Organelle proteomics of human, neonatal pig, healthy and diabetic mouse islets

Pancreatic islets, in which β cells constitute 50%-80% of the endocrine mass, are the body's site for glucose-regulated insulin secretion. Pancreatic islet dysfunction is a hallmark of type 2 diabetes. Nevertheless, the molecular changes that impair their function remain insufficiently understood. To determine how islet cellular organization supports function and how it deteriorates in disease, we used machine learning-based organelle proteomics to build an organelle atlas of pancreatic islets from mice, humans, and pig neonatal islet-like cells. This cross-species resource maps localization of ∼8,000 proteins, revealing conserved organelle organization and identifying previously unrecognized components of insulin granules. Applying our organelle proteomics workflow to db/db mice, a T2D model uncovered disease-associated changes affecting vesicular trafficking and mitochondrial amino acid metabolism. Additionally, we found a disassembly of the insulin-granule acidification machinery. Our atlas provides insights into islet subcellular organization and identifies processes whose disruption may drive pancreatic islet failure in T2D.