INTRODUCTION

Top Abstract Introduction Methods Results Discussion References

EXPOSURE OF THE HEART to one or more brief episodes of coronary artery occlusion, termed "ischemic preconditioning," markedly attenuates the severity of the deleterious effects of a subsequent ischemic event (19, 20). Initial studies of ischemic preconditioning focused on reduction of infarct size. However, Shiki and Hearse (23) demonstrated that an initial 5-min coronary occlusion significantly reduced the incidence of reperfusion arrhythmias generated by a second ischemic period in in situ rat hearts. It is now well established that ischemic preconditioning protects against both ischemia and reperfusion arrhythmias in a variety of species (23, 27). Although ischemic preconditioning clearly decreases the incidence of ventricular arrhythmias, it does not change the time course of arrhythmias (12, 27). This contrasts with the effects of preconditioning on infarct size, where a delay in myocardial necrosis is observed (20).

The mechanisms by which preconditioning exerts antiarrhythmic effects are poorly understood. Substances released during ischemia such as nitric oxide (30), prostanoids (29), and adenosine (1, 19) have been implicated, although their roles remain controversial (17, 18). A role for activation of ATP-sensitive potassium current in the cardioprotective and antiarrhythmic effects of preconditioning also has been suggested (3, 28).

Arrhythmias occurring in response to acute ischemia are believed to be caused primarily by reentry, whereas arrhythmias occurring upon reperfusion may be caused by oscillatory afterpotentials (OAP, also called delayed afterdepolarizations) in addition to reentry (6, 13, 22). OAP occur in response to Ca²⁺ overload and have been demonstrated in isolated tissue models of reperfusion (6, 7). There are multiple mechanisms for reentry. However, a delay in propagation that is longer than the effective refractory period (ERP) of the tissue that must be reexcited to complete the reentrant circuit appears to be a central requirement for reentry. It is possible that Ca²⁺ overload may also be a factor leading to reentry in reperfusion. Elevation of intracellular Ca²⁺ may cause conduction delays by reducing intercellular coupling (15, 16). Very little is known about the electrophysiological basis for the protective effect of preconditioning on arrhythmias. It is possible that ischemic preconditioning temporarily alters parameters such as ERP, action potential duration (APD), conduction velocity, or Ca²⁺ loading, which may contribute to generation or maintenance of arrhythmias. It is also possible that ischemic preconditioning modifies how these parameters change in response to a subsequent ischemia and reperfusion.

We have developed an isolated guinea pig ventricular tissue model of ischemia and reperfusion, which reliably generates arrhythmias and allows measurement of cellular electrophysiological activity with intracellular microelectrodes (6, 15). Late premature beats with characteristics attributable to OAP are observed in some preparations during reperfusion. However, closely coupled premature beats and ventricular tachycardia (VT) are the most common arrhythmias in both ischemia and reperfusion. We have presented evidence in previous studies which indicates that transmural reentry is the most likely mechanism for these rapid arrhythmias (6, 15). This model has been used to study cellular actions of antiarrhythmic and experimental drugs under conditions of ischemia and reperfusion (15, 16, 21). In the present study, we developed protocols that allowed us to utilize this preparation as an isolated tissue model of ischemic preconditioning.

The goals of the present study were 1) to determine whether brief periods of simulated ischemia and reperfusion protect against arrhythmias during a second, longer simulated ischemia and reperfusion in this model, 2) to determine how preconditioning modifies the cellular electrophysiological responses of this preparation to simulated ischemia and reperfusion, and 3) to identify possible antiarrhythmic effects of preconditioning in this isolated tissue model.

	METHODS

Top Abstract Introduction Methods Results Discussion References

Experiments in this study utilized an isolated tissue model of ischemia and reperfusion described in detail previously (6, 15). Briefly, guinea pigs weighing 350-650 g were killed by stunning followed immediately by carotid exsanguination, in accordance with the guidelines of the Canadian Council on Animal Care. Hearts were rapidly removed from the chest and were rinsed in "normal" Tyrode solution at 37°C. The composition of this Tyrode solution in mmol/l was 129 NaCl, 4 KCl, 0.9 NaH₂PO₄, 20 NaHCO₃, 0.5 MgSO₄, 2.5 CaCl₂, and 5.5 glucose. The pH was 7.4 when the solution had been gassed with 95% O₂-5% CO₂ (PO₂ = 487 mmHg; PCO₂ = 43 mmHg). Segments of right ventricular free wall were dissected in this solution at 37°C and were transferred to a tissue bath also at 37°C. For viability of intramural tissues, it was found essential that all dissections and preparation be conducted at 37°C. Preparations dissected at room temperature and then rewarmed to 37°C showed complete transmural conduction block and therefore were unable to generate transmural reentry (6). Right ventricular free walls were mounted on edge to permit simultaneous recording of endocardial and epicardial intracellular activity with standard microelectrode techniques. A high-gain electrocardiogram (ECG) also was recorded with Ag/AgCl wires at opposite ends of the tissue bath. The ECG was used to confirm that the microelectrode recordings were indicative of the activity of the whole preparation. Preparations were held in place ~1 mm above the floor of the tissue bath to allow circulation of Tyrode solution around all surfaces. The tissue bath was perfused with Tyrode solution at a flow rate of 30 ml/min.

Preparations were stimulated throughout the experimental protocol with trains of 15 rectangular pulses, 3 ms in duration, at two times diastolic threshold intensity. Stimulus trains were delivered to the endocardial surface of preparations and were separated by 3-s pauses. The basic cycle length during trains of stimuli was 500 ms. To determine the endocardial ERP, a premature stimulus was introduced after the last pulse of each train. Initially, the stimulus was delivered during the refractory period of the last regular driven beat. The test interval was then gradually prolonged in increments of 5-10 ms until the test stimulus induced an action potential that propagated to a recording electrode. The interval between the last regular stimulus and the earliest successful test stimulus was taken as a measure of the ERP.

Biological signals and a record of stimulation were displayed on an oscilloscope (model 5110; Tektronix) and were recorded on a microcomputer with a continuous data acquisition program (Axotape, Axon Instruments) following analog to digital conversion (TL1-125; Axon Instruments). Measurements of experimental data were made from recordings played back on the computer.

Experimental design. All preparations were equilibrated for at least 60 min in normal Tyrode solution. Control recordings were made at the end of this period, at which time all electrical parameters had stabilized. Preparations then were subjected to 15 min of superfusion with Tyrode solution modified to simulate selected conditions occurring in myocardial ischemia (hypoxia, hypercapnia, hyperkalemia, acidosis, lactate accumulation, and substrate deprivation; see Refs. 6 and 7). This "ischemic" Tyrode solution had the following millimolar composition: 123.0 NaCl, 8.0 KCl, 0.9 NaH₂PO₄, 6.0 NaHCO₃, 2.5 CaCl₂, 0.5 MgSO₄, and 20.0 sodium lactate. Ischemic Tyrode solution had a pH of 6.8 when the solution was bubbled with a 90% N₂-10% CO₂ gas mixture at 37°C (PO₂= 48 mmHg; PCO₂ = 64 mmHg). After the ischemic period, preparations were reperfused with normal Tyrode solution for 30 min. These durations of ischemia and reperfusion were chosen because, in this model, 15 min of ischemia conditions results in ~70-90% of preparations exhibiting ischemia and reperfusion arrhythmias and because electrical activity recovers to stable values within 30 min of reperfusion.

Preparations were subjected to one of three protocols to determine the electrophysiological effects of ischemic preconditioning. A control group of preparations was exposed to ischemic Tyrode solution for 15 min and then was reperfused for 30 min with normal Tyrode solution. The initial equilibration period in normal Tyrode solution was lengthened for the control group so that the total duration of the experiment was approximately the same as that for the preconditioned groups. Two groups of preparations were preconditioned either with a 2-min exposure to ischemic solution followed by a 5-min reperfusion or a 5-min exposure to ischemic solution followed by a 10-min reperfusion before being subjected to the same protocol used in the control group. These two preconditioning intervals were chosen because others have reported that a single 3- to 5-min ischemia reduces arrhythmias (23, 27) and because 5 min simulated ischemia in this model results in maximal changes in action potential configuration. In contrast, 2 min simulated ischemia causes only minimal electrophysiological changes and should provide a more moderate ischemic event for comparison.

The incidence of arrhythmias, APD at 90% repolarization, ERP, endocardial conduction time (CT), and transmural CT were measured during control, ischemic, and reperfusion periods. The interval between the stimulus and action potential upstroke at the endocardial electrode was taken as a measure of endocardial CT. The interval between the stimulus and the upstroke at the epicardial electrode was taken as a measure of transmural CT, although it is recognized that this measure includes components due to endocardial and epicardial spread of activity. Arrhythmias induced by ischemia and reperfusion have been documented in this model in previous publications and included premature ventricular beats (PVB) and VT. In this study, VT was defined as the occurrence of five or more consecutive PVBs.

The statistical significance of differences between groups for incidence of arrhythmias was determined using the ² test. Other data were compiled as means ± SE. Differences were determined with paired or unpaired Student's t-test or with the univariate mode of a repeated-measures analysis of variance in which multiple comparisons of means were made with the Bonferroni correction (SAS; SAS Institute). The latter analysis was needed because, although all preparations survived the entire protocol, some observations had missing data because of arrhythmias or development of conduction block. A P value less than 0.05 was considered statistically significant.

	RESULTS

Top Abstract Introduction Methods Results Discussion References

Induction of arrhythmias by simulated ischemia and reperfusion. Figure 1 shows representative recordings of activity from preparations exposed to ischemic conditions and reperfusion. In Fig. 1, A-F, traces show microelectrode recordings from endocardium and epicardium, the ECG, and a record of stimulus pattern. Figure 1A was recorded from a control preparation at the end of the equilibration period in normal Tyrode solution. Each stimulus initiated an impulse in the endocardium that arrived at the epicardial site after a delay. The ECG was biphasic and corresponded in time to the recorded action potentials. The ECG was quiescent throughout diastole.

View larger version (23K): [in this window] [in a new window]   Fig. 1.   Representative recordings from right ventricular preparations exposed to simulated ischemia and reperfusion. A-F: traces from top to bottom are intracellular recordings from endocardium (ENDO) and epicardium (EPI), electrocardiogram (ECG), and record of stimulation (STIM). Zero millivolts is indicated at the left of the action potential traces in each panel. A was recorded during the preischemic control period. B shows a bigeminal rhythm recorded at 1.5 min reperfusion after 15 min of simulated ischemia. C shows rapid tachycardia that developed at 2 min of reperfusion. With continued reperfusion, arrhythmias ceased (D). E, which was recorded from a different preparation, shows an example of transmural conduction block observed during simulated ischemia. F shows an example of sustained tachycardia induced during 5-min ischemic preconditioning period.

The preparation was then exposed to simulated ischemia for 15 min. This preparation did not exhibit arrhythmias during ischemic conditions; however, arrhythmias occurred when the preparation was reperfused with normal Tyrode solution. Figure 1, B and C, were recorded at 1.5 and 2 min of reperfusion, respectively, and illustrate two distinct types of arrhythmia seen upon reperfusion. Figure 1B shows examples of late PVB that occurred just in advance of the next stimulated beat. These PVB had characteristics suggesting that they were initiated by OAP (8): their occurrence depended on driven activity (i.e., they were triggered), their coupling intervals were ~80% of the driven cycle length, and the ECG was quiescent during the interval between the driven and ectopic beats. Shortly after this recording, rapid sustained VT occurred (Fig. 1C). Rapid arrhythmias and closely coupled PVB were associated with continuous ECG activity between beats and an alternating pattern of activation of endocardium and epicardium. These rapid arrhythmias were associated with marked prolongation of transmural CT and abbreviation of endocardial ERP and endo- and epicardial APD (quantitative data shown in APD and ERP and CT). In the example illustrated, VT briefly stopped spontaneously but began again when a late PVB was followed closely by a stimulated beat. Detailed evidence for transmural reentry as a likely mechanism for rapid VT and closely coupled PVB has been presented as part of a previous study conducted with this same model (6, 15). Figure 1D was recorded after 10 min of reperfusion. Arrhythmic activity had stopped, and the configuration of the action potentials and ECG had returned to near control.

Arrhythmias observed during ischemic conditions occurred as closely coupled premature beats or rapid tachycardia similar to those in Fig. 1C (not illustrated). Arrhythmias with characteristics of OAP rarely occurred during ischemic conditions. In addition, conduction block between endocardium and epicardium developed in 39% of preparations exposed to ischemic conditions. Figure 1E shows an example recorded from a different preparation than in Fig. 1, A-D, after 15 min of exposure to ischemic conditions. Both endocardial and epicardial microelectrode recordings showed a decreased resting potential with respect to their preischemic controls. The endocardial action potential was markedly abbreviated. Transmural conduction block had developed, thus activity initiated by stimulation of the endocardium was not recorded at the epicardial recording site. Although arrhythmias occurred in many preparations before transmural conduction block developed, arrhythmias were never observed during ischemic conditions when transmural conduction block had developed.

Most ventricular arrhythmias in this study occurred in response to the main ischemia and reperfusion. However, the short periods of ischemia and reperfusion used to precondition preparations also induced arrhythmias. Figure 1F was recorded from a preparation that was subjected to a 5-min preconditioning protocol. During the 5-min preconditioning ischemic periods, VT and PVB occurred in 50 and 13% of 15 preparations, respectively. Reperfusion after the 5-min preconditioning ischemia induced VT in 20% (3 of 15) and PVB in 7% of preparations (1 of 15). Arrhythmic activity was not observed during 2-min preconditioning periods in 15 preparations subjected to this protocol. However, the reperfusion period between the 2-min preconditioning ischemia and the subsequent main ischemic period was associated with VT in 13% (2 of 15) of preparations.

Figure 2 shows the incidence of arrhythmias occurring in response to exposure to ischemic conditions (Fig. 2A) and reperfusion (Fig. 2B) in control and preconditioned groups of preparations. The total incidence of preparations showing arrhythmias of any type is given for each group. In addition, the incidences of PVB and VT are presented separately. In control preparations (nonpreconditioned, n = 18), 67% of preparations exhibited arrhythmias (Fig. 2A). PVB were observed in 61% of control preparations, and VT occurred in 33% of preparations. The sum of these two values exceeds the total incidence of preparations showing arrhythmias (67%) because some preparations exhibited both types of arrhythmias. Upon reperfusion, arrhythmias occurred in 89% of control preparations (Fig. 2B). Reperfusion arrhythmias included PVB in 83% of preparations and sustained or nonsustained VT in 61% of preparations.

View larger version (21K): [in this window] [in a new window]   Fig. 2.   Effects of 2 and 5 min of ischemic preconditioning (PC) on the incidences of arrhythmias during subsequent 15-min ischemic period and after reperfusion. A shows the incidences of premature beats (PVB), ventricular tachycardia (VT), and arrhythmias of all types (Total) during 15-min ischemic period. The 5-min preconditioning protocol significantly decreased the incidence PVB but not VT or Total arrhythmias. B shows the incidence of reperfusion arrhythmias. The 2-min preconditioning protocol caused a trend toward a decrease in all categories of arrhythmia, whereas the 5-min preconditioning protocol significantly decreased the incidences of PVB, VT, and Total arrhythmias. * P < 0.05 and ** P < 0.01; n, no. of preparations.

Figure 2A also shows the incidence of arrhythmias during ischemic conditions after preconditioning. Neither preconditioning protocol significantly altered the incidence of total arrhythmias during ischemic conditions, although there was a trend toward a lower total incidence of arrhythmias in preparations preconditioned with a 5-min ischemic period. However, the 5-min preconditioning protocol significantly reduced the incidence of PVB induced by ischemia from 61% in the control group to 20% in the preconditioned preparations. A smaller nonsignificant reduction in the incidence of PVB from 61 to 40% was observed with the 2-min preconditioning protocol. Neither preconditioning protocol affected the incidence of VT during ischemia.

The incidence of reperfusion arrhythmias was affected more dramatically by preconditioning as shown in Fig. 2B. A large statistically significant antiarrhythmic effect was observed in preparations exposed to the 5-min preconditioning protocol. The incidence of total reperfusion arrhythmias decreased from 89% in control preparations to 47% in preconditioned preparations (P < 0.05), PVB decreased from 83 to 33% (P = 0.01), and VT decreased from 61 to 20% (P < 0.05). The 2-min preconditioning protocol caused a smaller nonsignificant decrease in the incidence of total arrhythmias as well as PVB and VT.

Transmural conduction block. Exposure of preparations to ischemic conditions resulted in gradual development of transmural conduction block. However, transmural conduction recovered early in reperfusion in all preparations exhibiting block. In the absence of preconditioning, 39% of preparations subjected to ischemia developed complete (3rd degree) transmural conduction block during the 15-min period of ischemic conditions, as indicated in Fig. 3. Preconditioning with 2- or 5-min ischemic periods reduced the incidence of complete conduction block during ischemic conditions to 20 and 7%, respectively. The reduction in the incidence of transmural conduction block observed with the 5-min preconditioning protocol was statistically significant (P < 0.05).

View larger version (26K): [in this window] [in a new window]   Fig. 3.   Effects of ischemic preconditioning on the incidence of transmural conduction block during simulated ischemia. Preconditioning for 2 min caused a decrease in the incidence of conduction block occurring during the ischemic period. However, this trend was not statistically significant. Preconditioning for 5 min significantly decreased the incidence of conduction block during 15 min of ischemic period. * P < 0.05.

APD and ERP. Figure 4 illustrates the effects of ischemia and reperfusion on mean endocardial (Fig. 4A) and epicardial (Fig. 4B) APD in the absence and presence of preconditioning. An initial value for APD measured after the equilibration period but before preconditioning is shown in Fig. 4, A-C. Values of APD determined throughout the main periods of ischemia and reperfusion are plotted as a function of time beginning at the start of the main ischemic period. Exposure of nonpreconditioned preparations to simulated ischemia resulted in rapid and marked shortening of endocardial and epicardial APD. Maximal abbreviation of APD occurred within 5-10 min of onset of ischemic conditions and persisted for the rest of the ischemic period. Reperfusion was accompanied by a gradual prolongation of mean APD in both endocardium and epicardium. Preconditioning with either 2- or 5-min periods of ischemia did not alter changes in endocardial or epicardial APD seen in response to subsequent ischemia and reperfusion.

View larger version (16K): [in this window] [in a new window]   Fig. 4.   Effects of simulated ischemia and reperfusion on action potential duration at 90% repolarization (APD90) and effective refractory (ERP) in control and preconditioned preparations. "C" on the abscissa designates data measured at the end of the initial equilibration period before preconditioning. APD90 was abbreviated during ischemic conditions and was prolonged upon repolarization in both endocardium (A) and epicardium (B). Ischemic conditions also caused a similar but smaller decrease in the duration of the endocardial ERP (C). ERP increased again upon reperfusion. Ischemic preconditioning for 2 or 5 min had no effects on the magnitudes or time courses of changes in APD90 or ERP. In A-C, symbol corresponds to an "n" of 14-18 for control, 12-15 for the 2-min preconditioning group, and 10-15 for the 5-min preconditioning group. Variation in n was the result of missing data when conduction block or arrhythmias occurred.

The effects of ischemia and reperfusion on endocardial ERP, with and without ischemic preconditioning, are shown in Fig. 4C. In the absence of preconditioning, ERP abbreviated rapidly during ischemic conditions and was gradually prolonged with reperfusion. Preconditioning for 2 or 5 min did not affect the response of endocardial ERP to ischemia and reperfusion significantly, although the recovery of ERP during early reperfusion tended to be more rapid in preparations subjected to 5 min of preconditioning compared with nonpreconditioned preparations.

The mean endocardial ERP was 99 ± 2 ms, and mean APD was 97 ± 5 ms in nonpreconditioned preparations during the preischemic period. Thus ERP corresponded closely to APD before onset of ischemic conditions. After 10 min of ischemic conditions, APD abbreviated to 51 ± 5 ms, whereas ERP only shortened to 70 ± 3 ms. Therefore, ischemic conditions resulted in significant postrepolarization refractoriness (P < 0.01). After 30 min of reperfusion, APD and ERP prolonged to 123 ± 6 and 115 ± 4 ms, respectively, thereby eliminating postrepolarization refractoriness. Because ischemic preconditioning did not alter the responses of either APD or ERP to ischemic conditions, preconditioning did not affect postrepolarization refractoriness.

Because activation of ATP-sensitive potassium current may be involved in ischemic preconditioning (3, 28), we also determined whether APD and ERP changed during the 5-min preconditioning protocol. These measurements were made only in experiments in which neither VT nor PVB occurred during the preconditioning ischemic period to avoid changes in APD and ERP attributable to changes in heart rate rather than ischemic conditions. Table 1 shows that the 5-min preconditioning protocol resulted in significant abbreviation of epicardial APD as well as endocardial APD and ERP. By the end of the 10-min reperfusion period separating preconditioning from the main ischemic period, the values of these parameters were not significantly different from control. Table 1 also shows that preconditioning significantly slowed transmural but not endocardial conduction.

CT. Figure 5 shows the effects of ischemic preconditioning on the responses of endocardial and transmural conduction to ischemic conditions and reperfusion. Figure 5A presents effects on endocardial CT measured as the time between the endocardial stimulus and the upstroke of the endocardial action potential. An initial value for endocardial CT measured after the equilibration period but before preconditioning is shown in Fig. 5, A and B. Values of endocardial CT determined throughout the main periods of ischemia and reperfusion are plotted as a function of time beginning at the start of the main ischemic period. Figure 5A shows that endocardial CT was not significantly affected by exposure to ischemic conditions or reperfusion. Furthermore, ischemic preconditioning did not significantly modify endocardial CT during preischemic, ischemic, or reperfusion periods.

View larger version (16K): [in this window] [in a new window]   Fig. 5.   Effects of ischemic conditions and reperfusion on endocardial and transmural conduction times (CT) in control and preconditioned preparations. A shows that endocardial CT was not significantly affected by exposure to ischemic conditions or reperfusion. Ischemic preconditioning for 2 or 5 min did not affect the response of endocardial CT to ischemia or reperfusion. B shows that transmural CT prolonged progressively with exposure to ischemic conditions and recovered gradually with reperfusion. Preconditioning for 2 or 5 min significantly attenuated the prolongation of conduction observed with ischemic conditions. * P < 0.05. Number of preparations in each group is the same as in Fig. 4.

In contrast to endocardial CT, transmural CT was markedly prolonged by ischemic conditions (Fig. 5B). In nonpreconditioned preparations, transmural CT more than doubled from 27 ± 2 ms immediately before ischemic conditions (time 0) to 58 ± 3 ms after exposure to ischemic conditions for 15 min (P < 0.01). Transmural CT remained significantly prolonged during the first 2 min of reperfusion when most reperfusion arrhythmias appear. With continued reperfusion, transmural CT gradually recovered toward control values.

Ischemic preconditioning greatly attenuated prolongation of transmural CT in response to ischemic conditions. Before onset of ischemic conditions, mean transmural CT in both preconditioned groups (26 ± 1 ms) was not significantly different from nonpreconditioned preparations. Preconditioning slowed the rate of development of prolonged transmural CT during ischemic conditions and resulted in a lower maximum CT at 15 min of ischemic conditions. Transmural CT was significantly shorter in both preconditioned groups compared with control, nonpreconditioned preparations throughout the ischemic period. By 15 min of ischemic conditions, transmural CT had increased only to 44 ± 3 and 41 ± 1 ms in preparations preconditioned for 2 and 5 min (P < 0.01 for both groups compared with control). In addition, transmural CT remained significantly different from control during the first 2 min of reperfusion, when most reperfusion arrhythmias first appear. Preconditioning for 5 min tended to cause greater attenuation of prolongation of transmural conduction than preconditioning for 2 min; however, this difference was not statistically significant.

	DISCUSSION

Top Abstract Introduction Methods Results Discussion References

In the present study, we have developed an isolated tissue model of ischemic preconditioning that permits correlation of the occurrence of arrhythmias with cellular electrophysiological changes. Brief exposure of preparations to simulated ischemia in advance of a more prolonged ischemic period followed by reperfusion resulted in significant reduction of arrhythmias in response to the main ischemia and reperfusion. The antiarrhythmic effects of preconditioning were more marked with reperfusion arrhythmias than with those occurring during the ischemic period and increased with the duration of ischemic preconditioning. Because the relationship between incidence of reperfusion arrhythmias and duration of the main ischemic period is "bell shaped," the protective effect of preconditioning on reperfusion arrhythmias could reflect a rightward shift in the bell-shaped curve (12). However, Lawson et al. (12) investigated this question directly in perfused rat hearts and saw no shift in the relationship between incidence of arrhythmia and duration of ischemia in response to preconditioning and concluded that the decrease in reperfusion arrhythmias reflected a true antiarrhythmic effect. Thus antiarrhythmic effects of preconditioning were demonstrable in this isolated tissue model and resembled those reported in in situ and in isolated perfused heart preparations in which preconditioning was induced by brief occlusions of coronary arteries (12, 23, 27).

The present study utilized simulated ischemia for both the preconditioning periods and the main ischemic period. In previous studies, we have demonstrated that this model of simulated ischemia closely mimics the electrophysiological effects observed in true ischemia and reperfusion. In addition, we have found that this model accurately mimics or predicts the pro- and antiarrhythmic effects of a wide range of drugs. However, when interpreting the present observations, it is important to note that many elements of true ischemia are not included in this model. Although this may be considered a limitation of the present model, it is also important that short exposures to the simulated ischemia did precondition the myocardium as demonstrated by the protective effects. Furthermore, the use of simulated ischemia has the advantage that roles of specific elements in preconditioning can be evaluated by adding or omitting specific conditions. For example, one might propose that preconditioning in the present model may involve brief intracellular acidosis followed by reperfusion with solution with physiological pH. This could result in enhanced Na⁺-H⁺ and Na⁺-Ca²⁺ exchange, leading to a brief rise in intracellular Ca²⁺ and activation of Ca²⁺-dependent protein kinase C. In future studies with this model, each of these steps can be manipulated to test this and other possible cellular mechanisms involved in preconditioning.

In the present study, ischemic preconditioning resulted in significant reduction in the incidence of PVB and transmural conduction block in ischemia and PVB plus VT in reperfusion. The antiarrhythmic effects of ischemic preconditioning were not accompanied by changes in the responses of endocardial or epicardial APD to ischemic conditions or reperfusion. Similarly, endocardial ERP of preconditioned preparations was not significantly different from ERP of control preparations at any point in ischemia and reperfusion, although there was a trend for more rapid recovery of ERP during early reperfusion in preparations preconditioned for 5 min. The most pronounced electrophysiological effect accompanying the antiarrhythmic effect of preconditioning was an attenuation of the marked depression of transmural conduction observed with ischemic conditions and early reperfusion. This effect was specific for transmural conduction. Endocardial CT was affected minimally by ischemic conditions as reported in earlier studies of this model (6, 15, 16, 21), and preconditioning for neither 2 nor 5 min caused any significant changes in endocardial CT relative to control.

The effect of preconditioning on transmural CT may provide the basis for the antiarrhythmic effect of preconditioning in this model of global ischemia and reperfusion. Most reperfusion arrhythmias appear during the first 2 min of reperfusion in this model. During this period, ERP is greatly abbreviated, transmural CT remains prolonged, and epicardial excitability rapidly recovers (transmural conduction block declines; see Refs. 6 and 15). In previous studies, we have presented evidence that the likeliest mechanism for most of the arrhythmias occurring in this model is transmural reentry (6, 15). For transmural reentry to occur, electrical impulses must propagate across the ventricular wall two times (from endocardium to epicardium and back). The total CT for this round trip must exceed the duration of the endocardial ERP for reexcitation of the endocardium to be successful. This is shown schematically in Fig. 6A, which shows the relationship between mean ERP and transmural CT at 2 min of reperfusion in control preparations. Mean anterograde CT from the site of endocardial stimulation to epicardium was 56 ms. Retrograde CT, 65 ms, was ~16% longer than anterograde CT, which resulted in a total round trip CT of 121 ± 15 (SE) ms. Thus total CT greatly exceeded the mean endocardial ERP of 77 ± 6 ms, and the probability of transmural reentry would be predicted to be high. Figure 6B illustrates the relationship between transmural CT and ERP in preparations subjected to 2 min of ischemic preconditioning. Total round trip CT was shortened to 89 ± 6 ms by preconditioning and exceeded ERP of 74 ± 6 by a much smaller value than in control preparations. Because these values are means with finite variances, the probability of successful reentry should be lower than in the control group. The relationship between transmural CT and ERP for preparations preconditioned for 5 min is shown in Fig. 6C. In this group, mean round trip transmural CT was 78 ± 6 ms, which was shorter than the mean ERP of 83 ± 7 ms. Again, because these are mean values, this does not eliminate the possibility of transmural reentry but should greatly decrease the probability of this event occurring.

View larger version (13K): [in this window] [in a new window]   Fig. 6.   Schematic representing relationships at 2 min of reperfusion between transmural CT and endocardial ERP in control preparations (A) and preparations preconditioned for 2 (B) and 5 (C) min. Numbers give mean values of ERP and transmural CT. See text for explanation.

The mechanisms by which ischemic conditions selectively prolong transmural CT and the mechanisms by which preconditioning attenuates this defect are not entirely clear. The differential effects of ischemic conditions on transmural and endocardial CT most likely represent differences in conduction transverse and longitudinal with respect to fiber orientation rather than differences in conditions between intramural and superficial tissues. Evidence for this comes from earlier studies of thin epicardial slices in which we measured conduction of impulses transverse and longitudinal to fiber orientation simultaneously (15, 16). Differences attributable to intramural versus superficial conditions are eliminated in these preparations, yet ischemic conditions caused significant prolongation of CT transverse to fiber orientation, whereas CT for the same impulse in the longitudinal direction was affected minimally. Because myocardial fibers are arranged in concentric spirals around the ventricles, transmural conduction must also proceed transverse to fiber orientation.

In perfused rabbit papillary muscles, severe ischemic damage is associated with cellular electrical uncoupling likely caused by closure of gap junctions (2, 11). Uncoupling may occur in response to elevation of intracellular free Ca²⁺, which is known to occur within a few minutes of onset of ischemia (14, 24) and which causes decreased conductance through gap junctions in the heart (5). We have proposed previously that ischemic conditions might prolong transmural conduction by increasing gap junction resistance through elevation of intracellular Ca²⁺ (15, 16). We proposed that this effect would disrupt conduction transverse to fiber orientation more than in the longitudinal direction because of the lower density of gap junctions between cells in the transverse direction (15, 16). In support of this hypothesis, we have found that agents that can decrease intracellular Ca²⁺ attenuate prolongation of transmural conduction by ischemic conditions (15, 16). It is possible that preconditioning attenuates prolongation of transmural CT through a similar effect. Indeed, Steenbergen et al. (25) have reported that preconditioning delays development of cytosolic Ca²⁺ overload in the myocardium. Tan et al. (26) have reported that ischemic preconditioning in perfused rabbit papillary muscles delays cellular electrical uncoupling.

Tan et al. (26) reported that conduction velocity decreased gradually to ~70% of control during 15 min of coronary occlusion in rabbit papillary muscles. Conduction in their study was measured parallel to fiber orientation and therefore might be comparable to endocardial conduction reported here. Because control endocardial CT was ~7 ms in the present study, a similar depression in conduction to that reported by Tan et al. (26) would amount to only 1 or 2 ms, which is compatible with our observations. Tan et al. (26) also reported that preconditioning did not affect the changes in conduction velocity that were observed during ischemia in papillary muscles. Data for conduction transverse to fiber orientation were not presented in their study (26).

Activation of ATP-sensitive potassium channels during ischemia may provide protective effects on mechanical and electrical activity upon reperfusion (3) and also may be involved in the protective effects of preconditioning on intercellular coupling (26). In addition, APD has been reported to shorten more rapidly in response to ischemia in preconditioned preparations (26). Glyburide blocks abbreviation of APD in response to the ischemic conditions in the model used in this study, which suggests that ATP-sensitive potassium current also is activated by our ischemic conditions (21). Because our preconditioning ischemic period significantly abbreviated both endo- and epicardial APD (Table 1), activation of ATP-sensitive potassium current might also be a step in the antiarrhythmic effects of preconditioning in this model. However, we observed no change in either endo- or epicardial APD during the main ischemic period after preconditioning. Thus greater activation of the ATP-sensitive potassium current during the main ischemic period did not appear to be a necessary step leading to the antiarrhythmic effects in the present model.

In contrast to our study, Tan et al. (26) reported greater abbreviation of APD in response to the main ischemia in preconditioned preparations. This might reflect the longer ischemic period used for preconditioning in their study or possibly a species difference. In addition, the end points of the studies differ greatly. Marked cellular uncoupling measured in rabbit papillary muscles correlates with irreversible ischemic cellular damage (2), whereas reperfusion in the present model induces arrhythmias but is followed by virtually complete recovery of normal electrical activity (6, 15). Furthermore, activation of ATP-sensitive potassium channels might relate more closely to protective effects of preconditioning on cellular damage and infarction rather than on electrical function (9, 10, 17, 28). For example, although activation of ATP-sensitive potassium channels may play a role in reduction of infarct size by preconditioning in dogs (9,10), Vegh et al. (28) were unable to demonstrate a protective role of these channels in suppression of arrhythmias by preconditioning in a similar canine model.

The antiarrhythmic effects of preconditioning might also involve other agents, including prostanoids (29), nitric oxide (30), or adenosine (19). We recently have shown that adenosine, acting at adenosine A1 receptors, decreases signs of Ca²⁺ overload in an isolated guinea pig ventricular myocyte model of ischemia and reperfusion (4). The mechanism of this protective effect is not clear. Preconditioning might decrease or delay Ca²⁺ overload by affecting Ca²⁺ influx, efflux, or redistribution. Interestingly, Zucchi et al. (31) have presented evidence that preconditioning reduces expression of ryanodine-sensitive release channels in rat plus the rate constant for Ca²⁺-induced Ca²⁺ release from the sarcoplasmic reticulum (31). They suggested that these changes might reduce elevation of free cytosolic Ca²⁺ during ischemia. A clearer understanding of how adenosine or other agents participate in the antiarrhythmic effects of preconditioning will require further experimental studies.

Both VT and PVBs induced in this model of ischemia and reperfusion may occur secondary to Ca²⁺ overload (6, 15, 16). Our observations with respect to transmural CT suggest that ischemic preconditioning may protect against arrhythmias by attenuating electrophysiological effects associated with Ca²⁺ overload in ischemia and reperfusion. Furthermore, these observations suggest that agents that reduce Ca²⁺ overload in ischemia and reperfusion should be explored for possible use as pharmacological preconditioning agents in the setting of ischemic heart disease.