Original Paper | Published:

The glucose dependence of Akt-transformed cells can be reversed by pharmacologic activation of fatty acid β-oxidation

Oncogene volume 24, pages 41654173 (16 June 2005) | Download Citation

Subjects

Abstract

Activation of the oncogenic kinase Akt stimulates glucose uptake and metabolism in cancer cells and renders these cells susceptible to death in response to glucose withdrawal. Here we show that 5-aminoimidazole-4-carboxamide ribonucleoside (AICAR) reverses the sensitivity of Akt-expressing glioblastoma cells to glucose deprivation. AICAR's protection depends on the activation of AMPK, as expression of a dominant-negative form of AMPK abolished this effect. AMPK is a cellular energy sensor whose activation can both block anabolic pathways such as protein synthesis and activate catabolic reactions such as fatty acid oxidation to maintain cellular bioenergetics. While rapamycin treatment mimicked the effect of AICAR on inhibiting markers of cap-dependent translation, it failed to protect Akt-expressing cells from death upon glucose withdrawal. Compared to control cells, Akt-expressing cells were impaired in the ability to induce fatty acid oxidation in response to glucose deprivation unless stimulated with AICAR. Stimulation of fatty acid oxidation was sufficient to maintain cell survival as activation of fatty acid oxidation with bezafibrate also protected Akt-expressing cells from glucose withdrawal-induced death. Conversely, treatment with a CPT-1 inhibitor to block fatty acid import into mitochondria prevented AICAR from stimulating fatty acid oxidation and promoting cell survival in the absence of glucose. Finally, cell survival did not require reversal of Akt's effects on either protein translation or lipid synthesis as the addition of the cell penetrant oxidizable substrate methyl-pyruvate was sufficient to maintain survival of Akt-expressing cells deprived of glucose. Together, these data suggest that activation of Akt blocks the ability of cancer cells to metabolize nonglycolytic bioenergetic substrates, leading to glucose addiction.

Introduction

The high glycolytic rate and increased glucose uptake of tumors were first reported by Warburg (for a review see Warburg, 1956). The mechanism of this high use of glucose under aerobic conditions is still debated. Warburg proposed that the high glycolytic phenotype was the consequence of mitochondrial respiratory defects, while others have suggested that this phenotype was an adaptive response to hypoxic conditions (Dang and Semenza, 1999). Recently, the highly glycolytic phenotype of cancer cells was reported to be induced directly by activation of the oncogene Akt (Elstrom et al., 2004), a serine/threonine kinase, which promotes anabolic metabolism, supports cell survival, and suppresses apoptosis through a variety of mechanisms. Akt's role in transformation has been considered to be mostly due to its effects on promoting cell survival. Despite its antiapoptotic role, Akt activation renders cancer cells absolutely dependent on the availability of glucose for their survival (Elstrom et al., 2004).

In nontransformed cells, AMP-activated protein kinase (AMPK) acts as an intracellular energy sensor that is activated by an increase in cellular AMP levels. Under stress conditions such as hypoxia, ischemia, and glucose deprivation, an increase in intracellular AMP allosterically activates AMPK to maintain the energy balance within the cell (Hardie, 2003). AMPK activation can both suppress anabolic pathways that consume energy and stimulate catabolic pathways that produce ATP. Mammalian AMPK has homologues in the budding yeast S. cerevisiae (SNF1) and in plants (SnRK) (Halford and Hardie, 1998; Carlson, 1999). Both AMPK homologues were originally identified by their ability to permit cells to metabolize alternative carbohydrates in response to glucose deprivation. However, mammalian cancer cells cannot adapt to glucose deprivation by metabolizing alternative carbohydrates the way yeast can. Several studies have observed that activation of AMPK can protect mammalian cells from stress-induced apoptosis, but the mechanisms of this protective effect are not yet fully understood (Culmsee et al., 2001; Hashimoto et al., 2002; Kato et al., 2002). AMPK has been shown to inhibit anabolic pathways such as protein translation and fatty acid and cholesterol synthesis, thus limiting ATP utilization during energetic stress. Recently, Inoki et al. suggested that inhibition of cap-dependent translation via AMPK-mediated TSC2 phosphorylation is critical for cell survival in response to ATP depletion (Inoki et al., 2003).

Here we show that following glucose depletion, the viability of Akt-transformed cells cannot be maintained simply by suppressing macromolecular synthesis. To maintain cell viability in response to glucose deprivation, Akt-transformed cells must begin to catabolize fatty acids to provide mitochondria with sufficient substrates to maintain oxidative phosphorylation and cell survival. Cancer cells lacking constitutively active Akt are able to induce sufficient AMPK-dependent β-oxidation following glucose restriction to maintain cell survival. In contrast, the endogenous levels of AMPK induced by glucose deprivation are not sufficient to reverse Akt suppression of β-oxidation in Akt-expressing cells unless supraphysiological levels of the AMP-analog 5-aminoimidazole-4-carboxamide ribonucleoside (AICAR) are used to activate AMPK. These data suggest that manipulations to limit glucose uptake or metabolism will be selectively toxic to Akt-transformed cells and that the therapeutic efficacy of such treatments may be compromised by the concomitant use of AMPK activators or other drugs that activate fatty acid β-oxidation. These insights may guide the development of target therapeutics against the metabolic abnormalities associated with constitutive activation of Akt.

Results

AMPK activation or Akt downregulation protects Akt-expressing cells from death induced by glucose deprivation

Cells that express constitutively active Akt depend on glucose availability for their survival (Elstrom et al., 2004). Two independent glioblastoma cell lines were utilized for this study, one of which possesses and one of which lacks constitutive Akt activity (LN18 and LN229, respectively). LN18 cells are susceptible to death induced by glucose deprivation, whereas LN229 cells are resistant to this stress (Figure 1a, b). In addition, two clones of LN229 cells stably expressing myristoylated Akt (myrAkt), a constitutively active form of Akt, were isolated. LN18, control transfected LN229, and LN229 cells expressing myrAkt (LN229+myrAkt) grow at similar rates in serum-free medium. In these conditions, while LN229 lacks active Akt, both LN18 and LN229+myrAkt cells possess constitutive Akt activation, as demonstrated by the presence of Akt phosphorylation at Ser473 (Figure 1c). Like LN18, LN229+myrAkt cells, when cultured in the absence of glucose, underwent massive apoptosis (Figure 1d).

Figure 1
Figure 1

AICAR activates AMPK and rescues Akt-expressing cells from death induced by glucose withdrawal. (a) Akt-transformed glioblastoma cells (LN18) and (b) non-Akt-transformed cells (LN229) were plated at 2 × 105 cells/well in serum-free medium in the presence of glucose (+glc; □), or in the absence of glucose with no drug (−glc; ) or with 1 mM AICAR (−glc+AICAR; •). Viability was determined by propidium iodide exclusion at the indicated times. (c) Total cell extracts were assayed by Western blot for phospho-Ser473 Akt (pS473-Akt), phospho-Ser79 ACC (pS79-ACC), phospho-Thr172 AMPK (pT172-AMPK) and actin. Lysates were prepared from cells cultured in serum-free medium in the presence or absence of glucose, with or without 1 mM AICAR for 8 h. (d) Viability of LN229 expressing myrAkt (LN229+Akt) cells was measured when cells were cultured in serum-free medium in the presence of glucose (+glc; □), or in the absence of glucose with no drug (−glc; ) or with 1 mM AICAR (−glc+AICAR; •). (e) LN18 cells were transfected with Akt siRNA (siRNA-Akt) or control siRNA (control). Viability, measured by propidium iodide exclusion, and Western blot for total Akt are shown. Data are presented as mean±s.d. of triplicate samples and are representative of three independent experiments

Since AMPK has been reported to be activated under conditions of low glucose, we investigated the induction of AMPK phosphorylation and activity in response to glucose withdrawal. To confirm activation of AMPK, phosphorylation of AMPK and ACC, a substrate phosphorylated by AMPK (Davies et al., 1990), was determined by immunoblot (Figure 1c). Under conditions of glucose withdrawal, all cell lines showed increased phosphorylation of AMPK and its substrate ACC. As quantitated by densitometry, the phosphorylation of the AMPK substrate ACC increased three-fold in LN18, 4.3-fold in LN229, and 3.5-fold in LN229+myrAkt cells, when the cells were incubated in glucose-free medium. Yet, despite this induction of AMPK activity, Akt-transformed cells failed to survive glucose withdrawal. Addition of AICAR, a widely used AMPK activator (Corton et al., 1995), to maximally elevate AMPK, led to comparable levels of S79-ACC phosphorylation in all three cell lines. In addition, treatment with AICAR significantly reversed the glucose dependence of cells with activated Akt (LN18 and LN229+myrAkt). At 2 days following glucose removal, the majority of Akt-expressing cells treated with AICAR remained viable, while there was extensive cell death in untreated Akt-expressing cells (Figure 1a, d). Finally, to confirm that the glucose addiction of LN18 cells was conferred by the constitutive activation of Akt, as it is in LN229+myrAkt, the cells were transfected with siRNA to suppress Akt expression. Transfected cells showed a significant downregulation of Akt protein and demonstrated a five-fold increase in the ability to survive glucose deprivation (Figure 1e).

AICAR-induced survival is dependent on activation of AMPK

To determine whether AMPK activation was required for the protective effect of AICAR in response to glucose deprivation, we stably expressed a dominant-negative form of AMPK (dnAMPK) in LN229+myrAkt cells. In cells expressing dnAMPK, activation of AMPK, in response to glucose deprivation and AICAR, was diminished, as represented by the reduced level of phosphorylation of AMPK on Thr172 and ACC on Ser79 (Figure 2c). This reduced level of AMPK activation in dnAMPK-expressing cells correlated with accelerated death during glucose withdrawal (Figure 2a). The expression of dnAMPK also impaired the ability of AICAR to protect cells from death induced by glucose deprivation (Figure 2b). The incomplete inhibition of endogenous AMPK and ACC phosphorylation in cells expressing dnAMPK may explain the fact that the effect of AICAR on viability was not completely blocked in these cells. The data suggest that AMPK activation is essential for AICAR-induced survival of Akt-expressing cells when cultured in the absence of glucose.

Figure 2
Figure 2

The protective effect of AICAR on cell survival is mediated by AMPK activation. (a) Percent viability following glucose withdrawal is presented for LN229 cells stably transfected with myrAkt in the presence of glucose (+glc; □), the absence of glucose (−glc; ), or the absence of glucose expressing dnAMPK (−glc+dnAMPK; •). (b) Percent viability following glucose withdrawal of LN229 cells expressing myrAkt in the presence of no drug (−glc; ), 1 mM AICAR (−glc+AICAR; □), or transfected with dnAMPK transgene and treated with 1 mM AICAR (−glc+AICAR+dnAMPK; ). Data are presented as mean±s.d. of triplicate samples and are representative of three independent experiments. (c) Western blots of phospho-Ser79 ACC (pS79-ACC), phospho-Thr172 AMPK (pT172-AMPK), and actin. Lysates from LN229 cells expressing myrAkt were prepared after cells were cultured in serum-free medium in the presence or absence of glucose/AICAR for 8 h

AICAR fails to protect cells from staurosporine-induced apoptosis

To assess whether AICAR acts as a general inhibitor of apoptosis, we tested its ability to protect cells treated with staurosporine, a broad kinase inhibitor (Omura et al., 1995). LN229 control cells or LN229 cells expressing myrAkt were cultured in serum-free medium containing glucose. Staurosporine (STS) induced cell death in both cell lines. AICAR did not protect cells from this death (Figure 3a), even though phosphorylation of AMPK and ACC were highly induced by AICAR (Figure 3b). Thus, AICAR's protective effect is selective for glucose deprivation-induced cell death and does not result from its ability to suppress apoptosis in general.

Figure 3
Figure 3

Activation of AMPK fails to protect cells from staurosporine-induced death. (a) Percent viability was measured for control LN229 cells (control) or LN229 cells expressing myrAkt (myrAkt) cultured in a serum-free medium for 48 h in the presence of no drug (no drug), 2 μM staurosporine (+STS), or 2 μM staurosporine and 1 mM AICAR (+STS+AICAR). Data are presented as mean±s.d. of triplicate samples and are representative of two independent experiments. (b) Western blot for phospho-Ser79 ACC (pS79-ACC), phospho-Thr172 AMPK (pT172-AMPK), and actin. Lysates were prepared from control LN229 cells or from cells expressing myrAkt cultured in serum-free medium in the presence of glucose with or without 1 mM AICAR for 24 h

Inhibition of mTOR does not protect cells from glucose deprivation-induced death

Protein synthesis is a process that increases consumption of cellular energy (Rolfe and Brown, 1997). During energy starvation, inhibition of translation may help maintain cellular homeostasis. Increasing evidence supports a role for AMPK in the inhibition of protein synthesis by downregulating the mTOR pathway (Horman et al., 2002; Inoki et al., 2003). Thus, we tested whether, during glucose withdrawal, the increased cell survival observed in AICAR-treated cells could be reproduced by inhibition of mTOR alone. 4E-binding protein 1 (4EBP1) is a downstream target of mTOR. Phosphorylation of 4EBP1 by mTOR stimulates cap-dependent translation by relieving the ability of 4EBP1 to suppress eIF4E. In the absence of glucose, LN229 cells expressing myrAkt displayed increased phosphorylation of 4EBP1 in comparison to wild-type cells. Further, myrAkt-induced phosphorylation of 4EBP1 was inhibited by AICAR to a comparable extent to that observed when cells were treated with 20 nM rapamycin, a specific inhibitor of mTOR (Figure 4a). However, unlike AICAR, rapamycin failed to protect Akt-expressing cells from death induced by glucose starvation (Figure 4c). The experiment was also repeated using increasing concentrations of rapamycin. No increase in survival was observed when myrAkt cells were treated with 40, 100, or 200 nM rapamycin (data not shown).

Figure 4
Figure 4

Rapamycin treatment blocks phosphorylation of 4EBP1 but does not protect Akt-expressing cells from glucose deprivation-induced death. (a) Control LN229 cells or LN229 cells expressing myrAkt were maintained in serum-free medium for 24 h before they were withdrawn from glucose in the presence or absence of 1 mM AICAR or 20 nM rapamycin (RAP) for 2 h. Lysates were analysed by Western blot for phospho-Ser473 Akt (pS473-Akt), phospho-Thr37/46 4EBP1 (pT37/46-4EBP1), and actin. Percent viability was measured daily for control LN229 cells (b), or LN229 cells expressing myrAkt (c), when cultured in medium containing no glucose in the presence of no drug (−glc; ), 20 nM rapamycin (−glc+RAP; •), or 1 mM AICAR (−glc+AICAR; □). Data are presented as mean±s.d. of triplicate samples and are representative of three independent experiments

Activation of Akt prevents AMPK-dependent changes in fatty acid metabolism in response to glucose withdrawal

To evaluate whether constitutive activation of Akt affects the ability of cells to adapt to glucose withdrawal by altering lipid metabolism, we determined the rate of fatty acid oxidation and fatty acid synthesis in LN229 control cells or LN229 cells expressing myrAkt (Figure 5a, b). In response to glucose withdrawal, control cells upregulated their rate of fatty acid oxidation and suppressed lipid synthesis. Under all conditions tested, LN229 cells with activated Akt demonstrated significantly reduced rates of fatty acid oxidation in comparison to control LN229 cells. In addition, Akt-expressing cells maintained significantly higher levels of fatty acid synthesis in the absence of glucose than control cells. During glucose withdrawal, only when Akt-transfected cells were treated with AICAR did the levels of fatty acid oxidation rise to levels comparable to those observed in glucose-deprived control cells. AICAR also suppressed fatty acid synthesis in glucose-deprived Akt-transfected cells to a level comparable to that of glucose-deprived control transfected cells. To determine whether the effect of AICAR on fatty acid oxidation was dependent on AMPK, Akt-expressing cells were cultured in the absence of glucose and treated with AICAR in the presence or absence of a dominant-negative form of AMPK (dnAMPK) (Figure 5c). The AICAR induction of fatty acid oxidation in Akt-transfected cells was significantly reduced by dominant-negative AMPK.

Figure 5
Figure 5

Akt activation affects AMPK-dependent changes in lipid metabolism following glucose withdrawal. (a) Fatty acid oxidation was determined by measuring 3H2O produced by control LN229 cells (control) or LN229 cells expressing myrAkt (myrAkt) incubated with [9,10-3H]palmitate for 6 h. Cells were incubated in serum-free medium containing glucose (+glc), no glucose (−glc), or no glucose and 1 mM AICAR (−glc+AICAR). Data are presented as mean±s.d. of triplicate samples and are representative of three independent experiments. (b) Fatty acid synthesis was assayed by measuring the amount of 14C incorporated into lipids after 2 h incubation with [1-14C]acetate. Prior to incubation with [1-14C]acetate, vector control and Akt-expressing cells were plated in medium containing glucose (+glc), no glucose (−glc), or no glucose and 1 mM AICAR (−glc+AICAR) for 1 h. (c) Fatty acid oxidation was measured in LN229+myrAkt parental cells (myrAkt+control) and cells expressing a dominant-negative form of AMPK (myrAkt+dnAMPK). Cells were cultured in the absence of glucose in the presence (−glc+AICAR) or absence (−glc) of 1 mM AICAR. Data are presented as mean±s.d. of triplicate samples and are representative of three independent experiments

Activation of fatty acid oxidation is necessary and sufficient to protect cells from death following glucose deprivation

To determine whether stimulation of fatty acid oxidation was required for the ability of AICAR to rescue Akt-expressing cells from cell death in response to glucose deprivation, we measured the viability of AICAR-treated cells when fatty acid oxidation was blocked by etomoxir (ETO), an inhibitor of CPT-1 (Lilly et al., 1992). CPT-1 is required for mitochondrial import of long-chain fatty acids to initiate β-oxidation. LN229 cells expressing myrAkt were cultured in the absence of glucose and treated either with AICAR alone or together with 60 μM ETO, a concentration that inhibited approximately 70% of AICAR-induced fatty acid oxidation (Figure 6c). In the absence of glucose, treatment with ETO blocked the ability of AICAR to promote cell survival in Akt-expressing cells (Figure 6a). ETO showed no toxicity when added to cells treated with AICAR in the presence of glucose (Figure 6b). ETO also reversed the protective effect of RNA interference against Akt in LN18 cells, suggesting that increased fatty acid oxidation is the mechanism for siRNA-mediated improvement of viability (Supplementary Figure 1).

Figure 6
Figure 6

Activation of fatty acid oxidation is both necessary and sufficient to protect Akt-expressing cells from death induced by glucose deprivation. (a) Percent viability was measured daily by propidium iodide exclusion for LN229 cells expressing myrAkt. Cells were plated at 2 × 105 cells/well and maintained in serum-free medium for 24 h before shifting to medium lacking glucose with no drug (−glc; ), with 1 mM AICAR (−glc+AICAR; □), or with 1 mM AICAR and 60 μM etomoxir (−glc+AICAR+ETO; •); or (b) cells were incubated in medium containing glucose in the presence of no drug (+glc; ), or 1 mM AICAR and 60 μM etomoxir (+glc+AICAR+ETO; •). (c) Fatty acid oxidation was determined as 3H2O production when cells were incubated with [9,10-3H]palmitate for 6 h. LN229 cells expressing myrAkt were maintained in serum-free medium for 24 h, then cells were shifted to glucose-free medium containing 1 mM AICAR (−glc+AICAR), or 1 mM AICAR in the presence of 60 μM etomoxir (−glc+AICAR+ETO). (d) Fatty acid oxidation measured as 3H2O production by cells expressing myrAkt when incubated with [9,10-3H]palmitate for 6 h. Cells were cultured in serum-free medium for 24 h, shifted to glucose-free conditions with no drug (−glc), or in the presence of 1 mM bezafibrate (−glc+BEZA). (e) Percent viability was measured by propidium iodide exclusion for myrAkt-expressing cells that were plated in serum-free medium for 24 h and then shifted to glucose-free medium with no drug (−glc; ), or treated with 1 mM bezafibrate (−glc+BEZA; •). Data are presented as mean±s.d. of triplicate samples and are representative of three independent experiments

To determine whether, in the absence of glucose, stimulation of fatty acid oxidation might be a sufficient mechanism to promote the survival of Akt-expressing cells, fatty acid oxidation was activated using bezafibrate, an activator of the peroxisome proliferator-activated receptor alpha (PPARα) (Kersten et al., 2000). Under conditions of glucose deprivation, treatment with bezafibrate (BEZA) promoted fatty acid oxidation (Figure 6d) and led to an increase in cell survival in Akt-expressing cells (Figure 6e). These results suggest that activation of fatty acid oxidation alone is sufficient to rescue Akt-expressing cells from glucose withdrawal-induced death.

Addition of methyl-pyruvate rescues viability during glucose deprivation

Death induced by glucose withdrawal in Akt-expressing cells could be due to Akt-mediated suppression of fatty acid catabolism, thus suppressing the generation of substrates required to support bioenergetics in the absence of glucose. Alternatively, Akt-induced changes in lipid metabolism could stimulate the production of lipid intermediates, such as ceramide or sphingosine, which might be proapoptotic in glucose-deprived cells. AMPK may protect these cells simply by stimulating the degradation of these proapoptotic lipids. To distinguish between these two possibilities we tested whether providing cells with an alternative oxidizable substrate could protect them from death upon glucose withdrawal. Methyl-pyruvate, a membrane permeable form of pyruvate, has been shown to be efficiently metabolized in the mitochondria resulting in increased ATP generation (Jijakli et al., 1996). Cells growing in glucose-free medium were supplied with 10 mM methyl-pyruvate. The addition of methyl-pyruvate rescued the viability of the Akt-expressing cells LN18 and LN229+myrAkt, when subjected to glucose withdrawal (Figure 7a, c).

Figure 7
Figure 7

Addition of methyl-pyruvate rescues viability of Akt expressing cell during glucose deprivation. Percent viability was measured by propidium iodide exclusion for LN18 (a), LN229 (b), and LN229+myrAkt (c) cells. Cells were withdrawn from glucose and incubated in the presence of no drug (−glc; ), or 10 mM methyl-pyruvate (−glc+pyruvate; •) for three days. Data are presented as mean±s.d. of triplicate samples and are representative of three independent experiments

Discussion

Cancer cells resist a variety of apoptotic stimuli and achieve independence from extracellular growth factors. Nevertheless, these cells still depend on a nutrient supply to proliferate and survive. Akt is an oncogene that promotes cell survival during growth factor deprivation. However, during glucose limitation, Akt paradoxically sensitizes cells to death. Studies aimed to understand the mechanisms of the glucose dependence of Akt-expressing cells have reported that glucose itself is critical to suppress apoptosis. Gottlob et al. (2001) proposed that the presence of glucose promotes the association of hexokinase with the voltage-dependent anion channel (VDAC) at the mitochondria, thus preventing the association of proapoptotic Bax with the channel and preserving mitochondrial integrity. Danial et al. (2003) showed that glucose may affect the Akt target Bad by regulating its phosphorylation and its association with glucokinase. In this report, we demonstrate that the glucose dependence of Akt-expressing cells reflects the inability of these cells to adapt to the bioenergetic stress of glucose withdrawal rather than to the actual depletion of glucose or glycolytic intermediates. When highly glycolytic Akt-expressing cells are withdrawn from glucose, their ability to switch from glucose to fatty acids as an alternative bioenergetic substrate is compromised, leading to energy failure and cell death. Pharmacologic activation of AMPK with AICAR overcomes the glucose dependence of Akt-expressing cells by shifting the balance between AMPK and Akt's effects on lipid metabolism, allowing the cells to utilize fatty acid oxidation as a survival adaptation.

AMPK-dependent protection from glucose deprivation-induced apoptosis has been suggested to involve inhibition of protein synthesis. Inoki et al. (2003) proposed that, during glucose deprivation, AMPK-dependent phosphorylation of TSC2 leads to inactivation of the mTOR pathways and protection from apoptosis. In our study, while activation of AMPK in Akt-expressing cells reduced the phosphorylation of 4EBP1, comparable inhibition of 4EBP1 phosphorylation by rapamycin failed to protect cells from death. These data suggest that, in Akt-expressing cells, inhibition of the anabolic mTOR pathway is not sufficient to protect cells from the energetic stress of glucose withdrawal.

In contrast, our data suggest that Akt-mediated suppression of fatty acid β-oxidation is responsible for the glucose dependence induced by constitutive Akt activation. First, expression of a constitutively active Akt inhibited fatty acid oxidation under all conditions tested (Figure 5a). Second, fatty acid oxidation could be induced by either bezafibrate or AICAR, and these treatments supported viability upon glucose withdrawal (Figures 6e and 1a, d). Third, siRNA inhibition of Akt expression enhanced cell survival upon glucose withdrawal of cells with activated Akt (Figure 1e), and this effect was reversed by ETO, a CPT1 inhibitor (Supplementary Figure 1). Oxidation of fatty acids is a major source of energy for heart and skeletal muscle, and in these tissues is regulated by energetic demand in an AMPK-dependent manner (Merrill et al., 1997). The present results indicate that activation of fatty acid oxidation is also an important mechanism for the survival of cancer cells when deprived of glucose.

Bolster et al. (2002) reported that AMPK activation reduced Akt phosphorylation and suggested that kinases upstream of Akt are AMPK targets. Direct cross-talk between Akt and AMPK as well as antagonizing actions have been proposed for these kinases (Inoki et al., 2003; Kovacic et al., 2003). In our study, we quantitated AMPK and ACC phosphorylation in Akt-expressing cells compared to control cells and found that neither AMPK nor ACC phosphorylation is suppressed in Akt-transformed cells. Nevertheless, activation of Akt antagonized the effects of endogenous AMPK on lipid metabolism.

Akt and AMPK can contribute to fatty acid oxidation and fatty acid synthesis by regulating multiple steps of these pathways. Mitochondrial oxidation of long-chain fatty acids depends on the action of CPT-1, which catalyses fatty acid transport into mitochondria. CPT-1 is allosterically inhibited by high levels of malonyl-CoA, the formation of which is catalysed by ACC, using acetyl-CoA as a substrate. AMPK phosphorylates and inhibits ACC leading to a fall in malonyl-CoA levels and activation of CPT-1. Upstream of ACC, ATP-citrate lyase (ACL) processes citrate exported from mitochondria to produce the cytosolic acetyl-CoA that serves as the substrate for the production of malonyl-CoA by ACC. ACL can be phosphorylated and activated by Akt (Berwick et al., 2002). Consistent with these findings, we observed that Akt-expressing cells have an increased phosphorylation of ACL (data not shown). Thus, constitutive activation of ACL could support increased acetyl-CoA and malonyl-CoA production, inhibiting fatty acid oxidation and inducing lipid synthesis in Akt-expressing cells.

On the other hand, it is also possible that Akt regulates other steps of lipid metabolism. For example, a constitutively active mutant of the transcription factor Foxa2, which is normally phosphorylated and inhibited by Akt, can enhance CPT-1 expression and fatty acid oxidation (Wolfrum et al., 2004). Therefore, it is possible that constitutive Akt activity suppresses expression of proteins required for fatty acid oxidation.

Depending on the metabolic demand, fatty acids can either be used for macromolecular synthesis or oxidized in the mitochondria for energy production. However, cellular transformation, by promoting cell cycle progression and cellular growth, would create a requirement for constant supply of fatty acids for synthesis of new cellular membranes, leading to a decrease in substrates for β-oxidation. Inhibition of enzymes involved in lipid synthesis has already been shown to suppress the growth of cancer cell lines; thus, these enzymes have been considered potential targets for cancer therapy. In this study, we have found that Akt-transformed cells are sensitive to inhibition of glycolysis because, as a consequence of their transformation, they are unable to efficiently oxidize fatty acids to maintain bioenergetics. This suggests that inhibitors of glucose uptake and/or metabolism may have therapeutic efficacy for treating Akt-transformed cells, because vegetative and nontransformed proliferating cells with a normal ability to use alternative substrates should be much less susceptible to glycolytic inhibitors. Furthermore, since pharmacologic activation of AMPK can reverse the glucose dependence of Akt-transformed cells, the data also suggest that clinically utilized AMPK activators, such as metformin, might enhance tumor cell survival when such drugs are utilized concomitantly with experimental agents that impair glucose utilization. Further studies are needed to define how oncogene-induced perturbations of cellular metabolism can be effectively exploited to improve the development and utilization of cancer therapeutics.

Materials and methods

Cell lines and cell culture

The glioblastoma cell lines LN18 and LN229 (kindly provided by Dr M Celeste Simon) have been described previously (Diserens et al., 1981; Van Meir et al., 1994). LN229 cells were stably transfected with a vector control or myristoylated Akt1 (myrAkt) as previously described (Elstrom et al., 2004). For all the experiments two myrAkt clones were tested. All the data represented in the figures utilize clone 1 (myrAkt-1) as LN229+myrAkt (Elstrom et al., 2004). All the cell lines were maintained as previously described (Elstrom et al., 2004). All experiments were performed following initial passaging of the cells in serum-free RPMI 1640 medium supplemented with 100 U/ml penicillin, 100 μg/ml streptomycin, and 2 mM L-glutamine for 24 h. For experiments involving glucose deprivation, cells were cultured in RPMI 1640 medium without glucose, but containing L-glutamine, for the time as indicated. Cell concentration was determined with a Coulter Z2 particle analyser. Cell viability was determined by the exclusion of 2 μg/ml propidium iodide (Molecular Probes) by a LSR flow cytometer (BD Biosciences).

Reagents

AICAR was obtained from Toronto Research Chemicals Inc. (North York, ON, Canada). Bezafibrate, ETO, staurosporine, fatty acid-free BSA (BSA), [9,10-3H]palmitate, Dowex 1X8-200 ion exchange resin, and methyl-pyruvate were obtained from Sigma (St Louis, MO, USA). Rapamycin was purchased from Calbiochem (San Diego, CA, USA). [1-14C]acetate was purchased from Amersham (Piscataway, NJ, USA).

Plasmids and transfection

The rat AMPK-α2 K45R construct (kindly provided by Dr Morris J Birnbaum) was previously described (Mu et al., 2001). The construct was subcloned into the EcoR1 site of pIRESpuro3 plasmid. Transfection was performed in LN229+myrAkt cells with Lipofectamine 2000 (Life technologies, Inc., Gaithersburg, MD, USA), and stable transfected cells (LN229+myrAkt+dnAMPK) were selected in medium containing puromycin.

Western blot analysis

Cells were lysed in RIPA buffer containing protease inhibitors (Complete; Roche Applied Science, Indianapolis, IN, USA) and phosphatase inhibitor cocktails I and II (Sigma). Protein concentrations were determined using the BCA Protein Assay Kit (Pierce Biotechnology; Rockford, IL, USA). Equal amounts of total protein (15 μg) were resolved on 10% NuPage Bis-Tris polyacrylamide gels (Invitrogen, Carlsbad, CA, USA). Nitrocellulose membranes were blocked in PBS containing 5% nonfat dry milk and 0.1% Tween-20 with the following antibodies: anti-phospho-Ser79 Acetyl CoA Carboxylase (Upstate, Lake Placid, NY, USA), anti-phospho-Ser473 Akt and anti-phospho-Thr37/46 4EBP1 (Cell Signaling Technology, Beverly, MA, USA), anti-phospho-Thr172 AMPK (gift from Dr Morris J Birnbaum, University of Pennsylvania, Philadelphia, PA, USA) (Mu et al., 2001), or anti-actin (Santa Cruz Biotechnology, Inc., Santa Cruz, CA, USA). Bands were detected using horseradish peroxidase-labeled secondary antibodies and enhanced chemiluminescence detection kit (Amersham).

siRNA

A heterogeneous mixture of short interfering RNAs (siRNA) was used to silence human Akt 1/2/3 kinases (New England BioLabs #N2005S). Cells were seeded at 1.5 × 105 cells/12 well plate and transfected with 75 nM Akt-siRNA or negative control siRNA using TransPass R1 transfection reagent (NEB). At 48 h following transfection cells were withdrawn from glucose for 24 h and viability was measured by propidium iodide exclusion.

Fatty acid oxidation assay

To measure fatty acid oxidation, we modified an assay in which oxidation of [9,10-3H]palmitate results in formation of 3H2O (Brivet et al., 1995). Briefly, cells were seeded in six-well plates at 1 × 106 cells/well in triplicate and incubated overnight in serum-free medium. The medium was then removed followed by addition of 1 ml medium containing 3 μCi/ml [9,10-3H]palmitate (70 nM), bound to 10 μM BSA. Incubation was carried out for 6 h at 37°C. Tritiated water was recovered by ion-exchange treatment on Dowex 1X8-200 columns (Moon and Rhead, 1987). Supernatants were applied to 3 ml-Dowex 1X8-200 columns and eluted with 2 ml water. 3H2O was quantified by liquid scintillation counting. In Figures 5a, c and 6d, fatty acid oxidation rates are shown as CPT-1 dependent oxidation. Counts due to nonspecific oxidation, as demonstrated by inability to be inhibited by 60 μM ETO, were subtracted from total counts.

Fatty acid synthesis assay

The rate of fatty acid synthesis was measured as 14C incorporation into lipids after incubation of cells with [1-14C]acetate. Briefly, cells were seeded in 12-well plates at 5 × 105 cells/well and incubated in serum-free medium for 24 h. Following 1-h incubation in the indicated conditions, 1 μCi [1-14C]acetate/500 μl medium was added to the cells. The reaction was carried for 2 h at 37°C. Cells were then washed with Tris-Cl (50 mM Tris-Hcl pH 7.4, 150 mM NaCl) twice. Lipids were extracted from cells by adding 500 μl hexane : isopropanol (3 : 2) and incubated for 30 min on a shaker at room temperature. Extracts were dried under nitrogen for 10 min and resuspended in 80 μl 99.9% chloroform. 14C incorporation into lipids was quantified by liquid scintillation counting.

References

  1. , , , and . (2002). J. Biol. Chem., 277, 33895–33900.

  2. , , and . (2002). J. Biol. Chem., 277, 23977–23980.

  3. , , , and . (1995). Ann. Clin. Biochem., 32 (Part 2), 154–159.

  4. . (1999). Curr. Opin. Microbiol., 2, 202–207.

  5. , , and . (1995). Eur. J. Biochem., 229, 558–565.

  6. , , and . (2001). J. Mol. Neurosci., 17, 45–58.

  7. and . (1999). Trends Biochem. Sci., 24, 68–72.

  8. , , , , , , , , , , and . (2003). Nature, 424, 952–956.

  9. , and . (1990). Eur. J. Biochem., 187, 183–190.

  10. , , , , and . (1981). Acta Neuropathol. (Berl.), 53, 21–28.

  11. , , , , , , , , , and . (2004). Cancer Res., 64, 3892–3899.

  12. , , , , and . (2001). Genes Dev., 15, 1406–1418.

  13. and . (1998). Plant Mol. Biol., 37, 735–748.

  14. . (2003). Endocrinology, 144, 5179–5183.

  15. , , , , , and . (2002). Biochem. Biophys. Res. Commun., 290, 263–267.

  16. , , , , , , , , and . (2002). Curr. Biol., 12, 1419–1423.

  17. , and . (2003). Cell, 115, 577–590.

  18. , , , , and . (1996). Arch. Biochem. Biophys., 335, 245–257.

  19. , , , , , and . (2002). Oncogene, 21, 6082–6090.

  20. , and . (2000). Nature, 405, 421–424.

  21. , , , , and . (2003). J. Biol. Chem., 278, 39422–39427.

  22. , , , and . (1992). Biochem. Pharmacol., 43, 353–361.

  23. , , and . (1997). Am. J. Physiol., 273, E1107–E1112.

  24. and . (1987). J. Clin. Invest., 79, 59–64.

  25. , , , and . (2001). Mol. Cell, 7, 1085–1094.

  26. , , and . (1995). J. Antibiot. (Tokyo), 48, 535–548.

  27. and . (1997). Physiol. Rev., 77, 731–758.

  28. , , , , , , , , and . (1994). Cancer Res., 54, 649–652.

  29. , , , and . (2004). Nature, 432, 1027–1032.

  30. . (1956). Science, 124, 269–270.

Download references

Acknowledgements

We thank Dr Morris Birnbaum for providing the dnAMPK plasmid and the AMPK antibody, Casey Fox, Aimee Edinger, and the rest of the Thompson laboratory for technical help, discussions and comments on the manuscript. This work was supported in part by grants from the NIH and NCI. RGS is supported by the Cancer Research Institute. RJD is supported by an NIH Training Grant (5T32-GM008638-08). GH is a Damon Runyon Fellow supported by the Damon Rynyon Cancer Research Foundation (DRG- #1714-02).

Author information

Affiliations

  1. Abramson Family Cancer Research Institute, Department of Cancer Biology, University of Pennsylvania, Philadelphia, PA 19104, USA

    • Monica Buzzai
    • , Daniel E Bauer
    • , Russell G Jones
    • , Ralph J DeBerardinis
    • , Georgia Hatzivassiliou
    • , Rebecca L Elstrom
    •  & Craig B Thompson
  2. Division of Human Genetics and Molecular Biology, The Children's Hospital of Philadelphia, Philadelphia, PA 19104, USA

    • Ralph J DeBerardinis

Authors

  1. Search for Monica Buzzai in:

  2. Search for Daniel E Bauer in:

  3. Search for Russell G Jones in:

  4. Search for Ralph J DeBerardinis in:

  5. Search for Georgia Hatzivassiliou in:

  6. Search for Rebecca L Elstrom in:

  7. Search for Craig B Thompson in:

Corresponding author

Correspondence to Craig B Thompson.

About this article

Publication history

Received

Accepted

Published

DOI

https://doi.org/10.1038/sj.onc.1208622

Supplementary Information accompanies the paper on Oncogene website (http://www.nature.com/onc)

Further reading