Integrative bioinformatics and machine learning reveal hub genes and immune signatures bridging type 2 diabetes mellitus, fracture susceptibility, and osteoblast differentiation dysfunction

To explore the causal link between T2DM and fracture, identify hub genes connecting T2DM, osteoblast differentiation, and fracture, and elucidate the molecular mechanisms behind T2DM-impaired fracture healing. We employed two-sample MR based on GWAS data to evaluate the causal effect of T2DM on fracture. By integrating gene expression data related to T2DM, fractures, and osteoblast differentiation, we systematically screened for hub genes using WGCNA, DEGs analysis, PPI network construction, and multiple machine learning algorithms. Subsequently, we performed functional annotation and validation of the identified hub genes through GSEA, cross-dataset validation, and immunocytes infiltration analysis. MR confirmed a positive causal effect of T2DM on fracture, with no horizontal pleiotropy detected. A total of 280 overlapping genes among T2DM, fracture, and osteoblast differentiation were identified. These genes were primarily involved in biological processes such as extracellular matrix organization, ossification, bone remodeling, and osteogenic differentiation, and were enriched in the lysosome, cell cycle, integrin, and AGE-RAGE pathways. Through PPI network and machine learning analyses, PLK1, CCNA2, and CHEK1 were identified as hub genes. GSEA revealed that hub genes were enriched in cell cycle regulation pathway. Cross-dataset validation further confirmed their significant upregulation in T2DM patients. Immune infiltration analysis indicated that neutrophils dominate the T2DM-associated immune microenvironment. T2DM has a positive causal association with fracture risk via multiple mechanisms, driven by abnormal cell cycle regulation and AGE-RAGE pathways. PLK1, CCNA2, and CHEK1 are hub genes connecting T2DM with osteoblast dysfunction and impaired fracture healing, positioning them as potential therapeutic targets.