Most solid tumours are hypoxic. Tumour cell proliferation and metabolism accelerate oxygen consumption. The low oxygen supply due to vascular abnormalisation and the high oxygen demand of tumour cells give rise to an imbalance, resulting in tumour hypoxia. Hypoxia alters cellular behaviour and is associated with extracellular matrix remodelling, enhanced tumour migration, and metastatic behaviour. In light of the foregoing, more research on the progressive and prognostic impacts of hypoxia on gliomas are crucial. In this study, we analysed the expression levels of 75 hypoxia-related genes in gliomas and found that a total of 26 genes were differentially expressed in The Cancer Genome Atlas (TCGA) database samples. We also constructed protein-protein interaction networks using the STRING database and Cytoscape. We obtained a total of 10 Hub genes using the MCC algorithm screening in the cytoHubba plugin. A prognostic risk model with seven gene signatures (PSMB6, PSMD9, UBB, PSMD12, PSMB10, PSMA5, and PSMD14) was constructed based on the 10 Hub genes using LASSO-Cox regression analysis. The model was verified to be highly accurate using subject work characteristic curves. The seven-gene signatures were then analysed by univariate and multivariate Cox. Notably, PSMB10, PSMD12, UBB, PSMA5, and PSMB6 were found to be independent prognostic predictive markers for glioma. In addition, PSMB6, PSMA5, UBB, and PSMD12 were lowly expressed, while PSMB10 was highly expressed, in the TCGA and GTEx integrated glioma samples and normal samples, which were verified through protein expression levels in the Human Protein Atlas database. This study found the prognostic predictive values of the hypoxia-related genes PSMB10, PSMD12, UBB, PSMA5, and PSMB6 for glioma and provided ideas and entry points for the progress of hypoxia-related glioma.
Keywords: Cox regression; glioma; hypoxia; risk prediction model; ubiquitin–proteasome system.
Copyright © 2021 Gao, Wang, Gu, Zhang and Wang.