Integrative bioinformatics and machine learning approach unveils potential biomarkers linking coronary atherosclerosis and glutamine metabolism-associated gene.

BACKGROUND: Atherosclerosis (AS), a chronic inflammatory disorder of the vasculature, remains the principal driver of cardiovascular disease, accounting for substantial morbidity, mortality, and healthcare burden worldwide. Beyond its vascular implications, recent research highlights the metabolic reprogramming of glutamine (Gln) as a central axis in disease biology. Glutamine metabolism, long recognized for its role in tumorigenesis, is now emerging as a critical determinant of clinical outcomes across diverse cancers, underscoring its broader relevance to pathological processes. METHODS: A bioinformatics analysis was performed in this work to discover and validate putative Gln-Metabolism genes (GlnMgs) linked with AS. GlnMgs were discovered using a combination of differential expression analysis with a collection of 43 potential GlnMgs. Gene Set Enrichment Analysis (GSEA) and Gene Set Variation Analysis (GSVA) were used to determine the possible biological roles and pathways of the discovered GlnMgs. Following that, Lasso regression and SVM-RFE techniques were used to identify hub genes and assess the diagnostic efficiency of the nine GlnMgs in differentiating AS. The relationship between hub GlnMgs and clinical features was also studied. Finally, the GSE43292 and GSE9820 datasets were used to validate the expression levels of the nine GlnMgs. RESULTS: Nine GlnMgs associated with AS were identified, namely NOXRED1, SIRT4, DDAH2, GOT1, MIR21, NOS3, CAD, ASRGL1, and GMPS. Functional enrichment analysis revealed their predominant involvement in key metabolic pathways, including the cellular amino acid metabolic process, glutamine family amino acid metabolic process, and α-amino acid metabolic process, underscoring their role in metabolic reprogramming during AS progression. Importantly, the diagnostic model constructed from these nine GlnMgs exhibited robust discriminatory power, achieving an AUC value of 0.980, thereby highlighting its potential as a highly reliable biomarker signature for AS. In parallel, immune infiltration analysis provided further mechanistic insight, revealing that M0 macrophages and memory B cells were significantly associated with the identified gene signature. These findings not only strengthen the diagnostic utility of the GlnMgs-based model but also suggest a pivotal link between glutamine metabolism and immune cell dynamics in shaping the atherosclerotic microenvironment. CONCLUSIONS: This study successfully discovered nine GlnMgs that are associated with AS. These findings provide valuable insights into potential novel biomarkers for AS and offer prospects for monitoring disease progression.