Gls Inhibitor CB-839 Modulates Cellular Metabolism in AML and Potently Suppresses AML Cell Growth When Combined with 5-Azacitidine

Glutamine (Gln) is required for growth and proliferation of several tumor types including AML. Glutaminase (GLS) is a mitochondrial enzyme that catalyzes conversion of Gln to glutamate (Glu), which provides carbons for the TCA cycle and regulates redox homeostasis through production of glutathione and NADH. CB-839 is a highly selective, reversible, allosteric inhibitor of GLS. In this study we studied metabolic and cellular consequences of GLS inhibition in AML cells cultured in normoxic or hypoxic conditions.

First, we performed metabolomic analysis of HL-60 cells co-cultured with bone marrow (BM)-derived mesenchymal stem cells (MSCs). Consistent with the known mechanism of GLS inhibition, CB-839 caused a rapid and extensive decrease in intracellular Glu in both HL60 and MSC and a corresponding increase in intracellular Gln in both cell types. Unexpectedly, CB-839-treated cells exhibited a rapid increase in intracellular and extracellular concentrations of multiple amino acids (Phe, kynurenine, Trp, Leu, Ile, Met, Tyr, Val, Thr, Ala, Gln, Asn, and His), possibly reflecting inhibition of global protein synthesis. CB-839 suppressed cysteine consumption from the extracellular compartment and caused rapid increase in intracellular taurine in HL-60 cells, suggesting altered redox homeostasis (Fig. 1A). CB-839 inhibited cellular growth of HL-60 and MV4;11 AML cells cultured alone or co-cultured with MSC, under conditions mimicking the BM microenvironment (Fig. 1B).

Stable isotope-resolved metabolomics (SIRM) analysis with ¹³C₅, ¹⁵N₂-Gln in HL-60 cells indicated that treatment with CB-839 severely hindered Gln anaplerosis to similar extent under normoxic or hypoxic conditions. Moreover, Gln is predominantly used to carry out oxidative metabolism. The enriched fraction of aspartate in treated cells dropped dramatically (to approximately 20% or less of the pool), suggesting that leukemia cells require Krebs cycle-derived oxaloacetate transamination for the generation of aspartate (Fig. 1C). Limiting Gln supply using CB-839 caused reduction in the concentration of alpha-ketoglutarate (α-KG) and the oncometabolite 2-hydroxyglutarate (2-HG), known to play a role in the pathogenesis of AML.

We have previously shown that the leukemic BM microenvironment is highly hypoxic (Benito PLoS One 2011), andhypoxia has been reported to induce production of the L-enantiomer of 2-HG (L-2HG) (Intlekofer Cell Metabolism 2015). In AML cells, hypoxia selectively induced the production of L-2HG measured by LC-MS/MS in HL-60 (6.2 fold) and OCI-AML3 cells (2.9 fold) with wt-IDH. This increase in L-2HG was potently inhibited by CB-839, implicating Gln as a source for L-2HG production by AML cells under hypoxia. HL-60 and OCI-AML3 AML cells produced very limited amounts of the D-enantiomer of 2HG (D-2HG), and neither hypoxia nor CB-839 significantly affected D-2HG levels. We recently reported that CB-839 increased hydroxymethylation (hmc) levels using a HELP-GT assay (Velez ASH 2015), and the implications of those observations are the subject of ongoing studies.

Prompted by the observation of increased hmc in response to CB-839 treatment, we next examined the efficacy of CB-839 in combination with the DNMT3A inhibitor 5-azacitidine (5-AZA). Treatment with 1µM CB-839 and escalating doses of 5-AZA caused additive or synergistic inhibition of cellular growth after 5 days of culture, both under normoxia and hypoxia, in AML cell lines (OCI-AML3, HL-60, MV4;11) and in primary AML cells (n=3) (Fig. 1D).

To test the efficacy of both compounds in vivo, we injected NSG-S mice with genetically engineered MV4;11/Luc cells. Bioluminescent imaging (BLI) demonstrated significantly reduced leukemia burden in treated groups compared to controls, more prominently in the CB-839 plus 5-AZA co-treated mice. CB-839 and 5-AZA co-treatment resulted in significant extension of survival compared with 5-AZA single agent, p<0.001 (Fig. 1E).

In summary, GLS inhibition causes AML growth arrest by multiple mechanisms, including inhibition of protein synthesis and disruption of redox homeostasis. Gln contributes to hypoxia-induced production of L-2HG and possibly epigenome regulation in AML, and concomitant blockade of GLS by CB-839 and DNMT3A with 5-AZA potently suppresses AML cell growth in vitro and in vivo. The clinical trial examining the efficacy of this combination is ongoing (Calithera, NCT02071927).

View largeDownload slide

View largeDownload slide

Close modal