The Cardiac Sodium Channel Safety Profile of Cenobamate at Therapeutic Concentrations: In Vitro Analyses

ABSTRACT

The US Food and Drug Administration (FDA) issued a Drug Safety Communication for the antiseizure drug lamotrigine (Lamictal), warning of the risk of QRS prolongation and cardiac arrhythmias. The FDA also requested that other antiseizure medications with a similar mechanism of action (e.g., block of neuronal sodium channels), including cenobamate, be evaluated for their risk as a cardiac sodium channel blocker. The purpose of this in vitro study was to determine the cardiac sodium channel blocking effects of cenobamate and its sodium channel classification. Follow‐up experiments were also performed with cenobamate at its highest free therapeutic concentration (66 μM). Experiments were performed using the manual patch clamp technique at physiologic temperatures. Despite its classification as a Class Ib and Id antiarrhythmic, cenobamate was found to have little effect on the cardiac sodium channel when tested in vitro at therapeutic levels. These findings provide further clarity and support for the clinical safety of cenobamate in patients without cardiovascular risk factors.

Study Highlights.

• What is the current knowledge on the topic?

  • ○
    In 2021, an FDA review of in vitro study data suggested an increased risk of arrhythmias and other serious adverse events in patients with heart disease who received the antiseizure medication lamotrigine. A Drug Safety Communication was issued for lamotrigine and the prescribing information was amended to add conclusions from in vitro testing that showed the drug could slow ventricular conduction, widening the QRS and inducing proarrhythmia, potentially leading to death.
• What question did this study address?

  • ○
    The FDA requested an evaluation of other medications with a similar mechanism of action (sodium channel blockade) in order to determine if any had similar cardiac effects, which included cenobamate. This study was conducted in vitro to determine the effect of cenobamate on the cardiac sodium channel.
• What does this study add to our knowledge?

  • ○
    The preclinical data indicate that maximum therapeutic concentrations of cenobamate had virtually no effect on the cardiac sodium channel.
• How might this change clinical pharmacology or translational science?

  • ○
    These results suggest that there is no preclinical evidence to indicate a potential risk of QRS prolongation or proarrhythmic effect with cenobamate at therapeutic doses. These data further support the well‐characterized safety profile of cenobamate in patients without cardiovascular risk factors and provide additional clarity.

1. Introduction

In 2021, the US Food and Drug Administration (FDA) issued a Drug Safety Communication for the antiseizure drug lamotrigine (Lamictal) [1, 2]. The lamotrigine prescribing information was revised to include the following language under 5.0 WARNINGS AND PRECAUTIONS, 5.4 Cardiac Rhythm and Conduction Abnormalities: “In vitro testing showed that LAMICTAL exhibits Class IB antiarrhythmic activity at therapeutically relevant concentrations…. Based on these in vitro findings, LAMICTAL could slow ventricular conduction (widen QRS) and induce proarrhythmia, which can lead to sudden death, in patients with clinically important structural or functional heart disease” [3].

Subsequently, the FDA requested that other antiseizure medications (ASMs) with a similar mechanism of action (MOA; eg, block of neuronal sodium channels), including cenobamate, be evaluated for their risk as a cardiac sodium channel blocker [1]. The study design requested by the FDA included defining the half‐maximal inhibitory concentration (IC₅₀) for the cardiac voltage‐gated sodium (Na_v) channel 1.5 (which provides an indication of blocking potency), the rate of drug unblocking recovery from block using the IC₅₀ concentration of drug, and a comparison of these parameters to known blockers of the cardiac sodium channel (e.g., antiarrhythmic drugs) [4]. The latter requirement was so that the antiarrhythmic class (Ia, Ib, Ic) of these ASMs could be assigned. The amount of cardiac sodium channel block is a critical determinant of the degree of QRS prolongation and the potential for proarrhythmic events: with an increased block of Na_v1.5, a greater possibility exists of QRS prolongation and a higher risk of proarrhythmic events [5, 6, 7]. In addition to the amount of block, the rate at which the blocking drug exits the channel (e.g., recovery from block) is important in defining the proarrhythmic potential of a sodium channel blocker [8, 9]. Drugs that unbind slowly exhibit an accumulation of block at elevated heart rates. The antiarrhythmic drug flecainide is an example of a potent cardiac sodium channel blocker and a drug that has slow recovery from block kinetics [10]. Flecainide, a Class Ic antiarrhythmic, is associated with QRS prolongation, ventricular arrhythmias, and an increase in mortality [11, 12]. In contrast, the Class Ib antiarrhythmic mexiletine is a weak Na_v1.5 blocker, has fast recovery from block kinetics, and is rarely associated with proarrhythmic events [13, 14, 15].

Cenobamate, an oral ASM approved for adults with focal seizures, exhibits a dual MoA that preferentially inhibits the persistent sodium current via blockade of voltage‐dependent sodium channels and functions as a positive allosteric modulator of γ‐aminobutyric acid type A (GABA_A) receptors at nonbenzodiazepine sites [16, 17, 18, 19]. A previous placebo‐controlled thorough QT study that assessed the electrocardiographic effects of cenobamate at therapeutic (200 mg/day) and supratherapeutic (500 mg/day) doses versus placebo in healthy adults showed no clinically relevant effect on the QRS interval (mean values within ±1.0 ms of placebo) [20]. However, a shortening of the QTc interval was observed at both therapeutic and supratherapeutic doses [20].

The purpose of this study was to characterize the effects of cenobamate on Na_v1.5 using the FDA‐requested study design as well as the more clinically relevant assessment of the Na_v1.5 effects of cenobamate at therapeutic plasma concentrations.

2. Methods

All chemicals and reagents were purchased from Sigma‐Aldrich Chemical Company (St. Louis, MO and Natick, MA) except where noted.

2.1. Na_v1.1–Na_v1.8 Assays

Sodium currents were measured in Chinese hamster ovary (CHO) cells (strain source, ATCC, Manassas, VA; substrain source, ChanTest Corp., Cleveland, OH) stably expressing human Na_v1.1, Na_v1.2, Na_v1.3, Na_v1.4, Na_v1.5, Na_v1.6, Na_v1.7, or Na_v1.8 cDNA. Na_v1.1–Na_v1.3 and Na_v1.6 are typically located in the central nervous system and peripheral nervous system, with Na_v 1.1–Na_v1.2 associated with epilepsy; Na_v1.4 is primarily located in skeletal muscle; Na_v1.5 is primarily located in the heart; and Na_v1.7–Na_v1.8 are located peripherally and associated with neuropathic pain [21, 22]. All sodium channel subtypes were tested to determine the potential selectivity of cenobamate. Automated patch‐clamp recordings were conducted at room temperature using the Ion Works Quattro system (version 2.0.2; Molecular Devices Corporation, Union City, CA). The intracellular solution was composed of 130 mM KCl, 4 mM MgCl₂, 0.5 mM EGTA, and 10 mM HEPES. The extracellular solution was composed of HEPES‐buffered physiological saline (HB‐PS) solution (137 mM NaCl, 4 mM KCl, 1.8 mM CaCl₂, 1 mM MgCl₂, 10 mM HEPES, 10 mM glucose; with a pH adjusted to 7.4 with NaOH plus 0.3% DMSO). Cells were depolarized from a holding potential of −80 to 0 mV (Na_v1.1, Na_v1.2, Na_v1.3, Na_v1.6, and Na_v1.7), −80 to −10 mV (Na_v1.4 and Na_v1.5), or −80 to 20 mV (Na_v1.8) for 14 pulses (20 ms for test pulse 1–10 and 12–14; 500 ms for test pulse 11). A pre‐pulse potential of −120 mV applied for 200 ms was included, and cells were returned to the holding potential of −80 mV (pulses 1–11) or −60 mV (test pulse 12–14) for 80 ms prior to evoking sodium currents. The pulse pattern was repeated before and 5 min after application of test article (cenobamate 0.3–1000 μM, SK Life Science Inc., Paramus, NJ), vehicle control (0.3% DMSO), or positive control (lidocaine 3000 μM).

Tonic block was calculated as follows:

$$\%\text{Block}\left(\text{Tonic}\right)=1-\left(\frac{I_{\mathrm{TP}1,\mathrm{TA}}}{I_{\mathrm{TP}1,\text{Control}}}\right)\times 100\%$$

where I _TP1,Control and I _TP1,TA are the inward peak sodium currents elicited by test pulse 1 in control and in the presence of test article, respectively. The same equation was used for experiments conducted at 10 Hz.

The inactivated state block, defined as the decrease in current amplitude in test pulse 12 due to the conditioning depolarizing pulse (test pulse 11), was calculated as follows:

$$\%\text{Block}\left(\text{Inactivation state}\right)=1-\left(\frac{I_{\mathrm{TP}12,\mathrm{TA}}}{I_{\mathrm{TP}12,\text{Control}}}\right)\times 100\%$$

where I _TP12,Control and I _TP12,TA are the inward peak sodium currents elicited by test pulse 12 in control and in the presence of test article, respectively.

Concentration‐response data for the blocks were fit to an equation of the following form:

$$\%\text{Block}=\%\mathrm{VC}+\left\{\frac{\left(100-\%\mathrm{VC}\right)}{\left[1+{\left(\frac{\left[\text{Test}\right]}{\mathrm{I}{\mathrm{C}}_{50}}\right)}^{\mathrm{N}}\right]}\right\}$$

where [Test] is the concentration of test article, IC₅₀ is the concentration of the test article producing half‐maximal inhibition, N is the Hill coefficient, % VC is the percentage of the current run‐down (the mean current inhibition measured during vehicle control experiments), and % Block is the percentage of ion channel current inhibited at each concentration of test article.

2.2. Cardiac Ion Channel Assays

2.2.1. hNa_v1.5

2.2.1.1. Drug and Solution Preparation

Cenobamate was obtained from SK Life Science Inc. Quinidine, flecainide, mexiletine, and ranolazine were purchased as powders from Sigma‐Aldrich. Ranolazine and mexiletine master stocks were made in de‐ionized H₂O. All other master stocks were made in DMSO. The master stocks were diluted in the external solution to create test concentrations. ATX‐II (Alomone Labs, Jerusalem, Israel) was added to the external solution at a concentration of 150 nM and used to elicit the late Na_v1.5 current. The external solution consisted of the following (in mM): 137 NaCl, 4.0 KCl, 1.0 MgCl₂, 1.8 CaCl₂, 10.0 HEPES, 11.0 dextrose adjusted to a pH of 7.4 with NaOH. The internal (pipette) solution consisted of the following (in mM): 130 KCl, 1.0 MgCl₂, 5.0 HEPES, and 7.0 NaCl adjusted to a pH of 7.2 with KOH.

2.2.1.2. Cell Line

Ready‐to‐use human embryonic kidney 293 (HEK293) cells (Nova Research Laboratories, New Orleans, LA) stably transfected with human Na_v1.5 cDNA were maintained in liquid nitrogen and each experimental day a new vial of cells was thawed and used.

2.2.1.3. Patch Clamp Protocols

Currents were measured using the whole‐cell variant of the patch clamp method [23]. Glass pipettes with resistance of ≈1–2 MΩ were pulled from borosilicate glass by a horizontal puller (Sutter Instruments, Novato, CA). Bath temperature was measured by a thermistor placed near the cell under study and were maintained by a thermoelectric device. An Axopatch 1‐B amplifier (Axon Instruments, Foster City, CA) was used for whole‐cell voltage clamping at 37°C ± 1°C. Creation of voltage clamp pulses and data acquisition was controlled by a computer running pCLAMP software (version 10.7; Axon Instruments). Flow rate for solution delivery was approximately 5–6 mL/min. Voltage protocols are provided for the Na_v1.5 current concentration‐response relationship, peak Na_v1.5 rate‐dependence, and peak Na_v1.5 recovery in the Supporting Information S1.

The peak or late Na_v1.5 concentration‐response relationship for the test article cenobamate (1–1000 μM) was compared with representative drugs of Class Ia (quinidine [3–100 μM]), Class Ib (mexiletine [10–300 μM]), Class Ic (flecainide [1–100 μM]), and Class Id (ranolazine [0.6–20 μM]). Tetrodotoxin (TTX, 30 μM) was applied at the end of some of the experiments to confirm the current was Na_v1.5 channel‐mediated.

2.2.1.4. Na _v 1.5 Recovery at Therapeutic Concentrations of Cenobamate

To characterize the peak Na_v1.5 recovery from block kinetics of cenobamate (SK Life Science Inc.), a free cenobamate concentration of 66 μM was chosen, as this represents a concentration within a clinically relevant range associated with the maximum recommended dose of cenobamate in humans (400 mg/day; 170 μM total) [24]. Whole‐cell patch‐clamp recordings were conducted at 37°C using an Axopatch 1‐B amplifier (Axon Instruments). Recovery was measured with a paired‐pulse protocol. Cells were depolarized from −120 to −15 mV for 500 ms (condition pulse), hyperpolarized to −120 mV in the interpulse interval of varying duration (1–1000 ms), and then depolarized to −15 mV for 20 ms (test pulse). Cells were paced once every 10 s, and experiments began after current was stable for at least 12 consecutive traces. Current was measured as the peak inward current during the test pulse.

Data were fit with a 2‐exponential function as follows:

$$y=\mathrm{A}1\times \exp \left(\frac{x}{\mathrm{t}1}\right)+\mathrm{A}2\times \exp \left(\frac{x}{\mathrm{t}2}\right)+y0$$

2.2.1.5. Data Handling and Statistical Analysis

Individual data and mean ± standard error of the mean (SEM) were reported. Data are presented as percent change of treated current amplitude as compared with the pre‐treated cell baseline current amplitude. This was measured as current amplitude after a steady‐state effect had been reached in the presence of the test article relative to steady‐state current amplitude before the test article was introduced (pre‐treated control). Each cell served as its own control. The test article effects were compared by a pair‐wise Student t test for significance (p < 0.05) using MicroCal Origin software (version 2021b; OriginLab Corp., Northampton, MA). Log‐linear plots were created of the mean percent blockade ± SEM at the concentrations tested. A nonlinear curve‐fitting routine was utilized to fit a three‐parameter Hill equation to the results using MicroCal Origin, version 2021b software. The equation is as follows:

$$y=V_{\max }\frac{x^n}{k^n+x^n}$$

where V _max=100, k, and n are unconstrained variables.

2.2.2. Human Ether‐A‐Go‐Go‐Related Gene (hERG) Channels

hERG recordings were performed with HEK293 cells stably expressing hERG cDNA (obtained from Dr. Craig January, University of Wisconsin Medical School) and were conducted in accordance with Good Laboratory Practice (GLP) regulations. Whole‐cell patch‐clamp recordings were conducted at 34°C–37°C with a Multiclamp 700A amplifier and pCLAMP software (version 9.0; Axon Instruments). The recording patch pipette filled with internal solution (115 mM K‐methanesulfonate, 10 mM KCl, 1 mM MgCl₂, 5 mM EGTA, 10 mM HEPES [VWR Amresco LLC, Solon, OH], 3 mM ATP‐Mg, 0.3 mM GTP‐Na, and 4 mM phosphocreatine [Acros Organics, Fair Lawn, NJ]) had a tip resistance of 1.2–1.5 MΩ. The extracellular solution was composed of the following: 140 mM NaCl, 10 mM HEPES, 4.2 mM KCl, 2 mM CaCl₂, 1 mM MgCl₂, and 10 mM d‐glucose, with a pH adjusted to 7.4 with NaOH. For each measurement, the test article (60–2000 μM cenobamate; SK Life Science Inc.), vehicle (0.3% DMSO), or positive control (0.1 μM cisapride; Tocris Bioscience, Bristol, UK) was applied in physiological salt solution for 3–6 min at a superfusion flow rate of 3.5–3.9 mL/min. Cells were depolarized from a holding potential of −80 mV to +40 mV for 2 s followed by repolarization to −50 mV for 1.5 s; this voltage protocol was repeated once every 10 s. IC₅₀ was calculated in SAS/STAT (v14.1) using a probit regression model.

2.2.3. Human L–Type Calcium (hCa_v1.2/β2/α2δ1) Channels

Ca_v1.2 recordings were performed using CHO cells (source strain: ATCC, Manassas, VA; source substrain: ChanTest Corp., Cleveland, OH) stably expressing human Ca_v1.2/β2/α2δ1 cDNA. The intracellular solution consisted of 130 mM Cs‐MES, 20 mM tetraethylammonium chloride, 1 mM MgCl₂, 10 mM EGTA, 10 mM HEPES, 4 mM MgATP, 0.3 mM TrisGTP, 14 mM Tris‐phosphocreatine, and 50 U/mL creatinine phosphokinase. The extracellular solution was composed of the following: 137 mM NaCl, 4 mM KCl, 1.8 mM CaCl₂, 1 mM MgCl₂, 10 mM HEPES, and 10 mM glucose, with a pH adjusted to 7.4 with NaOH (0.3% DMSO added). For each measurement, the test article (30 or 100 μM cenobamate; SK Life Science Inc.), vehicle (0.3% DMSO), or positive control (0.1 μM nifedipine) was applied in HEPES‐buffered physiological saline. Cells were depolarized from a holding potential of −80 mV to +10 mV for 200 ms; this voltage protocol was repeated at 10‐s intervals. Analysis of currents was performed using pCLAMP software (version 8.2; MDS‐AT, Sunnyvale, CA).

3. Results

3.1. Sodium Channel Panel (Na_v1.1–Na_v1.8)

Many widely utilized ASMs including cenobamate have blockade of neuronal sodium channels as a proposed MoA. The effect of cenobamate on various human sodium channels (Na_v1.1‐Na_v1.8) was characterized using automated patch clamp on stably expressing cells (n = 3–4). Results are shown in Table 1. Cenobamate showed more potent sodium channel inhibition in the inactivated state compared under tonic block conditions or 10 Hz pacing rates. Because Na_v1.5 was inhibited by cenobamate and block of Na_v1.5 has been suggested as a possible risk factor for QRS prolongation and possible pro‐arrhythmic effects, a series of experiments was conducted to more thoroughly characterize cenobamate's effect on this sodium cardiac current.

TABLE 1.

Effects of cenobamate on sodium channel subtypes.

	Type of block	Nav1.1	Nav1.2	Nav1.3	Nav1.4	Nav1.5	Nav1.6	Nav1.7	Nav1.8
Cenobamate IC50, μM
	Tonic	> 1000	> 1000	> 1000	> 1000	> 1000	> 1000	> 1000	> 1000
	10 Hz	879	> 1000	> 1000	601	203	> 1000	125	> 1000
	Inactivated state	144	146	81.5	112	76.7	118	23.3	143
Lidocaine (positive control), mean % inhibition ± SD at 3000 μM (n = 3–4 cells)
	Tonic	66.21 ± 6.22	65.78 ± 5.08	65.16 ± 3.11	84.12 ± 2.49	91.52 ± 2.05	69.78 ± 5.24	99.51 ± 1.01	99.08 ± 1.03
	10 Hz	87.02 ± 3.81	85.16 ± 3.12	90.15 ± 1.12	99.11 ± 0.32	100.94 ± 0.03	90.42 ± 1.99	100.61 ± 1.35	100.00 ± 0.71
	Inactivated state	99.46 ± 1.18	96.97 ± 1.41	96.80 ± 0.47	99.79 ± 0.15	101.61 ± 1.05	98.43 ± 1.12	98.00 ± 0.77	100.72 ± 1.71

Abbreviations: IC₅₀, half‐maximal inhibition concentration; SD, standard deviation.

3.2. Peak Na_v1.5 Recovery

Peak Na_v1.5 is the current responsible for conduction throughout the myocardium. The peak Na_v1.5 concentration–response relationship recorded at physiologic temperatures for cenobamate was characterized and compared with representative drugs of Class Ia (quinidine), Class Ib (mexiletine), and Class Ic (flecainide). The current–voltage relationship shown in Figure 1A indicates proper voltage control. The example current traces before and after the addition of 1000 μM cenobamate are shown in Figure 1B. Cenobamate was the weakest peak Na_v1.5 blocker with an IC₅₀ of 655 μM. This was approximately 23 times higher than the Class Ic drug flecainide (IC₅₀ = 27.7 μM), approximately 12 times higher than the Class Ia drug quinidine (IC₅₀ = 55.6 μM), and nearly 5 times higher than the Class Ib drug mexiletine (IC₅₀ = 136.6 μM) (Figure 1C).

FIGURE 1.

Effects of cenobamate on peak Na_v1.5 current. (A) Current–voltage relationship (n = 14). (B) Example current trace in control and after 1000 μM cenobamate. (C) Concentration‐response relationship for the indicated drugs (n = 4–6). Symbols are mean ± standard error of the mean. Fit values: Cenobamate IC₅₀ = 655.0 μM, Hill value = 0.9; mexiletine IC₅₀ = 136.6 μM, Hill value = 1.1; quinidine IC₅₀ = 55.6 μM, Hill value = 1.3; and flecainide IC₅₀ = 27.7 μM, Hill value = 1.2.

To further define the sodium channel class of cenobamate, the recovery kinetics of block were characterized. A concentration very near the calculated IC₅₀ value was used for cenobamate and the comparator drugs. The fit equation consisted of a fast time constant (t _fast), which largely represents recovery from channel inactivation, and a slow time constant (t _slow), which represents recovery from drug block. As indicated in Figure 2A, the recovery from drug block kinetics of cenobamate were virtually identical to the Class Ib drug mexiletine (see Figure 2 inset) and dramatically faster than either quinidine or flecainide (Table 2). The quality of fit determined by adjusted R square was better using a double exponential equation for control, quinidine, and flecainide. The difference in adjusted R square was not meaningfully different between a single and double exponential function for cenobamate and mexiletine. Therefore, the more parsimonious fit was chosen.

FIGURE 2.

Recovery and rate‐dependent effects for the indicated drugs. (A) Recovery from drug block kinetics (n = 4). Inset shows a magnification of control, cenobamate, and mexiletine recovery kinetics. Symbols are mean ± standard error of the mean. For comparison, the kinetics of recovery from inactivation (control) are also shown (n = 11). The fit equation is Y = A1 × exp(x/t1) + A2 × exp(x/t2) + y0. Fit values are given in Table 2. (B) Rate‐dependent block of peak Na_v1.5 (n = 4–5). Symbols are mean ± standard error of the mean. To account for current reduction in the absence of drug, the drug‐control value is also provided.

TABLE 2.

Recovery from drug block kinetics.

Drug	IC50 (μM)	A1	τ1 (ms)	A2	τ2 (ms)
Control	NA	0.7 ± 0.1	2.5 ± 0.3	0.3 ± 0.02	29.1 ± 3.5
Cenobamate	655	1.0 ± 0.006	52.5 ± 21.4	NA	NA
Mexiletine	136.6	0.97 ± 0.02	56.4 ± 2.1	NA	NA
Quinidine	55.7	0.6 ± 0.05	14.8 ± 1.9	0.3 ± 0.08	3741.5 ± 964.6
Flecainide	27.7	0.6 ± 0.08	12.5 ± 0.9	0.3 ± 0.09	10,223.7 ± 4281.9

Note: Data are mean ± standard error. n = 4–11.

Abbreviations: τ₁, time constant associated with the 1st component from an exponential fit; τ₂, time constant associated with the 2nd component from an exponential fit; A1, amplitude of the 1st component of an exponential fit; A2, amplitude of the 2nd component of an exponential fit; IC₅₀, half‐maximal inhibition concentration; NA, not applicable.

These recovery kinetics would suggest that cenobamate would have very little rate‐dependent current reduction. Using the same concentrations as in Figure 2A, the amount of block at 1 and 3 Hz was characterized. Cenobamate indeed showed very little current reduction even at a pacing rate of 3 Hz (Figure 2B). The effects of cenobamate were again very similar to the Class Ib drug mexiletine. In contrast, both quinidine and flecainide were associated with marked current reduction at both 1 and 3 Hz.

Attempts were made to fit the onset of block (τ_on) at pacing rates of either 1 or 3 Hz with the tested drugs. As shown in Figure 3, a steady‐state level of current reduction was achieved within 10–15 pulses. Unfortunately, the small amount of current reduction upon exposure to either cenobamate or mexiletine did not allow for fitting of the onset of block. However, τ_on could be fit for quinidine and flecainide (Figure 3).

FIGURE 3.

Rate‐dependent effects for the indicated drugs on peak Na_v1.5 amplitude. Cells were paced at either 1 or 3 Hz until a steady‐state reduction in current amplitude was achieved. Each line represents a different cell. Fitting the time course of current reduction with a single exponential function provides the time constant of block (τ_on) (inset Table). The fit equation was Y = A1 × exp(x/t1) + y0. NA, not applicable; τ_on, time constant at the onset of block at pacing rates of either 1 or 3 Hz. ^aCould not be fit.

The cenobamate results presented thus far have been characterized at concentrations well in excess of those used clinically. To provide more relevant characterization of the peak Na_v1.5 effects of cenobamate, we used a free concentration of 66 μM. As indicated in Figure 4A, a cenobamate concentration of 66 μM had no effect on peak Na_v1.5 current amplitude. Furthermore, the recovery from block kinetics using 66 μM cenobamate was not statistically different from control (p < 0.05). The τ_slow for cenobamate was 83.1 ± 13.9 ms, very similar to that observed for control (75.5 ± 13.2 ms). Cenobamate formulations were analyzed using a validated method, and concentrations were within ±20% of nominal concentrations.

FIGURE 4.

Effects of a free therapeutic concentration of cenobamate (66 μM) on peak Na_v1.5. (A) Example current trace before and after 66 μM cenobamate. (B) Recovery from block kinetics for cenobamate and control (n = 5). Symbols are mean ± standard error of the mean. The fit equation is Y = A1 × exp(x/t1) + A2 × exp(x/t2) + y0. Mean ± standard error of the mean fit values for control: t ₁ = 2.8 ± 0.7 ms and t ₂ = 75.5 ± 13.2 ms and for cenobamate: t ₁ = 3.8 ± 1.2 ms and t ₂ = 83.1 ± 13.9 ms. Values were not different (p < 0.05).

3.3. Late Na_v1.5

The effects of cenobamate on late Na_v1.5 were also examined at physiologic temperatures. Block of late Na_v1.5 can shorten the QT interval. Cenobamate blocked late Na_v1.5 current with an IC₅₀ of 403.8 μM, which was slightly more potent than block of peak Na_v1.5 (IC₅₀ = 655 μM; Figure 5). This was compared with an IC₅₀ of 8.2 μM for ranolazine, a relatively selective blocker of late Na_v1.5 and was included as the positive control.

FIGURE 5.

Concentration‐response relationships for cenobamate block of late Na_v1.5 (n = 4–5) and hERG (n = 3–4). Symbols are mean ± standard error of the mean. The Hill value for late Na_v1.5 is 0.6 and for hERG is 1.1.

3.4. hERG

Blocking of hERG can prolong the QT interval. At physiologic temperatures, cenobamate was characterized as a weak blocker of the hERG current, with an IC₅₀ of 1869 μM. In comparison, the positive control cisapride inhibited the hERG current by a mean (SEM) 84.6% (±1.3%) at 0.1 μM. These results were consistent with historical data and validated the performance of the assay system. In addition, cenobamate formulations were analyzed using a validated method, and concentrations were within ±15% of nominal concentrations.

3.5. hCa_v1.2

Cenobamate inhibited the hCa_v1.2/β2/α2δ1 calcium current by a mean (SD) 2.7% (±2.3%) at 30 μM and by 7.0% (±5.4%) at 100 μM. In comparison, nifedipine, used as a positive control, inhibited the hCa_v1.2/β2/α2δ1 calcium current by a mean (SD) 88.8% (±10.5%) at 100 nM; these results were consistent with historical data and validated the performance of the assay system.

4. Discussion and Conclusions

This study was undertaken to evaluate the cardiac Na_v1.5‐blocking effects of cenobamate at supratherapeutic concentrations as well as therapeutic free plasma concentrations at physiologic temperatures. These results were compared with the effects of representative drugs from Class Ia, Ib, and Ic antiarrhythmics. Compared with mexiletine with an IC₅₀ of 136.6 μM, quinidine with an IC₅₀ of 55.7 μM, and flecainide with an IC₅₀ of 27.7 μM, cenobamate was the weakest blocker amongst the group with an IC₅₀ of 655 μM (cenobamate approved maximum dose of 400 mg/day, equivalent to 66 μM free concentration). Using a concentration similar to the IC₅₀ value, the rate of drug recovery with cenobamate was similar to mexiletine and much faster than quinidine or flecainide. These results suggest that cenobamate has cardiac sodium channel effects similar to the Class Ib drug mexiletine. The IC₅₀ of 655 μM reported here differs substantially from the IC₅₀ of 87.6 μM for block of peak Na_v1.5 with cenobamate that was reported previously by Mateias et al. [25]. However, there are several key methodological differences between the studies that make them difficult to compare. Notable differences in the present study design compared with the previous study included conducting the experiments at physiologic vs. room temperature and using a pacing rate of 0.1 Hz vs. a more rapid pacing rate of 1 Hz [25], which can lead to a greater amount of current reduction. Also, in the present study, the source of cenobamate was supplied by the pharmaceutical manufacturer with a Certificate of Analysis. In addition, a validated formulation method was used to analyze cenobamate concentrations in the external solution obtained from the cell bath, verifying that the intended concentrations were obtained and further supporting the results.

In addition to the experiments and the antiarrhythmic class characterization performed using supratherapeutic concentrations, experiments were also performed using a concentration of 66 μM, which represented the highest therapeutic free exposure at the maximum recommended dose of 400 mg/day cenobamate in humans. At this concentration, cenobamate had virtually no effect on the cardiac sodium current, making it difficult to assign an antiarrhythmic classification. The lack of effects on cardiac sodium channel currents were also consistent with observed clinical data in healthy subjects [20]. In a placebo‐controlled thorough QT study that assessed the electrocardiographic effects of cenobamate at therapeutic (200 mg/day) and supratherapeutic (500 mg/day) doses (maximum recommended dose, 400 mg/day) versus placebo in healthy adults, cenobamate had no clinically relevant effects on the QRS interval (mean values within ±1.0 ms of placebo) [20]. In the same study, cenobamate was associated with a shortening of QTc: a concentration‐QTc analysis predicted a shortening of ∆∆QTcF by 9.85 and 17.14 ms at mean peak plasma levels of therapeutic (86.2 μM; free 33.6 μM) and supratherapeutic (239.0 μM; free 93.2 μM) doses, respectively [20]. This effect may be explained by cenobamate's block of late Na_v1.5, as the block of late Na_v1.5 has been shown to shorten QTc for other drugs, such as ranolazine, quinine, lidocaine, mexiletine, and flecainide [23].

A component of the preclinical evaluation of ASMs requested by the FDA was defining an IC₅₀ value for Na_v1.5 and using this value for further characterizations such as recovery from block of the sodium channel. Although it is common to use an IC₅₀ value to compare drug effects on ion channels, an IC₅₀ value alone is not appropriate to use when assigning potential clinical risk [26]. For example, results presented herein indicate that the cardiac sodium channel IC₅₀ for cenobamate is 655 μM, nearly 10 times higher than the free plasma C_max after the maximum recommended therapeutic dose. Similarly, the IC₅₀ for lamotrigine block is 280.2 μM, approximately 11 times higher than the free plasma C_max (24.5 μM) at the recommended therapeutic dose [27]. As with cenobamate, therapeutic concentrations of lamotrigine have not been associated with QRS prolongation as reported in a new analysis of the thorough QT clinical trial [27]. These preclinical data along with the previously reported clinical data in healthy subjects [20] suggest that there is little relationship between the cardiac sodium channel IC₅₀ and arrhythmic risk at clinically relevant concentrations for either of these drugs.

A limitation of this study and other studies of its type is that the correlation between cardiac sodium channel block and QRS prolongation in patients with normal cardiac function or those with compromised function is not known. While a therapeutic concentration of cenobamate had virtually no effect on the cardiac sodium channel in this in vitro study, there are not enough data to determine whether cenobamate may affect cardiac sodium channels in patients with ischemic or structural heart disease. Also, these analyses do not take into account potential pharmacokinetic drug–drug interactions that can lead to increased plasma concentrations. For instance, cenobamate is an inhibitor of cytochrome P450 isoenzyme 2C19 [28]. Concomitant administration with the sodium channel blocker phenytoin can increase phenytoin plasma exposure by a mean of 84% [16].

The results of this in vitro study showed that despite classification as a Class Ib and Id antiarrhythmic at supratherapeutic concentrations, cenobamate at therapeutic concentrations has little effect on the cardiac sodium channel. These data provide further clarity and support for the clinical safety profile of cenobamate in patients without cardiovascular risk factors.

Author Contributions

All authors wrote the manuscript. W.J.C. and S.M.M. designed the research. W.J.C. performed the research. W.J.C. and S.M.M. analyzed the data.

Conflicts of Interest

W.J.C.: Employee, Nova Research Laboratories. Nova Research Laboratories was contracted by SK Life Science Inc. to perform this study. S.M.M. and K.J.G: Employee, SK Life Science Inc.

Supporting information

Data S1: cts70378‐sup‐0001‐supinfo.docx.

Crumb W. J. Jr, Melnick S. M., and Glenn K. J., “The Cardiac Sodium Channel Safety Profile of Cenobamate at Therapeutic Concentrations: In Vitro Analyses,” Clinical and Translational Science 18, no. 10 (2025): e70378, 10.1111/cts.70378.