Occult Cardiotoxicity—Toxic Effects on Cardiac Ischemic Tolerance

Abstract

The outcome of cardiac ischemic events depends not only on the extent and duration of the ischemic stimulus but also on the myocardial intrinsic tolerance to ischemic injury. Cardiac ischemic tolerance reflects myocardial functional reserves that are not always used when the tissue is appropriately oxygenated. Ischemic tolerance is modulated by ubiquitous signal transduction pathways, transcription factors and cellular enzymes, converging on the mitochondria as the main end effector. Therefore, drugs and toxins affecting these pathways may impair cardiac ischemic tolerance without affecting myocardial integrity or function in oxygenated conditions. Such effect would not be detected by current toxicological studies but would considerably influence the outcome of ischemic events. The authors refer to such effect as “occult cardiotoxicity.” In this review, the authors summarize current knowledge about main mechanisms that determine cardiac ischemic tolerance, methods to assess it, and the effects of drugs and toxins on it. The authors offer a view that low cardiac ischemic tolerance is a premorbid status and, therefore, that occult cardiotoxicity is a significant potential source of cardiac morbidity. The authors propose that toxicologic assessment of compounds would include the assessment of their effect on cardiac ischemic tolerance.

Ischemic Tolerance and “Occult Toxicity”

In recent years, it has become evident that the outcome of cardiac ischemic events depends not only on the size, intensity, and duration of the ischemic stimulus but also, to a large extent, on the intrinsic defensive mechanisms of the myocardium itself (Ravingerová 2007). These protective mechanisms provide tolerance to myocardial ischemia/reperfusion (Peart and Headrick 2008). Ischemic tolerance is most potently induced by ischemic preconditioning: an adaptive response in which exposure to brief cycles of ischemia/reperfusion slows myocardial ATP consumption and markedly increases the viability of the heart when exposed to a subsequent ischemic injury (Murry, Jennings, and Reimer 1986; Reimer et al. 1986). The potency and reproducibility of ischemic preconditioning in all species examined has led to its establishment as the “gold standard” for studying the mechanisms and potential intensity of cardioprotection (Ferdinandy, Schulz, and Baxter 2007). The cardioprotective effect of preconditioning is biphasic: The first phase, referred to as “classic preconditioning” starts minutes after the preconditioning stimulus and lasts up to 2 hours (Murry, Jennings, and Reimer 1986). The second phase, “late preconditioning” or “second window of preconditioning,” starts 24 hours after the preconditioning stimulus and lasts up to 96 hours (Marber et al. 1993). As indicated by their time-course, classic preconditioning is mediated by modifications of existing proteins in cardiac myocytes, whereas late preconditioning results from adaptive transcriptional regulation and protein synthesis and requires alterations in the array of genes expressed by the myocardium (Bolli 2007). Induction of survival proteins such as Bcl-2 are required for the activation of late preconditioning (Rajesh et al. 2003).

Brief cycles of ischemia/reperfusion confer myocardial protection from reperfusion injury also when applied after a prolonged period of ischemia, a phenomenon referred to as “ischemic postconditioning” (Zhao et al. 2003). Other genetic, physiologic, and pharmacologic stimuli, besides ischemia, have also been found to confer cardioprotection by modulating ischemic tolerance (Chen et al. 2008; Ferdinandy, Schulz, and Baxter 2007; Ravingerová 2007). Thus, the basal status of cardiac ischemic tolerance and the recruitability of further cardioprotective mechanisms play a major role in determining the damage following an ischemic stimulus.

The treatment of acute myocardial infarction has dramatically evolved in the past two decades, mainly due to efficient means of tissue reperfusion and revascularization by fibrinolysis, percutaneous coronary intervention, or emergency coronary artery bypass surgery. Prompt reperfusion enables the salvage of myocardial tissue and has led to the idiom “time is muscle, and muscle is life” (Simoons et al. 1997). Although reperfusion is essential to salvage ischemic myocardium, it has the potential of causing further cell injury. Intrinsic defense mechanisms in the myocardium protect cardiac myocytes not only from ischemic injury but also from reperfusion injury. The term “ischemic tolerance” in the current literature often represents broader protective mechanisms that confer relative tolerance to the damage of ischemia/reperfusion (Ferdinandy, Schulz, and Baxter 2007). Thus, in the era of reperfusion therapy, the role of cardiac intrinsic defensive mechanisms in determining the outcome of ischemic events is triple: protecting cardiac tissue from ischemia, from reperfusion damages, and slowing the development of irreversible damage to cardiomyocytes, allowing essential time for therapeutic reperfusion.

Cardiac ischemic tolerance is determined by multiple genetic and environmental stimuli. Chen et al. (2008) have recently described mutations in the gene encoding aldehyde dehydrogenase 2 (ALDA-2), prevalent in the East Asian population, that lead to reduced cardiac ischemic tolerance. Myocardial ischemic tolerance is constantly modulated by metabolic, neuroendocrine, paracrine, and autocrine stimuli, transmitted via various signal transduction pathways (Ravingerová 2007) and affecting mainly myocardial mitochondria (Rajesh et al. 2003; McLeod, Pagel, and Sack 2005; Schwartz and Sack 2008). Increasing intrinsic myocardial ischemic tolerance has been proposed as a therapeutic target to combat ischemic heart disease (Yellon and Downey 2003; Ferdinandy, Schulz, and Baxter 2007; Schwartz and Sack 2008). As a mirror image, it is conceivable that different drugs, environmental pollutants, or other exogenous stimuli reduce basic ischemic tolerance and/or attenuate the important protective mechanisms of ischemic preconditioning by inhibiting signal transduction pathways and cellular components governing ischemic tolerance. Exposure to such compounds would be potentially harmful, but their effect would be occult: it would not be immediate and would not be apparent as long as the myocardium receives adequate supplies of oxygen and nutrients. However, it would render the heart more susceptible to subsequent ischemic injuries, increasing the potential significance of small or short ischemic events (Figure 1). Exposure to such compounds may accelerate irreversible tissue damage, so that reperfusion therapy becomes insufficient or too late.

Despite the wide attention given to changes in cardiac ischemic tolerance in the cardiologic, physiologic, and pharmacologic literature, data about ischemic tolerance is scarce in toxicological journals: a decrease in ischemic tolerance is still not considered a true cell damage, and therefore compounds that cause it are not considered toxic. Agents can affect ischemic tolerance either directly or indirectly by interfering with the recruitment of cardioprotective mechanisms that increase ischemic tolerance when needed, such as preconditioning or postconditioning.

The potentially harmful effect of compounds that reduce intrinsic cardiac protective mechanisms eludes classic toxicological studies and can be elucidated only by direct assessment of cardiac ischemic tolerance following exposure to an agent. Therefore, we refer to decreasing of cardiac ischemic tolerance by exposure to exogenous compounds as “occult cardiotoxicity.”

AIMS

The aims of the present review are

1. to summarize the main mechanisms that determine cardiac tolerance to ischemia/reperfusion, the methods to assess it, and the data on effects of drugs and toxins on it;

2. to express a view that low cardiac ischemic tolerance is a premorbid status, and therefore occult cardiotoxicity is a significant potential source of cardiac morbidity; and

3. to offer new approaches to assess occult cardiotoxicity and recommend that the toxicologic assessment of compounds would include the assessment of their effect on cardiac ischemic tolerance.

Assessment of Cardiac Ischemic Tolerance

In general, assessment of cardiac ischemic tolerance is based on the evaluation of the intensity of myocardial damage in response to ischemia/reperfusion. Different methods, in vivo, ex vivo, and in vitro have been used to assess cardiac ischemic tolerance. The lack of standard techniques may lead to discrepancies among results from different laboratories.

The terms ‘ex vivo’ and ‘in vitro’ are interchangeably used in some of the literature. In this text, in vivo methods refer to assessment of cardiac function and morphology in the intact animal; ex vivo methods refer to studies on isolated intact hearts removed from experimental animals, in Langendorff or working heart perfusion systems; and in vitro methods refer to methods that use isolated tissue fragments, isolated cardiac myocytes, cultured myocytes, or isolated organelles. In vivo and ex vivo methods are based on direct evaluation of the damage to the whole heart in response to infarction or to global ischemia. In vitro methods usually evaluate tissue damage by biochemical and physiologic correlates of viability or damage, such as tests for mitochondrial activity, leak of tissue proteins, or responsiveness of ion currents to exogenous stimuli. Two basic approaches are applied in examining the effect of compounds on ischemic tolerance ex vivo or in vitro: one involves both the exposure of the tissue to the tested compound and the assessment of ischemic tolerance ex vivo/in vitro. The other involves exposure of experimental animals to the tested compounds in vivo and then removal of the heart and evaluation of its ischemic tolerance ex vivo/in vitro.

In Vivo and Ex Vivo Methods—Endpoints of Myocardial Injury

Myocardial ischemic injury leads to three main direct adverse consequences (often referred to as “endpoints”): (1) loss of myocardial tissue due to irreversible damage to cardiomyocytes, by necrosis and programmed cell death; (2) arrhythmias; and (3) decrease in contractile function. These may lead to sudden death and to various degrees of heart failure. Classic in vivo or ex vivo experimental evaluation of cardiac ischemic tolerance is based on the quantitative assessment of at least one of these parameters following an ischemic injury. Thus, assessment of myocardial ischemic tolerance is based on the measurement of the decrease in infarct size, in the number of arrhythmic events, or in the postischemic contractile dysfunction of the surviving tissue or the whole heart.

Since the demonstration of the phenomenon of ischemic preconditioning (Murry, Jennings, and Reimer 1986), a decrease in infarct size has been widely used as an endpoint for assessment of ischemic tolerance (Schott et al. 1990; Zhao et al. 2003; Lochner, Genade, and Moolman 2003; Dikow et al. 2004; Liem et al. 2005; Kocsis et al. 2008). Decreasing infarct size is thus considered the gold standard for increasing cardiac ischemic tolerance by any physiologic or pharmacologic intervention. It was used in landmark studies of ischemic tolerance, including the following:

1. the demonstration of “postconditioning” by Zhao et al. (2003), who showed that the heart can be protected against the extension of ischemia-reperfusion injury if brief (10–30 sec.) coronary occlusions are performed at the beginning of the reperfusion;

2. the demonstration that whole body hyperthermia is associated with cardioprotection (Currie, Tanguay, and Kingma 1993);

3. the finding of various ligands and receptors that play a major role in the mechanism of preconditioning, including the demonstration of the pivotal role of adenosine and its specific receptor A1 in the cardioprotective process (G. S. Liu et al. 1991);

4. the demonstration that drugs, such as α-adrenergic agonists, δ-opioid agonists, nitric oxide donors, and estrogen, can increase cardiac tolerance to ischemia-reperfusion (occasionally referred to as “pharmacologic preconditioning”) (Hale and Kloner 1994; Hale, Birnbaum, and Kloner 1997; Sharma and Singh 1999; Fryer et al. 2000; Kukreja et al. 2005); and

5. the demonstration that disease states, such as uremia and chronic diabetes mellitus, as well as aging, are associated with low ischemic tolerance and decreased potential to undergo preconditioning and postconditioning (Forrat et al. 1993; Schulman, Latchman, and Yellon 2001; Dikow et al. 2004; Ebrahim, Yellon, and Baxter 2007a, 2007b; Przyklenk et al. 2008).

Arrhythmias often result from ischemic injuries. They may develop immediately after the ischemic stimulus, during recovery from ischemia, and/or during reperfusion. In cardiac ventricular tissue, they range from isolated ventricular premature beats, through ventricular tachycardia, to life-threatening ventricular fibrillation, accounting for most cases of sudden death resulting from myocardial infarction (Janse and Wit 1989). The number and duration of arrhythmias that follow an experimental coronary artery occlusion constitute a second criterion for cardiac tolerance to ischemia-reperfusion. Although the occurrence of arrhythmias does not necessarily involve irreversible damage to myocytes, most stimuli that decrease infarct size were also found to protect the heart from arrhythmias. These include preconditioning (Hagar, Hale, and Kloner 1991; Vegh, Komori, et al. 1992; Vegh, Szekeres, and Parratt 1992; Vegh, Szekeres, and Parratt 1991a, 1991b; Ravingerová et al. 2000, 2002; Ravingerová 2007), postconditioning (Kloner, Dow, and Bhandari 2006), whole body hyperthermia (Cornelussen et al. 1994, 1997), and pharmacologic agents that increase ischemic tolerance (“pharmacologic preconditioning”) (Tosaki et al. 1995; Fryer et al. 2000; Vegh and Parratt 2002).

Contractile dysfunction of viable myocytes, in the forms of myocardial stunning and myocardial hibernation, often follows ischemia and may contribute to the development of heart failure. Upon reperfusion, this tissue may fully recover and regain its contractile function. Enhanced ischemic tolerance may reduce postischemic contractile dysfunction. Thus, a decrease in the magnitude of postischemic contractile dysfunction serves as the third endpoint in the direct assessment of cardiac tolerance to ischemia-reperfusion (Banerjee et al. 1993; Sun et al. 1995; Takano et al. 2000). However, it should be borne in mind that functional recovery of the myocardium represents both loss of myocytes due to irreversible injury and delayed recovery of viable myocardium.

In the evaluation of ischemic tolerance, assessment of the consequences (infarct size, arrhythmias, and contractile dysfunction) is highly specific but not sensitive. It is the most direct way to evaluate ischemic tolerance. However, it requires sacrificing a large number of animals for the assessment of the effect of each stimulus on ischemic tolerance; it detects only dramatic effects on ischemic tolerance, such as those conferred by preconditioning and postconditioning, but not small changes (20%–30%); and it does not provide a clear numerical index that represents ischemic tolerance, which can be used in the screening for compounds that improve or worsen it. For such assessment, and for assessment of ischemic tolerance in human tissue, in vitro indirect correlates of cardiac ischemic tolerance have potential advantages.

In vitro methods for ischemic tolerance assessment employ small fragments of cardiac tissue or isolated myocytes and are based on comparisons of tissue viability (or damage) following ischemia/reoxygenation as compared to viability at control oxygenation. They enable measurements of fluxes of ATP and phosphocreatine in the tissue and evaluation of the biochemical pathways used in coping with ischemia/reoxygenation. The approach enables using identical cells, or tissue from the same heart, in both ischemic and oxygenated conditions. The ratio between the conditions (viability in ischemia/reperfusion divided by viability in control oxygenation) can provide a numerical index for ischemic tolerance. Thus, this approach is quantitative and sensitive. It enables testing human tissue obtained during cardiac surgery. It also enables large-scale screening of the effect of compounds on ischemic tolerance when tissue from experimental animals is used. Physiologic parameters such as velocity and force of contraction can also be assessed in vitro. Obviously, results from in vitro studies do not always correlate with results in vivo, and the correlation often depends on the in vivo parameter (endpoint) chosen for the comparison. Despite its potential for large-scale screening, sensitivity, and quantitative nature, only a few laboratories employ true in vitro approaches for assessment of cardiocyte ischemic tolerance, mainly for work on human tissue.

In the in vitro evaluation of ischemic tolerance, myocardial tissue is maintained in media containing glucose and aerated with oxygen. Ischemia is simulated by replacement of the bubbled oxygen with nitrogen (or another inert gas) and removal of glucose from the medium. Reoxygenation is simulated by restoration of oxygen and glucose supply (Zhang et al. 2000; Ghosh, Standen, and Galiñanes 2001; Loubani and Galinanes, 2001; Loubani, Ghosh, and Galiñanes 2003; Loubani, Hassouna, and Galiñanes 2004; Ad et al. 2005; Hassouna et al. 2006; Golomb et al. 2007). Figure 2 illustrates system of the in vitro experiment of simulated ischemia.

Biochemical markers of myocardial injury for in vitro assessment of cardiac ischemic tolerance include mainly the leak of creatine kinase or lactate dehydrogenase into the medium. A commonly used viability marker is the pigment 3-(4,5)-dimethylthiazol-2-yl)-2,5-diphenyl-tetrazolium bromide (MTT). In vitro evaluation has been used to examine the role of biochemical pathways in cardioprotection and the effect of different physiologic and pathologic conditions, as well as drugs, on cardiac ischemic tolerance. Physiologic and pathologic stimuli that have been studied by in vitro methods include preconditioning and its interaction with aging (Ovelgönne et al. 1996; Ghosh, Standen, and Galiñanes 2000b; Loubani and Galiñanes 2002b; Loubani, Ghosh, and Galiñanes 2003), aging per se (Mariani et al. 2000; Loubani, Ghosh, and Galiñanes 2003, O’Brien, Ferguson, and Howlett 2008), diabetes mellitus (Hassouna et al. 2006), and other pathological conditions leading to heart failure (Ghosh, Standen, and Galiñanes 2001). In vitro assessment of the effect of drugs on ischemic tolerance has been used for many agents including opioid agonists (Mühlfeld et al. 2006), adenosine agonists (Pomerantz et al. 2000), Na/H and Na/Ca exchanger inhibitors (Simm et al. 2008; MacDonald and Howlett 2008), isoflurane (Dworschack, Breukelmann, and Hannon 2004), nicorandil (Loubani and Galiñanes 2002a), cyclosporine A (Schneider et al. 2003), and sulfonylurea hypoglycemic agents (Cleveland et al. 1997; Schneider et al. 2003; Loubani et al. 2005; Hassouna et al. 2006).

Recently, we checked the effect of a known cardiotoxin, Bis(2-chloroethoxy) methane (CEM) on ischemic tolerance. CEM exerts cardiotoxic effect mainly through mitochondrial damage (Dunnick et al. 2004, 2006). Topical application of 400 and 600 mg/Kg/d cause time- and dose-dependent myocardial injury, expressed morphologically as perinuclear vacuolation (Dunnick et al. 2004). At a dose of 100 mg/Kg/d, CEM does not cause overt myocardial morphologic or functional damage. Using the in vitro approach of assessing tissue viability in ischemic versus oxygenated conditions, we found that this subtoxic dose of CEM decreases ischemic tolerance but does not affect myocardial viability in oxygenated conditions. Figure 3 (Golomb et al. 2007) illustrates the finding: panel A shows that 100 mg/Kg/d CEM had no effect on mitochondrial activity in oxygenated conditions, whereas 400mg/Kg/d decreased cardiac mitochondrial activity. However, in conditions of simulated ischemia/reoxygenation (panel B), the occult cardiotoxic effect of 100 mg/Kg/d CEM surfaced: MTT activity of hearts of these rats was approximately half that of control rats and similar to this of rats that received 400 mg/Kg/d CEM. Thus, 100 mg/Kg/d CEM did not affect cardiac viability in oxygenated conditions, but decreased ischemic tolerance, indicating the occult cardiotoxic effect of this CEM dose. In other words, low doses of the cardiotoxin, considered “subtoxic,” may damage the heart by reducing its protective capacity from ischemia. Furthermore, this result indicates that assessment of ischemic tolerance in vitro may be utilized as a sensitive tool to for evaluation of cardiotoxicity and possess advantages to current biochemical and morphological analyses.

It should be noted that the optimal experimental systems to examine compounds that improve cardiac ischemic tolerance differ from those utilized to study occult cardiotoxicity: In the former, the simulated ischemia/reoxygenation protocol should be such that it produces a significant injury in untreated tissue, so that treatment with an agent that improves ischemic tolerance would reduce the injury. On the other hand, in the latter, the simulated ischemia/reoxygenation protocol should cause only minimal injury in untreated tissue but would be significant in tissue that was preexposed to the occult cardiotoxic agent.

To summarize, various in vivo, ex vivo, and in vitro methods have been used for the assessment of cardiac ischemic tolerance. In vivo and ex vivo ischemic tolerance assessment is based on cardiac damage caused by an event of ischemia/reperfusion, wherein the damage is evaluated by one or more of the following endpoints: (1) infarct size; (2) nature, frequency, and duration of arrhythmic events; and (3) decreased cardiac contractility. A decrease in infarct size is considered the “gold standard” for cardioprotection. In vitro assessment of ischemic tolerance is based on biochemical markers, at full oxygenation compared to simulated ischemia/reoxygenation. The effect of stimuli on ischemic tolerance depends, to some extent, on the endpoint chosen for evaluation and the method chosen for its assessment. Different assessment methods usually give rise to similar results, though not identical. There is no standard single adopted method for the evaluation of myocardial susceptibility to ischemia/reperfusion.

Mechanisms of Ischemic Tolerance Modulation

The molecular/cellular basis of ischemic tolerance is multi-factorial. Ischemic tolerance is dynamic, constantly adapting to changing cellular environment. Various endocrine, neurocrine, paracrine, and autocrine stimuli have been shown to modulate cardiac ischemic tolerance, via different signal transduction pathways, activation of transcription factors, and reprogramming of the array of genes expressed by the myocardium. It is generally agreed that the most important organelle in ischemic tolerance modulation is the mitochondrion (Figure 4).

The mitochondrion modulates ischemic tolerance through (Sack 2006) the following:

1. regulation of cellular energy reserves, by modulation of ATP production rate and adjustment of ATP consumption through uncoupling proteins;

2. generation of oxygen free radicals and its adjustment to the balance with cellular antioxidants;

3. regulation of programmed cell death in response to injurious stimuli (Elmore 2007);

4. modulation of hypoxia-induced factors HIF-1α and HIF-2α, which regulate pathways of gene expression and protein degradation—the mitochondria act as oxygen sensors, determining the levels of active HIFs (Guzy et al. 2005); and

5. retrograde modulation of signal transduction pathways (Nemoto et al. 2000).

Early in the response to ischemia, mitochondrial membrane depolarization occurs, which probably results from inhibition of electron transfer and activation of uncoupling proteins. These changes contribute to a reduction in the production of reactive oxygen species (ROS), as well as in the uptake of calcium ions (Sack 2006). The study of ischemic preconditioning revealed that ischemia upregulates mitochondrial antioxidant defense mechanisms not only by inhibiting mitochondrial respiration but also through increased activity of mitochondrial uncoupling proteins (McLeod, Pagel, and Sack 2005) and by increasing levels of antioxidant enzymes (Yamashita et al. 1994). Thus, the mitochondrion is not only a passive responder to ischemia, which leads to cellular injury by ATP depletion and to reperfusion damage by formation of ROS. Rather, the mitochondrion actively and dynamically modifies the biological response to ischemia and reperfusion. Therefore, cellular damage due to ischemia can be viewed as a failure of the mitochondrial defense mechanisms that form ischemic tolerance.

The main mitochondrial modulators of ischemic and reperfusion injury are the mitochondrial permeability transition pore (mPTP), the mitochondrial ATP dependent potassium channel (mito-K-ATP), the mitochondrial connexin-43 (mito-Cx43), and the mitochondrial enzyme aldehyde dehydrogenase-2, so that ischemic tolerance depends on a status of low levels of formation and opening of mPTPs, on high levels of opening of mito-K-ATP, and on high enough activity of Cx-43 and aldehyde dehydrogenase-2.

Mitochondrial Permeability Transition Pore (mPTP)

mPTP is a protein pore, formed in mitochondrial membranes mainly in response to injury. Formation and opening of mPTP is a key event in the induction of programmed cell death in response to injury (Crompton 1999; Halestrap, Clarke, and Javadov 2004; Clarke et al. 2008). From an evolutionary point of view, mPTP probably led to an advantage by eliminating severely injured cells in tissues of young and healthy organisms, preventing them from forming injurious radicals, capable of damaging neighboring cells, thus preventing disturbances in cellular synchronization (e.g., arrhythmias) that stem from the existence of damaged living cells along the axis of cardiac electrical currents. However, in the context of more massive tissue injury such as an infarct, programmed cell death accentuates tissue damage. Excessive programmed cell death is a key event in various major pathological states, such as neurodegenerative disorders, stroke, and ischemic heart disease, so that its inhibition is a major challenge in cytoprotection. Molecular events occurring during ischemia play a major role in determining whether mPTP forms and opens during reperfusion. Ischemic preconditioning, drugs that mimic preconditioning (chemical preconditioning), and postconditioning attenuate mPTP opening (Ferdinandy, Schulz, and Baxter 2007). It was recently shown (Clarke et al. 2008) that inhibition of mPTP opening by ischemic preconditioning is probably mediated by decreased oxidative stress rather than mitochondrial protein phosphorylation. Thus, mPTP formation and opening is a final common pathway to cell death, modulated by cellular receptors and signal transduction pathways, which determine myocyte susceptibility to ischemia.

Mitochondrial ATP-Dependent Potassium Channels (Mito-K-ATP)

Potassium currents across the mitochondrial membrane affect mitochondrial membrane action potential, mitochondrial volume, ATP production, production of and responsiveness to uncoupling proteins, and calcium homeostasis. The inner mitochondrial membrane is impermeable to potassium, so that potassium cycling across the mitochondrial membrane is enabled by specific potassium pumps and channels (Ferdinandy, Schulz, and Baxter 2007). Mito-K-ATP is the main potassium pathway across the mitochondrial membrane. It is mostly closed at physiologic conditions of mitochondrial membrane potential of ~(−180mV). Mild depolarization of the mitochondrial membrane by Mito-K-ATP openers such as pinacidil, cromakalin, or levcromakalin confers an increase in mitochondrial respiration, ROS formation, volume, and calcium release (Holmuhamedov et al. 1998). Thus, mitochondrial functions are modulated by the level of opening of mito-K-ATP. Furthermore, mito-K-ATP opening has been shown to be a major end effector that determines the fate of cells after ischemia-reperfusion. Therefore, it has been suggested as the main end effector of ischemic preconditioning (Ardehali and O’Rourke 2005; Matejíková et al. 2009). Selective inhibition of mito-K-ATP has been shown to abolish the cardioprotective effect of ischemic preconditioning in different species (Gross and Auchampach 1992; Munch-Ellingsen et al. 2000; Sato, O’Rourke, and Marbán 1998; Sato et al. 2000; Ravingerová et al. 2002; Loubani and Galiñanes 2002a), and mito-K-ATP openers mimic preconditioning-induced cardioprotection (Grover et al. 1994; Garlid et al. 1997, 2003; Gross and Fryer 1999).

It should be noted that similar K-ATP channels are found also in the sarcolemma, mediating cytosolic potassium influx. Hypoxia leads to opening of sarcolemmal K-ATP channels, leading to shortening of the action potential and to increased osmotic load–potential arrhythmogenic effects. It has been suggested that opening of sarcolemmal K-ATP confers cardioprotection (Gross 1995). Since the elucidation of the pivotal role of mito-K-ATP in cardioprotection, the relative contribution of sarcolemmal K-ATP has been a subject of debate (Garlid et al. 1997; Y. Liu et al. 1998; Gross and Fryer 1999; Kita et al. 2000; Sato et al. 2000).

Connexin 43 (Cx-43)

Connexins are gap junction proteins, which are essential for many physiological processes, such as the flux of small molecules between cells, and coordinated depolarization of cardiac muscle. Cx-43 is the main myocardial connexin in gap junctions, especially in the intercalated disc. Cx-43 is also localized at the inner mitochondrial membrane. Mitochondrial Cx-43 is essential in the mitochondrial formation of reactive oxygen species, which leads to the activation of the signal transduction cascades of the cardioprotective effect of ischemic preconditioning (Boengler, Schulz, and Heusch 2006; Schulz et al. 2007).

Aldehyde Dehydrogenase-2 (ALDA-2)

The recent finding by Chen et al. (2008) of a mitochondrial enzyme, aldehyde dehydrogenase-2, whose activation increases ischemic tolerance, offers a new direction for studying the modulation of ischemic tolerance. Aldehyde dehydrogenase-2 is activated by protein kinase C, a main signal transduction pathway in modulating cardioprotection. Moreover, it has been previously shown that exposure to ethanol can increase cardiac ischemic tolerance (Chen, Gray, and Mochly-Rosen 1999). The finding that ethanol activates aldehyde dehydrogenase-2 provides a mechanism to explain this finding (Chen et al. 2008). The effect of additional signal transduction pathways on the expression and activity of this enzyme would probably shed new light on the potential effect of exogenous substances on ischemic tolerance. Thus, inhibition of aldehyde dehydrogenase-2 by exogenous substances may be a potential source of occult cardiotoxicity.

Mitochondria have been shown to be a prime target of various toxins, by morphologic and functional assays. For example, ZDV/3TC has been shown to damage mitochondria ultrastructurally (Figure 5). It is therefore conceivable that such compounds, even at low doses, would render the heart susceptible to ischemic injuries.

Significant Signal Transduction Pathways and Transcription Factors in Ischemic Tolerance

Multiple signal transduction pathways converge on the mitochondrion, and affect its different functions (Figure 4): regulation of ATP production, synthesis of uncoupling proteins, opening of mPTP, opening of mito-K-ATP, and the activity of Cx-43 and of aldehyde dehydrogenase-2. Signal transduction pathways that are activated by ischemic preconditioning, thus potentially conferring cardioprotection, have been subject to extensive investigation (Armstong 2004; Hausenloy and Yellon 2006). Signal transduction protein kinases proven to regulate cardiac ischemic tolerance are ones that transduce survival signals and include the following:

• Protein kinase C (PKC): activation of PKC through phospholipase C or D has been shown to be a crucial step in both early (Ytrehus, Liu, and Downey 1994) and late (Baxter, Goma, and Yellon 1995) preconditioning-induced cardioprotection. PKC activation leads to the opening of mito-K-ATP (Sato, O’Rourke, and Marbán 1998, Gross and Fryer 1999) and to activation of aldehyde dehydrogenase-2 (Chen et al. 2008). PKC isoforms that mediate ischemic tolerance modulation vary among studies, depending mainly on the species examined (Budas et al. 2007). In most studies, the PKC-ɛ isoform is a major mediator of cardioprotection.

• Phosphatidyl inositol 3 kinases (PI3K), their downstream kinase Akt, and associated pathway glycogen synthase kinase 3β (GSK-3β) constitute a major survival pathway (Tong et al. 2002; Juhaszova et al. 2004). Activated Akt phosphorylates antiapoptotic and proapoptotic proteins, activating the antiapoptotic substrates and deactivating the proapoptotic ones (Franke et al. 2003; Mocanu and Yellon 2007). PI3K/Akt pathway has been shown to play a major role in myocardial ischemic tolerance in all examined species (Hausenloy and Yellon 2004) and to mediate the cardioprotective effect conferred by ischemic preconditioning (Tong et al. 2002; Mocanu, Bell, and Yellon 2002; Hausenloy et al. 2005) and by postconditioning (Zhu et al. 2006). Pharmacologic enhancement of ischemic tolerance by various agents has been attributed to activation of this pathway, including insulin (Jonassen et al. 2001), statins (Bell and Yellon 2003a), bradykinin (Bell and Yellon 2003b), and erythropoietin (Bullard, Govewalla, and Yellon 2005).

• Mitogen activated protein kinases (MAPKs): MAPKs are highly conserved serine/threonine kinases that are activated by dual phosphorylation on a Thr-X-Tyr motif, in response to many stimuli, including growth factors, activation of G protein–coupled receptors, and environmental stresses. MAPKs modulate a variety of functions, including survival, growth, and proliferation. Some MAPKs are proapoptotic, whereas others mediate growth and survival. Thus, the balance between opposing activities of these kinases determine the fate of the cell. Three major MAPK cascades are known to modulate cardiac ischemic tolerance: p38 MAPKs, the extracellular signal–regulated kinases (ERK1 and ERK2), and c-Jun N-terminal kinases (JNK1 and JNK2) (Yue et al. 2000; Ping and Murphy 2000). In vitro and in vivo studies indicate that cardiomyocyte damage due to ischemia/reperfusion is partially mediated by proapoptotic MAPKs, and the ischemic tolerance offered by preconditioning is partly mediated by shifting the MAPK balances towards survival (Yue et al. 2000; Mocanu et al. 2000; Steenbergen 2002). Major transcription factors that mediate ischemic tolerance as downstream activators of these signal transduction pathways include nuclear factor κB (NF-κB) and signal transducers and activators 1 and 3 (STAT1 and STAT3) (Xuan et al. 1999, 2001).

Upstream of these kinases, Gi/q protein, activated by Gi/q protein coupled receptors such as adenosine A1 and A3 receptors, muscarinic, δ-opioid or α-adrenergic receptors have been shown to mediate ischemic tolerance in a wide variety of studies (Ravingerová 2007; Ferdinandy, Schulz, and Baxter 2007). Pretreatment with pertussis toxin, inactivating Gi/q protein, blocks the protective effects of preconditioning (Thornton, Liu, and Downey 1993; Schultz et al. 1998).

An additional, somewhat unexpected mediator of cardiac ischemic damage is the “guardian of the genome” and apoptosis mediating protein p53. Induction of p53 by ischemia or stress activates proapoptotic pathways and contributes to cellular damage. p53 inhibition following myocardial infarct has been shown to exert some beneficial effect (Matsusaka et al. 2006). Thus, maintenance of chronic expression and activity of proapoptotic proteins are a crucial defense mechanism against cancer that is obtained at a price: activation of such proteins may contribute to ischemic damage and increase susceptibility to ischemia. Temporary inhibition of such proteins following ischemia could increase ischemic tolerance in a beneficial manner. Given the well-established beneficial role of proapoptotic proteins such as p53, the design of inhibitors for them has not been widely sought but could become a new challenge for drug designers as a temporary cardioprotective agent (Vousden and Lane 2007). In general, since the basic mechanisms that confer cardioprotection are based on cell survival signaling, the design of chronic treatment to increase ischemic tolerance should be based on high organ specificity to avoid the risk of proliferative disorders in other organs.

Ligands That Increase Ischemic Tolerance (Pharmacological Preconditioning)

While the end effectors and general signal transduction pathways conferring ischemic tolerance are conserved, the effect of ligands on ischemic tolerance varies dramatically among studies according to species; the endpoint examined (infarct size versus arrhythmias versus loss of contractility); and the intensity of ischemia, reperfusion, and preconditioning protocol. However, in any species, ligands were found to affect ischemic tolerance: either to increase it, mimicking preconditioning, or to abolish the potential of ischemic preconditioning. Given the plethora and importance of the stimuli and signal transduction pathways that modulate ischemic tolerance, it would be expected that many drugs, as well as other exogenous stimuli, would affect it: some would activate similar pathways to preconditioning and increase ischemic tolerance, conferring “pharmacological preconditioning,” whereas others would attenuate these pathways, inhibit the recruitability of cardioprotective mechanisms, or directly reduce ischemic tolerance, conferring “occult cardiotoxicity.”

Two inducible intracellular enzymes have been shown to play a pivotal role in ischemic tolerance: inducible nitric oxide synthase (iNOS) and cyclooxygenase-2 (COX-2). These enzymes not only have been shown to be crucial for the cardioprotective effect conferred by late preconditioning but were also shown to be sufficient to induce prolonged ischemic tolerance when administered to the heart by viral vectors (Bolli 2007). Given the central role of nitric oxide and cyclo-oxygenase as targets in modern pharmacology, drugs affecting these enzymes are likely to affect cardiac ischemic tolerance (see below).

To summarize, cardiac ischemic tolerance is very dynamic, constantly adapting to changing environmental conditions and to signals to which cardiomyocytes are exposed. It is modulated by various receptors, through signal transduction pathways, transmitting signals to intracellular end effectors. The mitochondrion is the main intracellular regulating organelle. The main mitochondrial end effectors are the permeability transition pore (mPTP), K-ATP channels, and the recently recognized enzyme aldehyde dehydrogenase-2. In light of the pivotal role of the mitochondria in determining ischemic tolerance, mild mitochondrial dysfunction is likely to be reflected by a decrease in ischemic tolerance, with no effect on function in oxygenated conditions. Sensitive methods to assess mitochondrial function in conditions of simulated ischemia/reoxygenation serve for in vitro assessment of ischemic tolerance. Many ligands and receptors regulate ischemic tolerance, via the most fundamental survival signal transduction pathways, and therefore various drugs affect ischemic tolerance. Increasing ischemic tolerance would induce “pharmacologic preconditioning,” and decreasing it would cause “occult cardiotoxicity.”

Physiologic and Pathologic States of Altered Ischemic Tolerance

Most studies on ischemic tolerance and its mechanisms have been carried out using healthy experimental animals, in most cases young ones. It has become evident that physiological parameters such as aging (Ashton et al. 2003, 2006; Willems, Garnham, and Headrick 2003;Willems, Ashton, and Headrick 2005; Willems et al. 2005), caloric intake (Doenst et al. 1996; Long et al. 2002; Shinmura, Tamaki, and Bolli 2005, 2008), dietary ingredients (Venardos et al. 2005; Mancardi et al. 2009), and physical activity (Margonato et al. 2000; Hamilton et al. 2001) affect basic cardiac ischemic tolerance and its responsiveness to cardioprotection by ischemic and/or chemical preconditioning. Obviously, people differ in their basic ischemic tolerance and in their responsiveness to different cardioprotective stimuli according to genetic factors, such as polymorphism of aldehyde dehydrogenase-2 (Chen et al. 2008), their age, gender (Willems et al. 2005), and various environmental factors. This should be taken into account in attempts to design treatments for enhancing ischemic tolerance or to assess occult cardiotoxicity.

Data concerning the effect of common disease states on cardiac ischemic tolerance and on recruitablity of cardioprotective defense mechanisms are not abundant. In many cases, existing data are not conclusive and depend on the experimental model, the duration and severity of the disease, and the method used to assess ischemic tolerance. The literature on this subject has been thoroughly reviewed recently by Ferdinandy, Schulz, and Baxter (2007). Systemic diseases such as hypercholesterolemia, uremia, or diabetes mellitus and cardiac pathologies such as hypertensive pressure overload or heart failure are well known to be associated with increased mortality from ischemic disease. It is therefore intriguing to find out whether reduced ischemic tolerance contributes to this increased risk and, if so, which measures can be taken to enhance ischemic tolerance and avoid further deterioration of ischemic tolerance resulting from adverse effects of drugs or other treatment modalities. It should be emphasized that such disease states usually occur in the middle age or older population, who suffer a priori from a decrease in ischemic tolerance. Effect of drugs on ischemic tolerance may also differ between health and disease. A drug that reduces ischemic tolerance in young and healthy individuals may have a different effect in older patients with underlying diseases, and might even improve ischemic tolerance in these conditions.

Hypercholesterolemia

The effect of hypercholesterolemia on ischemic tolerance depends on the model of hypercholesterolemia used and on the duration of hypercholesterolemia. Conflicting results have been published. However, most evidence suggests that hypercholesterolemia impairs cardiac ischemic tolerance, and at least partly impairs the recruitability of cardioprotective mechanisms. Most important, both Kyriakides et al. (2002) and Ferdinandy’s group (Ungi et al. 2005) showed, in patients undergoing angioplasty, that hypercholesterolemic patients suffer from reduced ischemic tolerance compared to normo-cholesterolemic ones: in the latter study, hypercholesterolemia was associated with accelerated evolution of myocardial ischemia, delayed recovery on reperfusion, and decreased response to preconditioning (Ungi et al. 2005). In experimental models, 4 weeks of hypercholesterolemia in rabbits decreased ischemic tolerance, evidenced by increased infarct size (Jung et al. 2000). Hypercholesterolemic animals have been shown in some studies to have reduced ability to recruit the cardioprotective mechanisms induced by pacing (Szilvassy et al. 1995), short ischemic stimuli (Ueda et al., 1999) or α-adrenergic stimulation (Kocić et al. 1999). Studies in double knockout mice lacking LDL receptor and ApoE revealed decreased basal ischemic tolerance: 6 to 8 months of hypercholesterolemia led to increased infarct size (Li et al. 2001). On the other hand, preconditioning was found to be maintained in hypercholesterolemic rabbits (Kremastinos et al. 2000). A study with LDL receptor deficient mice showed that 2 weeks of hypercholesterolemia renders the myocardium more susceptible to ischemia/ reperfusion injury, whereas long-term hypercholesterolemia (12 weeks) confers cardioprotection (Girod et al. 1999). This cardioprotection might be attributed to repeated small ischemic events, eliciting preconditioning-like effects, or to other mitochondrial adaptive responses that take place only after prolonged exposure to hypercholesterolemia. Another study of cardioprotective mechanisms in rabbits suffering of hypercholesterolemia for six weeks showed loss of the cardioprotective mechanism of postconditioning, while maintaining the effect of preconditioning (Iliodromitis et al. 2006).

Uremia

Dikow et al. (2004) showed increased cardiac infarct size with moderate renal dysfunction in rat, in a manner unrelated to hypertension, sympathetic overactivity, or salt retention. We are not aware of studies that directly addressed the effect of renal dysfunction on recruitability of cardioprotective mechanisms, such as ischemic preconditioning, postconditioning, or pharmacologic preconditioning. Since erythropoietin was suggested to play a role in ischemic preconditioning, and to exert a similar cardioprotective effect per se (Baker 2005), it is likely that renal failure is associated with decreased ability for ischemic preconditioning that could, at least partly, be restored by administration of erythropoietin.

Diabetes Mellitus

Experimental data on the effect of diabetes mellitus on cardiac ischemic tolerance are contradictory. In some reports, diabetes confers relative ischemic tolerance, whereas in others it is unchanged or decreased, depending on the experimental model, type of diabetes, species, and duration of hyperglycemia.

Studies that followed changes in ischemic tolerance during the development of hyperglycemia revealed that the effect of diabetes mellitus on cardiac ischemic tolerance is biphasic. At the onset of diabetes, there is a period of increased ischemic tolerance, lasting several weeks, followed by a chronic status of decreased cardiac ischemic tolerance. This pattern of response to diabetes has been found in the streptozotocin model of type 1 diabetes (Tosaki et al. 1996, Ravingerová, Neckár, and Kolár 2003; Ma et al. 2006; Ravingerová 2007), of type 2 diabetes (Y. Liu et al. 1993), and by our group in the psammomys obesus model of type 2 diabetes (unpublished data). The decreased ischemic tolerance is also associated with attenuated responsiveness to cardioprotection by preconditioning in tissue from diabetic patients (Ghosh, Standen, and Galiñanes 2001; Lee and Chou 2003) and in experimental animals (Tosaki et al. 1996; Kristiansen et al. 2004).

Hypertensive Heart Disease/Left Ventricular Hypertrophy

Given the understanding that proper myocardial ischemic tolerance depends on mitochondrial function, and the well-known data from the Framingham study showing that left ventricular hypertrophy (LVH) constitutes a most powerful independent risk factor for cardiovascular morbidity and mortality (Levy et al. 1990), it is likely that LVH would be associated with decreased ischemic tolerance. The hypertrophic myocardium in pressure overload hypertrophy shows mitochondrial dysfunction, expressed by a shift from fatty acid β-oxidation toward glucose oxidation (Huss and Kelly 2005), resulting in decreased ATP production (Seccia et al. 1998) and enhanced lactate accumulation in response to ischemia (Allard et al. 1994), reduced antioxidant mitochondrial capacity, and increased generation of reactive oxygen free radicals (Batist et al. 1989). The basic cardiac tolerance to ischemia/reperfusion has long been thought to be reduced in some models of chronic hypertension, in association with the degree of LVH: hypertensive dogs with LVH had increased size of myocardial infarction compared to normotensive ones, with a threefold increase in sudden death (Koyanagi et al. 1982; Inou et al. 1987). Hearts from spontaneously hypertensive rats (SHR) (Kohya et al. 1995), from rats rendered hypertensive by nitric oxide inhibition (Hropot et al. 1994), and from rats rendered hypertensive by aortic constriction (Linz et al. 1996) show an increased incidence of arrhythmias on reperfusion or after ischemia. Using the endpoint of contractile function SHR (Besík et al. 2007) and aortic banded rats (Friehs and del Nido 2003) also showed reduced ischemic tolerance. On the other hand, studies in other models of hypertension do not support this view: nonpreconditioned hearts from mREN 27 rats and from normotensive control rats showed equal susceptibility to the effects of ischemia/reperfusion (Randall, Gardiner, and Bennett 1997). Similarly, dogs with compensated LVH (i.e., LVH without heart failure) after aortic banding did not differ from their control counterparts in their ischemic tolerance. Impairment of ischemic tolerance in the latter model was observed only when LVH progressed into heart failure (Gaasch et al. 1990). Infarct size of DOCA-salt hypertensive rats does not differ from that of their normotensive controls (Ebrahim, Yellon, and Baxter 2007a).

Most studies that addressed cardioprotective mechanisms in the hypertensive, hypertrophic heart show that ischemic preconditioning is intact in most models of hypertension. DOCA-salt hypertension in rats did not alter the infarct size limiting effect of ischemic preconditioning but led to a loss of the cardioprotective response to bradykinin (Ebrahim, Yellon, and Baxter 2007a). The cardioprotective effect on contractile function of ischemic preconditioning, adenosine, and muscarinic receptor agonist are intact in SHR (Boutros and Wang 1995) and in Dahl-salt sensitive rats (Butler, Huang, and Gwathmey 1999). The preconditioning response was found to be even enhanced in mREN 27 rats (Randall, Gardiner, and Bennett 1997). Only when heart failure develops (see below), or when the hypertrophy is long-standing, is a decrease in the preconditioning observed (Ebrahim, Yellon, and Baxter 2007b). In summary, compensated LVH in most studies is associated with decreased baseline ischemic tolerance, with preserved ability to recruit, at least partly, cardioprotective mechanisms such as ischemic preconditioning.

Heart Failure

Heart failure is associated with multiple cellular alterations: morphologic—such as fibrosis and myocyte hypertrophy; membranal—such as downregulation of adenosine and adrenergic receptors, changes in levels and locations of various signal transduction kinases, and in the ability to translocate and to phosphorylate kinases; impairment of excitation-contraction coupling, alterations in contractile proteins (Del Monte and Hajjar 2008), and mitochondrial changes, such as decreased activity of the electron transport chain (Buchwald et al. 1990) and transition of the energy substrate from fatty acids to glucose (Sack and Kelly 1998), all of which can affect ischemic tolerance and the recruitability of cardioprotective mechanisms.

Basic cardiac ischemic tolerance status in heart failure depends mainly on the endpoint examined. There is little doubt that ischemia becomes much more arrhythmogenic in the failing heart, in humans and in animal models (Chakko et al. 1989; Bril, Forest, and Gout 1991). When infarct size in the nonpreconditioned heart is measured, there are studies that show decreased (Hoskins et al. 1996) or unaltered infarct size (Miki et al. 2000) in heart failure, depending on the experimental model.

Ischemic preconditioning has been found to be impaired or lost in human (Ghosh, Standen, and Galiñanes 2001) and animal models of heart failure (Miki et al. 2000, 2003). Using an in vitro system of atrial trabeculae of patients undergoing cardiac surgery, and exposing the tissue to conditions of simulated ischemia/reoxygenation, Ghosh, Standen, and Galiñanes (2001) compared the damage to tissue obtained from patients suffering from heart failure (left ventricular ejection fraction [LVEF] < 30%) with that obtained from patients without heart failure (LVEF > 50%). The damage of simulated ischemia/reoxygenation was similar in these groups in the nonpreconditioned protocols. However, when protocols of preconditioning were applied, they failed to confer cardioprotection to the tissue of the heart failure patients, indicating that heart failure is associated also with failure of recruitment of the preconditioning response.

Assessment of Ischemic Tolerance in Human Atrial Tissue

The phenomenon of increased ischemic tolerance following ischemic preconditioning has been found to be powerful, reproducible, and to occur in all the examined species. To study ischemic tolerance in human tissue, mostly atrial biopsies have been utilized, obtained during cardiac surgical operations. Thanks to this approach, the phenomenon of ischemic preconditioning (Ghosh, Standen, and Galiñanes 2000b); the role of mito-K-ATP in preconditioning (Ghosh, Standen, and Galiñanes 2000a); and the roles in determining ischemic tolerance of ligands such as nitric oxide and adrenergic receptors (Zhang et al. 2000; Loubani and Galiñanes 2001), of signal transduction pathways, and of the mitochondrion as end effector (Loubani and Galiñanes 2002b; Hassouna, Matata, and Galiñanes 2004) and the effect of aging on the myocardium (Mariani et al. 2000; Loubani, Ghosh, and Galiñanes 2003) have all been shown and confirmed in human tissue. The premise behind these studies was that atrial tissue reflects the human myocardium, so that similar phenomena are likely to take place also in ventricular myocardium.

Decreased ischemic tolerance may also play a role in human atrial morbidity, mainly in atrial fibrillation, which is the most common cardiac arrhythmia. Experimental and clinical atrial fibrillation show association with mitochondrial structural and functional dysfunction (Ausma et al. 2001; Lin et al. 2003; Lévy and Sbragia 2005) and with decreased expression of receptors that mediate cardioprotective pathways such as δ-opioid receptors (Lendeckel et al. 2005). Studies of postoperative atrial fibrillation (POAF) after coronary artery bypass grafting indicate that this common complication is associated with decreased atrial ischemic tolerance: Ad et al. (2005) have shown that patients whose atria show low ischemic tolerance in vitro, as assessed by MTT reduction in ischemic versus oxygenated conditions, are the ones who develop atrial fibrillation in response to the stress of a cardiac operation (Figure 6). The morphologic correlate of atrial low ischemic tolerance is atrial perinuclear vacuolation, referred to as myolysis (Ad et al. 2001). Thus, POAF actually reflects a preoperative occult atrial morbidity of low ischemic tolerance, which becomes obvious only when the patient is exposed to the operative stress. So, on one hand, low atrial ischemic tolerance may lead to susceptibility to atrial fibrillation, and on the other, atrial fibrillation exerts mitochondrial and membranal effects that decrease atrial ischemic tolerance. Atrial fibrillation in itself significantly increases atrial oxygen demand, further increasing the relative ischemia experienced by the tissue. This may be the basis of the positive feedback cycle underlying the “remodeling of atrial fibrillation,” reflected in the idiom “atrial fibrillation begets atrial fibrillation” (Wijffels et al. 1995). All in all, reduced ischemic tolerance has been shown to set the basis of POAF and may play a role in the pathogenesis of chronic idiopathic atrial fibrillation.

Effects of Drugs and Toxins on Cardiac Ischemic Tolerance

Since ischemic tolerance is modulated by major ubiquitous ligands and receptors, signal transduction pathways, transcription factors, intracellular enzymes, and end effectors (Figure 4), chronic exposure to many drugs and toxins potentially affect it. Given the high worldwide cardiac ischemic morbidity, and the importance of ischemic tolerance in determining the outcome of cardiac ischemic events, we argue that efforts should be taken to quantify the effect on ischemic tolerance of almost any commonly used drug or widespread environmental toxins. Some groups of drugs and toxins deserve special attention in this respect follow:

• substances that affect mitochondrial function, and specifically ones that inhibit mitochondrial ATP-dependent K+ channels (mito-K-ATP);

• substances targeted to specifically inhibit enzymes along the regulation pathways of ischemic tolerance, such as cyclooxygenase-2 (COX-2), inducible nitric oxide synthase (iNOS), or PKC-ɛ;

• substances used in the treatment of risk factors for ischemic heart disease, such as hypercholesterolemia, diabetes mellitus, hypertension, or obesity; and;

• substances that are clearly epidemiologically associated with increased risk of cardiac ischemic morbidity and mortality, without a plausible mechanistic explanation for this association.

Substances Affecting Mitochondrial Function

Mitochondria constitute a primary target of many cardiotoxic drugs and toxins (Kang 2001). These include the most well-studied cardiotoxic effect of adriamycin (Berthiaume and Wallace 2006) as well as the classic environmental cardiotoxins that exert their effects through the metabolite thiodiglycolic acid, such as CEM (Dunnick et al. 2004), the drug ifosfamide (Visarius et al. 1998), mono-chloroacetic acid (Dunnick et al. 2004), chloroacetaldehyde (Joqueviel et al. 1997), trichloroethane (Yllner 1971), trichloroethylene (Anderson et al. 1987), 1,1-dichloroethylene (Anderson et al. 1987), cyclophosphamide (Joqueviel, Malet-Marino, and Martino 1997), vinylidene chloride (Jones and Hathway 1978), and vinyl chloride (Green and Hathway 1975). The mitochondrial dysfunction and cardiotoxicity exerted by these compounds are mostly dose dependent. Low, subtoxic doses of these compounds may lead to subtle mitochondrial dysfunction, expressed by low ischemic tolerance, with no effect on cellular viability in oxygenated conditions. This was exemplified in the case of the prototypic thiodiglycolic acid dependent cardiotoxin, CEM (Golomb et al. 2007).

The time dependence of the cardiac effect of these toxins has been associated with the recruitability of myocardial protective mechanisms. Dunnick et al. (2006) have shown that CEM exerts cardiotoxic effect during the first days of its application, but these effects are resolved after 16 days of application of the drugs. The authors suggested that the continuous exposure to the toxin lead to recruitment of cardioprotective mechanisms, such as induction of superoxide dismutase, and that these mechanisms diminish the chronic cardiotoxic effect.

Mito-K-ATP, which is a major end effector of cardioprotective mechanisms, constitutes the target of various drugs. Substances that open mito-K-ATP include pinacidil, levcroma-kalim, nicorandil, and cromakalim and might serve as enhancers of ischemic tolerance. Sulfonylureas and glinides inhibit the opening of K-ATP channels. K-ATP inhibitors serve as antidiabetic drugs, as they increase insulin secretion by closing the K-ATP channel in pancreatic β-cell membrane (Quast et al. 2004). These compounds have been shown to decrease ischemic tolerance and abolish classic and late preconditioning in patients, suggesting an occult toxic effect (Ferreira et al. 2005; Loubani et al. 2005; Bilinska et al. 2007). Indeed, epidemiological data indicate that patients treated with sulfonylureas were at higher risk of adverse cardiovascular outcomes than those treated with metformin (Evans et al. 2006). Therefore, K-ATP inhibitors that show pancreatic specificity are being sought, to decrease the potential occult cardiotoxic effect of these compounds. Glymepiride and glyclazide have been shown to be more beneficial than glybenclamide in this respect (Lee and Chou 2003; Loubani et al. 2005).

Inhibitors of Instrumental Enzymes for Ischemic Tolerance: Cyclooxygenase 2 (COX-2) Inhibitors

Since the discovery of the COX-2 molecule as an inducible cyclooxygenase, two lines of research were pursued to exploit its potential clinical value. On one hand, specific COX-2 inhibitors were designed and ignited a lot of enthusiasm as nonsteroidal anti-inflammatory agents that would preserve gastric mucosa. Indications for such inhibitors included not only treatment of chronic inflammatory disorders such as osteoarthritis, acute pain, rheumatoid arthritis (Matheson and Figgitt 2001), and dysmenorrhea (Harel 2004) but also treatment and prevention of cancer (Umar et al. 2003; Arber 2008) and a potential effect in psychiatric disorders (Müller, Riedel, and Schwarz 2004). On the other hand, cardiac researchers headed by Bolli (2007) have shown that cardiac COX-2 is cardioprotective and an obligatory mediator of late ischemic preconditioning, so that COX-2 inhibition by either NS-398 or celecoxib abolishes preconditioning-induced cardioprotection (Shinmura et al. 2000). Myocardial transduction of COX-2 expression (“gene transfer”) protected the heart against myocardial infarction in mice (Bolli 2007).

Despite the fact that the role of cardiac COX-2 in ischemic tolerance was already known, the finding of increased risk of myocardial infarct and cardiac mortality due to reofecoxib came as a surprise to the medical community. Still, most of the search for the mechanism of the adverse cardiovascular effect of coxibs was sought in prothrombotic effects of COX-2 inhibitors. However, it is emphasized that cardiac ischemic events should be considered not only as a thrombotic disorder but also as a failure of cardioprotective mechanisms conferring ischemic tolerance. It is therefore likely that a fundamental reason for the adverse cardiovascular effects of COX-2 inhibitors is that “they deprive the heart of its ability to shift to a preconditioned, protected phenotype in response to stress” (Bolli 2007). Thus, it could be argued that the mere definition of deprivation of the heart from ischemic tolerance as “occult toxicity” could have prevented or decreased the fiasco of the coxibs and the surprise from their adverse cardiovascular effects.

Another obligatory mediator of cardioprotection by ischemic preconditioning, upstream of COX-2, is iNOS. It should be noted that iNOS is involved in the pathogenesis of various disorders, so that iNOS inhibitors, such as SC-51 and aminoguanidine, are being sought and examined, not only for the treatment of septic shock, but also in asthma (Hansel et al. 2003). The potential effect of such compounds on cardiac ischemic tolerance, and the finding that aminoguanidine inhibits ischemic (Imagawa, Yellon, and Baxter 1999) and pharmacologic (Hattori et al. 2002) preconditioning, should be taken into consideration.

Substances Used in the Treatment of Risk Factors for Ischemic Heart Disease

Potential occult cardiotoxicity by substances used for treatment of cardiac risk factors such as diabetes mellitus, hypercholesterolemia, uremia, or hypertension requires special attention and awareness. These conditions are associated with both increased risk of thromboembolic events and, in some cases, with an a priori decrease in cardiac ischemic tolerance (Ferdinandy, Schulz, and Baxter 2007), so that cardiac viability may depend on remnant blood flow and remnant recruitability of protective mechanisms. Furthermore, in these conditions the heart is likely to be exposed to recurrent small and short ischemic events, leading to recruitment of cardioprotective mechanisms. Therefore, a decrease in basic cardiac ischemic tolerance or in the recruitability of adaptive mechanisms due to an adverse drug effect may be particularly harmful in patients with such risk factors.

The awareness to the potential occult cardiotoxic effect of oral antidiabetic drugs, such as sulfonylureas and glinidines, has led to the development and use of compounds that are more pancreatic specific, such as glymepiride and glyclazide (discussed above) (Lee and Chou 2003; Loubani et al. 2005). There is less awareness to this potential problem in the treatment of other cardiac risk factors.

The effect of statins on cardiac ischemic tolerance has been addressed in several studies. In most human observational (Dotani et al. 2000; Chan et al. 2002; Lindenauer et al. 2004) and randomized (Pasceri et al. 2004) studies, statins were found to protect the heart during cardiac procedures. Animal studies also showed a direct cardioprotective effect (Wayman, Ellis, and Thiemermann 2003). The cardioprotective effect of statins is mostly independent of their lipid-lowering properties and involves opening of mito-K-ATP channels and the induction of nitric oxide synthase and COX-2 (Tavackoli et al. 2004; Atar et al. 2006). However, the beneficial effect of statins is dose-dependent, in a manner that probably depends on the specific statin. In an ex vivo study, simvastatin, at a concentration of 25 μM, protected the isolated heart from ischemia/reperfusion induced contractile dysfunction, release of creatine kinase, and postischemic hyperpermeability. This effect almost disappeared at a concentration of 50 μM, and at 100 μM, simvastatin exacerbated the injury (Di Napoli et al. 2001). In another study, lovastatin at 50 μM or 15 mg/Kg/d was found to inhibit the cardioprotective mechanisms of ischemic preconditioning and postconditioning (Kocsis et al. 2008). Taken together, these data point to a dose-dependent effect of statins on cardiac ischemic tolerance and recruitability of cardioprotective mechanisms. At lower doses, statins confer cardioprotection both by correcting hyperlipidemia and by a direct effect on pathways and end effectors mediating preconditioning. At high doses, statins may have an occult cardiotoxic effect. To date, the main consideration in determining the dosage of statins administered to patients is the effect on blood cholesterol and on adverse effects, such as myopathy, rhabdomyolysis, or liver function disturbances. We argue that data should be collected to enable considering the effect of the drug dosage on cardiac ischemic tolerance as well and avoiding doses that might have an occult cardiotoxic effect.

Data about the effect of antihypertensive drugs on ischemic tolerance are scarce: studies are needed before conclusions can be offered concerning the potential effect of antihypertensive drugs, other than angiotensin converting enzyme (ACE) inhibitors, on ischemic tolerance. ACE inhibitors were found in various studies to improve ischemic tolerance, mainly through effects on the kinin system (Jaberansari et al. 2001; Lange et al. 2007). On the other hand, AT1 receptor blockade abolished the infarct size-limiting effect of ischemic preconditioning (Diaz and Wilson 1997), indicating that angiotensin II stimulation contributes to ischemic preconditioning, probably by activating protein kinase C (Y. Liu et al. 1995). Neutral endopeptidase inhibition by thiorphan does not affect basic cardiac ischemic tolerance, as determined by infarct size reduction, but potentiates preconditioning-induced cardioprotection via a bradykinin (B2) receptor-mediated mechanism (Nakano et al. 2002).

Substances Associated with Risk of Cardiac Ischemic Morbidity and Mortality

Exposure to drugs, toxins, and physical and chemical environmental agents has been associated with an increased risk of cardiac ischemic morbidity and mortality, although classic toxicologic studies have not shown a direct cardiotoxic effect. Examples range from exposure to particulate air pollution (Zanobetti and Schwartz 2007), through long-term effects of radiotherapy and chemotherapy administered for the treatment of Hodgkin’s lymphoma (Swerdlow et al. 2007), to exposure to lead (Jain et al. 2007), among many others. It is intriguing to postulate that in some of these cases, the exposure to the exogenous compound impairs cardiac ischemic tolerance through mitochondrial damage or through damage to the signal transduction pathways modulating ischemic tolerance, rendering the heart more susceptible to ischemic insults. Revealing the mechanism for this increased risk requires morphological, biochemical, and molecular techniques that directly assess mitochondrial function and reserves, such as electron microscopic analyses, MTT assays, and gene expression arrays. Recently, some such adverse effects of compounds have been clearly associated with a mitochondrial damage, referred to as mitochondrial cardiomyopathy. For example, morphometric and semiquantitative analysis of CD-1 mouse pup myocardial cells, following in utero and postnatal exposure to the antiretroviral drugs zidovudine and lamivudine, disclosed significant increases in the mean area and decreases in the mean number of cardiomyocytic mitochondria compared to controls. In addition, clusters of damaged mitochondria were more frequently seen by EM in treated animals than in controls (Figure 5) (Bishop et al. 2004). No changes were noted at light microscopic histopathology.

Another example is the exposure to environmental inhaled toxin present in air pollution. Studies with zinc, a common metal present in most ambient particulate matter, indicated that rats exposed by inhalation to 10, 30 or 100 μg/m³ of aerosolized zinc sulfate (ZnSO₄), 5 hours/day, 3 days/week for 16 weeks, did not show any histopathologic cardiac changes (Wallenborn et al. 2008). However, in the heart, cytosolic glutathione peroxidase activity decreased, while mitochondrial ferritin levels increased and succinate dehydrogenase activity decreased, suggesting a mitochondria-specific effect. Also, cardiac gene array analysis indicated small changes in genes involved in cell signaling, a pattern concordant with known zinc effects. Such occult mitochondrial damage is likely to account for high susceptibility to ischemic events.

To summarize, various genetic and environmental factors, including age and metabolic disease states, affect both cardiac ischemic tolerance, the ability to recruit protective mechanisms such as preconditioning and postconditioning, and cardiac responsiveness to drugs that confer cardioprotection. In conditions of low ischemic tolerance and low responsiveness to preconditioning stimuli, such as prolonged diabetes mellitus, uremia, and possibly hypercholesterolemia, it should be of particular importance to avoid exposure to potential occult cardiotoxic agents, which would render the myocardium even more prone to ischemia. Individuals with a low baseline ischemic tolerance are likely to suffer from adverse effects of drugs (or adverse environmental stimuli) that may further reduce their ischemic tolerance as a side effect. It is therefore of specific importance to recognize the effect of each drug on cardiac ischemic tolerance not only in young healthy individuals, but also in the disease states from which the patient suffers. Such information is still not available for most drugs. It should be emphasized that avoiding a side effect of occult cardiotoxicity by drugs should be a target in the design of the treatment given to each patient. To achieve this target, information should be available concerning the effect of drugs on ischemic tolerance, both in general and in the disease states for which the drugs are indicated.

Perspective: Means to Cope With the Occult Cardiotoxicity Problem

The understanding that ischemic tolerance plays a key role in the outcome of ischemic events, and that exposure to drugs and toxins may impair ischemic tolerance, raise the possibility that occult cardiotoxicity is a major environmental problem. Since cardiac ischemic tolerance is modulated by central and ubiquitous receptors, signal tranduction pathways, transcription factors, enzymes, and end effectors, almost any drug or toxin may affect it. Many drugs and toxins have been found to affect the level and activity of protein kinase C, nuclear factor κB, STAT 1 and 3, and the activity of enzymes that modulate the level of opening of ATP dependent potassium channels. Probably, there is hardly any drug or toxin that would not affect any of these crucial cellular components at any dose or duration. Thus, chronic administration of most drugs and toxins may affect cardiac ischemic tolerance.

To assess the magnitude of the problem of occult cardiotoxicity and cope with it certain steps should be taken.

First and most important, the problem should be recognized and taken seriously: a decrease in ischemic tolerance (or in the recruitability of cardioprotective mechanisms) by exposure to a substance should be regarded as a form of toxicity (we propose the use of the term “occult cardiotoxicity”) and not be tolerated as an acceptable side effect. Second, the effect of drugs and toxins on cardiac ischemic tolerance should be systematically assessed, so that quantitative information on the effect of substances on ischemic tolerance becomes available. In other words, assessing the effect on ischemic tolerance should be a part of the complete assessment of cardiotoxicity.

To achieve this, simple methods should be adopted for the assessment of ischemic tolerance and to enable large-scale studies to identify potential occult cardiotoxic substances. Such methods can be based on the comparison of cellular viability (or cellular damage) between oxygenation and simulated ischemia/ reoxygenation. Identification of molecular or biochemical correlates of low ischemic tolerance may considerably facilitate the identification of occult cardiotoxins. Once a potential occult cardiotoxin is identified, its effect on infarct size, arrhythmias, and cardiac function should be specifically assessed.

We must bear in mind that ischemic heart disease is often associated with other disease states such as hyperlipidemia, diabetes mellitus, or hypertension. Therefore, examination of the interactions of drugs/toxins with common cardiac and extracardiac disorders affecting ischemic tolerance should be considered as a more thorough examination of occult cardiotoxicity. Taking measures to study in advance the effect of exogenous substances, particularly drugs, on ischemic tolerance, together with recognition of the cellular mechanisms that govern ischemic tolerance, may help avoiding the exposure to occult cardiotoxins. Such measures may help prevent a second unpleasant surprise like the rofecoxib withdrawal.