ATM inhibition drives metabolic adaptation via induction of macropinocytosis

Abstract

Macropinocytosis is a nonspecific endocytic process that may enhance cancer cell survival under nutrient-poor conditions. Ataxia-Telangiectasia mutated (ATM) is a tumor suppressor that has been previously shown to play a role in cellular metabolic reprogramming. We report that the suppression of ATM increases macropinocytosis to promote cancer cell survival in nutrient-poor conditions. Combined inhibition of ATM and macropinocytosis suppressed proliferation and induced cell death both in vitro and in vivo. Supplementation of ATM-inhibited cells with amino acids, branched-chain amino acids (BCAAs) in particular, abrogated macropinocytosis. Analysis of ATM-inhibited cells in vitro demonstrated increased BCAA uptake, and metabolomics of ascites and interstitial fluid from tumors indicated decreased BCAAs in the microenvironment of ATM-inhibited tumors. These data reveal a novel basis of ATM-mediated tumor suppression whereby loss of ATM stimulates protumorigenic uptake of nutrients in part via macropinocytosis to promote cancer cell survival and reveal a potential metabolic vulnerability of ATM-inhibited cells.

Figure 1.

Inhibition of ATM increases glucose and glutamine consumption, while knockdown of glucose or glutamine transporters in ATM-inhibited cells has no effect. (A–C) Ovcar8 cells were treated with the ATM inhibitors (10 μM KU60019 or 1 μM AZD0156) for 24 h in RPMI-1640 + 0.1% FBS. (A) pChk2 and total Chk2 expression were determined by immunoblotting. β-actin was used as a loading control. One of three experiments is shown. (B and C) Glucose and glutamine uptake were determined. n = 9/group, one of three experiments is shown. Data represent mean ± SD. *P < 0.005; unpaired two-sided t test. (D) Ovcar3 cells were infected with lentivirus expressing short hairpin RNAs (shRNAs) targeting SLC2A1 (GLUT1) or SLC2A4 (GLUT4). shGFP was used as a control. mRNA expression was determined by RT-qPCR. n = 3/group, one of two experiments is shown. Data represent mean ± SD. *P < 0.01 vs. control; one-way ANOVA. (E) Same as D, but cells were treated with 10 μM KU60019 for 24 h in RPMI-1640 + 0.1% FBS, and glucose uptake using the fluorescent glucose analog 2NBDG was determined by flow cytometry. MFI = median fluorescence intensity. n = 3/group, one of three experiments is shown. Data represent mean ± SD. *P < 0.0001 vs. KU60019; ns = not significant; one-way ANOVA with Tukey’s multiple comparisons. (F) Ovcar8 cells were infected with lentivirus expressing short hairpin RNAs (shRNAs) targeting SLC1A5. shGFP was used as a control. mRNA expression was determined by RT-qPCR. n = 3/group, one of two experiments is shown. Data represent mean ± SD. *P < 0.0001 vs. control; one-way ANOVA. (G) Same as F, but cells were treated with 10 μM KU60019 for 24 h in RPMI-1640 + 0.1% FBS, and glutamine uptake was determined. n = 9/group, one of three experiments is shown. Data represent mean ± SD. *P < 0.005 vs. KU60019; ns = not significant; one-way ANOVA with Tukey’s multiple comparisons.

Figure S1.

Inhibition of ATM increases glucose and glutamine consumption in multiple cell lines; related to Fig. 1. (A) Ovcar3 and Ovcar10 cells were treated with the ATM inhibitor KU60019 (10 μM) for 24 h in RPMI-1640 + 1% FBS. pChk2 and total Chk2 expression were determined by immunoblotting. Vinculin was used as a loading control. One of three experiments is shown. (B) Ovcar3 cells were treated with the ATM inhibitor KU60019 (10 μM) for 24 h in RPMI-1640 + 1% FBS, and glucose and glutamine consumption was determined. n = 9/group, one of two experiments is shown. Data represent mean ± SD. *P < 0.005; unpaired two-sided t test. (C) Same as B, but for Ovcar10 cells. (D) Glucose uptake in the indicated cells by 2NBDG was determined by flow cytometry after treatment with the ATM inhibitor KU60019 (10 μM) for 24 h in RPMI-1640 + 1% FBS. MFI = median fluorescence intensity. n = 3/group, one of at least two experiments is shown. Data represent mean ± SD. *P < 0.05; unpaired two-sided t test. (E) Parental Ovcar3 cells were infected with lentivirus expressing short hairpin RNAs (shRNAs) targeting SLC2A1 (GLUT1) or SLC2A4 (GLUT4). shGFP was used as a control. Glucose uptake was determined using the fluorescent glucose analog 2NBDG by flow cytometry. n = 6/group, one of two experiments is shown. Data represent mean ± SD. *P < 0.0001 vs. control; one-way ANOVA. (F) Parental Ovcar8 cells were infected with lentivirus expressing shRNAs targeting SLC1A5. shGFP was used as a control. Glutamine uptake was determined. n = 9/group, one of two experiments is shown. Data represent mean ± SEM. *P < 0.0001 vs. control; one-way ANOVA.

Figure S2.

Inhibition of ATM, but not ATR, induces macropinocytosis; related to Fig. 2. (A) Ovcar8 cells were treated with the ATM inhibitor KU60019 (10 μM) in combination with EIPA (25 μM) for 24 h in RPMI-1640 + 0.1% FBS. Glucose and glutamine uptake were determined. n = 9/group, one of three experiments is shown. Data represent mean ± SD. *P < 0.005; unpaired two-sided t test. (B) Gating strategy for dextran uptake flow cytometry experiments is shown. (C) Ovcar8 cells were treated with the ATM inhibitor KU60019 (10 μM) for 2 h using the indicated media conditions and dextran uptake was determined by flow cytometry. N = 3/group, one of three experiments is shown. Data represent mean ± SD. *P < 0.01; unpaired two-sided t test. (D) Ovcar8 cells were treated with the ATM inhibitor AZD0156 (1 μM) for 2 h in basal media, and dextran uptake was determined by flow cytometry. n = 3/group, one of three experiments is shown. Data represent mean ± SD. *P < 0.005; unpaired two-sided t test. (E and F) Ovcar8 cells were infected with lentivirus expressing short hairpin RNAs (shRNAs) targeting ATM. shGFP was used as a control. (E) Immunoblot analysis of ATM. Vinculin was used as a loading control. One of five experiments is shown. (F) Dextran uptake in basal media was determined by flow cytometry. n = 3/group, one of five experiments is shown. Data represent mean ± SD. *P < 0.0001 vs. control; one-way ANOVA. (G) Ovcar8 cells were treated with the ATM inhibitor KU60019 (10 μM) for the indicated time points in basal media, and dextran uptake was determined by flow cytometry. Data shown are the ratio of % cells with high dextran uptake in KU60019-treated cells vs. vehicle controls at each time point. n = 3/group. Data represent mean ± SD. *P < 0.0001 vs. controls; unpaired two-sided t test. (H) Ovcar8 cells were treated with the ATM inhibitor KU60019 (10 μM) for 2 h in basal media, and 70 kD TMR-dextran uptake was determined by flow cytometry. N = 3/group, one of three experiments is shown. Data represent mean ± SD. *P < 0.0001; unpaired two-sided t test. (I–J) Ovcar8 cells were treated with the ATR inhibitor VE822 (1 μM) for 2 h in basal media. (I) Dextran uptake was determined by flow cytometry. n = 3/group, one of three experiments is shown. Data represent mean ± SD. ns = not significant; unpaired two-sided t test. (J) Immunoblot analysis of p-Chk1 and total Chk1. β-actin was used as a loading control. One of three experiments is shown. (K and L) Ovcar8 cells were infected with lentivirus expressing short hairpin RNAs (shRNAs) targeting ATR. shGFP was used as a control. (K) Immunoblot analysis of ATR. Vinculin was used as a loading control. One of three experiments is shown. (L) Dextran uptake in basal media was determined by flow cytometry. n = 3/group, one of three experiments is shown. Data represent mean ± SD. *P < 0.06; ns = not significant; one-way ANOVA. (M) Ovcar8, Ovcar3, and Ovcar10 cells were treated with the ATM inhibitor KU60019 (10 μM) for 2 h in basal media. Immunoblot analysis of p-Chk2 and total Chk2. β-actin was used as a loading control. One of three experiments is shown.

Figure 2.

Inhibition of ATM kinase activity induces macropinocytosis. (A) Ovcar8 cells were treated with the ATM inhibitor 10 μM KU60019 for 2 h in basal media, and 70 kD dextran uptake was determined by fluorescence microscopy. Scale bars = 25 μm (large images), 15 μm (insets). Arrowheads point to dextran-FITC. N > 20/group, one of three experiments is shown. Data represent mean ± SD. *P < 0.005 vs. KU60019; one-way ANOVA with Tukey’s multiple comparisons. (B) Ovcar8 cells were treated with the ATM inhibitor KU60019 (10 μM) alone or in combination with the macropinocytosis inhibitor EIPA (25 μM) or eHop-016 (2.5 μM) for 2 h in basal media, and dextran uptake was determined by flow cytometry. n = 3/group, one of at least four experiments is shown. (C and D) Ovcar3 (C) and Ovcar10 (D) cells were treated with the ATM inhibitor KU60019 (10 μM) alone or in combination with the macropinocytosis inhibitor EIPA (25 μM) for 2 h in basal media, and dextran uptake was determined by flow cytometry. n = 3/group, one of at least three experiments is shown. Data represent mean ± SD. *P < 0.0001 vs. KU60019; one-way ANOVA with Tukey’s multiple comparisons.

Figure 3.

Inhibition of macropinocytosis is a vulnerability of ATM-inhibited cells both in vitro and in vivo. (A and B) Ovcar8 cells were treated with the ATM inhibitor KU60019 (2 μM) or the macropinocytosis inhibitor EIPA (2 μM) alone or in combination for 3 d in RPMI-1640 + 5% FBS. (A) Proliferation was assessed by crystal violet staining. n = 3/group, one of three experiments is shown. Data represent mean ± SD. *P < 0.001; one-way ANOVA. (B) Apoptosis was assessed by Annexin V/7AAD staining. n = 3/group, one of three experiments is shown. Data represent mean ± SD. *P < 0.001; one-way ANOVA. (C) Ovcar8 cells expressing luciferase were injected intraperitoneally into nude mice (6–8 mice/group). At day 19, mice were randomized and thereafter treated daily through intraperitoneal injection with the ATM inhibitor KU60019 (10 mg/kg) or the macropinocytosis inhibitor EIPA (10 mg/kg) alone and in combination. Bioluminescence flux was determined using noninvasive IVIS imaging. Shown is the tumor growth curve over 30 d normalized to Day 19 tumors. Longitudinal analysis: combination vs. vehicle, P = 0.00141; combination vs. KU60019, P = 0.002334; combination vs. EIPA, P = 0.003603; linear mixed model fit REML. (D) Representative luciferase images of mice from each group at Day 48. (E) Quantification of D normalized to Day 19 tumors. Cross-sectional analysis: ANOVA with Wilcoxon rank sum test. (F) Representative cleaved caspase 3 immunohistochemistry in the indicated tumors. Scale bars = 100 μm (large images), 25 μm (insets). (G) H-Score of cleaved caspase 3 IHC. Data represent mean ± SEM. *P < 0.01; one-way ANOVA. (H) Representative pChk2 and total Chk2 immunoblot analysis of the indicated tumors. Vinculin was used as a loading control. (I) Representative confocal images of dextran-FITC in the indicated tumors. Scale bars = 10 μm (large images), 2.5 μm (insets). (J) Macropinocytotic index of dextran-FITC uptake in the tumors was assessed. Data represent mean ± SEM. *P < 0.0005 vs. KU60019; one-way ANOVA with Tukey’s multiple comparisons.

Figure S3.

Inhibition of macropinocytosis inhibits cell proliferation and increases cell death in ATM inhibitor-treated cells. Related to Fig. 3. (A) Ovcar8 cells were treated with the ATM inhibitor KU60019 (2 μM) or the macropinocytosis inhibitor EIPA (2 μM) alone and in combination for 3 d in RPMI-1640 + 5% FBS complete media, and dextran uptake was determined at the end of the experiment. n = 3/group, one of two experiments is shown. Data represent mean ± SD. *P < 0.05; one-way ANOVA. (B and C) Ovcar3 (B) or Ovcar10 (C) cells were treated with the ATM inhibitor KU60019 (2 μM) or the macropinocytosis inhibitor EIPA (2 μM) alone and in combination for 3 d in RPMI-1640 + 5% FBS complete media. Proliferation (left panel) was assessed by crystal violet staining. Apoptosis (right panel) was assessed by Annexin V/7AAD staining. n = 3/group, one of three experiments is shown. Data represent mean ± SD. *P < 0.01; one-way ANOVA. (D) Synergy analysis using Combenefit software using the LOEWE model. (E) Ovcar8 cells were treated with the ATM inhibitor KU60019 (2 μM) or the macropinocytosis inhibitor EIPA (1.25 μM) alone and in combination for 3 d in RPMI-1640 + 0.1% FBS complete media. n = 3/group, one of four experiments is shown. Data represent mean ± SD. *P < 0.005; one-way ANOVA. (F) Ovcar8 cells were treated with the ATM inhibitor KU60019 (2 μM) or the Rac1 inhibitor eHop-016 (1.75 μM) alone and in combination for 3 d in RPMI-1640 + 5% FBS complete media. Proliferation was assessed by crystal violet staining. One of four experiments is shown. Data represent mean ± SD. *P < 0.001; one-way ANOVA. (G) Ovcar8 cells were treated with the ATM inhibitor KU60019 (2 μM) or the Rac1 inhibitor eHop-016 (1.4 μM) alone and in combination for 2 d in RPMI-1640 + 0.1% FBS complete media. Proliferation was assessed by crystal violet staining. One of two experiments is shown. Data represent mean ± SD. *P < 0.0001; one-way ANOVA. (H) IMR90 fibroblasts (non-macropinocytotic) were treated with the macropinocytosis inhibitor EIPA (2 μM) for 3 d. Proliferation was assessed by crystal violet staining. Data represent mean ± SD. ns = not significant. Unpaired two-sided t test. One of two experiments is shown. (I) pChk2 and total Chk2 immunoblot analysis in all of the tumors. β-Actin was used as a loading control.

Figure 4.

ATM inhibitor-induced macropinocytosis leads to increased branched-chain amino acid uptake and affects mTORC1 activity. (A and B) Ovcar8 cells were treated with the ATM inhibitor KU60019 (10 μM) alone or in combination with the indicated metabolites (concentrations in Materials and methods) for 2 h in basal media, and dextran uptake was determined by flow cytometry. n = 3/group, one of at least three experiments is shown. Data represent mean ± SD. Dotted line indicates controls for each metabolite. *P < 0.05 vs. KU60019; one-way ANOVA with Tukey’s multiple comparisons. (C) Ovcar8 cells were treated with the ATM inhibitor KU60019 (10 μM) alone or in combination with the macropinocytosis inhibitor EIPA (25 μM) for 2 h in basal media, and isotopolog enrichment of BCAAs was assessed by mass spectrometry. n = 3/group, one of at least three experiments is shown. Data represent mean ± SD. *P < 0.05 vs. KU60019; one-way ANOVA with Tukey’s multiple comparisons. (D) BCAA metabolite abundance vs. ATM protein expression from DepMap.org. (E) Ovcar8 cells were treated with the ATM inhibitor KU60019 (10 μM) alone or in combination with 200 μM BCAAs for 2 h in basal media, and Western blotting was performed on the indicated proteins. Vinculin was used as a loading control. One of four experiments is shown. (F) Ovcar3 and Ovcar10 cells were treated with the ATM inhibitor KU60019 (10 μM) for 2 h in basal media, and Western blotting was performed on the indicated proteins. Vinculin was used as a loading control. One of three experiments is shown. (G) Fold change (Log₂FC) of metabolites in the ascites fluid in the indicated treatment groups.

Figure S4.

pChk2 protein expression negatively correlates with branched chain amino acids and positively correlates with mTORC1 activity; no change in alpha-ketoacids in ATM inhibited cells; tumor and interstitial fluid metabolites. Related to Fig. 4. (A and B) Ovcar8 cells were treated with the ATM inhibitor KU60019 (10 μM) alone or in combination with the macropinocytosis inhibitor EIPA (25 μM) for 2 h in basal media. (A) BCAA abundance was assessed by mass spectrometry. n = 8/group, one of at least three experiments is shown. Data represent mean ± SD. *P < 0.05 vs. KU60019; one-way ANOVA with Tukey’s multiple comparisons. (B) Relative percent isotopolog molar abundance (% isotopolog enrichment x total molar pool size). n = 8/group, one of three experiments is shown. Data represent mean ± SD. *P < 0.05 vs. KU60019; one-way ANOVA with Tukey’s multiple comparisons. (C) BCAAs metabolite abundance vs. pChk2 protein expression from depmap.org. (D) Ovcar8 cells were treated with the ATM inhibitor KU60019 (10 μM) for 2 h, and isotopolog enrichment of downstream BCAA metabolites (KMV/KIC, KIV) was assessed by mass spectrometry. n = 3/group, one of at least three experiments is shown. Data represent mean ± SD. ns = not significant; unpaired two-sided t test. (E) p4EBP1 and pS6K protein expression vs. pChk2 protein expression from PanCancer TCGA data. (F) Ovcar8 cells were treated with the ATM inhibitor KU60019 (2 μM) or the macropinocytosis inhibitor EIPA (2 μM) alone and in combination with and without 200 μM BCAAs for 3 d in RPMI-1640 + 5% FBS. Proliferation was assessed by crystal violet staining. One of two experiments is shown. Data represent mean ± SD. *P < 0.0001 vs. KU60019 and EIPA alone, #P < 0.0001 vs. combination treatment; one-way ANOVA. (G) Volcano plot of fold change (Log₂FC) of metabolites in ATM inhibitor KU60019-treated vs. control tumors or combination-treated (KU60019+EIPA) vs. KU60019-treated tumors. Dotted line equals P = 0.05. (H) Fold change (Log₂FC) of metabolites in the interstitial fluid in the indicated treatment groups.