Genetic variation in the interferon regulatory factor 5 (IRF5) gene affects systemic lupus erythematosus (SLE) susceptibility. However, association is complex and incompletely defined. We obtained fourteen European sample collections with a total of 1383 SLE patients and 1614 controls to better define the role of the different IRF5 variants. Eleven polymorphisms were studied, including nine tag single nucleotide polymorphisms (SNPs) and two extra functional polymorphisms. Two tag SNPs showed independent and opposed associations: susceptibility (rs10488631, P<10−17) and protection (rs729302, P<10−6). Haplotype analyses showed that the susceptibility haplotype, identified by the minor allele of rs10488631, can be due to epistasis between three IRF5 functional polymorphisms. These polymorphisms determine increased mRNA expression, a splice variant with a different exon 1 and a longer proline-rich region in exon 6. This result is striking as none of the three polymorphisms had an independent effect on their own. Protection was independent of these polymorphisms and seemed to reside in the 5′ side of the gene. In conclusion, our results help to understand the role of the IRF5 locus in SLE susceptibility by clearly separating protection from susceptibility as caused by independent polymorphisms. In addition, we have found evidence for epistasis between known functional polymorphisms for the susceptibility effect.
Systemic lupus erythematosus (SLE (MIM 152700)) is a disease that presents multiple unsolved challenges.1 Extensive research has identified many immunological abnormalities, but their significance is still unclear. It seems that defects in T- and B-cell tolerance play a critical role as well as defects in the clearance of apoptotic cells and of immunocomplexes. Recent evidences indicate that there is also an important involvement of the innate immune system involving plasmacytoid dendritic cells (pDCs), toll-like receptors (TLRs) and type I interferons (IFN).2 In SLE etiology, there is evidence for both an environmental and a genetic component.3, 4 This latter component accounts for the increased risk in siblings, which is about 20 times the risk in the population. However, only a few genetic factors had convincingly been identified. Here, we have explored the impact that genetic variation in the interferon regulatory factor 5 (IRF5 (MIM 607218)) locus has on SLE susceptibility.
The starting point in this line of research has been a hypothesis-driven genetic study based in the important role of type 1 IFN-regulated genes in SLE that found association of three SNPs in the 5′ end of IRF5.5 This association, observed in Swedish, Finish and Icelandic subjects,5 has been extended and confirmed in subsequent studies with Swedish, Spanish, Argentineans, European Americans, British, Danish, German and Korean samples.6, 7, 8, 9, 10 In every case, association has been clear indicating this is one of the major genetic factors in SLE susceptibility. Other aspects of this association have already been revealed, but more genetic investigation is required as association extends over many SNPs in the locus and it is unclear which are relevant. At least two promoters and nine mRNA variants have been described in IRF5.7, 8, 9, 11 Some of this variation is related with three functional polymorphisms in the genomic sequence. One of them, the rs2004640 SNP, influences the use of three alternative first exons because its T allele creates a donor splice site after exon 1B.8, 9 The rs10954213 SNP determines the length of the 3′-UTR.7 The A allele at this SNP creates a polyadenylation signal that results in a shorter 3′-UTR and a higher level of expression. Finally, an insertion/deletion polymorphism causes inclusion/loss of ten amino acids in frame that contribute to a proline-rich region of exon 6.9 Results from some previous genetic studies have indicated that SLE susceptibility was associated to a haplotype containing the donor splice site, T allele of rs2004640, and high RNA levels determined by rs10954213 A allele.7, 8 However, the role of the donor splice site has been questioned on functional grounds9 and association with the functional polymorphisms fails to explain but a very low fraction of association with the locus.
In our study, some of the previous questions have been solved by identifying two independent effects, susceptibility and protection, that depend on different polymorphisms. In addition, the susceptibility haplotype could be attributed to the epistatic effect of the three functional polymorphisms, each of them without any contribution in isolation.
DNA samples (Table 1) from 14 independent DNA collections from eight European countries were included in the study. Each collection contained approximately 100 SLE patients and 100 ethnically matched healthy controls, all of them Caucasians. A total of 1383 SLE patients and 1614 controls were analyzed. Females predominated both among SLE patients and controls. Significant heterogeneity in control genotypes and problems of Hardy–Weinberg equilibrium (HWE) led us to exclude from analysis samples from the Czech Republic.
A complete coverage of the IRF5 locus and neighboring sequences was intended with nine tag SNPs polymorphisms selected from HapMap12 (Figure 1; Supplementary Table 1). One of them was not in HWE in control samples, rs4728142 P=0.003, and was excluded from further analysis. Information lost was minimal because this SNP was not tagging any other SNP. Remaining SNPs were in HWE. All genotypes were obtained in a single laboratory with a 98.0% success rate obtained uniformly across cases and controls and different sample collections. Regenotyping was used liberally to check for accuracy of results.
Analysis of the linkage disequilibrium (LD) pattern showed high D' values between the eight tag SNPs indicating that they were in a single LD block (Supplementary Table 2). The r2 values in all the pair-wise comparisons were below 0.8 (mean value=0.28; maximum 0.76) as expected for tag SNPs. Therefore it is likely that variation in the IRF5 locus has been adequately covered and that it was done efficiently with a minimum of redundancy.
Wide and strong association of IRF5 with SLE susceptibility
For the joint analysis of differences between cases and controls, we used Mantel–Haenszel analysis. This analysis allows evaluating significance of the differences taking into account that samples come from different collections. Significant differences were observed in all eight tag SNPs (Table 2). They were very clear in three of them, rs729302, rs2004640 and rs10488631. They are, respectively, the most upstream SNP, placed 9 kb before the transcription start site; the SNP creating a splice donor site; and a SNP that is about 4 kb downstream the gene. The rs10488631 SNP showed, by much, the most extreme difference between cases and controls (G allele frequencies 19.0 and 10.1%, respectively). Given the large sample size of the joint analysis, the P-value for this difference was very low (P<10−17). Marked frequency differences between SLE patients and controls extended for about 25 kb. This together with the low r2 between the involved tag SNPs indicated that the IRF5 locus likely contains several polymorphisms modifying SLE susceptibility.
We checked for effect size heterogeneity between the sample collections using the odds ratio (OR) obtained with allelic frequencies in SLE cases and controls as effect size and the Breslow–Day as heterogeneity test (Table 2). There was significant heterogeneity in OR values for five SNPs. In each of the five SNPs, heterogeneity was caused by divergent results in a single sample collection. For the rs729302 SNP, the collection from Milan showed an increased frequency of the G allele in SLE patients in relation with the controls from the same town. For the other four SNPs, rs10488630, rs10488631, rs4731535 and rs13242262, heterogeneity was due to the samples of the German collection. Once data from the discordant collection for each of these SNPs were excluded from analysis, significant heterogeneity disappeared and OR estimates became more precise (Table 2), but interpretation of results did not change. We looked for possible causes of the discordant results by reviewing the accuracy of genotypes and the available clinical data from the two-outlier sample collections, but no likely cause was found. In subsequent analyses, we have considered whether the heterogeneity introduced by German and Milanese samples modifies the interpretation of results, and we have signaled this circumstance in each case.
Checking different genetic models by logistic regression showed that additive models accounted adequately for the effect of all tag SNPs, and they were significantly better than either dominant or recessive models in many of them (Supplementary Table 3). This was especially clear in the most strongly associated SNPs, rs10488631, rs2004640 and rs729302, where only the additive model fitted genotype effects.
Genotype association conditional in the other tag SNPs
Given the broad association appreciated, it became necessary to analyze the effects of each tag SNP conditional in other SNPs to distinguish between independent effects and those due to LD. Conditional logistic regression with additive models showed that only two SNPs, rs2004640 and rs729302, retained significant association when evaluated conditional on the most strongly associated SNP, rs10488631 (Figure 2a). Therefore, the effect of these two SNPs could have an independent contribution to SLE susceptibility. On the other side, association of rs10488631 with SLE was almost unaffected by accounting for any of the other SNPs, as shown in Figure 2b for the rs10488631 OR conditional on rs2004640, on rs729302 or on both SNPs. These results indicated that the large effect of the rs10488631 SNP was almost completely independent of any of the other tag SNPs.
Further analysis of the rs2004640 and rs729302 SNPs showed that none of them remained significant when considered conditional on the other SNP together with rs10488631 (Figure 2b). In consequence, only one of these two SNPs contributed independently to association, but it was unclear which of them as inclusion of one cancelled the effect of the other. Excluding samples from Germany and Milan modified this result: in this case, the rs729302 SNP remained significantly different whereas the rs2004640 SNP did not (Supplementary Figure 1).
Haplotype analysis of tag SNPs
We performed haplotype analysis with the eight IRF5 tag SNPs. There were only six haplotypes with a frequency over 5% (Table 3). They represented 83.4 and 82.4% of the chromosomes in SLE patients and controls, respectively. Frequency distribution of haplotypes was extremely different in SLE patients and controls (P=2.7 × 10−14). It showed readily two opposing effects: a susceptibility haplotype, #6, and two protective haplotypes, #1 and 2 (Figure 3). The most different discordant frequency was that of the susceptibility haplotype. It was identified by the minor allele, G, of the rs10488631 SNP. Haplotype #6 contained the common alleles at all the other SNPs except rs4731535. The common alleles were present in several other haplotypes and the minor allele of rs4731535 was included in two other haplotypes, a neutral, #3, and a protective haplotype, #2. A haplotype that has previously been reported as conferring SLE susceptibility, defined by the T allele of rs2004640 and the A allele of rs2280714,8 was split in three haplotypes, #4–6, in our analysis and only haplotype #6 showed association.
Two haplotypes were significantly less frequent in SLE patients than in controls, #1 and #2 (Figure 3). These differences were not due to an effect of balancing out the susceptibility effect of haplotype #6 because the two protective haplotypes together represented only about 20% of the total. The protective haplotypes were uniquely characterized by containing the minor allele, C, of the rs729302 SNP. All the other alleles in these two haplotypes were also present in neutral haplotypes. Especially interesting was the comparison of haplotypes #2 and #3 because they only differed in the allele of rs729302. In haplotype #2, which was protective, the rs729302 allele was C, whereas in the neutral haplotype #3 it was A. We explored more carefully the contribution of SNPs rs729302 and rs2004640 to the protective haplotypes. For this analysis, only three SNPs, rs729302, rs2004640 and rs10488631, were included in a new estimation of haplotype frequencies (Supplementary Table 4b). The table of haplotype frequencies showed that only the rare allele of SNP rs729302 correlated unambiguously with the protective haplotype.
In summary, association of the IRF5 locus was due to two independent and opposing effects that depended on haplotypes defined by the minor alleles of two SNPs, rs10488631 and rs729302.
Two additional functional polymorphisms in IRF5
Two new functional polymorphisms have been described in IRF5 during progress of our study, an insertion/deletion polymorphism and the rs10954213 SNP.7, 9 We have genotyped them in 486 control and 459 SLE samples from Slovakia, Rome, Naples, Hungary and Santiago with a 98.3% success rate and without significant deviation from HWE. The two were already well covered by the tag SNPs of our panel. SNP rs10954213 had r2 of 0.93 with SNP rs13242262, and the insertion/deletion polymorphism had r2 of 0.90 with SNP rs4731535. Therefore, results were very similar to the observed for these two tag SNPs.
The rs10954213 SNP was moderately different in SLE patients and controls (Table 4). The insertion/deletion polymorphism was not significantly different. When data from rs10954213 were considered conditional on the rs10488631 SNP, the association disappeared. The unconditional OR of this SNP, 0.82 (0.67–0.99), changed to a conditional OR of 0.94 (0.76–1.16). These results indicate that none of these two new functional polymorphisms have a significant contribution to SLE susceptibility by themselves.
Haplotype analysis considering all ten polymorphisms showed that the two new polymorphisms were irrelevant for the distribution of haplotype frequencies (Supplementary Table 5a). To focus on the three functional polymorphisms, we obtained the frequency distribution of haplotypes considering only these polymorphisms and the rs729302 and rs10488631 SNPs (Figure 4 and Table 5). There were six haplotypes with frequencies over 5% that accounted for 87.7 and 91.7% of all haplotypes in cases and controls, respectively. The effects of the individual SNPs were as described in the previous analysis: protection correlated with the minor allele of rs729302 and susceptibility with the minor allele of rs10488631. None of the putative functional polymorphisms correlated with any of the two effects, susceptibility or protection. However, there were evidences of an epistatic effect between them. The susceptibility haplotype, #6, included a combination of alleles for the three functional SNPs that was not present in any other common haplotype (even in haplotypes with frequencies over 2%). This combination, ‘T-in-A’, included the allele inducing alternative splicing (T of rs2004640), the insertion prolonging the proline-rich region of exon 6 and the short 3′-UTR associated with higher mRNA expression levels (A of rs10954213). None of the other common combinations of these three polymorphisms was associated with an increase in disease susceptibility. On the other side, the alleles corresponding to the three functional polymorphisms were irrelevant in the protective haplotypes.
The rs2004640 SNP and the two SNPs (rs4731535 and rs13242262) that were tag SNPs for the other two functional polymorphisms showed also a unique combination of alleles in the susceptibility haplotype #6 also when all the sample collections were included (Figure 3, Supplementary Table 4A). Therefore, results implicating an epistatic effect between the three functional polymorphisms were consistently observed.
Haplotype analysis restricted to females also showed the results that have been described above for the unstratified set of samples (Supplementary Tables 6A and B). We performed a correlation analysis between the rs10488631 and rs739302 genotypes and the 1997 ACR classification criteria for SLE, but no correlation was found. We also checked that inclusion of the rs4728142 SNP or of Czech Republic sample collection data, which were excluded from analysis as mentioned above, did not significantly modify the results.
Our results have confirmed the striking association of genetic variation in the IRF5 locus with SLE susceptibility.5, 6, 7, 8, 9, 10 Association is already very clear with P-values that make the results of an unprecedented strength in SLE genetics. It has been found in several populations of European Caucasians, in Argentineans, in Koreans and in a small number of Indo-Pakistani families. It is still unclear whether all ethnic groups will be affected similarly. We have included a diverse array of European Caucasian sample collections, and although there was some heterogeneity between them, it did not follow any identifiable pattern. Our results have shown the most extreme differences to date, probably due to the choice of tag SNPs, to the large number of samples and the use of a case-control design. However the SLE-associated haplotypes are present only in a fraction of SLE patients and it is unclear if they are from a specific subset of patients. Previous studies had already signaled, and we have confirmed this, that there are not specific SLE ACR classification criteria related with genetic variability in the IRF5 locus.5, 7, 8 However, data on disease activity and accumulated damage were not included in any of these analyses, and these SLE features are better candidates to be related with IRF5 variation because they have shown correlation with the overexpression of type 1 IFN response genes.13, 14
Our analysis has contributed to clarify the contribution of IRF5 to SLE in several ways. First, it has allowed the distinction of two independent and opposed effects in the locus. Magnitude of the susceptibility effect was larger than the protective one: the susceptible haplotype was about twice more frequent in SLE patients than in controls, whereas the protective haplotypes were near 1.5-fold less frequent. This is the first time the two opposed effects have been shown in a clear way and that they have been shown to be independent. Only a previous report mentioned a protective haplotype,7 but it was unclear whether the protective effect was related with the minor allele, G, of the rs2004640 SNP and, therefore, a consequence of the susceptibility effect associated with its major T allele. However, the G of rs2004640 in our study was present in three haplotypes and only two of them were significantly less frequent in SLE patients than in controls indicating that it could not account for SLE protection. In the other reports,5, 6, 8, 9, 10 an independent protective effect was not observed because they included less SNPs and their haplotype analysis lacked sensitivity.
Susceptibility was accounted for by the minor allele of one of the tag SNPs, rs10488631, which has not identifiable functional relevance. This SNP has already been studied in UK families without being remarked as special but only about 30% of the families were informative for this SNP and this, likely, hampered analysis.7 The SNP has a high r2 with ten other SNPs according to HapMap data (Supplementary Figure 2) that are, all of them, in the TNPO3 gene. As the minor allele of rs10488631 identifies unambiguously the susceptibility haplotype, it is clear that this haplotype extends into the TNPO3 gene. This raises the possibility that variation in this large gene, coding for a protein involved in transport to the nucleus of some members of the spliceosome, could play a role in SLE susceptibility. However, it seems unwarranted to relate this gene to SLE in the absence of any other evidence and because none of the SNPS showing high values of r2 with rs10488631 was predicted to have functional relevance by bioinformatics analysis. A better-grounded alternative is to consider rs10488631 as a tag SNP for the haplotype formed by the ‘T-in-A’ alleles of the three identified functional IRF5 polymorphisms. This haplotype includes the donor splice site for exon 1B, the two extra repetitions in exon 6 that extend a proline-rich region and the short 3′-UTR that determines increased mRNA expression of IRF5. What is most notable is that none of these three alleles seem able to affect SLE susceptibility by themselves or even in the combinations that are observed in common haplotypes. Results pointing to a similar conclusion have also been reported recently.15 It is unclear whether the three are required for increased SLE risk, it could be that ‘T-in-x’ or ‘x-in-A’ haplotypes will be enough as they are also restricted to the susceptibility haplotype #6. The latter haplotype, ‘x-in-A’ seems to be favored by the experiments that show that the donor splice SNP rs2004640 has little consequence in the pattern of expressed IRF5 variants.9 The main piece of evidence required to complete the hypothesis of epistasis is to define the role of the insertion/deletion polymorphism. The effects of the donor splice site and the 3′-UTR variants have already been sufficiently demonstrated.7, 8, 9 On the contrary, the effect of the different length of the proline-rich region has not been explored. This type of region has been involved in protein–protein interactions.16 It will also be necessary to find why IFR5 function is dramatically different when the functional alleles are in the same haplotype. In this respect, it has already been reported that each different IRF5 mRNA variant induces expression of a different pattern of IFN subtypes in virus-infected cell lines.11
Regarding the protective effect, the rs729302 SNP that identifies the protective haplotypes has not any other SNP in tight LD according to HapMap. It is located about 9 kb upstream the IRF5 gene and has no obvious consequence. Given this position, it is tempting to speculate that it could be in LD with unknown SNPs that affect IRF5 promoters.11
We think important to remark the significant contribution that HapMap-selected tag SNPs have had to our results. Coverage of variation in the IRF5 locus has been very good as evidenced not only in the LD analysis between the tag SNPs, but also by the two polymorphisms included in a second part of the analyses that were already well covered. This efficient coverage allowed us to explore the full locus with a small number of SNPs and permitted our new findings. Only, another report has used a comparable coverage and this is the report finding a more complete picture of the effects determined by IRF5 variation before ours.7 We have obtained some additional details because we used a case-control design with large sample size that is more powerful than a family association study; and because, thanks to the work of others, we were aware of the three functional polymorphisms in IRF5.
Although IRF5 is likely to have an important participation in SLE autoimmunity, most studies of this transcription factor have been related with response to viral infections and with its involvement in regulation of cell cycle and survival of cancer cells.17, 18, 19, 20 IRF5 is a critical mediator of signals triggered by TLRs.21, 22, 23 Its phosphorylation and nuclear translocation leads to the overexpression of many proinflammatory genes including cytokines, chemokines and type 1 IFN.22, 24, 25 The specific link of IRF5 to SLE susceptibility could reside in the antigen–antibody complexes containing DNA and RNA.1, 2 These complexes are taken by pDCs and interact with TLR7/8 and 9 leading to activation of DCs and of autoreactive B cells.
In conclusion, our data indicate that IRF5 is a major factor in SLE genetics, with variants playing both a susceptible and a protective role. Likely candidate for the susceptible effect is an epistatic haplotype determining increased mRNA expression of an IRF5 isoform including exon 1B and an extended proline-rich region, although it is unclear whether the three alleles are necessary. If this hypothesis is confirmed by other studies, it will be a very clear example of epistasis as none of the three IRF5 variants individually, and none of the other available haplotypes, has significant effect. The protective effect seems independent of any known functional polymorphism. Future research should address the mechanisms behind epistasis and the protective effect.
Materials and methods
We looked for DNA samples from about 100 SLE patients and 100 ethnically matched healthy controls from different European countries (Table 1). Most of these samples have been already described;26, 27 new collections were from Corunna in Spain, Martin in Slovakia and Groningen in the Netherlands. Significant heterogeneity and lack of HWE in several SNPs among controls led to the exclusion from further study of samples from the Czech Republic. Initial heterogeneity of genotypes of controls from Madrid disappeared after the inclusion of more controls. Patients and controls from the German collection have been included also in the study done by Kozyrev et al.9 A total of 1383 SLE and 1614 control DNA samples were obtained as detailed in Table 1. All SLE patients met the revised American College of Rheumatology classification criteria.28 Patients and controls provided written informed consent. Sample collection was approved by the respective ethical committees and the study was reviewed and approved by the Comite Etico de Investigacion Clinica de Galicia (Spain).
We have studied 11 polymorphisms in the IRF5 locus (Figure 1). Initially, we selected nine tag SNPs that covered variability in the IRF5 gene and neighboring sequences (from 10 kb upstream to 20 kb downstream) based on HapMap CEU data (Supplementary Table 1).12 For this selection, we fixed SNPs that have been shown to be associated with SLE susceptibility and used the pair-wise option of the Haploview software29 with a threshold of r2=0.8 and a minor allele frequency of 10%. These nine SNPs were amplified in a single PCR reaction done with the Qiagen Multiplex PCR kit (Qiagen, CA, USA) on 30 ng of genomic DNA. PCR products were purified by Exo-SAP digestion with Exonuclease I (Epicentre, Madison, WI, USA) and Shrimp Alkaline Phosphatase (Amersham Biosciences, Barcelona, Spain). Single-base extension reactions with the SNaPshot Multiplex Kit (Applied Biosystems, Foster City, CA, USA) were done. Samples were analyzed in the Abi Prism 3130xl Genetic Analyzer (Applied Biosystems).
Two functional polymorphisms in IRF5 that were described during this study were included.7, 9 The rs10954213 SNP was genotyped with a fluorogenic 5′ nuclease assay (TaqMan MBG Genotyping Assay, Applied Biosystems) performed in a total volume of 10 μl containing 25 ng of genomic DNA, 1 × TaqMan Universal PCR Master Mix, No Amperase UNG (Applied Biosystems) and 1 × Assay Mix containing primers and labeled probes. A Chromo4 real-time PCR system (MJ Research, Waltham, MA, USA) was used to run this assay. An insertion/deletion polymorphism of 30 bp in exon 6 has been genotyped by sizing the amplicons resulting from conventional PCR in 3% agarose gels. Sequences of primers and probes are available in Supplementary Table 1.
Analysis of results was based on the following programs: Haploview,29 Arlequin (106 permutations used thoroughly),30 Statistica 7.0 (StatSoft, Tulsa, OK, USA) and Phase 2.1.31 HWE was tested in control samples with a threshold of 0.05 without correction for multiple testing. The exact test of population differentiation32 implemented in Arlequin was used to explore differences in unphased genotypes between the control samples from each collection. Only controls were included because the objective of this analysis is to explore if the sampled populations can be considered in relation to IRF5 variation either as a single one or as multiple different populations. To account for possible variability in effect size between collections, case-control allele frequencies were compared with the Mantel–Haenszel test for the 2 × 2 contingency tables stratifying by collection of samples. In other words, response variable has been disease status; explanatory variable, allele frequency and control variable, the sample collection. Homogeneity of effect size across collections was assessed with the Breslow–Day test using as effect size the OR obtained with the allelic frequencies in SLE patients and controls. Likelihood ratio tests for the additive, dominant and recessive genetic models were obtained relative to the codominant model. Likelihoods for the fit of each model were calculated with univariate logistic regression. Multivariate logistic regression analysis was used to evaluate the effect of each SNP conditional on the remaining. A simplified additive model without interaction parameters was used for this analysis with genotype codes 0, 1 and 2 for genotypes AA, Aa and aa, respectively. Haplotypes were estimated with the Phase 2.1 software (default parameters) that implements a Bayesian algorithm. Further analysis was done by supervised removal of SNPs that did not discriminate between individually associated haplotypes. OR for each haplotype and its 95% confidence intervals (CI) were calculated taking as reference all chromosomes not bearing the haplotype. Assessment of the possible functional relevance of the SNPs was done with the PUPAsutie that includes several prediction software applications (http://pupasuite.bioinfo.cipf.es/).
Croker JA, Kimberly RP . SLE: challenges and candidates in human disease. Trends Immunol 2005; 26: 580–586.
Pascual V, Farkas L, Banchereau J . Systemic lupus erythematosus: all roads lead to type I interferons. Curr Opin Immunol 2006; 18: 676–682.
Nath SK, Kilpatrick J, Harley JB . Genetics of human systemic lupus erythematosus: the emerging picture. Curr Opin Immunol 2004; 16: 794–800.
Croker JA, Kimberly RP . Genetics of susceptibility and severity in systemic lupus erythematosus. Curr Opin Rheumatol 2005; 17: 529–537.
Sigurdsson S, Nordmark G, Goring HH, Lindroos K, Wiman AC, Sturfelt G et al. Polymorphisms in the tyrosine kinase 2 and interferon regulatory factor 5 genes are associated with systemic lupus erythematosus. Am J Hum Genet 2005; 76: 528–537.
Demirci FY, Manzi S, Ramsey-Goldman R, Minster RL, Kenney M, Shaw PS et al. Association of a common interferon regulatory factor 5 (IRF5) variant with increased risk of systemic lupus erythematosus (SLE). Ann Hum Genet 2006; 71: 308–311.
Graham DS, Manku H, Wagner S, Reid J, Timms K, Gutin A et al. Association of IRF5 in UK SLE families identifies a variant involved in polyadenylation. Hum Mol Genet 2007; 16: 579–591.
Graham RR, Kozyrev SV, Baechler EC, Reddy MV, Plenge RM, Bauer JW et al. A common haplotype of interferon regulatory factor 5 (IRF5) regulates splicing and expression and is associated with increased risk of systemic lupus erythematosus. Nat Genet 2006; 38: 550–555.
Kozyrev SV, Lewen S, Reddy PM, Pons-Estel B, Witte T, Junker P et al. Structural insertion/deletion variation in IRF5 is associated with a risk haplotype and defines the precise IRF5 isoforms expressed in systemic lupus erythematosus. Arthritis Rheum 2007; 56: 1234–1241.
Shin HD, Sung YK, Choi CB, Lee SO, Lee HW, Bae SC et al. Replication of genetic effects of interferon regulatory factor 5 (IRF5) on systemic lupus erythematosus in a Korean population. Arthritis Res Ther 2007; 9: R32.
Mancl ME, Hu G, Sangster-Guity N, Olshalsky SL, Hoops K, Fitzgerald-Bocarsly P et al. Two discrete promoters regulate the alternatively spliced human interferon regulatory factor-5 isoforms. Multiple isoforms with distinct cell type-specific expression, localization, regulation, and function. J Biol Chem 2005; 280: 21078–21090.
Consortium IH . A haplotype map of the human genome. Nature 2005; 437: 1299–1320.
Baechler EC, Batliwalla FM, Karypis G, Gaffney PM, Ortmann WA, Espe KJ et al. Interferon-inducible gene expression signature in peripheral blood cells of patients with severe lupus. Proc Natl Acad Sci USA 2003; 100: 2610–2615.
Bennett L, Palucka AK, Arce E, Cantrell V, Borvak J, Banchereau J et al. Interferon and granulopoiesis signatures in systemic lupus erythematosus blood. J Exp Med 2003; 197: 711–723.
Graham RR, Kyogoku C, Sigurdsson S, Vlasova IA, Davies LR, Baechler EC et al. Three functional variants of IFN regulatory factor 5 (IRF5) define risk and protective haplotypes for human lupus. Proc Natl Acad Sci USA 2007; 104: 6758–6763.
Li SS . Specificity and versatility of SH3 and other proline-recognition domains: structural basis and implications for cellular signal transduction. Biochem J 2005; 390: 641–653.
Barnes BJ, Moore PA, Pitha PM . Virus-specific activation of a novel interferon regulatory factor, IRF-5, results in the induction of distinct interferon alpha genes. J Biol Chem 2001; 276: 23382–23390.
Mori T, Anazawa Y, Iiizumi M, Fukuda S, Nakamura Y, Arakawa H et al. Identification of the interferon regulatory factor 5 gene (IRF-5) as a direct target for p53. Oncogene 2002; 21: 2914–2918.
Hu G, Mancl ME, Barnes BJ . Signaling through IFN regulatory factor-5 sensitizes p53-deficient tumors to DNA damage-induced apoptosis and cell death. Cancer Res 2005; 65: 7403–7412.
Yanai H, Chen HM, Inuzuka T, Kondo S, Mak TW, Takaoka A et al. Role of IFN regulatory factor 5 transcription factor in antiviral immunity and tumor suppression. Proc Natl Acad Sci USA 2007; 104: 3402–3407.
Schoenemeyer A, Barnes BJ, Mancl ME, Latz E, Goutagny N, Pitha PM et al. The interferon regulatory factor, IRF5, is a central mediator of toll-like receptor 7 signaling. J Biol Chem 2005; 280: 17005–17012.
Takaoka A, Yanai H, Kondo S, Duncan G, Negishi H, Mizutani T et al. Integral role of IRF-5 in the gene induction programme activated by toll-like receptors. Nature 2005; 434: 243–249.
Ouyang X, Negishi H, Takeda R, Fujita Y, Taniguchi T, Honda K et al. Cooperation between MyD88 and TRIF pathways in TLR synergy via IRF5 activation. Biochem Biophys Res Commun 2007; 354: 1045–1051.
Barnes BJ, Kellum MJ, Field AE, Pitha PM . Multiple regulatory domains of IRF-5 control activation, cellular localization, and induction of chemokines that mediate recruitment of T lymphocytes. Mol Cell Biol 2002; 22: 5721–5740.
Barnes BJ, Richards J, Mancl M, Hanash S, Beretta L, Pitha PM et al. Global and distinct targets of IRF-5 and IRF-7 during innate response to viral infection. J Biol Chem 2004; 279: 45194–45207.
Ferreiros-Vidal I, Gomez-Reino JJ, Barros F, Carracedo A, Carreira P, Gonzalez-Escribano F et al. Association of PDCD1 with susceptibility to systemic lupus erythematosus: evidence of population-specific effects. Arthritis Rheum 2004; 50: 2590–2597.
Ferreiros-Vidal I, D'Alfonso S, Papasteriades C, Skopouli FN, Marchini M, Scorza R et al. Bias in association studies of systemic lupus erythematosus susceptibility due to geographical variation in the frequency of a programmed cell death 1 polymorphism across Europe. Genes Immun 2007; 8: 138–146.
Hochberg MC . Updating the American College of Rheumatology revised criteria for the classification of systemic lupus erythematosus. Arthritis Rheum 1997; 40: 1725.
Barrett JC, Fry B, Maller J, Daly MJ . Haploview: analysis and visualization of LD and haplotype maps. Bioinformatics 2005; 21: 263–265.
Schneider S, Roesli D, Excoffier L . Arlequin ver 2000 A: Software for Population Genetics Data Analysis. Genetics and Biometry Laboratory, University of Geneva: Switzerland, 2000.
Stephens M, Smith NJ, Donnelly P . A new statistical method for haplotype reconstruction from population data. Am J Hum Genet 2001; 68: 978–989.
Goudet J, Raymond M, de Meeus T, Rousset F . Testing differentiation in diploid populations. Genetics 1996; 144: 1933–1940.
Cristina Fernandez-Lopez and Marta Picado have provided outstanding technical assistance. This work has been supported by Fondo de Investigacion Sanitaria, Instituto de Salud Carlos III (Spain), Grants 04/1651 and 06/0620 that are partially financed by the FEDER program of the EU and by Grants from the Xunta de Galicia. SR and CD were supported by Grant 00023728 of the Ministry of Health of the Czech Republic. RES and TW were supported by BMBF, KN Rheuma C2.12. Work by SD’A was supported by Telethon (Grant E1221) and the CARIPLO Foundation.
About this article
Cite this article
Ferreiro-Neira, I., Calaza, M., Alonso-Perez, E. et al. Opposed independent effects and epistasis in the complex association of IRF5 to SLE. Genes Immun 8, 429–438 (2007) doi:10.1038/sj.gene.6364407
- haplotype analysis
- systemic lupus erythematosus
Life Sciences (2019)
Rare variants in non-coding regulatory regions of the genome that affect gene expression in systemic lupus erythematosus
Scientific Reports (2019)
Association of IRF5 polymorphisms with increased risk for systemic lupus erythematosus in population of Crete, a southern-eastern European Greek island
Translational Research (2016)
IRF5, PTPN22, CD28, IL2RA, KIF5A, BLK and TNFAIP3 genes polymorphisms and lupus susceptibility in a cohort from the Egypt Delta; relation to other ethnic groups
Human Immunology (2015)