Exploration of core mitochondrial genes and epigenetic regulatory mechanisms in cardiac arrest via multi-omics integrated analysis

Abstract

There is a lack of effective molecular biomarkers for the early identification of the risk of cardiac arrest. Mitochondrial dysfunction is widely recognised as being involved in the pathophysiological process of cardiac arrest, but the upstream epigenetic regulatory mechanisms remain unclear. This study takes DNA methylation as its central focus. By systematically evaluating the genetic associations between methylation quantitative trait loci (mQTLs) in mitochondrial genes and the risk of cardiac arrest, and by tracking downstream changes in gene expression and protein abundance, the study aims to explore potential regulatory pathways linking epigenetic variations to molecular phenotypes and ultimately to disease. The study utilised mitochondrial genes from MitoCarta 3.0, obtaining corresponding methylation quantitative trait loci (mQTL) from GoDMC, gene expression quantitative trait loci (eQTL) from eQTLGen and GTEx, and protein quantitative trait loci (pQTL) from UKB-PPP and deCODE. These were used as instrumental variables for Mendelian randomisation (MR), summary-data-based Mendelian randomization (SMR) and colocalisation analysis, to systematically screen for multi-omics genetic signals in mitochondrial genes associated with cardiac arrest (from FinnGen). Based on the analysis results, multi-omics evidence was categorised, and single-cell expression profiles and machine learning models were utilised to perform functional annotation and assess the discriminatory power of candidate genes. The analysis identified five mitochondrial genes with a directional genetic association with cardiac arrest: GATM, GRHPR, NDUFS6, PMAIP1 and AGXT. Among these, the cg10760299 locus of GATM was negatively associated with the risk of cardiac arrest (OR = 0.875, 95% CI 0.832–0.921, p < 0.001), whilst gene expression (OR = 1.396, 95% CI 1.194–1.633, p = 0.005) and protein abundance (OR = 1.513, 95% CI 1.228–1.863, p = 0.015) were positively associated with the risk of cardiac arrest. Single-cell expression profiles indicated that key genes exhibited expression differences between cardiomyocytes and fibroblasts, whilst integrated learning models suggested that these key genes possess potential for disease discrimination. Through multi-omics integration analysis, this study identified association signals between methylation sites such as GATM/cg10760299 and the risk of cardiac arrest, providing new candidate targets for research into epigenetic biomarkers of cardiac arrest.