Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Synaptic transmission is depressed by acute intoxicating doses of ethanol (for a review, see Shefner, 1990). Ethanol-induced neuronal depression might result from attenuation of excitatory amino acid and/or enhancement of inhibitory -aminobutyric acid (GABA) receptor-mediated transmission. Electrophysiological and biochemical studies have revealed that mildly intoxicating ethanol concentrations selectively inhibit glutamate receptor function both in vitro and in vivo (Hoffman et al., 1989; Lovinger et al., 1989, 1990; White et al., 1990; Simson et al., 1991; Nie et al., 1993, 1994).

Ethanol enhancement of GABA receptor-mediated neurotransmission is more controversial, perhaps due to its differential effects in distinct brain regions. At the biochemical level, ethanol augments GABA-stimulated chloride flux in brain membrane vesicles (Allan and Harris, 1986) and in cultured spinal neurons (Suzdak et al., 1986; Ticku et al., 1986; Mehta and Ticku, 1988). At the cellular level, ethanol-induced facilitations of GABA inhibition have been reported in chick spinal cord (Celentano et al., 1988), rat dorsal root ganglion (Nishio and Narahashi, 1990), and cultured mammalian hippocampal and cortical neurons (Aguayo, 1990). However, ethanol produced no effect or antagonized GABA-mediated responses in the hippocampus (Carlen et al., 1982; Siggins et al., 1987), cerebellum (Harris and Sinclair, 1984; Siggins et al., 1987; Palmer and Hoffer, 1990), and locus ceruleus (Osmanovic and Shefner, 1990). Under some conditions, behaviorally relevant concentrations of ethanol potentiate pharmacologically isolated GABA_A (Weiner et al., 1994, 1997), but not GABA_B (Frye and Fincher, 1996), receptor-mediated inhibitory postsynaptic currents (IPSCs) in CA1. Similarly, low ethanol concentrations reproducibly enhance GABA_A inhibitory postsynaptic potentials (IPSPs) of CA1 hippocampal neurons but only when GABA_B receptors were pharmacologically blocked (Wan et al., 1996).

Because the activation of GABA_B receptors modulates N-methyl-D-aspartate (NMDA) responses (Mott and Lewis, 1994), we evaluated in vivo and in vitro the role of GABA_B receptors in ethanol-induced inhibition of NMDA responses in the ventral tegmental area (VTA), hippocampus, and nucleus accumbens (NAcc), three structures known to be sensitive to acute intoxicating levels of ethanol.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Subjects and Surgical Preparation: In Vivo Experiments. Male Sprague-Dawley rats (250-350 g) were anesthetized with halothane (3.0-4.0%) and placed into a stereotaxic apparatus. Body temperature was monitored and maintained at 37.0 ± 0.1°C by a feedback-regulated heating pad. We drilled holes in the skull to accommodate the placement of stimulating and recording electrodes. The dura was opened over recording sites to prevent breakage of micropipettes. Halothane anesthesia was maintained at 0.75% after surgery.

Extracellular Recordings: In Vivo Experiments. Extracellular potentials were recorded with a single 3.0 M NaCl-filled micropipette (5-10 M tip resistance; 1-2 µm i.d.) cemented 20 to 40 µm distal to a seven-barrel micropipette (30-80 M tip resistance) leading to an Axon Instruments (Burlingame, CA) Axoprobe-1A microelectrode amplifier/head stage. The microelectrode assemblies were oriented stereotaxically into the VTA [coordinates from bregma: 5.6-6.0 mm posterior (P), 0.5-1.0 mm lateral (L), 7.0-8.5 mm ventral (V)], dentate hilus (4.0 mm P, 2.5 mm L, 2.8-3.1 mm V), or NAcc (1.5-1.7 mm A, 1.0-1.5 mm L, 5.8-7.0 mm V) with a Burleigh piezoelectric microdrive. Evoked field potential and single-unit activity were filtered at 0.1 Hz to 10 kHz and 1 to 3 kHz (3 dB), respectively. Responses were displayed on analog and digital oscilloscopes. Only those spikes that had a greater than 3:1 signal-to-noise ratio were evaluated. Spikes were discriminated and converted to computer-level pulses for interspike-interval histogram, peristimulus spike histogram, or firing rate analysis by National Instruments (Austin, TX) NB-MIO-16 multipurpose analog/digital, digital input/output, and counter/timer data acquisition boards. Extracellular potentials were digitized at 20 kHz and 12-bit voltage resolution. We considered a cell to be bursting if its pattern of discharge was characterized by multiple action potentials over a short time period (10-15 ms) with spike amplitude decrement and spike interval increment.

Characterization of Hippocampal, VTA GABA, and NAcc Core Neurons. A detailed description of the characterization of hilar mossy cells and hilar interneurons in the dentate gyrus has been previously reported (Mayer and Henriksen, 1995). In brief, hilar mossy cells and hilar interneurons could be readily distinguished by firing pattern and response to perforant path stimulation. Hilar interneurons were nonbursting neurons activated by perforant path stimulation, although in a much different manner than hilar mossy cells. Hilar mossy cells, presumed to be feedback excitatory neurons, had pronounced bursting activity and were driven by perforant path input only when the stimulus level was threshold for activating dentate granule cells. When sufficiently strong, stimulation of the perforant path evoked a synchronized discharge of dentate granule cells, resulting in a population spike (Bliss and Lomo, 1973). When the stimulus intensity was adjusted to produce a population spike amplitude of 1.0 to 1.5 mV, a single cell discharge, not a burst, was seen about 80 to 90% of the time, occurring within 1.0 to 3.0 ms after the peak negativity of the evoked population spike. Hilar interneurons had a much lower threshold for discharge by perforant path stimulation than hilar mossy cells. Because dentate granule cells are believed to be subject to feedback inhibition via hilar interneurons, paired stimuli to perforant path resulted in inhibition of the conditioned population spike at interstimulus intervals of less than 40 ms. During this period of dentate granule cell inhibition, hilar interneurons are active. Moreover, hilar interneurons, unlike hilar mossy cells, are activated by perforant path stimulation at thresholds and latencies that are often shorter than that of dentate granule cells, suggesting that they are also feedforward inhibitory interneurons.

Stimulation: In Vivo Experiments. Constant current square-wave pulses (50-1000 µA; 0.15-ms duration; average frequency, 0.05 Hz) were generated by a constant current isolation unit coupled to a Grass Instruments (Quincy, MA) S88 stimulator and triggered by a MASTER-8 pulse generator or by computer. We characterized hilar mossy cells and hilar interneurons by stimulation of the perforant path with insulated, bipolar stainless steel (130-µm) electrodes located in the angular bundle (coordinates: 8.1 mm P, 4.2 mm L, 3.0 mm V). For driven activity in the NAcc, stimulating electrodes were placed in the basolateral nucleus of the amygdala (3.0 mm P, 5.0 mm L, 8.0 mm V).

Data Analysis: In Vivo Experiments. We acquired, analyzed, and processed data by customized National Instruments LabVIEW software on MacIntosh computers. Extracellularly recorded single-unit action potentials were discriminated by a peak detector digital processing algorithm. To determine changes in firing rate produced by in situ microelectrophoretic application of NMDA, we determined the area under the response by rectangular integration from baseline with IGOR Pro software. As a control, drug effects on NMDA activation were compared with the average NMDA response obtained during microelectrophoretic application of saline (+200 nA; average of two saline ejections: one before and one after drug testing). To determine single-unit modal (e.g., bursting versus nonbursting) activity, interspike interval histograms were generated and normalized to number of spikes before and after drug/experimental treatment (1.0-s epochs, 2000 spikes, 2-ms bin width). For determinations of the probability of the occurrence of an NAcc amygdala-driven spike across stimulus levels, we generated peristimulus spike histograms at 0.5-Hz stimulation and averaged over 40 trials (±100-ms epoch, 2-ms bin width). The number of driven spikes were determined at each stimulus level by rectangular integration using IGOR Pro software. The results were derived for control and drug treatment groups from calculations performed on the driven activity and NMDA activation of spontaneous activity and expressed as mean ± S.E. We compared results for each point before and after drug treatment by the two-tailed t test.

Drug Preparation and Delivery: In Vivo Experiments. For in situ drug application in the VTA, dentate hilus, and NAcc, 50 mM DL-2-amino-5-phosphonovalerate (APV), 1 mM baclofen, 40 mM CGP35348, 0.3 to 3.0 M ethanol, 0.5 mM muscimol, and 40 mM NMDA were dissolved in 0.9% saline, loaded into seven-barrel glass micropipettes (tip i.d., 1 µm), and microelectrophoretically administered (microelectro-osmotically administered in the case of ethanol) by current injection (5-200 nA). We obtained APV, baclofen, and muscimol from Research Biochemicals International (Natick, MA). Ethanol was purchased from the Remet Corporation (La Mirada, CA). CGP35348 and CGP55845 were obtained from Novartis Pharma (Basel, Switzerland). NMDA was obtained from Sigma Chemical Co. (St. Louis, MO).

Slice Preparation and Intracellular Recordings: In Vitro Experiments. Coronal NAcc slices were prepared from male Sprague-Dawley rats (100-170 g) that had been anesthetized with 1 to 2% halothane, as described previously (Nie et al., 1993, 1994). In brief, we rapidly removed the brain from the skull and transversely sectioned the NAcc with a vibratome to obtain 300- to 400-µm-thick slices. The slices were immediately transferred to a recording chamber for incubation in an interface configuration for 30 min, followed by complete submersion and continuous superfusion (2-4 ml/min) with warm (30-31°C), oxygenated (95% O₂, 5% CO₂) artificial cerebrospinal fluid (aCSF) of the following composition: 130 mM NaCl, 3.5 mM KCl, 1.25 mM NaH₂PO₄, 1.5 mM MgSO₄·7H₂O, 2.0 mM CaCl₂, 24 mM NaHCO₃, and 10 mM glucose. Intracellular glass micropipettes were filled with 3 M KCl (tip resistance, 60-100 M). We performed current-clamp or single-electrode voltage-clamp studies using an Axon Instruments Axoclamp 2A preamplifier. The electrode settling time and capacitance neutralization were continuously monitored with a separate oscilloscope. Neuronal recordings were taken from the NAcc core region at 0.7 to 2.2 mm anterior from bregma and surrounding, but especially ventromedial to the anterior commissure. We elicited EPSPs by local stimulation near the recording pipette with a bipolar stimulating electrode. Input-output curves were generated by applying stimuli of different intensities (0.05- to 0.1-ms pulse duration) at a rate of 0.1 Hz.

Data Analysis: In Vitro Experiments. Once stable responses were achieved, electrophysiological measures were taken at several time points before, during, and after ethanol superfusion. Continuous direct current recordings were stored on a polygraph, and selected records were digitized, stored, and analyzed on an IBM 486-type computer using Axon Instruments pCLAMP programs. We subjected the data to ANOVA with repeated measures and the Newman-Keuls post hoc test. We considered P < .05 to be statistically significant.

Drug Preparation and Delivery: In Vitro Experiments. The slices were treated with APV (30 µM) or 6-cyano-7-nitroquinoxaline-2,3-dione (CNQX; 10 µM) for about 30 min to isolate the non-NMDA- and NMDA receptor-mediated components of EPSPs, respectively. To enhance the NMDA component, the cell was held around 60 mV. We applied NMDA (200 µM) by pressure application (pipette tip diameter, ~2 µm; pressure, 5-15 psi) near the recorded neuron. Solutions were introduced in known concentrations into the slice chamber, without disrupting the flow of the superfusate, by means of a multiple valve system. To avoid loss of ethanol by evaporation, we diluted the solutions in gassed aCSF from sealed stock solutions of reagent-grade 95% ethyl alcohol in water immediately before administration. Usually, control recordings were taken for 15 to 20 min during superfusion of aCSF alone to establish baseline measures. Then, the superfusate was switched to aCSF plus ethanol for 5 to 15 min to develop a full ethanol effect for further electrophysiological measures. This period was followed by immediate washout with aCSF alone for 10 to 20 min because ethanol superfusion for longer than 15 to 20 min often led to lack of reversal of the ethanol effects. We obtained tetrodotoxin (TTX) from Calbiochem (San Diego, CA). CNQX was obtained from Tocris (Bristol, UK). Bicuculline methiodide was obtained from Sigma Chemical Co.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

GABA_B Receptor Antagonists Reduce Ethanol Inhibition of NMDA Activation of Hippocampal Neurons In Vivo. Hilar mossy cells were located from 160 to 350 µm ventral to sign reversal of the population EPSP and evinced relatively slow spontaneous activity (2.4 ± 0.3 Hz; range, 0.6 to 4.8 Hz; n = 23). Periodic (1 min on, 1 min off) microelectrophoretic application of NMDA (20 nA) produced a robust and reproducible enhancement of hilar mossy cell firing rate (Fig. 1A). Compared with control ejections of saline, NMDA activation of hilar mossy cell firing rate was significantly reduced by baclofen (89%; n = 18; P < .001), ethanol (94%; n = 13; P < .001), or muscimol (81%; n = 5; P < .001; Fig. 1B). Moreover, NMDA activation of hilar mossy cells was also reduced by the NMDA antagonist APV (98%; n = 8; P < .001), low-dose baclofen (73%; n = 3; P < .001), low-dose ethanol (75%; n = 3; P < .001), or 1.2 g/kg i.p. ethanol (75%; n = 6; P < .001). Microelectrophoretic application of CGP35348 alone did not significantly affect NMDA activation of hilar mossy cell firing rate (n = 12; P > .05) but blocked baclofen (n = 11; P > .05) or ethanol (n = 8; P > .05), but not muscimol (n = 3; P < .001), reduction of NMDA activation of hilar mossy cell firing rate (Fig. 1B).

View larger version (32K): [in this window] [in a new window]   Fig. 1.   GABAB receptor antagonists reduce ethanol inhibition of NMDA-activated hilar mossy cells in vivo. A, ratemeter record demonstrates the marked activation of an hilar mossy cell by in situ microelectrophoretic application of NMDA (20 nA). Microelectrophoretic application of APV (25 nA), baclofen (+50 nA), or ethanol (+100 nA) suppressed NMDA activation of this hilar mossy cell with subsequent block by the GABAB antagonist CGP35348 (100 nA). The symbols , #, and  indicate APV, baclofen, and ethanol effects on NMDA activation of this hilar mossy cell, respectively, and are for convenience in comparing differences between their effects before and after CGP35348. B, effects of baclofen, ethanol, and muscimol on NMDA activation of hilar mossy cells and the effects of CGP35348 on the inhibition produced by these agents (compared with saline control). Although CGP35348 had no effect alone, it significantly reduced both baclofen and ethanol, but not muscimol, depression of NMDA activation of hilar mossy cells. *P < .001.

View larger version (40K): [in this window] [in a new window]   Fig. 2.   GABAB receptor antagonists reduce ethanol inhibition of NMDA-activated hilar interneurons in vivo. A, ratemeter record demonstrates the marked activation of an hilar interneuron by in situ microelectrophoretic application of NMDA (20 nA). Although microelectrophoretic saline (+200 nA) had no effect on NMDA activation, microelectrophoretic baclofen (+50 nA), ethanol (+100 nA), or muscimol (+50 nA) markedly decreased NMDA activation of this hilar interneuron. In situ application of the GABAB antagonist CGP35348 (100 nA) blocked the inhibitory effects of baclofen and ethanol but had no effect on muscimol inhibition of NMDA activation. The symbols #, , and  indicate baclofen, ethanol, and muscimol effects on NMDA activation of this hilar interneuron, respectively, and are for convenience in comparing differences between their effects before and after CGP35348. B, effects of baclofen, ethanol, and muscimol on NMDA activation of hilar interneurons as well as the effects of CGP35348 on the inhibition produced by these agents (compared with saline control). Although CGP35348 had no effect alone, it significantly reduced both baclofen and ethanol, but not muscimol, depression of NMDA activation of hilar interneurons. *P < .001.

GABA_B Receptor Antagonists Reduce Ethanol Inhibition of NMDA Activation of VTA GABA Neurons In Vivo. VTA GABA neurons were phasic firing neurons with spike durations of less than 500 µs and relatively elevated firing rates (mean, 23.1 ± 2.1 Hz). Periodic (30-50 s on, 60-90 s off) microelectrophoretic application of NMDA (15 nA) produced a robust and reproducible enhancement of VTA GABA neuron firing rate (Fig. 3A). Compared with control ejections of saline, NMDA activation of VTA GABA firing rate was significantly reduced by in situ microelectrophoretic application of baclofen (79%; n = 19; P < .001) or ethanol (73%; n = 12; P < .001; Fig. 3B). Moreover, NMDA activation of VTA GABA neurons was also reduced by APV (93%; n = 6; P < .001), low-dose baclofen (48%; n = 21; P < .001), low-dose ethanol (54%; n = 19; P < .001), or i.p. ethanol (45%; n = 3; P < .001). Microelectrophoretic application of CGP35348 did not significantly affect NMDA activation of VTA GABA neuron firing rate (n = 8; P > .05) but significantly reduced baclofen (n = 6; P > .05) or ethanol (n = 6; P > .05) reduction of NMDA activation of VTA GABA neuron firing rate (Fig. 3B).

View larger version (34K): [in this window] [in a new window]   Fig. 3.   GABAB receptor antagonists reduce ethanol inhibition of NMDA-activated VTA GABA neurons in vivo. A, ratemeter record demonstrates the marked activation of a VTA GABA neuron by in situ microelectrophoretic application of NMDA (15 nA). Microelectrophoretic application of the GABAB agonist baclofen (+50-100 nA) or ethanol (+100-200 nA) suppressed NMDA activation of this VTA GABA neuron with subsequent block by the GABAB antagonist CGP35348 (100 nA). The symbols # and  indicate baclofen and ethanol effects on NMDA activation of this VTA GABA neuron, respectively, and are for convenience in comparing differences between their effects before and after CGP35348. B, effects of baclofen and ethanol on NMDA activation of VTA GABA neurons and the effects of CGP35348 on the inhibition produced by these agents (compared with saline control). Although CGP35348 had no effect alone, it significantly reduced both baclofen and ethanol depression of NMDA activation of VTA GABA neurons. *P < .01.

Potentiation of GABA_B Receptor-Mediated Inhibition of VTA GABA Neurons by Ethanol. Given the relatively stable firing of VTA GABA neurons, we studied the effects of microelectrophoretic baclofen and muscimol during continuous administration of microelectrophoretic ethanol (Fig. 4). Figure 4C summarizes the effects of baclofen and muscimol alone and in combination with ethanol. Both baclofen and muscimol significantly reduced VTA GABA neuron firing compared with saline (61 and 75%, respectively, P < .001; n = 3 each). During continuous ethanol administration, baclofen inhibition was significantly enhanced compared with baclofen alone (P < .5; n = 3); however, muscimol inhibition was significantly reduced compared with muscimol alone (P < .05; n = 3).

View larger version (21K): [in this window] [in a new window]   Fig. 4.   Ethanol potentiation of GABAB receptor-mediated inhibition of VTA GABA neurons in vivo. A and B, ratemeter records show the effects of low-dose microelectrophoretic baclofen, muscimol, and ethanol on the firing rate of a VTA GABA neuron. A, ratemeter record shows the inhibitory effects of low-dose microelectrophoretic baclofen (+25 nA) and muscimol (+25 nA) on the firing rate of this VTA GABA neuron. The inhibition produced by either drug was approximately 50%. B, ratemeter record is a continuation of the record in A and shows the inhibitory effects of continuous ethanol on the firing rate of this VTA GABA neuron and its effects on baclofen and muscimol inhibition of its firing rate. Although the inhibition by muscimol was moderately decreased, the inhibition of VTA GABA neuron firing by baclofen was clearly potentiated during the ethanol inhibition. C, effects of microelectrophoretic baclofen, muscimol, and ethanol on the firing rate of VTA GABA neurons. Baclofen, muscimol, or ethanol significantly reduced the firing rate of VTA GABA neurons compared with microelectrophoretic saline. *P < .001. The effects of baclofen and muscimol during continuous ethanol were compared with an adjusted ethanol baseline (i.e., ethanol control). The baclofen-induced inhibition of VTA GABA neuron firing was significantly greater during continuous ethanol administration than without ethanol (#P < .05 compared with baclofen alone), whereas the muscimol-induced inhibition was significantly less than when administered alone ( P < .05 compared with muscimol alone).

GABA_B Receptor Antagonists Have Mixed Effects on Ethanol Inhibition of NMDA Activation of NAcc Neurons In Vivo. The majority of amygdala-driven NAcc core neurons had very low or no spontaneous activity. The degree of activation of NAcc neurons by microelectrophoretic NMDA was less than that produced in hilar mossy cells, hilar interneurons, or VTA neurons (e.g., mean peak firing rate response, 9.8 ± 0.8 Hz versus 103.5 ± 5.1 Hz for hilar mossy cells), and depolarization block was the rule in most NAcc neurons if the duration of activation exceeded several seconds or if the iontophoretic current exceeded 20 nA.

View larger version (33K): [in this window] [in a new window]   Fig. 5.   Mixed effects of GABAB receptor antagonists on ethanol inhibition of NMDA-activated NAcc neurons in vivo. A, in situ microelectrophoretic application of NMDA activated this NAcc neuron; however, as in most NAcc core neurons, depolarization block occurred promptly after ejection, even at low current strengths (4 nA). Therefore, NMDA activation of NAcc neurons seldom exceeded 10 Hz. Microelectrophoretic application of ethanol (+200 nA) or baclofen (+100 nA) inhibited NMDA activation of this NAcc neuron; however, CGP35348 failed to antagonize baclofen and had inhibitory actions of its own. The symbol # marks baclofen effects on NMDA activation of this NAcc core neuron and is for convenience in comparing differences between its effects before and after CGP35348. B, effects of baclofen and ethanol on NMDA activation of NAcc core neurons and the effects of CGP35348 on the inhibition produced by these agents (compared with saline control). Baclofen and ethanol markedly inhibited, whereas CGP35348 moderately inhibited, the firing rate of NAcc core neurons compared with saline electrophoresis. The effects of baclofen and ethanol during CGP35348 were not significantly different from CGP35348 alone. *P < .001.

GABA_B Receptor Antagonists Reduce Ethanol Inhibition of NMDA Receptor-Sensitive Amygdala-Driven NAcc Core Neurons In Vivo. Because of the lack of NAcc core neuron spontaneous activity, their profound sensitivity to microelectrophoretic NMDA (i.e., brisk depolarization block), the significant inhibitory effects of CGP35348 alone on NMDA-activated NAcc neurons, and the mixed effects of CGP35348 on the ability of ethanol or baclofen to reduce NMDA activation of NAcc neurons, we evaluated the role of GABA_B receptors in amygdala-driven NAcc neurons. The discharge probability of NAcc neurons by stimulation of the amygdala was a function of stimulus intensity. Amygdala-driven NAcc spikes were studied at three different stimulus levels and expressed as percent occurrence: threshold, half-maximum, and maximum. Seven NAcc neurons demonstrated a minimum 80% reduction (from maximum) in spike occurrence after in situ microelectrophoretic application of APV (50 nA), and by this criterion, their discharge was considered to be a function of NMDA receptor activation. Both baclofen and ethanol markedly reduced the occurrence of NMDA receptor-sensitive amygdala-driven spikes across stimulus levels (Fig. 6A; P < .001; n = 7). Although CGP35348 had no significant effects (P > .05; n = 7), it blocked the inhibitory effects of ethanol and baclofen on amygdala-driven NAcc core neuron spike elicitation (Fig. 6B).

View larger version (17K): [in this window] [in a new window]   Fig. 6.   GABAB antagonists reduce ethanol inhibition of NMDA receptor-sensitive amygdala-driven NAcc neurons in vivo. Insets, representative NAcc core neuron driven by stimulation of the basolateral nucleus of the amygdala (BNA) before and after microelectrophoretic application of baclofen (+50 nA). A, occurrence of amygdala-driven NAcc spikes was studied at three stimulus levels: near threshold (50% firing), maximum (100% firing), and an intermediate level between threshold and maximum. In situ microelectrophoretic application of APV (50 nA), baclofen (+50 nA), and ethanol (+100 nA) markedly decreased the probability of firing (percent occurrence of spike) across stimulus levels. B, microelectrophoretic application of CGP35348 (100 nA) blocked ethanol and baclofen inhibition of amygdala-driven spikes. *P < .001.

GABA_B Receptor Antagonists Reduce Ethanol Inhibition of NMDA EPSPs in NAcc In Vitro. We performed in vitro studies in NAcc slices to evaluate the mechanisms underlying GABA_B receptor-mediated ethanol effects on NMDA responses in vivo. We recorded from a total of 35 neurons within the NAcc core at depths within the slice of 50 to 350 µm. These neurons had large resting membrane potentials averaging 84.8 ± 0.6 mV (range, 67 to 93 mV; n = 35) and current-evoked spikes averaging 118 ± 1.7 mV (n = 35). Local stimulation near the recording pipette evoked multicomponent EPSPs in NAcc core neurons. To investigate the interaction between GABA_B and NMDA receptors, we pharmacologically isolated NMDA EPSPs by superfusion of 10 µM CNQX and 30 µM bicuculline. Figure 7 shows such NMDA EPSPs in a representative NAcc core neuron evoked at three different stimulus levels. The residual CNQX-resistant EPSP component was voltage-sensitive and blocked by superfusion of the NMDA receptor antagonist APV, suggesting mediation by NMDA receptors. As reported previously (Nie et al., 1993), superfusion of 66 mM ethanol significantly decreased NMDA EPSPs across all stimulus levels (Fig. 7A; at threshold F_2,10 = 34.491, P = .0001; at half-maximal F_2,10 = 23.277, P = .002; at maximal F_2,10 = 12.995, P = .0017). Superfusion of CGP55845 significantly prevented the ethanol inhibition of EPSPs (Fig. 7B; threshold stimulus intensity: F_1,5 = 0.3379, P = .586; half-maximum: F_1,5 = 1.997, P = .216; maximum: F_1,5 = 1.046, P = .3532 in comparing CGP55845 alone as control with CGP55845 plus ethanol).

View larger version (15K): [in this window] [in a new window]   Fig. 7.   GABAB antagonists reduce ethanol inhibition of NMDA EPSPs in the NAcc in vitro. Insets, representative recordings elicited in an NAcc core neuron by local stimulation (arrow marks stimulus artifact) of submerged NAcc slices in the presence of the non-NMDA receptor antagonist CNQX (10 µM) and the GABAA antagonist bicuculline (30 µM) eliciting NMDA EPSPs, whose amplitude was a function of stimulus intensity (shown here at suprathreshold, 50% maximal, and maximal subspike EPSP). Superfusion of ethanol (66 mM) reduced NMDA EPSPs in this neuron by 40% with nearly complete recovery on washout. The GABAB antagonist CGP55845 (4 µM) alone slightly decreased NMDA EPSPs in this neuron but blocked ethanol inhibition of NMDA EPSPs. Subsequent superfusion of the competitive NMDA receptor antagonist APV (60 µM) markedly reduced NMDA EPSPs in this neuron, verifying involvement of NMDA receptors. Calibration bar is the same for all traces. A, effects of ethanol on NMDA EPSPs in the NAcc in vitro. Ethanol (66 mM) significantly reduced mean NAcc NMDA EPSP amplitudes by approximately 40% at all stimulus levels tested, with partial recovery on washout with aCSF. B, CGP55845 (4 µM) did not significantly affect NMDA EPSPs after washout but blocked ethanol inhibition of NMDA EPSPs. *P < .001.

GABA_B Receptor Antagonists Do Not Alter Ethanol Inhibition of NMDA-Induced Currents in NAcc In Vitro. To further evaluate whether the block of the ethanol reduction of NMDA EPSPs by CGP55845 occurred presynaptically or postsynaptically, we superfused slices with TTX to minimize the presynaptic action of NMDA. In the presence of 1 µM TTX, 10 µM CNQX, and 30 µM bicuculline, under voltage-clamp we evoked NMDA currents by local pressure application of NMDA (200 µM in the pipette; 2- to 4-s duration; 2-8 psi). Superfusion of 66 mM ethanol decreased NMDA currents with full recovery on washout. Ethanol-induced depression usually occurred within 1 to 3 min after ethanol reached the slice chamber, with a peak ethanol effect occurring at 4 to 8 min and recovery to control levels on washout for 6 to 15 min. Ethanol (66 mM) significantly depressed NMDA currents 35% (Fig. 8; F_2,12 = 15.925, P = .0008) but 4 µM CGP55845 did not significantly alter ethanol inhibition of NMDA currents (F_1,5 = 12.982, P = .0155 in comparing CGP55845 alone as control with CGP55845 plus ethanol).

View larger version (25K): [in this window] [in a new window]   Fig. 8.   Lack of effect of GABAB antagonists on ethanol inhibition of NMDA-evoked currents in the NAcc in vitro. Insets, inward currents (in voltage-clamp) elicited by pressure application of NMDA (200 µM in the pipette; 2 s; 6 psi) near the NAcc core neuron in the presence of TTX (1 µM), CNQX (10 µM), and bicuculline (30 µM). Holding potential, 60 mV; resting membrane potential, 87 mV. Recording pipette contained 3 M KCl. Input conductance was monitored by periodic voltage steps shown as brief negative deflections. Ethanol (66 mM) significantly reduced mean NAcc NMDA-induced currents. Superfusion of CGP55845 (4 µM) slightly decreased NMDA-evoked currents but did not block ethanol reduction of NMDA currents. *P < .001.

GABA_B Receptor Antagonists Do Not Alter Ethanol Inhibition of Non-NMDA EPSPs in NAcc In Vitro. We also examined the effects of CGP55845 on ethanol inhibition of non-NMDA EPSPs in NAcc slices. Using local stimulation, non-NMDA EPSPs were pharmacologically isolated from NMDA EPSPs by application of 30 µM APV and 30 µM bicuculline to the NAcc slice. In seven cells studied, superfusion of 66 mM ethanol significantly reduced the amplitude of non-NMDA EPSPs across all stimulus levels tested (Fig. 9A; threshold: F_2,14 = 5.035, P = .0225; half-maximum: F_2,14 = 9.560, P = .0024; maximum: F_2,14 = 6.741, P = .0089) compared with controls, with recovery on washout. Superfusion of 4 µM CGP55845 slightly decreased non-NMDA EPSPs but did not appear to affect ethanol inhibition of non-NMDA EPSPs (Fig. 9B; threshold stimulus intensity: F_1,7 = 1.729, P = .586; half-maximum: F_1,7 = 10.262, P = .0149; maximum: F_1,7 = 8.896; P = .0204 in comparing CGP55845 alone as control with CGP55845 plus ethanol).

View larger version (17K): [in this window] [in a new window]   Fig. 9.   Lack of effect of GABAB antagonists on ethanol inhibition of non-NMDA EPSPs in the NAcc in vitro. Non-NMDA EPSPs were obtained by local stimulation of the NAcc slice in the presence of bicuculline (30 µM) and APV (30 µM). A, ethanol significantly reduced NAcc non-NMDA EPSPs at all stimulus intensities (e.g., maximum: *P < .007; n = 6) with partial recovery on washout with aCSF. B, CGP55845 (4 µM) did not significantly affect non-NMDA EPSPs at higher stimulus intensities but reduced ethanol reduction of non-NMDA EPSPs at the lowest stimulus intensity.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

Activation of hilar mossy cells, hilar interneurons, VTA GABA neurons, and NAcc neurons by periodic application of NMDA in vivo was brisk and robust. Similar excitatory effects have been observed in these areas and others (Yang et al., 1996) with NMDA and other excitatory amino acids (Hu and White, 1996). In situ microelectrophoretic application of the GABA_B receptor agonist baclofen markedly depressed NMDA-activated hilar mossy cell, hilar interneuron, VTA GABA neuron, and NAcc neuron firing, suggesting that activation of postsynaptic GABA_B receptors inhibits NMDA receptor-mediated excitation or that activation of presynaptic GABA_B receptors inhibits NMDA receptor-mediated release of glutamate. Microelectrophoretic application of ethanol had similar inhibitory effects on these neurons, and microelectrophoretic application of the GABA_B antagonist CGP35348 consistently blocked both baclofen and ethanol reduction of NMDA-activated hilar mossy cell, hilar interneuron, and VTA GABA neuron firing, suggesting that ethanol enhances GABA_B receptor-mediated inhibition of NMDA receptor-mediated neurotransmission. However, CGP35348 had a slight inhibitory effect on NMDA activation on its own, suggesting some nonselective actions. This became especially evident in the NAcc, where CGP35348 often had pronounced inhibitory effects on NMDA activation of NAcc core neuron firing rate. Notwithstanding the differences, both baclofen and ethanol inhibition of amygdala-driven NMDA receptor-sensitive NAcc spikes was blocked by CGP35348, indicating that ethanol inhibition of NMDA neurotransmission was clearly mediated by GABA_B receptors in the NAcc as well as the hippocampus and VTA.

NAcc neurons were considerably more sensitive to depolarization block by NMDA than were hilar mossy cells, hilar interneurons, or VTA GABA neurons. The sensitivity of NAcc neurons to depolarization block by glutamate agonists is well known (Hu and White, 1996). One may speculate that NMDA receptor-mediated excitation is more pronounced in NAcc neurons at rest due to regionally specific NMDA receptor subunit compositions that are less sensitive to the voltage-dependent Mg²⁺ block of the NMDA receptor channel. The differential sensitivity of hippocampal, VTA, and NAcc neurons to NMDA-induced depolarization block, and perhaps to baclofen or ethanol inhibition of NMDA activation, may also be a function of how much the spontaneous firing rate is dependent on afferent input. For example, despite the fact that the firing rate of both VTA GABA neurons (Steffensen et al., 1998) and NAcc core neurons (Pennartz et al., 1991) is highly dependent on afferent excitatory synaptic input, VTA GABA neurons are relatively fast firing and NAcc neurons are mostly silent. Therefore, comparisons regarding the relative sensitivity of these neurons to ethanol or ethanol inhibition of NMDA activation of their firing have to be made in light of their intrinsic excitability and synaptic input.

Muscimol-induced reduction of NMDA activation of hilar mossy cells and hilar interneurons neurons was unaffected by CGP35348, demonstrating, at least in the hippocampus, the selectivity of CGP35348 to GABA_B receptor-mediated responses. A role for GABA_B receptors in ethanol actions on neuronal excitability is further supported by our findings demonstrating that baclofen, but not muscimol, inhibition of VTA GABA neuron firing rate was potentiated by ethanol. Surprisingly, muscimol reduction of VTA GABA neuron firing rate was actually reduced by ethanol. This result is in contrast to the many reports demonstrating ethanol potentiation of GABA_A responses. The subunit composition of the GABA_A complex appears to be important for alcohol effects on this receptor. Molecular studies have revealed a complex heterogeneity in the structure and pharmacology of GABA_A receptors. At least five different subunit families (, , , , and ) have been isolated, and several lines of evidence suggest that the potentiating action of ethanol requires the 2L subunit (Wafford et al., 1991; Harris et al., 1995). It has been demonstrated that regions with high densities of binding sites for the nonbenzodiazepine sedative zolpidem had a greater relative abundance of the 2L splice variant (Duncan et al., 1995) and that in these brain regions, ethanol was capable of enhanced responses to GABA (Criswell et al., 1995). Conversely, the density of zolpidem binding was significantly lower or absent in brain regions, such as the VTA, where ethanol did not enhance GABA_A transmission (Criswell et al., 1995). This may explain why ethanol did not potentiate muscimol inhibition of VTA GABA neurons; however, it does not explain the reduction of muscimol inhibition by ethanol. We can only speculate that the reduction in muscimol inhibition results from indirect GABA_A receptor-mediated inhibitory effects on VTA dopamine neurons that influence the activity of VTA GABA neurons.

We performed in vitro studies in the NAcc to examine the mechanisms underlying GABA_B receptor involvement in ethanol reduction of NMDA responses in vivo. Ethanol inhibition of NAcc NMDA EPSPs, but not NMDA-evoked currents, was blocked by superfusion of GABA_B antagonists, suggesting that GABA_B receptor-mediated inhibition of NMDA receptor-mediated neurotransmission was presynaptic. Because EPSPs evoked at normal resting potentials are likely to be primarily generated by non-NMDA glutamate receptors and because low ethanol concentrations (11-22 mM; equivalent to blood levels of 50-100 mg/100 ml) reduced these EPSPs (see also Nie et al., 1993) but did not alter responses to exogenous CNQX-sensitive, non-NMDA glutamate agonists, it seems likely that ethanol can act presynaptically to reduce EPSPs, perhaps by reducing glutamate release. This possibility is consistent with biochemical studies showing ethanol reduction of the release of several neurotransmitters, including glutamate (Martin and Swartzwelder, 1992). Of particular relevance, ethanol prevents NMDA-induced glutamate release in the striatum (Carboni et al., 1993).

The effect of ethanol on EPSPs in NAcc core neurons also involves postsynaptic sites (Nie et al., 1994). This possibility is supported by electrophysiological studies of hippocampal neurons in culture and slice preparations demonstrating that ethanol selectively inhibits NMDA-induced ionic currents (Lovinger et al., 1989, 1990). In addition, in isolated sensory neurons from adult rats, 50 mM ethanol significantly inhibited NMDA-activated currents but did not alter GABA-activated currents (White et al., 1990). We have demonstrated here and in previous reports (Nie et al., 1992, 1993, 1994) that low-dose ethanol reduces NMDA-activated currents. However, GABA_B antagonists did not alter ethanol reduction of postsynaptic NMDA responses, suggesting that GABA_B receptor modulation of NMDA receptor-mediated neurotransmission is mediated presynaptically.

The release of glutamate could be under independent control by presynaptic NMDA and non-NMDA receptors. Regulation of glutamate release by excitatory amino acid presynaptic receptors in the hippocampus and striatum has been shown to be a function of degree, with inhibition occurring under tonic conditions and strong potentiation occurring during exogenous administration of NMDA (Liu and Moghaddam, 1995) but not kainate (Chittajallu et al., 1996). Moreover, NMDA, but not non-NMDA, receptor-mediated EPSP facilitation is potentiated in the NAcc (Pennartz et al., 1991) and hippocampus (Muller and Lynch, 1988). A presynaptic explanation for the block of ethanol inhibition of NMDA responses, but not non-NMDA responses, is reasonable if ethanol increases GABA release onto glutamate terminals that contain GABA_B receptors that are negatively coupled to a positive feedback presynaptic NMDA, but not non-NMDA, receptor-mediated release of glutamate. Alternatively, stimulation may activate two different glutamatergic afferents: one consisting of terminals that contain both presynaptic NMDA and GABA_B receptors and one consisting of terminals that contain just presynaptic non-NMDA receptors.

Activation of GABA_B receptors has been shown to inhibit adenylyl cyclase activity (for a review, see Kuriyama et al., 1993), leading to reduced PKA activity. Therefore, antagonism of tonically activated presynaptic GABA_B receptors should then lead to enhanced cAMP levels and protein kinase A activity. This is evidenced by cerebellar data indicating that adenylyl cyclase and protein kinase A activation is necessary for ethanol potentiation of GABA responses (Freund and Palmer, 1996). GABA_B receptor activation has also been shown to reduce protein kinase C activation (Tremblay et al., 1995), postulated to be required for ethanol potentiation of IPSCs in hippocampal slices (Weiner et al., 1994). The fact that ethanol was without effect on pharmacologically isolated late hippocampal GABA_B IPSPs weakens the possibility that ethanol acts to release GABA (Morrisett and Swartzwelder, 1993; Wan et al., 1996). However, ethanol enhancement of hippocampal IPSCs (Weiner et al., 1994) could be obtained under conditions where late IPSCs were not apparent and PKC activity was enhanced. Our laboratory reported that ethanol enhancement of hippocampal GABA_Aergic IPSPs occurred only after blockade of GABA_B receptors (Wan et al., 1996), suggesting a complex interaction between ethanol and the two GABA receptor subtypes. Often, positive findings regarding ethanol actions on GABA neurotransmission depend on these and other conditions, including brain region, species, or GABA_A subunit compositions.

The inconsistency in evidence supporting a role for GABA synaptic transmission in ethanol effects may be contingent on the coupling of GABA_B receptors to second messenger systems, whether presynaptic, postsynaptic, or both. Activation of GABA_B receptors on glutamate terminals coactivated with GABA terminals may occur independently or in the absence of postsynaptic GABA effects on GABA_A or GABA_B receptors. Figure 10 schematizes potential sites for ethanol effects on excitatory synaptic transmission based on our findings and others. Central to this model is the hypothesis that ethanol increases the excitability of coactivated GABA neurons that inhibit excitatory synaptic transmission via GABA_B receptors located on glutamate terminals. The regulation of glutamate release that occurs with pathway stimulation is a net effect of glutamate autoreceptor activation by NMDA and non-NMDA receptor types as well as by GABA-mediated inhibition elicited by coincident stimulation of GABAergic pathways or by GABA interneurons activated by glutamate collaterals. This arrangement might explain why GABA_B receptor block attenuates ethanol reduction of NMDA EPSPs but not non-NMDA EPSPs. In this model, GABA_B receptors would be located on glutamate terminals containing both NMDA and non-NMDA receptors wherein only the NMDA receptors are inhibited by GABA_B activation through costimulated GABA release.

View larger version (32K): [in this window] [in a new window]   Fig. 10.   Schematic of proposed loci for ethanol effects on GABAB receptor regulation of glutamatergic synapses. Stimulation of both glutamatergic (GLU) and GABAergic fibers releases GLU onto target neurons as well as GABA onto GLU terminals to target neurons. Activation of presynaptic NMDA receptors on GLU terminals by exogenous administration of NMDA could increase GLU release perhaps by activation of second messenger systems involving protein kinase C (PKC) or protein kinase A (PKA). GABAB receptor activation by coactivated GABA release inhibits NMDA receptor-mediated GLU release. Block of ethanol inhibition of NMDA responses by GABAB receptor antagonists suggests that ethanol may increase release of GABA onto glutamate terminals containing GABAB receptors. Ethanol also has a postsynaptic effect on both NMDA and non-NMDA [-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid/kainate (A/K)] receptors that does not appear to be regulated by GABAB receptors.