Identification of Excellent Prognosis IDH Wildtype Glioblastomas Using Genomic and Metabolic Profiling

Abstract

High grade gliomas (HGGs) are aggressive brain tumors with altered cellular metabolism. HGGs can carry mutations in the tricarboxylic acid (TCA) cycle enzyme isocitrate dehydrogenase 1 (IDH1), conferring distinct biology and improved patient prognosis compared to IDH wildtype (wt) tumors. Using metabolomic analyses of tumor tissue, we previously showed that IDH wt and IDH mutant (IDH mut) tumors have unique metabolomic signatures that correlate with different survival outcomes. Among this cohort of 69 HGG samples, we identified two unique patient tumors that metabolically clustered with IDH mut tumors, but lacked both the IDH mutation and its product 2-hydroxyglutarate. We aimed to discover unique mutations in these two tumors that may impart an IDH mutant-like phenotype in the absence of an IDH1 or IDH2 mutation. Whole exome sequencing (WES) was performed on frozen tumor samples from two patients diagnosed as glioblastoma (GBM), IDH wt via Agilent v5 + IncRNA platform. Alignment to the hg38 genome and variant calling were completed using an accelerated implementation of GATK's BWA and MuTect2 algorithms from Sentieon. Variants were filtered based on supporting reads and variant allele thresholds, with synonymous variants and common SNPs removed. High-confidence variants were further filtered by membership in the four KEGG pathways associated with IDH1 and IDH2. Identified variants were corroborated with metabolomics data from the two unique IDH wt tumors compared with classical GBM IDH wt, oligodendrogliomas IDH mut and astrocytomas IDH mut to identify putative drivers of an IDH mutant-like metabolomic phenotype in these unique IDH wt tumors. Despite the lack of an IDH mutation, one patient survived 45.6 months and the other patient remains alive at last follow up 64 months post diagnosis, much longer than the 16-18-month median survival typical of patients with GBM IDH wt. WES of outlier IDH wt tumor samples revealed 65 unique mutations in the queried KEGG pathways, of which 34 had a variant allele frequency > = 0.15. These variants were processed in Gprofiler, confirming expected enrichment of the carboxylic acid metabolic biologic process, a functional gene set consisting of TCA genes, among these variants (p = 0.002, 3.6-fold enrichment). Accordingly, metabolite levels of intermediates of the TCA cycle, including malate and isocitrate were decreased in the outlier tumor samples compared to classic GBMs IDH wt (p<0.001). Presence of genetic alterations in key variants of the carboxylic acid metabolic biologic process (including ME1, GYP4F3, PTGIS, PFKL, PSPH, AKR1A1, HK2, NOS1) correlated with improved overall survival among GBM patients in the TCGA (p = 0.04). Laboratory validation of these findings in preclinical GBM models is ongoing. Disruption of the TCA cycle independent of an IDH mutation is associated with favorable survival in GBM. Pharmacologic inhibition of these pathways may be a promising strategy to improve GBM outcomes.