Journal of Molecular and Cellular Cardiology 47 (2009) 656–663

Contents lists available at ScienceDirect

Journal of Molecular and Cellular Cardiology

j ourna l homepage: www.e lsev ie r.com/ locate /y jmcc

Original article

Does small-conductance calcium-activated potassium channel contribute tocardiac repolarization?

Norbert Nagy a, Viktória Szűts a, Zoltán Horváth b, György Seprényi c, Attila S. Farkas a,b, Károly Acsai b,János Prorok a, Miklós Bitay d, Attila Kun a, János Pataricza a, Julius Gy. Papp a,b, Péter P. Nánási e,András Varró a,b, András Tóth a,⁎a Department of Pharmacology and Pharmacotherapy, University of Szeged, Szeged, Hungaryb Division of Cardiovascular Pharmacology, Hungarian Academy of Sciences, University of Szeged, Szeged, Hungaryc Institute of Biology, University of Szeged, Szeged, Hungaryd Department of Cardiac Surgery, University of Szeged, Szeged, Hungarye Department of Physiology, University of Debrecen, Debrecen, Hungary

⁎ Corresponding author. Dóm tér 12, P.O. Box 427 H-6762 545 681; fax: +36 62 544 565.

E-mail address: toth@phcol.szote.u-szeged.hu (A. Tó

0022-2828/$ – see front matter © 2009 Elsevier Inc. Adoi:10.1016/j.yjmcc.2009.07.019

a b s t r a c t

a r t i c l e i n f o

Article history:Received 11 April 2009Received in revised form 19 June 2009Accepted 16 July 2009Available online 24 July 2009

Keywords:SK2Action potentialApaminRepolarization reserveIntracellular calcium

Small-conductance calcium-activated potassium channels (SK channels) have a significant role in neurons.Since they directly integrate calcium handling with repolarization, in heart their role would be particularlyimportant. However, their contribution to cardiac repolarization is still unclear. A previous study reported asignificant lengthening effect of apamin, a selective SK channel inhibitor, on the action potential duration inatrial and ventricular mouse cardiomyocytes and human atrial cells. They concluded that these channelsprovide an important functional link between intracellular calcium handling and action potential kinetics.These findings seriously contradict our studies on cardiac “repolarization reserve”, where we demonstratedthat inhibition of a potassium current is not likely to cause excessive APD lengthening, since its decrease ismostly compensated by a secondary increase in other, unblocked potassium currents. To clarify thiscontradiction, we reinvestigated the role of the SK current in cardiac repolarization, using conventionalmicroelectrode and voltage-clamp techniques in rat and dog atrial and ventricular multicellularpreparations, and in isolated cardiomyocytes. SK2 channel expression was confirmed with immunoblottechnique and confocal microscopy. We found, that while SK2 channels are expressed in the myocardium, afull blockade of these channels by 100 nM apamin – in contrast to the previous report – did not causemeasurable electrophysiological changes in mammalian myocardium, even when the repolarization reservewas blunted. These results clearly demonstrate that in rat, dog and human ventricular cells under normalphysiological conditions – though present – SK2 channels are not active and do not contribute to actionpotential repolarization.

© 2009 Elsevier Inc. All rights reserved.

1. Introduction

Action potential repolarization is controlled by a fine balancebetween various transmembrane ionic currents which are essential todetermine the duration of the cardiac action potential (APD) [1,2]. Dueto this delicate interaction between the underlying currents cardiacaction potential repolarizes with a strong safety margin providing animportant protective mechanism against failure of a specific channelwhich might result in excessive lengthening of repolarization [3]. Sinceexcessive prolongation of the cardiac action potential is potentiallyarrhythmogenic, this strong protective safety mechanism, termedrepolarization reserve, plays a crucial role in stabilizing action potentialduration under a wide range of altered conditions [4,5].

20 Szeged, Hungary. Tel.: +36

th).

ll rights reserved.

Intracellular free calcium ([Ca2+]i) – beyond governing excitation–contraction coupling [6,7] – can also influence action potential durationthrough [Ca2+]i-sensitive ionic currents, such as the sodium-calciumexchanger, the [Ca2+]i-sensitive chloride current, [8,9]. Accordingly,massively elevated [Ca2+]i is often associated with cardiac arrhythmias[10]. The protective role of the repolarization reserve, therefore, is espe-cially important under pathological conditions when the excessivelyprolonged action potential is often combined with elevated [Ca2+]i[11,12].

Calcium-activated potassium channels are abundantly expressed invarious tissues, including brain, peripheral nerve, skeletal muscle,vascular endothelium, liver, kidney, and the heart [13]. These channelsare classified by their conductance: large (BK), [14] intermediate (IK),[15] and small-conductance (SK) [16] channels have been identified. SKchannels showhigh calcium-sensitivity,weak voltage-dependence, andvariable sensitivity to apamin, an inhibitor toxin isolated from beevenom [17–19]. Based on their apamin-sensitivity there are threemajor

657N. Nagy et al. / Journal of Molecular and Cellular Cardiology 47 (2009) 656–663

SK channel subtypes: SK1 channels are little affected, SK2 channels arehighly sensitive and SK3 channels are considered to be intermediate[17–19]. Expression of all subtypes of SK channels were demonstratedin various cardiac preparations including rat, murine and human hearts[17–19], however, their functional role in cardiac repolarization has notbeen extensively studied, and therefore is poorly understood. Based ontheir high calcium-sensitivity and relatively weak voltage-dependence,it can be expected that SK channels provide an important functionallink between calcium handling and electrical events of the surfacemembrane. Thus, these channels might modify action potential confi-guration by responding to beat-to-beat alterations of [Ca2+]i. In aprevious study Xu et al. found that SK channel blockade by apaminsignificantly lengthened APD inmurine atrial and ventricularmyocytes,and also in human atrial cells [17]. Implications of these observationsare very important, because the reported apamin-sensitive markedlengthening of APD suggested a crucial involvement of SK channels infrequency-dependent changes of APD. In addition based on thisobservation that APD lengthening by its inhibition can represent anantiarrhythmic mechanism which could initiate new pharmacologicalstrategies. Furthermore, SK channels can be expected to contribute tothe repolarization reserve – especially during the conditions of calciumoverload, a situation which may occur for example in heart failure.

The main goal of the present study was, therefore, to investigatethe role of the SK current in cardiac repolarization in the dog (aspecies having electrophysiological properties most relevant tohuman), and in the rat (a species having calcium handling and actionpotential characteristics similar to mouse) under physiological condi-tions [20,21]. We also had the opportunity to perform some tests onventricular preparations isolated from a healthy human heart. Weverified the expression of the SK2 channels in canine and rat myo-cardium, but no functional consequences of apamin application wereobserved.

Thus, our present results obtained in multicellular canine, rat andhuman cardiac preparations are in sharp contrast with previousfindings obtained in single myocytes isolated from human and murinehearts, and raised questions regarding the role of SK channels in cardiacrepolarization in general.

2. Materials and methods

2.1. Animal care

Studies were conducted in accordance with the standards of theEuropean Community guidelines on care and use of laboratory animals.All protocols have been approved by the Ethical Committee for Pro-tection of Animals in Research of the University of Szeged, Hungary(permit No. 54/1999 Oej).

2.2. Human patient

Undiseased heart obtained from an organ donor was explanted toobtain pulmonary and aortic valves for transplant surgery. Before car-diac explantation, the donor did not receive medication except furo-semide, dobutamine and plasma expanders. The investigation conformsto the principles outlined in the Declaration of Helsinki and allexperimental protocols were approved by Regional and InstitutionalHumanMedical and Biological ResearchEthics Committee, University ofSzeged, permit No. 717. (No. 63/97).

2.3. Protein isolation and Western blot analysis

Whole-cell tissue lysates were purified from the ventricularmyocardial tissues of adult hearts (Sprague–Dawley rats, n=6, andmongrel dogs, n=6) and also from canine and rat isolated ventricularmyocytes. Briefly, the tissue samples were cut into small pieces in Lysisbuffer (containing: 50 mM Tris–HCl, 1% nonidet P-40, 0.5% deoxycho-

late, 150 mM NaCl, 10 g/l PMSF, 5 μM leupeptin, 5 μM aprotinin, and5 μg/l Na-vanadate, and Protease inhibitor cocktail (Sigma), homoge-nized with polytron and centrifuged (10,000 g, 15 min) at 4 °C. Super-natants were collected and protein concentration was measured byLowry's method using BSA for standard. SDS-polyacrylamide gelelectrophoresis was performed in 10% acrylamide/bis-acrylamide gels.Fractionated proteins were transferred to polyvinylidene difluoride(Immobilon™-P membrane, Millipore) in transfer buffer (containing:25 mM Tris–HCl, 150 mM glycine, 20% methanol, pH=8.3). To avoidnonspecific binding, blots were blocked using TBST with 10% non-fatmilk (BioRad) and target antigens were labeled overnight with primaryantibodies, rabbit polyclonal anti-SK2 (Anti-KCa2.2, Alomone) ormurinemonoclonal α-sarcomeric actin (DAKO) at 4 °C. The binding of theprimary antibody was detected with horse radish peroxidase conju-gated anti-rabbit or anti-mouse secondary antibodies (DAKO), respec-tively, and visualized by enhanced chemoluminescence assay (ECL Pluskit, AmershamPharmacia Biotech). Optical densities of proteinbands onX-ray films were analyzed by using Image J and Excel programs.

2.4. Immunohistochemistry and confocal microscopy

Isolated canine and rat ventricular myocytes were fixed by acetone.Before staining the sampleswere rehydrated in calcium-free phosphatebuffered saline (PBS) and blocked for 120 min with PBST (PBS with0.01% Tween) containing 1% bovine serum albumin at room tempe-rature. Indirect immunofluorescence staining was performed usingrabbit polyclonal anti-SK2 (anti-KCa2.2, Alomone) primary antibody in1:50 dilution and a fluorescent secondary antibody, Alexa Fluor 448conjugated goat anti-rabbit IgG (Molecular Probes Inc) in 1:1000dilution. The 60 min period of incubation with the primary antibody atroom temperature was followed by a further 60 min incubation withthe secondary antibody. Between and after the incubation the sampleswere washed thoroughly with PBST. Control samples were incubatedonly with secondary antibody. For microscopic examination the cellswere mounted in Aqua Poly/Mount (Polysciences Inc.). The fluores-cence images of the immunostained samples were captured by anOlympus FV1000 confocal laser scanning microscope used with stan-dard parameter settings.

2.5. Recording of action potentials in multicellular cardiac preparations

Adult male Sprague–Dawley rats (weighing 150–200 g) and adultmongrel dogs of either sex (weighing 10–20 kg) were anticoagulatedwith sodium-heparin and anaesthesized with 30 mg/kg thiopental.Hearts were rapidly removed through right lateral thoracotomy andimmediately rinsed in ice-cold Krebs–Henseleit solution (containing inmM: NaCl 118.5, KCl 4.0, CaCl2 2.0, MgSO4 1.0, NaH2PO4 1.2, NaHCO3

25.0, and glucose 10.0) for 30 s. The pH of this solution was set to7.35±0.05when saturatedwith amixture of 95%O2 and 5% CO2. Actionpotentials were recorded from papillary muscle preparations excisedfromthe right ventricles of canine, rat, andhumanhearts, aswell as fromleft atrial trabeculae of canine and rat hearts at 37 °C. After excision, thepreparations were immediately mounted in the tissue chamber, havinga volume of 40 ml, and perfused with Krebs–Henseleit solution.

Preparations were continuously stimulated with an electrostimu-lator (Hugo Sachs Elektronik, model 215/II) using 2 ms long constantrectangular voltage pulses delivered through a pair of bipolar platinumelectrodes at a frequency of 1 Hz. Action potentials were recorded withconventionalmicroelectrode techniques. Sharpmicroelectrodes, havinga tip resistance of 10–20 MΩ when were filled with 3 M KCl, wereconnected to the amplifier (Biologic Amplifier, model VF 102). Thevoltage outputs from the amplifier were displayed on a dual beammemory oscilloscope (Tektronix, model 2230), and sampled at 40 kHzusing an analog-to-digital converter (Real TimeDevices Inc., model ADA3300). APD measured at 90% level of repolarization (APD90) wasobtained using a custom-made software (HSE-APES). Each preparation

658 N. Nagy et al. / Journal of Molecular and Cellular Cardiology 47 (2009) 656–663

was allowed for at least 60 min to equilibrate in Krebs–Henseleitsolution. The temperature of the superfusatewas kept constant at 37 °C.After stabilization of action potential parameters control actionpotentials were recorded from the surface region of the tissue samples,then apamin (100 nM) was superfused for 45 min and the measure-ments were repeated in the presence of apamin.

To ensure the physiological conditions of our preparations, controlAPD90 values of 200–220 ms for dogs, and 50–80 ms for rats at 1 Hzwere accepted as exclusion criteria (i.e. preparations having controlAPD90 out of these ranges were discarded). Furthermore, actionpotential amplitude had to be higher than 100 mV in both species.Attempts were made to maintain the same impalement during thewhole experiment. When impalement was dislodged, adjustment ofthe electrode was attempted. If the duration of the action potential ofthe re-established impalement deviated more than 5% from theprevious measurement, the experiment was discarded.

2.6. Isolation of rat and canine ventricular myocytes

Adult male Sprague–Dawley rats, weighing 150–200 g, wereretrogradely perfused at 37 °C with Krebs–Henseleit solution, contain-ing 2mMCaCl2, for 5min. The pH of this solution was set to 7.35±0.05by saturatingwith amixture of 95%O2 and 5%CO2. Then the superfusionwas switched to calcium-free Krebs–Henseleit solution for 8 min.Finally, the perfusate was completed with 0.5 g/l collagenase (Type I,Sigma), 0.5 g/l hyaluronidase, 200 μMCaCl2, and the heartwas perfusedfor a further 7 min with this solution. At the end of the enzymaticdissociation process the right ventricular myocardiumwas minced andgently agitated. Freshly isolated cells were placed in storage-solution(containing in mM: KOH 89.0, glutamate 70.0, taurine 15.0, KCl 30.0,KH2PO4 10.0, HEPES 10.0,MgCl2 0.5, glucose 10.0, EGTA0.5, pH=7.3) atroom temperature before use.

Single canine ventricular cells were obtained from hearts of adultmongrel dogs of either sex using the segment perfusion technique. Theanimals, weighing 10–20 kg, were anaesthetized with i.v. injection of30 mg/kg thiopental. After opening the chest the heart was removedand a segment of the left ventricular wall was perfused through theanterior descending coronary artery using a gravity flow Langendorffapparatus. The perfusatewasmodifiedMEMsolution (MinimumEssen-tial Medium Eagle, Joklik modification, Sigma, M-0518), supplementedwith1.2mMCaCl2, 10mMHEPES, 2.5 g/l taurine, 0.175 g/l pyruvic acid,and 0.75 g/l ribose (pH=7.2). After removing the blood the perfusatewas switched to nominally calcium-free MEM for 10 min. Dispersion ofcells was achieved by application of 0.5 g/l collagenase (Sigma type I) inthe presence of 75 μMCaCl2 for 40min. During this isolation procedurethe solutions were gassed with oxygen and the temperature wasmaintained at 35 °C. Finally, the left ventricular wall was minced andgently agitated. The cells, freshly released from the tissue,were stored atroom temperature before use. At least 60% of the cells were rod-shapedand showed clear striation when the external calcium was restored.

2.7. Measurement of apamin-sensitive current in single rat and canineventricular cells

One drop of the cell suspension was placed into a lucid chambermountedonthestageofan invertedmicroscope(Olympus,model IX71),and at least 10 min was allowed for myocytes to adhere before startingsuperfusion. External solution contained (inmM):N-methyl-glucamine140.0, KCl 4.0, MgCl2 1.0, glucose 5.0, and HEPES 10.0 (pH=7.4).Micropipettes were fabricated from borosilicate glass capillaries (ClarkElectromedical Instruments) using a microprocessor-controlled hori-zontal puller (Sutter Instruments). These electrodes had a resistance of1.5–2.5 MΩ when filled with pipette solution containing (in mM):potassium–gluconate 144.0, MgCl2 1.15, BAPTA 5.0, HEPES 10.0(pH=7.2). Free calcium concentration in the pipette solution was setto 900 nM calculated with WinMaxC software [22]. This software

enabled the calculation of the free [Ca2+]i levels using the appropriatemixture of CaCl2 and BAPTA. Due to the fast diffusion constants ofCa2+and BAPTA, we can presume that after cell dialysis the [Ca2+]iapproximates the Ca2+ level in the pipette solution. Nevertheless, sincethe [Ca2+]i may not fully equilibrate the adjusted pipette solution(900 nM), the [Ca2+]i is rather referred as “highly elevated” instead of“900 nM” in this condition.

Currents were recorded with Axopatch 1-D amplifier (AxonInstruments) using whole-cell configuration of the patch clamp tech-nique. Gigaseals were established with gentle suction and the cellmembrane beneath the tip was disrupted with further suction or byapplication of short electrical pulses. Series resistancewas compensatedfor 80%. Membrane currents were digitized with an analog-to-digitalconverter (Digidata 1440A), under software control (pClamp 10.0).Membrane currents were recorded from each cell before and afterapplication of 100 nM apamin. The external solution was sodium- andcalcium-free to eliminate sodium, calcium, and NCX currents. No ionchannel blocker was added to avoid any possible unspecific blockingeffect on the SK current. The applied voltage-clamp protocols andsolutions were identical to those used by Xu et al. [17].

Apamin-sensitive current was also investigated by the perforatedpatch clamp method in order to prevent changes in the intracellularmilieu of the cells, and to investigate the possible dynamic regulatoryrole of [Ca2+]i. The pipette solution contained (in mM): K-glutamate120, KCl 25, MgCl2 1, HEPES 10, and EGTA 5, pH 7.4 with KOH. Thispipette solution was supplemented with 200 μg/ml amphotericin B inorder to develop a reasonably good electrical access with the cellinterior. When amphotericin B penetrated into the cell membrane, theaccess resistance was 15–20 MΩ. The external solution contained (inmM): NaCl 138, KCl 4, MgCl2 1, CaCl2 1.8, NaHPO4 0.33, glucose 10,HEPES 10, pH7.4withNaOH. The cellswere loadedwith the fluorescentdye, Fluo 4-AM (5 μM), and calcium transients were recordedsimultaneously with the current.

2.8. Recording of [Ca2+]i transients in field-stimulated canine and ratventricular tissues

Canine and rat right ventricular tissue samples were loaded with25 μM Fluo 4-AM (Molecular Probes Inc.) at room temperature for50min. The tissue samplesweremounted in a lowvolumeQuickChangeimaging chamber (RC47FSLP, Warner Instruments) and paced initiallyat 1 Hz. Measurements were performed using an Olympus IX 71 typeinverted fluorescence microscope. Optical signals were recorded by aphoton counting photomultiplier module (Hamamatsu, model H7828)sampled at 1 kHz. The dye was excited at 480 nm, and fluorescenceemissionwasdetectedat 535nm.Opticalmeasurementswere governedby using Isosys software (Experimetria Ltd., Hungary). Frequency-dependent changes of [Ca2+]i transientsweremonitored after establish-ing steady-state conditions at each studied frequency (0.3, 1, and 3 Hz).

2.9. Drugs

All chemicals were purchased from Sigma except otherwiseindicated. Apamin was dissolved in 50 mM acetic acid yielding astock solution of 100 μM concentration, which was stored at −20 °C.Both AVE0118 (a gift from Aventis Pharma) and dofetilide (a gift fromGedeon Richter Ltd.) were dissolved in dimethyl sulfoxide, resultingin drug concentrations of 5 mM and 1 mM, respectively. These stocksolutions were stored at 4 °C. All solutions were made freshly on theday of experiment.

2.10. Statistics

Data are expressed as mean values±SEM. Student t-test for paireddata was used to compare results. Results were considered significantwhen p was less than 0.05.

Fig. 1. Expression of SK2 channel protein in canine and rat ventricular myocardium. (A)Representative immunoblots of SK2 protein obtained in canine (1, 2) and rat (3, 4) leftventricular myocardium obtained from multicellular tissues (A/a) and from isolatedmyocytes (A/b). Proteinswere purified from6 canine and 6 rat hearts and separatedusing10% SDS-PAGE. Protein expression were detected at a molecular weight of 60 kDa. (B–E)Confocal (left columns) and transmission microscopic (right columns) images capturedfrom the surface focus plane of isolated canine (B), (C) and rat ((D), (E) ventricularmyocytes. Panels (C) and (E) show immunostained cells, while the in (B) and (D) therespectivenegative controls, stainedonlywithAlexafluor 488-labeled secondary antibody,are depicted.

Fig. 2. Effect of 100 nM apamin on action potential configuration in multicellular cardiac preppotentials, recorded from the same preparations before and after 45min exposure to 100 nMap(DV), rat right ventricular papillary muscles (RV), human right ventricular papillary muscle (HVobtained for APD90 are shown in panel (F). Columns and bars represent mean data±SEM, num

659N. Nagy et al. / Journal of Molecular and Cellular Cardiology 47 (2009) 656–663

3. Results

3.1. Expression of SK2 channel protein in canine and rat ventricularmyocardium

As indicated by the representative Western blot samples, shown inFig. 1(A/a), SK2 channel protein was abundantly expressed in theventricular tissues of both dog and rat. Results of semiquantitative assayrevealed approximately similar levels of SK2 protein expression in thetwo species. To exclude the contribution of the SK2 proteins derivedfrom vessels, Western blot analysis was also performed in isolated ratand canineventricular cells. SK2proteinwas clearlydetectable at 60kDain both species, with similar protein levels (Fig. 1A/b).

For further confirmation, the distribution of SK2 channel proteinwasdetected directly in isolated myocytes. Confocal microscopic observa-tion of the immunostained cardiomyocytes also indicates comparablesurface distribution of SK2 channel protein in canine (Fig. 1C) and rat(Figs. 1E) ventricular cells. Respective negative controls are depicted inpanels (B) and (D).

3.2. Effect of apamin on action potentials recorded from canine, rat,and human multicellular cardiac preparations

Effect of 100 nM apamin on action potential configuration wastested in right ventricular papillary muscles and left atrial trabeculaeexcised from canine, rat, and healthy humanhearts using conventionalmicroelectrode techniques. All preparations were paced at a constantfrequency of 1 Hz in Krebs–Henseleit solution. Representative pairs ofaction potentials, recorded from the same preparations before andafter 45min exposure to 100 nM apamin, are depicted in panels (A–E)of Fig. 2, while the summarized results are presented in Fig. 2F. Asshown in Fig. 2, none of the preparations responded to the apaminexposure with alterations in action potential duration. These resultssuggest that SK channels – in spite of their abundant expression – fail

arations recorded at a stimulation frequency of 1 Hz. (A–E) Representative pairs of actionamin. The preparations exposed to apaminwere canine right ventricular papillarymuscles), canine left atrial trabeculae (DA), and rat left atrial trabeculae (RA). The average resultsbers in parentheses denote the number of preparations tested.

Fig. 3. (A–C) Effect of 100 nM apamin on action potential configuration in canine and rat ventricular muscles recorded at pacing frequencies of 0.3, 1 and 3 Hz. Representative pairs ofaction potentials taken from the same preparations before after the application of 100 nM apamin are presented. (D–F) Representative traces of intracellular calcium transientsrecorded from canine and rat ventricular tissues at the identical frequencies (0.3, 1 and 3 Hz).

660 N. Nagy et al. / Journal of Molecular and Cellular Cardiology 47 (2009) 656–663

to mediate a significant amount of ion current under physiologicalconditions (i.e. when studied in multicellular atrial and ventricularpreparations of the various species).

Since activation of SK channels is strictly dependent on elevation ofthe cytosolic free calciumconcentration, the effect of 100nMapamin onaction potential configurationwas also evaluated in canine and rat rightventricular papillary muscles paced with various stimulation frequen-cies. The representative pairs of action potentials, recorded from thesame preparations before and after the exposure to apamin andpresented in Figs. 3A–C, indicate that action potential duration wasnot modified by apamin at any of the pacing frequencies applied. Theaverage canine APD90 values were 215±5.0 versus 219±4.0 ms at0.3 Hz, 207±3.0 versus 209.5±3.5 ms at 1 Hz, and 188.0±3.0 versus

Fig. 4. Effect of apamin in preparations with attenuated repolarization reserve due to pretreatmApamin (100 nM) was applied in the presence of 300 nM dofetilide and 5 μM AVE0118, respe

188.5±3.5 ms at 3 Hz, respectively, before and after apamin (n=5).The respective values in rat preparations were 83.8±7.1 versus 83.2±8.1 ms at 0.3 Hz, 73.0±3.5 versus 72.6±2.5 ms at 1 Hz, and 71.2±2.9versus 72.6±3.4 ms at 3 Hz (n=5).

[Ca2+]i transients, recorded from right ventricular tissues of dogsand rats, paced with the frequencies of 0.3, 1, and 3 Hz are presentedin the right columns of Fig. 3. According to these graphs the highest[Ca2+]i values were observed at 3 Hz in canine myocytes, while at0.3 Hz in the case of rat ventricular tissues— following the predictionsof the positive and negative staircase phenomena, respectively,known to be characteristic of canine and rat ventricular myocardium.In spite of the elevated [Ca2+]i values seen at these frequencies noeffect of apamin on action potential duration was observed.

ent with dofetilide in canine (A) and AVE0118 in rat (B) ventricular myocardium at 1 Hz.ctively.

Fig. 5. The effect of 100 nM apamin on membrane currents recorded from voltage clamped canine (A) and rat (B) ventricular myocytes. Steady-state current–voltage relations wereobtained by plotting the end-pulsemembrane current as a function of the respectivemembrane potential (Vm).Measurementswere performed before and after the application of 100 nMapamin. The [Ca2+]i was highly elevated as a consequence of pipette solution buffered to 900 nM, as calculated by theWinMaxC software. During the experiments a different extent of cellshortening could be observed. Symbols and bars represent mean data±SEM, the number of myocytes were 7 in both groups. Error bars are missing when smaller than symbol size.

661N. Nagy et al. / Journal of Molecular and Cellular Cardiology 47 (2009) 656–663

In order to exclude the possibility that a strong repolarizationreservemay fully compensate for any apamin-induced actionpotentiallengthening, we blocked the major repolarizing currents prior to the

Fig. 6. Effect of 100 nM apamin on transmembrane currents recorded from canine and rat ((Aconditions. Currentwas obtained using a voltage ramp arising fromaholding potential of−50mcells were loadedwith 5 μMFluo 4-AM allowing for simultaneous recording of [Ca2+]i transienaccidental disruption of the patch. The investigation of the effect of apamin as a difference cur

application of apamin. These interventions are known to augmentchanges in action potential duration induced by the blockade of otheroutward currents. In canine papillary muscles the rapid delayed

) and (B) respectively, upper panel) ventricular myocytes under perforated patch clampV to+40mV, then hyperpolarizing to−100mVat a rate of 175mV/s (lower panels). The

ts ((A) and (B), middle panels). The pipette solution contained 5mMEGTA to indicate anyrent was attempted, however, no effect of apamin could be detected.

662 N. Nagy et al. / Journal of Molecular and Cellular Cardiology 47 (2009) 656–663

rectifier potassium current (IKr) was blocked by 300 nM dofetilide, inthe rat preparations 5 μM AVE0118 was used to inhibit the transientoutward potassium current (Ito) plus the ultra rapid delayed rectifiercurrent (IKur). While a substantial attenuation of the repolarizationreservewas obtained in the presence of these drugs (as confirmed by asignificant lengthening of APD90 in both species, shown in Figs. 4Aand B, even under these circumstances (i.e. when the repolarizationreserve was greatly compromised) apamin exposure failed to alteraction potential duration at all. In canine papillary muscles APD90 wasincreased by 300 nM dofetilide from 211±3.0 to 245±5.7 ms(pb0.05, n=5), and 247±6.0 ms was measured when 100 nMapamin was superfused in the presence of dofetilide. The respectivevalues in the case of rat preparations were: 70.8±5.2 ms in control,148.2±7.3 ms in the presence of 5 μM AVE0400 (pb0.05, n=5), and148.4±7.0 ms following further exposure to 100 nM apamin.

3.3. Effect of apamin on ion currents in single canine and ratventricular cells

To detect the effect of apamin on SK current in single cells, weapplied both the whole-cell and the perforated path configurations ofthe patch clamp technique and measured the apamin-sensitive currentin isolated canine and rat ventricular myocytes. Under whole-cellconditions, the free calcium concentration in the pipette solution wasset to 900 nM in these experiments in order to induce maximal acti-vation of SK channels. This pipette Ca2+concentration approximates theCa2+level which normally occurs during systole [23], therefore underthis condition the [Ca2+]i was highly elevated as compared to the dias-tolic Ca2+ level.

After performing whole-cell configuration, several minutes wereallowed to reach the complete dialysis. As a consequence of the subs-tantially elevated [Ca2+]i the cells typically showed different extent ofcell shortening. Experiments in which this procedure caused a signi-ficant change in the seal resistance and/or access resistance werediscarded. Test depolarizations, arising from the holding potentialof −50 mV, were applied to membrane potentials ranging from −120to+60mV for 150ms.Membrane currents recorded at the end of thesepulses and normalized to cell capacitance were plotted against therespective test potential. Since our external and pipette solutions weresodium- and calcium-free, and were also poor in chloride, the recordedmembrane currentwas carried exclusively bypotassium. These “steady-state” current–voltage relations, obtained before and after the exposureto100nMapamin,were fully identical in both canine and rat ventricularcells (Figs. 5A, B, respectively, n=7 for each) indicating that apaminfailed to activate any ion current in these myocytes throughout theentire voltage range tested.

Under perforated patch clamp conditions the cytosol was preventedfrom dialysis. The cells were loaded with the fluorescent dye, fluo-4AMto monitor the calcium release and membrane current simultaneously.The membrane was initially depolarized from the holding potentialof −50 mV to +40 mV, then it was hyperpolarized to −100 mV (at arate of 175 mV/s). Upon depolarization [Ca2+]i transients were clearlyobservable indicating normal functional Ca2+ homeostasis of the cells,however, the steady-state membrane current failed to alter afteradministration of 100 nM apamin in either species (Figs. 6A, B).

4. Discussion

Theoretically, SK current might be particularly important duringcardiac repolarization, since it canbe consideredas a direct linkbetweencalcium handling and membrane potential changes. Changes in theheart rate are known to be accompanied by alterations in both [Ca2+]iand APD [24]. This frequency-dependent modulation of APD is mainlyattributed to kinetic properties of the major repolarizing currents. If SKcurrent could be shown to operate in cardiac tissues under physiologicalconditions, it might evidently contribute to the frequency-dependent

adaptation of APD governed by changes in [Ca2+]i. Such integrationbetween calcium handling and repolarization might improve theadaptation of cardiac function. Furthermore, in a previous study signi-ficant changes in potassium channel expression, as well as in cardiacaction potential profile due to chronic changes in [Ca2+]i and/or calciumhandling proteins, were shown [25]. The aim of the present work,therefore, was to investigate the possible role of SK current in cardiacrepolarization in canine, rat and human cardiac tissues under physio-logical conditions. Our results verified the abundant expression of SK2channels in ventricularmyocardium, however,– in sharp contrast to therecent results of Xu et al. [17] – apamin, the potent inhibitor of SKchannels, failed to alter APD in any of the preparations examined,indicating that SK channelsmay have only negligible – if any – influenceon ventricular and atrial repolarization under physiological conditions.

In a previous study Xu et al. demonstrated that the SK2 channelexpression is more abundant in atrial than in ventricular tissue of themouse [17] suggesting that atrial myocardium should be moresensitive to apamin than the ventricular one. Indeed, they foundapamin to cause considerably larger prolongation of APD in atrial thanin ventricular preparations. These findings were also confirmed byTuteja et al. [19]. Furthermore, they have shown the SK channelisoforms differentially express in atrial and ventricular myocytes ofthe mouse. These findings may have important electrophysiologicalimplications. Although – based on their physiological significance –

we focused most of our efforts on the ventricular SK channels, ourinvestigations were extended to atrial myocardium as well, but –

similarly to results obtained in ventricular tissues – apamin wasineffective to alter action potential morphology in the atrial myocar-dium of rats and dogs.

The discrepancies between the present results and previousfindings [17] are unexpected and not well understood. Theoretically,they may be attributed to interspecies differences or the partiallydifferent experimental conditions. In our experiments action poten-tials were recorded with sharp microelectrodes in canine, rat, andhuman multicellular preparations, while Xu et al. used perforatedpatch clamp in isolated murine myocytes. It should be noted thataction potential recording from single myocytes – although anestablishedmethod – is probably not the best technique to investigateminor or fine changes in cardiac repolarization. In contrast tomulticellular preparations, the impairment of potassium channelsduring the enzymatic dissociation process cannot be ruled out in thecase of single cells, which may result in action potential instability,unexpected time dependent changes and attenuation of action of anion channel blocker. In further experiments, therefore, we tried tomeasure the apamin-sensitive current directly in voltage clampedsingle canine and rat myocytes. Since SK current is characterized byweak voltage sensitivity, it should be active in a wide range ofmembrane potential [17,18]. However, application of 100 nM apaminin our experiments contrast to those to Xu et al. which failed toinfluence the I–V curve at any of the membrane potentials studiedsuggesting that SK current has practically no role in ventricularrepolarization in isolated cardiomyocytes. The lack of effect of apaminin our patch clamp experiments was unexpected, because Xu et al.reported an apamin-sensitive current of 4 pA/pF in the voltage rangecorresponding to the action potential plateau [17]. This is roughlyequivalent with the total membrane current measured in ourmyocytes (see Figs. 5A and B).

An alternative explanation might be considered if physiological[Ca2+]i levels were not high enough to activate SK current of sufficientdensity to overcome normal repolarization reserve, whichmight largelycompensate for inhibition of SK current under physiological conditions.Inorder to clarify this possibility, apaminwas also tested in preparationshaving repolarization reserve markedly attenuated by blocking themajor repolarizing currents (IKr in dog, and IKur plus Ito in rat). Asexpected, these interventions caused significant lengtheningof APD, butthe subsequently applied apamin still had no effect on repolarization. In

663N. Nagy et al. / Journal of Molecular and Cellular Cardiology 47 (2009) 656–663

other words, apamin failed to alter cardiac repolarization even if themajor repolarizing currents were markedly compromised.

Experiments to study the role of [Ca2+]i on the effect of apaminwerealsodesigned.Ourfluorometricmeasurements revealed that thehighest[Ca2+]i levels could be observed at 3 Hz in canine, while at 0.3 Hz in ratventricularmyocytes. Nodifference in the effect of apaminwas seen as afunction of the stimulating frequency on action potential configurationsuggesting that even the highest physiological [Ca2+]i levelsmay still betoo low to activate SK channels. Similar conclusion can be drawn fromthe single cell experiments, where [Ca2+]i was highly elevated due to900 nM Ca2+present in the pipette solution (i.e. to a level close tonormal systolic [Ca2+]i values), or when dynamic changes in [Ca2+]iwere allowed under perforated patch clamp conditions. No effect ofapamin was seen in these experiments either.

Existence of some interspecies differences cannot be fully ruledout. Our measurements were mainly performed in canine and ratmyocardium, while those of Xu et al. were done in murine and humantissues [17]. Since the known electrophysiological properties of caninemyocytes are very similar to those of human [21,26], while themurineand rat myocytes are also almost identical regarding their actionpotential morphology and calcium handling [20], the lack of apaminaction in our experiments can hardly be explained by theseinterspecies differences. In addition, apamin remained ineffective inour single experiment performed in healthy human myocardium.

A recent study [27] by the same group further emphasized thefunctional role of the SK2 channels inmouse atrial myocytes, and raisedthe possibility that dysfunction of this channel may be an importantcontributor to the genesis of atrial fibrillation. These data are intriguingand offer several further pharmacological implications. However, wefeel important to emphasize that the ion channel composition, aswell asthe kinetics of the repolarization process in the mouse atria andventricles substantially differ from those in human, or in other,electrophysiologically more relevant experimental model, such as thedog. Therefore, it seems to be rather difficult to delineate the possiblerole of these channels in the induction and/or the maintenance of thehuman atrial fibrillation based on the data from this transgenic mousemodel. Consequently, the results of Li et al. [27] shouldalso beconfirmedon samples fromundiseased and in atrialfibrillating remodelled humanhearts or on more human relevant models. In the case, if SK2 channels,indeed, are proven to have an important function in remodelled humanatria. Our present data may suggest, that the pharmacologicalmodulation of the atrial SK2 channels can be beneficial in atrialfibrillation because of its ineffectiveness on the ventricles.

In summary, our data, in contrast to previous report seem toquestion the significance of thepreviousobservationsmade inmice, andthe reason of this discrepancy is not clear. We must conclude that SKchannels play only negligible – if any – role in cardiac repolarizationunder physiological conditions. However, the possibility that thesechannels may activate under special circumstances (such as calciumoverload, heart failure, or conditions of ischemia/reperfusion, or atrialfibrillation) cannot be ruled out. Our data has important implicationsregarding pharmacological speculations and possible drug develop-ment. Further experiments are required to investigate the possible roleof SK channels in diseased myocardium, before its pharmacologicalmodulation could be utilized.

Acknowledgments

We thankMiss Zsuzsanna Sebők for the skilful technical assistance.This work was supported by Hungarian National Research Foundation(GVOP-3.2.1, OTKA NI-61902, CNK77855 and F-61222), HungarianMinistry of Health (ETT353/2006), National Research and Develop-ment Programs (OM-00129/2007), European Community (EU FP6grant LSHM-CT-2005-018833, EUGeneHeart and EU FP6 grant LSHM-

CT-2006-018676, NORMACOR) and European Community (EU FP7grant ICT-2008-224381, preDiCT).

References

[1] Varró A, Lathrop DA, Hester SB, Nanási PP, Papp JG. Ionic currents and actionpotentials in rabbit, rat, and guinea pig ventricular myocytes. Basic Res Cardiol1993 Mar–Apr;88(2):93–102.

[2] Carmeliet E. K+ channels and control of ventricular repolarization in the heart.Fundam Clin Pharmacol 1993;7(1):19–28.

[3] Biliczki P, Virág L, Iost N, Papp JG, Varró A. Interaction of different potassiumchannels in cardiac repolarization in dog ventricular preparations: role ofrepolarization reserve. Br J Pharmacol 2002 Oct;137(3):361–8.

[4] Thomsen MB, Matz J, Volders PG, Vos MA. Assessing the proarrhythmic potentialof drugs: current status of models and surrogate parameters of torsades de pointesarrhythmias. Pharmacol Ther 2006 Oct;112(1):150–70.

[5] Tomaselli GF, Zipes DP. What causes sudden death in heart failure? Circ Res 2004Oct 15;95(8):754–63.

[6] Lederer WJ, Berlin JR, Cohen NM, Hadley RW, Bers DM, Cannell MB. Excitation–contraction coupling in heart cells. Roles of the sodium–calcium exchange, thecalcium current, and the sarcoplasmic reticulum. Ann NY Acad Sci 1990;588:190–206.

[7] Bers DM. Ca regulation in cardiac muscle. Med Sci Sports Exerc 1991 Oct;23(10):1157–62.

[8] Armoundas AA, Hobai IA, Tomaselli GF, Winslow RL, O'Rourke B. Role of sodium–calcium exchanger in modulating the action potential of ventricular myocytesfrom normal and failing hearts. Circ Res 2003 Jul 11;93(1):46–53.

[9] Altamirano J, Bers DM. Effect of intracellular Ca2+ and action potential durationon L-type Ca2+ channel inactivation and recovery from inactivation in rabbitcardiac myocytes. Am J Physiol Heart Circ Physiol 2007 Jul;293(1):H563–73.

[10] Ter Keurs HE, Boyden PA. Calcium and arrhythmogenesis. Physiol Rev 2007 Apr;87(2):457–506.

[11] Maltsev VA, Silverman N, Sabbah HN, Undrovinas AI. Chronic heart failure slowslate sodium current in human and canine ventricular myocytes: implications forrepolarization variability. Eur J Heart Fail 2007 Mar;9(3):219–27.

[12] Bassani RA. Transient outward potassium current and Ca2+ homeostasis in theheart: beyond the action potential. Braz J Med Biol Res 2006 Mar;39(3):393–403.

[13] Knaus HG, Garcia-Calvo M, Kaczorowski GJ, Garcia ML. Subunit composition ofthe high conductance calcium-activated potassium channel from smooth muscle,a representative of the mSlo and slowpoke family of potassium channels. J BiolChem 1994 Feb. 11.;269(6):3921–4.

[14] Meera P, Wallner M, Song M, Toro L. Large conductance voltage- and calcium-dependent K+ channel, a distinct member of voltage-dependent ion channelswith seven N-terminal transmembrane segments (S0–S6), an extracellular Nterminus, and an intracellular (S9–S10) C terminus. Proc Natl Acad Sci U S A 1997Dec 9;94(25):14066–71.

[15] JoinerWJ, TangMD,Wang LY, Dworetzky SI, Boissard CG, Gan L, et al. Formation ofintermediate-conductance calcium-activated potassium channels by interactionof Slack and Slo subunits. Nat Neurosci 1998 Oct;1(6):462–9.

[16] Kohler M, Hirschberg B, Bond CT, Kinzie JM, Marrion NV, Maylie J, et al. Small-conductance, calcium-activated potassium channels from mammalian brain.Science 1996 Sep 20;273(5282):1709–14.

[17] Xu Y, Tuteja D, Zhang Z, Xu D, Zhang Y, Rodriguez J, et al. Molecular identificationand functional roles of a Ca(2+)-activated K+ channel in human and mousehearts. J Biol Chem 2003 Dec 5;278(49):49085–94.

[18] Wang W, Watanabe M, Nakamura T, Kudo Y, Ochi R. Properties and expression ofCa2+-activated K+ channels in H9c2 cells derived from rat ventricle. Am J Physiol1999 May;276(5 Pt 2):H1559–66.

[19] Tuteja D, Xu D, Timofeyev V, Lu L, Sharma D, Zhang Z, et al. Differential expressionof small-conductance Ca2+-activated K+ channels SK1, SK2, and SK3 in mouseatrial and ventricular myocytes. Am J Physiol Heart Circ Physiol 2005 Dec;289(6):H2714–23.

[20] Gussak I, Chaitman BR, Kopecky SL, Nerbonne JM. Rapid ventricular repolarizationin rodents: electrocardiographic manifestations, molecular mechanisms, andclinical insights. J Electrocardiol 2000 Apr;33(2):159–70.

[21] Jost N, Virág L, Bitay M, Takács J, Lengyel C, Biliczki P, et al. Restricting excessivecardiac action potential and QT prolongation: a vital role for IKs in human ven-tricular muscle. Circulation 2005 Sep 6;112(10):1392–9.

[22] Patton C, Thompson S, Epel D. Some precautions in using chelators to buffer metalsin biological solutions. Cell Calcium 2004 May;35(5):427–31.

[23] Bers DM. Calcium fluxes involved in control of cardiac myocyte contraction. CircRes 2000 Aug 18;87(4):275–81.

[24] Bers DM. Excitation–Contraction Coupling and Cardiac Contractile Force. KluwerAcademic Press; 2001.

[25] Xu Y, Zhang Z, Timofeyev V, Sharma D, Xu D, Tuteja D, et al. The effects ofintracellular Ca2+ on cardiac K+ channel expression and activity: novel insightsfrom genetically altered mice. J Physiol 2005 Feb 1;562(Pt 3):745–58.

[26] Varró A, Baláti B, Iost N, Takács J, Virág L, Lathrop DA, et al. The role of the delayedrectifier component IKs in dog ventricularmuscle and Purkinjefibre repolarization.J Physiol 2000 Feb 15;523(Pt 1):67–81.

[27] Li N, Timofeyev V, Tuteja D, Xu D, Lu L, Zhang Q, et al. Ablation of a Ca2+-activatedK+ channel (SK2 channel) results in action potential prolongation in atrialmyocytes and atrial fibrillation. J Physiol 2009 Mar 1;587(Pt 5):1087–100.