11 Myocardial energetics

The heart is an example of a specialized type I (‘red’) striated muscle in that it is continuously active and reliant principally on aerobic metabolism for its energy supply. In normal humans at rest, the heart extracts 60–65% of O₂ available in the coronary circulation. This corresponds to an O₂ utilization rate of about 4.5µmol/min per gram wet weight, which may increase by three- to fourfold during exercise. This maximal physiological O₂ uptake is probably higher than any other organ. The majority of O₂ is utilized by the ventricles, particularly by the left ventricle because of its greater mass and pressure–volume product than the right ventricle. Although the heart (like all organs) is composed of several different cell types, ventricular cardiac myocytes (the contractile cells of the heart) constitute about 75% of the heart mass. These myocytes possess a well-developed myofibrillar apparatus and a capacity for aerobic metabolism that fit them for their major role in vivo, namely the rhythmic contraction which provides the force needed for the ejection of blood from the ventricles.

Although much of the myocyte volume is taken up by myofibrils, the mitochondria constitute about 30%. In the heart, these subcellular organelles regenerate the bulk of ATP from ADP and inorganic phosphate (P_i). ATP is an ‘energy transducing molecule’ which couples the energy available from fuel metabolism into external work (principally myofibrillar contraction but also processes involved in ion transport and biosynthetic pathways). Mitochondria possess an outer (OMM) and an inner (IMM) mitochondrial membrane, and the latter encloses the mitochondrial matrix into which multiple invaginations of the IMM (the cristae) protrude. The division of the mitochondria into the matrix and intermembrane space is functionally important in ATP regeneration. The ATP content of the heart is only sufficient to allow contraction for a few beats, and its supplies of endogenous fuels (e.g. glycogen, triglycerides) are limited given the amount of work it has to perform.

Thus, in order to maintain fuel oxidation, ATP regeneration, and muscle contraction, a highly developed coronary circulation and uninterrupted coronary blood flow are necessary to ensure adequate delivery of O₂ and fuels, and to remove the product of aerobic metabolism, CO₂. However, like all tissues, the heart contains phosphocreatine (PCr or creatine phosphate) which acts as a short-term ATP buffer during rapid increases in the rate of ATP utilization. Thus, under ‘stressed’ conditions, PCr transphosphorylates ADP to ATP in an equilibrium reaction catalysed by creatine kinase (Equation 11.1):

Even so, the concentration of PCr is only about twice that of ATP, and the utility of the creatine kinase reaction is limited to short-term buffering of ATP concentrations.

The heart is omnivorous and utilizes any metabolic fuel presented to it, within the constraints of metabolic regulation. The major substrates for oxidation in humans are lipid-derived fuels (principally long-chain fatty acids (LCFAs) such as palmitate, but also triglycerides and ketone bodies (acetoacetate and its reduction product, 3-hydroxybutyrate) in specialized circumstances) and the carbohydrate-derived fuels (glucose, lactate, and pyruvate). Although the heart can oxidize amino acids, these are of minor importance. The accepted dogma is that the human heart relies predominantly (70%) on lipid-derived fuels for its energy supply,¹ but the majority of investigations have involved postabsorptive subjects where lipid metabolism is generally more favoured than in the immediately postprandial state.

The initial stages of glucose metabolism are exclusively cytoplasmic (Fig. 11.1). Glucose is transported across the cardiac myocyte plasma membrane by two carriers, namely the type 1 and type 4 glucose transporters (GLUT1 and GLUT4).² GLUT1 provides a basal constitutive component whereas GLUT4 mediates inducible glucose uptake. Insulin increases glucose uptake by recruiting intracellular endosomal GLUT4 to the plasma membrane. Intracellular glucose is mainly used to provide energy for contraction but, in addition, can be polymerized to the intracellular storage carbohydrate glycogen. Although the heart has a limited capacity to store glycogen and its breakdown may not be quantitatively significant under normal conditions, it may provide energy for a limited period in pathological conditions (e.g. during myocardial infarction, when the supplies of exogenous fuels and O₂ are disrupted).

Each molecule of glucose is degraded through the glycolytic pathway to pyruvate, the chemical energy released allowing the net regeneration of two molecules of ATP (from ADP) per glucose molecule utilized. This so-called ‘substrate-level phosphorylation’ occurs at the phosphoglycerate kinase and pyruvate kinase steps. Additionally, two molecules of the electron acceptor nicotinamide adenine dinucleotide (NAD⁺) are reduced to NADH. Glycolysis can occur anaerobically (see below) but this is inefficient in terms of the quantity of ATP regenerated compared with that available from complete oxidation of glucose. Under the aerobic conditions normally existing in the heart, pyruvate is transported into the mitochondria, and the glycolytically derived electrons (as NADH) enter on a ‘shuttle mechanism’ (the malate/aspartate shuttle) regenerating cytoplasmic NAD⁺.

The heart can metabolize circulating lactate formed from anaerobic metabolism in other tissues. Lactate is first oxidized to pyruvate by the cytoplasmic enzyme, lactate dehydrogenase (Equation 11.2):

The pyruvate thus formed and the reducing equivalents enter oxidative metabolism in the same way as glycolytically derived pyruvate and NADH. Under anaerobic or hypoxic conditions where oxidative metabolism is reduced, the lactate dehydrogenase reaction is reversed, leading to formation of lactate and regeneration of cytoplasmic NAD⁺.

In the mitochondria, pyruvate is oxidized by the pyruvate dehydrogenase multienzyme complex (PDH-MEC) to acetyl-coenzyme A (acetyl-CoA) and CO₂, with reduction of NAD⁺ to NADH.³ This is a critical regulatory step in humans and many other organisms since it is physiologically irreversible and commits pyruvate (hence glucose) to oxidation. Before the PDH-MEC stage, glucose can be hepatically or renally resynthesized from lactate or pyruvate by gluconeogenesis. The activity of PDH-MEC is therefore subject to stringent regulation that couples its activity to the nutritional state of the animal. Thus, in the postabsorptive or fasted state, its activity is decreased to allow glucose conservation for tissues with an obligatory requirement for exogenous glucose, and this is brought about by increased reliance on lipid-derived fuels, with the converse occurring on feeding. This is one of the facets of the so-called glucose–fatty acid cycle or Randle cycle^4,–6 in which utilization of lipid fuels during fasting suppresses glucose utilization (see Fig. 11.1).

Acetyl-CoA then enters the nine-stage tricarboxylic (TCA) cycle (also known as the citric acid or Krebs cycle) by condensing with oxaloacetate to form citrate (the citrate synthase step) and CoA (Figs 11.1 and 11.2). By decarboxylation and oxidation (Fig. 11.2), two carbons of citrate are lost as CO₂, and the electrons are used to reduce O₂ to H₂O. The intermediate steps in the TCA cycle oxidations involve principally NAD⁺ as the initial e^– acceptor, and these reducing equivalents are used to drive ATP regeneration. The CO₂ lost originates from the oxaloacetate moiety of citrate rather than acetyl-CoA. The carbons donated by acetyl-CoA become part of the oxaloacetate backbone after the first turn of the TCA cycle and thus complete loss of the ‘acetyl-CoA carbon’ as CO₂ may require several turns of the cycle. However, intermediates of the TCA cycle are often used in biosynthetic reactions, and so the carbon may be retained.

Triglycerides, LCFAs, and ketone bodies are all capable of providing energy for the heart. LCFAs (principally palmitate) are present in the plasma either noncovalently bound to albumin or covalently bound as triglycerides which are in turn complexed with apolipoproteins. Ketone bodies are synthesized hepatically from LCFA but are present in the plasma only at low concentrations under ‘normal’ conditions. Compared with LCFA, they represent a relatively soluble, readily-diffusing, utilizable, nontoxic fuel and their concentrations increase during starvation or after exercise. Postexercise ketosis is effectively an ‘overshoot phenomenon’ whereby high plasma concentrations of LCFA used to fuel exercise continue to give rise to ketone bodies. In pathological conditions (e.g. untreated diabetes mellitus), their plasma concentrations rise to a much greater extent than in starvation. This results from the more extreme increases in plasma LCFA.

Like glucose, albumin-bound LCFAs enter the cardiac myocyte by a transporter-mediated process (principally the transmembrane CD36 protein in conjunction with a plasma membrane LCFA binding protein), and diffusion within the cytoplasm is facilitated by a cytoplasmic LCFA binding protein(s) (Fig. 11.1).² There are other pathways (diffusion through the plasma membrane, other LCFA transport proteins), but these are less important in the heart than CD36. Triglycerides are hydrolysed by the ecto-enzyme lipoprotein lipase on the capillary wall to form LCFAs (which then enter the myocyte) and glycerol, potentially a substrate for resynthesis of glucose by gluconeogenesis.

LCFAs and ketone bodies can only be catabolized aerobically, and their catabolism takes place exclusively in the mitochondria. LCFAs are first ‘activated’ in the cytoplasm by the formation of a covalent thioester bond with CoA (compare acetyl-CoA), and then cross the mitochondrial membranes. This involves the carnitine palmitoyl transferase I/carnitine palmitoyl transferase II system in the OMM and the IMM, respectively. The OMM transferase is an important point of control of LCFA metabolism. Once in the mitochondrial matrix, LCFA-CoA is reformed, and two-carbon fragments are successively removed as acetyl-CoA from LCFA (as LCFA-CoA) in a series of reactions known generically as β-oxidation. Ketone bodies, as acetoacetyl-CoA, are metabolized by a somewhat different process, to acetyl-CoA. Acetyl-CoA formed from β-oxidation or ketone body utilization is then oxidized through the TCA cycle (Figs 11.1 and 11.2). Because the commonest LCFAs contain even numbers of carbon residues which are ‘lost’ during the TCA cycle, these LCFAs generally cannot be utilized for net production of glucose.

The mechanism by which reduction of O₂ to H₂O by NADH is coupled to ATP resynthesis (electron transport and oxidative phosphorylation) remained a mystery long after the metabolic pathways of fuel metabolism had been elucidated. Initial work primarily attributed to Peter Mitchell led to the development of the ‘chemiosmotic hypothesis’.^7,8 Here, the transport of e^– down an electrochemical gradient of carriers on the internal face of the IMM is coupled to electrogenic vectorial pumping of protons into the space between the IMM and OMM. This sets up a ‘proton motive force’ which is energized by both a pH gradient and a potential difference across the IMM. The electrons pass to the next carrier, and the protons are vectorially directed into the intermembrane space. The ultimate e^– acceptor is molecular O₂ which is reduced to H₂O. The energy of the proton motive force flowing back down the energy gradient across the IMM is utilized to drive the resynthesis of ATP from ADP and P_i by the IMM ATP synthase. One consequence of the chemiosmotic hypothesis is that a precise relationship between O₂ consumed and ATP resynthesized (the P/O ratio, ATP molecules resynthesized from ADP/oxygen atom reduced) does not necessarily exist. Previously, the P/O ratio was thought to be stoichiometrically fixed to 3. However, the P/O ratio would remain at about 3 under normal circumstances, simply because of the free energy available and the efficiency of the process. The bulk of the mitochondrially generated ATP then exchanges with ADP in the cytoplasm, providing the energy for myofibrillar contraction and other processes. This exchange is mediated by the adenine nucleotide translocase of the IMM, though there is evidence that PCr and creatine kinase are also involved in transferring so-called ‘high-energy phosphate’ from the mitochondria to the cytoplasm.⁹

The combined metabolism of glucose through anaerobic glycolysis and the TCA cycle allows a much greater release of free energy than metabolism through anaerobic glycolysis alone. Whereas metabolism of glucose to lactate results in the regeneration of 2 mol ATP/mol glucose, the complete oxidation of pyruvate to CO₂ and water yields about 18 mol ATP/mol pyruvate (assuming a P/O ratio of 3). Thus, the complete oxidation of 1 mol of glucose results in the regeneration of (2 + 18 + 18), i.e. 38 mol, ATP and is thus far more efficient in terms of chemical energy released than its metabolism to lactate. However, it is absolutely dependent on a well-developed blood supply to provide not only O₂ but also glucose. For the LCFA palmitic acid [CH₃(CH₂)₁₄·CO₂^–], 2 mol ATP equivalents are used in the activation of 1 mol palmitate to palmitoyl-CoA. Thereafter, seven rounds of β-oxidation will regenerate approximately 35 mol ATP/mol palmitate, and produce 8 mol acetyl-CoA. The ensuing eight turns of the TCA cycle will regenerate about 96 mol ATP/mol palmitate, a net resynthesis of about 129 mol of ATP/mol palmitate.

LCFA and glucose do not produce the same amount of ATP/O₂ reduced to H₂O (though we are assuming a P/O ratio of 3). For complete oxidation, glucose requires 6 mol O₂/mol glucose, thus providing 6.33 mol ATP/mol O₂. In contrast, the complete oxidation of palmitate requires 23 mol O₂/mol palmitate, thus providing 5.61 mol ATP/mol O₂. In other words, glucose offers an ‘O₂ advantage’ of about 10%. Whilst this may not be a significant factor in normal hearts, it might offer a significant advantage in situations where O₂ delivery is compromised (e.g. heart failure). In fact, experimental studies in pigs¹⁰ suggest that glucose may offer as much as about a 50% advantage over LCFAs in terms of the myocardial O₂ uptake:left ventricle pressure–volume product because of factors additional to those conferred by simple energetic considerations. For example, ‘uncoupling’ of oxidative phosphorylation (reducing the proton motive force but maintaining O₂ consumption) by known endogenous uncouplers (for example, LCFAs, LCFA-CoA, and LCFA-carnitine, which are detergents and disrupt membrane structure) could reduce the P/O ratio.

Although the terms ‘hypoxia’ (inadequate O₂ supply) and ‘ischaemia’ (inadequate blood supply) are used interchangeably, they are not necessarily synonymous. Hypoxia may occur in the absence of ischaemia, whereas ischaemia inevitably involves an element of hypoxia. Ex vivo, the glucose-perfused heart can survive a degree of hypoxia,¹¹ and this is presumably also true in vivo. However, the heart cannot survive total anaerobiosis because anaerobic carbohydrate metabolism would have to increase about 20-fold to satisfy cardiac energy requirements. This is impossible given the maximal activity of the glycolytic pathway. As O₂ tension falls, glucose uptake and glycolytic flux are increased, and glucose is increasingly metabolized anaerobically. Glycolytic pyruvate is reduced by NADH to lactate by lactate dehydrogenase, and the NAD⁺ originating from glycolysis is regenerated (i.e. Equation 11.2 is reversed, being driven towards lactate production by the increased pyruvate, NADH, and H⁺ concentrations). Lactate production increases but, assuming it can be released into the perfusion medium, this does not present a major problem. The same is probably true in vivo where lactate can be released into the coronary circulation if perfusion is maintained. Two mechanisms operate in the short term to maintain ATP. First, ATP is buffered by the creatine kinase equilibrium (Equation 11.1). Any intracellular acidification through lactate production drives the equilibrium to the right, favouring ATP resynthesis. However, as stated above, the operation of the creatine kinase equilibrium is probably only important in short term ATP buffering. Secondly, glycolytic flux is increased. Here, the adenylate kinase equilibrium (Equation 11.3) plays a central role:  

Numerous NMR studies have shown that the concentration of ATP in muscle cytoplasm is about 8–10 mM. The total concentration of ADP is about a tenth of this (though much of the ADP pool exists in a protein-bound state in muscle). From mass action considerations, a small percentage fall in ATP and the necessarily larger percentage increase in free ADP will be ‘amplified’ in terms of percentage increases in AMP concentrations. The situation is not as simple as this because AMP concentrations are additionally affected by changes in pH, PCr, free Mg²⁺, and P_i. At physiological ATP concentrations, AMP is a powerful activator of 6-phosphofructo-1-kinase (PFK1), an important rate-controlling step of glycolysis, leading to an increase in glycolytic rate.

Glycogen breakdown is also increased by AMP (and catecholamine release) by activation of the glycogen phosphorylase step. The phosphorylated monosaccharidic glucose subunits enter glycolysis as glucose 6-phosphate (Fig. 11.1). Although each glycogen-derived ‘glucose molecule’ produces three molecules of ATP, one ATP per glucose molecule is expended in glycogen synthesis, so the net production of ATP per glucose molecule remains at two (as in simple glycolysis).

The most obvious pathological manifestation of cardiac hypoxia occurs during coronary artery disease and myocardial ischaemia. Myocardial infarction may totally deprive part of the heart of its blood supply, especially if there is no developed collateral circulation. In total ischaemia, the only intracellular metabolically available fuels are the limited free intracellular glucose and the glucose stored as glycogen. Essentially the same mechanisms operate to maintain ATP concentrations as those operating in hypoxia.

However, because lactate cannot be removed effectively from the myocyte, intracellular pH (pH_i) falls, and NADH accumulates. These eventually produce a profound inhibition of glycolysis because the 3-phosphoglyceraldehyde dehydrogenase step is inhibited by low pH and by product inhibition by NADH. In addition, binding of Ca²⁺ to the protein regulating myofibrillar contraction (the ‘Ca²⁺ sensor’, troponin C) is pH-sensitive and is disfavoured at lower pH_i values. Thus, a higher Ca²⁺ concentration is required to maintain an equivalent contractile force. This has energetic consequences for the ATP-consuming Ca²⁺ pumping involved in contraction, in that increased ATP consumption is necessary. Maintenance of ATP regeneration is impossible, ATP concentrations eventually fall, and contractile activity ceases.

Although complete four e^– reduction of O₂ to H₂O is the norm in oxidative phosphorylation, a small proportion of O₂ may be incompletely reduced by a single e^– to form superoxide anion radicals (O₂^{• –}) even under normal aerobic conditions.¹² This potentially exposes tissues to cytotoxic oxidative stress from O₂^{• –}and O₂^{• –}-derived reactive oxygen species (ROS). The proportion of O₂ reduced to O₂^{• –}  in vivo is difficult to estimate, but, in isolated mitochondria, it may be as high as 1–2%.¹² This is probably a maximum figure because of the optimal experimental conditions used and, in vivo, O₂^{• –} formation is probably much less. However, given that cardiac myocytes contain abundant mitochondria that are continuously respiring, it is unlikely to be negligible. Furthermore, production of ROS rises during ischaemia and is increased further in any subsequent reperfusion.^12,13. There are two major fates of O₂^{• –}. First, although it is not a particularly reactive species itself, O₂^{• –} can react very rapidly with NO, which is itself a radical (more accurately NO^•), to form highly reactive peroxynitrite anions (ONOO^–) which damages macromolecules. Secondly, O₂^{• –} spontaneously dismutes into H₂O₂ and O₂. Dismutation (step (1) in Equation 11.4 below) is enhanced by superoxide dismutase (SOD), of which there are two isoforms, mitochondrial Mn²⁺SOD or SOD2 and cytoplasmic Cu²⁺/Zn²⁺SOD or SOD1, thereby reducing formation of peroxynitrite. The importance of SOD2 is demonstrated by the finding that globally null SOD2 mice die in the early postnatal period, one of the pathological findings being a cardiomyopathy with clear fibrosis of the endocardium.¹⁴ H₂O₂ is less reactive than O₂^{• –}, and is rapidly decomposed into H₂O, and O₂ by catalase (step (2) in Equation 11.4):   

Even though it is rapidly destroyed, production of H₂O₂ may still be cytotoxic because, in the presence of transition metal ions (such as Fe²⁺ present within the myocyte), it will produce highly reactive hydroxyl radicals by the nonenzymic Fenton reaction. These species will damage macromolecules and react with metabolites.

The reasons underlying the transition of the heart into failure are probably manifold and multifactorial, but one aspect of the process involves cardiac energetics. Some of the changes seen in acute hypoxia and ischaemia contribute but there are additional changes, notably the long-term ‘metabolic remodelling’ of the heart. Partly because of ATP yield/O₂ (see above), the failing heart favours utilization of carbohydrates over lipid fuels and this involves changes in patterns of gene expression over the longer term.

In human HF, the levels of expression of genes associated with β-oxidation (medium-chain acyl-CoA dehydrogenase, long-chain acyl-CoA dehydrogenase), mitochondrial membrane fatty acid transport (the carnitine palmitoyl transferases), the TCA cycle (citrate synthase) and mitochondrial uncoupling protein 2 decline and, where examined, these changes are reflected at the protein level.^15,16 This suggests that the failing heart is less able to metabolize oxidatively. For carbohydrate metabolism, transcripts for the PDH kinases (the regulatory enzymes involved in the inhibition of PDH-MEC) and the GLUTs decline.¹⁶ Assuming that these changes are reflected at the protein level, the latter will disfavour carbohydrate utilization whereas the former should favour oxidative metabolism of glucose. It is impossible to predict changes in biochemical processes simply on the basis of transcript abundances. Leaving these findings aside, the mechanisms by which oxidative capacity of the heart is regulated are becoming clearer, and here the peroxisome proliferator-activated receptors (PPARs) and the PPAR coactivators (PGCs) are important.

The PPARs, i.e. PPARα, PPARβ/δ, and the alternatively spliced forms of PPARγ (PPARγ₁, PPARγ₂, PPARγ₃)¹⁷ and PGCs (PGC-1α, PGC-1β, and PERC)^18,–20 are central to the chronic regulation of oxidative metabolism. PPARs are nuclear receptors and transcription factors that act as ‘lipid sensors’, being activated by LCFAs and the products of their metabolism (Fig. 11.3). PPARγ is also the ‘receptor’ for the insulin-sensitizing thiazolidinedione drugs that are extensively used in the treatment of type 2 diabetes. PGC-1α acts as the master regulator of transcription by coactivating PPARs, and other transcription factors (nuclear respiratory factors, oestrogen-related receptors, retinoid X receptors) (Fig. 11.3). It is induced either at the level of transcription or is activated (either directly by post-translational modification or by release from repressor proteins) to promote a shift towards oxidative metabolism.

Whether metabolic remodelling is detrimental or beneficial is still debated,²¹ though the majority opinion believes that it is beneficial. The question of whether pharmacological agents that increase reliance of the early-stage failing heart on carbohydrate metabolism are beneficial then arises—or would a reversion to the normal state of greater reliance on fatty fuels be preferable? This has been the focus of many reviews,^18,21,–24 though it is probably too early to draw definite conclusions. The general approach involves inhibition of LCFA metabolism either by reduction of LCFA uptake into the mitochondria through inhibition of the carnitine palmitoyl transferases (Fig. 11.1) (etomoxir, oxfenicine, perhexiline) or inhibition of β-oxidation itself (Fig. 11.1) (ranolazine, trimetazidine).²² There have been a few small-scale clinical trials of these reagents which appear hopeful, but a clear correlation of benefit with reduced LCFA metabolism is not always apparent. This is clearly an area where more data are required. The β-blocker carvedilol, which is of proven benefit in HF patients, may reduce LCFA metabolism²⁵ and this may contribute to its superiority over some other β-blockers. Apart from drug-based therapies, increased cardiac oxidative metabolism and energy transfer from exercise training may also contribute to its beneficial effects in HF.²⁶

In partial or intermittent ischaemia, LCFA-CoA, and LCFA-carnitine accumulate because they cannot be oxidized. These are powerful detergents and they disrupt membranes to the detriment of cellular integrity. Triglycerides are synthesized because accumulating LCFA-CoA re-esterifies glycerol-3-phosphate (formed from the reduction of the glycolytic intermediate 3-phosphoglyceraldehyde by NADH):

In cardiomyopathies that develop in genetic conditions which interfere with lipid oxidation, triglyceride accumulation is evident.¹⁸ Furthermore, a reduced ability to oxidize lipids when they may be readily available (e.g. in diabetes) in early-stage HF may lead to ‘lipotoxicity’, a condition in which the accumulation of the intermediates of lipid oxidation and triglyceride is cytotoxic.^18,27,28

ATP, total adenine nucleotides (TAN; ATP + ADP + AMP) and total creatine (PCr + creatine) pools are progressively lost in the failing heart.^29,30 As the metabolic pools diminish, oxidative stress increases. In the experimental setting—a pacing-induced (volume overload) canine model of HF—the transition point from the initial compensatory phase to the maladaptive phase corresponds to a 30% loss of TAN.³¹ This transition point is also influenced by the total exchangeable phosphate pools (TEP; 2ATP + ADP + P_i + PCr + mitochondrial matrix GTP) and total creatine pool (PCr + creatine). Thus, at given values of TAN and TEP, the total creatine pool is adjusted to obtain the maximum available free energy from ATP hydrolysis and any changes (increases or decreases) in the total creatine pool will result in diminished performance unless accompanied by adjustments in TAN and TEP.³¹

At constant values of TAN, etc., proportionally small decreases in ATP concentrations are amplified as proportionally much larger increases in AMP concentrations by the adenylate kinase equilibrium (Equation 11.3). AMPK senses changes in AMP concentrations and functions as a ‘cellular fuel gauge’,^32,–35 and it may represent a therapeutic target in cardiovascular disease.³⁶ Activation of AMPK (by AMP) shifts tissues from an anabolic state, in which protein, glycogen and fatty acid synthesis are favoured, to a catabolic state in which glucose uptake, glycolysis, fatty acid oxidation, and mitochondrial biogenesis are favoured (Fig. 11.4). This is achieved by AMPK-mediated phosphorylation of target proteins to modify their biological activity.^32,–35 For example, AMPK phosphorylates the 6-phosphofructo-2-kinase/fructose-2,6-bisphosphatase bifunctional regulatory enzyme to stimulate the 6-phosphofructo-2-kinase activity (Fig. 11.1). Fructose 2,6-bisphosphate concentrations increase to activate the ‘rate-controlling’ PFK-1 step in glycolysis, thereby stimulating glycolytic flux.

AMPK is a heterotrimer composed of an α (α₁ or α₂), β (β₁ or β₂), and a γ (γ₁, γ₂, or γ₃) subunit (Fig. 11.4). The α subunit is the catalytic subunit (the major species in heart is α₂), the β subunit is a ‘scaffold’ which may be involved in binding to and quantitative sensing of glycogen, and the γ subunit is the regulatory adenine nucleotide binding subunit. AMPK activity is regulated in two ways. The γ subunit contains two adenine nucleotide binding sites (each composed of two ‘cystathionine β-synthase’ or ‘Bateman’ hemidomains) which may differ in their affinities for adenine nucleotides and in their relative effects. Binding of AMP activates AMPK whereas ATP competes with AMP to inhibit AMPK activity. Maximally, AMP binding increases AMPK activity by about 10-fold. In addition, phosphorylation of the α subunit (on a threonine residue, Thr-172) activates AMPK about 100-fold and this phosphorylation is favoured when AMPK is in its AMP-ligated state. There are a number of upstream protein kinases which reportedly effect this activation, but the best characterized is the tumour suppressor LKB1. Indeed, the tumour-suppressive properties of LKB1 probably involve an element of AMP-mediated inhibition of cell growth.

Ex vivo, AMPK is activated in both myocardial ischaemia and in ischaemia/reperfusion (AMP/ATP increased), and this increases fatty acid oxidation during reperfusion by inhibiting acetyl-CoA carboxylase (the first step in fatty acid synthesis).³⁷ Studies in experimental animals indicate that AMPK is activated in left ventricular hypertrophy and HF.^38,39 Whether activation of AMPK in acute stress (short-term ischaemia, ischaemia/reperfusion) or chronic (HF) stress is beneficial or detrimental is still debated.⁴⁰ Activation of AMPK is also seen in the heart following exercise⁴¹ and presumably will lead to mitochondrial biogenesis during which, by analogy with skeletal muscle, AMPK activates PGC-1.⁴² If this is the case, oxidative metabolism and mitochondrial biogenesis should be favoured during failure, quite the reverse of the demonstrated increased dependence of the failing heart on carbohydrate (and anaerobic?) metabolism.

Conversely, transgenic mice which cardiospecifically overexpress an inhibitory (‘dominant-negative’) form of the AMPK α₂ subunit (to limit activation of AMPK) have been engineered. Here, inhibition of AMPK activation is not detrimental and may even be beneficial in recovery from ex vivo ischaemia on reperfusion.⁴³ This conclusion concurs with an earlier study using hearts from AMPK α₂-null mice perfused ex vivo.⁴⁴ Here, although the ischaemic contracture in AMPK α₂-null hearts was more severe with an earlier onset than with wild-type hearts, the degree of functional recovery during reperfusion did not differ.

Other than the studies described in the previous section, is there any evidence that modulation of AMPK activity or its downstream effectors is valuable in heart disease?³⁶ The biguanide metformin has been used extensively in type 2 diabetes, and there is evidence that it ameliorates any associated HF. Metformin appears to be beneficial in experimentally induced HF.³⁹ Metformin is a weak ‘uncoupler’ of oxidative phosphorylation (P/O ratio decreased), leads to decreased ATP/ADP ratios, and to increases in AMP concentrations and activation of AMPK (Fig. 11.4). The thiazolidinediones (also used in type 2 diabetes) may also have a beneficial effect, though their use in HF may be contraindicated because of increased water retention. Although the target for these drugs is PPARγ, PPARγ itself may up-regulate release of the adipokine adiponectin from adipose tissue, leading to activation of AMPK in other tissues. Assuming that AMPK activation is beneficial in HF, direct activation of AMPK would seem desirable. In the laboratory setting, 5-amino-4-imidazolecarboxamide riboside (AICAR) is used, but this is unsuitable for clinical use. The AMPK activator A769662 has poor bioavailability, and effort is now being expended in synthesizing more useful derivatives.

Mutations in the γ₂ subunit of AMPK (expressed in both cardiac and skeletal muscle) give rise to an autosomal-dominant cardiac phenotype (characterized by a Wolff–Parkinson–White type premature excitation arrhythmia and glycogen accumulation).^34,45 Although originally thought to be a ‘cardiac hypertrophy’, the increased size and weight of the heart is caused by glycogen accumulation and the bound water associated with it. The precise cause of the arrhythmia is not understood but the assumption is that glycogen accumulation introduces alternative atrioventricular conduction pathways. Eight of the 10 individual mutations associated with the disease map to the two AMPK (γ subunit adenine nucleotide binding sites. Two (spontaneous) mutations give rise to a particularly severe phenotype leading to early infant mortality, and these are therefore not heritable. The precise cause of the glycogen accumulation is unclear. If AMP is unable to activate AMPK, cardiac anabolism would be favoured with concurrent synthesis of glycogen and lipids. Equally, if ATP is unable to inhibit AMPK, its basal activity would increase, favouring glucose uptake, and LCFA oxidation. However, operation of the Randle cycle (Fig. 11.1) could suppress glycolysis at the 6-phosphofructo-1-kinase step, and cause accumulation of glucose 6-phosphate. This might drive synthesis of glycogen. Indeed, a transgenic mouse model in which the AMPK γ subunit is mutated displays the characteristics of the latter scenario, with increased glucose uptake and fatty acid oxidation, and accumulation of glycogen.⁴⁶