Report

Molecular and Neuronal Substrate for the Selective Attenuation of Anxiety

See allHide authors and affiliations

Science  06 Oct 2000:
Vol. 290, Issue 5489, pp. 131-134
DOI: 10.1126/science.290.5489.131

This article has a correction. Please see:

Abstract

Benzodiazepine tranquilizers are used in the treatment of anxiety disorders. To identify the molecular and neuronal target mediating the anxiolytic action of benzodiazepines, we generated and analyzed two mouse lines in which the α2 or α3 GABAA(γ-aminobutyric acid type A) receptors, respectively, were rendered insensitive to diazepam by a knock-in point mutation. The anxiolytic action of diazepam was absent in mice with the α2(H101R) point mutation but present in mice with the α3(H126R) point mutation. These findings indicate that the anxiolytic effect of benzodiazepine drugs is mediated by α2 GABAA receptors, which are largely expressed in the limbic system, but not by α3 GABAAreceptors, which predominate in the reticular activating system.

Excessive or inappropriate anxiety can be controlled by enhancing inhibitory synaptic neurotransmission mediated by GABA (GABAergic inhibitory neurotransmission) using clinically effective benzodiazepine drugs (1). However, to date it has not been possible to identify the one or more GABAA receptor subtypes that mediate the attenuation of anxiety. Four types of diazepam-sensitive GABAA receptors can be distinguished on the basis of the presence of α1, α2, α3, or α5 subunits. These receptors can be rendered insensitive to diazepam in vitro by replacing a conserved histidine residue by arginine in the drug binding site (2,3). Introduction of the respective point mutation into mouse lines enables the pharmacological profile of benzodiazepine drugs to be attributed to defined receptor subtypes. Using this approach, we have attributed the sedative and amnesic properties of diazepam, but not its anxiolytic action, to α1 GABAA receptors (4,5). We hypothesized that the anxiolytic action of benzodiazepine drugs might be mediated by α2 or α3 GABAA receptors on the basis of their distinct neuroanatomical expression patterns. Whereas α2 GABAA receptors are preponderant in areas of the limbic system as well as in the cerebral cortex and striatum, α3 GABAA receptors are selectively expressed in noradrenergic and serotonergic neurons of the brainstem reticular formation, the basal forebrain cholinergic neurons, and GABAergic neurons in the reticular nucleus of the thalamus (6).

To distinguish the pharmacology of the α2 receptor subtype from that of the α3 subtype, we generated two mouse lines in which a point mutation was introduced into homologous positions of the α2 subunit gene (Fig. 1) and the α3 subunit gene (7), rendering the respective receptors insensitive to diazepam. [The point mutations were His101→ Arg and His126 → Arg, respectively, hence the mouse lines are denoted α2(H101R) and α3(H126R).] The mutants did not display an overt distinctive phenotype, bred normally, and expressed all subunits tested (α1, α2, α3, β2/3, and γ2) at normal levels (Fig. 2A) and with unaltered distribution. Ro15-4513 is an inverse agonist at the benzodiazepine binding site and binds to both diazepam-sensitive and diazepam-insensitive GABAA receptors. The proportion of diazepam-insensitive [3H]Ro15-4513 binding sites was increased from 5% in wild-type mice to 17% in α2(H101R) mice and to 11% in α3(H126R) mice [α2 wild-type controls: maximum number of binding sites B max = 0.08 ± 0.02 pmol/mg protein, dissociation constant K d = 4.3 ± 0.8 nM (n = 3); α2(H101R) mice:B max = 0.26 ± 0.01 pmol/mg protein,K d = 8.0 ± 1.3 nM (n= 3); α3 wild-type controls: B max = 0.06 ± 0.01 pmol/mg protein, K d = 4.9 ± 2.1 nM (n = 3); and α3(H126R) mice:B max = 0.11 ± 0.02 pmol/mg protein,K d = 6.5 ± 1.6 nM (n= 3)]. In line with the known distribution of the α2 subunit (6), novel diazepam-insensitive sites in α2(H101R) mice were visualized in all regions expressing α2 GABAAreceptors, as shown autoradiographically in parasagittal brain sections using 10 nM [3H]Ro15-4513 in the presence of 100 μM diazepam (Fig. 2B). Similarly, in α3(H126R) mice, the novel diazepam-insensitive [3H]Ro15-4513 binding sites displayed a distribution corresponding to that of the α3 subunit (Fig. 2B) (6). After immunoprecipitation with α2 or α3 subunit–specific antisera, [3H]Ro15-4513 binding revealed a decrease of more than three orders of magnitude in the affinity of diazepam to the α2 and α3 subunits from the respective mutant mice.

Figure 1

Targeting of the α2 subunit GABAA receptor gene. (A) Structure of wild-type and mutant alleles. Mutant allele 1 is obtained after gene targeting in mouse embryonic stem cells and introduced into the mouse germ line; breeding of these mice to EIIa-cre mice (16) results in excision of the neomycin resistance cassette (mutant allele 2). The 5′ and 3′ probes flanking the targeting vector are drawn as solid bars. His and Arg denote codons for histidine and arginine, respectively, at position 101 in exon 4. Polymerase chain reaction (PCR) primers P1, P2, and P3 are indicated. (B) Southern blot analysis of wild-type (wt) allele and mutant allele 1 (Mut1) in embryonic stem cells. (C) Genotyping offspring from a cross of a chimera and a mouse hemizygous for the EIIa-cre transgene. Top panel: Primers P1, P2, and P3 provide specific amplification products for each allele. Bottom panel: PCR primers UR26 and UR36 amplify the cretransgene. (D) Verification of the α2(H101R) point mutation by automated DNA sequencing.

Figure 2

Molecular characteristics of GABAA receptors in α2(H101R) and α3(H126R) mice. (A) Western blots of whole brain membranes from wild-type, α2(H101R), and α3(H126R) mice using antisera recognizing the α1, α2, α3, β2/3, and γ2 subunits. (B) Receptor autoradiography of diazepam-insensitive sites in wild-type, α2(H101R), and α3(H126R) brains. Parasagittal sections were incubated with 10 nM [3H]Ro15-4513 in the presence of 100 μM diazepam to reveal the diazepam-insensitive [3H]Ro15-4513 binding sites. In wild-type mice, diazepam-insensitive [3H]Ro15-4513 binding is due to α4 and α6 GABAA receptors. (C) GABA responses in cultured hippocampal pyramidal cells from α2(H101R) mice. The holding potential in the patch-clamp analysis was –60 mV; the chloride concentration was symmetrical. GABA was applied for 5 s. Hippocampal neurons from embryonic day 16.5 embryos were cultured for 10 to 14 days. **, P < 0.001 (Student'st test).

In cultured hippocampal pyramidal cells (8), the electrophysiological response to GABA (3 μM) was indistinguishable between cells from wild-type and α2(H101R) mice (Fig. 2C). However, the potentiation by diazepam (1 μM) was reduced in cells from α2(H101R) mice relative to cells from wild-type mice [17.6 ± 4.5% (n = 29) versus 48.1 ± 7.9% (n = 18), P = 0.001] (Fig. 2C); the remaining potentiation presumably can be attributed to GABAA receptors other than α2. The inverse agonistic action of Ro15-4513 (1 μM) in wild-type cells was converted into an agonistic response in cells derived from α2(H101R) mice [–39 ± 5.2% (n = 13) versus 11.7 ± 7.5% (n = 23), P = 0.003] (Fig. 2C). This is consistent with the switch in efficacy of Ro15-4513 from inverse agonism to agonism, as shown for recombinant α2(H101R)β3γ2 receptors expressed in HEK-293 cells (3).

The pharmacological importance of the mutated α2 and α3 GABAA receptors was assessed by comparing the diazepam-induced behavior of α2(H101R) and α3(H126R) mice with that of wild-type mice (9). When tested in a dose-dependent manner, the sedative, motor-impairing, and anticonvulsant actions of diazepam (10) were not impaired in either α2(H101R) mice or α3(H126R) mice relative to wild-type mice (Fig. 3).

Figure 3

Behavioral assessment of the sedative, motor-impairing, and anticonvulsant properties of diazepam in α2(H101R) and α3(H126R) mice relative to wild-type mice. (A) Dose-dependent inhibition of locomotor activity in wild-type and α2(H101R) mice [F(3,71) = 19.31,P < 0.001, n = 9 or 10 mice per group]. (B) Dose-dependent decrease in the latency to fall off the rotating rod (fixed 2 rpm) in wild-type and α2(H101R) mice [F(3,153) = 71.01, P < 0.001,n = 20 or 21 mice per group]. (C) Dose-dependent decrease of the percentage of mice developing tonic convulsions in wild-type (χ2 = 32.38, P< 0.001, n = 10 mice per group) and α2(H101R) mice (χ2 = 28.13, P < 0.001,n = 10 mice per group). (D) Dose-dependent inhibition of locomotor activity in wild-type and α3(H126R) mice [F(3,64) = 14.70, P < 0.001,n = 8 to 10 mice per group]. (E) Dose-dependent decrease in the latency to fall off the rotating rod (fixed 2 rpm) in wild-type and α3(H126R) mice [F(3,64) = 36.78, P < 0.001,n = 8 to 10 mice per group]. (F) Dose-dependent decrease of the percentage of mice developing tonic convulsions in wild-type (χ2 = 32.38, P< 0.001, n = 10 mice per group) and α3(H126R) mice (χ2 = 28.60, P < 0.001,n = 10 mice per group). Results are given as means ± SEM. +, P < 0.05; ++, P < 0.01; +++, P< 0.001 (Dunnett's post hoc comparisons or Fisher's exact tests). V, vehicle. The rotarod and pentylenetetrazole convulsion tests were performed according to Bonetti et al. (10). Locomotor activity was automatically recorded for 30 min. Mice were treated with either vehicle or diazepam (3, 10, and 30 mg/kg orally) 30 min before testing.

The anxiolytic-like action of diazepam in α2(H101R) and α3(H126R) mice was investigated in the light-dark choice test (11) and the elevated plus-maze test (12). In the light-dark choice test, the α2(H101R) mice did not show the behavioral disinhibition by diazepam that was apparent in wild-type mice. Diazepam up to 2 mg/kg body weight did not increase the time spent in the lit area in α2(H101R) mice relative to wild-type mice (P < 0.05 versus vehicle) (Fig. 4A). This effect was not due to a motor deficit in α2(H101R) mice, because no behavioral differences in the dark area were observed between wild-type and α2(H101R) mice under either vehicle or diazepam treatment. Furthermore, α2(H101R) mice retained the ability to display an anxiolytic-like response to ligands acting at GABAA receptor sites other than the benzodiazepine site. Sodium phenobarbital (15 mg/kg subcutaneously) induced a behavioral disinhibition in the light/dark choice test in α2(H101R) mice similar to that seen in wild-type mice [time in lit area, wild-type: vehicle, 70.5 ± 10.5 s, phenobarbital, 111.75 ± 10.0 s; α2(H101R): vehicle, 75.3 ± 11.5 s, phenobarbital, 110.9 ± 9.6 s;F(1,24) = 13.44, P < 0.01,n = 6 to 8].

Figure 4

Behavioral assessment of anxiolytic-like action of diazepam in α2(H101R) and α3(H126R) mice relative to wild-type mice. (A) Light/dark choice test. Diazepam dose-dependently increased the time spent in the lit area in wild-type mice [F(3,36) = 3.14, P < 0.05] but not in α2(H101R) mice [F(3,36) = 0.32, not significant] (n = 10 mice per group). (B andC) Elevated plus-maze. Diazepam (2 mg/kg) increased the percentage of time spent on the open arms and the number of entries on the open arms in wild-type mice (P < 0.01 and P< 0.05 versus vehicle) but not in α2(H101R) mice [F(1,32) = 4.31 and F(1,32) = 4.76,P < 0.05, respectively] (n = 8 to 10 mice per group). (D) Light/dark choice test. Both wild-type and α3(H126R) mice displayed a dose-dependent increase in the time spent in the lit area [F(1,70) = 14.74, P < 0.001, n = 9 or 10 mice per group]. (E andF) Elevated plus-maze. Diazepam (2 mg/kg) increased the percentage of time spent on the open arms and the number of entries on the open arms to the same extent in wild-type and α3(H126R) mice [F(1,36) = 26.52 and F(1,36) = 37.31,P < 0.001, respectively] (n = 10 mice per group). Results are given as means ± SEM. +, P < 0.05; ++, P < 0.01; +++, P < 0.001 (Dunnett's or Fisher's pairwise post hoc comparisons or Fisher's exact tests). V, vehicle; Dz, diazepam. The light-dark choice test was carried out as described (11) with an illumination of 500 lux. Mice were given vehicle or increasing doses of diazepam (0.5, 1, and 2 mg/kg orally). The elevated plus-maze was performed according to Lister (12) under an indirect dim-light illumination (< 10 lux). Vehicle or diazepam were administered 30 min before testing.

The absence of an anxiolytic-like effect of diazepam in α2(H101R) mice was confirmed in the elevated plus-maze test. In wild-type mice, diazepam facilitated the exploratory behavior by increasing both the amount of time spent (P < 0.01 versus vehicle) and the number of entries in the open arms (P < 0.05). In contrast, in α2(H101R) mice, diazepam failed to increase both parameters of exploratory behavior (Fig. 4, B and C). Again the failure was not due to motor impairment, because the motor activity in the enclosed arms was similar in α2(H101R) and wild-type mice irrespective of the treatment.

The potential contribution of α3 GABAA receptors to the anxiolytic-like activity of diazepam was examined in α3(H126R) mice. Both α3(H126R) and wild-type mice displayed similar dose-dependent anxiolytic-like responses to diazepam in the light/dark choice test (P < 0.01 versus vehicle) (Fig. 4D) and in the elevated plus-maze (P < 0.001 versus vehicle) (Fig. 4, E and F). These results indicate that the anxiolytic action of diazepam in wild-type mice does not involve interaction with α3 GABAAreceptors.

The anxiolytic-like action of diazepam is selectively mediated by the enhancement of GABAergic transmission in a population of neurons expressing the α2 GABAA receptors, which represent only 15% of all diazepam-sensitive GABAA receptors (13). The α2 GABAA receptor–expressing cells in the cerebral cortex and hippocampus include pyramidal cells that display very high densities of α2 GABAA receptors on the axon initial segment, presumably controlling the output of these principal neurons (14, 15). Our findings indicate that the α2 GABAA receptors are highly specific targets for the development of future selective anxiolytic drugs.

  • * These authors contributed equally to this report.

  • Present address: Department of Neurosciences, University of California, San Diego, 9500 Gilman Drive, La Jolla, CA 92093, USA.

  • To whom correspondence should be addressed. E-mail: rudolph{at}pharma.unizh.ch

REFERENCES AND NOTES

Stay Connected to Science

Navigate This Article