Js04.7.a analysis of rna classifies newly diagnosed glioblastoma patients and identifies patients vulnerable to targeted metabolic therapies

Abstract BACKGROUND Recent computational analysis of glioblastoma (GBM) single-cell RNA-sequencing data and bulk tumors identified four stable cellular states that embody metabolic (mitochondrial and glycolytic/plurimetabolic) and developmental (neuronal and proliferative/progenitor) attributes and generated a new GBM classification. Interestingly, the mitochondrial subtype depends on oxidative phosphorylation (OXPHOS), stratifies patients with specific metabolic pharmacologic vulnerabilities. Here, we set up and optimized RNA Sequencing of frozen and FFPE samples of GBM to apply the new proposed classification to a retrospective cohort of IDH wild type GBM patients. METHODS Tumor samples from IDH wild-type (wt) GBM patients diagnosed in Pitiè-Salpétrière Hospital from 2007 and 2019 and treated by first line standard radio-chemotherapy were analyzed by 3’ mRNA NGS. RESULTS Two hundreds thirty-two newly diagnosed (232) GBM IDH patients were analyzed. Median age was 60 years old (yo) (range 31-87 yo), median KPS at diagnosis was 80, Male/female ratio 1.5, pMGMT resulted hypermethylated in 48 out of 115 tested samples (42%). RNA analysis classified patients into molecular subgroups as it follows: 60 OXPHOS+, 48 Neuronal, 55 Proliferative and 58 Glycolytic; 11 unclassifiable. A subgroup of 28 patients was analyzed by RNA Sequencing from both FFPE and frozen tumors with concordant results in 25 patients (89%). Since F3T3 gene fusions correlate with hyperexpression of mitochondrial markers, we separately analyzed a subgroup of 20 IDHwt F3T3+ GBM patients. We found that F3T3+ are most frequently OXPHOS+ (50%) but not exclusively. Survival analysis shows significant longer overall survival (OS) of OXPHOS+F3T3- patients (22.9 months) compared to OXPHOS-F3T3-(18.4 months), (p=0.045) [including Neuronal patients (OS 18.4 months), Glycolytic (18.2 m), Proliferative 18.6 m)]. PFS after first line radio-chemotherapy did not differ significantly (p=0.11) between subgroups. Among F3T3+ patients, OXPHOS+ were long survivors with a median OS of 40.1 months while Glycolytic, Neuronal and Proliferative F3T3+ patients showed a worse prognosis similar to F3T3- patients (median OS 17.3 months; p=0.01). CONCLUSIONS In conclusions the RNA expression analysis classifies GBM patients from frozen and FFPE samples in metabolic subgroups. We identified OXPHOS+ GBM patients with more indolent clinical behavior compared to Glycolytic and other subgroups and with specific vulnerabilities to metabolic manipulations, as in the case of OXPHOS+ patients that could benefit from targeted metabolic therapies. RNA sequencing of FFPE is feasible in the setting of a prospectively clinical trial, aiming to select OXPHOS+ patients among newly diagnosed GBM patients, eligible for mitochondrial inhibitors add-on to first-line standard radio-chemotherapy (Optimum trial, NCT04945148).