Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

GABA, acting via GABA_A receptors, is the brain's major inhibitory neurotransmitter system in the central nervous system. Because the ability to enhance the ion channel activation of GABA_A receptors is a common feature of many sedative and hypnotic drugs, it is probable that the action on the GABA_A receptor complex is a molecular target for these drugs in the mammalian central nervous system (Charney et al., 2001). Nevertheless, many of these drugs, including ethanol and pentobarbital, may have multiple sites of action and the importance of specific GABA_A receptors in drug action is not clearly defined.

The GABA_A receptor has a pentameric structure, which is formed by the coassembly of subunit polypeptides that exist in a large multigene family (McKernan and Whiting, 1996; Barnard et al., 1998). There are at least 16 different members of the GABA_A receptor gene family, including 6, 3, 3, , , , and  subunits (Whiting et al., 1999). The largest population of GABA_A receptors in the rat brain has a subunit composition of ₁₂₂, whereas together ₂₃₂ and ₃₂/₃ constitute the next most prevalent subtypes (McKernan and Whiting, 1996). GABA affinity is mainly governed by  subunits (Smith et al., 2001), but affinity and efficacy at the benzodiazepine site are influenced by both  and  subunits (McKernan et al., 1995; Buhr et al., 1996; Wingrove et al., 1997) but not  subunits (Hadingham et al., 1993). The benzodiazepine site occurs at the interface of an  and a  subunit, with residues in both influencing modulation (Smith and Olsen, 1995). The presence of ₂ confers the classical benzodiazepine pharmacology to GABA_A receptors (Pritchett et al., 1989). In contrast,  subunits control loreclezole and etomidate sensitivity (Stevenson et al., 1995). Mouse strains lacking individual GABA_A receptor subunits provide insight regarding the role of GABA receptors. Mice lacking the ₂ subunit die shortly after birth (Gunther et al., 1995), whereas mice deficient the _2L (long splice variant) subunit are viable and show small increases in sleep time responses to midazolam and zolpidem, but responses to nonbenzodiazepine agents such as ethanol, etomidate, and pentobarbital are unchanged (Quinlan et al., 2000). ₃ null mice (/) did not show any changes in sleep times after administration of pentobarbital or ethanol, but they were more resistant to etomidate and midazolam (Quinlan et al., 1998). Mice lacking the ₆ subunit of the GABA_A receptor, which is expressed exclusively in cerebellar granule cells, have no major phenotypic abnormalities (Jones et al., 1997). Mice deficient the  subunit are also viable but show attenuated sensitivity to neuroactive steroids and epileptic seizures (Mihalek et al., 1999). Thus, deletion of some of the less abundant GABA_A receptor subunits reduces the action of some sedative drugs, but they do not provide evidence for a major role for GABA receptors in actions of many sedatives. It is important to note, however, that the only predominant GABA_A subunit that has been deleted is the ₂ and this proved lethal.

Recently, mice lacking the most predominant GABA_A receptor subunits, ₁ and ₂, were successfully generated (Sur et al., 2001; Kralic et al., 2002). Although the mice lack approximately 60% of the total number of brain GABA_A receptors, adult ₁ (/) and ₂ (/) mice do not display major phenotypic abnormalities or spontaneous seizures. ₁ (/) mice showed overexpression of the ₂ and ₃ subunit but lack approximately 40% of the GABA_A receptors despite this apparent compensation. In contrast, ₂ (/) mice displayed an equal reduction in all six  subunits and a loss of GABA_A receptors of about 50% (Sur et al., 2001).

In this study, we asked how this substantial decrease of benzodiazepine receptors and reduction of expression of different GABA receptor subtypes might affect receptor function by measuring agonist stimulated ³⁶Cl uptake in brain tissue from these mutant mice. For the behavioral studies, we used the loss of righting reflex and tested a GABA_A agonist, 4,5,6,7-tetrahydroisoxazolo(5,4-c)pyridin-3-ol (THIP), a nonselective benzodiazepine, flurazepam, and ₁-selective drug, zolpidem, a -selective drug, etomidate, and two drugs with activities on several channels in addition to GABA_A, ethanol and pentobarbital.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Animals. Null ₁ (/) and ₂ (/) allele mice were created using homologous recombination and genotyped as previously described in Sur et al. (2001). Homozygotes of the F2 generation were interbred avoiding any brother-sister mating, and homozygous colonies of ₁ (/), ₂ (/) and wild-type (+/+) mice were established. In this study, the mice from the F6 to F7 generations of this interbreeding have been used.

Loss of Righting Reflex. Sensitivity to ethanol (Aaper Alcohol and Chemical Co., Shelbyville, KT; 3.4, 3.6, and 3.8 g/kg, i.p.), flurazepam (Sigma-Aldrich, St. Louis, MO; 225.0 mg/kg, i.p.), zolpidem tartrate (gift from Dr. J. S. Vedo; Pharmacia, Skokie, IL; 60.0 mg/kg, i.p.), pentobarbital (Sigma/RBI, Natick, MA; 50.0 mg/kg, i.p.), etomidate (40.0 mg/kg, i.p.), and THIP (Sigma/RBI; 55.0 mg/kg, s.c.) was determined using the standard sleep time assay (Kakihana et al., 1966). Ethanol was diluted in 0.9% saline (20.0% w/v) and administered in doses adjusted by injected volumes. Flurazepam, zolpidem, and etomidate were dissolved in 3 to 4 drops of Tween 80 (Sigma-Aldrich) before saline was added and injected at 0.01 ml/g b.wt. Pentobarbital was dissolved in 0.9% saline and injected at 0.01 ml/g b.wt. THIP was dissolved in 0.9% saline and injected at 0.005 ml/g b.wt.

³⁶Cl Uptake. The procedure was carried out according to Allan and Harris (1986). Animals were killed by decapitation, and their brains were removed and placed in ice-cold buffer (145 mM NaCl, 5 mM KCl, 1 mM MgCl₂, 10 mM glucose, 1 mM CaCl₂, and 10 mM HEPES; adjusted to pH 7.5 with Tris base). The brain was homogenized by hand (10-12 strokes) using a glass-Teflon homogenizer (size C; Thomas Scientific, Swedesboro, NJ) in 5 ml of ice-cold assay buffer. The homogenate was centrifuged at 900g for 15 min. The supernatant was decanted, and the pellet was washed with 10 ml of assay buffer and centrifuged at 900g for 15 min. The final pellet was suspended in ice-cold assay buffer. Aliquots (0.2 ml) of membrane vesicles (0.6-0.8 mg of protein) were incubated for 5 min in a shaking water bath at 30°C. After this incubation, uptake was initiated by the addition of 0.2 ml of ³⁶Cl (2 mCi/ml assay buffer) (specific activity 12.8 mCi/g of Cl; obtained from ICN, Irvine, CA) containing 1 to 20 µM muscimol (final concentration). Three seconds after the addition ³⁶Cl influx was terminated by the addition of 4 ml of ice-cold assay buffer containing 100 µM picrotoxin, and rapid filtration under vacuum (10 mm Hg) onto a Whatman GF/C glass microfiber filter (Whatman, Clifton, NJ) using a Hoeffer manifold (Hoefer Scientific, San Francisco, CA) was performed. The filters were washed with an 8-ml assay buffer containing 100 µM picotoxin with the manifold towers removed. The amount of radioactivity on the filters was determined by liquid scintillation spectrometry. The amount of ³⁶Cl bound to the filter in the absence of membranes (no tissue blank) was subtracted from all values. Muscimol-dependent influx was defined as the amount of ³⁶Cl taken up while muscimol was present minus the amount of ³⁶Cl taken up in the absence of muscimol. The apparent potency (EC₅₀) and apparent efficacy (E_max) of muscimol was determined by the construction of a dose-response curve using five concentrations of the drug. The modulation of muscimol-stimulated ³⁶Cl uptake by ethanol (25 mM), pentobarbital (25 µM), and flunitrazepam (0.1 µM) was determined by comparing ³⁶Cl uptake in solutions of each different type of drug containing ³⁶Cl and 0.65 µM of muscimol for wild-type (+/+) cortex and ³⁶Cl and 1.0 µM of muscimol for cortex from ₁ (/) and ₂ (/) null mutant mice with solutions containing only ³⁶Cl and the corresponding muscimol concentrations for each genotype. These concentrations of muscimol induced a similar stimulation of uptake in all three genotypes. Net uptake was defined as the amount of ³⁶Cl uptake in the presence of muscimol, corresponding to each genotype concentration, plus the drug tested minus the ³⁶Cl uptake in the presence of muscimol alone.

Data Analysis. The results are expressed as mean ± S.E.M. The data are reported in minutes for sleep time. For ³⁶Cl uptake studies, E_max was determined from the dose-response data as the change in ³⁶Cl flux at the maximally effective concentration, and EC₅₀ was determined by linearly transforming the data (sigmoid curve analysis) using the GraphPad computer program (GraphPad Software, Inc., San Diego, CA). Statistical analysis of parameters was made using the Student's t test with Dunnet's correction for multiple comparisons (Winer, 1971) or by two-way ANOVA whenever appropriate. For all statistical comparisons, p  0.05 was considered significant.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

The duration of loss of righting reflex (sleep time) produced by ethanol was decreased in both null allele males compared with wild-type mice (P < 0.0001, dependence on genotype; P < 0.0001, dependence on dose, two-way ANOVA) (Fig. 1a). The residual response to ethanol was similar in ₁ and ₂ null mutant males. There were no differences in ethanol-induced sleep time in mutant females (Fig. 1b), however. Null allele mice of both strains did not differ from wild-type mice in their responses to pentobarbital (Fig. 1, c and d).

View larger version (40K): [in this window] [in a new window]   Fig. 1.   Ethanol- and pentobarbital-induced loss of righting reflex in mice lacking the 1 or 2 subunit of the GABAA receptor. a and c, show data for males. a, n = 10-12 (+/+); n = 10 1 (/); 2 (/) = 10 for each dose of ethanol. c, n = 10 for each genotype. b and d, show data for females. b, n = 10-12 (+/+); n = 10 1 (/); 2 (/) = 10 for each dose of ethanol. d, n = 10 for each genotype. Summary of statistics: a, males showed an effect of genotype (F2,85 = 27.1; P < 0.0001, two-way ANOVA) and dose (F2,85 = 42.1; P < 0.0001, two-way ANOVA). WT, wild-type.

Analysis of flurazepam and THIP-induced loss of righting reflex revealed that ₁ (/) and ₂ (/) null mutant mice showed markedly decreased sleep time in comparison with wild-type mice (Fig. 2, a and b). Both sexes of mutant mice showed similar changes in response.

View larger version (36K): [in this window] [in a new window]   Fig. 2.   THIP- and flurazepam-induced loss of righting reflex in mice lacking the 1 or 2 subunit of the GABAA receptor. a and c, show data for males. a, n = 6 (+/+); n = 7 1 (/); 2 (/) = 8. c, n = 9 (+/+); n = 8 1 (/) and 2 (/). b and d, show data for females. b, n = 9 (+/+); n = 7 1 (/); 2 (/) = 6. d, n = 9 for each genotype. *, p < 0.05; **, p < 0.01; ***, p < 0.001 mutant mice are different from wild-type mice (Student's t test with Dunnet's correction for multiple comparisons). #, p < 0.05; ###, p < 0.001 1 (/) null mutant mice are different from 2 (/) knockout mice (Student's t test with Dunnet's correction for multiple comparisons). WT, wild-type.

The duration of sleep time induced by etomidate was significantly decreased only in ₂ (/) males (Fig. 3a) and females (Fig. 3b). ₁ (/) mice of both sexes did not differ from wild-type animals in their sensitivity to etomidate. In contrast, ₁ (/) mice were more resistant to zolpidem (60 mg/kg) in comparison with wild-type animals (Fig. 3, c and d). There was no difference in resistance to zolpidem-induced sleep time between ₂ (/) and ₁ (/) females (Fig. 3d). On the other hand, ₂ (/) males showed resistance intermediate between ₁ (/) null mutant and wild-type males (Fig. 3c).

View larger version (37K): [in this window] [in a new window]   Fig. 3.   Etomidate- and zolpidem-induced loss of righting reflex in mice lacking the 1 or 2 subunit of the GABAA receptor. a and c, show data for males. a, n = 7 (+/+); n = 6 1 (/) and 2 (/). c, n = 11 (+/+); n = 12 1 (/); n = 13 2 (/). b and d, show data for females. b, n = 9 (+/+); n = 10 1 (/) and 2 (/). d, n = 9 (+/+) and 1 (/); n = 12 2 (/). *, p < 0.05; **, p < 0.01; ***, p < 0.001 mutant mice are different from wild-type mice (Student's t test with Dunnet's correction for multiple comparisons). ##, p < 0.01; ###, p < 0.001 1 (/) null mutant mice are different from 2 (/) knockout mice (Student's t test with Dunnet's correction for multiple comparisons). WT, wild-type.

To determine whether the deletion of ₁ or ₂ subunits changed the GABA_A receptor function, muscimol-stimulated ³⁶Cl influx was measured in cortical microsacs. Deletion of either the 1 or 2 subunits reduced the response to muscimol (Fig. 4a). The muscimol EC₅₀ was approximately 2-fold higher in both mutant strains than in wild-type mice. The deletion of the ₂ subunit significantly increased the slope of muscimol response. The maximal effect of muscimol was approximately 25% higher in membranes from wild-type animals compared with the null mutants (Table 1).

View larger version (41K): [in this window] [in a new window]   Fig. 4.   Effects of deletion of the 1 or 2 subunit of the GABAA receptor on muscimol-stimulated chloride flux measured in membranes from cortex. a, n = 11 for each genotype. b to d, n = 12 (+/+); n = 13 1 (/) and 2 (/). *, p < 0.05 mutant mice are different from wild-type mice (Student's t test with Dunnet's correction for multiple comparisons). #, p < 0.05 1 (/) null mutant mice are different from 2 (/) knockout mice (Student's t test with Dunnet's correction for multiple comparisons).

View this table: [in this window] [in a new window]   TABLE 1 Function of GABAA receptors in null mutant mice: parameters of muscimol-stimulated 36Cl uptake Values are the mean ± S.E.M. P values refer to statistical differences from wild-type mice (Student t test with Dunnet's correction for multiple comparisons).

Flunitrazepam enhanced muscimol-stimulated ³⁶Cl uptake equally in microsacs from wild-type and ₂ null mutant mice. Potentiation of muscimol response by flunitrazepam, however, was significantly higher in membranes from ₁ (/) mice than from wild-type mice (Fig. 4b). In contrast, there were no differences in potentiation of muscimol response by pentobarbital in the three genotypes (Fig. 4c). Ethanol potentiation of muscimol-stimulated ³⁶Cl influx was significantly decreased in ₂ (/) mice (Fig. 4d). There were no differences in these ethanol effects between males and females. Ethanol potentiated muscimol-stimulated ³⁶Cl influx by 1.70 ± 0.16- and 1.69 ± 0.14-fold in wild-type mice, 1.62 ± 0.22- and 1.75 ± 0.23-fold in ₁ (/) knockout mice, and 1.37 ± 0.07- and 1.15 ± 0.16-fold in ₂ null mutant mice (males and females, respectively).

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

Taken together, these results show that deletion of either the ₁ or ₂ subunits of GABA_A receptor reduced the behavioral effects of a wide variety of sedative hypnotic drugs (summarized in Table 2) and indicate that these mutant mice have a reduced number of functional brain GABA_A receptors.

View this table: [in this window] [in a new window]   TABLE 2 Summary of changes in drug-induced loss of righting reflex in 1 and 2 null mutant mice

The reductions in sleep time produced by the mutations are generally consistent with biochemical studies of brain and recombinant receptors, indicating that THIP and flurazepam affect most GABA_A receptors and thus show similar effects of deletion of either the ₁ or ₂ subunit. For etomidate, subtypes of the -subunit influence the potency with which etomidate potentiates GABA-evoked currents and the  isoform is a crucial determinant of the GABA-mimetic activity of this compound (Hill-Venning et al., 1997). Concerning the effects of etomidate on recombinant GABA_A receptors composed of ₁, _2S, and either ₁, ₂, or ₃ subunits, Sanna et al. (1997) showed that the order of potency after application of this drug was ₃ > ₂ > ₁. In accordance with these findings, the loss of righting reflex produced by etomidate was reduced only in the ₂ mutants not in the ₁ knockouts.

Zolpidem has a higher affinity for GABA_A receptors containing ₁ subunits than for other receptors (Pritchett et al., 1989; Sieghart, 1995), and sleep time produced by this drug was markedly decreased in ₁ null mutant mice. Nevertheless, ₂ knockout mice also showed a significant decrease in zolpidem-induced sleep time. This is perhaps surprising in view of the findings that there was a complete loss of high-affinity binding sites for zolpidem in ₁ (/) mice, whereas high-affinity zolpidem binding sites still accounted for 49% of total [³H]flumazenil sites in ₂ (/) brains (Sur et al., 2001). It should be noted, however, that at the high doses used to produce loss of righting reflex, it is likely that zolpidem is not completely selective for ₁-containing receptors.

In view of the clear changes in action of these GABAergic drugs, it was surprising that pentobarbital sleep time was not altered in the mutant mice, but this is consistent with the robust potentiation by pentobarbital of the GABA current in Purkinje cells isolated from these mutant mice (Sur et al., 2001). In addition, deletion of another  subunit, ₃, also failed to change pentobarbital sleep time (Quinlan et al., 1998). Pentobarbital has effects on a number of other ion channels (Wartenberg et al., 2001; Yamakura et al., 2001; Bachmann et al., 2002), and it is possible that these targets are more important for pentobarbital-induced loss of righting reflex than the GABA_A receptor.

Reduction of alcohol sleep time by deletion of either the ₁ or ₂ subunit is consistent with the idea that ethanol enhances GABA_A receptor function and thereby produces some of its behavioral effects (Harris, 1999). The effect of the null mutations was only seen in male mice, however, suggesting that females may employ additional mechanisms for production of ethanol effects or that the GABA_A receptor action of ethanol is less important in female mice. There is evidence for gender differences in alcohol sleep time and genetic influences upon that behavior (DeFries et al., 1989). In addition, neurosteroids derived from progesterone have been implicated in alcohol actions, including loss of righting reflex (Morrow et al., 2001). Thus, actions of ethanol such as loss of righting reflex likely involve multiple targets, including the GABA_A receptor.

Results from GABA receptor function measured in cortical membranes by chloride flux are generally consistent with previous biochemical and current behavioral studies of these mice. In particular, both knockout strains showed decreased maximum stimulation of muscimol-stimulated ³⁶Cl uptake. Sur at al. (2001) showed that both the ₁ (/) and ₂ (/) mice demonstrated a large (50-70%) widespread loss of GABA_A receptors measured by [³⁵S]t-butylbicyclophosphorothionate and [³H]muscimol binding and of benzodiazepine binding sites measured by [³H]flumazenil binding. Nevertheless, the 25% decrease in chloride flux (E_max) that we observed suggests that mutant mice compensate for loss of receptor subunits by either more efficient trafficking of receptors to the membrane surface or greater function of remaining receptors. In addition to the decrease in E_max, the mutant mice also showed an increased muscimol EC₅₀, likely a consequence of subunit substitution in the remaining receptors. The potency of muscimol varies with the type of  subunit incorporated into the GABA_A receptor complex in the order ₆ > ₅ = ₂ > ₁ > ₃ (Ebert et al., 1997) and substitution of ₃ for ₁ as a possible explanation for the decreased muscimol EC₅₀ in mutant mice. At first glance, the ability to potentiate the ³⁶Cl uptake by several different sedatives does not appear consistent with behavioral effects of these drugs. For example, potentiation of muscimol action by flunitrazepam was not changed in ₂ knockout mice and was increased in ₁ null mutant mice, despite decreased sleep times in both mutants. It should be noted, however, that we are measuring flunitrazepam effects on the remaining GABA_A receptors; thus, these potentiation values do not reflect the loss of GABA receptors. It appears that the reason for the decrease in behavioral actions of flurazepam is the decrease in number of GABA_A receptors, not a change in the properties of the remaining receptors. On the other hand, the increase in flunitrazepam potentiation of muscimol action in ₁ null mutant mice could reflect overexpression of ₂ and ₃ subunits in these mice (Sur et al., 2001). It is known that the benzodiazepine potentiation of GABA action is greater at ₃₃₂ subunits than at ₁ or ₅ containing receptors (Smith at al., 2001). In the case of ethanol, however, there is evidence for decreased sensitivity of the remaining receptors, at least in the ₂ null mutants.

The majority of gene-targeted mice are developed on a mixed genetic background of inbred mouse strains C57BL/6J and strain 129, and it is useful to compare the responses of wild-type and null allele mice to the pattern of responses found in the parental mouse strains used. In the present study, genes linked to the ₁ (/) and ₂ (/) mutations would be derived from strain 129/SvEv, whereas genes linked to the wild-type ₁ and ₂ genes would stem from C57BL/6J. Two recent articles (Simpson et al., 1997; Threadgill et al., 1997) have shown that there is substantial genetic variation among substrains of 129 inbred strain. There are no available data about effects of sedative drugs in the 129/SvEv substrain. Homanics et al. (1999) found that 129/SvJ and C57BL/6J inbred strains have similar hypnotic responses to ethanol and pentobarbital. 129/SvJ mice, however, are markedly less sensitive to the hypnotic effects of midazolam, zolpidem, and propofol (sleep times 30, 65, and 52% of the C57BL/6J values, respectively), whereas 129/SvJ are significantly less resistant to etomidate (30% greater sleep time than C57BL/6J). Because our data show a decrease of ethanol-induced sleep in ₁ (/) and ₂ (/) males and of etomidate-induced sleep only in ₂ (/) mice, these effects are most likely the result of mutations and not of flanking genes from the 129/SvEv strain.

Another concern with studies of null mutant mice is compensation (Crawley, 1996; Gerlai, 2001). It is likely that some compensatory changes occur in the brain to allow near normal brain function without ₁ or ₂ subunits. Our data showing only a 25% decrease in the maximal chloride flux through GABA_a receptors in the face of a larger loss of receptor density suggests one type of compensation. Interestingly, very similar results were recently obtained from independently generated ₁ knockout mice (Kralic et al., 2002).

In conclusion, these data show that removal of either ₁ or ₂ subunits of GABA_A receptors produce strong and specific decreases effects of different drugs. Overall, these data confirm the crucial role of the GABA_A receptor in mechanisms mediating sedative/hypnotic effects.