Convulsive action and toxicity of uremic guanidino compounds: behavioral assessment and relation to brain concentration in adult mice

96

Journal of the NeurologicalSciences, 112 (1992) 96-105 © 1992 Elsevier Science Publishers B.V. All rights reserved 0022-510X/92/$05.00

JNS 03858

Convulsive action and toxicity of uremic guanidino compounds: behavioral assessment and relation to brain concentration in adult mice R. D'Hooge ~, Y.-Q. Pei b, B. Marescau ~ and P.P. De Deyn

a

a Laboratory of Neurochemistry, Born-BungeFoundation, Universityof Antwerp (UIA), and Department of Neurology, Middelheim GeneralHospital, Antwerp, Belgium, and b Department of Pharmacology at Beijing Medical University, Beijing, People'sRepublic of China (Received 13 January, 1992) (Revised, received 4 May, 1992) (Accepted 14 May, 1992)

Key words: Guanidino compounus; Convulsants; Toxins; Chromatography; Epilepsy; Uremia; Mouse

Summary Four guanidino compounds that are known to accumulate in uremia, namely creatinine, guanidine, guanidinosuccinic acid and methylguanidine, were administered intraperitoneally and intracerebroventricularly to adult albino mice and the compounds epileptogenic and toxic properties were behaviorally assessed. After intraperitoneal injection, brain concentration of the compounds as a function of injected dose was monitored additionally. Guanidino compound brain concentration was determined by cation exchange chromatography with fluorescence ninhydrin detection. After systemic administration, especially guanidinosuccinic acid and methyiguanidine induced long-lasting generalized convulsions which gradually increased in severity. Increasing the dose injected intraperitoneally resulted in linear increase in brain concentration of the injected compounds, in parallel with increase in proportion of animals presenting with convulsions and/or severity of convulsions. Guanidinosuccinic acid brain concentration increased more slowly than that of the other 3 compounds and guanidinosuccinic acid also exerted its effect later than the others. Since none of the other metabolically related guanidino compounds determined was significantly increased in the brains of the injected animals, the observed behavior was most certainly induced by the compounds injected and not by some secondary metabolite. Epileptogenic properties of the four compounds were markedly and qualitatively different in systemic administration, but rather similar in intracerebral administration. A tentative epiieptogenic potency order was inferred from the combined behavioral and biochemical results. All 4 of the compounds tested displayed the ability to induce full-blown clonic-tonic convulsions and they did so in a dose-related manner. Guanidinosuccinic acid appeared to be slightly more potent than methylguanidine, but both guanidinosuccinic acid and methylguanidine were considerably more potent than guanidine. Creatinine was many times less potent than the other 3 guanidino compounds. Revised epileptogenic potency order on the basis of guanidino compound brain concentration after systemic administration as well as potency order after intracerebral administration paralleled the potency order of these compounds in their GABA antagonism reported earlier. It was therefore postulated that the GABA antagonism of uremic guanidino compounds could underlie their epileptogenic character. Moreover, these compounds could very likely be at the basis of the neurological complications including epilepsy of uremic patients in whom they accumulate in physiological fluids and brain.

Introduction Commonly known as guanidino compounds, many biologically active molecules contain one or more basic

Correspondence to: Dr. P.P. De Deyn, Laboratory of Neurochemistry, Born-Bunge Foundation, Universityof Antwerp (UIA), Universiteitsplein 1, B-2610 Antwerp, Belgium. Tel: (32-3) 820-2620; Fax: (32-3) 820-2248. Abbreviations: G, guanidine; GSA, guanidinosuccinic acid; MG, methylguanidine;CTN, creatinine.

guanidino groups (carbon atoms surrounded by 3 amino functions). Over 100 of these guanidino compounds have been found to occur in plants and animals naturally, and more than 10 different low molecular weight guanidino compounds were identified in the animal brain (Robin and Marescau 1985). The convulsive action and toxicity of guanidino compounds was reported as early as 1886 with the discovery of methylguanidine in decayed horse meat, which could cause dyspnea, muscle fibrillation and generalized convulsions (see Mori 1987). Since then, many guanidino compounds

97 were found to induce behavioral convulsions as well as epileptiform EEG discharges experimentally. These compounds include 3,-guanidinobutyric acid (Jinnai et al. 1966), guanidinoacetic acid, creatine, creatinine and creatine phosphate (Jinnai et al. 1969), a-N-acety!arginine (Mori and Ohkusu 1971), methylguanidine (Matsumoto et al. 1976), a-guanidinoglutaric acid (Shiraga and Mori 1982), a-keto-8-guanidinovaleric acid (Marescau et al. 1983), homoarginine (Yokoi et al. 1984) and 2-guanidinoethanol (Edaki et al. 1986). Guanidino compounds, which are generated as a result of protein and amino acid metabolism, depend upon renal function for their excretion. In patients with renal insufficiency or uremia, creatinine (CTN), guanidine (G), guanidinosuccinic acid (GSA), and methylguanidine (MG) accumulate in serum and cerebrospinal fluid (De Deyn et al. 1987; Marescau et al. 1989). As uremic patients often suffer from a wide range of neurological complications including epilepsy, the convulsive action and toxicity of guanidino compounds could hypothetically be one of the underlying causes of these complications. Moreover, De Deyn and Macdonald (1990) applied CrN, G, GSA and MG to mouse spinal cord neurons in primary dissociated cell culture and found that these uremic guanidino compounds inhibited the responses of the neurons to yaminobutyric acid (GABA) and glycine. Inhibition of GABAergic neurotransmission has indeed been suggested to underlie epilepsy (De Deyn et al., 1990; Schousboe 1990), but, since neurons cannot convulse as such, behavioral analysis of the actions of these four putative uremic toxins is still highly incomplete. We decided to perform this study dealing with the first characterization of convulsions induced by systemic and intracerebral injections of CTN, G, GSA and MG in adult mice. The convulsive action of GSA is reported for the first time, and the relative potencies of the 4 compounds are compared. In contrast to many other studies on the behavioral properties of convulsants, behavioral response is related to brain concentration of the compounds. Materials and methods

Experimental animals Prior to their use in the experiments, male and female Swiss Webster mice were kept under standard environmentally controlled conditions (12-h light/dark cycle, constant room temperature ;and humidity). For systemic administration of guanidino compounds, mice weighing 18-25 g were used, and for intracerebral administration we used full-grown 30-40 g mice. Chemicals and their administration Creatine hydrochloride (MW= 149.6), guanidine hydrochloride (MW = 95.5), guanidinosuccinic acid

(MW--175.2) and methylguanidine hydrochloride ( M W - 109.6) for systemic and intracerebral injection were purchased from Sigma Chemical Co. (St. Louis, USA). Guanidino compounds were also used as standards for guanidino compound determination: GSA, G, MG, guanidinoacetic acid, a-N-acetylarginine, argininic acid, fl-guanidinopropionic acid, 3,-guanidinobutyric acid, arginine and homoarginine were all purchased from Sigma Chemical Co. (St. Louis, USA), creatine and creatinine (CTN) from Merck (Darmstadt, Germany). a-keto-8-Guanidinovaleric acid was synthesized enzymatically according to the method of Meister (1954), which was modified as described earlier (Marescau et al. 1991). All other chemicals used were obtained from Merck and were of analytical grade. For intraperitoneal (i.p.) administration, GSA concentrations which exceeded its water solubility had to be used. Therefore, smooth suspensions of GSA in 30% polyethylene glycol solution were made with an agate mortar. The 3 other guanidino compounds easily dissolved in the polyethylene glycol solution. Concentrations were always calculated in such a manner that each animal received 0.! ml suspension or solution per i0 g body weight. Groups of 5-10 animals were injected i.p. with different doses of the compounds: CTN between 7 and 55 mmol/kg, G between 1 and 20 mmol/kg, GSA between 1 and 6 mmol/kg, a,ld MG between 0.5 and 9 mmol/kg. Each series of in~ctions commenced with a dose of maximal effect. For intracerebroventricular (i.c.v.) administration, the compounds were dissolved in saline (pH around 7.0 for all concentrations used) and delivered in a volume of 5 /~1 with a Hamilton microsyringe into the left lateral brain ventricle according to a technique modified from Herman (1975). Briefly, the animal's scalp was cut and the exposed skull was pierced with a small stainless steel drill (50/zl of a I% lidocaine solution was used as local anesthetic). The hole in the skull was made rather rostrally (I mm posterior to the coronal suture and 1 mm left to the sagittal suture) to avoid hippocampal damage. The animals were allowed at least 3 h rest for recovery and wearing off of the local anesthetic. An injection cannula, mounted on the microsyringe and fitted with a nylon stopper to ensure an injection depth of 3 mm, was slid through the hole perpendicular to the skull surface. While the animals were restrained by hand, the 5/zl volume was administered at an injection rate of I izl/5 sec and the cannula was kept in position for about 15 sec more. Post-mortem injection of black dye through the same hole in random control animals (n > 20) always revealed a homogenous filling of the lateral brain ventricles. Groups of 5-10 animals were injected i.c.v, with different doses of the compounds: CTN between 20 and 300/zmol/kg, G between 4 and 15/zmol/kg, GSAbetween 0.5 and 4.0 /zmol/kg, and MG between 0.5 and 7.0 /zmol/kg.

98

Each series of injections commenced with a dose of maximal effect, thereafter doses were decreased until zero effect. Each mouse was only used once for i.p. as for i.c.v, injection.

Systematic observation Immediately after injection of the guanidino compounds, the animals were placed in individual cylindrical plastic cages for observation of their behavior. Behavioral assessment of the physiological state of the mouse was performed according to Irwin (1968). The character, onset time latency, evolution and duration of presumed epileptic activity were noted within a 1-h observation period. As shown by Veliskovfi et al. (1990), systematic observation of convulsive behavior is much aided by the use of a scoring procedure. We assigned the following scores according to the severity of the convulsive response of the mouse: 0 = normal behavior; 1 - slight local spasmodic movements and twitching, most often in the mask, vibrissae, mouth and pinnae; 2 = more intense, but short lasting and isolated jerking, myoclonus and "pop-corn jumps", hyperactivity, rearing and slight jumping; 3 -- continuous, generalized convulsive activity lasting >__5 sec: jumping without resting, running fits and automatisms; 4 = generalized clonic attack: vigorous and incoordinated clonic movements of the body and limbs with falling and loss of righting reflex; 5 - tonic extension, often fatal to the animal.

Tissue preparation and determination of guanidino compound brain concentration Mice with guanidino compounds administered i.p. were decapitated immediately following the 1-h observation period (or earlier if death occurred within this period). Whole brains were quickly removed, rinsed in ice cold saline and stored at -75°C until analysis of guanidino compound brain concentration. Before analysis, brain tissue was homogenized at 0°C with a Potter homogenizer in 1 ml water and the obtained homogenate was preprocessed according to Marescau et al. (1986). Concentrations of CTN, G, GSA and MG as well as those of a-keto-8-guanidinovaleric acid, creatine, guanidinoacetic acid, a-N-acetylarginine, argininic acid, /3-guanidinopropionic acid, y-guanidinobutyric acid, arginine and homoarginine were determined according to an earlier described method (Marescau et al. 1992), modified from Hiraga arid Kinoshita (1981), with a Biotronik LC 5001 amino acid analyzer adapted for guanidino compound determination. Cation-exchange chromatography for separation of the guanidino compounds was followed by fluorescence ninhydrin detection using a Spectroflow 980 fluorometer (ABI Analytical Kratos Division, Ramsey, NJ) at an excitatory wave length of 301 nm, and an emission filter of 500 nm.

Statistics Convulsant and lethal doses in 50% of the animals (CDso and LDso, respectively) and their 95% confidence intervals for each of the 4 guanidino compounds and for both administration routes were determined by probit analysis or moving average interpolation (Finney 1971; Well 1952). Results were further analyzed according to Litchfield and Wilcoxon (1949). Significance levels were set at 5%.

ResuLts

Systemic administration of uremic guanidino compounds Solutions of CTN, G and MG and suspensions of GSA were administered i.p. to groups of 5-10 albino mice (Fig. 1). Although of different potency, all four chemicals shared the ability to induce generalized convulsions (score > 3). Highest applied doses of the compounds were always lethal to the animals and mortality was behaviorally related to respiratory depression and to convulsive action of the compounds since the animals died from severe clonic convulsions or from tonic extension and asphyxia. Even high doses of phenobarbital (80 mg/kg) did not block the clonic convulsions, but did protect the animals from the occurence of tonic convulsions and from the lethal effects induced by GSA and MG (data not shown). Dose-dependence was observed for each one of the chemicals. The injected dose of the guanidino compounds determined in effect the proportion of animals presenting with clonic convulsions. Sham i.p. injections of 0.1 ml 30% polyethylene glycol solution per 10 g body weight did not produce any lasting change in behavior (n = 10). CTN was clearly the least potent of the compounds tested both with respect to its convulsive action as to its toxicity. Threshold concentration to observe generalized convulsions (score 3) in some of the animals was more than 14 mmol CTN/kg. Doses of 7 or 14 mmol CTN/kg did produce stupor and behavioral suppression in some animals, but these doses did not induce convulsions nor behavioral toxicity. However, even doses over 14 mmol/kg did not induce vigorous convulsions. The animals only displayed slight clonic convulsions (score 3) preceded by a short excitatory phase lasting about 3 min and characterized by jumping and myoclonus. When such a reaction pattern occurred, the animals died with severe dyspnea within 10 min following CTN injection. While doses of 1 mmol G/kg failed to induce generalized clonic convulsions (score > 3), doses higher than 5 mmol/kg did produce severe clonic contractions with loss of righting reflex (score 4). About 7 min after injection of 10 mmol G/kg, the animals began breathing heavily with open mouth. Local spasmodic jerking of head and body rapidly increased to generalized

99 clonic contractions in limbs and body, which caused the animal to fall on its side, while the incoordinated movement of the animal's extremities continued. Eventually, the animals died, like those that were injected with CTN, 9-10 min after G injection with severe dyspnea but without tonic extension. Epileptogenic properties of GSA and MG were markedly more straightforward than those of CTN or G. However, the character of the convulsions induced by GSA was different from that of those induced by MG. Unless the animals died from severe clonic convulsions or from tonic extension, convulsions induced by these two compounds lasted far beyond the standard observation period, up to 3 h or more. With a latency of 5-10 min after i.p. injection of GSA suspension, doses of 6 mmol GSA/kg induced spasmodic twitches in face and limbs of the animals. About 15 min after the injection the animals suddenly started jumping up and down against the walls of the observatory cages with increasing vigor and intensity. This jumping developed into a running fit or into a full-blown clonic attack with loss of righting reflex, vigorous incoordinated clonie contractions in limbs and body, and sometimes rolling around the longitudinal axis. The animals often calmed down after this initial phase, but proceeded with typical generalized clonic convulsions. These generalized clonic convulsions comprised automatic crawling or grasping movements of forelimbs and hands with the head jerked backwards, hind legs broadly placed and toes spread. Often, the animals would foam at the mouth, some even bit their tongue, defecated and micturated, fell down but rose again or rolled around their axis. Between 30 and 50 min after

injection, the animals displayed another vigorous clonic attack (score 4), sometimes accompanied by vocalization and leading to tonic extension and death (score 5). Decreasing the dose of GSA decreased the severity of the convulsions as well as the proportion of animals displaying convulsions. Convulsive doses below 6 mmol GSA/kg induced typical generalized clonic convulsioas only (score 3 or 4), with suppressed tonic phase. A dose of 5 mmol MG/kg, i.p. injected as a solution, led to increased activity, rearing and local spasmodic movements of mask, head and sometimes limbs, 6 to 8 rain after injection. About 20-35 min after injection, a sudden vigorous clonic attack with loss of righting reflex occurred, often accompanied by vocalization, micturition and defecation. Typical generalized clonic convulsions, like those induced by GSA, commenced, regularly interrupted with intervals of 3-6 min by vigorous clonic attacks. Some of the animals died during second or third attack some with, some without tonic extension. For each of the four compounds, CDs0 for the induction of generalized convulsions (score > 3) and LDs0 were inferred from the behavioral observations (see Table 1). The potency of the compounds in the induction of generalized convulsions decreased in the order: MG > GSA > G > CTN. Potency ratios were as follows: between MG and GSA, 1.3; between GSA and G, 1.5; between G and CTN, 7.8 (values are ratios of highest CDs0 on lowest CDs0). Only CDs0 of CTN was significantly higher than that of G; the difference between CDs0 of MG, GSA and G were numerical, not statistical. GSA was the only compound able to induce clonic-tonic convulsions in a dose-dependent fashion

TABLE 1 CDso AND LDso OF 4 UREMIC GUANIDINO COMPOUNDS AFTER I.P. INJECTION WITH CORRESPONDING MEDIAN LATENCY AND BRAIN CONCENTRATION, AND AFTER I.C.V. INJECTION Data represent CDs0 (dose which causes generalized convulsions, score > 3, in 50% of the animals) and LD5o (lethal dose in 50% of the animals), calculated after i.p. and i.c.v, administration of uremic guanidino compounds to groups of 5-10 mice, with 95% confidence interval between parentheses and total number of animals per compound and per administration route between brackets. Media latency and brain concentration corresponding with CDs0 and LDso after i.p. injection of each of the four compounds are inferred by linear interpolation as described in the text. CTN

G

GSA

MG

Median latency (min) Brain concentration (nmol/g tissue)

25 (20-31) (n = 20) 12.7 1328 (1104- 1604)

3.2 (2.6-4.0) (n = 30) 11.5 183 (152-225 )

2.1 (1.6-2.6) (n = 35) 42.8 56 (42--70)

1.6 (1.2-2.2) (n = 35) 18.1 94 (74-123 )

LD50 (mmol/kg) Median latency (min) Brain concentration (nmol/g tissue)

25 (20-31) 12.7 1328 (1104-1604)

3.7 (3.0-4.6) 14.5 209 (173-256)

3.3 (2.5-4.3) >_60 90 (67-118)

4.3 (3.5-5.3) 40.2 225 (186-274)

1O1 (77-133) (n -- 33) 153 (109-215)

5.0 (3.8-6.6) (n = 25) 10.0 (8.3-12.1)

0.8 (0.5-1.2) (n = 30) 1.8 (1.3-2.6)

1.0 (0.6-1.5 ) (n = 25) 3.0 (1.9-4.6)

i.p. injected: CD5o (mmol/kg)

i.c.v, injected: CDso (~mol/kg) LDso (/zmol/kg)

100 after systemic administration. Thus, CD50 of 3.5 (2.74.6) mmol GSA/kg (95% confidence interval between parentheses) was calculated for the induction of fullblown clonic-tonic convulsions (score 5) by i.p. injections of GSA. LDs0 was lowest for GSA and highest for CTN. Toxicity order decreased as follows: GSA > G > MG > CTN. LDs0 of GSA, G and MG differed numerically,

CTN

$0'

O' 10 2

10 3

10 4

10 5

G

100 , , ~ ~

S°l~o . . .i . . . . . '"i0, ~ ~ O S A

~" i = I '

0 |0

l0

10 3

10 4

M/Os

~00

SO

0 10 3

!0 S

IP INJECTED (in I~mol/kg)

Fig. 1. Dose-dependent increase in severity of generalized convulsions and/or number of mice presenting with convulsions after intraperitoneai (i.p.) injection of one of the 4 uremic guanidino compounds: creatinine (top graph), guanidine (upper middle graph), guanidinosuccinic acid (lower middle graph) and methylguanidine (bottom graph). Abscissa depicts dose of compound injected i.p. in /zmol/kg, ordinate depicts proportion of animals presenting with generalized convulsions in % of total number of 5 to 10 animals per dose. White areas correspond to score 3 generalized convulsions, light shaded areas with score 4 clonic attacks, and dark shaded areas with score 5 clonic-tonic convulsions (scores are alloted as described in the text). Each series of injections commenced with a dose of maximal effect, thereafter doses were decreased until zero effect (doses above those of maximal effect did not increase the severity of the convulsions but killed the animals quickly after the injection; these doses are not shown in the figure).

not statistically; LD50 of CTN was significantly higher than those of the three other compounds. Median latency after i.p. injection of the compounds decreased with increasing doses of the guanidino compounds for the appearance of generalized convulsions (Fig. 2). Latency for the appearance of generalized convulsions (score > 3) was markedly longer for GSA than for the three other guanidino compounds. Median latency corresponding to the CDs0 of a compound was estimated by linear interpolation (see Table 1). Interpolated median latency of GSA was more than 2 times longer than that of the other compounds.

Brain concentration following i.p. injection of uremic guanidino compounds Brain concentrations of CTN, G, GSA and MG as well as those of 9 other metabolically relevant guanidino compounds were determined 1 h (or earlier if death occurred within the observation period) after i.p. injection of the four uremic guanidino compounds. In 5 control animals, sham i.p. injection of 30% polyethylene glycol solution (vehicle) yielded baseline values of guanidino compound brain concentration. CTN was present in the highest brain concentration: 144 + 7 nmol/g tissue; G in a concentration of 0.7 + 0.2 nmol/g tissue; GSA in a concentration of 0.4 + 0.2 nmol/g tissue; M G i n aconcentration of 0.3+0.1 nmol/g tissue (all values are means + SEM). Systemic injection of each of the four uremic guanidino compounds tested resulted in a linear increase in the brain concentration of the injected compound (Fig. 3). None of the 9 other guanidino compounds increased significantly during the hour following i.p. injection of one of the experimental guanidino compounds. Except in the case of CTN, differences in median mortality latency (parallel with the increase in injected dose) did not significantly alter the linearity of the increase in brain concentration of the compounds. Since the increase in brain of CTN within 1 h after i.p. injection seemed to level off beyond 15 mmol CTN/kg, these high doses were not included in our analysis of CTN brain concentration kinetics. Thus, very high correlation coefficients were calculated for the 4 compounds: R 2-- 0.998 for CTN; R 2 - 0.996 for G; R 2= 0.997 for GSA; R 2 - 0.989 for MG. Slopes were very similar for CTN, G and MG (47.6, 51.6 and 48.6 respectively) indicative of parallel kinetics. In contrast, slope of GSA brain concentration increase was markedly smaller (27.9) than those of the other three compounds. Brain concentrations corresponding to CDso and LD5o of the compounds were estimated by li~l~ar interpolation (see Table 1). Thus, revised epileptogenic potency order was constructed on the basis of brain concentrations: GSA > MG > G > CTN. Potency ratios were as follows: between GSA and MG, 1.68;

101 .~

60

!

.SO"

.~

3o10 !

i0 IP INJECTED (in mmol/kg)

100

Fig. 2. Dose-dependent decrease in median latency of appearance of score > 3 generalized convulsions after intraperitoneal (i.p.) injection of one of the 4 uremic guanidino compounds: creatinine (e), guanidine (o), guanidinosuccinic acid (m) and methylguanidine (El). Abscissa depicts dose of compound injected i.p. in mmol/kg, ordinate depicts median latency in min in a group of 5-10 mice per dose.

between MG and G, 1.95; between G and CTN, 6.35 (values are ratios of biggest CDso on smallest CDso). The difference between CDso of GSA and that of MG was numerical, not statistical. All other differences were statistically significant. Revised toxicity order decreased as follows: GSA > G > MG > CTN. Only the difference between LDso of G and MG was not statistically significant.

Intracerebral administration of uremic guanidino compounds CTN, G, GSA and MG were administered i.c.v, to groups of 5-10 albino mice (Fig. 4). Sham i.c.v, injections of saline produced generalized convulsions in one

animal only (total n > 25). Although again of different potency, all four compounds induced generalized convulsions (score ~ 3) dose-dependently. Latency was usually very small and did not seem to be influenced markedly by compound concentration.

Usually, convulsions appeared 20-30 sec after injection of the guanidino compounds. Low doses caused only local spasmodic movements and twitches in head, body and limbs of the animals. Higher doses led to an initial excitatory phase with fast circling in the direction opposite to the injection side, running fits and jumping up and down the walls of the observatory cages. Depending upon the dose injected, this initial phase increased to severe clonic convulsions with loss of righting reflex, increased salivation, micturition and defecation. Several clonic attacks occurred which sometimes led to tonic e~ension and death by asphyxia. Also, the animals often died from prolonged clonic convulsions. The severity of the convulsions as well as the proportion of animals presenting with generalized convulsions increased with increasing concentrations of the guanidino compounds injected i.c.v. The character of the convulsions induced by i.c.v, injections of each of the 4 experimental guanidino compounds was rather similar. Unless the animals died earlier, convulsions lasted 1 h or more. CDs0 for induction of generalized convulsions (score > 3) and LDs0 were inferred from the behavioral observations (see Table 1). Potency of the compounds in the induction of generalized convulsions decreased in the order: GSA > MG > G > CTN. The difference in CDs0 of GSA and that of MG was numerical, not statistical. All other differences were statistically significant. Potency ratios were as follows: between GSA and MG, 1.25; between MG and G, 5.0; between G and CTN, 20.2 (values are ratios of biggest CDs0 on smallest CDs0). Although of rather similar potency, GSA induced tonic convulsions more readily than MG. Toxicity after i.c.v, injection decreased in the order: GSA > MG > G > CTN. Only the difference between LDs0 of GSA and that of MG was not statistically significant.

1060

1OO0

Z = m.~g:

0

S

lO

i$ ;

5

100

lid INJECTED {in mmol/kg)

Fig. 3. Linear increase in brain concentration after intraperitoneal (i.p.) injection of one of the 4 uremic guanidino compounds: creatinine (e), guanidine (o), guanidinosuccinic acid ( I ) and methylguanidine (O). Abscissa depicts dose of each compound injected i.p. in mmol/kg, ordinate depicts brain concentration of the compound in nmol/g brain tissue. Brain concentrations are means + gEM of 5 mice per dose, gEMs smaller than the symbols are not shown.

102 Discussion

All 4 of the compounds tested displayed the ability to induce full-blown clonic-tonic convulsions and they did so in a dose-related manner. Since none of the other guanidino compounds was significantly increased in the brains of the injected animals, the observed behavior is most certainly induced by the compounds injected and not by some seconc .~ metabolite. InCTN 100

SO

0

~ 10 2

10 3

10 4

l0 s

10 6

G

_

5o

0

m

........

,

........

,

........

6

i! i I 0 10 2

10 3

10 4

l0 s

IN 6

ICV INJECTED(in nmol/kg) Fig. 4. Dose-dependent increase in severity of generalized convulsions and/or number of mice presenting with convulsions after intracerebroventricular (i.c.v.) injection of one of the 4 uremic guanidino compounds: creatinine (top graph), guanidine (upper middle graph), guanidinosuccinic acid (lower middle graph) and methylguanidine (bottom graph). Abscissa depicts dose of compound injected i.c.v, in nmol/kg, ordinate depicts proportion of animals pr~benQ,ng with generalized convulsions in % of total number of 5-10 animals per dose. White areas correspond to score 3 generalized convulsions, light shaded areas with score 4 clonic attacks, and dark shaded areas with score 5 clonic-tonic convulsions (scores are alloted as described in the text). Each series of injections commenced with a dose of maximal effect, thereafter doses were decreased until zero effect (doses above those of maximal effect did not increase the severity of the convulsions but killed the animals quickly after the injection; these doses are not shown in the figure).

creasing the dose injected i.p. resulted in linear increase in brain concentration of the injected compound, in paral.lel with increase in proportion of animals presenting with convulsions and/or severity of convulsions. However, only GSA and M,G can be considered systemic full-spectrum convulsants. If we compare the effects of GSA or MG with those of classical systemic chemical convulsants, important differences catch the eye (Stone 1972). After systemic administration, GSA and MG induce convulsions which gradually increase in severity and which last for many hours. A much used systemic convulsant like pentylenetetrazole also induces clonic-tonic convulsions, but with rapid onset and short duration. The unique epileptogenic properties of GSA and MG, which are, of course, endogenous compounds, may prove them useful for modelling epilepsy in future research of this disorder. CTN was significantly the least epileptogenic and the least toxic of the four guanidino compounds tested. Jinnai et al. (1969) applied creatinine intracisternally to rabbits and found that a dose of 13 mg CTN/kg induced clonic and tonic convulsions. We reported a CD5o after i.c.v, injection of 101 /zmol/kg (or 15.1 mg/kg) for the induction of generalized clonic convulsions in mice and higher doses induced tonic convulsions as well. Thus, CTN appears to induce clonic and tonic convulsions in rabbits and mice in similar concentrations. Systemic administration of CTN did not produce very severe clonic nor tonic convulsions in mice. In the case of CTN, i.p. administration of the compound appears to kill the animals before full expression of the convulsive potential of CTN. On the other hand, the relatively low convulsive action and toxicity of CTN could be a possible reason why this guanidino compound is tolerated in such high physiological concentrations. Indeed, in control animals CTN was found to be present in brain concentrations more than 200 times higher than those of G, GSA and MG. Although our values are somewhat higher than the mouse brain concentrations reported by Marescau et al. (1986), these authors also found this compound to be one of the 3 most abundant guanidino compounds in mouse, rat, rabbit and human brain, second to arginine and the metabolically related creatine. MG appeared to be more potent than GSA in the induction of generalized clonic convulsions after i.p. injection. However, GSA was injected as a suspension and is much less soluble than MG. Its systemic resorption must have been slower than that of MG and indeed, its brain concentration slope was less steep than those of CTN, G or MG, which were injected as solutions. Also indicative of the slow and difficult resorption of GSA is its long latency in the appearance of generalized clonic convulsions after i.p. injection as compared to that of the three other guanidino compounds tested. As i.p. injections of equal amounts of

103 GSA and MG results in a GSA brain concentration lower than that of MG, the actual potency order based upon brain concentration of the compounds 1 h after injection presents GSA as the most potent of the 4 compounds tested. Brain concentration corresponding tc CDso of GSA for induction of generalized clonic convulsions was not significantly smaller than that of MG, but GSA appeared to be slightly more potent than MG in the induction of generalized clonic convulsions after i.c.v, injections as well. Also, GSA was markedly more potent than MG in the induction of tonic convulsions after i.p. injection, but latter difference in potency could be due to differences in systemic resorption of the compounds. A tentative epileptogenic potency order can be inferred from the combined behavioral and biochemical results. Epileptogenic potency order appeared to decrease as follows: GSA > MG > G > CTN. De Deyn and Macdonald (1990) reported that CTN, G, GSA and MG rapidly and reversibly reduced GABA and glycine responses of mouse spinal cord neurons in cell culture. The potency order of these guanidino compounds in the inhibition of GABA responses reported by these authors is the same as the one inferred from the behavioral results presented in this paper. As the potency order of the compounds in the behavioral induction of epilepsy seems to mimic the potency order of these compounds in the inhibition of GABA responses, this latter inhibition could hypothetically underlie the convulsive action of these compounds. It has indeed often been suggested that inhibition of GABAergic neurotransmission leads to epilepsy and that many antiepileptic drugs agonize GABAergic function (Macdonald and Meidrum 1989; De Deyn et al. 1990; Schousboe 1990). For example, the convulsant pentylenetetrazole was shown to selectively antagonize GABA-mediated postsynaptic inhibition in cultured mammalian neurons (Macdonald and Barker 1977). Even in kindling induced epileptogenesis, which was thought to be based primarily upon increased excitatory amino acid neurotransmission, decrease in inhibition by GABA on the excitatory responses seems to play an important part (Kamphuis et a!. 1991). Moreover, other guanidino compounds, such as 8-guanidinovaleric acid and guanidinoethanesulfonic acid, are thought to induce convulsions through direct action on the GABA A receptor (Herranz et al. 1990; Mori 1987). In the case of CTN, G, GSA and MG, it was suggested that the inhibition of GABA and glycine responses by these guanidino compounds is due to blocking of the chloride channel associated with the inhibitory amino acid receptors (De Deyn and Macdonald 1990). Although our results could indicate predominant involvement of GABAergic neurotransmission in the epileptogenesis after injection of CTN, G, GSA and MG, the involvement of other neurotransmitter sys-

terns canpot be excluded. Indeed, GSA was found to be 10 times more potent than MG in the inhibition of GABA responses (De Deyn and Macdonald 1990), but GSA was only 1.25-1.68 times more potent than MG in behavioral epileptogenesis. This difference in potency could be due to differential effects of the compounds on neurotransmitter systems other than the GABAergic system. For example, MG was shown to inhibit acetylcholinesterase (Mori 1987), and GSA decreased excitatory amino acid mediated neurotransmission in rat hippocampal slices (D'Hooge et al. 1991). However, the effect of many guanidino compounds, including those studied here, upon neurotransmitter systems remains to be investigated further (Mori 1987). Evidently, GSA was the most toxic of the 4 uremic guanidino compounds tested. As we only observed for one hour, our results only give a clue as to the acute toxicity of these compounds, and chronic administration could give a totally different picture. Indeed, Yokozawa et al. (1989) gave daily i.p. injections of CTN, GSA and MG to rats with adenine-induced renal failure. As these authors have shown, chronic toxicity of MG was much higher than that of GSA, and CTN was many times less toxic than both GSA or MG. Unfortunately, Yokozawa et al. (1989) did not report upon the possible neurological effects of chronic administration of these uremic guanidino compounds. Certainly in the case of GSA and MG after i.p. injection, and for all 4 compounds after i.c.v, injection, acute toxicity of the compounds was related to their convulsive action since the animals died from prolonged clonic convulsions or from tonic extension, and since they could be.protected from the lethal effects of GSA and MG by pretreatment with phenobarbital. Our results also indicate toxic effects not primarily related to convulsions, and toxicity potency order was not parallel with epileptogenic potency order. Other systemic and neuronal effects could play a role in the toxicity of the compounds. For instance, Stein et al. (1968) first suggested the possible involvement of GSA in uremic bleeding, and Lonergan et al. (1971) showed that GSA decreases erythrocyte transketolase activity. Accumulation of toxic guanidino compounds was suggested to contribute to the complex neurological and hematological complications of patients with renal insufficiency (De Deyn et al. 1987). Of the different guanidino compounds De Deyn et al. analyzed in serum and cerebrospinal fluid of non-dialyzed uremic patients, CTN, G, GSA and MG levels were at least 10 times higher than those of control subjects. As yet, neurological complications in uremia, including epilepsy, remain unexplained (De Deyn 1989; Tyler 1975). The results presented here indicate that especially GSA displays its convulsive action and toxicity at similar concentrations as found in brain of uremic patients with neurological syrnptomatology (De Deyn

104

1989). Therefore, we postulate that these guanidino compounds could contribute to this symptomatology. In this paper, convulsions induced by systemic and intracerebral injections of the uremic guanidino compounds CTN, G, GSA and MG have been behaviorally assessed in adult albino mice and related to their brain concentration. It was postulated that the GABA antagonism of the uremic guanidino compounds could underlie their epileptogenic character. Moreover, these compounds could very likely be at the basis of the neurological complications including epilepsy of uremic patients in whom they accumulate in physiological fluids (Cohen 1970; De Deyn et al. 1987; Marescau et al. 1989). Thus, demonstrated toxicity of some of these guanidino compounds together with their clinical relevance should inspire further research into their pathophysiology and into possible ways to antagonize their action. Acknowledgements We acknowledge the expert technical assistance of I, Possemiers and F. Franck. Financial support was obtained from "Ministerie van Onderwijs van de Vlaamse Gemeenschap", University of Antwerp, Born-Bunge Foundation, OCMW Medical Research Foundation, United Fund of Belgium and National Fund of Scientific Research (NFWO grants D 11606 and CD F57).

References Cohen, B.D. (1970) Guanidinosuccinic acid in uremia. Arch. Intern. Med., 126: 846-850. De Deyn, F.P. (1989) Analytical studies and pathophysiologicai importance of guanidino compounds in uremia and hyperarginine. mia. Thesis, University of Antwerp (UIA). De Deyn, P.P. and R.L. Macdonald (1990) Guanidino compounds that are increased in cerebrospinal fluid and brain of uremic patients inhibit GABA and glycine responses on mouse neurons in cell culture. Ann. Neurol., 28: 627-633. De Deyn, P.P., B. Marescau, JJ. Cuykens, L. Van Gorp, A. Lowenthai and W.P. De Potter (1987) Guanidino compounds in serum and cerebrospinal fluid of non-dialyzed patients with renal insufficiency. Clin. Chim. Acta, 167: 81-88. De Deyn, P.P., B. Marescau and R.L. Macdonald (1990) Epilepsy and the GABA hypothesis. A brief review and some examples. Acta Neurol. Belg., 90: 65-81. D'Hooge, R., J. Manil, F. Colin and P.P. De Deyn (1991) Guanidinosuccinic acid inhibits excitatory synaptic transmission in CAI region of rat hippocampal slices. Ann. Neurol., 30: 622-623. Edaki, A., Y. Watanabe, I. Yokoi, H Shirags, T Fujii and A. Mori (1986) 2-Guanidinoethanol.induced convulsive seizures in rat. Neurosciences, 12: 106-107. Finney, DJ. (1911) Statistical Methods in Biological Assay, Griffin Press, London. Herman, Z.S. (1975) Behaviourai changes induced in conscious mice by intracerebroventricular injection of catecholamines, acetyl. choline and 5-hydroxytryptamine. Br. J. Pharmacol., 55: 351-358. Herranz, A.S., J.M. Soils, O. Herreras, N. Men~ndez, E. Ambrosio, L.M. C'r~nsanz and R. Martin del Rio (1990) The epileptogenic action of the taurine analogue guanidinoethanc sulfonate may be caused by a blockade of GABA receptors. J. Neurosci. Res., 26: 98-104. Hiraga, Y. and T. Kinoshita (1981) Post-column derivatization of

guanidino compounds in high-performance liquid chromatography using ninhydrin. J. Chromatogr., 226: 43-51. Irwin, S. (1968) Comprehensive observational assessment: la. A systematic, quantitative procedure for assessing the behavioral and physiologic state of the mouse. Psychopharmacologia, 13: 222-257. Jinnai, D., A. Sawai and A. Mori (1966) ~,-Guanidinobutyric acid as a convulsive substance. Nature, 212: 617. Jinnai, D., A. Mori, J. Mukawa, H. Ohkusu, M. Hosotani, A. Mizuno and L.C. Tye (1969) Biochemical and physiological studies on guanidino compounds induced convulsions. Jpn. J. Brain Physiol., 106: 3668-3673. Kamphuis, W., J.A. Gorter and F. Lopes da Silva (1991) A long-lasting decrease in the inhibitory effect of GABA on glutamate responses of hippocampal pyramidal neurons induced by kindling epileptogenesis. Neuroscience, 41: 425-431. Litchfield, J.T. and F. Wilcoxon (1949) A simplified method of evaluating dose-effect experiments. J. Pharmacol. Exp. Ther., 96: 99-113. Lonergan, E.T., M. Semar, R.B. Sterzel, G. Treser, M.A. Needle, L. Voyles L. and K. Lange (1971) Erythrocyte transketolase activity in dialyzed patients: a reversible metabolic lesion in uremia. N. Engl. J. Med., 284: 1399-1403. Macdonald, R.L. and J.L. Barker (1977) Pentylenetetrazole and penicillin are selective antagonists of GABA-mediated post-synaptic inhibition in cultured mammalian neurones. Nature, 267: 720-721. Macdonald, R.L. and B.S. Meldrum (1989) Principles of antiepileptic drug action. In: Levy, R., Mattson, R., Meldrum, B., Penry, J.K. and Dreifuss, F.E. (Eds.), Antiepileptic Drugs, Raven Press, New York, NY, pp. 59-83. Marescau, B., M. Hiramatsu and A. Mori (1983) a-Keto-& guanidinovaleric acid-induced electroencephalographic epileptiform discharges in rabbits. Neurochem. Pathol., 1: 203-209. Marescau, B., P.P. De Deyn, P. Wiechert, L. Van Gorp and A. Lowenthal (1986) Comparative study of guanidino compounds in serum and brain of mouse, rat, rabbit, and man. J. Neurochem,, 46:717-720. Marescau, B., P.P. De Deyn, P. Wiechert, M. Hiramatsu, L. Van Gorp, W,P. De Potter and A. Lowenthal (1989) Guanidino com. pounds in serum and cerebrospinal fluid of epileptic and some other neurological patients. In: Mori, A., Cohen, B.D. and Koide, H. (Eds.), Guanidines 2, Plenum Press, New York and London, pp. 203-212. Marescau, B., D.R. Deshmukh, M. Kockx, I. Possemiers, I.A. Qureshi, P. Wiechert and P.P. De Deyn (1992) Guanidino compounds in serum, urine, liver, kidney and' brain of man and some ureotelic animals. Metabolism, 41: 1-6. Matsumoto, M., K. Kobayashi, H. Kishikawa and A. Mori (1976) Convulsive activity of methylguanidine in cat and rabbits. IRCS Med. Sci., 4: 65. Meister, A. (1954) The a-keto analogs of arginine, ornithine and lysine. J. Biol. Chem., 206: 577-583. Mori, A. (1987) Biochemistry and neurotoxicology of guanidino compounds. History and recent advances. Pay. J. Biol. Sci., 22: 85-94. Mori, A. and H. Ohkusu (1971) Isolation of a-N-acetyI-L-arginine from calf brain and convulsive seizure induced by this substance. Adv. Neurol. Sci., 15: 303-306. Robin, Y. and B. Marescau (1985) Natural guanidino compounds. In: Mori, A,, Cohen~ B.D. and Lowenthal, A. (Eds.), Guanidines, Plenum Press, New York and London, pp. 383-438, Schousboe, A. (1990) Neurochemical alterations associated with epilepsy or seizure activity. In: Dam, M. and Gram, L. (Eds.), Comprehensive Epileptolo~, Raven Press, New York, NY, pp. 1-15. Shiraga, H. and A. Mori (1982) Convulsive activity of aguanidinoglutaric acid in rats. IRCS Med. Sci., 10: 855-856.

105 Stein I.M., Cohen B.D. and Horowitz H.I. (1968) Guanidinosuccinic acid: the "X" factor in uremic bleeding.'? Clin. Res., 16: 397. Stone W.E. (1972) Systemic chemical convulsants and metabolic derangements. In: Purpura, D.P., Penry, J.K., Tower, D.B., Woodbury, D.M. and Walter, R.D. (Eds.), Experimental Models of Epilepsy. A Manual for the Laboratory Worker, Raven Press, New York, NY, pp. 407-432. Tyler, H.R. (1975) Neurological aspects of uremia: an overview. Kidney Int., $2: 188-193. Vefiskov~, J., L. Velfsek, P. Mares and R. Rolcyta (1990) Ketamine suppresses both bicuculline- and picrotoxin-induced generalized

tonic-clonic seizures during ontogenesis. Pharmacol. Biochem. Behav., 37: 667-674. Weil, C.S. (1952) Tables for convenient calculation of median-effective dose (LDs0 or ED50) and instructions in their use. Biometrics, 8: 249-263. Yokoi, I., 3. Toma and A. Mori (1984) The effect of homoarginine on the EEG of rats. Neurochem. Pathol., 2: 295-300. Yokozawa, T., Z.L. Mo and H. Oura (1989) Comparison of toxic effects of methylguanidine, guanidinosuccinic acid and creatinine in rats with adenine-induced chronic renal failure. Nephron, 51: 388-392.