Variants at multiple loci implicated in both innate and adaptive immune responses are associated with Sjögren's syndrome

Abstract

Sjögren's syndrome is a common autoimmune disease (affecting ∼0.7% of European Americans) that typically presents as keratoconjunctivitis sicca and xerostomia. Here we report results of a large-scale association study of Sjögren's syndrome. In addition to strong association within the human leukocyte antigen (HLA) region at 6p21 (Pmeta = 7.65 × 10(-114)), we establish associations with IRF5-TNPO3 (Pmeta = 2.73 × 10(-19)), STAT4 (Pmeta = 6.80 × 10(-15)), IL12A (Pmeta = 1.17 × 10(-10)), FAM167A-BLK (Pmeta = 4.97 × 10(-10)), DDX6-CXCR5 (Pmeta = 1.10 × 10(-8)) and TNIP1 (Pmeta = 3.30 × 10(-8)). We also observed suggestive associations (Pmeta < 5 × 10(-5)) with variants in 29 other regions, including TNFAIP3, PTTG1, PRDM1, DGKQ, FCGR2A, IRAK1BP1, ITSN2 and PHIP, among others. These results highlight the importance of genes that are involved in both innate and adaptive immunity in Sjögren's syndrome.

Figure 1

Summary of genome-wide association results for 27,501 variants overlapping between DS1 and DS2 after imputation and meta-analysis. The −log₁₀(P) for each variant is plotted according to chromosome and base pair position. A total of seven loci (in red font) exceeded the GWS of P_meta < 5×10⁻⁸ (red dashed line). Suggestive threshold (P_meta < 5×10⁻⁵) is indicated by a blue dashed line.

Figure 2

Zoomed in plots of the meta-analysis results for the seven regions with P_meta<5×10⁻⁸. The HLA region (a) from Mb 28.4 to Mb 33.4 is shown for variants with P_meta<5×10⁻⁸ only. Strong linkage disequilibrium (LD) was observed with variants extending from Class I through Class II (pairwise LD with rs115575857 is given in blue). A second independent effect was identified in Class II peaking at rs116232857 (pairwise LD is depicted in red). The red insert on the left shows a further zoomed in view of the top associated variants in the Class II region. (b) The area of association surrounding IRF5 is shown, including the pairwise LD with the top overall variant, rs3757387, in blue. Pairwise LD with the top SNP, rs17339836, in the second effect is shown in red. STAT4 (c), IL12A (d), BLK (e), CXCR5 (f) and TNIP1 (g) regions are also provided and include pairwise LD with the top listed variant shown in blue. The black dashed lines represent P_meta = 5×10⁻⁸.

Figure 3

Identification of cis-eQTL in Sjögren’s syndrome-associated regions. Sjögren’s syndrome-associated variants in the HLA region within 100bp of RFX5 binding sites were evaluated for cis-eQTL with the expression of HLA genes in 222 European subjects. The top cis-eQTL in HLA-DRB6 (a), HLA-C (b), HLA-DPB1 (c), HLA-DQA1 (d), and HLA-A (e) are shown in the figure. We also identified cis-eQTL for top Sjögren’s syndrome-associated variants in IL12A (f), FAM167A-BLK (g, h), and TNIP1 (i). The risk genotype of each eQTL variant is plotted on the right side of each figure. Data are shown using box-and-whisker plots, in which distribution of the expression values are highlighted (median, upper quantile, lower quantile, maximum, and minimum). Each eQTL was evaluated using both a linear model and Analysis of Variance (ANOVA) model. Panel b only had one subject with the TT genotype; thus, we performed a t-test rather than an ANOVA to compare the TC to the CC genotype.