Multi-gene biomarkers reveal spatial organization and subpopulation-specific damage response in intrahepatic biliary epithelial cells
Kanakanui, K. G.; Hantelys, F.; Hrncir, H. R.; Bombin, S.; Gracz, A. D.
Show abstract
Background & AimsIntrahepatic biliary epithelial cell (BEC) heterogeneity remains challenging to define. Here, we sought to identify BEC subpopulations and biomarkers in mouse liver. MethodsWe performed scRNA-seq on Sox9EGFP+ liver epithelium from mice subjected to bile duct ligation (BDL) and sham controls. A machine learning algorithm, NSForest, identified minimal, multi-gene signatures for BEC subpopulations. These "metagenes" were validated using hybridization chain reaction (HCR) FISH in tissue sections from wild-type mice and on primary BECs expanded in vitro. Metagenes were used to match BDL subpopulations to their corresponding sham subpopulations for differential gene expression (DGE) analysis. ResultsWe identified 4 BEC subpopulations in sham controls, each associated with 1-2 gene metagenes. Spatial localization of metagene-defined BEC subpopulations by HCR FISH revealed heterogeneous cellular composition of intrahepatic bile ducts. BECs belonging to a given subpopulation were most likely to have neighbors of the same identity, forming homogenous cellular compartments within ducts. BDL downregulated subpopulation-specific genes and upregulated a damage-associated gene set. BDL samples also included a proliferative subpopulation not found in sham controls, which contained populations enriched for three of the four metagenes. All BEC subpopulations were also found in monolayers in vitro, where they clustered spatially with BECs of the same subtype. ConclusionsNovel metagene biomarkers of BEC subpopulations facilitated spatial localization of BECs in situ, identified subpopulation specific injury responses, and confirmed that BEC heterogeneity is preserved in vitro. The presence of locally homogenous BEC "neighborhoods" in vitro suggests some degree of BEC organization may be epithelial-autonomous.
Matching journals
The top 3 journals account for 50% of the predicted probability mass.