Back

Spatial organization of myofibroblastic and complement-secreting CAFs in neuroendocrine tumors

Niedra, H.; Springe, M. L.; Tiltina, K.; Peculis, R.; Saksis, R.; Nazarovs, J.; Ozolins, A.; Vilisova, S.; Senterjakova, N.; Gerina, A.; Konrade, I.; Pukitis, A.; Rovite, V.

2026-03-01 molecular biology
10.64898/2026.02.26.708188 bioRxiv
Show abstract

Neuroendocrine tumors are graded and classified largely by tumor cell-intrinsic features, yet the stromal microenvironment remains poorly defined across anatomical sites. We applied near single-cell spatial transcriptomics augmented with cell segmentation to eight treatment-naive neuroendocrine tumor primary tissues from pancreas, colon, appendix, and bile duct to build a spatially resolved stromal reference. Integration of stromal-enriched cell polygons identified ten transcriptional states shared across tumors, with a minority of niche-restricted clusters mapping to tumor/stroma and stroma/non-tumor boundaries. Within the shared fibroblast compartment, program scoring resolved four cancer-associated fibroblast states. Myofibroblastic and complement-secretory states dominated across samples, whereas inflammatory and antigen-presenting programs were consistently detected but weaker. Spatial mapping in desmoplastic tumors showed myofibroblastic fibroblasts enriched in collagen-dense regions, while complement-secretory fibroblasts localized preferentially to tumor-adjacent stromal interfaces. Pseudobulk differential expression and gene set enrichment analyses supported extracellular matrix remodeling in myofibroblastic fibroblasts and complement cascade activation in complement-secretory fibroblasts. Together, these findings demonstrate that anatomically distinct neuroendocrine tumors share a conserved yet spatially segregated stromal architecture, characterized by dominant matrix-producing and complement-enriched fibroblast states.

Matching journals

The top 11 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.