Single-cell profiling of pancreatic islets maps subtype-associated molecular alterations in type 2 diabetes.

AIMS/HYPOTHESIS: Type 2 diabetes exhibits remarkable clinical heterogeneity, represented by distinct pathological subtypes. However, the precise molecular alterations underlying these subgroups at single-cell resolution remain poorly understood. In this study, we integrated clinical metadata from both healthy control groups and individuals with type 2 diabetes with pancreatic islet single-cell transcriptomes to explore subtype-associated molecular alterations across mild age-related diabetes (MARD), mild obesity-related diabetes (MOD), severe insulin-deficient diabetes (SIDD) and severe insulin-resistant diabetes (SIRD).

METHODS

We applied the established type 2 diabetes clustering framework with independent validation against external centroids to evaluate subtype assignment accuracy across 43 pancreatic islet donors with type 2 diabetes. High-throughput single-cell RNA sequencing data comprising 131,083 cells from 32 pancreatic islet donors were subsequently analysed to systematically compare cellular composition and subtype-associated molecular perturbation patterns across these subgroups.

RESULTS

Four type 2 diabetes subgroups were identified, and validation against external centroids confirmed high assignment concordance, with subtype mapping sensitivities ranging from 85.0% to 100.0% in the full dataset and from 91.7% to 100.0% in the single-cell RNA sequencing sub-cohort. Notably, SIDD displayed the most pronounced beta cell perturbations, characterised by dual facets of beta cell dysfunction: the depletion of mature beta cell subclusters and the erosion of global cellular identity of beta cells. These alterations were accompanied by systematic activation of autophagy, apoptosis and endoplasmic reticulum stress pathways, while energy metabolism and insulin secretion were significantly compromised in SIDD individuals. In contrast, the SIRD phenotype manifested primarily as metabolic hyperactivation with largely preserved beta cell maturity. Additionally, MOD demonstrated moderate beta cell perturbations, characterised by enriched ribosomal function and protein synthesis activity. MARD displayed the mildest molecular disruption, consistent with its mild clinical phenotype.

CONCLUSIONS/INTERPRETATION: Our study provides an exploratory single-cell map of subtype-associated transcriptomic patterns in pancreatic islets across clinically defined type 2 diabetes subgroups. These findings serve as a valuable resource for future investigations exploring the cellular and molecular heterogeneity underlying type 2 diabetes pathogenesis.