单细胞分辨率下人远端胆管癌导管细胞的共有与分化转录特征

Background: While chronic pancreatitis (CP) has long been recognized as a risk factor for pancreatic ductal adenocarcinoma, recent work has shown an elevated risk of biliary tract malignancies in patients with pancreatitis. However, the underlying mechanisms and the genomic similarities that might explicate this association remain elusive. Inflammation is thought to be a precipitating risk factor for cholangiocarcinoma; however, whether the inflammatory milieu of chronic pancreatitis also contributes to this risk is unclear. To address this gap, we characterized the shared and unique genomic signatures between patients with chronic pancreatitis and those with distal cholangiocarcinoma (dCCA), compared to normal healthy individuals. Methods: We performed single cell RNA sequencing (scRNA-seq) from pancreata or tumor specimens obtained from 3 healthy human subjects, 3 patients with CP, and 2 individuals with dCCA. After ascribing cell-type identities and delineation of ductal cell clusters, we performed inference of copy number variations as well as functional perturbations at the gene and pathway level in ductal cell from the three patient cohorts - normal, CP and dCCA. Results: scRNA-seq revealed 35 different cell clusters across ~120,000 cells from the three patient cohorts. Based on semi-supervised learning, we identified 13 cell types in our data; additional analysis revealed that the Ductal cell cluster could be further sub-classified into two cell populations with disparate genomic signature. We further identified amplification/duplication of chromosomes 1, 7, 11, 12 and 18 in the Ductal A cluster from the dCCA cohort through two orthogonal approaches. Pseudobulking analysis suggested an upregulation of laminins in Ductal A subcluster, and Ligand-receptor analysis further revealed LAMC2 as a key target gene in the Ductal cell cluster, with putative fibroinflammation mediated by TGFB1 and FGF2 receptors. While laminin upregulation was unique to dCCA and not observed in CP, TGFB1 and FGF2 receptor-driven signaling was a shared feature in ductal cells across both CP and dCCA. Conclusions: We discovered a novel ductal cell population in dCCA that exhibits upregulation of laminins compared to normal pancreas and CP. Further still, the expression of ligand, receptor, and target genes in the dCCA extracellular matrix highlights how the TGFB1-FGF2-LAMC2 axis might be contributing to the desmoplastic process observed in dCCA tumors.