Hexokinases and Cardioprotection

Abstract

As mediators of the first enzymatic step in glucose metabolism, hexokinases (HKs) orchestrate a variety of catabolic and anabolic uses of glucose, regulate antioxidant power by generating NADPH for glutathione reduction, and modulate cell death processes by directly interacting with the voltage-dependent anion channel (VDAC), a regulatory component of the mitochondrial permeability transition pore (mPTP). Here we summarize the current state-of-knowledge about HKs and their role in protecting the heart from ischemia/reperfusion (I/R) injury, reviewing: 1) the properties of different HK isoforms and how their function is regulated by their subcellular localization; 2) how HKs modulate glucose metabolism and energy production during I/R; 3) the molecular mechanisms by which HKs influence mPTP opening and cellular injury during I/R; 4) how different metabolic and HK profiles correlate with susceptibility to I/R injury and cardioprotective efficacy in cancer cells, neonatal hearts, and normal, hypertrophied and failing adult hearts, and how these difference may guide novel therapeutic strategies to limit I/R injury in the heart.

Introduction

In the heart, ischemic preconditioning (IPC) is a process whereby repeated brief episodes of ischemia/reperfusion (I/R) protect the heart from injury during a subsequent prolonged I/R episode [1]. Although much research has been devoted to this phenomenon and its other variants, including pharmacologic preconditioning (PPC) and ischemic post-conditioning (IPoC), the underlying mechanisms of cardioprotection remain elusive. Common to all of these cardioprotective strategies is activation of a signaling cascade called the Reperfusion Injury Salvage Kinase (RISK) pathway involving phosphatidylinositol-3-kinase (PI3K), Akt, glycogen synthase kinase 3 beta (GSK3β) and other enzymes [2] (Fig. 1). In addition, cardioprotective signaling can also be transmitted via the Survivor Activating Factor Enhancement (SAFE) pathway involving the activation of tumor necrosis factor alpha (TNFα) and signal transducer and activator of transcription-3 (STAT-3) [3-6], which may crosstalk with the RISK pathway. Although the details are still fuzzy, these signals appear to converge ultimately on the mitochondrion [7], and avert cell death by inhibiting mitochondrial permeability transition pore (mPTP) opening in the inner mitochondrial membrane. Of note, activation of the RISK pathway depends on a modest burst of reactive oxygen species (ROS) production during the preconditioning ischemia, because if ROS scavengers are present during this period, cardioprotection is abolished [8]. When the RISK pathway is activated by pre- or post-conditioning, the massive ROS burst that typically occurs after prolonged I/R in the absence of pre- or post-conditioning is markedly attenuated [9]. This may be a major factor in averting the mPTP opening and cell death, since oxidative stress is one of the major factors promoting mPTP opening [10].

Fig. 1.

Overview of cardioprotective signaling by ischemic and pharmacologic pre-conditioning (PC) and post-conditioning (PoC). ROS-dependent PKCε activation (often triggered by ROS originating from mitochondria due to mitoK_ATP channel or transient low conductance mPTP opening) or other pathways such as adenosine receptor (AdR) stimulation lead to activation of the RISK pathway, initiating a signaling cascade that suppresses the massive ROS burst during reperfusion and prevents long-lasting mPTP opening. The Akt component of the RISK pathway enhances HK2 binding to mitochondria, contributing to an improved metabolic profile during prolonged ischemia and reduced ROS burst upon reperfusion characteristic of the cardioprotected state. Various activators (-A) and inhibitors (-I) of these steps are indicated. AR, adrenergic receptors; BkR, bradykinin receptors; OpR, opiate receptors; CsA, cyclosporine A; 5HD, 5-hydroxydecanoate; PI3K, phosphoinositol-3-kinase.

The finding that post-conditioning confers almost equivalent cardioprotection as pre-conditioning suggests that RISK pathway activation at any time point up and during early reperfusion is the main requirement for effective cardioprotection [11]. Once activated, the cardioprotected state has two phases. The early phase, attributed to acute post-translational modification of target proteins by the RISK pathway, lasts about 1-2 hours. Cardioprotection then dissipates, but returns again within 12-24 hours, and lasts for another 48-72 hours. This late phase is related to gene reprogramming and new protein synthesis triggered by RISK pathway activation [10].

How do hexokinases (HKs) fit into this picture? During low-flow ischemia or anoxia, glucose metabolism becomes the major source for ATP production via anaerobic glycolysis, and HKs mediates the first step in this process, namely the conversion of glucose to glucose-6-phosphate (G6P). G6P is a hub metabolite that can be directed to a number of catabolic or anabolic fates (Fig. 2). The main catabolic fate is glycolysis, which first generates ATP anaerobically via conversion to pyruvate and lactate, and then aerobically by mitochondrial oxidation of pyruvate and lactate. The main anabolic fates of G6P are two-fold: glycogen synthesis to store energy for deferred use, and the pentose phosphate shunt to generate ribose-5-phosphate (R5P). The conversion of G6P to R5P is a major source of cytoplasmic NADPH generation which is critical for maintaining antioxidant function by regenerating reduced glutathione (GSH) from oxidized glutathione (GSSG). R5P is also a precursor for synthesis of nucleotides, amino acids and fatty acids. Small amounts of G6P are also used by the hexosamine pathway for O-GlyNAtion of proteins, and some of the G6P directed to pyruvate is used in the TCA cycle for amino acid and fatty acid synthesis via anaplerosis.

Fig. 2.

Summary of catabolic, anabolic and antioxidant pathways in glucose metabolism. HK2 binding to mitochondria favors glycolysis, whereas cytoplasmic HK2 favors glycogen synthesis and the pentose phosphate shunt, which provides NADPH to reduce oxidized glutathione (GSSG) to reduced glutathione (GSH) and thereby bolsters antioxidant reserve.

In addition to acting as the gatekeeper for the metabolic and antioxidant roles of glucose, HKs have also been found to regulate mPTP opening directly. First described in the cancer field, the high-affinity HK isoforms HK1 and HK2 are strongly anti-apoptotic when bound to the outer mitochondrial membrane, putatively to the voltage-dependent anion channel (VDAC), an important regulator of the mPTP [12, 13].

Thus, HKs are multifunctional proteins that orchestrate metabolic, antioxidant and direct anti-cell death effects. These functions are also strongly influenced by the subcellular distribution of HKs, with mitochondrially-bound HK promoting glucose catabolism and anti-cell death effects, while cytoplasmic HK promotes anabolic usages and antioxidant effects (Fig. 2). The predominant HK isoform in adult heart, HK2, dynamically shuttles between the mitochondria and cytoplasm in response to changes in intracellular G6P, pH and the cardioprotective signaling pathway Akt [14-22]. These factors set the stage for HKs to have a critical influence on the susceptibility of the heart to I/R injury.

Here we summarize the current state-of-knowledge about HKs and their role in protecting the heart from I/R injury. We begin by discussing the properties of different HK isoforms and how their function is regulated by their subcellular localization. Next, we discuss how HK modulates glucose metabolism and energy production during I/R, and then review the molecular mechanisms by which HKs influence mPTP opening and cellular injury during I/R. Finally, we describe the differences in susceptibility to oxidative or I/R injury in cancer cells, neonatal hearts, and normal, hypertrophied and failing adult hearts in relation to their HK and metabolic profiles, and discuss potential therapeutic implications.

Properties of HK isoforms

HKs are comprised of a family of four isoforms. HKI and HK2 are the most abundant, with HKI (“the brain HK”) ubiquitous in most tissues, especially brain and red blood cells [23, 24] where glycolysis plays a critical role in energy production. In contrast, HK2 (“the muscle HK”) is found primarily in insulin-sensitive tissues such as adipocytes, adult skeletal and cardiac muscle [25, 26]. Few data are available for human heart, but a recent report indicates that in non-dilated human atrial tissue, HKI is the most abundant isoform [27]. In mouse and human skeletal muscle [25, 28], however, HKII accounts for >80% of total HK activity. Importantly, both HKI and HK2 contain a hydrophobic amino terminal mitochondrial binding motif, which is not present in the HK3 or HK4 (glucokinase) isoforms.

The idea that subcellular locations of HKI and HK2 are important in regulating glucose metabolism was first postulated by Wilson, who stated that “the Type I isozyme bound to actively phosphorylating mitochondria facilitates introduction of glucose into glycolysis, with the final stages of glucose metabolism occurring in the mitochondria. In contrast, Type II and to some extent Type III isozymes serve primarily anabolic function to provide G6P for glycogen synthesis or lipid synthesis via the pentose phosphate pathway” (see review [29]). Subsequent work by multiple investigators [14-22, 30, 31] has demonstrated that unlike other HK isoforms, the interaction of HK2 with mitochondria is not static, but is regulated by factors such as glucose, G6P and kinases such as Akt. Thus, a picture has emerged of HK2 as a multifunctional orchestrator of glucose metabolism: channeling G6P into glycogen and the pentose phosphate pathways when localized in the cytoplasm, and preferentially shuttling G6P to glycolysis and oxidative phosphorylation when bound to mitochondria [31, 32]. In contrast, HKI, due to its strong mitochondrial binding, primarily facilitates glycolysis [30, 31], although under some non-physiological conditions may also contribute to glycogen synthesis [33]. HK3 and HK4 are cytoplasmic, since they lack a mitochondrial binding motif, and serve primarily anabolic functions. HK4 (glucokinase), however, also has the ability to shuttle to the nucleus, perhaps playing a role in gene transcription/new protein synthesis.

HKI and HK2 are inhibited allosterically by their product, G6P, and their sensitivity to G6P inhibition decreases when HKs are bound to mitochondria [34, 35]. Physiological levels of orthophosphate (Pi) counter the G6P inhibition of HKI [24, 36, 37], but not HK2. In fact, Pi may cause further HK2 inhibition. Based on these observations, Wilson [29] proposed that “reciprocal changes in intracellular levels of G6P and Pi are closely associated with cellular energy status, and that the response of HK1 to these effectors adapts it for catalytic function by adjusting glucose flow into glycolytic metabolism. In contrast, HK2 serves primarily anabolic functions.”

G6P also plays a critical role in regulating HK2 binding to mitochondria. Elevated levels of intracellular G6P weaken HK2 binding to mitochondria, causing the enzyme to translocate to the cytoplasm and facilitate anabolic glucose metabolism. Acidosis similarly induces HK2 translocation from mitochondria to cytoplasm [38].

Regulation of glucose metabolism by HK during I/R

During acute myocardial ischemia, anaerobic glycolysis and glycogenolysis assume the central role for energy production when oxidative phosphorylation cannot occur because of a lack of oxygen. The shift to anaerobic metabolism entails rapid increases in glucose uptake, glycogenolysis, and glycolytic flux [39]. During severe ischemia, however, the accumulation of protons and glycolytic intermediates eventually inhibits glycolytic flux and anaerobic ATP production ceases after 20-30 min [40, 41]. The relative contribution of the glycolytic/glycogenolytic pathway to energy production is highly dependent on the severity of ischemia, with virtually no change up to moderate ischemia (reduction of coronary flow by <75%) to virtually 100% during total global ischemia. In the moderate ischemia case, glucose uptake remains unchanged, but glucose metabolism is directed from oxidation to lactate production [42]. Increased glucose uptake during ischemia is further stimulated by insulin [43]. Promoting glucose metabolism with glucose and insulin has been used successfully to protect hearts from I/R injury [44, 45], although the clinical utility has been limited [46].

In the adult heart, the subcellular localization of HK2 shifts during ischemia [47] due to the dissociation of HK2, but not HKI, from the mitochondria to cytoplasm [30, 48, 49] in response to intracellular acidification and G6P accumulation [38]. Upon reperfusion after a period of ischemia, the activity of HK is increased in both the cytosolic and mitochondrial compartments [50], probably promoted by activation of Akt signaling via the RISK pathway. Under these conditions of Akt-enhanced HK2 binding to mitochondria, the heart is in a cardioprotected state. After IPC, HK2 dissociates more slowly during a subsequent prolonged ischemic episode, and this correlates with a slower rate of ATP depletion than during ischemia in unprotected hearts [51]. The maintained fraction of HK bound to mitochondria may also serve to facilitate glycolysis during reperfusion, which plays a critical role in recovery [52].

Given the multiple roles of HK in orchestrating metabolic, antioxidant and direct cell death effects from different subcellular locations, it is natural to ask whether the parallels between the cardioprotective status and HK binding to mitochondria are mere associations or are causally-linked. A growing body of evidence now supports a causal relationship. HK binding to mitochondria in cardiomyocytes has been demonstrated to confer profound protection against cell death by preventing mPTP opening [18, 53, 54]. Following prolonged ischemia, cytochrome c release, ROS production and infarct size at the reperfusion have been shown to parallel the amount of HK dissociated from mitochondria [38]. Similar to IPC, several cardioprotective interventions acting through the RISK pathway also increase mitochondrial HK activity and reduce I/R injury in isolated rat hearts [55]. The cardioprotective actions of the volatile anesthetics isoflurane and sevoflurane, whose effects are dependent on the PI3K/Akt signaling [56], have been shown to increase mitochondrial HK activity in in vivo adult rat hearts [57]. Akt has been shown to directly phosphorylate HK2, which inhibits G6P-mediated dissociation from mitochondria and increases cell viability after stress [18, 58]. Finally, disrupting HK2 binding to mitochondria with targeted peptides has been found to increase I/R injury and attenuate cardioprotection by IPC, and genetic reduction of HK2 levels in heterozygous HK2 knock out mice increased susceptibility to I/R injury [59, 60].

Although HK localizes mainly to either the cytosol or mitochondria [29-31] and cardioprotection is associated with increased mitochondrial HK binding, some studies have identified a component of HK located in intracellular vesicles within the cardiomyocyte [48, 61], supporting older data that HK may also be associated with membrane structures other than mitochondria [62]. Although no changes in HK activity seems to occur in this microsomal compartment with ischemia or IPC, HK protein levels in this compartment during ischemia are affected by IPC [48]. It is conceivable that the HK in these microsomes is somehow involved in IPC, but further research will be necessary to elucidate the role, if any.

In the next section, we discuss the possible underlying molecular mechanisms by which HK binding to mitochondria may protect the heart from I/R injury.

Molecular actions of HK that impact mPTP opening during I/R

The mPTP protein complex

The mPTP is a multi-protein complex, whose pore-forming component is now believed to consist of F₀-F₁ ATP synthase dimers in the inner mitochondrial membrane [63]. The mPTP forms at contact sites between the inner and outer mitochondrial membranes. The other main proteins in the complex, which regulate pore function, include VDAC in the outer mitochondrial membrane, the adenine nucleotide translocator (ANT) in the inner mitochondrial membrane, cyclophilin D, and perhaps the mitochondria phosphate carrier (PiC). Other proteins with regulatory roles include the benzodiazepine receptor, creatine kinase and HK.

Once formed, mPTP can open in either a transient low conductance mode that may be protective [64-66], or a long-lasting high conductance mode that eventually irreversibly damages the mitochondria and promotes cell death signaling. The transient low conductance openings may be important for allowing mitochondrial to release accumulated matrix Ca by transiently depolarizing, thereby avoiding matrix Ca overload leading to long-lasting mPTP. Transient mPTP openings are also associated with a burst of superoxide production [67] that may contribute to the ROS-dependent triggering of the RISK pathway during IPC (Fig. 1), since blocking transient mPTP opening during IPC with CsA has been shown to abolish cardioprotection [68], similar to the effect of ROS scavengers administered during IPC [8].

Long-lasting mPTP opening, on the other hand, leads to: 1) an abrupt complete dissipation of the mitochondrial membrane potential, which converts mitochondria from ATP generators into ATP consumers as F₀-F₁ ATP synthase hydrolyzes ATP to pump protons out of the matrix in a futile effort to restore membrane potential; 2) loss of solutes with molecular masses <1.5 kDa, including key metabolic co-factors such as pyridine nucleotides essential for oxidative phosphorylation; 3) influx of water due to the high oncotic pressure of the matrix that swells the mitochondrial inner membrane and can eventually rupture the outer mitochondrial membrane rupture, releasing pro-apoptotic proteins such a cytochrome c [69, 70].

It has been shown that both HKI and HK2 interact with VDAC in the outer mitochondrial membrane [20, 71-73], and that this requires the presence of their N-terminal hydrophobic domain [74, 75]. The interaction of HK with mitochondria is known to protect against cell death in many cell types [76-78] and mPTP opening is inhibited by the interaction with HK2 [69]. Other proteins in the mPTP complex influence HK binding to VDAC. Inhibition of ANT with bonkregic acid, which locks the ANT in a closed state and is an mPTP inhibitor, reduced HKI and HK2 binding to isolated mitochondria, while treatment with atractyloside, which locks the ANT in an open conformation and is an mPTP inducer, increased HK binding [79]. Inhibition of CypD with the mPTP blocker cyclosporine A (CsA) also decreased HK2 binding to mitochondria, as did siRNA knockdown of CypD [80].

Since VDAC is both a component of the mPTP multiprotein complex and a key regulator of mitochondrial bioenergetics, its interaction with HK raises the following question: are the protective effects of mitochondrial HK binding against mPTP opening a direct effect on mPTP function, an indirect effect mediated by the metabolic effects of HK when bound to VDAC, or both? The evidence to date is limited but suggests that both direct (nonmetabolic) and indirect (metabolic) factors are important. For example, studies have shown that glucose must be present in order for mitochondrially-bound HK to offer protection against cell death [16, 53, 81], implying that enzymatic HK activity to generate G6P is critical. On the other hand, experiments overexpressing HK2 mutants lacking enzymatic activity, but with intact binding to mitochondria, conferred partial protection against H₂O₂-induced injury in cultured cardiomyocytes [18, 48, 53]. Conversely, HK2 mutants with intact enzymatic activity, but no mitochondrial binding, also exhibited partial protection, suggesting that cytoplasmic HK2 may confer some benefit as well. Whether this is also true in the setting of I/R injury, however, remains to be established.

Cytoprotective effects of HK during I/R

There are at least five mechanisms by which HK binding to VDAC may exert cytoprotective effects during I/R that avert long-lasting mPTP opening (Fig. 3 and see [26] for review). The first mechanism is a direct (nonmetabolic) action and the remaining four are indirect (metabolic) actions. They are:

Fig. 3.

Summary of possible mechanisms by which enhanced HK2 binding to VDAC in the preconditioned heart improves the metabolic profile and averts mPTP opening during I/R. IMS, intermembrane space; ANT, adenine nucleotide transporter; AK, adenylate kinase

1. Interference with translocation of pro-apoptotic proteins (Bax, Bad) to VDAC. In cellular studies employing oxidant stress, HK binding to mitochondria has been shown to be cytoprotective by structurally interfering with the ability of pro-apoptotic proteins such as Bax and Bad to translocate from the cytoplasm and bind to VDAC [82]. However, the relevance to I/R injury is unclear, since studies performed in isolated hearts [38, 59] or skeletal muscle [83] have not been able to show a correlation between HK-mediated protection and Bax/Bak levels at the mitochondria in the setting of I/R. Bax binding to VDAC is implicated in the formation of a megachannel in the outer mitochondrial membrane through which large molecules, including cytochrome c, can pass [84, 85]. Release of cytochrome c into the cytoplasm is a major trigger of the apoptotic signaling cascade, which feeds back to induce long-lasted mPTP opening [86]. In addition, cytochrome c plays an important antioxidant role [82], and its loss from the intermembrane space leaves mitochondria more susceptible to oxidative stress-induced mPTP opening during reperfusion.

2. Reduction in futile ATP consumption by reverse F₀-F₁ ATPase. When bound to VDAC under aerobic conditions, HK utilizes ATP to generate G6P in close proximity to the outer mitochondrial membrane, facilitating its glycolytic conversion to pyruvate as a substrate for oxidative phosphorylation. During ischemia, however, the lack of oxygen inhibits electron transport and the matrix rapidly depolarizes. Pyruvate can no longer be oxidized and is converted to lactate. G6P as well as other glycolytic intermediates accumulate intracellularly. Under these conditions, matrix F0-F1 ATPase operates in reverse, consuming ATP to pump protons in a futile attempt to restore membrane potential. Studies using F₀-F₁ ATPase inhibitors such as oligomycin show that this futile ATP consumption accounts for as much as 50-70% cellular ATP depletion during early ischemia [87, 88]. In this setting, VDAC acts as the critical gatekeeper for glycolytically-generated ATP to enter the IMS, where it is then transported by ANT into matrix to be consumed in this futile cycle. ATP and ADP can readily permeate the outer mitochondrial membrane when VDAC is in its open state, but are virtually impermeant when VDAC is closed [89]. It has been shown that HK binding promotes the closed state of VDAC, reducing the permeability to adenine nucleotides [90]. Under de-energized conditions, slowing of adenine nucleotide transport in and out of the mitochondria by inhibiting VDAC opening significantly attenuates myocardial ischemia-reperfusion injury [91]. Thus, one cytoprotective effect of HK binding to VDAC may be to reduce VDAC’s permeability to adenine nucleotides and thereby restrict the entry of glycolytic ATP into the intermembrane space and subsequently into the matrix via ANT, resulting in less futile ATP degradation by reverse F₀-F₁ ATPase. In addition, even when the VDAD pore is open, HK depletes local [ATP] in its immediate vicinity by dephosphorylating ATP to ADP, thereby favoring the entry of ADP, rather than ATP, into the IMS when the VDAC pore is open. This further reduces futile ATP consumption by reverse F₀-F₁ ATPase.

3. Facilitation of ATP production by adenylate kinase in the IMS. By hydrolyzing glycolytic ATP to ADP near the VDAC pore, HK facilitates entry of ADP, instead of ATP, into the IMS. The preferential transport of ADP into the IMS stimulates ATP generation by adenylate kinase, which catalyzes the reaction reaction 2ADP ↔ AMP + ATP. The regenerated ATP can then exit to IMS via VDAC to support cytoplasmic bioenergetic functions. Thus glycolytic ATP that would have been futilely consumed by reverse F₀-F₁ ATPase in the matrix is diverted to ATP regeneration by adenylate kinase.

4. Facilitation of glycolytic ATP generation. A fourth cytoprotective effect relates to the fact that in the process of dephosphorylating ATP to ADP before it can enter the IMS through VDAC, HK generates additional G6P, which is recycled to support additional glycolytic ATP generation.

5. Facilitation of antioxidant power by the pentose phosphate shunt. Alternatively, the G6P generated by mitochondrially-bound HK can be remain in the cytoplasm and be utilized by the pentose phosphate shunt to produce NADPH for regenerating GSH from GSSG, thereby enhancing antioxidant reserve [92].

The combination of these five factors may explain why IPC, by activating Akt and enhancing HK binding to VDAC, slows the rate of ATP decline during the subsequent prolonged ischemic episode [93], and also attenuates the massive burst of ROS that occurs upon reperfusion after prolonged ischemia by bolstering cytochrome c- and GSH-related antioxidant power. The combination of reduced Bax/Bad interaction with VDAC, less futile ATP consumption by reverse F₀-F₁ ATPase, enhanced adenylate kinase-mediated ATP generation in the IMS, and increased G6P production to facilitate both glycolytic ATP generation and antioxidant power via the pentose phosphate shunt may allow the heart to preserve a better metabolic profile during ischemia, and be better prepared to limit oxidative injury following reperfusion. In combination, these actions may avert long-lasted mPTP opening and limit cell death in pre-conditioned hearts (Fig. 3).

HK and metabolic profiles in cancer, developing hearts, and normal and diseased adult hearts

Differences in susceptibility to I/R injury and efficacy of IPC in various cells types in relation to their metabolic profiles may provide additional clues about the role of HK in cardioprotection. Specifically, cancer cells exhibit Warburg-like metabolic profiles characterized by enhanced anaerobic glycolysis, reduced glucose oxidation, upregulated glycolytic enzymes (including HK) and down-regulated oxidative phosphorylation enzymes, elements of which are shared by neonatal hearts, and hypertrophied or failing adult hearts (Table 1). Neonatal hearts, like cancer cells, are intrinsically less susceptible to I/R injury, and are not further protected by IPC like normal adult hearts [94]. Hypertrophied and failing adult hearts, on the other hand, show some elements of the Warburg metabolic profile, but are not necessarily less susceptible to I/R injury or as effectively cardioprotected by IPC compared to normal adult hearts [95-98]. Here we review the relationship between susceptibility to I/R injury, efficacy of IPC and the metabolic profiles in these different cell types.

Table 1.

Summary of HK isoforms, metabolic profiles, susceptibility to I/R injury and ability to protect by IPC in different settings.

Tissue	Predominant HK isoform	Metabolic Profile			Susceptibility to I/R injury	Protection by IPC
		Glycolysis	Glu Ox	FA Ox
Adult heart	HK2	Low	Low	High	High	High
Cancer cells	HK2	High	Low	Low	Low	?
Fetal/Neonatal heart	HK1	High	High	Low	Low	Low
Hypertrophied adult heart	HK2	High	Low	Low	High	High
Failing adult heart	HK2	Low	Low	Low	Low?	Low?
Diabetic adult heart	HK1	Low	Low	High	Low (short term diabetes)	Low
					High (long term diabetes)

FA= fatty acid; Glu=glucose; Ox=oxidation

Cancer

In 1956, Otto Warburg observed that the rate of anaerobic glycolysis was abnormally high in cancer cells, despite a smaller fraction of this glucose being oxidized by mitochondria due to a preferential channeling of pyruvate to lactate. This ‘Warburg effect’ indicates that cancer cells prefer anaerobic glycolysis for energy production, rather than mitochondrial oxidative phosphorylation of either glucose or fatty acids [99, 100]. The preference for anaerobic energy production by glycolysis may be important for allowing tumor cells to proliferate when O₂ supply is limited by the absence of a well-developed vascular supply,and also confers increased resistance to apoptosis.

HK plays a pivotal role in this metabolic shift and increased resistance to cell death, with the activity of HK2 and to some extent HKI being markedly elevated in rapidly growing tumors which exhibit high rates of glycolysis [71, 101-103]. Two factors contribute to this enhanced activity: increased expression of HK’s by as much as 100-fold, and a propensity for the HKs to bind to VDAC in the outer mitochondrial membrane in tumor cells [12, 103, 104]. Mitochondrial binding of the tumor HKs has been shown to provide the HK with preferential access to mitochondrially-generated ATP to phosphorylate glucose, [71] and to reduce its sensitivity to product inhibition by G6P [105], an important regulator of HK activity in normal cells [20, 105]. Secondly, HK binding to VDAC is integral to the protection against apoptosis in cancer cells by inhibiting cytochrome c release and mPTP opening [13], as described earlier.

The Warburg-like pattern of increased HK activity and redistribution to mitochondria is not solely the hallmark of rapidly growing cancer cells, but is also observed in other tissues including brain, skeletal muscle and heart at various developmental stages in health and disease.

Fetal and neonatal heart

An increase in HKI expression/activity signals a switch from pyruvate to glucose metabolism in the preimplantation embryo [106, 107], and HKI activity increases significantly in the morula and blastocyst stages [108-110]. From this point through the early perinatal period, cardiac energy metabolism relies heavily on glucose metabolism, facilitated by the continuous availability of glucose from the mother to the fetus through the placental circulation. Mitochondrial oxidative metabolism is poorly developed, with glycolysis as the primary energy source [111-114]. This high glycolytic profile seems to facilitate the proliferative state of developing fetal cardiomyocytes [111, 113, 114]. Even after birth, as mitochondrial metabolism matures, glycolysis is still the preferred energetic pathway for several weeks, driven in part by expression of insulin-independent isoforms of the glucose transporter (GLUT1) and HKI, with its strong binding to mitochondria. Thus, HKI is the predominant isoform in embryonic, fetal and neonatal hearts [115] facilitating their high glycolytic rate [30]. After birth, the transition from a continuous supply of placental glucose to an intermittent mixed carbohydrate-fat oral diet during nursing, induces a switch to the insulin-dependent isoforms GLUT4 and HK2 [30, 116], along with down-regulation of other glycolytic enzymes and upregulation of oxidative phosphorylation enzymes [117]. These changes confer a preference for fatty acids, lactate and ketone bodies over glucose in the adult heart.

The Warburg-like metabolic profile in the developing heart is associated with an increased resistance to ischemic/hypoxic injury as compared to adult myocardium [118-120]. This greater resistance to cell death may involve both the increased capacity of anaerobic glycolysis to maintain ATP and the stronger binding of HKI than HK2 to mitochondria, which suppresses mPTP opening [121]. Indeed, whereas in adult hearts, strengthening of HK2 binding to mitochondria is essential for the protective effects of IPC [59], this is not the case for neonatal hearts. In neonatal hearts, the inherently stronger binding of HKI to mitochondria is not further increased by IPC, and the intrinsically high resistance of the newborn heart to I/R injury also cannot be further increased by IPC or adaptation to chronic hypoxia [122]. Neonatal cardiac mitochondria also have been shown to be less sensitive to mPTP opening [123], further strengthening the idea that the high affinity binding of HKI to mitochondria in neonatal heart may contribute to the increased resistance to I/R injury.

Cardiac hypertrophy

Under normal aerobic conditions, the adult heart prefers fatty acid oxidation for energy production. Anaerobic glycolysis accounts for only about 5% of total ATP generation [124], and pyruvate and lactate oxidation less than 40%. Even when no exogenous fatty acids are provided to an isolated working heart, fatty acids derived from breakdown of endogenous lipids still account for substantial ATP production [117]. Cardiac hypertrophy induces a switch to increased anaerobic glucose utilization, with reduced or unchanged glucose oxidation and reduced fatty acid oxidation, i.e. a reappearance of the fetal Warburg-like metabolic pattern [98, 125, 126]. This metabolic change is attributed to both transcriptional down-regulation of oxidative phosphorylation genes (by master regulatory genes such as PGC-1α) and up-regulation of glucose metabolism genes (by master regulatory genes such as HIF-1 α), and by post-translational activation of AMP-activated protein kinase (AMPK), which increases glucose uptake and stimulates glycolysis. Glucose oxidation does not increase commensurately, however, since AMPK activation also reduces pyruvate oxidation by phosphorylating pyruvate dehydrogenase kinase (PDK), which inhibits pyruvate dehydrogenase (PDH).

In addition to glycolysis, other glucose metabolic pathways including the pentose phosphate pathway, the hexosamine biosynthesis pathway and anaplerosis are also altered in hypertrophied and failing adult hearts [127] [128].

As a key enzyme in glucose metabolism, HK plays an important role in the metabolic switch during hypertrophy [129]. Increased levels and activity of HK have been measured in hypertrophied hearts. [130] Moreover, it was recently shown in transgenic mouse models that overexpressing HK2 attenuated the response to a hypertrophic stimulus, [92] and, conversely, knocking down HK2 increased susceptibility [131]. HK’s role in promoting antioxidant function via the pentose phosphate pathway (Fig. 2) was proposed as the mechanism [92]. Despite the conversion to a metabolic profile resembling the neonatal pattern, however, hypertrophied adult hearts do not in general exhibit reduced susceptibility to I/R injury, and in most, but not all studies, can still be protected by IPC [95-97].

Heart failure

In the failing heart, the metabolic changes are more variable [128], influenced by both the pathogenesis and stage of heart failure at which the metabolic profile has been characterized. Both glycolytic and oxidative phosphorylation genes are usually down-regulated in severe heart failure, and glucose oxidation is reduced. However, fatty acid oxidation can be either increased or decreased (for review, see [98]). Thus, unlike the Warburg pattern characteristic of cancer, neonatal and hypertrophied adult hearts, failing adult hearts exhibit a spectrum of changes in which both energy production and energy transfer are impaired, described as an “energy-starved” profile [132]. Pharmacologic interventions that increase glucose oxidation (even when concomitantly inhibiting fatty acid oxidation) have been shown to be beneficial in failing hearts [133-141].

HK levels in heart failure are variable, and, instead of increased, may be unchanged [132] or reduced [142]. Nevertheless, some studies have reported that failing hearts are intrinsically less susceptible to I/R injury and less protected by IPC, resembling maximally cardioprotected neonatal hearts [143]. Whether HK overexpression is beneficial to the failing adult hearts is unknown at the present time, but is an important issue, since therapies to prevent I/R injury are highly clinically relevant to the human heart failure population.

Diabetic heart

Development of diabetes is associated with a shift in cardiac metabolism away from glucose metabolism towards fatty acid oxidation. This metabolic remodeling is accompanied by a significant decrease in cardiac HKII, but not HKI, protein content [62]. In addition, the diabetic heart displays an altered response to I/R and IPC, with short-term diabetes frequently offering protection against I/R but an attenuated response to IPC, whereas long-term diabetes results in both worsened outcome after I/R and loss of IPC protective effects [144]. Changes in HK may play a role in these altered responses of diabetic hearts to I/R and IPC. Indeed, altered cardiac HK cellular trafficking has been shown in the short-term type I diabetic heart [145], with the major difference in HK cellular trafficking between healthy and diabetic hearts occurring at early reperfusion after an IPC episode. Further research will be necessary to elucidate the role of HKs in diabetic cardiomyopathy.

Summary, conclusions and therapeutic implications

HKs are powerful orchestrators of glucose metabolism, determining whether glucose is directed catabolically to glycolysis, or anabolically to glycogen synthesis, the pentose phosphate shunt, where it can be utilized to increase antioxidant power and synthesize nucleotides, amino acids or fatty acids, or the hexosamine pathway mediating O-GlycNAtion signaling. The HK2 isoform which predominates in adult heart actively translocates between mitochondria and the cytoplasm to facilitate these catabolic or anabolic functions, respectively. When bound to VDAC in the outer mitochondrial membrane, HK1 and HK2 are highly protective against I/R injury by suppressing long-lasted mPTP openings. This protective effect of HK binding to VDAC can be explained by a combination of direct (nonmetabolic) actions which inhibit mPTP pore opening, and indirect (metabolic) effects which limit futile ATP consumption by mitochondria, enhance glycolytic ATP production by glycolysis and adenylate kinase, and increase antioxidant power. In the heart, IPC strengthens HK2 binding to mitochondria possibly via Akt signaling, a component of the RISK pathway, thereby increasing resistance to mPTP opening and cell death upon reperfusion. Disrupting HK binding to mitochondria with peptides has been shown to increase I/R injury and attenuate cardioprotection by IPC, and heterozygous HK2 knock out mice show increased susceptibility to I/R injury.

Based on the scope of direct anti-cell death and favorable metabolic actions of HK’s during I/R, it is intriguing to speculate that enhanced HK binding to VDAC may be one of the long sought after end-targets of cardioprotective signaling. If so, then strategies to strengthen the HK-VDAC interaction may have therapeutic promise. Overexpression of HK2 has been shown to protect cultured neonatal myocytes from oxidative injury [53], but similar experiments have yet to be carried out in adult hearts. HK1 overexpression may be particularly promising, since HK1 remains bound to mitochondria and does not normally translocate to the cytoplasm like HK2. HK1 is also the predominant isoform in neonatal hearts, which are highly resistant to I/R injury, probably because they are already maximally cardioprotected. Cancer cells predominantly upregulate HK2 in massive amounts which confers cytoprotective effects. In hypertrophied adult hearts, however, the more modest upregulation of HK2 in concert with other glycolytic enzymes does not appear to confer increased resistance to I/R injury as in cancer cells. The difference may be that in cancer cells, basal Akt activity is typically high, thereby reinforcing HK2 binding to mitochondria. This is consistent with the finding that hypertrophied hearts can still be cardioprotected by IPC, suggesting that HK2 binding to mitochondria is not maximized. It should now be possible to explore these issues and the therapeutic potential of HK-based gene delivery techniques in adult hearts, by comparing the ability of HK1 and HK2, and corresponding HK mutants which lack either enzymatic activity or the mitochondrial binding motif, to protect against I/R injury.

Highlights.

• Hexokinases orchestrate glucose metabolism, antioxidant function and cell death

• These functions are dynamically regulated by hexokinase’s subcellular localization

• Hexokinase expression is modified in a variety of disease states

• Hexokinase may be a key target of cardioprotective signaling in heart