Combined Analysis of ceRNA Regulatory Network and DNA Methylation in CAD

Abstract BackgroundThe mobidity and mortality of coronary artery disease (CAD) is increasing year by year. Hence it is urgent to probe into the molecular mechanism of CAD and seek new therapeutic strategies. The purpose of our study was to screen genes associated with the development of CAD by using bioinformatics tools and clinical samples. MethodsMicroarray datasets from the Gene Expression Omnibus (GEO) database of peripheral blood cells (PBLs) were chosen for this study, and candidate differentially expressed microRNAs (DEMs) were screened using the limma and weighted co-expression network analysis (WGCNA) packages in R (v4.0). Subsequently, we construct a competitive endogenous RNAs (ceRNA) network and perform enrichment analysis of genes in the network. Meanwhile, differentially methylated genes (DMGs) in PBLs were identified using the "ChAMP" package in a DNA methylation chip. We then constructed the methylation-associated ceRNA network in CAD. Eventually, the methylation levels of genes and the relationship with the expression of genes in ceRNA were validated in PBLs samples using the Illumina Methylation 850K chip and transcriptome sequencing, while gene expressions were verified by qRT-PCR. And the regulation of DNA methylation on gene expression was verified in the THP-1 cells treated with 5-Aza-2'-deoxycytidine (5-AZA). ResultsA total of 71 differentially expressed miRNAs were screened by both WGCNA and limma. Then the ceRNA network in CAD was constructed with 269 nodes and 705 edges, which were significantly enriched in the chemokine-mediated signaling pathway and so on. Furthermore, from 4354 identified DMGs in a methylation data, 34 methylation-associated differentially expressed genes (DEGs) and 1 differentially expressed lncRNA (DEL) were obtained. After verification of methylation experiments in study population A, three genes were found to have altered methylation consistent with the bioinformatics results. And these genes were correlated in terms of methylation and expression levels. Corresponding with the bioinformatics results, qRT-PCR results in validation set B also showed that the expression of AGPAT4 and FAM169A were significantly lower in CAD. In addition, 5-AZA treatment could increase the expression of AGPAT4 and FAM169A in THP-1 cells. ConclusionsOur study deepens the understanding of the molecular mechanisms underlying the pathogenesis of CAD and provides new ideas for its treatment.