blockade of the inward rectifying potassium current terminates ventricular fibrillation in the...

621

Blockade of the Inward Rectifying Potassium CurrentTerminates Ventricular Fibrillation in the Guinea Pig Heart

MARK WARREN, PH.D.,∗ PRABAL K. GUHA, M.D.,∗ OMER BERENFELD, PH.D.,ALEXEY ZAITSEV, PH.D., JUSTUS M.B. ANUMONWO, PH.D., AMIT S. DHAMOON, B.A.,

SUVEER BAGWE, M.D., STEVEN M. TAFFET, PH.D., and JOSE JALIFE, M.D.

From SUNY Upstate Medical University, Syracuse, New York, USA

IK1 Blockade and VF Dynamics. Introduction: Stable high-frequency rotors sustain ventricular fib-rillation (VF) in the guinea pig heart. We surmised that rotor stabilization in the left ventricle (LV) andfibrillatory conduction toward the right ventricle (RV) result from chamber-specific differences in func-tional expression of inward rectifier (Kir2.x) channels and unequal IK1 rectification in LV and RV myocytes.Accordingly, selective blockade of IK1 during VF should terminate VF.

Methods and Results: Relative mRNA levels of Kir2.x channels were measured in LV and RV. In addition,LV (n = 21) and RV (n = 20) myocytes were superfused with BaCl2 (5–50 µmol/L) to study the effects onIK1. Potentiometric dye-fluorescence movies of VF were obtained in the presence of Ba2+ (0–50 µmol/L)in 23 Langendorff-perfused hearts. Dominant frequencies (DFs) were determined by spectral analysis, andsingularity points were counted in phase maps to assess VF organization. mRNA levels for Kir2.1 andKir2.3 were significantly larger in LV than RV. Concurrently, outward IK1 was significantly larger in LVthan RV myocytes. Ba2+ decreased IK1 in a dose-dependent manner (LV change > RV change). In baselinecontrol VF, the fastest DF domain (28–40 Hz) was located on the anterior LV wall and a sharp LV-to-RVfrequency gradient of 21.2 ± 4.3 Hz was present. Ba2+ significantly decreased both LV frequency andgradient, and it terminated VF in a dose-dependent manner. At 50 µmol/L, Ba2+ decreased the averagenumber of wavebreaks (1.7 ± 0.9 to 0.8 ± 0.6 SP/sec · pixel, P < 0.05) and then terminated VF.

Conclusion: The results strongly support the hypothesis that IK1 plays an important role in rotor stabi-lization and VF dynamics. (J Cardiovasc Electrophysiol, Vol. 14, pp. 621-631, June 2003)

rotors, Kir2.1, Kir2.3, optical mapping, fibrillatory conduction, ventricular fibrillation

Introduction

Recent evidence suggests that ventricular fibrillation (VF)is organized by high-frequency rotors of cardiac excitation.1-7

In the guinea pig heart, VF is maintained by the activationof a single rotor located in the free wall of the left ventricle(LV), with fibrillatory conduction toward the right ventricle(RV),8 which results in predictable but heterogeneous dis-tribution of dominant frequencies (DFs) and an LV-to-RVfrequency gradient. Simulations using ionic models of car-diac tissue,9,10 as well as patch clamp analysis of LV andRV myocytes, strongly suggest that differences in the recti-fication of the outward background current play a key rolein allowing very-high-frequency rotors to stabilize in the LVbut not in the RV.8 However, to our knowledge, the identityof the channels responsible for those differences has not beendefinitely established.

We surmise that chamber-specific differences in the func-tional expression of channels responsible for the inward rec-

Supported by Grants RO1 HL70074 and P01 HL39707 from the NHLBIand by the Michel Mirowski International Fellowship in Cardiac Pacing andElectrophysiology (NASPE 2002-2003).

∗The first two authors contributed equally to this paper.

Address for correspondence: Jose Jalife, M.D., Department of Pharmacol-ogy, SUNY Upstate Medical University, 766 Irving Avenue, Syracuse, NY13210. Fax: 315-464-8000; E-mail: [email protected]

Manuscript received 6 January 2003; Accepted for publication 11 March2003.

tifying current (IK1) are important for the stability of VF inthe guinea pig heart. Specifically, we postulate that a rel-atively large outward component of IK1 abbreviates actionpotential duration (APD) in the LV and allows for stablehigh-frequency rotors, which maintain VF. Thus, we haveused a ribonuclease protection assay (RPA) to determine LVversus RV differences in mRNA levels of Kir2.x subfamilymembers responsible for IK1.11-14 In addition, we have usedthe whole-cell voltage clamp technique in the presence ofBa2+ to selectively block IK1

15,16 in LV and RV myocytesand to establish an appropriate range of concentrations thatwould reduce the outward component sufficiently to allowAPD prolongation without significantly reducing the restingmembrane potential. Finally, we conducted optical mappingexperiments to study Ba2+concentration-dependent effectson LV and RV frequencies of excitation, and VF organiza-tion and stability.

Methods

All experiments were performed according to the Guide-lines for the Care and Use of Laboratory Animals (NIH Pub-lication No. 85-23, revised 1996).

RNase Protection Assay

RNase protection assay was carried out in samples ob-tained from 12 hearts following a technique described previ-ously.17 Briefly, total RNA was extracted from the LV and RVof 12 guinea pigs (300–600 g) using Tri Reagent (MRC Inc.)following the manufacturer’s instructions. Antisense probes

622 Journal of Cardiovascular Electrophysiology Vol. 14, No. 6, June 2003

were designed to recognize coding regions of guinea pigKir2.x channels, and in vitro transcribed full-length mRNAfor each isoform was used as positive control. A probe for thehousekeeping gene cyclophilin was used as an internal con-trol. The RPA was performed using the Riboquant RPA kit(Pharmingen). Kir2.x signals were visualized by Phospho-rimager (Molecular Dynamics) and were quantified as a per-centage of the cyclophilin signal using Imagequant (Molec-ular Dynamics) software.

Isolation of Cardiac Myocytes

Cardiac myocytes were isolated as previously described.18

Briefly, 52 adult guinea pigs (400–500 g) were anesthetizedwith sodium pentobarbital [35 mg/kg intraperitoneally (IP)]immediately after administration of heparin (500 units IP).Following a midline thoracotomy, the hearts were quicklyremoved and the aorta was cannulated for retrograde per-fusion. After 5-minute perfusion with normal oxygenated(100% oxygen) HEPES-Tyrode’s solution at 37◦C (to re-move excess blood), a low-calcium medium was perfused for10 minutes. The heart then was perfused using an enzyme-containing medium (collagenase type II, 100 units/mL,Worthington Biochemistry) for 15 minutes and then using anincubation (KB) medium for another 15 minutes. Thereafter,perfusion was stopped, and the LV and the RV were removedfrom the Langendorff apparatus, cut into small chunks, andstored separately. Cells were freed by gentle mechanical ag-itation of the chunks and then incubated in the KB mediumfor another 45 minutes. The KB solution was gradually re-placed by Tyrode’s solution. This process was carried out in5-minute steps, with an increasing volume of the Tyrode’ssolution. The cells were placed on the stage of an invertedmicroscope (Nikon Diaphot) for electrophysiology.

Electrophysiologic Recordings and Analyses in Myocytes

All recordings were performed at 37◦ ± 1◦C using eitheran Axopatch-1D (voltage clamp) or Axoclamp-2B (currentclamp) amplifier (Axon Instruments, Union City, CA, USA).Electrode resistance was 1.5 to 2 M� when filled with thepipette solution. Voltage clamp signals were low-pass fil-tered (200–500 Hz) and acquired (1–5 kHz) using pCLAMP6.0. All voltage clamp recordings were corrected by –8 mV,the liquid junction potential in relation to the external so-lution [tip potentials in the suction pipette were measuredas the voltage difference between a pipette filled with inter-nal pipette solution and a 3M KCl microelectrode (referenceelectrode) after the bath solution had been changed fromnormal HEPES-Tyrode’s solution to pipette solution].8,18

Cell capacitance was measured using a brief hyperpolarizing(10 mV) pulse from a holding potential of −100 mV.

To study the background current density-voltage (I-V) re-lation, a slow voltage clamp ramp (1.6 mV/sec) was appliedfrom −110 mV to +20 mV (holding potential −80 mV).Nifedipine (2 µmol/L) was used to block L-type calcium cur-rents. Dose-response curves (5, 10, 20, 30, and 50 µmol/L)were constructed for BaCl2 blockade of LV (n = 21) andRV (n = 20) myocyte background currents. In a subset ofexperiments, Ba2+-sensitive current was determined in 11separate hearts by measuring the background current in my-ocytes before and after superfusion with 1 mmol/L BaCl2,and the difference current yielded the inward rectifier current

(IK1).16 Tetrodotoxin (TTX, 30 µmol/L) was used to block thevoltage-gated sodium current in this subset of experiments.

Isolated Heart Preparation

Twenty-three male guinea pigs (500–700 g) were hep-arinized and then anesthetized with sodium pentobarbital(35 mg/kg, IP). After midline thoracotomy, the heart wasexcised, Langendorff-perfused with normal oxygenated Ty-rode’s solution at 37.1◦ ± 0.5◦C (20 mL/min), and kept in abath at the same temperature. A nylon mesh was used to min-imize mechanical artifacts. The hearts were allowed to beatin sinus rhythm for 20 to 30 minutes, after which VF wasinduced by burst pacing (cycle length 20 msec) at randomventricular sites to establish the baseline behavior. Episodesof VF lasting at least 5 minutes were considered to be stable.Of 23 hearts mapped, 20 were used in the analysis of Ba2+ ef-fects on VF frequency and dynamics, as follows: 1 µM Ba2+,4 hearts; 3 µM Ba2+, 5 hearts; 12 µM Ba2+, 6 hearts for base-line, but only 5 hearts for Ba2+; and 50 µM Ba2+, 5 hearts.The other three hearts were used as control experiments inwhich Ba2+ was not introduced and the VF was maintainedfor 20 minutes (see Fig. 7).

Acquisition of Data from Isolated Hearts

The optical mapping system is described in detail else-where.19 Briefly, a small bolus of the voltage sensitive dye di-4-ANEPPS (10 µmol/L) was added to the coronary perfusionjust prior to VF induction and every 15 to 20 minutes there-after. This procedure was used in an effort to reduce poten-tial phototoxicity while preventing significant signal-to-noiserundown. In separate experiments, without optical mapping,we verified that the results of the optical experiments were notaffected by the dye.8 Four light sources were used to homo-geneously excite the dye fluorescence from the anterior andposterior walls. Two identical CCD cameras (DALSA, CA-D1-0128T-STDL, Waterloo, Ontario, Canada) captured theemitted signals at 600 frames/sec to obtain 5-second movies(64 × 64 pixels, ∼18 × 18 mm2).19 Activity was monitoredcontinuously by a volume-conducted ECG. Ten-second ECGepisodes were recorded (1,000 kHz) synchronously with theoptical movies. Fast Fourier transformation (FFT) was usedfor spectral analysis.

Analysis of Data from Isolated Hearts

DF maps

To construct the DF maps of each optical movie, spec-tral analysis was applied to the fluorescent signal recordedat each pixel. The frequency with the maximum power wasconsidered to be the DF and color-coded according to itsvalue.20 A domain was defined as an area on the DF map forwhich neighboring pixels shared the same DF. Only domains>1% of the total area recorded were considered for analy-sis (domains excluded were mainly in the periphery of thefield of view). To statistically analyze DF changes in eachgroup before and after Ba2+ perfusion, individual maps werealigned through their centers, rescaled to match their sizes,and then superimposed. The superimposed maps yielded foreach pixel a vector of DFs. The vectors that contained DFvalues from all individual maps provided two new maps witha pixel-by-pixel mean and standard deviation of the DFs forall the episodes analyzed.8 In addition, power spectra were

Warren et al. IK1 Blockade and VF Dynamics 623

obtained from the volume-conducted ECGs (e.g., see Fig. 8)to provide a global measure of the rate of the VF.

Phase maps

The instantaneous phase of the action potential recordedat each pixel was determined by transforming the originalsignals (5–50 Hz bandpass filtered) such that every spec-tral element was shifted by its corresponding quarter cycle(Hilbert transform21). Then, the instantaneous phase of thesignal was obtained from the inverse tangent of the ratio ofthe transformed signal to the original signal. The phase angle,with values between −π and π radians, was represented asa continuous color scheme from red to purple to constructa phase map in which the continuous spatial phase changereflected the process of excitation, repolarization, and recov-ery.3,8 A singularity point (SP) was defined as a point on thephase map were all phases converged.3 A phase singularityoccasionally becomes the center of a rotor, but more oftenit marks the fragmentation of a mother wave into daughterwavelets caused by the collision of the excitation front withrefractory tails of other waves.22,23 In hearts perfused with 3,12.5, and 50 µmol/L Ba2+, 300 frames of the phase movies atbaseline and after perfusion with Ba2+ were selected at ran-dom for each experiment, the location of the SPs was marked,and their spatial histogram was calculated for bins of 2 × 2pixels. The coordinate systems of the episodes analyzed foreach group before and after Ba2+ were aligned using the sameprocedure as for the DF map (see earlier), and a compositespatial histogram was created by counting the number of SPsin bins of 4 × 4 pixels. The mean number of SPs in eachpixel before and after perfusion with Ba2+ was averaged forall pixels to obtain the mean number of SP/sec · pixel.

Statistical Analysis

Patch clamp data and mRNA levels are given as mean ±SE. Optical data are presented as mean ± SD. Baseline val-ues were compared with those of Ba2+ using paired Student’st-test. Peak currents of LV and RV myocytes were comparedusing unpaired Student’s t-test. Groupwise concentration-dependent changes and time course of DFs and currents wereanalyzed using analysis of variance. P < 0.05 was consideredstatistically significant.

Results

Chamber-Specific Differences in IK1

It has been demonstrated that the so-called “strong” in-ward rectifying Kir2.x proteins are important molecular de-terminants of IK1.11-14 Thus, as an initial step toward estab-lishing the underlying mechanisms for the different outwardcomponents of the background current in LV and RV my-ocytes,8 we investigated chamber-specific differences in thelevel of expression of Kir2.x channels. In panel A of Figure 1are shown RPA data for Kir2.1 and Kir2.3 mRNAs from asingle experiment demonstrating denser bands for both genesin the LV. Panel B of Figure 1 clearly demonstrates that mR-NAs for both protein isoforms are significantly larger in theLV than in the RV (LV/RV for Kir2.1: 1.97 ± 0.45; Kir2.3:1.43 ± 0.08; P < 0.05 vs 1). Interestingly, neither Kir2.2nor Kir2.4 was present at significant levels in the ventriculartissue of either chamber (data not shown). The effectivenessof the Kir2.2 and Kir2.4 RPA probes was confirmed by us-

ing guinea pig brain as a positive control (Dhamoon et al.,unpublished data).

In the article by Samie et al.,8 it was presumed that theLV versus RV differences in the outward component of thebackground current were the result of differences in densityof IK1. Here we have conducted additional patch clamp exper-iments to test that idea. I-V relations were constructed afterthe application of a slow voltage clamp ramp (1.6 mV/sec)from −110 to +20 mV (holding potential −80 mV). PanelC of Figure 1 shows reversible inhibition of the current byBa2+. IK1 was measured as the difference between the cur-rent density at baseline and the current density during Ba2+(1 mmol/L). In panel D are shown average I-V relationshipsof IK1 in the LV and RV. The data demonstrate a larger out-ward component of IK1 in LV cells. Specifically, the outwardcurrent at −50 mV was 7.2 ± 0.5 pA/pF in the LV versus 4.3± 0.6 pA/pF in RV myocytes (P < 0.001).

Ba2+ Blockade of IK1

The next step of our study was to determine whether the ef-fects of micromolar concentrations of Ba2+, which is a selec-tive blocker of IK1,

15,16 are different in LV and RV myocytes.In Figure 2, panels A and B show the effects of Ba2+ (10 and50 µmol/L) on the voltage dependence of the backgroundcurrent densities measured in LV and RV myocytes, respec-tively. Outward (−50 mV, P < 0.001) and inward (−110 mV,P < 0.05) currents decreased in both cell types. The rever-sal potential (control −88.2 ± 4.0 mV) was not significantlyreduced in either type of cell at 10 µmol/L Ba2+. However,50 µmol/L Ba2+ induced a slight but significant (P < 0.001)reduction of the reversal potential in LV and RV myocytes,from −87.3 ± 4.4 mV to −78.5 ± 6.2 mV.

Panel C shows Ba2+ effects on peak outward current. Thecontrol values (n = 41) clearly confirm previous results8

showing that peak outward current is larger in LV than inRV myocytes (LV 7.3 ± 0.2 pA/pF vs RV 5.5 ± 0.1 pA/pF,P < 0.001). Ba2+ induced a concentration-dependent de-crease in the peak outward current density in both cell types.This effect was significant even at the lowest concentrationtested in the LV, i.e., 5 µmol/L Ba2+ reduced outward currentto 6.7 ± 0.2 pA/pF (P < 0.05). Asterisks indicate concentra-tions at which the outward current in LV myocytes was sig-nificantly larger than in RV myocytes. At 50 µmol/L Ba2+,differences between outward currents disappeared (LV 2.8± 0.3 pA/pF; RV 2.5 ± 0.2 pA/pF). Panel D shows an ex-ample of the effect of 10 µmol/L Ba2+on action potentialconfiguration. In four myocytes from two hearts, APD90 wassignificantly prolonged with respect to control (23.4% ± 4%,P < 0.004), whereas the resting membrane potential (control−79.6 ± 1.5 mV vs Ba2+ −78.6 ± 1.2 mV, P = NS) and max-imum upstroke velocity (control 274 ± 40 V/sec vs Ba2+ 275± 33 V/sec, P = NS) remained unchanged.

Effect of IK1 Blockade on VF Excitation Frequency

Previous experiments and computer simulations suggestthat IK1 plays an important role in VF dynamics.8,9 We testedthis hypothesis by perfusing the fibrillating heart with Ba2+concentrations that ranged between 1 and 50 µmol/L. PanelA of Figure 3 shows on the left a 2-second ECG episodefrom a representative experiment during VF under baselineconditions. The color DF map of the anterior wall in thisepisode is shown in the center. As demonstrated previously,8


Figure 1. A: Representative RPA showing bands corresponding to Kir2.1, Kir2.3, and cyclophilin (internal loading control) mRNA in the left ventricle (LV)and right ventricle (RV). B: mRNA levels of Kir2.1 (1.97 ± 0.45) and Kir2.3 (1.43 ± 0.08) in LV normalized with respect to RV (n = 12). Horizontal dashedline at 1 indicates equal mRNA level in the LV and RV. C: LV background current at baseline (gray line), after perfusion with 1 mmol/L Ba2+ (black line),and after washout (broken line). D: Ba2+ (1 mmol/L)-sensitive potassium currents in LV (n = 6) and RV (n = 5) myocytes.

although the ECG trace is highly irregular, the map demon-strates clearly demarcated DF domains, with the highest fre-quencies localized in the LV and a sharp LV-to-RV DF gra-dient of 21 Hz. On the right is a single pixel recording fromthe center of the largest LV domain, with its power spectrumshowing a distinct peak at 36 Hz. Panel B shows data obtainedafter 3-minute perfusion with 50 µmol/L Ba2+. The ECG onthe left shows that although the activity remained highly ir-regular, its frequency was much lower than in control. Inaddition, the DF map (center) was much more homogeneousin that the highest DF in the LV was greatly reduced andthe LV-to-RV DF gradient had disappeared. The single pixelrecording on the far right shows that the activity at that sitebecame somewhat more organized and its DF was reducedto 13 Hz.

Quantification of DF Distribution

The distribution of DFs shown in Figure 3 was consistentlyreproduced in all experiments. Panel A of Figure 4 shows themean DF maps (n = 4) of the anterior surface during VF in

baseline (left) and at 3 minutes of 1 µmol/L Ba2+ (right).At baseline, the highest DF domain (34.6 ± 4.3 Hz) waslocalized on the anterior LV; the lowest DF domain (13.5 ±1.7 Hz) was in the RV (P < 0.005). Thus, there was a largeleft-to-right DF gradient of 21.2 ± 4.3 Hz. Perfusion with1 µmol/L Ba2+ reduced slightly the DF, as shown by theselected area (squares) of four pixels within the LV (from31.3 ± 5.2 Hz to 26.6 ± 7.4 Hz) and within the RV (from14.2 ± 1.7 Hz to 11.3 ± 3.5 Hz). As shown in panel B, alarger effect was obtained with 50 µmol/L Ba2+ (n = 5),which changed the peak DF from 31.2 ± 5.2 Hz to 13.0 ±4.0 Hz.

The distribution of DFs at baseline was consistently higherin the LV compared to the RV.8 In the 20 experiments usedto study the effect of Ba2+, the baseline optical data hadan average DF of 30.6 ± 5.3 Hz and 14.6 ± 2.0 Hz in theLV and RV, respectively. There was also a difference in Ba2+dosage dependency between the two ventricles. The decreasein LV DF was highly dependent on the Ba2+ concentration(P < 0.01), as depicted in panel A of Figure 5. In contrast,the Ba2+ effect on the RV was small. As shown in panel B,


Figure 2. A: Voltage dependence of background current of left ventricular (LV) myocytes (n = 21) at baseline and at 10 and 50 µmol/L Ba2+. B: Meancurrent-voltage relationship of right ventricular (RV) myocytes (n = 20) at baseline and at 10 and 50 µmol/L Ba2+. C: Mean concentration-response curvesof peak outward current of LV and RV myocytes for 5 to 50 µmol/L Ba2+. D: Transmembrane action potentials from LV myocyte at baseline and at 10 µmol/LBa2+ (see text for details).

although the DF decreased from 15.8 ± 1.9 Hz to 11.8 ±2.3 Hz at 50 µmol/L, there was no significant concentrationdependence.

The Ba2+-induced reduction in dominant LV frequencystrongly supports the idea that a large outward componentof IK1 is central in the maintenance of high-frequency ro-tors during VF.8 As shown in panel C of Figure 5, Ba2+perfusion resulted in a marked concentration-dependent re-duction of the LV-to-RV frequency gradient. This effect waslargest at 50 µmol/L, to the extent that the gradient prac-tically disappeared (from 15.3 ± 4.2 Hz to 1.2 ± 2.5 Hz,P < 0.01).

To determine the distribution of excitation frequencies,we compared the average number of domains with distinctfrequencies in the absence (baseline) and in the presence ofBa2+. Significant reductions (P < 0.05) in the number of DFdomains were observed at 3 µmol/L (8.0 ± 1.4 to 6.1 ±0.9), 12.5 µmol/L (8.3 ± 1.8 to 6.2 ± 2.1), and 50 µmol/L(8.5 ± 1.6 to 5.2 ± 2.7), but not 1 µmol/L. These resultssupport our contention that the mechanism responsible forthe organization of excitation frequencies into distinct do-mains is related to a heterogeneous, chamber-specific distri-bution in the rectifying properties of the inward rectifyingcurrent.

Effect of IK1 Blockade on Wavelet Organization

Spatiotemporal organization22 in the form of patterns ofrecurrent periodic activity, including rotors (spiral waves),breakthroughs, waves entering the field of view from itsedges, or combinations thereof lasting for >25 cycles, waspresent in 45% of baseline VF episodes. Spiral waves werethe most frequent pattern observed (18% of total). In five VFepisodes, the periodic source lasted the entire 5 seconds ofthe movie. Ba2+ dramatically altered all such patterns. At3 µmol/L or higher, Ba2+ abolished the occurrence of spiralwaves and other type of periodic activity with lifespan largerthan three cycles.

Panel A of Figure 6 (top) shows phase maps obtained con-secutively during baseline VF at three instants of time corre-sponding to a complete rotation of a rotor near the apex of theLV (curved arrow). This rotor underwent >25 cycles of sta-ble, counterclockwise rotating activity at a period of 27 msec(37 Hz). The black circles highlight SPs about which sec-ondary waves were created when waves emanating from themain rotor collided against refractory tissue. Such collisionsusually gave rise to short-lived rotating activity. The maps atthe bottom show consecutive maps obtained at 50 µmol/LBa2+ in the same experiment during a complete rotation of


Figure 3. A: Global ECG (left), anterior wall dominant frequency(DF) map (middle), and single pixel recording with its power spec-trum and DF (right). B: Data from same experiment at 50 µmol/LBa2+ (3 min). Numbers and colors indicate local DFs (in Hz, seecolor scheme). Arrow indicates location of single pixel signal.

a meandering but shorter-lived spiral. The cycle length wasmuch longer (65 msec) and resulted in a frequency of 15 Hz.

To quantify the effects of Ba2+ on the dynamics of thewavelets, we measured the spatial distribution and lifespan

Figure 4. Mean anterior wall dominant frequency (DF) maps at baseline(left) and at 5 minutes 1 µmol/L Ba2+ (A: n = 4) and 50 µmol/L Ba2+ (B:n = 5). Numbers represent the mean DF at the selected areas (open squares)on the left and right ventricles. See text for detailed description.

of SPs on the anterior surface during 500-msec episodes.Panel B of Figure 6 shows spatial histograms depicting thedensity distribution of SPs at baseline (top) and at 50 µmol/LBa2+ (bottom). At the baseline (top), SPs were more denselypacked around the contour of the highest frequency domain(see Fig. 4B). Similar patterns were observed in the other ex-perimental groups under control conditions. The mean num-ber of SPs in the whole field of view was significantly re-duced after perfusion with 50 µmol/L Ba2+ from 1.7 ± 0.9 to0.8 ± 0.6 SP/sec · pixel (P < 0.01) and after perfusion with12.5 µmol/L Ba2+ from 2.3 ± 1.2 to 1.3 ± 1.2 SP/sec pixel(P < 0.01), but not after perfusion with 3 µmol/L Ba2+. Fur-thermore, consistent with the fact that Ba2+ reduced both thenumber of DF domains and the DF gradient, the distributionof SPs became more homogeneous throughout the ventricularsurface.

Ba2+-Induced VF Termination

Panel A of Figure 7 shows the evolution of the normal-ized average maximum DF frequency in each group. Clearly,whereas maximum DF was not significantly altered in threecontrol hearts, 3, 12.5, and 50 µmol/L Ba2+ produced veryrapid (maximum effect ≤5 min) and significant decreasesin DF (P < 0.01). Again, the largest effect occurred at50 µmol/L. Washout with normal Tyrode’s solution partiallyrestored the DF. As illustrated in panel B, the most strikingeffect of Ba2+ was termination of VF in a concentration-dependent manner. Specifically, 50 µmol/L Ba2+ terminatedVF in 5 of 5 experiments within 10 minutes; 1 µmol/L Ba2+did not terminate any of the episodes.

Although 50 µmol/L Ba2+ terminated VF reproducibly,usually sinus rhythm was not restored. In the example shownin Figure 8, VF was replaced by a different, much slowerarrhythmia. Panel A shows on the left an ECG episode ofbaseline VF with its FFT shown on the right. In panel B, at5 minutes, Ba2+ reduced the global frequency from 20 to


Figure 5. Dominant frequency (DF) at various concentrations of Ba2+. DFwas measured in selected areas (squared areas in Fig. 4) of the (A) anteriorleft ventricle (LV) and the (B) anterior right ventricle (RV). C: Difference inDF of LV and RV areas as a function of Ba2+ concentration.

11 Hz just before VF termination. In panel C, 1 minute afterVF termination, there was a very slow (∼4 Hz) and polymor-phic arrhythmia. The isochrone maps on the anterior (top) andposterior (bottom) walls were obtained during three consec-utive excitations and show beat-to-beat change of epicardialbreakthrough site. Similar results obtained in all experimentsindicate that Ba2+ may have induced slow spontaneous dis-charges at subepicardial locations, perhaps at varying siteswithin the His-Purkinje network.24

Discussion

The main new findings of this study are as follow. (1) Inthe guinea pig heart, the levels of mRNA encoding Kir2.1and Kir2.3 channel proteins are significantly greater in the

LV than in the RV, which helps explain the significant dif-ferences in the outward component of IK1 observed in thetwo ventricles. (2) The relative effect of selective blockadeby micromolar concentrations of Ba2+ on the outward com-ponent of IK1 is greater in LV than in RV myocytes. (3) Atthe range of concentrations known to selectively block IK1,Ba2+ perfusion terminates VF in the Langendorff-perfusedheart in a dose-dependent manner. (4) Prior to termination,Ba2+ reduces both high-frequency activity in the LV andthe LV-to-RV frequency gradient. (5) Concurrently, there aredose-dependent reductions in the number of DF domains,phase SPs, and wavelets, indicating that as the VF frequencygoes down, organization increases and larger areas of bothventricles activate at the same frequency. The results providethe first direct evidence for IK1 involvement in the mecha-nism of VF maintenance and demonstrate that pharmacologicblockade of this current can terminate reentry and fibrillation.

Outward Component of IK1 is Larger in LV

In a previous study, Samie et al.8 found that the outwardcomponent of the background current of LV myocytes is sig-nificantly larger than in RV myocytes. Here we demonstratefor the first time that this difference is concurrent with sig-nificantly higher levels of Kir2.1 and Kir2.3 mRNA in theLV compared to the RV. The difference is almost twofold forKir2.1, suggesting that Kir2.1 may play a more significantrole than Kir2.3 in determining outward currents through IK1in LV myocytes (see, for example, Dhamoon et al.25). Closeexamination of the data shown in Figure 1D shows that thereis a difference in IK1 current density in the outward, but notinward, direction. Previous studies have shown modulation ofboth outward and inward IK1.26 However, other studies haveshown modulation of only the inward component of IK1.27

Presently, we do not have an explanation for the differencesobserved in our study on the outward component alone. Nev-ertheless, it is becoming increasingly clear that Kir2.x mayform heteromeric channels. It is possible that the heteromericchannels may be different in the two chambers. Further ex-periments are needed to elucidate the mechanism by whichdifferential expression of the channel isoforms determinesthe chamber specific I-V relationships of IK1.

Whereas Liu et al.11 suggested that Kir2.2 is the largestcontributor to IK1 in the guinea pig heart, we did not findsignificant levels of Kir2.2 or Kir2.4 in LV or RV tissue. Onthe other hand, similar to our results, the mRNA data of Liuet al. showed that Kir2.1, but not Kir2.2, is the most abundantisoform in cardiac myocytes of the ventricles.

Barium Blockade of IK1

A characteristic of inward rectifying potassium channelsis their susceptibility to block by extracellular cations,28,29

and sensitivity to Ba2+ block is a hallmark feature of IK1.16

Voltage-dependent inhibition by Ba2+results from the inter-action of the cation with pore-lining amino acid residues.30-35

However, sensitivity to Ba2+ varies markedly among guineapig Kir2.x channel proteins.11 In the case of Kir2.1, Liu et al.measured a Kd for Ba2+ block of ∼30 µmol/L at −40 mV.However, other investigators have reported Kd values rangingbetween 2.2 µmol/L Ba2+ at −80 mV and 12.7 µmol/LBa2+ at −40 mV.31,35 Accordingly, at 5 µmol/L Ba2+, ourdata show significant decreases in the peak outward currentof IK1 in LV myocytes (Fig. 2C) and the DF of VF in the LV(Fig. 7A). At approximately 30 µmol/L Ba2+, the LV outward


Figure 6. A: Three phase snapshots during a complete rotation of a spiral at baseline (top) and at 50 µmol/L Ba2+ (bottom). Curved arrows show locationand chirality of mother singularity point (SP). Open circles indicate other short-lived SPs. B: Two-dimensional histograms depicting spatial distribution ofSP density at baseline (top) and at 50 µmol/L Ba2+ (bottom). Upper histogram shows contour of fast DF domain of map shown in Figure 4 (red).

Figure 8. Ventricular fibrillation (VF) is replaced by a Ba2+-induced slowectopic arrhythmia. A: ECG pattern (left) and power spectrum (right) duringbaseline VF. B: ECG pattern and power spectrum at 5 minutes of 50 µmol/LBa2+. C: Isochrone maps of anterior and posterior walls during slow (4 Hz)polymorphic arrhythmia, 1 minute after VF termination. Arrows indicateexcitations mapped.

current is reduced to half its maximal value (Fig. 2C). Onthe other hand, the sensitivity of IK1 in the RV myocytes toBa2+ is lower, as demonstrated by the convergence of thepeak outward currents of both LV and RV myocytes and DFsof both ventricles during VF at 50 µmol/L.

Role of IK1 Rectification in VF

Numerical work suggests that maintenance of a stable ro-tor depends on the strong repolarizing influence exerted byits core, which activates IK1 and leads to extreme abbrevia-tion of the APD in its proximity.9 Optical mapping duringVF in the guinea pig heart revealed the existence of a sta-ble high-frequency source consistently localized in the LV,which resulted in large gradients of DF of excitation andfibrillatory conduction from the LV to the RV.8 As shownhere, this frequency distribution correlates with differencesin rectification of IK1. These data strongly support numeri-cal predictions from simulations in which a two-dimensionalmodel of cardiac tissue was used to study the stability of ro-tors in cells whose IK1 I-V relation mimicked that of the LV orRV myocytes. Only the model with an outward component ofIK1 similar to that of LV myocytes was able to sustain stablehigh-frequency spirals.8

Our study shows that IK1 block has a highly signifi-cant effect on the various parameters that characterize VFdynamics. The marked dependence of the decrease of thehigh-frequency activity in the LV on the Ba2+ concentrationstrongly supports the hypothesis that a higher outward com-ponent of IK1 in LV cells allows stabilization of rotors thatact as sources for VF.8 The lesser effect of Ba2+ on IK1 in RVmyocytes is consistent with the small decrease in RV excita-tion frequency observed in all experiments and supports thenotion that a reduced frequency of the VF source in the LVallows for more organized activation of the RV and a reduced


Figure 7. A: Time course of Ba2+-induced changes in left ventricular dom-inant frequency (DF; normalized to the maximum during baseline) and afterwashout. All Ba2+ concentrations produced significant changes (P < 0.05)at 5 minutes, except 1 µmol/L. B: Percentage of hearts in which Ba2+ ter-minated ventricular fibrillation (VF) at each concentration.

degree of fibrillatory conduction. The possibility that Ba2+leads to further slowing of unstable rotors that might exist inthe RV during VF cannot be excluded.

Further supporting the concept that differences in IK1 rec-tification are the underlying mechanism of the spatial gra-dient of excitation frequencies, IK1 block induces a largeconcentration-dependent reduction of the LV-to-RV fre-quency gradient. Specifically, after perfusion with 1 µmol/LBa2+, the LV-to-RV gradient remains unchanged with respectto control (∼15 Hz). This result is not surprising because, at1 to 5 µmol/L Ba2+, there are large differences in the out-ward currents of LV and RV cells (Fig. 2C). By contrast,50 µmol/L Ba2+ abolishes almost completely the LV-to-RVexcitation frequency gradient. Note that at this concentration,the I-V relations of LV and RV myocytes are almost identi-cal (compare panels A and B of Fig. 2). Such a loss in thedispersion of inward rectification also is reflected as a morehomogeneous distribution of DF domains. A significant de-crease in the number of DF domains at ≥3 µmol/L Ba2+indicates a more organized activity within the ventricles,

which also is reflected by a reduction in the number andlifespan of SPs and wavelets.

It is important to note that blockade of outward currentby a given Ba2+ concentration need not be accompanied bya linearly related effect on a VF parameter, particularly be-cause measurements of current and measurements of VF dy-namics are necessarily made under completely different ex-perimental conditions and frequencies. For example, a 10%decrease in outward current (Fig. 2C) may result in a 55%decrease in LV DF (Fig. 5A). Similar to our results, Starmeret al.10 showed in computer simulations a dramatic effecton spiral wave dynamics and ECG pattern (from monomor-phic to polymorphic) when gK1 (IK1 maximal conductance)was varied by only 10%. On the other hand, comparisonof the data in Figure 7B with those of Figure 2C revealsthat the Ba2+ concentration dependence in the number ofVF episodes terminated correlated well with the concentra-tion dependence of changes in the outward component ofIK1. Nevertheless, caution must be exerted when attemptingto correlate changes in outward current with changes withVF dynamics.

Comparison with Other Data on VF

The multiple wavelet hypothesis assumes an underlyingheterogeneity of refractory periods (RPs) leading to frag-mentation of waves. In support of this view, Opthof et al.36

showed that heterogeneous distribution of RPs during pac-ing correlated with activation intervals during VF in the dog.Other studies consider dynamic wave instabilities sufficientto generate spatial variations in refractoriness that may leadto wave fragmentation.5-7,37,38 It has been reported that thereis a distribution of excitation frequencies from the apex to thebase of the guinea pig heart.39 Such a pattern was found to beinversely correlated with the APD measured during pacing,which decreased progressively from base to apex.40 The con-clusion was drawn, therefore, that ionic currents responsiblefor APD heterogeneity were influencing VF dynamics.39 Thesame study reported that there was no preferential directionof propagation of excitation wavefronts. Our results differsubstantially from such data, both in the degree of spatiotem-poral periodicity that was present in 45% of our hearts andthe DF of excitation waves. We report values of up to 40 Hzat baseline and control DF, compared with maximal valuesof 20 Hz reported by Choi et al.39 In our view, the discrep-ancy could be related to different experimental conditions anddefinition of DF.41 In any case, we submit that APD or RPmeasured under pacing conditions is unlikely to reflect APDduring VF, particularly at frequencies of 30 to 40 Hz, as con-sistently observed in the LV.8 Finally, our results are in closeagreement with two previous studies reporting an increaseduniformity in beat-to-beat morphology of epicardial unipo-lar electrograms, as well as a decrease in energy thresholdfor defibrillation in Ba2+ perfused open chest dog hearts.42,43

Although, as suggested by Dorian et al.,42,43 Ba2+ perfusionled to prolongation of RPs, it is likely that the increased elec-trogram uniformity and decreased energy for defibrillationwere the result of reduced VF frequency brought about byreduced IK1 in the outward direction.

Study Limitations

The limitations of the optical mapping technique and sig-nal analysis have been discussed previously.19,20 A further


limitation of this study is that IK1 blockade might indirectlyaffect the dynamics of other currents, particularly at highconcentrations. In this respect, 10 µmol/L Ba2+ did not alterthe reversal potential of IK1 or the action potential upstroke.Therefore, the marked reduction of VF frequencies after per-fusion with 3 or 12.5 µmol/L Ba2+, as well as the increasedwavelet organization after perfusion with 12.5 µmol/L Ba2+,can be attributed primarily to the reduced outward compo-nent of IK1. At 50 µmol/L, Ba2+ reduced current at all volt-ages and led to a slight but significant reduction of the re-versal potential of IK1 (∼9 mV). Although it is unlikely thatthe magnitude of depolarization would significantly reducesodium channel availability, upon termination of VF, Ba2+-induced depolarization and APD prolongation led to spon-taneous ectopic discharges, which resulted in a slow poly-morphic arrhythmias (Fig. 8). These effects are attributableto both the reduction in IK1 and the shift in reversal po-tential.44,45 Finally, Li et al.46 have described a fast (τ 1 ≈7.8 msec) transient outward inward rectifying potassium cur-rent (Ito.ir) in the guinea pig heart. Admittedly, the fast dynam-ics of this current would allow it to contribute to AP repolar-ization during high-frequency VF observed in the guinea pigheart. These authors argue that the molecular identity of thisnovel current is the human K+ channel TWIK-1, becauseamong other characteristics it shares a common sensitivityto Ba2+ block (IC50 = 100 µmol/L). However, the highersensitivity of Kir2.x to Ba2+ block (IC50 between 2.2 and12.5 µmol/L Ba2+) and our data showing half-maximal blockof IK1 (∼30 µmol/L Ba2+) lead us to consider that the maxi-mal changes induced in VF dynamics at concentrations as lowas 3 µmol/L Ba2+ perfusion indicate a preeminent role of IK1in determining high-frequency excitation patterns observedin the guinea pig heart during VF. Finally, it is important tonote that in the experiments presented here and elsewhere,8

myocytes were obtained at random from the RV and LV freewalls. We did not differentiate among subepicardium, suben-docardium, and midmyocardium, and we did not use cellsfrom the septum. Whether there is variability of Kir channelexpression across the wall is the subject of an ongoing studyin this laboratory but is beyond the scope of this report.

Acknowledgments: The authors thank S.F. Mironov, F.H. Samie, and J.Beaumont for fruitful discussions, and J. Jiang and You-Li for technicalassistance.

References

1. Davidenko JM, Pertsov AV, Salomonsz R, Baxter W, Jalife J: Stationaryand drifting spiral waves of excitation in isolated cardiac muscle. Nature1992;355:349-351.

2. Allessie MA, Bonke FI, Schopman FJ: Circus movement in rabbitatrial muscle as a mechanism of tachycardia. III. The “leading cir-cle” concept: a new model of circus movement in cardiac tissue with-out the involvement of an anatomical obstacle. Circ Res 1977;41:9-18.

3. Gray RA, Pertsov AM, Jalife J: Spatial and temporal organization duringcardiac fibrillation. Nature 1998;392:75-78.

4. Witkowski FX, Leon LJ, Penkoske PA, Giles WR, Spano ML, DittoWL, Winfree AT: Spatiotemporal evolution of ventricular fibrillation.Nature 1998;392:78-82.

5. Garfinkel A, Chen PS, Walter DO, Karagueuzian HS, Kogan B, EvansSJ, Karpoukhin M, Hwang C, Uchida T, Gotoh M, Nwasokwa O, SagerP, Weiss JN: Quasiperiodicity and chaos in cardiac fibrillation. J ClinInvest 1997;99:305-314.

6. Karma A: Electrical alternans and spiral wave breakup in cardiac tissue.Chaos 1994;4:461-472.

7. Chen PS, Wolf PD, Dixon EG, Danieley ND, Frazier DW, Smith WM,Ideker RE: Mechanism of ventricular vulnerability to single prematurestimuli in open-chest dogs. Circ Res 1988;62:1191-1209.

8. Samie FH, Berenfeld O, Anumonwo J, Mironov SF, Udassi S,Beaumont J, Taffet S, Pertsov AM, Jalife J: Rectification of the back-ground potassium current: A determinant of rotor dynamics in ventric-ular fibrillation. Circ Res 2001;89:1216-1223.

9. Beaumont J, Davidenko N, Davidenko JM, Jalife J: Spiral waves in two-dimensional models of ventricular muscle: Formation of a stationarycore. Biophys J 1998;75:1-14.

10. Starmer CF, Romashko DN, Reddy RS, Zilberter YI, Starobin J, GrantAO, Krinsky VI: Proarrhythmic response to potassium channel block-ade. Numerical studies of polymorphic tachyarrhythmias. Circulation1995;92:595-605.

11. Liu GX, Derst C, Schlichthorl G, Heinen S, Seebohm G, Bruggemann A,Kummer W, Veh RW, Daut J, Preisig-Muller R: Comparison of clonedKir2 channels with native inward rectifier K+ channels from guinea-pigcardiomyocytes. J Physiol 2001;532:115-126.

12. Nakamura TY, Artman M, Rudy B, Coetzee WA: Inhibition of rat ven-tricular IK1 with antisense oligonucleotides targeted to Kir2.1 mRNA.Am J Physiol 274:H892-H900.

13. Lopatin AN, Nichols CG: Inward rectifiers in the heart: An update onIk1. J Mol Cell Cardiol 2001;33:625-638.

14. Zaritsky JJ, Redell JB, Tempel BL, Schwarz TL: The consequences ofdisrupting cardiac inwardly rectifying K(+) current (I(K1)) as revealedby the targeted deletion of the murine Kir2.1 and Kir2.2 genes. J Physiol2001;533(Pt 3):697-710.

15. DiFrancesco D, Ferroni A, Viscentin S: Barium-induced blockade of theinward rectifier in calf Purkinje fibers. Pflugers Arch 1984;402:446-453.

16. Kubo Y, Baldwin TJ, Jan YN, Jan LY: Primary structure and func-tional expression of a mouse inward rectifier potassium channel. Nature1993;362:127-133.

17. Vaidya D, Tamaddon HS, Lo CW, Taffet SM, Delmar M, Morley GE,Jalife J: Null mutation of connexin43 causes slow propagation of ven-tricular activation in the late stages of mouse embryonic development.Circ Res 2001;88:1196-1202.

18. Delmar M, Michaels DC, Jalife J: Slow recovery of excitability and theWenckebach phenomenon in the single guinea pig ventricular myocyte.Circ Res 1989;65:761-774.

19. Baxter WT, Davidenko JM, Loew LM, Wuskell JP, Jalife J: Technicalfeatures of a CCD video camera system to record cardiac fluorescencedata. Ann Biomed Eng 1997;25:713-725.

20. Berenfeld O, Mandapati R, Dixit S, Skanes A, Chen J, Mansour MC,Jalife J: Spatially distributed dominant excitation frequencies revealhidden organization in atrial fibrillation in the Langendorff-perfusedsheep heart. J Cardiovasc Electrophysiol 2000;11:869-879.

21. Claerbout JF: Fundamentals of Geophysical Data Processing. NewYork: McGraw-Hill, 1976.

22. Chen J, Mandapati R, Berenfeld O, Skanes AC, Jalife J: High-frequencyperiodic sources underlie ventricular fibrillation in the isolated rabbitheart. Circ Res 2000;86:86-93.

23. Samie FH, Mandapati R, Gray RA, Watanabe Y, Zuur C, BeaumontJ, Jalife J: A mechanism of transition from ventricular fibrillation totachycardia: Effect of calcium channel blockade on the dynamics ofrotating waves. Circ Res 2000;86:684-691.

24. Berenfeld O, Jalife J: Purkinje-muscle reentry as a mechanism of poly-morphic ventricular arrhythmias in a 3-dimensional model of the ven-tricles. Circ Res 1998;82:1063-1077.

25. Dhamoon AS, Bagwe S, Guha P, Anumonwo JM, Taffet SM, Jalife J:Differential expression and whole-cell current rectification profiles ofguinea pig Kir2.x channels. Biophys J 2002;82:587a.

26. Murata Y, Fujiwara Y, Kubo Y: Identification of a site involved in theblock by extracellular Mg(2+) and Ba(2+) as well as permeation ofK(+) in the Kir2.1 K(+) channel. J Physiol 2002;544(Pt 3):665-677.

27. Yao JA, Jiang M, Fan JS, Zhou YY, GN: Heterogeneous changes in Kcurrents in rat ventricles three days after myocardial infarction. Cardio-vasc Res 1999;44:132-145.

28. Hagiwara S, Miyazaki S, Rosenthal NP: Potassium current and the effectof cesium on this current during anomalous rectification of the egg cellmembrane of a starfish. J Gen Physiol 1976;67:621-638.

29. Standen NB, Stanfield PR: A potential- and time-dependent blockadeof inward rectification in frog skeletal muscle fibres by barium andstrontium ions. J Physiol 1978;280:169-191.

30. Sakmann B, Trube G: Voltage-dependent inactivation of inward-rectifying single-channel currents in the guinea-pig heart cell membrane.J Physiol 1984;347:659-683.


31. Shieh RC, Chang JC, Arreola J: Interaction of Ba2+ with the pores ofthe cloned inward rectifier K+ channels Kir2.1 expressed in Xenopusoocytes. Biophys J 1998;75:2313-2322.

32. Eaton DC, Brodwick MS: Effects of barium on the potassium conduc-tance of squid axon. J Gen Physiol 1980;75:727-750.

33. Armstrong CM, Taylor SR: Interaction of barium ions with potassiumchannels in squid giant axons. Biophys J 1980;30:473-488.

34. Vergara C, Latorre R: Kinetics of Ca2+-activated K+ channels fromrabbit muscle incorporated into planar bilayers. Evidence for a Ca2+and Ba2+ blockade. J Gen Physiol 1983;82:543-568.

35. Alagem N, Dvir M, Reuveny E: Mechanism of Ba(2+) block of a mouseinwardly rectifying K+ channel: Differential contribution by two dis-crete residues. J Physiol 2001;534(Pt 2):381-393.

36. Opthof T, Misier AR, Coronel R, Vermeulen JT, Verberne HJ, FrankRG, Moulijn AC, van Capelle FJ, Janse MJ: Dispersion of refractorinessin canine ventricular myocardium. Effects of sympathetic stimulation.Circ Res 1991;68:1204-1215.

37. Panfilov AV: Spiral breakup as a model of ventricular fibrillation. Chaos1998;8:57-64.

38. Garfinkel A, Kim YH, Voroshilovsky O, Qu ZL, Kil JR, Lee MH,Karagueuzian HS, Weiss JN, Chen PS: Preventing ventricular fib-rillation by flattening cardiac restitution. Proc Natl Acad Sci USA2000;97:6061-6066.

39. Choi BR, Liu T, Salama G: The distribution of refractory periods influ-ences the dynamics of ventricular fibrillation. Circ Res 2001;88:E49-E58.

40. Laurita KR, Girouard SD, Rosenbaum DS: Modulation of ventricularrepolarization by a premature stimulus. Role of epicardial dispersion ofrepolarization kinetics demonstrated by optical mapping of the intactguinea pig heart. Circ Res 1996;79:493-503.

41. Berenfeld O, Pertsov AM, Jalife J: What is the organization of wavesin ventricular fibrillation? Circ Res 2001;89:E22.

42. Dorian P, Penkoske PA, Witkowski FX: Order in disorder: effectof barium on ventricular fibrillation. Can J Cardiol 1996;12:399-406.

43. Dorian P, Witkowski FX, Penkoske PA, Feder-Elituv RS: Barium de-creases defibrillation energy requirements. J Cardiovasc Pharmacol1994;23:107-112.

44. Malecot C, Coraboeuf E, Coulombe A: Automaticity of ventricu-lar fibers induced by low concentrations of barium. Am J Physiol1984;247:H429-H439.

45. Miake J, Marban E, Nuss HB: Gene therapy: Biological pacemakercreated by gene transfer. Nature 2002;419(6903):132-133.

46. Li GR, Yang B, Sun H, Baumgarten CM: Existence of a transientoutward K(+) current in guinea pig cardiac myocytes. Am J Physiol2000;279(1):H130-H138.

blockade of the inward rectifying potassium current terminates ventricular fibrillation in the...

Documents