Multi-omics interrogation of metabolic stress in cancer

Abstract

Cancer is associated with alterations to many pathways and cellular processes, including activation of pro-growth cell signaling pathways, resistance to cell death, activation of metastasis, and evasion of the immune system. Another key change in cancer is rewiring of cellular and mitochondrial metabolism to support cellular growth via energy generation, intake of molecular building blocks, and increased anabolic synthesis. While general metabolic changes in cancer are well characterized, the effect of cancer type-specific changes on metabolism are poorly understood. The goal of this dissertation was to identify metabolic vulnerabilities in specific cancer subtypes using data-driven approaches. In this dissertation, I present three studies that each utilize multi-omics approaches to characterize metabolic changes present in cancer. In the first project, we performed combined transcriptional and metabolic profiling of the mitochondrial stress response. Using integrated data analysis, we then generated mitochondrial stress signatures that describe discrete nodes of the mitochondrial stress response. Finally, we generated a workflow for quantifying these mitochondrial stress signatures in outside datasets, and utilized this to probe mitochondrial dysfunction in IDH1-mutant glioma. In the second project, we investigated the changes in lung tumorigenesis associated with aging using combined transcriptional and metabolic profiling of human patient samples. While we identified many shared genetic and metabolic changes in lung tumors independent of age, we identified age-dependent changes in metabolism that differ between tumors and healthy tissue. Furthermore, we identified the glutathione synthesis gene GCLC as differentially-regulated in young tumors compared to old tumors. Metabolic analysis highlighted changes to redox metabolism in tumors and with age, and identified changes in glutathione production associated with increased GCLC expression in young tumors. Using integrated multi-omics analysis, we then validated that GCLC expression is closely linked to glutathione levels in young patients, but not in old patients. In the final study, we investigated glutamate metabolism in triple negative breast cancer (TNBC). Given TNBC shows reduced expression of the glutamate enzyme GLUD1, we utilized bioinformatics approaches, combining human patient expression data and human-derived cell line metabolic data, to identify trends in TNBC with regard to GLUD1 expression and glutamate levels. We identified that the neurotransmitter GABA, alongside several of the genes encoding its receptors, were elevated in TNBC and specifically in concert with reduced GLUD1 expression. Together, these studies present methodologies for integrating transcriptional and metabolic changes in cancer to identify metabolic pathways that could be potential targets for cancer treatments.