Role of Sarcoplasmic Reticulum Calcium in Development of Secondary Calcium Rise and Early Afterdepolarizations in Long QT Syndrome Rabbit Model

Background L-type calcium current reactivation plays an important role in development of early afterdepolarizations (EADs) and torsades de pointes (TdP). Secondary intracellular calcium (Cai) rise is associated with initiation of EADs. Objective To test whether inhibition of sarcoplasmic reticulum (SR) Ca2+ cycling suppresses secondary Cai rise and genesis of EADs. Methods Langendorff perfusion and dual voltage and Cai optical mapping were conducted in 10 rabbit hearts. Atrioventricular block (AVB) was created by radiofrequency ablation. After baseline studies, E4031, SR Ca2+ cycling inhibitors (ryanodine plus thapsigargin) and nifedipine were then administrated subsequently, and the protocols were repeated. Results At baseline, there was no spontaneous or pacing-induced TdP. After E4031 administration, action potential duration (APD) was significantly prolonged and the amplitude of secondary Cai rise was enhanced, and 7 (70%) rabbits developed spontaneous or pacing-induced TdP. In the presence of ryanodine plus thapsigargin, TdP inducibility was significantly reduced (2 hearts, 20%, p = 0.03). Although APD was significantly prolonged (from 298 ± 30 ms to 457 ± 75 ms at pacing cycle length of 1000 m, p = 0.007) by ryanodine plus thapsigargin, the secondary Cai rise was suppressed (from 8.8 ± 2.6% to 1.2 ± 0.9%, p = 0.02). Nifedipine inhibited TdP inducibility in all rabbit hearts. Conclusion In this AVB and long QT rabbit model, inhibition of SR Ca2+ cycyling reduces the inducibility of TdP. The mechanism might be suppression of secondary Cai rise and genesis of EADs.


Introduction
Drug-induced acquired long QT syndrome and torsades de pointes (TdP) is one of the most serious adverse effects of medications. There is close relationship between risk of prolonged QT interval/TdP and inhibition of rapidly activating delayed rectifier potassium current (I Kr ) [1,2]. I Kr is encoded by the hERG gene, which is the gene encoding mutation in long QT syndrome type 2 [3]. However, the risk of TdP development is not always linked to prolonged QT interval. Amiodarone has been wildly used for ventricular and supraventricular tachyarrhythmias, especially in patients with heart failure. Although it causes significant QT prolongation, the risk of provoking TdP is relatively minor. The relationship between the risk of TdP and the severity of QT prolongation remains mysterious.
Early afterdepolarizations (EADs) are secondary depolarization during repolarization of action potential and usually developed in the presence of prolonged action potential duration (APD). L-type calcium current (I Ca,L ) reactivation is required in the initiation of EADs [4]. Previously we observed secondary intracellular calcium (Ca i ) rise in a rabbit heart failure model [5]. It has also been reported that I Ca,L blockade abolished EAD development in heart failure animal models [6]. Whether or not sarcoplasmic reticulum (SR) Ca 2+ cycling also plays a role in genesis of EADs and secondary Ca 2+ rise is still unclear. In this study, we hypothesized that SR Ca 2+ cycling inhibition suppressed secondary Ca i rise and development of EADs. We used a long QT syndrome rabbit model with atrioventricular block (AVB) creation and E4031 containing low-K + -low-Mg 2+ Tyrode's solution perfused to test the hypothesis. This study followed the previous studies of long QT rabbit models [7,8]. E4031 is a specific I Kr inhibitor and has been used for creation of long QT animal model.

Materials and Methods
The research protocol was approved by the Institutional Animal Care and Use Committee (IACUC) of Chang Gung Memorial Hospital (Permit Number: 2012121704) and conformed to the Guide for Use of Laboratory Animals. All surgery was performed under general anesthesia with ketamine, rompun and isoflurane, and all efforts were made to minimize suffering. Ten adult New Zealand white rabbits (2.5-3.5 kg) were used in this study.

Optical mapping and AVB model
We created AVB after harvesting rabbit hearts in this study using a modified atrioventricular node ablation method [5] and the same optical mapping method described previously [9,10]. In brief, the rabbits were generally anesthetized with intravenous injection of ketamine (8 mg/kg) and xylazine (8mg/kg).When the rabbits were fully anesthetized and unresponsive to physical stimuli, the hearts were rapidly harvested and Langendorff-perfused with 37°C standard Tyrode's solution of the following composition: 125 mM NaCl, 4.5 mM KCl, 0.5 mM MgCl 2 , 24 mM NaHCO 3 , 1.8 mM NaH 2 PO 4 , 1.8 mM CaCl 2 , 5.5 mM glucose and 100 mg/L albumin, equilibrated with 95% O 2 and 5% CO 2 in de-ionized water and with a pH of 7.40. We then performed radiofrequency ablation to create AVB using a 7Fr. quadripolar large-tip ablation catheter (Biosense Webster, Diamond Bar, CA, USA) at a output energy of 10~20W generated by a radiofrequency generator (Radionics RFG-3C, Radionics Inc., MA, USA). The target ventricular escape rate after AV node ablation was less than 60 beats per minute.
The hearts were stained with Rhod-2-AM (Molecular Probes, Carlsbad, CA, USA) for Ca i and RH-237 (Molecular Probes) for membrane potential (V m ). We used a laser light at a wavelength of 532 nm (Millennia, Spectra-Physics Inc., Santa Clara, CA, USA) to excite the fluorescence dyes. The emitted fluorescence was acquired and filtered (715 mm for V m and 580 nm for Ca i ) with twocharge-coupled device cameras (CA-D1-0128T, Dalsa Inc., Waterloo, Ontario, Canada) at 4 ms/frame temporal resolution and 128 x 128 pixels with spatial resolution of 0.35 x 0.35 mm 2 per pixel. To provoke development of TdP, the perfused solution was changed to a modified low-K + (2.

Experimental protocol
Pseudo-electrocardiography (pECG) was recorded using three electrodes placed at the left atrium, the posterior wall of the left ventricle and the right ventricle. A bipolar catheter was inserted into the right ventricular apex for pacing at twice threshold. Because of significantly prolonged APD in this long-QT syndrome model, the optical mapping signals were acquired at cycle lengths of 1000, 800, 600, 500, 400, 350, 300 ms and then down to the shortest 1:1 captured cycle length with a 20 ms step. A S1/S2/S3 short-long-short pacing protocol (S1 30 beats with S1-S1 500 ms, a long S1-S2 of 1000 or 2000 ms and a S2-S3 starting from 500 ms and gradually shortened to the ventricular effective refractory period) was used to induce TdP. Some TdP episodes were also induced by a short burst pacing following a long pause in this AVB model. After the baseline studies, E4031 (0.5 μM), ryanodine (1.0 μM) plus thapsigargin (1.0 μM), and nifedipine (2.0 μM) were administered subsequently, and the experimental protocols were repeated after each set of medications.

Data analysis
APD 80 was measured at the level of 80% repolarization of action potential. Secondary rise of Ca i is defined as the spontaneous increase of the Ca i at the downslope of the primary Ca i released. The amplitude of secondary Ca i rises is defined as the largest deviation from a line drawn between the onset and offset of the secondary Ca i rise as previously described [5]. TdP was defined as polymorphic ventricular tachycardia more than 3 beats. Continuous variables with normal distribution were expressed as the mean ± SEM, and categorical variables were expressed as number (percentage). Differences in continuous variables between different zones of the same heart with normal distribution were analyzed by paired t-test. One-way repeated measures ANOVA with post-hoc LSD analysis was used to compare continuous variables in the presence of different sets of medications with baseline values. The differences in categorical variables at different concentrations were tested using Cochran's Q test. Differences were considered significant when the probability value was < 0.05.

Inducibility of TdP in AVB model
We successfully obtained optical mapping recording in 10 rabbit hearts. There was no inducible TdP or EADs at baseline. We recorded spontaneous or pacing-induced TdP in 7 hearts (70%) in the presence of E4031. Fig 1 shows representative examples of spontaneous (panel A) and pacing-induced TdP (panel B, the second and the third subpanels). Panel A shows secondary Ca i rise and Ca i oscillation (red trace) during the spontaneous TdP. After ryanodine plus thapsigargin administration, the inducibility of TdP was significantly suppressed ( Fig 1B, the forth subpanel) and only 2 hearts had inducible TdP in the presence of ryanodine plus thapsigargin ( Fig 1C, 20%, p = 0.03). We further applied nifedipine and the inducibility of TdP was completely suppressed (0%). The bottom subpanel of Fig 1B shows a representative example in the presence of nifedipine. The results were compatible with previous findings that SR Ca 2+ cycling also played an important role in the genesis of EADs.

Action potential duration
Fig 2A shows representative action potential traces and APD maps at baseline, in the presence of E4031, further administration of ryanodine plus thapsigargin, and then nifedipine, respectively. APD 80 was longer in the presence of E4031 and was further prolonged after administration of ryanodine plus thapsigargin at both 1000 ms and 500 ms pacing cycle lengths (PCLs). Nifedipine shortened APD 80 . Because of significantly prolonged APD 80 , we were not able to pace the hearts at pacing cycles shorter than 400 ms after applying E4031 in most of the hearts. Statistical tests showed that E4031 significantly prolonged APD 80 (from 179 ± 5 ms to 298 ± 30 ms, p = 0.003, and from 178 ± 6 ms to 267 ± 21 ms, p = 0.002, at PCL 1000 ms and 500 ms, respectively, Fig 2B), and ryanodine plus thapsigargin further prolonged APD 80 (457 ± 75 ms and 357 ± 637 ms, at PCLs of 1000 ms and 500 ms, p = 0.007 and p = 0.003, respectively). Nifedipine significantly shortened the extremely prolonged APD 80 (307 ± 65 ms and 244 ± 37 ms, at PCLs of 1000 ms and 500 ms, p < 0.001 and p = 0.001, respectively). Although inhibition of SR Ca 2+ cycling further prolonged APD 80 , the inducibility of TdP was suppressed as shown in Fig 1C. Secondary Ca i rise Previously we reported that apamin induced secondary Ca i rise and EADs in failing hearts [5]. Optical mapping recording showed that EADs were initiated at the area with secondary Ca i rise. Therefore, we analyzed the amplitude of secondary Ca i rise and the relationship between secondary Ca i rise and TdP inducibility. Fig 3A shows representative V m -Ca i traces and secondary Ca i rise maps. Significant secondary Ca i rise developed at the basal area of the left ventricle in the presence of E4031. The amplitude of secondary Ca i rise was positively correlated with the APD with E4031. Fig 3B shows the average secondary Ca i rise at baseline, in the presence of E4031, further administration of ryanodine plus thapsigargin, and then nifedipine. At baseline, there was only minimal secondary Ca i rise (1.0 ± 1.0%), which was significantly enhanced by E4031 (8.8 ± 2.6%, p = 0.03). Ryanodine plus thapsigargin prolonged APD but significantly suppressed secondary Ca i rise (1.2 ± 0.9%, p = 0.02). Nifedipine partially restored the extremely prolonged APD and further suppressed secondary Ca i rise (0.4 ± 0.4%). As shown in Fig 3C,   without TdP. Note the 3 non-inducible hearts had relatively lower amplitude, although the animal number was too small for statistical testing.

Calcium decay
Whether calcium decay is important in the development of TdP is not clear. Fig 5A shows a representative example of the calcium decay at baseline, in the presence of E4031, further administration of ryanodine plus thapsigargin, and then nifedipine. The trend of calcium decay was similar to the trend of APD 80 . The average calcium decay were significantly prolonged by E4031 (Fig 5B, from 63 ± 1 ms to 100 ± 12 ms, p = 0.01) and then further prolonged after administration of ryanodine plus thapsigargin (205 ± 13 ms, p < 0.001). Nifedipine shortened the calcium decay (126 ± 22 ms, p < 0.001). Similarly, although inhibition of SR Ca 2+ cycling further delayed the calcium decay, the inducibility of TdP was not further enhanced.

Arrhythmia Pattern of Torsades de pointes
In this long-QT syndrome model, the TdP ventricular tachyarrhythmias usually resulted from beat to beat changes in wave propagation patterns initiated by EADs from the border of the largest Ca i rise area. Fig 6 shows an example of an episode of TdP following an intrinsic escape beat (beat 1). The beat 2 was an EAD beat initiated from the border of the largest secondary Ca i rise region. Because of significant spatially heterogeneous prolongation of APD, the shortcoupled EAD beat led to unidirectional conduction block and formed reentry. This arrhythmia pattern is consistent with a previous study reported by Asano et al [11].

Discussion
The major findings of this study include that inhibition of SR Ca 2+ cycling suppressed the inducibility of TdP. In this model, inhibition of SR Ca 2+ cycling did not shorten APD. The mechanism could be the suppression of secondary Ca i rise. We also demonstrated that EADs was initiated from border of the high secondary Ca i rises region with earlier Ca i pre-fluorescence. These results indicate that both SR Ca 2+ cycling and I Ca,L are important in development of secondary Ca i rise and EADs. Therfore, secondary Ca i rise might be a marker of torsadogenesis risk.

Effects of SR Ca 2+ cycling inhibitors on APD
The interaction between APD and Ca 2+ dynamics is complicated. APD and Ca i are bidirectionally coupled in cardiac tissue: a longer APD usually triggers a larger Ca 2+ release and a larger Ca 2+ release can either shortens or prolongs APD [12,13]. A larger Ca 2+ release enhances Na + -Ca 2+ exchange current to prolong APD, but it also potentiates Ca 2+ induced I Ca,L inactivation and enhances Ca 2+ -sensitive K + and Clcurrents to shorten APD. Ryanodine has been shown to slow down I Ca,L inactivation through blockade of SR Ca 2+ release [14]. Enhanced RyR phosphorylation is associated the development of EADs in a long QT 2 rabbit model [15]. The reduced Ca 2+ release leads with slow inactivation of I Ca,L leads to longer APD. Thapsigargin had been also shown to prolong APD through depletion of SR Ca 2+ , and patch clamp recording showed an increased total influx of Ca 2+ with a longer duration in the presence of thapsigargin [16]. Therefore, the net effects of SR Ca 2+ cycling blockade include prolongation of APD, slowing down I Ca,L inactivation and SR Ca 2+ reuptake.

The importance of SR Ca 2+ cycling on EADs genesis
The chain of Ca 2+ cycling includes I Ca,L , Na + -Ca 2+ exchange current and SR Ca 2+ release and reuptake. I Ca,L reactivation plays a central role in the development of EADs and TdP, and inhibition of I Ca,L is one of the preventive and therapeutic approaches to TdP [6,17,18]. Experimental and simulation studies have shown that I Ca,L may reactivate and reversely repolarize under the situation of reduced outward currents and/or increased inward currents [17]. A recent report showed that blockade of Na + -Ca 2+ exchange current suppresses genesis of EADs without interfering the reactivation of I Ca,L in a H 2 O 2 oxidative cell model [18]. It suggests that Na + -Ca 2+ exchange current may also play a role in EAD genesis. In this study, our data showed that SR Ca 2+ cycling inhibition suppressed EAD genesis even in the situation of longer APD. A possible mechanism is that the inhibition of SR Ca 2+ cycling leads to slowing of Ca 2+dependent I Ca,L inactivation, which reduces the possibility of available L-type Ca 2+ channels for reactivation and the degree of secondary rise of Ca i to attenuate EAD genesis. This study also confirms previous findings that any interference among the Ca 2+ cycling affects the genesis of TdP [19]. On the basis of this study and previous reports, we postulate that the interaction of L-type calcium channel, SR Ca 2+ release-reuptake and Na + -Ca 2+ exchanger is required to generate EADs.

Mechanisms of secondary Ca i rise
The formation of secondary Ca i rise is complicated. Priori et al. first proposed that abnormal Ca 2+ cycling can be the mechanism of EADs [20]. Piacentino et al. observed Ca 2+ influx during late portions of action potential in failing human cardiomyocytes using a voltage-clamp model [21]. Zeng and Rudy demonstrated that recovery and reactivation of I Ca,L is the mechanism of EADs [4]. It seems that I Ca,L plays the most important role in genesis of EADs. However, whether or not I Ca,L is the only factor in the EAD genesis was not very clear. Qu et al. investigated the mechanisms of EADs and ultra-long APD using a Luo and Rudy simulation model [22]. In the model, the formation of EADs and ultra-long APD is associated with alteration of window I Ca,L , speed of I K activation, slope of the steady-state inactivation curve of I Ca,L and pedestal I Ca,L [22]. SR Ca 2+ release or reuptake/extrusion can affect window I Ca,L , I K activation, inactivation of I Ca,L through Ca 2+ -dependent I Ca,L inactivation, modulating V m by Na + -Ca 2+ exchange current, regulating I Ks and Ca 2+ activated potassium and chloride currents. The mechanisms of secondary Ca i rise involve the interaction of I Ca,L , SR Ca 2+ cycling and Na + -Ca 2+ exchange current: SR Ca 2+ release causes inactivates I Ca,L ; then either sustained high V m or spontaneous Ca 2+ release-induced depolarization through Na + -Ca 2+ exchange current reactivates I Ca,L [19]. SR Ca 2+ cycling blockade reduces Ca 2+ release, subsequently slows I Ca,L inactivation to reduce secondary Ca i rise.

Arrhythmia pattern of TdP
Most EADs initiated from the border zone rather than the region with largest amplitude of Ca i rise in this study and the previous studies [5]. The mechanism of this phenomenon is not clear, and the reason could be the long refractory period of the myocardial tissue with the longest APD. Amplitude of Ca i rise is correlated with the APD, and the excessive prolongation of APD prevents immediate re-initiation of an action potential. The nearby myocardium with relatively shorter APD is available to initiate an EAD. Prolonged APD also leads to unidirectional block during the episodes of TdP. Therefore, the pattern of arrhythmias was focal trigger of EADs and reentry due to heterogeneously prolonged APD in this model.

Clinical implications
Prolongation of QT interval leading to TdP is one of the major adverse effects of medications. It has been reported that the risk of drug-induced QT prolongation and TdP is associated with I Kr blockade activity at the therapeutic level [23]. The concern leads to significant amount of drugs being withdrawn from the market or even never entering the market [24]. Some medications prolong QT interval but carry relatively low risk of TdP, such as amiodarone. Amiodarone has been reported to affect L-type Ca 2+ channel and SR Ca 2+ cycling [25,26]. It is possible that the effects of amiodarone on Ca 2+ homeostasis lead to anti-arrhythmic property with relatively less torsadogenic effects. The mechanisms can be explained partially by the effects of SR Ca 2+ blockade on suppressing inducibility of TdP. SR Ca 2+ homeostasis can also be one of the targets to manage TdP clinically.

Limitations
There are still some limitations in this study. We mapped only the epicardium of hearts and were not able to recognize some arrhythmias when EADs initiated from mid-myocardium, sub-endocardium or outside of mapping field. Purkinje-ventricular escape rhythm after AV node ablation depended on the level of ablation, and the escape rate was different among the hearts. Because of cardiac memory, the rate of escape rhythm might affect APD and the inducibility of TdP.

Conclusion
In this AVB and long QT rabbit model, inhibition of SR Ca 2+ cycling reduces the inducibility of TdP. The mechanism might be suppression of secondary Ca i rise, although inhibition of SR Ca 2+ cycling does not shorten APD. Nifedipine further inhibits the inducibility of TdP. These results indicate that both SR Ca 2+ cycling and I Ca,L are important in EAD genesis.