Journal of Pharmacological and Toxicological Methods 61 (2010) 178–191

Contents lists available at ScienceDirect

Journal of Pharmacological and Toxicological Methods

j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / j p h a r m t ox

Original article

Investigating dynamic protocol-dependence of hERG potassium channel inhibition at

37 °C: Cisapride versus dofetilide
James T. Milnes a,1, Harry J. Witchel a,2, Joanne L. Leaney b, Derek J. Leishman b,3, Jules C. Hancox a,⁎
a
Department of Physiology and Pharmacology, School of Medical Sciences, University Walk, Bristol, BS8 1TD, UK
b
Pfizer Global Research and Development, Sandwich Labs, Ramsgate Road, Sandwich, Kent CT13 9NJ, UK

a r t i c l e i n f o a b s t r a c t

Article history: Introduction: Pharmacological inhibition of cardiac potassium channels encoded by hERG (human ether-
Received 24 December 2009 à-go-go-related gene) is associated with QT interval prolongation and torsades de pointes arrhythmia.
Accepted 11 February 2010 Electrophysiological assays of hERG channel inhibition are integral to the safety testing of novel drug
candidates. This study was conducted to compare, for the high affinity hERG inhibitors dofetilide and
Keywords: cisapride, hERG blockade between action potential (AP) and conventional (step and step–ramp) screening
Action potential
waveforms. Furthermore, it evaluated dynamic (pulse-by-pulse) protocol-dependence of hERG channel
Cisapride
Dofetilide
inhibition by these drugs. Methods: Whole-cell patch-clamp recordings were made at 37 °C from hERG-
hERG expressing HEK 293 cells. Half-maximal inhibitory concentrations (IC50 values) for IhERG blockade were
Methods obtained using conventional voltage clamp and action potential clamp, using previously digitised ventricular
QT interval and Purkinje fibre (PF) AP waveforms. Results: A more marked variation in IC50 values with different
Torsades de points command waveforms was observed for cisapride (ranging from 7 to 72 nM) than for dofetilide (ranging
Screening from 4 to 15 nM), with higher IC50s obtained with AP than step or step–ramp commands. The two drugs
differed little from one another in effects on voltage-dependent activation; however, IhERG blockade by each
drug was initially voltage-dependent, but at steady-state was only voltage-dependent for cisapride. There
was comparatively little difference between the two drugs in effects on IhERG availability or time constants of
development of inactivation. Features of time-dependence of blockade and the use of protocols employing
varying rest periods in drug or commands of alternating duration highlighted a pronounced ability of
cisapride, but not dofetilide, to dissociate and reassociate from hERG on a pulse-by-pulse basis. Discussion:
Protocols described here that demonstrated dynamic variation (drug dissociation/reassociation) in hERG
channel current blockade at 37 °C for cisapride may have future value for investigating drug interactions
with the hERG channel. Downloadable digitised ventricular and PF AP waveforms that can be used in AP
clamp experiments also accompany this article.
© 2010 Elsevier Inc. All rights reserved.

1. Introduction assays of delayed action potential repolarisation, QT interval prolonga-

Diverse cardiac and non-cardiac drugs have been associated with the
polymorphic ventricular tachycardia torsades de pointes ((TdP); e.g
(Haverkamp et al., 2000; Redfern et al., 2003; Yap & Camm, 2003)). A
propensity for novel compounds to cause TdP is a major consideration
during safety testing during the drug discovery and development
process. Due to the fact that the incidence of TdP arrhythmia tends to be
low (Yap & Camm, 2003), a range of surrogate markers of arrhythmia
risk are used during the pre-clinical safety testing process, including
⁎ Corresponding author. Tel.: + 44 117 93312292.
E-mail address: jules.hancox@bristol.ac.uk (J.C. Hancox).
1
Present address: Xention Ltd., Iconix Park, London Road, Pampisford, Cambridge,
CB22 3EG, UK.
2
Present address: Brighton and Sussex Medical School, University of Sussex, Falmer
BN1 9PX, UK.
3
Present address: Lilly Research Laboratories, Greenfield Laboratories, P.O. Box 708,
Greenfield, IN 46140, USA.
tion and in vitro ion channel blockade (Friedrichs, Patmore, & Bass,
2005; Pugsley, Hancox, & Curtis, 2008; Gintant, 2008; Hancox, McPate,
El Harchi, & Zhang, 2008). A striking feature of virtually all drugs
associated with TdP is a propensity to inhibit the cardiac IKr/hERG
potassium channel, which normally plays a key role in ventricular action
potential repolarisation (Gintant, 2008; Hancox et al., 2008; Redfern
et al., 2003). Due to this, in vitro IKr assays comprise an essential early
part of an integrated risk assessment recommended in the guidelines
from the International Conference on Harmonization (ICH S7B Step 4;
discussed in (Friedrichs et al., 2005; Hancox et al., 2008)). Typically, such
assays involve assessment of the inhibition of ionic current carried by
recombinant hERG channels expressed in mammalian cell lines (see
(Gintant, 2008; Hancox et al., 2008) for recent reviews).
Due to the high susceptibility of hERG channels to pharmacological
blockade, many test compounds (up to ∼70%; (Shah, 2005)) are likely to
interact with hERG when tested over a wide range of concentrations,
raising the possibility of ‘false positives’ and an unduly high compound

1056-8719/$ – see front matter © 2010 Elsevier Inc. All rights reserved.
doi:10.1016/j.vascn.2010.02.007
 J.T. Milnes et al. / Journal of Pharmacological and Toxicological Methods 61 (2010) 178–191
attrition rate during drug safety assessment (Shah, 2005). An influential
concept to aid the interpretation of hERG inhibition data is that of the
‘cardiac safety index’ (ratio of half-maximal efficacy dose (ED50) to half-
maximal inhibitory concentration (IC50) for IKr/hERG blockade (Cavero
et al., 2000)) or ‘safety margin’ (the ratio between maximal effective free
plasma concentration (Cmax) and IKr/hERG IC50; (Redfern et al., 2003)).
In an extensive data evaluation by Redfern et al. (2003), drugs without
an incidence of TdP generally exhibited a safety margin of N30-fold, and
consequently a provisional safety margin of 30-fold was recommended
for drugs to treat serious debilitating diseases (Redfern et al., 2003;
Hancox et al., 2008). Clearly, measured IC50 values for hERG current
(IhERG) inhibition are key determinants of calculated safety margin
values, and it is significant in this regard that experimentally obtained
IhERG IC50s for some drugs can vary depending on measurement
conditions, including expression system (mammalian versus non-
mammalian), voltage stimulus protocol and experimental temperature
(Witchel, Milnes, Mitcheson, & Hancox, 2002; Kirsch et al., 2004; Yao
et al., 2005; Stork et al., 2007; Hancox et al., 2008). Whilst the use of a
mammalian cell expression system is recognised as important for the
pharmacological assessment of hERG IC50 (Witchel et al., 2002; Hancox
et al., 2008), there is at present no single prescribed voltage stimulus
pattern for IC50 assessment, and both voltage step and ramp protocols
are routinely used (Kirsch et al., 2004; Yao et al., 2005; Hancox et al.,
2008). Voltage step and ramp protocols differ significantly from cardiac
action potential (AP) waveforms, however, and it might be anticipated
that, for some drugs, IC50 values may differ between AP and conven-
tional step or ramp protocols used with mammalian hERG expression
systems.
The present study was designed (i) to compare, at physiological
temperature, IhERG blockade by two archetypal high affinity hERG
inhibitors between action potential, step and step–ramp voltage
protocols; (ii) to place any observed protocol-dependence of IC50
values in the context of other characteristics of blockade investigated
under the same recording conditions. The compounds chosen for this
investigation were the antiarrhythmic methanesulphonanilide drug
dofetilide and the non-cardiac drug cisapride (which was withdrawn
due to multiple reports of cardiac rhythm abnormalities (Henney,
2000)). Binding to the hERG channel by each of these drugs has been
mapped to residues within the channel's central cavity (Lees-Miller,
Fig. 1. Spread of published IC50 values for dofetilide and cisapride (mammalian expression systems). A) Chemical structure of dofetilide. B) Chemical structure of cisapride. C) Scatter
plot showing range of IC50 values for cisapride and dofetilide plotted with a logarithmic Y axis. For each plot, the long solid line represents the calculated mean IC50 value from the
data, whilst the long intermittent line represents the calculated median IC50 value. The shorter lines represent 25th and 75th percentile values. Tables 1 and 2 provide summary
information on the studies from which the IC50 values included in this figure come.
179
Duan, Teng, & Duff, 2000; Kamiya, Niwa, Mitcheson, & Sanguinetti,
2006; Mitcheson, Chen, Lin, Culberson, & Sanguinetti, 2000; Kamiya,
Niwa, Morishima, Honjo, & Sanguinetti, 2008), whilst a comparison of
published IC50 values from different studies (obtained from mamma-
lian cells; see Fig. 1 and Tables 1 and 2) indicates a markedly wider
spread of values for cisapride than dofetilide. The results of the
present study show a much wider range of measured IC50 values
between the different protocols for cisapride than for dofetilide, with
notably higher IC50 values with action potential than step/step–ramp
voltage waveforms. The results of further experiments indicate a
marked ability of cisapride, but not dofetilide, to interact with hERG
on a dynamic, pulse-by-pulse basis. In this study, protocols are
presented that facilitate the rapid identification of dynamic hERG
channel-drug interaction, and action potential waveforms that may be
useful for assessing IC50 values of hERG inhibitors are made available
to the reader.
2. Methods
2.1. Maintenance of mammalian cell lines stably expressing hERG
Experiments were performed on a cell line (Human Embryonic
Kidney; HEK 293) stably expressing hERG (provided by Prof. Craig
January, University of Wisconsin; (Zhou et al., 1998)). Cells were
passaged using a non-enzymatic agent (Splitase, AutogenBioclear)
and plated out onto small sterilised glass coverslips in 30 mm petri
dishes containing a modification of Dulbecco's Modified Eagle's
medium with Glutamax-1 (DMEM; Gibco, Invitrogen, Paisley, UK),
supplemented with 10% foetal bovine serum (Gibco), 400 μg ml− 1
gentamycin (Gibco) and 400 μg ml− 1 geneticin (G418; Gibco). The cells
were incubated at 37 °C for a minimum of 2 days prior to any electro-
physiological study.
2.2. Electrophysiological recording
Glass coverslips onto which hERG-expressing cells had been plated
were placed in a bath (0.5 ml volume) mounted on an inverted
microscope (Nikon Diaphot) and the cells were superfused with Normal
Tyrode's solution which contained (in mM): 140 NaCl, 4 KCl, 2.5 CaCl2, 1
180 J.T. Milnes et al. / Journal of Pharmacological and Toxicological Methods 61 (2010) 178–191

Table 1
Range of published IC50 values for dofetilide. The table shows details of selected studies in which hERG blocking potency of dofetilide was assessed using mammalian cell lines. Cell
line, experimental protocol/conditions and temperature information are given alongside IC50 values (rounded to nearest integer and highlighted in the table by use of bold font). In
temperature column (Temp °C), ‘RT’ denotes room temperature. Note that the table is not intended to show all relevant published data, rather a selection in order to show spread of
obtained IC50 values.

1st author Year Cell line Temp IC50 VHold Step Step Tails Tails S-to-S [K]o Config
(°C) (nM) (mV) (mV) (s) (mV) (s) (s) (mM)

Kupershmidt 2003 CHO-K1 20–22 15 −80 20 2 − 50 2 10 4 2 step

Snyders 1996 HEK 293 21–23 12 − 80 20 30 − 40 2 45 4 2 step
Rampe 1998 HEK 293 21–23 10 − 80 20 20 5 1 step
Rampe 1997 Mouse L Cell RT 15 −80 20 2 − 40 1.6 5 2 step
Kang 2004 CHO RT 11 − 80 20 2 − 40 2 40 5 2 step
Guth 2004 HEK 293 RT 27
Rezazadeh 2004 HEK 293 4 − 80 20 4 − 50 6 11 5 2 step
Weerapura 2002 CHO 34–36 10 0 − 110 0.025 0 5 1 step
Perrin 2008 HEK 293 51 − 80 20 3 − 110 0.5 10 5 2 step
Limberis 2006 HEK 293 43 − 80 40 2 − 50 2 1 2 step
Limberis 2006 HEK 293 46 −80 40 2 − 50 2 20 2 step

MgCl2, 10 Glucose, 5 HEPES, (titrated to pH 7.45 with NaOH). Drugs
were also added to this to make up test solutions at the final concen-
trations mentioned in the ‘Results’ text. Patch-pipettes (Corning 7052
glass, AM Systems Inc.) were pulled to resistances of 1–2 MΩ (Narishige
PP83) and fire-polished to 2.5–3.5 MΩ (Narishige, MF83). The internal
dialysis solution contained (in mM): 130 KCl, 1 MgCl2, 5 EGTA, 5 MgATP,
and 10 HEPES (titrated to pH 7.2 with KOH). The ‘pipette-to-bath’ liquid
junction potential measured for this filling solution was −3.2 mV. Since
this value was small, no corrections of membrane potential were made.
These external and pipette solutions have been used in previous hERG
channel studies from our laboratory (e.g. (Milnes, Crociani, Arcangeli,
Hancox, & Witchel, 2003; Ridley, Dooley, Milnes, Witchel, & Hancox,
2004; Duncan et al., 2007)). Whole-cell patch-clamp recordings of
membrane currents were made using an Axopatch 1D amplifier
(Axon Instruments) and a CV-4 1/100 headstage. Between 80 and 90%
of the electrode series resistance could be compensated. Voltage clamp
commands were generated using Clampex 8 (Axon Instruments). Data
were recorded via a Digidata 1200B interface (Axon Instruments) and
stored on the hard-disk of a Viglen computer prior to analysis. Data
digitisation rates were 2–25 kHz during all protocols and an appropriate
bandwidth of 2–10 kHz was set on the amplifier.
2.3. Data analysis and presentation
Data were analyzed using Clampfit 8 (Axon Instruments), Excel
2000 and Prism v.3 (Graphpad Inc.) software. Data are presented as
mean ± standard error of the mean (SEM). Statistical comparisons
were made using a two-tailed Student's t-test (paired or un-paired) or
one- / two-way analysis of variance (ANOVA) with a Bonferroni post-
test using Prism v.3 software (Graphpad Inc.). P values of less than
0.05 were taken as statistically significant.

Table 2
Range of published IC50 values for cisapride. The table shows details of selected studies in which hERG blocking potency of dofetilide was assessed using mammalian cell lines. Cell
line, experimental protocol/conditions and temperature information are given alongside IC50 values (rounded to nearest integer and highlighted in the table by use of bold font). ‘A’ —
calculated IC50 based on 2 published data points assuming an nH of 1. In temperature column (Temp °C), ‘RT’ denotes room temperature. Note that the table is not intended to show all
relevant published data, rather a selection in order to show spread of obtained IC50 values.

1st author Year Cell line Temp IC50 VHold Step Step Tails Tails S-to-S [K]o Config
(°C) (nM) (mV) (mV) (s) (mV) (s) (s) (mM)

Rampe 1997 Mouse L cell RT 45 − 80 20 2 − 40 1.6 5 2 step

Rampe 1997 Mouse L Cell RT 7 − 80 20 20 – – 20 1 step
Kirsch 2004 HEK 293 RT 26 − 80 20 2 −50 2 10 4 2 step
Mohammad 1997 HEK 293 22–23 7 − 80 10 10 −50 5 30 4 2 step
Mohammad 1997 HEK 293 22–23 64.9A − 80 − 20 4 − 50 5 4 2 step
Mohammad 1997 HEK 293 22–23 11.6A − 80 20 4 − 50 5 4 2 step
Rezazadeh 2004 HEK 293 4 − 80 20 4 − 50 6 11 5K 2 step
Guth 2004 HEK 293 22–23 5
Walker 1999 CHO-K1 20–22 16 − 75 25 3.9 − 55 5 10 4.8 2 step
Toga 2007 HEK 293 23 9 − 80 40 1 Ramp 0.5 /s 4 4 Step–ramp
Lin 2005 HEK 293 23 8 − 80 50 4 − 50 5 15 0 2 step
Lin 2005 HEK 293 23 24 − 80 50 4 − 50 5 15 5 2 step
Lin 2005 HEK 293 23 109 − 80 50 4 − 50 5 15 135 2 step
Lin 2005 HEK 293 23 174 − 80 50 4 − 50 5 15 0 Cs 2 step
Lin 2005 HEK 293 23 192 − 80 50 4 − 50 5 15 5 Cs 2 step
Lin 2005 HEK 293 23 196 − 80 50 4 − 50 5 15 135 Cs 2 step
Perrin 2008 HEK 293 21 − 80 20 3 − 110 0.5 10 5 2 step
Walker 1999 CHO-K1 37 24 − 75 25 3.9 − 55 5 10 4.8 2 step
Fossa 2004 HEK 293 35 15 − 80 20 1 Ramp 0.5 /s 4 4 Step–ramp
Martin 2004 HEK 293 37 18 − 80 0 3 − 50 4 15 5 2 step
Wang 2003 HEK 293 36 14 − 80 40 1 − 50 4 2 step
Furuta 2004 HEK 293 37 30 − 70 0 0.75 − 50 0.75 4 2 step
Chiu 2004 HEK 293 37 7 − 75 10 0.5 − 40 0.5 10 4 2 step
Kirsch 2004 HEK 293 37 23 − 80 20 2 − 50 2 10 4 2 step
Kirsch 2004 HEK 293 37 27 − 80 20 1 Ramp 0.5 /s 5 4 Step–ramp
Potet 2001 COS7 35 240 − 80 10 0.5 − 60 0.5 3 4 2 step
 J.T. Milnes et al. / Journal of Pharmacological and Toxicological Methods 61 (2010) 178–191 181

2.4. Equations where τ is the time constants for the decaying tail current. A represents
the total current fitted and C is the residual current component.
2.4.1. Equation 1 - fractional block
Fractional block of IhERG was calculated using the following equation: 2.4.7. Equation 7 - bi-exponential function
The decay of IhERG tails during current deactivation was fitted with
IhERG
Drug a bi-exponential equation of the form:
Fractional block = 1− ð1Þ
IhERG
Control
  !
−x −x
Where IhERG-Control and IhERG–Drug are membrane currents recorded y = As × exp + Af × exp +C ð7Þ
in the absence and presence of drug respectively. τs τf

2.4.2. Equation 2 - modified Hill equation where τs and τf represent the time constants for the slow and fast
The relationship between drug concentration and IhERG block by test components of the decaying tail current and As and Af represent the
compounds was determined by fitting data with a Hill equation of the total current fitted by each component, respectively. C described any
form: residual unfitted (non-deactivating) current component – typically at
or close to zero.
1
Y= ð2Þ
1 + 10½ðLogIC50 −X Þ H 
n

2.5. Drugs
where Y is fractional block at a given concentration of drug X; IC50 is
the concentration at which IhERG is half-maximally blocked and nH is Cisapride (Research Diagnostics Inc., New Jersey, USA), dofetilide
the Hill coefficient for the fit. [UK-068798] (Pfizer Global Research & Development, Sandwich, UK)
were dissolved in reagent-grade absolute ethanol or water to produce
2.4.3. Equation 3 - voltage dependence of IhERG activation stock solutions of 3 and 5 mM respectively and stored at −20 °C. Drugs
The voltage dependence of IhERG activation was determined by were serially diluted in vehicle prior to a final dilution of 1 in 10,000 in
fitting the values of the normalised tail currents with a modified Tyrode's solution to produce drug concentrations mentioned in the
Boltzmann equation of the form: ‘Results’ section. This ensured a constant final vehicle concentration of
b0.01% ethanol. Solutions containing drug were made up fresh each
IMax experimental day. During recordings, all drug-containing solutions were
I= 0
1 ð3Þ
applied to the cells under study using a home built, warmed, multi-
V1=2−Vm
1 + exp@ A barrelled solution application device (Levi, Hancox, Howarth, Croker, &
k Vinnicombe, 1996) capable of changing the bathing solution surround-
ing a cell in b1 s.
where I = IhERG tail amplitude following test potential Vm, IMax is the
maximal IhERG tail observed, V½ = potential at which IhERG was half 2.6. Voltage command protocols employed to produce concentration–
maximally activated and k is the slope factor describing the voltage- response relationships
dependent relationship.
Voltage command waveforms in this study are shown, overlaid
2.4.4. Equation 4 - voltage dependence of IhERG availability and to scale in Fig. 2. The voltage step protocol was comprised of a 2 s
0 1
depolarising command from a holding potential of −80 mV either to
B C +20 mV (as shown in Fig. 2), or −20 mV (not shown), followed by a 4 s
B C
B 1 C repolarising step to −40 mV to monitor IhERG tails. As in previous
Y = 1−B 0
 1C ð4Þ
B C studies from our laboratory (e.g. (Milnes et al., 2003; Milnes, Witchel,
B V1=2−Vm C
@
1 + exp@ AA Leaney, Leishman, & Hancox, 2006; McPate, Duncan, Hancox, & Witchel,
k
2008)), tail current amplitude was measured as the difference in
maximal outward tail current and current measured during a brief
where Y = IhERG availability (between 0 and 1) at test potential Vm, V½ = (50 ms) pre-pulse from −80 mV to −40 mV (which allows instanta-
potential at which IhERG was half maximally available and k is the slope neous leak current to be monitored in the nominal absence of IhERG
factor for the relationship.

2.4.5. Equation 5 - mono-exponential fit to the development of block

during a sustained depolarisation
Development of IhERG during a sustained depolarisation was fitted
with a mono-exponential function of the form:

Y = Span × ð1− expð−K × t ÞÞ + YMin ð5Þ

where Y is fractional block at time t (in seconds), Span is the difference

between the minimum block, YMin, and the extrapolated plateau
value. K is the rate-constant of the exponential association. From this a
τ value could be calculated as 1/Κ.

2.4.6. Equation 6 - mono-exponential function

Fig. 2. Voltage protocols used. The figure shows, to scale, superimposed voltage
The decay phase of rapidly inactivating IhERG was fitted with a commands used to study blocking potency of dofetilide and cisapride, including voltage
mono-exponential equation of the form: step, step–ramp, ventricular action potential and Purkinje fibre action potential

−x waveforms. An Excel file containing the ventricular and Purkinje fibre waveforms used
y = A × exp +C ð6Þ (each as columns of time and voltage data points) accompanies this article online and is
τ available for download.
182 J.T. Milnes et al. / Journal of Pharmacological and Toxicological Methods 61 (2010) 178–191
activation, (Milnes et al., 2003, 2006)). The start-to-start (S-to-S) interval
between successive applications of the protocol was 12 s (Milnes et al.,
2003). The ‘step–ramp’ protocol employed was comprised of a 1 s
depolarising step from the holding potential of −80 mV to +20 mV,
followed by a repolarising ramp voltage (velocity of 0.5 V s− 1). The S-to-S
interval between successive applications of the step–ramp protocol was
10 s. Action potential stimulus waveforms, applied using the action
potential voltage clamp technique (Zhou et al., 1998; Hancox, Levi, &
Witchel, 1998; Milnes et al., 2006), were a previously digitised guinea-
pig ventricular action potential (VAP) waveform (applied at S-to-S
intervals of both 2 s and 12 s) and a canine Purkinje fibre AP (PAP; applied
at a S-to-S interval of 2 s). We have shown previously that endogenous
‘leak’ current during AP clamp measurement of IhERG can be eliminated
using a p/4 protocol, in which currents elicited by a series of inverted,
scaled-down command waveforms are summated and used to enable
leak subtraction (see Hancox et al., 1998). The ‘p/n’ leak subtraction
function in pClamp was used here to perform leak subtraction during AP
protocols (cf Milnes et al., 2006). Capacitative current, which can overlap
ionic current during dynamic (AP and ramp) voltage changes, was
obviated by utilising membrane capacitance compensation on the
recording amplifier. As illustrated in Fig. 2, the shape and duration of
the physiological waveforms used differed markedly from those of either
step or step–ramp protocols. During VAP or step–ramp protocols block of
peak resurgent currents during either phase 3 of the action potential or
during the declining ramp were calculated (cf. (Milnes et al., 2006; Ridley,
Witchel, & Hancox, 2006; Alexandrou et al., 2006). During Purkinje action
potential commands the low plateau and shallow repolarisation of phase
3 does not produce prominent resurgent currents but a more bow-shaped
current profile (cf. (Lu et al., 2001; Milnes et al., 2006)). Block of peak
current during phase 3 of the PAP was calculated relative to the final
control current measurement (Milnes et al., 2006).
For each drug, concentration–response curves for the various
stimulus protocols were constructed using a range of drug concentra-
tions spanning 2.5 to 4 Log units. Each curve (derived from the data
using Eqs. (1) and (2)) was constructed using 5–8 different points
with 4–11 cells at each concentration. Isochronal concentration–
response curves were constructed to minimise and control for any
effect of current run-down. Cumulative concentration–response
experiments were not performed, with one cell typically being
exposed to one concentration of drug only. Prior to drug application,
and following attainment of whole-cell cell access, cells were subject
to repeated stimulation for 3–6 min to record stable control currents
using the particular voltage waveform under study. Cells were then
superfused with drug-containing solution. During drug superfusion
the appropriate command waveform was repeatedly applied, to
ascertain steady-state inhibition of IhERG for each concentration and
protocol. IhERG block was calculated following 8 or 6 min pacing,
respectively, in dofetilide or cisapride (complete and rapid solution
change around the cell occurred in b1 s; (Levi et al., 1996)). One
exception to this was using the VAP-12 s protocol with dofetilide. Due
to the very slow block development during this protocol, IhERG block
was calculated following 10 min exposure.
3. Results
3.1. Comparison of IC50 values between different stimulus protocols
Fig. 3 panel A shows the effect of dofetilide (30 nM, Fig. 3Ai) and of
cisapride (20 nM, Fig. 3Aii) on IhERG elicited by either a 2-step (Std
+ 20 mV) or VAP (S-to-S = 12 s) voltage command protocol. For
dofetilide, substantial IhERG inhibition was observed with both stimulus
protocols (Fig. 3Ai); however, for cisapride the extent of observed
inhibition was markedly greater for the 2-step protocol than with the VAP
waveform (Fig. 3Aii). Similar experiments were performed using each of
the six stimulus protocols investigated and concentration―response
relations were constructed as described in the ‘Methods’ (Section 2.6).
Fig. 3Bi shows plots of the IC50 values obtained for each drug with the
different stimulus protocols, with a more marked variation in the values
observed for cisapride (ranging from 7 to 72 nM) than for dofetilide
(ranging from 4 to 15 nM). IhERG IC50 data for the two drugs are re-drawn
as a scatter plot in Fig. 3Bii, on which are indicated mean and median IC50
values (long intermittent and solid lines, respectively), along with the
25th and 75th percentiles, (denoted by the shorter solid lines, as an
indication of variance). From Fig. 3Bi and Bii it is clear that cisapride
exhibited a much greater variation in IhERG IC50 with the different
protocols than did dofetilide, with protocols employing AP waveforms
yielding higher IC50 values. For each drug the IC50 values with the different
protocols were compared using a one-way ANOVA (P b 0.002 for
dofetilide and P b 0.0001 for cisapride), followed by a Bonferroni post-
test, the results of which are shown in Fig. 3C. Thus, statistically significant
variation of IC50 was observed for both drugs, but with a much wider
range of observed IC50 values for cisapride than for dofetilide.
Bi-exponential fitting (Eq. (7)) of tail current decay at − 40 mV
following “Std + 20 mV” protocols showed no significant effect of
either dofetilide or cisapride on fast or slow time constants nor on the
relative contributions of IhERG decay fitted by the τSlow or τFast when
compared with control (one-way ANOVA with a Bonferroni post-test;
data not shown). Subsequent experiments were performed with a
range of conventional (voltage step) protocols in order to identify
features of observed inhibition (and thereby protocols) that clearly
distinguished between the two drugs.
3.2. Voltage dependence of IhERG inhibition by dofetilide and cisapride
Current–voltage (I–V) protocols such as that shown as an inset to
Fig. 4 (which is identical to that used in previous studies from our
laboratory, (eg (Milnes et al., 2003, 2006; Ridley et al., 2006)) are
typically employed to compare the effects of drugs on the voltage
dependence of IhERG activation and to quantify the extent of inhibition
over a range of test potentials (in this case between −40 mV and
+50 mV). In the present study, on achieving the whole-cell configu-
ration this protocol was applied repeatedly every 3 min. Current
characteristics were stable by the fourth application of the protocol,
which was taken as the ‘Control’ for comparison with currents recorded
in the subsequent presence of drug. Cells were then superfused with
Tyrode's solution containing either dofetilide or cisapride and were
equilibrated in drug for N6 min while being held at −80 mV. The
protocol was then re-applied to record the currents in drug (“1st drug”).
Cells were then repeatedly paced to attain steady-state inhibition and
the I–V protocol applied again to establish steady-state current
recordings in the presence of drug (steady-state). Tail current
measurements were made relative to the instantaneous leak current
at −40 mV and fractional block and voltage dependence of activation
calculated using Eqs. (1) and (3) (Methods).
Fig. 4 panel A shows mean (±SEM) normalised IhERG tail amplitude
plotted as a function of step-depolarisation in control and at steady-state
in each drug and fitted with Eq. (3) (Fig. 4Ai for dofetilide; Fig. 4Aii for
cisapride). For dofetilide the V1/2 values derived from the fit to the data
were −20.4±0.8 mV and −22.9± 0.8 mV in control and dofetilide
respectively (t-test, Pb 0.05; n=6). For cisapride the V1/2 values derived
from the fit to the data and −19.9±0.6 mV and −23.1±1.0 mV in
control and cisapride respectively (t-test, Pb 0.05; n=6). Neither drug
altered significantly the slope factor for the fitted relations (see legend of
Fig. 4 for values). The small and similar effect of both drugs suggests that
effects on the voltage dependence of IhERG activation cannot account for
greater variation with protocol of cisapride block compared to that of
dofetilide.
Fractional block of IhERG tails was calculated at each potential in the
presence of drug during either the first protocol application (1st drug)
or at steady-state (SS) and is shown plotted in panel B (Fig. 4Bi for
dofetilide, Fig. 4Bii for cisapride). The inhibition of IhERG by both drugs
displayed significant voltage dependence during the first application
 J.T. Milnes et al. / Journal of Pharmacological and Toxicological Methods 61 (2010) 178–191 183

Fig. 3. Dofetilide and cisapride block of IhERG with different voltage protocols. (A) Effect of stimulus profile (lower panel) on block of IhERG. Current traces (upper panel) in the absence
and presence of either 30 nM dofetilide (Ai) or 20 nM cisapride (Aii). hERG currents were elicited in response to either a 2-step protocol or a VAP applied with a S-to-S of 12 and 2 s
respectively. (Bi) Summary of IhERG IC50 (mean ± SEM. 4-11 cells at each concentration; data points where n = 4 were outlining points not on the steepest portion of the curve) for
dofetilide and cisapride derived using 6 different stimulus protocols. (Bii) Scatter plot of hERG IC50s for dofetilide and cisapride derived from Bi. Median and mean of IC50s are
represented as long solid and dashed lines, respectively. Data variance is represented as the 25th and 75th percentile shown as the short-solid lines above and below the median.
(C) Summary Bonferroni post-test performed on data in panel Bi. N.S denotes non-significant. P b 0.05, 0.01, and 0.001 are shown as *, ** and *** respectively. Note that when Std + 20 mV
and −20 mV (square pulse) protocols were compared separately P b 0.05 (t-test).

of the protocol, with block increasing with increasing membrane
depolarisation (one-way ANOVA P b 0.05 and P b 0.0001 for dofetilide
and cisapride, respectively). This is consistent with the known gating-
dependence of IhERG blockade by both drugs (Kiehn, Lacerda, Wible, &
Brown, 1996; Walker et al., 1999; Mitcheson, Chen, Lin, et al., 2000).
However, at steady-state dofetilide block of IhERG was not voltage-
dependent (one-way ANOVA with Bonferroni post-test, N.S), but that
of cisapride remained significantly voltage-dependent (one-way
ANOVA with a Bonferroni post-test P b 0.05).
3.3. Effects of dofetilide on IhERG availability and inactivation time-course
The voltage dependence of IhERG availability and the time-course of
inactivation were investigated using “triple-pulse” voltage commands
184 J.T. Milnes et al. / Journal of Pharmacological and Toxicological Methods 61 (2010) 178–191

Fig. 4. Voltage dependence of IhERG block by dofetilide and cisapride. (A) Normalised current–voltage (I–V) relationships for IhERG tails in the absence and presence of dofetilide (Ai)
or cisapride (Aii). Protocol shown in inset (S-to-S = 12 s). Fits to the data yielded the following parameters: V1/2 = − 20.4 ± 0.8 mV and − 22.9 ± 0.8 mV in control and dofetilide
respectively, t-test P b 0.05, n = 6; and − 19.9 ± 0.6 mV and − 23.1 ± 1.0 mV in control and cisapride respectively, t-test P b 0.05 n = 6) and slope factor, k = 6.8 ± 0.7 v 6.2 ± 0.7 in
control and dofetilide respectively; t-test P N 0.7; and 6.7 ± 0.6 and 7.1 ± 0.9 in control and cisapride respectively, t-test N.S). (B) shows steady-state fractional block data for dofetilide
(Bi) and cisapride (Bii) plotted as a function of step potential. Fractional block data (mean ± SEM) of IhERG tails are shown for the first I–V protocol applied following equilibration in
drug while holding at −80 mV and following a period of pacing to achieve steady-state.

as shown in the insets above Fig. 5Ai and Bi. In order to investigate
voltage dependence of steady-state inactivation/availability, the
protocol shown above Fig. 5A (cf (Milnes et al., 2003; McPate,
Duncan, Milnes, Witchel, & Hancox, 2005)) was applied in the absence
and presence of each drug. Following an initial depolarisation to
+40 mV to activate and inactivate IhERG, the membrane potential was
then stepped briefly to a range (‘ladder’) of different potentials (to
enable recovery from inactivation), followed by a second depolarisa-
tion to +40 mV. The magnitude of outward IhERG evoked by the
second depolarisation to + 40 mV was then measured and normalised
to the maximal current observed during the protocol (cf (Milnes et al.,
2003; McPate et al., 2005)). Mean normalised current data are shown
(plotted against ‘ladder’ voltage) in Fig. 5 Ai and Aii for dofetilide and
cisapride, respectively. Data were fitted with Eq. (4) to obtain V1/2 and
k values (see Fig. 5 legend). Neither drug had a statistically significant
effect on the voltage dependence of hERG availability under our
experimental conditions.
In order to determine the effect of each drug on the time-course of
IhERG inactivation the cell membrane potential was stepped from
−80 mV to +40 mV (500 ms) to activate/inactivate fully IhERG prior
to a brief (2 ms) repolarisation to −100 mV to relieve fully inactivation,
with nominal deactivation observed. Membrane potential was then
stepped to a range of voltages to evoke rapidly inactivating transient
IhERG (−40 to +40 mV, see Fig. 5 panel B inset for protocol). Inactivation
time constants (τ ) were derived by fitting the rapidly inactivating
currents with Eq. (6). Measurements were first made in control and then
in the presence of drug. Mean (±SEM) τ values derived from these
experiments are plotted as a function of membrane potential (during
the third step of the protocol) in Fig. 5Bi and Bii for dofetilide and
cisapride respectively. Both drugs produced an overall significant effect
of decreasing the τ of inactivation over the voltage range studied
(P b 0.0001); however, post-hoc pair-wise testing of values at each test
potential (Bonferroni post-test) only yielded significant differences for
cisapride (Fig. 5Bii) at −10 and 0 mV. Considered collectively, the data
in Fig. 5 suggest that the two drugs had broadly similar effects and that
the differences in protocol-dependence of IhERG IC50 observed in Fig. 3
were not paralleled by differential effects of the two drugs on IhERG
availability or inactivation time-course.
3.4. Time-dependence of IhERG blockade on membrane depolarisation
Development of IhERG inhibition with time following channel
gating initiated by membrane depolarisation can be investigated
using either/both sustained membrane depolarisation and ‘envelope
of tails’ protocols (eg. (Walker et al., 1999; Milnes et al., 2003; Ridley
et al., 2004)). Time-dependent development of IhERG inhibition was
 J.T. Milnes et al. / Journal of Pharmacological and Toxicological Methods 61 (2010) 178–191 185

Fig. 5. Effect of dofetilide and cisapride on IhERG availability and time-course of IhERG inactivation. (A) Mean data from experiments to investigate steady-state voltage dependence of IhERG
availability in the absence and presence of dofetilide (Ai) and cisapride (Aii). Voltage protocol (start-to-start interval between successive sweeps of 6 s) is shown inset above (Ai) Fits to the
data yielded the following parameters: V1/2 = control −55.4± 0.7 and dofetilide −52.0± 0.80 mV, n = 10; t-test N.S and control −54.2 ± 0.7 and cisapride −55.0± 0.82 mV, n = 5; N.S)
and a slope, k, (−19.5± 0.6 in control −20.2± 0.8 in dofetilide, n = 10; N.S and −19.5± 0.6 in control −19.8± 0.8 in cisapride, n = 5; t-test N.S). (B) shows mean data from experiments to
investigate the effect of dofetilide (Bi) and cisapride (Bii) on the time-course of IhERG inactivation. Voltage protocol is shown inset above (start-to-start interval of 6 s). Rapidly inactivating
IhERG currents elicited on depolarisation to a range of potentials were fitted with a mono-exponential function yielding time constant (τ). Mean τ data are plotted as a function of membrane
potential. Data were compared using a two-way repeated measure ANOVA revealed a significant difference between control vs dofetilide (P b 0.001; n = 5) and control vs cisapride
(P b 0.0001; n = 6). # denotes P b 0.05 comparing τ values in control and drug with Bonferroni post-test.

assessed here using a 10 s depolarisation to 0 mV from −80 mV. The
protocol was applied first in the absence of drug and then again (repeated
at a S-to-S interval of 25 s) following drug equilibration, in the absence of
pulsing, in either dofetilide (60 nM) or cisapride (10 nM). Panel A of Fig. 6
shows representative current traces from these experiments. The ‘control’
and the first and fifth trace recorded in the presence of each drug are
shown (Fig. 6Ai — dofetilide; Fig. 6Aii — cisapride). Mean (±SEM)
fractional block was calculated at 100 ms intervals throughout the
depolarisation and the results of this are shown in panels 6Bi and 6Bii.
Block of IhERG by both drugs started at close to zero and developed
progressively with channel gating. Fractional block data were fitted with a
mono-exponential function. With the first depolarisation in the presence
of drug, IhERG developed with similar time-course for both drugs despite
the concentration of dofetilide used (60 nM) being ∼15-fold the IC50 (Std
+20 mV protocol), whereas cisapride was used at a concentration
(10 nM) close to the IC50 (Std+20 mV protocol). This is consistent with a
slower on-rate for dofetilide than for cisapride (the rate-constant values
derived from the fit to the mean data shown in Fig. 6B were: dofetilide — K
of 0.33 ± 0.02 s− 1, τ ∼ 3.0 s, n = 8; cisapride K = 0.47 ± 0.04 s−1,
τ∼2.1 s; n=7). A further four consecutive traces were recorded in the
presence of each drug; fractional block for the second and fifth sweeps
were calculated and are shown. In dofetilide the second protocol
application in drug resulted in a large degree of initial block at the start
with further small amount of block developing with time (Fig. 6Ai, Bi). By
contrast, in cisapride IhERG inhibition during the second application of the
protocol developed with a time-dependence that was quite similar to that
during the first pulse, also achieving similar levels of steady-state block at
the end of the 10 s voltage step (Fig. 6Aii, Bii). A mono-exponential fit to
development of IhERG inhibition by cisapride during the second
depolarisiong step gave a K value of 0.63±0.05 s−1 (τ=1.6 s). In
dofetilide, blockade during the fifth protocol application (5th sweep)
displayed little time-dependence with the exception of the initial few
hundred milliseconds (the current visible during the initial phase of
‘sweep 5’ could represent unblocked IhERG, or an endogenous transient
outward current present in the HEK 293 cells (Jiang, Sun, Cao, & Wang,
2002)). In contrast, for cisapride IhERG block still displayed marked time-
dependence throughout the entire protocol (see Fig. 6Aii and Bii). These
data suggest that, whilst inhibition of IhERG by both dofetilide and
cisapride is contingent upon channel gating, marked relief of inhibition
could occur for cisapride, but not dofetilide, between successive protocol
186 J.T. Milnes et al. / Journal of Pharmacological and Toxicological Methods 61 (2010) 178–191

Fig. 6. Time-dependent inhibition of IhERG by dofetilide and cisapride during a sustained depolarisation. (A) shows typical current records evoked in response to a sustained depolarization
from −80 mV to 0 for 10 s in the absence and the first recorded tracing following equilibration in drug (N 8 and 6 min for dofetilide (Ai) and cisapride (Aii) respectively). During
equilibration membrane potential was held at −80 mV. Following recording of the 1st current trace in drug a further 4 depolarizations were applied with a S-to-S= 25 s in the presence of
drug; the first and fifth recording traces are shown. (B) shows mean (±SEM) fractional block data from experiments identical to that shown in panel A calculated at 100 ms intervals and
plotted as a function of time (n = 7 and 5 for dofetilide and cisapride respectively). Time zero corresponds with depolarisation from −80 mV to 0. Mean fractional block data from the first
(1st) recording in drug were fitted with a mono-exponential association equation to give rate/time constant values given in the ‘Results’ text.

applications. This is consistent with the known propensity for methan-
sulphonanilides to become trapped in the IKr/hERG channel (Mitcheson,
Chen, & Sanguinetti, 2000; Carmeliet, 1992; Kamiya et al., 2006) and with
dissociation properties of cisapride observed at ambient temperature
reported from experiments employing hERG expression in Xenopus
oocytes (Stork et al., 2007).
Considered collectively, the preceding data are consistent with a
scheme in which IhERG inhibition by cisapride, but not dofetilide may
be relieved at negative membrane potentials under our recording
conditions. Further voltage-step protocols were devised to test this
proposition and are discussed below.
3.5. Protocols to examine relief of IhERG inhibition in the continued
presence of dofetilide and cisapride
The first protocol that we employed to investigate relief/recovery
from block in the maintained presence of dofetilide or cisapride is
shown schematically as an inset to Fig. 7. Briefly, membrane potential
was held at -100 mV to maintain channels in a closed conformation and
to facilitate rapid IhERG deactivation following test commands. A
conditioning train of depolarising pulses (10 pulses of 500 ms in
duration from −100 to +20 mV, frequency of 1 Hz) was applied to
attain a stable degree of IhERG inhibition by partial blocking concentra-
tions of dofetilide or cisapride (3 nM and 10 nM respectively). Following
the conditioning train, membrane potential was then held at −100 mV
for varying periods of time, before a test pulse (depolarisation to
+20 mV, repolarising step to -40 mV) to evoke IhERG tails. A voltage of
−100 mV would be anticipated to be sufficient for hERG channels to be
held in a resting/closed state; differing periods of time at this voltage
would therefore maintain channels in non-gated (i.e. resting) condi-
tions for differing extents of time. Fig. 7A shows representative current
records during the test pulse in the presence of each drug (conditioning
train not shown). As shown in Fig. 7Ai the IhERG tail amplitude in the
maintained presence of dofetilide did not change greatly with
progressively greater time spent at −100 mV before application of
the test pulse. By contrast, IhERG tail amplitude in the presence of
cisapride increased with progressively longer periods of time spent at
−100 mV (Fig. 7Aii). Mean IhERG tail amplitude (±SEM; normalised to
maximal tail amplitude during the sampling protocol) was plotted as a
function of time spent at −100 mV for each of dofetilide (Fig. 7Bi) and
cisapride (Fig. 7Bii). Plotted on the same axes are control data from cells
not exposed to drug. Normalised control tail current amplitudes and
those in the presence of dofetilide were similar, irrespective of the
duration of time spent at −100 mV. On the other hand, in the presence
of cisapride, IhERG tail amplitude increased with increasing time at
−100 mV, (i.e. fractional inhibition of IhERG changed by N20% over 8.5 s
at −100 mV). This time-dependent increase was described by a mono-
exponential function with a τ ∼ 5.8 s. The data in Fig. 7 point towards an
ability of cisapride but not dofetilide to dissociate from hERG channels
with increasing time spent at a negative holding voltage.
The respective ability of the two drugs to associate/dissociate from
the hERG channel on a dynamic pulse-by-pulse basis was investigated
further using the approach shown in Fig. 8. In these experiments
depolarising voltage commands, of either 500 ms or 2 s duration, were
applied from a holding potential of − 80 mV to + 20 mV, with
repolarising steps to −40 mV in order to observe IhERG tails. Shorter
and longer duration test pulses were alternated in the maintained
presence of drug. Start-to-start times for alternate pulses were adjusted
to maintain a similar period (6 s) at −80 mV between successive test
commands. Fig. 8Ai shows representative, concatenated traces at steady-
 J.T. Milnes et al. / Journal of Pharmacological and Toxicological Methods 61 (2010) 178–191 187

Fig. 7. Relief of IhERG blockade in maintained presence of drugs. A) Representative current traces recorded in the continued presence of either 3 nm dofetilide (Ai) or 10 nM cisapride
(Aii). A schematic diagram of a voltage command waveform applied to cells in the constant presence of drug is shown the inset. Ai shows representative IhERG traces elicited in the
presence of dofetilide in response to the protocol shown. Note the conditioning pulses are not shown. Dashed line indicates maximal observed tail current amplitude, showing that in
the continuing presence of dofetilide, IhERG tail current amplitude is relatively constant and independent of the duration of time spent at −100 mV prior to test pulse application. Aii
shows similar records to Ai, for cisapride In contrast, in the presence of cisapride (Aii), IhERG tail currents can be seen to increase with greater time spent at −100 mV. B) shows
summarised mean IhERG tail data (± SEM) normalised to the maximum tail current evoked in the presence of drug. Control data, recorded in the absence of drug (n = 4 cells)
exhibited a consistent magnitude of normalised tail current amplitude throughout the protocol (■). Similarly, in the presence of 3 nM dofetilide (Bi, ▲, n = 8 cells), normalised tail
current amplitude remained constant and cannot clearly be differentiated from control data. However, in the presence of 10 nM cisapride (Bii, ▲, n = 4 cells), normalised tail current
amplitude can be seen to increase as the duration at − 100 mV increases. Data were fitted with a mono-exponential fit (black solid line) yielding a K of 0.18 s− 1.

state in the presence of dofetilide, whilst Fig. 8Aii shows traces in the
presence of cisapride. For dofetilide, tail current amplitudes in the
presence of drug were similar between alternate short and long pulses.
However, it is evident that for cisapride, tail current amplitude was
smaller following 2 s than following 500 ms depolarising pulses to
+20 mV, and that this could be seen on an alternating, pulse-by-pulse
basis. Mean normalised data for dofetilide are shown in Fig. 8Bi and for
cisapride in Fig. 8Bii; in these plots short 500 ms duration pulses are
represented on the abscissa as odd number pulses and longer 2 s pulses
are shown as even number pulses. Control data were also recorded in the
absence of drug and alternating pulses under control conditions
produced almost identical tail current amplitudes (see Fig. 8Bi and Bii).
The pattern for IhERG tails recorded in the presence of dofetilide (3 nM or
6 nM) did not differ greatly from that in control. In contrast, however, in
the presence of cisapride IhERG tail amplitude displayed a marked
periodicity; with each cycle IhERG tails varied by about ∼30% between
successively alternating depolarising pulses. Dynamic, pulse-by-pulse
variation in the extent of observed IhERG inhibition by cisapride could be
exacerbated by reducing the duration of the shorter depolarising step
from 500 ms to 100 ms and by varying S-to-S intervals between
successive voltage commands (data not shown). The results with this
approach support further the notion that cisapride block could be
relieved in-between pulses, also providing evidence that pulse duration
was able to determine the extent to which block redeveloped on
depolarisation.
4. Discussion
4.1. Results in context
As a result of observations made in experiments performed using
mammalian (HEK 293) cells, two previous reports in this journal have
highlighted that the IhERG-blocking potency of some drugs varies with
voltage stimulus protocol and experimental temperature (Kirsch et al.,
188 J.T. Milnes et al. / Journal of Pharmacological and Toxicological Methods 61 (2010) 178–191

Fig. 8. Effect of alternating duration of depolarising command duration on IhERG tail block. A) representative consecutively recorded concatenated IhERG traces (upper traces) in
response to the voltage command shown (lower panel) applied in the maintained presence of either 6 nm dofetilide (Ai) or 10 nM cisapride (Aii). Alternating command waveforms
depolarised the cell to + 20 mV for either 500 or 2000 ms. Start-to-Start interval was set to allow a constant inter-pulse duration at −80 mV of 6 s. Ai) in the presence of dofetilide
IhERG tail amplitude remained constant irrespective of pulse duration. Intermittent line is drawn to facilitate comparison of IhERG tail amplitude. Aii) in the presence of cisapride tail
current amplitude varied greatly with depolarisation duration; arrows indicate cyclic variation in tail current amplitude in response to voltage protocol. B) shows normalised IhERG
tail data (± SEM) for experiments shown in A. Short (500 ms) and long (2 s) duration pulses correspond to odd and even pulses numbers respectively. Bi shows data from control
experiments (n = 8 cells) and in the presence of either 3 or 6 nM dofetilide (n = 8 and 11 cells respectively). Bii shows data from control experiments (re-drawn from (Bi)) and in the
presence of 10 nm cisapride (n = 6 cells).

2004; Yao et al., 2005). From investigating a panel of 15 drugs Kirsch et
al concluded that the use of a step–ramp protocol at ‘near physiological’
temperatures yields highly repeatable observations and a conservative
safety evaluation of IhERG inhibition (Kirsch et al., 2004). From studying a
set of five compounds, Yao et al subsequently suggested that the use of a
long-lasting (4.8 s) depolarising voltage-step protocol at room temper-
ature could suffice for drug screening in the majority of cases (Yao et al.,
2005). Neither of these important studies incorporated experiments
using physiological (AP) command waveforms for comparison with
conventional step or step–ramp waveforms.
By contrast, the focus of the present study was not to investigate a large
range of compounds, but rather to study two in detail: first to compare
potency of inhibition between AP and conventional voltage commands;
second, using a range of conventional protocols under identical recording
conditions to identify those which may help act as markers of the relative
protocol-sensitivity of the two chosen agents. Dofetilide was chosen as a
high affinity inhibitor and member of the archetypal selective IKr/hERG-
blocking methanesulphonanilide class, which tend to become trapped in
the channel on closure of the activation gate and consequently dissociate
from the channel only slowly (Carmeliet, 1992; Mitcheson, Chen, &
Sanguinetti, 2000; Stork et al., 2007). Cisapride was chosen as a high
affinity inhibitor for which the literature showed a comparatively wide
range of reported IC50 values (see Fig. 1 and Tables 1 and 2), whilst it is
known to bind to hERG within the channel pore at a site that is largely
similar to that of dofetilide (Lees-Miller et al., 2000; Kamiya et al., 2006,
2008; Mitcheson, Chen, Lin, et al., 2000). Thus, both compounds require
channel opening to access the inner cavity binding site and share as critical
molecular determinants of their binding site amino-acid residues Y652
 J.T. Milnes et al. / Journal of Pharmacological and Toxicological Methods 61 (2010) 178–191
and F656 on the S6 helices of the channel, and T623 and S624 in the pore
helix, differing only in the respects that V625 (pore helix), G648 (S6) and
V659 (S6) are involved in methansulphonanilide binding, but are not
critical for binding of cisapride (Lees-Miller et al., 2000; Mitcheson, Chen,
Lin, et al., 2000; Kamiya et al., 2006, 2008). Despite these similarities
between the two drugs, our experiments with different stimulus protocols
show a more marked effect of waveform on cisapride inhibition of IhERG
than on the action of dofetilide (evidenced by the wider range of observed
IC50 values). Differences between the two drugs could also be detected
with other protocols — the utility of which will be discussed below.
4.2. Comparative variation with protocol for cisapride and dofetilide
A recent study, conducted using the Xenopus oocyte expression
system at room temperature has provided evidence that (in contrast to
methanesulphonanilides) cisapride can dissociate from both open and
closed hERG channels (Stork et al., 2007). Although hERG channel kinetics
are known to differ between ambient and mammalian physiological
temperatures (Vandenberg et al., 2006) and although experiments with
Xenopus oocytes typically underestimate hERG-blocking potency
(Witchel et al., 2002), this observation nevertheless has qualitative
relevance to our own findings using a mammalian expression system at
37 °C. We observed not only significant protocol-dependence of the IC50
for IhERG inhibition by cisapride, but also that in the maintained presence
of drug cisapride was able to dissociate from closed channels at 37 °C. We
also observed a marked ability of the drug to dissociate and reassociate on
a pulse-by-pulse basis under our recording conditions — in marked
contrast with dofetilide. This difference in the ability of the two drugs to
dissociate from/reassociate with hERG channels is highly likely to
contribute significantly to the observed differences in blocking potency
between AP and conventional voltage commands for cisapride: the
comparatively short duration of AP waveforms compared to step or step–
ramp protocols means that access of the drug to gated (open /inactivated)
channel states is less extensive during an AP command than during step
or step–ramp commands. On the other hand, the propensity for dofetilide
to become trapped in the channel (Carmeliet, 1992) means that, in the
steady state, cyclical dissociation/reassociation of dofetilide is much less
likely to occur.
This scheme is also concordant with differences seen when the two
drugs were studied using different protocols. Thus, during a conventional
‘I–V’ protocol (Fig. 4) both drugs initially showed voltage-dependent
inhibition, whilst at steady-state this was the case only for cisapride,
presumably because accumulation of dofetilide block and drug retention
in the channel led at steady-state to an apparent lack of voltage
dependence of inhibition. Similarly, during a sustained depolarisation
(Fig. 6), both drugs showed time-dependent development of inhibition
during the first application of the protocol in drug, but whilst cisapride
still exhibited a largely similar profile by the fifth application of the
protocol, this was not the case for dofetilide. The protocol shown in Fig. 7
demonstrated definitively that under our conditions dofetilide showed
little if any dissociation when channels were held in the closed state in the
maintained presence of drug, whereas cisapride exhibited marked time-
dependent relief of inhibition. Finally, the protocols shown in Fig. 8 using
depolarisations of alternating short and long duration at a constant S-to-S
interval clearly differentiated between dofetilide and cisapride, to the
extent that the degree of inhibition varied on a pulse-by-pulse basis.
Moreover, at a long S-to-S interval cisapride inhibition of IhERG by
alternating short and long duration pulses could be made to vary ∼70%
between successive voltage commands, presumably due to the differing
amounts of time for the drug to associate with hERG channels during the
differing duration depolarising commands.
4.3. Protocol utility
All of the protocols used in this study may have utility in a research
context where mechanistic information regarding drug-hERG inter-
189
action is sought. In particular the protocols described in Figs. 7 and 8
may be of use in the wider study of dynamic drug dissociation/
reassociation in experiments at physiological temperature, whilst the
use of AP voltage clamp may feasibly be of supplemental value to
conventional protocols used to obtain concentration-response data.
On the other hand, in a commercial screening context where
throughput rate is an important consideration, an initial priority may
simply be to determine whether or not an agent can inhibit hERG and,
if so, to establish a concentration–response relationship. In this
regard, the approach recommended by Kirsch et al (Kirsch et al.,
2004) of using a step–ramp protocol at physiological/near physiolog-
ical temperatures – in order to gain a conservative estimate of a drug's
hERG-blocking propensity – is an entirely reasonable one. A question
arises, however, as to what happens in those instances where the
hERG-screen data that emerge are inconsistent with data from other
screens, particularly if hERG-blocking safety margin becomes an issue
(Redfern et al., 2003)? For example, if a novel hERG-blocking drug
were to undergo voltage-dependent cycles of dissociation and
reassociation (as seen here for cisapride), it is possible that their
effects might be over-estimated by standard hERG-blocking assays
depending on the stimulus frequency and “duty cycle”. In such cases,
the use of different/additional stimulus protocols in further hERG
experiments may be warranted (Yao et al., 2005). As demonstrated
here for cisapride, the use of AP command waveforms might in some
instances lead to significantly different IC50 values and it makes
intuitive sense that the use of physiological waveforms may have
some intrinsic value for evaluating drug action. One potential barrier
to the use of AP clamp for pharmacological studies may be a lack of
access to suitable AP waveforms. In this regard, we provide as a
supplement to this article ventricular and Purkinje fibre waveforms
that can readily be used in AP clamp experiments. Whilst it would
clearly be premature on the basis of our data with dofetilide and
cisapride alone to recommend the routine use of AP waveforms in
electrophysiological hERG screens, we would encourage experimen-
tation with AP clamp in hERG assays in order that the extent of its
utility for screening purposes can be established. Also, future direct
comparison may be warranted between IhERG-drug blocking potencies
obtained using AP clamp and those obtained using AP-like (abbrevi-
ated step–ramp (e.g. Gintant, 2000)) commands and/or rectangular
voltage commands with durations approximating those of physio-
logical waveforms. Such comparison may identify conventional
waveforms that yield IhERG-blocking potencies similar to those with
AP commands under a given set of protocol application conditions.
For compounds that transpire to show marked variation in IhERG-
blocking potency between different voltage protocols used to
determine IC50 values, further information may be gained by the use
of additional protocols used here. The use of a sustained-depolarisa-
tion protocol similar to that employed in Fig. 6 of this study, at a long
S-to-S interval (25 s was used here) is comparatively straightforward
and may provide some initial indication of those drugs that can
dissociate from the channel at rest and of those that become trapped,
whilst the protocols employed for Figs. 7 and 8 here will provide more
detailed information. A particular advantage of these protocols may
be their ability quickly to identify drugs that can associate/dissociate
rapidly from hERG channels. It is, however, vital for any such
experiments that voltage clamp is maintained throughout the entire
experiment. Indeed, both the data from our experiments at 37 °C with
cisapride and those obtained at ambient temperature by Stork
and colleagues for cisapride, droperidol, haloperidol and amiodarone
(Stork et al., 2007) raise a potential note of caution for those planar
patch-clamp approaches in which voltage clamp is not maintained
throughout the entire experiment. Thus, for drugs that can dissociate
from resting hERG channels, it is important that the duty cycle
between resting and gated (open/inactivated) states is well-con-
trolled during concentration–response studies and this requires
continuous voltage clamp.
190 J.T. Milnes et al. / Journal of Pharmacological and Toxicological Methods 61 (2010) 178–191
Acknowledgements
This work was sponsored by Pfizer Global Research and Develop-
ment. The authors thank Mrs Lesley Arberry for excellent technical
assistance.
Appendix A. Supplementary information
A supplementary file (containing ventricular and Purkinje fibre
action potential waveforms) associated with this article can be found, in
the online version, at doi:10.1016/j.vascn.2010.02.007. These wave-
forms may be used for AP voltage clamp experiments without restric-
tion, on the sole condition that the source is credited by appropriate
citation of this study.
References
Alexandrou, A. J., Duncan, R. S., Sullivan, A., Hancox, J. C., Leishman, D. J., Witchel, H. J.,
et al. (2006). Mechanism of hERG K+ channel blockade by the fluoroquinolone
antibiotic moxifloxacin. British Journal of Pharmacology, 147, 905−916.
Carmeliet, E. (1992). Voltage- and time-dependent block of the delayed rectifier K+
current in cardiac myocytes by dofetilide. Journal of Pharmacology and Experimental
Therapeutics, 262, 809−817.
Cavero, I., Mestre, M., Guillon, J. M., & Crumb, W. (2000). Drugs that prolong QT interval
as an unwanted effect: Assessing their likelihood of inducing hazardous cardiac
dysrhythmias. Expert Opinion on Pharmacotherapy, 1, 947−973.
Chiu, P. J., Marcoe, K. F., Bounds, S. E., Lin, C. H., Feng, J. J., Lin, A., et al. (2004). Validation
of a [3H]astemizole binding assay in HEK293 cells expressing HERG K+ channels.
Journal of Pharmacological Sciences, 95, 311−319.
Duncan, R. S., McPate, M. J., Ridley, J. M., Gao, Z., James, A. F., Leishman, D. J., et al. (2007).
Inhibition of the HERG potassium channel by the tricyclic antidepressant doxepin.
Biochemical Pharmacology, 74, 425−437.
Fossa, A. A., Wisialowski, T., Wolfgang, E., Wang, E., Avery, M., Raunig, D. L., et al. (2004).
Differential effect of HERG blocking agents on cardiac electrical alternans in the
guinea pig. European Journal of Pharmacology, 486, 209−221.
Friedrichs, G. S., Patmore, L., & Bass, A. (2005). Non-clinical evaluation of ventricular
repolarization (ICH S7B): results of an interim survey of international pharma-
ceutical companies. Journal of Pharmacological and Toxicological Methods, 52, 6−11.
Furuta, S., Kamada, E., Omata, T., Sugimoto, T., Kawabata, Y., Yonezawa, K., et al. (2004).
Drug–drug interactions of Z-338, a novel gastroprokinetic agent, with terfenadine,
comparison with cisapride, and involvement of UGT1A9 and 1A8 in the human
metabolism of Z-338. European Journal of Pharmacology, 497, 223−231.
Gintant, G. A. (2000). Characterization and functional consequences of delayed rectifier
current transient in ventricular repolarization. American Journal of Physiology, 278,
H806−H817.
Gintant, G. A. (2008). Preclinical Torsades-de-Pointes screens: Advantages and
limitations of surrogate and direct approaches in evaluating proarrhythmic risk.
Pharmacology and Therapeutics, 119, 199−209.
Guth, B. D., Germeyer, S., Kolb, W., & Markert, M. (2004). Developing a strategy for the
nonclinical assessment of proarrhythmic risk of pharmaceuticals due to prolonged
ventricular repolarization. Journal of Pharmacological and Toxicological Methods, 49,
159−169.
Hancox, J. C., Levi, A. J., & Witchel, H. J. (1998). Time course and voltage dependence of
expressed HERG current compared with native ‘rapid’ delayed rectifier K current
during the cardiac ventricular action potential. Pflugers Archiv. European Journal of
Physiology, 436, 843−853.
Hancox, J. C., McPate, M. J., El Harchi, A., & Zhang, Y. H. (2008). The hERG potassium
channel and hERG screening for drug-induced torsades de pointes. Pharmacology
and Therapeutics, 119, 118−132.
Haverkamp, W., Breithardt, G., Camm, A. J., Janse, M. J., Rosen, M. R., Antzelevitch, C.,
et al. (2000). Special article: The potential for QT prolongation and pro-arrhythmia
by non-antiarrhythmic drugs: Clinical and regulatory implications. Report on a
policy conference of the European Society of Cardiology. Cardiovascular Research,
47, 219−233.
Henney, J. E. (2000). From the Food and Drug Administration. JAMA, 283, 1131.
Jiang, B., Sun, X., Cao, K., & Wang, R. (2002). Endogenous Kv channels in human embryonic
kidney (HEK-293) cells. Molecular and Cellular Biochemistry, 238, 69−79.
Kamiya, K., Niwa, R., Mitcheson, J. S., & Sanguinetti, M. C. (2006). Molecular
determinants of HERG channel block. Molecular Pharmacology, 69, 1709−1716.
Kamiya, K., Niwa, R., Morishima, M., Honjo, H., & Sanguinetti, M. C. (2008). Molecular
determinants of HERG channel block by terfenadine and cisapride. Journal of
Pharmacological Sciences, 108, 301−307.
Kang, J., Chen, X. L., Wang, H., Ji, J., Reynolds, W., Lim, S., et al. (2004). Cardiac ion
channel effects of tolterodine. Journal of Pharmacology and Experimental Therapeu-
tics, 308, 935−940.
Kiehn, J., Lacerda, A. E., Wible, B., & Brown, A. M. (1996). Molecular physiology and
pharmacology of HERG. Single channel currents and block by dofetilide. Circulation,
94, 2572−2579.
Kirsch, G. E., Trepakova, E. S., Brimecombe, J. C., Sidach, S. S., Erickson, H. D., Kochan, M. C.,
et al. (2004). Variability in the measurement of hERG potassium channel inhibition:
Effects of temperature and stimulus pattern. Journal of Pharmacological and
Toxicological Methods, 50, 93−101.
Kupershmidt, S., Yang, I. C. -H., Hayashi, K., Wei, J., Chanthaphaychith, S., Peteren, C. I.,
et al. (2003). IKr drug response is modulated by KCR1 in transfected cardiac and
noncardiac cells. FASEB Journal, 17, 2263−2265.
Lees-Miller, J. P., Duan, Y., Teng, G. Q., & Duff, H. J. (2000). Molecular determinant of high
affinity dofetilide binding to HERG1 expressed in Xenopus oocytes: Involvement of
S6 sites. Molecular Pharmacology, 57, 367−374.
Levi, A. J., Hancox, J. C., Howarth, F. C., Croker, J., & Vinnicombe, J. (1996). A method for
making rapid changes of superfusate whilst maintaining temperature at 37 °C.
Pflugers Archiv. European Journal of Physiology, 432, 930−937.
Limberis, J. T., Su, Z., Cox, B. F., Gintant, G. A., & Martin, R. L. (2006). Altering extracellular
potassium concentration does not modulate drug block of human ether-a-go-go-
related gene (hERG) channels. Clinical and Experimental Pharmacology and
Physiology, 33, 1059−1065.
Lin, J., Guo, J., Gang, H., Wojciechowski, P., Wigle, J. T., & Zhang, S. (2005). Intracellular K+ is
required for inactivation-induced high affinity binding of cisapride to HERG channels.
Molecular Pharmacology, 68, 855−865.
Lu, Y., Mahaut Smith, M. P., Varghese, A., Huang, C. L. H., Kemp, P. R., & Vandenberg, J. I.
(2001). Effects of premature stimulation on HERG channels. Journal of Physiology,
537(3), 843−851.
Martin, R. L., McDermott, J. S., Salmen, H. J., Palmatier, J., Cox, B. F., & Gintant, G. A. (2004). The
utility of hERG and repolarization assays in evaluating delayed cardiac repolarization:
influence of multi-channel block. Journal of Cardiovascular Pharmacology, 43, 369−379.
McPate, M. J., Duncan, R. S., Hancox, J. C., & Witchel, H. J. (2008). Pharmacology of the short QT
syndrome N588K–hERG K+ channel mutation: Differential impact on selected class I and
class III antiarrhythmic drugs. British Journal of Pharmacology, 155, 957−966.
McPate, M. J., Duncan, R. S., Milnes, J. T., Witchel, H. J., & Hancox, J. C. (2005). The
N588K–HERG K+ channel mutation in the ‘short QT syndrome’: Mechanism of
gain-in-function determined at 37 °C.. Biochemical and Biophysical Research
Communications, 334, 441−449.
Milnes, J. T., Crociani, O., Arcangeli, A., Hancox, J. C., & Witchel, H. J. (2003). Blockade of
HERG potassium currents by fluvoxamine: Incomplete attenuation by S6 mutations
at F656 or Y652. British Journal of Pharmacology, 139, 887−898.
Milnes, J. T., Witchel, H. J., Leaney, J. L., Leishman, D. J., & Hancox, J. C. (2006). hERG K+ channel
blockade by the antipsychotic drug thioridazine: An obligatory role for the S6 helix residue
F656. Biochemical and Biophysical Research Communications, 351, 273−280.
Mitcheson, J. S., Chen, J., Lin, M., Culberson, C., & Sanguinetti, M. C. (2000). A structural
basis for drug-induced long QT syndrome. Proceedings of the National Academy of
Sciences of the United States of America, 97, 12329−12333.
Mitcheson, J. S., Chen, J., & Sanguinetti, M. C. (2000). Trapping of a methanesufonanlide
by closure of the HERG potassium channel activation gate. Journal of General
Physiology, 115, 229−240.
Mohammad, S., Zhou, Z., Gong, Q., & January, C. T. (1997). Blockage of the HERG human
cardiac K+ channel by the gastrointestinal prokinetic agent cisapride. American
Journal of Physiology, 273, H2534−H2538.
Perrin, M. J., Kuchel, P. W., Campbell, T. J., & Vandenberg, J. I. (2008). Drug binding to the
inactivated state is necessary but not sufficient for high affinity binding to human
ether-a-go-go-related gene channels. Molecular Pharmacology, 74, 1443−1452.
Potet, F., Bouyssou, T., Escande, D., & Baró, I. (2001). Gastrointestinal prokinetic drugs
have different affinity for the human cardiac human ether-a-go-go K+ Channel.
Journal of Pharmacology and Experimental Therapeutics, 299, 1007−1012.
Pugsley, M. K., Hancox, J. C., & Curtis, M. J. (2008). Perception of validity of clinical and
preclinical methods for assessment of torsades de pointes liability. Pharmacology
and Therapeutics, 119, 115−117.
Rampe, D., Murawsky, M. K., Grau, J., & Lewis, E. W. (1998). The antipsychotic agent
sertindole is a high affinity antagonist of the human cardiac potassium channel
HERG. Journal of Pharmacology and Experimental Therapeutics, 286, 788−793.
Rampe, D., Roy, M. L., Dennis, A., & Brown, A. M. (1997). A mechanism for the
proarrhythmic effects of cisapride (Propulsid): High affinity blockade of the human
cardiac potassium channel HERG. FEBS Letters, 417, 28−32.
Redfern, W. S., Carlsson, L., Davis, A. S., Lynch, W. G., MacKenzie, I., Palethorpe, S., et al.
(2003). Relationships between preclinical cardiac electrophysiology, clinical QT
interval prolongation and torsade de pointes for a broad range of drugs: evidence
for a provisional safety margin in drug development. Cardiovascular Research, 58,
32−45.
Rezazadeh, S., Hesketh, J. C., & Fedida, D. (2004). Rb+ flux through hERG channels
affects the potency of channel blocking drugs: Correlation with data obtained using
a high-throughput Rb+ efflux assay. Journal of Biomolecular Screening, 9, 588−597.
Ridley, J. M., Dooley, P. C., Milnes, J. T., Witchel, H. J., & Hancox, J. C. (2004). Lidoflazine is
a high affinity blocker of the HERG K+channel. Journal of Molecular and Cellular
Cardiology, 36, 701−705.
Ridley, J. M., Witchel, H. J., & Hancox, J. C. (2006). Clemastine, a conventional
antihistamine, is a high potency inhibitor of the HERG K+ channel. Journal of
Molecular and Cellular Cardiology, 40, 107−118.
Shah, R. R. (2005). Drug-induced QT interval prolongation — regulatory guidance and
perspectives on hERG channel studies. Novartis Foundation Symposium, 266, 251−280.
Snyders, D. J., & Chaudhary, A. (1996). High affinity open channel block by dofetilide of
HERG expressed in a human cell line. Molecular Pharmacology, 49, 949−955.
Stork, D., Timin, E. N., Berjukow, S., Huber, C., Hohaus, A., Auer, M., et al. (2007). State
dependent dissociation of HERG channel inhibitors. British Journal of Pharmacology,
151, 1368−1376.
Toga, T., Kohmura, Y., & Kawatsu, R. (2007). The 5-HT4 agonists cisapride, mosapride,
and CJ-033466, a novel potent compound, exhibit different human ether-a-go-go-
related gene (hERG)-blocking activities. Journal of Pharmacological Sciences, 105,
207−210.
 J.T. Milnes et al. / Journal of Pharmacological and Toxicological Methods 61 (2010) 178–191
Vandenberg, J. I., Varghese, A., Lu, Y., Bursill, J. A., Mahaut-Smith, M. P., & Huang, C. L.
(2006). Temperature dependence of human ether-a-go-go-related gene K+currents.
American Journal of Physiology. Cell Physiology, 291, C165−C175.
Walker, B. D., Singleton, C. B., Bursill, J. A., Wyse, K. R., Valenzuela, S. M., Qiu, M. R., et al.
(1999). Inhibition of the human ether-a-go-go related gene (HERG) potassium
channel by cisapride: Affinity for open and inactivated states. British Journal of
Pharmacology, 128, 444−450.
Wang, J., Della, P. K., Wang, H., Karczewski, J., Connolly, T. M., Koblan, K. S., et al. (2003).
Functional and pharmacological properties of canine ERG potassium channels.
American Journal of Physiology, 284, H256−H267.
Weerapura, M., Hebert, T. E., & Nattel, S. (2002). Dofetilide block involves interactions
with open and inactivated states of HERG channels. Pflugers Archiv. European
Journal of Physiology, 443, 520−531.
191
Witchel, H. J., Milnes, J. T., Mitcheson, J. S., & Hancox, J. C. (2002). Troubleshooting problems
with in vitro screening of drugs for QT interval prolongation using K+channels
expressed in mammalian cell lines and Xenopus oocytes. Journal of Pharmacological
and Toxicological Methods, 48, 65−80.
Yao, J. A., Du, X., Lu, D., Baker, R. L., Daharsh, E., & Atterson, P. (2005). Estimation of
potency of HERG channel blockers: impact of voltage protocol and temperature.
Journal of Pharmacological and Toxicological Methods, 52, 146−153.
Yap, Y. G., & Camm, A. J. (2003). Drug induced QT prolongation and torsades de pointes.
Heart, 1363−1372.
Zhou, Z., Gong, Q., Ye, B., Fan, Z., Makielski, J. C., Robertson, G. A., et al. (1998). Properties
of HERG channels stably expressed in HEK 293 cells studied at physiological
temperature. Biophysical Journal, 74, 230−241.