regions was polymorphic in four individuals when genomic DNA was assayed by PCR (Figure 5(b) ). Nor was this region polymorphic in the large cattle diversity panel. However, when cDNA from mammary tissue was used as template, two amplicons differing by three nucleotides were evident. When sequenced, these amplicons contained either four or five CAG codons in the poly-Q encoding region. Moreover, there were clear tissue-specific NFYA expression differences for these two variants (Figure 5(c) ). Only the longer variant was expressed in fat, spleen and liver while both variants were expressed in mammary and brain tissues. Thus, different splice donor or acceptor sites in the CAG repeat region of the gene were used to generate tissue specific mRNA sequence variation. Since the N-terminal region of NFYA, and in particular its poly-Q segment, possesses significant transcriptional activation activity, then the splice site variation could be generating tissue specific differences in the activity of this transcription factor [ 33 , 34 ].
Examination of the gene models for ABCF1 (ATP-binding cassette, sub-family F, member 1) [ 22 ], one of the polymorphic poly-Q encoding genes, as well as MLL2 and FOXP2 , revealed that their repeat regions also traversed splice sites. Moreover, several of the poly-Q encoding genes had repeats that were juxtaposed at or very close (< 5 bp) to splice sites ( ATXN3 (ataxin 3), FOXP1 (forkhead box P1), MED12 (mediator complex subunit 12), TNRC15 (trinucleotide repeat containing 15), TNRC4 (trinucleotide repeat containing 4), CXXC1 (CXXC finger 1), CDC2L6 (cell division cycle 2 -like 6) , ANKRD56 (ankyrin repeat domain 56), AMOT (angiomoton), while some repeats were contained within small alternatively spliced exons (e.g. CACNA1A, MED15 (mediator complex subunit 15), NCOR1; AMOT ). Thus, there is potential for poly-Q tracts and their allelic variants to influence alternative splicing through a number of possible mechanisms.
Additional file 5 shows UCSC bovine genome browser [ 22 ] representations of the 3' and 5' ends of two adjacent exons in MED12 . The exons are separated by an ~ 820 bp intron. Both exons encode large poly-Q tracts. Also included is protein sequence information for cow, dog, human and mouse as well as informative bovine ESTs. It is evident from the ESTs that there is considerable transcriptional diversity being generated from these two exons. This diversity could comprise the effects of polymorphic variation as well as alternative splicing because the ESTs are derived from a number of individuals. However, within one library from a single individual there are a number of very different ESTs (DN274709, DN540107, DN518332, DN521072, DN538689). The splice donor and acceptor sites in all of these cases are situated within CAG repeat tracts suggesting a direct involvement of these tracts in generating transcriptional diversity through their influence on alternative splicing. Alternatively, there is a possibility that the transcriptional diversity reflects heterozygosity for polymorphic alleles within this individual. However, this explanation cannot account for the full scope of the transcriptional diversity. | Discussion
The current study identified 178 bovine poly-Q encoding genes and demonstrated that a substantial number of the ranked genes were polymorphic in their repeat containing regions. Assessment of larger populations of cattle will be required to define the full repertoire of polymorphic poly-Q encoding genes as lower frequency alleles were probably excluded in the current analysis. The lack of conservation of the repeat regions associated with polymorphic variation in poly-Q encoding genes indicate that these are rapidly evolving gene regions [ 13 ]. Indeed, it has been estimated that the insertion-deletion mutation rate of microsatellites, such as CAG repeats in mammals, is approximately 100,000 fold greater than single nucleotide substitutions [ 14 ]. As alterations in the number of trinucleotide repeats do not cause frameshifts in coding sequences, variation in these regions both between and within species is probably much better tolerated than other types of indels. Longer CAG repeats with little interruption by the alternative glutamine encoding codon CAA, are more likely to be associated with higher allele diversity (Figure 3 ) [ 11 , 35 - 39 ]. Polymerase slippage, unequal crossing-over during replication, and repair associated mechanisms, are the most likely explanations for this increased genetic diversity [ 4 , 40 , 41 ]. The overall balance between the mutational mechanisms that promote repeat expansions and the accumulation of point mutations that decrease repeat purity and therefore decrease the probability of repeat expansion is likely to dictate the nature of the poly-Q tract within a gene and its propensity for allelic variation.
Although there are a number of human diseases that are caused by instability in the repeat regions of poly-Q encoding genes, particularly expansion in subsequent generations (anticipation) and during aging, there was no evidence for meiotic instability in the 16 polymorphic poly-Q encoding bovine genes that were tested. A similar result was also demonstrated for porcine HTT [ 42 ]. One likely explanation is that the identified bovine polymorphic gene variants represented normal alleles of relatively short lengths and not the extreme expansions characteristic of some poly-Q encoding genes that cause human diseases [ 1 , 6 , 11 ].
It has been argued that proteins containing repeats simply reflect the propensity for variation and that these regions are not subject to selection and have no functional impact [ 43 , 44 ]. These conclusions are at odds with substantial evidence demonstrating that amino acid repeat size in a wide spectrum of human genes impacts on molecular function and that these regions are indeed subject to strong selection [ 4 , 9 , 10 , 13 , 24 , 30 , 45 - 51 ]. In particular, Q-tracts have been shown to be autonomous length-dependent activators of DNA binding activities in transcription factors [ 46 , 47 , 49 ]. It is noted that there was strong enrichment for transcription factors in the bovine poly-Q encoding genes and thus it is likely that intra-species and inter-species poly-Q tract expansions (or contractions) directly impact the transcriptional activities of these proteins. This is an appealing concept as it links propensity for increased genetic diversity with subtle functional differences in a class of proteins that can amplify these genetic influences because they impact the transcriptional regulation of a much broader group of genes. Moreover, based on evidence linking normal polymorphic variations in two poly-Q encoding genes, RUNX2 and ALX4 , with morphological variation in dogs, it has been proposed that poly-Q encoded genes contribute raw genetic material for rapid evolutionary change in a species [ 10 , 14 , 15 , 52 ].
Those genes containing conserved repeats, which often tend toward shorter and less pure repeat sequences, may reflect strong functional constraints on the encoded proteins in these regions and therefore selective pressure against variation despite increased mutational opportunity. Consistent with this concept, these genes are rarely polymorphic in their repeat regions within a species population [ 11 ]. This information highlights local regions which may have important functions. The absence of polymorphic variation in the poly-Q tract of the bovine androgen receptor (AR), a co-activator for androgen dependent gene transcription, is interesting in view of the links between poly-Q variation in the human protein and androgen-dependent diseases, many of which influence reproductive success [ 53 , 54 ]. As the AR gene is carried on the X chromosome some of these diseases result in somatic mosaicism [ 35 ]. The long history of cattle domestication involving intensive sire-based selection, primarily focussed on growth and reproductive traits, may have selected against variation in the AR gene in cattle.
Comparison of allele frequencies in Bos taurus indicus and Bos taurus taurus breeds showed significant differences for 15 of 16 polymorphic poly-Q encoding genes. Since the repeat regions in these genes are subject to relatively high rates of mutation, then these data could suggest differences in the microsatellite mutational rates within the two breeds. This possibility seems unlikely as there were no systematic effects on the repeat sizes or the number of alleles for each gene. An alternative explanation is that the allele frequency differences reflect variation that has independently arisen since divergence of these two cattle breeds from an ancestral population. The frequencies of alleles for four genes, CACNA1A, FXC1, HTT and TBP were significantly different for Bos taurus taurus cattle specialised for milk or meat production. All of these genes have ubiquitous expression patterns (GEO database accessions; GDS182 (mouse) and GDS1402 (human)). TBP (TATA box binding protein) and HTT are transcriptional co-regulators located in the nucleus, CACNA1A is a subunit of a plasma membrane voltage dependent calcium channel and, FXC1 is a component of a hetero-oligomeric translocase complex present in the mitochondrial inner membrane where it plays a role in the import of proteins into the inner mitochondrial membrane. Thus, structural variations that impact on the functions of these proteins have potential to modify broad transcriptional activities and metabolism. These influences may be subject to different selective pressures in beef and dairy cattle populations. Notably, extreme polymorphic repeats in three of four of these genes ( CACNA1A , HTT and TBP ) are associated with a spectrum of human diseases [ 53 ].
FXC1 was the most polymorphic gene with 18 detected alleles. Although the evolutionary conserved FXC1 encoded protein has not been directly associated with any human diseases, a loss of function mutation in a related component of the mitochondrial hetero-oligomeric translocase complex, TIMM8A, has been linked with the neurodegenerative disorder, Mohr-Tranebjaerg Syndrome [ 53 ]. The extensive protein-protein interactions occurring in the inner mitochondrial membrane translocase complex suggest that variation in FXC1 protein sequence caused by polymorphisms in its poly-Q region may have potential for impacting mitochondrial function [ 55 ]. Alternatively, the positioning of the poly-Q tract near the N-terminus of this small protein may have insulated its primary functions from the influence of polymorphic variation.
The number of poly-Q encoding genes in the cow (178) was similar to human (201) especially taking into account that the former number was determined using a draft genome assembly which is missing representation of all genomic sequence (~90% coverage) and does not have completed gene models for all genes [ 19 ]. Surprisingly, there was only 61% (109/178) gene overlap between cow and human indicating substantial differences between the species, although the functional enrichments in both groups were similar [ 11 ]. The figure of 201 human poly-Q encoding genes (Q ≥ 5) is consistent with an independent figure of 158 human poly-Q genes (Q ≥ 4) in a dataset of 6,477 vertebrate orthologs [ 37 ]. However, the subset of 123 bovine CAG poly-Q genes (≥ (CAG) 5 ) was substantially larger than that estimated for humans (66-75) and rodents (53-58) [ 2 , 11 , 12 ]. This observation may partially reflect the impact of the large expansion of CAG repeats in cow genomic DNA compared with other mammalian species (~10 fold increase compared with humans) [ 19 ]. As the scale of the repeat expansions in bovine coding sequences is much less than the genome-wide figure, it is likely that coding sequence CAG repeats are subject to different selective pressures compared with non-coding intergenic CAG repeats, as has been previously reported [ 13 , 48 ].
A total of 26 bovine poly-Q encoding genes did not have murine and human orthologs with poly-Q tracts. This group of genes potentially highlights evolutionary adaptations that contribute to the unique biology of ruminants. While some of these encoded proteins contained unique poly-Q tracts, others were identified because there were no orthologs in either the mouse or human genomes or the protein models did not contain sequence corresponding to the orthologous region of the bovine poly-Q tract proteins. The latter case may reflect species specific differences in alternative splicing. More trivial explanations include potentially incomplete gene and protein models or that the Q-tracts in the human and murine orthologs did not reach the threshold size of five residues used to define a Q tract. Notably, these 26 genes were not different from all bovine poly-Q encoding genes in relation to encoded protein size or function.
Comparison of bovine and human poly-Q proteins revealed that the lengths of ortholog Q-tracts were generally similar (Figure 1 ). This result is different from the finding that rodent Q-tracts were generally shorter than human Q-tracts [ 2 ]. As the maximum length of the bovine CAG Q-tract was directly related to allelic diversity (Figure 3 ) then it may be concluded that the overall extent of polymorphic variation in the poly-Q encoding genes of cattle is likely to be similar to human populations.
Poly-Q encoding genes can potentially generate considerable transcriptional diversity through a variety of mechanisms. The tissue-specific mRNA splicing of NFYA alters the length of the poly-Q encoded region in the absence of polymorphic variation. The poly-Q region is part of the activation domain of this transcription factor and modifications to the region have marked effects on function [ 33 ]. Thus, the tissue specific splicing may be a means of generating NFYA functional variants which could impact the transcription of the larger repertoire of genes that are regulated by NFYA. Although no poly-Q tract polymorphic variants in NFYA were demonstrated in the current analysis, if they existed, such allelic variants could exert additional tissue specific functional influences through this splicing mechanism. Polymorphic repeat regions in other poly-Q encoding genes traverse splice sites (e.g. ABCF1 ) or are closely juxtapositioned with splice sites (e.g. MED12 ). Moreover, a considerable number of poly-Q encoding genes contained repeats immediately adjacent to or at splice sites. In the case of MED12 this was associated with considerable transcriptional variation likely caused by the influence of coding sequence CAG repeats on splicing. There are similarities between CAG repeats and splice donors (consensus sequence (A/C)AG at the 3' end of the exon) and splice acceptors (consensus sequence (C)AG at the 3' end of the intron). This similarity could cause additional mRNA splicing variation in poly-Q encoding genes. Notably the protein MBNL1 (muscleblind-like) is involved in pre-mRNA splicing site choice and it also binds strongly to transcribed CAG repeats [ 56 ]. This dual function reinforces the view that splice site choice can be influenced by coding sequence CAG repeats. The presence of poly-Q encoding tracts contained within small alternatively spliced exons for a considerable number of poly-Q encoding genes suggests that this is another mechanism which generates considerable transcriptional and functional diversity. Indeed, the latter is accentuated by the large size of many poly-Q encoding genes, which often encode multiple autonomous functional domains each encoded by discrete exons. Alternative splicing of these exons is frequent and can generate many combinatorial mRNAs for each of these genes, possibly producing a range of protein products with subtle functional variations and a diversity of tissue specific variants.
The bovine poly-Q proteins were strongly enriched for large multi-domain transcriptional regulators (Additional file 4 ). Moreover, many of these proteins participate in large and common physical interaction networks as well as an extensive network associated with their transcriptional regulation (Figure 4 ). The central importance of CREBBP in this nuclear network is emphasised by its ability to act as a transcriptional co-activator in conjunction with EP300 to help assemble large regulatory protein complexes at sites of active transcription. Both of these proteins have intrinsic histone acetylase activity, which is involved in modifying chromatin structure in nucleosomes and thereby regulating gene expression [ 53 ]. The broad tissue expression patterns of the proteins in these nuclear complexes indicate that they are co-expressed in cells and hence these regulatory networks are likely to be biologically operational, with their components acting cooperatively in large transcriptional regulatory complexes. There could be even greater functional complexity inherent in these complexes caused by polymorphic variation in some of these proteins. Indeed, there is potential for amplification of the phenotypic impact of poly-Q polymorphic variations of multiple components in this regulatory complex - an epistatic effect in genetics terms. Even in the absence of polymorphic variation, alternative splicing influenced by CAG tracts may generate poly-Q protein functional diversity that could impact the activities of large transcriptional regulatory complexes in the nucleus. This process could generate additional functional complexity tailored to the needs of specific tissues, developmental pathways or tissue responses to external stimuli. | Conclusions
A total of 178 poly-Q encoding genes have been identified in the bovine genome. By using a hierarchical prioritization process we established that at least 16 of 32 top ranked genes were polymorphic in regions encoding poly-Q tracts. There was significant correlation between the extent of allelic diversity and the length of the poly-Q tract. This information, in conjunction with the purity of the repeat region encoding the poly-Q tract and the lack of conservation of repeat length, provide indicators for the propensity for polymorphic variation in the full repertoire of poly-Q encoding genes. Unlike some human poly-Q encoding genes there was no evidence for repeat instability, which may be attributable to the lack of large repeat expansions in the normal cattle population. This may be an inherent biological characteristic of cattle or a reflection of strong selective pressures in breeding programs that exclude these variants. Polymorphic variations in bovine poly-Q proteins have strong potential to generate epistatic biological effects due to the involvement of these proteins in transcriptional regulatory complexes and protein-gene regulatory interactions. Thus, the bovine poly-Q encoding genes may be contributing to phenotypic variation in cattle populations both directly and indirectly through epistatic interactions. CAG repeat tracts in poly-Q encoding genes can also influence splice site choice and thereby increase the functional diversity of transcripts from these genes. | Background
About forty human diseases are caused by repeat instability mutations. A distinct subset of these diseases is the result of extreme expansions of polymorphic trinucleotide repeats; typically CAG repeats encoding poly-glutamine (poly-Q) tracts in proteins. Polymorphic repeat length variation is also apparent in human poly-Q encoding genes from normal individuals. As these coding sequence repeats are subject to selection in mammals, it has been suggested that normal variations in some of these typically highly conserved genes are implicated in morphological differences between species and phenotypic variations within species. At present, poly-Q encoding genes in non-human mammalian species are poorly documented, as are their functions and propensities for polymorphic variation.
Results
The current investigation identified 178 bovine poly-Q encoding genes (Q ≥ 5) and within this group, 26 genes with orthologs in both human and mouse that did not contain poly-Q repeats. The bovine poly-Q encoding genes typically had ubiquitous expression patterns although there was bias towards expression in epithelia, brain and testes. They were also characterised by unusually large sizes. Analysis of gene ontology terms revealed that the encoded proteins were strongly enriched for functions associated with transcriptional regulation and many contributed to physical interaction networks in the nucleus where they presumably act cooperatively in transcriptional regulatory complexes. In addition, the coding sequence CAG repeats in some bovine genes impacted mRNA splicing thereby generating unusual transcriptional diversity, which in at least one instance was tissue-specific. The poly-Q encoding genes were prioritised using multiple criteria for their likelihood of being polymorphic and then the highest ranking group was experimentally tested for polymorphic variation within a cattle diversity panel. Extensive and meiotically stable variation was identified.
Conclusions
Transcriptional diversity can potentially be generated in poly-Q encoding genes by the impact of CAG repeat tracts on mRNA alternative splicing. This effect, combined with the physical interactions of the encoded proteins in large transcriptional regulatory complexes suggests that polymorphic variations of proteins in these complexes have strong potential to affect phenotype. | Authors' contributions
RT conceived and designed the experiments, analysed the results and wrote the manuscript. VW performed the gene prioritization and genotyping. MH, RT and YS-L identified and annotated the bovine poly Q encoding genes. WB provided DNA samples from a population of cattle while MK and HR supplied trio samples. All authors read and approved the final manuscript.
Supplementary Material | Acknowledgements
We are grateful to Frank Nicholas (Sydney University) for suggestions and support. We also thank Blair Harrison and Rowan Bunch for assistance in sample collection. The research was supported by Dairy Australia through the Innovative Dairy Cooperative Research Center. | CC BY | no | 2022-01-12 15:21:38 | BMC Genomics. 2010 Nov 23; 11:654 | oa_package/76/9d/PMC3014979.tar.gz |
PMC3014980 | 21118519 | Background
An exciting challenge of biological research has been to understand phenotypic diversity within a species, which affects virtually every organ and cell type. In plants, this intraspecific diversity is often readily visible in the size, shape, color and number of flowers, fruits and seeds. Diversity can occur in every region of the gene, in coding regions or in regulatory sequences including upstream promoter and enhancer sequences, 5' and 3' UTRs and regulatory introns [ 1 , 2 ]. Changes in regulatory regions affecting allele expression and transcript amount can be simple, such as small and large indels, or more complex, such as transposon insertions and structural rearrangements. Molecular mechanisms responsible for the sequence modifications are replication errors, recombination and transposition. Although the majority of allelic variation is due to nucleotide polymorphisms, phenotypic differences can be caused by epigenetic modifications such as DNA methylation [ 3 , 4 ].
Sequence comparisons among inbred lines revealed that maize is a highly polymorphic species regarding genes and intergenic space [ 5 ]. Consequently, maize lends itself to studying the molecular basis of phenotypic variation. As an example for a detailed allelic analysis, we chose the p1 locus, which maps to the short arm of chromosome 1, for several reasons: 1) p1 produces a visible, quantitative phenotype in different tissues (Figure 1A ). It encodes a R2R3 Myb-like transcription factor that activates the structural genes c2 , chi1 and a1 of the phlobaphene biosynthesis pathway (Figure 1B ) [ 6 ]. Phlobaphenes, which are reddish flavonoid pigments, accumulate in male and female floral organs. 2) The p1 gene is dispensable for the organism. Loss-of-function or change-of-function alleles will not be eliminated from the gene pool. 3) The p1 gene is characterized by its tremendous allelic diversity. More than hundred p1 alleles with distinct spatial and temporal expression pattern are reported although only few are molecularly defined [ 7 ]. 4) Approximately 2.75 million years ago, the p1 gene arose as a tandem duplication of the p2 gene [ 8 ]. If the p2 gene is the older of the two, it probably is the orthologous gene copy to the p3 gene on chromosome 9 because maize arose by allotetraploidization about 5 mya [ 9 , 10 ]. Therefore, we also refer to p2 as ortholog and p1 as paralog. The p2 gene is not involved in phlobaphene pigmentation, but like p1 , is a QTL for maysin production (Figure 1B ) [ 11 ].
Phlobaphene pigmentation is most readily visible in the pericarp, i.e. the outer layer of the kernel, and the cob glumes. Traditionally, p1 alleles are phenotypically categorized and named based on expression in these tissues. The p1 gene designation is followed by a two-letter suffix that refers to pericarp and cob color, respectively. For instance, the P1-rr allele exhibits red pericarp and red cob glume pigmentation while the P1-rw allele has red pericarp and white or colorless cob glumes (Figure 1A ). Each phenotypic p1 group may consist of structurally very different alleles. Only few p1 alleles have been structurally determined of which only a small number has been completely or partially sequenced. P1-rr4B2 [ 12 ] and P1-rw1077 [ 13 ] are single copy genes that both were introgressed into the inbred line 4Co63. This inbred line contains a loss-of-function p1-ww allele. P1-wr in inbred line B73 is a multi-copy allele, consisting of 11 P1-wr tandem repeats that are flanked by p2/p1 and p1/p2 hybrid genes upstream and downstream of the cluster, respectively [ 10 ]. A large retroelement cluster is inserted in the 3' UTR of the p1/p2 hybrid gene.
The p1-ww alleles do not encode a functional P1 transcription factor; therefore pericarp as well as cob glumes are colorless (Figure 1A ). While loss-of-function alleles often result in deleterious or even lethal conditions for the organism, non-functional p1 alleles do not cause any impairment that would reduce the fitness of the mutant plant. The p1-ww alleles can vary in origin and structure. Most of the structurally known p1-ww alleles are derived from P1-rr by transposon insertions and/or excisions. The p1-ww1112 null allele, for example, arose from a transposon-induced recombination event between the 5.2-kb direct repeats, which led to the deletion of the entire coding sequence [ 14 ]. However, the origin of p1-ww allele in the inbred line 4Co63 is not known, but p1-ww[4Co63] is often used in genetic crosses. Brink, for instance, introgressed more than 100 p1 alleles in the inbred line 4Co63 [ 7 ]. Knowledge of the p1-ww[4Co63] sequence could help clarify whether p1-ww[4Co63] is derived from P1-rr [ 12 ], P1-wr [ 10 , 15 ], P1-rw [ 13 ] or even a different p1 allele and provide further insights into other intermediates of chromosomal rearrangements.
To shed light on the origin of p1 allelic variability, we analyzed here three p1 alleles in their chromosomal context, namely p1-ww[4Co63] , P1-rr4B2 and P1-rw1077 . First we resolve the structural organization of these p1 alleles and their corresponding p2 alleles on the single-nucleotide sequence level. Subsequently we compare their sequences also to the recently sequenced P1-wr[B73] cluster [ 10 ] to find large and small scale nucleotide polymorphisms that enable us to infer mechanisms for genome rearrangements. In particular, we focus on evolutionary changes in p1 alleles that occurred in the putative distal enhancer region and in the 3' UTRs. | Methods
Plant material
Seeds containing P1-rr4B2 and P1-rw1077 alleles, which were introgressed in a 4Co63 background, were thankfully provided by Tom Peterson, Iowa State University. The inbred lines B73 and 4Co63 carrying P1-wr and p1-ww alleles, respectively, were obtained from the Maize Genetics Cooperation Stock Center (maizecoop.cropsci.uiuc.edu) collection. Traditionally, p1 alleles are classified and named according to their pericarp and cob glume pigmentation, implicating that phenotypically similar but structurally different alleles share the same name. In this report, we use the inbred line where the p1 allele was originally described in as additional allelic designation such as P1-wr[B73] and p1-ww[4Co63] . Similarly, the inbred line will be used as allele description for p2 , for example p2[4Co63] . Whenever the p2 source is unknown, the name of the linked p1 allele will be added to p2 , such as p2[P1-rr4B2] and p2[P1-rw1077] .
p1-ww[4Co63] isolation and sequencing
The inbred line 4Co63 contains a p1-ww allele, according to the colorless pericarp and cob phenotype of 4Co63 ears. We constructed a size-restricted lambda library using a lambda DASH II/ Eco RI vector kit (Agilent Technologies) and Eco RI-digested 4Co63 genomic DNA. The lambda library was screened by hybridizing filters with probe 15 [ 12 ], which is derived from a distal enhancer fragment of P1-rr and is unique to p1 alleles. Two positively hybridizing lambda clones were isolated and subcloned into pBluescript II SK+ vectors (Agilent Technologies). Insert size and both end sequences of each clone were determined and found identical. A transposon minilibrary (Finnzymes) of one clone was constructed according to the manufacturer's instructions. Sequencing was performed with the ABI PRISM BigDye Terminator Cycle Sequencing Ready Reaction kit and an ABI 3730 capillary sequencer (Applied BioSystems). Sequence assembly and analysis were carried out using Lasergene (DNAstar) programs. Sequence gaps were closed by primer walking.
p2 amplification and sequencing
Genomic PCR was performed to amplify p2 alleles. PCR primers (see Table 1 ) (Figure 1B ) were designed based on p2 sequences from p2[p1-ww1112] [ 8 ], p2/p1[B73] and p1/p2[B73] [ 10 ]. The PCR-amplified products were cloned into pGEM-T Easy vector (Promega). The individual clones were completely sequenced using primers that are spanning the entire repeat length (approximately one primer every 300 bp, primer sequences available upon request). The sequencing reactions were carried out with the ABI PRISM BigDye Terminator Cycle Sequencing Ready Reaction kit and analyzed on an ABI 3730 capillary sequencer (Applied BioSystems). The sequences were assembled and evaluated with the Lasergene software (DNAstar).
Sequencing of P1-rr 3' noncoding region and flanking genes
The majority of P1-rr sequence was determined in P1-ovov1114 (orange variegated pericarp and cob) that is derived from P1-vv . The Ac element of P1-vv located in the second intron excised and reinserted 161 bp further upstream in the opposite direction [ 56 ], still allowing a considerable amount of phlobaphene accumulation in pericarp and cob. Similarly, P1-rr4B2 is a P1-rr revertant that also originated from P1-vv by Ac excision. When not otherwise specified, we use P1-rr (without additional allele designation) to refer to functional P1-rr alleles that are derived from the same P1-vv . Two Eco RI fragments, isolated from P1-ovov1114 , were cloned in lambda using two Eco RI recognition sites outside of P1-ovov1114 . The third site was provided by the Ac transposon [ 18 ]. The 3' fragment of 14.5 kb was further divided in two plasmids, SA206 and PA103, which we gratefully received from Thomas Peterson. A transposon minilibrary of both plasmids (Finnzymes) was constructed as per the manufacturer's protocol. Clones were sequenced using transposon primers, ABI 3730 capillary sequencers, and the ABI PRISM BigDye Terminator Cycle Sequencing Ready Reaction kit (Applied BioSystems). Both plasmids contain 12418 bp non-overlapping P1-rr and 3' flanking sequences.
Amplification and sequencing of p1 intergenic region and flanking genes
Genomic PCR was performed to amplify p1 intergenic region and flanking genes. PCR primers (see Table 1 ) were designed based on corresponding sequences from B73 [ 10 ] and P1-ovov1114 (this study). The PCR products were cloned, sequenced and analyzed as described above for p2 .
Amplification of 3' cDNA ends (3' RACE)
Total RNA was extracted from pericarp tissue 20 days after pollination and emerging silk with the RNeasy Plant Mini Kit (Qiagen). RNA was reverse-transcribed to cDNA using the GeneRacer Kit (Invitrogen) with the GeneRacer oligo(dT) primers. cDNA was PCR amplified with the GeneRacer 3' primer and a gene-specific primer (see Table 1 ). In general, 96 RT-PCR products per primer pair (but only 18 for P1-rr4B2 samples) were cloned into pGEM-T Easy vector (Promega) and sequenced with universal primers. DNA sequences were analyzed with Lasergene (DNAstar) software. Polyadenylation sites were only plotted in Figure 6A to 6C , when they occurred more than once.
Sequence annotation and GenBank accession numbers
The maize sequences were manually annotated using homology searches in various GenBank databases with multiple BLAST programs [ 57 ]. The sequences were submitted to GenBank and were assigned following accession numbers: p2[4Co63] : HM454271 , p2[P1-rr4B2] : HM454272 , p2[P1-rw1077] : HM454273 , p1-ww[4Co63] : HM454274 , p1-ww[4Co63] 3' flanking region: HM454275 , P1-ovov1114 3' end: HM454276 | Results
The structural organization of p1-ww[4Co63] and linked p2 gene
A partial genomic lambda library was constructed using Eco RI-digested 4Co63 DNA. Filters were screened with the probe p15 (Figure 2 ), which hybridizes to a distal enhancer region thought to be present in all p1 alleles at the time of p1-ww[4Co63] cloning. A lambda clone containing 11,073-bp genomic DNA was isolated and sequenced [GenBank: HM454274 ]. Interestingly, this sequence is 99.7% identical to the displaced p1/p2[B73] 3' UTR and its 3' intergenic region, starting from an Eco RI recognition sequence in the retrotransposon Opie , and ending in an Eco RI site in the retroelement Shadowspawn (Figure 2 ).
The extensive similarity with the p1/p2[B73] 3' end and intergenic region together with the identical Shadowspawn insertion suggests that both sequences continue to be similar past the end of the lambda clone. To confirm this assumption, we extended the sequence by genomic PCR from the Shadowspawn element to neighboring genes that are unrelated to p and therefore do not participate in potential p recombination events. PCR primer pairs were designed based on the equivalent P1-wr[B73] cluster, and PCR products were cloned and sequenced [GenBank: HM454275 ]. The analysis of 6,587 bp revealed that 4Co63 and B73 are virtually identical in this sequence; they consist of the 3' end of the Shadowspawn retroelement, a gene encoding a calmodulin-binding protein, part of a gene encoding a protein of unknown function, and intergenic regions (Figure 2 ). The calmodulin-binding protein, which in 4Co63 measures 361 aa, is 7 aa larger than in B73 and also contains two amino acid substitutions due to 3 indels and 10 SNPs. Based on maize and other EST data, this gene is transcribed and is very conserved in grass species such as rice, sorghum and barley.
Although we could not find any remnants of p1 in the 3' flanking regions, we also needed to extend the analysis to the 5' flanking regions. We knew that a functional p2 gene, as visualized in silk browning reactions, still had to be present. The browning reaction of freshly cut back silk correlates with silk maysin concentration and is therefore induced by both p genes [ 16 ]. To extend the sequences of p1-ww[4Co63] up to the p2 gene, we used a genomic PCR approach by taking advantage of existing p2 , p2/p1[B73] and p1/p2[B73] sequences for the primer design. We sequenced a total of 10,753 bp that include the complete p2 and flanking sequences [GenBank: HM454271 ] (Figure 3 ). No p1 fragments or traces were detected. The p2[4Co63] and p2[p1ww1112] alleles with their flanking sequences are 94.56% identical. They differ in 57 SNPs and multiple indels, of which the largest consists of a Heartbreaker MITE insertion of 317 bp in the second intron. Because the 1,008 bp coding sequences of the two duplicated genes only vary by one synonymous substitution each in exon 2 and 3, their deduced P2 protein sequences of 335 aa are identical (Additional file 1 : Supplemental Figure S1). RT-PCR experiments confirm that p2[4Co63] is expressed in silk tissue as predicted by the silk browning reaction (data not shown).
The sequence 246 bp downstream of the p2 stop codon is composed of a partial Eninu retroelement of 540 bp followed by a Ji retrotransposon, which covers the remaining 1,322 bp of the available sequence (Figure 3 and 4B ). The 4Co63 p2 allele differs from the B73 p2 sequences extracted from the p2/p1[B73] and p1/p2[B73] alleles in many SNPs and indels including transposon insertions, suggesting that p1-ww[4Co63] may not have arisen from P1-wr[B73] by recombination events in a direct lineage.
Interestingly, the retrotransposons at the 3' end of p2 , namely Eninu and Ji , and at the 5' end of the above described " p1-ww " lambda clone, namely Opie and Eninu , are identical to the retroelement cluster of P1-wr[B73] in sequence, insertion site and consequently target site duplications. Although we did not clone and sequence the complete retroelement cluster in p1-ww[4Co63] it is most likely that both clusters in 4Co63 and B73 are identical, at least in their initial transposition of Eninu and their nested insertions of Ji and Opie (Figure 3 ).
In brief, whereas p2 is present and functional in the 4Co63 inbred line, p1 coding and regulatory sequences are missing with the exception of the distal enhancer region. The structure of the p1-ww[4Co63] allele does not unambiguously point to a single known p1 allele where p1-ww[4Co63] is derived from, although, mechanistically, unequal crossing over between flanking sequences of the p1 gene could have been involved as discussed below.
The structural organization of P1-rr4B2 and linked p2 gene
How does the sequence arrangement of P1-wr[B73] and p1-ww[4Co63] including their flanking genes compare to P1-rr4B2 , a p1 single-copy allele that produces red pericarp and red glumes? P1-rr4B2 contains two large repeats flanking the coding sequence, which are about 5.2 kb in size [ 6 , 12 , 17 ]. Interestingly, the sequence upstream of the 5' large repeat contains fragments of Opie and Eninu retroelements inserted in the same position as in p1-ww[4Co63] and p1/p2[B73] as described above (Figures 2 and 3 ). Likewise, Eninu is bordered by the detached p 3' UTR sequence of 78 bp. Subsequently, P1-rr is highly similar to a single P1-wr[B73] copy with few exceptions: the upstream regulatory region is more complex in P1-rr than in P1-wr[B73] (Figure 2 ) and both sequences diverge shortly after the stop codon (see below). By sequencing two plasmids, SA206 and PA103, which contain the 3' large repeat and are derived from lambda clones used for the isolation of P1-rr 5' and coding sequences [ 18 ] (see Methods), we extended our P1-rr sequence analysis by 8,923 bp past the 3' large repeat and into flanking genes [GenBank: HM454276 ]. By aligning both large repeats we found 14 polymorphisms including the insertion of a transposable element of 1,616 bp in the 5' repeat (Figure 2 ). This element is flanked by 8-bp direct repeats (CCAGTGAG), which is typical for transposons of the hAT superfamily. The 3' large repeat following the upstream regulatory sequence resembles p1-ww[4Co63] but does not contain the Shadowspawn retrotransposon insertion (Figures 2 and 3 ). Furthermore, the final 4,341 bp of the plasmid insert, not related to P1-rr , are highly similar to the equivalent p1-ww[4Co63] and P1-wr[B73] sequences (Figures 2 and 3 ). The 3' flanking sequence contains one complete gene and one partial gene in opposite transcriptional orientation compared to P1-rr . The first gene, which is separated from P1-rr by 1,175 bp (measured from the end of the 3' P1-rr repeat to the stop codon), encodes the 4Co63-type calmodulin-binding protein consisting of 361 amino acids. No more than 609 bp of intergenic sequence divide the first from the second gene, of which only the final two exons are present in the plasmid clone.
The P1-rr sequence analysis revealed that P1-rr is located between the retroelement cluster and the gene encoding a calmodulin-binding protein. Most interestingly, the corresponding site in p1-ww[4Co63] and P1-wr[B73] is empty, i.e. this region does not contain a p1 gene copy (Figure 3 ). Based on the first maize p2 allele that was isolated from a line which contains the p1-ww1112 allele [ 8 ] we assume that a functional p2 allele of P1-rr4B2 is present upstream of the retroelement cluster because p1-ww1112 and P1-rr4B2 are both derived from the same allele. Furthermore, the p2[p1-ww1112] allele ends in Eninu and Ji retroelement fragments exactly like p2[4Co63] and p1/p2[B73] suggesting structural similarity among these alleles. Therefore, we decided to extend our sequence analysis to the p2 allele that is linked to P1-rr4B2 . We used the same genomic PCR strategy as described above to clone and sequence 10,423 bp of p2[P1-rr4B2] [GenBank: HM454272 ]. Indeed, the alignment of p2[P1-rr4B2] with p2[p1-ww1112] showed no SNPs but only four 1-bp indels that are not part of exons or putative regulatory sequences. Hence both p2 alleles are coding for an identical P2 protein (Additional file 1 : Supplemental Figure S1). As expected, p2[P1-rr4B2] is also flanked by Eninu and Ji retroelement sequences. Introgression of P1-rr4B2 in 4Co63 probably included p2 as well because p2[P1-rr4B2] differs from p2[4Co63] .
The structural organization of P1-rw1077 and linked p2 gene
The P1-rw allele specifies red pericarp and colorless cob glumes (Figure 1A ). In general, the structure of P1-rw1077 resembles P1-rr4B2 [ 13 ]. P1-rw1077 is a single-copy gene, which consists of a coding region flanked by two 6.3-kb direct repeats (Figure 2 ). The coding sequence of P1-rw1077 is chimeric in nature. While the 5' UTR is similar to p1 , the remaining coding region and adjacent Eninu and Ji retroelements (spanning about 6.9 kb) are p2 -like (Figure 3 ) [ 13 ]. Sequence alignments establish that the p2 fragment is more closely related to p2[P1-rr4B2] / [p1-ww1112] than to p2[4Co63] . Interestingly, the Ji retrotransposon is followed by a truncated P1-wr -like exon, which is not included in the P1-rw1077 transcript. This organization of sequences suggests that P1-rw1077 originated from a gene conversion event between p1 and p2 [ 13 ]. The P1-rw1077 sequence upstream of the 5' large repeat is very similar to the corresponding P1-rr4B2 region, suggesting that both alleles occupy the same chromosomal location. We confirmed this by PCR-amplification and sequencing of a 1,651-bp fragment that connects the 3' large repeat of P1-rw1077 with the gene encoding the calmodulin-binding protein. P1-rw1077 was introgressed in 4Co63, and indeed the 3' end of the intergenic region between the 3' large repeat and the neighboring gene is indistinguishable from 4Co63. Since P1-rw1077 and P1-rr4B2 occupy the same chromosomal position we wanted to find out whether the similarity extends to the region upstream of the retrotransposon cluster (Figure 3 ). We performed genomic PCR as described above to amplify and subsequently sequence 11,313 bp [GenBank: HM454273 ] that are 99.8% identical to the p2[4Co63] sequence. The 18 SNPs and 3 short indels, which are distributed over a consensus sequence of 10,703 bp, are not included in the p2[P1-rw1077] coding sequence and consequently do not alter the P2 protein sequence (Additional file 1 : Supplemental Figure S1). The polymorphisms between p2[P1-rw1077] and other p2 alleles suggest that this p2 sequence was introgressed together with P1-rw1077 into the 4Co63 background. This implies that the p2 part of P1-rw1077 , which is p2[P1-rr4B2] -like, is derived from a p2 source other than p2[P1-rw1077] . The p2[P1-rw1077] 3' sequence is also flanked by Eninu and Ji retroelement fragments, linking p2[P1-rw1077] to P1-rw1077 across the retrotransposon cluster (Figure 3 ).
In summary, while p1 alleles can be located on both sides of the retroelement cluster (Figure 3 ), complete p2 alleles have so far only been found upstream of the retroelement cluster.
Evolution of a putative distal enhancer by non-homologous end-joining and transposition
Because all known p1 alleles produce almost identical P1 proteins (Additional file 1 : Supplemental Figure S1), differential expression of p1 alleles could have evolved through changes in regulatory sequences, which control time-and tissue-specific p1 expression [ 17 ]. Sequences containing regulatory elements are only determined for P1-rr [ 19 , 20 ], but based on sequence similarities have likely the same function in other p1 alleles as well. While all known p1 alleles share the P1-rr promoter and proximal enhancer sequences, they differ in the sequence arrangement that contains the distal P1-rr enhancer. Comparing putative distal enhancer regions of p1 alleles reveals that the single P1-wr[B73] gene carries the simplest and therefore possibly the most ancestral form, which is confirmed by the presence of an almost identical enhancer region at the 3' intergenic region of the p2 gene in a wild relative of maize (Teosinte accession Zea mays ssp. parviglumis ) [ 8 ]. Complexity of this chromosomal region increased with P1-rw1077 and then P1-rr4B2 . Therefore, we can use the changes in sequence organization to explain the origin of the P1-rw1077 and P1-rr4B2 enhancer region within the P1-wr repeat context, where the 3' end of one copy equals the 5' end of the downstream copy.
P1-rw1077 is a complex allele that must have been shaped by multiple recombination events (Figure 3 ) [ 13 ]. Interestingly, the sequence following the p2 portion resembles the junction of two P1-wr[B73] copies in a head-to-tail arrangement, suggesting that P1-rw1077 arose from P1-wr -like tandem repeats. The sequence similarity between P1-rw1077 and P1-wr[B73] starts with the truncated exon 3. P1-rw1077 and P1-wr[B73] are identical until they diverge 1,001 bp after the truncated P1-rw1077 exon 3. The next 734 bp of P1-rw are of mixed origin and mostly unrelated to P1-wr . P1-rw continues its homology with P1-wr 3' of the 734-bp insertion, but not at the sequence where both alleles deviate from each other. Instead, the P1-rw sequence downstream of the insertion is identical to the region of P1-wr upstream of the insertion, i.e. the insertion is flanked by a 203-bp repeat sequence. The sequence after the point of divergence originated from an unknown Mu -like transposable element (MULE) in reverse orientation (Additional file 2 : Supplemental Figure S2). Based on BLASTN searches, the sequence consists of two MULE fragments that in a putative autonomous MULE would be separated by approximately 444 bp (for additional information on this new MULE family in the maize B73 genome see Additional file 3 ). While the initial 279 bp, starting with the TIR (GGAAAAAATT...), are derived from the MULE 3' end, the remaining 446-bp fragment stems from a sequence partially encoding the C-terminus of the MULE transposase. The final 9 bp (AACCTATGT) of the 734-bp insertion may represent filler DNA (see bottom panel of Figure 5 ). The 9-bp fragment is identical to a P1-wr[B73] sequence, which is located 27 bp downstream of the point of P1-rw1077 and P1-wr[B73] alignment. Filler DNA, which is usually found at repair sites of DNA double-strand breaks can be simple as described here or complex, consisting of a patchwork of multiple sequences. Filler DNA is associated with non-homologous end-joining and is usually derived from nearby sequences of either end of the break [ 21 - 23 ] (for additional information on the mechanism of NHEJ see Additional file 4 ).
P1-rr is structurally more complex than P1-rw1077 and a single P1-wr[B73] gene. P1-rw1077 and P1-rr contain the same MULE fragments and filler DNA inserted in exactly the same sequence position. However, a 1.2-kb duplication in P1-rr that partially includes the fragmented MULE suggests that P1-rr is derived from P1-rw1077 . This duplication results in the addition of a fourth exon, which is unique to P1-rr . A closer look at the P1-rr 3' UTR may help to shed light on the evolution of the P1-rr enhancer region (see bottom panel of Figure 6 ).
Whereas P1-rw1077 and P1-rr are identical in the initial 3' UTR, they diverge 35 bp following the stop codon. The next 13 bp (ATAATTGGGTCAC) in P1-rr originated from two separated P1-rw1077 sequences, 1,410 bp apart, implying a deletion event in P1-rr compared to P1-rw1077 . The 13-bp (ATAATTGGGTCAC) insertion in P1-rr can be assigned to P1-rw1077 sequences upstream and downstream of the deletion site. ATAATTGGG is duplicated 59 bp downstream and includes the first two bp of the MULE TIR. Obviously, the adjacent TCAC occurs frequently within the P1-rw1077 sequence. However, the closest TCAC can be located 21 bp upstream of the insertion site. The 13-bp P1-rr sequence subsequent of the point of divergence with P1-rw1077 is suggestive of filler DNA, indicating that a previous DNA double-strand break in P1-rw1077 was restored by the NHEJ pathway. A tandem duplication of 1,269 bp that comprises the majority of both MULE fragments and 3' flanking enhancer sequences generated the current P1-rr 3' end and enhancer region.
In summary, DNA double-strand breaks in a P1-wr -like tandem array were probably repaired by NHEJ events that could have resulted in the rearrangements and duplications of enhancer-carrying sequences and consequently in novel p1 alleles as discussed below.
The p alleles differ in their 3' UTRs and polyadenylation sites
Although p1 and p2 share nearly the same coding sequences, their downstream sequences vary remarkably (Additional file 5 : Supplemental Figure S3). Sequence alignments of p1 and p2 alleles revealed that the P1-rr4B2 and p2 divergence from P1-wr[B73] is caused by transposon insertions. In P1-rr4B2 , a Mu -like element was placed 109 bp downstream of the stop codon probably due to a deletion event (see above), and p2 alleles are followed by an Eninu retroelement 248 bp after the stop codon. The insertion sites close to the stop codon raise the question whether these transposable elements eliminated the transcription termination signals and the polyadenylation sites in the P1-rr4B2 and p2 3' UTRs. In general, the 3' UTR is also important for post-transcriptional regulation such as microRNAs and translational control, and gain or loss of cis elements within the 3' UTR could contribute to allelic diversity. Therefore, we decided to map the polyadenylation sites of these alleles.
The P1-wr[B73] coding sequence is not flanked by transposons, and its 3' UTR probably represents the original 3' UTR structure of all p alleles before transposon modifications. We performed 3' RACE experiments to identify the P1-wr[B73] 3' UTR using three different gene-specific primers (p1 race 5'-1 to 3, see Table 1 ) and three independent pericarp tissue sources. RNA was extracted 20 days after pollination (DAP) from plants that contain the P1-wr[B73] cluster. Since the combination of different primers and tissues gave the same result we merged the data sets as shown in Figure 4A . Interestingly, we detected 18 polyadenylation sites in P1-wr[B73] that are spanning 189 nt from 141 to 329 nt after the stop codon. However, 36% of transcripts are polyadenylated 301 nt after the stop codon. The P1-wr[B73] cluster consists of 11 P1-wr tandem repeats and a p2/p1[B73] hybrid gene that differ by few polymorphisms in their transcribed regions. As these 18 polyadenylation sites are not specific for a particular repeat, the polymorphisms apparently do not affect the polyadenylation signals.
RT-PCR results indicated that p2[4Co 63] , p2[P1-rr4B2] and p1/p2[B73] are expressed in silk tissue. Accordingly, we carried out 3' RACE experiments using total RNA from silk and three different primers (p2 race 5'-1 to 3, see Table 1 ), which hybridize to exon 3 of p2 . The RNAs extracted from p2[4Co63] , p2[P1-rr4B2] and p1/p2[B73] lines produced almost identical results with all PCR primers, which allows us to combine the data for ease of presentation (Figure 4B ). We found 19 polyadenylation sites in a 218-bp interval that is located between 139 and 356 nt past the p2 stop codon. Whereas seven minor polyadenylation sites (adding up to 21% of the total events) are upstream of the retrotransposon, 12 sites lie within the LTR, including the major site (33% of polyadenylated p2 mRNAs), which is 269 nt from the stop codon and 22 nt into the LTR. The sequence alignment between p2 and P1-wr[B73] shows that the main polyadenylation site of P1-wr[B73] is 87 bp past the point of p2 and P1-wr[B73] divergence. The equivalent p2 fragment was displaced by retroelement insertions, and therefore cannot serve its original function. Nevertheless, p2 was able to recruit alternative polyadenylation signals and sites located mostly in the Eninu LTR.
Subsequently, we performed 3' RACE experiments on P1-rr4B2 total RNA extracted from silk and one primer binding (p2 race 5'-3, see Table 1 ) to exon 3. This exon contains the 3' UTR of the alternatively spliced P1-rr4B2 transcript, which encodes the functional P protein. We sequenced significantly fewer clones compared to P1-wr[B73] and p2 and obtained fewer polyadenylation sites. While polyadenylation sites are distributed over 403 nt from 143 to 545 nt measured from stop codon, the first site is used most often (31%) (Figure 4C ). All seven polyadenylation sites are located in the MULE fragments, two within the TIR, the remainder in the transcribed part. Due to the partial deletion of the former 3' UTR alternative polyadenylation signals and sites had to be employed from adjacent sequences as described above. Note that the MULE borders P1-rr4B2 in opposite transcriptional orientation. A transcript from an intact member of the same MULE family could therefore produce antisense RNA that is complementary to P1-rr4B2 mRNA. | Discussion
Models for the evolution of p1 alleles
A distinguishing feature of the p locus is its tremendous allelic diversity, which makes it a preferable locus to study evolutionary changes and chromosomal dynamics on a larger and smaller scale. Although the grass family arose by an ancient whole genome duplication (WGD) event [ 24 ], the p gene has only a single ortholog in rice and sorghum, indicating that one copy was lost from the paleoploid ancestral genome. However, the more recent allotetrapolidization event, which formed the ancestor of maize about 5 mya [ 9 ], gave rise to two p copies located in the homoeologous regions of chromosomes 9 and 1. The copy on chromosome 1 was then duplicated in tandem 2.75 mya, thereby evolving into the current p2 (ortholog) and p1 (paralog) genes [ 8 , 10 ]. The bulk of retrotranspositions in most grasses occurred more recently. A series of nested insertions that split approximately 80 bp of the p 3' UTR occurred between 1.4 to 0.2 mya [ 10 ]. Although retroelements are highly repetitive in the genome, insertions of retroelements in a nested fashion create unique sequence junctions and become chromosomal markers [ 25 ]. However, we do not know whether the retroelements transposed into the paralog or ortholog repeat, or maybe even into a later-generated copy. A model proposed for the evolution of single-copy alleles states that the retroelement insertion occurred in the 3' UTR of p2 , thereby separating p1 , which turned into P1-rr and P1-rw [ 8 , 17 ] (Figure 3 ). In contrast, transposition into the 3' UTR of p1 retains the repeat structure and allows the amplification of additional copies by unequal crossover as suggested for the evolution of the multi-copy P1-wr[B73] allele [ 10 ] (Figure 3 ). Theoretically, only few recombination events are needed to transfer p1 and p2 sequences across the retroelement cluster. Therefore multi-copy alleles in a tandem array could have been created upstream, downstream or on both sides of the cluster simultaneously. These intermediate structures that enable us to discover the step-by-step evolution of all p alleles might still exist in the maize germplasm. Our current analysis allows us to present new and refined models for the evolution of p1-ww[4Co63] , P1-rr4B2 and P1-rw1077 .
Models for the evolution of p1-ww[4Co63]
p1-ww is a null allele because p1 specific sequences such as coding-, promoter-and proximal enhancer sequences are absent in 4Co63 (Figure 3 ). Is it possible that p1-ww[4Co63] represents a haplotype where the tandem duplication of the ancestral p gene never took place? According to this hypothesis, the nested retroelement insertions in 4Co63 that are identical to alleles containing p1 and p2 sequences must have happened before the p duplication event. However, since the p duplication occurred 2.75 mya, 1.37 million years before the first retroelement insertion, we can disregard this possibility [ 8 , 10 ]. Thus, the alternative explanation that a functional p1 allele was deleted to give rise to p1-ww[Co63] is more likely. The p1-ww[4Co63] structure does not reveal the functional p1 allele(s) and their deletion or recombination events that resulted in the current null allele. Considering that p1 alleles are located on both sides of the retroelement cluster, multiple recombination events could have occurred to create the p1-ww[4Co63] allele.
One possible scenario for the origin of p1-ww in inbred line 4Co63 is that the null allele, which carries a functional p2 , is derived from P1-wr[B73] . While unequal crossover among repeat sequences can lead to an increase of copy numbers, the alternative outcome is a reduction of repeats. During the evolution of the P1-wr[B73] allele, unequal crossover between the flanking genes of the cluster, namely p2/p1[B73] and p1/p2[B73] , could have caused the deletion of all P1-wr[B73] repeats, and would still have generated a functional p2 gene (Figure 7A ). However, p2 in 4Co63 differs by various SNPs and indels from the corresponding p2/p1[B73] and p1/p2[B73] sequences of P1-wr[B73] , indicating that P1-wr[B73] might not be the immediate progenitor for p1-ww[4Co63] .
p1-ww[4Co63] also could have evolved by a recombination event that involved two different p1 alleles. Unequal crossing over between p2 of P1-rr4B2 and p1/p2[B73] of P1-wr[B73] could have generated the current p1-ww[4Co63] structure and could have restored the p2 copy (Figure 7B ). Even then, the deletion of the original paralog would have been derived from the P1-wr[B73] allele. Nevertheless, this could not have happened recently (on an evolutionary time scale) because of sequence polymorphisms in the participating alleles. Interestingly, both p1-ww[4Co63] and P1-wr[B73] carry as a signature the Shadowspawn retroelement in the same position, indicating that p1-ww[4Co63] most likely derived from P1-wr[B73] in multiple steps.
In addition, we can envision a P1-rw -like allele, which is similar to P1-wr[B73] in the distal enhancer structure. Such a P1-rw allele has been described [ 17 ]. An unequal crossover between the large repeats flanking the coding regions duplicates the p1 gene or deletes the coding sequences, resulting in the p1-ww[4Co63] structure (Figure 7C ). This scenario resembles the origin of p1-ww1112 [ 14 ]. However, this model does not directly account for the Shadowspawn retroelement in p1-ww[4Co63] . All models demonstrate the complexity of the p locus and reveal the countless possibilities for recombination to occur whenever paralogous sequences are present.
Model for the evolution of P1-rw1077 and P1-rr4B2 , with focus on regulatory sequences
Despite the repeat structure of the P1-wr[B73] cluster, a single P1-wr[B73] copy has the least complicated p1 allele composition, followed by P1-rw1077 and then P1-rr4B2 . We hypothesize that P1-rw1077 originated from a P1-wr- like tandem array (Figure 5 ) because P1-rw1077 comprises a sequence fragment downstream of the p2 section that is virtually identical with the junction sequence of two P1-wr[B73] repeats in a head-to-tail assembly. This P1-wr tandem array could have been located on either side of the retroelement cluster.
A plausible sequence of events is as follows. A Mu -like element inserted into one of the P1-wr repeats 1,204 bp after exon 3 of the previous copy. Then an aberrant transposition event (abortive excision event) of this MULE caused a DNA double-strand break that enabled exonucleases to digest the unprotected DNA ends (Figure 5 ) thereby extending the gap into the adjacent P1-wr repeat. The deletion would have included MULE sequences (about 3.5 kb compared to a putative autonomous element) and almost the entire length of a P1-wr repeat (more than 12 kb). Non-homologous end-joining [ 21 - 23 , 26 ], copying a 9-bp sequence (AACCTATGT) that is located 27 bp downstream from the deletion endpoint, must have repaired the break. Due to the nature of tandem repeats, the large deletion described above results in small repeats of 203 bp that are flanking the MULE fragments. Interestingly, this duplication is part of a 1.2-kb sequence that contains the enhancer element of P1-rr .
A single P1-wr- like allele downstream of the retroelement cluster that is flanked by large repeats due to the retroelement insertion in p2 could have been converted into a tandem array by unequal crossover between the large repeats (Figure 7C ). Gene conversion events then could have transferred the altered region that originated at the 3' large repeat to the 5' large repeat where the distal enhancer sequence functions [ 17 ]. Alternatively, P1-rw1077 arose from P1-wr repeats upstream of the retrotransposon cluster. The sequence 3' of this cluster, which corresponds to the 3' intergenic region of p2 as found in the P1-wr[B73] cluster and p1-ww[4Co63] , is nearly identical with the 5' end of a P1-wr[B73] repeat over a stretch of 5.2 kb (Figure 2 and Additional file 6 : Supplemental Figure S4A). Due to this sequence similarity, a recombination event between p1-ww[4Co63] and the proposed P1-rw1077 precursor could have occurred that positioned P1-rw1077 downstream of the retrotransposon cluster (Additional file 6 : Supplemental Figure S4A). This arrangement assumes that the P1-rw1077 allele resembles p1-ww[4Co63] at the 5' end. After the recombination break point, P1-rw1077 has to be closer to P1-wr[B73] because, based on our model, P1-rw1077 is derived from P1-wr . Indeed, a sufficient amount of polymorphisms between the p1-ww[4Co63] and P1-wr[B73] alleles enables us to verify the predicted structure and to place the possible recombination site between 567 and 713 bp after the point of p1-ww[4Co63] and P1-wr[B73] alignment. Further recombination/gene conversion events contributed to the evolution of the present P1-rw1077 allele.
The presence of the MULE fragments and filler DNA in P1-rr in exactly the same sequence context as in P1-rw1077 agrees with our model that P1-rr continued to evolve from P1-rw1077 . In our model for the origin of P1-rr , we propose a second DNA double-strand break (DSB) that occurred in P1-rw1077 in between the stop codon and the MULE insertion (Figure 6 ). In contrast to the first DSB, there is no evidence for the participation of a TE, leaving the cause for the DSB unknown. Exonuclease activities expanded the gap until both ends were joined in a NHEJ fashion by synthesizing two short DNA pieces (filler DNAs) from sites close to the deletion end points into the repair site. The DSB repair caused a deletion of 1,410 bp across the repeat junction that spanned almost the entire sequence from the stop codon to the MULE fragments. Interestingly, this intermediate P1-rr structure can be found at the 3' end of P1-rr1088 , P1-rrCFS36 and P1-rwCFS342 [ 17 ].
The 5' transposon fragment happened to contain an 8-bp sequence close to the TIR (55-62 bp) that is present 1,269 bp further downstream as well. Unequal crossover between those 8 bp resulted in a tandem direct duplication of this 1,269 bp sequence. Accordingly, the final 318 bp of exon three, being part of the repeat, were replicated, too. A sequence at the 3' end of the first repeat was adopted as a splice acceptor site thereby generating a fourth exon. Although alternative splicing of exon 1, 2 and 4 has been reported, the protein product is of unknown function or may not have any function at all [ 27 ].
This putative evolutionary pathway explains how the P1-wr 3' UTR was almost entirely replaced by a MULE, how the fourth exon unique to P1-rr was generated and how the 1,269 bp SalI fragment containing the P1-rr distal enhancer was nearly completely duplicated (the initial 175 bp of the enhancer region are missing from the first repeat). Subsequently, gene conversion events could have placed part of the modified enhancer sequence of the downstream copy to the upstream large repeat [ 17 ]. Alternatively, if this P1-rr module arose on the P1-wr[B73] side of the retroelement cluster as we also discussed for P1-rw1077 , then a recombination event between p1-ww[4Co63] and the P1-rr ancestor could have transferred the P1-rr end to a position downstream of the retrotransposon cluster (Additional file 6 : Supplemental Figure S4B). The crossing over took place in the 595 bp sequence between the duplicated MULEs, which is why the repeat structure of P1-rr at the 5' end differs from the 3' end whereas they are identical in P1-rw1077 . Lastly, a 1.6 kb hAT -like transposable element inserted 340 bp upstream of the MULE or 159 bp 5' of the enhancer region. This transposition did not occur in P1-rr1088 [ 17 ]. Taken together, the novel distal enhancer structure of P1-rr could be the result of a MULE insertion and excision, deletion and repair by NHEJ, and duplication and deletion by recombination. This series of events from P1-wr to P1-rr confirms the sequential model of P1-rw and P1-rr evolution based on phylogentic analysis [ 17 ].
Function of the enhancer region rearrangements on p1 expression
When the p1 paralog was formed, it probably included the complete p coding sequence and the basal promoter that controls p expression in silk tissue. Then the paralog acquired two additional regulatory sequences adding equally to the basal expression in pericarp and glume. The enhancer sequences were identified and tested in transient and transgenic plants using P1-rr fragments fused to a GUS reporter gene [ 19 , 20 ]. A 1-kb sequence adjacent to the promoter contains a regulatory sequence termed proximal enhancer while a 1.2-kb fragment further upstream includes a distal enhancer (Figure 2 ). The proximal enhancer region corresponds mostly to a truncated MULE that captured part of a host gene in between the TIR [ 10 ]. The proximal enhancer region and the basal promoter sequence are virtually identical in all sequenced p1 alleles to date (Figure 2 ). In contrast, the distal enhancer region varies in all p1 alleles as described above. Therefore, we hypothesize that the different spatial and temporal expression patterns of p1 alleles are caused by distinct distal enhancer regions [ 17 ]. The distal enhancer as defined in P1-rr is located within a 1,269-bp Sal I fragment [ 19 , 20 ], out of which 671 bp are derived form the Mu -like transposon (Figure 2 ). Although this MULE fragment is missing in P1-wr[B73] , transgenes constructed from P1-wr upstream regulatory sequences linked to P1-rr cDNA produced red pericarp and cob glumes in transgenic plants [ 28 ], indicating that the enhancer sequence is included in the 589-bp region downstream of the MULE. Since this 589-bp region is duplicated in P1-rr , P1-rr has two enhancer sites that are separated by the MULE fragment. Additional P1-rr alleles, namely P1-rr1088 and P1-rrCFS36 , were shown to have the same enhancer structure as P1-rr4B2 with exception of the missing hAT insertion in P1-rr1088 [ 17 ]. Therefore, the hAT transposable elements inserted in the upstream copy of the enhancer region of P1-rr4B2 and P1-rrCFS36 obviously do not disrupt the enhancer sequence and function. Compared to P1-rr , P1-rw1077 has a deletion of 381 bp in the upstream repeat, which causes the loss of cob glume pigmentation [ 13 ]. Interestingly, two additional P1-rw alleles, P1-rwCFS302 and P1-rwCFS342 , lack the entire upstream repeat and the MULE fragment, thus having the identical enhancer arrangement as a single P1-wr[B73] copy [ 17 ]. Taken together, the analysis of three P1-rr and three P1-rw alleles revealed that P1-rr alleles contain two copies of the specific enhancer sequence while P1-rw alleles only have one [ 17 ]. Interestingly, this region coincides with a tissue-specific DNase I-hypersensitive site that remains closed in pericarp tissue of P1-pr , a silenced epiallele of P1-rr4B [ 29 ]; the P1-pr phenotype is shown in Figure 1A . It was reasoned that the upstream enhancer repeat that is missing in P1-rw1077 controls the glume-specific expression in a position-dependent manner, since the identical enhancer region is located 671 bp further downstream [ 13 ]. An alternative explanation was prompted by the fact that p1 expression in pericarp is weaker and delayed in P1-rw1077 compared to P1-rr . We hypothesize that the transcriptional strength of p1 alleles is correlated with the enhancer copy number, which is supported by similar findings in human upstream enhancers [ 30 ]. Consequently, P1-rw1077 produces less P1 protein than P1-rr in all expressing tissues. Also, each p1 allele is not expressed uniformly in female and male floral tissues within a plant. For example in P1-rr , p1 transcription is usually higher in pericarp than in cob glumes [ 15 ].
Therefore, we propose that the presence of only one distal enhancer site in the P1-rw1077 allele results in weak expression in pericarp tissue but no expression in cob glumes. Due to the duplication of the enhancer sequence as outlined in our model, p1 transcription in pericarp and glume tissue was equally elevated such that p1 is strongly expressed in pericarp and weakly expressed in glumes, thereby giving rise to P1-rr alleles. Note that comparisons with P1-wr alleles are not appropriate due to their post-transcriptional silencing, which potentially is repeat induced [ 31 ]. This model is supported by an analysis of the spatial expression pattern in transgenic plants where various p1 constructs were expressed only in few p expressing tissues, resembling P1-rr or P1-rw phenotypes. It was shown in these transgenic plants that p1 expression follows a spatial hierarchy that begins with pericarp and continues with cob glumes, husk, silk, and tassel glumes in decreasing order [ 32 , 33 ]. For instance, if the transgenes had been expressed in only one tissue, then it would have had to be in pericarp, in the case of two tissues then in pericarp and glumes, and so on.
The p alleles differ in their 3' UTR
Polyadenylation is involved in many facets of mRNA metabolism including enhancement of mRNA stability, transport of mRNA from the nucleus into the cytoplasm, and regulation of mRNA translation. Although polyadenylation signals in plants are less conserved than in mammals [ 34 ], three signals were identified in maize, rice, and Arabidopsis: the far upstream element (FUE, located -150 to -35 nt upstream of the cleavage site), the near upstream element (NUE, situated -35 to -10 nt upstream of the cleavage site) and the cleavage element (CE, positioned -10 to +15 nt upstream and downstreams of the cleavage site) [ 35 , 36 ]. As we have shown above, a fragmented MULE was placed adjacent to the P1-rr4B2 stop codon possibly due to a NHEJ event. All mapped polyadenlation sites of the P1-rr4B2 transcript are located within the MULE sequence, indicating that P1-rr4B2 successfully recruited alternative polyadenylation signals in the transposon. Similarly, a Mu insertion in the 3' UTR of the rf2a locus also resulted in the adoption of new polyadenylation signals and sites [ 37 ]. Retroelements, the most common transposons in maize, also insert in 3' UTRs without disrupting polyadenylation as demonstrated above for the p2 alleles. Our results suggest that polyadenylation in maize is a highly dynamic process which despite its importance for the cell is not tightly regulated. The large amount of polyadenylation sites found in our analysis of P1-wr[B73] transcripts that do not contain a transposon insertion supports this conclusion. A genome-wide analysis of genomic and transcript data could shed light on the mechanism of polyadenylation in maize and could establish the proportion of genes that terminate in transposable elements. Interestingly, it has been shown that many polyadenlylation signals in human and mouse genes have been derived from transposable elements [ 38 ]. Besides polyadenylation signals, transcriptional as well as translational regulators have been identified in the 3' UTR of plant and animal genes, and their gain or loss could cause allelic diversity. For example, targets of microRNAs are often located in 3' UTRs [ 38 , 39 ].
Gene copying events promote allelic diversity
Recombination is crucial for the evolution of genomes [ 40 , 41 ]. In particular, the non-homologous recombination pathway is frequently used to repair DNA double-strand breaks in somatic plant cells [ 26 ]. Previously, we reported a probable NHEJ event involved in the formation of the P1-wr[B73] cluster [ 10 ] that produced a hybrid gene due to the ligation of deletion end points located within two genes. Similarly, deletions and repair by NHEJ in the above mentioned alleles could have resulted in the restructuring of an enhancer region and formation of a novel 3' UTR.
The exceptional allelic variation at the p locus prompts the question about its similarities and differences to genes that exhibit less variation. We propose that the main cause for the diversity might lie in tandem gene amplification [ 8 , 17 , 42 , 43 ]. Once a gene underwent an initial tandem duplication, multiple unequal recombination events can follow as seen in the P1-wr[B73] multi-gene cluster [ 10 ]. A single crossing over or gene conversion event between misaligned paralogous gene copies can generate many new alleles including deletion and amplification derivatives. Interestingly, in plants such events can occur mitotically and can be transmitted into the next generation, thereby increasing allelic variation [ 44 ]. This explanation then implies that other loci exhibiting an increased allelic variation are multi-copy genes as well. Indeed, the complex r1 locus in maize is analogous to p1 in many aspects. The r1 locus, which also encodes a transcription factor, confers bluish anthocyanin pigmentation to various vegetative and floral plant tissues. Two r1 alleles, R-st and R-r , are molecularly well characterized. R-st contains various r1 genes, four of which are in tandem orientation [ 45 ]. R-r consists of one complete and three truncated r1 genes that originated from tandem duplication [ 46 , 47 ]. Comparable to p1 in complexity, both alleles undergo recombination and transposition events creating numerous derivative alleles. Paralogous gene copies in maize were also found at the pl1 [ 48 ] and a1 loci [ 49 ]. Especially the prolamine gene family with nearly 50 copies distributed over several chromosomes exemplifies the outcome of gene duplications [ 50 ]. Actually, a large proportion of genes are tandemly duplicated in Arabidopsis, rice, and maize [ 51 - 53 ]. Considering the amount of paralogous sequences and their possibilities to recombine, a single reference genome providing just one allele can obviously not reflect this allelic potential of the maize genome. Not surprisingly, a recent genomic comparison between the B73 and Mo17 inbred lines [ 54 ] revealed a large quantity of copy number variations and presence/absence variations confirming previous results [ 55 ]. Nonetheless, epialleles remain invisible in a traditional sequence comparison. Allelic diversity studies as presented here are essential for our understanding of the remarkably dynamic maize genome. | Conclusion
Allelic diversity is the source for evolution and domestication. While allelic variation in wild species ensures the best possible adaption to changing environmental conditions, humans have profited from allelic pools in crop plants by selecting phenotypic variations that best meet their needs. Alleles differ most often in small-scale nucleotide polymorphisms but also in large-scale sequence rearrangements. Maize has been shown to be a highly polymorphic species well suited to study genome dynamics and the underlying molecular mechanisms. In particular, the maize p locus with its well-established genetic history offers a tremendous amount of ancient allelic variations, some representing intermediate steps in large-scale sequence rearrangements. The tandemly duplicated p1 and p2 genes encode virtually identical Myb-like transcriptional activators, but only p1 controls the accumulation of reddish flavonoid pigments in maize female and male floral organs. Because all P1 proteins are almost identical, the phenotypic variation must be due to p1 regulation. Therefore, this locus represents an ideal example of how genomic rearrangements can contribute to novel regulatory elements.
Here, we used targeted genome sequencing to apply comparative genomics to the maize genome. Sequence alignments of orthologs and paralogs of different genotypes of a single genomic region allow us to reconstruct the repair of double strand breaks from transposition events within gene copies and their flanking regions. Such drastic invasions of new sequence elements in flanking regions result in the de novo creation of regulatory elements involved in the transcriptional and post-transcriptional regulation of gene expression that differentiate gene copies in their function. Interestingly, sequence chimerism in the 3' untranslated portion of the mRNA gave rise to multiple poly-A addition signals with similar strength, indicating a more relaxed sequence restriction of the 3' processing machinery than previously believed. | Background
The molecular mechanisms that modify genome structures to give birth and death to alleles are still not well understood. To investigate the causative chromosomal rearrangements, we took advantage of the allelic diversity of the duplicated p1 and p2 genes in maize. Both genes encode a transcription factor involved in maysin synthesis, which confers resistance to corn earworm. However, p1 also controls accumulation of reddish pigments in floral tissues and has therefore acquired a new function after gene duplication. p1 alleles vary in their tissue-specific expression, which is indicated in their allele designation: the first suffix refers to r ed or w hite pericarp pigmentation and the second to red or white glume pigmentation.
Results
Comparing chromosomal regions comprising p1-ww[4Co63] , P1-rw1077 and P1-rr4B2 alleles with that of the reference genome, P1-wr[B73] , enabled us to reconstruct additive events of transposition, chromosome breaks and repairs, and recombination that resulted in phenotypic variation and chimeric regulatory signals. The p1-ww[4Co63] null allele is probably derived from P1-wr[B73] by unequal crossover between large flanking sequences. A transposon insertion in a P1-wr -like allele and NHEJ (non-homologous end-joining) could have resulted in the formation of the P1-rw1077 allele. A second NHEJ event, followed by unequal crossover, probably led to the duplication of an enhancer region, creating the P1-rr4B2 allele. Moreover, a rather dynamic picture emerged in the use of polyadenylation signals by different p1 alleles. Interestingly, p1 alleles can be placed on both sides of a large retrotransposon cluster through recombination, while functional p2 alleles have only been found proximal to the cluster.
Conclusions
Allelic diversity of the p locus exemplifies how gene duplications promote phenotypic variability through composite regulatory signals. Transposition events increase the level of genomic complexity based not only on insertions but also on excisions that cause DNA double-strand breaks and trigger illegitimate recombination. | Authors' contributions
WG and JM conceived of the study. WG designed and carried out the experiments. WG analyzed the data. WG and JM wrote the paper. All authors read and approved the final manuscript.
Supplementary Material | Acknowledgements
We thank Hugo Dooner for critical reading of the manuscript. We are grateful to Thomas Peterson for kindly providing SA206 and PA103 plasmids and P1-rw1077 seeds. This work was supported by the Selman A. Waksman Chair in Molecular Genetics to JM. | CC BY | no | 2022-01-12 15:21:38 | BMC Genomics. 2010 Nov 30; 11:678 | oa_package/be/ae/PMC3014980.tar.gz |
PMC3014981 | 21126332 | Background
Several studies have attempted to compare protein and transcript expression levels to investigate the central dogma of the cell, i.e. the relation between DNA, RNA and protein content in a cell [ 1 - 8 ]. Microarrays have been the prevalent platform to measure the abundance of transcripts in a sample, although other technologies such as SAGE have also been employed. The corresponding protein abundance estimates have frequently been obtained through mass spectrometry or protein arrays. The resulting correlation coefficients in these comparative analyses have varied significantly, from 0.3 to 0.9, comparing 10 s of genes up to 1000 s of genes. Sub-groups representing functionally different Gene Ontology groups could, however, display both higher and lower correlations depending on their role in the cellular machinery [ 7 ].
To improve an estimate of correlation between RNA and protein molecules a more unbiased approach combined with a digital gene expression profile is needed. Massive DNA sequencing technology offers a new possibility to achieve a comprehensive and quantitative view of all genes being transcribed in a sample [ 9 - 11 ]. Here, we compare global IHC and IF protein expression in a human osteosarcoma cell line, U-2 OS (from the Human Protein Atlas program, HPA)[ 12 ], with massive DNA sequencing of the corresponding transcriptome (RNA-seq). | Methods
Cell cultivation
The osteosarcoma cell line U-2 OS (ATCC-LGC Promochem, Borås, Sweden) was cultivated in a 5% CO 2 environment at 37°C in McCoy's 5A media, as suggested by the provider, with the addition of 10% Fetal Bovine Serum (FBS) and an antibiotic/antimycotic solution (both from Invitrogen).
RNA sample preparation and cDNA synthesis
Cells were harvested and RNA was extracted using the RNeasy extraction kit as instructed by the manufacturer (Qiagen) and quality-assessed using the RNA nano kit on a BioAnalyzer 2100 (Agilent). The BioAnalyzer can interpret the generated data and score it with a RNA Integrity Number (RIN) ranging from 1 (very degraded) to 10 (no degradation). 10 μg of high-quality (RIN >9.5) total RNA was used as input material for depletion of ribosomal fragments using RiboMinus (Invitrogen). 250 ng ribosome-depleted RNA was adjusted to 4.5 μl in nuclease-free water and fragmented in 95°C for 20 minutes, after which it was immediately transferred to ice. 1 μl of biotinylated tagged random hexamers (Biotin-TEG-CTTTCCTCTCTATGGGCAGTCGGTGATNNNNNN, 1 pmol/μl, Operon) was added and the mixture was denatured at 70°C for 10 minutes and on ice for 2 minutes. During incubation, a cDNA-synthesis master mix consisting of (per sample) 6 μl 5x First-strand buffer (Invitrogen), 3 μl 0.1 M DTT (Invitrogen), 7.5 μl dNTP mixture (2 mM/dNTP) and 6.5 μl nuclease-free water was assembled on ice. Of this master mix, 23 μl was added to the denatured RNA:hexamer mixture together with 2 μl SuperScript III (Invitrogen) on ice. The first strand cDNA synthesis reaction was incubated at 20°C for 10 minutes followed by 37°C for 10 minutes and 42°C for 45 minutes. First-strand cDNA was purified by addition of 70 μl nuclease-free water using a MinElute spin column following the manufacture's instructions (Qiagen). Elution was carried out twice, each in 10 μl of EB-buffer, with a 180° rotation of the column in the centrifuge between the elutions. A second strand synthesis master mix was assembled on ice. This consisted of (per sample) 79 μl nuclease-free water, 30 μl 5x second-strand buffer (Invitrogen), 15 μl dNTP mixture (2 mM/dNTP), 1 μl 10 U/μl E. coli DNA Ligase (Invitrogen), 4 μl 10 U/μl E. coli DNA polymerase I (Invitrogen) and 1 μl 2 U/μl RNase H (Invitrogen). Of this, 130 μl was added to the ≈18 μl eluate containing the RNA:cDNA hybrid. The reaction was incubated at 16°C for 2 hours, after which 1.5 μl 3 U/μl T4 DNA Polymerase (New England Biolabs) was added, and the reaction was incubated for 5 more minutes at 16°C.
Enrichment and SOLiD DNA sequencing
20 μl of Streptavidin Dynabeads M-270 (Invitrogen) were washed in 50 μl 1× BW-buffer (20 mM Tris, 2 mM EDTA, 1 M NaCl) and pelletized using a MPC-6 magnetic particle concentrator (Invitrogen). To the pellet, the 150-μl second-strand syntheis reaction and 150 2× BW buffer were added, mixed by gentle vortexing and incubated on gentle rotation using a RotaMix (Elmi) at room temperature (≈22°C) for 15 minutes. The bead-DNA complex was then washed three times in 100 μl sterile deionised water and pelletized. A PNK-master mix consisting of (per reaction) 15 μl sterile deionised water, 2 μl PNK buffer (Invitrogen), 2 μl 10 mM ATP and 1 μl 10 U/μl PNK was assembled on ice and 20 μl was added to the pelletized beads and mixed by pipetting. The reaction was incubated at room temperature for 15 minutes. To create a blunt-ended dsDNA adapter sequence suitable for ligation to the beads, a mixture consisting of 440 μl sterile deionised water, 5 μl 100 pmol/μl RDV primer (AACTGCCCCGGGTTCCTCATTCTCT, MWG-Biotech), 5 μl 100 pmol/μl aRDV primer (AGAGAATGAGGAACCCGGGGCAGTT, MWG-Biotech) and 50 μl PNK buffer (Invitrogen) was assembled and incubated at 95°C for 3 minutes and allowed to cool to room temperature on a lab bench for 30 minutes. A ligation master mix consisting of 4 μl 5× Ligase buffer (Invitrogen), 1 μl 1 pmol/μl RDV:aRDV duplex, 14 μl sterile deionised water and 1 μl 3 U/μl T4 DNA Ligase was assembled on ice, added to the pelletized beads and mixed by pipetting. The reaction was incubated for 16 hours on a RotaMix at room temperature. The ligation reaction was washed three times in sterile deionised water and the beads were resuspended in 20 μl of sterile deionised water.
An amplification master mix was assembled, consisting of 5 μl 5× HF buffer (Finnzymes), 10 μl sterile deionised water, 5 μl dNTP mix (2 mM/dNTP), 1 μl 10 pmol/μl RDV primer (AACTGCCCCGGGTTCCTCATTCTCT, MWG-Biotech), 1 μl 10 pmol/μl LAmpFDV (CCACTACGCCTCCGCTTTCCTCTCTATGGGCAGTCGGTGAT, MWG-Biotech) and 1 μl 2 U/μl Phusion polymerase (Finnzymes). 23 μl amplification master mix and 2 μl beads were mixed denatured at 95°C for 30 seconds and cycled as follows: 30 seconds at 95°C, 30 seconds at 55°C, 30 seconds at 72°C for 16 cycles. After a final extension at 72°C for 10 minutes, the PCR product was purified using a MinElute column, following the manufacturers instructions with the elution step as described earlier.
The PCR product was subjected to emPCR and SOLiD sequencing following the manufacturers instructions (Life Technologies/Applied Biosystems). Fifteen million 35-base pair reads passed quality filters including filtering against adaptors.
Mapping of DNA reads and defining the transcriptome
The reads were mapped to the human genome (hg18) following the manufacturers instructions. A read was considered to be unique if is mapped to a single location with N mismatches and nowhere with N+1 or N+2 mismatches (i.e. a clear zone of 2). In total, 6 442 847 reads aligned uniquely to EnsEMBL genes. Only unique reads with a maximum of three mismatches were used to calculate expression values. A gene was considered present if at least one read fell entirely inside an exon of the gene. Raw sequence data has been deposited to the NCBI Short Read Archive with accession number SRA023713.1.
Production of Antibodies
The antibodies used in this study have been generated within the Human Protein Atlas program http://www.proteinatlas.org [ 16 ]. For each application using the generated antibodies, a standard set of categories has been established, and these are grouped into three main validation scores: (i) supportive, (ii) uncertain, or (iii) non-supportive [Additional file 1 : Supplemental table S1] [ 18 ].
Western blot analysis
Prior to immunohistochemistry and immunofluorescence the HPA antibodies were analysed by Western blot as previously described [ 16 ]. All membranes were incubated with the primary antibodies diluted 1:500. The secondary HRP-conjugated antibody (Swine Anti-Rabbit Immunoglobulin/HRP, DakoCytomation) was diluted 1:3000 and detection was carried out using a CCD-camera. Seven categories (1-87) for the Western blot validation have been proposed in which three (grade 1-3) are supportive , two (grade 4-5) are uncertain , and two (grade 6-7) are non-supportive .
Immunohistochemical analysis
Immunohistochemistry was performed on cell microarrays (CMA) where U-2 OS cells were represented, as previously described [ 18 ]. In brief, cells were harvested, fixed in formalin and dispersed into agarose. After histoprocessing and paraffin embedding of the cell pellets resulting in donor blocks, duplicate 0.6 mm punches were sampled and put into one recipient CMA. 4 micron sections were subsequently cut and immunohistochemically stained using an Autostainer Plus instrument (Dako, Glostrup, Denmark). An automated slide-scanner system, Scanscope T2 (Aperio Technology, Vista, CA, USA) was used to image the stained sections, resulting in digital images representing separated cell spots from the CMA.
The images were analyzed by an automated image analysis software, TMAx (Beecher instruments, Sun Praire, WI, USA), which identifies cells, measures the immunostaining and counts the fraction of stained cells, all according to processing logic previously described [ 19 ]. The measured staining intensity was further categorized into a IHC graded scale (negative, weak, moderate and strong). Here, the analyzed protein was considered "absent" when the staining intensity score was negative and "present" when the staining intensity score was weak, moderate or strong.
Immunofluorescence microscopy
Immunofluorescent stainings were performed as previously described [ 20 ]. Briefly, cells were seeded in 96-well glass bottom plates, fixed with paraformaldehyde and permeabilized with Triton X-100 before immunofluorescently stained. The entire procedure was automated using a pipetting robot. Besides the HPA antibody staining, organelle markers for microtubules, endoplasmic reticulum and nuclei were included. Image acquisition was performed manually using a LSM 510 Meta confocal laser-scanning microscope equipped with a 63x oil-immersion objective (Carl Zeiss GmbH, Jena, Germany). For each sample, two representative four-channel images were acquired. The laser power and detector gain were adjusted for each sample to obtain as good signal to noise ratios as possible and to use the entire dynamic range of the detector.
The images were visually inspected and annotated in terms of staining intensity and subcellular localization. A graded scale (negative, weak, moderate and strong) was used to categorize the staining intensity based on the used laser power and detector gain. For each antibody a validation score was set based on how well the observed subcellular localization agreed with information in the UniProt database. The IF validation scores consist of a nine-graded scale that can be merged into three main categories: supportive , uncertain or not supportive [Additional file 1 : Supplemental table S1]. Here, the analyzed protein was considered "absent" when the validation score was 7 (no staining, see [Additional file 1 : Supplemental table S1]) and "present" otherwise. | Results
The aim of this study was to compare the transcriptome of human U-2 OS cells with presence of the corresponding proteome. The transcriptome was extensively surveyed close to saturation by performing massive SOLiD DNA sequencing [Additional file 1 : Supplemental figure S1]. In total, approximately 15 million high quality 35-bp reads were obtained and mapped onto the human reference genome (hg18) and quantitative measures were computed on a per gene basis. Analysis of the transcription pattern demonstrated that the majority of all Ensembl genes (73.4%; 15536/21146 genes) were expressed in U-2 OS, i.e., a transcript being represented by at least one uniquely mapped read. The frequency distribution is presented in the additional information [Additional file 1 : Supplemental figure S2].
To create a comparative protein expression set, a non-redundant collection of antibodies and genes was assembled from the Human Protein Atlas [ 12 ]. In the initial collection of data, a high degree of protein presence was observed for both IHC and IF, demonstrating expressed proteins for 88.7% and 73.6% of all genes analyzed, respectively (Table 1 ). In the following analysis, all antibodies with protein expression data from both IHC and IF in the U-2 OS cell line were used. For the genes with more than one antibody directed towards the gene product, the best scoring IF antibody was selected according to a standard validation scheme [Additional file 1 : Supplemental table S1]. The assembled non-redundant set of antibodies was then used to collect corresponding immunohistochemistry and immunofluorescence information from U-2 OS, yielding the HPA subset. The HPA subset consists of 2749 Ensembl genes (with corresponding 2749 antibodies) that all have protein presence/absence information from both IHC and IF experiments (in the U2-OS cell line). Figure 1 shows the obtained data for gene NDUFS4 as an example of the input data for the three included platforms, IHC (A), IF (B) and RNA-seq (C).
Protein distribution and overlap with transcriptional data
Using the transcriptome sequencing strategy, we detect transcripts for 85.3% (2123+222) of the genes in the HPA subset (Table 1 ). A large overlap in expression is obvious comparing RNA-seq and the immunological assays. Figure 2 compares presence of proteins and transcripts, and it demonstrates that out of the HPA subset, RNA-seq detects 87.1% (2123/2123+315) of the IHC-detected proteins, and 87.2% (1771/1771+260) of the IF-detected proteins. These numbers are higher than what is expected by chance; a chi-square test results in p-values of 3.4 × 10 -13 and 2.6 × 10 -6 for IHC and IF, respectively. This supports a strong association between RNA and protein expression. The fact that approximately 13% of all detected proteins does not have a detectable transcript can indicate several different phenomena: (i) Some genes are very lowly expressed as transcripts, but efficiently translated into stable protein products or (ii) Some antibodies are cross-reactive, yielding false positive protein detection.
Interestingly, 9.4% (222/2123+222) of the genes in the HPA subset that were detected on the transcript level are not detected on the protein level. For IF, this number is 24.4% (574/1771+574). It is not clear if this is caused by a subset of genes that are transcribed but not translated, or if this is due to a limited sensitivity in the protein measurements. The fact that this fraction is higher for IF than IHC indicates that unspecific antibody-protein interaction in IHC in combination with a group of transcripts that do not undergo translation is the major contributor to this affect, since the sensitivity of IF is generally higher than that of IHC (see below).
Comparison over three technology platforms
For a more in-depth analysis, we investigated the expressed genes in a combined analysis of IHC, IF and RNA sequencing of the HPA subset (2749 genes). We show that 60.1% (1651 genes) of all investigated genes are detected by all platforms (Figure 3A ) and only 1.2% (34 genes) was not detected by any platform. If only one of the two proteins detection platforms is required to call presence on the protein level (in the case of one of them producing a false positive call), only 3.2% - 5.2% of all genes are not detected on the transcript or RNA level. Interestingly, 71% (1651+205)/(110+472+1651+205+55+120) of proteins detected by either IHC or IF were detected by both methods. In total, IHC detects more proteins than IF (2438 vs. 2031). The higher number of detected genes likely indicates a higher degree of false positives, since genes detected by IF and RNA-seq are more lowly expressed than genes detected by IHC and RNA-seq (Figure 3B , see below).
Since RNA-seq provides quantitative measures of gene expression levels, we investigated transcript levels for the genes in the subgroups defined in Figure 3A , where this was possible (Figure 3B ). This showed that the groups 'C' and 'BC' (detected in RNA-seq only and detected in IF and RNA-seq positive, respectively) were expressed at significantly lower levels than all genes combined (Kolmogorov-Smirnov (KS) test, p = 3.8 × 10 -4 and p = 4 × 10 -4 , respectively). This indicates that IF has a higher sensitivity to detect transcriptionally low expressed genes than IHC.
Next, we investigated the overlap between RNA and protein expression using the quantitative RPKM [ 10 ] values as a measure of transcript abundance. This measure is calculated by counting all reads that map to the exons of a gene and dividing by the length of the gene and total number of reads and is an expression value for each gene. The HPA subset was binned in 25 transcriptional levels, ranging from the top 5% of the bottom 5% expressed genes, as well as two transcriptional levels: the upper 50% and lower 50% expressed genes. Table 2 shows a 95% overlap between the upper 50% genes and protein presence based on IHC expression. For the lower 50%, this number drops to 82% and for IF, this overlap is 80% and 68% for the upper and lower intervals, respectively. Interestingly, for the smaller bins, this effect is very similar: The overlap is high (98% for IHC, 86% for IF) for the top 5% and remains relatively similar across the top 50% [Additional file 1 : Supplemental figure S3]. Furthermore, we chose the subset of antibodies that had the highest validation score (supportive staining) in Western blot [Additional file 1 : Supplemental table S1], and as expected, this yielded a slightly higher degree of overlap for both IHC and IF (Table 2 ) suggesting that some of the antibodies with a low validation score might be false positives.
Gene Set Enrichment Analysis
DAVID [ 13 , 14 ] is a tool that performs gene set enrichment analysis for several different categories (GO, KEGG, protein domains etc) and has the option to group similar categories into functional groups based on similarity. When the HPA subset (2749 genes) was analyzed for enrichment of gene categories against a background of all protein coding genes using DAVID some Gene Ontology categories emerged as over-represented. These include development, apoptosis, proteins related to direct protein sequencing, the cytoplasm and protein binding (data not shown).
Thus, since the HPA subset is somewhat biased from a gene category perspective, further gene set analysis of the subgroups was done with the HPA subset as background. We noticed that for genes in the ABC group (detected by all platforms, Figure 3A ), certain themes were enriched. Two category sub-clusters were significantly over-represented; intracellular proteins and nuclear proteins (Table 3 ). From a technical perspective, these proteins are located within the cells (or even within the nucleus) and are therefore equally well detected using either IHC or IF. As a contrast, in the BC group (IHC-, IF+, RNA+) we find that extracellular proteins are significantly enriched (a group of proteins usually not detected by IF). Given the higher sensitivity of IF, we might speculate that these are proteins destined for export that still reside within the cells, and thus are present at very low levels.
In the AB group (IHC+, IF+, RNA-), we notice that proteins related to glycosylation are significantly enriched. The apparent lack of correlation between RNA and protein expression for this GO category is not fully understood and requires further analysis to elucidate (Figure 4A , see below).
Western blot and IF validation score analysis
The Western blots performed within the HPA program are manually investigated and assigned a validation score based on the number of detected bands, approximate size of the bands etc [Additional file 1 : Supplemental table S1] using a standardized protein lysate panel. We analyzed these scores for all groups defined in Figure 3A . We observe that antibodies raised against genes in the A group (positive only in IHC) or the AB group (IHC+, IF+, RNA-) generally contain more low-scoring antibodies (KS-test, p = 1.6 × 10 -4 and KS-bootstrap test, p = 2 × 10 -3 ) (Figure 4A ). This suggests that some of these antibodies are staining the cells in an unspecific manner (false positives) and the RNA data can thus provide guidance for the validation of the corresponding antibodies. Western blot data was not available for our particular U-2 OS cell line analysed in this study, so a direct expression comparison between WB and other methods was not possible.
For IF images in the HPA program, a validation score is added after manual investigation. Comparative analysis of these scores could only be done for groups with staining according to IF, since a validation score of 7 is used to define absence. In the ABC group (present in all platforms), the fraction of antibodies receiving a supportive score [Additional file 1 : Supplemental table S1] is about three times higher than that in the B group (present only in IF) and is confirmed significant (KS-bootstrap test, 2 × 10 -3 ). | Discussion
A quantitative comparison of the transcriptome and the proteome in a cell (or tissue) is instrumental in deciphering fundamental regulatory pathways and mechanisms. The relationship between these two sets of biomolecules also has wide implications for the identification of new biomarkers and classifiers in the treatment of disease. Previous efforts to compare the transcript and protein abundance, at single or multiple gene levels, have demonstrated a great variation in outcome, attributed to both biological and technical issues [ 1 - 8 ]. From a technical perspective, transcriptome monitoring has recently been greatly improved by RNA-seq, which provides a digital and comparably unbiased profile of all genes transcribed in a cell. We show that almost 75% of all genes are represented by one or more transcripts, which is in line with other recent studies [ 10 ]. We also demonstrate that lowly expressed genes are less likely to be identified at the protein level.
The proteome has in most cases been represented by indirect methods such as arrays or by mass spectometry of a solubilized proteome. Here, we present a large-scale comparison of in-situ protein abundance in cells with the transcription levels provided by RNA-seq. The proteins are assessed by antibodies targeting the proteome, and the presence of a particular protein species is visualized by immunohistochemistry and immunofluorescence microscopy, from which an abundance is estimated. These are two semi-quantitative methods and we therefore performed the comparison at the level of presense/absense.
For 87% of all detected proteins, we were able to detect a corresponding transcript. For the remaining 13% of all detected proteins (where we found no transcript), it can indicate several different phenomena: 1) Some genes are very lowly expressed as transcripts, but efficiently translated into stable protein products or 2) Some antibodies are cross-reactive, yielding false positive protein detection. Both of there effected would yield result where proteins are detected but the transcript is not.
We observe approximately 2/3 overlap between the RNA and its corresponding protein species, irrespective of method. Even if one of the two methods for generation of protein expression data is considered false, only 3.2% - 5.2% of all investigated genes are not detected on the protein or transcript level. Overall, immunohistochemistry identifies slightly more proteins than immunofluorescence, and we suggest that this elevated detection is in part due to false IHC positives, as indicated by the fact that IF demonstrates higher sensitivity for the lowly expressed genes (Figure 3B ). Interestingly, the transcriptome-negative but IF- and IHC-positive group (8%), contained an over-representaion of genes related to glycosylation and needs to be confirmed with alternative methods such as mass spectrometry and RT-PCR. The group of genes detected by RNA-seq and IF but not IHC is significantly enriched for secreted proteins. Of 69 genes in this GO category, 33 are present in this group (p = 4.4 × 10 -7 ). | Background
An interesting field of research in genomics and proteomics is to compare the overlap between the transcriptome and the proteome. Recently, the tools to analyse gene and protein expression on a whole-genome scale have been improved, including the availability of the new generation sequencing instruments and high-throughput antibody-based methods to analyze the presence and localization of proteins. In this study, we used massive transcriptome sequencing (RNA-seq) to investigate the transcriptome of a human osteosarcoma cell line and compared the expression levels with in situ protein data obtained in-situ from antibody-based immunohistochemistry (IHC) and immunofluorescence microscopy (IF).
Results
A large-scale analysis based on 2749 genes was performed, corresponding to approximately 13% of the protein coding genes in the human genome. We found the presence of both RNA and proteins to a large fraction of the analyzed genes with 60% of the analyzed human genes detected by all three methods. Only 34 genes (1.2%) were not detected on the transcriptional or protein level with any method. Our data suggest that the majority of the human genes are expressed at detectable transcript or protein levels in this cell line. Since the reliability of antibodies depends on possible cross-reactivity, we compared the RNA and protein data using antibodies with different reliability scores based on various criteria, including Western blot analysis. Gene products detected in all three platforms generally have good antibody validation scores, while those detected only by antibodies, but not by RNA sequencing, generally consist of more low-scoring antibodies.
Conclusion
This suggests that some antibodies are staining the cells in an unspecific manner, and that assessment of transcript presence by RNA-seq can provide guidance for validation of the corresponding antibodies. | Conclustions
Our study demonstrates that RNA-seq is useful as a validation tool for the HPA program. In the case of previously uncharacterized proteins or conflicting data between protein array, Western blot, IHC and IF, we could use both the presence/absence call and the quantitative estimation of RNA molecules to decide on the quality of data. Indeed, presence of RNA does indicate that a corresponding protein can be expected in the analyzed sample. Furthermore, RNA-seq has been shown to efficiently identify splice variants [ 15 ], and this may also be used to discriminate between multiple antibodies directed towards different parts of the target protein.
We used the quantitative estimation of RNA molecules to show that the highest expressed genes are more likely represented by a corresponding protein. For the 50% highest expressed genes we detect the corresponding protein in more than 80% (IF) or 95% (IHC) of the cases. These numbers reveal a signficantly closer relationship between the presence of RNA and protein than what has been seen in previous studies. We believe that our data indicate that both RNA and proteins are expressed to a more significant extent than previously anticipated and that this means that cells are regulated at the level of protein abundance rather than on mere presence/absence. The next step in our understanding of the cell machinery will therefore require more sensitive and quantitative measures of proteins.
Authors' contributions
DK and LF analysed the data. DK performed low-level analysis of the RNA-seq data. LF performed low-level analysis of the IF and IHC data. DK and JL wrote the manuscript. EL and OE interpreted data. MU and JL designed the study. All authors read and approved the final manuscript.
Supplementary Material | Acknowledgements
The authors with to thank Dr. Johan Lindberg and Dr. Magnus Bjursell for valuable insights during data analysis. We would also like to thank the Uppsala Genome Center for SOLiD sequencing. This work was supported by the Knut and Alice Wallenberg Foundation, the Swedish Research Council and the Swedish Cancer Society. | CC BY | no | 2022-01-12 15:21:38 | BMC Genomics. 2010 Dec 2; 11:684 | oa_package/b0/b0/PMC3014981.tar.gz |
|
PMC3014982 | 21126341 | Background
Common bean ( Phaseolus vulgaris L.) is one of the oldest cultivated crops in the Americas and is the most important grain legume for human consumption with production more than double that of the second most important grain legume, chickpea ( Cicer arietinum L.) [ 1 ]. Common bean was domesticated in the Americas by indigenous people during pre-Colombian times. Archeological data suggest that bean was independently domesticated in different regions of the Americas including, the Andean region of South America [ 2 ], Argentina [ 3 ], and Mexico [ 4 , 5 ]. The oldest domesticated beans were found at archeological sites in each of these regions between 4300 and 8000 B.P. [ 6 , 7 ] Changes in bean plant phenotype as a result of domestication include but are not limited to growth habit, seed size, seed retention, and maturity[ 3 , 8 ]. However, the molecular events that underlie these differences in agronomic traits have not been elucidated.
Original classification of common bean germplasm was performed by Singh et al. [ 9 ] into two primary Centers of Domestication (COD); namely Middle American from Central and North America and Andean from South America. Primary CODs were further divided into Races based on geographic origin and genetic lineage. The Andean COD was subdivided into three Races: Nueva Granada (Columbia), Peruvian (Peruvian highlands), and Chilean (northern Chile and Argentina). The commercial market classes that represent Race Nueva Granada in the USA, include light red kidney, dark red kidney, white kidney, and cranberry beans. Andean beans such as Calima, Azufrado, sugar bean, and other mottled types are also widely grown in Africa and the Caribbean. Beans from the Middle American COD were domesticated in West-Central Mexico [ 10 ] and include Durango (central highlands of Mexico), Jalisco (coastal Mexico near the state of Jalisco), and Mesoamerican (lowland tropical Central America) Races. Market classes grown in the US that typify Race Durango include pinto, great northern, small red, and pink bean. Navy, small white, and black beans represent the Mesoamerica landrace. The organization of common bean into COD and Races is depicted in Figure 1 .
Current tools used to characterize genetic diversity within common bean include plant morphology [ 11 ], seed protein allozymes [ 11 - 14 ], random amplified polymorphic DNA[ 15 ], restriction fragment length polymorphisms [ 16 ], DNA sequence analysis [ 17 ], chloroplast DNA [ 18 ], and microsatellite markers [ 19 , 20 ]. While effective for evaluating genetic diversity of parent plants, these techniques are of limited use in breeding programs focused on crop improvement. Methods that can assess potential differences in biological function among cultivars on a genome wide basis have become available in high-throughput and low-cost formats. These 'omics' platforms provide an unprecedented opportunity to ultimately identify the underlying molecular mechanisms that account for traits of agronomic and nutritional importance.
The research presented herein utilized three 'omics' platforms (transcriptomics, proteomics, and metabolomics) to determine whether beans from Middle American and Andean COD differ in their patterns of gene transcription, protein expression, and/or small molecule synthesis and metabolism. To our knowledge, this is the first report that shows differences in gene transcription, protein expression, and metabolite profiles between beans from two COD using 'omics' techniques. Moreover the metabolomic approach was used to evaluate whether common bean cultivars within the same gene pool, referred to as a market class, differ in their profile of metabolites. These proof-in-principle experiments herald significant opportunities for utilizing 'omic' technologies in the rapid identification of traits with agronomic and nutritional importance that could serve to guide common bean breeding programs. | Methods
Bean Material
Navy bean (cv. Seahawk, cv. Norstar, and cv. Vista) and white kidney bean (cv. Beluga, cv. Silver Cloud, and cv. Lassen) cultivars representative of the two market classes were selected for this study to represent the Middle American and Andean COD, respectively. Seeds from each of the market classes used in this study were obtained from field grown plants at the Colorado State University Agricultural Research, Development and Education Center (ARDEC), Fort Collins, CO. The bean seed used for analysis within each omics approach was of the same physiological age but differed across omics platforms. Seeds for transcriptomics analyses were collected from plants at two and three weeks after anthesis (plant flowering), immediately frozen in liquid nitrogen and stored at -80°C. Seeds for proteomics and metabolomics were harvested at full maturity and air dried and stored at room temperature (RT, 22 ± 2°C) until use.
RNA isolation
RNA was extracted from navy and white kidney bean seeds from field grown plants in 2007. Seeds were collected in the field and immediately frozen as described above. Two cultivars, Silver Cloud (white kidney) and Norstar (navy) were selected for spotted array gene expression analysis. After storage, frozen seeds were separated from pods and RNA was isolated according to the TRIzol reagent protocol (Life Technologies, Gaithersburg, MD). The RNA sample was purified with the RNeasy Mini Kit (Qiagen, Valencia, CA) according to Affymetrix (Santa Clara, CA) instructions. RNA concentrations and purity were determined using an ND-1000 NanoDrop spectrophotometer (Thermo Scientific NanoDrop, Wilmington, DE). RNA integrity was evaluated by the Experion Bioanalyzer Automated Electrophoresis System (Bio-Rad Laboratories, Hercules, CA).
Microarray hybridization
Since a commercial Phaseolus vulgaris microarray was not available, a soybean ( Glycine max L. Merr.) array was selected because it was the closest phylogenetic relative among available arrays [ 29 ]. Although Glycine max and Phaseolus vulgaris differ in chromosome number and genome size (the soybean genome is twice as large as common dry bean), linkage mapping of DNA markers found an average conserved block length of 13.9 cM between the two genomes indicating high conversation and preservation [ 29 , 30 ]. Twenty-five μg of RNA was used to prepare complimentary DNA for spotted 2-color microarray analysis. Complimentary DNA (cDNA) was synthesized using the Superscript III kit (Invitrogen, Carlsbad, CA) and hybridized to a microarray developed for soybean [ 23 ] using the Genisphere Array 50 kit (Genisphere, Hatfield, PA). Briefly, RNA (29.5 μL) was mixed with dye-appropriate RT primer (1.5 μL) and heated to 80°C for 10 min, and transferred to ice for 2 min. Superase-in RNase inhibitor (1 μL) was added before 18 μL of reaction mix composed of 5× Superscript III first strand buffer (10 μL), 10 mM dNTP mix (2 μL), 0.1 M dithiotreitol (4 μL) and Superscript III enzyme (4 μL). After incubation at RT for 5 min, tubes were placed in an iQ iCycler (Bio-Rad) for 2 h at 50°C. The reaction was stopped with 0.5 M NaOH/50 mM EDTA (7 μL) followed by incubation at 65°C for 10 min and neutralized with 1 M Tris-HCl (10 μL). cDNA from the navy and white kidney bean samples were combined and purified using Geneclean Turbo kit (Qbiogene, Carlsbad, CA). The purified cDNA was stored at -20°C until use. Arrays were cross-linked by exposure to 65 mJ of UV irradiation and blocked by incubation in 250 mL prehybridization solution [5× SSC (18% Sodium Chloride, 9% Sodium Citrate) buffer, 0.1% SDS and 0.01% BSA solution] at 42°C for 1-2 h. The arrays were dried by centrifugation at RT for 5 min and placed into hybridization chambers and incubated at 42°C for 15 min. Formamide-based buffer (2×, 50 μL) and LNA dT blocker (2 μL) were added to the cDNA sample and heated for 10 min at 80°C. A cover slip was placed on the array and the cDNA mix was introduced between the cover slip and the array surface. The solution was distributed evenly on the array by capillary action. Water (15 μL) was added to the wells of the array chambers followed by incubation for 16 h at 42°C. The array was vigorously washed 3 times for 2 min, first with 2× SSC and 0.2% SDS (250 mL) followed by 2× SSC (250 mL) and then with 0.2× SSC (250 mL). The arrays were centrifuged for 3 min (1000 × g) to dry and warmed at 42°C for 15 min. A mix of 2× formamide-based buffer (42 μL), water (35 μL) and 3.5 μL each of Cy3 and Cy5 was prepared in the dark and incubated at 80°C for 10 min. The hybridization mix was then injected onto the soybean cDNA microarray. Arrays were incubated at 42°C for 3 h. After the second hybridization step, the arrays were washed and dried as described above and immediately scanned with a Genepix 4000B scanner (Molecular Devices, Sunnyvale, CA).
Scanned images were analyzed using GenePix Pro 6.0 software (Molecular Devices) where spot features were aligned for annotation. Poorly hybridized spots were discarded from analysis using the software parameters and the intensities of each dye were quantified. Fluorescence intensities were normalized and exported for statistical analyses.
Protein isolation and two-dimensional gel electrophoresis (2DGE)
2DGE was performed with bean extracts according to previously published methods [ 31 ]. Approximately 15 mg of bean seed samples from each of the six cultivars listed above were suspended in sample lysis buffer, containing 7 M urea, 2 M thiourea, 40 mM DTT, 2% CHAPS, 1% Pharmalyte pH 3-10 (GE Healthcare, Piscataway, NJ), and trace amount of bromophenol blue (BPB). Insoluble debris was removed after centrifugation for 30 min (1000 × g, 15°C). The supernatant was recovered and used for analysis. Protein concentration was quantified using the Bradford assay.
Bean seed protein (200 μg) was applied to reswelled, immobilized pH gradient (IPG) strips (pH 4-7L, 24 cm; GE Healthcare). The isoelectrofocusing conditions were as follows: 10V to 300 V for 3 h and at 5000 V for a total of 95 kVh at 20°C. After isoelectrofocusing, the strips were equilibrated in a buffer containing 30% glycerol, 6 M urea, 2% SDS, 10 mg/mL dithiothreitol (DTT) and 0.05 M Tris-HCl (pH 6.8) for 15 min and then for an additional 15 min in equilibration buffer in which 42.5 mg/mL iodoacetamide replaced the DTT. The strips were positioned at the top of 13-16% gradient polyacrylamide gels with 0.5% agarose containing Laemmli sample buffer [ 31 ]. SDS-PAGE was performed in Laemmli electrophoresis buffer at 150 V at RT. Proteins were stained with alkaline ammoniacal silver staining [ 32 ] scanned with a GS-800 Calibrated Densitometer (Bio-Rad) and analyzed using PDQuest v7.1.1 software (Bio-Rad).
Analysis of gels
Samples were analyzed using 2DGE to obtain quantitative protein profiles within the molecular weight range of 25 to 150 kDa. Each bean cultivar was run in duplicate. Protein spots were automatically detected using PDQuest v7.1.1 software. All spots were also manually confirmed. Images of duplicate gels were superimposed and a master gel generated for each of the six cultivars. Proteins were quantified using spot densitometry.
Comparisons of the 2DGE protein patterns generated an inclusion list for only those proteins that differed significantly between the two COD (greater than 2-fold change). Protein patterns were also compared between cultivars within each COD. Although out of the scope of this work, these spots can be used in the future for protein identification using MALDI TOF MS.
Metabolite extraction and analysis
It is imperative that the majority of the proteins within the dry bean samples are removed prior to metabolomic fingerprinting in order to avoid confounded results from possible protein fragments. The dried bean seeds were boiled to denature proteins and freeze dried for storage. The proteins were then precipitated during the metabolite extraction process using cold ethanol. The boiling procedure is as follows: 0.5 kg of bean was soaked in distilled water for 3 h at RT. After 3 h the beans were drained, rinsed thoroughly with deionized water, and blanched for 5 min at 93°C. Beans were boiled for 60 min in a 1.5% KCl solution using a pressure cooker. Finally, beans were freeze dried (Genesis SQ25LL, Virtis Company, Gardiner, NY) to powder form and stored at -80°C until use.
Metabolites were extracted using cold ethanol (65%, -20°C). Approximately 2.5 g of freeze dried bean powder was added to ethanol (65%, 50 mL) and vortexed to ensure complete mixing. The mixture was sonicated at RT for 2 h and centrifuged (1000 × g, 10 min) to separate the insoluble material from the ethanol extract. The extract was decanted into a clean conical tube and stored at -20°C until analysis up to 1 mo.
Ultra Performance Liquid chromatography-Mass Spectrometry (UPLC-MS)
Sample separation was performed using an Acquity UPLC under the control of MassLynx software (Waters, Millford, MA, USA). The sample set was randomized and held in an 8°C sample manager during the analysis. For each chromatographic run, a 1 μL sample injection was loaded to a 1.0 × 100 mm Waters Acquity UPLC BEH C18 column with 1.7 μm particle size held at 40°C. Separation was performed by reverse phase chromatography at a flow rate of 0.15 ml/min. The eluent consisted of water and methanol (Fisher, Optima ® LC/MS grade) supplemented with formic acid (Fluka, LC/MS grade) in the following proportions: Solvent A = 95:5 water:methanol + 0.1% formic acid; Solvent B = 5:95 water:methanol + 0.1% formic acid. The separation method is described as follows: 0.1 min hold at 100% A, 14.9 min linear gradient to 100% B, 5 min hold at 100% B, 1 min linear gradient to 100% A, and 1 min hold at 100% A. A blank injection of water and 15 min chromatographic run was preformed between samples to eliminate possible carryover of analytes and to re-equilibrate the column. This cleaning method is described as follows: 0.1 min hold at 100% A, 2.9 min linear gradient to 100% B, 1 min hold at 100% B, 3 min linear gradient to 100% A, and 8 min hold at 100% A for equilibration. The flow rate for all steps was held at 0.15 ml/min.
Eluate was directed to a Q-TOF Micro quadrupole orthogonal acceleration time-of-flight mass spectrometer controlled with MassLynx software (Waters/MicroMass, Millford, MA, USA) using electrospray ionization in the positive mode (ESI+). Mass data were collected between 50 and 1000 m/z at a rate of two scans per second with a 0.1 second interscan delay. The voltage and temperature parameters were tuned for general profiling as follows: capillary = 3000 V; sample cone = 30 V; extraction cone = 2.0 V; desolvation temperature = 300°C; and source temperature = 130°C. Mass spectral peaks were centered during acquisition producing centroid data. Leucine Enkephalin was infused via an orthogonal ESI probe and baffle system (LockMass) which allowed reference ions to be detected for a single half-second scan every 10 s in an independent data collection channel. The standard mass was averaged across 10 scans providing a continuous reference for mass correction of analyte data.
Chromatographic and spectral LC-MS peaks were detected, extracted, and aligned using MarkerLynx software (Waters, Millford, MA, USA). Chromatographic peaks were detected between 0 and 18 min with a retention time error window of 0.1 min. Apex track peak detection parameters were used, automatically detecting peak width and baseline noise. No smoothing was applied. To reduce the detection and inclusion of noise as data, an intensity threshold value of 40 and a noise elimination value of 6 were used. Mass spectral peaks were detected between 50 and 1000 m/z with a mass error window of 0.07 m/z. The de-isotoping function was enabled to eliminate the inclusion of isotopic peaks. A matrix of features as defined by retention time and mass was generated, and the relative intensity of all features, as determined by area, was calculated for all individual samples. Potential effects of technical variability were minimized by normalizing the intensity values to the total ion current (TIC) such that the summation of all feature intensities in each individual sample were equal.
Data analysis
For microarray analysis, data were imported into Partek Discovery Suite software (Partek, St. Louis, MO), PC and 1-way ANOVA (random effects) (p < 0.01) analyses were performed. For 2DGE, PC, hierarchal clustering based on Euclidian distances, and 1-way ANOVA (random effects) (p < 0.05) analyses were carried out using Partek Discovery Suite software. Finally, the LC-MS feature matrix was mean centered and imported into SIMCA-P+ software (Umetrics, Inc., Umeå, Sweeden). PC analysis was performed using Pareto scaling. Hierarchal clustering based on Euclidian distance and 1-way ANOVA (random effects) (p < 0.05) analyses were performed using Partek Discovery Suite software. Fold change for both proteomics and metabolomics was calculated using Partek Discovery Suite. | Results
Three 'omics' platforms to distinguish common bean COD
Experiments were conducted to determine whether or not the three 'omics' technology platforms (transcriptomics, proteomics, and metabolomics) would be capable of distinguishing dry bean cultivars from each of the two COD. The approach used was sequential; transcriptomics experiments were performed first, followed by proteomics, and then metabolomics, with increasing complexity of the experimental design as work progressed to each new platform. The white kidney bean and navy bean market classes representing the Andean and Middle American gene pools were selected as representative of the respective COD. Furthermore, both market classes are commercially important, and had a white pigmented seed coat thus limiting the likelihood of identifying qualitative differences between COD that were due solely to synthesis of different pigments.
Transcript profiling
Developing seeds from the Middle American COD navy bean cultivar Norstar and the Andean COD white kidney bean cultivar Silver Cloud were evaluated for differential expression of transcripts. Initially an Affymetrix soybean microarray was used for this purpose. However, using the unmodified array, we found that the high binding specificity of Affymetrix microarrays, resulted in a hybridization efficiency of P. vulgaris cDNA onto the soybean microarray (< 10%) that was too low for expression analysis. Although it has been recently reported that the use of masking biased probes overcomes this limitation [ 21 ], a spotted cDNA microarray for soybean was selected as an alternative approach because of their successful use in cross-species studies [ 22 ]. This cDNA soybean microarray, available through the Keck Center for Comparative and Functional Genomics at the University of Illinois, has been previously described [ 23 ]. Principal component (PC) analysis showed that navy and white kidney gene expression profiles differed when seeds were collected 2 wk after anthesis (Figure 2 ). The clones represented on the microarray were previously annotated as described by Vodkin et al. [ 23 ] and supplied with the microarray. Based upon that annotation, 330 clones were identified as differentially expressed between white kidney and navy at 2 wk after anthesis. The clones were selected by one-way ANOVA (p < 0.01) and a fold change cutoff in expression of 2.0. The majority of these clones (309) were up-regulated in white kidney bean and the remaining 21 were up-regulated in navy bean. Of these 330 differentially expressed clones, 30 were up-regulated in white kidney by at least 5-fold change and only 1 clone was up-regulated in navy by at least 5-fold change. The clones differentially expressed by 5-fold change were annotated and displayed in Table 1 . The majority of clones up-regulated by 5-fold change in white kidney have a role in transcription, translation, and protein synthesis/modification based on the annotations provided with the microarray. Examples of up-regulated white kidney clones include: ribosomal proteins and ubiquitin. The only clone up-regulated in navy has homology to a ripening induced protein in 3 wk old seeds, there were 13 clones that were differentially expressed between the two beans by greater than 2-fold change (Table 2 ). Three clones were up-regulated in white kidney, 1 of which was annotated as nucleoside diphosphate kinase. Ten clones were up-regulated in navy bean, but only 4 were identified in the annotation and are involved in metabolism. The observation that fewer genes were expressed at 3 wk old compared to 2 wk old likely reflects the reduced metabolism in the developing seed as it approached physiological maturity.
Proteomic profiling
Based on the transcript profiling data reported in the previous section, it was predicted that the profile of expressed protein would differ between cultivars representative of COD. Protein expression was assessed using a two-dimensional gel system instead of an LC-MS platform, so that a visual picture of similarities and differences in types and amounts of proteins expressed between COD could be obtained. Moreover, rather than limiting the analysis to one cultivar from each market class-COD category, three cultivars from each genepool were evaluated. The list of cultivars studied can be found in the Materials and Methods section. Master gels of the two dry bean germplasm (white kidney and navy) were created for purposes of exploratory analysis. Close visual inspection showed distinct differences in both the absolute number of proteins as well as differences in concentration (spot density) between COD. With the false discovery rate set at zero, 2186 spots were selected from the master gel using the PDQuest software.
Data from the three cultivars evaluated were combined and subjected to PC analysis (Figure 3a ). A heat map of spot densities showed visual differences in up- and down-regulated proteins when comparing white kidney bean and navy bean germplasm (Figure 3b ). One-way ANOVA analysis showed that 733 proteins were statistically different between the two COD (p < 0.05) with a fold change greater than 2. Of the 773 proteins, 282 proteins were up-regulated in white kidney bean and 254 proteins were up-regulated in navy bean.
Metabolomic fingerprinting
While the results of proteomic analyses were consistent with genetic differences between cultivars in each COD, these findings do not address the nature of these differences. It can be argued that the majority of the proteins detected participate in the synthesis, transformation, and degradation of both primary and secondary metabolites that play a role in plant architecture, reproduction, and defense against biotic and abiotic stresses [ 24 ]. The focus of the metabolomic analyses reported herein was on the small molecules having a mass less than 1000 Daltons. Our objective was to determine if these small molecule profiles differed between Andean and Middle American beans. The same market classes and cultivars evaluated in the proteomics experiments were metabolically fingerprinted. To perform unbiased metabolomic analyses, m/z values from 50-1000 were monitored using UPLC-ESI-MS revealing 6732 possible small molecules in positive mode. A heat map of differences in intensity of the small molecules between the cultivars from each COD is shown in Figure 4 where each column in Figure 4 corresponds to a specific cultivar. One-way ANOVA results showed that 472 small molecules corresponding to 7% of the total number of features detected were a minimum of 2 fold higher in Andean beans (p < 0.05); 487 small molecules, also approximately 7% of the total number of features detected, were 2 fold higher in Middle American beans (p < 0.05).
Cultivated varieties distinguished by metabolomic fingerprinting
Metabolomic fingerprinting was also utilized to assess differences in small molecule profiles among bean cultivars within market classes. Clear metabolite clustering of the cultivars within the two germplasms was observed using PC analysis and hierarchal clustering. As shown in Figure 5a , the bean cultivars were separated along PC 2 and PC 3, while PC 1 separated by COD. PC 1 explained 19.1% of the variation while PC2 and PC3 explained an additional 9.95% and 6.2% of the total variation among cultivars, respectively. Hierarchal clustering using Euclidian distances was used to distinguish among cultivars and between COD. The resulting dendrogram shows clustering of cultivars within COD (Figure 5b ) which mimics clustering seen in the proteomics analysis. One-way ANOVA analysis of the 6 cultivated varieties grouped by market class showed 542 features were statistically different (p < 0.05) with a fold change of at least 2 between the navy bean and white kidney market classes. The 542 features were further analyzed to determine which could differentiate cultivars within the market classes. In this case, 167 features were found to be statistically different (p < 0.05) within the white kidney bean cultivars; whereas, 246 features were found to be statistically different (p < 0.05) between the three navy bean cultivars. | Discussion
The elucidation of molecular differences underlying traits that permit the division of common bean germplasm into distinct centers of domestication has been largely unexplored by high-throughput 'omic' technologies. A triad of 'omic' technologies: transcriptomics, proteomics, and metabolomics, were used herein to determine whether differences in gene transcription, protein expression, or synthesis/metabolism of small molecules were of sufficient magnitude to distinguish among bean cultivars within and between two COD (Figure 1 ). As shown in Figure 2b , 3c , and 5a , distinct differences were observed among cultivars and between the two COD in all three 'omics' signatures.
The transcript profiling experiments revealed that chronological age of the seed from anthesis has a significant impact on the number of differentially expressed genes and the functions (gene ontology) associated with those genes. However, examination of the gene lists in Table 1 failed to provide a clear picture of the physiological distinctions between COD. One limitation to consider is that distinct P. vulgaris genes may exist without a soybean homologue to utilize during hybridization and thus a cross-species microarray may exclude unique dry bean transcripts. Reciprocally, the chance of false positives was increased in the cross-species use of the cDNA spotted arrays because of reduced hybridization stringency. The drawbacks associated with using a soybean array for common bean transcriptomics stem from reduced synteny between P. vulgaris and G. max . The most recent common ancestor for P. vulgaris and G. max was 19 Ma. At some time point following divergence, diploidization of the G. max genome occurred making its chromosome number 2n = 80 compared to 2n = 40 of P. vulgaris [ 25 ]. Genomic rearrangement including translocation or gene loss during diploidization of the soybean leads to reduced synteny between the two species and must be considered when attempting to determine chromosome overlap between the soybean and the common bean for genetic mapping. The use of the Affymetrix array platform as reported in Yang et al. [ 21 ] may serve to overcome some of the noted limitations, and it is worthwhile to note that the Whole-Genome Sequence of Common Bean project and the Ibero-American whole genome sequencing project include both COD [ 26 ]. With the entire common bean genome sequenced, a complete dry bean microarray can be synthesized and annotated to advance the transcriptomics platform Challenges in gene list interpretation also have been encountered using other organisms and have led to the development of alternative approaches to data analysis methods such as Gene Set Enrichment Analysis (GSEA) [ 27 ]. GSEA permits the investigator to use existing knowledge to test specific hypotheses about differences in pathways and networks of gene expression between two or more treatment conditions. Efforts to incorporate GSEA for market class stratification were limited by cross-species hybridization and the lack of dry bean and soybean gene lists. Consequently, no useful data was gained. This situation underscores the importance of current efforts to sequence the Phaseolus genome. When that work is completed, it will be possible to generate hypotheses for additional testing based on observations about plant morphology [ 11 ] or seed protein allozymes [ 11 - 13 ] that have been used to characterize genetic diversity within common bean. In addition, as work on metabolomic analyses continues and new compound libraries are constructed, metabolomic data such as that shown in Figure 4 and 5 will identify specific small molecules that distinguish cultivars within each COD and within a market class-COD grouping. That information will permit future hypothesis generation about the biosynthetic pathways that are induced or repressed to account for differences in metabolite profiles. Since differences in small molecule biosynthesis are likely to result not only from differences in gene transcription but also arise from differences in the translation of transcripts and post translational modifications of proteins, the potential contributions of proteomic analyses such as those shown in Figure 3 are clear. For future analyses, we recommend that an HPLC or UPLC separation approach be employed rather than 2DGE to interrogate the Phaseolus proteome because of the speed, throughput, and relative cost of the LC-based platform.
As one considers quantitative differences among the three 'omics' signatures that distinguished between the Andean and Middle American COD, it is readily apparent that marked differences exist in the number of features that distinguish between COD when tallying results across platforms. However we caution the reader that the purpose of this experimental approach was primarily the qualitative evaluation of distinguishing differences in transcript, protein, and/or metabolite expression between COD. While the next generation of experiments would be well served to use seeds matched for physiological age across all omics platforms, these initial experiments were performed using available mature seed with the exception of the initial work on transcript expression which required embryos to be harvested in the field and immediately frozen in liquid nitrogen.
Many factors other than the biology of the system could account, at least in part, for the numeric differences observed. For example, the use of the soy-array to identify differentially expressed transcripts or the use of only one of an almost endless number of assay conditions that have been developed for proteomic or metabolomic analyses are likely to make the numeric differences noted unreliable. Nonetheless, the magnitude of the observed differences indicates the potential merit of these approaches as a fertile source of information for plant breeders.
Overall, the 'omics' chosen for a particular set of experiments should depend on the question being asked. While considerable information can be gleaned from each 'platform, the biological activity assessed is interactive, and accordingly, the greatest insights are likely to emerge through the integration of complimentary data sets. Transcriptomics, proteomics, and metabolomics are just three of these [ 28 ]. | Conclusions
We have shown that two genetically diverse dry bean germplasm can be easily differentiated using a suite of three 'tools (transcriptomics, proteomics, and metabolomics). Using this panel of techniques we have provided a glimpse into a signature which can be used to determine COD. Furthermore, known genetic distances among cultivars and between COD were validated using proteomics and metabolomics. 'Omic' signatures unique to P. vulgaris germplasm may be useful to assess complex traits or reactions to biotic and abiotic stress and to incorporate genetic diversity in breeding efforts. This work compliments other techniques currently employed by breeders to assess dry bean genetic diversity. As signatures are further developed in future work, they have the potential to guide those engaged in crop improvement for selecting traits of agronomic and nutritional importance. | Background
Genetic diversity among wild accessions and cultivars of common bean ( Phaseolus vulgaris L.) has been characterized using plant morphology, seed protein allozymes, random amplified polymorphic DNA, restriction fragment length polymorphisms, DNA sequence analysis, chloroplast DNA, and microsatellite markers. Yet, little is known about whether these traits, which distinguish among genetically distinct types of common bean, can be evaluated using omics technologies.
Results
Three 'omics' approaches: transcriptomics, proteomics, and metabolomics were used to qualitatively evaluate the diversity of common bean from two Centers of Domestication (COD). All three approaches were able to classify common bean according to their COD using unsupervised analyses; these findings are consistent with the hypothesis that differences exist in gene transcription, protein expression, and synthesis and metabolism of small molecules among common bean cultivars representative of different COD. Metabolomic analyses of multiple cultivars within two common bean gene pools revealed cultivar differences in small molecules that were of sufficient magnitude to allow identification of unique cultivar fingerprints.
Conclusions
Given the high-throughput and low cost of each of these 'omics' platforms, significant opportunities exist for their use in the rapid identification of traits of agronomic and nutritional importance as well as to characterize genetic diversity. | Authors' contributions
MM carried out the metabolomics studies, statistical analysis associated with metabolomics analysis, analysis of proteomics data, and drafted the manuscript. VF carried out the transcriptomics studies, the statistics associated with the transcriptomics data and helped draft the manuscript. ML carried out the LC-MS work presented here, assisted MM with metabolomics analysis and helped edit and draft the manuscript. ER participated in the study design and helped draft the manuscript. HT participated in the study design, assisted with the statistical analysis and helped to draft the manuscript. MB participated in the study design and helped to draft the manuscript. All authors read and approved the final manuscript. | Acknowledgements
We thank Anne Hess for guidance regarding transcriptomics data transformation and analysis, Sophie Hermann for development of the microarray experiment protocols, SungGu Lee for his technical assistance in proteomic analyses, and Erica Daniell for harvesting the beans from ARDEC fields and RNA extraction for transcriptomics analysis. The work presented here is funded in part by the Colorado Agricultural Experiment Station, Fort Collins, Colorado and by AICR (grant #08A032). | CC BY | no | 2022-01-12 15:21:38 | BMC Genomics. 2010 Dec 2; 11:686 | oa_package/48/27/PMC3014982.tar.gz |
PMC3014983 | 21138572 | Background
Comparative studies are invaluable for understanding the evolution of complex traits. The reptile lineage has given rise to both a metabolically endothermic group (birds) and diverse ectothermic groups [ 1 ], thus providing a comprehensive system for comparative studies of metabolism, physiology, development and aging [ 2 ]. Ectothermic reptiles (e.g. turtles, crocodilians, tuatara, lizards, and snakes) exhibit extreme plasticity in their ability to modulate their metabolism in response to external stresses such as thermal and food stress [ 2 - 5 ]. Furthermore, they show extraordinary diversity in body structure (e.g. turtle shells, squamate limblessness), sex determining systems (e.g. temperature-dependent vs. genotypic), and sexual dimorphism. Snakes, in particular, have evolved dramatic evolutionary adaptations (e.g. venom, limblessness) and have been useful models for evolution and ecology [ 6 ], but the pursuit to understand these evolutionarily important traits at the molecular level has been limited by the molecular resources available. The heightened interest to utilize reptiles in molecular genetic studies is signified by the first two on-going ectothermic reptile genome projects (Anolis lizard and painted turtle) conducted by NIH [ 7 , 8 ], and the additional five ectothermic reptile genomes selected to be sequenced by BGI [ 9 ]. As well, the recent publications of viper venom gland transcriptomes [ 10 ], and the first ectothermic reptile linkage map (crocodile) [ 11 ] emphasize that developing genomic resources for this interesting group is of utmost importance. First glimpses into the evolution of reptile genomes (endothermic and ectothermic) have revealed unique genomic attributes such as microchromosomes, the evolution of gene structure and gene synteny [ 12 - 14 ], as well as dramatic evolutionary changes in functionally important genes [ 15 ].
Ultimately we are interested in understanding how reptiles use their genomes in sex-specific ways to respond to environmental and evolutionary pressures, and how these responses affect reproduction and aging. A necessary first step is to develop and characterize molecular resources for a species of interest. The main focus of this paper is to develop a garter snake transcriptome so it can be used as a reference for future studies. The western terrestrial garter snake ( Thamnophis elegans ) is an emerging model for understanding the molecular basis for the evolution of life-history trade-offs between cellular maintenance and longevity versus growth and reproduction [ 16 ]. Closely related populations of this species, found in the Sierra Nevada Mountains, harbor two ecotypes that contrast in physical appearance, physiological and behavioural stress response, natural lifespan, and reproductive traits. These two ecotypes can be contrasted as either "fast-living" or "slow-living" based on the overall life-history [ 17 - 20 ]. These ecotypes derive from both genetic and environmental differences [ 17 ], and selection acts strongly on divergent traits despite low levels of gene flow among these populations [ 21 , 22 ]. Additionally, males and females have different costs associated with reproduction, which is expected to cause sexual conflict at the genomic level. This conflict may be resolved through differential regulation of how these sexually antagonistic genes are utilized in each sex [ 23 ].
The paucity of molecular resources available for the garter snake and other ectothermic reptile species has limited our ability to identify the genetic basis for evolutionarily important traits, thus providing the inspiration behind this project. The application of new sequencing technologies, such as pyrosequencing, to traditional ecological models expands the horizon for ecological genomic studies [ 24 - 29 ]. Indeed, we are on the verge of elucidating the genetic basis of ecologically and evolutionarily relevant traits in natural populations. Here we apply this technology to develop a large scale, multi-tissue, multi-individual transcriptome using massively parallel sequencing with two goals in mind. Our first goal is to develop a molecular resource for the garter snake and make it available to the scientific community. This resource is a generalized transcriptome (i.e., RNA was pooled across ecotypes, populations, individuals, and tissues) for use as a reference for future studies. Our second goal is to identify sex-specific differences in the presence/absence of expressed transcripts by identifying transcripts that were present in the normalized library of one of the sexes but not the other. These data are accessible through the Garter Snake Transcriptome Browser at Indiana University CGB https://lims.cgb.indiana.edu/cgi-bin/gbrowse/telegans_bronikowski_2/ , the Bronikowski Lab Data Server http://eco.bcb.iastate.edu/ , and through NCBI Short Read Archive ( SRA010134 ). | Methods
Sampling
A total of 35 western terrestrial garter snake ( Thamnophis elegans ) individuals of varying sizes/ages (at least 1 year old) were included in this transcriptome (17 females and 18 males: see Additional file 1 for details on sampling). The snakes were either laboratory-born and raised or field-caught from seven populations at the northern end of the Sierra Nevada Mountains in Lassen County, California, which included the two life-history ecotypes as has been previously described [ 2 , 18 , 55 ]. Adult field-caught snakes were shipped live to Iowa State University. The laboratory snakes were 2-year old animals that had been used in a thermal experiment. Within 10 minutes of euthanizing, all organs and blood were either snap-frozen in liquid nitrogen and stored at -80°C, or put in RNA-later and kept at room temperature for 24 hours and then stored at -20°C. Organs included brain, gonads (from sexually mature individuals), heart, kidney, liver, spleen and blood (ISU IACUC 3-2-5125-J).
RNA isolation
Individual snake samples were pooled by tissue type and by sex for total RNA isolation using TriReagent and cleaned-up using Qiagen RNAeasy columns. The TriReagent manufacture's protocol was followed, but instead of precipitating the RNA, the supernatant from the chloroform step was added (0.75:1) to the RLT/BME buffer from the Qiagen RNAeasy kit. From this point on the Qiagen protocol was followed. Quality of the RNA was verified with a Bioanalyzer nanochip (Agilent). Total RNA quantity was determined by both the Bioanalyzer nanochip and a Nanodrop. RNA from each tissue type was pooled in equal amounts into their respective Male and Female samples, and concentrations determined by fluorometry using Quant-iT OliGreen (Invitrogen).
Library preparation and sequencing
Library preparations for GS FLX Titanium (Roche/454 Life Sciences) sequencing were developed in the Center for Genomics and Bioinformatics, Indiana University based partially on methods for use in GS FLX standard sequencing described in Meyer et al . [ 25 ], with modifications (K. Mockaitis, unpublished). Briefly, cDNA was synthesized from 630 ng of each total RNA pool (male, female) in a manner similar to ClontechTMSMART protocols, using primers optimized for the 454 sequencing process, and amplified by PCR to generate dsDNA. For partial normalization to reduce sequence coverage of high copy number transcripts, amplified cDNA was subjected to controlled in-solution hybridization and double-stranded nuclease (DSN) digestion using the Trimmer Direct kit (Evrogen) after reaction optimization. Normalized DNA was then fragmented by sonication, and ends enzymatically blunted and ligated to customized 454 adaptors. Amplification of ligation products exploited adaptor-mediated PCR suppression [ 25 ]. This method induces homo-adapted fragment hairpins, thereby severely limiting amplification of mis-ligated products. All amplification steps utilized high-fidelity polymerases. Final libraries were size selected by excision of the 500-800 bp fraction from agarose gels.
Emulsion PCR reactions were performed according to the manufacturer (Roche/454 Life Sciences). To optimize pyrosequencing throughput, prior to sequencing final libraries were titrated by emulsion PCR bead enrichment. Sequencing of male and female libraries was performed in parallel on a picotitre plate according to the manufacturer, and yielded 259 Mb (male) and 219 Mb (female) of sequence data in 1, 291869 reads with an average of 386 nts in length.
Sequencing adapters (A and B) were automatically removed from the reads using the signal processing software (Roche/454 Life Sciences). The reads were further cleaned and the adaptors removed by a program developed in-house at CGB, Indiana University http://sourceforge.net/projects/estclean/ . After cleaning, sequences ≤30 bp were removed from the dataset. Thus the final cleaned dataset before assembly contained 1, 238, 280 reads with an average length of 366 nts (Table 1 ).
Assembly and annotation
The pooled reads were mapped to the lizard and chicken genomes using the GS mapper v2.3 with 80% identity. For de novo assembly, the reads were assembled into contigs using the GS de novo assembler (NEWBLER v2.0.00.22, Roche/454 Life Sciences) with the default parameters (40 bp overlap; 90% identity), resulting in 82, 134 contigs and 134, 971 singletons (Table 1 ). We found that some singletons could be aligned to the contigs with 95% percent identity through the entire region except less than 10 bp from both 5' and 3' ends. We used Blast to map 5407 singletons to 2471 NEWBLER contigs. An additional attempt was made to assemble the remaining singletons using MIRA [ 32 ]. The recommended parameters for 454 reads were used, resulting in additional 14, 245 contigs and 93, 000 singletons remaining. Of these singletons, 339 reads were mapped to 380 contigs. Therefore, the final number of singletons is 92, 561 (Table 1 ).
For gene identification, contigs were compared to NCBI-NR protein database, HomoloGene, UniGene (Chicken) [ 56 ] databases using BlastX and tBlastX with cut-off e-values of 1e -5 , 1e -10 , 1e -20 and 1e -50 . Databases accessed in January 2010 were used for these analyses. Additionally the contigs were mapped to the draft Anolis lizard genome (AnoCar1.0) [ 8 ] and Ensemble annotated gene models using tBLATx and BlastX. Open reading frames (ORFs) were predicted using OrfPredictor [ 57 ] with the NR BLAST hits and NCBI's ORF Finder. Databases accessed in August 2010 were used for these analyses.
The NCBI-NR BlastX hits (e-value = 1e -5 ) for the NEWBLER contigs, MIRA contigs, and singletons were used with the program MEGAN [ 38 ] to map the sequences to the NCBI taxonomy (databases accessed on January 2010 for BlastX and February 2010 for taxonomy). Sequences that were assigned to the very tips of the branches outside of Chordata were considered to have originated from off-target species (i.e., not from T. elegans ). MEGAN was used to evaluate the GO annotation (GO Slim) assigned to each term (as of February 2010).
Additional clustering
We used two methods for additional clustering: homology clustering, and a newly developed method we refer to as graph-clustering or contig-graphs (Figure 1 ). Homology clustering is the grouping of singletons and contigs based on their BlastX hits in HomoloGene, and the draft lizard gene models (AnoCar1.0). The homology clustering was based on four different cut-off e-values: 1e -5 , 1e -10 , 1e -20 and 1e -50
The graph-clusters were assembled independent of any comparisons to other databases, but rather solely dependent on the information derived from the original reads. The GS de novo (NEWBLER, Roche/454 Life Sciences) assembly program is set up to allow for improved contig alignments when there is abundant alternatively spliced transcripts and gene duplication events by allowing reads to be split into two and each portion assigned to a different contig. We have developed a clustering method based on graph theory that uses this 'historical' information on how the reads were split and assigned into contigs (see Additional file 3 for full description). The underlying algorithm clusters contigs into network graphs where the contigs represent the nodes and the split reads are the edges. These graph-clusters can contain components that represent 1) a single gene with divergent alleles, 2) a single gene with alternatively spliced transcripts, 3) closely related genes within a gene family (gene duplications), or 4) any combination of the three (Additional file 3 ). In the case of alternatively-spliced transcripts, the nodes indicate exons, and the edges indicate the combinations of how these exons connect in the different transcripts (transcriptional paths). These exons could not be merged further based on similarity to each other. In the case of duplicated genes, parts of the transcripts were highly similar and merged into the same contig, but reads covering the regions of the duplicated genes that have diverged were split into different contigs (nodes). These nodes representing the diverged regions of the duplicated genes were still quite similar (assuming >80% but < 95% sequence identity). This similarity is used to distinguish if a contig-graph represents an alternatively-spliced gene or duplicated genes. Contig-graphs representing divergent alleles from the same gene are distinguished from duplicated genes by assuming that the alleles have > 95% sequence identity (see Additional file 3 for more details on the graph-clustering method).
For each component within a graph-cluster, the BlastX results for its contigs were summarized in order to classify the component into one of five categories. 1) None of the contigs had a homology hit; 2) some of the contigs had a homology hit, but to different files in NCBI; 3) some of the contigs had a homology hit, and they were to the identical file in NCBI; 4) all of the contigs had a homology hit, but to different files in NCBI; 5) all of the contigs had a homology hit, and they were all to the identical file in NCBI. These were summarized for each type of component.
Sequence variants
Sequence variants (SNPs and INDELs) were identified and their probability of being a 'true' variant based on Bayesian analysis using the program GIGABAYES [ 58 , 59 ]. Since NEWBLER aligns reads allowing gaps instead of substitutions, sequence variant callers cannot identify SNPs and INDELs accurately. MOSAIK [ 60 ] was used to realign reads against contigs. Because homopolymer errors are more prevalent with 454 sequences than ABI sequences, this has to be taken into account when calculating probabilities. INDELs called by GIGABAYES were filtered out if they were in homopolymer regions. For high confident sequence variants, we filtered out SNPs and INDELs with read coverage < 5 or > 100 and the probability < 0.9.
The relationship between the number of variants per length of the contig was tested using a regression analysis in R. Contigs in the 99 th percentile of number of variants/contig length were considered highly variable. The ratio between transitions and transversions ((TS + 1)/(TV+ 1)) was calculated for each contigs containing SNPs. The one is added to allow the ratio to be calculated for contigs that had a zero for either TS or TV values. For each contig-containing SNP, for which we had a predicted ORF, we calculated the ratio of non-synonymous (Ka) to synonymous (Ks) polymorphisms ((Ka + 1)/(Ks+ 1)). (python script available at http://eco.bcb.iastate.edu/ ).
Comparisons between female and male
The sex-of-origin of each read was tracked through the assembly so that the contigs could be classified as being a "female" contig (only containing female reads), a "male" contig (only containing male reads), or a "both" contig (containing both male and female reads) (Figure 1 ). To identify genes that are expressed by only one of the sexes, we compared the BlastX for each contigs and singleton. For example, if a BlastX hit was found in only female contigs and/or female singletons, and had no homology to male contigs or male singletons at e-values down to 1e -50 , then it was classified as female-specific. The same process was used to identify male-specific genes. These 190 genes were compared to RefSeq for all vertebrates. Their GO slim assignments for Biological Processes were statistically compared in Blast2Go to test for enrichment of particular GO terms using Fisher's Exact Test. | Results and Discussion
Sampling and 454 GS-FLX Titanium Sequencing
Our goal in sampling was to maximize the identification of unique transcripts, while capturing the diversity of expressed transcripts across tissues, individuals, populations, and stress conditions. Therefore, keeping male and female samples separate, we pooled RNA from 35 garter snakes ( T. elegans ) of varying sizes/ages (at least 1 year old) into two sex-specific RNA samples (sampling details in Additional file 1 ). The snakes were both laboratory-born and field-caught from seven focal populations of the Sierra Nevada Mountains in California. These sex-specific pools of RNA were used to develop normalized cDNA libraries that were sequenced on separate halves of a GS-FLX Titanium (Roche/454 Life Sciences) PicoTitre plate. We also obtained an extra quarter plate of male library reads for quality control assessment. This resulted in 446 Mbp of sequence data, 1.24 million reads (i.e. expressed sequence tags) averaging 366 bp in length after cleaning (Table 1 ; see Additional file 2 for size distribution of reads). The cleaned reads have been deposited in the NCBI Short Read Archive ( SRA010134 ).
Assembly and Annotation
The male and female reads were pooled for assembly, but the sex-of-origin for each read was tracked, which allowed contigs to be categorized as containing reads from both sexes, from males only, or from females only (Figure 1 ). Relative to the garter snake, the sequenced (draft) genome closest in evolutionary relationship is the Anolis lizard ( Anolis carolinensis ), which shared its most recent common ancestor with snakes ~215 million years ago [ 30 ]. The next evolutionarily closest species with a sequenced genome is the chicken ( Gallus gallus ), which shared its most recent common ancestor with snakes ~285 million years ago [ 31 ]. We used 454 gsMapper 2.3 to map the garter snake reads to the draft Anolis lizard and the chicken genomes. Of cleaned reads, 773,997 (62%) and 536,900 (43%), respectively, were mapped, even with the minimum percent identity of 80%. Those mapped reads were assembled to 255,211 and 175,279 contigs for the lizard and chicken genomes, respectively.
For de novo assembly, reads were assembled using GS de novo Assembler (NEWBLER v2.0.00.22; Roche) resulting in 82,134 contigs - with 84% of the reads being placed - with an average depth of 7.6× coverage (Table 1 ). We then used the program MIRA [ 32 ] for additional assembly of the remaining singletons. This resulted in 33,338 singletons (25%) being assembled into an additional 14,245 contigs. The remaining singletons were mapped to contigs in order to remove sequence redundancy. Overall, 86.8% of reads were assembled into 96,379 contigs with 7.5% of the reads remaining as singletons, and 5.7% of reads being discarded (Table 1 ). Because many more of the reads were place with the de novo assembly, we use the de novo assembly for all subsequent analyses.
The percentage of reads assembled de novo is similar to other studies that have applied pyrosequencing to non-model organisms [ 25 , 27 , 33 ]. The large number of contigs is likely due to the extensive diversity in the initial RNA samples - pooled across individuals and populations - in the form of sequence variants and alternative splicing. Different organs, different sexes, and different environmental/stress conditions are known to produce extensive alternative spliced transcripts in vertebrates [ 34 ]. This variation causes misalignments between reads arising from the same genomic region, preventing correct assembly by most algorithms. For this reason, NEWBLER was used for assembly because it splits reads at the boundaries of variation in order to build contigs reflecting both static and variable regions [ 35 ] (see Additional file 3 for graphical representation of Newbler-split reads and assembly).
The NEWBLER contigs, the MIRA contigs, and the singletons were compared to three reference databases for annotation using BlastX: NCBI non-redundant protein database (NR); NCBI HomoloGene; and UniGene (Chicken). The sequences were also compared to the 18,031 Ensemble annotated genes (17,672 coding and 359 pseudogenes) from the Anolis lizard draft genome (AnoCar1.0; http://uswest.ensembl.org/Anolis_carolinensis/ ) using tBLASTx. Over 24% of all sequences (contigs and singletons) identified a homologue in at least one of these reference databases at e-value 1e -5 (Figure 2 : annotated data can be downloaded from http://eco.bcb.iastate.edu/ ). As expected with longer sequences, a higher proportion of the contigs found an identity (34%) relative to the singletons (13%). Most of the contigs that found an identity in one database were identified in all four of the reference databases (Figure 2 ). This subset of contigs likely contains the more conserved genes that are well characterized and thus found in all of these databases. The number of contigs matching a homologue is similar or higher to other sequencing studies on non-model organisms [ 28 , 33 ].
The number of unique homologous genes identified in each database ranged from 26,232 in NCBI-NR to 12,953 in UniGene (Chicken) (Table 2 ). The BlastX results against the HomoloGene database and the Anolis lizard Ensemble annotation likely give the most realistic estimate of the number of unique genes in our dataset for which we could assign a homology ID, approximately 13,000. Naturally, the more quickly evolving genes and the snake-specific genes would be unlikely to find homologues in any of these databases. Thus, undoubtedly, there are uncharacterized genes yet to be discovered in the other 76% of the sequences for which we could not assign an ID based on these reference databases. ORF predictions indicate that an additional 97% of the non-annotated sequences had a predicted open reading frame of at least 30 bp, which suggests that these were transcribed from protein-coding genes. The majority of the GO annotations assigned to the snake sequences were for the biological processes of metabolism and regulation, although there were also a smaller number of sequences assigned to reproduction, behaviour, and stress response (see Additional file 4 for GO pie graphs).
Using tBlastX, 55,715 snake transcripts (contigs and singletons) were also mapped to the lizard draft genome (AnoCar1.0). To identify 5' and 3' UTRs and non-coding RNAs, these matches to the Anolis genome were compared to the matches to the AnoCar1.0 Ensemble annotation of coding (18,031) and non-coding (2939) RNA. We identified 2322 snake transcripts that contained 5'UTR (286 matched to 5'UTR only, the rest matched 5'UTR along with protein coding sequence and/or intergenic sequence), and 3680 that contained 3'UTR (1018 matched to 3'UTR only, the rest matched 3'UTR along with protein coding sequence and/or intergenic sequence). Furthermore, 2,534 of our snake transcripts matched the Anolis non-coding RNAs, and 36,188 matched other intergenic regions of the Anolis genome and therefore may also be non-coding RNAs (see Additional file 5 to access these transcripts). A Snake Transcriptome Browser (GBrowse) of the snake transcripts mapped against the Anolis draft genome (AnoCar1.0) has been set up to visualize the data, and to access the assembled and annotated data files https://lims.cgb.indiana.edu/cgi-bin/gbrowse/telegans_bronikowski_2/ .
To check for additional non-coding RNAs, we used RNAmmer 1.2 [ 36 ] and tRNAscan-SE 1.23 [ 37 ]. One contig was predicted to be 8 S rRNA, although this was not supported by the homology search through NCBI non-redundant protein sequences. Thus, it is likely a false prediction. Five singletons were identified as tRNAs: two sequences for tRNA Asp , one for tRNA His , one for tRNA Lys , and one sequence for tRNA Ser . Additionally, pseudogenes were predicted from 5 contigs and 11 singletons.
Phylogenetic assessment of BLAST results
We use the phylogenetic distribution of the Blast hits to identify off-target sequences and as a qualitative assessment of the assembly. Because we indiscriminately used all RNA isolated from six tissue types and blood, which would contain disease and commensal organisms, we expected our dataset to contain off-target species sequences that did not originate from the garter snake genome. To identify these off-target sequences we used the metagenomics program MEGAN [ 38 ]. MEGAN assigns a sequence to the lowest common ancestor of all its Blast assignments at a particular cut-off e-value. Thus sequences placed at the deeper nodes of the phylogeny represent more conserved genes relative to the genes on the leaves the tree that are more specific to that species. We used the output from BlastX (e-value = 1e -5 ) against the NCBI-NR for the contigs and singletons (categorized based on sex-of-origin) to map the hits on the NCBI taxonomic tree (Figure 3 ). We identified 979 sequences (0.5% of the dataset) that were likely from off-target species including 51 that assigned to bacteria, 135 that assigned to viruses, and 11 that assigned to fungi. Interestingly, there was a bias towards bacterial sequences originating from the male reads suggesting the possibility that one or more of the males in our sample may have had a high bacterial infection load.
Equal proportions of sex-specific contigs and singletons were assigned to the nodes of the taxonomic tree, which implies that our ability to map to NCBI was not biased towards one or the other sex (Figure 3 ). Overall, 88% of the assigned sequences were assigned to chordates or a lineage within chordates; and 62% were assigned to tetrapods or a lineage within tetrapods (Figure 3 ). The placement of sequences at a particular node is highly dependent upon the relative abundance (or presence) of sequences available in NCBI-NR from each taxonomic group. Although there are far fewer protein sequences from reptiles ( Sauropsida ) in NCBI-NR compared to mammals ( Mammalia ), both lineages had roughly equal number of genes assigned within them. The species with the most number of genes matched was Gallus gallus (chicken), which is the evolutionarily closest species with a completed genome in the NCBI-NR protein database.
As an additional evaluation of the quality of our assembly and sequencing, we identified potential chimeric sequences that are made-up of two unique sequences that have been concatenated such that the sequence had two highly significant (< 1e -20 ) NCBI-NR hits to different genes, which aligned to different ends of the contig or singleton. We identified 24 contigs and 2 singletons that had such signatures (i.e., representing < 0.025%). Of these, two of the contigs were adjacent mitochondrial genes representing a correct assembly as the mitochondrial genome is known to be transcribed in large multi-gene segments [ 39 ]. The other 22 contigs and 2 singletons seemed to be true chimeras based on visual inspection of the BlastX hits. These chimeric sequences could be due to misassembly, chimerization during the library construction, true biological gene fusion events, or errors in GenBank. Overall, the phylogenetic distribution of the BlastX hits and the low percentage of chimera sequences provide qualitative support for our assembly.
Additional clustering
Complex vertebrate transcriptomes are characterized by numerous alternatively spliced transcripts and transcripts from duplicated genes [ 34 , 40 ] making de novo assembly difficult. We used two methods for additional clustering of the contigs and singletons into reference gene sets: clustering based on homology, and clustering based on contig-graphs (Figure 1 ). Homology clustering uses BlastX against reference databases to group contigs and singletons into clusters that are likely to have originated from the same gene. Homology clustering of contigs and singletons based on the HomoloGene database produced 8771 - 13,346 homology clusters, depending on the e-value cut-off used (Table 2 ). At e-value 1e -20 , each of these homology clusters contained 1 to 93 contigs with an average of two contigs (see Additional file 6 Panels A and B for distributions of contigs and clusters). Additionally, ~2000 of the contigs assigned to two or more homology clusters (HomoloGene accession); these contigs likely belong to gene families or are contigs with a strong domain found in multiple HomoloGenes. For example, contig00258 assigned to four HomoloGene clusters, each of which contains an ubiquitin conjugating enzyme E2 catalytic domain (UBCc). Homology comparisons to the draft gene models from the lizard genome had a smaller range of minimal gene sets (Table 2 ). These homology clusters contain contigs and singletons that represent alternative alleles, alternatively spliced transcripts, and non-overlapping contigs from the same gene and possibly closely related genes in the same family (particularly at the lower e-value).
Output clusters from the contig-graphing approach were used to group NEWBLER-generated contigs based on read-assembly information (see methods and Additional file 3 for a complete description). For alternatively spliced transcripts, highly diverged alleles and repeat regions of duplicated genes, NEWBLER (454 GS Assembler) splits reads into separate contigs. During assembly, split reads are tracked and that information can be used to reconstruct how contigs may be related. This can be illustrated graphically using networks (i.e. graph-clusters) such that nodes in the graphs represent contigs and an edge between two nodes represents a read split between two contigs. This clustering is independent of the contigs' homology to other databases, and allows for further identification of divergent alleles, alternatively spliced transcripts, and gene families based on the structure of the network and nucleotide similarity of the contigs in the network (see Additional file 3 for more details of this method). We constructed 6,860 graph contig clusters containing 27,860 NEWBLER contigs. The graph-clusters contained on average four contigs (range 2-95) (see Additional file 6 panel C for distribution of contigs in graph-clusters). This clustering method can be used to further merge variable alleles of the same gene, construct alternatively spliced transcripts, and identify gene families. Within these graph-clusters, we identified 554 components that are predicted to represent highly divergent alleles of the same gene that can then be merged, 1,293 components that are predicted to represent alternative splicing events, and 158 components (pairs of duplicated contigs) that are predicted to represent duplicated genes (Figure 4 ).
Evaluating the BlastX NCBI-NR results of the contigs within each graph-cluster component revealed differences among the classes of components. As you would expect from alleles of the same gene, all the contigs from merged components (highly variable alleles of the same gene) were more likely to match the same gene in NCBI-NR than were contigs from alternatively spliced or duplicated gene components. Likewise, contigs in the 158 duplicated gene components were more likely to match different genes in NCBI-NR than either the merged components or the alternative spliced components (Figure 4 ).
As an example of the usefulness of this graph-clustering method, we highlight graph-cluster05625 (Figure 5 ) that consists of 27 contigs, of which nine contigs have identical hits in NCIB-NR, three contigs have different hits in NCBI-NR, and 15 contigs have no hits in NCBI-NR. Ten of these NCBI-NR matches are to the hypervariable immunological gene MHC class I, which is well documented as being under diversifying selection in other species [ 41 ]. This clustering method allowed us to identify additional contigs (not identified by our homology searches) that represent multiple alleles, alternatively spliced transcripts and a potential gene duplication event for a highly important immunological gene complex.
Sequence variants
One of our goals for including diverse individuals from multiple populations was to identify sequence variants (SNPs and INDELS) that would be useful in future studies. At a Bayesian probability of 90% we identified 126,946 variants: 95,295 SNPs and 31,651 INDELs in either single contigs (23,615) or across contigs that had been merged into graph-clusters (549). Over 110,000 of these variants had >99% probability of being a true variant rather than a sequencing error. Of the single contigs that had variants, the average number of variants per contig is 5.16 (SNPs = 3.84, INDELs = 1.32); with an average rate of one variant per 200 bp window (see Additional file 7 for details on variants). The number of variants (SNPs and INDELs) per contig had a weak but significant correlation with contig length (F = 1980, DF = 23,615, R 2 = 0.07, p-value < 0.01: Figure 6 panel A), with the scatter suggesting that some contigs represented highly conserved genes with low variation and some contigs with high levels of variation for contig length.
The degree and type of variability within a contig can indicate selection acting on the gene. As expected, large contigs with few variants tended to be from highly conserved genes including, MACF1 (microtubule-actin cross-linking factor 1), and myosin. We identified 236 contigs that were highly variable (in the 99 th percentile for highest ratio of variants per contig length) and, as expected of quickly evolving genes, few of these (19%) had homology hits in the BlastX searches, but 98% of them had predicted open reading frames (ORF) larger than 30 bp. This size criterion suggests that the ORFs were within true protein coding sequences. Of the 45 contigs that were highly variable and had a BlastX hit in NCBI-NR, at least 8 were from off-target sequences and 11 were associated with transposable elements, but there were also many sequences of interest that are known to be hypervariable and/or quickly evolving in vertebrate taxa. These included three immunological genes (MHC class I, complement factor-H related protein, interferon regulatory factor 7), a snake venom gene (venom factor 1), and a predicted pheromone receptor (see Additional file 8 for full annotation data on these genes of interest). Interestingly, there were also two highly variable genes involved in lipid metabolism or lipid oxidation: peroxisomal long-chain acyl-CoA thioesterase, and fatty acid desaturase 1. These genes may be particularly relevant for future studies on the evolution of metabolism and stress response in these garter snake populations.
Mutations to a nucleotide of a similar structure (i.e., TS: transitions - mutations from a purine to a purine, or from a pyrimadine to a pyrimidine) occur more often then transversions (TV: mutation from a purine to a pyrimadine or vice-versa). Thus, a TS/TV ratio <1 may reveal sequences subjected to diversifying selection [ 42 ]. We found 73, 836 TSs and 21, 459 TVs in this dataset. We identified 2, 165 contigs with a TS/TV <1. For SNPs within predicted coding regions, we determined whether they were non-synonymous polymorphisms (Ka) that changed the amino acid, or were synonymous polymorphisms (Ks). Overall, 29, 883 of the SNPs found in a coding region were non-synonymous and 23, 252 were synonymous. We found 8, 417 contigs (8.7% of all contigs) with a Ka/Ks ratio >1. This indicates that mutation(s) have changed the amino acid sequence more than would be expected under a neutral model, and that these genes may be under diversifying selection within or among these snake populations. The distributions of TS/TV and Ka/Ks are in Figure 6A . Of most interest are the 16 contigs at the intersection of Ka/Ks >1, TS/TV <1, and the 99 th percentile of highly variable contigs. Only three of these could be assigned a putative identification based on homology: the immune complement factor-H related protein, fatty acid desaturase 1, and a KRAB transcription factor. Revisiting the MHC class I graph-cluster05625 (Figure 5 ) that consists of 27 contigs, of the 20 contigs had variants 10 had Ka/Ks > 1. As predicted above, this further supports diversifying selection across this complex. Additional highly variable genes with high Ka/Ks ratios are likely to be targets of diversifying selection, potentially diversifying across the populations (or ecotypes) of the garter snakes.
Comparison between female and male
When the male and female reads were pooled and assembled into contigs, each read was tracked by the sex from which it was generated. Thus, the contigs and singletons could be classified on the origin of its reads. Focusing only on the sequences for which we could assign an ID based on homology, NCBI-NR BlastX hits (1e -50 ) were summarized based on whether they were unique to females (i.e., found only in female contigs and/or female singletons), were unique to males, or were composed of reads from both sexes. In this way we identified 190 genes (195 contigs) that were only present in the mRNA sequenced from one of the sexes (see Additional file 8 for full annotation data on these genes of interest). Of these 190 genes, 84 were expressed only in females and 106 were expressed only in males. While this is a relatively low number of genes that are sex-specific, recall that our sex-specific pools of RNA included seven tissue types and were normalized in order to maximize the number of unique transcripts. Therefore, these data are not quantitative differences in expression in a particular tissue type, but are presence/absence data across all seven tissues. Most other studies that look at sex-specific differences use microarrays (or RNA-seq) on non-normalized libraries from a particular tissue and thus have quantitative data to show difference in the levels of expression between the sexes. Indeed it is at the quantitative level that most genes are biased in their expression between the sexes [ 43 ]. Additionally, most genes that are biased in their expression between sexes are specific to a particular tissue. For example, a microarray study on chicken brain, gonad, and heart, identified ~13, 000 genes that had quantitative differences in one particular tissue, but only four genes were significantly different in all three tissues [ 44 ]. As well, sex-specific genes are typically quickly evolving; thus, homologues may not be present in the available databases [ 43 ].
The female-specific sequences were enriched for GO terms for "biosynthetic processes" relative to the male-specific sequences (FDR < 0.006, p-value < 0.0002: see Additional file 9 for distribution of GO terms by sex). Although many of these sex-specific genes were uncharacterized or classified as "predicted coding genes", some met our expectations for being sex-specific. For example, two of the male-specific genes are known to be involved in spermatogenesis (SPATA18, SPATA22) and two in sperm motility (CATSPERG, CATSPER2). Additionally, two female-specific genes are known to be involved in hormonal signalling and regulation (GnRH receptor and Irg1) [ 45 - 47 ]. Five of these 190 sex-specific genes also had a TS/TV ratio ≤ 1, male-specific: BTBD16, CATSPERG, CATSPER2; and female-specific: RAG1 and CDX4. Additionally, 13 of the sex-specific genes had a Ka/Ks ratio > 1, three of these overlapped with those that had a low TS/TV: male-specific CATSPERG, female specific CDX4 (also on the human X chromosome), and female specific RAG1.
These analyses suggest the CATSPER genes maybe under strong selective pressure. The CATSPER genes (1-4) produce proteins that form an ion channel that is specific to the sperm flagellum, and are required for hyperactivated motility during the fertilization process [ 48 - 50 ]. Mutations in the CATSPER genes cause infertility in mice and humans [ 49 - 51 ]. The CATSPERG gene encodes an additional protein that is associated with the CATSPER ion channel [ 52 ]. As with many reptiles, garter snakes mate multiply and have multiple paternity litters [ 53 ], which could lead to sperm competition and selective pressure on sperm function and motility. Further studies are needed to test the role of diversity in the CATSPER genes in reproductive fitness and sperm competition in this snake system. Although we hypothesize that these male-specific genes associated with sperm production and fertilization are being expressed in the gonads, we cannot address the question of tissue-of-origin with these data at this time. However, we are currently looking at gene-expression differences in individual-based and tissue-specific libraries.
Both snakes and birds have ZW/ZZ sex chromosomes, in which the female is the heterogametic sex. The chicken and snake are separated by ~285 million years and it has been well documented that reptile sex chromosomes have undergone drastic rearrangements over that time [ 54 ]. Therefore, whether any snake sequences aligned to the chicken sex chromosomes would be a very interesting result. Indeed, the snake sequences aligned to 15 genes on the chicken Z chromosome. Interestingly, four of these had female-specific expression and two had male-specific expression (see Additional file 8 for full annotation data of these genes of interest). It remains to be determined whether these genes reside on the garter snake Z (or W) chromosome. Equally interesting is that one gene known to be on the human X chromosome, CDX4, had female specific expression, had a TS/TV < 1, and a Ka/Ks > 1 suggesting it may be a sex-conflict gene that is quickly evolving in these snake populations.
Because the cDNA libraries used for sequencing were normalized, the identification of sex-specifically expressed genes is based on presence/absence rather than quantitative measures. Some of the genes identified here as sex-specific genes are likely under sex-biased expression, potentially the result of sex conflict resolved at the level of gene expression. Additionally, some of these sex-specific genes may reside on the garter snake sex chromosomes. Additional large-scale studies at the quantitative level will verify the sex-specific expression of these genes. | Results and Discussion
Sampling and 454 GS-FLX Titanium Sequencing
Our goal in sampling was to maximize the identification of unique transcripts, while capturing the diversity of expressed transcripts across tissues, individuals, populations, and stress conditions. Therefore, keeping male and female samples separate, we pooled RNA from 35 garter snakes ( T. elegans ) of varying sizes/ages (at least 1 year old) into two sex-specific RNA samples (sampling details in Additional file 1 ). The snakes were both laboratory-born and field-caught from seven focal populations of the Sierra Nevada Mountains in California. These sex-specific pools of RNA were used to develop normalized cDNA libraries that were sequenced on separate halves of a GS-FLX Titanium (Roche/454 Life Sciences) PicoTitre plate. We also obtained an extra quarter plate of male library reads for quality control assessment. This resulted in 446 Mbp of sequence data, 1.24 million reads (i.e. expressed sequence tags) averaging 366 bp in length after cleaning (Table 1 ; see Additional file 2 for size distribution of reads). The cleaned reads have been deposited in the NCBI Short Read Archive ( SRA010134 ).
Assembly and Annotation
The male and female reads were pooled for assembly, but the sex-of-origin for each read was tracked, which allowed contigs to be categorized as containing reads from both sexes, from males only, or from females only (Figure 1 ). Relative to the garter snake, the sequenced (draft) genome closest in evolutionary relationship is the Anolis lizard ( Anolis carolinensis ), which shared its most recent common ancestor with snakes ~215 million years ago [ 30 ]. The next evolutionarily closest species with a sequenced genome is the chicken ( Gallus gallus ), which shared its most recent common ancestor with snakes ~285 million years ago [ 31 ]. We used 454 gsMapper 2.3 to map the garter snake reads to the draft Anolis lizard and the chicken genomes. Of cleaned reads, 773,997 (62%) and 536,900 (43%), respectively, were mapped, even with the minimum percent identity of 80%. Those mapped reads were assembled to 255,211 and 175,279 contigs for the lizard and chicken genomes, respectively.
For de novo assembly, reads were assembled using GS de novo Assembler (NEWBLER v2.0.00.22; Roche) resulting in 82,134 contigs - with 84% of the reads being placed - with an average depth of 7.6× coverage (Table 1 ). We then used the program MIRA [ 32 ] for additional assembly of the remaining singletons. This resulted in 33,338 singletons (25%) being assembled into an additional 14,245 contigs. The remaining singletons were mapped to contigs in order to remove sequence redundancy. Overall, 86.8% of reads were assembled into 96,379 contigs with 7.5% of the reads remaining as singletons, and 5.7% of reads being discarded (Table 1 ). Because many more of the reads were place with the de novo assembly, we use the de novo assembly for all subsequent analyses.
The percentage of reads assembled de novo is similar to other studies that have applied pyrosequencing to non-model organisms [ 25 , 27 , 33 ]. The large number of contigs is likely due to the extensive diversity in the initial RNA samples - pooled across individuals and populations - in the form of sequence variants and alternative splicing. Different organs, different sexes, and different environmental/stress conditions are known to produce extensive alternative spliced transcripts in vertebrates [ 34 ]. This variation causes misalignments between reads arising from the same genomic region, preventing correct assembly by most algorithms. For this reason, NEWBLER was used for assembly because it splits reads at the boundaries of variation in order to build contigs reflecting both static and variable regions [ 35 ] (see Additional file 3 for graphical representation of Newbler-split reads and assembly).
The NEWBLER contigs, the MIRA contigs, and the singletons were compared to three reference databases for annotation using BlastX: NCBI non-redundant protein database (NR); NCBI HomoloGene; and UniGene (Chicken). The sequences were also compared to the 18,031 Ensemble annotated genes (17,672 coding and 359 pseudogenes) from the Anolis lizard draft genome (AnoCar1.0; http://uswest.ensembl.org/Anolis_carolinensis/ ) using tBLASTx. Over 24% of all sequences (contigs and singletons) identified a homologue in at least one of these reference databases at e-value 1e -5 (Figure 2 : annotated data can be downloaded from http://eco.bcb.iastate.edu/ ). As expected with longer sequences, a higher proportion of the contigs found an identity (34%) relative to the singletons (13%). Most of the contigs that found an identity in one database were identified in all four of the reference databases (Figure 2 ). This subset of contigs likely contains the more conserved genes that are well characterized and thus found in all of these databases. The number of contigs matching a homologue is similar or higher to other sequencing studies on non-model organisms [ 28 , 33 ].
The number of unique homologous genes identified in each database ranged from 26,232 in NCBI-NR to 12,953 in UniGene (Chicken) (Table 2 ). The BlastX results against the HomoloGene database and the Anolis lizard Ensemble annotation likely give the most realistic estimate of the number of unique genes in our dataset for which we could assign a homology ID, approximately 13,000. Naturally, the more quickly evolving genes and the snake-specific genes would be unlikely to find homologues in any of these databases. Thus, undoubtedly, there are uncharacterized genes yet to be discovered in the other 76% of the sequences for which we could not assign an ID based on these reference databases. ORF predictions indicate that an additional 97% of the non-annotated sequences had a predicted open reading frame of at least 30 bp, which suggests that these were transcribed from protein-coding genes. The majority of the GO annotations assigned to the snake sequences were for the biological processes of metabolism and regulation, although there were also a smaller number of sequences assigned to reproduction, behaviour, and stress response (see Additional file 4 for GO pie graphs).
Using tBlastX, 55,715 snake transcripts (contigs and singletons) were also mapped to the lizard draft genome (AnoCar1.0). To identify 5' and 3' UTRs and non-coding RNAs, these matches to the Anolis genome were compared to the matches to the AnoCar1.0 Ensemble annotation of coding (18,031) and non-coding (2939) RNA. We identified 2322 snake transcripts that contained 5'UTR (286 matched to 5'UTR only, the rest matched 5'UTR along with protein coding sequence and/or intergenic sequence), and 3680 that contained 3'UTR (1018 matched to 3'UTR only, the rest matched 3'UTR along with protein coding sequence and/or intergenic sequence). Furthermore, 2,534 of our snake transcripts matched the Anolis non-coding RNAs, and 36,188 matched other intergenic regions of the Anolis genome and therefore may also be non-coding RNAs (see Additional file 5 to access these transcripts). A Snake Transcriptome Browser (GBrowse) of the snake transcripts mapped against the Anolis draft genome (AnoCar1.0) has been set up to visualize the data, and to access the assembled and annotated data files https://lims.cgb.indiana.edu/cgi-bin/gbrowse/telegans_bronikowski_2/ .
To check for additional non-coding RNAs, we used RNAmmer 1.2 [ 36 ] and tRNAscan-SE 1.23 [ 37 ]. One contig was predicted to be 8 S rRNA, although this was not supported by the homology search through NCBI non-redundant protein sequences. Thus, it is likely a false prediction. Five singletons were identified as tRNAs: two sequences for tRNA Asp , one for tRNA His , one for tRNA Lys , and one sequence for tRNA Ser . Additionally, pseudogenes were predicted from 5 contigs and 11 singletons.
Phylogenetic assessment of BLAST results
We use the phylogenetic distribution of the Blast hits to identify off-target sequences and as a qualitative assessment of the assembly. Because we indiscriminately used all RNA isolated from six tissue types and blood, which would contain disease and commensal organisms, we expected our dataset to contain off-target species sequences that did not originate from the garter snake genome. To identify these off-target sequences we used the metagenomics program MEGAN [ 38 ]. MEGAN assigns a sequence to the lowest common ancestor of all its Blast assignments at a particular cut-off e-value. Thus sequences placed at the deeper nodes of the phylogeny represent more conserved genes relative to the genes on the leaves the tree that are more specific to that species. We used the output from BlastX (e-value = 1e -5 ) against the NCBI-NR for the contigs and singletons (categorized based on sex-of-origin) to map the hits on the NCBI taxonomic tree (Figure 3 ). We identified 979 sequences (0.5% of the dataset) that were likely from off-target species including 51 that assigned to bacteria, 135 that assigned to viruses, and 11 that assigned to fungi. Interestingly, there was a bias towards bacterial sequences originating from the male reads suggesting the possibility that one or more of the males in our sample may have had a high bacterial infection load.
Equal proportions of sex-specific contigs and singletons were assigned to the nodes of the taxonomic tree, which implies that our ability to map to NCBI was not biased towards one or the other sex (Figure 3 ). Overall, 88% of the assigned sequences were assigned to chordates or a lineage within chordates; and 62% were assigned to tetrapods or a lineage within tetrapods (Figure 3 ). The placement of sequences at a particular node is highly dependent upon the relative abundance (or presence) of sequences available in NCBI-NR from each taxonomic group. Although there are far fewer protein sequences from reptiles ( Sauropsida ) in NCBI-NR compared to mammals ( Mammalia ), both lineages had roughly equal number of genes assigned within them. The species with the most number of genes matched was Gallus gallus (chicken), which is the evolutionarily closest species with a completed genome in the NCBI-NR protein database.
As an additional evaluation of the quality of our assembly and sequencing, we identified potential chimeric sequences that are made-up of two unique sequences that have been concatenated such that the sequence had two highly significant (< 1e -20 ) NCBI-NR hits to different genes, which aligned to different ends of the contig or singleton. We identified 24 contigs and 2 singletons that had such signatures (i.e., representing < 0.025%). Of these, two of the contigs were adjacent mitochondrial genes representing a correct assembly as the mitochondrial genome is known to be transcribed in large multi-gene segments [ 39 ]. The other 22 contigs and 2 singletons seemed to be true chimeras based on visual inspection of the BlastX hits. These chimeric sequences could be due to misassembly, chimerization during the library construction, true biological gene fusion events, or errors in GenBank. Overall, the phylogenetic distribution of the BlastX hits and the low percentage of chimera sequences provide qualitative support for our assembly.
Additional clustering
Complex vertebrate transcriptomes are characterized by numerous alternatively spliced transcripts and transcripts from duplicated genes [ 34 , 40 ] making de novo assembly difficult. We used two methods for additional clustering of the contigs and singletons into reference gene sets: clustering based on homology, and clustering based on contig-graphs (Figure 1 ). Homology clustering uses BlastX against reference databases to group contigs and singletons into clusters that are likely to have originated from the same gene. Homology clustering of contigs and singletons based on the HomoloGene database produced 8771 - 13,346 homology clusters, depending on the e-value cut-off used (Table 2 ). At e-value 1e -20 , each of these homology clusters contained 1 to 93 contigs with an average of two contigs (see Additional file 6 Panels A and B for distributions of contigs and clusters). Additionally, ~2000 of the contigs assigned to two or more homology clusters (HomoloGene accession); these contigs likely belong to gene families or are contigs with a strong domain found in multiple HomoloGenes. For example, contig00258 assigned to four HomoloGene clusters, each of which contains an ubiquitin conjugating enzyme E2 catalytic domain (UBCc). Homology comparisons to the draft gene models from the lizard genome had a smaller range of minimal gene sets (Table 2 ). These homology clusters contain contigs and singletons that represent alternative alleles, alternatively spliced transcripts, and non-overlapping contigs from the same gene and possibly closely related genes in the same family (particularly at the lower e-value).
Output clusters from the contig-graphing approach were used to group NEWBLER-generated contigs based on read-assembly information (see methods and Additional file 3 for a complete description). For alternatively spliced transcripts, highly diverged alleles and repeat regions of duplicated genes, NEWBLER (454 GS Assembler) splits reads into separate contigs. During assembly, split reads are tracked and that information can be used to reconstruct how contigs may be related. This can be illustrated graphically using networks (i.e. graph-clusters) such that nodes in the graphs represent contigs and an edge between two nodes represents a read split between two contigs. This clustering is independent of the contigs' homology to other databases, and allows for further identification of divergent alleles, alternatively spliced transcripts, and gene families based on the structure of the network and nucleotide similarity of the contigs in the network (see Additional file 3 for more details of this method). We constructed 6,860 graph contig clusters containing 27,860 NEWBLER contigs. The graph-clusters contained on average four contigs (range 2-95) (see Additional file 6 panel C for distribution of contigs in graph-clusters). This clustering method can be used to further merge variable alleles of the same gene, construct alternatively spliced transcripts, and identify gene families. Within these graph-clusters, we identified 554 components that are predicted to represent highly divergent alleles of the same gene that can then be merged, 1,293 components that are predicted to represent alternative splicing events, and 158 components (pairs of duplicated contigs) that are predicted to represent duplicated genes (Figure 4 ).
Evaluating the BlastX NCBI-NR results of the contigs within each graph-cluster component revealed differences among the classes of components. As you would expect from alleles of the same gene, all the contigs from merged components (highly variable alleles of the same gene) were more likely to match the same gene in NCBI-NR than were contigs from alternatively spliced or duplicated gene components. Likewise, contigs in the 158 duplicated gene components were more likely to match different genes in NCBI-NR than either the merged components or the alternative spliced components (Figure 4 ).
As an example of the usefulness of this graph-clustering method, we highlight graph-cluster05625 (Figure 5 ) that consists of 27 contigs, of which nine contigs have identical hits in NCIB-NR, three contigs have different hits in NCBI-NR, and 15 contigs have no hits in NCBI-NR. Ten of these NCBI-NR matches are to the hypervariable immunological gene MHC class I, which is well documented as being under diversifying selection in other species [ 41 ]. This clustering method allowed us to identify additional contigs (not identified by our homology searches) that represent multiple alleles, alternatively spliced transcripts and a potential gene duplication event for a highly important immunological gene complex.
Sequence variants
One of our goals for including diverse individuals from multiple populations was to identify sequence variants (SNPs and INDELS) that would be useful in future studies. At a Bayesian probability of 90% we identified 126,946 variants: 95,295 SNPs and 31,651 INDELs in either single contigs (23,615) or across contigs that had been merged into graph-clusters (549). Over 110,000 of these variants had >99% probability of being a true variant rather than a sequencing error. Of the single contigs that had variants, the average number of variants per contig is 5.16 (SNPs = 3.84, INDELs = 1.32); with an average rate of one variant per 200 bp window (see Additional file 7 for details on variants). The number of variants (SNPs and INDELs) per contig had a weak but significant correlation with contig length (F = 1980, DF = 23,615, R 2 = 0.07, p-value < 0.01: Figure 6 panel A), with the scatter suggesting that some contigs represented highly conserved genes with low variation and some contigs with high levels of variation for contig length.
The degree and type of variability within a contig can indicate selection acting on the gene. As expected, large contigs with few variants tended to be from highly conserved genes including, MACF1 (microtubule-actin cross-linking factor 1), and myosin. We identified 236 contigs that were highly variable (in the 99 th percentile for highest ratio of variants per contig length) and, as expected of quickly evolving genes, few of these (19%) had homology hits in the BlastX searches, but 98% of them had predicted open reading frames (ORF) larger than 30 bp. This size criterion suggests that the ORFs were within true protein coding sequences. Of the 45 contigs that were highly variable and had a BlastX hit in NCBI-NR, at least 8 were from off-target sequences and 11 were associated with transposable elements, but there were also many sequences of interest that are known to be hypervariable and/or quickly evolving in vertebrate taxa. These included three immunological genes (MHC class I, complement factor-H related protein, interferon regulatory factor 7), a snake venom gene (venom factor 1), and a predicted pheromone receptor (see Additional file 8 for full annotation data on these genes of interest). Interestingly, there were also two highly variable genes involved in lipid metabolism or lipid oxidation: peroxisomal long-chain acyl-CoA thioesterase, and fatty acid desaturase 1. These genes may be particularly relevant for future studies on the evolution of metabolism and stress response in these garter snake populations.
Mutations to a nucleotide of a similar structure (i.e., TS: transitions - mutations from a purine to a purine, or from a pyrimadine to a pyrimidine) occur more often then transversions (TV: mutation from a purine to a pyrimadine or vice-versa). Thus, a TS/TV ratio <1 may reveal sequences subjected to diversifying selection [ 42 ]. We found 73, 836 TSs and 21, 459 TVs in this dataset. We identified 2, 165 contigs with a TS/TV <1. For SNPs within predicted coding regions, we determined whether they were non-synonymous polymorphisms (Ka) that changed the amino acid, or were synonymous polymorphisms (Ks). Overall, 29, 883 of the SNPs found in a coding region were non-synonymous and 23, 252 were synonymous. We found 8, 417 contigs (8.7% of all contigs) with a Ka/Ks ratio >1. This indicates that mutation(s) have changed the amino acid sequence more than would be expected under a neutral model, and that these genes may be under diversifying selection within or among these snake populations. The distributions of TS/TV and Ka/Ks are in Figure 6A . Of most interest are the 16 contigs at the intersection of Ka/Ks >1, TS/TV <1, and the 99 th percentile of highly variable contigs. Only three of these could be assigned a putative identification based on homology: the immune complement factor-H related protein, fatty acid desaturase 1, and a KRAB transcription factor. Revisiting the MHC class I graph-cluster05625 (Figure 5 ) that consists of 27 contigs, of the 20 contigs had variants 10 had Ka/Ks > 1. As predicted above, this further supports diversifying selection across this complex. Additional highly variable genes with high Ka/Ks ratios are likely to be targets of diversifying selection, potentially diversifying across the populations (or ecotypes) of the garter snakes.
Comparison between female and male
When the male and female reads were pooled and assembled into contigs, each read was tracked by the sex from which it was generated. Thus, the contigs and singletons could be classified on the origin of its reads. Focusing only on the sequences for which we could assign an ID based on homology, NCBI-NR BlastX hits (1e -50 ) were summarized based on whether they were unique to females (i.e., found only in female contigs and/or female singletons), were unique to males, or were composed of reads from both sexes. In this way we identified 190 genes (195 contigs) that were only present in the mRNA sequenced from one of the sexes (see Additional file 8 for full annotation data on these genes of interest). Of these 190 genes, 84 were expressed only in females and 106 were expressed only in males. While this is a relatively low number of genes that are sex-specific, recall that our sex-specific pools of RNA included seven tissue types and were normalized in order to maximize the number of unique transcripts. Therefore, these data are not quantitative differences in expression in a particular tissue type, but are presence/absence data across all seven tissues. Most other studies that look at sex-specific differences use microarrays (or RNA-seq) on non-normalized libraries from a particular tissue and thus have quantitative data to show difference in the levels of expression between the sexes. Indeed it is at the quantitative level that most genes are biased in their expression between the sexes [ 43 ]. Additionally, most genes that are biased in their expression between sexes are specific to a particular tissue. For example, a microarray study on chicken brain, gonad, and heart, identified ~13, 000 genes that had quantitative differences in one particular tissue, but only four genes were significantly different in all three tissues [ 44 ]. As well, sex-specific genes are typically quickly evolving; thus, homologues may not be present in the available databases [ 43 ].
The female-specific sequences were enriched for GO terms for "biosynthetic processes" relative to the male-specific sequences (FDR < 0.006, p-value < 0.0002: see Additional file 9 for distribution of GO terms by sex). Although many of these sex-specific genes were uncharacterized or classified as "predicted coding genes", some met our expectations for being sex-specific. For example, two of the male-specific genes are known to be involved in spermatogenesis (SPATA18, SPATA22) and two in sperm motility (CATSPERG, CATSPER2). Additionally, two female-specific genes are known to be involved in hormonal signalling and regulation (GnRH receptor and Irg1) [ 45 - 47 ]. Five of these 190 sex-specific genes also had a TS/TV ratio ≤ 1, male-specific: BTBD16, CATSPERG, CATSPER2; and female-specific: RAG1 and CDX4. Additionally, 13 of the sex-specific genes had a Ka/Ks ratio > 1, three of these overlapped with those that had a low TS/TV: male-specific CATSPERG, female specific CDX4 (also on the human X chromosome), and female specific RAG1.
These analyses suggest the CATSPER genes maybe under strong selective pressure. The CATSPER genes (1-4) produce proteins that form an ion channel that is specific to the sperm flagellum, and are required for hyperactivated motility during the fertilization process [ 48 - 50 ]. Mutations in the CATSPER genes cause infertility in mice and humans [ 49 - 51 ]. The CATSPERG gene encodes an additional protein that is associated with the CATSPER ion channel [ 52 ]. As with many reptiles, garter snakes mate multiply and have multiple paternity litters [ 53 ], which could lead to sperm competition and selective pressure on sperm function and motility. Further studies are needed to test the role of diversity in the CATSPER genes in reproductive fitness and sperm competition in this snake system. Although we hypothesize that these male-specific genes associated with sperm production and fertilization are being expressed in the gonads, we cannot address the question of tissue-of-origin with these data at this time. However, we are currently looking at gene-expression differences in individual-based and tissue-specific libraries.
Both snakes and birds have ZW/ZZ sex chromosomes, in which the female is the heterogametic sex. The chicken and snake are separated by ~285 million years and it has been well documented that reptile sex chromosomes have undergone drastic rearrangements over that time [ 54 ]. Therefore, whether any snake sequences aligned to the chicken sex chromosomes would be a very interesting result. Indeed, the snake sequences aligned to 15 genes on the chicken Z chromosome. Interestingly, four of these had female-specific expression and two had male-specific expression (see Additional file 8 for full annotation data of these genes of interest). It remains to be determined whether these genes reside on the garter snake Z (or W) chromosome. Equally interesting is that one gene known to be on the human X chromosome, CDX4, had female specific expression, had a TS/TV < 1, and a Ka/Ks > 1 suggesting it may be a sex-conflict gene that is quickly evolving in these snake populations.
Because the cDNA libraries used for sequencing were normalized, the identification of sex-specifically expressed genes is based on presence/absence rather than quantitative measures. Some of the genes identified here as sex-specific genes are likely under sex-biased expression, potentially the result of sex conflict resolved at the level of gene expression. Additionally, some of these sex-specific genes may reside on the garter snake sex chromosomes. Additional large-scale studies at the quantitative level will verify the sex-specific expression of these genes. | Conclusions
We have successfully sequenced the first large-scale, multi-organ transcriptome for an ectothermic vertebrate using pyrosequencing and de novo assembly. In the process, we use a method for graphically clustering contigs after NEWBLER assembly that allowed us to identify divergent alleles, alternatively spliced transcripts and gene families. We have identified a number of interesting genes that are sex-specifically expressed and/or that are predicted to be quickly evolving that beg for additional investigation. These are the starting points for genetic studies on evolution of metabolic and immune function, sexual conflict resolution, as well as the evolution of sex chromosomes.
This transcriptome is the most comprehensive set of published EST sequences available for an individual ectothermic reptile species. It has increased the number of nucleotide ESTs available for ectothermic reptiles 5×, and for snakes 50×. Additionally, we have identified over 100, 000 high confidence variants (SNPs and INDELs) that can be used for population genetic studies, and quantitative trait mapping in this and related species.
These sequence data are a tool for future gene expression experiments, and comparative transcriptomic, genomic, and metabolomic studies. They can assist interested researchers to address evolutionary- and ecological-genomic questions in this and other reptile species. Ongoing and future studies can use this generalized transcriptome as a reference for mapping quantitative expression and sequence data from experiments that use, for example, short-read sequencing technologies. By combining these new sequencing technologies our labs hope to gain insight into how these snake life-history ecotypes have evolved, as well as how sex-conflict genes evolve. Overall, this is a valuable resource for the study of evolutionary important traits at the molecular level. | Background
The reptiles, characterized by both diversity and unique evolutionary adaptations, provide a comprehensive system for comparative studies of metabolism, physiology, and development. However, molecular resources for ectothermic reptiles are severely limited, hampering our ability to study the genetic basis for many evolutionarily important traits such as metabolic plasticity, extreme longevity, limblessness, venom, and freeze tolerance. Here we use massively parallel sequencing (454 GS-FLX Titanium) to generate a transcriptome of the western terrestrial garter snake ( Thamnophis elegans ) with two goals in mind. First, we develop a molecular resource for an ectothermic reptile; and second, we use these sex-specific transcriptomes to identify differences in the presence of expressed transcripts and potential genes of evolutionary interest.
Results
Using sex-specific pools of RNA (one pool for females, one pool for males) representing 7 tissue types and 35 diverse individuals, we produced 1.24 million sequence reads, which averaged 366 bp in length after cleaning. Assembly of the cleaned reads from both sexes with NEWBLER and MIRA resulted in 96,379 contigs containing 87% of the cleaned reads. Over 34% of these contigs and 13% of the singletons were annotated based on homology to previously identified proteins. From these homology assignments, additional clustering, and ORF predictions, we estimate that this transcriptome contains ~13,000 unique genes that were previously identified in other species and over 66,000 transcripts from unidentified protein-coding genes. Furthermore, we use a graph-clustering method to identify contigs linked by NEWBLER-split reads that represent divergent alleles, gene duplications, and alternatively spliced transcripts. Beyond gene identification, we identified 95,295 SNPs and 31,651 INDELs. From these sex-specific transcriptomes, we identified 190 genes that were only present in the mRNA sequenced from one of the sexes (84 female-specific, 106 male-specific), and many highly variable genes of evolutionary interest.
Conclusions
This is the first large-scale, multi-organ transcriptome for an ectothermic reptile. This resource provides the most comprehensive set of EST sequences available for an individual ectothermic reptile species, increasing the number of snake ESTs 50-fold. We have identified genes that appear to be under evolutionary selection and those that are sex-specific. This resource will assist studies on gene expression and comparative genomics, and will facilitate the study of evolutionarily important traits at the molecular level. | Authors' contributions
TSS, SP, and AMB: conceived and designed the research plan. AMB collected animals; TSS and AMB: dissected tissues and isolated the RNA; KM: optimized the normalization and library preparation for Titanium platform and conducted the sequencing; J-HC: designed the bioinformatic analysis strategy and improved the assembly; HT: analyzed homology-based annotation and ORF prediction; YIY: analyzed the graph-clustering and sequence variants; JLV and TSS: conducted Ka/Ks analyses, additional annotation, and metagenomic analysis; TSS: drafted the manuscript. All authors contributed to the interpretation of the results and the content of the final manuscript.
Supplementary Material | Acknowledgements
We would like to acknowledge Steve Arnold for help collecting animals; Megan Manes and Courtney Wolken for laboratory assistance; James Ford for transcriptome library construction and sequencing; Suzanne McGaugh, Lex Flagel, Autum Pairett, Dan Warner, the Bronikowski Lab for comments, and three anonymous reviewers for their comments on improving this manuscript. Computer support was provided by the University Information Technology Services (UITS) and by The Center for Genomics and Bioinformatics computing group at IU. This research was funded by the National Science Foundation: IGERT in Computational Molecular Biology 0504304 (TSS), [IOS-0922528] (AMB); the National Research Foundation of Korea Grant funded by the Korean Government [NRF-2009-352-D00275] (HT), the ISU Laurence H. Baker Center (SRP), and the ISU Center for Integrative Animal Genomics (AMB). | CC BY | no | 2022-01-12 15:21:38 | BMC Genomics. 2010 Dec 7; 11:694 | oa_package/67/63/PMC3014983.tar.gz |
PMC3014984 | 21245923 | Introduction
In 1906 Jacques Loeb suggested that the synthesis of life is a significant goal of biology [1] . A century later, the construction of natural genomes from off-the-shelf chemicals [2] , [3] demonstrates significant progress toward achieving this goal. Until now, however, most advances in synthetic biology have relied on collections of parts – genes, proteins, and regulatory elements – derived from sequences that already exist in nature. Must the toolkit of life be so restricted? Natural sequences comprise only a miniscule fraction of the theoretical sequence space that is possible for genes and proteins. Indeed, a collection containing just a single molecule of every one of the 20 100 possible 100-residue proteins would fill a volume larger than a mole of universes [4] . From this enormous potential for diversity, natural selection has operated over billions of years to yield a relatively small collection of sequences: Life is sustained by only ∼4,000 genes in E. coli and approximately 5-fold more in humans [5] , [6] . These considerations might lead to the supposition that genes and proteins capable of sustaining life are somehow “special.” Is this true? Or might we find functional molecular parts in a collection of artificial sequences designed “from scratch” in the laboratory?
To address these questions, we probed the ability of unevolved sequence space to encode functions that enable cell growth. We designed and constructed a collection of artificial genes encoding approximately 1.5×10 6 novel amino acid sequences. Because folding into a stable 3-dimensional structure is a prerequisite for most biological functions, we did not construct this collection of proteins from random sequences. Instead, we used the binary code strategy for protein design, shown previously to facilitate the production of large combinatorial libraries of folded proteins [7] , [8] . In brief, the binary code strategy posits that stably folded proteins can be encoded by specifying the sequence pattern of polar and nonpolar residues (the binary pattern) to coincide with the exposed and buried parts of a structure, respectively. For example, to design a helical bundle comprising α-helices with hydrophilic faces exposed to aqueous solvent and hydrophobic faces buried in the interior of a protein, the binary pattern of polar ( P ) and nonpolar ( N ) residues should be PNPPNNPPNPPNNP , consistent with the structural repeat of 3.6 residues per alpha-helical turn [9] , [10] . For the current studies, we used the binary code strategy to design and construct a library of sequences designed to fold into 102-residue 4-helix bundles. Our strategies for protein design and selecting biological function are summarized in Figure 1 . Details describing the construction and biophysical characterization of the library are presented elsewhere [11] , [12] .
Here we demonstrate that novel proteins from this binary patterned library can rescue E. coli cells that lack certain natural proteins required for cell viability. The novel proteins are substantially less active, and may function by different mechanisms than the natural proteins they replace. Nonetheless, co-expression of several novel proteins rescues a strain in which multiple natural genes were deleted simultaneously. These findings show that an organism, which would otherwise not grow, can be sustained by macromolecules devised in the laboratory. | Materials and Methods
Construction of a library of de novo sequences
The design, construction, and characterization of a library of artificial genes encoding binary-patterned de novo proteins are reported elsewhere [11] , [12] . The library of gene inserts was digested with NdeI and BsrGI, as was plasmid pCA24NMAF2. Double-digested plasmid backbone was gel purified, and the library insert and vector backbone were ligated. Ligation products were transformed into electrocompetent XL1-blue (Stratagene) or MDS42recA (Scarab Genomics, LLC). Colonies were harvested by scraping from LB plates supplemented with 34 μg ml -1 chloramphenicol (cam), and plasmid DNA was purified from the resulting cell suspension. Aliquots from this library of plasmids were used to transform the auxotrophs. The size of this library—approximately 1.5×10 6 different sequences—was estimated by plating serial dilutions of the transformations of the initial ligation products on LB-cam [12] .
Isolation of de novo sequences that rescue auxotrophs
The collection of plasmids encoding our library of novel sequences, and two negative control plasmids, pCA24N- lacZ and pCA24NMAF2 (empty vector), were used to transform 27 E. coli auxotrophs. These were chosen from Keio strains that grow on LB, but not on M9-glucose. Following electroporation, cells were outgrown at 37°C, harvested by centrifugation, washed and resuspended in chilled M9 salts, and spread on both LB supplemented with 30 μg ml −1 kanamycin (kan), 34 μg ml −1 cam; and on M9-glucose supplemented with 50 μM IPTG, 30 μg ml −1 kan, 34 μg ml −1 cam. After several days at 33°C, plates were scored for number of colonies. In addition, a 1∶100 dilution of library-transformed cells was spread onto LB-cam plates to measure transformation frequency under conditions that were non-selective for auxotroph growth.
Growth assays
Single colonies were inoculated into LB (30 μg ml −1 kan, 170 μg ml −1 cam). Following overnight growth at 37°C, 1 ml of each culture was harvested by centrifugation at 4°C. Cell pellets were washed and resuspended in 1 ml chilled M9 salts, and the resulting cell suspensions were diluted 100-fold into M9-glucose (50 μM IPTG, 30 μg ml −1 kan, 34 μg ml −1 cam). Cultures were grown at 33°C and assayed by measuring absorbance at 600 nm.
Additional methods are described in Supporting Information Text S1 . | Results and Discussion
De novo designed sequences rescue single-gene knockout strains of E. coli
Previous work in our laboratory showed that several purified proteins from our binary patterned alpha-helical libraries can bind a cofactor and perform enzyme-like functions, including peroxidase, lipase, and esterase activities [13] . These findings led us to question whether proteins from this library might also provide biological functions that enable cell growth. To address this question, we tested the ability of binary patterned proteins to rescue strains of E. coli in which a conditionally essential function had been deleted. These strains were obtained from the Keio collection, which contains all the viable single-gene knockout strains of E. coli [5] . We tested 27 auxotrophs that grow on rich media, but fail to grow on minimal media (see legend for Table 1 ). Each auxotroph was transformed with a library of synthetic genes carried on the expression vector pCA24NMAF2 [14] . Typically, we obtained 5–10 million transformants, thereby ensuring reasonably deep coverage of the library of 1.5×10 6 de novo sequences. As negative controls, each auxotroph was transformed with the empty vector or the same vector expressing beta-galactosidase. Transformed cells were spread either on LB (rich) or on M9-glucose (minimal) media supplemented with isopropyl-beta-D-thiogalactoside (IPTG) to induce expression, and the formation of colonies was monitored ( Figures 1 & 2 ).
On rich media, all transformed cells grew, regardless of whether they received the control lacZ gene or a gene encoding a novel protein. As expected, on minimal media, auxotrophs transformed with the control plasmids failed to grow. (If the negative controls grew, the strain was considered ‘leaky’ and not used for further studies.) In most cases, auxotrophs transformed with the collection of novel sequences also failed to produce colonies, even after weeks of incubation. This indicates that proteins from the collection could not rescue these strains. However, for four auxotrophs – Δ serB , Δ gltA , Δ ilvA , and Δ fes – colonies formed on minimal plates after several days of incubation ( Figures 2A & 2B ). Growth of these cells under selective conditions suggests that a novel gene carried by the plasmid complements the deletion. The rates of complementation and the times required for colony formation are summarized in Table 1 .
To confirm that the colonies observed on selective media resulted from the uptake of a novel gene and not from an adaptive mutation on the chromosome, colonies were isolated by restreaking, and plasmid DNA was purified; this DNA was then used to transform naïve cells of the same auxotroph, and the transformed cells were once again spread on both LB and minimal plates. Growth of approximately the same number of colonies on selective plates as on rich plates indicates that the observed phenotype (growth on minimal) transferred with the genotype (the plasmid carrying the novel gene). These results are shown in Figure 2C and Figure S1 . To ensure that rescue was due to a de novo protein, and not some natural sequence that might have been picked up inadvertently during plasmid constructions, we isolated the ∼300 base pair fragment encoding the novel protein, recloned it into a new vector, and showed that the new clone also rescued the deletion strain.
The binary patterned sequences that rescue the four auxotrophs are shown in Figure 3 . (Additional sequences are shown in Figure S2 .) The novel sequences are designated according to the auxotroph they rescue, followed by a number (e.g. Syn-Fes-1 is a syn thetic sequence that rescues Δ fes .) For the Δ gltA strain, we isolated one novel protein that enabled cell growth. However, in the other cases, several sequences were isolated, suggesting that deletions of these genes are relatively easy to rescue. As shown by the color-coding in Figure 3 , the sequences of the biologically active proteins conform to our binary pattern for the design of 4-helix bundles. (As is often true for sequences in combinatorial libraries, spurious mutations were observed occasionally, but these do not interfere with the overall binary pattern – see Figure S2 ). NCBI-BLAST searches indicate the sequences of the de novo proteins are unlike those of any known natural proteins [15] .
The ability of each de novo protein to sustain cell growth was measured in cultures grown in minimal M9-glucose liquid media. Growth rates were compared for deletion strains expressing a de novo protein versus the same strains expressing either the negative control (LacZ) or the corresponding natural protein on the same pCA24NMAF2 vector. As shown in Figure 4 , the novel proteins enable cell growth under conditions where auxotrophs transformed with the control plasmid fail to grow at all. Cells relying on the de novo proteins grow significantly slower than those expressing the natural protein. (Fes is an exception because overexpression of E. coli Fes is toxic [14, Genobase ORF JW0576 - http://ecoli.naist.jp/GB6/info.jsp?id=JW0576 ].) It is not surprising that the unevolved novel sequences function at substantially lower levels than natural sequences selected by billions of years of evolution. Indeed, the relatively slow rates of cell growth enabled by these first-generation de novo sequences suggest that selections for faster growth might facilitate the evolution of more active proteins. Such experiments will provide an opportunity to test whether de novo progenitor sequences might lead to novel evolutionary trajectories.
The results shown in Figures 2 and 4 demonstrate that the de novo sequences enable cell growth; however, viability per se does not indicate that the novel proteins provide the same biochemical activities as the deleted natural proteins. Therefore, we devised a series of experiments to probe the functions of the de novo proteins.
Biological functions of the de novo proteins
The natural genes deleted in these four auxotrophs encode a range of functions (9). These are summarized below and depicted graphically in Figure S3 .
serB encodes phosphoserine phosphatase, responsible for the final step in serine biosynthesis: phosphoserine → serine + phosphate.
gltA encodes citrate synthase, which catalyzes an early step in glutamate biosynthesis: oxaloacetate + acetyl-coA → citrate → cis-aconitate → isocitrate → alpha-ketoglutarate → glutamate.
ilvA encodes biosynthetic threonine deaminase, which catalyzes the first step in the production of isoleucine from threonine: threonine → 2-ketobutyrate + ammonia → → → isoleucine.
fes encodes enterobactin esterase, which cleaves the iron-bound enterobactin siderophore, thereby enabling cells to acquire iron in iron-limited environments [16] .
In principle, the novel proteins could rescue these auxotrophs either by providing the same function as the deleted enzyme, or by one of the three alternative mechanisms described below:
First, we considered the possibility that an auxotroph in a biosynthetic pathway might be rescued by a de novo sequence that produces the end product via a novel bypass pathway. Although it seems unlikely that small, unevolved proteins could catalyze novel biosynthetic pathways, we nonetheless tested this possibility. If the novel sequences that rescue Δ serB, Δ gltA or Δ ilvA encode bypass pathways for the synthesis of serine, glutamate, and isoleucine, respectively, then these de novo sequences would also rescue cells deleted for enzymes that function at other steps in the these biosynthetic pathways. (Pathways are shown in Figure S3 ). This possibility was tested by the following three experiments:
De novo sequences that rescued Δ serB were transformed into Δ serC cells, which are deleted for phosphoserine aminotransferase. This enzyme converts phosphohydroxypyruvate to phosphoserine in the step prior to that catalyzed by the serB encoded enzyme. (Because the serC gene product is also involved in the biosynthesis of pyridoxine [9] , the selective media for this experiment was supplemented with pyridoxine.)
The sequence that rescued Δ gltA was transformed into Δ icd cells. This strain is also a glutamate auxotroph because it cannot catalyze the conversion of isocitrate to alpha-ketoglutarate, the direct precursor of glutamate.
The sequences that rescued Δ ilvA were transformed into Δ ilvC and Δ ilvD cells, which are deleted for isomeroreductase and dihydroxyacid dehydratase, respectively. Both these enzymes function downstream of ilvA in the biosynthesis of isoleucine. (Because the ilvC and ilvD gene products are also involved in valine biosynthesis [9] , the selective media for these experiments was supplemented with valine.)
In all cases, the de novo sequences did not rescue E. coli cells deleted for functions at other steps in the biosynthetic pathways. This demonstrates that the novel proteins do not enable bypass pathways for the synthesis of their respective end products.
Second, we considered the possibility that our de novo sequences might rescue the auxotrophs by altering the expression or activity of an endogenous E. coli protein. To assess this possibility, we relied on an exhaustive screen by Patrick et al. , who identified natural genes whose overexpression can rescue the deletion of noncognate genes in E. coli (i.e. ‘multicopy suppressors’) [17] . By screening the complete set of overexpressed E. coli ORFs, they found the following multicopy suppressors of our four strains: Δ ilvA can be rescued by overexpression of tdcB [18] , or emrD ; Δ serB can be rescued by overexpression of gph, hisB, or ytjC ; and Δ fes can be rescued by overexpression of thiL or setB . Δ gltA was not rescued by overexpression of any E. coli genes. To probe whether our novel sequences rely on these E. coli proteins for rescue, we tested whether our proteins rescue the following double deletion strains: Δ ilvA Δ tdcB, Δ ilvA Δ emrD, Δ serB Δ gph, Δ serB Δ ytjC, and Δ fes Δ setB . Transformation with the appropriate novel sequences showed that all five of these double knockouts were rescued by all of the corresponding artificial sequences listed in Figure 3 . The ability of our novel sequences to rescue these double deletions shows that the artificial proteins do not function by altering the expression or activity of E. coli proteins known previously to enable rescue.
The double mutant Δ fes Δ thiL cannot be constructed because deletion of thiL is lethal. The Δ serB Δ hisB strain was constructed, but since Δ hisB is itself an auxotroph, our novel sequences did not rescue this strain (successful rescue would require the unlikely occurrence of a single de novo protein replacing the activities of two conditionally essential genes.) Therefore, dependence on thiL and hisB could not be assessed. In addition, it remains possible that some other E. coli protein, which escaped selection by Patrick et al ., could act as a partner in the rescue mediated by our de novo sequences. With the exception of these caveats, our results demonstrate that the known multicopy suppressors are not required for the biological activities of our de novo proteins.
Third, we considered the possibility that the de novo proteins might rescue the deletion strains by a mechanism that does not depend on the specific sequences ( Figure 3 ), but instead involves global alterations in metabolism that are induced by the mere expression of foreign genes. For example, although our proteins were designed to fold into alpha-helical bundles, we considered the possibility that sequences isolated by our selections might be unfolded, and thereby induce a cellular stress response. To assess folding, we purified several proteins and measured their circular dichroism spectra. These spectra (shown in Figure S4 ) demonstrate the structures are predominantly alpha-helical, and similar to the spectra of designed 4-helix bundles solved previously by NMR [19] , [20] . Thus, rescue is not caused by unfolded sequences inducing a stress response.
We also note that if a generic stress response were responsible for rescue, one would expect all the Syn proteins to rescue all the deletions. This is not the case. Syn-Fes does not rescue Δ gltA , Syn-SerB does not rescue Δ ilvA , and so on. Thus, specific de novo sequences mediate the rescue of specific chromosomal deletions.
To demonstrate explicitly that rescue depends on a specific sequence, rather than a generic cellular response to the expression of foreign genes, we mutated one of the de novo proteins. The design of this mutant was based on an analysis of the common features among the de novo sequences that rescued Δ ilvA . Alignment of the Syn-IlvA sequences revealed two conserved polar residues: Glu36 and Lys42. We constructed the Lys42»Ala mutation in Syn-IlvA-1, and found that although this protein was expressed at the same level as the parental sequence, it fails to rescue Δ ilvA. (These results are shown in Figure S5 ) Thus, rescue is not simply due to expression of an exogenous gene; it is mediated by sequence-specific features of Syn-IlvA-1.
Fes differs from the other deletions because it is not involved in a biosynthetic pathway. Fes functions in iron acquisition. E. coli secretes the enterobactin siderophore (MW, 670 Da) into the media, where it binds iron, and is transported back into the cell. Because the affinity of enterobactin for iron is extremely high (K D ≈10 −35 M), release of the metal requires degradation of the siderophore. This is catalyzed by the Fes protein, enterobactin esterase [16] . The impact of the de novo syn-fes sequences on iron acquisition is dramatic: Elemental analysis shows that cells expressing the Syn-Fes proteins accumulate 6- to 10-fold more iron than control cells ( Figure S6 ). In principle, the de novo proteins could rescue Δ fes either by functioning as an esterase or by some alternative mechanism. For example, the de novo proteins could enable iron acquisition by being transported into the media, binding iron with high affinity, being re-transported back into the cell, and then releasing the tightly bound iron intracellularly. Although possible, this alternative mechanism seems unlikely for novel proteins that occur so frequently in a semi-random library.
Irrespective of whether the de novo proteins rescue auxotrophs by providing the same function as the deleted enzyme or by some alternative mechanism, we expected the artificial sequences – which were neither selected by evolution ( in vivo or in vitro ), nor explicitly designed for enzymatic activity – to have far lower levels of activity than naturally evolved sequences. To estimate this level of activity in vivo , we compared the growth rates of deletion strains expressing a high level (400X) of Syn-IlvA-1 to the same strain expressing a low level (1X) of the natural protein, encoded by IlvA. These experiments showed that even when the de novo protein is expressed at ∼400-fold higher levels than the natural protein, cells grow much more slowly. (Growth rates are shown in Figure S7 .) As expected, the de novo protein exhibits a very low level of biological activity.
Activity levels that are barely sufficient to sustain slow cell growth may not be detectable in vitro . Indeed, others have reported that although overexpression of one E. coli protein can sometimes rescue the deletion of another protein in vivo , these ‘moonlighting’ activities can be so low that they cannot be detected in vitro [21] , [22] . Despite these concerns, we attempted to assay biochemical activities in vitro using both cell lysates and purified proteins. Lysates are easy to prepare and are more likely to contain molecular partners (proteins, cofactors, or metals) present in vivo . The disadvantage of lysates, however, is that background activities may obscure the low-level activities of the de novo proteins. We assayed for phosphoserine phosphatase, enterobactin esterase, threonine deaminase, and citrate synthase activities in lysates from cells expressing the respective de novo proteins. In several cases, activity was observed, however, lysates from control cells also showed low levels of activity. This was not surprising, particularly for the Fes and SerB activities, since previous work in our laboratory showed that E. coli lysates contain nonspecific esterases and phosphatases [13] .
We also purified several of the de novo proteins. (To avoid contamination by the natural enzyme, purifications were from strains deleted for the natural gene.) We tested these purified proteins for the enzymatic activities deleted in the respective auxotrophs, but were unable to detect activity that was reproducibly above the controls. There are several reasons why such experiments might not demonstrate activity: (i) As noted above, the novel proteins have extremely low levels of activity. (ii) The de novo proteins may require cofactors. We do not know which cofactors might be required, as the novel proteins may use different cofactors than the natural enzymes. (iii) The novel proteins may function by partnering with an E. coli protein. For example, the specificity of an endogenous hydrolase might be altered by binding one of our helical bundles. If this were the case, activity would not be observed in preparations containing only the purified de novo protein. (iv) The Syn protein may function by a different mechanism than the deleted protein, and novel activities would not be detected in experiments designed to assay the natural enzyme.
While we do not yet know the precise mechanisms by which the novel proteins rescue these four deletions in E. coli, we have ruled out several alternative mechanisms including (i) bypass pathways, (ii) activation of known endogenous suppressors, and (iii) induction of a generic stress response (see above). The slow growth rates of cells relying on the novel proteins for survival indicate that regardless of the actual mechanism, the de novo proteins have very low levels of activity — as expected for sequences that were neither designed nor evolved for function. Irrespective of whether the de novo proteins catalyze the same biochemical reaction as the deleted natural protein, or function by some alternative mechanism (and whether they act directly or through interactions with endogenous proteins), elucidating the molecular basis of auxotroph rescue will enhance our understanding of the minimal functions necessary for cell growth.
Replacement of 0.1% of the E. coli genome
We showed above that artificial genes enable cell viability. Next, to assess whether portions of a genome can be replaced by genetic information devised in the laboratory, we probed the ability of several novel sequences to rescue the deletion of several genes simultaneously. A quadruple knockout strain – Δ fes Δ gltA Δ ilvA Δ serB – was constructed and designated QUAD. This strain has severe growth defects: While the parent strain forms colonies after overnight incubation on either nutrient-rich or nutrient-poor media, QUAD requires almost two days to form colonies on rich media, and as expected, does not form colonies on minimal media. To assess the ability of artificial genes to substitute simultaneously for the deleted functions, we transformed QUAD with derivatives of the pQLinkN vector [23] co-expressing four artificial genes shown above to rescue the four single-gene knockouts. The sequences used for this experiment were syn-fes-2 , syn-gltA-1, syn-ilvA-1 , and syn-serB-2 ( Figure 3 ). Transformed cells were plated on minimal media. As a control, we transformed QUAD with empty vector; no colonies were observed for the control. However, as shown in Figure 5 , transformation with the vector expressing all four artificial genes produced colonies. As discussed above, the individual Syn proteins are considerably less active than the deleted natural proteins. Therefore, it is not surprising that the simultaneous rescue of four deletions by four artificial proteins produced small colonies that required 13 days to appear. Nonetheless, the rescue of QUAD demonstrates that novel sequences bearing no similarity to naturally evolved proteins can compensate for deletion of a portion of the E. coli chromosome.
The E. coli genome contains approximately 4,000 genes [5] . Of these, only 290 are essential for survival under all conditions [24] . An additional 107 are required for growth in the nutrient-poor media provided by M9 salts with glucose [17] . Thus, only ∼400 genes are essential for growth in minimal media. Our results demonstrate that at least 1% of the E. coli genome that is absolutely required for growth under minimal conditions – and 0.1% of the entire E. coli genome – can be replaced by artificial genes designed de novo.
Conclusion
We have demonstrated that sequences designed de novo can provide functions necessary to sustain the growth of living cells. It should be emphasized that these macromolecules were isolated with relatively high frequency from a collection of sequences that were designed to adopt a stable globular fold [7] , [8] , [19] , [20] , [25] , but were not explicitly designed for function.
Both in terms of linear sequence and 3-dimensional structure, the novel proteins differ substantially from the natural proteins they replace: The binary patterned sequences in Figure 3 show no significant sequence similarity to any known proteins. At the structural level, the 102-residue 4-helix bundles specified by our design are much smaller and simpler than the structures of the four natural proteins deleted in the auxotrophs. (Structures are shown in Figure S8 ). Thus, although natural proteomes include diverse structures and topologies, our studies show that in some cases, cell growth can be sustained by much simpler structures.
The field of synthetic biology — and the related field of astrobiology — aim to go beyond questions about “what already exists” and attempt to probe “what might be possible?” The results described here suggest that the toolkit for synthetic biology need not be limited to genes and proteins that already exist in nature. We tested 27 deletion strains using a library of 1.5 million sequences designed to fold into 4-helix bundles. From these, we found 18 artificial sequences ( Figure 3 and Figure S2 ) that compensate for the deletion of 4 different genes, and provide biological functions enabling cell growth. Extension of this work to other deletion strains, using larger libraries, and designs that encode a greater diversity of protein structures can be expected to increase substantially the number of life-supporting functions that can be provided by genes designed de novo . The results described here showing that sequences designed in the laboratory can replace portions of the E. coli genome can therefore be viewed as an initial step toward the construction of artificial genomes. | Results and Discussion
De novo designed sequences rescue single-gene knockout strains of E. coli
Previous work in our laboratory showed that several purified proteins from our binary patterned alpha-helical libraries can bind a cofactor and perform enzyme-like functions, including peroxidase, lipase, and esterase activities [13] . These findings led us to question whether proteins from this library might also provide biological functions that enable cell growth. To address this question, we tested the ability of binary patterned proteins to rescue strains of E. coli in which a conditionally essential function had been deleted. These strains were obtained from the Keio collection, which contains all the viable single-gene knockout strains of E. coli [5] . We tested 27 auxotrophs that grow on rich media, but fail to grow on minimal media (see legend for Table 1 ). Each auxotroph was transformed with a library of synthetic genes carried on the expression vector pCA24NMAF2 [14] . Typically, we obtained 5–10 million transformants, thereby ensuring reasonably deep coverage of the library of 1.5×10 6 de novo sequences. As negative controls, each auxotroph was transformed with the empty vector or the same vector expressing beta-galactosidase. Transformed cells were spread either on LB (rich) or on M9-glucose (minimal) media supplemented with isopropyl-beta-D-thiogalactoside (IPTG) to induce expression, and the formation of colonies was monitored ( Figures 1 & 2 ).
On rich media, all transformed cells grew, regardless of whether they received the control lacZ gene or a gene encoding a novel protein. As expected, on minimal media, auxotrophs transformed with the control plasmids failed to grow. (If the negative controls grew, the strain was considered ‘leaky’ and not used for further studies.) In most cases, auxotrophs transformed with the collection of novel sequences also failed to produce colonies, even after weeks of incubation. This indicates that proteins from the collection could not rescue these strains. However, for four auxotrophs – Δ serB , Δ gltA , Δ ilvA , and Δ fes – colonies formed on minimal plates after several days of incubation ( Figures 2A & 2B ). Growth of these cells under selective conditions suggests that a novel gene carried by the plasmid complements the deletion. The rates of complementation and the times required for colony formation are summarized in Table 1 .
To confirm that the colonies observed on selective media resulted from the uptake of a novel gene and not from an adaptive mutation on the chromosome, colonies were isolated by restreaking, and plasmid DNA was purified; this DNA was then used to transform naïve cells of the same auxotroph, and the transformed cells were once again spread on both LB and minimal plates. Growth of approximately the same number of colonies on selective plates as on rich plates indicates that the observed phenotype (growth on minimal) transferred with the genotype (the plasmid carrying the novel gene). These results are shown in Figure 2C and Figure S1 . To ensure that rescue was due to a de novo protein, and not some natural sequence that might have been picked up inadvertently during plasmid constructions, we isolated the ∼300 base pair fragment encoding the novel protein, recloned it into a new vector, and showed that the new clone also rescued the deletion strain.
The binary patterned sequences that rescue the four auxotrophs are shown in Figure 3 . (Additional sequences are shown in Figure S2 .) The novel sequences are designated according to the auxotroph they rescue, followed by a number (e.g. Syn-Fes-1 is a syn thetic sequence that rescues Δ fes .) For the Δ gltA strain, we isolated one novel protein that enabled cell growth. However, in the other cases, several sequences were isolated, suggesting that deletions of these genes are relatively easy to rescue. As shown by the color-coding in Figure 3 , the sequences of the biologically active proteins conform to our binary pattern for the design of 4-helix bundles. (As is often true for sequences in combinatorial libraries, spurious mutations were observed occasionally, but these do not interfere with the overall binary pattern – see Figure S2 ). NCBI-BLAST searches indicate the sequences of the de novo proteins are unlike those of any known natural proteins [15] .
The ability of each de novo protein to sustain cell growth was measured in cultures grown in minimal M9-glucose liquid media. Growth rates were compared for deletion strains expressing a de novo protein versus the same strains expressing either the negative control (LacZ) or the corresponding natural protein on the same pCA24NMAF2 vector. As shown in Figure 4 , the novel proteins enable cell growth under conditions where auxotrophs transformed with the control plasmid fail to grow at all. Cells relying on the de novo proteins grow significantly slower than those expressing the natural protein. (Fes is an exception because overexpression of E. coli Fes is toxic [14, Genobase ORF JW0576 - http://ecoli.naist.jp/GB6/info.jsp?id=JW0576 ].) It is not surprising that the unevolved novel sequences function at substantially lower levels than natural sequences selected by billions of years of evolution. Indeed, the relatively slow rates of cell growth enabled by these first-generation de novo sequences suggest that selections for faster growth might facilitate the evolution of more active proteins. Such experiments will provide an opportunity to test whether de novo progenitor sequences might lead to novel evolutionary trajectories.
The results shown in Figures 2 and 4 demonstrate that the de novo sequences enable cell growth; however, viability per se does not indicate that the novel proteins provide the same biochemical activities as the deleted natural proteins. Therefore, we devised a series of experiments to probe the functions of the de novo proteins.
Biological functions of the de novo proteins
The natural genes deleted in these four auxotrophs encode a range of functions (9). These are summarized below and depicted graphically in Figure S3 .
serB encodes phosphoserine phosphatase, responsible for the final step in serine biosynthesis: phosphoserine → serine + phosphate.
gltA encodes citrate synthase, which catalyzes an early step in glutamate biosynthesis: oxaloacetate + acetyl-coA → citrate → cis-aconitate → isocitrate → alpha-ketoglutarate → glutamate.
ilvA encodes biosynthetic threonine deaminase, which catalyzes the first step in the production of isoleucine from threonine: threonine → 2-ketobutyrate + ammonia → → → isoleucine.
fes encodes enterobactin esterase, which cleaves the iron-bound enterobactin siderophore, thereby enabling cells to acquire iron in iron-limited environments [16] .
In principle, the novel proteins could rescue these auxotrophs either by providing the same function as the deleted enzyme, or by one of the three alternative mechanisms described below:
First, we considered the possibility that an auxotroph in a biosynthetic pathway might be rescued by a de novo sequence that produces the end product via a novel bypass pathway. Although it seems unlikely that small, unevolved proteins could catalyze novel biosynthetic pathways, we nonetheless tested this possibility. If the novel sequences that rescue Δ serB, Δ gltA or Δ ilvA encode bypass pathways for the synthesis of serine, glutamate, and isoleucine, respectively, then these de novo sequences would also rescue cells deleted for enzymes that function at other steps in the these biosynthetic pathways. (Pathways are shown in Figure S3 ). This possibility was tested by the following three experiments:
De novo sequences that rescued Δ serB were transformed into Δ serC cells, which are deleted for phosphoserine aminotransferase. This enzyme converts phosphohydroxypyruvate to phosphoserine in the step prior to that catalyzed by the serB encoded enzyme. (Because the serC gene product is also involved in the biosynthesis of pyridoxine [9] , the selective media for this experiment was supplemented with pyridoxine.)
The sequence that rescued Δ gltA was transformed into Δ icd cells. This strain is also a glutamate auxotroph because it cannot catalyze the conversion of isocitrate to alpha-ketoglutarate, the direct precursor of glutamate.
The sequences that rescued Δ ilvA were transformed into Δ ilvC and Δ ilvD cells, which are deleted for isomeroreductase and dihydroxyacid dehydratase, respectively. Both these enzymes function downstream of ilvA in the biosynthesis of isoleucine. (Because the ilvC and ilvD gene products are also involved in valine biosynthesis [9] , the selective media for these experiments was supplemented with valine.)
In all cases, the de novo sequences did not rescue E. coli cells deleted for functions at other steps in the biosynthetic pathways. This demonstrates that the novel proteins do not enable bypass pathways for the synthesis of their respective end products.
Second, we considered the possibility that our de novo sequences might rescue the auxotrophs by altering the expression or activity of an endogenous E. coli protein. To assess this possibility, we relied on an exhaustive screen by Patrick et al. , who identified natural genes whose overexpression can rescue the deletion of noncognate genes in E. coli (i.e. ‘multicopy suppressors’) [17] . By screening the complete set of overexpressed E. coli ORFs, they found the following multicopy suppressors of our four strains: Δ ilvA can be rescued by overexpression of tdcB [18] , or emrD ; Δ serB can be rescued by overexpression of gph, hisB, or ytjC ; and Δ fes can be rescued by overexpression of thiL or setB . Δ gltA was not rescued by overexpression of any E. coli genes. To probe whether our novel sequences rely on these E. coli proteins for rescue, we tested whether our proteins rescue the following double deletion strains: Δ ilvA Δ tdcB, Δ ilvA Δ emrD, Δ serB Δ gph, Δ serB Δ ytjC, and Δ fes Δ setB . Transformation with the appropriate novel sequences showed that all five of these double knockouts were rescued by all of the corresponding artificial sequences listed in Figure 3 . The ability of our novel sequences to rescue these double deletions shows that the artificial proteins do not function by altering the expression or activity of E. coli proteins known previously to enable rescue.
The double mutant Δ fes Δ thiL cannot be constructed because deletion of thiL is lethal. The Δ serB Δ hisB strain was constructed, but since Δ hisB is itself an auxotroph, our novel sequences did not rescue this strain (successful rescue would require the unlikely occurrence of a single de novo protein replacing the activities of two conditionally essential genes.) Therefore, dependence on thiL and hisB could not be assessed. In addition, it remains possible that some other E. coli protein, which escaped selection by Patrick et al ., could act as a partner in the rescue mediated by our de novo sequences. With the exception of these caveats, our results demonstrate that the known multicopy suppressors are not required for the biological activities of our de novo proteins.
Third, we considered the possibility that the de novo proteins might rescue the deletion strains by a mechanism that does not depend on the specific sequences ( Figure 3 ), but instead involves global alterations in metabolism that are induced by the mere expression of foreign genes. For example, although our proteins were designed to fold into alpha-helical bundles, we considered the possibility that sequences isolated by our selections might be unfolded, and thereby induce a cellular stress response. To assess folding, we purified several proteins and measured their circular dichroism spectra. These spectra (shown in Figure S4 ) demonstrate the structures are predominantly alpha-helical, and similar to the spectra of designed 4-helix bundles solved previously by NMR [19] , [20] . Thus, rescue is not caused by unfolded sequences inducing a stress response.
We also note that if a generic stress response were responsible for rescue, one would expect all the Syn proteins to rescue all the deletions. This is not the case. Syn-Fes does not rescue Δ gltA , Syn-SerB does not rescue Δ ilvA , and so on. Thus, specific de novo sequences mediate the rescue of specific chromosomal deletions.
To demonstrate explicitly that rescue depends on a specific sequence, rather than a generic cellular response to the expression of foreign genes, we mutated one of the de novo proteins. The design of this mutant was based on an analysis of the common features among the de novo sequences that rescued Δ ilvA . Alignment of the Syn-IlvA sequences revealed two conserved polar residues: Glu36 and Lys42. We constructed the Lys42»Ala mutation in Syn-IlvA-1, and found that although this protein was expressed at the same level as the parental sequence, it fails to rescue Δ ilvA. (These results are shown in Figure S5 ) Thus, rescue is not simply due to expression of an exogenous gene; it is mediated by sequence-specific features of Syn-IlvA-1.
Fes differs from the other deletions because it is not involved in a biosynthetic pathway. Fes functions in iron acquisition. E. coli secretes the enterobactin siderophore (MW, 670 Da) into the media, where it binds iron, and is transported back into the cell. Because the affinity of enterobactin for iron is extremely high (K D ≈10 −35 M), release of the metal requires degradation of the siderophore. This is catalyzed by the Fes protein, enterobactin esterase [16] . The impact of the de novo syn-fes sequences on iron acquisition is dramatic: Elemental analysis shows that cells expressing the Syn-Fes proteins accumulate 6- to 10-fold more iron than control cells ( Figure S6 ). In principle, the de novo proteins could rescue Δ fes either by functioning as an esterase or by some alternative mechanism. For example, the de novo proteins could enable iron acquisition by being transported into the media, binding iron with high affinity, being re-transported back into the cell, and then releasing the tightly bound iron intracellularly. Although possible, this alternative mechanism seems unlikely for novel proteins that occur so frequently in a semi-random library.
Irrespective of whether the de novo proteins rescue auxotrophs by providing the same function as the deleted enzyme or by some alternative mechanism, we expected the artificial sequences – which were neither selected by evolution ( in vivo or in vitro ), nor explicitly designed for enzymatic activity – to have far lower levels of activity than naturally evolved sequences. To estimate this level of activity in vivo , we compared the growth rates of deletion strains expressing a high level (400X) of Syn-IlvA-1 to the same strain expressing a low level (1X) of the natural protein, encoded by IlvA. These experiments showed that even when the de novo protein is expressed at ∼400-fold higher levels than the natural protein, cells grow much more slowly. (Growth rates are shown in Figure S7 .) As expected, the de novo protein exhibits a very low level of biological activity.
Activity levels that are barely sufficient to sustain slow cell growth may not be detectable in vitro . Indeed, others have reported that although overexpression of one E. coli protein can sometimes rescue the deletion of another protein in vivo , these ‘moonlighting’ activities can be so low that they cannot be detected in vitro [21] , [22] . Despite these concerns, we attempted to assay biochemical activities in vitro using both cell lysates and purified proteins. Lysates are easy to prepare and are more likely to contain molecular partners (proteins, cofactors, or metals) present in vivo . The disadvantage of lysates, however, is that background activities may obscure the low-level activities of the de novo proteins. We assayed for phosphoserine phosphatase, enterobactin esterase, threonine deaminase, and citrate synthase activities in lysates from cells expressing the respective de novo proteins. In several cases, activity was observed, however, lysates from control cells also showed low levels of activity. This was not surprising, particularly for the Fes and SerB activities, since previous work in our laboratory showed that E. coli lysates contain nonspecific esterases and phosphatases [13] .
We also purified several of the de novo proteins. (To avoid contamination by the natural enzyme, purifications were from strains deleted for the natural gene.) We tested these purified proteins for the enzymatic activities deleted in the respective auxotrophs, but were unable to detect activity that was reproducibly above the controls. There are several reasons why such experiments might not demonstrate activity: (i) As noted above, the novel proteins have extremely low levels of activity. (ii) The de novo proteins may require cofactors. We do not know which cofactors might be required, as the novel proteins may use different cofactors than the natural enzymes. (iii) The novel proteins may function by partnering with an E. coli protein. For example, the specificity of an endogenous hydrolase might be altered by binding one of our helical bundles. If this were the case, activity would not be observed in preparations containing only the purified de novo protein. (iv) The Syn protein may function by a different mechanism than the deleted protein, and novel activities would not be detected in experiments designed to assay the natural enzyme.
While we do not yet know the precise mechanisms by which the novel proteins rescue these four deletions in E. coli, we have ruled out several alternative mechanisms including (i) bypass pathways, (ii) activation of known endogenous suppressors, and (iii) induction of a generic stress response (see above). The slow growth rates of cells relying on the novel proteins for survival indicate that regardless of the actual mechanism, the de novo proteins have very low levels of activity — as expected for sequences that were neither designed nor evolved for function. Irrespective of whether the de novo proteins catalyze the same biochemical reaction as the deleted natural protein, or function by some alternative mechanism (and whether they act directly or through interactions with endogenous proteins), elucidating the molecular basis of auxotroph rescue will enhance our understanding of the minimal functions necessary for cell growth.
Replacement of 0.1% of the E. coli genome
We showed above that artificial genes enable cell viability. Next, to assess whether portions of a genome can be replaced by genetic information devised in the laboratory, we probed the ability of several novel sequences to rescue the deletion of several genes simultaneously. A quadruple knockout strain – Δ fes Δ gltA Δ ilvA Δ serB – was constructed and designated QUAD. This strain has severe growth defects: While the parent strain forms colonies after overnight incubation on either nutrient-rich or nutrient-poor media, QUAD requires almost two days to form colonies on rich media, and as expected, does not form colonies on minimal media. To assess the ability of artificial genes to substitute simultaneously for the deleted functions, we transformed QUAD with derivatives of the pQLinkN vector [23] co-expressing four artificial genes shown above to rescue the four single-gene knockouts. The sequences used for this experiment were syn-fes-2 , syn-gltA-1, syn-ilvA-1 , and syn-serB-2 ( Figure 3 ). Transformed cells were plated on minimal media. As a control, we transformed QUAD with empty vector; no colonies were observed for the control. However, as shown in Figure 5 , transformation with the vector expressing all four artificial genes produced colonies. As discussed above, the individual Syn proteins are considerably less active than the deleted natural proteins. Therefore, it is not surprising that the simultaneous rescue of four deletions by four artificial proteins produced small colonies that required 13 days to appear. Nonetheless, the rescue of QUAD demonstrates that novel sequences bearing no similarity to naturally evolved proteins can compensate for deletion of a portion of the E. coli chromosome.
The E. coli genome contains approximately 4,000 genes [5] . Of these, only 290 are essential for survival under all conditions [24] . An additional 107 are required for growth in the nutrient-poor media provided by M9 salts with glucose [17] . Thus, only ∼400 genes are essential for growth in minimal media. Our results demonstrate that at least 1% of the E. coli genome that is absolutely required for growth under minimal conditions – and 0.1% of the entire E. coli genome – can be replaced by artificial genes designed de novo.
Conclusion
We have demonstrated that sequences designed de novo can provide functions necessary to sustain the growth of living cells. It should be emphasized that these macromolecules were isolated with relatively high frequency from a collection of sequences that were designed to adopt a stable globular fold [7] , [8] , [19] , [20] , [25] , but were not explicitly designed for function.
Both in terms of linear sequence and 3-dimensional structure, the novel proteins differ substantially from the natural proteins they replace: The binary patterned sequences in Figure 3 show no significant sequence similarity to any known proteins. At the structural level, the 102-residue 4-helix bundles specified by our design are much smaller and simpler than the structures of the four natural proteins deleted in the auxotrophs. (Structures are shown in Figure S8 ). Thus, although natural proteomes include diverse structures and topologies, our studies show that in some cases, cell growth can be sustained by much simpler structures.
The field of synthetic biology — and the related field of astrobiology — aim to go beyond questions about “what already exists” and attempt to probe “what might be possible?” The results described here suggest that the toolkit for synthetic biology need not be limited to genes and proteins that already exist in nature. We tested 27 deletion strains using a library of 1.5 million sequences designed to fold into 4-helix bundles. From these, we found 18 artificial sequences ( Figure 3 and Figure S2 ) that compensate for the deletion of 4 different genes, and provide biological functions enabling cell growth. Extension of this work to other deletion strains, using larger libraries, and designs that encode a greater diversity of protein structures can be expected to increase substantially the number of life-supporting functions that can be provided by genes designed de novo . The results described here showing that sequences designed in the laboratory can replace portions of the E. coli genome can therefore be viewed as an initial step toward the construction of artificial genomes. | Conclusion
We have demonstrated that sequences designed de novo can provide functions necessary to sustain the growth of living cells. It should be emphasized that these macromolecules were isolated with relatively high frequency from a collection of sequences that were designed to adopt a stable globular fold [7] , [8] , [19] , [20] , [25] , but were not explicitly designed for function.
Both in terms of linear sequence and 3-dimensional structure, the novel proteins differ substantially from the natural proteins they replace: The binary patterned sequences in Figure 3 show no significant sequence similarity to any known proteins. At the structural level, the 102-residue 4-helix bundles specified by our design are much smaller and simpler than the structures of the four natural proteins deleted in the auxotrophs. (Structures are shown in Figure S8 ). Thus, although natural proteomes include diverse structures and topologies, our studies show that in some cases, cell growth can be sustained by much simpler structures.
The field of synthetic biology — and the related field of astrobiology — aim to go beyond questions about “what already exists” and attempt to probe “what might be possible?” The results described here suggest that the toolkit for synthetic biology need not be limited to genes and proteins that already exist in nature. We tested 27 deletion strains using a library of 1.5 million sequences designed to fold into 4-helix bundles. From these, we found 18 artificial sequences ( Figure 3 and Figure S2 ) that compensate for the deletion of 4 different genes, and provide biological functions enabling cell growth. Extension of this work to other deletion strains, using larger libraries, and designs that encode a greater diversity of protein structures can be expected to increase substantially the number of life-supporting functions that can be provided by genes designed de novo . The results described here showing that sequences designed in the laboratory can replace portions of the E. coli genome can therefore be viewed as an initial step toward the construction of artificial genomes. | Conceived and designed the experiments: MAF KLM LHB SRV MHH. Performed the experiments: MAF KLM LHB SRV. Analyzed the data: MAF KLM LHB SRV MHH. Contributed reagents/materials/analysis tools: MAF KLM LHB SRV MHH. Wrote the paper: MAF KLM LHB SRV MHH.
Current address: Energy Biosciences Institute, University of California, Berkeley, California, United States of America
Current address: Department of Biology, Massachusetts Institute of Technology, Cambridge, Massachusetts, United States of America
Current address: Departments of Anatomy and Neurobiology, Molecular and Cellular Biochemistry, University of Kentucky, Lexington, Kentucky, United States of America
Current address: Columbia University Medical Center, New York, New York, United States of America
A central challenge of synthetic biology is to enable the growth of living systems using parts that are not derived from nature, but designed and synthesized in the laboratory. As an initial step toward achieving this goal, we probed the ability of a collection of >10 6 de novo designed proteins to provide biological functions necessary to sustain cell growth. Our collection of proteins was drawn from a combinatorial library of 102-residue sequences, designed by binary patterning of polar and nonpolar residues to fold into stable 4-helix bundles. We probed the capacity of proteins from this library to function in vivo by testing their abilities to rescue 27 different knockout strains of Escherichia coli, each deleted for a conditionally essential gene. Four different strains – Δ serB , Δ gltA , Δ ilvA , and Δ fes – were rescued by specific sequences from our library. Further experiments demonstrated that a strain simultaneously deleted for all four genes was rescued by co-expression of four novel sequences. Thus, cells deleted for ∼0.1% of the E. coli genome (and ∼1% of the genes required for growth under nutrient-poor conditions) can be sustained by sequences designed de novo. | Supporting Information | We thank Wayne Patrick and Ichiro Matsumura for providing materials and advice at the outset of the project; Natividad Ruiz and Tom Silhavy for providing access to the Keio collection, and for guidance about creating the multiple knockouts; Joshua Lees and Fred Hughson for advice on the pQLinkN vector; Josh Rabinowitz and John Groves for advice; and Izhack Cherny for advice and encouragement. | CC BY | no | 2022-01-13 08:14:21 | PLoS One. 2011 Jan 4; 6(1):e15364 | oa_package/71/fb/PMC3014984.tar.gz |
PMC3014985 | 21245924 | Introduction
The genus Phlebovirus belongs to the Bunyaviridae family that includes four others genera, namely Hantavirus , Nairovirus , Orthobunyavirus , and Tospovirus . Phleboviruses have a worldwide distribution and are associated with a wide variety of arthropods (sandflies, mosquitoes, ticks). Rift Valley fever virus (RVFV), the prototype species of the genus, is endemic in Africa, where it is zoonotic, infecting mainly sheep, and causing severe disease with high rates death through abortion in sheep. During these outbreaks, RVFV can pass to human either directly via abortion products or via mosquito transmission, leading occasionally to potentially fatal meningoencephalitis and/or haemorrhagic fevers. Recent outbreaks occurred in the horn of Africa [1] – [2] and the virus also spread into the Arabic peninsula [3] . Outbreaks are directly correlated to rainfalls in these regions and thus, climate and vegetation data may be used to predict areas and periods at risk [4] . In the Mediterranean basin phleboviruses other than RVFV are well established, and seroprevalence can in some regions reach 20% in man [5] . Phleboviruses are mainly represented by Sandfly Fever Sicilian virus (SFSV), Sandfly Fever Naples virus (SFNV), Toscana virus (TOSV) and viruses more or less distantly related to SFNV and SFSV [6] , [7] , [8] . These phleboviruses are transmitted by phlebotomine flies in regions where the latter circulate. SFN- and SFS-like viruses can cause mild febrile illnesses (sometimes paucisymptomatic) which are likely to be largely underestimated due to the lack of diagnosis, little awareness among health professionals. Toscana virus is in the top three viral etiologies of aseptic meningitis in Italy, Spain and France [9] – [10] . In this respect, phleboviruses are representative of Bunyaviridae . Among the viral world, this family was the major responsible for emerging infectious diseases (EIDs) events between 1940 and 2004 ( figure 1 , data extracted from [11] ), even more represented than the Flaviviridae . Therefore, efforts should be brought not only in diagnosis but also in the understanding of the viral cycle to emphasize antiviral research against these viruses.
Phleboviruses, and in general members of the family Bunyaviridae , are enveloped spherical viruses of icosahedral geometry [12] – [13] with an 80–120 nm long diameter. Their genome is made of three segments [14] . Among these segments, two are single negative stranded (M and L) whereas the third one (S) can adopt an ambisense strategy of expression [15] . The L fragment is encoding the L protein that is carrying the RNA dependent RNA polymerase activity involved in the primary and secondary transcription generating mRNA and replicative intermediates, respectively. In complement, L is also presumably carrying the cap snatching activity required in the viral mRNA capping [16] . The M segment is coding for the glycoprotein precursor that is cleaved by host proteases in two structural domains G C and G N [17] . The S fragment can code for the non structural protein NSs as well as the nucleoprotein NP protein. The nucleoprotein is a 245 amino acids protein (for Rift Valley fever virus) that can bind to genomic RNA and replicative intermediates to form ribonucleoproteic complexes (RNP) of circular appearance [18] – [19] . The NP oligomerization and its RNA binding properties have not been extensively described until recently. NPs can dimerize with the involvement of their N-terminal domain and this NP-NP interaction does not depend on the presence of RNA [20] – [21] . To date, no physiological multimeric stage over dimerization has been clearly identified. Biophysical and structural studies of the NP alone would therefore provide insights into RNP formation process. For such studies, it would be beneficial to purify homogeneous preparations of monomeric nucleoprotein, or at least, NP assemblies of tractable, defined multimerization status.
Many previous studies reported the expression and purification of nucleoproteins for various applications. The NP protein of RVFV can be expressed in insect cells using the recombinant baculovirus technology, but the protein forms a high molecular weight RNP complex, as shown by size exclusion chromatography [22] . The NP proteins have already been produced in E. coli for ELISA experiments. The proteins were purified under denaturing conditions [23] , with a large N-terminal non cleavable tag [24] , or the purification procedure was stopped after the first affinity purification step [25] . Only recently, RVFV NP was purified by refolding the recombinant protein while the natively produced protein was considered as heterogeneous. The refolded protein lead to the first crystal structure determination of the phlebovirus NP [26] . In solution, the refolded protein behaved as a monomer and the NP crystallized as a dimer that was thought to occur naturally. Nevertheless, in this study, the role and the presence of the dimer in the NP oligomeric form observed by electron microscopy remained unclear.
In order to further understand the structural properties of phlebovirus NP, it is necessary to set up a process that would lead to the production of protein oligomers pure and homogeneous in size, as previously performed for the rabies NP [27] . Additionally, the process would include a tag removal to improve crystallization. To that aim, and based on the existing results, we decided to evaluate two strategies relying on bacterial recombinant expression for the production of several phlebovirus NP protein suitable for structural studies. These strategies already met success in large scale structural genomics projects. Firstly, the screening of N-terminal tags can drastically improve soluble expression [28] – [29] . Secondly, when proteins are reluctant to soluble expression, they can be expressed as inclusion bodies (IB) before being refolded in non-denaturing conditions [30] – [31] . In this study, both strategies will be performed in parallel, even if proteins can be expressed in the soluble fraction, in order to provide comparative data suitable to design optimized production protocols for NP proteins. In addition to the production of NPs suitable for structural studies, these data may highlight trends in the larger field of recombinant protein expression. | Materials and Methods
Cloning and protein expression
cDNA corresponding to the three Nucleoproteins (NP) of Rift Valley fever virus (RVFV, strain Smithburn DQ380157.1), Sandfly Fever Sicilian virus (SFSV, strain J04418.1) and Toscana virus (TOSV, strain AR2005) were amplified using two Polymerase Chain Reactions (PCR). A first amplification was performed using i) a forward primer carrying the coding sequence of the Tobacco Etch Virus (TEV) protease cleavage site followed by the 21 nucleotides of the NP sequences ( 5′ GAAAACCTGTACTTCCAGGGT- 21 nt 3′) and ii) a reverse primer carrying attB2 sequence for cloning by recombination, two stop codons and the 21 nucleotides long reverse complement sequence of the NP ( 5′ GGGGACCACTTTGTACAAGAAAGCTGGGTC TTATTA -21 nt reverse complement 3′). A second PCR was done on the first PCR product using the same reverse primer and a universal forward primer carrying the attB1 sequence as well as two additional nucleotides (TA) for the coding frame, and a sequence hybridizing the TEV protease cleavage site ( 5′ GGGGACAAGTTTGTACAAAAAAGCAGGCT TA GAAAACCTGTACTTCCAGGGT 3′ ). This second PCR products were cloned into the pDonR201 plasmid by recombination (Gateway, Invitrogen). The resulting entry clones were sequenced. The entry clones were then used as templates to clone the NP into two expression plasmids (see figure 2 ): pDest17 (Invitrogen), that allows the expression of the NP in fusion with a N-terminal Hexahistidine (6His) tag, removable with the inserted TEV protease cleavage site and pETG20A (kindly provided by Dr A. Geerlof) that allows the expression of the NP in fusion with a removable Thioredoxin-Hexahistidine (TRX-6His) tag. The resulting expression plasmids were transformed in C41 (DE3) E. coli strain (Avidis SA) carrying the pRARE plasmid (Novagen). For each construct, one liter of Terrific Broth (Athena Enzymes) containing 100 mg/l of ampicillin and 34 mg/l of chloramphenicol was inoculated with 30 ml of an overnight pre-culture. The bacteria were grown at 37°C up to OD 600 nm reached 0.8. Recombinant protein expression was then induced by adding 0.5 mM isopropyl β-D-1-thiogalactopyranoside (IPTG) and the culture temperature was dropped to 17°C for 16 hours. Cells were harvested by centrifugation at 4 000 g, 10 minutes. Cell pellets were then resuspended in 50 mM Tris buffer, 300 mM NaCl, 10 mM imidazole, 0.1% Triton, and 5% glycerol (pH 8.0). Lysozyme (0.25 mg/ml), phenylmethylsulfonyl fluoride (1 mM), DNase I (2 μg/ml), and EDTA-free protease cocktail (Roche) were added before performing a sonication step. The lysates were centrifuged at 12 000 g for 45 minutes. For both 6His and TRX-6His constructs, the supernatants were collected for the purification procedure in non-denaturing conditions, whereas for the 6His constructs, the pellets were used for the purification process in denaturing conditions, as described in figure 2 .
Purification under non-denaturing conditions
The recombinant soluble proteins were purified from the supernatant previously recovered using the Akta Xpress fast purification liquid chromatography system (GE Healthcare) as follows. The first purification step (immobilized metal affinity chromatography) was performed on a 5-ml His prep column (GE Healthcare). The clarified bacterial lysates were loaded at 5 ml/min. The columns were then washed with a washing buffer (50 mM Tris, 300 mM NaCl and 50 mM imidazole pH 8.0) and the proteins were eluted with 50 mM Tris, 300 mM NaCl and 500 mM imidazole pH 8.0. The elution fraction (15 ml) was then dialyzed against a buffer compatible with the TEV protease activity (Hepes 10 mM, NaCl 300 mM, pH 7.5). 6His tagged TEV protease mutant selected for optimized expression [32] was added to the protein samples in a 1/20 (w/w) ratio and cleavage was performed 16 hours à 4°C. The tag removal efficacy was evaluated on a Coomassie blue stained SDS Page gel by ImageJ software and was calculated on the band surface and intensity of the full lengths TRX-6His fusions: [100-(I AC S AC /I BC S BC )x100]% where I is the intensity of the band, S is the surface, “AC” is “After Cleavage” and “BC” is “Before Cleavage”. After cleavage, the solution was loaded on a 5-ml His prep column and the cleaved proteins were collected in the flow through, whereas 6His-TEV protease, TRX-6His tag and not cleaved fusions were retained onto the column. The cleaved proteins were further purified by a Size Exclusion Chromatography (SEC) in 10 mM Hepes, 300 mM NaCl, pH 7.5 using a 16/60 Superdex 75 (GE Healthcare). The SEC was calibrated using the LMW and HMW calibration kits (GE Healthcare) in order to convert elution volumes in Molecular Weights. The purity of the samples was checked on Coomassie blue stained SDS Page gels. The fraction of the low molecular weight oligomer of NP (LMWNP) amenable to protein crystallization was defined as: (PA LMWNP /PA ALL x100)% where PA is the “peak area” of the SEC chromatogram at OD 280 nm . The identity of the recombinant proteins were confirmed by matrix-assisted laser desorption ionization-time of flight mass (MALDI-TOF, Bruker Autoflex, Bruker Daltonics,) spectrometry after trypsin digestion. Finally, the oligomerization was also assessed on Coomassie blue stained SDS Page gels, with an upstream treatment with glutaraldehyde, as previously described [20] . Briefly, different collected fractions were incubated 30 minutes at room temperature with glutaraldehyde at a 0.05% concentration. The samples were then denatured 5 minutes at 95°C in SDS Page sample buffer (100 mM Tris pH 6.8, 25% glycerol, 10% sodium dodecyl sulfate, 5 mM β-mercaptoethanol, bromophenol blue) prior electrophoresis on SDS Page gels.
Purification under denaturing conditions
The pellets obtained from the centrifuged cell lysates were washed twice to purify the inclusion bodies. The pellets were resuspended in a first washing buffer (50 mM Tris, 25 mM Imidazole, 300 mM NaCl, 1 M urea, 0.1% Triton, pH 8) and centrifuged at 12 000 g during 30 minutes. The supernatant was removed and the new pellets were resuspended in a second washing buffer (50 mM Tris, 25 mM Imidazole, 300 mM NaCl, 1 M urea, pH 8). After a second centrifugation step (12 000 g, 30 minutes), inclusion bodies recovered from the pellets were solubilized in a denaturing buffer (50 mM Tris, 300 mM NaCl, 25 mM Imidazole, 8 M Guanidium, pH 8). The purity of the recombinant proteins was assessed on Coomassie blue stained SDS Page gels and their quantity was evaluated at OD 280 nm . Denaturing proteins were concentrated up to about 10 mg/ml and diluted 1/20 (v/v) in a refolding buffer (50 mM Sodium acetate, 100 mM KCl, 10 mM β-mercaptoethanol, pH 4.5) at 4°C overnight. The refolding volume did not exceed 60 ml (30 mg of the protein) in order to have protein quantities and volumes compatible with the downstream procedure. The observed aggregates after refolding were removed by centrifugation (10 000 g, 15 minutes) and the solution was further clarified using filtration on 0.22 μm filters. A preliminary refolding efficacy was calculated as follow: (100×[NP] Sol /[NP] BR )%, where “[NP] Sol ” is the NP concentration in the soluble fraction after refolding and [NP] BR is the NP concentration before refolding (0.5 mg/ml). The refolded proteins were then concentrated on Amicon Ultra 10 K (Milipore) up to 5 ml before being loaded on a 16/60 Superdex 75 (GE Healthcare) equilibrated with 10 mM Hepes, 300 mM NaCl, pH 7.5 for SEC purification and homogeneity analysis, as described previously.
Protein crystallization
Crystallization trials were initiated with the RVFV NP purified under non-denaturing conditions and TOSV NP purified from IB using a nano-drop dispenser (Honeybee; Genomic Solutions) in 96-well sitting drop plates (Greiner Bio One). Three commercial crystallization kits were tested at 20°C: Structure Screen combination, Stura footprints (Molecular Dimensions Limited), and Nextal SM1 (Qiagen). For each condition, three drops were done: 300, 200 or 100 nl were added to 100 nl of the crystallization solution. | Results
Cloning and expression strategy
The ORFs encoding NP proteins of RVFV, SFSV and TOSV were cloned in pDONR201 before being re-introduced in two plasmids for expression as a N-terminal tag fusion, as described in figure 2 . Based on their small size and ability to be purified by Immobilized Metal Affinity Chromatography (IMAC), only 6His (3.3 kDa) and TRX-6His (14.6 kDa) tags were selected for the tag screening although other tags such as 6His/MBP (Maltose Binding Protein), or 6His/GST (Glutathion S-Transferase) could be available and compatible with the cloning procedure [33] – [34] . Nevertheless, the latter tags were not tested because they are much larger than 6His or TRX-6His and might interfere in the oligomerization process by steric hindrance. When fused to a removable 6His tag, the NP proteins can follow two procedures. If the protein is expressed in the soluble fraction, it can be purified under non-denaturing conditions. The 6His tagged fusions can also be expressed as inclusion bodies for subsequent refolding. By contrast, the TRX-6His was used only for soluble expression for two reasons. Firstly, the TRX tag is expected to improve protein solubility. Secondly, if the TRX-6His tag needs to be refolded in a condition that isn't compatible with recombinant protein optimum, and vice versa , the presence of the tag would be deleterious for refolding efficacy. Protein expression was performed in only one culture condition that was chosen on the conclusions of a previous report [35] . Briefly, the two main culture parameters having an impact on soluble expression are the use of rare tRNA co-expressing strains to improve expression yields, and post-induction cultures at low temperatures to promote solubility. E. coli strain carrying a pRARE plasmid (Novagen) and bacterial growth at 17°C were thus selected.
Protein purification under non-denaturing conditions
For each NP fused with an Nterminus 6His tag, a part of the recombinant protein can be expressed in its soluble form, as shown by the amount of protein quantified from the soluble fraction after IMAC purification ( figure 3 , data summarized in table1 ). Among the 3 NPs, RVFV NP was the only one to be soluble at yields above 1 mg/L culture. This latter yield was arbitrarily defined as the threshold compatible with downstream crystallogenesis experiments. Therefore, among the 6His tagged fused proteins, only RVFV NP was further purified and characterized. When compared to the amount of the insoluble fraction (data in table 1 ), the soluble fraction of NPs corresponds to 1.5% of the overall expression (3.5 mg in the soluble fraction compared to 240 mg in the inclusion bodies for RVFV NP), or less for SFSV and TOSV NPs. By contrast, when fused to the solubilizing TRX-6His tag, the NPs are 7 to 20-fold more soluble, providing thus enough material for tag cleavage. Interestingly, the solubility trend observed with the 6His fusions (RVFV>TOSV>SFSV) is conserved with the TRX-6His tag.
The removal of the TRX-6His tag was then assessed based on two criteria. First, the remaining full length TRX-6His-NP can be quantified and compared to the quantities of the full length constructs before cleavage. Using this calculation, the best cleavage efficacy was observed for TOSV NP, for which no residual full length fusion protein was observed after the TEV protease cleavage, as shown in figure 4 . RVFV NP was also efficiently recovered since only 13% of the full length was reluctant to TEV protease cleavage. In contrast with to these two NPs, more than 90% of the full length TRX-6His-SFSV NP remained uncleaved after the incubation with TEV protease, resulting in about 10% cleavage yield. Following the cleavage, the protein solution is loaded again on a Nickel immobilized column. Theoretically, this step binds 6His tagged proteins (TRX-6His, full length fusions, and TEV protease) and separates them from the untagged protein (cleaved protein of interest) going through the column. Practically, most of the cleaved TOSV NP was found in the flow through fraction whereas the 6His tagged proteins and a small amount of cleaved TOSV NP were trapped on the column and released during elution ( figure 4 ). The purification of RVFV NP was not as efficient as for TOSV NP. Indeed, cross contaminations of full length and cleaved protein can be found in both flow through and elution pools. However, the cleaved RVFV NP represents about 90% of the flow through and is therefore amenable to SEC. For SFSV NP, the cleaved protein was observed at the elution of the IMAC with the TRX-6His-NP fusion. Moreover, only non cleaved NP went through the nickel immobilized column. It was thus concluded that the purification of SFSV NP failed with the TRX-6His construct and the process was aborted.
In summary, among the six fusion proteins tested from the soluble fraction, three constructs yielded amounts and homogeneity criteria, as shown in table 1 : 6His-RVFV NP, TRX-6His-RVFV NP and TRX-6His-TOSV NP.
6His-RVFV NP showed a degradation pattern after the IMAC purification, resulting in four sub-products observed from 26 to 30 kDa ( figure 3 ), whereas the expected size of the NP with the 6His Tag is 31 kDa. The oligomerization of the sub-products was then analyzed by SEC ( figure 5 , panel A). Based on the calibration curve, the main part of the 6His-RVFV NP eluted at 94 kDa that could correspond to a trimer of 6His-NP (theoretical MW: 93 kDa). A larger oligomer over 300 kDa was also observed. The oligomerization of the NP is independent to the protein degradation since the four cleaved products are almost equally distributed along the chromatogram ( figure 5 , panel A). Therefore, although the criteria in homogeneity and quantity were reached (see Table 1 ), crystallization trials were not launched because of protein degradation.
SEC was also performed for the NP of both RVFV and TOSV after release of the TRX-6His tag. RVFV NP mainly eluted at 93 kDa ( figure 5 , panel B) following the same trend as that of 6His-RVFV NP. In order to compare the oligomerization state of the RVFV NP in this protocol to a previous study [20] , several fractions from 300 to 30 kDa were treated with or without glutaraldehyde and analyzed on Coomassie blue stained SDS Page ( figure 6 ). The cross-linking with glutaraldehyde resulted in three major populations (monomers (1NP), dimers (2NP) and timers (3NP)) as well as to a lower extent tetramers and high molecular complexes on the top of the gel. The amounts of monomers, dimers, trimers and tetramers remained almost unchanged along the chromatogram. By contrast, crosslinked HMW complexes could be observed in the fractions that were collected at low elution volumes. Unlike the corresponding 6His tagged NP, the protein was pure and not degradated as judged by Coomassie Blue stained SDS Page ( Figure 6 ). Fractions corresponding to the major peak were pooled and the so-called LMW NP gathered 33% of the injected NP. The behavior of TOSV NP after the TRX-6His tag cleavage was different: most of the protein eluted at 193 kDa and the LMW NP corresponds to about 5%, leading to less than 1 mg of protein ( table 1 ). Nevertheless, the protein eluting at 193 kDa was pooled and concentrated for crystallogenesis experiments.
Purification under denaturing conditions and refolding of IB
More than 200 mg of each of the 6His-NP constructs were expressed as inclusions bodies ( table 1 ). Since the recombinant proteins were highly expressed, two standard washes of the inclusion bodies were sufficient to recover NPs that are more than 90% pure, as shown in the total fraction of the refolded NPs ( figure 7 ). Therefore, neither additional purification step nor optimized washes were needed before refolding. The theoretical isoelectric point (pI) for RVFV, SFSV and TOSV are respectively 9.8, 10.1 and 9.9. In order to refold 6His-NPs at a pH distant to the protein pI, it was decided to refold at a low pH and sodium acetate at pH 4.5 was selected. A first analysis of the protein refolding efficacy was performed by comparing the total amount of 6His-NP to be refolded (e.g. 30 mg for each NP) with the soluble and filtrated fraction after overnight dilution in the refolding buffer. Quantitative data ( table 1 ) were obtained by comparing OD 280 nm in the total and soluble fractions. Qualitative data (protein purity, and degradation) were evaluated using Coomassie blue stained SDS Page ( figure 7 ). More than 25% of RVFV and TOSV 6His-NP were recovered in the soluble fraction whereas the corresponding SFSV construct was completely insoluble after refolding.
The oligomerization states of RVFV 6His-NP after refolding was different to the ones observed for the soluble corresponding NPs. Most of the protein eluted at 134 kDa (4,3 molecules of 6His-NP) during the SEC. A minor peak at 56 kDa (1,8 6His-NP) was also observed ( figure 5 ). This latter population corresponds to 16% of the injected protein, leading to less than 1 mg of protein from the 30 mg used for refolding. The chromatogram of the refolded TOSV 6His-NP shows a homogeneous protein with peaks at 97 kDa (3 molecules of TOSV 6His-NP) and 33 kDa (1 TOSV 6His-NP). The NP eluting at 97 kDa corresponds to 63% of the injected protein (4 mg). The two protein pools corresponding to the peaks 97 and 33 kDa were separately used for crystallogenesis trials.
Protein crystallization
RVFV NP and TOSV NPs obtained from the non-denaturing production pipeline after TRX-6His removal, as well as the two pools (97 kDa and 33 kDa) of the refolded 6His-TOSV NP were finally concentrated in the SEC buffer up to 6.6 mg/ml, 6.3 mg/ml, 7.4 mg/ml and 3 mg/ml respectively. From the commercial kits, several crystal hits were obtained for RVFV NP ( figure 8 , panel A), and one condition lead to sea urchin crystals of the 97 kDa MW TOSV NP ( figure 8 , panel B). | Discussion
The initial objective of this study was to produce and purify phlebovirus NPs for structural studies. According to the previously reported studies, NPs were described as insoluble [23] or HMW complexes [22] . Oppositely to the latter strategy whose goal was to identify the viral organization of Rift Valley fever virus at the macromolecular level, our aim was to find a procedure to produce soluble and LMW NPs. Therefore, we decided to adopt a naive approach testing in parallel several methods and evaluate them on the basis of protein solubility and production of LMW NPs. With these two evaluation criteria, we devised an easy “Go/No Go” pipeline ( figure 2 ) that enables the comparison of the methods and the selection of the best one, in order to reach the required amount of homogeneous NP for crystallogenesis. For several reasons, it was decided to take advantage of the natural diversity within the viral genus. Firstly, screening several homologues can be considered as an additional experimental variable. It has already been demonstrated that it can improve the success rate to find at least one sequence suitable for structural studies [36] . Secondly, the cumulative data of several proteins could benchmark the protocol described in figure 2 . Finally, it could be of interest to see if phlebovirus NPs, share preliminary structural features such as oligomerization. Three phleboviruses belonging to distinct phylogenic clusters based on NP sequences [37] were therefore selected, the pairwise sequence identities of the NPs being 50% id. between RVFV and TOSV, 47% id. between RVFV and SFSV and 40% id. between TOSV and SFSV.
Among these three NPs, SFSV NP was the most reluctant to all strategies. Either the protein with a 6His tag is poorly soluble, or it cannot be refolded easily. When expressed in fusion with the TRX-6His sequence, the solubility issue is overcome but the fusion protein was almost refractory to TEV protease cleavage. Moreover, the cleaved and uncleaved forms co-eluted together on IMAC with no rational on the presence/absence of the 6His sequence. Altogether, these data suggest that even if the TRX-6His solubilizes the SFSV NP, the fusion protein organizes in soluble aggregates, as already observed with MBP when it is fused to the human papillomavirus E6 protein [38] . By contrast, the two RVFV and TOSV were soluble and easier to handle. However, sequence analysis for solubility in E. coli [39] did not allow predicting of the untracktability of SFSV compared to the two other NPs. Again, it suggests that the success rate of producing a class of proteins depends on the number of homologues tested experimentally. Alternatively, specific point mutations of poorly conserved amino acids between SFSV and the two others NPS might constitute a successful strategy to obtain soluble or stable SFSV NP. Such a study might help to point out critical residues like prolines or cysteines that can be involved in recombinant protein folding/misfolding.
TOSV NP was soluble in both 6His and TRX-6His constructs but the latter tag improved solubility about 20-fold. The tag was cleaved and the SEC chromatogram revealed two NP populations that could correspond to hexameric and trimeric complexes ( figure 5 , panel B). Surprisingly, the oligomerization states shifted to putative trimers and monomers (panel C) from the purification under denaturing conditions followed by refolding. A much more striking difference was observed for the RVFV NP. Both constructs (6His-NP and NP) from the non-denaturing pipeline eluted mainly as putative trimers ( figure 5 , panels A and B), whereas 6His-NP after refolding formed putative monomers, dimers and tetramers (panel C). This “2X” multimerization is in agreement with results obtained for refolded RVFV NP that lead to the structure determination of a NP dimer [26] but it differs from the “3X” multimerization of native and refolded TOSV NP and native RVFV NP. Our data suggest that the choice of the purification procedure ( i.e., denaturing vs non-denaturing conditions) has an impact on the oligomerization state going through two divergent directions and at this stage, we were not able to claim if the difference was due to a refolding artifact or an equilibrium between the two forms.
In a previous study, it was demonstrated that the RVFV NP can dimerize through an interaction involving the N-terminal domain [20] . When cross linking RVFV NP with glutaraldehyde, we did confirm that the protein expressed in E. coli is able to dimerize but also to produce tri- and tetramers ( figure 6 ). These latter oligomers almost disappeared in the HMW NP (>300 kDa), suggesting that HMW could behave like RNPs purified from infected Vero cells [20] . For the lower molecular weight complexes, the cross linking with glutaraldehyde resulted in the formation of a ladder that is probably due to a non specific reaction. Since SEC chromatograms of the soluble 6His-RVFV NP and RVFV-NP are almost superimposable, the 6His tag was not responsible for the oligomerization differences.
Two NP pools crystallized ( figure 8 ), one from the non-denaturing pathway (RVFV NP) and one from the refolding one (TOSV NP). In both cases, crystals were obtained with trimeric NPs. These results raise the importance to investigate the production under not only the native conditions but also the denaturing conditions for at least two raisons. Firstly, in contrast with to the soluble fraction of RVFV 6His-NP, refolded 6His-NP remained not degraded, even after refolding. It can be noticed that the degradation of the 6His-NP, when processed in non-denaturing conditions, occurred at an early stage of the production ( figure 3 ), probably during the culture growth or bacterial lysis, whereas the NPs produced as IB are prevented from bacterial proteolysis. Secondly, the putative trimer of TOSV NP that crystallized was obtained in higher amounts after refolding than in the soluble fraction that majorly lead to a heterogeneous oligomer preparation. Nevertheless, refolding from solubilized inclusion bodies met some limits in the purification of NP. When SFSV NP is not soluble and homogeneous even with the TRX tag, the refolding strategy also failed. Moreover, the renatured RVFV NP produced multimers that are different from RVFV and TOSV NPs purified under native conditions, in agreement with a previous study [26] . Although it remains unclear if the differences in the oligomerization states highlight a difference in protein folding, the structure of a phlebovirus NP as a trimer would certainly help understanding the oligomerization determinants of phlebovirus NPs.
Conclusion
The standard pipeline investigating conditions under both denaturing and non-denaturing conditions was proven to be efficient and could be applied to any recombinant protein. With no need of further refinement, two out of the three phlebovirus NPs were produced and purified in suitable amount and quality for crystallogenesis. Trimeric forms of RVFV and TOSV NPs yielded crystals. This result is a starting point for structural studies aiming at the elucidation of the RNA encapsidation mechanism, a targetable step for antiviral research [40] . | Conclusion
The standard pipeline investigating conditions under both denaturing and non-denaturing conditions was proven to be efficient and could be applied to any recombinant protein. With no need of further refinement, two out of the three phlebovirus NPs were produced and purified in suitable amount and quality for crystallogenesis. Trimeric forms of RVFV and TOSV NPs yielded crystals. This result is a starting point for structural studies aiming at the elucidation of the RNA encapsidation mechanism, a targetable step for antiviral research [40] . | Conceived and designed the experiments: V. Lantez, K. Dalle, B. Coutard. Performed the experiments: V. Lantez, K. Dalle. Analyzed the data: V. Lantez, B. Coutard, B. Canard. Contributed reagents/materials/analysis tools: R. Charrel, C. Baronti. Wrote the paper: B. Coutard, B. Canard.
Nucleoproteins (NPs) encapsidate the Phlebovirus genomic (-)RNA. Upon recombinant expression, NPs tend to form heterogeneous oligomers impeding characterization of the encapsidation process through crystallographic studies. To overcome this problem, we set up a standard protocol in which production under both non-denaturing and denaturing/refolding conditions can be investigated and compared. The protocol was applied for three phlebovirus NPs, allowing an optimized production strategy for each of them. Remarkably, the Rift Valley fever virus NP was purified as a trimer under native conditions and yielded protein crystals whereas the refolded version could be purified as a dimer. Yields of trimeric Toscana virus NP were higher from denaturing than from native condition and lead to crystals. The production of Sandfly Fever Sicilian virus NP failed in both protocols. The comparative protocols described here should help in rationally choosing between denaturing or non-denaturing conditions, which would finally result in the most appropriate and relevant oligomerized protein species. The structure of the Rift Valley fever virus NP has been recently published using a refolded monomeric protein and we believe that the process we devised will contribute to shed light in the genome encapsidation process, a key stage in the viral life cycle.
Author Summary
Phleboviruses have a worldwide distribution and are usually represented by their prototype Rift Valley fever virus that can have a great impact on health and economy in Africa. The genome of phleboviruses is a segmented negative strand RNA that is encapsidated by the nucleoprotein. The structure of the monomeric nucleoprotein has been recently published but it's not sufficient to decipher a convincing mechanism for the nucleoprotein oligomerization. In order to understand this key step in the virus life cycle, the purification of oligomers homogeneous in size would be a key step to launch structural studies. To that aim, a procedure relying on recombinant protein production in both denaturing and non-denaturing conditions was applied to three phlebovirus nucleoproteins. Although the best production pipeline differs for each protein, pure and homogeneous solutions of Rift Valley fever virus and Toscana virus nucleoproteins were successfully obtained. Both proteins, behaving as apparent trimers in solution, lead to protein crystallization, a starting point to understand the genome encapsidation through structural studies. | Supporting Information | The authors thank Dr H. Berglund and Dr A. Geerlof who kindly provided respectively the TEV protease expression plasmid and pETG20A. | CC BY | no | 2022-01-13 05:08:15 | PLoS Negl Trop Dis. 2011 Jan 4; 5(1):e936 | oa_package/fa/0e/PMC3014985.tar.gz |
PMC3014986 | 21245925 | Introduction
Plague arrived in Brazil during the Third Pandemic, in October 1899, imported by ship traffic to Santos, in São Paulo state, and was rapidly diffused to other coastal cities. By 1906, it had dispersed by means of land and sea commerce more broadly, and had become established in native rodent populations, particularly in the northeastern sector of the country [1] , [2] . Nonetheless, records of plague in Brazil are sparse through the 1920s, making detailed tracking of the pattern of spread of the disease in the region difficult or impossible [1] , [3] .
Only by around 1936 were data on plague and its control in Brazil regularly collated and archived. Based on analyses of these data (for 1936–1966), Baltazard [1] identified numerous distinct plague foci occurring in different environmental contexts, a viewpoint that was updated by Vieira & Coelho [4] , based principally on elevation. These foci appear to exist independently of one another in time and space [1] , and overall numbers of human cases varied from 20 to 100 until the 1970s. Since that time, all of the foci entered a period of relative inactivity, with few or no human cases [5] , [6] , [7] , [8] . The last significant outbreak in Brazil was in the late 1980s in Paraíba [9] .
The purpose of this contribution is to present a first range-wide analysis of the geography and ecology of plague transmission in northeastern Brazil using tools drawn from the emerging field of ecological niche modeling, which is beginning to see application to plague biology [10] , [11] . Although no recent plague transmission to humans has been recorded in this region, plague remains as a zoonosis across much of northeastern Brazil [12] , making a thorough understanding of its geographic distribution an ongoing priority. Here, we marshal new tools from quantitative biogeography in the form of ecological niche modeling approaches, which related known points of occurrence to raster geospatial GIS data layers to estimate the ecological niche of a species or other biological phenomenon, such as transmission of a disease [11] . The result is both a spatial prediction of areas of potential transmission and a first-order evaluation of environmental correlates of plague transmission in northeastern Brazil. | Methods
Input Data
Plague case occurrences
We drew occurrence data from the Serviço Nacional de Referência em Peste do Centro de Pesquisas Aggeu Magalhães, Recife, Pernambuco. The data cover the period 1966-present, and represent only cases with laboratory confirmation by means of bacteriological examination or serological testing [6] , [7] , [8] , [9] , [13] .
Case-occurrence data were referenced spatially to the particular town or ranch where the affected person lived, which we georeferenced by means of referring to municipal maps at diverse scales published by the Instituto Brasileiro de Geografia e Estatística. Of the total of 203 laboratory-confirmed plague cases, which fell into 157 distinct localities, 120 localities could be georeferenced with a spatial precision of 3 km or finer, and most to within 500–1000 m.
Environmental data
We sought fine-resolution geospatial data characterizing environmental variation across the landscapes of northeastern Brazil for the time period of interest. We estimated potential risk areas for plague case occurrences based on indirect, landscape-scale measures that are likely correlates of the actual environmental factors associated with the disease-vector-host interactions that affect the ecology of this disease [14] . In particular, we focused on correlates of key environmental dimensions related to precipitation, temperature, vegetation, topography and land use, so we explored two principal environmental data sets. First, we chose imagery from the Advanced Very High Resolution Radiomater [AVHRR; 15] satellite, in particular using multitemporal imagery summarized as Normalized Difference Vegetation Indices [henceforth “NDVI”] ; [ 15,16] for 12 individual months in the year spanning from April 1992 – March 1993, which was the period of availability of imagery closest to the period of plague transmission characterized by our occurrence data (a single year was used owing to limited availability of monthly data AVHRR sets). The NDVI data layers that we used are estimators of the photosynthetic mass presented within 1 km grid cells across the landscape [17] . We used individual monthly NDVI values to provide the model with information about seasonal variation in ‘greenness,’ which has been seen to be important in previous exercises in prediction of disease transmission geography [18] , [19] . Although clearly other means of summarizing seasonality in greenness of landscapes are available [e.g., 20] , the data sets and methodologies described in the latter paper have not been made broadly available.
Second, we explored the utility of fine-resolution interpolated climate dimensions [21] . These data, consisting essentially of monthly temperature and precipitation values averaged over 1950–2000, have been processed into 19 ‘bioclimatic variables’ that are thought to be more biologically relevant than the raw values [21] . Because of suspected high levels of intercorrelation among these variables, we explored correlations among them, and eliminated one member of each variable pair that showed high correlations, choosing the particular variable to eliminate based on ease of interpretation of the variable, leaving the following variables for analysis: annual mean temperature, mean diurnal range, maximum temperature of warmest month, minimum temperature of coldest month, annual precipitation, precipitation of wettest month, and precipitation of driest month.
Based on initial testing, we also explored inclusion of elevation as a predictor layer. A digital summary of elevational variation across the study area was obtained from the Hydro-1K data set [22] , with 1 km native spatial resolution. In sum, we analyzed NDVI and climate data sets, each with and without elevation data. All layers were clipped to the area within 300 km of known plague occurrences for analysis ( Figure 1 ).
Ecological Niche Modeling
Methods and approaches for estimating ecological niches from species' occurrence data have seen considerable exploration in recent years [23] , [24] . Outcomes of these tests have been mixed, with some serious criticisms of the algorithm used herein, the Genetic Algorithm for Rule-Set Prediction (GARP) [25] —these criticisms [23] , [26] , however, have been based either on misunderstandings of how to use the algorithm [27] or on artifactual differences in performance measures [28] , [29] . In reality, and when properly used and evaluated, GARP offers estimates of species' ecological niches that are highly robust to small sample size and to broad gaps in spatial coverage of landscapes in terms of input data [28] , [29] —for this reason, we used this approach throughout this study.
GARP is an evolutionary-computing method that estimates niches based on non-random associations between known occurrence points for species and sets of GIS coverages describing the ecological landscape. Occurrence data are used by GARP as follows: 50% of occurrence data points are set aside for an independent test of model quality (extrinsic testing data), 25% are used for developing models (training data), and 25% are used for tests of model quality internal to GARP (intrinsic testing data). Distributional data are converted to raster layers, and by random sampling from areas of known presence (training and intrinsic test data) and areas of ‘pseudoabsence’ (areas lacking known presences), two data sets are created, each of 1250 points; these data sets are used for rule generation and model testing, respectively.
The first rule is created by applying a method chosen randomly from a set of inferential tools (e.g., logistic regression, bioclimatic rules). The genetic algorithm consists of specially defined operators (e.g. crossover, mutation) that modify the initial rules, and thus the result are models that have “evolved”—after each modification, the quality of the rule is tested (to maximize both significance and predictive accuracy) and a size-limited set of best rules is retained. Because rules are tested based on independent data (intrinsic test data), performance values reflect the expected performance of the rule, an independent verification that gives a more reliable estimate of true rule performance. The final result is a set of rules that can be projected onto a map to produce a potential geographic distribution for the species under investigation.
Because each GARP run is an independent random-walk process, following recent best-practices recommendations [30] , for each environmental data set (see above), we developed 100 replicate random-walk GARP models, and filtered out 90% based on consideration of error statistics, as follows. The ‘best subsets’ methodology consists of an initial filter removing models that omit (omission error = predicting absence in areas of known presence) heavily based on the extrinsic testing data, and a second filter based on an index of commission error ( = predicting presence in areas of known absence), in which models predicting very large and very small areas are removed from consideration. Specifically, in DesktopGARP, we used a “soft” omission threshold of 20%, and 50% retention based on commission considerations; the result was 10 ‘best subsets’ models (binary raster data layers) that were summed to produce a best estimate of geographic prediction. We took as a final ‘best’ prediction for each species that area predicted present by any, most, or all 10 of these best-subsets models.
Predictive models of disease occurrence may be good or bad, but model quality can be ascertained only via evaluation with independent testing data, preferably which are spatially independent of the training data to avoid problems caused by spatial autocorrelation and nonindependence of points [28] . Because only data documenting presence of plague cases were available for this study (i.e., no data were available to document that plague was absent at particular sites), we used a binomial probability approach to model validation: we compared observed model performance to that expected under a null hypothesis of random association between model predictions and test point distribution. Because such tests require binary (i.e., yes-no) predictions, our first step was to convert raw (continuous) predictions to binary predictions. We considered three distinct thresholds: areas predicted as suitable by any (i.e., ≥1) of the 10 replicate best-subsets models (ANY), areas predicted as suitable by most (i.e., >5) of the 10 replicate best-subsets models (MOST), and areas predicted as suitable by all of the 10 replicate best-subsets models (ALL).
In the binomial tests, the number of test points was used as the number of trials, the number of correctly predicted test points as the number of successes, and the proportion of the study area predicted present as the probability of a success if predictions and points were associated at random [31] . All testing was carried out in a series of spatially stratified tests that are detailed below. These tests evaluated the ability of models to anticipate plague case distributions across unsampled areas, considering a model as validated if it predicts case distributions better than a “model” making random predictions. As such, these tests are considerably more stringent than simple random partitions of occurrence data or cross-validation exercises.
In view of the odd, focal distribution of northeastern Brazilian plague cases ( Figure 1 ), we carried out a series of tests of predictive abilities of models among the five foci that are easily discernable. In each case, we examined model predictivity in a k – 1 framework: with k = 5 foci, we tested all combinations of 4 foci by means of their ability to predict spatial distributions of plague cases in the fifth focus. Tests were developed within two spatial contexts—within 50 km and within 200 km—surrounding the known occurrences within the target focus.
Finally, we wished to develop a single overall model that represents the best-available picture of plague case-occurrence risk across northeastern Brazil, albeit not including statistical testing as above. This model was built using all occurrence data available. To assess uncertainty in these predictions based on all case-occurrence information, we built 100 models each based on a random 50% of the occurrence data chosen at random without replacement. These models thus capture the degree to which plague case-occurrence data availability may drive the results of the analyses, and we consider areas that are predicted consistently in all of these replicate analyses as most certain. We projected this model onto environments across eastern Brazil to provide a broader-extent visualization of the ‘niche’ of plague in northeastern Brazil.
Niche Characterization
To explore environmental factors associated with positive and negative predictions of suitability for DF transmission, we explored further the environmental correlates of the model based on all points. We plotted 1000 points randomly across areas of the municipalities predicted as absent or present by this model. We then assigned the value of each input environmental and topographic layer to each of the random points, and exported the associated attributes table in DBF format, which was then used for comparisons of environmental characteristics of areas predicted as suitable and unsuitable. | Results
The focal and discontinuous nature of plague case distributions in northeastern Brazil is at once visible in the raw distribution of the occurrence points derived at the outset of this study ( Figure 1 ). The discontinuities that have been assumed based on the clusters of known occurrences are supported by our ecological niche models, many of which show relatively small areas of highly suitable conditions separated by less-suitable areas (see, e.g., Figure 2 ). What is more, this result is manifested with or without elevation included in the analysis, and thus is not a simple consequence of topographic differences; it is also manifested in analyses based on both surface reflectance (NDVI) and climatic variables. As such, we interpret the discontinuity of plague distributions in northeastern Brazil as dependent on a multidimensional suite of environmental variables.
The model predictions in general performed quite well in anticipating plague case distributions in areas not included in model training. That is, plague not only occurs in discontinuous foci, but it also occurs under predictable and circumscribed environmental conditions, which is the basis for the success of the niche model predictions. The broadest panorama of results shows significant results dominating in the southwestern and northwestern foci ( Table 1 ). However, the frequency of significant results in these tests is clearly and linearly related to sample size on a log 10 scale ( P <0.05), suggesting that predictivity would be excellent throughout the region were sample size distributions to be more adequate.
Finally, visualizing plague distributions in environmental dimensions ( Figure 3 ), we see clear differences in the seasonal pattern of variation in greenness between areas predicted as suitable (i.e., suitability value of 10) and those predicted as unsuitable (value 0). That is, no marked seasonal variation is notable in areas predicted as unsuitable, whereas areas predicted as suitable show a marked elevation in greenness in April and May, and lower values thereafter, probably corresponding to patterns of rainfall (i.e., rainy season beginning in March, and ending by August).
Extending the model predictions across broader areas—namely all of northeastern and eastern Brazil—yields a picture of potential plague distribution across the region ( Figure 4 ). Because plague transmission to humans in Brazil is currently nil, and no broad-extent data are available regarding circulation among mammals, we have few means of testing the reality of these model projections. However, at least in the case of models based on climatic dimensions, the area predicted as suitable includes the Serra dos Orgãos sites from which plague has been documented [5] , [32] . | Discussion
The models that we developed for Brazilian plague offer several intriguing insights into plague distribution, ecology, and natural history in Brazil. However, understanding the limitations of these models is critical, prior to any detailed interpretation or exploration. First and foremost among the limitations of this study are the occurrence data used as input: we relied on human case-occurrence reports accumulated by the Serviço Nacional de Referência em Peste do Centro de Pesquisas Aggeu Magalhães and published in diverse scientific publications [6] , [7] , [8] , [9] , [13] . Our use of these data thereby assumes that human case-occurrences are representative of the ecological and environmental situations under which plague is maintained in the zoonotic world, which may not be the case, given the long chain of events necessary for a zoonotic occurrence to be represented in our data set (i.e., transmission to human, correct diagnosis, international reporting). On a finer scale, we also make the not-completely-satisfactory assumption that that place of residence (at the level of the ranch or settlement) is representative of the site of infection, which may be variably true depending on the particular social network and local economy.
One point that became clear in our analyses, confirming previous opinions, is that plague has a highly discontinuous and focal distribution in northeastern Brazil. Our initial suspicions that elevation played a significant role in creating these ‘islands’ were not supported, as analyses with and without elevation included as a predictor variable reconstructed the insular nature of the distribution. The NDVI-based analyses are particularly instructive, as they have no direct, mathematical relation to elevation [as do climate interpolations; 21] —rather, the discontinuous plague distribution in northeastern Brazil appears to reflect multidimensional qualities of the landscape and environment (which of course may be related biologically to elevation), rather than any simple univariate causation.
Previous studies had attributed the cause of plague focality in Brazil to elevation [1] . Baltazard [1] emphasized that Brazilian plague foci are independent—that is, that transmission appears to occur in uncorrelated patterns in different foci. Baltazard [1] also pointed out that these foci are all in elevated areas, and that they are subject to distinct precipitation regimes. Although plague has frequently shown long periods of apparent inactivity (i.e., no human cases), its reappearance at intervals nonetheless indicates its long-term persistence.
The foci are limited geographically, although their footprint can appear to expand during major outbreaks. These expansions appear to correspond to periods of particularly favorable conditions for plague transmission in the highland area, spreading out via valleys into the surrounding lowland areas. If these favorable conditions persist, taking the form of a prolonged winter, rodent host reproduction may be elevated, and plague may be able to spread beyond the limit of the highland areas into the dry sertão per se. This line of thinking led Baltazard [1] to consider the plague foci of Serra da Ibiapaba, Serra do Baturité, Serra do Machado, Serra de Uruburetama, Serra da Pedra Branca, Serra das Matas in northern Ceará (see Figure 1 ) as a single focus. Vieira and Coelho [4] , in contrast, argued that these foci should be treated as isolated and independent. Our analyses suggest that these foci are dependent on a broad suite of conditions, and are not simple or direct correlates of elevation.
Another factor that may play in the picture of focality is the presence of key rodent hosts for plague, including Necromys lasiurus (formerly placed in Bolomys and Zygodontomys ). Necromys is the rodent that is most abundant in northeastern Brazilian plague foci, and was considered as responsible for causing epizootics, from which the infection spreads to other species [1] . Given the distribution of this species, other species of rodents must be involved in plague maintenance farther south, for example in the Serra dos Órgãos, Rio de Janeiro state, Brazil. The relative roles of the distribution of the rodent hosts and the fleas ( Polygenis spp.) remain to be evaluated in detail. | Conceived and designed the experiments: JG ATP AA. Performed the experiments: JG. Analyzed the data: JG. Contributed reagents/materials/analysis tools: ATP AA. Wrote the paper: JG ATP.
Plague in Brazil is poorly known and now rarely seen, so studies of its ecology are difficult. We used ecological niche models of historical (1966-present) records of human plague cases across northeastern Brazil to assess hypotheses regarding environmental correlates of plague occurrences across the region. Results indicate that the apparently focal distribution of plague in northeastern Brazil is indeed discontinuous, and that the causes of the discontinuity are not necessarily only related to elevation—rather, a diversity of environmental dimensions correlate to presence of plague foci in the region. Perhaps most interesting is that suitable areas for plague show marked seasonal variation in photosynthetic mass, with peaks in April and May, suggesting links to particular land cover types. Next steps in this line of research will require more detailed and specific examination of reservoir ecology and natural history.
Author Summary
We analyzed the spatial and environmental distributions of human plague cases across northeastern Brazil from 1966-present, where the disease is now only rarely transmitted to humans, but persists as a zoonosis of native rodent populations. We elucidated environmental correlates of plague occurrences by way of ecological niche modeling techniques utilizing advanced satellite imagery and geospatial datasets to better understand the ecology and geography of the transmission of plague. Our analysis indicates that plague foci in Brazil are indeed insular as previously suggested. Furthermore, distribution of such foci are likely not directly dependent on elevation, and rather are contigent on climate and vegetation. Seasonality of zoonotic plague transmission is linked to variations of these ecological parameters- particularly the increase in precipitation and primary production of the rainy season. Spatial analysis of transmission events afford a broad view of potential plague foci distributions across northeastern Brazil and indicate that the epidemiology of plague is driven by a dynamic array of environmental factors. | We thank the staff of the Centro de Referência em Informação Ambiental (CRIA), of Campinas, Brazil, for hosting ATP during the early phases of this study, and particularly Marinez Siqueira, Vanderlei Canhos, Dora Canhos, and Priscila Gomes for their assistance with diverse challenges. | CC BY | no | 2022-01-13 05:08:15 | PLoS Negl Trop Dis. 2011 Jan 4; 5(1):e925 | oa_package/d2/6b/PMC3014986.tar.gz |
||
PMC3015018 | 20857495 | Material and Methods
Mice, diets and carcinogens
We bred heterozygous B6;129S7- Ptgs2 tm1Jed /J males to females (Jackson Laboratory, Bar Harbor, ME) to generate Cox-2 −/− , Cox-2 +/− and Cox-2 +/+ (WT) littermates. 16 Custom-formulated ZD and Zn-sufficient (ZS) diets (Harlan Teklad, Madison, WI) were identical except for the Zn content. 16 NQO was from Wako Chemicals (Richmond, VA) and NMBA from Midwest Research Institute (Kansas City, MI).
NQO-induced tongue carcinogenesis
The mouse studies were approved by The Ohio State University Animal Use Committee. Four-week old littermates were fed ZD or ZS diets to form six groups, namely, ZD: Cox-2 −/− ( n = 14), ZD: Cox-2 +/− ( n = 46), ZD:WT ( n = 19), ZS: Cox-2 −/− ( n = 16), ZS: Cox-2 +/− ( n = 37) and ZS:WT ( n = 25). After 4 weeks the mice were administered NQO in deionized water for tongue tumor induction (20 ppm for 19 weeks followed by 30 ppm for 7 weeks). At 26 weeks, the animals were sacrificed for tumor incidence analysis.
Expression profiling and related studies
Weanling Cox-2 −/− and Cox-2 +/+ mice were fed ZD or ZS diets to form four groups, namely, ZD: Cox-2 −/− ( n = 20), ZS: Cox-2 −/− ( n = 12), ZD:WT ( n = 12) and ZS:WT ( n = 12). After 9 weeks, 8 ZD: Cox-2 −/− mice were switched to a ZS diet to form the ZR: Cox-2 −/− group. After a week, all mice were sacrificed. This experimental regimen produced unbridled cell proliferation in ZD: Cox-2 −/− forestomach. 16 Tongue and forestomach were isolated and cut into two portions. One portion was formalin-fixed and paraffin-embedded for immunohistochemical (IHC) studies. Forestomach epithelia for expression profiling studies were prepared from the remaining portion by using a blade to strip off the submucosal layers and snap-frozen in liquid nitrogen.
We performed expression profiling of forestomach mucosa from ZD: Cox-2 −/− , ZS: Cox-2 −/− , ZD:WT mice and ZS:WT mice after 10 weeks of ZD or ZS diets ( n = 4 mice/group), using GeneChip® Mouse Genome 430 2.0 Array (Affymetrix, Santa Clara, CA). Total RNA was extracted from forestomach mucosa using TRIZOL reagent (Invitrogen, Carlsbad, CA). Five micrograms of total RNA was reverse transcribed into cDNA followed by in vitro transcription and labeling to produce biotin-labeled cRNA. The cRNA was hybridized to the arrays as described. 14
Expression data analysis
We used the Class Comparison analysis of BRB-Array Tools software version 3.7.0 (Biometric Research Branch, NCI) to identify differentially expressed mRNAs. The Robust Multichip Average method was performed. The array data were submitted to ArrayExpress (Accession number: E-TABM-778).
Gene ontology and pathway analyses
We used DAVID (Database for Annotation, Visualization and Integrated Discovery) 25 bioinformatics to identify relevant biological processes/functions from expression data captured by transcriptome analysis. Based on gene ontology, differentially expressed genes were grouped by scoring the statistical significance of predefined functional gene groups according to their functional similarity.
We used Ingenuity Pathway Analysis software (IPA, http://www.ingenuity.com ) to analyze probable network/pathway and functional group enrichment. For each data set, the selected genes were uploaded into the IPA application. Networks were then algorithmically generated based on gene–gene connectivity.
ZR and forestomach carcinogenesis in ZD:Cox-2 −/− mice
This mouse study was approved by the Thomas Jefferson University Animal Use Committee. Thirty-nine 4-week old Cox-2 −/− mice were fed a ZD diet to form the ZD: Cox-2 −/− group. After 4 weeks, the mice received three intragastric doses of NMBA (2 mg/kg body weight, twice weekly), a regimen that produced a high tumor outcome in ZD: Cox-2 −/− mice. 16 A day after the 3rd dose, 18 mice were switched to a ZS diet to form the ZR: Cox-2 −/− group, which were given an intragastric dose of Zn gluconate weekly for 14 weeks (0.04 mg Zn). The remaining ZD: Cox-2 −/− mice continued on ZD diet. All mice were sacrificed for tumor outcome analysis at 14 weeks of Zn intervention.
Tumor analysis
At autopsy, tongue, esophagus and forestomach were excised. Tumors greater than 0.5 mm were mapped. Tissues were formalin-fixed and paraffin-embedded for histopathologic/IHC studies.
Quantitative reverse transcriptase-polymerase chain reaction
Quantitative reverse transcriptase-polymerase chain reaction (qRT-PCR) was performed using the comparative C t method and predesigned probes on 7300 Real-time PCR System (Applied Biosystems, Foster City, CA). GAPDH was used to normalize RNA samples. 14
Immunoblotting
Proteins were separated by 10–14% Tris-HCl gel (Bio-Rad, Hercules, CA) as described. 13 GAPDH (Calbiochem, San Diego, CA) was used as a loading control.
IHC
IHC was performed as described. 13 – 16 The following antisera were used: mouse anti-proliferating cell nuclear antigen (PCNA) monoclonal (Thermo Scientific); rat anti-S100A8 monoclonal, goat anti-S100A9 monoclonal, and rat anti-receptor for advanced glycation end-products (RAGE) monoclonal (R&D Systems, Minneapolis, MN); rabbit anti-nuclear factor (NF)-κB p65 polyclonal (Abcam, Cambridge, MA), rabbit anti-NF-κB phospho-p65 (serine 276) polyclonal and rabbit anti-cyclin D1 monoclonal antiserum (Cell Signaling, Danvers, MA) or rabbit anti-p53 polyclonal antiserum (detects both mutated and wild-type proteins) (Leica Microsystems, Bannockburn, IL). Protein was localized by incubation with 3-amino-9-ethylcarbazole substrate-chromogen (AEC) (Dako, Carpinteria, CA) or 3,3′-diaminobenzidine tetrahydrochloride (DAB; Sigma-Aldrich, St. Louis, MO).
Immunoreactive scores were calculated by multiplying the percentage of positive cells by the grade of staining intensity. 15 The percentage of positive cells was evaluated as follows: 0 = 0–5%, 1 = 6–25%, 2 = 26–50%, 3 = 51–75% and 4 = 76–100%. The intensity of immunostaining was graded as follows: 0 = none, 1 = weak, 2 = moderate and 3 = intense. The PCNA labeling index (%) was calculated by dividing the number of PCNA-labeled nuclei by the total number of cells counted.
Zn measurement
Hair Zn content was determined by atomic absorption spectrometry. 16 Invariably, Zn levels were significantly lower in ZD than ZS samples. As examples, hair Zn levels were significantly lower in ZD: Cox-2 −/− than ZS: Cox-2 −/− mice at 10 weeks (array study) [130 μg/g (95% confidence interval [CI] = 124–135) vs . 172 μg/g (95% CI = 165–179), p = 0.002, n = 10/group) and at 26 weeks (NQO study); 111 μg/g (95% CI = 99–122) vs . 157 μg/g (95% CI = 146–168), n = 14 mice/group, p < 0.001].
Statistical analysis
Tumor multiplicity was analyzed by two-way analysis of variance (ANOVA). Differences among the groups were assessed using the Tukey-HSD post hoc t -tests for multiple comparisons. Tumor and carcinoma incidence rates were assessed by Fisher's exact test. CIs for the differences in incidence rates were calculated using the Wilson Score Method. 26 Statistical tests were two-sided and considered significant at p < 0.05. | Results
ZD enhances tongue carcinogenesis in Cox-2 deficient mice
NQO is a DNA adduct-forming agent that serves as a surrogate of tobacco exposure. 27 Nutritionally complete WT mice exposed to 10 ppm of NQO for 50 weeks did not develop tongue lesions. 28 At a high concentration of 100 ppm, however, WT mice developed malignant tongue and esophageal tumors. 29
To investigate whether a Zn-deficient condition eliminates the antitumor effect of genetic Cox-2 disruption in NQO-induced tongue carcinogenesis as it does in NMBA-induced forestomach carcinogenesis, 16 Cox-2 −/− , Cox-2 +/− , and WT mice on ZD vs . ZS diets were exposed to drinking water containing 20 ppm of NQO for 19 weeks followed by 30 ppm for another 7 weeks. At week 26, ZS: Cox-2 −/− and ZS: Cox-2 +/− mice had significantly lower tongue/forestomach tumor incidence than ZS:WT littermates ( Fig. 1 a , statistical data in Supporting Information Table 1 ). This result is consistent with those reported in nutritionally complete mice showing that Cox-2 absence protects against carcinogenesis. 16 , 21 , 30 Conversely, in ZD mice, genetic Cox-2 did not protect against carcinogenesis. ZD: Cox-2 −/− mice had significantly greater tongue/esophageal tumor incidence than ZD:WT littermates, and ZD: Cox-2 −/− and ZD: Cox-2 +/− mice showed significantly higher tumor multiplicity in all three sites (tongue, esophagus and forestomach) than ZD:WT controls (incidence, Fig. 1 a ; multiplicity, Fig. 1 b , Supporting Information Table 1 ). In addition, ZD led to large tumor size ( Fig. 1 c ) and malignant progression of tongue/esophageal/forestomach tumors in ZD: Cox-2 −/− and ZD: Cox-2 +/− mice compared with ZD:WT mice ( Fig. 1 d ), with statistical significance achieved for tongue SCC (ZD: Cox-2 −/− vs . ZD:WT, 35.7% [5 of 14] vs . 0% [0 of 19], p < 0.01; ZD: Cox-2 +/− vs . ZD:WT, 32.6% [15 of 46] vs . 0% [0 of 19], p < 0.01) ( Fig. 1 a , Supporting Information Table 1 ). These data demonstrated that prolonged ZD abolished the antitumor effect of COX-2 blockade in tongue tumor prevention and elicited tumors in multiple sites with progression to malignancy.
Among mice of the same genetic background, tongue tumor incidence/multiplicity and carcinoma incidence were significantly higher in homozygous ZD: Cox-2 −/− vs . ZS: Cox-2 −/− or heterozygous ZD: Cox-2 +/− vs . ZS: Cox-2 +/− mice, but not in ZD:WT vs . ZS:WT mice ( Fig. 1 a and b , Supporting Information Table 1 ), demonstrating that combined ZD and Cox-2 ablation led to a worse tumor outcome. These results are consistent with and extend our previous study in NMBA-induced forestomach carcinogenesis. 16
ZD per se induces an inflammatory gene signature in ZD:Cox-2 −/− forestomach
To test the hypothesis that ZD promotes carcinogenesis by activating cancer pathways not inhibited by genetic Cox-2 ablation, we performed transcriptome profiling of forestomach mucosa from ZD: Cox-2 −/− , ZS: Cox-2 −/− , ZD:WT and ZS:WT mice ( n = 4/group). We used forestomach rather than tongue because its epithelia can be readily separated from the muscularis layers without enzymatic digestion.
First, we examined the effect of ZD on gene expression changes in Cox-2 −/− forestomach and WT forestomach. By using a cutoff of p ≤ 0.05 and 2-fold difference in expression levels, we found 314 dysregulated probe sets in ZD: Cox-2 −/− vs . ZS: Cox-2 −/− forestomach (Supporting Information Table 2 ) but only 67 in ZD:WT vs . ZS:WT forestomach (Supporting Information Table 3 ). Thus, dietary ZD causes more extensive changes in gene expression in Cox-2 −/− than WT forestomach. A cohort of 36 genes, including the proinflammation mediators S100a8 / a9 , small proline-rich protein 2 Sprr2f / 2h , and keratins Krt6a / 8 / 19 , was common to both class comparisons, indicating that these genes were induced by ZD regardless of genotype.
Next, we compared the effect of Cox-2 deletion on gene expression changes in ZD forestomach and in ZS forestomach. With a cutoff of 2-fold difference, we found 90 dysregulated genes in ZD: Cox-2 −/− vs . ZD:WT forestomach (Supporting Information Table 4 ) but only 17 in ZS: Cox-2 −/− vs . ZS:WT forestomach (Supporting Information Table 5 ). There are no common changes in gene expression between these two class comparisons, and Cox-2 deletion causes fewer changes in ZS than ZD forestomach.
Our qRT-PCR data validated a total 12 selected genes for ZD: Cox-2 −/− vs . ZS: Cox-2 −/− forestomach; six genes for ZD:WT vs . ZS:WT, 7 genes for ZD: Cox-2 −/− vs . ZD:WT, and three genes for ZS: Cox-2 −/− vs . ZS:WT forestomach (Supporting Information Table 6 ).
Among the four class comparisons (Supporting Information Tables 2–5 ), ZD: Cox-2 −/− vs . ZS: Cox-2 −/− forestomach showed the most extensive changes in gene expression, a result consistent with their divergent tumorigenic potential 16 ( Fig. 1 ). Hierarchical clustering analysis of 45,000 transcripts revealed distinct expression patterns ( Fig. 2 a ) between hyperplastic ZD: Cox-2 −/− and nonproliferative ZS: Cox-2 −/− forestomach ( Fig. 2 b ). By further filtering the data using a cutoff of 4-fold difference, we identified a group of 63 genes (62 up- and 1 downregulated; Table 1 ). The most upregulated genes are implicated in the following processes: Sprr2h / 2f and Krt6a / 16 in cytoskeleton metabolism and S100a8 and S100a9 (upregulated 24- and 2.2-fold) in inflammatory/defense/immune responses. Interestingly, S100a8 / a9 were also upregulated 4.2- and 2.4-fold in ZD:WT vs . ZS:WT forestomach (Supporting Information Table 3 ). Because S100a8 / a9 overexpression is associated with ZD-induced rat esophageal preneoplasia, 14 the data that these same genes were upregulated by ZD in hyperplastic ZD: Cox-2 −/− forestomach indicate that they are relevant ZD-induced markers in early forestomach carcinogenesis.
DAVID bioinformatics reveals overrepresentation of inflammatory processes
To define the biological significance of the large lists of differentially expressed genes captured by our transcriptome profiling (Supporting Information Tables 2–5 ), we performed gene ontology functional group analyses for the four class comparisons using DAVID resources. 25 In preneoplastic ZD: Cox-2 −/− vs . ZS: Cox-2 −/− forestomach (Supporting Information Table 7 a ), we found significantly overrepresented biological processes only among the upregulated genes; including in particular, response to external stimulus comprising S100a8 /a9 and 14 genes ( p = 3.98E-004) and response to stimulus comprising S100a8 / a9 and 34 genes ( p = 5.04E-004). Thus, DAVID supports the premise that S100a8 / a9 are relevant markers associated with ZD-induced hyperplasia in ZD: Cox-2 −/− forestomach. Similarly, in ZD:WT vs . ZS:WT (Supporting Information Table 7 b ) and ZD: Cox-2 −/− vs . ZD:WT forestomach (Supporting Information Table 7 c ), significantly overrepresented processes were found only among the upregulated genes, including cytoskeleton and chemotaxis processes. By contrast, in ZS: Cox-2 −/− vs . ZS:WT forestomach (Supporting Information Table 7 d ), significantly overrepresented processes were found only among the downregulated genes that negatively modulated cell cycle and cytoskeleton processes. Together, the data revealed that dietary ZD and sufficiency led to distinct regulated processes in proliferation in Cox-2 −/− forestomach, a finding consistent with the divergent tumorigenic potential of ZD: Cox-2 −/− vs . ZS: Cox-2 −/− forestomach.
IPA reveals a NF-κB—centric network
To understand gene expression interactions in ZD: Cox-2 −/− vs . ZS: Cox-2 −/− forestomach ( Table 1 ) in the context of signaling pathways, we performed pathway analysis using IPA. We identified a nuclear factor (NF)-κB centric network of 35 genes, with 60% of the genes (21 of 35) from the upregulated genes that included S100a8 ( Fig. 2 c ). Because NF-κB is a transcription factor that regulates immune responses/cell proliferation and it is a link between inflammation and cancer development/progression, 31 our result that NF-κB showed connectivity to S100a8 predicted activation of a S100A8-NF-κB inflammatory pathway in ZD: Cox-2 −/− forestomach.
ZD activates S100A8 inflammatory signaling in preneoplastic ZD:Cox-2 −/− forestomach
We focused our study on S100A8 and its heterodimeric partner S100A9 because of their role in inflammation and cancer, 14 , 32 and their prominence among ZD-induced proinflammation markers in ZD: Cox-2 −/− forestomach (Supporting Information Table 7 a ). S100a8 / a9 genes encode the S100 family member calcium binding proteins. Interaction of S100A8/A9 ligands with their receptor RAGE triggers an intracellular NF-κB signaling cascade. 14 , 32
To determine if there is a link between S100A8 overexpression and downstream NF-κB signaling in preneoplastic ZD: Cox-2 −/− vs . ZS: Cox-2 −/− forestomach as predicted by IPA ( Fig. 2 c ), we analyzed expression of five signaling markers, namely, S100A8, S100A9, the RAGE receptor, NF-κB p65, and cyclin D1, by IHC. The cyclin D1 gene is a target of NF-κB activation. 33 IHC showed that all five markers were strongly expressed in hyperplastic ZD: Cox-2 −/− forestomach but weakly expressed in nonproliferative ZS: Cox-2 −/− forestomach ( Fig. 3 a ). The semi-quantitative mean immunoreactive scores of S100A8 and S100A9 protein were significantly higher in ZD: Cox-2 −/− than ZS: Cox-2 −/− forestomach [S100A8, 7.3 (95% CI = 5.5–9.1) vs . 1.6 (95% CI = 0.9–2.3), p < 0.001; S100A9, 6.6 (95% CI = 3.7–9.5) vs . 1.5 (95% CI = 0.9–2.1), p < 0.001, n = 8 mice/group). In addition, ZD: Cox-2 −/− forestomach overexpressed phospho-NF-κB p65, indicating activation and nuclear translocation of NF-κB p65 (Supporting Information Fig. 1 ). These data show that at the earliest stages of forestomach carcinogenesis ZD activates an alternative S100A8 inflammatory pathway not affected by genetic Cox-2 inhibition.
ZR reverses S100A8 inflammatory signaling and hyperplasia in ZD:Cox-2 −/− forestomach
Because ZR attenuates inflammation and reverses hyperplasia in early rat esophageal carcinogenesis, 14 we investigated this effect in ZD: Cox-2 −/− mice one week after switching to a ZS diet. In contrast to ZD: Cox-2 −/− forestomach that showed strong expression of all five S100A8 inflammatory signaling markers ( Fig. 3 a [left column]), ZR: Cox-2 −/− mouse forestomach had reduced or absent immunostaining of the same five markers ( Fig. 3 b ). Additionally, qRT-PCR analysis shows that S100a8 and S100a9 mRNA expression was significantly reduced in ZR: Cox-2 −/− vs . ZD: Cox-2 −/− forestomach ( Fig. 3 c ).
In addition, we determined the rate of cell proliferation in ZR: Cox-2 −/− vs . ZD: Cox-2 −/− forestomach by quantitative PCNA-lHC (PCNA: Fig. 3 b vs . 2 b ). PCNA is an endogenous cell proliferation marker. The PCNA-labeling index (%) was significantly lower in ZR: Cox-2 −/− than ZD: Cox-2 −/− forestomach ( Fig. 3 d ). Together, these data ( Fig. 3 b–d ) document that ZR effectively attenuated S100A8 inflammation and reversed the hyperplastic ZD: Cox-2 −/− phenotype.
ZD upregulates S100a8/a9 expression in preneoplastic ZD:Cox-2 −/− tongue
To determine whether ZD: Cox-2 −/− tongue, which shows high tumorigenic potential as does ZD: Cox-2 −/− forestomach 16 ( Fig. 1 ), also overexpresses the proinflammation genes S100a8 / a9 discovered in forestomach, we determined S100a8 / a9 mRNA and protein expression levels in tongue and forestomach from the four mouse groups (profiling studies) by qRT-PCR and immunoblotting. As in forestomach, S100a8 / a9 mRNA expression was strongest in ZD: Cox-2 −/− tongue, followed by ZD:WT tongue, and negligible in ZS: Cox-2 −/− and ZS:WT tongue ( Fig. 3 e , top). In parallel, S100A8/A9 protein expression was strong in both ZD: Cox-2 −/− tongue and forestomach but weak or absent in similar tissues of other mouse groups ( Fig. 3 e , bottom). These data suggests that in tongue and forestomach ZD activates similar inflammatory pathways that are not affected by COX-2 inhibition.
Activation of S100A8 and p53 inflammatory pathways accompanies malignant tumor progression in ZD:Cox-2 −/− and ZD:Cox-2 +/ − mice
We then went on to investigate whether during malignant tongue/forestomach tumor progression ( Fig. 1 ) S100A8 inflammatory signaling is in fact activated. In addition, we determined whether these carcinomas overexpress PCNA and p53 protein, two prognostic factors in human oral cancers. 34 , 35 The p53 tumor suppressor gene is mutated in approximately 50% of all human cancers, including oral-esophageal cancers 36 and divergent carcinogenic pathways mediated separately by NF-κB and p53 were reported in oral cancer. 37 Using IHC we examined expression of seven markers: PCNA, p53, and five S100A8—NF-κB signaling markers (S100A8, S100A9, RAGE, NF-κBp65 and cyclin D1). We analyzed a total of 15 ZD:Cox- 2 −/− tongue SCC and 6 ZD:Cox- 2 +/− forestomach SCC, as well as non-neoplastic ZS: Cox-2 −/− tongue and ZS: Cox-2 +/− forestomach ( n = 10/group).
ZD: Cox-2 −/− tongue SCC and ZD: Cox-2 +/− forestomach SCC showed high proliferative activity with abundant PCNA-positive nuclei in tumor areas and prominent accumulation of intensely stained p53-positive nuclei. Concurrently, these carcinomas displayed strong co-overexpression of all five S100A8—NF-κB signaling markers ( Fig. 4 ). In addition, these carcinomas overexpressed phospho-NF-κB p65 (Supporting Information Fig. 1 ), indicating activation and nuclear translocation of NF-κB p65. Collectively, these data demonstrate that under complete or partial genetic Cox-2 ablation, ZD stimulated RAGE-S100A8 inflammatory signaling cancer-and p53-associated response pathways, thereby driving malignant tumor progression and bypassing the antitumor effect of COX-2 blockade.
In sharp contrast, non-neoplastic ZS: Cox-2 −/− tongue and ZS: Cox-2 +/− forestomach showed basal cell proliferation with isolated occurrence of p53 protein, as well as low levels of expression of the same five S100A8 signaling markers ( Fig. 4 ), providing evidence that inflammatory pathways were not activated under conditions of COX-2 pathway blockade and ZS that protected against carcinogenesis.
ZR attenuates the inflammation and restores the antitumor effect of COX-2 blockade in cancer prevention
Finally, we investigated whether replenishing Zn can restore the antitumor effect of COX-2 blockade in tumor prevention. In a NMBA-induced forestomach carcinogenesis study, we showed that 14 weeks after ZR, ZR: Cox-2 −/− mice had significantly lower forestomach tumor incidence and multiplicity than ZD: Cox-2 −/− mice ( Fig. 5 a ). In addition, S100a8 / a9 mRNA expression was significantly lower in ZR: Cox-2 −/− vs . ZD: Cox-2 −/− forestomach ( Fig. 5 b ); S100A8/A9 protein expression was absent in ZR: Cox-2 −/− but strongly expressed in ZD: Cox-2 −/− forestomach ( Fig. 5 c ). Histopathologic and IHC studies show that ZR: Cox-2 −/− forestomach mucosa was typically thin, with PCNA-positive nuclei mainly in basal cells and weak to negligible immunostaining of the 4 inflammation-associated markers S100A8, S100A9, p53, and cyclin D1 ( Fig. 5 d ). In contrast, neoplastic ZD: Cox-2 −/− forestomach was highly proliferative, with PCNA-positive nuclei in many cell layers and strong overexpression of the same inflammation-associated markers ( Fig. 5 d ). Thus, ZR reverses preneoplasia ( Figs. 3 b –3 d ), and effectively restores the antitumor effect of Cox-2 ablation ( Fig. 5 ) by attenuating the inflammation. | Discussion
Increasingly cancers are treated with drugs that target specific pathways shown to be of pathogenetic significance. Our study shows that the antitumor effect of genetic disruption of Cox-2 in tongue cancer prevention is bypassed by Zn depletion ( Fig. 1 ), owing to activation of an alternative proneoplastic pathway that is not affected by COX-2 inhibition. Using a combination of techniques that included expression profiling, bioinformatics and investigation of identified markers in ZD: Cox-2 −/− mouse models of oral-esophageal cancers, our data document a mechanism for the inability of COX-2 blockade to prevent tumor growth under ZD conditions.
The hyperplastic ZD: Cox-2 −/− vs . ZS: Cox-2 −/− forestomach has a distinct signature ( Table 1 ). The pro-inflammation mediators S100a8 and S100a9 are upregulated 24-fold and 2.2-fold. In addition, the typical genes of the cornified envelope Sprr2h / 2f and Krt6A / 16 / 17 / 8 / 20 are upregulated 64- to 5.5-fold. Because simultaneous upregulation of S100A8 / A9 , SPRR2 and KRT6A / 16 / 17 is a common feature of human inflammatory skin diseases such as psoriasis 38 and atopic dermatitis, 39 this same signature in ZD: Cox-2 −/− forestomach indicates an association between inflammation and its highly hyperplastic phenotype. Our conclusion that S100a8 / a9 are relevant ZD-induced markers belonging to an inflammatory pathway that drives forestomach cell proliferation rather than an epiphenomenon of this process or of dietary Zn-deficit is supported by DAVID bioinformatics ( Table 7 a ).
S100A8/A9 have emerged as important markers for inflammation-associated cancers. 32 , 40 They are overexpressed in many human cancers, 32 including lung, colorectal, prostate, skin cancer, as well as HPV18-infected oral SCC. 41 The mechanistic role of S100A8/A9 in tumor biology is emerging. In a mouse skin cancer model, Gebhardt et al. 42 provided genetic evidence that S100A8/A9 binds to RAGE, and RAGE signaling sustains skin inflammation and promotes tumorigenesis. In the lung, S100A8/A9 induces the activation of serum amyloid A that activates NF-κB inflammatory signaling and facilitates metastasis. 43 In a colitis-induced mouse cancer model, S100A8/A9 and RAGE augment carcinogenesis 44 and in an inflammation-associated liver cancer model, S100A8/A9 are identified as NF-κB target genes and their overexpression promotes malignant progression. 45 Conversely, blockade of RAGE suppresses tumor growth and metastasis. 42 , 46
Our IHC data in ZD: Cox-2 −/− tongue and ZD: Cox-2 +/− forestomach carcinomas ( Fig. 4 ) demonstrate that under complete or partial blockade of COX-2 pathway dietary ZD activates an alternative cancer-associated RAGE-S100A8 inflammatory pathway. The finding that these same carcinomas showed high PCNA proliferative activity and prominent accumulation of p53 protein indicates that additional inflammation-associated cancer pathways are activated. The p53 tumor suppressor gene is mutated in human oral and esophageal cancer. 36 Mutated p53 protein has a prolonged half-life that leads to its accumulation in the nucleus. In this regard, human head and neck squamous cell cancer (HNSCC), which is a highly inflammatory, proliferative and aggressive cancer, 47 exhibits high levels of p53 expression, abundant cell proliferative activity, 34 , 35 as well as divergent carcinogenic pathways mediated separately by NF-κB and p53. 37
Chronic inflammation contributes to the development of ∼20% of all human cancers. The causes of inflammation are often unknown. 48 Our recent report in rat esophagus that dietary Zn regulates S100A8 expression and modulates the link between S100A8-RAGE and downstream NF-κB/COX-2 provides the first evidence that Zn has an inflammation-modulating role in essophageal cancer initiation/reversal. 14 Here we demonstrate that with COX-2 pathway blockade prolonged dietary ZD causes chronic inflammation in the tongue/forestomach by activating alternative inflammatory RAGE-S100A8/A9 and p53 response pathways, thereby fueling tumor progression and bypassing the antitumor effect of Cox-2 deletion. These new data provide a likely mechanism to explain the inefficacy of such targeted cancer therapy in oral-cancer patients, since many of these patients are frequently Zn-deficient. 4 – 7
Recent studies reported that Zn supplementation improves clinical outcomes in patients receiving radiotherapy for HNSCC, 49 as well as concomitant chemotherapy and radiotherapy for advanced nasopharyngeal carcinoma. 50 The present finding that ZR attenuates the inflammatory response and restores the antitumor effect of COX-2 blockade has important clinical implications. Thus, stratification of patients by Zn status would be useful, and a personalized cancer therapeutic paradigm that includes Zn may improve efficacy. | Additional Supporting Information may be found in the online version of this article.
Re-use of this article is permitted in accordance with the Terms and Conditions set out at http://wileyonlinelibrary.com/onlineopen#OnlineOpen_Terms .
Grant sponsors: NIH Grants R01CA118560 (LYYF) and R01 CA115965 (CMC)
S.-G. W. and C. T. contributed equally to this work
Zinc (Zn)-deficiency (ZD) is implicated in the pathogenesis of human oral-esophageal cancers. Previously, we showed that in ZD mice genetic deletion of cyclooxygenase-2 ( Cox-2 ) enhances N -nitrosomethylbenzylamine-induced forestomach carcinogenesis. By contrast, Cox-2 deletion offers protection in Zn-sufficient (ZS) mice. We hypothesize that ZD activates pathways insensitive to COX-2 inhibition, thereby promoting carcinogenesis. This hypothesis is tested in a Cox-2 −/− mouse tongue cancer model that mimics pharmacologic blockade of COX-2 by firstly examining transcriptome profiles of forestomach mucosa from Cox-2 −/− and wild-type mice on a ZD vs . ZS diet, and secondly investigating the roles of identified markers in mouse forestomach/tongue preneoplasia and carcinomas. In Cox-2 −/− mice exposed to the tongue carcinogen 4-nitroquinoline 1-oxide, dietary ZD elicited tongue/esophagus/forestomach carcinomas that were prevented by ZS. The precancerous ZD: Cox-2 −/− vs . ZS: Cox-2 −/− forestomach had an inflammatory signature with upregulation of the proinflammation genes S100a8 and S100a9 . Bioinformatics analysis revealed overrepresentation of inflammation processes comprising S100a8 / a9 and an nuclear factor (NF)-κB network with connectivity to S100A8. Immunohistochemistry revealed co-overexpression of S100A8, its heterodimeric partner S100A9, the receptor for advanced glycation end-products (RAGE), NF-κB p65, and cyclin D1, in ZD: Cox-2 −/− forestomach/tongue preneoplasia and carcinomas, evidence for the activation of a RAGE-S100A8/A9 inflammatory pathway. Accumulation of p53 in these carcinomas indicated activation of additional inflammatory pathways. Zn-replenishment in ZD: Cox-2 −/− mice reversed the inflammation and inhibited carcinogenesis. Thus, ZD activates alternative inflammation-associated cancer pathways that fuel tumor progression and bypass the antitumor effect of Cox-2 ablation. These findings have important clinical implications, as combination cancer therapy that includes Zn may improve efficacy. | Oral-esophageal squamous cell carcinomas (SCCs) are a major cause of cancer deaths worldwide. 1 Oral cancer, the major site being the tongue, causes a high mortality rate because of frequent development of a second primary esophageal cancer because of field cancerization effects. 2 Risk factors include alcohol consumption, tobacco and human papillomavirus (HPV). 3 The incidence of oral cancer is increasing, particularly in young adults without documented risk factors. 3 Epidemiologic and clinical studies have long implicated zinc (Zn)-deficiency (ZD) in the pathogenesis of oral-esophageal cancers in many populations. 4 – 6 ZD is associated with increased tumor size and poor disease prognosis. 4 , 7
Zn is required for the activity of many enzymes, for proper immune function, and for the conformation of many transcription factors that control cell proliferation, apoptosis, and signaling pathways. 8 , 9 Zn is known to undergo rapid ligand exchange reactions and is used as an information carrier in signal transduction pathways similar to calcium. 10 Consequently, ZD predisposes to disease by adversely affecting immune system, by increasing oxidative stress, and by increasing the generation of inflammatory cytokines. 11 Although the role of ZD as a causative factor of disease and as a determinant in disease progression is gaining attention, 12 the mechanisms underlying its protumorigenic effect, however, remain unclear.
In the rat, a ZD diet creates a precancerous condition in the upper digestive tract, including tongue, esophagus and forestomach (an expanded lower esophagus), by inducing proliferation 13 and gene expression changes, including overexpression of cyclooxygenase-2 ( Cox-2 ) and the proinflammation-genes S100 calcium binding protein a8 ( S100a8 ) and a9 ( S100a9 ). 13 , 14 ZD rats rapidly develop esophageal tumors after a single exposure to the environmental carcinogen N -nitrosomethylbenzylamine (NMBA) 15 and concurrent tongue, esophageal and forestomach tumors with exposure to the tongue carcinogen 4-nitroquinoline 1-oxide (NQO). 13 Zn-replenishment (ZR) reverses cell proliferation, corrects gene expression and inhibits carcinogenesis. 14 – 16
Targeted therapies that block molecules crucial to tumor growth are being explored in attempts to prevent or cure cancer. 17 The rationale for targeting the COX-2 pathway is supported by numerous studies. COX-2 is overexpressed in many human cancers, including esophageal and tongue SCC. 18 , 19 COX-2 catalyzes the formation of prostaglandins and is induced by factors implicated in carcinogenesis, including growth factors, inflammatory stimuli, oncogenes and tumor promoters. 20 The report that deletion of the Cox-2 gene in Apc knockout mice greatly reduces intestinal polyp formation provides genetic evidence that COX-2 plays a key role in tumorigenesis. 21 COX-2 selective inhibitors, celecoxib in particular, are being tested in clinical trials for the prevention of several cancers, 22 including esophageal cancer. 23 Although such targeted therapies have shown promising results in several cancers, their efficacy in oral-esophageal cancer has been limited. 24
Our previous work showed that in ZD rats pharmacologic COX-2 inhibition by the drug celecoxib did not prevent tongue carcinogenesis, and in ZD mice genetic Cox-2 deletion actually enhanced NMBA-induced forestomach tumorigenesis. 16 Aside from the result that ZD: Cox-2 −/− mouse forestomach overexpressed leukotriene A 4 hydrolase protein, indicating a shift of arachidonic acid to the 5-lipoxygenase pathway, mechanisms underlying this effect of ZD were not elucidated. We hypothesized that ZD adversely affects treatment outcome by stimulating pathways not inhibited by the pharmacologic blockade of COX-2. We tested this hypothesis in a ZD: Cox-2 −/− mouse oral-esophageal cancer model that mimics pharmacologic COX-2 blockade, using techniques that included transcriptome profiling, bioinformatics analyses, and investigation of the pathobiological roles of identified markers in murine tongue/forestomach preneoplasia and neoplasia. | Abbreviations
cyclooxygenase-2
database for annotation, visualization and integrated discovery
gene ontology
head and neck squamous cell cancer
immunohistochemistry
ingenuity pathway analysis
N -nitrosomethylbenzylamine
4-nitroquinoline 1-oxide
quantitative reverse transcriptase-polymerase chain reaction
receptor for advanced glycation end products
S100 calcium binding protein A8
S100 calcium binding protein A9
squamous cell carcinoma
nuclear factor-κB
proliferating cell nuclear antigen
zinc
Zn-deficiency
Zn-replenishment
Zn-sufficiency
Supplementary material | CC BY | no | 2022-01-12 15:46:50 | Int J Cancer. 2011 Jul 15; 129(2):331-345 | oa_package/ed/74/PMC3015018.tar.gz |
||
PMC3015047 | 21234374 | Osteoporosis is a common skeletal disease that increases the risk of fracture, with serious clinical and economic consequences. It can be diagnosed by dual-energy X-ray absorptiometry (DXA), even before a fracture has occurred, using the World Health Organization (WHO) criteria according to the patient's T-score. The WHO has also developed a fracture risk assessment tool (FRAX) to estimate the 10-year probability of major osteoporotic fracture (clinical spine, hip, proximal humerus, and distal forearm) and hip fracture, using clinical risk factors for fracture and femoral neck bone mineral density (BMD), if available. Cost-effective pharmacological agents that have been proven to reduce fracture risk in patients at high risk for fracture are now widely available. However, despite great progress in the management of osteoporosis, it remains a disease that is underrecognized and undertreated; if treatment is started, persistence is often poor, with only about 50% of patients who are prescribed medication for osteoporosis still taking it 1 year later. Even when treatment is taken correctly and for a sufficient length of time for the patient to benefit from reduction in fracture risk, there may nevertheless be limitations in effectiveness (note the lack of evidence for reduction in the risk of hip fractures or other nonvertebral fractures with some agents), limitations in the duration of therapy (e.g., no more than 24 months of lifetime teriparatide in the US), and concerns regarding long-term safety, such as atypical femur fractures and osteonecrosis of the jaw with bisphosphonates. For all of these reasons, the goal of reducing the global burden of osteoporotic fractures is not being fully achieved.
This special issue of the Journal of Osteoporosis describes new and emerging approaches to treatment that offer the potential to reduce the risk of fractures or manage their consequences better than what is currently observed in clinical practice. In recent years, our understanding of the pathophysiology of osteoporosis and the regulation of bone remodeling at the molecular level have undergone tremendous advances, leading to the investigation of drugs that target specific molecules in order to modulate the bone remodeling process. For example, the discovery that receptor activator of nuclear factor kappa B ligand (RANKL) is the principal regulator of osteoclastic bone resorption led to the development of denosumab, a fully human monoclonal antibody to RANKL. This potent antiresorptive agent, administered as a 60 mg subcutaneous injection every 6 months, recently received regulatory approval for the treatment of women with postmenopausal osteoporosis (PMO) at high risk for fracture. It has been shown to increase BMD, reduce bone turnover marker levels, and reduce the risk of vertebral fractures, hip fractures, and nonvertebral fractures in women with PMO.
Wnt signaling initiated by the binding of Wnt proteins to the low density lipoprotein-related protein (LRP5/6)-frizzled receptor complex has recently been recognized as an important upregulator of osteoblastic bone formation; sclerostin and Dickkopf-1 (DKK-1) are natural inhibitors of Wnt signaling. In this issue, J. J. Mason and B. O. Williams describe a rare genetic disorder, sclerosteosis, resulting from a mutation of the SOST gene that encodes for sclerostin, and van Buchem disease, a related disorder caused by a mutation closely linked to SOST on chromosome 17q11.2. Patients with sclerosteosis and van Buchem disease have high bone mass due to downregulation of sclerostin, suggesting that a therapeutic agent that downregulates sclerostin in a controllable fashion might be a potent osteoanabolic treatment for patients with osteoporosis. Mason and Williams review many of the studies that have enhanced our understanding of the regulators of Wnt signaling and lead to the investigation of compounds with potential therapeutic applications through their effects on sclerostin or DKK1. A fascinating new finding, yet to be fully elucidated, is that serotonin produced by enterochromaffin cells in the duodenum also downregulates Wnt signaling, raising the possibility that modulation of serotonin production or activity might also be an effective treatment for patients with osteoporosis. In a related paper by S. Silverman in this issue, the preclinical and clinical studies of sclerostin inhibition are presented.
The drugs used to treat osteoporosis are generally considered to be in 1 of 2 categories—antiresorptive (e.g, bisphosphonates) or osteoanabolic (e.g., teriparatide). Interestingly, some drugs may “uncouple,” at least in part, the closely related processes of bone resorption and formation. Strontium ranelate may be such a drug. Another, perhaps, is odanacatib, an investigational agent that inhibits cathepsin K, a protease produced by osteoclasts that is largely responsible for the degradation of the bone collagen matrix. J. L. Perez-Castrillon et al. review what is now known about the role of cathepsin K in health and disease, followed by data from phase 1 and phase 2 clinical trials with odanacatib. This drug is currently under investigation in a large phase 3 clinical trial to evaluate antifracture efficacy in women with PMO. Other papers in this issue cover new developments concerning skeletal heath in areas as diverse as bisphosphonate nanoparticles, melatonin, and thalassemia.
The papers in the issue were selected from many excellent submissions. We give our thanks to all authors of these submissions and to the numerous reviewers who kindly donated their time and expertise in helping select and revise those published here.
E. Michael Lewiecki E. Michael Lewiecki Manuel Diaz Curiel Manuel Diaz Curiel Joao Lindolfo Borges Joao Lindolfo Borges Annie Kung Annie Kung Maria Luisa Brandi Maria Luisa Brandi Hans Peter Dimai Hans Peter Dimai | CC BY | no | 2022-01-13 02:54:45 | J Osteoporos. 2010 Dec 26; 2010:318320 | oa_package/2f/69/PMC3015047.tar.gz |
|||||||
PMC3015048 | 21209746 | 1. Introduction
The volatile anesthetic drug isoflurane reduces respiratory resistance and can be applied to treat severe bronchospasm [ 1 , 2 ]. Though isoflurane is typically administered only for a short period for general anesthesia, it may also be a viable treatment option for severe bronchospasm and for sedation during critical care. However, little experience has been had with long-term treatment of isoflurane (>1 week) or use in patients with traumatic brain injury due to concerns regarding renal toxicity and increased intracranial pressure.
This case report illustrates the successful long-term application of isoflurane in a critically ill patient with several underlying comorbidities. | 3. Discussion
In our patient suffering from several morbidities, isoflurane allowed to treat live-threatening bronchospasm and provided long-term sedation. Bronchospasm was most likely caused by HSV-1 pneumonia. While chest X-ray and CT scan did not show alveolar infiltrates typical for bacterial pneumonia, pulmonary function was severely compromised due to airway hypersensitivity. High concentrations of HSV-1 DNA was detected in BAL, but no other sign of bacterial or fungal infection was found after microbiological cultures and DNA analyses.
Patients treated with immunosuppressive drugs are especially prone to HSV pneumonia and other viral infections. In patients who have HSV pneumonia and are undergoing mechanical ventilation, this is frequently associated with deteriorated lung function and weaning failure, even in the presence of normal X-ray findings [ 4 , 5 ]. HSV-1 pneumonia is associated with markedly decreased survival during ICU treatment [ 6 – 8 ]. When Linssen et al. examined patients with a positive BAL for HSV-1, they found a considerably increased 14-day mortality rate for patients with an HSV-1 load larger than 10 5 ge mL −1 [ 7 ]. In their study, mortality rose from 20% to 41% in patients with a viral load above that cutoff value. In some cases, HSV pneumonia may lead to severe obstructive lung disease.
Our case demonstrates the successful application of isoflurane in the ICU, for a prolonged period of time, for treatment of refractory bronchospasm. Though our patient had no history of asthma, the underlying HSV-1 pneumonia was associated with severe bronchospasm during mechanical ventilation. Over the course of the disease, intravenous sedative drugs did not achieve adequate sedation level. Especially during positioning and nursing care maneuvers, severe bronchospasm caused life-threatening cardiac depression. The application of bronchodilatory drugs had no beneficial effect.
In patients with traumatic brain injury, control of arterial pCO 2 levels is of upmost importance and excludes permissive hypercapnia as a treatment option. In this case, the additional application of the volatile anesthetic isoflurane enabled controlled mechanical ventilation and a stable control of pulmonary gas exchange. Finally, isoflurane treatment enabled successful bronchoscopy and BAL which resulted in diagnosis of the underlying infection.
While concerns have been raised about isoflurane possibly causing increased intracranial pressure, it has been shown during neurosurgical procedures that when applied below expired concentrations of 1.2%, isoflurane does not lead to any increase in intracranial pressure [ 9 , 10 ]. This makes these drugs suitable for sedation in traumatic brain injury.
Though volatile anesthetics have been shown to be effective in acute asthmatic bronchospasm, data on prolonged application is rare [ 11 ]. A common concern against the prolonged use of volatile anesthetics is their potential harm to renal function. During short-term use for surgical procedures, several perioperative studies could not find a deterioration of renal function in patients with pre-existing renal insufficiency after isoflurane anesthesia, but data for long-term application for more than one week is limited [ 12 – 15 ]. Sackey and coworkers compared isoflurane sedation in the ICU to sedation by midazolam in 40 patients for an observation period up to 96 hours [ 16 ]. In that study, sufficient sedation using isoflurane was reached without occurrence of hepatic or renal adverse events. In particular, patients receiving immunosuppressive therapy are known to have impaired renal function and decreased creatinine clearance. Our patient with a 12-year history of immunosuppression had experienced his first incident of acute renal failure one year prior to this event during diarrhea, and during our care he had experienced another episode after cardiopulmonary resuscitation which required renal replacement therapy. When isoflurane therapy was started, the patient was already recovering from acute renal failure, and renal replacement therapy had been stopped. Application of isoflurane did not deteriorate renal function and the time course of serum creatinine levels, blood urea, and creatinine clearance during isoflurane therapy can be seen in Figure 1 . At discharge, creatinine clearance was in normal range (75 mL/min).
Of note in this case was the use of the AnaConDa system for isoflurane delivery [ 17 ]. While most modern ventilators allow the appliance of sophisticated respiratory therapeutic regimes, they usually have no delivery method for volatile anesthetics. The AnaConDa system—which is essentially a modified heat-moisture exchanger—is directly connected to the patient's endotracheal tube and can be combined with any respirator. The device is approved for use with either isoflurane or sevoflurane, which are delivered by a syringe. Sackey and coworkers reported a remarkably lower consumption of isoflurane using the AnaConDa delivery method than from traditional vaporizer techniques and an environmental pollution below the internationally recommended long-term exposure limits which make side effects on ICU personnel unlikely [ 18 ].
While isoflurane, in this case, was able to control the patient's pulmonary situation and secure sufficient ventilation, the volatile anesthetic could not be safely withdrawn until the primary source of the bronchospasm was treated. Here, by day 6 of acyclovir antiviral therapy, inspiratory concentrations of isoflurane were able to be reduced stepwise, and isoflurane treatment was able to be stopped one day later.
To conclude, our case shows the successful long-term application of isoflurane, for a 12-day period, in a patient with severe HSV-1 pneumonia. The application of the volatile anesthetic permitted the therapy of refractory bronchospasm, enabled diagnosis of HSV-1 pneumonia, helped to control pCO 2 levels, and led to sufficient sedation. The use of isoflurane did not aggravate pre-existing renal insufficiency and was safely performed in a patient with traumatic brain injury. | Academic Editor: Piotr K. Janicki
A 47-year-old man with a history of heart transplant was admitted after severe traumatic brain injury and seizures. During mechanical ventilation, the patient developed bronchospasm that severely compromised respiratory function that led to cardiac arrest. After resuscitation, application of isoflurane through the Anaesthetic Conserving Device (AnaConDa) in the ICU successfully treated bronchospasm, provided adequate sedation, and enabled appropriate ventilation and diagnostic bronchoscopy. A subsequent bronchoalveolar lavage revealed a high amount of Herpes simplex DNA. Herpes simplex pneumonia was diagnosed and treated with acyclovir. Isoflurane treatment was applied for twelve days total without side effects on renal and cerebral function. The patient recovered quickly after the termination of sedation. At discharge, he was fully awake without focal neurological deficiency and his long-term outcome was excellent. This case demonstrates that isoflurane is a treatment option in life-threatening cases of bronchospasm and a safe option for long-term sedation. | 2. Case Report
A fully conscious 47-year-old man was transferred to the hospital after collapse and severe head injury. Twelve years prior to this incident, the patient underwent heart transplantation for treatment of dilatative cardiomyopathy. At admission, computer tomography (CT) revealed traumatic brain injury with an occipital skull fracture, left frontal contusion bleedings, and a traumatic subarchnoid hemorrhage. On the second day of hospitalization, he suddenly suffered from general seizures associated with severe arterial hypotension. After intubation and resuscitation, the patient was treated with anticonvulsants and sedative drugs. A new CT scan excluded any new cerebral injuries or bleedings.
The patient was then transferred to the ICU. In the presence of fever and clinical and laboratory signs of systemic inflammation, calculated antimicrobial treatment with clindamycin and ciprofloxacin was initiated. Routine clinical, radiological, and microbiological examinations, however, did not reveal any sources of infection. An attempt to reduce sedation to better assess neurological status resulted in severe bronchospasm and impaired lung function as indicated by wheezing and a markedly prolonged expiratory phase. Dynamic hyperinflation occurred and finally led to hypoxia and to cardiac arrest. Cardiopulmonary resuscitation was performed, and lung function was compromised with a paO2/FiO2 ratio of 225 at high airway pressures. Consequently, increasing doses of sedatives became necessary and a combination of midazolam (60 mg · h −1 ), ketamin (200 mg · h −1 ), clondine (120 μ g · h −1 ), and sufentanil (75 μ g · h −1 ) was needed to protect the patient from any stimuli during positioning and to prevent additional severe events of bronchospasm. We administered the bronchodilators salbutamol (1.5 mg) and ipratropium bromide (500 μ g) as aerosole three times per day, magnesium sulfate (20 mmol) intravenously and theophylline (400 mg loading dose over 20 min, then 40 mg · h −1 ) intravenously. Theophylline dose was then adjusted to maintain therapeutic blood levels between 10 and 20 mg · L −1 [ 3 ]. Both chest X-ray and spiral chest CT scan showed small pleural effusion but again no sign of pneumonia. Repeated microbiological examinations and cultures of tracheal aspirates, urine, or catheter tips did not document any pathogens. A diagnostic bronchoscopy for bronchoalveolar lavage (BAL) was halted due to life-threatening episodes of bronchial obstruction and bronchospasm.
On day 12 after admission, recurrent episodes of bronchospasm led to an increase of arterial pCO 2 to 63 mmHg. At this point, still unable to provide adequate sedation, we decided to start inhalative isoflurane therapy with the Anaesthetic Conserving Device Conserving Device (AnaConDa, Sedana Medical, Sundbyberg, Sweden). Inspiratory gas concentration was set to 0.6% which led to Ramsay-score of 5 to 6 points. This allowed appropriate ventilation and adjustment of pCO 2 to a concentration of 38 mmHg. IV sedation could then be reduced stepwise. Bronchoscopy was repeated successfully and revealed rigid mucus and a highly sensitive bronchial tree that reacted to mild stimuli with bronchoconstruction. Despite this impediment, we performed a bronchoalveolar lavage in the right band, left lower lobes (Segments 7 or 8, resp.). Microbiological cultures of the lavage did not reveal bacterial or fungal growth and polymerase chain reaction (PCR) for cytomegalovirus DNA was also negative. However, PCR for Herpes simplex Typ 1 (HSV-1) virus DNA revealed a high load of 9,300,000 genomic equivalents (ge) mL −1 . In the throat lavage 19,000,000 ge mL −1 , HSV-1 was found.
Following this discovery, antiviral treatment with acyclovir was initiated. Dosage was adjusted to 2 × 800 mg i.v. according to renal function. Over the next few days, both intensity and frequency of bronchospasm decreased. Six days after the beginning of the antiviral treatment, isoflurane was safely reduced in a stepwise manner and was stopped after a 12-day treatment period. Over the next few weeks, HSV-1 DNA concentrations consistently decreased in throat lavage to 6,200,000 ge mL −1 on day 8 and to 250,000 ge mL −1 on day 17 of acyclovir treatment. The patient had had a history of impaired renal function and suffered from acute renal failure on day 5 after admission, which made renal replacement therapy necessary from hospital day 5 to 10. However, when isoflurane was initiated, renal function had already recovered ( Figure 1 ), and the patient did not require renal replacement therapy any longer.
Neurologically, the patient recovered quickly after the termination of sedation. At discharge, he was fully awake and had no focal neurological deficiency.
Tracheostomy was performed on hospital day 19 to facilitate weaning from the ventilator. On day 39, the patient was eased off mechanical ventilation. He was transferred to a rehabilitation centre 40 days after hospital admission. At discharge, he was fully awake and had no focal neurological deficiency.
Seven months after the event, the patient was discharged from rehabilitation with no neurological deficiencies. Long-term outcome (as assessed by phone call) was excellent. | Conflict of Interests
The authors declare to have no conflict of interests. | CC BY | no | 2022-01-13 01:48:13 | Case Rep Med. 2010 Dec 16; 2010:746263 | oa_package/d2/60/PMC3015048.tar.gz |
|||
PMC3015055 | 20579004 | INTRODUCTION
Ghrelin, a gastric hormone, exerts orexigenic and pro-obesity effects by interacting with key brain circuits involved in appetite and energy balance ( Tschöp, Smiley & Heiman 2000 ; Wren et al . 2000 ; Cummings et al . 2001 ; Nakazato et al . 2001 ). The cloned receptor for ghrelin (GHS-R1A) is present, however, not only in discrete hypothalamic cell groups regulating energy balance but also in a number of other central nervous system (CNS) sites, such as the hippocampus, brainstem and reinforcement areas ( Guan et al . 1997 ) implying a role for ghrelin in brain reinforcement. Thus, ghrelin activates a key mesolimbic circuit involved in natural as well as drug-induced reinforcement, the cholinergic-dopaminergic reward link ( Jerlhag et al . 2006 , 2007 , 2008 ). This link encompasses the well-described dopamine projection from the ventral tegmental area (VTA) to the nucleus accumbens (N.Acc.) that forms part of the mesolimbic dopamine system, together with a cholinergic projection from the laterodorsal tegmental area (LDTg) to the VTA. By activating this reinforcement link, ghrelin may increase the incentive value of motivated behaviours such as food and drug seeking ( Abizaid et al . 2006 ; Jerlhag et al . 2006 ; Jerlhag et al . 2009 ). Indeed, alcohol reinforcement was absent in pharmacological and genetic models of suppressed ghrelin signalling ( Jerlhag et al . 2009 ).
Peripherally injected ghrelin also stimulates the mesolimic dopamine system ( Jerlhag 2008 ; Quarta et al . 2009 ), hypothesizing that peripherally produced ghrelin reaches deeper brain structures. Moreover, an important premise for the study is that the afferents regulate ghrelin's effects on the mesolimbic dopamine system involving GHS-R1A in VTA. In the present experiments therefore we sought to determine whether VTA administration of a GHS-R1A antagonist suppresses the locomotor stimulatory and accumbal dopamine releasing effects of peripheral ghrelin. The activity of VTA dopamine neurons is modulated by various afferents to the VTA including glutamate, opioids and orexin ( Kalivas, Churchill & Klitenick 1993 ; Wise 2002 ). Previously, a role for NMDA, opioid as well as orexin A receptors in different aspects of ghrelin-induced activation of the mesolimbic dopamine systems has been suggested ( Toshinai et al . 2003 ; Naleid et al . 2005 ; Abizaid et al . 2006 ). Inspired by the possibility to suppress natural and chemical drug reinforcement by agents that interrupt central ghrelin signalling, we also sought to determine whether ghrelin's ability to activate the mesolimbic dopamine system, as measured by locomotor stimulation and accumbal dopamine release, can be interrupted by pharmacological suppression of glutamatergic, opioid and orexin A systems. | MATERIALS AND METHODS
Animals
Adult post-pubertal age-matched male NMRI mice (8–12 weeks old and 25-30 g body weight; B&K Universal AB, Sollentuna, Sweden) were used for studies of locomotor activity and dopamine release as such studies are well-documented in this strain ( Jerlhag et al . 2006 ; Jerlhag et al . 2007 ; Jerlhag et al . 2008 ). Upon arrival the mice were allowed to habituate in groups of eight in standard cages (Macrolon III: 400 × 250 × 150 mm), for at least one week before initiation of the experiment. All mice were maintained at 20°C with 50% humidity and a 12/12 hour light/dark cycle (lights on at 7 am ). Tap water and food (Normal chow; Harlan Teklad, Norfolk, England) were supplied ad libitum , except during the experiments. All experiments were conducted during the day time when the mice are less active. Studies were approved by the Ethics Committee for Animal Experiments in Gothenburg, Sweden.
Drug administration
Acylated rat ghrelin (Bionuclear; Bromma, Sweden) was diluted in 0.9% sodium chloride (saline vehicle) and was administrated intraperitoneally (i.p.) (10 ml/kg body weight). The selected dose, 0.33 mg/kg, was determined previously, as it increases locomotor activity and accumbal dopamine release as well as induces a conditioned place preference in mice ( Jerlhag 2008 ). Ghrelin was administered 10 minutes prior to the initiation of the experiment (locomotor activity or microdialysis).
The dose of BIM28163 (Ipsen Biomeasure Inc, Milford, MA, USA), a GHS-R1A antagonist, has also been determined previously ( Halem et al . 2004 ; Jerlhag et al . 2009 ). BIM28163 was diluted in Ringer solution (NaCl 140 mM; CaCl 2 1.2 mM; KCl 3.0 mM and MgCl 2 1.0 mM) (Merck KgaA, Darmstadt, Germany) and was administered at a dose of 2.5 μg/side (uni- or bilaterally into the VTA) at 40 minutes prior to i.p. ghrelin/vehicle exposure. Previous studies have established that this compound is a GHS-R1A antagonist and fully inhibits ghrelin-induced GHS-R1A activation ( Halem et al . 2004 ).
The selected dose of AP5 (Sigma-Aldrich, Stockholm, Sweden), an N-methyl-D-aspartic acid (NMDA) receptor antagonist, was determined in a dose-response study where 0.5 μg/side (uni-or bilaterally into the VTA) was the highest dose not to affect locomotor activity per se ( Fig. 1 ). AP5 or Ringer vehicle were administered 10 minutes prior to i.p. ghrelin/vehicle administration. AP5 does not affect nicotinic acetylcholine receptors in the CNS ( Davies & Watkins 1982 ).
The selected dose of SB334867 (Tocris, Bristol, United Kingdom), an orexin A receptor antagonist, was determined in a dose-response study where 5 μg/side (bilaterally into the VTA) was the highest dose not to affect locomotor activity per se (data not shown). Doses in a similar range have previously been shown to block the cue-induced reinstatement of cocaine seeking ( Smith, See & Aston-Jones 2009 ). SB334867 or vehicle (10%-DMSO in Ringer vehicle; Merck KgaA) were administered 10 minutes prior to i.p. ghrelin/vehicle exposure.
Naltrexone, an unselective opioid receptor antagonist with some selectivity to the μ receptor, was diluted in saline vehicle. Naltrexone (1 mg/kg, i.p.) or saline vehicle were injected 30 minutes prior to i.p. ghrelin/vehicle. The dose was determined from previous studies in which doses in a similar range have been shown to block the reinforcing properties of alcohol in rodents ( Herz 1997 ). The rationale for administering by the i.p. route is that direct mesolimbic effects of nalrexone to interrupt ghrelin-induced reinforcement are unlikely, based on previous studies in which this antagonist had no effect on ghrelin-induced food intake when administered into discrete mesolimbic sites ( Naleid et al . 2005 ).
Intra-VTA injections were made using a volume of 0.5 μl/side via chronically implanted catheters, over 60 seconds. A 5 μl syringe (Kloehn, microsyringe; Skandinaviska Genetec AB, V. Frölunda, Sweden) was used to facilitate drug administration into the VTA. After each injection, the cannula was left in place for a further 60 seconds to facilitate diffusion. Potentially, the volume administered during intra-VTA injections could raise concerns about specificity due to possible leakage into neighboring structures. Previously, however, we found that only ‘on-target’ placements (in the VTA and LDTg but not in closely adjacent sites) resulted in significant effects of ghrelin or GHS-R1A antagonists on locomotor stimulation and dopamine release using a larger volume (1 μl) for mice ( Jerlhag et al . 2007 , 2009 ). Supportively, in the present study neither AP5 nor BIM28163 blocked ghrelin-induced locomotor stimulation nor accumbal dopamine release in a few mice in which the canulae were misplaced in neighbouring structures (data not shown). All drug challenges were part of a balanced design with regard to both the treatment order and the number of subjects per treatment.
Locomotor activity experiments
Peripheral administration of ghrelin has previously been shown to stimulate locomotor activity in mice ( Jerlhag 2008 ). Locomotor stimulation is, at least in part, mediated by an increase in the extracellular concentration of accumbal dopamine ( Engel et al . 1988 ). Locomotor stimulation has been suggested to be an homologous effect evolving from a common mechanism involving the mesolimbic dopamine system, implying that locomotor activity reflects reinforcement induced by drugs of abuse ( Imperato & Di Chiara 1986 ; Wise & Bozarth 1987 ; Engel et al . 1988 ). Thus, accumbal dopamine measurement experiments were conducted only after first establishing that aforementioned antagonist compounds suppress ghrelin-induced locomotor stimulation. All mice, except those treated with naltrexone, were implanted with bilateral guide cannulaes aiming at the VTA. The mice were anesthetized with isofluran (Isofluran Baxter; Univentor 400 Anaesthesia Unit, Univentor Ldt., Zejtun, Malta), placed in a stereotaxic frame (David Kopf Instruments; Tujunga, CA, USA) and kept on a heating pad to prevent hypothermia. The scull bone was exposed and two holes for the guide cannulas (stainless steel, length 10 mm, with an o.d./i.d. of 0.6/0.45 mm) and one for the anchoring screw were drilled. The VTA coordinates were 3.4 posterior to bregma, ±0.5 mm lateral to the midline and 1.0 mm below the brain surface ( Franklin & Paxinos 1996 ). All guide cannulae were surgically implanted four days prior to the experiment. After surgery the mice were kept in individual cages (Macrolon III). At the time of the experiment, a cannula for drug administration was inserted and extended another 3.8 mm ventrally beyond the tip of the guide cannula, aiming at the VTA.
Locomotor activity was registered in eight sound attenuated, ventilated and dim lit locomotor boxes (420 × 420 × 200 mm, Kungsbacka mät- och reglerteknik AB, Fjärås, Sweden). Five by five rows of photocell beams, at the floor level of the box, creating photocell detection allowed a computer-based system to register the activity of the mice. Locomotor activity was defined as the accumulated number of new photocell beams interrupted during a 60-minute period.
Before initiating the experiments, a dummy cannula was carefully inserted and retracted into the guide cannula to remove clotted blood and hamper the spreading depression. Mice were then allowed to habituate to the locomotor activity box one hour prior to drug challenge. In separte experiments, the effects of i.p. administered ghrelin on locomotor stimulation was investigated following intra-VTA administration of BIM28163, AP5 or SB334867 to mice. In subsequent experiments, naltrexone was injected i.p. prior to ghrelin. In the first locomotor activity experiment, BIM28163 (2.5 μg/side) or an equal volume (0.5 μl/side) of vehicle solution (Ringer) was administered locally and bilaterally into the VTA. Ghrelin (0.33 mg/kg) or an equal volume of vehicle solution (saline vehicle) was thereafter injected. The same experimental protocol was used for AP5 (0.5 μg/side), SB334867 (5 μg/side) and naltrexone (1 mg/kg, i.p.). All mice received drug treatment only twice (antagonist/vehicle and ghrelin/vehicle). Neither water nor food was available to the mice during the locomotor experiments. The activity registration started five minutes after the last injection and was subsequently measured for a 60-minute period. For intra-VTA administration only mice with guide cannulae placements in the VTA were included in the statistical analysis.
In vivo microdialysis and dopamine release measurements
For measurements of extracellular dopamine levels and overflow (that reflect dopamine release), mice were implanted unilaterally with a microdialysis probe positioned in the N.Acc. shell and a guide cannulae into the VTA. The probe and the guide cannula/e were positioned ipsilateral, and the location was randomly alternated to either the left or right side. The surgery was preformed two days prior to the experimental day as described above (see Locomotor activity experiments ). The coordinates for the N.Acc. shell were: 1.5 mm anterior to the bregma, ± 0.7 lateral to the midline and 4.7 mm below the surface of the brain and the coordinates for the VTA were 3.4 posterior to bregma, ±0.5 mm lateral to the midline and 1.0 mm below the brain surface ( Franklin & Paxinos 1996 ). At the time of the experiment a cannula for drug administration was inserted and extended another 3.8 mm ventrally beyond the tip of the guide cannula, aiming at the VTA. The exposed tip of the dialysis membrane (20 000 kDa cut off with an o.d./i.d. of 310/220 μm, HOSPAL, Gambro, Lund, Sweden) of the probe was 1 mm.
In separate experiments, the effects of intra-VTA administration of AP5, or in separate experiments BIM28163, on ghrelin-induced accumbal dopamine release were determined, involving microdialysis in freely moving mice. On the day of the experiment, a dummy cannula was carefully inserted and retracted into the guide cannula. The probe was thereafter connected to a microperfusion pump (U-864 Syringe Pump; AgnThós AB) and perfused with Ringer solution at a rate of 1.5 μl/minute. After one hour of habituation to the microdialysis set-up, perfusion samples were collected every 20 minutes. The baseline dopamine level was defined as the average of three consecutive samples before the first drug/vehicle challenge. After the baseline samples, the antagonist (AP5 or BIM28163) was administered locally into the VTA followed subsequently by a ghrelin (i.p.) injection. The dopamine levels in the dialysates were determined by HPLC with electrochemical detection. A pump (Gyncotec P580A; Kovalent AB; V. Frölunda, Sweden), an ion exchange column (2.0 × 100 mm, Prodigy 3 μm SA; Skandinaviska GeneTec AB; Kungsbacka, Sweden) and a detector (Antec Decade; Antec Leyden; Zoeterwoude, the Netherlands) equipped with a VT-03 flow cell (Antec Leyden) were used. The mobile phase (pH 5.6), consisting of sulfonic acid 10 mM, citric acid 200 mM, sodium citrate 200 mM, 10% EDTA, 30% MeOH, was vacuum filtered using a 0.2 μm membrane filter (GH Polypro; PALL Gelman Laboratory; Lund, Sweden). The mobile phase was delivered at a flow rate of 0.2 ml/minute passing a degasser (Kovalent AB), and the analyte was oxidized at +0.4 V.
After completion of the microdialysis experiments, the locations of the probe and guide cannulae were verified. Neither water nor food were available to the mice during the microdialysis experiment. Only mice with probe placement in the N.Acc. and guide cannulae in the VTA were included in the statistical analysis.
Verification of probe and cannula/e placement
After the locomotor activity and microdialysis experiments were completed, the location of the probe and/or cannula/e were verified. The mice were decapitated, probes were perfused with pontamine sky blue 6BX to facilitate probe localization, and the brains were mounted on a vibroslice device (752M Vibroslice; Campden Instruments Ltd, Loughborough, UK). The brains were cut in 50 μm sections and the location of the probe and/or cannula/e was determined by gross observation using light microscopy. The exact position (some correct and some misplaced) of the probe and/or guide cannula/e was verified ( Franklin & Paxinos 1996 ).
Statistical analyses
All locomotor activity data were evaluated by a two-way ANOVA followed by Tukey's HSD post-hoc tests comparing treatments. The microdialysis experiments were evaluated by a two-way ANOVA for repeated measures followed by Tukey's HSD post-hoc test for comparisons between different treatments and specifically at given time points. Data are presented as mean ± SEM. A probability value of P < 0.05 was considered as statistically significant. | RESULTS
Effects of intra-VTA administration of a GHS-R1A antagonist on ghrelin-induced locomotor stimulation and accumbal dopamine release in mice
First, the role of GHS-R1A receptors in the VTA for the reinforcing effects of ghrelin by tests of ghrelin-induced locomotor stimulation and, in separate studies, by measurement of ghrelin-induced dopamine release were investigated. The locomotor stimulatory and accumbal dopamine releasing effects of ghrelin were attenuated by local administration of the GHS-R1A antagonist BIM28163 into the VTA ( Fig 1a,b ), at a dose shown previously to have no effect on locomotor stimulation and accumbal dopamine release per se ( Jerlhag et al . 2009 ). Thus, ghrelin-induced locomotor stimulation ( P < 0.01) was attenuated by VTA administration of BIM28163 ( P < 0.01) in mice (F(3,25) = 5.45, P = 0.005: n = 6–8). In the microdialysis experiments a significant effect of systemic ghrelin to increase dopamine release in comparison to vehicle treatment was observed ( P = 0.003). Pre-treatment with BIM28163 attenuated the ghrelin-induced increase in dopamine release compared with vehicle pre-treatment in mice ( P = 0.001) (treatment F(3,26) = 6.39, P = 0.002; time F(13,338) = 1.77, P = 0.047; treatment-time interaction F(13,338) = 4.01, P < 0.001). This difference was evident at the time intervals 20–100 minutes ( P < 0.001: n = 7–8).
Effects of intra-VTA administration of an orexin A receptor antagonist or peripheral injection of an opioid receptor antagonist on ghrelin-induced locomotor stimulation in mice
The ghrelin-induced locomotor stimulation ( P < 0.01) was not affected by VTA administration of the orexin A receptor antagonist SB334867 ( P > 0.05) in mice (F(3,24) = 8.44, P = 0.005: n = 6–8) ( Fig. 2a ). Likewise, the ghrelin-induced locomotor stimulation ( P < 0.01) was not suppressed by i.p. injection of the opioid receptor antagonist naltrexone ( P > 0.05) in mice (F(3,28) = 6.01, P = 0.003: n = 8) ( Fig. 2b ).
Effects of intra-VTA administration of a NMDA receptor antagonist on ghrelin-induced locomotor stimulation and increased accumbal dopamine release in mice
Intra-VTA administration of the NMDA receptor antagonist, AP5, abolished the ghrelin-induced locomotor stimulation and accumbal dopamine release ( Figs 3a,b ), at a dose that had no effect per se ( Table 1 ). Specifically, the ghrelin-induced locomotor stimulation ( P < 0.01) was attenuated by VTA administration of AP5 ( P < 0.001) in mice (F(3,27) = 8.06, P < 0.001: n = 7-8). Moreover, systemic ghrelin increased dopamine release ( P < 0.001) and pre-treatment with AP5 attenuated the ghrelin-induced increase in dopamine release compared to vehicle pre-treatment ( P < 0.001) (treatment F(3,36) = 19.98, P < 0.001; time F(12,432) = 3.46, P < 0.001; treatment-time interaction F(12,432) = 5.49, P < 0.001). This difference was evident at the time intervals 20–180 minutes ( P < 0.001: n = 8–11).
Control experiments showed that neither VTA administration, the volume infused nor the antagonist per se had any effect on locomotor activity ( Figs 1a, 2a,b and 3a ) or accumbal dopamine release ( Figs 1b and 3b ).
Probe and cannula/e placements
For all locomotor activity experiments eight mice in each treatment group were implanted with bilateral guide cannulae, whereas 11 mice undertook surgery for each treatment group for the microdialyis experiments. All these mice were included in the experiments (locomotor activity or microdialysis studies). After the experiment the location of the probe and/or guide cannulae was verified and only mice with probe placement in the N.Acc. shell and/or cannula/e in the VTA were included in the statistical analysis ( Fig. 4 ). Neither AP5 nor BIM28163 suppressed ghrelin-induced locomotor stimulation or accumbal dopamine release in a few mice in which the canulae were misplaced in neighbouring structures. It should also be emphasized that in a few mice the probe was located outside the N.Acc. shell and in these mice no effect of ghrelin on accumbal dopamine release was observed (data not shown). | DISCUSSION
The present study demonstrates that the stimulatory effects of peripheral ghrelin on locomotor stimulation and accumbal dopamine release involve VTA GHS-R1A signalling as both effects were suppressed by VTA administration of a GHS-R1A antagonist (BIM28163). Moreover, ghrelin's ability to activate the mesolimbic dopamine system was suppressed by pharmacological blockade of glutamatergic receptors but not by blockade of opioid or orexin A receptors. Thus, the locomotor stimulating effect of ghrelin was not affected by intra-VTA administration of the orexin A antagonist (SB334867) or by peripheral administration of an opioid receptor antagonist (naltrexone). Finally, ghrelin-induced locomotor stimulation as well as accumbal dopamine release were suppressed by VTA administration of an NMDA receptor antagonist (AP5). Taken together these data suggest that systemic ghrelin activates the mesolimbic dopamine system via GHS-R1A in the VTA and that glutamate rather than orexin or opioid signalling is required for ghrelin to stimulate the mesolimbic dopamine system.
Given that transport of ghrelin across the blood-brain barrier into the brain is somewhat limited ( Banks et al . 2002 ), there have been suggestions that ghrelin may exert its central effects via vagal afferents ( Date et al . 2000 ) or by gaining access at circumventricular organs, such as the arcuate nucleus and area postrema. Studies using Fos protein to map ghrelin's central actions have shown that whereas peripheral administration of ghrelin and GHS-R1A agonists activate a rather limited population of cells in the arcuate nucleus ( Dickson, Leng & Robinson 1993 ; Hewson & Dickson 2000 ) and area postrema ( Bailey et al . 2000 ), additional hypothalamic cell groups were recruited after central administration ( Lawrence et al . 2002 ). In the present study, it was demonstrated that the actions of peripheral ghrelin on the midbrain dopamine system, reflected by locomotor stimulation and accumbal dopamine release, can be blocked by VTA administration of a GHS-R1A antagonist. Taken together with previous neuroanatomical and accumbal dopamine measurement studies showing that the target cells for ghrelin in the VTA include the dopaminergic cell group ( Abizaid et al . 2006 ; Jerlhag et al . 2006 ; Jerlhag 2008 ; Kawahara et al . 2009 ; Quarta et al . 2009 ), it seems likely that peripheral ghrelin directly activates the VTA dopamine system via GHS-R1A. The activation of the mesolimbic dopamine system by ghrelin may be due to the reported ability of the GHS-R1A to dimerize with the dopamine D1 receptor, both receptors expressed on dopamine neurons in the VTA, and thereby amplifies the dopamine signalling ( Jiang, Betancourt & Smith 2006 ). Central ghrelin signalling system, including the GHS-R1A, appears to be required for reinforcement induced by addictive drugs including cocaine ( Wellman, Davis & Nation 2005 ; Davis, Wellman & Clifford 2007 ; Tessari et al . 2007 ) and alcohol ( Jerlhag et al . 2009 ). The mesolimbic dopamine system appears to be a likely target for ghrelin also in man, as evidenced from functional MRI studies in which peripheral ghrelin altered the response of the ventral striatum to visual food cues ( Malik et al . 2008 ). Indeed, ghrelin may, via VTA GHS-R1A, increase the incentive value for natural as well as chemical reinforcements.
Neurotransmitters, including orexin, opioids as well as glutamate, have previously been shown to modulate the intake of natural and chemical reinforcers of the mesolimbic dopamine system as well as to regulate the activity of these VTA neurons ( Hoebel et al . 1989 ; Engel et al . 1992 ; Wise 2002 ; Thiele et al . 2003 ). Specifically, AP5 and SB334867, in a similar dose range, attenuates the reinforcing properties of drugs of abuse ( Herz 1997 ; Taber & Fibiger 1997 ; Smith et al . 2009 ). Here, neither the opioid receptor—nor the orexin A receptor—antagonist affected ghrelin-induced locomotor stimulation, indicating that these systems do not interfere with these effects of ghrelin. Consistent with our findings, the orexigenic response to ghrelin when administered into key mesolimbic dopamine structures such as the VTA has previously been shown to be independent of opioid receptor signalling ( Naleid et al . 2005 ). Peripheral injection of naltrexone has been shown to blocks the reinforcing properties of alcohol in rodents ( Herz 1997 ), suggesting that naltrexone passes the blood-brain barrier and has central effects. Even though opioid receptors, specifically the μ receptor, mediate drug-induced reinforcement this receptor appear to be less important for the ability of ghrelin to activate the mesolimbic dopamine system. Orexin-containing neurons have previously been suggested to regulate ghrelin-induced feeding ( Toshinai et al . 2003 ), but orexin A receptors do not appear to be crucial for the locomotor stimulatory effects of ghrelin. Collectively, these data suggest that ghrelin-induced activation of the mesolimbic dopamine system appears to be regulated via other mechanisms than its orexigenic properties. In the present study, it was shown that ventral tegmental NMDA receptors are required for ghrelin-induced locomotor stimulation and accumbal dopamine release. Supportively, the effects of ghrelin to increase the electrical activity of dopaminergic neurons in the VTA appears to be dependent on the excitatory glutamatergic input and also blockade of NMDA receptors in the VTA reduces food-induced accumbal dopamine release ( Taber & Fibiger 1997 ; Abizaid et al . 2006 ). NMDA receptors have also been shown to mediate the accumbal dopamine release observed when animals consume food after ghrelin administration ( Kawahara et al . 2009 ). As shown previously, ghrelin-induced reinforcement also involves nicotinic acetylcholine receptors in the VTA ( Jerlhag et al . 2006 ; Jerlhag et al . 2008 ). Collectively, these data suggest that neurotransmitters including acetylcholine and glutamate are required for ghrelin-induced reinforcement, which it has in common with natural as well as chemical reinforcers of the mesolimbic dopamine system. The mechanisms for the interaction between ghrelin, acetylcholine and glutamate in the VTA are still unclear. However, presynaptic nicotinic acetylcholine receptors have been shown to modulate the release of glutamate in the VTA, which via postsynaptic NMDA receptors regulate accumbal dopamine release ( Schilström et al . 1998 ; Schilström et al . 2000 ). Tentatively, ghrelin may increase the acetylcholine release, which via such mechanisms may indirectly activate the mesolimbic dopamine system. However, the possibility of the existence of presynaptic NMDA receptors on cholinergic neurons cannot be excluded ( Corlew et al . 2008 ). The possibility that ghrelin has an ability to rearrange the excitatory NMDA-mediated synaptic input in a manner that would increase the probability of activation of the dopamine neurons by other inputs should also be considered; thus findings in the hippocampus as well as VTA show that ghrelin has effects on synaptic plasticity ( Abizaid et al . 2006 ; Diano et al . 2006 ). The glutamatergic afferents to the VTA mainly originate from the prefrontal cortex, lateral hypothalamus, bed nucleus of stria terminalis, superior coliculus and the LDTg. This input regulates the activity of dopamine via NMDA receptors ( Schilström et al . 1998 ; Georges & Aston-Jones 2002 ; Geisler & Zahm 2005 ). It should however be emphasized that other afferents including GABA, serotonin and noradrenalin, known to modulate the activity of dopamine projections to the N.Acc. ( Wise 2002 ), may also have important roles for the ghrelin-induced activation of the mesolimbic dopamine system. Although, GABA A receptors in the VTA does not mediate the increase in accumbal dopamine observed after food consumption induced by ghrelin ( Kawahara et al . 2009 ).
In summary, the present study shows that the effects of peripheral ghrelin on locomotor stimulation and accumbal dopamine release in mice (that reflect direct actions of ghrelin at the level of the mesolimbic dopamine system, specifically the VTA) can be suppressed by an NMDA antagonist and are therefore likely to be under glutamatergic control. These data may have clinical implications since hyperghrelinemia is associated with addictive behaviours including compulsive overeating and alcohol use disorder ( Cummings et al . 2001 ; Kim et al . 2005 ; Kraus et al . 2005 ). It may therefore be proposed that ghrelin-responsive circuits at the level of the VTA, that appear to be sensitive to cholinergic and glutamatergic input, may serve as a novel pharmacological target for treatment of such addictive behaviours. | Re-use of this article is permitted in accordance with the Terms and Conditions set out at http://wileyonlinelibrary.com/onlineopen#OnlineOpen_Terms
Recently, we demonstrated that the central ghrelin signalling system, involving the ghrelin receptor (GHS-R1A), is important for alcohol reinforcement. Ghrelin targets a key mesolimbic circuit involved in natural as well as drug-induced reinforcement, that includes a dopamine projection from the ventral tegmental area (VTA) to the nucleus accumbens. The aim of the present study was to determine whether it is possible to suppress ghrelin's effects on this mesolimbic dopaminergic pathway can be suppressed, by interrupting afferent inputs to the VTA dopaminergic cells, as shown previously for cholinergic afferents. Thus, the effects of pharmacological suppression of glutamatergic, orexin A and opioid neurotransmitter systems on ghrelin-induced activation of the mesolimbic dopamine system were investigated. We found in the present study that ghrelin-induced locomotor stimulation was attenuated by VTA administration of the N-methyl-D-aspartic acid receptor antagonist (AP5) but not by VTA administration of an orexin A receptor antagonist (SB334867) or by peripheral administration of an opioid receptor antagonist (naltrexone). Intra-VTA administration of AP5 also suppressed the ghrelin-induced dopamine release in the nucleus accumbens. Finally the effects of peripheral ghrelin on locomotor stimulation and accumbal dopamine release were blocked by intra-VTA administration of a GHS-R1A antagonist (BIM28163), indicating that GHS-R1A signalling within the VTA is required for the ghrelin-induced activation of the mesolimbic dopamine system. Given the clinical knowledge that hyperghrelinemia is associated with addictive behaviours (such as compulsive overeating and alcohol use disorder) our finding highlights a potential therapeutic strategy involving glutamatergic control of ghrelin action at the level of the mesolimbic dopamine system. | Supported by the Swedish Research Council (K2006-21X-04247-33-3, K2007-54X-20328-013, K2010-55X-20328-04-3), the Alcohol Research Council of the Swedish Alcohol Retailing Monopoly, The Swedish Labour Market Insurance, Swedish council for tobacco research, the foundations of Wilhelm and Martina Lundgren, Knut and Alice Wallenberg, The Adlerbert Research, Tore Nilsson, Anders Otto Swärd, Magnus Bergvall, Torsten and Ragnar Söderberg, Längmanska art, European Union (FP7-HEALTH-2009-241592, FP7-KBBE-2009-3 (245009, NeuroFAST)) Novo Nordisk (GeA/AIR), ALF Göteborg (SU7601, SU76540), the Swedish Foundation for Strategic Research to Sahlgrenska Center for Cardiovascular and Metabolic Research (A305-188) and The Swedish Society of Medicine. Gun Andersson and Kenn Johannessen are gratefully acknowledged for expert and valuable technical assistance. We express our grateful thanks to Rakesh Datta, Michael D Culler and Jesse Dong (IPSEN) for contributions to the discovery and characterization of BIM28163 and also for generously supplying BIM28163.
Authors Contribution
EJ conducted the experiments and analyzed the data. All authors contributed to the writing of the manuscript. JAE designed research. All authors have critically reviewed content and approved final version submitted for publication. | CC BY | no | 2022-01-12 15:46:43 | Addict Biol. 2011 Jan; 16(1):82-91 | oa_package/28/dc/PMC3015055.tar.gz |
||
PMC3015056 | 21039362 | Methods
Fluid composition and study design
We used a new phosphate-containing solution for dialysis that in addition to standard electrolytes also contains 4.0 mmol of potassium and 1.2 mmol of phosphate (Phoxilium, Gambro Lundia AB, Lund, Sweden, Table 1 ). As a control, we used our routine dialysis solution that does not contain phosphate (Hemosol B0, Gambro Lundia AB, Table 1 ). At our ICU at Lund University Hospital we applied three regimes, half a year each, for all patients requiring CRRT treatment. The treatment mode used was CVVHDF. During the first period (group 1), all the patients received dialysate solution and replacement solution that did not contain phosphate (Hemosol B0), during the next half year (group 2), all patients requiring CRRT treatment received the phosphate-containing solution as dialysis solution and a phosphate-free replacement solution (Hemosol B0) and finally during the last half year period (group 3), the patients received the phosphate-containing solution both as a dialysis solution and as a replacement solution. Blood sampling was performed according to normal routines at our department, but the physicians in charge continuously modified the CRRT treatment settings and the phosphate supplementation according to the patients' ongoing clinical needs. The physicians treating the patients did not have knowledge of the study setting during the treatments, but they were fully aware of the fluids and their contents.
Patients and CRRT treatments
After acceptance by the Regional Ethical Review Board (DNR 570/2008), Lund University, Sweden, we evaluated retrospectively the first 14 patients who did not fulfill the exclusion criteria in each group, a total of 42 consecutive patients ( Table 2 ). Patients were excluded if they had chronic kidney disease, if they had received intermittent dialysis before the ICU stay, if the CRRT treatment lasted <10 h or if they were under the age of 18 years. In all patients, the Gambro Prismaflex CRRT machine with the CVVHDF modality and a Hospal M100 filter (Hospal Industrie, Meyziew, France) was used. The blood flow, the dialysis fluid flow, the replacement fluid flow, the anticoagulation used and the fluid removal were set according to the patients' conditions and requirements. Of the replacement fluid, 500 ml/h was postfilter and the rest was prefilter in each treatment according to the general standard at the department. Intravenous phosphate addition was prescribed when serum phosphate was <0.8 mmol/l, also according to the general standard at the department. Nutrition was given only if the patients were hemodynamically stable as parenteral or enteral or both during the study period.
Clinical parameters of the patients
All patients had regular measurements of serum sodium, potassium, ionized calcium, pH, p CO 2 and bicarbonate either from an arterial line or from a central venous line before the start and regularly every fourth hour all through the CRRT treatment during 1, 2, 3, 4 and maximum 5 days. For phosphate and magnesium analyses, blood samples were taken at 5:00 hours in the morning and at 17:00 hours in the afternoon before the start and during the CRRT treatment days. Blood samples were analyzed at the hospital laboratory (Laboratory for Clinical Chemistry, Lund University Hospital, Lund, Sweden). The results are presented as means±SD for normal distributed data, median and range for remaining data.
In addition, baseline characteristics of study patients and delivered CRRT were registered ( Table 3 ). Hypophosphatemia has been defined as condition where the serum phosphate level is <0.81 mmol/l.
Statistics
One-way repeated measures analysis of variance was used. The statistical program used was SigmaStat, version 3.5, for Windows XP. Differences were considered to be significant for P <0.05. | Results
Main diagnoses leading to intensive care of the study groups, baseline characteristics, timing of initiation of CRRT, ultrafiltration rate, duration of the CRRT and anticoagulation were statistically similar between the three groups ( Tables 2 and 3 ).
The incidence of hypophosphatemia occurred in 11 of 14 patients in group 1, where the patients did not receive the phosphate-containing dialysis solution, but received phosphate intravenously based on the serum phosphate values ( Fig. 1 ). In group 2, where the patients received a phosphate-containing solution as the dialysis solution and a phosphate-free solution as the replacement solution, five of 14 patients had at least one episode of hypophosphatemia ( Fig. 1 ). No episodes of hypophosphatemia were detected in group 3.
The serum phosphate level was 1.90 mmol/l at baseline and 0.99 mmol/l during CRRT treatment in group 1. In group 2, the corresponding values were 1.54 and 1.20 mmol/l, and in group 3, they were 1.83 and 1.43 mmol/l ( Table 5 ). However, due to the simultaneous intake of enteral/parenteral solutions, two of 14 of the patients in group 3 had a temporary increase in serum phosphate (<1.9 mmol/l), but there were no cases of hyperphosphatemia that required a withdrawal of the phosphate-containing dialysis solution from the treatment.
Phosphate intake was from nutrition and from intravenous supplementation. The average phosphate input calculated from the total nutrition was 18 mmol/CRRT treatment/day for all groups. Short-acting insulin was administered as infusion in order to achieve a blood glucose level between 5 and 8 mmol/l. Phosphate was supplemented intravenously if serum phosphate declined to <0.80 mmol/l. In group 1, the average phosphate supplementation was 10 mmol/CRRT treatment/day. In group 2, the average supplementation was 5 mmol/CRRT treatment and day, and in group 3, there was no intravenous supplementation.
There was a decline in phosphate levels during CRRT treatment in both groups 1 and 2. In group 1, serum phosphate declined constantly, although the patients received phosphate supplementation intravenously ( Fig. 2 ). In group 2, the phosphate levels were unstable and reached a low level in the end of the study. In group 3, where the patients received a phosphate-containing solution both as the dialysis solution and the replacement solution, none of the patients had episodes of hypophosphatemia ( Fig. 2 ). Phosphate remained stable in this patient population during the entire study.
For group 3, we also evaluated any adverse event in sodium, potassium and magnesium homeostasis. There was a significant increase in sodium ( P ≤0.001) and a slight increase in ionized calcium ( P =0.029), whereas potassium and magnesium remained stable during the entire study as summarized in Table 4 . A comparison between the study groups revealed that there was a significant difference in phosphate ( P ≤0.001), ionized calcium ( P =0.004) and bicarbonate ( P =0.045) between the groups ( Table 5 ). The pH and p CO 2 measurements were obtained at baseline and rapidly declined toward normal values after starting the CRRT treatment and remained essentially unchanged during the treatment in all groups ( Table 4 ). | Discussion
This study demonstrates that the new phosphate-containing dialysis solution is safe, reduces the variability of serum phosphate levels during CRRT and reduces the incidence of hypophosphatemia.
Various protocols for intravenous phosphate supplementation have been studied in the last 30 years. Today, there is a trend toward the use of larger and faster boluses of phosphate because of high failure of repletion (20–70%) and the need for additional phosphate administration. 24 – 31 Authors usually agree that larger amounts of phosphate are needed to correct total body deficit, but their fear of adverse reactions has prompted a restrained attitude. Recent studies on ICU patients confirm that a relatively rapid infusion of potassium phosphate is safe if baseline serum potassium is below 4.5 mmol/l. 1 The infusion rate is thus consequently limited by the serum potassium levels, and also by serum calcium levels, as high phosphate serum levels could induce hypocalcemia, as phosphate could precipitate with calcium in blood vessels and tissues. Recently, phosphate has been injected into dialysis solutions during treatment, but there is a risk of precipitation. 1 , 21 , 22 We evaluated whether phosphate-containing solutions for dialysis and replacement simplify phosphate replacement even further.
The frequency of phosphate disturbances in critically ill patients is high, although the figure varies considerably depending on the study, between 8.8% and 80%. 32 , 33 In our study, the incidence of hypophosphatemia during CRRT was 79% in the control group. The incidence of hypophosphatemia was lower (35%) in patients receiving the phosphate-containing solution as the dialysis solution and Hemosol B0 as the replacement solution. There were no episodes of hypophosphatemia in patients receiving only the phosphate-containing solution, but there was an incidence of mild (>1.9 mmol/l) hyperphosphatemia in 14% of these patients. This is probably due to the simultaneous intake of nutritional support, but the phosphate levels were only marginally elevated without physiological consequences and did not require withdrawal of the phosphate-containing dialysis solution from the patient.
The amount of phosphate required to correct total body deficit is variable and depends on the cause of hypophosphatemia and the chronicity of the process. 3 , 11 , 24 The many physiological rearrangements in ICU patients may explain the need for larger amounts required for repletion. In our study, phosphate addition was necessary for patients in groups 1 and 2. This is representative of our experience, where the majority of patients on CRRT require phosphate supplementation shortly after CRRT initiation despite nutritional support. Malnourished alcoholic patients have a larger deficit due to long-standing negative phosphate balance. Malnutrition induces a re-feeding syndrome that has been reported after only 48 h of fasting in the ICU. 2 The distribution volume for phosphate might be increased, 24 while insulin, carbohydrate, and catecholamine administration act to decrease the serum phosphate concentration. 2 , 3 , 11 , 34
The incidence of hypophosphatemia in this study was 10% of the 42 patients before the start of CRRT. The incidence of hyperphosphatemia was 52%, which is slightly lower than that found in other studies, where 65–80% of patients present hyperphosphatemia before the start of the CRRT. 8 , 35 Phosphate was within normal limits in 38% of patients even before the beginning of CRRT. The new dialysis fluid contains 30 mM bicarbonate compared with 32 mM in Hemosol B0. With prescribed CRRT clearances of ≥20 ml/kg/h, most acid–base disturbances can be managed with bicarbonate compositions of 25–35 mM. 36 The normal serum range for bicarbonate is 22–30 mmol/l, which was achieved in patients in all groups, although the increase in group 1 was most significant. Ionized calcium increased in all groups during treatment time, but the increase was less marked in groups 2 and 3. There is a possibility that the slow increase in bicarbonate and ionized calcium could reflect the differences in the ion content between Hemosol B0 and the phosphate-containing dialysis fluid, although these differences could also be due to the limitations of this study. The fact that this was a retrospective observational study, without other intervention but for the dialysis solutions, means that bias due to the patients' illnesses and due to other treatments could affect ion concentrations. Even the phosphate concentration in parenteral and enteral nutrition cannot be completely excluded, although this would not alter the basic results of the study.
The next objective of the study was to evaluate whether the administered phosphate could alter calcium and potassium homeostasis. Phosphate could theoretically precipitate with calcium, which could lead to hypocalcemia in the patients. We found that the phosphate-containing dialysis fluid did not induce hypocalcemia during CRRT. Another important concern was whether this fluid would cause potassium overload, as potassium phosphate is used instead of sodium phosphate. We found that the phosphate-containing dialysis fluid did not induce hyperkalemia either. Potassium phosphate was favored over sodium phosphate because potassium usually has to be added to CRRT solutions. In addition, there were no disturbances of magnesium levels.
The present study shows that by using the phosphate-containing fluid both as the dialysis fluid and the replacement fluid, we could eliminate the episodes of hypophosphatemia. The new phosphate-containing dialysis fluid simplified the phosphate control and avoided rapid phosphate fluctuations with intravenous bolus administration. | Re-use of this article is permitted in accordance with the Terms and Conditions set out at http://wileyonlinelibrary.com/onlineopen#OnlineOpen_Terms
Background
Hypophosphatemia occurs in up to 80% of the patients during continuous renal replacement therapy (CRRT). Phosphate supplementation is time-consuming and the phosphate level might be dangerously low before normophosphatemia is re-established. This study evaluated the possibility to prevent hypophosphatemia during CRRT treatment by using a new commercially available phosphate-containing dialysis fluid.
Methods
Forty-two heterogeneous intensive care unit patients, admitted between January 2007 and July 2008, undergoing hemodiafiltration, were treated with a new Gambro dialysis solution with 1.2 mM phosphate (Phoxilium) or with standard medical treatment (Hemosol B0). The patients were divided into three groups: group 1 ( n =14) receiving standard medical treatment and intravenous phosphate supplementation as required, group 2 ( n =14) receiving the phosphate solution as dialysate solution and Hemosol B0 as replacement solution and group 3 ( n =14) receiving the phosphate-containing solution as both dialysate and replacement solutions.
Results
Standard medical treatment resulted in hypophosphatemia in 11 of 14 of the patients (group 1) compared with five of 14 in the patients receiving phosphate solution as the dialysate solution and Hemosol B0 as the replacement solution (group 2). Patients treated with the phosphate-containing dialysis solution (group 3) experienced stable serum phosphate levels throughout the study. Potassium, ionized calcium, magnesium, pH, p CO 2 and bicarbonate remained unchanged throughout the study.
Conclusion
The new phosphate-containing replacement and dialysis solution reduces the variability of serum phosphate levels during CRRT and eliminates the incidence of hypophosphatemia. | T he majority of patients on continuous renal replacement therapy (CRRT) will require phosphate supplementation shortly after CRRT initiation. 1 One reason is that critically ill patients present several conditions predisposing hypophosphatemia such as sepsis, alcohol withdrawal, malnutrition, catecholamines, intravenous glucose infusion, hyperventilation, diuretics and rhabdomyolysis. 2 – 4 Another reason is the CRRT technique that achieves high clearance of small solutes, such as phosphate. 5 – 9 In addition, low serum phosphate levels may also occur in the setting of extracellular to intracellular shifts that occur with respiratory alkalosis, high blood concentrations of stress hormones (i.e., insulin, glucagon, adrenalin, cortisol) and with refeeding syndrome.
As phosphate is a constituent of enzymes and intermediate phosphorylated compounds, it plays a key role in cellular metabolism and is essential in several biological processes. Serum phosphate concentration is maintained between 0.81 and 1.45 mmol/l. By convention, hypophosphatemia is often graded as mild (<0.81 mmol/l), moderate (<0.61 mmol/l) and severe (<0.32 mmol/l). Severe hypophosphatemia has been linked to increased mortality in surgical intensive care patients 7 and was recently shown to serve as an independent mortality predictor in sepsis. 10 Symptoms of hypophosphatemia are usually only seen in patients with moderate or severe hypophosphatemia and include ventilatory muscle weakness, cardiac failure, insulin resistance, hemolysis, impaired platelet and white blood cell function, rhabdomyolysis, and, in rare cases, neurologic disorders. 3 , 11 – 17 However, all these alternations have been shown to reverse by simply correcting the phosphate levels. 3 , 7 , 12 , 13 , 18 – 20 Phosphate is supplemented intravenously in symptomatic patients, but phosphate has also been added directly to the dialysate and replacement fluids, 1 , 21 , 22 with a risk of precipitation with calcium.
The development of many electrolyte disturbances in the intensive care unit (ICU) could be prevented by the use of better adapted dialysis fluids. This study evaluated the possibility to achieve and maintain a normal phosphate balance over time in patients on CRRT by using a new phosphate-containing dialysis fluid and replacement fluid. | Competing interests : The authors have not disclosed any potential competing interests. The ICU department of Lund University Hospital has paid an ordinary price for the phosphate-containing dialysis solution used in the study. Gambro has approved a Research Grant, which has been paid to the ICU department of Lund University Hospital. Gambro has not had any insight in how the Research Grant has been used. An independent statistician has monitored the raw data and the statistics. | CC BY | no | 2022-01-12 15:46:43 | Acta Anaesthesiol Scand. 2011 Jan; 55(1):39-45 | oa_package/46/1b/PMC3015056.tar.gz |
||
PMC3015057 | 20977678 | Introduction
Eukaryotic cells contain very extensive intracellular membrane systems, and many vital cellular processes, such as metabolic reactions, signal transduction, cytoskeletal rearrangements, protein sorting and regulation of membrane dynamics, occur partially or entirely at membrane–cytosol interfaces. The main advantages of executing biochemical reactions on membranes include the limitation of substrate diffusion (i.e. limited to two dimensions instead of three) and the possibility of confining biochemical processes to restricted subcellular locations.
The containment of cellular processes to intracellular membranes requires the reversible assembly of protein complexes onto specific membranes. A group of phosphorylated derivatives of phosphatidylinositol (PtdIns), collectively known as phosphoinositides (PIs), are ideally suited for this task [ 1 ]. The PIs are generated and metabolized through the activities of a number of substrate-specific PI kinases and phosphatases, the dysfunctions of which are associated with various human diseases [ 2 ]. Of special interest in cancer research are two PI 3-kinases (i.e. kinases that phosphorylate the inositol headgroup in the 3-position) that have opposing roles in tumourigenesis. Class I PI 3-kinase (PI3K-I), on the one hand, is a well-known tumour promoting enzyme (or to be more precise, a small group of related enzymes) whose hyperactivity is strongly associated with carcinogenesis in humans [ 3 ]. Consistent with this, PTEN, a phosphatase that essentially reverses the 3-phosphorylation mediated by PI3K-I, is an important tumour suppressor [ 4 ]. The catalytic product of PI3K-I, PtdIns(3,4,5) P 3 , recruits several proteins involved in cell signalling to the plasma membrane, the most important one being the protein kinase AKT, which orchestrates various signalling cascades that promote cell growth and survival. On the other hand, class III PI 3-kinase (PI3K-III) is considered to be a tumour suppressor based on the findings that three of its accessory subunits, Beclin 1, UVRAG and BIF-1, have been independently identified as tumour suppressors whose partial or complete inactivation causes the increased occurrence of spontaneous tumours in mice and (probably) humans [ 5 – 7 ]. Recently identified molecular and cellular mechanisms that may serve to explain the tumour suppressor functions of PI3K-III are the topic of the present review. | Conclusions and perspectives
As discussed in the present review, PI3K-III may theoretically function as a tumour suppressor via at least three different mechanisms ( Fig. 3 ). The involvement of PI3K-III in autophagy maintains genome stability by eliminating damaged organelles that produce reactive oxygen species; the role of PI3K-III in endosomal fusion and sorting ensures correct downregulation of mitogenic signalling; and the correct function of PI3K-III in cytokinesis prevents aneuploidy. Further work is required to establish which of these three processes is most important in PI3K-III-mediated tumour suppression. It can also not be excluded that PI3K-III may act as tumour suppressor by additional means. For example, SARA, a mediator of transforming growth factor-β signalling, requires PtdIns3 P binding for its function [ 87 ], and the transforming growth factor-β signalling pathway does have a tumour suppressor role under most conditions [ 88 ]. Furthermore, PtdIns3 P binding subunits mediate membrane association of the retromer complex [ 89 ], which sorts cargoes such as mannose 6-phosphate receptors and Wntless (an accessory factor in Wnt secretion) for trafficking from endosomes to the biosynthetic pathway [ 89 , 90 ]. Even though there is no evidence so far that implicates the retromer in tumour suppression, this possibility cannot at present be discarded.
The involvement of PI3K-III in diverse cellular processes raises the question of how this complex is recruited to the correct membranes at the right time. There is evidence that PI3K-III is recruited to early and late endosomal membranes through interactions with the small GTPases RAB5 and RAB7, respectively [ 91 – 93 ]. Less is known about how PI3K-III is recruited to midbodies and autophagic membranes, although the latter is likely to be mediated by the autophagy-specific subunit ATG14 [ 69 ]. In addition, the finding that phosphorylation of VPS34 during mitosis causes its dissociation from Beclin 1 [ 18 ] provides a hint that post-translational modifications of PI3K-III may contribute to regulate its activity and specificity.
Although considerable efforts have been made to understand how PtdIns3 P is formed by PI3K-III, we are also beginning to learn about the metabolism of this lipid. Three known routes for PtdIns3 P metabolism have been described: degradation by lysosomal lipases, phosphorylation by the PtdIns3 P 5-kinase Fab1/PIKfyve, and dephosphorylation by 3-phosphatases [ 94 ]. It is worth noting that Fab1/PIKfyve is itself a PtdIns3 P binding protein [ 95 ], and that MTM1 and MTMR2, two phosphatases capable of dephosphorylating PtdIns3 P , can associate with PI3K-III on endosomal membranes [ 96 , 97 ]. This suggests a tight regulation of PtdIns3 P formation and turnover, and it will be interesting to determine whether PIKfyve and MTM1/MTMR2 may play any role in tumourigenesis.
Even though it is assumed that PI3K-III functions as tumour suppressor through its ability to produce PtdIns3 P at the correct intracellular membranes, this has not been formally demonstrated and, to date, we do not know whether the catalytic subunit of PI3K-III, VPS34, is a tumour suppressor. Further studies should clarify this, and it will also be important to establish whether the catalytic function of VPS34 is required for its eventual tumour suppressor function.
If we nevertheless accept that PI3K-III acts as tumour suppressor through PtdIns3 P generation, can this be exploited in cancer diagnosis and therapy? Because it is much easier to inhibit an enzyme pharmacologically than to boost its function, the tumour promotor PI3K-I is a more attractive drug target than the tumour suppressor PI3K-III. On the other hand, we know that PtdIns3 P can be dephosphorylated and that PI3K-III undergoes negative regulation [ 11 , 12 ], and alleviation of these inhibitory mechanisms might provide a viable strategy towards increasing the tumour suppressor activity of PI3K-III and its catalytic product in cancer treatment. More straightforwardly, knowing that PI3K-III subunits such as Beclin 1 and UVRAG are frequently downregulated or mutated in cancers [ 6 , 7 , 67 , 98 , 99 ], it will be interesting to perform systematic analyses of PI3K-III subunits and key PtdIns3 P effectors in various cancers. Such studies should reveal mutational and expression-based signatures that can be used to predict the outcome of disease, and to guide the choice of therapeutic regimens. | Re-use of this article is permitted in accordance with the Terms and Conditions set out at http://wileyonlinelibrary.com/onlineopen#OnlineOpen_Terms
Phosphorylated derivatives of the membrane lipid phosphatidylinositol (PtdIns), known as phosphoinositides (PIs), regulate membrane-proximal cellular processes by recruiting specific protein effectors involved in cell signalling, membrane trafficking and cytoskeletal dynamics. Two PIs that are generated through the activities of distinct PI 3-kinases (PI3Ks) are of special interest in cancer research. PtdIns(3,4,5) P 3 , generated by class I PI3Ks, functions as tumour promotor by recruiting effectors involved in cell survival, proliferation, growth and motility. Conversely, there is evidence that PtdIns3 P , generated by class III PI3K, functions in tumour suppression. Three subunits of the class III PI3K complex (Beclin 1, UVRAG and BIF-1) have been independently identified as tumour suppressors in mice and humans, and their mechanism of action in this context has been proposed to entail activation of autophagy, a catabolic pathway that is considered to mediate tumour suppression by scavenging damaged organelles that would otherwise cause DNA instability through the production of reactive oxygen species. Recent studies have revealed two additional functions of PtdIns3 P that might contribute to its tumour suppressor activity. The first involves endosomal sorting and lysosomal downregulation of mitogenic receptors. The second involves regulation of cytokinesis, which is the final stage of cell division. Further elucidation of the mechanisms of tumour suppression mediated by class III PI3K and PtdIns3 P will identify novel Achilles’ heels of the cell’s defence against tumourigenesis and will be useful in the search for prognostic and diagnostic biomarkers in cancer. | PI3K-III: a conserved lipid kinase complex
The catalytic subunit of PI3K-III was first identified as Vps34, an enzyme that mediates vacuolar protein sorting in the yeast Saccharomyces cerevisiae [ 8 ]. By contrast to PI3K-I, PI3K-III is conserved between yeast, plants and humans, and the human homologue of Vps34 is referred to as hVps34, PIK3C3 or VPS34. In the present review, the latter term is used. Subsequent work soon revealed that Vps34 is associated with a regulatory subunit, Vps15, a putative protein kinase [ 9 ]. More recently, Vps34 was found to engage in two functionally distinct complexes in yeast. One complex, consisting of Vps34, Vps15, Vps30 and Vps38, regulates endosomal sorting, whereas another complex, in which Vps38 is replaced by Atg14, is required for autophagy [ 10 ].
The two PI3K-III complexes in yeast have their human counterparts: one consisting of VPS34, VPS15, the Vps30 homologue Beclin 1 and the Vps38 homologue UVRAG, and the other containing ATG14 (also called hAtg14, Atg14L or Barkor) instead of UVRAG [ 11 – 14 ]. Thus, we can consider VPS34-VPS15-Beclin 1 as a core complex onto which the accessory subunits UVRAG and ATG14 can assemble in a competitive manner ( Fig. 1A ). In addition, the UVRAG containing complex can associate with Rubicon, a protein that negatively regulates the function of this complex in endosomal and autophagic trafficking [ 11 , 12 ], and with BIF-1 (also known as endophilin B1 or SH3GLB1), identified as a positive regulator of autophagy [ 5 ] ( Fig. 1B ). Although Rubicon can be isolated as a major constituent of UVRAG-containing PI3K-III complexes, this is not the case with BIF-1, which appears to be more transiently associated [ 12 , 15 ]. Nevertheless, the fact that BIF-1 contains an amino-terminal BAR domain predicted to have membrane-bending ability, as well as the finding that knockdown of BIF-1 inhibits the catalytic activity of PI3K-III, suggests that this protein is an important accessory subunit of mammalian PI3K-III [ 5 ]. BAR domain proteins have been found to associate with membranes of high curvature [ 16 ], and it is tempting to speculate that BIF-1 might direct PI3K-III activity to such membranes.
The recently solved crystal structure of Drosophila melanogaster Vps34 revealed that this PI3K has a considerably smaller ATP-binding pocket than class I PI3Ks [ 17 ]. This explains why several inhibitors of class I PI3Ks fail to inhibit class III PI3K (i.e. they are too bulky to fit into the ATP-binding pocket). More importantly, the structural insight offers a rationale for designing specific PI3K-III inhibitors in the future.
Even though considerable knowledge has been gained about the biochemical composition of PI3K-III, we still know little about the regulation of its catalytic activity. The catalytic activity appears to be stimulated by BIF-1, whereas Rubicon inhibits the overall function of PI3K-III [ 5 , 11 , 12 ]. However, the exact mechanisms of these regulations remain unsolved. An important clue to the regulation of VPS34 has emerged recently with the finding that VPS34 is phosphorylated on threonine 159 by the cyclin-dependent kinase Ckd1 during mitosis, and that this causes its dissociation from Beclin 1 and an inhibition of autophagy [ 18 ].
Recognition of PtdIns3 P by FYVE and phox homology (PX) domains
A breakthrough in our understanding of how PI3K-III and its catalytic product control cellular functions came with the identification of the FYVE (conserved in Fab1, YOTB, Vac1 and EEA1) domain, and the demonstration that this domain binds PtdIns3 P . The FYVE domain was originally identified as a zinc finger required for localization of the early endosomal autoantigen 1 (EEA1) to early endosomes [ 19 ]. The finding that wortmannin, a general PI3K inhibitor, prevents the localization of EEA1 to endosomes, provided a clue that EEA1 might be recruited by a 3-phosphorylated PI [ 20 ], and biochemical studies showed that the FYVE domains from yeast and mammalian proteins bind to PtdIns3 P with high specificity [ 21 – 23 ]. Further progress in deciphering the downstream functions of PtdIns3 P was obtained when the PX domain was identified as a second PtdIns3 P binding domain [ 24 – 27 ]. Although a few mammalian PX domains bind to other 3-PIs than PtdIns3 P , most of the PX domains bind specifically to PtdIns3 P with affinities comparable to those of FYVE domains [ 28 ]. The human genome encodes approximately 30 FYVE domain-containing proteins and some 45 PX domain-containing proteins that presumably mediate most (but not all) of the downstream functions of PtdIns3 P [ 29 ]. Additional PtdIns3 P -binding proteins that do not contain FYVE or PX domains include the Proppin/WIPI proteins, which bind PtdIns3 P [and to some extent the related PtdIns(3,5) P 2 ] through a WD40-repeat-containing β-propeller structure [ 30 , 31 ], and certain variant pleckstrin homology domains such as the GLUE (GRAM-like ubiquitin-binding in EAP45) domain [ 32 – 34 ].
Intracellular localization of PtdIns3 P
The identification of PtdIns3 P -binding domains offered the possibility to design probes that reveal the intracellular distribution of this lipid. Early work revealed that a single FYVE domain from the endosomal protein HRS binds PtdIns3 P with too low affinity to be useful as a probe. Remarkably, however, when two FYVE domains from HRS were fused in tandem (2xFYVE), the resulting construct could be used for imaging of cellular PtdIns3 P with high sensitivity and specificity, presumably as a result of an avidity effect [ 35 ]. The fact that 2xFYVE can be either transfected into cells as a fusion with green fluorescent protein or another tag, or expressed in bacteria and purified as a recombinant probe that can be used directly on fixed specimens, makes this probe very versatile for monitoring the distribution of PtdIns3 P . Subsequently, other probes have been used for monitoring PtdIns3 P , including the FYVE domain of SARA, which has an intrinsic ability to dimerize and therefore does not need to be expressed as tandem fusion [ 36 ], and certain PX domains [ 37 ]. In general, the various probes have yielded comparable results, although the 2xFYVE probe has been most rigorously tested for ligand specificity. The original studies using 2xFYVE by fluorescence and electron microscopy showed that the bulk of PtdIns3 P is associated with the limiting and intraluminal membranes of endosomes at steady-state [ 35 ]. Subsequent studies have revealed that PtdIns3 P can also be detected at the plasma membrane of cells stimulated with insulin or lysophosphatidic acid [ 38 , 39 ]. Because this pool of PtdIns3 P is generated through the activity of PI3K-II [ 40 ], which has so far not been implicated in tumour suppression, it will not be considered further in the present review. Recent analyses of starved yeast cells have revealed a strong localization of PtdIns3 P on autophagosomes, especially on the inner membranes [ 41 ], and, during induction of autophagy in mammalian cells, PtdIns3 P is formed on membranes of the endoplasmic reticulum (ER) [ 42 ]. The importance of this is discussed below.
PI3K-III and PtdIns3 P binding proteins in endosomal trafficking
Because Vps34 and Vps15 were originally identified as mediators of vacular protein sorting in yeast, the first characterized functions of PI3K-III and PtdIns3 P were those associated with endosomal trafficking ( Fig. 2A ).
PtdIns3 P and endosomal fusion
Yeast cells devoid of functional Vps34 or Vps15 secrete several hydrolases that normally are transported to the lysosome-like vacuole [ 8 , 43 ], suggesting that these proteins control trafficking between the secretory and endosomal pathways and/or between endosomes. A key effector of PtdIns3 P in endocytic trafficking is the FYVE-domain-containing protein Vac1, which interacts genetically and physically with the small GTPase Vps21, and with Vps45, a member of the Sec1/Munc18 family of proteins regulating the formation of SNARE complexes that mediate membrane fusion [ 44 , 45 ]. Indeed, studies of the mammalian homologues of Vac1, Vps21 and Vps45, termed Rabenosyn-5, RAB5 and VPS45, respectively, have revealed that these proteins function in tethering and fusion reactions in the early endocytic pathway [ 46 ]. Moreover, studies in Drosophila have revealed that Rabenosyn functions to bridge Rab5 with Vps45, thereby regulating the function of the SNARE protein Avalanche in endosomal fusion [ 47 ]. Interestingly, functional interference with Rabenosyn-5 and its interacting partners causes a loss of both epithelial and planar polarity [ 47 , 48 ]. The loss of epithelial polarity is a prevailing characteristic of carcinomas, and mutation of Rabenosyn indeed causes epithelial tumours in Drosophila . To date, the mechanisms by which Rabenosyn controls epithelial polarity have not been elucidated, whereas more detailed insight is available in the case of planar cell polarity. One consequence of interferring with Rabenosyn function is that Flamingo, a determinant of planar cell polarity through the noncanonical Wnt signalling pathway, accumulates in the cytoplasm instead of translocating to polarized membrane domains [ 48 ]. Accumulating evidence suggests a link between improper development of planar cell polarity and cancer [ 49 ], and even though it is still not clear whether Rabenosyn-5 is a tumour suppressor in mammals, the epithelial and planar cell polarity maintained by this RAB5 and PtdIns3 P effector has to be considered when dissecting the tumour suppressor activities of PI3K-III.
Early studies in mammalian cells have also revealed another important PtdIns3 P effector in endosomal tethering and fusion, namely EEA1, a protein that contains a Rab5 binding domain and a FYVE domain at its C-terminus, and a distinct Rab5 binding domain at its N-terminus [ 50 ]. EEA1 forms rod-shaped dimers through parallel coiled-coil interactions, and is well suited for tethering two opposing Rab5-containing membranes [ 51 ]. The exquisite localization of EEA1 to early endosomes is probably conferred by the coincident detection of PtdIns3 P and GTP-bound Rab5 [ 50 ]. EEA1 is structurally related to Rabenosyn-5, and also interacts with SNARE molecules [ 52 , 53 ]. In a remarkable reconstitution of Rab5-mediated fusion using liposomes and recombinant SNAREs and Rab5 effectors, the inclusion of PtdIns3 P could bypass the requirement for PI3K-III, as expected based on previous studies [ 54 ]. Importantly, the omission of either EEA1 or Rabenosyn-5 was sufficient to inhibit fusion strongly, indicating that these proteins, despite their structural relatedness, have nonredundant functions in endocytic membrane fusion.
PtdIns3 P and endosomal sorting to the degradative pathway
Consistent with the fact that Vps34 was originally identified as a mediator of protein sorting, studies of both yeast and mammalian cells have shown that PI3K-III is required for proper sorting of certain membrane proteins from endosomes to vacuoles/lysosomes [ 8 , 55 ]. Moreover, interference with the function of VPS34 in mammalian cells results in dilated late endosomes devoid of intraluminal vesicles (ILVs) [ 56 , 57 ]. Recent studies have revealed that not only VPS34, but also VPS15, Beclin 1, UVRAG and BIF-1 are required for proper degradation of endocytosed epidermal growth factor receptors in lysosomes, highlighting the involvement of an entire PI3K-III complex in endosomal sorting [ 15 ].
A mechanistic explanation for these findings has emerged with the discovery of the endosomal sorting complex required for transport (ESCRT) machinery [ 58 , 59 ]. This molecular machinery, which consists of at least four subcomplexes (ESCRT-0, -I, -II and -III), is recruited to endosome membranes where it recognizes ubiquitinated membrane proteins (e.g. activated growth factor receptors and membrane-anchored hydrolases) and sorts these into ILVs. Recent reconstitution studies employing giant unilamellar vesicles have revealed that ESCRT-0, which contains as many as five ubiquitin-binding domains, serves to sequester ubiquitinated cargoes, whereas ESCRT-I and -II, which also contain ubiquitin-binding domains, serve to form invaginations of the endosomal membrane [ 60 ]. Finally, ESCRT-III is recruited, cargo is deubiquitinated by deubiquitinating enzymes recruited by ESCRT-III [ 61 ], and ESCRT-III serves to sever the neck of the forming invagination, thereby securing the inclusion of cargo proteins within ILVs [ 60 ]. The main link between PI3K-III and the ESCRT pathway is the fact that Vps27/HRS, a core component of ESCRT-0, contains a FYVE domain that mediates its recruitment to endosomal membranes through binding PtdIns3 P [ 62 ]. Vps27/HRS in turn recruits ESCRT-I through interaction with the Vps23/TSG101 subunit, so the initial recruitment of ESCRT-0 to endosomal membranes via FYVE-PtdIns3 P interactions is crucial for the function of the ESCRT machinery. In addition, a subunit of ESCRT-II, Vps36/EAP45, contains a PtdIns3 P -binding GLUE domain that is predicted to contribute to the membrane recruitment of ESCRT-II [ 32 , 34 ]. The involvement of the ESCRT machinery in protein sorting and ILV biogenesis, as well as the notion that key subunits of this machinery require PtdIns3 P for their membrane recruitment, readily explains why interference with PI3K-III functions results in impaired protein sorting and causes endosomes devoid of ILVs.
PI3K-III and PtdIns3 P binding proteins in autophagy
Macroautophagy (referred to here as autophagy) is a process that involves the sequestration of cytoplasm by a double membrane called phagophore or isolation membrane, followed by fusion of the resulting autophagosome with endosomes and lysosomes [ 63 ] ( Fig. 2B ). The degradation of the sequestered cytosolic material in autolysosomes is beneficial for the cell, both under starvation conditions (when it is crucial to recycle free amino acids by degrading cytosolic proteins that are not housekeeping), during infection with cytosolic parasites, and under various stress conditions (e.g. those that result in the accumulation of cytosolic protein aggregates that are not readily degraded by proteasomes). The exact origin of the phagophore membrane is still a matter of debate, although there are strong arguments in favour of a contribution from both ER and mitochondrial membranes [ 42 , 64 , 65 ]. At least in the case of ER membranes, there is evidence for the production of PtdIns3 P during the early phase of autophagosome formation [ 42 , 66 ]. Several lines of evidence point to a crucial role for PI3K-III in autophagy [ 66 ], and immunoelectron microscopy of starved yeast cells using the 2xFYVE probe has revealed that PtdIns3 P is enriched on inner side of the phagophore during autophagosome formation [ 41 ].
Studies in yeast have revealed that a complex consisting of Vps34, Vps15, Vps30 and Atg14 is required for autophagy [ 10 ], and subsequent work in mammalian cells has shown a similar requirement for the mammalian homologues of these proteins [ 11 , 12 ]. In addition, an involvement of the Vps38 homologue UVRAG has been reported [ 7 ], which is surprising in light of the strong evidence that Vps38 mediates endosomal trafficking and not autophagy in yeast. A possible role of UVRAG in autophagy is supported by the independent identification of BIF-1, an interactor of UVRAG, as a regulator of autophagy [ 5 ]. On the other hand, monoallelic UVRAG mutations associated with microsatellite unstable colon cancer have no effect on autophagy, and the depletion of UVRAG has an undetectable effect on autophagy in HEK293 cells, whereas endosomal sorting is affected [ 67 ]. One explanation for the conflicting findings on UVRAG may stem from the involvement of UVRAG in fusion between autophagosomes (and late endosomes) and lysosomes through its interactions with the HOPS complex [ 68 ]. Except for the catalytic activity of VPS34, little is known about the specific functions of the various PI3K-III subunits in autophagy. However, recent evidence suggests that the specific function of ATG14 in autophagy may reflect the ability of this protein to target PI3K-III to ER membranes [ 69 ].
How does PtdIns3 P mediate autophagy? The only known PtdIns3 P effector in autophagy in yeast is the Proppin protein Atg18 [ 70 ], whose binding to PtdIns3 P is required for autophagy [ 71 ]. Although the exact function of Atg18 in autophagy remains to be clarified, current evidence suggests that this protein, in complex with Atg2, facilitates the recruitment of lipidated Atg8, a key effector in autophagosome formation to phagophore membranes [ 71 ]. A mammalian Atg18 homologue, WIPI2, is recruited to phagophore membranes along with ULK1, a protein kinase that positively regulates autophagy [ 72 ]. Interestingly, the depletion of WIPI2 leads to a strong accumulation of omegasomes, comprising ER-localized PtdIns3 P -containing structures positive for DFCP1 (double FYVE domain-containing protein 1) that are considered to act as platforms for autophagosome formation. This suggests a role for WIPI2 in the progression from omegasomes into autophagosomes.
PI3K-III and PtdIns3 P binding proteins in cytokinesis
A surprising finding when using a green fluorescent protein-tagged version of the 2xFYVE probe was that PtdIns3 P accumulates in the bridge separating two dividing cells, the so-called midbody. PtdIns3 P is frequently associated with the central, electron-dense part of the midbody, referred to as the midbody ring or the Flemming body, but can also be observed on small vesicles throughout the midbody region [ 73 ]. This localization of PtdIns3 P raises the question of whether its formation is required during cytokinesis, the final stage of cell division ( Fig. 2C ). Indeed, small interfering RNA-mediated knockdown of VPS34, as well as of the accessory PI3K-III components VPS15, Beclin 1, UVRAG and BIF-1, causes an increased proportion of cells undergoing cytokinesis, suggesting a role for PI3K-III in the completion of cytokinesis [ 15 , 73 ]. Small interfering RNA screening identified the large FYVE domain-containing protein FYVE-CENT (FYVE protein on centrosomes) as a PtdIns3 P effector in cytokinesis. FYVE-CENT localizes to the centrosome in interphase cells and translocates to the midbody during telophase. This translocation appears to be mediated by the microtubule-based motor protein KIF13A. The precise function of FYVE-CENT during cytokinesis remains to be characterized, although one clue arises from the finding that TTC19, a protein that associates with FYVE-CENT and accompanies it on its translocation from the centrosome to the midbody, interacts with the ESCRT-III subunit CHMP4B [ 73 ]. This is interesting because ESCRT-III has been proposed to mediate the final membrane abscission step during cytokinesis [ 74 , 75 ], and it is possible that TTC19, brought to the midbody by FYVE-CENT and KIF13A, may be a positive regulator of CHMP4B in cytokinesis. By analogy with its yeast counterpart Vps32/Snf7, CHMP4B is predicted to form spiral-shaped oligomers that constrict to mediate membrane severing [ 76 ], and TTC19 might serve to control this oligomerization.
PI3K-III and PtdIns3 P effectors in tumour suppression
The PI3K-III subunit Beclin 1 is monoallelically deleted in a large proportion of breast and ovarian cancers, and heterozygous beclin 1 knockout mice develop spontaneous mammary tumours [ 6 ]. These findings, combined with the observation that both Beclin 1 and its yeast homologue Vps30 mediate autophagy, suggest that Beclin 1 acts as a tumour suppressor because of its function in autophagy. In support of this idea, there is evidence that autophagy functions as tumour suppressor by scavenging damaged mitochondria and peroxisomes that would otherwise cause genomic instability through oxygen radical-induced DNA damage [ 77 ]. Further supporting the notion that PI3K-III acts as a tumour suppressor through its function in autophagy is the observation that two other PI3K-III accessory proteins identified as positive regulators of autophagy, UVRAG and BIF-1, are also tumour suppressors [ 5 , 7 ]. Even though these are compelling data, one obvious question arises regarding the role of PI3K-III in autophagy-mediated tumour suppression: are other autophagy regulators also tumour suppressors? One would expect that this would be the case but, so far, only one of the many other proteins implicated in autophagy regulation has been identified as a putative tumour suppressor, the protease ATG4C [ 78 ]. Because, among the more than 30 positive regulators of autophagy, three out of four identified tumour suppressors belong to the PI3K-III complex, the possibility exists that PI3K-III may mediate tumour suppression not by autophagy but by an alternative means.
Given the importance of endocytosis and lysosomal downregulation of growth factor receptors in attenuation of mitogenic cell signalling [ 79 ], one distinct possibility is that PI3K-III could (at least in part) exert its tumour suppressor function through its role in endosome fusion and endosomal receptor sorting. Although there is no direct evidence that this is the case, it is interesting to note that the PtdIns3 P -binding endosomal fusion regulator, Rabenosyn, is a tumour suppressor in flies [ 47 ], and the same is the case with multiple components of the ESCRT machinery that acts downstream of PtdIns3 P in the endosomal sorting of mitogenic receptors [ 80 – 83 ]. Arguing against this idea is the fact that Hrs, the PtdIns3 P binding ESCRT-0 protein that initiates further ESCRT recruitment to membranes, is not a tumour suppressor in Drosophila .
The recent discovery that PI3K-III regulates cytokinesis has offered a third possible explanation for the tumour suppressor activity of this enzyme complex [ 73 ]. Inhibition of PI3K-III activity not only causes an increased proportion of cells undergoing cytokinesis, but also an increase in bi- and multinucleate cells. Under certain conditions, tetraploidy may develop into aneuploidy, which is strongly associated with cancer. It is therefore likely that incomplete cytokinesis, as experienced when PI3K-III or the PtdIns3 P effector FYVE-CENT is functionally ablated, may represent a step in oncogenesis [ 84 ]. It is interesting to note that FYVE-CENT is frequently mutated in breast cancer [ 85 , 86 ], although it remains to be established whether this PtdIns3 P effector is a genuine tumour suppressor. | I thank my mentors Sjur Olsnes and Marino Zerial, and my excellent co-workers at the Institute for Cancer Research. Research in my laboratory is generously sponsored by the Norwegian Cancer Society, the Research Council of Norway, the South-Eastern Norway Regional Health Authority, the European Research Foundation, and by an Advanced Grant from the European Research Council.
Abbreviations
early endosomal autoantigen 1
endoplasmic reticulum
endosomal sorting complex required for transport
intraluminal vesicles
phosphoinositide
phosphoinositide 3-kinase
phosphatidylinositol
phox homology | CC BY | no | 2022-01-12 15:46:43 | FEBS J. 2010 Dec; 277(23):4837-4848 | oa_package/24/f6/PMC3015057.tar.gz |
|||
PMC3015060 | 20618342 | Introduction
Enteropathogenic Escherichia coli (EPEC) strains are Gram-negative pathogenic bacteria frequently associated with infantile diarrhoea in non-industrial countries ( Chen and Frankel, 2005 ; Croxen and Finlay, 2010 ). During the course of infection, EPEC adheres intimately to the surface of enterocytes and provokes microvillus effacement, leading to the formation of characteristic attaching and effacing (A/E) lesions ( Knutton et al ., 1987 ). The ability to form A/E lesions is associated with the LEE pathogenicity island ( McDaniel et al ., 1995 ), which encodes gene regulators, the adhesin intimin, a type III secretion system (T3SS), chaperones and effector proteins ( Garmendia et al ., 2005 ). In addition, EPEC strains express a diverse number of non-LEE-encoded effectors ( Iguchi et al ., 2009 ).
Following translocation, the effector proteins are targeted to different cellular compartments and subvert numerous cellular processes to sustain colonization and multiplication ( Garmendia et al ., 2005 ). Map, EspI (also known as NleA) and NleH1 are conserved effectors and major EPEC virulence factors ( Simpson et al ., 2006 ; Hemrajani et al ., 2008 ; Lee et al ., 2008 ). NleH1 inhibits NF-κB-dependent transcription through an N-terminal interaction with ribosomal protein S3 (RPS3) ( Gao et al ., 2009 ) and prevents apoptosis through a C-terminal interaction with Bax inhibitor-1 ( Hemrajani et al ., 2010 ). EspI is targeted to the Golgi apparatus where it interacts with the Sec23/24 complex and inhibits COPII-mediated vesicular transport between the endoplasmic reticulum and the Golgi apparatus ( Kim et al ., 2007 ; Lee et al ., 2008 ). As a result, EspI contributes to the disruption of cellular tight junctions ( Thanabalasuriar et al ., 2010 ). Map is targeted to the mitochondria via an N-terminal mitochondrial targeting sequence where it disrupts the mitochondrialmembrane potential ( Kenny and Jepson, 2000 ; Papatheodorou et al ., 2006 ). At early time points Map induces formation of transient filopodia at the bacterial attachment sites ( Kenny and Jepson, 2000 ) via its guanine exchange factor (GEF) activity for the Rho-GTPase Cdc42 ( Huang et al ., 2009 ). Moreover, via its C-terminus, which comprises a class I PSD-95/Disc Large/ZO-1 (PDZ)-binding motif (TRL), Map binds PDZ 1 of Na + /H + exchanger regulatory factor 1 (NHERF1) ( Alto et al ., 2006 ; Simpson et al ., 2006 ). PDZ domains are protein–protein recognition domains of about 90 amino acids, which are widely represented in the human genome ( Tonikian et al ., 2008 ). PDZ-binding motifs comprise a consensus S/T-X-Φ sequence (where X is any amino acid and Φ is any hydrophobic residues ( Songyang et al ., 1997 ). Importantly, PDZ binding motifs are also found at the C-termini of NleH1 and EspI (SKI and TRV respectively). Whereas no mammalian partner was thus far associated with the PDZ-binding motif of NleH1, a previous report demonstrated that EspI binds a number of PDZ-containing proteins, including NHERF1 and NHERF2 ( Lee et al ., 2008 ).
NHERF proteins (NHERF1 to 4) are present in abundance in the mammalian small intestine and colon ( Donowitz et al ., 2005 ) where they have a central role in trafficking, membrane retention, dimerization and regulation of ion channels and membrane proteins ( Shenolikar and Weinman, 2001 ; Shenolikar et al ., 2004 ). NHERF1 and 2 are closely related (51% amino acid identity) and the only members in the NHERF protein family which possess two PDZ domains and a C-terminal ezrin/radixin/moesin (ERM) binding domain (EBD) ( Reczek et al ., 1997 ; Yun et al ., 1998 ). The EBD domain links NHERF1 and 2 to the actin cytoskeleton thereby forming a scaffold for highly organized multimeric signalling complexes (reviewed in Lamprecht and Seidler, 2006 ). More than 30 binding partners have been reported for NHERF1 and 2 ( Shenolikar et al ., 2004 ), some are shared by more than one NHERF isoform ( Yun et al ., 1998 ; Sun et al ., 2000 ), while others are NHERF isoform-specific ( Shenolikar et al ., 2004 ).
In this paper we report the binding of three T3SS effector proteins (Map, EspI and NleH1) to a single eukaryotic target (NHERF2). We show that NHERF2 influences the trafficking and function of EspI, Map and NleH1, suggesting that it functions as a distribution hub at the plasma membrane. | Results
Map, EspI and NleH1 bind NHERF2
Previous reports have demonstrated that Map interacts with NHERF1 ( Alto et al ., 2006 ; Simpson et al ., 2006 ), and a global PDZ array screen indicated that EspI might bind NHERF1 and NHERF2 ( Lee et al ., 2008 ). In this study we screened target proteins, which are recognized by the PDZ-motifs of Map and NleH1 using the PDZ array. Overlaying a fusion consisting of the last 50 amino acids of Map and GST (GST-MapC50) on the PDZ-domain array confirmed that Map binds PDZ1 of NHERF1 ( Fig. 1A ). In addition, we found that Map interacted with PDZ2 of NHERF1 and both PDZ domains of NHERF2 ( Fig. 1A ). NHERF2 was also identified as interaction partner of Map by a yeast two-hybrid screen using full-length Map as bait and HeLa cDNA library as a prey (data not shown). Overlaying a fusion consisting of the last 150 amino acids of NleH1 and GST (GST-NleH1C150) (a GST-NleH1C50 fusion was unstable) on the PDZ-domain protein array revealed that NleH1 bound PDZ2 of NHERF2, although this interaction appeared weaker than either Map:NHERF2 or EspI:NHERF2 ( Fig. 1A ) ( Lee et al ., 2008 ).
In order to confirm and characterize the interaction of Map, EspI and NleH1 with NHERF2, we constructed and purified His-NHERF2 and MBP fusions of NHERF2 PDZ1, PDZ2 and EBD, which were immobilized on a membrane and overlaid with purified MBP-Map, His-EspI and GST-NleH1. Detection of bound proteins with effector-specific antiserum revealed that Map, EspI and NleH1 interacted with full-length NHERF2 ( Fig. 1B ). Map and EspI, but not NleH1, also bound the purified PDZ2 domain of NHERF2 ( Fig. 1B ). Although Map interacted with PDZ1 of NHERF2 on the PDZ array, neither Map nor EspI bound the purified PDZ1 or EBD domains ( Fig. 1B ). These results show that despite some differences between binding to the arrayed and fusion proteins, NHERF2 is targeted by Map, EspI and NleH1.
In a reciprocal experiment full-length Map, EspI and NleH1 were overlaid with His-NHERF2, which confirmed the interactions ( Fig. 1C, D and E ). This binding was dependent on the PDZ-binding motifs of Map and EspI as NHERF2 did not bind the C-terminally truncated effectors MapΔC3 ( Fig. 1C ) or EspIΔC7 ( Fig. 1D ). Unfortunately, due to protein instability of NleH1ΔC4, the role of the PDZ binding motif of recombinant NleH1 in binding NHERF2 could not be investigated.
NHERF2 co-immunoprecipitates EspI and NleH1 and colocalizes with Map, EspI and NleH1
Once we confirmed the interaction between NHERF2 and Map, EspI and NleH1 using recombinant proteins, we aimed to confirm the interaction by coimmunoprecipitation. Towards this end we generated a stable HeLa cell line constitutively expressing HA-tagged NHERF2 (HeLa–NHERF2). Immunofluorescence microscopy using anti-HA antibodies revealed that NHERF2 localized mainly at the plasma membrane (data not shown), in a pattern similar to that reported in un-polarized epithelial cells like A431 or HT1080 cells, which express NHERF2 endogenously ( Theisen et al ., 2007 ). The HeLa–NHERF2 cells were infected for 1 h with EPEC Δ map and EPEC Δ map expressing full-length Map or MapΔC3; EPEC Δ espI and EPEC Δ espI expressing full-length EspI or EspIΔC7; and EPEC Δ nleH and EPEC Δ nleH expressing full-length NleH1 or NleH1ΔC4. We were unable to co-immunoprecipitate NHERF2 and full-length Map or MapΔC3 (data not shown). Probing infected cell extracts with anti-EspI and anti NleH1 revealed that while both full-length EspI and EspIΔC7 were found in equal amounts, NleH1ΔC4 was unstable ( Fig. 2A and B ). Following cell lysis the supernatants were subjected to co-immunoprecipitation with anti-HA antibodies. While NHERF2 did not co-immunoprecipitate EspIΔC7 ( Fig. 2A ) and NleH1ΔC4 ( Fig. 2B ), both full-length EspI ( Fig. 2A ) and NleH1 ( Fig. 2B ) were co-immunoprecipitated.
In order to determine if the effectors colocalize with NHERF2, HeLa–NHERF2 cells were transfected with pRK5 ( Fig. 3 ) or pRK5 expressing myc-tagged Map, MapΔC3, NleH1 and NleH1ΔC4 as well as pGFP-EspI and pGFP-EspIΔC7. Analysing the transfected cells by immunofluorescence revealed that Map colocalized with mitochondria ( Fig. S1 ) and NHERF2 ( Fig. 3 ). Plasma membrane-associated EspI and NleH1 colocalized with NHERF2 ( Fig. 3 ). We observed no colocalization of NHERF2 and MapΔC3, EspIΔC7 or NleH1ΔC4 ( Fig. S2 ).
NHERF2 modulates Map-induced filopodia dynamics
We investigated the effect of NHERF2 expression on Map-induced filopodia. First, HeLa and HeLa–NHERF2 were transfected with pRK5 expressing myc-tagged Map or MapΔC3 and examined by scanning electron microscopy. This revealed that ectopic expression of Map resulted in filopodia formation in both cell lines, although the length and complexity of filopodia were more prominent on the HeLa–NHERF2 cells ( Fig. 4A ). Relatively small and sporadic filopodia were observed in either cell line ectopically expressing MapΔC3 ( Fig. 4A ). Next HeLa and HeLa–NHERF2 cells were infected with EPEC wild-type and the dynamics of filopodia formation and withdrawal was quantified by counting infected cells with filopodia over time. This revealed that at each time point more HeLa–NHERF2 cells with filopodia were observed compared with HeLa cells ( Fig. 4B ).
To investigate whether persistence of filopodia was due to the PDZ-mediated Map–NHERF2 interaction, the number of cells with filopodia was quantified following infection of HeLa and HeLa–NHERF2 cells with EPEC Δ map expressing Map or MapΔC3. Upon infection with EPEC Δ map expressing full-length Map, the fraction of infected HeLa–NHERF2 cells with filopodia was higher and decreased much slower compared with infected HeLa cells ( Fig. 4C ). This is consistent with the phenotype observed for infection with wild-type EPEC ( Fig. 4B ). In contrast, no difference in numbers of cells with filopodia or the kinetics of filopodia withdrawal was found between HeLa–NHERF2 and HeLa cells infected with EPEC Δ map expressing MapΔC3 ( Fig. 4C ). These results suggest that the expression of NHERF2 in HeLa cells modulates Map-induced filopodia formation and persistence, which is dependent on the PDZ-mediated Map–NHERF2 interaction.
NHERF2 accelerates trafficking of EspI to the Golgi
During the early stages of EPEC infection, localization of EspI shifts from the site of bacterial adhesion to the Golgi apparatus, and this trafficking is enhanced by its PDZ binding motif ( Lee et al ., 2008 ). In order to investigate the possible role of NHERF2 in trafficking of EspI to the Golgi, HeLa and HeLa–NHERF2 cells were infected for 15, 30, 60 and 90 min with EPEC Δ espI expressing EspI. The number of infected cells with EspI colocalizing with the Golgi protein GM130 was quantified by immunofluorescence microscopy ( Fig. 5A ). At 15 and 30 min post infection, significantly more HeLa–NHERF2 cells showed EspI staining at the Golgi apparatus compared with HeLa cells ( Fig. 5A ). To determine the importance of the EspI–NHERF2 interaction for trafficking of the protein to the Golgi apparatus, HeLa–NHERF2 cells were infected for 15, 30, 60 and 90 min with EPEC Δ espI expressing EspI or EspIΔC7 ( Fig. 5A ). Golgi staining of EspIΔC7 was significantly lower after 30 ( Fig. 5A and B ), 60 and 90 min post infection compared with cells infected with EPEC Δ espI expressing EspI ( Fig. 5A ).
Taken together, these results suggest that following translocation EspI interacts with NHERF2 via its PDZ-binding motif, which accelerates trafficking of EspI to the Golgi apparatus.
NHERF2 antagonises the anti-apoptotic activity of NleH1
NleH1 interacts with Bax inhibitor-1, which protects the cell against various apoptotic stimuli ( Hemrajani et al ., 2010 ; Robinson et al ., 2010 ). We investigated the impact of NHERF2 expression on the anti-apoptotic activity of NleH1. Low translocation levels of NleH1ΔC4 prevented an analysis of the importance of the PDZ-binding motif during infection. To circumvent this, we ectopically expressed NleH1 and NleH1ΔC4 in HeLa or HeLa NHERF2 cells, which were treated with the pro-apoptotic compound staurosporine (STS). The number of transfected apoptotic cells, marked by caspase-3 activation or nuclear condensation, was quantified by immunofluorescence counting. Significantly fewer HeLa cells transfected with pRK5-NleH1 displayed caspase-3 activation ( Fig. 6A and B ) or nuclear condensation ( Fig. 6C ) compared with pRK5-transfected HeLa cells or cells transfected with pRK5-NleH1ΔC4. Mock-transfected HeLa–NHERF2 cells displayed a similar phenotype as normal HeLa cells, showing that NHERF2 expression per se neither promotes nor prevents STS-induced apoptosis ( Fig. 6B and C ). Expression of NHERF2 did not influence the phenotype of cells expressing NleH1ΔC4. However, the inhibition of caspase-3 cleavage ( Fig. 6A and B ) and nuclear condensation ( Fig. 6C ) by NleH1 was significantly abolished in cells expressing HA-NHERF2 compared with the HeLa cell control. Taken together, these results show that even though the anti-apoptotic activity of NleH1 depends on its PDZ-binding motif, the interaction with NHERF2 is not the trigger for its anti-apoptotic signalling. The data rather suggest that NHERF2 counteracts the anti-apoptotic activity of NleH1, possibly by direct competition for its PDZ binding site or interference with its intracellular trafficking and localization. | Discussion
NHERF proteins are known to orchestrate intracellular protein trafficking, protein localization and cell signalling; regulation of ion transporters being the best-studied example. NHERFs, which are modular proteins comprising several protein–protein interaction domains, function as scaffolds for the assembly of multiprotein complexes at the plasma membrane. In particular, NHERFs are involved in tethering plasma membrane localized transporters such as NHE3, the apical form of the Na + /H + exchanger and CFTR, a cAMP-regulated Cl-channel, to the underlying actin cytoskeleton. NHERF1 and 2 bind the transporters via their PDZ domains and actin-associated ERM proteins via their C-terminal EBD ( Brone and Eggermont, 2005 ), thus linking the actin cytoskeleton to the plasma membrane ( Bretscher et al ., 2000 ). The assignment of individual functions to NHERF1 or 2 has proven to be difficult due to high homology and widely overlapping interactomes. Localization studies have shown that, in polarized Caco-2 cells, NHERF1 is predominantly found in the brush border (BB) whereas NHERF2 has some BB localization but is mainly found in the inter-microvillar clefts below the BB ( Donowitz et al ., 2005 ). The differential subcellular localization and tissue distribution of NHERFs might determine functional specificity and account for the observation from studies with knockout mice that NHERF1–3 are not functionally redundant. ( Seidler et al ., 2009 ). Despite their central position in diverse signalling processes, the role of NHERF proteins as versatile targets for bacterial effector proteins has not been fully investigated.
Although fulfilling unrelated intracellular functions Map, EspI and NleH1 share two common post-translocation features, binding to NHERF2 and trafficking to an intracellular membranous compartment. Map is targeted to the mitochondria ( Kenny and Jepson, 2000 ), EspI to the Golgi ( Gruenheid et al ., 2004 ) and NleH1 to the endoplasmic reticulum (ER) ( Hemrajani et al ., 2010 ). A PDZ array screen indicated that Map, EspI and NleH1 bind NHERF2. Using recombinant Map, EspI and NleH1 and a stable NHERF2 cell line revealed that all three effectors interact, in a PDZ binding motif-dependent manner, with NHERF2. Moreover, all three effectors colocalized with NHERF2 in HeLa–NHERF2 cells, and EspI and NleH1 were co-immunoprecipitated with NHERF2 from infected cells. Our efforts to co-immunoprecipitate Map and NHERF2 were not successful. The fact that expression of NHERF2 in HeLa cells modulated the intracellular trafficking and/or activity of all three effectors suggests that NHERF2 might form a molecular hub at the plasma membrane from which the effectors are distributed to their final destinations.
Map acts as a Rho-guanine nucleotide exchange factor (GEF) for Cdc42 ( Huang et al ., 2009 ), but is unique in the family of WxxxE motif proteins in possessing a PDZ-binding motif at the C-terminus ( Alto et al ., 2006 ; Bulgin et al ., 2010 ). Approximately 37% of human and mouse GEFs also possess a PDZ binding motif and there is increasing evidence that PDZ-based interactions are important for GEF function, and aid targeting of the GEF to a subcellular location ( Garcia-Mata and Burridge, 2007 ). Here removal of the C-terminal PDZ motif abolished the interaction of Map with NHERF2 and led to reduced filopodia formation and stability. Endogenous GTPases are modified posttranscriptionally by the addition of a lipid moiety to the C-terminus (farnesyl, geranyl, palmitoyl or methyl) ( Roberts et al ., 2008 ), which targets them to different membranous compartments. We suggest that the interaction of Map with the PDZ2 domain of NHERF2 may transiently anchor Map at the apical surface of an infected cell leading to filopodia formation beneath adherent bacteria, before the effector is targeted to the mitochondria.
EspI has been reported to inhibit COPII-dependent vesicular transport between the ER and Golgi leading to alterations in host cell protein secretion and tight junction integrity ( Kim et al ., 2007 ; Thanabalasuriar et al ., 2010 ). EspI dynamics and trafficking upon translocation are still poorly understood since the protein accumulates at the Golgi apparatus and not at the ER, where COPII vesicles are formed ( Kim et al ., 2007 ). Here we showed that the interaction between EspI and NHERF2 is mediated through the EspI C-terminal PDZ-binding motif, resulting in an increased rate of trafficking of EspI to the Golgi. This result is consistent with the potential role of NHERF2 in recycling of endogenous membrane proteins ( Donowitz et al ., 2005 ).
Recently, we found that NleH1 inhibits cell apoptosis by interacting with Bax inhibitor-1 and decreasing activation of caspase-3, nuclear condensation and mitochondrial hyper permeabilization ( Hemrajani et al ., 2010 ; Robinson et al ., 2010 ). NleH1 also binds to ribosomal protein 3 and reduces nuclear activity of NF-κB ( Gao et al ., 2009 ). Similar to Map and EspI, we found that NleH1 possesses a class I PDZ binding motif that regulates its anti-apoptotic function. Although a recombinant PDZ-binding motif mutant form of NleH1 could not be purified because of protein instability, immunofluorescence of cells transfected with the PDZ motif mutant demonstrated that the motif was critical to wild-type function of NleH1 and its interaction with NHERF2. However, the interaction with NHERF2 is not the trigger for NleH1-mediated anti-apoptotic signalling; on the contrary, binding of NleH1 to NHERF2 inhibits anti-apoptotic activity. This might represent a mechanism for the temporal control of NleH1 anti-apoptotic function. In addition, NleH1-NHERF2 interaction might enable NleH1 in general, or a fraction of the translocated protein, to perform a secondary, yet unidentified, activity before or in parallel to promoting cell survival.
The fact that three different T3SS effectors, which share little sequence similarity and have distinct functions, bind a single host cell scaffold protein is unprecedented. Moreover, while NleH1 binds NHERF2 only, EspI and Map bind both NHERF1 and NHERF2. It is tempting to speculate that EPEC and EHEC exploit endogenous BB-associated NHERFs to regulate spatial and temporal intracellular trafficking of Map, EspI and NleH1. Indeed, during infection of mice with wild-type C. rodentium NHERF2 was extensively recruited to the bacterial attachment sites (data not shown).
T3SS effectors function in a co-ordinated manner. As effectors from a single pathogen can complement or antagonize each other, the timing and location of their activity must be highly regulated. While timing of gene expression and effector protein translocation provides a regulatory mechanism at the bacterial level ( Mills et al ., 2008 ), ubiquitination followed by proteasomal degradation ( Kubori and Galan, 2003 ) and protein phosphorylation ( Kenny, 1999 ) represent mechanisms, which are exploited to control the activity of translocated effectors within host cells. In this study we identified a novel regulatory mechanism by which recruitment of NHERFs to the bacterial attachment site forms a distribution hub, which is exploited for membrane retention (Map) or trafficking (EspI). We propose that differential binding to NHERF1 and/or NHERF2 and to their different PDZ domains provides a mechanism to fine tune the activity of the hub. We aim to exploit the C. rodentium model to test this hypothesis in the context of host pathogen interactions in vivo . | Re-use of this article is permitted in accordance with the Terms and Conditions set out at http://wileyonlinelibrary.com/onlineopen#OnlineOpen_Terms
Enteropathogenic Escherichia coli (EPEC) strains are diarrhoeal pathogens that use a type III secretion system to translocate effector proteins into host cells in order to colonize and multiply in the human gut. Map, EspI and NleH1 are conserved EPEC effectors that possess a C-terminal class I PSD-95/Disc Large/ZO-1 (PDZ)-binding motif. Using a PDZ array screen we identified Na + /H + exchanger regulatory factor 2 (NHERF2), a scaffold protein involved in tethering and recycling ion channels in polarized epithelia that contains two PDZ domains, as a common target of Map, EspI and NleH1. Using recombinant proteins and co-immunoprecipitation we confirmed that NHERF2 binds each of the effectors. We generated a HeLa cell line stably expressing HA-tagged NHERF2 and found that Map, EspI and NleH1 colocalize and interact with intracellular NHERF2 via their C-terminal PDZ-binding motif. Overexpression of NHERF2 enhanced the formation and persistence of Map-induced filopodia, accelerated the trafficking of EspI to the Golgi and diminished the anti-apoptotic activity of NleH1. The binding of multiple T3SS effectors to a single scaffold protein is unique. Our data suggest that NHERF2 may act as a plasma membrane sorting site, providing a novel regulatory mechanism to control the intracellular spatial and temporal effector protein activity. | Experimental procedures
Bacterial strains and growth conditions
Escherichia coli strains used in this study are listed in Table 1 . Bacteria were cultured in Luria-Bertani (LB) medium or in Dulbecco's modified Eagle's medium (DMEM) supplemented with ampicillin (100 μg ml −1 ) or kanamycin (50 μg ml −1 ) as appropriate.
Plasmids
Plasmids and primers used in this study are listed in Tables 2 and 3 . For cloning into the bacterial expression vector pSA10 ( Schlosser-Silverman et al ., 2000 ), espI, espIΔC7 and nleH1ΔC4 were amplified from EPEC genomic DNA using primer pairs EspI-Fw1 and EspI-Rv1, EspIΔC7-Rv1 and NleH1-Fw1 and NleH1ΔC4-Rv, respectively. For generation of (His) 6 -tagged proteins, nherf2 was amplified from pBKCMV-HA-NHERF2 using primers NHERF2-Fw and NHERF2-Rv and espI was amplified from EPEC genomic DNA using primers EspI-Fw2 and EspI-Rv2; both PCR products were cloned into pET28a(+) (Novagen). For generation of MBP fusion proteins, map was amplified from pICC330 using primer pair Map-Fw2 and Map-Rv2, and PDZ1, PDZ2 and the ezrin-binding domain (EBD) of NHERF2 were PCR-amplified from pBKCMV-HA-NHERF2 using primer pairs PDZ1-Fw and PDZ1-Rv, PDZ2-Fw and PDZ2-Rv, EBD-Fw and EBD-Rv; all constructs were cloned into pMalc2X (New England Biolabs). For generation of GST fusion proteins, nleH1 and nleH1C150 (encoding the 150 C-terminal amino acids of NleH1) were amplified from EPEC genomic DNA using primer pairs NleH1-Fw2 or NleH1C150-Fw and NleH1-Rv2 respectively. Both constructs were cloned into pGEX-KG. For mapC50, the region encoding the 50 C-terminal amino acids of Map was amplified from EPEC genomic DNA using primer pairs MapC50-Fw or MapC50-Rv. The resulting fragment was cloned into the SmaI/EcoRI sites of pGEX-3X. For transfection experiments, map , nleH1 and nleH1ΔC4 were amplified from EPEC genomic DNA using primer pairs Map-Fw1 and -Rv1, NleH1-Fw1 and NleH1-Rv1 or NleH1ΔC4-Rv, respectively, and cloned into pRK5. Plasmid pRK5-MapΔC3 was generated by inverse PCR using pRK5-Map as template. espI and espI ΔC7 were amplified from EPEC genomic DNA using primer pairs EspI-Fw3 and EspI-Rv3 or EspIΔC7-Rv3, respectively, and cloned into pGFP-C2. All constructs were verified by DNA sequencing.
Protein expression and purification
Genes cloned into pET28a, pMALc2X and pGEX vectors were expressed in E. coli BL-21/DE3 star strain (Invitrogen) in presence of 1 mM isopropyl-β- d -thiogalactopyranoside (IPTG) at 37°C as described previously ( Crepin et al ., 2005 ). His-tagged proteins were purified using Ni 2+ agarose His-Bind Resin Column (Novagen), MBP-tagged proteins were purified on Amylose resin (New England Biolabs) and GST-tagged proteins were purified on glutathione-sepharose (GE Healthcare) using manufacturer's recommendations.
PDZ array screen
To assess the binding of the GST-MapC50 and NleH1C150 fusion proteins to the PDZ domain array, purified His-tagged PDZ domain fusion proteins were spotted as previously described ( Fam et al ., 2005 ; He et al ., 2006 ) at 1 μg per bin onto Nytran SuperCharge 96-grid nylon membranes (Schleicher & Schuell). The membranes were allowed to dry overnight and then blocked in ‘blot buffer’ (2% non-fat dry milk, 0.1% Tween-20, 50 mM NaCl, 10 mM Hepes, pH 7.4) for 30 min at room temperature. The arrays were then overlaid with control GST, GST-MapC50 or GST-NleH1C150 fusion proteins (100 nM in blot buffer) overnight at 4°C. The overlaid arrays were washed three times for five minutes each with 20 ml blot buffer, incubated with anti-GST horseradish peroxidase (HRP)-conjugated antibody (Amersham, 1:4000) for 1 h at room temperature, washed again three times for 5 min each with 20 ml blot buffer, and ultimately visualized via chemiluminescence with the ECL kit from Pierce.
Yeast two-hybrid
Yeast two-hybrid screen using Map as prey and cDNA library as bait was performed as described before ( Simpson et al ., 2006 ).
Western and far Western blot
Far Western blot was performed using 10 μg ml −1 of His-NHERF2, His-EspI, MBP-Map or GST-NleH1 in Tris buffered saline (TBS) 0.1% Tween20 with 1% skimmed milk followed by Western blot detection as previously described ( Simpson et al ., 2006 ). Western blots were performed following standard methods using rabbit anti-NHERF2 (raised against purified human His-NHERF2, Covalab), anti-EspI ( Lee et al ., 2008 ), anti-NleH (raised against purified His-NleH1, Covalab), mouse anti-Map (Immune Systems limited), anti-poly-Histidine (Sigma), anti-GST (AbCam) and anti-HA HRP-conjugated (Roche) antibodies. Goat anti-mouse and anti-rabbit IgG HRP conjugates (Jackson laboratories) were used as secondary antibodies.
Transfection and generation of NHERF2 stable cell line
HeLa cells were routinely maintained in DMEM containing 10% fetal calf serum (FCS) and 1 mM l -glutamine (Gibco) in a humidified atmosphere of 5% CO 2 at 37°C. Transfection with pBKCMV, pRK5 and pEGFP vectors and their derivatives was performed using lipofectamin 2000 (Invitrogen) according to the manufacturer's recommendations.
To generate the stable cell line, HeLa cells were grown to confluence in 75-cm 2 cell culture flasks and transfected with pBKCMV-HA-NHERF2 using lipofectamin 2000 (Invitrogen) as previously described ( Oh et al ., 2004 ). After 16 h, cells were trypsinized and 2 × 10 5 cells were sub-cultured in 100 mm Petri dishes in DMEM media supplemented with 800 μg ml −1 geneticin (Invitrogen). Cells were cultured for 10 days in the same Petri dish with repeated change of media for selection of resistant clones. Single clones were transferred to 96-well plates (Nalgene) and grown to confluence. Clones were then transferred to 24-well plates and replicate 24-well plates containing coverslips for expression analysis by immunofluorescence microscopy. Positive clones were trypsinized, sub-cultured in 25 cm 2 flasks and then stored at −80°C in DMEM supplemented with 10% dimethylsulfoxide (DMSO). In this article, the stable cell line is referred as HeLa–NHERF2.
Infection of HeLa cells
Cells were seeded onto glass coverslips 48 h prior to infection at a 70% confluence and maintained in DMEM supplemented with 10% FCS at 37°C in 5% CO 2 . Three hours before infection, cells were washed three times with PBS and the media replaced with fresh DMEM without FCS. To prime bacteria for infection, DMEM was inoculated with overnight cultures at a dilution of 1:50 of the appropriate bacteria as previously described ( Collington et al ., 1998 ). Recombinant protein expression was induced by addition of 1 mM of IPTG 30 min prior to infection. Infections were carried out at 37°C in 5% CO 2 at a multiplicity of infection (MOI) of 100 for each time point (15, 30, 60 or 90 min).
Co-Immunoprecipitation (Co-IP)
HeLa–NHERF2 cells were grown to confluence in 25 cm 2 cell culture flasks and infected for 1 h as described above. Cells were washed 3 times with PBS and lysed in 500 μl of Co-IP buffer [50 mM Tris-HCl, pH 7.5, 100 mM NaCl, 1% NP40, 0.5% sodium deoxycholate, 10% glycerol, protease inhibitors (Complete, Roche)]. The lysate was transferred to a pre-chilled eppendorf tube and centrifuged at 9200 g for 10 min at 4°C. The cleared lysate was transferred to a fresh pre-chilled eppendorf tube containing 15 μl of pre-equilibrated anti-HA agarose beads (Sigma) and incubated on a spinning wheel for 2 h at 4°C. The suspension was washed three times in Co-IP buffer followed by brief centrifugation at 9200 g at 4°C after each wash. Bound proteins were eluted using 80 μl of 100 μg ml −1 HA peptide (Sigma). Analysis of the samples was performed by Western blot as described before.
Immunofluorescence staining and microscopy
Following infection or transfection, the coverslips were washed three times in phosphate-buffered saline (PBS) and fixed with 3% paraformaldehyde (PFA) for 20 min before being washed 3 more times in PBS. The fixed cells were quenched for 20 min with PBS 50 mM NH4Cl, permeabilized for 4 min in PBS 0.25% Triton X-100 and washed three times in PBS. The samples were blocked for 1 h with PBS 5% bovine serum albumin (BSA) prior to incubation with primary and secondary antibodies. The primary mouse anti-HA (Covance) and anti-GM130 (BD transduction laboratories) as well as rabbit anti-Myc (AbCam) and anti-cleaved caspase-3 antibodies (Cell Signalling Technology) were used at a dilution of 1:200. Rabbit anti-O127 (kindly provided by Dr Roberto La Ragione, Veterinary Laboratory Agency, UK) was used at a dilution of 1:500. The samples were incubated with the primary antibody for 1 h, washed three times in PBS and incubated with the secondary antibodies for 1 h. Donkey anti-rabbit IgG conjugated to a Cy3, Cy5 or AMCA fluorophore and donkey anti-mouse IgG conjugated to Cy2, Cy3 or Cy5 fluorophores (Jackson ImmunoResearch) were used at a 1:200 dilution. Mouse anti-Myc FITC-conjugated antibody (Sigma) was used at a 1:100 dilution. Actin was stained using Rhodamine phalloidin (Invitrogen) at a 1:500 dilution and DNA was stained using Hoechst 33342 (Molecular Probes). All dilutions were prepared in PBS containing 5% BSA. Following three washes in PBS, coverslips were mounted on SuperFrost glass slides using Prolong Gold antifade reagent (Invitrogen) and visualized with a Zeiss Axioimager immunofluorescence microscope. All images were analysed using the Axiovision Rel 4.5 software.
Apoptosis assay
For apoptosis studies, 1 μM staurosporine (STS) (Calbiochem) was added to the media during 4 h and apoptotic cells (stained either for nuclear condensation or caspase-3 activation) were counted using immunofluorescence microscopy as described ( Hemrajani et al ., 2010 ).
Scanning electron microscopy (SEM)
HeLa and HeLa–NHERF2 cells transfected with pRK5-Map or pRK5-MapΔC3 were washed three times with PBS and fixed with 3% glutaraldehyde in PBS. The samples were then washed with PBS three times before being post fixed in 1% Osmium Tetroxide for 1 h. Following three washes in PBS and 15 min dehydration in graded ethanol solutions (50% to 100%), the samples were transferred to an Emitech K850 Critical Point drier and processed according to the manufacturer's instructions. The samples were coated with gold/palladium using an Emitech Sc7620 minisputter to a thickness of approximately 370 Å and examined at an accelerating voltage of 20 kV using a Jeol JSM-6390 electronic microscope.
Statistical analysis
Results are expressed as mean values ± standard deviation. All statistical tests were performed using the program Sigma Plot version 11.0 using the parametric Student's t -test. P < 0.05 was considered as significant. | We thank Erin Garcia for technical assistance. This work was supported by grants from the Australian Research Council and National Health and Medical Research Council (ELH), the Wellcome Trust and the Medical Research Council (GF).
Supporting information
Additional Supporting Information may be found in the online version of this article:
Fig. S1. Fluorescence microscopy of HeLa–NHERF2 cells transfected with pRK5 (Mock) and pRK5-map. Mitochondria were detected with Mitotracker (red) and Map was detected with anti-myc (green). Map colocalizes with the mitochondria. Scale bar = 10 μm.
Fig. S2. Fluorescence microscopy of HeLa–NHERF2 cells transfected with pRK5 (Mock), pRK5-mapΔC3, pEGFP-C2-espIΔC7 and pRK5-nleH1ΔC4. NHERF2 was detected with anti-HA (green) and effectors Map and NleH1 were stained with anti-myc (red). Green GFP signal was converted to red for EspI. Neither effector colocalized with NHERF2. Scale bar = 10 μm.
Please note: Wiley-Blackwell are not responsible for the content or functionality of any supporting materials supplied by the authors. Any queries (other than missing material) should be directed to the corresponding author for the article. | CC BY | no | 2022-01-12 15:46:43 | Cell Microbiol. 2010 Dec 3; 12(12):1718-1731 | oa_package/df/d6/PMC3015060.tar.gz |
||
PMC3015062 | 21069831 | The beauty of the developing embryo, and the awe that it inspires, lure many scientists into the field of developmental biology. What compels cells to divide, migrate, and morph into a being with a complex body plan? Evolutionary developmental biologists hold similar fascinations, with dynamics that take place on a grander timescale. How do phenotypic traits diverge over evolutionary time? This primer illustrates how a deep understanding of the basic principles that underlie developmental biology have changed how scientists think about the evolution of body form. The primer culminates in a conversation with David Stern, PhD, and Michael Shapiro, PhD, who discuss current topics in morphological evolution, why the field should be of interest to classic developmental biologists, and what lies ahead. Developmental Dynamics 239:3497–3505, 2010. © 2010 Wiley-Liss, Inc. | EVOLUTION OF A HYPOTHESIS
No one could mistake a human forelimb for a bird wing. Or could they? After peeling back layers of skin and muscle, one finds the same basic skeletal elements in both: scapula, humerus, radius and ulna, wrist, and digits. This morphologic homology supports the theory that the species' appendages stem from a common ancestor. However, even conserved skeletal elements vary in design. For example, to aid in flight, its wrist and digit bones are fused. Whale flippers also share the basic skeletal elements that birds and humans do, but they have additional phalanges for added length and flexibility. For decades, experimental evidence supported the then prevailing notion that the great variation observed among homologous traits is driven by mutations in the protein-coding regions of genes (Hoekstra and Coyne, 2007 ; Stern and Orgogozo, 2008 ).
Intuitively, protein-coding sequence would seem to be the most fertile ground from which mutations can execute changes in gene function. After all, in one fell swoop, a nonsynonymous change in amino acid sequence, or a large-scale genetic change such as gene loss, rearrangement, or duplication, can produce immediate effects. These changes can alter protein structure, stability, activity, and/or produce genes with novel function. Nevertheless, the so-called “coding mutation hypothesis” has fallen out of favor in recent years.
(R)EVOLUTIONARY DEVELOPMENTAL PRINCIPLES
The eventual demise in popularity of the coding hypothesis can be traced in part to the molecular techniques explosion of the 1980s. The ability to rapidly and thoroughly analyze DNA sequence and gene expression, and to manipulate gene expression and function, revealed that there are several core principles that guide much of embryonic development. Described below are three of the key principles that produced ripple effects that extended beyond classical developmental biology, eventually re-shaping the theory of phenotypic evolution.
Deep Homology: Making the Most of Pre-existing Information
The discipline of evolutionary development (evo-devo) was born from the discovery that single genes, and even entire gene networks, retain similar functions across species. One of the first examples illustrating the concept was the discovery that orthologs of the paired box transcription factor, Pax6, regulate development of two vastly different eye types: the vertebrate single lens eye and the Drosophila compound eye. Drosophila eyeless ( ey ) and mouse Pax6 not only share extensive sequence identity and similar expression patterns, but they are to a certain extent functionally interchangeable. The latter is demonstrated by the striking finding that misexpression of either mouse or fly Pax6 in Drosophila imaginal discs induces ectopic ommatidia (Fig. 1 A; Halder et al., 1995 ). What's more, although there are clear differences between mouse and Drosophila gene networks that regulate eye development, the networks do have several genes in common such as eyes absent/Eya and sine oculis/Six (Hanson, 2001 ). Based on these and similar findings, the phrase “deep homology” was coined to describe the concept that analogous structures are derived from similar genetic mechanisms (Shubin etal., 1997 ).
Beyond similar structures, regulatory genetic circuits can also appear in diverse anatomical contexts. Pax6 null mice also fail to form lens as well as nasal placodes (Hogan etal., 1986 ). Of interest, nasal development also involves shared expression of Pax6 , Six , and Eya (Purcell et al., 2005 ). Therefore, it is tempting to speculate that deployment of the gene program during nasal development may represent co-option of an ectodermal placode based program at a novel anatomical site. These and other data suggest that rather than building novel structures from scratch, nature instead creates morphological diversity by modulating existing information (Shubin et al., 2009 ).
Mosaic Pleiotropy: The Cons of Recycling
Many of the key players in conserved gene regulatory networks are “toolkit genes,” transcription factors and signaling molecules that are deployed repeatedly throughout embryogenesis. Their ultimate function depends upon when and where they are expressed, whether they have been integrated into different genetic regulatory networks, and other variations in the genetic landscape (epigenome, transcriptome, etc.). For example in addition to eye development, Pax6 also regulates neural, pancreatic, and pituitary development (van Heyningen and Williamson, 2002 ). As important developmental regulators, toolkit genes are poised to shape morphological evolution.
The question of how to change a gene that plays several developmental roles is a tricky one. Changes in coding sequence could wreak pleiotropic consequences that would likely be detrimental to the fitness of the organism. By extension, the common belief that gene duplication and divergence is a mechanism of change is also unlikely because developmental processes are generally sensitive to changes in gene dosage (Carroll, 2008 ). On the other hand, since genes that regulate physiological processes (i.e., metabolism, enzyme function, etc.) tend to reside at the terminal end of genetic regulatory networks, an animal may better tolerate these types of mutations in physiological genes (Stern and Orgogozo, 2008 ). Therefore, because the rules of the evolution of function may differ from the rules of the evolution of form, coding mutations are perhaps less likely to be driving forces of phenotypic evolution.
Modular cis -Regulatory Enhancers: A Means for Incremental Change
If not by coding changes, then what are the molecular mechanisms of phenotypic evolution? The answer lies in the observation that modular cis -regulatory elements (CREs) govern the complex expression pattern of many toolkit genes. Again using mouse Pax6 as an example, its expression is regulated by multiple enhancers: individual CREs direct expression in the pancreas, neural retina, diencephalon, and other sites (van Heyningen and Williamson, 2002 ; Kleinjan et al., 2004 ; Zhang et al., 2006 ). The expression patterns governed by individual CREs join to make a complete picture. Importantly for the theory of morphological evolution, removal of single elements can alter the picture without destroying it. Therefore, mutations in CREs are an effective means for altering specific aspects of gene expression, for rewiring genetic regulatory networks, and for creating phenotypic diversity (Carroll, 2008 ).
PROVING A HYPOTHESIS
While there are certainly exceptions to the cis -regulatory hypothesis (Hoekstra and Coyne, 2007 ), accumulating evidence supports that it is an important and often used mechanism of morphological change. Below are three examples from the literature that support the cis -regulatory hypothesis, and also introduce new hypotheses. These works operate under the notion that comparisons between closely related species are required to pinpoint the mutations that cause trait divergence.
Trichome Loss in Drosophila
Stern and colleagues explore the genetic basis for the loss of the hair-like trichomes from the dorsal side of each segment in larval Drosophila . The naked cuticle phenotype independently evolved at least three times: in D. sechellia , a member of the D. melanogaster species group, and in two members of the D. virilis species group, which diverged from D. melanogaster approximately 60 million years ago. Through genetic mapping, they found that variation in the gene shavenbaby ( svb ) caused the naked cuticle trait in D. sachellia (Sucena et al., 2003 ). What's more, because svb expression correlates precisely with altered trichome patterns in divergent species, the authors reasoned that regulatory changes in svb expression account for the phenotypic differences. Indeed, a follow-up paper shows that changes in three separate svb enhancers, that individually make small changes in trichome patterning, cumulatively cause loss of dorsal trichomes (McGregor et al., 2007 ).
Why might svb be a site for convergent evolution? Stern describes svb as an “input–ouput” gene that lies at an integral position within its genetic regulatory network (Stern and Orgogozo, 2009 ). If mutations were to occur in genes that regulate svb (i.e., wingless , hox , hedgehog ), they would be more likely to cause pleiotropic effects. If mutations were to occur in any of the dozens of svb -regulated genes that comprise the trichome cell differentiation program, they may produce morphological changes that are too specialized or too minute to offer any selective advantage. On the other hand, changes within svb itself can specifically alter an entire module of differentiation genes. For these reasons, Stern argues that input–output genes are hotspots for evolutionarily relevant mutations.
Pelvic Reduction in Sticklebacks
Here, Kingsley's group continues their search for mutations that cause an adaptive pelvic-reduced phenotype in the threespine stickleback fish, observed as a loss of the prominent serrated pelvic spine and hind fin skeleton (Shapiro etal., 2004 ; Chan et al., 2010 ). Fine genetic mapping localizes the cause of the phenotypic change to an intergenic region 23 kb upstream of Pitx1 , a conserved “toolbox” transcription factor that, among other functions, regulates hindlimb and hindfin development in vertebrates. Within the intergenic region, the authors identify a 2.5 kb “ Pel ” CRE that drives pelvic-specific expression of GFP in transgenic fish, and rescues pelvic structures when fused with a Pitx1 minigene in pelvic-reduced fish. Demonstrating the evolutionary importance of the region, 9 of 13 pelvic-reduced, and zero of 21 pelvic-complete natural stickleback populations have staggered deletions that overlap with the Pel CRE. They further show that the Pitx1 locus resides within an unusually flexible region of the genome that may be susceptible to double-stranded DNA breaks and repair, offering a possible explanation for this case of convergent morphological convergent evolution. The authors postulate that fragile DNA may also contribute to parallel evolution of other phenotypes.
Novel Spotted Wing Patterns in Drosophila
In this work, Carroll's group seeks the genetic mechanism for the evolution of a complex pattern of 16 vein-associated wing spots in Drosophila guttifera . They find that a single 277bp “ vein spot ” CRE upstream of yellow , a gene required for wing pigmentation, can drive GFP expression in the 16 spot pattern in transgenic D. guttifera . The absence of the vein spot CRE in D. melanogaster may be the reason that this species group does not bear wing spots. However, the vein spot CRE is present in D. deflecta which harbor a less complex variant of the D. guttifera pattern. Therefore, the cause of the different patterns in the two species groups may lie in the inducer of the vein spot CRE. In search of an inducer, the activity of the vein spot CRE was tested in the genetically well-defined D. melanogaster , and was found to direct reporter expression in a pattern similar to that of wg . Indeed, wg expression presages the spot patterns in D. guttifera and D. deflecta , and ectopic wg is able to drive ectopic black wing pigmentation in transgenic D. guttifera . Another important observation is that wing spots of multiple species always develop at physical wing landmarks: crossveins, campaniform sensillum, and vein fusion points.
From these and other data, the authors surmise that the D. guttifera wing spot pattern is a product of stepwise evolution. First, wg was expressed at a limited number of physical wing landmarks. Next wg became associated with pigmentation, partially through evolution of the vein spot CRE. Finally, wg evolved new sites of expression by adopting existing patterning information already present at additional physical wing landmarks. Of note is the fact that wg is a toolkit gene that has multiple functions in early development. D. guttifera can likely tolerate the extra spots of wg expression because they arise in the terminal stages of pupal wing development. Regardless, this view illustrates that an important feature of evolution is the layering of new information onto pre-existing gene activity patterns.
A CONVERSATION WITH THE EXPERTS
It remains to be seen whether the new hypotheses emanating from the works above will withstand the test of time. Moreover, as scientists discover additional principles of developmental biology, for example the guiding roles of epigenetics and noncoding RNAs, the theory of phenotypic evolution will continue to evolve. Evolutionary developmental biologists David Stern, PhD, and Michael Shapiro, PhD (Fig. 2 ), offer their perspectives on these and related topics in the field.
Developmental Dynamics: What is your lab's research focus?
David Stern: We have two foci at the moment. First, we study how development has evolved between closely-related Drosophila species to cause morphological divergence. Our most advanced project involves identifying the nucleotide changes of the shavenbaby locus that have led to divergence in larval morphology. Our finding, that multiple, small-effect mutations have accumulated in the cis -regulatory region of shavenbaby , has led us more recently to investigate questions of the structure and function of cis -regulatory regions. The second major focus of the lab involves studies of the evolution of behavior between closely-related Drosophila species. We are studying problems such as the evolution of courtship song. This is a new direction for the lab and these are early days. I am cautiously optimistic that we will have something interesting to say soon.
Michael Shapiro: We're interested in the developmental and genetic basis of evolutionary diversity among vertebrates. Our main focus is stickleback fish, which have served as important models for studies of behavioral, life history, and morphological variation for over a century. Beginning only approximately a decade ago, they have also emerged as important models for understanding the genetic and developmental changes that control ecologically relevant differences among populations. Sticklebacks are great models for this type of work because they show tremendous variation within and among species. Multiple species within the stickleback family have converged upon very similar adaptive phenotypes, giving us a unique chance to ask whether the same or different genes control similar evolutionary changes in independent evolutionary lineages.
Dev Dyn: What initially provoked your interest in this field?
D.S.: I came to evolutionary developmental biology through an unlikely path. I became interested in developmental biology while working on my PhD in the jungles of Malaysia on the behavior and evolution of soldier-producing aphids. Aphid colonies can produce multiple different phenotypic forms with the same genome and each individual aphid can develop along one of multiple possible developmental trajectories. Studying aphids in the field, I became increasingly interested in the mechanisms that allow aphids to develop into dramatically different body forms, such as winged versus unwinged forms, sexual versus asexual forms, and normal individuals versus soldiers. Also, the aphids I studied induced galls on trees. I spent most of my time in the field hunting down these galls. I think anyone who has worked on galls will tell you that it is easy to become obsessed with the question of how insects induce gall development on plants. This sterile field-dreaming about the mechanisms of development led me to the conclusion that I needed to learn more about the molecular mechanisms that generate diversity.
When I started looking for a postdoc in 1993, I had never studied developmental biology or molecular genetics and I was unaware of any of the recent literature on the evolution of development. I had read only work that approached development from an evolutionary perspective, such as the Maynard-Smith et al. review on constraints (Maynard Smith etal., 1985 ), Rupert Riedl's ideas on systems biology (Riedl, 1977 ), and books such as The Evolution of Individuality (Buss, 1987 ), and several of John Bonner's books (for example, Bonner, 1974 , 1988 ). While these piqued my interest in development, they did not provide a useful guide to contemporary developmental biology.
I stumbled into Michael Akam's lab for a postdoc through a series of serendipitous encounters in Cambridge, UK, catalyzed by my friend and collaborator, William Foster, an expert on aphid biology. I wish I could say that I studied the literature closely and chose to work with Michael after careful consideration of the best workers in the field. But, the truth is, I was flying blind and I had a fantastic first discussion with Michael, which convinced me that I should work with him. Of course, I doubt that an informed search would have led me to a better choice of person to introduce me to the study of developmental biology with an evolutionary twist.
M.S.: My route to this field was also different than most others in evo-devo and evolutionary genetics—I started out in paleontology. As an undergraduate student at Berkeley, I took a few paleontology courses and one of my professors was kind enough to take me on a summer field trip to dig dinosaurs and their contemporaries in the Southwest, and I was instantly hooked. Fieldwork remains an important part of my lab's work to this day, and I think it's important for student and postdocs to see their organism of choice in the real world rather than just in a tube in the lab. I gradually became more interested in the developmental processes that led to the types of diversity I saw in the fossil record, and for my dissertation project, I studied the developmental basis of limb reduction in a genus of Australian skinks that had different species with 5, 4, 3, or 2 fingers and toes. These lizards provided a great opportunity to find some molecular correlates of digit loss—for example, changes in the expression pattern of a key limb development gene—but we couldn't really get at the genetic basis of these limb changes because we couldn't interbreed the different species to take a forward genetic approach. In other words, we could see that gene expression was changing, but we couldn't definitively tell whether there were changes to the candidate genes we had chosen, or changes to an upstream signal affecting our candidates.
As I was getting close to finishing my dissertation, one of my committee members told me about a great new project that David Kingsley's lab had recently started on stickleback genetics, and that some of the populations of these fish were missing their “limbs.” The big advantage of sticklebacks is that you can do genetics and let the fish tell you which genes are important, rather than using a candidate approach like we did with the lizards. From there, we can test the developmental roles of the genes we identify. To me, this was an incredibly powerful approach to understanding which genes are actually important in adaptive evolution.
Dev Dyn: What three papers have most impacted your research and why?
D.S.: Limiting this response to anything close to three papers is difficult because I have learned so much from a huge number of papers. So, I will focus on three papers that are not usually on the radar of developmental biologists, but which influenced my thinking significantly.
As an undergraduate, I was fascinated by circadian rhythms and by Drosophila courtship song and I have a very clear memory of reading Zehring et al. ( 1984 ) when I was working in Chip Aquadro's lab at Cornell University in the 1980s. During my undergraduate years, I was fascinated by both genetics and the evolution of behavior and I very much hoped to combine these interests. This paper, which demonstrated that the aberrant behavior of a period mutant could be rescued by P-element transformation with the native locus, provided a clear indication of how to go about this kind of work. While this work has provided inspiration, even since my earliest days in science, it took me many years to find a tractable way to address these questions in an evolutionary framework. We are now undertaking this work on the evolutionary genetics of behavior and I cannot begin to express the thrill of finally working on questions that I have carried with me since my undergraduate days.
I could have chosen several of Cathy Laurie's extraordinary papers as one of my top three. One paper stands out (Stam and Laurie, 1996 ), because it was the first paper that provided a clean experimental dissection of a naturally occurring evolutionary variant. Previous work had implicated primarily an amino-acid polymorphism as the major, if not the only, determinant of variation in alcohol dehydrogenase activity in natural populations of D. melanogaster . Through a series of elegant molecular dissection experiments, this paper clearly showed that multiple additional polymorphisms, including multiple noncoding polymorphisms, within the Adh gene contribute to levels of Adh expression. I think this work was far ahead of its time and demonstrated the clear importance of cis -regulatory variation. Unfortunately, I think this work has been somewhat overlooked by workers in the field of evolutionary developmental biology.
I first read Boucher etal. ( 1992 ) when I started work on my book, approximately 5 years ago. I was shocked by the predictable spread of mutations in a population of HIV during antiviral treatment of infected individuals. While this extraordinary level of predictability probably results, at least in part, from the relatively small and simple genome of HIV, these observations inspired me to be open to the possibility that such predictability might be widespread. These, and similar, observations led me to think about why this predictability might exist. This is a problem I am still pursuing and which I think will be one of the unifying themes of evolutionary developmental biology in the future.
M.S.: The first paper we read in my undergraduate evolutionary biology class was “Evolution and Tinkering” by François Jacob ( 1977 ). The message was quite simple, but it had a profound impact on the way I thought about biology. In short, Jacob's argument was that evolution “acts” more like a tinkerer than an engineer, selecting for modifications to existing genes and processes to generate variation rather than designing new phenotypes from scratch. The paper was published in 1977 , long before the genomics era, yet it's still highly relevant. We now know, for example, that a rather large set of critical developmental genes—the so-called “genetic toolkit”—is highly conserved throughout the Metazoa. A lot of differences among organisms have more to do with when and where these genes are expressed, and slight tweaks to the coding regions, rather than complete overhauls of gene function or the evolution of completely new gene families. What we consider the evolution of novel structures often involves the redeployment of existing gene networks in new ways, not the evolution of entirely new genes. Jacob's paper is now the basis for the first writing assignment in the undergraduate course that I teach.
As an undergraduate and graduate student, I became very interested in vertebrate limb development and the developmental basis of limb diversity, and I was excited by the work of Pere Alberch and Neil Shubin (Shubin and Alberch, 1986 ), and John Saunders (Saunders, 1948 ). I was really blown away by the work coming out of Cliff Tabin's lab (Riddle et al., 1993 ), and a few other labs, that connected molecular mechanisms to the embryological processes that others had observed. This work gave me hope that we could eventually identify the genetic targets of evolutionary “tinkering” that led to diversity in vertebrate limbs and the skeleton in general.
Dev Dyn: Work from David's lab supports the theory that genes at integrative positions in developmental networks are genetic “hotspots” for phenotypic evolutionary changes. David, do hotspots reflect neutral processes or natural selection? Mike, what is your perspective on this theory?
D.S.: Since all neutral mutations fix in populations at the same rate, the only way hotspots could reflect neutral processes is if evolutionary hotspots were mutational hotspots. Since I don't think this is a good general explanation, this leaves only natural selection. There must be something about the mutations that occur at particular genes in genetic networks that make them more favorable, on average, than mutations at other genes in the network.
I think there are at least two facts that lead to this bias. The first is that mutations at different genes in a network can have widely different pleiotropic effects. In our work on shavenbaby , most, or perhaps all, mutations in genes that act upstream of shavenbaby —and that would result in similar phenotypic changes—would be very likely to have pleiotropic effects on segmentation generally. The second fact is that some genes “control” entire modules of morphogenesis. For example, shavenbaby activity regulates a large number of downstream effector genes that build trichomes. No individual gene acting downstream of shavenbaby can induce this entire module of morphogenesis.
Shavenbaby may be an extreme case, but the regulatory network containing shavenbaby is certainly similar, in overall architecture, to how we tend to envision hierarchies of gene regulation underlying development. That is, I think hotspot genes are probably abundant and that hotspots arise from the architecture of the genetic network. Natural selection is very discriminating and, if I were to go out on a limb, I think selection has led to a superabundance of mutations causing phenotypic evolution at hotspots.
M.S.: I think the idea of “hotspots” in integrative network positions is an important and testable hypothesis. David's work on svb certainly supports the idea, and to a lesser extent, the Pitx1 story in sticklebacks is consistent with it as well. I agree with David that we might see an overrepresentation of changes at these integrative positions, but I'd also like to see some additional examples before deciding that this is a predominant and predictable mechanism of evolutionary change. Examples from the evolution of certain phenotypes, such as changes in the sequence and expression of genes that control evolutionary variation in vertebrate pigmentation, provide potential counterexamples to part of this hypothesis. For instance, Mc1r is unquestionably a hotspot of repeated mutation leading to evolutionary variation (Hoekstra, 2006 ), but I would argue that it does not hold the same kind of integrative network position that David's work on svb demonstrates.
Dev Dyn: David Kingsley's lab showed that independent mutations in pitx1 contribute to pelvic variation in natural populations of threespine stickleback fish. Interestingly, the mutations occur in a genome region that is thought to be more susceptible to deletions (Chan etal. , 2010 ). Might this reflect a general trend?
D.S.: In general, natural selection, rather than mutation, is expected to be a much stronger force influencing the frequency of alleles in populations. So, the specific question about fragile regions is whether they are more likely to throw up adaptive variation than are other kinds of mutations. This seems rather unlikely to me. The most compelling data from studies of species difference in Drosophila , including some of our unpublished work, indicates that normally phenotypic evolution is not driven by unusual kinds of mutations.
It is unlikely that large populations are often mutation limited; they usually contain an abundance of mutations available for natural selection to act upon. However, species with smaller population sizes, like Darwin's finches and, perhaps, stickleback populations that invade new lakes, may be mutation limited. In this case, the increased frequency of mutations in fragile regions may bias evolution toward use of these mutations. Also, if selection acts strongly, like it appears to do in these stickleback populations, then mutations of rather large effect—such as mutations that cause dramatic changes to gene function—may tend to be selected more often than are mutations of more subtle effect.
I think that the specific ways in which genomes have evolved, and thus the resulting assemblage of mechanisms that guide development, is dependent on quirks of population history. This is a major theme of my recent book Evolution, Development, & the Predictable Genome (Stern, 2010 ).
M.S.: Again, I think we know relatively little about the nucleotide-level changes that are responsible for evolutionary diversity in metazoans. This is especially true for regulatory (as opposed to coding) changes, which likely account for a great deal of phenotypic diversity. With this in mind, I don't think we can yet determine whether the types of mutations seen in the Pitx1 regulatory region are common or unusual. These types of fragile regions are almost certainly present across metazoan genomes, and it will be interesting to see if the distribution of these sites corresponds to the genes that underlie adaptive variation. While fragile sites are probably subject to higher mutation rates, what's not clear is whether a disproportionate number of these mutations are selected and persist. It's also worth noting that transposable elements can potentially have similar effects on the genome as fragile sites. In addition to disrupting genes or their regulatory regions, these elements can lead to deletions in the genome, as well as unequal recombination. Transposable elements are likely important factors in human disease, and in the stunning diversity in the domestic dog, whose genome is littered with a particularly active type of these elements. It seems likely to me that these elements could play a role in natural diversity as well.
Dev Dyn: microRNAs are a means of manipulating gene expression that are unique to metazoans. As such, it has been put forth that they have been important for the evolution of complex body plans (Christodoulou et al ., 2010 ). What are your thoughts about miRNAs, or other noncoding RNAs, in evolution?
D.S.: My thoughts on this topic are similar to my thoughts on the contributions of all kinds of genomic changes to phenotypic evolution. Namely, we currently have too limited a set of data collected in a sufficiently unbiased manner to answer any of these questions with rigor. At the moment, we have examples of miRNA targets that have evolved in domesticated populations, but to my knowledge, no examples of miRNAs themselves, or miRNA targets in natural populations, that have contributed to natural variation. Given the fact that miRNAs obviously have evolved, it is almost certainly just a matter of time before such examples are found. I think it is probably unlikely that evolution of miRNAs played a large role in phenotypic evolution in natural populations, or we would probably have already discovered some examples.
But, I must emphasize that we are dealing currently with a small sample of the mutations known to contribute to phenotypic evolution in natural populations, and, for various reasons that I have discussed in other forums (Stern and Orgogozo, 2008 ), there are good reasons to believe that the vast majority of our current examples come from biased studies. For example, if an investigator maps a phenotypic difference to a genomic region carrying an obvious protein-coding candidate gene, then are they likely to look for alternative explanations? At the moment, the most compelling evidence for the molecular changes contributing to phenotypic evolution in the wild and between species does not include examples of miRNAs or other noncoding RNAs. But, I expect that they will be found eventually.
M.S.: I agree with David's caution and the biases that he highlights. Most of the work in evolutionary molecular genetics tends to focus on genes, either their coding regions or regulatory elements. Little is known about the evolutionary roles of noncoding RNAs, not because we think they aren't important, but rather because the field has paid little attention to the topic thus far.
Dev Dyn: How does the field of evolution shed light on mechanisms of development?
M.S.: As a colleague once told me, nature has been doing mutagenesis on a scale that even NIH funding can't touch. The enormous organismal variation we see in the wild must have resulted, at some point, from changes in developmental programs, and this giant, worldwide mutant screen gives us a tremendous resource with which to explore the molecular basis of developmental processes. Evolutionary studies can shed light on the generality of developmental processes—are the same genes always involved in the generation of similar phenotypes across different groups?—and can illuminate more basic developmental mechanisms as well. For example, the threespine stickleback was an ideal organism for finding the cis -regulatory region controlling pelvic expression of Pitx1 because this species shows dramatic, natural variation in pelvic morphology. This evolutionary study resulted in the identification of a cis -regulatory element that is surely present, yet has not been identified, in the more traditional models of vertebrate development.
On a more practical note, over the past few decades, comparative studies of developmental genetics have also led to the realization that the same gene networks are present throughout metazoans. This remarkable conservation has led to the adoption of a wide range of organisms, not just mice, as genetic models for normal and abnormal human development and variation.
D.S.: Perhaps the most important way that evolution sheds light on mechanisms of development is to recognize that all genomes have evolved, both in response to natural selection and through neutral processes. Genome architecture and function must reflect this history of selection and drift. I sometimes feel that there is a subtle tendency in the field of developmental biology to view the genome as an engineered product, instead of viewing the genome as the product of evolution, a product of bricolage, as François Jacob put it. Viewing the genome as a product of a long history of natural selection and drift immediately flags three important things that we should look for in the genome.
First, the contingency of evolution suggests that “solutions” to development that we observe today may appear like odd ways of doing things. Development may take three rights to make a left. For example, the Drosophila melanogaster male genitalia undergoes a 360° clockwise rotation during development to end up precisely where it would have ended up without rotating. Presumably this reflects an evolutionary history during which an ancestor of D. melanogaster developed with a genitalia that rotated 180° to facilitate back-to-back or face-to-face mating. There is, obviously, a set of developmental mechanisms in place to accomplish the 360° rotation of the genitalia that is, currently, a complete waste. But evolution built upon what was already there and now the poor fly is stuck with these wasteful mechanisms. It is difficult to argue that this is the optimal way for a genitalia to develop.
Second, much of the genome of most species has evolved through neutral processes. In recent years, Michael Lynch has been asking how, precisely, neutral processes might influence genome complexity, such as the origin of genes, the prevalence of gene duplication, and the like. His general result is that neutral processes provide an abundant source of genome complexity. These are tricky hypotheses to test robustly, but, again, the major point is that many aspects of the structure and function of genomes may not reflect adaptation according to any measure of optimality.
Third, against the backdrop of historical contingency and genetic drift, natural selection plays a widespread and discriminating role in sculpting genome structure and function. The extent of natural selection on genome function has only recently started coming into clear focus. For many years, most molecular evolutionary biologists assumed that selection (both negative and positive) acted primarily on nonsynonymous amino acid substitutions and that most of the noncoding regions of genomes were subject to neutral or near-neutral processes. It is now clear that natural selection (both negative and positive) is pervasive throughout many noncoding regions of metazoan genomes (Andolfatto, 2005 ). Many genomic regions show strong patterns of evolutionary conservation that cannot be explained by current functional assays. For the genes I have worked on, on a good day I like to think that we can explain maybe 5–10% of the evolutionarily conserved DNA associated with these genes.
Dev Dyn: What are some up-and-coming techniques that promise to propel the evo-devo field forward?
M.S.: My perception is that evolutionary developmental biology and evo-devo increasingly mean evolutionary genetics, and this is especially true for those of us interested in microevolutionary questions. With this in mind, I think the rapidly evolving set of genomics techniques are starting to have a huge impact on the field, and this impact will only increase. Genomics is hardly a new set of techniques, of course, but until recently these tools were largely unavailable to those of us studying weird organisms. Due to funding constraints, labs studying the genetic basis of human disease were the main beneficiaries of techniques that examined variation across the genome, but we're now starting to see studies emerge involving the genomic basis of adaptation in humans and a modest number of other organisms. The precipitous decrease in sequencing costs is making it easier (but still by no means trivial) to sequence, assemble, and annotate a reference genome for your favorite organism. This reference, coupled with comparative data from other individuals, populations, or species can be an important entry point to finding genomic regions under selection or associated with important phenotypes.
D.S.: Overall, we should look for dramatic improvements in the ability to manipulate gene structure and function in nonmodel systems. Improvements in transgenic technology are opening up opportunities to manipulate gene function in systems such as crustacea and a diversity of insects. I think we should keep an eye on engineered zinc fingers and TAL effectors for the ability to target manipulations to almost any genomic region in a wide variety of systems. These systems are gaining ground in model systems, and it is only a matter of time before they are transferred to nonmodel systems. These tools open up the possibility of performing homologous recombination in a diversity of organisms, which, really, will be the gold standard for how to test evolutionary hypothesis of functional molecular evolution.
I agree with Michael that genomics technologies are vital to progress on nonmodel systems, and I will reiterate his point that assembling and annotating new genomes is far from trivial, even with a big pot of money. In my opinion, the scale of this problem has been grossly underestimated; high sequence throughput does not equal easy genome assembly.
Finally, one area where high-throughput sequencing is making a big difference fast is in genetic mapping experiments. I think next generation sequencing, combined with some new approaches to thinking about genetic mapping, will quickly make most other genotyping platforms obsolete and will provide high resolution genetic mapping in a wide and interesting diversity of systems.
Dev Dyn: What are some exciting ideas that are emerging?
M.S.: I'm intrigued by the potential roles of epigenetics in the evolution of development. Strong evidence is emerging that nongenetic changes such as DNA methylation, which can have profound effects on gene expression and transposon activity, can be inherited transgenerationally. Epigenetic marking can occur in response to environmental stimuli, so environmental factors can potentially influence development for multiple generations. If these epigenetic changes occur in response to new habitats, then epigenetic process could play a key role in generating new phenotypes during adaptive radiations. These ideas are not new, but we still have few empirical examples of how these processes might work.
D.S.: In addition to questions that I have discussed above, such as our ignorance of the functions of large genome regions, I will add a few more favorites. First, the history of evolution is primarily a history of changes in the shapes and sizes of organs. We currently have almost no understanding of the mechanisms underlying evolutionary changes in shape and size. This ignorance reflects partly our still elementary understanding of shape and size control in model systems. I think that serious progress on these problems will require first some dedicated work on model systems.
Second, there has been vanishingly little work done on the role of precise changes in gene expression levels and timing to evolutionary transitions. Most work has focused on rather dramatic changes in gene expression patterns. This emphasis reflects, in part, what has been possible. As methods for quantifying temporal patterns of gene expression improve, I expect that we will see more work exploring the role of subtle changes in the temporal dynamics of gene expression in evolutionary transitions.
Third, based on some of our recent observations (Frankel et al., 2010 ), I expect that metazoan genomes encode lots of apparently redundant enhancers. Our recent findings suggest that some of the apparently redundant enhancers have been evolutionary conserved and contribute to phenotypic robustness. The developmental sources of robustness—or canalization, as Waddington called it—have been mysterious for decades and I think we are just beginning to get a window into these mechanisms. Robustness seems to reflect selection acting in the real world, a world of variable temperature, variable sources of nutrition, and of variable genomes. Studying development in the laboratory, under tightly controlled conditions, means that we are studying only a tiny fraction of how the genome functions. I foresee an explosion of studies into how the genome responds to variable environments and whether and how this contributes to developmental evolution. | Many thanks to David Stern and Mike Shapiro for their time, cooperation and insightful comments. | CC BY | no | 2022-01-12 15:46:43 | Dev Dyn. 2010 Dec 9; 239(12):3497-3505 | oa_package/c8/84/PMC3015062.tar.gz |
|||||
PMC3015063 | 20636473 | Introduction
Kinetoplastid parasites of the genus Leishmania cause a diverse spectrum of infectious diseases, the leishmaniases, in tropical and subtropical regions of the world ( Murray et al ., 2005 ; Chappuis et al ., 2007 ; Reithinger et al ., 2007 ). Mammalian-infective Leishmania species are divided into two subgenera, Leishmania ( Leishmania ) and Leishmania ( Viannia ), that partially differ in their developmental cycles within the female sand fly vector (reviewed in Lainson et al ., 1987 ). In both subgenera, however, Leishmania undergo transformation from the intracellular amastigotes taken up in the sand fly blood meal to flagellated promastigotes of different morphological forms (described below, using the terminology of Walters, 1993 and Cihakova and Volf, 1997 ). Completion of the parasite life cycle by transmission from vector to mammalian host requires promastigote differentiation into non-replicative metacyclic parasites. These forms are inoculated when the female sand fly takes a second blood meal ( Bates, 2007 ); the parasites enter resident dermal macrophages and transform into replicative amastigotes that can be disseminated to other tissues, often inducing immuno-inflammatory responses and persistent infection. The fate of these intracellular parasites determines disease type, which can range from cutaneous or mucocutaneous infection to diffuse cutaneous or the potentially fatal visceral leishmaniasis ( Murray et al ., 2005 ; Chappuis et al ., 2007 ; Reithinger et al ., 2007 ).
Metacyclogenesis in Leishmania , the end-point of parasite development in the vector, is induced in vitro by low pH and nutrient depletion, while reduced tetrahydrobiopterin levels may also act as a signal for parasite differentiation ( Cunningham et al ., 2001 ; Kumar et al ., 2007 ; reviewed in Bates, 2008 ). Metacyclic parasites display distinctive morphological and biochemical features: they have a small cell body and relatively long flagellum, are highly motile and resistant to human complement, therebyfacilitating parasite survival in the host following transmission ( Da Silva et al ., 1989 ). Complement resistance is associated with presence of an extensive glycocalyx composed chiefly of a complex lipid-anchored glycoconjugate, lipophosphoglycan (LPG; Turco and Descoteaux, 1992 ). Stage-specific expression of LPG and the parasite surface metalloprotease, GP63, have been monitored in sand flies by immunohistochemistry ( Davies et al ., 1990 ; Saraiva et al ., 1995 ). Importantly, it has been shown that modification of LPG during metacyclogenesis facilitates detachment of Leishmania major from the midgut in the sand fly species, Phlebotomus papatasi , and is essential for vector transmission ( Sacks and Kamhawi, 2001 ; Kamhawi, 2006 ). Relatively little is known, however, of parasite factors that promote later stages of development and their role in metacyclic transmission. Here we show that the stage-specific HASP (hydrophilic acylated surface protein) and SHERP (small hydrophilic ER-associated protein) proteins of L. major are essential for metacyclogenesis in the vector P. papatasi .
HASP and SHERP genes are encoded on the same segment of chromosome 23 (originally named the LmcDNA16 locus; Flinn and Smith, 1992 ) in both Old and New World Leishmania species, while similar but divergent sequences are found at the same location in the genome of L. ( Viannia ) braziliensis (D. Depledge, unpublished). The Leishmania -specific HASPs, characterized mainly by work on HASPB, are expressed on the plasma membrane of infective parasite stages only (metacyclics and amastigotes) and show both inter- and intra-specific variation, principally in their characteristic repetitive amino acid domains ( Flinn et al ., 1994 ; Rangarajan et al ., 1995 ; McKean et al ., 1997a ; Alce et al ., 1999 ; Denny et al ., 2000 ). The smallest HASPs, the HASPAs ( McKean et al ., 1997b ), lack these repeats, which are the only HASP domains showing some similarity to other proteins: specifically to the peptidoglycan- and immunoglobulin-binding domains of several bacterial surface proteins ( Flinn et al ., 1994 ). The HASPs are dually acylated by the N-terminal addition of myristate and palmitate, co- and post-translational modifications that are essential for plasma membrane trafficking ( Denny et al ., 2000 ). HASP function has been investigated in mutant parasites generated by targeted deletion of the whole LmcDNA16 locus ( McKean et al ., 2001 ). In vitro , these null parasites can undergo metacyclogenesis, are taken up by macrophages and survive in numbers comparable to wild type; in vivo they cause more rapid infection than wild-type parasites in susceptible BALB/c mice. In contrast, null parasites complemented by re-expression of the LmcDNA16 locus from an episome (that causes constitutive overexpression) are completely avirulent, probably due to pleiotrophic effects ( McKean et al ., 2001 ).
The LmcDNA16 null parasites described above are also deleted for the SHERP genes, found in close proximity within the HASP locus in all Leishmania species examined to date. SHERP is expressed in metacyclic parasites in culture, being the only well-validated protein marker exclusive to this stage (and not expressed in amastigotes; Knuepfer et al ., 2001 ). The SHERP open reading frame is expressed as a 6.2 kDa peripheral membrane protein that localizes, in wild-type metacyclics, to the cytosolic face of the ER and the outer mitochondrial membrane. Recent structural analysis suggests that SHERP is induced to fold by interaction with membrane lipids (B. Moore, unpublished) but the function of this unusual small protein is otherwise unknown.
Given the specific and high-level expression of HASP and SHERP products in metacyclics, and the critical role of these parasite stages in successful parasite transmission, we are using null and complemented mutant lines to investigate HASPB and SHERP function in the sand fly. Here, we confirm that stage-specific expression of each protein, as observed in culture, is also found in the vector although the detection of SHERP precedes that of HASPB. In contrast to earlier in vitro observations, however, loss of both proteins in the null parasites results in failure to produce metacyclics, decreased production of short promastigotes and lower colonization of the stomodeal valve (SV) region in late-stage infections in the sand fly. Conversely, complementation of the whole locus restores metacyclic production and SV colonization, while complementation with either HASPB alone or SHERP alone partially restores the wild-type phenotype. These data suggest that the HASP/SHERP proteins are critical for development of wild-type parasites in the sand fly and may therefore be essential in vector transmission. | Results
Expression of HASPB and SHERP during differentiation in culture
HASPB and SHERP expression have been shown previously to correlate with parasite differentiation in culture, using mixed populations of promastigotes grown from log to stationary phase and sampled at fixed time points ( Flinn et al ., 1994 ; Knuepfer et al ., 2001 ). To confirm the relative temporal expression of these two stage-regulated proteins, freshly differentiated L. major Friedlin promastigotes, derived from amastigotes isolated from the lymph nodes of susceptible mouse strains (as described in Depledge et al ., 2009 ), were subject to minimum passage in culture prior to analysis over a time-course. Growth is not synchronized under these conditions, as shown by microscopic analysis of the relative numbers of procyclic, pre-metacyclic and metacyclic parasites over time (see Experimental procedures ). Parasites were sampled at 24 h intervals from day 2 post-inoculation (when 99% of cells are procyclic) and total cell lysates (containing equivalent parasite numbers) analysed by immunoblotting with antibodies specific for HASPB and SHERP. As shown in Fig. 1A , both proteins show increased expression from day 4 (when 70% or cells are pre-metacyclic and 25% metacyclic) to day 7, when > 50% of viable parasites are of high motility, resistant to peanut lectin agglutination and can be classified as metacyclic promastigotes (see Fig. 2D ). Comparison of the relative signals for the two proteins by densitometry, however, shows that SHERP expression is stabilized by 6 days while HASPB expression continues increasing to day 7 ( Fig. 1B ), suggesting that there are temporal differences in the expression of each molecule in these mixed parasite populations.
Generation of new complemented lines for vector transmission studies
Previous analysis of genetically manipulated clones of L. major Friedlin either null for or complemented with the complete LmcDNA16 locus encoding HASP and SHERP genes failed to show a phenotype distinct to wild-type parasites in culture or after macrophage infection in vitro and in vivo ( McKean et al ., 2001 ). In these experiments, the ‘add-back’ parasites were avirulent, an observation interpreted as an overexpression phenotype due to the excessive amounts, and loss of regulated expression, of the HASP and SHERP proteins from the complementing plasmid. In Fig. 1C , comparing lysates of log-phase promastigotes of wild-type L. major Friedlin (FVI) with the three complemented lines (+HASPB, +SHERP and +cDNA16) described in McKean et al . (2001) after 3 days in culture, it is evident that both HASPB and SHERP are overexpressed in the single add-back lines. Similarly, the open reading frames of the LmcDNA16 locus (HASPB, HASPA and SHERP) are all overexpressed following complementation with the complete locus as compared with the wild-type parasites, in which only low levels of SHERP and the HASPs are detectable at this stage of the parasite life cycle ( Fig. 1C ). Before proceeding to sand fly transmission experiments therefore we generated a new complemented line expressing a single copy of the LmcDNA16 locus, coupled with a constitutively expressed NEO gene reintroduced into one allele of the original locus, thereby generating a heterozygous add-back parasite line. Correct genomic integration of the LmcDNA16 complementation cassette ( Fig. 2A ) was confirmed by Southern blot analysis ( Fig. 2B ). In the clones analysed here, the NEO probe hybridizes to a single fragment of 4.8 kb in the complemented double replacement clone (Kin) and this is absent from wild-type (FVI) and null (KO) parasites. The PAC gene is found on a single fragment of 3.8 kb in the null clone but is absent as expected in wild-type DNA. The PAC gene is also absent from the Kin clone, demonstrating that the PAC cassette has been replaced with NEO in these cells. As expected, neither HASPB nor SHERP is detected in the null clone. The HASP probe hybridizes in wild-type cells to fragments of 7.6 kb, 4.3 kb and 2.2 kb, corresponding to the expected sizes of HASPA2 , HASPA1 and HASPB respectively. The 4.3 kb and 2.2 kb bands ( HASPA1 and HASPB ) are also seen in the Kin clone, together with a band of 4.8 kb representing HASPA2 plus the inserted NEO gene. The SHERP probe is detected as two bands of 1.8 kb and 1.6 kb (corresponding to SHERP1 and SHERP2 respectively ) in both the wild-type and Kin clone. These data verify the genetic structure of the inserted add back (which has subsequently been confirmed by DNA sequencing and comparison to GenBank entry AJ237587.1).
Immunoblotting of lysates taken from the Kin line and wild-type parasites collected over a 7-day time-course was used to verify regulated expression of the HASPB and SHERP proteins in the complemented Kin parasites. As shown in the clones analysed in Fig. 2C , a similar temporal pattern of expression of HASPB and SHERP is detected in both wild-type and add-back cells, although the relative levels of expression are lower in the Kin parasites, as would be expected in a heterozygous add-back line.
To correlate these observations with expression specifically in metacyclic stages, parasites of the Kin and null lines, together with wild-type cells, were grown to 7 days in culture and then agglutinated with peanut lectin. Non-agglutinated parasites (defined as metacyclics in culture and in infected sandflies using the metacyclic LPG-specific antibody, 3F12; Sacks and da Silva, 1987 ; Saraiva et al ., 1995 ) were then analysed by indirect immunofluorescence using confocal microscopy. As shown in Fig. 2D , non-agglutinated parasites from all three lines have the typical metacyclic morphology of short body and relatively long flagellum. As a consequence, the nucleus and kinetoplast are located close together, as detected by DAPI staining of their DNA. All three lines are also cross-reactive with 3F12, which produces a punctuate staining pattern under the fixation conditions used with confocal microscopy. Only the wild-type and Kin parasites express HASPB, as expected. These observations using a defined metacyclic LPG marker confirm that the null parasites can undergo metacyclogenesis as reported previously ( McKean et al ., 2001 ).
To further characterize and compare the phenotype of the new Kin line with the null and wild-type parasites (independently from detection of either HASPB or metacyclic LPG), total glycoconjugate production was monitored in parasite lysates and secretory fractions of late-log-phase parasites (day 5 post-inoculation). The antibody used for detection, LT6 ( Ilg et al ., 1993 ), recognizes the galactose-mannose-phosphate disaccharide repeat units of LPG and secreted proteophosphoglycan (PPG; Rogers et al ., 2004 ). As shown in Fig. 2E , LPG is detected in total lysates from wild-type, KO and Kin parasites while secreted PPG is also abundant in the culture supernatant from all three lines. While this analysis would not detect structural differences between the different glycoconjugate fractions, these data demonstrate that PPG secretion is not compromised following deletion of the LmcDNA16 locus, at least in vitro . In vivo , PPG is secreted by short (leptomonad) promastigotes in Leishmania mexicana ( Rogers et al ., 2002 ), some of which differentiate into metacyclic parasites primed for transmission ( Bates, 2007 ). Our observation of PPG secretion by the KO L. major parasites described above suggests that in vitro differentiation can generate elongated nectomonads and short promastigotes in this line.
Development of LmcDNA16 mutant parasites in P. papatasi
To investigate a potential role for the HASP and SHERP genes in vector transmission, we used wild-type promastigotes of L. major Friedlin and the three mutant lines described in McKean et al . (2001) : Lmc DNA16 double-knockout or null (KO; ΔcDNA16::HYG/ΔcDNA16::PAC ), Lmc DNA16 double-knockout complemented with episomal HASPB (+HASPB; ΔcDNA16::HYG/ΔcDNA16::PAC [pTEX NEO HASPB] , as used in Fig. 1C ) and Lmc DNA16 double-knockout complemented with episomal SHERP (+SHERP; ΔcDNA16::HYG/ΔcDNA16::PAC [pTEX NEO SHERP] , as used in Fig. 1C ). In addition, the new add-back line described above, ΔcDNA16::HYG/ΔcDNA16::PAC/ΔPAC::cDNA16 (Kin), was used instead of the original episomally complemented parasites analysed in Fig. 1C (+cDNA16; McKean et al ., 2001 ). Female sand flies were infected by feeding through a chick-skin membrane on heat-inactivated rabbit blood containing promastigotes of each parasite line. Engorged sand flies were dissected over a time-course post-blood meal (PBM) and the location and rates of infection in the digestive tract determined by dissection and light microscopy. Parasite loads were also confirmed by quantitative PCR (qPCR) analysis of dissected guts.
Dissections of these sand fly females showed that all L. major lines were able to develop heavy infections in P. papatasi . However, quantitative differences in parasite load between the lines occurred after defecation on day 5 PBM and on day 12 ( Fig. 3 ). Results of qPCR analysis revealed that differences between the wild-type strain (FVI), double-knockout (KO) and complemented (Kin) lines in parasite loads were not significant in late-stage infections ( Fig. 4A ). However, infections with +HASPB were significantly less abundant than wild-type infections ( P = 0.0005, U = 197.0, Z = 3.470) while +SHERP infections were even less abundant than +HASPB infections ( P = 0.0339, U = 306.5, Z = 2.122; Fig. 4B ).
After escape from the peritrophic matrix, L. major development within the midgut of P. papatasi involves transformation of elongated nectomonads to short promastigotes (leptomonads in the terminology of Rogers et al ., 2002 ) and highly motile metacyclic forms. Parasites migrate to the thoracic part of the midgut and colonize the SV; forms that attach to the SV are called haptomonads. In the experiments described in Table 1 , significant differences were observed in the location of infections within the infected sand flies. Wild-type (FVI), complemented (Kin) and +HASPB lines showed heavy colonization of the SV with many haptomonads firmly adhering to its chitinous lining. Conversely, null parasites (KO) did not colonize the SV (only one female from 97 dissected by days 9 and 12 PBM had weak colonization of the SV), although the anterior part of the thoracic midgut near the SV, named the cardia, was reached by elongated nectomonads quite early after infective feeding (57% of infected flies by day 5 PBM). The +SHERP parasites colonized the SV in a small percentage of infected sand flies; only nine females from 82 dissected by 9 and 12 days PBM had short promastigotes weakly attached to the SV. Similarly to the null line, the parasite population of the +SHERP line consisted mostly of elongated nectomonads.
Differences between the five L. major lines in the presence and size of the four morphological categories (elongated nectomonads, short promastigotes, metacyclic promastigotes and round forms) were significant at all time intervals tested ( Fig. 5 , Table 2 ). However, the following similarities between lines were observed: (i) the FVI and +HASPB lines had shorter bodies and higher proportions of short and metacyclic promastigotes then the other three lines at 9 and 12 days PBM, (ii) the FVI and Kin lines had shorter bodies and lower numbers of elongated nectomonads than the other three lines at 5 days PBM, and (iii) the KO and +SHERP lines had longer bodies and higher proportions of elongated nectomonads than the other lines at 9 and 12 days PBM.
Although all five L. major lines were able to develop heavy late-stage infections in P. papatasi , the observed differences in both the locations of the infections in the fly and the parasite morphology allows the division of the tested lines into two groups. The first of these is composed of the wild-type FVI strain and the Kin and +HASPB lines, which all showed classical Leishmania development with transformation of elongated nectomonads to short promastigotes and metacyclic promastigotes and colonization of the SV in late-stage infections. Parasites of the second group, consisting of null and +SHERP mutants, remained mostly in the stage of elongated nectomonads and colonized the abdominal and thoracic midgut but not the SV.
To correlate the identity of parasite stages defined by morphology in the sand fly with expression of stage-specific markers, antibodies specific for HASPB, SHERP and metacyclic LPG were used for indirect immunofluorescence microscopy. The stage specificity of these reagents in the sand fly is illustrated in Fig. 6 which shows their reactions with wild-type L. major dissected from infected flies in late-stage infections. Staining with the 3F12 antibody specific for metacyclic LPG ( Sacks and da Silva, 1987 ) identified metacyclic promastigotes only, with no recognition of short promastigotes or elongated nectomonads ( Fig. 6 , images a–d). Similarly, anti-HASPB also showed specificity for metacyclic parasites and no cross-reactivity with the previous developmental stages ( Fig. 6 , images e–h). Conversely, while anti-SHERP also identified metacyclics, it also cross-reacted with short promastigotes but not with elongated nectomonads ( Fig. 6 , images i–l). With this knowledge, the same antibodies were used to monitor how the frequency of the various morphological forms differed between smears taken from individual flies ( Fig. 7 ).
At 5 days PBM, no positive reaction was detected with either the HASPB or SHERP antibodies (results not shown). In all Leishmania lines at later stages of infection (9 or 12 days PBM), 3F12 and anti-HASPB reacted only with parasite stages identified morphologically as metacyclics ( Fig. 7A and B ), as described above for the wild-type strain. Conversely, SHERP antibodies reacted not only with metacyclics but also with some short promastigotes of the wild-type strain and Kin line ( Fig. 7C and D ). These positively staining parasites may represent some intermediate forms with flagella not long enough to allow identification as metacyclic cells. Generally, the highest percentage of promastigotes morphologically identified as metacyclics was found in the wild-type FVI strain (23–36%) and the lowest in the null line (0–3%). The metacyclic LPG antibody, 3F12, positively identified metacyclic parasites in four of the five lines tested at 9 or 12 days PBM but no metacyclics were found in the null parasite infections. Anti-HASPB reacted positively with metacyclic parasites of three lines (FVI, +HASPB and Kin) but not with the null and +SHERP lines, as expected. Negative reaction of anti-SHERP antibodies was observed with KO and +HASPB lines as expected but, surprisingly, also with parasites of the +SHERP line, both as metacyclics and as short promastigotes. These results suggest not only that SHERP function may be required at an earlier stage than HASPB function but also that HASPB expression may be required for SHERP activity in vivo . | Discussion
In this article, we reveal a novel role for two previously described Leishmania proteins that show stage-specific expression, at both the RNA and protein level, during the parasite life cycle ( Flinn and Smith, 1992 ; Flinn et al ., 1994 ; Knuepfer et al ., 2001 ). As originally demonstrated under in vitro culture conditions, and here confirmed during development in P. papatasi , the HASPB and SHERP proteins are highly upregulated in infective metacyclic stages, with SHERP also showing low-level expression in the preceding short promastigote stage in the vector. Neither protein is detectable at earlier developmental stages in the sand fly, definitively confirming the stage-specific expression of these molecules. Low-level HASPB and SHERP signal can be observed in logarithmic phase culture due to a lack of synchrony in parasite growth and differentiation and the presence of a mixture of parasite stages in the starting inoculum. These practical issues are partially resolved by using low-passage parasites that exhibit higher expression levels of both proteins (see Fig. 1A ). Under these conditions, SHERP expression peaks before maximum HASPB expression, likely correlating with its detection in short promastigotes although these cannot be definitively identified in culture.
The observations above confirm the utility of HASPB as a marker for metacyclogenesis in L. major and suggest a vital role for both HASPB and SHERP either in this differentiation process and/or in the metacyclic parasite. HASPB in L. major is transported to the plasma membrane, a process requiring N-terminal modification byacylation ( Denny et al ., 2000 ). At this location, it can be detected externally on both the cell body and flagellum, most recently by imaging in live parasites (L. Maclean, unpublished). Our current model proposes that HASPB can be shed on macrophage entry, a process that could optimize presentation to the host immune system. Recombinant Leishmania donovani HASPB is a target for recognition by B-1 B cell-derived natural antibodies, with the resulting immune complexes triggering classical complement pathway activation, leading to IL-4 secretion, CD8+ T cell priming and vaccination against parasite infection in immuno-compromised mice ( Stager et al ., 2003 ). Whether HASPB is shed from metacyclics in the vector has not yet been established.
SHERP, in contrast to HASPB, is a peripheral membrane protein that is not expressed by parasites in the host. In wild-type promastigotes, SHERP is localized to the cytosolic face of the ER and the outer mitochondrial membrane ( Knuepfer et al ., 2001 ), with recent biochemical and structural analysis suggesting that membrane lipid interactions may drive the function of this unusual small protein (B. Moore, unpublished). Another potential focus for SHERP interactions is with the vacuolar ATPase protein complex, a membrane-localized pump that drives proton transport across the eukaryotic plasma membrane but also functions in acidification of subcellular compartments, including those within the endosomal/lysosomal system ( Sun-Wada et al ., 2004 ). This may be of particular significance in metacyclic parasites given the recent identification of autophagy as a key process in parasite differentiation and virulence ( Besteiro et al ., 2006 ; 2007 ;). In cultured L. major , dividing promastigotes are characterized by a multivesicular body-like network that matures into a lysosomal-like structure of high lytic capacity and low pH in metacyclic parasites. Given SHERP's intracellular localization to ER and mitochondrial membranes in L. major ( Knuepfer et al ., 2001 ) and the potential for both membrane types to be processed by autophagic digestion, perhaps SHERP plays a regulatory role in vacuolar acidification during autophagy in the vector.
Genetic analysis of parasite mutants has been critical for the analysis and confirmation of gene function in Leishmania species (e.g. Spath et al ., 2003 ; Ortiz et al ., 2007 ). However, parasites deleted for the LmcDNA16 locus containing HASP and SHERP genes failed to present a phenotype distinct from wild-type parasites when used directly to infect either cultured macrophages or susceptible BALB/c mice in vivo , while ‘add-back’ clones yielded phenotypic features consistent with protein overexpression ( McKean et al ., 2001 ). Of particular note, the null parasites were able to undergo metacyclogenesis in vitro , as monitored by analysis of stage-specific transcript patterns and resistance to agglutination with peanut lectin. These observations suggested that the products of the HASP/SHERP locus did not play a role in parasite differentiation ( McKean et al ., 2001 ). Importantly, the mouse infection experiments were initiated by high-dose needle injection of metacyclic-rich parasite populations rather than by experimental inoculation by sand fly bite, an approach that mimics parasite transmission in vivo .
Results presented in this article confirm that the LmcDNA16 null parasites described above can differentiate into metacyclics in vitro , as defined by expression of the 3F12 epitope on metacyclic LPG. We have now developed a new complemented Kin line expressing HASPB and SHERP proteins at similar levels to wild-type parasites and, importantly, at the correct stage of development. These parasites, together with the null line, have been used to infect female P. papatasi and parasite growth and differentiation measured following blood meal digestion, using both morphological and biochemical markers. The results of these analyses demonstrate conclusively that the genetic locus encoding HASPB and SHERP is essential for metacyclogenesis in the sand fly: null parasites accumulate at the earlier elongated nectomonad stage of development and do not colonize the SV. As a result, we predict that the null mutants cannot be transmitted to the host by sand fly bite since destruction of the vector SV has been described to facilitate the transmission of Leishmania and Trypanosoma parasites ( Volf et al ., 2004 ).
These observations, demonstrating that the LmcDNA16 null parasites can undergo metacyclogenesis in vitro but not in vivo , suggest either that the metacyclic phenotype observed in the vector and required for parasite transmission in vivo is not fully replicated in vitro or that other factors are critical for differentiation in the sand fly. Perhaps the loss of proteins expressed from this locus impacts on parasite adhesion, migration or sensitivity to midgut hydrolases or, alternatively, plays a role in establishment at the SV. Is the null phenotype observed in sand flies caused by the lack of HASPB and/or SHERP function? And what is the role of the HASPA genes ( McKean et al ., 1997b ) that are also present in the deleted LmcDNA16 locus and code for HASP proteins lacking the central repetitive domain of HASPB?
Using the single gene complemented lines (+HASPB and +SHERP), we have shown that HASPB alone can complement the null phenotype observed with the KO parasites, with these parasites able to complete classical development and colonize the SV in late-stage infections. However, it should be noted that +HASPB parasites overexpress the protein constitutively, a property shared by the +SHERP line with respect to SHERP expression. We therefore conclude that while HASPB is likely to be the dominant molecule in restoring the null phenotype, modulation of HASPB function by other factors in metacyclic parasites is not ruled out by these observations. Clearly, SHERP expression precedes HASPB expression and it is possible that the functions of these two proteins are linked during metacyclogenesis. An additional role for the HASPA proteins in this process cannot be discounted.
Delineating the function of all genes within the LmcDNA16 locus and their relative roles in parasite transmission will require an extended study utilizing a range of new transgenic lines expressing each gene product appropriately, both quantitatively and temporally. These approaches may reveal mechanistic details of metacyclogenesis within the sand fly and, as a consequence, underpin our understanding of transmission of the Leishmania parasite in vivo . | Re-use of this article is permitted in accordance with the Terms and Conditions set out at http://wileyonlinelibrary.com/onlineopen#OnlineOpen_Terms
The stage-regulated HASPB and SHERP proteins of Leishmania major are predominantly expressed in cultured metacyclic parasites that are competent for macrophage uptake and survival. The role of these proteins in parasite development in the sand fly vector has not been explored, however. Here, we confirm that expression of HASPB is detected only in vector metacyclic stages, correlating with the expression of metacyclic-specific lipophosphoglycan and providing the first definitive protein marker for this infective sand fly stage. Similarly, SHERP is expressed in vector metacyclics but is also detected at low levels in the preceding short promastigote stage. Using genetically modified parasites lacking or complemented for the LmcDNA16 locus on chromosome 23 that contains the HASP and SHERP genes, we further show that the presence of this locus is essential for parasite differentiation to the metacyclic stage in Phlebotomus papatasi . While wild-type and complemented parasites transform normally in late-stage infections, generating metacyclic promastigotes and colonizing the sand fly stomodeal valve, null parasites accumulate at the earlier elongated nectomonad stage of development within the abdominal and thoracic midgut of the sand fly. Complementation with HASPB or SHERP alone suggests that HASPB is the dominant effector molecule in this process. | Experimental procedures
Parasites
Leishmania major Friedlin (MHOM/IL/81/Friedlin/VI; FVI) and mutant parasites targeted for gene disruption/overexpression at the LmcDNA16 locus were maintained on blood agar slopes or in culture as previously described ( McKean et al ., 2001 ). Low-passage promastigotes (2–3 cycles post transformation from mouse-derived amastigotes) were inoculated at 1 × 10 5 ml −1 and the culture sampled daily. Parasite numbers, morphology and motility were monitored by light microscopy, counting > 100 parasites per line per time point and scoring for metacyclics (small body, flagellum ∼2× body length, high motility), pre-metacyclics (heterogeneous population of parasites showing shortening body and moderate motility) and procyclics (extended cell body ∼2× that of metacyclic parasites with relatively short flagellum).
The mutant lines used were the 4.8 Lmc DNA16 double-knockout (KO; ΔcDNA16::HYG/ΔcDNA16::PAC ), the 4.8 Lmc DNA16 double-knockout complemented with episomal HASPB (+HASP; ΔcDNA16::HYG/ΔcDNA16::PAC [pTEX NEO HASPB] ) and the 4.8 Lmc DNA16 double-knockout complemented with episomal SHERP (+SHERP; ΔcDNA16::HYG/ΔcDNA16::PAC [pTEX NEO SHERP]; McKean et al ., 2001 ).
A new complemented (knock-in, Kin) LmcDNA16 line was generated as described in Fig. 2 , by introduction of a linear fragment containing the complete LmcDNA16 locus plus constitutively expressed NEO gene into its original location on chromosome 23 in a heterozygous (LmcDNA16 locus single deletion) background, using homologous recombination as described in McKean et al . (2001) . Correct genomic integration was confirmed by DNA blotting, as described. Following rapid passage through susceptible BALB/c mice (as described in Depledge et al ., 2009 ), these parasites were also maintained in culture as described above.
Immunoblotting
Whole parasite lysates were separated by SDS-PAGE as described ( McKean et al ., 2001 ) and blots probed with polyclonal antisera against HASPB ( Flinn et al ., 1994 ) or SHERP ( Knuepfer et al ., 2001 ). A monoclonal antibody recognizing the constitutively expressed protein EF1α (clone CBP-KK1, Millipore) or the polyclonal L. major anti- N -myristoyl transferase ( Price et al ., 2003 ) were used to control for equivalent protein loading.
For PPG analysis, low-passage promastigotes were grown to late-log-phase and parasite samples collected for lysis and SDS-PAGE as described above. Culture supernatant was fractionated by sequential centrifugation at 2200 g (10 min at 3200 rpm) and 100 000 g (60 min at 30 000 rpm) for 10 min. The final supernatant was discarded and the pellet lysed in SDS-PAGE gel loading mix prior to SDS-PAGE analysis using a 3 cm, 4% stacking gel and a 12% resolving gel. Blots were probed with anti-LT6 (1:500; the kind gift of Paul Bates, Lancaster University) with detection by ECLplus (Amersham).
Sand flies and sand fly infections
The colony of P. papatasi was maintained at 26°C on 50% sucrose and 14 h light/10 h dark photoperiod as described previously ( Benkova and Volf, 2007 ). Sand fly females were infected by feeding through a chick-skin membrane on heat-inactivated rabbit blood containing 10 6 promastigotes ml −1 . Engorged sand flies were maintained in the same conditions as the colony and dissected 2, 5, 9 and 12 days PBM. The location of Leishmania infections in the sand fly digestive tract (foregut, SV, thoracic and abdominal midgut, and endoperitrophic and ectoperitrophic space) was determined by dissection and examination by light microscopy. Parasite loads were estimated by two methods: infections seen in the gut in situ were graded according to Myskova et al . (2008 ) as light (< 100 parasites per gut), moderate (100–1000 parasites per gut) and heavy (> 1000 parasites per gut). Alternatively, 30–40 guts from females with late infections (10–12 days PBM) were individually dissected into NET 50 and stored in −20°C for qPCR. Sand fly infection experiments were repeated four times for combinations of wild-type (FVI), KO and Kin lines and twice for combinations of FVI, +HASPB and +SHERP lines.
Quantitative PCR
Extraction of total DNA from dissected sand fly guts was performed using a DNA tissue isolation kit (Roche Diagnostics, Indianapolis, IN) according to the manufacturer's instructions and DNA was eluted in 100 μl of EB buffer. qPCR for detection and quantification of Leishmania parasites was performed in Bio-Rad iCycler & iQ Real-Time PCR Systems by using the SYBR Green detection method (iQ SYBR Green Supermix, Bio-Rad, Hercules, CA). The kinetoplast primers described by Mary et al . (2004) (forward primer 5′-CTTTTCTGGTCCTCCGGGTAGG-3′ and reverse primer 5′-CCACCCGGCCCTATTTTACACCAA-3′) were used (for more details see Myskova et al ., 2008 ). A series of 10-fold dilutions of Leishmania promastigote DNA, ranging from 10 4 to 10 −2 parasites per reaction, was used to mix with DNA from sand fly females. DNA from uninfected sand flies was used as a negative control.
Morphometry of parasites
Gut smears of L. major -infected females 5, 9 and 12 days PBM were fixed with methanol, stained with Giemsa and examined under the light microscope with an oil-immersion objective. One hundred and sixty randomly selected promastigotes from four sand flies/smears were measured in each combination of Leishmania line and time PBM. Body length, flagellar length and body width of parasites were measured and position of the kinetoplast in relation to the nucleus was examined. Four morphological forms were distinguished, based on the criteria of Walters (1993) and Cihakova and Volf (1997) : (i) short promastigotes: body length < 14 μm and flagellar length < 2 times body length; (ii) elongated nectomonads: body length ≥ 14 μm; (iii) metacyclic promastigotes: body length < 14 μm and flagellar length ≥ 2 times body length, and (iv) round forms: body width > 4 μm and body length ≤ 7.5 μm include also paramastigotes with kinetoplast lateral to the nucleus. We use here the term short promastigotes derived from the terminology of Walters (1993) (short nectomonad promastigotes) which is the older synonym of leptomonads (leptomonad promastigotes) proposed by Rogers et al . (2002) . Haptomonads cannot be distinguished in this study as they remain attached to the gut and cannot be measured on gut smears.
Indirect immunofluorescence
Metacyclic LPG was detected using 3F12 monoclonal antibody from mouse ascites fluid ( Sacks and da Silva, 1987 ). HASPB and SHERP proteins were detected using rabbit polyclonal anti-HASPB and anti-SHERP, both affinity-purified against recombinant protein ( Flinn et al ., 1994 ; Knuepfer et al ., 2001 ).
For cultured parasites, following agglutination with 100 μg ml −1 peanut lectin, non-agglutinated cells were fixed in 4% paraformaldehyde and permeabilized in 0.2% Triton X-100 prior to preparation of slides as previously described ( Denny et al ., 2002 ). Primary antibodies (undiluted 3F12 ascites; anti-HASPB at 1:200 in 1% BSA in PBS) were detected using Alexa Fluor 488- (goat anti-rabbit) or 594- (goat anti-mouse) conjugated secondary antibodies (Invitrogen). Samples were visualized by confocal microscopy using a Zeiss LSM 510 meta with a Plan-Apochromat 63×/1.4 Oil differential interference contrast objective lens and images acquired using LSM 510 version 3.2 software (Carl Zeiss, Jena, Germany).
Gut smears taken from infected sand flies were air-dried on glass slides and fixed with methanol. Non-specific binding was blocked with 1% BSA in PBS (Phosphate Buffered Saline, pH 7.4) for 20 min and the slides then washed and incubated for 30 min with antibodies. 3F12 ascites was used undiluted while anti-HASPB and anti-SHERP antibodies were diluted 1:200 in 1% BSA in PBS. After washing, slides were incubated for 1 h with goat anti-mouse polyvalent FITC-conjugated IgG (Sigma) in dilution 1:250 in 1% BSA in PBS (3F12 antibody assay) or with goat anti-rabbit FITC-conjugated IgG (Sigma) in dilution 1:160 in 1% BSA in PBS (anti-HASPB and anti-SHERP antibodies assays). Slides were then re-washed and after mounting in Vectashield with propidium iodide (Vecta Laboratories) examined under oil-immersion objective in Olympus BX51 fluorescent microscope. For each combination of Leishmania line, antibody and time PBM, 100 promastigotes were photographed with an Olympus camera, the images measured with Image J software and classified. These 100 parasites came from at least two different gut smears taken from different sand flies.
Statistical analysis
Measurements of parasites and the representation of morphological forms were compared using analysis of variance ( post hoc test) and chi-square test respectively. Results of qPCR were tested with non-parametric Mann–Whitney U -test and Kruskal–Wallis test ( anova ). All the statistical evaluations were performed with statistical software SPSS version16 and Statistica version 6.0. | We thank Paul Bates for kind provision of the LT6 antibody (originally generated by Thomas Ilg), Desikan Rangarajan who conducted the initial steps in the add-back cloning protocol, Johannes Doehl (BBSRC graduate student) who confirmed the full sequence of the Kin construct and Lorna Maclean for helpful discussion. This work was supported by the Wellcome Trust (Programme Grant 077503 to D.F.S.) and the Czech Ministry of Education (projects MSM0021620828 and LC06009). | CC BY | no | 2022-01-12 15:46:43 | Cell Microbiol. 2010 Dec; 12(12):1765-1779 | oa_package/b2/7f/PMC3015063.tar.gz |
||
PMC3015064 | 20846362 | Introduction
Among the non- Candida albicans Candida species ( Kaur et al. , 2005 ; Marcet-Houben & Gabaldón, 2009 ;), Candida glabrata is the species most closely related to Saccharomyces cerevisiae ( Fig. 1a ). In this review, we will discuss some of the properties that allow C. glabrata to exist as a human commensal. We will briefly discuss the infectivity, genetic changes, drug and stress resistance as well as the adherence of C. glabrata . A selection of traits supporting commensalism and/or pathogenicity was proposed earlier on for fungal and bacterial pathogens. This might occur in the environment ( Bliska & Casadevall, 2009 ), in confrontation with microbial communities or immune responses ( Seider et al. , 2010 ). A striking adaptation of C. glabrata might be its rapid growth and short generation time, which is tuned to 37 °C ( Fig. 1b ). Candida glabrata rarely penetrates tissue and is therefore exposed to microbial competitors on mucosal surfaces. Additionally, C. glabrata does not undergo a sexual cycle allowing the generation of resistant spores. We believe that these restrictions, combined with the host environment, favored the development of several specific traits, for example its high resistance to starvation. Compared with S. cerevisiae, C. glabrata shows only a small number of genetic adaptations. Considering its phylogenetic position ( Fig. 1a ), it can be assumed that C. glabrata 's ability to infect humans emerged independently from that of other Candida species.
After C. albicans, C. glabrata is the most common cause of vaginal and oral candidiasis. Candida glabrata is the second most important species in terms of candidiasis in the United States ( Pfaller & Diekema, 2004 ). In total, C. glabrata accounts for roughly 15–20% of all Candida infections, with this relative incidence increasing every year. In healthy individuals, C. glabrata is restricted because of the action of the innate immune system and the microbial communities, which counteracts dissemination by competition for nutrients and secretion of toxins ( Wargo & Hogan, 2006 ; Pamer, 2007 ;). Immunocompromised persons suffering from neutropenia (e.g. cancer or transplant patients) or persons treated with antibiotics are much more susceptible to C. glabrata infections ( Perlroth et al. , 2007 ; Caston-Osorio et al. , 2008 ; Cohen et al. , 2010 ;). Candida glabrata is more often found in elderly patients ( Hof, 2010 ). In murine infection models, C. glabrata infection is cleared and mice need to be rendered neutropenic to make them susceptible to pathogenic yeast infections ( MacCallum et al. , 2006 ). Together, this might indicate that C. glabrata is less virulent than C. albicans and mainly causes disease in severely ill patients. | Conclusion
Candida glabrata is an efficient human pathogen, which can infect immunocompromised and elderly persons. Its avid adherence to mammalian tissue and to other surfaces is based on the expression of a number of specific adhesins. Candida glabrata rarely penetrates tissue actively and therefore directly competes with other microbial agents on mucosal surfaces. This might have led to the selection of specific drug and oxidative stress-resistance traits. Somatic fungal cells have high resistance to starvation, a trait that might also support survival during engulfment in phagocytic cells. Prolonged survival of C. glabrata inside phagocytic cells might support the establishment of disseminated infection and thus directly relate to its success as a commensal and pathogen ( Fig. 3 ). Robust adherence, active suppression of fungicidal drugs, stress resistance and an enhanced ability to sustain prolonged starvation render C. glabrata superior to S. cerevisiae in causing disease. Altogether, we suggest that its adaptation to the mammalian host is rather due to slight genomic fine-tuning than due to radical large-scale changes. | The opportunistic human fungal pathogen Candida glabrata is closely related to Saccharomyces cerevisiae , yet it has evolved to survive within mammalian hosts. Which traits help C. glabrata to adapt to this different environment? Which specific responses are crucial for its survival in the host? The main differences seem to include an extended repertoire of adhesin genes, high drug resistance, an enhanced ability to sustain prolonged starvation and adaptations of the transcriptional wiring of key stress response genes. Here, we discuss the properties of C. glabrata with a focus on the differences to related fungi. | Be creative: genetic changes
Candida glabrata shares a recent common ancestor with several Saccharomyces species, and clearly belongs to a clade different from that of other Candida species which display particular features such as the recoding of the CUG codon to Serine ( Fig. 1a ). As a result of this evolutionary relatedness, most S. cerevisiae genes have orthologues in C. glabrata and the chromosomal structure in terms of gene order is largely conserved between the two species ( Dujon et al. , 2004 ; Marcet-Houben & Gabaldón, 2009 ;). Among many others, shared genomic features include a similar stress response ( Roetzer et al. , 2008 ) and a respiratory metabolism characterized by the lack of Complex I ( Marcet-Houben & Gabaldón, 2009 ). Despite this general resemblance, several differences in terms of gene content can be found, which might play an important role with respect to the phenotypic differences. Compared with S. cerevisiae, C. glabrata seems to have undergone an increased rate of gene loss from their common ancestor ( Dujon et al. , 2004 ). Notable absences include genes of the galactose, phosphate, nitrogen and sulfur metabolism. Similar to many other microbial pathogens ( Ehrlich et al. , 2008 ), such a reductive evolution could be related to C. glabrata 's adaptation as mammalian commensal and opportunistic pathogen. Indeed, C. glabrata relies on its host to overcome important auxotrophies such as those of nicotinic acid, pyridoxine and thiamine ( Kaur et al. , 2005 ). Conversely, C. glabrata encodes genes that are absent from other Saccharomyces , including a putative racemase of bacterial origin ( Marcet-Houben & Gabaldón, 2010 ). Other differences entail gene expansions that are specific to the C. glabrata lineage. Three of these specific expansions affect cell-wall organization and are probably related to adhesion properties of C. glabrata cells. These expansions include six copies of extracellular glycosylphosphatidylinositol-linked aspartyl proteases, eight copies of a α-1,3-mannosyltransferase involved in cell-wall biogenesis and a variable number of glycosylphosphatidylinositol-linked epithelial adhesin ( EPA ) genes located in subtelomeric regions. Importantly, this last family, which is required for epithelial cell adhesion and biofilm formation, seems to be a unique adaptation of C. glabrata because no significant homology can be detected in S. cerevisiae . Nevertheless, FLO genes in S. cerevisiae coding for glycosylphosphatidylinositol-linked flocculins are usually considered to be functional homologues of C. glabrata EPA genes ( De Las Penas et al. , 2003 ). Similarly to FLO genes, EPA genes are located in subtelomeric regions and are subject to SIR-mediated silencing ( Kaur et al. , 2007 ). Moreover, FLO genes have been found to be involved in cell adhesion and biofilm formation in S. cerevisiae ( Reynolds & Fink, 2001 ). An intriguing possibility is that FLO and EPA genes are evolutionarily related but their sequences have diverged beyond recognition. Such an accelerated divergence seems to be quite frequent in yeast genes ( Wolfe, 2004 ) and may have been favored by the extensive lineage-specific duplications undergone in the two lineages and by the acquisition of different tandemly repeated motifs.
Interestingly, despite being composed of different amino acids, the length of the tandemly repeated motifs in several FLO and EPA genes is similar (45 amino acids), perhaps reflecting a common selective pressure ( Thierry et al. , 2010 ). Finally, a recently discovered feature of the C. glabrata genome, one that is almost unique to this species, is megasatellites ( Thierry et al. , 2010 ). These long (135–417 nucleotides) repetitive sequences are enriched within genes coding for cell-wall proteins, including some EPA genes, and might also be related to the high degree of genomic plasticity observed in C. glabrata ( Polakova et al. , 2009 ). This genomic plasticity is likely to play an important role in the evolution of C. glabrata populations because a sexual cycle has not yet been observed. Remarkably, this apparent disappearance of the sexual cycle seems to be a common trait for most fungal pathogens ( Butler, 2010 ). Thus, differences in the genome sequence and its regulation provide important hints about the way in which two closely related yeasts have adapted to radically different environments.
Be sticky: adherence and persistence
Compared with S. cerevisiae, C. glabrata has lost some of the genes needed for several metabolic pathways, for example galactose utilization. However, it has also gained certain functions necessary for its commensal lifestyle. Unlike S. cerevisiae, C. glabrata is able to adhere not only to mammalian cells but also to other surfaces ( Cormack et al. , 1999 ). An elegant screen for factors required for biofilm formation identified CgRIF1, CgSIR4 as well as the protein kinase CgYAK1 and CgEPA6 encoding an adhesin ( Iraqui et al. , 2005 ). CgEpa6 is a member of the EPA gene family of glycosylphosphatidylinositol-anchored cell-wall proteins ( Castano et al. , 2005 ; Domergue et al. , 2005 ; Iraqui et al. , 2005 ; Kaur et al. , 2005 ;). Similar to the S. cerevisiae FLO lectin-like genes, the EPA genes are encoded in subtelomeric clusters and are subject to transcriptional silencing ( Castano et al. , 2005 ). Candida glabrata cells with mutations in the telomeric silencing factors Sir2, Sir3 and Sir4 are more adherent to cultured epithelial cells and show better colonization efficiency in kidneys ( Castano et al. , 2005 ; Domergue et al. , 2005 ;). Interestingly, the lack of nicotinic acid, a precursor of NAD + , triggers adhesin ( EPA6 ) expression ( Domergue et al. , 2005 ). Candida glabrata is an auxotroph for NAD + and utilizes nicotinamide riboside and nicotinic acid as NAD(+) sources during disseminated infection ( Ma et al. , 2007 ). The majority of all genes transcriptionally induced by niacin limitation are also regulated by the NAD(+)-dependent histone deacetylase Hst1 ( Ma et al. , 2009 ). Therefore, nicotinic acid and nicotinamide riboside are important host-derived regulatory signals. Transcript profiling of C. glabrata cells internalized by macrophages revealed increased expression of a gene family encoding extracellular glycosylphosphatidylinositol-linked aspartyl proteases ( YPS genes) ( Kaur et al. , 2007 ). YPS proteases are responsible for removing and releasing glycosylphosphatidylinositol-anchored cell-wall proteins and are thought to be necessary for cell-wall integrity and regulated adherence to mammalian cells. A recent survey identified several putative adhesin genes that had not yet been annotated ( de Groot et al. , 2008 ). Their role for C. glabrata virulence is currently unknown. It should also be noted that C. glabrata strains have different number and sequence of adhesin genes. Thus, C. glabrata uses signals from the host for growth and adherence and its adhesive repertoire is more variable and by far transcends that of S. cerevisiae .
Be adaptive: environmental stress response (ESR)
Vegetative cells of C. glabrata have high intrinsic stress tolerance. While C. albicans can survive on surfaces up to 4 months, C. glabrata can do so for even up to 5 months ( Kramer et al. , 2006 ). However, the persistence traits have not yet been explored systemically. Stress resistance might be a consequence of its natural habitat. Candida glabrata preferably grows on mucosal surfaces as biofilm. Mammalian mucosal areas cause nutrient shortage, osmotic and other stresses due to the presence of other microorganisms and protective mechanisms of the host ( Garside et al. , 2004 ; Pamer, 2007 ;). Furthermore, sexual reproduction as a way to generate stress-resistant spores has never been observed in C. glabrata (reviewed in Butler, 2010 ). So far, all clinical isolates of C. glabrata have been haploid, although it has the components of the mating machinery at its disposal ( Wong et al. , 2003 ). Candida glabrata might have lost the sexual reproduction cycle relatively recently or perhaps suppresses it in order to maximize proliferation.
Changing their transcriptional program is a major strategy of microorganisms to adapt to their immediate environment. In S. cerevisiae , the ESR comprises about 900 genes whose expression is coordinately altered to exposure of different types of stress ( Gasch et al. , 2000 ). Saccharomyces cerevisiae, C. albicans, Schizosaccharomyces pombe and C. glabrata live in different environments. Transcript analysis revealed that environmental stress (heat, hyperosmolarity, oxidative and starvation stress) induces a very similar pattern of regulated genes in these fungi ( Chen et al. , 2003 ; Enjalbert et al. , 2003 ; Gasch, 2007 ;). Environmental stress activates a variety of conserved signaling mechanisms. Each of them responds to a particular cue, for example oxidative stress or carbon source starvation. Among them, mitogen-activated protein kinase (MAPK) pathways are central for relaying stress and other environmental signals. The ESR of C. albicans and S. pombe is mainly regulated by the stress-activated MAPKs CaHog1 and Sty1, respectively ( Smith et al. , 2004 ; Enjalbert et al. , 2006 ;). CaHog1 (and most likely also C. glabrata CgHog1) is for full virulence ( Alonso-Monge et al. , 1999 ). In S. cerevisiae , the high osmolarity glycerol (HOG) MAPK pathway senses osmotic stress and, to some extent, also other stress types such as oxidative stress and acetate ( Smith et al. , 2010 ). The function of the C. glabrata HOG pathway seems to be very closely related to that of S. cerevisiae . However, unlike the S. cerevisiae HOG pathway, it also modulates resistance to longer chain weak organic acids such as sorbic acid ( Gregori et al. , 2007 ). Sorbic acid is a powerful activator of the stress transcription factors Msn2 and Msn4 in S. cerevisiae but not of the orthologous factors in C. glabrata ( Schüller et al. , 2004 ; Roetzer et al. , 2008 ;). Sorbic acid leads to rapid localization of CgMsn2-CFP to the nucleus in S. cerevisiae but not in C. glabrata . The difference suggests an as yet unknown weak acid signaling mechanism triggering the HOG pathway in C. glabrata . In S. cerevisiae , the activation of the HOG pathway is triggered by two redundant cell surface membrane sensors, Sho1 and Sln1. These genes have orthologues in C. glabrata . Initial genetic analysis of the first completely sequenced C. glabrata strain [CBS132, ATCC 2001 ( Dujon et al. , 2004 )] suggested that the Sln1 signaling branch was inactive because mutant strains lacking CgSHO1 displayed a high sensitivity to osmotic stress ( Calcagno et al. , 2005 ; Gregori et al. , 2007 ;). Further analysis revealed that a component of the Sln1 branch in this particular C. glabrata strain harbors a nonsense mutation ( Cgssk2 - 1 ), leading to its inactivation. Importantly, other strains (e.g. BG2) do not share this allele.
In S. cerevisiae , the partially redundant transcription factors Msn2 and Msn4 regulate many ESR genes in parallel to MAPK pathways. Msn2 and Msn4 also convey nutrient signals and are regulated by protein kinase A ( Görner et al. , 2002 ). Most Saccharomycotina species contain orthologues to MSN2 . One important functional similarity of ScMsn2 orthologues and paralogues is a conserved stress-regulated nuclear export signal (NES) ( Görner et al. , 1998 ; Roetzer et al. , 2008 ;). Msn2 orthologues from the CTG clade ( Fig. 1a ) do not have the stress-regulated NES and are most probably not involved in stress response. In the C. albicans , the Msn2-like protein (CaMsn4) lacks recognizable homology to the ScMsn2 NES and does not play any role in stress response ( Nicholls et al. , 2004 ). Because of selection against genetic redundancy, most gene pairs that originated during the whole-genome duplication (WGD) event lost one of the copies in subsequent lineages. Thus, retention of both Msn2 and Msn4 in many species (with the exception of Saccharomyces bajanus ) suggests different roles for these factors. Like other post-WGD species, C. glabrata encodes orthologues to both Msn2 and Msn4. CgMsn2/4 are important regulators for the CgESR but seem to be less relevant for glucose signaling ( Roetzer et al. , 2008 ). So far, CgMsn2/4 have not been involved in pathogenicity and thus might play a role in persistence outside the host ( Mundy & Cormack, 2009 ). These stress response factors might control the balance between stress resistance vs. growth required during feast and famine cycles ( Berry & Gasch, 2008 ).
Protective response to oxidative stress is important for a pathogen facing oxidative burst attacks of innate immune cells. Compared with common S. cerevisiae laboratory strains, C. glabrata possesses high intrinsic oxidative stress resistance, which is regulated, similar to that of S. cerevisiae , by the transcription factors CgSkn7, CgYap1 and CgMsn2/4 ( Cuellar-Cruz et al. , 2008 ; Roetzer et al. , 2010 ;). It also produces an indole-derived pigment contributing to oxidative stress resistance ( Brunke et al. , 2010 ). The loss of CgMsn2/4, CgYap1 and CgSkn7 did not affect the survival of C. glabrata when confronted with murine macrophages. Interestingly, the expression of the C. glabrata superoxide dismutase genes ( CgSOD1 / 2 ) is different from that of S. cerevisiae. CgSOD1 seems to be expressed constitutively, and both CgSOD1 and 2 are induced by glucose starvation. Candida albicans is equipped with five SOD genes with highly specialized roles during host contact ( Frohner et al. , 2009 ). A slight adaptation of SOD regulation adapts C. glabrata to an environment with a higher risk of oxidative stress. In addition, many genes of the oxidative stress regulon can also be induced by other stresses such as glucose starvation (our unpublished data). Another adaptation of transcriptional regulation between C. glabrata and S. cerevisiae is the regulation of the catalase gene CgCTA1 . The single catalase of C. glabrata combines both the different transcriptional regulation (carbon source vs. stress) and the different intracellular localization (cytoplasm vs. peroxisomes) of the two S. cerevisiae catalases ( Roetzer et al. , 2010 ). Thus, C. glabrata achieves a similar regulated and localized catalase activity, albeit by the regulation of one gene.
It was recently reported that certain Candida species can suppress the production of reactive oxygen species (ROS) produced by phagocytic cells in murine and human phagocytes, whereas internalization of S. cerevisiae cells enhanced ROS production ( Wellington et al. , 2009 ). Interestingly, in a cell-free system, S. cerevisiae was able to scavenge ROS in a manner similar to the one observed for C. albicans . Therefore, scavenging alone cannot account for the suppression of ROS production. We assume that an additional Candida -specific mechanism is responsible for active suppression. These results are supported by the finding that the transcription pattern of C. glabrata cells during phagocytosis did not show a prominent upregulation of oxidative stress-associated genes ( Kaur et al. , 2007 ). This might be the result of the efficient suppression and/or detoxification of ROS due to expressed protective enzymes. The role of the stress response for C. glabrata adaptation to the host environment is still being studied. So far, the available evidence points to a range of subtle changes of mechanisms known from S. cerevisiae .
Be flexible: phenotypic and morphological switching
Being largely immotile, fungi rely on passive dissemination methods and directed growth to expand into substrate. Candida glabrata primarily grows on surfaces and normally does not penetrate tissue. Under appropriate conditions, penetration into solid medium can be observed. For C. albicans , the switch between yeast and hyphae is one important contribution to virulence ( Kumamoto & Vinces, 2005 ). In contrast, and similar to S. cerevisiae, C. glabrata can form pseudohyphae in response to nitrogen source starvation ( Csank & Haynes, 2000 ). In S. cerevisiae , pseudohyphal growth is assumed to be a common event in search for nutrient-rich, optimal growth substrates. This morphological switch is regulated by a MAPK cascade, which has the transcription factor Ste12 as its final target ( Gancedo, 2001 ). CgSte12 is also essential for the nitrogen starvation-induced formation of pseudohyphae in C. glabrata . In a mouse model, Cgste12 Δ mutants show attenuated virulence ( Calcagno et al. , 2003 ). However, pseudohyphal forms of C. glabrata have not yet been found in clinical specimens, possibly because this is a rare event ( Kaur et al. , 2005 ).
Differently to pseudohyphal growth resulting from specific nutrient limitations, unusual cell aggregates have been observed for C. glabrata cells due to the loss of transcription factor CgAce2. These mutants have a cell separation defect and display a clumping phenotype that is similar to S. cerevisiae ace2 Δ mutants ( Kamran et al. , 2004 ). Interestingly, Cgace2 Δ mutants are hypervirulent in a mouse model and are able to escape from the vasculature and penetrate into tissue ( MacCallum et al. , 2006 ). Stead et al. (2010) found that lack of CgAce2 also changes the secretome of C. glabrata , which might contribute to its hypervirulence. The phenotype of the Cgace2 Δ mutant supports the assumption that C. glabrata has a limited but possibly important ability to actively invade tissues despite its preferred growth on surfaces.
Phenotypic variability is important for evading the immune system. For C. albicans , switching between white and opaque cells is important for virulence ( Kvaal et al. , 1997 , 1999 ). White cells are more suited for bloodstream infections, whereas opaque cells colonize skin surfaces ( Lachke et al. , 2003 ). Candida glabrata cells can undergo reversible phenotypic switching among phenotypes distinguishable by colony coloration on CuSO 4 -containing agar (white to very dark brown) and by the irregular wrinkle colony morphology ( Lachke et al. , 2002 ; Brockert et al. , 2003 ;). In a mouse model, dark brown cells were reported to have an advantage in colonizing the spleen and the liver ( Brockert et al. , 2003 ; Srikantha et al. , 2008 ;). The basis and molecular details of the reported phenotypic variability have not yet been explored in depth.
Be resistant: drug resistance
One of the reasons for the success of C. glabrata is its relatively high drug resistance, in particular toward different azole antifungals. In the last decade, C. glabrata emerged as a cause of mucosal and invasive fungal infections, in part due to its intrinsic or easily acquired resistance to azole antifungals. Therefore, echinocandins such as caspofungin are the treatment of choice. The global transcriptional response to azole treatment and the molecular mechanisms to this drug resistance are reminiscent of S. cerevisiae ( Vermitsky et al. , 2006 ). In S. cerevisiae , the zinc-cluster transcription factors ScPdr1 and ScPdr3 are central for regulating detoxification mechanisms based on transmembrane transporters such as ScPdr5 and ScSnq2. In C. glabrata , CgPdr1 activates the expression of CgCDR1 and CgCDR2 , which are drug efflux pumps of the ATP-binding cassette transporter type ( Sanglard et al. , 2001 ). Gain of function mutations in CgPDR1 enhances antifungal resistance and also increases virulence ( Ferrari et al. , 2009 ). Additionally, defects in mitochondrial function causing respiratory deficiency also increase azole resistance and this, in turn, is linked to the upregulation of both CgCDR1 and CgCDR2 ( Sanglard et al. , 2001 ). In both S. cerevisiae and C. glabrata , Pdr3 – and as a consequence drug resistance – is induced by the loss of the mitochondrial genome (S. Moye-Rowley, pers. commun.; Moye-Rowley, 2005 ). This might be a mechanism for a dedicated retrograde signal of drug-damaged mitochondria. Because the development of intrinsic drug resistance of C. glabrata preceded antifungal drugs, we speculate that these adaptations might be due to frequent exposure to toxic substances originating from the environment of the mucosal flora.
Be aware of the source: nutrients
Like other commensal or pathogenic microorganisms, C. glabrata faces phagocytic cells of the innate immune system during dissemination or after escaping from biofilms. Transcriptome studies revealed that C. albicans, C. glabrata and the encapsulated pathogenic yeast Cryptococcus neoformans , when engulfed by macrophages, switch their expression program to genes involved in the utilization of alternative carbon sources ( Lorenz et al. , 2004 ; Fan et al. , 2005 ; Kaur et al. , 2007 ;). This is a default response, indicating loss of contact to the preferred carbon source glucose in the environment. Therefore, phagocytosed cells are sealed from the environment and have to rely on their endogenous resources or on those acquired from the phagocytic cell. Each successful pathogen has a unique strategy to manage this situation ( Fig. 2 ). Unlike C. albicans , phagocytosed C. glabrata is unable to form hyphae and is trapped within the phagosome. Cryptococcus neoformans triggers phagosomal extrusions, which entails the escape of single cells without killing the phagocytes ( Alvarez & Casadevall, 2006 ). The transcriptional response to phagocytosis also induced genes involved in autophagy, peroxisome function and lipid metabolism ( Fan et al. , 2005 ). It is thought that carbon source and thus energy shortage is a predominant challenge for engulfed pathogenic fungi.
An important mechanism for survival during starvation conditions is autophagy ( Klionsky, 2005 ). Autophagy is a continuous recycling process of cellular constituents and even organelles ( Xie & Klionsky, 2007 ). Moreover, it is important for the virulence of several fungal pathogens of plants and animals ( Hu et al. , 2008 ). In contrast, C. albicans escapes nutrient depletion by forming hyphae and killing engulfing macrophages. Consistently, mutants unable to induce autophagy survived challenge by macrophages ( Palmer et al. , 2007 ). Candida glabrata trapped in a macrophage initially induced peroxisomes, most likely as part of the switch toward utilization of alternative carbon sources ( Roetzer et al. , 2010 ). These peroxisomes were subsequently degraded via the pexophagy pathway, a specific subtype of autophagy. Both selective and bulk autophagy contribute to the survival of C. glabrata during engulfment. Candida glabrata trapped inside the phagosome requires mobilization of resources for survival. Because C. glabrata is relatively resistant to starvation, macrophages apparently have additional strategies to kill the internalized fungal cells. Macrophages are known to cause oxidative damage to microorganisms, but engulfed C. glabrata cells do not show signs of a severe oxidative stress load ( Roetzer et al. , 2010 ). However, during maturation, the phagosome fuses with lysosomes in order to decrease the pH, which initiates enzymatic digestion of engulfed microorganisms ( Stuart & Ezekowitz, 2005 ). In vitro , only the exposure to a combination of low pH and carbon starvation caused loss of viability of C. glabrata . The autophagy-deficient atg11 Δ atg17 Δ mutant was more sensitive to this particular combination of conditions, both in vitro and in vivo . Thus, the energy-demanding equilibration of extracellular pH with limited energy resources might be the challenging environment C. glabrata is exposed to in the phagosome. | We regret that restrictions did not permit us to cite all important papers. We thank the anonymous referees for their very constructive comments. We are grateful to Sylvia Trnka, Wolfram Görner, Martina Dötsch and Christa Gregori for critically reading the manuscript. Work in the C.S. lab has been funded by the AMS, FWF B12-19966, the Herzfelder Foundation and Grant I031-B from the University of Vienna. T.G. is funded by grants from the Spanish Ministry of Science and Innovation (GEN2006-27784E, BFU2009-09168).
Statement
Re-use of this article is permitted in accordance with the Terms and Conditions set out at http://wileyonlinelibrary.com/onlineopen#OnlineOpen_Terms | CC BY | no | 2022-01-12 15:46:43 | FEMS Microbiol Lett. 2011 Jan; 314(1):1-9 | oa_package/a5/5b/PMC3015064.tar.gz |
|||
PMC3015065 | 20941788 | Introduction
Mutations in the SEC23B gene (Sec23 homolog B; MIM# 610512, 20p11.23) cause the vast majority of the congenital dyserythropoietic anemia Type II (CDA II MIM# %224100), an autosomal recessive disorder that represents the most common form of CDAs. Like other forms, it is characterized by a mild to moderate lifelong anemia, ineffective erythropoiesis, and morphologic abnormalities of mature red blood cells (RBC) and their precursors [ 1 ]. Splenomegaly and jaundice are the most evident clinical signs [ 2 , 3 ]. About 10% of patients require frequent transfusions or are dependent on them. Until now, it has not been possible to establish whether these cases are due to more severe mutations or whether they are due to other gene(s) causing the same phenotype, or to the interaction with other intra-erythrocytic defects, such as thalassemia [ 4 ].
CDA II is associated with the presence in the bone marrow of bi- or multinucleated late erythroid precursors, karyorrhexis, and pseudo-Gaucher cells. Diagnosis was, for a long time, achieved by bone marrow examination. It shows 5–10 times more erythroblasts than normal (erythroid hyperplasia) [ 1 , 3 ]; early erythroblasts are relatively normal, but more than 10% of all erythroid cells are binucleated with equal size of two nuclei or multinucleated [ 5 ]. Extensive morphological anomalies of CDA-II erythroblastic cells were observed using electron microscopy (EM), with the most significant being the presence of the so-called double membrane, which is observable in mature RBCs as well [ 6 , 7 ]. This typical aspect is due to residual endoplasmic reticulum [ 8 ]. Polyacrylamide gel electrophoresis in presence of sodium dodecyl sulphate (SDS-PAGE) revealed the presence of a thinner band 3 with an increased anodic mobility due to reduced glycosylation. This cardinal abnormality represents a key for the diagnosis [ 9 ] and could suggest a defect in vesicles trafficking.
SEC23B is an essential component of coat protein complex II (COPII)-coated vesicles that transport secretory proteins from the endoplasmic reticulum (ER) to the Golgi complex. In Saccharomyces cerevisiae this complex is well fine characterized: the coat comprises five subunits, the small GTPase Sar1 (Secretion associated, ras-related), the Sec23–Sec24 “inner coat” complex, and the Sec13–Sec31 “outer coat” complex. Sec23 is a GTPase-activating protein (GAP) that activates Sar1 [ 10 ], and the outer coat is responsible for stimulating this GAP-activity. Therefore, full assembly of the coat stimulates Sar1 GTP-hydrolysis activity [ 11 ]. This enables the coat to depolymerise and be recycled for another round of vesicle biogenesis.
The SEC23B gene spans ∼54 kb on human chromosome 20p11.23 and is composed of 20 exon regions codifying 767 residues arranged in five functional domains: zinc finger, trunk, β-sheet, helical, and gelsolin domain. Each domain interacts with at least one of three coat subunits, in particular, trunk domain is linked by Sec24, Sar1, and Sec31, gelsolin domain is linked by Sar1 and Sec31, and other three domains are linked only by Sec31.
To date, 34 causative mutations in SEC23B gene were described [ 12 – 15 ]. In this work, we described 19 novel variants localized along the entire coding sequence (eight missense, two frameshift, three nonsense, five splicing mutations, and one small deletion). This is the first update of SEC23B mutations involved in CDA II. Moreover, we attempted to characterize the putative effects of the amino acid substitutions on protein structure or function, as well as on mRNA stability by using bioinformatic analyses; we also tried to define the effect of mutations near splice site on mRNA processing alteration by using the same approaches. Finally, by defining the frequency of mutations and those of mutated alleles in each exon, we provided a powerful tool to improve the molecular diagnosis of CDA II. | Methods
Inclusion criteria and definition of the cohort used for mutational analysis
Twenty-eight CDA II patients (15 males and 13 females) from 21 unrelated families enrolled in the CDA II International Registry were investigated (Supporting Information Table Is ). The diagnosis of CDA II was based on the presence of mild to moderate anemia, ineffective erythropoiesis, and morphological abnormalities of the erythroblasts in the bone marrow. Confirmation of the diagnosis was made by mutation screening of the SEC23B gene or at least one of the following analyses: the revelation of the typical narrower band size and faster migration of the band 3 and band 4.5 proteins at SDS-PAGE; the demonstration of superficial appearance of reticulum-endothelial proteins (calreticulin, glucose regulated protein 78, protein disulphide isomerase) on membrane proteins by Western blot (WB) analysis; the presence of a discontinuous double membrane in mature erythroblasts by electron microscopy (EM) [ 5 ].
After signed informed consent, blood was obtained for genetic analysis from the probands. Blood from healthy control subjects was obtained after signed consent according to the Declaration of Helsinki. This project was approved by local ethical committee (University Federico II).
Genomic mutational screening
Genomic DNA preparation, mutational search, oligonucleotide primers design, and direct sequencing were performed as previously described [ 14 ]. Sequence primers are available on request ([email protected]). Nucleotide numbering reflects cDNA numbering with +1 corresponding to the A of ATG translation initiation codon in the reference sequence (Ensembl transcript ID: ENST00000377475). The initiation codon is codon 1.
In silico analyses
To perform a multiple sequence alignment, sequences homologous to human SEC23B were assembled in a multiple sequence alignment using database UniProtKB release 15.0 of March 25, 2009; program NCBI BLASTP 2.0.12 on http://services.uniprot.org ( http://services.uniprot.org/blast/ ). The alignment is made using sequences found in the Uniprot Knowledge-base. The aligned SEC23 sequences were from the species: human (Accession No. Q15437), mouse (Q9D662), Xenopus laevis (Q6DJE0), Danio rerio (Q6AZ98), Caenorhabditis elegans (Q9U2Z1), Saccharomyces cerevisiae (B3LKE0). The secondary structure of SEC23B protein was predicted by using Jpred, a web server that takes a protein sequence and from this predicts secondary structure using a neural network called Jnet ( http://www.compbio.dundee.ac.uk/www-jpred/ ). To localize the missense and nonsense alterations within the protein domains of SEC23B, we submitted the wild type protein sequence on ESyPred3D Web Server 1.0 ( http://www.fundp.ac.be/sciences/biologie/urbm/bioinfo/esypred/ ). This tool implements a homology modeling approach followed by a final analysis with MODELLER release 4 to build a 3D model of the submitted protein [ 16 ]. This routine includes the satisfaction of spatial and geometric restraints and a very fast dynamic annealing. The 3D structures were visualized using the molecular-graphics software Yasara ( http://www.yasara.org ).
The possible impact of the amino acid substitution on the structure or function of protein was predicted by using two bioinformatic tools specifically designed for interpretation of missense variants: PolyPhen ( http://genetics.bwh.harvard.edu/pph/and ) [ 17 ] and PANTHER ( http://www.pantherdb.org/tools/csnpScoreForm.jsp ). PolyPhen identifies homologues of the input sequence via BLAST search and computes the absolute value of the difference between profile scores of both allelic variants in the mutated position. Big values of this difference (PSIC score) may indicate that the studied substitution is rarely or never observed in the protein family. For the input form, we used the protein identifier of SEC23B (Q15437) from the UniProt database. PANTHER calculates the subPSEC (substitution position-specific evolutionary conservation) score based on an alignment of evolutionarily related proteins. This score is the negative logarithm of the probability ratio of the wild-type and mutant amino acids at a particular position. PANTHER subPSEC scores are continuous values from 0 (neutral) to about −10 (most likely to be deleterious). Data from PANTHER can be used to calculate the probability that a given variant will have a deleterious effect on protein function ( P deleterious ): a subPSEC score of −3 corresponds to a P deleterious of 0.5 [ 18 ].
To analyze the possible effect of the mutations near to 3′ and 5′ splice site junctions on splicing processing, we used a web server tool, Splice Site Prediction by Neural Network ( http://www.fruitfly.org/seq_tools/splice.html ). The output of the network is a score between 0 and 1 for a potential splice site.
Secondary structures of the full-length SEC23B mRNA, either wild type sequence that those mutated, were predicted by the program GeneBee ( http://www.genebee.msu.su/genebee.html ).
Frequency of the mutations assessment
The relative richness in mutations (i) has been calculated by the ratio of the mutations falling in each exon of SEC23B gene and the overall count of causative mutations identified until now; similarly, the frequency of mutated alleles in each exon (ii), by the ratio of the number of mutated alleles for each exon and the overall count of mutated alleles. For the mutations near to 3′ and 5′ splice site junctions, the exon flanking splice junction has been considered. The definition of “exon” includes about 200 bases upstream and downstream flanking the coding sequence. | Results
Patient's phenotype
Clinical findings of all 28 affected individuals enrolled in this study were shown in Supporting Information Table Is . The mean age of onset symptoms was 3 years. The mean Hb value was 10 g/dL before splenectomy, and mean ferritin value was 303 μg/L. Mean reticulocyte absolute count at diagnosis was 88797 × 10 6 /L. Clinical data shown here are indicative of a classical framework of CDA II; moreover, they are compatible with those published before [ 14 ]. For most of all patients SDS-PAGE revealed the typical narrower band size and faster migration of the band 3 and band 4.5 proteins; the diagnosis has been also confirmed by the demonstration of superficial appearance of reticulum-endothelial protein GRP78 on membrane proteins by WB.
Identification of novel mutations of SEC23B gene in CDA II patients
All enrolled patients were diagnosed at molecular level. Almost all patients carried two mutations in the compound heterozygous or homozygous state: only in four cases, mutations were not identified in both alleles. We identified a total of 19 novel different allelic variants (Table I ) (Fig. 1 A,B). Of these, eight were missense mutations: c.2270 A>C p.H757P, c.1832 G>C p.R611P, c.1858 A>G p.M620V; c.1735 T>A p.Y579N, c.1254 T>G p.I418M, c.1654 C>T p.L552F, c.1307 C>T p.S436L, c.1733 T>C p.L578P; 3 were nonsense mutations: c.1015 C>T p.R339X, c.1603 C>T p.R535X, c.367 C>T p.R123X; two were frameshift mutations: c.387 (del G) p.L129LfsX26, c.2150 (del C) p.A717VfsX7; 5 were splicing mutations: c.2149−2 A>G, c.1109+1 G>A, c.279+3 A>G, c.1109+5 G>A, c.221+31 A>G; only one consist of a deletion of a codon: c.1857_1859delCAT p.I619del. The 3D model of the protein has been built using the 3D structure 2NUP chain “A” (RCBS Protein Data Bank) as template. This template shares 84.5% identities with our query sequence (Fig. 1 A).
Analysis of 120 control chromosomes showed that these changes were absent in normal subjects. Mutated positions were mostly conserved in analyzed species from Homo sapiens to yeast (Fig. 1 B).
Evaluation of the causal role of mutations by in silico analyses
We evaluated in silico the possible effect of amino acid substitutions of the eight missense mutations by the empirically derived rules of PolyPhen tool. The effect of these alterations on protein structure was predicted to be “probably damaging”, i.e., they are with high confidence supposed to affect protein function or structure (Table I ). By using PANTHER tool, we partially confirmed the predicted functional significance of almost all missense mutations: only the M620V and the S436L variants are predicted to affect minimally protein function (subPSEC −3.2 and −3.09, respectively) (Table I ).
Moreover, two (c.1832 G>C p.R611P, c.1654 C>T p.L552F) out of the eight missense mutations were predicted to alter the secondary structure of SEC23B mRNA (GeneBee program) when compared to that predicted for wild type sequence. Particularly, the free energy of SEC23B mRNA was predicted to be affected in a different manner by the two nucleotide substitution (−640.8 kcal/mol for wild type allele; −650.8 kcal/mol for c.1832 C allele; −636.5 kcal/mol for c.1654 T allele) (Supporting Information Fig. 1 s, panels A-C-D). As positive control, we predicted also the secondary structure of c.325 A, the most frequent mutated allele so far described in CDA II patients, for which the free energy was predicted to be lower (−631.3 kcal/mol) when compared to wild type sequence (Supporting Information Fig. 1 s, panel B). This observation is consistent with published observation: in fact, E109K is already associated with protein instability, with less than 5% of protein detectable compared to wild type [ 12 ].
Splice Site Prediction by Neural Network tool predicted a donor splice site abolition for two out the five splice site mutations, c.1109+1 G>A and c.1109+5 G>A, both falling in intron between exon 3 and 4. Instead, for the mutation c.221+31 A>G was predicted the creation of a new donor splice site with a high score (0.99). Finally, the mutation c.279+3 A>G seems to induce only a limited score reduction of the predicted wild type donor site (from 1.00 to 0.95) (Table I ).
For only one splice site alteration, c.2149−2 A>G, this web server tool failed to predict the wild type acceptor splice site. For this variant, we used another bioinformatic tool, Gene Splicer ( http://www.cbcb.umd.edu/software/GeneSplicer/gene_spl.shtml ), whose prediction showed a mislocalization of the acceptor splice site (upstream of two positions) with a slight reduction of the score (from 6.5 to 6.3).
Frequency of the mutations
To date, 34 causative mutations of CDA II in SEC23B gene were described [ 12 – 15 ], localized along the entire coding sequence of the gene. Here, we described 19 novel variants, bringing the total to 53 causative mutations (Fig. 2 A).
Figure 2 B shows the relative richness of (i) distinct mutations for each exon (solid line), as well as the frequency of (ii) mutated alleles for each exon (dotted line). On the basis of first analysis (i), the greatest number of different mutations falls in the exon 14 (13.2%), 9 and 16 (9.4%), 2 and 5 (7.5%). When we analysed the frequency of mutated alleles for each exon (ii), we found that the largest percentage ranks within the exons 2 (20.5%) and 4 (28.5%), as the most frequent mutations R14W, E109K found in CDA II patients occur within the exons 2 and 4, respectively. | Discussion
CDA II is an autosomal recessive disorder affecting the normal differentiation-proliferation pathway of the erythroid lineage. Up to now, 4 articles described mutations in SEC23B gene as causative of CDA II [ 12 – 15 ]. Mutations previously identified in CDA II patients (Fig. 2 A) are heterogeneous and include frameshift, splicing, deletion, missense, and nonsense variations. In this study, we describe 28 cases from 21 unrelated families with at least one affected proband, in which the diagnosis was made mainly on the basis of biochemical and cytological parameters. In almost all cases, SEC23B gene molecular analysis confirmed the diagnosis. We identified a total of 19 novel different allelic variants localized along the entire coding sequence (Fig. 2 A, Table I ). Most of the patients are compound heterozygotes for two different mutations, while only two are homozygotes (F31P1, F46P1). Patients F33P1, F37P1, F38P1, and F40P1, indeed, show only one mutation (2149−2 A>G, I418M, R339X, and L552F, respectively) in the heterozygous state. However, we neither analyzed the sequence of the promoter region nor did we look for the presence of any large DNA deletions, even if the presence of heterozygous state of several polymorphisms seems to exclude this possibility. As we already assumed, we hypothesized that the elusive mutation lies in SEC23B , because a compound heterozygosity is more likely than two simple heterozygosities, supposing that the second mutation stands at a different locus [ 14 ]. Our cohort of patients showed clinical parameters consistent with what we published before, insofar as patients carrying two missense mutations tend to be more mildly affected compared to those with the association of one nonsense and one missense mutations [ 14 ] (Supporting Information Table Is ). Again, in no case did we find both alleles with a nonsense mutation. This confirm that the absence of SEC23B is supposed to be lethal. The pattern of these mutations emphasizes the high allelic heterogeneity of this condition, as most of the variations are inherited as private mutations. However, we also confirmed that in CDA II there are recurring alleles, as already demonstrated [ 14 ].
This study provided a brief description of a novel set of SEC23B mutations in CDA II disease, localized along the entire coding sequence and divided into: eight missense, two frameshift, three nonsense, five splicing mutations, and one amino acid deletion. The frameshift and nonsense mutations result in the formation of a shorter protein with the loss of some domains. Particularly, the nonsense c.1015 C>T p.R339X and the frameshift c.387 (delG) p.L129LfsX26 cause the loss of part of the trunk domain, fundamental for the interaction with the partners of the “pre-budding complex,” Sec24 and Sar1. The causative role of the missense changes was inferred first by the highly evolutionary conservation of the replaced residues (Fig. 1 B) and then by in silico analysis. We attempted to characterize the putative effects of the amino acid substitutions on protein structure or function by using PolyPhen tool. The empirically derived rules of this web server tool predicted for all missense mutations a “probably damaging” effect on protein structure, i.e., they are with high confidence supposed to affect protein function or structure (Table I ). Of note, the c.2270 A>C p.H757P substitution, which showed the highest PSIC score (3.488) by PolyPhen, is the first mutation identified to date that fall in the exon 20, only 10 amino acids prior to the end of the protein. Also PANTHER confirmed the probability that this variant might have a deleterious effect on protein function, with a P deleterious = 0.78 (subPSEC = −4.2). Since it has been already demonstrated that the most frequent mutation, c.325 G>A (p.E109K), lead to a protein instability [ 12 ], we decided to analyze the effect of all eight missense substitution on mRNA stability. For this purpose, we used a freeware software program, GeneBee, in which we loaded as the input file the full-length sequence of SEC23B mRNA, either wild type or mutated alleles. The server predicted that only two (c.1654 C>T p.L552F, c.1832 G>C p.R611P) out of the eight missense mutations might to alter the secondary structure of SEC23B mRNA when compared to that predicted for wild type sequence. Particularly, for the c.1654 C>T (p.L552F), the free energy of SEC23B mRNA was predicted to be higher (−636.5 kcal/mol) when compared to that of wild type (−640.8 kcal/mol); this value is suggestive of a reduced mRNA stability, which could result in reduced expression of SEC23B mRNA (Supporting Information Fig. 1 s, panels A–C). Instead, the substitution c.1832 G>C (p.R611P) seems to confer a higher stability to mRNA secondary structure, with a predicted free energy of −650.8 kcal/mol (Supporting Information Fig. 1 s, panel D). In the latter case, this variant could lead to the stabilization of mRNA resulting in increased expression of mutated allele. Of note, the free energy of our positive control, c.325 G>A (p.E109K), was predicted to be lower (−631.3 kcal/mol) when compared to wild type sequence (Supporting Information Fig. 1 s, panel B). Since we did not have the availability of the cDNA of patients with mutations close to splicing sites, we attempted to define the effect of mutations near splice site on mRNA processing alteration by using the web server tool, Splice Site Prediction by Neural Network. For both variants, c.1109+1 G>A and c.1109+5 G>A, the donor splice site abolition has been predicted. Instead, for the mutation c.221+31 A>G was predicted the creation of a new donor splice site with a high score (0.99).
To provide a useful information for molecular diagnosis of CDA II, we assessed the relative richness of any distinct mutations in each exon of SEC23B gene, as well as the frequency of mutated alleles for each exon, including also the bases flanking the coding sequence. As most of the mutations are the results of sporadic and independent events, we analysed the relative richness of the single mutations for each exon, to contribute to the improvement of the patient management, of the molecular screening and for providing a more aware genetic counselling. On the basis of this analysis, the most richly mutated exon is the 14 (13.2%), followed closely by exons 9 and 16 (9.4%). Of note, the exon 4, from which the most frequently mutated alleles in CDA II patients (E109K) stem, showed a relative richness of mutations of only 1.9%. However, when we analysed the frequency of mutated alleles for each exon, we found that the most frequently mutated exon 4 (28.5%), closely followed by the exon 2 (20.5%), according to the most frequent mutations E109K and R14W (exons 4 and 2, respectively). Both these available information could facilitate the molecular approach in terms of exon priorities to be analyzed.
Altogether, our results extend the pattern of SEC23B mutations associated with CDA II, propose putative models of causality for almost all the 19 novel variants identified and, finally, provide powerful information to improve the molecular diagnosis of CDA II. | Conflict of interest: Nothing to report.
Re-use of this article is permitted in accordance with the Terms and Conditions set out at http://wileyonlinelibrary.com/onlineopen#OnlineOpen_Terms .
Roberta Russo and Maria Rosaria Esposito contributed equally to this work.
SEC23B gene encodes an essential component of the coat protein complex II (COPII)-coated vesicles. Mutations in this gene cause the vast majority the congenital dyserythropoietic anemia Type II (CDA II), a rare disorder resulting from impaired erythropoiesis. Here, we investigated 28 CDA II patients from 21 unrelated families enrolled in the CDA II International Registry. Overall, we found 19 novel variants [c.2270 A>C p.H757P; c.2149−2 A>G; c.1109+1 G>A; c.387(delG) p.L129LfsX26; c.1858 A>G p.M620V; c.1832 G>C p.R611P; c.1735 T>A p.Y579N; c.1254 T>G p.I418M; c.1015 C>T p.R339X; c.1603 C>T p.R535X; c.1654 C>T p.L552F; c.1307 C>T p.S436L; c.279+3 A>G; c. 2150(delC) p.A717VfsX7; c.1733 T>C p.L578P; c.1109+5 G>A; c.221+31 A>G; c.367 C>T p.R123X; c.1857_1859delCAT; p.I619del] in the homozygous or the compound heterozygous state. Homozygosity or compound heterozygosity for two nonsense mutations was never found. In four cases the sequencing analysis has failed to find two mutations. To discuss the putative functional consequences of missense mutations, computational analysis and sequence alignment were performed. Our data underscore the high allelic heterogeneity of CDA II, as the most of SEC23B variations are inherited as private mutations. In this mutation update, we also provided a tool to improve and facilitate the molecular diagnosis of CDA II by defining the frequency of mutations in each exon. Am. J. Hematol., 2010. © 2010 Wiley-Liss, Inc. | Supplemental material | CC BY | no | 2022-01-12 15:46:43 | Am J Hematol. 2010 Dec; 85(12):915-920 | oa_package/61/0b/PMC3015065.tar.gz |
||
PMC3015066 | 21221171 | Introduction
Migraine is a common disease and a leading cause of disability [ 1 ]. In the United States, work loss due to migraine is estimated to cost $13 billion [ 2 ]. Calcitonin gene-related peptide (CGRP) is a neuropeptide that may play a key role in the pathophysiology of migraine [ 3 , 4 ]. Since CGRP receptor antagonists lack direct vasoconstrictor activity, this new mechanism of action may offer advantages over current triptan treatments, where cardiovascular liabilities are a major perceived risk [ 5 ], or may offer another treatment option in patients who do not respond to triptans.
Telcagepant (formerly known as MK-0974) is an orally available, selective and potent antagonist of the human CGRP receptor [ 6 – 8 ]. Telcagepant is currently in development for the acute treatment of migraine, and may have the potential to treat migraineurs with cardiovascular disease. The efficacy and tolerability of telcagepant has been demonstrated in phase 3 clinical trials at doses of 150 and 300 mg [ 9 , 10 ].
Non-clinical assessments of telcagepant in monkeys and rats suggested species dependent non-linear pharmacokinetics following oral doses of telcagepant. Intestinal first-pass metabolism was concluded to play a significant role [ 11 ]. A similar observation was made in early clinical studies [ 12 ] in which greater than dose proportional increases in exposure following oral administration were apparent. Here, the dose proportionality of telcagepant is formally characterized in a definitive oral dose proportionality study and compared to the results from a second study in which the dose proportionality following IV administration of telcagepant was characterized. | Material and Methods
Study Design
Two single-center, open-label, randomized studies were performed. The oral dosing study (Merck Protocol 029) was a 5-period crossover design with doses of 50, 150, 300, 450 and 600 mg. The IV study (Merck Protocol 019) was a 7-panel parallel design with doses of 5, 10, 25, 50, 100, 175, and 250 mg; the study used a parallel design because of dosing limitations imposed by the inclusion of Captisol® in the IV formulation (possible renal toxicity). The oral study was conducted at CPI Comprehensive Neuroscience, Inc., Miramar, FL, USA and the IV study was conducted at CEDRA Clinical Research, Austin, TX, USA. The studies were performed in conformance with legal requirements for the ethical conduct of research in human subjects and were approved by Independent Investigational Review Board, Inc., Plantation, FL (oral study) and IntegReview Ethical Review Board, Austin, TX (IV study). All subjects gave written informed consent.
Patients
Both studies enrolled healthy, non-smoking adult male and female subjects. Subjects were required to have a body mass index ≤33 kg/m 2 (oral study) or ≤35 kg/m 2 (IV study), and females were required to have a negative pregnancy test at screening and agree to remain abstinent or use appropriate double barrier contraception during the study. The oral study enrolled 20 subjects; 14 males ages 25 to 55 years, and 6 females, ages 39 to 43 years. One subject who completed the study but took concomitant medication was excluded from the pharmacokinetic analysis. The IV study enrolled 70 subjects (10 per dose group); 36 males ages 18–44 years and 34 females ages 19–43 years.
Drug Supplies
Oral telcagepant was supplied in soft gelatin (oral soft elastic) capsules manufactured by Banner Pharmacaps, Inc. The bulk drug used was telcagepant, monopotassium salt, ethanolate, supplied by Merck Research Laboratories. The IV formulation of telcagepant consisted of telcagepant, monopotassium salt, ethanolate infused as a 6 mg/mL solution (containing 35% Captisol®).
Study Procedures
Both studies were performed in specialist clinical research units. Concomitant medication was prohibited from 48 hours before dosing. For the oral 5-period, crossover study, the order in which the subjects received each dose was randomized according to a computer generated allocation schedule. Each subject received each dose with 240 mL of water at the same time in each period, after an overnight 8-hour fast. There was at least a 7-day washout interval between each treatment period. For the IV study, the treatment group to which subjects were assigned (10 per group) was enrolled sequentially. The study started with the lowest dose group (5 mg) and subsequent groups were dosed incrementally (10, 25, 50, 100, 175, and 250 mg) dependent on the tolerability and pharmacokinetic profile observed in the previous group. For each subject, after an overnight fast, the appropriate volume of IV solution was withdrawn from pre-filled vials and then infused via infusion tubing. Doses were administered at a constant rate of 20 minutes via a syringe-type, rate-controlled, infusion pump. In both studies, plasma samples were collected at predose and at specified time points up to 48 hours following dosing and subject safety was monitored by repeated clinical and laboratory evaluations, including assessment of adverse events.
Bioanalysis
Plasma samples were analyzed for telcagepant concentrations; details of the bioanalytical methods are published elsewhere [ 13 ]. Both analyte and internal standard were isolated from human plasma using solid phase extraction in 96-well format. The extracted analytes were analyzed by high performance liquid chromatography with tandem mass spectrometry. The lower limit of quantitation for telcagepant was 5 nM. The linear calibration curve range was 5 to 5000 nM.
Pharmacokinetic and Statistical Methods
The (apparent) terminal rate coefficient (λ) was estimated by regression of the terminal log-linear portion of the plasma concentration-time profile. The (apparent) terminal half-life was calculated as the quotient of ln(2) and λ. The area under the plasma concentration versus time curve to the last time point (AUC last ) was calculated using the linear trapezoidal method for ascending concentrations and the logarithmic trapezoidal method for descending concentrations. The area under the plasma concentration versus time curve (AUC 0-∞ ) was estimated as the sum of AUC last and the quotient of the last measured concentration and λ. The maximum concentration observed (C max ) and the time in which C max was observed (T max ) were assessed by inspection of the plasma concentration data. For the intravenous study, the actual volume of telcagepant delivered was estimated from the difference in mass of the filled syringe, before and after telcagepant administration. The clearance was calculated as the quotient of the dose and AUC 0-∞ . The volume of distribution at steady-state was calculated as the product of the clearance and mean residence time at steady-state. WinNonlin (Pharsight) 5.0.1 was used for all pharmacokinetic calculations.
For the oral study, the primary assessment of dose proportionality for telcagepant AUC 0-∞ was performed using the power model, as defined on the log scale below, in which μ is the overall mean, β is the overall slope, S j is the random effect of subject j and ε is the random error:
The slope (β) was estimated using a mixed-effects model with dose as a fixed effect and with subject as a random effect. The slope was estimated with an appropriately chosen contrast. On the original scale, the power model is expressed as AUC = α j (dose) β , with α j = exp(μ + S j + ε j ). Under this model, exact dose-proportionality is present when the true value of β = 1. The degree of dose proportionality over the dose range was estimated by calculating the 90% confidence interval for the slope.
To better illustrate the effect of the estimated slope, the true increase in AUC 0-∞ resulting from a 12-fold increase in dose (i.e., over the entire dose range studied) was estimated using the relationship AUC 600-mg /AUC 50-mg = 12 β , where β was the slope estimated from the power model. The 90% confidence interval for this ratio was also calculated by substituting the lower and upper bound of the 90% confidence interval computed for the slope into the relationship above. Based on the observed data, the method to estimate dose proportionality over the entire range was successively applied to smaller segments (50–450 mg, 50–300 mg) to estimate dose proportionality over other ranges of interest.
Similar analytic methods were used to analyze data from the IV study. Due to the parallel group design, the random effect of subject j was not included in the power model. To better illustrate the effect of the estimated slope, the true increase in AUC 0-∞ resulting from a 50-fold increase in dose (i.e., over the entire dose range studied) was estimated using the relationship AUC 250-mg /AUC 5-mg = 50 β , where β was the slope estimated from the power model. The 90% confidence interval for this ratio was also calculated by substituting the lower and upper bound of the 90% confidence interval computed for the slope into the relationship above. | Results
Single Oral Dose Pharmacokinetics
Telcagepant was rapidly absorbed with median time to peak plasma concentration (T max ) ranging from 1 to 2 hours ( Table 1 ). Telcagepant concentrations declined from the peak plasma concentration (C max ) in a biphasic manner ( Figure 1 ) with an apparent terminal half-life (t 1/2 ) of approximately 4 to 7 hours ( Table 1 ). T max was similar across doses. The apparent terminal half-life appeared to increase with dose; however, this may be confounded by the inadequate characterization of the terminal phase at lower doses because the concentration fell below the lower limit of quantification.
Table 2 contains point estimates and 90% confidence intervals for the slopes from the power model used to assess dose proportionality of telcagepant AUC 0-∞ , AUC 0-4h and C max over the dose ranges of interest. Greater than dose proportional increases were observed in the area under the concentration-time curve estimated to infinity (AUC 0-∞ ) and estimated to 4 hours post dose (AUC 0-4h ) over the range of 50 to 600 mg, 50 to 450 mg, and 50 to 300 mg. While the increase in AUC were greater than dose proportional over all ranges evaluated, C max was greater than dose-proportional over the range of telcagepant 50 to 600 mg, but approximately dose-proportional over the ranges of 50 to 300 mg and 50 to 450 mg. The estimated regression line and corresponding 95% confidence band are displayed graphically in Figure 2 for AUC 0-∞ .
Single IV Dose Pharmacokinetics
The pharmacokinetics of telcagepant after single IV dose administration of telcagepant to healthy subjects were evaluated over the dose range of 5 mg to 250 mg ( Table 3 ). Following single IV doses, telcagepant exhibited a multiphasic plasma concentration-time profile ( Figure 3 ). The terminal half-life appeared to increase with dose and plateau near the apparent terminal half-life estimate for an oral dose of 300 mg ( Table 3 ). The low half-life estimates from lower doses may not represent the terminal half-life as the concentration of telcagepant was near or below the lower limit of quantitation at those doses. A minor decrease in the clearance was noted with increasing doses, and the volume of distribution increased with dose, though these changes may reflect inaccuracies in the estimation of the terminal phase at the lower doses due to concentrations being below the lower limit of quantitation.
Table 4 contains point estimates and 90% confidence intervals for the slopes from the power model used to assess dose proportionality of telcagepant AUC 0-∞ , AUC 0-4h and C max over the 5 mg to 250 mg dose range. Following administration of a single IV dose of telcagepant, an approximately dose-proportional increase in AUC 0-∞ , AUC 0-4h , and C eoi was observed. The estimated regression line and corresponding 95% confidence band are displayed graphically in Figure 4 for AUC 0-∞ .
Safety and Tolerability
Telcagepant was generally well tolerated in both studies. No serious clinical or laboratory adverse experience were reported and no subject discontinued because of an adverse experience. There were no consistent treatment-related changes in laboratory, vital sign, or ECG parameters.
In the oral study, all clinical adverse experiences were mild and resolved. A total of 16 clinical adverse experiences were reported by 4 out of 20 subjects; 1 during a period in which the subject received a single dose of 150 mg, 6 during periods in which subjects received a single dose of 300 mg, 7 during periods in which subjects received a single dose of 450 mg, and 2 during periods in which subjects received a single dose of 600 mg. Those clinical adverse experiences which occurred more than once were constipation, upper respiratory tract infection (one was considered viral), headache, dizziness and nausea.
In the IV study, a total of 29 out of 70 subjects reported 47 adverse experiences. Of the 47 adverse experiences, 17 occurred prior to dosing and 30 occurred on-treatment. Twenty-five of the 47 total reported adverse experiences were related to the IV infusion (e.g., infusion pain). Among the 12 clinical adverse experiences which were on-treatment and were not related to the IV infusion, those which occurred more than once were headache and dizziness. | Discussion
In the field of migraine therapy, where most medication is self-administered by patients, the availability of an oral formulation is particularly important. Telcagepant is a novel, orally active and selective CGRP receptor antagonist being developed for the acute treatment of migraine. Telcagepant, at doses of 150 and 300 mg of the oral soft elastic liquid filled capsule formulation, demonstrated efficacy superior to placebo in human clinical studies [ 9 , 10 ]. The studies described here were conducted to understand the relationship between dose and plasma exposures following single oral doses of telcagepant capsules ranging from 50 mg to 600 mg and single IV doses of telcagepant capsules from 5 to 250 mg.
Following oral administration, telcagepant is rapidly absorbed with a median T max of 1 to 2 hours and an apparent terminal half-life of 4 to 7 hours [ 12 ]. Early clinical studies suggested greater than dose proportional increases in exposure following oral administration, which may indicate changes in clearance or bioavailability [ 12 ]. In non-clinical studies, oral dosing of telcagepant in rhesus monkeys was observed to result in much greater than dose proportional increases in exposure. The non-linear pharmacokinetics of telcagepant in rhesus monkeys was attributed to intestinal first-pass metabolism [ 11 ]. The initial clinical observations in the context of the non-clinical results prompted the present formal evaluation of the dose proportionality in a definitive oral dose proportionality clinical study and in an IV dose proportionality clinical study.
In the present study, the pharmacokinetics of telcagepant following oral administration to healthy adult subjects were non-linear with greater than dose proportional increases in AUC 0-∞ and AUC 0-4h over all ranges evaluated while C max was greater than dose proportional over the range of 50 to 600 mg. C max was approximately dose proportional over the ranges of 50 to 300 mg and 50 to 450 mg. The possibility that the non-linear oral pharmacokinetics observed in healthy adult subjects were due to interactions at the gastrointestinal level was evaluated by comparing these findings with the dose proportionality of telcagepant following single IV doses.
Following IV administration, approximately dose proportional increases in AUC 0-∞ , AUC 0-4h , and C eoi were observed with minimal change in the clearance. The terminal half-life appeared to increase with dose, and plateau near the apparent terminal half-life estimate for an oral dose of 300 mg; however, because the concentration of telcagepant was near or below the lower limit of quantitation at lower doses, it is unclear whether this trend is solely due to the pharmacokinetics of telcagepant. Instead, it is likely that the half-life estimates at lower doses do not represent the terminal half-life. A minor decrease in the clearance was observed with increasing doses while the volume of distribution increased with dose. These results suggest that the changes in distribution and elimination are minor, and these changes may not be meaningful in the context of dose proportionality. It is important to understand that due to safety limitations of an excipient in the IV formulation, it was not possible to dose subjects more than once (as in a cross-over study). Therefore, this study used parallel panels of subjects to construct the dose escalation, though it is expected that this design does not impact the conclusions to a meaningful extent.
Telcagepant is metabolized via CYP3A4 and is a substrate of P-glycoprotein in addition to being a time dependent inhibitor of CYP3A4. Both of these proteins are likely to play an important role in the bioavailability of telcagepant and the activity of each protein in the gut wall is theoretically saturable at clinically achievable drug concentrations. Thus, data from the oral dose proportionality study, data from the IV dose proportionality study, and knowledge of the molecular determinants of the disposition of telcagepant, are consistent with the possibility that greater than dose proportional increases in exposure of orally administered telcagepant may be a result of saturable metabolism/efflux of telcagepant in the gut. The relative bioavailabilty of telcagepant increases with dose.
The oral formulation of telcagepant used in the oral dose proportionality study (oral soft elastic liquid filled capsule) was an early formulation. However, this formulation demonstrated bioequivalent exposure, with a dose adjustment, with the anticipated tablet market formulation at two separate doses (140 mg of the tablet compared to 150 mg of the capsule and 280 mg of the tablet compared to 300 mg of the capsule, unpublished data). It is therefore expected that the conclusions about dose proportionality would apply to the anticipated market formulation.
In both studies, telcagepant was generally well tolerated in healthy male and female subjects when given as single oral doses, ranging from 50 to 600 mg, and single IV doses, ranging from 5 to 250 mg. This is in agreement with data from a clinical trial looking at the efficacy and tolerability of single oral doses of telcagepant up to 600 mg in migraine patients [ 8 ]. The long term safety and tolerability profile of telcagepant in migraine patients is the subject of ongoing studies. | Conclusion
In summary, telcagepant is generally well tolerated in young, healthy male and female subjects and exhibits a pharmacokinetic profile suitable for the acute treatment of migraine. The pharmacokinetics of telcagepant after an oral dose were observed to be non-linear with greater than dose proportional increases in exposure. When administered IV, the pharmacokinetics were approximately dose proportional. | Re-use of this article is permitted in accordance with the Terms and Conditions set out at http://wileyonlinelibrary.com/onlineopen#OnlineOpen_Terms
Introduction
Telcagepant (MK-0974) is a novel, orally active and selective CGRP receptor antagonist being investigated for acute treatment of migraine. Early clinical data suggested greater than dose proportional increases in exposure following oral administration. The aim of the present studies was to definitively characterize the oral and IV dose proportionality of telcagepant.
Methods
Healthy adult subjects were enrolled in two separate open-label randomized dose proportionality studies: 1) single oral dose crossover from 50 to 600 mg (N = 19); 2) single IV dose parallel group from 5 to 250 mg (N = 10 per dose). Blood samples were collected at time points from 0 to 48 hours postdose.
Results
Telcagepant was rapidly absorbed with a T max of approximately 1 to 2 hours after oral administration. The terminal half-life was approximately 8 to 9 hours after IV dosing and approximately 4 to 7 hours after oral dosing. Oral administration of telcagepant resulted in greater than dose proportional increases in exposure, while IV administration resulted in approximately dose proportional increases in exposure.
Conclusions
Telcagepant was generally well tolerated. Oral telcagepant exhibits non-linear pharmacokinetics. | All authors are responsible for the work described in this paper. All authors were involved in at least one of the following: conception, design, acquisition, analysis, statistical analysis, interpretation of data, drafting the manuscript and/or revising the manuscript for important intellectual content. All authors provided final approval of the version to be published.
The authors would like to acknowledge the contributions of Frederick A. Bieberdorf who provided and cared for subjects in the study and critically reviewed the study proposal and Sheila Erespe from Merck for her help with formatting and editing the manuscript.
Conflict of Interest:
Merck & Co., Inc. provided financial support for the study. Tae H. Han, Rebecca L. Blanchard, John Palcza, Ashley Martucci, Cynthia M. Miller-Stein, Deborah Panebianco, Ronda K. Rippley, Christopher Lines, and M. Gail Murphy are employees of Merck and own stock and/or hold stock options in Merck. Maria Gutierrez has received research support from Merck. | CC BY | no | 2022-01-12 15:46:43 | Arch Drug Inf. 2010 Dec; 3(4):55-62 | oa_package/1b/a9/PMC3015066.tar.gz |
|
PMC3015780 | 21278900 | A means for measuring levels of soluble huntingtin proteins in clinical samples is essential for assessing the biological effects of potential mutant huntingtin (mtHtt) modifying treatments being developed for Huntington’s disease (HD). We have optimized a previously described cell-based Homogeneous Time Resolved Fluorescence method that can measure soluble mtHtt and its ratio to the total Htt (tHtt) in blood buffy coats [1]. The results of the optimization and assay qualification indicate the assay to be specific for mtHtt in HD compared to Control subjects, highly sensitive, and technically and biologically reproducible. We therefore generated a Good Laboratory Practice Standard Operating Procedure which we validated, using 30 HD and 8 control buffy coat samples in which significant differences in mtHtt levels were found. We intend to deploy the assay to evaluate sample sets from observational and therapeutic studies enrolling HD subjects to further validate soluble mtHtt measurement by HTRF as a biomarker for HD and to explore its potential uses. | Introduction
The proximate cause of neuronal dysfunction and death in Huntington Disease (HD) is the presence of soluble mutant huntingtin ( mt Htt) protein or its cleavage products [1] [ [2] [3] . Consequently, soluble mt Htt is the most salient target for disease modifying therapies and measurement of soluble mt Htt is essential for evaluating the success of therapies intending to affect mt Htt levels. Using a cell based Homogeneous Time-Resolved Foerster Resonance Energy Transfer (HtrFRET) assay or HTRF, Weiss et al [1]have shown that mt Htt was specifically detectable in human brain and blood cell sub-populations, as well as in tissues and blood from HD mouse models, in correlation with the disease. Furthermore, using properties of Lumi4-Terbium cryptate (Tb) Weiss et al [4] have developed a triplex assay, in which mt Htt, as well as an epitope distant from the Htt N-terminus, hence an indication of total huntingtin ( t Htt), are measured concomitantly in the same sample.
This assay, applied to peripheral blood samples, hence easily accessible, from HD patients, could serve as a biomarker for assessing whether systemic therapies affect levels of mt Htt. We chose buffy coat (BC) as a blood subpopulation showing measurable Htt levels (Weiss et al [5] and that is also relatively easy to quantify and work with. To provide the necessary consistency, reliability, and quality control essential for a clinical diagnostic bioassay, we have optimized and qualified the assay for BC samples and developed Standard Operating Procedures (SOPs), so that the assay can be conducted under Good Laboratory Practice (GLP) conditions and in compliance with Federal Drug Administration (FDA) requirements. We first optimized the assay to measure soluble t Htt and mt Htt in human brain lysates from control and HD subjects. Parameters providing the highest specificity and sensitivity for mt Htt were selected and the assay was used to test sets of unblinded human BC samples from HD and control subjects, to qualify/ validate the assay in compliance with GLP. Following the assay qualification, a formal SOP consistent with FDA requirements for a potential diagnostic assay was issued. Using this SOP, we have started to test clinical samples to assess relative blood levels of mt/t Htt using blood samples from active single and multi-center clinical research studies. Our results using this SOP confirm that the assay is sensitive, linear within the dynamic range and reliable and we are deploying it as a GLP assay to explore its potential as a peripheral biomarker for HD.
Material and Methods
Antibodies: The monoclonal antibodies used in this cell-based HTRF assay are specific for selected epitopes on the Htt molecule. The antibodies include: 2B7 [5] ,which binds to the first 17 amino acids (aa) of normal and mt Htt; MW1 [6] ,which is specific for expanded polyglutamine sequences (polyQ) and binds to mt Htt, but not to normal Htt; and 2166 (Millipore Corp, Cat# MAB2166 [7] , which binds to a Htt epitope starting at aa 444, and recognizes both normal and mt Htt. 2B7 was conjugated to Tb (Cisbio), to serve as the donor for HTRF, while MW1 was bound to AlexaFluor 488 (AF488, Alexa Fluor 488, Invitrogen, Cat#A20181) and 2166 was conjugated to d2 (Cisbio) to serve as acceptors and to enable multiplexing between the donor and the two acceptors. Additional antibody pairs were also evaluated (not shown), however the antibodies mentioned above showed the largest HTRF signals and discrimination between the controls and HD and were therefore selected for the optimization study.
Human Samples: Human blood samples were collected at the MGH Huntington’s Disease Center by standard phlebotomy from consenting HD and control subjects participating in the Partners IRB approved REVEAL-HD study. Genotyping was confirmed by the MacDonald laboratory (Molecular Neurogenetics Unit Genotyping Resource) at MGH. Twenty-four of these samples were from 12 individuals in whom venipuncture was repeated within six weeks. The peripheral blood samples, collected into EDTA tubes, were centrifuged at 4°C for 10 minutes (min) at 1000 x g. BC fractions were isolated and centrifuged again for 20 min at 4°C at 15800 x g, then stored at -80°C until their analysis. The BC pellets were pre-diluted 1:2 in “lysis buffer” and an additional two serial two-fold dilutions were performed.
Brain Tissue Samples: Postmortem frontal cortex samples from HD and control patients obtained from the Brain Bank of the Alzheimer Disease Research Center (ADRC) at MGH were homogenized in 10x volume “lysis buffer” as above. Stocks of lysates were prepared, their protein concentrations determined and the stocks aliquoted and stored at -80 ° C for use as quality control (QC) samples and for assay development and troubleshooting purposes.
HTRF Assay : Our optimization process examined plate formats; the performance of various plate readers and filters, reaction volumes, buffer compositions, cell lysate and antibody concentrations, quality control and normalization approaches, and reaction incubation times. Optimal conditions utilized white, 384 well/plate (Perkin Elmer, Cat# 6008280), and a total volume of 15 μL/well. The cell lysates at 12.5 μL/well were derived from a 1:1 dilution of the BC samples and the two subsequent two fold dilutions, or from 9 μg/well of each control and HD brain lysate diluted in the lysis buffer. The fluorophore-conjugated antibodies were in a 2.5 μL/well reaction buffer, that included a final dilution of 1.25ng/well 2B7-Tb cryptate-labeled antibody and 2.35ng/well MW1-AF (different lot than previously mentioned) and 13ng/well 2166-d2-labeled antibodies in 50 mM NaH 2 PO 4 (NaH 2 PO 4 –H 2 O, Sigma Aldrich, S9638), 200 mM KF (Sigma-Aldrich, Cat# 402931-100G), 0.1% BSA (Sigma, Cat# A-7906) and 0.05% Tween 20 (Fisher, Cat# BP337-500). The concentrations of conjugated-antibodies are lot dependent, therefore their concentration is adjusted with each new lot to yield similar ratio of mt/t Htt of HD/control human brain lysates. The plates were spun for 3 min at 500 x g then incubated in the dark on an orbital shaker at 4°C for 2 hours (hr). The HTRF signals were read on a VICTORX5 plate reader (Perkin Elmer). The emission from Tb was captured at 615 nm, following optimization and we term the values, for simplicity, maximum (Max) signal. The ratio of signals at the 665/615 nm and 510/615 nm, respectively, represents a specific, artifact corrected determination of the two antibody pairs: 2166-d2/2B7-Tb and MW1-AF/2B7-Tb, simultaneously bound to Htt and reveal information about t Htt/Max and mt Htt/Max signal, respectively.
dsDNA Content was measured in all the wells (including the dilutions of each sample) of the assay following the reading of the HTRF signals using PicoGreen (Quant-iT PicoGreen dsDNA Quantitation Kit, Invitrogen, Cat# P11496). A standard curve of dsDNA was generated on every plate, to make sure that the samples measured are within the dynamic range of the standard curve. The plate was read at an excitation of 485 nm and an emission of 535 nm.
The protein concentrations of the human brain lysates and BC samples were determined using DC Bio-Rad, Protein Assay Reagents, per the kit instructions.
Purified 548Q25, 548Q46 and 548Q72 Htt fragments: Purification of recombinant huntingtin protein fragments was performed as previously described (Weiss et al [ [5] [4] ). We thawed the cryo-stored purified Htt fragments at room temperature, prior to spiking them into the control buffy coat sample at 1:1 ratio, yielding Htt fragments concentration of 300-0 ng/mL and a 1:4 final dilution of the control BC used. The HTRF assay of the spiked samples and controls proceeded as described above.
Results
Tb has high stability and long-lived fluorescence emission over a wide spectrum (480-720 nm) following its excitation at 340 nm. Its emission signal is captured at 615 nm-maximum (Max) signal. The energy transferred from the Tb excites the two acceptors: d2, with the peak emission at 665 nm (red) and AF488 (AF), with its peak emission at 510 nm (green). These filters were fine tuned as a result of optimization of the initial assay (Weiss et al [5] ). Time resolved detection based on Tb characteristics eliminates the background fluorescence, resulting in a more specific signal. The interaction between the fluorophores reflects the proximate binding of antibodies that are conjugated to their specific epitopes on the Htt molecule (Weiss et al [4] ) in this triplex assay. The antibodies we used in the HTRF assay, their conjugated fluorophores and their peak emissions and the antibodies’ epitopes on the Htt molecule are represented schematically in Figure (Fig.) 1:
HTRF Assay Schematic Summary
To qualify the HTRF assay and validate soluble mt Htt as a potential clinical biomarker, we developed and optimized a protocol based on the Weiss et al assay [5] , following GLP, to insure: standardization, reproducibility, clinical utility, the potential for regulatory acceptance, and to provide a basis for future modifications or variations. The immediate technical goal of the optimization process was to develop a protocol that provides the strongest detection of mt Htt (MW1/2B7) and mt/t Htt (MW1/2166) HTRF signal ratios, using human brain lysates as QCs. Several of the parameters evaluated experimentally for the clinical assay optimization are listed below.
Plate Format and Plate Color: We determined that 384 well plates gave a stronger specific signal compared to 96 well plates (results not shown), which is related to the optical interaction of the light, the fluorophores and the height of the reaction fluid (Cisbio, technical information). We compared black and white 384 well plates and evaluated the impact on the relative HTRF signals. We had hypothesized that black plates would reduce the fluorescent background signal better than white plates, however white 384 well plates showed consistently higher HTRF signals compared to black plates. This suggests that the time-resolution fluorescence conferred by the Tb, reduces the background sufficiently, irrespective of plate color and therefore white plates were used from that point forward.
Reaction Volume and Intra-assay Reproducibility : We compared (Fig. 2) the effects of different reaction volumes in the triplex HTRF using: 12 μL/well versus 15 μL/well in HD compared to control samples. Our results, represented as average (AVE) of HTRF raw signals ± standard deviation (STDEV) of the triplicates, indicate that at 15 μL/well there is better intra-assay reproducibility than at 12 μL/well for all the antibodies/fluorophores combinations: 2166-d2, 2B7-Tb and MW1-AF488, measured in the same well. The smaller coefficients of variation (CV) between the replicate wells increase confidence in the accuracy of the measurement and the signal seems more specific showing a higher signal with MW1-AF488 for HD compared to controls. Therefore, going forward we used 15 μL/well as the reaction volume.
Reaction Media (KF concentration): We compared (Table 1) concentrations of 83 mM KF to 200 mM KF in the reaction media (concentrations in use by different labs) and we found that 200 mM KF gave higher signal ratios, due to reduced non-specific binding in the control brain lysates, therefore we adopted 200 mM KF for our reaction media in the HTRF assay.
Table 1. Reaction Media- KF Concentration Control (CTRL) and HD brain lysates were tested using 384 white microtiter plates with 15 μL/well total volume. The antibodies concentrations were as described in the Material and Methods; the reaction media included 83 mM KF or 200 mM KF.
HTRF Reaction Time Course : Evaluation of the time course of the HTRF assay (Fig. 3) revealed an increase in the mt/t Htt (MW1-AF/2166-d2) of HD/control HTRF signal peaking at 2-3 hrs, followed by a decrease in this ratio from 9.2 at 3 hrs to 5.0 at 23 hrs, due mostly to an increase with time of the non-specific binding of MW1-AF488 in the CTRL brain. We therefore chose a 2 hr reaction time for the SOP. Stability of the signal within the 2- to 3 hrs incubation interval, indicates that the assay is robust.
Quality Control:
A. Consistency of Signal across the Plate: Plate based assays can provide inconsistent signals across the plate: top to bottom and/or left to right. Since our goal is to detect small changes in subjects with disease progression or changes due to a mt Htt altering therapy it is important to minimize and account for such potential variability. To this end we evaluated 384 well white plates from different manufacturers and we also assessed this parameter when we evaluated the different plate readers by testing brain lysates from control and HD subjects in triplicates at the four corners of the plates. The signal consistency is evaluated as the Z’ defined as:
For HTRF assays a Z’ value ≥0.5 is considered satisfactory.
Although minimal, we observed that the left to right variability was lower than top to bottom variability in the plates we determined experimentally to be best for our application. We therefore use only the top 60% of the plate, and the entire plate from left to right. To minimize and control for potential plate related variability we measured Htt in QC samples, control and HD brain lysates, in triplicates on both left and right extremes of each plate when testing clinical samples and calculated the Z’ based on a total of 6 replicates each for control and HD. A Z≥0.5 was determined as a QC’ acceptance criteria for the assay.
B. Signal Normalization: Since the binding to Htt of all the antibodies involved occurs simultaneously in the same cell lysate/well, we can normalize the signal for mt Htt to the signal for t Htt resulting in a ratio- mt/t Htt that identifies each sample. Furthermore, these ratios allow comparisons between different samples to assess the relative levels of mt/t Htt at the time of measurement. Since there is variability in the concentration of cells/mL of blood based on processing variations, sex, age, and other factors we make every attempt to compare the mt/t Htt ratios between samples corresponding to similar dsDNA content and/or protein concentrations, which are valuable QC factors. This is facilitated by measuring dsDNA, using PicoGreen, in all the wells, following the reading of the HTRF signals and by measuring the protein concentration of the lysates and comparing the resulting HTRF ratios at the sample dilutions corresponding to similar dsDNA and protein concentrations.
Plate Reader : We evaluated the performance of five different plate readers from three manufacturers and found differences (not shown) of up to 46% in regard to the highest HTRF signal ratios for the HD/control brain lysates and signal consistency across the plate (Z’ values). In addition, cost and technical support played a role in the decision process. We selected the VICTORX5 (Perkin Elmer) as our plate reader, and the assay qualification/validation and the analyses of the BC and QC samples that followed were performed using it. We started to test the clinical samples following the installation of Security Software and the implementation of Installation Qualification (IQ) and Operational Qualification (OQ) by the VICTOR5X manufacturer, as required by the FDA for a potential diagnostic assay.
Assay Qualification/Validation Using Clinical Buffy Coat and QC Samples: Using the optimized protocol and the VICTOR5X we proceeded to confirm the specificity, linearity and sensitivity of the assay. We then proceeded with assay qualification/validation to assess the technical and biological reproducibility of the assay using blood samples such as those intended for clinical use of the assay.
Blood samples from eight individuals including HD subjects and negative controls were divided into three equal sample sets. Each blood sample from each set was processed to obtain the BC and the samples were tested in triplicates along with two sets of QC samples on both the left and right edges of each plate. The three assays were performed over several days, by two operators, using replicate aliquots of the reagents.
Assay Specificity: We confirmed the assay’s specificity, using our GLP SOP protocol, as shown in Fig. 4 for human BC samples and the brain QC samples, and demonstrated a significant difference between HD and control subjects in regard to the mt Htt and mt/t Htt HTRF signal ratios.
Assay Linearity: Every BC sample was analyzed at 3 dilutions, corresponding to approximately 1:2, 1:4 and 1:8 final dilutions (Fig. 5). The control and HD brain lysates were used at a constant protein concentration (9 μg/mL) on every plate, a concentration that was determined experimentally (not shown) to be within the linear range for these samples and within the range of values obtained for BC samples in HTRF, dsDNA and protein measurements. The decrease in the signal for the control BC samples correlates most likely to the dilution of the matrix effect (the influence of other cell lysate components). We noticed a linear decrease of signal in HD samples with lower mt/t Htt. In samples with moderate to low Htt levels, the signal decreases in a linear fashion with the sample’s dilution, however, in samples with high Htt levels, the signal is usually inhibited at the lowest (1:2) dilution of the lysates, it peaks at the 1:4 dilution and then decreases, resulting in a linear dose response within 1:4 to 1:8 dilutions. The existence of a linear dose response curve in the HD BC samples provides the dynamic range necessary for the relative quantitation, in conjunction with QC criteria (dsDNA and protein concentration mentioned above). The inhibitory effect of concentrated lysates on the mt/t Htt HTRF signal is most likely due to excess of antigen (Htt) and/or matrix effect, as is often seen in such situations in ELISA and other immunoassays.
Relative Assay Sensitivity : Spiking of control BC sample with purified full-length mt Htt is not useful for absolute quantitation, since Htt will be affected by the aggregation of the mutant and normal protein in our testing conditions. Therefore, this is a semi-quantitative assay, in which we can only evaluate relative signals in HD samples compared to controls. To gain information on the recovery of full-length mt Htt and relative sensitivity of the measured mt Htt in our assay we spiked control brain lysates with increasing amounts of HD brain lysates to yield ratios as shown in Fig. 6 (ratios of CTRL: HD from 90:10 to 10:90).It is difficult to calculate the exact % recovery and implicit Limit of Quantitation (LoQ) sensitivity, since the control sample contributes as well to the signal, however a signal recovery of approximately 80-100% is apparent. The dynamic range resulting in a linear response occurs within 20-60% HD spiked into the control sample, again emphasizing the linear nature of the assay within this range.
To obtain an estimated numeric value for our HTRF measurements and to examine the effects of differing CAG lengths, we spiked purified N-terminus 548 aa Htt fragments: 548Q25, 548Q46, 548Q72, respectively into control BC. The final concentrations reached ranged between 300-0 ng/mL and the control BC final dilution was 1:4. We tested in parallel our QC set of control and HD brain lysates, as well as a known HD BC sample. Although these are just Htt fragments, and not the full-length molecule, it indicates the relative effect of the PolyQ length on the resulting HTRF signal ratios, as shown in Fig. 7.
As a result of spiking of the 548Qn fragments into a matrix similar to the clinical matrix, we observed that the dynamic range for mt Htt (lack of nonspecific signal in 548Q25) is within<100-16.1 ng/mL for 548Q46 and<100-16.1 ng/mL for 548Q72. The dynamic range for t Htt is 100-3.7 ng/mL for 548Q25, and 300-16.1 ng/mL for 548Q46 and 548Q72, a PolyQ dependent shift to higher concentrations. However, the dynamic range for mt/t Htt, as shown in Fig. 7is 33.4-16.1 ng/mL (the lower limit is imposed by the lower limits of quantitation for both mt Htt and t Htt), which is a narrow range.
Intra-Assay Variability : Fig. 8 is a representative example of the three assays performed for the assay qualification, demonstrating the reproducibility of the triplicate wells measured in the assay. The raw HTRF signals have a maximal Coefficient of Variation (CV) of <7%, which is low for a cell-based assay.
The Inter-Assay Variability : Fig. 9 illustrates the average of MW1/2B7, 2166/2B7 and MW1/2166 ratios from three experiments. The STDEV for the HD samples is higher than that for controls, and its CV < 28%, which is satisfactory for a cell-based assay. The Z’ for mt/t Htt (MW1-AF488/2166-d2) calculated from the corresponding values of two sets of control and HD brain lysates/plate and including each of the three experiments was Z’=0.6. This Z’, is therefore based on a total of six determinations/QC sample/plate/experiment x 3 experiments = a total of 18 replicates/QC sample, which is very satisfactory. The results from the intra and inter-assay variability analyses indicate high technical reproducibility.
Biological Reproducibility : We examined the biological reproducibility of the assay by measuring Htt in blood samples collected twice within a few weeks from the same individuals:
The results shown in Fig. 10 indicate excellent biologic consistency among the samples of the pairs tested, as confirmed by Pearson’s coefficient=0.88 and Lin’s concordance coefficient =0.87.
Following the assay qualification, we proceeded to write and issue a GLP compliant SOP. Using this SOP, we analyzed BC samples from 8 negative controls and 30 HD subjects (at different disease stages) in parallel to QC brain samples to demonstrate that this assay can differentiate between HD and controls based on the mt/t Htt levels.
Fig. 11: Soluble mt/t Htt Detection in HD and Control Buffy Coat Samples 8 control (CTRL) and 30 HD subjects (in different disease stages) BC samples from one clinical site and QC brain lysates were tested according to the SOP described (Material and Methods), in parallel to QC CTRL and HD brain lysates (not shown). Median ± standard deviation (STDEV) values of the control (CTRL) and the HD groups represented as ratios: mt Htt, t Htt and mt/t Htt are shown.
Measuring Clinical Buffy Coat Samples using SOP
: Using the GLP-SOP, we were able to confirm our original findings (Weiss et al [4] ) (Fig. 11) that the median mt/t Htt signals in BC samples from 30 HD subjects are higher than those measured in 8 control subjects. It is noteworthy that the samples tested were drawn from subjects at a single clinical site, ensuring minimal technical processing variability. The high standard deviation in the HD group reflects heterogeneity within this group which we hope to understand further by examining larger sample sets and clinical and laboratory covariates.
Discussion
Due to the importance of soluble mt Htt protein and its cleavage products in HD as both the causative agent and as a therapeutic target, its measurement in peripheral biological samples has been long sought. While conventional immunoassays, including ELISA, have been unable to reliably measure soluble Htt in biological samples, HTRF is able to detect it in peripheral blood samples, including cell lysates [5] . Furthermore, the multiplexing approach [4] enables simultaneous detection of several epitopes on the soluble Htt molecule, such that we can measure both an expanded polyglutamine epitope ( mt Htt) and an epitope distant from the polyglutamine tract ( t Htt) (Fig. 1) in human tissues. Weiss et al [5] have shown that this HTRF method is able to specifically distinguish HD from control samples. Since soluble Htt levels would be an essential marker for measuring the biological effects of therapeutics targeting Htt, we optimized the assay to comply with GLP requirements, paving the way for its use in clinical samples for its validation as a biomarker for HD [8] .
The optimization process we describe and quality control was enabled by the use of postmortem brain lysate samples from HD and control subjects. The selection criterion for optimal conditions for each variable was the highest ratio of HD/control when analyzing HTRF signal ratios of mt Htt and mt/t Htt. Based on this selection criterion we identified the 384 white plates with 15 μL/well as best suited for providing the strongest signal and enabling high throughput and the use of small sample volumes -15 μL (Fig. 2) with satisfactory intra-assay reproducibility. We determined that the reaction media needs to include the higher 200 mM KF, compared to 83 mM KF (Table 1) and that the reaction peaks between 2-3 hrs, while at 4 0 C (Fig. 3). We were able to identify Victor5X as a satisfactory plate reader for this assay, yielding optimal conditions with the following filters: excitation at 340 nm and emission at 665 nm, 615 nm and 510 nm (not shown). We proceeded with the assay qualification using clinical BC samples from HD and control samples complying with GLP standards. We determined that the optimized assay is specific in detecting mt Htt and higher mt/t Htt levels in HD compared to control samples (Fig. 4), sensitive in detecting at least 80% of HD’ soluble mt Htt spiked into control brain lysates (Fig. 6), and linear within the dynamic range identified in each sample by analyzing two fold dilutions (Fig. 5). The tendency of the Htt fragments to aggregate hampered the efforts to make this assay fully quantitative using Htt protein standards, however we attempted to obtain a relative quantitative sensitivity measurement for mt Htt that is normalized to the signal that estimates the presence of t Htt, thus mt/t Htt signals, using purified 548Qn Htt fragments (Weiss et al [5] [4] ) spiked into control BC samples (Fig. 7).The results using these Htt fragments indicate a narrow dynamic range for detecting mt/t Htt: 33.4-16.1 ng/mL for both the 548Q46 and 548Q72 Htt fragments. While it is impossible to know the absolute ranges we are indeed detecting by HTRF, serially diluting our BC samples, as we routinely do per our SOP, should enable us to capture the HTRF signals in spite of the narrow dynamic range. Based on the results presented with the spiked purified 548Qn Htt fragments, we are able to detect >3.7 ng/mL mt Htt and >16.1 ng/mL t Htt. These technical limitations could be circumvented only by purifying and spiking full-length soluble Htt into BC, thus finally determining the absolute quantitative soluble Htt measurements.
The measured HTRF ratios are compared between quality control brain lysate sets used at two ends of each plate with a Z’ ≥0.5 to ensure the quality of the analyses results. Analyses with this ratiometric approach offer the potential for informative use the HTRF assay in repeated measurements within subjects to evaluate the effects of disease progression or the effects of a potential therapy targeting Htt. To validate the measurement of soluble Htt as a clinical biomarker, we qualified the assay according to accepted GLP practices and found it to be technically reproducible with minimal (<7%) intra-assay (Fig. 8) and satisfactory (<28%, Fig. 9) inter-assay variability. The biological reproducibility between pairs of samples collected from subjects within a few weeks resulted in a coefficient of concordance >0.87 (Fig. 10), which is very promising for enabling longitudinal analyses. The results of our assay optimization led us to create a GLP compliant SOP, which we used to analyze clinical BC samples. The analysis of 30 HD samples in different disease stages and 8 control samples resulted in a specific discrimination of the two groups in regard to the mt Htt and the mt/t Htt ratios (Fig. 11). One possibility for the high standard deviation of mt Htt and the mt/t Htt ratios within the HD group is the heterogeneity of the subjects providing samples, who were selected without regard to disease severity or other clinical variables that could affect Htt levels. Variation in CAG length amongst the human samples could also play a role in variability between subjects, however since almost 90% of the expanded alleles from HD subjects are encompassed by a CAG repeat range of 40-48 (data from COHORT study, personal communication, Ray Dorsey), we don’t think that variability introduced by CAGn is likely to hamper the assay’s use as a biomarker. We plan to utilize sample sets from observational and therapeutic studies involving extensively characterized HD subjects to examine whether HTRF results are modulated by covariates such as age, gender, CAGn, disease stage, therapeutics and other biomarkers.
Acknowledgements:
Susan Maya, MGH, Keith Malarick, MGH, Caleb Dresser, MGH, Aleksey Kazantzev, PhD, MGH, Luisa Quinti PhD, MGH, David Oakes, PhD, University of Rochester, Mary McCarthy, Perkin Elmer, Greg Warner, PhD, Perkin Elmer, Alan Lopatin, Anna Sinsigalli, Cisbio, Diane Bowers, Cisbio, Christopher Beck,University of Rochester.
Funding Information:
Funding for this work was provided by NIH grants: NINDS P01NS58793 and U01NS071789 and by the Novartis Institute for Biomedical Research.
Competing interests:
The authors have declared that no financial interests exist.The 2B7-Tb antibody has been made available to us by Novartis for purchase through Cisbio.
Correspondence:
Correspondence should be addressed to Steven M. Hersch: [email protected]
| CC BY | no | 2022-01-13 05:57:35 | PLoS Curr. 2010 Dec 29; 2:RRN1205 | oa_package/69/f9/PMC3015780.tar.gz |
||||||
PMC3016240 | 21208471 | Background
Sex differences affect the physiological function of both gonadal and non-gonadal cellular systems. When gene expression was studied by microarray in a large number of mice, 55 to 72% of active genes showed sexual dimorphism in the liver, fat and muscle, and 13% of genes were sexually dimorphic in the brain [ 1 ]. These sex differences influence a variety of neural functions, both physiological and pathological.
One of the most robust sex differences is the estrogen-positive feedback, which signals the luteinizing hormone (LH) surge essential for ovulation. In post-pubertal females, rising levels of estradiol originating from developing ovarian follicles peak on proestrus, and induce the gonadotropin releasing hormone (GnRH) regulatory circuit to massively release GnRH, which stimulates estrogen primed gonadotrophs to release LH, resulting in ovulation and the formation of corpora lutea. Males, especially male rodents, do not exhibit this phenomenon. Their relatively constant levels of testosterone produce a negative feedback on the regulatory circuitry for GnRH release from the hypothalamus and gonadotropin release from the pituitary, an effect similar to that in females outside of proestrus. The inability of males to produce the estrogen positive feedback leading to a surge in LH has been attributed to the effects of androgen action on the central nervous system [ 2 - 7 ].
A mechanism for mediating estrogen positive feedback involves the synthesis of neuroprogesterone in the hypothalamus. Estradiol treatment of ovariectomized and adrenalectomized female rats increased hypothalamic progesterone levels and induced an LH surge [ 8 ]. Disruption of central (hypothalamic) progesterone synthesis blocked the LH surge in gonadally intact, cycling rats [ 9 ]. Interestingly, only adult females, which have an estrogen positive feedback mechanism, show an increase in hypothalamic progesterone in response to estradiol [ 8 , 10 ]. In other words, males and reproductive senescent females do not show an increase in hypothalamic progesterone synthesis. The cells responsible for the elevated neuroprogesterone levels in the hypothalamus after estradiol treatment are astrocytes [ 11 ].
In astrocytes from post-pubertal female rats, estradiol induces a rapid increase in free cytoplasmic calcium concentration ([Ca 2+ ] i ) that facilitates progesterone synthesis essential for positive estrogen feedback, the LH surge and ovulation in females [ 10 , 12 - 16 ]. We have not determined whether astrocytes derived from male rats similarly respond to estradiol stimulation by increasing [Ca 2+ ] i release and progesterone synthesis. The present experiments were performed to determine whether astrocytes derived from male and females respond differently to estradiol stimulation. In addition, Sry is expressed in the brain, and has been shown to directly influence the biochemical properties of the dopaminergic neurons of the nigrostriatal system and the specific motor behaviors they control [ 17 ]. To this end, the 'four core genotype' (FCG) mouse model, in which the sex chromosome complement is independent of gonadal phenotype [ 18 ], was used to determine whether sex differences are due to direct sex chromosome effects or to Sry transgene effects that determine gonadal differentiation and its dramatic influence on the sex steroid environment during early development. | Methods
All experimental procedures were approved by the Chancellor's Animal Research Committee at the University of California at Los Angeles.
Primary cell cultures
Primary hypothalamic astrocyte cultures pooled from two to six animals were obtained from 50-day-old adult Long-Evans rats (Charles River, Wilmington, MA, USA) and from 60-day-old adult mice (C57/Bl6 wild type and estrogen receptor-alpha (ERα) knockout (Jackson Laboratory, Bar Harbor, ME, USA) and C57BL/6J FCG mice (gift from Dr. Arthur Arnold, UCLA, Los Angeles, CA, USA)). FCG mice were obtained by breeding XX female mice with XY - Sry male mice, which possess a Y chromosome with the Sry gene deleted and a functional Sry transgene inserted onto an autosome. The presence of the Sry gene leads to differentiation of the indeterminate gonads into testes, and its absence results in formation of ovaries [ 19 , 20 ]. This cross generates four genotypes: XY - Sry gonadal males (XYM), XY - gonadal females (XYF), XX Sry gonadal males (XXM) and XX gonadal females (XXF) [ 21 ].
The hypothalamus was dissected with the following boundaries: rostral extent of the optic chiasm, rostral extent of the mammillary bodies, lateral edges of the tuber cinereum and the top of the third ventricle. Hypothalamic tissue was dissociated with 2.5% trypsin solution (Invitrogen, Eugene, OR, USA) and a fire polished glass Pasteur pipette. Cultures were grown in Dulbecco's modified Eagle's medium (DMEM)/F12 (Mediatech, Manassas, VA, USA) with 10% fetal bovine serum (FBS) (Hyclone, Logan, UT, USA) and 1% penicillin (10,000 IU/ml)-streptomycin (10,000 μg/ml) solution (PS) (Mediatech) at 37°C in 5% CO 2 . Astrocyte cultures were grown to confluency and purified from other glial cells [ 14 - 16 ] using a technique modified from McCarthy and de Vellis [ 22 ]. Sry expression in the astrocytes of FCG mice was analyzed and confirmed by reverse transcription (RT)-PCR.
Before the experiments, the DMEM/F12 medium with 10% FBS and 1% PS was removed, and primary astrocyte cultures were washed with Hanks' balanced salt solution (HBSS) (Mediatech), dissociated with a 2.5% trypsin solution and resuspended in DMEM/F12 medium with 10% FBS. Astrocytes were centrifuged for 3 minutes at 80 g , then the supernatant was removed and the pellet containing astrocytes resuspended. Astrocytes were counted, plated and incubated in DMEM/F12 medium with 10% FBS and 1% PS at 37°C in 5% CO 2 for 24 to 48 hours before Ca 2+ imaging and progesterone radioimmunoassay (RIA). For biotinylation, astrocytes were counted, plated and grown in flasks for 2 weeks before experimentation. Cultures were routinely checked for purity using immunocytochemistry for glial fibrillary acidic protein (Chemicon, Temecula, CA, USA) with Hoechst 3342 nuclear stain (Sigma-Aldrich, St. Louis, MO, USA). Cultures were determined to be > 95% pure astrocytes, as previously reported [ 14 - 16 ],
Intracellular Ca 2+ measurements
Astrocytes (5000) were plated onto 15 mm glass coverslips coated with 0.1 mg/ml poly-D lysine (Sigma-Aldrich) in 12-well culture plates and grown in DMEM/F12 medium with 10% FBS and 1% PS at 37°C in 5% CO 2 for 24 to 48 hours. The astrocytes were then starved of steroid for 18 hours by incubating in DMEM/F12 medium with 5% charcoal-stripped FBS at 37°C in 5% CO 2 . Before imaging, astrocytes were incubated for 45 minutes with HBSS and 4.5 μmol/l of the calcium indicator Fluo-4 AM (Invitrogen) dissolved in dimethyl sulfoxide (DMSO) and methanol. Glass coverslips were placed into a 50 mm chamber insert (Warner Instruments, Hamden, CT, USA) fixed into a 60 × 15 mm cell culture dish (Corning Inc., Corning, NY, USA) and placed into a quick exchange platform (QE-2; Warner Instruments) for imaging under a laser confocal microscope (Axioplan2-LSM 510 Meta; Zeiss, Thornwood, New York, NY, USA). Astrocytes were gravity perfused with HBSS, and media were removed by vacuum suction. Cyclodextrin encapsulated 17β-estradiol (1 nmol/l) (Sigma-Aldrich) was prepared in HBSS and used to induce [Ca 2+ ] i release. Controls were stimulated with HBSS only. For Fluo-4 AM imaging, a water immersion objective (IR-Achroplan 40X/0.80; Zeiss, Jena, Germany) was used with 488 nm laser excitation and emission monitored through a low-pass filter with a cutoff at 505 nm. The increase in Ca 2+ fluorescence (relative fluorescence units (RFU)) was calculated for each astrocyte as the difference between baseline fluorescence and peak response to drug stimulation.
Progesterone RIA
Approximately 500,000 astrocytes were plated into six-well culture plates and grown for 24 hours. Astrocytes were then starved of steroid in DMEM/F12 medium with 5% charcoal-stripped FBS for 18 hours before treatment with cyclodextrin encapsulated 17β-estradiol (1 or 100 nmol/l dissolved in HBSS) (Sigma-Aldrich) or HBSS with no steroids for 60 minutes at 37°C. After 1 hour of drug treatment, the supernatant for each well was collected and frozen at -20°C for up to 1 week before the RIA.
For the progesterone assay, samples were thawed, mixed with diethyl ether (Fisher Scientific, Fair Lawn, NJ, USA) and then mixed by vortex for 2 minutes. To freeze the aqueous layer, samples were placed into a methanol and dry ice bath. The upper ether layer was decanted into a separate tube and the ether was allowed to evaporate overnight. The extract was reconstituted in isooctane (Mallinkrodt Baker, Phillipsburg, NJ, USA) and a diatomaceous earth column (Celite columnTM; Celite Corp., Lompoc, CA, USA) with ethylene glycol as the stationary phase was used to isolate the progesterone. Progesterone was then eluted off the column using 4 ml of isooctane. Standards and samples (100 μl) were incubated with rabbit polyclonal antibody against progesterone (Sigma-Aldrich) for 30 minutes at 37°C. Tritium radiolabeled progesterone (2000 counts/minute/ml) was added and incubated for an additional 60 minutes at 37°C. Standards and samples were cooled to 4°C, and a 0.05% charcoal dextran solution (Sigma-Aldrich) added to remove all unbound progesterone. The mixture was centrifuged at 1500 × g for 15 minutes at 4°C. The supernatant was then collected for chromatographic detection of progesterone. All samples were run in duplicate, and sample progesterone concentrations determined by extrapolation from a curve determined from the progesterone standards.
Surface biotinylation
Primary astrocyte cultures were starved of steroids in DMEM/F12 medium with 5% charcoal-stripped FBS 12 hours before treatment with vehicle or 1 nmol/l of 17β-estradiol (Sigma-Aldrich) for 30 minutes at 37°C. Cells in each flask were washed three times with ice cold phosphate buffered saline (PBS) and incubated with freshly prepared biotin (0.5 mg/ml) (EZ-Link Sulfo-NHS-LC-Biotin; Pierce Biotechnology Inc., Rockford, IL, USA) in PBS for 30 minutes at 4°C with gentle agitation. Excess biotin reagent was quenched by rinsing the cells three times with ice cold glycine buffer (50 mmol/l glycine in PBS). Cells were scraped into 10 ml of PBS solution, transferred into a 50 ml conical tube and centrifuged at 500 × g for 3 minutes. The pellet was washed twice with ice-cold PBS and resuspended in 200 ml radioimmunoprecipitation assay (RIPA) lysis buffer containing the protease inhibitors 1 mmol/l phenylmethylsulfonyl fluoride, 1 mmol/l EDTA, 1 μg/ml pepstatin, 1 μg/ml leupeptin, 1 μg/ml aprotinin and 1 mmol/l sodium orthovanadate (all from Santa Cruz Biotechnology, Santa Cruz, CA, USA). The cells were homogenized by passing them through a 25-gauge needle, and the cell extract was centrifuged at 10,000 × g for 2 minutes at 4°C. The protein concentration of the supernatant was determined using the Bradford Assay (Bio-Rad, Hercules, CA, USA). Samples with equal protein concentration were added to a washed immobilized gel (NeutrAvidinTMGel; Pierce Biotechnology Inc.) for 2 hours at room temperature and then centrifuged at 1,000 × g for 1 minute. The beads were washed four times with 1 ml of RIPA buffer (Santa Cruz Biotechnology) containing the protease inhibitors previously mentioned. The bound proteins were eluted with SDS-PAGE sample buffer supplemented with 50 mmol/l dithiothreitol (DTT) for 1 hour at 37°C.
Western blotting
Samples were separated in a 10% Tris-HCl gel (Ready Gel; Bio-Rad) and transferred to polyvinylidene fluoride membranes (GE Healthcare, Piscataway, NJ, USA). ERα was detected with primary rabbit polyclonal antibody (1:1000) (C1355; Upstate Biotechnology, Inc., Lake Placid, NY, USA). β-actin was used as a loading control and detected by rabbit polyclonal antibody (1:5000) (Abcam, Cambridge, MA, USA). A secondary donkey anti-rabbit IgG (H+L) antibody (1:5000) (Jackson ImmunoResearch, West Grove, PA, USA) and an anti-biotin horseradish peroxidase-linked antibody (1:1000) (Cell Signaling Technology, Danvers, MA, USA) were then used (1.5 hour incubation). To estimate the molecular weight, samples were run alongside a biotinylated protein ladder (Cell Signaling Technology). Immunoreactive bands were visualized using an enhanced chemoluminesence (ECL) kit and ECL hyperfilm (GE Healthcare). Routine exposures varied from 0.5 to 2 minutes. The optical density of each immunoreactive band was determined. For each sample, immunoreactive ERα was normalized with β-actin to obtain the percentage of ERα protein to β-actin protein ratio (% relative ratio).
RT-PCR
Total RNA was extracted from several primary cultures of astrocytes from Long-Evans post-pubertal rats using Trizol reagent (Invitrogen), following the manufacturer's recommended protocol. To prevent DNA contamination, RNA was treated with DNase I (Ambion, Austin, TX, USA) at 37°C for 30 minutes followed by inactivation with DNase inactivation reagent (Ambion). Total RNA quality and concentration were assessed using a spectrophotometer (NanoDrop 1000; Thermo Scientific, Wilmington, DE, USA). RT was then performed using 1 μg of total RNA to synthesize single-stranded cDNA in a 20 μl reaction with a reverse transcriptase (SuperScript III; Invitrogen) and a combination of random hexamers and oligo (dT) 20 primers, following the manufacturer's protocol. Briefly, RT was performed at 50°C for 50 minutes, then the reaction terminated at 85°C for 5 minutes and the RNA destroyed with 1 μl of RNase H at 37°C for 20 minutes. cDNAs were subjected to PCR with primers specific to rat Sry (Fwd: 5-GCAGCGTGAAGTTGCCTCAAC-3 and Rev: 5-TGCAGCTCTAGCCCAGTCCTG-3) in an RT-PCR system (Mx3000p Real-Time PCR System; Stratagene, Santa Clara, CA, USA). PCR conditions used for amplification were as follows: initial denaturation at 94°C for 10 minutes, 35 cycles of denaturation at 94°C for 45 seconds, annealing at 60°C for 45 seconds and elongation at 72°C for 1 minute, with a final extension at 72°C for 7 minutes. Amplified products were separated by electrophoresis in a 2% agarose gel with ethidium bromide and visualized with ultraviolet light. Gel images were captured digitally to confirm product size and the absence of non-specific products. Negative controls (no cDNA) were included in every PCR run.
Statistical analysis
Data are presented as means ± standard error (SEM) in RFU, pg/ml or % relative ratio. Statistical comparisons were made using one-way analysis of variance (ANOVA) with Student-Newman-Keuls post hoc test when comparing means across at least three independent groups. For the FCG mice, mean comparisons and contrasts under the ANOVA model were made using the Tukey-Fisher least significant difference (LSD) criterion. Statistical calculations were carried out using SAS (version 9.2; SAS Institute, Cary, NC, USA) and GraphPad Prism (version 5; GraphPad Software, La Jolla, CA, USA) software programs. P < 0.05 was considered significant. | Results
Sex differences of [Ca 2+ ] i release in response to estradiol stimulation
Calcium imaging was used to confirm the sexual differentiation of hypothalamic astrocyte function in adult mice. Although 1 nmol/l estradiol induced a significant ( P < 0.05) [Ca 2+ ] i response in both female and male wild type astrocytes (ΔF Ca 2+ = 630 ± 13 RFU (n = 31) and ΔF Ca 2+ = 340 ± 17 RFU (n = 24), respectively) compared with control (ΔF Ca 2+ = 135 ± 6 RFU (n = 21) and ΔF Ca 2+ = 152 ± 7 RFU (n = 24), respectively), the estradiol-induced [Ca 2+ ] i release in male astrocytes was significantly( P < 0.05) smaller than the [Ca 2+ ] i release in female astrocytes (Figure 1 ). We previously reported that the estradiol-induced [Ca 2+ ] i response seen in female wild type astrocytes was abolished in female hypothalamic astrocytes from estrogen receptor-α knockout (ERKO) mice [ 12 ].
Sex differences in astrocytic progesterone synthesis to estradiol stimulation
We previously demonstrated that 1 nmol/l estradiol significantly increased progesterone synthesis in primary cultures of adult female hypothalamic astrocytes in rats [ 12 ]. Using the same culture conditions in adult female hypothalamic astrocytes obtained from mice, we now confirm that 1 and 100 nmol/l estradiol stimulates significant ( P < 0.05) progesterone synthesis (69.4 ± 2.0 pg/ml (n = 4) and 99.9 ± 13.0 pg/ml (n = 4), respectively) compared with control (21.6 ± 3.3 pg/ml (n = 4)) (Figure 2 ). Furthermore, 100 nmol/l estradiol stimulated greater progesterone synthesis compared with 1 nmol/l estradiol ( P < 0.05) (Figure 2 ). However, hypothalamic astrocytes from adult male mice did not synthesize progesterone above control levels (15.8 ± 0.8 pg/ml (n = 4)) when stimulated with estradiol at 1 or 100 nmol/l (22.2 ± 1.7 pg/ml (n = 4; P > 0.05 versus control) and 14.3 ± 2.4 pg/ml (n = 4; P > 0.05 versus control), respectively) (Figure 2 ). Adult male rat hypothalamic astrocytes similarly failed to synthesize progesterone when exposed to 1 or 100 nmol/l estradiol (20.0 ± 0.6 pg/ml (n = 4) and 18.9 ± 2.4 pg/ml (n = 4), respectively) compared with control (12.7 ± 2.9 pg/ml (n = 4; P > 0.05)). These results demonstrate sexual differentiation of hypothalamic astrocyte function in response to estradiol stimulation.
Sex differences do not result from the sex chromosome complement
Using FCG mice, we compared genetic sex chromosome effects versus Sry transgene effects. Baseline progesterone synthesis by control astrocytes from XYM mice (0.0 ± 5.3 pg/ml (n = 4)) was significantly lower than control astrocytes from XXF (21.8 ± 5.3 pg/ml (n = 4; P < 0.05)), XYF (16.0 ± 5.3 pg/ml (n = 4; P < 0.05)) and XXM mice (17.2 ± 5.3 pg/ml (n = 4; P < 0.05)), all of which were similar to each other ( P > 0.05) (Figure 3 ). Estradiol (1 nmol/l) significantly increased progesterone synthesis in both types of gonadal female (XXF and XYF) astrocytes (51.9 ± 5.3 pg/ml (n = 4; P < 0.05 versus control) and 47.4 ± 5.3 pg/ml (n = 4; P < 0.05 versus control), respectively), but failed to increased progesterone synthesis in gonadal male (XXM and XYM) astrocytes (16.6 ± 5.3 pg/ml (n = 4; P > 0.05 versus control) and 3.0 ± 5.3 pg/ml (n = 4; P > 0.05 versus control), respectively) (Figure 3 ). Comparison of progesterone changes after estradiol (1 nmol/l) versus control revealed a similar increase in progesterone synthesis in XXF and XYF astrocytes, whereas there was a lack of progesterone increase in both XXM and XYM astrocytes ( P < 0.05), suggesting a steroid-induced sex effect due to early differential gonadal (testes versus ovaries) hormone secretion.
Sex differences in astrocytic mERα trafficking in response to estradiol
We previously used surface biotinylation to demonstrate the presence of two ERα immunoreactive bands (66 kDa and 52 kDa) in the cell membrane of female wild type hypothalamic astrocytes, which were not present in female ERKO mouse astrocytes [ 23 ]. Immunoreactive mERα is transiently increased by estradiol exposure, reaching maximum levels after 30 minutes [ 23 ]. As in the previous study, the major ERα immunoreactive band was at 52 kDa, and was thus used for quantification. Basal levels of mERα were similar in both female and male wild type hypothalamic mouse astrocytes (27 ± 7% relative ratio (n = 3) and 22 ± 5% relative ratio (n = 3), respectively; P > 0.05) (Figure 4 ). Stimulation with 1 nmol/l estradiol for 30 minutes significantly increased mERα levels in female astrocytes (41 ± 10% relative ratio (n = 3; P < 0.05 versus control)). An equimolar concentration of estradiol did not change mERα levels in male astrocytes (28 ± 8% relative ratio (n = 3; P > 0.05 versus control)) (Figure 4 ). These results demonstrate a sex difference in the regulation of mERα trafficking in response to estradiol exposure in hypothalamic astrocytes. | Discussion
Although both male and female rodents have a well developed negative feedback mechanism regulating the release of GnRH and LH, one of the most robust sexually differentiated physiological responses is estrogen positive feedback, which induces the surge release of LH in response to estradiol stimulation. This phenomenon of estrogen positive feedback is a hallmark of various female animal species. For rodents, once the ability to produce estrogen positive feedback is lost during development, the deficit is permanent. In primates, including humans, many years of continuous estrogen exposure in males can result in an estrogen positive feedback response, although it is much attenuated [ 24 ]. According to the epigenetic theory of sexual differentiation of the brain, the sex difference in estrogen positive feedback is said to arise from the action of estradiol (aromatized from testosterone) during organization of the neural circuit(s) controlling GnRH neurons. Several mechanisms have been proposed to account for this differentiation, including the lack of estrogen-induced synaptic plasticity in the male arcuate nucleus [ 25 ] and an attenuated distribution of kisspeptin neurons in males [ 26 ]. Various structural sex differences that result from perinatal exposure to estradiol have been identified. In terms of regulating GnRH, males have greater postnatal apoptosis in the developing anteroventral periventricular nucleus (AVPV), a region crucial for estrogen positive feedback in females [ 27 - 29 ]. Although it is not clear whether such a sex difference in apoptosis is an important mechanism for elimination of estrogen positive feedback, it does support a role for postnatal sex steroids in organizing brain mechanisms involved in reproduction [ 30 ].
Over the past several years we have been investigating the role of neuroprogesterone in regulating estrogen positive feedback [ 31 ]. Experimental evidence is consistent with the hypothesis that synthesis of progesterone in the hypothalamus is needed for an LH surge and the transition from proestrus to estrus [ 8 , 10 ]. Significantly, estrogen treatment increases progesterone levels in the female but not male hypothalamus [ 8 , 31 ]. The cells responsible for neuroprogesterone production are hypothalamic astrocytes [ 11 ]. It has been reported that astrocytes are sexually differentiated by the neonatal hormonal milieu, especially estradiol, which profoundly influences their morphology and function [ 32 , 33 ]. Female astrocytes express ERs that are targeted to the cell membrane and directly regulated by estradiol levels [ 13 , 23 ]. Estradiol stimulated mERα transactivates the metabotropic glutamate receptor (mGluR) type 1a, which activates a phospholipase C-inositol trisphosphate (PLC-IP 3 ) cascade, resulting in the release of intracellular Ca 2+ from the smooth endoplasmic reticulum via an IP 3 receptor dependent mechanism [ 13 , 15 ]. The robust increase of [Ca 2+ ] i release stimulates neuroprogesterone synthesis in hypothalamic astrocytes, which is required for the LH surge [ 11 ]. Alternatively, astrocytes can potentially modulate GnRH neurons in the arcuate nucleus directly through ensheathment or unensheathment of synaptic connections or restriction of GnRH nerve terminal access to the portal vasculature, and indirectly through release of growth factors and prostagladins [ 34 , 35 ].
In the present study, we investigated astrocytic sex differences by examining the response to estradiol of adult hypothalamic astrocytes derived from male or female rats and mice. First, both male and female astrocytes have mERα that can be labeled using surface biotinylation [ 23 ]. Estradiol exposure increased the amount of ERα at the cell membrane in adult female astrocytes, but did not increase levels of mERα in male astrocytes. Second, male astrocytes did have an estradiol-induced [Ca 2+ ] i response, which was significantly attenuated compared with that of female astrocytes. These results are consistent with our previous observations that maximum [Ca 2+ ] i release is correlated with elevations in estradiol-induced mERα levels [ 23 ]. The importance of mERα was emphasized by the lack of estradiol-induced [Ca 2+ ] i release in female ERKO astrocytes [ 12 ]. This stimulation of [Ca 2+ ] i release involved mERα transactivation of mGluR1a, a G protein coupled receptor [ 15 ]. Furthermore, the estradiol-induced progesterone synthesis in adult female astrocytes similarly required this mERα-mGluR1a interaction [ 12 ]. Third, only astrocytes derived from adult female rats [ 10 , 12 , 14 ] and mice (present results) had an increase in progesterone synthesis, consistent with previous observations that only adult female rodents have increased levels of progesterone in the hypothalamus before the LH surge [ 8 ]. These results support the hypothesis that estrogen positive feedback requires a robust [Ca 2+ ] i release that triggers progesterone synthesis in the hypothalamus. The data also suggest that although a rise in [Ca 2+ ] i is necessary for progesterone synthesis, there appears to be a crucial concentration required, as male astrocytes have an attenuated [Ca 2+ ] i response that was unable to facilitate progesterone synthesis in the present study, consistent with previous reports in neonatal astrocytes [ 14 ] and in post-pubertal astrocytes, in which 0.1 nmol/l estradiol stimulated [Ca 2+ ] i release, but not progesterone synthesis [ 12 ].
Biological differences between males and females can result genetically from direct sex chromosome differences, developmentally through differential exposure to sex steroids during developmental 'organization', or functionally from acute 'activational' effects of gonadal steroids operating at many life stages, which can be controlled through gonadectomy [ 36 ]. Perinatal gonadal hormone secretions have been shown to have powerful and permanent actions on physiology, including pituitary function, gene expression in the brain and sexual behavior [ 37 - 40 ]. In spite of these epigenetic effects, several chromosomal dependent sex differences have been demonstrated in the brain. Specifically, the FCG mice model has demonstrated purely chromosomal XX versus XY differences in behaviors, including aggression, parenting, habit formation, nociception and social interactions [ 21 ]. For example, a chromosomal sex effect was demonstrated for vasopressin innervation of the septum [ 41 , 42 ]. At embryonic day 13, mesencephalic neurons express tyrosine hydroxylase, which is earlier than any sex steroid actions [ 43 ]. More importantly, neurons in the adult male substantia nigra were shown to express Sry , which maintained the expression of tyrosine hydroxylase, the rate limiting enzyme of catecholamine (dopamine) synthesis [ 17 ]. A reduction in Sry gene expression led to motor deficits in male rats, suggesting a function for Sry in the maintenance of dopamine neurons needed for motoric behaviors regulated by the nigrostriatal pathway that is affected in Parkinson's disease. These studies suggest that Sry directly affects the biochemical properties of the dopaminergic neurons of the nigrostriatal system and the specific motor behaviors they control.
Both male and female astrocytes have mERα, respond to estradiol stimulation by elevating [Ca 2+ ] i levels and synthesize progesterone. However, only in female astrocytes can estradiol increase the synthesis of progesterone (4- to 6.5-fold), a critical step in estrogen positive feedback [ 10 ]. Although we demonstrated a stark difference between male and female astrocytic response to estradiol, it was not clear whether this cellular differentiation was due to differences in the sex chromosome complement or to the presence of the Sry transgene with its influence on gonadal development and early sex steroid environment. Astrocytes from FCG mice were used to specifically differentiate the effects of sex chromosomes versus those of the Sry transgene. Animals with ovaries (XXF and XYF) had astrocytes in which estradiol facilitated progesterone synthesis, regardless of whether they had one or two X chromosomes. Conversely, mice with testes (XYM and XXM) were unresponsive to estradiol and did not increase progesterone synthesis. These results suggest a Sry transgene effect and not a sex chromosome effect on hypothalamic astrocyte response to estradiol. The effects from the Sry transgene could be due to direct effects of the Sry gene itself or its influence on gonadal differentiation and the sex steroid environment during early development. Interestingly, XYM from FCG mice synthesized little or no progesterone. This may reveal a potential chromosomal effect. However, male wild type astrocytes, without Sry translocation to an autosome, synthesized basal progesterone levels similar to female wild type astrocytes. Therefore, this difference could potentially be caused by the deletion and transgenic insertion of Sry , resulting in the inactivation of surrounding gene(s), positional effects or differential expression of the Sry transgene. Differences between wild type XY males and FCG XYM have been previously reported for mounting behavior, social exploration and concentration of tyrosine hydroxylase-immunoreactive neurons within the AVPV [ 41 ]. Unfortunately, the steroid profile of XYM in FCG mice has not yet been characterized and will require further experimentation. | Conclusions
Although there may have been a hint of a chromosomal sex difference in the basal level of progesterone synthesis, the overwhelming effect appeared to be from the Sry transgene, probably from its dramatic influence on gonadal differentiation and the steroid environment during early development. Wild type male astrocytes expressed Sry , and were sexually differentiated from wild type female astrocytes in terms of their [Ca 2+ ] i and progesterone responses to estradiol. These results are consistent with the stark sexual differentiation of estrogen positive feedback, which is dependent on the postnatal gonadal steroid environment. A potential mechanism for this sex difference in estradiol response was the trafficking of mERα in females but not in males. The estradiol-induced transient increase in mERα levels has been correlated with the robust [Ca 2+ ] i release necessary for progesterone synthesis. These results suggest that cell signaling in hypothalamic astrocytes is sexually differentiated, mainly as a result of postnatal gonadal steroid exposure, which may also mask the influence of possibly minor chromosomal effects. | Background
Reproductive functions controlled by the hypothalamus are highly sexually differentiated. One of the most dramatic differences involves estrogen positive feedback, which leads to ovulation. A crucial feature of this positive feedback is the ability of estradiol to facilitate progesterone synthesis in female hypothalamic astrocytes. Conversely, estradiol fails to elevate hypothalamic progesterone levels in male rodents, which lack the estrogen positive feedback-induced luteinizing hormone (LH) surge. To determine whether hypothalamic astrocytes are sexually differentiated, we examined the cellular responses of female and male astrocytes to estradiol stimulation.
Methods
Primary adult hypothalamic astrocyte cultures were established from wild type rats and mice, estrogen receptor-α knockout (ERKO) mice, and four core genotype (FCG) mice, with the sex determining region of the Y chromosome ( Sry ) deleted and inserted into an autosome. Astrocytes were analyzed for Sry expression with reverse transcription PCR. Responses to estradiol stimulation were tested by measuring free cytoplasmic calcium concentration ([Ca 2+ ] i ) with fluo-4 AM, and progesterone synthesis with column chromatography and radioimmunoassay. Membrane estrogen receptor-α (mERα) levels were examined using surface biotinylation and western blotting.
Results
Estradiol stimulated both [Ca 2+ ] i release and progesterone synthesis in hypothalamic astrocytes from adult female mice. Male astrocytes had a significantly elevated [Ca 2+ ] i response but it was significantly lower than in females, and progesterone synthesis was not enhanced. Surface biotinylation demonstrated mERα in both female and male astrocytes, but only in female astrocytes did estradiol treatment increase insertion of the receptor into the membrane, a necessary step for maximal [Ca 2+ ] i release. Regardless of the chromosomal sex, estradiol facilitated progesterone synthesis in astrocytes from mice with ovaries (XX and XY - ), but not in mice with testes (XY - Sry and XX Sry ).
Conclusions
Astrocytes are sexually differentiated, and in adulthood reflect the actions of sex steroids during development. The response of hypothalamic astrocytes to estradiol stimulation was determined by the presence or absence of ovaries, regardless of chromosomal sex. The trafficking of mERα in female, but not male, astrocytes further suggests that cell signaling mechanisms are sexually differentiated. | Competing interests
The authors declare that they have no competing interests.
Authors' contributions
JK carried out the intracellular Ca 2+ experiments, participated in the design and coordination of the study, performed the statistical analysis and drafted the manuscript. NH performed the primary cell culturing and carried out the progesterone radioimmunoassay experiments. GB performed the primary cell culturing and carried out the surface biotinylation and western blotting experiments. PD carried out the reverse transcription polymerase chain reaction. JC performed the primary cell culturing and participated in the design of the study. PM conceived of the study, participated in its design and coordination and drafted the manuscript. All authors read and approved the final manuscript. | Acknowledgements
This work was supported by the National Institutes of Health (NIH) Grant HD042635 and the Reproductive Scientist Development Program through NIH Grant HD00849-23 and Bayer HealthCare Pharmaceuticals. We thank Dr. Arthur Arnold at the University of California, Los Angeles, CA for kindly providing the FCG mice. | CC BY | no | 2022-01-12 15:21:43 | Biol Sex Differ. 2010 Nov 22; 1:7 | oa_package/85/43/PMC3016240.tar.gz |
PMC3016241 | 21208470 | Introduction
Sex differences in the brain are widespread but of variable magnitude. Differences in the size of specific structures or subnuclei are well characterized, as are sex differences in the density and number of excitatory and inhibitory synapses within particular brain regions. Many sex differences in the brain are induced during a perinatal sensitive period by oestradiol following its central aromatization from testicularly derived androgen precursors (for review see [ 1 ]). The most robust neuroanatomical sex differences are found in the brain areas directly involved in reproduction, such as the preoptic area, hypothalamus and spinal cord [ 2 - 6 ]. The impact of steroids on these brain regions across the life-span is codified in the Organizational/Activational Hypothesis of sexual differentiation first postulated over 50 years ago [ 7 ]. Gonadal steroids, including oestradiol, also potently regulate synaptic profiles in adult brain regions not directly associated with reproduction, such as the hippocampus [ 8 - 10 ] and the amygdala [ 11 ].
The hippocampus subserves important behavioural and physiological functions that are influenced by sex. There are subtle, identifiable, sex differences in hippocampal volume [ 12 ] and the morphology of hippocampal cells [ 13 - 15 ]. There are also subtle sex differences in hippocampal associated behaviours such as spatial learning strategies, stress responsivity, and the long-term impact of negative early life events [ 16 - 20 ]. There are also subtle but complex sex differences in the developing hippocampus that are associated with parameters responsive to oestradiol, including calcium entry in response to depolarizing GABA [ 21 - 24 ] and cell genesis [ 25 ].
The hippocampus is comprised of subregions (for example, dentate gyrus (DG), CA1 and CA3) and each subregion contains distinctive cell types characterized by distinctive rates of maturation. The pyramidal cells, the principle cells of Ammon's horn, are largely formed before birth [ 26 - 29 ]. In contrast, granule cells, which comprise the major cell type found in the DG, are predominantly born during the first 2 weeks of postnatal life [ 28 , 30 ]. The development of granule cells in the molecular layer of the DG is faster in males than females [ 30 ], suggesting a sex difference in cell proliferation.
Hormonally-mediated sex differences in cell death are central to the sexual differentiation of many brain areas (for review see [ 31 ]), but less is known about sex differences in cell genesis, particularly during development. However, we recently reported that neonatal males have more new cells as indicated by 5-bromo-2-deoxyuridine (BrdU) in the DG and CA1, relative to females, with no corresponding sex difference in the number of pyknotic cells [ 25 ]. Both oestradiol and testosterone treatment increase the number of BrdU+ cells in females to the level found in males. However, whether endogenous steroids mediate the sex difference in cell proliferation is unknown. It is also unknown whether developmental steroid effects on cell proliferation are transient or organized and, as a result, endure across the lifespan. Here we report that endogenous oestradiol can generate higher rates of cell proliferation in males but plays no role in females. Consistent with previous results, treatment of females with exogenous oestradiol increases cell proliferation both acutely and 3 weeks later, suggesting that oestradiol might have an organizational role in neurogenesis. The sex difference in neurogenesis reported here may reflect a critical time period in hippocampal development that contributes to sexual dimorphisms in the adult hippocampus. | Methods
Animals
Newborn male and female Sprague-Dawley rats were obtained from breeder females at the University of Maryland School of Medicine. The day of birth was defined as postnatal day 0 (PN0). The animals were housed under a 12:12 h light: dark cycle, with food and water freely available. All procedures were approved by the University of Maryland School of Medicine Institutional Animal Care and Use Committee and followed National Institutes of Health Guidelines.
Hormonal treatment of animals
On PN0, male and female rat pups were randomly distributed into different experimental groups and marked for identification by a subcutaneous (sc) ink injection in either the front or hind paws and injected sc with either oestradiol benzoate (100 μg ⁄ 0.1 mL in sesame oil), formestane (100 μg ⁄ 0.1 mL in sesame oil), tamoxifen (100 μg ⁄ 0.1 mL in sesame oil) or vehicle (sesame oil; see Figure 1 for the experimental timeline). The high dose of oestradiol is required in order to overcome the sequestering capacity of alpha-fetoprotein in the neonatal bloodstream [ 2 ] and is a dose routinely used by this laboratory to induce sexual differentiation of reproductive parameters [ 32 ]. Moreover, benzoate moiety does not alter the oestradiol's biological activity but does prolong its bioavailability. The dose of formestane has been previously used in this laboratory to decrease endogenous oestradiol in the neonatal brain to near undetectable levels [ 2 ]. The dose of tamoxifen was based on previously published literature using this drug in vivo [ 33 ]. Both formestane and tamoxifen readily cross the blood-brain-barrier [ 34 , 35 ]. Two hours after the steroid injection, all pups were injected intraperitoneally (ip) with BrdU on PN0 and PN1 (0.1 mL 0.9% sterile saline containing 100 mg ⁄ kg of BrdU), except for the experiments in which animals were injected with BrdU only on PN0. The rationale for administering BrdU 2 hours post hormone injection was based on previously published work in our laboratory in which a hormonal effect on BrdU+ cells was observed [ 25 ]. The injection sites were sealed with cyanoacrylate Vetbond Surgical Adhesive (3 M Animal Care Product, MN, USA).
Tissue collection
Pups were deeply anaesthetized with sodium pentobarbital Fatal Plus (250 mg/kg) and transcardially perfused with 0.9% saline until there was no blood trace and then fixed with 4% paraformaldehyde (PFA). Brains were removed and fixed for 24 h in 4% PFA, followed by 48 h in 30% sucrose in PFA before being sectioned coronally on a cryostat, with each slice being 45 μm thick. Slices used for quantification were separated from each other along the rostral/caudal axis so that two contiguous sections were not analysed in the same animal. Sections were collected in series such that each animal generated 4-5 series of sections with 6-8 sections per series obtained from each brain.
Immunohistochemistry
Free-floating tissue sections were rinsed with 0.1 M phosphate buffered saline (PBS) and then incubated with 3% hydrogen peroxide in PBS for 30 min. For BrdU immunohistochemistry, tissue sections were further incubated with 2N hydrogen chloride (HCl) for 60 min at 37°C in order to denature the DNA. After HCl incubation, the sections were rinsed with Borate buffer solution followed by PBS rinses. Finally, sections were incubated with 5% goat serum in PBS with 0.4% Triton X-100 (PBS-T) for 60 min followed by incubation in PBS-T with a monoclonal antibody against BrdU (1: 10,000, Caltag Laboratories, CA, USA) at room temperature (RT) for 60 min, then for 24 h at 4°C. Sections were rinsed in PBS and incubated with biotinylated anti-mouse secondary in PBS-T (1: 1000, Vector, Auckland, New Zealand) rinsed with PBS and incubated in Vectastain Elite ABC reagents (1: 1000, Vector). BrdU-positive cells were detected with diaminobenzidine (DAB) as chromogen, creating a dark brown colour in BrdU-positive nuclei.
For Ki-67 immunohistochemistry, tissue sections were treated as stated above with the omission of the HCl incubation step. Tissue sections were incubated with a polyclonal antibody against Ki-67 in PBS-T (1: 5000, Millipore, Darmstadt, Germany) for 60 min at RT and then 4°C for 48 h followed by biotinylated anti-rabbit secondary antibody (1: 1000). The Ki-67 positive cells were detected with DAB.
Double-label fluorescence immunohistochemistry
Fluorescent immunohistochemistry was used to quantify the extent of colocalization of BrdU and the neuronal specific marker, NeuN. Free-floating tissue sections were rinsed with 0.1 M PBS, incubated with 3% hydrogen peroxide in PBS for 30 min, rinsed and incubated with 2 N HCl for 60 min at 37°C. After HCl incubation, sections were rinsed with Borate buffer solution followed by PBS rinses, incubated with 0.3 M glycine in 0.4% Triton X-100 (PBS-T) for 30 min, and rinsed. Sections were co-incubated with primary antibodies against anti-rat BrdU (1:500, Abcam, Cambridge, UK) and anti-mouse NeuN (1: 500, Millipore) in PBS-T, which contained 10% bovine serum albumin (BSA), for 60 min at RT and overnight at 4°C then rinsed in PBS and incubated with biotinylated anti-rat secondary (1:300, Vector) in PBS-T for 90 min followed by co-incubation with streptavidin Alexa 488 (1:1000; Invitrogen, CA, USA) and anti-mouse Alexa 568 (1:500, Invitrogen) in PBS-T for 60 min in the dark. Rinsed tissue sections were mounted onto gelatin-subbed slides and coverslipped in the dark using Vector Vectashield.
Data analysis
Each subregion of the hippocampus (CA1, CA3 and DG) was analysed using a Nikon Eclipse E600 microscope and the Neurolucida Software System (Microbrightfield, VT, USA). Cell density estimates for each subregion were determined by counting the immunopositive BrdU+, and Ki-67+, cells from the pyramidal layer of CA1 and CA3 as well as granule cell layer of the hippocampus. Cell quantification consisted of using six gridded counting frames that were within the specific subregion of the hippocampus. Each individual counting frame measured 100 × 100 μm. For all three subregions, we calculated the number of immunopositive cells that were inside the counting frames, in both the left and right hemisphere of each section. This procedure was done in four to five sections throughout the rostral hippocampus of each animal. The total number of immunopositive cells across all counting frames was averaged to give one mean value per animal per subregion. All quantifications were performed using a 40× objective. Furthermore, our cell quantification was performed with adherence to appropriate stereological principles, which included but was not limited to: (1) the presence of a nucleus; (2) immunopositive cells were of homogeneously shape; and (3) labelling intensities of immunopositive cells were distinguishable from background staining. Figure 2A-D is a photomicrograph that includes cells considered BrdU+ and Ki67+ by both investigators conducting the analyses. Lastly, the sex and hormonal condition of the animals was unknown to the investigators conducting the analyses.
For immunofluorescence, each subregion of the hippocampus was analysed using a Nikon Eclipse 80i grid confocal microscope equipped with an OptiGrid structured light source. Volocity Grid Confocal software (Improvision, Warwickshire, UK) was used to perform the analysis of Z-stacks measuring 10-12 μm collected in two channels at 0.5 μm intervals using a 100× oil objective. BrdU+ cells were selected in each subregion and the number of cells co-expressing the neuronal marker NeuN+ and BrdU+ were quantified. The immunopositive cells were counted bilaterally within each subregion, in four sections (45 μm thick), throughout the hippocampus from each animal using an 80 × 100 μm counting frame.
Statistics
The detection of differences in the mean number of cells per region per animal between groups was determined using a three-factor ANOVA with sex, treatment and brain region as fixed factors followed by post hoc pairwise comparisons using P < 0.05 as the criterion for significance. All post hoc comparisons were performed using a Bonnferoni correction to control for familywise error. | Results
Neonatal oestradiol treatment increased BrdU+ cells in the newborn female but not the male hippocampus
In order to assess the effects of oestradiol on cell proliferation in developing hippocampus, pups were treated on PN0 with oestradiol followed 2 h later with a single injection of BrdU and euthanized 6 h post-BrdU injection (for procedural timeline see Figure 1-A ). A 2 × 2 × 3 ANOVA indicated a significant main effect of sex [ F (1,54) = 28.43, P < 0.001] confirming that males have a higher mean number of BrdU+ cells than females. There was also a main effect of treatment [ F (1,54) = 14.68, P < 0.001] with oestradiol increasing the mean number of BrdU+ cells over vehicle. The three-way interaction involving sex × treatment × brain region was also significant [ F (2,54) = 3.85, P < 0.03], due to the larger number of new cells being born in the DG compared to CA1 and CA3. Regardless of the total number of new cells, post hoc pairwise comparisons indicated vehicle treated females had significantly fewer BrdU+ cells in all three subregions of the hippocampus than either oestradiol treated females or males treated with vehicle or oestradiol (Figure 3A-C ; P < 0.01). In females, oestradiol treatment increased the number of BrdU+ cells to the equivalent level seen in males, whereas there was no effect of oestradiol on the number of BrdU+ cells in males ( P > 0.05).
Cells generated in the presence of oestradiol survive for both short- and long-term periods in the female hippocampus
In order to assess the short- and long-term survival of newborn cells in the neonatal hippocampus, pups were treated with oestradiol followed 2 h later with a single injection of BrdU on PN0 (for procedural timeline see Figure 1-B ). At PN4 (short-term period), males again had more BrdU+ cell than females [ F (1, 57) = 7.96, P < 0.007] and oestradiol treatment increased the number of BrdU+ cells in DG and CA1 [ F (1, 57) = 5.55, P < 0.03]. A significant three-way interaction [ F (2,57) = 4.76, P < 0.03] followed by post hoc comparisons showed control females had fewer BrdU+ cells in both the DG and CA1 than either oestradiol treated females or males (Figure 4A-C ; Post hoc , P < 0.01).
In order to assess whether cells generated under the influence of oestradiol survive over a long period of time, pups were treated with oestradiol and BrdU on PN0 and PN1, then sacrificed on PN21 (for procedural timeline see Figure 1C ). The same general pattern of effects observed on PN4 was again apparent on PN21, with males having more BrdU+ cells than females [ F (1, 66) = 6.99, P < 0.01]; oestradiol treatment increased the number of BrdU+ cells in all three subdivisions of the hippocampus of females but not in males [ F (2, 66) = 7.20, P < 0.002; Figure 5A-C , Post hoc , P < 0.01].
The majority of cells born in the presence of oestradiol become neurons
In order to determine the fate of cells born on the first few days of postnatal life, we performed fluorescent immunohistochemistry and quantified the proportion of BrdU+ cells that are also positive for NeuN, a marker of mature neurons. The quantification of BrdU+/NeuN+ cells at PN21 indicated that both male and female groups had proportionally similar numbers of BrdU+ cells that were co-labelled with NeuN in the pyramidal cell layer of CA3 and the granular cell layer of DG (73%; ± 2% in CA3 and 81%; ± 4% in the DG). However, males had, overall, ~10% more BrdU+/NeuN+ cells than females [ F (1, 33) = 8.73, P < 0.01] with the DG having the highest number of BrdU+ cells that co-labelled with NeuN in both sexes [ F (2, 33) = 5,28, P < 0.05]. Post hoc comparison following a significant three-way interaction [ F (2, 33) = 4.04, P < 0.03] revealed significantly fewer co-labelled BrdU+/NeuN+ cells in area CA1 for control females compared to either oestradiol treated females and oestradiol treated males or control males ( P < 0.05; Figure 6 ). In control females, 41% of cells in CA1 born on PN0-1 were identified as neurons on PN21. In contrast, for females treated neonatally with oestradiol, 77% of the surviving cells were NeuN+, which is comparable to the 80% and 78% observed in control and oestradiol treated males, respectively.
The sex difference in cell proliferation is still apparent at PN21 and increased in females by neonatal oestradiol treatment
In order to assess whether the neonatal oestradiol-induced increases in cell proliferation endure beyond the period of steroid exposure, we quantified cells expressing Ki-67, an endogenous protein that labels actively proliferating cells, on PN21. Post hoc comparison following a significant three-way interaction [ F (2,36) = 4.49, P < 0.02] revealed that control females had significantly fewer Ki-67+ cells in the DG, as compared to control males and both neonatal oestradiol treated females and males ( P < 0.01; Figure 7 ). There were no significant main effects for sex or treatment ( P's > 0.05).
Inhibiting aromatase activity or blocking oestrogen receptor binding reduces cell proliferation in the developing male but not in the female hippocampus
In order to assess the impact of endogenous oestradiol on cell proliferation, we quantified the number of BrdU+ cells in pups that were treated on PN0 and PN1 with formestane (an aromatase inhibitor), tamoxifen (an oestrogen receptor (ER) antagonist) or vehicle, followed 2 h later with an injection of BrdU (for procedural timeline see Figure 1D ). Treatment with formestane or tamoxifen significantly reduced the number of BrdU+ cell compared to controls [ F (2, 152) = 30.63, P < 0.001]; this effect was detected in all three subdivisions of the hippocampal complex [ F (2, 152) = 42.54, P < 0.001]. Post hoc comparison following a significant two-way interaction of sex × treatment [ F (2, 152) = 30.97, P < 0.001] indicated that the effect of formestane and tamoxifen treatment was restricted to males ( P < 0.01; Figure 8 ). | Discussion
In the present study, we found that exogenous oestradiol treatment promoted cell proliferation and survival in the neonatal female but not the male hippocampus, whereas antagonizing endogenous oestradiol synthesis or action reduced cell proliferation in the male but not in the female hippocampus. These results confirm and extend our previous report describing a sex difference and oestradiol-induced increase in the number of new cells in the neonatal hippocampus [ 28 ] by confirming a sex difference in cell genesis that is regulated by oestradiol. Our previous report included evidence of a male biased sex difference and oestradiol-induced increase in BrdU+/glial fibrillary acid (GFAP)+ expressing cells. We also quantified the number of NeuN+ cells on PN4, but did not co-label them with BrdU as there was insufficient time for neurons born on PN0 to differentiate. We found more NeuN+ cells in males, but there was no significant increase following oestradiol treatment at PN0-1. This is, again, most probably due to insufficient time for neurons born on PN0 to differentiate into neurons. Therefore our previous study did not assess neurogenesis per se , as opposed to the later developmental stage of PN21 used here. Moreover, many GFAP expressing cells ultimately become neurons [ 36 ] and it is possible the oestradiol-induced increase we observed in BrdU+/GFAP+ cells was a precursor to the later increase in BrdU+/NeuN+ cells seen here. The current analysis is a more accurate depiction of the fate of cells labelled at birth with BrdU under the influence of oestradiol, whereas our previous results depict the fate of new cells within a few days of birth.
There are two mechanisms by which an increase in new cell number can occur; (1) an increased rate of proliferation; and/or (2) a decreased rate of cell death. In our previous study we observed a significant increase in the number of new cells within 24 h of BrdU injection in males versus females and in females treated with oestradiol versus vehicle treated females. In the current study the same pattern was observed at the even shorter post-BrdU injection time point of 8 h. The half-life of BrdU is approximately 2 h [ 37 ] whereas the cell cycle requires approximately 24 h to complete [ 38 ]. Therefore, differences in the number of BrdU+ cells within 24 h of injection are generally interpreted as differences in the rate of cell proliferation and not cell death [ 37 ]. Our observation of more BrdU+ cells in males and oestradiol-treated females at both 8 h and 24 h post-BrdU injection is, therefore, most consistent with a hormonally mediated sex difference in cell proliferation. However, given that many new cells die shortly after being born and would not be detected here, a potential contribution of cell death to the sex differences observed cannot be entirely ruled out.
While BrdU labelling offers many advantages, such as long-term tracking of cell fate, there are also limitations associated with potential toxicity and non-specific labelling [ 37 ]. In contrast, Ki-67 is an endogenous protein that does not have any adverse effects on living cells and is expressed in all phases of the cell cycle except the resting phase and a short period at the beginning of the G1 phase [ 39 - 41 ]. Thus, Ki-67 is not present in quiescent and terminally differentiated cells and increased numbers of Ki-67 expressing cells is consistent with increased proliferation (but see [ 42 ]). In order to determine if sex differences or hormonal modulation of cell proliferation persisted outside of the early neonatal period we quantified the number of Ki-67 expressing cells in the hippocampus of 3-week-old animals and found males and neonatally oestradiol-treated females had more Ki-67+ cells than control females. This observation suggests a higher rate of cell proliferation was organized during the neonatal sensitive period. A precedent for the sexual differentiation of neurogenesis is evident in studies reporting differential sensitivity of adult males and females to exogenous oestradiol treatment or rates of cell birth and death in the DG. In the adult female hippocampus, oestradiol stimulates cell proliferation [ 43 ], enhances cell survival [ 44 ] and increases dendritic spine synapse density [ 45 - 48 ]. In contrast, the adult male hippocampus is insensitive to the spinogenesis or cell genesis inducing effects of oestradiol treatment [ 44 , 49 , 50 ]. Oestradiol effects are mediated via binding to the two oestrogen receptor isoforms, ERα and ERβ [ 51 , 52 ]. Both isoforms are distributed throughout the brain, including the hippocampus, and both ERs colocalize with the proliferative marker Ki-67 in the adult hippocampus [ 50 , 53 , 54 ], although, there is a region specific variation in their distribution [ 54 - 57 ]. Moreover, in the adult hippocampus, there are relatively low levels ERα, which is in contrast to elevated levels of ERα that occur during early postnatal development [ 58 - 60 ]. Nonetheless, it is unclear which receptor isoform is regulating the oestradiol's effects on cell proliferation in the developing hippocampus. Both the short-term increase in the number of BrdU cells and the higher density of Ki-67+ cells in oestradiol treated females indicate a change in proliferative rates of progenitor cells induced by oestradiol and this may be due to a change in the duration of the progenitor cell cycle. This was determined to be the case in neocortical neurogenesis [ 61 ]. Oestradiol recruits cells into the S-phase from either G1 or G0 phase, which effectively shortens the G-phase, and results in an increased rate of proliferation of dividing progenitor cells [ 61 ].
Both the amount of endogenous oestradiol and aromatase activity in the developing hippocampus are extremely low compared to the hypothalamus and do not appear to be sexually dimorphic [ 62 ], suggesting that hippocampal sensitivity to oestradiol is high and differs between the sexes, at least as indexed by hippocampal cell genesis. However, because oestradiol, the ER antagonist and the aromatase inhibitor were all administered systemically, it is possible the effects were not mediated directly at the hippocampus but, instead, were secondary to changes in other brain regions projecting to the hippocampus. Cholinergic neurons of the medial septum/diagonal band of Broca are essential for oestradiol-induced spinogenesis in adult CA1 hippocampus [ 63 ] and cholinergic input modulates maturation and integration of adult born DG granule cells [ 64 ]. Gonadally intact males release more acetylcholine into the hippocampus than females during locomotor tasks and this sex difference is organized by oestradiol during development [ 65 , 66 ]. The cholinergic system matures relatively early and more new septal cholinergic neurons are born in males during a brief period of gestation but the sex difference does not persist into adulthood [ 67 ]. Nonetheless, it is possible that the effects observed here are the results of oestradiol-induced acetylcholine release into the neonatal hippocampus during the early postnatal period. It is also possible that oestradiol is acting outside the central nervous system. In adults, systemic oestradiol alters the arterial cerebral blood flow in females, but not in males [ 68 - 71 ], via an interaction with nitric oxide [ 72 , 73 ]. Both ERα and ERβ, as well as the transmembrane G protein-coupled receptor, GPR30, have been identified in blood vessels [ 74 , 75 ] and have been implicated in the rapid vasodilator effects of oestradiol [ 76 ]. One of the most powerful stimulators of adult neurogenesis is exercise [ 77 , 78 ]; an effect believed to be at least partly due to enhanced blood flow and increased delivery of growth factors. Lastly, activation of the stress axis has negative effects on adult neurogenesis and the injection of BrdU and steroidal agents to neonates is undoubtedly stressful. However, given that the number of injections was carefully controlled for across groups, sex differences in stress responding can not entirely explain the current results.
Neurogenesis in the adult hippocampus is restricted to the proliferative zone of DG, with the majority of the new cells becoming granule neurons [ 79 , 80 ]. A notable difference in the profile of neurogenesis in the immature brain is the presence of ongoing proliferation in CA1 and CA3 of Ammon's horn. The colocalization of BrdU with NeuN in these areas indicates that the cells born early postnatally do become neurons but whether they become pyramidal neurons or interneurons is unknown. When, in development, this source of new cells is lost and neurogenesis becomes restricted to the DG is also unknown. An additional unknown is the ultimate role of these enduring neurons in the adult hippocampal function. We observed almost twice as many new cells being born in the neonatal male hippocampus compared to the female. However, when the overall size of the hippocampus is compared in males and females, either developmentally or in adulthood, the sex difference, while biased towards males, is of the order of 10%-12% [ 15 , 21 ]. The magnitude of the sex difference in new cells was just as strong at 3 weeks of age as when they were just a few days old. Therefore, it does not appear that these cells contribute significantly to the hippocampal volume. The early postnatal period is a time of olfactory imprinting and somatosensory stimulation from the dam. The intensity of maternal licking and grooming is greater toward male than female pups [ 81 ] and both the nature and function of olfactory learning at this time is likely to be different between the sexes. Whether the role of new neurons born during this period is related to olfactory or sensory learning remains to be determined. | Conclusion
The hippocampus is a critical brain region involved in a variety of cognitive functions (for example, learning and memory) and both the physiologic and emotional responses to stress. Abnormalities in the hippocampus are strongly associated with affective disorders such as major depressive disorder and schizophrenia [ 82 - 85 ], as well as neurologic diseases such as Alzheimers. The selective vulnerability of the hippocampus to hypoxia/ischaemia following stroke, both perinatally and in adulthood [ 86 ], further emphasizes the importance of this critical brain region and the need to understand the variables that impact upon it in both males and females. Many of the sex differences observed in the adult hippocampus appear to be the result of early life events, including the impact of gonadal steroid hormones on neurogenesis during the early postnatal period. | Background
Oestradiol is a steroid hormone that exerts extensive influence on brain development and is a powerful modulator of hippocampal structure and function. The hippocampus is a critical brain region regulating complex cognitive and emotional responses and is implicated in the aetiology of several mental health disorders, many of which exhibit some degree of sex difference. Many sex differences in the adult rat brain are determined by oestradiol action during a sensitive period of development. We had previously reported a sex difference in rates of cell genesis in the developing hippocampus of the laboratory rat. Males generate more new cells on average than females. The current study explored the effects of both exogenous and endogenous oestradiol on this sex difference.
Methods
New born male and female rat pups were injected with the mitotic marker 5-bromo-2-deoxyuridine (BrdU) and oestradiol or agents that antagonize oestradiol action. The effects on cell number, proliferation, differentiation and survival were assessed at several time points. Significant differences between groups were determined by two- or thee-Way ANOVA.
Results
Newborn males had higher rates of cell proliferation than females. Oestradiol treatment increased cell proliferation in neonatal females, but not males, and in the CA1 region many of these cells differentiated into neurons. The increased rate of proliferation induced by neonatal oestradiol persisted until at least 3 weeks of age, suggesting an organizational effect. Administering the aromatase inhibitor, formestane, or the oestrogen receptor antagonist, tamoxifen, significantly decreased the number of new cells in males but not females.
Conclusion
Endogenous oestradiol increased the rate of cell proliferation observed in newborn males compared to females. This sex difference in neonatal neurogenesis may have implications for adult differences in learning strategy, stress responsivity or vulnerability to damage or disease. | Abbreviations
BrdU: 5-bromo-2-deoxyuridine; DAB: diaminobenzidine; DG: dentate gyrus; ER: oestrogen receptor; GFAP: glial fibrillary acid protein; HCl: hydrogen chloride; ip: intraperitoneally; PBS: phosphate buffered saline; PFA: paraformaldehyde; PN: postnatal day; RT: room temperature; sc: subcutaneous.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
MB performed the immunohistochemistry, statistical analyses and drafted the manuscript. Both MB and JW treated the animals, processed the tissue samples and performed the cell counts. MM conceived the study and participated in its design and coordination. All authors read and approved the final manuscript. | Acknowledgements
This work was supported by R01 NS050525 awarded to MMM. | CC BY | no | 2022-01-12 15:21:44 | Biol Sex Differ. 2010 Nov 22; 1:8 | oa_package/8d/4b/PMC3016241.tar.gz |
PMC3016242 | 21162752 | Background
After the introduction of selective decontamination of the digestive tract (SDD) in intensive care unit patients in 1984, a reduction in the incidence of nosocomial infections in patients with endotracheal tubes was shown [ 1 ]. Prophylactic perioperative SDD to prevent nosocomial infection in patients undergoing an esophageal anastomosis is effective and easy to perform [ 2 ]. Topical application of nonresorbable antimicrobial agents to the oropharynx and gastrointestinal tract typically prevents secondary colonization with Gram-negative bacteria, S. aureus and yeast. Only selective antibiotics (both topically and systemically) without anti-anaerobic activity are usually used to maintain the anaerobic intestinal flora. These measures reduce the incidence of perioperative nosocomial infections [ 3 - 7 ]. Furthermore, SDD has been shown to be effective in the prevention of esophagojejunal anastomotic leakage after total gastrectomy and has therefore been proposed as a prophylactic measure [2;8;9]. However, SDD has not yet been widely accepted as standard procedure for patients undergoing an esophageal anastomosis, which represents a group of patients with relatively high risk of perioperative morbidity. Pulmonary complications are the major source of morbidity and mortality after esophageal resection [ 10 ]. Approximately 30% of patients who undergo esophagectomies suffer from respiratory complications, and 80% of these complications occur within the first five postoperative days [ 11 ].
Several risk factors for pulmonary infections in patients who undergo esophagectomies were identified. Elderly patients and those with chronic obstructive pulmonary disease are at risk for the development of major pulmonary complications that require interventions, such as antibiotic therapy, bronchoscopy or endotracheal intubation. Pneumonia is frequently caused by postoperative aspiration and is the major cause of death in these patients. Minor pulmonary complications that do not require interventional measures occur in almost all patients who undergo an esophageal resection [ 12 ].
Leakage of the esophageal anastomosis is another serious and potentially life-threatening complication. Together with postoperative pulmonary infections, anastomotic leakage is a major cause of death after esophageal resection [ 13 ].
Based on these findings and experiences with intensive care unit patients, SDD was introduced at our hospital as perioperative prophylaxis in patients undergoing an esophageal anastomoses, in a manner analogous to the approach described in Schardey et al. [ 2 ]. The intention was to reduce the incidence of postoperative infections and therefore reduce perioperative morbidity and mortality.
In a previously analyzed retrospective cohort, we showed a reduction in postoperative nosocomial infections in patients with SDD compared to patients without SDD (data not published at the time). Based on these positive results, SDD was routinely implemented as perioperative prophylaxis for all patients undergoing an esophageal anastomosis. Consequently, all patients receiving a total or partial esophagectomy, transhiatal esophagogastrectomy or a total gastrectomy were consistently treated with SDD; data were prospectively recorded to determine the perioperative morbidity and mortality (especially in patients who developed an anastomotic leakage) compared to a retrospective control group.
The aim of this study was to determine whether there was a benefit of SDD to postoperative morbidity and mortality in patients undergoing an esophageal anastomosis. For this purpose, the results of our prospective cohort were compared to the previously analyzed retrospective cohort of surgical patients at our institution who did not receive SDD. | Methods
Between January 2002 and December 2007, a total of 124 patients underwent elective partial or total esophageal resection at a tertiary referral hospital, and data were prospectively recorded and retrospectively analyzed. All patients undergoing an esophageal anastomosis after either total gastrectomy, transhiatal extended gastrectomy or a Merendino procedure were included in the study [ 14 ]. Patients who underwent the following procedures were excluded (n = 43): those who underwent a transthoracic esophagectomy were excluded because they had a cervical anastomosis and no intrathoracic anastomoses; patients having a subtotal gastrectomy with pouch reconstruction were excluded because a gastrojejunostomy was considered to be a different procedure. Additionally, patients undergoing a transmediastinal esophagectomy were excluded to separate proximal esophageal anastomoses from distal esophageal anastomoses because we believe these groups differ in terms of anastomotic leakage. Thus, a homogeneous cohort of 81 consecutive patients undergoing an esophageal anastomosis remained for further analyses.
For SDD, patients received a solution of polymyxin (100 milligrams), tobramycin (80 milligrams), vancomycin (125 milligrams) and nystatin (500 milligrams) by mouth four times a day. This solution was administered at a dose of 10 milliliters by mouth every six hours, and the treatment was continued intraoperatively and postoperatively via a gastric tube. The treatment was started on the morning of the day before surgery (they received at least four doses before surgery) and continued until the seventh postoperative day. All patients received total parenteral nutrition postoperatively for 7 days. On postoperative day 7, a radiological imaging study with oral water-soluble contrast was routinely performed to determine whether anastomotic leakage occurred. Anastomotic leakage was defined as an extravasation of the water-soluble contrast during the radiological study. If the radiological test was normal, the patients were allowed to start oral intake and SDD was stopped on day 7. Total parenteral nutrition was reduced when solid oral intake was started, which was usually on postoperative day 9. Pulmonary infection was diagnosed if two or more of the following clinical signs were present: fever, lung infiltrates on radiological imaging studies, purulent respiratory secretion or a positive culture from respiratory secretions. If a pulmonary infection was diagnosed, it was treated with tazobactam-piperacillin according to the hospital guidelines for the treatment of nosocomial pulmonary infections. All patients with suspected pulmonary infections underwent chest radiography to confirm the diagnosis.
All patients (retrospective and prospective, with or without SDD) received perioperative antibiotic prophylaxis of 2 grams of cefamandole intravenously and 500 milligrams of metronidazole intravenously 30 to 60 minutes before the start of the operation, according to the hospital's standard guidelines for perioperative antibiotic prophylaxis. Antibiotic prophylaxis was repeated intraoperatively if the duration of the procedure exceeded four hours.
The standard procedure for an esophageal anastomosis was an end-to-side esophagojejunostomy using a 25 millimeter circular stapling device. A short, crooked approximately 1-centimeter portion of the jejunum was reconstructed using an ENDO GIATM universal 45 millimeter stapling device.
The outcome was compared to a historic cohort (n = 121) that was treated from 1995 to 2002 at the same institution. Within this cohort, there were two subgroups: one that was treated with SDD (n = 53) and one that was not (n = 68). The surgeon chose whether or not a patient received SDD. Since July 2002, SDD was routinely used in all patients undergoing an esophageal anastomosis (eligible for this study: n = 81). Age, gender, body mass index (BMI), ASA score, anastomotic leakage, pulmonary infection and mortality rate were compared among the three groups. Perioperative morbidity was analyzed using the classification described by Dindo et al. [ 15 ].
All patients provided informed consent to the operative treatment, the administration of SDD and further analyses of data from their medical records. The medical ethics committee approved the analysis of these data.
Statistical analysis was performed using SPSS 11.5 software (SPSS Inc., Chicago, IL, USA). A two-sided p-value < 0.05 was considered statistically significant. Continuous data are expressed as mean ± standard deviation. Confidence intervals (95% CI) of binominal proportions were estimated according to Agresti and Coull. Mann-Whitney U tests and Kruskal Wallis tests were applied to compare continuous data. Chi-Square tests were used to compare proportions. | Results
Prospective cohort: Specification of sample
Between 2002 and 2007, a cohort of 81 patients who underwent an esophageal anastomosis (esophagojejunostomy) and received SDD was obtained. A total gastrectomy was performed in 41 patients (51%), transhiatal esophagogastrectomy was performed in 26 (32%) patients and a Merendino procedure was performed in 14 (17%) (14). The mean patient age was 63.1 ± 11.5 years; this cohort included 47 (58%) women.
Comparability of groups
In Table 1 , baseline characteristics of both the prospective and retrospective groups are provided. No significant differences in age, BMI or ASA classification were found between the three groups.
Retrospective analysis (Table 2 )
Patients from the retrospective subgroup with SDD were compared to patients who had no SDD. Significantly fewer patients with SDD than those without SDD died within the first 30 postoperative days. A mortality rate of 17.6% (12/68; 95% CI 10.2% - 28.5%) was found in the retrospective subgroup without SDD compared to the mortality rate of 1.9% (1/53; 95% CI 0.3% - 13.5%) in the subgroup with SDD (p = 0.005). Of the patients without SDD, 13.2% (9/68; 95% CI 6.9% - 23.5%) developed an anastomotic leakage, whereas only 3.8% (2/53; 95% CI 0.3% - 13.5%) of the retrospective subgroup with SDD developed an anastomotic leakage (p = 0.072). Of the patients who had no SDD, 25.0% (17/68; 95% CI 16.2% - 36.5%) developed postoperative pneumonia compared to 11.3% (6/53; 95% CI 4.9% - 22.9%) of the patients with SDD (p = 0.057). The postoperative length of hospital stay was shorter in the retrospective subgroup with SDD (21.8 ± 13.7 vs 25.2 ± 18.0 days, p = 0.145). According to the Dindos classification, overall morbidity was less severe in patients with SDD (p = 0.009) (Table 3 ).
Longitudinal analysis
The retrospective subgroup with SDD was compared to the prospective cohort with SDD (Table 2 ), and no statistically significant differences were found in terms of 30-day mortality: 1.9% (1/53; 95% CI 0.0% - 10.9%) in the retrospective subgroup compared to 1.2% (1/81; 95% CI 0.0% - 7.3%) in the prospective cohort (p = 0.761). In the retrospective subgroup, 11.3% (6/53; 95% CI 4.9% - 22.0%) developed postoperative pneumonia compared to 4.9% (4/81; 95% CI 1.6% - 12.4%) in the prospective cohort (p = 0.169). In the retrospective subgroup with SDD, 3.8% (2/53; 95% CI 0.3% - 13.5%) developed an anastomotic leakage, whereas 7.4% (6/81; 95% CI 3.2% - 15.5%) suffered from anastomotic leakage in the prospective cohort (p = 0.385). The length of postoperative hospital stay did not differ significantly between the prospective cohort and the retrospective subgroup with SDD (20.2 ± 10.9 vs 21.8 ± 13.7 days, p = 0.861).
Patients without SDD were compared to the patients of the prospective cohort with SDD (Table 2 ); the postoperative 30-day mortality rate was significantly lower in patients with SDD (1.2% vs. 17.6%; p = 0.001). Patients with SDD developed pneumonia postoperatively less often (4.9%; 4/81; 95% CI 1.6% - 12.5%) than patients without SDD (25.0%; 17/68; 95% CI 16.2% - 36.6%; p = 0.001). Fewer patients with SDD than patients without SDD developed an anastomotic leakage (6/81 vs. 9/68; 7.4% vs. 13.2%; p = 0.239). Patients with SDD in the prospective cohort had significantly fewer severe complications than patients without SDD in the retrospective subgroup (p = 0.001) (Table 3 ). The postoperative hospital stay was shorter in the prospective cohort with SDD than in patients without SDD (20.2 ± 10.9 vs 25.2 ± 18.0 days, p = 0.046).
Pooled subgroup analysis of patients with anastomotic leakage revealed a mortality rate of 55.6% (5/9) in the patients without SDD compared to 12.5% (1/8) in all patients with SDD (p = 0.064). | Discussion
This observational study corroborates the hypothesis that SDD reduces perioperative morbidity and mortality in patients who undergo a distal esophageal anastomosis [ 4 ]. In the present study, perioperative mortality was significantly reduced in patients undergoing a distal esophageal anastomosis when perioperative SDD was performed. Furthermore, significantly fewer patients developed perioperative pulmonary infections. The perioperative mortality rate of patients who postoperatively developed an anastomotic leakage was clearly lower in the SDD group, although this difference was not significant. However, we consider this result to be a strong indicator of the benefit of SDD. Anastomotic leakage is a potentially life-threatening postoperative complication. The mortality rate in patients with anastomotic leakage after total gastrectomy was reported to be as high as 45% [ 16 ]. In our patients without SDD, a mortality rate of 55.6% was found, whereas patients with SDD had a mortality rate of 12.5%. Furthermore, our results are comparable to the results of the study by Schardey et al., in which significantly fewer pulmonary infections were described, and a tendency of fewer anastomotic insufficiencies was found in patients receiving SDD [ 2 ].
SDD has not been widely adopted as a prophylactic measure in patients undergoing esophageal anastomoses. One possible reason for this might be the fear of complications of such a nonspecific antibiotic treatment. In the studies by Tetteroo et al., no increase in antibiotic resistance was found in patients receiving SDD found [4-7;17]. Additionally, the most recent Cochrane review on antibiotic prophylaxis reported no resistance associated with SDD in one trial [ 18 ]. In the study by de Jonge et al., a decreased development of resistance among SDD-treated patients in intensive care over a 27-month period was reported [ 19 ]. Although not measured, SDD-associated complications in our patients seldom occur; however, they most often present as nausea, which is considered to be a mild side effect. Additionally, no SDD-associated side effects were reported in a recently published study by Roos et al. [ 20 ]. Similar mortality rates and incidence of perioperative infections were found when SDD was employed in patients undergoing elective colorectal surgery compared to patients receiving the same operation but not receiving SDD. However, in that study, SDD was not applied in a standardized fashion, and pulmonary infections were not significantly reduced in the SDD group. We assume that patients have recurrent micro-aspirations after esophagus resection, and therefore, SDD seems to effectively prevent the development of pulmonary infections in our patient cohort.
We are aware of some relevant limitations of this study. The comparison of a prospective cohort with a retrospective cohort is problematic. However, apart from only a few data from the literature that address this specific problem, our retrospective patient cohort is the only source of data available. We aimed to minimize the risk of bias by statistically comparing all groups where no significant differences in the baseline characteristics were found. Furthermore, when comparing the retrospective subgroups with and without SDD, essentially the same outcome was obtained. However, the absence of evidence is not evidence for absence, and longitudinal analyses are always associated with a risk of bias. Furthermore, we assume that advancements in intensive care medicine and anesthesia have occurred over the course of data collection. This possible influence on perioperative morbidity and mortality was not addressed in the current study.
Despite the above-mentioned possible limitations of this study, our results show a strong tendency in favor of SDD for patients undergoing a distal esophageal anastomosis. We encourage institutions where SDD is not performed to conduct a double-blind randomized controlled prospective study to obtain results regarding SDD use with a higher level of evidence than provided here. | Conclusions
We conclude that SDD in patients undergoing an esophageal anastomosis may provide perioperative morbidity and mortality benefits. | Background
Selective decontamination of the digestive tract (SDD) to eliminate gram-negative bacteria is still not widely accepted, although it reduces the incidence of nosocomial infections. In a previous retrospective study, a clear benefit to perioperative morbidity, and a reduction in nosocomial infections were found in patients who underwent an esophageal anastomosis. Thus, SDD was applied routinely for esophageal anastomoses. We report the outcome of a cohort of 81 patients who underwent this treatment.
Methods
From 2002, patients who underwent an esophageal anastomosis (esophagojejunostomy) were prospectively recorded. Perioperatively, patients received polymyxin, tobramycin, vancomycin and nystatin by mouth four times a day. Outcome was compared to a control group that was treated before 2002 (68 patients without SDD and 53 patients with SDD). Postoperative morbidity and mortality were assessed.
Results
Between 2002 and 2007, 81 patients who underwent an esophageal anastomosis received SDD. Compared to a retrospective control group, patients with SDD had significantly less pneumonia (OR 0.06 (0.01-0.46), p < 0.001) and lower morbidity (OR 0.16 (0.05-0.49), p < 0.001). Furthermore, fewer anastomotic insufficiencies and complications were found. Similar results were found in the analysis of the patients treated before 2002.
Conclusions
SDD significantly reduces perioperative morbidity and mortality in patients who undergo a distal esophageal anastomosis compared to a historical control group. In patients with an anastomotic leakage, there was a strong tendency of SDD to reduce postoperative mortality. | Competing interests
The authors declare that they have no competing interests.
Authors' contributions
FN performed the data acquisition and drafted the manuscript. RW participated in the study design and performed the statistical analysis. FN and RW contributed equally to this manuscript. WK performed data acquisition of the retrospective cohort. MZ and JL critically revised the manuscript for important intellectual content. TS conceived the study and participated in the study design, drafting the manuscript and performing the revisions.
All authors read and approved the final manuscript.
Pre-publication history
The pre-publication history for this paper can be accessed here:
http://www.biomedcentral.com/1471-2482/10/36/prepub | Acknowledgements
We would like to acknowledge the support of the reviewers of BMC Surgery. | CC BY | no | 2022-01-12 15:21:44 | BMC Surg. 2010 Dec 16; 10:36 | oa_package/59/78/PMC3016242.tar.gz |
PMC3016243 | 21144047 | Introduction
Familial hypercholesterolaemia (FH) is primarily an autosomal dominant disorder, characterised by a lifelong elevation of serum cholesterol bound to low-density lipoprotein (LDL). The primary causative defects in approximately 85% of FH cases are mutations or deletions in the plasma membrane Low Density Lipoprotein Receptor (LDLR) encoding gene that is responsible for clearing LDL-cholesterol (LDL-C) from the blood stream by endocytosis and intracellular degradation [ 1 ]. Over 1000 different mutations in the LDLR gene on the distal short arm of chromosome 19 (p13.1-p13.3) have been described to date [ 2 ] and are recorded online at http://www.ucl.ac.uk/ldlr/Current/ [ 3 ]. The second gene responsible for fewer than 10% of FH cases encodes the ligand for LDLR, namely Apolipoprotein B-100 (ApoB-100), located on the short arm of chromosome 2 (p24) [ 4 ]. Mutations in this gene reduce ligand affinity for the receptors and cause reduced clearance of LDL particles resulting in hypercholesterolemia [ 5 ], albeit normal LDLR activity. A mutation in the codon for amino acid 3500 (CGG-to-CAG) was found to be a CG mutation hotspot associated with defective LDLs and hypercholesterolemia [ 6 ]. The pathophysiological consequences from LDLR or ApoB mutations are loss of protein function, which lead to monogenic FH. Defects in a third gene, located on the short arm of chromosome 1 (p34.1-p32), have also been identified to cause monogenic FH [ 7 ]. The convertase subtilisin/kexin type 9 (PCSK9)-gene codes for an enzyme that has also been called ''neural apoptosis regulated convertase 1'', which has been proposed to be involved in degrading the LDLR protein in the lysosome and thus preventing it from recycling [ 8 ]. Gain of function mutations in the PCSK9 gene could therefore cause increased degradation of LDLRs, reduced numbers of receptors on the surface of the cell, and monogenic FH. An autosomal recessive form of FH caused by loss of function mutations in the LDLRAP1 gene, which is located on the short arm of chromosome 1p35-36.1, has also been documented [ 9 ]. The clinical phenotype of the autosomal recessive form is similar to that of the classic homozygous FH caused by defects in the LDLR gene, but it is generally less severe and more responsive to lipid-lowering therapy (reviewed in [ 10 ]). This article focuses on LDLR-associated FH reviewing, the encountered obstacles, the achieved progress and the future prospectives of LDLR-gene therapy for this disease. | Methods of gene delivery
Gene transfer can be performed either ex vivo , involving isolation of autologous cells from the patient, their in vitro genetic modification and selection followed by reimplantation of the transduced cells, or it can be done in vivo , where the vector is delivered directly to the organ [ 83 ]. The advantage of the ex vivo approach is that the transduction/transfection conditions can be carefully controlled and optimised and individual clones with the most desirable characteristics can be isolated to eliminate unmodified cells or cells with deleterious mutations before re-implantation. While this approach is laborious and time consuming, it may also offer significantly greater safety and control with respect to vector mediated mutagenesis and possible germline transmission of the transferred genes, which is a risk of in vivo gene delivery. The disadvantages of the ex vivo approach are failure of cell engraftment and difficulties in returning the cells to the patient due to disease manifestations such as portal vein hypertension [ 83 , 84 ].
The in vivo approach eliminates the need for engraftment after re-implantation and is therefore easier to perform, more cost effective and may be more applicable for use in countries with limited laboratory resources. The gene transfer vector is injected into the bloodstream (systemic delivery) aiming at somatic cell delivery only or by use of specific cell targeting, preferentially to the tissues of interest (targeted delivery). Organ specific delivery of the gene transfer vector includes intrahepatic injection or selective intravasular application routes. Disadvantages of in vivo gene transfer are vector dilution, ectopic transgene expression and non-targeted, random, potentially genotoxic insertion into the host genome. | Conclusion
Several novel therapies have been developed recently to lower LDL-C in homozygous and heterozygous FH patients [ 57 - 65 , 211 ]. However, their major drawback is the need for life-long repeated administration in a similar manner to conventional pharmacological drugs. The advantage of gene therapeutic intervention over other therapeutic regimes is the potential for lifetime correction with a single vector administration. Yet, this goal still needs to be achieved. Despite the considerable progress, made in optimising the two most commonly used gene therapy vector-groups based on retro- and adenoviruses, neither vector has been found to be ideal for in vivo and/or ex vivo gene transfer. Vectors derived from AAV and LVs are very promising. However, the oncogenesis risk from semi-random integration into actively transcribing genes of the host by LVs [ 189 , 212 , 213 ], possible germline transmission [ 214 ] and some immunological reaction after AAV gene transfer in the human haemophilia-B trials [ 144 , 215 ] are critical drawbacks that require further vector development and improvement.
In addition to the LDLR gene augmentation approach, the successful use of the VLDLR as an effective surrogate lipoprotein receptor gene [ 139 - 141 , 153 , 216 , 217 ] for the complementation of mutated LDLR function in homozygote FH patients would also open an alternative therapeutic avenue, since it would avoid the immune problem in patients with no natural LDLR. Despite the fact that most of the pre-clinical and clinical studies were aimed at treatment of the homozygous form of FH, a minority of heterozygous FH patients, who are refractory to existing pharmacological therapy, are also possible targets. Therefore, once a safe and efficient transfer vector is developed and shown to be effective in homozygous FH, its application might be extended to severe heterozygous FH as well. Clearly there is also an urgent demand for safe and efficient vectors that would integrate into the host genome and provide long-term appropriate gene expression for in vivo and/or ex vivo gene therapy of FH and many other human diseases.
Future work will generally focus on making gene transfer vectors safer by improving their immunogenic, integration, expression and targeting profile. Reducing the inherent oncogenic danger of integrating vectors by engineering conditional suicide genes into the vector backbone to provide a self-destructive mechanism in case of oncogenesis or by targeting their integration into specific pre-defined benign genomic sites, i.e. by zinc-finger nuclease technology, may help achieving this goal [ 218 ]. In combinations with the above strategies, the use of ex vivo transduction to reduce vector spread can also improve the safety outcome, particularly, if autologous or induced pluripotent stem cells are the target. The application of viral or non-viral, integrating or non-integrating vectors for long term persistence in stem cells with self renewal and differentiation capacities will also be important perspectives for gene-based stem cell therapy. | Coronary artery diseases (CAD) inflict a heavy economical and social burden on most populations and contribute significantly to their morbidity and mortality rates. Low-density lipoprotein receptor (LDLR) associated familial hypercholesterolemia (FH) is the most frequent Mendelian disorder and is a major risk factor for the development of CAD. To date there is no cure for FH. The primary goal of clinical management is to control hypercholesterolaemia in order to decrease the risk of atherosclerosis and to prevent CAD. Permanent phenotypic correction with single administration of a gene therapeutic vector is a goal still needing to be achieved. The first ex vivo clinical trial of gene therapy in FH was conducted nearly 18 years ago. Patients who had inherited LDLR gene mutations were subjected to an aggressive surgical intervention involving partial hepatectomy to obtain the patient's own hepatocytes for ex vivo gene transfer with a replication deficient LDLR-retroviral vector. After successful re-infusion of transduced cells through a catheter placed in the inferior mesenteric vein at the time of liver resection, only low-level expression of the transferred LDLR gene was observed in the five patients enrolled in the trial. In contrast, full reversal of hypercholesterolaemia was later demonstrated in in vivo preclinical studies using LDLR-adenovirus mediated gene transfer. However, the high efficiency of cell division independent gene transfer by adenovirus vectors is limited by their short-term persistence due to episomal maintenance and the cytotoxicity of these highly immunogenic viruses. Novel long-term persisting vectors derived from adeno-associated viruses and lentiviruses, are now available and investigations are underway to determine their safety and efficiency in preparation for clinical application for a variety of diseases. Several novel non-viral based therapies have also been developed recently to lower LDL-C serum levels in FH patients. This article reviews the progress made in the 18 years since the first clinical trial for gene therapy of FH, with emphasis on the development, design, performance and limitations of viral based gene transfer vectors used in studies to ameliorate the effects of LDLR deficiency. | LDLR-associated FH
Owing to mutations in both alleles of the LDLR locus, homozygous LDLR-associated FH patients present with markedly elevated total serum cholesterol (>500 mg/dL, 13 mmol/L) and LDL-cholesterol levels (LDL-C, >450 mg/dL, 11.7 mmol/L). The deposition of insoluble cholesterol causes xanthomata on the tendons of the hands and feet, cutaneous planar and corneal arcus in early life [ 11 , 12 ]. Atheroma of the aortic root and valve can lead to myocardial infarction (MI) and sudden death before the age of 30 years. Coronary artery disease (CAD) is more common and more extensive in receptor negative patients (mutations that completely eliminate receptor functions) than in those with the receptor-defective type (mutations that partially inactivate receptor function), where there is residual receptor activity [ 12 , 13 ]. Heterozygous patients typically have a lower serum cholesterol level (250-450 mg/dL or 6.5-11.6 mmol/L) and LDL-C (200-400 mg/dL or 5.2-10.4 mmol/L) with positive age correlation. They develop the above clinical features at a less accelerated rate, but if untreated most suffer a severe MI and often sudden death or other cardiovascular events in the fourth or fifth decade of life. Due to several hormonal factors, approximately 80% of heterozygote men suffer from CAD, while only 20% to 30% of women are moderately affected [ 14 ].
In most investigated populations, the heterozygote form occurs in at least 1:500 and the homozygous form in one in one million individuals [ 15 ], although in some populations, for example the Afrikaner population in South Africa, heterozygosity is found in less than 1:80 individuals [ 16 , 17 ]. This unusual high frequency is due to founder effects and no heterozygote advantage has been identified. Heterozygous FH is therefore the most frequent clinically relevant Mendelian trait, being more frequent than homozygous cystic fibrosis and sickle cell anaemia.
Cholesterol levels alone are not sufficient to confirm a diagnosis of FH because blood cholesterol levels vary with age, gender and are population specific [ 18 ]. In addition, the range of blood cholesterol levels in FH overlaps with that of people with non-genetic multifactorial hypercholesterolaemia, which reduces diagnostic accuracy. Diagnostic criteria of FH, therefore, include clinical symptoms and laboratory findings as well as the family history of a dominant pattern of inheritance for either premature coronary heart disease or hypercholesterolaemia, (reviewed in [ 18 ]).
The human LDLR is a multi-component single-chain glycoprotein, which contains 839 amino acids in its mature form, encoded by a gene of 45 kb in length [ 19 ]. The gene contains 18 exons of which 13 exons code for protein sequences that show homology to other proteins such as the C9 component [ 20 ], Epidermal Growth Factor (EGF) [ 21 ], blood coagulation factor IX, factor X (FX) and protein C [ 22 - 24 ]. The mRNA transcript is 5.3 kb in length and encodes a protein of 860 amino acids. About half of the mRNA constitutes a long 3' untranslated region that contains two and a half copies of the Alu family of middle repetitive DNAs [ 25 ]. LDL-Receptors are expressed ubiquitously by almost all somatic cells under control of sterol negative feedback, mediated by three 16 bp imperfect repeats (sterol regulatory elements) and a TATA box like sequence in the promoter [ 26 ]. Their function is to bind to apolipoprotein ligands, apoB-100 and apoE. Uptake of LDL is mediated mainly through apoB-100 [ 27 ].
The mature human LDLR of 160 kDa is composed of five domains, Figure 1 . Exon 1 encodes a short 5' untranslated region and 21 hydrophobic amino acids that are not present in the mature protein. This sequence functions as a signal peptide to direct the receptor synthesising ribosomes to the Endoplasmic Reticulum (ER) membrane [ 25 ]. Other functional domains of the peptide correspond to the exons as indicated in Figure 1 , 41 bp of exon 17 plus exon 16 encode the transmembrane domain and the reminder of exon 17 together with exon 18 encode the cytoplasmic domain.
Analyses of LDLR-associated FH variants estimated that there were 1066 LDLR gene mutations/rearrangements, 65% (n = 689) of which were DNA substitutions, 24% (n = 260) small DNA rearrangements (<100 bp), and 11% (n = 117) large DNA rearrangements (>100 bp) [ 2 ]. The DNA substitutions and small rearrangements occur along the length of the gene, with 839 in the exons (93 nonsense variants, 499 missense variants and 247 small rearrangements), 86 in intronic sequences, and 24 in the promoter region. The highest proportion of exon variants occurs in the ligand binding domain (exons 2-6) and the EGF precursor domain (exons 7-14) [ 2 ].
Clinical management of FH
To date there is no cure for FH. The primary goal of clinical management in heterozygotes is to control hypercholesterolaemia by lifestyle modification and/or drug treatment in order to decrease the risk of atherosclerosis and to prevent CAD. Lifestyle modification involves educating patients to adhere to a low-fat diet, exercise and to reduce overweight or maintain an optimal body weight. An effective low-fat diet could lower LDL-C (LDL cholesterol) by 20% to 30% [ 28 - 30 ]. For patients who are not able to reach their LDL-C goal (<129 mg/dI, 3.31 mmol/L) on the lifestyle modification program, drug therapy is the next step. The current recommendations for LDL-C goals from the National Cholesterol Education Program Adult Treatment Panel III guidelines are <100 mg/dI, 2.586 mmol/L for patients with very high cardiovascular risk and <129 mg/dI, 3.31 mmol/L for patients with moderate cardiovascular risk [ 31 ]. The preferred and most effective lipid-lowering agents are the 3-hydroxy-3-methylglutaryl-coenzyme A (HMG-CoA) reductase inhibitors, more commonly known as statins [ 32 ]. Statins are the best tolerated medication in patients of all ethnic groups, both sexes, and generally, all ages. They also have an excellent safety profile over the now nearly 20 years of widespread clinical use, and have the highest level of patient adherence among available lipid-lowering agents with low incidence of side effects [ 33 ]. Because different statins have variable potency, the therapeutic outcome ranging from 20% to 60% reduction in LDL-C [ 32 ], depends on the particular statin used, the dose and the type of LDLR mutation. Despite the powerful effect of statins, they may not be appropriate for those who are best treated with non-systemic therapy (eg, young adults, women of childbearing age), who require only a modest reduction in LDL-C, or those with active liver disease or increased liver function test values and who predominantly have hypertriglyceridemia. Increasing the statin dose to 80 mg (rosuvastatin to 40 mg) is associated with a threefold increase in liver toxicity or myopathy [ 34 ]. Therefore, treatment with non-statin cholesterol lowering agents, for example bile acid resin [ 35 ], niacin [ 36 ], fibrate [ 37 ] or cholesterol absorption inhibitor [ 38 ], is recommended for these patients.
Bile acid binding resins are non-absorbable anion exchange resins that bind bile acids in the intestinal lumen, preventing their absorption from the ilium and therefore increasing their fecal excretion. The liver responds by up-regulating cholesterol 7-alpha hydroxylase, which increases the conversion of cholesterol to bile acids, thereby reducing the cholesterol concentration in the hepatocyte [ 39 ]. Gastrointestinal disturbances, and drug and/or fat-soluble vitamin malabsorption, which were associated with early generation bile acid resins, have been overcome with new generation agents [ 35 ]. Bile acid resins can lower LDL-C approximately 10% to 25% which is appropriate for patients who need only moderate LDL-C lowering [ 35 ].
Niacin, or nicotinic acid, is the oldest lipid-lowering drug dating back to the 1950s [ 39 ]. Depending on dose and formulation, LDL-C reductions of 12% to 20% maybe anticipated, along with good reductions in triglycerides and 17% to 31% increase in high-density lipoprotein cholesterol (HDL-C). The major drawback to niacin use is its side effects, which include itching, headaches and hepatotoxicity. It is contraindicated in patients with active liver disease or unexplained abnormalities in liver function tests [ 39 ].
A cholesterol absorption inhibitor, more commonly known as Ezetimibe, impairs dietary and biliary cholesterol absorption at the brush border of the intestine without affecting the absorption of triglycerides or fat-soluble vitamins [ 19 ]. It has been shown to be well tolerated and effective in lowering LDL-C when used as a monotherapy or when adding to statin therapy. Ezetimibe at a dose of 10 mg/day reduced LDL-C by approximately 17% with no adverse effect of myopathy or liver toxicity [ 40 , 41 ]. However, recently concerns have been raised in respect to an independent atherogenic property of this drug, which appears to counteract its cholesterol-reducing action [ 42 ].
For patients who do not respond to a maximum dose of a statin and those who develop side effects with higher doses, a combination therapy of statin with one of the above agent, rather than an increase in the statin to high doses, may be more effective in achieving LDL-C goals and improving CAD outcomes while remaining at an acceptable safety profile [ 43 ]. Adding a bile acid resin or niacin to the statin can reduce LDL-C by approximately 50%, depending on the choice of statin and dosage prescribed [ 44 , 45 ]. Co-administering 10 mg of ezetimibe with any dose of statin reduced LDL-C levels by an additional 25%, compared with the usual 6% attained by doubling the statin dose [ 46 ]. However, even after treatment with a combination therapy, the majority of homozygous and minority of heterozyogotes FH patients may still have extremely raised LDL-C serum levels [ 47 ] and their risk of CAD remains unacceptably high.
Surgical interventions involving a portocaval shunt or an ileal bypass have yielded transient lowering of plasma LDL in these patients [ 48 ]. The preferred treatment at present is an aggressive programme of plasma apheresis or LDL apheresis, a physical procedure in which LDL is selectively removed from the blood by passing plasma over columns that bind the LDL. A small number of angiographic regression studies have been conducted and each weekly or fortnightly treatment has been demonstrated to lower LDL-C levels by about 55% and to delay onset and progression of atherosclerosis [ 49 - 53 ].
The most significant but also most aggressive metabolic correction is orthotopic liver transplantation in homozygous patients [ 54 - 56 ]. However, the morbidity and mortality risks as well as scarcity of donated organs are serious limitations.
Several novel therapeutic approaches have also been developed recently to lower LDL-C, either as monotherapy or in combination with statins [ 57 ] including; squalene synthase inhibitors [ 58 ], microsomal triglyceride transfer protein inhibitors [ 59 , 60 ], siRNA for PCSK9 [ 61 ] or for apolipoprotein B-100 [ 62 ] silencing, antisense PCSK9 [ 63 ], and antisense apolipoprotein B-100 (more commonly known as Mipomersen sodium (ISIS 301012)) [ 64 , 65 ].
In August 2010, Genzyme Corp. and Isis Pharmaceuticals Inc. announced the completion of the four phase 3 clinical trials that are required in the initial United States and European of regulatory filings for mipomersen. Filings for therapeutic use in homozygous FH are expected in the first half of 2011 [ 66 ]. These double-blinded, placebo-controlled clinical trials have been conducted at several locations worldwide. They involve heterozygous FH patients [ 67 ], homozygous FH patients [ 68 ], and patients with severe hypercholesterolemia [ 69 ]. The latter are defined by LDL-C levels ≥200 mg/dL and baseline cardiovascular disease (CVD) or by LDL-C levels ≥300 mg/dL without CVD. The trials also include patients with high cardiovascular risk [ 70 ] and high cholesterol levels as defined by LDL-C levels ≥100 mg/dL who were already taking maximally tolerated lipid-lowering medications.
At the end of the study, these patients had an average LDL-C reduction of 36-37% with no serious adverse effects. The reductions observed in the study were in addition to those achieved with the patients' existing maximally tolerated statin regimens. The trial also met each of its three secondary endpoints with statistically significant reductions in apo-B, non-HDL-cholesterol and total cholesterol. All trials also demonstrated manageable safety and tolerability profile of mipomersen.
Although each of these novel therapies effectively lowers LDL-C, challenges remain for clinical development in the assessment of long-term safety.
Liver directed gene therapy for FH
Patients who have undergone liver transplantation and have experienced substantial reductions in LDL-C levels provide indirect evidence that gene therapy targeted towards the liver could be effective for this disease. While LDLR is expressed by the majority of body cells, hepatic reconstitution of LDLR expression alone may be sufficient for metabolic correction [ 71 , 72 ]. The liver is an attractive organ for FH gene delivery because of its large mass, its ability to synthesise large amounts of proteins, its central position in metabolism and its good accessibility through the portal vein [ 72 , 73 ].
The homozygous form of FH would be an excellent candidate for gene therapy since the plasma lipid profile, total cholesterol, LDL-C, HDL-C and LDL/HDL ratio, can be measured providing a convenient clinical endpoint to evaluate the response to therapy [ 71 , 72 ]. In addition, a sensitive non-invasive method using a scintillation camera is available to determine the location, magnitude, and duration of LDLR transgene expression which could provide functional transgene expression in gene therapy trials of FH [ 74 ]. Moreover, animal models are available, which include the Watanabe heritable hyperlipidemic (WHHL) rabbit [ 75 ], and rhesus monkeys [ 76 ], the ApoE-knockout (ApoE-/-) mouse [ 35 ], and the LDLR-knockout (LDLR-/-) mouse models [ 77 ]. The WHHL rabbit demonstrates hypercholesterolaemia due to natural deletion of 12 nucleotides in the LDL-binding domain of the LDLRcDNA [ 78 ]. This causes a delay in the post-translational processing of the 120 kDa LDLR-precursor to the 160 kDa mature form, leading to premature degradation of the mature form in the cytoplasm and consequently hypercholesterolaemia (700-1200 mg/dl at 12 months of age) [ 79 ]. The WHHL rabbit, therefore, demonstrates metabolic and clinical abnormalities similar to those in patients with FH and may be a more authentic FH model than the LDLR-/- or ApoE-/- mouse models [ 80 ] where the raised plasma cholesterol levels (225 ± 27 mg/dl) are lower unless the animals are subjected to a high cholesterol diet. There are also some intrinsic differences in the lipoprotein metabolism of mice compared to humans and rabbits. For instance, the main lipoprotein in plasma of FH patients and the WHHL rabbit is LDL, but in ApoE-/- mice [ 81 ] it is the VLDL fraction with apoB-48, and HDL and LDL in LDLR-/- mice [ 77 ]. The activity levels of the plasma cholesterol-ester transfer protein (CETP), which facilitates the transport of cholesteryl esters and triglycerides between the lipoproteins, and hence plays a role in LDL particle remodelling, are high in WHHL rabbits, although murine models lack this activity [ 80 , 82 ]. Consequently, HDL levels in the plasma are low in WHHL rabbits but high in mice and rats. In contrast, the ApoB-editing enzyme is not expressed in the liver of rabbits [ 80 ], but murine models do have ApoB-editing activity in the liver [ 81 ]. Therefore, apoB-48-containing VLDL is secreted from the liver in mice [ 80 , 81 ]. Selective breeding of WHHL rabbits resulted in coronary atherosclerosis-prone WHHL rabbits manifesting with features of coronary and aortic atherosclerosis and myocardial infarction, in contrast murine models are usually resistant to the development of myocardial infarction and features of coronary and aortic atherosclerosis [ 80 ]. For the above-described differences, the WHHL rabbit is thought to be a more authentic FH model similar to human subject (reviewed in [ 80 ]).
Gene transfer systems
In addition to the method chosen for delivery, successful treatment of FH would ideally require safe and efficient gene transfer vectors that provide appropriate and sustained levels of transgene expression and long-term survival of treated cells. The use of a liver specific promoter would be the most physiological approach to achieve this. However, because of present problems in transfection-efficiency, strong heterologous promoters are commonly used instead for proof of principles studies on the effectiveness of lipid-lowering. The development of more effective vectors to achieve this remains a formidable challenge to gene therapy. The properties required of such a vector system and those that should be avoided are listed in Table 1 .
Gene transfer vectors are generally classified under two categories; they are either non-viral or virus mediated gene transfer systems.
Non-viral gene transfer systems
Gene therapy vectors based on modified viruses are unquestionably the most effective gene delivery systems in use today. Their efficacy at gene transfer is however tempered by their potential toxicity [ 85 , 86 ]. An ideal vector for human gene therapy should deliver sustainable therapeutic levels of gene expression without compromising the viability of the host (at either the cellular or somatic level) in any way. Non-viral vectors are attractive alternatives to viral gene delivery systems because of their low toxicity, relatively easy production and great versatility [ 87 ].
Most of the non-viral vectors that have been described for gene therapy are based on complimentary DNA (cDNA) gene sequences driven by highly active promoters. The DNA in these vectors is typically formulated with cationic agents to form complexes, which protect the DNA and facilitate cell entry [ 87 , 88 ]. DNA can, however, be driven into cells by physical means and the liver is particularly amenable to gene delivery via hydrodynamic delivery. Mahato et al reported that a standard tail vein injection of naked DNA into mice resulted in almost no gene expression in major organs due to its rapid in vivo degradation by nucleases and clearance by the monocular phagocyte system [ 89 ]. However, a very rapid injection of a large volume of naked plasmid DNA solution (e.g. 5-10 μg of DNA in 2.5 ml saline, which is almost equivalent to the blood volume of the animal, within 5-7 seconds) via the same route induced efficient gene transfer particularly in the liver [ 90 ]. This procedure was applied in one of the first non-viral approaches to reverse hypercholesterolaemia in an FH model. In these experiments a DNA construct was produced which encoded a fusion-protein consisting of a soluble form of the LDLR combined with transferrin. The strategy of this approach was based on the ability of the fusion protein to be capable of binding both plasma LDL and the cellular transferrin receptor. When applied in vivo following hydrodynamic injection [ 91 ], this protein was shown to bind circulating plasma LDL and to mediate its clearance through the transferrin receptor on hepatocytes. Although the system proved functional, a statistically significant change in the lipoprotein profile of an animal model was not demonstrated and the possible immunogenicity of the fusion protein potentially precludes its utility.
In contrast to using a cDNA expression cassette, the use of a complete genomic DNA locus to deliver an intact transgene with its native promoter, exons, all intervening introns, and regulatory regions with flanking non-coding genomic DNA sequences may allow regulated complementation of LDLR deficiency in the liver of hypercholesterolaemic animals. In 2003, a bacterial artificial chromosome containing the entire LDLR genomic locus and based on the Epstein Barr Virus (EBV)-retention system was delivered to LDLR deficient Chinese hamster ovary cell line (CHOldlA7) [ 92 ], and achieved correction of the cells' deficiency phenotype [ 93 ]. This vector construct was able to mediate LDLR expression at significant levels in the CHOldlA7 cells and in human fibroblasts derived from FH patients for 3 months and to retain the classical expression regulation by sterol levels in these cells. These initial studies paved the way for the development a more sophisticated vector which utilised a Scaffold Matrix Attachment Region (S/MAR) rather than a potentially toxic viral component to provide episomal maintenance [ 94 ]. In this study the LDLR genomic locus was incorporated into an HSV-1 amplicon vector, which was shown to remain episomal for 11 weeks and provided the complete restoration of human low density lipoprotein receptor LDLR function in CHOldlA7 cells to physiological levels. The vector comprised the LDLR gene driven by 10 kb of the human LDLR genomic promoter region including the elements, which are essential for physiologically regulated expression. By utilizing the genomic promoter region it was demonstrated that long-term, physiologically regulated gene expression and complementation of receptor deficiency could be obtained in culture for at least 240 cell-generations. Importantly, this vector was shown to be sensitive to the presence of sterols or statins, which modify the activity of the LDLR promoter. These in vitro studies finally lead to the successful administration of genomic LDLR vectors in vivo via hydrodynamic delivery [ 95 , 96 ]. When administered hydrodynamically in mice it was demonstrated that efficient liver-specific delivery and statin-sensitive expression could be obtained for up to 9 months following delivery [ 96 ].
While the majority of studies were focused on treating FH by inhibition of hypercholesterolaemia through up-regulation of LDLR or other surrogate lipoprotein receptors (as will be discussed later), an alternative approach was to down-regulate apoB-100 LDLR-ligand or PCSK9 [ 63 ] expression. Down-regulation of apoB-100 by continues intravenous/subcutaneous administration of Mipomersen antisense apolipoprotein B-100 oligonucleotide had been attempted in several clinical trials achieving average LDL-C reduction of 36-37%. Although, the most common adverse event of these trials was erythema at the injection site due to the protocol [ 64 , 65 , 97 - 102 ], challenges remain for clinical development in the assessment of long-term safety.
Recombinant virus-based gene transfer systems
Recombinant viral vectors are usually more effective than non-viral vectors in mediating cell entry and nuclear transfer of therapeutic genes in the target cells. In addition natural tropism of viral envelopes and serotypes can be employed to achieve targeting selectivity for particular host cells. Most of these vectors have mechanisms to avoid intracellular degradation and overcome cellular and immunological barriers to the delivery of the genetic cargo. Generation of a virus vector requires the transformation of a potentially harmful virus from a pathogen into a gene transfer agent whilst retaining the viral infectivity. Hence, the first step is to make the vector replication defective (incapable of producing infectious viral particles in the host's target cells). Replication deficient viral vectors are developed by deletion of crucial genes in the virus genome, which are then generally replaced by the therapeutic gene. The elements removed in this way have to be provided in trans in order to support vector production. This can be achieved by use of a helper virus or a packaging (production) cell line transfected with the plasmids expressing the genes coding for the required structural virus components and replication proteins. Helper virus must be purified away from the final vector batches intended for safe gene delivery.
Recombinant viral vectors presently used are generally classified under two categories; integrating or non-integrating viral vectors [ 103 ]. This distinction is an important determinant of the suitability of each vector for a particular application. Present integrating vectors rely on random insertion of the transgenic DNA into the cell's genome, leading to stable integration and subsequent passage to the cell's progeny. This gene insertion via non-homologous end joining of vector DNA to that of the host using a virus integrase can be most efficiently achieved using retroviral or lenti-retroviral vectors. Adeno-associated viral vector integration is less frequent than that of the unmodified parent virus, which targets preferentially into chromosome 19 and does not show the locus specificity of wild-type virus. It is important to note that integration does not guarantee stable transgene expression due to host mediated gradual silencing of gene expression over time [ 104 ], immuno-elimination or physiological cell death of gene modified cells.
Non-integrative vectors such as adenovirus and herpes simplex viral vectors allow transient episomal expression of a foreign gene in the target cells. Because of their episomal maintenance, the transferred genes are usually lost over time by dilution at cell division in actively dividing cells or by degradation in non-dividing cells [ 105 ]. A non-integrative vector could ideally be delivered repeatedly if required, as long as no immunological reactions develop against the vector or transgenic protein. Unlike adenovirus and herpes simplex viruses, EBV is stably maintained without integration in permissive proliferating cells due to the EBV nuclear antigen 1 protein-mediated replication and segregation, providing long-term transgene expression [ 106 ]. It is however unlikely to be used in a clinical setting due to its association with Burkitt's lymphoma. The properties of the most commonly used viral vectors are summarised in Table 2 .
Retrovirus based vectors
Retroviruses (RVs) are a large family of enveloped RNA viruses, which are generally classified into three subfamilies; oncoretroviruses, lentiviruses and spumaviruses (foamy viruses). The RV particle is composed of two copies of an RNA genome held together primarily by a sequence called the dimer linkage, which is termed the leader region and found in some cases in gag coding region. The genome is surrounded by a spherical or cylindrical shaped core and an enveloped glycoprotein, (Figure 2-A ). Upon infection, RVs are able to convert their RNA genome in the host cytoplasm to DNA through reverse transcription (RT) [ 107 ].
The genome size of simple RVs is approximately 8-12 kb and comprises three main genes; the group specific antigen encoding gene ( gag ), the polymerase encoding gene ( pol ), and the envelope glycoprotein encoding gene ( env ), which are flanked by elements called long terminal repeats ( LTR s), Figure 2-B . The gag gene encodes the viral structural core proteins, which form the matrix, capsid and nucleocapsid, generated by protease cleavage of the Gag precursor protein. The pol gene expresses a complex of enzymes that are involved in particle maturation (protease), DNA metabolism (reverse transcriptase) and proviral integration (integrase). These enzymes are usually derived from the Gag/Pol precursor. The env gene encodes the surface glycoprotein and the transmembrane protein of the virion, which form a complex that interacts specifically with cellular receptor proteins. The genes in the viral DNA are bracketed by the LTR s, which define the beginning and the end of the viral genome. The LTR s are identical sequences that can be divided into three elements. U3 is derived from a sequence unique to the 3' end of the RNA, R is derived from a sequence repeated at both ends of the RNA, and U5 is derived from a sequence unique to the 5' end of the RNA. The genesis of the LTR elements lies in the process of reverse transcription. U3 contains most of the transcriptional control elements of the provirus (viral genome, which has integrated into the chromosomal DNA of a cell), which include the promoter and multiple enhancer sequences responsive to cellular and in some cases viral transcriptional activator proteins. The site of transcription initiation is at the boundary between U3 and R of the 5' LTR and the site of poly(A) addition is at the boundary between R and U5 at the 3' LTR , as shown in Figure 2-C . The other boundaries of U3 and U5 are determined by the primer binding site ( PBS ) and the polypurine tract ( PTT ), which are important for reverse transcription. Just downstream of the 3' end of the 5' LTR , is a short packaging sequence (Psi,Ψ), which extends into gag and is responsible for encapsidation of the two viral RNA genomes into the capsid. The att sequences at the ends of the 5' and 3' LTR s are necessary for proviral integration [ 107 ].
The life cycle of a RV starts with high affinity binding of the viral envelope glycoprotein to its receptor on the outer layer of the cell membrane. This interaction leads to the fusion of the lipid envelope surrounding the virus, with the target cell membrane. Cell entry of the viral capsid containing the RNA genome allows the reverse transcriptase enzyme to copy the viral RNA genome into a double-stranded DNA, which becomes associated with viral proteins to form what is called a pre-integration complex (PIC). The PIC translocates to the nucleus where the viral enzyme integrase, which is part of PIC, mediates integration of the provirus DNA sequence into the chromosomal DNA of the host cell. The inserted sequence (provirus) is flanked by complete copies of LTR sequences. The 5' LTR drives transcription of the RV genome, which gives rise to RNA that codes for the viral proteins Gag, Pol and Env as well as for the viral RNA genome, Figure 2-C . Gag and Gag/Pol proteins assemble as viral core particles at the plasma membrane which package the viral RNA genomes and bud from the cell membrane enveloped with plasma membrane lipid from the host, in which virus derived Env glycoproteins are embedded [ 107 ].
The first generation replication defective retroviral vectors were developed using Molony Murine Leukaemia Virus (MoMLV) as a prototype. Replication defective vector particles were produced by a deconstruction strategy, which aims to dissect/segregate the viral genome into two transcriptional units (plasmid constructs). They are the vector genome and the packaging constructs, Figure 3 . The vector genome retains all the necessary cis elements of the vectors and is generated by replacing viral protein encoding sequences later to be provided in trans with the transgene of interest. Both the Ψ signal that is essential for packaging of the vector genome into the capsid and the viral LTR s, which are necessary for proviral integration, remain in the vector genome construct. Expression of the transgenic protein is driven by the promoter in the U3 region of the 5' LTR . The packaging construct provides all of the viral proteins in trans to the vector genome construct (Gag, Pol and Env). The packaging signal is deleted from the packaging construct to prevent its incorporation into viral particles. When both the vector genome and packaging constructs are present in a producer cell, retroviral vector particles, which are capable of delivering the vector genome with its inserted gene into new target cells, are released [ 108 ]. The process of gene transfer by such a vector is referred to as transduction. When these two constructs are present in cells in an integrated form, the cell becomes a stable virus producer. Alternatively, virus can also be produced for a short period of time after transient co-transfection of the viral genome alongside with packaging plasmids.
The basic arrangement described above is functional but unsatisfactory for several reasons. Firstly, the sequence overlap that remains between the vector and packaging constructs could result in recombination to form infectious replication competent retrovirus (RCR). The overlap exists principally because extensive sequences of the gag gene are retained in the vector construct to enhance the efficiency of packaging, although Gag protein production is prevented by mutation. In addition, overlapping sequences also exist because the LTR s are retained in the packaging construct to provide both promoter and poly-adenylation sequences. Secondly, the early MoMLV based vectors were established in murine NIH 3T3 packaging cell lines, therefore, the possibility for RCR generation through recombination between vector constructs and defective endogenous MoMLV-like sequences present in the target cells cannot be excluded [ 109 ]. Thirdly, vector particles produced in murine cells can be sensitive to host compliment mediated inactivation after in vivo gene delivery [ 110 , 111 ].
In order to minimise the risk of RCR production, an improved vector system was designed by segregating the gag / pol and env genes present on the packaging construct, onto discrete expression units, Figure 4 . The risk of recombination was also further reduced by the use of heterologous envelope proteins that are derived from alternative viruses with no homology to parental virus sequences but are still able to be incorporated into the viral particle (a process referred as pseudotyping). Pseudotyping may also alter the tropism of the viral vector and can be used as a powerful tool for cell targeting different host tissues. Pseudotyping of MoMLV and other RVs with the murine ecotropic (recognising only receptors present on mouse cells), amphotropic (interacting with receptors on both mouse and human cells) or the vesicular stomatitis virus glycoprotein (VSV-G) envelopes (with broad host range including mammalian and even insect cells) has been achieved and proven useful [ 112 , 113 ]. In this improved virus production system, part of the gag sequence present on the first generation MoMLV vector is removed from the vector genome without significant loss of packaging efficiency (it was subsequently found that part of the gag region is essential for efficient vector packaging) [ 114 ], Figure 4 .
The problem of overlapping sequences between the vector and the packaging construct has been solved by using heterologous promoters and polyadenylation signals to drive structural gene expression from the packaging constructs. Strong heterologous promoters like cytomegalovirus (CMV) can provide high virus titre production circumventing the limited titre offered by he MoMLV LTR s that give low-level gene expression in producer cell lines not of murine origin. In the vector genome construct itself, heterologous promoters have been used to replace the 5' U3 promoter. In addition, the 3' U3 sequences can be significantly deleted as long as the sequences necessary for recognition by the integrase protein are retained. This is the basis of self inactivating (SIN) vectors where deletion of the viral promoter and enhancer regions in the 3' U3 are duplicated during reverse transcription in the 5' LTR to prevent LTR -driven transcription in infected host cells which could result in the expression of downstream inserted proto-oncogenes [ 115 ]. Transgene expression in these vectors is therefore typically and exclusively driven by an internal heterologous promoter, which allows the use of regulated and/or tissue specific expression. Finally, a non-murine producer cell line was used for vector production to prevent the possible generation of RCR through recombination with endogenous MoMLV-like sequences [ 110 ].
In the latest generation of RV based-vectors, improvements have also been made in the vector titre (number of colony-forming units per ml) by the development of transient plasmid co-transfection systems, which are capable of producing very high vector titres for a short period of time in the highly transfectable HEK 293 (human embryonic kidney epithelial cells) cell line [ 110 ]. Also some human cells used to generate packaging cell lines can produce a complement-resistant retroviral vector [ 111 ]. Transfection of HEK 293T cells using SV40 large T antigen to improve vector load and hence vector titre are used also to circumvent the cytotoxicity of the highly desirable VSV-G envelope that provides broad host range infection.
Recombinant MoMLV-based vectors produced by the strategy described above are efficient gene transfer vehicles, reaching transfer levels in vitro of close to 100%. They can be produced at a high titre (10 9 infectious units (lU)/ml) and have the capacity to infect a wide variety of dividing cells including hepatocytes. The RV vector genome can also provide transfer of RNA of approximately 7.5 kb in length.
The critical limitation to the use of RVs is their inability to infect non-dividing cells and as the liver is an only slowly proliferating tissue these vectors are not ideal for LDLR gene delivery to hepatocytes. Therefore, for direct in vivo transduction of the liver, cells have to be either in a naturally dividing state or to be induced to divide. Alternatively, the vectors can be used for ex vivo treatment.
Hypercholesterolaemia has been ameliorated by RV-based vectors using ex vivo gene delivery in numerous experimental studies. The original procedure used for liver-directed gene therapy of FH was based on the ex vivo approach, which involved re-infusion of autologous hepatocytes that had been removed from a WHHL rabbit and subjected to in vitro genetic correction with RV vectors based on MoMLV. Animals transplanted with LDLR transduced celIs demonstrated a 30-50% reduction in total serum cholesterol levels persistent for the duration of the experiment (122 days). Recombinant derived LDLR mRNA was detected in liver cells for 6 months. There was no apparent immunological response to the recombinant derived rabbit LDLR [ 116 ]. This study illustrated the potential of the ex vivo approach to ameliorate hyperlipidaemia associated with FH using a RV-based vector.
In preparation for human trials with RV-based vectors, the efficacy, safety and feasibility of ex vivo gene therapy for FH was further documented in non-human primates [ 117 , 118 ]. Three baboons were subjected to a partial hepatectomy and their hepatocytes were isolated, cultured, and transduced with a RV containing the human low-density lipoprotein (hLDLRcDNA) sequence. Infusion of the genetically modified hepatocytes was performed through a catheter that had been placed into the inferior mesenteric vein at the time of liver resection. The baboons tolerated the procedures and were monitored for up to eight months [ 117 ]. The safety and efficacy of the ex vivo approach for delivery of gene transduced hepatocytes via the mesenteric circulation was further documented in a canine model [ 118 ].
The above studies demonstrated the feasibility and safety of the ex vivo approach, which was then carried out on a human patient in the first clinical trial for FH published in 1994. In this trial a 29 year-old woman with a homozygous receptor defective FH was subjected to ex vivo gene therapy using an amphotropic RV-based vector expressing human LDLRcDNA under control of the CMV enhanced chicken β-actin promoter. The patient tolerated the procedure and in situ hybridisation of liver tissue four months after therapy revealed evidence for engraftment of transgene expressing cells. The patient's LDL/HDL ratio declined from 10-13 before vector delivery to 5-8 after vector delivery, an improvement that remained stable for the duration of the reported observation (18 months). However, kinetic studies of LDL metabolism including LDL binding, uptake and degradation were not presented [ 119 ]. This trial was severely criticised with respect to both the suitability of the patient for this therapeutic intervention and for the aggressiveness of the protocol, which involved a 25% hepatectomy [ 120 ].
Grossman et al then reported four additional homozygous FH patients subjected to a surgical resection of the left lateral segment of the liver and re-infusion of the genetically modified hepatocytes [ 121 ]. The patients tolerated the infusions of autologous hepatocytes well without complications. Liver biopsies performed four months after treatment revealed LDLR transgene expression in a limited number of hepatocytes by in situ hybridisation in all four subjects. One of four patients had a significant and prolonged reduction of about 20% in his LDL-C levels. Kinetic studies of the LDL metabolism demonstrated that LDL catabolism was increased in the same patient, which was consistent with increased LDLR expression [ 121 ]. The reason for the only marginally successful lowering of cholesterol levels and the variable metabolic responses observed in the five subjects studied are presumably due to low gene transfer efficiency or low expression levels [ 121 ]. The variable metabolic response observed following low-level genetic reconstitution in the five patients precluded a broader application of ex vivo liver-directed gene therapy with RV based vectors, pending improvement of vector efficiency. The following sections review the preclinical work towards this goal with alternative vector system.
Adenovirus based vectors
Adenoviruses (Ads) are icosahedral particles consisting of linear, double stranded DNA with a non-enveloped virion, (Figure 5-A ). There are at least 50 different human adenovirus serotypes with an approximate genome size of 36 kb. The Ad genome (Figure 5-B ) is intimately associated with viral proteins (core) and is packaged in the viral capsid, which consists primarily of three proteins; hexon, penton base and fibre-knob. After infection, the virus genome does not integrate into the host chromosomal DNA, instead it is replicated as an episomal (extra-chromosomal) element in the host nucleus [ 122 ].
Because of the ability of adenoviral vectors to infect a broad range of mammalian cell types regardless of their replication status, they have been widely used for a variety of gene transfer applications in vitro [ 123 ], in vivo [ 124 ] and in clinical trials [ 125 ]. Most adenoviral vectors currently used are derived from serotypes 2 or 5, which are endemic and cause upper respiratory tract infection in humans. Most human individuals have become immune-sensitised by natural infection during childhood [ 83 ].
Vectors derived from serotypes 2 and 5 enter the cells after attachment to the cellular receptor CAR (coxsackievirus and adenovirus receptor), through the knob of the fiber [ 126 ]. Virus entry occurs then through cIathrin-mediated endocytosis after binding of the penton base to integrins [ 127 ]. It is noteworthy that differences in the tropism of various Ad serotypes indicate that besides CAR, other cellular receptors also contribute, suggesting that the host range of Ad vectors can be altered by use of alternative serotypes.
The first generation of replication deficient Ad vectors was constructed by replacing one or two viral early (E1 and E2) genes, which are essential for viral replication, with the transcriptional cassette of interest containing an enhancer-promoter element and the desired gene. Vectors in such a configuration have a packaging capacity of 6.5-8.3 kb. The recombinant vectors are replicated in cells that express the products of the E1 and/or E2 genes. Purified high titre stocks of 10 11 -10 12 Ad particles per ml, can be generated and allow high efficiency Ad mediated gene transfer with strong tropism for the liver. Cells that were transduced with these vectors express adenoviral genes at low levels, in addition to the transgenic protein [ 128 ].
The utility of replication defective first-generation recombinant Ad to mediate hLDLR gene transfer in hepatocytes derived from FH patients was first examined and documented in 1993 [ 123 ], using the β-actin promoter. The level of recombinant-derived LDLR protein in transduced FH hepatocytes exceeded the endogenous levels by at least 20-folds.
Reversal of hypercholesterolaemia was then demonstrated in LDLR-/- mice fed with a high cholesterol diet after intravenous injection of a replication-defective Ad encoding the hLDLR driven by CMV promoter. This in vivo approach resulted in reduction of the elevated intermediate density lipoprotein (IDL)/LDL ratio to normal levels, four days after vector delivery [ 129 ]. Similarly, injection of a replication-defective Ad encoding the hLDLR driven by an optimised CMV promoter into the portal vein of WHHL rabbits, resulted in over-expression of hLDLR in the majority of hepatocytes that exceeded the levels in normal human liver by at least 10 fold. Transgene expression was stable for 7-10 days but diminished to undetectable levels within three weeks [ 130 ]. Similar studies were also conducted on WHHL rabbits with Ad vectors containing rabbit LDLRcDNA [ 131 ] or human LDLRcDNA [ 132 ]. These studies also resulted in strong but transient transgene expression. However, the high level of LDLR expression and substantial reduction of total and LDL cholesterol achieved by adenovirus LDLR gene transfer in these animal models led to a massive intracellular lipid (cholesterol and cholesterol ester) deposition in transduced cells [ 130 , 133 ]. This accumulation resulted from non-physiological over-expression of LDLR mediated by the Ad vector, causing pathological intracellular accumulation of the lipid that could not be compensated by the hepatic cell metabolism [ 133 , 134 ].
The transient expression was not solely due to the episomal nature of Ad infection but also a result of host immune responses against adenoviral proteins [ 124 , 135 , 136 ]. Co-administration of an Ad vector encoding hLDLR driven by a CMV promoter, with a blocking antibody directed against CD154 (CD40 ligand) to suppress immune responses against the vector and foreign transgene product in LDLR -/- mice, resulted in long-term expression of LDLR and maintained cholesterol levels within and below the normal range for at least 92 days post vector delivery. The loss of hLDLR expression in non anti-CD154-treated mice also demonstrated the importance of the host immune response against vector and transgene products [ 137 ].
In direct response to these immunological reactions and vector cytotoxicity, helper dependent adenovirus (HD-Ad) vectors were developed, in which additional viral coding sequences were deleted [ 138 ]. This also increases the insert capacity of the vector to approximately 30 kb. Nomura and colleagues [ 139 ] compared the efficiency of monkey LDLR gene therapy with that of monkey very low density lipoprotein receptor (VLDLR) gene therapy, using HD-Ad. High cholesterol diet fed LDLR-/- mice were injected with a single intravenous application of high (1.5 × 10 13 vector particles (vp)/kg) and low (5 × 10 12 vp/kg) doses of HD-Ad. Throughout the 24-week experiment, plasma cholesterol of LDLR-treated mice was lower than that of VLDLR-treated mice. Anti-LDLR antibodies developed in 2 of 10 mice treated with high-dose HD-Ad-LDLR but in none (0/14) of the other treatment groups. The antibody titre in the high-dose experiments was significantly above background, but was three orders of magnitude lower than that seen following first generation Ad-LDLR treatment, indicating that the marked pro-inflammatory adenoviral protein expression following FG-Ad-LDLR gene transfer could have acted as an adjuvant that stimulated antibody production in these mice. Long-term efficacy of low-dose HD-Ad-LDLR injected into 12-week old LDLR-/- mice was tested and after 60 weeks, atherosclerosis lesions covered approximately 50% of the surface of aortas of control mice whereas aortas of treated mice were essentially lesion-free. The lipid lowering effect of HD-Ad-LDLR lasted at least 108 weeks (>2 years) when all control mice had died [ 139 ].
Despite the reported improvements achieved by HD-Ad, the cytotoxic effect resulting from immune response to high titre (3.8 × 10 13 lU/mI) administration of a 2 nd generation adenoviral vector, which led to the unfortunate death of a patient in a non-FH clinical trial [ 125 ] stopped any further in vivo adenoviral vector delivery trials, pending improvement in vector design. In an attempt to address this issue, Jacobs and colleagues investigated the use of a relatively low dose (5 × 10 10 particles) of second generation E1E3E4-deleted adenoviral vectors for transfer of the LDLR or VLDLR, under control of the hepatocyte-specific human α 1 -antitrypsin promoter and 4 copies of the human apo E enhancer, into C57BL/6 LDLR-/- mice [ 140 ]. Evaluation was performed for 30 weeks after vector delivery in male and female mice fed either standard chow or an atherogenic diet. Compared to control mice, AdLDLR and AdVLDLR persistently decreased plasma non-HDL cholesterol in both sexes and on both diets and potently inhibit development of atherosclerosis in the ascending aorta. The non-physiologically regulated over-expression of LDLR or VLDLR, transferred by E1E3E4-deleted adenoviral vectors, significantly reduces tissue cholesterol levels in myocardium, quadriceps muscle, and kidney and does not lead to pathological intracellular accumulation of cholesterol and cholesterol esters in hepatocytes. The effectiveness of the vectors and expression cassette used in this study is stressed by the fact that, using vector doses that are 2-7.5-fold lower compared to those in other studies [ 139 , 141 ], equivalent results were obtained in terms of lipid lowering and reduction of atherosclerosis [ 140 ]. However, immune response to the vector system to evaluate potential development of neutralizing antibody or immune rejection to the transgene and/or vector has not been shown.
Adenoviral based vectors still remain the most efficient class of vector in terms of delivering to and expressing their genetic cargo in the cells of most tissues. However, because of their transient expression characteristics, while they remain useful for proof of principle for gene therapy they are not the vector of choice for the treatment of inherited monogenic diseases but will probably find application in the treatment of cancer in which cellular toxicity and immunogenicity might even enhance their anti-tumour effects [ 142 ].
Adeno-associated virus vectors
Vectors based on adeno-associated virus (AAV), a small (20-25 nm) non-enveloped DNA virus (Figure 6 ) that is non-pathogenic and replication-defective, have a number of attributes that make them suitable for gene transfer to the liver for the treatment of FH. A single administration of recombinant AAV (rAAV) into the liver results in long-term transgenic protein expression without toxicity in a variety of animal models [ 143 ]. These pre-clinical studies have lead to phase I/II trials of liver gene transfer for diseases such as haemophilia [ 144 ] for example, using AAV serotype 2, the first isolate to be characterised.
There are several current obstacles to AAV gene therapy that need to be addressed. Although AAV is not known to cause human disease, 85% of the adult population is sero-positive for AAV capsid proteins [ 145 ] and wild type (wt) AAV2 is endemic to humans. Thus most of the patients that participated in clinical trials are likely to have had pre-existing immunity to the serotype employed, as a result of prior natural infection. Cytotoxic T-cells resulting from wt-AAV infection can eliminate transduced cells and anti-AAV2 antibodies are able to block or reduce gene transfer with rAAV2 vectors. These factors may have limited transgenic hFIX protein expression in a recent hemophilia B gene therapy trial [ 144 ]. Switching the capsid protein to other AAV serotypes that are less prevalent in humans can overcome these immunological problems [ 146 ]. There are several AAV serotypes available that may prove useful in the future for clinical translation.
Currently a high multiplicity of infection is needed to achieve therapeutic AAV mediated gene transfer. Efficient transduction of target cells is blocked at several levels during AAV cell infection and movement of the vector into the nucleus. The host gender appears to be an important consideration, since in mice exogenous androgens can increase stable hepatocyte gene transfer in females to levels observed in male mice [ 147 ]. Strategies such as blocking endosomal degradation of AAV with proteasome inhibitors significantly improve AAV transduction in mice [ 148 , 149 ]. Switching AAV capsid proteins to an alternative serotype such as AAV8 can also enhance uncoating of the vector and release of the genome [ 146 ].
The dsDNA genome of the AAV vector can persist as an episomal element in transduced cells for long periods of time in a variety of molecular forms, including circular monomers, linear monomers and linear concatemers by head to tail recombination of the ITRs. Integration of single and concatemeric genomes into the chromosomal DNA of the host cells occurs at low frequency [ 103 ]. Because the transgene is predominantly expressed from the episomal form, expression usually declines over time due to dilution in the replicating cells or degradation in non-dividing cells [ 105 ]. A recent study found that administration of Ad 10-20 weeks after AAV gene transfer augmented AAV transgene expression two-fold by increasing the level of pro-viral mRNA [ 149 ] and this strategy may prove useful in clinical practice when transgenic protein expression levels fall.
Slow conversion of the virus single stranded (ss) to the double stranded (ds) DNA genome is another issue. After the AAV virus enters the nucleus, the virus single stranded DNA genome (ssDNA) is converted to a transcriptionally active double stranded DNA (dsDNA) [ 150 ]. Unless the conversion happens, ssDNA is lost rapidly after transduction, leading to a drop in transgenic protein expression. This rate-limiting conversion process can be circumvented by modifying the configuration of the provirus so that it is packaged as complementary dimer as opposed to the conventional ss [ 151 ]. This self-complementary (sc) AAV vector configuration has been shown to significantly improve gene transfer to the liver for human factor IX, achieving levels of stable transduction that are almost one order of magnitude higher than those achieved with an equivalent dose of comparable ssAAV [ 151 , 152 ]. Lowering the required dose of scAAV vector would be of benefit for safety considerations and for scaling up to clinical grade vector production. Modifying the promoter can alter the tissue-specific expression. Use of the liver-specific promoter, LP-1 for example in a self-complementary AAV2/3 vector driving the human factor IX (hFIX) protein, resulted in transgenic h(FIX) protein expression confined to the liver as detected by RT-PCR analysis [ 152 ]. This would be beneficial for FH gene therapy.
One of the earliest studies on AAV vectors for FH gene therapy found promising results. Reversal of hypercholesterolaemia was demonstrated in LDLR-/- mice fed with a high cholesterol diet after intraportal vascular injection of 1 × 10 12 AAV-2 vector particles encoding the murine VLDLR driven by the CMV enhanced chicken β-actin promoter [ 153 ]. Western blot analysis and immunohistochemistry revealed high levels of VLDLR expression in approximately 2-5% of cells of liver harvested at 3 and 6 months after vector delivery with a low vector DNA copy number of 1 copy/cell. Serum cholesterol progressively declined after vector administration and by 6 months, the aortic atherosclerotic lesion area was reduced 33% compared with control mice injected with saline. Phenotypic correction was incomplete however, primarily due to immune activation by the vector products and low efficiency of gene transfer mediated by AAV-2.
Lebherz and colleagues [ 154 ] compared the efficiency of AAV-2, -7 and -8 serotype vectors carrying the human LDLRcDNA expressed from a liver specific promoter based on the human thyroxin binding globulin [ 155 ]. A vector dose of 1 × 10 12 genome copies (gc) per mouse was injected into the portal veins of LDLR-/- mice that were fed a high-fat diet. Transduction efficiency was increased to 50 gc/cell and 10 gc/cell after treatment with an AAV-8 or AAV-7 vector respectively, compared with 2 gc/cell after administration of an AAV-2 vector. Animals receiving the AAV-LDLR serotype 7 and 8 achieved nearly complete normalization of serum lipids and failed to develop the severe atherosclerosis that characterized the untreated animals, with no apparent toxicity observed. Animals treated with the AAV-2 vector achieved partial lipid correction and only a modest improvement in atherosclerosis. Serotype 8 virus achieved stable transduction and expression of the transgene in up to 85% of the hepatocytes. These results are encouraging especially since no expression-terminating immune responses were detected [ 154 ]. There were similar findings in the apo-E mouse model of FH, where intravenous administration of AAV2/7- and AAV2/8-apoE vectors completely prevented atherosclerosis at 1 year [ 156 ].
Another approach using AAV vectors has been to try to counteract the development of atherosclerosis by gene transfer of interleukin-10 (IL10), an anti-inflammatory cytokine. Injection of AAV IL10 vector into the tail vein [ 157 ] of LDLR-knockout mice or into the tibial muscle [ 158 ] of apo-E deficient mice resulted in significantly lower levels of atherosclerosis.
More recently, a single intravenous injection of an AAV8 vector containing the mouse LDLR gene to a humanized mouse model of FH, the LDLR-/-Apobec-/- mouse, was found to significantly reduce plasma cholesterol and non-HDL cholesterol levels in chow-fed animals at low doses. Treated mice realized an 87% regression of atherosclerotic lesions with substantial remodeling, after 3 months compared to baseline mice [ 159 ].
In summary, modifying the AAV vector system by altering the capsid (reviewed in [ 160 ]), including dsDNA and using a liver specific promoter may result in long term, stable and liver specific AAV mediated transgenic protein expression which may be suitable for FH gene therapy.
Lentivirus based vectors
Lentiviruses (LVs) are a complex sub-group of RVs responsible for a variety of immunological and neurological diseases. Their biological and molecular and properties have been used to classify them as lenti-(sIow) retroviruses. They can be subdivided into primate and non-primate viruses. The primate viruses are the human and simian immunodeficiency viruses (HIV and SIV), and the non-primate viruses include the feline and bovine immunodeficiency viruses, the caprine arthritis/encephalitis virus, the visna/maedi/ovine progressive pneumonia virus, and the equine infectious anaemia virus (EIAV) [ 107 ]. As for all RVs, the LV genome consists of a positive-strand polyadenylated RNA of about 10 kb and includes three genes; gag , pol , and env organised in the 5' to 3' orientation. Lentiviruses have additional unique small ORFs located between pol and env at the 3' terminus, which contain genes for regulatory proteins [ 107 ].
Interest in LVs as putative gene transfer systems is derived from the fact that they have the potential to integrate efficiently into the genome of dividing and non-dividing cells providing the possibility for lifetime correction with a single administration of vector [ 161 , 162 ]. Unlike the RV pre-integration complex, which can only reach the target cell nucleus when the nuclear membrane is disrupted during mitosis, the lentiviral PIC contains nuclear localisation signals, which mediate their transport through nuclear membrane pores into the nucleus during the cell interphase [ 163 - 165 ].
Although integration of linear DNA episome, provirus precursor, is generally regarded as the end point of gene transfer, two circular episomal types with intact viral coding regions are also generated by cellular proteins from retro- or lenti-viruses and their derived vectors [ 166 ]. The first type circularizes by non-homologous recombination of end-joining to form a circular episome with two adjacent LTRs (2-LTR circular episome) [ 167 ]. The second type circularizes by homologous recombination within the LTRs to form a circular episome with a single LTR (1-LTR circular episome) [ 168 ]. It has been estimated that approximately one-third of linear lentiviral DNA become circular episomal forms and can express proteins and remain metabolically stable and transcriptionally competent in target cells, although, the single LTR circular episomal forms are more prevalent than 2-LTR circles [ 166 ].
Using a LV backbone, two types of vector system can be produced and used for gene transfer, the first are integrated lentivirus based vectors (ILV) and the second are integration deficient lentiviral vectors (IDLV). Initial research on the development of lentivirus-based vectors has focused mainly on HIV-1 derived integrated LV vectors as prototype. This is facilitated by the abundance of knowledge that has been accumulated on this virus since its recognition in 1984 as the causative agent of acquired immuno-deficiency syndrome. Like other virally derived vectors, the initial problem to overcome is to maintain viral infectivity but to render the virus replication deficient [ 161 , 162 ]. The LV based vector design is very similar to that of the three-plasmid co-transfection RV system based on MoMLV, described above. In addition, the emergence of a host immune response against lentiviral vectors has not been shown in most of the preclinical studies [ 169 - 173 ].
Due to the increased concern of insertional mutagenesis (IM) caused by integrating retro- and lentivirus based vectors (as will be discussed later), IDLVs has been thought of as a logical alternative to alleviate the risk of IM. IDLV particles can be generated by the use of integrase mutations that specifically prevent proviral integration resulting in the generation of increased levels of circular vector episomes in transduced cells, but not to compromise its other functions, Because these lentiviral circular episomes lack replication signals, they are gradually lost by dilution in the transduced actively dividing cells, but are stable for several months in transduced quiescent cells [ 174 - 176 ]. Compared to integrating lenti-vectors, IDLVs have a significantly reduced risk of causing IM, a lesser risk of generating RCRs, a reduced risk of transgene silencing [ 177 ], and also extremely low levels of integration (residual background integration frequencies of IDLVs in cultured cells through non-integrase pathways are within the range described for plasmid transfection (reviewed in [ 166 , 178 ]).
Recent studies using IDLVs have demonstrated effective gene transfer in the eye [ 176 ], brain [ 174 , 179 ], muscle [ 180 ], and to a lesser extent in the liver [ 181 ], albeit at lower expression levels than with integrating vectors. In addition to gene transfer, IDLVs are also proficient vectors for gene repair and can be converted into stable, replicating circular episomes. These properties, combined with their highly reduced risk of causing IM, have led to increased interest on IDLVs for gene transfer and therapy. Because of the possibility of mobilization by superinfection with replication competent viruses, it has been suggested that future IDL-based vectors should carry att mutations in addition to those in the integrase to minimize integration in the event of vector mobilization (reviewed in [ 166 , 178 ]). Long-term evidence for lack of genomic integration beyond residual levels warrants future investigation. To date, IDLVs have not been used for LDLR gene transfer and FH gene therapy.
Many labs including ours [ 182 ] experienced difficulties to produce infectious ILVs for transfer and expression of human LDLR under control of a ubiquitous promoter. However, based on the utilisation of a previously characterised liver specific promoter (LSP) [ 183 ], Kankkonen and colleagues were able to demonstrate for the first time the successful construction and production of high titre (1 × 10 9 IU) third-generation HIV-1 based lentiviral vectors encoding rabbit LDLR. LSP-driven transgene expression was detected after in vitro gene transfer into human hepatoma (HepG2) cells, but not after transfer into HeLa cells, HEK 293 cells, or WHHL rabbit skin fibroblasts [ 183 , 184 ]. In vivo injection of 1 × 10 9 infectious virus particles into the portal vein of WHHL rabbits resulted in liver-specific expression of the LDLR and clinical chemistry and histological analyses showed normal liver function and morphology during the 2-year follow-up without safety issues. This vector dose resulted in low transduction efficiency (<0.01%) but demonstrated on average a147 ± 7% decrease in serum cholesterol levels during the first 4 weeks, a 44 ± 8% decrease at 1 year and a 34 ± 10% decrease at the 2-year time point, compared to the control rabbits infected with HIV-green fluorescent protein. During this period, 70% of the rabbits treated with the liver specific lentiviral LDLR vector demonstrated a positive treatment effect with lowered plasma cholesterol levels (25 ± 8%). However, the detailed pattern of bio-distribution after HIV-vector mediated gene transfer, to evaluate potential risks for possible IM and germ-line transmission, has not been investigated.
Vector safety in gene therapy
The integration of RV and LV into the genome during gene therapy has caused concern because of the potential for vector-related deleterious side effects on the host. This is, in part, due to the fact that vector insertion occurs in a semi-random manner into actively transcribed genes. For RV vectors insertion preference is for gene promoter regions [ 185 - 188 ] whereas LVs appear to target the transcription unit of the gene [ 189 , 190 ] and therefore are believed less likely to cause effects on host gene expression following integration [ 191 - 194 ]. Genotoxicity by RV vectors associated with insertional mutagenessis (IM) has been studied for several years and the theoretical calculated estimates of mutagenesis at a haploid locus are supported by in vitro studies using model systems based on mutagenesis of the hprt locus or genes that control promotion of growth factor independence at frequencies between 10 -5 -10 -7 per provirus insertion [ 191 , 192 ]. Hence, the likelihood of adverse events caused by RV integration following therapeutic application was considered remote. Unfortunately and unexpectedly, however, development of clonal dominance has been observed in two patient trials that is attributed to RV mediated IM [ 86 , 195 - 199 ].
In an ex vivo trial carried out in France that used patients' own haematopoietic stem cells for transplantation after retroviral transduction to correct X-linked severe combined immuno-deficiency (X-SCID), clonally dominant clones have developed into leukaemias in 4 of these patients [ 85 , 86 , 196 ]. This also occurred in one patient in a British X-SCID trial [ 198 ]. In 4 of these cases integration by MoMLV is believed to have caused IM by inserting near the LMO2 gene [ 196 , 198 ]. In addition, insertions have been found in both BMI1 and CCND2 proto-oncogenes [ 196 , 198 ]. Although 5 out of the 20 patients that enrolled in the French and British trials have developed leukaemia it is difficult to understand clearly the events leading to this disease because of existing genetic abnormalities in the patients' cells that have also been identified. These include chromosomal translocations, gain-of-function mutations activating NOTCH1 , deletion of tumour suppressor gene CDKN2A , 6q interstitial losses, and SIL-TAL1 rearrangement [ 196 , 198 ].
In a more recent trial for chronic granulomatous disease (CGD) clonal dominance has also been attributed to retrovirus mediated IM 5 month after vector delivery in 2 patients [ 199 ]. Vector integrations activated the zinc finger transcription factor homolog's MDS1/EVI1, PRDM16 or SETBP1 raising concerns that this could eventually cause tumourgenesis. The first affected patient died 2.5 years after vector delivery as a result of a severe sepsis and the second patient has undergone allogeneic transplant [ 199 , 200 ].
In response to these findings, ex-vivo and in vitro models have been developed in order to examine RV and LV genotoxicity using haematopoietic cells. Ex vivo gene therapy using stem cells is considered a more controllable way of introducing genetic modification to the host than by direct systemic vector administration in vivo [ 201 - 205 ]. These models have confirmed that insertion of RV, and to a lesser extent SIN-RV and LV can contribute to leukaemic development [ 85 , 201 - 206 ]. Factors implicated in this process include the integrated vector copy number, integration sites, vector configuration and even the transgene carried by the vector [ 85 , 201 - 206 ]. Most recently, host cell transcription, in combination with the mutational potential of the vector, has been shown to be involved in the emergence of clonal dominance [ 206 , 207 ].
In our laboratory we have developed a model more suited to gene therapy for FH where vectors may be delivered directly in vivo. In this model vector application in utero is performed via the fetal mouse circulation that results in gene transfer to most organs, although the liver is mainly transduced [ 208 ]. We found that using a primate HIV-1 based vector carrying the human factor IX (hFIX) gene to correct haemophilia in a knockout mouse model of this disease comprehensive cure was achieved without adverse effects, however, the use of a non-primate EIAV vector driving hFIX gene expression led to a high frequency of liver tumours in these mice [ 209 ]. This model is still under development, and we have also obtained similar results with the non-primate feline immuno-defificiency (FIV) vector (Themis et al . unpublished data). Most importantly in these tumours, we find insertions within genes assigned as candidate genes involved in cancer development (within a 100 kb integration site window - the theoretical distance by which vector insertion is believed to influence expression of a gene carrying the integrated vector). More than 50% of these genes are registered in the Mouse Retroviral Tagged Cancer Gene Database (RTCGD) [ 210 ]. Furthermore, many genes carrying insertions have altered gene expression suggestive of IM by the non-primate LV. Hence, using in utero gene delivery where genes are in a highly active transcription state, we are able to sensitively detect adverse effects caused by vector integration.
The current models for vector associated genotoxicity all rely on the use of rodent cells as a measure of IM. As these cells are more predisposed to tumour development than human cells, each must be viewed with caution as reliable predictors for mutagenesis occurring in the clinic. The finding of vector genotoxicity in the clinic has, however, revived the use of models of genotoxicity to obtain useful information regarding safe vector design. They may also help to elucidate possible mechanisms relating to IM. In summary, the importance of genotoxicity assays to understand the cause and measure the risk of adverse effects by gene therapy of FH and indeed the treatment of any disease with these vectors cannot be overstated. With the current genotoxicity assays in place we are becoming more confident that gene therapy to FH homozygotes will be possible with minimal side effects.
Competing interest Disclosure
The authors declare that they have no competing interests.
Authors' contributions
FAA is responsible for conceiving this work and writing the manuscript. CC, SW, ALD, RH, and MT participated in the drafting of this manuscript. All authors read and approved the final manuscript.
Author's information
F. A. A. is an Assistant Professor of Genetics and Molecular Medicine, Department of Medical Genetics, Faculty of Medicine, Umm Al-Qura University, Al-Abedia Campus, Makkah 21955, Saudi Arabia. C. C. is an Emeritus Professor and Former Leader, Gene Therapy Research Group, Department of Molecular and Cell Medicine, Sir Alexander Fleming Building, Faculty of Medicine, Imperial College London, London SW7 2AZ, UK. S. W. is a Lecturer and Group Leader, Prenatal Gene Therapy Research Group, Department of Haematology, Haemophilia Centre and Haemostasis Unit, Royal Free and University College Medical School, London NW3 2PF, UK. R. H. is a Research Fellow and Gene Therapy Group Leader in the Section of Molecular Medicine in the Sir Alexander Fleming Building, NHLI, Imperial College London, London SW7 2AZ, UK. A. L. D. is a Senior Lecturer and Honorary Consultant in Obstetrics and Maternal/Fetal Medicine, and leads the Prenatal Cell and Gene Therapy Group, Institute for Women's Health, University College London and UCLH, 86-96 Chenies Mews, London, WC1E 6HX, UK. M. T. is a Lecturer and Group Leader, Gene Therapy and Genotoxicity Research Group, Brunel University, Heinz Wolff Building, Uxbridge, Middlesex, West London UB8 3PH, UK. | Acknowledgements
This work was supported by joint grants from the King Abdulaziz City for Science and Technology (KACST), the Saudi Basic Industries Corporation and Umm-Alqura University, Kingdom of Saudi Arabia. AD is receiving funding from the Department of Health's NIHR Biomedical Research Centres funding scheme. | CC BY | no | 2022-01-12 15:21:44 | Int Arch Med. 2010 Dec 13; 3:36 | oa_package/f5/b2/PMC3016243.tar.gz |
||
PMC3016244 | 21143947 |
In 1971, Judah Folkman published in the "New England Journal of Medicine" a hypothesis that tumor growth is angiogenesis-dependent and that inhibition of angiogenesis could be therapeutic [ 1 ]. This article also introduced the term antiangiogenesis to mean the prevention of new vessel sprouts from being recruited by a tumor.
In the last 35 years, it has been estimated that > 200 companies have worked and are still working in the area of angiogenesis and several of the compounds that modulate angiogenesis are currently being evaluated in clinical trials. A list of approved antiangiogenic drugs approved for clinical use is available in the Table 1 .
Even if the majority of pre-clinical studies have shown that the growth of all experimental tumors can be effectively inhibited by various antiangiogenic agents, the clinical benefits of antiangiogenic treatments are relatively modest, and in the majority of cases, the drugs merely slow down tumor progression and prolong survival by only a few more months.
The most promising antiangiogenic agents that are in clinical development at this moment include bevacizumab, the humanized anti-monoclonal antibody anti-VEGF approved for use in combination with cytotoxic agents [ 2 ], as well as small molecules receptor tyrosine kinase inhibitors (RTKIs), approved as single agents, and including sunitinib, an oral inhibitor of VEGFR-2, PDGFR, FlLT-3, and c-KIT, and sorafenib, an inhibitor of the Faf/MEK/Erk and the VEGFR and PDGFR signaling pathways. These agents are generally well tolerated, but the treatments may be accompained by distinct adverse effects, including hypertension and proteinuria.
In a communication in the 2003 ASCO Meeting, Hurwitz and co-workers reported that bevacizumab/IFL (irinotecan/fluorouracil/leucovirin) combination led to a significantly prolonged survival and had a better ability to shrink tumors that IFL alone. These results led the FDA to approve the use of bevacizumab in patients with metastatic colorectal cancer and Hurwitz and co-workers have published the results of this study in 2004 [ 3 ]. In December 2005, sorafenib received FDA approval for the treatment of renal cell carcinoma [ 4 ] while sunitinib received FDA approval in January 2006 for patients with gastrointestinal stromal tumors (GIST) and advanced kidney cancer [ 5 , 6 ].
Clinical studies have shown benefits in relapsed-free survival for metastatic colorectal cancer, advanced non-small cell lung cancer, renal cell carcinoma, hepatocellular carcinoma, metastatic breast cancer, GIST and in glioblastoma [ 7 , 8 ], but overall survival benefit has not yet been seen [ 9 ], with the exception of bevacizumab treatment in renal cell carcinoma as a single agent [ 10 ], or in metastatic breast cancer in combination with a taxane chemotherapy [ 11 ].
The most impressive clinical response occurred in the low dose bevacizumab plus chemotherapy with a statistically significant median overall survival (21.5 months) versus fluorouracil/leucovorin alone (13.9 months) or high-dose bevacizumab plus fluorouracil/leuocovirin (16.1 months) [ 12 ].
Autocrine VEGF signaling to promote malignant cell survival is also a common feature in haematological malignancies, suggesting that anti-VEGF/VEGFR targeted therapy would promote direct killing of tumor cells, as well as inhibit angiogenesis. VEGF-directed therapy has been investigated also in hematological malignancies, most commonly in acute myeloid leukemia, myelodysplastic syndrome, and in non-Hodgkin lymphoma.
Clinical trials involving anti-VEGF agents induce only a modest improvement in overall survival, measurable in weeks to just a few months, and various tumors respond differently in human patients to these agents.
These two principal findings could depend by different synergistic causes:
1) Lack of understanding of which patients will show the benefit of these agents and occurrence of drug resistance [ 9 , 13 , 14 ]. This is due to the absence of reliable surrogate markers of angiogenesis and antiangiogenesis to demonstrate the efficacy of antiangiogenic agents in clinical trials and for the monitoring of these agents [ 15 ].
2) Endothelial cells isolated from various tumors acquired genotype alterations, exhibiting aneuploidy, abnormal multiple chromosomes, and aberrant chromosomal architecture [ 16 ]. It has been proposed that proximity of tumor cells and endothelial cells within the tumor microenvironment may be responsible for the genotype alterations [ 17 ]. Genetic alteration of endothelial cells leads to altered antiangiogenic targets and resistance.
3) Antiangiogenic therapies may sometimes promote invasion and metastasis [ 18 ]. It has been demonstrated that sunitinib, a multi-targeted receptor tyrosine kinase inhibitor of VEGF and platelet derived growth factor (PDGF) signaling and the anti-VEGFR-2 antibody DC101 stimulated the invasive behavior of tumor cells despite their inhibition of primary tumor growth and increased overall survival in some cases [ 19 , 20 ].
4) Inherent or acquired resistance to anti-VEGF drugs can occur in patients, leading in some cases to a lack of response and in others to disease recurrence, although discontinuation of the therapy at the time of progression is a factor limiting the effectiveness of antiangiogenic therapy [ 21 ]. In the meantime, prolonged VEGF leads to vascular pruning and endothelial cell apoptosis, release of cytokines by host cells, which may promote tumor re-growth.
5) In most tumors, the vasculature is altered showing increased permeability, vessel dilatation, decreased/abnormal pericyte coverage and abnormal basement membrane structure. While VEGF neutralization can initially limit tumor proliferation due to its antiangiogenic effect, it can also result in transient vascular normalization with improved oxygenation and perfusion [ 22 ], favouring drug delivery. However, in gliomas normalization of the vascular bed involves restoration of the blood-brain barrier, thereby hampering, instead of enhancing, the delivery of therapeutic compounds to tumor cells [ 23 ].
6) Prolonged VEGF inhibition increases local hypoxia leading to systemic secretion of other angiogenic cytokines, such as FGF-2 and SDF-1α, which may promote cancer re-growth and metastasis [ 24 ]. An analysis of human breast cancer biopsies revealed that late-stage breast cancers expressed several angiogenic cytokines in contrast to earlier stage lesions, which preferentially expressed VEGF [ 25 ].
VEGF inhibitors are effective in the antiangiogenic treatment, but genetic mutations, vascular changes, up-regulation of other pro-angiogenic cytokines, promotion of invasion and metastasis, reduce their effectiveness. Further clinical investigations are needed to optimize antiangiogenic treatments in solid and hematological tumors management, as well as the identification of reliable markers that predict the relapse and the response to these therapies.
Competing interests
The Author declares that he has no competing interests. | Acknowledgements
Supported in part by MIUR (PRIN 2007), Rome, AIRC, Milan, and Fondazione Cassa di Risparmio di Puglia, Bari, Italy. | CC BY | no | 2022-01-12 15:21:44 | J Angiogenes Res. 2010 Dec 10; 2:27 | oa_package/da/de/PMC3016244.tar.gz |
||||||
PMC3016246 | 21138556 | Conclusions
Transposon-based technologies have enormous potential to develop powerful genomic tools with the vision of creating a bridge between physiology and genetics and to establish safe and inexpensive protocols for clinical gene transfer. Simple, plasmid-based vectors matched by a corresponding transposase source offer an easy and efficient method for germline transgenesis in laboratory animals and in large animal species for biotechnology. Furthermore, it is now both accessible and practical to generate highly complex libraries of gene knockouts in model species with a view to establishing new models of human disease for the annotation of disease pathways and for therapeutic and pharmaceutical intervention. The recently developed SB100X hyperactive transposon system yields highly efficient stable gene transfer after non-viral gene delivery into therapeutically relevant primary cell types, including stem cells, and thus may facilitate the clinical implementation of ex vivo and in vivo gene therapies. The next phase of preclinical research will focus on further refinement in large animal models to undertake SB -mediated transposition in vivo and to improve the safety profile of SB vectors by target-selected transgene integration into genomic 'safe harbors'. Although it remains to be seen whether the first clinical application of the SB system will result in a therapeutic effect, this trial will help validate the safety of this approach. The ongoing investigations will certainly prompt new ideas and new designs to be developed in this (ever) expanding universe of transposon technologies for genetic and cell engineering. | Transposable elements can be viewed as natural DNA transfer vehicles that, similar to integrating viruses, are capable of efficient genomic insertion. The mobility of class II transposable elements (DNA transposons) can be controlled by conditionally providing the transposase component of the transposition reaction. Thus, a DNA of interest (be it a fluorescent marker, a small hairpin (sh)RNA expression cassette, a mutagenic gene trap or a therapeutic gene construct) cloned between the inverted repeat sequences of a transposon-based vector can be used for stable genomic insertion in a regulated and highly efficient manner. This methodological paradigm opened up a number of avenues for genome manipulations in vertebrates, including transgenesis for the generation of transgenic cells in tissue culture, the production of germline transgenic animals for basic and applied research, forward genetic screens for functional gene annotation in model species, and therapy of genetic disorders in humans. Sleeping Beauty ( SB ) was the first transposon shown to be capable of gene transfer in vertebrate cells, and recent results confirm that SB supports a full spectrum of genetic engineering including transgenesis, insertional mutagenesis, and therapeutic somatic gene transfer both ex vivo and in vivo . The first clinical application of the SB system will help to validate both the safety and efficacy of this approach. In this review, we describe the major transposon systems currently available (with special emphasis on SB ), discuss the various parameters and considerations pertinent to their experimental use, and highlight the state of the art in transposon technology in diverse genetic applications. | Transposons as genetic tools
DNA transposons are discrete pieces of DNA with the ability to change their positions within the genome via a 'cut and paste' mechanism called transposition. In nature, these elements exist as single units containing the transposase gene flanked by terminal inverted repeats (TIRs) that carry transposase binding sites (Figure 1A ). However, under laboratory conditions, it is possible to use transposons as bi-component systems, in which virtually any DNA sequence of interest can be placed between the transposon TIRs and mobilized by trans -supplementing the transposase in the form of an expression plasmid (Figure 1B ) or mRNA synthesized in vitro . In the transposition process, the transposase enzyme mediates the excision of the element from its donor plasmid, followed by reintegration of the transposon into a chromosomal locus (Figure 1C ). This feature makes transposons natural and easily controllable DNA delivery vehicles that can be used as tools for versatile applications, ranging from somatic and germline transgenesis to functional genomics and gene therapy (Figure 2 ).
Transposons have been successfully used in plants and in invertebrate animal models, including Arabidopsis , rice, Caenorhabditis elegans [ 1 - 3 ] and Drosophila [ 4 - 6 ] for transgenesis and insertional mutagenesis, but until recently, there was no known transposon that was sufficiently active to be tailored as a tool for such purposes in vertebrates. This is because transposons tend to have limitations with respect to the species in which they can jump. In 1997, the Sleeping Beauty ( SB ) transposon system was engineered by molecular reconstruction of an ancient, inactive Tc1/ mariner -type transposon found in several fish genomes [ 7 ]. This newly reactivated element allowed highly efficient transposition-mediated gene transfer in major vertebrate model species without the potential risk of cross-mobilization of endogenous transposon copies in host genomes. This is because the genomes of major models lack endogenous transposon sequences with sufficient sequence similarity for mobilization by an exogenously supplied SB transposase. Indeed, SB has been successfully used as a tool for genetic modifications of a wide variety of vertebrate cell lines and species including humans [ 8 - 10 ].
During the past decade, other elements have been shown to catalyze efficient transposition in vertebrate model organisms. For example, the insect elements piggyBac [ 11 , 12 ] and Minos [ 13 , 14 ] catalyze efficient transposition in mammalian cells. Minos was also shown to be active in the basal chordate Ciona intestinalis [ 15 ]. Moreover, the reconstructed amphibian element Frog Prince [ 16 ], the reconstructed human Hsmar1 element [ 17 ], the reconstructed zebrafish transposon Harbinger3_DR [ 18 ], and the Tol1 [ 19 ] and Tol2 [ 20 ] elements isolated from the medaka fish have been found to be active in vertebrate species. Passport , a Tc1-family transposon isolated from a fish ( Pleuronectes platessa ), is active in a variety of vertebrate cells [ 21 ], and the Ac/Ds transposon originally discovered in maize by McClintock undergoes efficient transposition in zebrafish embryos [ 22 ]. Thus, the piggyBac , Minos and Ac/Ds elements appear to have a significantly wider possible host range than most other transposons. The basic criteria for the applicability of a transposon in any given model organism are 1) a sufficient level of transpositional activity in the given species, and 2) target site selection properties of the transposon, which are discussed below.
Hyperactive transposon systems
In evolutionary terms, the SB transposon represents a successful element that was able to colonize several fish genomes millions of years ago [ 7 ]. However, even successful transposons have not been selected for the highest possible activity in nature, because unlike viruses, they have to coexist with their hosts and, consequently, there is strong selective pressure to avoid insertional mutagenesis of essential genes. Indeed, although the resurrected SB element was sufficiently active to be mobilized in vertebrate cells, its relatively limited transpositional activity still presented a bottleneck for some applications. For example, requirements for transfection of primary cells and other hard-to-transfect cell types, or for remobilization of transposons from chromosomally resident single-copy donor sites, demanded an enzyme with more robust activity. Thus, enhancing transpositional activity has been one of the main targets for transposon vector development. To date, almost every single amino acid in the SB transposase has been changed in an attempt to increase its activity. Three main strategies have been applied to derive hyperactive mutants of the SB transposase: 'importing' amino acids and small blocks of amino acids from related transposases [ 23 - 25 ], systematic alanine scanning [ 26 ], and rational replacement of selected amino acid residues [ 24 ]. Together, these studies have yielded several, single amino acid replacements, each resulting in a relatively modest increase in transpositional activities. Unfortunately, the hyperactivity of most of the SB transposase mutants selected in immortalized cell lines did not translate to efficient stable gene transfer in primary cells in vivo [ 25 , 26 ].
Earlier studies established that certain combinations of amino acid replacements, each leading to hyperactivity, can yield a further enhancement in transpositional activity of the SB transposase [ 23 - 26 ], but guessing the correct combinations of variants out of the millions that are possible is like finding the correct combinations of numbers in a lottery. A high-throughput, PCR-based, DNA-shuffling strategy and screening of 2000 gene variants in mammalian cells produced a variant of SB that was 100-fold more potent in chromosomal insertion of a transgene than the originally reconstructed protein [ 27 ]. The use of SB100X demonstrated that it is possible to establish a transposon-based, non-viral vector system that is capable of stable gene transfer coupled with long-term gene expression at an efficiency comparable with that of viral strategies [ 27 ]. Thus, the hyperactive SB100X transposase holds great promise of offering broad utility in gene therapy and functional genomics.
Integration site preference
The insertion pattern of most transposons is nonrandom, showing characteristic preferences for insertion sites at the primary DNA sequence level, and 'hotspots' and 'cold regions' on a genome-wide scale. For example, for the primary DNA sequence, the Tol2 element does not appear to exhibit a pronounced preference for any sequence for insertion [ 28 ]. By contrast, the Harbinger3_DR transposon is highly specialized to integrate into the palindromic AAACACCWGGTCTTT consensus sequence [ 18 ], the piggyBac transposon targets the sequence TTAA, and all Tc1/ mariner transposons, including SB , Frog Prince , Minos and Hsmar1 , target their integration into TA dinucleotides. In the case of SB , this preference has been studied in detail, and palindromic AT repeats found to be the preferred sites for integration [ 29 ]. However, computational analyses revealed that target selection is determined primarily at the DNA structure level, not by specific base-pair interactions. For example, protein-induced deformability was shown to be associated with preferred SB insertion sites, whereas piggyBac and Tol2 integration sites lack such consistent, clear-cut structural patterns [ 30 , 31 ]. This suggested that integrations of SB will occur into any DNA available, depending on these preferences only, but this is not the case. In the context of chromatin, Tc1/ mariner elements have no or weak preference for transcription units, the 5' regulatory regions are not favored, and most hits in genes are localized within introns [ 29 , 32 ]. By contrast, piggyBac shows a greater propensity to integrate into transcription units, with a preference for insertion around transcription start sites [ 12 , 33 - 35 ], and the Tol2 transposon also shows a pronounced preference for integration close to transcriptional start sites [ 28 ]. This control of integration at the chromatin level is poorly understood. One possible explanation for this is the interaction of the transposase with unknown, chromatin-associated factors. Supporting this hypothesis, it has been shown that a host-encoded protein, lens epithelium derived growth factor (LEDGF), is involved in directing integration of human immunodeficiency virus (HIV) into active genes [ 36 ]. Taken together, the preferences of particular elements to integrate into expressed genes versus non-coding DNA, and their preferences for integration sites within genes are expected to be substantially different.
Integration site preference can greatly influence the utility of transposon vectors for different applications. For example, human gene therapy protocols require application of transposon vectors showing the least preference for target genes, for obvious safety reasons. The SB system (which shows close-to-random insertion site distribution) appears to best satisfy these needs, whereas the piggyBac and Tol2 systems (which prefer genes and their upstream regulatory regions for insertion) appear to be less favorable for potential therapeutic applications. Nevertheless, a systematic assessment of potential genotoxic effects associated with genomic integration of transposon vectors will need to be performed either in cell-based assays and/or in animal models to provide clinically relevant data.
Unlike in therapeutic applications, hitting genes by insertional elements is the goal with forward mutagenesis screens. However, the insertional biases associated with vector systems represent the main limitation to full genome coverage with individual transposon-based vectors. Thus, in this respect, the utility of transposons for mutagenesis is greatly enhanced by the availability of multiple alternative vector systems with distinct preferences for insertion, such as SB , Tol2 and piggyBac . Indeed, the propensity of Tol2 to insert close to transcriptional start sites of genes might be particularly advantageous for enhancer trapping [ 37 , 38 ], while the propensity of piggyBac to insert into transcription units supports genome-wide mutagenesis with gene trap cassettes [ 39 ].
Local hopping
'Local hopping' describes a phenomenon of chromosomal transposition in which transposons have a preference for landing into cis -linked sites in the vicinity of the donor locus. Local hopping seems to be a shared feature of 'cut and paste' transposons. However, the actual extent of hopping to linked chromosomal sites and the interval of local transposition varies. For example, the P -element transposon of Drosophila prefers to insert within ~100 kb of the donor site at a rate ~50-fold higher than in regions outside that interval [ 40 ]. Similarly, in germline mutagenesis screens in mice using SB , 30-80% of the transposons re-insert locally on either side of the transposon donor locus [ 41 - 43 ]. In contrast to the P -element, SB seems to have a much larger local transposition interval between 5 and 15 Mb [ 42 ].
The local hopping feature not only differs between different transposons, but a given transposon may show great variations in local hops in different hosts, and in different donor loci even in the same host. For example, about 50-60% of the reinserted Ac elements were found to be distributed within a 5-cM distance of the donor site in maize [ 44 , 45 ], and the frequency of local hopping greatly varies in Arabidopsis and tobacco, depending on the chromosomal location of the donor site [ 46 - 48 ]. Moreover, local hopping of the Ac element in tomato seems overall to be less prevalent than in maize [ 49 , 50 ], and there are species-specific differences in its tendency for local hopping out of different transposon donor loci [ 51 ]. This variation in local hopping of the same element could possibly be explained by varying affinity of the transposase to unknown, chromatin-associated factors in different hosts [ 52 ].
Local hopping can play a significant role in mutagenesis using chromosomally resident transposons. In practical terms, local hopping limits the chromosomal regions accessible to a transposon jumping out of a given chromosomal site [ 53 ]. To circumvent this limitation, establishing numerous 'launch pads' to initiate transposition out of different loci can be a viable strategy to increase coverage of gene mutations. On the other hand, local hopping can be useful for saturation mutagenesis within limited chromosomal regions.
Transposons and functional genomics
The post-genomic era presented the scientific community with the new challenge of functional annotation of every gene and identification of elaborate genetic networks. Diverse methods have been employed to address this task, including mutational analysis, which proved to be one of the most direct ways to decipher gene functions. There are versatile strategies for creating mutations, including insertional mutagenesis by discrete pieces of foreign DNA, which has the advantage that the inserted DNA fragment can serve as a molecular tag that allows rapid, usually PCR-based, identification of the mutated allele. Because the function of the gene in which the insertion has occurred is often disturbed, such loss-of-function insertional mutagenesis is frequently followed by functional analysis of mutant phenotypes. In many instances, retroviral vectors were used to introduce mutagenic cassettes into genomes, but their chromosomal insertion bias does not allow full coverage of genes [ 54 ]. The random integration pattern of the SB transposon, combined with its ability to efficiently integrate versatile transgene cassettes into chromosomes established this system as an extremely useful tool for insertional mutagenesis in both embryonic stem cells (ESCs) [ 34 , 55 ] and in somatic [ 56 , 57 ] and germline tissues [ 41 , 42 , 53 , 58 - 63 ] in animal models (Figure 2 ). There are several types of mutagenic cassettes that can be efficiently combined with transposon-based gene delivery for insertional mutagenesis. 5' gene-trap cassettes include splice acceptors and polyadenylation sequences so that transcription of genes can be disrupted upon vector insertion into introns (Figure 3A ) [ 54 ]. Often, such cassettes are also equipped with a reporter gene (usually, a fluorescent protein, β-galactosidase or antibiotic resistance) whose expression is dependent on correct splicing between exons of the trapped gene and the splice acceptor site carried by the transposon vector [ 64 , 65 ].
Insertional mutagenesis can be applied to cultured, germline-competent stem cells including ESCs and spermatogonial stem cells (SSCs) [ 39 , 66 ]. One advantage of this approach is the ability to perform preselection of modified ESC clones before generating mutant animals, and to differentiate selected clones into many different tissue types in vitro . It is possible to perform large-scale, transposon-based, insertional mutagenesis screens in ESCs and SSCs by simply transfecting or electroporating transposon donor and transposase expression plasmids into the cells. The amounts of the delivered plasmids can be adjusted to obtain the desired insertion frequencies per cell. In addition, transposons can also be remobilized from chromosomally resident loci and reintegrated somewhere else in the genome by transiently providing the transposase source; such excision-re-integration events can be monitored using double selection systems, in which excision activates the first and re-integration activates the second selection marker [ 43 ].
Because several aspects of physiology in rats have evolved to be more similar to humans than to mice, it would be desirable to use rat models in the process of functionally annotating the human genome by identifying the causative relationships between genes and disease phenotypes. As an important step towards this goal, an approach of establishing SB transposon-mediated insertional mutagenesis in rat SSCs was recently reported [ 66 ]. SB transposition can be used to tag and simultaneously mutate thousands of genes in culture, using gene-trap cassettes. Importantly, culture conditions maintain the potential of genetically manipulated SSCs to produce viable sperm cells. In that study, spermatogonial clones were transplanted to repopulate the testes of sterilized, wild-type recipient male rats. The stem cell genome was then passed on to transgenic offspring upon crossing of the recipient males with wild-type females (Figure 4 ). Although transposition events in a given target gene occur by chance, the tissue culture conditions allow screening for a large number of events. Transposition-mediated gene insertion and cell culture conditions thus allow generation of libraries of gene knockouts in rat SSCs (Figure 4 ). This technology has the potential to develop powerful genomic tools for use in the rat, offering the opportunity to create a bridge between physiology and genomics.
Another method in which transposons are used for insertional mutagenesis in animal models employs a 'jump-starter and mutator' scheme [ 42 , 58 , 61 ]. In this arrangement, mutator transgenic lines carry SB transposon-based gene-trapping vectors in the form of multicopy concatemers, whereas a jump-starter line expresses the transposase preferentially in the male germline [ 41 , 64 ]. Crossing of the two lines results in transposition in the germline of the F1 double-transgenic males, which are then repeatedly crossed with wild-type females to segregate the transposition events that occurred in their sperm cells to separate F2 animals. In the mouse system, a single sperm cell of an F1 male contains, on average, two transposon insertions [ 58 ], and up to 90% of the F2 progeny can carry transposon insertions [ 61 ]. The applicability of this approach has been demonstrated by the identification of mouse genes with either ubiquitous or tissue-specific expression patterns [ 42 , 64 , 67 , 68 ]. Recently, a similar system for SB insertional mutagenesis was also established in rats [ 62 , 63 ].
One cautionary note of launching transposition out of transposon arrays is that recombination between newly transposed transposon copies and the donor concatemer could lead to unwanted genomic rearrangements, as observed by Geurts et al . [ 68 ]. The most likely explanation for the rearrangements is that transposition out of a concatemer generates new transposase binding sites, linked either in cis (in the case of local hops) or in trans (in the case of transposition onto other chromosomes). However, because some transposon copies remain at the original donor locus, transposase can recombine chromosomal sequences that are located between the individual transposon units by hybrid element transposition (that is, the end of one transposon pairs with the opposite end of another transposon at a different location) [ 69 ], leading to deletions and translocations. Such chromosomal rearrangements are unlikely to occur if a single-copy donor is used. Thus, transposon systems sufficiently active for efficient transposition out of single-copy donors might eliminate the need for concatemeric donor sites in animal breeding schemes. Indeed, the Tol2 element was demonstrated to show transposition at reasonable efficiencies when launched from singly-copy donor sites in transgenic zebrafish [ 70 ]. In this context, the newly developed SB100X hyperactive system might also prove useful in future genetic screens.
Forward genetic screens do not necessarily need to depend on the breeding scheme described above; in some cases, a reasonable throughput in generation of transposon insertion mutants can be achieved by introducing the mutagenic transposon into individual animals, such as in zebrafish. A gene trap mutagenesis screen was recently employed to uncover genetic determinants of nicotine response in zebrafish, through a behavioral genetic screening paradigm [ 71 ]. Using standard transposase-mediated transgenesis protocols, Tol2 -based mutagenic vectors were co-injected into early zebrafish embryos by to generate a pool of mosaic F0 founder animals, which then underwent two successive rounds of crossing to generate homozygous mutant animals. Segregation of mutant animals from wild-type siblings was carried out using fluorescent reporters built into the gene trap cassettes. After profiling nicotine response in mutant versus wild-type fish, two mutants were identified out of a total of 102 fish lines screened [ 71 ]. This study emphasizes the utility of transposons for the discovery and functional annotation of genes relevant to human health in forward, phenotype-driven genetic screens in model species.
Transposon-based screens for cancer gene discovery
To induce gain-of-function mutations, transposon vectors can be equipped with oncogene trap cassettes that contain strong viral enhancers/promoters that can drive transcription outwards from the vector, thereby leading to overexpression of a full-length or truncated protein product of the trapped gene, as well as splice acceptor and polyA sites that lead to gene truncation with dominant phenotypes (Figure 3B ) [ 9 , 72 ]. SB vectors harboring oncogene traps have been successfully used in large-scale cancer gene discovery screens in experimental animals (Figure 5 ) [ 10 , 73 , 74 ]. In these studies, SB transposons were somatically mobilized from donor chromosomal concatemers, which contained either low (25) [ 56 ] or high (150-350) [ 57 ] numbers of the oncogene trap transposon. Dominant mutations in somatic tissues of double transgenic mice carrying a transposase source and the mutagenic transposons resulted in the generation of experimental tumors in cancer-predisposed [ 56 ] and wild-type [ 57 ] animals. In a follow-up study, Collier et al . demonstrated that a combination of low-copy oncogene trap lines with the SB11 transposase (an early-generation hyperactive SB variant) expressed from the Rosa26 locus could achieve whole-body transposon mobilization at rates sufficient to promote penetrant tumorogenesis without complications of embryonic lethality or genomic instability [ 75 ]. Thus, this approach can be successfully employed not only to identify novel cancer genes, but also combinations of cancer genes that act together to transform a cell.
Current efforts are concentrating on customized, tissue-specific screens for cancer development studies. The strategies employed to achieve this goal focus on establishing mouse lines that either conditionally express the transposase from tissue-specific promoters, or rely on generation of Cre recombinase-inducible transposase alleles that can be used in conjunction with mice that express Cre in a tissue-specific manner [ 76 - 78 ]. For example, this approach was addressed by Dupuy and co-workers [ 72 ], who were able to experimentally modify the spectrum of tumors by creating a Cre-inducible SB transposase allele (RosaSBase LsL ). With this strategy, they managed to overcome the obstacle of high embryonic lethality associated with ubiquitous SB transposase expression in the presence of the pT2/Onc2 oncogene trap [ 76 , 79 ], and to generate a model of germinal center B-cell lymphoma. They achieved this by activating SB transposase expression with an AidCre allele that drove Cre-mediated recombination in germinal center B-cells. In another approach, ubiquituous expression of the SB transposase was combined with the novel T2/Onc3 oncogene trap transposon vector. In that study, the MSCV (mouse stem cell virus) 5' long terminal repeat that was previously used to drive oncogene expression was replaced by the ubiquitously active CAGGS promoter, resulting in removal of the bias towards inducing mostly lymphomas and in reducing embryonic lethality. This strategy emphasizes that the change in the design of the mutagenic transposon (e.g. promoter choice) can have profound effects on the tumor types induced by transposition. Notably, this approach resulted in production of nearly 200 independent tumors of more than 20 types, and identification of novel, candidate cancer genes, suggesting that the combination of tissue-specific promoters and inducible transposase alleles could provide a fine mechanism of control in tumorogenesis studies.
Transposons as vectors for stable transgene integration and expression
The classic approaches to stable expression of foreign genes in vertebrates rely on physical methods of delivering gene constructs into cultured cells, such as transfection, electroporation and sonoporation, or microinjection into oocytes or fertilized eggs to generate germline transgenic animals. The main drawbacks of these approaches are the low rates of genomic integration and the unstable expression of the chromosomally integrated gene construct, which is believed to be associated with the phenomenon of concatemerization of the injected DNA before genomic integration [ 80 ]. Another particular problem in transgenic animals is that founders that develop from the injected oocytes or eggs are predominantly mosaic for the transgene, because integration generally occurs relatively late during embryonic development. In principle, all of these drawbacks can be circumvented by transposition-mediated gene delivery, as it can increase the efficiency of chromosomal integration and facilitates single-copy insertion events. Single units of expression cassettes are presumably less prone to transgene silencing than are the concatemeric insertions created by classic methods.
Transposon-based technologies can be exploited for gene transfer in cultured cells and in primary cell types, including stem cells (Figure 2 ). For example, transposons can be harnessed to integrate plasmid-based shRNA expression cassettes into chromosomes to obtain stable knockdown cell lines by RNA interference (Figure 3C ) [ 81 ]. Such technologies have been evaluated as a potential approach to the therapy of acquired immunodeficiency syndrome by stable RNA interference with SB vectors knocking down the CCR5 and CXCR4 cell surface co-receptors that are required for viral entry as a first step to confer resistance to HIV [ 82 ]. Both the SB and the piggyBac systems were shown to support efficient transposition in mouse [ 33 , 43 ] and human [ 83 , 84 ] ESCs. In a recent, elegant study, piggyBac -derived transgene vectors were introduced into human ESCs for the purpose of driving ESC differentiation toward a specific cell type [ 84 ]. The vectors included loss-of-function shRNA expression cassettes that could simultaneously knock down the expression of pluripotency genes and of genes that contribute to endodermal and mesodermal differentiation, plus a gain-of-function construct expressing Sox1 to direct differentiation towards neuroectoderm. What makes such complex gene-transfer experiments possible is the observation that several transgene constructs maintained on separate transposon vectors can be delivered simultaneously in 'multiplex' transposition reactions, in which the different constructs are simply mixed together and cotransfected into cells [ 85 ].
The recent discovery of induced pluripotent stem cells (iPSCs) by the expression of four key genes ( Oct4 , Sox2 , Klf4 and c- Myc ) in differentiated somatic cells holds enormous promise for future regenerative medicine [ 86 ]. Transposons are attractive vehicles for reversible production of iPSCs, because the excision step of the transposition reaction produced by transient re-expression of the transposase offers removal of the transgenes after completion of reprogramming, allowing subsequent differentiation of the iPSCs into various lineages in vitro [ 87 ]. Transposition-mediated generation of mouse and human iPSCs, and removal of the reprogramming factors from the pluripotent cells have already been achieved by the piggyBac system [ 88 ]. What makes piggyBac transposons especially attractive vectors for the production of transgene-free iPSCs is a special feature of these transposons: excision fully restores the sequence of the original wild-type locus [ 89 ], thereby allowing traceless removal of transgenes from the genome.
In vivo , co-injection of engineered transposons with transposase mRNA into fertilized oocytes can facilitate early integration events that potentiate successful transmission of the transgene through the germline to the next generation (Figure 2 ). This method has been employed to generate transgenic zebrafish with Tc3 [ 90 ], Mos1 [ 91 ], Tol2 [ 20 ] and SB [ 92 ]; transgenic Xenopus with SB [ 93 ] and Tol2 [ 94 ]; and transgenic mice with SB [ 27 , 95 - 97 ], piggyBac [ 11 ] and Tol2 [ 98 ]. In this context, an important step towards transgenesis with bacterial artificial chromosomes (BACs) has been made by the delivery of a ~70-kb BAC construct into zebrafish and mouse embryos with Tol2 [ 99 ]. Thus, transposons can evidently be used to stably deliver large transgene constructs together with complex regulatory regions, without the complications of DNA rearrangements and silencing associated with classic methods.
Any transgene vector system should provide long-term expression of transgenes. Transgenes delivered by non-viral approaches often form long, repeated arrays (concatemers) that are targets for transcriptional silencing by heterochromatin formation. In addition, long-term expression of transgenes delivered by retroviruses has been shown to be compromised by transcriptional silencing [ 100 ]. It was recently shown that the zinc finger protein ZFP809 bridges the integrated proviral DNA of the murine leukaemia virus and the tripartite motif-containing 28 transcriptional co-repressor in embryonic stem cells [ 101 ]. Thus, sequence elements in the vector itself can predispose the cargo for silencing. The cut and paste mechanism of DNA transposition results in a single copy of the transgene per insertion locus, thus concatemer-induced gene silencing is unlikely to be an issue with transposition-mediated gene transfer. Indeed, Grabundzija et al . found that transposon insertions delivered by the SB , Tol2 and piggyBac systems only rarely (< 4% of all insertions) undergo silencing in HeLa cells [ 28 ]. Furthermore, stable transgene expression observed in hundreds of independent insertions in this study suggests that these three transposon systems rarely target heterochromatic chromosomal regions for insertion, and that it is unlikely that certain sequence motifs in the transposon vectors are recognized by mediators of silencing in the cell. An additional factor that may provoke transgene silencing is the cargo DNA, particularly the type of promoter used to drive expression of the gene of interest. Indeed, it was previously shown that transgene constructs delivered into mouse cells using SB transposition can be subject to epigenetic regulation by CpG methylation and that a determinant of epigenetic modifications of the integrating transposon vector is the cargo transgene construct, with the promoter playing a major role [ 102 ]. However, with careful promoter choice, several studies have established that SB -mediated transposition provides long-term expression in vivo . For example, stable transgene expression from SB vectors was seen in mice after gene delivery in the liver [ 103 - 106 ], lung [ 107 , 108 ], brain [ 109 ] and blood after hematopoietic reconstitution in vivo [ 27 , 110 ]. Thus, although our understanding of all the factors that will ultimately determine the expressional fate of an integrated transposon is still rudimentary, it appears that transposon vectors have the capacity to provide long-term expression of transgenes both in vitro and in vivo .
Transposons as vectors for gene therapy
Considerable effort has been devoted to the development of gene delivery strategies for the treatment of inherited and acquired disorders in humans. A desirable gene therapy approach should 1) achieve delivery of therapeutic genes at high efficiency specifically into the relevant cells, 2) be adaptable to changing needs in terms of vector design, 3) minimize the risk of genotoxicity, and 4) be cost-effective.
Adapting viruses for gene transfer is a popular approach; for example, γ-retroviral and lentiviral vectors are efficient at integrating foreign DNA into the chromosomes of transduced cells and have enormous potential for lifelong gene expression [ 111 ]. A major concern of using retroviral vectors is the potential for mutagenic effects at the sites of genomic integration [ 112 - 114 ]. Indeed, insertional mutagenesis has been observed in clinical trials using a retroviral vector for gene therapy of X-linked severe combined immunodeficiency [ 112 , 114 , 115 ]. The clinical use of retroviral vectors can be curtailed because of the limited size of the payload, as multiple or large transgenes compromise the efficiency of viral reverse transcription and packaging. Finally, regulatory issues and the high costs associated with manufacture of clinical-grade retrovirus hamper their widespread translation into clinical practice. An ideal therapeutic vector would combine the favorable attributes of integrating viral vectors (that is, stable chromosomal insertion) while significantly reducing the potential for adverse events. Transposons could potentially offer such an alternative (Figure 2 ).
The advantage of SB transposon-based gene delivery is that, owing to stable genomic insertion of expression cassettes, it can lead to both long-term and efficient transgene expression in preclinical animal models [ 116 ]. Thus, the SB plasmid-based transposon system combines the advantages of viral vectors with those of naked DNA molecules. However, in contrast to viral vectors, transposon vectors can be maintained and propagated as plasmid DNA, which makes them simple and inexpensive to manufacture, an important issue for the implementation of future clinical trials. Further advantages of the SB system include its reduced immunogenicity [ 103 ], no strict limitation of the size of expression cassettes [ 24 ] and improved safety and toxicity profiles [ 87 , 117 - 119 ]. Because the transposition mechanism does not involve reverse transcription, DNA-based transposon vectors are not prone to incorporating mutations and can tolerate larger and more complex transgenes, including those containing repeat DNA motifs. Moreover, the use of SB -based gene delivery eliminates the risk of rearrangements of the expression cassette that, as part of a transposing unit of DNA, integrates into chromosomal DNA in an intact form [ 120 ]. Compared with retroviral systems, the SB vectors have an inherently low enhancer/promoter activity [ 117 , 118 ]. Inserting insulator sequences flanking the transcription units of the cargo to prevent accidental trans -activation of promoters of neighboring genes further increased the safety features of the SB system (Figure 3D ) [ 117 ]. Notably, the transposase can be provided as messenger RNA, thereby reducing the risk of 'rehopping' of the transposon-based vector [ 96 ]. Chromosomal integration of SB transposons is precise and random (see above), and no SB -associated adverse effects have been reported [ 116 , 120 , 121 ]. Of note, a precise integration mechanism, random integration pattern and negligible promoter/enhancer activity do not appear to be general features of all recombinase/transposon systems. For example, integration promoted by the bacteriophage-derived PhiC31 system was reported to generate chromosomal rearrangements [ 122 , 123 ]. The 5' TIR of the piggyBac transposon exhibits significant promoter activity in mammalian cells [ 124 ], and its genomic integration profile resembles that of integrating viral vectors [ 12 ], as described above.
The past few years have seen a steady growth in interest in applying the SB system for the treatment of several conditions including haemophilia A and B [ 103 , 104 , 106 , 107 , 125 ], junctional epidermolysis bullosa [ 126 ], tyrosinemia I [ 127 ], Huntington disease [ 128 ] sickle cell disease [ 129 ], mucopolysaccharidosis [ 105 , 130 ], cancer [ 109 , 131 ] and type 1 diabetes [ 132 ]. In addition, important steps have been made towards SB -mediated gene transfer in the lung for potential therapy of α-1-antitrypsin deficiency, cystic fibrosis and a variety of cardiovascular diseases [ 108 , 133 ]. Thus, the establishment of non-viral, integrating vectors has generated considerable interest in developing efficient and safe vectors for human gene therapy [ 116 , 120 , 134 - 136 ].
The SB100X hyperactive transposon system yields efficient stable gene transfer after non-viral gene delivery into therapeutically relevant primary cell types, including stem or progenitor cells. For example, the use of the SB100X system yielded robust gene transfer efficiencies into human hematopoietic progenitors [ 27 , 110 ], mesenchymal stem cells, muscle stem/progenitor cells (myoblasts) and iPSCs [ 137 ]. These cells are relevant targets for stem cell biology and for regenerative medicine and gene- and cell-based therapies of complex genetic diseases. Importantly, expression of the SB100X hyperactive transposase did not adversely influence the differentiation or function of these adult stem/progenitor cells, nor was there any evidence of any cytogenetic abnormalities [ 137 ]. In the context of iPSC technology, the ability to coax the differentiation of pluripotent stem cells into clinically relevant, transplantable cell types is a key step towards their ultimate use in clinical applications, especially because undifferentiated iPSCs pose an intrinsic tumorigenic risk [ 138 ]. It was recently demonstrated that SB transposon-mediated delivery of the myogenic PAX3 transcription factor into iPSCs coaxed their differentiation into MyoD+ myogenic progenitors and multinucleated myofibers [ 137 ], suggesting that PAX3 may serve as a myogenic 'molecular switch' in iPSCs, a finding that has implications for cell therapy of congenital degenerative muscle diseases, including Duchenne muscular dystrophy.
The first clinical application of the SB system is currently ongoing using autologous T cells genetically modified to redirect specificity for B-lineage malignancies [ 139 ]. Lymphocytes are a suitable initial platform for testing new gene transfer systems, as there have been hundreds of infusions of clinical-grade T cells genetically modified using viral and non-viral approaches without apparent genotoxicity [ 140 ]. The SB transposon tested in the first human application carries a chimeric antigen receptor (CAR) to render the T cells specifically cytotoxic toward CD19-positive lymphoid tumors [ 141 , 142 ]. The advantage of using the SB system for the genetic modification of T cells includes the reduced cost associated with manufacturing of clinical-grade DNA plasmids compared with recombinant viral vectors. This is particularly important when one considers that trials infusing CAR-positive T cells are only now beginning to demonstrate anti-tumor effects [ 143 , 144 ]. The higher enzymatic activity of SB100X might enable integration efficiencies comparable with that of retroviral vectors to be achieved for next-generation trials.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
Both authors contributed to drafting, reading and approving the final manuscript. | Acknowledgements
Work in the authors' laboratories was supported by EU FP6 (INTHER) and EU FP7 (PERSIST and InduStem), grants from the Deutsche Forschungsgemeinschaft SPP1230 'Mechanisms of gene vector entry and persistence', and from the Bundesministerium für Bildung und Forschung (NGFN-2, NGFNplus, iGene, InTherGD and ReGene). | CC BY | no | 2022-01-12 15:21:44 | Mob DNA. 2010 Dec 7; 1:25 | oa_package/cd/68/PMC3016246.tar.gz |
||||
PMC3016247 | 21172028 | Background
Hashimoto thyroiditis is the most common cause of hypothyroidism in the United States [ 1 ]. With a 5-10 time preference over men, the reported prevalence in white women is in the 1-2% range [ 2 ].
Etiology and pathogenesis of Hashimoto thyroiditis are still elusive. Moreover, little is known about progression of euthyroid to hypothyroid Hashimoto's. At least in children, disease progression from euthyroid to hypothyroid Hashimoto thyroiditis has been suggested [ 3 ]. Also, available evidence relating the progression of sub-clinical to overt hypothyroidism in adults has been rated as good [ 4 ]. Hence, it is conceivable that a euthyroid stage of Hashimoto thyroiditis exists and that progression to a full-blown disease stage is a matter of time.
Since there is growing evidence that unrecognized hypothyroidism is deleterious, early diagnosis of Hashimoto thyroiditis would be advantageous in predicting thyroid failure. Specifically, it is well known that maternal thyroid status assessment and treatment improves fetal outcomes and neuropsychological developmental of the newborn [ 5 ].
The University of Wisconsin Thyroid Multidisciplinary Clinic is a large referral site for thyroid diseases in the Midwest. A continuously increasing number of thyroid biopsies are being performed every year for cancer screening. Yet, Hashimoto thyroiditis is being too frequently diagnosed. The prevalence of Hashimoto thyroiditis is reported to be approximately twice that of type 1 diabetes. However, the prevalence of Hashimoto thyroiditis confirmed by cytology has never been documented in a large cohort of patients with ultrasound detectable nodules. To evaluate different aspects of thyroid physiopathology, a database of clinical features, ultrasound images and cytology results of patients referred for fine needle aspiration of thyroid nodules was prospectively developed. In this paper we probed our database for the frequency and characteristic of patients diagnosed with Hashimoto thyroiditis while being referred for thyroid cancer screening. | Methods
Thyroid Database
From March 2006 until September 2008, 811 patients underwent ultrasound guided fine needle aspiration (FNA) biopsy of thyroid nodules for screening of thyroid cancer. After excluding 50 patients who were either no-show or their specimens were non-diagnostic, we retrospectively studied 761 consecutive patients for which ultrasound guided thyroid FNA biopsies were performed at our clinic (Figure 1 ). All FNA samples were reviewed by cytopathologists at our institution.
The Hashimoto thyroiditis cohort consisted of 102 (13.4%) patients (659 out of 761 did not have cytological Hashimoto's diagnosis) for which 46 (6%) were identified as having clinical disease (i.e. diagnosed hypothyroid on thyroid hormone replacement and with cytological Hashimoto's diagnosis), 9 (1.2%) as having sub-clinical hypothyroidism (as defined by normal thyroid hormones with above normal [usually less than 10 ng/dl] TSH and with cytological Hashimoto's diagnosis) and 47 (6.2%) as having euthyroid autoimmunity (as defined by normal thyroid hormones with normal [0.45-4.12 ng/dl] TSH but with cytological Hashimoto's diagnosis). For all patients, data were collected for: age, gender, nodule size and number of nodules, levels of TSH/FT4, presence of TPO (thyroid peroxidase) autoantibodies, family history of Hashimoto's, goiter, and autoimmune diseases. The collection of patient's data and subsequent analysis was approved by University of Wisconsin Human Subjects Institutional Review Board.
Statistical Analysis
Data analysis was conducted using Excel X2 and Fisher's exact test and Student t-test were used when appropriate. Statistical significance was defined as P < 0.05. | Results
Patients Characteristics
Patients were referred to the University of Wisconsin Thyroid Multidisciplinary Clinic usually for evaluation of an already diagnosed thyroid problem. Thyroid nodules ≥ 1 cm were biopsied under ultrasound guidance mainly to screen for thyroid cancer. On-site cytopathology was available for every single case. Endocrine surgery referral was also available as needed during the same encounter. A database was started with the inauguration of the clinic and recruitment is ongoing. Of the first 811 patients' data analyzed, an unexpected large number of patients with positive Hashimoto thyroiditis cytology were identified. The analysis of our database revealed that from 761 patients, 102 (13.4%) had cytology proven Hashimoto thyroiditis. Patients' characteristics of this cohort are shown on table 1 . Hashimoto thyroiditis cohort consisted of 94 female and 8 male with mean (± SD) age of 47 ± 14 years. Out of these 102 patients with cytology-proven Hashimoto thyroiditis, 47 (6.2%) were euthyroid, 9 (1.2%) had sub-clinical hypothyroidism, and 46 (6%) had clinical hypothyroidism and were on thyroid hormone replacement (Table 1 and Figure 2 ). The number of nodules was 1.43 ± 0.54 average for the clinically hypothyroid subgroup, 1.77 ± 0.66 average for the sub-clinical subgroup (p < 0.05 when compared to clinically hypothyroid subgroup) and 1.58 ± 0.71 for the euthyroid subgroup. The size of the nodules was 1.42 ± 0.42; and 1.55 ± 0.81; for the clinical and sub-clinical subgroups respectively and 1.69 ± 0.74 (p < 0.05, also Table 1 ) for the euthyroid subgroup. Although differences between nodules sizes of clinically hypothyroid versus sub-clinical subgroups appear significant, the result of this comparison is skewed by the low number of patients in the sub-clinical subgroup as well as the wider range. Larger size of euthyroid subgroup nodules was however statistically significant when compared to clinically hypothyroid subgroup nodules both subgroups with similar number of nodules. Another significance difference was also observed in the comparison of TSH values between euthyroid and clinically hypothyroid subgroups as expected. The clinically hypothyroid subgroup had an average TSH of 8 ± 2.61 ng/dl while the euthyroid subgroup had a lower TSH of 2.18 ± 1.14 ng/dl. The sub-clinical subgroup had average TSH value of 6.73 ± 1.26 ng/dl. All nodules were reported as benign. Of note, although patients with clinical hypothyroidism were all taking thyroid hormone, none in the sub-clinical or the euthyroid subgroups were on thyroid hormone replacement. FT4 (free T4) in the clinically hypothyroid subgroup was found to be 1 ± 0.36; 0.96 ± 0.11 in the sub-clinical and 0.92 ± 0.11 in the euthyroid subgroup. FT4 however was not recorded in 52% of the clinically hypothyroid subgroup and in 56/57% of the sub-clinical/euthyroid subgroups respectively. Not recorded also, was the presence or absence of TPO autoantibodies in 30 of 46 patients with clinical hypothyroidism, 6 of 9 patients with sub-clinical and 34 of 47 with euthyroid Hashimoto's. However out of the recorded TPO autoantibodies, 13 of 16 (81.3%) were positive in the clinically hypothyroid subgroup while 8 of 16 (50%) were positive in the sub-clinical/euthyroid subgroups (also in Table 1 ). There were no significant differences among euthyroid, sub-clinical and clinically hypothyroid Hashimoto thyroiditis subgroups in terms of age, gender, family history of Hashimoto's, goiter, and/or autoimmune diseases (also in Table 1 ).
Hashimoto thyroiditis more common than expected
Although the prevalence of Hashimoto thyroiditis diagnosed by ultrasound guided FNA cytology has not been established, the prevalence found in our cohort was unexpectedly high in comparison with that reported for Hashimoto's hypothyroidism and subclinical disease combined (Figure 2 ). The reasons for these differences in prevalence estimates are not obvious. Hence we considered potential factors that might have biased the results. Among them, the clinical indications for the biopsies and the cytological criteria for Hashimoto thyroiditis diagnosis were reviewed for all cases. Clinical indications for biopsies of nodules were based on the American Thyroid Association guidelines for size criteria (see nodule sizes in Table 1 ) which were determined by ultrasound before and confirmed during the biopsy (Figure 3A as example). Standard cytological characteristics for diagnosis were also followed by our cytologists as described by Kini [ 6 ]. The cytological diagnosis of Hashimoto's thyroiditis relied on the presence of both an inflammatory infiltrate accompanied by thyroid follicular cells. Mixed population of lymphocytes, sometimes represented in the form of crushed lymphocytes and lymphoglandular bodies (cytoplasmic debris) together with the presence of Hurthle cells were described in all cytology reports (Figure 3B as example). All of these characteristics were considered in the diagnosis of each and every patient in our cohort. Lastly, our referral population is broad but representative of an area considered iodine sufficient. | Discussion
This is the first study to show such a high prevalence of Hashimoto thyroiditis diagnosed by ultrasound-guided FNA cytology on a somewhat large cohort of patients. Based on our study, the prevalence of cytology-proven Hashimoto thyroiditis appears to be >10% in patients with thyroid nodules. Given that the prevalence of thyroid nodules on ultrasonography or autopsy data is as high as 50% [ 7 ], such a high prevalence of Hashimoto thyroiditis diagnosed by cytology is noteworthy. More strikingly, the prevalence of euthyroid, non-previously diagnosed, cytology-proven Hashimoto thyroiditis (euthyroid autoimmunity), appears to be >5% in our study (Figure 1 and 2 ). This condition has not been previously defined. For comparison, its prevalence is similar to that of type 2 diabetes which is considered to be a health care crisis.
Weetman [ 2 ] reported clinical Hashimoto thyroiditis prevalence rate at 1 in 182 or 0.55% in the US. In the UK, Tunbridge et al [ 8 ] reported an overall Hashimoto thyroiditis prevalence of 0.8%. However, based on our study, the cytology of Hashimoto thyroiditis seems to be much more prevalent, at 13.4%. This difference may be partially explained by the fact that for diagnosing clinical Hashimoto thyroiditis, abnormally elevated TSH, low thyroid hormones [ 2 , 8 ] and the confirmatory presence of thyroid autoantibodies are usually accounted for. We hypothesized then, that cytological diagnosis of Hashimoto's may precede clinical diagnosis. Interestingly however, in most organ specific autoimmune diseases, humoral immunity heralds tissue infiltrative damage. Hence, we expected the diagnosis of Hashimoto's by cytology to be less but not more common than the antibody diagnosis especially on early stages of the disease (Figure 2 ).
In an attempt at differentiating better the apparent stages of Hashimoto thyroiditis, we divided our cytology-proven Hashimoto's cohort into three subgroups: clinically hypothyroid, sub-clinical and euthyroid. For these distinctions, we used the clinical definition of hypothyroid Hashimoto thyroiditis as that occurring in patients usually with TSH greater than 10 ng/dl at diagnosis, with low free thyroid hormones and on thyroid hormone replacement. Based on NHANES III study [ 9 ] normative data for TSH distribution (reference population of 13,344, 95% of TSH between 0.45-4.12 ng/dl), abnormally high TSH (but usually less than 10 ng/dl) with normal thyroid hormones was used to define sub-clinical hypothyroidism. Patients with those characteristics were assigned to the sub-clinical subgroup in our study. A third subgroup classified as euthyroid included patients with normal thyroid hormones and normal TSH.
Aside from the unexpected observation of the high prevalence of Hashimoto thyroiditis by cytology, especially in euthyroid patients, the lack of cytological correlation with TPO autoantibody positivity is noteworthy. As mentioned before, the hallmark in the diagnosis of Hashimoto thyroiditis is the presence of TPO autoantibodies [ 10 ]. Yet, only about half of the patients tested for anti-TPO in the euthyroid subgroup were positive. On the other hand and although non-statistically significant because of the small number reported, the clinically hypothyroid subgroup had most (13 out of 16 tested) patients positive for anti-TPO. Similarly, Poropatich et al., [ 11 ] found that anti-TPO and/or antithyroglobulin antibody titers were present in only 50% of the patients with euthyroid, cytology-proven Hashimoto thyroiditis, a finding never reproduced by these or other authors in the literature.
Given the wide range of normal values for TSH (1 fold) and the variability on the presence of TPO autoantibodies, it is conceivable that early Hashimoto's autoimmune process might be clinically missed. These issues, together with the awareness that sub-clinical and clinical hypothyroidism associates with cardiovascular and neuropsychiatric morbidities, make finding high prevalence of Hashimoto thyroiditis on cytology, especially in euthyroid patients clinically significant [ 12 - 14 ].
An important aspect of our findings is the fact that most of the patients in our cohort are pre-menopausal females. Past and more recent studies found that the risk of poor obstetrical outcome is increased with relative thyroxine deficiency [ 15 - 18 ]. Benhadi et al. recently showed that the risk of miscarriage, fetal and neonatal death is increased with higher TSH levels. Moreover, the risk of fetal loss occurred even when maternal free thyroxine levels were normal [ 16 ]. Based on our results of high prevalence of cytology-proven Hashimoto thyroiditis and especially high prevalence of euthyroid disease in pre-menopausal women, there might be need of at least follow up of thyroid function on cytologically-proven Hashimoto's in pre-conception through delivery stages.
Positive Hashimoto's cytology clearly represents a state in which inflammation of thyroid epithelium may lead to tissue remodeling. Hence the possibility of acquiring mutations while cells are dividing might be greater. Some studies have suggested that inflammation of thyrocytes could be linked to thyroid cancer [ 19 - 21 ]. For example, Larson et al. showed that phosphatidylinositol 3-kinase phosphorilates Akt, which in turn suppresses pro-apoptotic signals and promotes tumorigenesis, was increased in both Hashimoto thyroiditis and well differentiated thyroid cancer [ 19 ]. Furthermore, Borrello et al., showed that RET/PTC oncogene, when exogenously expressed in normal thyrocytes, induces expression of a large set of genes involved in inflammation and tumor invasion (cytokines, matrix-degrading enzymes and adhesion molecules) [ 20 ]. There seem to be however, controversial opinions regarding the association of Hashimoto thyroiditis and thyroid cancer. Some studies find Hashimoto thyroiditis as associated with thyroid cancer [ 22 , 23 ] as examples, while others do not report an association [ 24 , 25 ] as examples. Since higher TSH level in patients with thyroid nodules has been found to be associated with risk of differentiated thyroid cancer [ 26 , 27 ], we hypothesized that active remodeling of thyroid epithelium in cytological Hashimoto's may explain in part the increased risk for differentiated thyroid cancer observed in patients with elevated but within normal TSH [ 26 , 27 ]. | Conclusions
We report here that the pathogenesis of Hashimoto thyroiditis is present in an unexpectedly large number of individuals. Moreover, we found that many of these individuals carry a non-clinically manifested process. Biopsy proven Hashimoto thyroiditis in asymptomatic, clinically disease-free patients is a condition still undefined. There are no guidelines as to how to follow these patients. What is their risk of progressing to full-blown Hashimoto's? How would pregnancy affect their thyroid function? Are they at further risk of developing thyroid cancer since their TSH usually runs in the normal but higher quartile? Recognition of this entity emerges as necessary at least in patients at risk for hypothyroid complications (i.e. pregnancy). If the cytology of Hashimoto thyroiditis represents a chronic active inflammatory state, TSH elevations within normal limits may represent a manifestation of a common underlying process. | Background
Our Thyroid-Multidisciplinary Clinic is a large referral site for thyroid diseases. Thyroid biopsies are mainly performed for thyroid cancer screening. Yet, Hashimoto thyroiditis (HT) is being too frequently diagnosed. The prevalence of HT is reported as 0.3-1.2% or twice the prevalence of type 1 diabetes. However, the prevalence of HT confirmed by cytology is still uncertain. To evaluate different aspects of thyroid physiopathology including prevalence of Hashimoto's, a database of clinical features, ultrasound images and cytology results of patients referred for FNA of thyroid nodules was prospectively developed.
Methods
We retrospectively studied 811 consecutive patients for whom ultrasound guided thyroid FNA biopsies were performed at our clinic over 2.5 year period (Mar/2006-Sep/2008).
Results
The analysis of our database revealed that from 761 patients, 102 (13.4%) had HT, from whom 56 (7.4%) were euthyroid or had sub-clinical (non-hypothyroid) disease, and 46 (6%) were clinically hypothyroid.
Conclusions
This is the first study to show such a high prevalence of HT diagnosed by ultrasound-guided FNA. More strikingly, the prevalence of euthyroid HT, appears to be >5% similar to that of type 2 diabetes. Based on our results, there might be a need to follow up on cytological Hashimoto's to monitor for thyroid failure, especially in high risk states, like pregnancy. The potential risk for thyroid cancer in patients with biopsy-proven inflammation of thyroid epithelium remains to be established prospectively. However, it may explain the increased risk for thyroid cancer observed in patients with elevated but within normal TSH. | List of abbreviations used
HT: Hashimoto thyroiditis; FHx: Family history; FT4: Free T4; FNA: Fine needle aspiration; TPO: Thyroid peroxidase.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
All authors participated in the design of the study. AS and JCJ performed the statistical analysis. JCJ conceived of the study. All authors read and approved the final manuscript. | Aknowledgements
We would like to thank Drs. Haymart, Heras-Herzig, Gogineni, Bradley, and Poehls for their active participation in the ongoing University of Wisconsin Thyroid Database Study and Dr. Ernest L. Mazzaferri for critical review of this paper. This work was funded in part by NIH and VA grants (JCJ) and presented in part at the 2009 Endocrine Society Annual Meeting, USA. | CC BY | no | 2022-01-12 15:21:44 | Thyroid Res. 2010 Dec 20; 3:11 | oa_package/b1/c4/PMC3016247.tar.gz |
PMC3016248 | 21156050 | Introduction
This paper examines the issue of workforce stability and turnover in the context of policy attempts to improve retention of health workers. Staff turnover is often the primary topic for monitoring and research when retention is being examined, and can give insights into trends in outflow from the health care organisation. This is particularly relevant at a time of global HRH shortages [ 1 , 2 ]. However a focus only on turnover- on those who leave, is only part of the picture. Policy makers and managers also require an insight into why some staff stay, and what "staying" and workforce stability can contribute to, service delivery, staff workload, and the multiple dimensions of organisational performance, including costs. This paper argues that there are significant benefits to supporting policy makers and managers to develop a broader perspective of workforce stability and methods of monitoring it. The objective of the paper is therefore to contribute to developing a better understanding of workforce stability as a major aspect of the overall policy goal of improved retention of health workers. | Methods
The paper is based on a desk review of published and official sources, and analysis of workforce data from official sources. | Results and Discussion
There is a paucity of research which fully examines the complex interaction between staff turnover and organisational performance, especially quality of care in the health sector. There have been some exploratory studies, and some which have looked at staffing levels and outcomes, and examined turnover rates as one proxy measure for variations in staff satisfaction [ 5 - 9 ]. Few have taken the impact of turnover or stability rates on performance as the primary focus of examination.
Where such research has been conducted, mainly in developed countries, and focusing on nursing staff, there are some powerful messages for policy makers and managers. For example, one US study reported that health care organizations with the lowest nurse turnover rates (less than 12 percent) had the lowest risk-adjusted mortality scores, as well as the lowest severity-adjusted length-of-stay, and that for health care organizations with turnover rates in excess of 22 percent, the severity adjusted average length-of-stay was 1.2 days longer than those with the lowest turnover rates. The authors noted that whilst these findings do not establish a causal relationship, they did suggest that higher rates of turnover among the nursing staff probably lead to decreased efficiency and productivity, which in turn affects patient care [ 10 ]. Other detailed research studies are underway which will focus on the links between turnover and measures of outcome [ 11 , 12 ]; others have also identified the need to examine the impact of so-called staff "churn": a continuous high level of turnover, often accompanied by vacancies and reliance on short term cover by temporary staffing[ 13 ].
Whilst there may be relatively little research evidence on the impact of health workforce turnover/stability on care outcomes, there remains a need for organisations to be able to measure workforce turnover and stability, and assess its impact on costs..
Most research which focuses on health worker retention and which attempts to measure it uses turnover as a main indicator. Turnover, and the alternate terms of "attrition" or "wastage" are usually expressed in terms of the % of staff of a particular occupation or workplace who have left the organisation (or have moved jobs) within the last twelve months. The terms "attrition" and "wastage" are normally applied when the measurement is the % of staff who have left the organisation or system under scrutiny, whilst "turnover" is more often applied if all job moves (including those within the system) are being assessed [ 1 - 3 , 14 ].
This type of turnover data is routinely used as a method of comparing the "leaving" rates in different workplace units or organisations, and can be a method of benchmarking variations in rates across systems or organisations. A number of standard measures of turnover can be used, the main requirement being consistency of definition and application, to allow comparison over time and between employing units. The recent WHO/World Bank/USAID handbook [ 2 ] has suggested that the "workforce loss ratio" can be calculated by using number of workers who have left in the last year as numerator, and total number of health workers as the denominator . This is the standard approach- determining the loss over the year as a proportion of the total workforce [see also [ 14 ]].
If the reason for monitoring is to assess the ability of the organisation or system under scrutiny to retain staff it is important that any monitoring can differentiate between permanent and temporary moves, and what has been termed "involuntary" turnover or attrition (e.g. statutory retirement, ill health, death) and "voluntary" turnover, resignation or attrition [ 2 , 15 ]. Research using turnover or attrition data can give policy makers an insight into varying rates in different cadres of worker, and different reasons for attrition. One recent study, for example, highlighted that attrition of doctors and registered nurses in Kenya was much higher at provincial hospitals than at district hospitals or health centres, whereas the opposite pattern was found for laboratory staff and pharmacists [ 16 ]. It also found that resignation was the main factor in attrition of doctors and clinical officers, whilst the main reason for attrition of nurses was retirement.
One limitation of the standard measure of turnover (leavers in the year/total workforce) is that over the time period under measure it does necessarily differentiate well when comparing units with high and repeated turnover in a few posts, and those with lower turnover in more posts [ 13 , 17 ]. This can limit its utility as a measure of retention as it measures the leaving rate and gives less insight into how many staff are staying, and for how long. As noted earlier, a focus on turnover is a focus on the "leavers", and there should also be some consideration of "stayers" if retention is the focus. Developing an understanding of the stability in the workforce can provide managers and policy makers with a better understanding of the labour dynamics both internally, within the organisation, and externally, in terms of how the organisation connects to the wider labour markets. There are several different possible measures of workforce stability which can be used to give an insight into how many staff stay with the organisation, for what time period.
One indicator which can be used to assess workforce stability is the average years in post reported by staff group or work location [ 18 ]. A second measure of stability is to calculate the stability rate or "index" for each location, occupational group or profession. The stability index assesses the proportion of staff who were in post at the beginning of the year who were still in post at the end of the year. As data is collected over a longer period of time, stability indices for 1 year, 2 years (i.e the % of staff who have remained in post at end of two years), and longer can be calculated, giving a better insight into the relative rates of retention of different types of staff, or different organisations within the system under scrutiny. This type of measure may be of utility in most healthcare settings where HRH data collection is limited, rather than more sophisticated indicators that require
Table 1 gives an example of the use of stability index data. It is based on data collected by the National Health Service (NHS) in Scotland. There are 14 main regional employing organisations (Boards) within the NHS in Scotland and the Table presents the one and two year stability rates for all staff nationally, and in each of the fourteen Boards. The overall workforce (full time equivalent) is approximately 130,000, including approximately 11,000 doctors and 40,000 registered nurses. Boards vary significantly in size; the largest employs over 35,000 staff. Two year stability rates for nurses and doctors are also shown. The rates are calculated by estimating the percentage of staff that were in substantive posts at 30 September and who were still in substantive posts within the National Health Service (NHS) in Scotland in the same NHS Board and the same staff group a year later (index 1), and two years later (index 2).
The data is presented for each Board, in anonymised form, with a ranking of the two year stability rate for all staff (lowest stability rate at top- "Board A"). A quick examination of the data in Table 1 provides some immediate insights and underlines how useful stability data can be, particularly if it is collated beyond one year. For example it is evident that there is marked variation between Boards in the stability rate for all staff - the 2 year rate varies between 67% and 80.5%. The two year stability rates for doctors and for nurses are higher than the average for all staff groups. This is perhaps not unsurprising as these health professional groups are better paid and more likely to be located within career structures which will assist in retaining them in the health sector than are some "support" staff in clerical and administrative jobs; furthermore, as health professionals their skills will be less transferable to non health settings. There is also variation in the ranking of the two year rate- Board A which has the lowest stability rate for all staff, does not report the lowest rate for doctors (Board I) or nurses (Board N). Assessment of the data by size of Board can also give some indication of the possible impact of organisation size, but also has to be qualified by the recognition that "smaller" Boards (by workforce size) tend to be relatively rural, remote or island based.
This relatively straightforward measure of stability can assist in inform managers and policy makers about which occupations have relatively good or poor stability, or which organisations are retaining proportionately more staff for more time. As such it can point to "problem" areas. These may be areas with low stability where greater staff stability is required. The analysis can also identify organisations with high desired stability rates which may have best practice methods that can be identified and networked.
One of the benefits of this type of measure of stability is its utility. With some caveats about interpretation of data where the workforce populations are small [ 13 , 17 ] the stability rate is simple to measure, and simple to understand. It can be no more complex to collect data on stability than it is on turnover. Where new HR information systems are being developed, or current systems upgraded, consideration should be given to collating and standardising the use of the stability rate indicator. More complex measures of stability can also be used if an employing organisation wishes to develop a more detailed insight into stability and the length-of-service structure of the staff to facilitate inter-organisation comparisons [ 19 ].
Whilst routine collection and analysis of turnover and stability data can give an insight into level of staff retention and reasons for staying, or leaving, an estimate of the costs of turnover can assist in raising awareness of the need to improve retention, and can contribute to advocacy for policy interventions to improve retention [ 3 , 8 , 10 , 20 , 21 ]. In addressing issues related to costs of turnover, it is important first to note that there can benefits as well as costs to the employing organisation when it experiences staff turnover (see Table 2 ). The critical issue, as noted earlier, is that any unplanned and unnecessary ("voluntary") turnover should be prevented, if at all possible, in the interests of achieving greater stability in the workforce.
From a management perspective, the potential benefits of some level of turnover can include 'freeing up' of posts to allow new staff with new ideas and energy into the organisation, opportunities for cost reduction through staff reduction or deployment, and through 'losing' disaffected staff who would otherwise exhibit other forms of withdrawal behaviour, such as absence from work. There is however no consensus on what is the "ideal" level of turnover[ 14 ], and some potential benefits may be only short term.
However, as noted earlier in the paper it is clear that there can be a range of organisational costs and negative impacts on care associated with any unnecessary turnover of staff. Assessing the cost of turnover should therefore be an integral element in the approach to maintaining workforce stability. If the typical or average cost of turnover of staff is known, this can assist in assessing the impact on the organisation of different levels of turnover, and can also provide managers with an indication of the cost effectiveness of improving retention through turnover reduction (or stability increase) strategies.
There is little published data or information on turnover costs in the health sector and much of that which does exist stems from the United States. Most studies attempt to arrive at a cost per individual staff member "leaving", and then calculate a total organisational cost per annum. Costs per staff are usually examined in at least four components- separation costs (the costs incurred by the staff member leaving), temporary replacement costs (the costs of covering the post made vacant by the staff member leaving- e.g use of an agency staff, sue of overtime work by remaining staff etc), recruitment costs (the costs in advertising and selecting the replacement, and providing relocation costs), and induction costs (including "lost" productivity, until the replacement reaches the same level of productivity as the staff member who had left) [ 3 , 8 , 20 , 21 ]. Initial costs of training the worker are usually not included.
It should be recognised that actual turnover costs may vary significantly between individual employees, depending on the grade and experience of the worker, and on the replacement strategy used by the employer [ 3 , 20 , 21 ]. For experienced staff at senior level it is likely that turnover costs will be significantly higher than for junior staff, due to longer periods of induction and, in some specialities and occupations, because of skills shortages and difficulties in recruitment.
Turnover costs will also vary depending upon the replacement strategy being adopted (e.g. replacing an experienced worker with a less experienced worker is likely to lead to lower replacement costs, but lower productivity, in the short term at least) and are likely to vary according to the clinical and geographical setting.
There are a number of ways that the overall impact of staff turnover costs at organisation level can be illustrated, such as:
• percentage of paybill;
• cost per patient day;
• cost saving of reduction in turnover
Table 3 gives an example using % of paybill.. Using assumptions of 7 per cent turnover and turnover costs of $8,000 per nurse in an organisation which employs 500 nurses, this would be equivalent to turnover costs of $280,000 per annum. If the turnover was reduced to 5 per cent per annum, the illustrative example suggests a cost saving to the organisation of $80,000 per annum.
This approach to estimating turnover costs ensures that there is sufficient detail at the level of the individual 'leaver' to allow aggregated data to be used as a 'not less than' total cost to the organisation. The example in Table 3 above illustrates the potential magnitude of turnover costs at organisational level, and also reveals the potential cost savings which management could achieve by reducing turnover. Clearly, as noted above, not all turnover is "bad" for the organisation, but any turnover which could have been prevented through management action would have reduced turnover costs. Improving workforce stability can carry with it a benefit to the organisation in terms of reduced cost.
One recent study in the US [ 22 ] estimated that total turnover costs for a hospital system employing 5000 employees was between $US17 and $29 million. Another study [ 21 ] focusing on costs of individual turnover rather than organisational, highlighted nurse turnover cost assessments of between US$21,500 and US$31,500 per nurse, with a rough "rule of thumb" that the cost of staff nurse turnover normally fell between 0.75 and 2.0 times annual salary, depending on replacement strategy and other factors. | Results and Discussion
There is a paucity of research which fully examines the complex interaction between staff turnover and organisational performance, especially quality of care in the health sector. There have been some exploratory studies, and some which have looked at staffing levels and outcomes, and examined turnover rates as one proxy measure for variations in staff satisfaction [ 5 - 9 ]. Few have taken the impact of turnover or stability rates on performance as the primary focus of examination.
Where such research has been conducted, mainly in developed countries, and focusing on nursing staff, there are some powerful messages for policy makers and managers. For example, one US study reported that health care organizations with the lowest nurse turnover rates (less than 12 percent) had the lowest risk-adjusted mortality scores, as well as the lowest severity-adjusted length-of-stay, and that for health care organizations with turnover rates in excess of 22 percent, the severity adjusted average length-of-stay was 1.2 days longer than those with the lowest turnover rates. The authors noted that whilst these findings do not establish a causal relationship, they did suggest that higher rates of turnover among the nursing staff probably lead to decreased efficiency and productivity, which in turn affects patient care [ 10 ]. Other detailed research studies are underway which will focus on the links between turnover and measures of outcome [ 11 , 12 ]; others have also identified the need to examine the impact of so-called staff "churn": a continuous high level of turnover, often accompanied by vacancies and reliance on short term cover by temporary staffing[ 13 ].
Whilst there may be relatively little research evidence on the impact of health workforce turnover/stability on care outcomes, there remains a need for organisations to be able to measure workforce turnover and stability, and assess its impact on costs..
Most research which focuses on health worker retention and which attempts to measure it uses turnover as a main indicator. Turnover, and the alternate terms of "attrition" or "wastage" are usually expressed in terms of the % of staff of a particular occupation or workplace who have left the organisation (or have moved jobs) within the last twelve months. The terms "attrition" and "wastage" are normally applied when the measurement is the % of staff who have left the organisation or system under scrutiny, whilst "turnover" is more often applied if all job moves (including those within the system) are being assessed [ 1 - 3 , 14 ].
This type of turnover data is routinely used as a method of comparing the "leaving" rates in different workplace units or organisations, and can be a method of benchmarking variations in rates across systems or organisations. A number of standard measures of turnover can be used, the main requirement being consistency of definition and application, to allow comparison over time and between employing units. The recent WHO/World Bank/USAID handbook [ 2 ] has suggested that the "workforce loss ratio" can be calculated by using number of workers who have left in the last year as numerator, and total number of health workers as the denominator . This is the standard approach- determining the loss over the year as a proportion of the total workforce [see also [ 14 ]].
If the reason for monitoring is to assess the ability of the organisation or system under scrutiny to retain staff it is important that any monitoring can differentiate between permanent and temporary moves, and what has been termed "involuntary" turnover or attrition (e.g. statutory retirement, ill health, death) and "voluntary" turnover, resignation or attrition [ 2 , 15 ]. Research using turnover or attrition data can give policy makers an insight into varying rates in different cadres of worker, and different reasons for attrition. One recent study, for example, highlighted that attrition of doctors and registered nurses in Kenya was much higher at provincial hospitals than at district hospitals or health centres, whereas the opposite pattern was found for laboratory staff and pharmacists [ 16 ]. It also found that resignation was the main factor in attrition of doctors and clinical officers, whilst the main reason for attrition of nurses was retirement.
One limitation of the standard measure of turnover (leavers in the year/total workforce) is that over the time period under measure it does necessarily differentiate well when comparing units with high and repeated turnover in a few posts, and those with lower turnover in more posts [ 13 , 17 ]. This can limit its utility as a measure of retention as it measures the leaving rate and gives less insight into how many staff are staying, and for how long. As noted earlier, a focus on turnover is a focus on the "leavers", and there should also be some consideration of "stayers" if retention is the focus. Developing an understanding of the stability in the workforce can provide managers and policy makers with a better understanding of the labour dynamics both internally, within the organisation, and externally, in terms of how the organisation connects to the wider labour markets. There are several different possible measures of workforce stability which can be used to give an insight into how many staff stay with the organisation, for what time period.
One indicator which can be used to assess workforce stability is the average years in post reported by staff group or work location [ 18 ]. A second measure of stability is to calculate the stability rate or "index" for each location, occupational group or profession. The stability index assesses the proportion of staff who were in post at the beginning of the year who were still in post at the end of the year. As data is collected over a longer period of time, stability indices for 1 year, 2 years (i.e the % of staff who have remained in post at end of two years), and longer can be calculated, giving a better insight into the relative rates of retention of different types of staff, or different organisations within the system under scrutiny. This type of measure may be of utility in most healthcare settings where HRH data collection is limited, rather than more sophisticated indicators that require
Table 1 gives an example of the use of stability index data. It is based on data collected by the National Health Service (NHS) in Scotland. There are 14 main regional employing organisations (Boards) within the NHS in Scotland and the Table presents the one and two year stability rates for all staff nationally, and in each of the fourteen Boards. The overall workforce (full time equivalent) is approximately 130,000, including approximately 11,000 doctors and 40,000 registered nurses. Boards vary significantly in size; the largest employs over 35,000 staff. Two year stability rates for nurses and doctors are also shown. The rates are calculated by estimating the percentage of staff that were in substantive posts at 30 September and who were still in substantive posts within the National Health Service (NHS) in Scotland in the same NHS Board and the same staff group a year later (index 1), and two years later (index 2).
The data is presented for each Board, in anonymised form, with a ranking of the two year stability rate for all staff (lowest stability rate at top- "Board A"). A quick examination of the data in Table 1 provides some immediate insights and underlines how useful stability data can be, particularly if it is collated beyond one year. For example it is evident that there is marked variation between Boards in the stability rate for all staff - the 2 year rate varies between 67% and 80.5%. The two year stability rates for doctors and for nurses are higher than the average for all staff groups. This is perhaps not unsurprising as these health professional groups are better paid and more likely to be located within career structures which will assist in retaining them in the health sector than are some "support" staff in clerical and administrative jobs; furthermore, as health professionals their skills will be less transferable to non health settings. There is also variation in the ranking of the two year rate- Board A which has the lowest stability rate for all staff, does not report the lowest rate for doctors (Board I) or nurses (Board N). Assessment of the data by size of Board can also give some indication of the possible impact of organisation size, but also has to be qualified by the recognition that "smaller" Boards (by workforce size) tend to be relatively rural, remote or island based.
This relatively straightforward measure of stability can assist in inform managers and policy makers about which occupations have relatively good or poor stability, or which organisations are retaining proportionately more staff for more time. As such it can point to "problem" areas. These may be areas with low stability where greater staff stability is required. The analysis can also identify organisations with high desired stability rates which may have best practice methods that can be identified and networked.
One of the benefits of this type of measure of stability is its utility. With some caveats about interpretation of data where the workforce populations are small [ 13 , 17 ] the stability rate is simple to measure, and simple to understand. It can be no more complex to collect data on stability than it is on turnover. Where new HR information systems are being developed, or current systems upgraded, consideration should be given to collating and standardising the use of the stability rate indicator. More complex measures of stability can also be used if an employing organisation wishes to develop a more detailed insight into stability and the length-of-service structure of the staff to facilitate inter-organisation comparisons [ 19 ].
Whilst routine collection and analysis of turnover and stability data can give an insight into level of staff retention and reasons for staying, or leaving, an estimate of the costs of turnover can assist in raising awareness of the need to improve retention, and can contribute to advocacy for policy interventions to improve retention [ 3 , 8 , 10 , 20 , 21 ]. In addressing issues related to costs of turnover, it is important first to note that there can benefits as well as costs to the employing organisation when it experiences staff turnover (see Table 2 ). The critical issue, as noted earlier, is that any unplanned and unnecessary ("voluntary") turnover should be prevented, if at all possible, in the interests of achieving greater stability in the workforce.
From a management perspective, the potential benefits of some level of turnover can include 'freeing up' of posts to allow new staff with new ideas and energy into the organisation, opportunities for cost reduction through staff reduction or deployment, and through 'losing' disaffected staff who would otherwise exhibit other forms of withdrawal behaviour, such as absence from work. There is however no consensus on what is the "ideal" level of turnover[ 14 ], and some potential benefits may be only short term.
However, as noted earlier in the paper it is clear that there can be a range of organisational costs and negative impacts on care associated with any unnecessary turnover of staff. Assessing the cost of turnover should therefore be an integral element in the approach to maintaining workforce stability. If the typical or average cost of turnover of staff is known, this can assist in assessing the impact on the organisation of different levels of turnover, and can also provide managers with an indication of the cost effectiveness of improving retention through turnover reduction (or stability increase) strategies.
There is little published data or information on turnover costs in the health sector and much of that which does exist stems from the United States. Most studies attempt to arrive at a cost per individual staff member "leaving", and then calculate a total organisational cost per annum. Costs per staff are usually examined in at least four components- separation costs (the costs incurred by the staff member leaving), temporary replacement costs (the costs of covering the post made vacant by the staff member leaving- e.g use of an agency staff, sue of overtime work by remaining staff etc), recruitment costs (the costs in advertising and selecting the replacement, and providing relocation costs), and induction costs (including "lost" productivity, until the replacement reaches the same level of productivity as the staff member who had left) [ 3 , 8 , 20 , 21 ]. Initial costs of training the worker are usually not included.
It should be recognised that actual turnover costs may vary significantly between individual employees, depending on the grade and experience of the worker, and on the replacement strategy used by the employer [ 3 , 20 , 21 ]. For experienced staff at senior level it is likely that turnover costs will be significantly higher than for junior staff, due to longer periods of induction and, in some specialities and occupations, because of skills shortages and difficulties in recruitment.
Turnover costs will also vary depending upon the replacement strategy being adopted (e.g. replacing an experienced worker with a less experienced worker is likely to lead to lower replacement costs, but lower productivity, in the short term at least) and are likely to vary according to the clinical and geographical setting.
There are a number of ways that the overall impact of staff turnover costs at organisation level can be illustrated, such as:
• percentage of paybill;
• cost per patient day;
• cost saving of reduction in turnover
Table 3 gives an example using % of paybill.. Using assumptions of 7 per cent turnover and turnover costs of $8,000 per nurse in an organisation which employs 500 nurses, this would be equivalent to turnover costs of $280,000 per annum. If the turnover was reduced to 5 per cent per annum, the illustrative example suggests a cost saving to the organisation of $80,000 per annum.
This approach to estimating turnover costs ensures that there is sufficient detail at the level of the individual 'leaver' to allow aggregated data to be used as a 'not less than' total cost to the organisation. The example in Table 3 above illustrates the potential magnitude of turnover costs at organisational level, and also reveals the potential cost savings which management could achieve by reducing turnover. Clearly, as noted above, not all turnover is "bad" for the organisation, but any turnover which could have been prevented through management action would have reduced turnover costs. Improving workforce stability can carry with it a benefit to the organisation in terms of reduced cost.
One recent study in the US [ 22 ] estimated that total turnover costs for a hospital system employing 5000 employees was between $US17 and $29 million. Another study [ 21 ] focusing on costs of individual turnover rather than organisational, highlighted nurse turnover cost assessments of between US$21,500 and US$31,500 per nurse, with a rough "rule of thumb" that the cost of staff nurse turnover normally fell between 0.75 and 2.0 times annual salary, depending on replacement strategy and other factors. | Conclusion
Intuition would suggest that improving retention and stability of the health workforce brings benefits to staff, the organisation and those being cared for. The limited research available does provide some support for this, and gives insight into how health workforce stability can contribute to reduced costs, improved productivity and better care outcomes. Of equal importance, and more immediate utility in most organisations, is the possibility of introducing a measure or indicator of staff stability, and of examining the costs of staff turnover. These types of assessment can be valuable to managers and policy makers as they examine which policies may be effective in improving stability and retention, by reducing turnover. They can also be used as part of advocacy for the use of new retention measures. An additional benefit can be that the very action of setting up a local working group to assess the costs of turnover can in itself gives managers and staff a greater insight into the negative impacts of turnover, and can encourage them to work together to identify and implement stability measures. | This paper examines the issue of workforce stability and turnover in the context of policy attempts to improve retention of health workers. The paper argues that there are significant benefits to supporting policy makers and managers to develop a broader perspective of workforce stability and methods of monitoring it. The objective of the paper is to contribute to developing a better understanding of workforce stability as a major aspect of the overall policy goal of improved retention of health workers. The paper examines some of the limited research on the complex interaction between staff turnover and organisational performance or quality of care in the health sector, provides details and examples of the measurement of staff turnover and stability, and illustrates an approach to costing staff turnover. The paper concludes by advocating that these types of assessment can be valuable to managers and policy makers as they examine which policies may be effective in improving stability and retention, by reducing turnover. They can also be used as part of advocacy for the use of new retention measures. The very action of setting up a local working group to assess the costs of turnover can in itself give managers and staff a greater insight into the negative impacts of turnover, and can encourage them to work together to identify and implement stability measures. | Background
Why should the issue of health workforce stability be important? Retaining and developing the workforce ("talent management") is generally regarded as a major human resource objective for any organisation. In health care there is a general assumption that staff turnover (the opposite of stability), will negatively effect both access to care, and the level and quality of healthcare being provided. Turnover may reduce staffing and patient contact time; can add to organisational costs, if temporary cover for staff who leave (e.g. overtime pay) and recruitment of replacements incurs additional costs; and may reduce individual and organisational performance through the loss of experienced staff, and by undermining teamwork [ 3 , 4 ].
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
JB conceived, researched and wrote the paper | CC BY | no | 2022-01-12 15:21:44 | Hum Resour Health. 2010 Dec 14; 8:29 | oa_package/cc/6f/PMC3016248.tar.gz |
|
PMC3016249 | 21118540 | Background
Invariant natural killer-like T (iNKT) cells are placed ambiguously between adaptive and innate immune systems (reviewed in [ 1 , 2 ]). Derived from the thymus, expressing rearranged T-cell receptor (TCR) alpha and beta chains, they seem to belong to the adaptive immune system, while their receptor homogeneity, their continuous state of activation, their rapid secretion of large amounts of interferon (IFN)-γ and interleukin (IL)-4, their presumed recognition of invariant glycolipid self-ligands associated with the non-classical major histocompatibility (MHC) class I molecule, CD1 d, recall various aspects of innate immune recognition. Many features of iNKT cell behaviour are puzzling: their thymic ontogeny and relation to the classical pathways of T-cell mediated differentiation; the relative importance of endogenous and exogenous ligands in activation; and the polarity of their cytokine profile towards IFN-γ or IL-4 in relation to the activating ligand. However, in this report we address the basis for another characteristic of these enigmatic cells, namely their constitutive state of readiness to respond with massive cytokine production (reviewed in [ 3 ]). Using a sensitive endogenous reporter for IFN-γ production, we show that iNKT cells are constitutively and endogenously activated to IFN-γ production in the liver and kidney of normal mice. The activation is apparently restricted to these tissues, focal, spontaneous and independent of signals derived from bacteria or viruses. It is, however, dependent on the expression of CD1 d and on the presence of the classical iNKT cell receptor. The results suggest that the constitutive state of iNKT readiness may be maintained by intermittent local stimulation with endogenous ligands. | Methods
Mice and organ preparation
C57BL/6J mice were obtained from the specific pathogen-free mouse facility in the Institute for Genetics Uni. Köln. We are enormously indebted to the many colleagues and institutions that provided mice or organs for this study. TLR2 -/- [ 36 ], TLR4 -/- [ 37 ], TLR9 -/- [ 38 ], TLR2/4 -/- [ 39 ] and MyD88 -/- [ 40 ] mice were provided by Professor Dr Marina Freudenberg (Max Planck Institute for Immunobiology, Freiburg). IFN-γ -/- mice [ 8 ], IFN-receptor deficient mice (IFNAR, [ 10 ]), (IFNGR, [ 9 ]) and STAT-1 -/- mice [ 41 ] were provided by Dr Thomas Kolbe (University of Veterinary Medicine, Vienna). RAG-1 -/- mice [ 13 ] were provided by Dr Heike Weighardt (Technische Universitaet Muenchen, Munich). JHT mice [ 14 ] were provided by Dr Ari Waisman (Johannes Gutenberg-Universität Mainz, Mainz, Germany). Germ-free C57BL/6 mice were provided from three independent facilities, namely, University of Zürich, Zürich, Switzerland (Mr Rudolf Jörg), the Gulbenkian Institute of Science, Oeiras, Portugal (Dr Jocelyne Demengeot) and the Karolinska Institute, Stockholm. CD1d -/- mice [ 21 ] were provided by Professor Luc van Kaer (Vanderbilt University Medical Center, TN, USA). Jα18 -/- mice [ 22 ] were provided by Dr Maria C. Leite-de-Moraes, Hôpital Necker, Paris, France. MAVS/CARDIF-deficient mice [ 30 ] were provided by Dr Jürg Tschopp (Lausanne, Switzerland). MHC Aβ -/- mice [ 23 ] were provided by Dr Jocelyne Demongeot (Gulbenkian Institute of Sciences, Oeiras, Portugal). TAP1 -/- mice [ 24 ] were purchased from the Jackson Laboratory (Maine, US). For all assays, wild-type C57BL/6 mice from the same donor facility were used as controls. To control germ-free C57BL/6 mice, C57BL/6 mice from colonies maintained at SPF level from the same suppliers were used. All the above mice were on the C57BL/6J background. SWR/J, SJL, C57BR/J, C57L/J, RIIIS/J, A/J and C57BR/cdJ strains were purchased from The Jackson Laboratory (Maine, USA). In most cases, liver and kidneys were removed at the donor animal facility, immediately fixed in cold TBS/4% paraformaldehyde and shipped at 4°C by the quickest route to the Institute for Genetics in Cologne, Germany.
Tissue preparation
For paraffin sections, mouse tissues were fixed in TBS/4% paraformaldehyde at 4°C and dehydrated through an ethanol series at 4°C (50%, 70%, 90% and 96%). Tissues were then transferred into isopropanol and finally into a paraffin:isopropanol (1:1) solution at 60°C. The isopropanol was evaporated and fresh paraffin was then replenished several times at 60°C before the tissues (in paraffin) were moved to room temperature. The embedded tissues were cut with a microtome RM 2065 (Leica Microsystems, Wetzlar, Germany) into 6 μm thick serial sections. For cryosections, mouse liver was snap-frozen in liquid nitrogen and cut into consecutive serial sections (6 μm) using a cryotome CM 3050 S (Leica Microsystems).
Immunohistochemistry
Paraffin sections were de-waxed in xylene and post-fixed in 4% paraformaldehyde (1 h, RT). For Irga6 or Irgm3 staining, protein epitopes were demasked (10 min, boiling) in 10 mM citrate buffer (pH 6.0). For F4/80 or CD3 staining, protein epitopes were demasked by 0.1% Trypsin (SIGMA Type II) solution (0.1% CaCl 2 , pH7.8) for 30mins at 37°C or in 1 mM EDTA (pH8.0) for 10 min at 100°C, respectively. Sections were then saturated with quenching buffer (0.3% H 2 O 2, 20 min). After phosphate buffered saline (PBS) washing, the sections were probed either with rabbit anti-Irga6 antiserum 165/3 [ 42 ] or mouse anti-Irgm3 mAb (BD Transduction Laboratory) in DAKO diluent. HRP-staining was carried out with the horseradish peroxidase (HRP)-substrate kit HistoGreen (Linaris, Wetzlar, Germany). Nuclei were counterstained with Nuclear Fast Red. For Irga6 and F4/80 or CD3 double staining, Irga6 was first probed with 165/3 and stained with HRP-substrate AEC (Sigma, MO, USA). The sections were then washed with PBS (1 h) and subsequently probed with anti-F4/80 mAb (Serotec, NC, USA) or anti-CD3 mAb (Vector Laboratories, Linaris). The AP-staining was then carried out with AP-substrate kit III (Vector laboratory, Linaris). For staining of cryo-sections, sections were fixed first with cold acetone (10 min), before probing with anti-Irga6 antiserum 165/3 or anti-TCR Vβ8.1/8.2 mAb (BD Pharmingen, CA, USA). The sections were then stained with the HRP-substrate kit HistoGreen for Irga6 or with AP-substrate kit III for TCR Vβ 8. Samples were analysed using Zeiss Axioplan II microscope (Zeiss, Jena, Germany) equipped with SPOT RT slider digital camera (Diagnostic Instruments, MI, USA).
Quantitation of Irga6 expression foci in histological sections
Irga6 high expression foci were counted on stained 6 μm histological sections using 100× total magnification. Sections 300 μm apart were examined in order to avoid repeated counts of the same high expression patch. A total of 30 fields were counted from each organ, from which the mean number of high expression patches per field was estimated. The raw data from all such determinations are presented in Additional file 1 : Table S1.
Laser microdissection
Consecutive serial cryosections (6 μm) for liver were prepared using a cryotome CM 3050 S (Leica Microsystems) and sections were carried on PALM MembraneSlides (P.A.L.M. Microlaser Technologies, Bernried, Germany), which was covered with a polyethylene naphthalate membrane. In order to avoid the RNA degradation caused by the staining of sections with antibody, the following procedure was followed. Consecutive sections were numbered (1, 2, 3, 4...). All sections with odd-numbers (1, 3...) were on one slide (named 'A') and the sections with even numbers (2, 4...) on another slide (named 'B'). Both slides were then fixed in 70% ethanol (-20°C). The B slide was stored at -80°C and the A slide was stained for Irga6 with 165/3 using histogreen substrate. Geographical locations of Irga6 focal expression on the A slide were recorded photographically. The B slide was then stained only with Nuclear Fast Red, dehydrated in an ethanol series (70%, 96%, 100% 2 min each) and dried at 50°C. Irga6 focal expression was located on the B slide by referring to the photos of adjacent sections on the A slide. The Irga6 expression foci and non-focus region on the B slides were then collected by laser microdissection (LMD) equipment generously made available by Dr Margarete Odenthal (Institute for Pathology, University of Cologne, Cologne, Germany) consisting of an Interface Microbeam Mini laser (P.A.L.M Microlaser Technologies) and an Axiovert 135 microscope (Zeiss). Total RNA from the dissected samples (50-100 foci) was extracted with the RNeasy Micro kit (QIAGEN, Hilden, Germany) and cDNA was synthesized with half of the total RNA with the Superscript First-Strand Synthesis System for RT-PCR kit (Invitrogen, CA, USA).
Real-time PCR
PCR was carried out in a Light Cycler I System (Roche, Berlin, Germany) using a LightCycler SYBR Green I PCR kit (Roche). The 5' primers specific for Irga6 1A and 1B together with the common 3' primer on the coding exon were used to detect expression level of Irga6-1A and Irga6-1B transcript forms as target genes [ 4 ]. Mouse GAPDH gene was used as a reference gene and input control (for primers see Additional file 4 : Table S2). 2-4 μL of cDNA from materials collected by LMD was used as templates. Primer efficiency for Irga6-1A, -1B or mGAPDH was determined using liver cDNA dilution series (1, 1/10, 1/100...) as templates as described previously [ 43 ]. The proportional increase in Irga6-1A or -1B expression for 'foci' to 'non-foci' was then determined according to [ 43 ]. Melting curve analyses were performed in order to verify the amplification specificity. Each sample was tested in duplicate or triplicate. In order to determine whether the final PCR products amplified for Irga6-1A were -1A specific, 1A PCR products were cloned in to pGEM-T-easy (Promega, CA, USA) vector and sequenced. The percentage of -1A specific clones was then calculated. For Irga6 high expression foci, all clones amplified for Irga6-1A were -1A, while for the non-focal material, only 5.9% of clones amplified for Irga6-1A were -1A. The directly determined enrichment factor of 27.9 for Irga6-1A in Irga6 high expression foci (Figure 2C ) can, therefore, be multiplied a factor of 16.9 (100/5.9) in order to give a relative enrichment of over 470 in Irga6-1A compared with non-focal tissue.
Nested-RT-PCR
Total RNA from mouse liver or lymph nodes was extracted using the RNeasy Mini kit (QIAGEN) and cDNA was synthesized using the Superscript First-Strand Synthesis System for RT-PCR kit (Invitrogen). One mcrolitre cDNA from the liver and T cells or 2-4 μL of cDNA from laser dissected materials were used in each of the following RT-PCR reactions as templates. Primers located in different exons were used to generate specific products only from cDNA but not from genomic DNA (Additional file 4 : Table S2). For nested-PCR, 1 μL PCR product of the first round of PCR was used as template for the second PCR, making use of a nested primer and the common 3' or 5' primer. The cycle number was always 45 for each round of PCR. All final PCR-products were verified by sequencing.
TCR repertoire screening
In order to determine the TCR Vβ usage and Vα14 junctional diversity, nested-RT-PCR was performed. cDNA from more than 50 Irga6 cored patches collected by LMD or cDNA of 1000 lymphocytes from mesenteric and cervical lymph nodes were synthesized and nested-RT-PCR was performed as described in [ 4 ] The Vβ8 and Cβ primers [ 44 ] Vα14 and Cα primers [ 16 ], Vα2, Vα8, Vα17 primers [ 45 ] were described before (Additional file 4 : Table S2). The final products of nested-PCR were cloned into pGEM-T-easy (Promega) and sequenced. The sequences were compared to classical TCR V sequences in the database online using IMGT/V-QUEST http://imgt.cines.fr/ . Clones with a sequence identity of more than 95% to known TCR V sequences were identified as correct TCR V clones. Clones were classified into groups with same junctions. Representatives from each group were selected aligned and analysed with free software GeneDoc (Version 2.6.002; http://www.nrbsc.org ) and Vector NTI (Version 9, Invitrogen). See Additional file 5 : Table S3 for complete junction sequence data. | Results and discussion
We recently demonstrated the constitutive expression of the IFN-γ-inducible, immunity-related GTPase, Irga6, in hepatic parenchymal cells of normal mice [ 4 ]. This was attributed to the presence of a dedicated, liver-specific promoter associated with this innate immune resistance gene. During these studies we noticed that the expression of Irga6 in hepatic parenchymal cells was not uniform. Small focal patches, each consisting of a few to a few dozen contiguous cells, expressed very much more Irga6 than the general expression level (Figure 1A ). These foci of high Irga6 expression resembled those that we reported in the kidney parenchyma associated with tubular epithelium [ 4 ]. About 10% of high expression foci in the liver were characterized by a central accumulation of small mononuclear cells. We could establish immunohistologically that T cells, defined by CD3, and macrophage/DC lineage cells, defined by F4/80, were present in these mononuclear cell cores (Figure 1B ). The same cell types could be found adjacent to the patches of high Irga6 expression in the kidney cortical tubular epithelium, following closely the pattern of F4/80+ DC described in mouse kidney [ 5 ]. In subsequent analysis of the liver we distinguished between the minority of Irga6 expression patches with and the majority without, visible mononuclear cell cores (see Materials and Methods), as they clearly have different origins (see below). There are mononuclear cells associated with all kidney patches and the patches seem to have only one origin (see below). High Irga6 expression patches are not present in newborn mice but develop rapidly in both the kidney and liver between 1 week and 3 weeks after birth (Figure 1C ).
As Irga6 expression is induced by IFN-γ [ 6 , 7 ], we asked whether high expression patches were due to the local expression of IFN-γ. First, we examined the liver and kidneys of mice with genomic disruptions of components of the IFN response mechanism on the C57BL/6 background (Figure 2A Additional file 1 : Table S1. Both IFN-γ [ 8 ] and IFN-gamma receptor (IFNGR) [ 9 ] deficiencies eliminated all high Irga6 expression patches from the kidney and all Irga6 patches with mononuclear cores from the liver. The liver patches without cores were not significantly affected. IFNAR deficiency [ 10 ] had no marked effect on either organ. As expected, all kidney patches and liver patches with mononuclear cores were eliminated by STAT-1 deficiency. Surprisingly, the high expression patches without mononuclear cell cores were also absent in the STAT-1-deficient livers, even though the loss of type I and type II IFN receptors had no effect on them. We conclude that coreless liver patches are caused by the local action of another cytokine probably not of lymphoid origin, perhaps IFN-λ (IL-29) or IL-27 which both transduce signals via STAT1 in hepatocytes, but initiate signalling through distinct receptors [ 11 , 12 ]. Rag1 deficiency [ 13 ] behaved like IFN-γ or IFNGR deficiency, eliminating all Irga6 high expression foci with mononuclear cores from the liver, showing that lymphocytes with rearranged receptor chains were required for the focal expression patches with mononuclear cell cores but not for the coreless patches. It was not possible to assess high expression patches in kidney because of near universal high expression in these organs (Additional file 2 : Figure S1). No liver phenotype was detected in B-cell-deficient JHT mice [ 14 ], excluding B cells from further consideration. Again, excessive high expression was found in the kidneys (Additional file 2 : Figure S1).
Secondly, if IFN-γ is responsible for the focal induction of Irga6, there should be activation in the high expression patches of the IFN-γ-dependent promoter of the Irga6 gene, with the use of exon 1A, in contrast to the general hepatic parenchyma in which the liver-specific promoter is constitutively expressed with usage of exon 1B [ 4 ]. We therefore isolated high expression patches and areas of general expression level by laser microdissection and used quantitative real-time polymerase chain reaction (qRT-PCR) in order to identify the use of the IFN-γ-dependent 5'-untranslated 1A exon and the constitutively expressed 5'-untranslated 1B exon in the two tissue sites (Figure 2C ). Expression of the IFN-γ-dependent exon 1A was 27.5× higher by qRT-PCR in the high expression patches than in the general liver, while 1B exon expression was not altered (1.5×). However, when the Irga6-1A PCR products from the general liver were cloned and sequenced, only 5.9% of were specific (a consequence of the very low absolute expression of the IFN-γ-inducible transcript in non-focal hepatic parenchyma), while 100% of those cloned from the high expression patches were specific. Thus, the true excess of Irga6-1A in the high expression patches relative to the general parenchyma was 100 × 27.9/5.9 or about 470×, confirming the intense local induction by IFN-γ.
Thirdly, if IFN-γ is being secreted locally, affected cells should express not only Irga6 but also other members of the IFN-γ-inducible IRG protein family that are not constitutively expressed in hepatic parenchymal cells [ 4 ]. We therefore stained adjacent histological sections of liver and kidney for Irga6 and for Irgm3. Indeed, focal expression of Irgm3 coincided accurately with the high Irga6 expression patches (Additional file 3 : Figure S2).
Direct evidence for the expression of IFN-γ in the patches with mononuclear cell cores, but not in the coreless patches, was obtained by RT-PCR for IFN-γ transcripts in microdissected patch material. A strong signal for IFN-γ was recovered from only the patches with mononuclear cell cores (Figure 2D ). These experiments established beyond doubt that the liver and kidney contain reactive foci of local, T-cell-dependent IFN-γ production identified by the local high expression of Irga6.
The character of the small IFN-γ-secreting reactive foci in the liver and kidney suggested that they could be due to local immune activity stimulated by microbial material. We therefore analysed the liver and kidney for reactive foci from mice deficient in components of the Toll-le receptor (TLR) system (Figure 2B ). No effects were seen in mice deficient in TLR2, 4 or 9 or MyD88, arguing against bacterial components as the initiators of the foci. In order to test this conclusion further, we examined the liver and kidney from germ-free mice from three independent sources. No significant reduction in numbers of reactive foci in liver or spleen was detected (Figure 2B ). Bacterial infection is, thus, unlikely to be a stimulus for the reactive foci
Two further explanations were considered for the focal IFN-γ production. One was the local re-expression of an endogenous mouse mammary tumour virus (MMTV) open reading frame acting as a superantigen (reviewed in [ 15 ]). The other was local activation of iNKT cells. These two possibilities could be distinguished by a knowledge of the V gene usage of T cells in the mononuclear cores. In the case of a superantigen, specific Vβ families should be favoured, but no specific Vα usage, while in iNKT cells, a specific and diagnostic pattern of Vβ8/Vα14Jα18 could be anticipated [ 16 ]. We used laser microdissection to isolate and pool the mononuclear cell cores from many high Irga6 expression patches and employed semi-nested PCR on reverse-transcribed cDNA to amplify, and subsequently clone, Vα and Vβ from expressed TCRs. The result for Vβ usage was striking, namely, the exclusive usage of the three Vβ8 genes, with an excess of Vβ8.2 (Figure 3A and 3C ) in contrast to the diversity of Vβ genes recovered by the same technique from lymph node lymphocytes. We were able to confirm the high use of Vβ8 in mononuclear core T cells immunohistologically with a specific anti-Vβ8 monoclonal antibody (Figure 3B ). Vβ8 is uncommon, but not unknown, as a MMTV superantigen target [ 15 ] but is part of the diagnostic receptor combination of iNKT cells [ 16 ]. We therefore amplified by RT-PCR a number of specific Vα families from lymph node cells and from mononuclear cell cores (Figure 3D ). In lymph node T cells a weak signal for Vα14 was seen, compared with strong signals for Vα2 and Vα8. From the mononuclear core cells a strong signal was seen for Vα14, an unclear result for Vα2 and a weaker signal for Vα8. However, the true excess of Vα14 in the core cells was striking since the weak signal of Vα14 from lymph nodes proved, on cloning and sequencing, to be due, with a single exception, to the amplification of Vα11 which is closely related to Vα14. In contrast, all except one of the clones amplified for Vα14 from mononuclear cell cores of liver patches were indeed Vα14 (Figure 3D ). Thus, the TCR V-gene usage of the mononuclear cell cores was consistent with the TCRs of iNKT cells. This was further confirmed by showing that exclusively the canonical Jα18 was used with no length variation (Figure 3E ).
If the NKT function is dependent on a dedicated Vβ8/Vα14 TCR and is responsible for the high expression patches in liver and kidney, these should be reduced in mice carrying deletions in the TCR Vβ gene array that include the Vβ8 family of sequences [ 17 , 18 ]. We analysed the numbers of liver core patches and kidney high expression patches in inbred mouse strains of known TCR Vβ genotype, in each case comparing patch frequencies with C57BL/6. Tissues from the A/J and C57BRcdJ strains, with normal Vβ8 gene representation, were similar to C57BL/6 (Figure 4A ). However, five strains with Vβ8 deletions (SWR/J, SJL/J, RIIIS/J, C57BR/J and C57L/J) all showed reduction down to a more or less complete loss of cored liver patches and kidney patches. The residual kidney and cored liver patches in these strains may, perhaps, be attributed to usage of Vβ2 and Vβ7, V-genes that also contribute to a minority of iNKT receptors [ 19 ], although we did not find these subgroups in our analysis in C57BL/6 (Figure 3A ). SWR mice had an unexplained deficit of plain liver patches.
If the Irga6 high expression foci with mononuclear cores associated with high IFN-γ production were, indeed, due to iNKT cells, they should also be absent from mice deficient in the essential ligand-presenting molecule CD1 d [ 20 , 21 ] and in mice deficient in iNKT cells as a result of deletion of the diagnostic Jα18 T cell receptor segment [ 22 ]. Both these predictions were fulfilled, with a substantial loss of high Irga6 expression patches from the kidney and cored patches from the liver in both mutants (Figure 4B ). The patches of high Irga6 expression in the liver without mononuclear cores were not affected in either case. In order to find out whether CD4 + or CD8 + T cells play any additional role in the induction of high Irga6 expression patches, organs from mice lacking MHC Aβ b (CD4 + T cell deficient [ 23 ]) and TAP1 (CD8 + T cell deficient [ 24 ]) were also examined. Patches of high Irga6 expression were normal in liver and kidney from both strains. Thus, iNKT cells alone appear to be responsible for the high Irga6 expression foci.
We present direct evidence for the constitutive activation of mouse iNKT cells in two organs, the liver and the kidneys, generating small foci of high IFN-γ expression detected by the local high expression of an IFN-γ-inducible cytoplasmic protein, Irga6. In a previous study, we surveyed the pattern of constitutive Irga6 expression in a number of other organs and high expression foci were not seen [ 4 ]. The liver and kidneys thus appear to be special in this respect. Our results suggest a local activation event that triggers production of type II interferon by iNKT cells. The mouse liver sinusoids are known to contain a major population of NKT cells and NKT cells have been reported in the kidney, probably corresponding to the T cells identified in the tubulo-interstitial spaces [ 25 ]. Dendritic cells (DC) and hepatic stellate cells (Ito cells) [ 26 ] can both present exogenous α-GalCer via CD1 d to iNKT cells. Stellate cells of the liver sinusoids have also been shown to be IL-15-dependent activators of iNKT cells to proliferation in the absence of exogenous antigen [ 26 ]. An apparently similar cell type is also found in the tubular cortex of the kidney essentially co-localizing with dendritic cells [ 5 , 27 ]. Both DC and stellate cells are strong candidates for the activating cell implied by our experiments, but this does not, by itself, account for the focal triggering event. Our experiments with germ-free and TLR-deficient mice make bacterial products acting on DC TLRs [ 28 , 29 ] an unlikely cause. More plausible would, perhaps, be a locally reactivating endogenous retrovirus triggering DC or stellate cells and, thereby, iNKT cells locally. Arguing against a primary viral stimulus, we could show no deficit in liver cored high expression patches or kidney patches in organs from mice deficient in MAVS/CARDIF [ 30 ], the common activation pathway of the cytosolic RIG helicases that act as viral sensors (Figure 4C ).
It will be interesting to know how the spontaneous activation of iNKT cells and consequent constitutive IFN-γ production in liver and kidney of major inbred strains of mouse affects immunological function in these organs. NKT cells have been identified very recently as a minority population among normal resident lymphocytes in mouse kidney and both an excitatory role in ischaemia-reperfusion injury [ 31 ] and an inhibitory role in experimental glomerular nephritis induction by anti glomerular basement membrane antibody [ 32 , 33 ] have been described. Likewise, the large population of NKT cells that patrol the liver sinusoids [ 34 ] have been implicated in various liver disorders [ 35 ], but their spontaneous, focal activation has not been reported. Spontaneous focal activation of NKT cells in the liver and kidney with high local levels of IFN-γ will have to be considered in future studies of immune function in these organs. | Results and discussion
We recently demonstrated the constitutive expression of the IFN-γ-inducible, immunity-related GTPase, Irga6, in hepatic parenchymal cells of normal mice [ 4 ]. This was attributed to the presence of a dedicated, liver-specific promoter associated with this innate immune resistance gene. During these studies we noticed that the expression of Irga6 in hepatic parenchymal cells was not uniform. Small focal patches, each consisting of a few to a few dozen contiguous cells, expressed very much more Irga6 than the general expression level (Figure 1A ). These foci of high Irga6 expression resembled those that we reported in the kidney parenchyma associated with tubular epithelium [ 4 ]. About 10% of high expression foci in the liver were characterized by a central accumulation of small mononuclear cells. We could establish immunohistologically that T cells, defined by CD3, and macrophage/DC lineage cells, defined by F4/80, were present in these mononuclear cell cores (Figure 1B ). The same cell types could be found adjacent to the patches of high Irga6 expression in the kidney cortical tubular epithelium, following closely the pattern of F4/80+ DC described in mouse kidney [ 5 ]. In subsequent analysis of the liver we distinguished between the minority of Irga6 expression patches with and the majority without, visible mononuclear cell cores (see Materials and Methods), as they clearly have different origins (see below). There are mononuclear cells associated with all kidney patches and the patches seem to have only one origin (see below). High Irga6 expression patches are not present in newborn mice but develop rapidly in both the kidney and liver between 1 week and 3 weeks after birth (Figure 1C ).
As Irga6 expression is induced by IFN-γ [ 6 , 7 ], we asked whether high expression patches were due to the local expression of IFN-γ. First, we examined the liver and kidneys of mice with genomic disruptions of components of the IFN response mechanism on the C57BL/6 background (Figure 2A Additional file 1 : Table S1. Both IFN-γ [ 8 ] and IFN-gamma receptor (IFNGR) [ 9 ] deficiencies eliminated all high Irga6 expression patches from the kidney and all Irga6 patches with mononuclear cores from the liver. The liver patches without cores were not significantly affected. IFNAR deficiency [ 10 ] had no marked effect on either organ. As expected, all kidney patches and liver patches with mononuclear cores were eliminated by STAT-1 deficiency. Surprisingly, the high expression patches without mononuclear cell cores were also absent in the STAT-1-deficient livers, even though the loss of type I and type II IFN receptors had no effect on them. We conclude that coreless liver patches are caused by the local action of another cytokine probably not of lymphoid origin, perhaps IFN-λ (IL-29) or IL-27 which both transduce signals via STAT1 in hepatocytes, but initiate signalling through distinct receptors [ 11 , 12 ]. Rag1 deficiency [ 13 ] behaved like IFN-γ or IFNGR deficiency, eliminating all Irga6 high expression foci with mononuclear cores from the liver, showing that lymphocytes with rearranged receptor chains were required for the focal expression patches with mononuclear cell cores but not for the coreless patches. It was not possible to assess high expression patches in kidney because of near universal high expression in these organs (Additional file 2 : Figure S1). No liver phenotype was detected in B-cell-deficient JHT mice [ 14 ], excluding B cells from further consideration. Again, excessive high expression was found in the kidneys (Additional file 2 : Figure S1).
Secondly, if IFN-γ is responsible for the focal induction of Irga6, there should be activation in the high expression patches of the IFN-γ-dependent promoter of the Irga6 gene, with the use of exon 1A, in contrast to the general hepatic parenchyma in which the liver-specific promoter is constitutively expressed with usage of exon 1B [ 4 ]. We therefore isolated high expression patches and areas of general expression level by laser microdissection and used quantitative real-time polymerase chain reaction (qRT-PCR) in order to identify the use of the IFN-γ-dependent 5'-untranslated 1A exon and the constitutively expressed 5'-untranslated 1B exon in the two tissue sites (Figure 2C ). Expression of the IFN-γ-dependent exon 1A was 27.5× higher by qRT-PCR in the high expression patches than in the general liver, while 1B exon expression was not altered (1.5×). However, when the Irga6-1A PCR products from the general liver were cloned and sequenced, only 5.9% of were specific (a consequence of the very low absolute expression of the IFN-γ-inducible transcript in non-focal hepatic parenchyma), while 100% of those cloned from the high expression patches were specific. Thus, the true excess of Irga6-1A in the high expression patches relative to the general parenchyma was 100 × 27.9/5.9 or about 470×, confirming the intense local induction by IFN-γ.
Thirdly, if IFN-γ is being secreted locally, affected cells should express not only Irga6 but also other members of the IFN-γ-inducible IRG protein family that are not constitutively expressed in hepatic parenchymal cells [ 4 ]. We therefore stained adjacent histological sections of liver and kidney for Irga6 and for Irgm3. Indeed, focal expression of Irgm3 coincided accurately with the high Irga6 expression patches (Additional file 3 : Figure S2).
Direct evidence for the expression of IFN-γ in the patches with mononuclear cell cores, but not in the coreless patches, was obtained by RT-PCR for IFN-γ transcripts in microdissected patch material. A strong signal for IFN-γ was recovered from only the patches with mononuclear cell cores (Figure 2D ). These experiments established beyond doubt that the liver and kidney contain reactive foci of local, T-cell-dependent IFN-γ production identified by the local high expression of Irga6.
The character of the small IFN-γ-secreting reactive foci in the liver and kidney suggested that they could be due to local immune activity stimulated by microbial material. We therefore analysed the liver and kidney for reactive foci from mice deficient in components of the Toll-le receptor (TLR) system (Figure 2B ). No effects were seen in mice deficient in TLR2, 4 or 9 or MyD88, arguing against bacterial components as the initiators of the foci. In order to test this conclusion further, we examined the liver and kidney from germ-free mice from three independent sources. No significant reduction in numbers of reactive foci in liver or spleen was detected (Figure 2B ). Bacterial infection is, thus, unlikely to be a stimulus for the reactive foci
Two further explanations were considered for the focal IFN-γ production. One was the local re-expression of an endogenous mouse mammary tumour virus (MMTV) open reading frame acting as a superantigen (reviewed in [ 15 ]). The other was local activation of iNKT cells. These two possibilities could be distinguished by a knowledge of the V gene usage of T cells in the mononuclear cores. In the case of a superantigen, specific Vβ families should be favoured, but no specific Vα usage, while in iNKT cells, a specific and diagnostic pattern of Vβ8/Vα14Jα18 could be anticipated [ 16 ]. We used laser microdissection to isolate and pool the mononuclear cell cores from many high Irga6 expression patches and employed semi-nested PCR on reverse-transcribed cDNA to amplify, and subsequently clone, Vα and Vβ from expressed TCRs. The result for Vβ usage was striking, namely, the exclusive usage of the three Vβ8 genes, with an excess of Vβ8.2 (Figure 3A and 3C ) in contrast to the diversity of Vβ genes recovered by the same technique from lymph node lymphocytes. We were able to confirm the high use of Vβ8 in mononuclear core T cells immunohistologically with a specific anti-Vβ8 monoclonal antibody (Figure 3B ). Vβ8 is uncommon, but not unknown, as a MMTV superantigen target [ 15 ] but is part of the diagnostic receptor combination of iNKT cells [ 16 ]. We therefore amplified by RT-PCR a number of specific Vα families from lymph node cells and from mononuclear cell cores (Figure 3D ). In lymph node T cells a weak signal for Vα14 was seen, compared with strong signals for Vα2 and Vα8. From the mononuclear core cells a strong signal was seen for Vα14, an unclear result for Vα2 and a weaker signal for Vα8. However, the true excess of Vα14 in the core cells was striking since the weak signal of Vα14 from lymph nodes proved, on cloning and sequencing, to be due, with a single exception, to the amplification of Vα11 which is closely related to Vα14. In contrast, all except one of the clones amplified for Vα14 from mononuclear cell cores of liver patches were indeed Vα14 (Figure 3D ). Thus, the TCR V-gene usage of the mononuclear cell cores was consistent with the TCRs of iNKT cells. This was further confirmed by showing that exclusively the canonical Jα18 was used with no length variation (Figure 3E ).
If the NKT function is dependent on a dedicated Vβ8/Vα14 TCR and is responsible for the high expression patches in liver and kidney, these should be reduced in mice carrying deletions in the TCR Vβ gene array that include the Vβ8 family of sequences [ 17 , 18 ]. We analysed the numbers of liver core patches and kidney high expression patches in inbred mouse strains of known TCR Vβ genotype, in each case comparing patch frequencies with C57BL/6. Tissues from the A/J and C57BRcdJ strains, with normal Vβ8 gene representation, were similar to C57BL/6 (Figure 4A ). However, five strains with Vβ8 deletions (SWR/J, SJL/J, RIIIS/J, C57BR/J and C57L/J) all showed reduction down to a more or less complete loss of cored liver patches and kidney patches. The residual kidney and cored liver patches in these strains may, perhaps, be attributed to usage of Vβ2 and Vβ7, V-genes that also contribute to a minority of iNKT receptors [ 19 ], although we did not find these subgroups in our analysis in C57BL/6 (Figure 3A ). SWR mice had an unexplained deficit of plain liver patches.
If the Irga6 high expression foci with mononuclear cores associated with high IFN-γ production were, indeed, due to iNKT cells, they should also be absent from mice deficient in the essential ligand-presenting molecule CD1 d [ 20 , 21 ] and in mice deficient in iNKT cells as a result of deletion of the diagnostic Jα18 T cell receptor segment [ 22 ]. Both these predictions were fulfilled, with a substantial loss of high Irga6 expression patches from the kidney and cored patches from the liver in both mutants (Figure 4B ). The patches of high Irga6 expression in the liver without mononuclear cores were not affected in either case. In order to find out whether CD4 + or CD8 + T cells play any additional role in the induction of high Irga6 expression patches, organs from mice lacking MHC Aβ b (CD4 + T cell deficient [ 23 ]) and TAP1 (CD8 + T cell deficient [ 24 ]) were also examined. Patches of high Irga6 expression were normal in liver and kidney from both strains. Thus, iNKT cells alone appear to be responsible for the high Irga6 expression foci.
We present direct evidence for the constitutive activation of mouse iNKT cells in two organs, the liver and the kidneys, generating small foci of high IFN-γ expression detected by the local high expression of an IFN-γ-inducible cytoplasmic protein, Irga6. In a previous study, we surveyed the pattern of constitutive Irga6 expression in a number of other organs and high expression foci were not seen [ 4 ]. The liver and kidneys thus appear to be special in this respect. Our results suggest a local activation event that triggers production of type II interferon by iNKT cells. The mouse liver sinusoids are known to contain a major population of NKT cells and NKT cells have been reported in the kidney, probably corresponding to the T cells identified in the tubulo-interstitial spaces [ 25 ]. Dendritic cells (DC) and hepatic stellate cells (Ito cells) [ 26 ] can both present exogenous α-GalCer via CD1 d to iNKT cells. Stellate cells of the liver sinusoids have also been shown to be IL-15-dependent activators of iNKT cells to proliferation in the absence of exogenous antigen [ 26 ]. An apparently similar cell type is also found in the tubular cortex of the kidney essentially co-localizing with dendritic cells [ 5 , 27 ]. Both DC and stellate cells are strong candidates for the activating cell implied by our experiments, but this does not, by itself, account for the focal triggering event. Our experiments with germ-free and TLR-deficient mice make bacterial products acting on DC TLRs [ 28 , 29 ] an unlikely cause. More plausible would, perhaps, be a locally reactivating endogenous retrovirus triggering DC or stellate cells and, thereby, iNKT cells locally. Arguing against a primary viral stimulus, we could show no deficit in liver cored high expression patches or kidney patches in organs from mice deficient in MAVS/CARDIF [ 30 ], the common activation pathway of the cytosolic RIG helicases that act as viral sensors (Figure 4C ).
It will be interesting to know how the spontaneous activation of iNKT cells and consequent constitutive IFN-γ production in liver and kidney of major inbred strains of mouse affects immunological function in these organs. NKT cells have been identified very recently as a minority population among normal resident lymphocytes in mouse kidney and both an excitatory role in ischaemia-reperfusion injury [ 31 ] and an inhibitory role in experimental glomerular nephritis induction by anti glomerular basement membrane antibody [ 32 , 33 ] have been described. Likewise, the large population of NKT cells that patrol the liver sinusoids [ 34 ] have been implicated in various liver disorders [ 35 ], but their spontaneous, focal activation has not been reported. Spontaneous focal activation of NKT cells in the liver and kidney with high local levels of IFN-γ will have to be considered in future studies of immune function in these organs. | Conclusions
Invariant natural killer T (iNKT) cells are activated spontaneously in mouse liver and kidney. These activation events are local, probably extravascular, and result in the local secretion of IFN-γ. This in turn results in the local induction of IFN-γ-responsive genes in non-immune tissue cells, namely hepatic parenchymal cells and renal tubular epithelial cells. These activation events are spontaneous in the sense that they appear to be independent of exogenous pathogenic material of bacterial or viral origin. The cause of individual activation events is still unclear. In general, however, the results are consistent with a widely held view that iNKT cells are rendered spontaneously active by interaction with an endogenous ligand. In this study we demonstrate these activation events histologically. Further investigations may reveal the nature of the stimulatory event that lies at the heart of each activating focus. It will be important to generalise these findings to the human case, and to assess the implications of these findings for liver and kidney immunopathology. | Background
Invariant natural killer T (iNKT) cells differ from other T cells by their hyperactive effector T-cell status, in addition to the expression of NK lineage receptors and semi-invariant T-cell receptors. It is generally agreed that the immune phenotype of iNKT cells is maintained by repeated activation in peripheral tissues although no explicit evidence for such iNKT cell activity in vivo has so far been reported.
Results
We used an interferon (IFN)-γ-inducible cytoplasmic protein, Irga6, as a histological marker for local IFN-γ production. Irga6 was intensely expressed in small foci of liver parenchymal cells and kidney tubular epithelium. Focal Irga6 expression was unaffected by germ-free status or loss of TLR signalling and was totally dependent on IFN-γ secreted by T cells in the centres of expression foci. These were shown to be iNKT cells by diagnostic T cell receptor usage and their activity was lost in both CD1 d and Jα-deficient mice.
Conclusions
This is the first report that supplies direct evidence for explicit activation events of NKT cells in vivo and raises issues about the triggering mechanism and consequences for immune functions in liver and kidney. | Abbreviations
α-GalCer: alpha-galactosyl ceramide; F: foci; NF: non-focus region; iNKT: invariant natural killer-like T cell; FN: interferon; IFNAR: IFN type I receptor; IFNGR: IFN type II receptor; IL: interleukin; HRP: horseradish peroxidase; IRG: immunity-related GTPases; MHC: major histocompatibility; MMTV: mouse mammary tumour virus; PBS: phosphate buffered saline; PCR: polymerase chain reaction; qRT-PCR: quantitative reverse transcription PCR; TCR: T-cell receptor; TLR: Toll-like receptor.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
JZ and JCH identified the problem and conceived the experimental approach. JZ performed all the experimental work, counted the histological data and prepared the figures. JCH drafted the manuscript, which was completed collaboratively by JZ and JCH. Both authors read and approved the final manuscript.
Supplementary Material | Acknowledgements
We are very much indebted to Dr Margarete Odenthal, Institute of Pathology, University of Cologne, Cologne, Germany, for allowing us to make extensive use of her LMD facility. This work would not have been possible without the generosity of many scientists who provided knock-out and wild-type mice (itemized in Materials and Methods) and in some cases also enabled us to harvest tissues in their facilities. For this our thanks are due to Marina Freudenberg (Freiburg, Germany), Maria Leite de Moraes (Paris, France), Luc Van Kaer (Vanderbilt, TN, USA), Jocelyne Demengeot (Oeiras, Portugal), Jurg Tschopp (Lausanne, Switzerland), Rudolf Jörg (Zürich, Switzerland), Thomas Kolbe (Vienna, Austria), Heike Weighardt (Munich, Germany) and Ari Waisman (Mainz, Germany). We are grateful to two anonymous referees for suggesting that we identify iNKT cells more explicitly by examining tissues from mice deficient in Jα18, TAP and class II MHC. We owe special thanks to Pia Scholl, Institute for Genetics, for her devoted attention recently to the immunohistology of these as well as of the MAVS/CARDIF deficie.
This work was supported in part by the University of Cologne, Cologne, Germany, and in part by grants in the Schwerpunktprogramm 1110 and the collaborative research programme, SFB670 'Cell-autonomous Immunity', both from the Deutsche Forschungsgemeinschaft. Jia Zeng was supported in part from the International Graduate School in Genetics and Functional Genomics. | CC BY | no | 2022-01-12 15:21:44 | BMC Biol. 2010 Nov 30; 8:142 | oa_package/17/5b/PMC3016249.tar.gz |
PMC3016250 | 21176227 | Introduction
Cyclins are proteins which act as key controlling elements of the eukaryotic cell cycle. These proteins have some regions of homology such as the cyclin box and some other islands of homology outside the cyclin box [ 1 ]. In mammalian cells, cyclins bind to cyclin dependent kinases and form complexes that are involved in regulating different cell cycle transitions: cyclin-D-CDK4/6 complex for G1 progression, cyclin- E - CDK2 for the G1-S transition, cyclin-A-CDK2 for S phase progression and cyclin A/B-CDC2 for entry into M-phase. In addition to these functions, cyclins are also involved in some processes not directly related to the cell cycle. The importance of cyclin-CDK complexes in cell proliferation is underscored by the fact that deregulation in the function of these complexes is found in virtually the whole spectrum of human tumors and this comes from the fact that tumor-associated alterations in cyclins help to sustain proliferation independently of external mitogenic or anti-mitogenic signals [ 2 ]. In this review we are going to deal with the role of cyclins D and E in the development of cancer, since these cyclins have proved to be of great importance for cancer pathogenesis. | Conclusions
In conclusion, cyclins play a multifunctional and pivotal role in the pathogenesis of cancer. This is the reason why alterations in their structure and function through the influence of various pathways can lead to an array of cancer types. This discovery in combination with recent studies in genetically engineered mouse models implies their potential role in cancer therapy and especially targeted therapies. Despite the clinical applications of cell cycle specific chemotherapeutic agents there is still urgent need to develop novel drugs that are able to target multiple sites and pathways of the cell cycle [ 55 ]. | Cyclins are indispensable elements of the cell cycle and derangement of their function can lead to cancer formation. Recent studies have also revealed more mechanisms through which cyclins can express their oncogenic potential. This review focuses on the aberrant expression of G1/S cyclins and especially cyclin D and cyclin E; the pathways through which they lead to tumour formation and their involvement in different types of cancer. These elements indicate the mechanisms that could act as targets for cancer therapy. | Cyclins and cell cycle
Considerable effort over many years has been expended in order to understand the mechanisms that control normal cell cycles. This effort has resulted in a detailed - but not yet completed - picture of the cell cycle revealing that complex oscillations in the activation and inactivation of cyclin- dependent kinase complexes propel mammalian cells through the cycle. The levels of most CDKs are relatively constant during the cell cycle but their activities depend highly on the state and level of activation of their cyclin partners or other regulatory molecules [ 3 ].
The triggering factor for progression to S phase is a mitogenic signal. In response to mitogenic activation, cells synthesize D-type cyclins which form a holoenzyme with CDK4, CDK6. Cyclin D1 is the regulatory subunit whereas the CDKs are the catalytic subunit (figure 1 ). This assembly of proteins needs members of the Cip/Kip families of proteins which promote the activity of cyclin D dependent kinases and serve as inhibitors of CDK2. [ 4 ]. The active complex phosphorylates the pRB protein and leads to its inactivation. The inactivated pRB protein seperates from the complex of pRB and E2F transcription factors giving permission to genes required for S phase to be transcripted [ 3 ]. Cyclin E, cyclin A and DNA pol stand among these genes. Cyclin E binds to CDK2 leading to phosphorylation of substrates required for proper replication firing, centrosome duplication and histone biosynthesis [ 5 ]. Cyclin E and its partner, CDK2, can also further phosphorylate and inactivate pRB. Cyclin A binds to CDK2 and this complex phosphorylates CDC6 resulting in its relocalisation from the nucleus to the cytoplasm and in this way to its destruction. This procedure prevents CDC6 from assembling into origins of replication of DNA after G1. DNA re- replication is also avoided by the procedure where cyclin A -CDK2 phosphorylates MCM4 in the helicase complex and eventually inhibits its DNA helicase activity [ 6 ].
To summarize, such complicated, multilevel controls on expression and activation of cyclin/CDK complexes permit exquisite and necessary coordination of the cell cycle stages and thereby prevent from the formation of tumor cells [ 2 ].
Cyclin D and cancer
Cyclin D is solidly established as an oncogene with an important pathogenetic role in many human tumors. There are three highly homologous and almost indistinguishable biochemically D- type cyclins (D1, D2 and D3) in mammalian cells which are binded to either CDK4 or CDK6 in a tissue specific way. Among these types, cyclin D1 is the one most commonly expressed in several human cancers [ 6 ]. Cyclin D1 is a 35-kDa protein which is encoded by 5 exons situated at the region of chromosome band 11q13. In the aminoterminus of cyclin D1 appears a motif Leu - X - Cys - X - Glu (X represents any aminoacid) where pRB pocket domain binds. The carboxy terminus inhibits myogenic helix loop helix (HLH) protein function. HLH protein main action is to remove cells from the cell cell cycle (halt proliferation), so its inhibition by cyclin D1 leads the cell to G1 stage of the cell cycle. Repression by D cyclins appears to be independent of its effects on the cell cycle [ 7 ]. The protein is quite unstable with a half - life of less than 20 minutes; its degradation is ubiquitin proteosome- regulated [ 8 ].
Cyclin D1 is overexpressed in several human tumours. Chromosomal translocations, gene amplification and disruption of normal intercellular trafficking and proteolysis are the procedures which lead to accumulation of cyclin D1 in tumor cell nuclei and eventually to cyclin D1 overexpression in many tumours.
Chromosomal translocations are very common among parathyroid adenomas, B mantle cell lymphomas and multiple myelomas. Gene amplification (11q13) as a mechanism for aberrant overexpression of cyclin D1 is associated with non- small cell lung cancers, head and neck squamous cell carcinomas, pancreatic carcinomas, bladder cancer, pituitary adenomas and breast carcinoma. Emerging evidence suggests that nuclear retention of cyclin D1 resulting from altered nuclear trafficking and proteolysis is critical for the manifestation of its oncogenicity [ 9 ]. Disruption of the normal intracellular trafficking and proteolysis of the nuclear non - phosphorylatable cyclin results from a polymorphism in exon four of cyclin D1. This leads to a C-terminus that lacks the phosphor- acceptor site that targets cyclin D1 for cytoplasmic destruction [ 10 ].
Cyclin D oncogenic potential is manifested in several ways. As mentioned before, it leads to direct activation of CDK4/6. In addition to this function, cyclin D1/CDK complexes bind and sequester p21CIP1 and p27KIP1, which among others function as inhibitors of cyclin E/CDK2. [ 11 ]. In this way, both high expression of cyclin D1 and deregulated expression of cyclin E1 cooperate to increase tumour fitness. Another cyclin D1 function that can lead to tumour formation is the transcriptional control that does not involve CDKs. This function involves promoter recruitment of histone deacetylases (HDACs) and histone methyltransferases. Normally HDAC, by increasing the positive charge of histone tails and histone methylotransferases, through the methylation of histones, can both lead to high- affinity binding between histones and DNA backbone. In this way, DNA structure condenses and transcription is prevented [ 12 ].
Several groups have demonstrated that cyclin D1 can also act as a transcriptional co-factor for steroid hormone receptors such as estrogen receptor [ 13 ].
Besides tumour formation, cyclin D1 can also play a pivotal role in the invasiveness and the metastatic phenotype through the interactions between the malignant cell and the host environment. For example, overexpression of cyclin D1 through the activation of positive feedback loop of E2F-1 mediated transcription can lead to excessive expression of FGFR-1 (fibroblast growth factor receptor 1) [ 14 ]. FGFR up - regulation has been shown in several tumours such as brain, breast, prostate, thyroid, skin and salivary gland tumours. Additionally, cyclin D1 normally plays a regulatory role in angiogenesis and mithochondrial function. This suggests that deregulated cyclin D1 expression can contribute to the invasive and metastatic potential of a tumour, since mtDNA mutations can lead to development of metastases by overproduction of reactive oxygen species (ROS) [ 15 , 16 ].
The biological importance of these functions needs to be proved in vivo ; nevertheless it is obvious in concept that they could be of variable impact on tumour phenotype. Nevertheless, solitary cyclin overexpression is not sufficient for malignancy transformation. Additional cellular abnormalities are necessary for the tumour formation [ 17 ].
Table 1 describes the way that cyclin D is associated with several types of cancer.
Parathyroid adenomas are a common disease where cyclin D1 is overexpressed. The pericentromeric inversion of chromosome 11 places the 5' regulatory region of the PTH gene on 11p15 immediately upstream of cyclin D1 gene promoter. Many studies have taken place and they have demonstrated a cyclin D1 overexpression which varies between 20 - 40% [ 18 ]. Nevertheless, overexpression of cyclin D1 is also found in nonneoplastic proliferation of parathyroid gland, but not in the normal parathyroid tissue. The hormonal regulatory defect in parathyroid adenomas can be both primary and secondary to a defect in cellular - growth control indicated by cyclin D1 oncogene overexpression [ 19 ].
Papillary thyroid carcinoma is another malignant tumour where cyclin D1 is overexpressed. In addition to this, the level of cyclin D1 expression according to lymph node metastasis was statistically significant (P < 0.05). This fact indicates that cyclin D1 may be a useful marker for the evaluation of lymph node metastasis
In addition to solid tumours, overexpression of cyclin D1 has also been reported in certain lymphoid malignancies. Referring to B- cell non Hodgkin lymphomas, cyclin D1 was mainly overexpressed in mantle cell lymphomas and large B- cell lymphomas whereas the other subtypes showed normal cyclin D1 expression. Clinical signs (except for lymphadenopathy) and laboratory data (except for LDH) were not influenced by cyclin D1 overexpression which, nevertheless, proved to be associated with poor outcome of NHL patients [ 20 ]. More specifically, mantle cell lymphoma, which accounts for 5 - 10% of all non Hodgkin lymphomas, demonstrate chromosome translocations (t(11;14)) involving the immunoglobulin heavy chain IgH locus that lead to cyclin D1 deregulation [ 21 ]. Mantle cell lymphomas express variable levels of cyclin D1 at both transcript and protein levels. Overexpression of cyclin D2 and D3 has also been described. As far as mRNA forms produced by cyclin D1 gene (long - D1L, short - D1S) are concerned, the short version has been shown to be more related to blastoid histology than the long version (60% of D1 S and 9% of D1L) [ 22 ]. In addition to these notifications, cyclin D1 was also identified as a potentially important antigen for immunotherapy of mantle cell lymphoma as it was proved to be recognized by potent cytotoxic T cells when it was naturally presented by lymphoma cells in the context of HLA - A * 0201 molecules [ 23 ]. Another subtype of non Hodgkin lymphomas, known as diffuse large B - cell lymphoma, shows overexpression of D1 (2%) D2 (49%) and D3 (20%) cyclins [ 24 ]. A small subset of chronic lymphocytic leukemias overexpresses cyclin D1 in amounts that can be demonstrated by immunohistochemistry [ 25 ].
Cyclin D1 is solidly established as an oncogene with a pivotal role in pathogenesis of breast cancer. Besides gene amplification, cytoplasmic sequestration may also serve to regulate cyclin D1 activity in mammalian cancer cells [ 26 ]. Emerging evidence indicates that cyclin D1 may act, in part, through pathways which do not involve its role as a cell cycle regulator. One such function is the cyclin D1 contribution to cell adhesion and motility. So, it was demonstrated that cyclin D1/CDK4 complex interacts with filamin A (member of the actin - binding filamin protein family) and influences the migration and invasion potential of breast cancer cells [ 27 ]. CCDN1 amplification is found in 5 - 20% of primary breast cancers [ 28 ].
Cyclin E and cancer
Human cyclin E cDNA was identified in 1991 by screening human cDNA libraries for genes that could complement G1 cyclin mutations in yeast S. cerevisiae [ 29 ]. Cyclin E is derived from a gene on chromosome 19q12 → q13. This gene encodes a variety of polypeptides with molecular weights ranging from 39 to 52 kDa. The "regular" form contains the "cyclin box", a sequence set in amino acid position 129-215, which is partly common among the cyclins. In addition to the regular form, two splice variants and an isoform with 15 additional aminoacids at the N-terminus have also been described [ 30 ]. More recently six splice variants with the potential to produce cyclin E isoforms of substantially altered molecular weight have been found [ 31 ]. All splice variants with an intact cyclin box have the ability of binding and activation of CDK2 [ 32 ].
The deregulated expression and activity of cyclin E have been associated with a variety of cancers and it is considered to be involved in the oncogenic process [ 33 ]. The oncogenic activity of cyclin E is a result of several mechanisms. Cell cycle deregulation of cyclin E expression is common in some tumour cells leading to constitutive cyclin E expression and activity throughout the cell cycle. Overexpression of cyclin E can come from gene amplification in most cases [ 34 ]. For example cyclin E gene is amplified by 8 fold and its mRNA is overexpressed by 64 fold in a subset of breast cancer cell lines [ 35 , 36 ]. Defected degradation via the proteosome is another mechanism leading to cyclin E overexpression; the F- box proteins that target cyclin E for ubiquitination and as a result for degradation were discovered to be mutated in some cancers [ 37 ]. Cyclin E overexpression can lead to G1 shortening, decrease in cell size or loss of serum requirement for proliferation. This is the consequence of cyclin E normal function of S - phase induction; one of the pathways involving pRB has already been described; in addition cyclin E/CDK2 complexes have been proved to activate transcriptional regulators like human B - MYB and NPAT which are of great importance for cellular proliferation [ 38 ]. Besides the mechanisms already described, cyclin E demonstrates its oncogenic potential by a correlation with oncogenic viruses. HPV and especially HPV E7 protein which is implicated in cervix carcinoma, can lead to promotion of cyclin E - associated kinase activities through the interaction with p21 (which is a cyclin - dependent kinase inhibitor) [ 39 ]. CMV has a dual role in activating cyclin E through direct induction of cyclin E and inactivation of cyclin - dependent kinase inhibitors [ 40 ]. On the other hand, HIV - 1 halts cyclin E activity and causes G1 phase arrest, which encourages viral replication [ 41 ].
Cyclin E is a factor found in a variety of cancers like breast, ovarian, colorectal, bladder and other. In Table 2 the way cyclin E is associated with several types of cancer is depicted.
With regard to breast cancer, a remarkable number of studies have been driven. Altered expression of cyclin E occurs in 18 - 22% of the breast cancers and can serve as potential prognostic marker [ 42 ]. The expression of low molecular weight cyclin E derivatives has been investigated with great emphasis since they have been shown to be of great pathogenetic and prognostic importance for breast cancer. Low molecular weight isoforms are resistant to CKIs, bind more efficiently to CDK2 and can stimulate the cells to progress through the cell cycle more efficiently [ 43 , 44 ]. As a result resistance to anti - growth signals and genomic instability are more common. These forms have proved to be a remarkable marker of the prognosis of early -stage - node negative breast carcinoma [ 45 ].
Cyclin E has also been correlated with ovarian carcinomas, the fourth leading cause of cancer deaths among women in the United States. In ovarian carcinomas, cyclin E is overexpressed primarily in the low molecular weight isoforms [ 46 ] which are both biochemically and biologically hyperactive as mentioned before. The exact correlation between cyclin E overexpression and prognosis is controversial.
Cyclin E overexpression is also implicated in carcinomas at various sites along the gastrointestinal tract, but the most important sites are the stomach and the colorectal region [ 47 ]. As far as stomach cancers are concerned, cyclin E overexpression was found in 50 - 60% of gastric adenomas and adenocarcinomas [ 48 ]. Cyclin E was shown to be of independent prognostic significance in gastric carcinoma [ 49 ].
Regarding the colorectal carcinomas, cyclin E gene amplification is quite rare, estimated at the level of 10% [ 50 ]. Overexpression of cyclin E is detected in the early stages of the carcinogenic process promoting the morphological progression from adenoma to adenocarcinoma and the progression of early cancer [ 50 , 51 ]. Cyclin E was detected in both full length and low molecular weight forms in tumour and adjacent macroscopically normal mucosa.
Cyclin E overexpression has also been reported in melanomas. Cyclin E, in combination with other cell cycle regulators, has been proved to be of determinant significance for melanoma growth and/or transformation and this is indicared by the fact that cyclin E was not detected in benign naevi but it was easily detectable in most of the metastatic melanomas [ 52 ]. Another study has explicated the importance of the low molecular weight isoforms of cyclin E in melanoma formation. It was found that the isoforms were overexpressed in a subset of primary invasive and in metastatic melanomas but not in benign naevi. The low molecular weight forms are histologically active and can function as regulators of invasion and metastasis since they can form angiogenic tumors with prominent perineural invasion and increase the incidence of metastases in comparison to the full - length cyclin E [ 53 ].
Implication of cyclin E in the non small cell lung carcinomas has been the subject of several studies. The association with deeply invasive tumours and the function of cyclin E as an independent factor for poor prognosis were also proved by another study [ 54 ].
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
MS: partial English editing and correction, VP: partial English editing, IK: search of the literature, KX: editing and correction, IA: search of the literature, IP: search of the literature, KK: final editing and corrections. All authors read and approved the final manuscript. | CC BY | no | 2022-01-12 15:21:44 | World J Surg Oncol. 2010 Dec 22; 8:111 | oa_package/7f/1d/PMC3016250.tar.gz |
||||
PMC3016251 | 21134285 | Background
Left ventricular assist devices were approved as a bridge to transplantation therapy (BTT) in 1998 [ 1 ]. In 2001, they became a destination therapy in the United States [ 2 ]. Perioperative complications result from a combination of three factors: 1) intrinsic heart failure with secondary organ damage; 2) long-term effect of the implanted device; and 3) surgical techniques employed, including cardiopulmonary bypass (CPB). CPB has advantages such as inspecting left ventricle for thrombus, hemodynamic resuscitation, and removing fluid, including ultrafiltrations on CPB, but it may have unwanted effects as well, such as increasing systemic inflammatory response and transfusion requirements. To date, the experimental and clinical data comparing on-pump and off-pump coronary surgery suggest an affected cardiac function in favor of off-pump operations, which might be related to greater myocardial damage during hypothermic CPB operations [ 3 ]. We believe in order to improve the clinical outcome there is a need to ameliorate one of these factors. Although previous studies have demonstrated an advantage in reducing morbidity in off-pump coronary artery bypass-graft (CABG) patients, but not reducing mortality, it can be reasonably concluded that the same may hold true in sicker patients receiving off-pump left ventricular assist device (OP LVAD) implantation in high-risk heart failure patients [ 4 , 5 ]. This article analyzes our experience with the initial series of 29 consecutive patients undergoing placement of LVAD as both BTT and destination therapy, without the use of CPB, and measures postoperative mortality up to one year. | Methods
After obtaining approval from the Institutional Review Board, we retrospectively reviewed data from patients who were scheduled to undergo OP LVAD implantation at The Ohio State University Medical Center between October 2004 and May 2007. Postoperative outcomes included ventilation hours and postoperative hospital stays (days). Mean and standard deviation were calculated [ 5 ]. Early postoperative bleeding was in accordance with Interagency Registry for Mechanically Assisted Circulatory Support (INTERMACS) guidelines to ensure that the descriptions of adverse events in studies using mechanical assist devices are universal. The types of implanted devices included: 14 HeartMate II, 10 HeartMate XVE, 2 IVAD, and 1 DeBakey. The incidence rates of complications were calculated. Patient mortality was calculated at three time periods: 30 days, 90 days, and one year postoperatively. We performed a univariate analysis of all patient data to obtain descriptive statistics. Comparisons across categories were performed through Fisher's exact test.
Anesthetic Technique
We reviewed the intraoperative anesthetic chart in all study patients, which included the monitoring of arterial blood pressure, heart rate, pulmonary artery pressure, and central venous pressure. An intraoperative transesophageal echocardiogram was employed for assessment of right and left ventricular function, exclusion of associated pathologies that preclude OP implantation (intra-atrial shunts, ventricular thrombus, severe aortic insufficiency), and guidance in de-airing of the heart once the device was implanted. Heparin was administered to keep the kaolin activated clotting time (ACT) above 300 seconds. Inotropic medications were used to support right ventricular function according to the treating anesthesiologist. There was no point of care testing for transfusion requirements, which were determined by the treating physicians. The treating physician made clinical decisions about the transfusion of PRBCs or other blood products without clear logarithms.
Surgical Technique
A detailed surgical technique of OP LVAD implantation was first perfected after more than 30 intracorporeal ventricular assist devices were implanted off pump in bovine and ovine models for various studies including heart failure (unpublished data) due to the time- and cost-intensive nature of placing animals on CPB [ 6 ]. Our team's experience with this novel idea was then translated to our patients. Before starting the procedure, we performed transesophageal echocardiography to monitor underlying contraindications to OP LVAD implantation, such as severe aortic regurgitation, right-to-left shunts with worsening hypoxemia, and/or thrombus formation. Also, a primed bypass circuit was available in case of an emergent change to an on-pump procedure, such as hemodynamic instability. After median sternotomy, a pericardial well was created and a pre-peritoneal pocket formed by tunneling the driveline through the skin, with placement of the device in the pocket. After successful heparinization (ACT of 300 seconds), the ascending aorta was cannulated with an Embol-X cannula to collect debris, while the outflow graft was sewn into the ascending aorta in the standard fashion with a partial cross clamp. Opening of the inferior right pericardial edge to the inferior vena cava, with release of stay sutures, allowed cardiac positioning with the apex out of the chest as in an off-pump CABG procedure. Meanwhile, the patient was placed in the Trendelenburg position, slightly rotated towards the surgeon. An apical sewing cuff was sewn into the left ventricular apex by passing a large silk stitch into the center of the apex, while 2-0 pledgetted Ticron mattress sutures were placed in the apex and passed through the cuff. Three large Foley catheters were prepared and ready for use. The first was moistened and passed into the inflow elbow of the ventricular assist device, positioned so the tip exits the proximal end of the cannula. Next it was inflated with 10 to 15 ml of saline to plug the lumen of the cannula, while the proximal end of the Foley was clamped to stop back bleeding. The second Foley catheter was prepared similarly and used to help core the LV apex, making sure it was inflated and deflated while still providing occlusion of the lumen. The final Foley catheter was prepared similarly and used in case of rupture of one of the first two Foley balloons. The outflow graft was then attached to the LVAD and back bled to help de-air the pump. The apex was then cored with a coring knife that was passed through the Foley and silk stay suture. An incision was made into the LV apex and the Foley was placed into the cavity. The second Foley was inflated with 15 ml of saline and traction applied to bring the apex up and out of the chest cavity. The apex was cored without entering the Foley, and scissors were used to remove the remaining cuff. After coring the apex, the inflow elbow was connected by deflating and removing the balloon. To minimize ejection of blood out of the hole during balloon removal and inflow graft connection, the heart was arrested for 1 to 2 minutes with adenosine intravenous bolus (6 to 12 mg IV) or fibrillated using DC fibrillator (Cardiovascular Instrument Corp., Model 2039, Wakefield, MA, USA). A linear incision was made through the diaphragm, allowing the pump to sit in the preperitoneal space and then the diaphragm was closed around the inflow cannula, while the pump was back bled again from the aortic side for de-airing. | Results
The patients' baseline characteristics were measured in the intensive care unit prior to coming to the operating room (Table 1 ). The average patient age was approximately 50 years old and about two-thirds were male. Dilated cardiomyopathy was the most common etiology of heart failure and the average ejection fraction was approximately 15%. Of the 29 cases scheduled for OP LVAD, 27 cases were successfully performed with 18 patients as BTT and 7 as destination therapy. In one of the failed cases, the surgeon entered the right atrium during resternotomy. The second was a congenital patient with a previous Mustard operation who did not tolerate partial occlusion of the systemic great vessel. Both patients had successful implants with CPB.
Statistical analysis was performed on 27 patients. Of those analyzed, 4 patients presented with cardiogenic shock. No patients required unplanned right VAD placement. Subsequently, 5 patients had a cardiac transplantation. Of these 5 patients, 2 had a heart transplant 1 year after LVAD insertion; 1 patient was transplanted 2 months after the LVAD placement and died 8 months posttransplant from graft rejection; and the other 2 patients underwent heart transplants 2 and 5 months after LVAD implantation. We excluded those 3 patients who received transplants after less than one year when calculating the mortality. Two patients categorized as destination therapy became transplant eligible.
Eight of 27 patients experienced postoperative bleeding, which required a transfusion of more than 4 units of PRBCs for an overall incidence of 30%. One patient returned to the operating room for re-exploration due to excessive bleeding. Postoperative results are presented in Table 2 . The mean of postoperative hospital stays was 28 days with a standard deviation of 19. The average ventilation hour was 66 with a standard deviation of 93. There was no incidence of infection reported. There was 1 incidence of stroke, 6 of renal failure, 8 of respiratory failure, and 4 reintubations out of 27 operations. Causes of death in one year included sepsis with multi-organ failure (n = 6), coagulopathy with bowel necrosis (n = 1), and respiratory distress with right heart failure (n = 1). The percent of survival was 93% at 30 days, 76% at 90 days, and 67% at one year. According to a recent study on LVAD [ 6 ], the estimates of one-year survival rates were 68% for the continuous-flow LVAD and 55% for the pulsatile-flow LVAD. Our study showed similar one-year survival rates with continuous-flow LVAD and was better than the pulsatile-flow LVAD.
To investigate whether our OP LVAD technique improved over time, we compared the mortality rates of different implantation periods (Table 3 ). We divided the total study into three time intervals (2004-2005, 2006, and 2007). The death rate during this time decreased from 50% to 17%. Even though the trend was not statistically significant (p = 0.08, Fisher's exact test), we believe the limited sample size played a role. It will be promising and worthwhile to validate this trend in a large clinical trial and/or a multi-center study. | Discussion
This study reports the feasibility of a surgical technique for placement of LVADs through a median sternotomy incision without the use of CPB at our institution. This technique was used in the implantation of several different LVAD devices and was successful in 27 of 29 consecutive patients. Our initial anesthetic learning experience with OP LVAD insertion was satisfactory. Our one-year survival of OP LVAD insertion was 67%, which proved better than the Randomized Evaluation of Mechanical Assistance for the Treatment of Congestive Heart Failure (REMATCH) trial (52%) and the post-REMATCH era (56%). We believe this may be related to the avoidance of CPB, although this hypothesis has yet to be proven [ 2 , 4 ]. This was also comparable to the one-year survival of the most recent large trial using continuous-flow LVAD (68%), and better than pulsatile flow LVAD (52%), both on CPB [ 7 ]. However, our results were slightly less favorable compared with the one-year survival (73%) seen in recent trials conducted by Pagani et. al, using continuous-flow rotary LVAD using CPB in BTT patients, which correlated consistent improvements in patient outcome with newer device technology [ 8 ]. However, we believe that avoidance of CPB combined with improvements in device technology could provide greater benefits in future clinical outcomes.
In our series of 27 OP LVAD insertions, 9 patients experienced postoperative bleeding leading to transfusion of more than 4 units of PRBCs in the first 24 hours, a rate better than the historical REMATCH trial group (38%) in which all procedures were performed using CPB. In another study, using the same definition for massive perioperative bleeding as the INTERMACS registry and our study, which evaluated 222 patients with a Novacor LVAD implanted during CPB, massive perioperative bleeding occurred in 97 out of 222 patients and half of the total population returned for reoperation [ 9 , 10 ]. We believe this may be related to CPB and its deleterious effect and/or the Novacor device. However, this claim warrants further investigation in a randomized control trial with newer technology like the HeartMate II LVAD, which has different characteristics of flow.
Despite our initial encouraging results, this study has a number of limitations. While it does represent a consecutive series of patients who underwent the same surgical technique, it is a retrospective study that consists of a limited number of subjects and the hypothesis was generated after the surgical technique was established. In addition, due to the retrospective nature of the data collection, the transfusion trigger was not standardized within the first 24 hours postoperatively. Lastly, we could not do a case match control group from historical data in our institution for two reasons: 1) there was an upgrade in device technology, and 2) all of these cases were performed by one surgeon, who was using only this technique in 27 consecutive patients, and the historical group cases were performed by different surgeons on the CPB. | Conclusions
As we performed more procedures off pump, the perioperative team became more familiar with the surgical technique and anesthetic requirement, and this was reflected in our data, showing improvement over time in one-year survival post implantation. The death rate over one year decreased from 50% to 17% in our small series, although this trend was not statistically significant due to limitations in sample size; however, many events including postoperative care, etc. can affect the one year outcome. It did show promise in helping high-risk heart failure patients, and the next step is to perform this technique in a larger clinical trial to validate this hypothesis, which we believe is worthwhile in improving clinical outcomes of cases done off CPB. This would allow confirmation of our findings, while addressing the limitations of our retrospective analysis. In conclusion, based upon our findings, OP LVAD implantation is a feasible surgical technique, and combining this technique with improved device technology could provide even greater improvement in patient outcomes in the future. | Background
We hypothesize that implantation of left ventricular assist device through off-pump technique is feasible and has a comparable result to implantation on cardiopulmonary bypass and could improve one-year survival.
Methods
This retrospective, observational, single-center study was conducted on 29 consecutive patients at our institution who underwent off-pump left ventricular assist device implantation by a single surgeon.
Results
Twenty-seven procedures were performed successfully using the off-pump technique. The survival rate was 92% at 30 days, 76% at 90 days, and 67% at one year. We compared the one-year survival of different implantation periods, and divided our study into three time intervals (2004-2005, 2006, and 2007). There was a trend in reduction in number of deaths over one year that demonstrated a decrease in death rate from 50% to 17%, as well as improvement in our experience over time. However, this trend is not statistically significant (p = 0.08) due to limited sample size.
Conclusions
Based upon our findings, off-pump left ventricular assist device implantation is a feasible surgical technique, and combining this technique with improved device technology in the future may provide even greater improvement in patient outcomes. | Competing interests
The authors declare that they have no competing interests.
Authors' contributions
All authors have read and approved the final manuscript. HA: Designed study, analyzed data, and wrote manuscript. MA: Collected and analyzed data. JH: Analyzed data and wrote manuscript. GD: Provided anesthesia for cases, analyzed data, and wrote manuscript. LY: Provided statistics. BS: Invented technique and surgeon on cases. | Acknowledgements
We wish to acknowledge Tammy Yanssens, transplant coordinator at The Ohio State University Medical Center Ross Heart Hospital, for helping with data collection, and Keri J. Hudec, technical editor at The Ohio State University Medical Center Department of Anesthesiology. | CC BY | no | 2022-01-12 15:21:44 | J Cardiothorac Surg. 2010 Dec 6; 5:123 | oa_package/e0/b2/PMC3016251.tar.gz |
PMC3016252 | 21114651 | Introduction
Complement dysregulation is associated with several distinct patterns of glomerular pathology. Common to glomerular abnormalities associated with defective control of the alternative pathway is deposition of C3 in the absence of significant immunoglobulin ( 1 , 2 ). This pathological appearance typifies a number of conditions associated with genetic or acquired complement dysregulation, including dense deposit disease, C3 glomerulonephritis (C3GN) and CFHR5 nephropathy. ‘C3 glomerulopathy’ has recently been proposed as a new term under which this heterogeneous group of disorders can be classified ( 1 ).
C3GN is a feature of CFHR5 nephropathy, a familial renal disease characterized by persistent microscopic hematuria, synpharyngitic macroscopic hematuria and progressive renal failure ( 3 ). C3GN may be associated with membranoproliferative or mesangial proliferative features. Endemic in Cyprus, affected individuals have a heterozygous internal duplication in the CFHR5 gene. Previously, mutations in complement factor H, CD46 (membrane cofactor protein) and factor I were identified among patients with biopsy-proven C3GN ( 2 ), but CFHR5 nephropathy is the first description of C3GN associated with a mutation in the CFHR5 gene. CFHR5 is a member of the complement factor H (CFH) family, a group of highly related proteins encoded by genes located within the regulator of complement activation (RCA) gene cluster on chromosome 1. Comprising CFH, CFH-like protein (CFHL-1) and complement factor H-related proteins 1–5, the proteins are composed of individual domains termed short consensus repeats (SCRs), which display varying degrees of amino acid sequence similarity to each other. CFHR5 is a 65 kDa protein composed of nine SCRs, and the internal duplication in exons 2 and 3 characteristic of CFHR5 nephropathy results in an expressed protein with duplicated SCRs 1 and 2 respectively. Although the role of CFHR5 is not yet fully understood, its complement regulatory activity in vitro ( 4 ) and co-localization with renal complement deposits in vivo ( 5 ) suggest that it may play a role in complement regulation within the kidney. Furthermore, the mutant protein has been shown to have reduced affinity for glomerular-bound complement, raising the possibility of impaired targeting to complement within the kidney ( 3 ). We report the case of a 53-year-old gentleman with end-stage renal failure (ESRF) secondary to CFHR5 nephropathy, who underwent renal transplantation from a deceased donor and was found to have evidence of disease recurrence in a transplant biopsy 46 days later. | Discussion
To our knowledge this is the first description of recurrence of CFHR5 nephropathy in a transplant. The recurrence of CFHR5 nephropathy in an unrelated kidney demonstrates that local synthesis of normal CFHR5 by the kidney is not sufficient to prevent disease. However, we are aware of two other incompletely characterized cases with the CFHR5 mutation and renal disease, in which good allograft function was evident one decade after deceased donor renal transplantation. Firstly, a Cypriot male with a renal biopsy demonstrating C3GN reached ESRF at the age of 46. A deceased donor renal transplant was performed at the age of 48 and he died 12 years later from a myocardial infarction. The serum creatinine 3 months before his death was 89 μmol/L and the graft was never biopsied. He was a member of a family described previously and genotyping of his daughter and the offspring of his cousin demonstrated that he was an obligate carrier of the CFHR5 mutation (individual III-1 from family 1 in reference ( 3 )). Secondly, a 40-year-old Cypriot individual with ESRF and small kidneys, who presented with macroscopic hematuria at the age of ten but had been lost to follow up, received a deceased donor renal transplant at the age of 43. The transplant functioned well for 10 years until it was lost following atheroembolic complications of diagnostic coronary angiography. He returned to hemodialysis and died 6 years later. Subsequent molecular testing on stored genomic DNA demonstrated the presence of the CFHR5 mutation. There was no record of either native or allograft renal biopsy. These cases, together with the observations that the original disease takes many decades to cause ESRF, suggest that graft loss due to recurrence of CFHR5 nephropathy is not inevitable. Clearly larger studies will be required to establish the clinical course of CFHR5 nephropathy in renal transplantation. In summary, we describe the first reported case of recurrence of CFHR5 nephropathy in an unrelated renal transplant. Notably histological recurrence was demonstrable only 46 days after transplantation. | Complement factor H-related protein 5 (CFHR5) nephropathy is a familial renal disease endemic in Cyprus. It is characterized by persistent microscopic hematuria, synpharyngitic macroscopic hematuria and progressive renal impairment. Isolated glomerular accumulation of complement component 3 (C3) is typical with variable degrees of glomerular inflammation. Affected individuals have a heterozygous internal duplication in the CFHR5 gene, although the mechanism through which this mutation results in renal disease is not understood. Notably, the risk of progressive renal failure in this condition is higher in males than females. We report the first documented case of recurrence of CFHR5 nephropathy in a renal transplant in a 53-year-old Cypriot male. Strikingly, histological changes of CFHR5 nephropathy were evident in the donor kidney 46 days post-transplantation. This unique case demonstrates that renal-derived CFHR5 protein cannot prevent the development of CFHR5 nephropathy. | Case Presentation
A previously healthy British male with Cypriot ancestry was referred at the age of 36 with persistent microscopic hematuria, episodes of macroscopic hematuria coinciding with upper respiratory tract symptoms, and renal impairment (serum creatinine 178 μmol/L). He was not aware of any family history of renal disease at that time although relatives with CFHR5 nephropathy have subsequently been identified. Physical examination and blood pressure were normal. However urinalysis demonstrated 1+ blood and 1+ protein, and a renal biopsy was performed. Light microscopy revealed 30% glomerular obsolescence with a fibrous crescent in one glomerulus, and a few tubular red cell casts. Immunoperoxidase staining showed capillary wall C3 but notably was negative for IgA. Subendothelial electron dense deposits and rare subepithelial ‘humps’ were seen on electron microscopy (EM), as were mesangial deposits associated with an increase in mesangial cells and matrix. A further decline in renal function 6 years later led to a second biopsy ( Figure 1 ). This demonstrated large segmental scars, capillary wall thickening with double contours and mesangial cell interposition. Granular capillary wall C3 was again evident in the absence of immunoglobulin staining. On EM there were subendothelial and mesangial deposits, and rare subepithelial deposits. These histological features are consistent with C3 glomerulonephritis.
Over the following 4 years he suffered progressive renal impairment and required renal replacement therapy at the age of 47. Six years after commencing hemodialysis, during which he had further episodes of macroscopic hematuria, he received a deceased donor renal transplant. The donor was a 65-year-old female with no significant past medical history, who had a creatinine of 82 μmol/L. HLA matching demonstrated a 2,1,1 mismatch, but there were no recipient class I or II HLA antibodies at the time of transplantation, and this proceeded with no complications. Immunosuppression included alemtuzumab and corticosteroids peri-operatively, with tacrolimus monotherapy continued at discharge. Although his renal function initially improved, his creatinine stabilized at 186 μmol/L, and in the presence of persistent microscopic hematuria and following a single episode of macroscopic hematuria (with a urine protein: creatinine ratio of 56 mg/mmol), he underwent a renal transplant biopsy, 46 days after transplantation ( Figure 2 ).
Light microscopy showed occasional neutrophils in the capillary loops of one glomerulus and a small area of tubulointerstitial fibrosis. Immunoperoxidase staining showed capillary wall granular C3 and complement component 9 (C9), whilst electron microscopy showed increased mesangial matrix associated with scattered mesangial and subendothelial deposits (with new basement membrane beneath in two areas; Figure 2C ). Complement component 4d (C4d) staining was negative. The findings in the renal transplant were therefore consistent with a recurrence of his original disease. Serum C3 was 0.73 g/L (normal range 0.7–1.7 g/L) at the time of biopsy, whilst CFH and factor I levels were 320 mg/L (207% of normal control) and 32 mg/L (177% of normal control), respectively.
A second transplant biopsy performed 3 months later on account of a rise in the serum creatinine to 250 μmol/L, showed granular (mainly mesangial) C3 staining associated with subendothelial, mesangial and subepithelial deposits. Indirect immunofluorescence also showed the presence of glomerular complement components 5b (C5b)-9 ( Figure 2D ). Polymerase chain reaction (PCR) using genomic DNA isolated from peripheral blood monocytes and serum CFHR5 western blot analysis ( Figure 3 ) revealed the presence of the heterozygous internal duplication in the CFHR5 gene, confirming the diagnosis of CFHR5 nephropathy. | The authors would like to thank the staff of the Histopathology department and Dr. Jill Moss (Electron Microscopy Unit), Imperial College Academic Health Science Centre, London. K.A.V. is supported by a Kidney Research UK Clinical Research Training Fellowship. M.C.P. is a Wellcome Trust Senior Fellow in Clinical Science (WT082291MA), and E.G.d.J. is also funded by this fellowship.
Disclosure
Commercial Organizations
This manuscript was neither prepared nor funded in any part by a commercial organization.
Conflict of Interest
The authors of this manuscript have conflicts of interest to disclose as described by the American Journal of Transplantation . These are as follows:
PH Maxwell
Consultant: Bio Nano Consulting (Paid Director); Ownership: ReOx Ltd.—Director and stockholder;
Honoraria: Ipsen, Roche; Scientific Advisor: Roche Foundation for Anaemia Research (RoFAR); Other: Registrar (unpaid)—Academy of Medical Sciences, Board Member (unpaid)—Medical Education England, Chair Physiological Sciences Funding Committee (Paid)—Wellcome Trust.
DP Gale
Honoraria: Alexion
Abbreviations
complement component 3
complement component 3c
complement component 4d
complement component 5b
complement component 9
cluster of differentiation 46 (membrane cofactor protein)
complement factor H
CFH-like protein
complement factor H-related protein 5
C3 glomerulonephritis
electron microscopy
end-stage renal failure
micromole (s)
milligram (s)
millimole (s)
polymerase chain reaction
regulator of complement activation
short consensus repeat | CC BY | no | 2022-01-12 15:46:44 | Am J Transplant. 2011 Jan; 11(1):152-155 | oa_package/5e/f7/PMC3016252.tar.gz |
|||
PMC3016253 | 21162744 | Background
The use of pre- or peri-workout supplements among recreational and elite athletes have become increasingly popular due to studies suggesting improvements in aerobic and anaerobic performance and recommendations from expert panels in sports nutrition [ 1 ]. Among the most commonly used supplements for increasing muscular strength are those containing various creatine salts including creatine monohydrate [ 2 ], carbohydrate, protein [ 3 ], and amino acids [ 4 ], particularly branched chain amino acids (BCAA), for which evidence of effectiveness has been consistently seen in published studies [ 1 ]. Numerous studies have assessed the effectiveness of the individual supplements listed above, and have established a range of doses at which the specific supplement showed demonstrable effects. These studies have helped to establish minimal/threshold doses at which supplements exert their intended effects. Research data is most plentiful on supplementation with creatine monohydrate, carbohydrates, and protein and these three ingredients are consistently recommended by expert panels as ergogenic aids, and as such are the core constituent ingredients of many pre- and peri-workout supplements.
Based on the findings of such research and expert recommendations, supplement manufacturers have developed sports drinks combining the same three core ingredients and have added proprietary ingredients to be used in the peri-workout time period to increase muscle strength, lean mass, and/or endurance. Aside from the convenience of having multiple ingredients in one product, there is potential for the components to exert additive or synergistic effects. Because different dietary supplement products contain differing quantities of the core and proprietary components, it is often difficult to perform valid head-to-head studies. However, because most products purporting to build strength and/or endurance contain the same three core ingredients, and the preponderance of evidence suggests that these three ingredients are the most important contributors to observed ergogenic gains, then it is reasonable to assume that if similar quantities of the core ingredients were compared, a valid comparison could be made. If differences were found between two products, then a likely explanation for the difference would be some effect of the proprietary ingredients, since the core ingredients are matched by dose. Proprietary ingredients could contribute to a difference either by exerting independent effects or by enhancing the effects of the core ingredients in a differential way or both.
Size On Maximum PerformanceTM (SOmaxP) is a product manufactured by Gaspari Nutrition containing creatine, carbohydrate, whey protein and other proprietary ingredients, and was used during the peri-workout period only on the days when resistance training occurs. The comparator product was standardized to contain similar amounts of creatine, carbohydrate and whey protein. The study compared the effects of SOmaxP to a comparator product (CP), which was standardized to contain equal amounts of creatine (4 g creatine monohydrate), carbohydrate (39 g maltodextrin) and protein (7 g whey protein hydrolysate), and given with identical timing. We hypothesized that subjects in the SOmaxP groups would outperform the subjects in the CP during post-testing after adjusting for baseline differences. | Methods
Subjects
Twenty subjects, ten in each group, were randomized to receive either SOmaxP or CP during this 9-week study. Key elements of the inclusion criteria included: male or female subject in good health; aged between 18-45; a body fat of 10%-25% inclusive; who had undergone regular resistance training for at least two years; who had signed an informed consent; who were willing and able to comply with the training and supplement protocol; possessed normal vital signs; and had a fluent understanding of English. Physical activity levels and health history were determined using standardized questionnaires adapted from Kent State University, Purdue University, and Eastern Michigan University at baseline and weeks 3, 6 and 9. The protocol was in compliance with the Helsinki Declaration, and was approved by the IntegReview Ethical Review Board (Austin, TX). Although the inclusion criteria allowed for female subjects, no females enrolled in the study. The actual age range of subjects who participated in the study was 19-31 years.
Key exclusion criteria included: a history of various metabolic conditions or diseases; the concomitant use of a variety of medications, including but not limited to those with androgenic and/or anabolic effects; the use of nutritional supplements known to improve strength and/or muscle mass (e.g., creatine, HMB, androstenedione, DHEA, etc.) within six weeks prior to the start of the study; a weight gain or loss of more than 10 lbs. within the past 30 days; known allergy to any ingredients in SOmaxP Maximum PerformanceTM or CP; participation in other research studies within the last 30 days; the current use of tobacco products; and the presence of any orthopedic limitations or injuries.
Study Design
The study was a prospective, randomized, double-blind, parallel-group clinical trial. Subjects were matched into two groups according to body mass, age, and resistance training experience. Subjects were then randomly assigned (via the ABBA procedure [ 5 ]) to receive either SOmaxP or CP. Following informed consent and prior to the first testing session, a research nutritionist and Certified Strength and Conditioning Specialist (CSCS) met with each subject and discussed in detail the strength training regimen, and nutritional and supplement requirements for the study period.
Testing Sessions
Prior to pre-testing, subjects were instructed to refrain from heavy exercise for 48 hours and fast for at least 12-hours. The assessment of upper body muscular strength (1-RM) and repetitions to failure (RTF) testing was performed after a general warm-up of 3-5 minutes of light activity involving the muscle(s) to be tested (e.g., upper body ergometry prior to upper body strength testing). Next, the subject performed several minutes of static stretching exercises of the involved musculature. The subject then performed a specific warm-up set of 8 repetitions at approximately 50% of the perceived 1-RM followed by another set of 3 repetitions at 70% of the perceived 1-RM. Subsequent lifts were single repetitions of progressively heavier weights until failure. The initial increments in weight were evenly spaced and adjusted such that at least two single lift sets was performed between the three repetition warm-up set and the estimated 1-RM. At failure, a weight approximately midway between the last successful and failed lift was attempted. This process was repeated until the 1-RM was determined. The rest interval between sets was between 3-5 minutes (procedure modified from Brown et al., 2001) [ 6 ]. Results were obtained at baseline, and at week 3, 6 and 9. For testing at weeks 3, 6 and 9, in order to replicate pre-supplementation/baseline testing conditions as closely as possible, subjects were instructed to follow their previously recorded 3-day diet records, refrain from heavy exercise for 48 hours, and fast for at least 12-hours prior to the workout. Upper body muscle endurance was measured as the total repetitions completed during three successive sets of isotonic bench press at a load equal to 100% subjects' pre-testing body weight. Each set was separated by a one-minute rest period.
Body Composition Assessment
Body composition was assessed at baseline, and weeks 3, 6 and 9. Standing height was determined using a wall-mounted stadiometer. Body weight was measured using a SECATM Medical Scale. Lean mass and fat mass were assessed using dual energy x-ray absorptiometry (DEXA, General Electric LUNAR DPX Pro). For each subject, the same technician performed all four DEXA measurements.
Supplementation Protocol
SOmaxP contains creatine monohydrate (4 g), carbohydrate (39 g), and whey protein (7 g), and a number of proprietary ingredients. Subjects randomized to the SOmaxP group took 1 serving of SOmaxP + 30 ounces of water starting 10-15 minutes before the workout and finishing before the end of the workout, and used the product only on the days when resistance training occurs. The comparator product (CP) was standardized to contain equal amounts of creatine monohydrate (4 g), carbohydrate (39 g maltodextrin) and protein (7 g whey protein), and given with 30 ounces of water, with identical timing, and similarly used only on resistance training days. The CP was virtually indistinguishable in taste, color and consistency to SOmaxP. The supplements were prepared in powder form and packaged in coded generic containers for double-blind administration by an independent company (Command Nutritionals, Fairfield, NJ). Compliance to the supplementation protocol was monitored by a research nurse/dietician who contacted the study subjects on a weekly basis by telephone. Subjects were required to bring in their supplement bottles on workout days at weeks 3, 6 and 9 for visual inspection by study personnel to assess compliance with the protocol.
Side Effect Assessment
A questionnaire was completed at weeks 3, 6 and 9 (workout sessions 12, 24 and 36) to monitor individual changes in DOMS and assess potential adverse events and change in sleep habits, general attitude, irritability, appetite, thirst, muscle soreness, muscle cramping, stomach distress, and headache, as well as any other idiosyncratic responses to the supplementation/training protocol. If identified, events were recorded as adverse events. In addition, subjects were contacted on a weekly basis by phone contact to inquire if they had experienced any adverse events, and were told to call at any time during the study to report side effects.
Dietary (Nutrition) Monitoring
The research dietitian met with each subject to explain the proper procedures for recording dietary intake. Each subject's baseline diet (3-days: two weekdays & one weekend day) was analyzed using the NutraBase IV Clinical Edition, (CyberSoft, Inc., Phoenix, AZ) to determine its energy and macronutrient content. Additional 3-day diet records were analyzed at weeks 3, 6 and 9 to verify that eating habits had remained consistent throughout the study.
Resistance Training Protocol
All subjects followed a specific 4-day per week workout designed by a Certified Strength and Conditioning Specialist (CSCS). The workout involved training the upper and lower body twice per week using a 4-day split (i.e., upper body 1 , lower body 1 , upper body 2 , lower body 2 ) with gradual increases in volume and intensity. The workout consisted of at least 12 exercises, including but not limited to: bench press, lat pulldown, shoulder press, seated row, shoulder shrug, dip, biceps curl, triceps push down, leg press, squat, deadlift, lunge, leg curl, leg extension, and calf raise. For each exercise, subjects performed 3-6 sets of 8-15 repetitions with as much weight as they could handle with good form (typically 70-85% of the 1-repetition maximum). As subject strength and endurance improved, training resistances were progressively increased to maintain the required repetition range. Rest periods between exercises were 1-3 minutes, and between sets were 60-120 seconds. Training was conducted at the subject's local training facility, documented in training logs, and signed off by fitness instructors/gym personnel to verify compliance. Two different facilities were utilized and identical equipment was available at both facilities. In addition, at each session, the subject completed a physical activity question, which described their physical activity during the preceding month. A schematic of the training program is displayed below in Figure 1 .
Clinical Laboratory Chemical Analyses
Laboratory measures were performed at baseline, and weeks 3, 6 and 9. The tests included a complete blood count (CBC) with differential and platelet count, and a chemistry panel, which included sodium, potassium, chloride, carbon dioxide, calcium, AP, AST, ALT, bilirubin, glucose, blood urea nitrogen, creatinine, albumin, globulin, and estimated glomerular filtration rate, The lipid panel (total cholesterol, HDL- and LDL-cholesterol) was drawn at baseline and at week 9. Quest Diagnostics (Pittsburg, PA) was utilized to transport and analyze all blood samples.
Statistical Analysis
Separate analyses of co-variance (ANCOVA), using baseline scores as the covariate were used to analyze between-group differences in body composition, muscular performance, and clinical markers of safety. Data was considered statistically significant when the probability of a type I error was less than or equal to 0.05 (P ≤ 0.05). If a significant group, treatment and/or interaction was observed, least significant differences (LSD) post-hoc analyses were performed to locate the pair-wise differences between means. | Results
Demographics
The demographic characteristics of the two cohorts were similar, and these are presented in Table 1 . All 20 subjects were male, and the age range was 19-31 years. The mean values for age, height, weight, baseline fat percentage, blood pressure and resting heart rate were similar in the two cohorts.
Performance Measures
A summary of the performance and outcome measures at baseline ("Pre") and at week 9 session ("Post") are presented in Table 2 and discussed below. The values are the mean values per cohort at baseline and week 9. Figure 2 displays these data using the least square mean ANCOVA analysis for 1 RM. Figure 3 displays the ANCOVA for Repititions to Failure (RTF). Figure 4 displays the ANCOVA for percent body fat. Figure 5 displays the ANCOVA for lean mass. Figure 6 displays the ANCOVA for fat mass. Statistically significant differences between the SOmaxP and CP cohorts were observed for 1 RM (p = 0.019), RTF (p = 0.004), body fat percent (p = 0.028), lean mass (p = 0.049), and fat mass (p = 0.023).
The measures of muscular performance (1-RM and RTF total) increased in both the SOmaxP and CP cohorts, though by a higher percentage in the SOmaxP group. The 1 RM for the SOmaxP cohort increased from 233.5-283.5 lbs. [106.1-128.9 kg] from pre- to post-testing (21.4% increase), while the CP cohort increased from 256.5-292.5 lbs. [116.6-132.9 kg], (14.0% increase). The RTF for the SOmaxP cohort increased from 19.6 to 30.25 from pre- to post-testing (54.3% increase), while the CP cohort increased from 26.3 to 30.8 (17.1% increase).
Several measures of body composition differed statistically between the two cohorts, with the SOmaxP cohorts demonstrating favorable improvements. The body fat percentage in the SOmaxP group decreased from 16.8% to 15.5% from pre- to post-testing (7.7% decrease), while the CP cohort increased slightly from 16.5% to 16.9% (2.4% increase). Lean body mass increased in the SOmaxP group from 62.7 kg to 64.2 kg (2.4% increase), while the CP cohort increased marginally from 62.6 kg to 62.8 kg (0.3% increase). Body weight did not change significantly in either group, with the SOmaxP group experiencing a drop of 1.5 kg from a baseline of 81.1 kg to 80.8 kg (0.2 kg decrease), while the CP cohort gained 1.5 kg from a baseline of 79.9 kg to 80.2 kg (0.2 kg increase). Finally, in the SOmaxP cohort, fat mass decreased from 13.5 kg to 12.2 kg (9.6% decrease), while the CP cohort increased from 13.3 kg to 13.8 kg (3.8% increase). The percentage change from baseline (Post minus Pre × 100) in strength measures (RTF(t) and 1-RM) are presented in Figure 7 below, and similar changes in body composition measures (lean mass, body fat percentage and fat mass) are presented in Figure 8 .
There were no clinically meaningful changes in vital signs or laboratory results from baseline to Week 9. One subject experienced an adverse event. The subject was a 20 year-old male, (SOmaxP group) who experienced seasonal flu symptoms during Week 8 of the study. Symptoms included nausea, vomiting, and decreased appetite, and the events were not assessed as related to study product. Symptoms were resolved at the Week 9 post-testing visit. There were no significant changes in dietary intake for the subjects in either cohort, based on dietary diary evaluation. | Discussion
This double-blind, comparator study showed that nine weeks of supplementation with SOmaxP resulted in statistically significant improvements in muscular performance (1-RM and RTF), decreases in body fat and fat mass, and increases in lean mass, versus a comparator product matched with similar amounts of creatine, carbohydrate and whey protein. Both the SOmaxP and CP were well-tolerated, and there were no changes in laboratory measures or vital signs during the study. There were no adverse events assessed as related to either product, and no significant changes in body weight occurred during the study period in either group.
The SOmaxP cohort experienced an increase in strength and a concomitant increase in lean muscle mass and loss in body fat, without a significant change in body weight. These changes are consistent with a desired anabolic effect. Improvements in strength were also noted with the CP, though significantly less than with SOmaxP. The dose of creatine in this study (4 g/workout or 16 g/week) for both the SOmaxP and CP cohorts is lower than what is recommended by some of the more commonly described creatine protocols 1 , and yet strength gains were noted in both the SOmaxP and CP groups. Typical protocols recommend ingesting approximately 0.3 g/kg/day of creatine monohydrate for 5-7 days as a loading dose (e.g., 5 g 4 times per day), followed by 3-5 g/day thereafter [ 7 , 8 ]. A few studies have found that a loading period was not necessary for increasing muscle creatine (3 g/day for 28 days) [ 9 ], or muscle size and strength (6 g/day for 12 weeks) [ 10 , 11 ]. A loading dose was not used in this study for either cohort. Data from the current study show measurable strength gains at a creatine dose of 16 g/week without a loading dose.
The CP cohort gained strength, but only had a slight increase in lean mass, body fat % and body weight. A possible explanation for this is that the CP group, taking a similar 16 g/week of creatine monohydrate experienced physiologic changes sufficient to increase strength, but not sufficient to measurably increase lean mass. This finding is consistent with work by Rawson et al. (2010), who found that subjects who received low dose creatine (2.3 g/day or 16.1 g/week) for six weeks, experienced a significant increase in plasma creatine, and statistically significant enhanced fatigue resistance without weight gain compared to a matched placebo group [ 12 ].
There are several possible explanations for the statistically significant difference between the SOmaxP group and CP, and these may be explained in part by several of the proprietary ingredients. SOmaxP contains a large quantity of branched chain amino acids. Branched chain amino acids (BCAAs), particularly leucine, have been shown to have anabolic effects, presumably through reducing protein breakdown [ 13 ]. BCAAs have also been shown to increase the lactate threshold during an incremental exercise test in trained individuals [ 14 ]. Blood lactate concentrations increase significantly during intense exercise as anaerobic glycolysis becomes the dominant energy pathway [ 15 ]. In addition, the combined ingestion of protein and leucine with carbohydrate has been shown to increase post exercise muscle protein in male subjects [ 16 ]. BCAAs also activate key enzymes in protein synthesis [ 17 ], and act in a synergistic fashion with insulin to allow skeletal muscle to coordinate protein synthesis [ 18 ].
In addition, SOmaxP contains isomaltulose (palatinose) as part of its carbohydrate moiety. This carbohydrate is present in honey and has been associated with delayed digestion and absorption, which may account for the difference in body fat changes between the SOmaxP group and the CP group. Oizumi and colleagues (2007) developed a palatinose-based balanced formula (PBF) for use in human subjects with impaired glucose tolerance [ 19 ]. During a 12-week cross-over study of dietary intervention in 23 subjects with impaired glucose tolerance, the authors found that A 250 kcal can of PBF once per day had beneficial effects on serum free fatty acid levels and visceral fat area. Visceral fat area decreased by 17.1% in the PBF period compared to 5.1% in the control period. Abdominal fat area decreased by 7.7% in the PBF interval while gaining 3.7% in the control period. Free fatty acids decreased by 22% in the PBF intervention, while increasing by 18.7% during the control period, and the 2-hour post-prandial glucose level decreased by 15.7% in the PBF intervention group while increasing by 0.8% in the control period. A possible mechanism for this finding was described in an animal study by Matsuo et al. (2007), who found that a palatinose-based liquid formula suppressed postprandial glucose level and reduced visceral fat accumulation compared to a standard formula [ 20 ]. These data suggest that palatinose-based carbohydrates may have beneficial effects on fatty acid and glucose metabolism.
In addition, Achten et al. (2007) compared the oxidation rates from orally ingested sucrose and palatinose (250 kcal) during moderately intense exercise [ 21 ]. The authors found that in trained athletes cycling for 150 minutes at approximately 60% of VO 2 max experienced significantly lower oxygen consumption with palatinose compared to sucrose, resulting in a lower plasma insulin response at 30 minutes compared to sucrose. Subjects consumed either water or 1 of 2 carbohydrate solutions (sucrose or isomaltulose) providing 1.1 g/min of carbohydrate. The authors concluded that the lower carbohydrate delivery and a small difference in plasma insulin may have resulted in a higher endogenous carbohydrate use and higher fat oxidation during the isomaltulose trial than during the sucrose trial.
Another possible ingredient the SOmaxP that may contribute to the results of this study is L-ornithine-L-aspartate (LOLA), a substance shown to be effective in lowering blood ammonia concentration, particularly in patients with hepatic encephalopathy [ 22 ]. LOLA was administered at a dose of 20 g/day dissolved in 250 mL of 5% fructose solution and infused intravenously for a period of 4 hours during 7 consecutive days with a superimposed protein load at the end of the daily treatment period. Treatment was associated with a significant decrease in cerebral ammonia levels, which have been shown to be increased in subjects undergoing prolonged exercise [ 23 ]. Secher and colleagues (2008) reviewed the changes in cerebral blood flow and metabolism, and suggested that ammonia accumulation played a likely role in the development of what is known as central fatigue [ 24 ]. The efficacy of both oral and parenteral LOLA was confirmed by randomized, placebo-controlled, double-blind studies in patients with manifest hepatic encephalopathy and hyperammonemia [ 25 ]. The drug was able to reduce high blood ammonia levels induced either by ammonium chloride or protein ingestion or existing as a clinical complication of cirrhosis per se . Furthermore, LOLA improved performance in Number Connection Test-A as well as mental state gradation in patients with more advanced hepatic encephalopathy. Stauch et al (1998) found an improvement in cerebral ammonia levels compared to placebo using an oral dose of 6 gm per day [ 26 ].
In another published trial, LOLA decreased protein breakdown and stimulated protein synthesis in muscle in patients with hepatic encephalopathy [ 27 ]. The therapy had minimal side effects, increasing with higher intravenously administered dosages, and was well-tolerated after oral and parenteral administration. It is unclear if these results are generalizable to a healthy population, but the encephalopathy studies show that LOLA clearly has beneficial effects on the central nervous system and could possibly have an effect on central fatigue.
We acknowledge some limitations to the study. No females enrolled in the study, although some were approached for possible inclusion. The study group was small and homogenous, with a relatively tight age range, on the younger side of the eligibility criteria. No attempts were made to identify the physiologic mechanism for any differences between the two groups. The study attempted to control for the use of other supplements during the study, but did not perform any testing to verify non-use of other supplements. | Conclusions
The use of SOmaxP four times per week for nine weeks resulted in statistically significant improvements in strength, muscle endurance, lean muscle mass, and percentage body fat versus a comparator with identical quantities of creatine, whey protein and carbohydrate. Given that the quantities of the core components were identical, and these components are presumed to contribute most to ergogenic effects, the differences between the SOmaxP and CP groups may be due to additive or synergistic effects of the proprietary ingredients in SOmaxP. Additional research is needed to further elucidate these effects. A double-blinded, comparator controlled study of six weeks duration which includes muscle biopsy measurements is currently underway to examine and possibly help identify genetic and pharmacological mechanisms by which SOmaxP may exert these effects. | Background
The purpose of this study was to compare the effects of supplementation with Gaspari Nutrition's SOmaxP Maximum PerformanceTM (SOmaxP) versus a comparator product (CP) containing an equal amount of creatine (4 g), carbohydrate (39 g maltodextrin), and protein (7 g whey protein hydrolysate) on muscular strength, muscular endurance, and body composition during nine weeks of intense resistance training.
Methods
Using a prospective, randomized, double-blind design, 20 healthy men (mean ± SD age, height, weight, % body fat: 22.9 ± 2.6 y, 178.4 ± 5.7 cm, 80.5 ± 6.6 kg, 16.6 ± 4.0%) were matched for age, body weight, resistance training history, bench press strength, bench press endurance, and percent body fat and then randomly assigned via the ABBA procedure to ingest 1/2 scoop (dissolved in 15 oz water) of SOmaxP or CP prior to, and another 1/2 scoop (dissolved in 15 oz water) during resistance exercise. Body composition (DEXA), muscular performance (1-RM bench press and repetitions to failure [RTF: 3 sets × baseline body weight, 60-sec rest between sets]), and clinical blood chemistries were measured at baseline and after nine weeks of supplementation and training. Subjects were required to maintain their normal dietary habits and follow a specific, progressive overload resistance training program (4-days/wk, upper body/lower body split) during the study. An intent-to-treat approach was used and data were analyzed via ANCOVA using baseline values as the covariate. Statistical significance was set a priori at p ≤ 0.05.
Results
When adjusted for initial differences, significant between group post-test means were noted in: 1-RM bench press (SOmaxP: 133.3 ± 1.3 kg [19.8% increase] vs. CP: 128.5 ± 1.3 kg [15.3% increase]; p < 0.019); lean mass (SOmaxP: 64.1 ± 0.4 kg [2.4% increase] vs. 62.8 ± 0.4 kg [0.27% increase], p < 0.049); RTF (SOmaxP: 33.3 ± 1.1 reps [44.8% increase] vs. 27.8 ± 1.1 reps [20.9% increase], p < 0.004); and fat mass (SOmaxP: 12.06 ± 0.53 kg [9.8% decrease] vs. 13.90 ± 0.53 kg [4.1% increase], p < 0.024). No statistically significant differences in vital signs (heart rate, systolic and diastolic blood pressures) or clinical blood chemistries were noted.
Conclusions
These data indicate that compared to CP, SOmaxP administration augments and increases gains in lean mass, bench press strength, and muscular performance during nine weeks of intense resistance training. Studies designed to confirm these results and clarify the molecular mechanisms by which SOmaxP exerts the observed salutary effects have begun. Both SOmaxP and the CP were well-tolerated, and no supplement safety issues were identified. | Affiliations
S. Schmitz is not affiliated with any institution. J. Hofheins and R. Lemieux are associated with The Center for Applied Health Sciences, Division of Sports Nutrition and Exercise Science. Mr. Lemieux works as the strength coach for Kent State University.
Competing interests
Stephen Schmitz declares he has a potential competing interest as he is non-employee, part-time, paid consultant for Gaspari Nutrition, working specifically in the areas of dietary supplement adverse event monitoring and reporting for the company. Jennifer Hofheins and Robert Lemeiux declare that they are employed by the Center for Applied Health Sciences, which conducted the study. However, neither individual was compensated above and beyond their customary amount as a result of this study. Gaspari Nutrition is paying the JISSN article processing charges; however, no Gaspari Nutrition employee was involved in the writing of this article.
Authors' contributions
SS was the primary author of the manuscript. JH worked at the study site, was involved in subject recruitment, data collection and editing of the manuscript. RL developed the workout routine for the protocol. All three authors have read and approved the manuscript. | CC BY | no | 2022-01-12 15:21:44 | J Int Soc Sports Nutr. 2010 Dec 16; 7:40 | oa_package/53/2e/PMC3016253.tar.gz |
|
PMC3016254 | 21138552 | Background
We previously reported that allanxanthone C and macluraxanthone, two xanthones purified from Guttiferae trees, display in vitro antiproliferative and proapoptotic activities in leukemic cells from chronic lymphocytic leukemia (CLL) and leukemia B cell lines.
Results
Here, we investigated the in vivo therapeutic effects of the two xanthones in a xenograft murine model of human CLL, developed by engrafting CD5-transfected chronic leukemia B cells into SCID mice. Treatment of the animals with five daily injections of either allanxanthone C or macluraxanthone resulted in a significant prolongation of their survival as compared to control animals injected with the solvent alone ( p = 0.0006 and p = 0.0141, respectively). The same treatment of mice which were not xenografted induced no mortality.
Conclusion
These data show for the first time the in vivo antileukemic activities of two plant-derived xanthones, and confirm their potential interest for CLL therapy. |
To the Editor,
Despite recent therapeutic advances with the combination of purine analogs, alkylating agents and monoclonal antibodies, chronic lymphocytic leukemia (CLL) remains an incurable disease [ 1 - 3 ]. It is characterized by the clonal expansion of a population of CD5 + B lymphocytes and by the accumulation in the blood of leukemic cells that are quiescent but defective in their apoptotic program [ 2 , 4 ]. Thus, CLL is a disease of proliferation as well as accumulation. Treatments targeting both dividing and apoptosis-deficient quiescent cells might therefore improve the CLL patients' outcome [ 2 - 4 ]. A number of plant-derived compounds were found to exhibit in vitro capacities to either inhibit leukemic cell growth or induce apoptosis or both, but their clinical use was hampered by the lack of in vivo studies on animal models of CLL. However, some murine models recapitulating the human CLL disease were described lately, such as the TCL1 transgenic mouse model developing a CD5+ B cell lymphoproliferative disease typical of aggressive CLL [ 5 ]. We previously showed that several xanthones purified from african trees of the Guttiferae family display both antiproliferative and proapoptotic properties in cell lines derived from CLL and hairy cell leukemia (HCL), another chronic B-cell leukemia [ 6 ]. In addition, these compounds can induce the apoptosis of primary CLL cells in vitro through different mechanisms [ 6 ]. It seemed therefore crucial to determine whether some xanthones are capable of in vivo therapeutic effects in an animal model of CLL.
We selected two of the xanthones which were purified and characterized in our previous study [ 6 ] on the basis of their in vitro activities in CLL cells and their hardly detectable toxicity in B lymphocytes from healthy donors: (i) allanxanthone C, a xanthenedione that we have identified as acting by caspase activation, possibly through a mechanism involving inhibition of the NO pathway [ 4 ]; and (ii) macluraxanthone, originaly found to inhibit the growth of solid tumor cell lines [ 7 ] and moreover, capable of triggering the mitochondrial pathway of apoptosis in CLL cells [ 6 ]. Taking advantage of our previous data [ 8 ], we developed a xenograft mouse model by engrafting CD5-transfected human JOK-1 cells into SCID mice (Le Ster et al , submitted). Actually, it was demonstrated that transplantation of this cell line JOK-1 into SCID mice led to the establishment of a CLL model, allowing the evaluation of the antileukemic efficacy of fludarabine phosphate [ 9 ]. Furthermore, we reported that CD5 plays a prominent role in the control of CLL cell apoptosis through its distribution in lipid rafts and its interaction with the B-cell receptor [ 10 ]. Whereas CD5 is generally lost in long-term cultures of CLL cell lines, JOK-1/5.3 cells derived by stable transfection of the human CD5 gene into JOK-1 cells display a phenotype somewhat close to that of primary leukemic cells. The xenografted mice that we obtained developed a leukemia resembling the CLL type as defined by the French-American-British criteria.
We first verified that the xanthones were active on the JOK-1/5.3 cells used for engrafting the mice. Treatment with either allanxanthone C or macluraxanthone for 18 h resulted in a concentration-dependent inhibition of cell growth, peaking at respectively 40% and 70% with 40 μM (estimated by 3 H-thymidine uptake), in accordance with our previous data on CLL and HCL cell lines [ 6 ]. Both compounds induced the accumulation in the G 0 /G 1 phase of the cell cycle as compared to untreated cells ( P < 0.05) and decreased the percentages of cells in S and G 2 /M phases (evaluated by propidium iodide incorporation using flow cytometry and Multicycle AV program). Two other xanthones, 1,7-dihydroxanthone and α-mangostin which were inactive in our previous study [ 6 ] were used as negative controls. The proapoptotic capacities of allanxanthone C and macluraxanthone were also checked in JOK-1/5.3 cells by stimulation of phosphatidylserine externalization (quantified by annexin V-FITC binding), although these cells turned out to be less sensitive than primary CLL cells.
For the in vivo experiments, randomised groups of SCID CB-17 mice were inoculated with 10 7 JOK-1/5.3 cells (day 0). Xenografted mice were treated at days 3 to 7 with five daily injections of either allanxanthone C or macluraxanthone (5 mg/kg) or solvent alone as untreated control. The three groups of mice were then monitored daily and the survival was estimated according to the Kaplan-Meier's method (Figure 1 ). Mean survival times ± SE were 25.6 ± 0.6 days and 26.0 ± 1.7 days for respectively allanxanthone C and macluraxanthone-treated mice versus 20.2 ± 0.8 days for untreated control mice. These increases in survival (27% and 29% respectively) were significant with P values of 0.0006 for allanxanthone C group and of 0.0141 for macluraxanthone group as compared to control group (according to the Student's unpaired t-test). No significant difference was detected between the two groups of xanthone-treated mice ( P = 0.83). These results show that treatments of the xenografted mice with allanxanthone C and macluraxanthone resulted in a prolongation of their lifespan.
To check a toxicity of the xanthones, two groups of 5 mice which were not inoculated with JOK-1/5.3 cells were treated with either allanxanthone C or macluraxanthone according to the same protocol as before. No lethality was observed in these two groups of animals, suggesting an absence of toxicity of the xanthones per se under the treatment protocol used. This also favors that the deaths observed in the JOK-1/5.3-grafted mice were due to the presence of the leukemic cells, and that treatments with the xanthones were able to delay significantly these lethal effects.
In conclusion, results presented in this letter show for the first time that allanxanthone C and macluraxanthone purified from Guttiferaes are capable of in vivo antileukemic effects in a xenograft murine model of human CLL. These therapeutic activities of the natural compounds, of similar extent, occur without apparent toxicity. Although the comparison with known chemotherapeutic agents has to be performed, our data provide further confirmation that these xanthones might be used as new agents for the therapy of CLL and possibly allied chronic B cell malignancies. Experiments examining the effects of increasing doses and time of treatment as well as different schedules of administration are in progress in order to improve the therapeutic efficacy of the two xanthones. Studies of their exact mechanisms of action in primary CLL patients' cells are also considered in order to define therapeutic targets.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
SV performed in vivo studies, analyzed the data and revised the manuscript; KLS performed in vitro experiments; MM purified the xanthones. CBe contributed to design the study; PY designed the study; JPK designed the study, interpreted the data and revised the manuscript; CBi interpreted the data and wrote the manuscript. All authors read and approved the final manuscript. | Acknowledgements
We are grateful to Drs A.G.B. Azebaze (University of Douala, Cameroun) and A.E. Nkengfack (University of Youndé, Cameroun) for their invaluable contribution to the obtention of the xanthones. | CC BY | no | 2022-01-12 15:21:44 | J Hematol Oncol. 2010 Dec 7; 3:49 | oa_package/d8/eb/PMC3016254.tar.gz |
|||||
PMC3016255 | 21172004 | Background
Salmonella is worldwide one of the most important causes of zoonotic disease, associated with consumption of contaminated food. Most important food vehicles for Salmonella are eggs and egg products. The serotype typically associated with eggs and egg products is Salmonella Enteritidis. This serotype is highly prevalent in live poultry, both in broilers and laying hens. In live chickens, also many other Salmonella serotypes are detected, but egg-associated transmission to humans seems to be a characteristic that is specifically reserved for the serotype Salmonella Enteritidis, explaining why this particular serotype has caused a worldwide pandemic since the mid '80s [ 1 , 2 ].
Eggs can be contaminated on the outer shell or internally. Outer shell contamination is the result of environmental contamination after shedding of the pathogen by the animal. Internal egg contamination can in principle occur following by eggshell and eggshell membrane penetration, but it is believed that most often colonization of the reproductive tract and incorporation in the forming egg is the main route [ 3 ]. Egg formation in the oviduct takes about 24 hours during which the sequential addition of molecules takes place in different compartments of the reproductive tract while the forming egg is migrating from the ovary to the vagina, before being laid. Indeed, the yolk is produced in the ovary, the infundibulum (most upper part of the oviduct) captures the yolk, the magnum segment produces the egg white, the isthmus deposits the eggshell membranes, the uterus forms the shell and the vagina is involved in oviposition. Depending on the oviduct segment in which Salmonella bacteria are present, the bacteria can thus be incorporated in different parts of the egg. Yolk contamination would lead to extensive growth [ 4 ]. This would result in a drop in egg production. Therefore it is generally believed that eggs are mostly contaminated in the egg white, in which antibacterial factors are present that limit bacterial multiplication.
Until now, it has not been clearly explained why the serotype Enteritidis specifically has been implicated in egg contamination. It is clear however that Salmonella Enteritidis can contaminate the reproductive tract and seems to be a better colonizer of the oviduct environment as compared to other serotypes [ 5 ]. Secondly, it has been shown by comparing different strains from multiple serotypes that in general, Salmonella Enteritidis strains cope better with the antimicrobial properties of egg white as compared to other serotypes [ 6 ]. More efficient oviduct colonization and better survival in egg white are ideal characteristics for a pathogen that is transmitted to egg-consuming hosts. Despite this knowledge, the molecular mechanisms behind this behavior of Salmonella Enteritidis are still not unraveled. Some Salmonella genes that are involved in oviduct colonization and egg white survival have been identified, but genome-wide screening tools, complemented with in-depth studies on the role of specific genes, would be most welcome to get a better profile of the interaction of Salmonella Enteritidis with the oviduct tissue and egg white. Genes shown to be highly expressed in the oviduct or essential for oviduct colonization that are described until now are mostly major virulence genes that contribute to colonization of any organ, and are not exclusively related to oviduct colonization [ 3 ]. In the egg white, numerous antimicrobial compounds have been identified. These include molecules that degrade microbial components, such as lysozyme, antibacterial peptides, such as avian β-defensins, lipopolysaccharide binding and bactericidal/permeability increasing proteins (LBP-BPI proteins) that bind LPS and permeabilize the cytoplasmic membrane, such as ovocalyxin-36, molecules decreasing bioavailability of cations and vitamins, such as ovotransferrin, and many others [ 7 - 9 ]. Strikingly, the activity of these molecules to Salmonella and the potential protection mechanisms of Salmonella against these molecules are very poorly investigated under the conditions encountered by the bacteria it the chicken reproductive tract.
In this paper, it is proposed that stress-induced survival mechanisms enable the serotype Enteritidis to persistently colonize the oviduct without inducing damage and excessive inflammation, and to cope with the antimicrobial compounds and thus survive in egg white. It is proposed that studies should be carried out to a) analyze the virulence mechanisms that allow Salmonella Enteritidis to persistently colonize the oviduct tissue and b) to identify egg white molecules that have antibacterial activity against Salmonella and unravel the mechanisms that protect Salmonella Enteritidis against these environmental insults, thus enabling its survival in egg white. | Background
Egg-associated transmission to humans seems to be characteristic of the Salmonella serotype Enteritidis, explaining why this particular serotype has caused a worldwide pandemic since the mid '80s. Salmonella Enteritidis is much more capable to persistently colonize the laying hen reproductive tract and to survive in the hostile egg white, as compared to other serotypes.
Presentation of the hypothesis
It is hypothesized that stress-induced survival mechanisms enable the serotype Enteritidis to persistently colonize the oviduct without causing damage and excessive inflammation, and to cope with the antimicrobial compounds present in egg white.
Testing the hypothesis
To test the hypothesis first of all Salmonella Enteritidis genes that are essential for colonization of the oviduct and survival in eggs need to be identified. Comparative genomics tools should be used to identify genes or pathogenicity islands that are present in Salmonella Enteritidis and not in the multiple non egg-contaminating serotypes. High-throughput signature-tagged-mutagenesis approaches, coupled to micro-array detection of the genes that lead to an attenuated phenotype when mutated is proposed as an ideal tool to identify genes involved in oviduct colonization and egg white survival. Identifying the stressors and antibacterial molecules in the oviduct and in the egg white that limit colonization or survival of non-Enteritidis serotypes is a second important objective that can theoretically be achieved using screenings of expressed oviduct cDNA libraries for their antibacterial activity against strains from multiple serotypes. Finally, the effect of contact with these stressors in the oviduct or egg white on Salmonella gene expression will need to be analyzed, in order to clarify whether serotype Enteritidis-specific regulation of certain stress-survival pathways are either or not present.
Implications of the hypothesis
Knowledge on the pathogenesis of egg infections would furthermore give insights that might be extrapolated to other biological interactions, in which a highly specialized bacterial pathogen resists the host response in a specific biological niche. In addition, this info can be of value in developing early warning criteria to identify emerging egg-associated Salmonella strains and in developing safe live attenuated vaccine strains. | Presentation of the hypothesis
The Salmonella serotype Enteritidis has caused a worldwide pandemic due to its ability to persistently colonize the oviduct tissue of laying hens and survive in the hostile egg white environment. The Salmonella Enteritidis O- and H-antigens, and thus the LPS and flagellar antigens, are discriminating Enteritidis from many other serotypes (from other serogroups). It could thus very well be that the LPS structure of Salmonella serogroup D, containing the serotype Enteritidis, could be more resistant to penetration of antibacterial molecules, such as cationic antimicrobial peptides (CAMPs) present in egg white. However, this does not explain why strains from serotype Enteritidis and not other serogroup D serotypes have a tropism for eggs. It is hypothesized that stress-induced survival mechanisms play a role in the ability to persistently colonize the oviduct and to cope with the antibacterial molecules present in egg white. In other words, it is proposed that the serotype Enteritidis either contains Enteritidis specific genes or a specific regulation of stress-induced genes that enable these strains to live in close association with the laying hen oviduct and the egg white. As an example, successful strains could colonize the oviduct without causing overt disease, inflammation and tissue damage, and thus without a drop in egg production, and be incorporated in eggs during passage in the oviduct. Once inside the egg, these strains could respond to the environmental stresses they encounter by hitherto undefined stress-induced protection mechanisms, such as cell wall modifications or mechanisms that are used to export antibacterial compounds.
Testing the hypothesis
Testing of the hypothesis would require a multidisciplinary approach that, among others, should try to reply to the following questions: 1) which Salmonella Enteritidis genes are essential for colonization of the oviduct and survival in eggs?; 2) what are the stressors that Salmonella encounters in the oviduct and in the egg white and limit colonization or survival of other serotypes?; 3) what is the effect of contact with these stressors in the oviduct or egg white on Salmonella gene expression and does this transcriptome contain Salmonella Enteritidis specific gene responses or a Salmonella Enteritidis specific regulation of certain stress-survival pathways?
To identify Salmonella Enteritidis genes essential for oviduct colonization and egg white survival, different approaches, including many more than those proposed in this paper, can be used. First of all, one could analyze whether gene sequences that are present in Enteritidis but not in other serotypes exist and whether the encoded proteins have a role in the interactions with eggs or with the oviduct (cells). This would clearly require a comparative genomics approach using multiple serotype Enteritidis strains and strains from a variety of other serotypes. Earlier studies have already shown that no consistent large genomic differences exist between recent Salmonella Enteritidis isolates and isolated from the 1940's and 50's [ 10 ]. When comparing Salmonella Enteritidis with Salmonella Typhimurium and Salmonella Gallinarum, genes and genome islands were identified that were specific for the strains from the respective serotypes [ 11 ]. It has been shown that Salmonella Gallinarum is most likely a descendant of Salmonella Enteritidis, because both genomes are very similar with the exception that Salmonella Gallinarum has lost multiple genes and genomic islands during evolution [ 11 ]. This is of extreme interest because Salmonella Gallinarum is adapted to laying hens and, in contrast to Enteritidis, is not colonizing the gut to large extent, but colonizing and damaging the reproductive tract and other organs severely, resulting in clinical symptoms and death [ 12 ]. Searching for the function of the genes that are lost during Salmonella Gallinarum divergence would thus possibly give insights in the pathogenesis of egg infections, and result in the identification of genes that cause a less destructive phenotype, what would help in colonizing the reproductive tract without causing drops in egg production. Although the genomic comparisons with serotype Typhimurium strains can be interesting, this serotype has, although on a limited scale, also been implicated in egg infections and can clearly also colonize the oviduct of laying hens. It would however be of interest to compare the Salmonella Enteritidis genome with the genome of strains that are known not to persistently colonize the oviduct or contaminate eggs, such as Virchow, Hadar, other serogroup D strains, and multiple others. This may help to identify Salmonella Enteritidis specific genes that could potentially be involved in the adaptation to the oviduct and egg environment. Another way to identify genes involved in oviduct colonization or egg white survival could be the use of signature-tagged mutagenesis (STM)-related approaches, such as the microarray-based negative selection strategy that has been used by different authors to identify genes involved in colonization of mouse organs [ 13 , 14 ]. This method is based on the inoculation of a transposon-mutagenized library of Salmonella in a live animal, and a micro-array based detection of all transposon-inserted sequences in the Salmonella isolates recovered from the organ. Selective disappearance of mutants will thus yield a list of genes that are essential to colonize an organ. This method could clearly be of use to identify Salmonella Enteritidis genes involved in chicken oviduct colonization. Whatever method is used, it is clear that analyzing the behavior of deletion mutants in the genes that are identified by genomic comparisons or STM-related approaches as being potentially involved in oviduct colonization or egg white survival, will give the final proof for its actual role in these processes.
It is clear that the oviduct is a harmful environment for Salmonella . Indeed, Salmonella strains are colonizing any other organ, including the gut, liver and spleen, much better than the oviduct [ 5 , 15 , 16 ]. Moreover, metabolic mutants can colonize the gut and internal organs much longer than they can colonize the oviduct [ 17 ]. Egg white, or oviduct mucus, is clearly antibacterial and many antibacterial molecules from egg white have been described, but the anti- Salmonella activity of these molecules is under-investigated. It has been shown that Salmonella Enteritidis if far more capable to survive in egg white compared to strains from other serotypes [ 6 ], but the actual reason is unclear. It does not seem to relate only to the structural differences in LPS structure, because strains from the same serogroup D also survive less. Knowledge on the mechanisms by which egg white reduces the Salmonella load in eggs and more important, data on the specific protection mechanisms of Salmonella Enteritidis against these egg white stressors, could enhance our understanding of the actual reasons for the egg-derived pandemic. The differential response of Salmonella serotypes and strains to certain egg white specific factors, such as alkaline pH (8 to 9), can easily be tested, but it is reasonable to believe that one or a subset of egg white components are involved in the bacteriostatic or bactericidal activity against Salmonella . Isolating all these components and testing them one by one would be an impossible task seen the complexity of egg white. Although a challenge, it would be an option to test all egg white proteins at once after cloning an inducible oviduct cDNA library in Salmonella and testing viability after induction of expression. Chicken reproductive tract cDNA libraries have already been used to identify antibacterial eggshell proteins [ 18 , 19 ]. Recently, a high-throughput screening method was developed to identify antibacterial peptides [ 20 ]. The method uses inducible library expression in E. coli combined with a dye inclusion assay to monitor bacterial cell viability. It is thus suitable to identify antibacterial peptides or proteins. If this system would be converted to Salmonella and if an oviduct cDNA library could be properly expressed in Salmonella , this could open perspectives for antibacterial egg white protein identification. Of specific interest would be proteins that are antibacterial for non-Enteritidis strains, but not for Salmonella Enteritidis. Important in this regard in the recent finding that a Salmonella Enteritidis mutant in tolC, the outer membrane channel linked to multi-drug resistance (MDR) efflux pumps, is required to survive in egg white [ 21 ]. It could well be that Salmonella Enteritidis is capable to export specific antibacterial egg white molecules, while this trait is not present in other serotypes.
Identifying Salmonella Enteritidis specific gene sequences or Salmonella Enteritidis specific genes involved in oviduct colonization or egg white survival, and identifying antibacterial molecules in egg white that affect non-Enteritidis serotypes, if existing, would be a difficult but feasible approach. Even more complex would be that the selective advantage of Enteritidis would be due to differences in expression levels of Salmonella genes, present in any serotype, but differentially regulated by environmental triggers in the different serotypes. Measuring the expression profile of the Salmonella genome in the oviduct or eggs is possible using micro-array systems. Systems such as the in vivo expression technology (IVET) have been used already to identify genes highly expressed in the chicken oviduct [ 17 ]. Although measuring expression in the oviduct or eggs is feasible, the challenge would be to evaluate the significance of this upregulation. If, in an ideal situation, anti- Salmonella molecules present in the oviduct or egg white would be identified that affect most, if not all, non-Enteritidis serotypes but not Enteritidis strains, then it would be fairly easy to study the transcriptome and analyze the response to these stressors.
Implications of the hypothesis
It is astonishing that 30 years after the rise of the egg-associated pandemic caused by Salmonella Enteritidis, still the actual reasons for the close interaction of this serotype with eggs are not clarified. Methodologies now are in place to perform detailed studies on the molecular interactions between Salmonella strains, the oviduct tissue and egg white molecules. It is of utmost importance to understand the biological differences between the (un)ability of Salmonella strains and serotypes that are able vs those that are unable to colonize the laying hen oviduct and contaminate eggs. Published and unpublished data (from our own group and collaborators) point towards an interaction in which Salmonella Enteritidis is more capable to withstand the harmful environment of the oviduct and the attacks by antibacterial compounds present in egg white. It is hypothesized that this could be caused by stress-induced activation of a protection mechanism, what can be investigated using currently available methodologies. The implications of this hypothesis are multiple. First of all, identifying Salmonella Enteritidis specific mechanisms for persistent oviduct colonization and egg white survival would possibly aid in a more rapid identification of serotypes or strains that, in the future, would emerge as egg-associated strains. Thus a diagnostic tool to screen Salmonella strains for the ability to cause egg infections could be developed. Secondly, a potential application lies in genetic selection for laying hens that are capable to produce lethal concentrations of specific protective molecules in the egg white. It could well be that concentration dependant killing by an egg white molecule is responsible for killing of most Salmonella serotypes with the exception of Salmonella Enteritidis. In this case genetic selection could be used to select for a breed with higher concentrations of the antibacterial compound in the egg white. In addition, knowledge on genes essential to contaminate eggs is crucial for development of vaccines for laying hens. Genetically modified live attenuated vaccines could thus be developed without concerns that they would potentially contaminate eggs used for consumption. Knowledge on the pathogenesis of egg infections would furthermore give insights that might be extrapolated to other biological interactions, in which a highly specialized bacterial pathogen resists the host response in a specific biological niche.
List of abbreviations
None.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
Prof. F. Van Immerseel is research professor (BOF-ZAP) and coordinating the activities of a research group working on the pathogenesis of egg infections by Salmonella and wrote the hypothesis paper. | Acknowledgements
This hypothesis paper was only possible after understanding the basic elements of the pathogenesis of egg contamination by Salmonella , and thus the work of Jeroen De Buck, Inne Gantois, Ruth Raspoet and Rosalie Devloo, and the continuous discussions with other members of the team, especially Richard Ducatelle, is highly appreciated, as well as the work of numerous other researchers worldwide that increases our understanding on egg infections. Especially the collaboration with the Molecular Microbiology Group of the Institute of Food Research in Norwich (UK) has increased our knowledge on the molecular aspects of the interaction between Salmonella and the oviduct/egg. | CC BY | no | 2022-01-12 15:21:44 | Gut Pathog. 2010 Dec 20; 2:23 | oa_package/90/ba/PMC3016255.tar.gz |
||||
PMC3016256 | 21172038 | Background
The intestinal microbiota plays a critical role in the establishment and maintenance of healthy immune responses. Delayed colonisation of the infant gut with commensal bacteria or alterations in the microbiota profile are suggested to be strong risk factors for the development of immune-mediated chronic disorders such as allergic and autoimmune diseases. Mice raised in a germ- free environment fail to develop oral tolerance and have persistent Th2-dependent immune responses [ 1 ]. This immune deviation can be corrected by reconstituting the microbiota with a single bacteria species, but only if this occurs during the neonatal period. Similarly, infants with allergic disease have reduced numbers of beneficial Bifidobacteria species and increased numbers of pathogenic Clostridia and Staphylococci compared to non-allergic infants [ 2 - 4 ]. Moreover, these changes occur prior to the onset of allergy, suggesting a causal relationship between microbiota and healthy immune responses. These observations have led to the idea that probiotic bacteria - which have the potential to restore the intestinal microbiota balance - may be effective in preventing the development of chronic immune-mediated diseases.
Probiotics mediate their activity by a variety of mechanisms including altering the microbiota composition, maintaining epithelial barrier function and modulating mucosal and systemic immune responses [ 5 ]. Importantly, the effects of probiotic bacteria are species and strain-specific, so it is imperative to select probiotic bacteria with specific activities based upon known in vitro or in vivo effects that are relevant to the clinical context they will be applied to. Lactobacillus and Bifidobacteria species have been studied in greatest detail and are the most commonly used in research studies [ 6 ]. In particular, significant effort has been made in the examination of probiotic therapy for the prevention of allergic disease. Several clinical trials have demonstrated that a combined prenatal/early postnatal treatment has the greatest impact on reducing allergy symptoms, indicating the importance of early life interventions [ 7 - 9 ]. Prenatal supplementation with the probiotic Lactobacillus rhamnosus GG (LGG) was found to reduce the development of atopic dermatitis in high-risk infants by 2 years of age [ 7 ]. However, subsequent studies using the same or different probiotics have not confirmed these results, suggesting that there may be intrinsic differences between the probiotic strains used and study populations, reflecting the complexity of these studies [ 10 ].
Treatment with probiotics induces a variety of immunological effects on epithelial function, dendritic cells, Treg and T-helper responses. Examples of these effects include 1) enhanced epithelial barrier function via interactions with Toll-like receptors (TLRs) and modulation of epithelial cell signal transduction pathways that regulate cytokine production to promote anti-inflammatory responses [ 11 - 13 ]; 2) induction of tolerogenic DCs producing low IFNγ and elevated IL-10 [ 14 ]; 3) induction of Treg activity associated with increased TGF-β and IL-10 secretion by PBMCs [ 15 , 16 ]; 4) modulation of T helper responses [ 17 ]; 5) stimulation of IgA responses to oral and parenteral vaccines [ 18 - 20 ] and 6) modulation of immune factors within breast milk such as TGF-β, soluble CD14 and total IgA [ 14 , 21 ]. Genetic analysis of probiotic effects following treatment of healthy adults with L. casei and L. rhamnosus revealed upregulation of several key mucosal immune response genes encoding IFN-γ production and TLR3/9 expression while L. acidophilus was found to upregulate immunoregulatory genes including IL-17B, IRAK2 as well as several chemokines and cellular adhesion molecules in duodenal biopsies using a novel transcriptome analysis [ 22 ].
An important role for commensal (including probiotic) bacteria is the fermentation of dietary compounds leading to production of short-chain fatty acid (SCFA) metabolites. Several SCFAs are known to be produced such as butyrate, acetate and propionate. Levels of SCFA have been demonstrated to be reduced in inflammatory diseases, possibly as a result of an altered gut microbiota. These SCFAs have an important protective function in the intestine and exert anti-inflammatory effects in a number of animal models. Intestinal epithelial cells of germ-free mice had reduced SCFA receptor expression which was normalized following reconstitution with normal gut bacteria [ 23 ]. The SCFA, butyrate, also has an important role as an energy source for the mucosa in predominantly anaerobic environments such as the gastrointestinal tract as well as having other critical biological functions such as immune regulation [ 24 ]. Moreover, butyrate improves epithelial barrier integrity by modulating the expression of certain tight junction proteins such as cingulin, ZO proteins and occludin [ 25 , 26 ]. These combined effects promote establishment of a healthy microbiota and the development of a healthy gastrointestinal lymphoid tissue. Studies involving SCFA treatment using butyrate and acetate have yielded promising results, with amelioration of inflammatory lesions in mouse models of allergic airways disease and colitis [ 27 ]. This has led to the idea that probiotic supplementation could restore SCFA levels by modulating the microbial environment, representing a novel therapeutic or preventative strategy for chronic inflammatory diseases.
Probiotic metabolites such as butyrate are therefore an important class of therapeutic compounds. These SCFAs are also a well-known class of epigenetic drugs known as histone deacetylase inhibitors (HDACi) that have a central role as anti-cancer agents with strong anti-proliferative effects on tumour cells [ 28 , 29 ]. This activity of SCFAs may explain the anti-inflammatory effects observed for probiotic bacteria. Intriguingly, the principal HDACi compound is butyric acid (sodium butyrate) which inhibits most HDAC enzymes except class III and class II HDACs 6 and 10 [ 30 ]. Butyrate is one of the most potent HDACi in human colon cancer cell lines suggesting an integral role as anti-inflammatory derivatives of microbial fermentation in the colon [ 24 ]. Studies using fecal fermentation supernatants were found to be rich in butyrate and exhibited strong HDAC inhibitory properties in several colon cancer cell lines [ 31 ]. The HDACi activity of butyrate and propionate was associated with blockade of DC development primarily through the Na + coupled monocarboxylate transporter Slc5a8 [ 32 ].
Given the role of bacterial species in the production of SCFAs, probiotics may be considered as an alternative approach for the prevention or treatment of chronic inflammatory diseases. Evidence in support of this effect by probiotics is limited. The butyric acid-producing anaerobic bacterium, Faecalibacterium prausnitzii , is a novel probiotic used for the treatment of inflammatory bowel disease (IBD) [ 33 ]. The levels of this bacterium is lower in IBD and treatment of mice with F. prausnitzii led to a shift in microbiota composition, reduced inflammatory cytokine levels such as IL-2, increased the anti-inflammatory cytokine IL-10 and reduced colitis and mortality, suggesting that butyrate and other related metabolites may be critical in host protection [ 34 ]. The anti-inflammatory effects of SCFA-producing probiotic bacterium was further illustrated with Propionibacterium freudenreichii , which was found to shift the extracellular pH from 7.5 to 5.5 in the colon and was related to levels of the SCFAs, acetate and propionate [ 35 ]. It is possible that probiotic-derived SCFA metabolites such as butyrate produce HDACi effects that have multiple downstream effector functions on many target cells, including those of the immune system. This offers a promising intervention approach for the treatment of inflammatory conditions such as allergy and gastrointestinal disease. The exact mechanism(s) of action for probiotic bacteria have yet to be fully understood; therefore the epigenomic-modifying capacity of probiotics will be important in understanding how they mediate their health-promoting effects. | Background
The intestinal microbiota plays an important role in immune development and homeostasis. A disturbed microbiota during early infancy is associated with an increased risk of developing inflammatory and allergic diseases later in life. The mechanisms underlying these effects are poorly understood but are likely to involve alterations in microbial production of fermentation-derived metabolites, which have potent immune modulating properties and are required for maintenance of healthy mucosal immune responses. Probiotics are beneficial bacteria that have the capacity to alter the composition of bacterial species in the intestine that can in turn influence the production of fermentation-derived metabolites. Principal among these metabolites are the short-chain fatty acids butyrate and acetate that have potent anti-inflammatory activities important in regulating immune function at the intestinal mucosal surface. Therefore strategies aimed at restoring the microbiota profile may be effective in the prevention or treatment of allergic and inflammatory diseases.
Presentation of the hypothesis
Probiotic bacteria have diverse effects including altering microbiota composition, regulating epithelial cell barrier function and modulating of immune responses. The precise molecular mechanisms mediating these probiotic effects are not well understood. Short-chain fatty acids such as butyrate are a class of histone deacetylase inhibitors important in the epigenetic control of host cell responses. It is hypothesized that the biological function of probiotics may be a result of epigenetic modifications that may explain the wide range of effects observed. Studies delineating the effects of probiotics on short-chain fatty acid production and the epigenetic actions of short-chain fatty acids will assist in understanding the association between microbiota and allergic or autoimmune disorders.
Testing the hypothesis
We propose that treatment with specific probiotic bacteria under in vivo conditions would offer the ideal conditions to examine the microbiological, immunological and epigenetic mechanisms of action. Advances in epigenetic technology now allow investigators to better understand the complex biological properties of probiotics and their metabolites.
Implications of the hypothesis
Determining the precise mechanisms of probiotic action will lead to more specific and efficacious therapeutic strategies in the prevention or treatment of chronic inflammatory conditions. | Presentation of the hypothesis
Probiotics mediate pleiotropic effects however the precise mechanism(s) have yet to be fully elucidated. The 'epigenome targeting' hypothesis contends that the activity of probiotics and their metabolites are associated with chromatin remodeling as a function of intrinsic histone deacetylase inhibition. Modulation of gene transcription is an essential component of many biological processes. The onset of pathological conditions such as cancer and chronic inflammation often result from aberrant gene transcription. Integral to this process are epigenetic factors involving two critical enzymes, histone acetylases (HATs) and histone deacetylases (HDACs). These produce post-translational modifications of histone proteins and result in changes to chromatin structure and function [ 36 , 37 ]. While HATs serve to acetylate histones, conferring a 'relaxed' chromatin structure that allows transcriptional activation, HDACs has the opposite effect. HDAC enzymes repress transcription through tightening of the chromatin structure, excluding accessibility of transcription factors and other regulatory proteins to DNA and therefore the ability to influence gene expression [ 38 ].
The recent advances in understanding epigenetic processes in the maintenance of health allows us to investigate more precisely the mechanisms of action of novel therapies such as probiotics in defined clinical contexts. Critical transcriptional regulators of inflammation such as NF-κB and Foxp3 can now be studied closely to determine the impact of dietary/probiotic supplementation. The evidence presented above and our hypothesis provide for the first time a direct link between the anti-inflammatory and immunoregulatory properties observed for probiotics in terms of Treg induction, DC maturation and cytokine/chemokine secretion and the molecular events involved at the epigenetic level. We believe that the biological activity of probiotics are mediated through complex epigenetic changes that regulate the activation status of key transcription factors involved in host immunity.
Testing the hypothesis
Testing this hypothesis should be undertaken in an in vivo system such as the mouse, or preferably in humans using a randomized, double-blind, placebo-controlled study design. Probiotic bacteria or placebo should be given prenatally and/or post-natally to infants at high-risk of developing allergic disease. Clinical effects of the probiotic should be assessed by doctor-diagnosed allergy using defined criteria, skin prick test positivity, and SCORAD analysis of eczema severity. Immunological outcomes can be assessed by enumerating Treg and DC populations and the profile of cytokine secretion by cultured cord and peripheral blood mononuclear cells in response to SCFA metabolites. Levels of SCFA can be measured in stool or blood. Correlation between immunological and epigenetic changes could be assessed in the first instance by simple immunoblot analysis for histone hyperacetylation as well as real-time PCR for specific immune gene regulation in relevant cells. Genome-wide sequencing techniques such as mRNA-Seq (gene expression) and ChIP-Seq (specific epigenetic modifications including acetylation) will further our understanding of the critical biological functions of probiotics.
Implications of the hypothesis
Interest in probiotics has intensified in recent years due to their potential health benefits and as treatments for a wide range of inflammatory and immunological conditions. Probiotic bacteria have been demonstrated to modulate a variety of microbiological and immunological parameters although the precise mechanisms for these effects are yet to be fully elucidated. The ability of probiotics to alter intestinal microbiota composition and diversity suggest an important role for SCFAs such as butyrate. The demonstration of beneficial probiotic effects through as yet unidentified epigenetic mechanisms will have significant implications in the prevention and/or treatment of allergy and other chronic inflammatory diseases.
Competing interests
PVL, SSW and TCK declare that they have no competing interests. MLKT is a member of the medical advisory board of Nestle Nutrition Institute and has delivered speaker presentations for Wyeth/Pfizer.
Authors' contributions
PVL conceived and wrote the paper, SSW contributed to the manuscript, MLKT and TCK contributed to and critically appraised the manuscript. All authors read and approved the final manuscript. | CC BY | no | 2022-01-12 15:21:44 | Gut Pathog. 2010 Dec 21; 2:24 | oa_package/8b/9a/PMC3016256.tar.gz |
|||||
PMC3016257 | 21172001 | Background
Carers
Approximately six million (approximately one in ten) of the population of England and Wales are carers [ 1 ] with three in five people becoming carers at some point and numbers are expected to increase with the ageing population and greater numbers of severely disabled people living longer [ 2 ].
A carer (also sometimes known as informal carer or caregiver) is defined as:
'.... someone who, without payment, provides help and support to a partner, child, relative, friend or neighbour, who could not manage without their help. This could be due to age, physical or mental illness, addiction or disability. The term carer should not be confused with a care worker, or care assistant, who receives payment for looking after someone.' [3,p9).
Over half of these carers are women (58%) and approximately 175,000 of them are children. Nine hundred thousand carers care for over 50 hours a week [ 2 ].
It is now accepted that carers often suffer poorer mental and physical health as a result of their role [ 4 ]. For example, amongst carers of people with long-term conditions such as dementia and stroke, negative outcomes include emotional distress, reduced quality of life, burden and stress [ 5 - 7 ]. There is also some evidence that providing unpaid or informal care for a long time is associated with deterioration in both physical and psychological health [ 8 ].
It has been estimated that carers save the economy £87 billion annually [ 9 ]. The vital role played by these carers is now recognised by the government which has pledged to recognise and support carers [ 10 , 11 ]. The numbers of young carers has also been acknowledged [ 12 ].
General practice and carers
General practice can play a significant role in supporting carers. According to a joint report by the Royal College of General Practitioners (RCGP) and the Princess Royal Trust for Carers (PRTC):
'GPs and their teams are usually the first place that carers have contact with the National Health Service. They are uniquely placed to recognise that someone is, or is about to become, a carer.' [3,p2).
Research has also shown that carers believe that GPs are both well placed to support them and have the power to improve the quality of carers' lives [ 13 ]. It might therefore be expected that carers would visit their GPs more than the population as a whole, although the evidence here is inconclusive, possibly in part because of methodological issues [ 14 ].
Despite this there is little recent published UK research in this area. A postal survey nearly a decade ago (Simon and Kendrick 2001) [ 15 ] found that GPs and primary care teams thought they had a key, pro-active role to play supporting carers but that they lacked time, resources and training.
Research in Australia [ 16 ] focussing on GPs' perceptions of carer emotional needs highlighted that although GPs were aware of carers' increased emotional needs, the services they offered were almost exclusively 'practical '(including referring and directing carers to services to ensure the carer had practical assistance). GPs preferred to refer carers to community services although some did offer counselling described as 'informal sharing or 'coffee cup counselling'' (p4).
GP contracts in the UK currently give three points in the Quality and Outcomes Framework (QOF) for the establishment of a system to identify and refer carers to local authorities for assessment of their needs. The Quality and Outcomes Framework (QOF) rewards GP practices for how well they care for their patients and helps fund improvements in the care they deliver. It is based on performance against specified indicators or measures of achievement. Each indicator is worth a maximum number of points and GP practices are rewarded financially according to how many points they achieve. Primary Care Trusts also have a responsibility for establishing clinical governance protocols for the delivery of services by GPs which should include addressing and supporting the role of carers [ 17 ].
Aims
Using a self-completion questionnaire this study aimed to investigate GPs': attitudes to carers; awareness and knowledge of issues facing carers; perceived barriers to supporting carers and services offered to carers.
The study
The Department of Health (DH) in partnership with the Royal College of General Practitioners (RCGP) commissioned a pilot workshop training programme for GPs and other members of primary care teams across England to learn about carers. The findings reported here were gathered from questionnaires completed by GP participants prior to the workshops and was part of the programme's evaluation. | Methods
Setting and method
Six pilot workshops were held in the latter half of 2009 in six general practice Faculties (Bedfordshire and Hertfordshire; North West London; North West England; Vale of Trent; Yorkshire; Wessex). The aim of the workshops was to give participants a better understanding of the problems facing carers and the role that primary care might take in supporting them. Workshops were advertised by the RCGP and invitations were sent out from the relevant general practice Faculty. Where appropriate, workshop participants received credit for their Continuing Professional Development.
All participants were asked to complete the questionnaire when registering for the workshops. Questionnaires were distributed electronically prior to each workshop. Where participants had not completed it earlier, they were again asked to do so on arrival. Although representatives from a range of professions attended the workshops, only responses from the GPs are included here because of the small numbers representing other professional groups.
Questionnaire
The questionnaire was designed specifically for the study. Topics covered three sections: participant demographics (e.g. age, gender); practice background information (e.g. practice size, staff employed and location); attitudes towards carers; knowledge of issues facing carers and carer specific services offered by participants' practices. Questions were derived from literature focusing on carers and general practice and ideas raised in the 'Supporting Carers: An Action Guide for General Practitioners and Their Teams' (PRTC and RCGP 2008) [ 3 ]. Questions were phrased both positively and negatively to reduce the chances of response bias where participants may tend to agree rather than disagree with statements, a behaviour sometimes referred to as 'yeah saying' [ 18 ].
The questionnaire included both open-ended and closed questions using Likert scales and 'Yes/No' responses. This method was expected to capture a more complete picture of GP perspectives than using only one approach. The quantitative element provides numerical results whilst the open-ended questions gave the GPs the opportunity to articulate personal perspectives and describe experiences and issues not covered by the Likert scales.
Likert scales measured agreement with statements about, for example, attitudes towards carers and services offered. Participants could choose from 'strongly agree', 'agree', 'neither agree nor disagree', 'disagree', 'strongly disagree' and' don't know'.' For the purposes of the analysis responses were collapsed into 'agree', 'neither agree nor disagree', 'disagree' and 'don't know'. Where appropriate, findings (such as practice details) were subjected to descriptive statistics.
Open-ended questions concerned, for example, social and health issues facing carers and these responses were content analysed.
Questionnaires were piloted with five GPs. The only issues that arose related to the questionnaire layout. However, after the first batch of questionnaires was sent to participants, some objected to being required to specify their age. The questionnaire was therefore altered so participants described their age in ten year categories. Responses from the early questionnaires were retrospectively assigned to these categories.
Ethics approval
When contacted, the National Research Ethics Service advised that this study did not require ethical review because it was an educational evaluation, not research. It therefore did not require ethical approval. However the proposal was looked at by the local research ethics committee and the study was conducted following ethical principles such as informed consent and respect for confidentiality. | Results
Participants, reasons for attending and practice characteristics (Table 1 )
Questionnaires were completed by 78 of the 95 GPs attending the workshops (82% response rate). Three-quarters were female (76%) and over half (54%) were aged over 40 years.
Nearly a third (29%) had received previous training in issues in supporting carers. Approximately a third of these said it had been as a student and for a similar proportion training had been during their practice.
GPs were asked an open-ended question about why they attended the workshop and the main reasons they gave were to increase their own (38%) or their practices' (10%) knowledge. For example, they said they hoped to improve their own knowledge in issues facing carers and how best to identify them. Some explicitly mentioned they were attending with the intention of sharing the information gained with others in their practice. A quarter said they hoped to be able to improve the provision of services for carers (24%).
Participants came from varying sized practices. Numbers of GPs working in each practice ranged from two to 15 (median 6). Numbers of GP trainees per practice varied from none to six (median 1).
Patient list sizes ranged from 1700 to 18000 (median 8000). Two thirds of practices were identified as urban (65%), one in ten semi-rural (10%) and one in ten rural (10%).
The role of general practice in supporting carers, attitudes to carers and issues facing carers
Responses to Likert scales (Table 2 )
Responses to the Likert scales showed that GPs lacked confidence in their role (only 11% said they were confident) and the majority (89%) felt insufficiently trained in supporting carers. Less than half (45%) were confident that they could identify carers in their practice. A few (9%) regarded carers as 'sometimes a barrier in managing the healthcare of the cared-for person' and maintaining confidentiality of the care recipient was recognised as difficult by the majority of these GPs (92%).
Most of these GPs believe that general practice can play a role in supporting carers (77%) and that they should be pro-active in this role (93%). Approximately half said that they already take an active role here (53%). GPs said carers deserve more support (85%) although the majority (86%) agreed that 'supporting carers can be difficult'. Nevertheless more than eight in ten (84%) disagreed with: 'There is little point in referring carers to support services as they are unlikely to use them'. The majority (84%) agreed that carers 'should be a partner in the health care of their cared-for person'.
In terms of knowledge, awareness that carers are more likely to suffer from emotional problems was high. Three-quarters (75%) noted that they were aware of this. Two-thirds of GPs (64%) agreed that carers' all-cause mortality rate is increased whilst a fifth (21%) said they did not know.
Responses to open-ended questions
Health and social issues looked out for amongst carers (Tables 3 and 4 )
When asked about social issues amongst carers that they would look out for, approximately a quarter of GPs (27%) provided no answer. Amongst those that responded, the problems most frequently mentioned were financial (38%), isolation (32%) and mental health (21%). Just over one in ten (12%) mentioned each of the following: time for self, employment and support (Table 3 ).
Mental health issues were most frequently noted as carers' health issues. Some mental health problems were mentioned specifically (depression 51%, stress 12%, anxiety 10%) whilst some participants only referred to 'mental health' in general (18%). Neglect of carers' own health was also highlighted (24%) (Table 3 ).
When asked specifically about issues for young carers (Table 4 ) a range of responses were given. Nearly half of the GPs highlighted problems with education including missed schooling and poor academic performance (47%). Isolation was also commonly mentioned (40%). Other concerns included relationship problems with families or friends (18%) and depression (12%).
Services for carers and barriers to providing support (Tables 5 and 6 )
Few participants responded to the open-ended question concerning services offered specifically for carers. Three-quarters (73%) gave no response. Six percent (five GPs) already had a carers' register and the same number offered flu vaccinations, although they did not say if they were specifically offered to carers. Four GPs (5%) had a carers' pack or provided leaflets.
Approximately a quarter of GPs thought carers wanted emotional support (27%) and signposting or referral to relevant agencies (22%). Participants (19%) also thought carers would like information including advice about services (Table 5 ).
Over a quarter of GPs (28%) did not identify any barriers in supporting carers (Table 6 ). Those mentioning obstacles highlighted insufficient time (44%) and resources (23%). Lack of knowledge about carers was an issue for nearly one in five participants (18%). Difficulty in identifying carers was raised by one in ten. | Discussion
Against the backdrop of the likely future increase in numbers of carers because of an ageing population and more people living with improved healthcare so that more people live with disability for longer [ 2 ], it is important that general practice and GPs are in a position to support them. This study has demonstrated that although GPs regard general practice as an appropriate place to do this and think they should be pro-active in their role, they frequently lack confidence and training and sometimes knowledge to do it effectively. Key issues highlighted include the identification of carers - some GPs are aware that they are not identifying all carers in their practices and would like to be guided on how best to do this. Few services are currently being offered by general practices specifically for carers despite GPs' belief that carers want their support. The few carers' registers reported here by the GPs is all the more surprising given the financial incentives to have them.
There are several possible reasons for this. GPs may believe that as carers have no specific medical diagnosis and have social not medical problems, supporting them is perhaps only on the periphery of their role [ 19 ]. They may also think identification and support of carers may add to an already demanding workload and have concerns that carers may present problems they are unable to help with. However, the GPs in this study agree that carers deserve their support even if insufficient time and resources and concerns about confidentiality remain. Clearly the fact that carers and the person being cared for may not have the same GP or even attend the same practice complicates both identifying carers and may make communication more difficult.
There are noticeable differences between our findings and those of Simon and Kendrick (2001)[ 15 ]. Variations in question wording (such as open-ended questions versus checklists in Simon and Kendrick's study) and our sample's attendance at a carers workshop account for some of this, but the differences are still striking. On a positive note, the vast majority of our sample (90%) compared with a quarter in the earlier study said that they should be proactive in supporting carers. Approximately a third of our participants (29%) said they had received some previous training in carers compared with 10% nearly a decade ago. On the other hand, five GPs here (6%) reported keeping a carers' register. In Simon and Kendrick's study, a quarter recorded carer status. Although these activities are not identical, this suggests that identifying carers may remain a low priority. Also nearly half the GPs in Simon and Kendrick's study offered information to carers, compared with a handful here.
There are similarities with our findings and the research from Australia [ 16 ]. In both there is an apparent gap between what GPs think carers might like and what they offer them. Despite awareness of carers' emotional needs, GPs in Australia tended to refer carers to practical, rather than emotional support. Similarly here although a quarter of the GPs thought carers would like emotional support, the only service provided that might be directly expected to offer such support was carer support groups available at three practices. GPs here described fewer services for carers than in the Australian study but nearly all those mentioned here were also practical in their approach.
Nearly a decade ago it was suggested that primary care teams could support carers in a number of ways including acknowledging the problems they have and ensuring the general practice team are aware of them, flagging carers' notes so that GPs were aware of their caring role, acknowledging the role they play, treating them as team members and provision of information for the carer relating for example to the condition of the person the carer is looking after and information about benefits and services [ 19 ]. The fact that the DH and the RCGPs organised these workshops suggests greater awareness of the importance of carers and the significant role they play but our findings suggest that many of these earlier recommendations still stand. Although most GPs here did agree that carers should be a partner in the care of the recipient of care, little progress seems to have been made with the other recommendations.
However, there is very little recent published literature on the topic and it may be worth speculating why. Perhaps it is a symptom of lack of awareness of the major role played by carers or perhaps an indication of the uncertainly and maybe ambivalence amongst GPs about supporting this group.
Setting up carers registers is an important first step but if GPs are unaware of common problems amongst carers and feel they have insufficient knowledge and time to offer support, raising expectations may be detrimental. A possible approach here is increasing the numbers of carers' leads or champions in primary care teams [ 20 ]. Such carers champions are a member of the general practice staff who can recognising the needs and difficulties of carers and be able to offer them information and respond to their enquiries. The impact of this role has not been formally evaluated but they could assume much of the responsibility for supporting and signposting carers.
There remains work to be done. Repeating this survey with GPs who had not opted to attend training on carers would offer a wider perspective although the fact that the participants here had expressed an interest in the area but still lacked some knowledge and awareness of likely problems amongst carers (as suggested by the questions where they failed to answer or were unaware of some facts about carers) can also be seen to further the argument for the value of our findings. Better appreciation of GPs' perspectives and those of other primary care team members is required but research is also needed to identify what support carers need and want from primary care. Possibly adopting qualitative methods would allow a more in-depth understanding of all these perspectives. Research exploring the potential of roles such as carers' champions who might be well-placed to support carers whilst minimising the input from GPs would be beneficial.
Strengths and limitations
A strength of our study is that we achieved a high response rate for this group of participants [ 21 ] and the GP participants came from a range of areas geographically including both rural and urban populations and involved both large and small practices. The response rate was also high. Given the lack of research in the area, in many ways the study is exploratory since it investigated a wide range of issues many of which need further research.
A limitation is that the study participants had mostly chosen to attend a workshop on carers and may have been an atypical group motivated to support carers. It is impossible to know what impact this had on the findings but given the lack of research in the area, this is an important step in understanding GPs perceptions. Importantly these GPs were mostly attending the workshops because they wanted to know more about carers. There were a higher proportion of female GPs and they were slightly younger than might be expected [ 22 ] and nearly half said that they had an interest in carers and may therefore have been better informed compared with other GPs, although the current findings suggest they lack some knowledge in the area. | Conclusions
Without carers costs to the UK economy would increase hugely. GPs recognise general practice has an important role to play in supporting these carers but would appreciate both more training and support. Research is needed both to determine what support carers need and would like from general practice and to evaluate any new roles that are introduced. | Background
Approximately one in ten of the UK population are unpaid carers supporting a family member or friend who could not manage without their help, saving the UK economy an estimated £87 billion. This role is known to sometimes have a negative impact on carers and to require support both informally and from statutory services. General practice is a first point of contact for carers but research investigating general practitioners' (GPs') attitudes towards carers and awareness of issues facing carers is rare. This study therefore aimed to identify GPs' attitudes, awareness of issues, and perceptions of the barriers and enablers to provision of services.
Methods
Using a self-completion questionnaire distributed at a series of workshops, this study investigates GPs' attitudes to carers; awareness and knowledge of carers' issues; services offered in general practice and barriers to supporting carers.
Results
Seventy eight out of a total of 95 GPs (82% response rate) from a variety of areas in England completed the questionnaires. The GPs identified time, resources and lack of knowledge as barriers, but only 9% agreed with the statement that there is little support they can offer carers. However, nine in ten GPs (89%) feel they have insufficient training here and approximately half of them (47%) lack confidence that they are meeting carers' needs. Confidence in identifying carers is also low (45%). Issues that GPs would look out for amongst carers include emotional and physical health problems and financial and isolation difficulties. GPs specifically highlighted educational and isolation issues for young carers. Few services were described that targeted carers.
Conclusions
GPs recognise that they have an important role to play in supporting carers but would like training and support. Further investigation is needed both to determine how best to train and facilitate GPs and general practice teams in their role in supporting carers and to identify what carers need and want from general practice. Identifying carers' leads or carers' champions amongst practice staff is possibly one way forward. Given the proposed greater commissioning role for primary care, greater understanding here is particularly important. | Competing interests
The authors declare that they have no competing interests.
Authors' contributions
All authors contributed to the study design. NG was primarily responsible for the questionnaire construction. All authors participated in data collection and content analysis of open-ended questions. RH was primarily responsible for the quantitative analysis and coordinated the project. NG drafted the article. All authors read and approved the final manuscript.
Pre-publication history
The pre-publication history for this paper can be accessed here:
http://www.biomedcentral.com/1471-2296/11/100/prepub | Acknowledgements
To authors would like to thank The Department of Health who funded the study; the Royal College of General Practitioners who supported and facilitated data collection; the GPs who completed the questionnaires; Sue Barrett and Barbara Price from South-Thames Crossroads Care and Dr Pippa Oakeshott, Reader in General Practice, who acted as advisers on the project. | CC BY | no | 2022-01-12 15:21:44 | BMC Fam Pract. 2010 Dec 20; 11:100 | oa_package/14/ce/PMC3016257.tar.gz |
PMC3016258 | 21143883 | Background
Maternity care in the Netherlands is known for its high percentage of home births and for the independent position of midwives. In 2002 40.6% of all births took place in primary care, assisted by a midwife or general practitioner, 29.4% at home, 11.2% in hospital or birth centre, and 59.4% of all births took place in secondary care, under supervision of a gynaecologist in hospital [ 1 ].
The Dutch health care is organised in primary care, secondary care and tertiary care. For a patient the first point of contact with the health care system is primary care, which is freely accessible, close to people's homes, and general. Primary care providers as general practitioners (GPs) and midwives are gatekeepers for secondary care: hospitals and medical specialists. A patient needs a referral from a primary care practitioner to have access to a secondary care practitioner, a medical specialist who, in turn, can refer a patient to highly specialised tertiary care.
Most midwives in the Netherlands work in primary care where they are the lead professionals providing care to women with 'normal' or uncomplicated pregnancies. They are independent practitioners, like general practitioners or family doctors, and work in single-handed, duo-, or group-practices. In case of complications or an increased risk of complications during pregnancy, during labour or in the postpartum period, the midwife will refer her client to secondary care, where a gynaecologist will take over responsibility. The indications for referral have been agreed upon by all professional groups involved (gynaecologists, midwives and general practitioners) and are laid down in the so-called Obstetric Indication List [ 2 ]. A typical labour and maternity ward in a general hospital used to be staffed by obstetrical nurses, junior-doctors, sometimes a gynaecologist-in-training, and one or more gynaecologist/obstetricians, with occasionally a midwife.
In the last fifteen years the number of midwives practising in primary care has increased from 1.042 to 1.871 [ 3 ]. In addition, there have always been some midwives working in hospitals ("clinical midwives"), in academic settings and teaching hospitals, primarily to assist and coach gynaecologists-in-training. In the last fifteen years their number has increased threefold (from 192 to 573) with the largest increase since 2002 [ 3 , 4 ]. Several reasons are given for this increase. For instance, staff shortages on labour wards, leading to an increased demand for midwives to fill the vacancies of obstetric nurses and physicians-in-training. But also a growing preference among midwives for a salaried position with regular working hours (only possible in secondary care), especially after having been being self-employed in primary care for several years, during a period of increasing workload [ 5 , 6 ].
In the same period the attitude towards the role of midwives in secondary care has changed from assistant to semi-autonomous (but not independent) care provider. For instance, in 2000 the Dutch association of gynaecologists (NVOG) issued a statement, saying that the presence of midwives on the labour and delivery wards must: 'be seen in the light of an improvement of the quality of patient care on the delivery ward' [ 7 ]. And the Dutch association of midwives (KNOV) stated in 2002 that: 'the most important feature of the clinical midwife is that she, as specialist in physiological care, will guard the physiological approach of a patient with a medical indication within the clinical setting' [ 8 ]. More recently, in 2008, the NVOG stated that the clinical midwife is a valued addition to the obstetrical team because of her specific knowledge of the physiology of pregnancy, birth, and the puerperium [ 9 ].
Midwives working in hospitals have had the same 4-year vocational training at Bachelor level and have the same qualifications as primary care midwives in the Netherlands. That means they are qualified to assist healthy women with uncomplicated births only. But, because of working in a hospital as part of the obstetrical team and under supervision of a gynaecologist, they sometimes perform specific tasks or interventions - such as induction of labour - that are formally outside their competence. Since 2005 additional training, focussed on caring for women with pathology or an increased risk of complications, is available for midwives working in hospital. Since then, the debate about the role of hospital midwives has been between the value of their input as specialists in physiology - in which case they are expected to continue to be primarily involved in normal, uncomplicated labour and birth - and the value of enhancing their competency, so that some tasks of the gynaecologist can be shifted to the midwife, in which case they will increasingly be involved in more complicated births.
Until now, the actual involvement of midwives in maternity care in hospitals has remained invisible for outsiders, because in all statistics births in secondary care are registered as births assisted by gynaecologists. The aim of this study is to reveal the work of clinical midwives, to gain more insight in their role, and especially to determine whether that role is changing with the increase in the numbers of clinical midwives since 2002. We have three research questions. First: how many of the births in secondary care are assisted by midwives? Second: what are the characteristics of the women, the births, and the settings of midwife-assisted births in secondary care? And third: Are there any differences in the involvement of midwives before and after 2002? | Methods
Data source
The Netherlands Perinatal Registry (PRN) collects data from almost all births in the Netherlands. Perinatal data are collected in three separate registries: one for primary care (LVR1), one for secondary care (LVR2) and one for neonatal/paediatric care (LNR). In this retrospective study data of all births, registered in the LVR2 between 1 January 1998 and 31 December 2007, are used for analysis. The LVR2 registry starts at first contact (booking visit or referral from primary care) and contains complete perinatal data from 16.0 gestational weeks onwards. The coverage of the LVR2 is almost complete: in 1999 97% of all gynaecology partnerships provided data, in 2007 99%. Births registered in the LVR2 constitute 59% (1998) to 65% (2007) of all births occurring in the Netherlands [ 10 ]. In the Netherlands ethical approval is not required for this type of study (secondary analyses on anonymous data). Nevertheless the Steering Committee of the PRN approved this study. The PRN is registered at the Dutch Data Protection Authority.
Outcome measurement and determinants
The outcome measure in the analyses was the type of caregiver 'catching' the baby, differentiated in 'gynaecologist', 'gynaecologist-in-training', 'midwife' and 'other' (GP, nurse, other). On the LVR2 registration form a distinction is made between the person responsible for the care provided at the time of birth and the person 'catching' the baby. With births occurring in secondary care the gynaecologist is always the responsible care provider, but not always present at the birth. The person 'catching' the baby is the one present at the time of birth and is assumed to be the one directly involved and thus the one assisting with the birth.
The socio-demographic, obstetric and organisational aspects used in the analyses included: maternal age, parity, cultural/ethnic background, gestational age, process of birth, time of birth, type of hospital, and moment of referral (before onset of labour or during labour).
Statistical analysis
The number and frequency of births in secondary care, subdivided by different caregivers catching the baby, was presented for each year. We analysed which birth-related variables and background variables (maternal and hospital characteristics) were associated with the midwife being the person 'catching the baby'. This analysis was done for 1998 and 2007 separately to study the difference between both ends of the time scale. Following that, the data were combined in two periods: 1998-2002 and 2003-2007 to study the changes since 2002. Because the LVR2 contains nationwide population data and no sample data, differences between subgroups are absolute differences and statistical testing is not required.
For each socio-demographic, obstetric and organisational feature, the strength of its association with the outcome was first expressed as crude odds ratios (OR) with 95% confidence intervals (CI). We then adjusted the ORs in a multivariate logistic regression analysis to show the contribution of the examined feature in relation to the other characteristics. Interaction effects between parity and, subsequently, maternal age, process of birth and moment of referral were also examined with logistic regression analysis.
On average 9% of the records had missing data on the outcome measure (person catching the baby) and these records have been removed from the analyses. Zero to 2.3% of the characteristics were missing. Altogether, 3.2% of the babies from nulliparous mothers had missing values for one or more characteristics; for babies from multiparous mothers this percentage was 2.4%. These babies were not included in the multivariate logistic regression analysis.
All analyses were performed using a computer software package (SAS for Windows version 9.1; SAS Institute Inc., Cary, NC, USA).
Details of ethics approval
Ethical approval is not required for this type of study in the Netherlands.
The PRN is registered at the Dutch Data Protection Authority. | Results
Between 1998 and 2007 the proportion of births in secondary care increased from 59% to 65% of all births. Although the gynaecologist is the responsible care provider with all secondary care births, the person giving hands-on care is often someone else. In almost half of these births the gynaecologist-in-training was the one 'catching the baby', but since 2002 there is a strong increase in the number of births attended by a midwife. Table 1 shows an increase in the proportion of births attended by a midwife in secondary care from 8.3% in 1998 to 26.1% in 2007.
This increase in itself is no surprise, regarding the increasing numbers of hospital midwives. It merely illustrates that midwives are more and more replacing obstetrical nurses and GPs as well as gynaecologists as hands-on care givers on labour wards in secondary care. Figure 1 shows the relation between the number of midwives working in hospitals and the number of births in secondary care assisted by midwives. The correlation between these two is strong.
To get a clearer picture of the role of the midwife, a number of distinctions were made. First, we compared the outcome (midwife 'catching' the baby) in the group of women who were already in secondary care before the onset of labour and the group of women being referred during labour or birth. Both outcomes were comparable, with an increase of midwifery involvement from less than 10% in 1998 to approximately 25% in 2007, although there was a slightly (2-3%) larger involvement of midwives in the group of referrals during labour (Table 2 ).
Second, as is also shown in Table 2 we compared the involvement of midwives with spontaneous births and with assisted births (vacuum, forceps or Caesarean Section (CS)). The midwives' involvement with spontaneous births has increased from 13% to 41%, while the proportion of spontaneous births of all births in secondary care varied only slightly, from 61% in 1998, via 56% in 2001 to 61% in 2007 (Table 3 ).
Figure 2 shows the increase of the percentage of births in which the midwife was the one catching the baby for spontaneous vaginal births only. Differentiating within all spontaneous births in secondary care between births with and births without induction and/or augmentation of labour showed a decrease in the rate of induction (from 32% in 1998 to 25% in 2007, not in table) while showing an increase in the involvement of the midwife with births with induction of labour (from 11% in 1998 to 42% in 2007) and births with augmentation of labour (from 13% in 1998 to 40% in 2007). This is in both cases approximately the same increase as with all spontaneous births.
Table 2 also shows that the distinction according to time of birth for all births does not make a difference: midwives are just as often involved during evenings, nights and weekends as they are during the daytime on weekdays. However, if both previous distinctions are combined, the data show that midwives are more often involved with spontaneous births in the daytime on weekdays, than at other times of the week.
Table 4 shows the increased involvement of midwives over time in relation to background variables of the women and the hospitals. Midwives were least involved with women of 40 years or older, with women giving birth for the first time, with very preterm births and with births in non-academic teaching hospitals. Cultural/ethnic background showed no difference. The differences between age groups, parity groups, gestational age groups and hospitals and the lack of difference in the cultural/ethnic background groups, seen in 1998, were repeated in 2007.
To analyse the associations between socio-demographic, obstetric and organisational characteristics and outcome, the data were combined in two periods: 1998-2002 and 2003-2007. After controlling for background variables, midwives were more likely to assist with spontaneous births without induction or augmentation occurring in the daytime during weekdays and in academic and non-teaching hospitals (Table 5 ). We observed interaction between parity and gestational age, process of birth and referral respectively. Therefore the logistic regression analysis was performed for nulliparous and multiparous women separately.
The analysis showed that nulliparous women were more often assisted by midwives, when referred during labour, in the earlier period (1998-2002), but not in the later period (2003-2007). In the case of multiparae there is no difference between both periods, midwives were more often assisting women after referral during labour than after referral before the onset of labour. In both periods and for both groups of women midwives were less often assisting non-spontaneous births and births during out-of-office hours. The differences regarding different age groups and ethnicity are small. In both groups of women and in both time periods midwives are less often involved in pre-term birth and more often in post-term birth. In the later period the involvement of midwives in the different types of hospitals diverged less from each other than in the earlier period. That may be due to the fact that only after 2002 a considerable number of midwives were employed by non-teaching hospitals.
To find out whether the involvement of midwives with births in secondary care can help to prevent interventions, a prospective study is needed. But we did find that the steady increase after 2001 of the number of midwives working in hospitals coincided with a decrease in the percentage of instrumental births in secondary care (Table 3 ). This is not simply the result of an increasing number of births in secondary care, with a stable number of interventions, or of an increasing proportion of births in secondary care, with a stable proportion of assisted births, calculated over all births. Because referral to secondary care is regarded as an indication that an intervention is needed, the proportion of assisted births within secondary care, in stead of the proportion of assisted births calculated over all births, is expected to stay more or less equal or to increase, but certainly not to decrease. As can be seen in Table 1 and 3 , the number of births in secondary care first increased, with a peak of 116,000 children born in 2000, then decreased again. The number of assisted births (only occurring in secondary care) followed a similar pattern, with an increase leading to a peak in 2001, followed by a steady decrease as shown in Table 3 and the proportion of births in secondary care steadily increased. However, the rate of assisted births changed, not only in secondary care as is shown in Figure 3 but also in the population at large. The percentage of instrumental births, calculated on all births in the Netherlands decreased from 26,6% in 2001, via 25,6% in 2004, to 24,9% in 2007 (results not shown). | Discussion
This study has shown not only that more midwives were assisting with more births in secondary care in 2007 than they were in 1998, but also that their involvement was primarily with spontaneous births and births during day shifts rather than with complicated births and births during night or weekend shifts. We also found that increased involvement of midwives coincides with a decrease in the percentage of instrumental births in secondary care. To appreciate this it is important to realise that all women giving birth in secondary care have been referred to a gynaecologist because of perceived risks or complications. And this status of no longer being considered as low-risk, may trigger an interventionist response. In many cases interventions are indeed needed, but yet some, and in this dataset an increasing number, experience an uncomplicated birth, because the perceived risk did not manifest itself or could successfully be prevented.
We now can answer the research questions we formulated. The first was: how many of the births in secondary care are assisted by midwives? We found that the number has increased from 8 percent to 26 percent of all births in secondary care.
The second research question was: what are the characteristics of the women, the births, and the settings of midwife-assisted births in secondary care? The profile of a midwife-assisted birth in secondary care is that of a spontaneous birth, taking place on a weekday during office hours, in a non-teaching hospital. This profile reflects not only the formal competence of a midwife, but possibly also their work arrangements: primarily working during office hours and much less during evenings, nights and weekends. This profile also suggests that midwives have kept to their formal competence and have not been extending that by increasingly being involved with more complicated births.
The third question we posed was: Are there any differences in the involvement of midwives before and after 2002? The difference we found was in the quantity, not in the characteristics of the women or the births: there was no real difference between the groups of women served by midwives in the two time periods. Hospital midwives were before and after 2002 primarily involved with uncomplicated, spontaneous births. However, concurrent with the increasing involvement of midwives we found a decreasing number of instrumental births. Whether these two developments are related still has to be proven. There are indications from other studies that a relation exists. For instance, in a recent Cochrane review, comparing midwife-led care with other models of care (medical-led care or shared care), midwife-led care was associated with several benefits for mothers and babies, such as fewer episiotomies or instrumental births and increased chance of having a spontaneous vaginal birth and initiating breastfeeding [ 11 , 12 ]. Although the midwives working in hospitals in the Netherlands are not working in midwife-led care, their presence in the clinical setting may introduce elements of the midwifery model, such as increased continuity of care and reduction of unnecessary technology.
Our results show that the involvement of hospital midwives with spontaneous vaginal births has increased from 13% to 41%. That means however, that they are still not involved with more than half of the less complicated births, being approximately 60% of all births in secondary care, so there may be plenty of room for more midwives in hospitals.
Limitations of this study
This is a retrospective study, using existing data, collected in a standard form, not designed for these specific research questions. First, the variable used to indicate the involvement of a caregiver is the person 'catching' the baby. There is no absolute certainty that this person is indeed the one most closely involved with the birth, but it is the only variable available to differentiate between the caregiver responsible for the birth (i.e. the gynaecologist) and the person attending the woman. For instance, the assisting midwife or gynaecologist (in-training) may have let the partner catch the baby, which would have been noted as 'other'. On the other hand, some women giving birth in secondary care after referral during labour may have been attended by their own, primary care midwife. These incidences might be confounding the analysis somewhat, because we do not know how often that may have happened.
Second, there is no detailed information available about the hospital midwives in the registration used for this project. For example, we do not know how many of them work only in daytime shifts or in 24-hour shifts. | Conclusion
The analyses have shown that, although the involvement of midwives with births in secondary care has increased, they are primarily involved with relatively uncomplicated births and not with the more complicated births. But there are still twice as much uncomplicated births in secondary care not assisted by midwives. This may indicate that there is plenty of room for more midwives in hospitals. The analyses have also shown that, since 2002 the percentage of assisted births (vacuum, forceps, CS) has decreased, not only in secondary care but also when calculated as a proportion of all births registered in the Netherlands Perinatal Registry. The interesting question is, whether these two developments are related, but that still needs to be analysed. That analysis is not possible with these retrospective data.
These analyses are only a first attempt to shed more light onto the role of midwives working in hospitals in the Netherlands. For future studies more information is needed about the midwives in the different hospitals, their education, their attitudes and their actual involvement with births in secondary care. | Background
Most midwives in the Netherlands work in primary care where they are the lead professionals providing care to women with 'normal' or uncomplicated pregnancies, while some midwives work in hospitals ("clinical midwives"). The actual involvement of midwives in maternity care in hospitals is unknown, because in all statistics births in secondary care are registered as births assisted by gynaecologists. The aim of this study is to gain insight in the involvement of midwives with births in secondary care, under supervision of a gynaecologist. This is done using data from the PRN (The Netherlands Perinatal Registry), a voluntary registration of births in the Netherlands. The PRN covers 97% to 99% of all births taking place under responsibility of a gynaecologist.
Methods
All births registered in secondary care in the period 1998-2007 (1,102,676, on average 61% of all births) were selected. We analyzed trends in socio-demographic, obstetric and organisational characteristics, associated with the involvement of midwives, using frequency tables and uni- and multivariate logistic regression analyses. As main outcome measure the percentage of births in secondary care with a midwife 'catching' the baby was used.
Results
The proportion of births attended by a midwife in secondary care increased from 8.3% in 1998 to 26.06% in 2007, the largest increase involving spontaneous births of a second or later child, on weekdays during day shifts (8.00-20.00 hr) from younger mothers with a gestational age (almost) at term. After 2002, parallel to the growing numbers of midwives working in hospitals, the percentage of instrumental births decreased.
Conclusions
In 2007 more midwives are assisting with more births in secondary care than in 1998. Hospital-based midwives are primarily involved with uncomplicated births of women with relatively low risk demographical and obstetrical characteristics. However, they are still only involved with half of the less complicated births, indicating that there may be room for more midwives in hospitals to care for women with relatively uncomplicated births. Whether an association exists between the growing involvement of midwives and the decreasing percentage of instrumental births needs further investigation. | Competing interests
The authors declare that they have no competing interests.
Authors' contributions
Both authors contributed substantially to the design of the study. C.W.P.M.H. analysed the data, T.A.W. prepared the manuscript. Both authors read and approved the final manuscript.
Pre-publication history
The pre-publication history for this paper can be accessed here:
http://www.biomedcentral.com/1471-2393/10/80/prepub | Acknowledgements
This study is based on data from the Netherlands Perinatal Registry. We acknowledge all midwives, obstetricians, paediatricians, nurses and residents who routinely collect the perinatal data for the register. Furthermore, the authors are grateful for the support from their respective employers to make this study possible. | CC BY | no | 2022-01-12 15:21:44 | BMC Pregnancy Childbirth. 2010 Dec 9; 10:80 | oa_package/2f/73/PMC3016258.tar.gz |
Subsets and Splits