Role of GABA_A α5-containing receptors in ethanol reward: The effects of targeted gene deletion, and a selective inverse agonist

Abstract

GABA_A receptors containing α5 subunits have been suggested to mediate the rewarding effects of ethanol. We tested this hypothesis in mice with deletion of α5 subunits. α5 knockout mice did not differ from wildtypes in operant responding for 10% ethanol/10% sucrose, but responded less for 10% sucrose. The benzodiazepine (BZ) site inverse agonist, Ro 15-4513, has higher affinity for GABA_A receptors containing 5 subunits and dose-dependently (0–27 mg/kg, i.p.) reduced lever pressing for ethanol/sucrose in wildtype mice, but had less effect in knockout mice; lever pressing for sucrose was unaffected. These data suggest that α5 subunits are not essential for ethanol reward, but the reduction of operant responding for ethanol by Ro 15-4513 is mediated by α5-containing GABA_A receptors. In measures of ethanol consumption, α5 knockout mice did not differ from wildtypes at low ethanol concentrations (2–8%), but consumed less ethanol at higher concentrations; these differences were not attributable to increased behavioural disruption of the knockout by ethanol, since no differences were seen in sensitivity to ethanol's sedative or ataxic effects. Ro 15-4513's ability to reduce ethanol consumption was unaffected, suggesting that this effect is not mediated by the α5 subtype. Secondly, we tested the ability of a novel α5-efficacy-selective benzodiazepine receptor ligand, α5IA-II, that possesses greater inverse agonist activity at α5- than at α1-, á2- or α3-containing GABA_A receptors, to influence operant responding. α5IA-II (0.03–3 mg/kg) dose-dependently decreased lever pressing for 10% ethanol, the minimally effective dose of 1 mg/kg, corresponding to over 90% receptor occupancy, but did not affect lever pressing for 4% sucrose. Although inverse agonists acting at α5-containing receptors reduce ethanol self-administration, α5 subunits may not be essential to signaling ethanol reward.

Introduction

Ethanol achieves its effects on the central nervous system through diverse mechanisms, including interaction with several ligand-gated ion channels. A major contribution is played by GABA_A receptors, at which ethanol acts to facilitate chloride flux, resulting in enhanced neuronal inhibition. Compounds acting as inverse agonists at the benzodiazepine (BZ) binding site of GABA_A receptors (i.e. they reduce GABA-stimulated chloride flux, thereby reducing the membrane potential) may antagonize behaviorally sedative effects of ethanol (Bonetti et al., 1988, Dar, 1992, Suzdak et al., 1986, Wood et al., 1989), and it has also been suggested that such compounds may also reduce ethanol's euphoric and rewarding properties (June et al., 1994).

GABA_A receptors are a family of oligomeric proteins that show considerable heterogeneity in their regional distribution and function. It has been suggested that receptors containing the α5 subunit isoform may be particularly important in mediating the rewarding properties of ethanol. This conclusion is predicated on three lines of pharmacological evidence. First, the imidazobenzodiazepine, RY 023 which has 67-fold higher affinity for and is an inverse agonist at recombinant GABA_A receptors made up of β2, γ2 plus α5 subunits compared to the analogous receptors containing either an α1, α2 or α3 subunit (Liu et al., 1996, Liu et al., 1995), suppressed operant responding for ethanol reinforcers in the alcohol-preferring rat, when administered into dorsal hippocampal sites (June et al., 2001) rich in α5-containing receptors (Fritschy and Mohler, 1995, Pirker et al., 2000, Wisden et al., 1992). Higher doses of RY023 were required to reduce responding for saccharin, consistent with this latter effect being mediated by other GABA_A receptor subtypes (June et al., 2001). Second, RY 024, which has an α5 binding selectivity (76-fold) comparable to RY 023 but has greater α5 inverse agonism (Liu et al., 1996, Liu et al., 1995), decreases ethanol self-administration in rats when given systemically(McKay et al., 2004). In keeping with a role for α5-containing receptors in ethanol reward, chronic alcohol administration increases expression of α5 mRNA in hippocampus (Charlton et al., 1997, Devaud et al., 1995). Thirdly, the prototypic imidazobenzodiazepine inverse agonist, Ro 15-4513, which has 10–20-fold higher affinity for α5-compared to α1-, α2-, α3-, α4- or α6-containing receptors (Hadingham et al., 1993, Lingford-Hughes et al., 2002, Liu et al., 1996, Smith et al., 2001), has long been known to antagonize the rewarding effects of ethanol in operant self-administration tests (Bao et al., 1992, Glowa et al., 1988, Rassnick et al., 1993, Samson et al., 1989, Samson et al., 1987). The ability of Ro 15-4513 to disrupt responding for ethanol might thus be related to its action at α5-containing receptors.

Nevertheless, the conclusions that can be reached with the hitherto available pharmacological tools are limited. Liu et al. (1996) suggest that both RY 023 and RY 024 can exert actions, presumably inverse agonism, at other GABA_A subtypes, although it should be noted that to date only efficacy data for the α5 subtype has been presented (Liu et al., 1995). Similarly, the fact that Ro 15-4513 is proconvulsant (Lister and Nutt, 1988) and possesses inverse agonism at the α1, α2 and α3 subtypes (Benson et al., 1998, Hadingham et al., 1995, Kelly et al., 2002, Wafford et al., 1993) raises the possibility that the ability of Ro 15-4513 to disrupt ethanol reinforcement may represent actions at other receptor subtypes. An alternative approach to investigating the role of α5-containing subunits in ethanol's effects is to exploit the availability of mice in which the gene encoding subunits have been deleted (Collinson et al., 2002). If α5-containing receptors play an important role in mediating either the reinforcing effects of ethanol, or the ability of drugs like Ro 15-4513 to interfere with such reinforcing effects, then α5 knockout mice should show lower rates of responding for ethanol reward, and should be less sensitive than corresponding wildtype control mice to the disruptive effects of Ro 15-4513. To ensure that the α5 knockout mice were not generally different from wiltypes in the effects of ethanol, the pharmacological effects of ethanol on motor performance and operant performance were compared between the knockout and wildtype mice.

In addition, and as a complement to the knockout mice studies, we have used a novel compound, α5IA-II, which preferentially modulates the functions of α5-containing GABA_A receptors (21). This differs from the α5 binding-selective compounds RY 023, RY 024 and Ro 15-4513 in that α5IA-II possesses subtype selective efficacy. In other words, whilst α5IA-II binds with comparable affinity to the α1, α2, α3 and α5 GABA_A subtypes, it has much greater inverse agonist efficacy at the α5 compared to α1, α2 or α3 subtypes and consequently it's in vivo effects will be exerted primarily via α5-containing GABA_A receptors (see Table 1 and Fig. 1 for summary).