Glucose deprivation in tuberous sclerosis complex-related tumors
© Jiang et al; licensee BioMed Central Ltd. 2011
Received: 3 June 2011
Accepted: 21 October 2011
Published: 21 October 2011
Skip to main content
© Jiang et al; licensee BioMed Central Ltd. 2011
Received: 3 June 2011
Accepted: 21 October 2011
Published: 21 October 2011
Cancer cells possess unique metabolic phenotypes that are determined by their underlying oncogenic pathways. Activation of the PI3K/Akt/mTOR signaling cascade promotes glycolysis and leads to glucose-dependence in tumors. In particular, cells with constitutive mTORC1 activity secondary to the loss of TSC1/TSC2 function are prone to undergo apoptosis upon glucose withdrawal in vitro, but this concept has not been tested in vivo. This study examines the effects of restricting glucose metabolism by pharmacologic and dietary means in a tuberous sclerosis complex (TSC) tumor xenograft model.
Tumor-bearing mice were randomly assigned to receive unrestricted carbohydrate-free ("Carb-free") or Western-style diet in the absence or presence of 2-deoxyglucose (2-DG) in one of four treatment groups. After 14 weeks, tumor sizes were significantly different among the four treatment groups with those receiving 2-DG having the smallest tumors. Unexpectedly, the "Carb-free" diet was associated with the largest tumors but they remained responsive to 2-DG. PET imaging showed significant treatment-related changes in tumor 18fluorodeoxyglucose-uptake but the standard uptake values did not correlate with tumor size. Alternative energy substrates such as ketone bodies and monounsaturated oleic acid supported the growth of the Tsc2-/- cells in vitro, whereas saturated palmitic acid was toxic. Correspondingly, tumors in the high-fat, "Carb-free" group showed greater necrosis and liquefaction that contributed to their larger sizes. In contrast, 2-DG treatment significantly reduced tumor cell proliferation, increased metabolic stress (i.e., ketonemia) and AMPK activity, whereas rapamycin primarily reduced cell size.
Our data support the concept of glycolytic inhibition as a therapeutic approach in TSC whereas dietary withdrawal of carbohydrates was not effective.
Tuberous sclerosis is an autosomal dominant disorder characterized by multiple benign hamartomas and neoplasms caused by the disruption of a pair of tumor suppressor genes, TSC1 and TSC2, which encode for hamartin and tuberin, respectively . Mutations and epigenetic silencing of these genes have been reported in sporadic human cancers including epithelial tumors of the bladder, liver, and oral cavity as well as PEComas [2–6]. The TSC1 and TSC2 proteins negatively regulate mTOR Complex 1 (mTORC1) by inhibiting Rheb activity . Consequently, mTORC1 is constitutively activated in cells lacking TSC1 or TSC2. These findings led to the use of rapamycin and its analogs in the treatment of TSC and related disorders [8–11]. The effect of rapamycin is cytostatic, and tumors re-grow upon cessation of treatment. Long-term rapamycin can cause significant side effects, thus alternative approaches are being investigated.
Oncogenic pathways such as PI3K/Akt and Myc promote aerobic glycolysis and glutaminolysis, respectively, to provide adequate supplies of ATP and substrates for macromolecular synthesis [12–15]. The dependence of tumor growth on these metabolic events provides a basis for metabolic intervention as a strategy for controlling tumors . In this study, we examined the in vivo role of glucose deprivation in TSC-related tumors. Cells lacking hamartin or tuberin are prone to undergo apoptosis under low-glucose condition [17, 18]. mTORC1 enhances aerobic glycolysis and lactate production via up-regulation of HIF1α . TSC1/TSC2-null cells also exhibit impaired insulin-stimulated glucose uptake secondary to Glut4 mislocalization . Pathology associated with TSC such as angiomyolipoma and lymphangioleiomyomatosis display low FDG uptake on PET imaging despite increased glycolytic activity [20, 21]. The imbalance between energy supply and demand presents a rationale for targeting glucose metabolism to control mTORC1-mediated tumorigenesis.
Two common approaches to limit glucose metabolism in tumors include utilization of glycolytic inhibitors and dietary restriction. Compounds such as 2-deoxyglucose (2-DG) reduce cellular ATP levels and promote apoptosis especially in cells with mitochondrial respiration defects or under hypoxic condition [22, 23]. Early clinical experience suggests that 2-DG is safe up to a dose of 250 mg/kg , but efficacy has not been well documented. Dietary restriction of glucose/carbohydrate (e.g., Atkins-type diets) leads to relative hypoglycemia, hypoinsulinemia and ketonemia in humans . Ketogenesis is an ancient pathway of metabolic adaptation exploited when an organism experiences protracted energy stress . Critical tissues such as the brain and kidney can efficiently metabolize ketone bodies, but it is unclear if tumor cells exhibit such adaptation.
Here, we studied the effects of 2-DG and a carbohydrate-free ("Carb-free") diet on the growth of mTORC1-dependent tumors using a Tsc2-/- xenograft model. The inhibition of glycolysis using 2-DG resulted in reduced cell proliferation and suppressed tumor growth, thus confirming the sensitivity of TSC-related tumors to metabolic intervention. On the other hand, the "Carb-free" diet failed to promote ketogenesis and led to increased tumor size despite reduction in body weights. Our study highlights the differential effects of glycolytic inhibition and dietary glucose deprivation in modulating tumor metabolism and growth.
Composition of the Western and "Carb-free" diets*.
AIN93G Mineral mix/fiber
AIN93 Vitamin mix/fiber
At time of sacrifice, tumor sizes were significantly different among the 4 groups (ANOVA, p < 0.05) (Figure 2B). In pair-wise comparisons, tumor sizes of the Western+2DG and the 'Carb'-free groups were significantly different (p < 0.05) while the 2-DG-treated tumors showed strong trends towards smaller size compared to each of the respective diets (p = 0.06) (Figure 2B). Tumor size at the end of 14 weeks and growth rate of tumors were highest in the "Carb-free" group (Figure 2B, C). Analyses of variance did not reveal a significant 'interaction' between the two treatments (i.e., 2-DG and diet). Thus, our initial hypothesis of superior tumor control through the combination of glycolytic inhibition and carbohydrate restriction is not supported by the data. Instead, we found that 2-DG suppressed tumor growth regardless of diet composition. Unexpectedly, carbohydrate restriction led to a promotion in tumor size despite a loss in overall body weight.
Recognizing that the prescribed treatments may impact tumor glucose metabolism, we performed FDG-PET on a subset of the 12 mice (3 per group) bearing tumors of sufficient size (i.e., > 5 mm) for detection with PET imaging. Following an overnight fast, mice were given 200 μCi of 18FDG intravenously and scanned after a 45-minute uptake period. FDG activity was normalized to body weight and time of FDG injection and expressed as standard uptake value (SUV). Figure 2E illustrates representative PET images from each of the 4 treatments showing significant heterogeneity. We used the maximal SUV value within each tumor to indicate the metabolic activity of the lesions. The mean SUVmax of the LEF2 tumors were significantly different among the 4 groups with the highest values found in the Western diet group and the lowest in the Carb-free + 2-DG group (Figure 2D). While carbohydrate restriction and 2-DG reduced FDG uptake, SUVmax did not correlate with tumor size (compare Figures 2B with2D). The paradoxical increase in tumor size with reduced tumor SUVmax in the "Carb-free" group suggests the possibility of alternative energy sources besides glucose that were utilized by the Tsc2-/- tumor cells.
Finally, we compared the relative activities of the Akt/mTOR and AMPK pathways in LEF2 tumors following the four treatments. Tumor lysates uniformly showed high levels of phospho-S6 indicative of mTORC1 activation and low levels of phospho-Akt secondary to feedback inhibition  with no consistent diet/treatment-related effects (Figure 7D). Thus, the anti-tumoral effects of 2-DG was not secondary to mTORC1 inhibition. AMPK phosphorylation was most pronounced following exposure to the "Western diet+2-DG" treatment in keeping with its associated metabolic stress (e.g., ketonemia) and our in vitro findings (see Figure 1D, 4). However, this did not correlate with changes in the levels of phospho-S6, p27, or cyclin D1. Further, there was no consistent alterations in p62 or phospho-eIF2α expression in tumor lysates to suggest induction of autophagy or ER stress, respectively. In contrast, rapamycin abolished S6 phosphorylation, increased Akt phosphorylation and reduced p62 expression to an equal extent in tumors from the "Western" and "Carb-free"-diet groups (Figure 7E); this is consistent with the known effects of rapamycin in promoting autophagy .
We investigated two approaches aimed at restricting glucose metabolism in the treatment of TSC-related tumors. The rationale of our study is based on the observations that the loss of TSC1/TSC2 leads to an energy imbalance caused by mTORC1 hyperactivity that increases energy demand stemming from macromolecular synthesis and reduces energy supply as a result of impaired insulin-stimulated glucose uptake [19, 20]. Consequently, cells deficient in TSC1 or TSC2 are prone to undergo apoptosis upon glucose withdrawal or 2-DG treatment . The combination of low glucose concentration and 2-DG was more effective in limiting ATP, activating AMPK and inducing cell death in vitro (Figure 1). When applied to an in vivo model of Tsc2-null tumor, 2-DG suppressed tumor growth by reducing cell proliferation whereas a diet free of carbohydrate resulted in larger tumors with increased zones of liquefaction. We did not detect an 'additive' effect when the two treatments were given together. These findings indicate that the Tsc2-related tumors are sensitive to glycolytic inhibition in vivo, but the lack of tumor suppression following dietary restriction of carbohydrate could be a reflection of its inability to induce significant hypoglycemia. On the other hand, the increased necrosis in the "Carb-free" fed tumors was not due to an effect of the diet on tumor vasculature, ER stress or autophagy since the tumor expression of CD34, phospho-eIF2α(Ser51) and p62 were not significantly different among the 4 groups.
The ketogenic diet has been investigated for over two decades as a treatment for malignant brain tumors based on the notion that tumor metabolism favors aerobic glycolysis (i.e., Warburg effect) whereas tissues of critical organs such as the brain can utilize ketone bodies efficiently . In a short-term study, Marsh et al. reported a synergistic effect of 2-DG and a carbohydrate-free diet on inhibiting astrocytoma growth in a mouse model. They attributed the benefit of the diet to energy restriction resulting in ketogenesis and significant weight loss . When diet was given without restriction, its composition did not appear to significantly impact tumor growth in vivo . In our study, mice receiving the "Carb-free" diet weighed less than the Western diet group, but the difference was only ~10%, and the β-hydroxybutyrate levels did not rise significantly. Thus, unrestricted "Carb-free" diet per se did not create a state of metabolic "stress" in mice. In addition to the lack of caloric restriction, our "Carb-free" diet contained high protein content that can serve as an energy source via gluconeogenesis. Mavropoulos et al. also failed to show a significant anti-tumoral effect of the no-carbohydrate ketogenic diet compared with a Western diet in a prostate cancer xenograft model despite a decrease in serum insulin, IGF-1 and phospho-Akt levels . Interestingly, the ketogenic diet used in controlling seizures has been associated with disease progression in 3 of 5 human TSC patients . Together, these findings should raise awareness for more vigilant monitoring of TSC-related pathology in patients receiving the ketogenic diet for seizure management. At present, the effects of caloric restriction in TSC-related tumors are not known, but we surmise the possibility that a caloric-restricted diet with reduced glucose (exogenous or endogenous) availability may synergize with 2-DG to limit tumor growth.
Inhibition of glycolysis by 2-DG is currently being tested in clinical trials. 2-DG causes a depletion of cellular ATP that leads to the activation of AMPK. In combination with a Western diet, 2-DG treatment had the most severe metabolic stress in terms of hypoglycemia and ketogenesis. Consequently, we encountered a greater degree of toxicity with 3 of 10 mice experiencing significant weight loss before the completion of the experiment. This metabolic consequence was not observed in the mice receiving 2-DG while fed a "Carb-free" diet. Significant weight loss occurs when glycolysis is inhibited in which carbohydrate constitutes a significant source of energy as in the Western diet. When glucose becomes limiting, the body mobilizes fat stores for energy (i.e., lipolysis) resulting in loss of subcutaneous fat and weight, which is what we observed in a subset of the Western+2DG group. In contrast, the "Carb-free" diet provides abundant fat and protein as energy substrates, thus limiting the severity of weight loss. With respect to the Western diet+2DG treatment, AMPK(Thr172) phosphorylation was increased under in vitro and in vivo conditions, but we did not detect significant suppression of mTORC1 activity or changes in p27 and cyclin D1 expression. Therefore, the mechanism of 2-DG inhibition on cell proliferation in vivo remains undefined. Compared to rapamycin-treated samples, two distinct patterns of response emerged with 2-DG slowing proliferation and rapamycin reducing cell size. However, it is unlikely that the combination of 2-DG and rapamycin would be beneficial since the effects of mTORC1 inhibition would neutralize the glucose-dependence of the Tsc2-null cells. Indeed, Inoki et al. have shown that rapamycin and re-expression of Tsc2 in LEF2 cells rescued them from apoptosis under low-glucose condition .
Besides glucose, we found that ketone bodies can be utilized by cells in vitro to promote survival and growth although these effects were mild (Figure 3). In contrast, Fine et al. reported that ketone bodies inhibit cell growth by reducing ATP concentration in tumor cells that over-express UCP2 . We also showed that cell proliferation can be augmented by oleic acids in culture despite a reduction in cellular ATP, but saturated palmitic acid was toxic to cells. These findings are in keeping with those reported by Ricchi et al. in hepatocytes . Considered together, the Tsc2- tumor cells exhibit metabolic plasticity beyond their dependence on glycolysis. This may explain the lack of correlation between FDG uptake and tumor response. Further, glutamine has also been identified as an important nutrient in supporting cellular bioenergetics of the Tsc2-null cells . In future studies, targeting multiple metabolic pathways may be considered to maximize tumor control while minimizing toxicity.
Our study provides the first in vivo evidence demonstrating anti-tumoral effects of glycolytic inhibition in TSC2-related tumors. The predominant effect of 2-DG was to inhibit cell proliferation in contrast to the effect of rapamycin on cell size. A diet free of carbohydrate without caloric restriction was not effective in controlling TSC2-tumor growth, but changed the consistency of the tumors by inducing necrosis.
LEF2 cells were derived from an Eker rat renal tumor (Tsc2-/-), and the wild-type rat kidney-derived cells, RK3E, were purchased from ATCC. Both were cultured in DMEM/F12 medium containing 10% FBS fetal bovine serum (FBS), 100 units/ml penicillin G and 100 μg/ml streptomycin sulfate at 37°C in a humidified 5% CO2 incubator. Antibodies were purchased from the following sources: anti-BrdU (Dako), anti-CD34 (Cedarlane, Burlington, NC), anti-p62 and anti-α-tubulin (Sigma, St. Louis, MO). All remaining antibodies (anti-AMPK, anti-phospho-AMPK(Thr172), anti-S6, anti-phospho-S6, anti-p27, anti-cyclin D, anti-eIF2α, anti-Akt, anti-phospho-Akt, anti-cleaved PARP, and anti-cleaved-caspase 3) were from Cell Signaling (Danvers, MA). 2-deoxyglucose, oleic and palmitic acids were purchased from Sigma (St. Louis, MO). Rapamycin were obtained from EMD Biosciences (San Diego, CA).
Young (5-6 weeks), male, NOD.CB17-prkdcscid/J mice were purchased from JAX and fed regular chow for 2-3 weeks. At 8 weeks of age, animals were randomly allocated to 4 treatment groups: Western diet, Western+0.02% 2-DG, 'Carb-free' diet, 'Carb-free'+0.02% 2-DG. The diets with and without 2-DG were prepared and purchased from Animal Specialties (Hubbard, OR). After 4 days on treatment, 3 × 106 LEF2 cells were injected subcutaneously in the flank region of the mice. Animals were monitored for tumor growth and general health for 14 weeks. After an overnight fast, mice were sacrificed, tumors and sera were collected for analyses. For a subset of 12 animals, FDG-PET was performed (see below) and BrdU (50 μg/g) was injected intra-peritoneally once daily for two days before sacrifice. All work related to animals was in accordance with a protocol approved by the Institutional Animal Care Committee, University of Washington, Seattle.
Following an overnight fast, mice were anesthetized using isoflurane and given a retro-orbital injection of FDG (200 μCi). Mice were kept warm and anesthetized during a 45-minute uptake period. Following this, mice were imaged for 20 minutes on a Siemens Inveon Dedicated PET system at the micro-PET imaging facility (University of Washington). A transmission scan was taken of each mouse after the emission study. Images were reconstructed using the manufacturer supplied 3D OSEM MAP image reconstruction. Data were collected with both random, scatter and attenuation correction. A Beta smoothing parameter of 0.1 was used. The maximum uptake value for each tumor was captured from the reconstructed images by stepping through the tumor and measuring the maximum activity (nCi/cc) in all hotpots and was normalized to body weight and time of FDG injection (to account for decay) in order to calculate the standard uptake value (SUV).
Quantification of viable cells was performed using the trypan blue exclusion method. In brief, cells were harvested and collected following trypsin detachment and centrifugation. Cells were rinsed with PBS and then resuspended in 1 ml of PBS. A 10-μl aliquot of cell suspension was incubated with 10 μl 0.4% trypan blue stain (Invitrogen/Gibco) solution for 5 minutes at room temperature. Viable and nonviable cells based on absence and presence of intracellular dye, respectively, were counted by hemacytometer. Results represent the number of nonviable cells divided by the total number of cells counted and are expressed as percentages.
Cell apoptosis was measured by flow cytometry using the Annexin V kit (AbD Serotec, Raleigh, NC). In brief, RK3E and LEF2 cells were detached following treatment with 0.05% trypsin-EDTA for 10 minutes at 37°C, washed with PBS, pelleted by centrifugation, resuspended with pre-diluted binding buffer, and stained with Annexin V-FITC for 10 minutes in the dark at room temperature. Cells were washed and resuspended in binding buffer with Propidium lodide solution and analyzed by flow cytometry.
Cells were seeded on plates with 10% FBS with DMEM/F12 medium for overnight, then washed with PBS and replaced with 10% dialyzed FBS DMEM with 2 mM glucose, 1 mM HEPES. After 1 day in culture medium, cells were treated with 4 mM 2-Deoxyglucose, 1 mg/ml acetoacetate, or 2.5 mg/ml β-hydroxybutyrate for 6 hours. ATP levels were measured using a Luminescence ATP detection assay system (PerkinElmer, MA). Briefly, cells were lysed for 5 minutes with supplied lysis buffer in an orbital shaker at 700 rpm, then ATP detection substrate solution was added to the wells and agitated for 5 minutes. Plates were left in the dark for 10 minutes and luminescence was measured using a Molecular Devices Spectra Max M2 (Sunnyvale CA).
For fatty acid experiments, oleic acid-BSA and palmitic acid-BSA solutions were pre-made (2 moles of FFA to 1 mole BSA). Cells were washed with PBS and cultured in 10% charcoal striped FBS phenol-free DMEM with 1 mM Oleic acid-BSA, 1 mM palmitic acid-BSA, or 2 to 1 molar of oleic acid/palmitic acid (3 mM) for 24 hours.
Cells were grown on 24-well plates in 10% FBS DMEM/F12, treated with ketone bodies (1 mg/ml acetoacetate or 2.5 mg/ml β-hydroxybutyrate for 24 hours) or FFA (1 mM oleic acid, 1 mM palmitic acid or 3 mM oleic acid + palmitic acid for 24 hours) followed by MTT proliferation assay. Briefly, 100 μl of 5 mg/ml MTT was added to each well and incubated for 3.5 hours. 750 μl of MTT solvent (4 mM HCl, 0.1% Nondet P-40 in isopropanol) was added to each well and agitated for 15 minutes. Optical density was read at 600 nm using a Packard Spectracount microplate reader (PerkinElmer, Inc., Waltham, MA).
Blood was extracted via cardiac puncture immediately after sacrifice. Blood was spun for 15 minutes at 3000 rpm at 4°C. Plasma was analyzed for glucose, insulin, β-hydroxybutyrate, and triglycerides. Plasma glucose levels were measured colorimetrically using the glucose oxidase reagent (Pointe Scientific, Canton, MI). Plasma insulin levels were measured using the Linco ELISA (Millipore, Billerica, MA). β-hydroxybutyrate in plasma was measured using the KetoSite β-hydroxybutyrate dehydrogenase assay according to the manufacturer's instructions (Stanbio Laboratory, Boerne, TX). Plasma triglycerides were quantified via the Wako L-Type TG M colorimetric assay (Wako Diagnostics, Richmond, VA).
Portions of tumor were fixed in either neutral buffered formalin for routine histology and immunohistochemistry (IHC) or in methacarn (60% methanol, 30% chloroform, and 10% acetic acid) for bromo-deoxyuridine (BrdU) labeling. After fixation, tissue was embedded in paraffin. For routine histological analysis, 5-μm sections were cut from paraffin-embedded blocks and sections were deparaffinized, rehydrated, and washed before staining with hematoxylin QS and eosin (Vector Laboratories, Burlingame, CA) and mounted with Permount (Fischer Scientific, Santa Clara, CA). For CD34 staining, sections were deparaffinized, rehydrated, and washed with phosphate-buffered saline. After antigen retrieval in 0.1 M sodium citrate (pH 6.0) and quenching of endogenous peroxidase activity with 3% H2O2, samples were blocked with 5% normal horse serum (NHS) before incubation with primary antibodies overnight at 4°C. Negative controls were treated with 5% NHS without primary antibodies. Staining was detected using the Elite Vectastain ABC kit (Vector laboratories, Burlingame, CA) according to the manufacturers instructions. BrdU IHC was performed as described above with the following modifications: endogenous peroxidase activity was quenched with 0.3% H2O2/Methanol for 10 minutes and antigen retrieval consisted of trypsin digestion (1 mg/ml trypsin for 10 min at room temperature), followed by incubation in 2.5 M HCl for 10 min at 37°C. Mouse anti-BrdU diluted 1:40 was used as the primary antibody (Dako, Carpinteria, CA, USA).
Cells and tumor tissues were homogenized in ice-cold radioummunoprecipitation (RIPA) buffer (1% Nonidet P-40, 1% sodium deoxycholate, 0.1% SDS, 0.15 M NaCl, 10 mM Tris (pH 7.2), 0.025 M β-glycophosphate (pH 7.2), 2 mM EDTA, and 50 mM sodium fluoride) with protease and kinase inhibitors (0.05 mM AEBSF, 10 μg/ml aprotinin, 10 μg/ml pepstatin, 1 mM orthovanadate, 10 μg/ml leupeptin, 1 mM microcystin LR). The protein concentration was measured using the BCA Protein Assay (Pierce, Rockford, IL). Equal amounts of protein were separated by SDS-PAGE, transferred to Immobilin-P membranes (Millipore, Bedford, MA) and blotted with antibodies according to manufacturer recommendations.
Comparisons between two groups were analyzed using the Student t-test. For multi-group comparisons, analysis of variance (ANOVA) was used.
tuberous sclerosis complex
mammalian target of rapamycin complex 1
adenosine monophosphate-activated protein kinase
positron emission tomography
standard uptake value.
We thank Dr. Robert Miyaoka and staff at the Animal PET facility for conducting the FDG-PET studies, and the Mouse Metabolic Phenotyping Center (UW) for serum-based assays. This work was supported by grants from National Institutes of Health (R01 CA077882) and the Tuberous Sclerosis Alliance.
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.