The mitochondrial bioenergetic capacity of carcinomas

Energetic Metabolism of Proliferating Cells

Glycolysis and oxidative phosphorylation are the metabolic pathways relevant for energy provision during cellular proliferation (Fig. 1). Glycolysis, when compared with mitochondrial oxidative phosphorylation, is a very low-energy providing pathway (∼18-fold less) but supplies many of the metabolic intermediates required for the biosynthesis of macromolecules (Fig. 1). Proliferation by glycolysis is thus possible because cells, either from a tumor, an embryo or in culture, have a nonrestrained availability of metabolic precursors (glucose/glutamine) (Fig. 1) for the biosynthesis of their constituents. These pathways are tightly regulated during the cell cycle of the eukaryotic cell (13) and, whereas the buildup of cellular components at G1 (the oxidative phase of the cycle) requires the functional activity of mitochondria (14), progression through the synthesis (S-phase) and G2/M phases (so-called reductive metabolic phases) (13), is linked to the repression of mitochondrial function and the burst of glycolysis (1, 15). Indeed, when lymphocytes and thymocytes are stimulated to proliferate, aerobic glycolysis is sharply increased and oxygen consumption and the production of reactive oxygen species (ROS) sharply diminished (see (1) and references therein). The onset of glycolysis and the repression of mitochondrial respiration in S phase have the functional advantage of obliterating the production of ROS at a time when DNA molecules are most vulnerable to oxidative damage. Interestingly, Myc is a master regulator of the metabolic pathways that drive cell cycle progression (16) and cyclin D1, which represses mitochondrial activity (15), plays a prominent role in the metabolic transition at the G1/S boundary. In any case, a conclusion that stems from the metabolic trait of proliferating tissues, either cancerous or not, is that the expression of proteomic markers of energetic metabolism should differ when compared with nonproliferating tissues.

Integration of Glycolysis and Respiration

Energetic metabolism is delicately controlled by the availability of ATP, NADH and some metabolic intermediates (Fig. 1). In the cytoplasm, glucose is partially oxidized to pyruvate by glycolysis (Fig. 1). Pyruvate could have various fates, either it is reduced to lactate to regenerate NAD⁺ or enters the mitochondria where its carbon skeleton is completely oxidized to CO₂ by the sequential action of pyruvate dehydrogenase (PDH) and the enzymes of the TCA cycle (Fig. 1). The electrons obtained in oxidations and collected in NADH are transferred to the respiratory chain to generate the proton electrochemical gradient that will be used by the mitochondrial H⁺-ATP synthase for the synthesis of ATP (Fig. 1). In nonproliferating cells, the demand for glucose is low because the requirement for carbon skeletons is low and the complete oxidation of glucose by mitochondrial activity provides a high yield of ATP and NADH (Fig. 1). The cellular availability of ATP and NADH coordinate the down-regulation of the activity of key enzymes of energetic metabolism limiting the flux of matter and the production of biological energy through these pathways (Fig. 1) (1). On the contrary, proliferating cells take up considerably much larger amounts of glucose due to the continuous draining of metabolic intermediates (Fig. 1), the need for NADPH for biosynthetic purposes (Fig. 1) and the metabolic oscillations inherent to cell cycle progression (13) that boost the glucose consumption rate as well as its partial aerobic oxidation to lactate (1). The catabolism of glutamine is also of relevance during proliferation and in oncogenesis because replenishes the TCA cycle with carbon skeletons (17, 18).

Likewise, when cells have to adapt to a hypoxic environment or have a genetic alteration that impinges in the energy provision activity of mitochondria, the flux of glycolysis and concomitant production of lactate should increase to provide sufficient energy to maintain cellular functions. Activation/repression of glycolysis is driven by short-term allosteric regulation of PFK and PK (Fig. 1) as a result of changes in the cellular availability of ATP and NADH (Fig. 1). However, there are physiological situations in which the activation of the enzymes of glycolysis is not enough to cope with the cellular demand of carbon skeletons and energy required to sustain proliferation. The proliferation of hepatocytes during development of the fetal liver (12, 19) and of cancer cells in tumors (20, 21) provide two examples in which the induction of glycolytic genes and the repression of mitochondrial biogenesis accompany cellular proliferation.

The Bioenergetic Phenotype of Carcinomas

As discussed earlier, the energetic metabolism of any proliferating cell, cancerous or noncancerous, should be based on an active glycolysis. Gene expression and proteomic analysis have identified a common metabolic signature of cancer demonstrating that the increased expression of enzymes of the glycolytic pathway indeed represents a hallmark of malignant cells (see (1) and references therein). Furthermore, it is well established that tumor cells express embryonic-type isoforms of many of the enzymes of glycolysis, which is a very active pathway during development of the mammalian embryo (12). Recently, the switch to the embryonic M2 isoform of the glycolytic enzyme pyruvate kinase (Fig. 1) has been shown to promote tumorigenesis (22).

A debated question is whether the abnormal aerobic glycolysis of cancer cells results from an impaired bioenergetic activity of mitochondria as it was originally suggested by Warburg (9-11). There is a large body of data that sustain that mitochondria are structurally and functionally altered in the cancer cell (1, 4, 6, 7, 23-26). Consistent with an impaired bioenergetic function of mitochondria in cancer a large number of somatic mutations in mtDNA have been described in different neoplasias (Fig. 2) (25-27). Moreover, mutations in enzymes of the TCA cycle (Fig. 2) also link mitochondrial dysfunction with oncogenesis (28, 29).

The metabolic and enzymatic analogies that exist between the phenotype of cancer cells and embryonic tissues go far beyond the expression of protein isoforms of glycolytic enzymes and also include a similar trend in the expression of markers of the two pathways of energetic metabolism as well as the operation of the same mechanisms that control the biogenesis of mitochondria (12). Indeed, the expression level of β-F1-ATPase, the catalytic subunit of the H⁺-ATP synthase (Fig. 1), and thus a rate-limiting component of mitochondrial oxidative phosphorylation, inversely correlates with the expression of markers of the glycolytic pathway (GAPDH, HK, etc.) during development of the liver (12). Likewise, a test to assess the protein signature of energetic metabolism in normal and tumor biopsies derived from the same cancer patients confirmed that β-F1-ATPase expression is inversely correlated with the expression of GAPDH in most human carcinomas (5). The tumor drop in the β-F1-ATPase/GAPDH ratio suggests a deficit in the overall cellular activity of mitochondria in the cancer cell consistent with Warburg's postulates. This analysis was defined as the bioenergetic signature (1, 5). Recently, the quantitative determination of the bioenergetic signature revealed that tumors from different tissues and/or histological types have the same proteomic signature of energetic metabolism (30), indicating that cancer abolishes the tissue-specific differences in the bioenergetic phenotype of mitochondria. These findings have been confirmed in other laboratories (31) (and see (1) for extensive review of the literature).

Positron emission tomography using the glucose analog 2-deoxy-2[¹⁸F]-D-glucose as probe (FDG-PET) has conclusively demonstrated that the majority of human carcinomas fulfil the enhanced glycolytic criteria first observed by Warburg. Moreover, it has been demonstrated that the bioenergetic signature of cancer cells and tumors inversely correlate with the rates of aerobic glycolysis and the rates of glucose capture as assessed by FDG-PET imaging (32). These findings strongly support that the down-regulation of the bioenergetic function of mitochondria is part of the mechanism that triggers the increased glucose avidity of tumors (20, 32).

Analysis of the bioenergetic signature in large cohorts of different tumors have further emphasized that the expression of β-F1-ATPase affords an excellent marker of the prognosis of patients with breast (33, 34), colon (5, 31) and lung (32, 35) cancer. Likewise, a high tumor rate of FDG uptake is also a significant predictor of an unfavorable outcome in lung cancer (32). Moreover, the tumor expression of β-F1-ATPase is an independent marker of survival in breast and lung cancer as assessed by multivariate Cox regression analysis (32, 34). Recently, these findings have been confirmed in a different large cohort of patients with colon cancer (31). Thus, the alteration of the bioenergetic signature strongly supports a relevant role for the mitochondrial impairment of the cancer cell in progression of the disease. In addition, the bioenergetic signature also predicts the response to chemotherapy in various cancer cells (36, 37) as well as in colorectal tumors (31), consistent with the role of the H⁺-ATP synthase in signaling the execution of cell death (36) (and references therein).

Repression of Mitochondrial Biogenesis in Cancer

Cancer cells repress the biogenesis of mitochondria to acquire a glycolytic phenotype with little or no dependence on oxidative phosphorylation (21). Post-transcriptional regulation by translational control of the mRNA that encodes β-F1-ATPase (β-mRNA) in fetal liver (12, 38), during the cell cycle (39) and in rat (21) and human (40) carcinomas (Fig. 2) partially explain the abnormal biogenesis of mitochondria in these situations. The 3′UTR of β-mRNA is required for efficient translation of the transcript (21, 38, 41, 40) and translation masking of β-mRNA is observed in fetal rat liver (38) and in rat (21) and human (40) carcinomas (Fig. 2). In this context, the identification of the β-mRNA binding molecules that interact and presumably hamper translation of the transcript is of utmost importance to understand the bioenergetic phenotype of cancer cells. Recently, the AU-rich element (ARE) binding protein HuR, a central regulator of post-transcriptional gene expression, has been identified as a 3′UTR β-mRNA interacting protein (33). Although HuR plays no relevant role in regulating β-F1-ATPase expression (33), it affords a relevant independent marker of breast cancer prognosis and, when studied in combination with the bioenergetic signature of the tumor, allows the identification of patients with breast cancer, who are at high risk of disease recurrence (33). Contrary to the findings in breast, colon and lung cancer, in which the down-regulation of β-F1-ATPase expression is exerted at the level of translation (40) (Fig. 2), the silencing of β-F1-ATPase expression in leukemia is mediated by hypermethylation of the promoter of the gene (42) (Fig. 2). We suggest that alterations in the turnover of mitochondrial proteins are also likely to contribute to the abnormal bioenergetic phenotype of some human tumors.

Because the contribution that mitochondrial bioenergetics could have in cancer is still debated, we have recently generated colon cancer cell lines that express different levels of the β-F1-ATPase to assess the contribution of mitochondrial bioenergetics in cancer progression (20). The generated cells exhibit large ultrastructural, transcriptomic, proteomic and functional differences in their mitochondria and in their in vivo tumor forming capacity. We have confirmed that the activity of oxidative phosphorylation defines the rate of glucose utilization by aerobic glycolysis, supporting the relevance of the Pasteur Effect in cancer biology (Fig. 1) (20). The aggressive cellular phenotype, which is highly glycolytic, is bound to the deregulated expression of genes involved in energetic metabolism as well as in the regulation of the cell cycle, apoptosis, angiogenesis and cell adhesion (4, 7, 20). Remarkably, the molecular and ultrastructural analysis of the tumors derived from the different cell lines implanted highlighted that tumor promotion inevitably requires the selection of cancer cells with a repressed biogenesis and functional activity of mitochondria, i.e., the highly glycolytic phenotype is selected for tumor development. In others words, cancer cells with a functional bioenergetic activity of mitochondria are unable to promote tumor development (20) in agreement with findings that functional organelles halt proliferation (1, 3). In the same study, we demonstrate that the high glycolytic phenotype of the cells is a nongenetically acquired condition imposed by the cellular microenvironment that further provides a cell-death resistant phenotype (20). Indeed, the treatment of cancer cells with dichloroacetate partially restores the functional differentiation of mitochondria, halts proliferation and promotes tumor regression, further emphasizing the reversible nature of the metabolic trait of cancer and its potential Achilles heel (4, 20). Overall, these findings strongly support the tumor suppressor function of the bioenergetic activity of mitochondria (1).

The Impact of Cancer Genes on Energetic Metabolism

Studies aimed at establishing a link between malignant growth and the enforced glycolysis of cancer cells have uncovered diverse mechanisms that affect glycolysis, oxidative phosphorylation or both pathways (1, 8). For instance, gain-of-function mutations in the oncogene c-myc transactivate glycolytic genes as well as the glucose transporter GLUT1, enhancing both glucose uptake and lactate production (43) (Fig. 2). A direct role for c-myc in mitochondrial biogenesis and progression through the cell cycle and tumorigenesis has been described (16). Furthermore, deregulated c-myc also enhances glutaminolysis (44) (Fig. 2), which is a relevant pathway for cellular proliferation through the repression of miRNA 23a and miRNA23b (45). Deregulated c-myc is also responsible for the increased expression of the RNA binding proteins hnRNPI (PTB), hnRNPA1 and hnRNPA2 in human tumors (46) (Fig. 2). These proteins ensure the inclusion of exon 10 in the splicing of the pyruvate kinase gene rendering the PKM2 isoform (Fig. 2) (46) that diverts pyruvate from mitochondria and promotes aerobic glycolysis and tumorigenesis (22).

The rapid proliferation of cancer cells often generates a hypoxic microenvironment triggering the stabilization of the hypoxia inducible factor 1 alpha (HIF-1α) (47). HIF-1α induces the expression of many glycolytic enzymes, including LDHA and GLUT1 (Fig. 2), and cooperates with c-myc to induce the expression of pyruvate dehydrogenase kinase 1 (PDK1) and hexokinase II (HK-II), two additional paths that favor the glycolytic phenotype of cancer cells (48). Moreover, HIF-1α contributes to the repression of the biogenesis of mitochondria by inhibiting c-myc activity (49) and also affects mitochondrial respiration by switching the molecular composition of cytochrome c oxidase (50).

The oncogene Akt is a relevant molecule that signals both proliferation and resistance to cell death. Moreover, Akt produces a dose-dependent stimulation of glycolysis that correlates with tumor aggressiveness in vivo (Fig. 2) (51, 52). Mitochondrial respiratory defects promote the inactivation of PTEN, which leads to Akt activation resulting in drug resistance and survival advantage under hypoxic conditions (53).

Two independent studies have highlighted that the tumor suppressor p53 also has a metabolic role in cancer (Fig. 2) (54, 55). The product of a p53 targeted gene, TIGAR (TP53-induced glycolysis and apoptosis regulator) down-regulates glycolysis, and so, under conditions of p53 loss of function, glycolysis is increased (54). Likewise, loss of function of p53 results in decreased mitochondrial respiration and the metabolic shift to aerobic glycolysis (Fig. 2) (55). Recently, the tumor suppressor E3 ubiquitin ligase APC/C-Cdh1 has been shown to degrade the glycolysis promoting enzyme PFKFB3 and mutations in this gene are also likely to promote the shift of cancer cells to a Warburg phenotype (Fig. 2) (56).

Finally, epigenetic mechanisms that trigger the silencing of relevant genes involved in both gluconeogenesis and oxidative phosphorylation have been described to affect the metabolism of cancer cells (Fig. 2). Hypermethylation of the fructose-1,6-biphosphatase-1 (FBP1) promoter (57), an enzyme that antagonizes the flux of glycolysis (Fig. 1), and of the β-F1-ATPase gene promoter (42), that renders cells with a defective mitochondria have been reported (Fig. 2).

Overall, we suggest that the characterization of the mechanisms that promote the repression of mitochondrial activity in tumors will contribute to transform cancer into a chronic disease.