ORIGINAL ARTICLE
The Utility of hERG and Repolarization Assays in Evaluating
Delayed Cardiac Repolarization: Influence of
Multi-Channel Block
Ruth L. Martin, PhD, Jeff S. McDermott, MS, Heinz J. Salmen, BS, Jason Palmatier, BA,
Bryan F. Cox, PhD, and Gary A. Gintant, PhD
Abstract: Drug-induced delayed cardiac repolarization is a recog-
nized risk factor for proarrhythmia and is associated with block of IKr
(the potassium current encoded by the human ether-a- go-go-related
gene [hERG]). To evaluate the utility of 2 in vitro assays widely used
to assess delayed repolarization, we compared the effects of haloper-
idol and 9 structurally diverse drugs in a hERG and repolarization
(canine Purkinje fiber action potential duration [APD]) assay over
wide concentrations. Despite potent hERG current block (IC50 =
0.174 µM), haloperidol elicited a bell-shaped concentration–response
relationship for APD prolongation, with lesser prolongation (and re-
duced plateau height) observed with concentrations eliciting maximal
hERG block, consistent with multi-channel block at higher concen-
trations. Consistent with this hypothesis, APD prolongation with the
specific IKr blocker dofetilide was a) reduced by concomitant admin-
istration of nifedipine (calcium current block) and b) reversed by li-
docaine (late sodium current block). Additional studies demonstrated
prominent (>50%) hERG inhibition with most (9/10) drugs despite
wide APD changes (158% prolongation − 16% shortening), consis-
tent with multi-channel block. The poor correlation between hERG
and repolarization assays suggests that the hERG assay oversimpli-
fies drug effects on the complex repolarization process for drugs dem-
onstrating multi-channel block and that neither assay alone ad-
equately predicts proarrhythmic risk.
Key Words: hERG, torsade-de-pointes, action potential duration,
Purkinje fibers, cardiac repolarization, haloperidol, arrhythmias
(J Cardiovasc Pharmacol™ 2004;43:369–379)
R ecent warnings, relabeling, and withdrawals of some drugs
(including antihistamines, antifungals, and antipsychotic
agents) from the market for cardiovascular safety concerns
Received for publication March 6, 2003; accepted October 30, 2003.
From the Department of Integrative Pharmacology, Global Pharmaceutical
Research and Development, Abbott Laboratories, Abbott Park, IL.
Reprints: Gary Gintant, PhD, Department of Integrative Pharmacology
(R46R, Bldg AP-9), Abbott Laboratories, 100 Abbott Park Road, Abbott
Park, IL 60064-6119. E-mail: gary.gintant@abbott.com
Copyright © 2004 by Lippincott Williams & Wilkins
J Cardiovasc Pharmacol姠 • Volume 43, Number 3, March 2004
highlight the growing awareness of the potential risk for pro-
arrhythmia by noncardiovascular drugs. At issue is a rare,
drug-induced ventricular arrhythmia known as torsade-de-
pointes, which may lead to syncope or sudden cardiac death.1
The low incidence and potentially lethal consequences of this
drug-induced arrhythmia pose a challenge to physicians, the
pharmaceutical industry, and regulators alike. Dose-dependent
prolongation of the QT interval has been used as a surrogate
marker for proarrhythmia.2 However, the extent of QT prolon-
gation may be small (mean values sometimes <10 milliseconds
above baseline values near 400 milliseconds) and difficult to
detect above rate-dependent and diurnal variations in the QT
interval. As a consequence, preclinical assays to assess de-
layed repolarization have been employed to assess the poten-
tial risk for drug-induced proarrhythmia.
At present, two functional in vitro assays are routinely
employed when evaluating a drug’s potential to delay cardiac
repolarization.1,3–5 One assay (termed a repolarization assay)
characterizes drug-induced changes in the action potential du-
ration (APD) of cardiac tissues (such as syncytial Purkinje fi-
bers, papillary muscles, or isolated cardiac myocytes). An-
other approach evaluates drug-induced block of the rapidly
activating delayed rectifier current (typically either hERG cur-
rent expressed in heterologous systems or native IKr). The util-
ity of this ionic current assay is based on the fact that most
drugs that delay repolarization are associated with inhibition of
the delayed rectifier current IKr (see ref. 4); this current plays
a prominent role in defining terminal cardiac repolarization
and is encoded (at least in part) by the hERG gene.6,7 Using
either the repolarization or hERG ionic current assay, a con-
centration-dependent “signal” is generally considered as evi-
dence of proarrhythmic risk. However, the action potential re-
flects the effects of multiple ion channels, pumps, and
exchangers. Thus, multiple drug effects could mask or modu-
late the potentially detrimental effects of hERG current inhi-
bition, especially at higher drug concentrations.
Haloperidol is a butyrophenone antipsychotic linked
clinically to QT prolongation and torsade-de-pointes.8 Halo-
peridol has also been shown to block the cardiac delayed rec-
369
Martin et al
tifier hERG current in Xenopus oocytes,9 sodium channels,10
and is associated with block of calcium current.11 As part of a
routine preclinical evaluation of cardiac electrophysiologic ef-
fects of drug candidates, we studied the actions of haloperidol
(at and above therapeutic concentrations) in the hERG ionic
current and canine Purkinje fiber repolarization (APD) assay.
Despite potent concentration-dependent hERG block, haloper-
idol demonstrated a bell-shaped concentration–response curve
for APD prolongation over the same concentration range, con-
sistent with inhibition of multiple (non-hERG) cardiac cur-
rents. To explore this effect more fully, we compared haloperi-
dol’s effect on repolarization to that of the specific IKr blocking
agent dofetilide alone and in combination with either the cal-
cium channel blocking drug nifedipine (to mimic combined
hERG and calcium current) or the late sodium current blocking
drug lidocaine. A subsequent comparison of effects of 9 addi-
tional drugs in the hERG and APD assays revealed a mono-
tonic relationship between action potential prolongation and
hERG current inhibition for only 4 of 10 drugs, consistent with
multi-channel block affecting cardiac repolarization with in-
creasing drug concentrations. Two of 10 drugs eliciting promi-
nent hERG block (fluoxetine, verapamil) are not linked to ei-
ther QT prolongation or proarrhythmia clinically. Together,
these studies suggest that neither an APD repolarization assay
nor a hERG assay alone adequately predict delayed repolariza-
tion and proarrhythmic risk (especially for drugs affecting
multi-channel block) and that the hERG assay oversimplifies
drug effects on cardiac repolarization.
METHODS
hERG Ionic Current Studies
hERG channels stably expressed in human embryonic
kidney cells (HEK-293)12 were maintained in Minimal Essen-
tial Medium supplemented with 10% fetal bovine serum, 1%
penicillin-streptomycin, 2 mM L-glutamine, 0.1 mM nones-
sential amino acids, 1 mM sodium pyruvate, and 200 mg/mL
of G418 in a humidified atmosphere (95% air–5% CO2) at
37°C. Media were changed every 48 hours and cells were pas-
saged weekly. Cells were not allowed to become more than
80% confluent. For electrophysiology recordings, cells were
briefly trypsinized to release the cells from the plates, pelleted
by centrifugation (1000g), resuspended in culture media and
studied within 8 hours.
The bath solution contained (in mM): NaCl 140, KCl 5,
MgCl2 1, CaCl2 2, glucose 5, and HEPES 20, pH = 7.4. Patch
pipettes were constructed with borosilicate glass capillary
tubes (resistance 1.8–3.8 M⍀). The pipette solution contained
(in mM): K+ aspartate 125, KCl 20, EGTA 10, MgCl2 1,
HEPES 5, and MgATP 5 (pH = 7.3). Experiments were per-
formed at 36.5–37°C. Drugs were either diluted directly in
bath solution or were dissolved in dimethyl sulfoxide (DMSO)
before dilution in bath solution. Lower drug concentrations
370
J Cardiovasc Pharmacol姠 • Volume 43, Number 3, March 2004
were prepared by serial dilution; DMSO concentrations never
exceeded 0.1% (by volume). Cells studied were exposed to a
single concentration of drug, with current block measured
from drug-free control values.
Currents were recorded using either an Axopatch 200A
or Axopatch Multiclamp 700A along with pClamp data acqui-
sition software (version 8, Axon Instruments Inc., Union City,
CA, U.S.A.). Drug effects were assessed using a voltage clamp
protocol that stepped to −25, 0, 25, or 50 mV for 3 seconds,
followed by a step to −50 mV for 4 seconds from a holding
potential of −80 mV; clamp pulses were applied once every 15
seconds. IC50 values were calculated from tail currents mea-
sured at −50 mV following conditioning pulses to 0 mV (cho-
sen to mimic plateau potentials of canine Purkinje fibers). IC50
values were derived after adjusting for current run-down
evaluated from time-matched controls (E-4031, fluoxetine),
DMSO-vehicle (cisapride, haloperidol, indomethacin, moxi-
floxacin, terfenadine, telithromycin, verapamil), or lactobion-
ate vehicle (erythromycin). Current run-down in the presence
of DMSO (measured following approximately 6–8 minutes
exposure, analogous to drug-containing experiments) was ap-
proximately 9% and independent of the DMSO concentrations
tested [8.29 ± 5.94% (n = 18), 9.07 ± 4.60% (n = 18) and
8.36 ± 3.11% (n = 26) for 0.001%, 0.01% and 0.1% DMSO (by
volume, respectively)]. Current run-down for time (vehicle-
free) alone (4.02 ± 2.05% [n = 11]) was less but not statistically
different from DMSO control groups (ANOVA).
Action Potential Repolarization Studies
Protocols used to evaluate drug effects on the Purkinje
fiber action potential duration have been described previ-
ously.13 Briefly, free-running canine (approximately 1–3 years
of age, either gender, Marshall Farms) cardiac Purkinje fibers
were excised, placed in a warmed chamber, and superfused
(8–10 mL/min) with a Tyrode’s solution containing (in mM):
NaCl, 131; NaHCO3, 18; NaH2PO4, 1.8; MgCl2, 0.5; dextrose,
5.5; KCl, 4; CaCl2, 2 (aerated with 95% O2/5% CO2 [pH = 7.2
at room temperature]). Fibers were stimulated (2x threshold,
biphasic waveform, typically 1–2 milliseconds in duration) us-
ing platinum electrodes located in the chamber floor and im-
paled with 3M KCl-filled microelectrodes (resistance 10–30
M⍀); electrical activity was monitored using high-input im-
pedance electrometers (IE-210, Warner Instruments), re-
corded digitally (Digidata 1200, Axon Instruments), and ana-
lyzed using pClamp software. Studies were initiated after a
minimum 30-minute equilibration period with stimulation. Fi-
bers were considered suitable for study if (during stimulation
at 2 seconds basic cycle length) the following criteria are sat-
isfied: a) the membrane potential just prior to action potential
upstroke was more negative than −80 mV, b) the APD ranged
between 300 and 500 milliseconds (spanning approximately
1.2 standard deviations from the mean value of 405 millisec-
© 2004 Lippincott Williams & Wilkins
J Cardiovasc Pharmacol姠 • Volume 43, Number 3, March 2004 hERG Current, Repolarization Assays, and Multi-channel Block

onds), and c) the normal automatic rate did not exceed the
2-second stimulation cycle length.
Electrophysiological effects on repolarization were
evaluated during slow stimulation (2 seconds BCL [30 beats
per minute]) chosen because repolarization changes are typi-
cally exaggerated during slow stimulation (reverse use-
dependence14) and are also more likely to reflect changes dur-
ing bradycardia (a recognized risk factor for proarrhythmia
with drugs that delay repolarization). During experiments, fi-
bers were sequentially exposed to 3 ascending drug concentra-
tions (typically, 1-, 10-, and 100-fold clinically encountered
total plasma concentrations). The protocol consisted of pacing
in a control (drug-free) solution for 20–25 minutes at a basic
cycle length of 2 seconds, then consecutively paced at 800 mil-
liseconds BCL (75 bpm) during the transition to a 400 milli-
seconds BCL for 2–3 minutes at each stimulation rate. Action
potentials recorded at the end of each 25-minute equilibration
period were used to define drug effects and generate cumula-
tive concentration-response curves. The present in vitro stud-
ies were conducted in the absence of plasma proteins. For
those drugs demonstrating high plasma protein binding (for
example, haloperidol, fluoxetine, indomethacin, terfenadine,
and verapamil), it is likely that concentrations tested in vitro
are greater than free concentrations achieved in vivo. Differ-
ences in the free drug concentration in vitro versus in vivo will
depend upon the characteristics of plasma protein binding,
which can be highly nonlinear (and hence unpredictable) at
higher drug concentrations. For these drugs, multi-channel
blocking effects may occur at multiples of total plasma con-
centrations greater than expected from the present study.
For experiments with single drug exposure, the action
potential duration was measured as the time from the maxi-
mum upstroke velocity to a potential 10 mV more positive than
full repolarization from the average of 3 consecutive action
potentials. For later experiments with 2-drug exposure, the ac-
tion potential was measured to 90% of repolarization (APD90).
Plateau height was measured as the voltage 100 milliseconds
following the action potential upstroke (after transition of the
action potential notch but before the initiation of terminal re-
polarization). No drugs elicited significant depolarization of
the maximum diastolic potential.
Drugs evaluated previously for effects on Purkinje fiber
repolarization included cisapride, erythromycin, fluoxetine,
indomethacin, moxifloxacin, and terfenadine13; newly evalu-
ated compounds included the antipsychotic drug haloperidol,
the antibiotic telithromycin, and anti-anginal agent verapamil,
and the prototypic antiarrhythmic IKr channel blocking agent
E-403115 (Table 1). With the exception of indomethacin,
fluoxetine, and verapamil, all drugs have been linked clinically
to either QT prolongation and/or torsade-de-pointes) at or
above therapeutic concentrations on the basis of product label-
ing16 or critical reviews.17,18
To experimentally “simulate” the effects of multichan-
nel block on cardiac repolarization, we evaluated the actions of
the specific IKr channel blocking drug dofetilide19 in the ab-
sence and presence of either the L-type calcium current block-
ing agent nifedipine or the late sodium current blocking agent
lidocaine.20,21 A 100-nM concentration of dofetilide was cho-
sen to block IKr based upon IC50 values below 10 nM for block
of either native IKr or hERG current22,23; for nifedipine, a con-

TABLE 1. Concentrations Tested in hERG and Purkinje Fiber Repolarization Assays, Along with IC50 Values for hERG Tail
Current Block

Plasma Protein Ther. Conc.* Conc. Tested IC50 Hill

Drug Binding (%) (µM) (hERG and Purkinje, µM) (hERG, µM) Coefficient
Cisapride 98 0.17 0.001, 0.003, 0.01, 0.1; 0.1, 1, 10 0.0182 1.05
E-4031 0.02, 0.1, 0.2 0.048 1.09
Erythromycin 75–90 2.72 3.6, 36, 183, 132 158 0.77
Fluoxetine 94 2.26 0.58, 5.77, 57.8 0.99 1.06
Haloperidol 90 0.026 0.026, 0.26, 2.66 0.174 0.83
Indomethacin 99 5.59 5.59, 55.9, 279.5 >300 —
Moxifloxacin 50 10.7 9.1, 91.3, 456 58.5 0.92
Telithromycin 60–70 2.3 0.6, 6.15, 61.5; 6.15, 61.5 56.21 0.70
Terfenadine 97 0.009 0.01, 0.1, 0.5; 0.1, 1, 10 0.016 0.83
Verapamil 88–94 0.88 0.1, 1.0, 10 0.410 0.74
IC50 values determined from a minimum of three concentrations (n = 6 cells per datapoint) assuming 0 and 100% block at extreme concentration ranges. When
2 concentration ranges are listed, the first set indicates concentrations used to evaluate hERG block, while the second set (in italics) indicates concentrations
employed in the repolarization (APD) assay. Purkinje fiber repolarization study results from this laboratory have been reported previously for erythromycin,
fluoxetine, indomethacin, moxifloxacin, and terfenadine.13
*Therapeutic concentration represents total concentrations encountered clinically; concentrations tested did not consider possible effects of plasma protein
binding in vivo, which may reduce effective plasma concentrations for some highly bound drugs.

© 2004 Lippincott Williams & Wilkins 371

Martin et al
centration of 5 µM was chosen to reduce L-type calcium cur-
rent. Prior studies have demonstrated that nifedipine does not
affect hERG current expressed in HEK-293 cells,24 and the
structurally related compound nisoldipine does not block na-
tive cardiac K+ currents.25 Lidocaine (5 and 25 µM) was used
to block late sodium current; prior studies have shown that
these concentrations preferentially affect late sodium current
and modify repolarization at concentrations lower than those
required to inhibit the fast inward sodium current and action
potential upstroke.20
Statistical Analysis
Values are reported as mean ± standard error of the mean
(s.e.m.). Differences in APD parameters were evaluated using
repeated measures ANOVA followed by Dunnett’s post-hoc
test. IC50 values for hERG current block were obtained from
fits to the logistic equation of the form y = [(A1 − A2)/{1 +
(X/Xo)P }] + A2, where P represents power, Xo represents the
IC50 value, and A1 and A2 represent the initial (0) and final
(100%) values of block using Origin (version 6.0).
RESULTS
Effects of Haloperidol on hERG Current and
Purkinje Fiber Repolarization
Figure 1 characterizes the typical effects of haloperidol
on hERG current stably transfected in HEK-293 cells. Figure
1A illustrates the effect of 0.26 µM haloperidol on hERG cur-
rent during a 3-second depolarizing test pulse to 0 mV fol-
lowed by a 4-second repolarization step to −50 mV to elicit tail
currents. Haloperidol elicited time-dependent block of out-
ward current upon depolarization (consistent with open chan-
nel block), reduced tail current amplitude, and slowed tail cur-
rent kinetics. Block of hERG current by haloperidol was also
voltage-dependent. Figure 1B summarizes tail current block
with low and intermediate haloperidol concentrations (0.026
and 0.26 µM) following 3-second conditioning test pulses to
−25, 0.25, and 50 mV. Greater block is evident with stronger
depolarizing test pulses with both haloperidol concentrations.
Figure 1C summarizes concentration-dependent block of
hERG current evaluated from tail currents following condi-
tioning test pulses to 0 mV. Block of hERG tail current by
haloperidol was fit with an IC50 value 0.174 µM.
To compare hERG current block by haloperidol with its
effect on ventricular repolarization, we evaluated haloperi-
dol’s effects on canine Purkinje fiber repolarization. Changes
in the action potential configuration were evaluated during
slow stimulation (2 seconds BCL) to mimic the slow voltage
clamp protocol and to maximize repolarization delays; rates
slower than 2 seconds BCL were not possible due to normal
automaticity eliciting substantial phase 4 depolarization and
occasional premature beats. Figure 2 demonstrates the typical
372
J Cardiovasc Pharmacol姠 • Volume 43, Number 3, March 2004
effects of haloperidol on the canine Purkinje fiber action po-
tential. Figure 2A illustrates superimposed action potentials
recorded from a fiber exposed sequentially to 0.026, 0.26, and
2.66 µM haloperidol (concentrations matching those evaluated
in the hERG ionic current assay). Concentration-dependent
prolongation (extension of the plateau with minimal altera-
tion in the action potential configuration) was observed at the
lower and intermediate haloperidol concentrations (0.026 and
0.26 µM, respectively). These effects are qualitatively compa-
rable to those observed with specific IKr blocking drugs
E-4031 and dofetilide (unpublished observations). At the high-
est concentration (2.66 µM), haloperidol reduced the height at
the start of plateau and affected moderate “triangulation” of the
action potential. This resulted in action potential shortening
with the highest (compared with intermediate) haloperidol
concentration. Figure 2B summarizes the effects of haloperi-
dol on the action potential duration and plateau height from 6
fibers. Action potential prolongation was maximal at the inter-
mediate concentration (0.26 µM), and reduced at the highest
concentration (Fig. 2B, upper graph). In contrast, the plateau
height (measured 100 milliseconds after the upstroke) was sig-
nificantly depressed only at the highest haloperidol concentra-
tion (Fig. 2B, lower graph). Changes in the action potential
plateau configuration observed with the highest haloperidol
concentration suggest effects on additional ionic currents
(other than hERG) during the action potential plateau.
Multi-channel Block: Combined Effects of IKr
and Calcium Current Block
Changes in the action potential configuration with high
haloperidol concentration resemble those observed with L-
type calcium channel blocking agents verapamil and nifed-
ipine; namely, a reduction in plateau height and triangulation
of the action potential (unpublished observations). To investi-
gate the effects of simultaneous IKr and L-type calcium current
inhibition on cardiac repolarization, we evaluated changes in
Purkinje fiber repolarization during superfusion with the se-
lective IKr blocker dofetilide alone (0.1 µM) followed by su-
perfusion with dofetilide plus nifedipine (5 µM) to block cal-
cium current. Figure 3A illustrates representative changes in
the action potential configuration with dofetilide alone, and
during superfusion with dofetilide and nifedipine during slow
stimulation (2 seconds BCL). Superfusion with dofetilide
alone delayed final repolarization of the action potential to af-
fect substantial prolongation without affecting the plateau
height. Concomitant superfusion with nifedipine reduced the
plateau height, triangularized the action potential, and dramati-
cally reduced prolongation elicited by dofetilide. Figure 3B
summarizes the effects of dofetilide alone and in combination
with nifedipine on the action potential duration and plateau
height. While dofetilide alone prolonged APD by 77% (365 ±
25 [control] versus 646 ± 24 milliseconds [dofetilide]), the ad-
dition of nifedipine (in the continued presence of dofetilide)
© 2004 Lippincott Williams & Wilkins
J Cardiovasc Pharmacol姠 • Volume 43, Number 3, March 2004
mitigated these effects (22% prolongation versus control,
445 ± 27 milliseconds). Dofetilide alone elicited no change in
the plateau height (−3.6 ± 1.4 versus −2.6 ± 1.3 mV [control
versus dofetilide, respectively]). However, the addition of ni-
fedipine (in the continued presence of dofetilide) significantly
depressed the plateau height (−11.4 ± 2.6 mV compared with
© 2004 Lippincott Williams & Wilkins
hERG Current, Repolarization Assays, and Multi-channel Block
control value of −3.6 ± 1.4 mV). The maximum diastolic po-
tential and maximum upstroke velocity were unaffected in
these experiments. These results demonstrate that L-type cal-
cium current block can significantly attenuate action potential
prolongation elicited by IKr block.
The local anesthetic lidocaine acts to shorten the action
potential duration and lower the plateau height of Purkinje fi-
ber before affecting the action potential upstroke.20 (unpub-
lished observations). These effects have been attributed to
block of late sodium current prevalent in Purkinje fibers21,26
and ventricular midmyocardial myocytes.20,27 To evaluate the
effect of late sodium current inhibition on delayed repolariza-
tion elicited by IKr blockade, we superfused fibers with dofeti-
lide (0.1 µM) followed by dofetilide plus escalating lidocaine
concentrations (5 and 25 µM) during slow stimulation (2 sec-
onds BCL). Figure 4A illustrates representative changes in the
action potential configuration with dofetilide alone, and during
superfusion with dofetilide and the higher lidocaine concen-
tration (25 µM). As expected (and as shown above), dofetilide
alone delayed final repolarization without affecting the plateau
height. Concomitant superfusion with 25 µM lidocaine re-
duced the plateau height, triangularized the action potential,
and shortened the action potential duration compared with
control values. Figure 4B summarizes the effects of dofetilide
alone and in combination (with escalating lidocaine concen-
trations) on the action potential duration and plateau height.
The low lidocaine concentration reduced action potential pro-
longation elicited by dofetilide alone (100% prolongation by
dofetilide alone versus 47% prolongation in the presence of 5
mM lidocaine). The higher lidocaine concentration (25 µM)
abolished APD prolongation by dofetilide, affecting a modest
(9.5%) shortening that was not statistically different from con-
trol APD values. Neither dofetilide alone nor dofetilide plus
the lower lidocaine concentration affected a change in the pla-
teau height (−5.52 ± 2.38 mV versus −5.88 ± 2.09 mV versus
−8.04 ± 2.61 mV [control versus dofetilide versus dofetilide +
5 µM lidocaine, respectively]). However, the higher lidocaine
concentration significantly reduced the plateau height (−18.06
FIGURE 1. Haloperidol block of hERG current. (A) Typical
changes in hERG current with 0.26 µM haloperidol. The clamp
protocol consisted of a 3-second conditioning depolarizing
step to 0 mV, followed by a 4-second repolarizing step to ⳮ50
mV to record tail currents; ⳮ80 mV holding potential, with
pulses applied once every 15 seconds. Arrows indicate peak tail
current amplitude. (B) Voltage-dependent block of hERG tail
current was evident at low and intermediate concentrations of
haloperidol (0.026 and 0.26 µM); the highest concentration
tested (2.66 µM) blocked essentially all hERG tail current.
Mean (Ⳳ SEM) values for 6 cells represented for each concen-
tration and potential. (C) Haloperidol block of hERG tail cur-
rent was characterized with an IC50 of 0.174 µM; each point
represents mean values from 6 separate cells. Vehicle effects
were subtracted prior to curve-fitting.
373
Martin et al J Cardiovasc Pharmacol姠 • Volume 43, Number 3, March 2004

± 1.82 mV compared with control mean value of −5.52 ± 2.38 versus 727 ± 77, and 731 ± 71 V/s for 5 µM and 25 µM lido-
mV). The maximum upstroke velocity was unaffected by all caine, respectively). These results demonstrate that inhibition
lidocaine concentrations used in these experiments (745 ± 72 of late sodium current can reverse action potential prolonga-
and 691 ± 36 V/s for drug-free control and dofetilide periods tion elicited by IKr block in canine Purkinje fibers.

A Comparison of Assays: Action Potential

Prolongation versus hERG Current Block
To further evaluate the role of multi-channel block in
modulating delayed repolarization in vitro, we compared the
effects of haloperidol and 9 additional drugs in the hERG and
APD assay. Drugs selected included those linked to QT pro-
longation and proarrhythmia clinically as well as those not as-
sociated with delayed cardiac repolarization (Table 1). For
each drug, matched concentrations were evaluated in both the
hERG and APD assays; IC50 values for hERG block with all
drugs are shown in Table 1. Figure 5 plots mean values for
change in APD (abscissa) versus mean values for block of
hERG current (percent change, ordinate) for each of 10 drugs.
For most drugs (the exception being indomethacin), prominent
(>50%) concentration-dependent hERG block was observed
with higher concentrations (typically 10 and 100-fold total
plasma values encountered clinically). Of drugs shown to
block hERG, only 4 of 9 compounds (E-4031, erythromycin,
moxifloxacin, and telithromycin) elicited convincing concen-
tration-dependent APD prolongation along with incremental
hERG block. Four other drugs (haloperidol, fluoxetine, vera-
pamil, and terfenadine) elicited substantial concentration-
dependent hERG block despite minimal APD prolongation
(or, in some cases, APD shortening). The 1 drug demonstrat-
ing minimal hERG block (indomethacin) elicited APD short-
ening (−2.3%) at 50-fold therapeutic plasma concentrations
(279.5 µM).
It is evident that comparable levels of hERG block do not
correlate with the extent of action potential prolongation when
comparing different drugs. For example, action potential pro-
longation with cisapride was greater at the intermediate (1 µM)
versus higher concentration (57.6 ± 7.3 versus 32.4 ± 4.1% for

FIGURE 2. Effects of haloperidol on Purkinje fiber repolariza-

tion. (A) Typical effects of escalated haloperidol concentrations
(0.026, 0.26, and 2.66 µM) on the Purkinje fiber action po-
tential duration. The lower and intermediate haloperidol con-
centrations prolong repolarization by prolonging the plateau
phase of the action potential; the highest haloperidol concen-
tration depresses the plateau and triangulates the action po-
tential. (B) Summary of the effects of haloperidol on the action
potential duration (upper graph) and plateau height (mea-
sured 100 milliseconds following the action potential upstroke,
lower graph). Note the bell-shaped curve for APD prolonga-
tion with increasing haloperidol concentrations and the sig-
nificant reduction in plateau height at the highest concentra-
tion tested. Stimulation rate, 0.5 Hz. Results summarize 6
fibers, each obtained from a different heart. *P < 0.05 versus
control.

374 © 2004 Lippincott Williams & Wilkins

J Cardiovasc Pharmacol姠 • Volume 43, Number 3, March 2004
FIGURE 3. Calcium current block attenuates action potential
prolongation by IKr block. (A) Typical action potential changes
elicited by dofetilide alone (0.1 µM) and in the presence of
dofetilide plus nifedipine (5 µM). The significant action poten-
tial prolongation elicited by IKr block with dofetilide was sub-
stantially reduced during co-administration with the calcium
channel blocking agent nifedipine. Note the reduction of pla-
teau height during exposure to nifedipine. Panel B summarizes
the effects of dofetilide and dofetilide plus nifedipine on the
action potential duration and plateau height. Prolongation
with dofetilide (77%) was significantly reduced during the
subsequent co-administration of nifedipine (22%). Co-
administration of the 2 drugs also substantially reduced the
plateau height. Stimulation rate, 0.5 Hz. Results summarize 4
fibers, each obtained from a different heart. *P < 0.05 versus
control.
© 2004 Lippincott Williams & Wilkins
hERG Current, Repolarization Assays, and Multi-channel Block
1 and 10 µM concentrations, respectively) despite prominent
(>96% block) hERG block (IC50 value = 0.018 µM; Table 1).
Of all 10 compounds, only the prototypic antiarrhythmic agent
E-4031 elicited prominent APD prolongation (>25%) at thera-
peutic concentrations. While most (9/10) drugs demonstrated
concentration-dependent hERG block at or above therapeutic
concentrations, only 5 of 9 hERG blocking drugs (cisapride,
E-4031, erythromycin, moxifloxacin, telithromycin) affected
appreciable (>25%) prolongation of the action potential dura-
tion for matching concentrations.
DISCUSSION
Recent experiences suggest that delayed cardiac repolar-
ization by non-cardiovascular drugs predisposes to proarrhyth-
mia, including the potentially life-threatening cardiac arrhyth-
mia torsade-de-pointes.1,4 As these drug-induced events are
rare, emphasis is placed on in vitro preclinical assays to detect
delayed repolarization (a surrogate marker of proarrhythmia)
testing therapeutic and supratherapeutic drug concentrations
(the latter a risk factor for acquired long QT-syndrome). In
general, most drugs that delay repolarization reduce the de-
layed rectifier current, IKr, at or above therapeutic plasma con-
centrations.4 As the pore of the delayed rectifier channel in
humans is encoded by hERG,6,7 the hERG ionic current assay
(along with repolarization assays) has assumed prominence in
the in vitro evaluation of proarrhythmic liability.
The present study demonstrates concentration-
dependent block of hERG by haloperidol, characterized with
an IC50 value of 0.174 µM (Fig. 1). In contrast, APD prolon-
gation with haloperidol was greatest at 0.26 µM, with lesser
prolongation observed at a 10-fold greater concentration (Fig.
2). Haloperidol’s effect to delay repolarization at the 2 lower
concentrations was mirrored by the specific IKr blocking agent
dofetilide, while haloperidol’s effects on repolarization at the
high concentration (mitigated prolongation and plateau height
reduction) was mimicked by L-type calcium channel blockade
(by nifedipine) in the presence of continued IKr block by
dofetilide (Fig. 3). The mimicking of haloperidol’s effects by
coadministration of dofetilide and nifedipine is consistent with
multi-channel block by haloperidol mitigating the effects of
hERG block at supratherapeutic concentrations. Somewhat
unexpectedly, we also observed that lidocaine reduced and
then abolished action potential prolongation elicited by dofeti-
lide in a concentration-dependent manner (Fig. 4). Lidocaine’s
reversal of dofetilide’s effects was coincident with a reduction
in the height of the action potential plateau and mirrored ef-
fects observed with the combined administration of nifedipine
and dofetilide. The effect of lidocaine occurred without
changes in the maximum rate of rise of the action potential
upstroke, and is consistent with inhibition of late sodium cur-
rent that supports the plateau in Purkinje fibers and some ven-
tricular cells.20,21,26,27 Our results with lidocaine in canine Pur-
kinje fibers agree with an earlier rabbit Purkinje fiber study by
375
Martin et al
Abrahamsson and coworkers that showed APD prolongation
elicited by the iKr blocking agent almokalant was reversed by
lidocaine.28 However, our results with nifedipine are in dis-
agreement with analogous studies by Abrahamsson in which
nisoldipine did not reverse APD prolongation elicited by al-
mokalant. These later results may be due to a lower concentra-
tion of calcium channel blocker employed by Abrahamsson or
to species differences.
376
J Cardiovasc Pharmacol姠 • Volume 43, Number 3, March 2004
The modulation of delayed repolarization elicited by IKr
block by concomitant block of inward plateau current(s) (for
example, L-type calcium current or late sodium current) is
likely to be complex and involve at least 4 factors. First, IKr
activation is voltage dependent and spans the range of −40 to 0
mV for full activation in most species.6,7,29 Thus, a drug that
reduces (makes more negative) the plateau height by inhibiting
an inward plateau current will reduce IKr activation during the
action potential, thereby minimizing its contribution to repo-
larization and any further influence of IKr block. More negative
plateau potentials will also reduce the extent of IKr blockade
for drugs demonstrating voltage-dependent block (such as
haloperidol). Similarly, the contribution of IKs is likely to be
reduced with more negative plateau potentials (although acti-
vation of IKs is slower than IKr and is influenced more by ad-
renergic tone). Finally, reduction of ICa in the setting of IKr
block may also reduce the extent of calcium current reactiva-
tion later during the action potential plateau, thereby antago-
nizing the initiation of early afterdepolarizations implicated in
the genesis of torsade-de-pointes.28,30 Computer simulations
may be useful in elucidating these complex interactions.
It has been demonstrated that pretreatment with low con-
centrations of either lidocaine or nisoldipine directly inhibit
EADs while minimally affecting APD.28 Indeed, it has been
suggested that inhibition of inward plateau current(s) may be
protective against proarrhythmia with the novel antiarrhyth-
mic drug BRL-3287231, the antiarrhythmic drug verapamil,24
and the antidepressant drug fluoxetine.32,33 However, multi-
channel block does not always guard against proarrhythmia, as
excessive haloperidol concentrations have been linked to pro-
arrhythmia.34,35 Furthermore, inhibiting L-type calcium cur-
rent may reduce contractility, leading to other changes.32 Low-
ering the plateau height may also affect triangulation of the
ventricular action potential that, according to Hondeghem,
may be proarrhythmic.36 Lesser effects of IKr blockade on re-
FIGURE 4. Late sodium current block attenuates action poten-
tial prolongation by IKr block. (A) Typical action potential
changes elicited by dofetilide alone (0.1 µM) and in the pres-
ence of 25 µM lidocaine. Action potential prolongation elicited
by I K r block with dofetilide was reversed during co-
administration with lidocaine. Note the reduction of plateau
height during exposure to lidocaine. (B) The effects of dofeti-
lide and dofetilide plus lidocaine (5 and 25 µM) on the action
potential duration and plateau height. Prolongation with
dofetilide (100%) was significantly reduced during the subse-
quent co-administration of the lower lidocaine concentration
(5 µM, only 47% prolongation) and abolished by the higher
lidocaine concentration (25 µM, 9.5% shortening versus con-
trol). The higher lidocaine concentration also substantially re-
duced the plateau height. Stimulation rate, 0.5 Hz. Results
summarize 5 fibers, each obtained from a different heart. DoF,
dofetilide (0.1 µM); Dof + Lido 5 = 0.1 µM dofetilide + 5 µM
lidocaine; Dof + Lido 25 = 0.1 µM dofetilide + 25 µM lidocaine.
*P < 0.05 versus control.
© 2004 Lippincott Williams & Wilkins
J Cardiovasc Pharmacol姠 • Volume 43, Number 3, March 2004 hERG Current, Repolarization Assays, and Multi-channel Block

FIGURE 5. A comparison of APD prolongation and hERG current inhibition with 10 structurally diverse drugs. Plotted are changes
in APD (abscissa) versus hERG current block (ordinate). Matching concentrations of each drug were tested in each assay, with
concentrations tested representing therapeutic values and higher multiples (see Table 1). Nine of 10 drugs demonstrated
prominent concentration-dependent hERG block at supratherapeutic concentrations; however, only 4 demonstrated convincing
monotonic concentration-dependent APD prolongation with increasing hERG block (E-4031, erythromycin, moxifloxacin, and
telithromycin). Comparable levels of hERG block did not predict the extent of APD prolongation. Values represent means of a
minimum of 6 determinations per drug for each assay. Purkinje fiber repolarization study results from this laboratory have been
reported previously for 5 of the 10 compounds (erythromycin, fluoxetine, indomethacin, moxifloxacin, and terfenadine13). See
text for further discussion.

polarization resulting from multi-channel block with non-
hERG channels will depend upon the characteristics of drug
block of each channel, the properties and densities of indi-
vidual currents contributing to the tissue-specific action poten-
tial configurations, and patterns of electrical activity. Finally,
it must be recognized that delayed repolarization is a surrogate
marker for proarrhythmia and torsade-de-pointes that can only
be observed directly in intact heart preparations.
To our knowledge, this is the first study to systemically
compare matched concentrations of a diverse set of drugs from
different therapeutic areas in both a hERG and APD repolar-
ization assay. Our results demonstrate prominent hERG cur-
rent inhibition for 9 of 10 drugs at excessive concentrations
(Fig. 5, Table 1). One group of 4 drugs (E-4031, erythromycin,
moxifloxacin, and telithromycin) elicited concentration-
dependent prolongation of the action potential duration paral-
leling incremental hERG block. A second group of four drugs
(fluoxetine, haloperidol, terfenadine, and verapamil) mini-
mally prolonged (or in some cases, shortened) the canine Pur-
kinje fiber APD despite affecting substantial (>70%) hERG
block at supratherapeutic concentrations. Cisapride provides a
more complicated example, eliciting 94% block of hERG cur-
rent at a concentration of 0.1 µM, maximal APD prolongation
(57.6 ± 7.3%) at 1 µM, and lesser prolongation (32.4 ± 4.1%) at
10 µM concentration. These findings suggest that drugs that
inhibit hERG current may be further categorized into two
groups, with one group predominantly inhibiting only hERG
current (affecting prominent APD prolongation), and a second
group inhibiting multiple cardiac channels and thereby modu-
lating hERG current block (especially at high concentrations).
This conclusion is not surprising since one would expect
higher drug concentrations to foster block of multiple channels
as well as to potentially indirectly modulate block of other
channels (for example, reducing plateau height, thereby affect-
ing other currents later during the action potential).
We compared hERG block with changes in canine Pur-
kinje fiber repolarization, as it is not possible to routinely
evaluate hERG block and APD changes in native human ven-

© 2004 Lippincott Williams & Wilkins 377

Martin et al
tricular preparations. Indeed, the generally low hERG current
density in native cardiac preparations and presence of multiple
overlapping currents provide formidable obstacles for evalu-
ating hERG current inhibition in the same preparations as
those used to record action potentials. We would expect that
block of stably transfected hERG in HEK-293 cells would be
comparable to block of canine IKr in Purkinje myocytes be-
cause a) the cERG polypeptide is 97% homologous to its hu-
man counterpart (with 100% identity in the N-terminal PAS
domain, transmembrane segments, and cyclic nucleotide bind-
ing domain37), and b) cERG current shows similar sensitivity
to block by terfenadine, astemizole, and the methanesulfona-
nilide compound MK-499.38 It is unlikely that differences in
experimental conditions account for the lack of correlation be-
tween hERG block and APD prolongation in canine Purkinje
fibers since identical concentrations of drugs (from identical
lots) were evaluated in both assays under comparable experi-
mental conditions (temperature, slow activity, roughly compa-
rable voltage excursions). Whether the presence of channel
subunits significantly affects drug block of hERG current re-
mains unresolved.39,40 While it is well known that differences
in electrical activity exist between Purkinje fibers and various
regions of ventricular myocardium, it is uncertain how these
differences would affect the correlation between hERG and
APD assays with different tissue types. It has been shown that
high lidocaine concentrations reverse APD prolongation elic-
ited by almokalant in rabbit Purkinje fibers but not rabbit ven-
tricular papillary muscle.28 What is clearly demonstrated in
this study is that only a subset of drugs that elicit prominent
concentration-dependent hERG block elicit prominent con-
centration-dependent prolongation of the action potential du-
ration in vitro. These findings highlight the utility of repolar-
ization studies with native tissues as simultaneous, integrated
assays of drug effects on multiple cardiac ion channels.
In summary, the present studies demonstrate the diffi-
culty of assessing potential proarrhythmic risk on the basis of
concentration-dependent drug effects in 2 common in vitro
preclinical assays. Only 1 of 10 drugs tested (indomethacin)
does not block hERG at supratherapeutic concentrations. The
extent of hERG block is poorly correlated with action potential
prolongation in vitro, consistent with additional drug effects
on non-hERG channels (multi-channel block), especially at
high drug concentrations. While hERG block detects many
drugs linked to delayed repolarization and torsade-de-pointes,
it is not fully predictive of proarrhythmic risk; for example,
both fluoxetine and verapamil elicit prominent hERG block
but are not generally associated with either QT interval pro-
longation or torsade-de-pointes. Together, these observations
demonstrate that the hERG assay is overly simplistic in pre-
dicting delayed repolarization and torsade-de-pointes for some
drugs (especially at supratherapeutic concentrations), and that
a repolarization assay provides additional information to inter-
378
J Cardiovasc Pharmacol姠 • Volume 43, Number 3, March 2004
pret hERG assay results during the initial steps in the preclini-
cal evaluation of potential proarrhythmic risk.
ACKNOWLEDGMENTS
The authors thank Katina Daniell, Sandra Leitza, and
Gilbert Diaz for their unfailing provisions of HEK-293 cells;
Brian Ebert and Dr. Letty Medina for veterinarian assistance;
and Dr. James Sullivan for continued support.
REFERENCES
1. Haverkamp W, Breithardt G, Camm AJ, et al. The potential for QT pro-
longation and proarrhythmia by non-antiarrhythmic drugs. Clinical and
regulatory implications. Report on a policy conference of the European
Society of Cardiology. Eur Heart J. 2000;21:1216–1231.
2. Committee for Proprietary Medicinal Products (CPMP). Points to con-
sider: the assessment of the potential for QT-interval prolongation by non-
cardiovascular medicinal products. London: The European Agency for
the Evaluation of Medicinal Products; Dec.1997.
3. Hammond TG, Carlsson L, Davis AS, et al. Methods of collecting and
evaluating non-clinical cardiac electrophysiology data in the pharmaceu-
tical industry: results of an international survey. Cardiovasc Res. 2001;
49:741–750.
4. Malik M, Camm AJ. Evaluation of drug-induced QT interval prolonga-
tion. Implications for drug approval and labeling. Drug Saf. 2001;24:323–
351.
5. International Conference on Harmonisation of Technical Requirements
for Registration of Pharmaceuticals for Human Use. Draft Consensus
Guideline. Safety Pharmacology Studies for Assessing the Potential for
Delayed Ventricular Repolarization (QT Interval Prolongation) by Hu-
man Pharmaceuticals. Step 2 Release of the ICH Process. 7 Feb, 2002.
6. Sanguinetti MC, Jiang C, Curran ME, et al. A mechanistic link between an
inherited and an acquired cardiac arrhythmia: hERG encodes the IKr po-
tassium channel. Cell. 1995;81:299–307.
7. Trudeau MC, Warmke JW, Ganetzky B, et al. HERG, a human inward
rectifier in the voltage-gated potassium channel family. Science. 1995;
269:92–95.
8. Glassman AH, Bigger JT Jr. Antipsychotic drugs: Prolonged QTc inter-
val, torsade de pointes and sudden death. Am J Psychiatry. 2001;158:
1774–1782.
9. Suessbirch H, Schoenherr R, Heinemann SH, et al. The inhibitory effect
of the antipsychotics drug haloperidol on hERG potassium channels ex-
pressed in Xenopus oocytes. Br J Pharmacol. 1997;120:968–974.
10. Ogata N, Narahashi T. Block of sodium channels by psychotropic drugs in
single guinea-pig cardiac myocytes. Br J Pharmacol. 1989;97:905–913.
11. Buckley NA, Sanders P. Cardiovascular adverse effects of antipsychotic
drugs. Drug Saf. 2000;23:215–228.
12. Zhou Z, Gong B, Ye Z, et al. Properties of hERG channels stably ex-
pressed in HEK-293 cells studied at physiological temperatures. Biophys
J. 1998;74:230–241.
13. Gintant GA, Limberis JT, McDermott JS, et al. The canine Purkinje fiber:
an in vitro model system for acquired long QT syndrome and drug-
induced arrhythmogenesis. J Cardiovasc Pharmacol. 2001;37:607–618.
14. Hondeghem LM, Snyders DJ. Class III antiarrhythmic agents have a lot of
potential but a long way to go. Reduced effectiveness and dangers of re-
verse use-dependence. Circulation. 1990;81:686–690.
15. Katritsis D, Morgan J, Brachmann J, et al. Electrophysiological effects of
E-4031, a drug with selective class III properties, in man. PACE. 20(part
1):930–937.
16. Medical Economics Staff (ed). Physicians Desk Reference. Medical Eco-
nomics Company, Oradell, NJ; 2002.
17. Woosley RW. Available at: http://www.Torsades.org. Accessed Oct. 10,
2002
18. Aventis Pharma. Ketek (telithromycin) briefing document for the FDA
Anti-Infective Drug Products Advisory Committee Meeting. 4/26/2001.
© 2004 Lippincott Williams & Wilkins
J Cardiovasc Pharmacol姠 • Volume 43, Number 3, March 2004
http://www.fda.gov/ohrms/dockets/ac/01/briefing/3746b_02_FDA.pdf.
Accessed Oct. 5, 2002
19. Gwilt M, Arrowsmith JE, Blackburn KJ, et al. UK-68798: a novel, potent
and highly selective class III antiarrhythmic agent which blocks potas-
sium channels in cardiac cells. J Pharm Exp Ther. 1991;256:318–324.
20. Wasserstrom JA, Salata JJ. Basis for tetrodotoxin and lidocaine effects on
action potentials in dog ventricular muscle. Am J Physiol. 1988;254:
H1151–H1166.
21. Carmeliet E. Slow inactivation of the sodium current in rabbit cardiac
Purkinje fibers. Pflugers Arch. 1987;408:18–26.
22. Carmeliet EE. Voltage- and time-dependent block of the delayed K+ cur-
rent in cardiac myocytes by dofetilide. J Pharmacol Exp Ther. 1992;262:
809–817.
23. Yang T, Snyders D, Roden T. Drug block of IKr: Model systems and rel-
evance to human arrhythmias. J Cardiovasc. Pharmacol. 2001;38:737–
744.
24. Zhang DS, Zhou Z, Gong Q, et al. Mechanisms of block and identification
of the verapamil binding domain to hERG potassium channels. Circ Res.
1999;84:989–998.
25. Kass RS. Nisoldipine: a new, more selective calcium current blocker in
cardiac Purkinje fibers. J Pharmacol Exp Ther. 1982;223:446–456.
26. Gintant GA, Datyner NB, Cohen IS. Slow inactivation of a tetrodotoxin-
sensitive current in canine cardiac Purkinje fibers. Biophys J. 1984;45:
509–512.
27. Zygmunt AC, Eddlestone GT, Thomas GP, et al. Larger late sodium cur-
rent conductance in M cells contributes to electrical heterogeneity in ca-
nine ventricle. Am J Physiol. 2001;281:H689–H697.
28. Abrahamsson C, Carlsson L, Duker G. Lidocaine and nisoldipine attenu-
ate almokalent-induced dispersion of repolarization and early afterdepo-
larizations in vitro. J Cardiovasc Electrophysiol. 1996;7:1074–1081.
29. Gintant GA. Two components of delayed rectifier current in canine atrium
and ventricle. Does iKs play a role in the reverse rate-dependence of Class
III agents. Circ Res. 1996;78:26–37.
© 2004 Lippincott Williams & Wilkins
hERG Current, Repolarization Assays, and Multi-channel Block
30. January CT, Shorofsky S. Early afterdepolarizations: Newer insights into
cellular mechanisms. J Cardiovasc Electrophysiol. 1990;1:161–169.
31. Bril A, Gour B, Bonhomme M, et al. Combined potassium and calcium
channel blocking activities as a basis for antiarrhythmic efficacy with low
proarrhythmic risk: Experimental profile of BRL-32872. J Pharmacol
Exp Ther. 1996;276:637–646.
32. Pacher P, Magyar J, Szigligeti P, et al. Electrophysiological effects of
fluoxetine in mammalian cardiac tissues. Naunyn Schmiedebergs Arch
Pharmacol. 2000;361:67–73.
33. Thomas D, Gut B, Wendt-Nordahl G, et al. The antidepressant drug fluox-
etine is an inhibitor of human ether-a-go-go related gene (hERG) potas-
sium channels. J Pharmacol Exp Ther. 2002;300:543–548.
34. Metzger E, Friedman R. Prolongation of the corrected QT and torsades de
pointes cardiac arrhythmia associated with intravenous haloperidol in the
medically ill. J Clin Psychopharmacol. 1993;13:128–132.
35. Hunter N, Stern TA. The association between intravenous haloperidol and
torsades de pointes. Psychosomatics. 1995;36:541–549.
36. Hondeghem LM, Carlsson L, Duker G. Instability and triangulation of the
action potential predict serious proarrhythmia, but action potential dura-
tion prolongation is antiarrhythmic. Circ. 2001;103:2004–2013.
37. Zehelein J, Zhang W, Koenen M, et al. Molecular cloning and expression
of cERG, the ether a go-go-related gene from canine myocardium.
Pflugers Archiv. 2001;442:188–191.
38. Wang J, Della Penna K, Wang H, et al. Functional and pharmacological
properties of canine ERG potassium channels. Am J Physiol. 2003;284:
H256–H267.
39. Abbott GW, Sesti F, Splawski I, et al. MiRP1 forms IKr potassium chan-
nels with hERG and is associated with cardiac arrhythmia. Cell. 1999;97:
175–187.
40. Weerapura M, Nattel S, Chartier D, et al. A comparison of currents carried
by hERG, with and without coexpression of MiRP1, and the native rapid
delayed rectifier current. Is MiRP1 the missing link? J Physiol (Lond).
2002;540:15–27.
379