Keywords

FormalPara Key Points

  • Tumor cells exhibit an upregulation in glycolysis, glycogen metabolism, and gluconeogenesis as opposed to normal cells.

  • The metabolic reprogramming underlying the Warburg effect and other changes in glucose metabolism are driven by several oncogenes and tumor suppressors.

  • Numerous therapies based on cancer metabolism have been developed but have yet to show success in clinical trials.

FormalPara Introduction

Otto Warburg observed a peculiar phenomenon in 1924, unknowingly laying the foundation for the field of cancer metabolism. While his contemporaries hypothesized that tumor cells derived the energy required for uncontrolled replication from proteolysis and lipolysis, Warburg instead found them to rapidly consume glucose, converting it to lactate [1]. The significance of this finding, later termed the Warburg effect, went unnoticed by the larger scientific community at that time. The field of cancer metabolism lay dormant for almost a century awaiting advances in molecular biology and genetics which would later open the doors to new cancer therapies.

1 The Warburg Effect

1.1 Otto Warburg’s Early Studies of Normal Cellular Respiration

Warburg began his forays into research studying the oxygen consumption of sea urchin eggs, finding that the rate of respiration increased several fold after fertilization. He went on to further describe two processes that were crucial to cellular glucose metabolism: respiration and fermentation [2].

Most differentiated cells metabolize glucose through the TCA cycle under aerobic conditions. They then undergo oxidative phosphorylation to generate ATP (between 32 and 34 ATP molecules per glucose molecule) [3] (Fig. 1). While glycolysis produces only two net molecules of ATP per one molecule of glucose, the majority of ATP production occurs during the TCA cycle and oxidative phosphorylation. During these latter steps of respiration, the pyruvate molecule produced in glycolysis undergoes a series of reactions in the presence of oxygen. Without the presence of oxygen, cells undergo fermentation or anaerobic glycolysis, shunting the resultant pyruvate molecules to lactate production.

Fig. 1
figure 1figure 1

The Warburg effect depicted in proliferating tissue (right) in contrast with normal respiration in normal differentiated tissue (left)

Full size image

1.2 The Warburg Effect Is a Prominent Feature of Cancer Cell Metabolism

In 1927, Warburg studied the processes of respiration and fermentation in tumor cells. According to normal cellular respiration, glucose is converted to pyruvate and then enters the TCA cycle to undergo oxidative phosphorylation in the presence of oxygen. There should be minimal lactate production. However, in his in vivo and ex vivo studies, Warburg observed an increased glucose uptake and increased lactic acid production in tumor cells as compared to normal cells, even in the presence of oxygen [4]. This phenomenon, the metabolism of glucose to lactate despite the presence of adequate oxygen, is called the Warburg effect.

For Warburg, several questions remained unanswered, including why cancer cells would inefficiently shunt glucose to lactate production instead of to the TCA cycle, which would result in significantly higher ATP production. Warburg hypothesized that the lactate production in cancer cells was due to impairment of oxidative phosphorylation caused by mitochondrial damage [5].

There was debate surrounding this theory with disagreement arising particularly from Sidney Weinhouse, one of Warburg’s contemporaries. Using isotope tracing, Weinhouse’s experiments showed that the rates of oxidative phosphorylation in both normal cells and tumor cells are similar, suggesting that the mitochondria of tumor cells are intact [6]. Rather, tumor cells in oxygen-rich environments utilize both aerobic glycolysis and oxidative phosphorylation to sustain their rapid rates of proliferation. Only in hypoxic environments, such as the tumor core, do the rates of lactic acid production by anaerobic glycolysis overtake oxidative phosphorylation as the primary source of energy [7].

1.3 Metabolic and Genetic Reprogramming Underlying the Warburg Effect

With current advances in genetics and molecular biology, much of the past several decades of cancer research have been consumed by characterizing the genetic alterations which lead to the development of cancers. Warburg’s question regarding the cause of upregulation of aerobic glycolysis in cancer remained unanswered. However, cancer cells need not only a genetic switch but also metabolic building blocks and an energy source to undergo rapid proliferation. The recognition of the importance of this energy source allowed for the resurgence of cancer metabolism as a field that is closely related to tumor genetics. It is now understood that the metabolic reprogramming underlying the Warburg effect is driven by several oncogenes and tumor suppressors.

Some of the identified oncogenes, namely, Akt, PI3K, Ras, and VHL, act via the protein hypoxia-inducible factor 1α (HIF-1α) [8], resulting in the non-hypoxic expression of HIF-1α. In normal cells, HIF-1α becomes stabilized in a hypoxic environment to form a transcription factor involved in promoting glycolysis and suppressing oxidative phosphorylation [9]. HIF-1α, when present, upregulates the GLUT1 transporter to promote the retention of glucose inside cells in addition to upregulating hexokinase 2 (HK2), the enzyme which catalyzes the first committed step of glycolysis [10]. Typically, when oxygen is present, HIF-1α degrades in a concentration-dependent manner. In tumor cells, even in the presence of oxygen, high AKT and mTOR oncogenic activity promotes HIF-1α expression, leading to persistent transcription of the enzymes driving glycolysis and lactate production.

Oncogenic pathways have been found to work independently of HIF-1α to promote aerobic glycolysis as well, namely, the activation of oncogenes such as Myc, Ras, and Akt and the deactivation of tumor suppressors such as TP53 [8]. Like HIF-1α, Myc directly upregulates GLUT and HK2. The loss of TP53 function also upregulates GLUT expression. Additionally, TP53 deactivation indirectly leads to increased glycolysis. Without TP53 expression, TIGAR, a protein which causes shunting of glucose to the pentose phosphate pathway (PPP), is no longer upregulated, resulting in a greater flux of glucose through the glycolytic pathway [11].

2 The Role of Glycogen Metabolism and Gluconeogenesis in Tumor Growth

2.1 Glycogen Metabolism Is Upregulated in Several Cancers

Glycolysis is not the only component of glucose metabolism which plays a significant role in tumor growth. Glycogenolysis, the process by which glycogen is converted to glucose-1-phosphate (G1P) and then to glucose-6-phosphate (G6P) to enter the glycolytic pathway, provides another energy source for tumors in the face of nutrient stress (Fig. 2). Glycogen metabolism, although studied far less than glycolysis by cancer researchers, is upregulated in many cancer types including renal, breast, bladder, uterine, ovarian, skin, and brain cancers. The glycogen content of cancer cells has been demonstrated to be inversely proportional to the rate of replication [12]. Renal cell carcinoma which classically has clear cells on histology appears this way due to high glycogen content.

Fig. 2
figure 2figure 2

Glucose metabolism in cancer

Full size image

Advances in tumor genetics have again allowed for the characterization of tumor suppressor genes and oncogenes which have driven these changes in glycogen metabolism in tumor cells. The overexpression of the oncogene Rab25 has been demonstrated as a driver in increasing cellular glycogen stores via the AKT pathway [13]. In bladder cancer, the glycogen debranching enzyme AGL has been identified as a tumor suppressor. Additionally, deactivation of AGL leads to the accumulation of abnormal glycogen stores and promotes tumorigenesis in xenograft models [14].

Given this, Guo-Min Shen and colleagues studied glycogen metabolism in the setting of hypoxia in 2010. It was noted that glycogen accumulated in breast cancer cells after 24 and 48 hours of hypoxia due to HIF-1α induction of PPP1R3C, a glycogen synthase [15]. Later studies demonstrated that glycogen synthesis promotes cancer cell survival in the setting of hypoxic conditions [16]. Both glycogenolysis and glycogen synthesis enzymes appear to be upregulated by tumor cells with HIF-1α dependence including UGP2, PGM, GBE, GYS1, and PPP1R3C [17]. In vivo studies of suppression of glycogen synthase kinase 2 (GSK2) activity demonstrated a reduction in prostate tumor growth [18]. Glycogen metabolism is an important target of therapy given that cancer cells can utilize glycogen as an energy source even during nutrient deficiency due to poor angiogenesis [19].

2.2 Upregulation of Gluconeogenic Enzymes in Cancer

Gluconeogenesis is the process of generating glucose from carbon substrates which are not carbohydrates. There are two gluconeogenic enzymes which play important roles in cancer metabolism: phosphoenolpyruvate carboxykinase 1 (PCK1) and phosphoenolpyruvate carboxykinase 2 (PCK2). It has been demonstrated that TP53 inhibits both enzymes, meaning that the loss of TP53 upregulates these enzymes and gluconeogenesis [20]. It was also observed that the inhibition of mTOR in hepatocellular carcinoma and renal cell carcinoma cells directs the glycolytic flux towards lactate and gluconeogenesis with resultant tumor cell death via the downregulation of PCK1 [21].

3 Targeting Glucose Metabolism for Cancer Therapy (Fig. 3)

3.1 Therapies Targeting Glycolysis and the Warburg Effect

As discussed previously, over the latter half of the twentieth century, advances in molecular biology and the identification of oncogenes and tumor suppressors drew the attention of much of the anti cancer therapeutic efforts. It is true that genetic alterations drive uncontrolled replication in cancer cells, but it is important to recognize that a cancer cell is still dependent on nutrient availability. In the past two decades, there has been an upsurge in efforts to exploit the addiction of cancer cells to glucose and the Warburg effect [8]. Several enzymes in the glycolytic pathway have been targeted, some showing tumoricidal effects in vitro and in vivo. Unfortunately, there has been little clinical success given that glycolysis is crucial to the glucose metabolism of normal cells. The focus is targeting those elements of aerobic glycolysis which are more upregulated in cancer.

Fig. 3
figure 3figure 3

Potentially viable targets within glucose metabolism for cancer therapy

Full size image

As mentioned previously, glucose transporters (GLUT1-4) are upregulated in tumor cells by Myc and HIF-1α. Previous attempts with small molecule inhibitors of GLUT1 have seen in vitro tumoricidal effects in a renal cell carcinoma cell line [22] and hepatocellular carcinoma cell line [23]. However, GLUT1 is a prevalent glucose transporter in normal cells as well, which would likely preclude clinical success. Homozygous Glut1 deletion is embryonically lethal in mice, and heterozygous deletion has caused impaired motor activity and seizures [24]. A GLUT1 inhibitor called silibinin failed to demonstrate any reduction in prostate-specific antigen in a phase I clinical trial [25].

Hexokinase phosphorylates glucose to glucose-6-phosphate in the first committed step of glycolysis. Hexokinase 2 (HK2) is mostly expressed in cancer cells and is the primary hexokinase to function in tumors, so it is another potential therapeutic target. Experiments in which HK2 was systemically deleted have shown to be well tolerated in mice [26]. A glucose analog that competitively inhibits G6P isomerase in order to inhibit the phosphorylation of glucose, 2-deoxyglucose, had a phase I trial in combination with radiation therapy with good toleration in glioblastoma multiforme. However, a HK inhibitor called lonidamine failed to show any benefit in two phase III randomized trials [25].

Phosphofructokinase (PFK) is the enzyme which catalyzes the second committed step in glycolysis, the conversion of fructose-6-phosphate to fructose-1,6-bisphosphate (F1,6-BP). Although inhibiting PFK directly is not possible since it is crucial to glycolysis in normal cells, it may be feasible to target it indirectly. PFK is strongly allosterically activated by fructose-2,6-bisphosphonate (F2,6-BP). F2,6-BP is activated by another protein, PFKFB3, a target of HIF-1α. Attenuation of glycolysis was achieved in in vitro and in vivo studies with a small molecule PFKFB3 inhibitor called 3PO [27]. PFKFB3 inhibitors were also shown to reduce tumor angiogenesis [28].

In seeking a target that was more unique to cancer cell metabolism and central to the Warburg effect, Le et al. focused on lactate dehydrogenase A (LDHA) which reciprocally mediates the redox-coupled conversion between lactate with NAD+ and pyruvate with NADH [29, 30]. Elevated expression levels of LDHA are a hallmark of many types of tumors, including squamous head and neck cancer, colorectal cancer, and non-small cell lung cancer [30,31,32,33]. By perturbing the NADH/NAD+ ratio, a small molecular inhibitor of LDHA called FX-11 was shown to increase reactive oxygen species in tumor cells with subsequent cell death in not only in vitro studies but also pancreatic and lymphoma xenografts [34]. Several other LDHA inhibitors, such as gossypol, galloflavin, and N-hydroxyindole-based inhibitors, were tested in preclinical settings [34,35,36,37,38]. Among them, gossypol (AT-101), a non selective inhibitor of LDH, was tested in a phase I clinical trial targeting metastatic colorectal cancer (NCT00540722). Despite active investigations for developing LDH inhibitors, there is still a clinical need for highly selective and efficient LDH inhibitors, as gossypol shows off-target effects such as the inhibition of NADH-dependent enzymes (GAPDH) [36]. Although compounds targeting lactate metabolism have not yet been approved, it is clear that LDH-targeting strategies are promising approaches for cancer therapy.

3.2 Therapies Targeting Glycogenolysis and Gluconeogenesis

Significantly fewer therapies have been developed targeting glycogen metabolism or gluconeogenesis. Lee et al. inhibited glycogen phosphorylase in a pancreatic cell line with a compound called CP-320626 leading to tumor cell death with no effect on normal human fibroblasts [39]. Flavopiridol, another glycogen phosphorylase inhibitor, had safe and modest efficacy in clinical trials with prostate cancer, renal cell carcinoma, and colorectal carcinoma. However, flavopiridol is also a cyclin-dependent kinase inhibitor, so it is uncertain whether the anti tumor effects were purely from glycogen phosphorylase inhibition.

Conclusion

Currently, there are several challenges to metabolic cancer therapies. First, an understanding of the heterogeneity of metabolic phenotypes is only beginning to be established. Metabolic phenotypes likely vary based on tissue of origin, tumor microenvironment, primary versus metastatic tumors, and mutational differences. Second, there are limitations in translating in vivo mouse studies to clinical trials as is evidenced by the lack of success in advancing metabolic inhibitors through clinical trials up until this point. Third, there is the potential for metabolic inhibitors to be overcome by the adaptation of tumors to new energy sources. With renewed interest in cancer metabolism, the development of metabolic inhibitors will continue to grow, and it may be most effective to combine these therapies with other modalities of therapy in order to increase efficacy.