SNP Variants in Major Histocompatibility Complex Are Associated with Sarcoidosis Susceptibility-A Joint Analysis in Four European Populations

Abstract

Sarcoidosis is a multiorgan inflammatory disorder with heritability estimates up to 66%. Previous studies have shown the major histocompatibility complex (MHC) region to be associated with sarcoidosis, suggesting a functional role for antigen-presenting molecules and immune mediators in the disease pathogenesis. To detect variants predisposing to sarcoidosis and to identify genetic differences between patient subgroups, we studied four genes in the MHC Class III region (LTA, TNF, AGER, BTNL2) and HLA-DRA with tag-SNPs and their relation to HLA-DRB1 alleles. We present results from a joint analysis of four study populations (Finnish, Swedish, Dutch, and Czech). Patients with sarcoidosis (n = 805) were further subdivided based on the disease activity and the presence of Löfgren's syndrome. In a joint analysis, seven SNPs were associated with non-Löfgren sarcoidosis (NL; the strongest association with rs3177928, P = 1.79E-07, OR = 1.9) and eight with Löfgren's syndrome [Löfgren syndrome (LS); the strongest association with rs3129843, P = 3.44E-12, OR = 3.4] when compared with healthy controls (n = 870). Five SNPs were associated with sarcoidosis disease course (the strongest association with rs3177928, P = 0.003, OR = 1.9). The high linkage disequilibrium (LD) between SNPs and an HLA-DRB1 challenged the result interpretation. When the SNPs and HLA-DRB1 alleles were analyzed together, independent association was observed for four SNPs in the HLA-DRA/BTNL2 region: rs3135365 (NL; P = 0.015), rs3177928 (NL; P < 0.001), rs6937545 (LS; P = 0.012), and rs5007259 (disease activity; P = 0.002). These SNPs act as expression quantitative trait loci (eQTL) for HLA-DRB1 and/or HLA-DRB5. In conclusion, we found novel SNPs in BTNL2 and HLA-DRA regions associating with sarcoidosis. Our finding further establishes that polymorphisms in the HLA-DRA and BTNL2 have a role in sarcoidosis susceptibility. This multi-population study demonstrates that at least a part of these associations are HLA-DRB1 independent (e.g., not due to LD) and shared across ancestral origins. The variants that were independent of HLA-DRB1 associations acted as eQTL for HLA-DRB1 and/or -DRB5, suggesting a role in regulating gene expression.

Keywords: BTNL2; DRB1; HLA; SNP; haplotype; major histocompatibility complex; prognosis; sarcoidosis.

Figure 1

HLA-DRB1-SNP haplotype frequencies. HLA-DRB1 data were available for Finnish, Swedish, and Czech samples. Subjects with missing SNP genotype were excluded from the analysis. SNP order is the same as in the Table 2. The analysis shows population-specific differences in haplotype structures. Haplotype frequencies (2n) > 0.01 are presented.

Figure 2

The meta-analysis of non-Löfgren sarcoidosis (NL) vs. controls with Finnish, Swedish, Dutch, and Czech samples. The linkage disequilibrium (LD) information (r²) is shown by color-coded dots and the recombination rates where the peaks are the hotspots are based on the HapMap data. The LD structure (r²) in the current study for the top significant SNPs (I² < 25) are shown on the right. After adjusting for HLA-DRB1 alleles, the SNPs that remained significant are pointed with an arrow.

Figure 3

The meta-analysis of Löfgren (LS) vs. controls with Finnish, Swedish, and Czech samples. The linkage disequilibrium (LD) information (r²) is shown by color-coded dots and the recombination rates where the peaks are the hotspots are based on the HapMap data. The LD structure (r²) in the current study for the top significant SNPs (I² < 25) are shown on the right. After adjusting for HLA-DRB1 alleles, the SNPs that remained significant are pointed with an arrow.

Figure 4

The meta-analysis of sarcoidosis subgroups (resolved sarcoidosis, NLR vs. persistent sarcoidosis, NLP) with Finnish, Swedish, Dutch, and Czech samples. The linkage disequilibrium (LD) information (r²) is shown by color-coded dots and the recombination rates where the peaks are the hotspots are based on the HapMap data. The LD structure (r²) in the current study for the top significant SNPs (I² < 25) are shown on the right. After adjusting for HLA-DRB1 alleles, the SNPs that remained significant are pointed with an arrow.