Intratumoral heterogeneity contributes to cancer drug resistance, but the underlying mechanisms are not understood. Single-cell analyses of patient-derived models and clinical samples from glioblastoma patients treated with epidermal growth factor receptor (EGFR) tyrosine kinase inhibitors (TKIs) demonstrate that tumor cells reversibly up-regulate or suppress mutant EGFR expression, conferring distinct cellular phenotypes to reach an optimal equilibrium for growth. Resistance to EGFR TKIs is shown to occur by elimination of mutant EGFR from extrachromosomal DNA. After drug withdrawal, reemergence of clonal EGFR mutations on extrachromosomal DNA follows. These results indicate a highly specific, dynamic, and adaptive route by which cancers can evade therapies that target oncogenes maintained on extrachromosomal DNA.

Gliomas contain a small number of treatment-resistant glioma stem cells (GSCs), and it is thought that tumor regrowth originates from GSCs, thus rendering GSCs an attractive target for novel treatment approaches. Cancer cells rely more on glycolysis than on oxidative phosphorylation for glucose metabolism, a phenomenon used in 2-[ 18 F]fluoro-2-deoxy-D-glucose positron emission tomography imaging of solid cancers, and targeting metabolic pathways in cancer cells has become a topic of considerable interest. However, if GSCs are indeed important for tumor control, knowledge of the metabolic state of GSCs is needed. We hypothesized that the metabolism of GSCs differs from that of their progeny. Using a unique imaging system for GSCs, we assessed the oxygen consumption rate, extracellular acidification rate, intracellular ATP levels, glucose uptake, lactate production, PKM1 and PKM2 expression, radiation sensitivity, and cell cycle duration of GSCs and their progeny in a panel of glioma cell lines. We found GSCs and progenitor cells to be less glycolytic than differentiated glioma cells. GSCs consumed less glucose and produced less lactate while maintaining higher ATP levels than their differentiated progeny. Compared with differentiated cells, GSCs were radioresistant, and this correlated with a higher mitochondrial reserve capacity. Glioma cells expressed both isoforms of pyruvate kinase, and inhibition of either glycolysis or oxidative phosphorylation had minimal effect on energy production in GSCs and progenitor cells. We conclude that GSCs rely mainly on oxidative phosphorylation. However, if challenged, they can use additional metabolic pathways. Therefore, targeting glycolysis in glioma may spare GSCs.

SUMMARY Aerobic glycolysis (Warburg effect) is a core hallmark of cancer, but the molecular mechanisms underlying it remain unclear. Here, we identify an unexpected central role for mTORC2 in cancer metabolic reprogramming where it controls glycolytic metabolism by ultimately regulating the cellular level of c-Myc. We show that mTORC2 promotes inactivating phosphorylation of class IIa histone deacetylases that leads to the acetylation of FoxO1 and FoxO3, and this in turn releases c-Myc from a suppressive miR-34c-dependent network. These central features of activated mTORC2 signaling, acetylated FoxO and c-Myc levels are highly inter-correlated in clinical samples, and with shorter survival of GBM patients. These results identify a specific, Akt-independent, role for mTORC2 in regulating glycolytic metabolism in cancer.

Summary Small-molecule inhibitors targeting growth factor receptors have failed to show efficacy for brain cancers, potentially due to inability to achieve sufficient drug levels in the CNS. Targeting non-oncogene tumor co-dependencies provides an alternative approach, particularly if drugs with high brain penetration can be identified. Here we demonstrate that the highly lethal brain cancer glioblastoma (GBM) is remarkably dependent on cholesterol for survival, rendering these tumors sensitive to Liver X receptor (LXR) agonist-dependent cell death. We show that LXR-623, a clinically viable, highly brain-penetrant LXRα-partial/LXRβ-full agonist selectively kills GBM cells in an LXRβ- and cholesterol-dependent fashion, causing tumor regression and prolonged survival in mouse models. Thus, a metabolic co-dependency provides a pharmacological means to kill growth factor-activated cancers in the CNS.