Hormonally regulated α₄β₂δ GABA_A receptors are a target for alcohol

Here we report that low concentrations of alcohol (1–3 mM) increased Cl⁻ currents gated by a recombinant GABA_A receptor, α₄β₂δ, by 40–50% in Xenopus laevis oocytes. We also found greater hippocampal expression of receptors containing α₄ and δ subunits, using a rat model¹ of premenstrual² syndrome (PMS) in which 1–3 mM alcohol preferentially enhanced GABA-gated currents, and low doses of alcohol attenuated anxiety and behavioral reactivity. The alcohol sensitivity of δ-containing receptors may underlie the reinforcing effects of alcohol during PMS, when eye saccade responses to low doses of alcohol are increased².

Alcohol is an addictive recreational drug that reduces anxiety at low doses and causes sedation at high doses³. These effects are similar to those of drugs that enhance the function of GABA_A receptors, which gate the Cl⁻ currents that mediate most inhibitory neurotransmission in the brain. Acutely high doses of alcohol potentiate GABA-gated currents³ at both native³ and recombinant GABA_A receptors⁴, and chronically alter GABA_A receptor expression⁵. Low doses of alcohol have not been shown to directly modulate recombinant GABA_A receptors³, although there is indirect evidence for such effects at native receptors^6,7.

It has been suggested that discrepancies between the alcohol sensitivity of native and recombinant receptors may be due to their subunit composition¹. We therefore investigated the effects of low concentrations of alcohol on different subtypes of recombinant GABA_A receptors expressed in Xenopus oocytes. Low (1 mM) concentrations of alcohol selectively increased GABA-gated currents at α₄β₂δ receptors by 45 ± 11% (P < 0.05, Fig. 1a). Substitution of γ₂ for the δ subunit abolished this GABA-modulatory effect of alcohol (Fig. 1a), as did substitution of α₁ for α₄ when co-expressed with β₂δ (Fig. 1b). Furthermore, none of the other subunit combinations that yielded functional receptors (Supplementary Methods online) were sensitive to low concentrations of alcohol (Fig. 1a and b). Thus, co-expression of α₄ and δ subunits is necessary for potentiation by 1 mM alcohol.

Fig. 1.

Low concentrations of alcohol potentiate GABA responses at α₄β₂δ receptors. (a) The effects of alcohol on responses to GABA (EC₂₀ = 0.03 μM α₄β₂δ, 1.7 μM α₁β₂γ_2s, 5.6 μM α₄β₂γ_2s) were measured by two-electrode voltage-clamp recording at −70 mV in oocytes expressing recombinant GABA_A receptors. Inset, representative traces showing currents activated by GABA in the absence and presence of 1 mM alcohol. (b) Effects of 1 mM alcohol on GABA(EC₂₀)-gated currents at various GABA_A receptor subtypes (n = 8–20, *P < 0.05). Experiments were conducted according to Institutional Animal Care and Use Committee guidelines.

The α₄βδ GABA_A receptors are expressed at very low levels in most regions of the brain⁸. Predicting that physiological states that are associated with increased sensitivity to alcohol (such as PMS) may involve increased expression of α₄βδ, we used a rodent model of PMS¹ to test this hypothesis. With chronic in vivo administration and withdrawal of progesterone, we replicated the hormonal and behavioral facets of PMS². Expression of δ and α₄ subunit proteins in the hippocampus was three-fold higher after progesterone withdrawal (Fig. 2a and b, and Supplementary Fig. 1), as was δ-subunit mRNA (Supplementary Fig. 1). Co-assembly of these subunits was determined using co-immunoprecipitation (Fig. 2c) and verified by an increased efficacy of THIP, a GABA partial agonist, relative to GABA after progesterone withdrawal (Fig. 2d). This is characteristic of α₄βδ receptor expression⁹. Co-assembly of β₂ with α₄δ has not been tested directly but is suggested by reports of high levels of β₂ in areas rich in α₄ and δ subunits.

Fig. 2.

Progesterone withdrawal increases α₄βδ GABA_A receptors. (a) Left, western blot showing increased expression of the δ subunit (54 kDa), but not a control protein (GAPDH, 36 kDa) after progesterone withdrawal (P Wd) compared to control (Con). Right, mean values (n = 20–25, performed in triplicate). (b) Increases in α₄ protein following P Wd were prevented by in vivo administration of alcohol (0.5 g/kg × 3, intraperitoneally) during the final two hours of the withdrawal period (P Wd + Alc; n = 9–10). (c) Co-assembly of α₄ and δ GABA_A receptor subunits. After immunoprecipitation (IP) using protein A beads coupled to antibodies for the δ subunit (IP, δ) or a cytosolic protein (IP, Neg. con) membranes were probed with digoxygenin-labeled anti-α₄ on a western blot (n = 4 hippocampi, in duplicate). A prominent 67-kDa band was detected after P Wd, but was barely detectable under control conditions. (d) The maximum current produced by THIP compared to that produced by 10 mM GABA was 1.41 after P Wd (THIP EC₅₀ = 39 ± 2 μM), and 0.95 in control neurons (THIP EC₅₀ = 81 ± 6 μM). This was determined using whole-cell patch clamp recording in neurons from CA1 hippocampus (n = 10–15, *P < 0.05).

Concomitant with upregulation of α₄βδ receptors in the progesterone-withdrawal model, 1 mM alcohol produced a 81 ± 21% potentiation of GABA-gated currents recorded from hippocampal neurons in vitro (Fig. 3a). Higher concentrations (10 mM and greater) of alcohol were not as effective in potentiating GABA-gated currents. After progesterone withdrawal, low doses of alcohol (0.2–0.4 g/kg) administered in vivo also decreased the acoustic startle response, a measure of behavioral excitability, suggesting a greater anxiolytic effect of alcohol at this time (Fig. 3b).

Fig. 3.

Low doses of alcohol potentiate GABA-gated currents in hippocampal pyramidal neurons and decrease behavioral excitability after progesterone withdrawal. (a) GABA (10 μM)-gated currents recorded in the presence or absence of alcohol (0.1–10 mM), measured by whole-cell patch clamp recording after progesterone withdrawal (P Wd) or under control conditions (con). Inset, representative traces. Suppression of α₄ subunit expression (P Wd + α₄ suppressed), as described in Fig. 2b, abolished the stimulatory effects of low concentrations of alcohol after P Wd (n = 20–25 cells/concentration, 8–10 rats/group). (b) Low doses of alcohol administered to P Wd (but not control) rats significantly lowered the acoustic startle response, expressed relative to the saline control (n = 5–11, *P < 0.05, **P < 0.005).

To verify that native α₄βδ receptors were sensitive to low doses of alcohol, we repeated the progesterone withdrawal experiments after suppressing α₄ subunit expression. We have previously shown that, following progesterone withdrawal, injection of a GABA modulator prevents expression of the α₄ subunit¹. Thus, we used alcohol here to suppress expression of the α₄ protein following progesterone withdrawal (P Wd + Alc; Fig. 2b and Supplementary Fig. 2). Under these conditions, the GABA-modulatory effects of 1 mM alcohol were prevented in hippocampal neurons (Fig. 3a; P Wd + α₄ suppressed), again implicating α₄βδ receptors as sensitive to low doses of alcohol. There is also corroborating evidence that the reinforcing and anti-convulsant properties of alcohol are reduced in transgenic mice lacking the δ subunit¹⁰.

It is thought that α₄βδ receptors are expressed at extrasynaptic sites⁸ where they may dampen neuronal excitability, primarily by acting as a resistive shunt¹¹. Enhanced function of these receptors by low concentrations of alcohol in women with PMS would further decrease neuronal excitability, leading to behavioral stress-reduction. Blood alcohol levels of 1–3 mM may result from consumption of a half glass of wine¹² or less¹³. The reported increase in alcohol consumption and propensity for alcoholism in women with PMS¹⁴ may thus be accounted for by these enhanced reinforcing properties of alcohol. More broadly, it is conceivable that alterations in GABA_A receptors, perhaps including α₄β₂δ, are involved in the genetic predisposition for alcoholism in which there is an increased sensitivity to low doses of alcohol¹⁵.