The relative roles of mitochondrial (mito) ATP-sensitive K+ (mitoKATP) channels, protein kinase C (PKC), and adenosine kinase (AK) in adenosine-mediated protection were assessed in Langendorff-perfused mouse hearts subjected to 20-min ischemia and 45-min reperfusion. Control hearts recovered 72 ± 3 mmHg of ventricular pressure (50% preischemia) and released 23 ± 2 IU/g lactate dehydrogenase (LDH). Adenosine (50 μM) during ischemia-reperfusion improved recovery (149 ± 8 mmHg) and reduced LDH efflux (5 ± 1 IU/g). Treatment during ischemia alone was less effective. Treatment with 50 μM diazoxide (mitoKATP opener) during ischemia and reperfusion enhanced recovery and was equally effective during ischemia alone. A3 agonism [100 nM 2-chloro-N 6-(3-iodobenzyl)-adenosine-5′-N-methyluronamide], A1 agonism (N 6-cyclohexyladenosine), and AK inhibition (10 μM iodotubercidin) all reduced necrosis to the same extent as adenosine, but less effectively reduced contractile dysfunction. These responses were abolished by 100 μM 5-hydroxydecanoate (5-HD, mitoKATP channel blocker) or 3 μM chelerythrine (PKC inhibitor). However, the protective effects of adenosine during ischemia-reperfusion were resistant to 5-HD and chelerythrine and only abolished when inhibitors were coinfused with iodotubercidin. Data indicate adenosine-mediated protection via A1/A3 adenosine receptors is mitoKATP channel and PKC dependent, with evidence for a downstream location of PKC. Adenosine provides additional and substantial protection via phosphorylation to 5′-AMP, primarily during reperfusion.
- mitochondrial adenosine 5′-triphosphate-sensitive K+ channel
- protein kinase C
adenosine can protectthe myocardium from ischemic insult and initiate the protective phenomenon known as preconditioning. The molecular mechanisms involved are incompletely understood, although varied investigations implicate receptor-dependent activation of mitochondrial (mito) ATP-sensitive K+ (mitoKATP) channels (3, 13, 47) and protein kinase C (PKC) (20, 28, 31). Evidence also supports involvement of mitoKATP channels in adenosine receptor (AR)-mediated “delayed” cardioprotection (55). Non-receptor-mediated or metabolic effects of adenosine are not generally considered when protection against ischemia is examined, despite evidence for significant effects of these paths (1). Indeed, we (33, 35) recently acquired evidence of multiple mechanisms mediating protection in response to adenosine. The relative importance of receptor-dependent and -independent pathways in adenosine-mediated protection in ischemic hearts has not been adequately addressed. Furthermore, the protective roles of both mitoKATP channels and PKC have been questioned in recent studies. Cohen and colleagues (4) found that whereas G protein-coupled preconditioning stimuli protect via a mitoKATP channel-dependent process, adenosine is an exception. Additionally, there is evidence against an essential role for PKC in cardioprotective responses (25,30), and even evidence that adenosine can inhibit rather than activate PKC translocation in ischemic myocardium (8). Conflicting data also exist regarding the location of PKC upstream (20, 39, 54) versus downstream (23, 43,50-52) of mitoKATP channels in protective signaling cascades. These varied contradictory findings may reflect contributions of multiple parallel pathways to cardioprotection (19, 44, 45).
Given these uncertainties, the purpose of this study was to characterize protective effects of adenosine in ischemic-reperfused hearts, identify the roles of mitoKATP channels and PKC, and determine the importance of non-receptor-mediated “substrate” effects of adenosine (phosphorylation to 5′-AMP). Because conflicting data exist regarding the ability of adenosine to protect when supplied pre- (24, 37,46) versus postischemia (6, 9, 34), we compared the effects of adenosine before and during ischemia versus during ischemia and reperfusion.
All investigations described in this study conformed to the National Institutes of Health's Guide for the Care and Use of Laboratory Animals (NIH Publication No. 85-23, Revised 1996).
Langendorff-perfused heart preparation.
Hearts were isolated from 16- to 20-wk-old male C57/Bl6 mice (body wt = 24.5 ± 0.4 g, n = 298), anesthetized with 50 mg/kg pentobarbital sodium administered intraperitoneally, and perfused as described by us previously (13, 14). After thoracotomy was performed, the hearts were excised into ice-cold perfusion buffer, and the aorta was cannulated (20-gauge stainless steel) and perfused at 80 mmHg with Krebs buffer containing (in mM) 118 NaCl, 25 NaHCO3, 4.7 KCl, 1.2 KH2PO4, 2.5 CaCl2, 1.2 Mg2SO4, 11 glucose, and 0.6 EDTA. The buffer was equilibrated with 95% O2-5% CO2 at 37°C, giving a pH of 7.4 and Po 2
Hearts stabilized for 20 min at intrinsic heart rate were subjected to pacing at 420 beats/min (ventricular pacing, 2-ms square waves, 20% above threshold) for 10 min. Baseline function was assessed and 20-min global normothermic ischemia initiated, followed by 45-min reperfusion. In addition to untreated hearts (n = 15), experiments were conducted to identify roles and effects of adenosine and ARs, mitoKATP channels, and PKC. To identify effects of mitoKATP activation, hearts were treated with 50 μM diazoxide initiated 10 min before and during ischemia (n = 8), or initiated before ischemia and reinstated throughout reperfusion (ischemia and reperfusion,n = 9). This level of diazoxide has been shown to selectively modulate mitoKATP versus sarcolemmal KATP channels (11, 22, 39) and did not exert significant functional effects. To verify the role of mitoKATP channels in this response and test the importance of PKC, hearts subjected to 50 μM diazoxide during ischemia and reperfusion were cotreated with 100 μM of mitoKATPchannel blocker 5-hydroxydecanoate (5-HD, n = 8) or 3 μM of PKC inhibitor chelerythrine (n = 7). Effects of 100 μM 5-HD (n = 10) or 3 μM chelerythrine (n = 9) alone were also assessed. This concentration of 5-HD selectively blocks mitoKATP activity with little or no effect on sarcolemmal KATP channels (11, 22, 27,39). The level of chelerythrine employed is almost an order of magnitude greater than its inhibitor constant (K i) for PKC inhibition (15) yet below the level recently shown to interact with ARs (40).
To examine the effects of adenosine, hearts received 50 μM adenosine, a concentration previously shown to exert potent protective effects (34) and sufficient to substantially activate A1, A2A, A2B, and A3receptors. Adenosine was given before and during ischemia (n = 8) or during ischemia and reperfusion (n = 9). These responses were reassessed in hearts cotreated with 100 μM 5-HD (n = 8 for ischemia alone, n = 9 for ischemia and reperfusion) or 3 μM chelerythrine (n = 9 for ischemia alone, n = 8 for ischemia and reperfusion). To examine the effects of selective receptor activation, hearts were treated with 100 nM of the A1 adenosine receptor (A1AR) agonistN 6-cyclohexyladenosine (CHA, n = 8) or 100 nM of A3AR agonist 2-chloro-N 6-(3-iodobenzyl)-adenosine-5′-N-methyluronamide (Cl-IB-MECA, n = 8), initiated 10 min before ischemia and reinstated during reperfusion. The CHA concentration is greater than its K i at A1 receptors (∼1 nM) but severalfold lower than itsK i at A2 receptors (500 nM) (17). The Cl-IB-MECA concentration is also above itsK i at A3 receptors (of ∼1 nM ) and well below its K i at A1 and A2 receptors (500–1,000 nM) (16). Neither agonist modified flow in normoxic hearts, demonstrating lack of A2 receptor activation. Responses to CHA and Cl-IB-MECA were also assessed in hearts cotreated with 100 μM 5-HD (n = 8 for both CHA and Cl-IB-MECA) or 3 μM chelerythrine (n = 8 for both CHA and Cl-IB-MECA).
Effects of chelerythrine on receptor-mediated effects of adenosine were studied in a separate series of experiments. Specifically, hearts were stabilized and were either untreated (n = 8), treated for 5 min with 50 μM adenosine alone (n = 8), treated with 3 μM chelerythrine for 15 min (n = 7), or treated with 3 μM chelerythrine for 15 min with 50 μM adenosine infusion initiated after 10 min (n = 8). Maximal reductions in heart rate (A1AR-mediated bradycardia) and elevations in coronary flow (A2AR-mediated dilation) were assessed.
To examine the role of adenosine phosphorylation via adenosine kinase (AK), hearts were treated with 10 μM of the AK inhibitor iodotubercidin alone (n = 9), or in the presence of 100 μM 5-HD (n = 9) or 3 μM chelerythrine (n = 8). A 10 μM concentration of iodotubercidin will maximally block AK activity (18, 35). Protection with 50 μM adenosine during ischemia and reperfusion was also assessed in the presence of 10 μM iodotubercidin alone (n = 9), 10 μM iodotubercidin plus either 100 μM 5-HD (n = 8), or 3 μM chelerythrine (n = 8).
Determination of myocardial injury via enzyme efflux.
To assess necrosis, effluent was collected throughout reperfusion for LDH quantitation. Samples were stored at −80°C until analyzed enzymatically (Sigma; St. Louis, MO). Postischemic LDH efflux over 45-min reperfusion (IU/g) was determined by multiplying concentration (IU/ml) by effluent volume (ml/g).
Data are presented as means ± SE of individual experiments. Functional responses to ischemia-reperfusion and drug treatments were compared by multiway ANOVA with repeated measures. Postischemic LDH efflux was compared with one-way ANOVA. When differences were detected by ANOVA, a Tukey's post hoc test was employed for specific comparisons. In all tests, a value ofP < 0.05 was considered significant.
Functional response to ischemia-reperfusion.
Ventricular pressure development was high, and coronary flow was submaximal under normoxic conditions (Fig.1). Contractile function failed to recover to preischemic levels after 45-min reperfusion. In control hearts diastolic pressure remained elevated at ∼20 mmHg, and developed pressure recovered <50% of preischemic levels (Fig.1). After an initial hyperemic response, coronary flow recovered to ∼90% of preischemic levels in all groups (Fig.1 C).
Protection with adenosine, diazoxide, and AR agonism.
No differences in normoxic contractile function were detected between untreated hearts and those treated with diazoxide, adenosine, CHA, or Cl-IB-MECA (Figs. 1 and 2). Adenosine predictably elevated preischemic coronary flow (Fig. 1 C). Diazoxide provided cardioprotection when present during ischemia and reperfusion (Fig. 1). An almost identical protection was evident with the treatment during ischemia alone. Treatment with adenosine substantially improved postischemic contractile function, with a much greater degree of protection evident when supplied during ischemia and reperfusion versus ischemia alone (Fig.1). The effects of adenosine supplied during ischemia and reperfusion exceeded those for diazoxide. Protection was also evident with A1AR agonism with CHA, and A3AR agonism with Cl-IB-MECA (Fig. 2). Coronary flow was enhanced in adenosine-treated (Fig. 1 C) and Cl-IB-MECA-treated hearts (Fig. 2 C) during reperfusion. This dilation was not evident in any other group.
Myocardial LDH efflux in control hearts was ∼23 IU/g during 45-min reperfusion (Fig. 3). Adenosine, diazoxide, CHA, and Cl-IB-MECA all significantly reduced postischemic LDH efflux, with the greatest reductions evident for adenosine and Cl-IB-MECA. In contrast to the marked difference in functional recovery with adenosine treatment during ischemia and reperfusion versus ischemia alone (Fig. 1), there was a modest difference in LDH efflux between these two groups (Fig. 3).
Roles of mitoKATP channels, PKC, and AK in cardioprotection.
Infusion of the PKC inhibitor chelerythrine did not modify intrinsic ischemic tolerance (Fig. 4) and failed to modify receptor-mediated functional effects of 50 μM adenosine treatment (Table 1). Chelerythrine abrogated cardioprotection with diazoxide, ischemic adenosine, CHA, and Cl-IB-MECA (Figs. 3 and 4). Chelerythrine inhibited both anti-necrotic (Fig. 3) and functional effects (Fig. 4). Whereas chelerythrine also eliminated the anti-necrotic effects of adenosine treatment during ischemia and reperfusion, functional protection was only modestly reduced (Fig.4). Postischemic contractile recovery remained substantially elevated above that for untreated hearts. Enhanced postischemic coronary flows observed with adenosine and Cl-IB-MECA treatments were abolished by chelerythrine (Fig.4 C). Reflow in all other groups was unaltered.
The effects of mitoKATP channel blockade with 5-HD essentially mirrored those of chelerythrine. Treatment with 5-HD failed to modify intrinsic ischemic tolerance (Figs. 3 and 4) and abrogated cardioprotection with diazoxide, ischemic adenosine, CHA, and Cl-IB-MECA. Infusion of 5-HD eliminated the anti-necrotic effects of adenosine treatment during ischemia and reperfusion (Fig. 3), whereas substantial contractile protection was still evident (Fig. 4). As opposed to chelerythrine, 5-HD failed to alter postischemic reflow in adenosine and Cl-IB-MECA-treated hearts (Fig. 4 C).
Inhibition of adenosine kinase with 10 μM iodotubercidin exerted cardioprotection, and this effect was eliminated by 5-HD and chelerythrine (Fig. 5). Coinfusion of iodotubercidin with adenosine reduced the level of functional protection observed with adenosine alone but did not reduce the anti-necrotic effect (Fig. 5). Coinfusion of 5-HD or chelerythrine with iodotubercidin totally eliminated functional protection observed in adenosine-treated hearts (Fig. 5). The elevation in postischemic coronary flow with iodotubercidin was inhibited by PKC inhibition with chelerythrine (Fig. 5), as was the increase in flow with adenosine (Fig. 4).
The chief goals of this study were to characterize cardioprotective effects of exogenous adenosine and identify the relative importance of mitoKATP channels, PKC, and AK in observed protection. We assessed responses to exogenous adenosine and A1AR and A3AR agonism, testing the effects of mitoKATP channel inhibition and activation with a well-characterized blocker (5-HD) and opener (diazoxide) (11, 27,22, 39), PKC inhibition with the isoform nonspecific blocker chelerythrine (15, 43, 54), and AK inhibition with iodotubercidin (18, 35). Data reveal that ischemic tolerance in the murine heart is enhanced by adenosine, A1AR and A3AR agonism, AK inhibition, and mitoKATP channel activation. Effects of adenosine on necrosis appear primarily mediated during or shortly after the ischemic insult and occur via a mitoKATP channel and PKC-dependent process. Cardioprotection with A1AR and A3AR agonism, and AK inhibition is also mitoKATP and PKC dependent. However, adenosine treatment during ischemia and reperfusion also enhances outcome via a mitoKATP/PKC-independent path involving the activity of AK.
Characteristics of adenosine-mediated cardioprotection and role of AK.
To identify the effects of adenosine and ARs, we studied responses to exogenous adenosine, exogenous A1AR and A3AR agonism, and enhanced endogenous adenosine at the expense of phosphorylation to 5′-AMP. Data reveal exogenous adenosine protects against necrosis and contractile dysfunction when supplied during ischemia alone and provides further protection when supplied during reperfusion (Figs. 1 and 3). These observations contrast the effects of the mitoKATP channel activator diazoxide, which does not further enhance final recovery of function when supplied during ischemia and reperfusion versus ischemia alone (Fig. 1). Interestingly, postischemic diazoxide does accelerate initial functional recovery during the first 10–20 min of reperfusion. This more rapid recovery supports some short-term benefit from mitoKATP channel activation postischemia, potentially involving a reduction in stunning. We (13) previously observed beneficial effects of intrinsically activated mitoKATP channels in postischemic myocardium.
Previous studies (37, 41, 46, 53) suggest receptor activation before and during ischemia is essential in reducing stunning in vivo and in vitro. Conversely, other studies (6, 9,34) demonstrate specific protection during reperfusion. Our data indicate that the addition of adenosine during reperfusion does enhance recovery beyond that observed with adenosine during ischemia alone (Figs. 1 and 3), consistent with the observations of Gao et al. (9) in isolated mouse hearts. The final level of functional protection with adenosine supplied during ischemia and reperfusion markedly exceeds that with diazoxide, A1AR agonism, or A3AR agonism (Fig. 4), although these all comparably enhanced recovery of function during the initial 20 min of reperfusion (Figs. 1 and 2). Given reasonably high levels of CHA and Cl-IB-MECA employed, coupled with ischemic elevations in endogenous adenosine (34), it is likely A1AR and A3ARs are maximally activated during ischemia in the CHA and Cl-IB-MECA groups. Thus the greater functional protection with adenosine suggests protective processes additional to A1AR and A3AR activation. This effect does not involve A2AAR activation because these do not reduce injury in the current model (33). A non-receptor-mediated protective response is supported by our prior observation that antagonism of extracellular receptors does not eliminate adenosine-mediated functional protection (33). Additionally, Finegan et al. (6) showed that protection with adenosine is not eliminated by receptor antagonism whereas protection with selective receptor agonists can be abrogated.
Effects of iodotubercidin indicate that adenosine mediates some cardioprotection through phosphorylation to 5′-AMP (Fig. 5). When supplied alone, this AK inhibitor protects against ischemia-reperfusion, in agreement with earlier findings (33, 35). The protective response to iodotubercidin alone is related to enhancement of endogenous adenosine levels and resultant receptor activation, as demonstrated recently (33). An adenosine-receptor-mediated effect is also consistent with the comparable effects of PKC and mitoKATP channel inhibitors on responses to iodotubercidin and A1AR and A3AR agonists (Figs. 4 and 5). Despite protection with iodotubercidin alone, the drug significantly reduces protection with adenosine (Fig. 5). This somewhat paradoxical observation is explained by the fact that adenosine not only activates cell surface receptors but is also substrate for metabolic reactions, including phosphorylation by AK. Provision of iodotubercidin, either alone or in the presence of exogenous adenosine, enhances AR-mediated protection (33) while simultaneously blocking adenosine phosphorylation. The iodotubercidin-dependent reduction in protection with exogenous adenosine therefore indicates phosphorylation by AK contributes to protection. This agrees with the early findings of Bolling and colleagues (1), who concluded that metabolic substrate effects of adenosine are important in protection against ischemia.
Importance of mitoKATP channels and PKC in adenosine-mediated cardioprotection.
Although there is evidence that adenosine-mediated protection involves PKC and/or mitoKATP channels (20, 28, 31, 42, 47,49), the protective roles of these signaling elements have been questioned (4, 25, 30). In support of a role for mitoKATP channels, diazoxide provided protection similar to that with A1AR and A3AR agonism (Figs.1-3). More convincingly, the protective responses to CHA (an A1-selective agonist) (17, 24) and Cl-IB-MECA (an A3-selective agonist) (16, 21, 47) were entirely abrogated by 5-HD and chelerythrine (Figs. 3 and 4), and protection with iodotubercidin was also sensitive to these inhibitors (Fig. 5). These data collectively indicate that receptor-mediated protection is dependent on mitoKATP channels and PKC. These data also verify that levels of 5-HD and chelerythrine employed are sufficient to fully eliminate protection mediated via A1and A3 receptor agonism.
It is generally considered PKC lies upstream of mitoKATPchannels in the signaling cascade (20, 39, 54). However, there is also evidence mitoKATP channel-mediated protection is PKC dependent (23, 43, 50-52). We find that the PKC inhibitor chelerythrine blocks cardioprotection with the mitoKATP channel opener diazoxide (Figs. 3 and 4). This supports a downstream location of PKC distal to the target activated by diazoxide (i.e., mitoKATP channels). This observation agrees with findings of Ashraf and colleagues (43,50-52) in different models. Although this contrasts the conventional path in which PKC is upstream of mitoKATPchannels (20, 39, 54), our findings and those of Ashraf et al. do not exclude an upstream function of PKC. Either A1or A3AR may modify mitoKATP channel activity via a PKC-dependent process, with activated mitoKATPchannels subsequently initiating a protective cascade involving modulation of PKC. This is consistent with evidence for a “codependent” protective process requiring both mitoKATP channel and PKC activities (12).
Non-receptor-mediated protection with adenosine.
As already discussed, protection with adenosine markedly exceeds that with mitoKATP channel activation and A1AR or A3AR agonism. Furthermore, protection with adenosine during ischemia and reperfusion is not eliminated by 5-HD or chelerythrine, despite abrogation of responses to CHA, Cl-IB-MECA, ischemic adenosine, and diazoxide (Figs. 3 and 4). Reduced adenosine-mediated protection with iodotubercidin and elimination of protection after cotreatment with iodotubercidin and either 5-HD or chelerythrine indicates the response involves phosphorylation to 5′-AMP. The 5-HD and chelerythrine-resistant protective effect, evident with postischemic adenosine only, may reflect improved nucleotide pool repletion, as suggested by Bolling et al. (1), and/or modulation of AMPK activity. Activation of AMPK has been shown to protect against ischemic insult (5, 36). Cardioprotection via AMPK appears related to glucose metabolism (5), and studies show AMPK stimulates glucose transport and metabolism (38), potentially via modulation of ERK activity (2, 38). Iodotubercidin can inhibit effects of activated AMPK on glucose metabolism (38) and also inhibits ERK activation (7). The effects of iodotubercidin on AMPK are likely mediated via AK inhibition because AMPK is regulated by 5′-AMP levels, and AMP mimetics must also be phosphorylated by AK to modulate the enzyme. Thus the effects of iodotubercidin are consistent with a role for AMPK in adenosine-mediated protection. Irrespective of precise mechanism, our data collectively demonstrate that adenosine mediates substantial protection, which is not entirely explained by AR activation, is partly independent of mitoKATP and PKC activities, and is inhibited by a putative AK inhibitor.
An important limitation in the present study relates to the recent observation chelerythrine can inhibit ligand binding at, and activation of, ARs (40). Although inhibitory concentrations of chelerythrine exceed those employed here, we nonetheless tested this possibility. Because we found that chronotropic and vasodilatory effects of 50 μM adenosine are unaltered by 3 μM chelerythrine (Table 1), nonspecific antagonism of ARs does not appear to be a major confounding factor in the present study. Nonetheless, this may contribute to effects of chelerythrine in other studies, particularly when levels of inhibitor are >3–5 μM.
Another complicating factor relates to potential nonspecific effects of iodotubercidin, which may activate fatty acid oxidation (10) or inhibit protein kinases (26). However, nonspecific effects cannot explain responses to iodotubercidin in our model: effects of iodotubercidin are blocked by PKC and mitoKATP channel inhibition with chelerythrine and 5-HD (Fig. 5) and are inhibited by AR antagonism (35). These data support an AR-mediated action of iodotubercidin alone.
A final point to highlight is that protection observed with CHA is counter to prior studies in which we found no protection with the alternate A1AR agonistN 6-cyclopentyladenosine (33, 34). The explanation for this difference is not obvious. One possible confounding factor is that CHA may be less specific thanN 6-cyclopentyladenosine and therefore activate protective A3ARs. Alternatively (or in addition), there is evidence N 6-cyclopentyladenosine (but not CHA) is rapidly transported and catabolized by mammalian cells (32). This might limit efficacy of the drug as a cardioprotectant. Furthermore, uptake and phosphorylation of these types of adenosine analogs rapidly induces apoptosis (29). Such effects ofN 6-cyclopentyladenosine could counteract beneficial effects of A1AR activation. These possibilities remain to be directly tested.
In conclusion, our observations reveal that A1 and A3AR agonists mediate protection via a mitoKATPchannel and PKC-dependent process. Data support location of PKC downstream of mitoKATP channel activation in the protective cascade. This mitoKATP channel/PKC-dependent protection appears to occur primarily (but not exclusively) during ischemia versus reperfusion. Adenosine itself additionally mediates substantial protection, which is neither mitoKATPchannel nor PKC dependent and is inhibited by treatment with iodotubercidin, supporting a role for phosphorylation to 5′-AMP via AK. This response appears to occur primarily during the reperfusion period, and is consistent with the early studies of Bolling and colleagues (1) supporting protection via non-receptor-mediated effects of the purine nucleoside.
This work was supported by National Health and Medical Research Council of Australia Grant 145310 and by National Heart Foundation of Australia Grants G 98B 0080 and G 99B 0246. J. P. Headrick is the recipient of a career research fellowship from the National Heart Foundation of Australia.
Address for reprint requests and other correspondence: J. P. Headrick, Heart Foundation Research Centre, Griffith Univ. Gold Coast Campus, Southport, QLD 4217, Australia (E-mail:).
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
First published September 19, 2002;10.1152/ajpheart.00717.2002
- Copyright © 2003 the American Physiological Society