The role of mitochondrial dysfunction in the pathogenesis of atherosclerosis: A new exploration from bioinformatics analysis

Atherosclerosis (AS) is a complex cardiovascular disease associated with mitochondrial dysfunction (MD), which contributes to plaque formation and instability. This study explores the relationship and shared risk factors between the pathogenesis of AS and MD, aiming to advance preventive and therapeutic strategies for these comorbidities. Data from GSE28829, which includes 13 early and 16 advanced atherosclerotic plaque samples from human carotid arteries, were retrieved from the Gene Expression Omnibus database. Mitochondrial-related genes were sourced from the MitoCarta3.0 dataset. Differentially expressed genes were identified using the "limma R" package in R Studio. A gene co-expression network was constructed using the GeneMANIA database, and gene ontology and Kyoto Encyclopedia of Genes and Genomes pathway analysis were conducted using "clusterProfiler" R package. Candidate co-feature genes were identified using least absolute shrinkage and selection operator, random forest and support vector machine methods. Single sample gene set enrichment analysis on co-feature genes, and co-expression patterns and differential expression were visualized. Drug-protein interactions were predicted using the Drug Signature database, and molecular docking was used to select stable structures. A total of 571 differentially expressed genes and 15 interacting genes were obtained. Gene ontology functional enrichment primarily focused on pathways such as nuclear division and mitotic nuclear division, whereas Kyoto Encyclopedia of Genes and Genomes functional enrichment primarily focused on cell cycle, cellular senescence, and oocyte meiosis. 4 co-feature genes - ALDH1B1, CRY1, EFHD1, and NIPSNAP3B - were identified as potential diagnostic biomarkers using least absolute shrinkage and selection operator, random forest and support vector machine methods. These genes influence AS development through various biological pathways, with significant differences noted in pathways such as KRAS. 3 potential drugs, ISOSORBIDE DINITRATE, Benzaldehyde, and Hydroperoxycycloofosfamide - were identified as interacting with ALDH1B1, with interactions verified through molecular docking. This study demonstrates the relationship between AS and MD through bioinformatics and machine learning, identifying 4 key diagnostic genes and potential therapeutic drugs. These findings suggest avenues for new AS treatments. Future experimental validation further exploration AS-MD interactions and MD-based therapies are warranted.