Jiantao Ma, Chao-Qiang Lai, Xinmin S. Li, Meng Wang, Zeneng Wang, Jie Yao, Xiuqing Guo, Kent D. Taylor, Soyoung Lee, Russell P. Tracy, Durda Peter, Yongmei Liu, Jerome I. Rotter, Stephen S. Rich, M B Matthew Budoff, W. H. Wilson Tang, Joseph A. DiDonato, Jennifer A. Brody, Rozenn N. Lemaitre, Mandy Fretts, Nona Sotoodehnia, Bruce M. Psaty, José M. Ordovás, David S. Siscovick, Stanley L. Hazen, Dariush Mozaffarian
BACKGROUND: Trimethylamine N-oxide (TMAO) and its related metabolites have been linked to cardiovascular disease (CVD), but their impact on DNA methylation remains unclear. Investigating these relationships may clarify the role of epigenetic mechanisms in diseases. METHODS: This study analyzed data from 1,356 adults from the Cardiovascular Health Study (CHS) and the Multi-Ethnic Study of Atherosclerosis (MESA). Using stable-isotope dilution liquid chromatography with on-line electrospray ionization tandem mass spectrometry (LC-MS), we quantified TMAO and five related metabolites. DNA methylation levels were measured using Illumina BeadChip arrays. Epigenome-wide association analyses and meta-analyses were conducted across approximately 430,000 CpG sites. To explore the functional significance of the identified CpGs, we performed gene set enrichment analysis and Mendelian randomization (MR) analyses. RESULTS: We identified 143 metabolite-CpG pairs at FDR < 0.05, including four CpGs for TMAO (P ≤ 4.03e-7), 12 for betaine (P ≤ 1.19e-6), 53 for γ-butyrobetaine (P ≤ 6.11e-6), five for carnitine (P ≤ 5.42e-7), six for choline (P ≤ 2.81e-7), and 63 for crotonobetaine (P ≤ 7.25e-6). CpGs associated with γ-butyrobetaine showed moderate correlation with crotonobetaine-associated CpGs. In total, these metabolite-linked CpGs were mapped to 108 genes. Gene set enrichment analysis revealed 145 significantly enriched gene sets, including nine highly relevant to CVD risk. Furthermore, CpGs were enriched in 80 immunologic signature gene sets (FDR < 0.05). MR analysis identified three CpGs associated with coronary artery disease (CAD), including hypermethylation at cg18705301 (NDUFAF1), which was inversely associated with betaine levels and linked to a lower risk of CAD (P = 1.8e-5). CONCLUSION: This study identified specific DNA methylation sites associated with TMAO and related metabolites. These epigenetic changes may contribute to CVD risk through multiple pathways. Future research should validate these findings and explore their clinical implications.