Introduction

Top Abstract Introduction Methods Results Discussion References

Trans-2-en-NaVP is a major metabolite of VPA in humans that results from mitochondrial -oxidation (Baillie and Sheffels, 1995). In amygdala-kindled rats, VPA and trans-2-en-NaVP had approximately the same anticonvulsant efficacy, although both drugs produced motor impairment at effective doses (Löscher et al., 1988; Hönack et al., 1992). Trans-2-en-NaVP was more potent than VPA in the experimental treatment of myoclonic, clonic and tonic seizures in rodents and clonic seizures in dogs, when the two compounds were compared at their respective times of peak effect (Löscher et al., 1991a). The anticonvulsant efficacy of trans-2-en-NaVP occurred with less hepatotoxicity and teratogenic potential, compared with VPA (Hönack et al., 1992; Löscher et al., 1992, 1993). In addition, pharmacokinetic, histopathological and clinical chemistry data obtained from laboratory animal studies suggest that trans-2-en-NaVP might be a valuable substitute for VPA (Löscher et al., 1991, 1992, 1993).

Among other actions (Fariello et al., 1995), VPA has been reported to affect sodium-dependent neuronal activity. Voltage-clamp analysis in squid axons suggested that VPA (2 × 10² M) slowed sodium and potassium channel gating when applied internally (Fohlmeister et al., 1984). Voltage- and patch-clamp studies of vertebrate neurons demonstrated blockade of both sodium and potassium currents (Van Dongen et al., 1986; Zona and Avoli, 1990; Van den Berg et al., 1993). In addition, NaVP limited high-frequency SRF of sodium-dependent APs in mouse central neurons in cell culture at 37°C, at concentrations equivalent to clinically therapeutic free (not bound to proteins) plasma levels (McLean and Macdonald, 1986b). This led us to test the effect of trans-2-en-NaVP on SRF in the present study.

Less is known about the unsaturated metabolites of VPA, several of which have anticonvulsant efficacy in animals. Trans-2-en-NaVP is quantitatively predominant in adults. In snail neurons, trans-2-en-NaVP hyperpolarized more than, but lessened the frequency of depolarizing bursts less than, VPA (Altrup et al., 1992). Direct effects on sodium currents and sodium-dependent APs have not been reported. Both trans-2-en-NaVP and VPA diffuse mainly in extracellular space of the brain (Lucke et al., 1993) and may affect neurons in a time-dependent manner. Therefore, we examined the effects of acute (1-hr) and prolonged (24-48-hr) exposure to both trans-2-en-NaVP and VPA (here as NaVP) on intracellularly recorded electrophysiological properties of spinal cord and cortical neurons in dissociated monolayer cell culture.

The concentration range tested here included values from clinical pharmacokinetic studies. For 20 patients receiving VPA as monotherapy at an average daily dose of 1152 ± 661 ms/day (mean ± S.D.), the blood levels of VPA and trans-2-en-NaVP were 60.0 ± 22.6 µg/ml (~3.6 × 10⁴ M) and 3.0 ± 1.4 µg/ml (~1.8 × 10⁵ M), respectively (McLaughlin et al., 1992). VPA in plasma (free unbound fraction) may reflect the corresponding concentrations in tissue fluids and in brain (Vajda et al., 1981; Löscher et al., 1992). Vajda et al. (1981) studied nine neurosurgical patients receiving VPA orally as monotherapy, at doses between 1200 and 1600 mg/day, for at least 3 days before surgery for epilepsy. Sections of gray and white matter were obtained during surgery. CSF and plasma samples were taken simultaneously. VPA concentrations ranged from approximately 1.6 × 10⁴ to 9 × 10⁴ M in plasma, from 1.6 × 10⁵ to 2.3 × 10⁴ M in CSF and from 3.4 × 10⁵ to 1.6 × 10⁴ M in brain. Adkison et al. (1995) measured plasma and brain levels of VPA and its unsaturated metabolites in 24 patients, chronically receiving 740 to 4000 mg/day VPA, who subsequently underwent surgery for epilepsy. The mean plasma concentration of VPA was ~62.7 µg/ml (~3.8 × 10⁴ M), the free concentration was ~10.64 µg/ml (~6.4 × 10⁵ M) and the brain level was ~4.6 µg/g. The trans-2-en-NaVP level in plasma was ~2.9 µg/ml (~1.74 × 10⁵ M). A small free concentration of ~0.02 µg/ml (~0.12 × 10⁶ M) indicated extensive binding to plasma proteins. However, the brain tissue level was ~0.14 µg/g, indicating concentration in the brain over plasma. We tested both drugs at concentrations up to 60 mM.

We found that trans-2-en-NaVP and NaVP both limited SRF. The concentration dependence of limitation by both drugs shifted to the left with time of exposure. VPA was slightly more potent than trans-2-en-NaVP after prolonged exposure. Limitation by trans-2-en-NaVP occurred at concentrations that could be achieved with repeated oral dosing. Part of this work has been presented in abstract form (Wamil et al., 1994a).

	Methods

Top Abstract Introduction Methods Results Discussion References

Cell Culture

Spinal cord and neocortical cell cultures were prepared by published methods (Wamil and McLean, 1992). Briefly, embryonic mouse spinal cords (gestational day 13-14) and neocortices (day 14-16) were minced and dispersed, by trituration, to single cells and small clumps. The neurons were plated in collagen-coated dishes and maintained in vitro for 4 to 12 weeks before experimentation. The culture medium consisted of 80% (v/v) Eagle's minimal essential medium, 10% fetal calf serum (for the first week only), 10% heat-inactivated horse serum, 10 ng/ml 7S nerve growth factor and 1 ml/liter Mito Serum Extender (all supplements from Collaborative Research, Bedford, MA). Growth of non-neuronal cells was suppressed by brief treatment with 5-fluoro-2-deoxyuridine after 1 week. The Eagle's minimal essential medium was supplemented with 5.5 g/liter glucose and 1.5 g/liter sodium bicarbonate and was diluted to 300 to 325 mOsM with distilled water. After equilibration with the incubator atmosphere containing 10% CO₂, the pH of the culture medium was 7.4. Medium was changed twice weekly.

Experimental Protocols

General. The methods used here have been published in detail elsewhere (Wamil and McLean, 1992). For experiments, the culture medium was replaced with mDPBS with Mg⁺⁺ concentration elevated to suppress spontaneous synaptic activity (composition, in mM: NaCl, 143.4; KCl, 4.2; CaCl₂, 0.9; MgCl₂, 5.0-7.0; glucose, 5.6; in 9.5 mM sodium-phosphate buffer at pH 7.4). The culture dish was then placed in an aluminum block heated to 37°C, on the stage of an inverted phase-contrast microscope. Intracellular recordings of transmembrane potential were made during superfusion with drug-free and drug-containing solutions. A bridge circuit in the amplifier allowed simultaneous injection of current and recording of potential. The first derivative of the membrane potential (dV/dt) was obtained electronically, and the peak of the differentiated signal was proportional to _max. The _max indirectly reflected inward sodium current generating the upstroke of APs (for discussion, see McLean and Macdonald, 1983; Wamil and McLean, 1992).

Degree of repetitive firing. A series of depolarizing current pulses of 400-msec duration and variable amplitude were applied through the recording electrode to elicit overshooting APs in neurons with stable E_m values more negative than or equal to 45 mV. SRF of APs is defined as continuous high-frequency firing throughout one or more 400-msec depolarizations. All neurons in control medium demonstrated SRF. In drug-containing solutions, SRF could not be obtained from a percentage of neurons, depending on the drug concentration and the duration of drug exposure. In these neurons, a limited train (50-250 msec) of APs was evoked with a series of depolarizations above threshold and firing ceased for the remainder of the pulse after progressive decline of _max. With increasing depolarization past the point of maximal firing duration, the AP trains became shorter. This is referred to as limitation of SRF or limited firing. Limitation was judged to be present if firing could not be sustained throughout any of the depolarizations from the E_m or a less negative holding potential.

Drug application. Neurons were exposed to trans-2-en-NaVP or NaVP by superfusion with or without preincubation or by pressure application (Wamil and McLean, 1992, 1994). Drug application was terminated by washout of the drug with drug-free buffer. For superfusion experiments, control recordings were obtained from at least three neurons in each culture before drug application. If SRF was not present in all neurons, a new culture was obtained. After control recordings, the superfusate was changed to solution containing trans-2-en-NaVP or NaVP.

Drug preparation. Trans-2-en-NaVP was supplied in aqueous solution (100 mg/ml trans-2-en-NaVP; Desitin Arneimittel GmbH, Hamburg, Germany). That stock solution was diluted to desired concentrations in mDPBS. NaVP (RBI, Natick, MA) was dissolved in distilled water. Stock solutions were prepared at 10² g/ml (6 × 10² M) and then diluted to desired concentrations in mDPBS.

Statistics. Parametric data are given as means ± 1 S.E.M. The significance of differences between treatment groups was tested with the Student t test and Wilcoxon signed rank test. The IC₅₀ values for limitation of SRF and the significance of differences between values obtained at different times of exposure to trans-2-en-NaVP or NaVP were determined using computer-assisted linear regression analysis based on the method of Litchfield and Wilcoxon (1949).

	Results

Top Abstract Introduction Methods Results Discussion References

Passive membrane properties. During control or superfusion experiments with trans-2-en-NaVP and/or NaVP, the resting E_m in cortical or spinal cord neurons ranged from 40 to 80 mV. Spinal cord neurons in control mDPBS had a mean E_m of 60.7 ± 4.4 mV (n = 82). In a range of concentrations from 6 × 10⁹ to 6 × 10⁴ M (n = 121), the E_m averaged 58.4 ± 2.7 mV (n = 65). The difference was insignificant. R_in was 46.7 ± 2.9 M in control solution and 50.2 ± 4.1 M in trans-2-en-NaVP-containing solution (n = 29). In neocortical neurons, E_m was 53.1 ± 2.3 mV (n = 14) and did not change more than 2 mV during experiments with given cells; R_in was 47.8 ± 3.8 M (n = 10) and did not change substantially during exposure to either drug.

Concentration-dependent limitation of SRF by trans-2-en-NaVP and VPA. SRF was observed in spinal cord and cortical neurons in control solution. Acute (up to 1 hr) superfusion with trans-2-en-NaVP led to limitation of SRF, in a concentration-dependent manner. Limitation of high-frequency firing of APs occurred in spinal cord and cortical neurons (n = 36 and n = 10, respectively) with progressive reduction of _max before cessation of firing for the remainder of 400-msec depolarizations during superfusion with 6 × 10⁴ M trans-2-en-NaVP for 1 hr (fig. 1). The effect was reversible by hyperpolarization (fig. 1, A, 70 mV, and B, 72 mV) during exposure to the drug and by washout with drug-free mDPBS (fig. 1, POST).

View larger version (19K): [in this window] [in a new window]   Fig. 1.   Effect of trans-2-en-NaVP on SRF elicited by 400-msec depolarizing current pulses. Each row shows recordings from a single neuron (A, spinal cord neuron; B, cortical neuron). A, SRF was recorded before application of trans-2-en-NaVP (PRE). Pressure application of 6 × 104 M trans-2-en-NaVP for 3 min limited firing to 100 to 150 msec, with progressive decline in max, before cessation of firing for the remainder of the pulse (center left). Hyperpolarization resulted in recovery of SRF in the continuing presence of trans-2-en-NaVP (voltage-dependent unblock) (center right). SRF returned also after washout of trans-2-en-NaVP (POST). B, after SRF was recorded (PRE), this cortical neuron was exposed for 3 min to 6 × 104 M trans-2-en-NaVP. Limitation of SRF occurred within 150 msec of the depolarizing step (center left). Recovery was obtained by hyperpolarization (center right) and washout of trans-2-en-NaVP (POST). Top trace of each row, dV/dt. Bottom trace, intracellularly recorded transmembrane potential. Calibrations at bottom right apply throughout.

3

5

View larger version (16K): [in this window] [in a new window]   Fig. 2.   Concentration and time dependence of limitation of SRF by trans-2-en-NaVP (A) and NaVP (B). Ordinates, percentage of neurons capable of SRF. Abscissa, log concentration of trans-2-en-NaVP. A, percent SRF was decreased as a function of trans-2-en-NaVP concentration and time of exposure. B, acute (up to 1-hr) and prolonged (24-48-hr incubation and superfusion) exposure to NaVP reduced percent SRF of spinal cord neurons over a similar range of concentrations, with the exception of partial limitation acutely, compared with trans-2-en-NaVP. Numbers beside data points, number of neurons sampled at a given concentration.

4

5

Time-dependent limitation of SRF. Figure 3A shows the percentage of neurons with sustained firing as a function of time of exposure to a submaximal concentration of trans-2-en-NaVP. The percentage of spinal cord neurons with limited firing increased with time of exposure to trans-2-en-NaVP. About 50% of neurons had limited firing after exposure to 6 × 10⁷ M for 48 hr (fig. 3A). Fewer neurons were limited at different time intervals between 3 min and 48 hr (n = 36) (fig. 3A). With 6 × 10⁴ M trans-2-en-NaVP, 9 of 10 neurons were limited in 48 hr (fig. 3A).

View larger version (58K): [in this window] [in a new window]   Fig. 3.   A, the percentage of neurons capable of SRF decreased as a function of time of exposure to 6 × 107 and 6 × 104 M trans-2-en-NaVP. SRF was observed in all neurons in control solution (n = 32). Abscissa, times of exposure of different groups of neurons to trans-2-en-NaVP (3 min to 48 hr), including cells in a culture that had been exposed to trans-2-en-NaVP for 2 days and then changed to trans-2-en-NaVP-free medium for 5 days before re-recording (WASH). B, effects on cortical and spinal cord neurons were compared after 24 hr of exposure to trans-2-en-NaVP. Abscissa, log concentration of trans-2-en-NaVP. Numbers on bars, number of neurons sampled.

7

4

Different extents of limitation of SRF in cortical and spinal cord neurons by trans-2-en-NaVP. Sustained firing was observed in 10 cortical neurons incubated for 24 hr with 6 × 10⁸ M trans-2-en-NaVP. Firing was limited in 8 of 14 cortical neurons at 6 × 10⁷ M and in 5 of 10 cortical neurons at 6 × 10⁶ and 6 × 10⁵ M trans-2-en-NaVP (fig. 3B). A maximum of 70% limitation (7 of 10 neurons) was observed in cortical neurons exposed to 6 × 10² M trans-2-en-NaVP. All spinal cord neurons were limited at 6 × 10³ M trans-2-en-NaVP. In spinal cord neurons, under the same conditions, limitation was observed in 90% (9 of 10 neurons) at 6 × 10⁴ M (fig. 3B). Thus, cortical neurons appeared to be less sensitive to trans-2-en-NaVP, probably due to a lower frequency of AP firing (fig. 1) (see "Frequency Dependence").

Effect of the combination of trans-2-en-NaVP and NaVP on SRF. Figure 4 shows results obtained from neurons incubated for 24 hr with trans-2-en-NaVP, NaVP or both, at concentrations that limited SRF in a small percentage of neurons (fig. 2). Low concentrations of both drugs were selected to test for additive and hyperadditive effects of drug combinations. No hyperadditive effect was found using the three combinations shown in figure 4.

View larger version (39K): [in this window] [in a new window]   Fig. 4.   Effects of prolonged (24-hr) exposure to different concentrations of NaVP and/or trans-2-en-NaVP on spinal cord neurons. T-2-en, trans-2-en-NaVP.

Frequency dependence. The effect of 6 × 10⁴ M trans-2-en-NaVP on trains of APs depended on stimulation frequency. In control solution (fig. 5, Control, n = 8), stimulation at 1 and 100 Hz with 1-msec depolarizing current pulses elicited APs with nearly constant _max, despite some failures at 100 Hz. In neurons with limited firing during 400-msec pulses after overnight incubation with 6 × 10⁴ M trans-2-en-NaVP, many failures occurred even at 10 Hz (n = 10) (fig. 6A). The _max of APs that fired declined progressively, in a frequency-dependent manner (fig. 5). As shown in figure 6A, 82% of stimuli at 100 Hz fired APs in control solution; only 38% of stimuli fired APs at 100 Hz in solution containing 6 × 10⁴ M trans-2-en-NaVP (P < .05 vs. control by Wilcoxon signed rank test; for 17 neurons, E_m was 59.4 ± 4.7 mV). In four neurons with persistent unlimited SRF after overnight incubation with 6 × 10⁴ M trans-2-en-NaVP, _max declined slightly if at all at high frequencies of stimulation with trains of brief pulses. Stimulation of these neurons at 100 Hz resulted in few failures.

View larger version (20K): [in this window] [in a new window]   Fig. 5.   Frequency-dependent effect of trans-2-en-NaVP on AP firing. Trains of APs were elicited by 1-msec depolarizing current pulses of supratheshold amplitude at several frequencies after determination of whether firing was sustained (SRF) or limited during 400-msec depolarizations. Left, recordings from a neuron with SRF stimulated with trains at different frequencies in drug-free buffer. At frequencies greater than or equal to 50 Hz, some stimuli failed to elicit APs, but the max of APs that fired was not diminished significantly during the train. Right, after incubation for 24 hr and continued superfusion during experiments with 6 × 104 M trans-2-en-NaVP-containing solution, 18 of 20 neurons demonstrated limited repetitive firing and 2 of 20 neurons demonstrated SRF during 400-msec depolarizations. In a neuron with limited repetitive firing during 400-msec depolarizations, stimulation at rates faster than 5 Hz resulted in more failures than in control medium. At 100 Hz, max was diminished with repetitive firing and latency was prolonged. In a neuron with SRF during 400-msec depolarizations, little effect on max was noted during trains of 1-msec pulses at 100 Hz, although many stimuli failed to elicit APs (not shown). At similar latencies, APs in neurons with SRF had greater max than did APs in neurons with limited repetitive firing. Em values were constant throughout in each cell studied. Calibrations at bottom right apply throughout.

View larger version (29K): [in this window] [in a new window]   Fig. 6.   Effect of trans-2-en-NaVP on the percentage of stimuli at different frequencies that elicited APs and on absolute refractoriness. A, data from eight control neurons (Em = 60.4 ± 4.2 mV), showing that, at 1, 10 and 100 Hz, the percentage of AP failures was 95 ± 1.3, 92 ± 3.1 and 81.7 ± 5%, respectively. After exposure to trans-2-en-NaVP 104 g/ml (n = 9), the percentage of APs was 81.4 ± 6.5, 59.2 ± 8.9 (P < .05) and 39.5 ± 10.5% (P < .05), respectively. B, an absolute refractory period of 1.3 ± 0.2 msec in control solution was observed at 0.2-Hz frequency of two-pulse stimulation, to 3.2 ± 1.0 msec at 1 Hz. Data from 6 to 10 neurons exposed to 6 × 104 M trans-2-en-NaVP (Em = 53.4 ± 2.9 mV) show that the absolute refractory period was 4.4 ± 0.8 msec at 0.2 Hz (P < .01), 5.4 ± 1.0 msec at 0.5 Hz (P < .05) and 7.3 ± 1.8 msec at 1 Hz (P < .05). Numbers on bars, number of neurons sampled. * P  .05; ** P  0.1.

Effect of trans-2-en-NaVP on refractoriness. Refractoriness and recovery from inactivation were prolonged during exposure to NaVP (McLean and Macdonald, 1986b). To test the effect of trans-2-en-NaVP on these parameters, pairs of 1-msec depolarizing current pulses, identical in intensity and duration, were applied at variable intervals in neurons superfused with drug-free mDPBS or preincubated (12-24 hr) and superfused with 6 × 10⁴ M trans-2-en-NaVP. Figure 7 shows recordings from a spinal cord neuron (E_m = 60 mV) stimulated by paired pulses delivered every 1 sec at 1 Hz (maximum interval, 40 msec). In control solution, the absolute refractory period was 7 msec. During exposure to 6 × 10⁴ M trans-2-en-NaVP, this neuron had a 15-msec refractory period. Figure 6B compares refractoriness of trans-2-en-NaVP-treated neurons in which pairs of pulses were delivered at different intervals (0.2, 0.5 and 1 Hz) to each of nine neurons (E_m = 52.8 ± 4.5 mV). At the three intervals between delivery of paired pulses, the absolute refractory periods were 1.3 ± 0.2 (n = 5), 1.8 ± 0.4 (n = 7) and 3.2 ± 1.0 (n = 6) msec, respectively, in control solution. In trans-2-en-NaVP-containing solution, absolute refractoriness at the three intervals was 4.4 ± 0.8 (n = 5), 5.4 ± 1.0 (n = 10) and 7.3 ± 1.8 (n = 7) msec, respectively (fig. 6B). Absolute refractoriness in neurons treated with trans-2-en-NaVP was significantly different from that in control neurons at all three intervals (P < .01 at 0.2 Hz; P < .05 at 0.5 and 1 Hz; unpaired t test).

View larger version (30K): [in this window] [in a new window]   Fig. 7.   Trans-2-en-NaVP slowing recovery from inactivation. Pairs of 1-msec stimuli of identical amplitude were applied at variable interpulse intervals. The absolute refractory period was significantly longer in neurons exposed to 6 × 104 M trans-2-en-NaVP. Calibrations to the right apply throughout.

	Discussion

Top Abstract Introduction Methods Results Discussion References

Trans-2-en-NaVP has been shown to have anticonvulsant activity in animal models of partial and generalized seizures. The models include 1) amygdala kindling, a model of partial seizures secondarily generalized; 2) electroshock-induced seizures, a pharmacological model of tonic-clonic seizures; 3) generalized tonic-clonic seizures in genetically seizure-prone gerbils; 4) generalized absence-like activity in genetically seizure-prone rats; and 5) myoclonic and clonic seizures induced by pentylenetetrazole (Löscher et al., 1991; Semmes and Shen, 1991; Hönack et al., 1992). VPA, a broad-spectrum, clinically used, antiepilepsy medication, is effective in these models also. On this basis, trans-2-en-NaVP might be predicted to have a broad spectrum of clinical utility if adequate brain concentrations could be reached. The practical reason to determine this is that trans-2-en-NaVP is less toxic than VPA and could potentially become a substitute for the parent compound as a clinical antiepilepsy drug (Hönack et al., 1992).

To gain insight into the cellular mechanisms of action of trans-2-en-NaVP, we studied its effect on SRF of sodium-dependent APs by mouse central neurons in cell culture, an in vitro action that parallels efficacy against maximal electroshock-induced seizures in animals and partial seizures, with or without secondary generalization, clinically (for review, see Macdonald, 1983). Phenytoin, carbamazepine and VPA protect almost equally against these seizure types (Mattson et al., 1985, 1992). All three antiepilepsy drugs also limit SRF at concentrations equivalent to free (not bound to albumin) plasma levels that are therapeutically relevant (McLean and Macdonald, 1983, 1986a,b). High-frequency firing has been detected along subcortical pathways from a penicillin-induced cortical epileptogenic focus (Sypert and Reynolds, 1974). Limitation of such firing may be important in preventing the spread of seizures. Limitation by trans-2-en-NaVP was concentration, use and voltage dependent, as were the effects of the clinically used drugs.

Limitation of SRF by both trans-2-en-NaVP and NaVP was also time-dependent here. NaVP was slightly (about 4-fold) more potent in limiting SRF of spinal cord neurons after prolonged exposure. However, NaVP limited firing in only 75 to 80% of spinal cord neurons at the highest concentrations tested acutely. In previous experiments, inability of NaVP to limit firing in all neurons and inefficacy of 120 µM 2-en-VPA (not the trans-conformer) were observbed during exposures of <2 hr (McLean and Macdonald, 1986, fig. 4). Here, during superfusion for up to 60 min, a high concentration (60 mM) of trans-2-en-NaVP, but not NaVP, limited all exposed neurons. Both NaVP and trans-2-en-NaVP were more potent after prolonged exposure than they were acutely. Also, with time, NaVP became more potent than trans-2-en-NaVP in limiting SRF. A time-dependent shift in the concentration dependence of limitation of SRF was observed previously with gabapentin (Wamil and McLean, 1994), oxcarbazepine (Wamil et al., 1994), remacemide (Wamil et al., 1996) and phenytoin (M. J. McLean, unpublished observations). Exposure to 5 µg/ml gabapentin (equivalent to a therapeutic plasma level in clinical practice) for <5 min limited firing in only ~10% of neurons. Firing was limited in 40% of neurons exposed to this concentration for 1 hr, 60% exposed for 5 hr and ~80% exposed for 24 to 48 hr. These and the present findings suggest that prolonged exposure to investigational antiepilepsy compounds should be tested before the potential clinical significance of in vitro actions of the drugs is judged.

Mechanisms that could account for the delayed effects on SRF include slow uptake of trans-2-en-NaVP into neurons, slow access to a binding site or sites (possibly within or on the membrane and associated with the sodium channel), intracellular modification of the drug or its target site (e.g., phosphorylation of sodium channels) and/or modification of biochemical pathways. Perlman and Goldstein (1984) reported VPA to be a potent membrane-disordering agent. Because of the lipophilicity of VPA and trans-2-en-NaVP, this action should be fast and is less likely to account for the observed delay. The methods used here cannot distinguish among these possibilities, and we found no published data testing them critically in relation to efficacy. The delay parallels the delay to peak anticonvulsant activity of trans-2-en-NaVP (Löscher et al., 1993) and VPA (Wilder and Karas, 1982) in animals after i.v. administration. This is consistent with the interpretation that transport into neurons, an effect on enzymes (Löscher and Frey, 1977; Van der Laan et al., 1979; Löscher et al., 1991) or slow binding to a membrane site must occur before the anticonvulsant effect.

Overlap of the concentration dependence of limitation of SRF with clinically attainable concentrations suggests that limitation of SRF might be a clinically important mechanism of the protective efficacy of trans-2-en-NaVP. Plasma levels of trans-2-en-NaVP ranged between 370 and 600 µg/ml (~2.2-3.6 mM) in amygdala-kindled rats given 150 mg i.p. three times daily (Hönack et al., 1992). In epileptic patients, trans-2-en-NaVP derived from orally administered VPA actually accumulated in brain, relative to plasma concentrations, albeit at low levels (0.143 ± 0.026 µg/g vs. 0.022 ± 0.004 µg/ml) (Adkison et al., 1995). These low concentrations make it unlikely that accumulation of trans-2-en-NaVP derived from VPA is sufficient to enhance the effect of the parent compound. However, assuming accumulation after large systemic doses of trans-2-en-NaVP, brain concentrations as high as 14 to 25 mM might result at steady state. Such levels could make trans-2-en-NaVP a monotherapeutic antiepilepsy drug, if tolerable.

Prolonged incubation with trans-2-en-NaVP led to limitation of SRF in all spinal cord neurons tested at 1 mM, with an IC₅₀ value of 4.8 × 10⁵ M. The concentration-response curve stretched over 6 log units, suggesting that the desired concentration could be finely adjusted over time with large or small changes in oral dose. Acutely, exposure to 10 mM trans-2-en-NaVP limited firing in all neurons. The IC₅₀ for limitation was 1.3 × 10³ M (~216 µg/ml) acutely, but it was 4.8 × 10⁵ M (~8 µg/ml) after prolonged exposure, simulating chronic, repeated, clinical dosing. A maximum of 70% of cortical neurons had limited firing after prolonged exposure to 60 mM trans-2-en-NaVP, but firing of 50% of cortical neurons was limited at 6 × 10⁶ and 6 × 10⁵ M trans-2-en-NaVP (~1 and 10 µg/ml, respectively). Thus, systemic dosing could produce adequate CSF concentrations to limit firing rates in a large percentage of neurons. Presumably, as with phenytoin and carbamazepine, this could underlie anticonvulsant and antiepileptic efficacy, at least in part.

No hyperadditive effect was seen when trans-2-en-NaVP was combined with NaVP. Simple additive effects on SRF suggest a shared mechanism or mechanisms of antiepileptic action (Löscher et al., 1988; Hönack et al., 1992). The ability of trans-2-en-NaVP to limit SRF in vitro is shared by therapeutically relevant concentrations of several clinically used antiepilepsy compounds effective against partial seizures, with or without secondary generalization, including phenytoin (McLean and Macdonald, 1983), carbamazepine (McLean and Macdonald, 1986a), VPA (McLean and Macdonald, 1986b), felbamate (White et al., 1992), lamotrigine (Cheung et al., 1992), gabapentin (Wamil and McLean, 1994b) and oxcarbazepine (Wamil et al., 1994). Benzodiazepines (McLean and Macdonald, 1988) and barbiturates (Macdonald and McLean, 1986) limited SRF at concentrations encountered in treating status epilepticus. Phenobarbital protects many patients against partial seizures with or without secondary generalization. Thus, limitation of SRF is not the sole mechanism for efficacy against these seizure types. MK-801 (Wamil and McLean, 1992) and remacemide (Wamil et al., 1996) limited SRF at concentrations that are nontoxic in patients and that overlap the upper end of the range of concentrations required for each drug to block N-methyl-D-aspartate responses in vitro. Ethosuximide did not limit firing at 5 times the upper value in the clinically therapeutic range of concentrations (McLean and Macdonald, 1986b). Thus, limitation of SRF is unlikely to account for the broad spectrum of efficacy of trans-2-en-NaVP in animals and of felbamate and lamotrigine clinically.

AP firing frequency can be modulated by potassium and calcium currents. However, several of the present findings suggest an effect of trans-2-en-NaVP on sodium channels. These include 1) progressive reduction of _max until cessation of firing of APs, during both 400-msec depolarizing pulses and trains of 1-msec depolarizing steps at high frequency from a constant E_m; 2) prolongation of the absolute refractory period; and 3) prolongation of recovery from blockade. In addition, there was no effect on E_m and R_in. We did not observe the hyperpolarization observed by others in snail neurons (Altrup et al., 1992), perhaps because different conductances were present in the neurons studied here. A small percentage of cortical neurons showed no limitation of SRF when preincubated with trans-2-en-NaVP for 24 hr, perhaps due to the frequency dependence of block. Cortical neurons fired more slowly than did spinal cord neurons during 400-msec current pulses (fig. 1). Slow firing could result from activation of a calcium-dependent potassium conductance present in cortical, but not spinal cord, neurons. Nonetheless, cortical and hippocampal pyramidal neurons are capable of sustained high-frequency firing when this conductance is blocked, e.g., by biogenic amines (for review, see Nicoll, 1990). Thus, limitation of firing rates could occur in neurons in several different regions of the central nervous system. Proving the molecular mechanism of the effect on sodium-dependent APs will require further investigation with voltage (patch)-clamp techniques.

In conclusion, it appears that limitation of high-frequency AP firing could contribute to the anticonvulsant efficacy of trans-2-en-NaVP in animal models in which concentrations in the range of those required to limit SRF were achieved. Such concentrations are not achieved in patients when trans-2-en-NaVP is derived from VPA. Proof that effective levels of trans-2-en-NaVP could be achieved by oral dosing would require pharmacokinetic and efficacy testing in epileptic patients. The lower incidences of organ toxicity and teratogenic effects in animals with trans-2-en-NaVP than with VPA and the efficacy of trans-2-en-NaVP against a broad range of seizures in experimental models provide an impetus for such studies.