Identification of senescence-related biomarker for aortic dissection based on bioinformatics and machine learning algorithms.

Aortic dissection (AD) is a vascular surgical disease that seriously threatens human health. Due to a high misdiagnosis rate and unclear pathogenesis, it brings greater challenges to AD patients and vascular surgeons. This study aimed to explore sensitive diagnostic markers and potential therapeutic targets of AD from the perspective of cellular senescence, which has been our long-term concern. We downloaded the expression matrix of AD and control samples from the Gene Expression Omnibus database and obtained the senescence-related gene set. Differentially expressed genes (DEGs) in AD and control groups were analyzed by the "limma" package. Gene Ontology, Disease Ontology, and Kyoto Encyclopedia of Genes and Genomes enrichment analyses were carried out to demonstrate the enrichment function of DEGs. Two sensitive screening diagnostic marker algorithms, the least absolute shrinkage and selection operator and support vector machine-recursive feature elimination, were used to identify target genes in combination with senescence-related genes. Receiver operating characteristic curves for targeted genes were plotted to assess diagnostic efficacy. Gene set enrichment analysis was performed to preliminarily explore the pathways enriched in the 2 groups. CIBERSORT was used to look for differentially infiltrating immune cells, and Spearman analysis was carried out to explore the association between targeted genes and infiltrating immune cells. A total of 111 DEGs were identified, which were closely related to regulation of leukocyte migration, cellular transition metal ion homeostasis, transition metal ion homeostasis, and detoxification of copper ion from Gene Ontology and Kyoto Encyclopedia of Genes and Genomes analysis. Disease Ontology analysis found that these DEGs were mainly involved in acute myocardial infarction, vasculitis, and coronary artery disease. Hexokinase-3 (HK3) was identified as an overlapping gene of least absolute shrinkage and selection operator, support vector machine-recursive feature elimination, and senescence. HK3 showed good diagnostic efficacy (area under the receiver operating characteristic curve = 0.924 in the integrated cohort and 0.914 in the validation cohort). Correlation analysis results showed that HK3 was positively correlated with monocytes, neutrophils, and natural killer cells resting, while B cells naive, mast cells resting, and macrophages M1 were negatively correlated. HK3 was a potential biomarker for AD diagnosis and a target for precision therapy, and HK3 is also closely related to immune infiltration.