Document resource
Introduction The HLA genes are highly polymorphic, leading to numerous different HLA alleles. Certain HLA alleles are known to confer the risk of various autoimmune diseases, including SSc, and several HLA alleles interact with other gene(s), leading to non-additive risk modification. Such HLA-nonHLA gene interaction in SSc pathology has not been identified yet.Material and Methods We newly genotyped 966 SSc cases and 4095 controls from France. The genotype imputation was conducted on the TopMed imputation server. We conducted GWAS for the French SSc dataset and updated the EUR GWAS meta-analysis by incorporating the French results into the latest EUR GWAS meta-analysis dataset. To analyze HLA associations in the French dataset, we used the latest version of SNP2HLA (BEAGLE v4) to impute HLA alleles, amino acids, and SNPs within the MHC region in 1,523 SSc cases and 4,581 controls.Results The updated EUR GWAS meta-analysis identified one novel association in the ERAP1 (OR 0.87, 95% CI 0.83-0.92, P=3.0X10-8), which encodes a protein that trims peptides prior to their presentation by HLA-class I molecules. The lead SNP was an expression quantitative trait locus for ERAP1 in various cell types relevant to SSc pathology, such as whole blood, skin, fibroblasts, or EBV-transformed lymphocytes. Notably, the association was stronger than the EAS, not due to differences in allele frequencies (EAS_OR 0.96, 95%CI 0.88-1.04, P=0.31), indicating an EUR-specific association of this locus. Intriguingly, the same SNP has been reported for its association with psoriasis (Pso), but the direction of the effect is opposite between SSc and Pso. HLA-C*06, the most significantly associated HLA class I allele with Pso, with its odds ratio (OR) ~4.0-6.0 in multiple ancestries, was not strongly associated with SSc (OR 1.20). Furthermore, the associations in the MHC-I region showed a significant negative correlation between SSc and Pso (p=4.0X10-6), suggesting an overall opposite direction of associations in the MHC class I between SSc and Pso. On the other hand, we did not observe these patterns in non-HLA regions--effect size directions are opposite in only ~ half of the non-MHC lead variants for Pso. These data indicate that the ERAP1-HLA-I axis plays an opposing role in the development of SSc and Pso, two inflammatory skin disease entities.Conclusions The largest EUR GWAS meta-analysis for SSc provides insights into a novel disease mechanism of SSc, particularly in light of its unexpected association with Pso.Abstract P.002 Figure 1