Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Topiramate [TPM; 2,3:4,5-bis-O-(1-methyl-ethylidene)-36-D-fructopyranose sulfamate] is a novel anticonvulsant drug that has recently been licensed for clinical use in many European countries. Traditional animal tests indicate that it primarily acts by blocking the spread of seizure activity and that it has an anticonvulsant profile similar to that of phenytoin and carbamazepine (Shank et al., 1994). Accordingly, the clinical efficacy of TPM has been particularly established in patients with simple or complex partial seizures (Rosenfeld et al., 1997; Sachdeo et al., 1997), which are also known to respond to phenytoin and carbamazepine (Heller et al., 1995). The basic mechanism of action of TPM has been investigated in various experimental preparations, and preliminary results indicate that both excitatory (Coulter et al., 1993, Severt et al., 1995) and inhibitory (Brown et al., 1993; White et al., 1997) chemical neurotransmission can be modulated by the drug. Furthermore, TPM has been found to be capable of inhibiting repetitive firing and spontaneous bursting in cultured hippocampal neurons (Coulter et al., 1993), thus suggesting that it may act through a use-dependent inhibition of Na⁺ channels, like several other antiepileptic agents that have been shown to inhibit sustained firing in central neuron preparations (McLean and Macdonald, 1986). Direct evidence of the TPM action on transient Na⁺ current (I_Naf) was recently provided by Zona et al. (1997), showing that in cultured cerebellar neurons, this drug induces a voltage-dependent inhibition that becomes evident at depolarized membrane potentials.

A significant effect on Na⁺ currents has been demonstrated for several other "traditional" and "new" antiepileptic drugs, such as phenytoin, carbamazepine, valproate (Matsuki et al., 1984; Willow et al., 1985; Kuo and Bean, 1994a; Ragsdale et al., 1996), and lamotrigine (Cheung et al., 1992; Xie et al., 1995; Kuo and Lu, 1997). Together, these findings suggest that Na⁺ current inhibition may play a primary role in the mechanism of action of antiepileptic drugs, especially those that are effective against partial seizures. Experimental evidence suggests that antiepileptic drugs, such as phenytoin, carbamazepine (Matsuki et al., 1984; Willow et al., 1985; Kuo and Bean, 1994a; Ragsdale et al., 1996), and lamotrigine (Xie et al., 1995) exert a voltage-dependent effect on voltage-gated Na⁺ channels by preferentially acting on inactivated channels. This suggests that their inhibitory action is turned on by long-lasting depolarizing potentials, similar to those characterizing epileptic ictal phenomena, leaving the generation of physiological Na⁺-dependent action potentials (APs) unaffected.

An additional antiepileptic mechanism affecting Na⁺ currents was recently proposed by Chao and Alzheimer (1995) for phenytoin, which was found to selectively reduce the persistent fraction of Na⁺ currents (I_NaP) at concentrations lower than those capable of reducing the peak of the fast Na⁺ current. Because this small fraction of the Na⁺ current is known to play an important role in regulating the firing characteristics of neocortical pyramidal neurons (Stafstrom et al., 1985; Franceschetti et al., 1995; Crill, 1996; Fleidervish and Gutnick, 1996; Guatteo et al., 1996, Mantegazza et al., 1998) and has been found to contribute to the generation of epileptiform activity (Segal and Douglas, 1997), it can be considered to be a further significant target when studying the Na⁺ channel-operated effects of new antiepileptic drugs. We here report the results of voltage-clamp experiments performed on acutely dissociated neocortical neurons, with the aim of investigating the effects of TPM on both the transient and persistent components of Na⁺ current. Control intracellular recordings to verify the drug effect on neocortical neurons were made using slice experiments

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Slice and Cell Preparation. Ten- to 20-day-old Wistar rats (Charles River, Italy) were anesthetized with chloral hydrate (Fluka AG, Buchs, Switzerland) and decapitated. Then brains were removed and placed in ice-cold artificial cerebrospinal fluid (ACSF) of the following composition: 120 mM NaCl, 24 mM NaHCO₃, 1 mM CaCl₂, 1 mM NaHPO₄, 4 mM MgCl₂, 2.5 mM KCl, and 20 mM glucose, bubbled with 95% O₂ and 5% CO₂. Coronal slices with a thickness of 300 to 350 µm were cut from the dorsal frontoparietal cortex using a vibratome. The slices used for the current-clamp experiments were transferred to an interface chamber, perfused with ACSF, and allowed to equilibrate for 1 to 1.5 h before the electrophysiological recordings were started. The dissociated neurons for the patch-clamp experiments were prepared from slices kept for 10 to 15 min in modified ACSF [in which NaHCO₃ was partially substituted with 10 mM HEPES-NaOH bubbled with 95% O₂ and 5% CO₂ at 35°C, pH 7.4] with the addition of 1 mM kynurenic acid and 1 mg/ml Protease Type XIV to digest the extracellular matrix; after enzyme treatment, the slices were washed and stored in an enzyme-free solution. At the recording time, the neurons were dissociated using fire-polished Pasteur pipettes, plated in a Petri dish (Costar Corp., Cambridge, MA), left 3 to 5 min to allow attachment, and then bath perfused with modified ACSF (see below). Only large neurons with a pyramidal shape were selected for patch-clamp recordings.

Electrophysiological Recordings. The intracellular recordings in slices were performed on layer V pyramidal neurons using an IR-283 (Neuro Data Inst. Corp., New York) amplifier. Sharp electrodes were prepared using borosilicate glass capillaries (Clark Electromedical Inst.) and filled with 3 M K⁺ acetate (resistance 80-90 M). Only healthy neurons with a stable spontaneous resting membrane potential (V_rest) that was more negative than 60 mV, a stable firing level and overshooting APs were selected for the analysis. Voltage and current signals were displayed on a Tektronix storage oscilloscope and stored on a magnetic tape. To isolate the Na⁺ currents, after neuron impalement the slices were bath perfused by a solution of 60 mM NaCl, 60 mM tetraethylammonium Cl, 0.2 mM CaCl₂, 7 mM MgCl₂, 2 mM CdCl₂, 24 mM NaHCO₃, and 10 mM glucose (pH 7.4), bubbled with 95% O₂ and 5% CO₂.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Using the whole-cell configuration of the patch-clamp technique, voltage-clamp recordings were obtained from 40 pyramidal neurons acutely dissociated from rat sensorimotor cortex slices. The intracellular recordings in current-clamp configuration were made in 9 pyramidal neurons of layer V in slices of sensorimotor cortex.

In preliminary experiments, rather high TPM concentrations (100-400 µM) were tested on the Na⁺-dependent APs generated by isolated neocortical neurons, which were recorded in current-clamp configuration and stimulated by brief (1.5-ms) depolarizing current pulses. The main drug effect was a reduction in the long-lasting AP shoulder (Fig. 1), which becomes visible after K⁺ and Ca⁺⁺ current blockade and is sustained by I_NaP (Stafstrom et al., 1985); with 200 µM TPM concentration, the duration of AP shoulder was reduced from 32.1 ± 8.1 to 19.2 ± 6.3 ms (n = 3). Voltage-clamp experiments on dissociated neurons then were carried out to further characterize the effects of TPM on both the transient and persistent components of Na⁺ currents and to establish the effective drug concentration.

View larger version (16K): [in this window] [in a new window]   Fig. 1.   Effect of TPM 200 µM on an AP shoulder evoked in response to a brief depolarizing pulse (top trace) in a dissociated neocortical neuron and recorded in current-clamp configuration after K+ and Ca++ currents blockade. Holding potential, 85 mV.

TPM Effect on Persistent Fraction of Na⁺ Current. A first series of experiments was designed to evaluate the effect of TPM on the persistent fraction of Na⁺ current (I_NaP), which was evoked in dissociated pyramidal neurons using slowly rising ramps as voltage stimuli. Depolarizing ramp potentials (from 70 to +10 mV at a rate of 80 mV/s), which were slow enough to avoid transient Na⁺ channel opening, were consistently capable of evoking a small inward TTX-sensitive Na⁺ current, which is considered to be accounted for by I_NaP (Alzheimer et al., 1993; Brown et al., 1994). The I_NaP in cortical neurons typically begins to activate between 60 and 50 mV and has a broad peak at about 40 mV (Fig. 2A), as previously found by French et al. (1990), Alzheimer et al. (1993), and Brown et al. (1994). A contaminating outward current (Fig. 2A), which was insensitive to the K⁺ blockers and remained after the addition of 1 µM TTX, was frequently found to activate at more positive potentials (more than 25 mV) and was assumed to correspond to the cationic current described by Alzheimer (1994) in neocortical neurons.

View larger version (15K): [in this window] [in a new window]   Fig. 2.   A, INaP evoked by a slow voltage ramp stimulus (top trace) in a representative neuron, under control solution and in the presence of 25 µM TPM. B, percentage inhibition of INaP peak amplitude evaluated on six neurons. Inhibition percentages obtained at each concentration were fitted using a Hill equation (see text).

Effect of TPM on Transient Na⁺ Current (I_Naf). The experiments were performed after partially substituting NaCl with choline-Cl (see Materials and Methods) to reduce voltage-clamp errors due to the large I_Naf observed in pyramidal neocortical neurons. Under this conditions, the maximum peak amplitude of I_Naf never exceeded 2.5 nA (range, 0.4-2.5 nA).

View larger version (16K): [in this window] [in a new window]   Fig. 3.   INaf evoked under control conditions and after 30 µM TPM in a representative neocortical neuron perfused by low-Na+ ACSF (see text). No inhibition was found when INaf was evoked from holding potential of 70 mV (A), but a slight reduction became evident (B) when test pulse was evoked from 60 mV. C, normalized current-voltage plot of INaf (Control, ; TPM, ) evoked from holding potential of 60 mV (n = 4).

View larger version (20K): [in this window] [in a new window]   Fig. 4.   Effect of 100 µM TPM on INaf peak amplitude and on activation and steady-state inactivation of Na+ channels. A-C, INaf evoked as described in Fig. 3, under control conditions and in the presence of 100 µM TPM. The drug was capable of slightly inhibiting the INaf evoked in a neuron held at 70 mV (A); degree of inhibition increased when holding rose to 60 mV (B) but was prevented when membrane potential was hyperpolarized to 90 mV to deinactivate most of the Na+ channels (C). D, in the presence of 100 µM of TPM (), voltage-dependence curve of activation of Na+ channels (four neurons) did not significantly differ from that evaluated under control conditions (). E, steady-state inactivation curve under control conditions and in the presence of 100 µM TPM (four neurons). Peak values of INaf were plotted against prepulse potentials and fitted using a Boltzmann relationship, showing a significant drug-induced leftward shift in V1/2 without any significant changes in slope factor (labels as in D).

Effect of TPM on Activation and Inactivation of Na⁺ Channels. The voltage dependence of Na⁺ channel activation appeared to be unaffected by both 30 µM (not shown) and 100 µM TPM. The fast activation curves of the Na⁺ channels evaluated under control conditions and in the presence of 100 µM TPM in four neurons held at 60 mV are shown in Fig. 4D. The conductance of the Na⁺ channels (G_Na) was calculated by the equation G_Na = I_Naf/(V  V_Na), where I_Naf is the maximal inward current, V is the command potential, and V_Na is the theoretical reversal potential (calculated by means of the Nernst equation to be +33.8 mV with 15 mM Na⁺ in the external solution). The normalized conductance (G/G_max) was plotted against the command potential following the Boltzmann relation: G/G_max = 1/(1 + exp[(V_1/2  V)/k]), where G/G_max is the relative conductance, V_1/2 is the voltage at which half-maximal activation is reached, and k is the slope factor. Under control conditions, the half-maximal activation potential (V_1/2) was 18.2 mV and k was 9.2 mV/e-fold; in the presence of 100 µM TPM, V_1/2 was 20.0 mV and k was 8.9 mV/e-fold.

View larger version (32K): [in this window] [in a new window]   Fig. 5.   Time-dependent TPM inhibitory effect on INaf on a representative neuron. Under control conditions, the INaf peak appears to be slightly decreased by incoming slow inactivation (A); in the presence of TPM 100 µM, inhibition was greatly enhanced (B). C, plot of peak INaf values versus prepulse length. Inset, stimulus protocol. Peak values of INaf were normalized to amplitude of current evoked after first inactivating prepulse. Time courses were fitted by the exponential equation: y = y0 + A exp[(t  t0)/], where y0 is offset value, A is peak relative amplitude at t = t0 (corresponding to first pulse of 240 ms), and  is decay time constant. Under control conditions, y0 was 0.73 ± 0.02 and  was 258.2 ± 49.2 ms; in presence of the drug, y0 was 0.21 ± 0.04 and  was 344.6 ± 39.4 ms. D, change in time course of recovery from inactivation in the presence of 100 µM TPM; holding potential was 70 mV. Inset, stimulus protocol. Peak current amplitudes at various times of recovery were measured and plotted (control, ; TPM, ). To fit the data points, the following monoexponential equation was used: y = y0 + A[1  exp(t/)]. Under control conditions, y0 = 0.63 ± 0.02 and  = 70.63 ± 11.2 ms; with 100 µM TPM, y0 = 0.55 ± 0.02 and  = 182.4 ± 46.2 ms.

Effect of TPM on Na⁺-Dependent Potentials Recorded in Slices. The intrinsically bursting neurons recorded in slice experiments were perfused with 30 and 100 µM TPM dissolved in bath solution. When most of the hyperpolarizing K⁺ currents were blocked by means of a high extracellular TEA concentration (40 mM) and the Ca⁺⁺-dependent K⁺ currents were blocked by substituting Ca⁺⁺ ions with Co⁺⁺ or Mn⁺⁺, the intrinsically bursting pyramidal neurons discharged with a fast AP followed by a very long depolarizing shoulder arising from AP repolarizing phase at about 20 mV and lasting 800 to 2000 ms. The neurons were left at spontaneous V_rest (65.5 ± 0.7 mV) and small changes in membrane potential, occasionally occurring during the recording time, were adjusted to the control V_rest by injecting weak DC currents. TPM 30 µM did not modify the amplitude and shape of the fast Na⁺-dependent AP (Fig. 6A) but consistently shortened the depolarizing shoulders (Fig. 6B); the length of the depolarizing shoulders was 1347 ± 650 ms under control conditions (mean of four sweeps measured on six neurons), 782.5 ± 403.0 ms during the addition of TPM (p < .02), and recovered during wash-out (1083.9 ± 560 ms, p = .68). Conversely, TPM 100 µM slightly decreased the peak amplitude of the AP (Fig. 6A) and induced a progressive reduction to the point of almost complete suppression of the long-lasting depolarizing plateau after AP activation (Fig. 6B).

View larger version (20K): [in this window] [in a new window]   Fig. 6.   Effect of TPM on AP and AP shoulders recorded in slices. A, amplitude and time course of fast AP was unaffected by 30 µM TPM, but a slight decrease in amplitude concurrent with a slower time to rise (see first derivatives in inset) became evident when slices were perfused with 100 µM TPM. B, in same neuron reported in A, AP shoulder was evidently shortened during perfusion of 30 µM TPM and almost completely suppressed when drug concentration rose to 100 µM.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

The data from this study demonstrate that relatively low TPM concentrations are capable of significantly inhibiting both the fast and persistent components of Na⁺ currents in neocortical neurons: 25 to 30 µM TPM concentrations were found to inhibit I_NaP by about 20% in dissociated neocortical neurons and to reduce the I_Naf peak amplitude provided that it was evoked from a slightly depolarized holding potential. The therapeutic values of TPM blood levels have been little investigated in clinical trials; however, the little published data indicate that the single (Johannessen, 1997) or chronic (Perucca, 1997) administration of an effective daily dose of 400 to 1200 mg/day give rise to plasma levels ranging from 4 to 10 µg/ml, which are very close to the lower TPM concentrations used in the present study.

The evidence that relatively low TPM concentrations principally inhibit the persistent fraction of Na⁺ current when the neurons were stimulated starting from their spontaneous V_rest, currently close to 70 mV (Tseng and Prince, 1993; Franceschetti et al., 1995, 1998), is supported by the results obtained in the voltage-clamp experiments performed on dissociated neurons and confirmed by the current-clamp experiments on neocortical slices. TPM 30 µM did not affect "fast" APs in the layer V pyramidal neurons but consistently shortened the long-lasting AP shoulders, which are revealed by Ca⁺⁺ and K⁺ blockade and have been attributed to I_NaP. Different ionic mechanisms contribute to afterdepolarization in neocortical neurons (Friedman et al., 1992; Reuveni et al., 1993) and, among them, a cationic current, sensitive to phenytoin inhibition, recently described by Kang et al. (1998). However, the large plateaux potentials that were found to be blocked by the TPM have been previously demonstrated to be sustained by a persistent TTX-sensitive Na⁺ current (Stafstrom et al., 1985) that is prominent in layer V of IB neurons (Franceschetti et al., 1995).

It has been shown that I_NaP and I_Naf in neocortical neurons are generated by uniform channel population that can switch to different gating modes, so I_NaP can be sustained by the small fraction of Na⁺ channels that fail to inactivate despite membrane depolarization, thus giving rise to sustained burst of channel openings (Alzheimer et al., 1993; Brown et al., 1994). This opening mode can be activated by a slow rising depolarization (such as that induced by a slow ramp potential) and has been shown to underlie the depolarizing plateaux potentials that spontaneously occur in an in vitro model of seizure-like activity obtained in cultured hippocampal neurons (Segal, 1994; Segal and Douglas, 1997). Stabilization of Na⁺ channels in their inactivated state could prevent their late opening and thus account for the TPM-induced inhibition of I_NaP that we found in both dissociated neurons and slices. Moreover, our results with TPM agree with those previously obtained with phenytoin, which, at concentrations significantly lower than those needed to inhibit the I_Naf, has been found to reduce I_NaP amplitude in dissociated neocortical neurons (Chao and Alzheimer, 1995) and to prevent the late channel opening in the above-mentioned model of ictal epileptiform activity obtained in hippocampal neurons (Segal and Douglas, 1997).

Binding to the inactivated state of Na⁺ channels, capable of preventing any transition that would allow channel reopening, could account not only for the effect of TPM on I_NaP but also for most of its effects on I_Naf, which can be revealed by means of different inactivating procedures. This assumption is further supported by the finding that membrane hyperpolarization at 90 mV, which is suitable for removing most of the channel inactivation, was sufficient to prevent drug activity even when it was added at higher concentrations (100 µM). Experiments aimed at further clarifying the dynamics and the mechanisms through which TPM can influence Na⁺ channels were performed using a 100 µM drug concentration, which in our study approximated the concentration that inhibits the persistent fraction of Na⁺ current by about 50% and also was sufficient to slightly inhibit the I_Naf peak in neocortical neurons held at 70 mV. In the presence of the drug, the steady-state I_Naf inactivation curve was found to be significantly shifted to more hyperpolarized potentials without any significant slope changes. This suggests that TPM binding is capable of preventing channel transition from an inactivated to a deinactivated state during the depolarizing prepulse and that less positive subthreshold depolarizing events are needed to decrease Na⁺-mediated membrane excitability. A leftward shift in steady-state inactivation has been previously reported to be induced by several anticonvulsants and by other drugs acting on Na⁺ channels (i.e., antiarrhythmic agents and local anesthetics) (Willow et al., 1985; Kuo and Bean, 1994a; Ragsdale et al., 1996); moreover, it has been found to occur in cultured granular neurons exposed to TPM (Zona et al., 1997), thus further indicating that drugs capable of controlling intrinsic membrane excitability can act by favoring steady Na⁺ channel inactivation.

The effect of TPM on I_Naf was not only voltage but also time dependent: drug-induced inhibition of I_Naf regularly increased when the neuronal membrane was kept depolarized by progressively longer conditioning prepulses at 40 mV. Furthermore, the recovery from the inactivation appeared to be slower in the presence of the drug than under control conditions. Both of these effects may be attributable to preferential drug binding on the slowly inactivated channel state, which is known to occur during long-lasting depolarizations (Ruben et al., 1992), or conversely, they could be due to slowly occurring binding and unbinding to the fast inactivated or closed but "activated" channel states, as proposed by Kuo and Bean (1994b). According to the hypothesis formulated by these authors to elucidate the action of phenytoin (which is the more extensively studied of the anticonvulsants acting on Na⁺ channels), drug binding may be slow but tight to the "activated" channels (i.e., channels with at least one gating charge in the "on" position), thus supporting the time-dependent increase in the inhibitory effect and the delayed recovery in response to a deinactivating hyperpolarization. Different anticonvulsant molecules might have different affinities for the activated Na⁺ channels (or different speeds of binding and unbinding), which can establish the time course of drug action and presumably imply differences in therapeutic effects (Kuo and Lu, 1997).

TPM-induced I_Naf inhibition regularly increased with the length of the depolarizing prepulses, as has also been reported for phenytoin (Kuo and Bean, 1994a), but the recovery due to deinactivating hyperpolarizing pulses was found to be more rapid than that associated with phenytoin and apparently comparable to that of carbamazepine and lamotrigine (Xie et al., 1995; Kuo et al., 1997). This suggests that the drug may regularly increase its effectiveness during long-lasting depolarizing events by adding to the normally occurring slow Na⁺ channel inactivation, whereas its effect may be more quickly removed by incoming hyperpolarizations, which can be sustained by either synaptic inhibition or hyperpolarizing currents.

Na⁺ Current Inhibition as an Antiepileptic Effect of TPM. TPM has been reported to possess multiple mechanisms of action and to be capable of modifying both -aminobutyric acid_A- and kainate-evoked ion flow at very low concentrations (Severt et al., 1995; White et al., 1997), and the drug probably can act on different seizure types throughout its different effects (see Perucca, 1997, for review). The present data show that like some other anticonvulsants known to be especially effective against focal seizures, TPM can inhibit Na⁺ currents, and that this action may occur at potentially therapeutic concentrations. Moreover, there is evidence to indicate that such relatively low drug concentrations should not significantly affect the physiological generation APs in well polarized neurons, and the drug action would come into play as much as sustained depolarizing events or high-frequency AP discharges are capable of reducing the fraction of the "resting" Na⁺ channels. Possible contribution toward controlling membrane hyperexcitability comes from the ability of TPM to inhibit I_NaP because the channel opening mode sustaining this current fraction may play a part in the generation of ictal events (Segal, 1994). Moreover, because I_NaP starts activating at more negative potentials than I_Naf (French et al., 1990; Alzheimer et al., 1993; Segal and Douglas, 1997), its inhibition may prevent membrane depolarizations before a neuron starts to fire. Furthermore, a TPM-induced limitation of the intrinsic I_NaP dependent bursting behavior may contribute significantly to reduce epileptic synchronization (Chagnac-Amitai and Connors, 1989).