COMMENTARY

Uncoupling the Warburg effect from cancer

Ayaz Najafov and Dario R. Alessi1 Medical Research Council Protein Phosphorylation Unit, College of Life Sciences, University of Dundee, Dundee DD1 5EH, Scotland

remarkable trademark of most tumors is their ability to break A down glucose by at a vastly higher rate than in nor- mal tissues, even when oxygen is copious. This phenomenon, known as the Warburg effect, enables rapidly dividing tumor cells to generate essential biosynthetic building blocks such as nucleic acids, amino acids, and lipids from glycolytic intermediates to permit growth and du- plication of cellular components during Fig. 1. Mechanism by which APC/C-Cdh1 inhibits glycolysis and glutaminolysis to suppress cell pro- division (1). An assumption dominating liferation. APC/C-Cdh1 E3 ligase recognizes KEN-box–containing metabolic , such as PFKFB3 and research in this area is that the Warburg glutaminase-1 (GLS1), and ubiquitinates and targets them for proteasomal degradation. This inhibits effect is specific to cancer. Thus, much of glycolysis and glutaminolysis, leading to decrease in metabolites that can be assimilated into biomass, the focus has been on uncovering mecha- thereby suppressing proliferation. nisms by which cancer-causing mutations influence to stimulate glycol- ysis. This has lead to many exciting dis- inhibits glycolysis in proliferating noncancer active. Consistent with destruction being coveries. For example, the p53 tumor cells by mediating the degradation of two key mediated by APC-C-Cdh1, ablation of suppressor can suppress glycolysis through metabolic enzymes, namely 6-phosphofructo- the KEN-box prevents degradation of PFKFB3 (9, 10) and glutaminase-1 (8). its ability to control expression of key 2-kinase/fructose-2,6-bisphosphatase iso- Inhibiting the proteasomal-dependent metabolic , such as phosphoglycerate form 3 (PFKFB3) (9, 10) and glutaminase-1 degradation with the MG132 inhibitor mutase (2), synthesis of cytochrome C oxi- (Fig. 1) (8). PFKFB3 potently stimulates gly- markedly increases levels of ubiquitinated dase-2 (3), and TP53-induced glycolysis and colysis by catalyzing the formation of fruc- PFKFB3 and glutaminase-1 (8). More- apoptosis regulator (TIGAR) (4). Many tose-2,6-bisphosphate, the allosteric activator over, overexpression of Cdh1 to activate cancer-causing mutations lead to activa- of 6-phosphofructo-1-kinase (11). Glutamin- fi APC/C-Cdh1 decreases levels of PFKFB3 tion of the Akt and mammalian target of ase-1 is the rst in glutaminolysis, as well as glutmaninase-1 and concomi- rapamycin (mTOR) pathway that pro- converting glutamine to lactate, yielding bio- tantly inhibited glycolysis, as judged by foundly influences metabolism and ex- synthetic intermediates required for cell decrease in lactate production. This effect pression of metabolic enzymes to promote proliferation (12). is also observed when cells were treated glycolysis (5). Strikingly, all cancer cells APC/C is a cell cycle-regulated E3 with a glutaminase-1 inhibitor (6-diazo-5- but not nontransformed cells express ubiquitin ligase that promotes ubiquitina- oxo-L-norleucine) (8). The final evidence a specific splice variant of , tion of a distinct set of cell cycle proteins supporting the authors’ hypothesis is that termed M2-PK, that is less active, leading containing either a D-box (destruction box) proliferation and glycolysis is inhibited to the build up of phosphoenolpyruvate or a KEN-box, named after the essential after shRNA-mediated silencing of either (6). Recent work has revealed that re- Lys-Glu-Asn motif required for APC rec- ognition (13). Among its well-known sub- PFKFB3 or glutaminase-1 (8). duced activity of M2-PK promotes a These results are interesting, because unique glycolytic pathway in which phos- strates are crucial cell cycle proteins, such as cyclin B1, securin, and Plk1. By ubiq- unlike most recent work in this area, phoenolpyruvate is converted to pyruvate Colombo et al. (8) link the Warburg effect by a histidine-dependent phosphorylation uitinating and targeting its substrates to 26S proteasome-mediated degradation, to the machinery of the cell cycle that is of , promoting present in all cells rather than to cancer- assimilation of glycolytic products into APC/C regulates processes in late mitotic stage, exit from mitosis, and several events driving mutations. Further work is re- biomass (7). However, despite these ob- quired to properly define the overall im- in G1 (14). The Cdh1 subunit is the KEN- servations, one might imagine that the portance of this pathway, which has thus fi box binding adaptor of the APC/C ligase Warburg effect need not be speci c for far only been studied in a limited number and is essential for G1/S transition. Im- cancer and that any normal cell would of cells. It would also be of value to un- portantly, APC/C-Cdh1 is inactivated at need to stimulate glycolysis to generate dertake a more detailed analysis of how fi the initiation of the S-phase of the cell suf cient biosynthetic materials to fuel the rate of glycolysis and other metabolic expansion and division. Recent work by cycle when DNA and cellular organelles ’ pathways vary during the cell cycle of Salvador Moncada s group published in are replicated at the time of the greatest normal and cancer cells, by quantita- PNAS (8) and other recent work from the need for generation of biosynthetic mate- tively assessing the concentration of key same group (9, 10) provides exciting evi- rials. APC/C-Cdh1 is reactivated later at dence supporting the idea that the War- the mitosis/G1 phase of the cell cycle when burg effect is also required for the pro- there is a lower requirement for bio- Author contributions: A.N. and D.R.A. wrote the paper. liferation of noncancer cells. mass generation. The authors declare no conflict of interest. The key discovery was that the anaphase- Both PFKFB3 (9, 10) and glutaminase-1 See companion article on page 18868 in issue 44 of volume promoting complex/cyclosome-Cdh1 (8) possess a KEN-box and are rapidly 107. (APC/C-Cdh1), a master regulator of the degraded in nonneoplastic lymphocytes 1To whom correspondence should be addressed. E-mail: transition of G1 to S phase of the cell cycle, during the cell cycle when APC/C-Cdh1 is [email protected].

www.pnas.org/cgi/doi/10.1073/pnas.1014047107 PNAS Early Edition | 1of2 Downloaded by guest on September 30, 2021 metabolites at different stages of the through the pentose phosphate pathway, cer and normal cells are required. In the cell cycle. Strikingly, several other meta- resulting in insufficient regeneration of case of PFKFB3 and glutaminase-1, the bolic enzymes involved in biomass gener- reduced glutathione (10). These findings best way to address this issue would be to ation possess putative KEN-box motifs, imply that in nonproliferating cells such as generate knockin models, in which the namely , malate neurons, high APC/C-Cdh1 activity pre- KEN-boxes of PFKFB3 and glutaminase-1 dehydrogenase-1, and acetyl-CoA carboxyl- serves sufficient glucose levels to maintain are disrupted, to prevent recognition and ase-1 (8). It would also be interesting to antioxidant status at the expense of its ubiquitination by APC/C-Cdh1. The ex- determine whether expression and ubiq- utilization for biosynthetic purposes. pectation is that glycolysis should be uitination of these enzymes is also regulated As mentioned above, one of the key ways stimulated in the knockin models. It would by APC/C-Cdh1. It would also be worth- that the p53 tumor suppressor inhibits also be fascinating to define how these while to investigate whether any mutations glycolysis is to promote the expression of KEN-box–ablating mutations impacted on reported in human cancers disrupt the KEN- the TIGAR enzyme that breaks down other metabolic and survival pathways and boxes of PFKFB3, glutaminase-1, or any whether they predispose to development of the other enzymes mentioned above, Colombo et al. link of cancer. Similarly, it would be interesting because this might be expected to enhance to perform analogous studies with other glycolysis and proliferation. It would also the Warburg effect to metabolic knockin/knockout models, for be appealing to search for mutations that example, those that lack expression of affect the ability of APC/C-Cdh1 to rec- the machinery of the PFKFB3, TIGAR, or PK-M2 or that ex- ognize the KEN-box of PFKFB3/gluta- press a mutant or phosphoglycerate mu- minase-1 and/or ubiquitinate these cell cycle that is present tase incapable of catalyzing the conversion targets. It would also be interesting to of phosphoenolpyruvate to pyruvate. Such analyze the mechanism by which APC/ in all cells. experiments would provide much needed C-Cdh1 is regulated and whether activity data to define the relative importance of and/or levels of this complex are con- trolled by signaling pathways disrupted in fructose-2,6-bisphosphate (4). Loss of p53 these different pathways in regulating gly- cancer, such as the Akt–mTOR or p53 in tumors results in reduced TIGAR lev- colysis in normal as well as cancer cells. pathways. Interestingly, Cdh1 is reportedly els, thereby elevating fructose-2,6-bi- This knowledge could be exploited to phosphorylatedinvivoonserineand sphosphate and stimulating glycolysis. gain better insights into how drugs might threonine, as well as tyrosine residues on at PFKFB isoforms can also be directly acti- be developed that counteract the effects least 15 residues (http://www.phosphosite. vated by phosphorylation by kinases, such of cancer-causing mutations on stimulat- org/). The relevance of these modifica- as Akt, that are inappropriately activated ing glycolysis. It would be desirable to tions or upstream protein kinases that in many cancers (15). Taken together with develop compounds that inhibit the War- phosphorylate these sites is unknown. the findings that PFKFB3 is a key regula- burg effect in cancer cells that do not in- fl Nonproliferating neurons have high tor of glycolysis, these more recent studies uence the metabolism of normal tissues. levels of APC/C-Cdh1 activity, thereby provide valuable insights into how fruc- If the Warburg effect in normal pro- suppressing PFKFB3 expression and tose-2,6-bisphosphate levels are con- liferating cells is mainly mediated through maintaining low fructose-2,6-bisphosphate trolled. They also emphasize the im- the effects of APC/C-Cdh1 on regulating levels (10). Inappropriate stimulation of portance of fructose-2,6-bisphosphate in expression of metabolic enzymes, this may glycolysis in neuronal cells by overexpres- regulating flux through the glycolytic provide hope that it would be possible to sion of PFKFB3 or reducing expression of pathway, a phenomena that has been ap- develop drugs that suppress glycolysis in Cdh1 resulted in marked oxidative stress preciated for many decades (11). cancer cells without interfering with the and apoptosis (10). This was attributed to Firmer genetic data validating the im- ability of the APC/C-Cdh1 pathway to the activation of glycolysis reducing the portance of the different mechanisms by promote the Warburg effect in normal pool of glucose available to be oxidized which glycolysis can be stimulated in can- proliferating cells.

1. Hsu PP, Sabatini DM (2008) Cancer cell metabolism: 8. Colombo SL, et al. (2010) Anaphase-promoting com- 12. DeBerardinis RJ, Cheng T (2010) Q’s next: The diverse Warburg and beyond. Cell 134:703–707. plex/cyclosome-Cdh1 coordinates glycolysis and glu- functions of glutamine in metabolism, cell biology and 2. Kondoh H, et al. (2005) Glycolytic enzymes can modu- taminolysis with transition to S phase in human T cancer. Oncogene 29:313–324. – late cellular life span. Cancer Res 65:177 185. lymphocytes. Proc Natl Acad Sci USA 107:18868– 13. Pfleger CM, Kirschner MW (2000) The KEN box: An APC 3. Matoba S, et al. (2006) p53 regulates mitochondrial 18873. recognition signal distinct from the D box targeted by respiration. Science 312:1650–1653. 9. Almeida A, Bolaños JP, Moncada S (2010) E3 ubiquitin Cdh1. Genes Dev 14:655–665. 4. Bensaad K, et al. (2006) TIGAR, a p53-inducible regula- ligase APC/C-Cdh1 accounts for the Warburg effect by 14. Fang G, Yu H, Kirschner MW (1998) Direct binding of tor of glycolysis and apoptosis. Cell 126:107–120. linking glycolysis to cell proliferation. Proc Natl Acad CDC20 protein family members activates the ana- 5. Robey RB, Hay N (2009) Is Akt the “Warburg kinase”?- Sci USA 107:738–741. Akt-energy metabolism interactions and oncogenesis. phase-promoting complex in mitosis and G1. Mol Cell 10. Herrero-Mendez A, et al. (2009) The bioenergetic and Semin Cancer Biol 19:25–31. 2:163–171. 6. Christofk HR, et al. (2008) The M2 splice isoform of antioxidant status of neurons is controlled by continu- 15. Deprez J, Vertommen D, Alessi DR, Hue L, Rider MH pyruvate kinase is important for cancer metabolism ous degradation of a key glycolytic enzyme by APC/C- (1997) Phosphorylation and activation of heart 6- – and tumour growth. Nature 452:230–233. Cdh1. Nat Cell Biol 11:747 752. phosphofructo-2-kinase by protein kinase B and other 11. Hue L, Rider MH (1987) Role of fructose 2,6-bisphos- 7. Vander Heiden MG, et al. (2010) Evidence for an alter- protein kinases of the insulin signaling cascades. JBiol native glycolytic pathway in rapidly proliferating cells. phate in the control of glycolysis in mammalian tissues. Chem 272:17269–17275. Science 329:1492–1499. Biochem J 245:313–324.

2of2 | www.pnas.org/cgi/doi/10.1073/pnas.1014047107 Najafov and Alessi Downloaded by guest on September 30, 2021