Altered DNA methylation within DNMT3A, AHRR, LTA/TNF loci mediates the effect of smoking on inflammatory bowel disease

Abstract

This work aims to investigate how smoking exerts effect on the development of inflammatory bowel disease (IBD). A prospective cohort study and a Mendelian randomization study are first conducted to evaluate the association between smoking behaviors, smoking-related DNA methylation and the risks of Crohn's disease (CD) and ulcerative colitis (UC). We then perform both genome-wide methylation analysis and co-localization analysis to validate the observed associations. Compared to never smoking, current and previous smoking habits are associated with increased CD (P = 7.09 × 10^-10) and UC (P < 2 × 10^-16) risk, respectively. DNA methylation alteration at cg17742416 [DNMT3A] is linked to both CD (P = 7.30 × 10^-8) and UC (P = 1.04 × 10^-4) risk, while cg03599224 [LTA/TNF] is associated with CD risk (P = 1.91 × 10^-6), and cg14647125 [AHRR] and cg23916896 [AHRR] are linked to UC risk (P = 0.001 and 0.002, respectively). Our study identifies biological mechanisms and pathways involved in the effects of smoking on the pathogenesis of IBD.

Fig. 1. The association of gene regulated by smoking-related DNA methylation with the risk of Crohn’s disease.

This analysis was conducted basing on the data from the smoking EWAS (Joehanes R, et al., N = 15,907) and GWAS summary level data of IBD (de Lange KM, et al., N = 40,266, 12,366 UC cases). MR estimates were measured through the Wald ratio approach and combined using the IVW method if there were more than one independent instrumental variant. All tests were two-sided and adjusted for multiple comparisons. Manhattan plot showing the -log₁₀(p) of association between each CpG site with the risk of CD on the y-axis against genomic position on the x-axis. Gene names are italicized. The blue line means P = 0.05, and the red line means the threshold of false discovery rate correction. Source data are provided as Source Data for Figs. 1 and 2.

Fig. 2. The association of gene regulated by smoking-related DNA methylation with the risk of ulcerative colitis.

This analysis was conducted basing on the data from the smoking EWAS (Joehanes R, et al., N = 15,907) and GWAS summary level data of IBD (de Lange KM, et al., N = 45,975, 12,194 CD cases). MR estimates were measured through the Wald ratio approach and combined using the IVW method if there were more than one independent instrumental variant. All tests were two-sided and adjusted for multiple comparisons. Manhattan plot showing the -log₁₀(p) of association between each CpG site with the risk of CD on the y-axis against genomic position on the x-axis. Gene names are italicized. The blue line means P = 0.05, and the red line means the p-value threshold of FDR correction. Source data are provided as Source Data for Figs. 1 and 2.

Fig. 3. Regional plot of colocalization evidence of CpG site methylation and Crohn’s disease susceptibility.

Each dot represents a SNP. Gene names are italicized. Panel A displays the -log₁₀(p) of the SNP and mQTL in corresponding GWAS and EWAS. In the Panel B and C, the x-axis demonstrates the base position of the SNP and mQTL, and the y-axis is the -log₁₀(p) of the SNP in genome-wide association study and mQTL in genome-wide DNA methylation analysis, respectively. SNP that passed a colocalization threshold of posterior of H4 > 0.8 are highlighted in purple. Source data are provided as Source Data for Fig. 3.

Fig. 4. The study design of this study. CD, Crohn’s disease. UC, ulcerative colitis.

SNPs, single nucleotide polymorphisms. MR, Mendelian Randomization. GWAS, genome-wide association study. mQTL, methylation quantitative trait loci. IBD, inflammatory bowel disease. IBD-U, unspecific inflammatory bowel disease. DMP, differentially methylated position. EWAS, epigenome-wide association study.