Integration of Bioinformatics and Machine Learning Identifies ENO2 and MPP1 as Shared Molecular Nodes Linking Alzheimer’s Disease and Type 2 Diabetes Mellitus

Background: Alzheimer’s disease (AD) is frequently accompanied by type 2 diabetes mellitus (T2DM), and the two disorders share some common pathological mechanisms. However, the core genes responsible for this comorbidity and the therapeutic targets remain to be identified.Methods: Gene expression data were obtained from the Gene Expression Omnibus (GEO) database for AD (GSE5281) and T2DM (GSE76894), and differentially expressed genes (DEGs) were identified. Weighted gene co-expression network analysis (WGCNA) was performed to identify AD-associated modules. Four machine learning algorithms—Least Absolute Shrinkage and Selection Operator (LASSO) regression, Random Forest (RF), Support Vector Machine Recursive Feature Elimination (SVM-RFE), and Boruta—were used to screen core genes. Gene set enrichment analysis (GSEA), immune infiltration analysis using the CIBERSORT method, and drug prediction using the Comparative Toxicogenomics Database (CTD) were then conducted. In order to evaluate the binding mode and affinity of the predicted drugs with core targets, molecular docking was performed.Results: The number of DEGs identified in the AD dataset was 4970, and the number of DEGs identified in the T2DM dataset was 392. Cross-analysis of DEGs and WGCNA identified 15 overlapping disease-related genes. Two core genes, enolase 2 (ENO2) and membrane palmitoylated protein 1 (MPP1), were identified using four machine learning algorithms. Both genes were co-enriched in GSEA for leukocyte transendothelial migration, long-term potentiation/depression, citric acid cycle, oxidative phosphorylation, and glycolysis/gluconeogenesis. Immune infiltration analysis revealed that they participated in the disease process by regulating the immune microenvironment. Valproic acid (VPA) was selected as a candidate drug based on the results of the CTD drug prediction. The molecular docking results revealed that the binding free energies of VPA with ENO2 and MPP1 were –4.747 and –5.346, respectively, suggesting that VPA had a higher binding affinity for MPP1.Conclusion: ENO2 and MPP1 are shared molecular markers between AD and T2DM, and may be involved in the pathobiology of both diseases by regulating energy metabolism, synaptic function, and immune inflammation. VPA may exert potential therapeutic effects by targeting these two shared nodes, providing preliminary molecular evidence for the mechanisms underlying AD-T2DM comorbidity. Further validation in AD-T2DM comorbid cohorts is warranted.