Metabolic shifts toward glutamine regulate tumor growth, invasion and bioenergetics in ovarian cancer

Abstract

Glutamine can play a critical role in cellular growth in multiple cancers. Glutamine-addicted cancer cells are dependent on glutamine for viability, and their metabolism is reprogrammed for glutamine utilization through the tricarboxylic acid (TCA) cycle. Here, we have uncovered a missing link between cancer invasiveness and glutamine dependence. Using isotope tracer and bioenergetic analysis, we found that low-invasive ovarian cancer (OVCA) cells are glutamine independent, whereas high-invasive OVCA cells are markedly glutamine dependent. Consistent with our findings, OVCA patients' microarray data suggest that glutaminolysis correlates with poor survival. Notably, the ratio of gene expression associated with glutamine anabolism versus catabolism has emerged as a novel biomarker for patient prognosis. Significantly, we found that glutamine regulates the activation of STAT3, a mediator of signaling pathways which regulates cancer hallmarks in invasive OVCA cells. Our findings suggest that a combined approach of targeting high-invasive OVCA cells by blocking glutamine's entry into the TCA cycle, along with targeting low-invasive OVCA cells by inhibiting glutamine synthesis and STAT3 may lead to potential therapeutic approaches for treating OVCAs.

Figure 1. Glutamine addiction and cancer invasiveness are positively correlated in ovarian cancer cells (OVCA)

1. Gln deprivation effect on proliferation rate of a panel of OVCA cells for 24, 48 and 72 h. OVCAR3, IGROV1, OVCA429 cells were Gln independent; OVCA420 and OVCAR8 cells were moderately Gln dependent; and SKOV3, SKOV3ip and Hey 8 cells were Gln dependent. n ≥ 15.

2. Glucose deprivation effect on proliferation rate of OVCA cell lines for 24, 48, 72 h. n ≥ 15.

3. Gln deprivation effect on clonogenic formation in IGROV1 and SKOV3. n = 6.

4. Matrigel invasion assay was conducted to characterize invasiveness of OVCA cell lines. OVCAR3, IGROV1 and OVCA429 cells were noninvasive, OVCA420 and OVCAR8 cells were moderately invasive, while SKOV3, SKOV3ip and Hey 8 were highly invasive OVCA cell lines. n ≥ 6.

5. Correlation between proliferation rate at 72 h of OVCA cell lines under Gln‐depleted conditions with their corresponding number of invaded cells.

6. Western blot analysis of cell cycle proteins linked with growth rate (CDC2, Cyclin D1) and protein linked with metastatic potential (E‐cadherin) in OVCA cell lines. β‐actin was used as the loading control.

7. Genes in the glutaminolysis and tricarboxylic acid cycle metabolic pathways are associated with higher risk in OVCA patients. Shown is a metabolic network and the genes (rectangles) that produce the enzyme catalyzing the reactions in the network. The genes in the network are colored by their correlation with OVCA patient survival based on calculating their Cox hazard values (color key). The gene expression and clinical data were fitted with a Cox proportional hazard model to determine the hazard ratio for each gene in Gln/glucose metabolic network. A higher hazard ratio (poor outcome) was observed in gene products that catalyze reactions in Gln catabolism and higher expression of glycolytic genes culminated in better patient survival (low hazard ratio).

Data information: Data in (A–D) are expressed as mean ± SEM, *P < 0.05, **P < 0.01, ***P < 0.001. In (D), 1‐way ANOVA with Tukey's test was used to compare between cell lines, using OVCAR3 as the control.

Figure 2. Glutamine‐driven tricarboxylic acid (TCA) cycle metabolite abundances, reductive carboxylation and oxidative phosphorylation are significantly higher in high‐invasive than in low‐invasive ovarian cancer (OVCA) cells

1. A

   Schematic of carbon atom transitions using 1:1 mixture of ¹³C₆ glucose and 1‐¹³C‐labeled glucose. This allows estimations of the contribution of glucose toward TCA cycle metabolites and synthesis of glutamate and Gln. Black color represents labeled carbon on 1^st TCA cycle, yellow color represents labeled carbon on 2^nd TCA cycle, and blue color represents labeled carbon on 3^rd TCA cycle.

2. B–D

   Glucose's contribution to TCA metabolites and glutamate pool. Comparison of mass isotopomer distribution (MID) of malate (B), fumarate (C), glutamate (D) in high‐invasive (SKOV3) and low‐invasive (OVCAR3) cells cultured with a 1:1 mixture of ¹³C₆ glucose and 1‐¹³C‐labeled glucose.

3. E

   Schematic of carbon atom transitions using ¹³C₅‐labeled Gln. Black color represents labeled carbon of Gln before its entry into TCA cycle. Green color represents Gln's direct effect on canonical TCA cycle, red color represents Gln's effect on TCA cycle through reductive carboxylation.

4. F

   Gln's contribution to TCA cycle metabolites pool. Comparisons of MID of M5 glutamate, M4 fumarate, M4 malate, M4 citrate in OVCAR3 and SKOV3 cultured with ¹³C₅‐labeled Gln.

5. G

   Mass isotopomer analysis of isotopomers linked with reductive carboxylation. M5 citrate, M3 malate, M3 fumarate and M3 aspartate analysis indicates significantly higher reductive carboxylation fluxes in high‐invasive than in low‐invasive OVCA cells.

6. H

   Basal oxygen consumption rate (OCR) and extracellular acidification rate (ECAR) were measured for OVCAR3, OVCA420, SKOV3 and SKOV3ip cell lines. Basal OCR is a measure of OXPHOS activity, and basal ECAR is a measure of glycolysis activity.

7. I

   Using oligomycin, FCCP and antimycin, we estimated mitochondrial functional state in OVCA cells through maximal and reserve mitochondrial capacities.

8. J

   Using 2DG, we estimated glycolytic functional state in OVCA cells through maximal and reserve glycolytic capacities.

9. K

   Percentage of OXPHOS and glycolysis contribution to cancer cell's energetic demand.

10. L

    Relative metabolite abundances were measured using GC‐MS in OVCAR3 and SKOV3 cells.

11. M

    Extracellular uptake/secretion fluxes of amino acids involved in glutaminolysis (Glutamine: Gln; Glutamate: Glu; Alanine: Ala; Aspartate: Asp) were measured using ultra‐high‐performance liquid chromatography.

Data information: Data are expressed as mean ± SEM, n ≥ 9, *P < 0.05, **P < 0.01, ***P < 0.001. In (I) and (J), one‐way ANOVA with Tukey's test was used to compare between cell lines, using OVCAR3 as the control.

Figure 3. Glutamine has pleiotropic role of positively regulating respiration and maintaining redox balance selectively in high‐invasive ovarian cancer (OVCA) cells

1. Oxygen consumption rate (OCR) was measured in high‐invasive (SKOV3 and SKOV3ip), low‐invasive (IGROV1 and OVCAR3) and moderately invasive (OVCA420) cells in media containing Gln, aminooxyacetate (AOA) and α‐ketoglutarate (α‐KG). OCR was normalized with value before injections.

2. Analysis of proliferation of OVCA cell lines with range of concentrations of 6‐diazo‐5‐oxo‐L‐norleucine (L‐DON), a Gln analog which inhibits Gln's conversion to glutamate. Data were normalized with complete media conditions without L‐DON.

3. Analysis of proliferation of OVCA cell lines with range of concentrations of epigallocatechin gallate (EGCG, an inhibitor of glutamate dehydrogenase, which converts glutamate into α‐KG) used to inhibit glutamate entering the tricarboxylic acid (TCA) cycle. α‐KG was used to rescue reduction of cell proliferation by EGCG. Data were normalized with complete media conditions without EGCG.

4. Analysis of proliferation of OVCA cell lines when glutaminolysis was inhibited with range of concentrations of AOA. Data were normalized with complete media conditions without AOA.

5. Lactate and citrate generation from glucose through direct glycolysis or through both oxidative pentose phosphate pathway (PPP) and glycolysis. Black circles represent labeled carbons, empty circles represent unlabeled carbons, pink circles represent unlabeled carbons in oxidative PPP.

6. Lactate and citrate generation from glutaminolysis. Blue circles represent labeled carbons from malic enzymes, and green circles are labeled carbons from canonical TCA cycle.

7. Glucose deprivation effect on lactate secretion rate.

8. Comparison of M3 pyruvate and M3 lactate derived from either glucose (1:1 mixture of U‐¹³C₆ glucose and 1‐¹³C₁ glucose) or Gln (U‐¹³C₅ Gln) in high‐ and low‐invasive OVCA cells. Since all conditions have complete media, their total pyruvate and lactate content should be the same.

9. Comparison of M2 citrate labeling from glucose and M6 citrate labeling from Gln in OVCAR3 and SKOV3 cells. Total pyruvate and lactate content in all conditions should be equal as explained for (H).

10. Oxidative pentose phosphate pathway fluxes estimated using the mathematical relation (percentage of unlabeled M0 lactate—percentage of labeled M1 lactate) to represent relative oxidative PPP fluxes in OVCAR3 and SKOV3 cells cultured with 1:1 mixture of ¹³C₆ glucose and 1‐¹³C‐labeled glucose.

11. Role of Gln in maintaining mitochondrial membrane potential (MMP) levels in low‐ and high‐invasive cells.

12. Gln maintains ATP content selectively in high‐invasive OVCA cells.

13. Gln reduces reactive oxygen species induced by H₂O₂ in high‐invasive cells, but not in low‐invasive cells.

14. Gln maintains NADPH level in high‐invasive OVCA cells.

15. Gln increases ratio of NADPH/total NADPH ratio selectively in high‐invasive OVCA cells. Glucose is unable to provide enough reducing equivalents in high‐invasive OVCA cells.

16. Gln/glucose deprivation's effect on total glutathione and reduced glutathione level in cancer cells.

Data information: Data in (A–N) are expressed as mean ± SEM, n ≥ 9, **P < 0.01, ***P < 0.001. Data in (O, P) (n ≥ 6) are expressed as mean ± SEM, *P < 0.05, **P < 0.01, ***P < 0.001.

Figure 4. Regulating ratio of glutamine catabolism over anabolism reduces tumorigenicity

1. A

   Gene expression levels of Gln synthetase (GLUL) and glutaminase (GLS1) were determined using KyotoOv38, a database for gene expression of ovarian cancer cell lines.

2. B

   Western blot of the GLUL and GLS1 protein expression levels for OVCA cell lines.

3. C–E

   Progression free survival rate for OVCA patients, categorized according to GLS1/GLUL (C), GLUD1/GLUL (D), (GLS1 + GLUD1)/GLUL (E). The survival analysis is executed from comparison of upper quartile and lower quartile patients. (total patients n = 539).

4. F, G

   Converting Gln‐independent OVCA cells into Gln dependent by replacing glucose with galactose (F) or low glucose concentrations (L glucose) and 2‐DG (10 mM) (G).

5. H

   Galactose and low glucose's effect on oxygen consumption rate (OCR)/extracellular acidification rate.

6. I

   Combined approach of targeting high‐invasive OVCA cells by blocking glutamine's entry into the TCA cycle along with targeting low‐invasive OVCA cells by inhibiting glutamine synthesis leads to pronounced reduction of cell growth. Targeting GLUL using MSO under Gln deprivation decreases OVCA cell growth. Arbitrary Units (AU) is proportional to cell number (n ≥ 10).

Figure 5. Glutaminase activity is required for tumor growth and invasion in high‐invasive ovarian cancers (OVCAs)

1. A

   Kaplan–Meier curves of disease‐specific survival for patients with epithelial OVCA (n = 139) based on GLS1 and Gln synthetase protein expression. The log‐rank test (two sided) was used to compare difference between groups.

2. B

   Following transfection with either GLS1 siRNA or control siRNA, mRNA levels were assessed with qRT‐PCR.

3. C–H

   Therapeutic efficacy of siRNA‐mediated GLS downregulation: (C) tumor nodule (D) volume (E) pattern of invasion of low‐invasive OVCA cell line (IGROV1); (F) tumor nodule (G) volume and (H) pattern of invasion of SKOV3ip1 cells in nude mouse models. Following subcutaneous injection of nude mice with 2.0 × 10⁶ IGROV1 or SKOV3ip1 cells, mice were randomly allocated to one of the following groups: control siRNA‐DOPC or GLS1 siRNA‐DOPC. Treatment was started 3 days after tumor cell injection, and siRNA‐liposomes were administered twice weekly at a dose of 150 μg/kg body weight and continued for 2 weeks. At the time of sacrifice, mouse weight, tumor weight and tumor volume were recorded. Statistical analysis for tumor weights was performed by Student's t‐test. **P = 0.007 and **P = 0.008 compared with control siRNA‐DOPC.

Figure 6. Glutamine mediates oncogenic transformations in high‐invasive cells by regulating STAT3 activity

1. Comparison of OVCAR3 invasive capacity in Gln depleted and complete media conditions. Invasive capacity of SKOV3 is measured under complete media, Gln depleted media, and drugs inhibiting Gln's entry into tricarboxylic acid (TCA) cycle. L‐DON, BPTES, EGCG, AOA, and Rotenone decrease SKOV3's Matrigel invasive capacity. α‐Ketoglutarate (α‐KG) addition under glutamine deprivation or EGCG conditions rescues SKOV3's invasive capacity. Data are expressed as mean ± SEM, *P < 0.05, **P < 0.01, ***P < 0.001, n ≥ 4.

2. Activation of Stat3 through tyrosine‐705 phosphorylation (Stat3.pY705) is elevated in high‐invasive ovarian cancer (OVCA) cells. The phosphorylation level of EGFR and Erk 1/2 increases with increasing degree of invasiveness. Tyrosine kinase signaling pathway activities can be affected by metabolic stress and in high‐invasive cells that are glutamine and glucose deprived. Stat3.pY705 is reduced along with total Jak levels. Stat3.pY705 levels are only reduced in the OVCA420 cell lines upon glucose deprivation, while the regulation of Stat3 phosphorylation by metabolic stress is absent in the low‐invasive cell line, OVCAR3. Stat3 serine‐727 phosphorylation (Stat3.pS727) is also reduced along with phosphorylation level of Erk 1/2 in the high‐invasive cell lines and only in response to glutamine deprivation. β‐actin as loading control.

3. Cell proliferation in high‐invasive OVCA cell line is reduced when glutamine starved. Proliferation can be partially rescued by overexpression of transgenic Stat3 and a constitutively active mutant of Stat3, Stat3c, mean ± SD, n = 3, *P < 0.05. β‐actin as loading control.

4. α‐KG addition rescues STAT3 tyrosine phosphorylation and Jak1, but cannot rescue Stat3 serine‐727 phosphorylation. β‐actin as loading control. Western blot figures are cut from same gel, but lanes are rearranged to show complete media conditions in first column, glutamine deprivation in second and α‐KG addition in the third column. The full Western blot images of the gel and the detailed description are included in the source file.

5. OAA addition rescues STAT3 tyrosine phosphorylation. β‐actin as loading control. Western blot figures are cut from same gel, but lanes are rearranged to include only the relevant conditions. The full Western blot images of the gel and the detailed description are included in the source file.

6. BPTES and rotenone inhibit STAT3 phosphorylation at tyrosine 705 (Y705). β‐actin as loading control. Western blot figures are cut from same gel, but lanes are arranged to show only the relevant conditions. The full Western blot images of the gel and the detailed description are included in the source file.

7. Inhibition of Stat3 decreases OVCA metastasis. Treatment of SKOV3 cells with AG490 results in reduced invasion.

8. Gene expression levels of targets of STAT3 involved in invasion were determined using KyotoOv38 in five OVCA cells. High‐invasive OVCA cells had higher gene expression of invasive genes.

9. The addition of AG490, a Stat3 inhibitor enhances the effect of glucose and glutamine deprivation on proliferation in OVCA cell lines.

10. The addition of stattic, inhibitor of STAT3, has a combinational cytotoxic effect on both cell lines.

Data information: Data in (I) and (J) are expressed as mean ± SEM, n ≥ 8, *P < 0.05, **P < 0.01, ***P < 0.001. Source data are available online for this figure.

Figure 7. Glutamine's effect on metabolic rewiring in high‐invasive ovarian cancer (OVCA) cells is through STAT3 phosphorylation

1. A, B

   Effect of Gln deprivation on glucose uptake (A) and lactate secretion (B) fluxes in OVCAR3, OVCA420, SKOV3, SKOV3ip cells.

2. C, D

   Influence of AOA, α‐ketoglutarate (α‐KG) on glucose uptake (C) and lactate secretion (D) fluxes in OVCA cells.

3. E, F

   Effect of glucose deprivation on Gln, glutamate, aspartate and alanine uptake/secretion fluxes in high‐invasive (F) and low‐invasive (E) OVCA cells.

4. G

   AG490's effect on tyrosine and serine phosphorylation of STAT3 in low‐invasive (OVCAR3) and high‐invasive (SKOV3) OVCA cells.

5. H

   AG490 shifts low‐invasive OVCA cells from glycolysis to OXPHOS, whereas reverse was true for high‐invasive cells.

6. I

   AG490 increases oxygen consumption rate (OCR) for OVCAR3 and decreases OCR for SKOV3 cells.

7. J

   AG490 decreases extracellular acidification rate (ECAR) for OVCAR3 and increases ECAR for SKOV3 cells.

8. K, L

   AG490 decreases glucose uptake (K) and lactate secretion rate (L) for OVCAR3 and increases glucose uptake and lactate secretion rate for SKOV3.

Data information: Data in (A–D) are expressed as mean ± SEM, n ≥ 9. ***P < 0.001. Data in (E, F) are expressed as mean ± SEM, n = 6. *P < 0.05, **P < 0.01. Data in (H–L) are expressed as mean ± SEM, n ≥ 6, ***P < 0.001.

Figure 8. Glutamine's entry into tricarboxylic acid (TCA) cycle regulates ovarian cancer (OVCA) invasiveness

Schematic showing the shift in nutrient utilization in TCA cycle with increasing degree of invasiveness. Low‐invasive OVCA cells are glucose dependent for their TCA cycle pool. With increasing invasiveness in cancer cells, dominant nutrient which feeds the TCA cycle shifts from glucose to Gln. In high‐invasive OVCA cells, Gln dominates the TCA cycle. In low‐invasive OVCA cells, glucose activates Jak1, which activates STAT3 by tyrosine phosphorylation, thereby regulating glycolysis in cancer cells. In high‐invasive OVCA cells, besides glucose's role in activating STAT3 tyrosine phosphorylation, glutamine activates JAK1 through TCA cycle to further activate STAT3 by tyrosine phosphorylation and thus regulate glycolysis. Further, Gln activates Erk1/2, which subsequently activates STAT3 by serine phosphorylation selectively in high‐invasive OVCA cells. The serine phosphorylation of STAT3 enhances oxidative phosphorylation in mitochondria by interaction with mitochondrial complexes I and II, thereby increasing TCA cycle activity in high‐invasive OVCA cells.