Mitochondrial oxidative phosphorylation TRAP(1)ped in tumor cells

Abstract

Many tumors undergo a dramatic metabolic shift known as the Warburg effect in which glucose utilization is favored and oxidative phosphorylation is downregulated, even when oxygen availability is plentiful. However, the mechanistic basis for this switch has remained unclear. Recently several independent groups identified tumor necrosis factor receptor-associated protein 1 (TRAP1), a mitochondrial molecular chaperone of the heat shock protein 90 (Hsp90) family, as a key modulator of mitochondrial respiration. Although all reports agree that this activity of TRAP1 has important implications for neoplastic progression, data from the different groups only partially overlap, suggesting that TRAP1 may have complex and possibly contextual effects on tumorigenesis. In this review we analyze these recent findings and attempt to reconcile these observations.

Mitochondrial bioenergetics in tumor cells

Neoplasms profoundly reorganize their core metabolism to sustain growth in a dynamic environment where oxygen and nutrients are often limiting [1,2]. Glycolysis is elevated in many cancers and is uncoupled from oxygen availability (the Warburg effect) [3–6], favoring cancer cell growth under hypoxic conditions found in the interior of the tumor mass [3]. The enhanced use of glucose confers further advantages to neoplasms, providing essential intermediates for cell growth and proliferation by funneling metabolites into the pentose phosphate pathway (PPP) [7–9] and causes lactate efflux into the tumor microenvironment, which decreases extracellular pH and enhances the activity of several proinvasive factors [10,11]. Moreover, PPP induction contributes to the reinforcement of antioxidant defenses through the synthesis of NADPH, a key component of reactive oxygen species (ROS) scavenging systems [12], thus helping tumors face fluctuations in their redox equilibrium, which could be lethal [13].

Concomitant with upregulation of glycolysis, most tumor cells undergo a decrease in mitochondrial respiration, which could be secondary to upregulated glycolysis or brought about by tumor suppressor inactivation or oncogene activation [14–17]. Oxidative phosphorylation (OXPHOS) can be directly curtailed in tumor cells by increased glycolysis (the Crabtree effect) [18,19], via activation of the phosphoinositide 3-kinase (PI3K)/Akt pathway, activation of the transcription factors c-Myc [20] and hypoxia-inducible factor 1 (HIF1) [21,22], or inactivation of p53 [23]. In cancer cells, the interplay between deregulated signal transduction and changes in mitochondrial metabolism is complex and multifaceted. For example, c-Myc is a master anabolic inducer that triggers mitochondrial biogenesis, increases synthesis of acetyl-coenzyme A (CoA) (which contributes to histone acetylation and lipid biosynthesis), and stimulates catabolism of glutamine, which is a primary source of energy, nitrogen, and carbon to support the biosynthetic processes of neoplastic cells [20,24].

A comprehensive dissection of mitochondrial changes in tumor cells is beyond the scope of this review and several excellent papers can be found on this topic (e.g.,[9,10,25–30]. Here we propose the possibility that modulation of normal mitochondrial function can itself be a general activator of tumorigenesis. Recent observations suggest a key role for mitochondria in the tumorigenic process, mainly through genetic alterations in their bioenergetic machinery. Inactivating mutations of the tricarboxylic acid cycle (TCA) genes encoding succinate dehydrogenase (SDH) and fumarate hydratase (FH) are oncogenic [17,27] and contribute to neoplastic transformation by promoting accumulation of succinate and fumarate, respectively, which have been termed oncometabolites for their preneoplastic effects [31]. Loss of function mutations of SDH and FH occur in a panel of tumors including paraganglioma, pheochromocytoma, some forms of renal cell carcinoma (RCC), uterine fibroids, and skin leiomyomata, making SDH and FH unique examples of mitochondrial tumor suppressor genes. Both succinate and fumarate can lead to the stabilization of HIF1 [32,33], even under normal oxygen tension (pseudohypoxia). HIF1 stabilization contributes to neoplasia by promoting angiogenesis, epithelial–mesenchymal transition, and metabolic changes [34,35]. In addition, fumarate can bind reactive thiols of proteins in a process called protein succination, which was recently found to modulate the activity of enzymes involved in the redox regulation of tumor cells [17]. Recently it was discovered that gain-of-function mutations of mitochondrial isocitrate dehydrogenase (IDH) 1 and 2 cause accumulation of another oncometabolite, 2-hydroxyglutarate (2HG) in glioma, chondrosarcoma, cholangiocarcinoma, and acute myeloid leukemia. 2HG affects the activity of dioxygenases, mainly histone and DNA demethylases, causing profound epigenetic changes in cancer cells [36]. Furthermore, inactivating mutations in genes encoding respiratory complex subunits are associated with oncocytomas [37] and thyroid and prostate cancers [38,39].

Taken together, these observations indicate that genetically defined mutations in enzymes involved in mitochondrial energy metabolism disrupt OXPHOS and stimulate various processes associated with neoplastic transformation. However, these mutations are restricted to specific tumor types. Recently several papers have reported that the widely expressed mitochondrial chaperone TRAP1, which has been previously implicated as an antitumor molecular target, is an important modulator of the metabolic machinery of tumor cells, suggesting a more general role of TRAP1-mediated mitochondrial bioenergetic changes in tumorigenesis. A fuller understanding of the complex actions and regulation of TRAP1 in regulating mitochondrial metabolism could provide important information on the molecular mechanisms that mediate the metabolic adaptations of tumor cells.

TRAP1: the mitochondrial Hsp90

TRAP1, also referred to as Hsp75, is a molecular chaperone of the Hsp90 family. The human TRAP1 gene spans 60 kb and is located on chromosome 16p13. It comprises 18 exons, with 14 potential alternative transcripts and several nucleotide polymorphisms. At the protein level, it is predicted that differential splicing or amino acid changes can generate at least six major TRAP1 variants, but the biological meaning of this variability remains unclear. The main trap1 transcript encodes a protein of 704 amino acids comprising an N-terminal mitochondrion-targeting sequence, an ATPase domain and a C-terminal chaperone domain similar to that of cytosolic Hsp90s, which suggests functional similarity between TRAP1 and Hsp90. Hsp90 is a chaperone endowed with essential functions in priming client proteins for various biological processes, including protein–protein or protein–ligand interactions, subcellular trafficking, and control of protein maturation and stability [40,41].

Given the large number of Hsp90 clients (possibly 10% of the whole proteome), this chaperone plays a crucial role throughout the lifetime of cells. Hsp90 functions as a homodimer in conjunction with several cochaperones. A large body of experiments has established a model for its functional cycle, whereby conformational changes of the two protomers are regulated through rounds of ATP binding, hydrolysis, and release, although it is not fully understood how ATP hydrolysis is coupled to client maturation [40,41]. Less information has been collected on the biochemical activity of TRAP1, but the recent determination of the TRAP1 crystal structure [42] has greatly improved our understanding of its conformational cycle. ATP binding places TRAP1 in a high-energy, closed configuration with a peculiar asymmetry between the two protomers. An N-terminal ‘strap’ not found in Hsp90 further stabilizes this closed state [42]. Dissociation of ATP and chaperone reopening is predicted to be slower than ATP hydrolysis [43], whose energy is used during client remodeling in a two-step process: hydrolysis of the first ATP causes a change in protomer symmetry followed by rearrangement of the client-binding site, which is in turn coupled to structural changes in the client conformation; finally, the ADP-bound chaperone releases the client and eventually ADP [42]. Currently, no TRAP1 cochaperones have been identified, and until recently little was known of its client proteins. However, the finding that TRAP1 is primarily restricted to mitochondria [44], specifically in the matrix and associated with the inner membrane [45], has fundamental implications for its biological activity.

TRAP1: promoter or suppressor of neoplasia?

Although TRAP1 and its biological functions remain poorly investigated, recent results are beginning to place this chaperone at the center of mitochondrial physiology. TRAP1 has been implicated in critical mitochondrial pathways such as regulation of mitochondrial dynamics [46], mitophagy [47,48], protection from oxidative damage [49–51], and cell death [52]. Moreover, recent evidence suggests that TRAP1, and in particular its antioxidant activity, is involved in the pathogenesis of several disorders, including neurodegenerative diseases (Box 1) and cancer. It was observed that expression of TRAP1 is elevated in various tumor types with respect to surrounding nonmalignant tissues [52,53] (Table 1). In some neoplasms, such as prostate cancer or colorectal carcinoma, elevation of TRAP1 levels correlates with malignant progression and metastasis [53,54].

Box 1. Pathophysiology of TRAP1 in non-cancer settings.

TRAP1 transcripts are expressed at different levels in various tissues, including skeletal muscle, liver, heart, brain, kidney, pancreas, lung, and placenta [78]. More recent data, mainly obtained by immunohistochemistry, have identified the TRAP1 protein in the central nervous system, the gastrointestinal tract, the reproductive system, and many other tissues (see Table 1 and http://www.proteinatlas.org), but a detailed analysis of its expression pattern and levels in each tissue remains lacking.In addition to its role in cancer, TRAP1 may play a part in neurodegeneration. Evidence exists linking TRAP1 to Parkinson’s disease (PD), a disorder where poor quality control of mitochondria and unbalanced redox cycling play key roles [79]. Mutations in the E3 ubiquitin ligase Parkin or in the Ser/Thr kinase PINK1 cause inherited forms of PD, and both proteins participate in the maintenance of a healthy mitochondrial pool [80]. TRAP1 is a phosphorylation target of PINK1 [51] and has been found to rescue mitochondrial dysfunction in parallel with, or upstream of, Parkin in neuronal models where PINK1 is silenced [47,48]. TRAP1 also protects cells from oxidative toxicity caused by respiratory complex I inhibition via an α-Synuclein protein variant known to induce a genetic form of PD [81]. In further accord with its role as an antioxidant molecule, overexpression of TRAP1 protects rats from free radical generation, impairment of mitochondrial function, and brain infarction following cerebral ischemia [82]. The rise in ROS levels elicited by ischemic damage induces the permeability transition pore (PTP) a mitochondrial channel whose opening irreversibly commits cells to death [83]. Indeed, TRAP1 prevents cell damage in hypoxic cardiomyocytes by inhibiting PTP opening [84]. Furthermore, recessive mutations in the TRAP1 gene have been recently found in two families with congenital abnormalities of the kidney and urinary tract (CAKUT) and in three families where CAKUT is associated with congenital abnormalities in multiple organs [85], suggesting that TRAP1 mutations may be involved in the enigmatic etiology of these disorders.

Table 1.

TRAP1 expression levels in different tissue or cell types, either normal (upper panel) or transformed (lower panel)

Tissue/cell type	TRAP1 expression levelb	Refs
Central nervous system
Cerebral cortex, hippocampus, lateral ventricle, cerebellum	High (IHC)	The Human Protein Atlasa
Gastrointestinal tract
Oral mucosa, esophagus, stomach, duodenum, small intestine, rectum	High (IHC)	The Human Protein Atlas
Colon	Medium/High (IHC)	The Human Protein Atlas
Reproductive system
Testis, prostate, seminal vesicle	High (IHC)	The Human Protein Atlas
Breast, vagina, uterus, Fallopian tube, placenta	High (IHC)	The Human Protein Atlas
Uterus, cervix uteri, ovary	Medium/high (IHC)	The Human Protein Atlas
Prostate:
Epithelial cells: RWPE-1	Weak (WB)	[62]
Normal tissue	Negative (IHC)	[53]
Miscellaneous
Adrenal gland	High (IHC)	The Human Protein Atlas
Renal tubules	Positive (IHC)	[59]
Bladder	High (IHC)	The Human Protein Atlas
Tumor type/tumor cell line	TRAP1 expression level	Refs
Central nervous system
Glioma	Medium (IHC)	The Human Protein Atlas
Glioblastoma cell lines:
U-138 MG, U-251 MG, U-87 MG	Medium (IHC)	The Human Protein Atlas
LN229	Positive (WB)	[60,61]
Neuroblastoma cell line (SH-SY5Y)	Medium (IHC)	The Human Protein Atlas
Gastrointestinal tract
Colorectal carcinoma versus normal colon	Higher in 17/26 samples (WB, real-time PCR)	[55]
Colorectal carcinoma	Higher in lymph node metastasis (WB, IHC)	[54]
Colorectal carcinoma cell lines:
HCT116	Positive (WB)	[59,62]
HT-29	Increased in chemotherapy resistant cells (WB)	[55]
Reproductive system
Cervical carcinoma versus normal cervix Ovarian cancer (208 samples)	Lower (RNA); 55.3% high, 27.4% moderate, 12% low, 5.3% negative (IHC) correlates with favorable chemotherapy response	[59] [86]
Prostate:
K-Ras transformed epithelial cells: RWPE-2	Positive (WB)	[62]
Adenocarcinoma: PC3	Positive (WB)	[60,61]
High-grade intraepithelial neoplasia	High (IHC)	[53]
Gleason grade 3 and metastatic adenocarcinoma	High (IHC)	[53]
Gleason grade 4 or 5 and intraductal carcinoma	Heterogeneous (IHC)	[53]
Cell lines:
Ovarian adenocarcinoma: EFO-21	High (IHC)	The Human Protein Atlas
Endometrial adenocarcinoma: AN3-CA	Medium (IHC)	The Human Protein Atlas
Breast adenocarcinoma: MCF-7	Medium (IHC) Positive (WB)	The Human Protein Atlas, [63]
Breast adenocarcinoma: MDA-231	Positive (WB)	[61]
Cervical carcinoma: HeLa	Medium (IHC) Positive (WB)	The Human Protein Atlas, [59,62,63]
Miscellaneous
SAOS-2 osteosarcoma cell line	Positive; induced by oxidants (WB)	[50,62]
Raji Burkitt’s lymphoma cell line	Positive (WB)	[63]
Pheochromocytoma/paraganglioma cancer samples	Weak (RET/NF1/VHL mutations; WB) Positive (SDH mutations; WB)	[60]
Clear cell RCC (localized disease)	Weak (IHC)	[59]
Clear cell RCC (advanced disease)	Negative (IHC)	[59]
Normal bladder versus bladder cancer	Higher/equivalent (IHC)	[59]

^ahttp://www.proteinatlas.org.

^bAbbreviations: IHC, immunohistochemistry; WB, western blot analysis.

These observations have suggested that TRAP1 could have a role in favoring tumor progression. Accordingly, TRAP1 has been shown to protect cancer cells from treatment with various antineoplastic agents [50,55,56]. In addition, TRAP1 is a target of the transcription factor c-Myc [57], which is a key oncogene and driver of various neoplastic processes [58]. However, recent observations have also shown that TRAP1 expression inversely correlates with tumor stage in other cancers, such as clear cell RCC [59] (Table 1), suggesting that changes in TRAP1 expression during neoplastic progression may be dependent on the tumor type.

Therefore, to clarify what roles TRAP1 might play in tumorigenesis, it is essential to analyze its expression and biological functions using larger sample sizes and various cancer types and stages. Below we highlight the current knowledge regarding the role of TRAP1 in cancer.

TRAP1 is a bioenergetic master switch in tumor cells

It was recently proposed that TRAP1 maintains mitochondrial bioenergetic efficiency in tumor cells by inducing the electron transport chain (ETC), the TCA cycle, fatty acid oxidation, and amino acid synthesis [60], whereas TRAP1 inhibition triggers a signaling pathway involving AMPactivated protein kinase (AMPK) and ULK1 leading to mitophagy [61]. Furthermore, TRAP1 was found to stabilize respiratory complex II (also known as SDH) and knockdown of TRAP1 caused slight inhibition of SDH [60], suggesting that expression of TRAP1 in tumor cells may ensure a high level of cellular respiration even under nutrient starvation. In contrast to these observations, TRAP1 was found to inhibit SDH activity in various tumor and normal cell types [62]. Despite these different findings, both studies concluded that TRAP1 plays an important proneoplastic role. Differences in the experimental approach could account for the reported discrepancy between TRAP1 biochemical activities. Assays of SDH activity are extremely delicate, because the enzyme is a membrane-bound tetramer that requires a complex balance of cofactors to work properly; therefore, divergent results may reflect how the measurements were obtained. TRAP1-dependent slight activation of SDH was found in vitro on isolated enzyme [60], whereas SDH inhibition by TRAP1 was observed in situ (i.e., on permeabilized mitochondria [62]).

To affect the neoplastic process, changes in respiratory complex activity must result in a further outcome such as changes in redox equilibrium, oncometabolite levels, or respiration of the tumor cell. Independent observations revealed that TRAP1 expression dramatically decreases respiration, measured in vivo as the oxygen consumption rate (OCR) of cell monolayers, in diverse cell types. The specificity of this effect was further confirmed under multiple experimental conditions including RNAi with multiple TRAP1-directed short hairpin RNAs (shRNAs) and small interfering RNAs (siRNAs), re-expression of RNAi-resistant TRAP1 constructs, and comparisons between the OCRs of wild type fibroblasts and fibroblasts from TRAP1 knockout mice with or without reintroduced exogenous TRAP1 [59,62]. This TRAP1-dependent respiratory inhibition leads to multiple biochemical and biological consequences. For example, uncouplers cannot further increase the OCR [59,62], implying that cells lose their respiratory reserve capacity and must use biochemical routes alternative to OXPHOS for additional energetic requirements (e.g., supplying anabolic circuitries during proliferation). Accordingly, TRAP1 decreases oxygen-coupled ATP synthesis and accumulation of TCA cycle intermediates and shifts the burden of ATP production to glycolysis while inhibiting fatty acid oxidation [59,62]. Considered together, these bioenergetic features are in accord with a TRAP1-dependent switch towards a ‘Warburg’ phenotype. However, the mechanistic basis underlying this respiratory inhibition is not fully understood, because TRAP1 has been observed to inhibit either respiratory complex II/SDH [62] or respiratory complex IV [59]. Further investigations are clearly required to extend our knowledge of the interactions between TRAP1 and OXPHOS complexes.

A further layer of complexity is provided by a different set of experiments showing that, in parallel with slight TRAP1-dependent activation of complex II/SDH, silencing TRAP1 expression by RNAi only weakly affects respiration [60]. Moreover, TRAP1 inhibition by Gamitrinibs, a class of chimeric molecules in which the Hsp90 inhibitor 17-AAG is linked to a mitochondrion-targeting moiety [63], results in minimal changes in respiration, whereas at high concentrations Gamitrinibs induce abrupt mitochondrial uncoupling without any dose–response effect on the OCR [60]. Based on these data, it was proposed that TRAP1 maintains respiratory efficiency. However, sudden mitochondrial uncoupling could be due to nonspecific toxicity caused by Gamitrinib-mediated detachment of hexokinase II from mitochondria [61], which strongly induces permeability transition pore (PTP) opening and cell death in many tumor types [64–67].

In general, experiments conducted with Gamitrinibs must be considered with caution, because they are based on the assumption that standard Hsp90 inhibitors such as the geldanamycin derivative 17-AAG are excluded from mitochondria. However, whether 17-AAG can penetrate into mitochondria remains controversial [52,62]. Gamitrinibs might also alter mitochondrial homeostasis by disrupting Hsp90 function, either in the cytosol or in mitochondria [52], independently of any drug effect on TRAP1. For example, cytosolic Hsp90 is known to contribute to the mitochondrial targeting and import of various proteins [68–71]. Given these observations, it is highly desirable to confirm the data discussed above using selective TRAP1 inhibitors, which unfortunately remain to be identified.

TRAP1, respiration, and cancer: a matter of what and when

These observations indicate that the metabolic changes orchestrated by the activity of TRAP1 might profoundly influence neoplastic progression, even if the exact role played by TRAP1 remains controversial. Here we discuss mechanistic models that could account for the relationship between the bioenergetic functions of TRAP1 and the biological processes occurring in tumor cells.

According to one model, TRAP1 helps maintain mitochondrial protein folding and therefore promotes optimal mitochondrial function. TRAP1 might preserve bioenergetic activity and inhibit PTP opening, particularly under stress conditions encountered by the neoplastic cell, such as hypoxia and nutrient starvation (Figure 1A). This activity of TRAP1 might oppose the risk to mitochondrial proteins of redox stress-mediated inactivation as a result of oxidative damage. However, it must be emphasized that, in this context, the roles of TRAP1 and Hsp90 are indistinguishable, raising the possibility that the two proteins may have overlapping effects on mitochondrial function in this model. Moreover, an increase in OXPHOS activity strongly contributes to ROS generation, but increased OXPHOS is generally not observed during neoplastic transformation.

Figure 1.

Models of the roles played by tumor necrosis factor receptor-associated protein 1 (TRAP1) in neoplasia. (A) TRAP1 stimulates respiratory complex II [succinate dehydrogenase (SDH)], the tricarboxylic acid cycle (TCA) cycle, and mitochondrial metabolism in general, maintaining the correct folding of mitochondrial proteins even under stress conditions encountered by the tumor cell. This inhibition of mitochondrial damage favors neoplastic growth. (B) TRAP1 inhibits respiration and the ensuing reactive oxygen species (ROS) generation by downregulating the activity of both SDH and respiratory complex IV. As a consequence, succinate accumulates, stabilizing the transcription factor hypoxia-inducible factor 1 alpha (HIF1a) and inducing a pseudohypoxic phenotype. At the same time, the low levels of ROS shield the tumor cell from the risk of lethal oxidative insults. Taken together, these effects of TRAP1 activity may favor the early phases of neoplastic growth. (C) In later stages of tumorigenesis, the absence of TRAP1 prompts an increase in ROS levels, leading to increased cell motility and genomic instability and thus contributing to a more aggressive phenotype. Abbreviations: OXPHOS, oxidative phosphorylation; Q, coenzyme Q; cyt c, cytochrome c; OMM, outer mitochondrial membrane; IMM, inner mitochondrial membrane.

A second model proposes that metabolic modulation by TRAP1 directly promotes the neoplastic process. Inhibition of TRAP1 expression by RNAi abolishes tumor growth both in in vitro assays and in xenografts. Inhibition of SDH by TRAP1 leads to succinate accumulation, which is further enhanced during focus-forming experiments that mimic the accrual of neoplasm mass. In turn, succinate drives the stabilization of the HIF1α transcription factor independent of oxygen levels, thereby promoting glycolysis and tumor growth [62] (Figure 1B). Therefore, TRAP1 may induce a pseudohypoxic phenotype in tumor cells by increasing the level of the oncometabolite succinate, thereby sustaining the neoplastic process before hypoxic conditions are established by unrestrained malignant growth. Although xenograft models are revealing, these data must be confirmed in immunocompetent tumor models that more closely mimic natural neoplastic growth to determine the biological effect of TRAP1 on tumor growth. These experiments will also provide information on the role played by TRAP1 in the network of oncogenic signals. For instance, because TRAP1 expression is induced by Myc, it will be interesting to evaluate whether TRAP1 is a crucial oncogenic effector of Myc by following and modulating its expression in mouse models of Myc-driven tumorigenesis.

A third model suggests that TRAP1 acts as a tumor suppressor. In a selected panel of epithelial cancers (Table 1), TRAP1 expression correlates inversely with tumor grade and the absence of TRAP1 favors in vitro invasiveness [59]. TRAP1 interacts with and inhibits a mitochondrial fraction of the tyrosine kinase c-Src and the c-Src inhibitor dasatinib inhibits the increase in OCR and tumor cell invasion caused by knockout or knockdown of TRAP1. Therefore, the absence of TRAP1 deregulates mitochondrial c-Src, leading to enhanced activity of respiratory complex IV, increased ATP and ROS levels, and eventually c-Src- and ROS-dependent cell invasiveness (Figure 1C). Yet, the absence of observed spontaneous neoplastic events in TRAP1 knockout mice suggests that a tumor suppressor role of TRAP1 should be considered with caution.

These models suggest that the role of TRAP1 in tumorigenesis is likely to be important, if somewhat complex and perhaps contextual. Although contrasting results might easily be explained by differences in technical approach, the cell models used, or varying experimental conditions, it is plausible that the complex biological functions of TRAP1 could account for the observed differences. The inhibitory effects of geldanamycin derivatives on TRAP1 biochemical activities strongly suggest that TRAP1 regulates mitochondrial metabolism through its Hsp90-like chaperone activity, therefore regulating conformational changes among different activation states in client proteins. Hence, in analogy with Hsp90, TRAP1 might modulate the activity of numerous mitochondrial clients whose concerted regulation could finely tune mitochondrial bioenergetics. TRAP1 expression/activity might dynamically reflect varying metabolic cellular needs in the context of fluctuating environmental conditions. Thus, TRAP1 could affect the neoplastic process in specific tumor types or stages differently. For instance, clear cell renal cancers display decreased TRAP1 mRNA and protein expression during neoplastic progression [59], but a large majority of these malignancies are endowed with mutations in the VHL gene [72], which promotes HIF1 stabilization and glycolysis. If an important proneoplastic role of TRAP1 is to favor glycolysis over OXPHOS through the same HIF1 stabilization, selective pressure for TRAP1 activity would be lost during the progression of clear cell renal cancers.

TRAP1, oxidative stress, and tumorigenesis

A key to solving the riddle of TRAP1 in tumorigenesis may be its antioxidant function. Tumor cells are characterized by changes in redox equilibrium that favor increased steady-state accumulation of ROS [13]. Elevated ROS increases the risk of oxidative insults and of the ensuing lethal opening of the PTP; therefore, malignant cells must keep the PTP closed through deregulated mitochondrion-restricted kinase signaling [73]. TRAP1 could promote neoplastic growth by protecting tumor cells from ROS-dependent PTP opening. Indeed, TRAP1 was reported to interact with cyclophilin D, a mitochondrial chaperone known to activate the PTP, and prevent its ability to induce the PTP in several tumor cell models [52]. Notably, TRAP1 interactions with several respiratory complexes would keep the synthesis of ROS low and maintain the redox equilibrium of tumor cells. Although these regulatory events are poorly understood, it is reasonable to envision that they should be able to adapt quickly to changing environmental conditions. During the early phases of tumorigenesis, cells might require protection from excess ROS because adequate antioxidant defenses may not yet be fully established. Consequently, exposing cells to unbuffered oxidative stress could rapidly open the PTP and lead to cell death. Because neoplasms must adapt to conditions of oxygen paucity that occur during the growth of the primary tumor, cells might prioritize the establishment of a pseudohypoxic phenotype during early-stage tumorigenesis, producing higher levels of ROS in later stages of neoplastic growth. Shifting the redox equilibrium to produce increased amounts of ROS can be advantageous for advanced tumor cells, because free radicals damage nucleic acids and promote genetic instability [13], a crucial adaptive strategy for increasing cancer malignancy. Thus, reduced TRAP1 expression or activity could generate ROS and promote a more aggressive phenotype.

A further layer of regulation: post-translational modifications (PTMs) of TRAP1

Beyond its expression level, PTMs may modulate the ATPase-dependent chaperone activity of TRAP1, thereby differentially regulating TRAP1 activity in normal and cancer cells. Although there is no direct demonstration that PTMs serve any regulatory function for TRAP1, ample evidence for TRAP1 phosphorylation and acetylation exists in public phosphoproteomics and acetylome databases. In addition to TRAP1 phosphorylation by PINK1 (Box 1), TRAP1 has recently been found to interact with the mitochondrial subset of the kinase c-Src, blocking its stimulatory effects on OXPHOS. TRAP1 is tyrosine phosphorylated and this phosphorylation is inhibited by the Src inhibitor dasatinib [59].

Moreover, a rise in ROS levels favors cell migration and invasion [74] and TRAP1 depletion enhanced ROS-dependent cell migration and invasion in several tumor cell models in vitro [59]. Further, TRAP1 knockdown in neurons promotes downregulation of cell-adhesion molecules, consistent with a promotility/metastatic phenotype [75]. However, it should be mentioned that others have observed opposite effects on cell migration/invasion on compromising TRAP1 function [76,77], which is likely to reflect the altered metabolic environment found in diverse tumor types examined under distinct conditions.

Concluding remarks

This body of information highlights the importance of TRAP1 in tumor progression, despite the many questions that remain to be answered (Box 2). A more complete understanding of the role of TRAP1 in the process of tumorigenesis will require a careful mechanistic dissection of the complex links between the antiapoptotic effects of this chaperone, its role in modulating respiration and oxygen-coupled energy production, and its impact on cell motility. These activities are likely to be dynamic and reflect the changing energy needs of tumor cells during tumor initiation, growth, and metastasis, while also reflecting the degree of oxidative stress to which tumors are subjected during these processes. A more comprehensive understanding of the complex nature of TRAP1 function is also warranted before any therapies targeting this mitochondrial chaperone can be rationally advocated.

Box 2. Outstanding questions.

Moving forward, it will be important to address the following questions to gain a better understanding of TRAP1 function in cancer.

• In specific cancers, what is the relationship between changes in TRAP1 expression level and tumor stage and grade?

• What are the mechanisms of post-translational regulation of TRAP1 activity? What amino acids are modified?

• What are the various TRAP1 clients? Are there cochaperones that contribute to TRAP1 function?

• What are the bioenergetic functions of TRAP1; specifically:

  • How does TRAP1 affect respiration during neoplastic progression? Is this tumor-type dependent?
  • Does TRAP1 modulate the TCA cycle and lipid and amino acid metabolism at multiple points?

• What is the interplay among the biological functions of TRAP1 (metabolic changes and redox and oncometabolite regulation) and tumorigenesis?

• What role does TRAP1 play in in vivo tumor models established in immunocompetent hosts?