OUP user menu

Role of adenosine in delayed preconditioning of myocardium

G.F Baxter
DOI: http://dx.doi.org/10.1016/S0008-6363(02)00280-8 483-494 First published online: 15 August 2002


Myocardial protection conferred by ischemic preconditioning occurs in a bimodal time course. The early cardioprotection wanes rapidly and is succeeded by a delayed phase of protection reducing infarct development, myocardial stunning and arrhythmias. This ‘second window’ of preconditioning may be evident for up to 72 h. The current mechanistic paradigm for delayed preconditioning against infarction invokes roles for several freely-diffusible molecules, generated during the preconditioning period, that act in autocrine and/or paracrine fashion as triggers of cellular adaptation. These include adenosine, nitric oxide, reactive oxygen species and bradykinin. A role for adenosine receptor activation as a proximal molecular mechanism leading to delayed preconditioning against infarction was established in 1994. Pharmacological adenosine receptor blockade during preconditioning abolishes the acquisition of delayed protection, while transient adenosine A1 or A3 receptor activation fully recapitulates protection against infarction (but not against stunning or arrhythmias) 24 h later. Although nitric oxide is a co-trigger of delayed preconditioning, A1 agonist-induced delayed protection is independent of nitric oxide production. Adenosine receptor activation causes the activation of a complex protein kinase signalling cascade and, putatively, the subsequent activation of gene transcription. The induction or post-translational regulation of several proteins is associated with A1 agonist-induced delayed protection. These include the mitochondrial manganese-conjugated superoxide dismutase, and the 27-kDa heat shock protein. Opening of KATP channels during the index ischaemic event is an obligatory downstream event mediating A1 and A3 agonist induced delayed protection. However, the mechanism of sub-acute regulation of KATP channels following adenosine receptor activation is unknown. Evidence for induction of inducible nitric oxide synthase as a distal mechanism of A1 agonist-induced delayed protection is equivocal.

  • Adenosine
  • Receptors
  • Ischemia
  • Reperfusion
  • Preconditioning
  • CCPA, 2-Chloro-N6-cyclopentyladenosine
  • CGS21680, 2-[4-(2-Carboxyethyl)phenyl-ethylamino]-5′-N-ethylcarboxamidoadenosine
  • COX-2, Cyclo-oxygenase-2
  • HSP27, 27-kDa Heat shock protein
  • HSP72, 72-kDa Heat shock protein
  • IBMECA, N6-(3-Iodobenzyl)adenosine-5′-N-methyluronamide
  • eNOS, Endothelial nitric oxide synthase
  • GR79236, N-[(1S, trans)-2-Hydroxycyclopentyl]adenosine
  • 5-HD, Sodium 5-hydoxydecanoate
  • iNOS, Inducible nitric oxide synthase
  • KATP, ATP-Sensitive potassium channel
  • l-NAME, Nitro-l-argine methyl ester
  • MAP, Kinase mitogen activated protein kinase
  • MAPKAPK-2, MAP Kinase-activated protein kinase-2
  • Mn–SOD, Manganese-conjugated superoxide dismutase
  • NO, Nitric oxide
  • PD115199, 1,3-[Dipropyl-8-N-(2-diethylamino)ethyl]-N-methyl-4-(2,3,6,7-tetrahydro-2,6-dioxo)
  • benzenesulph-onamidexanthine PKC protein kinase C
  • 8-SPT, 8-(p-Sulphophenyl)theophylline
  • TTC, Triphenyltetrazolium chloride

Time for primary review 26 days.

1 Introduction

Soon after the original description of ischemic preconditioning [1], it was observed that the protection is critically time-dependent and disappears if the intervening period between preconditioning and the index ischemic insult is extended beyond 60–120 min [2,3]. Subsequent studies showed a recrudescence of the protected state in preconditioned hearts after 24 h [4,5]. This delayed or late preconditioning phenomenon has been tagged the ‘second window of protection’ (SWOP) [5]. Since 1993, considerable effort has been expended to discover the molecular basis of delayed preconditioning [6–8]. A mechanistic paradigm for delayed preconditioning is illustrated in Fig. 1. During the preconditioning stimulus, a number of factors are generated in myocardium which act as initiators or triggers of both the early and late forms of preconditioning. In relation to delayed preconditioning, the most important of these molecular triggers appear to be rapidly-generated ‘diffusible mediators’: adenosine, bradykinin and NO. Evidence supporting an ‘adenosine hypothesis’ of delayed preconditioning dates from 1994 [9]. Subsequently, the ‘nitric oxide (NO) hypothesis’ of delayed preconditioning was developed and extensively tested [10,11]. More recent evidence supports the involvement of bradykinin as a trigger of delayed preconditioning, signalling through an NO-dependent mechanism [12,13]. It is possible that other local paracrine triggers may participate in delayed preconditioning, including opioid peptides, cytokines and catecholamines. There is evidence that generation of intracellular reactive oxygen species, especially the superoxide anion, acts as a trigger of delayed preconditioning [14,15]. A relatively unexplored area is the role of a variety of early biochemical and biophysical alterations in the preconditioned myocardium which may act as regulatory switches of adaptation. These include changes in oxygen tension, intracellular pH, redox potential, osmotic alterations leading to cell swelling and membrane stretch. Together, these various triggers of adaptation initiate a complex signal transduction sequence that includes the activation of multiple kinases and transcription factors. The distinctive time course of delayed preconditioning [16,17], its abolition by protein synthesis inhibitors [18] and the association of protection with a number of inducible gene products together provide compelling evidence of a distal effector mechanism related to the acquisition of new proteins [6,7]. The focus of this review is the role of adenosine as a proximal component of the delayed preconditioning mechanism.

Fig. 1

Outline of the principal stages leading to cellular adaptation in delayed preconditioning. Several diffusible mediators, including adenosine, contribute critically to the trigger phase that initiates the adaptive response.

2 Adenosine A1 agonist-induced delayed protection

Early experimental evidence supporting the critical role of adenosine as a trigger of classic ischemic preconditioning came from Liu et al. [19] who showed that adenosine receptor blockade with 8-(p-sulphophenyl)-theophylline (8-SPT) abolished the infarct-limiting effect of preconditioning in rabbit heart. It was subsequently shown that transient adenosine A1 receptor (but not A2 receptor) activation with selective agonists reproduced the infarct-limiting effect of ischemic preconditioning in the rabbit, rat and dog [20–23]. Pharmacological evidence also supports a role of adenosine A3 receptor activation in classic ischemic preconditioning [24–26]. The recognition of delayed preconditioning occurred at a time when evidence implicating adenosine in classic preconditioning was unfolding rapidly.

In 1993, our work in the Hatter Institute in London was focused on the acquisition of heat shock proteins and antioxidant enzymes in the mechanism of delayed protection following transient oxidative stress. We were aware of no evidence to suggest that adenosine might regulate the induction or activities of these proteins and we believed that the early and delayed phases of protection were temporally and mechanistically distinct phenomena. A study was undertaken with the null hypothesis that adenosine was not involved as a trigger of delayed preconditioning [9]. Delayed preconditioning was induced in anesthetised rabbits, using 4×5-min occlusions of a left coronary artery branch, each separated by 10-min reperfusion periods. During this preconditioning protocol, animals were treated with an intravenous infusion of 8-SPT for 60 min or saline. Twenty four hours after this pretreatment, the animals were subjected to an index coronary artery occlusion of 30 min after which infarct size was assessed with triphenyltetrazolium chloride (TTC) staining. Ischemic preconditioning resulted in a significant limitation of infarct size compared to sham-operated controls. However, adenosine receptor blockade with 8-SPT during the preconditioning protocol completely abolished the cardioprotective effects observed 24 h later, prompting the conclusion that adenosine receptor activation during ischemic preconditioning was likely to be an essential trigger of the late protective response (see Fig. 2A).

Fig. 2

Principal experimental evidence for participation of adenosine as a trigger of delayed preconditioning and for delayed myocardial protection induced by selective activation of adenosine receptor subtypes. Panel (A) Effects of adenosine receptor blockade on delayed preconditioning in the rabbit. Rabbits were preconditioned (PC) or sham-operated (SHAM). During this period, they received either 8SPT or vehicle (saline). After 24 h they were subjected to 30-min left coronary artery occlusion and infarct size was determined by TTC staining. 8-SPT completely abolished the delayed protective effects of preconditioning, suggesting that endogenous adenosine is a trigger of adaptation. Adapted from Ref. [9]. Panel (B) Effect of 24-h pretreatment with the selective adenosine A1 receptor agonist CCPA. Rabbits were treated with i.v. boluses of saline or CCPA at 25, 50 or 100 μg/kg and after 24 h were subjected to 30-min left coronary artery occlusion. Treatment with CCPA conferred significant limitation of infarct size, suggesting that transient adenosine A1 receptor activation induces delayed myocardial protection 24 h later. Adapted from Ref. [9]. Panel (C) Effects of 24-h pretreatment of rabbits with A1 agonist CCPA (100 μg/kg), A2A agonist CGS21680 (100 μg/kg), or A3 agonist IBMECA (100 or 300 μg/kg). Twenty-four hours after a single bolus of these agents, conscious rabbits were subjected to 30-min left coronary artery occlusion and infarct size assessed. Pretretament with either CCPA or IBMECA induced significant limitation of infarct size. Adapted from Ref. [39]. Panel (D) Time course of delayed protection induced by transient adenosine A1 receptor activation. Saline (filled columns) or CCPA 100 μg/kg (open columns) was administered i.v. to rabbits. At the times indicated, they were subjected to myocardial infarction. Significant limitation of infarct size was observed at 24, 48 and 72 h after CCPA treatment. Adapted from Ref. [40]. *P<0.05 vs. respective control group.

The literature of classic preconditioning at that time suggested that adenosine induced its acute anti-ischemic actions through A1 receptor activation. In a cognate study, non-preconditioned naı̈ve rabbits received a single intravenous bolus of the highly selective adenosine A1 receptor agonist 2-chloro-N6-cyclopentyladenosine (CCPA) or saline vehicle, 24 h prior to index coronary artery occlusion. CCPA has a 10 000-fold selectivity for adenosine A1 versus A2 receptors and a subnanomolar receptor affinity. CCPA was shown to be a relatively short-acting compound, with reversal of the hemodynamic actions (bradycardia and systemic arterial hypotension) within 60–90 min. Increasing doses of CCPA in the range 25–100 μg/kg resulted in progressive reduction of infarct size compared to saline treated animals with the maximum protection observed at 100 μg/kg (Fig. 2B).

These early experiments described a previously unrecognised physiological action of adenosine A1 receptor activation. The prevailing consensus at that time was that adenosine is a short-lived mediator, inducing physiological effects of a relatively short nature. It was now clear that transient adenosine A1 receptor activation could trigger a response that was evident 24 h after the period of receptor occupancy. A subsequent study confirmed that this delayed action of A1 receptor activation was mediated in the myocardium because the late protection against ischemia–reperfusion injury was also evident in isolated Langendorff-perfused rabbit hearts [27]. This observation excluded any contributory mechanism involving systemic neurohormonal or blood-borne elements in the delayed cardioprotective effects of adenosine A1 receptor activation.

Since 1997, persuasive evidence has accrued for the critical involvement of NO as both a trigger and distal mediator of delayed preconditioning [6,10,11]. The elegant ‘NO hypothesis’ of delayed preconditioning, has stimulated a fundamental advance into the molecular pathways of delayed preconditioning and is reviewed more fully in this issue by Roberto Bolli (pp. 506–519). The important question of whether adenosine and NO act as independent triggers arises because some important actions of adenosine A1 receptor activation in cardiovascular cell types appear to be mediated by NO generation [28–31]. In the rabbit, Dana et al. [32] found that the NO synthase inhibitor nitro-l-argine methyl ester (l-NAME) administered prior to an A1 agonist did not prevent the development of late tolerance to ischemia, suggesting that adenosine A1 receptor activation and NO are independent co-activators of the delayed preconditioning response.

3 Adenosine A3 receptor agonist-induced delayed protection

A role for adenosine A3 receptor activation in triggering early ischemic preconditioning is supported by recent experiments undertaken with reportedly selective ligands. The majority of pharmacological studies imply an anti-ischemic effect of acute adenosine A3 receptor activation [24–26,33–36]. Although recent evidence suggests that the protective actions of A3 receptor agonists could be mediated by A1 receptor activation, this has not been definitively proven [37]. Moreover, studies in mice with knock-out of the A3 receptor gene, suggest that A3 receptor activation may have bipolar effects in myocardial ischemia–reperfusion, exacerbates ischemic injury and plays no role in the classic preconditioning response [38]. Of course, the interpretation of knock-out mouse studies can never be regarded as definitive. Deletion of a gene may result in the induction of multiple compensatory changes and in receptor knock-out mice, dissection of the physiological role of a receptor may not be possible if there is a high degree of redundancy in the signalling pathways. The contribution of A3 receptor activation to acute protection against ischaemia is, therefore, unclear. To what extent the acute actions of A3 receptor activation are influenced by the complexities of species differences in A3 receptor pharmacology also remains to be seen.

Against this background of growing controversy, there is limited evidence supporting a role of adenosine A3 receptor activation as a trigger for the delayed protective response in myocardium. Takano et al. [39] have undertaken a pharmacological and molecular characterisation of adenosine receptor participation in delayed protection. They reported that the A3 receptor agonist IB-MECA (100 or 300 μg/kg), given to rabbits 24 h before coronary artery occlusion, resulted in limitation of infarction, comparable to that seen with the A1 agonist, CCPA. The A2A receptor agonist CGS21680 was ineffective (see Fig. 2C). The distal molecular mechanisms appear to differ for A1 and A3 receptor activation (see below). It should be emphasised that there is no evidence to suggest that A3 receptor activation forms part of the endogenous pathway for delayed ischemic preconditioning. Indeed, the receptor antagonist 8-SPT used in the original study pointing to adenosine receptor participation in delayed preconditioning [9] has only weak affinity for A3 receptors yet it abolished the delayed protection. Although at present it seems likely that endogenous adenosine elicits protection via A1 receptor activation, definitive proof of A3 receptor involvement awaits studies of delayed ischemic preconditioning in A3 receptor-knockout mice (with the caveat stated previously) and in studies undertaken with reliably selective A3 receptor antagonists. In the meantime, the potential attraction of therapeutic A3 receptor agonist-triggered cardioprotection (early and late forms) lies in the apparent absence of vascular or cardiodynamic effects of A3 receptor agonists at cardioprotective doses.

4 Temporal characteristics of adenosine A1agonist-induced delayed cardioprotection

4.1 Time course of protection

The time course of the delayed protection against myocardial infarction following transient adenosine A1 receptor activation has been documented in the rabbit [40]. The animals were treated with a single intravenous bolus of CCPA 100 μg/kg or saline vehicle. After a period of 24–96 h, the animals were subjected to index coronary artery occlusion and infarct size was assessed with TTC staining. CCPA treated animals had significantly smaller infarct size compared to saline treated controls at 24, 48 and 72 h but no protection was observed at 96 h (Fig. 2D). The time course of protection after CCPA treatment in this model is identical to that of delayed cardioprotection following preconditioning with ischemia [16]

4.2 Tachyphylaxis and maintenance of the preconditioned state

Experimental transient A1 receptor activation clearly induces a late, long-lasting period of protection. From a therapeutic perspective it may be desirable to extend the duration of this protection still further. Patients with unstable angina for example, are at high risk of death or myocardial infarction following an unstable episode and this risk is particularly high in the first 1–2 weeks [41]. These patients may therefore benefit from pretreatment with agents that trigger or augment myocardial preconditioning over a period of several days or weeks and could maintain the myocardium in a protected state.

Downey's group addressed the possibility of extending the duration of classic preconditioning by using a continuous, high-dose infusion of CCPA in a rabbit model of infarction [42]. Rabbits treated with a 6-h infusion of CCPA showed a significant limitation of infarct size. However, no protection was observed after a 72-h infusion of CCPA. These workers concluded that myocytes become desensitised to the protective effects of CCPA with prolonged exposure, and this tachyphylaxis also rendered the heart refractory to the beneficial effects of ischemic preconditioning. The regulation of adenosine receptors has been studied in a number of different tissues and development of tachyphylaxis to adenosine A1 receptor agonists is both time- and concentration-dependent [43,44]. Thus, reducing the dosing frequency might reasonably be expected to delay or abolish the development of tachyphylaxis. We hypothesised that maintenance of a preconditioned state might be possible without development of tolerance by exploiting the natural time course of delayed protection.

Rabbits were treated intermittently with repeated intravenous boluses of CCPA 100 μg/kg or saline at 48-h intervals [45]. Forty eight hours after the fifth dose (day 10), the animals were subjected to index coronary artery occlusion and infarct size determination. To examine if the rabbits had developed tachyphylaxis to adenosine A1 receptor activation, a subgroup of animals was treated with a further bolus of CCPA 100 μg/kg and the hemodynamic response was monitored for 10 min prior to processing of the heart for infarct size evaluation. Intermittent dosing with CCPA over 10 days resulted in a 42% relative reduction in infarct size compared to vehicle pretreatment, confirming that repeated recruitment of the delayed protective mechanism at 48-h intervals did not result in a loss of protection. Furthermore, application of CCPA after 10 days of intermittent treatment resulted in unmodified hypotension and bradycardia, with no evidence of functional downregulation of A1 receptors.

Travers et al. [46] reported similar experiences with GR79236, another selective A1 agonist. In rabbits, GR79236 induced delayed protection against infarction 24 and 48 h after a single bolus treatment. Administration of GR79236 as a once-daily intravenous dose over a period of 7 days resulted in the maintenance of protection which was evident 24 h after the final dose [27]. These findings suggest that at least in the rabbit, the duration of delayed protection against infarction following pharmacological preconditioning with CCPA can be extended to a minimum of 10 days by intermittent dosing.

5 Adenosine and delayed protection against myocardial stunning

Both A1 and A3 receptor activation independently induce an acute attenuation of stunning in rabbit myocardium [47]. Although adenosine A1 receptor activation robustly induces late protection against infarction in several species, and a functional correlate of infarction in the isolated rabbit heart [27], protection against other endpoints of ischemia–reperfusion injury has not been demonstrated. Notably, in pigs and rabbits, the delayed protection afforded by ischemic preconditioning against myocardial stunning following a subsequent period of ischemia, was not abolished by pretreatment with non-selective adenosine receptor antagonists 8-SPT and PD115199 [48,49] or with the selective A1 agonist N-0861 [47]. Furthermore, transient adenosine A1 receptor activation with CCPA failed to elicit delayed preconditioning against stunning in a conscious rabbit model [48].

6 Signal transduction

6.1 PKC and tyrosine kinases

The intracellular signaling pathways mediating the delayed protective effects of adenosine A1 receptor activation are not completely understood. Although there is some similarity with the pathways identified for delayed ischemic preconditioning, there are divergences that are unsurprising, given the more complex nature of an ischemic trigger versus pharmacological A1 receptor activation. Adenosine receptors couple to protein kinase C (PKC) isoenzymes via the diacylglycerol-phospholipase C pathway [50]. PKC has been well established by pharmacological, biochemical and proteomic interrogations to be an obligatory mediator of delayed ischemic preconditioning [8,51–56]. Extensive studies by Ping and colleagues have established the role of PKC-ε and η isoenzymes, together with tyrosine kinases of the Src family, as obligatory signaling intermediates of ischemic preconditioning in rabbit myocardium [8,55,56]. Dana et al. [57] showed that early inhibition of PKC with chelerythrine chloride, or inhibition of tyrosine kinases with lavendustin A prior to CCPA, completely abolished the late infarct limitation. Thus, one or more PKC isoenzymes and tyrosine kinases play an essential role in mediating the delayed protection against infarction induced by A1 receptor activation. The PKC isoenzymes activated by CCPA were not identified in this study. Because the early inhibition of either PKC or tyrosine kinases completely abolished the delayed protection induced by CCPA, it is likely that the enzymes are activated sequentially rather than in parallel pathways. A role of early tyrosine kinase activation by CCPA is implicated in studies by Kukreja's group [58] in which treatment with genistein prior to CCPA abolished the delayed infarct limiting effect of CCPA in murine myocardium.

6.2 p38 MAP kinase

Among the mitogen-activated protein (MAP) kinase families, the role of stress-activated kinase, p38 MAP kinase, has received special attention in relation to A1 receptor-induced delayed protection. Adenosine treatment of the isolated rat heart was observed to cause rapid phosphorylation of MAP kinase-activated protein kinase-2 (MAPKAPK-2), a downstream substrate of p38 MAP kinase [59]. Dana et al. [57] observed that myocardial p38 MAP kinase catalytic activity was markedly elevated 24 h after treatment with CCPA. This increase in p38 activity correlated with phosphorylation of the small heat shock protein HSP27 (see below). The time course of this p38 MAP kinase induction is unknown as are the identity of the active isoenzyme and the mechanisms of its regulation. It is also unknown if this apparent increase in p38 MAP kinase activity is sustained throughout the period following A1 agonist stimulation until the period of protection 24–72 h later. If this is not the case, how this late increase in kinase activity occurred in the absence of any reinforcing stimulus such as ischemia remains to be determined [60].

Workers in Kukreja's laboratory [58] have reported that SB203580 given prior to CCPA abolished the late protection afforded by CCPA in isolated mouse heart, suggesting that early activation of p38 MAP kinase follows A1 receptor activation in this model. It is of particular interest that in this study A1 associated delayed protection required the phosphorylation of p38 MAP kinase at two time points: as an early event following CCPA administration and also as a late event during the ischemic episode 24 h later. Enhanced phosphorylation of p38 MAP kinase during index ischemia was observed in the hearts from mice pretreated with CCPA 24 h earlier. (This differs from the observation of Dana et al. who reported an increase in the basal activity of p38 MAP kinase before ischemia [57]). This enhanced phosphorylation of p38 MAP kinase was correlated with protection against infarction because SB203580 and genistein given before CCPA, abolished protection and also blocked the phosphorylation of p38 MAP kinase during the subsequent ischemia. A further intriguing observation was that this late phosphorylation of p38 MAP kinase in CCPA-pretreated hearts was abolished by 5-hydroxydecanoate (5-HD) given at the time of index ischaemia, suggesting that opening of the mitochondrial KATP channel participates in the distal signaling mechanism. Evidence for a role of mitochondrial KATP in adenosine-induced delayed protection is discussed in more detail below.

6.3 Transcription

Some important downstream elements of the signalling cascade have been described for delayed ischemic preconditioning and pharmacological preconditioning with nitric oxide donors, including those impinging upon the activation of the key transcription factors nuclear factor-κ B and AP-1 [61,62]. However, no studies to date have reported or dissected any transcriptional mechanisms for A1 agonist induced delayed protection.

7 Distal mediator(s) of A1and A3agonist-induced delayed protection

The gradual onset and prolonged duration of protection following ischemic preconditioning and adenosine agonist-induced delayed protection are compatible with cellular mechanisms involving the new synthesis and degradation of cardioprotective protein(s). In the global context of delayed protection studies, several inducible proteins associated with late protection have been proposed. However, it is essential to emphasise that although pharmacological A1 or A3 receptor activation fully recapitulates the late infarct-limiting effect of ischemic preconditioning, the precise molecular mechanisms controlling this protection may differ according to the nature of the triggering stimulus. With this caveat, most workers believe that the skeleton of the delayed preconditioning paradigm holds true for pharmacological delayed protection, involving an upstream kinase cascade, transcriptional regulation of target genes, and the accumulation and subsequent activation of the effector proteins. The relationship between these putative protein effectors of protection and A1 agonist induced protection are now discussed in approximate chronological order of their investigation.

7.1 27-kDa Heat shock protein (HSP27)

Heat shock proteins (HSPs) are a family of proteins upregulated by stressful stimuli such as hyperthermia, ischaemia–reperfusion or other oxidative stresses [63,64]. There has been interest in the potential of some of HSPs to act as the effectors of delayed ischemic preconditioning. Transfection [65] or transgenic overexpression [66–68] of HSP72 is associated with enhanced tissue tolerance to ischemia–reperfusion injury. In the early report of delayed preconditioning by Marber and colleagues [5], accumulation of myocardial HSP72 occurred in preconditioned hearts. While such associations are provocative, the available evidence is consistent with HSP72 being a marker of the cellular stress response but not an essential mediator of delayed ischemic preconditioning. The protein is not induced by A1 receptor agonist treatment [40,69].

The small HSP family member, HSP27, is a major substrate for MAPKAPK-2, downstream of p38 MAP kinase. Phosphorylation of HSP27 has been suggested to play an important role in the regulation of actin microfilament dynamics [70,71]. Furthermore, it has been demonstrated that overexpression of HSP27 in transfected adult cardiac myocytes resulted in enhanced resistance against ischemic injury [72]. Dana et al. examined the potential role for HSP27 phosphorylation in mediating A1 agonist induced delayed preconditioning [57]. Two-dimensional gel electrophoresis and Western blot analysis of rabbit myocardium pretreated 24 h previously with CCPA revealed an acidic shift in HSP27, corresponding to an increase in mono- and bi-phosphorylated isoforms. Furthermore, prior inhibition of either PKC or tyrosine kinase completely reversed this phosphorylation of HSP27. This association between post-translational modification of HSP27 and A1 receptor mediated delayed protection is intriguing. Although these data are not proof of a causal relationship, they are hypothesis-generating. Ideally the development of HSP27 deficient mice or the expression of mutant HSP27 protein will clarify the role of HSP27.

7.2 Superoxide dismutase

The mitochondrial antioxidant enzyme manganese–superoxide dismutase (Mn–SOD) has been the subject of considerable interest since the earliest reports of delayed ischemic preconditioning. The experimental evidence for involvement of Mn–SOD in preconditioning is reviewed by Masatsugu Hori in this issue (pp. 495–505). Hoshida et al. [73] reported that ischemic preconditioning of canine myocardium induced a biphasic increase in Mn–SOD activity, with the later peak corresponding to the period of delayed protection. Yamashita et al. reported induction of Mn–SOD induced by sublethal hypoxia [52] in neonatal rat cardiomyocytes at a time point that paralleled the appearance of delayed protection. Moreover, treatment with antisense oligonucleotides corresponding to the initiation site of Mn–SOD translation inhibited the increase in Mn–SOD content and activity, and abolished the increased tolerance to hypoxia 24 h after the preconditioning protocol. Pharmacological induction of delayed preconditioning in rat cardiomyocytes by α1 adrenoreceptor stimulation with noradrenaline was associated with induction and activation of Mn–SOD. The delayed protection was abolished by treatment with antisense oligonucleotide to Mn–SOD [74]. Adenosine A1 receptor activation with N6-(phenyl-2R-isopropyl)-adenosine (R-PIA) was found to upregulate several antioxidants acutely in rat cardiac myocytes over a 90–120-min period, including total SOD activity [75]. Dana et al. [76] examined a potential role for Mn–SOD in mediating the delayed cardioprotective effects of CCPA in rabbit myocardium. Myocardial samples from animals pretreated 24 h earlier with CCPA, had increased Mn–SOD activity compared to the controls. Furthermore, induction of delayed protection by CCPA in the rat heart was abolished by administration of antisense oligodeoxynucleotide to Mn–SOD.

7.3 ATP-sensitive K (KATP) channel

Much evidence implicates the ATP-sensitive potassium channel (KATP) in mediating classic preconditioning in a variety of experimental models and species [77,78]. The possible mechanisms by which KATP channels exert this protective effect are reviewed elsewhere in this issue by Brian O’Rourke (pp. 495–505).

Many studies point to important interactions between adenosine A1 receptor activation and the regulation of KATP channel opening. For example Randall et al. [79] and Cordeiro et al. [80] found enhanced responses of the KATP channel to either ischaemia or a pharmacological opener of the channel in the presence of exogenous adenosine. Liu et al. demonstrated a synergistic modulation of KATP currents by PKC and adenosine in isolated rabbit ventricular myocytes [81]. Furthermore, Yao et al. demonstrated an important synergistic relationship between adenosine and KATP channels in the memory of early ischemic preconditioning [82] The KATP channel-dependency of A1 agonist-induced late protection against infarction has been examined in three independent studies in rabbit myocardium [39,69,82]. In two studies it was observed that blockade of KATP channels with either glibenclamide or 5-HD prior to index coronary occlusion abrogated the infarct-limiting effect of CCPA pretreatment 24 h earlier [69,82]. In the study by Takano et al. [39], it was shown that the delayed protection conferred by either A1 or A3 receptor activation in rabbit myocardium was abrogated by 5-HD prior to the index coronary occlusion. In a human-derived Girardi cell model, examining responses to simulated ischemia, Carroll and Yellon demonstrated that exposure to adenosine enhanced cell viability following ischemia 24 h later (assessed by enzyme release and propidium iodide staining) [83]. The delayed protection induced by adenosine was abolished when the pretreated cells were incubated with 5-HD during the subsequent ischemic insult.

The mechanism by which transient activation of A1 or A3 receptors results in subacute modulation of KATP channels 24 h later is not known. Although pharmacological KATP channel blockers suggest pivotal involvement of KATP channel opening during the index ischemic insult, it is not known how opening of the channel is protective. Moreover it is not clear how altered regulation of its opening characteristics occurs 24 h or more after adenosine receptor activation. It is noteworthy that cardioprotection conferred by diverse physical and pharmacological stimuli, including ischemic preconditioning, heat stress, endotoxin derivatives, nitric oxide donors, adenosine A1 and A3 receptor agonists, and δ1-opioid receptor agonists, is effectively blocked by 5-HD. No delayed preconditioning studies, however, have attempted to directly correlate the induction or loss of protection with measurements of mitochondrial KATP channel function. Recent data from Downey and Cohen's laboratory challenges the distal effector role of KATP channel opening in classical preconditioning. KATP channel opening may be associated with the generation of reactive oxygen species provoking a radical re-assessment of this channel as a trigger or signal transduction component [84,85]. With regard to delayed protection, the obligatory nature of KATP channel opening during index ischemia is recognised but the precise mechanism is unclear. Kukreja's observation [58] that activation of p38 MAP kinase during index ischemia was dependent on KATP channel opening may point to a role of KATP channel-related oxidant signalling in delayed protection (see Fig. 3).

Fig. 3

A hypothetical mechanism by which adenosine receptor activation might lead to the late upregulation of proteins associated with the ischemia-tolerant phenotype. During preconditioning, adenosine is generated in the ischaemic tissue. Endogenous adenosine or selective pharmacological agonists activate A1 receptors. Selective pharmacological activation of A3 receptors also induces delayed protection. Following A1 receptor stimulation, early activation of several kinases is required, including PKC, tyrosine kinases and p38 MAP kinase. Roles for other kinases such as other MAP kinase families, JAK-STAT and PI-3 kinase can not be ruled out but no experimental evidence is extant. Transcriptional regulation of target genes is assumed but unproven: by extrapolation from ischemic preconditioning, activation of NF-κB is a likely mechanism. The late appearance of protection is consistent with the time course of new protein synthesis. Associations with A1 agonist-induced protection and the enhanced content or activity of several proteins have been described. These include HSP27 and Mn–SOD. Evidence for obligatory participation of iNOS is equivocal. There is evidence against participation of COX-2 in adenosine-induced delayed protection. Evidence for obligatory participation of KATP opening is strong. There is evidence for the late activation of p38 MAP kinase, possibly via reactive oxygen species signalling downstream of KATP channel opening. This could represent an important pathway for post-translational modification of nascent protein(s) essential for conferring the ischemia-tolerant phenotype. For example, HSP27 is phosphorylated by MAPKAPK-2, downstream of p38 MAP kinase.

7.4 Inducible nitric oxide synthase

An active area of controversy relates to the relationship between adenosine A1 receptor activation and subsequent induction of iNOS. There is compelling pharmacological and genetic evidence that delayed preconditioning induced by ischemia is dependent on the induction and activation of iNOS. This evidence is reviewed by Roberto Bolli elsewhere in this issue (pp. 506–519). However, there is no consensus that A1-agonist induced delayed protection is dependent on iNOS regulation. While one study showed that A1 agonist does not induce late protection in iNOS knockout mice [86], another study in the same mouse strain demonstrated the converse [87]. Moreover, in rabbits, late protection following A1 agonist treatment was not abolished by pharmacological iNOS inhibition with two structurally distinct iNOS inhibitors [32]. However, in another study by Auchampach and colleagues using an essentially similar pharmacological approach in rabbits, A1 agonist induced delayed protection was found to be inhibited by Nω-nitro-l-arginine a non-selective NOS inhibitor which would be expected to effectively inhibit all NOS isoenzymes, including iNOS [39]. It is of special interest that in these workers’ hands, A3 agonist induced late protection was not inhibited by Nω-nitro-l-arginine, implying a critical divergence in the distal mechanisms for A1 and A3 receptor-induced late protection.

7.5 Cyclo-oxygenase-2

An intriguing focus of delayed preconditioning research in Bolli's laboratory during the last 2 years has been recognition of the induction of cyclo-oxygenase-2 (COX-2). COX-2 inhibitors given prior to the index ischemic insult in mice and rabbits abrogate the delayed protection afforded by ischaemic preconditioning [88,89]. Pharmacological evidence suggests that NO from iNOS activates COX-2 through a cGMP-independent mechanism (Bolli, unpublished data reported at AHA Symposium, Seattle, August 2001). It has been shown that neither A1 nor A3 agonist induced delayed protection is contingent on upregulation of COX-2 activity [90]. In view of the apparent co-dependence and co-regulation of COX-2 and iNOS, the discord in the literature on iNOS upregulation following A1 requires emphasis. The relative contributions of these two co-regulated mediators in adenosine-induced delayed protection remain unclear and require further exploration. Moreover, the COX-2 products and distal mechanism that might influence responses to ischemia are unknown. KATP channel activity is known to be influenced by prostanoids and it is intriguing to speculate that, in relation to delayed ischemic preconditioning, channel opening may be distal to COX-2 activation. However, since adenosine-induced delayed protection appears to be KATP dependent and COX-2 independent, factors other than prostanoids must regulate KATP opening in adenosine-induced delayed preconditioning.

8 Conclusion

Adenosine is potent mediator of cardiovascular responses with actions on myocardium and conduction tissue, endothelium and vascular smooth muscle, and on the formed elements in blood. Adenosine is normally regarded as a short-acting mediator because it is rapidly deaminated or rephosphorylated. However, it is clear that in addition to acute effects, adenosine may also mediate sub-acute actions in myocardium, notably the induction of delayed tolerance to ischemia–reperfusion injury. This sub-acute cardioprotective effect of adenosine appears to be an important component of the development of delayed ischemic preconditioning in rabbit myocardium. The late development of an ischemia-tolerant state following ischemic preconditioning is completely abolished by adenosine receptor blockade and can be fully recapitulated by administration of selective A1 receptor agonists. However, the biochemical phenotype induced by pharmacological A1 receptor activation may differ subtly but significantly from that induced by ischemia. This finding may have important implications for our recognition of what constitutes the essential elements of the downstream pathways of protection, rather than epiphenomenal or co-incident changes. The identification of an A3 receptor mediated delayed protection pathway has the appeal of a pharmacological approach to protection devoid of hemodynamic perturbation. Ultimately, the encouraging experimental observation that the sub-acute cardioprotective action of A1 receptor activation is maintainable over a prolonged period provides ground for optimism that pharmacological induction of delayed preconditioning myocardium may have therapeutic potential.


The author gratefully acknowledges the Wellcome Trust and British Heart Foundation which have generously supported his work in this field.


  1. [1]
  2. [2]
  3. [3]
  4. [4]
  5. [5]
  6. [6]
  7. [7]
  8. [8]
  9. [9]
  10. [10]
  11. [11]
  12. [12]
  13. [13]
  14. [14]
  15. [15]
  16. [16]
  17. [17]
  18. [18]
  19. [19]
  20. [20]
  21. [21]
  22. [22]
  23. [23]
  24. [24]
  25. [25]
  26. [26]
  27. [27]
  28. [28]
  29. [29]
  30. [30]
  31. [31]
  32. [32]
  33. [33]
  34. [34]
  35. [35]
  36. [36]
  37. [37]
  38. [38]
  39. [39]
  40. [40]
  41. [41]
  42. [42]
  43. [43]
  44. [44]
  45. [45]
  46. [46]
  47. [47]
  48. [48]
  49. [49]
  50. [50]
  51. [51]
  52. [52]
  53. [53]
  54. [54]
  55. [55]
  56. [56]
  57. [57]
  58. [58]
  59. [59]
  60. [60]
  61. [61]
  62. [62]
  63. [63]
  64. [64]
  65. [65]
  66. [66]
  67. [67]
  68. [68]
  69. [69]
  70. [70]
  71. [71]
  72. [72]
  73. [73]
  74. [74]
  75. [75]
  76. [76]
  77. [77]
  78. [78]
  79. [79]
  80. [80]
  81. [81]
  82. [82]
  83. [83]
  84. [84]
  85. [85]
  86. [86]
  87. [87]
  88. [88]
  89. [89]
  90. [90]
View Abstract