Abstract
HERG (human-ether-a-go-go-related gene) encodes for a cardiac potassium channel that plays a critical role in defining ventricular repolarization. Noncardiovascular drugs associated with a rare but potentially lethal ventricular arrhythmia (Torsades de Pointes) have been linked to delayed cardiac repolarization and block of hERG current. This brief overview will discuss the role of hERG current in cardiac electrophysiology, its involvement in drug-induced delayed repolarization, and approaches used to define drug effects on hERG current. In addition, examples of hERG blocking drugs acting differently (i.e., overt and covert hERG blockade due to multichannel block) together with the utility and limitations of hERG assays as tools to predict the risk of delayed repolarization and proarrhythmia are discussed.
Introduction
Recent years have witnessed a heightened awareness regarding cardiovascular safety of drugs. In particular, a number of noncardiovascular drugs representing disparate chemical structures and therapeutic classes (including antihistamines, antibiotics, antifungals, antipsychotic, and gastrointestinal prokinetic agents) have been shown to delay ventricular repolarization. This effect has been linked to sudden cardiac death attributed to a rare, drug-induced polymorphic ventricular tachycardia termed Torsades de Pointes (TdP, Dessertennes, 1966) that may degenerate to ventricular fibrillation. In response to heightened awareness, actions from the pharmaceutical industry and regulatory agencies have included warnings, relabeling, and withdrawals from the marketplace. The reader is encouraged to read various articles reviewing the evolving scientific and regulatory perspectives (Haverkamp et al., 2000, Malik and Camm, 2001; Fermini and Fossa, 2003; Fenichel et al., 2004; Roden, 2004).
Drug-induced TdP occurs infrequently following administration of agents that alter or delay cardiac repolarization. Delayed repolarization is manifest as prolongation of the QT interval on the electrocardiogram. When properly controlled, QT prolongation may provide an antiarrhythmic property for therapeutic drugs specifically targeting the heart (see Shah and Hondeghem, 2005). However, drug-induced delayed repolarization with noncardiovascular drugs is generally considered as potentially harmful; such effects leading to QT interval prolongation have been termed acquired long QT syndrome. The rare occurrence of TdP and potentially lethal consequences necessitate the use of surrogate markers for this drug-induced event.
The most direct approach to evaluate drug-induced QT interval prolongation is to measure the interval (itself a surrogate marker). However, this approach has limitations related to the confounding effect of heart rate on the QT interval, the accurate measurement of the end of the interval, and the ability to reliably detect small changes in the QT interval (<5% of the interval) (Malik, 2002; Desai et al., 2003). In addition, while a link has been established between drug-induced QT prolongation and proarrhythmia (see later), the exact nature of that relationship is uncertain and likely varies with different drugs. Thus, additional surrogate markers are used to evaluate proarrhythmic risk related to altered ventricular repolarization; these include evaluation of drug effects on a cardiac repolarizing potassium current encoded by the hERG gene (termed hERG current studies), as well as studies that evaluate changes in the cardiac action potential configuration (termed action potential duration APD or repolarization studies). Congenital long QT syndromes, resulting from mutations in either cardiac potassium or sodium ion channel proteins or subunits, are also associated with QT prolongation, TdP, and sudden cardiac death (see Towbin and Vatta, 2001, Kass and Moss, 2003; Chiang, 2004 for reviews).
Ventricular repolarization is a dynamic electrophysiologic process that ensures the orderly termination of each heartbeat and modulates cardiac excitability and contractility. HERG is one of multiple cardiac ionic channels that provides repolarizing current and defines ventricular repolarization (Sanguinetti et al., 1995; Trudeau et al., 1995). Virtually all drugs that delay cardiac repolarization and cause TdP block the outward flow of repolarizing K+ current through hERG channels. Thus, hERG current block has become an accepted surrogate marker for cardiac proarrhythmia. This article reviews the role of hERG in ventricular repolarization and the utility of various hERG assays to predict altered repolarization and proarrhythmia. In addition, examples of hERG study results with clinically used drugs are provided to demonstrate the utility and weaknesses of the functional hERG current assay.
Primer on Ventricular Repolarization
The electrocardiogram (ECG) records the changing voltage on the body surface imparted by currents that flow throughout the heart. In general, the QT interval on ECG recordings represents the time from start of ventricular depolarization (Q-wave deflection) to end of ventricular repolarization (end of T-wave). As the time required for ventricular depolarization is short (typically < 60 msec in humans compared to 400 msec for the entire QT interval), the majority of the QT interval reflects ventricular repolarization. Thus, the QT interval is generally regarded as representing the duration of ventricular electrical activity, with QT prolongation predominantly reflecting delayed ventricular repolarization.
The cardiac action potential represents the characteristic configuration of transmembrane voltage (or potential) that occurs with each heartbeat (see Figure 1). This waveform represents the summation of multiple ion currents that turn on (activate) and off (deactivate or inactivate) with each beat. In the working ventricular myocardium, depolarization reflects the rapid activation of fast inward sodium current that depolarizes the membrane from the resting membrane potential to positive values to ensure rapid conduction; this current (analogous to sodium current in nervous tissue) is completed within a few milliseconds. On the ECG, the spread of ventricular depolarization is manifest as the spike-like QRS waveform. In contrast to depolarization, ventricular repolarization is much slower, taking a few hundred msec (much slower than nervous tissue) to bring the transmembrane potential back towards the resting potential. The slower time-course of repolarization is responsible for most of the approximately 400 msec QT interval duration.
Much of each ventricular action potential is spent during what is termed the “plateau” phase. During this time soon after the upstroke, a fine balance of declining inward (depo1arizing) current and increasing outward (repolarizing) currents is responsible for the relatively slow process of repolarization. As the plateau progresses, growing net outward current predominates, leading to terminal repolarization and the termination of the action potential (see Figure 1). Early during the plateau phase of the action potential, a depolarizing inward calcium current (carried through L-type calcium channels) is primarily responsible for sustaining membrane depolarization. Gradually, this current declines as a small but critical voltage-gated potassium current is activated, generating outward current, bringing the plateau to an end and initiating terminal repolarization (see Figures 1 and 2). This current has been termed IKr (“I” for current, “K” referring to potassium as the charge carrier, and ‘r” referring to the speed of current activation [specifically “r” for rapid, Sanguinetti and Jurkiewicz, 1990]). In humans, IKr flows through channels encoded by the human ether a-go-go-related gene hERG. Potent blocking drugs of IKr or hERG current (such as the specific blocking agent E-4031) have been used to delineate the contribution of this current during repolarization of the action potential.
HERG: What Is It?
HERG is an acronym for human ether-a-go-go-related gene, the gene that encodes the pore forming subunit of the delayed rectifier IKr channel in humans. The hERG channel is one of a family of ion channels first identified in a mutant Drosophila melanogaster fruitfly. In these fruitflies, exposure to ether anesthesia elicits leg shaking (hence the phrase “ether-a-go-go”). HERG was initially isolated by screening a human hippocampal cDNA library with a mouse homolog of ether-a-go-go, a Drosophila K+ channel gene. In humans, the gene resides at chromosone 7 q35-36; the channel is also known as KCNH2 (also referred to as Kv11.1). HERG was subsequently shown to be strongly expressed in heart (Curran et al., 1995), and to encode a K+ channel with properties similar to that of the rapidly activating delayed rectifier K+ current IKr (Sanguinetti et al., 1995; Trudeau et al.,1995).
An additional link between hERG and cardiac repolarization was provided by the finding of hERG mutations associated with a congenital long QT 2 syndrome linked to chromosome 7 (Curran et al., 1995). It is generally appreciated that most drugs that either delay repolarization or are associated with TdP block hERG current (see reviews cited above). Thus, block of hERG current has become an accepted surrogate marker for cardiac proarrhythmia and the evaluation of drug block of hERG current has come to play a pivotal role in the preclinical evaluation of potential proarrhythmic risk. In addition, a few drugs that delay repolarization may act by reducing the number of hERG channels in the plasma membrane to reduce hERG current density (Thomas et al., 2003; Ficker et al., 2004; Kuryshev et al., 2005).
The hERG channel is a K+ selective voltage-gated ion channel that activates upon depolarization and deactivates upon repolarization. Due to its inactivation properties, outward hERG current peaks later during cardiac repolarization (at less depolarized potentials, see Figure 2). Due to its unusual kinetic characteristics, hERG current also plays a unique role by providing current timed to suppress the initiation of premature beats (Smith et al., 1996; Spector et al., 1996; Lu et al., 2001). hERG current is monitored most readily using voltage clamp technique in which the transmembrane potential is experimentally controlled, and resulting ionic current flow across the membrane (the biological response) recorded (Figure 3). HERG current density in cardiac myocytes is small relative to many other currents, and thus difficult to measure in the presence of these larger time overlapping currents. However, hERG current is much more easily measured when it is the primary current (as when the pore forming subunit is expressed in heterologous systems such as HEK or CHO cells; Zhou et al., 1998). At least some drugs are thought to block hERG by binding to specific tyrosine and/or phenylalanine residues within the pore structure of the channel after gaining access from the intracellular space (Mitcheson et al., 2000; Milnes et al., 2003a; Witchel et al., 2004). HERG remains the only potassium channel target linked to delayed repolarization and a propensity towards proarrhythmia.
Approaches to Evaluate Drug Effects on HERG
Indirect Techniques To Monitor hERG Current
Various approaches have been developed to evaluate drug effects on hERG current. These can be conveniently divided into two categories based on whether direct or indirect measures of hERG current are made. Indirect approaches include binding assays, assays measuring ionic flux changes, and those detecting changes in membrane potential. Binding assays that measure displacement of potent, radiolabeled hERG ligands by drugs provide a convenient screening method to detect drug-hERG channel interactions. Such systems may conveniently use intact cells or cell membranes from heterologous expression systems transfected with the hERG channel and potent, radiolabeled hERG ligands such as dofetilide (Diaz et al., 2004). Similar studies have utilized displacement of other hERG channel ligands, including the methanesulfonanilide MK-499 (Wang et al., 2003) and astemizole (Chiu et al., 2004).
Another indirect approach employed to detect hERG channel-drug interactions is the ion flux assay. This assay measures reduced levels of radiolabeled Rb+ efflux (reflecting diminished outward K+ current through hERG channels) from hERG-transfected cells (Rezazadeh et al., 2004). Voltage-sensitive dyes have also been used to measure drug-induced membrane depolarization of hERG transfected cells in which hERG/IKr is the primary potassium current responsible for setting the resting membrane potential (Netzer et al., 2003). Results obtained with these indirect assays are less sensitive than those reported using direct measures of hERG current. The primary advantage of these indirect approaches is greater throughput achieved compared with assays that directly measure changes in hERG current. New approaches using planar patch techniques coupled to automation provide direct measures of hERG current and are poised to challenge the use of indirect techniques (Dubin et al., 2005, Guo and Guthrie, 2005). However, the ability of these automated systems to match the throughput of indirect screening assays remains to be proven.
The characterization of changes in hERG current measured directly remains the “gold standard” for evaluating drug effects. Figure 3 illustrates hERG current recorded from a HEK cell stably transfected with the hERG channel; under these experimental conditions, hERG is the predominant current (recorded without interfering contaminating currents). As hERG is a voltage-gated channel, it is activated and deactivated by changes in the transmembrane potential (see Figure 3B). Using voltage clamp techniques, the voltage is controlled (“clamped”) by the investigator, and transmembrane current that flows in response to the changes in the channel conformation is monitored. When this approach is applied to cells transfected with the hERG channel, increasing outward current is recorded upon membrane depolarization (representing current activation). Upon repolarization, a decreasing outward hERG current (termed hERG tail current) is recorded, representing channel deactivation. The amplitude of the tail currents recorded in the absence and presence of drugs can be compared to determine the IC50 values for hERG current block (that is, the concentration of drug required to block 50% of hERG current, see Figure 3C and 3D). Concentration-response curves are constructed to determine IC50 values for hERG current block.
General Limitations of Evaluations of HERG Channel Block
IC50 values characterizing drug block of hERG current provide a convenient way to compare the potency of drugs. However, it should be recognized that IC50 values for hERG block represent a simplification of potentially complex time-, voltage-, and state-dependent block and unblock of hERG current. Some drugs (such as dofetilide [Snyders and Chaudhary, 1996; Weerapura et al., 2002] and mesoridazine [Su et al., 2004]) are referred to as “open channel blockers,” because channel activation is required to reduce hERG current. It has been proposed that compounds may block open channels by binding to different sites within the channel cavity; the lower-affinity hERG blocker propafenone behaves as an “open state blocker” but does not appear to interact with the aromatic residue Tyr652 as do some higher affinity hERG blocking methanesulfonanilides (Witchel et al., 2004). A few drugs may block hERG exclusively by interacting with closed channels (fluvoxamine (Milnes, 2003a); peptide toxins have been shown to act preferentially in this manner (Milnes et al., 2003b). Thus, the configuration (and frequency of repetition) of the voltage clamp waveform may affect the potency of block as well as the time course of block and recovery, reflecting interactions with different states of the channel (see Tsujimae et al., 2004 for a comparison of quinidine and dofetilide block of hERG current).
It should also be noted that the choice of experimental preparation and conditions also affects determinations of the potency of hERG block. Studies measuring hERG current in oocytes injected with hERG mRNA typically report IC50 values less potent than those obtained using either HEK or CHO cells transfected with hERG. This effect is likely due to the yolk sac in oocytes acting as a drug “reservoir” (Witchel et al., 2002a). Temperature has been shown to affect IC50 value determinations for some compounds; a recent study by Kirsch et al. (2004) demonstrated that increasing the temperature from 22 to 35° C elicited a statistically significant decreased blocking potency for 2 drugs, increased potency for 4 drugs, and had no effect on 9 additional drugs evaluated using a 2-second step pulse clamp protocol; for the same change in temperature, Yao and colleagues also reported an increased potency for hERG block with 1 drug, decreased potency for a second drug, and no changes in 2 additional drugs (Yao et al., 2005).
Finally, attention must be paid to measuring bath concentrations of tested compounds, as drug levels may be lower than targeted due to degradation and adsorption to perfusion tubing (unpublished observations see also data cited in Brown, 2004). Such differences in protocols and techniques contribute to the range of IC50 values for compounds reported in the growing hERG pharmacologic literature (see for example Figure 4, also Kirsch et al., 2004), and demonstrate the necessity of routinely including an accepted and widely used positive control standard to monitor the assay sensitivity.
IC50 values for hERG block (obtained in the absence of plasma proteins) are compared to plasma drug concentrations to gain insight into safety margins for proarrhythmic risk (see next). However, the free concentration of drugs extensively bound to plasma proteins may be significantly lower than total drug concentration. Controversy exists regarding how to correct for the potential effects of plasma protein binding. For a series of 5 antipsychotic drugs (with plasma protein binding ranging from 83–99%), the safety margin for hERG block (assessed as the ratio of hERG IC50 values to total plasma concentrations) corresponded well with QT prolongation clinically; free (unbound) drug plasma levels did not provide as reliable a correlation (Kongsamut, 2002). In contrast, the ratio of hERG IC50 values to free plasma concentrations was predictive of QT prolongation for a series of seven fluoroquinolone antibiotics (weak plasma protein binding ranging from 20–50%, Kang et al., 2001). Given such discrepancies, it is recommended that total and (calculated) free drug plasma concentrations both be considered when reviewing safety margins in relation to hERG block.
Evaluation of Proarrhythmic Risk Based on hERG Current: hERG block in Context
While a general relationship between hERG blockade and proarrhythmic risk for most noncardiovascular drugs (as well as some cardiovascular drugs) is generally accepted, the characteristics of that relationship remains uncertain. Trends have been noted relating hERG block with the clinical QT prolongation and TdP. Redfern and colleagues (2003) summarized published in vitro electrophysiologic data from 52 compounds, comparing potency of hERG (or IKr) current block with clinical experience related to QT interval prolongation and reports of TdP in humans; these data were set against free plasma concentrations during clinical use (effective therapeutic plasma concentrations [ETPC unbound]).
For these comparisons, drugs were categorized into 5 groups based upon clinical experience, with higher numbers associated with reduced proarrhythmic risk: (1) Antiarrhythmic drugs expected to prolong cardiac repolarization); (2) Drugs withdrawn from the market due to TdP; (3) Drugs with measurable incidence/numerous TdP reports, (4) Drugs with isolated reports of TdP in humans; (5) No reports of TdP in humans. Drugs within each category were then characterized based on the ratio of their lowest reported hERG IC50 values vs. calculated unbound Cmax values, defining a “safety margin.” The dataset demonstrated that most drugs associated with TdP have IC50 values close to or superimposed upon free drug plasma concentrations found in clinical use. While a 30-fold margin between Cmax and hERG IC50 values was generally predictive of cardiac safety, the authors noted that increasing this margin should be considered to ensure cardiac safety, especially for future drugs aimed at nondebilitating diseases.
A closer evaluation of the dataset analyzed by Redfern and colleagues reveals that some drugs not associated with TdP in humans (categories 4 and 5) are characterized with hERG IC50 values comparable to calculated free plasma concentrations (e.g., verapamil). We use the term “overt hERG blocker” to classify drugs that either delay repolarization in integrated systems (APD repolarization or QT assays) or are linked to TdP at unbound (free) concentrations comparable to those that block hERG. These drugs elicit hERG block and QT prolongation over comparable concentration ranges. In contrast, we use the term “covert hERG blocker” to describe drugs that do not delay repolarization in integrated systems or elicit TdP but block hERG current at comparable unbound (free) concentrations. A comparison of select drugs representing overt and covert hERG blocking agents is instructive in defining the utility and limitations of the hERG functional current assay as a surrogate marker of proarrhythmic risk.
Dofetilide: An Example of an “Overt” hERG Blocking Drug
Dofetilide is a highly potent and specific blocker of IKr (hERG) current used to treat atrial fibrillation by blocking IKr in human atria. However, dofetilide also blocks IKr present in the ventricle. IC50 values for current block in the literature are in the low nanomolar range, and not far from plasma concentrations (see Figure 4). Dofetilide elicits concentration-dependent prolongation of the QTc interval at total plasma concentrations as low as 3.4 nM (Tikosyn product monograph) and is an example of an overt hERG blocking drug.
Cisapride and Terfenadine: Examples of Overt hERG Blocking Drugs Resulting from “Metabolic Multiplication”
Cisapride
Cisapride is a gastrointestinal prokinetic agent formerly used widely for the treatment of gastroesophageal reflux disease (GERD). First marketed in 1981, and ranked 73rd most dispensed drug in the United States in 1996 with over 5.4 million prescriptions filled, this drug was withdrawn from the market in 1997. Cardiotoxicity with cisapride is associated with high doses, or with concomitant use of drugs that inhibit the cytochrome P-450 3A4 enzyme system responsible for cisapride metabolism, leading to excessive plasma concentrations of the drug (Ahmad and Wolf, 1995, Pettignano et al., 1996). Increasing (and excessive) plasma concentrations of drugs arising after metabolic inhibition has been referred to as “metabolic multiplication.”
Figure 4 illustrates the therapeutic total plasma concentrations of cisapride, along with calculated free concentrations and reported IC50 values for hERG block. Cisapride can be considered an overt hERG blocking drug as supratherapeutic plasma levels that approach or exceed concentrations eliciting hERG block are associated with delayed repolarization and TdP.
Terfenadine
Terfenadine was one of the first non-cardiovascular drugs linked to TdP clinically and shown to prolong the QT interval. Initially marketed in 1981, terfenadine ranked 9th in total prescriptions dispensed in 1991, eventually being withdrawn from the market in 1997. Lessons learned with this drug highlight the potential risks resulting from metabolic multiplication. In most instances, terfenadine is rapidly and almost completely metabolized following oral administration by cytochrome P450 (CYP) 3A4 oxidases to an active compound essentially devoid of hERG blocking effects (fexofenadine).
However, in cases of inhibition of the CYP 3A4 oxidases (for example, with concomitant administration of ketoconazole), excessive concentrations of the parent compound terfenadine can result (see Figure 4), leading to concentration-dependent prolongation of the QT interval (Honig et al., 1993). Terfenadine was shown to block IKr in feline ventricular myocytes (Woosley, 1993); a link to the hERG channel as the molecular target was later demonstrated (Roy et al., 1996). Similar exacerbations of untoward drug effects have been reported with other hERG blocking drugs such as astemizole and cisapride (see review by Dresser, 2000). Terfenadine is an example of an overt hERG blocking drug in that supratherapeutic plasma concentrations attained after metabolic multiplication are comparable to IC50 values for hERG block and may elicit delayed repolarization and QT prolongation. These findings highlight the importance of relating hERG block to targeted therapeutic concentration and excessive plasma concentrations that may arise from adverse pharmacokinetic interactions.
Verapamil and Fluoxetine: Covert hERG Block Arising from Multichannel Block
The term “multi-channel block” refers to the reduction of other (non-HERG) cardiac currents by hERG-blocking drugs. Given the multiple ion channels, pumps, and exchanges that contribute to the cardiac action potential, it is likely that some drugs reduce additional cardiac currents other than hERG. As a result, these drugs may produce electrophysiologic effects distinctly different from drugs solely affecting hERG block, providing an ionic mechanism to explain covert hERG blockade. In vitro, delayed repolarization elicited by hERG block can be attenuated by reduction of either the L-type calcium current or late cardiac sodium current (see Martin et al., 2004); the former case is consistent with the electrophysiological effects of verapamil and fluoxetine, two covert hERG blockers.
Verapamil
Verapamil is an L-type calcium channel blocker with vasodilatory and antiarrhythmic actions first marketed in 1981 and used to treat angina pectoris, arrhythmias, and hypertension. In in vitro action potential studies, verapamil reduces the height of the action potential plateau and accelerates repolarization during the action potential plateau, both consistent with block of calcium current (Rosen et al., 1974). The use of this calcium channel blocking drug is not associated with either QT prolongation or risk of TdP. However, verapamil has been demonstrated to be a potent hERG blocker, with IC50 values for hERG block ranging from 0.094 to 3 μM; these values overlap therapeutic exposures (0.1–0.8 μM [total concentrations] and 0.01–0.08 μM [calculated free plasma concentrations], see Figure 4). A reported IC50 value for block of L-type calcium current block (0.164 μM) is also near values cited for hERG block.
The lack of association between hERG block and either QT prolongation or TdP with verapamil is consistent with the concomitant reduction of cardiac L-type calcium current mitigating the effects of hERG block. In general, reduction of calcium current (an inward depolarizing current during the action potential plateau) leads to a reduction in the height of the plateau and minimal effects (or abbreviation) of the action potential duration. Thus, a drug that reduces inward (depolarizing) calcium current could “balance” the diminished outward repolarizing current (resulting from hERG block) to minimize repolarization delays (and proarrhythmia). In in vitro studies with Purkinje fibers, delayed repolarization elicited by the specific IKr blocking drug dofetilide was significantly alleviated by the concomitant administration of the selective calcium channel blocking agent nifedipine (see Figure 5). Indeed, combined potassium and calcium channel blocking properties have been proposed as the basis for the lower proarrhythmic risk of the experimental antiarrhythmic compound BRL-32872 (Bril et al., 1996), especially at high concentrations shown to elicit less APD prolongation compared to lower concentrations. Thus, calcium channel block can attenuate the APD prolonging effects of hERG block.
Block of calcium current may also reduce proarrhythmic risk of hERG blockade by preventing the initiation of early afterdepolarizations (EAD’s) during prolonged action potentials. EAD’s are aberrant depolarizing “spikes” that arise from the plateau phase of prolonged action potentials (Fozzard, 1992); this triggered activity can propagate into surrounding tissue and may be responsible for initiating TdP (Yan et al., 2001). Drugs that reduce L-type calcium channel block may prevent TdP by preventing the EAD’s likely responsible for initiating this arrhythmia (January 1991).
Fluoxetine
Fluoxetine is an antidepressant drug first marketed in 1986 with extensive clinical experience (worldwide exposure estimated to be more than 38 million patients circa, 1999) that is not usually associated with either QT prolongation or TdP. Fluoxetine has also been reported to block hERG current (IC50 values ranging from 0.46 to 1.5 μM). Tese concentrations approximate the total therapeutic plasma concentrations (0.3–2.6 uM, Figure 4E). In addition, fluoxetine has also been demonstrated to block L-type calcium current at slightly higher concentrations (IC50 values of 2.8 and 5.4 uM, see Figure 4). Fluoxetine’s lack of association with either delayed repolarization or TdP is consistent with multichannel block; indeed, supratherapeutic fluoxetine concentrations (approximately 60 μM, well above the IC50 value for hERG block) actually shorten the canine cardiac Purkinje fiber action potential duration in vitro (Gintant et al., 2001). In the present regulatory environment, it has been argued that a drug with a hERG profile similar to the widely used antidepressant fluoxetine might either not be developed today and would be subject to special (and likely unnecessary) use restrictions (Fermini and Fossa, 2003).
HERG Block: One of Multiple Risk Factors for Proarrhythmia
Finally, it is generally accepted that TdP arises as a result of the unfortunate coincidence of multiple risk factors manifest at one time (Roden, 1998). A retrospective review of reports of drug-induced TdP from 249 patients with noncardiac drugs found at least one easily identified risk factor, with 71% of patients having 2 or more risk factors (Zeltser et al., 2003). The most prevalent risk factors were female gender, heart disease, hypokalemia, pharmacodynamic drug interactions (use of multiple drugs that affect repolarization), and pharmacokinetic interactions (potential interference with the metabolism of a QT-prolonging drug by a second medication). The presence of congenital mutations of cardiac ion channels may also play a role in increasing the risk of drug-induced TdP (Roden, 2004). All of these factors may reduce the ability of the ventricle to repolarize by reducing net outward current (referred to as “repolarization reserve,” Roden, 1998), rendering the heart more susceptible to delayed repolarization and potential proarrhythmia by hERG blocking drugs. The above observations dictate that the relative potency of hERG block represents only one of multiple factors that need to be considered when evaluating proarrhythmic risk. Thus, an evaluation of potential proarrhythmic effects of drugs based on hERG current block alone provides an incomplete view of a drug’s potential to affect the complex and integrated process of cardiac repolarization in vitro and in vivo.
Conclusion: hERG and Beyond
Torsades de Pointes is a rare arrhythmia that usually results from a simultaneous convergence of risk factors that lead to significant impaired or delayed cardiac repolarization. Drugs that block hERG, a critical cardiac repolarizing current involved in initiating terminal cardiac repolarization, act to prolong ventricular repolarization and have been linked with TdP. HERG is the only molecular potassium channel target linked to delayed repolarization and a propensity towards proarrhythmia. Thus, block of hERG current has become an accepted surrogate marker for TdP.
The potential proarrhythmic risk associated with hERG current block must first be considered in the context of potency of hERG block in relation to therapeutic plasma concentrations. Other factors that must be considered when assessing a drug’s potential proarrhythmia include: pharmacokinetic interactions (the potential for accentuated block due to excessive drug concentrations resulting from drug-drug interactions [e.g. metabolic multiplication], pharmacodynamic interactions (multiple drugs simultaneously reducing hERG current), and a changing electrophysiologic substrate (e.g., cardiac disease, hypokalemia). In the case of overt hERG blockers, drug effects are manifest predictably as concentration-dependent QT prolongation (and presumably greater risk for TdP) as drug concentrations approximate IC50 values for hERG block. In contrast, covert hERG blockers may not delay repolarization as a result of concomitant block of other (non-hERG) cardiac currents (e.g., L-type calcium current, late sodium current [multichannel block]) mitigating hERG block when therapeutic concentrations approximate IC50 values.
Finally, additional systemic (non-cardiac) “off-target” drug effects may further modulate hERG blockade (for example, drugs modulating autonomic influences to indirectly affect cardiac repolarization; see Fossa et al., 2005). Thus, hERG current block represents only one important part of a multifactorial integrated risk assessment that must include subcellular (hERG current assays), cellular (in vitro APD-repolarization assays) and in vivo (ECG-QT) components providing a balanced evaluation of proarrhythmic risk. Newer in vitro models are needed to quantitatively characterize the relationship between hERG block and proarrhythmic potential, especially for covert hERG blocking compounds.