Genetic Studies on Type I Interferon Pathway

REVIEW A compass that points to lupus: genetic studies on type I interferon pathway

C Kyogoku1,2 and N Tsuchiya3 1Department of Rheumatic and Autoimmune Diseases, University of Minnesota, Minneapolis, MN, USA; 2Division of Molecular Pharmacology and Pharmacogenomics, Kobe University Graduate School of Medicine, Kobe, Japan and 3Doctoral Program in Social and Environmental Medicine, Graduate School of Comprehensive Human Sciences, University of Tsukuba, Tsukuba, Japan

It was more than 20 years ago that patients with systemic lupus erythematosus (SLE) were first reported to display elevated serum levels of type I interferon (IFN). Since then, extensive studies revealed a crucial role for type I IFN in SLE pathogenesis. The current model proposes that small increase of type I IFN production by plasmacytoid dendritic cells (pDCs) is sufficient to induce unabated activation of immature peripheral DCs. IFN-matured DCs select and activate autoreactive T cells and B cells, rather than deleting them, resulting in peripheral tolerance breakdown, a characteristic feature of SLE. Furthermore, immune complexes provide an amplification loop to pDCs for further IFN production. In the past 5 years, high-throughput technologies such as expression profiling and single-nucleotide polymorphism (SNP) typing established the role of altered type I IFN system in SLE, and a detailed picture of its molecular mechanisms is beginning to emerge. In this review, we discuss two major lines of genetics studies on type I IFN pathway related to human SLE: (1) expression profiling of IFN-responsive genes and (2) disease-associated SNPs of IFN-related genes, especially IRF5 (IFN-regulatory factor 5). Lastly, we discuss how such genetic alterations in type I IFN pathway fit in the current model of SLE pathogenesis. Genes and Immunity (2007) 8, 445–455; doi:10.1038/sj.gene.6364409; published online 21 June 2007

Keywords: systemic lupus erythematosus; type I interferon; interferon signature; polymorphism, interferon regulatory factor 5; genetics

Introduction/historical perspective number of cases with SLE (0.7%).5,6 Another study reported that some patients treated with IFNa developed Systemic lupus erythematosus (SLE) is a chronic, other immune syndromes such as type I diabetes, inflammatory autoimmune disease that affects multiple psoriasis, inflammatory arthritis and Sjo¨gren syndrome,7 organs and is characterized by production of auto- which may share genetic background with SLE.8,9 antibodies to nuclear antigens (for example, histones, Conclusive evidence for a crucial role for IFN in SLE ssDNA, dsDNA, Ro, La) and immune complex (IC) pathogenesis has been provided over the past 5 years, formation. SLE is a complex disease that involves with the help of remarkable progress in comprehensive multiple, interacting genetic and environmental factors, gene-hunting technologies, such as microarray and high- the identification of which is a challenging task. To date, throughput SNP (single-nucleotide polymorphism) typ- only limited success has been achieved. ing. Using gene expression microarray, we have filed a Data linking type I interferon (IFN) to SLE were list of IFN-regulated mRNA transcripts that are ex- presented as early as 1979, when Hooks et al.1 first pressed at elevated levels in the blood cells of SLE observed elevated levels of IFN in SLE sera. Since then, patients.10–12 Furthermore, protein microarray experi- multiple literatures have supported this observation.2–4 ments revealed IFN-inducible protein expression profile In the later years, role of type I IFN in SLE was more in serum from SLE patients.13 High-throughput SNP clearly implicated by a series of studies in humans who typing platforms such as Illumina (Illumina Inc., San received IFNa as a therapeutic agent for viral hepatitis or Diego, CA, USA) and Sequenom (Sequenom Inc., San carcinoid tumors. Nearly a quarter of IFNa-treated Diego, CA, USA) enabled us to carry out B0.5 million subjects (22%) developed a positive blood test for genotypes a day with high fidelity. Together with a great antinuclear antibodies, and one in five (19%) developed effort in the recruitment of DNA samples from patients various symptoms of autoimmunity, including a small and controls, an increasing number of large-scale genetic association studies are being carried out. These recent advances in science and technology have led us to Correspondence: Dr N Tsuchiya, Doctoral Program in Social and understand better the role of type I IFN system in SLE. Environmental Medicine, Graduate School of Comprehensive Human Sciences, University of Tsukuba, 1-1-1 Tennodai, Tsukuba, 305-8575, Japan. Type I IFN in human SLE E-mail: [email protected] Received 21 May 2007; accepted 22 May 2007; published online 21 Type I IFNs were first identified as a cytokine to enhance June 2007 mammalian antiviral resistance.14 Viral infection rapidly Genetics of type I IFN in human SLE C Kyogoku and N Tsuchiya 446 induces leukocyte production of IFNa and IFNb, which therefore, may have an important role in activation of have a central role in host defense by triggering autoreactive T and B cells, together with TLR7/8/9 that inflammatory and antiviral responses. They also influ- are also expressed on B cells. ence cell proliferation, differentiation and survival of mature lymphocytes, class switching at immunoglobulin heavy chain loci, and activation of dendritic cells (DCs).15 Type I IFN in mouse lupus models Type I IFN genes cluster at human chromosome 9p22. In New Zealand black (NZB) X New Zealand white a b k There are at least 13 IFN genes, as well as IFN , IFN , (NZW) F1 female mice, polyinosinic polycytidylic acid t o 16 IFN and IFN , together comprising the type I IFNs. (poly I:C), a surrogate for dsRNA, accelerates the All type I IFN family members bind to common receptor development of lupus.39 Its parental strain NZB also complexes, IFNAR1 and IFNAR2, and initiate a JAK/ expresses lupus-like disease. Deletion of the IFNa/b STAT signaling cascade leading to the transcription of 17 receptor in NZB mice shows abrogation of a number of specific IFN-inducible genes. immune system alterations including autoantibody What causes elevated type I IFN expression observed production.40–42 in SLE? A strong candidate is IC. SLE patients have been Fas-defective MRL/lpr is another established lupus shown to exhibit increased levels of apoptosis as well as model.43 Injection of poly I:C in B6 lpr mice resulted in decreased clearance of apoptotic cell fragments, such as 18–20 increased severity of renal disease, higher titers of DNA and RNA, from plasma. Such endogenous autoantibodies, increase in serum Ig and accumulation immunostimulatory chromatin can drive autoantibody of activated lymphocytes. In addition, introduction of production and lead to generation of chromatin-contain- 21–26 homozygous mutation for type I IFN receptor resulted in ing ICs. Ronnblom et al. have investigated the dramatic decrease of ICs’ deposition in the kidney and a capacity of ICs in SLE sera to trigger IFN production reduced lymphadenopathy.41 by plasmacytoid dendritic cells (pDCs), which is the These observations support the importance of type I primary source of type I IFNs. They provided evidence IFN and are compatible with our understanding on that autoantibodies associated with apoptotic DNA and human SLE. However, significant differences are also RNA, particularly single-stranded CpG sequences in known between human SLE and mouse lupus models. In genomic DNA, can provide a stimulus that results in murine lupus, importance of IFNg, a type II IFN, has a IFN secretion. They also showed that internalization of been more highlighted, particularly at the site of g g g the ICs by Fc receptor II, but not Fc RI or Fc RIII, is lymphocyte and inflammatory cell infiltration in the required for most effective activation of IFNa production 43–45 27 kidney. On the other hand, in humans, peripheral by pDCs. blood expression profile appears to be preferentially Secondly, a virus could be an exogenous trigger for induced by IFNa, but not by IFNg,12,46 although a role of IFN pathway activation in SLE. After virus infection, IFNg as an inducer of organ damage cannot be excluded. unabated activation of IFN pathway, possibly caused by Data from murine lupus models have also supported a gene polymorphisms, can lead to differentiation of genetic contribution of IFN-regulated genes to lupus autoreactive B cells into antibody-secreting plasma 28 susceptibility. IFN-inducible protein 202 gene (ifi202)at cells. The antibodies cause amplification of chromatin- the Nba2 locus on chromosome 1 is expressed at high containing IC formation and may result in elevated type levels in NZB mice because of promoter region poly- I IFN production. morphisms.47 IFI202 is a transcription factor that influ- Toll-like receptor (TLR) is a key molecule in type I IFN ences proliferation, survival and differentiation of B cells signaling and has been a matter of interest for SLE 48 29 and induced by both IFNa and IFNg. Such an example researchers since its discovery in 1997. To date, 10 indicates the need to investigate human IFN-inducible members of human TLRs have been identified. TLRs genes as well as genes in its signaling pathway as recognize specific endogenous and exogenous microbial candidates for susceptibility genes to human SLE. ligands such as genomic DNA, single-stranded RNA (ssRNA), double-stranded RNA (dsRNA), lipopolysac- charide (LPS), flagellar proteins, lipopeptides and trigger Expression profiling of IFN-responsive activation of type I IFN pathway, which result in the eradication of invading pathogens.30,31 Among the TLR genes in SLE family, those that are expressed on pDCs (also known as Although reliable quantification of type I IFN-related interferon-producing cells) and trigger cell activation by gene products in SLE patients has been challenging both DNA, ssRNA or dsRNA are attractive candidates for at mRNA and protein levels, several groups have transducers of the elevated type I IFN signals in SLE. succeeded in providing a sensitive and relatively Human pDCs predominantly express TLR6, TLR7 and quantitative measure using microarray technology. TLR9;32–34 however, TLR6 does not induce IFNa/b.30,32 Comprehensive profiling of gene expression in peri- Therefore, TLR 7/8 (TLR8 is phylogenetically close to pheral blood mononuclear cells (PBMCs) and protein TLR7) that recognizes ssRNA35 and TLR9 that recognizes expression in sera from patients has consistently shown CpG DNA,36 are most relevant players in pDCs. In fact, the upregulation of IFN-inducible genes in SLE com- mouse studies showed that single-stranded CpG se- pared with normal controls.10,11,13,49–52 quences bind to toll-like receptor 9 (TLR9) expressed on B cells and DCs and cause activation and secretion of IFN signature in PBMCs various cytokines including type I IFNs.36,37 Baechler et al.10 studied gene expression microarray TLR3 is another member of the TLR family that profiles in 48 SLE patients and 42 controls to identify recognizes chromatin. It is stimulated by dsRNA and pathways that might be dysregulated in PBMCs of expressed in B and T cells but not in pDCs,38 and SLE patients. Several immune-related genes were up-

Genes and Immunity Genetics of type I IFN in human SLE C Kyogoku and N Tsuchiya 447 regulated in the SLE patients including Fc receptor for Authors finally hypothesized that high levels of IgA (FCAR, CD89), Fc receptor for IgG such as FcgRIIA systemic chemokines in active SLE, driven by type I (CD32) and FcgRI (CD64), TNFRSF6 (Fas/CD95), and IFN, lead to a state of ‘chemokine confusion’ that alters three molecules in the inflammatory IL-1 cytokine the normal trafficking and chemotaxis of leukocytes in pathway: IL-1b, IL-1 receptor II (IL-1RII) and IL-1 the body, contributing to the widespread inflammatory receptor antagonist. Notably, genes in the cluster of type immune responses observed in SLE. I IFN pathways (referred to as ‘IFN signature’) were upregulated in over half the patients (referred to ‘IFN- high’), whereas the others showed comparable levels to SLE-associated variations in IFN pathway the majority of controls (IFN-low). Most of the control genes subjects expressed low levels of IFN signature genes. In order to determine which genes in this cluster could It is not clear whether the type I IFN signature is a result be regulated by IFN treatment in vitro, PBMCs were of a dysregulated immune system or rather it reflects isolated from four healthy donors and cultured for 6 h genetic variation that is causally involved in the after stimulation with IFNa/b or IFNg. Overall, 23 of the pathogenesis of human SLE. In addition to the well- 161 genes that distinguish lupus patients from healthy established pathogenic role of genetic variation in major controls, were found to be IFN regulated,10 and 22 IFN histocompatibility complex, a number of rare and signature genes were strikingly upregulated in ‘IFN- common variants, such as SNP, deletion/insertion and high’ SLE patients.53 Among these, at least nine genes copy number variations are known to be important for (LY6E, OASL, IFIT1, IFIT4, STAT1, MX1, MX2, PLSCR1 disease pathogenesis. To identify such genetic variations, and IRF7) were replicated by other independently one of the conventional ways is to target genes in cell- carried out microarray studies.11,49–51 signaling pathways involved in a range of immuno- Baechler and co-workers10,53 also examined whether logical events. So far, several SLE-related gene polymorphisms are discovered in immune-signaling pathways, there was any clinical feature that distinguishes ‘IFN- 58 high’ group of patients from ‘IFN-low group’. The such as PDCD1 (PD-1) in the RUNX pathway, PTPN22 frequency of SLE who have a manifestation of renal (protein tyrosine phosphatase N22) in the T cell-signaling pathway,8 Fcg receptor II and III in immunoglobulin- disease, central nervous system lupus and hematological 59–62 deficit (particularly leukopenia, lymphopenia and dependent cell-signaling pathway and CR2 in complement pathway.63 thrombocytopenia) was significantly higher in the ‘IFN- 64 high’ group of patients as compared to the ‘IFN-low’, Recently, Sigurdsson et al. analyzed 44 SNPs in 13 suggesting that IFN signature can be a marker for genes from the type I IFN pathway in 679 Swedish, patients with severe SLE. Finnish and Icelandic patients with SLE, in 798 un- affected family members and in 438 unrelated control IFN signature in serum protein individuals for joint linkage and association with SLE. They identified association in TYK2 (tyrosine kinase 2) Inspired by the transcriptome analyses, Bauer et al.13 IRF5 P examined the levels of 160 serum proteins in 15 IFN- and (IFN-regulatory factor 5) with adjusted values of 1.1 Â10À5 for TYK2 rs2304256 SNP and high, 15 IFN-low SLE patients and 15 healthy controls Â À6 IRF5 using a high-throughput protein microarray technology, 7.9 10 for rs2004640 SNP. TYK2 binds to the type I IFN receptor complex with JAK1 and initiate a Luminex. Luminex assays are based on custom dual- antibody sandwich immunoassay arrays using xMAP JAK/STAT signaling cascade leading to the transcription of IFN signature genes.65 IRF5 is a member of IRF family technology (multi-analyte profiling beads) enabling the that plays an important role at downstream of the TLR– detection and quantification of multiple protein targets MyD88 signaling pathway as a master transcription simultaneously.54,55 Of the 160 analytes measured, 30 factor leading to the transcription of genes for inflam- showed significant differences between the ‘IFN-high’ a b 66,67 SLE group and controls. The extent of expression level matory cytokines as well as IFN and IFN . These findings were the first evidence that gave us an idea that difference was bigger in ‘IFN-high’ SLE than in ‘IFN-low SLE’; however, only 11 of 30 analytes were significantly type I IFN signature was, at least in part, controlled by gene polymorphisms and was causally involved in SLE different between ‘IFN-high’ and ‘IFN-low’ SLE. Nota- bly, 12 chemokines were included in the upregulated pathogenesis, rather than just a feature of disease phenotype. analytes (CCL2, CCL3, CCL7, CCL8, CCL17, CCL19, CXCL2, CXCL8, CXCL9, CXCL10, CXCL11 and CXCL13). Most of these chemokines (9/12) were tran- IRF5 scriptionally inducible by type I IFN in PBMC. In addition to IFN gene score used for defining ‘IFN- The human IRF family comprises nine members from high’/’low’,10 correlation between chemokine levels and IRF1 through IRF9. Each IRF contains a conserved DNA- various clinical measures/features of disease activity binding domain that contains a helix–turn–helix motif of was shown. For example, many of the IFN-regulated approximately 120 amino acids. This region recognizes a analytes showed strong positive association with the consensus DNA sequence known as the IFN-stimulated SLEDAI56 and the SLAM-R,57 two commonly used response element called ISRE (GAAANNGAAA[G/C] measures of global disease activity. CXCL11 (I-TAC), [T/C]), which was initially known as the IRF enhancer CXCL13 (BLC), CXCL10 (IP-10) and CCL3 (MIP-1A) (IRF-E). ISREs are found in the promoters of various IFN were observed at significantly high levels in serum of signature genes such as IFNa, IFNb, IL-12, CXCL10 and patients with renal disease. These data suggested that the IFIT1.68 Expression of IRF5 was detected in B cells and levels of serum chemokines might serve as convenient DCs69 and further enhanced by activation of type I IFN biomarkers for disease activity in SLE. pathway because promoter of IRF5 itself also possesses

Genes and Immunity Genetics of type I IFN in human SLE C Kyogoku and N Tsuchiya 448 ISRE.70 Among the IRF family members, IRF5 and IRF7 pathway. This finding represents an excellent example in are known to share the same signaling pathway that is which gene copy number variation plays an important initiated through TLR7/8 and TLR9. Unlike other TLRs role in disease pathogenesis, although the contribution of that are expressed on cell surface, TLR7/8 and TLR9 are the duplicated gene other than TLR7 requires further expressed in endosome and stimulated by internalized investigation. ssRNA and dsDNA, respectively.66 Efficient stimulation The mice deficient for IRF5 gene (Irf5À/À mice) were of TLR7/8/9 requires engagement of FcgRII or FcgRIII, initially reported to show severely impaired production carries of ICs from cell surface to endosomal compart- of pro-inflammatory cytokines in pDCs when stimulated ment.27,71–73 by various TLR ligands; however, IFNa induction was IRF5 is a cytoplasmic protein anchored on endosome normal.85 However, a recent report from the same by binding to MyD88. After ligation of TLR7/8/9 with investigators showed that production of IFNa and IFNb specific chromatin ligands, the MyD88-bound IRF5 is in splenic macrophages as well as serum IFN levels in activated by TRAF6 by an as-yet-unknown mechanism. response to viral infections were markedly decreased in Activated IRF5 dimerizes and translocates to the nucleus Irf5À/À mice, supporting the role of IRF5 in type I IFN led by NLS (nuclear localization signal), where it binds to production also in mice, although the contribution might ISRE in the promoter sequences and activates transcrip- be stimulation specific and cell type dependent.86 tion of IFNa/b and pro-inflammatory cytokines such as TNFa, IL-12 and IL-6, probably in cooperation with NF- kB.66,74 Another IRF family member IRF4 also binds to Association of IRF5 gene polymorphisms MyD88 at an overlapping region with IRF5-binding site, thereby serving as an antagonist of IRF5 by inhibiting with SLE the binding of IRF5 to MyD88 and attenuating the Association of IRF5 intron 1 SNP with SLE TLR–MyD88 signaling.66,75 In an effort to replicate the association of IRF5 Barnes et al.76 carefully examined the function of IRF5 rs2004640T with SLE,64 Graham et al.87 genotyped four by in vitro experiments using human cell lines. They sets of SLE cases and controls from the United States, showed that IRF5-mediated activation is virus specific. Spain, Sweden and Argentina (total of 1661 cases and Infection by Newcastle disease virus (NDV), vesicular 2508 controls) as well as 470 families with SLE. They stomatitis virus (VSV) and herpes simplex virus type I found a significant increase of the T-allele frequency in (HSV-1), but not Sendai virus, can activate IRF5, leading the patients; overall, four-cohort analysis showed 60.4% to induction of IFNa. Knockdown of IRF5 by siRNA in in cases versus 51.5% in controls (P ¼ 4.4 Â 10À16). In 470 2fTGH human fibroblast cell line reduces type I IFN SLE pedigrees, T allele showed significant overtransmis- induction in response to TLR7 ligand.77 IRF5 competes sion using the transmission disequilibrium test (219:153 with IRF7 in the formation of homodimers, and IRF5/ T:U, P ¼ 0.0006).87 IRF7 heterodimers result in a repression rather than When these data were combined with the previously enhancement of IFNa transcription.78 By means of published data from Swedish and Finnish cohorts,64 microarray experiments in BJAB human B cell lines, a robust and consistent association of the rs2004640T they demonstrated that 657 genes (568 upregulated and allele with SLE was observed. Overall odds ratios (OR) 89 downregulated) were modulated by IRF5 by at least showed 1.47 (95% confidence interval: 1.36–1.60), with twofold.79 These genes are considered to be either P ¼ 4.2 Â 10À21. The individuals carrying two copies of upregulated by direct binding of IRF5 to its promoter rs2004640T allele were at higher risk for SLE (OR ¼ 2.01, or secondarily upregulated by type I IFN, produced by P ¼ 3.7 Â 10À14) than individuals carrying one copy IRF5 activation. (OR ¼ 1.27, P ¼ 0.0031), suggesting a dosage effect of this allele.87 Graham et al. next investigated the potential function IRF5 in murine lupus of the rs2004640T allele. As shown in Figure 1, rs2004640T allele is located 2 bp downstream of the In addition to recent reports on human SLE, insights intron–exon border of exon 1B, creating a consensus GT from mouse studies also support the importance of gene donor site for alternative splicing. The splicing of IRF5 is polymorphisms in TLR–MyD88 signaling pathway in known to be highly complex.70 Transcription of multiple lupus pathogenesis. FcgRIIBÀ/À mice with B6 back- IRF5 isoforms is initiated at one of the three promoters, ground develop spontaneous SLE-like disease, charac- giving rise to transcripts starting either exon 1A, exon 1B terized by the production of autoantibodies and or exon 1C. In addition, the small sequences of insertion/ development of glomerulonephritis. Addition of the deletion variation, especially seen around exon 6, Yaa modifier to the FcgRIIB-deficient mice increases together with short/long length variation of 30-UTR, severity of disease.80,81 Yaa was first identified on Y produces in total 11 mRNA isoforms. These isoforms are chromosome derived from the SB/Le strain as an known to have different ability to stimulate production autoimmune accelerator;82 however, what elements of of IFNa and IFNb, and different expression pattern. Yaa genome are responsible for severe lupus had not Transcripts initiated at exon 1A and exon 1B are been known. Recently, Pisitkun et al.83 and Sabramanian constitutively expressed in pDCs and B cells, whereas et al.84 demonstrated that genomic duplication of X transcripts bearing exon 1C are inducible by type I chromosomal region encompassing TLR7 gene locus IFNs.70,87 Graham et al. demonstrated that individuals onto Y chromosome accounts for Yaa mutation. Thus, who are homozygous for the rs2004640 G allele male mice carrying Yaa express double dose of TLR7 expressed IRF5 isoforms containing exon 1A and exon gene, one from original X and another from duplicated Y 1C, but not exon 1B. In contrast, individuals with the chromosome, which leads to enhanced TLR7-signaling rs2004640T allele expressed transcripts containing exon

Genes and Immunity Genetics of type I IFN in human SLE C Kyogoku and N Tsuchiya 449

Figure 1 Three potentially functional IRF5 variations associated with SLE in Caucasians. There are three potentially functional variations in IRF5 that are associated with SLE in Caucasians: (1) exon 1b splicing donor site, (2) 30 amino acid indel of exon 6 PEST domain and (3) polyadenylation site in 30-UTR. These three variations define the risk haplotype 1 and protective haplotype 4 and 5 for SLE. Association of haplotypes with SLE were examined by meta-analysis using 555 complete trio pedigrees from United States and United Kingdom, and total 2188 cases and 3596 controls from United States, United Kingdom and Sweden.91

1B as well as other transcripts containing exons 1A and indel. Newly identified variants were genotyped in the 1C. These results implied that rs2004640T represents a HapMap CEU samples, and by combining the data with strong genetic risk factor for SLE probably through the the publicly available SNP genotype data from the influence on exon 1B splicing.87 However, since exons International HapMap Project, we were able to select 25 1A, 1B and 1C constitute 50-UTR, the function of these tagSNPs based on r240.8 and minor allele frequency different isoforms needs to be elucidated. Recent paper 41% threshold.92 Then we tested association of tagSNPs suggested that the translation efficiency of the isoform with disease risk in 555 trio and sib-pair SLE families containing exon 1B is low, making the interpretation of from the United States and the United Kingdom. After the molecular mechanisms difficult.88 extensive conditional logistic regression analysis (http:// pngu.mgh.harvard.edu/purcell//whap/), SNPs in three Search for the causative polymorphism in IRF5 distinct regions of IRF5 gene came out as an independent Sigurdsson et al.64 and Graham et al.87 focused on a risk factor for SLE. One was the region where rs2004640 limited number of SNPs already deposited in the is located, and rs2004640T allele that influences exon 1B database. In order to gain further insight into the splicing itself was the most likely causal variation. molecular mechanism of association, several groups The second region was within 30-UTR represented of investigators, including our own, have started to by rs10954213. The third region was represented by resequence the entire IRF5 genome to identify the rs2070197 SNP, which was B1.6 kb downstream of exon polymorphism that plays a causative role. 6 indel variation.91 In an entirely different research project, Cheung et al. carried out genome-wide testing of the 1 million HapMap phase I variants for association with gene IRF5 polyadenylation (poly A) site SNP was the second expression in Epstein-Barr virus (EBV)-transformed B genetic risk factor for SLE in Caucasians cells.89,90 They found 10 significant markers spanning Among four SNPs in 30-UTR region of IRF5, we found 90 kb of chromosome 7 where IRF5 gene is located. two SNPs (rs10954213 and rs10954214) that were in Among them, rs2280714 was identified as the strongest strong LD with rs2280714. Since rs2280714 was shown to cis-acting determinant of IRF5 expression. This finding be associated with high gene expression of IRF5,89,90 we led us to ask whether elevated expression marked by next examined the correlation between these 30-UTR rs2280714 is associated with the risk to SLE. Given that SNPs and IRF5 expression level. The mRNA expression rs2280714 is well downstream of IRF5 (B5 kb down- data of 210 HapMap samples (CEU, CHB, JAP and YRI stream of IRF5 30-UTR), and that its position is not even populations) collected at Sanger Institute (http:// in a recognizable regulatory region, we hypothesized www.sanger.ac.uk/humgen/genevar/) and mRNA ex- that there may be additional genetic variation in tight pression data set of 233 CEPH samples89 both revealed linkage disequilibrium (LD) with rs2280714 that controls that rs10954213 SNP was most strongly associated with the expression phenotype. gene expression among all the tagSNPs encompassing In order to more fully characterize genetic variation in the entire IRF5 gene. rs10954213 A allele was signifi- IRF5, we sequenced the exons, introns and 50 flanking cantly correlated with higher mRNA expression. There- region up to 1 kb.91 Most of the subjects were Caucasians. fore, rs10954213 was concluded to be the best predictor In total, 52 variants were observed, of which 20 were of IRF5 expression, and previously detected rs2280714 previously identified. Although none of the identified was considered to be a proxy. However, expression SNPs was a missense variation with a minor allele analysis using whole blood cell of SLE cases provided frequency 41%, we observed a 30 bp insertion/deletion evidence that rs2004640 also contributes to high level of (indel) of exon 6, which lead to 10 amino acid inframe IRF5 expression to some extent.91

Genes and Immunity Genetics of type I IFN in human SLE C Kyogoku and N Tsuchiya 450 If an SNP allele that is best correlated with high gene Exon 6 indel variation was the third genetic risk factor for SLE expression is also associated with disease risk, it makes in Caucasians perfect sense. That was indeed the case with rs10954213 By sequencing IRF5 genome, we observed 30 bp (10 in Caucasians. Our case–control study in the Caucasians amino acid) exon 6 indel variation (Figure 1). The exon 6 revealed that the A allele frequency was 67.2% in SLE indel was located in a proline-, glutamic acid-, serine- and 61.2% in control (P ¼ 9.1 Â10À5). and threonine-rich (PEST) domain of IRF5, a motif Next we examined the function of rs10954213 to previously implicated in protein stability and function explain how SNP affects the gene expression. By looking in the IRF family of proteins.97 As described above, we at the sequence, we realized that rs10954213 was located observed that rs2070197 confers the third genetic risk of within an evolutionarily conserved poly A signal in IRF5 for SLE, but we could not see any relevant the 30-UTR region of IRF5. The substitution of G for A functional importance in this as well as any other SNPs was predicted to disrupt a poly A þ signal sequence in the neighboring region. Therefore, we re-analyzed the (AAUAAA4AAUGAA). The AAUAAA motif in 30-UTR genetic data based on the hypothesis that rs2070197 is is well known to serve as a binding site for CPSF tagging other functional variation; most likely exon 6 (cleavage and polyadenylation specificity factor). During indel because of its potential function and relatively close RNA polymerase II transcription, CPSF binds to the location to it (B1.6 kb upstream). Finally by TDT analysis AAUAAA sequence, and together with cleavage stimu- in 555 trio and sib-pair SLE families from the United lation factors, forms a complex that cuts the mRNA States and the United Kingdom, we observed association strand 10–30 bp downstream of the poly A þ signal, and of exon 6 insertion with higher genetic risk for SLE only initiates polyadenylation of the transcripts.93–95 There- when conditioned on the exon 1B splicing SNP fore, we hypothesized that rs10954213 might influence (rs2004640) and poly A site SNP (rs10954213) (337:294 polyadenylation and result in different length and T:U, P ¼ 4.3 Â 10À3). The rs2070197 was shown to be a stability of the IRF5 mRNA. Specifically, we hypo- great predictor of exon 6 indel variation in Caucasians, thesized that A allele of rs10954213 might allow efficient because it tags particular haplotype composed of three polyadenylation at B12 bp downstream of AAUAAA SLE risk variations (rs2004640, exon 6 indel, rs1095213)91 motif, while the G allele favors the use of a distal poly (Figure 1). A þ site 648 bp downstream. The function of IRF5 exon 6 indel is putative and yet To test this hypothesis, we carried out northern under investigation. In addition to exon 6 indel, blotting and TaqMan quantitative PCR assay using alternative splice acceptor site variations for exon 6 IRF5 mRNA from EB-transformed cell lines and PBMC termed SS1 and SS291 that cause 48 bp indel at the of known genotype at rs10954213. As expected, cells beginning of exon 6 influence transactivation of down- homozygous for the A allele at rs10954213, carrying the stream genes.70,76,79 This variation might be worth taking wild-type AAUAAA on both alleles, expressed mainly into consideration in the future study. The conditional a short version of IRF5 mRNAs (termed ‘short’). In logistic regression analysis that includes these three contrast, cells homozygous for the G allele (AAUGAA) variations revealed that there is no additional SNP to expressed almost exclusively a longer mRNA (termed explain the full signal of IRF5 association to SLE. ‘long’) that utilized the second downstream poly A þ site91 (Figure 1). Haplotype analysis of three SLE risk variations in IRF5 This observation was also confirmed by Cunninghame We found three putative functional common variations Graham et al.96 using Affymetrix expression array. In fact, of IRF5, which affect isoform structure and/or expres- IRF5 mRNAs corresponding ‘short’ and ‘long’ were sion: (1) an SNP at the exon 1B splicing donor site already deposited in Genbank database. (rs2004640), producing transcripts containing exon 1B, To experimentally address the functional difference, (2) a 30 bp inframe indel of exon 6 that alters a PEST we cloned the two versions of IRF5 30-UTR, short and domain region, (3) an SNP in a conserved poly A signal long, into the downstream of the coding region of (rs10954213) associated with altered length of the IRF5 30- rabbit b-globin, and transfected 293 ‘Tet-off’ kidney cells UTR, and altered transcript levels91 (Figure 1). Although with expression plasmids. After suppression of new each variation was independently associated with risk to transcription by Doxycycline, we quantified the amount SLE to some extent, haplotype analysis in a large family of mRNA at the different time point by northern blotting. collection with 555 trio pedigrees from United States, The long mRNAs had about three times shorter half-life United Kingdom and a case–control sample totaling 2188 than short mRNAs. Estimated half-lives of short and case and 3596 control chromosomes from United States, long mRNA were 342788 and 125721 min, respec- United Kingdom and Sweden showed that distinct tively.91 Furthermore, we confirmed that cells carrying combinations of these three variants define three rs10954213 A/A genotype at poly A þ site (mainly different levels of risk for SLE. Among the five common express short mRNA) showed about fivefold higher level haplotypes defined, only haplotype 1 that has the ability of IRF5 protein than cells carrying rs10954213 G/G to express exon 1B isoforms (due to rs2004640), carries genotype (almost exclusively express long mRNA) by the exon 6 insertion and is expressed at high levels due to western blotting.91 the poly A þ variant was strongly associated with risk of These observations indicated that rs10954213 that SLE, appearing on 19.0% of SLE chromosomes compared influences IRF5 polyadenylation is the second causal to 11.9% of control chromosomes (OR ¼ 1.78, 95%CI: variation for the genetic risk to SLE in Caucasians. The 1.57–2.02, P ¼ 1.4 Â 10À19). Haplotypes 2 and 3, which risk allele produces short and relatively stable mRNA, carry only 2 of the 3 risk-associated functional alleles, and leads to higher level of mRNA and protein were average risk to SLE with no statistical significance expression, which may result in excessive production (OR ¼ 1.09 and 0.95, respectively). Haplotypes 4 and 5, of type I IFN production. which carry only 1 of the 3 risk-associated functional

Genes and Immunity Genetics of type I IFN in human SLE C Kyogoku and N Tsuchiya 451 alleles, and notably they both lack exon 1B isoform, T cells, in turn, stimulate autoreactive B cells to produce showed B25% reduction in risk with a protective effect autoantibodies. This results in the breakdown of toler- for SLE (OR ¼ 0.76 for both haplotypes, P ¼ 5.0 Â 10À8 ance to nuclear antigens, autoantibody secretion and IC and P ¼ 2.8 Â 10À5, respectively)91 (Figure 1). Recently, formation characteristic of SLE, and also provides these data were replicated by independent study in amplification loop for further stimulation of pDCs to Spanish, Argentine and German populations.88 produce greater quantities of type I IFNs. Alternatively, since TLR7/8/9 and IRF5 are also expressed in B cells, co-ligation of the B-cell receptor and TLR by ICs can directly stimulate development of autoreactive B cell to Type I IFN system in human SLE produce autoantibodies, where abnormal function of Figure 2 illustrates a model on the aberrant induction IRF5 allows B-cell tolerance breakdown against auto- of type I IFN pathway in human SLE. Here we focus on antigens. Polymorphisms in HLA-DRB1100 or reduction- how IRF5 and IFN signature stories can fit in the of-function polymorphisms in the B cell-inhibitory currently proposed model of SLE pathogenesis.16,98,99 receptor, FCGR2B,59,61 are likely to be involved in an The specific chromatin components such as ssRNA additive or epistatic manner in these processes. and dsDNA, either from infected viruses or internalized Secondly, upregulated production of type I IFN affects IC from apoptotic cell, are carried to endosome with a inflammatory cytokine-producing cells such as macro- help of Fcg receptors and trigger TLR7/8 and TLR9 phage, NK cells and cytotoxic CD8 þ T cells, to mediate signaling in pDCs, where IRF5 is either highly expressed cytotoxic effector function. Activation of these cells by and/or functionally potent because of gene polymorph- type I IFN results in tissue damage at the site of the cell isms. Activated IRF5 translocates to nucleus and stimu- encounter with pathogens. As shown in the IFN lates transcription of various inflammatory cytokines signature of serum protein in SLE, high expression of including type I IFNs. The unabated production of type I chemokines may disturb the normal trafficking and IFN further induces IFN-responsive genes termed ‘IFN chemotaxis of these cytotoxic effector cells to the correct signature’ in various types of immune cells and leads location of the body and contribute to the broader range to two key clinical features characteristic to SLE: (1) of inflammation and systemic tissue damage featured in autoantibody production (IC formation) and (2) inflam- SLE. The tissue damage also yields large numbers of mation and damage to systemic organs. nucleosomes that can be captured by pDCs and mature Firstly, upregulated production of type I IFN induces DCs, further amplifying the autoreactive process. the generation of mature DCs that capture nuclear The first trigger may be small increase in the type I IFN antigens and presents them to CD4 þ T cells. Activated levels; however, once an autoreactive process starts, such

Figure 2 A model on the role of IRF5 polymorphisms in the type I IFN system of human SLE. Virus infection and/or FcgR-captured immune complexes (ICs) trigger TLR signaling, in which IRF5 serves as a transcription factor. In genetically predisposed individuals, IRF5 is expressed at higher levels and/or functionally more potent in plasmacytoid dendritic cells (pDCs) and B cells, and leads to the upregulated production of inflammatory cytokines and/or type I IFNs. Type I IFNs induce ‘IFN signature’ in various immune cells (gray arrows). IFNa/b- matured DCs present autoantigens and activate autoreactive helper T cells and autoreactive B cells to produce autoantibodies, which results in IC formation. In B cells, high levels of IRF5 may contribute directly to produce autoantibodies after co-ligation of BCR-captured IC and TLRs. INFa/b-activated cytotoxic effector cells such as cytotoxic T cells (CTL), macrophages and NK cells may not be able to migrate to adequate locations of the body because of overproduction of chemokines, one of the features of IFN signature, and may contribute to inflammations in systemic organs. IC formation and nucleosomes from damaged tissue, in turn, may work on pDCs and mature DCs, respectively, to accelerate autoreacitve process (red arrows).

Genes and Immunity Genetics of type I IFN in human SLE C Kyogoku and N Tsuchiya 452 a self-activating system may exaggerate the strength of similar to human SLE.103 As IRF4 is known to compete type I IFN signaling activity in genetically predisposed with IRF5, thus serving as negative regulator of TLR individuals and may lead to the development of SLE signaling, attenuated function of IRF4 may lead to (Figure 2). increased function of IRF5. A recently reported human study showed that promoter SNP in IRF3 was protec- tively associated with development of SLE in Japanese.104 This SNP was associated with decreased mRNA expres- Future perspective sion of IRF3, suggesting that decreased transactivation of Type I IFN pathway is triggered as a normal event of IRF3 may be related to risk for SLE. immune system to protect our body from pathogens. Lastly, it is important to investigate the molecular However in SLE, genetic alterations in this pathway mechanisms that link the polymorphisms and the appears to lead to the sustained overproduction of type I development of SLE. The functional alteration caused IFN, which activates autoantigen-presenting cells (DCs), by a common genetic variation is, in itself, generally autoreactive T cells, autoreactive B cells and cytotoxic subtle and should be considered in the context of other effecter cells. Because most of the autoimmune responses genetic variations and environmental stimuli. Also, there featured in SLE, such as peripheral tolerance breakdown, is a certain limit of using primary cells from patients for nuclear autoantibody production, IC formation and experiment. Recently, more sensitive detection systems systemic tissue damage, can be explained at least in part are developed and resources of human cell lines with by impaired type I IFN system, it seems likely that type I annotated information by genome projects are available. IFN plays a major role in the pathogenesis. Taking full advantage of these resources, we need to take IRF5 is the first and so far the only gene in the type I a further step into the great challenge to delineate gene– IFN pathway whose genetic variations were confirmed gene interactions and gene–environment interactions to be associated with human SLE by large-scale associa- in the development of lupus and other autoimmune tion studies. However, most of the published studies diseases. analyzed Caucasian populations. In view of substantial difference in the allele frequencies and haplotype structure among populations, it is important to examine Asian and African populations for the association of IRF5 Acknowledgements with SLE. A recent report from Korea replicated the association of rs2004640 with SLE.101 We used an ethnic We are greatly indebted to Dr Robert R Graham collection from the Human Genome Diversity Panel (Harvard University, Massachusetts Institute of Techno- (HGDP)102 to gain information on the frequency of IRF5 logy) and Dr Timothy W Behrens (Genentech Inc.) for alleles in world populations. Importantly, the high-risk collaboration, supervision and inspiration. We also thank haplotype (haplotype 1) was rare/absent, whereas the Dr Emily C Baechler-Gillespie (University of Minnesota), protective haplotypes (haplotypes 4 and 5) were more Aya Kawasaki and Koki Hikami (University of Tsukuba) frequent, in the East Asian and African populations.91 for collaboration and helpful comments. Indeed, recent association studies including our own This work was supported by Grant-in-Aid for Scien- revealed substantial difference in the haplotype structure tific Research on Priority Areas ‘Applied Genomics’ from between Asian and Caucasian populations (submitted the Ministry of Education, Culture, Sports, Science and for publication). It would be of particular interest to see Technology of Japan, and the Grants from the Ministry of whether the same genetic effects of exon 6 indel and poly Health, Labour and Welfare of Japan, Japan Society for A site SNP are observed, and whether other polymor- the Promotion of Science (JSPS), Takeda Science Founda- phisms are involved, in populations other than the tion, and Japan Rheumatism Foundation. Caucasians. As represented by our study on IRF5 as well as other numerous studies, disease-associated gene variations are often related to gene expression and isoform structures. References Relevant variation in a gene is not necessarily one, but could be several, and the haplotype composed of 1 Hooks JJ, Moutsopoulos HM, Geis SA, Stahl NI, Decker JL, multiple variations may confer full genetic risk to Notkins AL. Immune interferon in the circulation of patients N Engl J Med 301 disease. Different levels of risk for disease in individuals with autoimmune disease. 1979; : 5–8. 2 Ytterberg SR, Schnitzer TJ. Serum interferon levels in patients can be defined by the haplotypes within a gene, as well with systemic lupus erythematosus. Arthritis Rheum 1982; 25: as genetic epitasis between multiple susceptible genes. 401–406. Taking these into consideration will help us search for 3 Preble OT, Black RJ, Friedman RM, Klippel JH, Vilcek J. new gene variations associated with disease risk and Systemic lupus erythematosus: presence in human serum of adequately understand the data from genetic association an unusual acid-labile leukocyte interferon. Science 1982; 216: study. 429–431. It is also important to search for genetic variations in 4 Bengtsson AA, Sturfelt G, Truedsson L, Blomberg J, Alm G, other genes that control type I IFN production or are Vallin H et al. Activation of type I interferon system in involved in downstream signaling pathway. There is systemic lupus erythematosus correlates with disease activity but not with antiretroviral antibodies. Lupus 2000; 9: evidence from a mouse study that loss of IRF4-binding 664–671. protein leads to the spontaneous development of a 5 Kalkner KM, Ronnblom L, Karlsson Parra AK, Bengtsson M, systemic autoimmune disorder characterized by the Olsson Y, Oberg K. Antibodies against double-stranded DNA accumulation of effector/memory T cells and IgG þ B and development of polymyositis during treatment with cells, autoantibody production and glomerulonephritis interferon. QJM 1998; 91: 393–399.

Genes and Immunity Genetics of type I IFN in human SLE C Kyogoku and N Tsuchiya 453 6 Ronnblom LE, Alm GV, Oberg K. Autoimmune phenomena IFN-alpha-producing cells by a lupus-related oligodeoxynu- in patients with malignant carcinoid tumors during inter- cleotide. Scand J Immunol 2001; 54: 543–550. feron-alpha treatment. Acta Oncol 1991; 30: 537–540. 26 Ronnblom L, Alm GV. A pivotal role for the natural 7 Gota C, Calabrese L. Induction of clinical autoimmune interferon alpha-producing cells (plasmacytoid dendritic disease by therapeutic interferon-alpha. Autoimmunity 2003; cells) in the pathogenesis of lupus. J Exp Med 2001; 194: 36: 511–518. F59–F63. 8 Kyogoku C, Langefeld CD, Ortmann WA, Lee A, Selby S, 27 Bave U, Magnusson M, Eloranta ML, Perers A, Alm GV, Carlton VE et al. Genetic association of the R620W poly- Ronnblom L. Fc gamma RIIa is expressed on natural morphism of protein tyrosine phosphatase PTPN22 with IFN-alpha-producing cells (plasmacytoid dendritic cells) human SLE. Am J Hum Genet 2004; 75: 504–507. and is required for the IFN-alpha production induced by 9 Alarcon-Riquelme ME. A RUNX trio with a taste for apoptotic cells combined with lupus IgG. J Immunol 2003; autoimmunity. Nat Genet 2003; 35: 299–300. 171: 3296–3302. 10 Baechler EC, Batliwalla FM, Karypis G, Gaffney PM, 28 Jego G, Palucka AK, Blanck JP, Chalouni C, Pascual V, Ortmann WA, Espe KJ et al. Interferon-inducible gene Banchereau J. Plasmacytoid dendritic cells induce plasma expression signature in peripheral blood cells of patients cell differentiation through type I interferon and interleukin with severe lupus. Proc Natl Acad Sci USA 2003; 100: 6. Immunity 2003; 19: 225–234. 2610–2615. 29 Medzhitov R, Preston-Hurlburt P, Janeway Jr CA. A human 11 Bennett L, Palucka AK, Arce E, Cantrell V, Borvak J, homologue of the Drosophila Toll protein signals activation of Banchereau J et al. Interferon and granulopoiesis signatures adaptive immunity. Nature 1997; 388: 394–397. in systemic lupus erythematosus blood. J Exp Med 2003; 197: 30 Kawai T, Akira S. TLR signaling. Cell Death Differ 2006; 13: 711–723. 816–825. 12 Kirou KA, Lee C, George S, Louca K, Papagiannis IG, 31 Kawai T, Akira S. Innate immune recognition of viral Peterson MG et al. Coordinate overexpression of interferon- infection. Nat Immunol 2006; 7: 131–137. alpha-induced genes in systemic lupus erythematosus. 32 Iwasaki A, Medzhitov R. Toll-like receptor control of the Arthritis Rheum 2004; 50: 3958–3967. adaptive immune responses. Nat Immunol 2004; 5: 987–995. 13 Bauer JW, Baechler EC, Petri M, Batliwalla FM, Crawford D, 33 Liu YJ. IPC: professional type 1 interferon-producing cells Ortmann WA et al. Elevated serum levels of interferon- and plasmacytoid dendritic cell precursors. Annu Rev regulated chemokines are biomarkers for active human Immunol 2005; 23: 275–306. systemic lupus erythematosus. PLoS Med 2006; 3: e491. 34 Colonna M, Trinchieri G, Liu YJ. Plasmacytoid dendritic cells 14 Isaacs A, Lindenmann J. Virus interference. I. The interferon. in immunity. Nat Immunol 2004; 5: 1219–1226. Proc R Soc Lond B Biol Sci 1957; 147: 258–267. 35 Heil F, Hemmi H, Hochrein H, Ampenberger F, Kirschning 15 Theofilopoulos AN, Baccala R, Beutler B, Kono DH. Type I C, Akira S et al. Species-specific recognition of single- interferons (alpha/beta) in immunity and autoimmunity. stranded RNA via toll-like receptor 7 and 8. Science 2004; 303: Annu Rev Immunol 2005; 23: 307–336. 1526–1529. 16 Stewart TA. Neutralizing interferon alpha as a therapeutic 36 Hemmi H, Takeuchi O, Kawai T, Kaisho T, Sato S, Sanjo H approach to autoimmune diseases. Cytokine Growth Factor et al. A Toll-like receptor recognizes bacterial DNA. Nature Rev 2003; 14: 139–154. 2000; 408: 740–745. 17 Aaronson DS, Horvath CM. A road map for those who don’t 37 Leadbetter EA, Rifkin IR, Hohlbaum AM, Beaudette BC, know JAK-STAT. Science 2002; 296: 1653–1655. Shlomchik MJ, Marshak-Rothstein A. Chromatin-IgG com- 18 Perniok A, Wedekind F, Herrmann M, Specker C, Schneider plexes activate B cells by dual engagement of IgM and M. High levels of circulating early apoptic peripheral blood Toll-like receptors. Nature 2002; 416: 603–607. mononuclear cells in systemic lupus erythematosus. Lupus 38 Hornung V, Rothenfusser S, Britsch S, Krug A, Jahrsdorfer B, 1998; 7: 113–118. Giese T et al. Quantitative expression of toll-like receptor 1-10 19 Herrmann M, Voll RE, Zoller OM, Hagenhofer M, Ponner BB, mRNA in cellular subsets of human peripheral blood Kalden JR. Impaired phagocytosis of apoptotic cell material mononuclear cells and sensitivity to CpG oligodeoxynucleo- by monocyte-derived macrophages from patients with tides. J Immunol 2002; 168: 4531–4537. systemic lupus erythematosus. Arthritis Rheum 1998; 41: 39 Reeves JP, Taurog JD, Steinberg AD. Polyclonal B-cell 1241–1250. activation of autoantibodies (CBA/N x NZB)F1 mice by 20 Casciola-Rosen LA, Anhalt G, Rosen A. Autoantigens polyinosinic polycytidylic acid. Clin Immunol Immunopathol targeted in systemic lupus erythematosus are clustered in 1981; 19: 170–180. two populations of surface structures on apoptotic keratino- 40 Santiago-Raber ML, Baccala R, Haraldsson KM, Choubey D, cytes. J Exp Med 1994; 179: 1317–1330. Stewart TA, Kono DH et al. Type-I interferon receptor 21 Vallin H, Blomberg S, Alm GV, Cederblad B, Ronnblom L. deficiency reduces lupus-like disease in NZB mice. J Exp Patients with systemic lupus erythematosus (SLE) have a Med 2003; 197: 777–788. circulating inducer of interferon-alpha (IFN-alpha) produc- 41 Braun D, Geraldes P, Demengeot J. Type I Interferon controls tion acting on leucocytes resembling immature dendritic the onset and severity of autoimmune manifestations in lpr cells. Clin Exp Immunol 1999; 115: 196–202. mice. J Autoimmun 2003; 20: 15–25. 22 Vallin H, Perers A, Alm GV, Ronnblom L. Anti-double- 42 Kono DH, Baccala R, Theofilopoulos AN. Inhibition of lupus stranded DNA antibodies and immunostimulatory plasmid by genetic alteration of the interferon-alpha/beta receptor. DNA in combination mimic the endogenous IFN-alpha Autoimmunity 2003; 36: 503–510. inducer in systemic lupus erythematosus. J Immunol 1999; 43 Nicoletti F, Di Marco R, Zaccone P, Xiang M, Magro G, 163: 6306–6313. Grasso S et al. Dichotomic effects of IFN-gamma on the 23 Bave U, Alm GV, Ronnblom L. The combination of apoptotic development of systemic lupus erythematosus-like syn- U937 cells and lupus IgG is a potent IFN-alpha inducer. drome in MRL-lpr/lpr mice. Eur J Immunol 2000; 30: 438–447. J Immunol 2000; 165: 3519–3526. 44 Richards HB, Satoh M, Jennette JC, Croker BP, Yoshida H, 24 Bave U, Vallin H, Alm GV, Ronnblom L. Activation of natural Reeves WH. Interferon-gamma is required for lupus ne- interferon-alpha producing cells by apoptotic U937 cells phritis in mice treated with the hydrocarbon oil pristane. combined with lupus IgG and its regulation by cytokines. Kidney Int 2001; 60: 2173–2180. J Autoimmun 2001; 17: 71–80. 45 Lawson BR, Prud’homme GJ, Chang Y, Gardner HA, Kuan J, 25 Magnusson M, Magnusson S, Vallin H, Ronnblom L, Alm Kono DH et al. Treatment of murine lupus with cDNA GV. Importance of CpG dinucleotides in activation of natural encoding IFN-gammaR/Fc. J Clin Invest 2000; 106: 207–215.

Genes and Immunity Genetics of type I IFN in human SLE C Kyogoku and N Tsuchiya 454 46 Kirou KA, Lee C, George S, Louca K, Peterson MG, Crow kinase 2 and interferon regulatory factor 5 genes are MK. Activation of the interferon-alpha pathway identifies a associated with systemic lupus erythematosus. Am J Hum subgroup of systemic lupus erythematosus patients with Genet 2005; 76: 528–537. distinct serologic features and active disease. Arthritis Rheum 65 Decker T, Stockinger S, Karaghiosoff M, Muller M, Kovarik P. 2005; 52: 1491–1503. IFNs and STATs in innate immunity to microorganisms. 47 Rozzo SJ, Allard JD, Choubey D, Vyse TJ, Izui S, Peltz G et al. J Clin Invest 2002; 109: 1271–1277. Evidence for an interferon-inducible gene, Ifi202, in the 66 Honda K, Taniguchi T. IRFs: master regulators of signalling susceptibility to systemic lupus. Immunity 2001; 15: 435–443. by Toll-like receptors and cytosolic pattern-recognition 48 Min W, Ghosh S, Lengyel P. The interferon-inducible receptors. Nat Rev Immunol 2006; 6: 644–658. p202 protein as a modulator of transcription: inhibition of 67 Barnes BJ, Moore PA, Pitha PM. Virus-specific activation of a NF-kappa B, c-Fos, and c-Jun activities. Mol Cell Biol 1996; 16: novel interferon regulatory factor, IRF-5, results in the 359–368. induction of distinct interferon alpha genes. J Biol Chem 49 Han GM, Chen SL, Shen N, Ye S, Bao CD, Gu YY. Analysis of 2001; 276: 23382–23390. gene expression profiles in human systemic lupus erythe- 68 Taniguchi T, Ogasawara K, Takaoka A, Tanaka N. IRF family matosus using oligonucleotide microarray. Genes Immun of transcription factors as regulators of host defense. Annu 2003; 4: 177–186. Rev Immunol 2001; 19: 623–655. 50 Crow MK, Wohlgemuth J. Microarray analysis of gene 69 Izaguirre A, Barnes BJ, Amrute S, Yeow WS, Megjugorac N, expression in lupus. Arthritis Res Ther 2003; 5: 279–287. Dai J et al. Comparative analysis of IRF and IFN-alpha 51 Ye S, Pang H, Gu YY, Hua J, Chen XG, Bao CD et al. Protein expression in human plasmacytoid and monocyte-derived interaction for an interferon-inducible systemic lupus asso- dendritic cells. J Leukocyte Biol 2003; 74: 1125–1138. ciated gene, IFIT1. Rheumatology 2003; 42: 1155–1163. 70 Mancl ME, Hu G, Sangster-Guity N, Olshalsky SL, Hoops K, 52 Baechler EC, Batliwalla FM, Reed AM, Peterson EJ, Gaffney Fitzgerald-Bocarsly P et al. Two discrete promoters regulate PM, Moser KL et al. Gene expression profiling in human the alternatively spliced human interferon regulatory factor-5 autoimmunity. Immunol Rev 2006; 210: 120–137. isoforms. Multiple isoforms with distinct cell type-specific 53 Baechler EC, Gregersen PK, Behrens TW. The emerging role expression, localization, regulation, and function. J Biol Chem of interferon in human systemic lupus erythematosus. Curr 2005; 280: 21078–21090. Opin Immunol 2004; 16: 801–807. 71 Means TK, Latz E, Hayashi F, Murali MR, Golenbock DT, 54 Shao W, Zhou Z, Laroche I, Lu H, Zong Q, Patel DD et al. Luster AD. Human lupus autoantibody-DNA complexes Optimization of rolling-circle amplified protein microarrays activate DCs through cooperation of CD32 and TLR9. J Clin for multiplexed protein profiling. J Biomed Biotechnol 2003; Invest 2005; 115: 407–417. 2003: 299–307. 72 Barrat FJ, Meeker T, Gregorio J, Chan JH, Uematsu S, Akira S 55 Perlee L, Christiansen J, Dondero R, Grimwade B, Lejnine S, et al. Nucleic acids of mammalian origin can act as Mullenix M et al. Development and standardization of endogenous ligands for Toll-like receptors and may promote multiplexed antibody microarrays for use in quantitative systemic lupus erythematosus. JExpMed2005; 202: 1131–1139. proteomics. Proteome Sci 2004; 2:9. 73 Boule MW, Broughton C, Mackay F, Akira S, Marshak- 56 Bombardier C, Gladman DD, Urowitz MB, Caron D, Chang Rothstein A, Rifkin IR. Toll-like receptor 9-dependent and CH. Derivation of the SLEDAI. A disease activity index for -independent dendritic cell activation by chromatin-immu- lupus patients. The Committee on Prognosis Studies in SLE. noglobulin G complexes. J Exp Med 2004; 199: 1631–1640. Arthritis Rheum 1992; 35: 630–640. 74 Honda K, Yanai H, Takaoka A, Taniguchi T. Regulation of the 57 Liang MH, Socher SA, Larson MG, Schur PH. Reliability and type I IFN induction: a current view. Int Immunol 2005; 17: validity of six systems for the clinical assessment of disease 1367–1378. activity in systemic lupus erythematosus. Arthritis Rheum 75 Negishi H, Ohba Y, Yanai H, Takaoka A, Honma K, Yui K 1989; 32: 1107–1118. et al. Negative regulation of Toll-like-receptor signaling by 58 Prokunina L, Castillejo-Lopez C, Oberg F, Gunnarsson I, Berg IRF-4. Proc Natl Acad Sci USA 2005; 102: 15989–15994. L, Magnusson V et al. A regulatory polymorphism in PDCD1 76 Barnes BJ, Kellum MJ, Field AE, Pitha PM. Multiple is associated with susceptibility to systemic lupus erythema- regulatory domains of IRF-5 control activation, cellular tosus in humans. Nat Genet 2002; 32: 666–669. localization, and induction of chemokines that mediate 59 Kyogoku C, Dijstelbloem HM, Tsuchiya N, Hatta Y, Kato H, recruitment of T lymphocytes. Mol Cell Biol 2002; 22: Yamaguchi A et al. Fcg receptor gene polymorphisms in 5721–5740. Japanese patients with systemic lupus erythematosus: con- 77 Schoenemeyer A, Barnes BJ, Mancl ME, Latz E, Goutagny N, tribution of FCGR2B to genetic susceptibility. Arthritis Rheum Pitha PM et al. The interferon regulatory factor, IRF5, is a 2002; 46: 1242–1254. central mediator of toll-like receptor 7 signaling. J Biol Chem 60 Tsuchiya N, Kyogoku C. Role of Fcg receptor IIb polymorph- 2005; 280: 17005–17012. ism in the genetic background of systemic lupus erythema- 78 Barnes BJ, Field AE, Pitha-Rowe PM. Virus-induced tosus: insights from Asia. Autoimmunity 2005; 38: 347–352. heterodimer formation between IRF-5 and IRF-7 modulates 61 Kono H, Kyogoku C, Suzuki T, Tsuchiya N, Honda H, assembly of the IFNA enhanceosome in vivo and transcrip- Yamamoto K et al. FcgRIIB Ile232Thr transmembrane poly- tional activity of IFNA genes. J Biol Chem 2003; 278: morphism associated with human systemic lupus erythema- 16630–16641. tosus decreases affinity to lipid rafts and attenuates 79 Barnes BJ, Richards J, Mancl M, Hanash S, Beretta L, Pitha inhibitory effects on B cell receptor signaling. Hum Mol Genet PM. Global and distinct targets of IRF-5 and IRF-7 during 2005; 14: 2881–2892. innate response to viral infection. J Biol Chem 2004; 279: 62 Tsuchiya N, Honda Z, Tokunaga K. Role of B cell inhibitory 45194–45207. receptor polymorphisms in systemic lupus erythematosus: a 80 Bolland S, Ravetch JV. Spontaneous autoimmune disease in negative times a negative makes a positive. J Hum Genet 2006; FcgRIIB-deficient mice results from strain-specific epistasis. 51: 741–750. Immunity 2000; 13: 277–285. 63 Wu H, Boackle SA, Hanvivadhanakul P, Ulgiati D, Grossman 81 Bolland S, Yim YS, Tus K, Wakeland EK, Ravetch JV. Genetic JM, Lee Y et al. Association of a common complement modifiers of systemic lupus erythematosus in FcgRIIB(À/À) receptor 2 haplotype with increased risk of systemic lupus mice. J Exp Med 2002; 195: 1167–1174. erythematosus. Proc Natl Acad Sci USA 2007; 104: 3961–3966. 82 Murphy ED, Roths JB. A Y chromosome associated factor in 64 Sigurdsson S, Nordmark G, Goring HH, Lindroos K, strain BXSB producing accelerated autoimmunity and Wiman AC, Sturfelt G et al. Polymorphisms in the tyrosine lymphoproliferation. Arthritis Rheum 1979; 22: 1188–1194.

Genes and Immunity Genetics of type I IFN in human SLE C Kyogoku and N Tsuchiya 455 83 Pisitkun P, Deane JA, Difilippantonio MJ, Tarasenko T, 93 Conne B, Stutz A, Vassalli JD. The 3’ untranslated region of Satterthwaite AB, Bolland S. Autoreactive B cell responses messenger RNA: A molecular. Nat Med 2000; 6: 637–641. to RNA-related antigens due to TLR7 gene duplication. 94 Edmonds M. A history of poly A sequences: from formation Science 2006; 312: 1669–1672. to factors to function. Prog Nucleic Acid Res Mol Biol 2002; 71: 84 Subramanian S, Tus K, Li QZ, Wang A, Tian XH, Zhou J et al. 285–389. A Tlr7 translocation accelerates systemic autoimmunity in 95 Mendez R, Richter JD. Translational control by CPEB: a murine lupus. Proc Natl Acad Sci USA 2006; 103: 9970–9975. means to the end. Nat Rev Mol Cell Biol 2001; 2: 521–529. 85 Takaoka A, Yanai H, Kondo S, Duncan G, Negishi H, 96 Cunninghame Graham DS, Manku H, Wagner S, Reid J, Mizutani T et al. Integral role of IRF-5 in the gene induction Timms K, Gutin A et al. Association of IRF5 in UK SLE programme activated by Toll-like receptors. Nature 2005; 434: families identifies a variant involved in polyadenylation. 243–249. Hum Mol Genet 2007; 16: 579–591. 86 Yanai H, Chen HM, Inuzuka T, Kondo S, Mak TW, Takaoka A 97 Levi BZ, Hashmueli S, Gleit-Kielmanowicz M, Azriel A, et al. Role of IFN regulatory factor 5 transcription factor in Meraro D. ICSBP/IRF-8 transactivation: a tale of protein– antiviral immunity and tumor suppression. Proc Natl Acad protein interaction. J Interferon Cytokine Res 2002; 22: 153–160. Sci USA 2007; 104: 3402–3407. 98 Pascual V, Farkas L, Banchereau J. Systemic lupus erythe- 87 Graham RR, Kozyrev SV, Baechler EC, Reddy MV, Plenge matosus: all roads lead to type I interferons. Curr Opin RM, Bauer JW et al. A common haplotype of interferon Immunol 2006; 18: 676–682. regulatory factor 5 (IRF5) regulates splicing and expression 99 Banchereau J, Pascual V. Type I interferon in systemic lupus and is associated with increased risk of systemic lupus erythematosus and other autoimmune diseases. Immunity erythematosus. Nat Genet 2006; 38: 550–555. 2006; 25: 383–392. 88 Kozyrev SV, Lewen S, Reddy PM, Pons-Estel B, Witte T, 100 Graham RR, Ortmann WA, Langefeld CD, Jawaheer D, Selby Junker P et al. Structural insertion/deletion variation in IRF5 SA, Rodine PR et al. Visualizing human leukocyte antigen is associated with a risk haplotype and defines the precise class II risk haplotypes in human systemic lupus erythema- IRF5 isoforms expressed in systemic lupus erythematosus. tosus. Am J Hum Genet 2002; 71: 543–553. Arthritis Rheum 2007; 56: 1234–1241. 101 Shin HD, Sung YK, Choi CB, Lee SO, Lee HW, Bae SC. 89 Morley M, Molony CM, Weber TM, Devlin JL, Ewens KG, Replication of the genetic effects of IFN regulatory factor 5 Spielman RS et al. Genetic analysis of genome-wide variation (IRF5) on systemic lupus erythematosus in a Korean in human gene expression. Nature 2004; 430: 743–747. population. Arthritis Res Ther 2007; 9: R32. 90 Cheung VG, Spielman RS, Ewens KG, Weber TM, Morley M, 102 Cann HM, de Toma C, Cazes L, Legrand MF, Morel V, Burdick JT. Mapping determinants of human gene expres- Piouffre L et al. A human genome diversity cell line panel. sion by regional and genome-wide association. Nature 2005; Science 2002; 296: 261–262. 437: 1365–1369. 103 Fanzo JC, Yang W, Jang SY, Gupta S, Chen Q, Siddiq A et al. 91 Graham RR, Kyogoku C, Sigurdsson S, Vlasova IA, Davies Loss of IRF-4-binding protein leads to the spontaneous LR, Baechler EC et al. Three functional variants of IFN development of systemic autoimmunity. J Clin Invest 2006; regulatory factor 5 (IRF5) define risk and protective 116: 703–714. haplotypes for human lupus. Proc Natl Acad Sci USA 2007; 104 Akahoshi M, Nakashima H, Shirakawa T. Roles of genetic 104: 6758–6763. variations in signalling/immunoregulatory molecules in 92 The International HapMap Consortium. A haplotype map of susceptibility to systemic lupus erythematosus. Semin the human genome. Nature 2005; 437: 1299–1320. Immunol 2006; 18: 224–229.

Genes and Immunity