Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

Excessive quantities of the excitatory amino acid, glutamate, are released from neuronal synaptic endings as a consequence of anoxia, seizures, and ischemia (Bradford, 1995). Glutamate in turn activates postsynaptic receptor subtypes, one of which is selectively responsive to the glutamate analog, N-methyl-D-aspartic acid (NMDA) (Harris, 1995). NMDA receptor stimulation results in the opening of a receptor-operated Ca⁺⁺ channel (Harris, 1995). Progressive accumulation of intracellular Ca⁺⁺ leads sequentially to seizure initiation, seizure spread (Bradford, 1995), and eventually neuronal death, i.e., the "excitotoxicity hypothesis" advanced by Olney (1978). Therapeutic agents have been sought that are capable of interrupting or preventing the pathological processes associated with an action at the NMDA receptor. For example, MK801 (dizocilpine) and CNS 1102 (aptaganel) possess exceptionally high affinities for the ionic channel subsite, the "phencyclidine or PCP-site", associated with the NMDA receptor. These compounds are effective broad-spectrum anticonvulsants and likewise prevent the pathological sequelae of cerebral ischemia (Chapman et al., 1990; Albers et al., 1992; Meldrum, 1992). However, preclinical and clinical testing of the two agents, especially MK801, revealed narrow separations between effective treatment doses and doses producing severe side effects, namely, impairments in motor function, learning, and memory, as well as abuse liability (Willetts et al., 1990; Rogawski, 1992; Hudzik and Palmer 1995; Grant et al., 1996). On the other hand, low-affinity NMDA receptor antagonists have shown promise for the treatment of stroke (Cregan et al., 1997; Palmer and Hutchison, 1997), Alzheimer's disease (Muller et al., 1995), and epilepsy (Rogawski et al., 1991; Rogawski, 1992; Palmer et al., 1995; Palmer and Hutchison, 1997). Low-affinity compounds exhibited more favorable therapeutic indices (Rogawski, 1992; Hudzik and Palmer 1995; Muller et al., 1995; Grant et al., 1996; Hudzik et al., 1996). Other mechanistic differences between the two classes of uncompetitive NMDA antagonists are that the low-affinity antagonists: 1) exhibited faster off-rate kinetics at the ion channel (Black et al., 1995, 1996; Muller et al., 1995; Subramaniam et al., 1996); 2) acted rapidly to retard the entry of NMDA-triggered [Ca⁺⁺]_in into cultured neurons, whereas high-affinity antagonists exerted a slow but complete block of the NMDA-triggered [Ca⁺⁺]_in response (Black et al., 1995, 1996); 3) possessed differences in regional binding affinity to NMDA receptors (Greene et al., 1996); and 4) displayed less specificity for NMDA receptor subtypes (Monaghan and Larsen, 1997).

The goals of the investigation were to develop a safe effective low-affinity NMDA receptor antagonist for possible clinical use in epilepsy and/or stroke. Desirable characteristics of the selected compound would include: chiral separation into the most favorable isomer, efficacy against animal seizure models, safety, and a predictable, uncomplicated pharmacokinetic/pharmacodynamic profile.

Based on their affinity to displace MK801 from rat synaptosomal membranes, more than 140 active compounds were tested in mice for effectiveness in the maximal electroshock (MES) test. Compounds were sought with favorable efficacy with selections limited to those with the best therapeutic indices, i.e., efficacy versus neurotoxicity ratios. [RS]-AR-R 15035AR ([RS]--phenyl-2-pyridine-ethanamine·2HCl) was identified from the screening paradigm as the most promising lead. Following chiral separation of the racemate and further testing of the enantiomers, the [S]-isomer, AR-R 15896AR (Fig. 1), was found to be superior to the [R]-isomer, AR-R 15895AR, in all of the models tested. AR-R 15896AR was further shown to be active in rodent models of focal and global ischemia, findings that support the ongoing Phase II clinical trials for patients with acute stroke (Palmer et al., 1996; Cregan et al., 1997). The preclinical anticonvulsant, acute safety, pharmacokinetic and pharmacodynamic profiles of AR-R 15896AR are reported here.

View larger version (7K): [in this window] [in a new window]   Fig. 1.   Structure of AR-R 15896AR.

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

All experiments for animal testing were approved by the Institutional Animal Care and Use Committee. The handling of rodents conformed to the guidelines of the U.S. Department of Agriculture through the Animal Welfare Act (Public Law 91-579). Astra Arcus USA is fully accredited by the Association for Assessment and Accreditation of Laboratory Animal Care (AAALAC).

Statistical procedures for the analysis of variance (ANOVA), the Newman Keuls range test, Fieller's theorem for comparing ratios, the Z test for comparisons of proportions of animals seizing in a population, the Mann-Whitney U test for nonparametric, nonsymmetrical data and the Student's two-tailed t test are referenced in Zar (1984). The resultant effective dose₅₀ (ED₅₀) and toxic dose₅₀ (TD₅₀) values, i.e., doses preventing seizures or neural impairment in 50% of the animals, were calculated using the AEL50 Program, which likewise gives the 95% confidence limits (Finney, 1971).

Materials

Mice. Male CF1 mice weighing 20 to 30 g (Harlan Laboratories, Indianapolis, IN) were housed under a 12-h light/dark cycle at 20 to 22°C with a relative humidity of 30 to 50%. Standard food (Purina Rodent Lab Chow 5001) and water were available ad libitum until the time of drug testing. Mice were quarantined in the institutional animal facility for at least 7 days before experiments began.

Rats. Young adult Sprague-Dawley rats weighing approximately 70 g were obtained from Harlan Laboratories. Unless otherwise specified, male rats were used for the majority of experiments. Rats were quarantined for a minimum of 7 days in the animal quarters under conditions similar to those described for mice.

Compounds. [RS]-AR-R 15035AR, [S]-AR-R 15896AR, and [R]-AR-R 15895AR (the suffix "AR" designates the dihydrochloride salt form) were synthesized at Astra Arcus USA. The following were provided as gifts from the manufacturers: felbamate (Carter Wallace, Cranbury, NJ), clonazepam (Roche Laboratories, Nutley, NJ), valproate (Abbott Laboratories, North Chicago, IL) and lamotrigine (GlaxoWellcome, Research Triangle Park, NC). Phencyclidine, flunarizine, strychnine, 4-aminopyridine, pentylenetetrazol, picrotoxin, bicuculline, kainate, N-methyl-DL-aspartate, phenobarbital, carbamazepine and phenytoin were purchased from Sigma. All drug doses were expressed as mg/kg of body weight calculated to the free base form of the compound. Drugs were routinely made up on the day of dosing by dissolution in distilled water or saline. Routes of administration were i.v. , p.o., i.p., or s.c. Dose-response curves consisted of at least five different concentrations of drug, with 10 animals normally used per dose.

Seizure Models

MES in Mice. For the MES test, a drop of saline containing 1% butacaine was applied to the eyes of mice pretreated with test compound or water. The MES was produced bicorneally (50 mA, 60 Hz for 0.2 s) using a constant current shock generator (model 11A; IITC, Landing, NJ). The instrument has a built-in high internal resistance to compensate for current loss through the animal. Seizures consisted of clonic flexion of the hind limb changing into tonic extension in 0.2 to 0.5 s and then into generalized clonic convulsions followed by depression and recovery. Mice were tested at 30 min following compound administration, except for i.v. dosing, for which testing occurred at 15 min (Swinyard and Woodhead, 1982; Palmer et al., 1995).

MES in Rats. At 24 h before testing of anticonvulsant compounds, rats were prescreened for their ability to develop full tonic extension in the MES test. The selection was necessary because approximately 15% of rats do not develop full tonic extension in the MES test. On the evaluation day, the rats received a stimulus of 150 mA (0.2-s duration) (model H electroshock stimulator obtained form Wahlquist Instrument Co., Salt Lake City, UT) delivered via corneal electrodes (eyes premoistened with 1% butacaine in saline). A dose-response curve was generated following i.p., p.o., s.c., or i.v. dosing. Rats were usually tested at 1 h postdose (Palmer et al., 1995). In a separate AR-R 15896AR dose-response test, the rats were sacrificed within 5 min after MES testing and their brains and blood removed and stored frozen until later determination of AR-R 15896AR concentrations.

Chemically Induced Seizures in Mice. Testing for efficacy against chemically induced convulsions occurred at 30 min postdose. Chemical convulsants were administered by various routes at doses that caused convulsions in 97 or 98% of the animals (ED_97,98 values). Following injections of chemical convulsants, mice were placed into individual clear plastic cages and usually observed continuously for 30 min. All anticonvulsants were administered i.p. for the 4-aminopyridine, NMDLA, kainate, pentylenetetrazol, and picrotoxin experiments and p.o. for the bicuculline and strychnine experiments.

4-Aminopyridine Tonic Seizure Test. 4-Aminopyridine (12 mg/kg, s.c.) was administered and mice were observed for the presence or absence of hind limb tonic extension (Yamaguchi and Rogawski, 1992; Cramer et al., 1994).

NMDLA Seizure and Mortality Test. NMDLA was injected (150 mg/kg, i.v.) into mice, and immediately afterward the animals were observed for 5 min for the presence or absence of the characteristic seizures (circling, myoclonic jerks, wild running, and finally severe clonic convulsions with loss of the righting reflex), and for the next 60 min for mortality as described (Czuczwar et al., 1985).

Kainate Seizure and Mortality Test. After injection of kainate (15 mg/kg, i.v.) the mice were observed for 5 min for the presence or absence of the characteristic seizures (clonic convulsions, persistent postural freezing, and/or hind limb tonic extension), which usually occur within 2 to 3 min and 30 min for protection against mortality (Braun and Freed, 1990; Cramer et al., 1994).

Bicuculline Seizure Threshold Test. Bicuculline (2.7 mg/kg, s.c.) was administered. Mice were observed for the presence or absence of the following: 1) clonic seizures, 2) tonic seizures, and 3) mortality (Swinyard and Woodhead, 1982).

Strychnine Seizure Pattern Test. Strychnine (1.2 mg/kg, s.c.) was administered, and mice were observed for the presence or absence of the hind limb extensor component of the characteristic spinal seizures. Measurements were also made regarding the latencies to seizures and mortality (Swinyard and Woodhead, 1982).

Pentylenetetrazol (PTZ) Seizure Threshold Test. PTZ (85 mg/kg, s.c.) was administered, and mice were observed for the presence or absence of clonic seizures. A threshold convulsion was described as an episode of clonic spasms lasting at least 5 s. Absence of this threshold convulsion for 30 min indicated that the test compound possessed the ability to raise the PTZ seizure threshold (Swinyard and Woodhead, 1982).

Picrotoxin Seizure Threshold Test. Picrotoxin (3.15 mg/kg, s.c.) was injected, and the animals were placed in isolated clear plastic cages and observed over the ensuing 45 min for the presence or absence of convulsions (Swinyard and Woodhead, 1982).

Kindled Seizures in Rats

Established Bicorneal Kindled Seizures. Rats were "kindled" by application of subthreshold electrical stimulation (8 mA, 60 Hz for 2 s, dosing two times per day, 5 h apart) applied via corneal electrodes until all animals exhibited grade IV to V seizure activity. All rats were kindled an equal number of times (11 total stimulations) although some demonstrated grade IV convulsions after fewer stimulations (~5 days). Seizures consisting of progressive development of symptoms are graded according to the following criteria established by Racine (1972) and previously used in our laboratory (Palmer et al., 1992, 1995):

• Grade Ifacial movements;
• Grade IIfacial movements with prominent head nodding;
• Grade IIIsymptoms as in II plus raising to the sitting position with mild forelimb clonus;
• Grade IVmarked rearing to a vertical position moving the head from side to side with prominent oral movements, and forelimb clonus; and
• Grade Vsymptoms as in IV progressing to falling and clonic seizures.

Development of bicorneal kindled seizures. Another study was designed to determine whether AR-R 15896AR would retard or prevent the kindling process. Animals were dosed p.o. [once per day (s.i.d.) for 5 days] with test compounds or saline and 1 h later received the first subthreshold stimulus; the second subthreshold stimulus was given 5 h after dosing. Seizure scoring was as described above. After 5 days the animals were granted a 2-day weekend washout (drug free period) followed by a final single stimulation on the following Monday (day 8). The purpose of the washout was to determine (assuming the compound was active) whether the presence of the drug was required to inhibit the seizures or whether the seizure process itself had been prevented (Palmer et al., 1992).

Febrile-Kindled Seizures in Weanling Rats. Rats were dosed p.o. with test agents or appropriate vehicle 1 h before placement in a water bath maintained at 45 ± 0.1°C for 4 min (Palmer et al., 1998). An adjustable perforated base was placed in the bath as a means of hind limb support to prevent forced swimming. The sides of the bath were of sufficient height to avert escape. Animals were kept under constant observation for the 4 min in the water bath or until appearance of symptoms. The time to onset of seizures was recorded. Zero time was defined as the end of the 4 min heat-stress session. For rats seizing while in the water, the remaining time to the end of the 4 min session was expressed as a negative number. A positive time indicated the seizure occurred after removal from the water. Seizure scoring was done according to Racine (1972). At the end of the 4-min period involving immersion or immediately after appearance of seizures, the animals were dried (incubator maintained at 28°C, ~30 min) and returned to the home cages. Groups of rats were tested 2 days per week over a 2-week period for a total of four immersions per rat. Group size was 12, unless otherwise noted. The following parameters of the experiment were analyzed:

1.	body weights taken on the respective test days;
2.	maximum seizure grade or score;
3.	time to seizure onset in seconds;
4.	seizure duration in seconds, and
5.	percent of animals seizing per testing session.

MES Tolerance Test

Mice. Mice were divided into three groups of between 50 and 60 animals each. One group received an p.o. dose of AR-R 15896AR (45 mg/kg, s.i.d.) for 4 days administered in water as vehicle. A second group received water, s.i.d. A third group was not dosed (to evaluate for effects of stress). On day 5 a dose-response curve (doses were 10, 12.5, 15, 20, 25, and 30 mg/kg, n = 10 mice per dose) was run on each group to determine the individual oral ED₅₀ values for protection in the MES test. An increase in the ED₅₀ value in the AR-R 15896AR-treated mice would indicate tolerance following a subchronic course of administration (Palmer et al., 1995). This testing paradigm has been used by us in the past to show tolerance to the low-affinity uncompetitive NMDA antagonist remacemide hydrochloride as well as phenytoin (Palmer and Hutchison, 1997).

Rats. Separate groups of rats (50 per group) received either water or AR-R 15896AR (24 mg/kg, s.i.d., represents 6 × oral ED₅₀ value) for 4 days. On day 5 a dose-response curve (doses were 1, 3, 6, 9, and 12 mg/kg, n = 10 rats per dose) was run on each group at 60 min postdose to determine the resultant ED₅₀ values for protection in the MES test. An increase in the ED₅₀ for protection in the MES test is suggestive of tolerance (Garske et al., 1991).

Proconvulsant Test in Mice

The metrazol seizure threshold test measures the minimal seizure threshold and is used to determine possible proconvulsant properties of compounds (Palmer et al., 1995).

Mice were pretreated with AR-R 15896AR, i.p. and 30 min later received controlled i.v. infusions of metrazol. The selected doses of AR-R 15896AR correspond to the i.p. ED₉₈ value (19 mg/kg) obtained from the MES test and a TD₃ value of 40 mg/kg obtained from the inverted screen test (see below for description of the latter test). To eliminate bias due to stress, the infusion needle with attached tubing was inserted into the tail vein, taped in place, the mouse then placed into a plastic cage (24 cm long × 24 cm wide × 16 cm high), and allowed to move about freely for 30 s. The convulsant solution (0.5% metrazol/0.9% NaCl plus 10 µg/ml heparin sodium) was infused at a constant rate of 0.34 ml/min by means of a Harvard model 901A syringe infusion pump. The time was recorded when the mouse exhibited the first twitch (time to first twitch) and the infusion continued until the mouse experienced 3 s of persistent clonus (time to clonus). The data were expressed as mean time in s ± S.E. to first twitch and to clonus. Comparisons were made to controls using the Student's two-tailed t test. An equal number of controls to experimental animals were run on a given day. Drug-treated animals were alternated with controls during the runs.

Therapeutic Index/Acute Safety

Mice. The inverted screen test was used to determine the acute TD₅₀ value in mice. The TD₅₀ value was the calculated endpoint at which 50% of the mice failed the task (Palmer et al., 1995). The apparatus consisted of six 13-cm square platforms of 0.6-cm wire mesh supported by metal bars, which in turn were mounted on a steel rod. The rod was supported at both ends and was inverted through an arc of 180°. Mice unable to climb to an upright position within 1 min were rated as failures. Testing occurred at 30 min following p.o. drug administration.

Rats. The gangplank escape test was used to determine the acute TD₅₀ following 60 min of p.o. administration of test compounds in rats (Garske et al., 1991). The apparatus consisted of a narrow board 1.25 cm wide × 63 cm long mounted 49 cm above the bench top. At one end there is an entry cubicle (13 × 13 × 15 cm high) open at the top and illuminated by a 52-W bulb suspended immediately overhead (21 cm above the plank). A larger box covering the plank was progressively darkened to an identical, but darkened, escape cubicle at the other end. The rat received a familiarization trial before drug testing. The escape cubicle takes advantage of the natural tendency of rats to seek a darker environment. A rat was rated as impaired if it fell from the plank

Screening for PCP Liability. The method was identical with that used previously (Hudzik et al., 1996). Compounds were administered to rats as multiples of the oral ED₅₀ value obtained from the MES test (5 naïve rats per dose). Animals were then placed in individual shoe box, Plexiglas cages and observed for approximately 5 h for the appearance of any of these five signs: ataxia, circling, head weaving, hyperactivity and retropulsion. Cumulative incidence scores (a maximum of 25, or all five signs in all five rats) were recorded. Phencyclidine was used as a reference standard and run side by side with each compound.

Pharmacokinetics

Dosing and Sampling Schedule. For the p.o. and i.v. pharmacokinetic study, in-dwelling catheters were surgically inserted into the jugular vein of anesthetized (ketamine cocktail) male rats. On the following day, AR-R 15896AR was administered as a single p.o. or i.v. (tail vein) dose of 5 mg/kg. For the i.v.-dosed rats, blood was taken from the jugular vein catheter at 2, 5, 10, 20, and 40 min and 1,2, 3, 4, 6, 8, 10, 12, and 24 h. For the p.o.-dosed rats, blood was taken at 5, 15, 30, 40, 50, min and 1, 2, 3, 4, 6, 8, 10, 12, and 24 h. The blood was placed in tubes containing anticoagulant and the plasma was obtained by low-speed centrifugation. Plasma samples from 10 rats were pooled from each time, and the concentration of AR-R 15896AR determined as described below.

Plasma and Brain Levels. Plasma samples were analyzed by liquid-liquid extraction. For sample analysis, 50 µl of solution containing 500 ng of internal standard (-methyl--phenyl-2-pyridine-ethanamine free base) was added to each tube containing 250 µl of plasma followed by the addition of 40 µl of 1 M NaOH. Mixing was followed by addition of 2 ml of 10% n-butanol/hexane, mixing for 10 min subsequent to 10 min low-speed centrifugation. The organic phase was removed and added to tubes containing 125 µl of 0.1 M HCl followed by mixing, centrifugation (822 × g) and removal of the organic layer. Added to the tubes was 15 µl of 1.0 M NaOH, followed by brief mixing and transfer to an HPLC autosampler. Standards were prepared in plasma from naïve rats at concentrations between 5 and 10,000 ng/ml. The method was specific and the peak height ratios versus concentration were linear over the measured range of standards. Samples were analyzed by HPLC using reverse phase chromatography with a mobile phase of 13.8% acetonitrile, 86% 0.05 M KH₂PO₄, pH 3.75, and 0.2% triethylamine (final pH adjusted to 3.75 with 85% H₃PO₄). The variable wavelength detector was setup at 260 nm with a 250- × 4-mm column (Lichrospher 60 RP-select B 5-µM particle size) heated to 40°C. Plasma-free samples were analyzed directly using similar HPLC conditions. The detection limit for AR-R 15035AR and its isomers is 5.0 ng/ml in plasma. Compound levels are reported as the free base.

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

MES Testing

Mice. The respective ED₅₀ values for [S]-AR-R 15896AR and [RS]-AR-R 15035AR for protection of mice against MES were roughly similar, 5.7 to 8.6 mg/kg (i.v.), to 17.3 to 17.5 mg/kg (p.o.), to 13.0 to 13.4 mg/kg (i.p.), to 9.4 to 8.6 mg/kg (s.c.). In contrast, the [R]-isomer, AR-R 15895AR, was consistently two- to threefold less potent (P < 0.0001; Student's two-tailed t test) when administered i.p., p.o. or i.v., but not following the s.c. route, where potency was only slightly less (P < 0.025) than for the racemate or the [S] isomer (Table 1).

Rats. The ED₅₀ values for [S]-AR-R 15896AR in the rat MES test were similar (2.5-4 mg/kg) with all four routes of drug administration. AR-R 15896AR tended to exhibit greater potency than the racemate, AR-R 15035AR, but none of the ED₅₀ values for the separate routes of administration were significantly different (Student's two-tailed t test). Both [S]-AR-R 15896AR and [RS]-AR-R 15035AR were considerably more potent than [R]-AR-R 15895AR in protecting rats from MES. For example, after p.o. dosing AR-R 15896AR was fivefold more potent (P < 0.0001) than AR-R 15895 (comparable ED₅₀ values were 3.7 versus 20.3 mg/kg, respectively). After i.v. administration both AR-R 15896AR and the racemate were twofold more potent (respective P values < 0.05 and < 0.02) than AR-R 15895AR (comparable ED₅₀ values were 2.6 and 2.7 versus 6.0 mg/kg, respectively). Moreover, following s.c. treatment, potencies for AR-R 15896AR and AR-R 15035AR were from two- to sixfold greater (P values < 0.0001) than for AR-R 15895AR (comparable ED₅₀ values were 2.5 and 5.9 versus 15.0, respectively) (see Table 2).

Mice: Time Course for Protection against MES. After receiving single p.o. doses of the approximate ED₉₈ value of the three compounds (AR-R 15896AR = 45 mg/kg, AR-R 15895AR = 115 mg/kg, and AR-R 15035AR = 50 mg/kg), mice were tested for protection against MES for up to 6 h (Table 3). For statistical comparisons of the proportion of animals seizing between the three compounds, the Z test (P > 0.05) (Zar, 1984) did not indicate marked differences among the respective time course profiles for protection. For all three compounds, maximal protection (100%) was observed from 30 to 90 min after dosing. By 2 h, protection was reduced from 100 to 40% for AR-R 15035AR, from 100 to 70% for AR-R 15896AR, and from 100 to 80% for AR-R 15895AR. At 6 h, protection ranged from 20 to 40% for all three compounds.

Rats: Time Course for Protection against MES. Single doses of approximately three times the oral ED₅₀ value of AR-R 15035AR and its enantiomers were compared side by side in rats for protection against MES over a course of 24 h (Table 4). The percentage of animals protected ranged from 60 to 100% during the initial 4 h, but by 8 h protection dropped to 30% for all three compounds. No protection was evident at 24 h. For comparison of the proportion of animals seizing in a population, the Z test (Zar, 1984) revealed no significant differences in the degree of protection among the three compounds. Similar studies were conducted with AR-R 15896AR alone using i.v., s.c., and i.p. routes of administration. With i.v. (7.5 mg/kg) and i.p. (12 mg/kg) dosing, 90 to 100% protection was observed over 4 h, decreasing to 70% (i.v.) and 50% (i.p.) protection at 6 h. Administration of AR-R 15896AR by the s.c. route yielded the shortest time course of protection: 100% at 30 min, 80 to 60% from 2 to 4 h, and 10% at 6 h. Another study directly compared male and female rats using 12 mg/kg of AR-R 15896AR administered p.o. Protection at 2 h was 100% for males and 90% for females; at 4 h protection was 90% for males and 100% for females; and at 6 h protection was 80% for both sexes. The proportional breakdown by gender of animals seizing was not significant (Z test, see Zar, 1984).

Chemically Induced Seizures in Mice

4-Aminopyridine. The test for efficacy against 4-aminopyridine seizures, following i.p. dosing of AR-R 15896AR and AR-R 15035AR yielded ED₅₀ values in the same range as were demonstrated to be effective against MES (Table 1). Based on the ED₅₀ value, the potency for AR-R 15895AR (32.4 mg/kg) was slightly less than half that for AR-R 15896AR (17.1 mg/kg) (P < 0.0001, Student' two-tailed t test). The differences in ED₅₀ values between AR-R 15895AR and the racemate (ED₅₀ value = 22 mg/kg) were also significant (P < 0.025). Because compounds active in the MES test are effective against 4-aminopyridine-induced seizures (Yamaguchi and Rogawski, 1992), phenytoin was tested as a positive control. The ED₅₀ value for i.p. dosed phenytoin was 7.5 mg/kg [95% confidence limits = 4 to 9].

Bicuculline. The -aminobutyric acid receptor antagonist, bicuculline (Kupferberg, 1989) produced clonic seizures followed by tonic extension and death. [S]-AR-R 15896AR, [RS]-AR-R 15035AR, and [R]-AR-R 15895AR (administered p.o.) were inactive in preventing clonic seizures; however, the compounds did inhibit tonic seizures in the following order of potency: AR-R 15896AR (ED₅₀ value = 9.8 mg/kg) AR-R 15035AR (ED₅₀ value = 29 mg/kg) > AR-R 15895AR (ED₅₀ value = 43.4 mg/kg) (see Table 1). Statistical analyses (Student's two-tailed t test) of the data showed that the inhibitory effects of AR-R 15896AR were more potent (P < 0.0001) when ED₅₀ values were compared to either AR-R 15895AR or AR-R 15035AR. Bicuculline-induced mortality was likewise prevented following pretreatment with AR-R 15896AR (ED₅₀ = 13 mg/kg), AR-R 15035AR (ED₅₀ = 43 mg/kg), and AR-R 15895AR (ED₅₀ = 69.5 mg/kg) (Table 1). The ED₅₀ value for protection versus mortality by AR-R 15896AR (13 mg/kg) was statistically less (P < 0.0001) than the ED₅₀ values for AR-R 15035AR (43 mg/kg) and AR-R 15895AR (69.5 mg/kg). Moreover, the ED₅₀ value for protection against lethality by AR-R 15035AR was significantly less (P < 0.05) than AR-R 15895AR. The positive control for this test was phenobarbital, which inhibited both the clonic and tonic components of bicuculline. The oral ED₅₀ value for phenobarbital was 21 mg/kg [95% confidence limits = 13 to 22].

NMDLA or Kainate. Following administration of either NMDLA or kainate, the majority (>95%) of the seizures occurred within 2 to 3 min. Pretreating mice with AR-R 15035AR and AR-R 15896AR (i.p.) resulted in prevention of NMDLA-induced seizures as well as the subsequent mortality. The respective ED₅₀ values were 42 and 54 mg/kg versus convulsions and 20.7 and 23.3 mg/kg versus mortality (Table 1). AR-R 15895AR was ineffective at the screening dose (75 mg/kg, i.p.). Both AR-R 15896AR and AR-R 15035AR were somewhat effective in suppressing kainate-induced seizures (respective ED₅₀ values = ~50 and 30.6 mg/kg) and the ensuing lethality (respective ED₅₀ values = 29.7 and 27.8 mg/kg). AR-R 15895AR was partially active in the kainate test40% of the animals exhibited no seizures at a dose of 75 mg/kg (i.p.); however, this appeared to be a threshold response because higher doses were less effective. This threshold response was also noticed with AR-R 15896AR (versus convulsions) and AR-R 15035AR (versus mortality) (Table 1). The potent uncompetitive NMDA receptor antagonist, phencyclidine, served as a positive reference standard: ED₅₀ values for prevention of NMDLA-elicited seizures and lethality were 6.3 mg/kg [95% confidence limits = 5 to 8] and 2.8 mg/kg [95% confidence limits = 1.6 to 3.8], respectively.

PTZ, Strychnine, or Picrotoxin. AR-R 15896AR was tested further for possible protection against seizures elicited by PTZ, strychnine, or picrotoxin. At the screening dose, p.o., of 40 mg/kg, mice were not protected from convulsions elicited by either PTZ or picrotoxin. The only protection noticed was after administration of high p.o. doses (50-100 mg/kg) in the strychnine test. The time to tonic hind limb extension induced by strychnine was modestly increased by 52% from 5.9 min in the controls to 9.1 min following pretreatment, p.o., with 75 to 100 mg/kg AR-R 15896AR. Moreover, the time to death was likewise increased by ~100% from 6.7 min in the controls to 13.4 min in the treatment group (P < 0.001, ANOVA followed by Newman Keuls analysis). Oral doses from 25 to 100 mg/kg AR-R 15896AR failed to protect mice from the clonic spinal seizures elicited by strychnine. At relatively high doses, the reference compound, phenobarbital, inhibited strychnine-induced clonic convulsions with an oral ED₅₀ value of 142 mg/kg [95% confidence limits = 124 to165]. With the other two tests, oral efficacies ED₅₀ values for phenobarbital were PTZ = 17 mg/kg [95% confidence limits = 14 to 21] and picrotoxin = 25 mg/kg [95% confidence limits = 17 to 33].

Kindled Seizures in Rats

Established Bicorneal Kindled Seizures. Once seizures were fully kindled, animals were administered, p.o., AR-R 15896AR or phenobarbital on the day of testing using either 6 or 10 times the ED₅₀ value from the MES test. Similarly valproate was administered at 1 or 3 times the ED₅₀ value. (ED₅₀ values for valproate and phenobarbital in the MES test are given in Table 7.) One hour after dosing, the subthreshold electrical stimulus was applied. AR-R 15896AR (24 and 40 mg/kg) and the lower dose of valproate (260 mg/kg) were ineffective in this test (Table 5). Seizures were significantly prevented with the higher dose of valproate (780 mg/kg, Mann-Whitney U test; Zar, 1984) and the two doses of phenobarbital (24 and 45 mg/kg).

Development of Bicorneal Kindled Seizures. Subchronic p.o. administration of AR-R 15896AR (12 or 24 mg/kg for 5 days) did not retard the development of kindled seizures. Valproate, when administered p.o. at either 260 or 780 mg/kg significantly impeded the formation of kindled seizures. The valproate treatment effects from the 2 by 10 ANOVA were: 260 mg/kg  F(1, 14) = 24.56, P = 0.0002; 780 mg/kg  F(1, 14) = 151.6, P = 0.00001. Moreover, at the higher dose, the seizure score after the valproate washout period (days 6 and 7) remained significantly lower than the controls (P < 0.0008) (Fig. 2).

View larger version (27K): [in this window] [in a new window]   Fig. 2.   Effect of AR-R 15896AR or valproate on the development of bicorneal kindling in rats. Animals were dosed orally (s.i.d./day/5 days) and 1 h later received the first subthreshold stimulation (8 mA, 60 Hz for 2 s), the second subthreshold stimulus was given 4 h after dosing. Stimulation was continued (two times per day) for a total of 5 days (10 kindles) at which time a drug free "washout period" was allowed for 2 days and on day 8 (11th kindle) the stimulus was given in the absence of drug. Seizures were scored using the criteria of Racine (1972). The two curves for valproate are significantly different than the control curve (P < 0.0002 for the low dose and P < 0.00001 for the high dose; ANOVA followed by Newman Keuls range test, n = 8 animals/dose).

Kindled Febrile Seizures in Weanling Rats. Seizures generally occurred when the animal was in the water bath; the onset consisted of tremors with the animal curling into a ball at which point the rat was removed from the bath and allowed to recover. In those animals exhibiting seizures, the time to onset was fairly consistent, occurring within ±30 s from the end of the 4 min exposure to the water bath. Seizure duration became more prolonged following repeated exposures to the bath. There were no differences between males (n = 14) or females (n = 10) regarding the time to seizure onset and seizure duration. The effect of heat stress significantly elevated the body temperatures of both females and males to approximately 43°C. This elevated body temperature did not fluctuate significantly following either single or multiple exposures and no differences between heat-stressed males and females were observed.

View larger version (25K): [in this window] [in a new window]   Fig. 3.   Effect of p.o. administered AR-R 15896AR (4 mg/kg) [- -] or MK801 (0.11 mg/kg) [--] or vehicle [--] on heat-stress induced seizures in weanling rats. Measurable end points are influence on body weight, time to seizure onset, maximum seizure grade and seizure duration. Animals (n = 12/dose) were pretreated at 60 min before heat stress, which consisted of exposure to a 45°C water bath for 4 min 2 days/week for 2 weeks (4 exposures per rat). Seizures were scored using the criteria of Racine (1972). Results were compared using an ANOVA followed by the Newman Keuls range test, the asterisk (*) denotes significance from control (P < 0.05). The comparative data for MK801 have been previously published (Palmer et al., 1998).

Proconvulsant Test in Mice

Shortening the latencies to first twitch and to clonus indicate the ability of a compound to manifest proconvulsant properties by lowering the threshold for metrazol seizures. Pretreating mice with AR-R 15896AR (19 or 40 mg/kg, i.p.) did not change the effects of i.v. infusion of metrazol regarding times to first twitch and full clonus (P > 0.05, Student's two-tailed t test) (Table 6).

Tolerance

Mice. There were no differences in the ED₅₀ values from the MES tests when comparisons were made among the three different treatment groups of mice evaluated on day 5 for possible indications of tolerance to AR-R 15896AR (Student's two-tailed t test). The resultant ED₅₀ values in mg/kg (in brackets below, 95% confidence limits), slopes ± S.E. of the dose-response curves were:

• Group I (naïve mice no treatment)14.4 [11 to 19], 4.7 ± 1.2;
• Group II (saline p.o. for 4 days)16.9 [14 to 20], 6.1 ± 1.5;
• Group III (AR-R 15896AR p.o. for 4 days)18.9 [16 to 22], 7.0 ± 1.7

Rats. Following p.o. administration of AR-R 15896AR daily for 4 days at 6 times the ED₅₀ value, there was a nonsignificant trend (P = 0.1; Student's two-tailed t test,) for the ED₅₀ value in the MES test to increase on day 5. The resultant ED₅₀ values in mg/kg (in brackets below, 95% confidence limits), slopes ± S.E. were:

1.	Controls (saline daily 4 days)3.4 [2 to 5], 2.2 ± 0.6;
2.	AR-R 15896AR (24 mg/kg/daily 4 days)6.6 [4 to 10], 2.6 ± 0.8

Therapeutic Index/Acute Safety

Mice. The p.o. toxic dose-50 (TD₅₀) values and i.v. median estimated lethal dose (MELD) values for AR-R 15896AR, AR-R 15895AR, and AR-R 15035AR, from the inverted screen and lethality tests, respectively, in mice fell within the overlapping 95% confidence limits of one another and therefore did not differ significantly (comparisons by the Student's two-tailed t test) (see Table 7). Respective TD₅₀ values in mg/kg were: AR-R 15896AR = 193.3 [95% confidence limits = 145 to 235], AR-R 15895AR = 142 [95% confidence limits = 80 to 226], and AR-R 15035AR [95% confidence limits =139 to 359]. The therapeutic indices, i.e., TD₅₀/ED₅₀ from the MES test were greater for AR-R 15035AR and AR-R 15896AR than for AR-R 15895AR (TI values of 14, 11.2, and 2.5, respectively). However, statistical comparisons of the TI values using Fieller's Theorem (Zar, 1984) revealed no significant differences between ratios. Table 7 also provides comparative data for clinically useful compounds that have been evaluated at Astra Arcus USA. For example, p.o. administered lamotrigine is the most potent compound regarding protection of mice in the MES test (ED₅₀ value = 6.1 mg/kg) and valproate is the least potent (ED₅₀ value = 631 mg/kg). The TI values for felbamate, lamotrigine, phenytoin, flunarizine, AR-R 15035AR and AR-R 15896AR ranged from 10 to 22. The compounds with the lowest TI values in mice (range 2-5) were AR-R 15895AR, valproate and phenobarbital.

Rats. TD₅₀ values for rats following single p.o. dosing regimens were determined from the gangplank escape test. At 1 h after dosing, the TD₅₀ values for AR-R 15035AR and AR-R 15895AR were greater than 1000 mg/kg, whereas that for AR-R 15896AR was 932.5 mg/kg [95% confidence limits 712 to 2361]. The resultant TI value for AR-R 15896AR was 252, whereas the estimated TI values for AR-R 15035AR and AR-R 15895AR were >100 and >49.5, respectively (Table 7). Comparative data for clinically useful drugs are also included in Table 7. With the exception of valproate, the therapeutic indices for these reference compounds were more favorable in rats than in mice.

PCP-like Actions. The three compounds were administered to rats in p.o. doses as multiples of the ED₅₀ values from the MES test. Doses 20 times (AR-R 15035AR), 30 times (AR-R 15895AR), and 50 times (Hudzik et al., 1996, for previously published findings for AR-R 15896AR) the respective ED₅₀ values produced no evidence of the five characteristic PCP-like behaviors (PCP behavioral scores were 0 out of a possible 25). Phencyclidine-dosed rats readily produced progressive behavioral signs: 1 × ED₅₀ value (score = 3 of 25); at 3 × ED₅₀ value (score = 8 of 25); at 6 × ED₅₀ value (score = 16 of 25); and at 10 × ED₅₀ value (score = 25 of 25). The previously determined ED₅₀ value for PCP in the MES test was 1.24 mg/kg, p.o. (Hudzik et al., 1996).

Pharmacokinetics of AR-R 15896AR

Following an i.v. bolus dose (5 mg/kg) in the rat, the disposition of AR-R 15896AR was best described by a two-compartment model with a distributive phase that was complete within 2 h postdose (Fig. 4). The terminal half-life (t_1/2) was estimated to be 4.75 h with a plasma clearance of 0.64 liters/h/kg (Table 8). The compound was distributed into a moderate volume (V₁ = 1.89 liters/kg and V_dss = 3.97 liters/kg) indicating significant tissue disposition. In rat, the oral bioavailability following 5 mg/kg AR-R 15896AR was 83% with a C_max and t_max of 720 ng/ml and 1.05 h, respectively (Fig. 4, Table 8).

View larger version (9K): [in this window] [in a new window]   Fig. 4.   The plasma kinetics of AR-R 15896AR in the rat after p.o. and i.v. dosing of 5 mg/kg. Plasma samples from 10 rats were pooled for each time point.

Following an i.v. bolus, AR-R 15896AR entered the brain rapidly and was in equilibrium with the plasma within 5 minutes after dosing (Fig. 5). The ratio of brain to plasma concentration (i.v. brain uptake index = 109%) and brain to plasma-free concentration remained constant over time.

View larger version (11K): [in this window] [in a new window]   Fig. 5.   The brain, plasma, and plasma-free kinetics of AR-R 15896AR in the rat after i.v. dosing of 10 mg/kg into the tail vein. The values are the mean concentration of compound ± S.E. from 3 rats per time point.

In order to determine whether efficacy of AR-R 15896AR in the MES test was directly correlated with a predictable pharmacokinetic profile and whether differences between i.v. or p.o. dosing would be evident, a separate study was performed. Rats received either p.o. or i.v. injections of AR-R 15896AR and were evaluated for protection in the MES test at 1 h or 3 h, respectively. The 1-h time point was selected for the p.o.-dosed rats because this time is near the t_max for AR-R 15896AR. For the rats dosed i.v., the 3-h time point was picked for testing because at this dose the plasma concentrations approximated the concentrations of AR-R 15896AR after an equal p.o. dose. The resultant concentrations of AR-R 15896AR in plasma and brain were proportional to the dose administered (Fig. 6). The dose-response curves were similar for both the i.v. and p.o. routes of administration. The calculated ED₅₀ values were 3.3 mg/kg [95% confidence limits = 2.1 to 4.4] for p.o. dosing and 3.9 mg/kg [95% confidence limits = 1.7 to 6.2] for i.v. dosing (Fig. 7, top). In addition, the brain concentrations of drug over the range of doses administered were similar for both the i.v. and p.o. dosed animals (Fig. 7, bottom). The EC₅₀ value for MES protection based on total brain concentration was approximately 25 µM.

View larger version (17K): [in this window] [in a new window]   Fig. 6.   Brain and plasma levels of AR-R 15896AR in rats sacrificed within 5 min after being tested for protection against MES. MES was administered 3 h after i.v. (top) and 1 h after p.o. (bottom) dosing (values are the mean ± S.E.; n = 10).

View larger version (17K): [in this window] [in a new window]   Fig. 7.   Dose-response data for male rats for protection against MES-induced seizures (n = 10 rats/data point). The rats were tested 1 h after p.o. or 3 h after i.v. dosing. The calculated ED50 values were 3.3 mg/kg [95% confidence limits = 2.1 to 4.4] for p.o. and 3.9 mg/kg [95% confidence limits = 1.7 to 6.2] for i.v.-dosed animals (top). The EC50 value for MES protection based on total brain concentration was approximately 25 µM (bottom).

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

The present investigations with the low-affinity uncompetitive NMDA receptor antagonist, [RS]-AR-R 15035AR, described the acute preclinical in vivo studies using animal models of seizures and safety, as well as pharmacodynamic profiling that contributed to the identification and selection of the [S] enantiomer, AR-R 15896AR, over the [R]-enantiomer, AR-R 15895AR, as the most favorable candidate for further consideration as an antiepileptic drug.

Anti-seizure Efficacy. Uncompetitive NMDA receptor antagonists are potent anticonvulsants, especially in the MES and NMDLA seizure tests (Czuczwar et al., 1985; Chapman et al., 1990; Rogawski et al., 1991; Rogawski, 1992; Bradford, 1995). These testing paradigms were used as rapid screens to evaluate efficacy of potential compounds previously identified from their respective receptor ligand binding profiles (displacement of labeled MK801) in rat synaptosomal preparations. The ability of AR-R 15035AR and respective isomers to prevent tonic hind limb extension in rodent seizure tests as a consequence of MES application or injections of 4-aminopyridine, bicuculline, and, to a lesser extent, strychnine, indicated the compounds act in part to suppress seizure spread. This phenomenon has been demonstrated with clinically useful drugs, namely, phenytoin, carbamazepine, phenobarbital, valproate, felbamate, lamotrigine, and flunarizine (Swinyard and Woodhead, 1982; Loscher and Schmidt, 1988; Kupferberg, 1989; see Table 7). The potency and acute safety profile for AR-R 15896AR, compared favorably to compounds currently in the clinic (comparisons of data generated at Astra Arcus USA are given in Table 7; see also Swinyard and Woodhead, 1982; Loscher and Schmidt, 1988; Kupferberg, 1989).

Pharmacokinetics. Predictable pharmacokinetic profiles are a desirable characteristic of a clinically useful anticonvulsant (Kupferberg, 1989). The pharmacokinetic evaluation of AR-R 15896AR did indeed reveal a favorable pharmacokinetic profile. Linear plasma levels of the compound were rapidly attained following either i.v. or p.o. administration in rats (Fig. 4). In fact, levels of AR-R 15896AR in rat brains exceeded that of plasma, whereas the ratio of brain-to-plasma concentration and brain-to-plasma-free concentration remained constant over time (Figs. 5 and 6). Moreover, there was a direct correlation between brain levels and efficacy of AR-R 15896AR in the MES test following either i.v. or p.o. dosing (Figs. 6 and 7), a result indicating a paucity of barrier mechanisms (absorption, distribution, excretion, first pass metabolism, etc.) limiting the bioavailability of p.o. administered AR-R 15896AR.

NMDA Receptor-mediated Events. The inhibition by AR-R 15896AR and its racemate, AR-R 15035AR, of NMDLA- and to a lesser extent kainate-elicited convulsions/subsequent mortality, provided a functional link to the original in vitro receptor ligand binding parameters for the compounds (Black et al., 1995). [RS]-AR-R 15035AR, [S]-AR-R 15896AR, and [R]-AR-R 15895AR all exhibited low affinities at the ion channel site of the NMDA receptor with K_i values of 3.5, 1.3, and 6.9 µM, respectively, for displacement of radiolabeled MK801 (MK801 K_i = 0.014 µM) (Black et al., 1995; see Palmer et al., 1992 for MK801 data).

Kindling. Glutamate is responsible for at least three phenomena associated with experimental epilepsy, namely, seizure initiation, seizure spread, and seizure kindling (Bradford, 1995). According to Bradford (1995), there appears to be a chronic tendency to an exaggerated release of glutamate occurring in the kindled epileptic focus during both epileptogenesis and full seizure expression. In general, uncompetitive NMDA receptor antagonists, including AR-R 15896AR (present study), are not markedly effective regarding inhibition of fully established kindled seizures but have been demonstrated to retard the development of seizures during the kindling process, an action also shared with valproate (present study), and clonazepam (McNamara et al., 1988; Rogawski et al., 1991; Palmer et al., 1992, 1995, 1998; Applegate et al., 1997). Even though both doses of valproate retarded the development of the kindling process, it was more active at the highest concentration representing three times the ED₅₀ value for efficacy in the MES test; and at this dose, side effects were evident. Moreover, at this dose alone valproate inhibited established kindled seizures, whereas phenobarbital was highly effective at the 3 × ED₅₀ dose. It's unlikely that AR-R 15896AR would be effective at doses higher than 40 mg/kg (10 times the ED₅₀ value) because in an earlier study with febrile convulsions in rats, the low-affinity uncompetitive NMDA receptor antagonist, remacemide hydrochloride, was ineffective at doses as high as 24 times the ED₅₀ value for the MES test (Palmer et al., 1998).

Safety Issues. Safety as been the major problem with development of effective therapy to treat epilepsy (Kupferberg, 1989). Moreover, in the present study AR-R 15896AR exhibited a remarkable safety profile in the rat when compared with clinically useful compounds (see Table 7). Clinical trials for epilepsy using higher affinity uncompetitive NMDA antagonists, MK801, and dextromethorphan were discontinued because of unacceptable psychotomimetic effects (Willetts et al., 1990; Rogawski, 1992; Muller et al., 1995; Grant et al., 1996). Neither AR-R 15035AR nor its respective [R]- and [S]-enantiomers, at doses an order of magnitude or higher than their respective MES ED₅₀ values, produced acute PCP-like behaviors in rats. In addition, AR-R 15896AR was not proconvulsant (present study) and had no effect on learning acquisition (Hudzik and Palmer, 1995) or motor behaviors (Palmer et al., 1996). In contrast to PCP and cocaine, rats did not self-administer AR-R 15896AR (Hudzik et al., 1996). In the present evaluations for acute safety, the therapeutic indices for AR-R 15896A