Septin 9 expression regulates ‘don't eat me’ signals and identifies an immune–epithelial class of intrahepatic cholangiocarcinoma

Abstract

Intrahepatic cholangiocarcinoma (iCCA) is a highly heterogeneous and aggressive liver cancer with limited therapeutic options. Precise classification and immunotherapy are perspectives to improve the treatments. We reported the role of septin 9 in apico‐basal polarity and epithelial‐to‐mesenchymal transition (EMT). Here, we aim to elucidate its role in iCCA. We analyzed single‐cell transcriptomes from human iCCA tumor cells based on phenotype and cell state. Knockdown of the septin 9 gene (SEPT9) was done using small interfering RNA (siRNA); interferon‐γ (IFN‐γ) stimulation was performed using different CCA cells; gene expressions were analyzed by reverse transcription and real‐time PCR analysis (RT‐qPCR); and immunofluorescence, immunoblotting, and flow cytometry were performed to assess the expression of proteins. The differential distributions of SEPT9 and vimentin (VIM) gene expressions allowed us to define specific cellular trajectories of malignant cells and thus identified distinct clusters of iCCA cells. One cluster was enriched in VIM and extracellular‐matrix (ECM) remodeling molecules, and another had high expression of SEPT9 and genes from the ‘don't eat me’ signal involved in immune escape. This antagonism between SEPT9 and VIM was confirmed by in vitro experiments. Notably, SEPT9 and ‘don't eat me’ gene expressions were inversely correlated to those of vimentin and the EMT markers. SEPT9 expression was upregulated by IFN‐γ and SEPT9 knockdown decreased expression of ‘don't eat me’ signal genes and increased expression of mesenchymal markers. Cancer Cell Line Encyclopedia (CCLE) transcriptome database analyses confirmed that iCCA cells enriched in septin 9 exhibit epithelial‐like features. This study revealed septin 9 as a cytoskeleton element of iCCA epithelial‐like cells and a regulator of the immune system response. It also brings new insights into the enigmatic relationship between EMT and immune response. Notably, we decoded a potential mechanism that could sensitize patients to immunotherapies.