Effective Modulation by Lacosamide on Cumulative Inhibition of INa during High-Frequency Stimulation and Recovery of INa Block during Conditioning Pulse Train

The effects of lacosamide (LCS, Vimpat®), an anti-convulsant and analgesic, on voltage-gated Na+ current (INa) were investigated. LCS suppressed both the peak (transient, INa(T)) and sustained (late, INa(L)) components of INa with the IC50 values of 78 and 34 μM found in GH3 cells and of 112 and 26 μM in Neuro-2a cells, respectively. In GH3 cells, the voltage-dependent hysteresis of persistent INa (INa(P)) during the triangular ramp pulse was strikingly attenuated, and the decaying time constant (τ) of INa(T) or INa(L) during a train of depolarizing pulses was further shortened by LCS. The recovery time course from the INa block elicited by the preceding conditioning train can be fitted by two exponential processes, while the single exponential increase in current recovery without a conditioning train was adequately fitted. The fast and slow τ’s of recovery from the INa block by the same conditioning protocol arose in the presence of LCS. In Neuro-2a cells, the strength of the instantaneous window INa (INa(W)) during the rapid ramp pulse was reduced by LCS. This reduction could be reversed by tefluthrin. Moreover, LCS accelerated the inactivation time course of INa activated by pulse train stimulation, and veratridine reversed its decrease in the decaying τ value in current inactivation. The docking results predicted the capability of LCS binding to some amino-acid residues in sodium channels owing to the occurrence of hydrophobic contact. Overall, our findings unveiled that LCS can interact with the sodium channels to alter the magnitude, gating, voltage-dependent hysteresis behavior, and use dependence of INa in excitable cells.


Introduction
LCS, a functionalized amino acid, is an antiepileptic and analgesic drug available orally and intravenously in clinical practice [1][2][3][4][5][6]. It is commonly administrated in patients with epilepsy and occasionally in patients with trigeminal neuralgia or other neuropathic pain [1,5,[7][8][9][10][11][12]. A growing body of evidence demonstrated the effectiveness and tolerability of LCS in patients with epilepsy [13][14][15][16][17][18][19]. Earlier reports have shown that this compound can decrease the frequency of interictal spikes as well as high-frequency oscillations in mesial temporal lobe epilepsy or refractory focal epilepsy [13,20]. The antiepileptic ability 2 of 17 of LCS was proposed mainly because it selectively enhances the voltage-dependence of the slow inactivation of Na + current (I Na ) [5,7,[21][22][23]. In addition to its antiepileptic effects, additional broad effects of LCS have been observed clinically [3,4]. For instance, a recent report showed that LCS could alter electrocardiographic changes in a status epilepticus animal model [24]. Another report showed that LCS could induce personality changes, which disappeared after its discontinuation [25]. Additional electrophysiological actions of LCS might explain its multiple therapeutic effects and need to be thoroughly established.
The voltage-gated Na + currents form voltage-gated sodium channels (Na V ) are critical for the generation and propagation of action potentials in excitable membranes [26,27]. These channels can briefly shift from the resting to the open state after depolarization, thereby allowing the flow of Na + from the extracellular solution into the cell under the driving forces of the electrical and chemical gradients. After opening briefly in a voltagedependent manner, the channels are shifted to the inactivated state(s), rendering I Na intense but brief [27]. In addition to the voltage-dependence of the slow inactivation of I Na , LCS has also been shown to enhance the frequency-dependent inhibition of I Na [28]. However, whether the accumulative inhibition of I Na inactivation during repetitive depolarization and/or recovery from I Na inactivation during the preceding conditioning pulse train could be perturbed by LCS has not been adequately investigated. In the current study, the following attempts were undertaken to evaluate how LCS could lead to any adjustments to the magnitude, gating kinetics, voltage-dependent hysteresis (Hys (V) ), and use dependence of I Na residing in two different excitable cells, pituitary tumor (GH 3 ) cells and neuroblastoma (Neuro-2a) cells. The observations could help to delineate the delicate modulation of functional activities in excitable cells occurring in vivo.

Effects of LCS on Voltage-Gated Na + Current (I Na )
For the first stage of experiments, whether LCS would exert any perturbations on I Na was tested. Pituitary tumor (GH3) cells were placed in Ca 2+ -free, Tyrode's solution, which contained 10 mM tetraethylammonium chloride (TEA) and 0.5 mM CdCl 2 to avoid interference by other types of ionic currents, such as K + and Ca 2+ currents. Upon membrane depolarization from a holding potential of −100 mV to −10 mV for 30 milliseconds (msec), the transient I Na (I Na(T) ) was elicited, and it could be attenuated dose-dependently by LCS ( Figure 1A). At one minute after adding 30 µM or 100 µM LCS, the I Na(T) amplitude measured at the beginning of the depolarizing pulse was attenuated to 1012 ± 29 pA (n = 8, p < 0.05) or 783 ± 22 pA (n = 8, p < 0.05) from a control value of 1212 ± 32 pA (n = 8), respectively. When LCS was removed, the current amplitude returned to 1207 ± 31 pA (n = 8).
In addition, late I Na (I Na(L) ) was measured at the end of the depolarizing test pulse. The extent of LCS-mediated inhibition of I Na(L) was higher than that at the start of the pulse (i.e., I Na(T) ). For example, 100 µM LCS decreased the I Na(L) amplitude from 74 ± 9 pA to 29 ± pA (n = 8, p < 0.05). Meanwhile, upon exposure to 100 µM LCS, the slow component in the inactivation time constant (τ) of I Na(T) decreased from 12.2 ± 2.1 to 8.1 ± 1.3 msec (n = 8, p < 0.05). The relationship between LCS concentration and the inhibition of I Na(T) and I Na(L) is constructed in Figure 1B. It demonstrated a differential dose-dependent LCS-mediated inhibition in I Na(T) and I Na(L) elicited by the rapid membrane depolarization. Based on a modified Hill equation described in Materials and Methods, the IC 50 values required for exerting a suppressive effect on I Na(T) and I Na(L) were further estimated as 78 and 34 µM, respectively, reflecting that these two values are distinguishable.   The inhibitory effects of LCS on I Na were verified in another kind of excitable cell: Neuro-2a cells. LCS (30 or 100 µM) also suppressed both the amplitude of I Na(T) and I Na(L) elicited by the rapid membrane depolarization (Figure 2A). Figure 2B shows a differential dose-dependent response of LCS-induced suppression of I Na(T) and I Na(L) in Neuro-2a cells. Based on a modified Hill equation, the IC 50 values to inhibit the I Na(T) and I Na(L) amplitude were 112 and 26 µM, respectively. Thus, LCS can suppress the magnitude of I Na(T) and I Na(L) in a time-and concentration-dependent manner both in GH 3 cells and Neuro-2a cells.

Effects of LCS on Persistent Na + Current (I Na(P) ) Triggered by Isosceles-Triangular Ramp Voltage (V ramp )
The persistent Na + current (I Na(P) ) is activated in the subthreshold voltage range and is also important in epilepsy by enhancing the repetitive firing capability of neurons. In this experiment, I Na(P) was elicited in GH 3 cells by an isosceles-triangular ramp voltage (V ramp ), which consisted of an upsloping (ascending) voltage from −100 mV to +50 mV followed by a downsloping (descending) voltage from +50 mV back to −100 mV in 1 s [29][30][31]. Voltage-dependent hysteresis (Hys (V) ) was observed in the instantaneous current-voltage (I-V) relationship of I Na(P) and presented as two distinct Hys(V) loops of the I Na amplitude: a high-(a counterclockwise direction) threshold loop and a low-(a clockwise direction) threshold loop ( Figure 3A). Figure 3B shows the time course of the inhibitory effect of LCS (30 or 100 µM) on I Na(P) amplitudes activated by double V ramp . Compared with the control situation without LCS, LCS reduced the I Na(P) amplitudes on the loop of Hys (V) ( Figure 3B,C). The 30 and 100 µM LCS attenuated the I Na(P) amplitudes at the level of −5 mV of the ascending limb to 245 ± 21 pA (n = 8, p < 0.05) and 212 ± 17 pA (n = 8, p < 0.05) from control values of 313 ± 24 pA (n = 8), and the I Na(P) amplitudes at the level

Effects of LCS on Persistent Na + Current (INa(P)) Triggered by Isosceles-Triangular Ramp Voltage (Vramp)
The persistent Na + current (INa(P)) is activated in the subthreshold voltage range and is also important in epilepsy by enhancing the repetitive firing capability of neurons. In this experiment, INa(P) was elicited in GH3 cells by an isosceles-triangular ramp voltage (Vramp), which consisted of an upsloping (ascending) voltage from −100 mV to +50 mV followed by a downsloping (descending) voltage from +50 mV back to −100 mV in 1 s [29][30][31]. Voltage-dependent hysteresis (Hys(V)) was observed in the instantaneous current- The panels on the right side indicate the expanded records (i.e., I Na(L) and I Na(T) ) from each purple dashed box. (B) Dose-dependent relationship of LCS on I Na(T) (peak component, blue squares) and I Na(L) (sustained component, red circles) evoked by short membrane depolarization from −100 to −10 mV (mean ± SEM; n = 8 for each point). The continuous gray line represents the best fit to a modified Hill equation, as described in Section 4. direction) threshold loop ( Figure 3A). Figure 3B shows the time course of the inhibitory effect of LCS (30 or 100 μM) on INa(P) amplitudes activated by double Vramp. Compared with the control situation without LCS, LCS reduced the INa(P) amplitudes on the loop of Hys(V) (Figures 3B,C). The 30 and 100 μM LCS attenuated the INa(P) amplitudes at the level of −5 mV of the ascending limb to 245 ± 21 pA (n = 8, p < 0.05) and 212 ± 17 pA (n = 8, p < 0.05) from control values of 313 ± 24 pA (n = 8), and the INa(P) amplitudes at the level of −60 mV of the descending limb to 151 ± 14 pA (n = 8, p < 0.5) and 101 ± 11 pA (n = 8, p < 0.05) from control values of 198 ± 17 pA (n = 8), respectively. The attenuation of INa(P) was reversed by application of tefluthrin (Tef, 10 μM), which is an activator of INa [30].

Effect of LCS on Window I Na (I Na(W) ) Elicited by a Short Ascending V ramp
The next question is if the magnitude of I Na(W) in response to the rapid ascending V ramp can be modified by LCS. The instantaneous I Na(W) was evoked by an ascending V ramp from −80 to +40 mV for 30 msec (i.e., a ramp speed of 4 mV/msec) [32,33]. As shown in Figure 4A, the amplitude of I Na(W) was markedly reduced within one minute while Neuro-2a cells were exposed to LCS. LCS (100 or 300 µM) decreased the amplitude of I Na(W) measured at the level of −10 mV from a control value of 602 ± 27 pA (n = 7) to 442 ± 21 pA (n = 7, p < 0.05) or 296 ± 18 pA (n = 7, p < 0.05), respectively. After washout of LCS, the current amplitude at −10 mV was returned to 594 ± 24 pA (n = 7). Figure 4B illustrates a summary graph demonstrating the changes in ∆area of V ramp -elicited I Na(W) measured at the voltage between −40 and +40 mV. It showed that LCS effectively diminished the I Na(W) 's area and Tef (10 µM) could reverse the LCS-mediated decrease in V ramp -elicited I Na(W) 's area.

Effect of LCS on Window INa (INa(W)) Elicited by a Short Ascending Vramp
The next question is if the magnitude of INa(W) in response to the rapid ascending Vramp can be modified by LCS. The instantaneous INa(W) was evoked by an ascending Vramp from −80 to +40 mV for 30 msec (i.e., a ramp speed of 4 mV/msec) [32,33]. As shown in Figure  4A, the amplitude of INa(W) was markedly reduced within one minute while Neuro-2a cells were exposed to LCS. LCS (100 or 300 μM) decreased the amplitude of INa(W) measured at the level of −10 mV from a control value of 602 ± 27 pA (n = 7) to 442 ± 21 pA (n = 7, p < 0.05) or 296 ± 18 pA (n = 7, p < 0.05), respectively. After washout of LCS, the current amplitude at −10 mV was returned to 594 ± 24 pA (n = 7). Figure 4B   . The voltage protocol is illustrated in the inset, and the downward deflection indicates instantaneous inward current (i.e., I Na(W) ) elicited by a short ascending V ramp . (B) Summary graph demonstrating attenuating effect of LCS (100 or 300 µM) and LCS (300 µM) plus Tef (10 µM) on the ∆area of I Na(W) (mean ± SEM; n = 7 for each point). Each area in this work was measured at the voltages ranging between −40 and +40 mV during the ascending V ramp . * Significantly different from control (p < 0.05), ** significantly different from LCS (100 µM) group (p < 0.05), and + significantly different from LCS (300 µM) group (p < 0.05).

Effect of LCS on the Cumulative Inhibition of I Na during a Train of Depolarizing Pulses
I Na(T) inactivation was previously demonstrated to accumulate before being activated during repetitive short pulses, which consisted of repetitive depolarization from −80 mV to −10 mV for 1 s with 40 msec in each pulse at a rate of 20 Hz [32,34]. To see the effects of LCS on the cumulative inhibition of I Na , the I Na(T) or I Na(L) amplitude with or without treatment of LCS was simultaneously measured at the beginning or the end of each depolarizing pulse. Without LCS, the I Na(T) or I Na(L) inactivation evoked by a 1 s repetitive depolarization showed decaying τ values of 86 ± 9 or 234 ± 22 msec (n = 8), respectively ( Figure 5). This indicates a pronounced time-dependent decay of I Na(T) or I Na(L), which can be fitted with a single-exponential process. Under 30 and 100 µM LCS, the exponential time course of I Na(T) or I Na(L) elicited by the same train of depolarizing pulses was shortened to 64 ± 7 msec (n = 8, p < 0.05) and 31 ± 6 msec (n = 8, p < 0.05), or to 143 ± 9 msec (n = 8, p < 0.05) and 104 ± 6 msec (n = 8, p < 0.05), respectively. Thus, that accumulative inactivation of the current can be strikingly enhanced in the cells upon LCS exposure apart from a decrease in I Na(T) and I Na(L) amplitude.
This indicates a pronounced time-dependent decay of INa(T) or INa(L), which can be fitted with a single-exponential process. Under 30 and 100 μM LCS, the exponential time course of INa(T) or INa(L) elicited by the same train of depolarizing pulses was shortened to 64 ± 7 msec (n = 8, P < 0.05) and 31 ± 6 msec (n = 8, p < 0.05), or to 143 ± 9 msec (n = 8, p < 0.05) and 104 ± 6 msec (n = 8, p < 0.05), respectively. Thus, that accumulative inactivation of the current can be strikingly enhanced in the cells upon LCS exposure apart from a decrease  (Figures 6A,B). In addition, the LCS-mediated reduction in decaying τ value of INa(T) could be partially reversed by veratridine ( Figure 6C). Veratridine was a sodium channel agonist and was recently reported to modify the gating of the human NaV1.7 channel [35]. The effects of LCS on the cumulative inhibition of I Na were also verified in Neuro-2a cells. Consistently, LCS (100 µM) diminished the I Na(T) magnitude and decreased the decaying τ value of I Na(T) during repetitive depolarizations from a control value of 79 ± 9 msec to 39 ± 5 msec (n = 7, p < 0.05) ( Figure 6A,B). In addition, the LCS-mediated reduction in decaying τ value of I Na(T) could be partially reversed by veratridine ( Figure 6C).
Veratridine was a sodium channel agonist and was recently reported to modify the gating of the human NaV1.7 channel [35].

Modification by LCS of the Recovery Process of INa(T) Inactivation following Conditioning Train of Depolarizing Stimuli
Earlier investigations have disclosed a unique type of recovery from INa(T) inactivation evoked by a train of the preceding conditioning depolarizing stimuli [34,36,37]. The preceding conditioning train was composed of twenty 40 msec pulses separated by 5 msec

Modification by LCS of the Recovery Process of I Na(T) Inactivation following Conditioning Train of Depolarizing Stimuli
Earlier investigations have disclosed a unique type of recovery from I Na(T) inactivation evoked by a train of the preceding conditioning depolarizing stimuli [34,36,37]. The preceding conditioning train was composed of twenty 40 msec pulses separated by 5 msec intervals at −80 mV for 1 s (Figure 7A, top). Following such a conditioning train, the I Na was produced by a two-step voltage protocol to measure the recovery time course of the current. The two-step voltage protocol included the first step that consisted of a 30 msec pulse from −80 to −10 mV and a second step that consisted of pulses for a variable length of time in a geometric progression (common ratio = 2) from −80 to −10 mV. The relative amplitude of recovery for current inactivation during this protocol was determined by the second step. intervals at −80 mV for 1 s (Figure 7A, top). Following such a conditioning train, the INa was produced by a two-step voltage protocol to measure the recovery time course of the current. The two-step voltage protocol included the first step that consisted of a 30 msec pulse from −80 to −10 mV and a second step that consisted of pulses for a variable length of time in a geometric progression (common ratio = 2) from −80 to −10 mV. The relative amplitude of recovery for current inactivation during this protocol was determined by the second step. In the control situation (i.e., without LCS), the recovery from INa(T) inactivation elicited by the preceding conditioning depolarizing stimuli was noticed to emerge in a biphasic manner including a rapid rising recovery phase followed by a late slow phase ( Figure 7A). Figure 7B. The experimental data points were well fitted with a sum of two exponential functions, i.e., fast (τfast) and slow time constant (τslow), as elaborated in Materials and Methods. Upon exposure to LCS (30 or 100 μM), both τfast and τslow in the recovery time course of current inactivation was increased (p < 0.05) compared to the control situation (Table 1). The blue or red smooth curve taken with or without the application of LCS was least-squares fitted with a two-exponential function, as detailed in Section 4, while the parameters are illustrated in Table 1.

Discussion
The principal finding in this study was that LCS could suppress INa in a time-, concentration-and frequency-dependent manner identified in GH3 and Neuro-2 cells. In the present investigations, the non-equilibrium voltage-dependent hysteresis (Hys(V)) of INa(P) was observed by an upright isosceles-triangular Vramp with a duration of 1 s (Figure 3). This implies that the INa(P) magnitude is contingent on the pre-existing state(s) or conformation(s) of the NaV channel. There are two types of triangular Vramp-elicited In the control situation (i.e., without LCS), the recovery from I Na(T) inactivation elicited by the preceding conditioning depolarizing stimuli was noticed to emerge in a biphasic manner including a rapid rising recovery phase followed by a late slow phase ( Figure 7A). It indicates a train of stimuli enabling the cells to resist the recovery from I Na(T) inactivation. The recovery time course of this biphasic-manner I Na(T) inactivation was constructed in Figure 7B. The experimental data points were well fitted with a sum of two exponential functions, i.e., fast (τ fast ) and slow time constant (τ slow ), as elaborated in Materials and Methods. Upon exposure to LCS (30 or 100 µM), both τ fast and τ slow in the recovery time course of current inactivation was increased (p < 0.05) compared to the control situation (Table 1).

Discussion
The principal finding in this study was that LCS could suppress I Na in a time-, concentration-and frequency-dependent manner identified in GH 3 and Neuro-2 cells. LCS induced differential inhibitions of I Na(T) and I Na(L) activated by a short depolarizing pulse. It also suppressed the high-or low-threshold amplitude of I Na(P) elicited by the upright isosceles-triangular V ramp at either the upsloping or downsloping limb leading to a striking reduction in Hys (V) strength of the current. The accumulative inhibition of I Na(T) and I Na(L) during a train of depolarizing pulses was substantially enhanced by LCS, and the values of τ fast and τ slow in the recovery time course during the preceding conditioning train of depolarizing pulses were increased. Taken together, LCS could modify the magnitude, gating properties, use-dependence, and Hys (V) behaviors of I Na , leading to the inhibition of I Na(T) , I Na(L) , I Na(P), and I Na(W) .
In the present investigations, the non-equilibrium voltage-dependent hysteresis (Hys (V) ) of I Na(P) was observed by an upright isosceles-triangular V ramp with a duration of 1 s (Figure 3). This implies that the I Na(P) magnitude is contingent on the pre-existing state(s) or conformation(s) of the Na V channel. There are two types of triangular V ramp -elicited I Na(P) , that is, low-threshold (i.e., activated at a voltage range near the resting potential) elicited upon the downsloping end of the triangular V ramp , and high-threshold (i.e., activated at a voltage range near the maximal I Na(T) ) elicited on the upsloping end of such V ramp [28]. LCS attenuated both types of triangular V ramp -elicited I Na(P) of the Hys (V) behaviors and Tef could effectively reverse this LCS-induced reduction in Hys (V) 's strength in the current. In the literature, this Hys (V) behavior of triangular V ramp -elicited I Na(P) has been demonstrated to link to the magnitude of background Na + currents closely, and I Na(L) and I Na(P) during an extended period of time are likely to share the same Na V channels [29,31,34,38,39]. This result indicates that LCS could diminish the background Na + currents conductance and reduce the subthreshold potential or depolarization drive in these excitable cells.
In addition, the time-dependent decline in I Na(T) and I Na(L) during a 20 Hz train of depolarizing voltage commands (i.e., 40 msec pulses applied from −80 to −10 mV at a rate of 20 Hz for 1 sec) was observed in an exponential fashion, as shown in Figures 5 and 6. This accumulative inhibition is a use-dependent property of I Na(T) and I Na(L) during rapid repetitive stimuli or high-frequency firing [32,34,[40][41][42][43]. LCS could enhance this accumulative inhibition through reducing not only the amplitude but also the τ value of I Na(T) and I Na(L) . The LCS-mediated enhancement was only reversed partially by veratridine, a sodium channel agonist. This means that LCS would lead to a "loss-of-function" change of Na + channels and keep the Na + channels in inactivated states during repetitive depolarization to prevent excessive excitability. In addition to LCS, a recent study demonstrated that other sodium blockers, such as lidocaine, also had the use-dependent inhibition of corneal nerve activity [44]. It is consistent with the current investigations.
It is important to emphasize that the LCS increased the τ fast and τ slow of recovery from the I Na(T) block elicited by the preceding conditioning train pulse and enlarged the value of A, as summarized in Table 1. This preceding conditioning train pulse consisted of rapid repetitive pulses, which mimics high-frequency firing on excitable cells and would cause a large fraction of Na V channels to shift toward the slowly recovering pool. The value of A indicates the fraction of the slow recovering pool of Na V channels. Taken together, LCS could cause a larger fraction of the Na V channels in an inactivated state and a longer recovery from inactivation after rapid repetitive stimuli or high-frequency firing. These results are partly relevant to the previous study that showed slow inactivation enhancement by LCS [21].
In this work, we further investigated how the protein of Na V could be delicately docked with LCS by using PyRx software (https://sourceforge.net/projects/pyrx/; accessed on 17 July 2022). The predicted binding sites with LCS are demonstrated in Figure 8. The LCS could form hydrophobic contacts with several residues, including Gly 76, Trp 77, Phe 80, Gln 122, Leu 125, and Leu 126. Phe 80 can interact with the N'-acetylamino acid N'-benzyl amide unit of LCS, Gly 76 and Trp 77 can dock to the linker that connects the two-aryl group, while Gln 122, Leu 125, and Leu 126 are noted to interact with the terminal aromatic ring. These three functional groups (i.e., N'-acetylamino acid N'-benzyl amide unit, the linker that connects the two aryl groups, and the terminal aromatic ring) in the molecule could be essential for its Na V -channel blocking activity [7]. However, as the 4 position of the structural moiety in safinamide (α-aminoamide) does not affect the slow inactivation of Na V channels, the interaction with Phe 80 appears to be unimportant. The detailed structure of this Na V channel, which is a particularly good exemplar for hNa V pharmacology, was shown in an earlier study [45]. These results indicate that LCS can favorably interact with the amino-acid residues of the Na V channel with an estimated binding affinity of −5.0. Kcal/mol, which is adjacent to the transmembrane region (i.e., position: 123-148) or membrane segment (i.e., position: 79-95) of the channel. Consequently, when LCS reaches the Na V channels on the membrane, the interactions may raise the structural or steric constraints, thereby resulting in a substantial decrease in Na V -channel openings. Consequently, LCS could reduce the subthreshold potential, enhance the accumulative inhibition during repetitive depolarization, and prolong the recovery from inactivation after repetitive depolarization in excitable cells.
The ionic composition of normal Tyrode's solution buffered by HEPES was as follows (in mM): NaCl 136.5, CaCl2 1.8, KCl 5.4, MgCl2 0.53, glucose 5.5, and HEPES-NaOH buffer (pH 7.4). During the measurements recoding K + currents, a patch electrode was filled with a solution (in mM): K-aspartate 130, KCl 20, MgCl2 1, KH2PO4 1, Na2ATP 0.1, Na2GTP 0.1, EGTA 0.1, and HEPES-KOH buffer (pH 7.2). To measure different patterns of voltage-gated Na + current, we substituted K + ions in the internal solution for equimolar Cs + ions, and the pH value in the solution was adjusted to 7.2 by adding CsOH. The pipette solution and culture media presently used were filtered with an Acrodisc ® syringe filter Note that the red arcs with spokes that radiate toward the ligand (i.e., LCS, in the center) represent the hydrophobic contacts.
In summary, LCS could suppress I Na in a time-, concentration-, and frequencydependent manner and modify the magnitude, gating properties, use-dependence, and Hys (V) behaviors of I Na , leading to inhibiting I Na(T) , I Na(L) , I Na(P), and I Na(W) . Consequently, LCS could reduce the subthreshold potential, enhance the accumulative inhibition during repetitive depolarization, and prolong the recovery from inactivation after repetitive depolarization in excitable cells.

Cell Preparations
Both the mouse neuroblastoma cell line, Neuro-2a (N2a, BCRC-60026), and the pituitary adenomatous cell line, GH 3 (BCRC-60015), were acquired from the Bioresource Collection and Research Center (Hsinchu, Taiwan). GH 3 cells were in Ham's F medium supplemented with 2.5% (v/v) fetal calf serum, 15% (v/v) horse serum, and 2 mM L-glutamine, while Neuro-2a cells were in Dulbecco's modified Eagle's medium with 10% (v/v) fetal bovine serum. These cells were maintained at 5% CO 2 in a 37°C water-jacketed incubator. Growth medium was replaced twice a week, and cells were split into subcultures once a week. Subcultures were made with trypsinization (0.025% trypsin solution (HyClone TM ) containing 0.01% sodium N, N-diethyldithiocarbamate and EDTA). Electrophysiological measurements were undertaken five or six days after cells were cultured up to 60-80% confluence [30].

Electrophysiological Measurements
During the few hours before the measurements, we harvested Neuro-2a or GH 3 cells with 1% trypsin-EDTA solution, and a few drops of cell suspension were rapidly placed into a custom-built recording chamber fixed on the stage of an inverted DM-IL microscope (Leica; Major Instruments, Tainan, Taiwan). We then suspended cells at room temperature (20-25 • C) in normal Tyrode's solution until cells attached to the chamber's bottom before the recordings were made. The pipettes used were fabricated from Kimax-51 glass tubing with an 1.5-1.8 mm outer diameter (#34500; Kimble, Dogger, New Taipei City, Taiwan) with a vertical two-stage puller (PP-83; Narishige, Taiwan Instrument, Tainan, Taiwan). When filled with different internal solutions, the electrodes presently used for measurements had a tip resistance of 3-5 MΩ. Ionic currents were examined in the whole-cell configuration of a modified patch-clamp technique with the use of either an Axoclamp-2B (Molecular Devices, Sunnyvale, CA, USA) or an RK-400 amplifier (Bio-Logic, Claix, France), as described elsewhere [28]. GΩ-seals were achieved in an all-or-nothing fashion and resulted in a dramatic improvement in the signal-to-noise ratio. The liquid junction potentials, which occur when the compositions in bath solution and those of the pipette internal solution are different, became zeroed shortly before GΩ-seal formation was achieved, and the whole-cell data were then corrected as previously desribed [30].

Data Recordings
The signals were monitored and digitally captured and stored online at 10 kHz or more in an ASUS ExpertBook laptop computer (Yuan-Dai, Tainan, Taiwan). For efficient analogto-digital (A/D) and digital-to-analog (D/A) conversion to proceed, a Digidata ® -1440A digitizer connected with a laptop computer via a USB 2.0 port was delicately operated by pClamp 10.6 software run under Microsoft Windows 7 (Redmond, WA, USA). Amplified current signals were low-pass-filtered at 2 kHz with an FL-4 four-pole Bessel filter (Dagan, Minneapolis, MN, USA). The voltage-clamp protocols comprising various rectangular and ramp waveforms were specifically designed and were thereafter imposed on the tested cells through D/A conversion. As pulse-train stimulation was needed, we used a dual output pulse stimulator (Astro-Med Grass S88X; Grass, West Warwick, RI, USA).

Data Analyses
To assess the dose-response curve of LCS-mediated inhibition on the peak (transient, I Na(T) ) and sustained (late, I Na(L) ) components of depolarization-activated I Na present in GH 3 or Neuro-2a cells, I Na 's were evoked by a 30 msec depolarizing pulse to −10 mV from a holding potential of −100 mV (indicated in the top part of Figures 1A and 2A), and current amplitudes obtained with or without the exposure to different LCS concentrations (3 µM-1 mM) were measured at the beginning (I Na(T) ) and end (I Na(L) ) of the depolarizing pulses. The concentration needed to suppress 50% of the current amplitude (i.e., IC 50 ) was appropriately determined according to the three-parameter logistic model (i.e., a modified form of the sigmoidal Hill equation) by use of goodness-of-fit assessments: In this equation, [LCS] = the LCS concentration; n H = the Hill coefficient; IC 50 = the concentration required for a 50% inhibition. Maximal inhibition (i.e., 1 − a) was also approximated in this formula.
By a two-step voltage protocol with varying interpulse intervals in a geometric progression (common ratio = 2), the recovery time course of I Na(T) from the block activated in response to the 1 sec conditioning pulse train was constructed, and the results acquired with or without the addition of LCS to GH 3 or Neuro-2a cells were thereafter drawn by plotting the relative I Na(T) amplitude (normalized with respect to the steady-state amplitude activated at 0.1 Hz). A basic assumption of the analyses is that the recovery time course of the current established under these experimental conditions can be reliably described by an exponential function [34]. Because the recovery time course in GH 3 cells exhibits a rapid rising recovery phase followed by a late slow phase, there should be at least two underlying exponential terms. Accordingly, the data points showing a recovery time course with or without the addition of LCS were fitted to the exponential function with the biexponential process, i.e., where y is the relative amplitude of I Na at time t, A or B is the relative amplitude of each exponential component, and τ fast or τ slow is the fast or slow time constant in the recovery of I Na block, respectively.

Curve-Fitting Approximations and Statistical Analyses
Linear or nonlinear curve fitting to experimental datasets in this work was made with the interactive least-squares procedure by using various tools, such as Microsoft Excel ®embedded "Solver" (Microsoft, Redmond, WA, USA) and the OriginPro ® 2021 program (OriginLab ® ; Scientific Formosa, Kaohsiung, Taiwan). The averaged results are presented as the mean ± standard error of the mean (SEM) with the sizes of independent samples (n) indicating cell numbers from which data were taken. The Student's t-test (paired or unpaired) between the two different groups was applied. When differences among more than two groups were encountered, we performed either analysis of variance (ANOVA)-1 or ANOVA-2 with or without repeated measures, followed by the post hoc Fisher's least significant difference test. Statistical significance (indicated with *, **, or + in the figures) was determined at a p value of <0.05.

Data Availability Statement:
The original data are available upon reasonable request to the corresponding author.