SL651498: an anxioselective compound with functional selectivity for alpha2- and alpha3-containing gamma-aminobutyric acid(A) (GABA(A)) receptors.

SL651498 [6-fluoro-9-methyl-2-phenyl-4-(pyrrolidin-1-yl-carbonyl)-2,9-dihydro-1H-pyrido[3,4-b]indol-1-one] is a novel pyridoindole derivative that displays high affinity for rat native GABA(A) receptors containing alpha(1) (K(i) = 6.8 nM) and alpha2 (K(i) = 12.3 nM) subunits, and weaker affinity for alpha5-containing GABA(A) receptors (K(i) = 117 nM). Studies on recombinant rat GABA(A) receptors confirm these data (K(i), alpha1beta2gamma2 = 17, alpha2beta2gamma2 = 73, alpha5beta3gamma2 = 215 nM) and indicate intermediate affinity for the alpha3beta2gamma2 subtype (K(i) = 80 nM). SL651498 behaves as a full agonist at recombinant rat GABA(A) receptors containing alpha2 and alpha3 subunits and as a partial agonist at recombinant GABA(A) receptors expressing alpha1 and alpha5 subunits. SL651498 elicited anxiolytic-like activity similar to that of diazepam [minimal effective dose (MED): 1-10 mg/kg, i.p.] in three conflict models, in the elevated plus-maze, the light/dark test, and the defense test battery in rats and mice. Results from activity tests and electroencephalogram analysis indicated that SL651498 induced muscle weakness, ataxia, or sedation at doses much higher than those producing anxiolytic-like activity (MED > or = 30 mg/kg, i.p.). Repeated treatment for 10 days with SL651498 (30 mg/kg, i.p., b.i.d.) in mice was not associated with the development of tolerance to its anticonvulsant effects or physical dependence. Furthermore, SL651498 was much less active than diazepam in potentiating the depressant effects of ethanol in mice. The "anxioselective" profile of SL651498 points to a major role for GABA(A) alpha2 subtype in regulating anxiety and suggests that selectively targeting GABA(A) receptor subtypes can lead to drugs with increased clinical specificity.

R-zacopride, a 5-HT3 antagonist/5-HT4 agonist, reduces sleep apneas in rats.

The effects of R-zacopride, a benzamide with potent 5-HT3 receptor antagonist and 5-HT4 receptor agonist properties, on spontaneous apneas were studied in 10 Sprague-Dawley rats by monitoring respiration and sleep for 6 h. R-zacopride (0.5, 1.0 and 10.0 mg/kg) suppressed spontaneous central apneas during non-rapid-eye-movement (NREM) sleep by 50% (P=.05 for 0.5 mg/kg, P=.02 for 1.0 mg/kg and P=.001 for 10.0 mg/kg dose vs. control), and during rapid-eye-movement (REM) sleep by 80% by all doses tested (P<.0007) for at least 2 h after intraperitoneal injection. We conclude that R-zacopride, over a 20-fold dose range, significantly reduces central apnea expression during NREM and REM sleep in the rat. The efficacy of this compound to suppress central apneas most probably arises from its antagonist actions at 5-HT3 receptors or from its mixed agonist/antagonist profile at 5-HT4/5-HT3 receptors.

Autoradiographic localization of [3H]thiocolchicoside binding sites in the rat brain and spinal cord.

Thiocolchicoside is used in humans as a myorelaxant drug with anti-inflammatory and analgesic activity. Recently we established the experimental conditions that allowed the identification of [3H]thiocolchicoside binding sites in synaptic membranes of rat spinal cord and cerebral cortex. The pharmacological characterization of these sites indicated that GABA and several of its agonists and antagonists, as well as strychnine, were able to interact with [3H]thiocolchicoside binding in a dose-dependent manner and with different affinities. In order to gain more insight into the nature and the anatomical distribution of the binding sites labeled by [3H]thiocolchicoside, in the present study we examined the localization of these sites on parasagittal and coronal sections of the rat brain and spinal cord, respectively, using receptor autoradiography. In the spinal cord an intense signal was observed in the gray matter, with the highest density occurring in the superficial layers of the dorsal horns. Strychnine completely displaced [3H]thiocolchicoside binding, whereas GABA only partially removed the radioligand from its binding sites. In the brain, specific binding occurred in several areas and was displaced by both GABA and strychnine. The distribution of [3H]thiocolchicoside binding sites in brain sections, however, did not match that found for [3H]muscimol. Furthermore, cold thiocolchicoside was not able to completely displace [3H]muscimol binding, and showed a different efficacy in the various areas labeled by the radioligand. We conclude that thiocolchicoside may interact with a subpopulation of GABA(A) receptors having low-affinity binding sites for GABA. Furthermore, the observed sensitivity to strychnine in the spinal cord indicates an interaction also with strychnine-sensitive glycine receptors, suggesting that the pharmacological effects of thiocolchicoside may be the result of its interaction with different receptor populations.

Characterization of [3H]thiocolchicoside binding sites in rat spinal cord and cerebral cortex.

Thiocolchicoside, a semi-synthetic derivative of the naturally occurring compound colchicoside with a relaxant effect on skeletal muscle, has been found to displace both [3H]gamma-aminobutyric acid ([3H]GABA) and [3H]strychnine binding, suggesting an interaction with both GABA and strychnine-sensitive glycine receptors. In order to gain further insight into the interaction of thiocolchicoside with these receptors, the binding of [3H]thiocolchicoside in rat spinal cord-brainstem and cortical synaptic membranes was characterized. [3H]Thiocolchicoside binding was saturable in both tissues examined. In spinal cord-brainstem membranes, we found a K(D) of 254 +/- 47 nM and a Bmax of 2.39 +/- 0.36 pmol/mg protein, whereas in cortical membranes, a K(D) of 176 nM and a Bmax of 4.20 pmol/mg protein was observed. A similar K(D) value was found in kinetic experiments performed in spinal cord-brainstem membranes. Heterologous displacement experiments showed that GABA and strychnine displaced the binding in a dose-dependent manner, whereas glycine was ineffective. [3H]Thiocolchicoside binding was also displaced by several GABA(A) receptor agonists and antagonists, but not by baclofen, flunitrazepam, guvacine, picrotoxin or by other drugs unrelated to GABA transmission. In spinal cord-brainstem, and to a lower extent, in cortical membranes, GABA and its analogs were not able to completely displace [3H]thiocolchicoside specific binding indicating that, besides GABA(A) receptors, thiocolchicoside can bind to another unidentified site. Unlabelled thiocolchicoside, however, completely displaced [3H]muscimol binding both in cortical and in spinal cord-brainstem synaptic membranes with an IC50 in the low microM range. Neurosteroids were found to modulate the binding in cortical but not in spinal cord-brainstem synaptic membranes. We conclude that [3H]thiocolchicoside binding shows a pharmacological profile indicating an interaction with the GABA(A) receptor. The different affinities for the GABA(A) receptor agonists and antagonists and sensitivity to neurosteroids obtained in the cerebral cortex and in the spinal cord may indicate a preferential interaction of the compound with a subtype of the GABA(A) receptor. The data also indicate that [3H]thiocolchicoside binds to another site(s), whose nature remains to be elucidated.

Preclinical profile of befloxatone, a new reversible MAO-A inhibitor.

Befloxatone, a novel oxazolidinone derivative, is a potent, selective and reversible monoamine oxidase A (MAO-A) inhibitor in vitro (K1A = 1.9-3.6 nM) and ex vivo (ED50 MAO-A = 0.02 mg/kg, p.o.). It does not interact with a large number of receptors, monoamine transporters or other amine oxidases. Binding studies with [3H]-befloxatone in rat brain sections show that it labels with high affinity (Kd = 1.3 nM) a single population of sites with the pharmacological characteristics and regional distribution of MAO-A. In the rat brain, befloxatone (0.75 mg/kg, i.p.) increases tissue levels of monoamines and decreases levels of their deaminated metabolites. Acute administration of befloxatone (0.75 mg/kg, i.p.) induces an increase in extracellular striatal dopamine and cortical norepinephrine but not cortical serotonin levels in the rat. Befloxatone (1 mg/kg, i.p.) potently inhibits the firing rate of serotonergic neurons, partially decreases the firing of noradrenergic neurons and has no effect on the firing of dopaminergic neurons (a mirror image of its effects on monoamine release in terminal regions), suggesting that the relative effects of befloxatone on monoamine release may be governed by autoreceptor-mediated control of monoaminergic neurons at the cell body level. Befloxatone (0.03-0.3 mg/kg, p.o.) exhibits potent activity in behavioural models predictive of antidepressant activity. Befloxatone (up to 1.5 mg/kg, p.o.) does not potentiate the pressor effects of orally administered tyramine at centrally active doses and duodenal [3H]-befloxatone binding is displaced by increasing doses of orally administered tyramine (0.1-40 mg/kg, i.p.). These results suggest that befloxatone is a potent reversible MAO-A inhibitor with antidepressant potential and a wide safety margin with regard to the potentiation of the pressor effect of tyramine.

Antagonist properties of eliprodil and other NMDA receptor antagonists at rat NR1A/NR2A and NR1A/NR2B receptors expressed in Xenopus oocytes.

We have studied the effects of a variety of N-methyl-D-aspartate (NMDA) antagonists acting at different sites of the NMDA receptor complex on NMDA-induced currents in Xenopus oocytes expressing heteromeric NR1A/NR2 and NR1A/NR2B receptors. The polyamine site antagonists eliprodil (IC50 = 3.0 microM) and ifenprodil (IC50 = 0.27 microM) antagonized NMDA responses at NR1A/NR2B receptors but not at NR1A/NR2A receptors (IC50 > 100 microM). The channel blockers dizocilpine, memantine and phencyclidine (PCP) were equally potent antagonists at both receptor subtypes whereas dextromethorphan was four times more potent at NR1A/NR2A receptors. The glycine site antagonists L-689,560 and 7-Cl-kynurenate were 10 times more potent at NR1A/NR2A than at NR1A/NR2B receptor subtypes. The selectivity of eliprodil and ifenprodil for the NR1A/NR2B receptor subtype may, at least partially, explain their favorable side effects profile.

Development of stable cell lines expressing different subtypes of GABAA receptors.

The experiments reported here were motivated by our interest to express in stably-transfected cells large amounts of recombinant rat GABAA receptors. For this, we developed an original two step selection strategy, in which the first step consisted of transfecting HEK 293 cells with rat GABAA receptor alpha and beta subunits. G 418 resistant colonies isolated at this step were screened for [3H] muscimol binding to select for those that coexpressed alpha- and beta-subunits. The best alpha and beta subunit expressing colony was then supertransfected with a plasmid coding for the gamma rat GABAA receptor subunit and a mutant DHFR gene. After a second round of selection, this time in presence of methotrexate, those colonies that coexpressed ternary alpha beta gamma GABAA receptor combinations were distinguished using [3H] flumazenil as a probe. This strategy was applied to the isolation of 3 GABAA receptor clones, alpha 1 beta 2 gamma 2s, alpha 3 beta 2 gamma 2s and alpha 5 beta 3 gamma 2s, that expressed relatively high levels of these proteins. These 3 cell lines exhibited pharmacological and functional properties similar to cells transiently-transfected with equivalent subunit combinations. These cell lines therefore provide attractive models with which to evaluate the intrinsic activity and potency of compounds at recombinant GABAA receptor subtypes.

Protracted treatment with diazepam reduces benzodiazepine1 receptor-mediated potentiation of gamma-aminobutyric acid-induced currents in dissociated rat hippocampal neurons.

Tolerance to benzodiazepines (BZs) is thought to involve alterations of the gamma-aminobutyric acid (GABA)A receptor as a result of the prolonged occupancy of its modulatory BZ recognition site. We used the whole-cell patch-clamp technique to compare the functional and pharmacological properties of GABAA receptors in acutely dissociated hippocampal neurons from the control or diazepam-tolerant rats. Administration of diazepam (15 mg/kg p.o.) twice a day for 10 days induced tolerance as demonstrated by the decreased potency of acute diazepam i.p. injections to protect against pentylenetetrazole-induced clonictonic convulsions (10.5% of tolerant rats protected by 0.1 mg/kg of diazepam against 55% of nontreated rats, 48 hr after the last dose of the chronic treatment). The specific current induced by 1 microM GABA in acutely dissociated hippocampal neurons 48 hr after withdrawal (10.5 +/- 1.3 microA/cm2) was similar to that observed in the control rats (8.7 +/- 0.8 microA/cm2). The EC50 value for GABA was unchanged by the chronic treatment [6.3 (5.4-7.1) and 7.5 (6.2-8.7) microM in neurons from the control and treated rats, respectively]. The potency of the nonselective allosteric modulator diazepam to stimulate Cl- currents was identical in cells from treated rats [EC50 values of 25 (20-30) and 34 (26-41) nM in the control and treated rats, respectively; P < .05], but the potency of the selective BZ1-site ligand zolpidem was decreased [EC50 values of 99 (88-111) and 267 (221-313) nM in the control and treated rats, respectively; P < .05]. The maximal potentiation of the GABA-induced current was significantly decreased with diazepam (maximal potentiation: 168.0 +/- 16.2 and 124.0 +/- 8.9% in the control and treated rats, respectively). These results suggest that tolerance to diazepam is accompanied in hippocampal neurons by a decrease in BZ1 binding sites and in the functional coupling of BZ/GABA recognition sites.

Zolpidem functionally discriminates subtypes of native GABAA receptors in acutely dissociated rat striatal and cerebellar neurons.

The whole-cell patch-clamp technique was used to compare the properties of native GABAA receptors in Purkinje and striatal neurons acutely dissociated from neonatal rat brains (7-11 days old). In symmetrical chloride concentrations and at a negative holding voltage, GABA induced inward currents in a concentration-dependent manner with EC50 values of 4 and 8 uM in Purkinje and striatal neurons, respectively. Diazepam potentiated the current induced by 1 uM GABA in Purkinje and striatal neurons with EC50 values of 28 and 42 nM and maximal potentiations of 128 and 182%, respectively. Zolpidem potentiated this GABA-induced current in Purkinje and striatal neurons with EC50 values of 33 and 195 nM and maximal potentiations of 189 and 236%, respectively. These results show that zolpidem, in contrast to diazepam, functionally discriminates subtypes of native GABAA receptors. Zolpidem has greater affinity for GABAA receptors containing omega 1 (Purkinje cells) than for those with omega 2 (striatum) sites and has higher intrinsic activity at these receptors than diazepam. These properties of zolpidem may contribute to its hypnoselective profile.

Antagonist properties of the stereoisomers of ifenprodil at NR1A/NR2A and NR1A/NR2B subtypes of the NMDA receptor expressed in Xenopus oocytes.

The NMDA receptor antagonist ifenprodil contains two asymmetric centres which give rise to four stereoisomeric forms of this molecule. The inhibitory effects of each of these stereoisomers on recombinant NMDA receptors expressed from NR1A/NR2A and NR1A/NR2B subunit combinations were studied in Xenopus oocytes by voltage-clamp recording. All four ifenprodil stereoisomers were potent antagonists at NR1A/NR2B (IC50 < 0.8 microM), but weak antagonists at NR1A/NR2A receptors (IC50 > 100 microM). In heteromeric NR1A/NR2B receptors, (+) erythro- and (-) threo-ifenprodil (IC50 0.21 and 0.22 microM, respectively) were about 4 times more potent than (-) erythro- and (+) threo-ifenprodil (IC50 0.81 and 0.76, respectively). These results show that the stereoisomers of ifenprodil exhibit a weak though significant stereoselectivity at the NR1A/NR2B NMDA receptor subtype.

Block of P-type Ca2+ channels by the NMDA receptor antagonist eliprodil in acutely dissociated rat Purkinje cells.

The effect of eliprodil on P-type Ca2+ channels was investigated in acutely dissociated rat Purkinje neurons, by using the whole-cell patch-clamp technique. Eliprodil inhibited in a reversible manner the omega-agatoxin-IVA-sensitive Ba2+ current elicited by step depolarizations from a -80 mV holding voltage (IC50 = 1.9 microM). The Ba2+ current showed steady-state inactivation (V1/2 = -61 mV) which was shifted toward more positive values when the intracellular Ca2+ buffering was increased. In these conditions, the potency of eliprodil was decreased (IC50 = 8.2 microM), suggesting a modulation by intracellular Ca2+ of the eliprodil blockade. The potency of eliprodil was not modified at more depolarized holding potentials and was not dependent on the frequency at which the step-depolarizations were applied (0-0.2 Hz) indicating a lack of voltage and use dependence of the eliprodil blockade. When eliprodil was applied in the patch-pipette at a concentration which causes full block when applied externally, the Ba2+ current amplitude was not affected and external application of eliprodil was still efficacious, indicating an extracellular location of the binding site. Analysis of the time course of recovery from Ca2+ channel blockade obtained by concomitant application of eliprodil with Cd2+, omega-agatoxin-IVA or fluspirilene, indicated that these later compounds did not interact with eliprodil, suggesting that eliprodil acts at a different site. These results demonstrate that eliprodil blocks P-type Ca2+ channels in cerebellar Purkinje neurons and suggest that this property may contribute to its neuroprotective activity.

Differential effects of midazolam and zolpidem on sleep-wake states and epileptic activity in WAG/Rij rats.

Hypnotic drugs are known to possess antiepileptic activity. Therefore, the effects of the benzodiazepine hypnotic midazolam (10 mg/kg) and the novel imidazopyridine hypnotic zolpidem (10 mg/kg) on sleep-wake states and on the number of spike-wave discharges were evaluated in WAG/Rij rats. Rats of this strain are considered to be a model for generalized absence epilepsy. Animals were implanted with chronic monopolar EEG electrodes and, after recovery from surgery, the EEG was recorded for 6 h during the dark period on 3 consecutive days. Sleep recordings were analyzed using Hjorth's parameters and number and duration of spike-wave discharges were visually determined. It was found that both drugs facilitated nonREM sleep at the cost of wakefulness. Both hypnotics also reduced the number and duration of spike-wave discharges. The initial decrease after midazolam, however, was followed by a rebound reflecting a poorer quality of vigilance expressed as an increase in spike-wave discharges. The strong antiabsence activity of zolpidem mimics that of midazolam and is well correlated with their equipotent hypnotic action and anticonvulsant effect in the isoniazid test.

Modulation of the gamma-aminobutyric acid type A receptor by the antiepileptic drugs carbamazepine and phenytoin.

We report here that carbamazepine and phenytoin, two widely used antiepileptic drugs, potentiate gamma-aminobutyric acid (GABA)-induced Cl- currents in human embryonic kidney cells transiently expressing the alpha 1 beta 2 gamma 2 subtype of the GABAA receptor and in cultured rat cortical neurons. In cortical neuron recordings, the current induced by 1 microM GABA was enhanced by carbamazepine and phenytoin with EC50 values of 24.5 nM and 19.6 nM and maximal potentiations of 45.6% and 90%, respectively. The potentiation by these compounds was dependent upon the concentration of GABA, suggesting an allosteric modulation of the receptor, but was not antagonized by the benzodiazepine (omega) modulatory site antagonist flumazenil. Carbamazepine and phenytoin did not modify GABA-induced currents in human embryonic kidney cells transiently expressing binary alpha 1 beta 2 recombinant GABAA receptors. The alpha 1 beta 2 recombinant is known to possess functional barbiturate, steroid, and picrotoxin sites, indicating that these sites are not involved in the modulatory effects of carbamazepine and phenytoin. When tested in cells containing recombinant alpha 1 beta 2 gamma 2, alpha 3 beta 2 gamma 2, or alpha 5 beta 2 gamma 2 GABAA receptors, carbamazepine and phenytoin potentiated the GABA-induced current only in those cells expressing the alpha 1 beta 2 gamma 2 receptor subtype. This indicates that the nature of the alpha subunit isoform plays a critical role in determining the carbamazepine/phenytoin pharmacophore. Our results therefore illustrate the existence of one or more new allosteric regulatory sites for carbamazepine and phenytoin on the GABAA receptor. These sites could be implicated in the known anticonvulsant properties of these drugs and thus may offer new targets in the search for novel antiepileptic drugs.

Mizolastine, a novel selective histamine H1 receptor antagonist: lack of sedative potential on the EEG in the rodent.

The sedative potential of mizolastine, a new, potent and selective antagonist of histamine H1-receptors, has been evaluated in the rodent with EEG techniques. In chronically implanted rabbits, sedation was observed in ECoG recordings after intravenous injection of terfenadine (1-10 mg/kg) and loratadine (0.3-3 mg/kg) but not after intravenous injection of astemizole or mizolastine (1-10 mg/kg). In freely moving implanted rats, mizolastine and cetirizine (10 mg/kg i.p.) did not modify the sleep-wakefulness pattern recorded during the dark period nor did mizolastine alter the sleep architecture recorded in rats during the light period. In contrast, during the dark-period recording, astemizole, loratadine and terfenadine (10 mg/kg i.p.) increased the total duration of slow-wave sleep; this sleep-facilitating effect had a late onset of action, beginning 3 h after drug injection. In conclusion, the results obtained with astemizole, cetirizine, loratadine and terfenadine demonstrate their low sedation potential in the rat, and suggest that the absence or low incidence of sedation seen in humans with these drugs may be due to their limited ability to cross the blood brain-barrier, especially at recommended therapeutic doses. Mizolastine appears to be devoid of sedative effects in our experimental models irrespective of the route of administration used. These results predict a lack of sedative action in humans with mizolastine at therapeutic doses.

The NMDA receptor antagonist eliprodil (SL 82.0715) blocks voltage-operated Ca2+ channels in rat cultured cortical neurons.

The effect of the non-competitive NMDA receptor antagonist eliprodil on NMDA receptor- and voltage-operated Ca2+ currents was investigated in rat cultured cortical neurons by using the whole-cell patch clamp technique. With neurons voltage-clamped at -40 mV, eliprodil reduced in a concentration-dependent manner the inward current induced by N-methyl-D-aspartate (NMDA) (10 microM) in the presence of D-serine with an IC50 of 0.67 microM (Imax = 83%). Eliprodil also blocked the total inward Ba2+ current carried in part by L- and N-type Ca2+ channels with an IC50 of 1.48 microM (Imax = 87%). These results suggest that the neuroprotective properties of eliprodil could be due to its combined ability to antagonize the NMDA receptor- and voltage-operated Ca2+ channels.

Influence of zolpidem, a novel hypnotic, on the intermediate-stage and paradoxical sleep in the rat.

Paradoxical sleep (PS) in mice, rats, and cats is preceded and sometimes followed by a short-lasting stage characterized by cortical high-amplitude spindles and hippocampal low-frequency theta rhythm. This intermediate stage (IS) seems to correspond to a transient functional disconnection of the forebrain from the brainstem. Pentobarbital and benzodiazepines greatly extend IS at the expense of PS, which is suppressed. Zolpidem, a new imidazopyridine hypnotic, was studied at 2.5, 5, and 7.5 mg/kg IP. At 2.5 mg/kg, which is already a true hypnotic dose, it only decreased PS during the first 2 h of recording with a rebound during the following 4 h of recording. At 5 mg/kg, zolpidem increased the number and total duration of IS episodes, increased IS episodes not followed by PS, and increased PS latency of occurrence. PS amount was decreased during the first three h with a rebound in the next 3 h. At 7.5 mg/kg, the total amount of PS was also decreased. The eye movement number and theta rhythm frequency of PS were unchanged. These results show that zolpidem produces less disruption of the association between IS and PS than do previous hypnotics.

Evaluation of the stability and quality of sleep using Hjorth's descriptors.

Hjorth's descriptors (NSD) (activity, mobility, and complexity) provide a useful tool for evaluating micro- and macrostructural elements of sleep. In rats, the automatic analysis of sleep recordings by means of NSD calculated from SM and Vis derivations allow different sleep stages to be discriminated. Activity distribution over the total recording permits the definition of a global effect on the EEG. Moreover, the quality of sleep can be evaluated by the variations of the SM activity distribution by considering different classes of amplitude. Index and rate of unstable amplitude segments (UAS) constitute useful parameters to analyse the stability and quality of sleep with different models of insomnia and after pharmacological treatment. The UAS in rats can be compared to CAPs in humans. The NSD are also able to define quality and stability of human sleep. EEG analysis using NSD provides a novel perspective for the analysis of the stability and quality of rat and human sleep. This microstructural analysis of sleep appears to be a useful tool for pharmaco-EEG studies.

Effects of a magnesium-deficient diet on sleep organization in rats.

The influence of magnesium, one of the most important cations in the vertebrate body, on the sleep-wakefulness cycle and ECoG patterns in chronically implanted rats recorded during the light period over a 6-hour period was investigated. Two groups of rats were studied. Group 1 (6 rats): after a control period of 2 weeks, the rats were maintained for 9 weeks on a Mg(2+)-deficient diet. Group 2 (5 rats): after a control period of 2 weeks, the rats were maintained for 7 weeks on a Mg(2+)-deficient diet followed by 4 weeks on a normal diet (recovery period: weeks 8-11). Mg(2+)-deficient diet for 9 weeks induced sleep and ECoG time-dependent alterations. After 6-7 weeks on a Mg(2+)-deficient diet (n = 11) sleep analysis showed a significant increase of wakefulness (+50%) at the expense of slow wave sleep (-24%) but paradoxical sleep was not significantly modified. After 9 weeks of a Mg2+ deficient diet, sleep was disorganized: light sleep and polyspikes occurred indicating an increase in neuronal excitability. When Mg2+ was reintroduced in food and water, sleep organization and ECoG recordings were restored to their original patterns. Our findings which are in line with previous clinical and pharmacological observations provide conclusive arguments for the neuroprotective effect of magnesium ions in neurologic disorders and epileptiform activity. Mg2+ deficiency induces ECoG alterations in the rat which bear some similarities with those seen in neurogenic spasmophilic syndromes in man.