Naunyn-Schmiedeberg’s Arch Pharmacol (1998) 357: 401–407

© Springer-Verlag 1998

ORIGINAL ARTICLE

Giovanna Schmid · Roberta Sala Giambattista Bonanno · Maurizio Raiteri

Neurosteroids may differentially affect the function of two native GABAA receptor subtypes in the rat brain

Received: 13 October 1997 / Accepted: 30 December 1997

Abstract Hippocampal noradrenergic and cerebellar glutamatergic axon terminals are known to possess GABAA receptors mediating, respectively, enhancement of noradrenaline (NA) and glutamate release. It has been recently found that the hippocampal receptor is benzodiazepine-sensitive, whereas the cerebellar receptor is insensitive to benzodiazepine agonists. We here tested the effects of neurosteroids on these two native GABAA receptors using superfused rat hippocampal and cerebellar synaptosomes. Allopregnanolone (3α,5α-P), at nanomolar concentrations, potentiated the GABA-induced [3H]-NA release from superfused hippocampal synaptosomes; in the absence of GABA, the steroid was ineffective up to 10 µM. The enhancement by GABA of the K+-evoked [3H]-D-aspartate release from cerebellar synaptosomes also was potentiated by nanomolar 3α,5α-P; in addition, at 1–10 µM, the steroid increased [3H]-D-aspartate release in the absence of GABA. Both in hippocampus and cerebellum the potentiations of the GABA effects produced by nanomolar 3α,5α-P were abolished by dehydroepiandrosterone sulphate (DHEAS). Added up to 10 µM, DHEAS could not inhibit the effects of GABA alone. The enhancement of [3H]-D-aspartate release elicited by 3 µM 3α,5α-P in the absence of added GABA was antagonized completely by bicuculline and picrotoxin and halved by DHEAS. To conclude, 3α,5αP, at nanomolar concentrations, behaves as a positive allosteric GABA modulator at both the GABAA receptors under study. Low micromolar 3α,5α-P can directly activate the cerebellar receptor, whereas the hippocampal GABAA receptor is insensitive to the neurosteroid alone. DHEAS appears to be a pure antagonist at the neurosteroid allosteric sites. Along with the previously observed differential sensitivity to benzodiazepines, the present data strengthen the idea that the two receptors investigated represent native subtypes of the GABAA receptor having distinct pharmacology, neuronal localization and function.

G. Schmid · R. Sala · G. Bonanno · M. Raiteri (✉) Istituto di Farmacologia e Tossicologia, Università di Genova, Viale Cembrano 4, I-16148 Genova, Italia

Key words Noradrenaline release · glutamate release · hippocampus · cerebellum · allopregnanolone · dehydroepiandrosterone sulphate

Introduction The GABAA receptor is a multigene family (α1-α6, β1-β3, γ1-γ3 and δ) that is formed by co-assembly of different membrane-spanning subunits originating an integral chloride selective channel. The expression of each of these subunits differs across CNS regions and it is therefore very likely that GABAA receptors exist as multiple subtypes (Macdonald and Olsen 1994; Barnard 1996). In addition to bind GABA, which leads to direct opening of the chloride channel, the GABAA receptor is the target of several pharmacologically and clinically important drugs (and endogenous modulators) including benzodiazepines, barbiturates and neuroactive steroids which interact with distinct allosteric sites on the receptor complex. These allosteric modulators are thought to bind differentially to GABAA receptor subtypes (Biggio et al. 1992; Macdonald and Olsen 1994; Lüddens et al. 1995; Rabow et al. 1995). Several naturally-occurring steroids can affect the function and modulate the binding of GABAA receptors (Majewska et al. 1986; Harrison et al. 1987; Simmonds 1991; Hawkinson et al. 1994a, 1996; Concas et al. 1996; see, for reviews Paul and Purdy 1992; Lambert et al. 1995). In addition, electrophysiological experiments have shown that concentrations of neuroactive steroids higher than those required for enhancement of GABA effects can directly activate GABAA receptors in the absence of GABA (Cottrell et al. 1987; Peters et al. 1988; Kokate et al. 1994). Evidence for heterogeneity of neurosteroid binding sites derives from different experimental approaches including functional and ligand binding assays with native GABAA receptors in brain membranes (Morrow et al. 1990; Prince and Simmonds 1993; Hawkinson et al. 1994b; Olsen and Sapp 1995). On the other hand, no subunit specificity was observed in electrophysiological studies performed upon

402

recombinant GABAA receptors expressed in human embryonic kidney cells (Puia et al. 1990). Therefore, the issue of the heterogeneity of steroid binding sites remains unresolved. Certain endogenous steroids, such as dehydroepiandrosterone sulphate (DHEAS) and pregnanolone sulphate, have been reported to act predominantly as GABAA receptor antagonists (Gee et al. 1988; Majewska et al. 1990; Demirgören et al. 1991; Spivak 1994) although the site of binding for these antagonists has not been defined. Nerve ending preparations isolated from rat hippocampus and cerebellum respectively contain, inter alia, noradrenergic and glutamatergic nerve terminals endowed with presynaptic receptors of the GABAA type modulating noradrenaline (Fung and Fillenz 1983; Bonanno and Raiteri 1987) and glutamic acid (Gallo et al. 1981; Levi and Gallo 1981) release. Using transmitter release as a functional response, it was recently found that the hippocampal receptor is sensitive to benzodiazepine agonists, whereas the cerebellar receptor is benzodiazepine-insensitive, suggesting that the two receptors represent native subtypes of the GABAA receptor present in the adult brain (Schmid et al. 1996). In this work we compared the sensitivity of the hippocampal and cerebellar GABAA receptors to naturally-occurring neurosteroids.

Materials and methods Animals. Adult male Sprague Dawley rats (CD-COBS, Charles River, Italy) weighing 200–250 g were used. Animals were housed at constant temperature (22±1°C) and relative humidity (50%) on a regular light-dark schedule (light 7.00 a.m.–7 p.m.). Food and water were freely available. Preparation of the synaptosomal fractions. Rats were killed by decapitation, the brains were rapidly removed and the cerebella or hippocampi were dissected. Crude synaptosomal fractions (subsequently referred to as synaptosomes) were prepared as described previously (Raiteri et al. 1984) with some modifications. Briefly, the tissue was homogenized (25 strokes at 90 rpm in about 2 min for cerebellar synaptosomes and 12 strokes at 900 rpm in about 1 min for hippocampal synaptosomes) in 40 volumes of 0.32 M sucrose, buffered at pH 7.4 with phosphate. The homogenate was first centrifuged at 1000 g for 5 min; synaptosomes were then isolated from the supernatant by centrifugation at 12000 g for 20 min. All the above procedures were performed at 0–4°C. The synaptosomal pellet was finally resuspended in a physiological medium having the following composition (mM): NaCl 125, KCl 3, MgSO4 1.2, CaCl2 1.2, NaH2PO4 1.0, NaHCO3 22, glucose 10 (aeration with 95% O2 and 5% CO2 at 37°C); pH 7.2–7.4. Release experiments. Synaptosomes were incubated at 37°C for 15 min in the presence of 0.08 µM [3H]-noradrenaline ([3H]-NA; experiments with hippocampal nerve endings) or with 0.12 µM [3H]-Daspartate ([3H]-D-ASP; experiments with cerebellar nerve endings). Hippocampal synaptosomes were labeled in the presence of 0.1 µM of the serotonin uptake inhibitor 6-nitro-quipazine to avoid false labeling of serotonergic nerve terminals. At the end of the incubation, identical aliquots of the synaptosomal suspensions were layered onto microporous filters at the bottom of parallel superfusion chambers maintained at 37°C (Raiteri et al. 1974). Superfusion was started at a rate of 0.5 ml/min with standard medium aerated with 95% O2 and 5% CO2.

Fig. 1 Effect of 3α,5α-P on the basal release of [3H]-NA from hippocampal synaptosomes (upper panel) and on the K+-evoked overflow of [3H]-D-ASP from cerebellar synaptosomes (lower panel). Synaptosomes, prelabeled with the radioactive tracer, were exposed to GABA at the end of the first fraction collected ([3H]-NA release) or concomitantly to the depolarizing stimulus ([3H]-D-ASP release); 3α,5α-P was introduced 8 min before GABA. When tested alone 3α,5α-P was added at the end of the first fraction collected or concomitantly to the depolarizing stimulus. See Materials and methods for additional technical details. Data represent means ±SEM of 4–10 experiments run in quadruplicate (four superfusion chambers for each condition) on different days. * P<0.01 when compared to the effect of GABA alone; ❍ P<0.05 and ❍❍ P<0.001 when compared to the 35 mM K+-evoked overflow of [3H]-D-ASP; ● P<0.05 when compared to the effect of GABA on the depolarization evoked [3H]-DASP overflow Hippocampal synaptosomes. After 36 min of superfusion to equilibrate the system, four 3-min fractions were collected. Synaptosomes were exposed to GABA at the end of the first fraction collected. Allopregnanolone (3α,5α-P) or DHEAS was added 8 min before GABA. In some experiments (see Fig. 1), 3α,5α-P was added at the end of the first fraction collected. Fractions collected and superfused synaptosomes were counted for radioactivity. Cerebellar synaptosomes. After 36 min of superfusion fractions were collected according to the following scheme: two 4-min samples (basal release) before and after one 8-min sample (K +-evoked release). A 90-s period of depolarization was applied after the first sample had been collected. Depolarization of synaptosomes was performed with 35 mM KCl, substituting for an equimolar concentration of NaCl. GABA or 3α,5α-P was added to the superfusion medium concomitantly with the depolarizing stimulus. 3α,5α-P (when used as an allosteric modulator), DHEAS, bicuculline, picrotoxin or 2-(3’-carbethoxy-2’-propenyl)-3-amino-6-paramethoxy-phenyl-pyridazinium bromide (SR 95531) was added 8 min before. Fractions collected and superfused synaptosomes were counted for radioactivity. Calculation. Neurotransmitter efflux in the fractions collected was calculated as a percentage of the total tissue tritium content at the onset of the fraction considered. In the experiments of [ 3H]-NA release, the effects of drugs on the spontaneous release of [3H]NA were eval-

403 uated by performing the ratio between the efflux in the third fraction collected (in which the maximum effect of GABA was reached) and that of the first fraction. This ratio was compared to the ratio obtained under control conditions. The depolarization-evoked overflow of [3H]-D-ASP was estimated by subtracting the transmitter content of the basal release from the release evoked in the 8-min fraction collected during and after the depolarization pulse. Drug effects were evaluated as the ratio of the depolarization-evoked overflow in the presence of the drug vs. that calculated under control conditions. Appropriate controls were always run in parallel. Data were compared by one-way ANOVA followed by MannWhitney U-test. Drugs. [3H]-noradrenaline ([3H]-NA; specific activity, 39 Ci/mmol) and [3H]-D-aspartate ([3H]-D-ASP; specific activity, 24 Ci/mmol) were obtained from Amersham Radiochemical Centre (Buckinghamshire, UK). GABA, allopregnanolone (3α,5α-P), (+)-bicuculline and picrotoxin were obtained from Sigma Chemical Co. (St. Louis, Mo., USA). Dehydroepiandrosterone sulphate was obtained from RBI (Natick, Mass., USA). 2-(3’-carbethoxy-2’-propenyl)-3-amino-6paramethoxy-phenyl-pyridazinium bromide (SR 95531) was a gift from SANOFI (Brussels, Belgium). 3α,5α-P and DHEAS were solubilized in dimethylsulphoxide to a concentration of 10 mM and then diluted to working concentrations with physiological medium. The maximum concentration of dimethylsulphoxide used (0.1%) did not modify, on its own, the release of [3H]-NA or [3H]-D-ASP or their modifications by GABA.

Results It was previously shown that GABA, acting at GABAA receptors, increased the basal release of [3H]-NA from rat hippocampal synaptosomes and the K+-evoked release of [3H]-D-ASP from rat cerebellar synaptosomes exposed in superfusion to varying concentrations of GABA. The release of [3H]-NA evoked by high-K+ or the basal release of [3H]-D-ASP, respectively, were not increased by GABA (Schmid et al. 1996 and references therein). To investigate the action of neurosteroids at these two receptors, two concentrations of GABA were chosen which, based on the concentration-response curves previously obtained (Schmid et al. 1996) produced comparable enhancements of [3H]-NA and [3H]-D-ASP release. Since the use of low concentrations of GABA led to somewhat variable neurosteroid modulations, we used higher concentration of GABA also considering that neurosteroids, differ-

ently from benzodiazepines, have been reported to produce, in combination with GABA, effects higher than the maximal effects produced by the amino acid alone (Turner et al. 1989; Horne et al. 1993; Goodnough and Hawkinson 1995). Figure 1 (upper panel) illustrates the effects of the neurosteroid 3α,5α-P on the GABA-evoked release of [3H]-NA and on the basal release of the [3H]-catecholamine (in the absence of GABA) from superfused rat hippocampal synaptosomes. The release of [3H]-NA elicited by 5 µM GABA was unaffected by 0.01 µM 3α,5α-P. The neurosteroid potentiated GABA slightly, though not significantly, at 0.03 µM. The positive modulation by 3α,5α-P was significant at 0.1 µM, the maximal effect (about 90%) being reached between 0.1 and 0.3 µM. With no GABA added to the superfusion medium, 3α,5α-P (0.01–10 µM) did not modify the basal release of [3H]-NA. As shown in the lower panel of Fig. 1, the GABA(3 µM)-induced potentiation of the release of [3H]D-ASP evoked by K+-depolarization from superfused rat cerebellar synaptosomes was further increased by 3α,5α-P. The pattern of the 3α,5α-P effect, in the presence of GABA, was similar to that observed with [3H]-NA release from hippocampal synaptosomes: 0.1 µM of 3α,5α-P increased the effect of 3 µM GABA by about 100%. Moreover, in the absence of GABA, 3α,5α-P was able to enhance the depolarization-evoked release of [3H]-D-ASP from cerebellar synaptosomes when added at 1–10 µM. No such enhancement could be seen at concentrations (0.1–0.3 µM) at which 3α,5α-P potentiated the effect of GABA. The potentiation by 0.1 µM 3α,5α-P of the GABA(5 µM)-evoked [3H]-NA release from hippocampal nerve endings was prevented by DHEAS, a neurosteroid reported to be a non-competitive blocker of GABAA receptors (Majewska et al. 1990; Spivak 1994). Added at 1 µM, DHEAS abolished the effect of 0.1 µM 3α,5α-P. However, DHEAS (1 or 10 µM) had no effect, on its own, on the basal release or the GABA(5 µM)-evoked release of [3H]-NA (Table 1). DHEAS (1 µM) also blocked the potentiation by 0.1 µM 3α,5α-P of the GABA effect on the release of [3H]-D-ASP

Table 1 Effect of allopregnanolone and DHEAS on the GABA-induced potentiation of the basal release of [3H]-noradrenaline from hippocampal synaptosomes Drugs

% of potentiation

5 µM GABA 5 µM GABA + 0.1 µM 3α,5α-P 5 µM GABA + 0.1 µM 3α,5α-P + 0.1 µM DHEAS 5 µM GABA + 0.1 µM 3α,5α-P + 1 µM DHEAS 5 µM GABA + 1 µM DHEAS 5 µM GABA + 10 µM DHEAS 10 (µM DHEAS

30.6 54.2 47.0 31.1 32.0 30.3 6.2

Synaptosomes were exposed to GABA at the end of the first fraction collected; 3α,5α-P and DHEAS were introduced 8 min before GABA. See Materials and methods for additional technical details. Data represent means ±SEM of n experiments run in triplicate (three

superfusion chambers for each condition) on different days. * P<0.01 when compared to the effect of GABA on the spontaneous [3H]-NA release; ** P<0.05 when compared to the effect of GABA + 3α,5α-P

± ± ± ± ± ± ±

1.8 4.0 * 3.0 * 4.2 ** 2.2 4.1 5.0

n 12 8 6 4 4 4 3

404 Table 2 Effect of allopregnanolone and DHEAS on the GABA-induced potentiation of the K+-evoked overflow of [3H]-D-aspartate from cerebellar synaptosomes Drugs

% of potentiation

3 µM GABA 3 µM GABA + 0.1 µM 3α,5α-P 3 µM GABA + 0.1 µM 3α,5α-P + 0.1 µM DHEAS 3 µM GABA + 0.1 µM 3α,5α-P + 1 µM DHEAS 3 µM GABA + 1 µM DHEAS 3 µM GABA + 10 µM DHEAS 10 (µM DHEAS

23.8 49.3 40.0 23.0 20.9 20.1 2.2

Synaptosomes were exposed to GABA concomitantly with the depolarizing stimulus; 3α,5α-P and DHEAS were introduced 8 min before GABA. See Materials and methods for additional technical details. Data represent means ±SEM of n experiments run in triplicate.

* P<0.01 when compared to the effect of GABA on the depolarization-evoked [3H]-D-ASP overflow; ** P<0.05 when compared to the effect of GABA + 3α,5α-P

evoked by K+ from cerebellar synaptosomes (Table 2). DHEAS (added at 1 or 10 µM) did not modify either the K+-evoked release of [3H]-D-ASP or the effect of GABA on the K+-evoked [3H]-D-ASP release. We finally tried to characterize the [3H]-D-ASP releasing effects displayed by 3 or 10 µM 3α,5α-P in the absence of GABA. As shown in Fig. 2, the enhancements of the K+evoked [3H]-D-ASP release elicited in cerebellar synaptosomes by 3 or 10 µM 3α,5α-P could be abolished by bicuculline or SR 95531, two antagonists at the GABA recognition site of the GABAA receptor, and by picrotoxin, an antagonist at the Cl– channel. When added on their own at 10 µM bicuculline, SR 95531 or picrotoxin did not affect significantly the K+-evoked release of [3H]-D-ASP (% change of overflow vs. control: 3.7±6, n = 4; –0.8±1.9, n = 8; 2.3±1.9, n = 6, respectively) in line with the idea that, in our superfusion system, endogenous GABA or/and neurosteroids released during K+ depolarization are removed before GABAA receptor activation. Figure 2 also shows that the effect of 3 µM 3α,5α-P was partly (about 50%) prevented by 10 µM DHEAS. This antagonism by

DHEAS appears to be maximal, as 30 µM of the compound did not further decrease the effect of 3α,5α-P.

Fig. 2 Effect of bicuculline, SR 95531, picrotoxin or DHEAS on the 3α,5α-P-induced potentiation of the K+-evoked overflow of [3H]-D-ASP from cerebellar synaptosomes. Synaptosomes were exposed to 3α,5α-P concomitantly with the depolarizing stimulus; antagonists were introduced 8 min before 3α,5α-P. See Materials and methods for additional technical details. Data represent means ±SEM of 4–10 experiments in quadruplicate. * P<0.05 and ** P<0.001 when compared to the effect of 3α,5αP on the depolarization-evoked [3H]-D-ASP overflow

± ± ± ± ± ± ±

3.4 3.0 * 2.5 * 3.0 ** 3.7 4.9 5.0

n 10 6 5 4 3 3 3

Discussion Native GABAA receptor subtypes and the neurosteroid allosteric sites It has been recently proposed that the GABAA receptors here investigated represent two pharmacologically distinct subtypes of the GABAA receptor present in the adult rat CNS. The hippocampal receptor is expressed by noradrenergic neurons and is sensitive to benzodiazepine agonists. The cerebellar receptor is expressed by glutamatergic (granule) neurons and is benzodiazepine-insensitive (Schmid et al. 1996; see also Korpi et al. 1993 and references therein). Moreover, the function of the two GABAA receptors is voltage-dependent: the hippocampal receptor can evoke NA release only when the membrane on which it is

405

localized i under resting potential (Fung and Fillenz 1983; Bonanno and Raiteri 1987), whereas the cerebellar receptor is silent at resting potential but mediates increase of glutamate release when the membrane is depolarized (Gallo et al. 1981; Levi and Gallo 1981). GABAA receptor-induced depolarization has often been described (Cherubini et al. 1991; Gonzalez et al. 1992; Staley et al. 1995; see, for a review, Kaila 1994). The mechanisms involved (which lie beyond the scope of the present work) have been discussed in detail by the authors above. The first result of the present work shows that nanomolar concentrations of 3α,5α-P can enhance the effects of GABA at both the hippocampal and the cerebellar receptor. The potentiation by 3α,5α-P at the two receptor subtypes appears quantitatively similar in terms of maximal effect and approximate affinity. Since neuroactive steroids are known to allosterically enhance GABAA receptor-mediated effects at nanomolar concentrations (Simmonds 1991; Paul and Purdy 1992; Lambert et al. 1995), the effects elicited by low nanomolar concentrations of 3α,5α-P, in the presence of GABA, can be interpreted as due to activation by 3α,5αP of a neurosteroid allosteric site present on both the hippocampal and the cerebellar GABAA receptors. As to the relations between allosteric effects of nanomolar steroid concentrations and subunit composition of GABAA receptors, results from molecular biology studies with various subunit combinations are in part conflicting, but it seems that neurosteroids, unlike benzodiazepines, do not exhibit an absolute GABAA receptor subunit specificity (Lambert et al. 1995; Rabow et al. 1995). The two native GABAA receptors here characterized may represent a good example. In fact, the hippocampal and the cerebellar GABAA receptors display differential sensitivity to benzodiazepines and to drugs acting at the benzodiazepine site (Schmid et al. 1996), most likely due to different subunit composition (Doble and Martin 1992; Laurie et al. 1992; Korpi et al. 1993), but possess similar neurosteroid allosteric sites modulating the effects of GABA. Differential direct activation of GABAA receptors by 3α,5α-P As shown in Fig. 1, concentrations of 3α,5α-P higher than those required to potentiate GABA increased the release of [3H]-D-ASP, but not that of [3H]NA, in the absence of GABA. One may object that, in the cerebellum, 3α,5α-P does not act on its own, but by enhancing the effect of endogenous GABA released during superfusion. This is, in our opinion, unlikely. In fact, the technique employed (a thin layer of synaptosomes up-down superfused on a microporous filter; Raiteri et al. 1974; Raiteri and Levi 1978) has often been shown to permit immediate removal by the superfusion fluid of the endogenous compounds released. With no drug added, the GABAA receptor biophase should therefore remain virtually GABA- and neurosteroidfree. Indeed classical GABAA receptor antagonists had no

effect, on their own, on [3H]-D-ASP release and DHEAS did not antagonize the releases evoked by GABA alone; furthermore, maximal potentiation of the GABA effect was observed at 0.1 µM of 3α,5α-P, whereas the steroid alone was ineffective at 0.3 µM, thus excluding involvement of endogenous GABA released by high K+. On the other hand, previous studies with different systems (Cottrell et al. 1987; Morrow et al. 1987, 1990; Puia et al. 1990; Kokate et al. 1994), had shown that steroids, at low micromolar concentration, can evoke directly GABA-mimetic responses. Where neurosteroids bind to elicit these direct effects is not clear. It should be noted that, in most cases, the ’steroid only’ effects are sensitive to bicuculline. This also occurs in our system, where the effect of 3α,5α-P alone on [3H]-DASP release was abolished by bicuculline and by SR 95531, both competitive antagonists at the GABA recognition site of the GABAA receptor, as well as by the chloride channel blocker picrotoxin. Based on these data, it would seem justified to propose that 3α,5α-P acts at the GABAA receptor on glutamatergic cerebellar terminals by mimicking GABA, i.e. as an agonist at the GABA recognition site. However, this apparently simple idea is contrasted by previous binding studies showing that the sites for steroid binding differ from the GABA-binding site (see Macdonald and Olsen 1994) and by results of a very recent work with mutated GABAA receptors supporting the view that steroids do not bind to the GABA-binding site when they directly gate the channel of the GABAA receptor (Ueno et al. 1997). Altogether, the data available seem compatible with the view that 3α,5α-P acts at two sites on the cerebellar GABAA receptor: a) a high-affinity, DHEAS-sensitive site mediating allosteric potentiation of GABA effects; b) a low-affinity site mediating direct GABA-like effects, in absence of GABA. Moreover, our results suggest that these direct effects of 3α,5α-P may be allosterically enhanced by 3α,5α-P itself when acting at the high-affinity site. In fact, the effect of 3 µM 3α,5α-P was inhibited (by a maximum of 50%) by DHEAS (Fig. 2), a likely antagonist at the highaffinity neurosteroid site (see below). The lack of effect of 3α,5α-P alone at the hippocampal GABAA receptor suggests that the low-affinity bicucullinesensitive neurosteroid site may not be present on all GABAA receptors. Thus, in the particular system here studied, both the hippocampal and the cerebellar receptors possess an high-affinity allosteric neurosteroid site, while only the latter receptor is endowed with a low affinity site where neurosteroids bind and elicit direct GABA-like effects. As previously discussed (Schmid et al. 1996), the benzodiazepine-sensitive hippocampal and the benzodiazepineinsensitive cerebellar GABAA receptors seem to be α1 subunit- and α6 subunit-containing receptors, respectively. Interestingly, results from recombinant receptor experiments show that 3α,5α-P enhanced the effects of GABA in both α1 β2 γ2 and α6 β2 γ2 receptors, while the action of the neurosteroid without added GABA was more pronounced in the α6-containing than in the α1-containing receptor (Korpi and Lüddens 1993).

406

Is DHEAS an antagonist at the allosteric neurosteroid site? DHEAS is one of a variety of neurosteroids thought to be formed de novo in mammalian brain (Corpechot et al. 1983; Baulieu and Robel 1990). The steroid has been defined as an allosteric antagonist of the GABAA receptor (Majewska et al. 1990). Although the site of binding for DHEAS has remained elusive, the inhibitory effects of the steroid on GABA-induced currents in cultured neurons were found to be consistent with a non-competitive specific antagonistic action (Demirgören et al. 1991; Spivak 1994). The results obtained with DHEAS in the present work lend themselves to a different interpretation. As shown in Tables 1 and 2, DHEAS (1 or 10 µM) did not affect the GABA-induced release of [3H]-NA or [3H]-D-ASP, in the absence of 3α,5α-P; however, at 1 µM or less, DHEAS prevented the positive allosteric activity of nanomolar 3α,5αP on the GABA effects. Thus it seems that DHEAS behaves as an antagonist of the allosteric effects elicited by 3α,5αP, but not of the effects elicited by GABA in the absence of 3α,5α-P. Different interpretations of the mechanism of action of DHEAS may arise from the different experimental conditions employed. With our technique of superfusion, the effects of the drugs under study added to the medium can hardly be affected by endogenous compounds (steroids, benzodiazepine-like substances, GABA etc.) released by the tissue. In other preparations, including brain slices or cultured cells, such endogenous compounds may remain in the GABAA receptor biophase and influence the effects of the drugs under study. In particular, endogenous agonists at the GABAA neurosteroid sites released from the tissue would lead to overestimation of GABA-evoked effects taken as controls; addition of DHEAS will decrease these effects by blocking the neurosteroid site, but this event could be interpreted as a non-competitive antagonism of the GABA response. To conclude, based on our results and the above considerations, DHEAS may represent an antagonist at the neurosteroid allosteric high-affinity binding site on the GABAA receptor. Since DHEAS is an endogenous neurosteroid, it could be speculated that the compound is synthesized to antagonize excessive activation of the modulatory neurosteroid site, without directly affecting the primary GABA-mediated transmission. Administration of DHEAS should result in increased neuronal excitability, an effect that was indeed observed during iontophoretic application of DHEAS onto guinea-pig brain neurons (Carette and Poulain 1984). Circadian variations in the DHEAS contents of the rat brain were observed (Baulieu et al. 1987), and this suggested that DHEAS could act as an endogenous analeptic (Demirgören et al. 1991). It was also found that DHEAS can enhance retention of long-term memory in mice (Roberts 1990; Flood et al. 1992). Acknowledgements This work was supported by grants from the Italian MURST and from the Italian CNR. The authors wish to thank Mrs. Maura Agate for her collaboration in preparing this manuscript.

References Barnard EA (1996) The molecular biology of GABAA receptors and its applications. In: Tanaka C, Bowery NG (eds) GABA: Receptors, transporters and metabolism. Birkhäuser, Basel, Switzerland, pp 129–143 Baulieu E-E, Robel P (1990) Neurosteroids: a new brain function? J Steroid Biochem Mol Biol 37: 395–403 Baulieu E-E, Robel P, Vatier O, Haug M, Le Goascogne C, Bourreau E (1987) Neurosteroids: pregnenolone and dehydroepiandrosterone in the brain. In: Fuxe K, Agnati LF (eds) Receptor-receptor interaction, a new intramembrane integrative mechanism. MacMillan, Basingstoke, pp 89–104 Biggio G, Concas A, Costa E eds (1992) GABAergic synaptic transmission. Raven Press, New York Bonanno G, Raiteri M (1987) Release-regulating GABAA receptors are present on noradrenergic nerve terminals in selective areas of the rat brain. Synapse 1: 254–257 Carette B, Poulain P (1984) Excitatory effect of dehydroepiandrosterone, its sulphate ester and pregnenolone sulphate, applied by iontophoresis and pressure, on single neurones in the septo-preoptic area of the guinea pig. Neurosci Lett 45: 205–210 Cherubini E, Gaiarsa JL, Ben Ari T (1991) GABA: an excitatory transmitter in early postnatal life. Trends Neurosci 14: 515–519 Concas A, Mostallino MC, Perra C, Lener R, Roscetti G, Barbaccia ML, Purdy RH, Biggio G (1996) Functional correlation between allopregnanolone and [35S]-TBPS binding in the brain of rats exposed to isoniazid, pentylenetetrazol, or stress. Br J Pharmacol 118: 839–846 Corpéchot C, Synguelakis M, Talha S, Axelson M, Sjövall J, Vihko R, Baulieu E-E, Robel P (1983) Pregnenolone and its sulfate ester in the rat brain. Brain Res 270: 119–125 Cottrell GA, Lambert JJ, Peters JA (1987) Modulation of GABAA receptor activity by alphaxalone. Br J Pharmacol 90: 491–500 Demirgören S, Majewska MD, Spivak CE, London ED (1991) Receptor binding and electrophysiological effects of dehydroepiandrosterone sulfate, an antagonist of the GABAA receptor. Neuroscience 45: 127–135 Doble A, Martin IL (1992) Multiple benzodiazepine receptors: no reason for anxiety. Trends Pharmacol Sci 13: 76–81 Flood JF, Morley JE, Roberts E (1992) Memory-enhancing effects in male mice of pregnenolone and steroids metabolically derived from it. Proc Natl Acad Sci USA 89: 1567–1571 Fung S-C, Fillenz M (1983) The role of pre-synaptic GABA and benzodiazepine receptors in the control of norepinephrine release in rat hippocampus. Neurosci Lett 42: 61–66 Gallo V, Levi G, Raiteri M, Coletti A (1981) Enhancement by GABA of glutamate depolarization-induced release from cerebellar nerve endings. Brain Res 205: 431–435 Gee KW, Bolger MB, Brinton RE, Coirini H, McEwen BS (1988) Steroid modulation of the chloride ionophore in rat brain: structureactivity requirements, regional dependence and mechanism of action. J Pharmacol Exp Ther 246: 803–812 Gonzalez MP, Oset-Gasque MJ, Bugeda J, Arce C, Castro E, Parramón M (1992) GABAA receptor modulates catecholamine secretion from bovine chromaffin cells depending on the membrane potential. Neuroscience 47: 487–494 Goodnough DB, Hawkinson JE (1995) Neuroactive steroid modulation of [3H]muscimol binding to the GABAA receptor complex in rat cortex. Eur J Pharmacol Mol Pharmac Sect 288: 157–162 Harrison NL, Majewska MD, Harrington JW, Barker JL (1987) Structure-activity relationships for steroid interaction with the γ-aminobutyric acidA receptor complex. J Pharmacol Exp Ther 241: 346–353 Hawkinson JE, Kimbrough CL, Belelli D, Lambert JJ (1994a) Correlation of neuroactive steroid modulation of [35S]t-butylbicyclophosphorothionate and [3H]flunitrazepam binding and γ-aminobutyric acidA receptor function. Mol Pharmacol 46: 977–985

407 Hawkinson JE, Kimbrough CL, McCauley LD, Bolger MB, Lan NC, Gee KW (1994b) The neuroactive steroid 3α-hydroxy-5β-pregnan-20-one is a two-component modulator of ligand binding to the GABAA receptor. Eur J Pharmacol Mol Pharmacol Sect 269: 157–163 Hawkinson JE, Acosta-Burruel M, Kimbrough CL, Goodnough DB, Wood PL (1996) Steroid inhibition of [3H]SR 95531 binding to the GABAA recognition site. Eur J Pharmacol 304: 141–146 Horne AL, Harkness PC, Hadingham KL, Whiting P, Kemp JA (1993) The influence of the γ2L subunit on the modulation of responses to GABAA receptor activation. Br J Pharmacol 108: 711–716 Kaila K (1994) Ionic basis of GABAA receptor channel function in the nervous system. Prog Neurobiol 42: 489–537 Kokate TG, Svensson BE, Rogawski MA (1994) Anticonvulsant activity of neurosteroids: correlation with γ-aminobutyric acidevoked chloride current potentiation. J Pharmacol Exp Ther 270: 1223–1229 Korpi ER, Lüddens H (1993) Regional γ-aminobutyric acid sensitivity of t-butylbicyclophosphoro[35S]thionate binding depends on γ-aminobutyric acidA receptor α subunit. Mol Pharmacol 44: 87–92 Korpi ER, Kleingoor C, Kettenmann H, Seeburg PH (1993) Benzodiazepine-induced motor impairment linked to point mutation in cerebellar GABAA receptor. Nature 361: 356–359 Lambert JJ, Belelli D, Hill-Venning C, Peters JA (1995) Neurosteroids and GABAA receptor function. Trends Pharmacol Sci 16: 295–303 Laurie DJ, Seeburg PH, Wisden W (1992) The distribution of 13 GABAA receptor subunit mRNAs in the rat brain. II. Olfactory bulb and cerebellum. J Neurosci 12: 1063–1076 Levi G, Gallo V (1981) Glutamate as a putative transmitter in the cerebellum stimulation by GABA of glutamic acid release from specific pools. J Neurochem 37: 22–31 Lüddens H, Korpi ER, Seeburg PH (1995) GABAA benzodiazepine receptor heterogeneity: neurophysiological implications. Neuropharmacology 34: 245–254 Macdonald RL, Olsen RW (1994) GABAA receptor channels. Annu Rev Neurosci 17: 569–602 Majewska MD, Harrison NL, Schwartz RD, Barker JL, Paul SM (1986) Steroid hormone metabolites are barbiturate-like modulators of the GABA receptor. Science 232: 1004–1007 Majewska MD, Demirgören S, Spivak CE, London ED (1990) The neurosteroid dehydroepiandrosterone sulfate is an allosteric antagonist of the GABAA receptor. Brain Res 526: 143–146 Morrow AL, Suzdak PD, Paul SM (1987) Steroid hormone metabolites potentiate GABA receptor-mediated chloride ion flux with nanomolar potency. Eur J Pharmacol 142: 483–485 Morrow AL, Pace JR, Purdy RH, Paul SM (1990) Characterization of steroid interactions with γ-aminobutyric acid receptor-gated chloride ion channels: evidence for multiple steroid recognition sites. Mol Pharmacol 37: 263–270 Olsen RW, Sapp DW (1995) Neuroactive steroid modulation of GABAA receptors. In: Biggio G, Sanna E, Costa E (eds) GABAA

receptors and anxiety: from neurobiology to treatment. Raven Press, New York, pp 57–74 Paul SM, Purdy RH (1992) Neuroactive steroids. FASEB J 6: 2311–2322 Peters JA, Kirkness EF, Callachan H, Lambert JJ, Turner AJ (1988) Modulation of the GABAA receptor by depressant barbiturates and pregnane steroids. Br J Pharmacol 94: 1257–1269 Prince RJ, Simmonds MA (1993) Differential antagonism by epipregnanolone of alphaxalone and pregnanolone potentiation of [3H]flunitrazepam binding suggests more than one class of binding site for steroids at GABAA receptors. Neuropharmacology 32: 59–63 Puia G, Santi MR, Vicini S, Pritchett DB, Purdy RH, Paul SM, Seeburg PH, Costa E (1990) Neurosteroids act on recombinant human GABAA receptors. Neuron 4: 759–765 Rabow LE, Russek SJ, Farb DH (1995) From ion currents to genomic analysis: recent advances in GABAA receptor research. Synapse 21: 189–274 Raiteri M, Levi G (1978) Release mechanisms for catecholamines and serotonin in synaptosomes. In: Ehrenpreis S, Kopin I (eds) Reviews of neuroscience. Raven Press, New York, pp 77–130 Raiteri M, Angelini F, Levi G (1974) A simple apparatus for studying the release of neurotransmitters from synaptosomes. Eur J Pharmacol 25: 411–414 Raiteri M, Bonanno G, Marchi M, Maura G (1984) Is there a functional linkage between neurotransmitter uptake mechanisms and presynaptic receptors? J Pharmacol Exp Ther 231: 671–677 Roberts E (1990) Dehydroepiandrosterone (DHEA) and its sulfate (DHEAS) as neural facilitators: Effects on brain tissue in culture and on memory in young and old mice. A cyclic GMP hypothesis of action of DHEAS and DHEAS in nervous system and other tissue. In: Kalimi M, Regeson W (eds) The biologic role of dehydroepiandrosterone (DHEA). Walter de Gruyter, Berlin, pp 13–42 Schmid G, Bonanno G, Raiteri M (1996) Functional evidence for two native GABAA receptor subtypes in adult rat hippocampus and cerebellum. Neuroscience 73: 697–704 Simmonds MA (1991) Modulation of the GABAA receptor by steroids. Semin Neurosci 3: 231–239 Spivak CE (1994) Desensitization and noncompetitive blockade of GABAA receptors in ventral midbrain neurons by a neurosteroid dehydroepiandrosterone sulfate. Synapse 16: 113–122 Staley KJ, Soldo BL, Proctor WR (1995) Ionic mechanisms of neuronal excitation by inhibitory GABAA receptors. Science 269: 977–981 Turner DM, Ransom RW, Yang JSJ, Olsen RW (1989) Steroid anesthetics and naturally occurring analogs modulate the γ-aminobutyric acid receptor complex at a site distinct from barbiturates. J Pharmacol Exp Ther 248: 960–966 Ueno S, Bracamontes J, Zorumski C, Weiss DS, Steinbach JH (1997) Bicuculline and gabazine are allosteric inhibitors of channel opening of the GABAA receptor. J Neurosci 17: 625–634