by Ebun Omoyinmi, Paola Forabosco, Raja Hamaoui, Annette Bryant, Anne Hinks, Simona Ursu, Childhood Arthritis Prospective Study (CAPS) , BSPAR study group , Childhood Arthritis Response to Medication Study (CHARMS) , Lucy R. Wedderburn, Wendy Thomson, Cathryn M. Lewis, Patricia Woo
The cytokine IL-10 and its family members have been implicated in autoimmune diseases and we have previously reported that genetic variants in IL-10 were associated with a rare group of diseases called juvenile idiopathic arthritis (JIA). The aim of this study was to fine map genetic variants within the IL-10 cytokine family cluster on chromosome 1 using linkage disequilibrium (LD)-tagging single nucleotide polymorphisms (tSNPs) approach with imputation and conditional analysis to test for disease associations. Methodology/Principal Findings
Fifty-three tSNPs were tested for association between Caucasian paediatric cohorts [219 systemic JIA (sJIA), 187 persistent oligoarticular JIA (pOJIA), and 139 extended OJIA (eOJIA) patients], and controls (Wellcome Trust control cohort, WTCCC2). Significant association with sJIA was detected at rs1400986 in the promoter of IL-20 (odds ratio 1.53; 95% CI 1.21–1.93; p?=?0.0004), but in no other subtypes. Imputation analysis identified additional associated SNPs for pOJIA at IL-20 and IL-24, including a rare, functional, missense variant at IL-24 with a p?=?0.0002. Penalised logistic regression analysis with HyperLasso and conditional analysis identified several further associations with JIA subtypes. In particular, haplotype analysis refined the sJIA association, with a joint effect at rs1400986 and rs4129024 in intron 1 of MAPKAPK2 (p?=?3.2E-5). For pOJIA, a 3-SNP haplotype including rs1878672 in intron 3 of IL-10 showed evidence for association (p?=?0.0018). In eOJIA, rs10863962 (3'UTR of FCAMR) and rs12409577 (intron of IL-19) haplotype showed some evidence of association (p?=?0.0003). Conclusions
This study supports previous association of IL-20 with sJIA. Haplotype analyses provided stronger association signals than single point analyses, while a penalised logistic regression approach also suggested multiple independent association signals. Replication studies are required to confirm or refute these findings. The results indicate that combined effects with unknown/rare variants remain to be characterised in JIA, and represent a possible example of synthetic association in this region.