logy 532 (2006) 44–49www.elsevier.com/locate/ejphar

European Journal of Pharmaco

Effects of estradiol on cardiac ion channel currents

Clemens Möller ⁎, Rainer Netzer

Cellular Assays, Evotec AG, Schnackenburgallee 114, D-22525 Hamburg, Germany

Received 13 October 2005; received in revised form 3 January 2006; accepted 10 January 2006Available online 9 February 2006

Abstract

Steroids are known to exert direct and indirect effects on cardiovascular functions, and women have been found to be more susceptible to QTprolongation than men. Although many clinical studies have been performed, the effects of steroids on cardiac repolarization are not yet fullyunderstood. We examined the effects of 17-beta-estradiol (estradiol) on the major cardiac currents that are correlated to clinical observations ofarrhythmias. Effects on the two major currents responsible for repolarization of the cardiac action potential (mediated by the human ether-à-gogorelated gene (HERG) product), and by the potassium channel Q1 (KCNQ1) co-expressed with the potassium channel accessory subunit E1(KCNE1) were examined, as well as effects on the sodium inward current (mediated by the sodium channel 5A (SCN5A) and generating the rapidupstroke of the action potential). A concentration-dependent effect of estradiol on the KCNQ1/KCNE1-mediated potassium current was observed.The half-maximal inhibition concentration (IC50) of estradiol on the KCNQ1/KCNE1 ion channel was calculated to 1.13±0.23 μM. The HERG-mediated potassium and the SCN5A-mediated sodium currents, however, were only slightly reduced by estradiol at concentrations of up to30 μM. This suggests that alterations of the cardiac action potentials by steroids may be mediated by interaction with the KCNQ1/KCNE1ion channel.© 2006 Elsevier B.V. All rights reserved.

Keywords: QT prolongation; HERG; KCNQ1; Patch-clamp; Estradiol

1. Introduction

Torsades de pointes is a type of life-threatening arrhythmia,preceded by a markedly prolonged QT interval in theelectrocardiagram. Women have been found to be moresusceptible to developing this kind of ventricular tachycardiathan men (Makkar et al., 1993; Kawasaki et al., 1995). A role ofsteroids, particularly of estrogens, has therefore been proposedto contribute to the prolonged repolarization phase observed inwomen. However, in spite of a greater sensitivity of women toQT-prolonging drugs (Rodriguez et al., 2001), an effect ofestradiol on the QT interval could not be clinically shown (Hulotet al., 2003; Nowinski et al., 2002). The higher risk of females todevelop torsades de pointes is not yet fully understood.

The action potential of the human heart is governed by theflux of sodium, calcium and potassium ions through a variety ofion channels of different biophysical properties (Netzer et al.,2003). The currents through the human ether-à-gogo related

⁎ Corresponding author. Tel.: +49 40 560 81 438; fax: +49 40 560 81 222.E-mail address: clemens.moeller@evotec.com (C. Möller).

0014-2999/$ - see front matter © 2006 Elsevier B.V. All rights reserved.doi:10.1016/j.ejphar.2006.01.006

gene (HERG) product and the potassium channel Q1 (KCNQ1)co-expressed with potassium channel accessory subunit E1(KCNE1) coded ion channels are of major importance for theduration of the repolarization phase, and thereby for the length ofthe QT interval. While HERG forms the alpha subunit of therapidly activating IKr channel, a coassembly of KCNQ1 andKCNE1 forms the slowly activating IKs channel. Malfunction ofone or both of these ion channels can lead to a delayedrepolarization and thus to a prolonged QT interval. Such aprolonged QT interval has been observed as a congenital form(arising from mutations in KCNQ1 (denoted LQT1), KCNE1(LQT5), or HERG (LQT2)), and as a drug-induced form causedby compounds interacting with one of these channels (reviewedby, e.g., Ackerman and Clapham, 1997; Keating and Sangui-netti, 2001; Roden, 2004). So far, nearly all cases of drug-induced QT prolongation have been correlated to a blockade ofthe HERG potassium channel (reviewed by, e.g., Roden, 2004).It has been discussed in the literature, whether HERG requiresbeta-subunits (like KCNE2) to form the native IKr channel(Abbott et al., 1999). A detailed analysis performed byWeerapura et al. (2002) showed that the biophysical properties

45C. Möller, R. Netzer / European Journal of Pharmacology 532 (2006) 44–49

of HERG were altered when it was co-expressed with KCNE2.These alterations were shown to not affect the physiologicallyrelevant outward current of the ion channel. Also thepharmacological effects of compounds on HERG remainedunaltered when HERG was co-expressed with KCNE2 (Weer-apura et al., 2002).

The SCN5A-mediated sodium current gives rise to the rapidupstroke at the beginning of the action potential. Malfunction ofthe SCN5A ion current inactivation has particularly beenobserved in a congenital form and has also been correlated to aprolonged QT interval (denoted LQT3) and clinical observa-tions of cardiac arrhythmias.

To elucidate gender differences in the cardiac repolarizationphase, we examined the effects of estradiol on the two major ionchannels responsible for terminating the action potential of thehuman heart, and on the SCN5A mediated sodium current.

2. Materials and methods

2.1. Molecular biology

The cDNA coding for the KCNQ1 (GenBank Acc. No.U40990) was cloned into the pcDNA6-vector (Invitrogen,Leek, Netherlands). The cDNAs coding the human ether-à-gogo related gene product (HERG; GenBank Acc. No. U04270)and the KCNE1 (GenBank Acc. No. M26685) were cloned intothe pcDNA3-vector (Invitrogen, Leek, Netherlands). A C-terminal (in case of HERG, KCNE1), or N-terminal (in case ofKCNQ1), epitope-tag was introduced via PCR. Plasmids weresequenced and subsequently introduced into cells using thetransfection reagent DMRIE-C (liposome formulation of thecationic lipid 1,2-dimyristyloxypropyl-3-dimethyl-hydroxyethyl ammonium bromide-cholesterol), according to themanufacturer's instructions (Gibco BRL, Eggenstein, Germany,Cat. No. 10459-014). Stably transfected cells were selected inthe presence of 800 μg/ml G418 (Gibco BRL, Eggenstein,Germany) and 5 μg/ml blasticidin (Invitrogen, Leek, Nether-lands, Cat. No. R210-01). Clonal cell lines were established.Expression of protein was analysed by means of immunoflu-orescence using antibodies directed against the epitope-tag. Thefunctional expression of the ion channels was validatedelectrophysiologically.

cDNA coding for the human SCN5A gene product (SCN5A;GenBank Acc. No. M77235) was cloned into a standard vector.The plasmid was sequenced and subsequently introduced intocells. Electrophysiological recordings were performed approx-imately 20–32 h after transfection.

2.2. Cell culture

Cells were grown at 37 °C and 5% CO2 in 25 ml flasks(Nunc, Roskilde, Denmark, Cat. No. 147589) in 6 ml MEMALPHA Medium (Gibco BRL, Eggenstein, Germany, Cat. No.22571) supplemented with 10% (v/v) heat inactivated fetal calfserum, 1% (v/v) P/S/G-solution (Gibco BRL, Eggenstein,Germany, Cat. No. 10378-016) and G-418 (17 μl/ml medium ofG-418 solution; Gibco BRL, Eggenstein, Germany, 50 mg/ml,

Cat. No. 10131, supplemented with 11% (v/v) 1 M HEPES,pH 7.4).

2.3. Electrophyisology and data analysis

2.3.1. General procedurePatch-clamp experiments were performed in the voltage-

clamp mode (Hamill et al., 1981) and whole-cell currents wererecorded. Extracellular solutions contained: NaCl 130 mM, KCl5.4 mM, CaCl2 1.8 mM, MgCl2 1 mM, Glucose 5 mM, Hepes10 mM adjusted to pH 7.4; intracellular solutions contained:KCl 130 mM, MgCl2 1 mM, Glucose 5 mM, EGTA 1 mM,MgATP 2 mM, Hepes 10 mM adjusted to pH 7.2. 17-beta-Estradiol (estradiol, Sigma-Aldrich, Cat. No. E8875) wasdissolved in DMSO (dimethyl sulfoxide) stock solutions at1000× final concentration, aliquotted, and frozen. For investiga-tions, estradiol stocks were dissolved in extracellular buffer toyield the final concentration, with 0.1% vehicle (DMSO). ForKCNQ1/KCNE1 investigations, the perforated patch configura-tion (Horn andMarty, 1988) with amphotericin B as pore-formingagent was used. The perforated patch configuration wasperformed to maintain intracellular cascades, needed to preventthe current from rundown. Investigations on the HERG and on theSCN5A mediated currents were performed in the conventionalwhole-cell configuration. Patch pipettes were pulled (puller fromScience Products GmbH, Hofheim, Germany) from borosilicateglass tubes (GC 150, Clark Electromedical Instruments, Pang-bourne, UK). Current signals were amplified and digitised by anEPC patch-clamp amplifier (HEKA-Electronics, Lambrecht,Germany), stored and analysed on a personal computer usingthe Pulse/Pulsefit software (HEKA, Lambrecht, Germany).Experiments were conducted at room temperature (21±2 °C).

For investigating effects and reversibility of compounds oncardiac ion channel currents, the following stimulation proto-cols were applied:

2.3.2. Stimulation protocol for the KCNQ1/KCNE1-mediatedcurrent

From a holding potential of −80 mV cells were depolarisedto +40 mV for 1 s (see inset in Fig. 1A). The currentamplitude at the end of the test pulse to +40 mV was used forthe analysis. Currents were leak-current corrected. Stimulationfrequency was 0.1 Hz.

2.3.3. Stimulation protocol for the HERG-mediated currentFrom a holding potential of −80 mV cells were depolarised

for 1 s to +20 mV, followed by a 1 s partial repolarisation backto −40 mV (see inset in Fig. 2A). Tail outward currentamplitudes at −40 mV were analysed. Currents were leak-current corrected. Stimulation frequency was 0.1 Hz.

2.3.4. Stimulation protocol for the SCN5A-mediated currentFrom a holding potential of −80 mV cells were depolarised

to −10 mV for 25 ms (see inset in Fig. 3A). The peak inwardcurrent amplitude at −10 mV and the time constant of thecurrent decay were analysed. A P/4 leak current subtraction wasperformed. Stimulation frequency was 0.33 Hz.

Fig. 2. Effect of 10 μM estradiol on the HERG-mediated potassium current. (A)Two superimposed original HERG current traces recorded in absence (lowertrace) and presence (upper trace) of 10 μM estradiol. Inset: stimulation protocol.(B) Current amplitude is plotted against time. Onset (indicated by the longdashed line) and offset (indicated by the short dashed line) of test compoundapplication time. The extrapolated time course of current amplitude undervehicle conditions is depicted as the solid line.

Fig. 1. Effect of 10 μM estradiol on the KCNQ1/KCNE1-mediated potassiumcurrent. (A) Two superimposed original KCNQ1/KCNE1-current tracesrecorded in absence (upper trace) and presence (lower trace) of 10 μM estradiol.Inset: stimulation protocol. (B) Current amplitude is plotted against time. Onset(indicated by the long dashed line) and offset (indicated by the short dashed line)of test compound application time. The extrapolated time course of currentamplitude under vehicle conditions is depicted as the solid line.

46 C. Möller, R. Netzer / European Journal of Pharmacology 532 (2006) 44–49

2.3.5. Data evaluation and determination of IC50 valueEndogenous rundown of potassium currents was corrected

by calculating an extrapolated time course under vehicleconditions using a biexponential fit of the equation

Y ¼ a� expð−cxÞ þ b� expð−dxÞ ð1Þto the first 20 stimuli.

Effects are expressed as a fraction of the remaining currents,where 1.0 corresponds to no effect, and 0.0 to full currentblockade.

An estimated concentration–response relation was calculat-ed by non-linear least-squares fits of the equation

f ¼ 1=ð1þ ðC=IC50ÞnHÞ ð2Þto the individual data points. The Hill coefficient (nH) and thehalf-maximal inhibiting concentration (IC50) were calculated bythe fitting routine.

3. Results

3.1. Effects of estradiol on the KCNQ1/KCNE1-mediatedpotassium current

KCNQ1/KCNE1 currents were investigated using thedescribed protocol. In Fig. 1, current recordings from a CHO

cell stably transfected with the KCNQ1/KCNE1 ion channel areshown in the absence and in the presence of 10 μM estradiol.Fig. 1A shows the initial stimulus recorded in the absence ofestradiol, and 5 min after addition of 10 μM estradiol andcontinuous stimulation. The time course of the experimentalcurrent amplitude, and the rundown correction (Eq. (1)), isshown in Fig. 1B. A significant current reduction after additionof 10 μM estradiol is observed. The effect of the hormone ispartly reversible during a wash period of 5 min (Fig. 1B), butdid not reach steady state during this period. Concentrations of0.1 μM, 1 μM, and 10 μM were tested (n=3–4 each). Currentinhibitions were (mean±SD): 3±5% (0.1 μM), 50±11%(1 μM) and 87±9% (10 μM). The IC50 value (Eq. (2)) for theinhibition of the KCNQ1/KCNE1-mediated potassium currentwas calculated to 1.13±0.23 μM, and the Hill coefficient to1.02±0.21.

3.2. Effects of estradiol on the HERG-mediated potassiumcurrent

HERG currents were investigated using the describedprotocol. In Fig. 2, current recordings from a CHO cell stably

Fig. 3. Effect of 10 μM estradiol on the SCN5A-mediated sodium current. (A)Two superimposed original SCN5A current traces recorded in absence (lowertrace) and presence (upper trace) of 10 μM estradiol. Inset: stimulation protocol.(B) Current amplitude is plotted against time. Onset (indicated by the longdashed line) and offset (indicated by the short dashed line) of test compoundapplication time.

47C. Möller, R. Netzer / European Journal of Pharmacology 532 (2006) 44–49

transfected with the HERG potassium channel are shown in theabsence and in the presence of 10 μM estradiol. The initialstimulus recorded in the absence of the hormone, and 5 minafter addition of 10 μM estradiol, are depicted in Fig. 2A. The

Fig. 4. Vehicle control experiments performed on KCNQ1/KCNE1, SCN5A, and HE(black) vehicle addition for KCNQ1/KCNE1 (A), SCN5A (B) and HERG (C). (D) Amis plotted against time. Onset (indicated by the long dashed line) and offset (indicat

time course of the current amplitude and the rundown correctionis shown in Fig. 2B. Upon addition of up to 30 μM estradiol,only a minor current reduction of the HERG-mediated outwardcurrent was observed. On average, experiments with addition ofestradiol showed a leak and rundown corrected currentreduction of 12±1% (10 μM, n=3) and 19±6% (30 μM, n=5).

The maximum solubility of estradiol in extracellular bufferwas reached at N30 μM and therefore higher concentrationscould not be investigated.

3.3. Effects of estradiol on the SCN5A-mediated sodiumcurrent

SCN5A currents were investigated using the describedprotocol. In Fig. 3, current recordings from a CHO cell transientlytransfected with the SCN5A sodium channel are shown in theabsence and in the presence of 10 μM estradiol. The initialstimulus recorded in the absence of the hormone, and 5 min afteraddition of 10 μM estradiol, are depicted in Fig. 3A. The timecourse of the current amplitude is shown in Fig. 3B. Uponaddition of up to 30 μM estradiol, only a minor current reductionof the SCN5A-mediated inward current was observed. Onaverage, experiments with addition of estradiol showed a leakand rundown corrected current reduction of 1%±7% (10 μM,n=3) and 7%±2% (30 μM, n=3). No change of the currentinactivation kinetics was observed upon addition of up to 30 μMestradiol (Fig. 3A). As noted above, the maximum solubility ofestradiol in extracellular buffer was reached at N30 μM andtherefore higher concentrations could not be examined.

3.4. Vehicle control experiments

Vehicle control experiments were performed for theKCNQ1/KCNE1 ion channel, the HERG ion channel and the

RG mediated currents. Two superimposed current traces before (grey) and afterplitude of KCNQ1/KCNE1 (△), SCN5A (▽), and HERG (○) mediated currentsed by the short dashed line) of vehicle application.

Fig. 5. Concentration–response relation for the block of the KCNQ1/KCNE1,the HERG, and the SCN5A ion channel by different concentrations of estradiol.Data points are given as mean±SEM from 3 to 5 experiments each and Eq. (1)was fitted to them.

48 C. Möller, R. Netzer / European Journal of Pharmacology 532 (2006) 44–49

SCN5A ion channel under the same conditions as those ofestradiol and resulted in current reductions of 6±11% (n=3,KCNQ1/KCNE1), 1±4% (n=4, HERG), and 5±5% (n=3,SCN5A) (Fig. 4).

3.5. Concentration–response relations for the effect of estradiolon cardiac ion channels

Fig. 5 shows the concentration–response relation for theblock of the KCNQ1/KCNE1 ion channel by 0.1 μM, 1 μM and10 μM, for the HERG ion channel by 1, 3, 10 and 30 μM, andfor the SCN5A sodium channel by 10 μM and 30 μM ofestradiol. Eq. (2) was fitted to the data points and yielded anIC50 of 1.13±0.23 μM, and a Hill-coefficient of 1.02±0.21 forthe inhibition of the KCNQ1/KCNE1 potassium channel byestradiol. Due to only a minor current inhibition uponapplication of up to 30 μM estradiol, no IC50 values for theHERG potassium and the SCN5A sodium channels wereobtained.

4. Conclusions and discussion

Comparing the data recorded in the presence of estradiol andthe vehicle control data, it can be concluded that the potassiumcurrent in CHO cells transfected with the human cardiacKCNQ1/KCNE1 ion channel gene is reduced in a concentra-tion-dependent manner by estradiol, with a significant inhibi-tion upon application of 1 μM and 10 μM. From the datarecorded, the IC50 value is calculated to 1.13 μM. Waldegger etal. (1996) previously observed an effect of the hormone on theslowly activating potassium ion channels expressed afterinjection of KCNE1-RNA into Xenopus oocytes. The directeffect of estradiol on the human cardiac KCNQ1/KCNE1 ionchannel, as observed in the present study, is consistent with thisearlier investigation.

The HERG potassium channel is a notorious off-target evenof some marketed drugs. Interactions of compounds with theHERG channel are responsible for some observations of cardiacside effects. An interesting finding in the present study is thatestradiol affected the human cardiac KCNQ1/KCNE1 ion

channel, but only a minor effect was observed upon applicationof up to 30 μM estradiol on the HERG and on the SCN5Achannel.

Female gender is a risk factor for drug-induced QTprolongation (Rodriguez et al., 2001). Both the KCNQ1/KCNE1 ion channel and the HERG ion channel contribute tothe repolarization reserve of the cardiac action potential,determining the QT interval length. An inhibition ormalfunction of any of these ion channels can lead to QTprolongation. The estradiol plasma level in women haspreviously been found to cover a wide range of 105±34 pMduring menses, and 750±277 pM during the pre-ovulatoryphase (Hulot et al., 2003). Significantly higher steady-stateand peak estradiol plasma concentrations can be reached undertreatment with steroid substitution therapy, and duringpregnancy. The present study suggests a potential directinhibition of the human cardiac KCNQ1/KCNE1 ion channelby estradiol. It can be concluded that at least one member ofthe class of steroids has a potential to affect the actionpotential of the human heart, and that this effect arises from aninteraction of the hormone with the KCNQ1/KCNE1 ionchannel.

The direct physiological consequence of the interaction ofestradiol with KCNQ1/KCNE1 is hard to predict, but it may bespeculated that interaction of estradiol or other steroids with thehuman cardiac KCNQ1/KCNE1 ion channel could lead to agreater sensitivity to drug effects on both the KCNQ1/KCNE1and the HERG ion channels. More importantly, our findingsmay be of importance for preclinical drug discovery anddevelopment projects for the steroid class of compounds. Theinvestigation of compound effects, particularly for steroids, onthe KCNQ1/KCNE1 ion channel could add valuable informa-tion to the cardiac safety profile of a drug, which is typicallyevaluated by focussing on interaction with the HERG ionchannel.

To fully understand how sex hormones affect the actionpotential of the human heart, further studies are required. Theeffects of other hormones (particularly of progesterone) on the ionchannels investigated here need to be addressed. Also the effectsof hormones on other ion channels responsible for the generationof the action potential as well as their effects on ion channelexpression require further investigations.

Acknowledgements

We thank Drs. Dietlind Koch and Andreas Scheel (Evotec)for valuable discussions, Mss. Irene Schlobohm, Anja vonNordheim-Hansen, Heike Deisemann, Sandra Linge, andSandra Riedel (Evotec) for technical assistance and Ms KarenHinson-Rehn (Evotec) for critically reading the manuscript.This project was supported in part by the BMBF (Grant No.0313310C).

References

Abbott, G.W., Sesti, F., Splawski, I., Buck, M.E., Lehmann, M.H., Timothy, K.W., Keating, M.T., Goldstein, S.A., 1999. MiRP1 forms IKr potassium

49C. Möller, R. Netzer / European Journal of Pharmacology 532 (2006) 44–49

channels with HERG and is associated with cardiac arrhythmia. Cell 16;97(2), 175–187.

Ackerman, M.J., Clapham, D.E., 1997. Ion channels-basic science and clinicaldisease. N. Engl. J. Med. 336, 1575–1586.

Hamill, O.P., Marty, A., Neher, E., Sakmann, B., Sigworth, F.J., 1981. Improvedpatch clamp techniques for high-resolution current recording from cells andcell-free membrane-patches. Pflügers Arch. 391, 85–100.

Horn, R., Marty, A., 1988. Muscarinic activation of ionic currents measured by anew whole-cell recording method. J. Gen. Physiol. 92, 145–159.

Hulot, J.S., Demolis, J.L., Riviere, R., Strabach, S., Christin-Maitre, S., Funck-Brentano, C., 2003. Influence of endogenous oestrogens on QT intervalduration. Eur. Heart J. 24, 1663–1667.

Kawasaki, R., Machado, C., Reinoehl, J., Fromm, B., Baga, J.J., Steinman, R.T.,Lehmann, M.H., 1995. Increased propensity of women to develop torsadesde pointes during complete heart block. J. Cardiovasc. Electrophysiol. 6,1032–1038.

Keating, M.T., Sanguinetti, M.C., 2001. Molecular and cellular mechanisms ofcardiac arrhythmias. Cell 104, 569–580.

Makkar, R.R., Fromm, B.S., Steinman, R.T., Meissner, M.D., Lehmann, M.H.,1993. Female gender as a risk factor for torsades de pointes associated withcardiovascular drugs. JAMA 270, 2590–2597.

Netzer, R., Bischoff, U., Ebneth, A., 2003. HTS techniques to investigate thepotential effects of compounds on cardiac ion channels at early-stages ofdrug discovery. Curr. Opin. Drug Discov. Dev. 6, 462–469.

Nowinski, K., Pripp, U., Carlstrom, K., Landgren, B.M., Schenck-Gustafsson,K., Bergfeldt, L., 2002. Repolarization measures and their relation to sexhormones in postmenopausal women with cardiovascular disease receivinghormone replacement therapy. Am. J. Cardiol. 90, 1050–1055.

Roden, D.M., 2004. Drug-induced prolongation of the QT interval. N. Engl. J.Med. 4 (350), 1013–1022.

Rodriguez, I., Kilborn, M.J., Liu, X.K., Pezzullo, J.C., Woosley, R.L., 2001.Drug-induced QT prolongation in women during the menstrual cycle.JAMA 285, 1322–1326.

Waldegger, S., Lang, U., Herzer, T., Suessbrich, H., Binder, K., Lepple-Wienhues, A., Nagl, U., Paulmichl, M., Franz, H.B., Kiesl, L., Lang, F.,Busch, A.E., 1996. Inhibition of minK protein induced K+ channels inXenopus oocytes by estrogens. Naunyn-Schmiedeberg's Arch. Pharmacol.354 (6), 698–702.

Weerapura, M., Nattel, S., Chartier, D., Caballero, R., Hebert, T.E., 2002. Acomparison of currents carried by HERG, with and without coexpression ofMiRP1, and the native rapid delayed rectifier current. Is MiRP1 the missinglink? J. Physiol. 540 (Pt 1), 15–27.