Abstract Transcriptomic classification has been used to molecularly characterize glioblastoma (GBM) but has failed to predict survival and inform on pharmacologic vulnerability. Here, we developed a computational approach for the unbiased identification of the core biological pathways that optimally classify individual glioma cells and bulk tumors. Using single cell RNA-sequencing data from 36 high-grade gliomas, we uncovered four transcriptional states that exist along two evolutionary axes, a metabolic axis including mitochondrial and glycolytic/pluri-metabolic and a neurodevelopmental axis including proliferative/progenitor and neuronal states. The activation of the same set of biological pathways independently stratified primary GBM into four subtypes, among which the mitochondrial subgroup was associated with the most favorable clinical outcome. By integrating genomic, transcriptomic, DNA methylation, microRNA and proteomics analysis, we found that mitochondrial GBM was enriched with coherent gain-of-function of mitochondrial genes and loss-of-function alterations targeting glycolysis and alternative metabolic programs, suggesting that this subgroup may fail to produce compensatory metabolism. Mitochondrial GBM relied exclusively on oxidative phosphorylation for energy production whereas the glycolytic/pluri-metabolic subtype was sustained by concurrent activation of multiple metabolic fluxes including aerobic glycolysis, amino acid consumption and lipid synthesis and storage. Deletion of SLC45A1, a gene coding for a glucose-H+ symporter on chromosome 1p36.23, emerged as the truncal genetic alteration most significantly associated with mitochondrial GBM. Reintroduction of SLC45A1 in mitochondrial GBM cells harboring SLC45A1 gene deletion induced cytoplasmic acidification, loss of cell fitness and growth arrest. The strict dependency of mitochondrial GBM on mitochondrial respiration was associated with excessive generation of reactive oxygen species and unique sensitivity to inhibitors of oxidative phosphorylation. Collectively, this work presents a classification of GBM that informs clinical outcome and identifies patients who are more likely to benefit from therapies targeting metabolic vulnerabilities. Citation Format: Luciano Garofano, Simona Migliozzi, Young Taek Oh, Fulvio D'Angelo, Ryan D. Najac, Aram Ko, Brulinda Frangaj, Francesca Pia Caruso, Kai Yu, Jinzhou Yuan, Wenting Zhao, Anna Luisa Di Stefano, Franck Bielle, Tao Jiang, Peter Sims, Mario L. Suvà, Fuchou Tang, Xiao-Dong Su, Michele Ceccarelli, Marc Sanson, Anna Lasorella, Antonio Iavarone. Pathway-based classification of glioblastoma uncovers a mitochondrial subtype with therapeutic vulnerabilities [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2021; 2021 Apr 10-15 and May 17-21. Philadelphia (PA): AACR; Cancer Res 2021;81(13_Suppl):Abstract nr 97.
Read full abstract