Brief ischemic episodes protect the heart from subsequent

Transcription

1 Selective Pharmacological Agents Implicate Mitochondrial but Not Sarcolemmal K ATP Channels in Ischemic Cardioprotection Toshiaki Sato, MD, PhD; Norihito Sasaki, MD, PhD; Jegatheesan Seharaseyon, PhD; Brian O Rourke, PhD; Eduardo Marbán, MD, PhD Background Pharmacological evidence has implicated ATP-sensitive K (K ATP ) channels as the effectors of cardioprotection, but the relative roles of mitochondrial (mitok ATP ) and sarcolemmal (surfacek ATP ) channels remain controversial. Methods and Results We examined the effects of the K ATP channel blocker HMR1098 and the K ATP channel opener P-1075 on surfacek ATP and mitok ATP channels in rabbit ventricular myocytes. HMR1098 (30 mol/l) inhibited the surfacek ATP current activated by metabolic inhibition, whereas the drug did not blunt diazoxide (100 mol/l)-induced flavoprotein oxidation, an index of mitok ATP channel activity. P-1075 (30 mol/l) did not increase flavoprotein oxidation but did elicit a robust surfacek ATP current that was completely inhibited by HMR1098. These results indicate that HMR1098 selectively inhibits surfacek ATP channels, whereas P-1075 selectively activates surface K ATP channels. In a cellular model of simulated ischemia, the mitok ATP channel opener diazoxide (100 mol/l), but not P-1075, blunted cellular injury. The cardioprotection afforded by diazoxide or by preconditioning was prevented by the mitok ATP channel blocker 5-hydroxydecanoate (500 mol/l) but not by the surfacek ATP channel blocker HMR1098 (30 mol/l). Conclusions The cellular effects of mitochondria- or surface-selective agents provide further support for the emerging consensus that mitok ATP channels rather than surfacek ATP channels are the likely effectors of cardioprotection. (Circulation. 2000;101: ) Key Words: mitochondria potassium ischemia preconditioning Brief ischemic episodes protect the heart from subsequent lethal ischemic injury (ischemic preconditioning, IPC). 1 Although the precise mechanism of IPC remains elusive, much attention has focused on the potential role of ATPsensitive K (K ATP ) channels as the effectors of protection. Cardiac myocytes contain 2 distinct K ATP channels: the classic one in the surface membrane (surfacek ATP channel) 2 and another in the mitochondrial inner membrane (mitok ATP channel). 3 Although the cardioprotective effects were initially attributed to surfacek ATP channels, the effects of surfacek ATP channels on excitability cannot account for the protection. 4 7 Recent studies provide further evidence that mitok ATP channels rather than surfacek ATP channels are the dominant players. Diazoxide, a selective mitok ATP channel opener in cardiac myocytes, is cardioprotective. 8,9 The mitok ATP channel blocker sodium 5-hydroxydecanoate (5HD) 10,11 can prevent diazoxide-induced cardioprotection 8,9 and can block genuine IPC Activation of protein kinase C, which figures prominently in the signal transduction cascade of IPC, potentiates mitok ATP channel opening. 10 In light of such new information, the role of surfacek ATP channels in cardioprotection needs to be reevaluated. A selective opener and a blocker of mitok ATP channels, namely diazoxide and 5HD, have been identified. A missing link, however, has been the absence of selective agonists or antagonists of surfacek ATP channels. In the present study, we first examined the effects of the K ATP channel blocker HMR and the K ATP channel opener P on K ATP channels in rabbit ventricular myocytes. Our results show that both HMR1098 and P-1075 target surfacek ATP but not mitok ATP channels. Using these surface-selective agents, we investigated whether surfacek ATP channels are involved in cardioprotection. Methods This investigation conformed to the Guide for the Care and Use of Laboratory Animals published by the US National Institutes of Health (NIH publication 85-23, revised 1985). Flavoprotein Fluorescence and Electrophysiology of Rabbit Ventricular Myocytes Ventricular myocytes were isolated from rabbits 17 and cultured on coverslips in M199 with 5% FBS at 37 C. Experiments were performed the next day. Mitochondrial matrix redox state, reported by the fluorescence of FAD-linked enzymes, 18,19 was used to index mitok ATP channel Received September 20, 1999; revision received November 19, 1999; accepted December 6, From the Institute of Molecular Cardiobiology, Johns Hopkins University, Baltimore, Md. Dr Sato is now at the Department of Physiology, Oita Medical University, Oita, Japan. Correspondence to Eduardo Marbán, MD, PhD, Director, Institute of Molecular Cardiobiology, Johns Hopkins University, Ross 844/720 Rutland Ave, Baltimore, MD marban@jhmi.edu 2000 American Heart Association, Inc. Circulation is available at

2 Sato et al Mitochondrial K ATP Channels in Cardioprotection 2419 activity. 8 Cells were superfused with solution containing (in mmol/l) NaCl 140, KCl 5, CaCl 2 1, MgCl 2 1, and HEPES 10 (ph 7.4 with NaOH) at room temperature ( 22 C). Endogenous flavoprotein fluorescence was excited with a xenon arc lamp with a bandpass filter centered at 480 nm. Emitted fluorescence was recorded from 1 cell at a time at 530 nm by a photomultiplier tube and expressed as a percentage of the DNP-induced fluorescence. In some experiments, flavoprotein fluorescence was measured during whole-cell patch-clamp experiments to administer drugs through the pipette (cf Figure 3). For whole-cell patch recordings, the internal pipette solution contained (in mmol/l) potassium glutamate 120, KCl 25, MgCl 2 0.5, K-EGTA 10, HEPES 10, and MgATP 1 (ph 7.2 with KOH). Currents were elicited every 6 seconds from a holding potential of 80 mv by 2 consecutive steps to 40 mv (for 100 ms) and 0 mv (for 380 ms). Current amplitude at 0 mv was measured 200 ms into the pulse to quantify surfacek ATP channel activity. In some experiments (eg, Figure 6), whole-cell currents and flavoprotein fluorescence were recorded simultaneously, and flavoprotein fluorescence was excited during the 100-ms step to 40 mv. Functional Expression of K ATP Channels and Electrophysiology Details of the functional expression of K ATP channels in HEK 293 cells have been described previously. 11 Plasmid DNA (3 g total) containing Kir6.1 or Kir6.2 was cotransfected with either SUR2B or SUR2A cdna into HEK cells by use of lipofectamine (Gibco) 18 hours after the cells were split. Mouse Kir6.1, provided by Prof Y. Kurachi (Osaka University, Japan), and rabbit Kir6.2 (GenBank AF006262) were cloned into vector pgfp-ires. Rat SUR2A, supplied by Prof S. Seino (Chiba University, Japan), was expressed in the mammalian vector pcmv6. Mouse SUR2B, supplied by Prof Y. Kurachi, was cloned into the expression vector pcdna3. Electrophysiological recordings were made 48 hours after transfection with solutions identical to those used in rabbit ventricular myocytes (see above). Voltage ramps from 100 to 60 mv were applied over 100 ms every 6 seconds from a holding potential of 80 mv. The current at 0 mv was measured to assay K ATP channel activity. Experiments were performed at room temperature ( 22 C). Simulated Ischemia and Cellular Injury A cell-pelleting model of ischemia modified from Vander Heide et al 4 was used to quantify myocyte injury. In brief, adult rabbit ventricular cells were washed with incubation buffer: (in mmol/l) NaCl 119, NaHCO 3 25, KH 2 PO 4 1.2, KCl 4.8, MgSO 4 1.2, CaCl 2 1, HEPES 10, glucose 11, and taurine 58.5, supplemented with 1% BME amino acids and 1% MEM nonessential amino acids (ph 7.4 with NaOH). Aliquots (0.5 ml) of suspended cells were placed into a microcentrifuge tube and centrifuged for 10 seconds. Approximately 0.25 ml of excess supernatant was removed to leave a thin fluid layer above the pellet, and 0.2 ml of mineral oil was layered on the top to prevent gaseous diffusion. After 60 minutes or 120 minutes, 5 L of cell pellet was sampled through the oil layer and mixed with 75 L of 85 mosm hypotonic staining solution: (in mmol/l) NaHCO , KH 2 PO 4 0.4, KCl 2.7, MgSO 4 0.8, and CaCl 2 1, with 0.5% glutaraldehyde and 0.5% trypan blue. Cells permeable to trypan blue were counted and expressed as a percentage of the total cells counted ( 300 for each sample). In the control group, cells were pelleted and sampled at 60 or 120 minutes. For the diazoxide-treated or P-1075 treated groups, diazoxide (100 mol/l) or P-1075 (30 mol/l) was added to the solution 15 minutes before the pelleting. Cells treated with diazoxide in the presence of 500 mol/l 5HD or in the presence of 30 mol/l HMR1098 were likewise pelleted and sampled at 60 minutes. Once applied, drugs were not washed out and thus were present throughout the simulated ischemia. For the IPC group, the cells pelleted were incubated under oil for 10 minutes and then removed from beneath the oil with a pipette and resuspended in fresh buffer for 30 minutes. Subsequently, the cells were pelleted again and subjected to a prolonged period of simulated ischemia. In the control group, cells were subjected only to the Figure 1. Effect of HMR1098 on DNP-activated surfacek ATP current. A, Time course of current measured at 0 mv. Lines indicate periods when cell was exposed to DNP (100 mol/l) and HMR1098 (30 mol/l). B, Summarized data from 5 cells. *P 0.05 vs current before HMR1098. CONT indicates control group. prolonged period of simulated ischemia without IPC. For the 5HD-treated and HMR1098-treated groups, either 5HD (500 mol/l) or HMR1098 (30 mol/l) was added to the incubation buffer 10 minutes before the IPC. All 4 conditions were tested simultaneously in each of 6 replications. The small percentage of cells ( 18%) that were nonviable at the beginning of the experiment were mostly rounded and had been damaged as a consequence of the enzymatic isolation process. The osmotic fragility of cells induced by ischemia was quantified as percentage of the vital cells at the beginning of each experiment. In nonpelleted control cells suspended in oxygenated buffer with or without drugs, there was no change in the percentage of stained cells measured after 120 minutes of incubation. Pelleting experiments were performed at 37 C. Chemicals Diazoxide and DNP were purchased from Sigma Chemical Co. 5HD was purchased from Research Biochemicals International. HMR1098 was a gift from Hoechst Marion Roussel (now Aventis Pharmaceuticals), and P-1075 was a gift from Leo Pharmaceutical Products. Diazoxide and P-1075 were dissolved in DMSO before being added into the experimental solution. The final concentration of DMSO was 0.1%. Statistical Analysis All data are presented as mean SEM, and the number of cells or experiments is shown as n. Statistical analysis was performed with ANOVA combined with the Fisher post hoc test. Values of P 0.05 were considered significant. Results Effect of HMR1098 on SurfaceK ATP and MitoK ATP Channels We first verified the inhibitory effect of HMR1098 on surfacek ATP channels by whole-cell patch clamp. Figure 1A shows surfacek ATP current (I K,ATP ) elicited by exposure to 2,4-dinitrophenol (DNP, 100 mol/l). Although DNP eventually increased I K,ATP, subsequent application of 30 mol/l HMR1098 suppressed I K,ATP. As summarized in Figure 1B, HMR1098 (30 mol/l) inhibited DNP-induced I K,ATP from to na (P 0.05, n 5). The effects of HMR1098 on mitok ATP channels were examined by measuring mitochondrial matrix redox potential. Figure 2A shows the time course of flavoprotein fluorescence in a cell exposed twice to diazoxide, a selective mitok ATP channel opener in heart cells. 8,20 Diazoxide (100 mol/l) induced reversible oxidation of the flavoproteins. A second application of diazoxide in the presence of HMR1098

3 2420 Circulation May 23, 2000 Figure 2. Effect of HMR1098 on diazoxide (DIAZO)-induced flavoprotein oxidation. A, DIAZO (100 mol/l) reversibly oxidized flavoprotein. HMR1098 (30 mol/l) did not affect oxidative effect of DIAZO. Lines indicate periods when cell was exposed to drug. B, Summarized data for diazoxide-induced flavoprotein oxidation measured in absence and presence of HMR1098. (30 mol/l) once again increased the flavoprotein fluorescence, and the degree of oxidation was identical to that achieved during the first exposure to diazoxide. As summarized in Figure 2B, diazoxide (100 mol/l) reversibly increased flavoprotein oxidation to 41 8% of the DNP value (n 4). HMR1098 (30 mol/l) did not alter the effect of diazoxide (42 9% of the DNP value, n 4). To verify that the lack of effect of HMR1098 on diazoxideinduced flavoprotein oxidation did not result from inadequate diffusion of the drug to mitochondria, we measured flavoprotein fluorescence after including HMR1098 (30 mol/l) in the patch pipette. Figure 3A shows that after 10 minutes in the whole-cell configuration, exposure to diazoxide (100 mol/l) still induced flavoprotein oxidation. This effect of diazoxide could be blocked by 5HD (500 mol/l), a specific mitok ATP channel inhibitor. 10,11 Subsequent reapplication of diazoxide in the absence of 5HD once again increased flavoprotein oxidation. Despite the presence of HMR1098 in the pipette, diazoxide increased flavoprotein oxidation to 46 6% of the DNP value (n 5, Figure 3B). This degree of oxidation is comparable to that observed in the absence of HMR1098 (cf Figure 2). The results indicate that HMR1098 selectively inhibits surfacek ATP channels but not mitok ATP channels. Figure 3. Effect of diazoxide (DIAZO) on flavoprotein fluorescence measured in whole-cell patch-clamp configuration. Patch pipette contained 30 mol/l HMR1098. A, DIAZO (100 mol/l) reversibly increased flavoprotein oxidation. Oxidative effect of DIAZO was inhibited by 5HD (500 mol/l). B, Summarized pooled data for DIAZO-induced flavoprotein oxidation when HMR1098 was administered via pipette. Intracellular HMR1098 did not alter oxidative effect of diazoxide. Figure 4. Effects of P-1075 on surfacek ATP current. A, Time course of current measured at 0 mv. Lines indicate periods when cell was exposed to P-1075 (30 mol/l) and HMR1098 (30 mol/l). B, Summarized data from 4 cells. **P 0.01 vs control (CONT) and P-1075 HMR1098. C, Dose-response curve for P-1075 in rabbit ventricular myocytes. Each point constitutes measurements from 4 to 5 cells. D, Dose-dependent effects of P-1075 on Kir6.2 SUR2A (cardiac-type) and Kir6.1 SUR2B K ATP (vascular smooth muscle type) channels expressed in HEK cells. Each point constitutes measurements from 5 to 10 cells. Effect of P-1075 on SurfaceK ATP and MitoK ATP Channels We then examined the effects of the K ATP channel opener P on surfacek ATP and mitok ATP channels. P-1075 is a derivative of the cyanoguanidine K ATP channel agonist pinacidil, which is known to open both mitok ATP and surfacek ATP channels. 8 Figure 4, A and B, shows that P-1075 (30 mol/l) significantly increased I K,ATP (P 0.01, n 4). Subsequent application of 30 mol/l HMR1098, which we have found to be a selective surfacek ATP channel blocker, suppressed I K,ATP completely. The EC 50 for P-1075 to activate I K,ATP in rabbit ventricular myocytes was 13.4 mol/l (Figure 4C), but cardiovascular effects have been described at much lower concentrations. 21 To investigate the molecular basis of P-1075 action, we expressed the cardiac and vascular smooth muscle isoforms of surfacek ATP channels heterologously in human embryonic kidney (HEK) 293 cells. Figure 4D shows dose-response curves for the agonist effect of P-1075 on Kir6.2 SUR2A (cardiac type) 22 and

4 Sato et al Mitochondrial K ATP Channels in Cardioprotection 2421 Figure 5. Effect of P-1075 on mitochondrial matrix oxidation. A, Diazoxide (DIAZO, 100 mol/l) reversibly oxidized flavoproteins, whereas P-1075 (30 mol/l) did not oxidize flavoproteins. Lines indicate periods when cell was exposed to drugs. B, Summarized data for flavoprotein oxidation. DIAZO(1) and DIAZO(2) indicate first and second exposures to diazoxide (100 mol/l), respectively. Kir6.1 SUR2B (vascular smooth muscle type) 23 K ATP channels. In cells expressing the cardiac-type Kir6.2 SUR2A channels, P-1075 effectively increased I K,ATP in a concentration-dependent manner (EC 50 for P mol/l). Conversely, P-1075 activated Kir6.1 SUR2B K ATP channels at nanomolar concentrations (EC nmol/l). Thus, the low-dose effects previously described 21 are unlikely to reflect activation of cardiac surface K ATP channels. The effects of P-1075 on mitok ATP channels were examined by measurement of mitochondrial redox potential. Figure 5 shows that diazoxide (100 mol/l) induced reversible oxidation of the mitochondrial matrix to 44 4% of the DNP value (n 5). Subsequent exposure to P-1075 (30 mol/l) had no effect (2 1% of the DNP value, n 5), whereas diazoxide once again increased flavoprotein oxidation (39 5% of the DNP value, n 4). Even when very high (100 mol/l) or low (100 nmol/l) concentrations of P-1075 were applied, the drug failed to elicit any flavoprotein response (not shown). These results indicate that P-1075 selectively activates surfacek ATP channels without affecting mitok ATP channels. To verify further the specificity of diazoxide and P-1075 for mitok ATP and surfacek ATP channels, respectively, we measured flavoprotein fluorescence and membrane current simultaneously. Figure 6A and 6B shows the effects of diazoxide and P-1075 in a representative experiment. Diazoxide (100 mol/l) induced reversible oxidation of flavoproteins but did not affect I K,ATP. In contrast, exposure to P-1075 (30 mol/l) failed to increase flavoprotein oxidation but did elicit I K,ATP. As summarized in Figure 6C and 6D, unlike diazoxide, P-1075 activated only I K,ATP. Effects of HMR1098 and P-1075 on Simulated Ischemia and Cellular Injury Using the mitochondria- or surface-selective agents, we examined the role of mitok ATP and surfacek ATP channels for ischemic cardioprotection. The mitok ATP channel opener diazoxide (100 mol/l) significantly decreased the percentage of cells stained after 60 minutes of simulated ischemia (from 32 3% to 17 3%, P 0.001, n 5), and this protection was completely prevented by the mitok ATP channel blocker 5HD (500 mol/l) (Figure 7A). In contrast, the surfacek ATP channel blocker HMR1098 (30 mol/l) did not prevent the cardioprotection by diazoxide (from 38 4% to Figure 6. Effects of diazoxide and P-1075 on flavoprotein fluorescence and I K,ATP. A and B, Simultaneous measurements of flavoprotein fluorescence and I K,ATP. Diazoxide (DIAZO, 100 mol/l) reversibly oxidized flavoprotein without affecting membrane current. P-1075 increased I K,ATP but not flavoprotein fluorescence. C and D, Summarized data for flavoprotein oxidation and I K,ATP from 5 cells. 18 1%, P 0.001, n 4) (Figure 7B). In a separate series of experiments (Figure 7C), simulated ischemia for 60 and 120 minutes stained 35 2% (n 5) and 42 2% (n 5) of cells, respectively. Inclusion of diazoxide (100 mol/l) significantly decreased the percentage of cells stained, to 18 2% Figure 7. Summarized data for percentage of cells stained after 60 minutes or 120 minutes of simulated ischemia. Cells stained after simulated ischemia were plotted as a percentage of total viable cells before ischemia. A, Effect of 5HD on diazoxide-induced cardioprotection. CONT indicates control group; DIAZO, diazoxide (100 mol/l)-treated group; and DIAZO 5HD, diazoxide plus 5HD (500 mol/l) treated group. B, Effects of HMR1098 on diazoxideinduced cardioprotection. DIAZO HMR1098 indicates diazoxide plus HMR1098 (30 mol/l) treated group. C, Comparative effects of diazoxide and P-1075 on ischemic cardioprotection. P1075 indicates P-1075 (30 mol/l) treated group. Data obtained from parallel experiments are summarized in each panel. **P vs CONT group.

5 2422 Circulation May 23, 2000 Figure 8. Effects of 5HD and HMR1098 on ischemically preconditioned cells. Cells stained after 60 and 120 minutes of simulated ischemia were plotted as percentage of total viable cells before ischemia. CONT indicates nonpreconditioned control group; IPC, preconditioned group; IPC 5HD, 5HD (500 mol/l) added 10 minutes before preconditioning; and IPC HMR1098, HMR1098 (30 mol/l) added 10 minutes before preconditioning. **P vs CONT group. (n 5) after 60 minutes and 24 2% (n 5) after 120 minutes of simulated ischemia (P versus control group). In contrast, the selective surfacek ATP channel opener P-1075 (30 mol/l) did not alter the extent of stained cells as a consequence of ischemia (34 4% after 60 minutes, 38 4% after 120 minutes). These results indicate that mitok ATP but not surfacek ATP channels are involved in pharmacological cardioprotection. We examined the effects of 5HD and HMR1098 on simulated IPC. As summarized in Figure 8, IPC significantly decreased the percentage of cells stained during 60 minutes (from 26 4% to 13 2%, P 0.005, n 6) and 120 minutes (from 35 3% to 20 3%, P 0.005, n 6) of simulated ischemia. 5HD (500 mol/l) added 10 minutes before preconditioning abolished the protection (25 3% after 60 minutes and 34 3% after 120 minutes ischemia, respectively). In contrast, HMR1098 (30 mol/l) did not interfere with IPC after 60 minutes (15 2%, P 0.005, n 6) or 120 minutes (20 3%, P 0.005, n 6) of simulated ischemia. These results indicate that the cardioprotection afforded by IPC is mediated by mitok ATP channels, not surfacek ATP channels. Discussion The cardioprotective effects of IPC are abolished by antagonists of K ATP channels and are mimicked by agonists of K ATP channels On the basis of these studies, an early hypothesis proposed that opening of surfacek ATP channels abbreviates action potential duration, thereby reducing cellular calcium overload and preserving viability in ischemic myocardium. However, this hypothesis cannot account for IPC, because abbreviation of action potential duration is not necessary for protection. Grover et al 5 showed that dofetilide attenuated the shortening of action potential duration but did not abolish IPC in dogs. Yao and Gross 6 demonstrated that low-dose bimakalim produced cardioprotection without any effect on monophasic action potential duration. Furthermore, IPC occurs even in unstimulated (electrically quiescent) cardiac myocytes, in which action potential abbreviation cannot be a factor. 4,7 MitoK ATP channels share some pharmacological properties with surfacek ATP channels, while possessing a distinct pharmacological profile. Garlid et al 20 demonstrated that diazoxide opens mitok ATP channels 2000-fold more potently than surfacek ATP channels in cardiac myocytes. Although direct effects of diazoxide on mitochondrial energy metabolism in pancreatic -cells have been proposed, 27 previous studies in our laboratory demonstrated that diazoxide oxidizes the mitochondrial matrix redox potential via opening of mitok ATP channels in rabbit hearts, whereas surfacek ATP channels are quite resistant to this drug 8 ; the specificity of diazoxide for mitok ATP channels is verified in Figure 6 of the present article. Moreover, we reported that 5HD completely and reversibly blocks the oxidative effect of diazoxide without affecting surfacek ATP channels, indicating that 5HD is an effective and specific blocker of mitok ATP channels 8,10,11 (although this may not be the case in other species 28,29 ). When these drugs are used as pharmacological tools to activate or to inhibit mitok ATP channels, growing evidence supports the idea that mitok ATP channels rather than surfacek ATP channels are the dominant players for cardioprotection. 8,9 Moreover, in animal models in vivo, diazoxide mimics, 30 whereas 5HD abolishes, the infarct size limiting effect of IPC These findings motivated us to reevaluate the role of surfacek ATP channels in cardioprotection. The sulfonylthiourea HMR1098 has been reported to be a cardioselective K ATP channel blocker, blocking K ATP channels in cardiac muscle cells with 10- to 50-fold higher potency than in pancreatic -cells with little effect on the coronary vasculature. 15 We demonstrated that HMR1098 inhibited surfacek ATP channels activated by metabolic inhibition (Figure 1) and by the surfacek ATP channel opener P-1075 (Figure 4). However, HMR1098 did not affect diazoxide-induced flavoprotein oxidation (Figure 2). Furthermore, direct application of HMR1098 to the cytoplasm by inclusion in the pipette failed to prevent the oxidizing effect of diazoxide (Figure 3). These results, taken together, indicate that HMR1098 selectively inhibits surfacek ATP channels without affecting mitok ATP channels. The K ATP channel opener P-1075 activated I K,ATP in rabbit ventricular myocytes; I K,ATP was blocked by HMR1098 (Figure 4). Moreover, the molecularly defined cardiac surfacek ATP channel (Kir6.2 SUR2A) 22 expressed in HEK cells was effectively activated by P Indeed, P-1075 showed similar potencies in activating native surfacek ATP channels and expressed Kir6.2 SUR2A channels (EC 50 sof13 mol/l and 2.5 mol/l, respectively). P-1075 activated Kir6.1 SUR2B (vascular smooth muscle type) 23 channels 100-fold more potently than Kir6.2 SUR2A channels (EC 50 of 102 nmol/l). Conversely, P-1075 did not affect mitochondrial oxidation state measured with or without invasion by patch pipettes or their contents (Figures 5 and 6). These results indicate that, unlike diazoxide, P-1075 selectively activates surfacek ATP but not mitok ATP channels. We previously reported that diazoxide (50 mol/l) decreased myocyte death in a cellular model of simulated ischemia. 8 In the present study, we verified the cardioprotective effects of diazoxide (Figure 7). Diazoxide-induced cardioprotection was prevented by the mitok ATP channel blocker 5HD. In contrast, the surfacek ATP channel blocker HMR1098 did not abolish the cardioprotective effect of diazoxide (Figure 7B). Furthermore, we found that the surfacek ATP channel opener P-1075 did not

6 Sato et al Mitochondrial K ATP Channels in Cardioprotection 2423 produce cardioprotection (Figure 7C). Sargent et al 21 reported that, in Langendorff-perfused rat hearts, P-1075 increases coronary flow and protects ischemic myocardium at nanomolar concentrations. Consistent with their findings, P-1075 does activate the vascular smooth muscle type (Kir6.1 SUR2B) K ATP channel at nanomolar concentrations, whereas the cardiac-type (Kir6.2 SUR2A) K ATP channel is affected only in the micromolar range (Figure 4D). However, we found that the surfacek- ATP channel opener P-1075 (30 mol/l) did not produce cardioprotection (Figure 7C). Although the reason for this discrepancy is unknown, our results imply that the direct activation of surfacek ATP channels in rabbit myocytes does not underlie myocyte cardioprotection. IPC is present in all species examined, including humans. 31 Compelling evidence suggests that mitok ATP channels rather than surfacek ATP channels may serve as end effectors of IPC. However, another study reported that digoxin preserves the subsarcolemmal ATP and inhibits surfacek ATP channels, thereby abolishing IPC in rabbit hearts. 32 To clarify these discordant data, further studies using the selective agonist and antagonist of surfacek ATP or mitok ATP channels are necessary. Our present results demonstrate that the mitok ATP channel blocker 5HD but not the surfacek ATP channel blocker HMR1098 abolished the cardioprotection in a cellular model of preconditioning. These results divorce the cardioprotective effect of IPC from activation of surfacek ATP channels. In conclusion, our observations with diazoxide, P-1075, 5HD, and HMR1098 provide the first definitive pharmacological evidence that surfacek ATP channels are not involved in cardioprotection in isolated rabbit myocytes. By extension, the present data provide further support for the emerging consensus that mitok ATP channels are the end effectors of preconditioning specifically in rabbit heart. Acknowledgments This study was supported by the National Institutes of Health (R37-HL to Dr Marbán) and by a Banyu Fellowship in Lipid Metabolism and Atherosclerosis (to Dr Sato). Dr Marbán holds the Michel Mirowski, MD, Professorship of Cardiology of the Johns Hopkins University. References 1. Murry CE, Jennings RB, Reimer KA. Preconditioning with ischemia: a delay of lethal cell injury in ischemic myocardium. Circulation. 1986; 74: Noma A. ATP-regulated K channels in cardiac muscle. Nature. 1983; 305: Inoue I, Nagase H, Kishi K, Higuti T. ATP-sensitive K channel in the mitochondrial inner membrane. Nature. 1991;352: Vander Heide RS, Rim D, Hohl CM, Ganote CE. An in vitro model of myocardial ischemia utilizing isolated adult rat myocytes. J Mol Cell Cardiol. 1990;22: Grover GJ, Dalonzo AJ, Dzwonczyk S, Parham CS, Darbenzio RB. Preconditioning is not abolished by the delayed rectifier K blocker dofetilide. Am J Physiol. 1996;40:H1207 H Yao Z, Gross GJ. Effects of the K ATP channel opener bimakalim on coronary blood flow, monophasic action potential duration, and infarct size in dogs. Circulation. 1994;89: Liang BT. Direct preconditioning of cardiac ventricular myocytes via adenosine A 1 receptor and K ATP channel. Am J Physiol. 1996;40:H1769 H Liu Y, Sato T, O Rourke B, Marban E. Mitochondrial ATP-dependent potassium channels: novel effectors of cardioprotection? Circulation. 1998;97: Garlid KD, Paucek P, Yarov-Yarovoy V, Murray HN, Darbenzio RB, D Alonzo AJ, Smith MA, Grover GJ. Cardioprotective effect of diazoxide and its interaction with mitochondrial ATP-sensitive K channels: possible mechanism of cardioprotection. Circ Res. 1997;81: Sato T, O Rourke B, Marbán E. Modulation of mitochondrial ATPdependent K channels by protein kinase C. Circ Res. 1998;83: Hu H, Sato T, Seharaseyon J, Liu Y, Johns DC, O Rourke B, Marbán E. Pharmacological and histochemical distinctions between molecularly defined sarcolemmal K ATP channels and native cardiac mitochondrial K ATP channels. Mol Pharmacol. 1999;55: Auchampach JA, Grover GJ, Gross GJ. Blockade of ischaemic preconditioning in dogs by the novel ATP dependent potassium channel antagonist sodium 5-hydroxydecanoate. Cardiovasc Res. 1992;26: Hide EJ, Thiemermann C. Limitation of myocardial infarct size in the rabbit by ischaemic preconditioning is abolished by sodium 5-hydroxydecanoate. Cardiovasc Res. 1996;31: Schultz JJ, Qian YZ, Gross GJ, Kukreja RC. The ischemia-selective K ATP channel antagonist, 5-hydroxydecanoate, blocks ischemic preconditioning in the rat heart. J Mol Cell Cardiol. 1997;29: Gögelein H, Hartung J, Englert HC, Schölkens BA. HMR 1883, a novel cardioselective inhibitor of the ATP-sensitive potassium channel, I: effects on cardiomyocytes, coronary flow and pancreatic -cells. J Pharmacol Exp Ther. 1998;286: Xu X, Tsai TD, Lee KS. A specific activator of the ATP-inhibited K channels in guinea pig ventricular cells. J Pharmacol Exp Ther. 1993;266: Liu Y, Gao WD, O Rourke B, Marban E. Synergistic modulation of ATP-dependent K currents by protein kinase C and adenosine: implications for ischemic preconditioning. Circ Res. 1996;78: Chance B, Salkovitz IA, Kovach AG. Kinetics of mitochondrial flavoprotein and pyridine nucleotide in perfused heart. Am J Physiol. 1972; 223: Hajnoczky G, Robb-Gaspers LD, Seitz MB, Thomas AP. Decoding of cytosolic calcium oscillations in the mitochondria. Cell. 1995;82: Garlid KD, Paucek P, Yarov-Yarovoy V, Sun X, Schindler PA. The mitochondrial K ATP channel as a receptor for potassium channel openers. J Biol Chem. 1996;271: Sargent CA, Sleph PG, Dzwonczyk S, Normandin D, Antonaccio MJ, Crover GJ. Cardioprotective effects of the cyanoguanidine potassium channel opener P J Cardiovasc Pharmacol. 1993;22: Inagaki N, Gonoi T, Clement JP IV, Wang CZ, Aguilar-Bryan L, Bryan J, Seino S. A family of sulfonylurea receptors determines the pharmacological properties of ATP-sensitive K channels. Neuron. 1996;16: Yamada M, Isomoto S, Matsumoto S, Kondo C, Shindo T, Horio Y, Kurachi Y. Sulfonylurea receptor 2B and Kir6.1 form a sulfonylurea-sensitive but ATP-insensitive K channel. J Physiol (Lond). 1997;499: Gross GJ, Auchampach JA. Blockade of ATP-sensitive potassium channels prevents myocardial preconditioning in dogs. Circ Res. 1992; 70: Armstrong SC, Liu GS, Downey JM, Ganote CE. Potassium channels and preconditioning of isolated rabbit cardiomyocytes: effects of glyburide and pinacidil. J Mol Cell Cardiol. 1995;27: Mizumura T, Nithipatikom K, Gross GJ. Bimakalim, an ATP-sensitive potassium channel opener, mimics the effects of ischemic preconditioning to reduce infarct size, adenosine release, and neutrophil function in dogs. Circulation. 1995;92: Grimmsmann T, Rustenbeck I. Direct effects of diazoxide on mitochondria in pancreatic -cells and on isolated liver mitochondria. Br J Pharmacol. 1998;123: Moritani K, Miyazaki T, Miyoshi S, Asanagi M, Zhao LS, Mitamura H, Ogawa S. Blockade of ATP-sensitive potassium channels by 5-hydroxydecanoate suppresses monophasic action potential shortening during regional ischemia. Cardiovasc Drugs Ther. 1994;8: Sakamoto K, Yamazaki J, Nagao T. 5-Hydroxydecanoate selectively reduces the initial increase in extracellular K in ischemic guinea-pig heart. Eur J Pharmacol. 1998;348: Baines CP, Liu GS, Birincioglu M, Critz SD, Cohen MV, Downey JM. Ischemic preconditioning depends on interaction between mitochondrial K ATP channels and actin cytoskeleton. Am J Physiol. 1999;276: H1361 H Cohen MV, Downey JM. Ischemic preconditioning: can the protection be bottled? Lancet. 1993;342: Haruna T, Horie M, Kouchi I, Nawada R, Tsuchiya K, Akao M, Otani H, Murakami T, Sasayama S. Coordinate interaction between ATP-sensitive K channel and Na,K -ATPase modulates ischemic preconditioning. Circulation. 1998;98: