Abstract
Aortic dissection (AD) involves complex interactions among amino acid, glucose, and lipid metabolism, exacerbating aortic inflammation and extracellular matrix (ECM) degradation, coupled with smooth muscle cell (SMC) dysfunction (phenotypic alteration, aging, apoptosis). To explore AD pathogenesis, we integrated single-cell RNA sequencing (scRNA-seq), metabolomics, machine learning, and Mendelian randomization to investigate SMC changes and gene-metabolite interactions. ScRNA-seq data (GSE213740, GSE155468) were analyzed for cell clustering and pseudo-time trajectories via Seurat and Monocle2. Metabolomics (9 samples: 6 AD, 3 controls) and machine learning validated key genes/metabolites, with Mendelian randomization assessing causal links. Nine cell subsets and 2000 variable genes were identified, with SMCs central to AD via cholesterol metabolism. APOE and PLTP were key genes; metabolomics highlighted cholesterol esters (CEs) and triglycerides (TGs) as critical metabolites. Machine learning confirmed APOE/PLTP's high predictive accuracy (AUC: 0.796-0.989). Mendelian randomization linked elevated CEs and TGs to increased AD risk (IVW: P = .04 and P = .02, respectively). This study establishes a gene-metabolite network where APOE and PLTP regulate CEs/TGs, influencing SMC function and AD progression, offering potential therapeutic targets.