C-Myc Activates Multiple Metabolic Networks to Generate Substrates for Cell-Cycle Entry

Home , Carboxylation

Oncogene (2009) 28, 2485–2491 & 2009 Macmillan Publishers Limited All rights reserved 0950-9232/09 $32.00 www.nature.com/onc REVIEW c-Myc activates multiple metabolic networks to generate substrates for cell-cycle entry

F Morrish1, N Isern2, M Sadilek3, M Jeffrey4 and DM Hockenbery1

1Clinical Research Division, Fred Hutchinson Cancer Research Center, Seattle, WA, USA; 2WR Wiley Environmental Molecular Sciences Laboratory, High Field NMR Facility, Paciﬁc Northwest National Laboratory, Richland, WA, USA; 3Department of Chemistry, University of Washington, Seattle, WA, USA and 4Advanced Imaging Research Center, University of Texas Southwestern, Dallas, TX, USA

Cell proliferation requires the coordinated activity of and supply of metabolic intermediates required for cell cytosolic and mitochondrial metabolic pathways to division are less well understood. Of the known provide ATP and building blocks for DNA, RNA and transcription factors exerting cell-cycle control, a likely protein synthesis. Many metabolic pathway genes are candidate for this regulatory function is the c-myc targets of the c-myc oncogene and cell-cycle regulator. oncogene, due to the diversity of known target genes However, the contribution of c-Myc to the activation of involved in both cell-cycle and metabolic pathways cytosolic and mitochondrial metabolic networks during (Zeller et al., 2006). Induction of Myc occurs rapidly in cell-cycle entry is unknown. Here, we report the metabolic response to growth factors stimulating cell-cycle entry, fates of [U-13C] glucose in serum-stimulated mycÀ/À and with mRNA levels increasing 20-fold within the first 2 h myc þ / þ fibroblasts by 13C isotopomer NMR analysis. We after serum addition (Dean et al., 1986). Global analysis demonstrate that endogenous c-myc increased 13C labeling of downstream targets reveals that Myc potentially of ribose sugars, purines and amino acids, indicating regulates up to 15% of human genes through direct partitioning of glucose carbons into C1/folate and pentose binding to target gene sequences (Zeller et al., 2006), and phosphate pathways, and increased tricarboxylic acid may have even broader effects on gene expression cycle turnover at the expense of anaplerotic flux. Myc through global chromatin remodeling (Knoepfler expression also increased global O-linked N-acetylgluco- et al., 2006). samine protein modification, and inhibition of hexosamine Notably, Myc regulates key genes in glycolysis, C1/ biosynthesis selectively reduced growth of Myc-expressing folate, purine, pyrimidine and lipid metabolism (Dang cells, suggesting its importance in Myc-induced prolifera- et al., 2006; Zeller et al., 2006). In addition, Myc targets tion. These data reveal a central organizing function for genes required for mitochondrial DNA replication and the Myc oncogene in the metabolism of cycling cells. The transcription (Li et al., 2005). Hence, Myc supports the pervasive deregulation of this oncogene in human cancers bigenomic transcriptional regulation necessary for may be explained by its function in directing metabolic mitochondrial biogenesis, a key event during cell-cycle networks required for cell proliferation. entry (Mandal et al., 2005; Morrish et al., 2008). In Oncogene (2009) 28, 2485–2491; doi:10.1038/onc.2009.112; summary, the network of known Myc gene targets published online 18 May 2009 suggests that Myc is involved in orchestrating the changes in cell metabolism necessary for cell-cycle entry. Keywords: NMR; isotopomer; stable isotope labeling; However, to the best of our knowledge, there have been O-linked N-acetylglucosamine; glucose no studies of central carbon metabolic flux or carbon partitioning associated with physiological Myc expression. Our goal in this study was to address this question.

Introduction Results and discussion Cell proliferation is an essential component of normal growth and development, and as such, is highly Myc increases oxidative metabolism of glucose during regulated by hormonal and environmental factors. cell-cycle entry Although our knowledge of cell-cycle regulatory genes Rat1A fibroblasts with mycÀ/À and myc þ / þ genotypes is advanced, the links between the cell-cycle machinery were incubated with [U-13C]-labeled glucose starting at 12 h after serum addition to quiescent cells. The Correspondence: Dr DM Hockenbery, Clinical Research Division, metabolic fates of 13C atoms can be assessed in a wide Fred Hutchinson Cancer Research Center, 1100 Fairview Avenue range of metabolites based on the resonant absorption North, D2-190, Seattle, WA 98109, USA. signals of 13C nuclei when placed in an external magnetic E-mail: [email protected] 13 Received 9 November 2008; revised 1 February 2009; accepted 18 field. The positions of C tracer in the carbon backbone February 2009; published online 18 May 2009 of metabolites vary according to the proportional fluxes Metabolic flux analysis of Myc-dependent cell-cycle entry F Morrish et al 2486 of alternate metabolic pathways, which can be derived mentary Figure S1). This was confirmed by quantitative by measurements of the relative abundance of the mass isotopomer analysis using LC/MS/MS (liquid different labeling patterns. NMR spectroscopy was chromatography/mass spectrometry/mass spectrome- performed on cell extracts after 4 h of labeling to try). We detected a 20% increase in the fractional approximate steady state for analysis of 13C-labeled contribution of 13C-labeled glutamate to the total metabolites. This allowed us to compare the metabolic glutamate pool in myc þ / þ cells compared to mycÀ/À þ / þ profiles during the G0- to S-phase transition of myc cells (Figure 1c). Glutamate is in equilibrium with a- cells, which enter S phase at 16 h, and mycÀ/À cells, which ketoglutarate and can be used to assess tricarboxylic have delayed S-phase entry (Schorl and Sedivy, 2003; acid (TCA)cycle flux using the program tcaCALC Tollefsbol and Cohen, 1990). (UT Southwestern Medical Center, Dallas, TX, USA) Figure 1a shows representative 1H-decoupled, 13C (Malloy et al., 1990). In our experiments, the multiplet NMR spectra from cell extracts. 13C-incorporation was resonances of glutamate were well resolved (Supplemen- observed for multiple metabolites, with the most tary Figure S1)and the areas of the glutamate C2, C3 prominent peaks corresponding to glutamate and lactate and C4 resonances were used to calculate relative flux of resonance signals. Inspection of the glutamate spectra glucose carbons into the TCA cycle by pyruvate indicated higher levels of 13C-labeled glutamate and dehydrogenase (PDH)and pyruvate carboxylase (PC). glutamine in myc þ / þ compared to mycÀ/À cells (Supple- Results of this analysis, illustrated in Figure 1b,

TMPSA myc-/-

6 3 2

8 7 4 5 1312 9 14 11 10 1

200 175 150 125 100 75 50 25 0

TMPSA

myc+/+ 2 4 3 13 8 14 6 12 9 7 11 10 5 1

200 175 150 125 100 75 50 25 0

1.2 70.0% myc-/- 1 myc+/+ 60.0% p<0.05 0.8 50.0% 0.6 40.0% 0.4 30.0%

Relative value 0.2 20.0% 0 Acetyo-CoA Oxaloacetate PDH PC Acetyl-CoA 10.0% derived from derived from derived from % Fractional Contribution 0.0% glucose pyruvate endogenous oxidation carboxylation substrates myc-/- myc+/+ Figure 1 13C NMR analysis of 13C glucose metabolism in mycÀ/À and myc þ / þ cells during cell-cycle entry. (a) 1H-decoupled 13C spectra. Key: (1)C3 alanine, (2)C3 lactate, (3)C3 glutamate, (4)C4 glutamate, (5)C2 glycine, (6)C5 proline, (7)C2 glutamate, (8)C2 lactate, (9)ribose and sugar moieties, (10)C4 0 NXP, (11)C2 adenine, (12)C1 glutamate, (13)C5 glutamate, (14)C1 lactate. TMPSA is internal standard. NXP refers to the mono-, di- or triphosphate form of any nucleotide. (b)tcaCALC-derived relative ﬂuxes. (c)Percentage contribution of 13C-labeled glutamate to total glutamate by LC/MS/MS mass isotopomer analysis.

Oncogene Metabolic flux analysis of Myc-dependent cell-cycle entry F Morrish et al 2487 demonstrate an approximately sevenfold greater PDH C2 position in the adenine base, originating from N10- flux in myc þ / þ cells compared to mycÀ/À cells. We also formyl-tetrahydrofolate, as well as increased incorpora- evaluated the contributions of [U-13C] glucose and tion of 13C into the ribose moiety of nucleotides, endogenous substrates to acetyl-CoA entering the demonstrated by a threefold increase in ribose C40 TCA cycle (Figure 1b). Analysis by program tcaCALC labeling (Figure 2d). Folate is required for one-carbon indicated that both [U-13C] pyruvate and endogenous transfer reactions in purine and thymidylate biosynth- unlabeled substrates each supplied approximately twice esis, as well as epigenetic modification of DNA and as much acetyl-CoA in myc þ / þ cells compared to mycÀ/À proteins. Expression of B50% of genes involved in cells. folate metabolism is increased following Myc induction In contrast, there is higher relative metabolism of (Morrish et al., 2008), and SHMT2 is one of only two pyruvate by the anaplerotic enzyme PC for the mycÀ/À Myc target genes results shown to complement growth cells, resulting in a substantially higher PC/PDH flux defects in mycÀ/À cells (Nikiforov et al., 2002). Treat- ratio (4.3 compared to 0.37 for myc þ / þ cells). In keeping ment with the dihydrofolate reductase inhibitor, metho- with their higher relative PC flux, mycÀ/À cells had a trexate, inhibited growth of myc þ / þ cells more sixfold greater contribution of 13C pyruvate to the total effectively than mycÀ/À cells (Figure 3c). Similar results oxaloacetate pool by pyruvate carboxylation. These were obtained using mycÀ/À cells transfected with results identify significant qualitative differences be- inducible Myc (mycÀ/À MycER)(Figure 3c).These tween myc þ / þ and mycÀ/À cells in substrate partitioning results provide additional evidence of the close relation- and pathway flux into the TCA cycle during preparation ship between nucleotide synthesis and Myc-dependent for S-phase entry, consistent with observed differences cell proliferation (Mannava et al., 2008), and establish in serum-induced changes in mitochondrial mass and direct links between increased glucose metabolism and respiration (Morrish et al., 2008). nucleotide synthesis in Myc-expressing cells. The Myc-dependent alterations in TCA cycle and related fluxes can be explained by changes in Myc- Myc increases phosphocholine levels regulated gene expression. Induction of Myc (MycER) Phosphocholine is an intermediate in phosphatidylcho- in mycÀ/À cells results in temporal repression of PDK2 line (PtdCho)synthesis and is required for membrane and PDK4 (Morrish et al., 2008), which would lead to biosynthesis during cell-cycle entry (Jackowski, 1994). increased pyruvate oxidation by decreasing inhibitory Because phosphocholine is difficult to quantitate in 13C phosphorylation of PDH. Expression of PC is also spectra due to spectral overlap, we used 1D-1H spectra temporally decreased by Myc induction, reinforcing the to evaluate phosphocholine content. We found a allocation of pyruvate to supply the mitochondrial threefold increase in phosphocholine levels in serum- acetyl-CoA pool. These results, together with the effects stimulated myc þ / þ cells compared to mycÀ/À cells of Myc on mitochondrial bioenergetics during cell-cycle (Figure 3a). Myc induction increases choline kinase entry, point to a key function for Myc in augmenting expression (Morrish et al., 2008), and elevated phos- mitochondrial capacity and directing substrates to phocholine levels are associated in neuroblastomas with mitochondria during the transition from maintenance N-Myc amplification (Peet et al., 2007). The rate- energy requirements of quiescent cells to the active limiting enzyme in PtdCho synthesis, CTP:phosphocho- biosynthesis of cycling cells (Morrish et al., 2008). line cytidylyltransferase (CCT/PCYT1A), is also an Myc target gene (Li et al., 2003), suggesting Myc increases glycerophospholipid synthesis. The increased TCA cycle Myc increases glucose carbon flux into branch metabolic þ / þ pathways of glycolysis flux observed in myc cells may also have a function Metabolism of glucose supplies intermediates for in increased glycerophospholipid synthesis by generat- amino-acid synthesis, C1/folate metabolism, and pen- ing acetyl groups required for de novo fatty acid tose phosphate and hexosamine biosynthetic pathways, synthesis. In preliminary results, we observed that fatty acid synthesis from [U-13C] glucose is increased in in addition to lactate generation. We found evidence for þ / þ increased glycolytic flux as 13C labeling of lactate was myc cells (Morrish, unpublished data). These results mycÀ/À greater in myc þ / þ cells compared to mycÀ/À cells may be relevant to the delayed cell-cycle entry of (Figure 2a). This result was expected, as multiple cells, as net accumulation of PtdCho near the glycolytic genes are targets of Myc (Osthus et al., G1/S boundary is necessary for membrane synthesis 2000), and we have previously shown increased conver- (Jackowski, 1994). PtdCho is also a source of lipid second messengers involved in cell-cycle progression sion of glucose to pyruvate and lactate in Rat1A myc þ / þ cells compared to mycÀ/À cells (Morrish et al., 2008). (Hunt and Postle, 2004). When we evaluated the spectra for additional 13C- enriched intermediary metabolites, we found increased O-GlcNAc posttranslational modification is regulated levels of 13C-labeled alanine and glycine, derived from by Myc pyruvate and the glycolytic intermediate 3-phosphogly- Glycolytic (fructose-6-phosphate)and TCA cycle inter- cerate, respectively (Figure 2b and c). Glycine is a mediates (glutamine, acetyl-CoA)are substrates for the precursor for de novo purine synthesis as well as a hexosamine biosynthesis pathway (HBP). Because the carbon donor for C1/folate metabolism required for 13C data demonstrated increased glycolytic and TCA nucleotide biosynthesis. We observed 13C labeling of the cycle flux in serum-stimulated myc þ / þ cells, we looked

Oncogene Metabolic ﬂux analysis of Myc-dependent cell-cycle entry F Morrish et al 2488

Figure 2 13C labeling of selected metabolites in mycÀ/À and myc þ / þ cells. 13C enrichment at individual carbons (left). Data are calculated from integrated peak areas relative to internal standard. Mean of three experiments. Signiﬁcant differences (*Po0.05, **Po0.001)assessed by Student’s t-test. Corresponding NMR spectra (right)for ( a)lactate, ( b)alanine, ( c)glycine and ( d)NXP ribose C40 and adenine C2.

for global differences in serum-induced protein mod- tions and an early response to serum was not observed ification by O-linked N-acetylglucosamine (O-GlcNAc) (Figure 3c). Hence, early Myc-associated increases in in myc þ / þ and mycÀ/À cells. Immunoblotting of cell protein O-GlcNAcylation are not explained by upregu- extracts using an O-GlcNAc-specific antibody detected lation of OGT expression. multiple bands, with an overall temporal increase after To examine the function of O-GlcNAcylation in Myc- serum addition in myc þ / þ cells. Several bands were dependent proliferation, we treated cells with 6-diazo-5- detected at lower intensity in serum-starved cells, oxo-L-norleucine (DON), an inhibitor of glutamine:- whereas others appeared with serum stimulation fructose-6-phosphate aminotransferase (GFAT), the (Figure 3b). The number of O-GlcNAc-modified pro- first and rate-limiting enzyme of the hexosamine path- teins detected in mycÀ/À cells was reduced and the serum way. Both myc þ / þ and induced mycÀ/À MycER cells response significantly attenuated. were more sensitive to growth inhibition at 48 h, Increased O-GlcNAc protein modification may suggesting an important function for the HBP during result from increased expression of modifying enzyme, proliferation in myc þ / þ cells (Figure 3d). O-GlcNAc O-GlcNAc transferase (OGT), and/or increased hexo- protein modification is a dynamic process analogous to samine synthesis. OGT protein expression was reduced phosphorylation (Hart et al., 2007). Interference with as demonstrated by immunoblotting in myc þ / þ cells O-GlcNAc modification has been shown to disrupt compared to mycÀ/À cells under serum-deprived condi- cell-cycle progression (Slawson et al., 2005). Several

Oncogene Metabolic ﬂux analysis of Myc-dependent cell-cycle entry F Morrish et al 2489

Figure 3 Metabolic pathways required for cell-cycle progression in myc þ / þ cells. (a)Proton spectra for phosphocholine (3.2 p.p.m.). (b)Immunoblots showing O-GlcNAcylated proteins and O-GlcNAc transferase (OGT)expression with serum stimulation. Cyclin E and tubulin are positive and loading controls. Densitometry values are indicated below each lane (ImageQuant, GE Healthcare, Piscataway, NJ, USA). (c)Dose–response assay of cell growth by HO33342 staining for methotrexate and 6-diazo-5-oxo-L-norleucine (DON). TGR represents myc þ / þ cells. Relative growth rates of untreated cell lines: mycÀ/À 0.69, mycER 0.83. Representative of three experiments in triplicate.

HBP genomic loci are bound by Myc, including GNS the TCA cycle, the pentose phosphate pathway and and NAGPA (Li et al., 2003)and HBP genes are enriched C1/folate metabolism (Figure 4). The obligate function in expression profiling of Myc-induced pancreatic neo- of mitochondria in C1/folate metabolism raises the plasia (Lawlor et al., 2006). In addition to transcriptional possibility that Myc-induced mitochondrial biogenesis regulation of the HBP, Myc may increase the supply of provides for biosynthesis of key intermediate essential substrates for hexosamine biosynthesis as a metabolites in addition to bioenergetic requirements consequence of the increased glycolytic and TCA cycle for cell-cycle progression. This extensive programming fluxes demonstrated by 13C isotopomer analysis. of metabolism may be unique to Myc. For example, In conclusion, our results demonstrate the broad HIF-1a directs increased glycolytic metabolism, effects of endogenous Myc expression on glucose carbon but at the expense of anabolic synthesis and mitochon- flux in both catabolic and anabolic pathways, in drial respiration (Lum et al., 2007). Finally, the response to serum stimulation, including glycolysis, observation of Myc effects on O-GlcNAc posttransla-

Oncogene Metabolic ﬂux analysis of Myc-dependent cell-cycle entry F Morrish et al 2490

Figure 4 Myc regulation of glucose metabolism may provide key intermediates, energy and reducing power for cell proliferation. 13C-labeled metabolites from this study are boxed. Dashed arrows link mitochondrial metabolites with reversible (double-ended arrows) or irreversible (single arrows)reactions. Both glycolysis and the mitochondrial tricarboxylic acid (TCA)cycle generate metabolic intermediates, in addition to ATP, providing building blocks for protein, lipid and nucleic acid synthesis. Posttranslational protein modiﬁcation may be subject to substrate-level control, including extramitochondrial acetyl-CoA, derived from intramitochondrial pyruvate and fatty acid metabolism, and glucosamine-6-phosphate, derived from fructose-6-phosphate and glutamine.

tional protein modification provides an interesting P41 RR02301, the MDL database (www.mdl.imb.liu.se)and example of the potential intersection of metabolome the Human Metabolome database (www.hmbd.ca). A portion and transcriptome networks for signal amplification and of this research was performed at EMSL, a national scientific integration. user facility sponsored by the Department of Energy’s Office of Biological and Environmental Research at Pacific North- west National Laboratory. This work was funded by Acknowledgements RO1CA106650-02 (DH). Development of the program tca- CALC (University of Texas Southwestern Medical Center) We thank John Sedivy for cell lines. This work utilized the was supported by H47669-16, a Department of Veterans MMC database supported by NIH grants R21 DK070297 and Affairs Merit Review Award to CR Malloy, and RR02584.

References

Dang CV, O’Donnell KA, Zeller KI, Nguyen T, Osthus RC, Li F. Knoepﬂer PS, Zhang XY, Cheng PF, Gafken PR, McMahon SB, (2006). The c-Myc target gene network. Semin Cancer Biol 16: Eisenman RN. (2006). Myc inﬂuences global chromatin structure. 253–264. EMBO J 25: 2723–2734. Dean M, Levine RA, Ran W, Kindy MS, Sonenshein GE, Campisi J. Lawlor ER, Soucek L, Brown-Swigart L, Shchors K, Bialucha CU, (1986). Regulation of c-myc transcription and mRNA abundance by Evan GI. (2006). Reversible kinetic analysis of Myc targets in vivo serum growth factors and cell contact. J Biol Chem 261: 9161–9166. provides novel insights into Myc-mediated tumorigenesis. Cancer Hart GW, Housley MP, Slawson C. (2007). Cycling of O-linked beta- Res 66: 4591–4601. N-acetylglucosamine on nucleocytoplasmic proteins. Nature 446: Li F, Wang Y, Zeller KI, Potter JJ, Wonsey DR, O’Donnell KA 1017–1022. et al. (2005). Myc stimulates nuclearly encoded mitochondrial Hunt AN, Postle AD. (2004). Phosphatidylcholine biosynthesis inside genes and mitochondrial biogenesis. Mol Cell Biol 25: the nucleus: is it involved in regulating cell proliferation? Adv 6225–6234. Enzyme Regul 44: 173–186. Li Z, Van Calcar S, Qu C, Cavenee WK, Zhang MQ, Ren B. (2003). A Jackowski S. (1994). Coordination of membrane phospholipid global transcriptional regulatory role for c-Myc in Burkitt’s synthesis with the cell cycle. J Biol Chem 269: 3858–3867. lymphoma cells. Proc Natl Acad Sci USA 100: 8164–8169.

Oncogene Metabolic flux analysis of Myc-dependent cell-cycle entry F Morrish et al 2491 Lum JJ, Bui T, Gruber M, Gordan JD, DeBerardinis RJ, Covello KL Osthus RC, Shim H, Kim S, Li Q, Reddy R, Mukherjee M et al. et al. (2007). The transcription factor HIF-1alpha plays a critical (2000). Deregulation of glucose transporter 1 and glycolytic gene role in the growth factor-dependent regulation of both anaerobic expression by c-Myc. J Biol Chem 275: 21797–21800. and aerobic glycolysis. Genes Dev 21: 1037–1049. Peet AC, McConville C, Wilson M, Levine BA, Reed M, Dyer SA Malloy CR, Sherry AD, Jeffrey FM. (1990). Analysis of tricarboxylic et al. (2007). 1H MRS identifies specific metabolite profiles acid cycle of the heart using 13C isotope isomers. Am J Physiol 259: associated with MYCN-amplified and non-amplified tumour sub- H987–H995. types of neuroblastoma cell lines. NMR Biomed 20: 692–700. Mandal S, Guptan P, Owusu-Ansah E, Banerjee U. (2005). Schorl C, Sedivy JM. (2003). Loss of protooncogene c-Myc function Mitochondrial regulation of cell cycle progression during develop- impedes G1 phase progression both before and after the restriction ment as revealed by the tenured mutation in Drosophila. Dev Cell 9: point. Mol Biol Cell 14: 823–835. 843–854. Slawson C, Zachara NE, Vosseller K, Cheung WD, Lane MD, Hart Mannava S, Grachtchouk V, Wheeler LJ, Im M, Zhuang D, GW. (2005). Perturbations in O-linked beta-N-acetylglucosamine Slavina EG et al. (2008). Direct role of nucleotide metabolism in protein modification cause severe defects in mitotic progression and C-MYC-dependent proliferation of melanoma cells. Cell Cycle 7: cytokinesis. J Biol Chem 280: 32944–32956. 2392–2400. Tollefsbol TO, Cohen HJ. (1990). The protein synthetic surge in Morrish FM, Neretti N, Sedivy JM, Hockenbery DM. (2008). The response to mitogen triggers high glycolytic enzyme levels in human oncogene c-Myc coordinates regulation of metabolic networks to lymphocytes and occurs prior to DNA synthesis. Biochem Med enable rapid cell cycle entry. Cell Cycle 7: 1054–1066. Metab Biol 44: 282–291. Nikiforov MA, Chandriani S, O’Connell B, Petrenko O, Kotenko I, Zeller KI, Zhao X, Lee CW, Chiu KP, Yao F, Yustein JT et al. Beavis A et al. (2002). A functional screen for Myc-responsive genes (2006). Global mapping of c-Myc binding sites and target reveals serine hydroxymethyltransferase, a major source of the one- gene networks in human B cells. Proc Natl Acad Sci USA 103: carbon unit for cell metabolism. Mol Cell Biol 22: 5793–5800. 17834–17839.

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

Oncogene