Identification of an energy metabolism-related six-gene signature for distinguishing and forecasting the prognosis of low-grade gliomas.

Abstract

Low-grade gliomas (LGG) account for 20-25% of all gliomas. In this study, we assessed whether metabolic status was correlated with clinical outcomes in LGG patients using data from The Cancer Genome Atlas (TCGA). LGG patient data were collected from TCGA, and the Molecular Signature Database was used to extract gene sets related to energy metabolism. After performing a consensus-clustering algorithm, the LGG patients were divided into four clusters. We then compared the tumor prognosis, function, immune cell infiltration, checkpoint proteins, chemo-resistance, and cancer stem cells (CSC) between the two groups with the greatest prognostic difference. Using least absolute shrinkage and selection operator (LASSO) analysis, an energy metabolism-related signature was further developed. Energy metabolism-related signatures were applied to identify four clusters (C1, C2, C3, and C4) using a consensus-clustering algorithm. C1 LGG patients were more related to the synapse and had higher CSC scores, more chemo-resistance, and a better prognosis. C4 LGG was observed to have more immune-related pathways and better immunity. We then identified six energy metabolism-related genes (PYGL, HS3ST3B, NNMT, FMOD, CHST6, and B3GNT7) that can accurately predict LGG prognosis not only as a whole but also based on the independent predictions of each of these six genes. The energy metabolism-related subtypes of LGG were identified, which were strongly related to the immune microenvironment, immune checkpoint proteins, CSCs, chemo-resistance, prognosis, and LGG advancement. A signature of genes involved in energy metabolism could help to distinguish and predict the prognosis of LGG patients, and a promising method to discover patients that may benefit from LGG therapy.