Background Juvenile idiopathic arthritis (JIA) is the most common arthritic disease of childhood and is caused by a combination of genes and environment. In the last few years great advances have been made in dissecting the genetic basis of JIA with now 17 confirmed susceptibility loci identified. One of these loci, the MHC region, has been established for many years but the complexity and the broad linkage disequilibrium across the region has rendered fine-mapping associations challenging. Novel imputation strategies can now be utilized to impute HLA classical alleles and amino acids across the region.

Objectives The aim of this work was to gain a greater understanding of the associations across the HLA region in all JIA subtypes.

Methods Using the dense genotype data obtained from the analysis of the Illumina ImmunoChip in 4574 JIA cases (not including the systemic-onset cases) and 14412 controls, we imputed HLA classical alleles (2-digit and 4-digit resolution) and amino acids across the MHC region (Chr6:29-34Mb) using the SNP2HLA algorithm and a large reference panel. We performed univariate analysis for all markers across the region and tested all amino acids in HLA-DRB1 by performing an omnibus test of amino acid residues for each position. Conditional analysis, including the most significant marker as a covariate, was performed to identify independent effects. Analysis was repeated in the individual subtypes (persistent oligoarthritis (pO) n=1751, extended oligoarthritis (eO) n=658, RF negative polyarthritis (RF-P) n=1525, RF positive polyarthritis (RF+P) n=337, enthesitis related arthritis (ERA) n=183, psoriatic arthritis (jPsA) n=112).

Results The omnibus test for all amino acids across HLA-DRB1 showed most significant association at HLA-DRB1 amino acid (AA) 67 and conditioning on all residues at AA 67 found significant association remaining at HLA-DRB1 AA 13 (the glycine residue most associated), suggesting two independent effects in HLA-DRB1. There was evidence for further effects in HLA-DRB1. Conditioning on all alleles of HLA-DRB1 found additional independent effects at HLA-DPB1-0201 and HLA-A AA 95. These associations hold across RF-P and both pO and eO subtypes, HLA-DRB1 AA 13 is most strongly associated in RF+P with the histidine residue driving the association. In the ERA subtype HLA-B27 showed the strongest association. For jPsA no HLA markers reached genome-wide significance.

Conclusions Analysis of the MHC region in the largest cohort of JIA cases and controls studied to date has found the strongest association with the HLA-DRB1 region, with additional multiple independent effects.

Acknowledgements Juvenile Arthritis Consortium for Immunochip (JACI)

Disclosure of Interest None declared