The modulatory effects of the anxiolytic etifoxine on GABA_A receptors are mediated by the β subunit

Abstract

The anxiolytic compound etifoxine (2-ethylamino-6-chloro-4-methyl-4-phenyl-4H-3,1-benzoxazine hydrochloride) potentiates GABA_A receptor function in cultured neurons (Neuropharmacology 39 (2000) 1523). However, the molecular mechanisms underlying these effects are not known. In this study, we have determined the influence of GABA_A receptor subunit composition on the effects of etifoxine, using recombinant murine GABA_A receptors expressed in Xenopus oocytes. Basal chloride currents mediated by homomeric β receptors were reduced by micromolar concentrations of etifoxine, showing that β subunits possess a binding site for this modulator. In oocytes expressing α₁β_x GABA_A receptors (x=1, 2 or 3), etifoxine evoked a chloride current in the absence of GABA and enhanced GABA (EC10)-activated currents, in a dose-dependent manner. Potentiating effects were also observed with α₂β_x, β_xγ_2s or α₁β_xγ_2s combinations. The extent of potentiation was clearly β-subunit-dependent, being more pronounced at receptors containing a β₂ or a β₃ subunit than at receptors incorporating a β₁ subunit. The mutation of Asn 289 in the channel domain of β₂ to a serine (the homologous residue in β₁) did not significantly depress the effects of etifoxine at α₁β₂ receptors. This specific pattern of inhibition/potentiation was compared with that of other known modulators of GABA_A receptor function like benzodiazepines, neurosteroids, barbiturates or loreclezole.

Introduction

In the mammalian brain, the inhibitory activity of type A gamma-aminobutyric acid receptors (GABA_ARs) can be potently enhanced by allosteric modulators such as benzodiazepines, neurosteroids or barbiturates. This property underlies the pharmacological treatment of diverse neuropsychiatric disorders like anxiety, insomnia or muscular spasms. Unfortunately, classical benzodiazepines display major side effects including anterograde amnesia, impairment of locomotor activity, ethanol-potentiation, tolerance and rebound effects on cessation of treatment (Woods and Winger, 1995). In contrast, neurosteroids exhibit less detrimental side effects but are not yet exploited on a large scale for their therapeutic potential (Gasior et al., 1999). Etifoxine (2-ethylamino-6-chloro-4-methyl-4-phenyl-4H-3,1-benzoxazine hydrochloride, Stresam®), a molecule structurally unrelated to benzodiazepines and neurosteroids, has also revealed anxiolytic properties in rodents (Boissier et al., 1972, Schlichter et al., 2000) and humans (Servant et al., 1998), without sedative, myorelaxant and mnesic side effects at anxiolytic concentrations (Micallef et al., 2001).

Recent binding and electrophysiological experiments have demonstrated that etifoxine binds to GABA_ARs via an allosteric site which differs from that of benzodiazepines and neurosteroids (Verleye et al., 1999, Schlichter et al., 2000, Verleye et al., 2001, Verleye et al., 2002). In addition, etifoxine increases GABAergic neurotransmission by an indirect mechanism involving the activation of peripheral (mitochondrial)-type benzodiazepine receptors and, very likely, an enhancement of neurosteroid synthesis (Schlichter et al., 2000).

Mammalian GABA_ARs are hetero-pentamers assembled from distinct subunit families with multiple subtypes (α_1–6, β_1–3, γ_1–3, δ, ϵ, π and θ). The most abundant native receptors, however, are composed of only α, β and γ subunits with presumably a 2α:2β:1γ stoichiometry (Sieghart et al., 1999, Mehta and Ticku, 1999, Knight et al., 2000). The subunit composition of GABA_ARs has important functional implications for the effects of positive modulators like benzodiazepines and neurosteroids. Both α and γ subunits are required for the high-affinity binding of benzodiazepines, thus suggesting that the binding site is located at the α–γ interface (Sigel and Buhr, 1997). By contrast, the modulation of GABA responses by neurosteroids does not require the presence of α or γ subunits but depends on the presence of a β subunit (Mehta and Ticku, 1999). The location of the binding site of etifoxine and its mechanism of action are unknown at present. In order to address these questions, as a first step, we used a recombinant strategy to determine what types of subunits underlie its modulatory effects. For this purpose, we analysed the ability of etifoxine to modulate GABA-evoked chloride currents in Xenopus oocytes expressing one (β_x), two (α₁β_x, β_xγ_2s) or three (α₁β_xγ_2s) murine GABA_AR subunits. The α₁ and γ_2s subunits were selected for co-expression with β subunits because they are the most abundant representatives of their subunit family in many regions of the rodent brain and, therefore, are likely to co-assemble with a β subunit to produce native GABA_AR subtypes (Pirker et al., 2000). In complementary experiments, we also examined the effects of replacement of α₁ by α₂ and the effects of mutating β₂ N-289, a residue that is known to control the interaction of several other positive modulators with GABA_A receptors (Wafford et al., 1994, Belelli et al., 1997, Halliwell et al., 1999).