Anticonvulsant Effects of the Glycine/NMDA

Home , Glycine receptor agonist

Br. J. Pharmacol. (1994), 112, 97-106 '"I Macmillan Press Ltd, 1994 Anticonvulsant effects of the glycine/NMDA receptor ligands D-cycloserine and D-serine but not R-(+)-HA-966 in amygdala-kindled rats 'Wolfgang Loscher, 2Piotr Wlaz, Chris Rundfeldt, Halina Baran & Dagmar Honack Department of Pharmacology, Toxicology, and Pharmacy, School of Veterinary Medicine, Binteweg 17, 30559 Hannover, Germany 1 The effects of the glycine/NMDA receptor partial agonists, D-cyCloserine and (+)-HA-966 and the full agonist, D-serine, on focal seizure threshold and behaviour have been determined in amygdala- kindled rats, i.e. a model of focal (partial) epilepsy. The uncompetitive NMDA receptor antagonist, MK-801, was used for comparison. 2 The high efficacy glycine partial agonist, D-cycloserine, did not alter the threshold for induction of amygdaloid afterdischarges (ADT) at doses of 20-80 mg kg-' i.p., but significant ADT increases were determined after application of higher doses (160 and 320 mg kg-'). The ADT increases after these high doses were long-lasting; significant elevations were still observed 2 days after drug injection. Determina- tion of D-cycloserine in plasma and brain tissue showed that it was rapidly eliminated from plasma. Compared to peak levels in plasma, only relatively low concentrations of D-cycloserine were measured in brain tissue. 3 The low efficacy glycine partial agonist, (+)-HA-966, 10-40 mg kg-l i.p., did not alter the ADT or seizure recordings (seizure severity, seizure duration, afterdischarge duration) at ADT currents. How- ever, the drug dose-dependently increased the duration of postictal behavioural and electroencephalographic depression in kindled rats. At the higher dose tested, postictal immobilization was dramatically increased from 3 min to about 120 min. This might indicate that glutamatergic activity is decreased postictally, which is potentiated or prolonged by (+)-HA-966. 4 Like D-cycloserine, the glycine receptor full agonist, D-serine, injected bilaterally into the lateral ventricles at a dose of 5 pmol, significantly increased the ADT, while no effect was seen at a lower dose (2.5 pmol). 5 The anticonvulsant effects observed with D-cycloserine were completely antagonized by combined treatment with (+)-HA-966, indicating that the effects of D-cycloserine were mediated by the glycine/ NMDA receptor complex. 6 MK-801, 0.1 mg kg-', did not alter the focal seizure threshold or seizure recordings at ADT current, but induced marked phencycidine(PCP)-like behavioural alterations, such as hyperlocomotion, stereotypies and motor impairment. No PCP-like behaviours were observed after D-cycloserine, D-serine or (+)-HA-966. High doses of (+)-HA-966 induced moderate motor impairment in kindled rats. 7 The long lasting increases in seizure threshold observed after the high efficacy glycine partial agonist, D-cycloserine but not the low efficacy partial agonist, (+)-HA-066, may suggest that the effects of D-cycloserine are mediated by adaptive changes in the NMDA receptor complex in response to glycine receptor stimulation. 8 Pharmacological intervention at the strychnine-insensitive glycine receptor by high-efficacy partial agonists with systemic bioavailability may be an effective means of increasing seizure-threshold without concomitantly inducing PCP-like adverse effects. Keywords: Glycine receptors; NMDA receptors; glutamate; epilepsy; MK-801; dizocilpine; stereotypies; locomotor activity; motor impairment; postictal depression

Introduction The strychnine-insensitive glycine site of the N-methyl-D- advantage of glycine antagonists in this respect might be that aspartate (NMDA) receptor complex has generated an enor- they have a larger therapeutic window than antagonists of mous amount of interest since it was first described seven years the NMDA receptor and associated ion channel (Carter, ago by Johnson & Ascher (1987). It is now clear that oocupa- 1992). Indeed, glycine antagonists, such as 7-chlorokynurenic tion of the glycine site by an agonist is an absolute require- acid, were shown to antagonize or attenuate convulsions in ment for NMDA receptor activation and that glycine is, in different seizure models, including seizures induced by effect, a NMDA receptor co-agonist (Kemp & Leeson, 1993). NMDA, audiogenic stimulation, electroshock or amygdala- Because of this crucial role of the glycine site for NMDA kindling, without producing phencycidine(PCP)-like receptor activation, glycine receptor antagonists are thought behavioural adverse effects (Croucher & Bradford, 1990; to have several therapeutic indications for diseases, including Koek & Colpaert, 1990; Tricklebank & Saywell, 1990; cerebral ischaemia and epilepsy, in which glutamatergic Croucher & Bradford, 1991; Baron et al., 1992). Based on overactivity is thought to be involved (Carter, 1992). One the role of NMDA receptors in the initiation and propaga- tion of seizures (Dingledine et al., 1990; LUscher, 1993), the NMDA potentiation by glycine and the anticonvulsant 'Author for correspondence. activity of NMDA and glycine antagonists, it has been pro- 2Permanent address: Department of Pharmacology, Veterinary posed that glycine or glycine agonists such as D-serine are Faculty, School of Agriculture, Lublin, Poland. proconvulsant (Foster & Kemp, 1989). However, in contrast 98 W. LOSCHER et al. to that proposal, several reports have shown that 1-amino-I- nula placements was made by localization of a dye which had carboxycyclopropane, which is almost a full agonist (up to been infused into the lateral ventricles. 90% efficacy of glycine), D-cycloserine, which has about 40-70% efficacy of glycine, and glycine itself have anticon- Kindling vulsant activity when administered to animals in high doses (Wood et al., 1966; Mayer et al., 1971; Lapin, 1981; Toth et After a postoperative period of 2 weeks, constant current al., 1983; Peterson & Boehnke, 1989; Skolnick et al., 1989; stimulations (500 IiA, 1 ms, monophasic square-wave pulses, Stark et al., 1990; Tricklebank & Saywell, 1990; Peterson, 50 s-' for 1 s) were delivered to the amygdala at intervals of & Schwade, 1993). Furthermore, glycine and D-serine were 1 day until 10 sequential stage 5 seizures were elicited. found to potentiate the anticonvulsant effect of clinically Seizure severity was classified according to Racine (1972): (1) established antiepileptic drugs, while L-serine was ineffective immobility, eye closure, twitching of vibrissae, sniffing, facial (Peterson, 1991). Since these data conflict with the current clonus; (2) head nodding associated with more severe facial theory of the role of glutamate receptors in seizures and clonus; (3) clonus of one forelimb; (4) rearing, often accom- anticonvulsant drug actions, it has been proposed that the panied by bilateral forelimb clonus; (5) rearing with loss of anticonvulsant effect of glycine agonists and high efficacy balance and falling accompanied by generalized clonic partial agonist cannot be explained entirely by their actions seizures. After kindling acquisition, the electrical suscep- at the glycine site of the NMDA receptor (Kemp & Leeson, tibility of the stimulated region for triggering of paroxysmal 1993). However, Peterson (1992) demonstrated that the neuronal activity (threshold for induction of afterdischarges, anticonvulsant effect of D-cycloserine in the maximal elect- ADT) was recorded using an ascending step procedure roshock seizure (MES) test in rats could be antagonized by (Freeman & Jarvis, 1981). The initial current intensity was the glycine antagonist, 7-chlorokynurenic acid. 10 ,A, and the current intensity was increased in steps of In the present study, we were interested to examine if about 20% of the previous current at intervals of 1 min until D-cycloserine also exerts anticonvulsant effects in amygdala- an afterdischarge of at least 3 s duration was elicited. Deter- kindled rats, i.e. a widely used model of complex-partial mination of ADTs was repeated at intervals of 2-3 days seizures with secondary generalization to clonic convulsions. until all animals exhibited reproducible seizure thresholds. For comparison with D-cycloserine, we used the uncom- Since almost all fully kindled animals exhibited generalized petitive NMDA receptor antagonist, MK-801 (dizocilpine), seizures (stage 4-5) at the ADT current, it was not necessary the full glycine receptor agonist, D-serine, and the low to determine the threshold for generalized seizures (GST) efficacy partial glycine agonist (+)-HA-966 (R-(+)-3-amino- separately. In fully kindled rats, the ADT determined with I-hydroxypyrrolid-2-one), which has less than 8% of the interstimulation intervals of 1 day was not different from agonist, glycine (Henderson et al., ADT values determined with interstimulation intervals of efficacy of the endogenous 1 min, thus demonstrating that the short interstimulation 1990). interval did not bias ADT determinations. In addition to ADT, in fully-kindled rats the following parameters of kindled seizures were measured at stimulation with the ADT Methods current: seizure severity was classified as described above. Seizure duration was the duration of the limbic (stage 1-2) Animals and motor seizures (stage 3-5). Afterdischarge duration was the total time of spikes (with an amplitude of at least twice Female Wistar rats (Harlan-Winkelmann, Borchen, Ger- the amplitude of the prestimulus recording and a frequency many), weighing 210-230 g, were used. The animals were greater than 1 s-') in the EEG recorded from the site of purchased from the breeder at a body weight of about 200 g. stimulation. Furthermore, the period of postictal immobiliza- Following arrival in the animal colony, the rats were kept tion ('freezing'; defined as the period of immobility between under controlled environmental conditions (ambient tem- the end of the afterdischarge until the onset of locomotion) perature 24-25°C, humidity 50-60%, 12/12 h light/dark was recorded. cycle, light on at 06 h 00 min) for at least 1 week before being used in the experiments. Standard laboratory chow Evaluation of drug effects on focal seizure threshold (Altromin 1324 standard diet) and tap water were allowed ad lib. All experiments were done between 08 h 00 min and The effects of MK-801, D-cycloserine, and (+)-HA-966 were 12hOOmin. compared in groups of 6-10 fully kindled rats by determination of the ADT, i.e. the most sensitive measure of anticon- Preparation of animals vulsant effects on focal seizure activity in kindled rats. The drugs were injected i.p. 1 h (D-cycloserine), 2.5 h (MK-801) The rats were anaesthetized with chloral hydrate (360 mg or 0.5 h ((+)-HA-966) prior to amygdala stimulation. The kg-', i.p.) and received stereotaxic implantation (according effect of D-serine on ADT was determined 60 min after to the surgery methods described in the atlas of Paxinos & bilateral i.c.v. administration of 2.5 or 5 fmol (half of these Watson, 1986) of one bipolar electrode in the right doses injected per site). These doses of D-serine were infused basolateral amygdala. Coordinates for electrode implantation i.c.v. in a volume of 4 fl (i.e. 2 pl per site) over a period of were AP -2.2, L -4.8, V -8.5. All coordinates were 1 min per site. The injection needle was left in place for an measured from bregma. Skull screws served as the indifferent additional 1-min period at the end of the infusion to allow reference electrode. The electrode assembly was attached to diffusion of the solution into the ventricles. Dosages and the skull by dental acrylic cement. In one group of rats, pretreatment times were chosen on the basis of previous intracerebroventricular (i.c.v.) injection guide cannulae were dose-effect and time-course experiments with the drugs in implanted in addition to the bipolar amygdala electrode. In rats (Loscher & Honack, 1991a; Vartanian & Taylor, 1991; each of these animals, two 0.6 mm guide cannulae were Peterson, 1992). The control ADT was determined 2-3 days bilaterally implanted 0.5 mm dorsal to the left and right before and after each drug treatment, and the next drug lateral ventricle. Coordinates for guide cannula implantation experiment in the same group of rats was only undertaken if were AP -0.8, L ± 1.5, V -2.9. A 0.35 mm stylet was the post-drug ADT was not significantly different from the placed in the cannulae to prevent clogging when not in use. pre-drug ADT. For control determinations, rats received i.p. The guide cannulae were secured together with the electrode or i.c.v. injection of vehicle (saline) with the same pretreat- assembly with dental acrylic cement. For i.c.v. drug injection, ment time used for drug testing. For all drug experiments, at a 0.35 mm injection needle that extended 1 mm beyond the least 4 days were interposed between 2 drug injections in the dorsal end of the guide cannula was inserted into each guide same group of rats in order to avoid alterations in drug cannula. In preliminary experiments, verification of the can- potency due to cumulation or tolerance. D-CYCLOSERINE AND (+)-HA-966 IN THE KINDLING MODEL gg

Evaluation of behavioural effects seizure readings in the same group of rats was calculated by the Wilcoxon signed-rank test for paired replicates. Behavioural alterations after administration of test drugs were determined at different times after injection of each Drugs compound up to 2 min before amygdala stimulation. For all observations, rigorous observational protocols described D-Cycloserine and D-serine were purchased from Sigma elsewhere were used (Loscher & Honack, 1991b; 1992). (Munich, Germany). (+ )-HA-966 (R-( + )-3-amino-l-hyd- Hyperlocomotion, head weaving (swaying movements of the roxypyrrolid-2-one) was generously supplied by Merz & Co. head and upper torso from side to side for at least one (Frankfurt/M, Germany) and MK-801 (dizocilpine maleate) complete cycle; i.e. left-right-left), stereotyped sniffing, biting, by Merck Sharp & Dohme (Rahway, NJ, U.S.A.). All drugs licking or grooming, reciprocal forepaw treading ('piano were freshly dissolved in saline before each experiment and playing'), stereotyped rearing, hyperexcitability (as indicated were injected i.p. in a volume of 2-3 ml kg-' or i.c.v. in a by increased reactions to noise or handling), tremor, abduc- volume of 4y1L. tion of hind limbs, reduction of righting reflexes, flat body posture, circling, Straub tail and piloerection were scored using a ranked intensity scale where 0 = absent, 1 = Results equivocal, 2 = present and 3 = intense. Behavioural alterations other than those decsribed above were recorded Effect of MK-801, D-cycloserine, (+ )-HA-966, and separately. Ataxia was quantitated by a rating system as D-serine on thefocal seizure threshold (ADT) in described recently (Loscher et al., 1987). In short, animals amygdala-kindled rats were taken out of the cage, placed in an open field, observed for about 1 min and ataxia was rated as follows: (1) slight The uncompetitive NMDA receptor antagonist, MK-801, ataxia in hind-legs (tottering of the hind quarters); (2) more 0.1 mg kg-', i.p., did not alter the ADT of kindled rats pronounced ataxia with dragging of hind legs; (3) further (Figure 1) or seizure severity, duration of seizures and after- increase of ataxia and more pronounced dragging of hind discharges recorded at ADT currents (Table 1). The partial legs; (4) marked ataxia, animals lose balance during forward glycine agonist, D-cycloserine, did not significantly alter the locomotion; (5) very marked ataxia with frequent loss of ADT at doses of 20, 40 or 80 mg kg-' (Figure 1). At balance during forward locomotion; (6) permanent loss of 80 mg kg-', only 1 of 9 kindled rats (not included in the righting reflexes, but animal still attempts to move forward. In addition to rating of motor impairment in the open field, impaired motor function was quantitated by the rotarod test of Dunham & Miya (1957). The rotarod test was carried out 240- with a foam rubber coated rod of 6 cm diameter which rotated at 8 r.p.m. Neurological deficit was indicated by inability of the animals to maintain their equilibrium for at 220- least 1 min on the rotating rod. The kindled rats were trained prior to drug experiments to remain on the rod. After drug 200- treatment, rats which were not able to maintain their equilibrium on the rod for I min were put again on the rod a 0 180- 1 I-, further two times. Only animals which were not able to L.) remain on the rod during 3 subsequent 1 min trials were V 160- considered to exhibit deficit. For time course - I neurological -c 140- studies, the rotarod test was performed at different times 0) after drug injection immediately after rating of adverse effects Co in the open field. 120- I I I" C.)a) I -C) I T Determination ofD-cycloserine in plasma and brain 4- -1 In two groups of 6 rats each (non-kindled but age-matched 80 - with the kindled rats), D-cycloserine was injected at a dose of -C 320 mg kg-'. One group of rats was killed after 15 min, the 60- other after 60 min. A third group of rats received i.p. saline and served as control. Blood was sampled for preparation of plasma. The brains were rapidly removed and samples of the 40- frontal cortex were weighed and homogenized by an Ultra- Turrax homogenizer in 2 ml of ice-cold 80% ethanol. Further 20- processing of plasma and brain samples was as described recently (Rundfeldt & Loscher, 1992; Loscher et al., 1993). MK DCS DCS DCS DCS DCS HA HA HA DS DS was determined in and brain i"i D-Cycloserine HIiL0.1 20 40 80 160 10 plasma by high 320 20 40 2.5 5 performance liquid chromatography (h.p.l.c.) after precolumn derivatization by means of o-phthaldialdehyde/2-mercapto- Figure 1 Effect of MK-801 (MK), D-cycloserine (DCS), D-serine ethanol essentially as described recently for endogenous (DS) or (+)-HA-966 (HA) on focal seizure threshold (ADT) deter- amino acids (Rundfeldt & Ldscher, 1992; Loscher et al., mined by electrical stimulation of the amygdala in amygdala-kindled 1993). rats. Data are from experiments with groups of 8-11 fully kindled rats (except 40 mg kg-' DCS, which is from a group of 6 animals). The doses of drugs are shown below each column in mg kg-i, i.p. Statistics (for DS in gmol, i.c.v.). Control ADTs (with i.p. or i.c.v. injection of 2-3 before All data are as means ± s.e.mean. In some saline) were determined days each drug experiment; given experiments, average pre-drug control ADT was 27.8 ± 1.6 pA (mean ± s.e.mean single animals responded differently from the other animals of 11 pre-drug control threshold determinations in groups of 6-11 of the group. In order to maintain normal distribution of rats). Drug data are shown as mean percentage (+s.e.mean) of the data, such outliers were not involved in presentation of individual pre-drug control threshold. Significant differences to pre- group means and deviation, but data from such animals are drug control are indicated by asterisks (*P<0.05; **P<0.01). mentioned in the text. Significance of differences between Pretreatment times were 2.5 h (MK), I h (DCS, DS) or 0.5 h (HA). 100 W. LOSCHER et al.

Table 1 Effects of MK-801, D-cycloserine, (+)-HA-966, D-serine, and a combination of D-cycloserine and (+)-HA-966 on ictal and postictal recordings at ADT currents in kindled rats Seizure recordings at ADT Afterdischarge Postictal Seizure severity Seizure duration duration immobilization Drug Dose score sec sec sec Control 4.8 0.1 43.2 7.2 78.0 ± 16.4 72.3 ± 13.8 MK-801 0.1 mg kg-' i.p. 4.7 0.2 44.7 ± 6.2 62.2 ± 16.7 64.4 14.3 Control 4.8 0.1 65.0 6.8 121.7 ± 13.1 113.9 23.2 D-Cycloserine 20mgkg-' i.p. 4.7 0.2 54.7 5.9 95.9 ± 19.7 58.9 ± 14.2* Control 4.2 ± 0.7 58.5 10.9 89.2 ± 20.5 56.0 13.2 D-Cycloserine 40mgkg-' i.p. 4.5 0.2 51.5 ±6.4 90.2 ± 17.9 66.3 + 18.2 Control 4.9 0.1 46.7 6.6 77.9 ± 11.3 166.7 i 32.5 D-Cycloserine 80 mg kg-' i.p. 4.4 0.4 46.7 6.1 85.5 ± 12.4 186.0 ± 39.9 Control 4.9 0.1 52.6 5.4 84.2 ± 11.7 142.2 26.1 D-Cycloserine 160mgkg-Ii.p. 4.2 0.4 50.4 ± 5.8 92.3 ± 19.5 174.9 ± 51.4 Control 4.9 0.1 55.8 ± 4.2 86.2 ± 10.0 172.6 27.9 D-Cycloserine 320 mg kg-' i.p. 4.3 0.3 48.1 ± 5.7 106.3 ± 28.1 196.8 34.8 Control 4.6 0.3 50.9 4.7 81.2 ± 10.3 218.2 ± 39.2 (+)-HA-966 l0mgkg-' i.p. 5 46.9 4.9 67.9 ± 9.8 511.3 235.4 Control 4.9 0.1 45.3 ± 4.6 79.0 ± 12.8 143.2 ± 26.9 (+ )-HA-966 20 mg kg-' i.p. 4.0 ± 0.6 41.1 ±5.6 66.9 ± 11.8 1839.3 ± 878.2* Control 5 49.9 ± 3.2 81.6 ± 6.2 206.6 ± 37.6 (+ )-HA-966 40mgkg-' i.p. 4.3 0.4 38.4 4.4 59.9 ± 7.9* 7411.9 ± 1661.4** Control 5 56.8 ± 4.3 80.8 ± 10.0 181.9 ± 12.8 D-Serine 2.5 pmol i.c.v. 5 55.8 ± 3.7 82.9 ± 5.5 219.1 ± 11.6 Control 5 53.7 ± 3.2 82.5 ± 4.5 190.7 ± 11.5 D-Serine 5.0 pmol i.c.v. 4.9 0.09 55.4 ± 4.9 81.5 ± 4.5 201.4± 18.1 Control 4.5 0.3 55.4 ± 3.5 114.7 ± 25.5 150.3 ± 26.5 D-Cycloserine 160mgkg-' i.p. plus (+)-HA-966 20 mg kg- ' i.p. 4.5 0.3 57.5 ± 2.3 82.4 ± 7.6 285.3 ± 48.9** Data are from experiments with groups of 9-11 fully kindled rats (except 40mgkg-' D-cycloserine, which is from a group of 6 animals). Control ADTs (with i.p. or i.c.v. injection of saline) were determined 2-3 days before each drug experiment; average pre-drug control ADT was 27.5 ± 1.1 fLA (mean ± s.e.mean of 12 pre-drug control threshold determinations in groups of 6-11 rats). Pretreatment times were 2.5 h (MK-801), I h (D-cycloserine, D-serine) or 0.5 h ((+)-HA-966). In case of combined treatment with D-serine and (+)-HA-966, both drugs were given 60min prior to ADT determination. At the individual pre-drug and drug ADT current, severity and duration of seizures as well as duration of amygdaloid afterdischarges were recorded. Furthermore, the duration of postictal immobilization ('freezing') was determined. All data are shown as means ± s.e.mean. Significant differences to pre-drug control are indicated by asterisks (*P <0.05; **P<0.01). ADT = afterdischarge threshold. average data shown in Figure 1) exhibited a marked increase (not illustrated). In contrast, no ADT increases were seen in ADT in response to D-cycloserine. When the dose of with MK-801, (+)-HA-966, D-serine or vehicle at post-drug D-cycloserine was increased to 160 or 320 mg kg-', sig- (or post-vehicle) control recordings (not illustrated). nificant ADT increases were found 1 h after administration (Figure 1). Seizure severity, seizure duration and afterdis- Effect of combined treatment with D-cycloserine and charge duration recorded at ADT currents were not changed ( + )-HA-966 on the focal seizure threshold (ADT) in by treatment with D-cycloserine (Table 1). amygdala-kindled rats The low efficacy partial glycine agonist, (+)-HA-966, did not significantly alter the focal seizure threshold at doses of In this experiment, (+ )-HA-966, 20 mg kg-', and D- 10, 20 or 40 mg kg-' (Figure 1). At the two higher doses, cycloserine, 160 mg kg-', were administered together, and the only 1 of 9-10 animals per group (not included in the ADT was recorded 1 h after injection as well as 2 and 5 days average data shown in Figure 1) displayed marked threshold later (Figure 3). In contrast to the findings with D-cycloserine increases in response to (+)-HA-966. The severity or dura- alone (Figures 1 and 2), the ADT was significantly decreased tion of seizures recorded at ADT currents was not altered by after 1 h and not increased after 2 days (Figure 3). Thus, (+)-HA-966 (Table 1). (+ )-HA-966 completely antagonized the anticonvulsant effect With the full glycine agonist, D-serine, significant increases in response to treatment with D-cycloserine. in ADT were recorded 1 h after i.c.v. injection of a dose of 5 gimol, but not 2.5 tmol (Figure 1). Seizure parameters Effect of (+ )-HA-966 on postictal immobilization recorded at ADT currents were not altered by treatment with D-serine (Table 1). Following termination of a fully kindled (stage 4-5) seizure, The significant increase in ADT seen after high doses of amygdala-kindled rats show a typical period of behavioural D-cycloserine was very long-lasting. Thus, it was still present depression ('freezing'), during which the animals do not move 2 days after drug administration (Figure 2). In order to prove but infrequently show facial clonus or head nodding. This the reproducibility of this long lasting alteration in ADT, the postictal immobilization is associated with postictal depres- experiment with 160 mg kg-' D-cycloserine was repeated in sion (i.e. flattened activity) in the electroencephalographic two other groups of kindled rats: in both groups, the ADT recording from the amygdala; the depressed electro- was significantly increased 2 days after drug administration encephalographic activity being interrupted by short periods D-CYCLOSERINE AND (+ )-HA-966 IN THE KINDLING MODEL 101 601 60 -

50 50 - R j. o 40 I1 cn al I 40 - -0ja is 30- .5 EC I I C.) I 01) I 4CCo 2 20- C.) 30 - 0)

tm 10 - 0) 20 -

0- Control DCS DCS (2 days 160 o: - a 320 c,> 1v c Vu. pre-drug) c-0,~~~~~~~~~~~oC' u m O co 0-0C4 ) - - (-A - a- X CQ 10

Figure 2 Duration of the increase in focal seizure threshold (ADT) after administration of high doses of i-cycloserine (DCS) in amygdala-kindled rats. Data are mean ± s.e.mean of a group of 9 fully kindled rats with reproducible ADTs. DCS was administered at a 2 after ADT control Control DCS Control Control i.p. dose of 160 mg kg' days pre-drug (2 days 160 + (2 days (5 days recording. The ADT shown for 'DCS 160' was determined I h after pre-drug) HA 20 post-drug) post-drug) drug administration. Two days after drug injection (2 days post- drug), the ADT was still significantly elevated. Five days after drug Figure 3 Effects of combined treatment with D-cycloserine (DS; injection (5 days post-drug), ADTs were not different from pre-drug 160mg kg-') and (+)-HA-966 (HA; 20mg kg-') on focal seizure control. Two days later, DCS was administered at a dose of threshold (ADT) in kindled rats. Data are means + s.e.mean of a 320 mg kg-', i.p. and ADT was determined after I h ('DCS 320'). group of 10 fully kindled rats with reproducible ADTs. Both drugs Similar to the experiment with 160mg kg-', the increase in ADT were administered consecutively 2 days after pre-drug ADT control was still present 2 days after drug injection (2 days post-drug), but recording. The ADT shown for 'DCS 160 + HA 20' was determined reached control level after 5 days. Significant differences from pre- I h after i.p. drug administration. Significant difference from pre- drug control are indicated by asterisks (*P<0.05; **P<0.01). The drug control is indicated by asterisk (*P<0.05). Two and 5 days time of day of ADT determinations was the same for all after drug administration, the ADT was not significantly different experiments. At all control days, saline was injected I h prior to from pre-drug control. The time of day of ADT determinations was ADT determination. Saline injection alone did not induce any altera- the same for all experiments. At all control days, saline was injected tions in ADTs (not illustrated). I h prior to ADT determination. of spiking only during the infrequent head nodding or facial weaving, ataxia, piloerection, and infrequent circling, ab- movements. As shown in Table 1, the postictal behavioural ducted hind limbs, reciprocal forepaw treading, flat body depression normally lasts for about 1-4min after termina- posture and Straub tail. Most of these behavioural alteration of a kindled seizure and afterdischarge. Thereafter, tions reached their maximal intensity after 1.5-2.5 h. The animals resume normal behaviour and electroencephalo- most intense behaviours, i.e. motor impairment, hyper- graphic activity. (+)-HA-966 significantly increased the dura- locomotion and stereotypies (head weaving) are shown in tion of postictal immobilization in a dose-dependent manner Figure 4. Motor impairment was characterized by ataxia in (Table 1). At the highest dose, viz. 40 mg kg-', (+)-HA-966 the open field and inability of the rats to remain on the increased the duration of freezing almost 40 fold, resulting in rotarod. None of these MK-801-induced behavioural altera- a total freezing time of 124 min! Concomitantly with the tions were seen after saline injection (not illustrated). D- prolongation of behavioural depression, the flattened activity Cycloserine was almost devoid of any behavioural effects in the EEG (postictal depression) was prolonged (not illus- (Figure 4). Some animals showed moderate piloerection and trated). In contrast to (+)-HA-966, MK-801, D-cycloserine Straub tail after high doses but nothing more. (+)-HA-966 or D-serine did not increase the duration of postictal induced moderate motor impairment at the highest dose immobilization (Table 1) or postictal EEG depression. D- tested (Figure 4). Furthermore, piloerection and Straub tail Cycloserine significantly decreased the period of postictal were observed in some rats. Combined treatment with D- behavioural depression at 20 mg kg-' but not at any of the cycloserine and (+)-HA-966 did not induce more marked other doses tested (Table 1). After combined treatment with adverse effects than single drug treatment (Figure 4). Neither (+)-HA-966 and D-cycloserine, the effect of (+)-HA-966 on D-cycloserine nor (+)-HA-966, nor combined treatment with postictal immobilization appeared to be reduced (Table 1), both drugs induced PCP-like behaviours, such as hyper- but it has to be considered that pretreatment time of (+)- locomotion or stereotypies, at any dose tested. Furthermore, HA-966 was 60 min in this experiment instead of 30 min in no signs of proconvulsant activity were seen with D- the experiments with (+)-HA-966 alone. cycloserine at any time after treatment, including the 1-h interval prior to ADT determination. In this period, electro- Behavioural effects of MK-801, D-cycloserine, encephalographic recordings from the amygdala showed nor- ( + )-HA-966 and D-serine in kindled rats mal activity without any indication of paroxysmal activity (not illustrated). Similarly, no proconvulsant activity was MK-801, 0.1 mg kg'I, induced the characteristic PCP-like seen after i.c.v. injection of D-serine. Indeed, all rats behaved behavioural syndrome, consisting of hyperlocomotion, head normally after D-serine administration (Figure 4). 102 W. LOSCHER et al.

Plasma and brain concentrations of D-cyclosAerine after Discussion i.p. injection in rats D-Cycloserine was discovered in 1954 in a culture of Strep- Following i.p. administration of D-cycloserine, 320 mg tomyces orchidaceus. Found to be a broad-spectrum anti- high plasma levels of 3427 nmol ml-' (350 L,gmlr 1) kg'er,were biotic with a moderate degree of effectiveness in vitro against determined after 15 min (Table 2). Plasma levels dec- the tubercle bacillus, its most important clinical application lined to 2199 nmol ml-' after 60 min. From tihishrapidlydecline, a has been in the treatment of tuberculosis (Mandell & Sande, plasma half-life of about 70 min can be estimlated. In con- 1990). Although the drug was found to be relatively nontoxic trast to plasma, levels of D-cycloserine in brgsin tissue in- in experimental animals (Tettenborn, 1988), in clinical use the creased from 15 to 60 min after administraticn (Table 2). administration of high doses of D-cycloserine (1 g per day) to Consequently, the brain/plasma ratio was onbly 0.13 after tuberculosis patients has been attended by symptoms of 15 min but 0.36 after 60 min. Peak brain levels c after neurotoxicity, such as drowsiness, somnolence, hyperflexia, 60 min were 800 nmol g' (80 Ig g'1), corre*spondingobtained to mental confusion, psychotic states and tonic-clonic or about 23% of peak plasma levels. absence seizures (Storey & McLean, 1957; Mandell & Sande, 1990). Anticonvulsant effects of D-cycloserine were first reported by Fust et al. (1958) and Mayer et al. (1971) who (Score) (%) found that the drug had a moderate anticonvulsant effect in

4- 6--100 fatal pentylenetetrazole (PTZ) shock in mice. More recently, a) 5 -- 80 Peterson & Schwade (1993) showed that D-cycloserine 4- (50-300mgkg-', i.p.) dose-dependently blocked tonic but E. 260 not clonic PTZ seizures in rats with an ED50 of 109 mg kg-', i.p. The anticonvulsant effect was stereospecific in that L- aL3 -j I cycloserine was ineffective. Similar results were reported for -- the MES test in rats, in which D-cycloserine, administered 4 either i.p. or i.c.v., induced significant anticonvulsant activity 2- with a time of peak effect of 1-2 h and an ED50 of C3- 153 mg kg', i.p. or 5 jmol, i.c.v. (Peterson, 1992). C0'. Shortly after the discovery of the stimulatory action of glycine at the NMDA receptor, it was found that D- cycloserine binds to the glycine/NMDA modulatory site (Ki 2 2.3 pmol 1' compared to 0.2 pmol 1' for glycine) and acts as a partial agonist of relatively high activity (cf., Carter, 0. 1992; Kemp & Leeson, 1993). In in vivo experiments in mice - -- and rats, D-cycloserine was found to possess cognition- enhancing properties, which were related to the partial -a 3- agonistic effects of this drug at the glycine/NMDA site 0 (Flood et al., 1992; Schuster & Schmidt, 1992; Sirvio et al., 0 1992). Maximum cognition-enhancing effects were seen at i.p. 2- T, doses of 10-20 mg kg-'. In the present experiments in kind- a) led rats, no behavioural or anticonvulsant effects were >. 1 0 observed after administration of D-cycloserine at doses of 20, M C C C CSDSH AH 40 or 80 mg kg-'. Similarly, Peterson (1992) did not obtain 0 4 8 6 2 1 04 0- en )S HA + DCS anticonvulsant effects in the MES test at doses below 5 20 + 160 100mg kg-' D-cycloserine. Significant increases in focal seizure threshold in kindled rats were found when the dose of Figure 4 Behavioural adverse effects of MK-8101 (MK), D- D-cycloserine was increased to 160 or 320mg kg-', which is cycloserine (DCS), D-serine (DS), (+)-HA-966 (HA)>, or combined consistent with the dose-range reported by Peterson (1992) treatment with HA and DCS in amygdala-kindled rat's. The doses of and Peterson & Schwade (1993) to exert anticonvulsant drugs are shown below each column in mgkg-', i.p. (for DS in effects in the MES and PTZ tests in rats. In contrast to the jimol, i.c.v.). Data are mean + s.e.mean of 9-11 ful kindled rats present data, Peterson & Schwade found that D- per experiment (except data for DCS 40, which are fi oy (1993) had no effect on kindled and were determined immediately prior to ADT determinations. The cycloserine significant amygdala severity of motor impairment was scored by behaviouiralratings (left seizures in doses up to 400 mg kg-'. However, Peterson & side of ordinate scale; open columns) and also detetrmined in the Schwade (1993) used amygdala stimulation with a fixed sup- rotarod test (right side of ordinate scale; solid coluimns). For the rathreshold current of 400liA, whereas we determined the rotarod test, the percentage of animals which did not pass the test is individual focal seizure threshold before and after application given. Hyperlocomotion and stereotypies (head weavirig) were scored of D-cycloserine, which may explain the differences between by severity ratings. Other behavioural alterations not included in the experimental results. figure, such as circling and flat body posture, were onily infrequently After systemic application, D-cycloserine is distributed observed with MK-801 and not at all with any of tiie other drugs. throughout body fluids and tissues. There seems to be no For pretreatment times of drugs and further details, see legend to Figure 1. appreciable blood-brain barrier to the drug, since cerebro- spinal fluid concentrations in tuberculosis patients are about 50% of those in plasma (Iwainsky, 1988). However, the Table 2 Levels of D-cycloserine in plasma and birain tissue present data on D-cycloserine levels in plasma and brain after i.p. administration of D-cycloserine in rats demonstrate that in rats, only about 20% of peak plasma Concentration of D-cycloserine concentrations appear in the brain, which may explain the Time after Plasma Frontal cortex Brain/plasma relatively high doses of D-cycloserine necessary to induce administration (nmol ml- ') (nmol g- I) ratio anticonvulsant effects. A similar low brain penetration rate has previously been determined for D-cycloserine in rats by 15 min 3427 ± 81.9 450 ± 118 ± ± 0.360.36 Crema & Bert6 (1960). Consistent with the present data, peak 60 min 2199 75.1 800 122 brain levels were determined after 60 min, and brain levels Data (means ± s.e.mean) are from experiments with two rapidly declined thereafter with a half-life of about 2.2 h groups of 6 rats each. D-Cycloserine was adminis;tered at a (Crema & Berte, 1960). In view of the limited penetration of dose of 320 mg kg-', i.p. D-cycloserine from blood to brain in intact animals, the D-CYCLOSERINE AND (+ )-HA-966 IN THE KINDLING MODEL 103 higher neurotoxicity of D-cycloserine in tuberculosis patients devoid of such anticonvulsant activity. This was proven with compared to laboratory animals (Tettenborn, 1988) may thus (+ )-HA-966, a highly selective ligand for the strychnine- be a consequence of disturbed blood-brain-barrier function. insensitive glycine site (Singh et al., 1990a). Tested in doses As shown by the present data, the increase in focal seizure of 10-40 mg kg-', (+)-HA-966 did not exert any significant threshold in kindled rats following D-cycloserine lasted for effects on focal seizure threshold or severity and duration of several days. In view of the rapid elimination of D-cycloserine seizures recorded at seizure threshold. Observation of in rats, this long-lasting increase in seizure threshold is behaviour showed that (+)-HA-966 induced motor impair- difficult to explain by a direct anticonvulsant effect of the ment in this dose-range, indicating that pharmacologically drug mediated by the glycine site. However, the fact that active doses had been tested. Indeed, (+ )-HA-966 was both the acute and the long-lasting increase of focal seizure recently reported to exert effects via NMDA receptors at in response to D-cycloserine were antagonized by treatment doses of 3- 30 mg kg-' in rats (Hutson et al., 1991; Dunn et with the selective glycine receptor ligand, (+ )-HA-966, al., 1992; Bristow et al., 1993). In the only previous study in strongly indicate that these effects of D-cycloserine were due which (+)-HA-966 was examined in amygdala-kindled rats, to interactions with the glycine receptor. the compound was injected into the amygdala (Croucher & One explanation for the long-lasting increase in seizure Bradford, 1991). Moderate (20-30%) increases in GST were threshold following administration of D-cycloserine would be found 20 min after intra-amygdaloid injections of 50-100 adaptive changes such as desensitization of glycine/NMDA nmol (+)-HA-966, while seizure duration or afterdischarge receptors in response to the high efficacy partial agonist. duration were not altered (Croucher & Bradford, 1991). Recent experiments with the glycine partial agonist ACPC Although we did not determine the GST in our experiments, (1-aminocyclopropane-l-carboxylic acid), which has similar most animals showed generalized motor seizures at ADT intrinsic activity (60-80%) as D-cycloserine (Rao et al., 1990) currents, which were not affected by systemic treatment with have suggested that the NMDA receptor complex can be (+ )-HA-966. Thus, the present experiments demonstrated desensitized by treatment with a high efficacy partial glycine that, in contrast to D-cycloserine and D-serine, (+)-HA-966 agonist, resulting in significant protection against cerebral was not capable of exerting anticonvulsant effects in the ischaemia and NMDA-induced convulsions and death (von kindling model. Similar findings were reported for i.c.v. Lubitz et al., 1992; Skolnick et al., 1992). Such adaptive administration of the glycine antagonist, 7-chlorokynurenic changes in the NMDA receptor complex would also explain acid in amygdala-kindled rats (Morimoto & Sato, 1992), the anticonvulsant effect of D-cycloserine found by Peterson thus suggesting that glycine antagonists or low efficacy (1990) and the anticonvulsant effects of other glycine agonists partial agonists are ineffective in this model of partial or high efficacy partial agonists reported by other groups (see epilepsy. Instead, as shown by the present data, (+)-HA-966 Introduction). However, the present finding that no pro- antagonized the anticonvulsant effect induced by D-cyclo- longed increase in seizure threshold is induced by the glycine/ serine. NMDA receptor full agonist D-serine (see below) would Previous studies with systemic administration of (+)-HA- argue against receptor desensitization as the responsible 966 in other seizure models have shown that very high doses mechanism for the long duration of D-cycloserine's anticon- are needed to block NMDA-induced seizures (ED50 900 mg vulsant effect. kg-') or tonic seizures in the MES test (ED50 106 mg kg-') in Alternative explanations for the 'paradoxical' anticonvul- mice (Singh et al., 1990a; Vartanian & Taylor, 1991). We did sant effects of the glycine high efficacy partial agonist, D- not try such high doses in kindled rats, since in the dose cycloserine, include the possibility that it might act by range of 10-40 mg kg-', (+)-HA-966 already caused marked potentiating the activity of an endogenous glutamate anta- increases in the duration of postictal immobilization. At the gonist, such as kynurenic acid (Kemp & Leeson, 1993), since highest dose tested in the present study (40 mg kg-'), rats recent experiments by Norris et al. (1992) have shown that were immobilized for more than 2 h following the kindled D-cycloserine, 320 mg kg- , significantly potentiated the seizure. The behavioural and EEG depression following kind- ability of MK-801 to raise seizure threshold. Furthermore, it led and other types of seizures is thought to represent has been suggested that NMDA receptor-mediated release of decreased or inhibited neuronal firing following seizure neurotransmitters such as noradrenaline and GABA may be activity (Goddard et al., 1986). Several neurotransmitter involved in the anticonvulsant effects of agonists or high systems, including GABA, adenosine and opioids, have been efficacy partial agonists at the strychnine-insensitive glycine involved in the mechanisms underlying these postictal altera- site (Peterson, 1991). tions (Dragunow, 1986; Goddard et al., 1986; Loscher, 1989). Interestingly, as with D-cycloserine, significant increases in For instance, similar prolongation of postictal depression as focal seizure threshold were also produced by the glycine obtained in the present study with (+ )-HA-966 was receptor full agonist D-serine, but the duration of this effect previously found with systemic administrations of morphine was much shorter than with D-cycloserine. D-Serine is a and adenosine analogues in kindled rats (Frenk et al., 1979; selective agonist of the strychnine-insensitive glycine receptor Whitcomb et al., 1990). It has been proposed that the with a similar affinity for the receptor to glycine (Kemp & mechanisms involved in postictal behavioural and EEG Leeson, 1993). The effects of D-cycloserine and D-serine in depression are responsible for spontaneous seizure arrest and kindled rats substantiate the theory that the strychnine- may forestall the onset of the next seizure or reduce the insensitive glycine receptors are not saturated at physiological severity of subsequent seizures (Dragunow, 1986; Goddard et concentrations of glycine and that exogenously administered al., 1986; Loscher, 1989). In addition to behavioural and glycine agonists can influence the glycine receptor activity in electroencephalographic depression, the postictal phase of vivo (Singh et al., 1990b; Peterson, 1991; Carter, 1992; Kemp kindled rats is characterized by intermittent postictal spiking, & Leeson, 1993). To our knowledge, the present data are the which has been related to a paroxysmal activation of first demonstration of a direct anticonvulsant effect of D- inhibitory systems (Frenk et al., 1979; Engel et al., 1981). serine. In contrast to D-cycloserine, the blood-brain barrier is However, as reported previously (e.g. Engel et al., 1981; almost impermeable to D-serine so that the compound has to Loscher & Honack, 1990a), this postictal spiking is often be administered centrally. However, oral administration of associated with signs of limbic seizure activity (facial clonus, very large doses of D-serine (2100 mg kg-') in rats was shown head nodding) which could indicate an overlap of both ictal to enhance the anticonvulsant effect of phenobarbitone, car- and postictal phenomena during the phase of behavioural bamazepine and phenytoin (Peterson, 1991). depression. The finding that the low efficacy partial glycine If stimulation of glycine/NMDA receptors is essential for agonist, (+ )-HA-966, markedly enhanced the duration of the anticonvulsant effect observed after administration of the postictal behavioural and EEG depression in kindled rats high efficacy partial glycine agonist D-cycloserine and the full may indicate that the glycine modulatory site is involved in agonist D-serine, then a low efficacy partial agonist should be these postictal events and that (+ )-HA-966 possibly acts as a 104 W. LOSCHER et al. glycine antagonist by potentiating and/or prolonging a post- associated with hyperactivation of NMDA receptor function, ictal decrease of glutamatergic activity. such as epilepsy and ischaemia. However, in view of the fact The lack of anticonvulsant activity of the uncompetitive that high doses of D-cycloserine induce seizures in man NMDA receptor antagonist, MK-801, in fully kindled rats (Storey & McLean, 1957), this strategy should be explored confirmed previous reports on this drug (McNamara et al., with caution, although glycine site antagonists, such as 1988; Loscher & H6nack, 1991a). MK-801 was primarily 7-chlorokynurenic acid and NMDA receptor antagonists are included in the present experiments for direct comparison of known to induce (pro)convulsant effects at high doses as well PCP-like behavioural effects with adverse effects in response (Myslobodsky et al., 1981; Klockgether et al., 1988; Koek & to the glycine site ligands. Neither D-cycloserine nor (+)- Colpaert, 1990; LUscher & H6nack, 1990b). The anticonvul- HA-966 produced PCP-like behavioural alterations in kind- sant activity of D-cycloserine and D-serine and the unex- led rats, confirming previous observations with these drugs in pected findings with the low efficacy partial glycine agonist non-kindled rodents (Hutson et al., 1991; Peterson, 1992; (+)-HA-966 in kindled rats substantiate that we are far from Bristow et al., 1993). Only in one study with i.c.v. injection in understanding the complex interactions between NMDA mice, 5,7-dichlorokynurenic acid, (+)-HA-966 and D-cyclo- receptor functions and ictal and postictal events. In view of serine were reported to induce head weaving, but the peak the fact that amygdala-kindling is thought to represent a intensity of the response was never more than 12% of that predictive model of complex partial seizures, i.e. the most seen with MK-801 (Tricklebank & Saywell, 1990). We have common type of epileptic seizures in man (Lbscher, 1993), recently shown that amygdala-kindled rats are more suscep- the lack of anticonvulsant activity of (+)-HA-966 combined tible to PCP-like adverse effects, including head weaving, of with the pronounced increase in postictal depression found in NMDA antagonists than non-kindled rats, most probably the present study might suggest that glycine antagonists or because of the kindling-induced changes in glutamate recep- low efficacy partial agonists such as (+)-HA-966 are of tor function (LUscher & H6nack, 1991b,c). The present data limited value for clinical development as antiepileptic drugs. seem to indicate that this higher susceptibility of kindled rats However, in view of the fact that the tools currently in use to behavioural adverse effects of NMDA receptor antagonists are far from perfect, with affinities for the glycine/NMDA does not extend to glycine site ligands, although motor receptor often well in excess of 1 lM, final judgment about impairing effects of (+ )-HA-966 were observed at lower the usefulness of glycine site antagonists should await doses than reported previously for non-kindled rats (Bristow development of compounds with high affinity and selectivity. et al., 1993). In conclusion, the present findings indicate that the strychnine-insensitive glycine receptor can be manipulated by We thank Dr W. Danysz (Merz + Co., Frankfurt/M, Germany) for helpful discussions and Mrs D. Pieper and Mrs M. Gramer for high efficacy partial agonists such as D-cycloserine to effect technical assistance. The study was supported by a grant (Lo 274/2- changes in the NMDA receptor complex leading to long- 3) from the Deutsche Forschungsgemeinschaft. P.W. acknowledges lasting decreases in seizure susceptibility. As recently sug- financial support by the TEMPUS programme of the Commission of gested from experiments with the partial glycine agonist the European Communities and H.B. the financial support of the ACPC (von Lubitz et al., 1992; Skolnick et al., 1992), this Austrian Science Research Fund (Charlotte-Biihler-Habilitations- strategy might be interesting for treatment of disorders stipendium No H017-MED).

References BARON, B.M., HARRISON, B.L., MCDONALD, I.A., MELDRUM, B.S., ENGEL, J. Jr., ACKERMANN, R.F., CALCECOTT-HAZARD, S. & PALFREYMAN, M.G., SALITURO, F.G., SIEGEL, B.W., SLONE, KUHL, D.E. (1981). Epileptic activation of antagonistic systems A.L., TURNER, J.P. & WHITE, H.S. (1992). Potent indole- and may explain paradoxical features of experimental and human quinoline-containing N-methyl-D-aspartate antagonists acting at epilepsy: a review and hypothesis. In Kindling 2. ed. Wada, J.A. the strychnine-insensitive glycine binding site. J. Pharmacol. Exp. pp. 193-217. New York, U.S.A: Raven Press. Ther., 262, 947-956. FLOOD, J.F., MORLEY, J.E. & LANTHORN, T.H. (1992). Effect on BRISTOW, L.J., HUTSON, P.H., THORN, L. & TRICKLEBANK, M.D. memory processing by D-cycloserine, an agonist of the NMDA/ (1993). The glycine/NMDA receptor antagonist, R-(+)-HA-966, glycine receptor. Eur. J. Pharmacol., 221, 249-254. blocks activation of the mesolimbic dopaminergic system induced FOSTER, A. & KEMP, J.A. (1989). Glycine maintains excitement. by phencyclidine and dizocilpine (MK-801) in rodents. Br. J. Nature, 338, 377-378. Pharmacol., 108, 1156-1163. FREEMAN, F.G. & JARVIS, M.F. (1981). The effect of interstimulation CARTER, A.J. (1992). Glycine antagonists: regulation of the NMDA interval on the assessment and stability of kindled seizure receptor-channel complex by the strychnine-insensitive glycine thresholds. Brain Res. Bull., 7, 629-633. site. Drugs Future, 17, 595-613. FRENK, H., ENGEL, J. Jr., ACKERMANN, R.F., SHAVIT, Y. & LIEBES- CREMA, A. & BERTA, F. (1960). t0ber die Verteilung und Stoff- KIND, J.C. (1979). Endogenous opioids may mediate post-ictal wechsel des Cycloserins bei der Ratte. Naunyn-Schmied. Arch. behavioral depression in amygdala-kindled rats. Brain Res., 167, Pharmacol., 239, 475-480. 435-440. CROUCHER, M.J. & BRADFORD, H.F. (1990). 7-Chlorokynurenic FUST, B., BOHNI, E., PELLMONT, B., ZBINDEN, G. & STUDER, A. acid, a strychnine-insensitive glycine receptor antagonist, inhibits (1958). Experimentelle Untersuchungen mit D-cycloserin. Schweiz. limbic seizure kindling. Neurosci. Lett., 118, 29-32. Zeitschr. Tuberk., 15, 129-157. CROUCHER, M.J. & BRADFORD, H.F. (1991). The influence of GODDARD, G.V., DRAGUNOW, M., MARU, E. & MACLEOD, E.K. strychnine-insensitive glycine receptor agonists and antagonists (1986). Kindling and the forces that oppose it. In The Limbic on generalized seizure thresholds. Brain Res., 543, 91-96. System: Functional Organization and Clinical Disorders. ed. DINGLEDINE, R., McBAIN, C.J. & McNAMARA, J.O. (1990). Doane, B.K. & Livingston, K.E. pp. 95-108. New York, U.S.A.: Excitatory amino acid receptors in epilepsy. Trends Pharmacol. Raven Press. Sci., 11, 334-338. HENDERSON, G., JOHNSON, J.W. & ASCHER, P. (1990). Competitive DRAGUNOW, M. (1986). Adenosine: the brain's natural anticonvul- antagonists and partial agonists at the glycine modulatory site of sant? Trends Pharmacol. Sci., 7, 128-130. the mouse N-methyl-D-aspartate receptor. J. Physiol., 430, DUNHAM, N.W. & MIYA, T.S. (1957). A note on a simple apparatus 189-212. for detecting neurological deficit in rats and mice. J. Am. Pharm. HUTSON, P.H., BRISTOW, L.J., THORN, L. & TRICKLEBANK, M.D. Assoc., 46, 208-209. (1991). R-(+)-HA-966, a glycine/NMDA receptor antagonist, DUNN, R.W., FLANAGAN, D.M., MARTIN, L.L., KERMAN, L.L., selectively blocks the activation of the mesolimbic dopamine WOODS, A.T., CAMACHO, F., WILMOT, C.A., CORNFELDT, M.L., system by amphetamine. Br. J. Pharmacol., 103, 2037-2044. EFFLAND, R.C., WOOD, P.L. & CORBETT, R. (1992). Stereoselec- tive R-(plus) enantiomer of HA-966 displays anxiolytic effects in rodents. Eur. J. Pharmacol., 214, 207-214. D-CYCLOSERINE AND (+)-HA-966 IN THE KINDLING MODEL 105

IWAINSKY, H. (1988). Mode of action, biotransformation and phar- NORRIS, D.O., MASTROPAOLO, J., O'CONNOR, D.A., NOVITZKI, macokinetics of antituberculosis drugs in animals and man. In M.R. & DEUTSCH, S.I. (1992). Glycinergic interventions potentiate Antituberculosis Drugs. ed. Bartmann, K. pp. 399-553. Berlin, the ability of MK-801 to raise the threshold voltage for tonic Germany: Springer-Verlag. hindlimb extension in mice. Pharmacol. Biochem. Behav., 43, JOHNSON, J.W. & ASCHER, P. (1987). Glycine potentiates the 609-612. NMDA response in cultured mouse brain neurones. Nature, 325, PAXINOS, G. & WATSON, C. (1986). The Rat Brain in Stereotoxic 529-531. Coordinates, 2nd ed. Sydney, Australia: Academic Press. KEMP, J.A. & LEESON, P.D. (1993). The glycine site of the NMDA PETERSON, S.L. (1991). Anticonvulsant drug potentiation by glycine receptor - five years on. Trends Pharmacol. Sci., 14, 20-25. in maximal electroshock seizures is mimicked by D-serine and KLOCKGETHER, T., TURSKI, L., SCHWARZ, M., SONTAG, K.-H. & antagonized by 7-chlorokynurenic acid. Eur. J. Pharmacol., 199, LEHMANN, J. (1988). Paradoxical convulsant action of a novel 341-348. non-competitive N-methyl-D-aspartate (NMDA) antagonist, tilet- PETERSON, S.L. (1992). 7-Chlorokynurenic acid antagonizes the amine. Brain Res., 461, 343-348. anticonvulsant activity of D-cycloserine in maximal electroshock KOEK, W. & COLPAERT, F.C. (1990). Selective blockade of N-methyl- seizures. Epilepsy Res., 13, 73-81. D-aspartate (NMDA)-induced convulsions by NMDA PETERSON, S.L. & BOEHNKE, L.E. (1989). Anticonvulsant effects of antagonists and putative glycine antagonists: relationship with MK-801 and glycine on hippocampal afterdischarge. Exp. phencyclidine-like behavioral effects. J. Pharmacol. Exp. Ther., Neurol., 104, 113-117. 252, 349-357. PETERSON, S.L. & SCHWADE, N.D. (1993). The anticonvulsant LAPIN, I.P. (1981). Antagonism of glycine to seizures induced by activity of D-cycloserine is specific for tonic convulsions. Epilepsy L-kynurenic, quinolinic acid and strychnine in mice. Eur. J. Res., 15, 141-148. Pharmacol., 71, 495-498. RACINE, R.J. (1972). Modification of seizure activity by electrical LOSCHER, W. (1989). GABA and the epilepsies. Experimental and stimulation. II. Motor seizure. Electrencephalogr. Clin. Neuro- clinical considerations. In GABA: Basic Research and Clinical physiol., 32, 295-299. Applications. ed. Bowery, N.G. & Nistic6, G. pp. 260-300. RAO, T.S., CLER, J.A., COMPTON, R.P., EMMETT, M.R., MICK, S., Rome, Italy: Pythagora Press. SUN, E.T., IYENGAR, S. & WOOD, P.L. (1990). Neuropharma- LOSCHER, W. (1993). Basic aspects of epilepsy. Curr. Opin. Neurol. cological characterization of 1-aminocyclopropane-l-carboxylate Neurosurg., 6, 223-232. and 1-aminocyclobutane-l-carboxylate. Ligands of the N-methyl- LOSCHER, W. & H6NACK, D. (1990a). The effect of interstimulation D-aspartate-associated glycine receptor. Neuropharmacology, 29, interval on the assessment of anticonvulsant drug potency in fully 305-309. kindled rats. Epilepsy Res., 7, 182-196. RUNDFELDT, C. & LOSCHER, W. (1992). Development of tolerance LOSCHER, W. & HONACK, D. (1990b). High doses of memantine to the anticonvulsant effect of vigabatrin in amygdala-kindled (l-amino-3,5-dimethyladamantane) induce seizures in kindled but rats. Eur. J. Pharmacol., 213, 351-366. not in non-kindled rats. Naunyn-Schmied. Arch. Pharmacol., 341, SCHUSTER, G.M. & SCHMIDT, W.J. (1992). D-Cycloserine reverses 476-481. the working memory impairment of hippocampal-lesioned rats in LOSCHER, W. & HONACK, D. (1991a). Anticonvulsant and a spatial learning task. Eur. J. Pharmacol., 224, 97-98. behavioral effects of two novel competitive N-methyl-D-aspartate SINGH, L., DONALD, A.E., FOSTER, A.C., HUTSON, P.H., IVERSEN, acid receptor antagonists, CGP 37849 and CGP 39551, in the L.L., IVERSEN, S.D., KEMP, J.A., LEESON, P.D., MARSHALL, G.R., kindling model of epilepsy. Comparison with MK-801 and car- OLES, R.J., PRIESTLEY, T., THORN, L., TRICKLEBANK, M.D., bamazepine. J. Pharmacol. Exp. Ther., 256, 432-440. VASS, C.A. & WILLIAMS, B.J. (1990a). Enantiomers of HA-966 LOSCHER, W. & HONACK, D. (1991b). The novel competitive N- (3-amino-l-hydroxypyrrolid-2-one) exhibit distinct central ner- methyl-D-aspartate (NMDA) antagonist CGP37849 preferen- vous system effects: (+)-HA-966 is a selective glycine/N-methyl- tially induces phencyclidine-like behavoral effects in kindled rats: D-aspartate receptor antagonist, but (-)-HA-966 is a potent attenuation by manipulation of dopamine, alpha-l and ?-butyrolactone-like sedative. Proc. Nati. Acad. Sci. U.S.A., 87, serotoninlA receptors. J. Pharmacol. Exp. Ther., 257, 1146-1153. 347-351. LOSCHER, W. & H6NACK, D. (1991c). Responses to NMDA receptor SINGH, L., OLES, R.J. & TRICKLEBANK, M.D. (1990b). Modulation antagonists altered by epileptogenesis. Trends Pharmacol. Sci., 12, of seizure susceptibility in the mouse by the strychnine-insensitive 52. glycine recognition site of the NMDA receptors/ion channel com- LOSCHER, W. & HONACK, D. (1992). The behavioural effects of plex. Br. J. Pharmacol., 99, 285-288. MK-801 in rats: involvement of dopaminergic, serotonergic and SIRVIO, J., EKONSALO, T., RIEKKINEN, P., LAHTINEN, H. & RIEK- noradrenergic systems. Eur. J. Pharmacol., 215, 199-208. KINEN, P. (1992). D-Cycloserine, a modulator of the N-methyl- LOSCHER, W., HONACK, D. & HASHEM, A. (1987). Anticonvulsant D-aspartate receptor, improves spatial learning in rats treated efficacy of clonazepam and the J-carboline ZK 93423 during with muscarinic antagonist. Neurosci. Lett., 146, 215-218. chronic treatment in amygdala-kindled rats. Eur. J. Pharmacol., SKOLNICK, P., MARVIZON, J., JACKSON, B., MONN, J., RICE, K. & 143, 403-414. LEWIN, A. (1989). Blockade of N-methyl-D-aspartate induced LOSCHER, W., HORSTERMANN, D., HONACK, D., RUNDFELDT, C. convulsions by 1-aminocyclopropanecarboxylates. Life Sci., 45, & WAHNSCHAFFE, U. (1993). Transmitter amino acid levels in 1647-1655. rat brain regions after amygdala-kindling or chronic electrode SKOLNICK, P., MILLER, R., YOUNG, A., BOJE, K. & TRULLAS, R. implantation without kindling: evidence for a pro-kindling effect (1992). Chronic treatment with l-aminocyclopropanecarboxylic of prolonged electrode implantation. Neurochem. Res., 18, acid desensitizes behavioral responses to compounds acting at the 775-781. N-methyl-D-aspartate receptor complex. Psychopharmacology, MANDELL, G.L. & SANDE, M.A. (1990). Drugs used in the 107, 489-496. chemotherapy of tuberculosis and leprosy. In Goodman and Gil- STARK, L., PETERSON, S.L. & ALBERTSON, T.E. (1990). Phar- man's The Pharmacological Basis of Therapeutics. ed. Goodman macological and toxicological evaluation of potential antiepileptic Gilman, A., Rall, T.W., Nies, A.S. & Taylor, P. pp. 1146-1164. drugs in the kindling model of epilepsy. In Kindling 4. ed. Wada, New York, U.S.A.: Pergamon Press. J.A. pp. 267-281. New York, U.S.A.: Raven Press. MAYER, O., JANKU, I. & KRSAK, M. (1971). Die zentralen Wir- STOREY, P.B. & MCLEAN, R.L. (1957). A current appraisal of cyclo- kungen des Cycloserins im Tierexperiment. Arzneim.-Forsch., serine. HAM & CT, 4, 223-232. (Drug Res.) 21, 298-303. TETTENBORN, D. (1988). Experimental pharmacology and tox- McNAMARA, J.O., RUSSELL, R.D., RIGSBEE, L. & BONHAUS, D.W. icology of antituberculosis drugs. In Antituberculosis Drugs. ed. (1988). Anticonvulsant and antiepileptogenic actions of MK-801 Bartmann, K. pp. 307-398. Berlin, Germany: Springer-Verlag. in the kindling and electroshock models. Neuropharmacology, 27, TOTH, E., LAJTHA, A., SARHAN, S. & SEILER, N. (1983). Anticonvul- 563-568. sant effects of some inhibitory amino acids. Neurochem. Res., 8, MORIMOTO, K. & SATO, M. (1992). NMDA receptor complex and 291-302. kindling mechanisms. In Molecular Neurobiology of Epilepsy. ed. TRICKLEBANK, M.D. & SAYWELL, K. (1990). Behavioural properties Engel, J.Jr., Wasterlain, C., Cavaheiro, E.A., Heinemann, U. & of antagonists acting at the glycine modulatory site on the Avanzini, G. pp. 297-305. Amsterdam, Netherlands: Elsevier NMDA receptor/ion channel complex. Soc. Neurosci. Abstr., 16, Science Publishers. 462. MYSLOBODSKY, M.S., GOLOVCHINSKY, V. & MINTZ, M. (1981). Ketamine: convulsant or anti-convulsant? Pharmacol. Biochem. Behav., 14, 27-33. 106 W. LOSCHER et al.

VARTANIAN, M.G. & TAYLOR, C.P. (1991). Different stereoselectivity WHITCOMB, K., LUPICA, C.R., ROSEN, J.B. & BERMANN, R.F. of the enantiomers of HA-966 (3-amino-l-hydroxy-2-pyrrolidi- (1990). Adenosine involvement in postictal events in amygdala- none) for neuroprotective and anticonvulsant actions in vivo. kindled rats. Epilepsy Res., 6, 171-179. Neurosci. Lett., 145, 133-122. WOOD, D.J., WATSON, J.W. & STACEY, E.N. (1966). A comparative VON LUBITZ, D.K.J.E., LIN, R.C.-S., MCKENZIE, R.J., DEVLIN, T.M., study of hyperbaric oxygen-induced and drug-induced convul- MCCABE, R.T. & SKOLNICK, P. (1992). A novel treatment of sions with particular reference to GABA metabolism. J. global ischemia with a glycine partial agonist. Eur. J. Pharmacol., Neurochem., 13, 361-370. 219, 153-158. (Received October 19, 1993 Revised January 13, 1994 Accepted January 19, 1994)