Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Pharmacological K_ATP activation is associated with cardioprotection and may simulate some aspects of ischemic preconditioning (Auchampach et al., 1991; Gross and Auchampach, 1992; Grover et al., 1994; Armstrong et al., 1995). Numerous K_ATP subtypes exist and seem to be differentially expressed in various tissues (Inoue et al., 1991; Atwal et al., 1993; Grover et al., 1995a; Chutkow et al., 1996; Inagaki et al., 1996). Pharmacological data suggest that mitochondrial K_ATP activation is critical for cardioprotection and that this channel is distinct from sarcolemmal channels (Garlid et al., 1996, 1997; Liu et al., 1998; Nakai et al., 2001). Clear pharmacological separation between smooth muscle relaxation and cardioprotection has been reported for several K_ATP openers such as BMS-180448 and BMS-191095 (Atwal et al., 1993; Grover et al., 1995b, 2001; Rovnyak et al., 1997) (chemical structures shown in Fig. 1). BMS-191095, in particular, is very selective with no vasorelaxant activity while retaining the cardioprotective efficacy of nonselective agents such as cromakalim or pinacidil (Grover et al., 2001). In addition, there are no electrophysiological effects in isolated rat or guinea pig hearts, suggesting a lack of effect on cardiac sarcolemmal K_ATP. Further evidence for selectivity was the lack of effect of BMS-191095 on whole cell myocyte K_ATP current (Grover et al., 2001). Interestingly, the protective effects of this compound were completely abolished by glyburide and 5-HD.

View larger version (12K): [in this window] [in a new window]   Fig. 1.   Chemical structures of BMS-180448, cromakalim, and BMS-191095.

Recent data suggest the importance of mitochondrial K_ATP in mediating cardioprotection (Garlid et al., 1997; Liu et al., 1998), and these studies have been primarily done using diazoxide. Although diazoxide opens cardiac mitochondrial K_ATP with 1000-fold selectivity compared with cardiac sarcolemmal channels, it is a potent sarcolemmal K_ATP opener in vascular smooth muscle (Garlid et al., 1997). Diazoxide, therefore, will significantly reduce arterial blood pressure well before cardioprotective doses are achieved. Recently published data show that BMS-191095 selectively opens mitochondrial K_ATP without affecting sarcolemmal channels in vascular smooth muscle, heart or pancreatic -cells (Grover et al., 2001). BMS-191095 would be expected to be devoid of vasodilator and proarrhythmic activity unlike nonselective agents, and such activity is contraindicated during acute myocardial ischemia (Chi et al., 1990; Belin et al., 1996; Grover et al., 2001). This study with BMS-191095 was done in vitro or in isolated hearts and smooth muscle ex vivo (Grover et al., 2001). Although previous work on related K_ATP openers show that the in vitro and in vivo potencies correlate well (D'Alonzo et al., 1995a; Grover et al., 1995a, 1996), the goal of this study was to compare the cardioprotective, electrophysiological, and hemodynamic properties of BMS-191095 in vivo. This was done in canine models of ischemia and reperfusion as well as proarrhythmia models. Our data showed that BMS-191095 reduced infarct size in a dose-dependent manner while being devoid of cardiac electrophysiological and hemodynamic effects. No proarrhythmogenic activity was observed, which is consistent with selectivity for mitochondrial channels.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Canine Model of Ischemia and Reperfusion

Mongrel dogs of either sex (15-24 kg) were anesthetized with i.v. sodium pentobarbital (30 mg/kg), and a catheter was placed into the right femoral artery for later collection of blood samples for blood gases and reference blood flow analysis. This technique has been described previously (Grover et al., 1995a). Another catheter was placed into the right femoral vein for supplementation of anesthesia and drug infusion. A Mikrotip catheter pressure transducer (Millar, Houston, TX) was placed into the left femoral artery and was advanced into the aortic arch for the measurement of arterial blood pressure. An endotracheal tube was placed into the trachea and the animals were artificially respired, and this was maintained such that eucapnia and normoxia were observed throughout the study.

A left thoracotomy was performed at the 5th intercostal space and the heart was exposed. The left circumflex coronary artery (LCX) was isolated proximal to its first branch, and a silk suture was placed around it for later occlusion. A catheter was placed into the left atrial appendage for dye (area at risk measurement) and radioactive microsphere injection.

The animals were allowed to stabilize for 5 to 10 min at which time an arterial blood sample was removed anaerobically for measurement of blood gases using a Radiometer (ABL4, Copenhagen, Denmark) blood gas analyzer. The blood gases were adjusted to normoxic and eucapnic levels by adjustment of the ventilator. Arterial blood pressure and heart rate were measured at baseline once the eucapnia was achieved. At this time myocardial blood flow was measured using radioactive microspheres (¹¹³Sn, ⁵⁷Co, ⁸⁵Sr, or ⁴⁶Sc; 15 ± 3 µm; PerkinElmer Life Sciences, Boston, MA). Before coronary occlusion, the animals were divided into five groups: 1) vehicle-treated animals (i.v., saline, n = 6) starting 10 min before LCX occlusion and given over a total of 25 min; 2) BMS-191095-treated animals (8 µg/kg/min, n = 6); 3) BMS-191095-treated animals (25 µg/kg/min i.v., n = 6); 4) BMS-191095-treated animals (80 µg/kg/min i.v., n = 6); 4) BMS-191095-treated animals (139 µg/kg/min i.v., n = 6); and 5) BMS-191095-treated animals (139 µg/kg/min i.v.) + 150 µg/kg/min 5-HD (150 µg/kg/min, intracoronary, n = 6). BMS-191095 or vehicle was started 10 min preocclusion and the total dose given as 0.2, 0.6, 2.0, or 3.5 mg/kg (over a total of 25 min). 5-HD was given directly into the LCX starting 10 min before ischemia and for an additional 10 min into ischemia for a total dose of 3 mg/kg. After the first 10 min of drug infusion, the LCX was completely occluded for a total of 90 min. Myocardial blood flow was again measured 40 min after the initiation of LCX occlusion and in this model is a measure of collateral blood flow into the ischemic region. At 90 min after occlusion, the LCX was reperfused. After 1 h of reperfusion, regional myocardial blood flow was again measured.

The reperfusion was continued for a total of 5 h at which time the LCX was cannulated and perfused at the animals' existing pressure with Ringer's lactate for determination of the area at risk. Patent blue violet dye (1 mg/kg of a 10-mg/ml solution) was injected into the left atrial catheter and the heart was quickly excised and the atria trimmed. The ventricles were cut transversely into 0.5-cm slices. The borders of the area at risk (no stain) were delineated and separated and the slices were incubated at 37°C for 30 min in a 1% solution of 2,3,5-triphenyl tetrazolium chloride in phosphate-buffered saline. This agent stains viable tissue red, whereas infarcted tissue is not stained and becomes white or gray in color. The myocardial area at risk and infarct areas of interest were measured using computerized planimetric techniques. The tracings were digitized using Applescan (Apple Computer Inc., Cupertino, CA) and then area was determined on a Macintosh IIcx computer using Macdraft software (Innovative Data Design, Concord, CA). The infarct size was expressed as a percentage of the left ventricular area at risk. Myocardial blood flow was calculated by taking myocardial pieces from the subepicardial and subendocardial halves of the ischemic and nonischemic regions (four pieces from each area) of the left ventricular free wall. The center of the ischemic region was used for the ischemic regional blood flow determinations. The radioactivity in the tissue pieces as well as the reference blood samples were determined in an Autogamma 8000 gamma counter (Beckman Coulter, Inc., Irvine, CA), and tissue flows were calculated from these counts. Reference blood withdrawal rate was 9 ml/min.

A separate study was done to ascertain plasma concentrations of BMS-191095 at times relevant to ischemic protection. Conscious mongrel dogs of either sex (15-24 kg) were infused i.v. with a total dose of 0.2 mg/kg (n = 3) or 0.6 (n = 3) mg/kg BMS-191095. The right cephalic vein of each dog was cannulated for drug infusion and the left cephalic vein was cannulated for blood withdrawal. The animals were restrained using a sling. This infusion was continued for a total of 25 min. Blood was withdrawn from the dogs at 10, 25, and 40 min after initiation of drug infusion. The 40-min time represented a point in which BMS-191095 was washing out for 15 min. It also is equivalent to 30 min into the ischemic interval for the infarct size studies, and thus is a relevant time in terms of cardioprotective efficacy. Venous blood was collected into sterile Vacutainer tubes containing EDTA. The tubes were centrifuged at 20°C at 2700 rpm for 17 min. The plasma was transferred to 5-ml opaque plastic test tube, sealed, and frozen. All plasma samples were assayed for BMS-191095 concentrations by a validated liquid chromatography/mass spectroscopy method. The lower limit of quantification of the analytical method was 0.75 ng/ml BMS-191095 in plasma.

Electrophysiological and Arrhythmia Models in Dogs

Electrophysiological Characterization in Dogs. Male mongrel dogs (15-25 kg, n = 13) were anesthetized with dial urethane (0.35 ml/kg i.v.) and were intubated and mechanically ventilated (model 613; Harvard Apparatus, South Natick, MA) with room air sufficient to maintain eucapnia. The right femoral artery and vein were cannulated to measure systemic blood pressure and to infuse drug, respectively. A lead II ECG was continuously monitored. The arterial catheter was connected to a pressure transducer (model P23XL; Spectromed, Oxnard, CA) and associated amplifier and chart recorder (TA 4000; Gould, Cleveland, OH) to monitor arterial pressure. A left thoracotomy was performed at the 5th intercostal space. Stimulating and recording electrodes were placed on the left atrium and left ventricle such that the heart could be paced from the atrium and stimuli administered (model PGEN; N.B. Datyner, Stoney Brook, NY, and voltage to current converter; SIS, Princeton, NJ) to the ventricle while simultaneously monitoring ventricular electrogram. The left carotid artery was isolated and a quadrapolar catheter was inserted into the vessel and positioned to record the His-bundle electrogram. All waveforms were displayed on a chart recorder (TA4000; Gould).

Effect in a PES Model. Seventeen fasted (12 h) mongrel dogs (16-25 kg) of either sex were anesthetized with thiopental (12 mg/kg i.v.), given atropine (15 mg/kg i.m.), and intubated. The animal was connected to an inhalational anesthesia machine (Narkovet Deluxe; North American Drager, Telford, PA), and anesthesia was maintained with isoflurane (1-2%) delivered with oxygen (100%) at 4 to 10 respirations/min to maintain normocapnia. An ECG harness was connected to the dog, and a standard lead II ECG (VSM Monitor; Physio Control, Redmond, WA) was monitored.

Drug Preparation

Dial urethane was prepared as follows: 40% urethane (Sigma-Aldrich), 20% 5,5 diallyl barbituric acid (Sigma-Aldrich), and 40% of ethyl urea (Aldrich Chemical Co., Milwaukee, WI). The chemicals were placed in a graduated cylinder and distilled water added to achieve a proper concentration. The solution was transferred to 100-ml amber bottles, stored at room temperature, and protected from light. Hearts were stained with 2,3,5-triphenyltetrazolium chloride (Sigma-Aldrich) prepared as a 1% solution in phosphate-buffered saline, pH 7.4 (Sigma-Aldrich). BMS-191095 was prepared (20 mg/ml) fresh on the day of use in polyethylene glycol 400 (Fisher Scientific, Fair Lawn, NJ).

Statistics

Comparisons between treatments for infarct size, hemodynamics, and regional myocardial blood flow electrophysiological parameters were done using a factorial analysis of variance. For the analysis of variance studies, a Newman-Keuls post hoc test was used. Differences were deemed significant if p < 0.05. Inducibility differences were determined using a Fisher's exact test. All values were expressed as mean ± S.E.M.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Effect of BMS-191095 on Infarct Size in Dogs. The infarct size and area at risk data for BMS-191095 and vehicle-treated animals are shown in Fig. 2. The left ventricular area at risk (shown as percentage of left ventricle) was similar for all groups, indicating comparable anatomy. Infarct size expressed as a percentage of the area at risk was approximately 70% in vehicle-treated animals and was significantly reduced by BMS-191095 in a dose-dependent manner, with the lowest efficacious dose being approximately 0.6 mg/kg. The infarct size in the 3.5 mg/kg BMS-191095-treated group was reduced by approximately 70% relative to vehicle-treated animals. The infarct size data were also calculated as a percentage of reduction from the respective vehicle-treated group. From these data we calculated an ED₂₅ for infarct size reduction. ED₂₅ was defined as the dose causing 25% reduction in infarct size from vehicle-treated group values and 25% reduction was chosen as a minimal effective dose. The ED₂₅ for BMS-191095 was found to be 0.4 mg/kg. 5-HD abolished the protective effect of 3.5 mg/kg BMS-191095 such that infarct size was 66 ± 6% of area at risk, which is not significantly different from vehicle-treated animals (Fig. 2)

View larger version (48K): [in this window] [in a new window]   Fig. 2.   Effect of increasing doses of BMS-191095 on infarct size expressed as a percentage of the area at risk (AR). BMS-191095 reduced infarct size in a dose-dependent manner starting at the 0.6-mg/kg dose. The area at risk (expressed as percentage of LV) was similar for all groups.

View larger version (31K): [in this window] [in a new window]   Fig. 3.   Plasma BMS-191095 concentrations within the cardioprotective range (0.6 mg/kg) and below the cardioprotective range (0.2 mg/kg). Plasma concentrations were between 0.4 and 1 µM at the relevant cardioprotective dose. This agrees well with the cardioprotective potency calculated in vitro.

Electrophysiological Characterization of BMS-191095. BMS-191095 had no significant effects on His-bundle conduction (Table 3). Also, there were no significant changes in ventricular ERP, CT, CV, wavelength, or APD90 with BMS-191095 treatment (Table 4). No changes in AERP were observed (data not shown). There were no rate-dependent changes in ventricular CT or CV observed in these studies. As expected, decreases in ventricular ERP and wavelength were rate-dependent. There were no significant changes in wavelength in both treatment groups, and any decreases were associated with decreases in BCL and not administration of compound. The data in Tables 3 and 4 are expressed as percentages of change from predrug values.

PES Model. BMS-191095 produced no significant changes in ventricular ERP (148 ± 4 and 140 ± 4 ms), APD (152 ± 4 and 153 ± 4 ms), conduction time (56 ± 3 and 52 ± 3 ms), or wavelength (55 ± 3 and 56 ± 3 mm) relative to control values in either the normal or ischemic regions of the myocardium, respectively (Table 6). There was also no change in QT-intervals from a control value of 178 ± 4 ms. To normalize for differences in heart rate, some of the animals were tested at their intrinsic rate. In the BMS-191095 treatment group, none (0%) of the animals converted to an inducible state. After final ligation of the LCX, BMS-191095 did not cause any mortality. Historical data with vehicle show approximately 70 to 80% survival using this protocol. BMS-191095 caused no hemodynamic effects at any dose tested (data not shown).

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

Pharmacological activation of K_ATP is associated with cardioprotection in numerous models of ischemia and reperfusion (McPherson et al., 1993; Yao and Gross, 1994; Mizimura et al., 1995; Liu et al., 1998). The pharmacological cardioprotection is seen in these studies with distinct K_ATP opener chemotypes. The cardioprotective effects K_ATP openers are abolished by glyburide and 5-HD. Although the broad class of K_ATP openers exert cardioprotective effects, there is some diversity in terms of the pharmacological profiles of K_ATP openers with some being potent vasodilators, some having effects on pancreatic -cells, and others having no effect on cardiac action potential duration (Atwal et al., 1993; Inagaki et al., 1996; Rovnyak et al., 1997). Studies from several laboratories showed a poor correlation between action potential shortening and cardioprotection for compounds and preconditioning, suggesting that sarcolemmal activation may not be important (Yao and Gross, 1994; Grover et al., 1995b; Grover and Sleph, 1995; Hamada et al., 1998). Structure-activity studies using benzopyran and cyanoguanidine analogs showed a clear delineation between vasodilator and APD shortening effects versus cardioprotection (Atwal et al., 1993; Rovnyak et al., 1997). Despite this unusual profile, the cardioprotective effects of these selective agents were completely abolished by K_ATP blockers.

One K_ATP opener, diazoxide, is a potent vasodilator despite weak effects on APD shortening (Faivre and Findlay, 1989). Garlid showed that diazoxide is fairly selective for cardiac mitochondrial K_ATP compared with cardiac sarcolemmal channels (Garlid et al., 1996). Within the dose range that it specifically opens cardiac mitochondrial K_ATP, diazoxide still exerted cardioprotective effects (Garlid et al., 1997). Nonselective K_ATP openers such as bimakalim also opened mitochondrial channels (Garlid et al., 1996), although this occurred within the dose range where sarcolemmal activation was observed. Mitochondrial K_ATP have been shown to be important in mediating pharmacological cardioprotection and preconditioning (Liu et al., 1998; Nakai et al., 2001)

BMS-191095 was shown to be highly selective for mitochondrial K_ATP and has no effects on smooth muscle or pancreatic cells (Grover et al., 2001). Although this compound is of great interest, only in vitro and ex vivo (Langendorff hearts) studies have been reported previously. The goal of the present study was to show efficacy and selectivity of BMS-191095 in vivo. This was done by comparing hemodynamic, electrophysiological, and cardioprotective activity in dogs and correlating this with activity expected for a mitochondrial-specific opener.

BMS-191095 reduced infarct size in a clear, dose-dependent manner. We were able to calculate a potency that was described as the dose causing a 25% reduction in infarct size (ED₂₅). ED₂₅ was selected because it is approximately the minimal reduction in infarct size needed to see statistically relevant changes. The ED₂₅ value of 0.4 mg/kg probably represents the minimally efficacious dose for this compound. The plasma concentrations for efficacious doses of BMS-191095 were consistent with in vitro potency data showing low micromolar potency (Grover et al., 2001). The cardioprotective effect of the highest dose of BMS-191095 tested was abolished by 5-HD, further confirming the mechanism of protection as being mitochondrial K_ATP.

This cardioprotective activity was not associated with any hemodynamic change, distinguishing BMS-191095 from first generation compounds such as cromakalim and diazoxide. The hemodynamic effects of these first generation agents preclude their clinical use due to the toxic effects of hypotension (Belin et al., 1996). Although diazoxide is selective for cardiac mitochondrial K_ATP relative to cardiac sarcolemmal channels, its potent vasodilator effect precludes its use as a cardioprotectant clinically. Reperfusion myocardial blood flow was increased in the BMS-191095-treated animals, but this may be related to improved viability and metabolic demand from salvaged tissue rather than a direct vasorelaxant effect. This is confirmed by the loss of the enhanced reperfusion blood flow effect of BMS-191095 by coadministration with 5-HD. 5-HD will not block direct vasodilator effects of K_ATP openers (McCullough et al., 1991). This is further confirmed by the lack of blood flow changes in the nonischemic regions. It must be stated that it is possible that some additional actions for BMS-191095 at doses at or above BMS-191095 exist. This may explain the lack of clear dose dependence for these effects at the lower doses.

The degree of cardioprotection seen for BMS-191095 was profound particularly because the 90-min coronary occlusion model is so severe. Ischemic preconditioning is not highly efficacious in dog in this model (Gumina et al., 1999). The 70% reduction in infarct size also compares favorably to the degree of protection for sodium/hydrogen exchange inhibitors seen in the 90-min coronary occlusion model in dogs (Gumina et al., 1999). This degree of protection has not typically been found for other K_ATP openers, perhaps due to their lack of selectivity (Grover, 1994) and consequent limitation of dosing.

APD shortening or changes in refractory periods may also be contraindicated due to increased propensity for arrhythmogenesis. Selective openers of mitochondrial K_ATP would be expected to be devoid of electrophysiological effects typically seen for nonselective agents (Cole et al., 1991; D'Alonzo et al., 1992; Hiraoka and Furukawa, 1998). Doses of BMS-191095 that were within the cardioprotective range had no effects on atrial and ventricular conduction, ventricular refractory period, or APD configuration. As would be expected of an agent devoid of such electrophysiological changes, no proarrhythmic activity was observed. The lack of effect of BMS-191095 in proarrhythmia models confirms the observations that this compound has no effects on APD and therefore is devoid of cardiac sarcolemmal K_ATP activation. BMS-191095 had no effects in the PES model and a model of sudden cardiac death. Therefore, it is possible to retain the cardioprotective effects of K_ATP openers while effectively eliminating potentially undesirable proarrhythmic activity.

Identification of the importance of mitochondrial K_ATP in the pathogenesis of myocardial ischemia and preconditioning was a critical step forward. First generation K_ATP openers are nonselective and shorten APD, relax smooth muscle, and exert cardioprotective effects (Edwards and Weston, 1993; Ashcroft et al., 1996). Structure-activity work clearly showed a separation between smooth muscle relaxant and cardioprotective activities for several chemotypes (Grover et al., 1995). BMS-191095 is completely devoid of vasodilator activity but nevertheless retains cardioprotective activity and has recently been shown to be a selective mitochondrial K_ATP opener that not only explains its lack of effect on blood pressure and cardiac APD but also confirms the central role for this channel in cardioprotection. From a clinical viewpoint, mitochondrial selective K_ATP openers will have a significantly greater therapeutic window compared with cromakalim or diazoxide due to less hemodynamic or proarrhythmic activity. Also, agents such as BMS-191095 will be potentially useful research tools.