Question

How does the omission of succinate affect the initial velocity of the reaction?  What does this result indicate about the effects of substrate concentration on enzyme-catalyzed reactions? Is there any succinate present in your isolated mitochondria?

Answer 1

Succinate is at the cross-road of several metabolic pathways.

Succinate is involved in the formation and elimination of reactive oxygen species.

Succinate is also involved in epigenetics and tumorigenesis.

Succinate plays a role in endo- and paracrine modulation and inflammation.

We review succinate as a metabolite or a signal.

Succinate is an important metabolite at the cross-road of several metabolic pathways, also involved in the formation and elimination of reactive oxygen species. However, it is becoming increasingly apparent that its realm extends to epigenetics, tumorigenesis, signal transduction, endo- and paracrine modulation and inflammation. Here we review the pathways encompassing succinate as a metabolite or a signal and how these may interact in normal and pathological conditions.

Succinic acid (butanedioic acid) was discovered in 1546 by Georgius Agricola, a German chemist who purified it from amber by dry distillation. The central role of succinate in biochemistry and bioenergetics became apparent through the pioneering work of Albert Szent-Györgyi. Szent-Györgyi described the dehydrogenation of succinate to fumarate by succinate dehydrogenase (SDH) and the connection of this enzyme to cytochromes and Warburg's respiratory enzyme and oxygen (Albert Szent-Györgyi, oxidation, energy transfer, and vitamins, Nobel Lecture December 11, 1937). Subsequently, Sir Hans Krebs completed the succinate–fumarate–malate–oxaloacetate pathway adding citrate, isocitrate and α-ketoglutarate to the citric acid cycle (CAC); thus succinate was placed into one of the most important cyclic pathways in metabolism, a collecting pool of catabolic and a starting point of many anabolic processes [41]. However, succinate is far from being just a mere substrate for SDH responsible for donation of two hydrogens to the respiratory chain, as it is also localized at the crossroads of metabolic pathways, illustrated in Fig. 1. Succinate is a product of substrate-level phosphorylation materialized in the CAC; it is involved in a macrophage-specific metabolic pathway generating itaconate, and is also a downstream product of the α-ketoglutarate dehydrogenase complex, a heavily regulated multi-subunit complex. Furthermore, succinate metabolism is intertwined with the metabolism of branched-chain amino acids among other metabolites, as well as heme synthesis, ketone bodies utilization, and the GABA shunt. In addition, succinate is a critical mediator of the hypoxic response, an important piece of the puzzle regarding tumorigenesis and is involved in protein succinylation, a newly discovered posttranslational modification. SDH plays an important role in reactive oxygen species (ROS) homeostasis of cells producing superoxide and H2O2 but also contributing in their elimination. Finally, succinate was discovered to exert its effects outside cells in a para- and endocrine mode, mediated by the expression of at least one plasmalemmal succinate receptor type. All of the genes coding for enzymes participating in succinate-related pathways are subjects of mutations leading to various pathologies including oxidative stress, tumor formation, neurodegeneration or hypoxia. The present review discusses the role of succinate and its precursors under normal and pathological conditions.

Succinate in metabolism

1. Succinate dehydrogenase in the citric acid cycle and electron transport chain

Succinate dehydrogenase (also known as complex II) couples two major pathways in mitochondria, the citric acid cycle and the respiratory chain, both being essential for oxidative phosphorylation. This enzyme complex is composed of six subunits encoded by SDHA, SDHB, SDHC, SDHD, SDHAF1 and SDHAF2, see Table 1. SDHAF1 and SDHAF2 are coding for associated accessory factors. SDH is the only complex in the respiratory chain without mtDNA-encoded polypeptides. The complex is embedded in the inner mitochondrial membrane and exhibits a matrix-facing- and a membrane-integrated domain performing two distinct chemical reactions: i) reversible oxidoreduction of succinate to fumarate which is associated with the soluble mitochondrial matrix-associated domain and ii) reversible oxidoreduction of CoQ to CoQH2, a task allocated to the membrane-embedded domain. SDH is the only respiratory complex lacking proton pumping activity; as such, the reduction of CoQ to CoQH2 results in the electron transfer from SDH's covalently bound FADH2 to molecular oxygen yielding six protons, which are liberated on the outer surface of the inner membrane. Thus, the maximal P/O ratio (the number of ATP moles formed per one atom oxygen reduced to water by 2e− flowing in the mitochondrial electron transfer chain) is 1.7 for succinate, i.e., less than that for NAD+-linked substrates which is 2.7. This calculation is based upon current structural information of mammalian ATP synthase

SDH exhibits a complex multi-subunit structure, for which heme plays a crucial role in the assembly [108]. SDHA is a flavoprotein (~ 70 kDa) harboring an FAD prosthetic group; SDHB is an iron–sulfur protein exhibiting 2Fe:2S, 4Fe:4S and 3Fe:4S iron–sulfur centers, while SDHC (also known as cytochrome b 560 protein_ and SDHD (a transmembrane protein) are the membrane integrated domains, having one heme b group at the SDHC/SDHD interface sandwiched in between the transmembrane helices. SDHC and SDHD anchor SDHA and SDHB to the membrane. SDH complexes form tetramers; the crystal structure of SDH expressed in the porcine heart has been determined in Ref. [196]. Two CoQ binding sites have been proposed [78]: a proximal Qp located on the matrix side which affords sensitivity of the complex to theonyl trifluoroacetone TTFA, and atpenin A5 and a distal Qd site (distal to [Fe–S] clusters) located near the intermembrane space. The midpoint potentials of the complex's functional groups are FAD: − 79 mV, [2Fe–2S]: 0 mV, [4Fe–4S]: − 260 mV, [3Fe–4S]: + 60 mV, coQ: + 113 mV, heme b: − 185 mV [196]. Regarding the mechanistic aspects of the SDH action, succinate is oxidized to fumarate on SDHA and reduces FAD to FADH2. In turn, FADH2 transfers electrons to consecutive iron–sulfur centers localized in SDHB. Subsequently, the quinone binding site receives electrons and protons probably from the 3Fe–4S cluster. The details and catalytic role of the distal CoQ site and the heme b group have not been yet elucidated.

2.

Succinate and substrate-level phosphorylation

Succinate-CoA ligase (SUCL), also known as succinyl coenzyme A synthetase, or succinate thiokinase, catalyzes the conversion of succinyl-CoA and ADP (or GDP) to CoASH, succinate and ATP (or GTP), a process also known as substrate-level phosphorylation, (SLP) [90]. The enzyme is a heterodimer composed of an invariant α subunit encoded by SUCLG1 and a substrate-specific β subunit, encoded by either SUCLA2 or SUCLG2. This dimer combination results in either an ATP-forming (EC 6.2.1.5) or a GTP-forming SUCL (EC 6.2.1.4). ΔG of either reaction is ~ 0.07 kJ/mol and therefore, reversible [115]. Unlike bacterial SUCL, in which multiple levels of regulation exist [24], [25], [208], in mammals there has been only one such report, and that is a ten-fold enhancement of ATP-hydrolytic activity by inorganic phosphate [157]. There, it was argued that because inorganic phosphate is known to stabilize the SUCL dimer [65] and increase the Vmax of ATP hydrolysis, it is reasonable to consider that the forward reaction (i.e. ATP- or GTP-forming) kinetics are also enhanced by inorganic phosphate. However, the effect of inorganic phosphate was already saturating at 5 mM, and acknowledging that its concentration in the matrix may well exceed 10 mM [116], [117] it is not expected to be a major regulator in SUCL activity in vivo. Having said that and mindful of the low ΔG of the reaction catalyzed by SUCL, it is safe to assume that it is influenced by the concentration of its reactants by the law of mass-action.

Indeed, as shown in a string of publications by our laboratory, an excess of succinate forces the reaction of SUCL towards ATP (or GTP) hydrolysis, abolishing mitochondrial SLP [39], [40], [43], [97], [98]. To address SLP in intact mitochondria (whether they are isolated from tissues or accessed by cell permeabilization protocols) one needs to dwell on certain bioenergetic principles of phosphorylation. Mitochondria can be both ATP producers and consumers [42], [127], [156], [169], [170], [222]. ATP for hydrolysis could either derive from within the matrix or from the extramitochondrial compartment. Hydrolysis of ATP in the matrix occurs primarily by a reverse-operating F0–F1 ATP synthase; import of extramitochondrial ATP into the matrix may occur by a reverse-operating adenine nucleotide translocase (ANT). The directionalities of the F0–F1 ATP synthase and that of ANT depend on several parameters [39], [40], [43], [130]. The value of the mitochondrial membrane potential (ΔΨm) at which the F0–F1 ATP synthase shifts from ATP-forming to ATP-hydrolysing is termed “reversal potential”, and it is designated as ‘Erev_ATPase’. By the same token, the value of ΔΨm at which the ANT shifts from a ‘forward’ i.e. bringing ADP into the matrix in exchange for ATP to a ‘reverse’, i.e. bringing ATP into the matrix in exchange for ADP, mode of operation is termed ‘Erev_ANT’. Derivations of Erev_ATPase and Erev_ANT are elaborated in Ref. [39].

The values of Erev_ANT and Erev_ATPase as a function of ΔΨm and matrix ATP/ADP ratio are depicted in Fig. 2. In Fig. 2, four spaces are discerned; ‘A’, ‘B’, ‘C’ or ‘D’. Mitochondria exhibiting such ΔΨm and matrix ATP/ADP ratio pair values that would place them in the ‘A’ space (green) produce ATP through the synthesizing operation of the F0–F1 ATP synthase, using ADP from a forward-operating ANT. On the other hand, mitochondria exhibiting such ΔΨm and matrix ATP/ADP ratio pair values that would place them in the ‘B’ space (orange) hydrolyze ATP, but that cannot come from the extramitochondrial compartment, because the ANT is not reversed. For those mitochondria that exhibit ΔΨm and matrix ATP/ADP ratio pair values placing them in the ‘C’ space (gray), both the F0–F1 ATP synthase and the ANT operate in reverse mode and therefore gain access to extramitochondrial ATP pools. Mitochondria exhibiting such ΔΨm and matrix ATP/ADP ratio pair values that would place them in the ‘D’ space (brown) is only a theoretical consideration, and has never been experimentally reproduced.

Depolarized mitochondria can exist in the ‘B’ space and thus avoid draining of cytosolic ATP pools only if they exhibit the capacity of adequate SLP, so that they can maintain sufficiently high matrix ATP/ADP ratios and ‘stay away’ from the ‘C’ space. Since SLP in the matrix is materialized exclusively by SUCL, the importance of the directionality of this reaction is immediately recognized. Mitochondria exhibiting such matrix ATP/ADP and ΔΨm values that would place them in the ‘A’, ‘C’, or ‘D’ space may or may not exhibit SLP.

To interrogate the directionality of SUCL in situ, we devised a biosensor test [43]. This test lies on the principle that the adenine nucleotide exchange through the ANT is electrogenic, since one molecule of ATP4 − is exchanged for one molecule of ADP3 − [99]. In sufficiently energized mitochondria, export of ATP in exchange for ADP ‘costs’ ~ 25% of the total proton motive force [4]. Therefore, during the forward mode of ANT, abolition of its operation by an ANT inhibitor such as carboxyatractyloside (cATR) or bongkrekic acid (BKA) will lead to a gain of ΔΨm (typically measured fluorimetrically using a fluorescent potentiometric probe), whereas during the reverse mode of ANT, abolition of its operation by the inhibitor will lead to a loss of ΔΨm. An explanatory scheme of this test is depicted

By using this biosensor test, the role of succinate in abolishing substrate-level phosphorylation in intact mitochondria can be easily demonstrated. The results of a typical experiment are shown in Fig. 3B, C. As shown in Fig. 3B, mitochondria respired on glutamate and malate (black trace), or succinate (red trace), or glutamate and malate and succinate (blue trace), and allowed to fully polarize (solid traces). State 3 respiration was initiated by ADP; in an air-tight chamber mitochondria eventually run out of oxygen, verified by recording ‘zero’ levels of dissolved oxygen (dashed traces). Anoxia also coincided with the onset of an additional depolarization leading to a clamp of ΔΨm at ~− 100 mV. At this membrane potential, F0–F1 ATP synthase was operating in reverse mode, i.e. pumping protons out of the matrix at the expense of matrix ATP hydrolysis. In mitochondria respiring on substrates supporting substrate-level phosphorylation, i.e. glutamate and malate [43], inhibition of the ANT caused a moderate repolarization. This implied that the translocase was still operating in the forward mode. In contrast, mitochondria respiring on succinate alone or glutamate and malate and succinate, reacted to ANT inhibition with an immediate and complete depolarization in anoxia. Likewise, in the presence of atpenin A5, a specific inhibitor of SDH [134] which leads to accumulation of succinate in the matrix, ANT inhibition induced a depolarization in mitochondria that had been respiring even on substrates supporting substrate-level phosphorylation (Fig. 3C, magenta trace). It is therefore inferred that succinate, either by exogenous addition or due to a built-up attributed to inhibition of SDH, abolishes SLP

3. Succinate and ketone bodies utilization

Ketone bodies (KB) encompass acetone, acetoacetate and β-hydroxybutyrate, molecules synthesized in the liver from acetyl-CoA; the term “ketone body” is actually a misnomer, because β-hydroxybutyrate is not an actual ketone. Acetone is a degradation product of acetoacetate and is not an energy source, although it is a strong inducer of gluconeogenesis [91]. KB are formed after prolonged starvation and are almost exclusively catabolized in the brain and heart. Acetoacetate and β-hydroxybutyrate are in equilibrium with the NAD+/NADH ratio through the reversible reaction catalyzed by β-hydroxybutyrate dehydrogenase (Fig. 1). Catabolism of KB commences with the thioesterification of acetoacetate by succinyl-CoA:3-ketoacid CoA transferase (SCOT) using succinyl-CoA as CoA donor, yielding acetoacetyl-CoA and succinate. With the subsequent thioesterification of acetoacetyl-CoA by thiolase using CoASH, two molecules of acetyl-CoA will form that can enter the citric acid cycle. As it is evident, KB utilization is in equilibrium with succinyl-CoA/succinate ratio. It is therefore reasonable to suggest that KB utilization should lead to an increase in succinate. Indeed, ketogenic diets markedly boost the protein expression of hypoxia-inducible factor-1alpha (HIF-1α) [162], [225], presumably due to an elevation in succinate that in turn inhibits prolyl hydroxylases that constitutively degrade this transcription factor (see below under Section 5). By the same token, an excess of succinate would inhibit KB utilization; product inhibition of 3-oxoacid CoA-transferase occurs with a Ki < 1 mM for succinate [61], [62], [184]. The concentration of succinate in the mitochondrial matrix fluctuates from ~ 0.5 mM in normoxia, to almost 6 mM under ischemic and hypoxic conditions [20], [21], [63]. It can therefore be postulated, that immediately after ischemia/hypoxia, or when SDH operation is affected by genetic aberrations (see below under Section 5) leading to elevations in succinate concentration in the matrix, KB utilization would be inhibited.

4. Succinate and heme metabolism

Heme is the prosthetic group of hemoglobin, myoglobin, cytochromes and several enzymes [114], [135], [218]. Its synthesis commences by the condensation of glycine and succinyl-CoA to δ-aminolevulinate by ALA synthetase (Fig. 1), for which there are two 2 tissue-specific isoenzymes: a housekeeping enzyme encoded by the ALAS1 and an erythroid tissue-specific enzyme encoded by ALAS2. Succinyl-CoA can be derived either from the reaction catalyzed by the α-ketoglutarate dehydrogenase complex (KGDHC), or reversal of SUCL, or from catabolism of biomolecules converging to propionyl-CoA (see below under Section 2.5). It is noteworthy that ALAS2 specifically associates with the ATP-forming subunit of SUCL, SUCLA2, implying that SUCL-derived succinyl-CoA is perhaps the prevailing one, at least in erythroid tissues [68]. There, provision of succinate from SDH for thioesterification to succinyl-CoA by SUCL could be critical in maintaining heme synthesis. On the other hand, cancer cells exhibiting a deficiency in fumarase leading to accumulation of fumarate and succinate are vitally dependent on an intact heme catabolism: indeed, silencing of HMOX1, the gene encoding for heme oxygenase 1, an enzyme that degrades heme to biliverdin, reduces colony-forming capacity in fumarase-deficient cells [67].

A link between impaired heme synthesis and KGDHC deficiency (or upstream of this enzyme complex) can be inferred by the work of Atamna and colleagues [13], [14]. Finally, biotin deficiency which is expected to affect negatively the activity of the five biotin-dependent carboxylases (pyruvate carboxylase, propionyl-CoA carboxylase, 3-methylcrotonyl-CoA carboxylase, acetyl-CoA carboxylase 1 and 2) may also lead to an impairment of heme synthesis, partially due to the defect in propionyl-CoA metabolism, see Ref. [15] and below under ‘succinate and the metabolism of biomolecules converging to propionyl- and methylmalonyl-CoA’.

5. Succinate and the metabolism of biomolecules converging to propionyl-and methylmalonyl-CoA

Methionine, threonine, isoleucine, valine, propionate, odd-chain fatty acids and cholesterol are catabolized in the citric acid cycle entering as succinyl-CoA through propionyl → methylmalonyl-CoA, mediated by the sequential actions of propionyl-CoA carboxylase and methylmalonyl-CoA mutase [147], (Fig. 1). Increases in propionyl-CoA and methylmalonyl-CoA may cause secondary metabolic aberrations due to their ability to inhibit steps in urea cycle [49], [72], [193], gluconeogenesis [79], [178], [179] and the glycine cleavage system [84], [85]. Thus, inhibition of SDH leading to a built-up of succinate that in turn results in product inhibition of SUCL and upstream pathways leading to elevations in propionyl-CoA and methylmalonyl-CoA, potentially result in inhibitions of the aforementioned pathways. Indeed, patients suffering from SDH or fumarase deficiency have also been reported to exhibit hyperammonemia and hypoglycemia [209], [175]. Propionyl and/or methylmalonyl aciduria or acidemia has not been reported but it is possible that it has not been sought for these very rare disorders. However, methylmalonyl acidemia is a consistent finding in patients suffering from SUCL deficiency [33]. It is also noteworthy that methylmalonic acid is an inhibitor of SDH [59], [200], therefore, a vicious cycle between methylmalonate and SDH mediated by an elevation in succinate concentration, can be envisaged.

.6. Succinate and itaconate metabolism

Another intermediate that can be metabolized by SUCL but also inhibits SDH activity and as a consequence of this, interferes with succinate metabolism, is itaconate. Itaconic acid (2-methylidenebutanedioic acid, methylenesuccinic acid, CAS registry number: 97-65-4) is an unsaturated dicarboxylic acid that has been identified in a number of metabolomic studies using activated macrophages [195], Mycobacterium tuberculosis-infected lung tissue [185], urine and serum samples [105], and glioblastomas [221]. Cells of macrophage lineage produce itaconic acid from cis-aconitate (an intermediate in the citrate → isocitrate catalysis by aconitase, Fig. 1) through an enzyme exhibiting cis-aconitate decarboxylase activity, coded by the immunoresponsive gene 1 (Irg1, NM_001258406.1) [47], [131], [194]. Irg1-mediated itaconate production contributes to the antimicrobial activity of macrophages by inhibiting isocitrate lyase, an enzyme of the glyoxylate shunt [126], [152]. The glyoxylate pathway is essential for the survival of bacteria growing on fatty acids or acetate [83], and it is absent in animals. Starting with the work of Henry A. Lardy reported over 50 years ago [3], [217], and recently revisited [47], [131] including by our group [137], it is established that itaconate can be converted by SUCL to itaconyl-CoA at the expense of ATP (or GTP) (Fig. 1), and is also a weak competitive inhibitor of SDH [3], [28], [56], [64]. Lardy's work also suggested that itaconate can be oxidized in a malonate-sensitive manner implying metabolism through SDH, but due to a lack of hydrogen on the α-carbon of itaconate a double bond cannot be formed and therefore, itaconate cannot be processed directly by SDH. There are two possibilities by which it can be converted to product(s) suitable for oxidation by the SDH: i) saturation of itaconate to methylsuccinate which can be processed by SDH [1], [2]; ii) itaconate hydroxylation yielding hydroxymethyl-succinate, though to the best of our knowledge this is still a theoretical possibility. Decarboxylation or isomerisation of itaconate would yield products that cannot be further metabolized by SDH.

Because of all of the above, itaconate effectively abolishes SLP due to: i) a ‘CoA trap’ in the form of itaconyl-CoA that negatively affects the upstream supply of succinyl-CoA from KGDHC; ii) depletion of ATP (or GTP) which is required for the thioesterification by SUCL; and iii) inhibition of SDH leading to a build-up of succinate which shifts SUCL equilibrium towards ATP (or GTP) utilization. Therefore, Irg1-expressing cells of macrophage lineage trade their capacity of mitochondrial SLP for producing itaconate during mounting of an immune defense [137]. Under these circumstances, Irg1-expressing cells are further expected to accumulate succinate, which plays a significant role as an inflammatory signal

7. Succinate and the GABA shunt

The ‘GABA shunt (Fig. 1) is a pathway commencing from glutamate being converted to GABA by the cytosolic glutamate decarboxylase (4.1.1.15), encoded by either GAD65 or GAD67 [188]. GABA may also arise by metabolism of putrescine [183] or enter the cytoplasm from the extracellular space. GABA transaminates with α-ketoglutarate to form glutamate and succinate semialdehyde by the mitochondrial GABA transaminase (GABA-T, EC 2.6.1.19). Succinate semialdehyde will get dehydrogenated by succinate semialdehyde dehydrogenase (EC 1.2.1.24) yielding succinate, and thus enter the citric acid cycle. Succinate semialdehyde can also be converted to γ-hydroxybutyrate by succinic semialdehyde reductase [158], an enzyme that is primarily localized in astrocytes and microglia of the human brain [87], [167]. γ-hydroxybutyrate is a neurotransmitter and a psychoactive drug, but is also used as a general anesthetic as well as in the treatment of alcohol withdrawal [8], [50], [111]. A deficiency of succinic semialdehyde dehydrogenase will lead to accumulation of γ-hydroxybutyrate [154], leading to epilepsy and other neurological dysfunctions [94].

The enzymes required for a complete ‘GABA shunt’ are present in specific neurons but also glial cells in the adult human brain; furthermore, they are known to be expressed in cells of macrophage lineage, and in pancreas.

In neurons, GAD65 is specifically localized in the nerve terminals of GABA-ergic cells [38]. Exocytosed GABA exerts an inhibitory effect on synaptic transmission by interacting with ionotropic receptors on the postsynaptic membrane, resulting in an increased chloride conductance ensuing an inward chloride current and a consequent hyperpolarization [102]. However, in neonatal hippocampal neurons the electrochemical gradient of chloride is outward directed therefore, opening the receptor channels is associated with depolarization; hence, at this developmental stage, GABA is an excitatory neurotransmitter [38]. GAD67 is expressed throughout the cytosol of GABA-ergic neurons and may be involved in metabolic functions pertinent to the GABA shunt [124]. Astrocytes express both isoforms of glutamate decarboxylase [190], GAD67 (but not GAD65), GABA-T as well as GABA receptors and transporters [133], [211]. Succinate semialdehyde dehydrogenase is expressed in both neurons and glial cells in the adult human brain.

Recently, cells of macrophage lineage have also been reported to express enzymes of the GABA shunt [132], [198]. In these cells, the shunt was reported to be an important pathway for elevation of succinate upon lipopolysaccharide treatment. In pancreas, GAD65 is known to mediate GABA formation in islet β-cells, acting as an intra-islet transmitter regulating hormone release [216]. In β-cells, GABA induces membrane depolarization and insulin release, whereas in α-cells GABA results in hyperpolarization and suppression of glucagon secretion [216]. This differential response of plasmalemmal potential to GABA is dependent on the expression of K+/Cl− cotransporter-2 (KCC2) [122] which is present in α-cells, but not β-cells [54].

Obviously, the GABA shunt produces less ATP from glutamate to succinate than the glutamate-α-ketoglutarate-succinyl-CoA-succinate pathway. In the GABA shunt only succinic semialdehyde dehydrogenase would contribute to the ATP production by the reduction of 1 mol of NADH worth of ~ 2.7 mol ATP. In the other pathway NADH production by α-ketoglutarate dehydrogenase reaction could produce 1 mol of NADH (2.7 mol ATP) and substrate level phosphorylation by SUCL contributing with a further 1 mol of ATP. Glutamate can be either deaminated by glutamate dehydrogenase producing NAD(P)H or transaminated, thus producing 3.7–6.6 mol of ATP.

3. Succinate in signal transduction

3.1. Succinate and protein succinylation

Succinylation is the mechanism by which a succinyl group is added to a residue of a protein, usually a lysine but it could be also an arginine or histidine, depending on pH. Chemical succinylation (using succinic anhydride) has been used for over 50 years [129], [149] for the purpose of determining the importance of amino groups to a protein's function. It is also extensively used in food industry in order to modify the physicochemical properties of proteins (such as emulsion and foaming capacity, water- and oil-absorption capacity, swelling power and solubility, to name a few) and hence widen their potential applications in beverages, bakery products, instant meals, creams, etc. [6], [11], [227], [230]. However, recently, succinylation of proteins in lysine residues has been shown to occur as a posttranslational modification in vivo [229], in which a succinyl moiety is attached on a lysine through an amide (isopeptide) bind. Isotopic labeling showed that the succinyl moiety may originate from succinate, however, in these experiments extremely high concentrations of succinate were used (80–160 mM) [229]. Furthermore, it was more recently shown that succinyl-CoA originating from KGDHC is the precursor for lysine succinylation, and that a defect in SUCL leads to over-succinylation of proteins [220]. On the other hand, protein succinylation was also found to extensively occur in cytosolic and nuclear proteins [220], implying that another succinyl-metabolite drives succinylation in the cytoplasm and nucleus, or, a yet to be identified enzyme yielding succinyl-CoA exists outside mitochondria. So far, the consensus is that the process of succinylation of proteins in vivo is non-enzymatic [214], [215] and therefore dependent on the concentration of succinyl-CoA by the law of mass-action; although this concept has not been unequivocally proven and there has been at least one report claiming that KGDHC acts as a trans-succinylase [70], it is extremely likely that it is true, given that enzymatic reactions of short-chain lysine acylations, such as acetylation, propionylation and butyrylation, use their corresponding high-energy CoAs, a notion that has been hypothesized well before the discovery of this posttranslational modification [96]. Therefore, metabolic scenarios revolving around succinate are bound to influence the levels of succinyl-CoA, and as a consequence of this, physiologically regulate the function of thousands of proteins [220]. Yet, the physiological regulatory consequences of succinylation remain to be discovered. Furthermore, excessive alterations in the level of protein succinylation may lead to pathology. Indeed, mutations in isocitrate dehydrogenase isoform 1 causing it to form R-2-hydroxyglutarate lead to inhibition of SDH with an ensuing accumulation of succinyl-CoA and hypersuccinylation promoting cancerous metabolism and apoptotic resistance [113]. On the other hand, desuccinylation occurs enzymatically by sirtuin 5, a mitochondrial protein acting on hundreds of matrix proteins, thus implicated in the regulation of energy metabolism [139], [212], [215].

3.2. Succinate in inflammation

Immune cells alter their metabolism in response to varying conditions, a maneuver which is crucial for proper immune function [153], [198]. A multitude of alterations lead to accumulation of succinate by any of the three, non-mutually exclusive mechanisms: i) inhibition of SDH activity due to an overall decrease in oxidative phosphorylation [101], ii) increased glutamine metabolism through anaplerosis of α-ketoglutarate into the citric acid cycle [198] or partially through upregulation of the ‘GABA shunt’, which is present in cells of macrophage lineage, yielding succinate [132], [198], and iii) inhibition of SDH through production of itaconate (see above under Section 2.6). In turn, succinate will signal the elevation in HIF-1α concentration (through either transcription factor stabilization (see below under Section 5) or other, distinct regulatory mechanisms [103] with downstream activation of several targets [26], [66], [148], or activate succinate receptors (see below under Section 3.3). A well-established paradigm is that like this, HIF-1α favors differentiation of T lymphocytes into pro-inflammatory Th17 cells and attenuates regulatory T cell development [53].

3.3. Succinate receptors: succinate receptor 1 (SUCNR1)

In 2004, a previously orphan G-protein coupled receptor (GPCR) GPR91 [223] was identified as succinate receptor [81]. Its stimulation by succinate resulted in IP3 signalization [81] with calcium mobilization and transient phosphorylation of extracellular regulated kinase (ERK)1/2 mediated by Gq and Gi [168]. By searching for natural ligands of GPR91, He et al. [81] found that pig kidney cell fractions were able to elevate intracellular calcium concentration in GPR91 expressing cells. The natural ligand was identified as succinate and GPR91 was renamed to SUCNR1. Succinate is normally present in the blood plasma at about 5 μM concentration [104]. The first biological effect of succinate to be identified mediated by SUCNR1 was hypertension via stimulation of renin release [81], recently reviewed by Peti-Peterdi et al. [155]. The receptors' expression is high on dendritic cells [171], [132] resulting in production of pro-inflammatory cytokines, enhancing activation of T helper cells. In human monocyte-derived dendritic cells, succinate stimulated migration draining lymph nodes [171]. In kidney, SUCNR1 can be activated by the succinate present in the glomerular filtrate [168]. In white adipose tissue it is hypothesized that SUCNR1 can be a sensor for dietary energy [125]. In liver stellate cells, succinate receptors react to succinate released from ischemic hepatocytes resulting in cellular activation [48]. In the brain, SUCNR1 is expressed on the surface of retinal ganglion cells and stimulate vascular endothelial growth factor (VEGF) expression [73]. Expressing succinate receptor in retinal pigment epithelium prevented the dry form of age-related macular degeneration [60]. Recently, SUCNR1 was found to be located in cardiomyocytes and in a right ventricular hypertrophy (RVH) model, succinate treatment aggravated the development of RVH. Inhibition of the succinate receptor signal transduction pathway protected against the consequences of the succinate treatment [226]. Overall, it is apparent that the succinate receptor SUCNR1 exhibits a widespread distribution in the body. It is therefore not surprising that succinate-mediated para- and endocrine signal transduction affords metabolic (white adipose tissue), immunological (stellar cells, T helper cells) or cardiac effects, influences blood pressure or retinal functions, plays a role in the metabolic diseases, but also contributes in the complications of diabetes, heart failure and liver damage.

4. The role of succinate in mitochondrial ROS homeostasis

Recently, Dröse published a comprehensive review regarding the role of SDH in ROS production and elimination [58]; hereby we will cover this topic from succinate as a metabolite point of view

4.1. The role of succinate in mitochondrial ROS production

In order to clarify the possibility of the involvement of an enzyme in ROS (superoxide, H2O2, hydroxyl radical) formation, the structural characteristics of the enzyme and the pathways associated with need to be addressed. The SDHA subunit possesses a FAD prosthetic group, which is a possible candidate for ROS formation [146]. Almost all flavoenzymes are known to produce ROS, reviewed in Ref. [9], [10], [203]. Possible further sites for ROS production are the iron–sulfur clusters, which are found not only in complex II but complexes I and III as well [119]. Heme group of cytochromes is able to participate in auto-oxidation reactions, or like cytochrome c can participate in superoxide scavenging [16], [32]. It is to be noted that the function of heme groups in SDH from the catalytic point of view is not entirely clear, thus their participation in ROS homeostasis is essentially unknown. The immediate electron acceptor associated with the SDH is ubiquinone. Ubiquinone accepts electrons from the iron–sulfur centers. To the best of our knowledge, this reduction involves a semiquinone intermediate [173], [207], a compound exhibiting a strong tendency for auto-oxidation.

4.2. Succinate and reverse electron transport (RET) in state 4 respiration

In state 4 respiration there is a near equilibrium between the proton motive force and the redox spans of complexes III and I, thus at the expense of inward flux of protons, electrons can flow backwards, Fig. 4A. When mitochondria are supplied with saturating levels of succinate, membrane potential is high and electron flow is slow because all of the ADP is converted to ATP. Part of the flow of electrons from succinate is used to reverse electron transfer through complex I. Reverse flow through complex I will reduce NAD+ to NADH [37]. Part of the electron flow follows the canonical pathway from CoQ, complex III, cytochrome c, to complex IV and oxygen reduction. The reverse flow of electrons generates very high rates of H2O2 [100]. Under these conditions, the sources of H2O2 have not been completely clarified yet. Complex I could be one of them, because this type of H2O2 production is highly sensitive to complex I inhibitors, such as rotenone. There are at least two additional putative ROS forming sites in complex I.; the [Fe–S] clusters and the FMN. During reverse electron transport (RET) however, not only complex I but also matrix NAD+ can be reduced. This reduction will initiate further ROS formation on the E3 subunit of pyruvate dehydrogenase or a-ketoglutarate dehydrogenase complex [192], [202].

4.3. Succinate-mediated ROS production in the presence of ADP (state 3 respiration)

Under phosphorylating conditions entry of protons through the ATP synthase results in a decrease of electrochemical proton gradient, Fig. 4B. The reduced proton motive force cannot drive the inward flux of protons through complex I and III, thus the conditions for RET and consequently to RET-mediated ROS production are not favorable, therefore, the rate of ROS production is smaller than in state 4 respiration. The probability of RET occurring in mitochondria in situ is low because succinate is mainly generated by NAD+-dependent dehydrogenases in the citric acid cycle and net flow of electron through complex I is always forward [141]. However, ischemic accumulation of succinate was shown to control reperfusion injury through mitochondrial ROS formation [45]. In the presence of ADP, ΔΨm is diminished by about 25 mV [44], thus the conditions for RET are also not favorable. In some studies and reviews, ROS generated by an intact, fully operational SDH is considered as negligible [17], [189]. Quinlan et al. however, described that mitochondria respiring on succinate produce H2O2 under conditions when the contribution of respiratory complexes I and III in ROS formation was excluded and the main source of ROS is SDH itself [164]. It was further concluded that ROS production by the SDH, FAD reduction is essential. Finally, Quinlan et al. also showed that SDH was able to produce H2O2 in the reverse reaction when CoQ pool was reduced by electrons supplied from the oxidation of α-glycerophosphate.

4.4. Succinate-mediated ROS production in respiration-inhibited mitochondria

Under pathological conditions, different parts of the respiratory chain can be inhibited. Inhibition of various parts of the respiratory chain provides information about the mechanisms of mitochondrial ROS production. Specifically:

4.4.1. Complex I inhibition

The conditions of RET have already been discussed above. During RET, the very high rate of succinate mediated ROS production is drastically decreased by the complex I inhibitor, rotenone. However, in the presence of uncoupler, while succinate-supported mitochondria show a high rate of oxidation with a low rate of H2O2 production, inhibition of complex I results in an immediate increase of ROS formation. Although addition of rotenone elevates mitochondrial NADH level, it neither slows down oxygen consumption nor changes ΔΨm to a considerable extent, thus it does not reestablish RET. Under these conditions, malate is formed in the citric acid cycle from succinate and gets oxidized thus rotenone would inhibit the forward flow of electrons in complex I, resulting in an increased ROS production [205].

4.4.2. Complex II inhibition

Complex II inhibitors are extremely useful tools for exploring the sites and mechanisms responsible for ROS formation.

3-Nitropropionic acid (3-NPA) is an irreversible inhibitor of the SDHA's succinate binding site. In Hela cells induction of apoptosis was shown to be hindered by 3NPA, whereas TTFA caused a time- and dose-dependent increase of apoptosis [109], reviewed in Ref. [75]. The contribution of SDH in terms of ROS metabolism has not been addressed by the use of 3NPA; instead, malonate, TTFA and atpenin A5 have been used more extensively. Malonate is an inhibitor acting competitively on the carboxylate site (succinate binding site) of SDHA. Malonate effectively inhibits ROS production generated by low concentrations of succinate in the absence of RET.

Both (TTFA) and atpenin A5 bind to the Q site located at the matrix-membrane interface and TTFA may bind to the second, putative, distal Q site (Qp) as well [58]. Atpenin A5 was a more effective inhibitor of glycerophosphate stimulated ROS production (electrons were directed from CoQ to complex II flavin site) than that of malonate, indicating again that the flavin of SDHA was the most important site of ROS production in the complex II [164].

4.4.3. Complex III inhibition

In mitochondria oxidizing succinate during state 3, inhibition of complex III with antimycin stimulates ROS production robustly. This stimulation is attributed to a set of reactions known as CoQ cycle reviewed in Ref. [206]. According to Trumpower, an unstable semiquinone is formed in the Qo site (near to the outer surface of the inner membrane) responsible for superoxide formation. Recently, this model was further refined by the elegant work in Martin Brand's laboratory [163].

4.5. The role of succinate in mitochondrial ROS elimination

Most publications dealing with mitochondrial ROS homeostasis focus on ROS generation and neglect the capacity of mitochondria to dismutate superoxide and eliminate H2O2. Simple and convincing experiments have already demonstrated the lack of ROS accumulation in intact mitochondria and the impairment of ROS scavenging mechanism as a prerequisite to detect net release of H2O2 [57], [144], [191], [231]. Early studies in the 70's and 80's showed that succinate exerted an inhibitory effect on lipid peroxidation in mitochondria/mitoplasts/submitochondrial particles, generated by NADPH/ADP/Fe3 + complexes [136] or by organic hydroperoxides [23]. These pro-oxidants strongly stimulated malondialdehyde (the end-product of membrane lipid peroxidation) formation, and peroxidation of membrane lipids dissipated mitochondrial membrane potential due to the emergence of physical holes in the lipid bilayer [23]. The protective role of succinate as mitochondrial respiratory substrate was attributed primarily to the reduction of CoQ pool [35]. NAD+ dependent substrates were less efficient in inhibiting of lipid peroxidation because strong peroxidative stimuli permeabilized membranes releasing NAD+; thus with decreasing coenzyme concentration, oxidation of substrates slowed down and the antioxidant CoQH2 formation was decreased. Succinate however, was oxidized by the membrane-bound SDH and the enzyme activity was only moderately sensitive to oxidants [201], therefore it was more efficient in maintaining CoQH2 pool, maintaining membrane integrity [204].

4.6. Succinate and ROS elimination of mitochondria

In mitochondria at least 12 enzymes participate in ROS elimination [203]. Except superoxide dismutases and catalase, all of the remaining enzymes playing a role in ROS elimination require respiring mitochondria in a direct- or indirect manner. Probably one of the most important factors in H2O2 elimination is the availability of NADPH. NADPH is required for the regeneration of glutathione peroxidase, thioredoxin reductase and the peroxiredoxin system. The contribution of succinate oxidation to NADPH generation compared to other respiratory substrates is relatively small. None of the enzymes participating in the regeneration of NADPH (transhydrogenase, NADP+-dependent isocitrate dehydrogenase, malic enzyme, glutamate dehydrogenase) use succinate as a substrate. Thus, the H2O2 scavenging activity of isolated mitochondria supported with succinate is lower than that of mitochondria supported by substrates like glutamate and malate, or pyruvate and malate [213], [231].

5. Succinate in tumorigenesis

5.1. Mutations of genes encoding for SDH subunits associated with hereditary pheochromocytoma/paraganglioma syndromes

Succinate dehydrogenase (SDH, EC 1.3.5.1) is an enzyme complex with four subunits encoded by four nuclear genes: SDHA, SDHB, SDHC and SDHD. Germline mutations (detected in DNA isolated from peripheral blood samples) of genes encoding for SDH subunits have been shown to represent genetic susceptibility for the development of familial pheochromocytomas and paragangliomas (Pheo/PGL). The chromosomal localization, length and mutations of SDHx genes are summarized in Table 1. These tumors arise from the adrenal medulla and in paraganglia of the autonomous nervous system. Based on genetic background to date, five types of familial paraganglioma/pheochromocytoma syndrome (PGL1-5) have been described. First, in 2000, mutations of SDHD gene were identified causing PGL type 1 syndrome. Later, mutations of the SDHC (PGL type 3), SDHB (PGL type 4), SDH5 also called SDHAF2, encoding for the SDH assembly factor 2 that is required for the assembly of the FAD prosthetic group to complex II (PGL type 2) and SDHA (PGL type 5) were also detected [12], [18], [31], [80], [142]. Mutation of SDHx genes causing PGL was somehow unexpected because, earlier mutation of the SDHA gene was found in a family with Leigh syndrome (an early-onset progressive neurodegenerative disorder linked to defects causing impaired oxidative phosphorylation and no paraganglioma [29], [52]). In addition, mutations of SDHAF1 caused an isolated complex II deficiency presenting as a specific leukoencephalopathic syndrome [69] again without clinical evidence of tumors. The clinical presentation showed a rapidly progressive psychomotor regression after a 6- to 11-month disease-free interval with lack of speech development, followed by spastic quadriparesis and partial loss of postural control with dystonia. In muscle and fibroblasts a ~ 20–30% residual activity of SDH was demonstrated, while other mitochondrial enzyme activities were normal. The molecular genetic analysis performed in these two families identified two missense mutations affecting the LYR-motif of SDHAF1 and both mutations were detected in homozygous form [69]. Contrary to these syndromes, SDH mutations causing hereditary Pheo/PGL are heterozygous mutations inherited in an autosomal dominant manner, demonstrating that the SDH genes function as classical tumor-suppressor genes.

As seen in Table 1 the genes which are mostly mutated are SDHB and SDHD, and only few mutations have been detected in SDHC; mutations in SDHA or SDHAF2 are extremely rare among patients with Pheo/PGLs.

Based on available genotype–phenotype correlations it seems that SDHD and SDHC mutations are being detected more frequently in cases with a positive family history, while SDHB mutations are more common in sporadic cases [151].

Mutations in SDHB, SDHC and SDHD can be found through the genes and no clear mutation hotspots can be distinguished, however, mutations affecting the conserved regions and showing a minor allele frequency (MAF) < 0.01 are pathogenic. In addition, based on bioinformatic analysis of reported mutations it became evident that more than half (41 out of 79) of pathogenic missense SDHB mutations affect the highly conserved Fe–S clusters or the L(I)YR motifs which are important for incorporation of the Fe–S cluster into SDHB (Fig. 5). Very recently, Saxena et al. [176] characterized the biochemical and genomic consequences of the R46Q mutation in SDHB-deficient UOK269 primary kidney cancer cell line. They showed that in the absence of SDHB, respiration was undetected, succinate was elevated, hypoxia-inducible factor 1α (HIF1α) but not HIF2α was increased, glutamine was the main source of fuel for the citric acid cycle and a strong DNA CpG island methylator phenotype was detected [176]. The detailed description of pathomechanisms associated with SDHx mutations is summarized

No significant difference was observed in median age at diagnosis of the first tumor in carriers of SDHB and SDHD mutations, but it must be noted that in patients younger than 20 years of age the most frequently mutated gene was SDHB (23 of 32 patients with metastatic disease exhibited a SDHB mutation) [95]. Metastatic potential of the SDHB associated Pheo/PGL was also demonstrated in adult patients where the risk for malignancy in SDHB mutation carriers has been reported between 24%–71.4% [7], [30], [199].

In addition to the malignant potential related to the SDHB mutations multiple primary tumors and intraabdominal, extraadrenal PGLs are also associated with SDHB mutations [95]. On the contrary, PGLs localized in the head and neck region are frequently associated with SDHD and SDHC mutations. Of note, carriers of SDHD mutations can present frequently with multiple tumors, with an incidence varying between 30 and 74% [19], [138].

5.2. Other tumors associated with SDHx mutations

Beside Pheo/PGL, other tumors were also shown to harbor SDHx mutations. Of these, Carney–Stratakis syndrome (Carney dyad) is an autosomal dominantly inherited tumor syndrome with PGLs and gastrointestinal stromal tumors (GIST) where similarly to hereditary Pheo/PGL syndromes, germline mutations and deletions of SDHB, SDHC and SDHD accompanied with loss of heterozygosity in tumor tissues have been detected [128], [150]. Renal cell carcinoma with oncocytic features is another tumor type associating with SDHB and SDHD mutations [138], [166], [210]. In these diseases, mutations have been demonstrated at germline level and in tumor tissues the loss of a wild type allele underlines the classical tumor suppressor role of these genes.

Apart from the disease causing mutations (with allele frequencies-(MAF) < 0,01), variants (MAF > 0,01) of the SDHx genes have been detected in thyroid C-cells hyperplasia [118], papillary thyroid cancers [138], [199], pituitary adenoma [19], [55], prolactinoma [143], pancreatic neuroendocrine tumors [55], bronchial carcinoid [123], chronic lymphocytic leukemia [165], adrenal neuroblastoma [34], and in patients with Cowden-like syndrome (a rare disease presenting with breast, thyroid, uterine benign and malignant diseases) [140]. In these later cases it is unclear whether the identified SDH mutations/variants are the sole cause for disease manifestation or represent genetic modifiers and additional genetic, epigenetic and/or environmental factors may also contribute to tumorigenesis. Further genetic studies performed in well-characterized patients will clarify the genetic role of these variants in tumorigenesis.

5.3. Pathomechanisms associated with SDHx mutations

Multiple mechanisms have been proposed for tumorigenic processes suffering from SDHx mutations. However, it has to be mentioned that tumourigenesis or cell survival at the very least was completely unexpected in cells lacking SDH enzyme activity. Therefore the “real” mechanism for tumorigenesis is still under debate. Several aspects and hallmarks associating with specific survival mechanisms have been detected in SDHx mutated tumors, but the complex picture is still obscure.

First of all, the main biological aspect of SDHx related tumors is that they are mostly benign lesions, except in approximately half of the SDHB mutations associated cases. This very important feature has to be kept in mind when we are evaluating the tumourigenesis process. For classical tumor-suppressor molecules (i.e. p53, PTEN, VHL, etc.) the tumorigenic processes have been determined in specific animal models having the respective tumor suppressor gene knocked-down or mutated. The resulting phenotype in these animals resembles those observed in patients having mutations in the respective tumor suppressor gene, hence contributing fundamentally to the understanding of the function of a particular cancer gene [228]. On the other hand, more pronounced effects leading to cell survival, failure of apoptosis and tumor formation were detected after silencing these tumor suppressors.

Regarding the SDHx deficient animal models the clinical phenotypes observed were only partly informative for tumourigenesis. First, the SDHD knockout animal model was developed. Homozygous SDHD−/− animals died at early embryonic stages while heterozygous SDHD+/− mice showed a general, non-compensated deficiency of succinate dehydrogenase activity without any other major alterations, including tumors [160]. The phenotypes observed in other SDHx deficiency models were similar only to a model carrying a missense SDHC mutation showing increased ROS levels and oxidative damage in the mitochondria, mitochondrial respiratory chain dysfunction and decreased body size [159]. Based on these reports the dynamism of tumorigenesis in SDHx mutated cases together with mechanisms leading to malignancy observed in relation with SDHB mutations are still unclear.

5.4. Neoplastic growth conferred by an excess in TRAP1-mediated inhibition of SDH

An alternative way for tumorigenesis implicating succinate but no mutations in SDH is through the action of tumor necrosis factor receptor-associated protein-1 (TRAP1), an evolutionarily conserved chaperone of the Hsp90 family [5] and a bona fide inducible target of the proto-oncogene c-Myc [46]. TRAP1 is overexpressed in the mitochondrial matrix of a multitude of tumor cell types [92]. There, as it has been thoroughly investigated by the group of Andrea Rasola, it supports tumor progression by suppressing mitochondrial respiration via inhibition of SDH [177]. The contribution of TRAP1 to neoplastic potential is so robust that if its expression in cancer cell lines is silenced by RNAi, the cells loose the ability to form foci; accordingly, tumor cells transfected with short-hairpin RNA directed against TRPA1 loose the ability to develop tumor masses when injected into nude mice [177]. The effect of TRAP1 affording neoplastic potential is unfolding only upon entrance into the mitochondrial matrix and is firmly attributed to succinate accumulation and HIF1α stabilization through inhibition of SDH, even in the absence of hypoxic conditions [177]. More recently though, the group of Rasola also discovered that SDH inhibition by TRAP1 is oncogenic not only by conferring stabilization of HIF1α, but also through a build-up of antioxidant defenses [77], protecting against the opening of the mitochondrial permeability transition pore, a non-selective high-conductance channel which allows the flux of water and other molecules up to 1500 Da across the inner mitochondrial membrane [22], [27].

5.5. Succinate as an oncometabolite inhibiting α-KG-dependent dioxygenases

In the past few years, the concept of ‘oncometabolite’ has been introduced in the field of cancer metabolism. This term is reserved for succinate, fumarate and 2-hydroxyglutarate, the latter being a product of mutant isocitrate dehydrogenase [121]. These three metabolites are ‘connected’ through their structural similarity to α-ketoglutarate, a competitive inhibitor of a 60-strong family of α-ketoglutarate-dependent dioxygenases [120], [174]. Here, we will only consider the oncometabolic potential of succinate; for a comprehensive review regarding all three oncometabolites, the reader is referred to the review by Nowicki and Gottlieb [145].

Insufficiency of SDH activity in tumor tissues harboring SDH mutations (or TRAP1 overexpression) lead to accumulation of succinate. However, not only succinate increases, but also elevation in succinate/fumarate ratio can be detected in these tumors, suggesting that these molecules can be used as a metabolic biomarker for identification of SDHB/D associated Pheo/PGLs [110]. Succinate may function as a driver of tumorigenesis through multiple mechanisms.

5.5.1. Pseudo-hypoxia mechanism as a result of inhibition of PHD

The first evidence regarding the potential activation of the HIF1α pathway in SDHx related Pheo/PGL came from a case report showing accumulation of succinate and stabilization of HIF1α in an SDHD mutated pheochromocytoma [71]. A subsequent whole genome gene expression study performed in hereditary Pheo/PGLs showed that tumors with SDHB or SDHD mutations clustered in the same group as VHL-associated tumors, and that this cluster exhibited a significantly different gene expression profile compared to those detected in RET and NF1 associated tumors [51]. The common feature of overlapping genes was that most of them contain hypoxia response element (HRE) in their promoters. This site is where the transcription factor HIF1α binds. It was well known that the von Hippel–Lindau protein (pVHL) binds to HIF1α causing its proteasomal-mediated degradation [106]. Therefore, it was concluded that the main pathway contributing to tumorigenesis in SDHx mutated tumors was mediated by HIF1α. In addition, the similar phenotypes of tumors observed in VHL syndrome (including pheochromocytomas, paragangliomas and renal cell cancer) further supported this observation. In addition, Pollard et al. showed that tumors associating either with SDHx or FH (the gene encoding fumarate hydratase) mutations accumulated Krebs cycle intermediates and exhibited an increase of HIF1α expression [161] suggesting that there might be a link between succinate accumulation and HIF1α expression [180].

The mechanism leading to stabilization of HIF1α in SDHx mutated tumors was discovered by Selak et al. using siRNA mediated silencing of the SDHD in HEK293 cells [180]. They showed that the accumulated succinate itself can inhibit prolyl hydroxylase (PHD). This enzyme is responsible for hydroxylation of HIF1α causing its degradation. Therefore in SDH-mutated tumors, succinate accumulation through inhibition of PHD causes HIF1α stabilization (Fig. 1). Later, this group clarified that the pseudo-hypoxia observed in SDH-suppressed cells was independent from oxidative stress and was attributed solely to accumulation of succinate [181]. This result is not completely supported by others showing that reactive oxygen species (ROS) from impaired complex II may also contribute to tumorigenesis in SDHx-mutated tumors [74], [82], [186], [197].

5.5.2. Succinate and reactive oxygen species (ROS)

SDH has an important role in mediating ROS formation in mitochondria, see above under Section 4. Nevertheless, several groups working with SDH deficient models in Caenorhabditis elegans, yeast or hamster showed ROS production from this complex [74], [182], [186], [197]. In addition, very recently, Saito et al. [172] demonstrated in PC12 cells that SDHB silencing resulted in loss of complex II activity, followed by increased expression of tyrosine hydroxylase (TH; the rate-limiting enzyme in catecholamine biosynthesis) and catecholamine secretion. An elevated production of ROS, nuclear stabilization of HIF1α and increased expression of anti-apoptotic Bcl-2 were also detected. As a biological consequence, resistance against apoptosis was confirmed. These effects were abolished by pre-treatment with the ROS scavenger N-acetyl cysteine suggesting that increased ROS may function as a signal transducer necessary for induction of pseudo-hypoxic state [172]. These data further support earlier results which demonstrated that mutation of the SDHC [88] resulted in increased oxidative stress, apoptosis failure and tumorigenesis [76]. These biochemical consequences may explain the malignant potential observed in SDHB mutation carriers.

5.5.3. Epigenetic effects through inhibition of histone demethylases (HDM) and the ten-eleven translocation (TET) family of 5-methylcytosine (5mC) hydroxylases

As mentioned above, accumulation of succinate causes inhibition of PHD and consequently stabilization of HIF1α. However, succinate may inhibit other dioxygenases as well, some of which are directly involved in tumorigenesis. Histone demethylases (HDM) and the ten-eleven translocation (TET) family of 5-methylcytosine (5mC) hydroxylases (TET1-3) have been shown to be inhibited by succinate and fumarate accumulation [107]. These enzymes play a central role in epigenetic regulation, reviewed in [174]. TET enzymes are DNA demethylases catalyzing the oxidation of 5- methylcytosine (5-mC) to 5-hydroxymethylcytosine (5-hmC), the initial step in the DNA demethylation pathway [174]. Jumonji C domain-containing histone lysine demethylases (KDM2-7) are epigenetic regulators of chromatin. The first study demonstrating this mechanism was performed in a yeast model of paraganglioma, and the results were later confirmed in human embryonic kidney cells (T293 cells) [187]. Subsequently, by using pharmacological inhibition of SDH activity in various cell lines and siRNA-based gene silencing of SDHD and SDHB in HEK293 and Hep3B cells, it was shown that a decrease in SDH activity leads to an increased methylation of histone H3 which was reversed by overexpressing the H3K27me3-specific Jmjd3 histone demethylase [36]. In 2012 Xiao et al. confirmed these results in Hela cells and, in addition, using SDHx mutants demonstrated that indeed pathogenic FH and SDH mutants inhibit histone demethylation and hydroxylation of 5mC [224].

Translating these observations into clinicopathological studies, global DNA methylation profiles were evaluated in SDHx-associated and KIT (encoded by c-KIT gene) tyrosine kinase pathway-mutated GIST samples. It was demonstrated that SDH-deficient GIST had a significantly higher hypermethylation compared to the KIT-mutant group. This study detected the methylation pattern by high throughput methylation array using DNA samples isolated from formalin-fixed paraffin embedded tumor tissues [93]. This study was the first to prove that indeed the hypermethylator phenotype associates with SDHx mutations. A similar study was performed in a large cohort of Pheo/PGL tumor tissues. Methylome analysis categorized the tumor samples into three different classes presenting with distinct clinical features and mutational status. SDHx-related tumors displayed a hypermethylator phenotype, associated with downregulation of key genes involved in neuroendocrine differentiation. Moreover, by using an in vitro cell line system it was shown that decitabine (a synthetic DNA “demethylating” molecule) treatment reversed the biochemical and migratory phenotype caused by succinate accumulation in SDH-deficient mouse chromaffin cells [112]. In aggregate, it was concluded that succinate may be a link between genetic and epigenetic regulation and the hypermethylator phenotype may be responsible for the SDHB mutations-related malignancy [112]. A very recent study evaluated the expression of 5mC, 5-hydroxymethylcytosine (5hmC), TET1, H3K4me3, H3K9me3, and H3K27me3 on tissue microarrays containing Pheo/PGL (n = 134) and hereditary and sporadic smooth muscle tumors (n = 56) in comparison to their normal counterparts, by immunohistochemistry. 5hmC was lost in SDH- and FH-deficient tumors and loss of 5hmC was associated with nuclear exclusion of TET1, a known regulator of 5hmC levels [86].

From the above, it is concluded that SDHx mutations associated mechanisms leading to tumourigenesis is very complex, and several pathways and cross-talks have been proposed. By reviewing all these reports together with observations obtained in tumor samples we suggest that these mechanisms are not mutually exclusive. The key factor in tumorigenesis is the accumulation of succinate which in turn inhibits the α-KG-dependent dioxygenases leading to either stabilization of HIF1α, or hypermethylation of histones and DNA. Based on these results novel therapeutics can be developed which are desperately needed for malignant SDHB-associated cases. Finally, it is worth considering that the very same mechanisms operating in succinate-mediated tumorigenesis are also involved in the benefits of hypoxia as a therapy for mitochondrial diseases [89].