Renal artery denervation prevents ventricular arrhythmias in long QT rabbit models

Long QT syndrome (LQTS) is commonly presented with life-threatening ventricular arrhythmias (VA). Renal artery denervation (RDN) is an alternative antiadrenergic treatment that attenuates sympathetic activity. We aimed to evaluate the efficacy of RDN on preventing VAs in LQTS rabbits induced by drugs. The subtypes of LQTS were induced by infusion of HMR-1556 for LQTS type 1 (LQT1), erythromycin for LQTS type 2 (LQT2), and veratridine for LQTS type 3 (LQT3). Forty-four rabbits were randomized into the LQT1, LQT2, LQT3, LQT1-RDN, LQT2-RDN, and LQT3-RDN groups. All rabbits underwent cardiac electrophysiology studies. The QTc interval of the LQT2-RDN group was significantly shorter than those in the LQT2 group (650.08 ± 472.67 vs. 401.78 ± 42.91 ms, p = 0.011). The QTc interval of the LQT3-RDN group was significantly shorter than those in the LQT3 group (372.00 ± 22.41 vs. 335.70 ± 28.21 ms, p = 0.035). The VA inducibility in all subtypes of the LQT-RDN groups was significantly lower than those in the LQT-RDN groups, respectively (LQT1: 9.00 ± 3.30 vs. 47.44 ± 4.21%, p < 0.001; LQT2: 11.43 ± 6.37 vs. 45.38 ± 5.29%, p = 0.026; LQT3: 10.00 ± 6.32 vs. 32.40 ± 7.19%, p = 0.006). This study demonstrated the neuromodulation of RDN leading to electrical remodeling and reduced VA inducibility of the ventricular substrate in LQT models.

www.nature.com/scientificreports/ Congenital long QT syndrome (LQTS) is a congenital cardiac disorder that prolongs ventricular repolarization leading to ventricular arrhythmias (VAs) or sudden cardiac death (SCD). Beta-blocking (BB) agents are the cornerstone in preventing cardiovascular (CV) events in LQTS, however, 32% of symptomatic VAs and 14% of cardiac arrests were identified in a 5-year follow-up 1 . The implanted cardioverter defibrillator (ICD) reduces the incidence of SCD but cannot prevent the occurrence of life-threatening VAs. Therefore, the demand for an alternative approach is required for drug-refractory LQTS patients. Recently, neuromodulation such as left cardiac sympathetic denervation (LCSD) successfully reduced the VAs and SCD 2,3 . However, 23% of symptomatic VAs and 95% of LCSD-related adverse effects were documented 4,5 . Alternatively, renal artery denervation (RDN) is another effective neuromodulation with favorable results in suppressing life-threatening VAs [6][7][8] . Previous studies have demonstrated the efficacy of RDN, including reduced VA occurrence in electrical storms, prevention of VAs in heart failure (HF), and myocardial infarction subjects [9][10][11] . Despite favorable results from previous research, the exact role and mechanism of RDN in LQTS remained questionable. Finally, the similar ion channels between rabbit and human hearts were demonstrated in previous research was the reason for choosing a rabbit model many LQTS research. We aimed to investigate the cardiac electrophysiological (EP) properties and arrhythmogenicity of RDN as an alternative approach to prevent VAs in the LQTS rabbit model. www.nature.com/scientificreports/ urements from the baseline to the sequentially increased drug doses for LQT subtype induction. In the LQT1 and LQT1-RDN groups, the QTc intervals were gradually prolonged under sequential doses of HMR-1556 when compared with those at the baseline within each group ( Table 2). Similar results of QTc interval prolongation were found in the LQT2, LQT2-RDN, LQT3, and LQT3-RDN groups under sequentially increased drug doses of LQT induction when compared to those at the baseline within each group ( Table 2). Under the same LQT subtype induction dose, the QTc intervals between the LQT and LQT-RDN groups showed no difference in all the LQT subtypes (LQT1, LQT2, and LQT3) ( Table 2). Figure 2 demonstrated the effective refractory periods (ERP) of LV and RV at 2 times and 10 times pacing threshold under sequentially increased drug doses of LQT induction in each group, respectively. In LV and RV, the ERPs during 2 times and 10 times pacing threshold revealed no difference between the baseline and those with sequentially increased HMR-1556 dose within the LQT1 and LQT1-RDN groups, respectively ( Fig. 2a, b). In LV and RV, the ERPs during 2 times and 10 times pacing threshold revealed no difference between the LQT1 and LQT1-RDN groups at baseline and each HMR-1556 doses, respectively ( Fig. 2a, b). Figure 3a demonstrated the percentages of QTc interval prolongation in the LQT1 and LQT1-RDN groups under sequential HMR-1556 infusion. Under HMR-1556 0.04 mg/kg/min infusion, the percentages of QTc interval prolongation in the LQT1 and LQT1-RDN groups were significantly longer than those in the baseline, respectively (Fig. 3a).

Cardiac EP study and VA inducibility test.
In LV and RV, the ERPs during 2 times and 10 times pacing threshold gradually prolonged under sequentially increased erythromycin doses in the LQT2 and LQT2-RDN groups with significant difference between those at the baseline and final erythromycin doses within each group, respectively (Fig. 2c, d). In LV, the ERPs during 2 times and 10 times pacing threshold of the LQT2-RDN groups in each erythromycin dose were significantly shorter than those in the LQT2 groups, respectively (Fig. 2c). Figure 3b demonstrated the percentage of QTc interval prolongation in the LQT2 and LQT2-RDN groups under sequential erythromycin infusion. Under erythromycin 133, 266, and 400 nmol/kg/min infusion, the percentages of QTc interval prolongation in the LQT2 group were significantly longer than those in the baseline, respectively (Fig. 3b). Under erythromycin 266 and 400 nmol/kg/min infusion, the percentages of QTc interval prolongation in the LQT2-RDN group were significantly longer than those in the baseline, respectively (Fig. 3b). Under erythromycin 266 and 400 nmol/ kg/min infusion, the percentages of QTc interval prolongation in the LQT2 group were significantly longer than those in the LQT2-RDN group, respectively (Fig. 3b).
In LV, the ERPs during 2 times and 10 times pacing threshold gradually prolonged under sequentially increased veratridine doses in the LQT3 and LQT3-RDN groups with significant difference between those at the baseline and veratridine 2 µM/kg/min infusion within each group, respectively (Fig. 2e). In RV, the ERPs during 2 times and 10 times pacing threshold gradually prolonged under sequentially increased veratridine doses in the LQT3 and LQT3-RDN groups, respectively (Fig. 2f). In RV, the ERPs during 2 times and 10 times pacing threshold revealed a significant difference between those at the baseline and veratridine 4 µM/kg/min infusion within the LQT3-RDN group, respectively (Fig. 2f). Figure 3c demonstrated the percentage of QTc interval prolongation in the LQT3 and LQT3-RDN groups under sequential veratridine infusion. Under veratridine 1, 2, and 4 µM/kg/min infusion, the percentages of QTc interval prolongation in the LQT3 group were significantly longer than those in the baseline, respectively (Fig. 3c). Under veratridine 2 and 4 µM/kg/min infusion, the percentages of QTc interval prolongation in the LQT3-RDN group were significantly longer than those in the baseline, respectively (Fig. 3c).
Two rabbits in the LQT1 group developed spontaneous VAs at baseline, whereas the LQT1-RDN group revealed no spontaneous arrhythmias. The VAs inducibility in the LQT1-RDN group was significantly lower than in the LQT1 group (Fig. 4). Two rabbits in the LQT2 group developed spontaneous VAs, whereas 1 rabbit in the LQT2-RDN group developed spontaneous VA. The VA inducibility in the LQT2-RDN group was significantly lower than in the LQT2 group (Fig. 4). Finally, there were no documented spontaneous VAs in both the LQT3 and LQT3-RDN groups. The VA inducibility in the LQT3-RDN group was significantly lower than in the LQT3 group (Fig. 4).

Discussion
The main findings of this study are 1. The QTc interval prolongations were less prominent in the LQT-RDN groups than those in the LQT groups after LQT subtype induction; 2. The RDN decreased the burden of spontaneous and inducible VAs. Neuromodulation via RDN potentially prevented VAs through electrical remodeling in Table 2. The QTc interval measurement (ms) from baseline to the sequentially increased doses of all LQTS subtype induction (LQTS induction drugs and doses are described in the "Methods" section).  15 . Our study revealed comparable results with trends of VERP prolongation and significant reduction of VA inducibility. In addition, the administration of anesthetic agents was an important consideration during the experimental procedure 16,17 . It is suggested that the pentobarbital agents may potentially affect the circadian and cardiac rhythm leading to confounding results. However, an animal study has demonstrated the limited effects on repolarization and VA burden under anesthetic agents. The dissociation between the pharmacologic target of pentobarbital and the CV physiological actions was identified in a mice study. Therefore, the use of pentobarbital revealed a stable VA threshold and was considered one of the suitable models for cardiac EP studies.
In LQTS patients, an elevated sympathetic tone may trigger ventricular repolarization leading to fatal VAs, therefore, antiadrenergic agents have been the main goal for LQTS treatment. Despite the promising results from antiadrenergic agents, not all LQTS subtypes benefit from the current treatment, thus alternative therapy may offer additional insights on LQTS treatment. A previous study demonstrated renal sympathetic stimulation can increase both the systemic and cardiac sympathetic tone leading to potential VAs and SCD, whereas the RDN showed cardiac protective effects by stabilizing the ventricular substrate through decreased sympathetic nerve density 18 . Moreover, Chen et al. reported a positive correlation between sympathetic nerve density and fatal VAs 18 . The underlying mechanism of RDN includes the reduced secretion of norepinephrine (NE) by more than 80% after the RDN procedure 18,19 . In our previous research, the RDN significantly reduced the NE and epinephrine levels in the renal parenchymal tissues leading to the downregulation of sympathetic activity 10,15 . Moreover, the RDN results in significant stellate ganglion remodeling and influences afferent nerve fibers of the ganglia in the brainstem 18,20 .
In our study, the incidence of VA inducibility was lower significantly in all 3 subtypes of LQT-RDN groups than those in the LQT groups. Interestingly, our study showed results of reduced VA burden through RDN in the LQT3-RDN group sheds new light on the therapeutic strategies. In LQT3 subjects, the role of sympathetic activation in VA events is controversial and the efficacy of antiadrenergic agents on LQT3 subjects is uncertain. Wilde et al. demonstrated the benefits of RDN which attenuated QTc interval prolongation but prevented bradycardia in LQT3 subjects 21 . Our study revealed similar results in which the LQT3 group showed a lower VA burden and the SR CL prolongation was less prominent in comparison with the LQT1 and LQT2 groups. Importantly, the RDN in the LQT3-RDN group successfully decreased VA burden without affecting the heart rate which could be a potential substitute for an antiadrenergic agent in LQT3 subjects.
In this study, the major limitation was the different physiological and autonomic mechanisms between the rabbit and humans. An LQTS rabbit model induced by drugs cannot fully represent the real-world clinical conditions of congenital LQTS patients 22 . Second, the telemetric tools were not included in this study. Our study investigated the LQTS animal model induced by drugs, therefore it is not able to have telemetric monitoring in an ambulatory animal. However, previous investigations have proven the elevated sympathetic tone leading to increased VA inducibility which was compatible with our findings 23,24 . Third, our study was performed under GA rather than a conscious state which could potentially affect the results. However, an LQTS animal study and our previous RDN studies showed favorable results demonstrating the role of RDN 15,25,26 . Finally, autonomic blockades were not applied in this study to evaluate the intrinsic autonomic tone. Previous studies have demonstrated the RDN approach attenuated the sympathetic tone via elimination of peri-renal sympathetic neurons leading to reduced production of catecholamine levels 27 . The efficacy of RDN in the heart was demonstrated in our previous studies, in which the RDN led to a decreased density of sympathetic neurons of the myocardium and attenuated VA inducibility 10,15,26 . These studies have fully demonstrated the interaction between cardiac sympathetic tone and the RDN approach.
In summary, our findings suggested that RDN affects the sympathetic activity leading to QTc interval prolongation and reduction of VA burden in all the subtypes of LQTS models. These results offer additional insights on the characteristics of neuromodulation through RDN and highlight the potential benefits of RDN as an alternative antiadrenergic treatment in LQTS.  inhibitor that inhibits the outward K + current for myocardial repolarization. The erythromycin is a rapid component of delayed rectifier K + current (IKr) inhibitor that inhibits the outward K + current for myocardial repolarization. The veratridine is an inhibitor of sodium (Na 2+ ) channel inactivation that results in the opening of the Na 2+ channel during myocardial depolarization. Previous studies have proven the validity of the LQTS rabbit model induced by drugs selected in this study 28 . Detailed LQTS subtype induction and doses are described in the experimental procedure section. A total of 44 male New Zealand white rabbits (weight 2.5-3.0 kg, from Shulin Breeding facility, New Taipei, Taiwan) at 12 weeks of age were randomized into 6 groups of LQT1 (n = 8), LQT1-RDN (n = 7), LQT2 (n = 8), LQT2-RDN (n = 7), LQT3 (n = 7), and LQT3-RDN (n = 7). All rabbits underwent 2 anesthetic procedures, the first procedure was under intraperitoneal and intravenous anesthesia for RDN or sham (exposure of the abdominal cavity without RDN), and the second procedure was under general anesthesia (GA for the final experiment. In the first anesthetic procedure, the LQT1, LQT2, and LQT3 groups underwent surgical opening followed by the closure of the abdominal cavity without RDN, whereas the LQT1-RDN, LQT2- Surgical and chemical RDN. The RDN procedure was performed 1 month before the experimental procedure. We selected a combination of surgical and chemical RDNs. This technique has been described in our previous publications with a confirmed RDN effect 15,26 . In brief, each rabbit underwent fasting for 1 night before the surgery. All rabbits were anesthetized with induction by intraperitoneal injection of sodium pentobarbital (40 mg/kg) and maintained by intravenous injection of xylazine (1 mg). The renal arteries were approached through a mid-abdominal incision. Under surgical RDN, bilateral renal arteries were surgically denervated by cutting all visible nerves in the area of the renal hilus and by removing 1 cm of the adventitia from the renal artery. Under chemical RDN, the area was then moistened with a 20% phenol solution for 15 min (mins). After the procedure, the abdomen was closed layer by layer. All rabbits received antibiotic agents (gentamycin 5 mg/ kg) immediately after the procedure. Pain-controlling agents (ibuprofen 2 mg/kg) were placed in their water for 3 consecutive days. The final experimental procedure was performed after 4 weeks of maturation.
Experimental procedure. During preparation, a warming blanket was used to maintain the rabbit's body temperature. All animals were anesthetized with an intraperitoneal injection of sodium pentobarbital (40 mg/ kg) followed by endotracheal intubated with ventilation. Inhalation of isoflurane 2% was administrated for maintaining GA. An intravenous (IV) line was set up for the medication infusion and fluid supplement. Thoracotomy was performed after subcutaneous injection of 5 ml lidocaine 1% (10 mg/ml). After local anesthesia, a pericardial incision was performed to expose the epicardial surface. Electrocardiogram (ECG) measurements and cardiac EP study were performed in all rabbits. The ECG signals were amplified by a standard amplifier (Lab System TM PRO EP Recording System, Bard, MA, USA, filter setting:0.1-150 Hz). The QT interval and corrected QT interval (QTc) (Bazett's formula) were measured at baseline and in each concentration of drug infusion.
Drug infusion of HMR-1556, erythromycin, and veratridine were administrated to mimic the acute onset of LQT1, LQT2, and LQT3, respectively. In the LQT1 and LQT1-RDN groups, HMR-1556 was administered in 4 sequentially increased drug doses to mimic the LQT1 syndrome. The initial IV infusion rate of HMR-1556 was 0.005 mg/kg/min for 30 min followed by an increased infusion rate of 0.01, 0.02, and 0.04 mg/kg/min for 30 min sequentially. In the LQT2 and LQT2-RDN groups, erythromycin was administered in 3 sequentially increased drug doses to mimic the LQT2 syndrome. The initial IV infusion rate of erythromycin was 133 nmol/kg/min for 30 min followed by an increased infusion rate of 266 and 400 nmol/kg/min for 30 min sequentially. In the LQT3 and LQT3-RDN groups, veratridine was administered in 4 sequentially increased drug doses to mimic the LQT3 syndrome. The initial IV infusion rate of veratridine was 0.5 µM/kg/min for 30 min followed by an increased infusion rate of 1, 2, and 4 µM/kg/min for 30 min sequentially. Cardiac EP study and VA inducibility test was done for all drugs under each concentration.
Cardiac EP study was performed by a custom-made stimulator (Model 5325, Medtronic, Ltd, Minneapolis, USA) that delivered constant-current pulses of 1 ms duration. The test was performed at baseline and during each concentration of drug infusion in all groups. Programmed electrical stimulation was delivered to the multielectrode catheter distal tips at 2 and 10 times the pacing thresholds at the RV and LV. Eight consecutive stimuli (S1S1 = 300 ms cycles) followed by a premature stimulus (S1S2). The S1S2 was initially decreased from 200 ms by decrements of 10 ms and then 1 ms when achieving an ERP, respectively. The VA inducibility test was performed by burst S1S1 ventricular pacing (cycle length decreasing from 250 ms to failure of 1:1 ventricular capture) was