Diastolic Intracellular Calcium-Membrane Voltage Coupling Gain and Postshock Arrhythmias: Role of Purkinje Fibers and Triggered Activity

Abstract

Rationale:

Recurrent ventricular arrhythmias after initial successful defibrillation are associated with poor clinical outcome.

Objective:

We tested the hypothesis that postshock arrhythmias occur due to spontaneous sarcoplasmic reticulum Ca release, delayed afterdepolarization (DAD), and triggered activity (TA) from tissues with high sensitivity of resting membrane voltage (V_m) to elevated intracellular Ca (Ca_i) (high diastolic Ca_i-voltage coupling gains).

Methods and Results:

We simultaneously mapped Ca_i and V_m on epicardial (n=14) or endocardial (n=14) surfaces of Langendorff-perfused rabbit ventricles. Spontaneous Ca_i elevation (SCaE) was noted after defibrillation in 32% of VT/VF at baseline and in 81% during isoproterenol infusion (0.01~1 μmol/L). SCaE was reproducibly induced by rapid ventricular pacing and inhibited by 3 μmol/L of ryanodine. The SCaE amplitude and slope increased with increasing pacing rate, duration, and dose of isoproterenol. We found TAs originating from 6 of 14 endocardial surfaces but none from epicardial surfaces, despite similar amplitudes and slopes of SCaEs between epicardial and endocardial surfaces. This was because DADs were larger on endocardial surfaces due to higher diastolic Ca_i-voltage coupling gain, compared to those of epicardial surfaces. Purkinje-like potentials preceded TAs in all hearts studied (n=7). I_K1 suppression with CsCl (5 mmol/L, n=3), BaCl₂ (3 μmol/L, n=3), and low extracellular potassium (1 mmol/L, n=2) enhanced diastolic Ca_i-voltage coupling gain and enabled epicardium to also generate TAs.

Conclusions:

Higher diastolic Ca_i-voltage coupling gain is essential for genesis of TAs and may underlie postshock arrhythmias arising from Purkinje fibers. I_K1 is a major factor that determines the diastolic Ca_i-voltage coupling gain.

Postshock arrhythmias, including ventricular tachycardia (VT) and ventricular fibrillation (VF), can occur after initial successful defibrillation shocks. During cardiopulmonary resuscitation, recurrent spontaneous VF occurred in over half of the patients after initial successful defibrillation, and was associated with poor outcomes.¹ Electrical storm, a condition in which multiple temporally related episodes of spontaneous VT/VF occur, is known to be associated with high morbidity and mortality.^{2, 3} The mechanisms of postshock arrhythmias are unclear, although some reports implicate Purkinje fibers as sources of spontaneous VT/VF.^{2, 4} It is well established that delayed afterdepolarizations (DADs) result from Ca-activated transient inward currents (I_ti), which are induced by spontaneous sarcoplasmic reticulum (SR) Ca release.⁵ DADs can be induced in Purkinje fibers and atrial/ventricular myocytes.⁶ However, the vulnerability to DADs is higher in Purkinje fibers than other types of cardiac myocytes.⁷ Possible explanations for the greater vulnerability in Purkinje fibers to DADs include a higher propensity for SR Ca release than working myocardium, or a larger membrane potential (V_m) response to elevated intracellular Ca (Ca_i) during diastole (diastolic Ca_i-voltage coupling gain). The differential coupling gain may produce differential vulnerability to DADs. The purpose of this study was to test the hypotheses that (1) DAD-mediated triggered activities (TAs) secondary to spontaneous SR Ca release underlie the mechanism of postshock arrhythmias, and (2) the high diastolic Ca_i-voltage coupling gain in Purkinje fibers predisposes these cells to DADs which initiate TA after successful defibrillation.

Methods

An expanded Methods section is available in the online Data Supplement.

Hearts of New Zealand White rabbits (n=45) were perfused with 37°C oxygenated normal Tyrode's solution. We simultaneously mapped Ca_i and V_m using Rhod-2 AM and RH237 on epicardial (n=14) or endocardial (n=14) surface of ventricles. In additional hearts, we mapped epicardial surfaces to explore the effect of I_K1 suppression (n=8), and endocardial surfaces to study the effect of ryanodine (n=3) and I_Kr blocker, E-4031(n=2). Transmembrane potential (TMP) was recorded in an additional 4 rabbits. Cytochalasin D (10 μmol/L, n=7) or blebbistatin (10 to 20 μmol/L, n=15) or both (n=23) was used to inhibit contraction during optical and TMP recordings (online Supplement Figure I).

Atrioventricular block was created in all hearts with cryoablation. VF was induced with burst ventricular pacing, and defibrillated with transvenous electrodes. To characterize spontaneous Ca_i elevation (SCaE), we performed ventricular pacing for 100 beats at a pacing cycle length (PCL) of 600, 500, 400, 300, 200 ms, and the minimum cycle length with 1:1 pacing capture. When substantial SCaEs emerged, pacing at the same PCL with different numbers of paced beats (50, 200, 300, and 400) were performed (n=6) to study a dependence of SCaE on the pacing duration. Isoproterenol at various concentrations (0.01 to 1.0 μmol/L) was administered, and defibrillation and pacing protocols were repeated. We defined the amplitude of baseline V_m and Ca_i transient as 1 arbitrary unit (AU). A focal ectopy with DAD at the earliest activation site (the onset of the optical action potential was required to precede the QRS onset of pseudo-ECG if it was available) was defined as TA.

Results

Electrophysiological Characteristics of SCaE

Figure 1A shows pseudo-ECG recording (left panel) and Ca_i optical signal (right panel) in a fibrillation-defibrillation episode. The red arrowhead indicates SCaE, which occurred 286 ms after the shock. We found SCaEs after successful defibrillation in 11/34 (32%) of induced VF episodes (VF duration: 25 ± 20 seconds, shock intensity: 153 ± 80 V). Postshock SCaEs were noted more frequently (38/47, or 81%) during isoproterenol infusion. SCaEs were also induced after cessation of rapid ventricular pacing, without defibrillation shocks. The amplitude and slope of SCaE and the coupling interval from the last paced beat to the onset of SCaE depended on the PCL (Figure 1B). Shorter PCL elicited higher amplitude and slope of SCaE with shorter coupling intervals. Longer pacing durations increased the amplitude and slope of SCaE (Figure 1C). The incidence of SCaE was 0%, 0%, 0%, 5%, 44%, and 61% at baseline and 0%, 0%, 37%, 72%, 98%, and 100% during isoproterenol infusion after 100-beat pacing at a PCL of 600, 500, 400, 300, 200 ms, and minimal 1:1 capture PCL (145 ± 22 ms), respectively. The magnitude of SCaE were enhanced by isoproterenol in a dose-dependent manner (Figure 1D), and the coupling interval of SCaE was shortened with higher dose of isoproterenol (online Supplement Figure II). To determine whether defibrillation shocks affect SCaE, we compared SCaEs with or without S2 shock applied on the refractory period after 100-beat rapid S1 pacing (n=9). The SCaE magnitude was not changed by the S2 shock, indicating intracellular Ca accumulation during tachycardia (rather than shock itself) is responsible for SCaE (online Supplement Figure III). SCaEs were completely suppressed by 3 μmol/L of ryanodine (n=3, online Supplement Figure IV).

Figure 1.

Spontaneous Ca_i elevation (SCaE). The optical tracings in the figure were taken from a pixel where the maximal SCaE was recorded in the mapped region. A, Pseudo-ECG (pECG) recordings during successful defibrillation of VF. Optical tracing of intracellular Ca (Ca_i) revealed type A defibrillation, followed by SCaE (arrowhead). B, Dependence of SCaE on pacing cycle length (PCL). *P < 0.05 vs. 400 ms; **P <0.01 vs. 400 to 600 ms. “1:1” indicates the shortest PCL associated with 1:1 capture. C, Dependence of SCaE on pacing duration at a fixed PCL of 200ms. **P < 0.01 vs. 50 beats. D, A dose-dependent effect of isoproterenol (ISO) on SCaE. **P < 0.01 vs. control (CTR). Bar graphs represent means ± SEM; E = escape beat; CI = coupling interval from the last stimulus to the onset of SCaE.

SCaEs occurred in all preparations during isoproterenol infusion, and were present on both epicardial and endocardial surfaces. The amplitude, slope, and coupling interval of SCaE measured at the sites where the maximal SCaE was recorded were comparable between epicardial and endocardial preparations (amplitude: 0.16 ± 0.04 AU vs. 0.16 ± 0.71 AU, slope: 0.77 ± 0.44 AU/s vs. 0.96 ± 0.41 AU/s, coupling interval: 395 ± 83 ms vs. 345 ± 110 ms, P = NS for all).

Differential V_m Responses to SCaE Between Epicardial and Endocardial Sites

Despite similar SCaEs on the epicardial and endocardial surfaces, changes in V_m accompanying SCaEs were significantly larger on endocardial surface. Figure 2 shows typical examples of V_m responses to SCaE. SCaEs occurred widely throughout the tissue after rapid pacing during isoproterenol infusion. We randomly selected 10 pixels with apparent SCaEs from each cardiac surface to compare V_m changes during the SCaEs (n=8). SCaEs were found on both cardiac surfaces (red arrowheads). However, there was little change in V_m on epicardial surface, while larger V_m elevations consistent with DADs (black arrows) were observed at some pixels on endocardial surface (Figure 2A). The relationship between the magnitude of SCaEs and DADs is shown in Figure 2B. Notably, for the same magnitude of SCaE, DADs were larger on endocardial surface than epicardial surface, although there were significant heterogeneous responses of V_m to SCaE on both surfaces. If we define “diastolic Ca_i-voltage coupling gain” as the ratio of the DAD magnitude to SCaE magnitude, the gain was greater on the endocardial surface (0.13 ± 0.08 (calculated by amplitude) and 0.19 ± 0.13 (calculated by slope)) than the epicardial surface (0.04 ± 0.05 (by amplitude) and 0.05 ± 0.06 (by slope)) (P < 0.001 for both). At baseline, DADs were found in only 1 heart (4%), although SCaEs were noted in 61% of hearts. With beta-adrenergic stimulation, DADs appeared in 21% and 79% of epicardial and endocardial preparations, respectively (P = 0.004).

Figure 2.

Differential membrane potential (V_m) responses to Ca_i changes between epicardial and endocardial surfaces. The black and red lines indicate optical tracings for V_m and Ca_i, respectively. A, After rapid ventricular pacing, there were SCaEs (arrowheads) at multiple sites. Note that the SCaE at epicardium failed to induce significant changes of V_m. In contrast, the similar magnitude of SCaE resulted in delayed afterdepolarization (DAD, arrow) at endocardium. B, The relationship between the amplitude of DAD and SCaE (upper panel) and between the slope of DAD and SCaE (lower panel). RV = right ventricle; LV = left ventricle; S = interventricular septum; PM = papillary muscle.

Figure 3 shows the effect of diastolic Ca_i-voltage coupling gain on genesis of TA. In Figure 3A, SCaEs became progressively larger with progressively shorter PCLs on both cardiac surfaces, but rate-dependent increase in DAD amplitude was observed only on endocardial surface. When the number of paced beats was further increased at the PCL that maintained 1:1 capture, DAD reached an activation threshold and TA developed from endocardial surface. No ectopy occurred on epicardial surface, and an ectopic beat conducted from the outside of the mapped region was seen after the SCaE. Figure 3B illustrates SCaEs after defibrillation with postshock ectopic beats. At the early site in epicardium for the postshock beat (dark blue region), no SCaE occurred (not shown). However, a large SCaE (red arrowhead) was found at a site away from the early site (white cross) with only a tiny DAD (black arrow). Consequently, this site was passively activated. In contrast, postshock SCaE accompanied large DADs on endocardial surface, developing TAs. During isoproterenol infusion, TAs arose from the mapped field of endocardial surface in 6 of 14 hearts (43%), while none of 14 hearts showed TAs from epicardial surface (P = 0.008). In 2 of 6 hearts with TAs, VTs were reproducibly induced by rapid pacing. Figure 4A shows an example of triggered VT. Pacing at the PCL of 300 ms induced a single TA. The TA elicited another SCaE that was weaker than the first SCaE, and could not produce a large enough DAD to cause another TA. When the PCL was shortened to 200 ms, larger SCaEs continuously evoked TA, producing VT. The phase plot analysis of Ca_i against V_m displayed clockwise rotation exclusively at the VT origin (Figure 4B), indicating Ca_i event occurred before V_m event at the VT origin site. The activation propagated to the rest of ventricle where Ca_i elevation was secondary to V_m depolarization and the phase plot analyses showed counterclockwise rotations. Snapshots of ratio images during the VT confirmed this relationship between Ca_i and V_m (Figure 4C). Ca_i elevation was detected before V_m elevation (Ca_i prefluorescence, white arrow). Subsequently, the activation wavefront of V_m spread faster than that of Ca_i.

Figure 3.

Effect of diastolic Ca_i-voltage coupling gain on genesis of triggered activity (TA). A, Simultaneous V_m and Ca_i optical tracings after pacing at various PCLs for the number of pacing beats indicated in parenthesis. Different line colors were used to mark different episodes. The optical tracings were obtained at the pixel where the maximal SCaE was recorded. On endocardial surface, this site coincided with the earliest site of the ectopic beat (PCL 160 ms for 300 beats). B, Ca_i and V_m optical tracings recorded at the maximal SCaE sites (white cross) after defibrillation.

Figure 4.

VT originating from endocardial surface. A, (Left panel) A single ectopy (VPC) induced by 100-beat pacing at a PCL of 300 ms. Optical tracings were obtained at the earliest site of the VPC. Preceding SCaEs (arrowheads) and DADs (upward arrows) suggest DAD-mediated TA as the mechanism of the VPC. (Right panel) At a PCL of 200 ms, VT was induced. The earliest activation occurred before the QRS onset of pseudo-ECG. All VT beats had the same origin as that of the VPC. Gradual disappearance of the VT confirms the mechanism is not reentry. Every SCaE accompanied DAD and TA during the VT, but not TA at the VT termination (asterisk). B, Ca_i elevation preceded V_m elevation only at the VT origin. C, Snapshots of Ca_i and V_m ratio maps at times from 10 ms before to 40 ms after the QRS onset of the VT beat. Note Ca_i prefluorescence at the VT origin (white arrow). E = escape beat.

Single or double ectopic beats were observed after successful defibrillation in 5/34 episodes (15%) at baseline, and 12/47 episodes (26%) with isoproterenol. Post-shock VTs (>3 consecutive beats) occurred during isoproterenol infusion in 13/47 episodes (28%) (VT duration: 5.1 ± 3.9 seconds, 3 monomorphic, 10 polymorphic). We found that at least one beat of the VT arose from the mapped region of endocardial surfaces in 4 episodes of post-shock VT but none from epicardial surfaces.

Heterogeneous Distribution of SCaEs, DADs and TAs

SCaEs were not a local phenomenon, but usually appeared over a large area. The area involving SCaE became larger as PCL was shortened. SCaEs might occupy the entire mapped field for short PCLs during isoproterenol infusion. The magnitude of SCaE varied depending on the site, resulting in a spatial heterogeneity of SCaE (Figure 5A). The spatial distribution of SCaE was independent of pacing site (online Supplement Figure V). DADs appeared in areas where large SCaEs were noted. DADs were more appreciable on endocardial surface than on epicardial surface. The maximal amplitude of DADs was significantly larger on endocardial surface (0.057 ± 0.037 AU) than epicardial surface (0.008 ± 0.014 AU, P<0.001). Figure 5B shows the relationship between the maximal SCaE in the mapped area and corresponding DADs. Note that TAs or VTs originated only from endocardial surface due to larger DADs resulting from a higher diastolic Ca_i-voltage coupling gain (We did not use the amplitude of SCaE for these analyses because the occurrence of TAs prevented detection of the maximum SCaE). Figure 5C depicts the distribution of the maximal SCaE. The maximum SCaEs occurred at various sites, without apparently clustering. However, TAs arose frequently from the papillary muscle (PM) or near the summit of the interventricular septum, suggesting increased diastolic Ca_i-voltage coupling gain at those sites.

Figure 5.

Heterogeneous distribution of SCaEs. A, The SCaE maps on epicardial and endocardial surfaces for different PCL. Note that SCaEs occupy nearly all mapped area on both surfaces after 100-beat pacing at the shortest PCL. DADs were evident in the DAD maps for the endocardial surface. B, DAD slopes as a function of SCaE slopes measured at the maximal SCaE in the epicardial (open circles) and endocardial (closed circles) surfaces. Note that TAs (yellow circles) or VTs (red circles) occurred only on endocardial surfaces. C, Distribution of the maximal points of SCaE for each rabbit heart. Black dots depict SCaE only, blue dots for SCaE with DAD, yellow dots for SCaE with TA, red dots for SCaE with VT.

Figure 6A further illustrates the importance of diastolic Ca_i-voltage coupling gain in arrhythmogenesis. Figure 6Aa shows SCaEs (red arrowheads) induced VT at a cycle length of 378 ms. DADs (black arrows) and TAs were present at the VT origin (white cross). The magnitudes of the slope of the SCaEs ranged from 1.7 to 1.0 AU/s in this example. Figure 6Ab shows an idioventricular rhythm with a cycle length of 632 ms observed on the endocardial surface of another rabbit. Of note, SCaEs were present at a site (white cross) away from the early site of the idioventricular rhythm (blue color on isochrones). At the former site, the SCaEs with a slope of 1.0-1.3 AU/s failed to induce any changes of V_m. These findings suggest spatial inhomogeneity of diastolic Ca_i-voltage coupling gain was present within endocardial surface, and the differential coupling gain influenced genesis of ventricular arrhythmias.

Figure 6.

TA from papillary muscle (PM). A, Spatially heterogeneity of diastolic Ca_i-voltage coupling gain. Isochronal maps and optical tracings during VT originating from the PM (panel a) and idioventricular rhythm (panel b). Optical tracings were obtained at the white cross on each isochronal image. SCaEs (red arrowheads) generated DADs (black arrows) and TAs at the VT origin site, while SCaEs were dissociated from the idioventricular rhythm at the non-origin site. B, TA originating from the PM studied with zoom lens (upper panels). Arrow points to large DAD on papillary muscle. Lower panels show snapshots of wavefront propagation with the time of shock as time 0. Schematic diagram shows propagation pattern of two wavefronts from the PM (fast: blue lines, slow: brown lines).

Contribution of Purkinje Fibers to TA Development

The occurrence of TAs on endocardial surface and spatial heterogeneity of diastolic Ca_i-voltage coupling gain suggest Purkinje fibers may be the site of TA, since Purkinje fibers do not cover the endocardial surface uniformly (online Supplement Figure VI). Figure 6B shows optical tracings at the PM display SCaE and DAD. Note that there were two activation wavefronts arising from the PM. One was a fast wavefront conducting away along preferential pathways and reaching the distal part of the ventricle (white arrows). The other was a slow and centrifugally spreading wavefront (yellow arrows) that occurred later and failed to reach the distal area because it collided with the returning fast wavefront at the skirt area of the PM. The former and latter are consistent with wavefront propagation through the Purkinje fiber network and myocardium, respectively. These data are also shown in online Supplement Figure VII and Movies I and II. It is unlikely that this characteristic activation pattern resulted simply from anisotropic myocardial fiber orientation inserting into the PM, because pacing from the PM could not reproduce this pattern (online Supplement Figure VII and Movie III). To confirm the contribution of Purkinje cells on the optical signals, we used E-4031, a selective I_Kr blocker, in 2 additional hearts. E-4031 is known to prolong APDs more in Purkinje cells than in ventricular myocytes.⁸ The results (as shown in online Supplement Figure VIII) suggest that Purkinje fibers significantly contribute to the inscription of the endocardial optical tracings in our experimental conditions.

To further ascertain the role of Purkinje fibers in the generation of DADs and TAs, we performed simultaneous recordings of local bipolar electrograms and dual optical signals. Figure 7A shows a fibrillation-defibrillation episode followed by polymorphic VT. The first beat of the VT after defibrillation arose from the PM. Note that the upstroke of the optical action potential coincided with the timing of Purkinje-like potential (PP) recorded at the PM. When a bipolar electrode was placed on early sites of the TA beat (n=3) or the interventricular septum (n=4), the PPs preceding ventricular electrograms were found during TAs in all 7 hearts studied. The PPs were noted in 20/36 episodes (56%) after defibrillation and 70/96 episodes (73%) after pacing (online Supplement Figure IX).

Figure 7.

Role of Purkinje fibers. A, Polymorphic VT after defibrillation. Simultaneous recordings of local bipolar electrograms (BEG) and optical tracings at the PM are shown. The first DAD after defibrillation was large enough to generate TA at this site. Consequently, the first beat of the VT arose from the PM. PP = Purkinje-like potential. B, Simultaneous recordings of transmembrane potentials (TMP) and BEG during pacing-induced TA (asterisks). The top TMP tracing was from Purkinje cells because the phase zero of the action potential was coincidental with PP on BEG. Large DAD (filled arrowhead) was noted prior to the onset of action potential. The middle and bottom tracing shows action potentials with phase zero occurring at the timing of ventricular deflection on BEG. There was little (unfilled arrowhead) or no DADs on these TMP recordings. A red vertical line indicates the QRS onset.

Figure 7B shows TMPs obtained at the site as close to the recording bipolar electrode pair as possible. The top tracing of Figure 7B shows that the phase zero of the action potential occurred simultaneously with PP, indicating that this action potential originated from Purkinje fibers. A large DAD (filled arrowhead) preceded the onset of the action potential. TMP recorded in deeper layer of the endocardium or in the epicardium had little or no DADs, and the onset of action potential occurred at the time of ventricular electrogram. These findings indicate that Purkinje fibers are more prone to develop large DADs and TAs than ventricular myocytes on the endocardium or epicardium.

Role of I_K1 for Determining Diastolic Ca_i-Voltage Coupling Gain

Although our data suggest Purkinje fibers are more susceptible to DADs than ventricular myocytes, the underlying mechanism is unclear. Since inward rectifying potassium current (I_K1) is smaller in Purkinje fibers,⁹ an equivalent I_ti will cause a larger amplitude of DAD with lower I_K1.¹⁰ Therefore, we hypothesized that I_K1 might play a key role in higher diastolic Ca_i-voltage coupling gain in Purkinje fibers. To test this hypothesis, we examined the effect of I_K1 suppression on SCaE and DAD on epicardium by cesium chloride (CsCl, 5 mmol/L, n=3), barium chloride (BaCl₂, 3 μmol/L, n=3), and 1 mmol/L extracellular potassium (low [K⁺]_o, n=2). CsCl or BaCl₂ did not exert a significant effect on SCaEs but significantly increased DADs (DAD amplitude: 0.012 ± 0.008 AU to 0.042 ± 0.016 AU, P = 0.013, DAD slope: 0.05 ± 0.03 AU/s to 0.27 ± 0.14 AU/s, P = 0.018), indicating diastolic Ca_i-voltage coupling gain was enhanced CsCl and BaCl₂ (Figure 8). Low [K⁺]_o enhanced SCaE, but diastolic Ca_i-voltage coupling gain was also increased (online Supplement Figure X). Notably, TA was never observed from epicardium without I_K1 suppression (n=14), while TA occurred from epicardium in 6 of 8 hearts (75%) after I_K1 suppression (P <0.001). When we examined the overall incidence of VTs including VTs originating from the outside of the mapped area, VT was induced in 10/28 hearts (36%) with isoproterenol alone and in 6/8 hearts (75%) with isoproterenol and I_k1 suppression. Among 51 VT episodes after I_K1 suppression, 9 episodes (24%) included at least one beat arising from epicardium.

Figure 8.

Effect of I_K1 suppression. A, (Upper panels) Ca_i and V_m optical tracings at epicardium. I_K1 suppression by CsCl (5 mmol/L) augmented the DAD despite little change in the SCaE. (Lower panels) Changes in optical action potential and Ca_i transient with CsCl (PCL 200 ms). B, the magnitude of SCaE and DAD before and after CsCl or BaCl₂ (left panel) and incidence of TA at epicardium with or without I_K1 suppression (right panel). *P < 0.05; **P <0.001. C, VT after I_K1 suppression. All VT beats except 2^nd beat arose from the epicardial surface of right ventricular outflow tract (asterisks) where SCaEs (arrowheads) and DADs (arrows) preceded the QRS onsets of the VT.

Discussion

SCaE

SCaEs were observed both after defibrillation shocks and after rapid ventricular pacing. The latter findings are consistent with that observed by Fujiwara et al.¹¹ Ca_i signals obtained with our system represent averaged Ca_i from multiple myocytes. Faster Ca_i wave propagation and more synchronized Ca_i waves would produce more rapidly rising upstrokes of the averaged Ca_i signal, which is consistent with the PCL-dependent increase in the slope of SCaE. We propose that SCaEs at a tissue level correspond with Ca_i waves at a cellular level. In line with this proposal, SCaEs were eliminated by ryanodine treatment.

Katra et al¹² reported similar spontaneous Ca_i elevations in a canine wedge model. They observed larger Ca_i elevations at endocardium than epicardium, whereas we found comparable SCaEs on both cardiac surfaces. This difference might be due to different species and/or experimental protocols. They used an I_Ks blocker to enhance Ca entry and removed Purkinje fibers to focus on the transmural difference in spontaneous Ca_i elevation. Rate of Ca uptake was slower in the endocardium than epicardium, which promoted Ca_i-overload more greatly at endocardium. We did not evaluate the difference in Ca uptake properties because our optical signals from endocardial surface include Ca_i transients from Purkinje as well as endocardial cells. Our findings indicate vulnerability to DAD and TA is higher in Purkinje fibers than ventricular myocardium in intact whole hearts. Katra et al¹² also demonstrated that a sufficiently large spontaneous Ca_i elevation was necessary for genesis of TA. This holds true for our model, but it should be emphasized that diastolic Ca_i-voltage coupling gain is another important factor for development of TAs.

SCaE and DAD

Despite large SCaEs, large DADs were not consistently observed. As compared with single cells, Ca_i waves were less likely to cause large DADs because neighboring cells act as an electrotonic current sink.^{11, 13, 14} A second explanation is that each optical pixel contains a large number of cells. If only Purkinje fibers develop substantial DADs, the DAD amplitude might be reduced artificially by spatial averaging. In the example shown in Figure 6B, we were able to zoom in the lens to obtain a greater spatial resolution of the TA. As expected, the amplitude of DAD was larger in Figure 6B than in other examples because there are fewer cells per pixel. In addition, when diastolic Ca_i-voltage coupling gain was increased by I_K1 suppression, DADs also became more apparent because more cells participated in developing DADs in that pixel.

Diastolic Ca_i-Voltage Coupling Gain

The “gain” in cardiac excitation-contraction (E-C) coupling is defined as the ratio of the total flux through the SR Ca release channels (ryanodine receptor) to that through I_Ca,L.¹⁵ However, reverse E-C coupling (i.e., spontaneous SR Ca release leading to V_m depolarization by I_ti) also occurs to generate DADs.^5,14 Based on our results, DAD amplitude is determined by at least two factors: the magnitude of the whole cell Ca_i transient during a diastolic Ca_i wave, and the sensitivity of resting V_m to the elevated whole cell Ca_i (diastolic Ca_i-voltage coupling gain). Two major elements of the latter are the amplitude of I_ti generating the depolarization and the strength of background outward potassium current opposing depolarization, primarily the I_K1. I_ti components such as the Na-Ca exchange current (I_NCX) and Ca-activated Cl current have a similar or lower expression levels and current density in Purkinje cells compared with ventricular myocytes,^{16, 17} whereas I_K1 density is reduced in Purkinje cells.⁹ We found that, with comparable SCaE, I_K1 suppression allowed DADs to occur in epicardial ventricular myocytes. Taken together, these findings suggest that reduced I_K1, rather than augmented Ca_i release or increased I_NCX, accounts for the greater diastolic Ca_i-voltage coupling gain in Purkinje cells than ventricular myocytes. Reducing I_K1 with CsCl or BaCl₂ both depolarizes resting V_m, and decreases outward current in the nonlinear negative slope region of I_K1 conductance. The experiments with low [K⁺]_o suggest that the latter effect is mainly responsible for increasing the Ca_i-voltage coupling gain, since low [K⁺]_o, which decreases outward I_K1 but hyperpolarizes resting V_m, similarly increased the Ca_i-voltage coupling gain.¹⁸

The importance of diastolic Ca_i-voltage coupling gain is strengthened by the observation that procedures for enhancing Ca entry into myocytes (e.g. high dose of isoproterenol, aggressive pacing protocol) potentiated SCaEs, but did not induce TAs at epicardium without I_k1 suppression. This highlights the importance of diastolic Ca_i-voltage coupling gain as a key contributor to the development of DADs.

Clinical Implications

Our observations show that DAD-mediated TAs in Purkinje fibers can cause recurrence of ventricular arrhythmias following successful defibrillation. While spontaneous VF was not observed in these normal hearts, our previous study showed that recurrent spontaneous VF can occur in failing hearts after successful defibrillation.¹⁹ Sympathetic blockade is the most effective medical treatment for electrical storm,³ and successful radiofrequency catheter ablation of VF can be achieved at the site with Purkinje-like potentials.^{2, 4} Our observations may explain why these treatments are effective.

Catecholaminergic polymorphic VT (CPVT) is an inherited disease characterized by adrenergically mediated VT and VF. Optical mapping study using intact RyR2^R4496C+/−hearts revealed Purkinje system served as sources of focal arrhythmias in CPVT, albeit defective ryanodine receptors are present in all cardiac cells.²⁰ These results can be explained by higher diastolic Ca_i-voltage coupling gain of Purkinje cells than ventricular myocytes.

Study limitations

We created atrioventricular block because conducted sinus beats frequently prevented us from analyzing SCaEs, except at sites near the TA origin (online Supplement Figure XI).

In 23 rabbit ventricles, we used both cytochalasin D and blebbistatin to obtain sufficient suppression of contraction during isoproterenol infusion. Although blebbistatin had little effect on action potential morphology and Ca_i transients,²¹ cytochalasin D can prolong action potential duration and affect Ca_i transients.²² However, incidence and characteristics of SCaEs were comparable among experiments using cytochalasin D, blebbistatin, or both drugs.

We used CsCl and BaCl₂ to inhibit I_K1, but neither was a specific I_K1 blocker. Yet, a major effect at the concentration of CsCl and BaCl₂ used in this study was I_K1 inhibition.^{23, 24} Because I_K1 was crucial in not only stabilizing but also maintaining the resting V_m, abnormal automaticity would occur if the resting V_m of epicardium was depolarized to less than −60 mV.⁶ However, the concentrations of CsCl and BaCl₂ in this study had little effect on the resting V_m,^{25, 26} and low [K⁺]_o enhanced induction of TAs despite its hyperpolarizing effect on resting V_m.¹⁸ Therefore, DAD-mediated TA was a likely mechanism for emergence of the epicardial ectopic activities after I_K1 suppression.

Conclusions

SCaEs can emerge in the ventricles after prolonged rapid ventricular pacing or after successful defibrillation. Higher diastolic Ca_i-voltage coupling gain in Purkinje fibers is a likely mechanism underlying the spatial heterogeneity of susceptibility to DADs and TAs. I_K1 is a key determinant for the diastolic Ca_i-voltage coupling gain, and genesis of DAD-mediated triggered arrhythmias. These findings indicate Purkinje fibers are important arrhythmogenic substrates after an initial successful defibrillation.

Non-standard Abbreviations and Acronyms

Ca_i

intracellular calcium

CPVT

catecholaminergic polymorphic ventricular tachycardia

DAD

delayed afterdepolarization

E-C

excitation-contraction

I_K1

inward rectifying potassium current

I_NCX

sodium-calcium exchange current

I_ti

Ca-activated transient inward currents

[K⁺]_o

extracellular potassium concentration

PCL

pacing cycle length

PM

papillary muscle

SCaE

spontaneous intracellular calcium elevation

SR

sarcoplasmic reticulum

TA

triggered activity

TMP

transmembrane potential

VF

ventricular fibrillation

V_m

membrane potential

VT

ventricular tachycardia