Comparison of Mitochondrial Phenotypes Across Cancers with Different Tissues of Origin

Abstract

Cancer metabolism is typically characterized by aerobic glycolysis, an increase in glucose uptake and lactate production despite adequate sufficient oxygen availability to support mitochondrial oxidative phosphorylation (OXPHOS). The mitochondrial contribution to cancer has remained an important open question, with some suggesting that mitochondrial damage must be present to enable aerobic glycolysis and others suggesting mitochondrial OXPHOS is vital to cancer progression. Recent evidence of the oncogenic consequences to metabolic reprogramming caused by mutations to mitochondria-related genes has increased interest in understanding the interplay between metabolism and cancer, and how this may be targeted to improve therapy.

The aim of this dissertation was to determine whether there was a consistent mitochondrial phenotype across several models of two different cancers: acute myeloid leukemia (AML) and hepatocellular carcinoma (HCC). Both AML and HCC have extremely poor prognoses (5-year survival 20-25%), heterogeneous patient populations, and limited curative treatment options. Mitochondrial phenotyping was accomplished through application of a comprehensive, diagnostic biochemical workflow that integrated several assessments of mitochondrial function under physiologically relevant stimuli with mitochondrial-targeted mass spectrometry-based proteomics. The application of this platform was first used to validate a proteomics-based normalization factor that accounts for mitochondrial content across different preparations to improve comparisons of mitochondrial function between tissues. AML and HCC mitochondria were found to have a common phenotype in which maximal respiratory capacity was similar to their respective noncancer tissues, but respiration was inhibited in the presence of physiological ATP demand states. Different mechanisms contributed to this outcome in either cancer, suggesting that this may be a conserved feature of metabolic reprogramming in cancer. The detection of this phenotype was entirely dependent upon the use of mitochondrial interrogation methods that simulate in vivo energy states, supporting the tremendous importance of the integration of bioenergetics with physiology.