Rare variants and autoimmune disease Jonathan Massey and Steve Eyre

Transcription

1 BRIEFINGS IN FUNCTIONAL GENOMICS. VOL 13. NO ^397 doi: /bfgp/elu011 Rare variants and autoimmune disease Jonathan Massey and Steve Eyre Advance Access publication date 9 May 2014 Abstract The study of rare variants in monogenic forms of autoimmune disease has offered insight into the aetiology of more complex pathologies. Research in complex autoimmune disease initially focused on sequencing candidate genes, with some early successes, notably in uncovering low-frequency variation associated with Type 1 diabetes mellitus. However, other early examples have proved difficult to replicate, and a recent study across six autoimmune diseases, re-sequencing 25 autoimmune disease-associated genes in large sample sizes, failed to find any associated rare variants. The study of rare and low-frequency variation in autoimmune diseases has been made accessible by the inclusion of such variants on custom genotyping arrays (e.g. Immunochip and Exome arrays). Whole-exome sequencing approaches are now also being utilised to uncover the contribution of rare coding variants to disease susceptibility, severity and treatment response. Other sequencing strategies are starting to uncover the role of regulatory rare variation. Keywords: autoimmunity; immunochip; complex disease; rare variants; low-frequency variants INTRODUCTION In recent years, there have been great advances in our understanding of the genetics underlying various autoimmune diseases. However, the vast majority of these discoveries come from genome-wide association studies (GWAS), designed and powered to detect common variation, defined as having a minor allele frequency (MAF) of >5% in the population. Many of the variants discovered have small effect sizes (typically odds ratios <1.5) and together explain only a small proportion of disease heritability. There have been many hypotheses suggested for the missing heritability phenomena in complex diseases, for example, epigenetics, epistasis, gene environment interactions and rare variation (defined here as MAF < 0.5%). Between the two extremes of rare and common variants, studies of low-frequency variants, typically with an MAF range of 0.5 5%, have seen the most success in determining the contribution of less common genetic variants to autoimmune disease. MONOGENIC STUDIES AND RARE VARIANTS IN AUTOIMMUNITY The association of rare variants in autoimmune diseases has, until recently, been confined to monogenic/mendelian forms of disease, where these variants have large effect sizes and a high penetrance. A good example of this is autoimmune polyendocrinopathy syndrome type I and rare variant defects in the autoimmune regulator gene [1]. Important insights have been gained in understanding the aetiology of certain autoimmune diseases, particularly systemic lupus erythematosus (SLE), from studies of rare variants in these forms of disease. By discovering the cause of Mendelian forms of disorders that are symptomatically similar to SLE, advances have been made into which pathways play an important role in the more complex disorder. For example, three prime repair exonuclease 1 (TREX1), thought to play a role in DNA repair, is associated with monogenic Aicardi Goutières syndrome [2], a disease that exhibits Corresponding author. J. Massey, Arthritis Research UK Centre for Genetics and Genomics, Centre for Musculoskeletal Research, Manchester Academic Health Science Centre, The University of Manchester, Manchester, Oxford Road, Manchester M13 9PT, UK. Tel.: þ ; Fax: þ ; jonathan.massey@manchester.ac.uk Jonathan Massey is a postdoctoral research associate in the Centre for Musculoskeletal Research at the University of Manchester, where his research focused on the contribution of rare genetic variants to inflammatory arthritis susceptibility. Steve Eyre is a senior research fellow in the Centre for Musculoskeletal Research at the University of Manchester. His research interests are on the genetics and genomics of a range of rheumatological disorders, with a specific focus on inflammatory joint disease. ß The Author Published by Oxford University Press. All rights reserved. For permissions, please journals.permissions@oup.com

2 Rare variants and autoimmune disease 393 phenotypic overlap with SLE. More recently, it was discovered that rare coding TREX1 mutations are present in 0.5% of SLE patients [3]. This not only provides mechanistic insight into both diseases but also opens the possibility of rare variants defining sub-phenotypes of complex diseases. Such discoveries can also provide information into the fundamental workings of the immune system, which has implications for a number of diseases, e.g. juvenileonset SLE patients have mutations in protein kinase C delta, a gene crucial for regulating B-cell tolerance [4]. EARLY CANDIDATE GENE-BASED STUDIES Rare variant studies were initially conducted in genes associated with disease through prior evidence, such as GWAS. Low-frequency variants were discovered to be associated with susceptibility to rheumatoid arthritis (RA) at the TNFAIP3 locus using a candidate gene approach [5]. The study of rare variants in complex autoimmunity is a nascent area and researchers have reason to be cautious. Some other early success stories have proven difficult to validate, such as the association in multiple sclerosis (MS) with rare variants in CYP27B1 [6], which has not been replicated in an independent study [7]. Similarly, a study of rare variants in sialic acid acetylesterase with multiple autoimmune diseases [8] has been strongly refuted [9]. Hunt et al. [9] outlined some of the difficulties in studying rare variants, drawing attention to methodological issues, such as the use of principal component analysis, widely used for the common variant analysis, which must be used with caution when studying rare variants; an unidentified substructure using these methods will likely confound rare variant studies. AUTOIMMUNE DISEASES AND DENSE-MAPPING ARRAYS More recently, there have been efforts to make the study of low-frequency and rare variation more accessible by the inclusion of such variants in the design of modern genotyping arrays. The Illumina Immunochip is a custom genotyping array designed to fine-map all the GWAS-associated regions (186 in total) in 11 immune-mediated diseases. The content of this custom array included single nucleotide polymorphisms (SNPs) from the 1000 Genomes project and various re-sequencing data. Around lowfrequency variants (18% of the total SNPs) and rare variants (16% of the total SNPs) [10] were available for genotyping. So far, large-scale studies utilising these arrays have identified many lowfrequency variants associated with autoimmune diseases, for example, with TYK2 in RA [11] and juvenile idiopathic arthritis [12], but have so far had less success with the identification of rarer variant associations. AUTOIMMUNITY, RARE VARIANTS AND EXON-FOCUSED ARRAYS In addition, arrays have now been released by Affymetrix and Illumina to survey variations in the exome, containing many more rare and lowfrequency variants. The Illumina HumanExome Beadchip contains variants, mainly protein-altering, selected from > exome and genome sequences [13]. In a study of insulin processing and secretion traits, one of the first applications of this array [13], of the variants polymorphic in the Finnish population studied, the majority (47 722; 53.1%) were found to be rare (MAF < 0.5%). This study failed to identify any associations with rare variants, although it did highlight low-frequency variants in both previously identified (SGSM2 and MADD) and novel genes (TBC1D30, KANK1 and PAM). In inflammatory disorders, there is currently one published example of a study using exome array technology in Takayasu s arteritis [14]. Whilst this study did uncover associations with both IL12B and HLA-B, it failed to reveal any insights with rare variants. The genotype clustering algorithms used for common variants on arrays are well established, but it remains the main technical consideration when studying lower-frequency variants in this way. The introduction of newer algorithms, such as opticall and zcall, has improved the situation, but researchers should be aware of the correct application and their limits. AUTOIMMUNE DISEASE AND TARGETED RE-SEQUENCING PROJECTS The study of rare variants in Type 1 diabetes mellitus (T1DM) [15] can be regarded as the seminal success

3 394 Massey and Eyre story of lower-frequency variants in autoimmunity. Nejentsev et al. sequenced the exons and regulatory sequences of 10 T1DM candidate genes in pools of DNA from 480 cases and 480 controls of UK origin. Association testing of discovered variants, within these pools, identified low-frequency variants in IFIH1 and potentially in CLEC16A. Two SNPs from each of these genes were genotyped in another UK cohort of 8379 cases and controls. Both IFIH1 variants showed strong association with T1DM, whereas the CLEC16A SNPs failed to replicate; 3165 families, with either one or two affected offspring, were used as an additional validation to test the association of all IFIH1 variants discovered from sequencing. This revealed four functionally relevant low-frequency variants (MAF in controls %) that were independently associated with lowered disease risk. In a similar effort to Nejentsev et al., Diogo et al. [16] performed exon sequencing of 25 candidate genes, located within RA risk loci, in pools of DNA from 500 cases and 650 controls (European origin). Using gene-based burden tests, they identified an accumulation of rare non-synonymous variants in IL2RA and IL2RB that were exclusive to cases. Since the T1DM publication [15], the extent to which rare variants contribute to autoimmune disease susceptibility has been placed into question [17]. Here, in one of the largest rare variant studies to date, the exons of 25 genes, identified from autoimmune GWAS studies, were selected and subsequently sequenced in UK residents of Northern European ancestry. This included cases from six autoimmune disease cohorts (coeliac disease, T1DM, Crohn s disease, MS, psoriasis and autoimmune thyroid disease). Of the 2990 variants discovered in protein-coding regions, the majority (68.9%) only occurred in one or two individuals. Using gene-based testing, they were unable to show any significant rare-variant enrichment in these 25 genes for affected individuals. However, in coeliac disease, using a single-variant analysis method, they were able to demonstrate an association with two low-frequency (albeit with MAFs of 4.97%) variants in the NCF2 gene. This represented a novel association for this disease, but it was only omitted from discovery in the earlier Immunochip due to missing data [17]. The authors of the paper conclude that rare coding variants are unlikely to contribute greatly to the missing heritability of autoimmune diseases and warn against whole-exome sequencing projects in complex autoimmune diseases. Despite this negative assessment of the role of rare variants in autoimmunity, there have been success stories (Table 1). Furthermore, although groundbreaking, the Hunt et al. [17] study is not exhaustive. The authors concentrated on 25 loci that showed evidence of pleiotropic association (associated in at least two diseases) and had, in most cases, a single strong candidate immune-related gene. Additionally, the loci were required to be available on the Illumina Immunochip. The six autoimmune diseases chosen are quite heterogeneous in pathology. This heterogeneity could influence the pleiotropy and hence the selection of genes. The selection of only immunerelated genes does not allow for the distinct mechanisms of each disease, which might have different rare variant contributions. Equally, the selection of genes based on inclusion on the Immunochip array does not allow for the study of immune-related genes in other areas of the genome. Therefore, there might yet be utility in sequencing for rare variants by whole exome/genome approaches in individual pathologies. EXOME RE-SEQUENCING A more comprehensive methodology for studying likely disease-causing rare variants, compared with exome array analysis, is to re-sequence all exonic DNA in disease cohorts. Tang et al. [24], studying psoriasis in a Chinese population, used a combination of the whole-exome sequencing approach and targeted sequencing of immune-related genes (622 genes identified from the GWAS catalogue, i.e. not restricted to Immunochip loci). The whole-exome study did not identify any significantly associated genes, but the authors conclude this might have been due to lack of power in this discovery phase (781 cases and 676 controls). They did, however, identify significantly associated low-frequency, independent variants in IL23R and GJB2. Additionally, they found suggestive evidence for rare variants in FUT2 and TARBP1. However, they failed to replicate the findings of a previous study of psoriasis in a mixed-ancestry cohort, which identified associated rare variants in CARD14 [22]. The whole-exome approach has been used in Crohn s disease and ulcerative colitis to identify

4 Rare variants and autoimmune disease 395 Table 1: Rare and low-frequency variant associations with autoimmune diseases Disease Method Rare or low-frequency Gene/region Reference T1DM Candidate sequencing Low IFIH1 [15] Inflammatory Candidate sequencing Low IL23R [18] bowel disease Candidate sequencing Rare NOD2 [19, 20] Whole-exome sequencing Rare (ulcerative colitis) and PRDM1 [21] low (Crohn s disease) Psoriasis Candidate sequencing Rare CARD14 [22] GWAS/Immunochip Low TYK2 [23] Whole-exome and candidate Low IL23R, GJB2 [24] sequencing RA Candidate sequencing Rare IL2RA, IL2RB [16] Immunochip Low TYK2 [11] Juvenile idiopathic Immunochip Low TYK2 [12] arthritis SLE Genotyping Low BLK [25] Genotyping Low TREX1 [3] Behc et s disease Candidate sequencing Rare and low IL23R, TLR4, MEFV [26] Systemic sclerosis Immunochip Rare IL12A [27] Primary biliary Immunochip Low TYK2 [28] cirrhosis Rare IL12A, 16p13 Coeliac disease Immunochip Low CD28-CTLA4-ICOS, [10] RGS1, PTPN2 Rare 16p13 low-frequency PRDM1 variants associated with disease (42 cases and 5 controls whole-exome sequenced, followed-up by genotyping in > cases) [21]. This is in addition to the previously identified low-frequency and rare variants, from targeted candidate gene sequencing studies, in IL23R [18] and NOD2 [19, 20]. Ellinghaus et al. [21] conclude that exome rare variants in inflammatory bowel disease are unlikely to contribute significantly to disease susceptibility. They do leave open the possibility of the importance of rare variation in non-coding/ regulatory regions. RARE REGULATORY VARIATION The importance of rare regulatory variation is partially supported by three studies conducted using the Immunochip. In coeliac disease ( cases and controls) [10], there were two low-frequency variant associations upstream of RGS1 and CD28, a low-frequency association in an intron of PTPN2 and a rare intergenic variant at the 16p13 locus. In primary biliary cirrhosis (2861 cases) [28], there was an associated intergenic rare variant at the 16p13 locus, located in an open-chromatin peak, and a rare variant located upstream of the IL12A gene. Evidence for an association within this IL12A intergenic region has been further supported by a study in systemic sclerosis (5850 cases) [27]. Using genotype imputation, regulatory low-frequency variants have also been demonstrated in SLE, where a missense substitution in BLK decreased the protein half-life [25]. With the increase in the regulatory characterisation of intergenic regions, for example, with cellspecific eqtl, ChIP-SEQ and DNase hypersensitivity experiments, collated on databases such as ENCODE, it is hoped that it will become easier to decipher the effects of rare regulatory variation. RARE VARIANTS AND TREATMENT RESPONSE Following on from rare, Mendelian forms of disease, it could be hypothesised that rare variants may well drive rare forms of common disease either in subphenotypes, rare adverse events or disease outcomes. Response to drug therapy for autoimmunity is one area in which the study of rare variants is being actively pursued. Nelson et al. [29] sequenced 202 drug target genes in people and found an abundance of rare, functional variants. It has been hypothesised that because of the evolutionary history of humans, many pharmacogenetic traits, unlike

5 396 Massey and Eyre common traits such as autoimmune disease susceptibility, have escaped natural selection. This lack of constraint may have allowed pharmacogenetic-associated variants to rise in frequency [30], making them more amenable to study. Methotrexate is a drug used to treat many autoimmune diseases, such as RA, psoriatic arthritis, SLE and Crohn s disease. It is also used in cancer therapy. Ramsey etal. [31] studied the clearance of methotrexate in 699 children with acute lymphoblastic leukaemia in relation to variants in SLCO1B1, which had been identified in a previous GWAS [32]. They found that both common and rare variants contributed to the phenotype, but that rare variants had larger effect sizes. Rare variants have also been implicated with pharmacogenetic traits in diseases outside of the autoimmunity arena [33, 34]. CONCLUSIONS AND FUTURE DIRECTIONS Projects to uncover rare variation contributing to autoimmune diseases are ongoing. In this effort, studies have been conducted in isolated populations, such as Iceland and the decode project, in diseases, such as Alzheimer s disease [35], and provided a framework for such studies in autoimmunity. Following the recommendations by Hunt et al. [9], family-based studies are being increasingly used to study rare variant associations with autoimmune diseases, such as the discovery of novel low-frequency variants in TYK2 in MS [36]. Whole-genome sequencing approaches with large numbers of disease cohorts are now becoming a reality, offering the possibilities of hypothesis-free rare variant analysis with greater coverage than before, particularly in regulatory regions. The problems of distinguishing real variation from sequencing errors at low sequencing coverage will improve in line with sequencing and bioinformatics advances. Re-sequencing and array-based studies will be further enhanced with ongoing efforts to generate large numbers of control data, from projects such as UK10K [37]. This will improve both the power of studies and the imputation accuracy of low-frequency variation. In conclusion, it is currently unclear how much rare variants will contribute to our understanding of autoimmune disease heritability. The studies presented here give a glimpse of what can be achieved with careful design and sufficient power. It is likely that rare variants will play a limited, but important, role in the post-gwas era. Key points Successes in ascertaining rare variant associations important in autoimmune diseases have largely been localised to Mendelian forms of disease. Success in common, complex autoimmune diseases has been limited to low-frequency variant associations rather than rare variants, with a few notable exceptions. Recent efforts have questioned the extent of the impact of rare variants in autoimmune diseases. More exhaustive approaches, including larger more focused cohorts, incorporating re-sequencing and whole exome arrays, are currently being pursued to better understand the contribution of rare variants to disease subsets, prognosis and treatment response. FUNDING This work was supported by Arthritis Research UK grant number References 1. Nagamine K, Peterson P, Scott HS, et al. Positional cloning of the APECED gene. Nat Genet 1997;17: Crow YJ, Hayward BE, Parmar R, et al. Mutations in the gene encoding the 3-5 DNA exonuclease TREX1 cause Aicardi-Goutieres syndrome at the AGS1 locus. Nat Genet 2006;38: Namjou B, Kothari PH, Kelly JA, et al. Evaluation of the TREX1 gene in a large multi-ancestral lupus cohort. Genes Immun 2011;12: Belot A, Kasher PR, Trotter EW, etal. Protein kinase cdelta deficiency causes mendelian systemic lupus erythematosus with B cell-defective apoptosis and hyperproliferation. Arthritis Rheum 2013;65: Bowes J, Lawrence R, Eyre S, et al. Rare variation at the TNFAIP3 locus and susceptibility to rheumatoid arthritis. Hum Genet 2010;128: Ramagopalan SV, Dyment DA, Cader MZ, et al. Rare variants in the CYP27B1 gene are associated with multiple sclerosis. Ann Neurol 2011;70: Cortes A, Field J, Glazov EA, et al. Resequencing and finemapping of the chromosome 12q13-14 locus associated with multiple sclerosis refines the number of implicated genes. Hum Mol Genet 2013;22: Surolia I, Pirnie SP, Chellappa V, et al. Functionally defective germline variants of sialic acid acetylesterase in autoimmunity. Nature 2010;466: Hunt KA, Smyth DJ, Balschun T, etal. Rare and functional SIAE variants are not associated with autoimmune disease risk in up to 66,924 individuals of European ancestry. Nat Genet 2012;44: Trynka G, Hunt KA, Bockett NA, et al. Dense genotyping identifies and localizes multiple common and rare variant association signals in celiac disease. Nat Genet 2011; 43: Eyre S, Bowes J, Diogo D, et al. High-density genetic mapping identifies new susceptibility loci for rheumatoid arthritis. Nat Genet 2012;44: Hinks A, Cobb J, Marion MC, et al. Dense genotyping of immune-related disease regions identifies 14 new

6 Rare variants and autoimmune disease 397 susceptibility loci for juvenile idiopathic arthritis. Nat Genet 2013;45: Huyghe JR, Jackson AU, Fogarty MP, et al. Exome array analysis identifies new loci and low-frequency variants influencing insulin processing and secretion. Nat Genet 2013;45: Terao C, Yoshifuji H, Kimura A, et al. Two susceptibility loci to Takayasu arteritis reveal a synergistic role of the IL12B and HLA-B regions in a Japanese population. AmJ Hum Genet 2013;93: Nejentsev S, Walker N, Riches D, et al. Rare variants of IFIH1, a gene implicated in antiviral responses, protect against type 1 diabetes. Science 2009;324: Diogo D, Kurreeman F, Stahl EA, et al. Rare, low-frequency, and common variants in the protein-coding sequence of biological candidate genes from GWASs contribute to risk of rheumatoid arthritis. AmJ Hum Genet 2013; 92: Hunt KA, Mistry V, Bockett NA, etal. Negligible impact of rare autoimmune-locus coding-region variants on missing heritability. Nature 2013;498: Momozawa Y, Mni M, Nakamura K, etal. Resequencing of positional candidates identifies low frequency IL23R coding variants protecting against inflammatory bowel disease. Nat Genet 2011;43: Lesage S, Zouali H, Cezard JP, etal. CARD15/NOD2 mutational analysis and genotype-phenotype correlation in 612 patients with inflammatory bowel disease. AmJ Hum Genet 2002;70: Rivas MA, Beaudoin M, Gardet A, etal. Deep resequencing of GWAS loci identifies independent rare variants associated with inflammatory bowel disease. Nat Genet 2011;43: Ellinghaus D, Zhang H, Zeissig S, et al. Association between variants of PRDM1 and NDP52 and Crohn s disease, based on Exome sequencing and functional studies. Gastroenterology 2013;145: Jordan CT, Cao L, Roberson ED, et al. Rare and common variants in CARD14, encoding an epidermal regulator of NF-kappaB, in psoriasis. AmJ Hum Genet 2012;90: Tsoi LC, Spain SL, Knight J, et al. Identification of 15 new psoriasis susceptibility loci highlights the role of innate immunity. Nat Genet 2012;44: Tang H, Jin X, Li Y, et al. A large-scale screen for coding variants predisposing to psoriasis. Nat Genet 2014;46: Delgado-Vega AM, Dozmorov MG, Quiros MB, etal. Fine mapping and conditional analysis identify a new mutation in the autoimmunity susceptibility gene BLK that leads to reduced half-life of the BLK protein. Ann Rheum Dis 2012;71: Kirino Y, Zhou Q, Ishigatsubo Y, et al. Targeted resequencing implicates the familial Mediterranean fever gene MEFV and the toll-like receptor 4 gene TLR4 in Behcet disease. Proc Natl Acad Sci USA 2013;110: Mayes MD, Bossini-Castillo L, Gorlova O, et al. Immunochip analysis identifies multiple susceptibility Loci for systemic sclerosis. AmJ Hum Genet 2014;94: Liu JZ, Almarri MA, Gaffney DJ, et al. Dense fine-mapping study identifies new susceptibility loci for primary biliary cirrhosis. Nat Genet 2012;44: Nelson MR, Wegmann D, Ehm MG, et al. An abundance of rare functional variants in 202 drug target genes sequenced in 14,002 people. Science 2012;337: Cirulli ET, Goldstein DB. Uncovering the roles of rare variants in common disease through whole-genome sequencing. Nat Rev Genet 2010;11: Ramsey LB, Bruun GH, Yang W, et al. Rare versus common variants in pharmacogenetics: SLCO1B1 variation and methotrexate disposition. Genome Res 2012;22: Trevino LR, Shimasaki N, Yang W, et al. Germline genetic variation in an organic anion transporter polypeptide associated with methotrexate pharmacokinetics and clinical effects. J Clin Oncol 2009;27: Irvin MR, Zhang Q, Kabagambe EK, et al. Rare PPARA variants and extreme response to fenofibrate in the Genetics of Lipid-Lowering Drugs and Diet Network Study. Pharmacogenet Genomics 2012;22: Drake KA, Torgerson DG, Gignoux CR, et al. A genomewide association study of bronchodilator response in Latinos implicates rare variants. J Allergy Clin Immunol 2014;133: Jonsson T, Atwal JK, Steinberg S, et al. A mutation in APP protects against Alzheimer s disease and age-related cognitive decline. Nature 2012;488: Dyment DA, Cader MZ, Chao MJ, etal. Exome sequencing identifies a novel multiple sclerosis susceptibility variant in the TYK2 gene. Neurology 2012;79: Muddyman D, Smee C, Griffin H, et al. Implementing a successful data-management framework: the UK10K managed access model. Genome Med 2013;5:100.