S. K. NORRIS 2 and A. E. KING

Transcription

1 /97/ $03.00/0 THE JOURNAL OF PHARMACOLOGY AND EXPERIMENTAL THERAPEUTICS Vol. 281, No. 3 Copyright 1997 by The American Society for Pharmacology and Experimental Therapeutics Printed in U.S.A. JPET 281: , 1997 The Stereo-Isomers of the Anticonvulsant ARL 12495AA Limit Sustained Repetitive Firing and Modify Action Potential Properties of Rat Hippocampal Neurons in Vitro 1 S. K. NORRIS 2 and A. E. KING Department of Physiology, Worsley Medical and Dental Building, University of Leeds, Leeds LS2 9NQ, U.K. Accepted for publication February 25, 1997 Because of the resistance of some forms of epilepsy to treatment using conventional anticonvulsants and the development of refractoriness as a consequence of long-term medication there is a need for effective anticonvulsants with low toxicity. Strategies for the development of novel anticonvulsants have focused on the suspected underlying central pathophysiologies such as decreased inhibitory ( -aminobutyric acid mediated) or increased excitatory (L-glutamate mediated) neurotransmission and abnormal neuronal excitability. The synthetic compound remacemide hydrochloride ((S,R)-2-amino-N-(1-methyl-1,2-diphenylethyl)-acetamide hydrochloride) has been favorably assessed for its antiepileptic and neuroprotective potential (Palmer et al., 1992, 1993) and has entered phase II clinical trials. Studies with radiolabeled remacemide hydrochloride in a number of animal species including human indicate that ARL 12495AA ((S,R)- 1-methyl-1,2-diphenylethylamine-monohydrochloride) is a major desglycinate metabolite of remacemide hydrochloride, Received for publication October 7, This work was supported by Astra Charnwood, Loughborough, UK. 2 Current address: School of Physiology and Pharmacology, University of New South Wales, NSW 2052, Australia. ABSTRACT The effects of the resolved enantiomers of the anticonvulsant ARL 12495AA ((S,R)-1-methyl-1,2-diphenylethylamine-monohydrochloride), (S)-ARL and (R)-ARL 12495, on (1) sustained repetitive firing and (2) action potential properties of rat hippocampal neurons were assessed. Whole-cell currentclamp recordings were made from CA1 neurons in slices of adult rat brain. Sustained repetitive firing was evoked by injection of long duration (500 msec) depolarizing ( pa) current pulses. Sustained repetitive firing was inhibited by (S)-ARL and by (R)-ARL 12495; the threshold concentration was 5 M reaching a near maximum at 400 M. Comparing the potencies of the two isomers, IC 50 values of 55 and 39 M were calculated for (S)-ARL and (R)-ARL 12495, respectively. The actions of the two drugs on neuronal firing were not therefore markedly stereoselective. Examination of individual spike properties revealed a concentration-related ( M) and time-dependent increase in the spike duration by (S)-ARL and (R)-ARL The spike amplitude and rate-of-rise were attenuated significantly by these two drugs. Both isomers decreased the after-hyperpolarization after a single spike and after trains of spikes. No clear stereoselectivity was demonstrable for the effects of the two enantiomers on action potential properties. Possible mechanisms of action for (S)-ARL and (R)-ARL including partial blockade of voltage-sensitive sodium channels and modulation of potassium channels are considered. The possibility that multiple mechanisms of action contribute to the therapeutic efficacy of the anticonvulsant ARL 12495AA is discussed. and that ARL 12495AA also possesses anticonvulsant properties (Palmer et al., 1992, 1993). Although some progress has been made in understanding the complex etiology of epilepsy and in the development of therapeutically useful drugs, a major consideration is the determination of central actions that may relate to clinical efficacy. Considering the anticonvulsant efficacy of remacemide hydrochloride and ARL 12495AA it was proposed that a contributory mechanism of action may be partial blockade of the NMDA receptor-ion channel complex. This suggestion was based on the fact that both the parent compound and its metabolite ARL 12495AA are effective against MES and NMDA-induced seizures in rats and mice (Stagnitto et al., 1990; Garske et al., 1991; Palmer et al., 1992). Also, remacemide hydrochloride did not prevent seizures in mice elicited by bicuculline or picrotoxin (Palmer et al., 1993). Both compounds are noncompetitive antagonists of [ 3 H]-dizocilpine (MK 801) binding to rat brain homogenates (Palmer et al., 1992). The metabolite ARL 12495AA has been shown to be more potent than remacemide hydrochloride against MESand NMDA-induced seizures (Palmer et al., 1992) and at displacing [ 3 H]-dizocilpine binding (IC 50 values of 0.48 and ABBREVIATIONS: CSF, cerebrospinal fluid; ADP, after-depolarization; AHP, after-hyperpolarization; I K(Ca), calcium-sensitive potassium current; NMDA, N-methyl-D-aspartate; MES, maximal electroconvulsive shock; SRF, sustained repetitive firing. 1191

2 1192 Norris and King Vol M for ARL 12495AA and remacemide hydrochloride, respectively (Palmer et al., 1992). Such observations have raised the possibility that some of the anticonvulsant activity of remacemide hydrochloride may be attributable to effects mediated by ARL 12495AA. Efficacy in the MES test predicts effectiveness against generalized tonic/clonic seizures (Upton, 1994), and drugs used in the treatment of generalized seizures such as phenytoin, carbamazepine and valproate, have been shown to limit high frequency firing of fast sodium-dependent action potentials in central neurons (McLean and Macdonald, 1983; 1986a, b). Thus, an alternative and perhaps overlapping mechanism of action that has been considered for some anticonvulsants is a direct or indirect interaction with neuronal sodium channels. Diazepam, for example, is believed to augment inhibitory synaptic transmission but at clinically relevant doses it limits the generation of sodium-dependent action potentials in cultured spinal neurons (Backus et al., 1991). Remacemide hydrochloride and ARL 12495AA have been reported to limit action potential firing of mouse cultured neurons (Wamil et al., 1992), and ARL 12495AA has been shown to inhibit veratridine-evoked (and hence sodium channel-mediated) release of glutamate from mouse cortical slices (Srinivasan et al., 1994). Such observations suggest that modulation of sodium channel activity may be relevant to the anticonvulsant efficacy of ARL 12495AA. Recent studies on ARL 12495AA have reported stereoselectivity, i.e., the activity of the resolved isomer (S)-ARL is higher compared to (R)-ARL The displacement of [ 3 H]-dizocilpine binding by ARL 12495AA is stereoselective, (S)-ARL is more than an order of magnitude more potent than (R)-ARL (IC 50 values of 0.25 and 3.4 M, respectively) (Palmer et al., 1992). In autoradiographic studies, (S)-ARL is more potent than (R)-ARL at displacing [ 3 H]-dizocilpine binding in the CA1 region of rat brain slices (IC 50 values of 1.5 and 18.4 M, respectively) (Porter and Greenamyre, 1995). In the same study, remacemide hydrochloride only weakly displaced [ 3 H]-dizocilpine binding in the CA1 area with an IC 50 value of 968 M. A similar order of potency has been reported for inhibition of NMDA-evoked currents in rat cultured hippocampal neurons, i.e., (IC 50 values); (S)-ARL (0.6 M) (R)-ARL (4 M) (S)-remacemide hydrochloride (67 M) (R)-remacemide hydrochloride (75 M) (Subramaniam et al., 1993). In the light of these findings, the aim of our study was first, to determine whether both enantiomers of ARL 12495AA were able to limit SRF of rat CA1 hippocampal neurons in vitro and second to determine any differences in the potency of (S)-ARL compared to (R)-ARL Methods Brain Slice Preparation Adult female Wistar rats ( g) were anesthetized with urethane (2 g kg 1, i.p.) then killed by decapitation. The brain was rapidly removed and placed in ice-cold artificial CSF of the following composition (concentrations in mm); NaCl 128, KCl 1.9, KH 2 PO 4 1.2, MgCl 2, 1.3, CaCl 2 2.4, NaHCO 3 26, glucose 10, ph 7.4, saturated with 95% O 2 and 5% CO 2. Serial coronal sections (350 to 400- m thick) of the hippocampus were cut in ice-cold artificial CSF using a vibrating microtome (Vibratome, Intracel. Ltd., Royston, Herts, U.K.). Slices were allowed to equilibrate in artificial CSF at room temperature for 1 hr. A single slice was then submerged in a tissue bath and perfused continuously with artificial CSF (21 23 C, 4 ml min 1 ). Experimental Procedure Drug preparation. Drugs were dissolved in distilled water to give stock solutions of 10 mm. Stock solutions were made up freshly each day then diluted in artificial CSF to the desired concentration and perfused through separate gravity-fed inlets for at least 15 min. Drug removal was achieved by returning to perfusion with drug-free artificial CSF. (S)-ARL and (R)-ARL (also known as ARL 12859AE and ARL 12860AE, respectively) were kindly provided by Astra Charnwood (Loughborough, U.K.) as the maleate salts. The structures for these compounds have been published in Palmer et al., (1992). All other compounds were purchased from Sigma-Aldrich Co. Ltd., Poole, Dorset, U.K. Electrophysiological recordings. Whole-cell recordings were made from neurons with a resting membrane potential more negative than 50 mv and with action potentials that overshot 0 mv. Recordings were made without visualization of the neurons. The patch pipettes were pulled from thin-walled fiber-filled borosilicate glass (1.2 mm o.d.) on a horizontal puller (Brown-Flaming, Sutter Instruments Co., Novato, CA). The patch pipettes were back-filled with a solution of the following composition (concentrations in mm); K-gluconate 125, NaCl 15, MgCl 2 2, EGTA 11, HEPES 10, ATP 1.5, GTP 0.2. The ph of the solution was adjusted to 7.4 with 1MKOH so that the final concentration of K was 140 mm. The resistance of the patch pipettes in artificial CSF was 4 to 6 M. Current-clamp recordings were made using an Axopatch-1D amplifier (Axon Instruments, Foster City, CA). Signals were filtered at 2 khz, digitized by a data recorder (Instrutech Corp., Great Neck, N.Y.) and stored on video tape (VHS, Sanyo, London, U.K.). Signals were monitored using a digital oscilloscope (Tektronix, London, U.K.) and captured on-line with an IBM-AT computer running SIGAVG software (CED, Cambridge, U.K.). Stimulation protocol. Repetitive firing was evoked by injection of rectangular depolarizing current pulses ( pa, 500 msec) from the resting membrane potential. A series of pulses of increasing amplitude were applied to each neuron to determine its currentfrequency relationship. (S)-ARL or (R)-ARL was then perfused for 15 min before the current pulses were repeated in the presence of the drug. Neuronal input resistance was calculated from the steady-state voltage response to a small amplitude hyperpolarising current pulse (20 40 pa, 500 msec). Data Analysis Data were sampled at 20 khz for analysis of single action potential properties and at 5 khz for SRF analysis. Statistical significance (95% confidence limits, i.e., P.05) was determined using twotailed, paired or unpaired Student s t test. Concentration-response relationships were obtained by plotting the drug concentration vs. the number of spikes elicited by a maximal depolarizing current pulse ( pa). The data were normalized to allow for differences in maximal firing. Concentration-response curves were fitted to the data by least squares regression to the equation Y Y max /(1 (IC 50 /[drug]) n ), where IC 50 is the concentration resulting in a 50% block of SRF and n is a slope factor, using an iterative procedure on an IBM-AT computer with Sigmaplot 5.0 software (Jandell Corporation, Erkrath, Germany). Y max was constrained to be 100. All data are presented as the mean S.E.M. Spike amplitude was measured from the firing threshold to the peak of the spike. Spike duration was measured from the firing threshold to the same voltage level on the downstroke of the spike. The rate-of-rise was calculated from the 20 to 80% amplitude values of the spike. The amplitude of the fast AHP after a single spike was measured from firing threshold level in response to a threshold current pulse. The

3 1997 Actions of (S)- and (R)-ARL amplitude of the slow AHP after a train of action potentials was measured 100 msec after the termination of a maximal depolarizing pulse. Results Effects on passive membrane properties. Whole-cell recordings were made from CA1 neurons with a mean S.E.M. resting membrane potential of mv and mean input resistance of M (n 22). No concentration tested of either (S)-ARL or (R)-ARL had any statistically significant effect on membrane potential. For example, the membrane potential was mv (n 5) in the presence of 400 M (S)-ARL 12495, and mv (n 5) in the presence of 400 M (R)-ARL At the highest concentration tested (400 M) each isomer increased neuronal input resistance; (S)-ARL caused an increase to M (n 5, P.001) and (R)-ARL caused an increase to M (n 5, P.001). However, at the lower concentration of 40 M, which was subsequently found to be near the IC 50 for each drug for SRF limitation (see below), neither isomer had any significant effect on input resistance; the input resistance was M (n 4) after 40 M (S)-ARL and M (n 5) after 40 M (R)-ARL Effects on sustained repetitive firing. All sampled neurons fired trains of fast action potentials with no evidence of spike frequency accommodation on injection of depolarizing current pulses (fig. 1). Over the range of currents tested ( pa) and under the present recording conditions, all neurons displayed almost linear current-frequency relationships (not illustrated). Typical maximal firing frequencies were 28 to 32 Hz with pulses of 200 to 400 pa. Superfusion with (S)-ARL (4 400 M) (fig. 1A) or (R)-ARL (4 400 M) (fig. 1B) significantly reduced the total number of action potentials evoked by any given depolarizing pulse. In the neuron of figure 1A, 40 M (S)-ARL reduced the firing frequency from 30 to 14 Hz and in the neuron of figure 1B a similar concentration of (R)-ARL reduced the frequency to 16 Hz. The effects of both drugs were slowly reversible and required a minimum wash period of 30 min; the re-establishment of SRF in a neuron that was exposed to 120 M (S)-ARL and then returned to control artificial CSF is illustrated in figure 2A. This limitation of SRF was apparently concentration related. At lower concentrations (4 40 M) of either drug the limitation of SRF was associated with an increased interspike interval (see for example the middle records of fig. 1) although at higher concentrations ( M) there was in addition a marked accommodation of firing frequency (fig. 1; lower traces). SRF was completely abolished by 400 M Fig. 1. (S)- and (R)-ARL limit sustained repetitive firing. Figure 1 shows typical voltage responses of two hippocampal neurons on injection of a current pulse (200 pa, 500 msec) from each neurons resting membrane potential (indicated beneath each trace). It can be seen that both neurons fired repetitively for the duration of the pulse with little spike frequency adaptation (i.e., sustained repetitive firing; SRF). Perfusion with (S)-ARL (A) or (R)-ARL (B) reduced the number of spikes evoked by the same current pulse. At a concentration of 40 M (middle traces) each isomer limited SRF principally by increasing the interspike interval, although at 400 M each isomer caused extreme firing frequency adaptation. Fig. 2. The limitation of SRF by (S)- or (R)-ARL is reversible and concentration dependent. Upper panel, SRF evoked by a 200 pa depolarising current step in control conditions (left trace) is limited to one spike after perfusion with 120 M (S)-ARL (middle trace). After a prolonged (30 min) perfusion with drug-free solution there was partial recovery of SRF (right trace). Lower panel, Concentrationresponse curves for (S)-ARL (filled circles) and (R)-ARL (open circles) were constructed as described in Methods. In this and all subsequent figures each point is the mean S.E.M. of data from at least five neurons. The IC 50 values for limitation of SRF were 55 M for (S)-ARL and 39 M for (R)-ARL

4 1194 Norris and King Vol. 281 (S)-ARL in three of five neurons and by 400 M (R)- ARL in two of five neurons. The concentration-response relationships where drug concentration is plotted against the number of spikes elicited by a maximal depolarizing current pulse ( pa) are illustrated in figure 2. Comparing the two drugs, IC 50 values were calculated (see Methods ) as 55 M for (S)-ARL and 39 M for (R)-ARL The similarity in these values for the enantiomers does not indicate marked stereoselectivity. Effects on spike amplitude, rate-of-rise and duration. (S)-ARL or (R)-ARL altered the properties of individual spikes. Figure 3A shows a typical example of the actions of (S)- ARL (40 and 400 M) (fig. 3A) on the first spike within a train evoked by a maximal depolarizing pulse. Because the effects of (S)-ARL and (R)-ARL on spike shape were equivalent, only examples for (S)-ARL are illustrated. In this neuron, it is apparent that (S)-ARL reduced first spike amplitude and reduced the rate-of-rise while increasing first spike duration. Note also the raised firing threshold (as indicated by the dashed lines in fig. 3A) after the drug. The quantified data for the effects of either drug over the concentration range of 12 to 400 M on these three first spike parameters are presented in the histograms of figure 3, B, C and D. The most consistent action of either drug was a concentration-related increase in the first spike duration (fig. 3D) with a threshold concentration of 12 M for either enantiomer. A trend for a reduction of the first spike rate-of-rise was evident (fig. 3C) but the data was highly significant (P.01) for either (S)-ARL or (R)-ARL only at the highest concentrations tested (400 M). Considering the first action potential amplitude, although an effect was observed with 12 M (R)- ARL a significant reduction by both enantiomers occurred at concentrations of 120 M (P.05) and 400 M (P.01). No clear stereoselectivity in the effects of these drugs on the parameters of the first spike was evident. A striking feature of the effects of either enantiomer was that the reduction of spike amplitude and rate-of-rise by (S)- or (R)-ARL became enhanced over time with successive spikes in a train (fig. 4). Typically under control conditions and (fig. 4, A and D, upper traces), the amplitude and rate-of-rise of successive spikes decreased to a level that was then maintained for the duration of the pulse. This decrement was more pronounced after superfusion of (S)-ARL Fig. 3. (S)- or (R)-ARL alters individual spike properties. A shows the effects on one neuron of (S)-ARL (40 and 400 M) on the first spike evoked by a depolarizing pulse of 200 pa. (S)-ARL reduced the amplitude and the rate-of-rise of the spike although increasing the duration of the spike. The firing threshold (as indicated by the dashed line in each trace) was also raised following perfusion with (S)-ARL at both concentrations. Similar effects on spike properties were observed with (R)-ARL (not shown). The effects of (S)- ARL (open bars) and (R)-ARL (filled bars) on spike amplitude (B), rate-of-rise (C) and duration (D) were concentration dependent. Statistical significance indicated by *P.05, **P.01, ***P.001.

5 1997 Actions of (S)- and (R)-ARL Fig. 4. The reduction of spike amplitude and rate-of-rise by (S)- or (R)-ARL is time dependent. A shows SRF of a neuron (evoked by a 200 pa, 500-msec pulse from 60 mv) in control conditions (upper trace) and after perfusion with (S)-ARL at 120 M (lower trace). In the absence of the drug (F) spike amplitude (B) and rate-of-rise (C) decrease to a sustained level. In the presence of 40 M (å)or120 M( ) (S)-ARL the reduction in spike amplitude and rate-of-rise are enhanced. D shows SRF of another neuron (evoked by a 250 pa, 500-msec pulse from 66 mv) in control conditions (upper trace) and after perfusion with 120 M (R)-ARL (lower trace). E and F show the time-dependent reduction of spike amplitude and rate-of-rise by 40 M (å) and 120 M ( ) (R)-ARL Scale bars apply to A and D. Statistical significance indicated as before (fig. 4A, lower trace) or (R)-ARL (fig. 4D, lower trace). The quantified data for the effects of 40 and 120 M (S)-ARL or (R)-ARL on the decrement of spike amplitude (fig. 4, B and E) and rate-of-rise (fig. 4, C and F) are presented; after the drugs, spike amplitude and rate-ofrise decreased more rapidly to reach significantly lower final values. As evident in figure 4, A and D, the enhancement of spike duration in the presence of either isomer also became more pronounced with successive spikes in a train. In control conditions, the duration of the first spike was msec (n 14) and the duration of the last (i.e., 15th) spike was msec (n 10). After 120 M (S)-ARL the duration of the first and last (i.e., 3rd) spikes was increased significantly to msec (n 5) and msec (n 3), respectively. Similarly, after perfusion with 120 M (R)-ARL the first spike duration was msec (n 5) and the last (i.e., 2nd) spike duration was (n 3). All values with either drug at 120 M were highly significantly different to control (P.01). Effects on spike AHP. (S)- and (R)-ARL reduced the fast AHP after a single spike evoked by a threshold current pulse (fig. 5A). The threshold concentration for reduction in fast AHP amplitude was 4 M for (S)-ARL and 12 M for (R)-ARL and was concentration-related over the range 4 to 400 M. In addition, the slow AHP after a train of spikes was reduced by (S)- or (R)-ARL (fig. 5B) over an equivalent concentration range. In some neurons abolition of either the fast or slow AHP unmasked an after-depolarization (ADP; shown as negative values in graphs of fig. 5). With high concentrations (120 and 400 M) of either drug the slow AHP was converted to an ADP in nearly every recorded neuron. Discussion Our data show that the resolved enantiomers of ARL 12495AA, namely (S)-ARL and (R)-ARL 12495, limited SRF of rat CA1 hippocampal neurons in vitro. The IC 50 values for SRF limitation were 55 and 39 M for (S)-ARL and (R)-ARL 12495, respectively. If the inhibition of SRF by ARL 12495AA is stereoselective then one isomer should be substantially more potent than the other at inhibiting SRF. As the potencies of the two enantiomers are similar to each other it is concluded that the effects of ARL 12495AA on high-frequency firing are not stereoselective. In

6 1196 Norris and King Vol. 281 Fig. 5. (S)- and (R)-ARL reduce AHP amplitude. (S)-ARL (open bars) and (R)-ARL (filled bars) caused a concentration-dependent reduction in the amplitude of the AHP after a single spike (A). A single spike was evoked with a threshold pulse and the AHP amplitude was measured as described in Methods. The inset in A shows a single rheobase spike in the absence and presence (arrow) of 40 M (S)-ARL In addition, (S)-ARL (open bars) and (R)-ARL (filled bars) caused a concentration-dependent reduction of the amplitude of the AHP after a train of spikes. The inset shows the effect of 120 M (S)-ARL (arrow) on the AHP after a train of spikes evoked by a 160 pa, 500-msec pulse from 64 mv. Statistical significance indicated as before. this respect, these data are in agreement with Wamil et al. (1996) who showed that remacemide hydrochloride and the racemate ARL 12495AA limited SRF of mouse cultured spinal cord neurons and that the stereo-isomers of ARL 12495AA were equipotent. These authors reported IC 50 values for SRF limitation of 7.9 M for remacemide hydrochloride, 3.3 M for (S)-ARL and 3.5 M for (R)-ARL These values for cultured spinal neurons are an order of magnitude more potent than the IC 50 values found in this study of CA1 neurons. The reason for this difference is unclear but could reflect species variation, differential neuronal susceptibility, variations in the recording conditions (e.g., temperature) and drug exposure time. From previous studies, it was suggested that limitation of SRF is a characteristic effect of anticonvulsant drugs such as phenytoin and carbamazepine that can partially block voltage-sensitive sodium channels (Willow et al., 1985; Lang et al., 1993). Although no direct evidence exists for such an action by (S)-ARL and (R)-ARL 12495, the concentration-related reduction of spike amplitude and rate-of-rise produced by each of the two enantiomers is consistent with sodium channel blockade. In particular, the spike rate-of-rise is often used as an indirect measure of the maximum sodium conductance of excitable membranes (Strichartz and Cohen, 1978; Catterall, 1987). Although the amplitude and rate-ofrise of the first spike were not significantly reduced by 40 M

7 1997 Actions of (S)- and (R)-ARL (S)- or (R)-ARL 12495, succeeding spikes in the train were significantly reduced by this concentration (which is similar to the IC 50 for SRF limitation). Mechanistically, phenytoin, carbamazepine and lamotrigine have each been shown to prolong recovery of sodium channels from inactivation (Willow et al., 1985; Lang et al., 1993) and it is possible that these enantiomers of ARL 12495AA act similarly. (S)- and (R)-ARL significantly increased spike duration, an effect that was due partly to a slower rise time after the drugs but mostly to a lengthening of the repolarization time. Because spike repolarization is dependent mainly on potassium channel activation it is possible that (S)- and (R)-ARL can modulate such channels. (S)- and (R)-ARL also reduced the amplitude of the fast and slow AHP which is further indirect evidence that each compound may modulate potassium conductances. However, the amplitude of the slow AHP depends critically on Ca influx during the preceding train of spikes to activate calcium-sensitive potassium currents (I K(Ca) ), therefore any reduction of the slow AHP may be a consequence of reduced SRF and not due to a reduction of I K(Ca) per se. Spike broadening is not elicited by anticonvulsants such as phenytoin and lamotrigine that are thought to block sodium channels only (McLean and Macdonald, 1983; Lees and Leach, 1993). Blockade by (S)- and (R)-ARL of potassium channels in addition to sodium channel blockade is consistent with the reported effects of (S,R)-ARL 12495AA on neurotransmitter release from mouse cortical slices (Srinivasan et al., 1994). This group found that (S,R)-ARL 12495AA inhibited glutamate release evoked by either veratridine (which prevents sodium channel inactivation) or potassium. By comparison, lamotrigine (which blocks sodium but not potassium channels) inhibits veratridine- but not potassium-evoked glutamate release from rat cerebral cortex slices (Leach et al., 1986). Voltage-clamp studies will be required to establish direct blockade of ionic channels by (S)- or (R)-ARL Recently, Wamil et al. (1996) reported that ARL 12495AA when applied to mouse spinal cord neurons at a concentration equivalent to its IC 50 for limitation of SRF in these neurons caused a use-dependent block of NMDA-induced responses. Other studies on hippocampal neurons have indicated a noncompetitive block of the NMDA receptor ion channel complex by the resolved enantiomers (Subramaniam et al., 1996). Given these proposed interactions with NMDAactivated channels, sodium and potassium channels, (S,R)- ARL 12495AA and its enantiomers are certainly not selective ion channel blockers. Paradoxically, it may be their interaction at more than one site that accounts for effectiveness against seizures in animal models. Indeed, blockade of sodium and NMDA receptor ion channels are synergistic in that both actions would be expected to result in dampening of excessive excitatory activity. Partial blockade of sodium channels theoretically would (1) stabilize presynaptic membranes, thereby inhibiting the pathological and potentially excitotoxic release of neurotransmitters such as L-glutamate and (2) stabilize postsynaptic membranes thereby limiting high frequency firing. Partial blockade of NMDA receptor ion channels would reduce NMDA receptor-mediated excitation thereby limiting excessive Ca entry. Which of these drug actions provides for the primary mechanism underlying the anticonvulsant action? Because inhibition of NMDA currents by (S)-ARL12495 has an IC 50 of 0.6 M (Subramaniam et al., 1993) then it is tempting to propose a primary action against abnormal excitatory amino acid-mediated excitation. However, this conclusion may be based on a premature and overly simplistic assumption about the aetiology of human epilepsy. Blockade of multiple ion channel types by an antiepileptic drug is not unprecedented. Felbamate may block sodium and NMDA receptor ion channels as it has been shown to limit SRF of mouse spinal cord neurons (White et al., 1992) and to inhibit NMDA-evoked currents in rat hippocampal neurons (Rho et al., 1994). However, blockade of potassium currents and a slowing of membrane potential repolarization by (S)- or (R)-ARL would be expected to enhance neuronal excitability, ultimately opposing the effects of sodium and NMDA ion channel blockade. Given that there is significant broadening of spikes by (S)- or (R)-ARL at concentrations that limit SRF it is probable that potassium channel blockade does occur in vivo. Why this excitatory action of ARL 12495AA does not appear to counteract the drugs antiepileptic actions is not clear. Perhaps the combined effects of sodium and NMDA channel blockade are sufficient to outweigh the effect of potassium channel block or it may be that inhibitory neurons are more susceptible than excitatory neurons to potassium channel blockade by (S,R)-ARL 12495AA. The question of the therapeutic relevance of any observed electrophysiological action of anticonvulsants applied acutely to isolated central nervous system preparations is not easily addressed. For clinically available antiepileptics such as lamotrigine and phenytoin, it is claimed that calculated free plasma levels would be in the concentration range sufficient to limit SRF (Cheung et al., 1992, McLean and MacDonald, 1983). In the case of the anticonvulsants remacemide hydrochloride and ARL 12495AA, any comparison is limited by the fact that the pharmacokinetics of the maximum dose tolerated in humans and its antiepileptic efficacy remains to be established. The pharmacokinetic profiles of (S)- and (R)- ARL12495 are also unknown. Remacemide hydrochloride is 77% bound to plasma protein (Scheyer et al., 1992) thus measured plasma concentrations undoubtedly will be higher than actual brain levels. In rats, the concentration of ARL 12495AA found in whole brain homogenates of rats effectively treated against MES (i.v. 20 mg kg 1 ) was approximately 10.9 g g 1 of tissue estimated to be equivalent to 52 M (J. Farmer, Astra Charnwood, Loughborough, U.K., personal communication) and this is similar to the IC 50 values of 55 and 39 M for (S)- and (R)-ARL reported here. However, these values should be viewed cautiously because an average whole brain concentration may not indicate the concentration localised at the active site. In summary, (S)-ARL and (R)-ARL effectively limited neuronal SRF. However, unlike the effects of ARL 12495AA at hippocampal NMDA receptor ion channels (Subramaniam et al., 1993, 1996), these effects on hippocampal firing properties were not markedly stereoselective. The likelihood that the anticonvulsant activity of ARL 12495AA is mediated by partial blockade of both sodium channels and NMDA receptor ion channels warrants further investigation. Acknowledgments The authors thank Astra Charnwood, Loughborough, U.K. for the gifts of (S)- and (R)-ARL 12495, J. Daniel and S. Shaw for technical assistance and D. Johanson for the photography.

8 1198 Norris and King Vol. 281 References BACKUS, K. H., PFLIMLIN, P. AND TRUBE, G.: Action of diazepam on the voltagedependent Na current. Comparison with the effects of phenytoin, carbamazepine, lidocaine and flumazenil. Brain Res. 548: 41 49, CATTERALL, W. A.: Common modes of drug action on Na channels: local anesthetics, antiarrhythmics and anticonvulsants. Trends Pharmacol. Sci. 8: 57 65, CHEUNG, H., KAMP, D. AND HARRIS, E.: An in vitro investigation of the actions of lamotrigine on neuronal voltage-activated sodium channels. Epilepsy Res. 13: , GARSKE, G. E., PALMER, G. C., NAPIER, J. J., GRIFFITH, R. C., FREEDMAN, L.R., HARRIS, E. W., RAY, R., MCCREEDY, S. A., BLOSSER, S. A., WOODHEAD, J.H., WHITE, H. S. AND SWINYARD, E. A.: Preclinical profile of the anticonvulsant remacemide and its enantiomers in the rat. Epilepsy Res. 9: , LANG, D. G., WANG, C. M. AND COOPER, B. R.: Lamotrigine, phenytoin and carbamazepine interactions on the sodium current present in N4TG1 mouse neuroblastoma cells. J. Pharmacol. Exp. Ther. 266: , LEACH, M. J., MARDEN, C. M. AND MILLER, A. A.: PHARMACOLOGICAL STUDIES ON LAMOTRIGINE, A NOVEL POTENTIAL ANTIEPILEPTIC DRUG: II. NEUROCHEMICAL STUDIES ON THE MECHANISM OF ACTION. EPILEPSIA 27: , LEES, G. AND LEACH, M. J.: Studies on the mechanism of action of the novel anticonvulsant lamotrigine (Lamictal) using primary neuroglial cultures from rat cortex. Brain Res. 612: , MCLEAN, M. J. AND MACDONALD, R. L.: Multiple actions of phenytoin on mouse spinal cord neurons in cell culture. J. Pharmacol. Exp. Ther. 227: , MCLEAN, M.J.AND MACDONALD, R. L.: Sodium valproate, but not ethosuximide, produces a use- and voltage-dependent limitation of high frequency repetitive firing of action potentials of mouse central neurons in culture. J. Pharmacol. Exp. Ther. 237: , 1986a. MCLEAN M. J. AND MACDONALD, R. L.: Carbamazepine and 10,11-epoxycarbamazepine produce a use- and voltage-dependent limitation of rapidly firing action potentials of mouse central neurons in cell culture. J. Pharmacol. Exp. Ther. 238: , 1986b. PALMER, G. C., MURRAY, R. J., WILSON, T.C.M.,EISMAN, M. S., RAY, R.K., GRIFFITH, R. C., NAPIER, J. J., FEDORCHUK, M., STAGNITTO, M. L. AND GARSKE, G. E.: Biological profile of the metabolites and potential metabolites of the anticonvulsant remacemide. Epilepsy Res. 12: 9 20, PALMER, G. C., CLARK, B. AND HUTCHISON, J. B. Antiepileptic and neuroprotective potential of remacemide hydrochloride. Drugs Future 18: , PORTER, R. H. P. AND GREENAMYRE, J. T.: Regional variations in the pharmacology of NMDA receptor channel blockers: Implications for therapeutic potential. J. Neurochem. 64: , RHO, J. M., DONEVAN, S. D. AND ROGAWSKI, M. A.: Mechanism of action of the anticonvulsant felbamate: Opposing effects on N-methyl-D-aspartate and -aminobutyric acid A receptors. Ann. Neurol. 35: , SCHEYER, R. D., CRAMER, J. A., LEPPIK, I. E., PELLOCK, J. M., HOCHHOLZER, J.M., SPENCER, M., OHARA, K. AND MUIR, K. T.: Remacemide elimination after initial and chronic dosing. Clin. Pharmacol. Ther. 51: 189, SRINIVASAN, J., RICHENS, A. AND DAVIES, J. A.: The desglycine metabolite of remacemide (FPL 12495) inhibits the release of glutamate from mouse cortical slices. Br. J. Pharmacol. 112: 630P, STAGNITTO, M. L., PALMER, G. C., ORDY, J. M., GRIFFITH, R. C., NAPIER, J.J., BECKER, C. N., GENTILE, R. J., GARSKE, G. E., FRANKENHEIM, J. M., WOODHEAD, J. H., WHITE, H. S. AND SWINYARD, E. A.: Preclinical profile of remacemide: A novel anticonvulsant effective against maximal electroshock seizures in mice. Epilepsy Res. 7: 11 28, STRICHARTZ, G.R. AND COHEN, I. S.: V max as a measure of G Na in nerve and cardiac membranes. Biophys. J. 23: , SUBRAMANIAM, S., DONEVAN, S. D. AND ROGAWSKI, M. A.: Effects of the anticonvulsant remacemide and its des-glycine metabolite on NMDA receptor currents in rat hippocampal neurons. Epilepsia 34 (Suppl 6): 117, SUBRAMANIAM, S., DONEVAN, S. D. AND ROGAWSKI, M. A.: Block of the N-methyl- D-aspartate receptor by remacemide and its des-glycine metabolite. J. Pharmacol. Exp. Ther. 276: , UPTON, N.: Mechanisms of action of new antiepileptic drugs: rational design and serendipitous findings. Trends Pharmacol. Sci. 15: , WAMIL, A. W., MCLEAN, M. J., CHEUNG, H. AND HARRIS, E.: Remacemide and its desglycinated metabolite limit sustained high-frequency repetitive firing of action potentials of mouse central neurons in vitro. Epilepsia 33: 85P, WAMIL, A. W., CHEUNG, H., HARRIS, E.W.AND MCLEAN, M. J.: Remacemide HC1 and its metabolite, FPL 12495AA, limit action potential firing frequency and block NMDA responses of mouse spinal cord neurons in cell culture. Epilepsy Res. 23: 1 14, WHITE, H. S., WOLF, H. H., SWINYARD, E. A., SKEEN, G.A. AND SOFIA, R. D.: A neuropharmacological evaluation of felbamate as a novel anticonvulsant. Epilepsia 33: , WILLOW, M., GONOI, T. AND CATTERALL, W. A.: Voltage clamp analysis of the inhibitory actions of diphenylhydantoin and carbamazepine on voltagesensitive sodium channels in neuroblastoma cells. Mol. Pharmacol. 27: , Send reprint requests to: Dr. A. E. King, Department of Physiology, Worsley Medical and Dental Building, University of Leeds, Leeds LS2 9NQ, U.K.