Pyruvate Dehydrogenase Inhibition Leads to Decreased Glycolysis and Increased Reliance on Gluconeogenesis in Acute Myeloid Leukemia

Background:

Acute myeloid leukemia (AML) is an aggressive disease characterized by poor outcomes and chemotherapy resistance. Devimistat (CPI-613®) is a novel agent that inhibits two key tricarboxylic acid (TCA) cycle enzymes, pyruvate dehydrogenase (PDH) and α-ketoglutarate dehydrogenase complexes (KGDH). Our lab has shown that genetic and pharmacologic inhibition of PDH sensitizes cells to chemotherapy and recent clinical trials of chemotherapy plus devimistat have been promising. However, the means by which AML cells adapt to PDH inhibition during devimistat treatment is unknown.

Methods:

Murine AML cells were used to knockout PDH through the CRISPR-CAS9 system. Viability assays were done using EZQuant or Cell-Titer Glo assays. Glycolytic flux was assessed by measurement of the extracellular acidification rate (ECAR) using the XF24 Extracellular Flux Analyzer. Glucose uptake was measured by Glucose Uptake-Glo assay. Protein levels were assessed by Western blot. Gene expression was measured by qPCR. Amino acid dependencies were assessed by incubating cells in Hank's Balanced Salt Solution with devimistat and either asparagine or glutamate for 4 hours followed by return to complete media and viability assayed after 72-hours.

Results:

Previous studies have shown that glycolysis can compensate for mitochondrial inhibition so we determined glycolytic rates during PDH inhibition. We have shown previously that genetic PDH loss decreases ECAR levels consistent with decreased glycolytic flux. Devimisat treatment also decreased ECAR levels in K562 and OCI cells in a dose dependent manner. We next examined glucose uptake. PDH knockout cells showed decreased glucose uptake compared to control. This could be due to reduced glucose import or reduced glucose retention. The expression of glucose importer GLUT1 was significantly decreased in cells with genetic or pharmacologic inhibition of PDH. LDHA expression was also reduced in PDH KO cells and devimistat treated cells, contributing to the ECAR decrease. Hexokinase II (HKII) traps glucose in AML cells and docks at the mitochondrial membrane. It can be dislodged with mitochondrial stress and subsequently degraded via the proteasome. PDH knockout cells and devimistat treated cells demonstrated reduced levels of HKII that could be rescued with the proteasomal inhibitor bortezomib. Taken together this data shows that genetic and pharmacologic inhibition of PDH decreases glycolysis by decreasing glucose import and retention. To assess the necessity of the diminished glycolytic activity we treated PDH knockout and devimistat treated cells with the glycolytic inhibitor 2-deoxy-D-glucose. Both devimistat treated and PDH KO cells were exquisitely sensitive to glycolysis inhibition. We next assessed the importance of gluconeogenic amino acids. We specifically looked at glutamate and asparagine as they have been shown to be highly utilized in cancers. Asparagine but not glutamate rescued AML cells from devimistat treatment. Asparaginase which catalysis the conversion of asparagine to aspartic acid, thus limiting the amount of asparagine in the cells sensitizes cells to both genetic and pharmacologic PDH inhibition. Asparagine can be converted into oxaloacetate which can be converted into phosphoenolpyruvate (PEP) for gluconeogenesis. Phosphoenolpyruvate carboxykinase (PCK2) catalyzes the rate limiting step in gluconeogenesis converting oxaloacetate to PEP which can travel backwards through glycolysis. We found that PCK2 was highly upregulated in both the PDH knock cells and devimisat treated cells. Indicating that with the reduced glucose import and retention the cells rely on gluconeogenesis from amino acids to maintain needed glycolytic intermediates.

Conclusions:

This data shows that PDH inhibition leads to diminished glycolytic rate, increased reliance on amino acid metabolism and increased gluconeogenesis. This in turn suggests that glycolysis inhibition and amino acid deprivation in combination with devminisat mediated PDH inhibition should be explored in clinical trials.