Methods: Groups of male Swiss Webster mice (n greater or equal to 10/group) were given either vehicle (intralipid, 10 ml *symbol* kg sup -1, control groups) or propofol (50 mg *symbol* kg sup -1, test groups) injected intraperitoneally. Five min after injection, convulsions were induced with either bicuculline (1.36-5.44 nmoles), picrotoxin (0.21-1 nmol), N-methyl-D-aspartic acid (0.51-2 nmol), quisqualic acid (1-10 nmol), kainic acid (0.252-2 mole), or strychnine (1.35-10.78 nmol) injected intracerebroventricularly. The number of animals with convulsions after each dose was recorded. Analysis of statistical significance was based on the log-probit lines of the quantal dose-response for the respective control and test groups, calculated 50% effective doses (ED50), the potency ratios (ED50^higher/ED50^lower) and their 95% confidence limits.

Results: Propofol pretreatment decreased the potency ratio of both bicuculline (0.47, 95% confidence interval = 0.23-0.94) and picrotoxin (0.61, 0.47-0.79), signifying an anticonvulsant effect. Conversely, propofol pretreatment significantly enhanced the convulsive potency of kainic acid (potency ratio and 95% confidence interval = 1.66, 1.21-2.29), quisqualic acid (3.17, 1.98-5.09), and strychnine (1.76, 0.79-3.89).

Conclusions: Current results suggest that propofol augments the paroxysmal motor phenomenon induced by kainic acid, quisqualic acid, and strychnine. This action may be, at least partly, responsible for the motor manifestations reported after propofol administration. These in vivo results on modulation of gamma-aminobutyric acid, glycine, alpha-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid, and kainate receptor-mediated transmission may be of significance in understanding the mechanism of propofol action at the excitatory and inhibitory amino acid receptors.

Key words: Hypnotics: propofol. Receptors, excitatory neurotransmission: alpha-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid; kainic acid; N-methyl-D-aspartic acid; quisqualic acid. Receptors, inhibitory neurotransmission: bicuculline; gamma-aminobutyric acid; picrotoxin.

MANY anesthetics possess anticonvulsant properties, yet many of these same drugs can also cause seizures. [1,2] For preclinical anticonvulsant screening it is customary to test drug effect against a wide range of known chemical convulsants. Because convulsants act by known discrete mechanisms and neurotransmitters, they can help determine the spectrum and mechanism of test drug action. [3] A recent strategy used in anticonvulsant drug development is to block synaptic excitation or to enhance synaptic inhibition. [4,5] Picrotoxin and bicuculline induce convulsions by decreasing gamma-aminobutyric acid (GABA)-ergic inhibitory neurotransmission. N-methyl-D-aspartic acid, kainic acid, and quisqualic acid produce convulsions by enhancing the glutamate receptor-mediated excitatory neurotransmission. Both types of agents are useful in screening new anticonvulsant drugs. [3] 

Propofol is a rapidly effective, short-acting intravenous anesthetic. Although some initial data suggested that propofol had no effect on the minimal electrical threshold required to induce seizures in mice, [6] propofol was subsequently shown to protect mice [7] and rats [8] against electroshock-induced seizures and against seizures induced by some chemoconvulsants in mice, [7] rats, [9] and rabbits. [10,11] However, propofol-induced "seizurelike" excitatory effects including clonic movements, extensor rigidity, and opisthotonos have also been reported in mice, [12] rats, [13] and dogs. [14] The basis of these apparently contradictory reports has not yet been explained. In the meantime, the use of propofol in patients with seizure disorders remains controversial. [15] 

In laboratory animals, propofol has been tested against seizures induced by chemoconvulsants like pentylenetetrazol, [7,9-11] penicillin G, [11] bupivacaine, [16] or by electroshock. [6-8] To further understand the spectra and mechanism of the anticonvulsant profile of propofol, we assessed the effect of propofol pretreatment on convulsions induced by picrotoxin and bicuculline, which decrease GABAergic inhibitory neurotransmission, strychnine, which blocks glycinergic inhibition, and by N-methyl-D-aspartic acid, kainic acid, and quisqualic acid, which enhance the glutamate receptor-mediated excitatory neurotransmission.

The study protocol was approved by the Institutional Animal Care and Use Committee. Male Swiss Webster mice (Germantown, NY) weighing 20-25 g were housed five per cage in a room with controlled temperature (22 plus/minus 2 degrees Celsius), humidity, and artificial light (0630-1900). Animals had free access to food and water and were used after a minimum of 4 days acclimatization to the housing conditions. All chemoconvulsant drugs were purchased from Sigma Chemical Co. (St. Louis, MO).

Test group animals were injected intraperitoneally with a bolus (50 mg *symbol* kg sup -1) of propofol (Diprivan, Stuart Pharmaceuticals, Wilmington, DE) in a uniform volume of 10 ml *symbol* kg sup -1. For this purpose, contents of a freshly opened ampule of propofol was mixed in 1:1 ratio with the vehicle (Intralipid 10%, KabiVitrum, Clayton, NC). Control group animals were injected with an equal volume of vehicle. Five minutes after injection of vehicle or anesthetic, animals received a dose of the convulsant. The dose of anesthetic and the time for convulsant injections were based on the results of a published study, [7] which indicated that 50 mg *symbol* kg sup -1 is the median effective dose of propofol that significantly affects the minimum convulsant dose of pentylenetetrazol and the peak of the behavioral effects is reached, 5 min after intraperitoneal administration. The dose ranges of chemoconvulsants used were: bicuculline (1.36-5.44 nmol), picrotoxin (0.21-1 nmol), N-methyl-D-aspartic acid (0.51-2 nmol), quisqualic acid (1-10 nmol), kainic acid (0.252-2 nmol), and strychnine (1.35-10.78 nmol). Bicuculline solution was prepared just before use by dissolving in 0.05 N hydrochloric acid. All other convulsant drug solutions were freshly prepared by dissolving in sterile water. The convulsants were injected intracerebroventricularly in an uniform volume of 10 micro liter using the method employed by us in several other studies. [17-23] Briefly, intracerebroventricular injections were made using a 27-G 1/4-inch needle attached to a 500-micro liter Hamilton syringe assembled on an automatic dispenser (PB 600, Hamilton Company, Reno, NV). The needle was fitted with polyethylene tubing leaving 3 mm of the needle tip exposed. The point of injection was on an imaginary line drawn through the anterior base of the ears and 2 mm from an imaginary mid-sagittal line.

After intracerebroventricular injection, animals were placed individually in polycarbonate cages (33 x 15 x 13 cm) covered with a lid for observation by a person unaware of the drugs or doses used. Animals were observed for 30 min after intracerebroventricular injection and were scored as showing motor paroxysms when one or more of the following signs were present [17,24] : (1) wild running or jumping; (2) myoclonic convulsions (i.e., isolated, jerky limb movements); and (3) clonic convulsions (i.e., repetitive movements involving all limbs simultaneously, usually accompanied by loss of righting reflex).

Statistical significance of the data was analyzed by the log-probit method of Litchfield and Wilcoxon using the PHARM/PCS computer program. [25] Calculated 50% effective dose (ED50) values from the seizure tests are estimates of the dose at which 50% of animals exhibit seizures, whereas the potency ratio is the ratio of higher/lower ED50 values. [25] The computer program used also provides the 95% confidence limits associated with the ED50 and potency ratio.

Vehicle administration did not affect behavior of animals. Fifteen to twenty percent of mice had some staggering gait within 2-3 min after the propofol injection but continued to be active. No animal showed loss of righting reflex after propofol. The only observable behavioral sign 5 min after propofol administration was that the animals were less agitated when handled.

The effect of propofol pretreatment on GABA-antagonist-induced convulsions is shown in Table 1. Convulsions induced by bicuculline and picrotoxin were dose-dependent in control and propofol-treated groups. However, propofol pretreatment shifted the dose effect line to the right (Figure 1). The decrease in convulsant potency ratio after propofol pretreatment was significant (P < 0.05) for bicuculline and picrotoxin (Table 2).

The effect of propofol pretreatment on excitatory amino acid agonist-induced convulsions is provided in Table 3. Convulsions induced by N-methyl-D-aspartic acid, quisqualic acid, and kainic acid, were dose-dependent both in control and propofol-treated groups. Propofol pretreatment shifted the dose effect line to the left (Figure 2). The increase in potency ratio resulting from propofol pretreatment was significant (P < 0.05) for kainic acid and quisqualic acid but not for N-methyl-D-aspartate acid induced convulsions (Table 4).

The effect of propofol pretreatment on glycine antagonist-induced convulsions is provided in Table 5. In controls, strychnine-induced convulsions were dose-dependent. Propofol pretreatment significantly increased the potency ratio (Table 5).

The present results clearly indicate that propofol attenuates the convulsive mechanisms at picrotoxin and bicuculline sensitive GABAergic pathways and extends the list of chemoconvulsants against which the anticonvulsant effect of propofol has been shown. Picrotoxin and bicuculline convulsions result from decreasing GABAergeic inhibitory neurotransmission. Picrotoxin blocks GABA-activated chloride channels whereas bicuculline antagonizes GABA^Areceptors by competing with GABA for its binding sites. Some electrophysiologic data indicate that propofol enhances GABA-mediated synaptic inhibition in vitro. [26,27] Propofol appears to activate the GABA^Areceptor-chloride inophore complex [28] and to potentiate binding of [sup 3 Hydrogen]-flunitrazepam to the GABA^Areceptor. [29] Our results demonstrate an anticonvulsant effect of propofol against picrotoxin and bicuculline and thus provide in vivo behavioral evidence for propofol-induced enhanced GABAergic transmission. It is noteworthy that in vivo results are significant because they permit discrimination of changes measured in the in vitro paradigm that may not have physiologic significance. In fact, in vivo results provide the closest representation of human behavioral response to anesthetic drugs. [30] 

In contrast to its effect on GABA antagonist-induced motor effects, propofol significantly augments the motor manifestation of some excitatory amino acid receptor agonists. Excitatory amino acid receptors are usually classified by agonists that selectively activate them. Glutamate, the prototype agonist, nonselectively stimulates all excitatory amino acid receptor subtypes. Based on the selective agonists, the three major ligand gated subtypes of excitatory amino acid receptors identified and classified are, N-methyl-D-aspartate acid, alpha-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid, and kainic acid receptors. The alpha-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid receptor is also called the quisqualate receptor because it is also activated by quisqualate. Our in vivo results suggest that propofol facilitates excitatory amino acid receptor-mediated responses. Propofol modulation appears to be more potent at alpha-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid and kainic acid receptor subtypes than at N-methyl-D-aspartate acid receptor-related mechanisms. However, some in vitro data suggest that propofol inhibits excitatory amino acid agonist-mediated effects on the neuronal channel currents [31] and Calcium sup +2 entry into synaptosomes. [32] This disparity between the in vivo and in vitro results emphasizes the importance of in vivo paradigms in assessing the mechanism of anesthetic action. [33] Unlike in vitro studies, multiple sites of modulation of motor effects are possible in vivo. For example, by increasing conductance at the GABA-associated channel and depolarizing the membrane, propofol may allow the excitatory amino acid agonists to exert a proportionately larger effect or, propofol may have presynaptic effects in upstream neurons.

An important concern when using propofol for neuroanesthesia relates to reports of alleged seizure activity induced by propofol. [15] In fact, propofol is the most recent anesthetic agent to be implicated as a potential cause of iatrogenic convulsions. The issue of whether the excitatory behavior associated with propofol-induced anesthesia is indeed an epileptiform seizure or not remains controversial. [15,34,35] Present results suggest that propofol augments the paroxysmal motor phenomenon observed after kainic acid, quisqualic acid, and strychnine by acting on excitatory and glycinergic receptors/pathways. This action may indeed be at least partly responsible for the motor manifestations reported after propofol administration. Thus, these results provide some insight into the inconsistent effects of propofol in epileptic patients as well as the occasional motor excitation seen in nonepileptic patients. A comparative evaluation of thiopental, etomidate, or ketamine with propofol, using the present experimental protocol could provide better a perspective on the practical implications of the observed augmented motor effects after propofol administration.

Strychnine is a specific glycine antagonist. Like GABA, glycine is a ubiquitous amino acid in the central nervous system that functions as an inhibitory neurotransmitter. While GABA acts primarily in the cerebral cortex, glycine inhibits subcortical areas of the brain. Glycine activates a strychnine-sensitive Chlorine sup - channel in postsynaptic membranes. Some in vitro evidence suggests propofol potentiates glycine-activated currents in embryonic spinal neurons. [26] In an earlier in vivo study, [12] mice showed an increased incidence of motor manifestations when a subconvulsant dose of strychnine was administered 5 min after propofol. In another report, intravenous propofol injected in female mice showed anticonvulsant effect when challenged with intravenous infusion of strychnine. [36] As proposed earlier, methodologic differences could be the basis for disparities in results on the effect of propofol on strychnine-induced convulsions. [36] 

Opisthotonos characteristically occurs from strychnine poisoning in most laboratory animals and in humans due to strychnine-induced reduction in the reciprocal inhibition existing between antagonistic muscles. In humans, opisthotonos has been observed occasionally after propofol administration. In fact, although many anesthetics produce clinically observed seizure activity, opisthotonos and other athetoid movements in the postoperative period are almost uniquely associated with the use of propofol. The Medline database contains at least 17 reports of propofol-induced opisthotonos. [37] Attempts to treat this disorder with propofol itself have been unsuccessful. [37,38] Based on the present and earlier experimental results in animals, [12] propofol-induced opisthotonos may be caused by an upregulated strychninelike glycinergic inhibition especially at subcortical sites in susceptible patients. In fact, a speculative therapeutic strategy recommends that propofol use be avoided while treating opisthotonos. [37,39] Our results provide some in vivo data to support this recommendation.

In summary, the present in vivo results provide evidence for modulation of GABA, glycine, and glutamate receptor-mediated events by propofol. While the role of GABA-mediated events in the molecular mechanism of anesthetics has been extensively investigated, [30] to date there has been very little research on the effects of anesthetics on the subtypes of excitatory amino acid receptors. [33,40] Only recently, in vitro studies implicating excitatory amino acid receptor subtypes in the molecular mechanism of anesthetics have started appearing. [41,42] However, in vivo results are probably more significant in correlating molecular events that are influenced by the anesthetics and the physiologic effects that subsequently occur. Hence, the present in vivo results suggesting modulation of alpha-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid and kainate receptor-mediated events by propofol are significant to understanding the mechanism of the anesthetic action at the excitatory amino acid receptor sites.

The authors thank Dr. Gilbert J. Grant, Dr. Mark I. Zakowski, and Dr. Robert R. Ben-Harari for critical reading of the manuscript.