Integrating bioinformatics and machine learning analyses to identify immune-related secretory proteins and therapeutic small-molecule drugs in calcific aortic valve disease with type 2 diabetes.

INTRODUCTION: Type 2 diabetes mellitus (T2DM) is a globally prevalent metabolic disease, and emerging studies have revealed its strong association with calcific aortic valve disease (CAVD). Chronic inflammation, oxidative stress, and immune dysregulation induced by hyperglycemia in T2DM may accelerate CAVD progression, although the molecular mechanisms remain unclear. METHODS: We integrated and analyzed four CAVD and two T2DM gene expression datasets from the GEO database. Through differential gene expression analysis, weighted gene co-expression network analysis (WGCNA), and secretory protein screening, we identified shared pathogenic genes between T2DM and CAVD. Protein-protein interaction (PPI) networks, functional enrichment analysis, and Connectivity Map (cMAP) prediction were conducted to identify potential therapeutic targets. A diagnostic model was constructed using 113 machine learning algorithms, and immune infiltration analysis was performed using CIBERSORT. The expression of key genes was validated in clinical valve tissue samples via RT-qPCR, Western blotting, and immunohistochemistry. RESULTS: A total of 13 intersecting genes were identified as potential secretory biomarkers. The diagnostic model built with four key genes (CDH19, COL1A2, PRG4, and SPP1) showed excellent predictive performance (average AUC = 0.95). Immune infiltration analysis revealed significant differences in macrophage and T cell subtypes between CAVD and controls. CDH19 was downregulated, while COL1A2, PRG4, and SPP1 were significantly upregulated in T2DM-associated CAVD tissues. Among the candidate compounds, phorbol-12-myristate-13-acetate (PMA) emerged as a top therapeutic molecule potentially capable of reversing pathological gene expression. CONCLUSION: Our study identifies key secretory proteins and immune signatures in T2DM-associated CAVD and proposes a novel diagnostic model with strong clinical applicability. These findings offer new insights for early diagnosis and personalized treatment strategies in CAVD patients with T2DM.