Effects of DNA Methylation on Cardiovascular Disease, Target Organ Damage, and their Risk Factors in African Americans

Abstract

Cardiovascular disease (CVD) is the leading cause of mortality among US adults, and African Americans have a higher burden of CVD morbidity and mortality than any other racial group. Identification of novel CVD biomarkers is essential to better identify at-risk individuals, advance precision medicine, and inform efforts to reduce CVD burden. DNA methylation (DNAm) is an epigenetic mechanism that captures genetic influences as well as imprints of lifestyle and environmental exposures throughout the life course. DNAm patterns may help identify biological mechanisms contributing to CVD pathogenesis. This dissertation explores the effects of DNAm and DNAm-based epigenetic age acceleration (EAA) on cardiometabolic risk factors, atherosclerosis, CVD incidence, and target organ damage from hypertension (TOD) in African American participants from the Genetic Epidemiology Network of Arteriopathy (GENOA). In Aim 1, we used univariate and multivariate linear mixed models to assess the epigenome-wide association between DNAm sites (CpGs) and measures of TOD in the heart (left ventricular mass index (LVMI) and relative wall thickness (RWT)), kidneys (estimated glomerular filtration rate (eGFR) and albuminuria), and brain (white matter hyperintensity). LVMI, RWT, and albuminuria were each associated with one CpG in univariate models, and seven CpGs were associated with TOD measures in the multivariate (pleiotropy) model (false discovery rate (FDR) < 0.1). Mendelian randomization analysis provided evidence of a causal pathway between three CpGs and eGFR. In Aim 2, we assessed the associations between four measures of EAA, 10 cardiometabolic risk factors, and CVD incidence. We then evaluated whether EAA improved the predictive accuracy of two clinically-used CVD risk scores: the Framingham risk score (FRS) and the Atherosclerotic Cardiovascular Disease risk equation (ASCVD). Increased biological aging, as assessed by EAA, was associated with worse cardiometabolic risk profile, but the associations between each of the four EAA measures differed across cardiometabolic risk factors. GrimAge acceleration (GrimAA) was associated with CVD incidence (hazard ratio per 5-year increase 1.47, 95% CI: 1.05 – 2.01, P = 0.024) after adjusting for traditional CVD risk factors. GrimAA improved model fit over clinical risk scores using likelihood ratio tests (P = 0.013 for FRS, P = 0.008 for ASCVD), did not improve C statistics (P > 0.05), and marginally improved net reclassification index (NRI) which assesses reassignment of risk categories (NRI = 0.055, 95% CI: 0.040 – 0.071 for FRS; 0.029, 95% CI: 0.006 – 0.064 for ASCVD). In Aim 3, we evaluated in prospective analyses whether EAA measures, previously identified atherosclerosis-associated CpGs, and methylation risk scores (MRSs) derived from these CpGs are associated with single- or multi-site atherosclerosis (coronary artery calcification (CAC), abdominal aorta calcification (AAC), and ankle-brachial index (ABI)). One and six CpGs were associated with AAC and multi-site atherosclerosis, respectively (FDR < 0.1). Both a 5-year increase in GrimAA as well as a one unit increase in the MRS for carotid artery plaque were associated with a 1.6-fold increase in AAC and 0.7 units increase in multi-site atherosclerosis (score range: 0-12) after adjusting for CVD risk factors (Bonferroni-adjusted P < 0.05). Together, these studies support the premise that DNAm plays an important role in CVD and TOD and is a promising biomarker that may improve risk assessment in African Americans.