Background/Purpose: – Numerous juvenile idiopathic arthritis (JIA) risk loci have been identified, overwhelmingly from cohorts of children of European ancestry (EA). The extent to which these regions confer risk in individuals of non-European ancestry is unknown, although JIA occurs in children of African ancestry as frequently as in those of European ancestry {Schwartz et al, J Rheum 1997). We posit that the functional mechanism of the true risk-driving variants should generally span ancestries. Thus, identifying the causal variants on risk haplotypes in EA will facilitate assessment of risk in non-European populations, minimizing the effects of linkage disequilibrium.     

Methods: Using published JIA risk regions from EA cohorts (3,939 children with JIA and 14,412 healthy controls), we selected regions containing single nucleotide polymorphisms (SNPs) that met FDR P< 0.05. In these regions, we identified those SNP residing in open chromatin in multiple immune cell types (neutrophils, CD4+ T cells) using ATACseq and regulatory regions identified using H3K4me1 and H3K27ac marks. We further filtered these SNPs for those located in regions of bi-directional transcription initiation characteristic of non-coding regulatory regions detected using dREG to analyze global run-on sequencing (GROseq) data in resting human CD4+ T cells. We examined the allele frequencies of these SNPs in multiple ancestries using NIH Ldlink/Ldhap.

Results: We identified 846 SNPs meeting FDR P< 0.05 spanning 24 known JIA risk regions and that were situated within regions of open chromatin in neutrophils or CD4+ T cells.  Of these SNPs, 138 SNPs were within regions of bi-directional transcription initiation characteristic of non-coding regulatory regions. These 138 SNPs are therefore strong causal candidates. We found that 86% of these SNPs could be identified in individuals of African ancestry identified in Ldlink, although there were differences in allele frequency. In individuals of African ancestry, 18 flipped major/minor allele frequency compared to EA and 49 had increased allele frequency of 0.05 in AA. Thus, although the allele frequencies differed across ancestries in some cases, most of our 138 candidate causal SNPs were identified in both EA and AA. 

Conclusion: Plausible causal SNPs situated within known JIA risk loci can be identified in individuals of both African and European ancestry, suggesting that, within these loci, the same genetic mechanisms driving risk are likely to be operating.  Our functional prioritization pipeline of JIA-associated SNPs identified 138 high priority candidate SNPs for investigation across ancestral groups. Coupled with the knowledge of allele frequencies, analyzing a priori functionally-SNPs greatly increases the statistical power to detect associations in other ancestries. Current work is focused on exploring these associations in the UK Biobank data, emphasizing SNPs where allele frequency and sample size provides sufficient statistical power.

« Back to ACR Convergence 2024

ACR Meeting Abstracts - https://acrabstracts.org/abstract/juvenile-idiopathic-arthritis-genetic-risk-haplotypes-relevance-to-children-of-african-ancestry/