Targeting glutaminolysis has antileukemic activity in acute myeloid leukemia and synergizes with BCL-2 inhibition

Abstract

Cancer cells require glutamine to adapt to increased biosynthetic activity. The limiting step in intracellular glutamine catabolism involves its conversion to glutamate by glutaminase (GA). Different GA isoforms are encoded by the genes GLS1 and GLS2 in humans. Herein, we show that glutamine levels control mitochondrial oxidative phosphorylation (OXPHOS) in acute myeloid leukemia (AML) cells. Glutaminase C (GAC) is the GA isoform that is most abundantly expressed in AML. Both knockdown of GLS1 expression and pharmacologic GLS1 inhibition by the drug CB-839 can reduce OXPHOS, leading to leukemic cell proliferation arrest and apoptosis without causing cytotoxic activity against normal human CD34(+) progenitors. Strikingly, GLS1 knockdown dramatically inhibited AML development in NSG mice. The antileukemic activity of CB-839 was abrogated by both the expression of a hyperactive GAC(K320A) allele and the addition of the tricarboxyclic acid cycle product α-ketoglutarate, indicating the critical function of GLS1 in AML cell survival. Finally, glutaminolysis inhibition activated mitochondrial apoptosis and synergistically sensitized leukemic cells to priming with the BCL-2 inhibitor ABT-199. These findings show that targeting glutamine addiction via GLS1 inhibition offers a potential novel therapeutic strategy for AML.

Figure 1

Glutamine controls mitochondrial OXPHOS in AML. (A) A total of 11 AML cell lines were cultured with or without Gln (4 mM) for 48 hours. Apoptosis was quantified by flow cytometric analysis of Annexin V binding. (B) OCI-AML2 and MOLM-14 cell lines were cultured for 48 hours with or without Gln (4 mM) and αKG (5 mM) as indicated, and apoptosis was quantified based on Annexin V binding. (C) OCI-AML2 and MOLM-14 cell lines were cultured for 48 hours with or without Gln (4 mM), and the OCR was measured using a Seahorse XF96 extracellular flux analyzer, under both basal conditions and after the addition of increasing doses of CCCP (0.25, 0.5, and 0.5 µM) and antimycine (10 µM), as indicated. Histograms show data that are representative of 3 independent experiments at baseline or after CCCP (cumulative dose, 1 µM). (D) MOLM-14 cell lines were cultured for 48 hours with or without Gln (4 mM) and αKG (5 mM) as indicated, and the OCR was measured as in C. Histograms show data that are representative of 3 independent experiments. *P < .05, **P < .01, ***P < .001.

Figure 2

GAC protein is prominently expressed in AML and modulates the OCR. (A) Analysis of human leukemic cell lines. AML cells from 8 patients and normal CD34⁺ HPCs from 4 healthy donors were analyzed by western blotting using anti-GLS1, anti-GLS2, and anti-ACTIN antibodies. (B) MOLM-14 (left) and OCI-AML2 (right) cells were transfected with a lentiviral vector expressing a doxycycline-inducible GLS1 shRNA (#5) construct. Stably infected cell lines were established by puromycin selection. After 2 days of doxycycline exposure, the OCR was measured using a Seahorse XF96 extracellular flux analyzer under both basal conditions and after the addition of CCCP and antimycin, as indicated. The inhibition of GAC expression was controlled in western blots using anti-GAC antibody. Histograms show data that are representative of 3 independent experiments. (C) OCI-AML2 cells were cultured with or without CB-839 (1 µM), BPTES (10 µM), or compound 968 (10 µM) for 6 hours, and the OCR was measured using a Seahorse XF96 extracellular flux analyzer. Histograms show data that are representative of 3 independent experiments. (D) OCI-AML2 cells were cultured with or without CB-839 (1 µM), BPTES (10 µM), or compound 968 (10 µM) for 4 hours, and 5 × 10⁶ cells were washed twice in cold PBS; the pellet was frozen and each indicted metabolite was measured. (E) OCI-AML2 cells were cultured with or without CB-839 (1 µM) and αKG (5 mM) for 6 h, and the OCR was measured. Histograms show data that are representative of 3 independent experiments. (F) OCI-AML2 cells were transfected with a lentiviral vector expressing a doxycycline-inducible V5-tagged GAC^WT or GAC^K320A construct. Stably infected cell lines were established by puromycin selection. After 6-hour doxycycline exposure with or without CB-839, the OCR was measured under both basal conditions and after the addition of CCCP and antimycin, as indicated. Histograms show data that are representative of 3 independent experiments (G) OCI-AML2 cells stably infected with GAC^WT or GAC^K320A were cultured for 6 hours with or without doxycycline or CB-839 (1 µM) and were analyzed by western blotting using anti-GAC and anti-ACTIN antibodies. *P < .05, **P < .01, ***P < .001.

Figure 3

Targeting glutaminase activity inhibits AML cell proliferation. (A) OCI-AML2 and MOLM-14 shGLS1#5 leukemic cells were seeded at 5 × 10⁵ cells/mL with or without doxycycline; viable cells were counted manually following trypan blue staining on days 3, 5, and 7. Each experiment was performed independently ≥3 times. (B) OCI-AML2, MV4-11, MOLM-14, HL-60, and OCI-AML3 cells transfected with shGLS1#5 or #7 were seeded at 5 × 10⁵ cells/mL with or without doxycycline; viable cells were counted following trypan blue staining on days 3, 5, and 7. Each experiment was performed independently ≥3 times. (C) OCI-AML2 and MOLM-14 shGLS1#5 leukemic cells were seeded at 5 × 10⁵ cells/mL with or without doxycycline, and αKG (5 mM) and viable cells were counted following trypan blue staining on day 7. (D) A total of 10 cell lines were seeded at 5 × 10⁵ cells/mL with or without CB-839 (1 µM); viable cells were counted following trypan blue staining on days 3, 5, and 7. (E) OCI-AML2 cells were seeded at 5 × 10⁵ cells/mL with or without CB-839 (1 µM) and αKG (5 mM); viable cells were counted following trypan blue staining on days 3, 5, and 7. (F) CD34⁺ HPCs from 4 healthy donors were seeded at 5 × 10⁵ cells/mL with and without CB-839 (1 µM); viable cells were counted following trypan blue staining on days 2, 3, 4, and 5. *P < .05, **P < .01, ***P < .001.

Figure 4

GLS1 controls AML survival by modulating the TCA cycle. (A) ShGLS1#5 was induced with doxycycline for 4 days in OCI-AML2, HL-60, MV4-11, MOLM-14, and OCI-AML3 leukemic cells, and apoptosis was evaluated based on Annexin-V binding. (B) CD34⁺ HPCs from 3 healthy donors were transfected with a lentiviral vector that expressed a noninducible GLS1 shRNA (#5), and transfected cells were selected with puromycin for 2 days; 3 days later, apoptosis was evaluated based on Annexin-V staining, and protein extracts were immunobloted with anti-GAC and anti-ACTIN antibodies. (C) MOLM-14 shGLS1 cells, MOLM-14 shCTR cells (9 mice each), and OCI-AML2 shGLS1 cells or OCI-AML2 shCTR cells (9 mice each) were intravenously injected into NSG mice (2 × 10⁶ cells per mouse). Mice were treated with doxycycline by oral gavage. The Kaplan-Meier survival curves of mice treated with doxycycline are shown for both cell lines. (D) The spleens of the mice injected with the OCI-AML2 shGLS1 and OCI-AML2 shCTR cell lines were measured. (E) A total of 11 cell lines were cultured for 4 days with or without CB-839 (1 µM), and apoptosis was evaluated based on Annexin-V staining. (F) OCI-AML2 and HL-60 cells were cultured for 4 days with or without CB-839 (1 µM), and dmαKG (5 mM) and apoptosis was evaluated based on Annexin-V staining. (G) OCI-AML2 cells that were stably infected with GAC^WT or GAC^K320A were cultured for 2 days with or without doxycycline or CB-839 (1 µM). Apoptosis was evaluated based on Annexin-V staining. (H) AML samples from 10 patients and CD34⁺ HPCs from 6 healthy donors were cultured for 4 days with or without CB-839 (1 µM). Apoptosis was evaluated based on Annexin-V staining. *P < .05, **P < .01, ***P < .001.

Figure 5

GLS1 inhibition induces mitochondrial apoptosis and sensitizes leukemic cells to priming with ABT-199. (A) OCI-AML2 WT cells or cells transfected with shGLS1#5 were cultured with or without CB-839 (1 µM) or doxycycline as indicated. Apoptosis was evaluated by western blotting using anti-CASPASE-8, cleaved CASPASE-3, anti-PARP, and anti-ACTIN antibodies. (B) OCI-AML2 cells were cultured with or without CB-839 (1 µM) and with or without a CASPASE-8 inhibitor (IETD) or a pan-caspase inhibitor (QVD). Protein extracts were immunoblotted using anti-CASPASE-8, anti-cleaved CASPASE-3, and anti-ACTIN antibodies. Mitochondrial depolarization was evaluated using TMRE staining and apoptosis was quantified based on Annexin V staining. (C) OCI-AML2 cells were cultured for 1 day with or without the indicated doses of CB-839 and ABT-199; apoptosis was evaluated based on Annexin V staining. (D) MOLM-14 cells were cultured for 1 day with or without the indicated doses of CB-839 and ABT-199; apoptosis was evaluated based on Annexin-V staining. Histograms show data that are representative of 3 independent experiments. The response of the combination was compared with its single agents against the widely used Loewe model for drug-with-itself dose additivity using Chalice software and presented as an isobologram. *P < .05, **P < .01, ***P < .001.