Differential effects of energy stress on AMPK phosphorylation and apoptosis in experimental brain tumor and normal brain
© Mukherjee et al. 2008
Received: 18 November 2007
Accepted: 12 May 2008
Published: 12 May 2008
AMP-activated protein kinase (AMPK) is a known physiological cellular energy sensor and becomes phosphorylated at Thr-172 in response to changes in cellular ATP levels. Activated AMPK acts as either an inducer or suppressor of apoptosis depending on the severity of energy stress and the presence or absence of certain functional tumor suppressor genes.
Here we show that energy stress differentially affects AMPK phosphorylation and cell-death in brain tumor tissue and in tissue from contra-lateral normal brain. We compared TSC2 deficient CT-2A mouse astrocytoma cells with syngeneic normal astrocytes that were grown under identical condition in vitro. Energy stress induced by glucose withdrawal or addition of 2-deoxyglucose caused more ATP depletion, AMPK phosphorylation and apoptosis in CT-2A cells than in the normal astrocytes. Under normal energy conditions pharmacological stimulation of AMPK caused apoptosis in CT-2A cells but not in astrocytes. TSC2 siRNA treated astrocytes are hypersensitive to apoptosis induced by energy stress compared to control cells. AMPK phosphorylation and apoptosis were also greater in the CT-2A tumor tissue than in the normal brain tissue following implementation of dietary energy restriction. Inefficient mTOR and TSC2 signaling, downstream of AMPK, is responsible for CT-2A cell-death, while functional LKB1 may protect normal brain cells under energy stress.
Together these data demonstrates that AMPK phosphorylation induces apoptosis in mouse astrocytoma but may protect normal brain cells from apoptosis under similar energy stress condition. Therefore, using activator of AMPK along with glycolysis inhibitor could be a potential therapeutic approach for TSC2 deficient human malignant astrocytoma.
List of abbreviations
AMP-activated protein kinase
AMP-activated protein kinase kinase
mammalian target of rapamycin
Dulbecco's modified Eagle's medium
terminal deoxynucleotidyl transferase-mediated dUTP nick end labeling
AMP-activated protein kinase (AMPK) is a primary regulator of the cellular response to lowered ATP levels in eukaryotic cells [1, 2]. AMPK is a serine/threonine protein kinase and a member of the Snf1/AMPK protein kinase family . The activity of AMPK requires phosphorylation of the alpha subunit on Thr-172 in its activation loop by one or more upstream kinases (AMPKK) [3–5]. AMPK phosphorylation down regulates ATP consuming processes like the synthesis of fatty acids, cholesterol, and proteins, while up-regulating ATP producing catabolic pathways like fatty acid oxidation and glucose uptake. Previous studies indicate that AMPK has both pro and anti-apoptotic effects on eukaryotic cells [6–16]. The pro-apoptotic action of activated AMPK is associated with activation of stress kinases and caspase-3 [7–9, 13, 17]. In primary b cells, prolonged stimulation of AMPK induces apoptosis through an activation of c-Jun-N-terminal kinase (JNK) . On the other hand, pharmacological activation of AMPK protects thymocytes from dexamethasone-induced apoptosis  and Rat-1 fibroblasts from serum withdrawal induced apoptosis .
AMPK is an anti-growth molecule because of its relationship with two tumor suppressor genes: LKB and TSC2 (tuberous sclerosis complex 2). LKB mutations cause Peutz-Jeghers syndrome, an autosomal dominant disorder characterized by multiple hamartomatous polyps in colon and other parts of the gastrointestinal tract [18–20]. LKB1 is the upstream activating kinase for the stress-responsive AMP-activated kinase, and provides a link between regulators of cellular metabolism and cell proliferation in cancer . There are several reports that LKB1 activates AMPK and thus serves as the principal AMPKK [6, 22, 23]. LKB1 protects cells from apoptosis in response to agents that elevate intracellular AMP. LKB1-deficient mouse embryonic fibroblasts are defective in AMPK activation and undergo apoptosis under conditions that elevate AMP. Recent observations indicate that Ca2+/calmodulin – dependent protein kinase kinases also regulate AMPK in cell lines lacking expression of LKB1 [24, 25].
Under energy stress, AMPK phosphorylates TSC2 on T-1227 and S-1345 which regulate cell size . Activation of TSC2 by AMPK- dependent phosphorylation prepares cells for an unfavorable growth environment and protects cells from death. Glucose deprivation induces apoptosis in TSC2-/- and TSC1-/- cells . Since mTOR is a key downstream target of TSC1/TSC2, inhibition of mTOR by rapamycin suppresses apoptosis in those cells following energy depletion . LKB1 may activate TSC2 via the AMPK [26, 27]. mTOR inhibitors have anti-neoplastic potential since mutations in LKB1, TSC2, or PTEN tumor suppressor genes produce aberrant mTOR activation in certain neoplastic conditions [28–31].
Dagon et al  recently found that a 40% dietary restriction increased hippocampal AMPK activity, induced neurogenesis, and improved cognition. On the other hand, a 60% dietary restriction over activated AMPK, reduced cognition, and induced neural apoptosis. We show here that phosphorylation of AMPK at Thr-172 produced apoptosis in TSC2 deficient CT-2A mouse astrocytoma under 40% dietary caloric restriction, whereas less AMPK phosphorylation with no apoptosis was seen in contra-lateral normal brain. This finding was further supported by in vitro observations where energy stress produced differential phosphorylation of AMPK in CT-2A cells and in astrocytes and where phosphorylation of AMPK enhanced apoptosis in CT-2A cells.
Differential effects of energy stress on ATP depletion, AMPK activation and apoptosis in CT-2A tumor cells and normal astrocytes in vitro
To determine the number of apoptotic cells, we stained the cells with Annexin-V, which specifically labels apoptotic cells. Figure 1c shows that glucose withdrawal induced apoptosis in 48% of CT-2A cells at 18 hours. A significant number of astrocytes were found Annexin and PI positive in the glucose containing medium which could be due to their high basal turnover rate and confluency. Glucose withdrawal had no significant effect on apoptosis in astrocytes. The increased number of living astrocytes was likely due to decrease in their proliferation rate in glucose free media. Caspase-3 activation in CT-2A cells was preceded by upstream caspase-9 cleavage after 6 to 8 hours of glucose and glutamine deprived conditions (Figure 1d). Caspase-9 cleavage was associated with cytochrome c release in the cytosolic part of CT-2A cells in 8 hours of glucose/glutamine free conditions (figure 1e). Caspase-9 cleavage and cytochrome c was not detected in astrocytes in the indicated time period (data not shown). Treatment of CT-2A cells with 3.0 mM glucose caused a marked elevation of AMPK phosphorylation in 18 hours, whereas astrocyte was stressed in low glucose condition in 48 hours (Figure 1f). CT-2A cells usually die after 48 hours with 3.0 mM glucose. Treatment of astrocytes with 3.0 mM glucose caused no AMPK phosphorylation at Thr-172 in 18 hours. AMPK phosphorylation was observed in both CT-2A and astrocytes following treatment with 2-DG, which inhibits glucose utilization (Figure 1g and 1h). In CT-2A, phosphorylation was significantly greater than in astrocytes following 2-DG treatment. These results indicate that reduced energy substrates produced significantly greater stress and apoptosis in CT-2A tumor cells than in normal astrocytes.
AMPK stimulation caused apoptosis in CT-2A cells
Differential effects of energy stress on AMPK activation and apoptosis in CT-2A brain tumor and normal brain in vivo
Effects of energy stress on the mTOR/S6K protein synthesis pathway in CT-2A tumor cells and astrocytes
Role of energy stress on LKB1 expression in normal brain
Pharmacological stimulation and inhibition of AMPK in astrocytes and CT-2A cells
TSC2 expression in CT-2A and astrocytes in vitro and in vivo: TSC2 is required for the protection of astrocytes from apoptosis under energy stress
Our results indicate that the response of AMPK to energy stress in brain tumors is fundamentally different from the response to a similar energy stress in the neural parenchyma. Because CT-2A cells are more highly glycolytic than normal astrocytes, the rate of ATP depletion and level of AMPK activation was greater in CT-2A cells than in normal astrocytes. Although the pro-apoptotic and the anti-apoptotic effects of AMPK are well documented, our current findings provide novel evidence that activation of AMPK produces apoptosis in brain tumor tissue and tumor cells, but prevents apoptosis in normal brain tissue and astrocytes. Our findings also suggest that LKB1, an up-stream activator of AMPK, may play a role in the anti-apoptotic effects of AMPK in normal brain under energy stress.
The coordination of cellular energy status and survival is currently a major field of interest in cell biology. Although both normal brain tissue and brain tumors are highly dependent on glucose for survival, brain tumors are more susceptible to the effects of glucose deprivation than normal brain tissue . AMPK monitors the energy status of a cell and can initiate appropriate responses to ATP depletion during energy metabolic stress [2, 33]. Our current findings indicate that AMPK phosphorylation is significantly greater in CT-2A astrocytoma cells than in normal astrocytes after 24 hr of glucose deprivation. The enhanced AMPK phosphorylation is associated with death in the CT-2A cells, but not in the astrocytes under normal energy condition. These findings are also consistent with our in vivo observations where a 40% dietary restriction, which lowers blood glucose levels, caused a significant elevation of AMPK phosphorylation in both tumor and in the brain. AMPK phosphorylation was associated with caspase 3 cleavage in CT-2A brain tumor, but not in contra-lateral normal appearing brain tissue.
Previous studies showed that AMPK activation with either AICAR or low glucose causes apoptosis in cultured rat b cells or in insulin producing MIN6 cells [13, 14]. The pro-apoptotic effects of AMPK activation in various cancer cells includes inhibition of fatty acid synthase and induction of stress kinase and caspase-3 [7–9, 14, 17]. On the other hand, AMPK activation also had a protective effect on stress injured cells in heart ischemia and reperfusion injury [34, 11, 35]. In contrast to CT-2A cells under energy stress where AMPK activation was associated with enhanced apoptosis, we found that AMPK activation protected normal astrocytes form apoptosis under energy stress. This came from findings that compound C, an inhibitor of AMPK activation, enhanced astrocyte apoptosis under energy stress. On the other hand, stimulation of astrocytes with AMPK stimulator significantly increased the number of astrocytes under very low glucose conditions. A linkage between the LKB1 and the TSC2 tumor suppressors may account in part for the apparent opposite effects of AMPK activation on cell survival and cell-death.
LKB1 is an upstream AMPKK which phosphorylates and activates AMPK under normal physiological conditions [6, 36]. Mutations of LKB1 occur in Peutz-Jeghers syndrome and enhance susceptibility to tumor formation [18–20]. Shaw et al  found that LKB1 deficient cells are hypersensitive to apoptosis induced by energy stress. On the other hand, it is essential to protect cells from apoptosis in response to agents that elevate intracellular AMP. The loss of LKB1 in tumors can result in increased cell growth, but cells lacking LKB1 are resistant to transformation and readily undergo apoptosis under energy stress condition . We found that LKB1 is not expressed in CT-2A tumor tissue, but is expressed in normal brain tissue contra-lateral to the tumor. We also found that 40% dietary restriction caused a significant elevation of LKB1 expression in normal brain tissue. The lack of LKB1 in CT-2A tumor cells might explain the aberrant growth of these cells under in vivo and in vitro conditions. Since LKB1 is absent in CT-2A tumor, other AMPKKs could be responsible for the stimulation of AMPK under energy stress. Hurley et al  investigated other AMPKK in three LKB1 deficient cancer cell lines and found that Ca2+/calmodulin dependent protein kinase kinases regulated AMPK activity in those cell lines. Rattan et al  found that AMPK can be a potential target for treatment of various cancers independent of the functional tumor suppressor genes, LKB1. In our findings the presence of LKB1 in normal brain and induction of apoptosis by inhibiting AMPK in normal astrocyte potentially suggest that LKB1/AMPK may play a role in their survival under energy stress.
Cell survival under energy stress is dependent in part on the ability to conserve energy through inhibition of protein synthesis. Protein synthesis utilizes approximately 20%–25% of the total cellular energy and is coordinated with cellular energy status . Activation of AMPK inhibits protein synthesis by suppressing the functions of multiple translation regulators including S6K, 4E-BP1, and eEF2 in response to energy starvation [26, 39]. The mechanism of S6K inhibition by AMPK involves the mammalian target of rapamycin (mTOR) pathway [27, 40–43]. We found that 2-DG inhibited both mTOR and S6K phosphorylation in normal astrocytes, but not in CT-2A cells. Furthermore, rapamycin treatment suppressed glucose deprivation induced cell-death in CT-2A cells suggesting that high mTOR activity is responsible for cell-death under energy starvation condition. Our results support the previous findings of Inoki et al  who found that rapamycin treatment suppressed glucose-deprivation induced cell-death in TSC2 -/- epithelial cells.
Like LKB1, TSC2 is another tumor suppressor gene that influences protein synthesis downstream of AMPkinase and upstream of mTOR [26, 44, 45]. Inoki et al  proposed that phosphorylation of TSC2 by AMPK protect cells from energy deprivation induced cell-death. In our study, TSC2 knockdown astrocytes are hypersensitive to apoptosis under energy stress condition. Absence of both LKB1 and TSC2 in CT-2A cells well explain the uncontrolled mTOR activity and lack of protection from cell-death under severe energy stress. Shaw et al  also found that LKB1 is required for repression of mTOR under low ATP conditions in cultured cells in an AMPK and TSC2 dependent manner and LKB1 mutant mice show elevated signaling downstream of mTOR. These observations together with our findings suggest that the LKB1-AMPK-TSC2-mTOR pathway underlies the adaptive ability of normal cells to conserve energy when food is scarce.
We previously showed that moderate CR (40%) enhanced apoptosis in experimental mouse and human brain tumors while enhancing the overall health and vitality of the tumor-bearing mice . Though the pro-apoptotic effects of CR in brain tumors occur in large part from reduced glycolytic energy that tumors rely upon for growth, its effects on normal brain were not clear. Also, the molecular mechanisms involved in the global therapeutic approach of CR in brain tumors were unknown. We also proposed that alternative energy fuels like ketones for the protection of normal brain under energy stress condition . The bioenergetic transition from glucose to ketones during CR might have a direct influence on AMPK activation and cell survival for the normal brain.
In conclusion, we provide evidence for a dual role of AMPK activation in the protection of normal brain from energy stress while also killing the brain tumor cells. Further studies are required to better define the role of AMPK as a potential regulator of energy metabolism and apoptosis in normal and tumor tissues.
Phosphorylated and non – phosphorylated AMPK(Thr172), mTOR(Ser2448), pS6 kinase (Thr389), caspase-9, caspase-3, PARP, cytochrome-c antibodies, lysis buffer (10×), and rapamycin (mTOR inhibitor) were purchased from Cell Signaling, MA. Anti LKB1 was purchased from SantaCruz Biotechnology, CA. β-actin was purchased from Novus Biologicals, CO. 2 deoxyglucose (2-DG) from Sigma, MO, AICAR (AMPK stimulator), compound C (AMPK inhibitor) was from Calbiochem, CA. Calcein was purchased from Molecular Probe (Carlsbad, CA).
Cell culture conditions
The CT-2A, mouse astrocytoma cell line was established as previously described . The C8-D1A, astrocyte cell line was purchased from American Type Culture Collection (ATCC), VA. Both CT-2A tumor and astrocyte cell lines were originated from C57BL/6J mice. Cell lines were maintained in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal bovine serum and 0.5% penicillin/streptomycin (Sigma, St. Louis, MO). The cells were cultured in a CO2 incubator with a humidified atmosphere containing 95% air and 5% CO2 at 37°C. For ATP depletion, the cells were stimulated with 2-DG and different amount of glucose and glutamine containing media. For 2-DG treatment, the culture media was replaced with serum free DMEM containing 6 mM glucose 24 hours before 2-DG stimulation. The cells were then stimulated with 50 mM of 2-DG that was dissolved in serum free DMEM for indicated times. For glucose starvation, the cells were pretreated with serum free DMEM for 24 hours and were then cultured with 6.0 mM glucose and/or 4.0 mM glutamine and, 0 mM glucose and 0 mM glutamine in serum free DMEM for 24 hours. Additionally, 3.0 mM glucose and 2.0 mM glutamine, 1.0 mM glucose and 0.66 mM of glutamine in serum free DMEM were used for indicated time.
For rapamycin treatment, CT-2A cells were cultured with serum free media for 24 hours and were then incubated with 3.0 mM glucose containing media in the presence or absence of 20 nM rapamycin for 24 hours. For compound C treatment, astrocytes were cultured in 3.0 mM glucose containing media in the presence or absence of 1.0 μM compound C for 24 and 48 hours. For AICAR treatment, CT-2A and astrocytes were cultured in 6.0 mM glucose containing media for 24 hours and in 1.0 mM glucose containing for 72 hours in the presence and absence of 0.5 and 1.0 mM AICAR.
To harvest cells, the flasks were washed once with phosphate buffered saline (PBS) and treated with trypsin-EDTA for 2 minutes. Cells were gently pipetted off the flask with PBS, transferred to 15 ml conical centrifuge tubes, centrifuged for 3 minutes at 1000 rpm, and pellets were washed again with PBS. Final pelletes were lysed with lysis buffer and were centrifuged at 8,100 × g for 20 min at 4°C. The lysates were collected and stored at -80°C for protein analysis. For caspase analysis, both the floating and the adhered cells were collected and lysed. ATP determination – Five thousand cells per well were seeded in 96 wells plate and allowed to adhere for 24 hours. Cells were washed with serum free media and replaced with fresh media with 6.0 mM glucose and without glucose. ATP levels were measured using the ViaLight Plus kit (Cambrex, Charlescity, IA) that is based on the bioluminescent measurement of ATP in presence of luciferase, Mg2+ and O2. Luminescence was measured with a microplate luminometer with an integration time of 5s/well.
siRNA against the mouse TSC2 (mouse) was bought from SantaCruz and they were transfected using Lipofectamine 2000 (Invitrogen, CA). Transfections were carried out using the concentrations of 50 and 100 nM for 24 and 48 hours. For each experiment astrocytes were grown for 36–48 hours with siRNA, transfected and control astrocytes were incubated for additional 24 hours in the presence and absence of glucose and glutamine and cell lysates were collected for western blot.
Western Blot Analysis
Tumor cells and tissues were homogenized in ice-cold lysis buffer (Cell Signaling Technology, MA) containing 20 mM Tris-HCl (pH 7.5), 150 mM NaCl, 1 mM Na2EDTA, 1 mM EGTA, 1% Triton, 2.5 mM NaPPi, 1 mM α-glycerophosphate, 1 mM Na3 VO4, 1 μg/ml leupeptin, and 1 mM phenylmethylsufonyl fluoride. Tissue lysates were transferred to Eppendorf tubes, mixed on a rocker for 1 hr at 4°C, and then centrifuged at 8,100 × g for 20 min. Supernatants were collected and protein concentrations were estimated using the Bio-Rad DC protein assay. Approximately 30–50 μg of total protein from cells and 10–20 μg of protein from tissues for each sample were loaded on a 12% SDS-polyacrylamide gel (Bio-Rad, CA) and electrophoresed. Proteins were transferred to a PVDF immobilon TM-P membrane (Millipore, MA) overnight at 4°C and blocked in 5% nonfat powered milk in Tris-buffered saline with Tween 20 (pH 7.6) for 3 h. Blots were then probed with different indicated antibodies overnight at 4°C with gentle shaking. The blots were then incubated with appropriate whole horseradish peroxidase-conjugated secondary antibody at room temperature. Bands were visualized using enhanced chemiluminescence plus system (Amersham, UK). Blots were reprobed with β-actin antibody used as a loading control.
Measurement of apoptosis by Annexin V staining
CT-2A and astrocyte cells (including floating cells) grown in 6 well plates were collected following mild trypsinization. According to manufacturer's (BD Bioscience, Pharmingen, CA) protocol, trypsinized cells were washed once with PBS, and they were resuspended in 100 ul of Annexin binding buffer mixed with 5 ul of Annexin V fluorescein conjugate and propidium Iodide (PI). Resuspended cells were incubated at room temperature in the dark for 15 minutes. Labeled cells were analyzed by FACS (Beckman coulter, CA).
Mice and experimental brain tumors
Mice of the C57BL/6J (B6) strain were obtained from the Jackson Laboratory (Bar Harbor, ME). The mice were propagated in the animal care facility of the Department of Biology of Boston College, using the animal husbandry conditions described previously . The syngeneic mouse brain tumor, CT-2A, was originally produced by implantation of a chemical carcinogen, 20-methylcholanthrene, into the brains of B6 mice [50, 52]. The CT-2A tumor arose in the cerebral cortex and was characterized as a malignant anaplastic astrocytoma . The morphological, biochemical, and growth characteristics of the CT-2A brain tumor have been previously described . Male B6 mice (8–12 weeks of age) were used as tumor recipients. All animal experiments were carried out with the ethical committee approval in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals and approved by the Institutional Care Committee.
Brain tumor implantation
Briefly, B6 mice were anaesthetized with Avertin (0.1 ml/10 g body weight) intra-peritoneally and their heads shaved and swabbed with 70% ethyl alcohol under sterile conditions. Small tumor pieces (about 1 mm3, estimated using a 1 mm × 1 mm grid) from the donor tumor were implanted into the right cerebral hemisphere of anaesthetized recipient mice as we previously described [49, 53].
The mice were group housed prior to the initiation of the experiment and were then separated and randomly assigned to either a control group that was fed ad libitum (AL) or to a caloric restricted (CR) group that was fed a total CR of 40% (60% of the control group). Within each experiment, the AL-fed and CR-fed mice were matched for age and body weight. Each mouse was housed singly in a plastic shoebox cage with a filter top and was given a cotton nesting-pad for warmth. CR was initiated 24 hours after tumor implantation and was continued for 13 days after implantation. Total CR maintains a constant ratio of nutrients to energy, i.e., the average daily food intake (grams) for the AL fed mice was determined every other day and the CR-fed mice were given 60% of that quantity on a daily basis [54, 55]. All mice received PROLAB RMH 3000 chow (LabDiet, Purina, Richmond, IN) that contained a balance of mouse nutritional ingredients and delivered 4.1–4.4 Kcal/g gross energy according to the manufacturer's specifications. Body weights of all the mice were recorded every other day.
In situ apoptotic cell detection (TUNEL)
Apoptotic cells were detected using the ApopTag in situ detection kit TUNEL (terminal deoxynucleotidyl transferase mediated deoxyuridine triphosphate biotin nick end labeling) (Oncor, Gaithersberg, MD) as we previously described [49, 55].
We thank to Michael Kiebish for technical assistance. This work was supported in part from NIH grants (HD39722) and (CA102135) to T.N.S, a grant from the American Institute of Cancer Research, and the Boston College Expense Fund.
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.