Identification and analysis of DNA methylation inflammation related key genes in intracerebral hemorrhage

Abstract

Abstract Background: Inflammation and DNA methylation have been reported to play key roles in intracerebral hemorrhage (ICH). The proposed study intended to investigate new diagnostic biomarkers associated with inflammation and DNA methylation through comprehensive bioinformatics approaches. Methods: GSE179759 and GSE125512 were sourced via the Gene Expression Omnibus (GEO) database, and 3222 inflammation-related genes (IFRGs) were downloaded from the Molecular Signatures Database (MSigDB). Key differentially expressed methylation-regulated and inflammation-related genes (DE-MIRGs) were achieved by overlapping methylation-regulated differentially expressed genes (MeDEGs) between ICH patients and control samples, module genes from Weighted Correlation Network Analysis (WGCNA), and the IFRGs. The functional annotation of DE-MIRGswas performed by Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) resources. A protein-protein interaction (PPI) network was further constructed to clarify the interrelationships between the different DE-MIRGs. The key genes were categorized by Least Absolute Shrinkage Selection Operator (LASSO), and support vector machine recursive feature elimination (SVM-RFE), and subsequently performed Gene Set Enrichment Analysis (GSEA). Results: A number of 22 DE-MIRGs were acquired among 451 MeDEGs, 3222 IFRGs and 302 module genes, and they were mainly enriched in GO terms of wound healing, blood coagulation and hemostasis; KEGG pathways of PI3K-AKT signaling pathway, Focal adhesion, and Regulation of actin cytoskeleton. A PPI network with 22 nodes and 87 edges was constructed based on the 22 DE-MIRGs, and 11 of them were selected for the following key gene selection. Moreover, 2 key genes (SELP and S100A4) were obtained according to LASSO and SVM-RFE. Finally, SELP was mainly enriched in Cell morphogenesis involved in differentiation, Cytoplasm translation, and Actin binding of GO terms, and the KEGG pathway including Edocytosis, Focal adhesion, and Platelet activation. S100A4 was major enriched in GO terms including Mitochondrial inner membrane, Mitochondrial respirasome, and Lysosomal membrane; Oxidative phosphorylation, Regulation of actin cytoskeleton, and Chemical carcinogensis-reactive oxygen species in KEGG pathways. Conclusion: 22 DE-MIRGs were identified associated with inflammation and DNA methylation between ICH patients and normal controls, and 2 key genes (SELP and S100A4) were obtained and regarded as the biomarker for ICH, which could provide the research foundation for the further pathological mechanism investigation of ICH.