Targeting mitochondrial glutaminase activity inhibits oncogenic transformation

Abstract

Rho GTPases impact a number of activities important for oncogenesis. We describe a small molecule inhibitor that blocks oncogenic transformation induced by various Rho GTPases in fibroblasts, and the growth of human breast cancer and B lymphoma cells, without affecting normal cells. We identify the target of this inhibitor to be the metabolic enzyme glutaminase, which catalyzes the hydrolysis of glutamine to glutamate. We show that transformed fibroblasts and breast cancer cells exhibit elevated glutaminase activity that is dependent on Rho GTPases and NF-κB activity, and is blocked by the small molecule inhibitor. These findings highlight a previously unappreciated connection between Rho GTPase activation and cellular metabolism and demonstrate that targeting glutaminase activity can inhibit oncogenic transformation.

Figure 1. The small molecule 968 inhibits cellular transformation

(A) Left: NIH 3T3 cells were transiently transfected with oncogenic Dbl and cultured for 14 days in 5% calf serum, while treated with different benzo[a]phenanthridinones (designated 384, 335, 968, 537, and 343) (10 μM each). Cells were fixed with 3.7% formaldehyde in PBS and stained with crystal violet for counting foci. Right: 968 was serially diluted (10, 5, 2.5, and 1.25 μM) and evaluated for its ability to inhibit focus formation. (B) NIH 3T3 cells were stably transfected with Dbl and grown in DMEM supplemented with 1% calf serum and the indicated amounts of 968. After 6 days, the cells were counted. 100% represents the number of Dbl-transformed cells counted in the absence of 968 (27.5 × 10⁴ cells). Data represent the average of 3 experiments (± s.d.). (C) Chemical structures of the benzo[a]phenanthridinone derivatives examined for their effects on Dbl-induced focus formation (1A and S1B). (D) Control NIH 3T3 cells were cultured in DMEM supplemented with 10% calf serum in 6 well plates, and were either treated with 10 μM 968 or 335, or untreated. At the indicated times, the cells were counted. Data represent the average of 3 experiments (± s.d.). (E) Photomicrographs of Dbl-transfected NIH 3T3 cells (bottom panels) and control NIH 3T3 cells (top panels) cultured in 10% calf serum and treated with either DMSO (vehicle control) or 10 μM 968.

Figure 2. Effects of 968 on the transforming activity of constitutively active Rho GTPases

(A) NIH 3T3 cells stably expressing hemagglutinin (HA)-tagged Cdc42(F28L), Rac(F28L), RhoC(F30L), or vector control cells, either treated with 10 μM 968 or untreated, were grown in soft agar (plus DMEM supplemented with 10% calf serum). Top: Relative expression of the HA-tagged GTPases. Bottom: Cells were scored after 14 days and plotted as the percentage of the total number of colonies greater than 50 μm in diameter. Data represent the average of 3 experiments (± s.d.). (B) Cells were cultured in DMEM supplemented with 10% calf serum with or without 10 μM 968, for 6 days, and then were trypsinized and counted. Data represent the average of 3 experiments (± s.d.). (C) Cells were cultured in DMEM supplemented with 1% calf serum, treated with 10 μM 968 or untreated, and counted at the indicated times. Data represent the average of 3 experiments (± s.d.). (D) Cells were serum-starved, treated with 10 μM 968 or untreated, and seeded in MilliCell upper chambers containing growth factor-reduced Matri-gel. After 24 hours at 37°C, the migratory cells were fixed, stained with GIEMSA, and counted. Data represent the average of 3 experiments (± s.d.).

Figure 3. Rho GTPases are hyper-activated and are important for the growth of breast cancer cells

(A) Lysates from MDA-MB231 cells, SKBR3 cells, and HMECs, were prepared and incubated with GST fused to the limit Rho-binding domain on Rhotekin (GST-RBD). The top panels show the relative levels of RhoA-GTP and RhoC-GTP that were co-precipitated with GST-RBD from the indicated cells, as detected by Western blotting with an anti-RhoA monoclonal antibody and an anti-RhoC polyclonal antibody. The middle panels compare the relative expression of RhoA and RhoC in whole cell lysates (WCL) from the different cells and the bottom panel shows the relative input of GST-RBD. (B) MDA-MB231 cells were transfected with control siRNA or siRNAs targeting RhoA and RhoC. Top: Shown are the efficiencies of the siRNAs targeting RhoA and RhoC as assessed by Western blot analysis using anti-RhoA and anti-RhoC antibodies. Bottom: Cells were grown for the indicated number of days in RPMI 1640 medium supplemented with 1% fetal bovine serum and counted. Data represent the average of 3 experiments (mean ± s.d.). (C) MDA-MB231 cells transfected with control siRNA or siRNAs targeting RhoA and RhoC were grown in soft agar and scored after 10 days. Histograms show the percentage of the total number of colonies greater than 50 μm in diameter. Data represent the average of 3 experiments (mean ± s.d.). (D) SKBR3 cells were transfected with control siRNA or siRNAs targeting RhoA and RhoC. Top: Shown are the efficiencies of the siRNAs targeting RhoA and RhoC. Bottom: Cells were grown in RPMI 1640 medium supplemented with 1% fetal bovine serum for the indicated number of days and counted. Data represent the average of 3 experiments (mean ± s.d.). (E) SKBR3 cells transfected with control siRNA or siRNAs targeting RhoA and RhoC were grown in soft agar and scored after 10 days. Histograms show the percentage of the total number of colonies greater than 50 μm in diameter. Data represent the average of 3 experiments (mean ± s.d.).

Figure 4. Effects of 968 on the growth and invasive activity of human breast cancer cells

(A) MDA-MB231 cells, SKBR3 cells, and NIH 3T3 cells stably expressing Dbl, were either treated with 10 μM 968 or untreated, and grown in soft-agar as in 2A. Data represent the average of 3 experiments (± s.d.). (B) MDA-MB231 and SKBR3 cells were cultured in RPMI 1640 medium supplemented with 10% fetal bovine serum in 12 well plates for the indicated number of days in the presence or absence of 10 μM 968, and then counted. Data represent the average of 3 experiments (± s.d.). (C) Breast cancer cells were cultured in RPMI 1640 medium supplemented with 10% fetal bovine serum, and HMECs were cultured in MEGM complete medium in 12 well plates, for 6 days in the presence or absence of 10 μM 968, and counted. Data represent the average of 3 experiments (± s.d.). (D) Breast cancer cells were cultured in RPMI 1640 medium supplemented with 1% fetal bovine serum, treated with 10 μM 968 or untreated, and counted at the indicated times. Data represent the average of 3 experiments (± s.d.). (E) Breast cancer cells were serum-starved, treated with 968 or untreated, and seeded in MilliCell upper chambers containing growth-factor-reduced Matri-gel. After 24 hours at 37°C, the migratory cells were fixed, stained with GIEMSA, and counted. Data represent the average of 3 experiments (± s.d.). (F) HMECs were cultured in MGEM complete medium in 12 well plates for the indicated number of days in the presence and absence of 10 μM 968, and then counted. (G) Photomicrographs of HMECs cultured in MGEM complete medium treated with either DMSO (vehicle control) or 10 μM 968.

Figure 5. Glutaminase serves as a target for 968

(A) The E. coli-expressed mouse ortholog of human GAC was pre-incubated with the indicated amounts of 968 (●) or 335 (○) for 30 minutes at 37°C and then assayed. 100% = 620 moles of glutamine hydrolyzed per minute per mole of enzyme. Data represent the average of 3 experiments (± s.d.). Top panel: The biotin-labeled, active moiety of 968 linked to streptavidin-agarose beads, or control beads alone, were incubated with NIH 3T3 cell lysates transiently expressing V5-tagged mouse GAC. Following precipitation of the beads and re-suspension, the samples were analyzed by Western blotting with anti-V5 antibody. (B) NIH 3T3 cells stably expressing HA-Cdc42(F28L), cells stably expressing HA-Cdc42(F28L) transfected with control siRNA or siRNAs targeting both isoforms of mouse KGA, or control cells, were grown in DMEM supplemented with 1% calf serum. Top: Efficiencies of siRNAs targeting both isoforms of KGA, and relative levels of HA-Cdc42 in the different cells. Bottom: Growth profiles for the indicated cell lines. Data represent the average of 3 experiments (± s.d.). (C) NIH 3T3 cells stably expressing Cdc42(F28L), cells stably expressing Cdc42(F28L) transfected with control siRNA or siRNAs targeting both isoforms of mouse KGA, and control (vector) cells, were grown in soft agar and scored after 10 days. Histograms show the percentage of the total number of colonies greater than 50 μm in diameter. Data represent the average of 3 experiments (mean ± s.d.). (D) Breast cancer cells transfected with control siRNA or siRNAs targeting both isoforms of KGA were grown in RPMI 1640 medium supplemented with 1% fetal bovine serum. Top: Efficiencies of siRNAs targeting both isoforms of KGA. Bottom: Growth profiles for the indicated cell lines. Data represent the average of 3 experiments (± s.d.). (E) The indicated breast cancer cell lines were transfected with control siRNA or with siRNAs targeting both isoforms of KGA and then grown in soft agar and scored after 10 days. Data represent the average of 3 experiments (mean ± s.d.).

Figure 6. The relationship between the glutamine dependence of transformed/cancer cells and their sensitivity to 968

(A) MDA-MB231 cells and SKBR3 cells were grown in 10% fetal bovine serum in the presence or absence of glutamine for the indicated number of days and counted. (B) NIH 3T3 cells stably expressing Dbl and cells stably expressing Cdc42(F28L) were grown in 10% calf serum in the presence or absence of glutamine for the indicated number of days and counted. (C) NIH 3T3 cells stably expressing HA-tagged Dbl alone, and cells stably expressing HA-Dbl that were transfected with control siRNA or siRNAs targeting both isoforms of KGA, were cultured in DMEM supplemented with 1% calf serum in the presence and absence of 7 mM dimethyl α-ketoglutarate (DM-α-KG). Top: Efficiencies of siRNAs targeting both isoforms of KGA, and relative levels of HA-Dbl in the different cells. Bottom: Growth profiles for the indicated cell lines. Data represent the average of 3 experiments (± s.d.). (D) NIH 3T3 cells stably expressing Dbl were cultured in DMEM supplemented with 1% calf serum in the presence of 10 μM 968 alone or together with 7 mM DM-α-KG. Data represent the average of 3 experiments (± s.d.). (E) NIH 3T3 cells transiently expressing Dbl were assayed for focus formation in the presence of 10 μM 968 alone or together with 7 mM DM-α-KG. (F) MDA-MB231 cells were grown in 1% fetal bovine serum in the presence of 10 μM 968 alone or together with 7 mM dimethyl α-ketoglutarate (DM-α-KG) for the indicated number of days and counted. Data represent the average of 3 experiments (mean ± s.d.). (G) SKBR3 cells were grown in 1% fetal bovine serum in the presence of 10 μM 968 alone or together with 7 mM dimethyl α-ketoglutarate. Data represent the average of 3 experiments (± s.d.).

Figure 7. Effects of 968 on B lymphoma cells and Cdc42(F28L)-transformed cells over-expressing GAC

(A) P493 B lymphoma cells were cultured in RPMI 1640 medium supplemented with 10% fetal bovine serum in the presence of the indicated concentrations of 968. (B) P493 B lymphoma cells were cultured in RPMI 1640 medium supplemented with 10% fetal bovine serum in the presence of the inactive 968-analog 335. (C) P493 B lymphoma cells (2 × 10⁷) were injected s.c. into the flank of SCID mice. At day 12 the tumors were ~170 mm³, at which time treatments with 968 (200 μg per injection) were begun by daily intraperitoneal injections (200 μL) for 12 days. Control animals were treated with vehicle (DMSO in PBS). N = 7. The p value on day 24 = 0.000375. (D) Left: Control NIH 3T3 cells, NIH 3T3 cells transiently expressing Dbl, cells stably expressing Cdc42(F28L) and transiently expressing mouse GAC, cells stably expressing Cdc42(F28L), cells transiently expressing GAC, and cells stably expressing Cdc42(F28L) and transiently expressing Dbl, were examined for focus-forming activity. Right: Quantification of foci. Data represent the average of 3 experiments (± s.d.). (E) Left: Focus-forming assays performed on NIH 3T3 cells stably expressing Cdc42(F28L), cells stably expressing Cdc42(F28L) and transiently expressing Dbl, and cells stably expressing Cdc42(F28L) and either transiently expressing wild-type mouse GAC or the GAC(S291A) mutant. Right: Quantification of foci. Data represent the average of 3 experiments (± s.d.).

Figure 8. Role of glutaminase activity in cellular transformation

(A) Mitochondrial fractions were prepared from equivalent numbers of control NIH 3T3 cells and cells stably expressing Dbl, Cdc42(F28L), Rac(F28L), or RhoC(F30L) that had been treated with 10 μM 968 or untreated. Top: Relative amounts of KGA in the different mitochondrial preparations as assessed using an antibody that recognizes both KGA isoforms. Bottom: The different mitochondrial fractions were assayed for basal glutaminase activity. 100% = 680 nM glutamine hydrolyzed per minute per 10⁵ cells. Data are the average of 3 experiments (± s.d.). (B) Mitochondrial fractions were prepared from equivalent numbers of the indicated breast cancer cells that had been treated with 10 μM 968 or untreated. Top: Relative amounts of KGA and VDAC in the mitochondrial preparations. Bottom: The different mitochondrial fractions were assayed for basal glutaminase activity. 100% = 750 nM glutamine hydrolyzed per minute per 10⁵ cells. Data represent the average of 3 experiments (± s.d.). (C) Mitochondrial fractions from NIH 3T3 cells stably expressing Dbl that were cultured for 2 days in the presence or absence of 2 μM BAY 11-7082, or transfected with control siRNA or siRNAs targeting p65/RelA. Top: Relative amounts of KGA in the mitochondria from the indicated cells as assessed using an antibody that recognizes both KGA isoforms, and relative efficiencies of two siRNAs targeting p65/RelA. Bottom: The basal glutaminase activity for the different mitochondrial fractions. 100% represents the activity for untreated Dbl-transformed cells. Data represent the average of 2 experiments (± range). (D) Mitochondrial fractions were prepared from HMECs, and SKBR3 cells that had been treated with 2 μM BAY 11-7082 or untreated, or transfected with siRNAs targeting p65/RelA. Top: Relative amounts of KGA in the different mitochondrial fractions as assessed using an antibody that reorganizes both KGA isoforms, and relative efficiencies of two siRNAs targeting p65/RelA. Bottom: The basal glutaminase activity for the different mitochondrial fractions. 100% represents the activity for untreated SKBR3 cells. Data represent the average of 2 experiments (± range). (E) Top: V5-GAC was transiently expressed in control NIH 3T3 cells or NIH 3T3 cells stably expressing Dbl that were treated with either 2 μM BAY 11-7082, 10 μM 968, or untreated, and then immunoprecipitated from the different samples. Bottom: Basal activity for V5-GAC immunoprecipitated from the different cells. Data represent the average of 3 experiments (± s.d.). (F) V5-GAC was transiently expressed in NIH 3T3 cells stably expressing Dbl. The V5-GAC was immunoprecipitated, treated with alkaline phosphatase for 1 hour at 37°C (or untreated), and assayed for activity in the presence and absence of phosphate.