Guinea-pigs of either sex weighing 300 to 400 g were euthanized under anesthesia with sodium pentobarbital 125 mg/kg i.p. The hearts were quickly removed and a standard strip (16×8 mm) of right ventricular myocardium was dissected from the free wall and placed in a special perfusion chamber bath (volume of 5 ml) partitioned into 2 compartments by a thin latex membrane. This latex membrane is perforated at its bottom allowing the ventricular strip to be passed through and so to be divided into two zones, called the normal zone and the altered zone, respectively. The preparation was pinned, endocardial surface upward, to the silicon base of the bath.#

This double compartment allowed the two parts of the same ventricular strip to be independently superfused at a rate of 2 ml/min. The continuity of the partition was tested at the end of each experiment by means of dye injection (methylene blue) into one of the compartments.#

Temperature at the level of the double chamber, including that of incoming fluids, was controlled and maintained to 36.5±0.5 °C by a circulating thermostat-controlled bath (Polystat 5HP, Bioblock, France).#

Studies were performed in Tyrode's solution oxygenated with a mixture of 95% oxygen and 5% carbon dioxide (pO2 and pCO2 at 510±20 and 34±2 mm Hg, respectively). The composition of the Tyrode's solution was (in mmol/l): Na⁺, 135; K⁺, 4; Ca²⁺, 1.8; Mg²⁺, 1.0; H2PO4⁻, 1.8; HCO3⁻, 25; Cl⁻, 117.8; and glucose, 5.5. The pH was 7.35±0.05 (fitted with diluted HCl). Modified Tyrode's solution mimicking ischemia were also used. Ischemic-like solution differed from the normal one by (1) increased extracellular potassium concentration (8 or 12 mmol/l), (2) decreased HCO3⁻ concentration (9 mmol/l) leading to decreased pH (6.90±0.05), (3) decreased pO2 (80 mm Hg instead of 510 mm Hg) by replacement of 95% O2 and 5% CO2 by 95% N2 and 5% CO2, and (4) complete withdrawal of glucose. We also used a modified ischemic-like solution (with HCO3⁻ concentration of 9 mmol/l, pH 6.90±0.05 and pO2 80 mm Hg) where extracellular potassium concentration was normal (4 mmol/l). Tyrode's solution modifications are believed to reproduce in vitro the electrophysiological abnormalities induced in vivo by ischemia (Morena et al., 1980) and have been used previously (Rouet et al. 2000; Ducroq et al. 2005).#

Dofetilide (gift of Pfizer Central Research, Sandwich, UK), was first diluted in ethanol–HCl (0.05 N) and then in Tyrode's solution at 10 nmol/l. Ethanol 0.05 N had no effect on electrophysiological parameters. The two compartments, at appropriate time intervals, were superfused with drug containing solutions during both ischemic-like and reperfusion periods.#

The preparations were stimulated at a frequency of 1 Hz via bipolar Teflon-coated steel wire electrodes positioned near the two ventricular strip extremities either in the normal or the altered zone. Stimulation was applied either in one or the other half of the muscle preparation with a home-built commutator. Stimuli were rectangular pulses, 2 ms in duration and twice the diastolic threshold intensity (around 2–2.5 V) delivered by a programmable stimulator (SMP-310, Biologic, France). During the protocol, stimulation was stopped whenever spontaneous repetitive responses occurred, and an extrastimulus, 2 ms in duration and twice the intensity of the basic stimulus, was applied every four stimulations in an attempt to elicit triggered repetitive responses by a progressive increase in 5 ms steps of the time interval between the stimulus and the extrastimulus. The first interval between the stimulus and the extrastimulus was 50 ms, short enough to consider the ischemic-like phase induced reduction of action potential durations at 90% of repolarization.#

Transmembrane action potentials were recorded simultaneously in both ventricular regions by intracellular glass microelectrodes filled with KCl 3 mol/l (tip resistance 10 to 30 MΩ) coupled to Ag/AgCl microelectrode holders leading to the double input stage of a high impedance capacitance-neutralizing home built amplifier. The two reference silver–silver chloride electrodes were positioned in the superfusate of each chamber, close to the preparation. Action potentials were monitored on a digital memory oscilloscope (Gould Instrument Systems Inc., USA) and digitized by a device of automatic acquisition and processing of the action potential (DATAPAC, Biologic, France). The following action potential parameters were automatically recorded and measured: resting membrane potential, maximal upstroke velocity of action potential, action potential amplitude, and action potential duration measured respectively at 50% and 90% of full repolarization.#

During a 120 min equilibration period, the two compartments of the tissue bath were superfused with normal Tyrode's solution and the right ventricular muscle was stimulated at a frequency of 1 Hz at 36.5±0.5 °C. Thereafter, one chamber was superfused for 30 min with a modified Tyrode's solution (ischemic-like period with [K⁺]o=4, 8 or 12 mmol/l, respectively K4, K8 and K12 groups) and then returned for 30 min to normal Tyrode's solution superfusion (reperfusion period), while the second compartment remained in normoxic conditions and superfused with normal Tyrode's solution all along the experiment (lasting 60 min). Then either dofetilide 5 or 10 nmol/l was superfused in the two zones. Dofetilide was superfused at the beginning of simulated-ischemia and all along during ischemic-like and reperfusion phases. Thus, the electrophysiological effects of dofetilide were investigated, by means of intracellular glass microelectrodes on: (i) action potential parameters simultaneously recorded in normal and altered zones; (ii) the incidence of arrhythmias during simulated-ischemia and reperfusion. A total of 60 preparations were used, all with their respective control period, distributed into K4, K8 and K12 groups at both 5 and 10 nmol/l dofetilide.#

The following types of arrhythmias were recorded: (1) myocardial conduction blocks, (2) triggered repetitive responses induced by a single extrastimulus, (3) spontaneous repetitive responses such as extrasystoles (1 to 3 spontaneous action potentials), salvos (4 to 9 spontaneous action potentials) and sustained arrhythmias (>10 spontaneous action potentials), as previously described (Rouet et al. 2000; Ducroq et al. 2005).#

Data were expressed as mean±S.E.M. and percentages of variations with respect to initial values measured before initiation of the ischemic-like phase. Preparations were discarded if, before the onset of simulated-ischemia, the following characteristics were not at least obtained: resting membrane potential −80 mV, maximal upstroke velocity 150 V/s and action potential durations at 90% of repolarization 100 ms.#

In the two compartments, each cell served as its own control and significance of differences in absolute values was determined using analysis of variance (ANOVA) for repeated measures followed by Dunnet's test as compared to initial values. Significance of differences between groups was determined using 2-factor ANOVA. The Fisher's exact test was used for comparison of nonparametric categorical data. Differences were considered significant when P<0.05. The action potential parameters and the ratio of experiments exhibiting arrhythmias were analyzed from 20 experiments in each group (2 concentrations of dofetilide in each of 3 [K⁺]o groups). The difference between normal and altered zones of action potential duration was analyzed in the following way: the values of action potential durations at 90% of repolarization were firstly normalized in each experiment to exclude the influence of individual differences at the beginning of each experiment. In this respect, action potential durations at 90% of repolarization in each compartment at 0, 10, 20 and 30 min of ischemia-like conditions was expressed as a percentage of the value observed before initiation of ischemia-like conditions. Then, the resulting action potential durations at 90% of repolarization in altered zone was subtracted from the corresponding data in normal zone.#

Table 1 shows that simulated-ischemia induced a significant membrane depolarization and changes in all parameters measured in all groups. Under control conditions, there were significant decreases of resting membrane potential, by −7±3% (P<0.01), −20±3% (P<0.01) and −25±3% (P<0.01), after 30 min of simulated-ischemia in presence of 4, 8 and 12 mmol/l [K⁺]o, respectively. In the same way, action potential amplitude was reduced by −8±2% (P<0.05), −21±3% (P<0.01) and −26±4% (P<0.01) and maximal upstroke velocity by −21±6% (P<0.05), −55±6% (P<0.01) and −40±14% (P<0.01). Moreover, action potential durations at 50% and 90% of repolarization were decreased by simulated-ischemia [action potential durations at 50% of repolarization: −37±8% (P<0.01), −55±6% (P<0.01) and −55±6% (P<0.01) and action potential durations at 90% of repolarization: −27±5% (P<0.01), −44±5% (P<0.01) and −52±5% (P<0.01) in presence of 4, 8 and 12 mmol/l [K⁺]o, respectively]. Finally, as illustrated in Figs. 1 and 2, the higher the [K⁺]o the more dramatic were these action potential alterations (P<0.001 between groups).#

A 60 min Tyrode's superfusion in normal zone did not modify any action potential parameters. Table 2 shows that, in normal zone, dofetilide was devoid of significant effects on resting membrane potential, action potential amplitude and maximal upstroke velocity. Moreover, dofetilide 5 nmol/l induced an increase of action potential durations at 50% of repolarization of +4±3% (NS), +7±2% (P<0.05), +14±5% (P<0.01) and of action potential durations at 90% of repolarization of +9±3% (NS), +7±2% (P<0.05), +14±3% (P<0.01) in K4, K8 and K12 groups, respectively after 60 min of superfusion. In the same way, dofetilide 10 nmol/l induced an increase of action potential durations at 50% of repolarization of +12±5% (P<0.01), +9±6% (P<0.01), +14±3% (P<0.01) and of action potential durations at 90% of repolarization of +16±6% (P<0.01), +11±3% (P<0.01), +19±5% (P<0.01) in K4, K8 and K12 groups, respectively.#

Results showed that whatever the [K⁺]o present during simulated-ischemia, alterations of action potential duration were not counterbalanced by the class III properties of dofetilide (5 and 10 nmol/l) as compared with control group (Table 1). Indeed in the K4 group, action potential durations at 90% of repolarization was decreased by −27±5% (P<0.01), −35±5% (P<0.01) and −37±3% (P<0.01) under control conditions and in presence of dofetilide 5 and 10 nmol/l, respectively. In the same way, action potential durations at 90% of repolarization was decreased by −44±5% (P<0.01), −47±5% (P<0.01) and −39±6% (P<0.01) in the K8 groups and by −52±5% (P<0.01), −63±4% (P<0.01) and −55±3% (P<0.01) in the K12 groups.#

As summarized in Tables 1 and 2, reperfusion phase in controls allowed a return of the action potential parameters recorded in the altered zone to the values observed in the normal zone in the K4, K8 and K12 groups. Moreover, while a class III effect of dofetilide was not observed during simulated-ischemia, prolongation of action potential duration was observed during reperfusion in altered zone in presence of dofetilide 5 and 10 nmol/l, particularly in the K12 group [where action potential durations at 90% of repolarization was increased by +15±6% (P<0.05) and +16±5% (P<0.01) in presence of 5 and 10 nmol/l dofetilide, respectively].#

Fig. 3 (comparison between control lines in panels A, B, C) shows that [K⁺]o increase during ischemia led to worsening of dispersion of action potential durations at 90% of repolarization between altered and normal zones (P<0.01 between groups). Dofetilide only slightly increased this difference as compared to control groups. In the K4 group (Fig. 3A) an increase in action potential durations at 90% of repolarization difference between altered and normal zones was promoted by dofetilide 10 nmol/l at 30 min ischemia as compared to control group (P<0.05). In the same way, action potential durations at 90% of repolarization difference between altered and normal zones in K12 group (Fig. 3C) was significantly increased by dofetilide 10 nmol/l at 20 min ischemia as compared to control group (P<0.05).#

The different forms of ventricular arrhythmias observed during both simulated-ischemia and reperfusion are illustrated in Fig. 4 and reported in Tables 3 and 4. In K4 and K12 groups, a high incidence of electrical disturbances was observed during both simulated-ischemia and reperfusion. In contrast, the incidence of sustained activities and salvos was reduced during ischemia in K8 group as compared with both K4 and K12 groups. This beneficial effect of moderately increased [K⁺]o on salvos was maintained during the following reperfusion phase.#

In K4 group, dofetilide 10 nmol/l reduced the occurrence of sustained activities as compared to control group (P<0.05). It also seemed to exhibit antiarrhythmic properties on the global incidence of spontaneous repetitive responses (8/10 experiments showed arrhythmias in control group versus 4/10 in presence of dofetilide 10 nmol/l) although this did not reach statistical significance. Neither proarrhythmic nor antiarrhythmic effects of dofetilide (5 and 10 nmol/l) were observed in K8 and K12 groups during simulated-ischemia.#

During the reperfusion phase, in the K4 group, dofetilide 10 nmol/l significantly reduced the incidence of sustained activities (P<0.05) and abolished the occurrence of salvos (P<0.01). Dofetilide was devoid of effects on arrhythmias during reperfusion in K8 group. In K12 group, dofetilide (5 and 10 nmol/l) reduced the incidence of sustained activities (P<0.05).#

Ischemia induced a membrane depolarization concomitant with a decrease in APA from the lowest [K⁺]o. This effect was worsened by higher [K⁺]o thus suggesting that extracellular potassium during simulated-ischemia is not entirely responsible for cell membrane depolarization. High [K⁺]o increases gK1 and this might be counteract by other factors in ischemia. Moreover, a decrease in action potential duration occurred during simulated-ischemia. In the same way, these modifications were increased by high [K⁺]o. Activation of KATP channels plays an important role in action potential duration decrease during ischemia (Hamada et al., 1998). On the other hand, it was shown that action potential duration decrease during high [K⁺]o was greater than during ischemia, probably because of IK1 blockade by other ischemic components i.e. acidosis and lysophosphatidylcholine accumulation (Watanabe et al., 1997). In our study, action potential duration shortening should not be solely attributed to IKATP activation as it was accentuated by high [K⁺]o. IKr, which is enhanced by high [K⁺]o (Yang and Roden, 1996), may also be implicated in this phenomenon. In contrast, although it was enhanced by high [K⁺]o, IK1 may not be implicated in this phenomenon because of the IK1 inhibition by low intracellular ATP (Shieh et al., 1996).#

Our results demonstrate that high [K⁺]o made the ischemic-like induced arrhythmias worse. Both normal and higher [K⁺]o favored the initiation of severe arrhythmias whereas intermediate [K⁺]o limited their occurrence. This suggests that arrhythmias in K4 and K12 groups had a different pathophysiologic mechanism.#

Both re-entrant and non re-entrant mechanisms have been proposed for the initiation of ischemia-induced arrhythmias in vivo (Caralis and Perez-Stable, 1986). The development of re-entry requires the combination of several factors: an area of unidirectional conduction block associated with conduction velocity slowing and local differences in refractoriness (Caralis and Perez-Stable, 1986). In accordance with these conditions, re-entrant pathway around the border zone between altered and normal zones seems to be the most likely mechanism for arrhythmias in K12 group. However, one needs to recognize the limitations, especially geometry and extension, of tissues investigated in vitro (Rouet et al. 2000; Ducroq et al. 2005), inasmuch as re-entry is concerned. With these limitations in mind, our results showed that [K⁺]o increase enhanced action potential duration difference between normal and altered zones (Fig. 3). Moreover, it has also been proved that high [K⁺]o associated with hypoxia and acidosis led to conduction velocity slow-down (Veenstra et al., 1987).#

In contrast, arrhythmias observed in K4 group may well originate from abnormal automaticity linked with delayed afterdepolarizations. Indeed, several previous findings support this hypothesis: the participation of delayed afterdepolarizations in ischemia-induced arrhythmias is well-established and high [K⁺]o abolishes triggered activities and abnormal automaticity (Katzung and Morgenstern, 1977). Moreover, it has been suggested that delayed afterdepolarizations are more likely to play a role in arrhythmia generation when tissues are moderately rather than severely ischemic (Coetzee et al., 1987).#

The low incidence of ventricular arrhythmias observed in K8 group fits with these hypothesis: the increase in [K⁺]o may inhibit the onset of delayed afterdepolarizations but it might be insufficient to allow the onset of re-entry.#

It is however important to indicate that in the model used here and elsewhere (Rouet et al. 2000; Ducroq et al. 2005), we could not record delayed afterdepolarizations. This may be related to the scarce presence of Purkinje fibers in right ventricular strips dissected from the free wall of the guinea-pig heart.#

It has been demonstrated that reperfusion-induced arrhythmias resulted from both re-entrant and non-reentrant mechanisms (Pogwizd and Corr, 1987). The heterogeneous return of electrical activity contributes to the setting of multiple promoting sites for re-entrant arrhythmias (Corr and Witkowski, 1984). Moreover, during the early reperfusion phase, calcium overload (Coetzee et al., 1987), accumulation of cyclic AMP during previous ischemia (Yoshida et al., 2000) and enhanced sympathetic activity in vivo (Curtis and Hearse, 1989), may all lead to the development of delayed afterdepolarizations.#

Curtis and Hearse (1989) have demonstrated that interactions between the non-ischemic and the reperfused tissue can be ruled out in arrhythmogenesis and that arrhythmias may arise from within the reperfused tissue itself. Thus, in the present work, re-entrant circuits around the border zone are not likely to play a part in reperfusion-induced arrhythmias. As seen during ischemia, a low arrhythmias incidence was observed in K8 group as compared to both K4 and K12 groups. This may be explained by: (i) arrhythmogenesis during simulated-ischemia may have consequences on the incidence of arrhythmias during reperfusion or (ii) similar to ischemia-induced arrhythmias, different mechanisms may be involved in reperfusion-induced arrhythmias in K4 and K12 groups.#

As previously described (Rouet et al., 2000), dofetilide exhibited pure class III drug effects in normal zone but not in altered zone during ischemia probably because (i) action potential lengthening effect of dofetilide was overwhelmed by the activation of KATP channels (Rouet et al., 2000), (ii) IKr blockade by dofetilide was impaired by the increased [K⁺]o (Yang and Roden, 1996). Indeed, it was shown that [K⁺]o increase from 4 to 8 mmol/l led to a significant decrease in the efficacy of dofetilide to block IKr: dofetilide concentrations required to 50% IKr inhibition were 11.2±1.9 nmol/l and 79±32 nmol/l in presence of 4 and 8 mmol/l [K⁺]o, respectively (Yang and Roden, 1996).#

Probably because of individual differences, the slight preventing effect of dofetilide 5 nmol/l on early ischemic action potential duration at 90% of repolarization difference between normal and altered zones observed in the present work (Fig. 3C) was not as marked as this was previously observed (Rouet et al., 2000).#

The present results also demonstrate that dofetilide was antiarrhythmic during both simulated-ischemia (K4 group) and reperfusion (K4 and K12 groups). On the one hand, the antiarrhythmic effects of dofetilide during simulated-ischemia could not be explained by its class III effects since dofetilide did not prolong action potential duration in normal zone in the first 30 min of experiments. Moreover, the antiarrhythmic effects observed in presence of dofetilide during reperfusion may be related to action potential duration lengthening in normal and altered zones. By this way, dofetilide may interrupt re-entrant pathways and propagation of abnormal automatism. Finally, we point out that in K8 group, where the incidence of arrhythmias was low in the control period, the effects of dofetilide were neutral.#

Although our study was performed in an adequate model to investigate the effects of specific IKr blockers, such as E-4031, considering IKr density and kinetics and prevention of reverse frequency dependence (Lu et al., 2001), further investigations are needed not only with dofetilide but also with other IKr and/or IKs blockers, to see if the guinea-pig right ventricular myocardium is entirely relevant to approach human-targeted arrhythmogenesis and its possible modification. On the other hand, we have recently observed in the same model (Ducroq et al., 2005) that dofetilide 10 nmol/l plus HMR 1556 1 nmol/l used to mimic IKr and IKs blocking properties of azimilide enabled a less severe shortening of action potential duration at 90% of repolarization at 30 min of ischemic-like conditions (−43±9%), as compared with azimilide 0.5 μmol/l (−64±5%, P<0.05) but similar to what seen with azimilide 0.1 μmol/l (−53±5%) and controls (−52±6%). During reperfusion, 2/9 (22%) preparations had sustained activities, which was less than what observed in controls (5/10, 50%) and with azimilide 0.5 μmol/l (0/10, 0%), although not statistically different (respectively, P=0.35 and P=0.21). Lack versus homogenous class III effects of azimilide in respectively simulated-ischemia and reperfusion may thus explain its different efficacy on arrhythmias in guinea-pig ventricle, pointing to this tissue-type appropriateness to investigate drugs which modulate both components of the delayed outward rectifier current, although prevention of reperfusion arrhythmias calls for other than just IKr and IKs blocking properties of azimilide (Ducroq et al., 2005).#

It is important to interpret the results of the study, considering that this in vitro model does not reflect exactly the electrophysiological changes occurring during ischemia and reperfusion in the human heart since: (i) the composition of the extracellular bath is continuously controlled by a constant perfusion of normal or modified Tyrode's solution whereas, in the early stage of myocardial infarction in humans, the ionic alterations may differ among patients; (ii) other components of the extracellular environment such as lysophosphatidylcholine and arachidonic acid, which were not taken into account in the present work, may accumulate during ischemia; (iii) sympathetic and parasympathetic drives are absent; (iv) action potentials were recorded in only one cell in each zone even when electrical disturbances such as conduction blocks may occur in other myocardial areas than those including the recorded cells; (v) the questions arise as to whether re-entry may occur in a strip of ventricular muscle with a small mass and short distances and only on narrow sharp boundary between normal and altered zones, on how would the impulse re-enter in those conditions, where would the area of unidirectional block be, how much the conduction velocity should decrease and how about the local differences in refractoriness.#

Collectively however our results suggest that dofetilide should be beneficial in the context of ischemia-induced arrhythmias. This has a potential extension to patients with chronic ischemic disease and/or myocardial infarction although attention needs to be paid, common with other class III agents, to pre-existing QT prolongation.#