Coeliac Disease Research Center, Faculty of Medicine and Health Technology, Tampere University, Tampere, Finland.
Research Programs Unit, Immunobiology, and the Haartman Institute, Department of Molecular Genetics, University of Helsinki, Helsinki, Finland.
J Hum Genet. 2021 Jun;66(6):613-623. doi: 10.1038/s10038-020-00888-5. Epub 2021 Jan 15.
The phenotype of coeliac disease varies considerably for incompletely understood reasons. We investigated whether established coeliac disease susceptibility variants (SNPs) are individually or cumulatively associated with distinct phenotypes. We also tested whether a polygenic risk score (PRS) based on genome-wide associated (GWA) data could explain the phenotypic variation. The phenotypic association of 39 non-HLA coeliac disease SNPs was tested in 625 thoroughly phenotyped coeliac disease patients and 1817 controls. To assess their cumulative effects a weighted genetic risk score (wGRS39) was built, and stratified by tertiles. In our PRS model in cases, we took the summary statistics from the largest GWA study in coeliac disease and tested their association at eight P value thresholds (P) with phenotypes. Altogether ten SNPs were associated with distinct phenotypes after correction for multiple testing (P ≤ 0.05). The TLR7/TLR8 locus was associated with disease onset before and the SH2B3/ATXN2, ITGA4/UBE2E3 and IL2/IL21 loci after 7 years of age. The latter three loci were associated with a more severe small bowel mucosal damage and SH2B3/ATXN2 with type 1 diabetes. Patients at the highest wGRS39 tertiles had OR > 1.62 for having coeliac disease-related symptoms during childhood, a more severe small bowel mucosal damage, malabsorption and anaemia. PRS was associated only with dermatitis herpetiformis (P = 0.2, P = 0.02). Independent coeliac disease-susceptibility loci are associated with distinct phenotypes, suggesting that genetic factors play a role in determining the disease presentation. Moreover, the increased number of coeliac disease susceptibility SNPs might predispose to a more severe disease course.
由于不完全清楚的原因，乳糜泻的表型存在很大差异。我们研究了已确定的乳糜泻易感性变异（SNP）是否单独或累积与不同的表型相关。我们还测试了基于全基因组关联（GWA）数据的多基因风险评分（PRS）是否可以解释表型变异。在 625 名经过详细表型分析的乳糜泻患者和 1817 名对照中，测试了 39 个非 HLA 乳糜泻 SNP 的表型关联。为了评估它们的累积效应，构建了加权遗传风险评分（wGRS39），并按三分位数分层。在我们的病例 PRS 模型中，我们采用了乳糜泻最大 GWA 研究的汇总统计数据，并在八个 P 值阈值（P）下测试了它们与表型的关联。经过多重检验校正（P≤0.05）后，共有 10 个 SNP 与不同的表型相关。TLR7/TLR8 基因座与疾病发病前相关，SH2B3/ATXN2、ITGA4/UBE2E3 和 IL2/IL21 基因座与 7 岁后相关。后三个基因座与更严重的小肠黏膜损伤相关，SH2B3/ATXN2 与 1 型糖尿病相关。wGRS39 最高三分位数的患者在儿童期发生乳糜泻相关症状、小肠黏膜损伤更严重、吸收不良和贫血的 OR 值>1.62。PRS 仅与疱疹样皮炎相关（P=0.2，P=0.02）。独立的乳糜泻易感性基因座与不同的表型相关，表明遗传因素在决定疾病表现方面发挥作用。此外，增加的乳糜泻易感性 SNP 数量可能会导致更严重的疾病过程。
