Icariin, a Novel Blocker of Sodium and Calcium Channels, Eliminates Early and Delayed Afterdepolarizations, As Well As Triggered Activity, in Rabbit Cardiomyocytes

Icariin, a flavonoid monomer from Herba Epimedii, has confirmed pharmacological and biological effects. However, its effects on arrhythmias and cardiac electrophysiology remain unclear. Here we investigate the effects of icariin on ion currents and action potentials (APs) in the rabbit myocardium. Furthermore, the effects of icariin on aconitine-induced arrhythmias were assessed in whole rabbits. Ion currents and APs were recorded in voltage-clamp and current-clamp mode in rabbit left ventricular myocytes (LVMs) and left atrial myocytes (LAMs), respectively. Icariin significantly shortened action potential durations (APDs) at 50 and 90% repolarization (APD50 and APD90) and reduced AP amplitude (APA) and the maximum upstroke velocity (Vmax) of APs in LAMs and LVMs; however, icariin had no effect on resting membrane potential (RMP) in these cells. Icariin decreased the rate-dependence of the APD and completely abolished anemonia toxin II (ATX-II)-induced early afterdepolarizations (EADs). Moreover, icariin significantly suppressed delayed afterdepolarizations (DADs) and triggered activities (TAs) elicited by isoproterenol (ISO, 1 μM) and high extracellular calcium concentrations ([Ca2+]o, 3.6 mM) in LVMs. Icariin also decreased INaT in a concentration-dependent manner in LAMs and LVMs, with IC50 values of 12.28 ± 0.29 μM (n = 8 cells/4 rabbits) and 11.83 ± 0.92 μM (n = 10 cells/6 rabbits; p > 0.05 vs. LAMs), respectively, and reversed ATX-II-induced INaL in a concentration-dependent manner in LVMs. Furthermore, icariin attenuated ICaL in a dose-dependent manner in LVMs. The corresponding IC50 value was 4.78 ± 0.89 μM (n = 8 cells/4 rabbits), indicating that the aforementioned current in LVMs was 2.8-fold more sensitive to icariin than ICaL in LAMs (13.43 ± 2.73 μM; n = 9 cells/5 rabbits). Icariin induced leftward shifts in the steady-state inactivation curves of INaT and ICaL in LAMs and LVMs but did not have a significant effect on their activation processes. Moreover, icariin had no effects on IK1 and IKr in LVMs or Ito and IKur in LAMs. These results revealed for the first time that icariin is a multichannel blocker that affects INaT, INaL and ICaL in the myocardium and that the drug had significant inhibitory effects on aconitine-induced arrhythmias in whole rabbits. Therefore, icariin has potential as a class I and IV antiarrhythmic drug.

Icariin, a flavonoid monomer from Herba Epimedii, has confirmed pharmacological and biological effects. However, its effects on arrhythmias and cardiac electrophysiology remain unclear. Here we investigate the effects of icariin on ion currents and action potentials (APs) in the rabbit myocardium. Furthermore, the effects of icariin on aconitine-induced arrhythmias were assessed in whole rabbits. Ion currents and APs were recorded in voltage-clamp and current-clamp mode in rabbit left ventricular myocytes (LVMs) and left atrial myocytes (LAMs), respectively. Icariin significantly shortened action potential durations (APDs) at 50 and 90% repolarization (APD 50 and APD 90 ) and reduced AP amplitude (APA) and the maximum upstroke velocity (V max ) of APs in LAMs and LVMs; however, icariin had no effect on resting membrane potential (RMP) in these cells. Icariin decreased the rate-dependence of the APD and completely abolished anemonia toxin II (ATX-II)-induced early afterdepolarizations (EADs). Moreover, icariin significantly suppressed delayed afterdepolarizations (DADs) and triggered activities (TAs) elicited by isoproterenol (ISO, 1 µM) and high extracellular calcium concentrations ([Ca 2+ ] o , 3.6 mM) in LVMs. Icariin also decreased I NaT in a concentration-dependent manner in LAMs and LVMs, with IC 50 values of 12.28 ± 0.29 µM (n = 8 cells/4 rabbits) and 11.83 ± 0.92 µM (n = 10 cells/6 rabbits; p > 0.05 vs. LAMs), respectively, and reversed ATX-II-induced I NaL in a concentration-dependent manner in LVMs. Furthermore, icariin attenuated I CaL in a dose-dependent manner in LVMs. The corresponding IC 50 value was 4.78 ± 0.89 µM (n = 8 cells/4 rabbits), indicating that the aforementioned current in LVMs was 2.8-fold more sensitive to icariin than I CaL in LAMs (13.43 ± 2.73 µM; n = 9 cells/5 rabbits). Icariin induced leftward shifts in the steady-state inactivation curves of I NaT and I CaL in LAMs and LVMs but did not have a significant effect on their activation processes. Moreover, icariin had no effects on I K1 and I Kr in LVMs or I to and I Kur in LAMs. These results revealed for the first time that icariin is a multichannel blocker that affects I NaT , I NaL and I CaL in the myocardium and that the drug had significant inhibitory effects on aconitine-induced arrhythmias in whole rabbits. Therefore, icariin has potential as a class I and IV antiarrhythmic drug.

INTRODUCTION
Icariin (C 33 H 40 O 15 , molecular weight = 676.7), the chemical structure of which has been reported by Tao et al. (2013), is a flavonoid monomer extracted from Herba Epimedii. It has been confirmed to have a variety of pharmacological and biological effects, including anti-inflammatory (Xu et al., 2010;Tao et al., 2013), antioxidant (Liu et al., 2004;Huang et al., 2014), anti-tumor (Wang et al., 2011;Tan et al., 2016), and neuroprotective effects . It was recently reported that icariin protected H9c2 cells from apoptosis by inhibiting endoplasmic reticular stress and the reactive oxygen speciesdependent JNK and p38 pathways Zhou et al., 2014). Icariin was also found to ameliorate cardiac remodeling and left ventricular dysfunction in rats with heart failure by attenuating matrix metalloproteinase activity and myocardial apoptosis (Song et al., 2011). Furthermore, icariin protected the heart from ischemia-reperfusion injury through PI3K-Akt signaling pathway activation (Ke et al., 2015). Additionally, Sun et al. (2011) found that icariin facilitated the differentiation of mouse embryonic cells into cardiomyocytes. The results of these studies indicate that icariin has cardioprotective effects. However, the effects of icariin on APs and ion channels in cardiomyocytes have not been reported. Thus, the aim of the present study was to investigate the effects of icariin on action potentials (APs), ion currents in cardiomyocytes, as well as arrhythmias in whole rabbits, and to further investigate the medicinal value of icariin for the treatment of heart diseases.

Cardiomyocyte Isolation
The animal experiments performed in this investigation conformed to the Guide for Care and Use of Laboratory Animals of Hubei Province, China, and the study protocol was approved by Experimental Animal Ethics Committee of Wuhan University of Science and Technology. Hearts from adult New Zealand white rabbits (1.5-2 kg) of either sex were quickly Abbreviations: AP, action potential; LAM, left atrial myocyte; LVM, left ventricular myocyte; APD, action potential duration; APD 50 and APD 90 , APD at 50 and 90% repolarization; V max , maximum upstroke velocity of AP; APA, AP amplitude; RMP, resting membrane potential; RD, rate dependence of the APD; RRD, reverse rate dependence of the APD; ATX-II, anemonia toxin II; EAD, early afterdepolarization; DAD, delayed afterdepolarization; TA, triggered activity; ISO, isoproterenol; [Ca 2+ ] o , extracellular calcium concentration; I NaT , transient sodium current; I CaL , L-type calcium current; I NaL , late sodium current; I K1 , inward rectifier potassium current; I Kr , rapid component of delayed rectifier potassium current; I to , transient outward potassium current; I Kur , ultra-rapid delayed rectifier potassium current; CL, cycle length; ventricular premature contraction (VPC); ventricular tachycardia (VT); ventricular fibrillation (VF). removed and retrogradely perfused by the Langendorff method, as described previously (Wu, 2005), with Ca 2+ -free Tyrode solution containing the following compounds (in mM): 135 NaCl, 5.4 KCl, 1.0 MgCl 2 , 10 glucose, 0.33 NaH 2 PO 4 , and 10 HEPES, pH 7.4 with NaOH for 5 min. Then, hearts were perfused with Ca 2+ -free Tyrode solution containing collagenase type I (1 g/l) and bovine serum albumin (BSA, 1 g/l) for 30-40 min before being perfused with KB solution for another 5 min. After perfusion, the left ventricle and left atrium were isolated and gently agitated in KB solution. The cardiomyocytes were filtered through a nylon mesh and stored in KB solution containing the following compounds (in mM): 70 KOH, 40 KCl, 20 KH 2 PO4, 50 glutamic acid, 20 taurine, 0.5 EGTA, 10 glucose, 10 HEPES, and 3.0 MgSO4, pH 7.4 with KOH. All solutions used in this study were saturated with 95% O 2 and 5% CO 2 and were maintained at 37 • C.

AP Recordings
For AP recording, quiescent and Ca 2+ -tolerant cardiomyocytes were bathed in standard Tyrode solution. The patch pipette solution contained the following reagents (in mM): 110 Kaspartate, 30 KCl, 5 NaCl, 10 HEPES, 0.1 EGTA, 5 MgATP, 5.0 creatine phosphate, and 0.05 CAMP, pH 7.2 with KOH. When filled with pipette solution, the electrode resistance was in the range of 1.5-2.5 M . APs were induced in current-clamp mode by 1.5-fold diastolic threshold current pulses of 5 ms in duration at different pacing cycle lengths (CLs).

Ion Current Recordings
Currents were recorded with a patch-clamp amplifier (EPC9, Heka electronic, Lambrecht, Pfalz, Germany) and were filtered at 2 kHz and digitized at 10 kHz.
The bath solution used for I NaL recording contained the following compounds (in mM): 135 NaCl, 5.4 CsCl, 1.0 MgCl 2 , 10 glucose, 0.33 NaH 2 PO 4 , 0.3 BaCl 2 , 10 HEPES, and 1.8 CaCl 2 , pH 7.4 with NaOH, and 1 µM nicardipine was added to the bath solution to block I CaL . The pipette solution used for this experiment was the same as that used for I NaT recording I NaL was recorded using a 300-ms depolarization pulse at a HP of −90 mV, followed by pulses with potentials that were increased from −80 mV to +60 mV in 10-mV increments, and was measured at 200 ms in depolarization testing pulse.
The bath solution (except nicardipine) used for I CaL recording was the same as that used for I NaL recording. The electrode was filled with an internal solution containing the following compounds (in mM): 80 CsCl, 60 CsOH, 40 aspartate acid, 0.65 CaCl 2, 5.0 HEPES, 10 EGTA, 5.0 MgATP, and 5.0 Na 2 -creatine phosphate, pH 7.2 with CsOH. I CaL was determined using 300ms voltage steps with potentials that were increased from −40 mV to +50 mV in 5-mV increments at 0.5 Hz. For the steadystate inactivation protocol, I CaL was determined using 2,000-ms conditional prepulses with potentials that were increased from −50 mV to 0 mV in 5-mV increments-using a HP of −40 mV-followed by a 300-ms test pulse at 0 mV.
For I K1 recording, the cells were bathed with Tyrode solution, and 1 µM nicardipine was used to block I CaL . The internal solution contained the following compounds (in mM): 140 KCl, 1.0 MgCl 2 , 5.0 K 2 ATP, 10 EGTA, and 5.0 HEPES, pH 7.3 with KOH.
The bath solution and pipette solution used to record I Kur were the same as those used to record I to , but the pulse protocol was different from that used to record I to (see the Results Section).

Aconitine-Induced Arrhythmias in Whole Rabbits
Twenty healthy New Zealand rabbits were randomly divided into two groups (n = 10 for each group): normal saline (NS) and icariin. In the NS group, saline was injected intraperitoneally within half an hour before the experiment. In the icariin group, 3 mg/kg icariin was injected intraperitoneally within half an hour before the experiment. At the beginning of the experiments, both groups of rabbits were anesthetized with xylazine (7.5 mg/kg, i.m.) and ketamine (30 mg/kg, i.v.) through ear vein injection. A standard limb lead II electrocardiogram (ECG) was recorded using the BL-420F data acquisition and analysis system (Chengdu TaiMeng, Sichuan, China) for 120 min following the application of 2 µg/kg/min aconitine, which was injected by a constant velocity pump and used to induce arrhythmias. The onset time and onset dosage of aconitine that induced ventricular premature contraction (VPC), ventricular tachycardia (VT) and ventricular fibrillation (VF) were measured.

Drugs and Reagents
Icariin (purity >97%) was obtained from Sigma Aldrich (Saint Louis, MO, USA). Collagenase type I and CsCl were purchased from Gibco (GIBCO TM, Invitrogen Co., Paisley, UK). BSA and HEPES were obtained from Roche (Basel, Switzerland), and the other chemicals were obtained from Sigma Aldrich (Saint Louis, MO, USA). Dimethyl sulfoxide (DMSO) was used to dissolve icariin to obtain a 1 mM stock solution. The final concentration of the DMSO added to the bath solution was less than 0.1%.

Data Analysis
Fitmaster (v2x32, HEKA) was used for data analysis, and the figures were plotted by Origin 8.0 (OriginLab Co., MA, USA). All data were expressed as the mean ± SD. Data pertaining to the I NaT and I CaL steady-state activation and steady-state inactivation relationships were fitted by the Boltzmann equation, is the membrane potential, V 1/2 is the half-activation and half-inactivation potential, k is the slope factor, and Y is relative conductance (G/G max , steady-state activation) and relative current (I/I max , steady-state inactivation). The dose-response relationship curves for the effects of icariin on I NaT and I CaL were fitted to the Hill equation, (I control −I drug )/I drug = E max /[1 + (IC 50 /C) n ], where I control and I drug represent the amplitude of I NaT and I CaL obtained in the absence and presence of icariin, respectively, E max is the maximum inhibition, IC 50 is the concentration of icariin at which its half-maximum inhibitory effects are exerted, C is the concentration of icariin, and n is the Hill coefficient. Current density was calculated by dividing the current amplitude by the cell capacitance. The statistical significance of the differences between two groups was determined by Student's t-test, and mean comparisons among multiple groups were performed by one-way analysis of variance (ANOVA) followed by Bonferroni's test. P < 0.05 was considered significant. Figure 1A shows the representative morphologies of a single isolated left ventricular myocyte (LVM, left) and left atrial myocyte (LAM, right). The rod-shaped LVM had glossy and smooth edges, as well as the typical transverse striations. The LAM was more slender than the LVM.

Effects of Icariin on Action Potentials
APs were consecutively recorded by 5-ms and 1.5-fold threshold current pulses at 1 Hz in the absence and presence of icariin. Icariin attenuated AP amplitude (APA) and the maximum upstroke velocity (V max ), shortened action potential durations (APDs) at 50 and 90% repolarization (APD 50 and APD 90 , respectively) in a concentration-dependent manner in LVMs and LAMs. However, icariin had no significant effects on resting membrane potential (RMP) at concentrations of 5 and 10 µM ( Figure 1B; Table 1).

Effects of Icariin on Cellular Arrhythmias
In the present study, we used 10 nM anemonia toxin II (ATX-II) and a stimulation frequency of 0.25 Hz to elicit early afterdepolarizations (EADs) in LVMs. ATX-II significantly lengthened the APD from 179.78 ± 18.64 ms to 1186.44 ± 93.13 ms and induced EADs in 7 of 10 cells (70%; n = 10 cells/5 rabbits; Figures 2A-C), and 20 µM icariin decreased the APs prolonged by ATX-II from 1186.44 ± 93.13 ms to 360.08 ± 41.95 ms and completely abolished the EADs induced by ATX-II in seven cells. In another group, to elicit delayed afterdepolarizations (DADs) and triggered activities (TAs) in LVMs, we added 1 µM isoproterenol (ISO) to the external solution and the extracellular calcium concentration was elevated to 3.6 mM following a baseline pacing CL of 9,000 ms and on top of that 15 beats with a stimulation frequency of 2.5 Hz. DADs were noted in 6 of 9 cells (3 rabbits; 66.7%), and TAs were noted in 3 of 9 cells (33.3%). Administration of 10 µM icariin significantly suppressed the ISO-induced DADs and completely abolished the ISO-induced TAs ( Figure 2D).

Effects of Icariin on I NaT and I NaL
When the effects of icariin on I NaT reached a steady state (3 min), the next concentration of the drug could be added to the external recording solution. Icariin (1, 5, 10, and 20 µM) reduced I NaT in a dose-dependent manner in LVMs and LAMs. Figures 3A,B show the representative recordings for I NaT in LVMs and LAMs, . Data pertaining to APD 90 from 30 sequential curves were averaged. The averaged data for different pacing CLs are shown, n = 9 cells/4 rabbits. & and && p < 0.05 and 0.01 vs. a pacing CL of 500 ms; ** p < 0.01 vs. control at the same pacing CL; ## p < 0.01 vs. 5 µM icariin at the same pacing CL.  respectively, and Figure 3C shows the corresponding currentvoltage relationships in LVMs and LAMs. The IC 50 values for I NaT in LVMs and LAMs were 11.83 ± 0.92 µM (n = 10 cells/6 rabbits) and 12.28 ± 0.29 µM (n = 8 cells/4 rabbits; p > 0.05 LAMs vs. LVMs; Figure 3D) (Figures 3F,H). However, it had no significant effects on the activation process in LVMs and LAMs (Figures 3F,H).
To identify I NaL , we recorded current before and after the application of 4 µM TTX using 300-ms depolarization pulses with potentials ranging from a HP of −90 mV to a potential of −20 mV. TTX (4 µM) had no significant effects on I NaT but decreased the amplitude of I NaL from −0.39 ± 0.004 pA/pF to 0.023 pA/pF (n = 6 cells/3 rabbits; p < 0.01 vs. control), indicating that the TTX-sensitive current was I NaL . Administration of 10 nM ATX-II significantly enhanced I NaL , an effect that was reversed by administration of 1, 10, 20, and 40 µM icariin (n = 6 cells/4 rabbits; Figures 4A,C). The percentage inhibitions by 1, 10, 20, and 40 µM icariin of ATX-II augmented I NaL were 7.8 ± 1%, 29 ± 6.4%, 43.68 ± 5.6%, and 61.4 ± 5.7%. Figure 4B shows the representative current recordings of I NaL at −20 mV that are shown in Figure 4A.

Effects of Icariin on I CaL
To elicit I CaL , we clamped LVMs at −40 mV and then depolarized the cells to +5 mV for 300 ms at 0.2 Hz. As shown in Figure 5A, the I CaL run-down phenomenon lasted for approximately 5 min after membrane rupture in the control condition and then reached a steady state for 15 min (n = 5 cells/2 rabbits). I CaL decreased by 8.5% during the this 5-min period. We performed To investigate the efficiency of the effects of icariin on I CaL in LVMs, we recorded the current sequentially. As shown in Figure 5B, 10 µM icariin was added to the bath solution after the first (1st) current curve (control). I CaL decreased rapidly between the tenth (10th) current curve (45 s after perfusion with icariin) and the thirteenth (13th) current curve (60 s after perfusion with icariin) and then decreased gradually until it reached a steady state (the twenty-seventh current curve). Icariin was washed out after 27th current curve (130 s after perfusion with icariin). I CaL increased rapidly between the 27th current curve and the thirtieth (30th) current curve and then increased gradually until it reached its maximum value (82%) at the fifty-fifth current curve (270 s after perfusion with icariin). The summary data are shown in Figure 5C (n = 10 cells/4 rabbits). The above results indicate that icariin rapidly and reversibly inhibited I CaL in LVMs.
When the effects of icariin on I CaL reached a steady state (2.5 min), the next concentration of the drug could be added to the bath solution. Figure 5D shows the representative I CaL recordings in LVMs after sequential treatments of 0.1, 1, 5, 10 µM icariin and 1 µM nicardipine. Icariin decreased I CaL in a concentration-dependent manner in LVMs, with an IC 50 of 4.78 ± 0.89 µM (n = 8 cells/4 rabbits; Figure 5G). Nicardipine (1 µM) almost completely inhibited I CaL in LVMs in the presence of 10 µM icariin, indicating that I CaL was the nicardipine-sensitive current. Figure 5E shows the representative I CaL recordings in LAMs after sequential treatments of 1, 5, 10, and 20 µM icariin.

Effects of Icariin on Main Potassium Currents
To elicit I K1 in LVMs, we clamped the cells at −40 mV (to inactivate their sodium channels) and depolarized them from −120 mV to +50 mV in 5-mV increments for 400 ms at 0.5 Hz. As shown in Figures 6A,B, icariin (10 and 40 µM) had no effect on I K1 (n = 18 cells/8 rabbits). I Kr in LVMs was elicited using a 3-s depolarization pulse whose potential was increased from a HP of −40 mV to a potential of 50 mV in 10-mV increments before returning to a potential of −40 mV for 5 s. Only the I Kr tail-current (I Kr−tail ) was measured. Icariin (10 and 40 µM) . Consecutive recordings of I CaL evoked by a 300-ms depolarization from a holding potential of −40 mV to 5 mV at 0.2 Hz. The 1st sweep represents the control condition, the 2nd to 27th sweeps represent the period in which icariin exerted its effects on I CaL , and the 27th-55th sweeps indicate the period in which the effects of icariin on I CaL were reversed. The inset represents the time course of the entire process, including the control condition, the icariin perfusion period, and the icariin (Continued)
Representative current recordings of I CaL evoked according to the steady-state inactivation protocol in LVMs (H) and LAMs (J) in the absence and presence of 10 µM icariin. (I). Steady-state activation (n = 14 cells/7 rabbits) and steady-state inactivation (n = 10 cells/4 rabbits) curves of I CaL in LVMs before and after icariin application. (K). Steady-state activation (n = 12 cells/7 rabbits) and steady-state inactivation (n = 10 cells/6 rabbits) curves of I CaL in LAMs before and after icariin application. had no significant effects on I Kr−tail (n = 12 cells/5 rabbits; Figures 6C,D). I to in LAMs was elicited by 400 depolarization voltage steps with potentials that were increased from −80 mV to +50 mV in 10-mV increments, followed by a conditional test in which −40 mV was administered for 100 ms to block sodium currents. Forty micrometer icariin had no significant effect on I to in LAMs (n = 15 cells/6 rabbits; Figures 6E,F). I Kur in LAMs was elicited by an 80-ms prepulse whose potential was increased from a HP of −50 mV to a potential of 30 mV (to inactivate I to ), followed by 140-ms test pulses with potentials that were increased from −40 mV to +60 mV in 10-mV increments-using a HP of −50 mV-after a 50-ms interval before returning to −30 mV. Figure 6G shows the I Kur current-voltage relationship in LAMs in the absence and presence of icariin (20 and 40 µM). Icariin had no significant effect on I Kur (n = 1 5 cells/5 rabbits).

Effects of Icariin on Aconitine-Induced Arrhythmias
In the NS group, VPC, VT, and VF were observed in all 10 rabbits. In the icariin group, VPC, VT and VF occurred in 9, 4 and 1 of 10 rabbits, respectively. Compared with the NS group, icariin application prior to aconitine administration increased the onset time (Figures 7A,B,D) and onset dosage ( Figure 7C). The administration of icariin attenuated the incidence of arrhythmias induced by aconitine ( Figure 7E) and rabbit mortality ( Figure 7F).  (Figure 7). In this study, icariin decreased V max of APs and shortened APDs (APD 50 and APD 90 ) in a concentration-dependent manner in LVMs and LAMs. The abovementioned decrease in APA and V max , which may be associated with the inhibitory effects of I NaT , can reduce conduction velocities, resulting in reentry blockade (Baba et al., 2005). Moreover, the APD shortening induced by icariin may be closely related to I CaL inhibition because icariin does not affect I K1 , I Kr , I to , and I Kur , which also play important roles in maintaining APD. Some drugs displays reverse rate dependence (RRD) of APD property, that is, the effect of a drug to prolong APD may be greater at slow than at fast heart rate, and vice versa. The findings of previous studies suggest that RRD of APD can be induced by enhancing I CaL and inhibiting I Kr or I K1 (Bosch et al., 1998;Virag et al., 2009). RRD of APD enhancement leads to an increase in the cardiac transmural dispersion of the repolarization (Osadchii, 2013), which subsequently facilitates the occurrence of reentrant arrhythmias (Coronel et al., 2009;Maoz et al., 2014). In the present study, icariin attenuated I CaL but had no effect on I Kr or I K1 , indicating that icariin might diminish or not produce RRD. These results suggest that icariin has increased antiarrhythmic efficiency compared with other drugs and that it is safer than its counterparts.
Sodium channels are known as the key targets of class I antiarrhythmic drugs. I NaT is the main depolarization current in AP phase 0 and plays an important role in myocardial excitability and propagations (Goldin, 2002). In this study, icariin decreased the amplitude of I NaT , which caused a decrease in Na + influx. Therefore, the results of this study indicate that icariin can relieve intracellular Na + overload and exerts class I antiarrhythmic drug effects.
I NaL is involved in the AP plateau phase (Kiyosue and Arita, 1989). A variety of pathological conditions, such as ischemia and hypoxia (Saint, 2006), cardiac hypertrophy and heart failure (Valdivia et al., 2005;Guo et al., 2014), can increase I NaL , resulting in an elevated intracellular sodium concentration ([Na + ] i ), as well as a subsequent increase in the intracellular Ca 2+ concentration ([Ca 2+ ] i ) as a result of the activity of a reverse Na + /Ca 2+ exchanger (NCX), leading to Ca 2+ overload resulting in arrhythmia (Kihara and Morgan, 1991;Haigney et al., 1992;Yeh et al., 2008;Tang et al., 2012). On the other hand, increases in I NaL can effectively lengthen the APD, resulting in EADs (Undrovinas et al., 1999). The authors of previous studies found that inhibiting I NaL significantly prevented arrhythmias such as ventricular tachycardia and ventricular fibrillation (Pezhouman et al., 2014;Markandeya et al., 2016). Therefore, I NaL is considered a new target for the treatment of arrhythmias (Undrovinas and Maltsev, 2008). In the present study, icariin reversed the increases in I NaL induced by ATX-II (a known I NaL opener), decreased I CaL , shortened the APD, and suppressed the EADs induced by ATX-II in LVMs. The percentage inhibitions by 1, 10, 20, and 40 µM icariin of ATX-II augmented I NaL were 7.8 ± 1%, 29 ± 6.4%, 43.68 ± 5.6%, and 61.4 ± 5.7%. The percentage inhibitions by 3, 6, and 9 µM ranolazine of ATX II augmented I NaL were 24 ± 6%, 44 ± 8%, and 62 ± 4% . The inhibitory effects of icariin on ATX-II augmented I NaL is weaker than ranolazine (a known I NaL blocker). Icariin can inhibit I CaL and shorten APD, thus we concluded that icariin might inhibit ATX-II-induced arrhythmias by blocking I NaL and I CaL .
I CaL is one of the major inward currents in phase 2 of the AP and regulates Ca 2+ -related physiological processes (Benitah et al., 2010). Extracellular Ca 2+ flows into cardiomyocytes mainly through L-type calcium channels and subsequently causes elevations in [Ca 2+ ] i, which causes the sarcoplasmic reticulum to release large amounts of Ca 2+ into the cytosol, a phenomenon known as Ca 2+ -induced Ca 2+ release, which increases [Ca 2+ ] i further. A large number of studies have shown that various pathological conditions, including ischemia/reperfusion injury (de Diego et al., 2008) and heart failure (Casini et al., 2009), are associated with [Ca 2+ ] i abnormalities, especially intracellular Ca 2+ overload, which plays a crucial role in the genesis of arrhythmias such as ventricular and atrial fibrillation (Kihara and Morgan, 1991;Yeh et al., 2008). Therefore, inhibiting I CaL can facilitate [Ca 2+ ] i reductions that suppress arrhythmias in the above pathological conditions. In this study, icariin decreased the amplitude of I CaL , which caused a decrease in Ca 2+ influx. Therefore, icariin exerts class IV antiarrhythmic drug effects by inhibiting I CaL . DADs and TAs can be induced by [Ca 2+ ] i overload caused by the application of ISO and high [Ca 2+ ] o (Shutt et al., 2006;Sicouri et al., 2013). In the present study, icariin significantly suppressed DADs and TAs in LVMs, possibly by inhibiting I CaL . Moreover, the inhibition of I CaL induced by icariin in LVMs was 2.8-fold stronger than that induced by icariin in LAMs. Thus, icariin shows a degree of ventricular selectivity with respect to its inhibitory effects on I CaL .
Elevations in [Ca 2+ ] i increase I NaL by activating the CAMK II and PKC pathways Wu et al., 2015). The increased I NaL elevates [Na + ] i , which increases [Ca 2+ ] i by activating a reverse NCX (Kihara and Morgan, 1991;Haigney et al., 1992;Yeh et al., 2008;Tang et al., 2012). The cellular response may cause or aggravate arrhythmias. In the present study, icariin inhibited both sodium currents (I NaT and I NaL ) and I CaL , which blocked the cellular response more effectively, indicating that icariin may be a more effective antiarrhythmic drug than established medications. I Kr is an important outward current in AP repolarization. Decreases in I Kr lengthen the APD and lead to QT interval prolongation. A variety of noncardiovascular drugs can block I Kr , thereby inducing long QT syndrome and torsade de pointes (TdPs) (Viskin et al., 2003). For example, grepafloxacin, a quinolone antibiotic, was withdrawn from the American drug market because it blocked I Kr significantly and caused excessive QT interval prolongation, resulting in TdPs (Anderson et al., 2001). Therefore, the authors of another study measured I Kr antagonist potency to evaluate the proarrhythmic effects of new drugs (Kim et al., 2016) and found that it did not affect I Kr . In this study, icariin showed no effect on I Kr . Thus, we deemed the compound a safer drug than its established counterparts.
Aconitine, a specific sodium channel agonist, sustained activation of the sodium channels and induced intracellular Na + accumulation leading to intracellular Ca 2+ overload through NCX (Peper and Trautwein, 1967). Moreover, icariin can augment I CaL directly causing intracellular Ca 2+ overload, which may eventually result in arrhythmias (Zhou et al., 2013). In the present study, we found that icariin increased the onset time and onset dosage of aconitineinduced VPC, VT and VF in whole rabbits. It also decreased the incidence of aconitine-induced VT and VF, as well as mortality in rabbits. The above results indicate that icariin shows cardioprotective effects against aconitine-induced arrhythmias. The cardioprotective effects may be due to reduction of I NaT , I NaL and I CaL .

CONCLUSION
In summary, we found for the first time that icariin exerted class I and IV antiarrhythmic agent effects and moderately inhibited I NaL . Icariin inhibits aconitine-induced arrhythmias in whole rabbits. Icariin also suppressed EADs or DADs and TAs induced by ATX-II or ISO and high [Ca 2+ ] o , respectively, by inhibiting I NaT , I NaL , and I CaL , but had no effect on I K1 , I Kr , I to, and I Kur , especially I Kr , which may indicate that icariin is a safer drug than its counterparts. Thus, icariin may have promise as an agent used in the clinical treatment of arrhythmia.

AUTHOR CONTRIBUTIONS
JM designed the research. WJ, MZ, and ZC performed the experiments. ZL, JH, PPZ, YT and PHZ analysis the data. WJ wrote the main text and prepared all of the figures. All authors reviewed and approved this manuscript.

ACKNOWLEDGMENTS
We thank American Journal Experts (AJE) for editing the English of this manuscript.