Review Article | Published:

Coupling Krebs cycle metabolites to signalling in immunity and cancer

Nature Metabolismvolume 1pages1633 (2019) | Download Citation


Metabolic reprogramming has become a key focus for both immunologists and cancer biologists, with exciting advances providing new insights into the mechanisms underlying disease. There is now extensive evidence that intermediates and derivatives of the mitochondrial Krebs cycle—metabolites traditionally associated with bioenergetics or biosynthesis—also possess ‘non-metabolic’ signalling functions. In this review, we summarize the non-metabolic signalling mechanisms of succinate, fumarate, itaconate, 2-hydroxyglutarate isomers (d-2-hydroxyglutarate and l-2-hydroxyglutarate) and acetyl-CoA, with a specific focus on how such signalling pathways alter immune cell and transformed cell function. We believe that the insights gained from immune and cancer cells that are summarized here will also be useful for understanding and treating a range of other diseases.


In the past 5 years, there has been a remarkable increase in the knowledge of how intracellular metabolic changes in tumours and especially in immune cells are not only linked to energy demand and biosynthesis, but also to discrete effector mechanisms that alter cell behaviour in specific ways. An area of particular focus has been on the Krebs cycle (also known as the tricarboxylic acid (TCA) cycle or the citric acid cycle (CAC)), the primary oxidative pathway for acetyl-CoA and the pathway for generation of the reducing equivalents NADH and FADH2 in aerobic organisms. Importantly, NADH and FADH2 are required to transfer electrons to the mitochondrial respiratory chain, also known as the electron transport chain (ETC), a series of enzyme and coenzyme complexes found along the inner mitochondrial membrane (IMM). Transfer of electrons along the ETC occurs via several redox reactions to facilitate the generation of a proton (H+) electrochemical potential gradient, which subsequently drives the maintenance of a high, energy-rich ATP (ATP) to adenosine diphosphate (ADP) ratio via FoF1-ATP synthase. This process is referred to as oxidative phosphorylation (OXPHOS) and requires oxygen (O2).

The TCA cycle itself operates in the mitochondrial matrix and is an amphibolic pathway that acts as an important nexus for the integration of multiple catabolic and anabolic pathways, such as glycolysis and gluconeogenesis. As depicted in Fig. 1, the pathway consists of eight enzymes, namely citrate synthase (CS), aconitase (ACO2), isocitrate dehydrogenase (IDH), α-ketoglutarate dehydrogenase (α-KGDH), succinyl-CoA synthetase, succinate dehydrogenase (SDH), fumarate hydratase (FH) and malate dehydrogenase (MDH). The first reaction, an irreversible aldol condensation, is catalysed by CS and extends the four-carbon oxaloacetate to six-carbon citrate, with the additional two carbons derived from acetyl-CoA. In the second step, ACO2 catalyses the reversible stereo-specific isomerisation of citrate to isocitrate via cis-aconitate in a two-step reaction. To achieve this, ACO2 employs a dehydration–hydration mechanism. First, citrate is protonated at position C3 before deprotonation occurs at position C2 to form cis-aconitate (dehydration). Second, cis-aconitate flips 180° to allow the dehydration and hydration steps to occur on opposite faces of the intermediate, ensuring that the stereochemistry (2R, 3S) of the reaction is correct. Subsequently, cis-aconitate is converted to isocitrate (rehydration). The third step, catalysed by IDH, requires NAD(P)+ and results in the oxidation of isocitrate to oxalosuccinate, generating NAD(P)H. Oxalosuccinate is subsequently decarboxylated to form the five-carbon α-ketoglutarate (α-KG) and CO2. In the fourth step, α-KGDH catalyses the oxidative decarboxylation of α-KG to form succinyl-CoA, NADH and CO2. Succinyl-CoA synthetase then catalyses the hydrolysis of succinyl-CoA, coupled to the condensation of GDP and inorganic phosphate (Pi) or of ADP and Pi to form succinate and either GTP or ATP, respectively, which are energetically equivalent. As such, the reaction catalysed by α-KGDH is referred to as substrate-level phosphorylation and constitutes the fifth step in the Krebs cycle. In the sixth step, SDH, also complex II of the ETC, catalyses the oxidation of succinate to fumarate. SDH, which uses FAD as a prosthetic group, generates FADH2 from the oxidation of succinate. The two electrons carried by FADH2 are subsequently used to reduce ubiquinone (Q) to ubiquinol (QH2). In the seventh step, FH catalyses the hydration of the α,β-unsaturated carbonyl compound fumarate to form l-malate. The final step in the cycle is catalysed by MDH and results in the oxidation of malate to generate NADH. This step also regenerates oxaloacetate, which can then be re-used by CS to enable continuation of the cycle.

Fig. 1: The Krebs cycle.
Fig. 1

The Krebs cycle is a metabolic pathway operating in the mitochondrial matrix of all aerobic organisms. Breakdown of nutrients, such as glucose, generates acetyl-CoA, which can then be funnelled into this pathway. For a full cycle to be completed, a series of ten enzymatic reactions are required. These reactions are catalysed by the mitochondrial enzymes CS (1), IDH (2), ACO2 (3), α-KG dehydrogenase (OGDH) (4), succinyl-CoA synthetase (5), SDH (6), FH (7) and MDH (8). The primary function of the TCA cycle is to generate reducing equivalents, such as NADH and FADH2 (produced by SDH). NADH and FADH2 can then transfer electrons to the ETC to drive OXPHOS and the production of ATP via FoF1-ATP synthase.

Intriguingly, Krebs cycle–derived metabolites have also been attributed both signalling functions and impact on multiple processes critical for immune cell activation and cellular transformation. The biological targets of succinate, itaconate, fumarate, 2-hydroxyglutarate (2-HG) and acetyl-CoA are of direct relevance to immunity and/or tumorigenesis, as they inhibit specific enzymes or drive covalent modifications of proteins to alter their functions. A common outcome of these processes is epigenome remodelling and altered expression of specific sets of genes that can govern both immune cell effector functions and tumour deveopment. The complexities of metabolic rewiring may ultimately lead to new therapeutic modalities that could have a major impact on the pathogenesis of multiple diseases previously linked to metabolic reprogramming. The field is on the cusp of a major paradigm shift in the understanding of how intracellular metabolic changes lead to disease, presenting the exciting prospect of new approaches to complement or even replace current therapeutic approaches for several diseases with an unmet therapeutic need. This review will discuss the extensive evidence for the non-metabolic functions of succinate, itaconate, fumarate, 2-HG and acetyl-CoA, focusing on immune cell signalling and carcinogenesis. Finally, the therapeutic opportunities that have arisen or that may arise from understanding these signalling pathways will also be examined.

Succinate as a signal in macrophages and tumour cells

As mentioned, succinate is generated in the mitochondrial matrix in a reversible three-step reaction catalysed by the TCA cycle enzyme succinyl-CoA synthetase, with concomitant production of a high-energy nucleoside triphosphate. Recent data indicate that succinate is a pro-inflammatory metabolite that accumulates in macrophages treated with lipopolysaccharide (LPS) and interferon-γ (IFN-γ)1,2. As depicted in Fig. 2, succinate has been shown to act through several pathways to exert its inflammatory effects: through generation of mitochondrial reactive oxygen species (mtROS), by activation of hypoxia-inducible factor-1α (HIF-1α; discussed below) and through ligation of the G-protein-coupled receptor succinate receptor 1 (SUCNR1).

Fig. 2: The diverse signalling roles of succinate.
Fig. 2

Succinate levels are elevated in response to LPS stimulation, in synovial fluids from patients with RA, in the circulation in models of diet-induced obesity and in adipose tissue in response to hypoxia and hyperglycaemia (1). Succinate oxidation (as well as direct product inhibition) and ROS HIF-1α, which binds to the HRE in the IL-1β promoter (2). Succinate is the ligand for the G-protein-coupled SUCNR1. Ligand binding induces Gi and Gq signalling cascades (3). SUCNR1 ligation on dendritic cells induces cell migration (4) and acts in synergy with TLR ligands to increase IL-1β expression, and IL-1β and LPS further enhance SUCNR1 expression (5). Endogenously generated succinate is released, for example via SLC13a3 and SLC13a5 (SLC13a3/5) in NSCs, and binds to SUCNR1 on the same or nearby SUCNR1-expressing cells (6). Activation of SUCNR1 on NSCs induces the secretion of anti-inflammatory prostaglandin E2 (7). Succinate regulates the activity of the JMJDs and the TET family of 5mC hydroxylases, which play a role in histone and DNA demethylation, respectively, and can thereby remodel the epigenome (8). Another consequence of succinate accumulation is the modification of proteins by lysine succinylation (9).

Succinate and SUCNR1-mediated signalling in immunity

In 2004, He et al.3 made an important discovery when they found that the Krebs cycle–derived metabolites succinate and α-KG acted as ligands for the orphan G-protein-coupled receptors GPR91 (now known as SUCNR1) and GPR99 (now known as OXRG1), respectively. The authors demonstrated that ligation of SUCNR1 by succinate acted as an important hypertensive agent by modulating the renin–angiotensin system, independently of its bioenergetic role3. Since this seminal discovery, important roles for SUCNR1 in regulating numerous physiological processes implicated in health and disease have emerged, notably in the prevention of age-related macular degeneration (AMD)4 and immune cell function5. SUCNR1 is highly expressed in kidney, liver, spleen and small intestine3, and on dendritic cells5, and ligand binding induces Gi and Gq signalling cascades3. SUCNR1 ligation on dendritic cells or the U937 macrophage cell line induces cell migration, suggesting that succinate can act as a chemokine5. Extracellular succinate addition also acts in synergy with toll-like receptor (TLR) ligands to increase tumour necrosis factor-α (TNF-α) and interleukin-1β (IL-1β) expression5 and increases the capacity of dendritic cells to act as antigen-presenting cells. Priming of dendritic cells with succinate and antigen simultaneously elevates antigen-specific T cell activation, increasing TNF-α and IFN-γ production from these cells5. SUCNR1 is also expressed on tuft cells in the intestine, where its signalling promotes type 2 immunity to certain infectious agents, such as tritrichomonad protists, and drives small intestine remodelling via a tuft cell–ILC2 circuit6,7,8.

Littlewood-Evans et al.9 have demonstrated that succinate can act in an autocrine and paracrine manner to increase IL-1β production in macrophages. Following LPS treatment, both intracellular succinate levels and the cell surface expression of SUCNR1 are increased. Interestingly, endogenously generated succinate is released through an undefined mechanism and binds to SUCNR1 on the same or nearby SUCNR1-expressing cells to amplify IL-1β production. IL-1β itself further enhances SUCNR1 expression, fuelling this cycle of cytokine production. Macrophage activation and IL-1β production are decreased in the absence of SUCNR1 both following LPS treatment and in a model of antigen-induced arthritis. This correlates with an observed increase in succinate in the synovial fluid from patients with rheumatoid arthritis (RA), suggesting that chronically elevated succinate is pathological in this setting10. Intriguingly, HIF-1α expression is also increased in synovial joints of patients with RA11, suggesting that succinate may act through two routes to increase inflammation in RA: SUCNR1 and HIF-1α.

Tissue and circulating succinate are also elevated in other chronic inflammatory settings, such as models of diet-induced obesity12. It has been suggested that exposure of adipocytes to hypoxia and hyperglycaemia (for example, during obesity) induces succinate release from adipose tissue in mice13, and this is associated with macrophage infiltration into the adipose tissue. The intracellular pathway driving succinate accumulation in these cells or its release from these cells is unclear. In the absence of SUCNR1, a reduction in absolute numbers of macrophages, but not inflammatory signalling, was observed, and this resulted in an improvement in adipose tissue inflammation and glucose tolerance in mice following high-fat diet feeding. In contrast to previous reports, the authors observed no effect of succinate alone on macrophage infiltration, suggesting that additional signals present in the hypoxic environment may also be required for succinate-induced chemotaxis. More recently, succinate has been implicated as a systemic signal for thermogenesis. It was demonstrated that brown adipocytes possess a unique capacity to sequester extracellular succinate, which serves to increase UCP1 activity as a consequence of elevated mtROS14. Interestingly, the authors demonstrate a role for shivering muscle in the generation of succinate following cold exposure. Succinate enters the circulation and is sequestered by the brown adipose tissue. Muscle has been demonstrated to produce succinate in other contexts too15. For example, it has been demonstrated that succinate is elevated in the plasma of marathon runners, and it is possible that muscle may act as a source tissue15.

Succinate has also been implicated in the pathogenesis of neuroinflammation. In a mouse model of multiple sclerosis (MS), transplantation of neural stem cells (NSCs) reduced succinate levels in the cerebrospinal fluid and thereby decreased infiltration of damaging macrophages and microglial cells16. Mechanistically, ligation of SUCNR1 on the surface of NSCs upregulates the expression of the dicarboxylate co-transporter SLC13a3 and SLC13a5, correlating with increased uptake by the NSCs and thereby scavenging succinate away from pro-inflammatory immune cells. Activation of SUCNR1 on NSCs also induces the secretion of anti-inflammatory prostaglandin E2, further contributing to their anti-inflammatory phenotype.

On the basis of these data, it may be predicted that SUCNR1 antagonists will be protective in settings where inflammation is exacerbated, such as graft rejection or autoimmunity. A selective antagonist for human and rat SUCNR1 has been shown to be protective in hypertension17. Similarly, potent agonists have recently been synthesised, which will deepen our understanding of the signalling pathways affected by this important receptor18. Whether these compounds will have therapeutic potential remains to be explored.

HIF-1 and succinate in inflammation and tumorigenesis

Another mechanism by which succinate exerts its signalling properties involves the transcription factor HIF-1, as shown in Fig. 2. HIF-1 is a transcriptional regulator, central in the response to hypoxia19. HIF-1 was first discovered by Semenza and Wang20 in 1992, and it was found to promote erythropoiesis by binding the erythropoietin gene enhancer. HIF-1 transcriptionally directs a switch in metabolism from oxidative phosphorylation to glycolysis, which allows rapid ATP production without the need for oxygen21. It is well-established that HIF-1 activity promotes tumour progression. Tumour environments are often oxygen-deprived, and such hypoxic conditions lead to activation of HIF-1. HIF-1 binds to hypoxia response elements (HREs) in target genes and increases the expression of numerous glycolytic enzymes and glucose transporter-1 (GLUT1) and GLUT3. It also targets genes involved in erythropoiesis, angiogenesis and proliferation.

HIF-1 is composed of a β-subunit that is constitutively active and an oxygen-sensitive α-subunit. HIF-1α is tightly regulated by prolyl hydroxylases (PHD), a class of α-KG-dependent dioxygenases (α-KGDDs) that convert α-KG to succinate and use molecular oxygen to hydroxylate HIF-1α. In the presence of oxygen, conserved proline residues in the HIF-1α subunit are hydroxylated by hypoxia-inducible factor prolyl hydroxylase 1 (PHD1)–PHD3 (ref. 22). Hydroxylation provides a recognition site for the binding of Von Hippel-Lindau (VHL) E3 ubiquitin ligase, which ubiquitinates HIF-1α and subsequently targets it for proteasomal degradation23. This degradation is prevented in conditions of limited oxygen, allowing for HIF-1α translocation to the nucleus, dimerization with HIF-1β and binding of target genes19.

HIF-1α is vital for the switch to glycolysis observed in M1 macrophages and dendritic cells following stimulation. It upregulates the expression of several glycolytic genes and aids in the switch to glycolysis by maintaining levels of nicotinamide adenine dinucleotide (NAD+), a vital co-factor in the glycolytic pathway. It does so by increasing the expression of lactate dehydrogenase (LDH), which reduces pyruvate to lactate and consequently generates NAD+ that will be reduced to NADH by an active glycolytic pathway. It also increases expression of pyruvate dehydrogenase kinase (PDK), which phosphorylates and inhibits pyruvate dehydrogenase (PDH); this limits production of acetyl CoA, which enters the Krebs cycle24.

While the primary role of HIF-1α is to induce a switch to glycolysis, which is likely to be important at sites of inflammation where oxygen levels are low19, hypoxia is not essential for HIF-1α activation in immune cells. Ligation of TLR4 by LPS can transcriptionally induce HIF-1α, as well as a range of HIF-1α target genes under hypoxic and also normoxic conditions2, in a nuclear factor-kappa B (NF-κB)-dependent manner25.

LPS can also stabilize HIF-1α indirectly by increasing succinate levels. Evidence for the role of succinate in HIF-1α stabilization first came from cancer studies. An elevation in succinate has been reported in rare hereditary tumours, such as paragangliomas (PGLs) and gastrointestinal stromal tumours (GISTs), that possess mutations in the genes encoding SDH26. The PHDs generate succinate from α-KG, and importantly succinate (acting via product inhibition) will inhibit the PHDs and stabilize HIF-1α (ref. 27). An additional mechanism by which succinate accumulates in tumours involves the mitochondrial chaperone tumour necrosis factor receptor–associated protein 1 (TRAP1), which is highly expressed in many tumours28. TRAP1 decreases SDH activity, resulting in succinate accumulation. More recently, however, it has been demonstrated that succinate accumulation in tumours promotes angiogenesis via upregulation of vascular endothelial growth factor (VEGF) in a HIF-1α-independent manner. Succinate instead activates extracellular regulated kinase 1 (ERK1) and ERK2 and signal transducer and activator of transcription 3 (STAT3) in a SUCNR1-dependent manner29.

Analogous to tumours, LPS-induced succinate was shown to both directly and indirectly (via ROS) inhibit PHD activity in macrophages, resulting in stabilization of HIF-1α (ref. 2). Like succinate, ROS are critical regulators of HIF-1α activity. ROS can stabilize HIF-1α by inducing non-enzymatic decarboxylation of α-KG30 and by oxidizing iron (Fe2+) to Fe3+, two required PHD co-factors31. ROS also impair the activity of factor inhibiting HIF (FIH), another member of the α-KGDD family31, which hydroxylates an asparagine residue in the carboxy terminus of HIF-1α, resulting in a decrease in transcriptional activity. It should be noted that hypoxia and ischaemia, as well as decreased oxidative phosphorylation, drive succinate accumulation in the mitochondrial matrix32,33. As such, following ischaemia or hypoxia, or when SDH operation is affected by genetic aberrations (such as tumours harbouring SDH mutations), elevated succinate can potentiate the hypoxic phenotype by promoting further HIF-1α stabilization.

LPS-induced HIF-1α expression not only promotes glycolysis but is also directly pro-inflammatory. The finding that hypoxia was capable of inducing IL-1β production in astrocytes in an NF-κB-independent, HIF-1α-dependent manner34 led to the discovery of an HRE in the IL-1β promoter2. The sustained induction of IL-1β in response to LPS was shown to require HIF-1α and to occur under normoxic conditions. The loss of HIF-1α in macrophages decreased TNF-α, IL-1β and IL-1α production but did not affect production of the anti-inflammatory cytokines IL-4 or IL-10 (ref. 35). A critical consequence of succinate elevation in response to LPS is the potentiation of inflammatory signalling and, in particular, IL-1β, in a HIF-1α-dependent manner2. This requires both succinate oxidation by the enzyme SDH and an increase in mitochondrial membrane potential. These factors combine to drive pro-inflammatory ROS production, HIF-1α stabilization and increased IL-1β following LPS treatment2,36. Succinate reciprocally decreases anti-inflammatory cytokines, such as IL-10 and IL-1RA, but whether this is HIF-1α-dependent remains to be explored36. Limiting succinate oxidation with the prodrug dimethylmalonate (DMM), which releases malonate, a potent SDH inhibitor, profoundly represses LPS-induced HIF-1α, IL-1β and ROS and boosts IL-10 and IL-1RA. ROS production in this setting is believed to be a result of reverse electron transport (RET) at complex I of the ETC. Supporting this, the complex I inhibitor rotenone significantly decreased LPS-induced ROS. It should be noted that if the ETC was operating in the forward direction, then rotenone would boost RET-independent ROS production. To further investigate the potential role for RET in this process, the authors36 examined macrophages from mice expressing an alternative oxidase (AOX) from Ciona intestinalis37. AOX provides a pathway to oxidise excess electrons that build up in the Q pool (as a result of succinate accumulation, for example) that would otherwise contribute to mtROS production. AOX expression in macrophages also impaired LPS-induced ROS production. These data suggest that complex I of the ETC is an important site of ROS production in macrophages and that such ROS production may depend on RET. In other conditions, complex I can generate significant ROS when operating in the forward direction (dependent on the NADH/NAD+ ratio). Deletion of a subunit of complex I, NDUFS4, is sufficient to generate substantial ROS in macrophages, most likely as a result of impaired electron shuttling down the ETC and thereby enhanced ROS production, which will in turn promote inflammation38. This is unlikely to be RET-dependent ROS production as the membrane potential, which much be significantly elevated to drive ROS via RET, is decreased in NDUFS4-deficient macrophages39. Furthermore, these studies were performed in the absence of a stimulus like LPS, which is known to drive Q pool reduction (another requirement for RET). These data suggest that complex I can generate ROS through a variety of mechanisms in macrophages to promote inflammation.

The relationship between HIF-1α and succinate in macrophages has been extensively studied; however, this relationship is less well explored in other cell types, particularly adaptive immune cells, such as CD4+ T cell subsets. The initial observation that LPS drives profound succinate accumulation in macrophages may account for the innate immune cell bias. Whether other stimuli are capable of driving a training phenotype similar to that of β-glucan warrants further investigation.

Succinate regulates the epigenome and innate immune memory

α-KG and succinate can regulate the activity of other members of the α-KGDD family of enzymes, in particular those involved in the regulation of histone and DNA demethylation, the α-KG-dependent Jumonji-C-domain-containing histone demethylases (JMJDs) and the ten eleven translocation (TET) family of 5mC hydroxylases, which play a role in DNA demethylation. Like the PHDs, these enzymes are subject to product inhibition by succinate, and therefore, their activity is dependent on the ratio of α-KG to succinate40. In this way, succinate and α-KG can remodel the epigenome. Methylation marks on histones can both positively and negatively influence gene expression, while DNA methylation is typically associated with an open chromatin state and active transcription; therefore, epigenetic changes can alter gene function, acting as on/off switches. Importantly, these changes in gene expression are heritable and appear to be intimately linked to the metabolic state of cells41. Succinate, fumarate and α-KG may also indirectly regulate the activity of histone demethylases through their effects on HIF-1α, which can bind to and induce the expression of certain histone demethylases, including JMJD1A, JMJD2C and JMJD2B, which are histone 3 lysine 9 (H3K9) demethylases41.

Another critical consequence of altering the epigenome is the newly emerging concept of innate immune training. It has been recently revealed that innate immune cells demonstrate a form of immunological memory and have the ability to respond more robustly to a second potentially unrelated stimulation42. At a molecular level, this involves epigenetic modifications, leading to stronger gene transcription upon re-stimulation as opposed to gene recombination that occurs in the adaptive memory response. Epigenetic marks can persist after the stimulus that induced them has resolved and are therefore stable signals that can be sustained for days or even longer. Such innate immune training has been demonstrated to occur after training of human monocytes in vitro with β-glucan, a component of Candida albicans, and subsequent treatment with LPS; following this, elevated cytokine production was observed upon re-stimulation with LPS43. Interestingly, in addition to potentiating cytokine production in response to LPS, training with β-glucan has been shown to induce epigenetic upregulation of genes involved in glycolysis, such as HIF-1α, hexokinase (HK) and pyruvate kinase, and to increase succinate levels44. HIF-1α reciprocally enhanced trained immunity in response to β-glucan. Inhibition of HIF-1α with ascorbate impaired the training effect induced by β-glucan and decreased TNF-α production in these cells, demonstrating the importance of HIF-1α for enhanced immunity. Training with β-glucan was shown to be protective against Staphylococcus aureus infection, and this effect was abrogated in HIF-1α-deficient mice. As shown in Fig. 2, succinate and other metabolites may therefore be capable of influencing the epigenome through effects on HIF-1α and perhaps subsequently on IL-1β, which has also been demonstrated to induce trained immunity in monocytes44. Whether other stimuli are capable of driving a similar training phenotype as β-glucan does warrants further investigation.

Succinylation as a covalent modification

Another consequence of dysregulated succinate metabolism is the recently identified post-translational modification (PTM), lysine succinylation. This modification is caused by the accumulation of succinyl-CoA, which can result from SDH inhibition and succinate accumulation39. Treatment of mouse fibroblasts with the SDH inhibitor 3-nitropropionic acid increases succinylation45. This modification induces a 100-Da change in mass, comparable to that of two well-established lysine modifications: acetylation and dimethylation. Importantly, succinylation masks the positive charge on lysine, likely resulting in a significant conformational change in the protein. Western blot analysis of whole-cell lysates revealed that this modification is evolutionarily conserved and that substrates are numerous46, including proteins involved in cellular metabolism45. Succinyl-proteome profiling in bacteria47, plants48,49 and HeLa cells all point toward metabolic pathways as key targets for this PTM. A study in yeast identifies histones as targets of this PTM; mutation of succinylation sites was found to reduce cell viability50.

The enzyme responsible for succinylation is yet to be identified, and indeed it is possible that this reaction occurs non-enzymatically by direct reaction between succinyl CoA and the modified protein48. However, sirtuin 5 (SIRT5), which was previously thought to function primarily as a deacetylase, has been shown to have potent desuccinylase activity51. Interestingly, SDHA is a target of lysine succinylation. SIRT5-deficient mice had significantly increased SDH activity, suggesting that succinylation positively regulates its activity45. This PTM appears to be LPS-inducible. LPS decreases SIRT5 expression in macrophages and increases protein succinylation2.

The α-ketoglutarate dehydrogenase complex (KGDHC) has also been suggested to mediate succinylation in an α-KG-dependent manner. Inhibition of KGDHC reduces succinylation of proteins in neuronal cells52. The authors identify the pyruvate dehydrogenase complex (PDHC), IDH and FH as targets of succinylation, with succinylation decreasing IDH activity and increasing FH activity53.

Succinylation can also modulate macrophage function. Succinylation of Lys311 of pyruvate kinase M2 (PKM2), a key glycolytic enzyme required for the shift to glycolysis in activated macrophages, was shown to limit its activity by promoting its tetramer-to-dimer transition54. The authors demonstrate that SIRT5 desuccinylates and activates PKM2, which limits IL-1β production. Conversely, SIRT5-deficient mice exhibit hypersuccinylation and increased IL-1β. There are many aspects of this PTM that require further investigation. The breadth of the targets of succinylation and indeed the enzyme catalysing this PTM remain to be determined, as well as precisely how this PTM alters protein function.

These data demonstrate that succinate can have a profound impact on cellular function, acting both intracellularly via ROS, HIF-1α and succinylation and histone and DNA modification and extracellularly via SUCNR1.

Itaconate as a novel anti-inflammatory metabolite

One of the most striking examples of a Krebs cycle–derived metabolite acting as a signal in immunity is itaconate1,55,56,57,58. This previously unidentified dicarboxylic acid was first discovered in 1836 as a product of citric acid distillation by the Swiss chemist Samuel Baup59. In 1840, it was independently synthesised by the decarboxylation of cis-aconitate, resulting in the introduction of its current name, itaconic acid, an anagram of aconitic acid59. While itaconic acid has long been used in the industrial arena for the synthesis of various polymers, the first (unwittingly) relevant discovery regarding the role of itaconate in mammalian biology came in 1995 when Lee and colleagues59,60 cloned immunoresponsive gene 1 (Irg1), a gene potently upregulated in LPS-activated peritoneal macrophages. However, it was not until 2011 that the production of itaconic acid in a mammalian system, and the first suggestion that it may play a role in cellular immunity, was uncovered61. Intriguingly, it was described in two separate immune contexts, both in the lungs of Mycobacterium tuberculosis (Mtb)-infected mice61 and secreted into the supernatant of LPS-activated RAW264.7 cells (a macrophage cell line)62. A key discovery in the field of itaconate biology followed when Michelucci and colleagues63 identified immunoresponsive gene 1 (IRG1) as the enzyme responsible for itaconate synthesis in both mouse and human macrophages. This subsequently led to the renaming of IRG1 to cis-aconitate decarboxylase (CAD). Although a role for itaconate as an anti-bactericidal agent has previously been suggested (see refs. 63,64,65,66), more recently, itaconate has been found to be an important immunomodulatory metabolite56,67,68,69.

Itaconate is an endogenous SDH inhibitor

Despite the known anti-bacterial function of itaconate, its role in regulating macrophage function remained virtually unexplored until 2016. Since 2016, several key studies have uncovered a role for itaconate as a crucial anti-inflammatory metabolite that negatively regulates the inflammatory response and cytokine production56,67,68,69, as shown in Fig. 3. Importantly, Lampropoulou et al.56 first demonstrated that bone-marrow derived macrophages (BMDMs) activated by LPS and pre-treated with a cell-permeable methyl ester derivative of itaconate, dimethyl itaconate (DI), exhibited potent inhibition of pro-inflammatory mediators, including nitric oxide (NO), ROS and the cytokines IL-6, IL-12p70 and IL-1β. Furthermore, Irg1-deficient BMDMs, in which genetic deletion of Irg1 completely abolished itaconate synthesis, exhibited a significant increase in the production of IL-12, IL-6 and NO and, under conditions that activate the NLRP3 inflammasome, increased IL-1β and IL-18. An increase in HIF-1α, a critical regulator of aerobic glycolysis and IL-1β in macrophages, was also observed in Irg1-deficient BMDMs stimulated with LPS. As such, this study was the first to highlight a role for itaconate as an anti-inflammatory metabolite.

Fig. 3: Itaconate is a thiol-reactive anti-inflammatory metabolite.
Fig. 3

In LPS-activated macrophages, mitochondrial IRG1/CAD (which requires autocrine/paracrine IFN-β signalling) synthesises itaconate from cis-aconitate (1), whereby it can inhibit SDH (2) or be exported out of the mitochondria via mitochondrial carrier proteins (3). In the cytosol, itaconate alkylates KEAP1, a novel PTM termed 2,3-dicarboxypropylation, and GSH to form an itaconate-GSH adduct, or 2,3-dicarboxypropyl-GSH (4). This in turn activates the anti-inflammatory and anti-oxidant transcription factors NRF2 and ATF3 (5). Levels of the metabolite 2,3-dicarboxypropyl cysteine (itaconate–cysteine adduct) increase, which is indicative of the turnover of 2,3-dicarboxypropylated targets (6). Activation of NRF2 acts to negatively regulate the pro-inflammatory cytokine IL-1β (7), while activation of ATF3 inhibits IκBξ and IL-6 (8). Itaconate has also been shown to boost ROS levels in certain contexts and promote tumour growth (9).

Mechanistically, Lampropoulou et al.56 suggested that the ability of itaconate to modulate inflammation arose from its ability to competitively inhibit SDH and succinate oxidation56. Intriguingly, itaconate was first shown to act as a competitive SDH inhibitor more than 60 years ago70, which prompted speculation that it was an endogenous SDH inhibitor, akin to malonate, in macrophages. Supporting this notion, exogenous treatment of macrophages with DI and of RAW264.7 cells and A549 cells (a lung adenocarcinoma cell line) with itaconic acid were found to induce succinate accumulation, indicative of SDH inhibition55,58. Likewise, itaconic acid was found to competitively inhibit purified SDH56. Crucially, LPS-stimulated Irg1-deficient BMDMs displayed significantly attenuated succinate accumulation and an increased oxygen consumption rate (OCR), further supporting the role of itaconate in SDH inhibition and metabolic rewiring in macrophages55,56. As such, itaconate inhibition of SDH was proposed to limit inflammation by blocking the generation of ROS derived from RET at complex I, as previously discussed. Together these papers offer a molecular explanation for succinate accumulation1, and the breakpoint observed in the TCA cycle by Jha et al.1, whereby isotopic tracing using U-13C-glutamine showed that approximately 35% of the pool of succinate, but only 22% of malate, could be attributed to glutamine anaplerosis in BMDMs treated with LPS, suggesting inefficient succinate-to-fumarate transition at SDH.

However, it must also be noted that, compared with malonate, itaconate is a relatively weak competitive inhibitor of SDH55,57, which raises the question as to whether any additional mechanisms could contribute to the anti-inflammatory effects of itaconate. Furthermore, DI, the itaconate derivative used in these studies, is not metabolised to itaconate intracellularly. It has also been shown that exogenous addition of 13C-itaconate to RAW264.7 cells was sufficient to increase intracellular levels of unlabeled succinate (suggestive of SDH inhibition), with no evidence of cellular uptake71. It is possible that the observed effect of itaconic acid on succinate levels could be receptor-mediated, as posited by El Azzouny and colleagues71, especially considering that the discovery of Krebs cycle intermediate signalling via GPCRs, such as that observed for succinate and SUCNR1 (refs. 3,71). It is also important to note that itaconate is an α,β-unsaturated carbonyl compound, making it a potential Michael acceptor that could undergo nucleophilic attack by the thiolate ion of protein cysteine residues; as such, the methyl esterification of the proximal carboxyl group conjugated to the alkene is predicted to increase the electrophilicity of itaconate. In this way, the effects of DI could possibly be attributed to electrophilic inactivation of metabolic enzymes, as opposed to an increase in intracellular itaconate71. Supporting this, Lampropoulou et al.56 reported an upregulation of genes involved in phase II conjugation, glutathione conjugation and biological oxidations from an RNA sequencing (RNA-seq) screen performed in DI-treated BMDMs.

Itaconate alkylates redox-sensitive cysteine residues

Independently, Mills et al.68 also noted that itaconate could potentially alkylate cysteine residues to form a 2,3-dicarboxypropyl adduct, as shown in Fig. 3. As such, it was hypothesised that the anti-inflammatory effects of itaconate could be due to alkylation of the key redox-sensing protein kelch-like ECH-associated protein 1 (KEAP1) and activation of the anti-inflammatory and anti-oxidant transcription factor nuclear factor, erythroid 2 like 2 (NFE2L2, also known as NRF2)68,72. DI was first employed as an experimental tool and was shown to potently induce NRF2 and several NRF2 target genes in unstimulated and LPS-activated BMDMs. This strongly suggested that DI was indeed a thiol-reactive metabolite; however, esterification on the 1-position would render it an activated Michael acceptor, akin to dimethyl fumarate (DMF), and this could result in rapid activation of NRF2 along with many other reactive intracellular thiols68,73. To overcome the limitations of DI, a new itaconate derivative, 4-octyl itaconate (4-OI), was synthesised in which the octyl ester group was located on the carboxyl group distal to the alkene. Importantly, itaconate and 4-OI were shown to have similar thiol-reactivity profiles, whereas DI was shown to significantly and acutely deplete glutathione (GSH) levels, confirming that it was indeed an activated Michael acceptor. Furthermore, the ability of DI to induce NQO1 activity (encoded by a prototypical NRF2 target gene) was diminished when pre-incubated with GSH; however, the ability of 4-OI to induce NQO1 activity remained unaffected. These findings suggest that 4-OI represents a suitable cell-permeable itaconate surrogate for probing the physiological function of itaconate68. Importantly, pre-treatment of unstimulated and LPS-activated macrophages with 4-OI still resulted in a significant increase expression of in NRF2 and NRF2 target genes, namely Heme oxygenase 1 (Hmox1), suggesting that itaconate is an endogenous signal governing NRF2 activation. This finding was corroborated by Bambouskova et al.67, who demonstrated a decrease in LPS-induced NRF2 stabilization in Irg1-deficient BMDMs, confirming that endogenous itaconate stabilises NRF2.

Further supporting the hypothesis that itaconate is a thiol-reactive metabolite, a significant increase in the levels of the metabolite 2,3-dicarboxypropyl cysteine (itaconate-cysteine adduct), akin to the breakdown product of fumarate-mediated protein succination S-(2-succino)cysteine (2SC), was observed in LPS-activated BMDMs68. Mechanistically, 4-OI has been shown to alkylate KEAP1 on several cysteine residues, including cysteine 151 (C151), a principal redox-sensing cysteine important for activation of NRF2 (refs. 68,74). Furthermore, through a quantitative unbiased proteomic screen performed in both LPS- (which will drive intracellular itaconate accumulation) and 4-OI-treated macrophages, several metabolic and inflammatory proteins were found to be alkylated, representing the first demonstration of this novel post-translational modification (PTM) termed 2,3-dicarboxypropylation. Pre-treatment of LPS-activated BMDMs with 4-OI also resulted in a significant attenuation of IL-1β levels, an effect that was largely abrogated in Nrf2-deficent macrophages. It should also be noted that NRF2 can limit LPS-induced inflammatory gene transcription in a redox-independent manner through direct ligation of these genes and inhibition of RNA polymerase II recruitment72. Whether itaconate can mediate cytokine inhibition in this manner remains to be explored. 4-OI was also shown to be protective in an LPS lethality model in vivo, prolonging survival, improving body temperature regulation and decreasing pro-inflammatory cytokine production. In addition, a novel negative feedback loop between IRG1, itaconate and IFN-β was elucidated, whereby the induction of Irg1 and itaconate were largely dependent on autocrine–paracrine type I IFN signalling68. Intriguingly, pre-treatment of LPS-activated macrophages with 4-OI resulted in a significant attenuation of IFN-β levels and IFN-stimulated genes (ISGs), such as Isg20, an effect that was at least partially dependent on NRF2. As such, this study confirmed itaconate as an anti-inflammatory metabolite and provided a unique insight into its mechanism of action with the discovery of a novel PTM. Further work is required to understand the scope of this PTM and its effect on protein function.

In addition to NRF2 regulation, a NF-κB inhibitor zeta-activating transcription factor 3 (IκBξ–ATF3) inflammatory axis has recently emerged as a target of itaconate and its more electrophilic derivatives67. Bambouskova and colleagues (2018) demonstrated that DI could induce electrophilic stress and NRF2 expression in macrophages via depletion of intracellular GSH to form a DI–GSH adduct, akin to 2,3-dicarboxypropylation, as previously discussed. Interestingly, DI is reported to selectively inhibit LPS-induced IL-6 while TNF-α levels remain unaffected, an effect shown to be mediated through inhibition of IκBξ the major transcription factor governing secondary transcriptional responses to TLR activation67. Mechanistically, inhibition of IκBξ by DI was shown to be mediated by activation of the transcription factor ATF3 and occurred at the level of Nfkbiz translation, as assessed by eIF2α phosphorylation, in a NRF2-independent manner.

The use of itaconate ester derivatives as endogenous itaconate mimics poses limitations due to varying electrophilicities, depending on whether the ester group lies on the carboxyl group proximal or distal to the methylene. To assess this, Bambouskova et al.67 synthesised 1-ethyl itaconate (1EI) and 4-ethyl itaconate (4EI) and tested their effects on IκBξ. The authors demonstrate that 1EI (an activated Michael acceptor) could inhibit IκBξ; however, 4EI (which has similar reactivity to endogenous itaconate) could not inhibit IκBξ unless GSH synthesis was inhibited by buthionine sulfoximine (BSO). This highlights the importance of acute electrophilic stress that is required to induce this novel pathway. To determine the endogenous relevance of itaconate to the IκBξ–ATF3 inflammatory axis, the authors utilised Irg1-deficient macrophages; however, no differences were observed in IκBξ levels between Irg1–/– and wild-type macrophages stimulated with LPS. Even so, the authors posited that elevated itaconate levels may affect the induction of IκBξ in LPS-tolerised macrophages. Although the induction of IκBξ was much lower in LPS-tolerised macrophages, there was an increase in Irg1-deficient macrophages, suggesting that at later time points when itaconate has built up to sufficient levels (>18 h), it may induce low levels of electrophilic stress by modifying GSH, which subsequently regulates the tolerisation process67. As such, a unique link between electrophilic stress and the IκBξ–ATF3 inflammatory axis was uncovered that may be exploited for therapeutic gain.

A role for IRG1 and itaconate in the tolerisation process and as a negative regulator of inflammation is supported by the finding that IRG1 suppresses the production of the pro-inflammatory cytokines TNF-α, IL-6 and IFN-β in LPS-tolerised BMDMs75. Li et al.75 attributed the effect of IRG1 on cytokine production to ROS-mediated induction of A20, a negative regulator of TLR signalling. Supplementation of ROS in Irg1-deficient BMDMs increased A20 expression and abolished the break of endotoxin tolerance. More recently, IRG1 was also found to be significantly increased in A20-deficient macrophages, which suggests an interesting feedback loop may exist between IRG1 and A20 expression76. Intriguingly, Li et al.75 also observed increased IFN-β and IRF3 signalling in LPS-tolerised, Irg1-silenced macrophages, supporting the existence of a negative feedback loop between IRG1, itaconate and IFN-β, as previously mentioned. IRG1 has also been shown to mediate the immunomodulatory effects of CO-releasing molecule 2 (CORM2)-induced HMOX1 on TNF-α production in both LPS-activated macrophages and in a murine endotoxin shock model77, and it also appears to play a major role in embryo implantation in the womb, a process thought to require immunosuppression78,79,80. More recently, an important role for myeloid-derived IRG1 (and presumably itaconate) in Mtb infection was uncovered. Nair et al.69 convincingly demonstrated an increase in pro-inflammatory cytokine production, including IL-6 and IL-1β, in Mtb-infected Irg1-deficent mice; this effect was independent of its ability to act as an anti-bactericidal agent. Interestingly, IRG1 was shown to specifically impair neutrophil recruitment to the lungs, curtailing excessive lung inflammation and pathology—an effect thought to be a consequence of itaconate-mediated transcriptional regulation of the inflammatory response. These data suggest that itaconate may dampen the immune response of Mtb-infected mice to limit immunopathology.

Exploration of the role for itaconate in immune cell activation is still in its infancy, and the profound levels to which it is produced in macrophages suggest that it is likely to be central to macrophage function. Further work to decipher its potentially diverse roles and whether it has important signalling functions outside macrophage biology are ongoing

Itaconate in macrophage–tumour crosstalk

Most recently, an intriguing study in peritoneal tumours (B16 melanoma and ID8 ovarian carcinoma) has demonstrated that tumour-infiltrating macrophages release itaconate, which potentiates tumour growth81. Itaconate was shown to boost OXPHOS and mtROS generation and subsequent MAPK activation in tumour-infiltrating macrophages (Fig. 3). The effect of itaconate on ROS production is consistent with a previous study in zebrafish macrophages, which demonstrated that IRG1 was essential for OXPHOS-driven mtROS production and bactericidal killing82. Furthermore, Weiss et al.81 demonstrated that ROS production in response to itaconate regulates NRF2, suggesting that there are two mechanisms for itaconate-induced NRF2 activation: directly via KEAP1 2,3-dicarboxypropylation and indirectly via ROS production (which would then modify KEAP1). NRF2 therefore appears to be an important target for itaconate. Furthermore, IRG1 (and presumably itaconate) levels were also shown to be markedly increased in glioma tissue. This was associated with a decrease in the microRNA miR-378, which targets Irg1, as well as poorer overall survival and clinicopathological parameters. Overexpression of miR-378 suppressed glioma tumour growth both in vitro and in vivo, the epithelial–mesenchymal transition (EMT) and metastasis83. These data are consistent with previous reports demonstrating that Irg1 acts as an oncogene, driving glioma pathogenesis84. As such, itaconate serves as another example of crosstalk between macrophages and tumours in the context of metabolic reprogramming.

Fumarate as a signal of the immune system and tumours

Fumarate is a Krebs cycle intermediate generated through the oxidation of succinate by SDH. It is also produced as a breakdown product of tyrosine metabolism and from the urea and purine nucleotide cycles85. Fumarate is subsequently converted to malate by FH, which catalyses the reversible hydration of fumarate to malate both within the Krebs cycle and in the cytosol, as shown in Fig. 4.

Fig. 4: Fumarate is an oncometabolite and epigenetic modifier.
Fig. 4

Under conditions of FH deficiency, fumarate levels markedly increase, leading to perturbations in mitochondrial OXPHOS. Elevated fumarate acts as a competitive inhibitor of SDH, inhibits complex I via succination of key (Fe-S) cluster biogenesis proteins (1) and inhibits ACO2 via succination through redox-sensitive cysteine residues (2). In the cytosol, increased fumarate can succinate KEAP1 and GSH to activate NRF2 (3), while fumarate also acts to competitively inhibit PHDs and stabilise HIF-1α (4). In the nucleus, fumarate acts as an epigenetic modifier, whereby it can inhibit the KDM family of histone demethylases (5) and the TET family of DNA demethylases (6), which act as a signal to induce EMT.

Fumarate regulates the epigenome and immune cell function

Fumarate has recently emerged as an important inflammatory signal in innate immune training, as outlined above44,86. Importantly, Arts et al.86 demonstrated that the accumulation of fumarate in β-glucan-treated monocytes was essential for trained immunity by enhancing cytokine production upon re-stimulation with LPS. Furthermore, increased fumarate levels were shown to be driven by enhanced shunting of glutamine into the TCA cycle, otherwise referred to as glutamine anaplerosis, as ascertained by a combination of biochemical techniques and the use of the glutaminase inhibitor BPTES87. Fumarate is a known epigenetic regulator that exerts its modulatory effects primarily through inhibition of α-KGDDs86,88. To determine the role of increased fumarate levels on the epigenetic landscape, Arts et al.86 used the cell-permeable fumaric acid ester monomethyl fumarate (MMF), which was shown to alter histone methylation similarly to β-glucan training. Furthermore, treatment with MMF augmented TNF-α and IL-6 secretion in β-glucan-trained macrophages stimulated with LPS. The authors proposed that fumarate accumulation acted to inhibit the KDM5 family of histone demethylases, which subsequently increased the levels of H3K4me3, a marker of active gene transcription, at the promoters of both Tnf and Il6, thus providing the first link between Krebs cycle rewiring and epigenetic regulation in an inflammatory context86.

Outside of trained immunity, fumarate has previously been reported to accumulate in LPS-activated macrophages1,2. Using an integrated high-throughput transcriptional metabolic profiling and analysis pipeline (CoMBI-T), Jha et al.1 uncovered the induction of an inflammatory argininosuccinate shunt, a metabolic pathway that links the Krebs cycle to the urea cycle. Induction of this pathway occurred via an increase in the expression of argininosuccinate synthase (ASS1), which catalyses the formation of argininosuccinate from aspartate, citrulline and ATP89. Together with the argininosuccinate lyase (ASL), ASS1 is responsible for the synthesis of the semi-essential amino acid arginine in most bodily tissues. Importantly, ASS1 also constitutes the rate-limiting step in the aspartate–argininosuccinate shunt and l-arginine biosynthesis. Considering that ASL cleaves argininosuccinate into arginine and fumarate, the authors proposed that this metabolic pathway accounted for the elevated fumarate levels observed in macrophages.

The electrophilic properties of fumarate derivatives can also regulate T cell function. Blewett et al.90 demonstrated that dimethyl fumarate (DMF), a potent electrophile currently used to treat psoriasis and relapsing-remitting MS, but not MMF inhibited the activation of human and mouse T cells. Using a proteomic approach, the authors discovered a range of proteins that regulate T cell function and are sensitive to covalent modification by DMF, including protein kinase C θ (PKCθ). Mechanistically, DMF blocked the association of PKCθ with the costimulatory receptor CD28 and T cell activation.

More recently, Kornberg and colleagues91,92 suggested that a potential consequence of elevated fumarate levels is negative regulation of glycolysis via succination (see below) of the active site cysteine residue (C152) in GAPDH. This suggestion arose from the observation that the cell-permeable fumarate esters DMF and MMF modify and irreversibly inhibit GAPDH enzymatic activity and aerobic glycolysis in peripheral blood mononuclear cells (PBMCs) from mice and patients with MS treated with DMF92. Induction of aerobic glycolysis is a key marker of activated immune cells and represents an important phenotypic switch to facilitate proliferation and immune cell effector functions. Two pro-inflammatory T cell subsets implicated in the pathogenesis of multiple sclerosis (MS), T helper 1 (TH1) and TH17 cells, rely heavily on aerobic glycolysis, and treatment with DMF or MMF inhibited their development in vitro, as previously reported92,93,94,95. Importantly, inhibition of aerobic glycolysis by the specific GAPDH inhibitor heptelidic acid (HA) and 2-deoxyglucose (2-DG) recapitulated the immunomodulatory effects of DMF in vitro, while in vivo administration of HA inhibited the development of experimental autoimmune encephalomyelitis (EAE), a murine model of MS92. As such, this represents an important proof-of-principle study supporting the concept that metabolically targeted therapies may be successfully used to treat inflammatory-driven diseases and also providing mechanistic insight into the anti-inflammatory action of DMF and MMF. Intriguingly, endogenous fumarate was also found to succinate GAPDH in both mouse and human macrophages, suggesting that elevated fumarate levels may act as an endogenous anti-inflammatory signal to limit inflammation; however, the precise role of fumarate accumulation remains an open question to be explored. It is important to note that while HA and DMF have proven to be beneficial in the treatment of debilitating diseases such as MS, prolonged and/or systemic inhibition of glycolysis is likely detrimental to the host and will impair immune (both innate and adaptive) cell function, which in the context of infection would be unfavourable. Further work is required to fully understand the effect of prolonged glycolytic inhibition.

Fumarate and cellular transformation

In addition to its role as an inflammatory regulator, fumarate can also be viewed as an oncometabolite96. Given its central role in energy metabolism, FH was long considered a housekeeping enzyme97. Consistent with this view, homozygous FH loss in fumaric aciduria, an autosomal recessive metabolic disorder, leads to very severe neurological disorders and is fatal in early childhood98. However, in 2002 it was shown that mutations of FH are the cause of hereditary leiomyomatosis and renal cell cancer (HLRCC), a rare cancer predisposition syndrome characterized by benign tumours of the smooth muscle of skin and uterus, and a very aggressive form of renal cancer87. These findings indicate that not only do cells survive the loss of FH, but in specific tissues, they can also undergo transformation to tumour cells. How FH loss leads to cancer has been a long-standing question in the field.

The defining biochemical feature of FH loss is the accumulation of fumarate, which has been implicated in tumorigenesis96. To date, several biological functions have been ascribed to fumarate. For instance, accumulated fumarate can bind and inactivate reactive thiol residues of proteins and peptides in a process called succination. This protein modification was originally described in diabetes99. This function of fumarate derives from its unsaturated dicarboxylic acid structure, which makes it a mildly electrophilic molecule that can be involved in a Michael addition with nucleophiles, such as reactive thiol residues of proteins. The first insights into the relevance of succination in patients with HRLCC was proposed in 2011, when it was shown that fumarate could bind reactive thiol residues and inactivate KEAP1, the negative regulator of the master antioxidant transcription factor NRF2, leading to a powerful antioxidant response100,101, as shown in Fig. 4. Later, it was also shown that fumarate binds glutathione (forming succinic-glutathione (succinicGSH)) and thereby depletes this antioxidant molecule, leading to unscheduled oxidative stress95,102. Of note, fumarate-dependent oxidative stress has also been shown to drive renal cells to senescence in an FH-deficient mouse model, and its bypass by the co-ablation of p21 increased malignant transformation in the kidneys103. Consistent with an important role of senescence bypass for the development of tumorigenesis, recent data in patients with HLRCC shows that CDKN2A (also known as P16), a key player in senescence, is hypermethylated and suppressed in tumours. Fumarate has been shown to cause succination of the mitochondrial aconitase ACO2, leading to an additional truncation of the TCA cycle104, and also of iron-responsive element binding protein 2 (IRP2), promoting the transcription of transferrin105. Finally, it has been shown that fumarate can drive succination of several members of the family of iron sulfur (Fe-S) cluster biogenesis proteins, causing a depletion of the Fe-S clusters required for the activity of mitochondrial enzymes, including complex I106. All these reactions appear to be independent of enzymatic catalysis, driven by the high level of fumarate accumulation and irreversible. Succination is a hallmark of FH-deficient tumours, and an antibody against succinated proteins has been proposed as a diagnostic marker for HLRCC107.

As mentioned, fumarate also inhibits a variety of α-KGDDs, enzymes involved in multiple cellular processes ranging from metabolism to signalling and epigenetics88,108,109, as shown in Fig. 4. Examples of α-KGDDs inhibited by fumarate are the PHDs, which are negative regulators of HIFs110. Accordingly, fumarate causes HIF stabilization even under normoxic conditions, that is, pseudohypoxia, a hallmark of FH-deficient tumours110. α-KGDDs are also involved in the de-methylation of DNA, RNA and histones. Recent data have demonstrated that fumarate acts as a powerful inhibitor of the TET family of DNA demethylases and of a series of histone demethylases88,108. In this context, fumarate is considered an important epigenetic modifier (reviewed in ref. 111). It has also been shown that fumarate accumulation leads to the inhibition of TET-dependent DNA demethylation of a class of anti-metastatic miRNAs, the miR200 family, promoting an EMT112, a process involved in cancer initiation and metastasis113. Of note, FH-deficient renal cancers are characterized by a distinct DNA hypermethylation phenotype, which also affects the tumour suppressor CDK2NA114, indicating that, by allowing senescence escape, this epigenetic activity of fumarate may have a critical role in tumour etiology.

Finally, fumarate and FH were recently proposed to be key players in the response to DNA damage. Originally described in yeast115, FH has a ‘moonlighting’ role in the nucleus that capitalizes on the epigenetic-modifying function of fumarate. Upon DNA damage, FH localizes in the nucleus, where it is phosphorylated by the catalytic subunit of DNA-PK (DNA-PKcs)116. Here, via its reverse activity, FH generates fumarate that inhibits the histone lysine demethylase KDM2B, which in turn facilitates the binding of proteins involved in DNA repair. In support of this unexpected function of fumarate, it has recently been shown that the aberrant accumulation of this metabolite in HLRCC correlates with increased endogenous damage, lower DNA repair efficiency and increased sensitivity to poly-ADP ribose polymerase (PARP) inhibitors117. Of note, this biological function of fumarate is also shared with succinate and 2-HG (see below). It has also recently been shown that FH loss in renal cancer cells leads to increased DNA damage and ionising radiation (IR) resistance due to a fumarate-dependent inhibition of mitotic entry after IR, even in the presence of unrepaired damage118. Overall, these lines of evidence provide a potential model for how fumarate accumulation promotes genomic alterations that could give rise to cancer formation in patients with HLRCC, cooperating with the other above-described oncogenic signals it elicits.

2-hydroxyglutarate in immunity and cancer

The next intermediate we will consider is 2-HG, which exists in two enantiomeric forms, l-2-HG and d-2-HG104, as shown in Fig. 5. Like itaconate, 2-HG was first discovered in the 1800s (in this instance by the German biochemist Karl Heinrich Ritthausen); however, it did not attract much attention until recently when its physiological function was uncovered119. Interest in these enantiomers heightened in the 1980s when they were linked to two rare but clinically related diseases, now termed l-2-hydroxyglutaric aciduria (L2-HGA) and d-2-hydroxyglutaric aciduria (D2-HGA), which manifest early in childhood and usually lead to severe disability and mental retardation in adulthood120,121,122. The cause of these so-called 2-hydroxyglutaric acidurias (2-HGAs) was attributed to germline mutations in l-2-HG dehydrogenase (L2HGDH) and d-2-HG dehydrogenase (D2HGDH), encoding the mitochondrial enzymes responsible for the conversion of 2-HG to α-KG and thus for the breakdown of 2-HG123,124,125. In the case of D2-HGA, about 50% of patients present with D2HGDH mutations, while the remaining half possess mutations in the gene encoding the Krebs cycle enzyme IDH2, which converts isocitrate to α-KG126. A third 2-HGA, in which both enantiomers accumulate, that arises from germline mutations in the mitochondrial citrate carrier SLC25A1 has also been discovered. This provides an intriguing link between mitochondrial citrate metabolism and 2-HG production127 that remains to be explored. In recent years, two landmark genomic studies carried out in human glioma and acute myeloid leukaemia (AML) found mutations in IDH1 and IDH2, and it is now appreciated that the genes encoding these enzymes (which drive elevated production of d-2-HG) are the most frequently mutated metabolic genes in human cancer128,129,130. Furthermore, the literature suggesting that these metabolites play a role in immune cell regulation is ever expanding.

Fig. 5: 2-HG is an oncometabolite and epigenetic modifier.
Fig. 5

2-HG exists as two enantiomers, l-2-HG and d-2-HG. Under conditions whereby IDH1/2 are mutated (mIDH1/2), this mutant enzyme preferentially acts on α-KG, as opposed to isocitrate, to generate d-2-HG, and in this context, d-2-HG accumulates to high levels (1); however, under normal conditions, d-2-HG levels are maintained at low concentrations via D2HGDH (2). In response to hypoxia or acidic pH, LDHA and MDH2 can promiscuously generate l-2-HG from α-KG (3), whose levels are usually maintained at low concentrations via L2HGDH (4). Accumulation of d-2-HG and/or l-2-HG in the nucleus results in the competitive inhibition of the KDM family of histone demethylases (5) and the TET family of DNA demethylases (6), thus acting as an important epigenetic modifier driving tumorigenesis and as a regulator of T cell immunity.

2-HG governs T cell differentiation and anti-tumour immunity

While the role of 2-HG in tumorigenesis has been a heavy focus of research over the last several years, an unexpected role for this unusual metabolite has more recently begun to emerge as a regulator of T cell function and inflammation. It has become increasingly appreciated that different T cell subtypes undergo dramatic metabolic remodelling that governs their proliferation, differentiation and effector functions131. TH17 cells, derived from naive CD4+ T cells, are one such subset that rely heavily on a switch from OXPHOS to glycolysis, a process that requires the activation of HIF-1α and induction of RORγt132,133. On the other hand, induced regulatory T (iTreg) cells, also derived from naive CD4+ T cells, do not require HIF-1α (refs. 132,133). Interestingly, Xu et al.133observed increased levels in of d-2-HG in TH17 cells when compared to iTreg cells. The elevated levels of d-2-HG were associated with increased DNA methylation levels at the Foxp3 locus, the master transcription factor driving iTreg differentiation, resulting in its repression133. Furthermore, TH17 cell differentiation could be initiated by exogenous addition of d-2-HG to naive CD4+ T cells or by genetic silencing of TET1 and TET2, established targets of 2-HG133,134,135. Mechanistically, d-2-HG production was driven by the conversion of glutamate to α-KG by the aspartate aminotransferase GOT1. Inhibition of GOT1 by aminooxyacetic acid (AOAA), which decreases d-2-HG levels, promoted Foxp3 expression and the differentiation of TH17 cells to iTregs133. Lastly, AOAA ameliorated disease progression in EAE, a disease in which TH17 cells mediate pathology. This study identified a crucial role for 2-HG as an epigenetic modifier governing T cell differentiation and suggests that therapeutic strategies aimed at lowering its levels may represent a promising approach for treatment of TH17-mediated autoimmune disorders.

In addition to CD4+ T cells, an unexpected role for l-2-HG in the promotion of tumour killing via CD8+ T cells, which are cytotoxic T lymphocytes important for anti-tumour immunity, has emerged136. Like TH17 cells, CD8+ T cells are highly dependent on the activation of HIF-1α and glycolysis, which is responsible for mediating trafficking into hypoxic tumour environments and inflamed tissue137,138. Interestingly, Tyrakis et al.136 observed a dramatic increase in l-2-HG levels, which reached up to 1.5 mM, in response to T cell receptor triggering in mouse CD8+ T cells. This increase occurred under normoxic conditions and was dependent on HIF-1α stabilization. Mechanistically, HIF-1α stabilization induced LDHA expression and activity, which promiscuously converted glutamine-derived α-KG to l-2-HG136. Accumulation of l-2-HG altered CD8+ T cell differentiation through modulation of both the histone and DNA methylation landscape of the cell and also through HIF stability. Furthermore, exogenous treatment of CD8+ T cells with l-2-HG greatly enhanced the in vivo proliferation, persistence and anti-tumour capacity of adoptively transferred CD8+ T cells136. 2-HG enantiomers are therefore beginning to emerge as important signalling moieties linking metabolic reprogramming, epigenetic alterations and effector functions of immune cells.

IDH mutations promote d-2-HG accumulation and tumorigenesis

Genomic studies have established that somatic mutations in IDH1 and IDH2 are perhaps the first genetic events to occur during tumorigenesis in human glioma and AML, and these appear to initiate pathogenesis by a common mechanism114,119,139,140. The most commonly occurring cancers thought to arise from mutant IDH1 and IDH2 (mIDH1/2) include glioma, cartilaginous tumours, AML, angioimmunoblastic T cell lymphoma (AITL) and intrahepatic cholangiocarcinoma (ICC). Additionally, mutations in these genes have also been reported to sporadically arise in prostate cancer, melanoma, medulloblastoma and hepatocellular carcinoma (HCC)119. Remarkably, almost all IDH1/2 mutations occur in a few unique locations in the active site of the enzymes, which result in the acquirement of a neomorphic enzymatic activity converting α-KG to the oncometabolite d-2-HG141,142,143. As with succinate and fumarate, d-2-HG acts as a competitive inhibitor of α-KGDDs, including but not limited to the TET family and JMJD family of DNA and histone demethylases, respectively144,145. d-2-HG, produced by mIDH1/2, has been shown to drive a CpG island methylator phenotype (G-CIMP) in both glioma and ICC tumours to promote tumorigenesis and has also been found to occupy the active site of histone demethylases, thereby increasing histone methylation in primary gliomas144,146,147,148. Furthermore, Flavahan et al.149 demonstrated that the IDH-driven G-CIMP phenotype occurs at cohesion and CCCTC-binding factor (CTCF)-binding sites, compromising this methylation-sensitive insulator protein, and enables a constitutive enhancer to interact aberrantly with and activate the glioma oncogene PDGFRA.

Emphasis has often been placed on epigenetic alterations elicited by elevated d-2-HG as a driver of tumorigenesis in IDH-mutant cancers; however, it is also apparent that it can antagonise other α-KGDDs. Although the half-maximal inhibitory concentration (IC50) of d-2-HG toward KDM family members is in the low micromolar range (24–106 μM), it can also inhibit FIH and PHD2, with IC50 values of 1.5 and 7.3 mM, respectively144. In IDH1/2-mutant gliomas, d-2-HG has been reported to accumulate to levels up to 35 mM; as such, d-2-HG could inhibit α-KGDDs to affect alternative cellular signalling pathways141. Indeed, d-2-HG accumulation in IDH-mutant glioma has been found to inhibit PHD hydroxylase and stabilise HIF-1α (ref. 150), and HIF-1α stabilization has been shown to occur in IDH1-mutant brain tissue128. l-2-HG is also a potent inhibitor of α-KG-dependent enzymes. It has been reported that under conditions of hypoxia, l-2-HG is selectively produced by mammalian cells151. The authors demonstrate that this is independent of IDH1/2 and is primarily mediated by LDHA and malate dehydrogenase (MDH) via ‘promiscuous’ reduction of α-KG and is enhanced by an acidic pH152. Under acidic conditions, the protonation of α-KG promotes binding to LDHA and l-2-HG production. Functionally, l-2-HG is sufficient to drive histone methylation, in particular histone 3 lysine 9 (H3K9me3), and can also promote HIF-1α stabilization in normoxic conditions and therefore may assist with hypoxic adaptation. These data suggest that l-2-HG might represent a metabolic response to certain environmental stimuli, including hypoxia and acidosis, that drives changes in cellular signalling and function. Although it has been reported that d-2-HG is a less potent inhibitor of α-KGDDs than l -2-HG, the concentrations at which it accumulates under pathogenic conditions suggest HIF-1α stabilization and impaired collagen biogenesis may contribute to tumorigenicity150,152,153,154,155. Intriguingly, another possible mechanism by which d-2-HG drives tumorigenesis is through the promotion of genetic instability, predisposing cells to oncogenic transformation119. d-2-HG has been shown to inhibit α-KG-dependent alkylation repair homologs, ALKBH2 and ALKBH3, which are critical DNA damage repair enzymes that detoxify 1-methyladenine (1 mA) and 3-methylcytosine (3mC) lesions in DNA derived from alkylating agents119,144,156. Indeed, IDH-mutant cells display reduced DNA repair kinetics, accumulate DNA double-stranded breaks (DSBs) and are sensitised to alkylating agents117,119,144,156. Furthermore, d-2-HG has been proposed to promote DNA damage by altering the expression of genes involved in DNA repair, namely the DNA damage sensor ATM (ataxia-telangiectasia mutated), whereby an accumulation of repressive histone methylation markers, H3K9 and H3K27, were found at the ATM promoter in Idh-mutant mouse haematopoietic stem cells and human AML samples157. As such, these findings directly link mIDH-derived d-2-HG accumulation to DNA damage, genetic instability and carcinogenesis.

Alternative d-2-HG targets in cancer development

In addition to the role of α-KGDD inhibition by d-2-HG in the regulation of histone and DNA methylation and genomic instability, d-2-HG has been found to affect several other cellular pathways that may also contribute to cancer initiation and progression119. Of note, d-2-HG produced by mIDH1 in low-grade glioma was found to activate mechanistic target of rapamycin (mTOR) signalling158. Mechanistically, it was proposed that d-2-HG destabilised DEPTOR, a negative regulator of mTORC1 and mTORC2, partly through inhibition of the histone demethylase KDM4A158. Furthermore, metabolomic profiling of mIDH1/2 gliomas revealed significant alterations in cellular metabolism, an effect recapitulated with the exogenous addition of d-2-HG159. This metabolic rewiring resulted in a decrease in a common dipeptide found in the brain N-acetyl-aspartyl-glutamate (NAAG), and NAAG levels were found to be significantly lower in mIDH-mutant human glioma tissue159. Likewise, mutations in IDH were also found to reprogram pyruvate and Krebs cycle metabolism, lower dependence on oxidative mitochondrial metabolism and silence LDHA expression, all suggesting that cellular reprogramming elicited by d-2-HG may play a prominent role in the pathogenesis of IDH-mutant tumours160,161,162. However, the mechanism by which d-2-HG orchestrates these events remains to be explored. Furthermore, d-2-HG activates NF-κB in bone marrow stromal cells in an IκB kinase–independent fashion, thus generating a stromal niche for IDH-mutant AML development163. Intriguingly, d-2-HG has also been found to be a competitive inhibitor of SDH, a key metabolic enzyme and tumour suppressor as previously discussed, resulting in the accumulation of succinate, impaired mitochondrial oxygen consumption and increased protein succinylation164. As such, the initiation and pathogenesis of IDH-mutant tumours may share parallels with SDH-mutant tumours. d-2-HG has also been found to bind the DNA methyltransferase DNMT1, which subsequently represses receptor-interacting protein kinase 3 (RIP3). Finally, d-2-HG also impairs necroptosis and the small GTPase CDC42, leading to suppression of mixed lineage kinase 3 (MLK3)-mediated apoptosis165,166. These findings therefore suggest that both suppression of RIP3-mediated necroptosis and MLK3-mediated apoptosis by d-2-HG may contribute to IDH-mutant carcinogenesis165,166.

Citrate as a source of acetyl-CoA for histone acetylation

One final metabolite to consider in the context of cellular signalling is acetyl-coenzyme A (acetyl-CoA), derived from citrate. Acetyl-CoA is an energy-rich intermediate that provides acetyl groups to the Krebs cycle through the generation of citrate, which is further oxidised for energy production. Acetyl-CoA is generated from the breakdown of carbohydrates, fatty acids and proteins though glycolysis, β-oxidation and the degradation of the amino acids leucine, isoleucine and tryptophan, respectively85. Acetyl-CoA serves as a building block for the synthesis of lipids, ketone bodies and amino acids. As fatty acid synthesis occurs in the cytosol, acetyl-CoA must exit the mitochondria to fulfil this function. To do so, following its conversion to citrate, it exits the mitochondria via the mitochondrial citrate carrier (CIC) and is then converted back to acetyl-CoA (and oxaloacetate) by the enzyme ATP-citrate lyase (ACL). It is this enzyme that endows acetyl-CoA with much of its signalling capacity in the context of tumour and immune cell biology.

Acetyl-CoA (derived from citrate) has been shown to drive histone acetylation, and this process can have a profound impact on both tumour and immune cell function, as shown in Fig. 6. Chromatin structure is regulated in part through modification of histones, including histone acetylation167. It has been demonstrated in mammalian cells that histone acetylation is dependent on ACL-induced production of acetyl-CoA168, with ACL silencing decreasing the degree of histone acetylation. One of the earliest studies to identify a role for ACL in histone acetylation demonstrated a requirement for the ACLY gene (encoding ACL) in response to growth factor stimulation and during differentiation, for example, the differentiation of murine 3T3- L1 pre-adipocytes into adipocytes. Cells depleted of ACL contained visibly less lipid than control cells. This was linked to decreased expression of genes involved in glucose uptake and glycolysis, which are required for adipocytes to engage in fat storage. It was shown that during adipocyte differentiation, global histone acetylation is determined by glucose availability through an ACL-dependent pathway168. These seminal findings highlight a clear link between nutrient sensing, metabolism and histone acetylation, which has now been extended to immune and tumour cell function.

Fig. 6: Acetyl-CoA as an epigenetic modifier in immunity and cancer.
Fig. 6

Acetyl-CoA, an energy-rich intermediate derived from citrate (and pyruvate), has been shown to drive histone acetylation, and this process can have a profound impact on both tumour and immune cell function. To facilitate this, citrate exits the mitochondria via the CIC (1) and is then converted back to acetyl-CoA by the enzyme ATP-citrate lyase (ACL) (2). In T cells, acetyl-CoA has been shown to enhance acetylation of histone H3 acetylation at lysine 9 (H3K9Ac) and to promote transcription of IFNG (3). In addition, IL-2 treatment of CD4+ T cells promotes phosphorylation of ACL and acetylation of cell cycle genes to promote T cell growth (4). Oncogenic Akt has also been shown to phosphorylate ACL to promote acetylation and epigenomic remodelling of cancer cells, a hallmark of human glioma and prostate cancer (5).

Acetyl-CoA and histone acetylation regulate immunity

Acetyl-CoA levels and histone acetylation have been implicated in the regulation of immune cells. It is well-established that glycolysis is required for effector T cell function, and a recent study from Peng et al.169 has provided mechanistic insight into this requirement. The authors demonstrate that LDHA, an enzyme that catalyses the reversible conversion of pyruvate and NADH to lactate and NAD+, is induced in activated T cells and enhances histone acetylation of IFN-γ via the maintenance of acetyl-CoA levels to promote its expression. Mice with a T cell-specific LDHA deficiency produced less IFN-γ and were protected from immunopathology. Mechanistically, LDHA-deficient T cells showed decreased acetyl-CoA levels and decreased histone H3 acetylation at the lysine 9 residue (H3K9Ac), a histone mark associated with active transcription, on the IFN-γ promoter. Furthermore, artificial acetyl-CoA reduction, through inhibition of ACLY, decreased IFN-γ expression and acetate supplementation, which augmented acetyl-CoA production, and corrected H3K9Ac marks and IFN-γ expression in LDHA-deficient T cells.

The mechanism of IFN-γ regulation proposed by Peng et al.169 differs from that proposed by Chang et al.170, who suggest that aerobic glycolysis boosted IFN-γ production by engaging GAPDH and preventing its binding to the three prime untranslated region (3′-UTR) of Ifng, thereby enhancing IFN-γ translation. These assays are technically demanding to perform, and this discrepancy remains to be clarified. However, it should be noted that the study from Peng et al.169 employed the use of an artificial construct—the bovine growth hormone 3′-UTR—rather than direct manipulation of the Ifng 3′-UTR itself. Multiple mechanisms involving acetyl-CoA may therefore exist to regulate Ifng expression. Activated T cells, which display elevated glycolysis, may remodel their chromatin, thereby generating Ifng mRNA that can be robustly translated, as GAPDH is engaged in glycolysis and no longer binds the Ifng mRNA. These data demonstrate the complex relationship between metabolism and gene regulation. It will be interesting to speculate whether acetyl-CoA can boost histone acetylation and transcription of other cytokines, or indeed other immune or tumour cells.

ACL and histone acetylation have also been implicated in T cell growth171. An unbiased mass spectrometry–based study of the nuclear phosphoproteome of resting and IL-2-treated CD4+ T lymphocytes revealed that ACL is phosphorylated in response to IL-2 treatment171. Pharmacological or genetic ablation of ACL function impaired IL-2-promoted T cell growth. The authors demonstrate that ACL is required to enhance histone acetylation levels and induce the expression of cell cycle regulating genes in response to IL-2. They speculate that this is as a result of acetyl-CoA generation, but do not experimentally demonstrate this concept.

Acetyl-CoA alters the epigenome of tumour cells

In 2014, Lee et al.172 demonstrated that acetyl-CoA levels and histone acetylation in tumours are regulated by oncogenic activation of protein kinase B (known as Akt) and subsequent ACL phosphorylation. This was one of the first reports to implicate metabolic reprogramming mediated by oncogenic activation in the regulation of the cancer cell epigenome. Importantly, histone acetylation and phosphorylated Akt levels correlate significantly in human glioma and human prostate cancer, and levels of histone acetylation have been shown to be predictive of which patients later exhibit biochemical failure, with lower levels predictive of a worse outcome. These data suggest that histone acetylation levels might be a valuable biomarker to predict tumour relapse. Histone acetylation has also been implicated in the development of human non-small cell lung carcinoma (NSCLC)173,174. Specifically, aberrant histone acetylation is reported to play an important role in EMT and metastasis in lung cancer173. More recently, Mi et al.174 also identified an acetyllysine-binding protein, YEATS2, that is amplified in NSCLC and binds to acetylated histone H3 (H3K27Ac) to promote a transcriptional program required for cancer cell growth and survival. Further study is needed to understand how global histone acetylation levels impact tumour growth and progression, as well as response to treatment.

ACL and acetyl-CoA levels and histone acetylation have also been shown to play an important role in the DNA damage response in tumour cells175. ACL is phosphorylated downstream of Akt activation following DNA damage, and this facilitates histone acetylation at double-strand break (DSB) sites, enabling breast cancer early onset 1 (BRCA1) recruitment and DNA repair by homologous recombination. ACL deficiency results in impaired BRCA1 recruitment to sites of DSBs and suppressed homologous recombination. The mechanisms of Akt activation by DNA damage signalling and the mechanisms that regulate levels of nuclear ACL, which is critical for its role in regulating BRCA1 recruitment, are less clear.

Therapeutic opportunities

The emerging roles of Krebs cycle intermediates in the pathophysiology of biomedically important processes raises the possibility of the generation of a new class of therapies focused on these metabolites, as shown in Fig. 7. This concept has already warranted investigation and was first illustrated following the discovery that succinate is accumulated during ischaemia and is then oxidised upon reperfusion to drive the mitochondrial superoxide formation that contributes to ischaemia–reperfusion (IR) injury33,174. Inhibiting the mitochondrial enzyme SDH during ischaemia by preloading tissues with the SDH inhibitor DMM prevented succinate accumulation during ischaemia and protected against tissue damage upon reperfusion33. Malonate itself was also protective when added at reperfusion176,177. This work indicates that simple inhibitors of Krebs cycle enzymes could be used to prevent pathological damage. The production of succinate is also enhanced during inflammation2, acting as a signal from mitochondria to the cytosol to activate pro-inflammatory gene expression while also generating ROS as a further pro-inflammatory redox signal36. As the mechanism of mitochondrial superoxide generation by succinate oxidation during inflammation seems to be the same as that during IR injury (RET)36, it is perhaps unsurprising that DMM could also decrease inflammation in a mouse model of sepsis36. Hence, simple molecules designed to inhibit SDH are a novel class of anti-inflammatory and anti–IR injury compounds.

Fig. 7: Therapeutic opportunities targeting metabolic signalling pathways.
Fig. 7

The SDH inhibitor malonate and the malonate prodrug DMM (hydrolysed to release malonate) ameliorate damage associated with succinate accumulation during ischaemia (1). SUCNR1 receptor inhibitors, such as compound 5g, are in development (2). Enasidinib, which targets mutated IDH2 to reduce 2-HG production, is leading to the treatment of AML (3). 4-OI, DI and DMF are anti-inflammatory and may have therapeutic potential (4). Inhibitors of α-KGDDs are an emerging target, but selectivity is an issue; for example, dimethyloxaloylglycine (DMOG) acts on TETs but also PHDs (5). Specific inhibition of KDMs has also been challenging, but EPT-103182 has shown promise in early pre-clinical work (6). The specific PHD inhibitor Vadadustat has shown promise in chronic kidney disease (7).

As well as acting within the cell, a portion of the succinate that accumulates during ischaemia and inflammation is released from the cell into the extracellular environment through poorly defined processes. This occurs following IR injury178,179, inflammation16 or cold exposure14. The diverse consequences of released succinate remains to be explored, but one likely role is as an inflammatory signal, acting via SUCNR1 (refs. 5,9,180,181). There is considerable interest in developing drugs that can counteract the effects of succinate on this receptor180. Although SUCNR1 receptor inhibitors, such as compound ‘5g’, are in their infancy, they may provide further understanding as to the signalling role of succinate and a method of blunting a cascade response during inflammation. Circulating Krebs cycle intermediates are particularly interesting potential drug targets.

As previously discussed, mitochondrial metabolites have an impact on metabolism following their export from the mitochondrial matrix to the cytosol and nucleus, and α-KGDDs are major targets for the action of mitochondrial metabolites, such as succinate and 2-HG7,8,180,182. High levels of succinate or 2-HG inhibit α-KGDDs. These include the PHDs, the TET DNA demethylases and the JMJDs109,183,184. Inhibitors of α-KGDDs are often non-specific, either targeting multiple α-KGDD family members, such as dimethyloxaloylglycine (DMOG) acting on not only TETs but also PHDs185. Specific inhibition of histone lysine demethylases (KDMs) has also been challenging186. EPT-103182 has shown promise in early pre-clinical work in an osteosarcoma cell line; however, it has yet to be investigated in clinical trials187. However, specific PHD inhibitors have been developed, such as Vadadustat, and these have shown promise in chronic kidney disease, increasing erythropoiesis and providing a means of attenuating HIF in other pathologies188. The development of more effective drugs that affect the interactions of Krebs cycle intermediates with α-KGDDs is an area of considerable current interest189.

A further therapeutic angle has revealed following the discovery of mutations in certain Krebs cycle enzymes that result in the generation of novel metabolites—the best example being mIDH2. Interestingly, the drug Enasidenib has been developed for AML, which binds to the mutated, but not the wild-type, form of IDH2 to prevent d-2-HG production190.

The generation of itaconate from the Krebs cycle intermediate cis-aconitate contrasts with most other metabolite signals emanating from mitochondria in that it primarily acts as a protective, anti-inflammatory metabolite63,67,68. As itaconate acts as an endogenous protective and anti-inflammatory agent, there has been interest in developing drugs that enhance its levels within cells. 4-OI is an itaconate mimic that was protective in a mouse model of sepsis, decreasing cytokine production and increasing survival68. The therapeutic potential of DI was also demonstrated, as in vivo administration of DI resulted in an amelioration of skin pathology in a mouse model of psoriasis, a disease also currently treated by the related electrophile DMF67,68,73. Thus, the development of new therapies designed to fine-tune the cell delivery and activity of NRF2 activators is a promising therapeutic strategy.

In summary, the emerging role of Krebs cycle intermediates in many central biomedical pathways opens up the possibility of extending mitochondrial pharmacology182,191,192,193 to manipulating the levels and impacts of mitochondrial metabolites both within the matrix and the rest of the cell, or following their release from the cell into the circulation.

Outstanding questions and concluding remarks

While the understanding of the role of metabolites as signalling molecules has come a long way, it is still in its infancy. The clear roles for succinate and HIF as pro-inflammatory signals, l-2-HG in driving anti-tumour immunity, fumarate as an anti-inflammatory and pro-tumorigenic signal and acetyl-CoA as an epigenetic regulator reveal some of the potential targets uncovered in this burgeoning field, yet many unanswered questions remain. One critical point to consider is the limited human and in vivo data that are available at present. These studies prove especially challenging owing to the rapid rate at which steady-state metabolism changes. Appropriate methods of euthanasia and tissue extrusion are paramount to obtain an accurate snapshot of the in vivo metabolic status of cells and tissues, making relevant and comparable analyses extremely challenging. Also, studies in human immune cells are currently very limited. Understanding the cellular location of metabolite accumulation as well as how these cells may signal in a paracrine manner remain to be explored. For example, the relevant ratios of the different metabolites, which may vary over time and in different compartments of the cell, will be informative. It is also likely that unidentified receptors and transporters exist, which are capable of recognising metabolites to initiate downstream signalling events or to facilitate cell entry from the extracellular milieu, respectively. Whether metabolites such as succinate, itaconate and acetyl-CoA serve as biomarkers or drivers of disease or tumour progression remains to be further explored. Of note, some of the major findings on the function of oncometabolites were obtained from rare cancer syndromes. However, many tumours appear to maintain a functional TCA cycle, as observed in human lung and brain tumours in vivo194,195,196,197, while in some instances ETC dysfunction can act as a barrier to tumorigenesis198. As such, the intriguing role of these metabolites in driving rare cancer syndromes may represent an exception to the rule for tumorigenesis (if such a rule exists), and it will be crucial to understand whether some of the processes elicited by oncometabolites occur in other sporadic cancers and in the absence of inactivating mutations of FH, IDH, or SDH.

Interestingly, many of the metabolites discussed share converging molecular functions (including activation of HIF and NRF2 and alterations in chromatin structure), but we still do not know how specific outcomes emerge from these ‘unspecific’ functions in different cell types. It is possible that the function of these metabolites is highly context-dependent and the inherent chromatin landscape and biochemical set-up of a cell might be a key determinant of the response to the accumulation of these metabolites. On this note, it possible that the often-opposing roles of metabolites in the context of tumour progression and immune cell activation may be detrimental to the host. For example, succinate, which promotes a potentially tumour-fighting, pro-inflammatory state also drives HIF-1α activation in tumours, which is associated with cancer progression2,26. On the flip side, it would be interesting to explore if pathogens exploit or manipulate immunometabolism of the host for their own need or indeed if they produce metabolites that can either acts as pathogen- and damage-associated molecular patterns and/or impair immune cell function. It is likely that the field is only at the tip of the iceberg in terms of the breadth of both metabolites with signalling capacity and indeed the entire metabolome152. This field still remains in its infancy compared to proteomic and transcriptomic analysis, but liquid chromatography approaches to detect metabolites are constantly evolving, enabling both the detection of a broader range of metabolites and more unbiased analyses, which are likely to uncover novel metabolic signals. In the future, exciting new insights into how the Krebs cycle connects with signalling and perhaps how this cycle may serve as the ultimate determinant of cell function in health and disease will emerge.

Additional information

Publisher’s note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.


  1. 1.

    Jha, A. K. et al. Network integration of parallel metabolic and transcriptional data reveals metabolic modules that regulate macrophage polarization. Immunity 42, 419–430 (2015).

  2. 2.

    Tannahill, G. M. et al. Succinate is an inflammatory signal that induces IL-1β through HIF-1α. Nature 496, 238–242 (2013).

  3. 3.

    He, W. et al. Citric acid cycle intermediates as ligands for orphan G-protein-coupled receptors. Nature 429, 188–193 (2004).

  4. 4.

    Chouchani, E. T. et al. A unifying mechanism for mitochondrial superoxide production during ischemia–reperfusion injury. Cell. Metab. 23, 254–263 (2016).

  5. 5.

    Rubic, T. et al. Triggering the succinate receptor GPR91 on dendritic cells enhances immunity. Nat. Immunol. 9, 1261–1269 (2008).

  6. 6.

    Lei, W. et al. Activation of intestinal tuft cell-expressed Sucnr1 triggers type 2 immunity in the mouse small intestine. Proc. Natl Acad. Sci. USA. 115, 5552–5557 (2018).

  7. 7.

    Nadjsombati, M. S. et al. Detection of succinate by intestinal tuft cells triggers a type 2 innate immune circuit. Immunity 49, 33–41.e7 (2018).

  8. 8.

    Schneider, C. et al. A metabolite-triggered tuft cell–ILC2 circuit drives small intestinal remodeling. Cell 174, 271–284.e14 (2018).

  9. 9.

    Littlewood-Evans, A. et al. GPR91 senses extracellular succinate released from inflammatory macrophages and exacerbates rheumatoid arthritis. J. Exp. Med. 213, 1655–1662 (2016).

  10. 10.

    Kim, S. et al. Global metabolite profiling of synovial fluid for the specific diagnosis of rheumatoid arthritis from other inflammatory arthritis. PLoS One 9, e97501 (2014).

  11. 11.

    Hollander, A. P., Corke, K. P., Freemont, A. J. & Lewis, C. E. Expression of hypoxia-inducible factor 1α by macrophages in the rheumatoid synovium: implications for targeting of therapeutic genes to the inflamed joint. Arthritis Rheum. 44, 1540–1544 (2001).

  12. 12.

    Sadagopan, N. et al. Circulating succinate is elevated in rodent models of hypertension and metabolic disease. Am. J. Hypertens. 20, 1209–1215 (2007).

  13. 13.

    van Diepen, J. A. et al. SUCNR1-mediated chemotaxis of macrophages aggravates obesity-induced inflammation and diabetes. Diabetologia 60, 1304–1313 (2017).

  14. 14.

    Mills, E. L. et al. Accumulation of succinate controls activation of adipose tissue thermogenesis. Nature 560, 102–106 (2018).

  15. 15.

    Lewis, G. D. et al. Metabolic signatures of exercise in human plasma. Sci. Transl. Med. 2, 33ra37 (2010).

  16. 16.

    Peruzzotti-Jametti, L. et al. Macrophage-derived extracellular succinate licenses neural stem cells to suppress chronic neuroinflammation. Cell Stem Cell 22, 355–368.e13 (2018).

  17. 17.

    Bhuniya, D. et al. Discovery of a potent and selective small molecule hGPR91 antagonist. Bioorg. Med. Chem. Lett. 21, 3596–3602 (2011).

  18. 18.

    Geubelle, P. et al. Identification and pharmacological characterization of succinate receptor agonists. Br. J. Pharmacol. 174, 796–808 (2017).

  19. 19.

    Tannahill, G. M. & O’Neill, L. A. The emerging role of metabolic regulation in the functioning of Toll-like receptors and the NOD-like receptor Nlrp3. FEBS Lett. 585, 1568–1572 (2011).

  20. 20.

    Semenza, G. L. & Wang, G. L. A nuclear factor induced by hypoxia via de novo protein synthesis binds to the human erythropoietin gene enhancer at a site required for transcriptional activation. Mol. Cell. Biol. 12, 5447–5454 (1992).

  21. 21.

    Semenza, G. L. Regulation of oxygen homeostasis by hypoxia-inducible factor 1. Physiology (Bethesda) 24, 97–106 (2009).

  22. 22.

    Epstein, A. C. R. et al. C. elegans EGL-9 and mammalian homologs define a family of dioxygenases that regulate HIF by prolyl hydroxylation. Cell 107, 43–54 (2001).

  23. 23.

    Maxwell, P. H. et al. The tumour suppressor protein VHL targets hypoxia-inducible factors for oxygen-dependent proteolysis. Nature 399, 271–275 (1999).

  24. 24.

    Pearce, E. L. & Pearce, E. J. Metabolic pathways in immune cell activation and quiescence. Immunity 38, 633–643 (2013).

  25. 25.

    Frede, S., Stockmann, C., Freitag, P. & Fandrey, J. Bacterial lipopolysaccharide induces HIF-1 activation in human monocytes via p44/42 MAPK and NF-κB. Biochem. J. 396, 517–527 (2006).

  26. 26.

    Wu, W. & Zhao, S. Metabolic changes in cancer: beyond the Warburg effect. Acta Biochim. Biophys. Sin. (Shanghai) 45, 18–26 (2013).

  27. 27.

    Selak, M. A. et al. Succinate links TCA cycle dysfunction to oncogenesis by inhibiting HIF-α prolyl hydroxylase. Cancer Cell. 7, 77–85 (2005).

  28. 28.

    Rasola, A., Neckers, L. & Picard, D. Mitochondrial oxidative phosphorylation TRAP(1)ped in tumor cells. Trends Cell Biol. 24, 455–463 (2014).

  29. 29.

    Mu, X. et al. Oncometabolite succinate promotes angiogenesis by upregulating VEGF expression through GPR91-mediated STAT3 and ERK activation. Oncotarget 8, 13174–13185 (2017).

  30. 30.

    Guzy, R. D., Sharma, B., Bell, E., Chandel, N. S. & Schumacker, P. T. Loss of the SdhB, but not the SdhA, subunit of complex II triggers reactive oxygen species-dependent hypoxia-inducible factor activation and tumorigenesis. Mol. Cell. Biol. 28, 718–731 (2008).

  31. 31.

    Hagen, T. Oxygen versus reactive oxygen in the regulation of HIF-1α: the balance tips. Biochem. Res. Int. 2012, 436981 (2012).

  32. 32.

    Folbergrová, J., Ljunggren, B., Norberg, K. & Siesjö, B. K. Influence of complete ischemia on glycolytic metabolites, citric acid cycle intermediates, and associated amino acids in the rat cerebral cortex. Brain Res. 80, 265–279 (1974).

  33. 33.

    Chouchani, E. T. et al. Ischaemic accumulation of succinate controls reperfusion injury through mitochondrial ROS. Nature 515, 431–435 (2014).

  34. 34.

    Zhang, W. et al. Evidence that hypoxia-inducible factor-1 (HIF-1) mediates transcriptional activation of interleukin-1β (IL-1β) in astrocyte cultures. J. Neuroimmunol. 174, 63–73 (2006).

  35. 35.

    Peyssonnaux, C. et al. Cutting edge: essential role of hypoxia inducible factor-1α in development of lipopolysaccharide-induced sepsis. J. Immunol. 178, 7516–7519 (2007).

  36. 36.

    Mills, E. L. et al. Succinate dehydrogenase supports metabolic repurposing of mitochondria to drive inflammatory macrophages. Cell 167, 457–470.e13 (2016).

  37. 37.

    El-Khoury, R. et al. Engineering the alternative oxidase gene to better understand and counteract mitochondrial defects: state of the art and perspectives. Br. J. Pharmacol. 171, 2243–2249 (2014).

  38. 38.

    Jin, Z., Wei, W., Yang, M., Du, Y. & Wan, Y. Mitochondrial complex I activity suppresses inflammation and enhances bone resorption by shifting macrophage-osteoclast polarization. Cell. Metab. 20, 483–498 (2014).

  39. 39.

    Murphy, M. P. How mitochondria produce reactive oxygen species. Biochem. J. 417, 1–13 (2009).

  40. 40.

    Bénit, P. et al. Unsuspected task for an old team: succinate, fumarate and other Krebs cycle acids in metabolic remodeling. Biochim. Biophys. Acta 1837, 1330–1337 (2014).

  41. 41.

    Brigati, C. et al. Inflammation, HIF-1, and the epigenetics that follows. Mediators Inflamm. 2010, 263914 (2010).

  42. 42.

    Benn, C. S., Netea, M. G., Selin, L. K. & Aaby, P. A small jab—a big effect: nonspecific immunomodulation by vaccines. Trends Immunol. 34, 431–439 (2013).

  43. 43.

    Saeed, S. et al. Epigenetic programming of monocyte-to-macrophage differentiation and trained innate immunity. Science 345, 1251086 (2014).

  44. 44.

    Cheng, S. C. et al. mTOR- and HIF-1α-mediated aerobic glycolysis as metabolic basis for trained immunity. Science 345, 1250684 (2014).

  45. 45.

    Park, J. et al. SIRT5-mediated lysine desuccinylation impacts diverse metabolic pathways. Mol. Cell 50, 919–930 (2013).

  46. 46.

    Zhang, Z. et al. Identification of lysine succinylation as a new post-translational modification. Nat. Chem. Biol. 7, 58–63 (2011).

  47. 47.

    Xie, L. et al. First succinyl-proteome profiling of extensively drug-resistant Mycobacterium tuberculosis revealed involvement of succinylation in cellular physiology. J. Proteome Res. 14, 107–119 (2015).

  48. 48.

    Feng, S. et al. Succinyl-proteome profiling of Dendrobium officinale, an important traditional Chinese orchid herb, revealed involvement of succinylation in the glycolysis pathway. BMC Genomics 18, 598 (2017).

  49. 49.

    He, D. et al. Global proteome analyses of lysine acetylation and succinylation reveal the widespread involvement of both modification in metabolism in the embryo of germinating rice seed. J. Proteome Res. 15, 879–890 (2016).

  50. 50.

    Xie, Z. et al. Lysine succinylation and lysine malonylation in histones. Mol. Cell. Proteomics 11, 100–107 (2012).

  51. 51.

    Du, J. et al. Sirt5 is a NAD-dependent protein lysine demalonylase and desuccinylase. Science 334, 806–809 (2011).

  52. 52.

    Wang, Y. et al. KAT2A coupled with the α-KGDH complex acts as a histone H3 succinyltransferase. Nature 552, 273–277 (2017).

  53. 53.

    Gibson, G. E. et al. α-ketoglutarate dehydrogenase complex–dependent succinylation of proteins in neurons and neuronal cell lines. J. Neurochem. 134, 86–96 (2015).

  54. 54.

    Wang, F. et al. SIRT5 desuccinylates and activates pyruvate kinase M2 to block macrophage IL-1β production and to prevent dss-induced colitis in mice. Cell Rep. 19, 2331–2344 (2017).

  55. 55.

    Cordes, T. et al. Immunoresponsive gene 1 and itaconate inhibit succinate dehydrogenase to modulate intracellular succinate levels. J. Biol. Chem. 291, 14274–14284 (2016).

  56. 56.

    Lampropoulou, V. et al. Itaconate links inhibition of succinate dehydrogenase with macrophage metabolic remodeling and regulation of inflammation. Cell Metab. 24, 158–166 (2016).

  57. 57.

    Meiser, J. et al. Pro-inflammatory macrophages sustain pyruvate oxidation through pyruvate dehydrogenase for the synthesis of itaconate and to enable cytokine expression. J. Biol. Chem. 291, 3932–3946 (2016).

  58. 58.

    Németh, B. et al. Abolition of mitochondrial substrate-level phosphorylation by itaconic acid produced by LPS-induced Irg1 expression in cells of murine macrophage lineage. FASEB J. 30, 286–300 (2016).

  59. 59.

    Okabe, M., Lies, D., Kanamasa, S. & Park, E. Y. Biotechnological production of itaconic acid and its biosynthesis in Aspergillus terreus. Appl. Microbiol. Biotechnol. 84, 597–606 (2009).

  60. 60.

    Lee, C. G., Jenkins, N. A., Gilbert, D. J., Copeland, N. G. & O’Brien, W. E. Cloning and analysis of gene regulation of a novel LPS-inducible cDNA. Immunogenetics 41, 263–270 (1995).

  61. 61.

    Shin, J. H. et al. 1H NMR-based metabolomic profiling in mice infected with Mycobacterium tuberculosis. J. Proteome Res. 10, 2238–2247 (2011).

  62. 62.

    Sugimoto, M. S. H. Y. et al. Non-targeted metabolite profiling in activated macrophage secretion. Metabolomics 8, 624–633 (2011).

  63. 63.

    Michelucci, A. et al. Immune-responsive gene 1 protein links metabolism to immunity by catalyzing itaconic acid production. Proc. Natl. Acad. Sci. USA. 110, 7820–7825 (2013).

  64. 64.

    McFadden, B. A. & Purohit, S. Itaconate, an isocitrate lyase-directed inhibitor in Pseudomonas indigofera. J. Bacteriol. 131, 136–144 (1977).

  65. 65.

    Naujoks, J. et al. IFNs modify the proteome of Legionella-containing vacuoles and restrict infection via IRG1-derived itaconic acid. PLoS Pathog. 12, e1005408 (2016).

  66. 66.

    Williams, J. O., Roche, T. E. & McFadden, B. A. Mechanism of action of isocitrate lyase from Pseudomonas indigofera. Biochemistry 10, 1384–1390 (1971).

  67. 67.

    Bambouskova, M. et al. Electrophilic properties of itaconate and derivatives regulate the IκBξ–ATF3 inflammatory axis. Nature 556, 501–504 (2018).

  68. 68.

    Mills, E. L. et al. Itaconate is an anti-inflammatory metabolite that activates Nrf2 via alkylation of KEAP1. Nature 556, 113–117 (2018).

  69. 69.

    Nair, S. et al. Irg1 expression in myeloid cells prevents immunopathology during M. tuberculosis infection. J. Exp. Med. 215, 1035–1045 (2018).

  70. 70.

    Ackermann, W. W. & Potter, V. R. Enzyme inhibition in relation to chemotherapy. Proc. Soc. Exp. Biol. Med. 72, 1–9 (1949).

  71. 71.

    El Azzouny, M. et al. Dimethyl itaconate is not metabolized into itaconate intracellularly. J. Biol. Chem. 292, 4766–4769 (2017).

  72. 72.

    Kobayashi, E. H. et al. Nrf2 suppresses macrophage inflammatory response by blocking proinflammatory cytokine transcription. Nat. Commun. 7, 11624 (2016).

  73. 73.

    Brennan, M. S. et al. Dimethyl fumarate and monoethyl fumarate exhibit differential effects on KEAP1, NRF2 activation, and glutathione depletion in vitro. PLoS One 10, e0120254 (2015).

  74. 74.

    Saito, R. et al. Characterizations of three major cysteine sensors of Keap1 in stress response. Mol. Cell Biol. 36, 271–284 (2015).

  75. 75.

    Li, Y. et al. Immune responsive gene 1 (IRG1) promotes endotoxin tolerance by increasing A20 expression in macrophages through reactive oxygen species. J. Biol. Chem. 288, 16225–16234 (2013).

  76. 76.

    Van Quickelberghe, E. et al. Identification of immune-responsive gene 1 (IRG1) as a target of A20. J. Proteome Res. 17, 2182–2191 (2018).

  77. 77.

    Jamal Uddin, M. et al. IRG1 induced by heme oxygenase-1/carbon monoxide inhibits LPS-mediated sepsis and pro-inflammatory cytokine production. Cell Mol. Immunol. 13, 170–179 (2016).

  78. 78.

    Cheon, Y. P., Xu, X., Bagchi, M. K. & Bagchi, I. C. Immune-responsive gene 1 is a novel target of progesterone receptor and plays a critical role during implantation in the mouse. Endocrinology 144, 5623–5630 (2003).

  79. 79.

    Luan, H. H. & Medzhitov, R. Food fight: role of itaconate and other metabolites in antimicrobial defense. Cell. Metab. 24, 379–387 (2016).

  80. 80.

    Sherwin, J. R. et al. Identification of genes regulated by leukemia-inhibitory factor in the mouse uterus at the time of implantation. Mol. Endocrinol. 18, 2185–2195 (2004).

  81. 81.

    Weiss, J. M. et al. Itaconic acid mediates crosstalk between macrophage metabolism and peritoneal tumors. J. Clin. Invest. 128, 3794–3805 (2018).

  82. 82.

    Hall, C. J. et al. Immunoresponsive gene 1 augments bactericidal activity of macrophage-lineage cells by regulating β-oxidation-dependent mitochondrial ROS production. Cell. Metab. 18, 265–278 (2013).

  83. 83.

    Shi, H. Z., Wang, D., Sun, X. N. & Sheng, L. MicroRNA-378 acts as a prognosis marker and inhibits cell migration, invasion and epithelial-mesenchymal transition in human glioma by targeting IRG1. Eur. Rev. Med. Pharmacol. Sci. 22, 3837–3846 (2018).

  84. 84.

    Pan, J. et al. Immune responsive gene 1, a novel oncogene, increases the growth and tumorigenicity of glioma. Oncol. Rep. 32, 1957–1966 (2014).

  85. 85.

    Frezza, C. Mitochondrial metabolites: undercover signalling molecules. Interface Focus 7, 20160100 (2017).

  86. 86.

    Arts, R. J. et al. Glutaminolysis and fumarate accumulation integrate immunometabolic and epigenetic programs in trained immunity. Cell. Metab. 24, 807–819 (2016).

  87. 87.

    Tomlinson, I. P. et al. Germline mutations in FH predispose to dominantly inherited uterine fibroids, skin leiomyomata and papillary renal cell cancer. Nat. Genet. 30, 406–410 (2002).

  88. 88.

    Xiao, M. et al. Inhibition of α-KG-dependent histone and DNA demethylases by fumarate and succinate that are accumulated in mutations of FH and SDH tumor suppressors. Genes Dev. 26, 1326–1338 (2012).

  89. 89.

    Delage, B. et al. Arginine deprivation and argininosuccinate synthetase expression in the treatment of cancer. Int. J. Cancer 126, 2762–2772 (2010).

  90. 90.

    Blewett, M. M. et al. Chemical proteomic map of dimethyl fumarate-sensitive cysteines in primary human T cells. Sci. Signal. 9, rs10 (2016).

  91. 91.

    Angiari, S. & O’Neill, L. A. Dimethyl fumarate: targeting glycolysis to treat MS. Cell Res. 28, 613–615 (2018).

  92. 92.

    Kornberg, M. D. et al. Dimethyl fumarate targets GAPDH and aerobic glycolysis to modulate immunity. Science 360, 449–453 (2018).

  93. 93.

    Brück, J., Dringen, R., Amasuno, A., Pau-Charles, I. & Ghoreschi, K. A review of the mechanisms of action of dimethylfumarate in the treatment of psoriasis. Exp. Dermatol. 27, 611–624 (2018).

  94. 94.

    Mills, E. A., Ogrodnik, M. A., Plave, A. & Mao-Draayer, Y. Emerging understanding of the mechanism of action for dimethyl fumarate in the treatment of multiple sclerosis. Front. Neurol. 9, 5 (2018).

  95. 95.

    Zheng, L. et al. Fumarate induces redox-dependent senescence by modifying glutathione metabolism. Nat. Commun. 6, 6001 (2015).

  96. 96.

    Sciacovelli, M. & Frezza, C. Oncometabolites: unconventional triggers of oncogenic signalling cascades. Free Radic. Biol. Med. 100, 175–181 (2016).

  97. 97.

    Rustin, P. et al. Inborn errors of the Krebs cycle: a group of unusual mitochondrial diseases in human. Biochim. Biophys. Acta 1361, 185–197 (1997).

  98. 98.

    Allegri, G. et al. Fumaric aciduria: an overview and the first Brazilian case report. J. Inherit. Metab. Dis. 33, 411–419 (2010).

  99. 99.

    Blatnik, M., Thorpe, S. R. & Baynes, J. W. Succination of proteins by fumarate: mechanism of inactivation of glyceraldehyde-3-phosphate dehydrogenase in diabetes. Ann. NY Acad. Sci. 1126, 272–275 (2008).

  100. 100.

    Adam, J. et al. Renal cyst formation in Fh1-deficient mice is independent of the Hif/Phd pathway: roles for fumarate in KEAP1 succination and Nrf2 signaling. Cancer Cell 20, 524–537 (2011).

  101. 101.

    Ooi, A. et al. An antioxidant response phenotype shared between hereditary and sporadic type 2 papillary renal cell carcinoma. Cancer Cell 20, 511–523 (2011).

  102. 102.

    Sullivan, L. B. et al. The proto-oncometabolite fumarate binds glutathione to amplify ROS-dependent signaling. Mol. Cell 51, 236–248 (2013).

  103. 103.

    Linehan, W. M. et al. Comprehensive molecular characterization of papillary renal-cell carcinoma. N. Engl. J. Med. 374, 135–145 (2016).

  104. 104.

    Ternette, N. et al. Inhibition of mitochondrial aconitase by succination in fumarate hydratase deficiency. Cell Rep. 3, 689–700 (2013).

  105. 105.

    Kerins, M. J. et al. Fumarate mediates a chronic proliferative signal in fumarate hydratase-inactivated cancer cells by increasing transcription and translation of ferritin genes. Mol. Cell Biol. 37, e00079-17 (2017).

  106. 106.

    Tyrakis, P. A. et al. Fumarate hydratase loss causes combined respiratory chain defects. Cell Rep. 21, 1036–1047 (2017).

  107. 107.

    Bardella, C. et al. Aberrant succination of proteins in fumarate hydratase-deficient mice and HLRCC patients is a robust biomarker of mutation status. J. Pathol. 225, 4–11 (2011).

  108. 108.

    Laukka, T. et al. Fumarate and succinate regulate expression of hypoxia-inducible genes via TET enzymes. J. Biol. Chem. 291, 4256–4265 (2016).

  109. 109.

    Salminen, A., Kauppinen, A. & Kaarniranta, K. 2-Oxoglutarate-dependent dioxygenases are sensors of energy metabolism, oxygen availability, and iron homeostasis: potential role in the regulation of aging process. Cell. Mol. Life Sci. 72, 3897–3914 (2015).

  110. 110.

    Isaacs, J. S. et al. HIF overexpression correlates with biallelic loss of fumarate hydratase in renal cancer: novel role of fumarate in regulation of HIF stability. Cancer Cell. 8, 143–153 (2005).

  111. 111.

    Yang, M., Soga, T., Pollard, P. J. & Adam, J. The emerging role of fumarate as an oncometabolite. Front. Oncol. 2, 85 (2012).

  112. 112.

    Sciacovelli, M. et al. Fumarate is an epigenetic modifier that elicits epithelial-to-mesenchymal transition. Nature 537, 544–547 (2016).

  113. 113.

    Brabletz, T., Kalluri, R., Nieto, M. A. & Weinberg, R. A. EMT in cancer. Nat. Rev. Cancer 18, 128–134 (2018).

  114. 114.

    Brat, D. J. et al. Comprehensive, integrative genomic analysis of diffuse lower-grade gliomas. N. Engl. J. Med. 372, 2481–2498 (2015).

  115. 115.

    Yogev, O. et al. Fumarase: a mitochondrial metabolic enzyme and a cytosolic/nuclear component of the DNA damage response. PLoS Biol. 8, e1000328 (2010).

  116. 116.

    Jiang, Y. et al. Local generation of fumarate promotes DNA repair through inhibition of histone H3 demethylation. Nat. Cell Biol. 17, 1158–1168 (2015).

  117. 117.

    Sulkowski, P. L. et al. Krebs-cycle-deficient hereditary cancer syndromes are defined by defects in homologous-recombination DNA repair. Nat. Genet. 50, 1086–1092 (2018).

  118. 118.

    Johnson, T. I., Costa, A. S. H., Ferguson, A. N. & Frezza, C. Fumarate hydratase loss promotes mitotic entry in the presence of DNA damage after ionising radiation. Cell Death Dis. 9, 913 (2018).

  119. 119.

    Ye, D., Guan, K. L. & Xiong, Y. Metabolism, activity, and targeting of d- and l-2-Hydroxyglutarates. Trends Cancer 4, 151–165 (2018).

  120. 120.

    Chalmers, R. A. et al. d-2-hydroxyglutaric aciduria: case report and biochemical studies. J. Inherit. Metab. Dis. 3, 11–15 (1980).

  121. 121.

    Duran, M., Kamerling, J. P., Bakker, H. D., van Gennip, A. H. & Wadman, S. K. l-2-Hydroxyglutaric aciduria: an inborn error of metabolism? J. Inherit. Metab. Dis. 3, 109–112 (1980).

  122. 122.

    Kranendijk, M., Struys, E. A., Salomons, G. S., Van der Knaap, M. S. & Jakobs, C. Progress in understanding 2-hydroxyglutaric acidurias. J. Inherit. Metab. Dis. 35, 571–587 (2012).

  123. 123.

    Rzem, R. et al. A gene encoding a putative FAD-dependent l-2-hydroxyglutarate dehydrogenase is mutated in l-2-hydroxyglutaric aciduria. Proc. Natl. Acad. Sci. USA 101, 16849–16854 (2004).

  124. 124.

    Struys, E. A. et al. Mutations in the d-2-hydroxyglutarate dehydrogenase gene cause D-2-hydroxyglutaric aciduria. Am. J. Hum. Genet. 76, 358–360 (2005).

  125. 125.

    Topçu, M. et al. l-2-Hydroxyglutaric aciduria: identification of a mutant gene C14orf160, localized on chromosome 14q22.1. Hum. Mol. Genet. 13, 2803–2811 (2004).

  126. 126.

    Kranendijk, M. et al. IDH2 mutations in patients with d-2-hydroxyglutaric aciduria. Science 330, 336 (2010).

  127. 127.

    Nota, B. et al. Deficiency in SLC25A1, encoding the mitochondrial citrate carrier, causes combined d-2- and l-2-hydroxyglutaric aciduria. Am. J. Hum. Genet. 92, 627–631 (2013).

  128. 128.

    Dang, L. & Su, S. M. Isocitrate dehydrogenase mutation and R-2-hydroxyglutarate: from basic discovery to therapeutics development. Annu. Rev. Biochem. 86, 305–331 (2017).

  129. 129.

    Mardis, E. R. et al. Recurring mutations found by sequencing an acute myeloid leukemia genome. N. Engl. J. Med. 361, 1058–1066 (2009).

  130. 130.

    Parsons, D. W. et al. An integrated genomic analysis of human glioblastoma multiforme. Science 321, 1807–1812 (2008).

  131. 131.

    MacIver, N. J., Michalek, R. D. & Rathmell, J. C. Metabolic regulation of T lymphocytes. Annu. Rev. Immunol. 31, 259–283 (2013).

  132. 132.

    Dang, E. V. et al. Control of TH17/Treg balance by hypoxia-inducible factor 1. Cell 146, 772–784 (2011).

  133. 133.

    Xu, T. et al. Metabolic control of TH17 and induced Treg cell balance by an epigenetic mechanism. Nature 548, 228–233 (2017).

  134. 134.

    Figueroa, M. E. et al. Leukemic IDH1 and IDH2 mutations result in a hypermethylation phenotype, disrupt TET2 function, and impair hematopoietic differentiation. Cancer Cell. 18, 553–567 (2010).

  135. 135.

    Losman, J. A. et al. R-2-hydroxyglutarate is sufficient to promote leukemogenesis and its effects are reversible. Science 339, 1621–1625 (2013).

  136. 136.

    Tyrakis, P. A. et al. S-2-hydroxyglutarate regulates CD8+ T-lymphocyte fate. Nature 540, 236–241 (2016).

  137. 137.

    Doedens, A. L. et al. Hypoxia-inducible factors enhance the effector responses of CD8+ T cells to persistent antigen. Nat. Immunol. 14, 1173–1182 (2013).

  138. 138.

    Finlay, D. K. et al. PDK1 regulation of mTOR and hypoxia-inducible factor 1 integrate metabolism and migration of CD8+ T cells. J. Exp. Med. 209, 2441–2453 (2012).

  139. 139.

    Suzuki, H. et al. Mutational landscape and clonal architecture in grade II and III gliomas. Nat. Genet. 47, 458–468 (2015).

  140. 140.

    Watanabe, T., Nobusawa, S., Kleihues, P. & Ohgaki, H. IDH1 mutations are early events in the development of astrocytomas and oligodendrogliomas. Am. J. Pathol. 174, 1149–1153 (2009).

  141. 141.

    Dang, L. et al. Cancer-associated IDH1 mutations produce 2-hydroxyglutarate. Nature 462, 739–744 (2009).

  142. 142.

    Ward, P. S. et al. The common feature of leukemia-associated IDH1 and IDH2 mutations is a neomorphic enzyme activity converting α-ketoglutarate to 2-hydroxyglutarate. Cancer Cell 17, 225–234 (2010).

  143. 143.

    Yan, H. et al. IDH1 and IDH2 mutations in gliomas. N. Engl. J. Med. 360, 765–773 (2009).

  144. 144.

    Chowdhury, R. et al. The oncometabolite 2-hydroxyglutarate inhibits histone lysine demethylases. EMBO Rep. 12, 463 (2011).

  145. 145.

    Xu, W. et al. Oncometabolite 2-hydroxyglutarate is a competitive inhibitor of α-ketoglutarate-dependent dioxygenases. Cancer Cell. 19, 17–30 (2011).

  146. 146.

    Noushmehr, H. et al. Identification of a CpG island methylator phenotype that defines a distinct subgroup of glioma. Cancer Cell. 17, 510–522 (2010).

  147. 147.

    Turcan, S. et al. IDH1 mutation is sufficient to establish the glioma hypermethylator phenotype. Nature 483, 479–483 (2012).

  148. 148.

    Wang, P. et al. Mutations in isocitrate dehydrogenase 1 and 2 occur frequently in intrahepatic cholangiocarcinomas and share hypermethylation targets with glioblastomas. Oncogene 32, 3091–3100 (2013).

  149. 149.

    Flavahan, W. A. et al. Insulator dysfunction and oncogene activation in IDH mutant gliomas. Nature 529, 110–114 (2016).

  150. 150.

    Zhao, S. et al. Glioma-derived mutations in IDH1 dominantly inhibit IDH1 catalytic activity and induce HIF-1α. Science 324, 261–265 (2009).

  151. 151.

    Intlekofer, A. M. et al. Hypoxia induces production of l-2-hydroxyglutarate. Cell Metab. 22, 304–311 (2015).

  152. 152.

    Intlekofer, A. M. et al. l-2-Hydroxyglutarate production arises from noncanonical enzyme function at acidic pH. Nat. Chem. Biol. 13, 494–500 (2017).

  153. 153.

    Nadtochiy, S. M. et al. Acidic pH Is a metabolic switch for 2-hydroxyglutarate generation and signaling. J. Biol. Chem. 291, 20188–20197 (2016).

  154. 154.

    Sasaki, M. et al. d-2-hydroxyglutarate produced by mutant IDH1 perturbs collagen maturation and basement membrane function. Genes Dev. 26, 2038–2049 (2012).

  155. 155.

    Tarhonskaya, H. et al. Non-enzymatic chemistry enables 2-hydroxyglutarate-mediated activation of 2-oxoglutarate oxygenases. Nat. Commun. 5, 3423 (2014).

  156. 156.

    Wang, P. et al. Oncometabolite d-2-hydroxyglutarate inhibits ALKBH DNA repair enzymes and sensitizes IDH mutant cells to alkylating agents. Cell Rep. 13, 2353–2361 (2015).

  157. 157.

    Inoue, S. et al. Mutant IDH1 downregulates ATM and alters DNA repair and sensitivity to DNA damage independent of TET2. Cancer Cell 30, 337–348 (2016).

  158. 158.

    Carbonneau, M. et al. The oncometabolite 2-hydroxyglutarate activates the mTOR signalling pathway. Nat. Commun. 7, 12700 (2016).

  159. 159.

    Reitman, Z. J. et al. Profiling the effects of isocitrate dehydrogenase 1 and 2 mutations on the cellular metabolome. Proc. Natl. Acad. Sci. USA 108, 3270–3275 (2011).

  160. 160.

    Chesnelong, C. et al. Lactate dehydrogenase A silencing in IDH mutant gliomas. Neuro-Oncol. 16, 686–695 (2014).

  161. 161.

    Grassian, A. R. et al. IDH1 mutations alter citric acid cycle metabolism and increase dependence on oxidative mitochondrial metabolism. Cancer Res. 74, 3317–3331 (2014).

  162. 162.

    Izquierdo-Garcia, J. L. et al. IDH1 mutation induces reprogramming of pyruvate metabolism. Cancer Res. 75, 2999–3009 (2015).

  163. 163.

    Chen, J. Y. et al. The oncometabolite R-2-hydroxyglutarate activates NF-κB-dependent tumor-promoting stromal niche for acute myeloid leukemia cells. Sci. Rep. 6, 1 (2016).

  164. 164.

    Li, F. et al. NADP+–IDH mutations promote hypersuccinylation that impairs mitochondria respiration and induces apoptosis resistance. Mol. Cell 60, 661–675 (2015).

  165. 165.

    Jiang, B. et al. IDH1 mutation promotes tumorigenesis by inhibiting JNK activation and apoptosis induced by serum starvation. Cell Rep. 19, 389–400 (2017).

  166. 166.

    Yang, Z. et al. 2-HG inhibits necroptosis by stimulating DNMT1-dependent hypermethylation of the RIP3 promoter. Cell Rep. 19, 1846–1857 (2017).

  167. 167.

    Kurdistani, S. K. & Grunstein, M. Histone acetylation and deacetylation in yeast. Nat. Rev. Mol. Cell Biol. 4, 276–284 (2003).

  168. 168.

    Wellen, K. E. et al. ATP-citrate lyase links cellular metabolism to histone acetylation. Science 324, 1076–1080 (2009).

  169. 169.

    Peng, M. et al. Aerobic glycolysis promotes T helper 1 cell differentiation through an epigenetic mechanism. Science 354, 481–484 (2016).

  170. 170.

    Chang, C. H. et al. Posttranscriptional control of T cell effector function by aerobic glycolysis. Cell 153, 1239–1251 (2013).

  171. 171.

    Osinalde, N. et al. Nuclear phosphoproteomic screen uncovers ACLY as mediator of IL-2-induced proliferation of CD4+ T lymphocytes. Mol. Cell. Proteomics 15, 2076–2092 (2016).

  172. 172.

    Lee, J. V. et al. Akt-dependent metabolic reprogramming regulates tumor cell histone acetylation. Cell. Metab. 20, 306–319 (2014).

  173. 173.

    Zhang, L., Liu, Z., Ma, W. & Wang, B. The landscape of histone acetylation involved in epithelial-mesenchymal transition in lung cancer. J. Can. Res. Ther. 9, S86–S91 (2013).

  174. 174.

    Mi, W. et al. YEATS2 links histone acetylation to tumorigenesis of non-small cell lung cancer. Nat. Commun. 8, 1088 (2017).

  175. 175.

    Sivanand, S. et al. Nuclear acetyl-CoA production by ACLY promotes homologous recombination. Mol. Cell 67, 252–265.e56 (2017).

  176. 176.

    Valls-Lacalle, L. et al. Succinate dehydrogenase inhibition with malonate during reperfusion reduces infarct size by preventing mitochondrial permeability transition. Cardiovasc. Res. 109, 374–384 (2016).

  177. 177.

    Valls-Lacalle, L. et al. Selective inhibition of succinate dehydrogenase in reperfused myocardium with intracoronary malonate reduces infarct size. Sci. Rep. 8, 2442 (2018).

  178. 178.

    Kohlhauer, M. et al. Metabolomic profiling in acute ST-segment-elevation myocardial infarction identifies succinate as an early marker of human ischemia-reperfusion injury. J. Am. Heart Assoc. 7, e007546 (2018).

  179. 179.

    Zhang, J. et al. accumulation of succinate in cardiac ischemia primarily occurs via canonical Krebs cycle activity. Cell Rep. 23, 2617–2628 (2018).

  180. 180.

    Ariza, A. C., Deen, P. M. & Robben, J. H. The succinate receptor as a novel therapeutic target for oxidative and metabolic stress-related conditions. Front. Endocrinol. (Lausanne) 3, 22 (2012).

  181. 181.

    Hamel, D. et al. G-protein-coupled receptor 91 and succinate are key contributors in neonatal postcerebral hypoxia-ischemia recovery. Arterioscler. Thromb. Vasc. Biol. 34, 285–293 (2014).

  182. 182.

    Jean, S. R., Ahmed, M., Lei, E. K., Wisnovsky, S. P. & Kelley, S. O. Peptide-mediated delivery of chemical probes and therapeutics to mitochondria. Acc. Chem. Res. 49, 1893–1902 (2016).

  183. 183.

    An, J., Rao, A. & Ko, M. TET family dioxygenases and DNA demethylation in stem cells and cancers. Exp. Mol. Med. 49, e323 (2017).

  184. 184.

    Vissers, M. C., Kuiper, C. & Dachs, G. U. Regulation of the 2-oxoglutarate-dependent dioxygenases and implications for cancer. Biochem. Soc. Trans. 42, 945–951 (2014).

  185. 185.

    Zhang, J. et al. Effect of TET inhibitor on bovine parthenogenetic embryo development. PLoS One 12, e0189542 (2017).

  186. 186.

    Hatch, S. B. et al. Assessing histone demethylase inhibitors in cells: lessons learned. Epigenetics Chromatin 10, 9 (2017).

  187. 187.

    Maes, T., Carceller, E., Salas, J., Ortega, A. & Buesa, C. Advances in the development of histone lysine demethylase inhibitors. Curr. Opin. Pharmacol. 23, 52–60 (2015).

  188. 188.

    Pergola, P. E., Spinowitz, B. S., Hartman, C. S., Maroni, B. J. & Haase, V. H. Vadadustat, a novel oral HIF stabilizer, provides effective anemia treatment in nondialysis-dependent chronic kidney disease. Kidney Int. 90, 1115–1122 (2016).

  189. 189.

    Yeh, T. L. et al. Molecular and cellular mechanisms of HIF prolyl hydroxylase inhibitors in clinical trials. Chem. Sci. 8, 7651–7668 (2017).

  190. 190.

    Yen, K. et al. AG-221, a first-in-class therapy targeting acute myeloid leukemia harboring oncogenic IDH2 mutations. Cancer Discov. 7, 478–493 (2017).

  191. 191.

    Smith, R. A., Hartley, R. C., Cochemé, H. M. & Murphy, M. P. Mitochondrial pharmacology. Trends Pharmacol. Sci. 33, 341–352 (2012).

  192. 192.

    Smith, R. A., Hartley, R. C. & Murphy, M. P. Mitochondria-targeted small molecule therapeutics and probes. Antioxid. Redox Signal. 15, 3021–3038 (2011).

  193. 193.

    Yousif, L. F., Stewart, K. M. & Kelley, S. O. Targeting mitochondria with organelle-specific compounds: strategies and applications. Chembiochem 10, 1939–1950 (2009).

  194. 194.

    Fan, T. W. et al. Altered regulation of metabolic pathways in human lung cancer discerned by 13C stable isotope-resolved metabolomics (SIRM). Mol. Cancer 8, 41 (2009).

  195. 195.

    Sellers, K. et al. Pyruvate carboxylase is critical for non-small-cell lung cancer proliferation. J. Clin. Invest. 125, 687–698 (2015).

  196. 196.

    Maher, E. A. et al. Metabolism of [U-13C]glucose in human brain tumors in vivo. NMR Biomed. 25, 1234–1244 (2012).

  197. 197.

    Hensley, C. T. et al. Metabolic heterogeneity in human lung tumors. Cell 164, 681–694 (2016).

  198. 198.

    Joshi, S. et al. The genomic landscape of renal oncocytoma identifies a metabolic barrier to tumorigenesis. Cell Rep. 13, 1895–1908 (2015).

Download references


This work was supported by a grant to M.P.M. from the Medical Research Council UK (MC_U105663142) and a Wellcome Trust Investigator award to MPM (110159/Z/15/Z). The O’Neill laboratory acknowledges the following grant support: European Research Council (ECFP7-ERC-MICROINNATE), Science Foundation Ireland Investigator Award (SFI 12/IA/1531), GlaxoSmithKline Visiting Scientist Programme and The Wellcome Trust (oneill-wellcometrust-metabolic, grant number 205455).

Author information


  1. School of Biochemistry and Immunology, Trinity Biomedical Sciences Institute, Trinity College Dublin, Dublin, Ireland

    • Dylan G. Ryan
    •  & Luke A. O’Neill
  2. MRC Mitochondrial Biology Unit, University of Cambridge, Cambridge, UK

    • Michael P. Murphy
    •  & Hiran A. Prag
  3. MRC Cancer Unit, University of Cambridge, Hutchison/MRC Research Centre, Cambridge, UK

    • Christian Frezza
  4. Department of Cancer Biology, Dana-Farber Cancer Institute, Harvard Medical School, Boston, MA, USA

    • Edward T. Chouchani
    •  & Evanna L. Mills
  5. Department of Cell Biology, Harvard Medical School, Boston, MA, USA

    • Edward T. Chouchani
    •  & Evanna L. Mills


  1. Search for Dylan G. Ryan in:

  2. Search for Michael P. Murphy in:

  3. Search for Christian Frezza in:

  4. Search for Hiran A. Prag in:

  5. Search for Edward T. Chouchani in:

  6. Search for Luke A. O’Neill in:

  7. Search for Evanna L. Mills in:


All authors wrote, edited and approved final manuscript.

Competing interests

The authors declare no competing interests.

Corresponding author

Correspondence to Luke A. O’Neill.

About this article

Publication history