Mitochondrial dysfunction is a key feature of type 2 diabetes and is closely linked to ageing, a major risk factor for the disease. This study investigated islet cell composition and mitochondrial oxidative phosphorylation protein expression in pancreatic tissue from older donors (≥62 years) with and without type 2 diabetes, matched for age, sex, and BMI. Fixed human pancreatic tissue sections were immunolabelled for insulin, glucagon, NDUFB8 (complex I), MTCO1 (complex IV), and VDAC1 (a mitochondrial mass marker) to quantify islet composition and mitochondrial protein levels. A machine learning-based single-cell segmentation pipeline enabled high-resolution profiling of individual cell populations within islets. In type 2 diabetes, islets exhibited an increased alpha:beta cell ratio, altered spatial organisation with fewer beta-beta and more alpha-alpha interactions, and a significantly higher proportion of bi-hormonal cells co-expressing insulin and glucagon. Within beta cells, we observed significant changes in mitochondrial protein expression, including reduced complex I and elevated complex IV levels. Unsupervised clustering of mitochondrial expression patterns identified three distinct beta cell expression clusters. Donors with type 2 diabetes showed a marked shift in the distribution of beta cells across clusters, with increased proportions of beta cells exhibiting low complex I and high complex IV expression. These results highlight significant alterations in islet architecture and mitochondrial protein expression associated with type 2 diabetes, providing new insights into the mechanisms underlying type 2 diabetes.
Keywords: bi-hormonal cells; islet architecture; machine learning; mitochondrial dysfunction; oxidative phosphorylation; single-cell analysis and type 2 diabetes.