The Anti-Epileptic Drugs Lamotrigine and Valproic Acid Reduce the Cardiac Sodium Current

Anti-epileptic drugs (AEDs) are associated with increased risk of sudden cardiac death. To establish whether gabapentin, lamotrigine, levetiracetam, pregabalin, and valproic acid reduce the Nav1.5 current, we conducted whole-cell patch-clamp studies to study the effects of the five AEDs on currents of human cardiac Nav1.5 channels stably expressed in HEK293 cells, and on action potential (AP) properties of freshly isolated rabbit ventricular cardiomyocytes. Lamotrigine and valproic acid exhibited inhibitory effects on the Nav1.5 current in a concentration-dependent manner with an IC50 of 142 ± 36 and 2022 ± 25 µM for lamotrigine and valproic acid, respectively. In addition, these drugs caused a hyperpolarizing shift of steady-state inactivation and a delay in recovery from inactivation. The changes on the Nav1.5 properties were reflected by a reduction in AP upstroke velocity (43.0 ± 6.8% (lamotrigine) and 23.7 ± 10.6% (valproic acid) at 1 Hz) and AP amplitude; in contrast, AP duration was not changed. Gabapentin, levetiracetam, and pregabalin had no effect on the Nav1.5 current. Lamotrigine and valproic acid reduce the Nav1.5 current density and affect its gating properties, resulting in a decrease of the AP upstroke velocity. Gabapentin, levetiracetam, and pregabalin have no effects on the Nav1.5 current.


Introduction
Anti-epileptic drugs (AEDs) are a mainstay of epilepsy treatment and are also prescribed for behavioral problems and psychiatric disorders [1]. These drugs exert their anti-convulsant actions through various mechanisms, including the blocking of neuronal sodium channels [2]. Of clinical relevance, we and others found that AED use is associated with an increased risk of sudden cardiac death (SCD) due to cardiac arrhythmia [3,4]. The often-used drug carbamazepine is an example of such a drug [4]. We recently demonstrated that carbamazepine blocks the cardiac sodium channel Na v 1.5 [4].
SCN5A encoded Na v 1.5 is the most prominent sodium channel in the heart [5,6]. It is responsible for the rapid upstroke of the cardiac action potential (AP) and regulates impulse propagation in the heart. Na v 1.5 block is a plausible mechanism contributing to the elevated SCD risk of cardiac arrhythmia and SCD of carbamazepine [7] because drugs that block the Na v 1.5 increase SCD risk [8,9]. This insight was first gained in the landmark Cardiac Arrhythmia Suppression Trial in which patients randomized to the class 1c cardiac antiarrhythmic drugs (potent Na v 1.5 blockers) flecainide or encainide suffered excess mortality rates due to SCD compared to placebo-treated patients [10]. On the other hand, Na v 1.5 block by AEDs is plausible given that the Na v 1.5 shares great homology with neuronal sodium channel isoforms [11,12].
In view of these observations, the aim of our study was to establish whether other AEDs than carbamazepine also block the Na v 1.5. Here, we studied the five AEDs which,

Patch-Clamp Recording
We applied the whole-cell configuration of the patch-clamp technique using an Axopatch 200B amplifier (Molecular Devices, San Jose, CA, USA). Borosilicate glass patch pipettes (GC100F-10; Harvard Apparatus, Cambourne, UK) had a resistance of 2-3 MΩ after filling with the pipette solution. All signals were low-pass filtered (5 kHz) and digitized at 33 kHz. Series resistance was compensated by ≥80%, and AP potentials were corrected for the calculated liquid junction potential [16] by an offline 15 mV shift in potential toward more negative values. In order to obtain steady-state conditions, signals were recorded after a stable stimulation period, i.e., under baseline conditions, and 5-7 min after the application of AEDs.

Sodium Current Measurements
The Na v 1.5 current was measured in single HEK cells using a pipette solution containing (in mM) 10 NaF, 10 CsCl, 110 CsF, 11 EGTA, 1.0 CaCl 2 , 1.0 MgCl 2 , 2.0 Na 2 ATP, 10 HEPES (pH adjusted to 7.2 with CsOH), and a bath solution containing (in mM) 20 NaCl, 120 CsCl, 1.8 CaCl 2 , 1.0 MgCl 2 , 5.5 glucose, and 5.0 HEPES (pH adjusted to 7.4 with CsOH) [17]. The Na v 1.5 current was measured at room temperature in response to depolarizing voltage steps from a holding potential of −120 mV (cycle length of 5 s). I Na was defined as the difference between peak and steady-state current. The dose-response curves were fitted by the Hill equation: Y = 1/[(1 + (IC 50 /X) n )], where Y is the current normalized to baseline condition, IC 50 is the dose required for 50% current block, and n is the Hill coefficient. The Na v 1.5 (in)activation current was measured with a double-pulse protocol, as detailed in Section 3.2, below. During the first depolarizing pulses (P1), I Na activates, and the currents analyzed here are used to determine current-voltage (I-V) relationships and the voltage dependency of activation. For the latter, the I-V relationships were corrected for driving force and normalized to maximum peak current. The second pulse (P2) is used to determine the voltage dependency of inactivation and currents were normalized to the largest I Na . Voltage dependence of activation and inactivation curves were fitted with the Boltzmann function: y = 1/{1 + exp [(V-V 1/2 )/k]}, where V 1/2 is the midpoint of channel (in)activation, and k is the slope factor of the (in)activation curve. The use-dependent block was determined by application of 30 activating pulses from −120 to −20 mV at a frequency of 4 Hz, as detailed in Section 3.2. The Na v 1.5 currents were normalized to the current of the first pulse. The length of recovery from inactivation was measured with a double-pulse protocol with two depolarizing steps (P1 and P2) from −120 to −20 mV and a variable interpulse interval (see Section 3.2). Currents measured during P2 were normalized to currents measured during P1. Recovery from inactivation was fitted by a double-exponential function: , where τ f and τ s are the fast and slow time constants of recovery from inactivation, respectively, and A f and A s are the fractions of fast and slow recovery from inactivation, respectively.

Preparation of Antiepileptic Drugs
All AEDs (purity ≥98%) were purchased from Sigma-Aldrich. Lamotrigine and valproic acid were dissolved in dimethyl sulfoxide (DMSO) to produce a 1 M stock solution. The stock solution was stored at −20 • C and freshly diluted in the bath solution to the desired concentration just before use. The concentration of DMSO in the final solution was less than 0.33% and this does not affect cardiac ion channels [18,19]. Pregabalin, gabapentin, and levetiracetam were freshly dissolved in the bath solution to the desired concentration just before use. The Na v 1.5 current was measured at baseline conditions and after the wash-in of AEDs at concentrations of 1, 10, 30, 100, 300, or 1000 µM; these concentration ranges surrounded the therapeutic concentrations of the AEDs, as indicated by the yellow parts in Figure 1B-F [20][21][22][23][24][25][26][27].

Statistical Analysis
Values are presented as mean ± SEM. Curve fitting and statistics was performed using Prim8 GraphPad (GraphPad Software, San Diego, CA, USA). One-Way ANOVA or Two-Way ANOVA was used to assess the statistical significance of the differences among multiple groups. One-way repeated measures (RM) ANOVA was used, followed by pairwise comparison using the Holm-Sidak's multiple comparisons test or One-way RM ANOVA on Ranks (Friedman test), followed by Dunn's multiple comparison test for post hoc analyses when data were not normally distributed. Differences between the two groups were tested using paired Student's t-tests or Two-Way RM ANOVA, followed by pairwise comparison using the Holm-Sidak's multiple comparisons test. Details on normalization are given in Methods or in the figure legends. p < 0.05 was considered to be statistically significant.
polarizing pulses from −120 to −40 mV ( Figure 1A) to HEK293 cell with stable Nav1.5 expression and tested various drug concentrations including the therapeutic concentrations, which are indicated as yellow parts in Figure 1B-F. We found that lamotrigine and valproic acid reduced the Nav1.5 current density in a concentration-dependent manner ( Figure 1E,F). The average IC50 of lamotrigine and valproic acid were 142 ± 36 µM (n = 6-7) and 2022 ± 25 µM (n = 5), respectively. The tested concentrations of gabapentin, levetiracetam, and pregabalin had no effect on the Nav1.5 current density ( Figure 1B-D).

Effects of Lamotrigine (100 μM) on Gating Properties of Nav1.5 Channels
Second, we tested if the decrease in Nav1.5 in response of 100 µM lamotrigine (close to IC50) is accompanied by changes in gating properties in HEK293 cells. Figure 2A shows typical Nav1.5 currents under baseline conditions and in the presence of 100 µM lamotrigine measured over a wide range of depolarizing voltages (for protocol, see Figure 2A, inset). The average current-voltage (I-V) relationships are shown in Figure 2B. Lamotrigine induced a similar amount of block over the whole voltage range measured and a hyperpolarizing shift in both activation and inactivation ( Figure 2C). The V1/2 of channel activation occurred at −52.7 ± 1.7 mV in the absence, and −56.7 ± 1.5 mV in the presence, of lamotrigine (n = 9, p < 0.05). The V1/2 of channel inactivation were −92.9 ± 1.5 mV and −99.5 ± 2.6 mV, respectively (n = 9, p < 0.05, Table 1). The slope of the inactivation curve was significantly changed after the application of lamotrigine from 6.0 ± 0.4 to 7.7 ± 0.6 mV (n

Inhibition of the Na v 1.5 Current by Lamotrigine and Valproic Acid in a Concentration-Dependent Manner
First, we measured the effects of gabapentin, levetiracetam, pregabalin, lamotrigine, and valproic acid on the Na v 1.5 current density in HEK293 cells. We applied 100 ms depolarizing pulses from −120 to −40 mV ( Figure 1A) to HEK293 cell with stable Na v 1.5 expression and tested various drug concentrations including the therapeutic concentrations, which are indicated as yellow parts in Figure 1B-F. We found that lamotrigine and valproic acid reduced the Na v 1.5 current density in a concentration-dependent manner ( Figure 1E,F). The average IC 50 of lamotrigine and valproic acid were 142 ± 36 µM (n = 6-7) and 2022 ± 25 µM (n = 5), respectively. The tested concentrations of gabapentin, levetiracetam, and pregabalin had no effect on the Na v 1.5 current density ( Figure 1B-D).

Effects of Lamotrigine (100 µM) on Gating Properties of Na v 1.5 Channels
Second, we tested if the decrease in Na v 1.5 in response of 100 µM lamotrigine (close to IC 50 ) is accompanied by changes in gating properties in HEK293 cells. Figure 2A shows typical Na v 1.5 currents under baseline conditions and in the presence of 100 µM lamotrigine measured over a wide range of depolarizing voltages (for protocol, see Figure 2A, inset). The average current-voltage (I-V) relationships are shown in Figure 2B. Lamotrigine induced a similar amount of block over the whole voltage range measured and a hyperpolarizing shift in both activation and inactivation ( Figure 2C). The V 1/2 of channel activation occurred at −52.7 ± 1.7 mV in the absence, and −56.7 ± 1.5 mV in the presence, of lamotrigine (n = 9, p < 0.05). The V 1/2 of channel inactivation were −92.9 ± 1.5 mV and −99.5 ± 2.6 mV, respectively (n = 9, p < 0.05, Table 1). The slope of the inactivation curve was significantly changed after the application of lamotrigine from 6.0 ± 0.4 to 7.7 ± 0.6 mV (n = 9, p < 0.05, Table 1). To study the rate-dependent effects of lamotrigine, we applied a double-pulse protocol with an interpulse interval of 50 ms ( Figure 2D) and found that the reduction in the Na v 1.5 current density at rising pulse numbers increased more in the presence of lamotrigine ( Figure 2E). Consistent with this observation, this was attended by delayed recovery from steady-state inactivation ( Figure 2E, Table 1) with τ f and τ s significantly changed from 11.2 ± 1.7 to 17.1 ± 3.6 ms, and from 134.8 ± 21.5 to 657.7 ± 125.1 ms, respectively (n = 7, p < 0.05, Table 1).
Biomedicines 2023, 11, x FOR PEER REVIEW 5 of 13 = 9, p < 0.05, Table 1). To study the rate-dependent effects of lamotrigine, we applied a double-pulse protocol with an interpulse interval of 50 ms ( Figure 2D) and found that the reduction in the Nav1.5 current density at rising pulse numbers increased more in the presence of lamotrigine ( Figure 2E). Consistent with this observation, this was attended by delayed recovery from steady-state inactivation ( Figure 2E, Table 1) with τf and τs significantly changed from 11.2 ± 1.7 to 17.1 ± 3.6 ms, and from 134.8 ± 21.5 to 657.7 ± 125.1 ms, respectively (n = 7, p < 0.05, Table 1).

Effects of Valproic Acid (2000 µM) on Gating Properties of Na v 1.5 Channels
Third, we studied the effects of 2000 µM valproic acid (close to IC 50 ) on the Na v 1.5 current in HEK293 cells, similar to that obtained for lamotrigine. Valproic acid reduced the Na v 1.5 current density ( Figure 3A,B). Figure 3C showed that valproic acid did not induce a statistically significant shift in voltage dependency of activation (V 1/2 from −50.6 ± 1.5 to −50.5 ± 2.7 mV) (n = 15, p = NS) or slope of activation (Table 1). However, valproic acid induced a significant shift in steady-state inactivation (V 1/2 from −92.7 ± 1.4 to −99.0 ± 1.9 mV) (n = 15, p < 0.05, Figure 3C, Table 1). And the slope of the inactivation curve was significantly changed after the application of valproic acid from 6.6 ± 0.2 to 5.9 ± 0.3 mV (n = 9, p < 0.05, Table 1). Valproic acid had modest effects on the rate of recovery from inactivation ( Figure 3D,E, Table 1), affecting τ f mildly (from 11.3 ± 2.1 to 17.5 ± 3.6 ms, p < 0.05), but not τ s (from 165 ± 18.0 to 200.8 ± 34.5 ms, n = 7, p = NS). Accordingly, valproic acid significantly affected the rate of reduction of the Na v 1.5 current density at repetitive pacing with an interpulse interval of 50 ms (n = 10, p < 0.05, Figure 3D).

Effects of Lamotrigine and Valproic acid on Action Potentials Properties
In a final patch clamp experiment, we studied the effects of lamotrigine (100 µM) and valproic acid (3000 µM) on APs elicited in rabbit ventricular cardiomyocytes in order to verify our findings regarding the effects of these AEDs on the Na v 1.5 current in HEK293 cells, and to investigate possible drug effects on other ion channels. Figure 4A showed typical APs at 1 Hz under baseline conditions and in the presence of 100 µM lamotrigine; average AP parameters are summarized in in Figure 4B-D. Lamotrigine caused statistically significant decreases in dV/dt max and APA in a frequency-dependent manner ( Figure 4C,E). dV/dt max was significantly decreased at all rates, e.g., by 43.0 ± 6.8% (from 309.7 ± 24.3 to 176.6 ± 19.0 V/s, n = 7, p < 0.05) at 1 Hz, and by 70.1 ± 8.2% (from 281.2 ± 29.0 to 84.0 ± 11.5 V/s, n = 7, p < 0.05) at 3 Hz (n = 7, p < 0.05) ( Figure 4C). Similarly, APA decreased by 2.1 ± 0.5% (from 126.0 ± 1.3 to 123.4 ± 1.6 mV, n = 7, p < 0.05) at 1 Hz, and by 7.8 ± 1.1% (from 122.3 ± 1.9 to 112.8 ± 2.0 mV, n = 7, p < 0.05) at 3 Hz ( Figure 4E). The reduction in dV/dt max was larger at higher stimulation frequencies (n = 6-7, p < 0.05), consistent with the reduced rate of recovery from inactivation of Na v 1.5 ( Figure 2D,E). Lamotrigine did not change RMP or APD 90 . The absence of effects on RMP and APD 90 indicates that other ionic currents were virtually unaffected.

Discussion
Our main findings were: (1) lamotrigine and valproic acid inhibited the Nav1.5 current in a dose-dependent manner, while gabapentin, levetiracetam, and pregabalin had no effect at the doses tested; (2) lamotrigine and valproic shifted the voltage dependency of inactivation and slowed the recovery from inactivation; (3) lamotrigine and valproic acid reduced dV/dtmax and APA in rabbit cardiomyocytes with a larger amount of reduction at fast pacing rates; and (4) lamotrigine and valproic acid did not impact other AP properties, except for modest reduction of RMP by valproic acid.
Our observations are largely consistent with reports on the effects of lamotrigine and valproic acid on neuronal sodium channels, reflecting the high homology between cardiac

Discussion
Our main findings were: (1) lamotrigine and valproic acid inhibited the Na v 1.5 current in a dose-dependent manner, while gabapentin, levetiracetam, and pregabalin had no effect at the doses tested; (2) lamotrigine and valproic shifted the voltage dependency of inactivation and slowed the recovery from inactivation; (3) lamotrigine and valproic acid reduced dV/dt max and APA in rabbit cardiomyocytes with a larger amount of reduction at fast pacing rates; and (4) lamotrigine and valproic acid did not impact other AP properties, except for modest reduction of RMP by valproic acid.
Our observations are largely consistent with reports on the effects of lamotrigine and valproic acid on neuronal sodium channels, reflecting the high homology between cardiac and neuronal sodium channels [11,12]. We observed a concentration-dependent blockade of the Na v 1.5 current by lamotrigine, and similar effects were reported in both cardiac sodium channels expressed in HEK293 cells [28] and neuronal sodium channels expressed in CHO cells [29]. Meanwhile, lamotrigine reduced the density of voltage-gated sodium current in rat cerebellar granule cells and induced a hyperpolarizing shift in the voltage dependency of inactivation [30], which is consistent with our findings. Our observed lamotrigine-induced delay in recovery from inactivation is also mirrored by similar effects on cardiac sodium channels and neuronal sodium channels [29]. Previous reports on the effects of valproic acid on neuronal sodium channels were also consistent with our findings. For instance, valproic acid reduced sodium current density in the nodal membrane of peripheral nerve fibers of Xenopus laevis by 54% at a dose (2.4 mM) which is very close to our IC 50 value of the Na v 1.5 current block (2.0 mM) [31]. Moreover, valproic acid (2 mM) shifted the voltage dependence of inactivation to more hyperpolarized potentials in cortical neurons [32]. Furthermore, valproic acid (1 mM) reduced the sodium current density and slowed the recovery from inactivation in rat hippocampal neurons [33]. While some studies showed that valproic acid had no effect on fast neuronal sodium current, the concentrations used in these studies were mostly lower than therapeutic concentrations [34,35].
The antagonizing effects of lamotrigine and valproic acid on the Na v 1.5 current, dV/dt max , and APA are here reported for concentrations (100 µM lamotrigine, 3000 µM valproic acid) that exceed the upper limit of the therapeutic range of lamotrigine (59 µM) and valproic acid (867 µM) [27], but only by a factor of 2-3 (while a reduction in sodium current density by lamotrigine already started at concentration within its therapeutic range). Thus, these effects may be clinically relevant because these somewhat elevated concentrations may occur in clinical practice, e.g., at (mild) overdoses. A drug-induced blockade of the human Na v 1.5 current could reduce cardiac excitability (as evidenced by a widened QRS complex of the ECG) and increase mortality risk [36]. Accordingly, a review of case reports of lamotrigine overdose found that lamotrigine overdose may be associated with ECG changes (QRS widening) and cardiac arrhythmias (wide complex tachycardia, complete heart block) which are consistent with a reduced Na v 1.5 current [37]. Another way in which our findings may have relevance in routine clinical care is that specific subgroups of patients may have elevated vulnerability to the Na v 1.5 blocking effects of lamotrigine or valproic acid [37]. In these individuals, even plasma concentrations within the therapeutic ranges may cause clinically significant effects on cardiac electrophysiology, leading to cardiac arrhythmia and SCD. Enhanced vulnerability may stem from acquired comorbidities and/or from inherited susceptibility. The concept that acquired comorbidities may permit the occurrence of fatal cardiac arrhythmia and SCD upon use of Na v 1.5 current blocking drugs was discovered in the Cardiac Arrhythmia Suppression Trial, in which patients randomized to the Na v 1.5 current blockers flecainide or encainide suffered excess SCD rates compared to placebo-treated patients [10]. In a meta-analysis, it was discovered that this risk occurred in patients who have increased vulnerability to this adverse drug effect due to comorbidities associated with reduced Na v 1.5 function, such as myocardial ischemia/infarction and heart failure [38]. This insight has prompted the recommendation in authoritative clinical guidelines to screen patients for the presence of these comorbidities and withhold these drugs from patients who have them [39]. On the other hand, inherited susceptibility may stem from carrying variants in genes that encode subunits of the Na v 1.5 channel, in particular, variants in SCN5A, the gene that encodes its α-subunit. Loss-offunction mutations in this gene underlie the Brugada syndrome [40] and cardiac conduction disease [41], inherited cardiac arrhythmia syndromes associated with elevated SCD risk. Accordingly, mutations in SCN5A were found in a series of patients with epilepsy who suffered unexplained and autopsy-negative SCD (sudden unexplained death in epilepsy), and these mutant genes, when heterologously expressed in CHO-K1 cells, produced altered Na v 1.5 channel functional properties [42]. Of interest, one of these patients used lamotrigine at the time of SCD [43]. In view of these observations, when prescription of lamotrigine or valproic acid is considered, it could be prudent to first investigate whether the patient has any acquired or inherited condition that would increase the vulnerability to excessive Na v 1.5 channel block which could culminate in SCD. This strategy would mirror the strategies in routine cardiology practice to screen patients on these conditions before cardiac drugs that block Na v 1.5 channels (e.g., flecainide or other class I antiarrhythmic drugs) are considered [39], and to withhold these drugs from patients with Brugada syndrome or those who carry SCN5A mutations [44]. Screening for inherited vulnerability may be facilitated by the rapidly increasing availability of widespread DNA testing. Before full implementation of DNA testing, inquiring about the presence of familial SCD during simple history taking may already be informative because of the familial nature of SCD [45]. Conversely, the use of lamotrigine and valproic acid may not confer increased risk of cardiac arrhythmias and SCD in patients without enhanced vulnerability.
In any case, when we consider our observations in view of previous pharmacoepidemiologic studies into the associations between AED use and the risk of SCD [3,46,47], we conclude that blocking effects on Na v 1.5 channels do not fully account for the increased risk of SCD associated with epilepsy [48]. Our conclusion derives from the fact that our observed effects on Na v 1.5 currents were only partly consistent with the findings in these pharmacoepidemiologic studies. For instance, while we report that lamotrigine blocks Na v 1.5 currents, this drug was not associated with increased SCD risk in a study of Eroglu et al. [46]. While the comparator in that study was valproic acid, the SCD risk of lamotrigine actually tended to be smaller (but statistical significance was not reached). In the studies of Bardai et al. [46] and Hookana et al. [47], the numbers were too small to draw conclusions on possible effects of lamotrigine on SCD risk. A recent review also found that there is not sufficient evidence to support or refute the notion that lamotrigine is associated with increased SCD risk [49]. For valproic acid, increased SCD risk was reported by Hookana et al., but not by Bardai et al. (a possible effect on SCD risk was not studied by Eroglu et al., who used valproic acid as comparator). Conversely, while we found that gabapentin, levetiracetam, and pregabalin had no effects on Na v 1.5 currents, Eroglu et al. found that pregabalin conferred higher SCD risk than valproic acid, while gabapentin and levetiracetam also tended to have higher SCD risk (but it was not statistically significantly). Similarly, Bardai et al. reported higher SCD risk for gabapentin (for levetiracetam, the statistical power was too small to draw meaningful conclusions). Still, Hookana et al. reported no elevated SCD risk for gabapentin and pregabalin. We conclude from these comparisons that other mechanisms beyond the Na v 1.5 block also contribute to the elevation in SCD risk in epilepsy, as previously reported [50].

Conclusions
Lamotrigine and valproic acid reduce the Na v 1.5 current by reducing its current density and changing its gating properties; these effects are reflected in changes in AP properties. Gabapentin, levetiracetam, and pregabalin have no effects on the Na v 1.5 current.