LAUSR

Background Ferroptosis is an autophagy-dependent form of cell death, sometimes called “ferritinophagy”. Its related pathway has been proven to regulate the programmed death of glioma stem cells. Mining autophagy-dependent ferroptosis-related gene (AD-FRG) signature could facilitate the discovery of mechanisms and therapeutic targets showing drug resistance to chemotherapeutic drugs. Methods We exhaustively searched HADB, MSigDB and FerrDb datasets and obtained 25 genes confirmed to exist in autophagy and ferroptosis death pathways. Glioma gene expression and clinicopathological data were collected from TCGA and CGGA datasets. Results Lasso regression and Cox regression analysis were carried out to construct a nine AD-FRGs signature (SIRT1, MTDH, HSPB1, CISD2, HMOX1, ATG7, MTOR, PRKAA2 and EIF2AK4). ROC curve showed that nine genes signature could effectively predict 1- (AUC = 0.869), 3- (AUC = 0.922) and 5-year (AUC = 0.870) survival rates. Immunohistochemical images confirmed the protein expression level of the gene model. The prognostic nomogram of risk score, age, WHO grade, isocitrate dehydrogenase (IDH) wild-type condition, 1p/19q co-deletion state was built. The calibration curve demonstrated that the prediction of the nomogram is highly consistent with the actual results. Moreover, tumor microenvironment analysis showed that the high-risk group was associated with high immune infiltration status and high tumor purity. Correlation analysis showed that the expression of SIRT1, CISD2 and HSPB1 might be related to macrophage infiltration and immunotolerance in glioma tissues. Conclusion Based on autophagy-dependent ferroptosis-related genes, we established gene signature and nomogram that maybe effectively predict the overall survival rate of glioma and correlate with the immunosuppressive tumor microenvironment (TME).