- Email: [email protected]
The hypoxic core: a possible answer to the cancer paradox
BBRC Biochemical and Biophysical Research Communications 299 (2002) 676–680 www.academicpress.com
The hypoxic core: a possible answer to the cancer paradox Michael Guppy* Biochemistry and Molecular Biology, School of Biomedical and Chemical Sciences, University of Western Australia, 35 Stirling Highway, Crawley, WA 6009, Australia Received 4 November 2002
Abstract There are many differences, at all levels of organization, between cancerous and normal cells. Two of these (oxygen delivery and glucose metabolism) are related and manifest as low intercellular oxygen tensions (pO2 ) and a glycolytic metabolic profile in tumours and/or cancer cells. It is becoming increasingly apparent that these characteristics of cancer combine to enhance both the survival and aggressiveness of cancer cells, and that they can adversely impact on some forms of treatment. But they are also exploited in current strategies of detection and monitoring of cancers. These are therefore characteristics with important implications for the crucial balance between the aggression and growth characteristics of a tumour, and our ability to detect and treat it. The interactions and the hierarchy of events leading to these manifestations are complex, not fully understood, and involve a pivotal and intriguing paradox. This paradox results in a seemingly contradictory state in which the most dangerous tumours are those that are the most hypoxic, but also those that are the most angiogenic. This review is a synthesis of the available data into a feasible hypothesis which offers a possible resolution of this paradox and provides a testable paradigm for tumour behaviour. Ó 2002 Elsevier Science (USA). All rights reserved. Keywords: Cancer; Tumour; Hypoxia; Glycolysis; PET; HIF; pO2 ; FDG; Angiogenesis; VEGF
Cancer cells are hypoxic in vivo Vaupel et al. [1] have produced a comprehensive compilation of the data on oxygen tensions in normal and tumour tissues. In human primary tumours as compared to normal tissue; the microcirculation is compromised, tissue oxygenation is not regulated according to metabolic demand, and the pO2 of the tissue is markedly lower, linked to size (a tumour with a volume of 2 ml is extremely hypoxic), often close to 0 mm Hg, and heterogeneous. Metastatic lesions are generally similar, but may be even more hypoxic, and in recurrent tumours there is a higher proportion of the tumour that is hypoxic compared to the respective primary tumour.
Cancer cells may have upregulated glycolysis Consistent with the concept of a hypoxic cancer cell is the view that the metabolism of cancer cells is more * Fax: +61-8-9380-1148. E-mail address: [email protected].
glycolytic than that of normal cells. Glycolysis was first linked to cancer 70 years ago [2] and the data were summarized in a 1956 paper [3] in which Warburg suggested that aerobic glycolysis was a phenomenon peculiar to tumours. (Note that the general perception of glycolysis is that it is activated only in the absence of oxygen, the Pasteur effect. But the same pathway commonly operates (a) in most cells and (b) in the presence of oxygen. It produces lactate, is termed aerobic glycolysis, and can account for a significant proportion of ATP turnover in the presence of oxygen [4].) Investigations into tumour intermediary metabolism over the past 70 years have led to a general perception that cancer metabolism, even under normoxic conditions, is predominantly fueled by glucose (via aerobic glycolysis). This issue however is confusing and controversial as (a) the current perception is that cancer cells inherently undergo aerobic glycolysis, a scenario that does not necessarily invoke hypoxia, (b) there is no obvious or consistent evidence from studies using recent technology that the glycolytic enzymes or glucose transporters are overexpressed in cancers (e.g., 2D-PAGE [5], proteomic analysis [6], and cDNA micro array [7]), (c) what little
0006-291X/02/$ - see front matter Ó 2002 Elsevier Science (USA). All rights reserved. PII: S 0 0 0 6 - 2 9 1 X ( 0 2 ) 0 2 7 1 0 - 9
M. Guppy / Biochemical and Biophysical Research Communications 299 (2002) 676–680
rigorous and comprehensive data there are show that the ATP production in transformed cells under normoxic conditions is mostly oxidative [8,9], and (d) it has been shown in Chinese hamster lung fibroblasts that a high rate of aerobic glycolysis is not required for the malignant phenotype [10]. Nevertheless, the current perception (which may be correct) is that a cancer cell in vivo is hypoxic and makes ATP by glycolysis. Such a scenario has important implications for the detection, monitoring, and treatment of cancer, and for our understanding of the paradoxical behaviour of cancers in vivo.
Detection, monitoring, and treatment The perceived glycolytic character of cancer cells is being increasingly exploited by the glucose-based positron emission tomography (PET) detection technique which is based upon the assumption that tumours take up glucose more rapidly than the surrounding tissue. The compound used in this technique is 18-fluorodeoxyglucose (FDG). The interest in this technique is demonstrated by the 134 papers containing the words ÔcancerÕ and ÔpositronÕ in Current Contents between January and June, 2002. There is no doubt that this is an effective detection technique and it is now also being used to monitor treatment and to predict outcomes. But there are many known factors that influence FDG uptake and some of these can be confounding. FDG uptake is influenced by glucose concentration, by hypoxia, by the heterogeneity of the tissue (proportion of viable and necrotic cells), by the volume of the tissue, by fibrosis and by the inflammatory response. In addition, it may be that some malignancies have lower inherent FDG uptake than others, or are associated with greater levels of inflammatory change, and there is evidence to suggest that cancers handle fluoroglucose differently from glucose [11,12]. As a result of these uncertainties, (a) the efficacy of PET varies with the cancer type and stage [11,12], (b) there can be significant false negatives and positives, for example, in breast tissue [13], and (c) correlations between FDG uptake and a variety of putatively related parameters can be either absent, surprisingly weak, or totally contradictory between studies [14,15]. Due to these limitations, research is currently being directed towards more specific and sensitive molecules to use with the PET technology. The issue of a link between glycolysis and cancer is therefore key to the FDG PET technique, and as discussed above, has yet to be resolved. For example, is the PET technique effective because cancer cells are inherently glycolytic, or because they are hypoxic and thus, as a result of a Pasteur effect, glycolytic? In this regard it is certainly the case that FDG uptake increases under hypoxia in MCF-7 cells [16]. If the PET technique is
677
effective only for hypoxic cancers, might this explain why the efficacy varies with the cancer type, size, and stage? Another possible scenario is that cancer cells take up glucose rapidly (via the glucose transporter Glut-1), but convert it to lipid instead of processing it through glycolysis. This scenario would render the PET technique effective, but would not require an upregulation of glycolysis in cancer tissue. Many cancers do have unusually high levels of fatty acid synthase [17], and there is evidence from studies in which the transporter is directly measured to suggest that (a) increased Glut-1 expression is associated with some types of tumours [18] and (b) that its expression correlates inversely with survival [19]. But the Glut-1 issue is unresolved as its expression can be extremely variable in a particular tumour type [20], and the data on the correlation of Glut-1 expression with FDG uptake are inconsistent [14,15,21]. Hypoxia also has implications for cancer treatment, as hypoxic cells are radiation resistant [22]. The reason for this resistance is that only under normoxia does oxygen react with and result in the permanent incorporation of the DNA radical produced by the radiation. Under hypoxic and reducing conditions the damage can be reversed. As a result, disease-free survival time after radiotherapy is related to the pO2 of the tumour. This has led to various treatment strategies such as attempts to oxygenate tumours during radiation, and the search for compounds that are converted to toxic radicals only under hypoxic and reducing conditions [22,23].
The paradox: the relation between pO2 , angiogenesis, and prognosis Radiation therapy aside, patient survival is also directly related to the average pO2 of the tumour. But paradoxically, survival is inversely related to the degree of tumour angiogenesis, a process that should lead to an increase in the pO2 . This paradox has been specifically noted several times over the last four years [24–27], and the current explanation is that the abnormal angiogenesis which is typical of tumours [1] does not result in increased functional perfusion and the associated increase in pO2 [25,26]. However Giatromanolaki and Harris [27] raise the possibility that aggressive cancers might outgrow their blood supply, suggesting that hypoxia may be related to increased metabolic demand, rather than, or as well as, an abnormally differentiated vasculature. I would like to propose an explanation that builds on the ideas of Giatromanolaki and Harris [27] and involves the hypoxia-inducible factor (HIF-1). The HIF transcriptional system is part of the process by which mammalian cells respond to oxygen availability. The mechanism appears to be conserved among metazoans as a similar complex mediates the response to hypoxia in Caenorhabditis elegans [28]. HIF-1 regulates the
678
M. Guppy / Biochemical and Biophysical Research Communications 299 (2002) 676–680
expression of at least 30 genes via a hypoxia response element. HIF-1 is a heterodimer comprising a constitutively expressed b subunit and an a subunit whose expression is regulated by pO2 . Regulation of the expression of the a subunit is through a degradation pathway initiated by a proline hydroxylase which is sensitive to pO2 . The proline hydroxylated subunit is targetted by a ubiquitin-protein ligase. HIF-1a is therefore degraded under normoxic conditions, but stable under hypoxic conditions, and can appear within minutes of a cell becoming hypoxic [26,29–31]. Rapidly accumulating evidence suggests that HIF-1a is a significant factor in the establishment, development, growth, and metastasis of lethal cancers. Its expression is correlated with tumour grade and the von Hippel-Lindau tumour suppressor plays a key role in the oxygen-dependent degradation of HIF-1a [31]. Until recently, the general consensus was that HIF-1a is not usually detected in normal tissue. But this view is changing as our detection techniques improve [32], and as it is found that HIF-1a expression is also regulated, under normoxic conditions, by peptides such as insulin [33] and stimuli such as mechanical stress [34]. Nevertheless, HIF-1a is detected in most malignant, but not benign tumours, and many of the products of genes transactivated by HIF-1a are overexpressed in human tumours. These include proteins involved in adaptation to hypoxia, angiogenic proteins that support tumour vascularization and proliferation (e.g., VEGF), and proteins that putatively play a role in metastasis and invasion [25,35]. HIF-1a therefore links hypoxia, angiogenesis, the cancer phenotype, and intriguingly, aerobic glycolysis, as there is recent evidence to suggest that HIF-1a can accumulate under aerobic conditions if pyruvate is being produced from glucose [36]. But this only partially resolves the paradox mentioned previously, and in some ways re-emphasizes it, as HIF-1a expression correlates with vascularity [25]. So how does a hypoxic cancer, which is synthesizing HIF-1a, stimulating the production of VEGF, and promoting angiogenesis, remain in this seemingly contradictory state? This is a crucial question as this state characterizes the most aggressive, fast-growing cancers with the poorest prognoses. Recent data from confrontation cultures consisting of non-tumour and tumour spheroids may offer an answer [37]. The data show that angiogenesis in the non-tumour tissue results in improved oxygen supply, down-regulation of HIF-1a and VEGF, but no change in growth rate. In contrast, upon vascularization of the tumour spheroid, the growth rate increases dramatically, the pO2 of the tumour decreases, and there is a concomitant upregulation of HIF-1a and VEGF. The authors suggest that the initial angiogenesis of the tumour enables the outer layer of the tumour to increase its oxidative metabolism and to proliferate, and that it does so at the expense of the core which remains hypoxic, but not
necrotic. The core continues to express HIF-1a, VEGF, and all the other proteins whose genes are targets of HIF-1a, while the outer layer is perfused and grows rapidly. These are the aforementioned paradoxical tumours; irregularly perfused, hypoxic but chronically angiogenic, aggressive, fast growing, metastasizing, and expressing HIF-1a and VEGF. The crucial character of an aggressive cancer is therefore the hypoxic core, so a feasible hypothesis is that cancer cells are glycolytic only under hypoxic conditions, they are not inherently glycolytic. Therefore not all cancers will be glycolytic and the contribution of glycolysis to ATP turnover will vary with the degree of hypoxia. This is consistent with the variable nature of the FDG PET data and with the data from the two aforementioned studies of cancer metabolism [8,9]. But an integral part of this hypothesis must also be the fact that the cells in the hypoxic core adapt to hypoxia, rather than becoming necrotic, and the mechanism by which this occurs would be one of the roles of the multifaceted HIF-1a. The abnormal expression of HIF1a in tumours has already been mentioned above. But as well as this quantitative difference between cancer and normal tissue, the adaptation to hypoxia may involve a relationship between pO2 and the expression of HIF-1a which is specific to cancer cells.
Conclusion There are four questions that need to be addressed. 1. What is the metabolic profile of cancer cells, under normoxia and hypoxia? The perception, as noted above, is that cancer cells are glycolytic, even under normoxic conditions. But despite the general confidence in the literature and the efficacy of the PET technique, the data are equivocal. 2. Is there a hypoxic core in tumours, or perhaps more importantly, is there a highly oxidative shell, and what proteins are being produced in these two areas? 3. If there is a hypoxic core, are the cells specifically adapted to hypoxia compared to those in the oxidative shell, or are all cancer cells preadapted to survive hypoxia? 4. When, with regard to pO2 , does HIF-1a protein appear in the cells in the hypoxic core of a tumour? There has been one study relating HIF-1a expression to pO2 in HeLa cells [38], showing that HIF-1a is expressed only below 6% oxygen (note that oxygen levels in hypoxic tumours in vivo are extremely variable, but are commonly below 3% [1]). But it is not known whether this relationship is different between normal and tumour cells. These data are necessary if we are to understand cancer biology and the cancer process. This will in turn
M. Guppy / Biochemical and Biophysical Research Communications 299 (2002) 676–680
lead to an understanding of the mechanisms that form the basis of many of the current diagnostic and treatment strategies, and will in turn facilitate the improvement of these strategies and suggest new ones.
[16]
[17]
References [1] P. Vaupel, O. Thews, D.K. Kelleher, M. Hoeckel, Current status of knowledge and critical issues in tumour oxygenation. Results from 25 years research in tumour pathophysiology, Adv. Exp. Med. Biol. 454 (1998) 591–602. [2] O. Warburg, Ist die aerobe Glykolyse spezifisch fuer die Tumoren?, Biochem. Z. 204 (1929) 482–483. [3] O. Warburg, On the origin of cancer cells, Science 123 (1956) 309– 314. [4] M. Guppy, S. Kong, X. Niu, S. Busfield, S.P. Klinken, Method for measuring a comprehensive energy budget in a proliferating cell system over multiple cell cycles, J. Cell. Physiol. 170 (1997) 1–7. [5] K.F. Czerwenka, M. Manavi, J. Hosmann, J. D, I. Kerstin, D. Pischinger, W.B. Battistutti, M. Behnam, E. Kubista, Comparative analysis of two-dimensional protein patterns in malignant and normal human breast tissue, Cancer Detect. Prev. 25 (2001) 268– 279. [6] G.A. Chen, T.G. Gharib, C.C. Huang, D.G. Thomas, K.A. Shedden, J.M.G. Taylor, S.L.R. Kardia, D.E. Misek, T.J. Giordano, M.D. Iannettoni, M.B. Orringer, S.M. Hanash, D.G. Beer, Proteomic analysis of lung adenocarcinoma: identification of a highly expressed set of proteins in tumors, Clin. Cancer Res. 8 (2002) 2298–2305. [7] A.C.O. Vertegaal, H.B. Kuiperij, T. van Laar, V. Scharnhorst, A.J. van der Eb, A. Zantema, cDNA micro array identification of a gene differentially expressed in adenovirus type 5- versus type 12transformed cells, FEBS 487 (2000) 151–155. [8] J.A. Mares-Perlman, E. Shrago, Energy substrate utilization in freshly isolated Morris hepatoma 7777 cells, Cancer Res. 48 (1988) 602–608. [9] M. Guppy, P. Leedman, X. Zu, V. Russel, Contribution by different fuels and metabolic pathways to the total ATP turnover of proliferating MCF-7 breast cancer cells, Biochem. J. 364 (2002) 309–315. [10] A. Franchi, P. Silvestre, J. Pouyssegur, A genetic approach to the role of energy metabolism in the growth of tumor cells: tumorigenicity of fibroblast mutants deficient either in glycolysis or in respiration, Int. J. Cancer 27 (1981) 819–827. [11] E.K.J. Pauwels, E.J.C. Sturm, E. Bombardieri, F.J. Cleton, M.P.M. Stokkel, Positron-emission tomography with [18F]fluorodeoxyglucose. Part 1. Biochemical uptake mechanism and its implication for clinical studies, J. Cancer Res. Clin. Oncol. 126 (2000) 549–559. [12] D.A. Mankoff, J.R. Bellon, Positron-emission tomographic imaging of cancer: glucose metabolism and beyond, Semin. Radiat. Oncol. 11 (2001) 16–27. [13] K. Murthy, M. Aznar, C.J. Thompson, A. Loutfi, R. Lisbona, J.H. Gagnon, Results of preliminary clinical trials of the positron emission mammography system PEM-I: a dedicated breast imaging system producing glucose metabolic images using FDG, J. Nucl. Med. 41 (2000) 1851–1858. [14] N. Avril, M. Menzel, J. Dose, M. Schelling, W. Weber, F. Janicke, W. Nathrath, M. Schwaiger, Glucose metabolism of breast cancer assessed by 18F-FDG PET: histologic and immunohistochemical tissue analysis, J. Nucl. Med. 42 (2001) 9–16. [15] R. Bos, J.J.M. van der Hoeven, E. van der Wall, P. van der Groep, P.J. van Diest, E.F.I. Comans, U. Joshi, G.L. Semenza, O.S. Hoekstra, A.A. Lammertsma, C.F.M. Molthoff, Biologic corre-
[18]
[19]
[20]
[21]
[22] [23]
[24]
[25]
[26] [27]
[28]
[29] [30]
[31]
[32]
[33]
[34]
679
lates of 18 fluorodeoxyglucose uptake in human breast cancer measured by positron emission tomography, J. Clin. Oncol. 20 (2002) 379–387. P. Burgman, J.A. OÕDonoghue, J.L. Humm, C.C. Ling, Hypoxiainduced increase in FDG uptake in MCF7 cells, J. Nucl Med. 42 (2001) 170–175. J.N. Thupari, M.L. Pinn, F.P. Kuhajda, Fatty acid synthase inhibition in human breast cancer cells leads to malonyl CoAinduced inhibition of fatty acid oxidation and cytotoxicity, Biochem. Biophys. Res. Commun. 285 (2001) 217–223. S. Chang, S. Lee, C. Lee, J.I. Kim, Y. Kim, Expression of the human erythrocyte glucose transporter in transitional cell carcinoma of the bladder, Urology 55 (2000) 448–452. M. Younes, D. Juarez, L.V. Lechago, S.P. Lerner, Glut 1 expression in transitional cell carcinoma of the urinary bladder is associated with poor patient survival, Anticancer Res. 21 (2001) 575–578. P.R. Wachsberger, E.L. Gressen, A. Bhala, S.B. Bobyock, C. Storck, R.A. Coss, D. Berd, D.B. Leeper, Variability in glucose transporter-1 levels and hexokinase activity in human melanoma, Melanoma Res. 12 (2002) 35–43. E.M. Marom, T.A. Aloia, M.-B. Moore, M. Hara, J.E. Herndon, D.H. Harpole Jr., P.C. Goodman, E.F. Patz Jr., Correlation of FDG-PET imaging with Glut-1 and Glut-3 expression in earlystage non-small cell lung cancer, Lung Cancer 33 (2001) 99–107. J.M. Brown, The hypoxic cell: a target for selective cancer therapy, Cancer Res. 59 (1999) 5863–5870. P. Vaupel, M. Hockel, Blood supply, oxygenation status and metabolic micromilieu of breast cancers: characterization and therapeutic relevance, Int. J. Oncol. 17 (2000) 869–879. G.L. Semenza, Regulation of mammalian O2 homeostasis by hypoxia-inducible factor 1, Annu. Rev. Cell Dev. Biol. 15 (1999) 551–578. G.L. Semenza, Hypoxia, clonal selection, and the role of HIF-1 in tumor progression, Crit. Rev. Biochem. Mol. Biol. 35 (2000) 71– 103. G.L. Semenza, HIF-1 and tumor progression: pathophysiology and therapeutics, Trends Mol. Med. 8 (2002) S62–S67. A. Giatromanolaki, A.L. Harris, Tumour hypoxia, hypoxia signaling pathways and hypoxia inducible factor expression in human cancer, Anticancer Res. 21 (2001) 4317–4324. H. Jiang, R. Guo, J.A. Powell-Coffman, The Caenorhabditis elegans HIF-1 gene encodes a bHLH-PAS protein that is required for adaptation to hypoxia, Proc. Natl. Acad. Sci. USA 98 (2001) 7916–7921. G.L. Semenza, HIF-1, O2 , and the 3 PHDs: how animal cells signal hypoxia to the nucleus, Cell 107 (2001) 1–3. U.R. Jewell, I. Kvietikova, A. Scheid, C. Bauer, R.H. Wenger, M. Gassmann, Induction of HIF-1a in response to hypoxia is instantaneous, FASEB J. 15 (2001) 1312–1314. N. Masson, C. Willam, P.H. Maxwell, C.W. Pugh, P.J. Ratcliffe, Independent function of two destruction domains in hyoxiainducible factor-a chains activated by prolyl hydroxylation, EMBO J. 20 (2001) 5197–5206. D.M. Stroka, T. Burkhardt, I. Desbaillets, R.H. Wenger, D.A.H. Neil, C. Bauer, M. Gassmann, D. Candinas, HIF-1 is expressed in normoxic tissue and displays an organ-specific regulation under systemic hypoxia, FASEB J. 15 (2001) 2445–2453. D.P. Stiehl, W. Jelkmann, R.H. Wenger, T. Hellwig-Burgel, Normoxic induction of the hypoxia-inducible factor 1a by insulin and interleukin-1b involves the phosphatidylinositol 3-kinase pathway, FEBS Lett. 512 (2002) 157–162. C.H. Kim, Y.S. Cho, Y.S. Chun, J.W. Park, M.S. Kim, Early expression of myocardial HIF-1a in response to mechanical stresses. Regulation by stretch-activated channels and the phosphatidylinositol 3-kinase signaling pathway, Circ. Res. 90 (2002) E25–E33.
680
M. Guppy / Biochemical and Biophysical Research Communications 299 (2002) 676–680
[35] E.K. Rofstad, Microenvironment-induced cancer metastasis, Int. J. Radiat. Biol. 76 (2000) 589–605. [36] H. Lu, R.A. Forbes, A. Verma, Hypoxia-inducible factor 1 activation by aerobic glycolysis implicates the Warburg effect in carcinogenesis, J. Biol. Chem. 277 (2002) 23111–23115. [37] M. Wartenberg, F. Donmez, F.C. Ling, H. Acker, J. Hescheler, H. Sauer, Tumor-induced angiogenesis studied in confrontation
cultures of multicellular tumor spheroids and embryoid bodies grown from pluripotent embryonic stem cells, FASEB J. 15 (2001) 995–1005. [38] B.-H. Jiang, G.L. Semenza, C. Bauer, H.H. Marti, Hypoxiainducible factor 1 levels vary exponentially over a physiologically relevant range of O2 tension, Am. J. Physiol. 271 (1996) C1172– C1180.
The hypoxic core: a possible answer to the cancer paradox Michael Guppy* Biochemistry and Molecular Biology, School of Biomedical and Chemical Sciences, University of Western Australia, 35 Stirling Highway, Crawley, WA 6009, Australia Received 4 November 2002
Abstract There are many differences, at all levels of organization, between cancerous and normal cells. Two of these (oxygen delivery and glucose metabolism) are related and manifest as low intercellular oxygen tensions (pO2 ) and a glycolytic metabolic profile in tumours and/or cancer cells. It is becoming increasingly apparent that these characteristics of cancer combine to enhance both the survival and aggressiveness of cancer cells, and that they can adversely impact on some forms of treatment. But they are also exploited in current strategies of detection and monitoring of cancers. These are therefore characteristics with important implications for the crucial balance between the aggression and growth characteristics of a tumour, and our ability to detect and treat it. The interactions and the hierarchy of events leading to these manifestations are complex, not fully understood, and involve a pivotal and intriguing paradox. This paradox results in a seemingly contradictory state in which the most dangerous tumours are those that are the most hypoxic, but also those that are the most angiogenic. This review is a synthesis of the available data into a feasible hypothesis which offers a possible resolution of this paradox and provides a testable paradigm for tumour behaviour. Ó 2002 Elsevier Science (USA). All rights reserved. Keywords: Cancer; Tumour; Hypoxia; Glycolysis; PET; HIF; pO2 ; FDG; Angiogenesis; VEGF
Cancer cells are hypoxic in vivo Vaupel et al. [1] have produced a comprehensive compilation of the data on oxygen tensions in normal and tumour tissues. In human primary tumours as compared to normal tissue; the microcirculation is compromised, tissue oxygenation is not regulated according to metabolic demand, and the pO2 of the tissue is markedly lower, linked to size (a tumour with a volume of 2 ml is extremely hypoxic), often close to 0 mm Hg, and heterogeneous. Metastatic lesions are generally similar, but may be even more hypoxic, and in recurrent tumours there is a higher proportion of the tumour that is hypoxic compared to the respective primary tumour.
Cancer cells may have upregulated glycolysis Consistent with the concept of a hypoxic cancer cell is the view that the metabolism of cancer cells is more * Fax: +61-8-9380-1148. E-mail address: [email protected].
glycolytic than that of normal cells. Glycolysis was first linked to cancer 70 years ago [2] and the data were summarized in a 1956 paper [3] in which Warburg suggested that aerobic glycolysis was a phenomenon peculiar to tumours. (Note that the general perception of glycolysis is that it is activated only in the absence of oxygen, the Pasteur effect. But the same pathway commonly operates (a) in most cells and (b) in the presence of oxygen. It produces lactate, is termed aerobic glycolysis, and can account for a significant proportion of ATP turnover in the presence of oxygen [4].) Investigations into tumour intermediary metabolism over the past 70 years have led to a general perception that cancer metabolism, even under normoxic conditions, is predominantly fueled by glucose (via aerobic glycolysis). This issue however is confusing and controversial as (a) the current perception is that cancer cells inherently undergo aerobic glycolysis, a scenario that does not necessarily invoke hypoxia, (b) there is no obvious or consistent evidence from studies using recent technology that the glycolytic enzymes or glucose transporters are overexpressed in cancers (e.g., 2D-PAGE [5], proteomic analysis [6], and cDNA micro array [7]), (c) what little
0006-291X/02/$ - see front matter Ó 2002 Elsevier Science (USA). All rights reserved. PII: S 0 0 0 6 - 2 9 1 X ( 0 2 ) 0 2 7 1 0 - 9
M. Guppy / Biochemical and Biophysical Research Communications 299 (2002) 676–680
rigorous and comprehensive data there are show that the ATP production in transformed cells under normoxic conditions is mostly oxidative [8,9], and (d) it has been shown in Chinese hamster lung fibroblasts that a high rate of aerobic glycolysis is not required for the malignant phenotype [10]. Nevertheless, the current perception (which may be correct) is that a cancer cell in vivo is hypoxic and makes ATP by glycolysis. Such a scenario has important implications for the detection, monitoring, and treatment of cancer, and for our understanding of the paradoxical behaviour of cancers in vivo.
Detection, monitoring, and treatment The perceived glycolytic character of cancer cells is being increasingly exploited by the glucose-based positron emission tomography (PET) detection technique which is based upon the assumption that tumours take up glucose more rapidly than the surrounding tissue. The compound used in this technique is 18-fluorodeoxyglucose (FDG). The interest in this technique is demonstrated by the 134 papers containing the words ÔcancerÕ and ÔpositronÕ in Current Contents between January and June, 2002. There is no doubt that this is an effective detection technique and it is now also being used to monitor treatment and to predict outcomes. But there are many known factors that influence FDG uptake and some of these can be confounding. FDG uptake is influenced by glucose concentration, by hypoxia, by the heterogeneity of the tissue (proportion of viable and necrotic cells), by the volume of the tissue, by fibrosis and by the inflammatory response. In addition, it may be that some malignancies have lower inherent FDG uptake than others, or are associated with greater levels of inflammatory change, and there is evidence to suggest that cancers handle fluoroglucose differently from glucose [11,12]. As a result of these uncertainties, (a) the efficacy of PET varies with the cancer type and stage [11,12], (b) there can be significant false negatives and positives, for example, in breast tissue [13], and (c) correlations between FDG uptake and a variety of putatively related parameters can be either absent, surprisingly weak, or totally contradictory between studies [14,15]. Due to these limitations, research is currently being directed towards more specific and sensitive molecules to use with the PET technology. The issue of a link between glycolysis and cancer is therefore key to the FDG PET technique, and as discussed above, has yet to be resolved. For example, is the PET technique effective because cancer cells are inherently glycolytic, or because they are hypoxic and thus, as a result of a Pasteur effect, glycolytic? In this regard it is certainly the case that FDG uptake increases under hypoxia in MCF-7 cells [16]. If the PET technique is
677
effective only for hypoxic cancers, might this explain why the efficacy varies with the cancer type, size, and stage? Another possible scenario is that cancer cells take up glucose rapidly (via the glucose transporter Glut-1), but convert it to lipid instead of processing it through glycolysis. This scenario would render the PET technique effective, but would not require an upregulation of glycolysis in cancer tissue. Many cancers do have unusually high levels of fatty acid synthase [17], and there is evidence from studies in which the transporter is directly measured to suggest that (a) increased Glut-1 expression is associated with some types of tumours [18] and (b) that its expression correlates inversely with survival [19]. But the Glut-1 issue is unresolved as its expression can be extremely variable in a particular tumour type [20], and the data on the correlation of Glut-1 expression with FDG uptake are inconsistent [14,15,21]. Hypoxia also has implications for cancer treatment, as hypoxic cells are radiation resistant [22]. The reason for this resistance is that only under normoxia does oxygen react with and result in the permanent incorporation of the DNA radical produced by the radiation. Under hypoxic and reducing conditions the damage can be reversed. As a result, disease-free survival time after radiotherapy is related to the pO2 of the tumour. This has led to various treatment strategies such as attempts to oxygenate tumours during radiation, and the search for compounds that are converted to toxic radicals only under hypoxic and reducing conditions [22,23].
The paradox: the relation between pO2 , angiogenesis, and prognosis Radiation therapy aside, patient survival is also directly related to the average pO2 of the tumour. But paradoxically, survival is inversely related to the degree of tumour angiogenesis, a process that should lead to an increase in the pO2 . This paradox has been specifically noted several times over the last four years [24–27], and the current explanation is that the abnormal angiogenesis which is typical of tumours [1] does not result in increased functional perfusion and the associated increase in pO2 [25,26]. However Giatromanolaki and Harris [27] raise the possibility that aggressive cancers might outgrow their blood supply, suggesting that hypoxia may be related to increased metabolic demand, rather than, or as well as, an abnormally differentiated vasculature. I would like to propose an explanation that builds on the ideas of Giatromanolaki and Harris [27] and involves the hypoxia-inducible factor (HIF-1). The HIF transcriptional system is part of the process by which mammalian cells respond to oxygen availability. The mechanism appears to be conserved among metazoans as a similar complex mediates the response to hypoxia in Caenorhabditis elegans [28]. HIF-1 regulates the
678
M. Guppy / Biochemical and Biophysical Research Communications 299 (2002) 676–680
expression of at least 30 genes via a hypoxia response element. HIF-1 is a heterodimer comprising a constitutively expressed b subunit and an a subunit whose expression is regulated by pO2 . Regulation of the expression of the a subunit is through a degradation pathway initiated by a proline hydroxylase which is sensitive to pO2 . The proline hydroxylated subunit is targetted by a ubiquitin-protein ligase. HIF-1a is therefore degraded under normoxic conditions, but stable under hypoxic conditions, and can appear within minutes of a cell becoming hypoxic [26,29–31]. Rapidly accumulating evidence suggests that HIF-1a is a significant factor in the establishment, development, growth, and metastasis of lethal cancers. Its expression is correlated with tumour grade and the von Hippel-Lindau tumour suppressor plays a key role in the oxygen-dependent degradation of HIF-1a [31]. Until recently, the general consensus was that HIF-1a is not usually detected in normal tissue. But this view is changing as our detection techniques improve [32], and as it is found that HIF-1a expression is also regulated, under normoxic conditions, by peptides such as insulin [33] and stimuli such as mechanical stress [34]. Nevertheless, HIF-1a is detected in most malignant, but not benign tumours, and many of the products of genes transactivated by HIF-1a are overexpressed in human tumours. These include proteins involved in adaptation to hypoxia, angiogenic proteins that support tumour vascularization and proliferation (e.g., VEGF), and proteins that putatively play a role in metastasis and invasion [25,35]. HIF-1a therefore links hypoxia, angiogenesis, the cancer phenotype, and intriguingly, aerobic glycolysis, as there is recent evidence to suggest that HIF-1a can accumulate under aerobic conditions if pyruvate is being produced from glucose [36]. But this only partially resolves the paradox mentioned previously, and in some ways re-emphasizes it, as HIF-1a expression correlates with vascularity [25]. So how does a hypoxic cancer, which is synthesizing HIF-1a, stimulating the production of VEGF, and promoting angiogenesis, remain in this seemingly contradictory state? This is a crucial question as this state characterizes the most aggressive, fast-growing cancers with the poorest prognoses. Recent data from confrontation cultures consisting of non-tumour and tumour spheroids may offer an answer [37]. The data show that angiogenesis in the non-tumour tissue results in improved oxygen supply, down-regulation of HIF-1a and VEGF, but no change in growth rate. In contrast, upon vascularization of the tumour spheroid, the growth rate increases dramatically, the pO2 of the tumour decreases, and there is a concomitant upregulation of HIF-1a and VEGF. The authors suggest that the initial angiogenesis of the tumour enables the outer layer of the tumour to increase its oxidative metabolism and to proliferate, and that it does so at the expense of the core which remains hypoxic, but not
necrotic. The core continues to express HIF-1a, VEGF, and all the other proteins whose genes are targets of HIF-1a, while the outer layer is perfused and grows rapidly. These are the aforementioned paradoxical tumours; irregularly perfused, hypoxic but chronically angiogenic, aggressive, fast growing, metastasizing, and expressing HIF-1a and VEGF. The crucial character of an aggressive cancer is therefore the hypoxic core, so a feasible hypothesis is that cancer cells are glycolytic only under hypoxic conditions, they are not inherently glycolytic. Therefore not all cancers will be glycolytic and the contribution of glycolysis to ATP turnover will vary with the degree of hypoxia. This is consistent with the variable nature of the FDG PET data and with the data from the two aforementioned studies of cancer metabolism [8,9]. But an integral part of this hypothesis must also be the fact that the cells in the hypoxic core adapt to hypoxia, rather than becoming necrotic, and the mechanism by which this occurs would be one of the roles of the multifaceted HIF-1a. The abnormal expression of HIF1a in tumours has already been mentioned above. But as well as this quantitative difference between cancer and normal tissue, the adaptation to hypoxia may involve a relationship between pO2 and the expression of HIF-1a which is specific to cancer cells.
Conclusion There are four questions that need to be addressed. 1. What is the metabolic profile of cancer cells, under normoxia and hypoxia? The perception, as noted above, is that cancer cells are glycolytic, even under normoxic conditions. But despite the general confidence in the literature and the efficacy of the PET technique, the data are equivocal. 2. Is there a hypoxic core in tumours, or perhaps more importantly, is there a highly oxidative shell, and what proteins are being produced in these two areas? 3. If there is a hypoxic core, are the cells specifically adapted to hypoxia compared to those in the oxidative shell, or are all cancer cells preadapted to survive hypoxia? 4. When, with regard to pO2 , does HIF-1a protein appear in the cells in the hypoxic core of a tumour? There has been one study relating HIF-1a expression to pO2 in HeLa cells [38], showing that HIF-1a is expressed only below 6% oxygen (note that oxygen levels in hypoxic tumours in vivo are extremely variable, but are commonly below 3% [1]). But it is not known whether this relationship is different between normal and tumour cells. These data are necessary if we are to understand cancer biology and the cancer process. This will in turn
M. Guppy / Biochemical and Biophysical Research Communications 299 (2002) 676–680
lead to an understanding of the mechanisms that form the basis of many of the current diagnostic and treatment strategies, and will in turn facilitate the improvement of these strategies and suggest new ones.
[16]
[17]
References [1] P. Vaupel, O. Thews, D.K. Kelleher, M. Hoeckel, Current status of knowledge and critical issues in tumour oxygenation. Results from 25 years research in tumour pathophysiology, Adv. Exp. Med. Biol. 454 (1998) 591–602. [2] O. Warburg, Ist die aerobe Glykolyse spezifisch fuer die Tumoren?, Biochem. Z. 204 (1929) 482–483. [3] O. Warburg, On the origin of cancer cells, Science 123 (1956) 309– 314. [4] M. Guppy, S. Kong, X. Niu, S. Busfield, S.P. Klinken, Method for measuring a comprehensive energy budget in a proliferating cell system over multiple cell cycles, J. Cell. Physiol. 170 (1997) 1–7. [5] K.F. Czerwenka, M. Manavi, J. Hosmann, J. D, I. Kerstin, D. Pischinger, W.B. Battistutti, M. Behnam, E. Kubista, Comparative analysis of two-dimensional protein patterns in malignant and normal human breast tissue, Cancer Detect. Prev. 25 (2001) 268– 279. [6] G.A. Chen, T.G. Gharib, C.C. Huang, D.G. Thomas, K.A. Shedden, J.M.G. Taylor, S.L.R. Kardia, D.E. Misek, T.J. Giordano, M.D. Iannettoni, M.B. Orringer, S.M. Hanash, D.G. Beer, Proteomic analysis of lung adenocarcinoma: identification of a highly expressed set of proteins in tumors, Clin. Cancer Res. 8 (2002) 2298–2305. [7] A.C.O. Vertegaal, H.B. Kuiperij, T. van Laar, V. Scharnhorst, A.J. van der Eb, A. Zantema, cDNA micro array identification of a gene differentially expressed in adenovirus type 5- versus type 12transformed cells, FEBS 487 (2000) 151–155. [8] J.A. Mares-Perlman, E. Shrago, Energy substrate utilization in freshly isolated Morris hepatoma 7777 cells, Cancer Res. 48 (1988) 602–608. [9] M. Guppy, P. Leedman, X. Zu, V. Russel, Contribution by different fuels and metabolic pathways to the total ATP turnover of proliferating MCF-7 breast cancer cells, Biochem. J. 364 (2002) 309–315. [10] A. Franchi, P. Silvestre, J. Pouyssegur, A genetic approach to the role of energy metabolism in the growth of tumor cells: tumorigenicity of fibroblast mutants deficient either in glycolysis or in respiration, Int. J. Cancer 27 (1981) 819–827. [11] E.K.J. Pauwels, E.J.C. Sturm, E. Bombardieri, F.J. Cleton, M.P.M. Stokkel, Positron-emission tomography with [18F]fluorodeoxyglucose. Part 1. Biochemical uptake mechanism and its implication for clinical studies, J. Cancer Res. Clin. Oncol. 126 (2000) 549–559. [12] D.A. Mankoff, J.R. Bellon, Positron-emission tomographic imaging of cancer: glucose metabolism and beyond, Semin. Radiat. Oncol. 11 (2001) 16–27. [13] K. Murthy, M. Aznar, C.J. Thompson, A. Loutfi, R. Lisbona, J.H. Gagnon, Results of preliminary clinical trials of the positron emission mammography system PEM-I: a dedicated breast imaging system producing glucose metabolic images using FDG, J. Nucl. Med. 41 (2000) 1851–1858. [14] N. Avril, M. Menzel, J. Dose, M. Schelling, W. Weber, F. Janicke, W. Nathrath, M. Schwaiger, Glucose metabolism of breast cancer assessed by 18F-FDG PET: histologic and immunohistochemical tissue analysis, J. Nucl. Med. 42 (2001) 9–16. [15] R. Bos, J.J.M. van der Hoeven, E. van der Wall, P. van der Groep, P.J. van Diest, E.F.I. Comans, U. Joshi, G.L. Semenza, O.S. Hoekstra, A.A. Lammertsma, C.F.M. Molthoff, Biologic corre-
[18]
[19]
[20]
[21]
[22] [23]
[24]
[25]
[26] [27]
[28]
[29] [30]
[31]
[32]
[33]
[34]
679
lates of 18 fluorodeoxyglucose uptake in human breast cancer measured by positron emission tomography, J. Clin. Oncol. 20 (2002) 379–387. P. Burgman, J.A. OÕDonoghue, J.L. Humm, C.C. Ling, Hypoxiainduced increase in FDG uptake in MCF7 cells, J. Nucl Med. 42 (2001) 170–175. J.N. Thupari, M.L. Pinn, F.P. Kuhajda, Fatty acid synthase inhibition in human breast cancer cells leads to malonyl CoAinduced inhibition of fatty acid oxidation and cytotoxicity, Biochem. Biophys. Res. Commun. 285 (2001) 217–223. S. Chang, S. Lee, C. Lee, J.I. Kim, Y. Kim, Expression of the human erythrocyte glucose transporter in transitional cell carcinoma of the bladder, Urology 55 (2000) 448–452. M. Younes, D. Juarez, L.V. Lechago, S.P. Lerner, Glut 1 expression in transitional cell carcinoma of the urinary bladder is associated with poor patient survival, Anticancer Res. 21 (2001) 575–578. P.R. Wachsberger, E.L. Gressen, A. Bhala, S.B. Bobyock, C. Storck, R.A. Coss, D. Berd, D.B. Leeper, Variability in glucose transporter-1 levels and hexokinase activity in human melanoma, Melanoma Res. 12 (2002) 35–43. E.M. Marom, T.A. Aloia, M.-B. Moore, M. Hara, J.E. Herndon, D.H. Harpole Jr., P.C. Goodman, E.F. Patz Jr., Correlation of FDG-PET imaging with Glut-1 and Glut-3 expression in earlystage non-small cell lung cancer, Lung Cancer 33 (2001) 99–107. J.M. Brown, The hypoxic cell: a target for selective cancer therapy, Cancer Res. 59 (1999) 5863–5870. P. Vaupel, M. Hockel, Blood supply, oxygenation status and metabolic micromilieu of breast cancers: characterization and therapeutic relevance, Int. J. Oncol. 17 (2000) 869–879. G.L. Semenza, Regulation of mammalian O2 homeostasis by hypoxia-inducible factor 1, Annu. Rev. Cell Dev. Biol. 15 (1999) 551–578. G.L. Semenza, Hypoxia, clonal selection, and the role of HIF-1 in tumor progression, Crit. Rev. Biochem. Mol. Biol. 35 (2000) 71– 103. G.L. Semenza, HIF-1 and tumor progression: pathophysiology and therapeutics, Trends Mol. Med. 8 (2002) S62–S67. A. Giatromanolaki, A.L. Harris, Tumour hypoxia, hypoxia signaling pathways and hypoxia inducible factor expression in human cancer, Anticancer Res. 21 (2001) 4317–4324. H. Jiang, R. Guo, J.A. Powell-Coffman, The Caenorhabditis elegans HIF-1 gene encodes a bHLH-PAS protein that is required for adaptation to hypoxia, Proc. Natl. Acad. Sci. USA 98 (2001) 7916–7921. G.L. Semenza, HIF-1, O2 , and the 3 PHDs: how animal cells signal hypoxia to the nucleus, Cell 107 (2001) 1–3. U.R. Jewell, I. Kvietikova, A. Scheid, C. Bauer, R.H. Wenger, M. Gassmann, Induction of HIF-1a in response to hypoxia is instantaneous, FASEB J. 15 (2001) 1312–1314. N. Masson, C. Willam, P.H. Maxwell, C.W. Pugh, P.J. Ratcliffe, Independent function of two destruction domains in hyoxiainducible factor-a chains activated by prolyl hydroxylation, EMBO J. 20 (2001) 5197–5206. D.M. Stroka, T. Burkhardt, I. Desbaillets, R.H. Wenger, D.A.H. Neil, C. Bauer, M. Gassmann, D. Candinas, HIF-1 is expressed in normoxic tissue and displays an organ-specific regulation under systemic hypoxia, FASEB J. 15 (2001) 2445–2453. D.P. Stiehl, W. Jelkmann, R.H. Wenger, T. Hellwig-Burgel, Normoxic induction of the hypoxia-inducible factor 1a by insulin and interleukin-1b involves the phosphatidylinositol 3-kinase pathway, FEBS Lett. 512 (2002) 157–162. C.H. Kim, Y.S. Cho, Y.S. Chun, J.W. Park, M.S. Kim, Early expression of myocardial HIF-1a in response to mechanical stresses. Regulation by stretch-activated channels and the phosphatidylinositol 3-kinase signaling pathway, Circ. Res. 90 (2002) E25–E33.
680
M. Guppy / Biochemical and Biophysical Research Communications 299 (2002) 676–680
[35] E.K. Rofstad, Microenvironment-induced cancer metastasis, Int. J. Radiat. Biol. 76 (2000) 589–605. [36] H. Lu, R.A. Forbes, A. Verma, Hypoxia-inducible factor 1 activation by aerobic glycolysis implicates the Warburg effect in carcinogenesis, J. Biol. Chem. 277 (2002) 23111–23115. [37] M. Wartenberg, F. Donmez, F.C. Ling, H. Acker, J. Hescheler, H. Sauer, Tumor-induced angiogenesis studied in confrontation
cultures of multicellular tumor spheroids and embryoid bodies grown from pluripotent embryonic stem cells, FASEB J. 15 (2001) 995–1005. [38] B.-H. Jiang, G.L. Semenza, C. Bauer, H.H. Marti, Hypoxiainducible factor 1 levels vary exponentially over a physiologically relevant range of O2 tension, Am. J. Physiol. 271 (1996) C1172– C1180.