Nitric oxide, superoxide, and peroxynitrite in myocardial ischaemia-reperfusion injury and preconditioning

Abstract

There appears to be a controversy in the study of myocardial ischaemia-reperfusion injury and preconditioning whether nitric oxide (NO) plays a protective or detrimental role. A number of findings and the interpretation of the results to date do not support such a controversy. An understanding of the latest developments in NO, superoxide (O₂⁻·) and peroxynitrite (ONOO⁻) biology, as well as the various ischaemic animal models utilized is necessary to resolve the apparent controversy. NO is an important cardioprotective molecule via its vasodilator, antioxidant, antiplatelet, and antineutrophil actions and it is essential for normal heart function. However, NO is detrimental if it combines with O₂⁻· to form ONOO⁻ which rapidly decomposes to highly reactive oxidant species. There is a critical balance between cellular concentrations of NO, O₂⁻·, and superoxide dismutase which physiologically favour NO production but in pathological conditions such as ischaemia and reperfusion result in ONOO⁻ formation. In contrast, exposure of the heart to brief episode(s) of ischaemia markedly enhances its ability to withstand a subsequent ischaemic injury. The triggering of this endogenous cardioprotective mechanism known as preconditioning requires both NO and O₂⁻· synthesis. However, preconditioning in turn attenuates the overproduction of NO, O₂⁻· and ONOO⁻ during a subsequent episode of ischaemia and reperfusion, thereby protecting the heart. Here we review the roles of NO, O₂⁻·, and ONOO⁻ in both ischaemia-reperfusion injury and preconditioning.

Myocardial ischaemia-reperfusion andpreconditioning

Ischaemic heart disease, a major cause of mortality in industrialized societies, is characterized by insufficient blood supply to regions of the myocardium which leads to tissue necrosis (infarction). Ischaemic heart disease develops as a consequence of many diseases, including hypertension, atherosclerosis, hyperlipidaemia, and diabetes. The treatment of this condition has entered a new era where mortality can be approximately halved by procedures which allow the rapid return of blood flow, i.e. reperfusion, to the ischaemic zone of the myocardium. Reperfusion, however, may lead to further complications such as diminished cardiac contractile function and arrhythmias (see for review: Braunwald, 1985). Therefore, development of cardioprotective agents which improve myocardial function, decrease the incidence of arrhythmias, lessen the necrotic tissue mass, and delay the onset of necrosis during ischaemia/reperfusion is of great clinical importance. Previous attempts to attenuate the consequences of ischaemia/reperfusion injury with pharmacological tools have been largely unsuccessful. However, the heart was found to be able to adapt to ischaemic stress (Murry et al., 1986). Ischaemic preconditioning is a well-described adaptive response in which brief exposure to ischaemia markedly enhances the ability of the heart to withstand a subsequent ischaemic injury (see for review: Przyklenk & Kloner, 1998). The discovery of this endogenous cardioprotective mechanism encouraged the exploration of new ways to protect the ischaemic/reperfused myocardium. The exact cellular mechanism of preconditioning is still a question of debate, however, among several other mediators, NO, superoxide (O₂⁻·), and peroxynitrite (ONOO⁻) certainly play a role in both ischaemia/reperfusion injury and in the development of the cardioprotective effect of preconditioning (see for reviews: Baxter & Ferdinandy, 2001; Bolli, 2001; Bolli et al., 1998; Schulz et al., 2001).

Nitric oxide: cardioprotective or detrimental?

There appears to be a controversy in the study of myocardial ischaemia-reperfusion injury and preconditioning whether NO plays a protective or detrimental role. Findings from our laboratories and others, and interpretation of the results to date, do not support such a controversy. An understanding of the latest development in NO, O₂⁻· and ONOO⁻ biology as well as the various ischaemic animal models utilized is necessary to resolve the apparent controversy.

Coronary endothelium, endocardial endothelium, cardiac nerves, and cardiomyocytes of the normal heart are all sources of the basal production of NO by Ca²⁺-dependent NO synthases (NOS) (Curtis & Pabla, 1997; Pabla & Curtis, 1996). This serves a number of important physiological roles in the regulation of cardiac function including coronary vasodilation, inhibiting platelet and neutrophil actions, antioxidant effects, modulation of cardiac contractile function, and inhibiting cardiac oxygen consumption (Hare & Comerford, 1995; Xie & Wolin, 1996).

NO plays protective roles in the ischaemic heart by several mechanisms (Figure 1) such as stimulating soluble guanylate cyclase and thus reducing [Ca²⁺]_i partly through activation of cGMP-dependent protein kinase, terminating chain propagating lipid radical reactions caused by oxidative stress (Rubbo et al., 1994), and by inhibiting the activity of platelets and neutrophils and their adhesion to the endothelial surface (Kubes et al., 1991; Radomski et al., 1987). NO (either via NOS acitivity or through NO donors) protects against the toxic effects of exogenously supplied ONOO⁻ in the coronary circulation (Villa et al., 1994), on platelets (Moro et al., 1994), or in liposomes (Rubbo et al., 1994). Whether NO-induced inhibition of mithochondrial respiration and modulation of apoptosis is protective or detrimental in ischaemia/reperfusion is not clear in the literature. NO through its interaction with components of the mitochondrial respiratory chain might function not only as a physiological regulator of cell respiration, but also as a modulator of the generation of reactive oxygen species by mitochondria, thereby affecting mechanisms of cell survival or death (Moncada & Erusalimsky, 2002). It should be noted that the exact role of NO in apoptosis is not known as NO has been shown to exert both proapoptotic and antiapoptotic effects in the myocardium (Gao et al., 2002a; Iwai-Kanai et al., 2002; Sam et al., 2001; Taimor et al., 2000; Weiland et al., 2000). The controversy is possibly due to the lack of ONOO⁻ measurements in these studies.

Figure 1.

Cellular mechanisms of nitric oxide (NO), superoxide (O₂⁻·), and peroxynitrite (ONOO⁻) actions. NO is an important cardioprotective molecule via its vasodilator, antioxidant, antiplatelet, and antineutrophil actions and it is essential for normal heart function. However, NO is detrimental if it combines with O₂⁻· to form ONOO⁻ which rapidly decomposes to highly reactive oxidant species leading to tissue injury. There is a critical balance between cellular concentrations of NO, O₂⁻·, and superoxide dismutase (SOD) which physiologically favor NO production but in pathological conditions such as ischaemia and reperfusion result in ONOO⁻ formation. ONOO⁻ is detoxified if it combines with reduced glutathione (GSH) or other thiols to form S-nitrosoglutathione (GSNO) or other nitrosothials, an NO donor molecule. MMP – matrix metalloproteinase; NOS – NO synthase; PARP – poly-ADP ribose synthase; XOR – xanthine oxidoreductase.

Toxicity of NO via reaction with O₂⁻· toform ONOO⁻

NO is necessary for normal cardiac physiology, but it is potentially toxic in excess concentrations. It is now understood that many of the toxic actions of NO are not directly due to NO itself but are mediated via production of ONOO⁻, the reaction product of NO with O₂⁻· (Figure 1; Beckman & Koppenol, 1996; Rubbo et al., 1996). Although formation of ONOO⁻ has been shown to contribute to several cardiovascular pathologies including ischaemia/reperfusion injury, the possible physiological role of ONOO⁻ is not known.

Possible sources of O₂⁻· in the heart include NAD(P)H oxidoreductases (Mohazzab et al., 1997), xanthine oxidoreductases (Hille & Nishino, 1995), mitochondrial electron transport chain activity, activated neutrophils, arachidonic acid metabolism, and auto-oxidation of certain tissue metabolites. The highly oxidative heart muscle is subjected to its own basal production of O₂⁻· (Boveris & Chance, 1973) and NO (Hare & Comerford, 1995). Mitochondrial Mn-superoxide dismutase (SOD), cytosolic Cu-Zn SOD, extracellular Cu-Zn SOD, glutathione (GSH), uric acid, and catalase are the more important endogenous antioxidants.

NO and O₂⁻· combine at a reaction rate which is only limited by diffusion to form ONOO⁻. ONOO⁻ at pH<8 is protonated to form the unstable intermediate, peroxynitrous acid, which spontaneously decomposes to yield highly reactive oxidant species (Figure 1). Understanding the balance between local concentrations of NO, O₂⁻·, and SOD is critical in understanding NO biology and its potential toxicity in the form of ONOO⁻ (Beckman & Koppenol, 1996; Rubbo et al., 1996). One must consider the competition between NO and SOD for O₂⁻·. Under normal physiological conditions in vascular endothelium, [NO] is ∼10 nM, [SOD] is ∼1 μM. Therefore, the reaction rate of O₂⁻· to form ONOO⁻ is only a fraction of the dismutation rate of O₂⁻· by SOD, and, as a result, very little ONOO⁻ is formed. However, at maximal vascular rates of NO production (i.e. which may occur during acute reperfusion of ischaemic tissue or during inducible NOS expression), [NO] is⩾1 μM. Therefore, the formation of ONOO⁻ will predominate over the dismutation of O₂⁻·. Under a variety of cellular stress conditions where NO production is upregulated in the endothelium and cardiomyocytes, either by acute increase in Ca²⁺-dependent NOS activity via changes in shear stress or high [Ca²⁺]_i which occurs during acute reperfusion and ischaemia, or via de novo expression of inducible NOS in cells or tissues as a result of the action of pro-inflammatory cytokines, one predicts the formation of ONOO⁻ (Csonka et al., 1999; Depre & Hue, 1994; Ferdinandy et al., 2000).

Pathophysiology of ONOO⁻ in myocardialischaemia-reperfusion

Acute injury

We have shown first that ONOO⁻ is produced during the acute reperfusion of ischaemic hearts and that drugs which inhibit ONOO⁻ formation or antagonize its toxicity protect the heart from reperfusion injury (Yasmin et al., 1997; Cheung et al., 2000b). We have also developed a novel, simple and effective methodology to detect cardiac ONOO⁻ based on the reaction of ONOO⁻ with tyrosine to form dityrosine (Ferdinandy & Schulz, 2001a; Yasmin et al., 1997). Rapid generation of ONOO⁻ during reperfusion of the ischaemic heart has also been detected using luminol chemiluminescence in the perfusate and anti-nitrotyrosine labelling of myocardial proteins (Wang & Zweier, 1996). We showed that low concentrations of the NOS inhibitor N^G-monomethyl-L-arginine (L-NMMA), or a cell permeable SOD mimetic, MnTBAP protected the hearts from ischaemia-reperfusion injury. The beneficial effect of L-NMMA fell within a narrow range of concentrations and was lost at higher concentrations which further reduced coronary flow (Yasmin et al., 1997). Our data also showed that a NO donor, at subvasodilatory concentration, protected hearts from endogenous ONOO⁻-mediated injury. This study provided the first mechanistic evidence of how either NO donors by detoxifying ONOO⁻ or NOS inhibitors by inhibiting ONOO⁻ formation are able to reduce ischaemia-reperfusion injury (Yasmin et al., 1997).

Myocardial infarction (late injury, neutrophil dependent)

What is the importance of the neutrophil-independent free radical release during acute reperfusion, as opposed to the neutrophil-dependent reperfusion events which occur several hours later? It should be noted that the late influx of neutrophils does not substantially contribute to the development of infarction, however, the literature is somewhat controversial in this regard (Williams et al., 1994; Baxter, 2002; Gao et al., 2002b; Jordan et al., 1999). One very important defect in the heart resulting from acute reperfusion following ischaemia is endothelial injury, which is manifested by reduced endothelium-dependent vasodilator response within the first minute of reperfusion (Siegfried et al., 1992). We speculate that this damage is self-inflicted, resulting from the burst of endogenous ONOO⁻ from the vascular endothelium immediately at reperfusion. This would cause enhanced susceptibility of the endothelial surface to neutrophil and platelet adhesion, platelet aggregation and neutrophil activation, events which are normally inhibited by the physiological production of endothelium-derived NO (Kubes et al., 1991; Radomski et al., 1987). Villa et al. (1994) showed that bolus injection of ONOO⁻ into isolated hearts acutely inhibited endothelium-dependent coronary vasodilation. Myocardial ONOO⁻ generation has also been shown to occur during the neutrophil-dependent phase of ischaemia-reperfusion injury, seen 5 h after reperfusion, using an in vivo model of regional ischaemia in rats (Liu et al., 1997). An understanding of the oxidative damage which occurs during acute reperfusion following ischaemia is thus crucial to help devise strategies to reduce the possible subsequent neutrophil-mediated damage in the later stages of reperfusion injury.

Cellular targets of ONOO⁻

The possible downstream targets of ONOO⁻ which mediate its toxicity are several (Figure 1). Its highly reactive decomposition products at physiological or acidic pH can attack protein (oxidation of sulfhydryls, nitration of tyrosine residues), lipids (formation of lipid peroxides), and DNA (strand breakage). This results in the depletion of low molecular weight antioxidants such as glutathione and most often the inhibition of several enzyme activities including superoxide dismutase, aconitase and other enzymes of the mitochondrial respiratory chain, creatine kinase, Ca²⁺-ATPase, Na⁺-K⁺-ATPase, glutathione peroxidase, prostacyclin synthase, α-antiproteinase and many others, just to name a few (see for reviews: Beckman & Koppenol, 1996; Ronson et al., 1999; Rubbo et al., 1996; Szabó, 1996). As a result of DNA strand breakage activation of the NAD⁺ consuming DNA repair enzyme poly-ADP ribose synthase contributes further to the depletion of cellular energy stores (Szabó, 1996). ONOO⁻ has been also shown to cause irreversible inhibition of the mitochondrial respiratory chain (Brown, 2001; Xie & Wolin, 1996) and to trigger apoptosis of cardiac myocytes (Arstall et al., 1999; Beckman, 1999). An important question is which of these multiple effects represents an early event in ONOO⁻-mediated toxicity, at a time point prior to the onset of irreversible energy depletion, cellular damage and death.

Interestingly, ONOO⁻ may also activate some proteins, again due to its ability to react with critical thiol residues. At low micromolar concentrations Maeda's group have shown that ONOO⁻ activates the zymogen form of matrix metalloproteinases (MMPs) (Okamoto et al., 1997; 2001). This family of proteolytic enzymes, best known for their ability to degrade and remodel the extracellular matrix, are now recognized to have a variety of novel actions as proteases, both physiological and pathological, on substrates apart from the extracellular matrix. As we showed that ONOO⁻ biosynthesis peaks in isolated rat hearts in the first minute of reperfusion following ischaemia and contributes to contractile dysfunction (Yasmin et al., 1997) we addressed whether MMP activation could play a role in the detrimental actions of ONOO⁻ in the heart. We found that MMP-2 is activated and released from the reperfused heart, peaking in the first minute of reperfusion, and that inhibition of MMP-2 activity functionally protected the heart (Cheung et al., 2000a). Moreover, infusion of ONOO⁻ into isolated rat hearts resulted in MMP-2 activation which preceded the onset of mechanical dysfunction, an effect which could be abolished either with glutathione or an inhibitor of MMP activity (Wang et al., 2002a). MMP-2 was found to colocalize with the contractile protein regulatory element troponin I in cardiac myocytes and upon ischaemia-reperfusion injury was found to be responsible for the proteolytic cleavage of troponin I (Wang et al., 2002b), one of the contributory factors to altered contractility in myocardial ischaemia/reperfusion injury (Bolli & Marban, 1999). It is thus apparent that an important target of the early phase of ONOO⁻-induced oxidative stress injury in the heart is activation of MMP-2 and its resultant proteolysis of novel intracellular targets, including troponin I.

Can ONOO⁻ protect the heart againstmyocardial ischaemia-reperfusion?

The literature is somewhat controversial as to whether ONOO⁻ is either cytotoxic or cytoprotective (Ferdinandy & Schulz, 2001b; Vinten-Johansen, 2000). It seems that deleterious effects of ONOO⁻ in the heart are predominantly observed in in vitro or ex vivo crystalloid buffer perfused systems, however, ONOO⁻ is cardioprotective if applied in vivo. However, this controversy derives solely from the rapid reactions of exogenously administered ONOO⁻ with a variety of biomolecules. Simply, exogenous administration of ONOO⁻ does not accurately reflect pathological conditions in which intracellular generation of ONOO⁻ is enhanced.

Due to the very short half life of ONOO⁻ at physiological pH, it has very little chance to reach its cellular targets when it is applied via the blood, as it rapidly reacts with plasma proteins and thiols such as GSH. Thus, ONOO⁻ is likely to be detoxified before it has a chance to reach tissues downstream of the injection site, let alone the intracellular compartment (Ishida et al., 1999). ONOO⁻ oxidizes thiols (Cheung et al., 1998), e.g. it reacts with GSH to form the nitrosothiol, nitrosoglutathione, a NO donor (Figure 1; Mayer et al., 1995; Prendergast et al., 1997; Villa et al., 1994). By this reaction ONOO⁻ is a vasodilator (Villa et al., 1994; Wu et al., 1994) and inhibitor of platelet aggregation (Moro et al., 1994). Thus nature has a built-in mechanism to transform toxic ONOO⁻ into a NO donor. Indeed, hearts with enhanced endogenous GSH levels are less susceptible to ischaemia-reperfusion injury (Kirshenbaum & Singal, 1993) and micromolar concentrations of GSH added to the perfusate protects isolated hearts from post-ischaemic contractile dysfunction through the reduced formation of ONOO⁻ at reperfusion (Cheung et al., 2000b). Exogenously administered ONOO⁻ was also shown to inhibit leukocyte-endothelial cell interactions and to protect against ischaemia-reperfusion injury in rats in vivo (Lefer et al., 1997). Intraventricular infusion of ONOO⁻reduced myocardial infarct size and preserved coronary endothelium after ischaemia and reperfusion in cats (Nossuli et al., 1997), which effects were mediated by the intermediate formation of S-nitrosothiols (Nossuli et al., 1998).

Exogenously applied ONOO⁻, however, has been shown to be detrimental to cellular functions when it was applied in crystalloid buffer systems, in which the concentrations of extracellular antioxidants and both free and protein bound thiols are limited. In this case, exogenous ONOO⁻ and its toxic metabolites have a greater chance to reach their cellular targets and cause injury, this of course being dependent upon concentration and the duration of exposure. We have shown that continuous infusion of 40 but not 4 μM ONOO⁻ into isolated working rat hearts perfused with Krebs-Henseleit buffer impaired cardiac contractile function within 45 min (Schulz et al., 1997). Authentic ONOO⁻ inhibited contractile function in cardiac myocytes (Ishida et al., 1996) and in isolated rat papillary muscle (Digerness et al., 1999). Administration of the ONOO⁻ generator SIN-1 exerted either cardiodepression in crystalloid buffer-perfused or cardioprotection in blood-perfused rat hearts (Ma et al., 2000).

One can conclude that exogenous ONOO⁻ is toxic when applied in crystalloid perfused hearts, however, it can be protective under experimental conditions in which ONOO⁻ first reacts with thiol groups, thereby forming NO donors. It is questionable whether this mechanism of ONOO⁻ detoxification is always of the capacity to eliminate a sufficient portion of ONOO⁻ at the site of its endogenous formation, especially under conditions of ischaemia/reperfusion. To our knowledge, there is no literature showing any tissue protective effect of endogenously formed ONOO⁻. In contrast, many studies show that enhanced formation of ONOO⁻ in the myocardium is cytotoxic to the heart and contributes to ischaemia/reperfusion injury in isolated rat hearts (Wang & Zweier, 1996; Yasmin et al., 1997) and anaesthetized rats (Liu et al., 1997), and in other pathological models of myocardial dysfunction in several species in vivo or ex vivo (Ferdinandy et al., 1999; 2000; Ishiyama et al., 1997; Oyama et al., 1998; Sakurai et al., 1999; Weinstein et al., 2000) including myocardial inflammation in humans (Kooy et al., 1997). Many of these studies showed a correlation between endogenous ONOO⁻ formation and deterioration of cardiac function.

Taken together, there is a consensus in the literature that endogenously formed ONOO⁻ contributes to the injury to the heart as a result of ischaemia/reperfusion and other cardiac pathologies.

NO is cardioprotective in ischaemia-reperfusion injury

A number of ex vivo studies using isolated crystalloid-buffer perfused hearts have shown that enhancing NO levels, either by applying NO donors, NO-dependent vasodilators, L-arginine supplementation, angiotensin converting enzyme inhibitors, or pretreating the animal with endotoxin analogues which among several other effects enhance the activities of myocardial inducible NOS, cyclo-oxygenase, and antioxidant enzymes, functionally protect the heart from acute ischaemia/reperfusion injury and/or infarct size development (Masini et al., 1991; Massoudy et al., 1995; Schoelkens & Linz, 1992; Xi et al., 1999; Yasmin et al., 1997). Whether this protective effect is due to the antioxidant properties of NO, lowering [Ca²⁺]_i via activation of guanylate cyclase, modulating mitochondrial respiration, influencing apoptosis, or other mechanisms is not completely understood at the moment.

In vivo studies show that NO donors improve the recovery of mechanical function and/or reduce infarct size following ischaemia-reperfusion. Subvasodilatory doses of NO donors (Siegfried et al., 1992), L-arginine (Weyrich et al., 1992) or agonists of endothelium-derived NO (Richard et al., 1995) were shown to be protective (see for review: Grisham et al., 1998). In vivo studies using endothelial NOS knockout mice enhance the notion that the basal release of NO in the heart is an important endogeneous cardiac protectant, particularly in regards to the prevention of neutrophil sticking and platelet activation. Infarct size in endothelial NOS knockout mice was larger than in the wildtype controls, with higher P-selectin expression and significantly more neutrophils in hearts from endothelial NOS knockout mice (Jones et al., 1999a). Yang et al. (1999) showed that the protective effect of an angiotensin converting enzyme inhibitor in ischaemia-reperfusion injury in wildtype mice was lost in the corresponding endothelial NOS knockout mice. In contrast, an ex vivo study showed that the functional recovery of hearts from endothelial NOS knockout mice was improved in comparison to wildtype controls (Flogel et al., 1999), suggesting that NO and thus ONOO⁻ generated from endothelial NOS contributes to ischaemia/reperfusion injury seen in early reperfusion.

Inhibition of NOS: can it also be protective in ischaemia-reperfusion injury?

Ex vivo studies

Despite evidence that supplementation of NO can protect hearts from ischaemia-reperfusion injury, reports that inhibition of NOS can also prevent such injury began to appear. Most importantly, all studies using NOS inhibitors must be evaluated in regard to the concentration, potency and efficacy of the NOS inhibitor used, as well as to experimental details (i.e. constant flow versus constant pressure perfusion, global versus low-flow ischaemia, etc.). Low to moderate inhibitory concentrations of L-NAME (3 μM) or L-NMMA (30 μM) protected isolated working rabbit hearts from ischaemia/reperfusion injury when given prior to the onset of global, no-flow ischaemia (Schulz & Wambolt, 1995). We later showed that the mechanism of cardioprotection by L-NMMA in constant pressure-perfused rat hearts subjected to ischaemia-reperfusion was by reducing the formation of ONOO⁻ at reperfusion (Yasmin et al., 1997). Depre et al. (1995) showed that L-NMMA at concentrations between 0.001–10 μM significantly protected isolated rabbit hearts perfused at constant flow and subjected to low-flow ischaemia. They suggested that the protective effect of L-NMMA was by the enhancement of glycolysis during ischaemia, however, they could not exclude a protective effect during reperfusion itself. Another study using constant flow-perfused hearts suggested additional protective effects of L-NAME (30 μM), including the stimulation of adenosine release which occurs only at such high levels of NOS inhibition (Woolfson et al., 1995). A study in isolated rat hearts which used 1 mM L-NAME (Wang & Zweier, 1996), supramaximal in terms of blocking NO biosynthesis, may have included protective effects unrelated to NOS inhibition, as L-NAME in particular has been shown to be a muscarinic receptor antagonist at concentrations greater than 100 μM (Buxton et al., 1993).

In vivo studies

Essentially there is no study to date which carefully assesses the dose-dependent actions of NOS inhibitors in myocardial ischaemia-reperfusion injury in vivo. This is an obvious gap in the literature. Studies using high doses of NOS inhibitors (i.e.>IC₅₀), which significantly raise mean arterial blood pressure, uniformly provide evidence of a worsened outcome in terms of infarct size or function post ischaemia reperfusion (Hoshida et al., 1995; Williams et al., 1995). Interestingly, a study in open-chested rats subjected to only 4 min of regional ischaemia and 4 min of reperfusion showed that low doses of L-NAME caused a marked reduction in mortality due to ventricular fibrillation. This protective effect was abolished by co-administration of L-arginine and potentiated by co-administration of SOD, suggesting that reperfusion arrythmias may be due to the generation of ONOO⁻ (Ohoi & Takeo, 1996).

Myocardial adaptation to ischaemia:preconditioning

Preconditioning confers a remarkable cardioprotection in a variety of species including humans (see for review: Przyklenk & Kloner, 1998), although the cardioprotective effectiveness of ischaemic preconditioning might be attenuated in the heart during aging and some disease states such as hyperlipidaemia and diabetes (see for review: Ferdinandy et al., 1998). Preconditioning can be elicited by different sublethal stress signals, such as brief periods of ischaemia, hypoxia, rapid electrical pacing, heat stress, or administration of bacterial endotoxin, etc. The cardioprotective effect of preconditioning shows two distinct phases. The early phase is manifested within minutes after the preconditioning stimulus and has a duration of less than 3 h. The late phase is characterized by a slower onset (⩾20 h) and a duration of up to 72 h. Both phases of preconditioning involve reduction of necrotic tissue mass (infarct size), improvement of cardiac performance and reduction of arrhythmias following ischaemia and reperfusion (see for reviews: Baxter & Ferdinandy, 2001; Ferdinandy et al., 1998; Przyklenk & Kloner, 1998).

Among several other mediators, NO, oxygen free radicals, and antioxidant enzymes have been suggested to be, and also refuted as, key triggers and mediators of preconditioning. NO has been also suggested to play a role in the mechanism of other well known triggers of preconditioning, such as adenosine, bradykinin, and opioids (see for review: Bolli, 2001). There is still a considerable debate regarding the exact cellular mechanism of ischaemic preconditioning (see for review: Schulz et al., 2001). The mechanisms of early and late preconditioning seem to be different. Further discrepancies are generally attributed to species differences, different stimuli to induce preconditioning, and different study end-points, i.e. myocardial function, arrhythmias, or infarct size. Understanding the cellular pathways involved in the ischaemic adaptation of the myocardium may lead to the development of ‘preconditioning mimetic' drugs for patients suffering from ischaemic heart disease.

NO, O₂⁻·, and ONOO⁻ in earlypreconditioning

The role of NO in early preconditioning was suggested first by Vegh et al. (1992). They demonstrated that 10 mg kg⁻¹ L-NAME administered both before and after preconditioning abolished the antiarrhythmic effect of preconditioning in a coronary occlusion model of anaesthetized, open chest dogs (Vegh et al., 1992). Bilinska et al. (1996) reported that infusion of the NO donors nitroglycerin (500 μM) or SIN-1 (10 μM) for 5 min before coronary occlusion mimicked the antiarrhythmic effect of ischaemic preconditioning in isolated rat hearts. Using the electron spin resonance technique to directly measure cardiac NO content we have shown that a decrease in basal cardiac NO content before preconditioning, due to either in vivo pretreatment with 1 mg kg⁻¹ L-nitroarginine (Ferdinandy et al., 1996), experimental hypercholesterolaemia (Ferdinandy et al., 1997b), or selective depletion of neurotransmitters including NO from cardiac sensory neurons (Ferdinandy et al., 1997a) leads to the loss of pacing-induced preconditioning in isolated working rat hearts. However, preconditioning, in turn, markedly decreased the accumulation of NO in the heart tissue during subsequent ischaemia and reperfusion in isolated working rat hearts (Csonka et al., 1999). In the presence of 4.6 μM L-nitroarginine, a nonvasoactive concentration in this model which reduced basal NO synthesis, preconditioning failed to protect against ischaemia-reperfusion and failed to attenuate NO accumulation produced by ischaemia-reperfusion. When L-nitroarginine was applied after the preconditioning protocol, the effect of preconditioning on test ischaemia-reperfusion and ischaemic NO accumulation was not affected (Csonka et al., 1999; Ferdinandy et al., 1997b). These results prove that intact NO biosynthesis is required for the triggering mechanism of preconditioning and further show that the cardioprotection provided by preconditioning involves a mechanism which decreases the accumulation of NO in the myocardium during ischaemia and reperfusion. In accordance with this, Woolfson et al. (1995) reported that the limitation of infarct size after inhibition of NO synthesis by L-NAME shares a common mechanism with ischaemic preconditioning in rabbit hearts. Furthermore, preliminary studies by Wang & Zweier (1997) showed that preconditioning decreases NO synthesis during subsequent ischaemia and this is associated with the protective effect of preconditioning in isolated rat hearts.

The nature of preconditioning-induced inhibition of NO synthesis is not known. Preconditioning may decrease the rate of enzymatic and/or nonenzymatic (Zweier et al., 1995) NO production during ischaemia-reperfusion by altering cellular pH and the availability of cofactors and/or arginine for NO synthesis, or may possibly stimulate the formation of endogenous NOS inhibitors (Vallance et al., 1992).

In contrast to the aforementioned studies, Lu et al. (1995) reported that 10 mg kg⁻¹ L-NMMA or L-NAME did not affect the antiarrhythmic effect of preconditioning in anaesthetized rats with coronary occlusion/reperfusion. Weselcouch et al. (1995) showed that in rat hearts, 30 μM L-NAME did not interfere with the effect of preconditioning on post-ischaemic myocardial function. In an isolated rabbit heart study, 100 μM L-NAME failed to block the infarct size limiting effect of preconditioning, nor had it any effect on infarct size without preconditioning. However, enhancement of exogenous NO levels by pretreatment with a NO donor induced preconditioning via an antioxidant pathway (Nakano et al., 2000). In these studies, surprisingly, the different NOS inhibitors neither interfered with preconditioning nor the outcome of ischaemia-reperfusion without the preceding preconditioning. Since neither NO generation nor NOS activities were determined in these studies and only a single dose of NOS inhibitors were used, it is difficult to interpret these negative results.

The role of NO in the mechanism of other well established triggers of early preconditioning such as adenosine and bradykinin has not been well established. Little is known whether NO plays a significant role in adenosine-induced early cardioprotection (Cargnoni et al., 1999; Gao et al., 2000; Gattullo et al., 1999; Mubagwa & Flameng, 2001). It seems that bradykinin, a well known endothelium-dependent vasodilator, triggers cardioprotection via an NO-independent pathway in rats and rabbits (Bugge & Ytrehus, 1996; Goto et al., 1995).

The possible role for oxygen free radicals in preconditioning was suggested by Tanaka et al. (1994) who showed that the antioxidants SOD or mercaptopropionyl glycine were both able to inhibit the protective effect of preconditioning on infarct size in rabbits. Osada et al. (1991) reported that the antiarrhythmic effect of preconditioning was lost when the preconditioning ischaemia was applied in the presence of both SOD and catalase in isolated rat hearts. Tritto et al. (1997) showed that a 5 min infusion of O₂⁻· generated by purine/xanthine oxidase prior to ischaemia-reperfusion resulted in a reduction of infarct size in rabbit hearts, similar to the effect of preconditioning. Activation of mitochondrial K_ATP channels has been shown in several studies to play a role in preconditioning (Liu et al., 1998; Schulz et al., 2001), however, it was recently suggested that opening of mitochondrial K_ATP channels triggers the preconditioned state by generating free radicals (Pain et al., 2000). These studies strongly suggest that the generation of free radicals during preconditioning stimuli is necessary to trigger the protective machinery of preconditioning. Preconditioning, in turn, attenuates the increased free radical synthesis during subsequent ischaemia-reperfusion (Tosaki et al., 1994). Others, however, showed that preconditioning was not affected by SOD in rabbit hearts (Iwamoto et al., 1991; Omar et al., 1991), however, it can be explained by the very low cell permeability of SOD.

As the majority of studies show that NO and oxygen free radicals are both required to elicit preconditioning, it was plausible to speculate that formation of ONOO⁻ is an important oxidative stimulus to trigger cellular adaptive mechanisms. Altug et al., (2000; 2001) have found that brief exposure of isolated rat hearts to 1 μM exogenous ONOO⁻ was capable of mimicking the beneficial effects of ischaemic preconditioning, and this was abolished by the administration of the antioxidant mercaptopropionylglycine. Others have shown that administration of a ONOO⁻ generating system triggered early preconditioning in rabbits in vivo. As exogenous ONOO⁻ may not properly reflect the effect of endogenous ONOO⁻ formation, we have recently measured endogenous ONOO⁻ formation during preconditioning induced by three brief cycles of ischaemia and reperfusion and also during subsequent test ischaemia/reperfusion in isolated working rat hearts to clarify the possible contribution of endogenous ONOO⁻ to ischaemic preconditioning. When test ischaemia and reperfusion were preceded by preconditioning, ONOO⁻ formation was markedly attenuated upon reperfusion (Csonka et al., 2001). We have also found that the first brief period of ischaemia/reperfusion significantly enhanced endogenous ONOO⁻ formation, which was reduced after subsequent cycles of brief ischaemia/reperfusion (Csonka et al., 2001). These findings are in accordance with that of Liaudet et al. (2001) showing that in mice and rats with coronary occlusion, myocardial preconditioning attenuates reperfusion-induced activation of poly-ADP ribose synthase. This is possibly due to inhibition of ONOO⁻ formation, as ONOO⁻ is a well-known activator of poly-ADP ribose synthase (Szabó et al., 1996).

Taken together, these results show that ONOO⁻ formation during ischaemia/reperfusion might act as a trigger for preconditioning, but preconditioning in turn decreases formation of ONOO⁻ upon subsequent cycle(s) of ischaemia/reperfusion (Figure 2).

Figure 2.

Increased synthesis of nitric oxide (NO) and superoxide (O₂⁻·) leads to a pronounced formation of the toxic metabolite peroxynitrite (ONOO⁻) upon reperfusion thereby leading to myocardial injury (A). Preceding short periods of ischaemia/reperfusion during preconditioning, however, results in a moderate formation of reactive oxygen species which triggers a cardioprotective mechanism leading to subsequent inhibition of ONOO⁻ formation (B). A similar mechanism is suspected during late preconditioning, however, expression of iNOS and antioxidant enzymes such as superoxide dismutase (SOD) also occurs. Changes in cardiac NO, O₂⁻·, and ONOO⁻ and the exact role of increased iNOS are questions of debate (C). Bars denote periods of ischaemia, arrows indicate formation of reactive oxygen species upon reperfusion or expression of enzymes, font size shows relative amount of NO, O₂⁻·, and ONOO⁻ formation.

NO, O₂⁻·, and ONOO⁻ in late preconditioning

In a conscious rabbit model of preconditioning with repetitive coronary occlusion/reperfusion, Bolli et al. (1997) reported that L-nitroarginine, given either during preconditioning or 24 h later, abrogated the protective effect of preconditioning and that the selective inducible NOS inhibitors aminoguanidine or S-methylisothiourea abolished preconditioning only when applied 24 h after preconditioning. They also showed that preconditioning induces an increase in inducible NOS mRNA levels in the ischaemic regions of the rabbit heart, and that this induction is triggered by increased generation of NO during the preconditioning stimulus (Jones et al., 1999b). Targeted disruption of the inducible NOS gene in mice led to a complete blockade of late preconditioning (Guo et al., 1999). Dexamethasone or selective inducible NOS inhibitors inhibited the late effect of preconditioning on infarct size in anaesthetized rabbits (Imagawa et al., 1999) and on arrhythmias in anaesthetized dogs (Kis et al., 1999). These results suggest a dual role of NO in late preconditioning, as intact NO synthesis by endothelial NOS is necessary to trigger late preconditioning and NO derived from inducible NOS is a mediator of late protection (Figure 2, see for review: Bolli et al., 1998). However, the exact role of inducible NOS in late preconditioning is not known as neither NO levels, nor NOS activities were determined in the aforementioned studies.

The role of NO in the mechanism of other well-known triggers of late preconditioning, such as adenosine and bradykinin has been investigated by several laboratories. The late infarct size limiting effect of bradykinin has been shown to be NO-dependent in a rat coronary occlusion/reperfusion model (Ebrahim et al., 2001). The role of NO in adenosine-induced late cardioprotection is somewhat controversial. Adenosine A(1) receptor stimulation has been shown to induce late preconditioning via an iNOS-dependent pathway in mice (Kudo et al., 2002). The lack of protective effect of adenosine A(1) receptor activation in iNOS gene-knockout mice suggests a direct cause-and-effect relationship of iNOS in adenosine-induced late cardioprotection (Zhao et al., 2000). However, another study showed that delayed protection against myocardial infarction observed 24 h after pharmacologic preconditioning with an adenosine A1 agonist occurred independently of either early generation of NO or subacute induction of inducible nitric oxide synthase in rabbits (Dana et al., 2001). An interesting study showed that in iNOS knock-out mice, eNOS plays a role in delayed A(1) receptor triggered preconditioning (Bell et al., 2002). Others have demonstrated that A(1) receptor-induced late preconditioning uses an NOS-dependent pathway whereas A(3) receptor-induced late preconditioning is mediated by an NOS-independent pathway in rabbits (Takano et al., 2001). In contrast, the inhibition of iNOS with S-methylisothiourea and targeted disruption of the iNOS gene also abolished the protective effect of adenosine A(3) receptor stimulation in mice hearts which shows an essential role of iNOS in A(3) adenosine receptor-induced late preconditioning (Zhao & Kukreja, 2002). The controversial findings can be related to the use of single concentration of inhibitors of adenosine receptor and NOS in the aforementioned studies.

The involvement of oxygen free radicals in late preconditioning was suggested by Sun et al. (1996) who showed that a combination of antioxidants (SOD, catalase, mercaptopropionyl-glycine) infused during the preconditioning stimulus completely abolished the late effect of preconditioning on stunning in conscious pigs. Zhou et al. (1996) reported that isolated rat myocytes preconditioned by anoxia or with the administration of O₂⁻· induced late cytoprotection which was characterized by increased MnSOD activity and decreased O₂⁻· production. Takano et al. (1998) demonstrated that i.v. infusion of the NO donors DETA/NO or SNAP induces cardioprotection 24 h later in conscious rabbits and this effect was lost when they were infused with mercaptopropionyl-glycine. This suggests that the mechanism whereby NO induces preconditioning involves the generation of oxidant species, possibly ONOO⁻.

Taken together, these results suggest that both NO and oxygen free radicals are necessary to trigger late preconditioning. Xuan et al. (2000) have suggested that endothelial NOS is the source of NO during the trigger phase of delayed preconditioning. Since both free radical scavengers and inhibitors of endothelial NOS abolish delayed preconditioning, it has been proposed that the formation of ONOO⁻ is an important upstream event upon the triggering mechanism of delayed preconditioning (Bolli, 2000). Preliminary studies by Emani et al. (1999) have reported that administration of a ONOO⁻ generating system to rabbits evoked a delayed protective response 24 h later, although interpretation of this study is difficult because under these conditions ONOO⁻ could serve as a donor of NO as discussed in detail above. It should be noted that most of the aforementioned conclusions are based mainly on pharmacological studies. To elucidate the exact role of NO, O₂⁻·, and ONOO⁻ in preconditioning requires further biochemical evidence.

Conclusions/Relevance

Severe myocardial ischaemia and reperfusion results in the deterioration of myocardial function and leads to the development of infarction. However, brief episodes of ischaemia trigger an endogenous cardioprotective mechanism known as preconditioning. The role of NO in cardiac ischemia and preconditioning is complex.

In order to understand NO biology, we need to know its relationship with key oxidants such as O₂⁻·. The study of ONOO⁻ in the heart will lead to a better understanding of basic physiological and pathological mechanisms relevant to cardiac ischaemia and reperfusion injury and give new insight to novel therapeutic targets and strategies for its treatment or prevention. Specifically targeting ONOO⁻ is an exciting new strategy to protect the heart from oxidant stress in ischaemia and reperfusion injury.

Abbreviations

GSH

glutathione

K_ATP

ATP-sensitive potassium channel

L-NAME

N^G-nitro-L-arginine-methyl-ester

L-NMMA

N^G-monomethyl-L-arginine

MMP

matrix metalloproteinase

NO

nitric oxide

NOS

nitric oxide synthase

O₂⁻·

superoxide

ONOO⁻

peroxynitrite

SOD

superoxide dismutase