Pretreatment with Pinacidil Promotes Arrhythmias in an Isolated Tissue Model of Cardiac Ischemia and Reperfusion

Reperfusion of ischemic cardiac tissue is associated with myocardial damage and a high incidence of arrhythmias (Pogwizd and Corr, 1987). Paradoxically, pre-exposure of the myocardium to short periods of ischemia is protective and can decrease myocardial damage and arrhythmias induced by subsequent prolonged ischemia followed by reperfusion (Kloner and Jennings, 2001). This phenomenon, called ischemic preconditioning, was first described in 1986 (Murry et al., 1986).

The mechanism that underlies the protective effects of preconditioning is not fully understood but is believed to involve opening of cardiac ATP-sensitive potassium channels (K_ATP). K_ATP channels in both the sarcolemma and mitochondria have been implicated in preconditioning (for review, see Gross and Fryer, 1999; Grover and Garlid, 2000; O'Rourke, 2000; Gross and Peart, 2003); however, whether activation of sarcolemmal or mitochondrial K_ATP channels is responsible for protective effects is controversial. Some studies have suggested that the beneficial effects of preconditioning are mediated by activation of K_ATP channels in the mitochondrial inner membrane (for review, see Gross and Fryer, 1999; Grover and Garlid, 2000; O'Rourke, 2000; Gross and Peart, 2003). In contrast, other studies have presented evidence that activation of sarcolemmal K_ATP channels is cardioprotective in ischemia and reperfusion (for review, see Gross and Fryer, 1999; Grover and Garlid, 2000; Gross and Peart, 2003).

Pretreatment of cardiac preparations with pinacidil, a non-specific K_ATP channel opener, has been reported to exert protective effects similar to ischemic preconditioning (Cole et al., 1991; Tanno et al., 2001). In Langendorff-perfused rabbit hearts, pretreatment with pinacidil prior to global ischemia resulted in a reduction of infarct size and improved contractility in reperfusion (Tanno et al., 2001). A study in isolated guinea pig free wall preparations reported that pretreatment with pinacidil before no-flow ischemia improved the recovery of developed tension and abolished severe arrhythmias in reperfusion (Cole et al., 1991). In contrast, other studies found that pinacidil pretreatment promoted arrhythmias induced by global ischemia and reperfusion (D'Alonzo et al., 1998) or hypoxia and re-oxygenation (Fischbach et al., 2003, 2004) in Langendorff-perfused hearts.

The basis for these conflicting observations is not clear. Pinacidil pretreatment has been reported to reduce Ca²⁺ overload in an isolated ventricular myocyte model of chemically induced hypoxia and re-oxygenation, which the authors suggested would be a protective action in ischemia and reperfusion (Baczko et al., 2004). Since intracellular Ca²⁺ overload has been associated with induction of cardiac arrhythmias (Clusin, 2003), this effect of pinacidil would also be expected to be antiarrhythmic. On the other hand, pinacidil causes marked abbreviation of action potentials, which are believed to promote ventricular arrhythmias caused by re-entry (Pasnani and Ferrier, 1992; Di Diego and Antzelevitch, 1993). Whether the pro- or antiarrhythmic effects of pinacidil will dominate in the setting of ischemia and reperfusion is not clear. Therefore, the overall goal of this study is to investigate the effects of pinacidil pretreatment on induction of arrhythmias by simulated ischemia and reperfusion and to relate these effects to changes in electrophysiological parameters that may provide a substrate for induction of arrhythmias.

This study uses an isolated right ventricular free wall model of simulated ischemia and reperfusion developed in this laboratory (Ferrier and Guyette, 1991; Li and Ferrier, 1991; Pasnani and Ferrier, 1992; Heisler and Ferrier, 1996). This model exhibits arrhythmias in response to ischemia and reperfusion and allows intracellular recordings to be made with microelectrodes. Thus, changes in incidence of arrhythmias can be correlated with changes in cellular electrophysiology. In addition, this model exhibits an antiarrhythmic effect in response to ischemic preconditioning (Zhu and Ferrier, 1998).

Materials and Methods

The present study conforms to guidelines published in the Canadian Council on Animal Care Guide to the Care and Use of Experimental Animals (CCAC, 1984, 1993). Experiments were conducted in ventricular tissues from 350- to 500-g albino guinea pigs obtained from Charles River (St. Constant, QC, Canada). Animals were anesthetized with sodium pentobarbital (120 mg/kg i.p.) and coadministered with heparin (3300 IU/kg i.p.) to inhibit blood coagulation. Hearts were rapidly excised and rinsed with Tyrode's solution (37°C) of the following composition: 129.0 mM NaCl; 4.0 mM KCl; 0.9 mM NaH₂PO₄; 20.0 mM NaHCO₃; 2.5 mM CaCl₂; 0.5 mM MgSO₄; 5.5 mM dextrose, pH 7.4, gassed with 95% O₂/5% CO₂. Segments of right ventricular free wall, approximately 1 cm², were dissected at 37°C in Tyrode's solution and mounted in a tissue bath as previously described (Ferrier and Guyette, 1991). In some experiments, right ventricular papillary muscles were dissected and mounted in the bath. Tissues were superfused at a rate of 30 ml/min with Tyrode's solution at 37°C for 1 h before data collection began.

Throughout the experiments, right ventricular free wall preparations were stimulated through bipolar silver electrodes placed on the endocardial surfaces of the preparations. Stimuli were delivered in repeated trains of 15 pulses, separated by 3-s pauses (Pulsar 6I; Fredrick Haer Co., Bowdoinham, ME). Stimuli were 3-ms pulses and were delivered at a frequency of 2 Hz at twice the diastolic threshold. Two high-resistance intracellular microelectrodes (18–24 MΩ), filled with 2.7 M KCl, were used to record action potentials from sites on the endo- and epicardial surfaces. Microelectrode recordings were amplified with an Axoclamp 2A amplifier (Axon Instruments Inc., Union City, CA). The endocardial recording was taken from a site at least 5 mm away from the site of stimulation. In addition, an ECG was recorded with two Ag/AgCl wires immersed in the Tyrode's solution at opposite ends of the tissue bath. The ECG was amplified with a P15 A.C. preamplifier (Grass Instruments, Quincy, MA). Biological signals were monitored on a model 5110 oscilloscope (Tektronix, Beaverton, OR). Action potentials, ECGs, and stimulus trains were also digitized with a TL1-125 analog-to-digital converter (Axon Instruments) and recorded on a computer with a continuous data acquisition and analysis program (Axotape; Axon Instruments).

After equilibration in Tyrode's solution, preparations were superfused for 15 min with Tyrode's solution modified to simulate conditions of ischemia (e.g., hypoxia, hypercapnia, acidosis, hyperkalemia, lactate accumulation, and no glucose). The composition of this “ischemic” Tyrode's solution was: 123.0 mM NaCl; 8.0 mM KCl; 0.9 mM NaH₂PO₄; 6.0 mM NaHCO₃; 2.5 mM CaCl₂; 0.5 mM MgSO₄; 20 mM sodium lactate; pH 6.8, gassed with 90% N₂/10% CO₂. Tissues were then reperfused for 30 min with normal Tyrode's solution. In experiments in which the effects of pinacidil were determined, tissues were exposed to 100 μM pinacidil 10 min in advance of superfusion with ischemic Tyrode's solution. This concentration of pinacidil was selected on the basis of preliminary experiments that indicated that 100 μM was the lowest concentration that caused significant shortening of action potential duration in both endo- and epicardium. The duration of exposure to pinacidil was 5 min. Next, tissues were superfused with drug-free Tyrode's solution for 5 min and then exposed to simulated ischemia and reperfusion as described above. Pinacidil was dissolved in DMSO. The final concentration of DMSO in Tyrode's solution was 0.03%. Therefore, in control experiments a 5-min exposure to 0.03% DMSO was substituted for pinacidil exposure.

The incidence of arrhythmias, action-potential duration at 90% repolarization (APD₉₀), for both endo- and epicardium, effective refractory period (ERP), endocardial and transmural conduction times, and conduction block were measured at set intervals throughout the experiments. Endocardial conduction times were determined by measuring the time between the stimulus and the action-potential upstroke recorded by the endocardial microelectrode. Transmural conduction times were determined by measuring the interval between the stimulus and the action-potential upstroke recorded by the epicardial microelectrode. ERP was measured by interpolating a test stimulus during the 3-s pause following the train of 15 stimuli. Initially, the test stimulus was introduced 500 ms after the last regular stimulus. The test stimulus was then moved earlier in 5- to 10-ms steps, on repetitions of the trains of stimulation, until the test stimulus no longer initiated an action potential. The shortest interval that initiated an action potential was taken as a measure of the ERP. Measurements of ERP were made in papillary muscles where premature stimulation did not induce arrhythmias.

Statistical significance of differences between groups for incidence of transmural conduction block and for incidence of arrhythmias was determined with a Fisher's exact test or χ² test. Differences in mean data for APD₉₀, ERP, and conduction times were evaluated with the univariate mode of repeated measures analysis of variance, in which multiple comparisons were made with the Bonferroni correction (SAS; SAS Institute, Cary, NC). The latter analysis compensated for missing data points resulting from conduction block or arrhythmias at specific times during ischemia or reperfusion. Differences were considered significant for p < 0.05.

Results

Figure 1A shows a schematic of the right ventricular free wall preparation and the positions of recording and stimulating electrodes used in this study. Representative microelectrode and ECG recordings are shown in panels B to E, along with a recording of the stimulus pattern. Figure 1B was recorded during the preischemic control period. Each stimulus elicited action potentials in the endocardium that propagated to the epicardium, and no arrhythmic activity was observed. Figure 1C was recorded after 5-min exposure to ischemic Tyrode's solution and shows marked abbreviation of action potentials in both endo- and epicardium. During ischemic conditions the delay between endo- and epicardial activation increased. In this example, no arrhythmias were present at 5 min of ischemia. Figure 1D, recorded after 1 min of reperfusion with normal Tyrode's solution, shows two bigeminal complexes in which the spontaneous beats were closely coupled to the stimulated beats. This was followed by initiation of sustained tachycardia. The bigeminal complexes and tachycardia were accompanied by corresponding ECG activity. Figure 1E shows that, in this example, the arrhythmias had subsided, and APD₉₀ had recovered to near preischemic values after 10 min of reperfusion.

Effects of ischemia and reperfusion on endocardial APD₉₀ in the absence and presence of pinacidil pretreatment are shown in Fig. 2A. In the absence of pinacidil (open symbols), endocardial APD₉₀ abbreviated rapidly with exposure to ischemic conditions. APD₉₀ recovered to control levels within approximately 10 min of reperfusion. In contrast, when preparations were exposed to 100 μM pinacidil for 5 min during the preischemic period, endocardial APD₉₀ decreased significantly compared with untreated preparations. This abbreviation of APD₉₀ persisted during washout of drug. In addition, the abbreviation of APD₉₀ in response to ischemia was significantly greater than in untreated preparations. This marked abbreviation of endocardial APD₉₀ recovered slowly during reperfusion but reached control values by 15 min of reperfusion.

Figure 2B illustrates the effects of ischemia and reperfusion on the APD₉₀ of action potentials recorded from the epicardium in the absence and presence of pinacidil pretreatment. Epicardial APD₉₀ shortened significantly with exposure to ischemic conditions in the absence of pretreatment with pinacidil. APD₉₀ recovered to control values during the first 10 min of reperfusion. In pinacidil-pretreated preparations, epicardial APD₉₀ abbreviated during exposure to pinacidil and recovered only slightly during washout of drug. The subsequent exposure to ischemic conditions caused a rapid and profound abbreviation of epicardial APD₉₀. Recovery of epicardial APD₉₀ in reperfusion was slow and incomplete in pinacidil-pretreated preparations. In Fig. 2B, data points are missing between 15 and 25 min of the protocol because transmural conduction block occurred during ischemia.

Abbreviation of APD₉₀ is expected to provide conditions that would promote re-entrant arrhythmias, as long as ERP also is abbreviated. In initial experiments, we attempted to determine ERP in right ventricular free wall preparations; however, the premature stimuli used to determine ERP frequently triggered arrhythmias in ischemia. Therefore, to determine whether abbreviation of ERP accompanies abbreviation of APD₉₀, we examined effects of pinacidil and ischemia in small right ventricular papillary muscles in which premature test stimuli did not trigger arrhythmias. Figure 3A shows mean effects of pinacidil pretreatment and initiation of ischemic conditions on APD₉₀ and ERP in papillary preparations. Both APD₉₀ and ERP abbreviated in parallel in response to pinacidil and to ischemia. Figure 3B shows that there was a close correlation between APD₉₀ and ERP. Thus, changes in APD₉₀ provide a good estimate for changes in ERP in this model.

Changes in conduction times may also contribute to induction of arrhythmias in ischemia and reperfusion. Therefore, we measured endocardial conduction times from the stimulus to the endocardial action-potential upstroke, as well as transmural conduction times from the stimulus to the epicardial action-potential upstroke (Fig. 4). Figure 4A shows endocardial conduction times during ischemia and reperfusion in the presence and absence of pinacidil pretreatment. Endocardial conduction times were very short and showed virtually no prolongation in response to ischemia or reperfusion. Pinacidil pretreatment had no significant effect on endocardial conduction times. Figure 4B shows that, in contrast to endocardial conduction, transmural conduction was clearly affected by ischemic conditions and pinacidil pretreatment. Mean transmural conduction times were longer than endocardial conduction times and increased significantly in response to exposure to ischemic conditions in the absence of pinacidil pretreatment. Transmural conduction times gradually recovered to control values during reperfusion. Pinacidil had no effect on transmural conduction times prior to exposure to ischemic conditions; however, pinacidil-pretreated preparations showed a rapid increase in transmural conduction time early in ischemia, followed by complete conduction block. Conduction block terminated near the end of ischemia or early reperfusion; however, transmural conduction times were significantly prolonged when conduction first resumed. With continued reperfusion, transmural conduction times eventually recovered to control values.

The incidence of transmural conduction block in ischemia was significantly different between control and pinacidil-pretreated preparations. In the absence of pinacidil pretreatment, transmural conduction block occurred in 33% of preparations during ischemia. In contrast, transmural conduction block occurred in 100% of preparations pretreated with pinacidil (p < 0.05, n = 9 control; 12 pinacidil-pretreated preparations). The mean times at which conduction block started in control and pinacidil-pretreated preparations were 7.4 ± 1 min and 6.4 ± 0.4 min, respectively. The mean times at which conduction resumed in control and pinacidil-treated preparations were 12.3 ± 1.9 min and 15.7 ± 0.8 min, respectively. Neither the times at which block started nor the times at which it stopped were statistically different between control and treated tissues.

Figure 5 shows the incidence of arrhythmias occurring in absence and presence of pinacidil pretreatment. Arrhythmias included premature beats, as well as nonsustained and sustained tachycardias. In the preischemic period, only a low incidence of premature beats was observed. In the absence of pinacidil, an increase in arrhythmias was observed during ischemia and reperfusion, although this was not statistically significant. Pinacidil did not change the incidence of arrhythmias prior to ischemia and reperfusion; however, there was a significant increase in arrhythmias in ischemia and reperfusion in preparations pretreated with pinacidil.

Two types of arrhythmias are observed in this model of ischemia and reperfusion: afterpotential-induced activity and re-entrant rhythms (Ferrier and Guyette, 1991). Figure 6A was recorded from a control preparation in early reperfusion and shows arrhythmic activity with characteristics typical of afterpotential-mediated activity. Stimulated beats are indicated by arrows. This recording shows a period of bigeminal rhythm in which each stimulated beat was followed by spontaneous beat with a long coupling interval. The ECG recording was quiescent during the interval between the stimulated and spontaneous beats. We have previously shown that this rhythm is associated with the occurrence of oscillatory afterpotentials (Ferrier et al., 1985; Ferrier and Guyette, 1991). Figure 6B was recorded from a pinacidil-pretreated preparation in reperfusion. The first stimulus triggered a rapid tachycardia that was terminated by the second stimulus. The third stimulus triggered a sustained rapid tachycardia. Here, the spontaneous beats occurred with very short coupling intervals, and the ECG activity was continuous throughout the arrhythmia. During these arrhythmias, spontaneous endocardial and epicardial action potentials occurred alternately. We have previously demonstrated that this type of arrhythmia is most likely caused by transmural re-entry (Ferrier and Guyette, 1991). In the present study, pinacidil caused a change in the type of arrhythmia observed in reperfusion. In the absence of pinacidil pretreatment, nearly all spontaneous beats were coupled to stimulated beats and exhibited characteristics typical of arrhythmias initiated by oscillatory afterpotentials (Fig. 6C). In contrast, none of the pinacidil-pretreated preparations exhibited arrhythmias with this configuration. In pinacidil-pretreated tissues, 100% of preparations exhibited rapid, closely coupled arrhythmias with characteristics typical of transmural re-entry (Fig. 6C). Arrhythmias with a re-entrant pattern were only observed in one preparation in the absence of pinacidil pretreatment.

Discussion

The goals of this study were to 1) investigate whether pretreatment of ventricular tissues with pinacidil would be pro- or antiarrhythmic in the setting of ischemia and reperfusion, and 2) determine the cellular electrophysiological basis for the effects of pinacidil pretreatment on arrhythmias. Pinacidil pretreatment resulted in enhanced abbreviation of APD₉₀ during ischemia and reperfusion. In addition, pinacidil pretreatment significantly slowed transmural conduction and increased the incidence of transmural conduction block during ischemia. These effects of pinacidil pretreatment were accompanied by a significant increase in arrhythmias in ischemia and reperfusion in the isolated tissue model of ischemia and reperfusion used in the present study. These arrhythmias exhibited properties characteristic of re-entrant rhythms. Thus, the effects of pinacidil were predominantly proarrhythmic in this model.

Pinacidil pretreatment resulted in several changes in electrophysiological properties of ventricular muscle. In the absence of ischemic conditions, pinacidil shortened both endocardial and epicardial APD₉₀, as expected on the basis of previous reports (Gross and Peart, 2003). Interestingly, although pinacidil was washed out before initiation of ischemic conditions, pretreatment markedly potentiated action-potential shortening in ischemia and early reperfusion. The mechanism for this effect is not clear. Abbreviation of APD₉₀ was enhanced, although the effects of pinacidil pretreatment on APD₉₀ were dissipating by the time that ischemia was initiated. Therefore, the concentration of pinacidil remaining in the tissue must have been very low. Thus, the potentiation of action-potential shortening in ischemia suggests that ischemia must enhance pinacidil's actions. One might propose that the partial depolarization that occurs in this preparation during ischemia (Pasnani and Ferrier, 1992) might augment effects of pinacidil on APD₉₀. However, unlike several other potassium channel openers, pinacidil's action is not voltage-dependent (Magnon et al., 1998). Another possibility is that ischemia sensitizes K_ATP channels to pinacidil. Indeed, Fan et al. (1990) have shown that the effects of pinacidil on K_ATP channels are inversely related to intracellular ATP levels. Thus, if ATP levels fall during ischemic conditions, the tissues may be sensitized so that residual amounts of pinacidil in the tissue may be sufficient to further shorten APD₉₀ in ischemia.

In addition to enhancing abbreviation of APD₉₀, pinacidil pretreatment also further depressed transmural conduction in ischemia and early reperfusion. As a result, all pinacidil-pretreated preparations exhibited transmural conduction block in ischemia, whereas only 33% of control preparations exhibited transmural block. In contrast, endocardial conduction was not depressed by pinacidil pretreatment. This differential effect of pinacidil pretreatment on endocardial and transmural conduction may reflect differences in intercellular resistance and fiber orientation. Transmural conduction is largely perpendicular to fiber orientation, whereas endocardial spread of activity can occur along fibers through end-to-end connections. Since cells have a higher density of gap junctions at end-to-end connections compared with side-to-side connections between cells, axial resistance is lower for propagation parallel to fibers than for propagation transverse to fiber orientation (Clerc, 1976). Activation of K_ATP channels by pinacidil would be expected to decrease membrane resistance, which would shorten the space constant. This may further impair propagation transverse to fiber orientation, where axial resistance is high.

Interestingly, pinacidil has been shown to inhibit gap junction communication in astrocytes (Velasco et al., 2000). If pinacidil pretreatment also inhibits gap junctions in cardiac tissue, this may reduce the number of available gap junctions to less than that required to maintain conduction transverse to fiber orientation. On the other hand, the high density of gap junctions at end-to-end connections between cells may minimize effects on conduction parallel to fiber orientation. The mechanism by which pinacidil affects gap junctions is not clear; however, pinacidil has recently been shown to uncouple oxidative phosphorylation by reducing the proton gradient across the inner mitochondrial membrane (Holmuhamedov et al., 2004). This action would be expected to further reduce ATP levels in ischemia and might therefore decrease Ca²⁺ uptake by the sarcoplasmic reticulum. This, in turn, would cause an increase in free cytosolic Ca²⁺, which would be predicted to further increase gap junction resistance. Indeed, in previous studies, we have linked selective depression of transmural conduction in ischemia to intracellular Ca²⁺ overload and uncoupling of gap junctions (Ferrier and Guyette, 1991; Li and Ferrier, 1991; Thomas et al., 1995, 1996).

In earlier studies with this model of ischemia and reperfusion, we identified two types of arrhythmias (Ferrier and Guyette, 1991). Rhythms in which spontaneous beats were coupled to stimulated beats at long intervals during which the ECG was quiescent were attributed to triggered activity initiated by oscillatory afterpotentials (delayed afterdepolarizations) (Fig. 6A). In contrast, arrhythmias with very closely coupled beats, alternate activation of endocardium and epicardium, and continuous ECG activity bridging the interval between sequential beats during the arrhythmia were ascribed to transmural re-entry (Fig. 6B). In the present study, pinacidil pretreatment clearly promoted rapid arrhythmias with characteristics of transmural re-entry. This observation supports findings in perfused heart preparations in which pinacidil treatment promoted very rapid ventricular tachyarrhythmias (D'Alonzo et al., 1998; Fischbach et al., 2003, 2004).

In contrast to the studies above, pinacidil pretreatment of perfused guinea pig right ventricles abolished tachyarrhythmias (Cole et al., 1991). It is not clear why pinacidil promoted arrhythmias in some studies but suppressed rapid arrhythmias in the latter study. However, the present study used 100 μM pinacidil for pretreatment, whereas the study by Cole et al. (1991) investigated effects of 1 to 10 μM pinacidil. On the other hand, the studies by Fischbach et al. (2003, 2004) and D'Alonzo et al. (1998) reported proarrhythmic effects in preparations pretreated with 1 to 30 μM pinacidil. The mechanism of arrhythmia present in these models also may affect the outcome of pretreatment with pinacidil. Our study suggests that pinacidil suppresses arrhythmias with characteristics of afterpotentials but promotes rapid re-entrant arrhythmias. If the arrhythmias generated in the Cole et al. (1991) study were mediated by afterpotentials, abbreviation of APD₉₀ may have suppressed arrhythmias by decreasing sarcoplasmic reticulum Ca²⁺ overload (Clusin, 2003). In contrast, the preparation used in the present study provides a substrate for transmural re-entry (Ferrier and Guyette, 1991; Li and Ferrier, 1991; Pasnani and Ferrier, 1992; Heisler and Ferrier, 1996). In this setting, abbreviation of APD₉₀ would be expected to promote rapid arrhythmias caused by transmural re-entry (Ferrier and Guyette, 1991; Li and Ferrier, 1991; Pasnani and Ferrier, 1992; Heisler and Ferrier, 1996).

The relationship between the cellular effects of pinacidil pretreatment and induction of transmural re-entry is shown schematically in Fig. 7. Figure 7A shows the relationships between conduction times and ERP in untreated preparations at 5 min of reperfusion. Values are based on mean measured data, and ERP is estimated from APD₉₀, as these were shown to closely parallel one another in the present study. When action potentials were initiated by stimulation of the endocardium, mean conduction time to the epicardial recording site was 30 ms. Based on previous studies with this preparation (Ferrier and Guyette, 1991; Li and Ferrier, 1991; Pasnani and Ferrier, 1992; Heisler and Ferrier, 1996), we estimated the return conduction time to be 10% longer, which gives a round-trip conduction time of 63 ms. Since the mean ERP of the endocardial tissue was 83 ms, the likelihood of activity re-exciting the endocardium and completing a reentrant loop would be very low. In contrast, after pretreatment of preparations with pinacidil, the relationship between conduction times and ERP was altered (Fig. 7B). The mean transmural conduction time from endocardium to epicardium was 32 ms. Round-trip conduction time was estimated at 67 ms, which is longer than the mean ERP of 57 ms in pinacidil-pretreated preparations. These conditions would greatly increase the likelihood that endocardial reactivation would be successful and re-entry would occur.

Previous studies have reported either pro- or antiarrhythmic effects of K_ATP activation by pinacidil pretreatment in advance of ischemia and reperfusion (Cole et al., 1991; D'Alonzo et al., 1998; Tanno et al., 2001; Fischbach et al., 2003, 2004). The present study demonstrates a cellular electrophysiological basis for these differing observations by showing that pinacidil pretreatment in an isolated tissue model of ischemia and reperfusion promotes rapid re-entrant arrhythmias but reduces the incidence of arrhythmias with characteristics of afterpotential-mediated activity. The increased incidence of re-entrant arrhythmias was associated with changes in APD₉₀ and transmural conduction that would promote transmural re-entry. Abbreviation of APD₉₀ and ERP can likely be attributed to activation of K_ATP channels. Prolongation of transmural conduction times also may involve activation of K_ATP channels; however, it is possible that effects of pinacidil on gap junctions also contribute to this effect. The decrease in the incidence of afterpotential-mediated arrhythmias also might be linked to the abbreviation of APD₉₀, which would be expected to reduce sarcoplasmic reticulum Ca²⁺ overload. However, in this model of ischemia and reperfusion, pinacidil pretreatment was pre-dominantly proarrhythmic because it significantly increased re-entrant arrhythmias. In addition, pinacidil pretreatment promoted transmural conduction block. These proarrhythmic effects might limit the usefulness of this agent in cardioprotection in ischemic heart disease. In future studies, it will be interesting to determine whether agents that are more selective for mitochondrial K_ATP channels will provide cardioprotection without promoting re-entrant arrhythmias.