Investigating the shared genetic architecture between multiple sclerosis and inflammatory bowel diseases

Abstract

An epidemiological association between multiple sclerosis (MS) and inflammatory bowel disease (IBD) is well established, but whether this reflects a shared genetic aetiology, and whether consistent genetic relationships exist between MS and the two predominant IBD subtypes, ulcerative colitis (UC) and Crohn's disease (CD), remains unclear. Here, we use large-scale genome-wide association study summary data to investigate the shared genetic architecture between MS and IBD overall and UC and CD independently. We find a significantly greater genetic correlation between MS and UC than between MS and CD, and identify three SNPs shared between MS and IBD (rs13428812), UC (rs116555563) and CD (rs13428812, rs9977672) in cross-trait meta-analyses. We find suggestive evidence for a causal effect of MS on UC and IBD using Mendelian randomization, but no or weak and inconsistent evidence for a causal effect of IBD or UC on MS. We observe largely consistent patterns of tissue-specific heritability enrichment for MS and IBDs in lung, spleen, whole blood and small intestine, and identify cell-type-specific enrichment for MS and IBDs in CD4⁺ T cells in lung and CD8⁺ cytotoxic T cells in lung and spleen. Our study sheds light on the biological basis of comorbidity between MS and IBD.

Fig. 1. Summary of pairwise genetic correlations estimated using linkage disequilibrium score regression (LDSC) with and without constrained intercept.

Bars represent the point estimates of genetic correlation for each disease pair. Error bars represent the 95% confidence intervals (CIs) of the estimated genetic correlations. See Table S1 for complete LDSC results. MS: multiple sclerosis. IBD: inflammatory bowel disease. CD: Crohn’s disease. UC: ulcerative colitis. Source data are provided with this paper.

Fig. 2. Local genetic correlations (r_g) between multiple sclerosis (MS) and inflammatory bowel disease (IBD; top), ulcerative colitis (UC; middle) and Crohn’s disease (CD; bottom), respectively.

Left: average local r_g estimates for each disease pair in regions harbouring disease-specific risk variants (p < 5 × 10⁻⁸), regions harbouring shared risk variants (“intersection”) and all other regions (“neither”). Local genetic correlations with estimates less than −1 or greater than 1 were forced to −1 or 1, respectively. Error bars represent the 95% confidence intervals (CIs), calculated using a jack-knife method. Right: density distribution of local r_g estimates for each disease pair in disease-specific regions (red, green), intersection regions (blue) and other (purple) regions. For MS-IBD, 45, 45, 8, and 572 regions were included in the ‘MS-specific’, ‘IBD-specific’, ‘intersection’, and ‘neither’ categories; for MS-UC, 38, 23, 7, and 456 regions were included in the ‘MS-specific’, ‘UC-specific’, ‘intersection’, and ‘neither’ categories; for MS-CD, 39, 32, 7, and 480 regions were included in the ‘MS-specific’, ‘CD-specific’, ‘intersection’, and ‘neither’ categories. Source data are provided with this paper.

Fig. 3. Summary of bi-directional Mendelian Randomisation (MR) analyses between multiple sclerosis (MS) and each of inflammatory bowel disease (IBD), ulcerative colitis (UC) and Crohn’s disease (CD).

Green: Causal Analysis Using Summary Effect estimates (CAUSE); dark blue: Generalised Summary-data-based Mendelian Randomisation (GSMR); red: MR-Egger; orange: inverse variance weighting (IVW); purple: weighted median; light blue: weighted mode. Error bars represent the 95% confidence intervals (CIs) for the associated MR point estimates. See Tables S4, S6 for complete details of the MR analyses. Source data are provided with this paper.

Fig. 4. Tissue type-specific enrichment of single nucleotide polymorphism (SNP) heritability for multiple sclerosis (MS), inflammatory bowel disease (IBD), ulcerative colitis (UC) and Crohn’s disease (CD) in immune tissues estimated using stratified linkage disequilibrium score regression (S-LDSC).

Negative log10 p-values of coefficient Z-scores for each individual test (two-tailed Z-test) are displayed on the x axis. The grey and pink dotted lines represent the FDR threshold <5% and Bonferroni corrected threshold for multiple comparisons, respectively. Original indicates results from S-LDSC analyses adjusted for the baseline model and the set of all genes. Conditional indicates results from conditional S-LDSC analyses that additionally adjusted for the set of genes specifically expressed in the three non-focal tissues (e.g. small intestine–terminal ileum, lung and whole blood in analyses of spleen). Source data are provided with this paper.

Fig. 5. Cell-type-specific enrichment of single nucleotide polymorphism (SNP) heritability for multiple sclerosis (MS), inflammatory bowel disease (IBD), ulcerative colitis (UC) and Crohn’s disease (CD) in immune tissues estimated using stratified linkage disequilibrium score regression (S-LDSC).

Cell types are included if they showed FDR-significant enrichments in at least one disease. Negative log10 p-values of coefficient Z-scores for each individual test (two-tailed Z-test) are displayed on the x-axis. The grey and pink dotted lines represent the FDR threshold <5% and Bonferroni corrected threshold for multiple comparisons, respectively. Original indicates results from S-LDSC analyses adjusted for the baseline model and the set of all genes. Conditional indicates results from conditional S-LDSC adjusted for the baseline model, the set of all genes, the set of genes specifically expressed in the three non-focal tissues (e.g. small intestine–terminal ileum, lung and whole blood in analyses of spleen) and the set of genes highly expressed in other FDR-significant cell types of the same tissue and same disease (e.g. spleen: B/T doublet and spleen: CD8⁺ cytotoxic lymphocytes in analyses of spleen: B hypermutation cells). Source data are provided with this paper.