Functional diversity of GABA-activated Cl- currents in Purkinje versus granule neurons in rat cerebellar slices.

In rat cerebellar slices, we compared whole-cell recordings of spontaneous inhibitory postsynaptic currents (sIPSCs) with Cl- currents resulting from pulses of GABA (1 mM, < 2 ms) to outside-out patches from Purkinje and granule neurons. sIPSCs in Purkinje cells decayed with a single fast exponential, as previously reported, whereas in granule cells sIPSC decay was best described by the sum of a fast and a slow exponential curve, with a variable contribution of the slow component to the peak current. GABA pulses to nucleated patches from granule cells elicited Cl- currents with decays similar to sIPSC decays, whereas in patches from Purkinje neurons GABA pulses produced Cl- currents decaying largely with a fast component, but often followed by a slower exponential. GABA concentration steps produced rapidly desensitizing currents in patches from both cerebellar neurons. In distinct cerebellar neurons, specific functional properties of GABAA receptors may relate to the presence of distinct receptor subtypes.

Neurosteroids act on recombinant human GABAA receptors.

The endogenous steroid metabolites 3 alpha,21dihydroxy-5 alpha-pregnan-20-one and 3 alpha-hydroxy-5 alpha-pregnan-20-one potentiate GABA-activated Cl- currents recorded from a human cell line transfected with the beta 1, alpha 1 beta 1, and alpha 1 beta 1 gamma 2 combinations of human GABAA receptor subunits. These steroids are active at nanomolar concentrations in potentiating GABA-activated Cl- currents and directly elicit bicuculline-sensitive Cl- currents when applied at micromolar concentrations. The potentiating and direct actions of both steroids were expressed with every combination of subunits tested. However, an examination of single-channel currents recorded from outside-out patches excised from these transfected cells suggests that despite the common minimal structural requirements for expressing steroid and barbiturate actions, the mechanism of GABAA receptor modulation by these pregnane steroids may differ from that of barbiturates.

Nongenomic regulation of glutamatergic neurotransmission in hippocampus by thyroid hormones.

Thyroid hormones (THs) are well known for their genomic effects but recently attention has focused also on their nongenomic actions as rapid modulators of membrane receptors. Here we show that thyroxine (T4) and 3,3',5'-l-triiodothyronine (T3) rapidly decrease N-methyl-d-aspartate (NMDA)-evoked currents in rat hippocampal cultures with potency in the micromolar range. The effect is not mediated by glutamate or glycine binding sites as an increase in agonist or glycine concentration does not alter TH potencies. Furthermore THs' effect on NMDA receptors is independent of voltage and of subunit composition. The mechanism of THs' antagonistic effect does not involve PKC phosphorylation of NMDA receptors since neither blocking nor stimulating PKC changed THs' modulation. T3, but not T4, inhibits also kainate-evoked currents in hippocampal neurons in culture. In hippocampal pyramidal neurons in slice, T3, but not T4, significantly reduced the frequency of miniature excitatory postsynaptic currents (mEPSCs) without affecting their amplitude and decay. In cultured rat cortical neurons THs prevented glutamate-induced neuronal death at concentrations similar to those effective on glutamatergic receptors. Taken together our data show for the first time that THs can rapidly affect ionotropic glutamatergic receptors in hippocampal neurons, an effect that could have an important role in their modulation of brain function in physiological and pathological states.

Evidence that the beta-acids fraction of hops reduces central GABAergic neurotransmission.

Humulus lupulus (hops) is traditionally used as a tranquilizing herbal remedy. Here, we investigated the in vivo and in vitro effect of hop beta-acids on central nervous system function. Oral administration of beta-acids (5-10mg/kg) in rats produced an increased exploratory activity in the open field, a reduction in the pentobarbital hypnotic activity and a worsening of picrotoxin-induced seizures. When dosed at 10mg/kg, beta-acids increased, in the elevated plus maze, open arm entries reducing in parallel those in closed arms. In the forced swimming test, we observed a reduction in the immobility time that could suggest an antidepressant-like activity. Electrophysiological studies performed on cerebellar granule cells in culture showed that the beta-acids fraction decreased GABA-evoked current in a dose-dependent way. The effect was not inhibited by the benzodiazepine antagonist Ro 15-1788. Benzodiazepine receptors involvement was also excluded by [(3)H]-Ro 15-1788 binding assay. In conclusion, the behavioral effects of beta-acids fraction could be explained by a reduction in the GABAergic activity although we cannot rule out the involvement of other neurotransmitter systems.

Pro-cognitive activity in rats of 3-furan-2-yl-N-p-tolyl-acrylamide, a positive allosteric modulator of the α7 nicotinic acetylcholine receptor.

BACKGROUND AND PURPOSE: α7 nicotinic acetylcholine receptors (α7 nAChRs) may represent useful targets for cognitive improvement. The aim of this study is to compare the pro-cognitive activity of selective α7-nAChR ligands, including the partial agonists, DMXBA and A-582941, as well as the positive allosteric modulator, 3-furan-2-yl-N-p-tolyl-acrylamide (PAM-2). EXPERIMENTAL APPROACH: The attentional set-shifting task (ASST) and the novel object recognition task (NORT) in rats, were used to evaluate the pro-cognitive activity of each ligand [i.e., PAM-2 (0.5, 1.0, and 2.0 mg·kg(-1) ), DMXBA and A-582941 (0.3 and 1.0 mg·kg(-1) )], in the absence and presence of methyllycaconitine (MLA), a selective competitive antagonist. To determine potential drug interactions, an inactive dose of PAM-2 (0.5 mg·kg(-1) ) was co-injected with inactive doses of either agonist - DMXBA: 0.1 (NORT); 0.3 mg·kg(-1) (ASST) or A-582941: 0.1 mg·kg(-1) . KEY RESULTS: PAM-2, DMXBA, and A-582941 improved cognition in a MLA-dependent manner, indicating that the observed activities are mediated by α7 nAChRs. Interestingly, the co-injection of inactive doses of PAM-2 and DMXBA or A-582941 also improved cognition, suggesting drug interactions. Moreover, PAM-2 reversed the scopolamine-induced NORT deficit. The electrophysiological results also support the view that PAM-2 potentiates the α7 nAChR currents elicited by a fixed concentration (3 µM) of DMXBA with apparent EC50 = 34 ± 3 µM and Emax = 225 ± 5 %. CONCLUSIONS AND IMPLICATIONS: Our results support the view that α7 nAChRs are involved in cognition processes and that PAM-2 is a novel promising candidate for the treatment of cognitive disorders.

On the putative physiological role of allopregnanolone on GABA(A) receptor function.

To obtain definitive evidence for a physiological allosteric modulatory role for endogenous brain ALLO on GABA(A) receptor function, we studied GABA(A) receptor activity under conditions in which the concentration of endogenous brain ALLO was decreased by about 80% for longer than 5 h following the administration of SKF 105111- 17beta-17-[bis (1methylethyl) amino carbonyl] androstane-3,5-diene-3-carboxylic acid (SKF), a potent inhibitor of 5alpha-reductases Type I and II. We used the in situ patch-clamp technique to record GABA-evoked currents and spontaneous inhibitory postsynaptic currents (sIPSCs) from pyramidal neurons in neocortical slices of vehicle- or SKF-treated mice. The potency, but not the efficacy, of exogenously applied GABA was decreased in slices from mice treated with SKF. When neocortical slices were treated in vitro for 3 h with 10 microM SKF, ALLO was also reduced (25-30%) and in addition, the GABA dose-response curve was shifted to the right; however this shift was not as marked as the shift in the slices obtained from mice treated with SKF, in keeping with the smaller decrease of the ALLO content in these slices. Furthermore, direct application of ALLO to these slices shifted the dose-response curve of GABA back toward a non-SKF treated profile. We then analyzed GABAergic sIPSCs in neocortical slices obtained from vehicle or SKF-treated mice. Mean decay time and charge transfer were significantly reduced by SKF treatment. The decay of sIPSCs was best fitted by two exponentials, but only the fast component was decreased in the SKF group. Direct application of ALLO (100 nM) normalizes the sIPSC kinetics in slices from ALLO depleted mice. No changes were detected in the amplitude or frequency of sIPSCs. These data demonstrate that endogenous ALLO physiologically regulates spontaneously induced Cl(-) current by acting on a specific recognition site, which is probably located on GABA(A) receptors (a receptor on a receptor), thereby prolonging inhibitory currents by facilitating conformational transition of the GABA-gated Cl(-) channel to an open state.

Brain allopregnanolone regulates the potency of the GABA(A) receptor agonist muscimol.

Allopregnanolone (ALLO), a potent positive-allosteric modulator of the action of GABA at GABA(A) receptors, is synthesized in the brain from progesterone by the sequential action of two enzymes: 5alpha-reductase and 3alpha-hydroxysteroidoxidoreductase. The concentration of ALLO in various parts of the mouse brain varies substantially, from 15 pmol/g in the olfactory bulb, to approximately 6 pmol/g in the frontoparietal cortex, and 2.7 pmol/g in the cerebellum. The systemic administration of 48 micromol/kg of the Type I and Type II 5alpha-reductase inhibitor, (17beta)-17-[bis (1-methylethyl) amino carbonyl)] androsta-3, 5-diene-3-carboxylic acid (SKF 105,111), reduced brain ALLO content by 80-90% in 30 min; the rate constant (k) of ALLO decrease in each brain area can be utilized to establish the rate of ALLO biosynthesis, which is higher in the olfactory bulb (62 pmol/g/h) than in the frontoparietal cortex (24 pmol/g/h) or cerebellum (11 pmol/g/h). The duration of the righting reflex loss elicited by the potent GABA(A) receptor agonist muscimol was reduced in SKF 105,111-treated ALLO-depleted mice. SKF 105,111 treatment had no effect on muscimol metabolism or on brain levels of pregnenolone and progesterone; however, the brain levels of 5alpha-DHP, the precursor of ALLO, were also decreased. Administration of ALLO at a dose of 15 micromol/kg i.p. by itself did not alter the muscimol-induced loss of the righting reflex; but it completely blocked the effect of SKF 105,111. To elucidate the possible molecular mechanism by which a decrease of brain ALLO content can shorten the duration of the righting reflex loss elicited by muscimol, we patch-clamped neocortical pyramidal neurons of mice pretreated with SKF 105,111 or vehicle, and studied the efficiency of muscimol in eliciting Cl- currents. The current amplitude was significantly smaller in neurons from SKF 105,111-treated mice, especially at lower doses (0.1-1 microM) of muscimol, and the muscimol dose-response (0.1-10 microM) relationship displayed cooperativity (nH=1.4). These data suggest that ALLO synthesized in brain plays an important physiological permissive role in the modulation of GABA-gated Cl- channel function.

Synthesis and anticonvulsant activity of novel and potent 2,3-benzodiazepine AMPA/kainate receptor antagonists.

We have previously shown that 1-aryl-3,5-dihydro-7, 8-methylenedioxy-4H-2,3-benzodiazepin-4-ones (3) possess marked anticonvulsant properties and antagonize seizures induced by 2-amino-3-(3-hydroxy-5-methylisoxazol-4-yl)propionic acid (AMPA) in analogy to the structurally related 1-(4-aminophenyl)-4-methyl-7, 8-methylenedioxy-5H-2,3-benzodiazepine (1, GYKI 52466), a well-known noncompetitive AMPA/kainate receptor antagonist. We now report the synthesis of 3-(N-alkylcarbamoyl)-1-aryl-3,5-dihydro-7, 8-methylenedioxy-4H-2,3-benzodiazepin-4-ones (4a-h) and 1-aryl-3, 5-dihydro-7,8-methylenedioxy-4H-2,3-benzodiazepine-4-thiones (5a-c). The activity of all compounds, intraperitoneally (ip) injected, was evaluated against audiogenic seizures in DBA/2 mice and against seizures induced by maximal electroshock (MES) and pentylenetetrazole (PTZ) in Swiss mice. Some of the new compounds 4 and 5 showed remarkable anticonvulsant activity, and their toxicity, as evidenced by the rotarod test, is lower than that of 1. The time course of anticonvulsant activity of derivatives 4b and 5b,c was studied and compared to that of 1 and 3b,c. Compounds 4a,b and 5a-c antagonize seizures induced by AMPA and kainate (KA) and their anticonvulsant activity is reversed by pretreatment with aniracetam. Using the patch-clamp technique, the capability of derivatives 3c, 4b, and 5c to antagonize KA-evoked currents in primary cultures of granule neurons was tested and compared with that of the parent compounds 1 and 1-(4-aminophenyl)-3, 4-dihydro-4-methyl-3-methylcarbamoyl-7,8-methylenedioxy-5H-2, 3-benzodiazepine (2, GYKI 53655).

The density and distribution of six GABAA receptor subunits in primary cultures of rat cerebellar granule cells.

In cultured cerebellar granule neurons (seven days in vitro) the expression of GABAA receptor subunits was quantified by using freeze-fracture immunocytochemical techniques with antibodies that specifically recognize the alpha 1, alpha 6, beta 2-3, gamma 2 and delta subunits of the GABAA receptor. In some experiments we have also used a less specific antibody that recognizes several alpha receptor subunits (alpha-total). The specificity of these antibodies was verified in human embryonic kidney cell line no. 293 cells transfected with complementary DNAs codifying for various GABAA receptor subunits. The most abundant labeling in granule cells was generated by the antibody against the beta 2-3 subunits (approximately 44 colloidal gold particles/microns2), while the specific antibodies against alpha 1 and alpha 6 subunits show a labeling of about 16 colloidal gold particles/microns2. The alpha-total antibody shows a labeling of approximately 37 gold particles/microns2. Both the gamma 2 and delta antibodies show a labeling of about 10 gold particles/microns2. In granule cells, the relative proportion of the label density revealed with antibodies against alpha-total, beta 2-3, gamma 2 and delta subunits is approximately 4:4:1:1. Assuming that one molecular form of the alpha subunit is assembled in a GABAA receptor, it can be estimated that in granule cells about 50% of receptors include the alpha 1 subunit. A similar relative abundance can be estimated for the alpha 6 subunit. The proportion of GABAA receptors containing the gamma 2 or delta subunits can be estimated to be about 50% in each case. Cerebellar granule cells express various abundances of GABAA receptor subunits which can be estimated by freeze-fracture immunocytochemistry. Fifty to sixty percent of these subunits form small receptor clusters, which appear to be associated with neuronal cytoskeleton proteins.

Pharmacological profile of apigenin, a flavonoid isolated from Matricaria chamomilla.

Dried flowers of Matricaria chamomilla L. are largely used to provide sedative as well as spasmolytic effects. In the present study, we examined in particular the pharmacological property of a fraction isolated from a methanolic extract of M. chamomilla, which was identified by HPLC-MS-MS analysis as apigenin. By radioreceptor binding assays, we demonstrated the ability of the flavone to displace a specific radioligand, [(3)H]Ro 15-1788, from the central benzodiazepine binding site. Electrophysiological studies performed on cultured cerebellar granule cells showed that apigenin reduced GABA (gamma-aminobutyric acid)-activated Cl(-) currents in a dose-dependent fashion. The effect was blocked by co-application of Ro 15-1788, a specific benzodiazepine receptor antagonist. Accordingly, apigenin reduced the latency in the onset of picrotoxin-induced convulsions. Moreover, apigenin injected i.p. in rats reduced locomotor activity, but did not demonstrate anxiolytic, myorelaxant, or anticonvulsant activities. The present results seem to suggest that the inhibitory activity of apigenin on locomotor behaviour in rats cannot be ascribed to an interaction with GABA(A)-benzodiazepine receptor but to other neurotransmission systems, since it is not blocked by Ro 15-1788.

Terfenadine prevents NMDA receptor-dependent and -independent toxicity following sodium channel activation.

Exposure of cultured cerebellar neurons to terfenadine prevented the N-methyl-D-aspartate (NMDA) receptor-mediated early appearance (30 min) of toxicity signs induced by the voltage sensitive sodium channel (VSSC) activator veratridine. Delayed neurotoxicity by veratridine (24 h) occurring independently from NMDA receptor activation was also prevented by terfenadine. Terfenadine did not protect from excitotoxicity following direct exposure of neurons to glutamate. Our results suggest that terfenadine may modulate endogenous glutamate release following activation of VSSCs.

Freeze-fracture immunocytochemical study of the expression of native and recombinant GABAA receptors.

To assess the density and distribution of native and recombinant GABAA receptors we used label-fracture and fracture-flip technologies combined with immunocytochemistry using monoclonal and polyclonal Abs directed against the extracellular domain of the GABAA receptor protein located in the freeze-fracture replicas. In cortical neurons there is a high density of GABAA receptors on both soma and dendrites with some areas were the density of receptors is higher, but there are no well defined clusters. In cerebellar granule cells most of the receptors are distributed in round clusters both in neurites and soma. In astroglial cells the receptor density is lower than in neurons and only occasionally they appear in clusters. In cells transfected with cDNAs encoding for various molecular forms of GABAA receptor subunits, the receptor density is moderate when cDNAs for alpha, beta and gamma subunits are cotransfected; however, on cells cotransfected with cDNAs for beta and gamma subunits the receptor density is significantly lower. Recombinant receptors appear randomly distributed and occasionally they aggregate in small groups.

Triazolam is more efficacious than diazepam in a broad spectrum of recombinant GABAA receptors.

Benzodiazepine-induced modifications of GABA (gamma-aminobutyric acid) activated Cl- currents were studied in native GABAA receptors expressed in neonatal rat brain cortical neurons in primary cultures and in recombinant GABAA receptors expressed in transformed human embryonic kidney cells (293) after a transient transfection with cDNAs encoding for different molecular forms of alpha, beta, and gamma subunits of GABAA receptors. The efficacy of triazolam in cortical neurons was higher than that of diazepam. In transfected cells, triazolam showed a greater efficacy as a positive modulator of GABA-elicited Cl- currents in alpha 1 beta 1 gamma 1, alpha 1 beta 1 gamma 2, alpha 1 beta 1 gamma 3, alpha 6 beta 1 gamma 2 and alpha 1 beta 3 gamma 2 receptors than diazepam, except in alpha 3 beta 1 gamma 2 receptors where diazepam was more efficacious. When triazolam and diazepam were applied together to GABAA receptors assembled by transfecting cDNAs encoding for alpha 1 beta 1 gamma 1 subunits, the action of triazolam was curtailed in a manner related to the dose of diazepam. In recombinant receptors assembled with alpha 1 beta 1 gamma 1 receptors, maximally active doses of triazolam were more efficacious than those of clonazepam, alpidem, zolpidem, diazepam or bretazenil.

Modulation of kainate--activated currents by diazoxide and cyclothiazide analogues (IDRA) in cerebellar granule neurons.

1. Patch-clamp technique was used in primary cultures of cerebellar granule neurons to study the modulation of the cyclothiazide analogue (IDRA21) and of the diazoxide derivative (IDRA 5) on KA-evoked currents. 2. The dose-response of kainic acid (KA) reveals an EC50=90 microM and an Hill coefficient of 1.3. IDRA 21 and cyclothiazide potentiate KA-evoked current in a dose dependent way, being cyclothiazide more potent but less efficacious than IDRA 21. Conversely IDRA 5 acts as a negative modulator of KA evoked -current. 3. Application of IDRA 21 and cyclothiazide results in a current potentiation of 125+/-18% and 80+/-12% respectively, while IDRA 5 decreases KA-current (-21+/-5%). Coapplication of cyclothiazide and IDRA 21 produces a potentiation of 110+/-17%, suggesting a competition of the two drugs for the same site. 4. In the same experimental model we studied the ability of IDRA compounds of promoting toxicity through AMPA-receptor activation. Under basal conditions AMPA treatment (50 microM for 1 hour) results in a negligible excitotoxicity. 5. In contrast similar treatment with AMPA + IDRA 21 (1 mM) or + IDRA 5 (1 mM) or + cyclothiazide (100 microM) induces citotoxicity. The neurotoxic damage induced by IDRA 21 and cyclothiazide is blocked by GYKI 53655 (50 microM) and by NBQX (10 microM). Interestingly GYKI and NBQX are ineffective in reducing IDRA 5 toxicity.

Evidence that total extract of Hypericum perforatum affects exploratory behavior and exerts anxiolytic effects in rats.

Clinical trials have extensively reported the ability of Hypericum perforatum extracts to exert a significant antidepressant activity. Hypericin, the main constituent of H. perforatum extract, is no more regarded as the active principle of the antidepressant activity of the drug. Hence, the question of which constituents are involved in the basic activity of the total extract, is still waiting for an answer. In the present study we focused our attention on the potential anxiolytic activity of H. perforatum total extract, and of some pure components such as protohypericin and a fraction containing hypericin and pseudohypericin. Herein we report that the total extract of H. perforatum increases the locomotor activity in the open field and exerts anxiolytic activity in the light-dark test, whereas the single components did not show any effect. Interestingly, the anxiolytic activity of the total extract was blocked by pretreatment of rats with the benzodiazepine antagonist Flumazenil, hence suggesting an implication of benzodiazepine receptor activation in the anxiolytic effect of H. perforatum extract. Electrophysiological studies, performed to gain more information on the mechanism of action, showed that hypericin reduced the GABA-activated chloride currents, while pseudohypericin did an opposite effect. Furthermore, both hypericin and pseudohypericin inhibited the activation of NMDA receptors.

Thyroid hormones modulate GABA(A) receptor-mediated currents in hippocampal neurons.

Thyroid hormones (THs) play a crucial role in the maturation and functioning of mammalian central nervous system. Thyroxine (T4) and 3, 3', 5-L-triiodothyronine (T3) are well known for their genomic effects, but recently attention has been focused on their non genomic actions as modulators of neuronal activity. In the present study we report that T4 and T3 reduce, in a non competitive manner, GABA-evoked currents in rat hippocampal cultures with IC50s of 13±4µM and 12±3µM, respectively. The genomically inactive compound rev-T3 was also able to inhibit the currents elicited by GABA. Blocking PKC or PKA activity, chelating intracellular calcium, or antagonizing the integrin receptor αVß3 with TETRAC did not affect THs modulation of GABA-evoked currents. THs affect also synaptic activity in hippocampal and cortical cultured neurons. T3 and T4 reduced to approximately 50% the amplitude and frequency of spontaneous inhibitory synaptic currents (sIPSCs), without altering their decay kinetic. Tonic currents evoked by low GABA concentrations were also reduced by T3 (40±5%, n=14), but not by T4. Similarly, T3 decreased currents elicited by low concentrations of THIP, a low affinity GABAA receptor agonist that preferentially activates extrasynaptic receptors, whereas T4 was ineffective. Thus, our data demonstrate that T3 and T4 selectively affect GABAergic phasic and tonic neurotransmission. Since THs concentrations can be regulated at the level of the synapses these data suggest that the network activity of the whole brain could be differently modulated depending on the relative amount of these two hormones. This article is part of a Special Issue entitled 'Trends in neuropharmacology: in memory of Erminio Costa'.

Evidence that isopropylthioxanthone (ITX) is devoid of anxiolytic and sedative effect.

Isopropylthioxanthone (ITX) is a well-known photo-initiator in ultraviolet light-cured inks frequently used in milk packaging materials, yoghurt, ready-to-feed infant formula, and other drinks. Traces of ITX have been found in milk and, as a consequence, there was considerable interest in studying the biological activity of this molecule and its potential hazard for the human health. Although the ITX genotoxic effects have been excluded by the European Food Safety Authority (EFSA), the US Environmental Protection Agency (USEPA) is still examining its possible toxic potential depending on a dose-effect ratio. Little is known about the ITX activity on the function of the central nervous system and cerebral neurotransmitters. Using behavioural, biochemical, and electrophysiological tests, the authors have found that: (1) ITX did not exert an in vivo anxiolytic or sedative effect when administered orally to rats; (2) ITX did not affect the binding characteristics of central and peripheral benzodiazepine receptors studied in vitro; and (3) ITX did not influence the ability of gamma-Aminobutyric acid (GABA) to increase the chloride channel permeability studied by patch clamp technique in a single neuron of cultured cerebellar granule cells.

Influence of recombinant gamma-aminobutyric acid-A receptor subunit composition on the action of allosteric modulators of gamma-aminobutyric acid-gated Cl- currents.

gamma-Aminobutyric acid (GABA)-activated Cl- currents in neonatal rat cortical neurons and in cultured cells engineered for the expression of specific molecular forms of the GABAA receptor alpha, beta, and gamma subunits, were recorded with the patch-clamp technique in the whole-cell configuration. The effects of various allosteric modulators of GABAA receptors were determined. Diazepam and clonazepam showed greater efficacy as positive modulators of GABA-elicited currents in alpha 2 beta 1 gamma 2 or alpha 3 beta 1 gamma 2 receptors than in alpha 1 beta 1 gamma 2 or alpha 5 beta 1 gamma 2 receptors or in cortical neurons. Alpidem was more efficacious at alpha 1 beta 1 gamma 2 or alpha 2 beta 1 gamma 2 receptors than at alpha 1 beta 1 gamma 2 or alpha 5 beta 1 gamma 2 receptors or in cortical neurons. Conversely, zolpidem was equally efficacious for all these receptors except for alpha 5 beta 1 gamma 2. Both imidazopyridines (alpidem and zolpidem) were virtually ineffective at modulating the GABA response of alpha 5 beta 1 gamma 2 receptors and in almost all the receptors assembled from alpha 1, alpha 2, alpha 3 or alpha 5 subunits together with beta 1 and gamma 1 subunits. The beta-carboline derivatives methyl-6,7-dimethoxy-4-ethyl-beta-carboline-3-carboxylate (DMCM) and methyl-beta-carboline-3-carboxylate (beta-CCM) elicited a positive allosteric modulation of alpha 1 beta 1 gamma 1 or alpha 2 beta 1 gamma 1 receptors, whereas they acted as negative allosteric modulators at nearly all other receptors tested, as they do in cortical neurons. Although the positive allosteric modulation by beta-carbolines never exceeded a doubling of the GABA response, DMCM was more efficacious at alpha 1 beta 1 gamma 1 receptors and beta-CCM was more efficacious at alpha 2 beta 1 gamma 1 receptors. DMCM was inactive at alpha 3 beta 1 gamma 1 receptors, whereas beta-CCM was virtually inactive at alpha 5 beta 1 gamma 1 receptors. The benzodiazepine 4'-chlorodiazepam, which is a negative modulator resistent to flumazenil inhibition, acted at all the various GABAA receptors that contained a gamma subunit.