Effects of 2-[N-(4-Chlorophenyl)-N-methylamino]-4H-pyrido[3.2-e]-1,3-thiazin-4-one (YM928), an Orally Active α-Amino-3-hydroxy-5-methyl-4-isoxazolepropionic Acid Receptor Antagonist, in Models of Generalized Epileptic Seizure in Mice and Rats

Epilepsy is one of the most common neurological diseases, affecting approximately 1% of the population (Sander and Shorvon, 1987). It is commonly accepted that a significant proportion of patients suffer from seizures resistant to drug treatment (Dam, 1986). The ratio of intractable patients has not been reduced despite the marketing of several new antiepileptics (AEDs) in recent years (Cramer et al., 1999). Thus, the development of new AEDs is expected.

Epilepsy is characterized by epileptic seizure caused by excessive neuronal discharge. Possible mechanisms of AEDs are the enhancement of inhibitory neuronal transmission and suppression of excitatory neuronal transmission. The mechanisms of most AEDs presently available can be categorized as blockade of voltage-dependent Na⁺ channels, potentiation of GABAergic transmission, and blockade of T-type Ca²⁺ channels (Macdonald and Kelly, 1995; Löscher, 1998). Compounds that selectively block excitatory neurotransmitter receptors are not used as AEDs.

Glutamate is a major excitatory neurotransmitter in the vertebrate central nervous system. Excessive activation of glutamate receptors is thought to be involved in the generation and propagation of epileptic seizures (Chapman, 1998), and glutamate receptor antagonists have been proposed as potential new AEDs. Glutamatergic receptors consist of ionotropic subtypes [N-methyl-d-aspartate (NMDA) receptors, α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) receptors, and kainate receptors] and metabotropic receptors. Competitive (Croucher et al., 1982) and noncompetitive (Troupin et al., 1986) antagonists for NMDA receptors have been reported to possess anticonvulsant action in preclinical seizure models. However, adverse effects such as psychosis hamper their clinical use (Rogawski and Porter, 1990; Sveinbjornsdottir et al., 1993). Selective antagonists of AMPA/kainate receptors are also candidates for new AEDs. AMPA receptors mediate fast excitatory neuronal transmission, and the anticonvulsant effects of selective antagonists such as NBQX (Ikonomidou and Turski, 1997), YM90K (Shimizu-Sasamata et al., 1996), GYKI52466 (Smith et al., 1991; Yamaguchi et al., 1993), and talampanel (Czuczwar et al., 1998) have been reported. However, the properties of these AMPA receptor antagonists have not been optimized for the treatment of epilepsy. For example, the short half-life of NBQX restricts its use in chronic diseases such as epilepsy (Gill et al., 1992; Chizh et al., 1994); a decrease in plasma concentration of talampanel is seen on concomitant use with enzyme-inducing AEDs (Langan et al., 2003); and nephrotoxicity is reported for NBQX (Xue et al., 1994). Thus, the development of novel AMPA antagonists has been sought.

2-[N-(4-Chlorophenyl)-N-methylamino]-4H-pyrido[3.2-e]-1,3-thiazin-4-one (YM928) is a selective, noncompetitive AMPA receptor antagonist. It inhibited kainate-induced neurotoxicity, AMPA-induced intracellular Ca²⁺ increase, and AMPA-induced current in rat hippocampal cultures. Oral administration of YM928 was effective against sound-induced seizure in DBA/2 mice (Ohno et al., 2003). However, because genetically epilepsy-prone rodents sometimes exhibit false-positive responses to nonanticonvulsant compounds, in the present study we examined the effect of YM928 in the maximal electroshock (MES) seizure test, pentylenetetrazol (PTZ)-induced seizure test, and strychnine (STR)-induced seizure test, all widely used animal models in the screening of potential AEDs, and in the AMPA-induced seizure test. The anticonvulsant effect of YM928 was compared with those of established AEDs. Effect on motor coordination was examined using the rotarod test. To assess the development of tolerance, effects of subchronic administration were assessed in the electroshock seizure threshold test.

Materials and Methods

All experiments were performed in accordance with the guidelines of the Animal Ethical Committee of Yamanouchi Pharmaceutical Co., Ltd.

Animals. Male Wistar rats (Japan SLC, Hamamatsu, Japan), weighing 180 to 220 g, were used for the MES- and PTZ-induced seizure tests. Male Fisher rats (F344/DuCrj; Japan Charles River, Yokohama, Japan), weighing 190 to 260 g, were used for the electro-shock seizure threshold test. Male ICR mice (Japan SLC), weighing 27 to 43 g, were used for the MES-, PTZ-, STR-, and AMPA-induced seizure tests. The animals were given free access to standard diet (CE-2; CLEA Japan, Tokyo, Japan) and tap water. Rats were kept in groups of three to five and mice in groups of 10 at a controlled temperature (23 ± 3°C) and humidity (55 ± 10%) with a 13-h light cycle (lights on from 7:30 AM to 8:30 PM).

Seizures Induced by Electroshock in Mice and Rats. Mice were stimulated with corneal electrodes from a stimulator (MK-800; Muromachi Kikai, Tokyo, Japan) using a suprathreshold current (50 Hz, 50 mA, 0.2 s). Rats were stimulated with auricular electrodes from a stimulator (MK-810; Muromachi Kikai) using a suprathresh-old current (50 Hz, 40 mA, 0.2 s). The electrodes were placed in 0.9% sodium chloride solution before application. Tonic hind limb extension (limb extension exceeding a 90° angle with the plane of the body) was used as the criterion of convulsion. Drugs were administered orally 60 min before the stimulus. ED₅₀ values, the dose at which tonic hind limb seizures were prevented in 50% of animals, and 95% confidence interval were calculated (n = 10/group).

Seizures Induced by PTZ in Mice and Rats. PTZ at a dose of 100 mg/kg in mice and 70 mg/kg in rats was injected subcutaneously 60 min after the oral administration of test compounds. The animals were observed for 30 min after injection and wild running, clonic seizures, tonic seizures, and respiratory arrest were monitored. ED₅₀ values and 95% confidence interval of clonic seizure were calculated (n = 9-10/group).

Seizures Induced by STR in Mice. STR at a dose of 0.8 mg/kg was injected subcutaneously 60 min after the oral administration of test compounds. The animals were then observed for 30 min after injection and wild running, clonic seizures, tonic seizures, and respiratory arrest were monitored. ED₅₀ values and 95% confidence interval of tonic extension seizure were calculated (n = 10/group).

Seizures Induced by AMPA in Mice. For the intracerebroventricular (i.c.v.) injection of AMPA, cannulae were prepared according to a previously reported method (Nakajima et al., 1994). Briefly, mice were anesthetized with pentobarbital (2.5-3.0 mg/mouse i.p.), and the skull was exposed. The tip of the injection apparatus was placed at 0.9 mm lateral and 0.7 mm posterior to the bregma, at a depth 3.0 mm below the surface of skull. It was held in place with dental cement applied to the exposed skull surface. After the operation, mice were housed individually to avoid damage to the injection apparatus. After a recovery period of at least 6 days, test compounds were injected orally. Sixty minutes later, AMPA at a dose of 1 μg/mouse was injected at a volume of 4 μl. Wild running, clonic seizures, tonic seizures, and respiratory arrest were monitored for 10 min. ED₅₀ values and 95% confidence interval of clonic seizure were calculated (n = 8/group).

Rotarod Performance in Mice. Animals that were able to remain on a rotarod apparatus revolving at 5 rpm for 120 s were initially selected for the evaluation. Test compounds were then administered orally and rotarod performance was retested 60 min later. Mice that were not able to remain on the apparatus for 60 s in three trial sessions were termed to have motor impairment. The number of mice having motor impairment was counted, and TD₅₀ values and 95% confidence interval were calculated (n = 10/group).

Electroshock Seizure Threshold Test, Tolerance Study in Rats. On day 1, the threshold for seizures induced by electroshock in rats was determined via auricular electrodes by means of a current stimulator with a sine wave stimulus (50 Hz, 0.2 s). YM928 at doses of 2, 4, and 8 mg/kg p.o. was administered orally 120 min before stimulation (n = 16-20/group). Stimulus intensity was varied by an up-and-down method in which the intensity of current was lowered or raised if the preceding animal did or did not show hind limb extension, respectively. From days 2 to 15, YM928 was administered once a day, and on day 16 the threshold was determined 120 min after the injection of YM928. On days 17 and 23, the threshold was determined without drug administration. The same animals were repeatedly used throughout the study.

Drugs. YM928 and diazepam were synthesized by Yamanouchi Pharmaceutical Co., Ltd. Lamotrigine was obtained from Glaxo Wellcome Research and Development Ltd. (Hertfordshire, UK). Gabapentin was obtained from Parke-Davis (Ann Arbor, MI). Carbamazepine and valproate were purchased from Wako Pure Chemicals (Osaka, Japan). Phenytoin was purchased from Dainippon Pharmaceutical (Osaka, Japan). Phenobarbital was purchased from Sanko Seiyaku Kogyo (Tokyo, Japan). Ethosuximide, STR, and AMPA were purchased from Sigma-Aldrich (St. Louis, MO). PTZ was purchased from Tokyo Kasei (Tokyo, Japan). PTZ, STR, and AMPA were dissolved in saline. Valproate and ethosuximide were dissolved in distilled water. All other test compounds were suspended in 0.5% methylcellulose solution. Drugs were administered at a dosing volume of 10 ml/kg in mice and 2 ml/kg in rats. Doses are expressed in terms of base.

Statistical Analysis. ED₅₀ values for anticonvulsant tests and TD₅₀ values for the rotarod test were calculated by the method of probit analysis. In the subchronic administration test, seizure threshold and standard deviation were calculated by the reported method (Kimball et al., 1957). Student's t test was used to determine the difference in threshold on the first and last days of administration. Differences were considered significant when p < 0.05.

Results

Anticonvulsant Activity in Mice. The electroshock produced hind limb extension in all animals that received vehicle solution. YM928 at doses of 2 to 30 mg/kg p.o. dose dependently suppressed the tonic extensions, with complete inhibition observed at doses of 15 to 30 mg/kg (Fig. 1A). ED₅₀ value of YM928 was 7.4 mg/kg. Carbamazepine, phenytoin, lamotrigine, valproate, phenobarbital, diazepam, and gabapentin showed anticonvulsant activity (Table 1). Ethosuximide, which blocks T-type Ca²⁺ channels, was ineffective at the dose range tested.

PTZ (100 mg/kg) induced clonic seizures in all mice treated with vehicle. YM928 at doses of 3 to 30 mg/kg p.o. dose dependently suppressed the clonic seizures, with complete suppression observed at 20 to 30 mg/kg (Fig. 1B). The ED₅₀ value was 9.6 mg/kg. Valproate, phenobarbital, diazepam, and ethosuximide were effective (Table 1). The ED₅₀ value of YM928 in the MES test did not differ significantly from that in the PTZ-induced seizure test. Valproate and phenobarbital inhibited MES- and PTZ-induced seizures with similar potencies. In contrast, the effect of diazepam against PTZ-induced seizure was much more potent than that in MES-induced seizure. Carbamazepine, phenytoin, and lamotrigine, which exert a blocking action on voltage-dependent Na⁺ channels, and gabapentin were not active.

STR (0.8 mg/kg) induced seizures in all mice treated with vehicle. YM928 at doses of 5 to 50 mg/kg p.o. dose dependently suppressed the tonic extension, with complete suppression observed at 50 mg/kg (Fig. 1C). ED₅₀ value was 14.0 mg/kg. Valproate, carbamazepine, phenytoin, lamotrigine, diazepam, ethosuximide, and gabapentin showed anticonvulsive activity (Table 1). No anticonvulsant activity was detected for phenobarbital.

In the AMPA-induced seizure model, mice showed wild running, clonic seizures, and tonic seizures immediately after injection. YM928 at doses of 0.5 to 20 mg/kg p.o. dose dependently suppressed the clonic seizures (Fig. 1D). The ED₅₀ value was 5.5 mg/kg. Valproate, carbamazepine, phenytoin, and lamotrigine showed anticonvulsant activities (Table 1). Under the present experimental condition, the potencies of YM928 in MES- and AMPA-induced seizure tests were similar. In contrast, carbamazepine, phenytoin, and lamotrigine were less potent in the AMPA-induced seizure test. No anticonvulsant activity was detected for phenobarbital, diazepam, and ethosuximide.

Anticonvulsant Activity in Rats. The Antiepileptic Drug Development Program proposed that anticonvulsant action be confirmed in another rodent species (Porter et al., 1984). The effects of YM928 and reference compounds were therefore examined in rats. Electroshock induced tonic extension in all rats treated with vehicle. YM928 at doses of 1 to 10 mg/kg p.o. dose dependently suppressed the tonic extension induced by electroshock (Fig. 2A). The ED₅₀ value was 4.0 mg/kg. PTZ (70 mg/kg) induced clonic seizures in all rats treated with vehicle. YM928 at doses of 2 to 20 mg/kg p.o. suppressed the clonic seizure induced by PTZ (Fig. 2B). The ED₅₀ value was 6.2 mg/kg. The effects of YM928 were compared with those of AEDs. Carbamazepine, phenytoin, lamotrigine, valproate, phenobarbital, and diazepam showed anticonvulsant activity in electroshock seizure test (Table 2). Valproate, phenobarbital, diazepam, and ethosuximide were effective in the PTZ-induces seizure test (Table 2). The anticonvulsive profiles of AEDs were similar to those in mice.

Rotarod Performance in Mice. YM928 significantly impaired rotarod performance in mice (TD₅₀ = 22.5 mg/kg; Fig. 3). The protective index (P.I., rotarod TD₅₀/MES ED₅₀) was calculated to be 3.0. TD₅₀ values of reference compounds were carbamazepine, 120 mg/kg; phenytoin, 240 mg/kg; lamotrigine, 63.7 mg/kg, valproate, 1120 mg/kg; phenobarbital, 49.9; diazepam, 9.1 mg/kg; and ethosuximide, 721 mg/kg (Table 1). The P.I. values of these compounds were 0.7 to 17.2. Gabapentin did not induce motor disturbances at a dose of 800 mg/kg (Table 1).

Electroshock Seizure Threshold Test, Tolerance Study in Rats. To examine the effect of YM928 after repeated administration, seizure threshold was measured on the first and last days of subchronic administration in rats. Based on the ED₅₀ value of YM928 in the rat MES seizure test, YM928 at doses of 2 to 8 mg/kg were used. Seizure threshold was not significantly altered after repeated administration of YM928 (Fig. 4). Seizure threshold was reexamined 1 and 7 days after the subchronic administration to assess the possibility of rebound (a decrease in the threshold of convulsions compared to that of the control group). Seizure threshold in rats injected with YM928 was not different from those receiving vehicle (Fig. 4, days 17 and 23).

Discussion

The present study clearly demonstrates that YM928 exerts anticonvulsant actions in models of generalized seizure. Oral administration of YM928 protected against seizures induced by electroshock in mice and rats. Anticonvulsants that block voltage-dependent Na⁺ channels show potent effects in the MES test, and the test is generally thought to be a model of generalized tonic-clonic seizure in human (Macdonald and Kelly, 1995; Löscher and Schmidt, 1988). YM928 is therefore expected to be valuable in the treatment of generalized tonicclonic seizures. The electric current used in the mice experiments (50 mA, 0.2 s) was of the same intensity as that commonly reported by other researchers, whereas that for rats (40 mA, 0.2 s) was weaker: here, intensity was set just above threshold so that YM928 dosage in the subchronic seizure test could be adjusted according to results. ED₅₀ values for carbamazepine and phenobarbital were comparable with those in a previous study using stimulation at 150 mA (Porter et al., 1984). In contrast, ED₅₀ for diazepam was markedly lower in our study, suggesting that the efficacy of diazepam in the MES test critically depends on experimental conditions.

YM928 showed anticonvulsant effects in the PTZ-induced seizure test also. Anticonvulsant efficacy against seizures induced by PTZ is thought to correlate with that against human generalized absence/myoclonic seizures (Löscher and Schmidt, 1988; Macdonald and Kelly, 1995). Considering the efficacy of YM928 against PTZ-induced seizures, a beneficial effect of YM928 in these seizure types is also expected. It was recently pointed out that results in the MES and PTZ tests do not always predict clinical efficacy of novel AEDs against generalized tonic-clonic and absence seizures, respectively (Löscher, 1998). Conclusive characterization is therefore dependent on clinical evaluation. Lamotrigine and gabapentin were not active against clonic seizure induced by PTZ. Previous study reported that lamotrigine and gabapentin had no effect against PTZ-induced clonic convulsions but dose dependently antagonized tonic convulsions induced by PTZ (Dalby and Nielsen, 1997). Therefore, it should be stressed that profile of test compounds against PTZ-induced seizure depends on seizure type and our results should be compared with those monitoring clonic seizure.

Most AEDs tested in this study showed preferential efficacy in either the MES- or PTZ-induced seizure test, but the potency of YM928 against the MES did not significantly differ from that against PTZ-induced seizure, suggesting that various seizures could be treated by single dosage of YM928. In the STR-induced seizure test, YM928 showed an anticonvulsant effect. All compounds except phenobarbital suppressed seizure induced by STR. YM928 had anticonvulsant effects in all tests examined, suggesting that it is a broadspectrum antiepileptic compound. Effects on the AMPA-induced seizure test were investigated to assess anti-AMPA activity in vivo. YM928 significantly reduced seizure induced by AMPA. Carbamazepine, phenytoin, and lamotrigine, which are not thought to inhibit neuronal transmission by acting selectively at AMPA receptors, also showed marked effects. In the present study, all AEDs were tested 60 min after administration. ED₅₀ values of these AEDs may not show the peak effects.

In our previous study, YM928 blocked AMPA-induced increases in intracellular Ca²⁺ and AMPA-induced inward currents in rat hippocampal cultures. YM928 also blocked kainate-induced toxicity noncompetitively and showed little effect on NMDA- and veratridine-induced increases in intracellular Ca²⁺ in rat hippocampal culture (Ohno et al., 2003). These findings indicate that the mechanism of YM928's anticonvulsant effect is most likely its antagonistic action at AMPA receptors. AMPA receptor antagonists have been reported to show anticonvulsant effects in generalized seizure models. Both quinoxalinedione derivatives such as NBQX (Ikonomidou and Turski, 1997) and YM90K (Shimizu-Sasamata et al., 1996) as well as 2,3-benzodiazepines such as GYKI52466 (Smith et al., 1991; Yamaguchi et al., 1993) and talampanel (Czuczwar et al., 1998) have been shown to be anticonvulsive when administered intraperitoneally. The present results are consistent with these previous studies and reveal that YM928 is well absorbed orally.

The effect of YM928 on motor coordination was examined in the rotarod test. The protective index of YM928 was 3.0, suggesting that YM928 can exert antiepileptic effects with only minor motor disturbances. In contrast, NBQX and GYKI showed anticonvulsant effects in MES and PTZ threshold tests (Löscher and Hönack, 1994) and fixed intensity tests (Yamaguchi et al., 1993) only at doses inducing ataxia or muscle relaxation. Contrary to these previous studies, our present study indicates that AMPA antagonists can exert antiepileptic effects with only minor motor disturbances. Antagonists of the NMDA receptor-channel complex also induce motor disturbances at anticonvulsant doses (Carter, 1994). AMPA antagonists may thus have an advantage in terms of safety profile when studied clinically. To confirm this conclusion, NBQX, GYKI52466, and YM928 should be examined under the same experimental conditions.

The development of tolerance to the anticonvulsant action of some of AEDs has been reported in animal studies (File, 1983; De Sarro et al., 1992; Rundfeldt et al., 1995) and clinical investigations (Oxley, 1986). In the subchronic administration test, YM928 did not alter the threshold of electronic seizure, suggesting that YM928 does not induce tolerance. In this test, Fischer rats were used and the threshold was measured 120 min after the administration of YM928. The plasma concentrations of YM928 at a dose of 10 mg/kg were not significantly changed after 28 days of treatment under these conditions (day 1, 2.59 ± 0.40 μg/ml; day 28, 2.80 ± 0.16 μg/ml; mean ± S.D., n = 4). On this basis, these conditions were used to cancel the effect of pharmacokinetic alteration. Seizure thresholds were monitored 1 and 7 days after repeated administration. Seizure threshold in the YM928-treated group was not significantly different from that of the control group. These results indicate that seizure susceptibility is not changed after the withdrawal of YM928.

The effects of AMPA antagonists on animal models of stroke are well documented. NBQX (Sheardown et al., 1990; Gill et al., 1992), YM90K (Shimizu-Sasamata et al., 1996; Kawasaki-Yatsugi et al., 1998), and GYKI52466 (Le Peillet et al., 1992) have been shown to be neuroprotective in global and focal ischemia models in rats. The AMPA receptor is also thought to be involved in the pathogenesis of amyotrophic lateral sclerosis (Plaitakis, 1990; Rothstein et al., 1993). Recently, it was reported that NBQX ameliorated the oligodendrocyte damage in experimental autoimmune encephalomyelitis, a model of multiple sclerosis (Pitt et al., 2000; Smith et al., 2000). YM928 may be effective in these acute and chronic neurodegenerative disorders.

In summary, oral administration of YM928 showed anticonvulsant actions in the MES-, PTZ-, STR-, and AMPA-induced seizure tests. The broad action of YM928 was similar to that of valproate. These effects were produced without apparent motor impairment, and no diminution of anticonvulsant effect was seen on subchronic administration. The findings of this study show that YM928 has potential value as a novel antiepileptic drug.