The isolated rat heart with regional ischemia caused by coronary artery occlusion allows the generation and examination of ischemia-induced and reperfusion-induced arrhythmias, eg, ventricular fibrillation. 1,2 This experimental preparation has been used extensively for the study of the pathophysiology of myocardial ischemia and reperfusion and for the pharmacologic assessment of the antiarrhythmic activity of newly developed drugs, and many of its advantages and limitations have been characterized. 3 The present study addresses one of its possible limitations.
The potassium channel that conducts the rapid component of the delayed rectifier potassium current (IKr) is expressed in the rat heart. 4 However, in rat ventricular myocytes, IKr is functionally inactive and does not play a role in the repolarization process. 5,6 This contrasts with the human ventricular myocyte, in which IKr is one of the main repolarizing currents. 7 In the rat ventricle, functionally the most important repolarizing current is the transient outward K+ current (Ito), 8 with the inward rectifier K+ current (IK1) contributing to the terminal phase of repolarization. 9 The repolarization process of the rat myocyte and the shape of the corresponding QT interval in the electrocardiogram (ECG) therefore differ markedly in rats versus humans. 10 Nevertheless, drugs that block Ito11,12 or IK19,13 consistently widen QT in the rat heart, as they do in other species. 9 Thus, despite anomalies in repolarization, it has been argued that QT widening in rat hearts, as in man, results from repolarization delay. 9
In isolated ventricular myocytes, simulated ischemia consistently shortens cardiac action potential duration. 14 Thus, one of the more disconcerting aspects of the Langendorff rat heart preparation is the tendency for ischemia to transiently widen QT interval. This observation is consistent and had been reported by all those who have examined it. 6,15,16 Interestingly, this effect is not restricted to rats, and appears to be common to all regionally ischemic isolated Langendorff-perfused hearts in which data are available, ie, rabbit and primate. 9 It also occurs in blood-perfused and Krebs solution–perfused rat hearts. 17 These data are all consistent with the suggestion that conduction delay, such as in response to Na+ channel blockade, may give rise to QT widening in the Langendorff preparation 18,19 and account for ischemia-induced QT widening. If so, the model would be compromised as a bioassay for detecting drug-induced repolarization delay.
Despite these data, we contend that ischemia-induced and drug-induced QT widening result from repolarization delay rather than slowed conduction. There are two reasons for this contention. First, the temporal pattern of ischemia-induced QT widening in the Langendorff preparation (rat included) is not progressive and monomorphic as would be expected if it resulted from slowed conduction. Instead, it is complex, with a peak in QT width after approximately 10 minutes of ischemia, followed by a shortening, then a return to baseline by approximately 30 minutes. 20 Second, elevation of extracellular K+, which causes partial depolarization and inactivation of Na+ channels, 14 and hence slowing of conduction, 21 actually shortens QT interval in isolated hearts 9 rather than widening it. Finally, ischemia- and drug-induced QT widening in isolated rat hearts is matched by an equivalent widening of local monophasic action potential duration, 22 which is not explicable on the basis of conduction delay.
Thus, we have tested the hypothesis that QT widening in the rat Langendorff preparation represents repolarization delay and not conduction slowing. To achieve this, we examined the effects of three Na+ channel–blocking drugs—quinidine, lidocaine and flecainide—on QT before and during regional ischemia in isolated rat hearts perfused with Krebs solution containing 3 m M or 5 m M K+. Elevation of the extracellular K+ concentration by causing partial depolarization of the sarcolemma potentiates the conduction slowing effects of class I antiarrhythmic drugs, 23–25 so, if Na+ channel blockade prolongs QT interval in this preparation, the QT-widening effect of the three drugs would be expected to be augmented by elevation of the K+ content of the perfusion solution. The data were derived from a larger experiment on the antiarrhythmic effects of class I agents. 26
Animals and General Experimental Methods
The animal-handling protocol was in accordance with the Guidance of the Operation on the Animals (Scientific Procedures) Act 1986, London, UK. Male Wistar rats (180–250 g; Bantin and Kingman, UK) were anesthetized with pentobarbitone (60 mg/kg intraperitoneally) mixed with 250 IU sodium heparin to prevent blood clot formation in the coronary vasculature. Hearts were excised and placed into ice-cold solution containing (in m M) NaCl, 118.5; NaHCO3, 25.0; MgSO4, 1.2; NaH2PO4, 1.2; CaCl2, 1.4; KCl, 3 (or 5 where indicated), and glucose 11.1; then perfused according to Langendorff, with solution delivered at 37°C and pH 7.4. All solutions were filtered (5-μm pore size) before use. Perfusion pressure was maintained constant at 70 mm Hg. A unipolar ECG was recorded by implanting one stainless-steel wire electrode into the center of the region to become ischemic, with a second connected to the aorta. A traction-type coronary occluder consisting of a silk suture (Mersilk 4–0) threaded through a polythene guide was used for coronary occlusion. The suture was positioned loosely around the left main coronary artery beneath the left atrial appendage. Regional ischemia was induced by tightening the occluder.
Four sets of experiments were performed and each consisted of four groups (quinidine, lidocaine, flecainide, and vehicle control). The basis for choice of drug concentrations is explained later. In the first set of experiments, hearts were perfused with the lower concentrations of the drugs, ie, quinidine 0.79 μM, lidocaine 3.88 μM, flecainide 0.74 μM, or vehicle (time-matched control). In the second set of experiments, hearts were perfused with the higher concentrations of the drugs, ie, quinidine 7.90 μM, lidocaine 12.93 μM, flecainide 1.48 μM, or vehicle (time-matched control). In these first two sets of experiments, all perfusion solutions were prepared by using modified Krebs solution containing 3 m M K+. In the third and fourth sets of experiments, again, the same lower and higher drug concentrations were used, together with vehicle controls, but the Krebs solution was modified to contain 5 m M K+. Each group consisted of 6 hearts; thus, a total of 96 hearts were used.
In all four sets of experiments, hearts were perfused for an initial 5 minutes with control solution, then solution was switched in a blinded fashion to one of the four test solutions: control (vehicle), quinidine, lidocaine, or flecainide. The choice of solution was made by reference to a randomization table. After a further 5 minutes of perfusion, the left main coronary artery was occluded for 30 minutes. Randomization was achieved by coding each group with a letter of meaning that was unknown to the operator. Blinded analysis was achieved by using stock solutions prepared by a second operator who did not participate in heart perfusion or data analysis.
Individual measures of coronary flow and ECG variables were taken 1 minute before and 1 minute after the introduction of drug perfusion or vehicle, 1 minute before and each minute after coronary occlusion for 5 minutes, then every 5 minutes thereafter for 25 minutes.
The choice of drug concentrations was based on the following. In clinical studies aimed at evaluating drug effects on ventricular arrhythmias, peak blood concentrations (unbound fraction plus plasma/protein bound fraction) have been determined following typical drug dosage. These concentrations have been reported to be 7.9 μM for quinidine, 27 12.93 μM for lidocaine, 28 and 1.48 μM for flecainide. 29 These concentrations are within the so-called human therapeutic range for these agents. 30 All three drugs are bound to plasma proteins: quinidine 90%, lidocaine 70%, and flecainide 50%. 31 Thus, the plasma/protein unbound concentrations associated with these clinically relevant dosages are approximately 0.79 μM, 3.88 μM, and 0.74 μM for quinidine, lidocaine, and flecainide, respectively. Therefore, we chose to study the mean peak unbound and mean peak total plasma concentrations (lower and higher concentrations, respectively). It was found recently that only the mean peak total plasma concentrations of these drugs, which are inappropriately high concentrations, prevented ischemia-induced phase 1 ventricular fibrillation in Langendorff-perfused rat hearts. 26
Measurement of Involved Zone Size and Coronary Flow
At the end of the experiment, the size of the involved zone (the region subjected to ischemia) was quantified using the disulphine blue dye exclusion method 1 and expressed as percent total ventricular weight. Coronary flow was measured by timed collection of coronary effluent. Values of coronary flow in the uninvolved tissue were calculated from the total coronary flow and the weights of the involved and uninvolved zones as described previously. 2
Any heart with a sinus rate less than 250 beats per minute or a coronary flow more than 18 mL/min/g or less than 7.5 mL/min/g 6 minutes before the onset of ischemia (before the start of perfusion with drug or vehicle) or an involved zone of less than 30% or more than 50% of total ventricular weight was excluded. All excluded hearts were replaced to maintain equal group sizes.
The ECG was recorded using a MacLab system. From the ECG, the PR interval, RR interval, and QT interval (measured at the point of 90% repolarization with on-screen cursors) were obtained. Measurement of all variables was performed in a blinded manner. In the first instance, no attempt was made to correct QT for heart rate because previous studies have shown that QT interval is not rate-dependent in the rat heart. 9
Drugs and Materials
Drug stocks were prepared fresh each week and perfusion solutions were prepared fresh each day. Vehicle stock was an 8-mL solution containing 6 mL ethanol and 2 mL water. The control solution contained 0.24 mL of vehicle stock in 1 L modified Krebs solution (ie, 0.018% ethanol). The drug solutions were prepared from 8 mL of a drug stock. Drug stocks for low concentrations consisted of 10 mg quinidine, 35 mg lidocaine, or 11.7 mg flecainide dissolved in 6 mL ethanol plus 2 mL water, whereas drug stocks for high concentrations consisted of 100 mg quinidine, 116.7 mg lidocaine, 23.4 mg flecainide dissolved in 6 mL ethanol plus 2 mL water, such that 0.24 mL of these stocks dissolved in 1 L of modified Krebs solution yielded the drug solutions (quinidine 0.79 μM or 7.90 μM, lidocaine 3.88 μM or 12.93 μM, and flecainide μM 0.74 or 1.48 μM). Thus, all solutions contained the same amount of ethanol (0.018%).
All salts were reagent-grade chemicals from Sigma Chemical. Water for preparing perfusion solution was supplied using a reverse-osmosis system (Milli-RO 10 and Milli-Q 50; Millipore) and had a specific resistivity of more than 18 mOhm.
Gaussian distributed variables were expressed as means ± SEM. When the effects of K+ concentrations on the baseline parameters were analyzed, continuous data from two independent samples were compared with independent-samples t tests. When the effects of K+ concentrations on the QT90 intervals and the heart rate values of drug-free control groups were analyzed, data from independent samples were subjected to two-way, repeated-measurements analysis of variance with one “between-subject factor” (ie, K+, 3 m M vs. 5 m M) and with one “within-subject factor” (ie, time). If a significant interaction was found, pairwise comparisons were performed based on estimated marginal means. When drug effects were analyzed, continuous data from independent samples were subjected to three-way, repeated-measures analysis of variance with two between-subject factors (ie, K+, 3 m M vs. 5 m M; and treatment, control vs. quinidine, lidocaine, or flecainide) and with one within-subject factor (ie, time). If a significant interaction was found, pairwise comparisons were performed based on estimated marginal means. A P value < 0.05 was taken as indicative of a statistically significant difference between values.
Effects of Altering the K+ Concentration of the Perfusion Solution on Baseline Variables
At baseline, hearts perfused with Krebs solution containing 5 m M K+ (pooled groups, n = 48) had significantly higher heart rates, shorter QT90 intervals, and greater coronary flow compared with those perfused with 3 m M K+ (Table 1). The K+ content of the perfusion solution did not affect the duration of PR intervals.
To find out whether the longer QT90 interval duration resulted from the lower heart rate in hearts perfused with 3 m M K+, the relationship between baseline heart rate and QT90 interval was examined by regression analysis. The accumulated data from experiments performed previously by our group (unpublished data) using 308 isolated rat hearts perfused with 3 m M K+ (Fig. 1A) and the data from the hearts in the present study (Fig. 1B, C) showed that there is only a weak correlation between the baseline heart rate and QT90 intervals in isolated rat hearts perfused with either 3 m M or 5 m M K+. Although the correlation was statistically significant, the weak correlation coefficients and shallow regression slopes suggest that there is no biologically significant relationship between these variables, and hence no requirement for correcting QT for heart rate in isolated rat hearts, as suggested previously. 9 This means the effects of K+ on QT and heart rate occur in parallel rather than in series (ie, independently) in the rat Langendorff preparation.
QT90 Intervals and Heart Rate in Controls
During ischemia, control QT90 intervals remained significantly shorter while the heart rates remained significantly higher in hearts perfused with 5 m M versus 3 m M K+ (Fig. 2). The QT90 intervals increased markedly during the first 3 minutes of ischemia (P < 0.05 vs. baseline in both groups;Fig. 2A). This was followed by a gradual shortening of QT90 until it returned to baseline level (P > 0.05 vs. baseline in both groups) at minute 20 of ischemia. Heart rate also changed in parallel in the two groups, decreasing during ischemia (P < 0.05 vs. baseline in both groups;Fig. 2B). Experience has shown that this is in fact a model-dependent artifact that is unrelated to ischemia (it began before the onset of ischemia) and is complete 15 minutes after the start of perfusion. 32 However, although the QT90 intervals and heart rates both changed in parallel between the groups of control hearts perfused with 3 m M and 5 m M K+, importantly, the changes in QT90 did not parallel the changes in heart rate, confirming that QT90 intervals are not determined by, and do not depend on, the heart rate in Langendorff-perfused rat hearts (Fig. 2).
Effect of Na+ Channel Blockers on QT90 Interval
Of the lower concentrations of the three drugs, only quinidine widened the QT90 interval, and this occurred only at the end of ischemia (Figs. 3A–D). The higher concentrations of quinidine and flecainide widened the QT90 interval, but this effect was significantly greater in hearts perfused with 3 m M versus 5 m M K+ (Figs. 3E–H), ie, K+ surmounted the QT90 interval–widening effect. Lidocaine was without effect at either K+ concentration.
During early ischemia (3–5 minutes), the higher concentration of quinidine added extra QT90 interval prolongation to that seen when the perfusion solution contained 3 m M K+, whereas lidocaine and flecainide were without effect (Figs. 3C, G). Only in the groups perfused with quinidine was the QT90 interval–widening effect sustained during ischemia. In all other groups, any QT90 prolongation seen during early ischemia was followed by a gradual shortening (Figs. 3D, H).
Effect of Na+ Channel Blockers on Heart Rate
In all drug groups, heart rate decreased significantly during the interval between the start of perfusion and the initiation of ischemia (Table 2) to a similar extent to that seen in the merged control groups (Figure 2B). The drug- and time-matched heart rates were significantly higher in hearts perfused with 5 m M versus 3 m M K+ (Table 2). None of the drugs at their lower concentrations affected heart rate before or during ischemia, regardless of the K+ content of the Krebs solution. In hearts perfused with Krebs solution containing either 3 m M or 5 m M K+, the higher concentration of quinidine reduced heart rate compared with controls, but only at the end of ischemia (at 25 minutes), whereas lidocaine and flecainide did not influence heart rate, even at their higher concentrations, at any time during the experiments regardless of the K+ content of the Krebs solution (Table 2).
Coronary Flow and Mean Involved Zone Size
None of the drugs affected coronary flow before or during ischemia and none affected mean involved zone size (data not shown).
Ischemia caused a transient widening of QT interval followed by a shortening during the following 30 minutes. Class I antiarrhythmic drug–induced QT widening was most pronounced with quinidine, whereas lidocaine was without effect. Baseline QT intervals correlated inversely with the K+ content of the Krebs solution, as it does with serum K+ in rats in vivo. 33–35 Elevation of Krebs solution K+ content shortened the QT interval. Moreover, it also reduced or surmounted any class I drug–induced QT widening. Elevated extracellular K+ concentration causes partial depolarization, which inactivates Na+ channels and consequently slows impulse conduction. 14 These findings reveal that it is not tenable that Na+ channel blockade and conduction slowing play any role in ischemia-induced or drug-induced QT widening in the isolated Langendorff-perfused rat heart, contrary to previous suggestions. 18,19 The following elaborates on this and identifies possible reasons to question it.
K+, K+ Channels, and QT Interval in Rat Hearts
First, it is important to emphasize that it is well-established that drugs that selectively block K+ channels are known to widen QT interval in isolated rat hearts, and that this effect is attributable to K+ channel blockade. 6,9,13,22,36 Moreover, elevation of extracellular K+ can surmount drug-induced QT widening in a manner explicable by an interaction at the level of K+ channels.
In two previous studies, switching perfusion from 8 m M to 3 m M K+ did not widen QT interval in nonischemic isolated perfused rat hearts. 9,36 However, in these particular studies, the QT intervals were measured at 100% repolarization. As a result of the extreme difficulty in identifying the precise time of restoration of isoelectric voltage in the rat ECG, determination of QT at 100% repolarization is a hazardous practice that is no longer followed. 37 Nevertheless, in previous studies, when QT interval had been widened by perfusion with K+ channel–blocking drugs, switching perfusion from 3 m M to 8 m M K+ was able to shorten QT interval at 100% repolarization. 9,36 In the present study, findings were similar for the effects of quinidine and flecainide on QT90.
Elevation of extracellular K+ increases the conductance of several K+ channels, which can be expected to evoke shortening of the duration of the action potential and QT interval. 14 Conversely, hypokalemia inhibits IK1 in rat ventricular myocytes, which can be expected to result in QT prolongation. 38 Likewise, pharmacologic blockade of IK1 with tacrine and RP58866 widens QT interval in isolated rat hearts. 9,13 Therefore, reduced IK1 conductance might play a role in producing QT interval widening in rat hearts perfused with the lower K+ concentration. In summary, the relationship between QT and K+ channels in rat hearts is well established.
K+, Heart Rate, and QT Interval in Rat Hearts
Heart rate is perceived to be a confounding factor in the interpretation of effects of ions and drugs on repolarization. 14 It is important to consider whether this plays any part in diminishing the strength of our conclusions.
In the present study, the K+ content of the perfusion solution correlated positively with heart rate. This mild tachycardic effect of K+ has been reported previously 2 and is probably attributable to an enhanced conductance of the If Na+ current in sinoatrial cells. 39 The QT interval–widening effect of reducing the K+ content of the perfusion solution did not result from the coincident decrease in heart rate, as only a weak correlation was found between the baseline heart rate and QT values, and none at all was found when analysis was restricted to hearts perfused with a single Krebs solution K+ content. Reducing heart rate from 400 to 200 beats per minute can be expected to widen QT interval by only approximately 15 milliseconds in the isolated Langendorff-perfused rat heart (Fig. 1).
Thus, K+ affects heart rate and QT interval independently in the rat Langendorff preparation. This means that the QT interval–widening effect of quinidine and flecainide was the result of an effect of the drugs on the myocardium and not secondary to any reduction in heart rate. However, it is likely that heart rate did modulate the QT interval–widening effects of these drugs. The potential relevance of this is discussed later.
K+, Coronary Flow, and QT Interval in Rat Hearts
Other factors may also confound interpretation. For example, under appropriate circumstances, myocardial stretch can affect repolarization. 14 Elevation of the K+ content of the perfusion solution was associated with an increase in coronary flow. This weak vasodilatory effect of K+ in isolated rat hearts has been reported previously, 2,40 but the mechanism responsible is not established. Drugs that increase coronary flow do not necessarily lengthen QT (eg, verapamil and mibefradil 37); therefore, it is unlikely that an increase in coronary flow served as a reason for the widened QT interval. Changes in shear stress leading to stretch-induced QT changes can also be ruled out, as shear stress is constant in the constant pressure Langendorff model even if coronary flow is altered by vasoactive drugs. 41
Ischemia and QT Interval in Rat Hearts
Interpretation of the meaning of drug-induced QT interval widening in the rat Langendorff preparation is made difficult by the observation of ischemia-induced transient QT interval widening. In most experimental models of ischemia, adenosine triphosphate (ATP)–dependent K+ channels (KATP) open when intracellular ATP levels decrease. This results in action potential duration (and QT interval) shortening. 14 However, in the present study, QT intervals increased markedly in all groups during the first 3–5 minutes of ischemia. This was followed by gradual returning of QT interval to the baseline level by the 20th minute of ischemia in all groups except the ones perfused with quinidine. This curious observation is not unique to our laboratory: marked prolongation of QT interval and action potential duration have been observed by others during the first 3–5 minutes of regional 6,16,42 and global 15 ischemia in isolated Langendorff-perfused rat hearts. Workman et al 16 suggested that KATP channels do not open in Langendorff-perfused rat hearts during regional ischemia, as the KATP channel blocker glibenclamide did not affect ischemia-induced prolongation of the action potential duration, whereas the KATP channel opener Ro-31-6930 prevented it.
Ischemia-induced QT widening is not a rat-dependent phenomenon; it has been observed in rabbit and primate Langendorff-perfused hearts. 9 However, global ischemia in isolated working rat hearts has been reported to shorten QT intervals, and this was prevented by the selective KATP channel blocker HMR 1098. 43 Therefore, it would appear that the mode of perfusion (Langendorff vs. working mode), rather than the species, determines the effect of ischemia on QT interval in isolated hearts. One possible explanation for this is that the unloaded nature of the Langendorff preparation predisposes the heart to QT widening during ischemia. There is a possible explanation for this because ischemia-induced KATP channel opening (which will shorten action potential duration and QT interval) may depend on myocardial stretch, which differs between loaded (working) and unloaded (Langendorff) perfused hearts. This is supported by reports that stretch enhances KATP channel opening 44 and that the shortening of the action potential duration upon stretch is more pronounced during acute ischemia. 45
Although a lack of stretch-induced and/or KATP channel opening-induced shortening of the action potential duration may explain the lack of QT shortening in the Langendorff preparation, it does not explain the widening of QT interval seen during the first 10 minutes of ischemia. One possible explanation is that ischemia itself blocks certain cardiac potassium currents. There is some evidence to support this because Henry et al 46 showed that simulated ischemia initially widened action potential duration in isolated rat ventricular myocytes as a result of reduced IK1 and Ito K+ currents. Whether this also explains QT interval widening in ischemic rabbit and primate hearts 9 remains to be determined.
In summary, the net balance of effects on the different cardiac K+ currents would appear to be sufficient to explain not only the classical effects of ischemia on QT in vivo, but also the range of (sometimes unexpected) effects observed in different heart perfusion preparations.
Effect of Class I Antiarrhythmic Drugs on QT Interval
It would be of great value to be able to establish that class I drug–induced QT interval widening is attributable to blockade of repolarizing currents rather than an uncertain combination of this action plus blockade of inward Na+ current. Lidocaine did not affect QT interval in the present experiments, whereas quinidine and flecainide at their higher concentrations widened the QT interval. The effect of the latter drugs was most pronounced when the perfusion solution contained the lower-K+ concentration selected for study. Furthermore, quinidine and flecainide widened QT interval independently of heart rate. Quinidine has been reported to lengthen QT intervals in isolated perfused rat hearts dose-dependently in several published studies. 47,48 Likewise, in rats in vivo, lidocaine has been shown to have no effect on QT whereas flecainide and quinidine caused QT widening. 49,50 These findings are in accordance with those of the present study. In contrast, Suzuki et al 18 reported that lidocaine widened QT intervals dose-dependently in another in vivo study in rats. However, in this study, lidocaine widened QT only in the precordial chest ECG leads at right anterior chest positions, whereas the drug did not prolong the action potential duration but decreased the maximum upstroke velocity in single rat myocardial cells in the same study. 18 Therefore, although it was suggested that lidocaine-induced QT interval prolongation was caused by a delay in local ventricular conduction time in the right ventricle, 18 it is difficult to justify this suggestion because it fails to explain why decreased left ventricular conduction delay did not result in QT prolongation in other leads. Importantly because lidocaine was without effect on the QT interval in the present experiments, and elevation of the K+ concentration of the perfusion solution failed to potentiate the effects of any of the representative Na+ channel blockers on the QT interval, we would argue that it is highly unlikely that Na+ channel block and conduction slowing play any role in drug-induced QT prolongation in the Langendorff-perfused rat heart.
It is useful to have a hypothetical mechanism for class I drug–induced QT widening ruled out. However, it would also be valuable to be able to identify the genuine mechanism. Unlike lidocaine, quinidine and flecainide block K+ channels (eg, Ito) in rat ventricular myocytes at similar concentrations to those that block Na+ channels. 11,38 K+ channel blockade results in QT prolongation, 51 so the simplest explanation for quinidine- and flecainide-induced QT prolongation in the present experiments is K+ channel blockade.
Lowering the K+ content of the perfusion solution potentiated the QT interval–widening effect of quinidine and flecainide. Likewise, the QT interval–widening effects of the K+ channel blockers tacrine and tedisamil were enhanced by lowering the K+ content of the perfusion solution in previous studies using the same preparation. 9,36 Although Hirota et al 34,38 reported that the K+ channel blocking and QT interval–widening effects of quinidine were not affected by the extracellular K+ concentration in rat ventricular myocytes and in rats in vivo, Yang and Roden 52 demonstrated that lowering the extracellular K+ concentration markedly reduced the quinidine concentration producing 50% inhibition of IKr in single myocytes. Despite the rat lacking functional IKr, 4–6 the present and previously published studies with Langendorff-perfused rat hearts 9,36 reveal that the K+ dependence of drug-induced QT interval lengthening resembles that seen in man and in other species that express IKr, such as guinea pig, rabbit, and dog. Thus, these data suggest that the modulating effect of extracellular K+ on drug-induced QT widening represents an interaction at the level of K+ channels, rather than Na+ channels.
Overall, these data strongly suggest that class I drug–induced QT interval widening is attributable solely to K+ channel blockade.
Correlation Between Heart Rate and QT with or Without Drugs
As noted earlier, heart rate has the ability to modulate the effects of drugs on QT interval, and this needs to be accounted for when evaluating the mechanism of QT interval widening. Although drug-free QT interval does not correlate with heart rate in isolated rat hearts, drug effects on QT may be either heart rate–dependent 13,53 or heart rate–independent 9 in this preparation. This can be explained by assuming that, in the absence of drugs, changes in heart rate affect the relative contribution of each outward current to the repolarization process without changing the total outward current, which remains constant over a range of heart rates. Thus, a drug that selectively inhibits a specific repolarizing current among the several that exist may exert its maximal effect at the heart rate at which the inhibited current is the most prominent repolarizing current, but the heart rate dependence may be altered or lost, depending on the rate dependence of the kinetics of the interaction between the drug and the channel. This implies that promiscuous molecular selectivity (ie, the wider variety of repolarizing currents a drug inhibits) will lessen the heart rate–dependence of a drug's QT interval–widening effects. The present results support this proposition, as quinidine and flecainide inhibit multiple K+ channels 11,38 and both drugs were found to widen QT interval over a wide range of heart rates.
Heart rate dependence of drug effects on QT could have been avoided by adding extra groups of hearts paced at one of several fixed heart rates. However, regardless of the outcome, doing this would not have materially affected the conclusions of the study. It is true that, when a drug widens QT interval and slows heart rate, the QT effect will be exaggerated if the drug has more pronounced K+ channel–blocking actions at a slower heart rate, as noted earlier. However, the extent to which the QT interval–widening effect of the drug is caused by bradycardia-induced rate-dependent K+ channel block as opposed to rate-independent K+ channel block is immaterial, as it does not address the key question of whether the drug-induced QT interval widening occurred because of block of K+ channels or because of the other main molecular effect of the drug block of Na+ channels.
In addition, the nature of the rate dependence of the K+ versus Na+ channel block of the drugs studied actually substantiates the primary conclusion we have made about the effects of the drugs on QT (effects being caused by K+ channel blockade). None of the drugs possess reverse rate-dependent Na+ channel–blocking actions. This means that we are certain that, when quinidine slowed heart rate, this did not exacerbate its ability to block Na+ channels. Therefore, rate-dependent effects of mixed Na+/K+ channel–blocking drugs on Na+ channels are inversely related to their rate-dependent effects on QT interval. This inverse relationship is most emphatically casual, not causal. Conversely, the rate-dependent effect of mixed Na+/K+ channel–blocking drugs on K+ channels is causal: rate acts as a multiplier of drug effects on the QT interval. Bearing this in mind, the fact that the greatest QT interval widening in the present study occurred with the most bradycardic drug, and the fact that that perfusion with elevated K+ solution (at a concentration insufficient to affect drug-free QT interval) reversed this, provides further evidence to support the conclusion that drug-induced QT interval widening is not caused by Na+ channel block.
The Langendorff preparation is unusual in that ischemia causes a transient widening of QT interval. Nevertheless, the preconception that the rat Langendorff preparation is compromised because it is susceptible to QT interval widening as a consequence of ischemia-induced (and class I drug–induced) Na+ channel blockade and conduction delay, is incorrect. Instead, it would appear that QT interval widening in this preparation is exclusively the result of K+ channel blockade and repolarization delay, as it would seem to be in other preparations, whether in vitro or in vivo. Moreover, the relative lack of influence of variations in heart rate, myocardial stretch, and coronary flow on QT interval in the rat Langendorff preparation means that the preparation is ideal for screening for drug effects on cardiac K+ channels (other than IKr).
The authors thank the Soros Foundation for covering the living costs of András Farkas during this study in London, UK. The authors also thank Dr. Krisztina Boda for her generous help in statistical analysis and Hugh Clements-Jewery for his helpful comments during the preparation of the manuscript.
1. Curtis MJ, Hearse DJ. Reperfusion-induced arrhythmias are critically dependent upon occluded zone size: relevance to the mechanism of arrhythmogenesis. J Mol Cell Cardiol. 1989; 21:625–637.
2. Curtis MJ, Hearse DJ. Ischemia-induced and reperfusion-induced arrhythmias differ in their sensitivity to potassium: implications for mechanisms of initiation and maintenance of ventricular fibrillation. J Mol Cell Cardiol. 1989; 21:21–40.
3. Curtis MJ. Characterisation, utilisation and clinical relevance of isolated perfused heart models of ischemia-induced ventricular fibrillation. Cardiovasc Res. 1998; 39:194–215.
4. Wymore RS, Gintant GA, Wymore RT, et al. Tissue and species distribution of mRNA for the IKr
channel, erg. Circ Res. 1997; 80:261–268.
5. Tande PM, Bjornstad H, Yang T, et al. Rate-dependent class III antiarrhythmic action, negative chronotropy, and positive inotropy of a novel Ik
blocking drug, UK-68,798: potent in guinea pig but no effect in rat myocardium. J Cardiovasc Pharmacol. 1990; 16:401–410.
6. Rees SA, Curtis MJ. Selective IK
blockade as an antiarrhythmic mechanism: effects of UK66,914 on ischemia and reperfusion arrhythmias in rat and rabbit hearts. Br J Pharmacol. 1993; 108:139–145.
7. Iost N, Virág L, Opincariu M, et al. Delayed rectifier potassium current in undiseased human ventricular myocytes. Cardiovasc Res. 1998; 40:508–515.
8. Varró A, Lathrop DA, Hester SB, et al. Ionic currents and action potentials in rabbit, rat, and guinea pig ventricular myocytes. Basic Res Cardiol. 1993; 88:93–102.
9. Rees SA, Curtis MJ. Specific IK1
blockade: a new antiarrhythmic mechanism? Effect of RP58866 on ventricular arrhythmias in rat, rabbit, and primate. Circulation. 1993; 87:1979–1989.
10. Gussak I, Chaitman BR, Kopecky SL, et al. Rapid ventricular repolarization in rodents: electrocardiographic manifestations, molecular mechanisms, and clinical insights. J Electrocardiol. 2000; 33:159–170.
11. Slawsky MT, Castle NA. K+
channel blocking actions of flecainide compared with those of propafenone and quinidine in adult rat ventricular myocytes. J Pharmacol Exp Ther. 1994; 269:66–74.
12. Clark RB, Sanchez-Chapula J, Salinas-Stefanon E, et al. Quinidine-induced open channel block of K+
current in rat ventricle. Br J Pharmacol. 1995; 115:335–343.
13. Rees SA, Curtis MJ. Tacrine inhibits ventricular fibrillation induced by ischemia and reperfusion and widens QT interval
in rat. Cardiovasc Res. 1993; 27:453–458.
14. Carmeliet E. Cardiac ionic currents and acute ischemia: from channels to arrhythmias. Physiol Rev. 1999; 79:917–1017.
15. Perchenet L, Kreher P. Mechanical and electrophysiological effects of preconditioning in isolated ischemic/reperfused rat hearts. J Cardiovasc Pharmacol. 1995; 26:831–840.
16. Workman AJ, MacKenzie I, Northover BJ. Do KATP
channels open as a prominent and early feature during ischemia in the Langendorff-perfused rat heart? Basic Res Cardiol. 2000; 95:250–260.
17. Clements-Jewery H, Hearse DJ, Curtis MJ. The isolated blood-perfused rat heart: an inappropriate model for the study of ischemia- and infarction-related ventricular fibrillation. Br J Pharmacol. 2002; 137:1089–1099.
18. Suzuki J, Tsubone H, Sugano S. Characteristics of electrocardiographic changes with some representative antiarrhythmic drugs in adult rats. J Vet Med Sci. 1991; 53:779–787.
19. Sun LS, Legato MJ, Rosen TS, et al. Sympathetic innervation modulates ventricular impulse propagation and repolarization in the immature rat heart. Cardiovasc Res. 1993; 27:459–463.
20. Ridley PD, Curtis MJ. Anion manipulation: a new antiarrhythmic approach. Action of substitution of chloride with nitrate on ischemia- and reperfusion-induced ventricular fibrillation and contractile function. Circ Res. 1992; 70:617–632.
21. Buchanan Jr, JW Saito T, Gettes LS. The effects of antiarrhythmic drugs, stimulation frequency, and potassium-induced resting membrane potential changes on conduction velocity and dV/dtmax
in guinea pig myocardium. Circ Res. 1985; 56:696–703.
22. Rees SA, Curtis MJ. Further investigations into the mechanism of antifibrillatory action of the specific IK1
blocker, RP58866, assessed using the rat dual coronary perfusion model. J Mol Cell Cardiol. 1995; 27:2595–2606.
23. Oshita S, Sada H, Kojima M, et al. Effects of tocainide and lidocaine on the transmembrane action potentials as related to external potassium and calcium concentrations in guinea-pig papillary muscles. Naunyn Schmiedebergs Arch Pharmacol. 1980; 314:67–82.
24. Matsuo S, Koh Y, Sato R, et al. Electrophysiological effects of flecainide on guinea pig ventricular muscle in high [K+]o, acidosis and hypoxia. Jpn Heart J. 1987; 28:539–554.
25. Nakaya H, Shishido Y, Oyama Y, et al. Voltage-dependent modification of Vmax
recovery from use-dependent block by pirmenol in guinea pig papillary muscles: comparison with other class I drugs. J Cardiovasc Pharmacol. 1992; 19:140–147.
26. Farkas A, Curtis MJ. Limited antifibrillatory effectiveness of clinically relevant concentrations of class I antiarrhythmics
in isolated perfused rat hearts. J Cardiovasc Pharmacol. 2002; 39:412–424.
27. Dinh HA, Murphy ML, Baker BJ, et al. Efficacy of propafenone compared with quinidine in chronic ventricular arrhythmias. Am J Cardiol. 1985; 55:1520–1524.
28. Lie KI, Wellens HJ, van Capelle FJ, et al. Lidocaine in the prevention of primary ventricular fibrillation. A double-blind, randomized study of 212 consecutive patients. N Engl J Med. 1974; 291:1324–1326.
29. Woosley RL, Siddoway LA, Duff HJ, et al. Flecainide dose-response relations in stable ventricular arrhythmias. Am J Cardiol. 1984; 53:59B–65B.
30. Singh BN, Opie LH, Marcus FI. Antiarrhythmic agents. In: Opie LH, ed. Drugs for the heart.
3rd ed. Philadelphia: WB Saunders, 1991:180–216.
31. Benet LZ, Oie S, Schwartz JB. Design and optimization of dosage regimens: pharmacokinetic data. In: Hardman JG, Limbird LE, Molinoff PB, et al, eds. Goodman & Gilman's The Pharmacologic Basis of Therapeutics,
9th ed. New York: McGraw-Hill, 1996:1707–92.
32. Avkiran M, Curtis MJ. Independent dual perfusion of left and right coronary arteries in isolated rat hearts. Am J Physiol. 1991; 261:H2082–H2090.
33. Akita M, Kuwahara M, Tsubone H, et al. ECG changes during furosemide-induced hypokalemia in the rat. J Electrocardiol. 1998; 31:45–49.
34. Hirota M, Ohtani H, Hanada E, et al. Effects of hypokalemia on arrhythmogenic risk of quinidine in rats. Life Sci. 1998; 62:2159–2169.
35. Yelamanchi VP, Molnar J, Ranade V, et al. Influence of electrolyte abnormalities on interlead variability of ventricular repolarization times in 12-lead electrocardiography. Am J Ther. 2001; 8:117–122.
36. Rees SA, Tsuchihashi K, Hearse DJ, et al. Combined administration of an IK(ATP)
activator and Ito
blocker increases coronary flow independently of effects on heart rate, QT interval
, and ischemia-induced ventricular fibrillation in rats. J Cardiovasc Pharmacol. 1993; 22:343–349.
37. Farkas A, Qureshi A, Curtis MJ. Inadequate ischemia-selectivity limits the antiarrhythmic efficacy of mibefradil during regional ischemia and reperfusion in the rat isolated perfused heart. Br J Pharmacol. 1999; 128:41–50.
38. Hirota M, Ohtani H, Hanada E, et al. Influence of extracellular K+
concentrations on quinidine-induced K+
current inhibition in rat ventricular myocytes. J Pharm Pharmacol. 2000; 52:99–105.
39. Frace AM, Maruoka F, Noma A. External K+
conductance of the hyperpolarization-activated current in rabbit cardiac pacemaker cells. Pflugers Arch. 1992; 421:97–99.
40. Hearse DJ, Stewart DA, Braimbridge MV. Hypothermic arrest and potassium arrest: metabolic and myocardial protection during elective cardiac arrest. Circ Res. 1975; 36:481–489.
41. Ellwood AJ, Curtis MJ. The role of shear stress-independent release of nitric oxide in mediating drug-induced coronary vasodilatation can be assessed in the Langendorff constant pressure perfusion preparation. Med Sci Res. 1998; 26:79–81.
42. Curtis MJ, Garlick PB, Ridley PD. Anion manipulation, a novel antiarrhythmic approach: mechanism of action. J Mol Cell Cardiol. 1993; 25:417–436.
43. Gogelein H, Ruetten H, Albus U, et al. Effects of the cardioselective KATP
channel blocker HMR 1098 on cardiac function in isolated perfused working rat hearts and in anesthetized rats during ischemia and reperfusion. Naunyn Schmiedebergs Arch Pharmacol. 2001; 364:33–41.
44. Van Wagoner DR, Lamorgese M. Ischemia potentiates the mechanosensitive modulation of atrial ATP-sensitive potassium channels. Ann N Y Acad Sci. 1994; 723:392–395.
45. Horner SM, Lab MJ, Murphy CF, et al. Mechanically induced changes in action potential duration and left ventricular segment length in acute regional ischemia in the in situ porcine heart. Cardiovasc Res. 1994; 28:528–534.
46. Henry P, Popescu A, Puceat M, et al. Acute simulated ischemia produces both inhibition and activation of K+
currents in isolated ventricular myocytes. Cardiovasc Res. 1996; 32:930–939.
47. Morgan DJ, Huang JL. Albumin decreases myocardial permeability of unbound quinidine in perfused rat heart. J Pharmacol Exp Ther. 1994; 268:283–290.
48. Huang JL, Morgan DJ. Influence of perfusion flow rate on uptake and pharmacodynamics of quinidine in isolated perfused rat heart. J Pharm Sci. 1994; 83:119–123.
49. Barrett TD, Hayes ES, Walker MJ. Lack of selectivity for ventricular and ischemic tissue limits the antiarrhythmic actions of lidocaine, quinidine and flecainide against ischemia-induced arrhythmias. Eur J Pharmacol. 1995; 285:229–238.
50. Barrett TD, Hayes ES, Yong SL, et al. Ischemia selectivity confers efficacy for suppression of ischemia-induced arrhythmias in rats. Eur J Pharmacol. 2000; 398:365–374.
51. Haverkamp W, Breithardt G, Camm AJ, et al. The potential for QT prolongation and pro-arrhythmia by non-anti-arrhythmic drugs: clinical and regulatory implications. Report on a Policy Conference of the European Society of Cardiology. Cardiovasc Res. 2000; 47:219–233.
52. Yang T, Roden DM. Extracellular potassium modulation of drug block of IKr
. Implications for torsade de pointes and reverse use-dependence. Circulation. 1996; 93:407–411.
53. Curtis MJ, Barsby RW, Forster R. Protection by ITF 1300, a heparin ion-pair complex, against arrhythmias induced by regional ischemia and reperfusion in the isolated rat heart: possible mechanism of action. J Cardiovasc Pharmacol. 1995; 25:643–651.