Deregulation of Glucose Transporter 1 and Glycolytic Gene Expression by c-Myc*

Unlike normal mammalian cells, which use oxygen to generate energy, cancer cells rely on glycolysis for energy and are therefore less dependent on oxygen. We previously observed that the c-Myc oncogenic transcription factor regulates lactate dehydrogenase A and induces lactate overproduction. We, therefore, sought to determine whether c-Myc controls other genes regulating glucose metabolism. In Rat1a fibroblasts and murine livers overexpressing c-Myc, the mRNA levels of the glucose transporter GLUT1, phosphoglucose isomerase, phosphofructokinase, glyceraldehyde-3-phosphate dehydrogenase, phosphoglycerate kinase, and enolase were elevated. c-Myc directly transactivates genes encoding GLUT1, phosphofructokinase, and enolase and increases glucose uptake in Rat1 fibroblasts. Nuclear run-on studies confirmed that the GLUT1 transcriptional rate is elevated by c-Myc. Our findings suggest that overexpression of the c-Myc oncoprotein deregulates glycolysis through the activation of several components of the glucose metabolic pathway.

Unlike normal mammalian cells, which use oxygen to generate energy, cancer cells rely on glycolysis for energy and are therefore less dependent on oxygen. We previously observed that the c-Myc oncogenic transcription factor regulates lactate dehydrogenase A and induces lactate overproduction. We, therefore, sought to determine whether c-Myc controls other genes regulating glucose metabolism. In Rat1a fibroblasts and murine livers overexpressing c-Myc, the mRNA levels of the glucose transporter GLUT1, phosphoglucose isomerase, phosphofructokinase, glyceraldehyde-3-phosphate dehydrogenase, phosphoglycerate kinase, and enolase were elevated. c-Myc directly transactivates genes encoding GLUT1, phosphofructokinase, and enolase and increases glucose uptake in Rat1 fibroblasts. Nuclear run-on studies confirmed that the GLUT1 transcriptional rate is elevated by c-Myc. Our findings suggest that overexpression of the c-Myc oncoprotein deregulates glycolysis through the activation of several components of the glucose metabolic pathway.
To form a three-dimensional multicellular spheroid mass, neoplastic cells alter their metabolism such that they are able to survive and grow in the hostile microenvironments created by the decreased blood flow found in tumor vasculature (1,2). The most striking feature of tumor cells is the production of large amounts of lactic acid, which is due to the glycolytic conversion of glucose to lactic acid even in the presence of oxygen (3). This is often accompanied by an increased rate of glucose transport (4 -6).
Glucose is a major regulator of gene transcription. In particular, it stimulates transcription of genes encoding glycolytic and lipogenic enzymes in adipocytes and hepatocytes through the carbohydrate response element (ChoRE), 1 a 5Ј-CACGTG-3Ј motif (7)(8)(9)(10)(11). The ChoRE is similar to the core binding site for the transcription factors USF2 (12), which is implicated in glucose metabolism, TFE3, and the hypoxia-inducible transcription factor (HIF). Hence, the ChoRE serves to integrate physiological signals through transcription factors to regulate glucose metabolism.
During tumor formation, adaptation to hypoxia may be mediated by the HIF-1 family of transcription factors, which induce angiogenesis and other metabolic changes. (2,13,14). It is notable that glucose transport and transporter mRNA are induced in cells transformed by ras or src oncogenes (5). The c-myc oncogene is activated in a variety of pathways that are important in controlling cell growth and tumorigenesis. (15)(16)(17)(18). Intriguingly, the ChoRE sequence matches the core E-box (5Ј-CACGTG-3Ј) binding site for c-Myc, which binds E-boxes of target genes to stimulate transcription (2,18,19). Previous work showed that c-Myc directly up-regulates the expression of the lactate dehydrogenase gene (LDH-A) (20), which is important in the transformed phenotype (anchorage-independent growth) of cells that overexpress c-Myc (20,21). In addition to LDH-A, we report here the deregulation of GLUT 1 and several glycolytic genes by c-Myc.

EXPERIMENTAL PROCEDURES
Cell Culture and Transfection-Rat fibroblasts were cultured in 5% CO 2 at 37°C in DMEM supplemented with 10% fetal bovine serum (FBS; Life Technologies, Inc.) and antibiotics. Rat1a and Myc-transformed Rat1a-Myc fibroblasts were as described previously (22). Human lymphoid cells were cultured in Iscove's modified Dulbecco's medium with 10% FBS and antibiotics. Cells lacking c-myc (HO15), heterozygous HET15, and parental lines (TGR) (a gift of John Sedivy, Brown University) were cultured in DMEM supplemented with 10% FBS and antibiotics (23). HO15 cells were transfected with either MLVmyc or MLV empty vectors (20) using Lipofectin (Life Technologies, Inc.) according to the manufacturer's protocol. The human GLUT1 cDNA (24) was kindly provided by Drs. M. Mueckler (Washington University, St. Louis, MO) and C. Heilig (The Johns Hopkins University, Baltimore, MD).
A Rat1 cell line expressing a fusion protein of c-Myc and the human estrogen receptor (MycER) (a gift of J. M. Bishop, University of California, San Francisco) was grown in DMEM with 10% FBS, penicillin, and streptomycin (25). Myc activity was induced when confluent My-cER cells were treated with 0.25 M 4-hydroxytamoxifen (4-HOTM; Research Biochemicals, Natrick, MA) for the times indicated as described (20,26). To block protein synthesis, 10 M cycloheximide (CHX) was added to the cells 30 min prior to 4-HOTM treatment.
Adenoviral in Vivo Gene Transfer-Animal studies were approved by The Johns Hopkins University School of Medicine Animal Use and Care Committee. Adenoviral constructs (a gift from W. El-Deiry, University of Pennsylvania, Philadelphia, PA) containing either LacZ (Ad/LacZ) or c-myc (Ad/c-myc) coding sequence were as described (27). Two-monthold male BALB/c mice, weighing 20 -25 g, were intravenously injected with 4 ϫ 10 9 plaque-forming units (pfu) of either Ad/LacZ or Ad/c-myc. Mice were then sacrificed by exposure to CO 2 on days 3, 4, and 5 after viral injection. Their livers were quickly removed and processed for * This work was supported by the Alexander and Margaret Stewart Trust Fund, a research grant from the Maryland American Cancer Society, and National Institutes of Health Grant CA 51497. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

RNAs or frozen in liquid nitrogen.
RNA Analysis-Total RNA was isolated by guanidium thiocyanate lysis followed by cesium chloride centrifugation. 15-g aliquots of RNA were used in Northern blot analysis (22). IMAGE Consortium cDNA clones (28) were obtained (Research Genetics, Inc.), and the inserts were isolated and 32 P-labeled by Random-Primer synthesis using a Random-PrimeII kit (Stratagene).

2-[ 3 H]Deoxyglucose (2-DG)
Uptake-Uptake experiments were conducted on confluent Rat1a and MycER cells. Cells were treated with 0.25 M 4-HOTM (Research Biochemicals, Natrick, MA) 1 h before 0.1 mM 2-DG with 10 nM [ 3 H]2-DG (final concentration) were added. At the end of a 10-min incubation, cells were washed with ice-cold phosphatebuffered saline. Cells were than lysed by the addition of 2 ml of 10 mM NaOH containing 0.1% Triton X-100, and a 100-l aliquot was assayed for 3 H by liquid scintillation counting (30). The protein content was measured by the method of Bradford (31).

Myc Induces GLUT1 and Glycolytic Gene Expression in Myctransformed Fibroblasts and in Hepatocytes in
Vivo-To identify genes whose expression is deregulated in the presence of increased ectopic c-Myc, c-Myc-transformed Rat1a cells were studied (Fig. 1A). Using this system of isogenic cell lines, GLUT1, phosphofructokinase (PFK), glyceraldehyde-3-phosphate dehydrogenase (GAPDH), phosphoglucose isomerase (GPI), phosphoglycerate mutase (PGM), and enolase showed elevated mRNA levels in c-Myc-transformed cells as compared with the nontransformed Rat1a cells. The expression levels of aldolase A, aldolase C, and triose-phosphate isomerase (TPI) were decreased, and the expression levels of hexokinase I, hexokinase II, and phosphoenolpyruvate carboxykinase were unchanged. Rat vimentin was used as a control (32).
To determine whether c-Myc could induce GLUT1 and glycolytic gene expression in vivo, we injected adenoviruses expressing either LacZ or c-Myc into mice. Tail-vein injection of adenoviruses has the advantage of very high hepatic clearance, which yields highly efficient gene delivery to the liver (33). Four days after injection, 40% of hepatocytes were positive for ␤-galactosidase in LacZ adenovirus-treated animals (data not shown). 2 c-Myc protein levels were 10-fold elevated in Myc adenovirus-treated animals (Fig. 1B). GLUT1, PFK, GAPDH, PGM, and enolase mRNAs were significantly induced in the Myc-expressing livers as compared with LacZ-expressing livers at 4 and 5 days after injection of viruses (Fig. 1C). These observations suggest that transient expression of Myc in vivo induces hepatic expression of the same glycolytic genes that are induced in the Rat1a system.
Glycolytic Genes That Are Direct Targets of c-Myc-To determine whether the genes displaying increased or decreased levels of mRNA expression are transcriptionally activated by Myc, we used a previously described Rat1 fibroblast line expressing a protein that fuses Myc to the estrogen receptor ligand binding domain (MycER). With exposure of cells expressing the MycER protein to estrogenic compounds such as 4-hydroxytamoxifen (4-HOTM), the ligand-bound MycER protein translocates to the nucleus. The MycER protein then activates target genes without requiring new intervening protein synthesis (26,34). Hence, prior exposure of MycER cells to the protein synthesis inhibitor cycloheximide (CHX) would not block activation or repression of direct target genes by 4-HOTM. Activation of MycER by 4-HOTM caused induction of GLUT1, PFK, and enolase mRNAs in the presence of CHX at 2 and 4 h ( Fig. 2A). Note that mRNA levels are diminished in the CHX ϩ TM samples at the 6 h time point; this is likely due to decreased MycER protein (10% of 0 h level) with CHX treatment (data not shown). GAPDH, GPI, and PGM expression were not increased in the MycER system. Thus, GAPDH, GPI, and PGM do not meet this criterion for direct targets of c-Myc. Among the potentially down-regulated genes identified in Rat1a-Myc cells, only TPI mRNA moderately decreased in MycER cells exposed to 4-HOTM. However, a decrease in TPI mRNA was also seen 2 S. Kim, Q. Li, C. V. Dang, and A. Lee, unpublished observations.

c-Myc Activates Glycolytic Genes 21798
with CHX treatment alone. The ribosomal phosphoprotein (36B4) mRNA level was shown to be unresponsive to estrogen and independent of cell cycle progression (35,36) and was therefore used as a control. To account for experimental variation, a separate tamoxifen induction experiment was performed, and GLUT1 expression in MycER cells was determined (Fig. 2B). Following c-Myc induction, GLUT1 expression behaves in a manner that is consistent with a direct c-Myc target.
To determine whether 4-HOTM itself may directly increase the expression of GLUT1, PFK, and enolase independent of the MycER protein, we treated Rat1a cells lacking MycER with 4-HOTM, with or without CHX, and performed Northern analysis. The mRNA levels of these genes were unchanged upon addition of 4-HOTM and cycloheximide in the absence of the MycER protein (Fig. 2C). Hence, the induction of enolase, GLUT1, and PFK expression in the Rat1a-MycER cells by 4-HOTM is dependent on MycER activity.
Expression of GLUT1 and Glycolytic Genes in Cells Lacking c-Myc-To further authenticate GLUT1, PFK, and enolase as c-Myc targets, we sought to determine their expression in Rat1 fibroblasts lacking c-myc (Fig. 3). All three transcripts showed decreased levels in c-myc null cells as compared with wild type parental cells. The expression of PFK and enolase are elevated in c-myc null cells rescued by a constitutively overexpressed c-Myc (HO15-Myc) when compared with empty vector control transfected myc null cells (HO15-MLV). GLUT1 expression, however, was only slightly elevated in the c-Myc-overexpressing HO15 cells as compared with control MLV transfected cells, which display elevated GLUT1 as compared with the parental HO15 cells. Replicate experiments yielded similar results, in which the HO15-MLV cells have elevated expression of all three genes as compared with the HO15 parental cell line. The cause for this elevation in our control HO15-MLV cells is not known. Thus, in this genetically defined system Myc levels parallel the expression of these three target genes.
Among the three up-regulated genes, GLUT1 is most frequently implicated in tumorigenesis (35)(36)(37)(38). We therefore chose to characterize GLUT1 further. Nuclear run-on experiments demonstrated an enhanced transcriptional rate of GLUT1 in Rat1a-Myc cells as compared with Rat1a fibroblasts (Fig. 4A). The nuclear run-on signals for enolase and PFK were low and insufficient for interpretation (not shown). These results, nevertheless, underscore the activation of GLUT1 by c-Myc at the transcriptional level.
We further sought to determine whether GLUT1 expression parallels that of c-myc in Burkitt's lymphoma cells that are characterized by c-myc gene activation by chromosomal translocation. Both c-Myc-transformed lymphoblastoid (CB33-Myc) and Burkitt's lymphoma cell lines, Ramos and ST486, have elevated c-Myc protein levels (20) that are associated with elevations of GLUT1 mRNA levels compared with the nontransformed lymphoblastoid CB33 cells (Fig. 4B). We thus observed a correlation between c-Myc expression and the endogenous levels of GLUT1 mRNA in these Myc-transformed human lymphoid cells.
Increased Glucose Uptake in Myc-transformed Cells-To determine whether increased Myc activity influences glucose transport, [ 3 H]2-DG uptake was studied in MycER Rat1 cells as well as Rat1a cells lacking the MycER system. 4-HOTM increased [ 3 H]2-DG uptake in MycER cells, while it had no influence in Rat1a cells (Fig. 5). DISCUSSION Normal mammalian cells use oxygen to generate energy from glucose and other metabolites through oxidative phosphorylation. In conditions of oxygen deprivation, normal cells rely on glycolysis to generate energy by converting glucose to lactic acid. Neoplastic transformation, however, alters glucose metabolism with enhanced conversion of glucose to lactic acid, thereby making tumor cells less dependent on oxygen. The molecular basis for physiological and pathological regulation of glucose metabolism is beginning to emerge with the identification of transcription factors that regulate glycolytic genes (2). The ability of c-Myc, USF1, and HIF-1 transcription factors to regulate the LDH-A promoter through a Myc consensus binding site 5Ј-CACGTG-3Ј suggests that these transcription factors converge onto common cis elements (5Ј-RCGTG-3Ј) that are found in regulatory sequences of glycolytic genes (20). Because hypoxia physiologically induces glycolytic gene expression through HIF-1 binding sites, it stands to reason that glycolysis might also be activated by USF or c-Myc. Hence, we sought to determine whether c-Myc regulates genes that are also regulated by HIF-1.
In this study, we observe that the glucose transporter GLUT1, PFK, GAPDH, GPI, PGM, and enolase are up-regulated by c-Myc in fibroblasts. In addition, we used adenovirusmediated gene transfer to the liver as a means of studying in vivo gene expression and determined that these genes are also up-regulated by c-Myc in vivo. Our findings agree with a previous study of transgenic mice that also suggests the in vivo regulation of hepatic glycolysis by c-myc (41), although GLUT2, but not GLUT1, was induced in the transgenic livers. While our observation supports the hypothesis that c-Myc induces GLUT1 expression in the liver in vivo, the physiological significance of this induction is unclear because GLUT2 is the dominant form of hepatic glucose transporter. The use of transgenic mice, however, cannot establish whether c-Myc regulates certain genes directly. In our study, only GLUT1, PFK, and enolase behave as direct target genes in the MycER system. While some HIF-1 and c-Myc target genes overlap, the two sets of targets remain distinct. For example, the mRNA levels for HKI and adolase are unchanged or decreased by c-Myc, but are hypoxia inducible (2,13,14). GPI is increased by c-Myc but is unaffected by hypoxia. Furthermore, we have reported that c-Myc overexpression down-regulates VEGF, a HIF-1-mediated hypoxia-inducible gene (42).
We observed that c-Myc directly induces GLUT1 and increases glucose uptake in the Rat1 MycER cells. In addition to LDH-A, GLUT1 is an intriguing c-Myc target when its role in oncogenesis is considered (37,38). Elevation of GLUT1 and c-Myc RNA are among the earliest changes in gene expression after Ha-ras T24 transformation of Rat1a fibroblasts (39). Reduction of GLUT1 expression through antisense GLUT1 RNA suppresses NIH 3T3 cell transformation by N-ras (40). Hence, the induction of GLUT1 expression by c-Myc may play an important role in tumor glucose metabolism.
The specific roles of PFK and enolase in tumorigenesis are less clear. Glycolytic flux is controlled at multiple steps of glycolysis, such that interruption of a non-rate-limiting step may affect overall flux (43). PFK is a rate-limiting glycolytic enzyme due to its intricate allosteric responses to ATP/ADP ratios and therefore may be critically important for the control of glycolytic flux. Enolase has not been directly implicated in human cancers; however, it is intriguing to note that a negative transactivating factor of the c-myc promoter (myc-binding protein 1) may be identical to enolase A (44,45). While the role for enolase in regulating c-myc expression remains to be confirmed and further studied, this finding suggests that the induction of enolase by c-Myc may increase glycolysis and also serves as a negative feedback onto c-myc gene expression.
In summary, we observe that while both c-Myc and HIF-1 up-regulates GLUT1 and glycolytic gene expression, the sets of target genes for these transcription factors are distinct. HIF-1 physiologically induces glycolytic gene expression in hypoxic cells that also undergo growth arrest. By contrast, c-Myc stimulates glucose uptake, glycolysis, and overall metabolism as well as activates the cell cycle machinery, which are all necessary for cell proliferation (46). The widespread deregulation of the c-myc gene in human cancers, therefore, may be a major contributor to the enhanced tumor glycolysis known as the Warburg effect, which may confer growth advantage to tumor cells deprived of oxygen.