Adenosine Receptor-Mediated Cardioprotection

Abstract

Adenosine is a purine nucleoside, which is produced primarily through the metabolism of adenosine triphosphate (ATP), therefore its levels increase during stressful situations when ATP utilization increases. Adenosine exerts potent cardioprotective effects on the ischemic/reperfused heart, reducing reversible and irreversible myocardial injury. Adenosine receptors (ARs) are G-protein-coupled receptors, and 4 subtypes exist—A₁, A_2A, A_2B, and A₃, all of which have been shown to be cardioprotective. Adenosine receptors are expressed on multiple cardiac cells, including fibroblasts, endothelial cells, smooth muscle cells, and myocytes. Activation of both A₁ and A₃ receptors prior to ischemia has been shown in multiple experimental models to reduce ischemia/reperfusion-induced cardiac injury. Additionally, activation of the A_2A receptor at the onset of reperfusion has been shown to reduce injury. Most recently, there is evidence that the A_2B receptor has cardioprotective effects upon its activation. However, controversy remains regarding the precise timing of activation of these receptors required to induce cardioprotection, as well as their involvement in ischemic preconditioning and postconditioning. Adenosine receptors have been suggested to reduce cell death through actions at the mitochondrial ATP-dependent potassium (K_ATP) channel, as well as protein kinase C and mitogen-activated protein kinase (MAPK) signaling. Additionally, the ability of ARs to interact has been documented, and several recent reports suggest that these interactions play a role in AR-mediated cardioprotection. This review summarizes the current knowledge of the cardioprotective effects of each AR subtype, as well as the proposed mechanisms of AR cardioprotection. Additionally, the role of AR interactions in cardioprotection is discussed.

Introduction

Adenosine is a purine nucleoside catabolite of adenosine triphosphate (ATP) whose concentration increases in the extracellular space during imbalances in myocardial O₂ supply and demand. During such conditions, adenosine exerts multiple effects on the heart, such as slowing of sinoatrial (SA) and/or atrioventricular (AV) node conduction (negative dromotropic effect), resulting in decreased heart rate (negative chronotropic effect), coronary artery vasodilation, and attenuation of the metabolic and functional effects of β-adrenergic receptor stimulation (antiadrenergic effect). The physiological effects of adenosine aid in restoration of myocardial O₂ supply–demand balance.

Myocardial ischemia, such as that which occurs during a coronary occlusion or conditions of induced ischemia during open heart surgery, is the most severe form of myocardial O₂ supply–demand imbalance. Evidence generated over the past 30 years indicates that adenosine can exert multiple beneficial effects during myocardial ischemia, including effects independent of the physiological responses described above. In fact, initial studies examining adenosine’s beneficial effects focused on the hypothesis that adenosine infusion protected the ischemic heart by accelerating postischemic ATP synthesis via the purine salvage pathway. Subsequent studies have since provided substantial support for the hypothesis that adenosine’s effects are mediated by the activation of 4 G-protein-coupled receptor (GPCR) subtypes—A₁, A_2A, A_2B, and A₃—all of which appear to be expressed in mammalian ventricular myocardium. This review article will critically examine our current understanding of the role of these 4 receptor subtypes, presented in the order in which they were discovered, in mediating adenosine’s beneficial effects in ischemic/reperfused myocardium.

Adenosine A₁ Receptors and Cardioprotection

The A₁ receptor was the first adenosine receptor (AR) subtype to be implicated in the cardioprotective effects of adenosine. The adenosine A₁ receptor (A₁R) is a 36-kDa protein that couples primarily to G_i in the heart and has a high affinity for adenosine. The initial evidence for the expression of A₁R in mammalian ventricular myocardium was provided by Lohse et al, ¹ based on the binding of tritiated 8-cyclopentyl-1,3-dipropylxanthine (DPCPX) to bovine myocardial membranes. A year later, this laboratory used the same approach to confirm the expression of A₁R in isolated rat ventricular myocytes. ² Myocardial adenosine A₁R is also located on coronary vascular smooth muscle and cardiac fibroblasts.^3–5 The A₁R remains the only AR subtype to be identified on cardiomyocytes by radioligand binding.

In 1990, our laboratory provided the first evidence that stimulation of the adenosine A₁R prior to ischemia could protect the ischemic heart. ⁶ We reported that R-N6-(phenyl-2R-isopropyl)-adenosine (PIA), an adenosine A₁R agonist, but not phenylaminoadenosine, an adenosine A₂ receptor agonist, significantly delayed the onset of ischemic contracture in the globally ischemic isolated perfused rat heart preparation. The beneficial effect of PIA was identical to that of adenosine, and both the effects of adenosine and PIA were blocked by an A₁R antagonist. The hearts were paced at a constant rate in these studies, indicating that the beneficial effects of A₁R stimulation were independent of its negative dromotropic/chronotropic effects.

Subsequent studies in the next several years indicated that administration of adenosine A₁R agonists prior to ischemia improved postischemic function and reduced infarct size in both isolated hearts and intact animals in multiple species (rat, rabbit, dog, pig, and mouse).^6–13 A₁R cardioprotection is blocked by pertussis toxin pretreatment, consistent with the hypothesis that the effects of cardiac A₁R stimulation are mediated via coupling to a G_i protein. ¹⁴ Additional investigations in several species have demonstrated that adenosine A₁R agonists can protect isolated ventricular myocytes against injury and death induced by simulated ischemia, hypoxia, and other stresses,^15–17 consistent with aforementioned evidence that A₁R are expressed on ventricular myocytes.

In all of the above studies, adenosine A₁R agonists were administered prior to ischemia. There have been fewer studies examining the effects of adenosine A₁R agonists during reperfusion, and these findings have been conflicting. The first of these studies was conducted by Norton et al, ⁷ who reported that low and intermediate doses of the A₁ receptor agonist cyclopentyladenosine (CPA) administered for 65 minutes, beginning 5 minutes prior to reperfusion, reduced myocardial infarct size after 48 hours in rabbits. Finegan et al ¹⁸ reported that the A₁ receptor agonist N6-cyclohexyladenosine (CHA, 50 nmol/L) administered at reperfusion improved the recovery of function in isolated ejecting rat hearts as effectively as when administered prior to ischemia. Similar beneficial effects of reperfusion A₁ receptor agonists were reported by Louttit et al ¹¹ using the A₁ agonist GR 79236 (N-[(1S, trans)-2-hydroxycyclo-entyl]adenosine) in porcine myocardium. Urmaliya et al recently demonstrated in the isolated mouse heart that 100 nmol/L CPA improved cardiac function and reduced infarct size when administered for the first 15 minutes of reperfusion. ¹³ In contrast, other studies have concluded that A₁ receptor agonists administered during reperfusion are not protective. Thornton et al ⁸ reported that PIA and GR 79236 did not reduce myocardial infarct size in intact rabbits. In this study, PIA was infused for only 5 minutes, beginning 2 minutes before reperfusion, at a dose (915 μg/kg), which produced very profound bradycardia and hypotension that persisted for at least 1 hour into reperfusion. The lack of protection could have been due to severe hemodynamic depression and coronary steal. In contrast with Louttit et al, ¹¹ Smits et al, ¹⁰ and Baxter et al ¹⁹ reported that reperfusion treatment with GR 79236 did not reduce infarct size in pigs and rabbits, respectively. These conflicting results with this same agonist could be due to differences in the concentrations and durations of A₁R agonist administration and/or the length of reperfusion.

Adenosine A₁ Receptors, Ischemic Preconditioning, and Ischemic Postconditioning

Initial studies of A₁R agonist cardioprotection in the early 1990s coincided with the first reports of the phenomenon of ischemic preconditioning (IPC), in which brief periods of ischemia–reperfusion prior to a prolonged period of ischemia can significantly reduce myocardial infarct size. While interstitial fluid adenosine levels increase during the brief preconditioning occlusion, and brief adenosine and infusions can mimic the infarct size-reducing effect of IPC, ²⁰ there is conflicting evidence regarding the role of the A1R in IPC. Liu et al showed that 2 AR antagonists blocked the protective effects of IPC in rabbit hearts, and IPC protection was mimicked by an A₁R agonist, leading the authors to conclude preconditioning was mediated by the A₁R. ²¹ However, the next year these results could not be replicated in rat heart. ²² The results of subsequent studies indicated that the A₁R antagonist DPCPX did not block the beneficial effects of IPC in rats ²³ and rabbits,^24–26 but did block IPC in dogs ²⁷ and pigs. ¹¹ A subsequent report indicated that DPCPX did not block IPC in canine myocardium in which a multicycle IPC protocol was used. ²⁸ Thus, it is possible that blocking A1 receptors, at least in canine myocardium, may simply raise the threshold for IPC. Evidence of this threshold effect with respect to blocking the A1R has not been described however in other species. There also appear to be differences between the beneficial effects of IPC and A₁R agonists.^24,29 In fact, selective A₁ receptor agonist preconditioning, in which the A₁R activation is only transient, cannot be accomplished in vivo because of the long half-life of these agents. Results in A₁R knockout (KO) mice are also conflicting, as Lankford et al ³⁰ reported that deletion of this receptor blocked IPC in isolated mouse hearts, whereas Eckle et al ³¹ observed no loss of in vivo IPC in A₁R KO mice. Interestingly, differing results were obtained even though both groups used multicycle IPC protocols. Thus, at the present time the role of the A1R in mediating the cardioprotective effect of IPC remains controversial and may be species dependent.

Similar to the conflicting reports on the efficacy of reperfusion A₁ agonist treatment, there are conflicting theories regarding the role of the A₁R in ischemic postconditioning. Ischemic postconditioning is the phenomenon in which brief cycles of ischemia–reperfusion at the onset of reperfusion reduce infarct size. Kin et al ³² first reported that the A₁ receptor antagonist DPCPX did not block ischemic postconditioning in an in vivo rat model of regional myocardial ischemia. Similar results were observed by Philipp et al ³³ in intact rabbits. However, a report by Donato et al ³⁴ indicated that DPCPX did block the cardioprotective effect of ischemic postconditioning in isolated rabbit hearts. What must be kept in mind with these studies is that DPCPX, even at doses as low as 200 nmol/L, can block adenosine A_2B receptors. ²⁸ Subsequently, Xi et al ³⁵ reported that ischemic postconditioning was blocked in isolated mouse hearts from A₁R KO mice.

In summary, although there is universal agreement that A₁R activation prior to ischemia is cardioprotective, the effects of A₁R agonists during reperfusion and the role of the A₁R in ischemic preconditioning and postconditioning remain controversial, possibly due to species differences.

Adenosine A_2A Receptors and Cardioprotection

The A_2A receptor (A_2AR) is a 45- to 55-kDa protein that couples to G_s in the heart and has a high affinity for adenosine. It has been known for years that adenosine A_2A receptors play a major role in adenosine’s coronary vasodilatory properties, and expression of this receptor in coronary endothelial cells, ³⁶ as well as in coronary smooth muscle cells ³⁷ has been confirmed with radioligand-binding assays. Along with vascular cells, the adenosine A_2AR is expressed in additional cell types in the myocardium. The A_2AR gene is expressed in mast cells, ³⁸ neutrophils, ³⁹ and CD4+ T cells. ⁴⁰ Evidence for adenosine A_2AR expression on the myocyte has also been demonstrated. Bruce Liang and colleagues showed that adult rat ventricular myocytes express adenosine A_2AR messenger RNA (mRNA), ⁴¹ and using immunoblotting our laboratory has shown that adult rat ventricular myocytes express adenosine A_2AR protein. ⁴² A_2AR expression was also shown by Marala and Mustafa using immunoblotting in porcine ventricular myocytes. ⁴³ Our laboratory recently demonstrated the expression of the A_2AR gene in mouse ventricular myocytes as well. ⁴⁴

Unlike the adenosine A₁R, activation of the adenosine A_2AR prior to ischemia does not result in the reduction of myocardial ischemia–reperfusion injury.^8,45–47 In contrast, the cardioprotective effects of the adenosine A_2AR occur when activated at the onset of reperfusion. This initial discovery was published by Norton and colleagues ⁷ using an in vivo rabbit model of regional myocardial ischemia. The selective A_2AR agonist, 2-p-(2-carboxyethyl)phenethylamino-5′-N-ethylcarboxyamidoadenosine (CGS-21680) was administered intravenously at 3 different concentrations (0.001, 0.01, and 0.1 mg/min), starting 5 minutes before reperfusion and continuing for the first 60 minutes of reperfusion (following a 30-minute coronary occlusion). Infarct size after 48 hours reperfusion was reduced with intermediate and high doses of CGS-21680, but not with the low dose. The timing of A_2AR activation at reperfusion appears to be critical, as Boucher et al ⁴⁸ observed that delaying CGS-21680 treatment for as little as 5 minutes after reperfusion in an in vivo rabbit model of regional myocardial ischemia prevented A_2AR-mediated cardioprotection. Similar protective effects of reperfusion A_2AR activation with CGS-21680 and other A_2AR agonists were subsequently shown in vivo in canine, pig, rabbit, rat, and mouse models.^49–59

Although the numerous in vivo studies on A_2AR cardioprotection implicate anti-inflammatory effects, the results of additional studies suggest that this protection could be due in part to direct myocardial effects. As previously mentioned, there is evidence for A_2AR expression in cardiomyocytes,^{41–44,60,61} although the majority of evidence indicates that this receptor exerts no direct effects on cardiac function in normal hearts. We reported that an intracoronary infusion of the A_2AR agonist CGS-21680 increased regional coronary blood flow in an in vivo porcine preparation, but had no effects on regional ventricular function or load-insensitive measurement regional contractility (preload-recruitable stroke work area [PRSWA]). After 15 minutes of coronary occlusion and 2 hours of reperfusion (no irreversible injury), the same intracoronary dose of CGS-21680 significantly increased PRSWA. ⁵³ Additionally, Toufektsian et al showed that infusion of A_2AR agonist ATL146e at 1, 3, and 6 hours post-reperfusion preserved global cardiac function in ischemic/reperfused mice. ⁶² These studies suggest that at least some component of A_2AR-mediated cardioprotection appears to be mediated by cardiomyocyte A_2AR. However, there is very limited evidence that A_2A receptor agonists exert cardioprotective effects during reperfusion in isolated perfused hearts, as the majority of evidence indicates that this approach is ineffective.^{54,55,59,63,64}

A_2A Receptors and Postconditioning

As described above, there is ample evidence that pharmacological activation of the A_2AR just prior to or at the onset of reperfusion is cardioprotective. Additionally, the A_2AR has been implicated in the phenomenon of ischemic postconditioning. The basis of this hypothesis is that the brief, interrupted cycles of reperfusion prevent the rapid washout of extracellular adenosine, thus prolonging AR activation by the endogenous agonist. Kin et al ³² were the first group to test this hypothesis. They reported that washout of adenosine was delayed for the first 2 minutes of reperfusion in postconditioned isolated perfused mouse hearts. In this same study, the authors also observed that the AR antagonist ZM241385, which has a high affinity for A_2AR, blocked the infarct size-reducing effects of ischemic postconditioning in intact rats. It has also been shown that genetic deletion of the adenosine A_2AR attenuates the cardioprotective effects of ischemic postconditioning. ⁶⁴ These findings suggest that endogenous adenosine may act via the A_2AR to promote cardioprotection during ischemic postconditioning.

Adenosine A₃ Receptors and Cardioprotection

The adenosine A₃ receptor (A₃R) is a 36-kDa GPCR that couples to G_i and G_q and has a low affinity for adenosine, although the specifics of this coupling in the heart have not been examined to date. Zhou et al ⁶⁵ published the first report of A₃R gene expression in rat heart. Salvatore et al reported A₃R mRNA in human myocardium a year later. ⁶⁶ There is 1 report to date documenting A₃R gene expression in rabbit cardiomyocytes, ⁶⁷ but this has not been verified in cardiomyocytes of other species. One of the problems in studying A₃R expression in the heart is its very low expression level and the lack of validated, specific antibodies. There are some significant species differences in A₃R structure and pharmacology,^68,69 which may contribute to difficulties in identifying the expression and functional effects of this receptor in the heart.

Indirect evidence for the expression of A₃R in cardiomyocytes is based on the results from various receptor agonists and antagonists. Included in these reports are several studies in neonatal rat cardiomyocytes, suggesting that A₃R stimulation produces direct cardioprotective effects.^16,70 Studies to date in adult cardiomyocytes have been limited. Maddock et al ⁷¹ reported that low concentrations (1 nmol/L) of 2-chloro-N ⁶ -(3-iodobenzyl)adenosine-5′-N-methylcarboxamide (Cl-IB-MECA), which exhibits a strong selectivity for A₃R, decreased hypoxia/reoxygenation-induced apoptosis and necrosis in adult rat ventricular cardiomyocytes when administered at the onset of reoxygenation. Wan et al ⁷² examined the effects of the novel A₃R agonist CP-532,903 [N(6)-(2,5-dichlorobenzyl)-3′-aminoadenosine-5′-N-methylcarboxamide] in adult mouse ventricular cardiomyocytes. These authors reported that CP-532,903 increased whole cell ATP-dependent potassium (K_ATP) channel current in wild-type (WT) cardiomyocytes but not in myocytes isolated from A₃R KO mice.

Initial reports that the A₁R antagonist DPCPX could not block IPC^23–26 or the infarct-reducing effect of the nonselective agonist APNEA (N6-[2-(4-aminophenyl)ethyl]adenosine) ²⁵ led to the hypothesis that the A₃R stimulation was cardioprotective. A subsequent report by Auchampach et al provided more direct support for this hypothesis. ⁷³ This group reported that the AR agonist IB-MECA (N ⁶ -(3-iodobenzyl)adenosine-5′-N-methyluronamide), which exhibits some selectivity for A₃R, exerted cardioprotective effects when administered prior to ischemia in conscious rabbits in the absence of hemodynamic effects. This latter property of A₃R agonists is a major aspect of the interest in A₃R cardioprotection, since adenosine and agonists for other ARs do exert hemodynamic effects, which may limit their potential as pharmaceutical agents. Additional studies with more selective agonists in other species now fully support the hypothesis that A₃R activation prior to ischemia exerts cardioprotective effects.^63,72–79 Included in these reports are studies conducted in isolated perfused hearts of various species, indicating that A₃R cardioprotection can occur via direct myocardial effects.

Adenosine A₃R activation also appears to be capable of reducing myocardial ischemia–reperfusion injury when administered at the onset of reperfusion. This effect has been observed both in isolated perfused hearts^71,76,80 and in in vivo preparations.^77,79,81 Subsequent studies provided evidence that the cardioprotective effect of A₃R activation in vivo could be due to A₃R-mediated inhibition of neutrophil superoxide generation and chemotaxis.^82,83 The most recent of these studies indicated that the selective A₃R agonist Cl-IB-MECA failed to exert a beneficial effect during reperfusion in A₃R KO mice or in chimeric mice lacking the expression of the A₃R in bone marrow-derived cells. ⁷⁷ However, these mechanistic observations cannot explain how A₃R agonists exert a cardioprotective effect during reperfusion in isolated perfused hearts devoid of circulating bone marrow cells. Also recall that Maddock et al ⁷¹ provided evidence that selective A₃R activation during reoxygenation can exert a direct protective effect on adult cardiomyocytes.

Adenosine A₃ Receptors, Preconditioning, and Postconditioning

Early studies in the mid-to-late 1990s examining the role of ARs in the phenomenon of IPC suggested a role for the A₃R, which had only recently been identified. This evidence was based on reports that although IPC could be mimicked by exogenous adenosine, its beneficial effects could not be blocked by the relatively selective A₁R antagonist DPCPX in several species.^25–27,84 These observations were supported by several reports that high doses of the nonselective receptor antagonist 8-sulfophenyltheophylline (8-SPT) were required to block IPC,^25,84 consistent with the fact that rodent A₃R are relatively insensitive to blockade by methylxanthine-type AR antagonists such as DPCPX.^68,69

There have been few, if any, studies using selective A₃R antagonists to directly address the role of A₃R in IPC and ischemic postconditioning, possibly due to the aforementioned species-dependent differences in the A₃R. Two groups have observed no loss of in vivo IPC in A₃R KO mice,^31,85 although one group of investigators ⁸⁵ reported that global genetic deletion of the A₃R actually reduced basal infarct size. In the only published report that we could find using an A₃R antagonist, Kin et al ³² reported that the A₃R selective antagonist, MRS1523, significantly blunted ischemic postconditioning in an in vivo rat model of regional myocardial ischemia.

In summary, there is significant evidence that A₃R activation exerts a cardioprotective effect both prior to ischemia and during reperfusion, although the role of this receptor in IPC and ischemic postconditioning is less clear.

Adenosine A_2B Receptors and Cardioprotection

The A_2B receptor (A_2BR), a 37-kDa GPCR which couples to G_s/G_q, was the fourth AR subtype identified in mammalian myocardium, and to date there is much less information available on the precise role of this receptor compared to the other AR subtypes. The only acknowledged function for cardiac A_2BR is the modulation of coronary flow, which is consistent with the expression of this receptor in coronary artery tissue (vascular smooth muscle and/or endothelium).^36,86 Adenosine A_2BR are also expressed in cardiac fibroblasts.^87,88 There are 2 reports that A_2BR mRNA is present in rat and mouse ventricular cardiomyocytes,^44,64 but there is no physical evidence to confirm the expression of A_2BR protein in cardiomyocytes. The A_2BR, like the A₃R, is a low-affinity receptor, but with large increases in adenosine levels during myocardial ischemia–reperfusion and other stresses, this receptor could play a role in cardioprotection.

One of the first studies implicating A_2BR in the modulation of myocardial ischemia–reperfusion was conducted by Auchampach et al. ²⁸ These authors reported that 2 adenosine A₁R antagonists, DPCPX and BG 9928 (1,3-dipropyl-8-[1-(4-propionate)-bicyclo-[2,2,2]octyl)]xanthine), which exhibit high affinity for canine adenosine A_2BR, reduced myocardial infarct size ∽40% in an in vivo canine model of regional myocardial ischemia. Protection with these antagonists occurred irrespective of whether they were administered prior to ischemia or prior to reperfusion. In contrast, the A₁R antagonist BG 9719 (1,3-dipropyl-8-[2-(5,6-epoxy-s-norbornyl)]xanthine, which exhibits lower affinity for canine A_2BR, had no effect on infarct size. The authors concluded that the infarct size-reducing effects of DPCPX and BG 9928 could be via the inhibition of adenosine A_2BR.

In contrast to the above conclusion, the results of additional studies with AR agonists suggest a protective role for A_2BR activation. The initial report for an A_2BR cardioprotective effect was by Philipp et al ³³ who reported that the nonselective agonist 5′-(N-ethylcarboxamide) (NECA) reduced infarct size in intact rabbits, an effect that was blocked by the A_2BR antagonist MRS1754 but not by A_2AR antagonist 8-(13-chlorostyryl) caffeine (CSC). Subsequent studies by 3 groups indicate that the novel A_2B receptor agonist BAY 60–6583 administered prior to or at the onset of reperfusion reduces myocardial infarct size in isolated rabbit and rat hearts and intact mice.^89–91 The administration of BAY 60–6583 prior to ischemia has also been reported to reduce myocardial infarct size in intact mice and rats.^31,92

There are also only a very limited number of studies on the cardioprotective role of the A_2BR in IPC and ischemic postconditioning. Solenkova et al ⁹³ reported that the A_2BR antagonist MRS1754 blocked the cardioprotective effect of IPC in isolated rabbit hearts, and Eckle et al ³¹ reported that IPC failed to reduce infarct size in A_2BR KO mice, although it appeared that basal infarct size was increased significantly in these animals. In contrast, Maas et al ⁹² reported that the deletion of the A_2BR in mice and pharmacological inhibition using ATL-801 in rats did not alter the ability of ischemic preconditioning to reduce infarct size. Both Eckle et al ³¹ and Maas et al ⁹² used multiple cycle IPC. In the only study to date on the role of A_2BR in ischemic postconditioning, Methner et al ⁹¹ reported that MRS1754 blocked ischemic postconditioning in an in situ mouse model. Thus, at the present time the role of the A_2BR in ischemic preconditioning remains unclear.

Mechanisms of AR-Mediated Cardioprotection

Numerous mechanisms of adenosine-mediated cardioprotection have been proposed over the past 25 years. Initial ideas focused on metabolic mechanisms, with the original hypothesis being that adenosine was rephosphorylated to ATP via the purine salvage pathway. ⁹⁴ Exogenous and endogenous adenosine may exert their beneficial effects via this mechanism. However, given that AR agonists are not readily metabolized, it is unlikely that this mechanism plays a role in their cardioprotective effects. Adenosine A₁R agonists have been shown to modulate metabolism, such as glucose oxidation during reperfusion, ¹⁹ but there is little, if any, data to indicate that the beneficial effects of A_2A, A₃, and A_2B receptor agonists are mediated via direct metabolic effects.

Adenosine Triphosphate–Dependent Potassium Channels

Studies on the mechanisms of AR-mediated cardioprotection have focused primarily on various intracellular signal transduction pathways. Since the cardioprotective effects of adenosine A₁R activation were the first to be recognized, the majority of mechanistic studies have been on this receptor. Initial mechanistic studies were focused on the activation of K_ATP channels and specific protein kinase C (PKC) isoforms. The former hypothesis, which was the focus of numerous investigations in the 1990s, was based on a study by Kirsch et al ⁹⁵ who reported that adenosine A₁R were coupled to sarcolemmal K_ATP channels in neonatal rat ventricular myocytes. However, these results were obtained in only 7 of 16 experiments, under low-ATP conditions. Several subsequent studies in rat and rabbit ventricular myocytes yielded conflicting results on the ability of adenosine and/or A₁R agonists to increase sarcolemmal K_ATP channel activity.^96–100 In addition, there have been only a limited number of studies (with conflicting reports) supporting a role for A₁R stimulation of K_ATP channels during myocardial ischemia–reperfusion.^46,101–103

Conflicting studies on the link between A₁R and sarcolemmal K_ATP channels, along with the identification of a mitochondrial K_ATP channel ¹⁰⁴ led to the subsequent hypothesis that mitochondrial K_ATP channels played a key role in cardioprotection. Although studies using several pharmacological agents implicated mitochondrial K_ATP channels in protection via multiple GPCRs, ¹⁰⁵ pharmacological studies did not support such a mechanism for adenosine A₁R cardioprotection.^106,107 At this point, it appears that the roles of sarcolemmal and/or mitochondrial K_ATP channels in adenosine A₁R-mediated cardioprotection are uncertain at best.

The only other AR subtype linked to stimulation of K_ATP channels is the A₃R. Similar to the A₁R, there have been a limited number of studies addressing this hypothesis, although the evidence is much more consistent. At least 4 studies have been conducted addressing the role of K_ATP channels in mediating A₃R receptor cardioprotection, and the results of all 4 studies support this hypothesis.^{72,75,108,109} Tracey et al ¹⁰⁸ reported that A₃ receptor agonist-mediated cardioprotection was blocked by agents which have been used to implicate both sarcolemmal and mitochondrial K_ATP channels. Wan et al ⁷² provided direct support for A₃R stimulation of sarcolemmal K_ATP channels in adult mouse ventricular myocytes and also observed that A₃ receptor agonist cardioprotection was blocked in mice lacking the pore-forming subunit of the sarcolemmal K_ATP channel (Kir6.2).

Protein Kinase C Signaling

Another signaling mechanism implicated in A₁R cardioprotection is the activation of multiple PKC isoforms. This hypothesis is based on observations that adenosine and A₁R agonist effects can be blocked with various PKC inhibitors.^99,110–114 Additionally, there are several reports demonstrating the effects of A₁R agonists on PKC isoform translocation in normal intact myocardium and/or isolated ventricular myocytes.^115–119 All of these studies support the hypothesis that A₁R activation increases the translocation of PKC-ϵ from cytosolic to membrane fractions. Two of these studies^116,117 indicated no effect of an A₁R agonist on PKC-δ translocation in rat cardiomyocytes, whereas Henry et al ¹¹⁵ did report evidence of PKC-δ translocation in the same preparation. A more recent report indicated that A₁R stimulation in rat cardiomyocytes increased the translocation of both of these PKC isoforms to caveolin 3-enriched membrane fractions. ¹¹⁹ Thus, in contrast to the role of K_ATP channels, there are solid and consistent data supporting a role for one or more PKC isoforms in adenosine A₁R-mediated effects in cardiac tissue. However, there still remain few, if any, studies examining the effects of adenosine A₁R activation on PKC isoform translocation during myocardial ischemia–reperfusion.

There is no evidence to date that reduction of acute myocardial ischemia–reperfusion injury via activation of other AR is associated with or mediated by PKC isoforms. Several recent reports have linked PKC to adenosine A_2BR cardioprotection,^33,90 but the primary hypothesis in these studies is that PKC activation during early reperfusion following ischemic preconditioning is necessary for the activation of the A_2BR or its downstream signaling, rather than PKC being downstream of the A_2BR. This was also suggested by Liang with A₁ and A₃ receptors in cultured myocytes. ¹²⁰ However, this hypothesis is based on pharmacological studies, and there is currently no biochemical evidence for this pathway in the heart. To date, there is very little information available on the mechanism of A_2BR cardioprotection.

Anti-Inflammatory Actions

The results of numerous studies examining the beneficial effects of A_2AR agonists in various animal disease models have led to the hypothesis that A_2AR activation exerts strong anti-inflammatory effects.^50,57–59 Initial in vivo studies of A_2AR-mediated cardioprotection were consistent with this hypothesis, as A_2AR agonists decreased neutrophil adherence and infiltration in in vivo models of ischemia–reperfusion. Jordan et al ⁵⁰ reported that a low dose (2 μg/kg per min) of CGS-21680 reduced infarct size in the dog by nearly 50%. The reduction in infarct size was associated with decreased neutrophil accumulation, and in vitro studies indicated that CGS 21680 reduced both neutrophil adherence to endothelium and superoxide anion production. More recent reports have concluded that the beneficial effects of A_2AR activation could be due to effects on CD4+ T-helper lymphocytes and mast cells. Activation of the adenosine A_2AR on CD4+ T cells has been shown to prevent neutrophil accumulation and reactive oxygen species production in the coronary endothelium. ⁵⁷ The same laboratory subsequently provided evidence that resident mast cell degranulation is prevented by the A₂AR, which resulted in infarct size reduction in isolated perfused mouse hearts. ⁵⁸ Interestingly, while A₃R activation is cardioprotective, 2 initial reports demonstrated the proinflammatory effects of the A₃R.^121,122 Cerniway and colleagues ¹²² showed that targeted deletion of the A₃R actually reduced infarct size in mice by 60% and attributed this to the loss of proinflammatory effects of the A₃R on mast cells. However, subsequent studies indicate that infarct size reduction induced by A₃R agonists administered during reperfusion appears to be due to A₃R anti-inflammatory effects.^81–83

MAP Kinase, Protein Kinase B Signaling

Most recently, studies addressing the mechanisms of AR-mediated cardioprotection have focused on the role of various kinases, such as mitogen-activated protein kinase (MAPK) and PI3 kinase/protein kinase B (AKT) signaling pathways. It has been reported that all 4 ARs, when expressed in chinese hamster ovary(CHO) and/or human embryonic kidney (HEK) cells, are capable of activating one or more of these kinases. ¹²³ Additionally, administration of NECA to CHO cells expressing A_2BR was shown to increase cAMP response element-binding protein (CREB) phosphorylation, ¹²⁴ and stimulation of skeletal muscle cells with NECA has also been shown to induce CREB phosphorylation. ¹²⁵ There is significant evidence, both in intact hearts and in isolated cardiomyocytes, that agonists for all 4 ARs can activate at least 1 MAPK and/or AKT.^70,126–129 There are also reports that pharmacological inhibitors of these pathways block AR-mediated cardioprotection.^{48,93,130–132} Interestingly, although A₁R cardioprotection appears to be due to the activation of one or more protein kinases, there is evidence that A₁R activation in nonischemic myocardium is associated with increased protein phosphatase 2A (PP2a) activity.^133–135 Whether the A₁R modulates PP2a activity during myocardial ischemia–reperfusion, and what role, if any, this plays in A₁ cardioprotection has yet to be determined.

Unresolved Aspects of Adenosine Receptor-Mediated Cardioprotection

As detailed above, there is significant evidence that all 4 AR subtypes exert beneficial effects in ischemic–reperfused myocardium in multiple species. There are, however, several aspects of AR cardioprotection that remain controversial and/or unresolved. For example, although one multicenter animal study concluded that the administration of an A₁R agonist during reperfusion was not cardioprotective, ¹⁹ there are other studies indicating that this is not the case.^7,11,13,17 Additionally, as previously mentioned, the role of the A₁R in ischemic preconditioning is controversial, with some groups showing that the A₁R antagonist DPCPX blocks the beneficial effects of IPC^11,27 and other groups showing no such effect.^23–26 It is possible that these conflicting results were obtained due to the dose of DPCPX used (as mentioned, this antagonist may also block A_2BR ²⁸ ), as well as to the differences in IPC protocols. Species differences may also exist in the role of A₁R in ischemic preconditioning, as the protective effects of IPC were blocked with DPCPX in canine and pig^11,27 but not rat or rabbit.^23–26 With respect to the A_2AR, it is not clear why the long-lasting effects of the A_2AR agonist CGS-21680 exert protection only during reperfusion and not when administered prior to ischemia. Finally, although there is substantial evidence that at least 3 AR subtypes (A₁, A_2A, and A₃) are expressed in ventricular myocytes, and some initial observations suggesting that A_2BR may be expressed in cardiomyocytes^44,89 as well, it is not clear why ventricular cardiomyocytes would express multiple AR subtypes that all exert beneficial effects during ischemia–reperfusion, potentially by stimulating the same signaling pathways.

Some of these controversies may be related to the differences in the concentrations of the different agonists used and the duration of the agonist treatment. An examination of several studies suggests that optimal A_2AR agonist-mediated protection occurs with low doses infused for at least 1 hour, and the results of some studies indicate that A_2AR protection can be induced without significant hemodynamic effects.^52,136 Higher concentrations are associated with significant hypotension and reflex tachycardia and could also induce coronary steal. Higher concentrations of agonists also raise the issue of selectivity. There are no AR subtype-specific agonists or antagonists; rather, these agents are for the most part selective for a particular subtype over only a narrow range of concentrations.

One possible explanation for the apparent differences in the timing of AR-mediated cardioprotection is the largely unrecognized role of endogenous adenosine. Adenosine accumulates in ischemic tissue to levels that are likely to reach the micromolar range, and these levels can remain high during the early minutes of reperfusion. Endogenous adenosine will bind to AR subtypes with varying affinities—A₁ and A_2A receptors have a high affinity for adenosine, whereas A_2B and A₃ receptor subtypes have a low affinity for adenosine. Endogenous adenosine may modulate the effects of exogenous agonists. One source of endogenous adenosine is ecto-5′-nucleotidase, also referred to as CD73, which generates adenosine via the dephosphorylation of extracellular adenosine monophosphate (AMP). Genetic deletion of CD73 has been reported to block both ischemic preconditioning and A_2B receptor agonist-mediated postconditioning.^31,91 Urmaliya et al recently published that A₁R agonist postconditioning in the isolated mouse heart was abolished in both A_2AKO mice and WT mice receiving A_2A or A_2B receptor antagonists, and concluded that this was due to the loss of activation of A₂ receptor subtypes by endogenous adenosine. ¹³

The evidence supporting all 4 subtypes in adenosine cardioprotection could be related to redundancy in signal transduction pathways. As previously noted, all 4 ARs have been shown to activate ERK 1/2 when transfected in CHO cells. ¹²³ Additionally, separate studies have suggested that ARs are capable of activating multiple kinases such as p38 and AKT in both transfected cells and native tissues.^70,126–129 It is possible that optimal AR-mediated cardioprotection requires activation of more than 1 receptor subtype, in order to ensure the most advantageous activation of intracellular signaling pathways. This could also be due to AR subtypes differentially activating signaling in specific subcellular compartments.

Interactions Among ARs in Cardioprotection

As previously mentioned, ARs are expressed in multiple cell types in the heart. While ARs are classically considered cardioprotective, they have been shown to have differential effects on particular cell types. For example, the A₁R has been shown to promote cardioprotection when activated prior to ischemia, while activation of the A_2AR is cardioprotective at the onset of reperfusion. A proposed mechanism for A_2AR protection is through its anti-inflammatory effects, such as reduction of neutrophil polymorphonuclear leukocytes (PMN) adherence to endothelial cells. However, PMNs also express adenosine A₁R, whose activation is associated with the promotion of neutrophil adherence and superoxide generation.^137,138 In fact, there are at least 2 reports that these effects may contribute to myocardial reperfusion injury.^139,140 Thus, it is possible that the opposing effects of PMN adenosine A₁ and A_2A receptor activation may explain some of the confusion regarding the effects of adenosine and AR agonists on infarct size when administered during reperfusion.

Increasing evidence suggests that interactions among receptor subtypes may contribute to AR-mediated cardioprotection. It has been reported that the A_2AR antagonist 4-(-2-[7-amino-2-{2-furyl}{1,2,4}triazolo{2,3-a}{1,3,5}triazin-5-yl-amino]ethyl)phenol (ZM241385), which can block both A_2A and A_2B receptors, blocked the cardioprotective effects of A₁R agonist pretreatment in an in vivo rat preparation. ⁴⁷ Two more recent reports suggested that activation of both the A_2AR and the A_2BR are required for A₁R agonist protection when administered at reperfusion in mouse heart, ¹³ as well as during ischemia in the cardiac cell line H9c2(2-1). ¹⁴¹ Xi et al ⁸⁹ reported that postconditioning in isolated rat hearts with the nonselective AR agonist NECA required simultaneous stimulation of both A_2A and A_2B receptors. A similar conclusion was reached by Methner et al ⁹¹ based on their findings in intact mice. These observations need to be further investigated to determine the mechanistic basis for these apparent interactions. The implicit assumption in these recent observations is that these interactions are occurring at the level of the ventricular cardiomyocyte, since there is evidence that all 4 AR subtypes may be expressed in cardiomyocytes, however this has not yet been determined.

Conclusions

All 4 AR subtypes (A₁, A_2A, A_2B, and A₃) are expressed in the heart, and there is significant evidence that all 4 subtypes exert cardioprotective effects. Consistent findings in multiple species and experimental models indicate that activation of both A₁R and A₃R prior to ischemia are protective, whereas activation of A_2AR and A_2BR are beneficial during reperfusion. It is also accepted that activation of A_2AR prior to ischemia is not protective. Several other aspects of AR-induced cardioprotection are less clear including specific signaling mechanisms by which these beneficial effects are mediated, and whether these beneficial effects are all mediated via direct effects on ventricular cardiomyocytes. This latter aspect is supported by some studies in isolated cardiomyocytes, however the A₁R remains the only subtype confirmed by radioligand binding in ventricular cardiomyocytes.

The results of several recent studies suggest that the presence and activation of more than one receptor subtype is needed for optimal AR cardioprotection, raising the possibility that there may be interactions among ARs. This is an intriguing idea that warrants additional studies, since there is growing evidence of GPCR heterodimerization.^142,143 An additional area that should be examined is the role of ARs in the postinfarct heart, given the expression of ARs in numerous cell types, such as fibroblasts, macrophages, mast cells, and inflammatory cells that play a role in cardiac remodeling.