The emerging role of fumarate as an oncometabolite

1Cancer cells exhibit characteristic “hallmarks” of malignancy including increased proliferation, survival, and in particular dysregulated metabolism. There is abundant evidence showing that cancer cells produce energy through a high rate of glycolysis in the cytoplasm, in marked contrast to the process in most normal cells, which employ a relatively low rate of glycolysis followed by oxidation of pyruvate in the mitochondria. Although Otto Warburg postulated that this switch in cellular metabolism was the fundamental cause of cancer, most cancer research since has focused on mutations in, and roles of, oncogenes and tumor suppressors in the onset and progression of cancers. The development and application of highly sensitive new technologies such as mass spectrometry and nuclear magnetic resonance combined with metabolic labeling and profiling have increased our understanding of the complexities of normal and dysregulated cellular metabolism, particularly when linked with powerful computing programs that allow for the integration and interrogation of data. Furthermore, cancer associated mutations have been identified in genes of known metabolic function; namely isocitrate dehydrogenase 1 and 2 (IDH1 and 2), succinate dehydrogenase (SDH) and fumarate hydratase (FH; ). Consequently, there has been renewed interest in Warburg’s hypothesis and the link between dysregulated metabolism and cancer.

2The term oncometabolite has only recently been coined and assigned with confidence to R-2-hydroxyglutarate ((R)-2HG), the reduced form of 2-oxoglutarate (2OG). (R)-2HG is a by-product produced by gain-of-function mutations of IDH1 and IDH2, which normally catalyze the reversible NADP+-dependent oxidative-decarboxylation of isocitrate to produce 2OG in the cytoplasm and mitochondria, respectively. IDH mutations have been found in 75% of low grade gliomas and secondary glioblastoma multiforme and approximately 20% of acute myeloid leukemia. 2HG acts as a competitive inhibitor to multiple 2OG utilizing 2-oxygenases, including prolyl hydroxylases (PHDs), histone demethylases, and the TET family of 5-methylcytosine (5mC) hydroxylases. In gliomas, (R)-2HG accumulation caused by oncogenic IDH mutations enhances DNA methylation and epigenetic remodeling, which stalls cell differentiation and thereby primes cells for malignancy.

3How should we define an oncometabolite? Using (R)-2HG as an example, one could propose that an oncometabolite is a small molecule component (or enantiomer) of normal metabolism whose accumulation causes metabolic dysregulation and consequently primes cells allowing future progression to cancer. There are likely to be numerous and complex interacting steps in this process including inhibition, disruption or activation of pathways each of which will require detailed investigation. Nevertheless, the concept of oncometabolites is novel and exciting and offers a real and innovative route into therapies for a variety of cancers. Here we will discuss evidence implicating fumarate as an oncometabolite in FH-deficient cells. Furthermore, we will highlight where these studies have provided useful insights into cell metabolism.

4Germline loss-of-function mutations in the Krebs cycle enzyme FH predispose affected individuals to benign cutaneous and uterine leiomyomata, renal cysts and aggressive collecting duct and Type 2 papillary renal tumors in hereditary leiomyomatosis and renal cell cancer (HLRCC; ). However, the exact mechanisms leading to FH-associated oncogenesis remain to be elucidated.

5Fumarate hydratase is a highly conserved homotetrameric protein located and functioning in both the mitochondria and the cytosol. In the mitochondria, FH catalyses the hydration of fumarate to generate malate as part of the Krebs cycle. This pathway is not only essential for the production of cellular energy, but also forms a central metabolic hub to generate macromolecular precursors. In the cytosol, FH has been proposed to participate in a number of pathways where fumarate can be produced, including the urea cycle and the purine nucleotide cycle. Both forms of FH are encoded by the same transcript; localization of the protein is effected by cleavage of the resulting propeptide into two smaller peptides, one retaining the N-terminal mitochondrial targeting sequence (MTS) and one that is released into the cytoplasm. How FH is localized within the cell and the exact role the enzyme plays in different cellular compartments have not been elucidated fully and this will certainly be a focus for future research. Currently, it is unclear whether the mitochondrial Krebs cycle defect is responsible for oncogenesis, or if other mechanisms contribute, such as fumarate accumulation (Figure 1). To address this question we have used a conditional Fh1 (the ortholog of human FH) knockout mouse model and a panel of four mouse embryonic cell lines (MEFs) derived from this: wild-type MEFs, Fh1-deficient MEFs (Fh1-/-, Fh1KO), and Fh1-deficient MEFs in which there is stable re-expression of either full length, mitochondrial-targeted FH (Fh1-/- + FH), or cytoplasmic FH (Fh1-/- + FHΔMTS; ). These have been used to investigate the importance of FH in both the mitochondria and the cytoplasm and to unravel some of the complex consequences of FH loss for cellular, tissue and whole animal physiology with successful extrapolation into FH-deficient human tumors. Immunofluorescence studies with the MEFs described above have demonstrated that Fh1 loss results in a striking change in the morphology of mitochondria, which become much enlarged. This phenotype reinforces the observation that mitochondrial dysfunction is associated with FH deficiency; but the precise reasons for this and the consequences for the mitochondria and the cell remain to be determined. It could be postulated that disruption to the Krebs cycle leads to alterations in mitochondrial membrane potential and permeability of the outer membrane and increased autophagy; all aspects of cell biology and physiology that can, and should, be investigated.

6Fumarate hydratase-deficient cells and tumors have been shown to accumulate fumarate to very high levels with multiple consequences including the activation of oncogenic pathways. In Fh1 deficient MEFs the level of fumarate is approximately 8–10 fmol/cell as measured by 1H magnetic resonance spectroscopy metabolite analysis and no fumarate can be detected by this technique in either wild-type MEFs or Fh1-/- + FH MEFs. Perhaps surprisingly only very low levels (approximately 1 fmol/cell) can be detected in Fh1-deficient MEFs complemented with extramitochondrial FH (Fh1-/- + FHΔMTS), although the defect in aerobic metabolism is not corrected. Currently, we are undertaking metabolomic analyses to confirm these observations in MEFs by alternative techniques (capillary electrophoresis time-of-flight mass spectrometry;, ) and to extend the studies to mouse and human tissues lacking FH. It would be interesting to determine the relative levels of fumarate under a variety of physiological conditions in different cellular compartments; mitochondrial versus cytoplasm – especially since cytoplasmic “rescue” effects such a dramatic reduction in the overall cellular fumarate levels and in the nucleus, given the proposed role for FH in the DNA damage response in yeast.

7Others had postulated previously that FH-associated tumorigenesis might be driven by the upregulation of a number of oncogenic pathways by hypoxia inducible factor (HIF; ). Indeed, it has been shown that fumarate competitively inhibits 2OG-dependent oxygenases, particularly the HIF PHDs, thus mimicking hypoxia (pseudohypoxia), stabilizing the HIF complex and potentially activating its oncogenic target genes.

8Hypoxia inducible factor is stabilized in human tumors in HLRCC, in Fh1-deficient MEFs and in the hyperplastic renal cysts that develop in mice following targeted inactivation of Fh1. Gene expression analysis in all these tissues revealed strong signatures of HIF activation. Furthermore, both succinate and fumarate inhibit PHD enzymatic activities in vitro and cell-permeable esters of 2OG reactivate the enzymatic activity of the PHDs and alleviate the pseudohypoxia caused by succinate or fumarate accumulation. However, using a mouse model in which Fh1 inactivation in renal tubular cells was combined with inactivation of Hif-1α, Hif-2α, or both Hif-α isoforms; hyperplastic cyst formation was shown to be Hif independent (and separately Phd independent). Indeed combined inactivation of Fh1 and Hif-1α greatly exacerbated the cystic hyperplasia. While this suggests that the effect of HIF may be discounted in the early events of fumarate-mediated oncogenesis it neither precludes a role in tumorigenesis for long-term stabilization of HIF nor its consequent activation of multiple oncogenic pathways.

9This is by no means an end to the story as recent evidence has shown that fumarate (and succinate) inhibit the activity or function of other members of the 2OG oxygenase superfamily, including histone demethylase enzymes (HDMs) and TET proteins which are critical in epigenetic regulation of gene expression. Despite the identification of cancer-associated mutations in both classes of these enzymes, a direct causal role in oncogenesis is yet to be determined.

10In addition to its role as an allosteric regulator of 2OG-dependent oxygenases, fumarate is also an endogenous electrophile and reacts spontaneously with cysteine residues in proteins by a Michael addition reaction to form S-(2-succinyl) cysteine (2SC), a process termed succination. Accumulation of cellular fumarate has been shown to correlate directly with an increase in succinated proteins. It has been proposed that this results from mitochondrial stress in adipocytes during adipogenesis, when cultured in high glucose medium, in adipose tissue of obese type 2 diabetic mice and in skeletal muscle of streptozotocin-induced type 1 diabetic rats. Mechanistically, it has been proposed that nutrient excess from hyperglycemia results in high a NADH/NAD+ ratio, leading to feedback inhibition of oxidative phosphorylation and accumulation of mitochondrial intermediates including fumarate, which in turn causes protein succination. Targets for succination include the glycolytic enzyme glyceraldehyde-3-phosphate dehydrogenase, adiponectin, cytoskeletal proteins, and endoplasmic reticulum chaperone proteins. Furthermore, evidence suggests that succination of these proteins in cells may impair their functions.

11Compared to the situation in diabetes, protein succination is predictably more severe in FH-deficiency due to the significantly higher levels of fumarate accumulation. Immunohistochemical analysis of FH-deficient tumors and cysts has shown a striking direct relationship between FH inactivation and an increase in 2SC proteins, which is absent in non-HLRCC tumors and normal tissue controls and has provided a potentially robust diagnostic biomarker for FH-deficiency in cells and tissues. We hypothesize that succination resulting from FH deficiency targets multiple proteins and may, at least in part, account for the altered metabolism and oncogenic drive observed in HLRCC, as exemplified by the succination of Kelch-like ECH-associated protein 1 (KEAP1). Evidence for KEAP1 succination came from the observation that there is striking upregulation of the nuclear factor (erythroid-derived 2)-like 2 (NRF2)-mediated antioxidant signaling pathway in our murine Fh1 deficient renal cyst model, mouse embryonic fibroblasts as well as human FH-deficient cells and tissues. NRF2 controls the adaptive response of cells to oxidative and electrophilic stress, through the activation of target genes containing antioxidant response elements (AREs) while KEAP1 is the substrate recognition subunit of a Cul3-based E3 ubiquitin ligase complex and a major cellular electrophile sensor. In the absence of electrophiles, the homodimeric KEAP1 interacts with an NRF2 monomer, promoting its ubiquitylation and proteasomal-mediated degradation. KEAP1 has been shown to be succinated on two critical cysteine residues (Cys155 and Cys288) in FH-deficient cells, which disrupts its interaction with NRF2, resulting in stabilization and accumulation of nuclear NRF2. This allows binding to AREs and consequent activation of downstream target genes involved in defense against reactive oxygen species. The activation of the NRF2-mediated antioxidant pathway is a clear point of focus for future work; especially as activating NRF2 mutations and inactivating KEAP1 mutations are prevalent in many cancer types and oncogene-induced Nrf2 transcription promotes tumorigenesis in mice. NRF2 may contribute to tumor development by enabling FH-deficient cells to tolerate high levels of exogenous or endogenous oxidants, thus promoting their survival.

12Succination may result in the disrupted function of multiple proteins and offers a unique mechanism by which fumarate may lead to dysregulated cellular metabolism and act as an oncometabolite. Clearly screens need to be undertaken to identify other candidate succination targets which have cysteine residues critical for their function and are associated with oncogenic signaling or metabolic pathways.

13The Krebs cycle dysfunction caused by loss of FH activity poses significant challenges to cells in meeting energy requirements, in the generation of macromolecular precursors and in survival. Studies, in part contradictory, using a number of cellular models, have identified a variety of mechanisms by which FH-deficient cells may deal with these problems. Impaired respiration and upregulation of aerobic glycolysis have been observed in FH-deficient cell lines and tissues, presumably as an adaptation to meet cellular energy requirements by producing ATP independently of the TCA cycle. Elevated glutaminolysis has been observed and stable isotope labeling studies of an Fh1-deficient murine renal cell line have suggested that glutamine is the major carbon source for the Krebs cycle. These authors have also proposed upregulation of the heme biosynthesis pathway as a means of removing excess carbon from the dysregulated Krebs cycle whilst permitting partial mitochondrial NADH generation. Enhanced glycolysis and glutaminolysis are both stereotypical features of transformed cells and may prime FH-deficient cells toward malignancy. Separately, partial reversal of the Krebs cycle, so called glutamine-dependent reductive carboxylation, has been observed in human carcinoma lines including UOK262 cells, deficient in FH. By this mechanism 2OG is reductively carboxylated by IDH isoforms to generate isocitrate, followed by its subsequent metabolism to produce citrate, oxaloacetate and acetyl coenzyme A (AcCoA). AcCoA is crucial for fatty acid synthesis and protein acetylation while oxaloacetate is reduced to malate to compensate for decrease in the levels of these metabolites due to Krebs cycle blockage. Such a mechanism would allow cells with FH deficiency and impaired oxidative phosphorylation to maintain cell growth. There are some anomalies between these various proposed adaptive responses perhaps relating to the cellular model systems employed. More comparable analyses need to be conducted, ideally both in vitro and in vivo and in combination with metabolite and transcriptome profiling. Clearly, however, the adaptive response to fumarate accumulation of FH-deficient cells through alterations of primary metabolism may contribute to oncogenic transformation. Glycolysis, glutaminolysis, anaplerosis and the urea cycle may all be relevant for FH-deficiency and a greater understanding of these and their inter-relationships in normal and dysregulated cell metabolism are vital.

14Despite evidence for fumarate as an oncometabolite, in other circumstances this metabolite has been shown to exhibit cytoprotective roles. For example, it has been reported that in yeast cytoplasmic FH translocates to the nucleus following DNA damage. There it activates the damage response to double strand breaks, a process that can be complemented by high concentrations of fumarate in the absence of FH enzymatic activity. Elevated nuclear FH has also been detected in HeLa cells following irradiation damage suggesting that human cytosolic FH may have a similar function. Additionally, it has also been shown that elevated fumarate in Fh1 cardiac knockout mice greatly reduced the amount of heart tissue damage following ischemic-reperfusion injury. This is achieved by diverting amino acids into the Krebs cycle, thus maintaining ATP levels, stabilizing Nrf2 and consequently activating the Nrf2 antioxidant pathway. Perhaps the key to the opposing roles of fumarate as an oncometabolite or in a protective role lies in the exact cellular concentrations of fumarate and its cellular compartmentalization, as this metabolite on its own is essential for normal functioning of the Krebs cycle. Therefore, it would be of immense value to be able to determine the endogenous levels of fumarate in different cells and under different stress conditions. Technical difficulties to undertaking this include heterogeneity in tissue samples and, more significantly, the lack of effective methods to accurately quantify small molecule metabolites such as succinate and fumarate in sub-cellular compartments, e.g., mitochondria and nucleus, where local metabolite levels could be important.

15The shift in focus of cancer research to one of trying to understand how altered cellular metabolism and cancer are linked has highlighted how woefully ignorant we are about the complexities and interrelationships of cellular metabolic pathways and how these are altered under conditions of a variety of stress agents. However, studies into rare genetic disorders associated with metabolism are beginning to provide real insights into the adaptive responses of cells and dysregulated metabolism associated with cancer.

16The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.