Washington University School of Medicine in St. Louis St. Louis, MO
V. Sviderskiy1, V. Vasudevaraja2, L. G. Dubois3, J. Stafford4, E. K. Liu2, J. Serrano5, R. Possemato2, and M. Snuderl6; 1Washington University in St. Louis, Saint Louis, MO, 2NYU Grossman School of Medicine, New York, NY, 3Federal University of Rio de Janeiro, Rio de Janeiro, Brazil, 4University of Vermont, Burlington, VT, 5New York University School of Medicine, New York, NY, 6NYU Grossman School of Medicine, New York City, NY
Purpose/Objective(s): Malignant gliomas represent a disproportionate source of cancer-related morbidity and mortality in adults and children. High-grade malignant gliomas such as adult IDH wild-type glioblastoma (GBM) and pediatric histone H3 mutated high-grade gliomas (HGG – H3) have less than an 18-month median survival despite aggressive therapy. Prior molecular characterization of adult and pediatric tumors suggests that genetically and epigenetically these tumors differ considerably. However, the metabolic differences between adult and pediatric tumors has not been assessed and may represent an opportunity to better understand these subgroups and identify novel therapeutic targets. Materials/
Methods: A total of 114 adult and pediatric primary gliomas including 18 pediatric HGG-H3 samples were freshly collected and then frozen in -800C. LC/MS based metabolomics was performed on the lipid and polar layers after sample processing. Samples with adequate remaining tissue were also collected for transcriptomic analysis with normalized read counts used for analysis of the data. MetaboAnalyst (version 3.2.0) R package was used for metabolic data preprocessing, batch correction, and differential metabolomic analysis of polar and lipid metabolites separately. Normalized batch corrected data was used to perform unsupervised principal component analysis (PCA) and generate heatmaps by examining variance and calculating Z-scores respectively. Results: Principal component analysis demonstrated clear separation of adult GBM and pediatric HGG-H3 tumors in both the polar and lipid metabolite data, suggesting key metabolic differences exist between these two subgroups. Of the polar metabolites, glucose was the most highly enriched in pediatric HGG-H3 mutant tumors and other pediatric tumors compared to adult GBM tumors, which instead had elevated levels of lactate. These data suggest differences in availability of glucose and/or utilization of glucose for downstream pathways between these subtypes. We find that SLC2A4, which encodes the insulin-stimulated glucose transporter GLUT4, is overexpressed in the HGG-H3 samples and may facilitate this difference. Our data support adult GBM tumors instead rely more on fatty acid oxidation, as they have an abundance of acyl carnitines and have significant enrichment of transcripts needed for oxidative phosphorylation. Conclusion: Our findings suggest intrinsic differences exist in the metabolism of adult glioblastomas and high-grade pediatric gliomas, despite their histological similarities, such as the presence of microvascular proliferation and necrosis. Our data has the potential to provide new insights into metabolic vulnerabilities of these subgroups for potential targeted therapies.
