Abstract
GABAA receptors have been implicated in mediating several acute effects of ethanol including anxiolysis, ataxia, sedation/hypnosis, and anticonvulsant activity. Ethanol sensitivity of neurons has been associated with expression of α1 subunit-containing receptors. The objective of this study was to determine the contribution of α1 subunit containing receptors to ethanol responses in comparison to neurosteroids and other anesthetics using GABAA receptor α1 subunit knockout mice. Deletion of α1 subunit-containing receptors did not alter the anxiolytic, ataxic, anticonvulsant, or hypnotic effects of ethanol or acute functional tolerance to ethanol but did increase sensitivity to the locomotor-stimulating effects of ethanol. The ability of ethanol to potentiate muscimol-stimulated chloride uptake and ethanol clearance was also not altered following α1 subunit deletion. The anticonvulsant and hypnotic effects of neurosteroids as well as their potentiating effect on GABA-mediated Cl− uptake were unaltered in α1−/− mice. The hypnotic effect of pentobarbital, etomidate, and midazolam were reduced, whereas the effect of ketamine was enhanced in α1−/− mice. Thus, GABAA receptor α1 subunit-containing receptors appear to influence the motor-stimulating effect of ethanol and the sedative/hypnotic effects of some anesthetics, but not ethanol. These receptors do not appear to be necessary for most ethanol responses, suggesting involvement of other GABAA receptor subtypes or other targets altogether.
GABAAreceptors form the major inhibitory neurotransmitter system expressed in the CNS and are the targets of several classes of drugs including alcohols, inhalation anesthetics, neuroactive steroids, barbiturates, and benzodiazepines. GABAA receptors are heteromeric protein complexes consisting of several homologous membrane-spanning glycoprotein subunits. GABAAreceptor subunits cloned from the mammalian CNS have been divided into classes, with some containing several isoforms: 6α, 4β, 3γ, 1δ, 1ε, 1π, and 1θ (Sieghart and Sperk, 2002). Furthermore, α1 subunits are the most abundant α subunit variant expressed in brain and may serve as important targets for ethanol (McKernan et al., 1991a,b).
Although the behavioral effects of acute systemic ethanol exposure (i.e., anxiolytic, anticonvulsant, ataxic, and sedation/hypnosis) are well established, the neuronal mechanisms underlying these actions remain elusive. Similarities between the acute effects of benzodiazepines (BZDs), neurosteroids, and ethanol implicate GABAA receptors as a site of action for ethanol. Indeed, several findings suggest GABAA receptor involvement in the sedative/hypnotic and anxiolytic actions of ethanol, similar to those of BZDs (for review, see Grobin et al., 1998). At the functional level, physiologically relevant concentrations of ethanol (20–60 mM) potentiate muscimol-stimulated chloride uptake in cerebral cortical synaptoneuorosomes of rats (Suzdak et al., 1986a; Morrow et al., 1988a), mouse cerebellar microsacs (Allan and Harris, 1987), and cultured spinal cord neurons (Mehta and Ticku, 1988). Ethanol potentiation of GABAA receptor-mediated chloride uptake as well as the behavioral effects of ethanol are blocked by GABAA receptor antagonists and inverse agonists (Suzdak et al., 1986b; Ticku and Kulkarni, 1988). Nevertheless, direct interaction of ethanol with neuronal GABAAreceptors using patch-clamp recording techniques has rarely been observed at these ethanol concentrations (Frye et al., 1994; Crews et al., 1996; Marszalec et al., 1998). One recent study demonstrates ethanol (1 mM) enhancement of GABA responses in progesterone-withdrawn rats that exhibit increased α4δ subunit expression (Sundstrom-Poromaa et al., 2002). Although it remains inconclusive whether ethanol's actions through GABAAreceptors are direct or indirect, the subunit composition of the receptor may be an important determinant of ethanol's effects on the CNS.
It has been proposed that the acute effects of ethanol are mediated through α1-containing receptors. Studies using the α1 subunit-selective ligand zolpidem determined a high correlation between [3H]zolpidem binding to neurons and sensitivity to ethanol, suggesting that α1-containing receptors may be involved in mediating ethanol-induced neuronal inhibition (Breese et al., 1993). Ethanol was found to enhance neuronal responses to GABA in the neocortex, medial septum, inferior colliculus, substantia nigra pars reticulata, ventral pallidum, and cerebellum, but not in the hippocampus or ventral tegmental area (Palmer and Hoffer, 1990; Soldo et al., 1994). This localization of ethanol-sensitive GABAA receptors is closely correlated to brain regions exhibiting high-affinity zolpidem binding (Breese et al., 1993;Criswell et al., 1995). Moreover, these brain areas were found to preferentially express α1, β2, and γ2 subunits suggesting that the α1β2γ2 subtype confers ethanol sensitivity (Criswell et al., 1995, 1997).
The effects of chronic ethanol consumption on GABAA receptor function and subunit expression further supports the role of subunit composition in determining ethanol sensitivity. Chronic ethanol exposure produces reduced efficacy of muscimol agonists, BZDs, and ethanol, as observed in functional assays (Allan and Harris, 1987; Morrow et al., 1988b). These changes in GABAA receptor function following chronic ethanol consumption are associated with significant increases in α4, γ2s, and γ1 levels and a decrease in α1 subunit expression (Devaud et al., 1997). Furthermore, the observed changes in subunit expression may explain the behavioral tolerance to ethanol, cross-tolerance to BZDs and sensitization to the anticonvulsant action of neurosteroids following chronic ethanol consumption (Grobin et al., 1998).
Increasing evidence for the role of GABAAreceptors in the acute actions of ethanol have led to the production of several subunit-specific knockout mouse lines used to explore the contribution of subunit subtypes to ethanol action. Although knockout mice of the α6 and γ2L subunits exhibited normal responses to ethanol, δ subunit knockout mice were less sensitive to some ethanol responses, possibly due to reduced sensitivity to neurosteroids (Homanics et al., 1997; Homanics et al., 1999; Mihalek et al., 1999,2001). Initial characterization of α1−/− mice revealed a dramatic loss of GABAA receptor number, BZD sites, a reduction in muscimol-stimulated Cl− uptake, increased susceptibility to bicuculline-induced seizure and a pathologic tremor (Kralic et al., 2002b). Furthermore, loss of α1-containing receptors resulted in a compensatory increase in α2 and/or α3-containing receptors that may contribute to altered functional and behavioral responses to BZD site agonists including reduced sensitivity to zolpidem, which might predict reduced responses to ethanol (Vicini et al., 2001; Kralic et al., 2002b). The goals of the present studies were to identify the effects of α1 subunit deletion on functional and behavioral responses elicited by ethanol, neurosteroids, and anesthetics.
Materials and Methods
Mouse Production.
Genetically engineered male and female mice were produced as described (Vicini et al., 2001) and raised in colonies at the University of North Carolina-Chapel Hill and the University of Pittsburgh. Briefly, the control α1 allele was an unrecombined floxed allele in which exon 8 of the α1 gene was flanked by loxP sites. The amount and distribution of α1 protein produced from this floxed allele does not differ from a true wild-type allele (unpublished observations). The knockout α1 allele was a floxed allele following Cre-mediated recombination. This recombined allele has been demonstrated to be a true null allele (Vicini et al., 2001; Kralic et al., 2002b). Control (α1+/+) and knockout (α1−/−) mice of F4 to F6 generations on a C57BL/6J X Strain 129Sv/SvJ hybrid genetic background were derived from heterozygous breeding pairs. All mice were genotyped by Southern blot analysis of tail DNA (Vicini et al., 2001). After weaning, mice were group housed, given free access to standard rodent chow and water, and maintained on a 12-h alternating light/dark schedule, with lights on at 7:00 AM. All studies were conducted with mice between 7 to 12 weeks of age, were carried out in accordance with the Guide for the Care and Use of Laboratory Animals by the U.S. National Institutes of Health, and were approved by the Institutional Animal Care and Use Committees at the Universities of North Carolina-Chapel Hill and Pittsburgh.
Chloride Uptake Assay.
Following decapitation, brains were immediately removed and placed in ice-cold saline from which cerebral cortices were isolated. Seven sets of cortices per genotype were pooled for each experiment. Synaptoneurosomes were prepared and Cl− uptake was conducted as previously described (Morrow et al., 1988). The synaptoneurosomal pellet was resuspended in 6.6 volumes of ice-cold assay buffer (20 mM Hepes, 118 mM NaCl, 4.7 mM KCl, 1.2 mM MgSO4, 2.5 mM CaCl2, pH 7.4) for a final protein concentration of approximately 5 mg/ml. The homogenate (200 μl) was aliquotted per assay tube and preincubated at 30°C for 12 min. Muscimol-stimulated Cl− uptake was initiated by addition of 0.2 μCi 36Cl (PerkinElmer Life Sciences, Boston, MA) in the presence of an EC30concentration of muscimol (2.5 and 5 μM for α1+/+ and α1−/−samples, respectively) alone or in conjunction with THDOC (1 nM–10 μM) (Steraloids, Newport, RI) or 30 mM ethanol (AAPER Alcohol and Chemical Co., Shelbyville, KY). The respective EC30 of muscimol was used for each genotype since it was previously determined by conducting concentration-response curves that the EC50 andEmax of muscimol were altered following α1 subunit deletion (EC50: 3.7 ± 0.6 and 7.9 ± 1.0 μm; Emax: 30.4 ± 1.1 and 21.7 ± 0.7 nmol of Cl−/mg of protein in α1+/+ and α1−/− mice, respectively) (Kralic et al., 2002b). The solution was vortexed and uptake terminated after 5 s by addition of 4 ml of ice-cold assay buffer containing 100 μM picrotoxin with rapid vacuum filtration over S&S no. 32 filters (Schleicher and Schuell Biosciences, Keene, NH) using a single manifold filter. The synaptoneurosomes were washed twice with 4 ml of buffer, the filter allowed to dry, and radioactive counts determined by liquid scintillation spectroscopy. Chloride uptake was measured in the absence of muscimol and subtracted from all tubes to determine muscimol-stimulated chloride uptake at the respective EC30 concentrations. Net potentiation by drugs was obtained by subtracting muscimol-stimulated chloride uptake from total uptake obtained from drug plus muscimol. Concentration-response curves were evaluated using nonlinear regression by Prism (GraphPad Software, San Diego, CA) to obtain the EC50 andEmax values and compared between genotype by Student's t test.
Ethanol Metabolism and Clearance.
Mice were tested for ethanol clearance and metabolism following injection of ethanol (3.5 g/kg; i.p.). Blood was collected from the retro-orbital sinus at 1 and 3 h postinjection. Plasma ethanol levels (BECs) were determined using a commercial kit (Procedure 333-UV; Sigma-Aldrich, St. Louis, MO). Data were analyzed using SigmaStat and compared by two-way ANOVA.
Open-Field and Elevated Plus Maze.
Naive mice were evaluated for basal anxiety levels and locomotor activity as well as for the anxiolytic and locomotor stimulant effects of ethanol using the elevated plus maze and open-field exploratory observation. Mice were transported to the testing room the night before testing. Animals were weighed and tested between 9:00 and 11:00 AM. Ethanol (0.5, 1.0, or 1.5 g/kg) or saline was injected i.p. after which mice were returned to their home cage for 10 min. Mice were then placed on the center platform of the elevated plus maze and five anxiety- and motor-related behaviors were recorded over a 5-min period: 1) number of open arm entries, 2) time in open arms, 3) number of closed arm entries, 4) time in the closed arms, and 5) total number of arm entries. Following plus maze testing, mice were placed back in their home cages for a 5-min period. Subsequently, mice were tested in the open-field exploratory assay. Each mouse was placed facing outward in a corner of an 18 × 18-inch clear Plexiglas box with a grid of 4.5-inch squares under the floor. There were sixteen total squares, with four squares being in the center not bounded by the wall of the box. The mice were observed over a 10-min period, and the number of crossings were recorded manually. Each time the mouse entered a square with a majority of its body, a crossing was recorded, with differentiation between inner and outer squares being noted. Data were analyzed using SigmaStat using two-way ANOVA with Tukey's post hoc test.
Accelerating Rotarod.
Mice were tested for motor coordination and for the ataxic effects of low dose ethanol on an accelerating rotarod (model 7650; Ugo Basile, Comerio, Italy). To test motor coordination and motor learning, each mouse performed a single daily trial for 9 consecutive days. Saline was administered i.p 10 min before the trial on day 8 to habituate the mice to injection. On the final day (trial 9), each mouse was injected with ethanol (1.5 g/kg; ip) and tested 10 min later on the rotarod. All trials were conducted for a total of 180 s, during which the rotarod increased in speed from 5 to 40 rpm. The time each mouse remained on the rod was recorded. Data were analyzed with SigmaStat using repeated measures two-way ANOVA with Tukey's post hoc test.
Acute Functional Tolerance.
Acute functional tolerance assays the rapid development of tolerance to ethanol and is based on the ataxic effects of ethanol (Erwin and Deitrich, 1996). Mice were acclimated to a stationary 2.5-cm diameter rod for 1 min (Rotarod, model 7650, Ugo Basile). The mice were then injected with ethanol (1.75 g/kg; i.p.) and tested again on the stationary rod for balance. Once the mice could remain on the rod for 60 s (t1), a blood sample was collected retro-orbitally to determine the first blood-ethanol concentration (BEC1) and then immediately injected with 2.0 g/kg ethanol. The mice were subsequently tested every 5 min for the ability to remain on the rod for 60 s. Once this was achieved (t2), a second blood sample was drawn for blood ethanol determination (BEC2). Acute functional tolerance (AFT) was determined as the difference between BEC2 and BEC1. Parameters AFT, BEC1, BEC2, t1, andt2 were statistically analyzed by Student's test using SigmaStat.
Bicuculline-Induced Seizure Threshold Test.
Seizure thresholds were determined at the beginning of the light cycle from 8:00 AM to 12:00 PM in a room adjacent to the colony room under normal lighting conditions, with low-level white background noise as previously described (Devaud et al., 1995). Mice were injected i.p. with ethanol (2 or 3 g/kg), allopregnanolone (4, 8 or 16 mg/kg) or respective vehicles (saline or 20% 2-hydroxypropyl-β-cylcodextrin/saline) in a 10 ml/kg volume 60 and 15 min, respectively, before seizure threshold determination. Mice were restrained in a Plexiglas plunger-style mouse restraint (Braintree Scientific, Inc., Braintree, MA). Threshold determination was made by constant lateral tail vein infusion via a 28-gauge butterfly needle of bicuculline (Sigma-Aldrich) dissolved in 0.1 N HCl and diluted with isotonic saline to a final concentration of 0.05 mg/ml, pH 7. The solution was infused at a constant rate of 0.5 ml/min; the endpoint taken as the first myoclonic jerk of the head and neck. This time point precedes forepaw clonus and generalized tonic/clonic convulsions. Each animal was tested once. Seizure thresholds were determined by experienced observers who were blind to the experimental conditions. Seizure thresholds were calculated from the time of infusion × dose of bicuculline per body weight and presented as milligrams per kilogram bicuculline. Data were analyzed using SigmaStat using two-way ANOVA with Tukey's post hoc test.
Loss of Righting Reflex.
Mice were tested for the duration of the loss of the righting reflex (i.e., sleep time) in response to ethanol (3.0 and 3.5 g/kg; Pharmco, Brookfield, CT), the BZD midazolam (75 mg/kg; ESI Lederle, Philadelphia, PA), the neurosteroid pregnanolone (8 mg/kg; Sigma-Aldrich), the barbiturate pentobarbital (45 mg/kg; Abbott, Chicago, IL), and the anesthetics propofol (4 mg/kg; AstraZeneca, Wilmington, DE), ketamine (150 mg/kg; Fort Dodge, Fort Dodge, IA), and etomidate (Bedford Laboratories, Bedford, OH; 20 mg/kg) as previously described (Mihalek et al., 1999). Doses were chosen based on their capacity to induce hypnosis. All agents were injected i.p. except for pregnanolone and propofol, which were injected intravenously via the retro-orbital sinus. Upon becoming ataxic, mice were placed on their backs in a v-shaped trough, and the time “down” was noted. Mice were monitored until they could right themselves three times in 30 seconds, and the time was recorded when they passed this criteria. A heat lamp and random rectal temperature measurements were used to ensure normothermia. Mice were tested in three groups with a 7 day period between each drug: Group 1 received midazolam and etomidate; group 2 received diazepam, propofol, pregnanolone, and ketamine; group 3 received ethanol only. All assays were performed by an investigator blinded to the genotypes of the animals being tested. Mice in the ethanol, ketamine, and etomidate experiments were determined to have a misplaced injection and excluded from data analysis if the time to lose the righting reflex after injection was greater than 4 min (Ponomarev and Crabbe, 2002). Mice in all other sleep time experiments were determined to have a misplaced injection and excluded from data analysis if the time to lose the righting reflex following injection was greater than two standard deviations from the group mean. Data were evaluated using a two-way ANOVA with sex and genotype as factors. If no effect of sex was observed, data were collapsed within genotype and reanalyzed by ANOVA.
Results
Ethanol and THDOC Potentiation of Muscimol-Stimulated Chloride Uptake.
Previous results demonstrated that the potency and efficacy of muscimol-stimulated Cl− uptake was reduced in α1−/− mice (Kralic et al., 2002a). Ethanol and THDOC potentiation of muscimol-stimulated chloride uptake was measured in synaptoneurosomal preparations of cerebral cortex from α1+/+ and α1−/− mice using respective EC30 concentrations of muscimol. Ethanol and THDOC potentiated the effects of muscimol on chloride uptake in α1+/+ mice, confirming previous reports (Suzdak et al., 1986b; Allan and Harris, 1987; Morrow et al., 1988b). Net potentiation by ethanol (30 mM) over EC30 muscimol-stimulated chloride uptake was similar between α1+/+ and α1−/− (3.2 ± 0.6 and 3.8 ± 0.8 nmol of Cl−/mg of protein), respectively (Fig.1A). THDOC (1 nM–10 μM) potentiated chloride uptake in both α1+/+ and α1−/− mice in a dose-dependent manner resulting in similar potency of THDOC (522 ± 76 and 486 ± 59 nM) and maximal efficacy (15 ± 1 and 14 ± 1 nmol of Cl−/mg of protein) (Fig. 1B).
Ethanol Metabolism and Clearance.
To determine whether α1−/− mice differed from their α1+/+ littermates with respect to ethanol pharmacokinetics, the BEC at 60 and 180 min was measured after i.p. injection of 3.5 g/kg ethanol. Female mice exhibited significantly higher BECs than male mice (235 ± 9 and 186 ± 10 mg/dl, respectively) at the 180 min time point only [two-way ANOVA; gender:F(1,22) = 14.3, p < 0.001]. Nevertheless, there was no effect of genotype on the BEC at 60 min (α1+/+, 362 ± 16 and α1−/−, 341 ± 16 mg/dl) or 180 min (α1+/+, 203 ± 9 and α1−/−, 218 ± 10 mg/dl), thereby allowing valid comparisons between genotypes for behavioral responses to ethanol.
Elevated Plus Maze.
Baseline performance on the elevated plus maze did not vary significantly with genotype, as previously reported (Kralic et al., 2002a). Total arm entries, a measure of locomotor activity, were increased by ethanol in both α1+/+ and α1−/− mice (Fig. 2A) [two-way ANOVA; dose:F(3,155) = 9.3, p < 0.001]. α1−/− mice were more sensitive to the locomotor-activating effects of ethanol at the 1.0-g/kg dose compared with α1+/+ mice, however (Fig. 2A) (Tukey's post hoc, p < 0.05). To determine whether the deletion of α1 subunits altered the anxiolytic effect of ethanol, the behavior of α1+/+ and α1−/− mice was evaluated following ethanol administration (0.75, 1.0, and 1.5 g/kg) in the elevated plus maze. As presented in Fig. 2, ethanol administration produced an anxiolytic effect in both α1+/+ and α1−/− mice. Ethanol increased both the percentage of open arm entries (Fig. 2B) [two-way ANOVA; GT:F(1,155) = 14.3, p < 0.001; dose: F(3,155) = 13.7,p < 0.001] and the percentage of time spent in open arms (Fig. 2C) [two-way ANOVA; dose:F(3,155) = 13.5, p < 0.001]. Since no effect of gender was detected on any parameter, data from both genders was collapsed for these analyses.
Exploratory Activity.
Baseline exploratory activity in the open field did not vary significantly with genotype. Sensitivity to the locomotor stimulating effect of ethanol, however, was enhanced in α1−/− mice. Ethanol (0–1.5 g/kg) administration produced a 65 to 125% increase in total square entries (Fig. 3A) [two-way ANOVA; GT:F(1,149) = 10.0, p < 0.01; dose: F(3,149) = 6.7,p < 0.001; GT × dose:F(3,149) = 4.0, p < 0.01] in α1−/− with no change in α1+/+ mice. The anxiolytic effect of ethanol assessed as the percentage of center square entries did not significantly vary with dose or genotype (Fig. 3B). Since no effect of gender was detected in this test, data from both genders was collapsed for these analyses.
Accelerating Rotarod.
Since motor coordination involves proper cerebellar GABAA receptor function and cerebellum contains a relatively high proportion of α1-containing receptors, the rotarod test was employed to determine whether deletion of α1 subunit-containing receptors affected baseline motor coordination and the ataxic effects of acute ethanol administration. Over the first seven trials of baseline measurements, the ability of the mice to remain on the rod increased over the trials, but there was no significant difference between α1+/+ and α1−/− mice with respect to the amount of time able to remain on the rod in six of seven trials [two-way repeated measures ANOVA, Trial: F(6,412) = 33.78, p < 0.001; genotype × trial:F(6,412) = 3.24, p < 0.01] (data not shown). Compared with females, however, males showed reduced ability over all seven trials to remain on the rod, although mice of both genders reached similar criterion by trial 7 [two-way repeated measures ANOVA, gender:F(1,412) = 13.56, p < 0.001; trial: F(6,412) = 33.0,p < 0.001; gender × trial:F(6,412) = 3.23, p < 0.01] (data not shown). Control α1+/+ and α1−/− mice exhibited a similar performance on the rotarod following saline injection (trial 8) and a similar ataxic effect following ethanol (1.5 g/kg, i.p.) administration (trial 9; Fig.4) [two-way repeated measures ANOVA; trial: F(1,116) = 25.98,p < 0.001]. Since no effect of gender was detected, data from both genders was collapsed for analysis of trials 8 and 9.
Acute Functional Tolerance.
To determine whether GABAA α1-containing receptors contribute to the development of tolerance to an acute injection of ethanol, acute functional tolerance to ethanol was measured. The test for acute functional tolerance to ethanol measured four parameters:t1, the time to regain balance on a stationary rod after an initial injection of ethanol; BEC1, blood ethanol concentration at t1;t2, time to regain balance after a second ethanol injection; and BEC2, blood ethanol concentration att2. Comparison of α1+/+ and α1−/− mice revealed no difference in the motor ataxic effects of ethanol att1 ort2 or respective BECs. Furthermore, the AFT parameter was similar between genotypes (Table1). Since no effect of gender was detected on any parameter, data from both genders was collapsed for these analyses.
Anticonvulsant Effects of Allopregnanolone and Ethanol.
The bicuculline-induced seizure threshold assay was employed to determine the anticonvulsant effect of ethanol and allopregnanolone in α1+/+, α1+/−, and α1−/− mice. Our previous study demonstrated that bicuculline seizure thresholds were reduced in α1−/− mice (Kralic et al., 2002a). Ethanol administration (2 and 3 g/kg) produced an anticonvulsant effect in all genotypes (Fig. 5A) [two-way ANOVA; GT:F(2,59) = 21.0, p < 0.001; dose: F(2,59) = 21.08,p < 0.001]. At the highest dose, ethanol increased the seizure threshold to a similar magnitude; furthermore, the potency of ethanol did not vary with genotype. Allopregnanolone administration (4, 8, and 16 mg/kg) produced a dose-dependent anticonvulsant effect in all genotypes (Fig. 5B) [two-way ANOVA; GT:F(2,85) = 31.1, p < 0.001; dose: F(3,85) = 16.81,p < 0.001]. At the highest dose, allopregnanolone increased the seizure threshold to a similar magnitude; furthermore, the potency of allopregnanolone did not vary with genotype. Since no effect of gender was detected in either study, data from both genders was collapsed for these analyses.
Drug-Induced Hypnosis.
The loss of righting reflex (LORR) assay was conducted to determine whether the hypnotic effects of ethanol (3 and 3.5 g/kg, i.p.), pregnanolone (8 mg/kg), pentobarbital (45 mg/kg), midazolam (75 mg/kg), etomidate (20 mg/kg), ketamine (150 mg/kg), and propofol (33 mg/kg) were altered following deletion of α1 subunit-containing receptors. As shown in Fig.6, the duration of LORR induced by ethanol, pregnanolone, and propofol did not differ between α1+/+ and α1−/− mice. The duration of pentobarbital and midazolam-induced LORR were reduced in α1−/− mice [Student's t test,p < 0.001 and p < 0.05, respectively]. The hypnotic effect of etomidate was greater in female α1+/+ than male α1+/+mice and reduced by deletion of α1 subunits in female α1−/− mice only [two-way ANOVA; GT:F(1,30) = 5.8, p < 0.05; gender: F(1,30) = 7.6,p < 0.05]. The duration of LORR induced by ketamine was greater in α1−/− mice than in α1+/+ mice [Student's t test,p < 0.05]. Data from both genders was collapsed for all drugs except etomidate for these analyses since no effect of gender was detected.
Discussion
Previous examination of mice with a targeted deletion of the GABAA receptor α1 subunit revealed loss of ∼50% of all GABAA receptors and altered functional and behavioral responses to BZD site agonists (Kralic et al., 2002a). Surprisingly, deletion of all α1-containing receptors did not alter the ability of ethanol or allopregnanolone to potentiate GABAA receptor function. Moreover, the anxiolytic, ataxic, anticonvulsant, or hypnotic effects of ethanol and the development of acute functional tolerance were unaffected by deletion of α1 subunits. The anticonvulsant and hypnotic effects of neurosteroids were unaltered following deletion of α1 subunits, as predicted by results of neurosteroid potentiation of GABAA receptor function. In contrast, α1−/− mice were more sensitive to the locomotor stimulating effects of ethanol. These results suggest that most of the effects of ethanol and neurosteroids are not dependent upon the expression of α1-containing GABAAreceptors. Low-dose ethanol responses such as the stimulating effects may be dependent upon α1-containing receptors, however. Moreover, the duration of LORR induced by ethanol, pregnanolone, and propofol were unaffected by deletion of α1 subunits, whereas the effects of pentobarbital, midazolam, etomidate, and ketamine were significantly altered, demonstrating the selectivity of α1 subunit deletion on drug responses.
The absence of any effect of α1 subunit deletion on the ability of ethanol to potentiate GABAA receptor-mediated chloride uptake suggests that ethanol's actions are not selectively mediated through α1 subunit-containing GABAAreceptors. These results do not support findings in which α1 subunit expression/zolpidem responses predicted the sensitivity of neurons to ethanol (Criswell et al., 1995; Duncan et al., 1995). Furthermore, our results suggest that the reduction in α1 subunit expression following chronic ethanol consumption does not underlie the decreased sensitivity or tolerance to the functional effects of acute ethanol (Morrow et al., 1988a; Devaud et al., 1995, 1997).
As with ethanol, neurosteroid potentiation of GABA-mediated chloride uptake was not altered following deletion of α1 subunits. This result supports findings from recombinant receptor systems in which there was no or only a modest contribution of α subunit subtypes to the pharmacological profile of neurosteroids (Puia et al., 1993; Lambert et al., 2001). The δ subunit, whose expression is unaltered in α1−/− mice (Kralic et al., 2002b), appears to be influential in mediating sensitivity to neuroactive steroids in vivo as well as sensitivity to the anticonvulsant effects of ethanol (Mihalek et al., 1999, 2001). The presence of δ subunits in the receptor have been shown to confer increased neurosteroid efficacy in vitro (Brown et al., 2002).
Both ethanol and neurosteroids have been shown to produce anticonvulsant effects in several different seizure paradigms (Kokate et al., 1994; Finn et al., 1995; Devaud et al., 1996). The lack of effect of α1 subunit deletion on the anticonvulsant effects of these compounds suggests that their actions are not selective for α1-containing receptors. Furthermore, the absence of any effect of α1 subunit deletion helps to define which changes in GABAA receptor subunit expression may underlie the increased sensitivity to neurosteroids following chronic ethanol consumption (Devaud et al., 1996). Therefore, the decrease in α1 subunit expression following chronic ethanol consumption is unlikely to contribute to the increased sensitivity to the anticonvulsant effects of neurosteroids (Devaud et al., 1997). Studies in δ subunit knockout mice suggest that the anticonvulsant of ethanol is influenced by the presence of the δ subunit (Mihalek et al., 2001). Furthermore, the absence of any effect on the functional and anticonvulsant response of neuroactive steroids, selective modulators of GABAA receptors, suggests that the reduction in GABAA receptor number does not account for changes in the anticonvulsant activity of other compounds (e.g., BZDs) (Kralic et al., 2002a).
The surprisingly normal behavioral responses to ethanol observed in α1−/− mice suggest that mechanisms involving indirect interaction of ethanol with GABAAreceptors through induction of intermediary proteins, endogenous modulators, or action on non-GABAergic systems altogether may be responsible for many behavioral actions of ethanol. These results are also compatible with other GABAA receptor subunits being directly involved in ethanol's mechanism of action. Ethanol may also exert its effects by modulating post-translational modification pathways since targeted deletion of protein kinase Cγ and protein kinase Cε reduced and increased, respectively, the sensitivity to acute ethanol (Harris et al., 1995; Hodge et al., 1999;Bowers et al., 2001). In addition, the ability of systemic ethanol administration to induce physiologically relevant levels of allopregnanolone is required to observe certain behaviors and electrophysiological effects of ethanol (Morrow et al., 1999; Khisti et al., 2000; VanDoren et al., 2000). Since various ethanol effects are postulated to involve neurosteroid intermediates, it is noteworthy that α1 subunit deletion lacked an effect on responses to either modulator. This observation is consistent with the hypothesis that both of these modulators have indirect actions on GABAA receptors (Morrow et al., 2001). Although adaptations in GABAA receptor subunit expression, notably α2 and α3, could compensate for the loss of α1 expression in knockout mice masking changes in ethanol sensitivity that would be otherwise observed, this outcome would argue against subunit selectivity of ethanol (Kralic et al., 2002a,b). Lastly, since ethanol has been shown to affect other neurotransmitter systems, the behavioral effects of ethanol described may be mediated by a non-GABAergic system mechanism.
The differential effects of α1 subunit deletion on the hypnotic effects of the anesthetics tested may suggest selectivity of anesthetics for α1-containing receptors or the importance of adaptations in GABAA receptors or other systems. The selective effects of α1 subunit deletion on the hypnotic effects of pentobarbital and etomidate over ethanol, pregnanolone, and propofol suggest a contribution of α1-containing receptors to these actions. The enhanced hypnotic effects of ketamine, anN-methyl-d-aspartate antagonist, suggest a compensatory adaptation of the glutamatergic system in response to the loss of GABAA receptors or a demonstration of the imbalance in CNS excitability resulting from α1 subunit deletion. The reduced hypnotic effect of midazolam in α1−/− mice was unexpected since it had been previously shown that diazepam, another member of the BZD class, exhibited enhanced effects in the same test (Kralic et al., 2002a). These results suggest that these compounds may have unique pharmacological profiles such as selectivity for type I and II BZD sites. Indeed, midazolam, rather than diazepam, is prescribed for its anesthetic properties. Since etomidate has been shown to act upon β3-containing receptors for high-dose effects (Jurd et al., 2002), the reduced effect of etomidate in the LORR assay is supported by the reduced expression of β2/3 in α1−/− mice (Kralic et al., 2002a).
The hypnotic effects of some of these same drugs have been studied in an independently generated α1 knockout mouse line (Blednov et al., 2003). Results with zolpidem and some benzodiazepines are consistent between the different mouse lines (Kralic et al., 2002a; Blednov et al., 2003). Surprisingly, the sleep time in response to some drugs appears to differ between the two α1 knockout mouse lines. For example, Blednov et al. (2003) report a reduction in ethanol-induced sleep time in males only, whereas we did not detect a difference in either sex. We observed a reduction in etomidate-induced sleep time in female knockouts and in pentobarbital sleep time in both sexes, whereasBlednov et al. (2003) observed no difference in either sex for both of these drugs. The differences between drug responses of the two α1 knockout mouse lines are likely the result of differences in breeding strategies or genetic backgrounds or less likely the result of differences in experimental conditions.
In conclusion, we have demonstrated that deletion of α1-containing receptors does not significantly alter the functional and most behavioral responses to ethanol or neurosteroids, suggesting that the pharmacological effects of these compounds are not selectively mediated by α1-containing receptors. Moreover, loss of over half of GABAA receptor expression following deletion of α1 subunits does not alter most of the behavioral effects of acute ethanol administration and suggests that the molecular basis of ethanol actions remains elusive. In contrast, the hypnotic effects of midazolam, pentobarbital, and etomidate are altered by deletion of α1 subunits, emphasizing that subunit specificity can play a role in the responses of many GABAergic modulators.
Acknowledgments
We would like to thank Joanne Steinmiller for expert technical assistance.
Footnotes
-
Supported by National Institutes of Health Grants AA09013 and AA11605 to A.L.M. and GM52035, GM47818, and AA10422 to G.E.H.
-
DOI: 10.1124/jpet.102.048124
- Abbreviations:
- CNS
- central nervous system
- BZD
- benzodiazepine
- BEC
- blood-ethanol concentration
- ANOVA
- analysis of variance
- AFT
- acute functional tolerance
- THDOC
- tetrahydrodeoxycorticosterone
- LORR
- loss of righting reflex
- GT
- genotype
- Received December 19, 2002.
- Accepted January 30, 2003.
- The American Society for Pharmacology and Experimental Therapeutics