Potentiation by glycine of anticonvulsant drugs in maximal electroshock seizures in rats

Neuropharmaeology Vol. 29, No. 4, pp. 399409, Printed in Great Britain. All rights reserved

1990 Copyright

0

0028-3908/90 $3.00 + 0.00 1990 Pergamon Press plc

POTENTIATION BY GLYCINE OF ANTICONVULSANT DRUGS IN MAXIMAL ELECTROSHOCK SEIZURES IN RATS S. L. PETERSON,‘*J. P. TRZECIAKOWSKI,’ G. D. FRYE’ and H. R. ADAMS~ ‘Department of Medical Pharmacology and Toxicology, Texas A&M University, College Station, Texas 77843, U.S.A. and ‘Department of Clinical Pathology, Scott and White Memorial Hospital and Clinic, 2401 South 31st St, Temple, Texas 76508, U.S.A. (Accepted 29 September 1989)

Summary-This study evaluated the potentiation by glycine of anticonvulsant drugs in maximal electroshock seizures in rats. Administered alone, glycine (40 mmol/kg, p.0.) induced no anticonvulsant effect or neurotoxicity. Administered together with the anticonvulsants, glycine significantly enhanced the anticonvulsant potency of phenobarbital and carbamazepine. Glycine also potentiated the anticonvulsant actions of MK-801 and diazepam but did not improve the selectivity of the drugs, as effective doses were still associated with neurotoxicity. Glycine did not potentiate phenytoin or sodium divalproate. Administration together with glycine had no significant effect on the concentrations of phenobarbital or carbamazepine in the brain. Administration together with phenobarbital had no relevant effect on the concentration of glycine in the brain but administration of glycine and carbamazepine together resulted in an increased concentration of glycine in the hippocampus and brainstem. These findings indicate a possible glycine-sensitive component in the mechanism of action of phenobarbital, carbamazepine and diazepam in maximal electroshock seizures. Although the mechanism may not be mediated by a glycine-GABA interaction, the evidence does implicate a possible interaction between glycine and anticonvulsant drugs at NMDA receptors. Key words-glycine,

phenobarbital,

carbamazepine, diazepam, maximal electroshock, rats.

Administered by itself, glycine has been shown to possess modest anticonvulsant properties. Glycine has been tested in seizures induced by hyperbaric oxygen or thiosemicarbazide in rats (Wood, Watson and Stacey, 1966), seizures induced by 3-mercaptopropionic acid (3-MPA) in mice (Seiler and Sarhan, 1984b; Toth, Lajtha, Sarhan and Seiler, 1983), seizures induced in DBA/2 audiogenic seizure-susceptible mice (Toth et al., 1983), strychnine-induced seizures in mice (Seiler and Sarhan, 1984b) and ky nurenine-induced seizures in mice (Lapin, 198 1). Although unimpressive when used alone, glycine consistently augments the anticonvulsant effects of G.L\BA-mimetic drugs in seizures induced by pharmacological impairment of the central gamma-aminobutyric acid (GABA) system. When glycine is combined with one of the GABA transaminase inhibitors, gamma-vinyl-GABA (GVG), ethanolamine-O-sulfate or gabaculline, against 3-MPA-induced seizures in mice, it exerts an anticonvulsant response which is greater than the response to either of the individual agents (Seiler and Sarhan, 1984a). Glycine also enhances the anticonvulsant effects of GVG in 3-MPAinduced seizures in rats (Seiler and Sarhan, 1984a). The anticonvulsant effects of 4,5,6,7-tetrahydroisoxa*To whom proofs and correspondence

should be sent.

zololo[4,5-Clpyridine-3-01, (THPO, an inhibitor of the reuptake of GABA) and muscimol are also enhanced by glycine in 3-MPA-induced seizures (Seiler and Sarhan, 1984a; Seiler, Sarhan, Krogsgaard-Larsen, Hjeds and Schousboe, 1985). In addition, glycine and GVG act synergistically to protect against bicuculline and pentylenetetrazol-induced seizures (Seiler and Sarhan, 1984b). Glycine also potentiates the activity of clinicallyeffective anticonvulsants. Glycine appears to potentiate the anticonvulsant action of phenobarbital and phenytoin in audiogenic seizure-susceptible DBA/2 mice, as well as 3-MPA-induced seizures in mice (Toth and Lajtha, 1984). Glycine enhances the anticonvulsant effect of phenobarbital in pentylenetetrazol-induced seizures in mice (Toth and Lajtha, 1984). In kindled amygdala seizures in rats, a dose of glycine with no anticonvulsant effect by itself significantly potentiates the anticonvulsant effect of phenobarbital and diazepam (Peterson, 1986). In vitro studies have demonstrated that glycine enhances the activity of the anticonvulsant MK-801 ([ +]-5-methyl-10,ll -dihydro-SH-dibenzo[a,d]cyclohepten-5,10-imine) (Robinson and Coyle, 1987); MK-801 is a noncompetitive antagonist of the excitatory amino acid N-methyl-D-aspartic acid (NMDA) receptor (Wong, Knight and Ransom, 1987) and is

399

S. L. PETERSON et al.

400

effective against maximal electroshock seizures (Clineschmidt, Martin and Bunting, 1982). Although the in vivo seizure tests, cited previously, indicate a GABA-glycine interaction, it is possible that glycine potentiates the NMDA antagonist properties of anticonvulsant drugs. The purpose of this study was to evaluate the interaction between glycine and clinically-effective anticonvulsants in maximal electroshock seizures in rats. The evidence presented above indicates that glycine has weak anticonvulsant properties by itself but potentiates the activity of anticonvulsant drugs. However, the relationship with the anticonvulsant drugs is not well characterized due to the diverse models of epilepsy which have been tested and the variation in doses of glycine, as well as routes and times of administration. The present experiment systematically evaluated the glycine-anticonvulsant interaction in maximal electroshock seizures, a standardized rodent model of grand ma1 epilepsy (Swinyard and Woodhead, 1982), with emphasis on consistent routes of administration, as well as the use of optimal doses and times of administration of drugs. In the situations where glycine was found to potentiate significantly the clinically-effective anticonvulsants, a further investigation was undertaken to determine whether the effect was the result of a pharmacokinetic interaction. METHODS

Animals Male, Sprague-Dawley rats, obtained from Harlan, Inc. and initially weighing 50-75 g, were used for this experiment. The animals were maintained in a climate-controlled vivarium on a 12 hr lightdark cycle and were allowed free access to food and water. Maximal electroshock and seizure response Maximal electroshock seizures were induced by passing a 60 cps, 150 mA and 0.2 set duration current generated by a Wahlquist stimulator (Salt Lake City, Utah) through saline-soaked cornea1 electrodes. Two endpoints were determined to evaluate the seizure response. The first endpoint was the abolition of tonic hindhmb extension. This was considered to occur when, after the stimulus, the hindlimbs failed to extend beyond a 90” angle to the torso. The second endpoint involved the timing of the stimulus-induced flexion and extension periods, so that the extension/flexion (E/F) ratio could be calculated (Swinyard, 1972). Flexion was considered to be the period from the end of the stimulus until the forelimbs extended to at least an angle of 90” to the torso. Extension was the period from the end of flexion until the ears returned to the normal position, indicating the end of the whole body tonus. The duration of the extension period did not depend on the occurrence of tonic hindhmb extension.

Determination of neurotoxicity Each animal was tested for neurological deficit during the 5 min period preceding each seizure test. A battery of 3 tests was administered to each animal and failure to pass 2 of the tests was taken as an endpoint of neurological deficit or drug-induced neurotoxicity (Swinyard and Woodhead, 1982). The neurological tests were as follows: (1) Rotorod. The animal was placed on a 6cm diameter rod that rotated at 8 rpm. Inability to remain on the rod for one minute during any of 3 separate trials was considered a failure; (2) Position Sense. The right hindlimb was lowered over the edge of a table and the inability to correct the position of the limb within 5 set was considered a failure; (3) Gait and Stance. The animal was placed on the surface of a table so that the gait could be observed and any circular or staggering gait, abnormal limb or torso posture, tremor, hyperactivity, somnolence or catalepsy was considered a failure. The administration of glycine and anticonvulsant drugs Glycine (Sigma, St Louis, Missouri) was administered orally by orogastric intubation as a 3 M solution in 0.9% saline. Glycine was administered at various times, prior to the seizure testing, as indicated in the Results. All anticonvulsants were administered in a volume of 4ml/kg weight by intraperitoneal injection at a previously determined time of peak effect (Swinyard 1982). Sodium phenobarbital and Woodhead, (Sigma) was dissolved in 0.9% saline while diazepam (Hoffman-LaRoche, Nutley, New Jersey), carbamazepine (Sigma), sodium divalproate (Abbott, North Chicago, Illinois), phenytoin (Sigma) and MK-801 (Merck, Sharpe and Dohme, West Point, Pennsylvania) were administered in 2% carboxymethylcellulose (Sigma). The doses of phenobarbital, divalproate and phenytoin were calculated as the free base of the drug. Anticonvulsant testing procedure The animals were randomly divided into groups of 10 and were tested for seizures twice. For the first seizure test, half of each group (n = 5) received a dose of the anticonvulsant assigned to the group, while the other half (n = 5) received the same dose of anticonvulsant plus glycine. For the second seizure test, which occurred 72 hr later, each half of the group received treatment with the drug that the other half received on the first test day. In this way each animal served as its own control when determining the interaction between glycine and a dose of anticonvulsant. In the experiment when the anticonvulsant effects of glycine by itself were determined, each animal was tested with 2 different doses of glycine during 2 seizure tests, 72 hr apart.

Glycine in maximal electroshock Analysis of the content of anticonvulsant, taurine in brain and serum

glycine and

After decapitation, the animals were exsanguinated into a collection vial. The blood was centrifuged in a Sorvall C60 at 15,000 g for 15 min, to separate the serum from the clot. Two 0.5 ml aliquots of serum were collected, one for determination of the content of glycine and taurine in serum and the other for the content of the anticonvulsant in serum. Both samples were frozen until assayed by the methods described below. After decapitation, the brain tissue was rapidly removed from the skull and the right parietal cortex, right hippocampus, cerebellum and brain stem were rapidly dissected free on an ice-cold petri dish. Each structure was sonically homogenized in distilled water (10 ~1 H,O/mg tissue) and frozen at -70°C until assayed for the content of glycine, taurine and anticonvulsant, as described below. Determinations of glycine and taurine were carried out using a modification of the method of Lindroth and Mopper (1979), with minor modifications (McGowan, Frye and Breese, 1986). Aliquots of 50 ~1 of homogenates of tissue were added to sufficient absolute ethanol to yield a 70% solution, homogenized and then centrifuged to remove precipitates of protein. Sample supernatants or known standard solution of taurine and glycine were spiked with S-aminovaleric acid solution, which served as an internal standard. Samples and standards were derivatized by reacting, for one minute, equal aliquots with o-phthaldialdehyde (OPT) solution (250 mg OPT in 2ml ethanol and 2 ml 2-mercaptoethanol, added to 46 ml 0.4 M borate buffer pH 10.4). After mixing in a Vortex mixer, lo-100 ~1 of the derivatized mixture were injected onto a C-18 reverse-phase high performance liquid chromatography (HPLC) column (Waters uBondapak). The amino acid-OPT derivatives were eluted using a two-pump gradient system (LDC/Milton Roy) that began with a mobile phase of 30% methanol/O.1 M NaH,PO, (apparent pH 6.0) and linearly increased the concentration of methanol to 70% over a period of 15 min. The amino acid derivatives were quantified by fluorometric detection (Kratos Schoeffel, FS 970 HPLC fluorometer), using an excitation wavelength of 330nm and recording all wavelengths above 413 nm. The content of protein of each sample was determined by the Bradford (1976) method using the BioRad Protein Assay Kit (BioRad, Richmond, California). Concentrations of glycine and taurine in supernatants from tissue were estimated from 6-point standard curves and expressed as nmol/mg of tissue protein or mmol/ml of serum. High performance liquid chromatography methods were also used to determine concentrations of phenobarbital and carbamazepine in serum and brain tissue (McDonald and Martan, 1978; Olson and Schmidt, 1980). Aliquots of 100~1 of supernatant of tissue,

401

serum or known solutions of anticonvulsant, were combined with 100 ~1 of glacial acetic acid, 50 ~1 of 5-(p-methylphenyl)-5-phenylhydantoin (used as the internal standard) and 1.0 ml of dichloromethane. The mixture was shaken for 2-3 min and then centrifuged for 5 min (1OOg). The aqueous phase was aspirated and discarded, after which the organic phase was evaporated to dryness under a stream of air in a 50°C water bath. The residual was dissolved in 20 ~1 of a mobile phase (16% methanol and 16% acetonitrile in 3.4 mM phosphoric acid). Ten p 1 of the mobile phase was injected onto a C-18 reversedphase column (Supercosil LC-18, Supelco 5-8230, 150 x 4.6 mm). The concentrations of the drugs were quantified by comparison with a 4-point standard curve, using peak-height ratios of the unknown and internal standard at 195 nm. Concentrations are expressed as pgg/mg of tissue protein or as pg/ml of serum. Statistical

analysis

Statistical analysis, between the E/F ratios produced by the anticonvulsants with and without glycine, was determined by multivariate and univariate repeated measures analysis of variance (ANOVA). In addition, a nonlinear regression analysis was performed on an IBM PC-XT computer, using the E/F ratios from the two sets of data generated from each combination of anticonvulsant and glycine. The regression program utilized a modified Gauss-Newton algorithm (Trzeciakowski, 1987) to fit the E/F dose-response data to the following model: E/F = M/(1 + (ED,,/[anticonvulsant

dose])s) + B (1)

where B refers to the E/F response, extrapolated to zero dose of anticonvulsant, M refers to the E/F level of response approached asymptotically as the dose of anticonvulsant was increased, ED,, refers to the dose of anticonvulsant required to decrease the E/F ratio to a level equal to (B - M)/2 and S refers to the slope of the dose-response curve. For each subject and treatment with an anticonvulsant, the data obtained in the absence (control) and presence of glycine were fitted simultaneously to equation (1). In subsequent analyses, 15 additional models were tested, in which the 4 parameters (M, ED,, B and S), either alone or in combination, were allowed to take on separate values for each of the two groups (control and glycine-treated). The null hypothesis assumed that the model with the least number of parameters provided the best fit of the data. Acceptance of the null hypothesis would indicate that there was no significant difference between the control and glycinetreated dose-response curves. Models containing additional parameters could displace simpler models in the ranking only if they were shown to improve

S. L. PETERsoN et al

402

0 hz9 eZa

significantly the fit of the data. This was determined by constructing the following F-test: F = K=,

- ~~,PF,IlK~F,

-

DF,W,I

(2)

In equation (2) SS refers to the error (residual) sum of squares, DF refers to the degrees of freedom and the subscripts 1 and 2 refer to the simpler and more complex models, respectively. Significance of the F value was tested at (DF, - DF,) and DF, degrees of freedom. On the E/F scale, a value of zero was considered as the maximum possible reduction in the intensity of seizures. When a treatment with drug blocked the

tonic
individual

means.

RESULTS

Maximal

electroshock

Hours after 9lyCine Fig. I. The effect of glycine on the E/F ratio. All animals were tested for seizures twice at intervals of 72 h. The 23 control animals received 2 consecutive untreated seizure tests. The glycine-treated rats received the same dose but were tested for seizures at 2 different times after administration. There were no significant (ANOVA) differences among the average E/F ratios of the groups.

(3)

where B refers to the baseline value, computed by the best-fitting version of equation (1). In this way, 100% protection would correspond to an E/F ratio of zero. Nonlinear probit analysis (Systat ver. 3.0, SYSTAT Inc., Evanston, Illinois) was employed to determine the TD,, and confidence intervals of the quanta1 neurotoxicity data. The Mantel-Hansel test was used to establish further statistical significance in differences between the neurotoxicity, induced by the anticonvulsant with and without glycine. The TD, used for calculation of the safety ratio was determined from the nonlinear probit analysis. Multivariate and univariate repeated measures ANOVA were used to determine significant differences in the content of glycine and anticonvulsant in the serum and various tissues of the brain. Scheffe’s multivariate contrasts were used to test differences between

untreated, control seizures (n- 23) 30 mmol/kg glycins (p.0.) Ln ~101 40 mmol/kg glycine (p.0.) In ~201

seizure response

The E/F ratio provided a consistent measurement of the severity of the maximal electroshock seizure response. As shown in Figure 1, there was no significant difference (ANOVA) between the E/F ratios in a group of untreated rats (n = 23) tested twice at the 72 hr interval. Although the E/F ratio was shown to be a reliable measurement of seizure response, tonic hindlimb extension was found to be a poor measurement. For example, of the 23 untreated rats, indicated in Figure 1, only 4 and 7 animals responded with tonic hindlimb extension in the first and second seizure tests, respectively. Therefore, due to the low frequency of tonic hindlimb extension, the E/F ratio was used to determine the seizure response through the experiment.

Anticonvulsant

activity of glycine

The administration of glycine alone induced no significant alterations in the maximal electroshock seizure response (Fig. 1). Neither the 30 nor 40 mmol/kg oral doses of glycine, administered 1, 2, 4 or 8 hr prior to the seizure test, significantly altered the E/F ratio when compared to each other, or the untreated rats (ANOVA). No neurotoxicity was induced by either dose of glycine at any time of administration. Larger doses were not tested because the volume of the orally-administered, near saturation solution of glycine was approaching the gastric capacity of the rats. To establish a time of peak effect for an interaction of glycine with anticonvulsants, the 40mmol/kg oral dose of glycine was tested with 10 mg/kg (i.p.) phenobarbital (approximate ED,,). As shown in Figure 2, glycine, administered 4 and 8 hr prior to the seizure test, significantly [F( 1,72) = 13.41 enhanced the anticonvulsant effect of phenobarbital (administered 30min prior to the seizure test). Glycine administered 1 or 2 hr prior to the seizure test did not enhance the anticonvulsant effect of phenobarbital. The E/F ratios induced by phenobarbital did not differ significantly at the 4 times tested. Therefore, to maximize the observation of an interaction between glycine and an anticonvulsant, should it occur, the 40 mmol/kg (3.0 g/kg) oral dose of glycine was administered 4 hr prior to the seizure test in all of the following experiments. Interaction

between glycine and anticonvulsant

Glycine (40 mmol/kg, p.o., 4 hr) significantly potentiated the anticonvulsant effect of phenobarbital (i.p., 0.5 hr). By nonlinear regression analysis, it was determined that the data best fitted the model for 2 curves with separate ED,, values (Fig. 3A) (F(1,15) = 39.71. The ED,, values for phenobarbital, with and without glycine, as determined by regression

Glycine in maximal electroshock 0 I

1

2

4

ofter

glycine

8

The effect of glycine on the E/F ratio, induced by IO mg/kg phenobarbital (PB). At a given hour after administration, the same 10 rats were treated with phenobarbital and tested with and without glycine. To control for possible diurnal variations in the effect of drugs, the animals not treated with glycine were tested for seizures within 30 min of those that were treated with glycine. Phenobarbital was administered 30 min prior to the seizure test. Note that phenobarbital induced an anticonvulsant effect, as the average E/F ratio indicated here, was approximately half of those shown in Figure 1. Glycine, administered 4 and 8 hr prior to the seizure test, signi~~ntiy (ANOVA) enhanced the anticonvulsant action of phenobarbital. Fig, 2.

(AI

_ . 0’

were 13.2and 10.7 mg/kg, respectively (Table 1). Further evaluation of the data by ANOVA for repeated measures determined that glycine induced significant differences in the E/F ratios at the 2.5, 7.5, 10 and 20mg/kg doses of phenobarbital (Fig. 3A). The phenobarbital-induced neurotoxicity was not significantly altered by glycine, as determined by the Mantel-Haensei test, as well as by the nonlinear probit analysis where the confidence intervals were found to overlap (Fig. 3A and Table 1). Glycine increased the protective index (TDS/EDSo) for phenobarbital from 2.7 to 3.4 and increased the safety ratio (ED,,/TD3) from 0.37 to 0.53 (Table 1). Glycine (40 mmol/kg, p.o., 4 hr) significantly potentiated the anticonvulsant effect of carbamazepine (i.p., 0.5 hr). Analysis of nonlinear regression determined that the data best fitted the model for 2 curves with separate ED,, values (Fig. 3B) [f;(1,9) = 24.53. The EDSo values for carbamazepine with and without glycine, as determined by the regression analysis, were 10.1 and 6.8 mg/kg, respectively (Table I). Additional evaluation by ANOVA for repeated measures determined that glycine induced significant differences in the E/F ratios at the 2.5, 5.0, 7.5, 10 and 15mg/kg doses of carbamazepine (Fig. 3B). The ~arbamazepine-induced neurotoxi~ity was enhanced significantly by glycine as determined by the analysis,

PB (10 mg/kg. Lp.. ~-10) PB + Gly(40 mmol/kg. p.o., n=10) t P ( 0.05

Hours

403

-

100

100

80

PB t GLY

1 I

00

10

Phenobarbital

G

r;

1

100

$0

Carbamazepine

(mg/kg)

100

(mg/kg)

(DI

(C) 100

E

0.0

MK-801

60

1 0.5

10

Diarepom

10.0

0.1

Imp/kg)

MK-601

Imp/kg)

Fig. 3. The effect of glycine (GLY) on the anticonvulsant action of (A) phenobarbital (PB), (B) carbamazepine (CBZ), (C) diazepam (DZP) and (D) MK-801. In all cases a 40 mmol/kg dose of glycine was administered orally, 4 hr prior to the seizure test. At a given dose of anticonvulsant, the same 10 rats were tested with and without glycine. Each animal was tested for neurological deficit during the 5 min period preceding each seizure test. Asterisks (*) indicate significant differences between the responses to a dose of anticonvulsant as determined by ANOVA. The percentage protection curves were determined by nonlinear regression analysis. using the E/F ratios. Although the percentage toxicity curves are presented using a percentage scale, nonlinear probit analysis was used to evaluate the statistical significance of the quanta1 neurotoxicity responses.

S. L. PETERSON

404

Mantel-Haensel test [x’(l) = 6.31 (Fig. 3B, Table 1). The TD,, for carbamazepine was reduced significantly from 56.9 to 45.0mg/kg, in the presence of glycine, as the confidence intervals did not overlap (Table 1). Although glycine reduced the TDSO for carbamazepine, the protective index was nevertheless increased from 5.6 to 6.6 and the safety ratio (ED,,/TD,) increased from 0.66 to 1.18 (Table 1). Diazepam was found to be only partially effective against maximal electroshock seizures, at doses that induced neurotoxicity. However, the anticonvulsant effect of diazepam (i.p., 1 hr) was enhanced by glycine (40 mmol/kg, p.o., 4 hr) as nonlinear regression analysis determined that the data best fitted the model for 2 curves with separate EDSo values [F(l,7) = 36.41 (Fig. 3C). The analysis also determined that the ED,, values for diazepam, with and without glycine, were 3.7 and 7.0 mg/kg, respectively (Table 1). The ANOVA for repeated measures established significant glycine-induced differences in the E/F ratios at the 2.5, 5.0, 15 and 20 mg/kg doses of diazepam (Fig. 3C). Glycine did not significantly alter the diazepaminduced neurotoxicity, but enhanced the protective index for diazepam from 1.4 to 2.2 (Table 1). A safety ratio could not be calculated because an ED,, was not found. Glycine (40 mmol/kg, p.o., 4 hr) significantly potentiated the anticonvulsant effect of MK-801 (i.p., 1.Ohr). By nonlinear regression analysis it was determined that the data best fitted the model for 2 curves with separate ED,, values (Fig. 3D) (F(1,9) = 201. The ED,, values for MK-801, with and without glycine, were 0.17 and 0.19 mg/kg, respectively (Table 1). Further analysis of the data by ANOVA for repeated measures determined that glycine induced significant differences in the E/F ratios at the 0.0625, 0.125 and 0.25 mg/kg doses of MK-801 (Fig. 3D). The MK-801-induced neurotoxicity was not significantly altered by glycine (Fig. 3D, Table 1). Glycine induced very little improvement in the poor protective index of MK-801. Glycine did not significantly affect the anticonvulsant activity of phenytoin or sodium divalproate in maximal electroshock seizures (data not shown). In

et

al,

both cases the analysis by nonlinear regression established that none of the proposed models provided a significantly better fit of the data than the assumption of all common parameters. Evaluation of the E/F ratios by ANOVA for repeated measures determined that there were no significant differences between the doses with and without glycine for phenytoin (7 doses between 3.75 and 45.0 mg/kg, i.p., 1 hr) and for only one dose (300 mg/kg) of sodium divalproate (6 doses between 50 and 400 mg/kg, i.p., 0.25 hr). The effect of glycine on the neurotoxicity of phenytoin and sodium divalproate were not tested because glycine did not significantly affect the therapeutic effect of the drugs. Concentrations of glycine and anticonvulsant in brain and serum

Since potentiation by glycine of responses to anticonvulsant drugs might result from pharmacokinetic interactions, the content of glycine and the anticonvulsants in brain and serum were determined in untreated rats, as well as in animals treated with glycine, anticonvulsant or both. The doses of phenobarbital and carbamazepine administered were the ED,,‘s determined from the dose-response curve for glycine and anticonvulsant, determined by the nonlinear regression analysis. The animals were treated with glycine and anticonvulsant drug using the same times and routes of administration as during the seizure testing. Analysis of the glycine (40 mmol/kg, p.o., 4 hr) and phenobarbital (10 mg/kg, i.p., 0.5 hr) test groups indicated a significant main effect of glycine [F(1,14) = 306.21 and a significant interaction between glycine and tissue [F(4,76) = 311.81. The Scheffe multiple comparison procedure established significant increases in the content of glycine in the serum and cerebellum in both the glycine and glycine plus phenobarbital groups, as compard to the control group (Fig. 4A). No significant effects of the phenobarbital or interaction between glycine and phenobarbital on the content of glycine in the various tissues were observed. There were no significant changes in the content of taurine in the various tissues

Table I. Interaction of glycine with the anticonvulsant and neurotoxic effects of phenobarbital (PB). carbamazepine (CBZ), diazepam (DZP) and MK-801 in maximal electroshock seizures HOW PB PB plus glycine CBZ CBZ plus glycine DZP DZP plus glycine MK-801 MK-801

0.5 0.5 4.0 0.5 0.5 4.0 1.0 1.0 4.0 I.0 I.0

plus glycine

4.0

ED, fmeike)

TD,, (meike)

PI

Safety ratio

13.2

35.2 (33.1-37.2)

2.7

0.37

10.7’ 10.1

36.2 (34.7-37.8) 56.9 (54.4-59.4)

3.4 5.6

0.52 0.66

6.8’ 7.0

45.0’ (43.846.2) 9.6(8.1-l 1.2)

6.6 I.4

I.18 -

3.7* 0.19

8.2 (6.8-9.8) 0.124(0.119XI.128)

2.2 0.7

-

0.17’

0.156(0.113~).121)

0.9

-

Protective index (PI) was calculated as the TD,,/EDs, safety ratio as TD,/ED,, *Significantly different from the anticonvulsant, administered alone.

Glycine in maximal electroshock in the glycine and phenobarbital test group (data not shown). For the glycine (40 mmol/kg, p.o., 4 hr) and carbamazepine (6.8 mg/kg, i.p., 0.5 hr) test groups, significant main effects of glycine [F(1,20) = 148.91 and carbamazepine [F(1,20) = 5.91, as well as significant interactions between glycine and tissue [F(4,80) = 128.71, carbamazepine and tissue [F(4,80) = 6.21 and glycine, carbamazepine and tissue [F(4,80) = 3.31 were observed. As observed in the phenobarbital groups, the a posteriori tests established that glycine, administered alone, significantly increased the content of glycine in the cerebellum and serum, as compared to the control group (Fig. 4B). The treatment with glycine plus carbamazepine resulted in a significantly increased content of glycine in the

405

hippocampus, brain stem and serum, as compared to the control group (Fig. 4B). The treatment with glycine plus carbamazepine significantly increased the content of glycine in the hippocampus but reduced the content of glycine in the cerebellum and serum as compared to the glycine-treated rats (Fig. 4B). There was also a significant interaction between carbamazepine and tissue [F(4,80) = 4.31, involving the content of taurine in the brain tissues. Specifically, the treatment with glycine and carbamazepine increased the content of taurine in the hippocampus, as compared to the control group (370.6 f48.6 vs 250.6 f 4.5 ng/mg protein) and decreased the content of taurine in the cerebellum, as compared to the glycine-treated animals (170.4 f 10.2 vs 206.4 + 4.7 ng/mg protein).

(A) 0

con

rnPlY

m

Pb

m

*

r-b + gb

P c 0.05

“S con *

T

*

5000

1 4000

3000

2000

1000

E

CTX

k f ._ 4

0 CER

HPC

-

BS

Serum

: a! In z .

(B) 0 con m Cbz

m

clr

*

P c 0.05

“J con

1x3

cbz + gly

*

P c 0.05

“S ply

Q .E 0 a 5000

d 5 E

4mO

3000

2000

1000

>

0 CTX

HPC

CER

BS

serum

Fia. 4. The effect of glycine (gly), administered alone and in combination with (A) phenobarbital (pb), or
406

S. L. PETERSON et al.

For the glycine and phenobarbital test groups, a significant main effect of glycine [F(l,lO) = 8.51 interaction between glycine and tissue [F(4,40) = 5.71 was found with a reduced content of phenobarbital in the serum of rats treated with glycine plus phenobarbital (Table 2). Glycine had no significant effect on the content of carbamazepine in the brain tissues or serum in the glycine and carbamazepine test groups (Table 2).

their mechanism of action in maximal electroshock seizures. Glycine did not alter the distribution of either phenobarbital or carbamazepine in the brain. Although the administration of glycine was associated with a decreased concentration of phenobarbital in the serum, this was not reflected by any significant reductions in the content of phenobarbital in the regions of the brain tested. Therefore, the potentiation by glycine of phenobarbital and carbamazepine was probably not the result of a pharmacokinetic interaction, that produced an enhanced concentration of those anticonvulsants in brain. Phenobarbital did not systematically alter the distribution of glycine in brain to an extent that would explain the potentiated anticonvulsant effect. After the administration of glycine an increase in glycine in the serum and cerebellum was observed regardless of the presence or absence of phenobarbital. The content of glycine in the other regions of the brain was unaffected by phenobarbital. Administered alone, glycine did not produce an anticonvulsant effect, so it is doubtful that the increase in glycine in the cerebellum, observed with and without phenobarbital, contributed to the enhanced response to phenobarbital. Carbamazepine appeared to produce specific alterations in the central distribution of glycine. As observed in the phenobarbital and glycine test groups, glycine administered alone increased the content of glycine in both the serum and the cerebellum. When carbamazepine and glycine were administered together, there was still a significant increase in glycine in the serum but the increase was less than when glycine was administered alone. Carbamazepine also prevented the glycine-induced increase in glycine in the cerebellum. Since the combination of carbamazepine and glycine induced a greater anticonvulsant effect than the administration of either compound alone, the carbamazepine-induced decrease in the content of glycine in the cerebellum supports the conclusions drawn from the data for phenobarbital, suggesting that the content of glycine in the cerebellum is not relevant to the control of maximal electroshock seizures. Perhaps the most interesting finding was that the administration of carbamazepine and glycine together resulted in significant increases in the content of glycine in the hippocampus and the brainstem. The increases in the

DISCUSSION

Orally administered glycine was found to potentiate the anticonvulsant effects of phenobarbital, carbamazepine and diazepam in maximal electroshock seizures in rats. Glycine alone (40 mmol/kg) induced no neurotoxicity or significant anticonvulsant effect, which indicates that the enhancement of the response by anticonvulsant drugs was the result of a potentiation or synergistic interaction. The potentiation of phenobarbital appeared to be selective (Swinyard and Woodhead, 1982) since both the protective index and the safety ratio were increased by glycine and because the TD,, was not significantly altered by glycine. Glycine also potentiated carbamazepine and although glycine significantly reduced the TD,,, the protective index and safety ratio were increased, which indicates that the selectivity of carbamazepine altered. Because the anticonvulsant was not potentiation did not appear to be the result of a glycine-induced alteration in the distribution of phenobarbital or carbamazepine in the regions of the brain evaluated, the results provide evidence for a glycine-sensitive component in the anticonvulsant mechanism of action of phenobarbital and carbamazepine in maximal electroshock seizures. Diazepam has been reported to be effective in maximal electroshock seizures but only at doses associated with minimal neurotoxicity (Chweh, Swinyard, Wolf and Kupferberg, 1985; Krall, Penry, White, Kupferberg and Swinyard, 1978). In the present study, glycine enhanced the anticonvulsant effect but did not improve the selectivity of diazepam. Because of the apparent lack of a specific anticonvulsant mechanism, the pharmacokinetic interactions between glycine and diazepam were not investigated. The lack of interaction between glycine and either phenytoin or sodium divalproate suggests that those drugs do not have a glycine-sensitive component in

Table 2. Effect of glycine on the concentration of phenobarbital hitmocamous (HP0 cerebellum

(PB) and carbamazepine (CBZ) in serum, cortex (CTX), (CER) and brainstem (BS)

Anticonvulsant Serum (pglml) PB PB plus glycine CBZ CBZ plus glycine

14.3 * 0.4 13.2 & 0.2* 1.7~0.111 2.0 * 0.088

CTX 0.63 0.62 0.079 0.105

F k f f

0.05 0.02 0.007 0.022

administered

in brain tissue (pg/mg

HPC 0.68 0.65 0.055 0.077 alone.

+ k k +

0.05 0.09 0.006 0.008

protein)

CER 0.65 0.62 0.061 0.060

k i i k

0.3 0.2 0.005 0.007

BS 0.65 0.60 0.085 0.088

f + + f

0.03 0.03 0.005 0.005

Glycine in maximal electroshock content of glycine in those regions of the brain were associated with a content of glycine in the serum that was significantly less than that when glycine was administered alone. These findings might indicate a selective accumulation of glycine induced by carbamazepine in the hippocampus and brainstem. Since the hippocampus has the greatest density of strychnine-insensitive glycine binding sites, while the brainstem has the greatest density of binding sites for strychnine (Bristow, Bowery and Woodruff, 1986), the enhanced content of glycine in those areas may be related to the enhanced anticonvulsant effect of carbamazepine. Carbamazepine and phenobarbital produced little or no change in the content of taurine in the central nervous system (CNS). Although taurine is known to be anticonvulsant (Huxtable, 1981), it is doubtful that changes in the content of taurine in the CNS played a significant role in the potentiation by glycine of carbamazepine or phenobarbital. However, the finding does indicate that although glycine enters the brain by a nonsaturable, nonspecific carrier mechanism (Oldendorf and Szabo, 1976) the large doses administered in this experiment may not have distorted the levels of other neuronally important amino acids in brain. It has been hypothesized that glycine potentiates anticonvulsant drugs that act by a GABAergic mechanism, since glycine has been shown to potentiate anticonvulsants with GABAmimetic mechanisms, particularly when tested against convulsive agents that impair GABA systems (Seiler and Sarhan, 1984a; Seiler and Sarhan, 1984b; Seiler et al., 1985). Recent in uivo studies suggest that facilitation of GcZBA-mediated neurotransmission is involved in the anticonvulsant mechanism of action of diazepam (C’hweh et al., 1985). Although the results of the present study would support a potentiation by glycine of the diazepam-induced facilitation of GABA-mediated neurotransmission, it would not appear to be a selective anticonvulsant effect, since the response was associated with significant neurotoxicity. In vitro experiments also provide evidence that an augmentation of GABA-responses is associated with the anticonvulsant effects of phenobarbital (Macdonald and Barker, 1979; Schultz and Macdonald, 1984). However, in uivo experiments indicate that selective anticonvulsant effects of phenobarbital may occur independently from GABA-mediated neurotransmission (Ulloque, Chweh and Swinyard, 1986) so the potentiation by glycine of phenobarbital found in the present experiment, may not be mediated at the G4BA receptor complex. Any hypothesis of a glycine-GABA interaction in maximal electroshock seizures is further weakened by the results of the experiments comparing the administration of glycine, together with sodium divalproate and carbamazepine. Valproic acid is proposed to possess Gj4BA-mimetic properties by increasing the content of GABA in nerve terminals (Kupferburg, 1985;

407

Loscher, 1985), yet this agent was not potentiated by glycine. Carbamazepine is not thought to possess GABA-mimetic properties (Swinyard, White and Wolf, 1988), yet it was potentiated by glycine. Thus, these data do not provide support for a glycineinduced enhancement of the efficiency of the GABA receptor-chloride channel complex transducer, as an important mechanism in the potentiation by glycine of anticonvulsant drugs in maximal electroshock seizures. An alternative hypothesis concerning potentiation by glycine of anticonvulsants proposes an interaction of glycine with NMDA receptors. This is based largely on the observation that MK-801 was potentiated by glycine in the maximal electroshock seizures, an effect that would be predicted from previous in vitro studies (Reynolds and Miller, 1988; Wong et al., 1987). Although there is currently no direct evidence of a clinically-effective interaction of an anticonvulsant drug with the NMDA receptor complex, phenobarbital has been shown to antagonize glutamate in vitro (Macdonald and Barker, 1979) and to inhibit glutamate-induced seizures in mice (Stone and Javid, 1983). In addition, MK-801 potentiates the anticonvulsant activity of phenobarbital in electroshock seizures (Kulkarni and Ticku, 1989), suggesting an interaction involving the NMDA receptor-cation channel complex. However, any such hypothesis concerning the involvement of the NMDA receptor-cation channel complex in the observed potentiation by glycine of anticonvulsant drugs must remain speculative until such time as specific antagonists of the NMDA-coupled glycine receptor and the strychnine-sensitive glycine receptors have been tested. This is particularly important in view of the recent report that glycine may have a proconvulsant action at NMDA receptors in spinal convulsions of mice (Larson and Beitz, 1988). Tonic hindlimb extension was found to be an unreliable measure of the severity of maximal electroshock seizures in Sprague-Dawley rats. The stimulus parameters used in the present experiment were initially reported to induce tonic hindlimb extension in approximately 95% of Sprague-Dawley rats (Laffan, Swinyard and Goodman, 1957). More recently, the percentage of Sprague-Dawley rats that display tonic hindlimb extension has been shown to be much smaller (Buterbaugh, 1978; Novack, Stark and Peterson, 1979) and reportedly varies between vendors and shipments (Browning and Nelson, 1985). The E/F ratio is an alternative means of determining the severity of maximal electroshock seizures (Swinyard, 1972), where an increase in duration of the extension phase and a decrease in the duration of the flexion period indicate the reduced severity of seizures (Laffan et al., 1957). Numerous studies have measured the duration of various phases of electroshock-induced seizures to evaluate the severity of electroshock seizures (Ramer and Pinel, 1976; Browning, Turner, Simonton and Bundman, 1981;

408

s. L. F'ETERSONet d.

Browning and Nelson, 1985), as well as the effect of anticonvulsant drugs on the electroshock seizure response (Novack et al., 1979; Berman and Adler, 1984; McNamara, Russell, Rigsbee and Bonhaus, 1988; Tortella, Ferkany and Pontecorvo, 1988). Nevertheless, there may be significant differences between the effects of anticonvulsant drugs on the E/F ratio and the incidence of tonic hindlimb extension. Such differences might account for the lack of an interaction between glycine and phenytoin, reported in the present study and the significant potentiation of phenytoin by glycine as demonstrated in mice, using the occurrence of tonic hindlimb extension as the measure of the severity of seizures (Toth and Lajtha, 1984). Studies are currently underway to test this hypothesis. Drugs which are effective against maximal electroshock seizures have excellent clinical efficacy against grand maf convulsions (Krall et al., 1978; Swinyard and Woodhead, 1982). The results of this study indicate that glycine can enhance the effect of phenobarbital and carbamazepine in grand mal epilepsy, without increasing the incidence of neurological side effects. This may be of particular interest in epilepsies that are controlled only by toxic doses of anticonvulsant drugs. Analogs of glycine and glycine are essentially nontoxic to humans, as daily doses in totals of grams have been administered with no apparent side effects (McGeer, Eccles and McGeer, 1978; Roach and Carlin, 1982). In that regard, the glycine prodrug milacemide, produces an impressive increase in the seizure threshold of hyperbaric oxygen-induced seizures (Youdim, Kerem and Dudevani, 1988) and should be tested for the potentiation of anticonvulsant drugs. With the identification of specific anticonvulsants, that are potentiated by glycine, future studies will determine whether the chronic administration of glycine can further enhance phenobarbital and carbamazepine in maximal electroshock seizures. Acknowledgemenu-The

authors acknowledge the gifts of diazepam from Hoffman-LaRoche, sodium divalproate from Abbott and MK-801 from Merck, Sharpe and Dohme. We thank Terry Baumgart for the anticonvulsant analysis and Annette Fincher for the amino acid analysis. We are especially grateful to Lisa Boehnke and Ruth Riegel for their diligent, conscientious assistance during the seizure and neurotoxicity testing. This work was suported by Public Health Service grants NS 24566 and AA 06322. REFERENCES

Berman E. F. and Adler M. W. (1984) The anticonvulsant effect of opioids and opioid peptides against maximal electroshock seizures in rats. Neurophartnacology U: 367-371.

Bradford M. (1976) A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Analyt. Biochem. 72: 248-254.

Bristow D. R., Bowery N. G. and Woodruff G. N. (1986) Light microscopic autoradiographic localization of

[‘Hlglycine and [‘HIstrychnine binding sites in rat brain. Eur. J. Pharmac. 126: 303-307. Browning R. A. and Nelson D. K. (1985) Variation in threshold and pattern of electroshock-induced seizures in rats depending on site of stimulation. Life Sci. 37: 2205-2211. Browning R. A., Turner F. J., Simonton R. L. and Bundman M. C. (1981) Effect of midbrain and pontine tegmental lesions on the maximal electroshock seizures pattern in rats. Epilepsia 22: 583-594. Buterbaugh G. G. (1978) Effect of drug modifying central serotonergic function on the response of extensor and nonextensor rats to maximal electroshock. Life Sci. 23: 2393-2404. Chweh A. Y., Swinyard E. A., Wolf H. H. and Kupferberg H. J. (1985) Effect of GABA agonists on the neurotoxicity and anticonvulsant activity of benzodiazepines. Life Sci. 26: 737-744. Clineschmidt B. V., Martin G. E. and Bunting P. R. (1982) Anticonvulsant activity of (+ )-5-methyl-l&I I-dihydroSH-dibenzo la.dl cvcloheuten-5.10-imine (MK-801). a substance wiih’ potent an&convulsant, central sympathomimetic, and apparent anxiolytic properties. Drug Deoef. Res. 2: 123-124. Huxtable R. J. (1981) Insights of function: Metabolism and pharmacology of taurine in the brain. In: The Role of

Peptides

and

Amino

Acids

as

Neurotransmitters

(Lombardini J. B. and Kenny A. D., Eds), pp. 53-97. Alan R. Liss, New York. Krall R. L., Penry J. K., White B. G., Kupferberg H. J. and Swinyard E. A. (1978) Antiepileptic drug development: II. Anticonvulsant drug screening. Epilepsia 19: 409-428. Kulkarni S. K. and Ticku M. J. (1989) Interaction between gabaergic anticonvulsants and the NMDA receptor antagonist MK-801 against MES- and picrotoxin-induced convulsions in rats. Life Sci. 44: 1317-1323. Kupferberg H. J. (1985) Sodium valproate. In: Antiepileptic Drugs: Mechanisms of Action (Glaser G. H., Penry J. K. and Woodbury D. M., Eds), pp. 6433654. Raven Press, New York. Laffan R. J., Swinyard E. A. and Goodman L. S. (1957) Stimulus intensity, maximal electroshock seizures, and potency of anticonvulsants in rats. Arch. Int. Pharmacodyn. 111: 6&69.

Lapin I. P. (1981) Antagonism of glycine to seizures induced by L-kynurenine, quinolinic acid and strychnine in mice. Eur. J. Pharmac. 71: 495498.

Larson A. A. and Beitz A. J. (1988) Glycine potentiates strychnine-induced convulsions: role of NMDA receptors. J. Neurosci. 8: 3822-3826. Lindroth P. and Mopper K. (1979) High performance liquid chromatographic determination of subpicomolar amounts of amino acids by precolumn fluorescence derivatization with o-phthaldialdehyde. Analyt. Chem. 51: 1667-1674. Loscher W. (1985) Valproic acid. In: Handbook of Experimental Pharmacology (Frey H. H. and Junz D. Eds), pp. 507-536. Springer, Berlin. Macdonald R. L. and Barker J. L. (1979) Anticonvulsant and anesthetic barbiturates: different postsynaptic actions in cultured mammalian neurons. Neurology 29: 432-447. McGowan T. J., Frye G. D. and Breese G. (1986) Evidence for site specific ethanol actions in the CNS. Alcohol Drug Res. 6: 423-429.

McDonald D. M. and Martan L. J. (1978) A simultaneous high performance liquid chromatographic analysis of the most common anticonvulsants and their metabolites in the serum. J. Analyt. Toxicol. 2: 127-133. McGeer P. L., Eccles J. C. and McGeer E. G. (1978) Molecular Neurobiology of the Mammalian Brain, p. 229. Plenum Press, New York.

Glycine in maximal electroshock McNamara J. O., Russell R. D., Rigsbee L. and Bonhaus D. W. (1988) Anticonvulsant and antiepileptic actions of MK-801 in the kindling and electroshock models. Neuropharmacology 27: 563-568.

Novack G. D., Stark L. G. and Peterson S. L. (1979) Anticonvulsant effects of benzhydryl piperazines on maximal electroshock seizures in rats. J. Pharmac. exp. Ther. 208: 480484.

Oldendorf W. H. and Szabo J. (1976) Amino acid assignment to one of the three blood-brain amino acid carriers. Am. J. Physiol. 230: 94-98.

Olson D. C. and Schmidt G. J. (1980) The analysis of NBS antiepilepsy drug standards using liquid chromatography. Chromat. Newslett. 8: 15-17.

Peterson S. (1986) Glycine potentiates the anticonvulsant action of diazepam and phenobarbital in kindled amygdaloid seizures in rats. Neuropharmacology 25: 1359-1363.

Ramer D. and Pine1 J. P. J. (1976) Progressive intensification of motor seizures produced by periodic electroconvulsive shock. Expl Neurol. 51: 421433. Reynolds I. J. and Miller R. J. (1988) Multiple sites for the regulation of the N-methyl-p-aspartate receptor. Molec. Pharmac. 33: 581-584.

Roach E. S. and Carlin C. (1982) N,N-Dimethylglycine for epilepsy. N. Engl. J. Med. 307: 1081-1082. Robinson M. B. and Coyle J. T. (1987) Glutamate and related acidic excitatory neurotransmitters: from basic science to clinical application. FASEB J. 1: 446455. Schultz D. W. and Macdonald R. L. (1984) Barbiturate enhancement of GABA-mediated inhibition and activation of chloride ion conductance: correlation with anticonvulsant and anesthetic actions. Brain Res. 20% 1777188. Seiler N. and Sarhan S. (1984a) Synergistic anticonvulsant effects of a GABA agonist and glycine. Gen. Pharmac. 15: 367-369. Seiler N. and Sarhan S. (1984b) Synergistic anticonvulsant effects of GABA-T inhibitors and glycine. Arch. Pharmac. 326: 49-57.

Seiler N., Sarhan S., Krogsgaard-Larsen P., Hjeds H. and Schousboe A. (1985) Amplification by glycine of the anticonvulsant effect of THPO, a GABA uptake inhibitor. Gen. Pharmac. 16: 509-511.

409

Stone W. E. and Javid M. J. (1983) Effects of anticonvulsants and other agents on seizures induced by intracerebral L-glutamate. Brain Res. 264: 165-167. Swinyard E. A. (1972) Electrically induced convulsions. In: Experimental Models of Epilepsy (Purpura D. P., Penry J. K., Tower D., Woodbury D. M. and Walter R., Eds), pp. 434456. Raven Press, New York. Swinyard E. A., White H. S. and Wolf H. H. (1988) Mechanisms of anticonvulsant drugs. ISI Arias of Science: Pharmacology pp. 95-98. Swinyard E. A. and Woodhead J. H. (1982) Experimental detection, quantification, and evaluation of anticonvulsants. In: Antiepileptic Drugs (Woodbury D. M., Penry J. K. and Pippenger C. E., Eds), pp. 111-126. Raven Press, New York. Tortella F. C., Ferkany J. W. and Pontecorvo M. J. (1988) Anticonvulsant effects of dextrorphan in rats: possible involvement in dextromethorphan-induced seizure protection. Lfi Sci. 42: 2509-2514. Toth E. and Lajtha A. (1984) Glycine potentiates the action of some anticonvulsant drugs in some seizure models. Neurochem. Res. 12: 1711-1718. Toth E., Lajtha A., Sarhan S. and Seiler N. (1983) Anticonvulsant effects of some inhibitory neurotransmitter amino acids. Neurochem. Res. 8: 291-302. Trzeciakowski J. P. (1987) Inhibition of guinea pig ileum contractions mediated by a class of histamine receptor resembling the H, subtype. J. Pharmac. exp. Ther. 243: 874-880.

Ulloque R. A., Chweh A. Y. and Swinyard E. A. (1986) Effects of gamma-aminobutyric acid (GABA) receptor agonists on the neurotoxicity and anticonvulsant activity of barbiturates in mice. J. Pharmac. exp. Ther. 237: 468472.

Wong E. H. F., Knight A. R. and Ransom R. (1987) Glycine modulates [3H]MK-801 binding to the NMDA receptor in rat brain. Eur. J. Pharmac. 142: 487488. Wood D. J., Watson J. W. and Stacey E. N. (1966) A comparative study of hyperbaric oxygen-induced and drug-induced convulsions with particular reference to GABA metabolism. J. Neurochem. 13: 361-370. Youdim M. B. H., Kerem D. and Duvdevani Y. (1988) The glycine-prodrug, milacemide, increases the seizure threshold due to hyperbaric oxygen; prevention by ldeprenyl. Eur. J. Pharmac. 150: 381-384.