Noninvasive low-frequency electromagnetic stimulation of the left stellate ganglion reduces myocardial infarction-induced ventricular arrhythmia

Noninvasive magnetic stimulation has been widely used in autonomic disorders in the past few decades, but few studies has been done in cardiac diseases. Recently, studies showed that low-frequency electromagnetic field (LF-EMF) might suppress atrial fibrillation by mediating the cardiac autonomic nervous system. In the present study, the effect of LF-EMF stimulation of left stellate ganglion (LSG) on LSG neural activity and ventricular arrhythmia has been studied in an acute myocardium infarction canine model. It is shown that LF-EMF stimulation leads to a reduction both in the neural activity of LSG and in the incidence of ventricular arrhythmia. The obtained results suggested that inhibition of the LSG neural activity might be the causal of the reduction of ventricular arrhythmia since previous studies have shown that LSG hyperactivity may facilitate the incidence of ventricular arrhythmia. LF-EMF stimulation might be a novel noninvasive substitute for the existing implant device-based electrical stimulation or sympathectomy in the treatment of cardiac disorders.


Effect of LF-EMF stimulation on serum norepinephrine and LSG function.
In comparison with group baseline, serum norepinephrine was decreased from 180.3 ± 6.8 pg/ml to 162.5 ± 5.8 pg/ml at 30 min later and to 160.3 ± 5.2 pg/ml at 90 min later in the LF-EMF group, whereas kept a stable level in the Control group (Fig. 5A). Furthermore, the systolic blood pressure increase in response to LSG stimulation was kept a baseline level in the Control group (Fig. 5B), whereas significantly attenuated by LF-EMF in the LF-EMF group at a voltage of 20-30 V as compared to group baseline (Fig. 5C). Take 25 V for example, the maximal systolic blood pressure increase induced by LSG stimulation was decreased from 88.3 ± 15.4% to 43.1 ± 6.2% (P < 0.01) at 90 min later, whereas kept at about 90% in the Control group (Fig. 5B,C). Figure 6A shows the representative examples of LSG neural activity at baseline, 30 min after LF-EMF stimulation, 90 min after LF-EMF stimulation and 15 min after acute myocardial infarction. Figure 6B,C demonstrates that no significant difference was shown both in the frequency and the amplitude of LSG neural activity between the Control group and the LF-EMF group. As compared to group baseline, LF-EMF stimulation resulted in a significant decrease in LSG neural activity at 30 min and 90 min later, whereas no significant change was caused by sham LF-EMF stimulation (Fig. 6B,C). Furthermore, as compared to baseline, the neural activity was significantly increased after acute myocardial infarction in the Control group (Frequency: 62.5 ± 5.2impulse/min vs 112.2 ± 8.1impulse/min, P < 0.01; Amplitude: 0.18 ± 0.03 mV vs 0.33 ± 0.05 mV, P < 0.01) but kept at a comparable level in the LF-EMF group (Frequency: 60.8 ± 4.8impulse/min vs 65.6 ± 4.8impulse/min, P > 0.05; Amplitude: 0.19 ± 0.02 mV vs 0.18 ± 0.02 mV, P > 0.05).

Discussion
In the present study, we applied LF-EMF at the body surface of LSG. Both the ventricular electrophysiological parameters (APD 90 , PCL, Smax) and autonomic neural activity (serum norepinephrine, LSG function and LSG neural activity) were significantly affected by LF-EMF stimulation. Furthermore, the acute myocardial infarction-induced increased neural activity of LSG was significantly attenuated and the VAs was significantly reduced by LF-EMF. These findings suggested that exposure the LSG to LF-EMF might significantly reduce the neural activity of LSG, therefore reducing the incidence of VAs.
Previous studies have shown that activation of LSG facilitates, whereas inhibition of LSG protects against VAs 4,10 . In the past two decades, TMS has been widely used in clinical neurology 11,12 . Amounts of studies have  shown that high-frequency stimulation increases cortical excitability, whereas low-frequency stimulation decreases neuronal excitability 11,12 . Recently, studies also demonstrated that TMS might affect the cardiac rhythm by modulating the autonomic nervous system 7 . Scherlag et al. 8 showed that high-frequency magnetic stimulation of the vagal nerves might induce atrial tachycardia and atrial fibrillation, which was eliminated after propranolol and atropine injection. Low-frequency stimulation of the vagal nerves, however, reduced the heart rate and decreased the voltage required to induce atrioventricular conduction block 8 . Furthermore, recent study demonstrated that exposure the heart to the LF-EMF might significantly suppress atrial fibrillation and the mechanism might be by modulating the neural activity of atrial ganglionated plexus 9 . In the present study, we found that exposure the LSG to the LF-EMF significantly reduced the serum norepinephrine, neural activity of LSG and VAs. All these indicate that noninvasive LF-EMF might reduce VAs by facilitating the autonomic rebalance, but what underlie the beneficial effects of LF-EMF on LSG was poorly defined.
In the present study, we suggested some possible mechanisms underlying the suppressing of LSG neural activity. Firstly, TMS, as an effective treatment for patients with neural disorders, has been implicated long-lasting  therapeutic effects after the cessation of TMS treatment 13 . Most researchers have contributed these effects to be long-term depression (LTD) and long-term potentiation (LTP) cause the duration of the effects seemed to implicate changes in synaptic plasticity 13 . LTD is caused by low-frequency stimulation or the stimulation of a postsynaptic neuron, whereas LTP is caused by high-frequency stimulation or the stimulation of a presynaptic neuron 13 . Ca+ + signal, which is known to regulate membrane excitability and modulate second messengers related to multiple receptors and signal transduction pathways, has been shown to be the major determinant whether LTD or LTD arises 14,15 . Recently, Scherlag et al. 8 also suggested that LTP or LTD was existed cause exposure the chest to the low-frequency electromagnetic field for 35 mins might result in the suppression of atrial fibrillation for 3 to 4 hours after the application of LF-EMF. In the present study, we also found that pretreatment with LF-EMF might significantly attenuated the acute myocardial infarction-induced activation of LSG neural activity and VAs, suggesting that LTP or LTD might be a potential explain for the salutary effects of LF-EMF stimulation. Secondly, previous studies have shown that TMS might also affect the expression levels of various receptors and other neuromediators, such as β -adrenoreceptors, dopamine 11,16,17 . In the present study, serum norepinephrine was significantly decreased after exposure to the LF-EMF, indicating that modulating the neurotransmitters might be one of the underlying mechanisms underlying the salutary effects of LF-EMF stimulation. Thirdly, previous studies also showed that TMS might also modulate dentritic sprouting (axon growth) and the density of synaptic contacts, and the authors suggested that these results are associated with the Brain-derived neurotrophic factor (BDNF)-tyrosine kinase B (TrkB) signaling system 18,19 . BDNF, as the most abundant neurotrophin in the brain, was reported to be a major contributor to the N-methyl-D-aspartate receptor-dependent LTP and LTD processes 20 . Wang et al. 21 demonstrated that low-frequency TMS might reduce BDNF levels. High-frequency stimulation, however, might increase serum BDNF levels and the affinity of BDNF for TrkB receptors. Furthermore, previous studies also showed that trancranial stimulation might result in the changes in neural-related proteins, such as c-fos and tyrosine hydroxylase, which are closely related with the neural remodeling processes 6,13,20 . Autonomic neural remodeling, however, plays a key role in the initiation and maintenance of VAs 4,10 . All these implicate that modulating autonomic neural remodeling might be another mechanism of the antiarrhythmic effect of LF-EMF stimulation. Fourthly, the above mainly shows the underlying mechanisms of LF-EMF stimulation, but how can the LSG perceive the LF-EMF remains unknown. During the past few decades, many mechanisms, which might provide the basis for how the animals detect magnetic fields, have been proposed 22 . However, the magnetoreceptors have not been identified with certainty in any animal, and the mode of transduction for the magnetic sense remains unknown 23 . Recently, Xie et al. hypothesized that the putative magnetoreceptor, the iron-sulphur cluster protein, might combine with the magnetoreception-related photoreceptor cryptochromes to form the basis of magnetoreception in animals and this was corroborated in pigeon retina 24 . Furthermore, Zhang et al. further showed that the cells which had been transfected iron-sulphur cluster protein might response to the remote magnetic stimulation 25 . All these indicate that the iron-sulphur cluster protein might be the potential magnetoreceptor for the animals to detect the magnetic fields.
Though the present study showed wonderful results, but there are some limitations in this study. First, anesthesia with pentobarbital might affect the autonomic nervous system. However, this could be counteracted cause anesthesia was maintained continuously during the whole surgery and conducted in a same fashion in both groups. Second, the coil used in this study is too large to achieve LSG-targeted stimulation without affecting the surrounding tissues. It would be a great step forward if the coils could be technically improved. Third, we only observed the effect of LF-EMF in acute canine model. Fourth, we mainly focused on the autonomic nervous system imbalance, one of the major contributors of post-infarction VAs, cause we intervened the LSG with LF-EMF in this study. It's a great limitation that some other major factors, like area at risk, infarct size, degree of collateral flow and the possibility of any preconditioning pathway were not involved in this study. However, previous studies have shown that LSG activation might facilitate the incidence of VAs, whereas pre-emptive or post-ischemic/ infarction LSG inhibition by blockage or denervation might decrease the incidence of VAs and improve the infarct size, collateral flow, contractile force both in animals [26][27][28][29] and patients 30,31 . Furthermore, studies have shown that LSG stimulation might increase the likelihood of early or delayed afterdepolarization development and the initiation of reentry, thereby resulting in the incidence of VAs [32][33][34] . In this study, LF-EMF stimulation of the LSG might significantly inhibit the neural activity of LSG, thereby reducing the incidence of VAs. Therefore, it's reasonable to refer that improving the above factors might also be the potential mechanisms underlying the beneficial effects of LF-EMF stimulation, but further studies with optimized parameters and all-round considerations are required in the future.
In conclusion, the present study showed that LF-EMF stimulation might significantly reduce the neural function and neural activity of LSG. Exposure the LSG to the LF-EMF might be a feasible method for preventing the acute myocardial infarction-induced VAs. However, larger studies with optimized parameters should be done in the chronic models to verify the beneficial effect of LF-EMF stimulation.

Methods
Animal preparation. Sixteen canines weighing between 20 and 25 kg were included in this study. The experiments were approved by the Animal Ethics Committee of Wuhan University under approval number 2015-0445 and followed the guidelines outlined by the Care and Use of Laboratory Animals of the National Institutes of Health. All surgeries were performed under anesthesia with sodium pentobarbital at an initial dose of 30 mg/kg and a maintenance dose of 60 mg/h. The depth of anesthesia was evaluated by monitoring corneal reflexes, jaw tone, and alterations in cardiovascular indices. The body surface electrocardiogram was recorded throughout the experiment with a computer-based Lab System (Lead 2000B, Jingjiang Inc., Wuhan, China). The core body temperature of the dogs was kept at 36.5 ± 1.5 °C. Left thoracotomy was conducted at the fourth intercostal space. At the end of the experiment, canines were a lethal dose of pentobarbital (100 mg/kg, iv).

LF-EMF.
Repeated LF-EMF was supplied by the magnetic stimulation machine (YRD CCY-I, YiRuiDe Inc., Wuhan, China) with the curve 8 coil located at the body surface of the LSG (Fig. 1A). The LSG was stimulated by intermittent (8 s ON, 10 s OFF) LF-EMF stimulation with the frequency set at 1 HZ and intensity at approximately 90% of motor threshold (Fig. 1B). Motor threshold was defined as the lowest electromagnetic intensity that induced muscle contractions in the proximal forepaw and shoulder.
Monophasic action potential recording. Monophasic action potentials from the left ventricle were recorded with a custom-made Ag-AgCl catheter. A dynamic steady state pacing protocol (S1S1) was performed to determine action potential duration alternans 35 . The pulse train was delivered at an initial cycle length slightly shorter than the sinus cycle length and the drive train of stimuli was maintained for 30 s to ensure a steady state, and then a 2-min interruption was taken to minimize the pacing memory effects. After that, another pulse train with the PCL decreased by 10 ms was delivered until action potential duration alternans appeared. Action potential duration alternans was defined as Δ APD 90 ≥ 10 ms for ≥ 5 consecutive beats 36 . The monophasic action potential recordings were analyzed by the LEAD 2000B work station system (Lead 2000B, Jingjiang Inc. China). The APD 90 was defined as the 90% repolarization duration and the diastolic interval was the time interval from the previous APD 90 point to the activation time of the following beat. As described in previous studies, the dynamic action potential duration restitution curves were constructed from (Diastolic interval, APD 90 ) pairs using Origin 8.0 (OriginLab, Co., Northampton, MA, USA) 35,37 . Slope of the shortest diastolic interval was defined as S max .
Measurements of heart rate variability. Spectral power for heart rate variability was analyzed on 5-minute electrocardiogram recording segments and an autoregressive algorithm was used to analyze digitized signals from the electrocardiographic recordings. The following power spectral variables were determined: HF, LF and LF/HF 38 . Neural recording from the LSG. To record the neural activity of the LSG, one tungsten-coated microelectrode was inserted into the fascia of the LSG and one ground lead was connected to the chest wall. The signal of the LSG was recorded with a PowerLab data acquisition system (8/35, AD Instruments, Australia) and amplified by an amplifier (DP-304, Warner Instruments, Hamden, CT, USA). The band-pass filters were set at 300 Hz to 1 kHz and the amplification ranges from 30 to 50 times 39 . The neural activity, deflections with a signal-to-noise ratio greater than 3:1, was manually determined as described in our previous studies 39-41 . LSG function. LSG function was measured as the LSG stimulation-induced maximal change in systolic blood pressure as described in our previous study 38 . High frequency stimulation (20 Hz, 0.1 ms pulse duration) was applied to the LSG using a stimulator (Grass-S88; Astro-Med, West Warwick, RI, USA). The voltage ranged from 20 V to 30 V and increased by 5 V. To eliminate the residual effect of the LSG stimulation, each stimulation should be less than 30 s and the next stimulation should be not be taken until the blood pressure returned to a normal level.