Myc Regulates a Transcriptional Program That Stimulates Mitochondrial Glutaminolysis and Leads to Glutamine Addiction

Myc regulates a transcriptional program that stimulates mitochondrial glutaminolysis and leads to glutamine addiction

David R. Wisea, Ralph J. DeBerardinisb, Anthony Mancusoa, Nabil Sayeda, Xiao-Yong Zhangc, Harla K. Pfeifferc, Ilana Nissimd, Evgueni Daikhind, Marc Yudkoffd, Steven B. McMahonc, and Craig B. Thompsona,1

aDepartment of Cancer Biology, Abramson Cancer Center, University of Pennsylvania, Room 451, Biomedical Research Building II/III, 421 Curie Boulevard, Philadelphia, PA 19104-6160; bDepartment of Pediatrics and McDermott Center for Human Growth and Development, University of Texas Southwestern Medical Center, Dallas, TX 75390; cDepartment of Cancer Biology, The Kimmel Cancer Center, Thomas Jefferson Medical College, Philadelphia, PA 19107; and dDepartment of Pediatrics, Children’s Hospital of Philadelphia, 34th Street and Civic Center Boulevard, Philadelphia, PA 19104

Contributed by Craig B. Thompson, October 10, 2008 (sent for review September 12, 2008) Mammalian cells fuel their growth and proliferation through the to exploit the glucose addiction of cells transformed by PI3K catabolism of two main substrates: glucose and glutamine. Most of mutation for cancer therapy are currently being investigated (4). the remaining metabolites taken up by proliferating cells are not In addition to glucose, glutamine can be an essential nutrient for catabolized, but instead are used as building blocks during ana- cell growth and viability (8, 9). In vitro addiction to glutamine as a bolic macromolecular synthesis. Investigations of phosphoinositol bioenergetic substrate was first observed in HeLa cells but it was not 3-kinase (PI3K) and its downstream effector AKT have conﬁrmed found to be a universal property of cancer cell lines. In cancer that these oncogenes play a direct role in stimulating glucose patients, some tumors have been reported to consume such an uptake and metabolism, rendering the transformed cell addicted to abundance of glutamine that they depress plasma glutamine levels glucose for the maintenance of survival. In contrast, less is known (10, 11). Despite these observations, the high rates of glutamine about the regulation of glutamine uptake and metabolism. Here, metabolism and addiction exhibited by some cancer cells are poorly we report that the transcriptional regulatory properties of the understood. Recently, we reported that glioma cells can exhibit oncogene Myc coordinate the expression of genes necessary for glutamine uptake and metabolism that exceeds the cell’s use of cells to engage in glutamine catabolism that exceeds the cellular glutamine for protein and nucleotide biosynthesis (12). In such cells, requirement for protein and nucleotide biosynthesis. A conse- the excess glutamine metabolites produced were found to be quence of this Myc-dependent glutaminolysis is the reprogram- secreted as either lactate or alanine. This high rate of glutaminolysis ming of mitochondrial metabolism to depend on glutamine catab- was found to be beneficial because it provided the cell a high rate olism to sustain cellular viability and TCA cycle anapleurosis. The of NADPH production that was used to fuel lipid and nucleotide ability of Myc-expressing cells to engage in glutaminolysis does biosynthesis (see supporting information (SI) Fig. S1 for schematic not depend on concomitant activation of PI3K or AKT. The stimu- of glutaminolytic pathway). However, not all tumor cells exhibit lation of mitochondrial glutamine metabolism resulted in reduced glutaminolysis. This suggested that the use of glutamine as a glucose carbon entering the TCA cycle and a decreased contribu- bioenergetic substrate is not induced as an indirect consequence of tion of glucose to the mitochondrial-dependent synthesis of phos- cell growth, but as a direct consequence of a specific oncogenic pholipids. These data suggest that oncogenic levels of Myc induce event. a transcriptional program that promotes glutaminolysis and trig- In the present studies, the oncogenes known to contribute to gers cellular addiction to glutamine as a bioenergetic substrate. malignant transformation of glial cells were tested for the ability to induce glutaminolysis. Here, we report that the glutaminolytic cancer ͉ mitochondria phenotype exhibited by tumor cells correlates with a cellular addiction to glutamine metabolism for the maintenance of cell any cancer cell lines depend on a high rate of glucose uptake viability. In contrast to glucose, glutamine uptake was not found to Mand metabolism to maintain their viability despite being be under the direct or indirect control of the PI3K/AKT pathway. maintained in an oxygen-replete environment (1). This metabolic Inhibitors of either PI3K or AKT, despite suppressing glucose phenotype, first observed by Otto Warburg, has been termed metabolism in a dose-dependent fashion, had no effect on the aerobic glycolysis (2). Initially, this high rate of glycolysis was glutaminolytic phenotype. In contrast, high level expression of Myc believed to result from mutations that impair the ability of cancer was required to maintain the glutaminolytic phenotype and addic- cells to carry out oxidative phosphorylation (3). However, such tion to glutamine as a bioenergetic substrate. When an inducible defects appear to be rare in spontaneously arising tumors (4). Myc transgene was introduced in mouse embryonic fibroblasts Recent studies have suggested that activating mutations in phos- (MEF), induction of Myc expression resulted in the induction of phoinositol 3-kinase (PI3K) and its downstream effector AKT glutamine transporters, glutaminase, and lactate dehydrogenase A induce the transformed cell to take up glucose in excess of its (LDH-A). Induction of these key regulatory genes involved in bioenergetic needs (5). The resulting high rate of glycolytic metab- glutaminolysis correlated with the Myc-induced increases in glu- olism leads to the conversion of mitochondria into synthetic or- tamine uptake and glutaminase flux. This increase in glutamine ganelles that support glucose-dependent lipid synthesis and nonessential amino acid production (6, 7). Glycolytic pyruvate that Author contributions: D.R.W., R.J.D., A.M., N.S., X.-Y.Z., E.D., M.Y., S.B.M., and C.B.T. accumulates in excess of cellular bioenergetic and synthetic needs designed research; D.R.W., R.J.D., A.M., N.S., X.-Y.Z., H.K.P., I.N., and E.D. performed is converted to lactate and secreted. A consequence of this meta- research; D.R.W., R.J.D., A.M., N.S., X.-Y.Z., H.K.P., I.N., E.D., and M.Y. analyzed data; and bolic conversion is that cells become addicted to glucose for their D.R.W., R.J.D., A.M., N.S., X.-Y.Z., and C.B.T. wrote the paper. ATP production and survival as available lipids and amino acids are The authors declare no conﬂict of interest. redirected from use as bioenergetic substrates and committed to 1To whom correspondence should be addressed. E-mail: [email protected]. use in anabolic synthesis (5). These data suggest that cancer cell This article contains supporting information online at www.pnas.org/cgi/content/full/ nutrient uptake and metabolism may be under the direct control of 0810199105/DCSupplemental. the oncogenic signaling pathways that transform the cell. Strategies © 2008 by The National Academy of Sciences of the USA

18782–18787 ͉ PNAS ͉ December 2, 2008 ͉ vol. 105 ͉ no. 48 www.pnas.org͞cgi͞doi͞10.1073͞pnas.0810199105 Downloaded by guest on September 29, 2021 Fig. 1. Glutamine catabolism in the human glioma line SF188. (A) Protein synthesis is a minor fate of glutamine carbon. SF188 cells were cultured in medium 14 supplemented with 0.01% [ C-U5]glutamine relative to unenriched glutamine 14 for4h.[ C-U5]glutamine in SSA precipitated protein (striped bar) and total glutamine consumed from the medium (gray bar) are presented as the mean Ϯ standard deviation (SD) of four independent experiments. (B) The requirement for glutamine can be satisﬁed by alpha-ketoglutarate. SF188 cells were allowed to plate in complete medium and then cultured in either glutamine-depleted medium (Ϫ glutamine), complete medium (ϩ glutamine), or glutamine-depleted medium supplemented with 7 mM dimethyl ␣-ketoglutarate (Ϫ glutamine ϩ ␣-ketoglutarate). Cell viability was determined at the time points shown by trypan blue dye exclusion. The data presented are the mean Ϯ SD of triplicate samples. Representative data from one of three independent experiments are Fig. 2. PI3K/Akt signaling regulates the consumption of glucose but not of shown. glutamine. SF188 cells stably expressing Bcl-xL were treated with AktiVIII at doses ranging from 0–20 ␮M. Medium was collected and analyzed for glucose, lactate, glutamine, and ammonia. The rates shown were calculated from the difference uptake was not a compensatory response to increased glutamine in metabolite concentration between the medium at the time point shown and incorporation into proteins as a result of Myc-induced protein fresh medium. The data points presented are the mean Ϯ SD of triplicate samples. synthesis, because most of the additional glutamine carbon taken up after Myc induction was secreted as lactate. Myc-induced repro- gramming of intermediate metabolism resulted in glutamine ad- location of a variety of nutrient transporters (5, 13). We therefore diction, despite the abundant availability of glucose. Glutamine investigated whether the PI3K/AKT pathway might also function to addiction correlated with Myc-induced redirection of glucose car- up-regulate glutamine uptake and metabolism. To determine bon away from mitochondria as a result of LDH-A activation. As whether PI3K/AKT signaling contributes to glutamine metabolism, a result, Myc-transformed cells became dependent on glutamine the effects of the PI3K inhibitor LY294002 or the AKT inhibitor, anapleurosis for the maintenance of mitochondrial integrity and AKT inhibitor VIII, on glucose and glutamine uptake were studied. TCA cycle function. Introduction of a Myc-shRNA hairpin reversed SF188 cells with a Bcl-xL transgene were used in this study to the glutamine dependence of Myc-transformed cells. In addition, prevent apoptosis induced by drug treatment. As presented in Fig. Myc-transformed cells were sensitive to inhibitors of glutamate 2, AKT inhibitor VIII suppressed glucose metabolism and lactate conversion to ␣-ketoglutarate in a Myc-dependent fashion and this production in a dose-dependent fashion. In contrast, neither glu- sensitivity could be reversed by supplying cells with a cell-penetrant tamine metabolism nor ammonia production was inhibited by AKT form of the mitochondrial substrate ␣-ketoglutarate. Taken to- inhibitor VIII. If anything, there was a compensatory up-regulation gether, these results suggest that glutamine addiction can be a direct of glutamine metabolism in response to increasing doses of the consequence of Myc-induced transformation. inhibitor. Similar results were observed using the PI3K inhibitor LY294002 (data not shown). Results The Human Glioma Line SF188 Depends on Glutamine Catabolism to Myc Can Regulate Glutaminolysis. Another oncogene associated with Maintain Viability. When SF188 cells were cultured in the presence a poor prognosis in glial tumors is Myc (14). SF188 cells were of 14C-labeled glutamine, Ͻ15% of the glutamine the cells took up originally isolated from a patient whose tumor displayed amplifi- from the medium was incorporated into newly synthesized protein cation of Myc (15). Western blot analysis of the SF188 cells used in (Fig. 1A). Despite the fact that only a small fraction of the glutamine these studies revealed Myc protein expression in excess of that was used for anabolic synthesis, SF188 glioma cells were unable to observed in proliferating fibroblasts or a tumor cell line lacking CELL BIOLOGY survive in glutamine-deficient medium despite the presence of 25 amplified Myc (Fig. 3A). To determine whether Myc is required for mM glucose in the medium (Fig. 1B). ␣-ketoglutarate is the the maintenance of oxidative glutamine metabolism in SF188 cells, glutamine metabolite that enters the mitochondrial TCA cycle. The the effect of suppressing Myc using shRNA was investigated. SF188 replacement of glutamine with a cell-penetrant form of ␣-ketoglu- cells were transduced with a lentivirus containing shRNA against tarate (dimethyl ␣-ketoglutarate) in the medium suppressed the cell MYC (shMYC) or a lentivirus containing a control shRNA (shC- death observed when the cells were cultured in glutamine-deficient TRL) and the rate of glutamine consumption and ammonia medium (Fig. 1B). These data suggest that glutamine addiction does production was examined. shMYC cells had an Ϸ80% reduction in not result from glutamine’s role as an amide donor in nucleotide their Myc level (Fig. 3A). This level of Myc reduction lead to a biosynthesis or as the source of nitrogen for the maintenance of statistically significant reduction in glutamine consumption (P Ͻ nonessential amino acid production because dimethyl ␣-ketogluta- 0.01) and ammonia production (P Ͻ 0.05) (Fig. 3B). rate is devoid of nitrogen groups and cannot participate in these glutamine-dependent reactions. Myc Activates the Transcription of Genes Required for Glutamine Uptake and Metabolism. Myc’s transforming properties depend on PI3K/AKT Signaling Regulates the Consumption of Glucose but Not of its ability to bind to DNA and modify gene transcription (16). By Glutamine in Glioma Cells. Previous work has demonstrated that the using quantitative RT-PCR (qPCR), we observed that shMYC cells PI3K/AKT pathway can regulate the expression and surface trans- expressed significantly lower levels of the high affinity glutamine

Wise et al. PNAS ͉ December 2, 2008 ͉ vol. 105 ͉ no. 48 ͉ 18783 Downloaded by guest on September 29, 2021 Furthermore, the glutamine uptake in vehicle-treated cells began to reach an asymptote by 6 h, whereas the glutamine metabolized by the 4-hydroxy tamoxifen treated cells increased linearly over the culture period (Fig. 4E). Thus, the ability of Myc-induced cells to metabolize glutamine does not appear to be saturatable over this period as would be predicted for glutaminolysis, which ends not in the cellular accumulation of glutamine-derived metabolites over time, but in the secretion of glutamine-derived lactate into the medium. Despite increasing glutamine consumption, Myc induction did not increase the proliferative expansion of normal MEF under the same conditions. After 24 h of 4-hydroxy tamoxifen or vehicle treatment of cells that were initially plated at 1 ϫ 105 cells per well the day before induction, the control cells increased to 5.2 Ϯ 0.5 ϫ 105 cells per well (mean Ϯ SD) whereas the Myc- induced cells had only increased to 3.5 Ϯ 0.5 ϫ 105 cells per well (mean Ϯ SD).

Myc Diverts Glucose Away from Mitochondrial Metabolism. Previous studies have suggested that proliferating nontransformed cells Fig. 3. Myc activates the transcription of genes involved in glutamine uptake maintain de novo phospholipid biosynthesis from glucose (7). and metabolism. (A) Myc protein is over-expressed in SF188 cells. Western blot However, when MEF stably expressing MycER were treated with reveals over-expression of c-Myc in SF188 glioblastoma cells compared with MEF 4-hydroxy tamoxifen, the use of glucose as a precursor for phos- and another glioblastoma cell line, LN229. (B) Myc is required for glutamine pholipid synthesis was suppressed (Fig. 5A). Concomitantly, an metabolism. SF188 cells were transduced with either lentiviral shRNA against Myc (shMYC) or Luciferase (shCTRL) for 3 days. Glutamine and ammonium levels in the increased amount of the glucose-derived carbon was secreted from medium were analyzed using the Nova Flex and are presented as the mean Ϯ SD the cell as lactate (Fig. 5B). In contrast, the contribution of of triplicate samples. Data from one of five independent experiments are shown. glutamine to phospholipid synthesis in Myc-induced cells was Knockdown of Myc protein is depicted in (A). (C) Myc is required for the expres- maintained and even increased despite increased secretion of sion of the proximal enzymes of glutaminolysis. RNA was extracted from shMYC glutamine-derived lactate (Figs. 4D and 5C). Together, these data and shCTRL cells and quantified using quantitative RT-PCR (qPCR). The bars demonstrate that Myc-induction leads to the diversion of glucose- shown are normalized to a ␤-actin control and represent the mean Ϯ SD of derived pyruvate away from mitochondria, its conversion to lactate, triplicate samples. Representative data from one of two independent experi- and secretion from the cell. As a result, Myc induction enhances ments are shown. EIF1A is included as a negative control. (D) Myc is enriched at cellular dependence on glutamine to maintain phospholipid syn- the regulatory binding sites of genes involved in glutamine uptake. Sheared chromatin from fixed and lysed SF188 cells was immunoprecipitated using the thesis and TCA cycle anapleurosis. antibodies indicated. Precipitated DNA fragments were quantified by qPCR. The data presented are the mean Ϯ SD of triplicate samples. The schematic shows the The Glutamine Addiction Exhibited by SF188 Glioma Cells Is Myc- location of Myc-bound E-box elements within the genomic loci of ASCT2 and SN2. Dependent. The above data suggest that Myc is both necessary and potentially sufficient for the glutaminolytic metabolism exhibited by SF188 cells. To confirm that Myc is also involved in the glutamine importers ASCT2 and SN2 (P Ͻ 0.01) without expressing signifi- addiction observed by these cells, SF188 cells were transduced with cantly lower levels of the control transcript EIF1A (Fig. 3C). either a lentivirus containing a MYC-shRNA (shMYC) or a control Furthermore, when Myc antibodies were used to perform chroma- shRNA (shCTRL). The resulting cells were incubated in glu- tin immuno-precipitation (ChIP), Myc was found to selectively bind tamine-depleted or complete medium. Cells transduced with to the promoter regions of both ASCT2 and SN2 (Fig. 3D). This MYC-shRNA had a statistically significant increase (P Ͻ 0.01) in selectivity was comparable to that of the established Myc target their resistance to glutamine starvation relative to cells transduced CYCLIN D2 (Fig. 3D). Thus, Myc appears to bind to the promoter with a control shRNA (Fig. 6A). elements of glutamine transporters and this binding is associated As a further confirmation that this glutamine addiction is Myc- with enhanced levels of glutamine transporter mRNA. dependent, the cells were treated with aminooxyacetate (AOA), a chemical inhibitor of glutamate-dependent transaminases that con- Myc Activates Glutaminolysis in MEF. The above data demonstrate vert glutamate into ␣-ketoglutarate in the glutaminolytic pathway that Myc transcription contributes to the high level of glutaminolysis (18). AOA selectively induced the death of the cells transduced with exhibited by SF188 glioma cells. We next wanted to determine control shRNA without affecting the viability of cells transduced whether the induction of Myc transcription was sufficient to induce with MYC-shRNA (P Ͻ 0.01). Although AOA is a well- increased glutamine metabolism. To address this question, immor- characterized inhibitor of the transaminases (19), a chemical inhib- talized MEF that stably express a 4-hydroxy tamoxifen-inducible itor can have nonspecific effects on the viability of cells. To confirm MycER construct (17) were analyzed to study the effects of Myc the specificity of AOA’s effects, the ability of dimethyl ␣-ketoglu- activation on glutamine metabolism. Treatment of cells with 4-hy- tarate to reverse the AOA-induced toxicity to SF188 cells was droxy tamoxifen for 24 h resulted in increased levels of the examined. Addition of 7 mM dimethyl ␣-ketoglutarate completely transcripts for not only the glutamine transporter ASCT2, but also suppressed the death induced by AOA treatment of SF188 parental for glutaminase (P Ͻ 0.005), the enzyme that deamidates glutamine and control transduced cells (P Ͻ 0.01) (Fig. 6B and data not to glutamate, resulting in its intracellular capture, and LDH-A (P Ͻ shown). 0.005), which converts glutamine-derived pyruvate into lactate (Fig. 4A). These increases in mRNA levels correlated with enhanced Discussion functional activity. Treatment with 4-hydroxy tamoxifen resulted in The factors that regulate glutamine uptake and metabolism during statistically significant increases in glutamine uptake (P Ͻ 0.05) cell growth and transformation have remained poorly understood. (Fig. 4B), glutaminase flux (P Ͻ 0.005) (Fig. 4C), and production In this manuscript, we provide evidence that oncogenic levels of of glutamine-derived lactate (P Ͻ 0.05) (Fig. 4D) in the MEF Myc reprogram intermediate metabolism, leading to glutamine expressing MycER. As a result, the rate at which glutamine was addiction for the maintenance of mitochondrial TCA cycle integ- consumed from the medium was significantly greater (Fig. 4E). rity. Previous work has demonstrated that LDH-A induction by

18784 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0810199105 Wise et al. Downloaded by guest on September 29, 2021 Fig. 4. Myc activates glutaminolysis in MEF. (A) Oncogenic levels of Myc induce the expression of genes involved in glutaminolysis. qPCR analysis of target genes from total RNA isolated from MEF MycER treated with 200 nM 4-hydroxytamoxifen (4-OHT) or vehicle (EtOH) for 24 h. The bars shown are normalized to an internal ␤-actin control and represent the mean Ϯ SD of triplicate samples. Repre- sentative data from one of three independent experiments are shown. (B) On- cogenic levels of Myc induce glutamine uptake. MEF MycER treated as in (A) were cultured for 1 min with medium supple- 14 mented with [U- C5]glutamine. Uptake of the label was quantified by scintillation counting of the cellular lysate. The data presented are the mean Ϯ SD of triplicate samples. (C) Oncogenic levels of Myc induce flux through glutaminase. MEF MycER treated as in (A) were cultured for 8 h with medium supplemented with L-[␥-15N]glutamine. Glutaminase activity was determined by measuring the 15 ϩ isotopic enrichment of NinNH4 in the culture medium by GC-MS. The bars shown represent the mean Ϯ SD of triplicate cultures. (D) Oncogenic levels of 13 Myc induce the flux of glutamine into lactate. MEF MycER treated as in (A) were cultured for6hinmedium supplemented with 4 mM [U- C5]glutamine. The medium 13 13 13 13 was subsequently removed and analyzed with C NMR spectroscopy. [2,3- C]lactate is metabolically derived from [U- C5]glutamine, while [3- C]lactate is metabolically derived from the natural abundance of [1-13C]glucose and [6-13C]glucose. The data presented are the mean Ϯ SD of triplicate samples. (E) Oncogenic levels of Myc induce the consumption of glutamine from the medium. The glutamine concentration in medium from MEF MycER treated as in (A) was analyzed at the time points shown by the Nova Flex. The data points shown represent the mean Ϯ SD of triplicate samples.

Myc is required for Myc-transformation (20). This results in diver- substrate is glutamine. Myc-transformation leads to conversion sion of glucose-derived pyruvate into lactate. Despite this, Myc- from glucose to glutamine as the oxidizable substrate used to transformed cells display an increased mitochondrial mass and maintain TCA cycle activity and cell viability. Myc binds to the promoters and induces the expression of several key regulatory increased rate of O2 consumption (21). Furthermore, Morrish et al. (22) have reported that Myc-over-expressing cells are exquisitely genes involved in glutaminolytic metabolism. Our studies suggest that supraphysiological levels of Myc associated with oncogenic sensitive to inhibition of the mitochondrial electron transport chain. transformation are both necessary and sufficient for the induction To explain this apparent paradox, they suggested that mitochon- of glutaminolysis to levels that result in glutamine addiction. drial respiration might be maintained by catabolizing alternative Yuneva et al. (23) have reported that some, but not all, Myc bioenergetic substrates. In this article, we report that the alternative transformants were dependent on glutamine. They also demon- CELL BIOLOGY

Fig. 5. Myc diverts glucose away from mitochondrial metabolism in MEF. (A) Oncogenic levels of Myc suppress the contribution of glucose to phospholipid synthesis. MEF MycER treated as in Fig. 4A were cultured with medium supplemented with D-[U-14C]-glucose for 8 h. After the culture period, lipids were harvested and 14C enrichment in phospholipids (PL) was determined by scintilla- Fig. 6. The glutamine addiction exhibited by SF188 glioma cells is Myc- tion counting. The bars shown represent the mean Ϯ SD of triplicate samples. dependent. (A) Myc-suppressed SF188 cells are resistant to glutamine starva- Representative data from one of three experiments are shown. (B) Oncogenic tion. shMYC and shCTRL SF188 cells, described in Fig. 3B, were allowed to plate levels of Myc induce lactate production. The lactate concentration in the medium in the presence of glutamine and then cultured in the absence of glutamine. from MEF MycER treated with 4-OHT or EtOH for 18 h was quantiﬁed by using the Cell viability was determined at the time points shown by trypan blue dye Nova Flex Metabolite Analyzer. Each time point is the mean Ϯ SD of triplicate exclusion. The data points shown represent the mean Ϯ SD of triplicate samples. Representative data from one of three experiments are shown. (C) samples. (B) Myc-suppressed SF188 cells are resistant to an inhibitor of glu- Glutamine’s contribution to phospholipid synthesis is maintained in the presence taminolysis. shMYC and shCTRL SF188 cells, described in Fig. 3B, were allowed of oncogenic levels of Myc. MEF MycER treated as in Fig. 5A were cultured with to plate in the presence of glutamine and then were treated with 500 ␮M medium supplemented with L-[U-14C]-glutamine for 8 h. After the culture period, aminooxyacetate (AOA). Cell viability was determined at the time points lipids were harvested and 14C enrichment in PL was determined by scintillation shown by trypan blue dye exclusion. AOA- treated shCTRL cells were also counting. The bars shown represent the mean Ϯ SD of triplicate samples. Rep- treated with 7 mM dimethyl ␣-ketoglutarate (AOA ϩ ␣-ketoglutarate). The resentative data from one of three independent experiments are shown. data points shown represent the mean Ϯ SD of triplicate samples.

Wise et al. PNAS ͉ December 2, 2008 ͉ vol. 105 ͉ no. 48 ͉ 18785 Downloaded by guest on September 29, 2021 strated that over-expression of Bcl-2 suppressed the death of of the transaminase inhibitor AOA to induce the death of Myc- Myc-transformants deprived of glutamine. Cell types and cell lines transformed cells but not isogenic cells in which Myc is suppressed vary greatly in their level of expression in Bcl-2 family members and by a MYC-shRNA provides evidence that such an intervention this may account for the differences observed between cell lines. would have a selectively toxic effect on Myc-transformed cells. Consistent with this, when SF188 cells were transfected with Bcl-xL, Myc-activation/amplification is one of the most common oncogenic they underwent cell cycle arrest but did not die when deprived of events observed in a wide variety of cancers and is known to drive glutamine (data not shown). Like Yuneva et al. (23), we also found the progression of human lymphomas (28, 29), neuroblastoma (30), that cell-penetrant TCA cycle intermediates could suppress Myc- and small cell lung cancer (31). Despite this, therapeutics that induced apoptosis. Together, these results suggest that glutamine inhibit the transcriptional properties of Myc have so far eluded drug addiction does not result from the use of glutamine as an amine discovery efforts. The identification of Myc’s role in glutaminolysis donor, but rather because glutamine metabolism is essential to may provide a number of enzymatic targets through which to maintain mitochondrial integrity and function in Myc-transformed selectively impair the growth and survival of Myc-transformed cells. tumor cells. Finally, the present results demonstrate that nutrient Whether levels of Myc expression induced in response to mito- uptake in mammalian cells is under distinct and specific regulation genic stimulation also play a critical role in glutamine uptake and as a result of the properties of known oncogenes. Previous results metabolism in the growth of nontransformed cells remains to be have suggested that an activating mutation in PI3K or AKT determined. At lower levels of expression, Myc-induced increases in facilitate the uptake and metabolism of glucose in an mTOR- protein synthesis and cell growth will also stimulate the incorpo- dependent manner (13). In addition, induction of HIF-1 in response ration of glutamine into newly synthesized proteins (24, 25). There to hypoxia or mitochondrial ROS can also reprogram the intracel- are undoubtedly additional signaling pathways that contribute to lular fate of glucose (6). The present studies suggest that oncogenic the regulation of glutamine uptake. Cells lacking oncogenic Myc Myc activation selectively induces addiction to glutamine, the other levels, while not glutamine dependent, still take up sufficient major catabolic substrate used by mammalian cells to maintain glutamine to fuel both nucleotide and protein biosynthesis for cell bioenergetics during cell growth and proliferation. Whether other growth and proliferation. Perhaps that is what is most surprising essential nutrients are under similar control by these or other about the current results. Little of the glutamine uptake stimulated oncogenic signaling pathways remains to be determined. by Myc is used for macromolecular synthesis. Previous 13C-NMR studies found that during glutaminolysis, Ͼ60% of glutamine- Methods derived carbon is released from the cell as either lactate or CO2 Cell Culture and Media. SF188 cells (UC Brain Tumor Research Center, SF, CA) and (12). Although the TCA cycle was also replenished by glutamine, SV40-immortalized MEF stably transfected with MycER (a gift from Drs. AT only 5% of glutamine fluxing through the TCA cycle was incor- Tikhonenko and R.A. Amaravadi of University of Pennsylvania, Philadelphia, PA) porated into fatty acids. Here, we show that only 15% of the were cultured in DMEM (Invitrogen), 10% FBS (Gemini Biosystems), 100 units/ml ␮ glutamine carbon taken up by the cell is incorporated in protein. Penicillin, 100 g/ml Streptomycin, 25 mM glucose, and 6 mM L-glutamine at 37°C ina5%CO2 incubator. To avoid depleting oxygen, all experiments were carried Nevertheless, the additional stimulation of glutaminolysis by onco- out under subconﬂuent conditions. For glutamine starvation experiments, genic levels of Myc results in cellular addiction to glutamine. DMEM without glutamine (Invitrogen) was supplemented with 10% dialyzed FBS The ability of Myc to induce glutaminolysis does have a poten- (Gemini Biosystems). For metabolic tracing experiments, DMEM without glu- tially beneficial effect for the transformed cell. Glutaminolysis tamine and with 10% dialyzed FBS was supplemented with either L-glutamine 13 14 results in the robust production of NADPH, thus providing an that was unenriched or with L-[U- C5]glutamine (Isotec), [U- C5]glutamine (GE energy source for a wide variety of synthetic reactions required for Amersham), and [␥-15N]glutamine (Cambridge Isotope Laboratories). DMEM 14 cell growth (12). It has long been believed that the major source of without glucose (Sigma) was supplemented with [U- C6]glucose (Sigma). To NADPH production during cell growth occurs through the oxida- activate MycER, cells were incubated with 200 nM 4-hydroxytamoxifen for 24 h. tive arm of the pentose phosphate shunt (26). However, recent Lentiviral transduction, qPCR, immunoblotting, and metabolite analysis were evidence suggests that transformed cells exhibiting aerobic glycol- performed as previously described (6, 12, 33, 34). ysis derive the majority of their ribose biosynthesis through the Glutamine Uptake. Cells were incubated with 4 mM glutamine and 1 ␮M nonoxidative arm of the pentose phosphate shunt (27). Under these 14 [U- C5]glutamine in 2 ml of medium in a 6-well plate for 1 min at 37°C in a 5% conditions, G6PD cannot be used to produce a supply of NADPH CO2 incubator. After incubation, the medium was aspirated, and cells were to support macromolecular synthesis of fatty acids or nucleotides. washed three times with unlabeled medium, after which cells were lysed with 200 Without a compensatory mechanism to generate NADPH, de novo ␮l of a 0.2%SDS/0.2N NaOH solution, incubated for 1 h, neutralized with 10 ␮lof nucleotide synthesis using ribose produced in the nonoxidative arm 2N HCl, and analyzed with a beta scintillation counter (PerkinElmer Life of the pentose phosphate shunt would rapidly lead to intracellular Sciences) (32). depletion of NADPH. The ability of Myc to stimulate NADPH production through glutamine-dependent degradation provides the NMR Analysis. Cells were grown to 80% conﬂuency, after which the culture medium was removed and replaced with medium that contained 4 mM transformed cell with a mechanism to produce the quantities of 13 NADPH needed to meet the demands of cell proliferation. [U- C5]glutamine (and no unenriched glutamine) for 6 h. The resulting medium was then analyzed in a 20-mm NMR tube with a 9.4 Tesla spectrometer at 100.66 The data presented here also demonstrate that glutamine uptake MHz. A 90° excitation pulse was applied every 20 seconds (fully relaxed), with is controlled independently of glucose uptake. Although the PI3K/ broad-band decoupling used only during 13C data acquisition (no NOE enhance- AKT pathway plays a major role in regulating glucose uptake, it ment). Spectra were acquired with 16384 points, a bandwidth of 25,000 Hz and does not appear to be required for the uptake and catabolism of 3,000 excitations. Free induction decays were apodized with exponential multi- glutamine in Myc-transformed cells. Thus, the two main bioener- plication (3 Hz line broadening). Peak intensities in Fourier transformed spectra getic substrates used by proliferating cells appear to be under were determined with Nuts NMR (Acorn NMR, Livermore, CA). Carbon-3 of independent control by oncogenic signaling pathways. The remark- lactate derived from glutamine produced a doublet (21.5 and 20.3 ppm) due to able ability of Myc and Akt to cooperate in transforming cells may splitting from 13C at carbon-2, whereas lactate derived from the natural abun- 13 result in part from their ability to complement each other in dance C of glucose, produced a singlet at 20.9 ppm. All peaks were well resolved from each other. stimulating the uptake of these two critical nutrients. In conclusion, the results presented here provide evidence that Lipid Biochemistry. To determine the rate of lipid synthesis from glucose or Myc transformation is associated with induction of a level of glutamine, 1–4 ϫ 105 MEF were plated in 6-well plates and cultured with DMEM 14 glutamine metabolism that results in glutamine addiction. Such supplemented with either 2.2 ␮M D-[U- C6] glucose (GE Amersham) or 0.54 ␮M 14 addiction may ultimately be exploited through the use of inhibitors L-[U- C5]glutamine (GE Amersham), both supplemented to 0.01% of their re- of the enzymes involved in the glutaminolytic pathway. The ability spective unenriched nutrients. After 8 h, cells were trypsinized, washed in PBS,

18786 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0810199105 Wise et al. Downloaded by guest on September 29, 2021 and lysed in 0.4 ml of 0.5% Triton X-100. Total lipids were extracted and phos- tion. The recovered counts were sensitive to 5 ␮g/ml cycloheximide. The glu- pholipids were collected using TLC (33). Total incorporated label in the phospho- tamine consumption rate was then calculated using the GLN2 glutamine assay kit lipid fraction was analyzed with a scintillation counter (PerkinElmer Life Sciences). (Sigma). The data presented are the mean Ϯ SD from four independent experiments. ChIP Analysis. SF188 cells were plated on 15-cm dishes and were fixed in 1% formaldehyde. Chromatin was sheared to an average size of 500–1,000 bp by Gas Chromatography–Mass Spectrometry. Cells were plated at 1.2 ϫ 106 per 6-cm sonication (30 times with 30-s pulses, on a Diagenode Bioruptor). Lysates corre- dish. At 80% confluency, they were fed with 1.5 ml of DMEM containing 4 mM sponding to 5–10 ϫ106 cells were rotated at 4°C overnight with 2 ␮g of polyclonal l-[␥-15N]glutamine (Cambridge Isotope Laboratories). Every 2 h, medium was antibodies specific for c-MYC (sc-764, Santa Cruz Biotechnology) or normal rabbit collected to determine the concentration of NH3 using the Nova Biomedical Flex IgG. Precipitated DNA fragments were quantified using PCR. Analyzer. An aliquot was used to determine isotopic enrichment in NH3 with published methods (35).

Akt Inhibitor. SF188 cells stably expressing Bcl-xL were plated at a density of 4 ϫ 105 cells in a 6-well plate format, and incubated in DMEM with 10% FCS for 36 h AOA Inhibitor Experiments. SF188 cells were treated with AOA (Sigma) at doses before treatment with AKT inhibitor VIII (Calbiochem) at 0, 0.1, 2, 5, and 10 ␮M ranging from 500 nM – 500 ␮M and viability was assessed 24 h post-treatment concentrations in triplicate. After 8 h, medium samples were collected for me- (18). Five hundred ␮M was the lowest dose that killed a significant fraction of the tabolite analysis, and viable cells were counted using trypan blue exclusion. cells. SF188 cells with Myc or control shRNA were replated 3 days post viral transduction. After allowing to plate overnight, cells were treated with 500 ␮M Contribution of Glutamine to Protein Synthesis. SF188 cells were cultured in for 24 h, and then viability was assessed using trypan blue dye exclusion. 14 medium supplemented with 0.01% [U- C5]glutamine relative to unenriched glutamine for 4 h. Cells were washed three times with PBS, extracted using 0.5% ACKNOWLEDGMENTS. We thank members of the Thompson laboratory for Triton-X, acidified using 10% (vol/vol) of 35% (wt/v) sulfosalicylic acid, and then thoughtful discussions and Tullia Lindsten and Tamar Schwartz for their spun down at 13,000 rpm for 15 min at 4°C. The pellet was then resolubilized in review of the manuscript. This work was supported by grants from the National Cancer Institute; the National Institutes of Health, including K08 NaOH at 37°C. 14C incorporated into the protein product were quantified using DK072565 (to R.J.D.) and HD26979 and NS054900 (to M.Y.); the Abramson a scintillation counter (PerkinElmer Life Sciences). Efficiency of protein recovery Family Cancer Research Institute (C.B.T.); a Medical Scientist Training grant 14 was controlled for by calculating the recovery of C BSA (Sigma) added after lysis. and a Cancer Research Institute grant (to D.R.W.). S.B.M. was supported by 14 Recovery of intracellular [U- C5]glutamine in its free form was controlled for by grants from the National Institutes of Health and Commonwealth Universal 14 calculating the recovery of [U- C5]glutamine added just before SSA precipita- Research Enhancement Program, Pennsylvania Department of Health.

1. Warburg O (1956) On the origin of cancer cells. Science 123:309–314. 20. Lewis BC, et al. (2000) Tumor induction by the c-myc Target Genes rcl and lactate 2. Warburg O (1925) u¨ber den Stoffwechsel der Carcinomzelle. J Mol Med 4:534. dehydrogenase A. Cancer Res 60:6178–6183. 3. Warburg O (1956) On respiratory impairment in cancer cells. Science 124:269–270. 21. Li F, et al. (2005) Myc stimulates nuclearly encoded mitochondrial genes and mito- 4. Kroemer G, Pouyssegur J (2008) Tumor cell metabolism: Cancer’s Achilles’ Heel. Cancer chondrial biogenesis. Mol Cell Biol 25:6225–6234. Cell 13:472–482. 22. Morrish F, Neretti N, Sedivy JM, Hockenbery DM (2008) The oncogene c-Myc coordi- 5. Elstrom RL, et al. (2004) Akt stimulates aerobic glycolysis in cancer cells. Cancer Res nates regulation of metabolic networks to enable rapid cell cycle entry. Cell Cycle 64:3892–3899. 7:1054–1066. 6. Lum JJ, et al. (2007) The transcription factor HIF-1alpha plays a critical role in the 23. Yuneva M, Zamboni N, Oefner P, Sachidanandam R, Lazebnik Y (2007) Deficiency in growth factor-dependent regulation of both aerobic and anaerobic glycolysis. Genes glutamine but not glucose induces MYC-dependent apoptosis in human cells. J Cell Biol Dev 21:1037–1049. 178:93–105. 7. Bauer DE, Hatzivassiliou G, Zhao F, Andreadis C, Thompson CB (2005) ATP citrate lyase is 24. Johnston LA, Prober DA, Edgar BA, Eisenman RN, Gallant P (1999) Drosophila myc an important component of cell growth and transformation. Oncogene 24:6314–6322. regulates cellular growth during development. Cell 98:779–790. 8. Coles NW, Johnstone RM (1962) Glutamine metabolism in Ehrlich ascites-carcinoma 25. Iritani BM, Eisenman RN (1999) c-Myc enhances protein synthesis and cell size during cells. Biochem J 83:284–291. B lymphocyte development. Proc Natl Acad Sci USA 96:13180–13185. 9. Eagle H (1955) Nutrition needs of mammalian cells in tissue culture. Science 122:501– 26. Biaglow JE, et al. (2000) G6PD deficient cells and the bioreduction of disulfides: Effects 514. of DHEA, GSH depletion and phenylarsine oxide. Biochem Biophys Res Commun 10. Klimberg V, McClellan JL (1996) Glutamine, cancer, and its therapy. Am J Surgery 273(3):846–852. 172:418–424. 27. Serkova N, Boros LG (2005) Detection of resistance to imatinib by metabolic profiling: 11. Chen MK, Espat NJ, Bland KI, Copeland EM, III, Souba WW (1993) Influence of pro- Clinical and drug development implications. Am J Pharmacogenomics 5:293–302. gressive tumor growth on glutamine metabolism in skeletal muscle and kidney. Ann 28. Dalla-Favera R, et al. (1982) Human c-myc onc gene is located on the region of Surg 217:655–666, discussion 666–667. chromosome 8 that is translocated in Burkitt lymphoma cells. Proc Natl Acad Sci USA 12. DeBerardinis RJ, et al. (2007) Beyond aerobic glycolysis: Transformed cells can engage 79:7824–7827. in glutamine metabolism that exceeds the requirement for protein and nucleotide 29. Taub R, et al. (1982) Translocation of the c-myc gene into the immunoglobulin heavy synthesis. Proc Natl Acad Sci USA 104:19345–19350. chain locus in human Burkitt lymphoma and murine plasmacytoma cells. Proc Natl 13. Edinger AL, Thompson CB (2002) Akt maintains cell size and survival by increasing Acad Sci USA 79:7837–7841. mTOR-dependent nutrient uptake. Mol Biol Cell 13:2276–2288. 14. Ben-Porath I, et al. (2008) An embryonic stem cell-like gene expression signature in 30. Schwab M, et al. (1983) Amplified DNA with limited homology to myc cellular onco- poorly differentiated aggressive human tumors. Nat Genet 40:499–507. gene is shared by human neuroblastoma cell lines and a neuroblastoma tumour. 15. Trent J, et al. (1986) Evidence for rearrangement, amplification, and expression of Nature 305:245–248. c-myc in a human glioblastoma. Proc Natl Acad Sci USA 83:470–473. 31. Nau MM, et al. (1985) L-myc, a new myc-related gene amplified and expressed in human small cell lung cancer. Nature 318:69–73. 16. Dang CV (1999) c-Myc target genes involved in cell growth, apoptosis, and metabolism. CELL BIOLOGY Mol Cell Biol 19:1–11. 32. Bode BP, Souba WW (1994) Modulation of cellular proliferation alters glutamine 17. Thomas-Tikhonenko A, et al. (2004) Myc-transformed epithelial cells down-regulate transport and metabolism in human hepatoma cells. Ann Surg 220:411–424. clusterin, which inhibits their growth in vitro and carcinogenesis in vivo. Cancer Res 33. DeBerardinis RJ, Lum JJ, Thompson CB (2006) Phosphatidylinositol 3-kinase-dependent 64:3126–3136. modulation of carnitine palmitoyltransferase 1A expression regulates lipid metabo- 18. Moreadith RW, Lehninger AL (1984) The pathways of glutamate and glutamine lism during hematopoietic cell growth. J Biol Chem 281:37372–37380. oxidation by tumor cell mitochondria. Role of mitochondrial NAD(P)ϩ-dependent 34. Patel JH, McMahon SB (2007) BCL2 is a downstream effector of MIZ-1 essential for malic enzyme. J Biol Chem 259:6215–6221. blocking c-myc-induced apoptosis. J Biol Chem 282:5–13. 19. Rej R (1977) Aminooxyacetate is not an adequate differential inhibitor of aspartate 35. Brosnan JT, et al. (2001) Alanine metabolism in the perfused rat liver. Studies with 15N. aminotransferase isoenzymes. Clin Chem 23:1508–1509. J Biol Chem 276:31876–31882.

Wise et al. PNAS ͉ December 2, 2008 ͉ vol. 105 ͉ no. 48 ͉ 18787 Downloaded by guest on September 29, 2021