Abstract
INTRODUCTION: Atherosclerosis (AS) and atrial fibrillation (AF) are clinically intertwined cardiovascular disorders, yet their shared pathophysiological mechanisms remain poorly understood. Mitochondrial dysfunction and immune dysregulation have been implicated separately in both diseases, but their potential crosstalk in driving comorbidity is unexplored. This study aims to uncover common molecular pathways linking AS and AF, focusing on mitochondrial metabolism and immune response, and to identify diagnostic biomarkers using integrative bioinformatics approaches. METHODS: Transcriptomic datasets of AS and AF were analyzed through GO and KEGG enrichment to identify shared biological themes. WWGCNA prioritized 35 mitochondrial genes associated with both diseases. Three machine learning algorithms (LASSO, SVM, and random forest) were applied to screen hub genes, followed by validation in independent datasets. Immune infiltration and single-cell transcriptomic analyses were conducted to assess immune-microenvironment interactions and gene expression at cellular resolution. RESULTS: GO/KEGG analyses revealed dominant enrichment in mitochondrial oxidative phosphorylation and immune-inflammatory pathways for AS and AF. WGCNA identified 35 mitochondrial hub genes, with MRPS23 and CASP8 emerging as key candidates via machine learning. MRPS23 showed significant downregulation in AS tissues and cellular heterogeneity in scRNA-seq (single-cell RNA sequencing), while both genes exhibited strong correlations with immune cell subsets in both diseases. CONCLUSION: This study establishes MRPS23 as a novel biomarker bridging mitochondrial dysfunction and immune dysregulation in AS and AF comorbidity. Its decline in AS suggests a potential role in mitochondrial ribosomal integrity loss driving metabolic stress, while immune-microenvironment interactions highlight its pleiotropic impacts on inflammatory cascades. These findings advance the "mitochondria-immune axis" paradigm for cardiovascular comorbidity, offering MRPS23 as a dual-disease therapeutic target. Further validation in preclinical models and clinical cohorts is needed to translate these insights into diagnostic or therapeutic applications.