TGF-β based risk model to predict the prognosis and immune features in glioblastoma

BACKGROUND: Transforming growth factor-β (TGF-β) is a multifunctional cytokine with an important role in tissue development and tumorigenesis. TGF-β can inhibit the function of many immune cells, prevent T cells from penetrating into the tumor center, so that the tumor cells escape from immune surveillance and lead to low sensitivity to immunotherapy. However, its potential roles in predicting clinical prognosis and tumor microenvironment (TME) immune features need to be deeply investigated in glioblastoma (GBM). METHODS: The TCGA-GBM dataset was obtained from the Cancer Genome Atlas, and the validation dataset was downloaded from Gene Expression Omnibus. Firstly, differentially expressed TGF-β genes (DEGs) were screened between GBM and normal samples. Then, univariate and multivariate Cox analyses were used to identify prognostic genes and develop the TGF-β risk model. Subsequently, the roles of TGF-β risk score in predicting clinical prognosis and immune characteristics were investigated. RESULTS: The TGF-β risk score signature with an independent prognostic value was successfully developed. The TGF-β risk score was positively correlated with the infiltration levels of tumor-infiltrating immune cells, and the activities of anticancer immunity steps. In addition, the TGF-β risk score was positively related to the expression of immune checkpoints. Besides, the high score indicated higher sensitivity to immune checkpoint inhibitors. CONCLUSIONS: We first developed and validated a TGF-β risk signature that could predict the clinical prognosis and TME immune features for GBM. In addition, the TGF-β signature could guide a more personalized therapeutic approach for GBM.