To explore whether common genetic variation influences sJIA susceptibility, we performed a genome-wide association study (GWAS) of sJIA in collaboration with investigators from the International Childhood Arthritis Genetics Consortium (INCHARGE). INCHARGE includes investigators from major pediatric rheumatology centers in North America, South America, and Europe. The INCHARGE sJIA study population includes 988 children with sJIA and 8010 control subjects for whom we have produced a dataset of over 6.7 million genetic markers or single nucleotide polymorphisms (SNPs). To analyze this dataset, we divided the study population into 9 groups by country of origin and examined genetic associations with sJIA separately in each stratum. The association results were then meta-analyzed to identify genetic variants whose influence on sJIA disease risk transcended ethnicity. During this reporting period, we published the first manuscript from this study, which definitively identified the major histocompatibility complex (MHC) locus as the greatest genetic risk factor for sJIA (Ombrello MH et al. 2015 Proceedings of the National Academy of Sciences USA). The strongest sJIA-associated variant, rs151043342, was located nearest to the human leukocyte antigen (HLA)-DRA gene (p = 2.8 times 10 to the negative 17, odds ratio = 2.6). This variant was in strong linkage disequilibrium with the classical HLA allele, HLA-DRB1*11 (p = 2.7 times 10 to the negative 16, odds ratio = 2.3) and the haplotype of HLA-DRB1*11/HLA-DQA1*05/HLA-DQB1*03 (p = 6.4 times 10 to the negative 17, odds ratio = 2.3), both of which were also strongly associated with sJIA. Combinatorial analysis of MHC locus SNPs and classical HLA alleles revealed that the disease-associated SNPs and HLA types were co-inherited as a part of a 243 kilobase sJIA-associated haplotype that included the HLA-DRA, HLA-DRB family, HLA-DQA1, and HLA-DQB1 genes. Finally, conditional analysis adjusting for this primary association signal found that rs12722051, which encodes a tyrosine to phenylalanine substitution at position 48 of HLA-DQA1, influenced sJIA risk independent of rs151043342 (conditional p = 1 times 10 to the negative 5, odds ratio 0.7). Based upon these findings, we are undertaking a series of computational and experimental investigations to determine the mechanism(s) through which variants of the class II HLA region participate in the pathophysiology of sJIA. We have also recently completed our primary analyses of the full sJIA GWAS dataset, which has identified (in addition to the HLA locus) 21 novel sJIA candidate susceptibility loci (p less than 1 times 10 to the negative 5), the strongest of which was a susceptibility locus in a non-coding region of chromosome 1. In addition, we identified several important susceptibility loci that may immediately nominate novel therapeutic strategies. Given that this study provided the first characterization of sJIA's genetic architecture, we used the sJIA GWAS association data to perform formal statistical comparisons of the genetic underpinnings of sJIA and other JIA subtypes. By comparing association data of sJIA with those of oligoarticular, seronegative polyarticular, and rheumatoid factor positive polyarticular JIA, we demonstrated that the genetic architecture of sJIA is unique and distinct from these other subtypes. Based on the unique inflammatory phenotype of sJIA, combined with a complete divergence of the sJIA genetic architecture from those of the other forms of JIA, we suggested that the classification of this condition as a JIA subtype be reconsidered. A manuscript describing these results is currently under review. Additionally, we have continued to investigate the role of rare genetic variants in sJIA pathogenesis using targeted deep resequencing. Among the resequencing targets examined are genes with previously reported associations with sJIA, genes for which at least a modest association with sJIA (p less than 1 times 10 to the negative 5) was identified in a preliminary analysis of our sJIA GWAS, and genes known to cause monogenic autoinflammatory syndromes or familial hemophagocytic lymphohistiocytosis. Although in the preliminary stages, we have identified numerous interesting novel and ultra-rare variants within the INCHARGE sJIA collection and we are currently undertaking mechanistic investigations to elucidate the roles of these variants in sJIA pathophysiology. Finally, during this reporting period we continued to investigate Behcet's disease (BD), attempting to elucidate the mechanism(s) through which its strongest risk factor, HLA-B*51, influences disease risk. First, based on our recent work, we hypothesized that HLA-B*51 may contribute to BD risk through its interaction with its killer immunoglobulin-like receptor (KIR) ligands, KIR3DL1 and KIR3DS1. To examine this possibility, we determined KIR3DL1/KIR3DS1 genotypes in 1709 BD cases and 1799 healthy controls. Although we found that the effect of HLA-B*51 on BD risk did not involve KIR3DL1/KIR3DS1, we unexpectedly identified an HLA-B*51-independent role for KIR3DS1 alleles in the ocular manifestations of BD. A manuscript describing this work is forthcoming in Genes & Immunity. Second, we have recently shown that variants in endoplasmic reticulum-related aminopeptidase 1 (ERAP1), which trims peptides for loading onto HLA class I molecules, influence BD risk through an interaction with HLA-B*51. Additionally, we have recently reported that ERAP1 is encoded by a series of haplotypes/allotypes. In the present study, we used haplotypic association analysis to demonstrate that a single allotype of ERAP1 was associated with BD. The protein encoded by the disease-associated allotype has low enzymatic activity, suggesting that a hypoactive ERAP1 protein contributes to BD risk by altering the peptide repertoire available for binding to HLA-B*51. This work was recently published in Annals of the Rheumatic Diseases.
Showing the most recent 10 out of 16 publications
