Abstract

Cholangiocarcinoma (CCA) has a complex tumor microenvironment that critically influences tumor progression and therapeutic resistance. Glycosylation abnormalities have been linked to cancer growth and progression. This work was designed to develop a prognostic model based on glycosylation-related genes (GRGs) for predicting CCA outcomes and immunotherapy responses. Glycosylation patterns in macrophage subsets of CCA were analyzed via scRNA-seq. Key genes were identified by integrating module genes from WGCNA and DEGs. A risk model for CCA was established utilizing LASSO Cox regression. In vitro tests were conducted to validate the function of PGK1. The immune checkpoint blockade group exhibited elevated M1 signature scores and higher glycosylation levels. A risk model incorporating five genes (ANXA3, PGK1, PLAUR, CREB5, SPP1) for CCA was established. The high macrophage glycosylation-related risk score group had a considerable infiltration of M0 macrophages. In vitro experiments confirmed that PGK1 advanced glycation end products accumulation, drove M2 polarization of macrophages, and increased CCA cell proliferation and migration. This work proposes a glycosylation-based risk model for predicting CCA prognosis and directing therapeutic strategies. PGK1 is highlighted as a potential therapeutic target in CCA.