Back

Single cell and spatial characterization of the human pancreas reveals drivers of beta cell dysfunction in cystic fibrosis

Mummey, H.; Corban, S.; Lucero, J.; Melton, R.; Pittigher, M.; Johnston, V.; Deutsch, G.; D'Antonio-Chronowska, A.; Wang, A.; Hull-Meichle, R.; Gaulton, K. J.

2025-12-16 genomics
10.64898/2025.12.14.694120 bioRxiv
Show abstract

Cystic fibrosis (CF) causes severely damaged pancreas morphology and results in high rates of cystic fibrosis-related diabetes (CFRD), but the cellular pathways and regulatory programs driving CFRD pathogenesis in the pancreas are not understood. In this study, we performed single cell multiomic and spatial profiling of 2.04M cells in the pancreas from 23 non-disease and CF donors and defined regulatory programs and tissue niches of specific cell types and sub-types in CF. A high-mucin sub-type of ductal cells had relatively preserved abundance, altered localization and up-regulated stress, secretion and pro-fibrotic activity in CF, and conventional ductal cells showed evidence for transition to the high-mucin state. Increased inflammatory and fibrotic pathway activity in CF was linked to closer proximity and crosstalk between immune and stellate cells in specific niches. Beta cells had extensive genomic changes in CF, including increased stress and insulin secretion-related processes, and these changes were broadly distinct from those in type 1 and type 2 diabetes. Islets in CF preferentially localized near large adipose tissue and collagen structures, and both areas were strongly linked to beta cell loss in CF due to signaling from adipocytes, macrophages, stellate, and high-mucin ductal cells. Overall, our results reveal cellular drivers of pancreatic dysfunction in CF and offer new in-roads to preserving beta cells in CFRD.

Matching journals

The top 4 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.