Background and aims: Intrahepatic cholangiocarcinoma (ICC) is the second most common primary liver cancer whose incidence is increasing globally. Majority of ICC patients are diagnosed at an advanced stage with limited treatment options. However, the high tumor heterogeneity of ICC restricts the efficacy of available systematic therapies. We aim to dissect the tumor heterogeneity of ICC using high-resolution single-cell sequencing technique combined with spatial transcriptomics to help develop novel therapeutic strategies.
Methods: We performed single-cell RNA sequencing in 26 tumor samples from 23 ICC patients and spatial transcriptomics with 8 tumor sections from 8 ICC patients to explore the tumor heterogeneity of ICC. Bulk RNA sequencing cohort from our center (n=157), and two public datasets (ICGC JAPAN and ICGC SINGAPORE) were used as validation.
Results: We discovered that malignant cells in ICC harvested wide heterogeneity with diverse transcriptomic profiles, among which we identified a basal-like tumor cell subpopulation characterized by the expression of basal epithelial related genes including KRT5, KRT6A, and KRT17. Compared with other tumor cells, the highly expressed genes in basal-like subpopulation enriched in MET signaling and extracellular matrix organization associated with tumor invasion. Spatial transcriptomics data also confirmed the existence of basal-like tumor cell subpopulation in ICC tumor. Moreover, we explored the association between basal-like tumor cell subpopulation and prognosis in ICC patients. The results showed that a worse overall survival was observed in patients with a high level of basal-like gene signature expression than those with a low level (p < 0.001). We then divided our ICC cohort into basal-high group (n=8) and basal-low group (n=18) according to the percentage of basal-like tumor cells. As expected, conspicuous cancer associated fibroblasts (CAFs) enrichment was observed in basal-high group tumors, especially matrix CAFs and inflammatory CAFs. Cell-cell communication analysis further exhibited significant HGF-MET interaction between inflammatory CAFs and basal-like tumor cells. Consistently, inflammatory CAFs were the major source of HGF in tumor microenvironment and increased MET expression was found in basal-like tumor cells, indicating the HGF-MET axis may participate in basal-like phenotype formation of tumor cells.
Conclusion: We identified an aggressive basal-like tumor cell subpopulation correlated with poor prognosis in ICC. MET pathway may contribute to the invasive phenotype transition of basal-like tumor cells and serve as a new therapeutic target for ICC.

Changyi Liao

Appearances