Cardiac metabolism as a driver and therapeutic target of myocardial infarction

Abstract Reducing infarct size during a cardiac ischaemic‐reperfusion episode is still of paramount importance, because the extension of myocardial necrosis is an important risk factor for developing heart failure. Cardiac ischaemia‐reperfusion injury (IRI) is in principle a metabolic pathology as it is caused by abruptly halted metabolism during the ischaemic episode and exacerbated by sudden restart of specific metabolic pathways at reperfusion. It should therefore not come as a surprise that therapy directed at metabolic pathways can modulate IRI. Here, we summarize the current knowledge of important metabolic pathways as therapeutic targets to combat cardiac IRI. Activating metabolic pathways such as glycolysis (eg AMPK activators), glucose oxidation (activating pyruvate dehydrogenase complex), ketone oxidation (increasing ketone plasma levels), hexosamine biosynthesis pathway (O‐GlcNAcylation; administration of glucosamine/glutamine) and deacetylation (activating sirtuins 1 or 3; administration of NAD+‐boosting compounds) all seem to hold promise to reduce acute IRI. In contrast, some metabolic pathways may offer protection through diminished activity. These pathways comprise the malate‐aspartate shuttle (in need of novel specific reversible inhibitors), mitochondrial oxygen consumption, fatty acid oxidation (CD36 inhibitors, malonyl‐CoA decarboxylase inhibitors) and mitochondrial succinate metabolism (malonate). Additionally, protecting the cristae structure of the mitochondria during IR, by maintaining the association of hexokinase II or creatine kinase with mitochondria, or inhibiting destabilization of FOF1‐ATPase dimers, prevents mitochondrial damage and thereby reduces cardiac IRI. Currently, the most promising and druggable metabolic therapy against cardiac IRI seems to be the singular or combined targeting of glycolysis, O‐GlcNAcylation and metabolism of ketones, fatty acids and succinate.


| INTRODUC TI ON
Cardiac metabolism changes rapidly during a sudden ischaemic episode of the heart, with the oxygen shortage repressing oxidative metabolism of fatty acids (FA), carbohydrates, ketones and amino acids, and activating anaerobic glycolysis to spare the use of limited oxygen. During reperfusion with the wash-in of oxygen and the wash-out of ischaemic metabolites, there is an abrupt normalization of intracellular pH and a specific start-up of oxidative metabolism for the various substrates, with ongoing changes in what is now aerobic glycolysis. These metabolic changes during ischaemia and early reperfusion are not merely passive bystanders, but determine to a large extent the actual injury developing in the heart following an ischaemic episode. Modulation of these metabolic changes, therefore, offers the opportunity for developing therapy against cardiac IRI, as was already shown by Sodi-Pallares in 1962 using potassium-insulin-glucose administration during myocardial infarction. 1 Since these earlier studies, it has become clear that metabolic therapy for IRI has travelled a rather bumpy road, without the development of a proven effective clinical metabolic therapy as of yet.
Here, we review the current literature concerning this topic, going from the well-known metabolic pathways to novel metabolic targets that go beyond the general metabolism of glucose and FA, and focus on important processes such as acidosis, ketone oxidation, succinate accumulation, mitochondrial F O F 1 -ATPase, energy transfer pathways, protein O-GlcNAcylation and acetylation as novel metabolic targets for treating IRI.

| G ENER AL A S PEC TS OF C ARD IAC ME TABOLIS M IN HE ALTHY HE ART
The healthy heart is a true omnivore in that it can degrade various energy-containing substrates. The major cardiac fuels for respiration are fats (triglyceride and long-chain fatty acids), carbohydrates (glucose, lactate and cardiac glycogen) and ketone bodies (acetoacetate and β-hydroxybutyrate) (Figure 1). Regulation of substrate use by the heart is to a large extent determined by the amount of substrate delivered to the heart (ie plasma concentration), the number of specific substrate transporters present in the cell membrane (CD36/Fat for fatty acids, GLUT1/4 for glucose and MCT1/2 for lactate and ketone bodies) and the activities of the metabolic enzymes and substrate/products/cofactors present in the enzymatic pathways. [2][3][4] It should thereby be realized that a high plasma concentration of one substrate usually competes and inhibits the use of other substrates.
For example, high plasma fatty acid levels will impair glycolysis and glucose oxidation, 4 or increasing plasma lactate levels will impair fatty acid oxidation and glycolysis. 5 Although in general fatty acids contribute more than carbohydrates to ATP generation in the heart, this depends critically on (patho) physiological, nutritional and hormonal state. For example, when insulin, lactate and fatty acids are present at normal physiological concentrations in the ex vivo-perfused rodent heart, the contribution of carbohydrates can be higher than that of fatty acids. [6][7][8] Ketone bodies, when provided at physiological plasma concentrations (<0.3 mmol/L), contribute less than 5% to cardiac ATP generation. 9 Each substrate is finally broken down to acetyl coenzyme A (acetyl-CoA) that feeds the tricarboxylic acid (TCA) cycle to produce reducing equivalents (NADH and FADH 2 ) that then feed the electron transport chain to build a proton gradient to drive the F O F 1 -ATP/synthase to make ATP. During normoxia, approximately 90% of the ATP produced is derived from mitochondrial oxidative breakdown of substrates, with cytosolic glycolysis only contributing ~5%-10% of total ATP production. Only during total ischaemia does glycolysis become the major supplier of ATP, when glycogen stored in the heart starts feeding glycolysis. The malate/ aspartate shuttle (MAS; Figure 1) is critical for maintaining glycolytic rate because it is the main pathway to recycle glycolysis-produced NADH through the mitochondria into cytosolic NAD + to maintain glycolytic activity.
It should be noted that for cardiac metabolism, it is not only the flux through the metabolic pathway that regulates cardiac function and physiology, but also the level of its intermediary metabolites.
Typical examples of regulating metabolic intermediates are as follows: a) acetyl-CoA, partly regulating the acetylation status and thereby function of proteins, b) succinate, build-up during ischaemia, of novel specific reversible inhibitors), mitochondrial oxygen consumption, fatty acid oxidation (CD36 inhibitors, malonyl-CoA decarboxylase inhibitors) and mitochondrial succinate metabolism (malonate). Additionally, protecting the cristae structure of the mitochondria during IR, by maintaining the association of hexokinase II or creatine kinase with mitochondria, or inhibiting destabilization of F O F 1 -ATPase dimers, prevents mitochondrial damage and thereby reduces cardiac IRI. Currently, the most promising and druggable metabolic therapy against cardiac IRI seems to be the singular or combined targeting of glycolysis, O-GlcNAcylation and metabolism of ketones, fatty acids and succinate.

K E Y W O R D S
ischemia, metabolism, mitochondria partly determining reactive oxygen production upon reperfusion, and c) UDP-GlcNAc, as end product of the hexosamine biosynthetic pathway. Although only a small part of glucose (<0.01% of glycolytic rate) 10 is shuttled into this pathway, the generated O-linked attachment of N-acetylglucosamine moiety onto proteins affects protein function significantly.
Finally, spatial and temporal distribution of metabolic enzymes can also be important effectors of cardiac metabolism. One classical example relates to hexokinase II (HKII) that shuttles between mitochondria and cytosol, depending on nutritional and pathophysiological status. When bound to mitochondria (mtHKII), HKII increases glycolysis, impairs in vivo cardiac oxygen consumption, impairs fatty acid oxidation, facilitates growth processes and protects mitochondria against injury. 7,11,12 Another example is creatine kinase (CK), the enzyme making ATP from phosphocreatine (PCr) and thereby functioning as an important local buffer for ATP breakdown. Cytosolic and mitochondrial isoforms of CK form a metabolic signalling network that is needed for rapid energy (ATP) transfer between ATP producers (mitochondria) and users (muscle contraction and ion pumps), preventing measurable decreases in ATP during changes in cardiac work. 13,14

| Glycolysis during ischaemia
During ischaemia, glycolysis is increased due to (i) increased glucose extraction from the blood (going from 1% during normoxia to 30% extraction during severe low flow), (ii) increased translocation of GLUT transporters to the sarcolemma, (iii) increased glycogenolysis, due to stimulation of glycogen phosphorylase A by diminishing glucose, ATP, glucose 6-phosphate and increases in AMP and Ca 2+ , (iv) removing citrate-mediated inhibition of the glycolysis-controlling enzyme 6-phosphofructokinase-1 (PFK-1) and (v) AMP-activated F I G U R E 1 Summary of the proposed pathways of cardiac metabolism covered in this review. Many of the discussed pathways show protection against IRI (green arrows) or are protective if blocked (red arrows). CK, creatine kinase; CPT, carnitine palmitoyltransferase; HKII, hexokinase II; MCT, monocarboxylate transporter; MPT, mitochondrial pyruvate transporter; mPTP, mitochondrial permeability transition pore; OGT, O-GlcNAc transferase protein kinase (AMPK) activation, resulting in both increasing GLUT transport capacity at the membrane and activation of 6-phosphofructokinase-2 (PFK2), resulting in fructose 2,6-bisphosphate production, the most potent allosteric stimulator of PFK1. [15][16][17][18] Activation of glycolysis during low-flow conditions makes use of all these mechanisms, whereas, during no-flow conditions, only mechanisms (iii), (iv) and (v) are active. Of note, although most pre-clinical studies have employed the no-flow ischaemia model, low-flow ischaemia may better reflect the clinical condition of myocardial infarction, where residual flow is often present. 19 Following long-lasting glycolytic activity, glycolysis can subsequently be inhibited by accumulation of its products: lactate, protons and NADH. 15 As long as glycolysis can proceed and generate ATP during the entire oxygen-limited condition, it offers protection against IRI. Because during low-flow conditions lactate and protons can still partly be removed and glucose uptake proceeds, inhibition of glycolysis is much delayed as compared to no-flow conditions. This explains why during even extreme low-flow ischaemia (5% of normal flow), an extended 40-min period of lowflow induces hardly any IRI (<3% cell death), 20 whereas 40-minute no-flow induces 20x more cell death. 21 Low-flow ischaemia does induce damage when glucose is replaced by another substrate or taken out of the perfusate, supporting the concept that it is the glycolytically derived ATP during ischaemia that protects against IRI. 22,23 Increasing glycolysis during low-flow ischaemia through, for example, increasing levels of glucose and insulin in the perfusate or boosting endogenous glycogen is associated with delayed contracture development and protection against IRI. 24 The start of contracture development during ischaemia is a read-out of the time when glycolysis stops 22 and the free energy of ATP hydrolysis, the ∆G ATP , falls below a critical level to support the ATPase activity of ion pumps and cross-bridge cycling. 26 Also during no-flow ischaemia activating or prolonging glycolysis is most frequently associated with protection against IRI. In line, AMPK-deficient mouse models are characterized by decreased glucose uptake and glycolysis, increased ATP depletion during ischaemia, more rapid and severe ischaemic contracture, increased cell death during reperfusion and poorer post-ischaemic contractile recovery. 27,28 Previous studies also clearly showed that either increasing pre-ischaemic glycogen 29 or activating glycolysis through redox control by niacin-induced lowering of NADH/NAD 30 resulted in delayed ischaemic contracture and reduced IRI. Similarly, the pharmacologic overactivation of AMPK reduces cardiomyocyte death and ameliorates post-ischaemic function recovery. 31,32 However, there are experimental conditions where increasing pre-ischaemic glycogen and/or a delayed contracture development were actually associated with increased IRI. 23,33,34 Although it is not completely clear what sets these conditions apart, it could be related to excessive glycogen loading (and thus excessive accumulation of glycogen breakdown products and consequently low pH during ischaemia) 23 due to elevated insulin before ischaemia (thereby inhibiting non-ischaemic glycogen breakdown) 16 as compared to no insulin before ischaemia. 33 The presence of more insulin before ischaemia can also result in impaired activation of the reperfusion injury salvage kinase (RISK) pathway during reperfusion, 35 possibly explaining the increased IRI in these experimental conditions. Mechanisms explaining increased infarct size with increased accumulation of glycolytic breakdown products during ischaemia are a) detachment of hexokinase II from mitochondria (mtHKII) due to high levels of G6P and low pH 11,33,36 and/or b) increased ischaemic Na + loading due to increased proton production by excessive glycogen breakdown, resulting in increased Ca 2+ overload upon reperfusion. 23

| Glycolysis during reperfusion
The strongest cardioprotective intervention for protecting against cardiac IRI, observed across all species, concerns ischaemic preconditioning (IPC). IPC relates to the application of short, non-lethal periods of ischaemia before the long, lethal period of ischaemia, resulting in a 75% reduction in infarct size. 37 IPC activates glycolysis during normoxia and reperfusion, through activation of AMPK and Akt, translocation of GLUT4 transporters to the cell membrane and translocation of HKII from cytosol to mitochondria ( Figure 1). 7,[38][39][40][41][42][43] Interestingly, no protection by IPC is observed in the absence of glucose. 39 Taken together, the data strongly suggest that IPC protective mechanism is mediated through activated aerobic glycolysis during early reperfusion. An increased glycolysis may mediate protection through increased activity of mtHKII, knowing that glucose phosphorylation by mtHKII is needed for protection against mitochondrial damage and cell death. 44,45 Additionally, older literature has suggested that protection by glycolysis is due to the preferential use of glycolytically produced ATP by ion pumps and restoration of ionic homeostasis during reperfusion. 46 Finally, increasing glycolysis during early reperfusion contributes to maintain a low pH, which in turn exerts multiple protective effects, including the prevention of mitochondrial permeability transition pore (mPTP) opening, 47 one of the mechanisms proposed for the cardioprotective effect of post-conditioning. 48 It is especially an increased ratio of glycolysis over glucose oxidation that induces this net proton production, 49 and may, for example, explain the decreased infarct size following IR in high-fat diet-induced obesity. 50 It is thereby noteworthy that this reduction in infarct size with increased glycolysis uncoupling from glucose oxidation is opposite to the negative effect this uncoupling exerts on cardiac mechanical function, resulting in a decreased cardiac performance and mechanical efficiency. 49 However, the decreased mechanical function upon reperfusion may actually contribute to the reduction in infarction, knowing that a slow recovery of energy requirement in early reperfusion is cardioprotective 51 and that there is often a dichotomy between recovery of mechanical function and cell death following IR. 52,53 Besides 'conditioning interventions' to activate aerobic glycolysis, several pharmacologic cardioprotective agents are also known to increase glucose uptake/glycolysis (metformin, insulin, volatile anaesthetics, adenosine, NO donors, fructose 1,6-diphosphate, nicotinamide mononucleotide (NMN), HIF1α stabilizers and AMPK activators). Some of these agents should ideally be applied at the end of the ischaemic period or at the onset of reperfusion, in order to prevent excessive glycogen accumulation, resulting in excessive accumulation of glycogen breakdown products during the ischaemic episode, thereby offsetting the beneficial effects of an activated glycolysis during early reperfusion.
In summary, most studies report that increasing glycolysis during ischaemia and early reperfusion confers protection against cardiac IRI, making glycolysis a potential important metabolic goal for IRI therapy.

| Hexosamine biosynthesis pathway (HBP)
Beyond glycolysis, glucose can be used by accessory metabolic pathways such as the hexosamine biosynthesis pathway (HBP) (Figure 1 [62][63][64][65] Notably, the treatment was also cardioprotective when applied at the onset of reperfusion, which increases its clinical relevance. 66 Finally, IPC performed in the ex vivo-perfused heart or in vivo by left anterior descending artery ligation stimulates protein O-GlcNAcylation and administration of an OGA inhibitor prior to surgery reduces infarct size. 66,67 This was further validated in humans submitted to remote IPC. 68 These data suggest that the cardioprotective effect of IPC may be mediated at least in part by an increase in OGT expression and activity. 67 The molecular mechanisms involved in the cardioprotective action of O-GlcNAcylation are not fully understood. It has several effects, including (i) the modification of Ca 2+ handling, 64  It would be worthwhile to investigate these novel strategies in humans.

| The malate-aspartate shuttle
In the healthy adult myocardium, the malate-aspartate shuttle (MAS) constitutes the main pathway for transportation of redox products from glycolysis in the cytosol into the mitochondrial matrix over the impermeable inner mitochondrial membrane ( Figure 1). 70 Under normal conditions, the shuttle capacity is high. The shuttle also meets elevated glycolytic demand in a variety of physiological and pathophysiological processes including cardiac hypertrophy. 71 Hence, shuttle activity is modifiable. Because of its central role as a regulatory mechanism in the energy metabolism of the cardiomyocytes, it constitutes a potential target for induction of cardioprotection.
The MAS has gained conceptual interest as a means to modify mitochondrial function, because transient shut down of metabolism by blocking the cytosolic-mitochondrial crosstalk via the MAS may induce cardioprotection ( Figure 1). As a proof of concept, transient aminooxyacetate (AOA) administration can induce reversible MAS inhibition. AOA is a non-specific competitive inhibitor of various amino acid transaminases 72 but in in situ heart models, AOA primarily leads to reversible inhibition of the MAS. 73,74 However, recent data also show that AOA directly reacts with alpha-keto acids (pyruvate, alpha-ketoglutarate, oxaloacetate, etc) to form stable oximes with unknown functional consequences as of yet. 75 Cardioprotection by MAS inhibition using AOA can be induced both prior to and during an ischaemic insult. 76,77 In the normal heart, MAS inhibition reduces mitochondrial complex I-linked respiration and glycolysis. 76 Even though the effect of MAS inhibition predominantly involves the glycolytic pathway, MAS inhibition also yields cardioprotection in a setting with free FA as additional substrate next to glucose. 78 MAS inhibition reduced succinate-induced ROS production (see 3.5.5. for mechanism) by limiting transport of substrate to the succinate build-up. 79 The attenuated succinate oxidation during reperfusion reduces reverse electron flow at complex I and hence reduces oxidative stress and cellular damage. This key mechanism is further supported by the ability of MAS inhibition to preserve post-ischaemic complex I respiration. 77 However, the mechanism does not explain that MAS inhibition also reduces infarct size when administrated in the late phase of an ischaemic event, when it is unlikely to reduce succinate substantially. 77

| Lysine acetylation of cardiac proteins
The adduction of acyl groups to protein lysine residues (acylation) is a biologically significant post-translational modification. Although many such acylations exist (eg succinylation, glutarylation and malonylation), here we focus on the most widely studied, acetylation.
Early studies on the cardiac acetyl-proteome revealed a preponderance of metabolic and mitochondrial targets, 80,81 leading to speculation that lysine acetylation is an important metabolic regulator.
However, recent studies in hyperacetylation models have revealed a minimal impact on bioenergetics, 82 consistent with studies at the molecular level. 83 An important consideration in this regard is acetylation stoichiometry, with little understood about the precise relationship between site occupancy and enzyme activity.
While there are thousands of kinases and phosphatases in a typical cell, enzymes that regulate acetylation number far fewer. Most work on acetyltransferases has focused on those in the nucleus that regulate histones (see Ref. 84 for recent review). However, more recently other acetylation pathways have been uncovered. In particular, GCN5L1 was identified as a mitochondrial lysine acetyltransferase expressed in the heart, 85 and its deletion confers a number of cardiac pathologic phenotypes. 86 Furthermore, direct non-enzymatic lysine acetylation from acetyl-CoA (and corresponding acylation from other acyl-CoAs) has been found to occur in the cell, 87 such that the concept of 'acyl-carbon stress' is now well established to be partly mediated by excessive lysine acetylation. 88,89 On the deacetylation side, in addition to the histone deacetylases (HDACs), the sirtuin (SIRT) class of NAD + -dependent lysine deacylases has emerged as key mediators of cardioprotection ( Figure 1). Of the 7 mammalian SIRTs, SIRT1 and SIRT3 are robustly expressed in the heart and have been shown to play a direct role in several cardioprotective paradigms. 39,[90][91][92][93][94] Unlike other tissues where SIRT1 is mostly nuclear, cardiomyocyte SIRT1 resides mostly in the cytosol 91 and is thus positioned to interact with the cardioprotective signalling machinery.
Furthermore, a decline in SIRT1 activity is proposed to play a role in the loss of IPC protection with ageing. 98 In the context of metabolism, it was shown that the metabolic remodelling that occurs in acute IPC is critically dependent on SIRT1. 99 SIRT1 is also proposed to be an intermediate signal in cardioprotection via other stimuli such as phosphodiesterase (PDE) inhibitors. 10 Given the original discovery of the sirtuins as putative mediators of the longevity effects of caloric restriction (CR) 11 it has also been posited that the cardioprotective benefits of CR may be mediated via SIRT1. 12 Similar cardioprotective benefits have also been described for mitochondrial SIRT3, 92,93 with its pharmacologic activation also shown to confer acute cardioprotection. 13 Among the more prominent SIRT3 deacetylation targets is the mPTP regulator cyclophilin D (CypD), with deacetylation at lysine 166 (mouse) required for the cardioprotective effects of SIRT3. 14 Although activation of SIRT1 or SIRT3 may appear attractive as a pharmacologic protective strategy, it should be cautioned that the specificity of many SIRT-activating drugs is uncertain, with stilbenes such as resveratrol plus several commercial SIRT1-activating drugs called into question. 15 In addition, inactivation of the NAD + -recycling enzyme NAMPT phenocopies SIRT1 inhibition in preventing cardioprotection, 93,106 suggesting that NAD + availability is a critical determinant for the cardioprotective efficacy of SIRT1. However, while the potential of boosting NAD + bioavailability via delivery of precursors such as nicotinamide mononucleotide (NMN) or nicotinamide riboside (NR) has shown considerable promise in pre-clinical studies, 95,107 the wide variety of non-SIRT metabolic pathways that rely on the NAD + /NADH redox couple 18 suggests that side effects of such compounds may outweigh their cardioprotective benefits. 19 Interestingly, it may well be that part of the cardioprotective effects of NAD + -boosting strategies mainly involves redox-controlled activation of glycolysis. 19 In line, we recently showed that increased protein acetylation by metabolic over-fuelling dramatically reduced both insulin-and AMPK-mediated glucose uptake in cardiomyocytes, presuming the importance of reducing protein acetylation for promoting glucose metabolism during IRI. 110,111 Overall, while promoting deacetylation, for example by targeting SIRT1 activation, is an attractive avenue for cardioprotection, considerable effort is required for the development of specific drugs to achieve this end.

| Oxygen consumption
A general characteristic of irreversible ischaemia, that is long (≥25 minutes) periods of ischaemia with cell death occurring, is the fast and complete recovery of oxygen consumption and therefore metabolic recovery during early reperfusion. At the same time, the recovery of mechanical function is lagging behind and still severely depressed. Thus, there is metabolic recovery that is uncoupled from mechanical recovery. In contrast, with reversible ischaemia, thus short (≤20 minutes) periods of ischaemia without cell death, recovery of oxygen consumption is much slower and still coupled to mechanical function following reversible ischaemia. [112][113][114] This fast recovery of post-ischaemic metabolism was associated with increased mitochondrial Ca 2+ in early reperfusion with irreversible ischaemia. [112][113][114][115] The fast recovery of oxygen consumption can possibly be explained by Ca 2+ activation of pyruvate dehydrogenase, resulting in increased carbohydrate (glucose, lactate, pyruvate) oxidation during the first hour of reperfusion. 116  as a primary mechanism to combat IRI.

| Glucose oxidation
Glucose oxidation, which actually is oxidation of the end product of glycolysis, that is pyruvate, occurs in the mitochondria. The flux from glycolysis to glucose oxidation is mostly controlled by the activity of the PDH complex located in the mitochondrial matrix. Glucose oxidation is completely shut down during ischaemia; its degree of reactivation during reperfusion depends, for example, on the availability of FA for oxidation because FA compete with glucose for oxidation.
During early reperfusion, FA oxidation is increased due to activation of AMPK by the ischaemia. 122,123 FA oxidation reduces glucose oxidation because of the Randle cycle, in which acetyl-CoA derived from FA oxidation inhibits PDH. 124 However, activation of PDH by the mitochondrial Ca 2+ overload occurring at reperfusion may overcome this inhibition and result in higher glucose oxidation than in normoxic conditions. 125 Thus, glucose oxidation is reduced during early reperfusion following moderate ischaemia, but may increase following more severe or longer ischaemia. 114,126 Several ex vivo (ie in isolated perfused hearts) and in vivo studies have shown that stimulation of glucose oxidation during reperfusion is associated with a better recovery of myocardial function and a lesser reperfusion injury. In the ex vivo setting, interventions at reperfusion that increase glucose oxidation and improve the recovery of myocardial function may directly target the PDH flux, by activating PDH activity with dichloroacetate, [127][128][129][130] or by the law of mass action, provisioning extra pyruvate 129 or increasing glucose uptake by administration of high glucose/high insulin at reperfusion 131 or through overexpression of the GLUT1 transporter. 132 Indirectly, inhibition of FA oxidation [133][134][135] or uptake 136,137 may prevent inhibition of glucose oxidation by reversing the Randle mechanism.
Importantly, IPC, the most robust cardioprotective intervention, was also shown to increase glucose oxidation during reperfusion, 138,139 although other studies found no changes. 140 A major limitation of ex vivo experiments is that they do not allow assessment of long-term post-ischaemic myocardial salvage. Indeed, the improved recovery of function observed during the short reperfusion period (usually one hour) might reflect only speeding up of the mechanical recovery, associated with the improved re-energization observed. [132][133]136 Nevertheless, several ex vivo studies observed a reduction of myocardial necrosis biomarkers associated with increased glucose oxidation in response to glucose addition 141 128,130 Although complete glucose metabolism, that is glycolysis + oxidation, is proton neutral, glycolysis not followed by pyruvate oxidation generates two protons. 150 During reperfusion, the glycolysis/oxidation uncoupling worsens, aggravating the proton overload, which drives the Na + and Ca 2+ overload leading to cardiomyocyte necrosis and/or opening of the mPTP. 151  In conclusion of this section, glucose oxidation appears to be a promising cardioprotection target that seems to have been overlooked by clinical studies. A search of the ClinicalTrials.gov database with the keywords 'myocardial infarction' or 'reperfusion injury' and 'glucose oxidation' failed to retrieve a single study. Acute stimulation of glucose oxidation at reperfusion with well-tolerated agents such as dichloroacetate or reconstituted HDL would be worth trying in a clinical context.

| Fatty acid metabolism
The energy metabolism in heart heavily relies on fat oxidation.  metabolism have resulted in disturbances in cardiac function. 160 This highlights the importance of FA as essential substrates for the heart energy production to maintain normal cardiac function under chronic conditions. Nevertheless, the use of small molecular inhibitors of enzymes in fatty acid transport and metabolism pathways has demonstrated that inhibition of FA oxidation in general reduces damage induced by IR [160][161][162][163][164][165][166][167] (Table 1). It should be noted, however, that effects of FA on cardiac IRI are critically dependent on FA concentrations, with detrimental effects commonly only observed at rather high levels (>0.8 mmol/L) of FA. 168 Although older literature has reported much higher FA plasma levels in cardiac patients, these high levels were likely the result of ongoing lipolysis in the test tube due to the use of heparin in these patients; preventing test tube lipolysis shows that these high FA levels are usually not observed during human in vivo IR episodes. 169 Finally, FA may also contribute to cardiac IRI due to its dislodging effects on hexokinase II-mitochondrial binding in the heart. 170

| Ketones
Beta-hydroxybutyrate (BHB) and acetoacetate (AcAc) represent the two main ketone bodies. 171 BHB, which is produced by degradation of FA in liver and then transported to extrahepatic tissues (including the heart), is one of the important metabolic substrates for energy production during fasting. 172 BHB is not only a metabolic intermediate, but also possesses a variety of signalling functions. 173 It has been shown to inhibit Class I histone deacetylases (HDACs), increasing therefore histone acetylation, and thereby induces the expression of genes that restrain oxidative stress. 174 BHB also suppresses sympathetic nervous system activity and reduces total energy expenditure and heart rate by inhibiting short-chain fatty acid signalling through G protein-coupled receptor 41 (GPR41). 175  oxidation, but less when compared to glucose. 178,179 Additionally, the oxidation of ketone bodies also yields potentially higher energy than fatty acid oxidation, keeping ubiquinone oxidized, which raises redox span in the electron transport chain (ETC) and makes more energy available to synthesize ATP. A recent ex vivo cardiac study employing direct measurements of cardiac efficiency showed no effects of ketones on cardiac efficiency. 9 However, increasing the supply of ketones to the heart did increase total ketone body oxidation without a decrease in any other metabolic pathway. 9 Therefore, ketone bodies are considered an alternative and efficient energy source in myocardium, especially in failing hearts. 180 Increased circulating ketone bodies have been previously reported in patients with congestive heart failure. 181 In addition to the beneficial effects of ketone bodies in heart failure, they have been shown to have beneficial effects in IRI by reducing myocardial infarct size, either by increased levels during fasting or when they were administered exogenously. Because of concentration-dependent dynamics, increases in ketone bodies during fasting can elevate the rate of BHB utilization. 182 It has been reported that fasting increased the myocardial BHB/AcAc ratio reflecting altered mitochondrial redox state and fasting of rats for only 24 hours improved the post-ischaemic recovery of contractile function and reduced the lactate dehydrogenase release in isolated hearts subjected to global IR. 183 In another study, short-term fasting increased the concentration of BHB and BHB/AcAc ratio compared to controls, limited the infarct size and reduced the total number of premature ventricular complexes and the duration of ventricular tachycardia occurring at early reperfusion. 184 However, low concentration of endogenous ketone bodies failed to preserve the myocardial ATP levels whereas exogenous supplementation (to 40 times the original concentration) prevented the loss of ATP by ischaemic injury. 185 Administration of exogenous BHB 60 minutes before the start of ischaemia reduced in vivo myocardial infarct size and apoptosis in rats subjected to IR. 186 It was recently demonstrated that starting in vivo BHB treatment at reperfusion and continuing administration for the next 24 hours of reperfusion using minipumps reduced infarct size, attenuated apoptosis in myocardium and preserved cardiac function of IR in mice. The above-mentioned beneficial effects were attributed to reduced mitochondrial formation of ROS, enhanced ATP production, attenuated mitochondrial swelling and partly restored mitochondrial membrane potential in myocardium. 187 In summary, in the experimental IRI context, ketone bodies may confer cardioprotective effects possibly due to altered mitochondrial redox state resulting from increased ketogenesis, up-regulation of crucial OXPHOS mediators and reduction of oxidative stress.

| Succinate and ROS
A new mechanism is emerging whereby the production of mitochondrial ROS is considered a highly orchestrated, metabolite-driven process early in IRI. The citric acid cycle metabolite, succinate, is extensively accumulated during ischaemia and is rapidly oxidized upon reperfusion. 79,188,189 In vivo, succinate likely accumulates via the reduction of fumarate by succinate dehydrogenase (SDH or complex II) reversal. A lack of the terminal electron acceptor, oxygen, maintains a reduced CoQ pool, and additionally, the pH of ischaemic tissue is lowered. 51,79,190 Upon reperfusion, succinate is rapidly oxidized to fumarate, and together with ETC activity restarting, the reduced CoQ pool provides electrons for the ETC complex to proton pump, establishing a large proton motive force (Δp). The large Δp and highly reduced CoQ pool, together with depleted adenine nucleotides, drive reverse electron transport (RET) through mitochondrial complex I, resulting in the production of superoxide at the flavin mononucleotide (FMN) site ( Figure 1). 191,192 The mitochondrial ROS produced, together with impaired calcium handling, activate downstream pathways, resulting in mitochondrial permeability pore formation and, ultimately, cell death.
Not all of the succinate that has accumulated at ischaemia is oxidized by SDH, as it has been suggested that a proportion is released from the cell. 75 Succinate was significantly elevated in the blood of patients with an acute ST-elevation myocardial infarct 193 suggesting its release into the bloodstream upon reperfusion. This opens the intriguing possibility that changes in mitochondrial metabolites during IRI could be involved in paracrine signalling, complementary to the signalling role succinate plays in the immune system. 194 Accumulation of succinate, with its derivative succinyl-CoA, also leads to protein succinylation. 195 The Sirt5, which has limited deacylase activity, also catalyses the removal of succinyl groups from proteins. In line, it has been shown that the increase in IRI in The concept that FoF1-ATP/synthase is the anatomical support for H + dissipation and cell death integrates other well-accepted mPTP regulators, 210 including CypD, which has been recently proposed to reduce mito-ATP synthase supramolecular assembly, thereby increasing mPTP opening probability. 211

| Energy transfer pathways
Mitochondria are central players of cellular energy metabolism, especially in striated oxidative muscles and heart. This is much more transfer pathways ensure precise feedback signalling between contraction workload and oxygen consumption in mitochondria. Each sarcomere has its own corresponding mitochondrion, which together with phosphotransfer system and feedback metabolic signalling creates the intracellular energy unit (ICEU). [214][215][216] An important characteristic of the heart is its metabolic stability, as reflected by the apparent invariability of intracellular concentration of ATP and phosphocreatine (PCr) in spite of the variable workloads, corresponding rates of ATP oxidative synthesis and myofibrillar hydrolysis. 217 Under conditions of total ischaemia, the PCr concentration falls rapidly and heart contraction ceases, but ATP concentration stays almost stable decreasing only by 10% at the end of the first minute of ischaemia. 217 Colocalization of BetaII-tubulin and VDAC in heart muscle is functionally related to the ability of creatine to stimulate OXPHOS due to functional coupling between mitochondrial creatine kinase (MtCK) and adenine nucleotide translocase (ANT). 218 Key events observed after acute ischaemia-reperfusion (IR) and chronic ischaemia are the decrease (or loss) in the stimulatory effect of creatine and decrease in diffusion restrictions for ATP and ADP at the level of the mitochondrial outer membrane (MOM), which is mediated by BetaII-tubulin 219,220 The disruption of mitochondrial interactions with cytoskeleton will result in decreased intracellular compartmentalized energy transfer and the loss of probability of interaction between mitochondria and BetaII-tubulin.
In adult cardiomyocytes, octameric mitochondrial creatine kinase (MtCK) binds electrostatically to the negatively charged cardiolipins of the mitochondrial inner membrane sharing the same cardiolipin patches with ANT. 221 It is possible that the unaltered octameric isoenzyme of MtCK ensures the stability of its molecular interaction with ANT after IR. 220 Alternatively, the IR-induced alteration of mitochondrial increased MOM permeability might influence kinetic properties of the MtCK. 220 The predominant HK isoform in adult heart, HK2, dynamically shuttles between the mitochondria and cytoplasm, resulting in increased glycolysis when bound to mitochondria. 222 Besides that, mitochondrially bound HK2 is an important player in the field of voltage-dependent anion channel (VDAC) interactions with regulatory proteins and its functional coupling with OXPHOS. It was recently shown that IR disrupts interactions between VDAC, ANT and HK2 through nitration of tyrosine residues in VDAC and ANT, contributing to mitochondrial and cellular dysfunction following IR. 223 Contact sites between MIM and MOM have a substantial role in transporting ATP, generated within the mitochondria, to the cytosol as PCr. Binding between CK and HK2 may stabilize these contact sites, and loss of HK2 during ischaemia leads to contact site breakage and decreased rates of extramitochondrial PCr synthesis. 36,224,225 The interplay between energy transfer pathways and different binding sites for tubulin and hexokinase to VDAC may be one of the targets of IPC.
The connections of mechanisms of energy transfer pathways and cardioprotection are not clear yet; one component participating in this system is the AK system together with K ATP channel. The AK-catalysed phosphotransfer system would promote K ATP channel opening primarily by accelerating conversion of ATP to ADP, whereas CK systems would predominantly facilitate conversion of ADP to ATP and K ATP channel closure. 226 The alterations in MOM permeability for adenine nucleotides seem to be an important feature of cardiac ischaemic and IR injuries, as it regulates the energy transfer pathways. IPC induced redistribution of high-energy phosphoryl transfer and increased phosphotransfer reactions (creatine kinase, glycolysis), leading to improved intracellular metabolic communication and preservation of cellular ATP synthesis and ATP consumption following IR. 227 These phosphoryl fluxes correlated tightly with post-ischaemic functional recovery 227 and pre-conditioning-induced energetic remodelling, improving contractile performance following IR. The study of intracellular phosphotransfer reactions is still not fully unravelled and will need more sophisticated studies, before it can be used in the development of an injury-tolerant state in cardiomyocytes.

| CON CLUS ION
We have outlined some of the important metabolic changes occurring during ischaemia and subsequent reperfusion. While many aspects happen in the entire cell, the mitochondria are a clear focus within many of these metabolic changes. Such changes include alterations in fatty acid and succinate metabolism, along with F O F 1 -ATP/synthase activity, resulting in increased mitochondrial ROS production and subsequent opening of mPTP. Additionally, glycolysis, hexosamine biosynthesis, glucose oxidation ketone metabolism and the malate/aspartate shuttle are all processes which directly affect metabolite levels and mitochondrial pathways.
All of which have shown to elicit cardioprotective effects when altered.
While many of the outlines metabolic processes are ideal drug targets for ischaemia/reperfusion injury, further studies are necessary to fully understand underlying mechanisms and establish potential therapies. In this context, it is vital to test possible pharmacologic interventions at a clinically relevant time-point at the end of ischaemia as well as in pre-diseased and aged models. 228 Some of the outlined mechanisms have already been shown to be relevant in man 193 and targeted successfully in many species, including large animals. 189,197 Others are still awaiting a relevant drug target suitable for translation in patient with an acute MI. Recent methodological advances in detecting metabolic changes within the heart will make these efforts easier to achieve. Furthermore, the observed metabolic changes are not limited to cardiac I/R injury, but could play important roles in many physiological and pathophysiological situations, such as exercise, inflammation and cancer.

D I SCLOS U R E S
None.

AUTH O R CO NTR I B UTI O N S
Each author wrote part of the review and edited the entire manuscript.