Establishment and validation of five autophagy-related signatures for predicting survival and immune microenvironment in glioma

Background Gliomas, especially Glioblastoma multiforme, are the most frequent type of primary tumors in central nervous system. Increasing researches have revealed the relationship between autophagy and tumor, while the molecular mechanism of autophagy in glioma is still rarely reported. Objective Our research aims to conform the autophagy-related genes (ARGs) implicated in the development and progression of glioma and improve our understanding of autophagy in glioma. Methods 20 candidate ARGs were screened through the protein-protein interaction network. We also downloaded the publicly accessible glioma data for 665 individuals from TCGA and 970 individuals from CGGA with RNA sequences and clinicopathological information. Subsequently, univariate and multivariate Cox regression analysis identified 5 key ARGs among the 20 candidate genes as key prognostic genes for survival, GSEA and immune response analysis. Results ATG5, BCL2L1, CASP3, CASP8, GAPDH were identified as key ARGs in our research. Further studies showed that the high-risk population was linked to a dismal prognosis and suggested an immune-inhibitory microenvironment. GSEA results demonstrated that high risk population was closely related to DNA repair, hypoxia pathways, implicated in immunosuppression and carcinogenesis. Through CMap, we finally identified 14 candidate drugs for the ARG high risk population. Conclusions This study established and verified an ARG risk model, which can serve as an independent predictor for prognosis, reflect on the strength of the immune response and predict the potential drugs in glioma. Our findings offer new understandings of ARG molecular mechanism and promising therapeutic targets for glioma treatment.