Abstract
Coronary artery disease (CAD) is a complex condition, the development of which involves interaction between both genetic and environmental factors. Epigenetic mechanisms, including DNA methylation, may contribute significantly to disease pathogenesis. The DNA methylation profile of CAD has been relatively poorly described in the current literature. Therefore, an opportunistic post-hoc analysis of an Otago-based cardiovascular disease epigenome-wide association study (EWAS) cohort was conducted to investigate differential methylation patterns at cytosine-phosphate-guanine dinucleotide (CpG) sites associated with coronary artery disease.
Genome-wide DNA methylation profiles of 487 males taken from whole blood samples were analysed using the Illumina Infinium HumanMethylation450 BeadChip array, which assessed 456,279 CpG sites across the human genome. A post-hoc EWAS based on CAD status was conducted on this cohort which included 203 CAD cases and 284 controls. Methylation data was analysed using principal component analysis and multivariate regression.
Of all CpG sites analysed, 6.4% (29,336 CpG sites) demonstrated genome-wide significant associations with CAD (P-value<5x10-6), after adjustment for age and blood cell composition. The top three CpG sites (cg22671939, cg00589493, cg14042137) showed particularly strong associations with CAD, consistent across all adjustment models for conventional risk factors of CAD. Evaluation of genes associated with the top 1500 CpG sites using Metascape pathway enrichment analysis identified biological pathways with potential CAD-related plausibility relating to vascular morphology, immune processes, and nervous systems.
Additionally, a form of sensitivity analysis was conducted to evaluate the ability of the dataset to detect existing differential methylation patterns associated with the risk factors of age, smoking exposure, and type II diabetes mellitus. These post-hoc EWAS produced results that demonstrated concordance with epigenetic associations purportedly associated with the specific phenotypes, as reported in previously published EWAS. While suggesting that the CAD EWAS dataset may have the sensitivity with which to detect real associations, such validation also demonstrates the possible robustness of CpG site methylation as disease biomarkers. Epigenetic markers may have potential in disease risk prediction, guidance of disease management decisions, and development of precision medicine.
A large number of differentially methylated CpG sites associated with CAD were identified by this CAD EWAS. These DNA methylation markers may have potential as indicators of, and treatment targets for, coronary artery disease. However, further study in a de novo CAD EWAS setting is required to validate the associations identified by this post-hoc discovery EWAS, and broaden generalisability of study findings.