Mechanisms of shock-induced arrhythmogenesis. during acute global ischemia

Transcription

1 AJP-Heart Articles in PresS. Published on February 21, 2002 as DOI /ajpheart Mechanisms of shock-induced arrhythmogenesis during acute global ischemia by Yuanna Cheng, 1 Kent A. Mowrey, 1 Vladimir Nikolski, 1,2 Patrick J. Tchou, 1 Igor R Efimov. 1,2 from 1 Department of Cardiovascular Medicine, Cleveland Clinic Foundation, Cleveland, Ohio Department of Biomedical Engineering, Case Western Reserve University, Cleveland, Ohio Running title: Shock-induced arrhythmogenesis in ischemia This work was presented in part at the 73 th Annual Scientific Session of the American Heart Association and published in abstract form in Circulation: 2000, 102 (18): II-339. Corresponding author: Yuanna Cheng Department of Cardiovascular Medicine Desk FF1-06 Cleveland Clinic Foundation 9500 Euclid Avenue Cleveland, OH Phone: (216) Fax: (216) chengy@ccf.org Copyright 2002 by the American Physiological Society.

2 Cheng et al, Shock-induced arrhythmogenesis in ischemia 2 ABSTRACT Little is known about the mechanisms of vulnerability and defibrillation under ischemic conditions. We investigated these mechanisms in 18 Langendorff-perfused rabbit hearts during 75% reduced-flow ischemia. Electrical activity was optically mapped from the anteriorepicardium during right ventricular shocks applied at various phases of cardiac cycle while an excitation-contraction de-coupler 2,3-butanedione monoxime (BDM, 15 mm) was used to suppress motion artifact caused by contraction of the heart. During ischemia, vulnerable window width increased (30-90% of action potential duration (APD) in control to % of APD in ischemia). Moreover, arrhythmias severity increased along with the reduction of APD (176 9 ms in control, ms in ischemia, p<0.01) and increased dispersion of repolarization (45 17 ms in control, ms in ischemia, p<0.01). Shock-induced virtual electrode polarization was preserved. Depolarizing (contrary to hyperpolarizing) response time-constants increased. Virtual electrode-induced wavefronts of excitation had much more tortuous pathways leading to wavefront fractionation. Defibrillation failure at all shock strengths was observed in 4 hearts. Optical mapping revealed that the shock extinguished arrhythmia, however, the arrhythmia selforiginated after an isoelectric window of msec. Conclusions: In most cases, virtual electrode-induced phase singularity (VEIPS) was responsible for shock-induced arrhythmogenesis during acute global ischemia. Enhancement of arrhythmogenesis was associated with an increased dispersion of repolarization and altered de-excitation. In 4 hearts, arrhythmogenesis could not be explained by VEIPS. Keywords: cardiac vulnerability, defibrillation, ischemia, voltage sensitive dye, optical mapping

3 Cheng et al, Shock-induced arrhythmogenesis in ischemia 3 INTRODUCTION Uncovering the mechanism by which strong electric shocks extinguish life-threatening arrhythmias has been challenging researchers for many years since its discovery in 1899 (34). The last decade of research resulted in a significant improvement of our understanding of the basic mechanisms of defibrillation summarized in the Virtual Electrode Polarization Hypothesis of Defibrillation, which offers new insight into the effects of the electric shocks on the myocardium (15). This success was primarily due to the earlier theoretical and experimental advancements, which resulted in the formulation of the bidomain formalism of cardiac syncytium (19, 28, 44) and the fast fluorescent mapping of electrical activity in the heart (10). The virtual electrode polarization hypothesis of defibrillation (15) is based upon numerous theoretical (18, 39,41, 43) and experimental (6,13, 14, 20, 26, 29, 50) studies. According to this hypothesis, the shock induces both positive and negative changes in pre-shock transmembrane potential. The success or failure of the shock is determined by this shock-induced polarization pattern. The induction of shock-induced reentrant arrhythmia is via a virtual electrode-induced phase singularity mechanism (VEIPS). However, nearly all of these studies investigated defibrillation/vulnerability in the normal non-ischemic myocardium. Defibrillation/vulnerability under ischemic conditions was not intensively studied, despite the fact that up to 70% of implantable cardioverter defibrillator patients have some form of coronary artery disease (21, 53). Behrens et al (3) presented evidence of increased vulnerability during external shocks associated with increased heterogeneity of ventricular repolarization in acute global ischemia. Yet the defibrillation threshold was not affected by acute myocardial ischemia in their study. Knisley and Holley (24) characterized the response of the ischemic tissue to the electric shock and reported observations of virtual electrodes under ischemic conditions. Yet, the role of virtual electrodes in defibrillation failure and shock-induced vulnerability in ischemia remains unclear.

4 Cheng et al, Shock-induced arrhythmogenesis in ischemia 4 Our goal was to investigate shock-induced vulnerability and defibrillation under the conditions of acute global ischemia produced by reduced flow in the Langendorff-perfused rabbit heart using voltage-sensitive dye and imaging techniques.

5 Cheng et al, Shock-induced arrhythmogenesis in ischemia 5 METHODS Experimental preparation Langendorff-perfused hearts (n=18) from young rabbits (age: days) were used in the present study. A detailed description of heart preparation has been previously published (12-14). Briefly, the rabbit was anesthetized with nembutal then the heart was removed and placed on a modified Langendorff apparatus for retrograde perfusion. A custom made 10-mm platinum coil electrode (Guidant Corp. MN) was inserted into the right ventricular cavity through the pulmonary artery. A second similar electrode was positioned in the bath 1-2 cm above and 1-2 cm behind the heart. The heart was stained with a gradual injection of 350 L of stock 1.25 mg/ml solution of voltage-sensitive dye di-4-anepps (Molecular Probes, OR) in DMSO (Dimethyl Sulfoxide, Fisher Scientific, NJ) over minutes. An excitation-contraction de-coupler 2,3- butanedione monoxime (BDM, Fisher Scientific) (22) was added to the perfusate with the following composition (mm): BDM 15, NaCl 128.2, CaCl 2 1.3, KCl 4.7, MgCl , NaH 2 PO , NaHCO 3 25, and glucose 11. Global acute ischemia was produced by rapid reduction of the flow rate by 75 % from 20 ml/min in control to 5 ml/min. Ischemia-induced electrophysiologic changes were continuously monitored with electrograms and optical imaging. A relatively steady state of action potential duration (APD) was reached between minutes of flow reduction in the hearts. Therefore, data collection was performed after 30 minutes of flow reduction during ischemia and data was compared with control data collected at the normal flow rate of 20 ml/min. Optical mapping The fluorescence was excited by a DC-powered light source at nm. The emission was collected above 610 nm by a 16x16 element square matrix of photodiodes, coupled to a computerized data conditioning and acquisition system. Data was filtered at 1 khz and

6 Cheng et al, Shock-induced arrhythmogenesis in ischemia 6 sampled at a rate of 1894 frames per second, yielding temporal resolution of 528 sec. The field of view was 16 x 16 mm in all experiments. Optical action potentials (AP) were recorded before, during and after the application of shocks. Typical data scans were 1-2 sec and included the last basic beat AP, the onset of the next AP, and the shock-induced response. A single lead electrocardiogram was recorded using two Ag/AgCl probes placed ~1 cm from the right and left side walls of the glass chamber relative to a Ag ground placed at the bottom of the chamber. This lead configuration produced ECG qualitatively similar to the lead I of the body surface ECG. Experimental protocol The heart was positioned in a temperature-controlled, water-jacketed glass chamber with the anterior wall facing the optical apparatus. Figures 1-3 show typical fields of view (square line, 16 x 16 mm) seen by the photodiode array. The heart was paced at a basic cycle length (CL) of 300 ms by electrical stimuli of twice the diastolic pacing threshold strength and 2 ms stimulus duration from the apex of the heart. Truncated exponential monophasic shocks of 8 ms in duration were delivered from a 150 F capacitor defibrillator HVS-02 (Ventritex, CA) between the two electrodes described above. In 13 out of 18 experiments, shocks with 100-V strength were applied at various phases of the cardiac cycle. 100-V shock was chosen because in our previous study, we reported the shock-induced vulnerability in the structurally normal heart and demonstrated that 100-V shock could induce 100% of arrhythmias when applied at the certain phases of APD (52). We extended this investigation under ischemic condition in this study. Triggering of shocks at variable coupling intervals from the last pacing stimulus was performed via a custom pacing program embedded into the data acquisition and analysis program, as previously described (13).

7 Cheng et al, Shock-induced arrhythmogenesis in ischemia 7 Classification of shock-induced arrhythmia In some cases no extrasystole resulted from the shock. Occurrence of one or more extrasystole was defined as shock-induced arrhythmia, which can be further divided into nonsustained or sustained ones. Non-sustained shock-induced arrhythmia was defined as an arrhythmia that self-terminated in less than a minute. Sustained shock-induced arrhythmia was defined as an arrhythmia that lasted more than 1 minute and had a CL less than 160 ms which required a defibrillation rescue shock to terminate it. Time-constants of cellular response to shock A set of experiments was conducted in 5 of 18 rabbits to quantify the transmembrane response to a defibrillation shock; the cellular time constant ( ) was calculated during delivery of the shock in control and during the ischemia. Monophasic shocks (8 ms in duration, 150 F) of 100-V, 130-V, 160-V, 190-V and 220-V were applied at 25%, 50% and 75% of APD. The polarizations were approximated for both control and ischemia with single-exponential fits using the Levenberg-Marquardt method (33). To automate the processing of large amounts of data, a custom program was developed using Microsoft Visual C++ to automatically analyze the data with the ability for manual review and correction. The was calculated only in the virtual electrode areas away from the shock electrode because electroporation created near the shock electrode (1) would contaminate the cellular response, and the virtual electrode areas have been shown to provide the substrate for shock-induced arrhythmogenesis and defibrillation failure (6, 13). Since the shock electrode in this study was always inserted in the right ventricular cavity and seen at the left edge of field of view (see figure 1-3), was calculated from optical traces at the right side of field of view only. To more accurately and objectively define the areas of virtual electrodes at a distance from the shock electrode and thus those traces accepted for analysis, the following exclusion criteria were followed:

8 Cheng et al, Shock-induced arrhythmogenesis in ischemia 8 1. The virtual electrode polarization area directly above the shock electrode was not included in the analysis to avoid contamination of cellular responses by electroporation. 2. To improve the fidelity of measurement, only strong transmembrane responses to the shock (amplitude > 10 mv) were considered. 3. Also to improve fidelity of measurement, only traces with a signal-to-noise ratio above 75 were considered. A total of individual responses to a total of 300 shocks (150 shocks in control and 150 shocks in ischemia) were included in the analysis: during control and during ischemia from 5 hearts. Data analysis and visualization The signal analysis software programs used in this study were previously described (13). These programs automatically calculated from all 256 optical recordings maps of activation (37), repolarization (16) and action potential duration (APD) (52). Briefly, activation, repolarization and APD were calculated as following: For each optical channel we defined depolarization (activation) time as the time difference between the pacing stimulus and the maximum of the first derivative of the action potential upstroke. Assuming the base potential as 0% and maximum potential as 100% we defined repolarization time as the time difference between the pacing stimulus and the time when an optical signal repolarized to 10% level. APD was calculated as a difference between repolarization and depolarization times. Thus the changes in resting potential and action potential amplitude (APA) brought by global ischemia will not affect calculated time maps of activation, repolarization and APD. Since APD was calculated using an APD 90 definition described above, all APD 90 values will be referred to as APD throughout this report. For transmembrane potential voltage maps, we assigned 0% for the resting potential and 100% for the maximum APA and expressed all voltage maps relative to the last basic beat for each individual channel (% APA). Activation time (AT) was subtracted from coupling interval (CI), defined as

9 Cheng et al, Shock-induced arrhythmogenesis in ischemia 9 the time difference between the stimulus and the shock application, in order to calculate percent APD in each channel at which the shock was applied, according to the following formula: %APD = (CI AT)/APD x 100. The vulnerable window was defined by excluding the coupling interval at which arrhythmia incidence was significantly (p<0.05) below 50% (52). Dispersion of repolarization was defined as the difference between the shortest and longest repolarization times across the field of view (3). The gradient of the transmembrane potential and its derivative was calculated using five-point algorithm similar to that used in our previous report to calculate the conduction velocity (37). An isoelectric window was calculated from the field of view as a delay between the start of the shock application and the upstroke of 1 st shock-induced response. Figure 1 illustrates three main types of analysis used in our study. The left panel shows a photograph of the heart superimposed with an isochronal map (orange map) of post-shock activation during arrhythmogenesis. The time scale starts at the time of shock withdrawal. This approach is limited by its two-dimensional nature, being able to represent only a single sweep of a wavefront across a field of view. Therefore, a three-dimensional stack plots was also used to visualize multiple rotations of reentry of the post-shock activation in one figure. To build such plot, each of the shock-induced response from 256 optical channels was low-pass filtered at 50 Hz, differentiated, and normalized relative to the last basic beat response (dv m /dt) max from the same recording site. The horizontal x-y plane represented the field of view and the vertical z- dimension corresponded to time, increasing from top to bottom. Using IDL software (Research Systems, Boulder, Co), 3-dimensional volumes were built by stacking the 2-dimensional plots of normalized derivatives of optical recordings. Then, an isosurface was constructed using a density threshold value of 30%. The shown stackplot visualizes the first two reentrant beats of the arrhythmia, one of which is also shown in the left panel as a conventional isochronal map. Arrhythmogenesis was better understood with maps of transmembrane potential at the end of the shock, shown at the top of the stack-plot. This map revealed the virtual electrode polarization

10 Cheng et al, Shock-induced arrhythmogenesis in ischemia 10 developed by the shock. In this case the right ventricle was positively polarized relative to the pre-shock potential, while the left ventricle was negatively polarized. Statistical analysis Group data was expressed as mean values ± standard deviations (SD). Statistical comparisons were performed using the paired or unpaired t-test. Differences were considered significant when p < 0.05.

11 Cheng et al, Shock-induced arrhythmogenesis in ischemia 11 RESULTS Effects of acute global ischemia on the electrical activity Figure 2 shows a photograph of the experimental preparation used in the present study (A) and representative optical recordings of action potentials during different phases of acute ischemia (B). During ischemia, APD progressively shortened and became more triangular than that in control recordings. Figure 2C shows reduction of the mean APD during ischemia over the time. Shown data represent an average of 256 optical recordings collected from the preparation shown in this figure. Note after min of ischemia, APD reduction became more or less stabilized. Figure 3 shows maps of activation, repolarization and APD during control and ischemia from another representative preparation. The activation pattern remained similar with only a slight slowing of the conduction in ischemia compared to that in control. The average conduction time across the filed of view (minimum activation time subtracting from maximum activation time) was ms and ms (p=0.061, 256 channels x 13 rabbits) in control and after 30 minutes of ischemia, respectively. In contrast, repolarization underwent significant changes during acute ischemia. The average repolarization time was significantly reduced. Repolarization in control was ms. Repolarization after 30 minutes of ischemia was ms (p<0.01, 256 channels x 13 rabbits). Dispersion of repolarization increased (p<0.01, ms in control, ms in ischemia, 256 channels x 13 rabbits). The repolarization pattern became highly heterogeneous as can be easily seen in the repolarization and APD maps in figure 3. Furthermore, in all hearts (n=13), we found a significant (p<0.01) reduction of APD 90 : ms in control, ms after 30 min of ischemia. These findings confirm previous observations during acute ischemia in various species (3, 11, 38, 40, 46, 47).

12 Cheng et al, Shock-induced arrhythmogenesis in ischemia 12 Shock-induced vulnerability is enhanced by acute ischemia Figure 4 shows the incidence of shock-induced arrhythmia provoked by 100-V, 8-ms monophasic shocks delivered at various phases of APD. In this case we considered all types of arrhythmia, including sustained and non-sustained. In control, arrhythmia incidence was observed mainly during the T-wave (APD>50%), while in ischemia it was evident at any phase of APD. The width of vulnerable window increased from 30-90% of APD in control to % of APD in ischemia. Average incidence was 36% and 61%, in control and ischemia, respectively (p<0.01). These data summarized a total of 307 and 152 shocks from 13 hearts under control and ischemic conditions, respectively. Noteworthy, out of all shock-induced arrhythmias, average incidence of sustained arrhythmias were significantly different: 16% and 53%, in control and ischemia, respectively (p<0.01). Thus, in control in this preparation, arrhythmias were primarily self-terminating, lasting less than 1 minute. In contrast, in ischemia arrhythmias were primarily sustained and required defibrillation. Virtual electrode polarization and de-excitation in acute ischemia Figure 5 shows contour maps of the transmembrane voltage at the end of 100-V, 8-ms shocks applied at 15%, 40% and 60% of APD juxtaposed with isochronal maps of post-shock activation under control and ischemic conditions. All maps were recorded from the same field of view in the representative heart, shown in figure 2. A -100-V cathodal shock produced the virtual electrode polarization pattern with an area of positive polarization near the shock electrode (virtual cathode) and adjacent area of negative polarization (virtual anode). No significant qualitative differences were evident in the virtual electrode polarizations patterns at the end of the shock between control and ischemia applied at the same phase of APD. However, there was a difference in the occurrence of the post-shock break-excitation. In control, as we have recently described (52), a shock applied during absolute refractory period (upper-left panel: 15% of APD) failed to produce de-excitation in the area of negative

13 Cheng et al, Shock-induced arrhythmogenesis in ischemia 13 polarization (virtual anode). This was due to the all-or-none repolarization effect (30;49). The term de-excitation is used to describe the situation in which the negative polarization is strong enough to abolish the action potential and restore excitability. In most cases the lack of an extensive de-excited region precluded break-excitation from occurring. In contrast, shocks applied late in the APD (upper-right panel: 60% of APD) de-excited large region of the heart. This de-excitation provided the substrate for formation of a wide wavefront of break-excitation, which resulted in a reentrant arrhythmia via VEIPS. Shocks applied during the plateau phase (upper-middle panel: 40% of APD) were able to de-excite a small region at the apex (lowermiddle part of the field of view) providing a substrate for a small wavefront of break-excitation. This wavefront barely survived propagating in a fractioned fashion, between non de-excited regions. In ischemia, shock-induced de-excitation appeared at any phase of APD, and occupied a larger area available for post-shock break-excitation. As evident from activation maps, breakexcitation was induced at any phase of shock application and resulted in arrhythmia. Figure 6 further illustrates the genesis of wavefront of break-excitation. We have recently suggested that wavefront of break-excitation arises when two conditions are met: (a) sufficient de-excitation is produced (6) and (b) critical transmembrane voltage gradient at the boundary between virtual anode and virtual cathode is reached (7). Figure 6 shows the distribution of the amplitude of transmembrane voltage gradient immediately after the shock (within 0.5 ms). These panels represent the same data as in the corresponding panels of figure 5. As seen in the six upper panels, both the control and ischemia transmembrane voltage gradients reached similar amplitudes in the same area between the virtual anode and virtual cathode. Yet, wavefronts of break-excitation (lower six panels) were not generated in all cases. Under control conditions, a decrease in the prematurity of shock application resulted in a decrease of the width of the wavefront until no wavefront was generated (control panel: 15% of APD). No significant delay was observed between shock withdrawal and wavefront generation, which remained within 20

14 Cheng et al, Shock-induced arrhythmogenesis in ischemia 14 ms. In contrast in ischemia, a wavefront was generated at any phase. However, its first appearance was delayed with decrease in prematurity of shock application from 12 ms to 53 ms. The large isoelectric window (53 ms), defined as a delay between shock and epicardial excitation, could be the result of a delayed breakthrough of a intramural propagation of virtual electrode induced scroll waves that occurs inside of the bulk of the myocardium. There was also a difference in the kinetics of the responses to the shocks. Figure 7 shows time-constants of shock-induced responses for different shock polarities, strengths and times of application. Virtual cathode polarization (depolarizing cellular responses) was consistently faster in control compared with ischemia for almost all strengths and times of application. However, statistical significance was not reached for virtual anode area (hyperpolazing cellular responses). Observation of defibrillation failure during ischemia due to self-fibrillation: evidence of isoelectric window. In a majority of hearts (9 out of 13) under ischemic conditions, mechanisms of shockinduced arrhythmogenesis appeared similar to that under control conditions. Namely, arrhythmia was induced by the virtual electrode induced phase singularity mechanism (13). Figure 8 illustrates a representative example. This record was performed from the same heart as in figure 1. Notice that the time scale is the same as in the stack plot of the figure 1. Comparison of these two figures illustrates similarities between the two types of arrhythmogenesis under control and ischemic conditions: both are reentrant waves rotating in clock-wise direction. However, this comparison also illustrates significant differences. Under ischemic conditions the pathway was significantly more tortuous and discontinuous compared with control. This was due to shortening of APD, significant dispersion of repolarization, and local deceleration of conduction, which caused a higher degree of fractionation of conduction wavefronts. However, in several hearts (4 out of 13) in addition to the virtual electrode induced phase singularity mechanism of arrhythmogenesis, there were also defibrillation failures. These cannot

15 Cheng et al, Shock-induced arrhythmogenesis in ischemia 15 be explained by this mechanism, because of a significant isoelectric window or delay between the shock and epicardial excitation during failed defibrillation (5). Interestingly, this type of defibrillation failure was observed in ischemic hearts only after a number of applied shocks, which could potentially have caused additional damage to the heart. Repeated attempts to rescue these hearts led to an acceleration of re-fibrillation with shortening of isoelectric window ( ms in average). These 4 hearts could not be rescued and the experiments were eventually terminated. Figure 9 illustrates optical recordings from one of these 4 hearts. The trace in upper panel shows the fibrillation terminated by a biphasic rescue shock (17) followed by self-refibrillation. No pacing stimulus was applied after rescue shock. An isoelectric window of 450 ms is clearly seen in the optical trace. Such extended delay before self-refibrillation cannot be explained by a delay of break-excitation in the bulk of myocardium to the epicardium. Seven expanded superimposed traces in lower panel show 1 st and 2 nd beats following the isoelectric widow. They were taken from depolarized region shown in figure 10 marked as 1-7. The first post-shock beat (left column of figure 10) appeared near the left-ventricular apex and spread normally in the right side of the field of view, but significantly slower in the left side of the field of view (next to the shock electrode and selected with a circle). Maps of repolarization and APD50 revealed that the area of slow conduction was also an area of delayed repolarization. The second beat (middle column of figure 10) revealed that conduction delay in this area doubled. Maps of repolarization and APD50 showed that this small area was significantly depolarized (optical traces from this region are shown in lower panel of fig. 9) and remained unexcitable when the third beat arrived. Stack plot (right column) shows that the third wavefront fractionated at this island of depolarized tissue (red arrow in the stack plot) and formed reentry, which rapidly deteriorated into fibrillation. The movie in the online data supplement illustrates in detail the onset of fibrillation via initial formation of reentry and further deterioration via its break-up.

16 Cheng et al, Shock-induced arrhythmogenesis in ischemia 16 DISCUSSION In the present study we confirmed that global acute ischemia produces a significant reduction of APD and increase of dispersion of repolarization in the whole rabbit heart, which is in agreement with findings from others (3, 11, 38, 40, 46, 47). Also in an agreement with the observations by Behrens et al (3) during externally delivered shocks, we demonstrated that ischemia resulted in a significant increase of vulnerability to shock-induced arrhythmias during internal defibrillation-strength shocks. This was evident from widening of the vulnerable window (from 30-90% of APD in control to % of APD in ischemia) and increased propensity of sustained arrhythmias (16% in control versus 53% in ischemia). Besides ventricular repolarization heterogeneity as an important contributing factor, we speculate that an enhanced susceptibility to de-excitation due to ischemia might also contribute to such dramatic difference in vulnerability. Normally, de-excitation during the absolute refractory period is of an all-or-nothing nature, while during the relative refractory period it is of a gradual nature (23, 49, 52). Only the latter provides the substrate for arrhythmogenesis, while the former is antiarrhythmic. That explains the vulnerable window in the normal condition. Ischemia makes it possible to de-excite cells in a gradual fashion during nearly the entire refractory period, which leads to a possibility of arrhythmogenesis at any coupling interval of shock application in ischemic hearts. However, additional studies are required to explore this hypothesis at cellular level. Our data indicate (see figures 5, 6 and 8) that during ischemia virtual electrode-induced phase singularity mechanism is responsible for shock-induced arrhythmogenesis in majority of cases: break-excitation wavefront were produced at the areas of maximum gradient between virtual cathode and virtual anode as during control. However, these wavefronts had much more tortuous pathways and increased instability due to increased dispersion of activation and repolarization. These instabilities lead to wavefront fractionation and development of ventricular fibrillation.

17 Cheng et al, Shock-induced arrhythmogenesis in ischemia 17 The virtual electrode polarization pattern remained grossly similar in ischemia compared with that in control (figure 5). However, there was a different effect of ischemia on depolarizing and hyperpolarizing cellular responses ( + / - ), with + being increased by ischemia and - unaltered. We speculate that increased + in ischemia might be related to increase in intracellular resistance caused by gap junction uncoupling (32, 36, 25, 51). Resulted slowing conduction of wavefront of activation is considered as a risk factor for arrhythmogenesis, which may contribute to defibrillation failure of electric shock. However, the exact cellular mechanism of the difference in + / - during ischemia remains unknown. Further pharmacological studies are required to explore the ionic currents involved. Furthermore, we observed persistent self re-fibrillation in 4 hearts after a number of shocks. Presence of a significant isoelectric window ( ms) indicates that the refibrillation was not due to shock-induced break excitation. We observed that in these hearts conduction in the area adjacent to the shock electrode was significantly slowed, which contributed to post-shock wave fractionation and reentry, and deteriorated into fibrillation (figure 10). Presence of large isoelectric windows (> 300 ms) observed only under ischemic conditions and only in 4 hearts cannot be explained by intramural propagation of virtual electrode induced scroll waves as under normal and some ischemic condition (lower-left panel, figure 6) which was < 60 ms (8). Noteworthy that this type of failure was not observed until a certain number of shocks was delivered to the ischemic heart. We never observed this type of failure and large (> 60 ms) isoelectric window in non-ischemic hearts. This suggests that shock-induced damage superimposed with ischemia provides the substrate for either focal or reentrant mechanism of arrhythmogenesis (or both) not directly related to shock-induced break-excitation. Additional studies are required to elucidate the exact mechanisms of this type of defibrillation failure.

18 Cheng et al, Shock-induced arrhythmogenesis in ischemia 18 Study limitations 1. Shock-induced arrhythmias are complex 3-dimensional phenomena. Therefore, the 2- dimensional mapping technique used in this study provides only limited insights into the mechanism. For example, our ability to reveal the mechanism of isoelectric window is limited. Nevertheless, our data strongly indicate that the observed phenomenon can be extended to the 3- dimensional situation. Unfortunately, no experimental technique is available at present to assess the 3-dimensional map of electrical activity. Computer simulations may provide the missing link. 2. We used excitation-contraction uncoupler 2,3-butanedione monoxime (BDM) in this study. On one hand, BDM is known to have antifibrillatory effect (27, 35) and the protective effect of BDM from ischemia/reperfusion injury is also well known (2, 4, 45). These effects of BDM could bias our estimates of vulnerable window toward smaller degree. However, even these protective effects of BDM exist, our study indicates that fibrillation was readily inducible under ischemia while it was not readily inducible under control conditions. On the other hand, BDM is a general phosphatase acting on numerous proteins (9, 31, 42). It is reported that BDM has various other effects. For example, a recent report suggests that the Na/Ca exchanger is also inhibited by BDM by unknown mechanisms, other than phosphatase activity (48). Thus caution should be taken in interpretation of present data from this study. 3. The present study was performed under experimental constraints. During ischemia, the vulnerable window and the value of were determined over period of time during which there were ongoing changing electrophysiologic states of the ventricular myocardium. It is very difficult, if not impossible, to keep myocardial ischemia at equilibrium over time. Measurements were begun 30 min after the onset of ischemia because the APD duration was significantly altered at this time and tended toward a quasi equilibrium for the subsequent period of time. We measured the vulnerable window and in a separate set of experiments in an attempt to reduce the time of data collection during ischemia to a minimum.

19 Cheng et al, Shock-induced arrhythmogenesis in ischemia In the present study, we investigated the effects of global rather than regional ischemia. Global ischemia affects the heart in a more homogeneous way compared to that of regional ischemia, such as acute myocardial infarction that occurs in a distinct territory of the heart. This may contribute to discrepancies in the exact pattern of virtual electrodes to that of global ischemia. We are presently working on regional ischemia/infarct model to address this limitation. Conclusions: In the present study, we confirmed in our model a well-known effect of ischemia resulting in dramatic reduction of action potential duration along with increased dispersion of repolarization. We observed the widened the vulnerable window and increased severity of arrhythmias during ischemia. The shock-induced virtual electrode polarization pattern remains similar between the control and ischemia. In ischemia, depolarizing response time-constants increased while hyperpolarizing response time-constants unaltered. Virtual electrode-induced phase singularity mechanism was responsible for shock-induced arrhythmogenesis during ischemia in majority of cases. However, in ischemia, virtual electrode-induced wavefronts of break-excitation had much more tortuous pathways in three-dimensional myocardium and increased instability leading to wavefront fractionation. Increased propensity to shock-induced arrhythmias in ischemia is due to the increased dispersion of repolarization and altered deexcitation. We further demonstrated that in some cases persistent self-reinitiating arrhythmia occurred following an isoelectric window under acute ischemic conditions after rescue shocks first terminated the arrhythmia. This was probably due to either the focal or reentrant mechanism of arrhythmogenesis (or both) accelerated by discontinuous conduction and wavefront fractionation, which was not directly related to shock-induced break-excitation.

20 Cheng et al, Shock-induced arrhythmogenesis in ischemia 20 ACKNOWLEDGMENTS The study was supported in part by the National Institutes of Health grant R01-HL59464 (IR Efimov) and the American Heart Association Southern and Ohio Valley Research Consortium ( W to IR Efimov and V to Y Cheng). We thank Brian Wollenzier for expert assistance in experiments and Michael Tchou for assistance in data analysis. We are grateful to Guidant Corporation for providing us with custom-made leads.

21 Cheng et al, Shock-induced arrhythmogenesis in ischemia 21 REFERENCES 1. Al-Khadra AS, Nikolski V, and Efimov IR. The role of electroporation in defibrillation. Circ.Res. 87: , Bauza G, Le Moyec L, and Eugene M. ph regulation during ischaemia-reperfusion of isolated rat hearts, and metabolic effects of 2,3-butanedione monoxime. J Mol.Cell Cardiol. 27: , Behrens S, Li C, and Franz MR. Effects of myocardial ischemia on ventricular fibrillation inducibility and defibrillation efficacy. J.Am.Coll.Cardiol. 29: , Boban M, Stowe DF, Kampine JP, Goldberg AH, and Bosnjak ZJ. Effects of 2,3- butanedione monoxime in isolated hearts: protection during reperfusion after global ischemia. J Thorac.Cardiovasc.Surg. 105: , Chen PS, Shibata N, Dixon EG, Wolf PD, Danieley ND, Sweeney MB, Smith WM, and Ideker RE. Activation during ventricular defibrillation in open-chest dogs. Evidence of complete cessation and regeneration of ventricular fibrillation after unsuccessful shocks. J Clin Invest 77: , Cheng Y, Mowrey KA, Van Wagoner DR, Tchou PJ, and Efimov IR. Virtual Electrode Induced Re-excitation: A Basic Mechanism of Defibrillation. Circ.Res. 85: , Cheng Y, Nikolski V, and Efimov IR. Reversal of repolarization gradient does not reverse the chirality of shock-induced reentry in the rabbit heart. J.Cardiovasc.Electrophysiol. 11: , Cheng Y, Nikolski V, and Efimov IR. Virtual electrode induced transmural and midmyocardial scroll-waves during internal defibrillation shocks. Circulation 102(18), II (Abstract).

22 Cheng et al, Shock-induced arrhythmogenesis in ischemia De Tombe PP, Burkhoff D, and Hunter WC. Comparison between the effects of 2-3 butanedione monoxime (BDM) and calcium chloride on myocardial oxygen consumption. J Mol.Cell Cardiol. 24: , Dillon S and Morad M. A new laser scanning system for measuring action potential propagation in the heart. Science 214: 453-6: Donaldson RM, Nashat FS, Noble D, and Taggart P. Differential effects of ischaemia and hyperkalaemia on myocardial repolarization and conduction times in the dog. Journal of Physiology 353: , Efimov IR, Aguel F, Cheng Y, Wollenzier B, and Trayanova N. Virtual Electrode Polarization in the Far Field: Implications for External Defibrillation. Am J Physiol 279: H1055-H1070, Efimov IR, Cheng Y, Van Wagoner DR, Mazgalev T, and Tchou PJ. Virtual electrodeinduced phase singularity: a basic mechanism of failure to defibrillate. Circ.Res. 82: , Efimov IR, Cheng Y, Biermann M, Van Wagoner DR, Mazgalev T, and Tchou PJ. Transmembrane voltage changes produced by real and virtual electrodes during monophasic defibrillation shock delivered by an implantable electrode. J.Cardiovasc.Electrophysiol. 8: , Efimov IR, Gray RA, and Roth BJ. Virtual Electrodes and De-excitation: New Insights into Fibrillation Induction and Defibrillation. J Cardiovasc Electrophysiol 11: , Efimov IR, Huang DT, Rendt JM, and Salama G. Optical mapping of repolarization and refractoriness from intact hearts. Circulation 90: , Feeser SA, Tang AS, Kavanagh KM, Rollins DL, Smith WM, Wolf PD, and Ideker RE. Strength-duration and probability of success curves for defibrillation with biphasic waveforms. Circulation 82: , 1990.

23 Cheng et al, Shock-induced arrhythmogenesis in ischemia Fishler MG. Syncytial heterogeneity as a mechanism underlying cardiac far-field stimulation during defibrillation-level shocks. J.Cardiovasc.Electrophysiol. 9: , Geselowitz DB and 3. d. Miller WT. A bidomain model for anisotropic cardiac muscle. Ann Biomed Eng 11: , Gillis AM, Fast VG, Rohr S, and Kleber AG. Spatial changes in transmembrane potential during extracellular electrical shocks in cultured monolayers of neonatal rat ventricular myocytes. Circ Res 79: : Gold MR, Peters RW, Johnson JW, and Shorofsky SR. Complications associated with pectoral cardioverter- defibrillator implantation: comparison of subcutaneous and submuscular approaches. worldwide jewel investigators. J Am Coll Cardiol 28: : Gray RA, Jalife J, Panfilov A, Baxter WT, Cabo C, Davidenko JM, and Pertsov AM. Nonstationary vortexlike reentrant activity as a mechanism of polymorphic ventricular tachycardia in the isolated rabbit heart. Circulation 91: , Hall AE and Noble D. Transient responses of Purkinje fibers to non-uniform currents. Nature 199: , Holley LK and Knisley SB. Transmembrane potentials during high voltage shocks in ischemic cardiac tissue. Pacing Clin Electrophysiol 20: (Pt 2): , Kleber AG, Riegger CB, and Janse MJ. Electrical uncoupling and increase of extracellular resistance after induction of ischemia in isolated, arterially perfused rabbit papillary muscle. Circ.Res. 61: , Knisley SB, Hill BC, and Ideker RE. Virtual electrode effects in myocardial fibers. Biophys.J. 66: , Lee MH, Lin SF, Ohara T, Omichi C, Okuyama Y, Chudin E, Garfinkel A, Weiss JN, Karagueuzian HS, and Chen PS. Effects of diacetyl monoxime and cytochalasin D on

24 Cheng et al, Shock-induced arrhythmogenesis in ischemia 24 ventricular fibrillation in swine right ventricles. Am.J.Physiol Heart Circ.Physiol 280: H2689-H2696, Muler AL and Markin VS. [Electrical properties of anisotropic neuromuscular syncytia. I. Distribution of the electrotonic potential]. Biofizika 22: , Neunlist M and Tung L. Spatial distribution of cardiac transmembrane potentials around an extracellular electrode: dependence on fiber orientation. Biophys.J. 68: , Noble D. The Initiation of the Heart Beat. Advancement of Science 23: , Papp Z, van der Velden J, and Stienen GJ. Calpain-I induced alterations in the cytoskeletal structure and impaired mechanical properties of single myocytes of rat heart. Cardiovasc.Res. 45: , Peters NS, Coromilas J, Severs NJ, and Wit AL. Disturbed connexin43 gap junction distribution correlates with the location of reentrant circuits in the epicardial border zone of healing canine infarcts that cause ventricular tachycardia. Circulation 95: , Press WH, Teukolsky SA, Vetterling WT, and Flannery BP. Numerical Recipies in C. Cambridge, Cambridge University Press Prevost, JL and Battelli F. Sur quel ques effets des dechanges electriques sur le coer mammifres. Comptes Rendus Seances Acad Sci 129: 1267, Riccio ML, Koller ML, and Gilmour RF Jr. Electrical restitution and spatiotemporal organization during ventricular fibrillation. Circ.Res. 84: , Saffitz JE, Schuessler RB, and Yamada KA. Mechanisms of remodeling of gap junction distributions and the development of anatomic substrates of arrhythmias. Cardiovasc.Res. 42: , Salama G, Kanai A, and Efimov IR. Subthreshold stimulation of Purkinje fibers interrupts ventricular tachycardia in intact hearts. Experimental study with voltage-sensitive dyes and imaging techniques. Circ.Res. 74: , 1994.

25 Cheng et al, Shock-induced arrhythmogenesis in ischemia Salama G, Kanai A, Huang DT, Efimov IR, Girouard SD, and Rosenbaum DS. Hypoxia and Hypothermia Enhence Spatial Heterogeneiteis of Repolarization in Guinea Pig Hearts: Analysis of Spatial Correlation of Optically Recorded Action Potential Durations. J.Cardiovasc.Electrophysiol. 9: , Sepulveda NG, Roth BJ, and Wikswo JP. Current injection into a two-dimensional anisotropic bidomain. Biophys.J. 55: , Shaw RM and Rudy Y. Electrophysiologic effects of acute myocardial ischemia. a mechanistic investigation of action potential conduction and conduction failure. Circ Res 80: : Sobie EA, Susil RC, and Tung L. A generalized activating function for predicting virtual electrodes in cardiac tissue. Biophys.J. 73: , Stapleton MT, Fuchsbauer CM, and Allshire AP. BDM drives protein dephosphorylation and inhibits adenine nucleotide exchange in cardiomyocytes. Am J Physiol 275: H1260- H1266, Trayanova NA, Roth BJ, and Malden LJ. The response of a spherical heart to a uniform electric field: a bidomain analysis of cardiac stimulation. IEEE Trans Biomed.Eng. 40: , Tung L. A bidomain model for describing ischemia myocardial DC potentials Massachusetts Institute of Technology. (Thesis/Dissertation). 45. Vanoverschelde JL, Janier MF, and Bergmann SR. The relative importance of myocardial energy metabolism compared with ischemic contracture in the determination of ischemic injury in isolated perfused rabbit hearts. Circ.Res. 74: , Vermeulen JT, Tan HL, Rademaker H, Schumacher CA, Loh P, Opthof T, Coronel R, and Janse MJ. Electrophysiologic and extracellular ionic changes during acute ischemia in failing and normal rabbit myocardium. J Mol.Cell Cardiol 28: , 1996.

26 Cheng et al, Shock-induced arrhythmogenesis in ischemia Watanabe I, Kanda A, Engle CL, and Gettes LS. Comparison of the effects of regional ischemia and hyperkalemia on the membrane action potentials of the in situ pig heart. Experimental Cardiology Group, University of North Carolina at Chapel Hill. J.Cardiovasc.Electrophysiol. 8: , Watanabe Y, Iwamoto T, Matsuoka I, Ohkubo S, Ono T, Watano T, Shigekawa M, and Kimura J. Inhibitory effect of 2,3-butanedione monoxime (BDM) on Na(+)/Ca(2+) exchange current in guinea-pig cardiac ventricular myocytes. Br.J Pharmacol. 132: , Weidmann S. Effect of current flow on the membrane potential of cardiac muscle. J Physiol 115: , Wikswo JP, Lin SF, and Abbas RA. Virtual electrodes in cardiac tissue: a common mechanism for anodal and cathodal stimulation. Biophys.J. 69: , Yamada KA, McHowat J, Yan GX, Donahue K, Peirick J, Kleber AG, and Corr PB. Cellular uncoupling induced by accumulation of long-chain acylcarnitine during ischemia. Circ.Res. 74: 83-95, Yamanouchi Y, Cheng Y, Tchou PJ, and Efimov IR. The Mechanisms of Vulnerable Window: The Role of Virtual Electrodes and Shock Polarity. Canadian Journal of Physiology & Pharmacology 79: 25-33, Zipes DP and Roberts D. Results of the international study of the implantable pacemaker cardioverter-defibrillator. a comparison of epicardial and endocardial lead systems. the pacemaker-cardioverter- defibrillator investigators. Circulation 92: 59-65: 1995.

27 Cheng et al, Shock-induced arrhythmogenesis in ischemia 27 Figure Legends Figure 1. Three main types of data visualizations. Left panel: Two-dimensional 5-msec isochronal map (orange) of post-shock activation recorded under control condition. Time-scale shows elapsed time since shock withdrawal. In this case, shock (-100-V, 8-ms) delivered at 60% of APD resulted in a well-defined wavefront formed postshock reentry. The red arrow indicates the direction of propagating wavefront. The experimental preparation and the field of view (black square, 16 x 16 mm) superimposed with post-shock activation map recorded from it are shown at the bottom. Right panel: Two-dimensional 5% isolines map (blue) of transmembrane potential at the end of the same shock. A three-dimensional stackplot visualizing activation wavefronts in the field of view during shock-induced arrhythmogenesis is shown below. See text for details. RV: right ventricle. LV: left ventricle. Figure 2. Experimental preparation and effect of ischemia on action potential duration. A. Photograph of the preparation. The large open box shows the 16x16 mm field of view. The small black box shows a recording site, illustrated in panel B. The bipolar electrode, which was used for pacing, is shown at the apex of the heart. The shock electrode in this study was always inserted in the right ventricular cavity (see red wire), positioned at the left edge of the field of view. B. Representative optical recordings of action potentials during control (0 min), 10, 15 and 30 minutes of low-flow global ischemia. C. Plot of mean action potential duration during the ischemia over the time. Data represent an average of 256 optical recordings.

28 Cheng et al, Shock-induced arrhythmogenesis in ischemia 28 Figure 3. Maps of activation, repolarization and action potential duration during control and ischemia. Left panel: Photograph of the preparation superimposed with a 16x16 mm field of view and a 5- msec isochronal map of activation. Right panel: 5 msec isochrones maps: Rows from top to bottom are (1) activation, (2) repolarization, and (3) action potential duration (APD90). Columns from left to right are (1) control, (2) 15 min after ischemia, and (3) 30 min after ischemia. Figure 4. Incidence of shock-induced arrhythmias versus phase of shock application. Arrhythmias were induced by 100 V / 8 msec monophasic shocks delivered at various phases of APD. See text for more details. Figure 5. Virtual electrode-induced arrhythmogenesis in ischemia versus in normal heart. A monophasic, cathodal shock (-100 V, 8 msec) was applied at 15%, 40% and 60% of APD. Figure shows transmembrane voltage maps at the end of the shock expressed as % APA relative to the last basic beat for each individual channel and post-shock activation maps under both the control and ischemic conditions. Starting times in the each of maps of post-shock activation represent the time elapsed after the shock withdrawal. Figure 6. Relation between postshock transmembrane voltage gradient and wavefront origin. Top panels: postshock distribution of the absolute amplitude of the transmembrane voltage gradient (black band). Bottom panels: the distribution of dv m /dt normalized to the (dv m /dt) max of a last basic action potential upstroke recorded at the same location (gray band). Numbers in ms of each panel represent the times elapsed after shock withdrawal. The panels show the analysis of the same data presented in corresponding panels of figure 5.

29 Cheng et al, Shock-induced arrhythmogenesis in ischemia 29 Figure 7. Summary of cellular responses to the monophasic shocks applied at the different phases of APD. Shocks at the intensities of 100-V, 130-V, 160-V, 190-V, and 220-V were delivered at 25%, 50% and 75% of APD. Star (*) represents significant difference (p<0.05). Figure 8. Stackplot of shock-induced arrhythmogenesis during ischemia. A cathodal monophasic shock (-100-V, 8-ms) induced arrhythmia. Please compare with the stackplot of arrhythmogenesis in control (right-bottom panel in figure 1). Orientation of the field of view, time-scale, and position of the electrode are the same as in figure 1. See text for detail. Figure 9. Optical evidence of "isoelectric window". Upper panel: Representative example of 1 of 4 hearts, in which defibrillation was not successful at any shock strength. The location of the optical trace is shown in figure 10 in dark blue square. The isoelectric window (red line) was 450 ms. A spatio-temporal pattern of the shock-induced arrhythmia (600 ms black arrow) is visualized in figure 10. Lower panel: Seven superimposed traces of arrhythmias following the isoelectric window. These traces were chosen along the depolarized region shown in figure 10 marked as 1-7. Figure 10. Spatio-temporal maps of self-fibrillation following a post-shock isoelectric window. The data represents maps from the optical recordings illustrated in figure 9. Maps of activation, repolarization and APD50 are shown for the first and second non-paced arrhythmia beats. The red circle shows an area of slow conduction, which precipitated in wavefront fractionation and reentry as shown during beat #3 in the stackplot (right panel) with the red arrow. The dark blue square in the right panel of the repolarization map represents the location of the optical trace shown in the upper panel of figure 9. The white bar numbered from 1 to 7 represents the location of seven optical traces shown in the lower panel of figure 9.

30 Cheng et al, Shock-induced arrhythmogenesis in ischemia ms isochrones 5% isolines % %APA 50% 15% Apex Apex LV RV RV Base Time X Y Apex LV RV Base Base 250 ms Base LV 16 mm Figure 1, Cheng et al

31 Cheng et al, Shock-induced arrhythmogenesis in ischemia 31 A B B 100 ms min APD90 (ms) C min Figure 2, Cheng et al

32 Cheng et al, Shock-induced arrhythmogenesis in ischemia 32 Activation Control 15 min ischemia 30 min ischemia ms Repolarization ms APD90 ms Field of view: 16 x 16 mm msec isochrones Figure 3, Cheng et al

33 Cheng et al, Shock-induced arrhythmogenesis in ischemia 33 Control Ischemia Arrhythmia Incidence (%) Phase of shock application (% APD) Vulnerable windows Figure 4, Cheng et al

34 Cheng et al, Shock-induced arrhythmogenesis in ischemia 34 15% of APD 40% of APD 60% of APD 105% 100% 90% Ischemia Control Activation Vm Activation Vm No post shock activation %APA 40% 115% %APA 35% 105 ms %APA 25% 116 ms % %APA 20% 111 ms %APA 15% 125 ms 5 105% %APA 10% 120 ms 5% isolines 5 ms isochrones 5% isolines Figure 5, Cheng et al

35 Cheng et al, Shock-induced arrhythmogenesis in ischemia 35 15% of APD 40% of APD 60% of APD Post-shock transmembrane voltage gradient Control mv/mm Ischemia 0 Post-shock wavefront Control (dv m/dt) (dv m/dt) max ms 14 ms 12 ms 0.5 Ischemia 0 Figure 6, Cheng et al 53 ms 21ms 12 ms

36 Cheng et al, Shock-induced arrhythmogenesis in ischemia 36 (ms) Virtual Cathode * * * * * * * * * * * * * Shock Strengths (Volts) Virtual Anode Control Ischemic % of APD 50% of APD 75% of APD Figure 7, Cheng et al

37 Cheng et al, Shock-induced arrhythmogenesis in ischemia ms Figure 8, Cheng et al

38 Cheng et al, Shock-induced arrhythmogenesis in ischemia ms shock 1 isoelectric window (450 ms) 600 ms st beat 2nd beat 200 ms Figure 9, Cheng et al

39 Cheng et al, Shock-induced arrhythmogenesis in ischemia 39 1st beat 2nd beat 65 Activation Repolarization ms 0 85 ms st beat 2nd beat 600 ms APD50 ms - missing data mm Figure 10, Cheng et al