Endoplasmic reticulum stress and glucose metabolism gene signature associated with type 2 diabetes mellitus

Type 2 diabetes mellitus (T2DM) involves pancreatic β-cell dysfunction, systemic insulin resistance, and dysregulated cellular stress responses. Endoplasmic reticulum (ER) stress and glucose metabolism are closely interconnected in pancreatic islets, but their combined value for transcriptomic-based diagnostic modeling remains unclear. Public pancreatic islet transcriptomic datasets were analyzed. GSE76894, including 19 T2DM and 84 control samples, was used as the training dataset, and GSE38642, including 9 T2DM and 54 control samples, served as an independent validation dataset. Differentially expressed genes were intersected with curated ER stress and glucose metabolism–related gene sets. Functional enrichment and machine-learning analyses were performed to identify candidate diagnostic genes. A total of 2,781 differentially expressed genes were identified in the training dataset, including 1,322 upregulated and 1,459 downregulated genes. Intersection with curated pathway-related gene sets yielded 20 ER stress and glucose metabolism–related differentially expressed genes. Enrichment analysis highlighted metabolic regulation, kinase activity, carbohydrate metabolism, and FoxO, HIF-1, and MAPK signaling pathways. Logistic regression, least absolute shrinkage and selection operator regression, and support vector machine analysis identified a nine-gene signature consisting of EP300, IRS1, IL1B, IAPP, AKT1, IGF1, VEGFA, HNF1A, and HSPA5. The model showed high classification performance in the training dataset and moderate discriminatory performance in the validation dataset. Exploratory immune-infiltration analysis suggested an association between IL1B and immune-related signals in pancreatic islets. This study identified a pathway-informed nine-gene signature associated with T2DM status in pancreatic islet transcriptomic datasets. These findings support the potential value of integrating ER stress and glucose metabolism gene sets for molecular classification and provide a basis for future biomarker validation.