|Home | About | Journals | Submit | Contact Us | Français|
Genetic variants of the interferon (IFN) regulatory factor 5 (IRF5) gene are associated with systemic lupus erythematosus (SLE) susceptibility. The contribution of these variants to IRF-5 expression in primary blood cells of SLE patients has not been addressed, nor has the role of type I IFN. The aim of this study was to determine the association between increased IRF-5 expression and the IRF5 risk haplotype in SLE patients.
IRF-5 transcript and protein levels in 44 Swedish patients with SLE and 16 healthy controls were measured by quantitative real-time PCR, minigene assay, and flow cytometry. The rs2004640, rs10954213, rs10488631 and the CGGGG indel were genotyped in these patients. Genotypes of these polymorphisms defined a common risk and protective haplotype.
IRF-5 expression and alternative splicing were significantly upregulated in SLE patients versus healthy donors. Enhanced transcript and protein levels were associated with the risk haplotype of IRF5; rs10488631 gave the only significant independent association that correlated with increased transcription from non-coding exon 1C. Minigene experiments demonstrated an important role for rs2004640 and the CGGGG indel, along with type I IFNs in regulating IRF-5 expression.
This study provides the first formal proof that IRF-5 expression and alternative splicing are significantly upregulated in primary blood cells of SLE patients. The risk haplotype is associated with enhanced IRF-5 transcript and protein expression in SLE patients.
Systemic Lupus Erythematosus (SLE) is a multi-system complex autoimmune disease driven by genetic risk factors and environmental stressors. Although the etiopathogenesis of SLE is still incompletely understood, patients display elevated type I IFN in their serum and IFNα-induced gene transcripts in their blood cells (1–3). The increase correlates with both disease activity and severity suggesting a role for IFNα in disease pathogenesis (4). It is still not known whether the type I IFN expression signature is a general biomarker of an activated/dysregulated immune system or rather it reflects primary genetic variation causal to disease pathogenesis. Recent data from Niewold et al. provides initial support for the latter hypothesis indicating that high serum IFNα activity is a heritable risk factor (5). Several studies have recently reported genetic association between the transcription factor IFN regulatory factor 5 (IRF5) and SLE. IRF-5 is a DNA binding protein that controls inflammatory and immune responses (6–7). It has been shown to be involved in host defense against viruses; more recent data demonstrate its critical role in Toll-like receptor (TLR) signaling (8–9). Additional studies in mice show that production of IFNα, IFNβ, and IL6 in response to sera or IgG-RNA immune complexes from lupus patients is dependent on TLR7, IRF5 and IRF7 (10). Besides its function in the host immune response to pathogens, IRF-5 regulates the expression of genes involved in apoptosis, cell cycle, and cell adhesion, all of which play critical roles in autoimmunity (11–15).
Genetic association studies examining single-nucleotide polymorphisms (SNPs) in the IRF5 gene have defined distinct haplotypes that confer either susceptibility to (risk), or protection from, SLE in persons of varying ethnic ancestry (16–21). A primary hypothesis is that these haplotypes will be characterized by functional changes in IRF-5 transcription, resulting in altered IRF-5 protein/isoform expression, IRF-5-mediated transcription and biological activity. The rs2004640 T allele, located within the 5′-untranslated region (UTR) of IRF5, was the first SNP identified showing convincing association with SLE (16) and later studied for its functional consequence on IRF-5 expression (19). SNP rs2004640 was shown to create a 5′ donor splice site in the alternate Ex1B resulting in enhanced expression of Ex1B-associated transcripts and decreased Ex1C utilization (19). Two new isoforms were isolated from peripheral blood mononuclear cells (PBMC) of rs2004640 heterozygous donors suggesting that a given haplotype may drive expression of unique IRF-5 isoforms (19). A subsequent report suggested that transcription from Ex1B was negligible in SLE patients with different haplotypes of the rs2004640 SNP since only isoforms initiating from Ex1A were identified (20). SNP rs10954213, located within the 3′-UTR of IRF5, was shown to provide an alternate polyA+ signal shortening the length of the 3′-UTR and enhancing IRF-5 stability (22).
Little is still known of the functional significance of IRF5 gene polymorphisms and their association with expression or contribution to biological function that may implicate them in SLE disease pathogenesis. Important insight was gained recently by Niewold et al. demonstrating an association between serum IFNα activity and an IRF5 risk haplotype in SLE patients (23). We recently reported that variants of IRF5 with the highest posterior probabilities of being causal in SLE were a SNP rs10488631 located 3′ of IRF5, and a novel CGGGG insertion-deletion (indel) polymorphism located 64 bp upstream of the first untranslated exon (Ex1A) (24). The CGGGG indel is diallelic and part of a polymorphic repetitive DNA-stretch that consists of either three or four CGGGG repeats; the insertion of one CGGGG repeat is the risk allele (4xCGGGG). The CGGGG indel explained the association signal from multiple SNPs in the IRF5 gene, including rs2004640 and rs10954213, previously considered to be causal variants in SLE (24) (Fig. 1A). Here, we examine the functional association of the risk alleles of these four polymorphisms (CGGGG indel, rs2004640, rs10488631, and rs10954213) with IRF-5 transcript and protein expression in SLE.
Our study included 44 Swedish patients with SLE and 16 controls. The patients originate from the rheumatology clinic at Uppsala University Hospital in Sweden and consist of 35 women and 9 men with a mean age at disease onset of 31.4 ± 11.2 years and a mean age at blood sampling of 46.9 ± 13.4 years. Each of the patients fulfilled at least four of the classification criteria for SLE as defined by the American College of Rheumatology (ACR) with a median number of ACR criteria of 5 (range 4–8) (25). Clinical disease activity at the time of blood sampling was assessed with the modified SLE disease activity index (mSLEDAI-2K) where complement and anti-DNA antibodies were omitted (26–27). The median mSLEDAI-2K value was 1 (range 0–26) showing that most patients had low or no clinical disease activity; a few patients had high disease activity at the time of blood collection. The controls were from the same regions as the patients and were age and sex-matched. All study subjects provided informed consent to participate in the study, and the study was approved by institutional and regional ethical boards at both institutions.
Subjects were genotyped at the rs2004640, rs10954213, and rs10488631 SNPs in 250ng of DNA extracted from blood samples of study subjects using the Illumina Golden Gate assay (Illumina, San Diego, CA), as described (28). The 5 bp CGGGG indel of IRF5 was genotyped by PCR amplification followed by size separation using 4% agarose gels, or using an ABI 3770 capillary sequencer (Applied Biosystems, Foster City, CA).
PBMCs were isolated from peripheral blood of genotyped SLE patients and healthy donors by Ficoll gradient centrifugation. Total RNA was isolated with the Qiagen RNeasy mini kit and 0.5 μg of RNA was reverse-transcribed (29). Semi-quantitative PCR was performed with primers provided in Supplementary Table 1. PCR products were analyzed by gel electrophoresis and individual bands excised, purified and sequenced (29).
Q-PCR was performed with custom-designed primers (Supplementary Table 1). Total IRF-5 was measured by Taqman assay using PerFecta qPCR supermix, a dual-labeled BHQ oligonucleotide probe (Biosearch Technologies, Novato, CA) and primers amplifying conserved exons 8–9. GAPDH was measured as an endogenous control for the Taqman assay (Applied Biosystems). IRF-5 transcripts were measured using Ex1-specific forward primers (29), Ex2-specific reverse primers, and Power SYBR Green PCR Master Mix (Applied Biosystems). IFNA transcripts were similarly detected by SYBR green assay. β-actin served as an endogenous control for SYBR green assays. All reactions were performed in the 7300 Real-Time PCR System (Applied Biosystems), and conditions were as follows: for Taqman assays, 1 cycle of 45°C (5 min) and 95°C (3 min), followed by 40 cycles of 95°C for 15 sec and 60°C for 45 sec; for SYBR Green assays, 1 cycle of 45°C (3 min) and 95°C (10 min), followed by 40 cycles of 95°C for 15 sec and 60°C for 1 min. Relative mRNA expression was calculated using the delta Ct method with normalization to β-actin or GAPDH.
Cell lysates from PBMC pellets were prepared by resuspending in equal volumes of SDS loading buffer (29). Samples were denatured and proteins separated by 10% SDS–PAGE, transferred to nitrocellulose membrane and probed with anti-human IRF5 monoclonal antibodies (mAbs) (M01; Novus Biologicals, Littleton, CO) or anti-tubulin Abs (Cell Signaling, Danvers, MA) at a 1:1000 dilution (24). Immunoreactive protein complexes were visualized with enhanced chemiluminescence (ECL) reagents (GE Healthcare, Piscataway, NJ).
Serum IFNα levels were measured by a dissociation-enhanced lanthanide fluoroimmunoassay (DELFIA); on the solid phase were anti-IFNα mAbs LT27:273 and LT27:293, and for detection europium-labeled LT27:297 anti-IFNα mAb (30). The detection level of the assay was ≥1 U IFNα/ml and the standard was calibrated against the NIH reference leukocyte IFNα GA-23-902-530.
5′-UTR or Ex5-7 minigene constructs were generated by PCR from genotyped genomic DNA of 2 healthy donors or 11 SLE patients (24). The pcDNA3 minigene constructs were transfected to Hek 293 cells using Lipofectamine 2000 reagent (Invitrogen, San Diego, CA) with or without the indicated SR or snRNP proteins. ASF/SF2 and SRp40 were from Robert Lafyatis (Boston University); U1, SRp20 and SFRS10/tra2Beta were from Jane Wu (Vanderbilt University). 24 hr post-transfection, cDNA was synthesized as described (24) and subjected to 35 PCR cycles using Taq DNA polymerase (Invitrogen). Primers (Supplementary Table 1) were derived from sequences of the pcDNA3 vector or specific for each non-coding first exon amplified through Ex2 (29). Q-PCR assays were carried out as described above and normalized using the geometric mean of pcDNA3 vector or β-actin; semiquantitative RT-PCR was carried out as described above and PCR fragments amplified from Ex1A through Ex2 or Ex7.
1.5×106 PBMCs were washed and resuspended in PBS containing 2% heat-inactivated fetal bovine serum (FBS) and incubated with fluorochrome-conjugated mAbs to cell surface markers at 4°C for 20 min. Cells were washed with PBS and fixed in 2% formaldehyde (Fisher Scientific, Houston, TX). Cell populations were stained as follows: CD14-PE (monocytes), CD19-APC (B cells), CD8-APC (T cells), CD56-APC (NK cells), CD123-PE/BDCA2-APC (plasmacytoid dendritic cell, pDC), and BDCA1-PE/CD11c-APC (myeloid dendritic cell, MDC). All antibodies were from BD Pharmingen (San Jose, CA) except anti-BDCA1 and 2 (Miltenyi Biotec, Auburn, CA). For intracellular flow, cells were permeabilized with 0.1% Triton-X (Fisher Scientific) in PBS, washed with PBS/2% FCS and stained with rabbit anti-human IRF5 Abs (32.5 ng/μl; Cell Signaling) followed by FITC-labeled anti-rabbit IgG secondary Abs (10 ng/μl; BD Pharmingen). Isotype controls used rabbit IgG (R&D Systems, Minneapolis, MN) in place of anti-IRF5. After staining, cells were washed and fixed in 2% formaldehyde in PBS. Data were acquired using a FACSCalibur flow cytometer and analyzed using CellQuest software (BD Biosciences).
The unpaired two-tailed t-test or non-parametric Mann-Whitney U test was used (GraphPad or JMP7 software) to compare quantified IRF-5 and/or IFNα expression between two genotyped subgroups or healthy versus SLE patients depending on the data distribution. The two-way ANOVA was used to compare IRF-5 expression in immune cell subpopulations of genotyped patients. Linkage disequilibrium (LD) and haplotype frequencies were determined using the Haploview 4.1 software.
In humans, IRF-5 exists as multiple alternatively spliced isoforms, each having different effects on type I IFN gene expression (29). Six of these isoforms, identified by their utilization of distinct non-coding first exons (Ex1A, 1B and 1C) and the Ex6 indel (Fig. 1B), are expressed in PBMC of healthy donors and SLE patients (20,29). In order to determine whether IRF-5 expression levels differ in SLE patients versus healthy donors, we performed Q-PCR in purified PBMC. Fig. 2A shows that total IRF-5 expression was significantly upregulated in SLE patients compared to healthy donors (p=0.0084). Transcription from Ex1A (p=0.0103) and Ex1D (p=0.0365) was significantly upregulated in patient cells; transcription from Ex1C was low in cells from patients and controls while Ex1B was negligible.
Since transcription from Ex1A differed most between groups, we examined the profile of Ex1A-associated transcripts. RT-PCR from Ex1A through Ex7, which allows for amplification of the Ex6 indel, showed low levels of single IRF-5 isoforms in unstimulated healthy donor PBMC (Fig. 2B). Amplification of IRF-5 from SLE patients gave enhanced IRF-5 expression and alternative splicing that yielded multiple transcripts. Individual bands were cut, purified and sequenced to identify isoforms (Fig. 2B). A number of new isoforms from SLE patients, shown by the asterisks, have been cloned and sequenced that consist of intronic insertions (represented by high molecular weight bands). Similar results were found when analyzing Ex1C and 1D transcripts; however, these data varied more between patients (data not shown). The biological function of these new isoforms is currently unknown, but they appear to be expressed in SLE patients and not healthy donors.
Enhanced levels of small nuclear ribonucleoproteins (snRNPs) and Serine/Arginine-rich (SR) proteins, which are components of the spliceosome, have been detected in SLE patients (31–32). We determined by minigene assay whether overexpression of some of these components might contribute to alterations in IRF-5 transcription and/or alternative splicing in SLE. snRNP U1 or SR proteins ASF/SF2, SRp20, SRp40 and SFRS10/tra2Beta were co-transfected with healthy donor minigenes consisting of the IRF-5 5′-UTR or Ex5-7 to Hek293 cells lacking endogenous IRF-5 (13,33). Enhanced transcription from Ex1A and increased alternative splicing of Ex5-7 was observed upon co-transfection with distinct snRNPs or SR proteins (Fig. 2C). Co-transfection with U1 snRNP gave the most significant upregulation of IRF-5 expression from both minigenes suggesting its contribution to enhanced alternative splicing. All SR proteins and snRNPs were similarly expressed in Hek293 cells, as determined by immunoblotting (data not shown).
Few studies have extended past genetic associations to show functional significance of IRF5 gene variants (16–24). We focused on determining the biological function of two IRF5 variants independently associated with SLE, the CGGGG indel and rs10488631, and two variants correlated by LD with the CGGGG indel, rs2004640 and rs10954213 (24). LD and haplotype frequencies based on these four polymorphisms were determined in a larger cohort of genotyped SLE patients (n=131) from the rheumatology clinic at Uppsala University Hospital (24). The full risk haplotype H2 contains the risk alleles of the four polymorphisms – 4xCGGGG indel, the T-allele of SNP rs2004640, the A-allele of SNP rs10954213, and the C-allele of SNP rs10488631; H3 is the corresponding full protective haplotype with the non-risk alleles of the polymorphisms (Figs. 3A&B). IRF-5 expression in SLE patients carrying the H2 or H3 haplotype in homozygous form was analyzed by Q-PCR. Data in Fig. 3C demonstrate that total IRF-5 levels and transcription from each non-coding first exon was upregulated in SLE patients with the H2 (n=6) versus H3 haplotype (n=7). Lack of statistical significance is likely due to the small number of patients having these rare haplotypes (ANOVA p=0.17). Expression of total IRF-5 in patients carrying the H2 haplotype in homozygous form was increased by 33% as compared to patients homozygous for the H3 haplotype; expression from Ex1A, 1C, and 1D was increased by 38%, 84%, and 40%, respectively. Expression from Ex1B was undetectable.
Patients with varying haplotypes (H1–H9) were stratified by individual risk alleles to determine the independent effects of each variant on IRF-5 expression. Single marker analyses gave a trend of enhanced IRF-5 expression in patients with the risk alleles of the CGGGG indel or SNPs rs2004640 or rs10954213; stratifying by rs10488631 gave the only significant association (p=0.0155 for Ex1C) (Fig. 3D). Further stratification of each variant with one or two other variants did not provide additional significance; although combinations of the risk alleles of rs10488631 and the CGGGG indel or rs2004640 showed a suggestive association (data not shown).
To determine whether IRF-5 transcript levels in patients carrying the H2 haplotype correlate with protein expression, PBMC from SLE patients were stained with antibodies to extracellular cell surface molecules and intracellular IRF-5. Cell type-specific expression was analyzed by flow cytometry. Fig. 4A shows that IRF-5 expression was significantly upregulated in patients carrying the homozygous H2 versus the H3 haplotype (p=0.005). IRF-5 expression was increased by >25% in monocytes and MDCs from patients with the homozygous H2 haplotype, and to a lesser extent in B cells and PBMC. IRF-5 expression was also increased in pDCs from these patients (data not shown), yet this data was less consistent due to difficulties in obtaining sufficient cell numbers (30, 34). IRF-5 expression in monocytes gave the only significant difference between groups (p=0.037). We next examined the contribution of IRF5 risk alleles to increased protein expression in MDC and monocytes from genotyped patients. Data in Fig. 4B demonstrate that the risk alleles of the CGGGG indel, rs10954213 and rs10488631 contribute to enhanced IRF-5 expression in both cell types. Independent stratification of IRF-5 expression to the CGGGG indel or rs10954213 was not possible since all patients examined in this experiment carried the risk alleles of both variants. Additional samples were used for immunoblot analysis to confirm the contribution of individual variants to IRF-5 expression. Data in Fig. 4C show a significant association between IRF-5 protein expression and the risk alleles of rs10954213 and rs10488631. These data support an upregulation of both IRF-5 transcript and protein expression in patients with the H2 haplotype.
The difference in IRF-5 expression that was observed between healthy donors and SLE patients (Fig. 2A) was significantly larger than that observed between patients with the H2 and H3 haplotypes (Fig. 3C). To begin to address this discrepancy, we stratified IRF-5 expression by the protective alleles of the SNP rs10488631 and the CGGGG indel in healthy donors and SLE patients. We focused on these two variants since each was shown to be independently associated with SLE (24) and data presented here indicate a contribution to IRF-5 expression. We demonstrate that IRF-5 expression is significantly upregulated in SLE patients, independent of the risk alleles, providing strong evidence that additional factors/variants contribute to its expression (Fig. 5).
Analysis of IRF-5 expression in patients with H1–H9 haplotypes indicate that the risk allele of the 3′-UTR SNP rs10488631 is the primary independent contributor to IRF-5 expression; however, we can not rule out an effect of 5′-UTR polymorphisms since patients carrying only the risk allele of rs10488631 are rare. In order to determine whether the CGGGG indel and/or rs2004640 function to regulate IRF-5 gene transcription independent of 3′ SNPs, we used the minigene reporter assay (24). 5′-UTR constructs were generated from patients with the risk alleles at the CGGGG indel and rs2004640 (H1/H2 haplotypes) or the protective alleles at the CGGGG indel and rs2004640 (H3/H4 haplotypes) (Fig. 3B). Similar to data from RT-PCR analysis of IRF-5 expression in healthy donor PBMC (29), only low levels of Ex1A-associated transcripts were detected in cells transfected with healthy donor minigenes (Supplementary Fig. 1). Analysis of IRF-5 Ex1-specific utilization revealed a distinct expression pattern where only Ex1A and 1D transcripts were significantly upregulated in patients with the H1/H2 haplotype (Fig. 6A). Transcription from Ex1B was not detected in constructs with either protective or risk genotypes; Ex1C transcripts were detected, albeit to a low extent. Minigene data from two patients homozygous risk for the CGGGG indel and protective for rs2004640 (H7/H8) indicated a differential effect on transcription from Ex1A or 1D. Additional samples will be required to define independent function; current results suggest that the CGGGG indel is associated with transcription from Ex1D, while rs2004640 defines transcription from 1A (Fig. 6A and data not shown).
While primary genetic variation may be causal to SLE, high levels of serum IFNα activity are a heritable risk factor (5) and may contribute to IRF-5 expression (Fig. 2A). We have demonstrated previously that IRF-5 transcription from Ex1C, but not 1A or 1B, was regulated by type I or II IFNs through a functional IFN-stimulated response element in its promoter (29). To determine whether 5′-UTR risk variants alter IRF-5 transcriptional regulation by IFNα, cells were treated with IFNα. Ex1C-associated transcripts were upregulated independent of genotype (Fig. 6B). Ex1D utilization was also enhanced by IFNα, yet this was genotype-dependent, as the IFN response was negligible in patients with protective H3/H4 haplotypes (Fig. 6C). Ex1A and 1B utilization was unchanged after treatment (data not shown). To determine whether a specific IRF-5 haplotype was associated with IFNA expression in SLE patients, Q-PCR was performed along with the serum IFNα immunoassay. Given the small number of samples examined, only a suggestive correlation between the CGGGG risk allele and increased IFNA transcripts and IFNα protein expression was observed (Fig. 6D).
The recent identification of IRF5 as an SLE susceptibility gene has provided a critical link between the type I IFN pathway and disease pathogenesis. Given the role of IRF-5 in host defense and in type I IFN gene regulation (6–9), it is reasonable to hypothesize that genetic variants associated with IRF-5 expression may contribute to increased IFNα levels in SLE patients. Here, we demonstrate a direct association of IRF-5 transcript levels with SLE. Total IRF-5 expression and expression from non-coding Ex1A and 1D were significantly upregulated independent of genotype in primary PBMC from SLE patients compared to healthy donors (Fig. 2). IRF-5 alternative splicing was also enhanced in SLE patients (Fig. 2B) providing the first evidence that IRF-5 isoforms specific to SLE may be identified. IRF-5 exists as multiple alternatively spliced isoforms, each with distinct cell type-specific expression, regulation, and function (29). Previous studies have shown that the identical polypeptides V3 (Ex1C) and V4 (Ex1A) are the primary inducers of virus-mediated IFNα (29). These isoforms are highly expressed in pDC of healthy donors. The mechanism(s) of enhanced IRF-5 alternative splicing in SLE is not known but data presented here provide initial evidence that U1 snRNP, SRp20 and SRp40 may be involved. snRNPs are components of the spliceosome and are essential for the removal of introns; SR proteins are critical splicing factors involved in regulating/selecting splice sites. snRNPs, specifically U1 snRNP, are the major SLE autoantigens in addition to dsDNA (35–36). U1 snRNP was recently shown to induce robust type I IFN (36–37). The ability of U1 to alter the utilization of Ex1A and splicing of Ex5-7 suggests that it may also be a critical factor in regulating IRF-5 alternative splicing. Equally important, SRp20 and p40 have been shown to be autoantigens in patients with SLE (32,38) and overexpression in our minigene assays showed enhanced transcription and alternative splicing of IRF-5. Additional experiments are necessary to confirm an association of autoantigen levels in serum of SLE patients with IRF-5 expression and/or alternative splicing.
Significant effort has gone into the replication and expansion of IRF5 genotype data in SLE (16–22). Multiple risk and protective haplotypes have been predicted, yet information is significantly lacking regarding the functional consequence of these polymorphisms. We identified an IRF5 risk haplotype, containing the risk alleles of SNPs rs2004640, rs10954213, rs10488631, and the CGGGG indel (24), that in part explains the association between SLE and elevated IRF-5 expression. Transcripts were increased 30–85% in patients having the risk (H2) versus protective haplotype (H3) in homozygous form. The most significant correlation with the H2 haplotype came from utilization of Ex1C. Contrary to published data by Graham et al. (19), detection of Ex1B-associated transcripts was negligible in either healthy donors or SLE patients. Our data instead support findings by others indicating an overall low utilization of Ex1B (20,29). Transcription from Ex1C is generally negligible in unstimulated healthy donor PBMC, yet stimulation with IFNα leads to enhanced expression (29). We determined that the risk alleles of rs10488631, and not the other variants, were independently associated with increased IRF-5 expression from Ex1C (20,22). SNPs rs2004640 and rs10954213 have been associated with IRF-5 transcription in HapMap samples; however, neither were independently associated (19–20,22,39–40). It is difficult to determine the biological significance of single polymorphisms since patient samples contain multiple polymorphisms that may or may not be associated with the disease. Furthermore, SLE patients that are homozygous for a single risk variant are rare. Indeed, the CGGGG polymorphism accounts for the association signal observed from rs2004640 and rs10954213 (24). Minigene reporter assays demonstrated that risk alleles of both the CGGGG indel and rs2004640 were associated with increased transcription from Ex1A and 1D (Fig. 6A). The fact that neither of these variants gave independent association with IRF-5 expression may be due to their being masked by risk alleles of rs10954213 or rs10488631, or due to a genuine joint effect by these alleles.
By FACS and immunoblot analysis, we show a direct correlation between increased IRF-5 transcription and protein expression in blood cells of SLE patients (Fig. 4). Similar to Q-PCR data, the risk allele of rs10488631 was the best independent predictor of IRF-5 protein expression; the rs10954213 risk allele was also significantly associated. Observed differences between IRF-5 transcript and protein expression associated with rs10954213 may be due to its function in stabilizing IRF-5 proteins (22).
While our data support an association of the H2 haplotype with enhanced IRF-5 expression in SLE, these may not be the only variants contributing to expression; other factors independent of genotype may also contribute. Stratification of IRF-5 expression by rs10488631 and the CGGGG indel in healthy donors and SLE patients support this idea (Fig. 5). Differences in expression between these groups could be a result of functional variants that have yet to be identified. The other possibility is that disease-associated factors contribute to IRF-5 expression. Niewold et al. (23) reported that an IRF5 risk haplotype (rs2004640, rs3807306, rs10488631, and rs2280714) was associated with increased serum IFNα activity. High IFNα levels detected in SLE patients could be a factor contributing to enhanced IRF-5 expression (11,29). Minigene experiments confirmed that both genotype and soluble IFNα contribute to IRF-5 expression (Figs. 6B & C). Patients with SLE have long been known to display elevated type I IFN in their serum. We provide the first detailed evidence supporting a role for the IRF5 risk haplotype in directing expression of both IRF-5 and IFNα, whereby IRF-5 expression is linked to circulating IFNα levels in patients.
Figure 1. Minigenes generated from healthy donors express low levels of Ex1A-associated transcripts. RT-PCR analysis of an in vitro generated IRF5 minigene construct from genomic DNA of a healthy donor. Human Hek 293 cells were transiently transfected with healthy donor control (minigene) or left untransfected or transfected with empty vector (mock). Total RNA was isolated and Ex1-specific forward primers and Ex2 reverse primers were used for PCR amplification of IRF-5. H2O, neg. control; +, positive control DNA.
Table I. Primers used for PCR amplification
We thank Anne Trönnberg for excellent technical assistance and Rezvan Kiani for collecting blood samples. This work was supported by grants from the National Institute of Health (NIH)/National Institute of Arthritis and Musculoskeletal and Skin Diseases (NIAMS; 5R03AR054070) and the Arthritis Foundation (to B.J.B.); The Alliance for Lupus Research, the Swedish Research Council, the Dana Foundation, the Swedish Rheumatism Association, and the Gustafsson Foundation and the King Gustaf V 80th Birthday Foundation (to L.R.); the Swedish Research Council for Medicine and Knut and Alice Wallenberg Foundation (to A-C.S.).