Re:Mitochondrial uncoupling in cancer cells: Liabilities and opportunities

[문형철]

beyond reason

Review

Mitochondrial uncoupling in cancer cells: Liabilities and opportunities☆

Author links open overlay panelGyorgyBaffy

https://doi.org/10.1016/j.bbabio.2017.01.005

Get rights and content

Under an Elsevier user license

open archive

Highlights

•

Uncoupling proteins are members of the solute carrier family of transporters located in the inner mitochondrial membrane

•

With the exception of UCP1-mediated thermogenesis in brown adipose tissue, the physiological role of UCPs remains debated

•

UCP2 is a key player in cellular energy metabolism through modulating the flow of TCA cycle and inducible proton leak

•

UCP2 may confer chemoresistance on cancer cells through metabolic reprogramming and reduced oxidative stress

•

Inhibition of UCP2 has been shown to boost the effects of chemotherapy and may represent a novel anti-cancer strategy

Abstract

Acquisition of the endosymbiotic ancestor of mitochondria was a critical event in eukaryote evolution. Mitochondria offered an unparalleled source of metabolic energy through oxidative phosphorylation and allowed the development of multicellular life. However, as molecular oxygen had become the terminal electron acceptor in most eukaryotic cells, the electron transport chain proved to be the largest intracellular source of superoxide, contributing to macromolecular injury, aging, and cancer. Hence, the ‘contract of endosymbiosis’ represents a compromise between the possibilities and perils of multicellular life. Uncoupling proteins (UCPs), a group of the solute carrier family of transporters, may remove some of the physiologic constraints that link mitochondrial respiration and ATP synthesis by mediating inducible proton leak and limiting oxidative cell injury. This important property makes UCPs an ancient partner in the metabolic adaptation of cancer cells. Efforts are underway to explore the therapeutic opportunities stemming from the intriguing relationship of UCPs and cancer. This article is part of a Special Issue entitled Mitochondria in Cancer, edited by Giuseppe Gasparre, Rodrigue Rossignol and Pierre Sonveaux.

• Previous article in issue
• Next article in issue

Keywords

Cancer cell

Metabolic flexibility

Mitochondrial respiration

Uncoupling

Hypoxia-inducible factor

Chemoresistance

1. Introduction

Just when it seemed that evolution had it all figured out — endosymbiosis-derived mitochondria provided a highly efficient mechanism to harness the energy of organic carbon, sufficient atmospheric oxygen has accumulated to make molecular oxygen the penultimate electron acceptor, and eukaryotic life has started to flourish on Earth — cancer entered the scene. Cancer is primarily a metazoan experience, a crisis of multicellular existence [1]. Cancer cells reject cooperation as they seek self-autonomy, evasion of growth control and programmed death, replicative immortality, and a limitless ability to spread and invade, eventually resulting in the demise of their host [2].

Mitochondria are prominently positioned in the history of cancer. These organelles were the prerequisite of eukaryotic complexity [3]. Mitochondria developed a precise balance between the molecular pathways of bioenergeticsand biosynthesis. In addition, however, mitochondria have become a major source of oxidative stress since superoxide is an inherent byproduct of their respiration in the presence of molecular oxygen [4]. This problem is augmented by disparate needs of energy metabolism in cancer cells and quickly becomes a liability when the rules of physiology are ignored to achieve greater metabolic flexibility that provides competitive benefits over normal cells. Uncoupling proteins (UCPs) may loosen the tight link (i.e., coupling) between mitochondrial respiration and ATP synthesis, allowing separate regulation of bioenergetic and biosynthetic requirements while limiting superoxide generation. Consequently, UCPs may have a key role in the adaptive metabolic response of cancer cells.

2. The contract of endosymbiosis: rise of the mitochondria

2.1. Evolutionary aspects

The pivotal event of endosymbiosis resulted in the first eukaryotic cell about 2 billion years ago (2 Gya) in the Paleoproterozoic period when initial oxygenation of the Earth's atmosphere has begun but molecular oxygen (O2) had a limited presence [1], [5]. Thus, the early evolution of α-proteobacterial and archaeal ancestors took place in a mostly anaerobic environment. Charcoal evidence suggests that a second rise in atmospheric oxygen to levels allowing combustion (~ 17% pO2) did not occur until much later (~ 0.5 Gya), in the Phanerozoic period [6]. Fossil records of early eukaryotes indicate that their radiation into plants, fungi, and metazoa may have started long before our planet became truly oxygenated [7]. In all likelihood, anaerobic biochemistry had a prodigious presence in early eukaryotes.

Perhaps the greatest benefit of endosymbiosis was the development of mitochondria, the double-membrane organelles that became the energetic powerhouse in eukaryotic cells [5]. Acquisition of the α-proteobacteriaprovided the Archaean host with components of the electron transport chain (ETC), a molecular system that extracts metabolic energy by using terminal electron acceptors (e.g., proton, nitrates, and later O2) obtained from the environment in a process termed respiration [7], [8]. Mitochondria represent a large surface area of internal membranes with ETC components encoded by myriads of small α-proteobacterial genomes (mitochondrial DNA) that offer high ATP yield for the evolutionary transition from single cells to multicellularity [3]. However, this evolutionary achievement has a major downside since the ETC is also the largest intracellular source of superoxide, contributing to macromolecular injury, aging, and cancer [9], [10]. Hence, the ‘contract of endosymbiosis’ represents a fundamental compromise for multicellular life.

2.2. Oxidative phosphorylation

The mitochondrial machinery of oxidative phosphorylation consists of four protein complexes (I to IV) in the ETC in addition to complex V (ATP synthase), all located in the inner mitochondrial membrane (Fig. 1). Most of our current understanding of oxidative phosphorylation is derived from Mitchell's chemiosmotic hypothesis [11], [12]. The oxidative metabolism of glucose, fatty acids, and amino acids provides the ETC with reducing equivalents (NADH for complex I and FADH2 for complex II) that pass electrons to successive acceptors in a series of exergonic reactions until they reach O2 at the level of complex IV (cytochrome c oxidase). Meanwhile, protons are pumped out into the inter-membrane space to create a protonmotive force or electrochemical gradient (Δψm) across the mitochondrial inner membrane. The energy of this proton gradient drives complex V that converts ADP to ATP. The latter is transported to the cytosoland exchanged for new ADP through the mitochondrial ATP/ADP exchanger or adenine nucleotide translocator (ANT). If mitochondrial respiration is impaired, ATP synthase and ANT may work in reverse mode fueled by mitochondrial ATP uptake and hydrolysis to maintain Δψm by pumping protons out of the matrix [13].

1. Download : Download high-res image (208KB)
2. Download : Download full-size image

Fig. 1. Mitochondrial oxidative phosphorylation and uncoupling.

Electron transport chain complexes (I to IV) pass substrate-derived electrons onto molecular oxygen (respiration) and build a chemiosmotic gradient (Δψm) by translocating protons (+) across the inner mitochondrial membrane (blue line). Proton flow back to the matrix fuels the conversion of ADP to ATP by complex V (ATP synthase) and facilitates electrogenic ADP/ATP exchange by the adenine nucleotide translocator (ANT). Slow electron transport due to high Δψm (substrate excess), impaired activity of complex IV (hypoxia), or limited ATP synthesis (insufficient ADP) results in premature electron spinoff and partial reduction of O2 into superoxide (O2•-), which may be released both into the matrix and the cytosol. Uncoupling proteins(UCPs) are specialized mitochondrial carriers that mediate inducible proton leak (possibly in exchange for C4 metabolites derived from the tricarboxylic acid cycle) and limit the rate of superoxide generation by removing some of the physiologic constraints of mitochondrial respiration.

Major mitochondrial functions such as protein and metabolite transport, oxidative phosphorylation, ion homeostasis, or initiation and execution of apoptosis critically depend on Δψm [14]. Oxidative phosphorylation must occur in a coordinated (coupled) fashion, otherwise a mismatch between the inflow of metabolic substrates and the amount of ADP available in the mitochondrial matrix may lead to adverse changes in Δψm. The near-equilibrium hypothesis of oxidative phosphorylation predicts that mitochondrial respiratory rate is determined by substrate supply ([NADH]/[NAD+]), the mitochondrial phosphorylation potential ([ATP]/[ADP][Pi]), and the activity of cytochrome c oxidase ([c2 +]/[c3 +], where c2 + and c3 +denote reduced and oxidized cytochrome c, respectively) [15]. Substrate supply and ATP utilization stimulate respiration rates with an opposing effect on Δψm. When the rate of NADH production is not matched by the rate of phosphorylation (e.g., complex IV activity is impaired due to low pO2 or there is lack of ADP), Δψm may become too high and slows down the ETC flow of electrons with more chance for premature electron leakage and excessive superoxide generation [16].

2.3. Mitochondrial ROS generation

Production of superoxide anion through incomplete reduction of O2 is an important collateral event of mitochondrial electron transport. Mitochondria generate superoxide through the escape of electrons at complex I from iron-sulfur groups, at complex II from flavin-containing proteins, and at complex III from the ubiquinone (Q) cycle [17]. It is estimated that about 2% of the oxygen consumed by the mitochondria is reduced by the bifurcated electrons to form superoxide, which is subsequently converted to hydrogen peroxideand other reactive oxygen species (ROS) [18]. Measurements in isolated mitochondria indicate that 70–80% of the superoxide formation may be related to the generation of semiquinone, a highly reactive intermediate of the Q cycle in complex III. Since semiquinone is located at the outer side of the inner membrane (Qo), complex III is the key source of mitochondrial ROSdirectly released to the cytosol. Unimpeded electron flux gives the electrons less time to reside at sites where superoxide is generated, while higher Δψm due to excess substrate supply or mismatched ATP synthase activity results in longer half-life of ETC intermediates, adding to the risk of ROS generation [19]. Since rapidly growing and proliferating cancer cells process large amounts of metabolic substrates while facing a variable degree of hypoxia, they are particularly susceptible to these physiologic restraints.

It is difficult to overstate the significance of mitochondrial ROS production in eukaryotic evolution. Once operating in increasingly oxygenated environments, the α-proteobacterial endosymbiont had become a potentially lethal weapon to the wellbeing of its archaeal host. Thus, eukaryotic cells evolved a variety of protective cellular mechanisms such as the use of superoxide dismutase and redox buffering systems, partial beta-oxidation of long-chain fatty acids in the peroxisomes, and transfer of α-proteobacterial genes to the nuclear genome to avoid DNA damage [20]. As discussed below, UCPs with their ability to modulate mitochondrial ROS generation represent a special form of protection from oxidative stress.

2.4. Mitochondria under hypoxia

For an extended period, our eukaryotic ancestors lived in a world where oxygen was limited or unavailable. To this day, O2 as the terminal electron acceptor has remained an opportunity rather than an inevitable choice for some eukaryotic species. For instance, the malaria parasite Plasmodium falciparum has a fully functional ETC in their mitochondria, but generates ATP via substrate-level phosphorylation by fermenting lactate [5]. As eukaryotes evolved into multicellular life, however, molecular oxygen has become indispensable for most species and their survival in severely hypoxic or anoxic environments is no longer sustainable. Facultative anaerobeorganisms such as Saccharomyces cerevisiae exhibit the Pasteur effect: while these yeast cells produce ATP by the energetically more favorable oxidative phosphorylation in the presence of oxygen, they readily switch to fermentation in its absence [5]. The explanation of Pasteur effect lies in the allosteric inhibition of phosphofructokinase, a key glycolytic enzyme, by the ATP and citrate produced in the mitochondria.

Interestingly, cytochrome c oxidase has a high affinity for O2 (Km < 2 mm Hg), which means that cells may maintain normal levels of O2 consumption unless substantial hypoxia develops [21]. Accordingly, the respiratory rate is essentially independent of pO2 and remains a factor of the rate of ATP utilization (i.e., ADP supply) until the environment is nearly anoxic [22]. However, respiratory rates become independent of ADP availability in uncoupled mitochondria. The role of uncoupling in this situation is to protect Δψm from rising too high by allowing an ADP-independent electron flow within the ETC (a.k.a. state 4 respiration). Under these conditions, ETC activity has no metabolic constraints other than pO2. Paradoxically, hypoxia may aggravate mitochondrial ROS production due to a relative lack of the terminal electron acceptor O2 and slow electron transport, which prolongs the half-life of ETC intermediates such as semiquinone (of course, no superoxide is formed in full anoxia). Even a modest increase in proton leak may reduce Δψm and limit the formation of superoxide, indicating that mitochondrial uncoupling is a powerful tool to control this process [23]. However, uncoupling of ETC from ATP synthesis to protect from ROS represents a potentially costly bioenergetic compromise. In addition, ROS have important benefits as they may act as signaling mediators in numerous molecular pathways of cellular adaptation [24].

3. Mitochondrial uncoupling proteins

3.1. Origins of UCPs

Uncoupling proteins, along with the mitochondrial ATP/ADP exchanger ANT, belong to the SLC25 group of the large solute carrier (SLC) family of transporters [25]. SLC25 proteins are widely found in eukaryotes and they are characterized by a signature threefold repeat (tripartite) structure that includes a total of six transmembrane α-helices. The presence of UCPs in Acanthameoba castellanii, which is phylogenetically located next to the divergence of plants, animals, and fungi, suggests that these specialized mitochondrial carriers emerged no later than in the last common eukaryotic ancestor, making it likely that they exist in all eukaryotes [26]. For the same reason, UCPs had a long time to evolve separately, which probably explains the diversity of cellular functions carried out by these proteins [27].

Similar to other members of the nuclear-encoded SLC25 family, UCPs are targeted to the mitochondria and become embedded in the inner mitochondrial membrane as homodimers [28]. Here they may interact with other membrane-bound macromolecules such as the components of mitochondrial ETC. The prototype uncoupling protein UCP1 is a highly abundant constituent of brown adipose tissue, making up to 5% of its total mitochondrial protein content [29]. UCP1 has a pivotal role in non-shivering thermogenesis by increasing the permeability of mitochondrial inner membrane to protons and dissipating heat from the metabolic energy accumulated in brown adipose tissue [30]. UCP1-mediated proton leak is responsive to cold or various hormonal activators and it is greatly enhanced by fatty acids, while purine nucleotides such as ATP and GDP have an inhibitory effect. The precise mechanism by which UCP1 is able to move protons through the inner membrane is not completely understood [30]. Importantly, the highly specialized role of UCP1 probably evolved as an extreme form of proton conductance and there may be more ancestral functions shared by other UCPs.

Two additional UCPs, UCP2 and UCP3 were identified in 1997 by cloningstrategies in the mouse and human genomes based on partial sequence homology with UCP1 [31]. The description of UCP4 and UCP5 followed soon thereafter [32], [33] with further UCPs identified in disparate groups of eukaryotes [34]. The expression‎ of UCPs is regulated both at transcriptional and translational levels. Human UCP2 is a 308 amino acid protein with molecular weight of 33 kDa [35]. UCP2 is fairly ubiquitous, although its protein abundance is one to two magnitudes smaller when compared to the amount of UCP1 in brown adipose tissue. UCP3 is specific for cardiac and skeletal muscle, while UCP4 and UCP5 are mainly expressed in the brain [34].

3.2. Physiology of UCPs

The physiological roles of UCPs beyond those of UCP1 remain hotly debated [34], [36], [37]. It has been pointed out that UCPs in unicellular eukaryotes could not have a thermoregulatory role in the absence of a meaningful temperature gradient between the cell and external environments [28]. Many studies suggest that a more fundamental effect of UCPs could be the lowering of Δψm with a subsequent protection from ETC-derived generation of ROS [38]. Based on it broad tissue distribution, UCP2 has been under particular scrutiny as a potential system-wide regulator of mitochondrial oxidative stress [39]. Loss-of-function studies indicate that UCP2 is not responsible for basal proton leak across the mitochondrial inner membrane [38]. However, UCP2 mediates variable amounts of net proton conductance when activated by free radicals such as superoxide or the lipid peroxidation end-product hydroxynonenal. Moreover, the half-life of UCP2 is only about 30 min, which enables this molecule to mediate rapid biological responses [40].

The idea of UCP2 as a major regulator of mitochondrial oxidative stress is supported by extensive experimental evidence. Thus, absence of UCP2 is associated with increased ROS production in a variety of cell types, while enhanced UCP2 expression‎ was found to protect cells from injury and death related to oxidative stress [41], [42]. The notion that protection from mitochondrial ROS generation is an inherent biological function of UCP2 as a mitochondrial anion carrier is consistent with the concept of ‘mild uncoupling’ [43]. In an effort to explain the physiological impact of uncoupled mitochondrial respiration, Skulachev proposed that a small decrease in Δψm due to controlled mitochondrial proton leak (i.e., mild uncoupling) may not substantially affect the rate of ATP production, but still be sufficient to limit mitochondrial superoxide production and serve as a cellular defensemechanism from oxidative injury [43]. The idea of physiological uncoupling has been extended to other mitochondrial carriers that are involved in proton exchange such as the ANT and the aspartate/glutamate antiporter [44].

3.3. Novel concepts about UCP2

Recent findings suggest that the biological role of UCP2 in cell metabolismmay not be limited to inducible proton conductance [37]. Involvement of UCP2 in lipid metabolism was proposed based on observations in liver cells where the presence or absence of UCP2 was linked to intracellular lipid accumulation [45]. Later, UCP2 was implicated in the fasting-induced shift of substrate utilization from glucose oxidation to fatty acid oxidation [46], [47]. In addition, UCP2-deficient macrophages exhibit lower rates of mitochondrial oxidation of glutamine [48]. Moreover, lack of UCP2 in murineembryonic fibroblasts results in increasing glucose dependence and decreased rates of fatty acid oxidation [37]. These observations suggest that UCP2 could promote a metabolic switch from glucose oxidation to glutamine and lipid utilization. The idea was then forwarded that the impact of UCP2 on substrate utilization may spare pyruvate for biosynthetic needs and prevent its complete mitochondrial breakdown [49]. According to this proposition, UCP2 is a metabolic tool that makes aerobic cells become glycolytic and use glutamate or fatty acids for their need of mitochondrial energy production, which is rather similar to the metabolic changes seen in cancer cells.

In a recent work, silencing of UCP2 in HepG2 human liver cancer cells grown in glucose medium predictably resulted in increased Δψm and ATP/ADP ratios when compared to wild-type cells, while these parameters changed in opposing direction if cells were grown in glutamine medium, indicating less efficient oxidative phosphorylation [50]. Surprisingly, mitochondrial levels of TCA intermediates 2-oxoglutarate, aspartate, and L-malate were higher in both glucose and glutamine medium, suggesting that UCP2 promotes the cataplerotic efflux of these C4 metabolites out of the mitochondria. Moreover, UCP2 reconstituted in liposomes and expressed in yeast revealed that TCA intermediates are exchanged for cytosolic phosphate by a proton-assisted mechanism [50]. These findings identify UCP2 as a transport protein that augments cytosolic utilization of glucose over its mitochondrial oxidation. In this scenario, however, increased glutaminolysis becomes essential as it helps a highly cataplerotic TCA cycle to replenish oxaloacetate and may also permit increased rates of fatty acid oxidation [51], [52]. It should be pointed out that the proposed C4 transport function of UCP2 involves the movement of protons, which is consistent with many prior observations. These findings are highly relevant to the biology of cancer cells and confirmation for broader applicability is awaited [50].

4. Breach of the endosymbiotic contract: the rise of cancer

Cancer cells face a major evolutionary dilemma: to exploit the energetic benefits of oxidative phosphorylation or switch back to substrate-level phosphorylation that provides more metabolic flexibility [53]. The aerobic biochemistry of mitochondria is regulated with precision to meet the needs of the eukaryotic cell with high-yield generation of ATP and macromolecular biosynthesis. Cancer cells tend to avoid physiologic regulatory constraints and seize most of the environmental resources shared with their normal cell counterparts. Thus, cancer cells adopt a proliferative phenotype that is increasingly reminiscent of our prokaryotic ancestors. On this backward way along the phylogenetic tree, cancer cells are ready to hijack any molecular tool that may support their growth and survival [54]. Based on what we know about the biology of UCPs, it is important to understand how they may fit this uncompromising, competitive agenda of cancer cells and whether this association can identify UCPs as a suitable molecular target of oncotherapy.

4.1. Metabolic reprogramming in cancer cells

Eukaryotic cells derive their energy from a series of successive biochemical reactions in the cytosol and mitochondria (Fig. 2). These pathways also provide a platform for the biosynthesis of macromolecules such as proteins, lipids, and nucleic acids. In quiescent cells, pyruvate from the glycolytic breakdown of glucose as a preferential substrate is transported to the mitochondrial matrix where it is converted to acetyl-CoA by the pyruvate dehydrogenase (PDH) complex. This step commits pyruvate to complete oxidation in the mitochondria. The metabolic energy of acetyl-CoA is captured in the TCA cycle in the form of NADH and FADH2 to feed the ETC and fuel ATP synthesis. An alternative NADH source is acetyl-CoA gained from fatty acid oxidation. In proliferating cells, much of the carbon in the TCA cycle is used for biosynthetic pathways of non-essential amino acids and fatty acids, resulting in continuous efflux of intermediates (cataplerosis). To sustain the TCA cycle, there has to be a proportionate influx of intermediates (anaplerosis) from pyruvate converted to oxaloacetate by the pyruvate carboxylase or from glutamine sequentially converted into alpha-ketoglutarate (glutaminolysis).

1. Download : Download high-res image (586KB)
2. Download : Download full-size image

Fig. 2. Uncoupling proteins and metabolic reprogramming in cancer cells.

Major differences in the relative significance of key biochemical pathways in the cytosoland mitochondria of normal cells and cancer cells are illustrated by the width of gray background. Preferential metabolic nodes are shown in yellow. In normal cells (left panel), glucose uptake (1) and glycolysis (GL) are strictly regulated and moderately active. Biosynthetic needs are proportionate to physiologic demands. The glycolysis end product pyruvate is transported to the mitochondria (2) and converted by the pyruvate dehydrogenase complex (PDH) to acetyl-CoA, which is oxidized in the tricarboxylic acid cycle (TCA). This process yields NADH for the electron transport chain (ETC), which extrudes protons (+) to the mitochondrial intermembrane space and sustains the mitochondrial membrane potential (Δψm) that drives the ATP synthase (FoF1), making oxidative phosphorylation the main source of ATP in normal cells. In cancer cells (right panel), glucose uptake and glycolysis rates are vastly upregulated and yield large amounts of pyruvate that is mostly converted to lactate by the lactate dehydrogenase(LDH) and transported out of the cell (3). Much of ATP production in cancer cells is derived from glycolysis. High biosynthetic needs in cancer cells require an increasingly active pentose phosphate pathway (PPP) and a cataplerotic TCA cycle, which is mainly replenished (anaplerosis) by α-ketoglutarate resulting from glutamine uptake (4) and by oxaloacetate derived from pyruvate via the pyruvate carboxylase (PC). Since ADP in cancer cells is primarily converted to ATP by glycolysis in the cytosol, mitochondrial ADP uptake through the adenine nucleotide translocator (ANT) and subsequent ATP synthesis is restricted (Crabtree effect) and less able to match mitochondrial respirationand proton pumping activity. To avoid mitochondrial hyperpolarization (high Δψm) and excessive superoxide generation by the ETC, cancer cells may rely on induced proton leak by uncoupling proteins, primarily UCP2, which supports metabolic flexibility by allowing high rates of mitochondrial biosynthesis without the constraints of oxidative phosphorylation. Recent data suggest that efflux of C4 metabolites from the TCA cycle promoted by UCP2 may contribute to this process.

Metabolic and bioenergetic needs are very different in cancer cells characterized by rapid growth and sustained proliferation [51], [52]. In the 1920's Warburg described a link between mitochondria and cancer. He found that cancer cells preferentially utilize glycolysis over oxidative phosphorylation to produce energy even in the presence of oxygen [55]. This phenomenon, termed aerobic glycolysis or the ‘Warburg effect’ is in essence a lack of the Pasteur effect and has been recognized as a major hallmark of cancer. It has become increasingly clear that aerobic glycolysis is rarely a sign of defective oxidative phosphorylation as originally presumed [56]. Shortly after Warburg's observations, high rates of glycolysis in cancer cells (and in non-cancerous proliferating cells) were found to inhibit mitochondrial respiration, a phenomenon later termed the ‘Crabtree effect’ [57]. Crabtree suggested that glucose-induced inhibition of oxygen consumption explains the Warburg effect and there is no need to consider a mitochondrial respiratory defect. More recently it was suggested that the Crabtree effectresults from competition between oxidative phosphorylation and glycolysis for ADP and inorganic phosphate [58].

Much of the energetic needs in cancer cells are met by increased glycolysis. Accordingly, the rate of glucose entry to cancer cells is 20-to-30-fold higher than in normal cells [59]. In addition, key glycolytic enzymes such as hexokinase are heavily upregulated in cancer cells. Most of the glycolytic end product pyruvate in cancer cells is converted to lactate by the lactate dehydrogenase (LDH) and exported out of the cell (e.g., through the monocarboxylate transporter MCT4). Glycolysis supplies the pentose phosphate shunt (hexose monophosphate pathway), a gateway for nucleotide synthesis and a source of NADPH needed for many biosynthetic reactions and for sufficient redox capacity [52]. Importantly, cancer cells may abandon oxidative phosphorylation but retain functional mitochondria and keep Δψm in the physiological range for full biosynthetic competence. Vigorous biosynthesis requires intensified anaplerosis to replenish the TCA cycle, which involves pyruvate carboxylase and glutaminolysis, the latter accounting for the avidity of cancer cells for glutamine [60], [61].

The Warburg effect is not absolute and glycolysis typically accounts for no more than 60% of total cellular ATP production in most cancer cells [62]. Targeted depletion of mitochondrial DNA that encodes several ETC components has been shown to weaken cancer cells both in vitro and in vivo [61]. Exclusive conversion of glucose to lactate has also been questioned by observations in glioblastoma, lung cancers, and liver cancers [63], [64], [65]. The extent to which cancer cells may utilize the Warburg phenotype may depend on the stage of malignancy. An interesting example for the metabolic flexibility of cancer cells is the reverse Warburg effect, described in metastatic cancer [66]. In this condition, some cancer cells take up high-energy mitochondrial fuels (L-lactate and ketone bodies) released by the ROS-mediated destruction of neighboring stromal cells. These fuels are then converted to acetyl-CoA and then fully oxidized in cancer cell mitochondria [66]. Additional observations indicate that cancer cells exhibit hierarchical metabolic heterogeneity whereby MCT4-expressing tumor cells perform glycolysis and secrete lactate via MCT4 and MCT1-expressing cells import lactate via MCT1 and metabolize it within their mitochondria [67]. Thus, cancer cells may re-engage their oxidative phosphorylation machinery as they reach advanced levels of malignancy.

4.2. Cancer cells in adversity: oxidative stress and hypoxia

Cancer cells typically encounter harsh microenvironments due to nutrient limitation, hypoxia, and mechanisms of host defense. Most cancer cells are characterized by increased levels of intracellular ROS [68]. This generates a substantial degree of oxidative stress in cancer cells, which has pleiotropic outcomes and contributes to heterogeneous cell populations. ROS promote genomic instability by causing mutations and other types of DNA injury. Higher levels of ROS facilitate the progression of chronic inflammation to malignant transformation [69]. In addition, oxidative modifications of redox-sensitive transcription factors such as NF-κB or intermediate signaling molecules such as PKC, ERK, and JNK may stimulate pathways of cell growth and cell survival [24], [70], [71]. By contrast, ROS also activate cellular effectors such as the p53 tumor suppressor protein that initiate cell death or cause growth arrest in response to oxidative stress and excessive amounts of ROS may inflict fatal structural damage [72], [73]. Balancing these opposing biological effects of ROS is an important adaptive ability of cancer cells [74].

A substantial amount of increased ROS in cancer cells is derived internally from the mitochondrial ETC. Indeed, there are several reasons why cancer cells may have a chronic mismatch between their rates of mitochondrial respiration and ATP synthesis: (1) cancer cells with an accelerated metabolism rapidly accumulate reducing equivalents such as NADH that may generate an electron flux exceeding ATP utilization rates; (2) cytochrome c oxidase have limited access to the terminal electron acceptor O2 in hypoxia; and (3) excess needs of biosynthetic pathways leave no ADP available for the mitochondria (Crabtree effect) [1], [51], [52]. With any and all of these changes, there is a higher probability of electron spinoff with premature and incomplete reduction of O2 into superoxide. While higher ROS levels may initially provide cancer cells with a competitive edge due to genomic instability and pro-oncogenic signaling, increasing oxidative stress ultimately results in the cell's demise unless escalation of ROS is controlled (Fig. 3).

1. Download : Download high-res image (126KB)
2. Download : Download full-size image

Fig. 3. Metabolic flexibility and oxidative stress in cancer cells.

Schematic diagram illustrating the effect of mitochondrial uncoupling proteins (UCPs) on ROS levels in cancer cells. Physiologic constraints result in tightly balanced mitochondrial energy metabolism and low-level oxidative stress in normal cells. Mismatch in this balance increasingly promotes ROS generation in cancer cells that favor metabolic flexibility and mitigate oxidative stress by additional protective mechanisms such as upregulation of uncoupling proteins. Note that low ROS levels may quickly escalate with mismatched energy metabolism in normal cells, which are less robust than cancer cells.

Since blood vessel formation and capillary supply cannot keep pace with rapid tumor growth, cancer cells often encounter different levels of tissue oxygenization, ranging from normoxia (pO2 ~ 2–4%), through hypoxia, to anoxia (pO2 < 0.1%) [75]. Cancer cells located a few hundred μm from their feeding capillaries routinely face severe hypoxia [76]. Adaptation to hypoxic environments is therefore essential for cancer cells and hypoxia-inducible factor (HIF) plays an important role in this process [77]. HIF is a key transcriptional regulator that mediates a complex homeostatic response to hypoxia in all metazoan species [78]. Several members of the HIF family have been identified, HIF-1 being the most characterized. HIF-1 is a heterodimer consisting of a hypoxia-responsive HIF-1α subunit and a constitutively expressed HIF-1β subunit [78]. HIF-1α has a high turnover rate and is primarily regulated at the protein level by proteasomal degradation [79]. In normoxia, HIF-1α is hydroxylated by cytosolic enzymes with a prolyl hydroxylase domain (PHD). HIF-1α is additionally hydroxylated at an asparagine residue by factor-inhibiting HIF (FIH), a member of the 2-oxoglutarate- and O2-dependent hydroxylase family [80]. These covalent modifications allow HIF-1α to be recognized by the von Hippel-Lindau tumor suppressor protein (pVHL) and targeted for degradation, which prevents HIF-1 dimerization and binding to transcriptional coactivators such as p300 and CREB-binding protein [81].

In hypoxia, HIF-1α escapes hydroxylation and recognition by pVHL, forms a dimer with HIF-1β and translocates into the nucleus [82], [83]. While PHD inactivation occurs at moderately low pO2, deeper levels of hypoxia may be needed to block FIH. Thus, asparagyl hydroxylation may act as a safety switch to prevent accidental activation of hypoxia-regulated genes. It has been speculated that the spectrum of gene expression‎ in response to HIF-1α action may be linked to the variable sensitivity of prolyl and asparagyl hydroxylases to hypoxia [84]. Of note, the HIF pathway is also responsive to non-hypoxic stimuli such as hormones and growth factors. Indeed, several tumors display high HIF activity even in the presence of sufficient oxygen [83], [85]. This situation, known as pseudohypoxia, is most evident in cancer cells with a loss of one of the tumor suppressor proteins pVHL, succinate dehydrogenase or fumarate dehydrogenase [86].

The paramount importance of HIF-1 stems from its ability to singlehandedly coordinate a number of genes encoding key components of metabolic adaptation in cancer cells [77], [87]. HIF regulates the balance between glycolytic and oxidative metabolism, tailoring the specific needs of rapidly growing cancer cells in a microenvironment that bears much similarity to the anaerobic world of early eukaryotic ancestors. HIF-1 induces glucose transporters (e.g., GLUT1 and GLUT3) and activates key glycolytic enzymes (e.g., hexokinase II), while it diverts pyruvate from breakdown in the TCA cycle by activating pyruvate dehydrogenase kinase 1 (PDK1) and hence inactivating PDH. HIF-1 also activates LDH and MCT4, facilitating the production of lactate from pyruvate and its transport out of the cell. HIF-1 inhibits mitochondrial biogenesis via Myc degradation and promotes mitochondrial autophagy (mitophagy) by inducing BNIP3 and NIX [88], [89]. These changes produce a strong Warburg phenotype and help maintain a balance between TCA cycle flux and ETC capacity, with less mitochondrial ROS generation. At the same time, HIF-1 also facilitates changes aimed at improving ETC function. HIF-1 activates transcription of the cytochrome coxidase gene encoding COX4-2 and LON, a mitochondrial protease that is required for degradation of COX4-1. This adaptive response maximizes the efficiency of respiration since COX4 subunit switching tailors cytochrome coxidase activity to reduced O2 availability [90].

5. UCPs and cancer

5.1. UCPs as biomarkers of oncogenesis and chemoresistance

As evidence began to accumulate about the role of UCP2 in cellular energy metabolism and mitochondrial oxidative stress, the link between UCP2 and cancer has increasingly become the focus of investigations [91]. Consistent with the biology of ROS, UCP2-deficient mice showed persistent inflammation and activation of NF-κB activation in their colon with a predisposition for chemically induced tumorigenesis [92]. Repression of UCP2 levels in the estrogen receptor (ER)-positive breast cancer cell lineMCF-7 by estrogen, a key factor of breast cancer initiation, was linked to oxidative injury and cell proliferation [93]. By contrast, stronger UCP2 expression‎ was found in human colorectal cancer compared to peritumoral tissue controls and UCP2 levels increased with the degree of neoplastic changes along the colon adenoma-carcinoma sequence [94]. Similar correlation of UCP2 levels with clinical stages of colon cancer was also described [95]. Increased UCP2 expression‎ was subsequently described in leukemia as well as in cancers of the breast, ovaries, bladder, esophagus, pancreas, kidney, testicles, lung, prostate and skin [96], [97]. These observations suggest that upregulation of UCP2 is widespread in cancer with a link to neoplastic progression.

Experimental reports on the association between carcinogenesis and UCPs are not limited to UCP2. In prostate cancer cell lines induced for osteoblastogenic and adipogenic differentiation, UCP1 expression‎ correlated with disease progression from primary to bone metastatic cancers [98]. UCP4 positivity was a reliable marker of lymph node metastases in breast cancer [99]. Postmenopausal breast cancer that simultaneously displayed a low ER status and higher UCP5 expression‎ had a worse prognosis [100]. In an interesting study, UCP3 transgene was targeted to the basal epidermis to investigate the effect of nutrient wasting on skin carcinogenesis due to excessive mitochondrial uncoupling and futile mitochondrial respiration[101]. Mechanistic studies revealed that UCP3 overexpression‎ resulted in increased fatty acid oxidation, breakdown of the phospholipid membrane and impaired Akt recruitment to the plasma membrane, while Akt overexpression‎ overcame the effect of forced uncoupling [101]. Whether this strategy has a therapeutic potential remains to be seen.

Overexpression‎ of UCP2 in a human colon cancer cell line was shown to blunt the pro-oxidative and pro-apoptotic effects of topoisomerase I inhibitor CPT-11 and conferred chemoresistance to tumor xenografts [102]. Interference with ROS-responsive post-translational modification of the tumor suppressor p53 and a shift towards the glycolytic phenotype contributed to prosurvival advantage in these cancer cells [102]. Thus, UCP2 may promote chemoresistance by abrogating the ROS-mediated positive feedback loop of the p53 response.

5.2. Mitochondrial uncoupling and the hypoxia response in cancer cells

The role of mitochondrial uncoupling in HIF stabilization adds to the controversial relationship of UCP2 and cancer. Hagen et al. proposed a model in which reduced mitochondrial respiratory rates (e.g., by nitric oxide-mediated blockade of cytochrome c oxidase) may lead to intracellular redistribution of O2 and subsequent activation of cytosolic prolyl hydroxylases, accounting for HIF-1 degradation and the failure to register hypoxia [103]. Accordingly, factors that stimulate ETC activity such as substrate excess or uncoupling are likely to promote HIF stabilization. Thus, increased UCP2 activity in cancer cells may contribute to an excessive HIF response despite normoxic conditions (pseudohypoxia) through intracellular O2 redistribution. A competing proposal invoked the role of mitochondrial ROS in the regulation of HIF based on observations that ρ° cells, depleted of their mitochondrial DNA and unable to generate ROS, failed to stabilize HIF-1α [104]. This mechanism was unequivocally demonstrated in studies with cytochrome b cybrids, which are incapable of oxygen consumption while still generate ROS under hypoxia [105]. Further studies with site-specific inhibitors (leaving ETC activity intact) confirmed that abrogation of ROS generation by complex III is sufficient to inhibit HIF stabilization [106]. These observations explain HIF-1 activation by the hypoxia-induced surge in mitochondrial ROS production, but assume an opposing effect on HIF regulation by UCPs and mitochondrial uncoupling in general (Fig. 4). However, how UCPs affect the HIF response in tumor tissues facing deeper hypoxia or near-anoxia remains unclear. In addition, limiting ROS-mediated HIF-1 stabilization is not always a beneficial strategy since in some instances higher HIF-1α expression‎ is a good prognostic factor, correlating with lower cancer stage or improved patient survival, as it has been seen in a variety of cancers including non-small-cell lung cancer, head and neck squamous cell carcinoma, and neuroblastoma [107].

1. Download : Download high-res image (228KB)
2. Download : Download full-size image

Fig. 4. Uncoupling proteins and the stabilization of hypoxia inducible factor.

In normoxia, alpha subunit of hypoxia-inducible factor 1 (HIF-1α) is tagged for proteasomal degradation by prolyl and asparaginyl hydroxylases (PHD and FIH) in the cytosol. Hypoxia inhibits these enzymes, allowing heterodimerization and nuclear translocation of HIF-1α. The role of uncoupling proteins (UCPs) in this process remains controversial. UCPs stimulate mitochondrial respiration and promote HIF-1α stabilization through intracellular O2 redistribution and relative cytosolic hypoxia. However, the effect of UCPs on mitochondrial superoxide generation may weaken ROS signals, which are known to prevent HIF-1α from degradation.

5.3. UCPs as therapeutic targets in cancer

Several in vitro studies indicate that inhibition of UCP2 by genetic or chemical means may provide a unique opportunity to enhance the efficacy of chemotherapeutic agents and kill cancer cells by exposing them to additional oxidative injury (Table 1). Manipulation of endogenous UCP2 levels by overexpression‎ or silencing altered the sensitivity of various human hepatocarcinoma cells to gemcitabine-induced mitochondrial ROS production and apoptosis [108]. Intrinsic UCP2 mRNA abundance correlated with IC50 values for gemcitabine toxicity in cell lines derived from pancreatic adenocarcinoma, non-small-cell lung cancer, and bladder carcinoma [109]. UCP2 overexpression‎ strongly decreased gemcitabine-induced mitochondrial superoxide formation and protected these cancer cells from apoptosis. Interestingly, UCP2 expression‎ was stimulated by gemcitabine treatment in this study, suggesting that UCP2 may also play a role in acquired chemoresistance [109]. In experiments with human-derived colon cancer cells, UCP2 inhibition by the naturally occurring agent genipin augmented the effects of cisplatin [110] while use of the small molecule UCP inhibitor chromanes increased the toxicity of arsenic trioxide and anthracyclinederivatives [111]. Combination of cytotoxic agents with inhibition of UCP2 by genipin or mRNA silencing in cell lines from pancreatic adenocarcinoma and breast cancer resulted in increased ROS levels, cell growth inhibition and autophagy [112], [113]. Glutathionylation of UCP2 by diamide sensitized drug resistant leukemia cells to a variety of chemotherapeutic agents [114]. In vivo confirmation of these promising results will hopefully open the road to novel treatment strategies in cancer.

Table 1. Effect of UCP2 inhibition on chemotherapy in human cancer cells.

Compilation of recent studies in which in vitro combination of chemotherapeutic agents with genetic or chemical inhibition of UCP2 resulted in increased rates of cytotoxicity and cell death.

Cell lines	Mode of UCP2 inhibition	Chemotherapeutic agent with enhanced effect	Reference
Hepatocellular carcinoma (HuH6, Hep3B, HepG2)	mRNA silencing	Gemcitabine	[108]
Pancreatic adenocarcinoma (PaCa44), non-small cell lung cancer (A541), bladder carcinoma (RT112)	Genipin, mRNA silencing	Gemcitabine	[109]
Pancreatic adenocarcinoma (PaCa44)	Genipin	Gemcitabine	[112]
Colon adenocarcinoma (HT-29, SW-620)	Genipin	Cisplatin	[110]
Colon adenocarcinoma (HT-29)	Chromanes (CSIC-E-379)	Arsenic trioxide, cisplatin, doxorubicin	[111]
Colon adenocarcinoma (HCT-116)	Genipin	Cisplatin	[119]
Acute promyelocytic leukemia (HL-60/MX2)	Genipin	Doxorubicin, epirubicin	[120]
Acute promyelocytic leukemia (HL-60/MX2)	Glutathionylation (diamide)	Podofilox, all-trans-retinoic acid, mitomycin C, doxorubicin, quercetin	[114]
Breast cancer (MCF-7, T47D)	Genipin, mRNA silencing	Tamoxifen, cisplatin	[113]
Breast cancer (MCF-7, MCF7/LCC9)	miR-214	Tamoxifen, fulvestrant	[121]
Lung cancer (A549, H460)	mRNA silencing	Paclitaxel	[118]

Some of the observations with regard to the impact of UCP2 on cancer cells have yielded controversial results. Esteves et al. found that, contrary to what one would have expected, UCP2 overexpression‎ in melanoma, glioblastoma and pancreatic cancer cell lines resulted in cell cycle dysregulation and altered cell proliferation signaling [115]. These authors found that the changes were due to decreased HIF stabilization and activation of the AMPK signaling pathway, while they found no change in intracellular ROS levels and apoptosis rates. Furthermore, UCP2 overexpression‎ in these experiments had no effect on Δψm and mitochondrial O2 consumption, making it unlikely that the described changes were related to mitochondrial uncoupling [115]. Disparate observations on the role of UCP2 in cancer may stem from its controversial impact on HIF stabilization as well as from the pleiotropic effects of mitochondrial ROS, which are presumed to be a major biological target of UCPs.

Since ROS signaling is involved in both oncogenic and tumor suppressor pathways, modulation of ROS levels by UCPs may yield diverse outcomes in a given type of cancer cell depending on the balance between pro-survival and anti-survival effects [24], [116]. Some of these effects can be evoked by using low doses of chemical uncouplers that dissipate Δψm in a controllable manner. For instance, limited uncoupling in prostate and colon cancer cell lines by the weak acid protonophore FCCP (carbonyl cyanide 4-(trifluoromethoxy) phenylhydrazone) is able to induce markedly altered phosphorylation patterns within signaling pathways involving Erk p44/p42, Akt, AMPK, and mTOR [117]. Additional complexity emerged from the studies of Su et al., who found that low UCP2 expression‎ in lung cancer predicts a poor response to paclitaxel chemotherapy. In a series of experiments using lung cancer cell lines with high and low UCP2 expression‎, these authors concluded that UCP2 levels negatively correlate with the presence of p53 mutation, which is known to confer tolerance to higher ROS levels on cancer cells [118].

6. Conclusions

Cancer cells arise in metazoan organisms and gain competitive advantage over normal cells by evading physiologic restraints that regulate multicellular life. Metabolic reprogramming is a hallmark of this complex process, aimed at altering the hierarchy of biochemical pathways in cancer cells to meet their requirements of excessive growth and proliferation. Mitochondria represent a major obstacle to this agenda. These organelles are a major hub of energy metabolism in eukaryotic cells that have tightly controlled oxidative phosphorylation machinery balanced with molecular pathways of biosynthesis. Cancer cells must break this balance to accomplish their new priorities, which often lead to markedly increased mitochondrial ROSgeneration. UCPs are able to control this inherent byproduct of mitochondrial respiration, making them a potential accomplice of cancer cells. Although the precise mechanisms by which UCPs exert their biological effects remain to be fully understood, rapidly accumulating evidence indicates that these mitochondrial anion transport proteins may represent a unique molecular target in the treatment of cancer.

Disclosure

The author discloses no conflict of interest.

Transparency document

Download : Download Acrobat PDF file (1MB)

Transparency document.

References

[1]

A.F. Davila, P. ZamoranoMitochondria and the evolutionary roots of cancer

Phys. Biol., 10 (2013), p. 026008

CrossRefGoogle Scholar

[2]

D. Hanahan, R.A. WeinbergHallmarks of cancer: the next generation

Cell, 144 (2011), pp. 646-674

ArticleDownload PDFView Record in ScopusGoogle Scholar

[3]

N. Lane, W. MartinThe energetics of genome complexity

Nature, 467 (2010), pp. 929-934

CrossRefView Record in ScopusGoogle Scholar

[4]

M.D. Brand, L.F. Chien, E.K. Ainscow, D.F. Rolfe, R.K. PorterThe causes and functions of mitochondrial proton leak

Biochim. Biophys. Acta, 1187 (1994), pp. 132-139

ArticleDownload PDFView Record in ScopusGoogle Scholar

[5]

M. Muller, M. Mentel, J.J. van Hellemond, K. Henze, C. Woehle, S.B. Gould, R.Y.Yu, M. van der Giezen, A.G. Tielens, W.F. MartinBiochemistry and evolution of anaerobic energy metabolism in eukaryotes

Microbiol. Mol. Biol. Rev., 76 (2012), pp. 444-495

View Record in ScopusGoogle Scholar

[6]

C.M. Belcher, J.C. McElwainLimits for combustion in low O2 redefine paleoatmospheric predictions for the Mesozoic

Science, 321 (2008), pp. 1197-1200

CrossRefView Record in ScopusGoogle Scholar

[7]

M. Van Der Giezen, T.M. LentonThe rise of oxygen and complex life

J. Eukaryot. Microbiol., 59 (2012), pp. 111-113

View Record in ScopusGoogle Scholar

[8]

S. BerryThe chemical basis of membrane bioenergetics

J. Mol. Evol., 54 (2002), pp. 595-613

View Record in ScopusGoogle Scholar

[9]

S. Papa, V.P. SkulachevReactive oxygen species, mitochondria, apoptosis and aging

Mol. Cell. Biochem., 174 (1997), pp. 305-319

CrossRefView Record in ScopusGoogle Scholar

[10]

M. Inoue, E.F. Sato, M. Nishikawa, A.M. Park, Y. Kira, I. Imada, K. UtsumiMitochondrial generation of reactive oxygen species and its role in aerobic life

Curr. Med. Chem., 10 (2003), pp. 2495-2505

CrossRefView Record in ScopusGoogle Scholar

[11]

P. MitchellCoupling of phosphorylation to electron and hydrogen transfer by a chemi-osmotic type of mechanism

Naturwissenschaften, 191 (1961), pp. 144-148

CrossRefView Record in ScopusGoogle Scholar

[12]

D.G. Nicholls, S.J. FergusonBioenergetics: An introduction to the chemiosmotic theory

(2 ed.), Academic Press, Place Published (1992)

Google Scholar

[13]

A. Chevrollier, D. Loiseau, P. Reynier, G. StepienAdenine nucleotide translocase 2 is a key mitochondrial protein in cancer metabolism

Biochim. Biophys. Acta, 1807 (2011), pp. 562-567

ArticleDownload PDFView Record in ScopusGoogle Scholar

[14]

G. Solaini, G. Sgarbi, A. BaraccaOxidative phosphorylation in cancer cells

Biochim. Biophys. Acta, 1807 (2011), pp. 534-542

ArticleDownload PDFView Record in ScopusGoogle Scholar

[15]

M. Erecinska, D.F. WilsonRegulation of cellular energy metabolism

J. Membr. Biol., 70 (1982), pp. 1-14

CrossRefView Record in ScopusGoogle Scholar

[16]

G.C. BrownControl of respiration and ATP synthesis in mammalian mitochondria and cells

Biochem. J., 284 (Pt 1) (1992), pp. 1-13

CrossRefView Record in ScopusGoogle Scholar

[17]

I. FridovichSuperoxide anion radical (O2 −), superoxide dismutases, and related matters

J. Biol. Chem., 272 (1997), pp. 18515-18517

View Record in ScopusGoogle Scholar

[18]

A. Boveris, B. ChanceThe mitochondrial generation of hydrogen peroxide. General properties and effect of hyperbaric oxygen

Biochem. J., 134 (1973), pp. 707-716

CrossRefView Record in ScopusGoogle Scholar

[19]

L. Casteilla, M. Rigoulet, L. PenicaudMitochondrial ROS metabolism: modulation by uncoupling proteins

IUBMB Life, 52 (2001), pp. 181-188

View Record in ScopusGoogle Scholar

[20]

D. SpeijerHow the mitochondrion was shaped by radical differences in substrates: what carnitine shuttles and uncoupling tell us about mitochondrial evolution in response to ROS

BioEssays, 36 (2014), pp. 634-643

CrossRefView Record in ScopusGoogle Scholar

[21]

N.S. Chandel, G.R. Budinger, S.H. Choe, P.T. SchumackerCellular respiration during hypoxia. Role of cytochrome oxidase as the oxygen sensor in hepatocytes

J. Biol. Chem., 272 (1997), pp. 18808-18816

View Record in Sco