The purpose of the present study was to comparatively evaluate human HERG currents and QT intervals following challenge with suspected torsadogenic and nontorsadogenic drugs. Various concentrations of 14 different drugs were initially evaluated in terms of their relative potency to block IHERG in stably transfected human embryonic kidney cells. Four general categories of drugs were identified: high-potency blockers (IC50 < 0.1 μM) included lidoflazine, terfenadine, and haloperidol; moderate-potency blockers (0.1 μM < IC50 < 1 μM) included sertindole, thioridazine, and prenylamine; low-potency blockers (IC50 > 1 μM) included propafenone, loratadine, pyrilamine, lovastatin, and chlorpheniramine; and ineffective blockers (IC50 > 300 μM) included cimetidine, pentamidine, and arsenic trioxide. All measurements were performed using similar conditions and tested acute drug effects only (<30 min of drug exposure per measurement). Since two of the drugs that were ineffective IHERG blockers, arsenic trioxide and pentamidine, have been associated with cardiac repolarization delays (QT interval lengthening) and torsades de pointes ventricular arrhythmias in patients, we chose to evaluate them further using the isolated perfused rabbit heart model. Neither arsenic trioxide nor pentamidine had any significant effect on QT intervals in this model, even at relatively high (micromolar) concentrations. Similar results were obtained for loratadine in this model. When the hearts were challenged with a known torsadogenic drug such as cisapride, significant QT lengthening was rapidly induced. These results demonstrate that arsenic trioxide and pentamidine are essentially devoid of direct acute effects on cardiac repolarization or inhibition of IHERG.
QT interval prolongation and an associated severe life-threatening ventricular arrhythmia, torsade de pointes (TdP), are a high-priority cause for concern in drug development and regulatory safety evaluation (Fenichel et al., 2004). Of the drugs recently removed from the U.S. market, one of the most common causes has been QT interval-related cardiac toxicity. This toxicity was discovered either after approval during clinical use or in late-stage clinical trials rather than in early drug development, with significant resultant difficulties. Notably, women have tended to be much more susceptible than men to these cardiotoxic drug effects (Makkar et al., 1993; Lehmann et al., 1997).
Given the medical and economic consequences of this issue, the International Conference on Harmonization established an Expert Working Group to draft guidance recommending the incorporation into drug development of preclinical models predictive of QT interval prolongation and proarrhythmia. This draft guidance, ICH S7B, was published in revised form for comment in the Federal Register in June 2004 (Attachment 5). ICH S7B recommends a testing strategy comprised of both in vitro and in vivo assays considered likely to be predictive for drug-induced QT interval prolongation and proarrhythmia.
A large number of drugs from a wide variety of classes, including antihistamines, antipsychotics, antiarrhythmics, antibiotics, and gastrointestinal prokinetic agents, have been associated with the syndrome of TdP, a potentially fatal form of ventricular cardiac arrhythmia. Many of the drugs that have been associated with drug-induced development of TdP arrhythmias have also been shown to block the rapid component of the delayed rectifier repolarizing potassium current, IKr, in ventricular cardiomyocytes (Nattel, 1999). The major channel protein responsible for IKr is encoded by the human ether-a-go-go-related gene (HERG) gene, also known as KCNH2 (Curran et al., 1995). Zhou et al. (1998) demonstrated that expression of the HERG gene in stably transfected HEK cells produces a current, IHERG, that has the characteristics of IKr. This cell line has proved to be useful for studies examining pharmacological blockade of IKr. Since HEK cells produce few other currents, the results obtained from these cells are typically much easier to acquire and interpret than are those from studies that use freshly isolated ventricular cardiomyocytes. There are a growing number of drugs being monitored by the FDA and others because of their suspected involvement in the development of TdP-like cardiac arrhythmias. At present, there are limited systematic data available regarding the influence of many suspected arrhythmogenic compounds on IHERG. Thus, the purpose of the present study was to determine whether or not a subset of these compounds with a range of risk for TdP arrhythmias can block IHERG and, if so, what their relative potency is.
We comparatively evaluated 14 compounds for their ability to block IHERG in a stably transfected cell culture model system. Many of these compounds, including lidoflazine, terfenadine, haloperidol, sertindole, thioridazine, prenylamine, propafenone, pyrilamine, pentamidine, and arsenic trioxide, have been associated with clinical long QT syndrome and TdP arrhythmias in patients. A subset of these drugs (lidoflazine, terfenadine, haloperidol, sertindole, thioridazine, and propafenone) have been shown previously to be relatively potent blockers of IHERG (Hanley and Hampton, 1983; Connolly et al., 1983; Drolet et al., 1999), whereas there were limited and sometimes conflicting reports regarding the effects of loratadine, prenylamine, pyrilamine, pentamidine, and arsenic trioxide on IHERG (Guijarro et al., 1976; Drolet et al., 2004; Ficker et al., 2004). The present study, therefore, set out to conduct a systematic comparison of all of these compounds to determine their relative potencies for blocking IHERG. Several “negative control” compounds, including lovastatin, chlorpheniramine, and cimetidine, were also evaluated under similar conditions.
Two compounds in particular, arsenic trioxide and pentamidine, had little or no effect on IHERG, although both of these compounds have been associated with long QT syndromes and TdP-like arrhythmias in patients (Wharton et al., 1987; Bibler et al., 1988; Stein et al., 1991; Eisenhauer et al., 1994; Ohnishi et al., 2000; Barbey and Soignet, 2001; Unnikrishnan et al., 2001). Thus, to further evaluate the potential for these compounds to influence cardiac repolarization and cause arrhythmias, we also examined their influence in the isolated perfused (Langendorff) rabbit heart model. The isolated rabbit heart has been shown previously to be sensitive to torsadogenic drugs (Asano et al., 1997; Drici et al., 1999; Liu et al., 1999) and, hence, would fit into the preclinical strategy of early detection of torsadogenic hazard. Our results suggest that neither arsenic trioxide nor pentamidine have any significant acute effect on cardiac repolarization.
Materials and Methods
Drugs and Chemicals. Racemic methadone was obtained from Eli Lilly & Co. (Indianapolis, IN). Lidoflazine was obtained from Research Diagnostics (Flanders, NJ). Loratadine was obtained from Schering Plough (Kenilworth, NJ). All other drugs and chemicals were purchased from Sigma-Aldrich (St. Louis, MO). For each compound tested, a concentrated stock solution (>10 mM) was prepared by dissolving the powder in deionized Milli-Q water or dimethyl sulfoxide. Small aliquots of the concentrated stocks were immediately frozen and stored at –80°C. Aliquots were thawed immediately before use and diluted to the desired final concentration in Tyrode's solution.
Cell Culture. HEK 293 cells that stably express HERG were obtained from Dr. Craig January (Zhou et al., 1998). These cells were maintained using the culture conditions previously described by Zhou et al. (1998).
Whole-Cell Patch Clamp. Coverslips with stably transfected HERG-expressing HEK cells were placed in a recording chamber (ΔT3 Dish; Bioptechs, Butler, PA), mounted on the stage of an inverted microscope (Olympus IX50; Olympus, Tokyo, Japan). The superfusion rate was 0.5 ml/min. The time needed to change the solution near the cell was estimated to be less than 0.2 s. Output signals from the amplifier were digitized using a DigiData 1200 A/D, D/A board in conjunction with an IBM (White Plains, NY) compatible computer. This system was used to generate command pulses and to acquire and analyze the data.
Whole-cell voltage-clamp measurements were obtained by ruptured patch-clamp technique. For this purpose, the Axopatch 200B Amplifier (Molecular Devices, Sunnyvale, CA) was used. Recording electrodes (2–5 MΩ resistance) were made from thin-walled borosilicate glass TW150F (WPI, Sarasota, FL). Measurements were checked for possible rundown during the experiment. Creation of voltage-clamp command pulse protocols and data acquisition were controlled by pCLAMP software (version 8.01; Molecular Devices) installed on a personal computer (Compaq Deskpro 4000; Compaq, Palo Alto, CA).
The bath solution consisted of 137 mM NaCl, 5.1 mM KCl, 2 mM CaCl2, 10 mM HEPES, 1 mM MgCl2, and 10 mM glucose. The pH was adjusted to 7.4 by addition of NaOH. The pipette filling solution contained 140 mM KCl, 4 mM Mg-ATP, 5 mM EGTA, 1 mM MgCl2, and 10 mM HEPES, pH adjusted to 7.2 with KOH. Experiments were performed at room temperature, 22.0 ± 0.5°C.
To study dependence of steady-state block of HERG channels on drug concentration in HEK cells, membrane potential was switched from holding –80 to + 20 mV for 2 s following return to –50 mV for 6 s in intervals of 15 to 30 s. Tail currents were measured at –50 mV in control and in the presence of the drug at concentrations determined empirically.
For each batch of cells, this protocol was applied for 3 min to rule out possible rundown of HERG currents. Drug testing was performed only on cells showing less than 5% of decline from initial value of tail current. In addition, we performed vehicle control experiments in parallel with drug testing to ensure that the repeated recordings and the duration of the experiment did not significantly influence the amount of HERG current produced.
Each cell served as its own control. All raw measurements of tail currents were performed using CLAMPFIT program, a part of pCLAMP software (version 8.1; Molecular Devices). Results were transferred to Origin 6.1 (OriginLab Corp., Northampton, MA) and/or Microsoft Excel (Microsoft, Redmond, WA) spreadsheets for further analysis.
Animals. New Zealand White rabbits (3–4 months old, 3–3.5 kg) were obtained from Covance Laboratories, Inc. (Denver, PA). This investigation conforms to the Guide for the Care and Use of Laboratory Animals published by the U.S. National Institutes of Health (publication 85-23, revised 1996), and all experiments were conducted in accordance with the guidelines of the Georgetown University Animal Care and Use Committee.
Isolated Perfused Rabbit Heart Preparation. The isolated perfused rabbit heart preparation was performed essentially as we have described previously (Liu et al., 2003). Following anesthesia with pentobarbital (30 mg/kg i.v.), heparin (10,000 units i.v.) was administered to prevent clotting. The heart was then rapidly removed through a median sternotomy incision and placed in an ice-cold Tyrode's solution. After the removal of extraneous fat and connective tissue, the aortic root was cannulated for retrograde perfusion. The heart was then mounted in a Langendorff perfusion apparatus. The right atrium was removed, and the AV node was cauterized to ablate endogenous pacing. The heart was paced from an electrode positioned in the ventricular endocardium. The perfusate was maintained at 37°C by immersion in a temperature-controlled bath containing the perfusion solution and delivered to the aortic inflow cannula by a Masterflex pump at a constant flow rate of 10 ml/min in a noncirculating manner. Hearts were paced at a fixed cycle length of 400 ms (twice diastolic threshold) using a Pulsemaster A300 cardiac stimulator and the Stimulus Isolator A360 (World Precision Instruments, Inc., Sarasota, FL). Four silver-silver chloride electrode pellets were positioned in a simulated “Einthoven” configuration with the reference and “foot” electrodes fixed beneath the heart on the walls of a tissue bath. All hearts were perfused with Tyrode's solution to equilibrate the preparation for 30 to 60 min prior to the baseline measurements.
The signals were amplified by an ECG amplifier that allowed the simultaneous recording of three signals, filtered selectively for 60-Hz noise, and ECG recordings were acquired on a strip chart with a paper speed of 100 mm/s. All signals were digitized at sampling rate of 2 kHz and stored on a computer hard drive and CD for later analysis.
After baseline ECG recordings were obtained, the QT interval and QRS duration were observed until three separate recordings 2 to 3 min apart with stable measurements (<5% change) have been obtained. Each test drug was evaluated over a range of concentrations, beginning with a low concentration and sequentially increasing the concentration after 30 min. All measurements were made during the last 5 min of perfusion with each drug concentration. The QT values recorded at the end of each perfusion period were used for comparative analysis.
Statistical Analyses. Unless otherwise noted, all data are expressed as mean ± S.E.M. One-way analysis of variance was used for multiple comparisons, and the Student's t test was used when only two groups were compared. p < 0.05 was required to reject the null hypothesis.
We tested 14 drugs for their ability to block cardiac HERG K+ currents in a stably transfected HEK cell line (Zhou et al., 1998). The drugs were tested using the whole-cell patch-clamp recording technique as described previously (Katchman et al., 2002). For each drug tested, we calculated the concentration that inhibited 50% of IHERG (IC50) by comparing the amount of IHERG present in the presence of a given concentration of drug relative with the amount of IHERG present in the same cell prior to the administration of drug. These results are summarized in Table 1.
The drugs are listed in order of potency for blockade of IHERG. Lidoflazine, terfenadine, and haloperidol were by far the most potent HERG inhibitors of the compounds tested, with IC50 values in the midnanomolar range. Notably, all three drugs blocked IHERG in a use-dependent manner, as exemplified by the results shown for terfenadine in Fig. 1. For comparison, control tracings recorded under similar conditions but in the absence of any drugs are shown in Fig. 2. In contrast to the recordings performed in the presence of terfenadine and other use-dependent blockers of IHERG (Fig. 1; Table 1), there was no significant change in peak IHERG tail current density following repeated (n = 20–30) measurements in the absence of drugs (Fig. 2).
Some of the drugs showing intermediate potency for IHERG block, such as sertindole and prenylamine, were also found to act in a use-dependent manner. Most of the less potent drugs tested did not show any signs of use dependence. For example, lovastatin blocked IHERG in a concentration-dependent manner that showed no signs of use dependence (Fig. 3).
For three of the drugs tested, cimetidine, pentamidine, and arsenic trioxide, we were unable to derive IC50 values because little or no inhibition of IHERG was observed in the presence of these drugs. In the cimetidine example, partial block (< 50%) of IHERG was achieved when the drug was applied at the highest concentration tested (10 mM). Upon removal of cimetidine (washout), nearly complete recovery of current was observed (data not shown).
Pentamidine, on the other hand, appears to represent a special case. Although there was no significant inhibition of IHERG at concentrations up to 1 mM pentamidine (p > 0.05, n = 11), we observed a much faster decay of the tail currents in the presence of pentamidine (Fig. 4). Upon removal of pentamidine (“washout”), we observed nearly complete recovery of the decay rate (Fig. 4, bottom set of tracings in top). Subsequent application of the known IHERG blocker, haloperidol (1 μM), caused strong suppression of IHERG tails within seconds after administration. Although we could not derive an IC50 for pentamidine block of peak HERG tail currents due to the weak effect of this drug on IHERG tail amplitude (Fig. 4, middle), we were able to calculate an IC50 of 205 μM for its effect on the decay rate (τ) of IHERG deactivation (Fig. 4, bottom).
Of the compounds tested in this study, arsenic trioxide appeared to be the least effective blocker of IHERG. At the highest concentration tested (300 μM), weak inhibition of IHERG tail currents was seen, but this was not found to be significant (p > 0.05, n = 9). There was no question that IHERG could be blocked in these cells by other drugs since application of either terfenadine (1 μM) or haloperidol (1 μM) readily blocked IHERG in the same cells (Fig. 5). Thus, arsenic trioxide does not appear to directly influence IHERG at concentrations ≤ 300 μM.
To determine whether arsenic trioxide and pentamidine may directly cause repolarization delays through blockade of endogenous cardiac K+ currents or other mechanisms, we examined the effects of these drugs on QT interval duration in the isolated perfused (Langendorff) rabbit heart preparation. Examples of our ECG recordings for these experiments are shown in Fig. 6. Arsenic trioxide (10 μM) had little effect on the ECG, yielding QT durations that were not distinct from control (Fig. 6A). In contrast, subsequent perfusion with methadone (3 μM) led to a small but noticeable 4.4% increase in QT duration from 225 (control) to 235 ms. Methadone's effect could be partially reversed by removal of the drug (washout). When the same heart was then challenged with cisapride (1 μM), QT duration increased to 255 ms (13.3% increase relative to control). Thus, the heart was clearly responsive to known HERG-blocking drugs such as methadone (Katchman et al., 2002) and cisapride (Bran et al., 1995) but not to arsenic trioxide.
A similar finding was observed for pentamidine (Fig. 6B). Although a mild increase of 5 ms in QT duration was observed in the presence of pentamidine in this experiment, there was little concentration dependence associated with this response since QT duration was similar in the presence of both concentrations (3 and 10 μM) of pentamidine tested in this experiment. In contrast, subsequent perfusion with prenylamine clearly lengthened QT duration in a concentration-dependent manner, thereby indicating that the QT interval was responsive to a known HERG-blocking drug (prenylamine), whereas pentamidine had little effect under similar conditions. Subsequent experiments supported these examples by showing that neither arsenic trioxide nor pentamidine elicited any significant fluctuations in QT interval duration across broad concentration ranges, extending to 50 and 30 μM, respectively (Fig. 7, A and B, respectively).
Because loratadine also showed some use-dependent ability to block IHERG at micromolar concentrations, and because there are conflicting reports about loratadine's effects on IHERG in different laboratories (Ducic et al., 1997; Crumb, 2000; Davie et al., 2004), we also evaluated loratadine using the Langendorff-perfused rabbit heart model. Our data show that loratadine has no significant effect on QT length at concentrations up to and including 50 μM (Fig. 7C). The only apparent effect was a tendency for the QT to become shorter in the presence of loratadine, but this effect was not found to be significant (p > 0.05, n = 4). No arrhythmias or other ECG abnormalities were observed during perfusion with loratadine in these experiments.
In contrast, methadone, which was previously shown to block IHERG in a concentration-dependent fashion (IC50 = 10–20 μM) (Katchman et al., 2002; Kornick et al., 2003) led to significant increases in QT interval duration in this model (Fig. 8). A clear trend toward increased QT duration was observed in the presence of methadone at concentrations as low as 1 to 3 μM (10–12% increase relative to control; Fig. 8B), with significant increases apparent at 10 (21 ± 5%, n = 4, p < 0.05) and 30 (44 ± 12%, n = 4, p < 0.001) μM. No arrhythmias or other ECG abnormalities were observed following methadone perfusion in these experiments.
We tested 14 drugs for their ability to block IHERG in a stably transfected cell line. Our data suggest that these drugs could generally be segregated into four groups based on their potency and mechanism of IHERG inhibition. For example, the most potent IHERG blockers (IC50 < 100 nM) were lidoflazine, terfenadine, and haloperidol, which all interfered with IHERG in a use-dependent manner. A second category of compounds (sertindole, thioridazine, and prenylamine) blocked IHERG with moderate potency (IC50 = 0.1–1 μM) with some use dependence. The third and largest category consisted of low-potency IHERG blockers (IC50 = 1–100 μM) that, with one exception (loratadine), did not induce any apparent use dependence with respect to IHERG inhibition. This third category includes propafenone, loratadine, pyrilamine, lovastatin, and chlorpheniramine. The fourth and final category consists of three compounds: cimetidine, pentamidine, and arsenic trioxide, which were essentially inactive with respect to IHERG blockade in our assays even when evaluated at excessively high concentrations. Although several of the drugs tested here have been shown to block IHERG in previous studies, this is the first report showing that prenylamine, pyrilamine, and chlorpheniramine can block IHERG in a concentration-dependent manner.
Although prenylamine has long been associated with clinical QT lengthening and TdP-like arrhythmogenic events (Guijarro et al., 1976; Oakley et al., 1980), there do not appear to be any published reports evaluating its influence on IHERG directly. Thus, the present study provides the first quantitative assessment of prenylamine with respect to IHERG. The IC50 for this block was 590 nM, similar to that of sertindole and thioridazine, suggesting that it could be contributory to the observed cardiac repolarization and arrhythmogenic problems associated with prenylamine use. Although we did not perform comprehensive evaluation of prenylamine's actions using the Langendorff model, we did observe QT lengthening by prenylamine in each of the three experiments that were performed with this drug (e.g., see Fig. 6B).
There are few reports of QT problems or ventricular arrhythmias and no reports of IHERG blockade by pyrilamine or chlorpheniramine. Like chlorpheniramine, we included lovastatin as a “negative control” but found that it also blocked IHERG in a concentration-dependent manner with an IC50 of 7 μM, a value that is remarkably close to that (12.5 μM) recently reported for IHERG block in another study (Wible et al., 2005). Although each of these compounds did, in fact, block IHERG in a concentration-dependent manner, the concentrations required to achieve such inhibition are far above (>100-fold) typical plasma concentrations for these drugs in patients. Thus, the IHERG results reported for these compounds are consistent with their cardiac safety record clinically.
Loratadine is a widely used nonsedating antihistamine that is thought to be generally “safe” clinically (Woosley, 1996) despite apparently conflicting reports about its ability to block IHERG and IKr (Ducic et al., 1997; Crumb, 2000; Davie et al., 2004). Our IHERG results with loratadine (IC50 = 4 μM) are remarkably similar to those reported by Davie et al. (2004) (IC50 = 5 μM). In contrast, Crumb (2000) reported that loratadine had much greater potency with respect to block of IHERG amplitude (IC50 = 173 nM), whereas Ducic et al. (1997) reported that loratadine had little or no effect on IHERG in microinjected Xenopus oocytes and IKr in isolated ventricular myocytes at concentrations up to 1 μM. To help resolve the issue of whether or not loratadine has a significant effect on cardiac repolarization, we used the isolated Langendorff-perfused rabbit heart model to assess ECG effects. As our results showed (Fig. 7C), loratadine did not significantly alter QT duration even during challenges with relatively high (micromolar) concentrations, thereby indicating that it had little or no effect on the duration of cardiac repolarization in this model. These results are reminiscent of those reported by Davie et al. (2004), who showed that despite micromolar blockade of IHERG in transfected cells, loratadine had no appreciable effect on either action potential duration or IKr in guinea pig ventricular myocytes.
The fourth category of compounds (cimetidine, pentamidine, and arsenic trioxide) was relatively ineffective at blocking IHERG. Despite these observations, both pentamidine (Wharton et al., 1987; Bibler et al., 1988; Eisenhauer et al., 1994) and arsenic trioxide (Ohnishi et al., 2000; Barbey and Soignet, 2001; Unnikrishnan et al., 2001) have been associated with cardiac repolarization delays (QT lengthening) and TdP-like arrhythmias in patients. Recent reports have suggested that although these drugs may not interfere with IHERG directly, they may nevertheless cause reductions in IHERG via interference with trafficking of HERG proteins to the membrane (Ficker et al., 2004; Cordes et al., 2005; Kuryshev et al., 2005). In the case of pentamidine, our results are consistent with these previously published data. Although we observed a dose-dependent decline in the τ of IHERG deactivation with pentamidine, this effect required extremely high concentrations (IC50 = 205 μM) that are far above the maximal plasma concentration (Cmax = 10–30 nM) (Vohringer and Arasteh, 1993) reported in patients taking recommended therapeutic doses of pentamidine.
In the case of arsenic trioxide, however, there are independent reports suggesting that it lengthens action potential duration and rate-corrected QT intervals in vivo in guinea pig hearts (Chiang et al., 2002) and that it blocks IHERG in transfected Chinese hamster ovary cells at low micromolar concentrations (Drolet et al., 2004). In contrast, our data suggest that arsenic trioxide has no significant acute effect on IHERG in stably transfected HEK cells at concentrations up to 300 μM.
The lack of a QT response for pentamidine or arsenic trioxide in the isolated perfused rabbit heart model is consistent with the our finding that neither of these compounds had much effect on IHERG at relevant concentrations and suggests that neither of these compounds cause any direct acute effect to block IHERG. In contrast, Drolet et al. (2004) have published data suggesting that submicromolar concentrations of arsenic trioxide were sufficient to achieve significant inhibition of IHERG. We did not see this effect in our experiments, although certain differences in assay systems exist. For example, the Drolet study (2004) used transiently transfected Chinese hamster ovary cells, whereas we used stably transfected HEK cells in the present study. Conceivably, these differences and/or other methodological differences could account for the apparent discrepancy in our results with respect to direct blockade of IHERG by arsenic trioxide. Further testing is required to resolve this apparent discrepancy.
In contrast, our arsenic trioxide results are consistent with those recently published by Ficker et al. (2004). Both we (present study) and Ficker et al. (2004) showed that arsenic trioxide has no significant direct acute effect on IHERG in transfected cells. Although we have not examined the effects of prolonged exposure to arsenic trioxide, Ficker et al. (2004) demonstrated that 24-h exposure to arsenic trioxide leads to a concentration-dependent decrease in IHERG, with the IC50 for this effect being ∼1.5 μM. From these and other data, Ficker et al. (2004) have suggested that arsenic trioxide may reduce IHERG by interfering with HERG protein trafficking. As indicated above, similar observations regarding HERG protein trafficking interference have been made recently with pentamidine (Cordes et al., 2005; Kuryshev et al., 2005), thus suggesting that these two drugs may act through similar mechanisms with respect to their arrhythmogenic potential (Eckhardt et al., 2005).
Alternatively, it is possible that drugs like arsenic trioxide and pentamidine have arrhythmogenic activity that is independent of IHERG and QT. It has recently been suggested, for example, that QT interval may not be the best marker for ventricular proarrhythmia potential (Brown, 2004; Shah and Hondeghem, 2005). Shah and Hondeghem (2005) have instead proposed a novel “TRIaD” (triangulation, reverse use dependence, instability of action potential, and dispersion) analytical method. Interestingly, there does appear to be T-wave morphology changes that could be indicative of “triangulation” during some of our drug perfusions (e.g., prenylamine; see Fig. 6B). It may be interesting to see if the TRIaD method will be predictive of arrhythmogenic potential for arsenic trioxide and pentamidine, but additional data are needed to make such a determination. Other potential mechanisms such as indirect effects that could lead to proarrhythmic conditions should also be evaluated.
In summary, our data show that unlike many other torsadogenic and suspected torsadogenic drugs, neither pentamidine nor arsenic trioxide appear to directly block the channel to interfere with IHERG or cause QT prolongation acutely in well-established model systems, even at relatively high nontherapeutically relevant concentrations. A possible mechanistic explanation for the clinical QT and torsadogenic effects of these compounds may reflect impaired trafficking of HERG proteins (Eckhardt et al., 2005), as suggested by recent studies (Ficker et al., 2004; Cordes et al., 2005; Kuryshev et al., 2005). Further evaluation of how these and other non-HERG blockers may elicit dangerous cardiac arrhythmias is warranted.
- Received July 27, 2005.
- Accepted November 7, 2005.
This work was supported by grants from the United States Food and Drug Administration Center for Drug Evaluation and Research and Office of Women's Health, by the National Institutes of Health (HL58743), and by the Center for Education and Research on Therapeutics, Agency for Healthcare Research and Quality (U18 HS010385).
ABBREVIATIONS: TdP, torsade de pointes; HERG, human ether-a-go-go-related gene; HEK, human embryonic kidney; ECG, electrocardiogram; As2O3, arsenic trioxide.
- The American Society for Pharmacology and Experimental Therapeutics