Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

The neuronal mechanisms underlying the central actions of acute ethanol exposure are not understood completely, although numerous ethanol effects have been investigated (Bloom and Siggins, 1987; Siggins et al., 1987a; Harris and Allan, 1989; Siggins et al., 1990; Harris et al., 1995b). In the central nervous system (CNS), much attention has focused on the effects of ethanol on synaptic transmission (Berry and Pentreath, 1980; Siggins et al., 1987a; Shefner, 1990; Nie et al., 1993, 1994); many studies suggest that the synapse is the neuronal site most sensitive to ethanol. It is generally agreed that glutamate and -aminobutyric acid (GABA) are the principal neurotransmitters mediating, respectively, excitatory and inhibitory synaptic transmission in the brain. Ethanol has been found to reduce the activity of the glutamatergic system (Lovinger et al., 1989, 1990; Lovinger, 1993; Nie et al., 1993, Nie et al., 1993, 1994; Martin et al., 1995) and, in some CNS regions, to enhance the activity of GABAergic systems (Celentano et al., 1988; Deitrich et al., 1989; Aguayo and Pancetti, 1994; Mehta and Ticku, 1994); these effects probably contribute to ethanol-elicited neuronal depression in the CNS.

GABA acts on at least two classes of receptors: GABA_A and GABA_B receptors. Molecular studies have revealed a complex heterogeneity in the structure and pharmacology of GABA_A receptors, because at least five different subunit families (, , , , and ) have been isolated (Schofield et al., 1987; Pritchett et al., 1989; Shivers et al., 1989). The differences in subunit composition have important functional implications for the pharmacology of GABA_A receptors and the action of ethanol because ethanol enhancement of GABA_A receptor activation has been controversial (Mancillas et al., 1986; Siggins et al., 1987b; Celentano et al., 1988; Deitrich et al., 1989; White et al., 1990; Proctor et al., 1992; Aguayo and Pancetti, 1994). Often, positive findings depend on satisfaction of certain conditions, such as the activation of protein kinase C (PKC; Weiner et al., 1994; Harris et al., 1995a; Macdonald, 1995) or the study of different brain areas, neuron types or regions (Proctor et al., 1992; Soldo et al., 1994), species, or GABA_A subunit compositions (Wafford and Whiting, 1992; Harris et al., 1995c). Recently, our laboratory reported that ethanol enhancement of GABA_Aergic inhibitory postsynaptic potentials (IPSPs) in hippocampal pyramidal neurons only occurred after blockade of GABA_B receptors (Wan et al., 1996), suggesting a complex interaction between ethanol and the two GABA receptor subtypes.

The neuronal effects of glutamate are brought about by two receptor classes: ionotropic receptors, which are ligand-gated channels passing cationic currents, and metabotropic receptors, which are coupled to transduction systems via G proteins (Watkins et al., 1991; Pin and Duvoisin, 1995). Interestingly, low glutamate concentrations can enhance GABA responses in hippocampus (Stelzer and Wong, 1989), and GABA_A receptors can mediate excitatory as well as inhibitory synaptic events in central neurons, depending on the presence of endogenous glutamate at appropriate levels (Michelson and Wong, 1991). These data suggest that important interactions exist between glutamate and GABA receptors in the CNS.

Behavioral studies indicate that the nucleus accumbens (NAcc) is a brain region involved in the rewarding effects of ethanol (Koob and Bloom, 1988). Both glutamate and GABA are major neurotransmitters in this area. Glutamatergic afferents from cortex, subiculum, and amygdala terminate on the most abundant cell type in the NAcc: medium spiny neurons. These neurons are most likely GABAergic (Bolam et al., 1983) and probably project to each other as well as to extrinsic sites. Recently, we reported that low ethanol concentrations significantly reduced glutamatergic synaptic transmission in rat NAcc neurons (Nie et al., 1993, 1994). To further clarify the effects of ethanol on amino acid transmitter systems in the NAcc, we now have examined GABA responses and their possible interactions with ethanol and glutamate in the core NAcc using intracellular recording in a slice preparation. We found that ethanol enhanced responses to exogenous GABA in about half the NAcc neurons studied, suggesting heterogeneity with respect to their GABA receptors. In the ethanol-sensitive neurons, glutamate and a metabotropic receptor agonist also enhanced the ethanol-induced potentiation of GABA responses. These and other results suggest that ethanol and glutamate interactions with GABA receptors may involve metabotropic receptors, perhaps operating through a second messenger linkage.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Slice Preparation. We prepared NAcc slices from male Sprague-Dawley rats (100-170 g) as previously described (Yuan et al., 1992). Briefly, we cut coronal slices 350- to 400-µm thick on a vibrating cutter (Vibroslice; Campden Instruments, Silbey, UK) and placed them in ice-cold (3-5°C) artificial cerebrospinal fluid (ACSF) gassed with carbogen (95% O₂, 5% CO₂) and of the following composition: NaCl, 130 mM; KCl, 3.5 mM; NaH₂PO₄, 1.25 mM; MgSO₄·7H₂O, 1.5 mM; CaCl₂·2H₂O, 2 mM; NaHCO₃, 24 mM; and glucose, 10 mM. We immediately transferred the slices to a recording chamber where they were incubated in an interface configuration for 30 min with their upper surfaces exposed to warmed, humidified carbogen. The slices were then completely submerged and continuously superfused with ACSF at a constant rate (2-4 ml/min) for the remainder of the experiment. The inner chamber had a total volume of ~0.5 ml; at the superfusion rates used, 90% replacement of the chamber solution could be obtained within 1 to 2 min (Siggins et al., 1987b). During testing, we maintained the bath temperature constant at 31-35°C.

Electrophysiology. We filled intracellular glass micropipettes with 3 M KCl (tip resistance, 60-110 M), performed single-electrode voltage-clamp studies with an Axoclamp 2A preamplifier (Axon Instruments Inc., Foster City, CA), and, in discontinuous voltage-clamp recording, continuously monitored the electrode settling time and capacitance neutralization on a separate oscilloscope. We took neuronal recordings from the NAcc core region at levels 2.2 to 0.7 mm from bregma and surrounding, but ventromedial to, the anterior commissure, using a "blind" approach. We accepted all successfully impaled [as judged by long-term stability of resting membrane potentials (RMPs) > 60 mV] neurons into the study sample. We applied GABA (1 or 5 mM in the pipette) locally near the recorded neuron by pressure (pipette tip diameter, 2.5-4 µm; pressure, 3-10 psi; duration, 2-4 s). GABA currents were always elicited in the presence of tetrodotoxin (TTX; 1 µM) to minimize presynaptic effects. In some cells, we also superfused 6-cyano-7-nitroquinoxaline-2,3-dione (CNQX) or DL-2-amino-5-phosphonovalereate (D-APV) to block non-N-methyl-D-aspartate (NMDA) or NMDA receptors. In most neurons, the GABA currents were highly reproducible for up to 90 min using GABA pressure application at intervals of 2 to 3 min (Fig. 1A). The cells were held near RMPs (approximately 83 mV); thus, with Cl-containing recording pipettes, GABA currents were inward (depolarizing) in direction. After stable responses were achieved, we took various current and voltage measurements at several time points before, during, and after ethanol superfusion. Continuous d.c. recordings were stored on polygraph paper, and selected records were digitized, stored, and analyzed on an 80486 computer using the pClamp programs (Axon Instruments).

View larger version (27K): [in this window] [in a new window]   Fig. 1.   GABA-induced inward currents in voltage-clamped NAcc core neurons are stable over time and blocked by the GABAA receptor antagonist bicuculline. A, pressure (2 s, 6 psi) application of GABA (5 mM in the pipette) elicits inward currents in the presence of 1 µM TTX. The GABA currents were reproducible for up to 90 min if the ejection intervals were ~2 to 3 min. Recording pipettes contained 3 M KCl. Holding potential: 84 mV = RMP. In this and subsequent figures, smaller inward deflections are responses to 10 mV hyperpolarizing steps to monitor input conductance. B, GABA currents in another NAcc core neuron in the presence of TTX as in part A. In control conditions, GABA applied by pressure (5 mM in the pipette, 2 s, 8 psi) near the recorded neuron elicits inward currents. Bicuculline (Bic; 30 µM) superfusion for 5 min almost totally blocks the GABA currents, with recovery on washout, suggesting mediation by GABAA receptors. Recording pipette filled with 3 M KCl. Holding potential: 90 mV = RMP. Under these conditions, the small residual inward current during bicuculline applications is probably due to incomplete block of GABAA receptors; we have not seen evidence of postsynaptic GABAB receptors in NAcc core neurons (Berton et al., 1998). In this and subsequent figures, bars over current records indicate duration of GABA application.

Drug Treatment. We introduced ethanol and other drugs in known concentrations into the slice chamber by a multiple valve system, without disrupting the flow of the superfusate. To avoid loss of ethanol by evaporation, the solutions were diluted in gassed ACSF from sealed stock solutions of reagent grade 95% ethyl alcohol in water immediately before administration. The usual testing protocol was: recording of membrane currents with periodic local application of GABA for 10 to 15 min during superfusion of ACSF alone ("control"); followed by switching to ACSF with drug [e.g., ethanol, glutamate, CNQX, D-APV, -methyl-4-carboxyphenylglycine (MCPG)] while repeating GABA application for 5 to 20 min; then switching again to ACSF alone for 15 to 35 min with subsequent GABA current measures ("washout"). We defined ethanol potentiation of GABA responses as at least a 10% increase in peak GABA current. We obtained TTX from Calbiochem (La Jolla, CA), CNQX and MCPG from Tocris Cookson (Bristol, UK), D-APV from Research Biochemicals (Wayland, MA), ethanol from the Remet Corp. (La Mirada, CA), and all other drugs from Sigma Chemical Co. (St. Louis, MO).

Data Analysis. Data are expressed as mean ± S.E. For statistical analysis, we used ANOVA for repeated measures, followed by the Newman-Keuls post hoc test. We accepted P < .05 as statistically significant.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Neuronal Sample. We studied a total of 84 neurons within the NAcc core area (see Fig. 1 of Yuan et al., 1992) at depths within the slice of 40 to 380 µm. As reported previously, these neurons had large RMPs averaging 83 ± 1.1 mV (mean ± S.E.; range: 66 to 93 mV; n = 67), and current-evoked spikes averaging 115 mV (range: 100 to 125 mV; n = 67). In most cells, stable recordings could be maintained for up to 2 to 3 h.

Ethanol Effects on Membrane Properties. In general, superfusion of intoxicating concentrations of ethanol (11, 22, 44, 100, or 200 mM) had little reversible or reproducible effect on membrane potential or input slope resistance of NAcc neurons (Nie et al., 1993). Of 67 neurons, only two showed a small (2-3 mV) hyperpolarizing response, and three showed small (2-3 mV) depolarizing responses. In the remaining 62 cells, ethanol had no measurable effect on membrane potential. Therefore, analysis of ethanol effects on GABA responses could proceed without the confounding effects of direct potential changes.

Ethanol Effects on GABA Currents. GABA currents were evoked in the presence of 1 µM TTX to minimize presynaptic effects. In our conditions, exogenous GABA evoked reproducible inward currents (Fig. 1A) in NAcc core neurons that were nearly totally blocked by 30 µM bicuculline (Fig. 1B), indicating mediation by GABA_A receptors. The small residual currents that are sometimes seen are likely due to incomplete blockade of GABA_A receptors at these low bicuculline concentrations, rather than GABA_B receptor effects (see Fig. 1B legend). With superfusion of ethanol (11-200 mM), GABA currents were increased as follows: in 17% of tested cells with 11 mM ethanol, in 33% of cells with 22 mM ethanol, in 45% with 44 mM ethanol, in 50% of cells with 100 mM ethanol, and in 22% of cells with 200 mM ethanol (Table 1). In the remaining neurons, ethanol had no effect (Fig. 2) or slightly decreased GABA currents (<10%). In most cells, ethanol potentiation of GABA currents occurred within 3 to 14 min (usually 3-7 min, counting the 1-2 min "dead-time" of the superfusion tubing) and recovered to control levels on washout for 5 to 15 min, or even disappeared despite continuing superfusion of ethanol. Thus, in four cells, the potentiation occurred within 3 to 5 min, then disappeared, suggesting the development of rapid tolerance to ethanol in these NAcc neurons (Wan et al., 1996).

View larger version (31K): [in this window] [in a new window]   Fig. 2.   Ethanol enhances GABA currents in some NAcc neurons but not others. A, the ethanol-sensitive type of NAcc core neuron. Voltage-clamp recording in the presence of 1 µM TTX. In control conditions, pressure application (5 mM, 2 s, 6 psi) of GABA near the recorded neuron elicits inward currents. Ethanol (EtOH; 44 mM) superfusion for 12 min enhances the GABA currents to 135% of control, with recovery on washout (14 min). Recording pipette filled with 3 M KCl. Holding potential: 88 mV; RMP: 90 mV. B, the second type of NAcc core neuron (ethanol-insensitive). GABA currents recorded in the presence of 1 µM TTX. Control: pressure application of GABA (5 mM in the pipette, 2 s, 5 psi) near the recorded neuron induces inward currents. Ethanol (44 mM) superfusion for 6 and 12 min has little effect on the GABA currents. Holding potential: 78 mV = RMP. Recording pipette filled with 3 M KCl.

View larger version (22K): [in this window] [in a new window]   Fig. 3.   Concentration-response relationships: comparisons of effects of ethanol superfusion on GABA currents, averaged (±S.E.) between all cells tested and only those showing ethanol sensitivity. Ethanol 11 to 200 mM enhanced the GABA currents in a bell-shaped fashion. In both total cells and ethanol-sensitive cells, the 11 (only shown in "total cells"), 22, and 200 mM ethanol application did not significantly alter the GABA currents. However, across all cells, 44 and 100 mM significantly enhanced GABA currents (P = .004, P = .0015, respectively). In the cells showing ethanol sensitivity, 44 and 100 mM ethanol also significantly enhanced GABA currents (P = .0027, P = .0004, respectively). Statistics by ANOVA (see Results): *P < .001; **P < .01, compared with control.

Glutamate Enhances GABA Currents and Their Potentiation by Ethanol. To determine possible mechanisms underlying the variability of interactions between GABA and ethanol, we asked whether GABA responses could be altered by glutamate in NAcc neurons, as reported previously for hippocampal neurons (Stelzer and Wong, 1989). In the presence of 1 µM TTX, 20 µM glutamate enhanced (>10%) GABA responses in 6 of 11 cells studied to a mean of 119 ± 3.3% of control (Fig. 4A). In the remaining cells, glutamate had no measurable effect on GABA responses. At this low concentration, glutamate had no effect on RMPs in any tested neuron, suggesting that enhancement of GABA responses by glutamate occurs at concentrations below those required for its known excitatory (ionotropic) action.

View larger version (27K): [in this window] [in a new window]   Fig. 4.   Ethanol and glutamate interactions with GABA responses in the NAcc. A, representative current records from an NAcc neuron. In control ACSF, GABA application (5 mM in the pipette, 3 s, 8 psi) elicits inward currents. Glutamate (20 µM) enhances the peak GABA currents to 128% of control. B, in another neuron, ethanol (EtOH; 44 mM, 8 min) enhances the peak GABA currents to 132% of control. Superfusion of 20 µM glutamate (Glu; 10 min) together with 44 mM ethanol (24 min) further enhances the GABA currents to 159% of control. After washout for 20 min, 20 µM glutamate (8 min) alone, like ethanol, enhances the GABA currents to 117% of control, with some recovery on washout (Washout; 12 min). KCl-filled pipette. Holding potential: 84 mV = RMP. C, mean (±S.E.) GABA current amplitudes from 9 cells, treated as in part B. Whereas the increase in mean GABA currents by ethanol alone (117 ± 5.2% of control) was not statistically significant, that produced by ethanol plus glutamate (136 ± 9.2% of control) was significant (P = .01; F(1,8) = 11.3). Superfusion of glutamate alone enhanced GABA responses to 109 ± 4.7% of control (not significantly different from control).

View larger version (21K): [in this window] [in a new window]   Fig. 5.   Ionotropic glutamate receptor antagonists do not alter ethanol-GABA interactions. A, voltage-clamp recordings of an NAcc core neuron in the presence of 1 µM TTX. In control conditions, GABA pressure application (5 mM in the pipette, 2 s, 5 psi) elicits inward currents. Ethanol (EtOH; 44 mM) superfusion for 12 min enhances the GABA currents to 127% of control. After washout of ethanol and in the presence of 30 µM D-APV (30 min) to block NMDA receptors, GABA still elicits inward currents. In the presence of D-APV for 42 min, superfusion of ethanol (EtOH; 44 mM, 12 min) still enhances the GABA currents to 126% of control. Rec; ording pipette filled with 3 M KCl. Holding potential: 82 mV = RMP. B, current records of another NAcc neuron in the presence of 1 µM TTX. In control conditions, GABA (5 mM in the pipette, 2 s, 9 psi) pressure application elicits inward currents. Ethanol (44 mM) superfusion for 10 min enhances the GABA currents to 150% of control, with recovery on washout. In the presence of 10 µM CNQX for 30 min, to block AMPA/kainate receptors, GABA still induces inward currents, and superfusion of ethanol (44 mM, 10 min) together with CNQX (40 min) still enhances GABA currents to 144% of control, with recovery on washout. Recording pipette filled with 3 M KCl. Holding potential: 87 mV = RMP.

The Role of Metabotropic Receptors. Because of the lack of effect of D-APV and CNQX, we examined the possible involvement of metabotropic glutamate receptors in the effects of ethanol and glutamate on GABA currents. In eight NAcc cells, superfusion of the metabotropic glutamate receptor agonist trans-1-aminocyclopentane-1,3-dicarboxylic acid (trans-ACPD; 5 µM) had no effect on RMP, but enhanced GABA currents in four cells (Fig. 6). In testing two of these trans-ACPD-sensitive cells, subsequent superfusion of the metabotropic receptor antagonist MCPG (1 mM) blocked the trans-ACPD enhancement of GABA responses in both cells (Fig. 6). Furthermore, in all five ethanol-sensitive NAcc cells studied, superfusion of MCPG (1 mM) significantly blocked the ethanol enhancement of GABA currents (Fig. 7), suggesting that ethanol enhancement of GABA effects in NAcc could be regulated or mediated by postsynaptic metabotropic glutamate receptors.

View larger version (13K): [in this window] [in a new window]   Fig. 6.   The metabotropic glutamate receptor agonist trans-ACPD also enhances GABA currents in NAcc core neurons; GABA currents evoked in the presence of 10 µM CNQX and 30 µM D-APV in a NAcc neuron. Pressure application of GABA (5 mM in the pipette, 3 s, 8 psi) near the recorded neuron elicited inward currents. trans-ACPD 5 µM superfusion for 12 min enhanced the GABA currents to 125% of control, with recovery on washout (18 min). MCPG (1 mM) almost totally blocked the trans-ACPD enhancement of GABA currents in this neuron. Bars over current records indicate GABA application. Recording pipette filled with 3 M KCl. Holding potential: 85 mV = RMP.

View larger version (26K): [in this window] [in a new window]   Fig. 7.   MCPG blocks ethanol enhancement of GABA currents in NAcc. A, GABA currents evoked in a representative NAcc neuron in the presence of 10 µM CNQX, 30 µM D-APV, and 1 µM TTX. Superfusion of 44 mM ethanol for 12 min enhances the GABA currents. In the same neuron, superfusion of MCPG (1 mM) blocks ethanol enhancement of GABA currents (right panel). Recording pipette filled with 3 M KCl. Holding potential: 87 mV = RMP. B, pooled data from all five neurons studied with MCPG. Ethanol significantly (**P < .01) increased the mean GABA current amplitudes, and this effect was totally reversed by 1 mM MCPG (in the continued presence of ethanol) to less than control levels.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

In this study, we found that ethanol enhanced GABA responses in 17 to 50% of NAcc core neurons, but had no effect or slightly decreased (<10%) GABA responses in the remaining neurons (depending on concentration). These and other data suggest the existence of a subset of NAcc core neurons with ethanol-sensitive GABA_A receptors. In these ethanol-sensitive neurons, ethanol potentiation of GABA responses also could be enhanced further by glutamate. In the ethanol-insensitive neurons, glutamate plus ethanol had no effect on GABA responses. The NMDA and non-NMDA ionotropic glutamate receptor antagonists had no effect on either the percentage of cells showing ethanol enhancement of GABA responses or the degree of ethanol enhancement of GABA currents in these neurons. The metabotropic glutamate receptor agonist trans-ACPD mimicked glutamate and ethanol by enhancing GABA responses in half the NAcc neurons tested. Interestingly, the metabotropic glutamate receptor antagonist MCPG blocked the ethanol enhancement of GABA responses.

As noted in the Introduction, there have been many electrophysiological studies of ethanol effects on GABAergic systems. Although ethanol has been shown to enhance GABA_A receptor-activated events in neurons or isolated preparations of several CNS regions (see the Introduction), such an ethanol-GABAergic interaction (e.g., of GABA_A-mediated IPSPs or responses to exogenous GABA) in native neurons has not been universally observed (Siggins et al., 1987b; Osmanovic and Shefner, 1990; White et al., 1990; Frye et al., 1994). For example, in the hippocampus, there is still some disagreement about whether ethanol can affect the GABA_A receptor/chloride channel complex (see, e.g., Aguayo and Pancetti, 1994, using cultured mouse hippocampal neurons versus the in vivo and in vitro studies of rat hippocampus by Harrison et al., 1987; Siggins et al., 1987b; Morrisett et al., 1991; Proctor et al., 1992; Peoples and Weight, 1994). Our laboratory initially saw no specific effect of systemic ethanol on responses to iontophoretic GABA in rat hippocampal neurons in vivo (Mancillas et al., 1986) and little effect or an inhibitory action on evoked IPSPs or GABA-induced hyperpolarizations in rat hippocampal slices (Siggins et al., 1987b). However, our more recent hippocampal studies using pharmacological isolation of GABA_A-IPSPs suggest that ethanol can reproducibly enhance those IPSPs, but only when GABA_B receptors are blocked (Wan et al., 1996) and probably via a presynaptic mechanism (Siggins et al., 1999).

Our data showing ethanol enhancing GABA responses in some NAcc neurons are generally consistent with those of others studying different native neuronal preparations (Deitrich et al., 1989; Harris and Allan, 1989; Aguayo and Pancetti, 1994) and specific GABA_A receptor subunit compositions expressed in Xenopus oocytes (Wafford et al., 1991; Whitten et al., 1996). It is possible that the existence of ethanol-sensitive and -insensitive NAcc neurons reflect different GABA_A receptor subunit compositions, as with studies showing that the potentiating action of ethanol required the _2L subunit (Wafford et al., 1991; Wafford and Whiting, 1992). One possible explanation for the lack of ethanol-GABA interaction in some NAcc neurons, that rapid tolerance in the ethanol-insensitive cells masks a fleeting increase in GABA currents (see below), could also derive from differences in GABA_A receptor subunit compositions.

Another possible mechanism for the inconsistent actions of ethanol on GABA_A receptors is that ethanol or some regulator (e.g., a metabotropic link) alters the activity of a calcium- or phospholipid-dependent protein kinase (e.g., PKC), which in turn alters the sensitivity of the GABA receptor via phosphorylation. Several reports suggest that ethanol enhancement of GABA responses requires phosphorylation of a GABA_A receptor subunit by PKC (Harris et al., 1995a; Macdonald, 1995). It may be that only half the NAcc core neurons we tested (the ethanol-sensitive neurons) contained the PKC phosphorylation sites or the related metabotropic machinery necessary to render GABA_A receptors responsive to ethanol. Our current studies are using protein kinase inhibitors to test this hypothesis.

Based on RMP, magnitude of responses to locally applied GABA, and GABA current kinetics, we found evidence for two different types of neurons in the NAcc core. The GABA_A receptor antagonist bicuculline blocked GABA currents in both types, suggesting mediation by GABA_A receptors in both. Although we cannot eliminate other factors, based on the amplitude and duration of GABA currents, the amount of GABA and its concentration gradient reaching the neuron may be important factors in these differences. The differences might also indicate that the GABA currents in "ethanol-insensitive" neurons have an early, ethanol-sensitive, but rapidly desensitizing component that is unresolved under our conditions of GABA application by pipette in a slice preparation.

We also found that ethanol enhancement of GABA currents in some cells lasted only 4 to 6 min, despite continued ethanol superfusion, suggesting rapid tolerance for this effect in some NAcc neurons. This finding agrees with several reports (Durand et al., 1981; Gilliam, 1989; Khanna et al., 1990; Ghosh et al., 1991; Palmer et al., 1992) and also with data from our laboratory showing rapid tolerance to the ethanol enhancement of IPSPs in some hippocampal neurons (Wan et al., 1996). This tolerance could explain why ethanol effects were found in some but not all cells, and it contrasts with the ethanol reduction of NMDA-induced currents that shows no apparent acute tolerance in NAcc neurons (Nie et al., 1994).

The inverted U-shaped dose-response curve for ethanol potentiation of GABA responses in NAcc may seem unusual for ethanol effects. This ethanol-GABA receptor interaction is different from the more standard dose-response relationship seen for ethanol inhibition of NMDA currents in the same NAcc slice preparation, which shows no inverted-U shape but rather a standard asymptotic depression at all higher doses (Nie et al., 1994). This may be related to the relative lack of short-term tolerance seen in the ethanol-NMDA interactions (Nie et al., 1994), versus a more obvious short-term tolerance to ethanol-GABA receptor interactions in some NAcc neurons. Thus, the tolerance mechanism in NAcc neurons for GABA receptor-ethanol interactions may be different from that of NMDAR-ethanol interactions, possibly because of the involvement of metabotropic system(s) in the former. This might also suggest that the ethanol inhibition of some NAcc neurons via GABAergic mechanisms might be more fleeting than for the ethanol reduction of NMDA receptor-mediated glutamatergic transmission. The need for a metabotropic step may also explain the sometimes slow development of the ethanol effects seen in our in vitro slice paradigm and in related slice studies (Proctor et al., 1992; Soldo et al., 1994).

Electrophysiological data from Stelzer and Wong (1989) showed that glutamate can enhance GABA responses in central neurons. Recently, this laboratory also reported a suppression of GABA_A receptor responses by NMDA application (Chen and Wong, 1995), suggesting that NMDA receptors do not mediate glutamate's enhancement of GABA responses NAcc. Our results agree with these findings: neither D-APV nor CNQX blocked glutamate enhancement of GABA responses in NAcc core neurons. In the ethanol-sensitive NAcc neurons, glutamate plus ethanol further potentiated GABA responses, and this type of NAcc neuron was more sensitive not only to ethanol but also to glutamate. The reason for this correlation is not known, but it is possible that the factor(s) responsible for ethanol sensitivity, such as a specific GABA_A receptor subunit composition, also confer glutamate sensitivity.

The metabotropic glutamate receptor agonist trans-ACPD enhanced GABA responses in half the NAcc neurons tested, suggesting mediation of the glutamate effect by metabotropic glutamate receptors. We also found that the mGluR antagonist MCPG blocked the ethanol enhancement of GABA responses, further supporting the regulation of ethanol-GABA_A receptor interactions by metabotropic receptors, perhaps through endogenously released glutamate. It is possible that these interactions involve protein kinase A or C, an idea consistent with cerebellar data suggesting that adenylyl cyclase or PKA activation is necessary for ethanol potentiation of GABA responses (Freund and Palmer, 1996) and that ethanol potentiation of IPSPs in hippocampal slices may require GABA_A receptor subunit phosphorylation by PKC (Weiner et al., 1994). Further studies are needed to determine whether these metabotropic and kinase mechanisms also generalize to GABA_A receptors of other central neurons (Siggins et al., 1999).

As to the functional and behavioral significance of our findings, the reported role of the NAcc in alcohol and drug reinforcement and dependence may suggest that ethanol potentiation of inhibitory GABAergic function in some NAcc neurons could underlie some aspect of these phenomena. In addition, ethanol (Nie et al., 1993, 1994) and other reinforcing drugs (Yuan et al., 1992) inhibit excitatory glutamatergic synaptic transmission in accumbens neurons. These combined data suggest that animals may work to inhibit their accumbens neurons, and ethanol may facilitate such inhibition. Furthermore, data in this study suggest that metabotropic mechanisms may be involved in these ethanol actions. Because most NAcc neurons are themselves GABAergic and thus inhibitory, disinhibition of "downstream" neurons (for example, ventral pallidum, thalamus) could play a role in ethanol reinforcement or dependence.