Cardioprotective Effect of Diazoxide and Its Interaction With Mitochondrial ATP-Sensitive K+ Channels
Possible Mechanism of Cardioprotection
Abstract Previous studies showed a poor correlation between sarcolemmal K+ currents and cardioprotection for ATP-sensitive K+ channel (KATP) openers. Diazoxide is a weak cardiac sarcolemmal KATP opener, but it is a potent opener of mitochondrial KATP, making it a useful tool for determining the importance of this mitochondrial site. In reconstituted bovine heart KATP, diazoxide opened mitochondrial KATP with a K1/2 of 0.8 μmol/L while being 1000-fold less potent at opening sarcolemmal KATP. To compare cardioprotective potency, diazoxide or cromakalim was given to isolated rat hearts subjected to 25 minutes of global ischemia and 30 minutes of reperfusion. Diazoxide and cromakalim increased the time to onset of contracture with a similar potency (EC25, 11.0 and 8.8 μmol/L, respectively) and improved postischemic functional recovery in a glibenclamide (glyburide)-reversible manner. In addition, sodium 5-hydroxydecanoic acid completely abolished the protective effect of diazoxide. Whole-myocyte studies showed that diazoxide was significantly less potent than cromakalim in increasing sarcolemmal K+ currents. Diazoxide shortened ischemic action potential duration significantly less than cromakalim at equicardioprotective concentrations. We also determined the effects of cromakalim and diazoxide on reconstituted rat mitochondrial cardiac KATP activity. Cromakalim and diazoxide were both potent activators of K+ flux in this preparation (K1/2 values, 1.1±0.1 and 0.49±0.05 μmol/L, respectively). Both glibenclamide and sodium 5-hydroxydecanoic acid inhibited K+ flux through the diazoxide-opened mitochondrial KATP. The profile of activity of diazoxide (and perhaps KATP openers in general) suggests that they protect ischemic hearts in a manner that is consistent with an interaction with mitochondrial KATP.
The KATP openers are a structurally diverse class of agents that, among other activities, protect ischemic myocardial tissue.1,2 These protective effects are abolished by KATP blockers such as glibenclamide (glyburide) and 5-HD. It was originally thought that KATP openers protected ischemic myocardium by shortening the APD, thereby producing a “cardioplegic” effect. Although ATP is conserved during ischemia with these agents,3,4 this protective action is not accompanied by significant cardiodepression,2 which is not consistent with a cardioplegic effect. Recent studies by Yao and Gross5 showed cardioprotection in dogs with a dose of bimakalim that did not shorten monophasic APD. We found a similar lack of correlation between monophasic APD and cardioprotection for cromakalim.6 Pyranyl cyanoguanidine analogues that retain glibenclamide-reversible cardioprotective effects have been discovered, but they are relatively devoid of APD shortening effects or the ability to enhance whole-myocyte K+ currents.7,8 These agents are also weak vasodilators, suggesting this activity to be unimportant for cardioprotection.
These data suggest the possibility of a site of action that is distinct from sarcolemmal KATP. KATP are known to be expressed in mitochondrial membranes, where they function in the control of mitochondrial volume and energetics.9,10 Interestingly, benzopyran KATP openers have been shown to open mitochondrial KATP within their cardioprotective concentration range, suggesting the possibility that mitochondrial KATP is the site of cardioprotective action.11,12
Diazoxide is an potent opener of KATP in pancreatic cells, but compared with other KATP openers, diazoxide has relatively little effect on cardiac APD.13 Recent studies from Garlid’s laboratory12 showed that diazoxide opened bovine cardiac sarcolemmal KATP in the high micromolar range (K1/2, 855 μmol/L), whereas it opened mitochondrial KATP within a submicromolar range (K1/2, 0.4 μmol/L), making this agent a potentially useful tool for separating the activities of the two KATP.12 Conversely, cromakalim was a potent opener of both sarcolemmal and mitochondrial channels. The first goal of the present study was to determine whether diazoxide retains the cardioprotective activity of cromakalim, thereby implicating a role for mitochondrial KATP. With this in mind, we determined the cardioprotective potency of diazoxide (which has not previously been determined) in isolated rat hearts and compared this activity with its ability to open rat cardiac mitochondrial KATP. A second goal of the present study was to determine the effect of 5-HD, which has little effect on sarcolemmal KATP,14 on the cardioprotective action of diazoxide or on the mitochondrial KATP opening activity of diazoxide.
Materials and Methods
Isolated Rat Heart Model of Ischemia and Reperfusion
Male Sprague-Dawley rats (400 to 500 g) were anesthetized using 100 mg/kg IP sodium pentobarbital. The trachea was intubated, and then the jugular vein was injected with heparin (1000 U/kg). While the rats were being mechanically ventilated, their hearts were perfused in situ via retrograde cannulation of the aorta. The hearts were then excised and quickly moved to a Langendorff apparatus, where they were perfused with oxygenated Krebs-Henseleit solution containing (mmol/L) NaCl 112, NaHCO3 25, KCl 5, MgSO4 1.2, KH2PO4 1, CaCl2 1.2, glucose 11.5, and pyruvate 2 at a constant perfusion pressure (85 mm Hg). A water-filled latex balloon attached to a metal cannula was then inserted into the left ventricle and connected to a Statham pressure transducer for measurement of left ventricular pressure. The hearts were allowed to equilibrate for 15 minutes, at which time EDP was adjusted to 5 mm Hg and this balloon volume was maintained for the duration of the experiment. Preischemia or predrug function, heart rate, and coronary flow (extracorporeal electromagnetic flow probe, Carolina Medical Electronics) were then measured. Contractile function was calculated by subtracting EDP from left ventricular peak systolic pressure, resulting in LVDP. Cardiac temperature was maintained throughout the experiment by submerging the hearts in 37°C buffer, which was allowed to accumulate in a stoppered heated chamber.
After equilibration, the hearts were subjected to one of several treatments. Hearts were treated with vehicle (0.04% DMSO, n=7), 1 to 100 μmol/L cromakalim (n=6 per group), or 1 to 100 μmol/L diazoxide (n=6 to 8 per group). The respective drug treatments were given for 10 minutes and were included in the perfusate. At this time, the hearts were subjected to 25 minutes of global ischemia and 30 minutes of reperfusion. Ischemia was initiated by completely shutting off perfusate flow. At the end of the reperfusion period, contractile function, coronary flow, and LDH release were measured. The respective drugs were given only before global ischemia and were not given during reperfusion. Severity of ischemic/reperfusion damage was determined from the time to the onset of contracture during global ischemia, recovery of contractile function at 30 minutes into reperfusion, and cumulative LDH release into the reperfusate. The fluid in the organ chamber originates from the coronary effluent; therefore, the effluent was collected during reperfusion to calculate LDH release. Time to contracture was defined as the time (minutes) during global ischemia in which the first 5 mm Hg increase in EDP was observed, and these data were used to calculate cardioprotective potency because of the concentration-dependent profile of this index. Cardioprotective potency was expressed as EC25, which is the concentration causing a 25% increase in time to onset of contracture relative to vehicle-treated hearts and is a reliable index of potency for pharmacological comparison of drugs.3
Another study was done to determine whether the preischemic and postischemic effects of diazoxide are abolished by the KATP blocker glibenclamide. For this study, rat hearts were isolated and prepared as described above. They were pretreated with vehicle (0.04% DMSO, n=5), 30 μmol/L diazoxide (n=5), or 30 μmol/L diazoxide+0.3 μmol/L glibenclamide (n=5), 10 μmol/L cromakalim (n=5), or 10 μmol/L cromakalim+0.3 μmol/L glibenclamide (n=5) for 10 minutes, and the hearts were rendered globally ischemic for 25 minutes and reperfused for 30 minutes. Functional and biochemical end points were determined as described above. The concentrations of diazoxide and cromakalim were chosen to be approximately equivalent in terms of cardioprotective activity.
We also determined the effect of 5-HD on the cardioprotective activity of diazoxide. Rat hearts were pretreated with vehicle (0.04% DMSO, n=5), 30 μmol/L diazoxide (n=5), or 100 μmol/L 5-HD+30 μmol/L diazoxide (n=5). Drug treatment lasted for 10 minutes, after which the hearts were rendered globally ischemic for 25 minutes and reperfused for 30 minutes. Functional and biochemical end points were measured as described above. These studies are important because of the observation that 5-HD abolishes the protective effects of KATP openers while having no effect on their APD shortening activity, showing a lack of blocking activity on cardiac sarcolemmal channels.14
Although no work has been done on the effect of diazoxide on mitochondrial KATP in species other than rats and cattle, it would be prudent to show that the protective effects of diazoxide can be seen in other species. We determined the effect of diazoxide on necrosis and recovery of function in isolated rabbit hearts subjected to global ischemia and reperfusion. Rabbit hearts were pretreated for 10 minutes with vehicle (0.04% DMSO, n=13), 30 μmol/L cromakalim (n=13), 30 μmol/L diazoxide (n=13), or 30 μmol/L diazoxide+100 μmol/L 5-HD (n=13). The hearts were then subjected to 50 minutes of total global ischemia followed by 30 minutes of reperfusion. Time to contracture, postischemic recovery of contractile function, and LDH release were measured as described above. Rabbit hearts are more resistant to global ischemia compared with rat hearts; therefore, longer periods of ischemia are required to achieve similar degrees of ischemic damage.
Electrophysiological Recording: Whole-Cell K+ Currents
In addition to cardioprotection studies, the relative effect of cromakalim and diazoxide on K+ currents in rat ventricular myocytes was studied to determine whether a dissociation between cardioprotection and sarcolemmal currents could be observed. Rat ventricular myocytes were dissociated using previously described methods.15 Currents were recorded using the whole-cell configuration of the patch-clamp technique.16 Electrodes (2 to 5 MΩ) were fabricated from borosilicate glass. Voltage-clamp protocols were generated, and data were acquired using Pulse software (HEKA) in conjunction with a HEKA EPC9 amplifier (Lambrecht). Voltage ramps from −100 to 40 mV were applied from a holding potential of −40 mV (to inactivate Na+ current). Resultant currents were filtered at 3 KHz using an analog four-pole Bessel filter. The bath solution contained (mmol/L) NaCl 140, KCl 4, MgCl2 1, CaCl2 2, glucose 10, and HEPES 5, pH 7.4. Nisoldipine (1 μmol/L) was added to the bath solution to inhibit L-type Ca2+ current. The pipette solution contained (mmol/L) KCl 125, MgCl2 2, CaCl2 2, NaCl 10, EGTA 10, HEPES 5, glucose 10, and ATP 1, pH 7.2 with KOH. All experiments were performed at 33°C to 34.5°C. Stock solutions of diazoxide and cromakalim (0.1 mol/L) were prepared using DMSO and then diluted in bath solution as required.
Isolated Rat Heart Model for APD Determination
Whereas a separation between cromakalim and diazoxide was observed for whole-myocyte K+ currents, it was also important to determine the effects of these agents on K+ currents in ischemic tissue. To accomplish this, rats were anesthetized, and their hearts removed and perfused as described above. The hearts were instrumented to record the ECG and epicardial MAPs (Franz epicardial Langendorff probe, EP Technologies) throughout the experiment. ECG and MAP signals were routed to a chart recorder (TA4000, Gould) and an oscilloscope (DL1200, Yokogawa). The ambient temperature around the preparation was maintained by a heated vessel (37±0.2°C, FE 2, Haake). A cannula was placed inside the right atria to allow superfusion of the tissue during subsequent global ischemia in an effort to maintain electrical activity. Platinum leads were also placed on the heart to pace the ventricles once the atrial superfusion failed to afford adequate rate control. When atrial superfusion failed to maintain proper sinoatrial nodal activity, ventricular pacing was initiated at 5 Hz.
A total of 33 hearts were used for the present studies. Each heart was given 10 minutes of equilibration time. After equilibration, ECG and action potentials were recorded on chart paper run at 200 mm/s, and coronary flow was noted. Hearts were then given vehicle (0.1% DMSO, n=13), cromakalim (10 μmol/L, n=10), or diazoxide (30 μmol/L, n=10). At the end of 10 minutes of respective compound administration, data were again collected. The hearts were then rendered globally ischemic by stopping flow through the perfusion cannula. Atrial superfusion was immediately initiated with drug-free buffer (37°C) at a flow rate of 10 to 20 mL/min to maintain intrinsic heart rate. Data were collected every minute or until (1) tachyarrhythmias were initiated, which precluded data collection, (2) hearts were no longer able to be ventricularly paced, or (3) action potentials became of extremely low amplitude or were sufficiently deformed and irregularly shaped to preclude a valid measurement.
Effect of Diazoxide on Rat Cardiac Mitochondrial KATP
Purification of Inner Membrane Vesicles From Rat Heart Mitochondria
Rat heart mitochondria were prepared according to Matlib et al17 and inner membrane vesicles were prepared as described by McEnery et al.18 In brief, mitochondria were suspended at 20 mg protein/mL in 220 mmol/L d-mannitol, 70 mmol/L sucrose, 0.5 mg/mL bovine serum albumin, and 20 mmol/L potassium HEPES, pH 7.4. The suspension was sonicated for 20 seconds at 55 W and then cooled on ice for 2 minutes. After eight sonication/cooling cycles, the suspension was centrifuged at 10 000g for 15 minutes at 4°C, and the supernatant was centrifuged at 120 000g for 30 minutes at 4°C. The resulting pellet was washed in 150 mmol/L potassium phosphate, 25 mmol/L potassium EDTA, pH 7.9, 1 mmol/L ATP, 0.5 mmol/L dithiothreitol, and 5% ethylene glycol (PA buffer). Vesicles were incubated in PA buffer containing 3 mol/L guanidine-HCl to remove F1-ATPase and other membrane-bound proteins. Treated vesicles were centrifuged at 140 000g for 30 minutes at 4°C, and the pellet was resuspended in PA buffer. This wash cycle was repeated twice. The final membrane pellet was resuspended to 10 mg protein/mL in 250 mmol/L sucrose, 1 mmol/L TEA+-EDTA, and 50 mmol/L Tris-HCl, pH 7.2, and stored at −70°C until needed.
Purification, Reconstitution, and Assay of Rat Heart Mitochondrial KATP
Purification and reconstitution followed protocols described previously.19,20 Inner mitochondrial membrane proteins were solubilized in 3% Triton X-100, 20% glycerol, 0.1% β-mercaptoethanol, 0.2 mmol/L TEA+-EGTA, 1 mmol/L MgCl2, and 50 mmol/L Tris-HCl, pH 7.2. After incubation for 60 minutes at 4°C, the mixture was centrifuged at 120 000g for 30 minutes. One milliliter of the supernatant containing 2 mg of extracted proteins was loaded onto a 1-mL DEAE-cellulose column that had been equilibrated with column buffer containing 1% Triton X-100, 0.1% β-mercaptoethanol, 1 mmol/L TEA+-EDTA, and 50 mmol/L Tris-HCl, pH 7.2. The active fraction was eluted at 300 mmol/L KCl, was desalted by passing it through a Sephadex G-250 to 300 column (1:20 [vol/vol] ratio), and was concentrated by filtration.
The purified mitochondrial KATP fraction was added to a 10:1 mixture of l-α-lecithin (Avanti) and cardiolipin in a buffer containing 10% octylpentaoxyethylene, 300 μmol/L PBFI, 100 mmol/L TEA+-SO4, 0.14 mmol/L KCl, 1 mmol/L TEA+-EDTA, and 25 mmol/L TEA+-HEPES, pH 6.8. This mixture was loaded onto a Bio-Beads SM-2 column (Bio-Rad) (1:4 [vol/vol] ratio), incubated for 40 minutes at 4°C, and then centrifuged at 800g for 2 minutes. Eluted sample was loaded onto a new Bio-Beads SM-2 column, incubated, and centrifuged as described before, and an eluted sample was passed twice through Sephadex G-25–300 columns (1:10 [vol/vol] ratio). The resulting stock of proteoliposomes was stored on ice during the experiment.
Stock proteoliposomes (15 μL) were added to 2 mL of external medium containing 150 mmol/L KCl, 1 mmol/L TEA+-EDTA, and 25 mmol/L TEA+-HEPES, pH 7.4. In the assay, electrophoretic K+ flux was initiated by 1 μmol/L FCCP to provide charge compensation via H+ flux. K+ flux into proteoliposomes was determined by linear regression of the initial rate of K+ uptake, and K+-dependent fluorescence of intraliposomal PBFI was calibrated for each preparation.9,19
Assays of K+ Flux in Proteoliposomes Containing Reconstituted KATP Isolated From Bovine Heart Mitochondria and Sarcolemma
The purpose of this part of the study was to confirm the relative mitochondrial selectivity of diazoxide. This was done in bovine hearts because rat hearts provide insufficient material with which to prepare reconstituted sarcolemmal channels. Inner mitochondrial membrane proteins were solubilized in 3% Triton X-100 and fractionated on DEAE cellulose. KATP, which eluted with 250 mmol/L KCl, was reconstituted into proteoliposomes exactly as previously described.9,19 Sarcolemmal KATP was solubilized, purified, and reconstituted into proteoliposomes using the same protocols, except that its activity was found in the 100 mmol/L KCl fraction.20 Internal medium contained 300 μmol/L PBFI, 0.14 mmol/L KCl, 1 mmol/L TEA+-EDTA, 25 mmol/L TEA+-HEPES, and 100 mmol/L TEA+-SO4 (pH 6.8). Vesicles were added (in a final concentration of 0.38 mg lipid/mL) to external medium containing 150 mmol/L KCl, 3 mmol/L MgCl2, 0.5 mmol/L ATP, and 25 mmol/L TEA+-HEPES (pH 7.4) at 25°C. Sarcolemmal vesicles were prepared from the left ventricular muscle of fresh bovine heart according to a modification21 of the method of Jones.22 Internal and external media were as described for mitochondrial KATP, except that choline was substituted for TEA+ ion, which inhibits K+ flux through cell KATP. In the assay, electrophoretic K+ flux was initiated by 1 μmol/L FCCP, which catalyzes proton flux to provide charge compensation. K+ flux was determined by linear regression of the initial changes in the K+-dependent fluorescence of intraliposomal PBFI, which was calibrated for each preparation.
Chemicals and Drugs
PBFI was purchased from Molecular Probes Inc. Cromakalim was synthesized in the Department of Chemistry at Bristol-Myers Squibb. All other compounds were obtained from Sigma Chemical Co.
Differences with respect to time and treatment were discerned using a factorial ANOVA. A Newman-Keuls, Dunnett’s, or Tukey’s post hoc test was used for comparisons. All data are presented as the mean±SEM, and significant differences were determined at the P<.05 level.
Effect of Diazoxide on Severity of Ischemic/Reperfusion Damage in Rat Hearts
The effect of diazoxide on preischemic and postischemic contractile function and coronary flow is shown in Table 1⇓. Diazoxide had little effect on heart rate or LVDP before ischemia. Diazoxide significantly increased coronary flow at 30 and 100 μmol/L. During reperfusion, significant bradycardia and contractile dysfunction were observed in vehicle-treated hearts, suggesting severe ischemic/reperfusion injury. Diazoxide attenuated the reperfusion bradycardia and significantly enhanced the recovery of contractile function in a concentration-dependent manner. Coronary reflow was significantly reduced in vehicle-treated hearts. Reflow was only slightly improved by diazoxide. The effect of diazoxide or cromakalim on the time to onset of contracture is shown in Fig 1⇓. Diazoxide increased the time to contracture in a concentration-dependent manner, with an EC25 of 11.0 μmol/L (n=6 to 8 per group). Cromakalim increased the time to contracture, with an EC25 of 8.8 μmol/L (n=6 or 7 per group), which is similar in cardioprotective potency to diazoxide. The effect of diazoxide on reperfusion EDP and LDH release is shown in Fig 2⇓. Reperfusion EDP and LDH release were elevated in vehicle-treated hearts. Diazoxide significantly reduced reperfusion EDP and LDH starting at 1 and 3 μmol/L, respectively, and they were reduced in a concentration-dependent manner. The functional and LDH data for cromakalim are not shown, but similar data have been published in detail previously.2 For these parameters, cromakalim was slightly more potent as a cardioprotectant compared with diazoxide, as was seen for time to contracture.
We also determined the effect of glibenclamide on the preischemic and postischemic cardioprotective activity of diazoxide (Fig 3⇓, Table 2⇓). At 30 μmol/L, diazoxide significantly increased preischemic coronary flow. Glibenclamide completely abolished this effect. Diazoxide significantly reduced LDH release during reperfusion, and this protective effect was completely abolished by glibenclamide. Diazoxide also significantly enhanced postischemic functional recovery, and glibenclamide completely abolished this protective effect (Table 2⇓). Cromakalim at 10 μmol/L improved postischemic contractile function and reduced cumulative LDH release to a degree similar to that observed for 30 μmol/L diazoxide. The cardioprotective effect of cromakalim was completely abolished by glibenclamide.
The effect of the more selective KATP blocker, 5-HD, on the cardioprotective activity of diazoxide is shown in Table 3⇓ and Fig 4⇓. Diazoxide exerted a significant protective effect, as measured by functional recovery and a reduced LDH release. 5-HD completely abolished this cardioprotective effect. The preischemic coronary dilator effect of diazoxide was not attenuated by 5-HD, which is consistent with the selective nature of its KATP blockade. Previous studies have shown both 5-HD and glibenclamide alone (at the concentrations used in the present study) have no effect on severity of ischemic/reperfusion damage in our isolated rat heart model.2,14,23,24
A small study was conducted to show whether diazoxide could also protect ischemic tissue in another species. We chose rabbits because their hearts can be readily used for Langendorff preparations and clear determinations of ischemic severity can be made. As shown in Table 4⇓, diazoxide was protective in rabbit hearts, as measured by an enhanced recovery of postischemic contractile function and reduced LDH release. Diazoxide significantly increased preischemic coronary flow and had no effect on cardiac function at this time. A similar response was noted for cromakalim. In addition to the protective effects on reperfusion function and necrosis, diazoxide significantly (P<.05) increased the time to contracture during ischemia from 42.9±0.9 minutes in vehicle-treated hearts to 49.5±0.3 minutes in diazoxide-treated hearts. Cromakalim increased the time to contracture (47.9±0.7 minutes) to a degree similar to that found with diazoxide at the 30 μmol/L concentration. At 100 μmol/L, 5-HD completely abolished the cardioprotective effects of diazoxide (Table 4⇓). 5-HD completely abolished the increase in time to contracture (42.7±0.8 min) observed with diazoxide alone.
Effect of Diazoxide and Cromakalim on Whole-Cell Currents
Consistent with previous studies,25 stimulation of isolated voltage-clamped rat ventricular myocytes with a slow voltage ramp evoked an N-shaped current response (Fig 5⇓). Exposure of the cells to 100 μmol/L diazoxide had little affect on the current-voltage relationship; current measured at +40 mV was 0.56±0.09 nA in control and 0.59±0.09 nA in the presence of diazoxide (n=6, Fig 5⇓). Addition of 300 μmol/L diazoxide produced a robust increase in current at voltages positive to ≈−85 mV in 8 of 18 cells tested. At 1000 μmol/L, diazoxide evoked a large response in 6 of 8 cells tested (Fig 5A⇓ and 5B⇓). The activated current was reversible on washout. Interestingly, the higher concentrations of diazoxide, especially 1000 μmol/L, partially inhibited the control currents (Fig 5A⇓).
The reversal potential of the activated current indicates that diazoxide evoked a K+ current. Moreover, the current-voltage relationship observed in the presence of diazoxide is similar to that previously reported for the K+ channel activator rimakalim in rat cardiac myocytes.20 Currents activated by diazoxide (1000 μmol/L) were fully blocked by the subsequent addition of glibenclamide (3 μmol/L, n=6, Fig 5A⇑) but refractory to inhibition by 5-HD (100 μmol/L, n=4). Cromakalim was found to be considerably more potent than diazoxide, producing a small increase in current at 10 μmol/L in 5 of 8 cells tested, and a robust response at 30 μmol/L in all cells tested (n=7, Fig 5B⇑). A comparison of the concentration-response curves indicates that diazoxide was ≈50-fold less potent than cromakalim as a K+ channel activator in isolated rat ventricular cells.
Effect of Diazoxide on Normal and Ischemic APD in Rat Hearts
Since 10 μmol/L cromakalim and 30 μmol/L diazoxide were found to be equivalent in terms of their ability to protect ischemic rat hearts, we compared the effect of these concentrations on APD in normal and ischemic rat hearts. Vehicle-treated hearts displayed an early ischemic APD lengthening before shortening to 25±6% at 6 minutes after occlusion (Fig 6⇓). The only statistically significant shortening occurred at 5 and 6 minutes after ischemia in the vehicle-treated hearts. In the 8 hearts that showed tachyarrhythmias, the mean time to onset was 406±64 seconds. The remaining preparations were discontinued for reasons stated above (loss of reliable action potential measurements and inability to pace); therefore, the n value for all time periods may not be the same. No effect was seen on preischemic APD values after 10 minutes of exposure to 30 μmol/L diazoxide. Although diazoxide-treated hearts maintained the early portion of the APD lengthening (1-minute data) seen in the vehicle group, at 2 and 3 minutes after occlusion diazoxide did slightly shorten APD relative to vehicle. At later time points (4 to 6 minutes), diazoxide did not significantly reduce APD relative to vehicle-treated hearts. Mean onset time to tachyarrhythmias was 265±40 seconds in diazoxide-treated hearts. Cromakalim caused a 7±3% shortening of preischemic APD, but this was not statistically significant. On occlusion, cromakalim caused an abolition of the early lengthening seen with vehicle and diazoxide, and APD shortening reached a nadir that was 67±10% less than its control value at 5 minutes of ischemia. All points collected after occlusion were significantly different from corresponding control, diazoxide, or vehicle values. A mean duration of 146±10 seconds was noted for the onset to tachyarrhythmias for cromakalim, and most preparations invariably showed fibrillatory activity early in the protocol. It is important to note that the arrhythmias noted in the vehicle- and diazoxide-treated groups were mostly tachycardic, and at no time was the cycle length statistically different among groups, irrespective of the need for pacing. Basic cycle lengths were 220±9, 215±9, and 213±8 milliseconds for vehicle, diazoxide, and cromakalim, respectively. These data taken together show that there is little relationship between APD shortening and cardioprotection for diazoxide and cromakalim.
Effects of Diazoxide on Reconstituted Bovine Heart Mitochondrial and Sarcolemmal KATP Activity
Fig 7⇓ contains the diazoxide (Fig 7A⇓) and cromakalim (Fig 7B⇓) concentration-response curves for stimulation of K+ flux in vesicles reconstituted with KATP purified from bovine heart mitochondria (solid circles, Fig 7⇓) and sarcolemma (open circles, Fig 7⇓). Cromakalim was a potent activator of K+ flux in both preparations (Fig 7B⇓). Observed K1/2 values for cromakalim were 1.6±0.1 μmol/L (n=5) for mitochondrial KATP and 18±2 μmol/L (n=3) for sarcolemmal KATP. The reconstituted KATP from the two sources responded differently to diazoxide. Observed K1/2 values for diazoxide were 0.80±0.03 μmol/L (n=4) for mitochondrial KATP and 840±25 μmol/L (n=3) for sarcolemmal KATP. In both preparations, KATP openers activated K+ flux only when it was inhibited by ATP. Therefore, mitochondrial KATP is ≈1000-fold more sensitive to diazoxide than sarcolemmal channels. These data confirm previously published work from this laboratory.12 Also, compared with cromakalim, diazoxide is ≈50-fold less potent at activating sarcolemmal KATP, which is consistent with the findings seen with whole-cell currents.
Effect of Diazoxide and Cromakalim on Rat Heart Mitochondrial KATP: Effect of KATP Blockade
Since the ischemia studies were performed in isolated rat hearts, we also determined the effect of diazoxide on rat cardiac mitochondrial KATP. As shown in Fig 8A⇓, diazoxide and cromakalim are potent activators of the ATP-inhibited mitochondrial KATP from rat heart. Observed K1/2 values were 0.49±0.05 μmol/L for diazoxide (n=3) and 1.1±0.1 μmol/L for cromakalim (n=3). These values are similar to those previously obtained with reconstituted mitochondrial KATP from rat liver and bovine heart.12
To mimic the in situ pharmacological experiments, we included MgATP, diazoxide (10 μmol/L), and glibenclamide or 5-HD in the assay medium and varied the inhibitor concentration. As demonstrated in Fig 8B⇑, K+ flux through the diazoxide-opened channel was inhibited by low concentrations of glibenclamide (K1/2, 56 nmol/L) and 5-HD (K1/2, 83 μmol/L). Similar results were obtained in two experiments using 10 μmol/L cromakalim, with glibenclamide inhibiting with K1/2 of 92 nmol/L and 5-HD inhibiting with K1/2 of 31 μmol/L. In the absence of ATP and K+ channel openers, glibenclamide inhibited K+ flux with K1/2 of 40 nmol/L, whereas 5-HD had no effect up to 500 μmol/L, the highest concentration tested.
Structurally diverse KATP openers exert cardioprotective effects in various animal models of ischemia.2 These protective effects are exerted directly on ischemic myocytes and are independent of vasodilatory activity. The cardioprotective effects of KATP openers are universally abolished by KATP blockers such as glibenclamide. The ability of glibenclamide to abolish cardioprotective effects is selective for KATP openers.23 KATP openers may mimic an endogenous cardioprotective mechanism, since many laboratories have shown cardiac preconditioning to be abolished by KATP blockers.26–28
It was tempting to conclude that the cardioprotective effects of KATP openers were due to APD shortening and the consequent reduction of Ca2+ entry into myocytes. KATP openers do conserve ATP during ischemia, which is at least consistent with a cardioplegic effect for KATP openers.3,4 Data from Gross’s laboratory5 as well as our own6 suggested a lack of correlation for ischemic APD shortening and cardioprotection. Monophasic APD measurements were used in these studies; therefore, regional changes in APD could have gone undetected. Subsequently, we performed studies showing a lack of correlation between intracellular APD measurements and glibenclamide-reversible cardioprotection for the pyranyl cyanoguanidine BMS-180448.7 This compound is interesting because it is less potent as a vasodilator compared with cromakalim but is identical as a cardioprotectant.27 Interestingly, BMS-180448 is poor at opening single KATP in cardiomyocytes, and the mechanism for the glibenclamide reversibility of its cardioprotective action is not clear. We hypothesized that an intracellular KATP could be mediating cardioprotection.
A KATP that appears to be opened by benzopyran KATP openers such as cromakalim is expressed in mitochondrial membranes.9–12 This channel was clearly shown to be inhibited by glibenclamide.9 This channel was also shown to be inhibited by ATP and ADP and to be dependent on [K+]o, and there appeared to be little dependence of pH on channel function.9 The function of mitochondrial KATP is to regulate mitochondrial volume, which, in turn, is thought to regulate electron transport.12 Recent results from Garlid’s laboratory12 have also shown that diazoxide opens liver mitochondrial KATP in the low micromolar range. Diazoxide is an interesting compound because it is potent at opening pancreatic KATP but is relatively weak at shortening cardiac APD.1,13 Results shown in the present study also demonstrated diazoxide to be ≈2000-fold more potent than sarcolemmal KATP at opening mitochondrial KATP. This makes diazoxide a potentially important tool for dissecting the role of mitchondrial KATP in heart. With this in mind, we determined the effect of diazoxide in a rat model of ischemia and reperfusion. Potent cardioprotective effects exerted by diazoxide would strongly implicate a role for mitochondrial KATP.
Diazoxide protected ischemic/reperfused rat hearts with a potency in the low micromolar range. This protection was concentration dependent and was similar in profile to other KATP openers tested in this model, although it is slightly less potent than agents such as cromakalim and aprikalim.2,24 The protective effects of a high concentration of diazoxide were completely abolished by glibenclamide; however, they were not clearly correlated with APD shortening either before or during ischemia. Diazoxide had no effect on APD before ischemia and was significantly less potent than cromakalim during ischemia. Ischemic APD shortening was slightly enhanced by diazoxide early into ischemia; however, the ultimate amount of shortening was similar to vehicle. Diazoxide (30 μmol/L) was found to be comparable to 10 μmol/L cromakalim in the degree of cardioprotection. Historical EC25 values for cromakalim (time to contracture) are 5 to 9 μmol/L,2,8 which are comparable to the values found for diazoxide. The potency of these two agents for opening mitochondrial KATP was also similar. Therefore, diazoxide and cromakalim differ markedly in their effects on APD as well as on the opening of reconstituted sarcolemmal KATP, and the common thread is that both open cardiac mitochondrial KATP with high potency. These studies were important in documenting a separation of sarcolemmal KATP opening from cardioprotection in which an index of such opening (APD) was measured during ischemia. A similar separation was seen for whole-cell myocyte currents. Diazoxide activated IKATP but was ≈50 times less potent than cromakalim. Moreover, the diazoxide-activated current was refractory to block by 5-HD (100 μmol/L). Similar results were previously reported for cromakalim-activated IKATP in guinea pig myocytes, where the activated currents were also fully blocked by glibenclamide but insensitive to 5-HD.14 Thus, sarcolemmal KATP is sensitive to block by glibenclamide but not 5-HD.
Further evidence for a role for mitochondrial KATP in mediating the cardioprotective effects of diazoxide was seen in the 5-HD studies. Previous studies showed that 5-HD does not abolish the APD shortening or vasodilator effects of KATP openers but completely abolishes their cardioprotective effects.14 As further confirmation of the importance of mitochondrial KATP, we determined the effect of 5-HD on the pharmacological actions of diazoxide. 5-HD completely abolished the cardioprotective effects of diazoxide but did not affect its coronary dilator activity, which is consistent with its effects on other KATP openers.14 5-HD also significantly inhibited the ability of diazoxide to open reconstituted mitochondrial KATP. These data strongly suggest an important role for mitochondrial KATP in mediating the cardioprotective effects of KATP openers. The purity of mitochondrial pellets and sarcolemmal membranes was verified as described previously.9
The data in the present study, as well as previously reported work, suggest the possibility that differences between KATP may be not only at the level of cell type but also at the level of organelles. The data collected in the present study are consistent with the idea that the relevant cardioprotective site for KATP openers is mitochondrial KATP. Pyranyl cyanoguanidine analogues that protect ischemic myocardium without affecting sarcolemmal K+ currents have been found. These agents protect hearts in a glibenclamide-reversible manner, suggesting an involvement with KATP. Interestingly, agents such as BMS-180448 are less potent as vasodilators compared with cromakalim, while being equipotent as cardioprotectants. Diazoxide is a potent pancreatic KATP opener, unlike BMS-180448 and cromakalim.29–32 Interestingly, diazoxide does not appear to be potent at reducing cardiac APD,13 although results have been variable and have not been measured in ischemic myocardium.33 These results suggest multiple subtypes of KATP or receptors (if not of KATP itself).
Although these data are consistent with the mitochondrion as a potential site of cardioprotective activity, it is still unclear how the opening of mitochondrial KATP can exert protective effects. Agents such as diazoxide are protective in isolated rat heart models of total global ischemia, in which oxidative phosphorylation is essentially inhibited. Nevertheless, KATP openers such as cromakalim significantly conserve ATP and inhibit accumulation of AMP during global ischemia, suggesting reduced ATP hydrolysis. Mitochondrial KATP are thought to be involved with mitochondrial volume control and energetics. Therefore, it is likely that mitochondrial KATP activation may act to inhibit ATP wastage.34 There may be an interaction between mitochondrial KATP and inefficient ATP wastage by mitochondrial ATP hydrolase activity. This plausible mechanism, which remains to be investigated, would be consistent with the ability of KATP openers to conserve ATP without major effects on cardiac function. Published data from our laboratories showed that BMS-180448 preserved mitochondrial ultrastructural integrity as well as the efficiency of oxygen utilization, further suggesting mitochondrial protection.35 Interestingly, the histological necrosis data in that study35 were well correlated with the surrogate marker of LDH release, and later studies36 further validated these findings.
Selected Abbreviations and Acronyms
|5-HD||=||sodium 5-hydroxydecanoic acid|
|APD||=||action potential duration|
|FCCP||=||carbonyl cyanide p-(trifluoromethoxy)phenylhydrazone|
|I KATP||=||glibenclamide-sensitive K+ current|
|K1/2||=||concentration in which ΔJ/ΔJmax=0.5 (50% of maximal K+ flux [J])|
|KATP channel(s)||=||ATP-sensitive K+ channel(s)|
|LVDP||=||left ventricular developed pressure|
|MAP||=||monophasic action potential|
|PBFI||=||K+-binding benzofuran isophthalate|
This study was supported in part by grant GM-31086 (Dr Garlid) from the National Institutes of Health.
- Received March 24, 1997.
- Accepted September 23, 1997.
- © 1997 American Heart Association, Inc.
Grover GJ. Protective effects of ATP-sensitive potassium channel openers in experimental myocardial ischemia. J Cardiovasc Pharmacol. 1994;24(suppl 4):S18–S27.
Grover GJ, Newburger J, Sleph PG, Dzwonczyk S, Taylor SC, Ahmed SZ, Atwal KS. Cardioprotective effects of the potassium channel opener cromakalim: stereoselectivity and effects on myocardial adenine nucleotides. J Pharmacol Exp Ther. 1991;257:156–162.
McPherson CD, Pierce GN, Cole WC. Ischemic cardioprotection by ATP-sensitive potassium channels involves high-energy phosphate preservation. Am J Physiol. 1993;265:H1809–H1818.
Yao Z, Gross GJ. Effects of the KATP channel opener bimakalim on coronary blood flow, monophasic action potential duration, and infarct size in dogs. Circulation. 1994;89:1769–1775.
Paucek P, Mironova G, Mahdi F, Beavis AD, Woldegiorgis G, Garlid KD. Reconstitution and partial purification of the glibenclamide-sensitive, ATP-dependent K+ channel from rat liver and beef heart mitochondria. J Biol Chem. 1992;267:26062–26069.
Paucek P, Yarov-Yarovoy V, Sun X, Garlid KD. Physiological and pharmacological activators of the mitochondrial KATP channel. Biophys J. 1995;68:A145. Abstract.
Garlid KD, Paucek P, Yarov-Yarovoy V, Sun X, Schindler PA. The mitochondrial KATP channel as a receptor for potassium channel openers. J Biol Chem. 1996;271:8796–8799.
McCullough JR, Normandin DE, Conder ML, Sleph PG, Dzwonczyk S, Grover GJ. Specific block of the anti-ischemic actions of cromakalim by sodium 5-hydroxydecanoate. Circ Res. 1991;69:949–958.
Hamill OP, Marty A, Neher E, Sakmann B, Sigworth FJ. Improved patch-clamp techniques from high resolution current recording from cells and cell-free membrane patches. Pflugers Arch. 1981;395:85–100.
Matlib MA, Rouslin W, Vaghy PL, Schwartz A. Isolation of cardiac muscle mitochondria. Methods in Pharmacology. New York, NY: Plenum Publishing Corp; 1984;5:25–37.
McEnery MW, Hullihen J, Pedersen PL. Fo ‘proton channel’ of rat liver mitochondria. J Biol Chem. 1989;264:12029–12036.
Paucek P, Garlid KD. Isolation, and reconstitution of two distinct KATP channels from bovine heart sarcolemma, and mitochondria. Biophys J; 1993;64, A311. Abstract.
Periyasamy SM, Kakar SS, Garlid KD, Askari A. Ion specificity of cardiac sarcolemmal Na+/H+ antiporter. J Biol Chem. 1990;265:6035–6041.
Sargent CA, Smith MA, Dzwonczyk S, Sleph PG, Grover GJ. Effect of potassium channel blockade on the anti-ischemic actions of mechanistically diverse agents. J Pharmacol Exp Ther. 1991;259:97–103.
Auchampach JA, Grover GJ, Gross GJ. Blockade of ischemic preconditioning in dogs by the novel ATP dependent potassium channel antagonist sodium 5-hydroxydecanoate. Cardiovasc Res. 1992;26:1054–1062.
Schulz R, Rose J, Heusch G. Involvement of activation of ATP-dependent potassium channels in ischemic preconditioning in swine. Am J Physiol. 1994;267:H1341–H1352.
Toombs CF, Moore TL, Shebuski RJ. Limitation of infarct size in the rabbit by ischemic preconditioning is reversible with glibenclamide. Cardiovasc Res. 1993;27:617–622.
Ciampolillo F, Tung DE, Cameron JS. Effects of diazoxide and glyburide on ATP-sensitive K+ channels from hypertrophied ventricular myocytes. J Pharmacol Exp Ther. 1992;260:254–260.
Rousslin W, Broge C-W. Mechanisms of ATP conservation during ischemia in slow and fast heart rate hearts. Am J Physiol. 1993;264:C209–C216.
Grover GJ, Dzwonczyk S, Monticello T. Comparative effects of cromakalim and diltiazem in ischemic hypertrophied and non-hypertrophied rat hearts. Am J Physiol. 1996;271:H174–H182.