Functional characterization of the sjögren's syndrome-associated locus DDX6-CXCR5
Recommended Citation
Khanam S, Joachims ML, Means N, Omdal R, Wahren-Herlenius M, Alevizos I, Witte T, Jonsson R, Rischmueller M, Rhodus NL, Montgomery C, Ng WF, Nordmark G, Adrianto I, Sivils K, and Lessard C. Functional characterization of the sjögren's syndrome-associated locus DDX6-CXCR5. Ann Rheum Dis 2018; 77:1267.
Document Type
Conference Proceeding
Publication Date
6-1-2018
Publication Title
Ann Rheum Dis
Abstract
Background: Sjögren's syndrome (SS) is a chronic, heterogeneous disease with hallmark features of auto-inflammation and autoantibody production. We previously identified association between the DDX6-CXCR5 locus and SS surpassing genome-wide significance.
Objectives: This study aims to determine the mechanism by which this association contributes to disease.
Methods: Fine mapping and imputation allowed enrichment of existing genetic datasets from a total of 1916 SS cases and 3684 controls with 971 testable variants in the DDX6-CXCR5 interval. Candidate variants were prioritized using statistical and bioinformatics approaches. Electromobility shift assays (EMSAs) and pull-downs (PDs) followed by mass spectrometry (MS) were used to determine allelic-specific differences in binding using lysates from HSB-2 (T), Jurkat (T), Reh (B), Ramos (B), Daudi (B), THP-1 (monocyte), and HEK 293T (epithelial) cells.
Results: Bioinformatic analysis of the top associated variants after imputation (rs7125066 and rs7119038) in the DDX6-CXCR5 region did not yield evidence of regional functionality. However, 46 other candidates that span the region of association were identified through imputation. Chromatin methylation pattern data from the Roadmap database showed several variants in this region were within transcription start sites or enhancer elements depending on the cell type and state. Using RegulomeDB, Haploreg, and other databases, rs4938572, rs12365699, rs57494551, and rs10892294 showed strong evidence affecting binding and/or expression of one or more target genes in the region and were selected for further study. Using EMSAs, statistically significant increase in binding for the risk allele as compared to the non-risk allele was found for rs4938572 (p<0.01) for cell lysates from HSB-2, Reh, THP-1, Ramos, Jurkat, and Daudi but not HEK 293T cells. While decreased binding of risk allele for rs10892294 and rs12365699 were found for Ramos cells only, rs57494551 showed decreased binding for both Ramos and THP-1 cell lysate. The risk allele of rs57494551 also decreases binding using THP-1 cell lysate. Preliminary MS analysis showed several immune related transcription factors likely binding this region, including: IRF8, GTFTI, RFX5, IKZF3, PAX5, and IKZF1. Cell type-specific expression of the genes in the region shows the expression of DDX6 spans many immune cells subsets while the CXCR5 expression is more restricted.
Conclusions: The genetic variants in the DDX6-CXCR5 locus that were selected based on bioinformatics data resulted in a difference in the direction of binding, with the risk allele all residing on the same risk haplotype. The change in binding of protein.to each associated variant likely would alter expression of both protein coding genes in a cell/context specific manner. Ongoing studies will assess the regulatory role of these sequences using luciferase assays and CRISPR/Cas9 based genetic modification of target SNPs in various cell types.
Volume
77
Issue
Supplement 2
First Page
1267
