Background/Purpose: While genome-wide association studies have provided valuable information about genetic risk for juvenile idiopathic arthritis (JIA), we are still unable to determine the actual genetic variants that exert the biological effects that confer risk.  This is because the tag SNPs used to identity the risk loci are in linkage disequilibrium with multiple other SNPs, the majority of which have no effect on disease risk.  The problem is further complicated by the fact that the JIA risk haplotypes are highly enriched for functional genomic elements other than genes, including H3K4me1/H3K27ac-marked enhancers.  Reasoning that the relevant (causal) variants on the JIA haplotypes are likely to be those that have demonstrable effects on gene expression, we used a massively parallel reporter assay (MPRA) to screen >4,500 genetic variants on the JIA risk haplotypes.

Methods: We used the MPRA (Tewhey et al, Cell 2016; 165, 1519) in 5×10⁸Jurkat T cells.  We synthesized 180mer oligos that represent the genetic variants of interest. These oligos were then barcoded and cloned into green fluorescence protein (GFP) reporter vectors, which were transfected into the Jurkat T cells. After 24 hr, GFP-mRNA was isolated and sequenced.  Since the oligos are barcoded, the effects of each variant on gene expression could be determined. Tag counts from the sequencing of the barcodes were summed for each variant. After read-depth normalization, activity estimates for each variant were calculated as a log-fold change of the tag abundance in the RNA relative to the number of tags observed in the plasmid library.  We identified sequences that exhibit regulatory activity using DESeq2, which leverages a negative binomial and pooled variance estimates based on abundance within the plasmid pool. Finally, we will identified variants that show differential activity by comparing the log of the RNA/plasmid ratios between alleles using Welch’s t-test. We repeated the experiment using 5 replicates.

Results: Our libraries contained 4,914 good-quality oligos sufficient for analysis.  We identified 52 SNPs in which the test variant demonstrated higher activity than the common allele.  As predicted, many of these SNPs were situated in H3K27ac-marked, non-coding regions (e.g., an intronic enhancer in LNPEP, intergenic enhancers in the IRF1locus). Of particular interest was the presence of 3 SNPs within an H3K4me1/H3K27ac-marked region situated betweenHLA-DRB1and HLA DQA1. Using a conventional reporter assay, we verified that this region (chr6:32,571,604-32,573,132) has enhancer activity.  When tested individually, each of the 3 SNPs identified on the MPRA altered the functional activity of this enhancer (p< 0.04). Finally, we interrogated publically available promoter capture HiC data from CD4+ T cells  and identified multiple genes, including non-HLA genes (e.g., AGER, C6orf10, TAP2), that physically interact with this enhancer region.

Conclusion: We demonstrate the feasibility of screening large numbers of genetic variants on the JIA risk haplotypes for functional activity.  This screening process elucidates previously unappreciated features on the established risk loci, such as the presence of an intergenic enhancer within the HLA class IIlocus.      

« Back to 2020 Pediatric Rheumatology Symposium

ACR Meeting Abstracts - https://acrabstracts.org/abstract/a-massively-parallel-reporter-assay-screen-of-genetic-variants-on-jia-haplotypes-reveals-variants-associated-with-altered-function-of-an-intergenic-enhancer-in-the-hla-class-ii-locus/