Abstract
Aberrant expression of coding genes of the V-ATPase subunits has been reported in glioma patients that can activate oncogenic pathways and result in worse prognosis. However, the predictive effect of a single gene is not specific or sensitive enough. In this study, by using a series of bioinformatics analyses, we identified five coding genes (ATP6V1C2, ATP6V1G2, TCIRG1, ATP6AP1 and ATP6AP2) of the V-ATPase that were related to glioma patient prognosis. Based on the expression of these genes, glioma patients were sub-classified into different prognosis clusters, of which C1 cluster performed better prognosis; however, C2 cluster showed more malignant phenotypes with oncogenic and immune-related pathway activation. The single-cell RNA-seq data revealed that ATP6AP1, ATP6AP2, ATP6V1G2 and TCIRG1 might be cell-type potential markers. Copy number variation and DNA promoter methylation potentially regulate these five gene expressions. A risk score model consisted of these five genes effectively predicted glioma prognosis and was fully validated by six independent datasets. The risk scores also showed a positive correlation with immune checkpoint expression. Importantly, glioma patients with high-risk scores presented resistance to traditional treatment. We also revealed that more inhibitory immune cells infiltration and higher rates of "non-response" to immune checkpoint blockade (ICB) treatment in the high-risk score group. In conclusion, our study identified a five-gene signature from the V-ATPase that could sub-classify gliomas into different phenotypes and their abnormal expression was regulated by distinct mechanisms and accompanied with immune microenvironment alterations potentially act as a biomarker for ICB treatment.