Knowledge-Based Therapeutics for Tricarboxylic Acid (TCA) Cycle-Deficient Cancers

Abstract

With the foundation pre-laid, research in the new millennium has readily excavated and expanded upon the architectural framework laid out by Otto Warburg's seminal work in a new wave of “westward expansion,” ever widening our understanding of cancer metabolism beyond the telescopic vision seen over a century ago. On this path, the unique circuitry of the cancer metabolic program has been elucidated, illuminating mutations of conserved cellular pathways implicated in tumorigenesis. Paramount among these are mutations in tricarboxylic acid cycle enzymes, succinate dehydrogenase, and fumarate hydratase, leading to deleterious accumulations in metabolic intermediates, “oncometabolites,” the pilots of the disease process. In this work, we seek to reflect on the advancements in the field in recent years, updating knowledge on the exact biochemical mechanisms at the helm of the tumor, providing rationale for clinical trials currently underway, and anticipating directions for the future on this expansive frontier.

Nearly a century since its inception, the idea that cancer cells can alter their functional phenotype has been ratified and lately has become a revived topic of interest in the field of metabolism. Through a lull since Otto Warburg's discovery at the beginning of the twentieth century, numerous mutations in metabolic enzymes have been cataloged in recent years to show that either loss- or gain-of-function of specific metabolic enzymes play a critical role in tumorigenesis (Warburg et al. 1927). It was recognized that up- or down-regulation of metabolic enzymes across the cell's functional network are implicated in tumors’ metabolic rewiring, yet none are so interwoven into the cellular regulatory system as those of the tricarboxylic acid (TCA) cycle. This amphibolic enzymatic process is the crossroads of the cell's energetic and biosynthetic capacity. The TCA cycle is involved in anabolism, heme and iron–sulfur complex biosynthesis, urea cycle, as well as in the metabolism of fatty, amino, and nucleic acids (Arnold and Finley 2023). Among the many enzymes in this pathway, three are paradigms to discussion: succinate dehydrogenase (SDH), fumarate hydratase (FH), and isocitrate dehydrogenase (IDH) (the latter is discussed in Gunn and Losman 2023). The genes coding these enzymes are critical to cell survival. Therefore, gain- or loss-of-function mutations distinctly impede their activity and are detrimental to cellular and organismal fate. If humans inherit maternal and paternal loss-of-function mutated alleles of these critical genes, they are prone to severely deleterious pathologies. Leigh syndrome, neonatal dilated cardiomyopathy, and neurodegeneration with ataxia and late-onset optic atrophy (NDAXOA) are examples of inborn errors of metabolism caused by biallelic germline mutations in human SDHA, one of the four subunits of the SDH enzyme. As opposed to biallelic genetic disorders, in cancer predisposition syndromes, patients are born with one mutant and one normal allele of a tumor suppressor gene and then suffer a second somatic “hit” (Knudsonian two-hit hypothesis), by which cells lose their heterozygosity and become susceptible to a multitude of cancers (Knudson 1971). Forming the basis of this work, SDH and FH will serve as the paradigm for our understanding of the consequences of mutations in metabolic tumor suppressor genes. The tumorigenic processes associated with SDH- and FH-related syndromes act in a multifactorial mode, culminating in a cancer phenotype, via (1) pseudohypoxia, the process by which hypoxia-response genes are transcribed irrespective of the presence of true hypoxia, (2) DNA and histone hypermethylation, and (3) reactive oxygen species (ROS). Mutations in SDH and FH result in deleterious amassments of succinate and fumarate, respectively. These accretions of TCA cycle intermediates are highly correlated with incidence of specific cancers, thus earning the moniker “oncometabolites” (Chowdhury et al. 2011). Of note, IDH-related cancers are not associated with hereditary syndromes and only one allele is somatically mutated, thus presenting with a gain-of-function phenotype (Parker and Metallo 2015).

Oncometabolites generally act as competitive inhibitors of dioxygenases, a family of enzymes that split molecular oxygen (O₂) into a hydroxyl group (-OH) on their substrates (Fig. 1) while the second oxygen atom oxidizes α-ketoglutarate (αKG) to succinate (McNeill et al. 2002; Nguyen and Durán 2016). Some dioxygenases are functionally attuned to molecular oxygen sensing and their inhibition acclimates the cell to pseudohypoxia, while other dioxygenases act as histone or DNA demethylases (Losman et al. 2020). Pseudohypoxia is mediated by the hypoxia-independent stabilization of the labile α subunits of hypoxia-inducible factors (HIFs) HIF1A and HIF2A. The levels of these proteins are tightly controlled posttranslationally through the hydroxylation of specific proline residues within their oxygen-dependent degradation domain via the oxygen sensor dioxygenases known as HIF-prolyl hydroxylases (PHDs) (Marxsen et al. 2004). HIF hydroxylation enables the binding of the E3 ubiquitin ligase von Hippel–Landau (VHL), which consequently targets HIF for degradation in an oxygen-dependent manner (Hayashi et al. 2019). Dysfunctions of this pathway can occur at many levels. One example is the classic VHL disease, caused by the germline loss of function of the VHL gene, leading to unadulterated HIF expression. In the case of SDH or FH loss of function, the inhibition of PHDs by succinate and fumarate, respectively, prevents VHL binding to HIF, and results in pseudohypoxia.

Figure 1.

Schematic representation of oncometabolite-driven gene-expression changes. Accumulations of succinate and fumarate, due to succinate dehydrogenase (SDH) or fumarate hydratase (FH) deficiencies, respectively, result in aberrant dioxygenases functioning, which causes two major changes in the cancer cell, which further feed into one another, propelling the cell to epithelial–mesenchymal transition. (1) Hypermethylation—both histone hypermethylation via inhibited KDMs and DNA hypermethylation via inhibited TETs. (2) Pseudohypoxia—via inhibited prolyl hydroxylases (PHDs), leading to constituently stabilized hypoxia-inducible factor (HIF) and subsequently induced hypoxia-response element (HRE)-regulated genes. (Figure generated with BioRender, https://biorender.com.)

At large, chromatin (structured DNA and histones) remodeling is achieved by chemical alterations, regulating the accessibility of the chromatin to the transcriptional machinery of the cell. These include DNA cytosine-methylation, and many histone modifications such as lysine and arginine-methylation, acetylation, ubiquitylation, ADP-ribosylation, and SUMOylation (Sulkowski et al. 2020). In a similar fashion to PHD inhibition, the burgeoning tides of oncometabolites also block demethylation pathways, resulting in hypercytosine methylation on DNA and hyperlysine methylation on histones (Nowicki and Gottlieb 2015). The equilibrium of DNA methylation is balanced by the ebb and flow of DNA methlytransferases (DNMTs) and DNA demethylases, 10-11-translocation methylcytosine dioxygenases (TETs). TETs catalyze the αKG-dependent hydroxylation reaction of 5-methylcytosine to 5-hydroxymethylcytosine (5-hmc), and further into unmethylated cytosine and methanol (Inoue et al. 2016). As αKG-dependent dioxygenases, TETs can be inhibited by oncometabolites, thus disturbing the balance of DNA methylation. Importantly, TET 1 and TET2 are known tumor suppressors, whose direct genetic loss of function leads to tumorigenic processes, particularly of mixed lineage leukemia (MLL) and acute myelogenous leukemia (AML), respectively (Huang et al. 2013; Tulstrup et al. 2021). Similar to DNA, histone methylation programs are orchestrated by the contrasting activities of histone methyltransferases and histone lysine and arginine demethylases. Of note, many lysine demethylases, especially those containing the Jumonji-like domain—JMJ, are αKG-dependent dioxygenases (Pollard et al. 2008; Johansson et al. 2014).

Multiple theories have also been put forward to answer how ROS may contribute to HIF-A stabilization under normoxic conditions, specifically in the context of SDH and FH mutations. It was suggested that ROS induces oxidation of iron cofactors of PHDs, while others suggest that ROS results in the up-regulation of HIF-A protein translation (Movafagh et al. 2015). Micro-RNAs (miRs) also take part in the hypoxia regulatory circuits. MiR-210, a HIF target, was shown to alter levels of iron-sulfur scaffold protein, ISCU, and subsequently induce ROS (Favaro et al. 2010). Furthermore, under acute hypoxia, HIF1 and HIF2 employ miR-210 to reduce the accumulation of HIF3, a dominant-negative regulator of HIF1 (Serocki et al. 2018). Similarly, miR-147 was shown in vitro to stabilize HIF1 via targeting of HIF3 (Wang et al. 2016). Taken together, the tumorigenic mechanism behind these cancers is the intersection between altered metabolism and aberrant gene expression.

While downstream mechanistic contribution of SDH or FH loss (and of IDH oncogenic mutations) to cancer converge on rather similar outcomes such as pseudohypoxia and hypermethylation phenotypes, there are apparent differences in the tissue of origin from which tumors with these different mutations arise. These differences can be the result of multiple factors. First, varied oncometabolites may have different effects on specific dioxygenases, and with that, on the transcriptomic and phenotypic outcome. Second, beyond their inhibitory effect on dioxygenases, the consequences of specific genetic alterations may have different metabolic adaptations and/or biochemical outcomes. Third, considering the housekeeping functions of the metabolic tumor suppressor genes, the ability of specific tissues to survive their genetic losses would dictate the ability of those tissues to feel and respond to the oncogenic effects of the accumulated oncometabolites.

THE BIOCHEMISTRY AND PATHOPHYSIOLOGY OF SUCCINATE DEHYDROGENASE

SDH, otherwise known as complex II of the electron transport chain (ETC), is an enzyme that serves as part of the oxidative phosphorylation machinery. It is a heterotetrameric protein complex that consists of two main components: (1) a catalytic core in the mitochondrial matrix, which is formed by two subunits, SDHA (c5p15.33) and SDHB (c1p36.13), as well as (2) a membrane component, formed by two additional subunits, SDHC (c1q23.3) and SDHD (c11q23.1), which anchor the enzyme complex to the inner mitochondrial membrane (Rutter et al. 2010; Bardella et al. 2011). The catalytic core of the enzyme aids in the organized transfer of electrons to ubiquinone. The membrane component works to stabilize the binding of ubiquinone to the enzyme via histidine ligands and a heme b moiety. Together with SDH assembly factors 1 and 2 (SDHAF1 [c19q13.12] and SDHAF2 [c11q12.2]), all of the above genes fall under the term SDHx (Dubard Gault et al. 2018).

Aberrations of each of the aforementioned enzyme subunits and their assembly factors are associated with specific cancer pathologies, namely, paraganglioma (PGL), pheochromocytoma (PHEO), gastrointestinal stromal tumors (GIST), renal cell carcinoma (RCC), and pituitary tumors (Pawlu et al. 2005; Fang et al. 2021; Loughrey et al. 2022). Of note, SDHB-associated tumors present with the most aggressive disease course and mortality (Amar et al. 2007). SDHC-related pathologies have been less extensively studied; however, a recent case series from across the United Kingdom details the mutation's relation to head and neck PGL, extra-adrenal PGL, and PHEO (Williams et al. 2022). SDHD is unique in that it is maternally imprinted, meaning that the maternal allele is silenced while the paternal allele is activated, which presents an interesting caveat to loss of heterozygosity in SDH-deficient cancers (Hensen et al. 2004; Bayley et al. 2014; Burnichon et al. 2017). Additionally, SDHD mutation is associated with more head and neck PGL than SDHB (Pawlu et al. 2005). Mutations of SDHA, unlike other SDHx genes, are rarely associated with hereditary PGL or PHEO (Burnichon et al. 2010). Rather, they are more often linked to GIST (Casey et al. 2017).

THE BIOCHEMISTRY AND PATHOPHYSIOLOGY OF FUMARATE HYDRATASE

The 10-exon-long FH gene is located at 1q42.1 in the human genome (Schmidt et al. 2020). It is formed by 510 amino acids folded into a structure with three functional domains: Central, amino-terminal lyase 1, and carboxyl-terminal fumarase C (Allegri et al. 2010). FH can be divided into two classes. Class I is dimeric and iron-dependent with an accompanying iron sulfur cluster, whereas class II is homotetrameric and iron-independent, mainly found in humans (Feliciano et al. 2016). Expression of this gene is differentially processed to form either a mitochondrial or cytosolic protein, distinguishable by pyroglutamic acid on the mitochondrial form and N-acetylalanine on the cytosolic form (Picaud et al. 2011). The mitochondrial isoform is of interest due to its utility in the TCA cycle and the cytosolic isoform is important due to its ability to metabolize fumarate from various metabolic processes like the urea cycle, and purine metabolism. Fumarate itself, a chemically reactive molecule, modifies cysteine residues on proteins and peptides like glutathione (Zheng et al. 2015). Germline mutations in the FH gene have been associated with hereditary leiomyomatosis and renal cell cancer (HLRCC), mainly type 2 renal papillary cancer (PRCC2), as well as cutaneous leiomyomas and uterine fibroids, which can be the presenting feature in women (Casey et al. 2020).

EMPLOYING KNOWLEDGE-BASED THERAPEUTICS

Many of the biochemical consequences of FH and SDH deficiency are alike. Therefore, similar therapeutic approaches related to general bioenergetic status or to the consequences of oncometabolite accumulation within these tumors should be considered. Nonetheless, there are major differences between the metabolic activities outside the mitochondria with FH and SDH. While SDH is relegated to the TCA cycle and ETC functions, FH has additional cytosolic and nuclear activities and, unlike succinate, fumarate itself is both substrate and product in many catalytic and noncatalytic, non-TCA cycle reactions (Yang et al. 2012a; Jiang et al. 2015; Eniafe and Jiang 2021; Solaimuthu et al. 2022). These unique traits may generate further diagnostic and therapeutic opportunities in the management of FH-deficient cancers.

Unfortunately, to date, no clinically specific targeted therapies have been approved for use in FH- and SDH-deficient cancers; however, the next section will recount where the current clinical entry points stand. To elaborate on these opportunities, the following will focus on targeted strategies toward (1) metabolic remodeling unique to these cancers, (2) oncometabolite-induced rewiring of the transcriptome, and (3) the biochemical consequences of nonenzymatic reactions due to increased fumarate in the cell.

METABOLIC REMODELING

Cancers are able to adapt their metabolism to feed the ever-growing tumor. In this light, TCA cycle-deficient cells specifically modulate the activities of energetic processes, particularly glycolytic flux, for the purpose of supporting energetic demands, enhanced cellular growth, and proliferation (Yang et al. 2012b). An effective way to target fermentative glycolysis and cytosolic NAD⁺ recycling (required for effective glycolysis within the tumor) is to inhibit pyruvate reduction to lactate via lactate dehydrogenase (LDH) (Vazquez et al. 2016). Indeed, when LDH activity was genetically silenced with shRNA in xenografted FH-deficient tumors made from human cancer cells, a significant decrease in proliferation rate and tumor size was recorded (Xie et al. 2009). Nevertheless, despite many approaches and successful preclinical reports, LDH inhibitors did not yet progress to clinical use (Sharma et al. 2022). In addition to the increased glycolysis, TCA cycle-deficient cancers display other metabolic alterations. For example, decreased activity of the energy-sensing enzyme AMP-dependent protein kinase (AMPK) and subsequent decreased phosphorylation of acetyl CoA carboxylase, leads to increased fatty acid synthesis (Linehan and Rouault 2013). Whether such metabolic alterations create therapeutic opportunities is yet to be evaluated preclinically. Fatty acid synthesis inhibitors have been tested clinically (NCT03179904), but not in the context of SDH or FH deficiency.

Glutamine metabolism is a hallmark of cancer metabolism, by way of its sheer abundance and centrality in the donation of carbon and nitrogen to biosynthetic pathways critical to proliferation of the tumor (for example, amino, fatty, and nucleic acid synthesis), leading some cancer cells to become “glutamine addicted” (Masisi et al. 2020). Additionally, glutamine is implicated in ROS management via glutathione synthesis. Glutaminase plays an integral role in the production of glutamate from glutamine, which can further be modified to αKG to fuel the TCA cycle (Chen and Cui 2015). With all of these implicated pathways, therapeutic approaches under development target glutaminase in multiple cancers, including SDHB-associated PGLs and PHEOs (Table 1). In an in vitro PHEO model, a significantly increased GLS-1 expression in cells depleted of SDH function was reported (Sarkadi et al. 2020). Furthermore the selective GLS-1 inhibitor, bis-2-(5-phenylacetamido-1,3,4-thiadiazol-2-yl)ethyl sulfide (BPTES) decreased the number of proliferating, SDH-abrogated cells. In support of this in vitro study, SDHB-related cancers display up-regulated glutamine metabolism (Lussey-Lepoutre et al. 2015). In SDH-deficient tumors, glutaminolysis can further support succinate production, and its oncometabolic functions (Moreno et al. 2020). To this end, a small molecule, allosteric and selective glutaminase inhibitor, telaglenastat (formerly known as CB-839), was investigated in SDH-deficient tumors, among other solid tumors, in a clinical trial (NCT02071862). Telaglenastat was well tolerated and showed signs of anticancer activity in a limited cohort of SDH-deficient tumors; however, the current cohort was too small to draw significant conclusions (Harding et al. 2021). Recent literature also points to the added tumor-reducing benefit of combining GLS-1 inhibition with disruption of either (1) mTOR with everolimus, or (2) VEGF/MET/AXL with cabozantinib (Yang et al. 2019). It was suggested that such combination therapy reduced overall consumption of glucose and glutamine, resulting in antiproliferative effects, culminating in decreased production of ATP and cellular building blocks (Emberley et al. 2021). CANTATA (NCT03428217), a randomized, clinical, placebo-controlled, double-blind trial studied the synergy between telaglenastat and cabozantinib in RCC (Tannir et al. 2022). While preclinical studies showed promising results, the aforementioned combination failed to promote the efficacy of both agents in the context of RCC regression in general, showing equal progression-free survival in telaglenastat + carbozantinib versus placebo + carbozantinib. However, no separate analysis for SDH-deficient RCC was performed and numbers of SDH-deficient RCC were low (Tannir et al. 2022). ENTRATA (NCT03163667) studied the concomitant effect of everolimus and telaglenastat in RCC (Lee et al. 2022). The combination was well tolerated and improved progression-free survival. Nevertheless, no GLS-1 therapies (monotherapy or combined therapy) are known to benefit TCA cycle-deficient RCC to date.

Table 1.

Clinical trials for tricarboxylic acid (TCA) cycle-deficient cancers mentioned in the text

Disease	Trial ID	Title	Summary	Dates	Progress
FH-def (general) SDH-def (general) SDH-def GIST	NCT02071862	Study of the glutaminase inhibitor CB-839 in solid tumors	Exploiting tumors’ dependence on glutamine, CB-839 used to inhibit tumors’ glutaminase. Among other cancers, FH-deficient tumors, SDH-deficient GIST, and SDH-deficient non-GIST cancers were treated.	Feb 26, 2014–Jul 20, 2022	Completed
FH-def RCC SDH-def RCC	NCT04068831	Talazoparib and avelumab in participants with metastatic renal cell carcinoma	Patients received talazoparib (PARPi) and avelumab in context of either VHL-, SDH-, or FH-deficient RCC.	Aug 28, 2019–ongoing	Active, not recruiting
FH-def RCC	NCT04387500	Sintilimab injection combined with inlyta in fumarate hydratase- deficient renal cell carcinoma	Examining ORR and PFS (over 3 yr each) in response to sintilimab (PD-1i) injection and treatment with inlyta (tyrosine kinase inhibitor) in FH-deficient RCC.	May 14, 2020– ongoing	Active, recruiting
Nonspecific	NCT03428217	CANTATA: CB-839 with cabozantinib vs. cabozantinib with placebo in patients with metastatic renal cell carcinoma (CANTATA)	Nongenotyped RCC treatment with cabozantinib and either CB-839 or placebo to see differences between PFS in 18 mo, and OS in 36 mo.	Feb 9, 2018– Aug 23, 2021	Completed
Nonspecific	NCT03163667	CB-839 with everolimus vs. placebo with everolimus in participants with renal cell carcinoma (RCC) (ENTRATA)	Nongenotyped ccRCC treatment with everolimus and either CB-839 or placebo to see differences between PFS in 11.2 mo, and OS in 30.4 mo.	May 23, 2017–Sep 15, 2022	Completed
SDH-def GIST	NCT03556384	Temozolomide (TMZ) in advanced succinate dehydrogenase (SDH)-mutant/deficient gastrointestinal stromal tumor (GIST)	Examine the efficacy of TMZ in advanced SDH-deficient GIST via measures of ORR in 6 mo and PFS in 4 yr, OS in 4 yr, and detailed summary of adverse events over 6 mo.	June 14, 2018–ongoing	Recruiting
Nonspecific	NCT04394858	Testing the addition of an anticancer drug, olaparib, to the usual chemotherapy (temozolomide) for advanced neuroendocrine cancer	Nongenotyped neuroendocrine tumors (with future intention to segregate SDH deficiency) treated with temozolomide (tyrosine kinase inhibitor) plus olaparib (PARP-i) to observe PFS, OS, OR, and incidence of adverse events over 5 yr.	May 20, 2020–ongoing	Recruiting

Urea cycle is also dysregulated in FH-deficient cancers. Normally, arginosuccinate lyase (ASL) functions to create fumarate and arginine from arginosuccinate. However, in the context of high fumarate levels, the reaction is reversed by Le-Chatelierian equilibrium ­principles, resulting in a flux toward the precursor arginosuccinate (Vazquez et al. 2016). The arginine shunted into arginosuccinate production is taken from exogenous sources, while the fumarate is abundant due to the nature of the FH deficiency. This mechanism aids to sequester fumarate in the tumor cell in arginosuccinate and prevent it from reaching toxic levels, allowing permissive tumor growth. This delineates exogenous arginine as a limiting factor in the reaction. Therefore, it is targetable by way of either circulating arginine depletion or its import blockade for the obligate-arginine-auxotrophic tumor. To this end, pegylated arginine deiminase (ADI-PEG 20) was used in vitro to deplete extracellular arginine and the subsequent formation of arginosuccinate, resulting in decreased cellular proliferation of FH-deficient cancer cells (Zheng et al. 2013). Recently, ADI-PEG 20, when combined with spermidine analog GC7, N1-guanyl-1,7-diaminoheptane, was shown to prevent resistance to ADI-PEG 20 treatment through GC7-based TCA cycle inhibition (Carpentier et al. 2023). Furthering this to patients, there was a phase II/III clinical trial (NCT02709512) that adds ADI-PEG 20 to SOC (therapy including combination of permetrexed and cisplatin) treatment to patients with nonepithelioid malignant pleural mesothelioma portends to starve tumor cells of arginine, leading to cell death (Szlosarek et al. 2023). Hence, it is of interest to explore this therapy in the context of TCA cycle-deficient tumors, as these tumors may have defunct arginine production and, therefore, clinical use, providing a tolerable side effect profile in this specific patient population.

A potentially effective cancer therapeutic strategy identifies genes that become essential following a loss of a specific tumor suppressor, an approach known as “synthetic lethality.” Considering the major metabolic consequences of FH loss and fumarate level induction, several synthetic lethal approaches were theorized for targeted therapy. One of the major metabolic consequences of FH loss is increased redox stress, exacerbated by glutathione succination (generated by a chemical reaction between fumarate and reduced thiol [-SH]) and managed by Nrf2 induction (Fig. 2B; Sullivan et al. 2013; Zheng et al. 2015). Additionally, SDHB-silenced tumors showed similar reliance on Nrf2-led expression profile to combat increased redox stress (Liu et al. 2020). This state allows for the therapeutic exploitation via synthetic lethality of several redox mediators. It was demonstrated that heme oxygenase (HO-1, enzyme and HMOX1, gene) is synthetically lethal with FH (www.ncbi.nlm.nih.gov/gene/3162#bibliography). HMOX1 is itself an Nrf2 target gene and, as such, an important antioxidant. Hence, selective killing of FH-deficient cancer cells can be achieved by specific inhibition of HO-1 (Frezza et al. 2011; Podkalicka et al. 2020). Another important Nrf2 target is the cystine transporter and ferroptosis mitigator, SLC7A11 (Koppula et al. 2021). Ferroptosis is a unique type of nonapoptotic, programmed cell death that is iron-dependent. Cystine, comprised of two oxidized cysteine molecules, is a plasma-circulating molecule that is uptaken by Nrf2-activated cells and subsequently reduced intracellularly to its constituent molecules, supporting glutathione biosynthesis (Vazquez et al. 2016). Indeed, it was demonstrated that FH-deficient cells in which glutathione is succinated and Nrf2 is induced, are dependent on SLC7A11 (Kerins et al. 2018). Despite the potential sensitization of FH-deficient cells to redox induction and ferroptosis-inducing agents, to our knowledge, there are no ongoing clinical studies with HO-1 or SLC7A11 inhibitors.

Figure 2.

Schematic representation of molecular and cellular level changes caused by succination. (A) Chemical structure of succination: fumarate and cysteine combining to form S-(2-succinic)cysteine (2-SC). (B) Model of cellular pathways impacted by succination: Succination of glutathione and KEAP1 balances intracellular reactive oxygen species (ROS) levels through Nrf2 induction. GPX4 succination prevents efficient management of lipid peroxidation, sensitizing cells to ferroptosis. Dotted lines correspond to reactions that are inhibited by succination. (Figure generated with BioRender, https://biorender.com.)

Importantly, cells with a TCA cycle-deficient phenotype partially compensate for their telescoped TCA cycle through an altered production of the nonessential amino acid, aspartate, via pyruvate carboxylase that produces oxaloacetate, the precursor for aspartate, directly from pyruvate (Mullen et al. 2014; Cardaci et al. 2015). A small molecule tool compound, ZY-444, is a pyruvate carboxylate inhibitor, which demonstrated effective tumor growth inhibition in a preclinical model for breast cancer (Lin et al. 2020). By extension, this or similar pyruvate carboxylate inhibitors may have use in the targeting of TCA cycle-deficient tumors.

Cells under perpetual redox stress depend on NADPH production for the maintenance of reduced glutathione and for surviving redox stress. The oxidative pentose phosphate pathway (oxPPP) is a major player in the reduction of NADP⁺ to NADPH. In line with that, FH-deficient cells demonstrate dependency on oxPPP (Adam et al. 2011; Yang et al. 2013). Targeting glucose-6-phosphatase dehydrogenase (G6PD), the first and rate-limiting enzyme of the oxPPP, has been explored as a potential therapeutic opportunity. Until recently, DHEA, an endogenous steroid, was thought to act as an inhibitor of G6PD; however, this is not the case (Ghergurovich et al. 2020). To the best of our knowledge, no further inhibitors of G6PD have been tested in the clinic, especially not within the context of TCA cycle-deficient cancers.

THERAPEUTIC OPPORTUNITIES DUE TO ONCOMETABOLITES

Accretions of oncometabolites elaborate unique phenotypes in SDH- and FH-deficient cancers, hinging on chromatin modification and genetic reprogramming. Targeted therapies and clinical trials, informed by twenty-first century scientific advances, have been elaborated to improve outcomes in these TCA cycle-deficient cancers. At first, tyrosine-kinase inhibitors (TKIs) had been highlighted as a therapeutic option for SDH-deficient GIST (Nannini et al. 2021). This is thought to be due to the lack of SDH, driving a hypoxic response via HIF-dependent transcription and consequent angiogenic cascade mediated by the vascular endothelial growth factor (VEGF) pathway. This hypothesis was initially confirmed in cells treated with sunitinib, a tyrosine kinase inhibitor with antiangiogenic properties via VEGFR inhibition (Cassol et al. 2014; Boikos et al. 2016). Within GIST, angiogenic genes were found to be up-regulated, suggesting that targeted therapies to VEGF and its receptors may have clinical utility (Favier et al. 2012). Of note, the first-line therapy for GIST overall, imatinib, serving as a c-kit inhibitor, is relatively ineffective in treating wild-type c-kit or PDGFR tumors (WT-GIST), half of which are SDH-deficient (Belinsky et al. 2017; Nannini et al. 2021). Therefore, specific therapies for TCA cycle-deficient cancers were tested based upon the up-regulation of angiogenic genes. Time to tumor progression was lengthened and multiple instances of disease stabilization were achieved upon treatment with sunitinib versus imatinib (Janeway et al. 2009). An observational study supported these findings by showing that four patients who experienced disease progression while under monotherapy with imatinib were stabilized with sunitinib and achieved disease control of their GIST (Liu et al. 2017). In agreement with the oncometabolic role of succinate, transcriptome analysis highlighted that the HIF2a transcriptional signature was up-regulated in SDH-deficient PHEO and PGL (Favier et al. 2012). This is clinically significant, as targeted therapies to nullify the above transcription factor's response to the pseudohypoxic tumor microenvironment may prove useful. Belzutifan, an FDA-approved drug that impedes the interaction between HIF2a and its cotranscription factor ARNT/HIF1B, decreases tumor burden in VHL-deficient renal cancer (Courtney et al. 2018; Moog et al. 2020; Jonasch et al. 2021). Considering the pseudohypoxic phenotype, it is important to research the effects of this drug on TCA cycle-deficient tumors. Currently underway (NCT04924075) is a phase II study testing belzuitfan in advanced PGL, advanced GIST, and solid tumors, all with pseudohypoxia-related genetic alterations.

The importance of TET inhibition by oncometabolites has been typified in TCA cycle-deficient tumors in our understanding of the hypermethylation phenotype. Succinate- or fumarate-mediated HIF2a activation and TET2 inhibition generates a positive feedback loop between pseudohypoxia and hypermethylation (Fig. 1), in which HIF2a stimulates methylation processes while TET2 inhibition by the oncometabolites maintain an hypermethylation state that enables epithelial-to-mesenchymal transition (EMT)-mediated by HIF (Sciacovelli et al. 2016; Morin et al. 2020). O⁶-methylguanine-DNA methyltransferase (MGMT) functions to repair DNA via the removal of alkyl groups from nucleotides, usually guanine. Without this important function, cells are vulnerable to alkylating agents. Extensive hypermethylation, particularly of the MGMT promoter region in glioblastoma, has been associated with increased sensitivity to temozolomide, an alkylating chemotherapeutic (Hegi et al. 2005; Ravegnini and Ricci 2019; Giger et al. 2022). There are ongoing studies to identify induced vulnerabilities associated with hypermethylation phenotype mediated by TET inhibition (Morin et al. 2020). However, temozolomide sensitivity in SDHx-deficient PHEO and PARA was assessed in a retrospective study detailing that temozolomide had limited efficacy (Perez et al. 2022). Interestingly, it has been reported that in SDH-deficient GIST, similarly to glioblastoma, MGMT hypermethylation of the promoter's CpG island is associated with lower expression of MGMT (Hadoux et al. 2014). Such a hypermethylation phenotype is likely associated with TET and JmJ domain-containing histone demethylase inhibition by succinate (Letouzé et al. 2013). Therefore, NCT03556384, a phase 2 clinical study, assesses the efficacy of temozolomide on advanced SDH-deficient GIST with fulfillment of primary objectives expected in 2023.

Syrosingopine, an inhibitor of the monocarboxylate transporters 1 (MCT1) and MCT4, was used in conjunction with metformin, a mild oxidative phosphorylation inhibitor, to induce synthetic lethality via accumulation of lactate intracellularly (Benjamin et al. 2018). A similar approach has been taken to clinical trial via the MCT1 inhibitor AZD-3965 (NCT01791595) to induce intracellular lactate, thereby decreasing intracellular pH and disrupting the feedback networks of glycolysis (McNeillis et al. 2020). While these drugs have not yet been studied in the context of TCA cycle-deficient cancers, their mechanistic approach may be exploited.

An additional reported role for oncometabolites is their direct role in inducing DNA damage. Oncometabolites can inhibit homology-directed repair (HDR) pathways via the inhibition of KDM4B, a lysine demethylase, leading to increased hypermethylation of histone 3 lysine 9 (H3K9) (Sulkowski et al. 2018). Subsequently, the augmented hypermethylation of H3K9 adjacent to DNA damage sites disguises the signal needed for synchronous HDR function (Sulkowski et al. 2020). It has been suggested that Tip-60 is activated at double-strand break (DSB) sites where it complexes with ATM to repair DNA damage, and increased levels of oncometabolites make ATM less active (Sun et al. 2005). It was theorized that oncometabolites increase the methylation status of H3K9m3e3 so that the spike needed at DSBs is not detected above the new high threshold (Sulkowski et al. 2020). With disordered HDR mechanisms, cells become increasingly sensitive to PARP inhibitors such as olaparib and BMN-673 (Sulkowski et al. 2018). To this end, PARP inhibitors are currently being used in the clinic for PHEO and PGL in TCA cycle-deficient tumors (NCT04394858).

In recent years, our understanding of miRs has guided new and innovative therapies. For example, targeted reduction of the oncogenic miR-155 by the oligonucleotide anti-miR-155, cobomarsen, was shown to reduce tumor burden in cutaneous T-cell lymphoma and was therefore put to the test in subsequent clinical trials (Bartolucci et al. 2022). PHEO and PGLs associated with germline SDHx mutations were found to overexpress the HIF target, miR-210, both in vitro and in patient samples (Tsang et al. 2014). This is useful because it identifies miR-210 as a marker for pseudohypoxia within these TCA cycle-deficient cancers. MiR-210 functions by controlling certain vascular remodeling genes such as ephrin A3 (a ligand for ephrin receptor, associated with angiogenesis) (Hu et al. 2010). The sheer abundance of miR-210 may serve as reason enough to explore its essentiality in TCA cycle-deficient cells and to develop anti-miR-210 therapies for these tumors.

Many RNA epitranscriptome modulators (RNA demethylases) may also be affected by increased levels of oncometabolites (Mauer et al. 2019). These modulators act on RNA molecules and manage the dynamics of translation. For example, oncometabolites resulting from FH mutation inhibit FTO (fat mass and obesity-associated protein), the RNA demethylase acting on N⁶-methyladenosine, whereas other RNA demethylases are spared, including those acting on N¹-methyladenosine and 5-formylcytosine (researchfestival.nih.gov/2023/posters/rewiring-rna-methylation-oncometabolite-fumarate-renal-cell-carcinoma). While FTO has been implicated in cancer, the mechanistic understanding underlying potential therapeutic opportunities is yet to be fully elucidated.

NONENZYMATIC ONCOMETABOLIC ROLE OF FUMARATE

FH-deficient cells accumulate fumarate and succinate and may present with dysfunction of complexes I and II of the ETC (Tyrakis et al. 2017). On the one hand, high levels of intracellular fumarate cause FH-deficient tumors to operate similarly to SDH-deficient tumors via pseudohypoxia and hypermethylation (Navarro et al. 2022). However, an important complexity to note is that fumarate is an inherently more reactive molecule than succinate.

Fumarate has been found to disrupt normal cellular function via the nonenzymatic covalent modification of cellular proteins (Fig. 2A) to form S-(2-succinyl)cysteine (aka 2-succinocysteine [2SC]) in a posttranslational process aptly named “succination” (Alderson et al. 2006). These processes induce perturbations of multiple reactions that generate potential biomarkers for FH-deficient tumors. Of note, a recent study discovered unique fumarate-altered metabolites in the circulation of patients harboring FH-deficient RCC (Zheng et al. 2023). These metabolites, succinyl-adenosine and succinic-cysteine (in a free amino acid format, not as a posttranslational protein decoration), enable disease detection at early respectable stages, and, therefore, effective monitoring of liquid biopsies may become a life-saving clinical practice. Immunohistochemical staining of 2SC-decorated proteins is a useful diagnostic technique in the pathological confirmation of FH-deficient cancers (Trpkov et al. 2016). However, unlike periodic monitoring of plasma metabolites, the visualization diagnostic approach depends on attaining biopsies and cannot be used for early detection or to monitor disease progression. Nevertheless, nonenzymatic protein succination can play a mechanistic role in the tumorigenic process and may expose vulnerabilities of FH-deficient tumors. For instance, the inhibition via succination of the E3 ubiquitin ligase kelch-like ECH-associated protein 1 (KEAP1) prevents the degradation of Nrf2 and then acts via a “feedforward” mechanism to reinforce an antioxidant environment ideal for tumor growth (Adam et al. 2011; Sullivan et al. 2013; Sourbier et al. 2019; Rogerson et al. 2023). From an immune standpoint, Nrf2 has been shown to inhibit interferon β and quells the innate immune response, setting an anti-inflammatory scene for the tumor to grow (Ryan et al. 2022). Nrf2-inhibitors have been elaborated to induce a proapoptotic response (Pouremamali et al. 2022). With this knowledge, investigators have sought to systematically attack FH-deficient tumors with ROS-based therapies. Bortezamib, a potent ROS inducer via proteasomal dysfunction, has been shown to counteract the innate antioxidant mechanisms of FH-deficient cancers and lower tumor burden when combined with cisplatin in preclinical studies (Sourbier et al. 2010). Succinated, therefore defunct, glutathione peroxidase 4 (GPX4) is also seen in FH-deficient cells (Kerins et al. 2018). GPX4 normally protects cells from lipid peroxidation, as such, its succination by fumarate further sensitizes cells to ferroptosis (Kerins et al. 2018; Stockwell 2019). Counterintuitively though, FH-deficient cells express increased levels of GPX4 and are hypersensitive to direct GPX4 inhibitors RSL3 and ML162 (Kerins et al. 2018). Fumarate accumulation has also been linked to inhibition of PTEN via posttranslational succination at its cysteine 211, precluding the succinated PTEN moiety from anchoring to the plasma membrane, thereby leading to up-regulation of phosphoinositide 3-kinase (PI3K)/protein kinase B (AKT) signaling pathway (Ge et al. 2022). This allows for potential therapeutics targeting the deregulated PI3K/AKT signaling via PI3K and AKT inhibitors. To date, AKT inhibitors have not shown much promise due to toxic side effects (Duan et al. 2020). However, PI3K inhibitors have been tried preclinically and showed efficacy (Van Looy et al. 2014). Nevertheless, treatment with PI3K inhibitors induced adaptive mutations in PTEN that nullified the treatment efficacy (Papa and Pandolfi 2019). It would be of interest to learn whether PI3K activation via PTEN cysteine succination would be more sensitive to PI3K inhibitors and less likely to develop resistance.

METABOLITES AS DIAGNOSTIC BIOMARKERS

Through the above discussion, we have elaborated on the link between genetic deficiencies of the TCA cycle and the unique metabolic landscape of the tumor and circulation. Such dramatic metabolic changes can be harnessed to monitor tumor burden and, with that, treatment efficacy. Magnetic resonance spectroscopy has been used serially to monitor metabolites in vivo, specifically succinate and fumarate levels in patients with SDH and FH deficiency, respectively (Lussey-Lepoutre et al. 2020; Wu et al. 2022). This works by detecting peaks of compounds of interest that do not overlap with other compounds to identify a clear signal, correlated to the compound of interest. This method has been proven compatible within a diverse array of TCA cycle-deficient cancers (Branzoli et al. 2023). Beyond traditional measures of tumor size, longitudinal metabolic imaging analyses may offer a noninvasive approach to record response to treatment (Moog et al. 2022). Beyond metabolic imaging, circulating metabolites may also be used as diagnostic biomarkers to monitor tumor burden and pharmacodynamic biomarkers to monitor treatment efficacy. Indeed, succinyladenosine and succinocysteine, the two circulating metabolites mentioned above, normally serve as classifiers for FH-deficient RCC, and were also used to effectively monitor disease progression and treatment response (Zheng et al. 2023).

DISCUSSION

Succinate dehydrogenase and fumarate hydratase are two key enzymes involved in the TCA cycle. Mutations in these enzymes have been associated with the development of various types of cancer, including paragangliomas, pheochromocytomas, renal cell carcinoma, and leiomyomas. The mutations in these enzymes lead to the accumulation of succinate and fumarate, respectively, which can have profound effects on cellular metabolism and lead to a variety of oncometabolic changes, including but not limited to the reconstruction of the tumor cell's vital metabolic processes, the adaptation of its expression profile, and the deleterious accumulation of reactive intermediates in the cell. Clinical trials have been conducted to explore the potential of targeting these oncometabolic changes for therapeutic purposes. Notable approaches include the targeting of glycolytic flux, glutamine metabolism, ferroptosis, TET inhibition, among others. Current therapies for tumors with SDH and FH mutations are limited. Surgical resection is often the first-line treatment for localized tumors, and there are no specific targeted therapies approved for TCA cycle-deficient cancers. However, advances in our understanding of the oncometabolic changes associated with SDH and FH mutations may lead to the development of more effective targeted therapies in the future.

CONCLUSION

Through the lens of this review, we wish to bring to the forefront the most recent advances in both the biological underpinnings of TCA cycle-deficient cancers, the rationale for clinical trials targeting vulnerabilities in these cancers as they stand in the early 2020s, and future prospects for therapeutic developments in avenues yet to be explored in-clinic. To date, there are currently no targeted therapies approved for tumors with SDH and FH mutations and many of our treatment options currently swing with the surgical knife, rather than tried and proven biological remedies. With hope for the future, advances in our understanding of the underlying oncometabolic changes associated with mutations of TCA cycle-deficient cancers may lead to the development of more effective combined therapeutic approaches, improving outcomes for afflicted patients.