Inhibition of HERG potassium channels by cocaethylene: a metabolite of cocaine and ethanol

Abstract

Objective: To investigate the mechanism by which cocaethylene, a metabolite of cocaine and alcohol inhibits a cardiac delayed rectifier potassium channel. Methods: The cDNA of the HERG potassium channel that underlies I_Kr in humans was transiently expressed in tsA201 cells and currents recorded using the patch clamp technique. Results: The cocaethylene inhibition of HERG is concentration-dependent with an IC₅₀ of 4.0 μM. The inhibition increases over the range of voltages where the channels activate suggesting that cocaethylene binding may be linked to the activation or opening of the channels. Cocaethylene slows the deactivation of the tail current indicating that drug-modified channels are stabilized in the open conformation. Cocaethylene also accelerates inactivation but has no effect of the recovery from inactivation. Conclusions: Cocaethylene inhibits HERG by binding to the activated or open channels and by modulating the kinetics of inactivation. The cocaethylene inhibition of the channels occurs within the range of concentrations detected in the plasma of humans following the ingestion of cocaine and alcohol and is likely to contribute to the potent cardiotoxicity of this drug combination.

Time for primary review 27 days.

This article is referred to in the Editorial by C.A. Karle and J. Kiehn (pages 6–8) in this issue.

1 Introduction

Cocaine is a widely abused drug and its use has been linked to a high incidence of cardiac arrhythmias and sudden death [1]. Although the mechanisms underlying these fatalities are not known, the majority is believed to result from cardiovascular complications [2]. In humans, cocaine acts as a cardiac stimulant increasing heart rate and blood pressure [3,4]. These cardiovascular effects are amplified when cocaine is consumed with alcohol [5] and this drug combination is more cardiotoxic [6] and significantly increases the risk of sudden death [7]. Statistics show that more than half of the people who use cocaine also report the simultaneous use of alcohol [8], which is believed to prolong and enhance the euphoric effects of cocaine [9].

Alcohol alters the metabolism of cocaine in vivo, slowing its degradation and causing the production of cocaethylene, a metabolite of cocaine and ethanol [10]. In whole-animal studies, cocaethylene depresses myocardial function and slows the electrical conduction within the heart, increasing both the QRS and QT intervals [11,12]. The widening of the QRS complex is consistent with a slowing of myocardial conduction and may result from an inhibition of cardiac sodium channels [13]. The increase in the QT interval generally reflects a delay in myocardial repolarization.

The delayed rectifier current plays a key role in myocardial repolarization and is an important determinant of action potential duration. The delayed rectifier current is composed of two components; a slowly activating outward current (I_Ks) and a rapidly activating inward rectifying current (I_Kr) [14]. The recently cloned HERG channel displays similar rectification [15] and pharmacology [16] as the native I_Kr. Reduction of HERG currents by naturally occurring mutations, or inhibition by drugs, prolongs the cardiac action potential and produce long QT syndromes that increase the likelihood of arrhythmia and sudden cardiac arrest [17]. The cocaethylene-induced increase in the QT interval closely resembles the effects of drugs that inhibit HERG channels suggesting that a reduction in the delayed rectifier current may contribute to the repolarization abnormalities observed during the combined use of alcohol and cocaine.

In this study, the effect of cocaethylene on the cardiac HERG potassium channel was investigated. HERG channels are inhibited by cocaethylene with an IC₅₀ of 4.0 μM. Cocaethylene selectively inhibits the current over the range of voltages where channels activate, accelerates inactivation, and slows deactivation. The data suggest that cocaethylene inhibits HERG current by binding to open channels and by modulating inactivation.

2 Methods

2.1 Expression in cultured cells

A standard calcium phosphate precipitation procedure was used to transfect tsA201 cells with HERG cDNA [18]. Within 24 h of transfection, tsA201 cells express 1–5 nA of outward current at +20 mV and up to 10 nA of outward tail current at −80 mV. Some, but not all, tsA201 cells express a relatively small amount of endogenous K⁺ current (50–200 pA). These outward currents are biphasic at +20 mV with a rapidly inactivating and small (≈30 pA) sustained components and the current rapidly deactivates at hyperpolarized voltages. In these studies, the measurement of HERG current was restricted to tsA201 cells expressing >1 nA of current at +20 mV and cells expressing large amounts of the endogenous current (>100 pA) are routinely discarded. Using these criteria, the endogenous currents do not interfere with the accurate measurement of HERG current.

2.2 Electrophysiology

Whole-cell patch recordings were made using sylgard-coated (Dow Corning, Midland, MI, USA) electrodes fashioned from Corning 8161 glass (Wilmad Glass, Buena, NJ, USA). Series resistance was less than 2 MΩ and was 80% compensated. Currents were recorded using an Axopatch 200A amplifier and pclamp software (Axon Instruments, Burlingame, CA, USA). Holding potentials were −80 mV unless otherwise stated. Internal solution consisted of (in mM): 120 KCl, 5 EGTA, 10 HEPES titrated to a pH of 7.4 using KOH. External solution consisted of (in mM): 136 NaCl, 2 KCl, 1 CaCl₂, 1 MgCl₂, 10 HEPES titrated to pH 7.4 with NaOH. The bath temperature was maintained at 25°C using a Medical Systems TC-202 temperature controller (Medical System, Greenvale, NY, USA). Data are reported as the mean±standard error (S.E.M.) and plotted using sigmaplot (Jandel Scientific, Chicago, IL, USA).

3 Results

3.1 Cocaethylene inhibition of HERG

The inhibitory effect of cocaethylene on HERG current was monitored by applying depolarizing pulses during bath application of the drug (Fig. 1A). In the absence of cocaethylene, depolarization causes the channels rapidly activate and inactivate resulting in small currents at depolarized voltages. The large rising phase of the tail current results from the rapid recovery from inactivation and reopening of the channels during repolarization. The slow decay of the tail current reflects the normal closing of the channels. These properties of HERG heterologously expressed in tsA201 cells are consistent with previous studies of these channels [15].

The onset of cocaethylene inhibition of HERG current is rapid, reaching equilibrium within 60 s of the start of the perfusion. Cocaethylene (0.1–100 μM) reduces HERG currents in a concentration-dependent fashion (Fig. 1B). The cocaethylene inhibition was quantified by measuring the peak amplitudes of the tail currents. The currents measured in the presence of cocaethylene were normalized to control current (i.e. drug-free conditions) and plotted versus concentration (Fig. 1C). The smooth curve is a fit of the cocaethylene inhibition to a single site model with an IC₅₀ of 4.0±0.2 μM (n=3). A similar estimate of the IC₅₀ of cocaethylene inhibition is obtained from the reduction in current amplitude of the depolarizing prepulses (IC₅₀=3.4±0.3 μM, n=3).

3.2 Cocaethylene shifts the activation of HERG

In addition to reducing HERG currents, cocaethylene appears to shift the activation of the channels toward more hyperpolarized voltages. Fig. 2B shows a family of currents measured between −40 and +60 mV and the tail currents measured at −80 mV. The peak amplitudes of the tail currents were normalized to the control current measured at +40 mV and plotted versus voltage (Fig. 2D). The amplitude of the tail current increases between −40 and +20 mV: reflecting the voltage-dependent activation of the channels. The smooth curves are fits to the Boltzmann function with a midpoint and slope factor of −12.0±0.2 and 7.7±0.2 mV, respectively (n=6).

Cocaethylene (5 μM) reduces the peak amplitude and slows the decay of the tail current (Fig. 2C). In addition, cocaethylene slightly accelerates the decay of the currents at depolarized voltages suggestive of more rapid inactivation. The effects of cocaethylene on the inactivation and deactivation are considered in more detail in the next sections.

The voltage-dependent activation in the presence of cocaethylene was assessed from the peak amplitude of the tail currents and has a midpoint and slope factor of −20.1±0.3 mV and 7.1±0.3 mV (Fig. 2C) (n=10). In five paired experiments the midpoint of activation was significantly shifted by −8.3±0.8 mV with no change in voltage sensitivity (t-test, n=5, P<0.05). A similar shift in the gating of cardiac I_Kr current has been observed with antiarrhythmic drugs suggesting that this may be a common mechanism of many different HERG channel inhibitors [19,20].

The underlying cause of the hyperpolarizing shift in activation is not known. One possibility is that cocaethylene may preferentially bind to activated but not closed channels. Stabilizing channels in a combination of open and inactivated states would tend to bias the equilibrium away from the closed state, which could contribute to the apparent shift in gating. This was further investigated by examining the voltage sensitivity of the cocaethylene inhibition. An estimate of the steady state inhibition was obtained by comparing the control and drug-modified currents near the end of the depolarizing prepulses. The percent inhibition [(1−I_CE/I_CTRL)·100] was calculated from the ratio of the currents measured before and after application of cocaethylene and plotted versus voltage (Fig. 2D). The cocaethylene inhibition increases for depolarizations greater than −30 mV reaching a maximum of 67.6±2.7% at +10 mV (n=6). This increase in cocaethylene inhibition parallels the voltage-dependent activation of the channels suggesting that activation or opening may play an important role in cocaethylene binding. By contrast, the apparent hyperpolarizing shift in activation and the increase in the inhibition with depolarization are inconsistent with models in which cocaethylene appreciably binds to closed channels.

3.3 Cocaethylene slows the deactivation of HERG

In addition to reducing currents, cocaethylene also slows the time course of deactivation (Fig. 2C). To further investigate the effect of cocaethylene on deactivation the tail currents at −60 mV were measured before (top) and after (bottom) application of 5 μM cocaethylene (Fig. 3). The dashed line is the cocaethylene-modified tail current normalized to the control. The decay of the current measured after application of cocaethylene is sluggish by comparison to the control consistent with a slower rate of deactivation. In control experiments the decay of the tail current was found to be biexponential with fast and slow time constants of 131±6 and 952±41 ms (n=17). After application of 5 μM cocaethylene the tail current is well fitted by two components with time constants of 295±39 and 1240±120 ms (n=12). In paired t-tests, both the fast and slow time constants were found to significantly increase after addition of cocaethylene (P<0.05, n=12). The slow deactivation induced by cocaethylene cannot be attributed to an unusually slow rate of recovery from inactivation, which at −60 mV is not altered by the drug (Fig. 5). The data indicate that cocaethylene binding stabilizes the channels in the open state, thus slowing the closing of the channels.

3.4 Cocaethylene promotes rapid inactivation

Cocaethylene tends to selectively reduce the amplitude of the plateau currents at depolarized voltages (>+40 mV) by promoting a slow decay of the current. An example of this drug-enhanced decay is shown in Fig. 2B and C, where the time course and extent of the current relaxation at +60 mV increases after application of cocaethylene. To quantify this decay the amplitude of the plateau current measured at the end of the depolarizing pulse was normalized to the respective peak current from the same recording. In control experiments, the plateau currents are reduced by 32.0±4.5% with respect to the peak currents (n=7). In the absence of drug, this time-dependent reduction in current amplitude reflects the inactivation of the channels. After application of cocaethylene, the plateau is reduced 58.8±3.4% by comparison to the paired peak current (n=7). The relative amplitude of the plateau current is significantly reduced by cocaethylene (paired t-test, P<0.05) consistent with a more extensive time-dependent current decay.

The cocaethylene-induced decay of the current at depolarized voltages may be linked to changes in channel gating. At +60 mV, HERG channels are maximally activated and the recruitment of additional channels to the open state cannot further contribute to cocaethylene binding. However, in HERG channels strong depolarization favors inactivation and the observed reduction in steady state current coincides with the range of voltages where the currents display significant rectification [21]. Cocaethylene may act by promoting more rapid inactivation. The inactivation time course was measured using a standard triple pulse protocol (Fig. 4A). The decay of the current measured during the test pulses (Fig. 4B and C) is well fitted by a single exponential reflecting the time course of inactivation. In the absence of drug, the inactivation time constants decrease with depolarization, consistent with an increasing rate of inactivation (Fig. 4D). Cocaethylene (5 μM) accelerates the decay of the currents at all voltages. This more rapid inactivation may account for the slow drug-induced decay of the HERG current observed at depolarized voltages (Fig. 2C).

3.5 Cocaethylene does not alter the recovery from inactivation

The effect of cocaethylene on recovery from inactivation was measured using a standard double-pulse protocol (Fig. 5A). At test voltages more negative than −60 mV, the tail currents are biphasic with the rapid rising component reflecting the recovery from inactivation and the slow decay the deactivation of the channels (Fig. 5B). At voltages more depolarized than −60 mV the tail currents are monophasic, reflecting the recovery from inactivation. Depending on the test voltage, the currents were fitted with either one (>−60 mV) or two (<−60 mV) exponentials and the recovery time constants plotted versus voltage (Fig. 5D). Between −140 and −50 mV cocaethylene does not alter the time course of recovery from inactivation. At more depolarized voltages the drug-modified channels appear to recover more rapidly. However, at these voltages the time course of recovery is strongly influenced by the kinetics of inactivation, which is accelerated by cocaethylene (Fig. 4). When the recovery time course is measured at voltages where the contribution of inactivation is minimized (−60 mV), cocaethylene has no effect on the recovery from inactivation. This finding is inconsistent with models where cocaethylene simply acts by stabilizing the channels in the inactivated state, which predicts a reduced rate of recovery from inactivation.

3.6 Cocaethylene inhibition of the S631A mutant channel

The role of inactivation in the cocaethylene inhibition of HERG channels was further investigated by examining its effect on a mutant HERG channel (S631A) that has disrupted inactivation [22]. Removing inactivation dramatically increases the amplitude of the currents measured at depolarized voltages and abolishes the characteristic ‘hooked’ tail current observed when the cell is returned to a hyperpolarized voltage (Fig. 6A). Cocaethylene reduces the amplitude of the S631A current but does not significantly alter the time course (Fig. 6A) suggesting that cocaethylene rapidly inhibits HERG channels near the beginning of the voltage pulse. The cocaethylene inhibition is concentration-dependent with an IC₅₀ of 24.1±4.8 μM (n=5) (Fig. 6B) which is 6-fold greater than that of the wild-type channel (IC₅₀=4.0 μM). This finding is consistent with the data showing that inactivation contributes to the cocaethylene inhibition of the channels (Fig. 4). However, it is important to note that although the S631A mutation weakens the observed inhibition, cocaethylene completely inhibits the currents at high concentrations. Inactivation of the channels is not an absolute requirement for cocaethylene binding.

3.7 Ethanol inhibition of HERG

Cocaethylene inhibition of HERG channels is likely to contribute to the potent cardiotoxicity observed when cocaine and alcohol are consumed. However, in addition to the synthesis of cocaethylene, other mechanisms, such as a direct effect of ethanol on the channel and the potential interaction between ethanol and cocaethylene may also contribute to this inhibition. The direct effect of ethanol on HERG current is shown in Fig. 7A. The currents are shown before and after bath application of 50 mM ethanol, an anesthetizing concentration of the drug [23]. Ethanol reduces HERG currents by 17.6±0.9% (n=7), an effect that is partially reversed during washout. In the presence of 50 mM ethanol, cocaethylene (5 μM) further reduces the current by 56.5±1.3% (n=6). This is statistically indistinguishable from the 53.5±1.5% cocaethylene inhibition of the current measured in the absence of ethanol (n=7) indicating that ethanol does to potentiate the inhibitory effects of cocaethylene. The reduction in current amplitude produced by the combination of ethanol and cocaethylene (65.7±1.7%, n=7) is greater than that of either drug alone indicating that to some extent the inhibitory effects of these drugs are additive.

As discussed in the preceding section cocaethylene causes a time dependent decay in the current at depolarized voltages and slows the time course of current deactivation (Fig. 7B). By contrast, ethanol reduces the amplitudes of the current but does not alter the kinetics. Cocaethylene and ethanol appear to inhibit HERG channels by different mechanisms. The data indicate that at anesthetizing concentrations, ethanol inhibits HERG channels and may contribute to the cardiotoxicity observed during its use with cocaine.

4 Discussion

Cocaethylene is a potent metabolite that is synthesized in humans when cocaine and alcohol are combined. A prominent effect of cocaethylene in the heart is an increase in the QT interval, consistent with delayed cardiac repolarization [11]. HERG channels underlie the rapidly activating component of the delayed rectifier K current that plays a critical role in the repolarization of the heart. Cocaethylene inhibits HERG channels (IC₅₀=4.0 μM) at concentrations that are comparable to those detected in the plasma of victims that died following the use of cocaine and alcohol [24,25] and is likely to contribute to the potent cardiotoxicity observed when cocaine and alcohol are used in combination.

The cocaethylene inhibition of heterologously expressed HERG channels is voltage sensitive, increasing for steps more depolarized than −40 mV and reaching a maximum at approximately +10 mV (Fig. 2D). At physiological pH, cocaethylene is positively charged and electrostatic interactions could contribute to binding. However, the relative voltage insensitivity of the inhibition at potentials more depolarized than +10 mV is difficult to reconcile with a simple electrostatic interaction, which predicts a continuous increase in the inhibition with progressive depolarization. Clearly, a model of cocaethylene binding based solely on electrostatic interactions cannot account for the observed voltage dependence. Cocaethylene inhibition increases over the range of voltages where the channels activate (Fig. 2D), suggesting that the cocaethylene binding may derive a significant fraction of its voltage sensitivity through coupling to channel gating. Unfortunately, at depolarized voltages the open and inactivated conformations of HERG channels are in rapid equilibrium, making it difficult to unequivocally identify the state(s) with which cocaethylene interacts.

Tail current experiments show that cocaethylene significantly slows deactivation indicating that drug-modified channels close slowly at hyperpolarized voltages. This effect is reminiscent of quaternary ammonium pore blockers of K channels that are known to interfere with deactivation [26]. These quaternary ammonium compounds slow deactivation because the blocker must dissociate and diffuse out of the pore before the activation gate can close. Cocaethylene binding may prevent channel closing by a ‘foot in the door’ mechanism similar to what has been proposed for quaternary ammonium compounds [26]. Overall, the good correlation between the channel activation and the development of cocaethylene inhibition (Fig. 2C), and the slowed deactivation suggest an important role for the open state in cocaethylene binding. Whether this involves cocaethylene block of the pore or an allosteric inhibition of permeation is currently not known.

The role of inactivation in cocaethylene binding was also investigated. At +60 mV, the normalized plateau current measured in the presence of cocaethylene is significantly reduced by comparison to that of drug-free controls (Fig. 2B and C). The data suggest that the more extensive decay of the current observed in the presence of cocaethylene may result from enhanced inactivation. This was tested by directly measuring the effect of cocaethylene on the time course of inactivation. Cocaethylene accelerates the decay of HERG currents over a wide range of voltages (Fig. 4). This is unlikely to result from a slow block of open channels because the cocaethylene inhibition of the inactivation-deficient S631A mutant displays no time dependence (Fig. 6B). The data suggest that the more rapid decay observed in the presence of cocaethylene is linked to inactivation. However, the data are inconsistent with a model in which cocaethylene simply binds to and stabilizes the channels in the inactivated conformation. Stabilizing the channels in the inactivated state predicts a reduced rate of recovery from inactivation, a result not observed in these studies (Fig. 5). The data suggest that the more rapid inactivation does not result from an increase in cocaethylene binding to inactivated channels and is consistent with recent work indicating that the cocaine binding is not substantially altered by manipulations that disrupt inactivate [21]. Overall, cocaethylene appears to act by inhibiting open channels and by promoting more rapid inactivation. Both mechanisms are likely to contribute to the cocaethylene-induced inhibition of HERG currents.

4.1 Comparison with previous studies

Many drugs inhibit HERG channels by initially interacting with the open channels [27] and inactivation is proposed to further stabilize drug binding [28]. Similar to the antiarrhythmic drugs, opening of the channel appears to play an important role in regulating cocaethylene binding. However unlike other HERG channel inhibitors, the recovery of cocaethylene-modified channels is rapid, a finding that is inconsistent with a large increase in the binding affinity as the channels inactivate. The data do not provide strong support for an increase in cocaethylene binding to inactivated channels.

The inhibition of HERG channels by cocaine, the parent compound of cocaethylene, has recently been investigated [21,29]. Chemically, cocaine and cocaethylene are similar except for the substitution of an ethyl moiety, which causes the later drug to be slightly more hydrophobic. The IC₅₀ of cocaine inhibition (≈6 μM) is not substantially different from that of cocaethylene. Similar to cocaine, cocaethylene inhibition of HERG increases over the range of voltages where the channels open, causes a hyperpolarizing shift in activation, and promotes more rapid inactivation. A recent study has shown that a quaternary derivative of cocaine preferentially inhibits HERG when applied to the cytoplasmic side of the channel [29]. The high dependence of cocaethylene and cocaine inhibition on activation and the preferential binding of cocaine from the cytoplasmic side is characteristic of many HERG channel inhibitors [30,31].

4.2 Physiological relevance

Statistics show that approximately 75% of the population that regularly use cocaine simultaneously consume alcohol [8], which is believed to potentiate and prolong the effects of the drug [9]. In the presence of ethanol, the metabolism of cocaine is altered, resulting in the synthesis of cocaethylene rather than the weaker cocaine metabolites. The role of cocaethylene in producing the potent cardiotoxicity associated with the combined use of cocaine and alcohol is not known. In humans, the plasma levels of cocaethylene measured during the use of alcohol and cocaine are reported to be equivalent to, and in some cases, greater than the levels of cocaine [32,33]. Concentrations up to 2 μM have been measured in the plasma of emergency room patients that tested positive for cocaine and alcohol [33] and even higher concentrations (5 μM) are detected in postmortem tissue samples [25]. Cocaethylene inhibits HERG channels (IC₅₀=4.0 μM) well within the range of concentrations detected in the plasma of humans following the use cocaine and alcohol.

Ventricular dysrhythmias, including torsade de pointes, have been observed following acute administration of cocaethylene in animal studies [11]. Similar ventricular arrhythmias are commonly reported in humans suffering from genetic long QT syndromes or following treatment with antiarrhythmic drugs [17]. Cocaethylene inhibition of HERG channels may weaken cardiac repolarization and thereby contribute to the higher incidence of arrhythmias and sudden death associated with the consumption of cocaine and alcohol.

Acknowledgements

I would like to thank Drs. Richard Horn and Manuel Covarrubias and for helpful comments on the manuscript. This work was supported by a grant from the American Heart Association (9730216N).

References