Abstract
As one of important components of tumor microenvironment, CAFs (cancer-associated fibroblasts) play a vital role in the development and metastasis of bladder cancer. The present study aimed to develop a CAFs-related gene signature to predict the prognosis of patients and the response to chemotherapy and immunotherapy based on research of multidatabase. Expression data and clinical information were obtained from TCGA and GEO databases. Different bioinformatic and statistical methods were combined to construct the robust CAFs-related gene signature for prognosis. The model was explored from four aspects: single-cell source, immune infiltration, correlation with cancer-related genes and pathways, and prediction of drug response. After screening, five genes (BNC2, LAMA2, MFAP5, NID1, and OLFML1) related to CAFs were used for constructing the signature to divide patients into high- and low-risk groups. Patients in low-risk group had better prognosis and multidatabase analysis confirmed the predictive value. The five genes were mainly expressed by fibroblasts and involved in regulation of pathways related with glycolysis, hypoxia, and epithelial-mesenchymal transition (EMT). BNC2, LAMA2, and NID1 were strongly associated with drug sensitivity. Moreover, the immunological status was different between high- and low-risk groups. High-risk patients had poor response to chemotherapy or immunotherapy. The CAFs-related gene signature might help to optimize risk stratification and provide a new insight in individual treatment for bladder cancer.