Materials and Methods

Transverse hippocampal slices (400 μm) were prepared from 4- to 6-week-old HAS and LAS rats, and ILS and ISS mice using a Sorvall (Newtown, CT) tissue chopper. Submerged slices were incubated in a submersion chamber consisting of a grid of small, square compartments with plastic netting attached to the bottom, suspended in a 250-ml beaker and covered with a loose-fitting plastic lid. This chamber was maintained at a constant temperature of 31–33°C in aerated (95% O₂, 5% CO₂) artificial cerebrospinal fluid containing 126 mM NaCl, 3 mM KCl, 1.5 mM MgCl₂, 2.4 mM CaCl₂, 1.2 mM NaH₂PO₄, 11 mM glucose, and 26 mM NaHCO₃. Slices were left in this chamber for at least 90 min after the dissection. For recordings, slices were transferred to a submersion recording chamber maintained at a constant temperature of 31–33°C and superfused with aerated artificial cerebrospinal fluid at 2 ml/min. Slices were allowed to equilibrate in the recording chamber for a few minutes before electrophysiological recordings were begun.

GABA_A IPSCs were recorded from CA1 neurons using the whole-cell patch-clamp technique. Recording electrodes were constructed from borosilicate glass (1.5 mm o.d., 0.86 i.d.; Sutter Instrument Co., Novato, CA) and had resistances of 6 to 9 MΩ when filled with the patch pipette solution. The patch pipette solution contained 125 mM potassium-gluconate (Fluka, Buchs, Switzerland), 5 mM KCl, 10 mM HEPES (Fluka), 0.1 mM CaCl₂, 1 mM potassium-EGTA (Fluka), 2 mM MgCl₂, 2 mM magnesium-ATP, and 0.2 mM Tris-GTP (pH = 7.3 adjusted with KOH; 290 mOsm) and was kept on ice until immediately before use. Series resistances ranged from 10 to 41 MΩ (average 30 ± 1.5 MΩ). The average change in the series resistance for all cells from the beginning of the control to the end of the washout period was 15.6 ± 1.2%, and all cells in which the change was >25% were excluded from subsequent analysis.

All recordings were made in the presence of 20 μM 6,7-dinitroquinoxaline-2,3-dione (DNQX) and 50 μMdl-(−)-2-amino-5-phosphonovaleric acid (APV) to block excitatory postsynaptic currents. Synaptic stimulation was delivered using a bipolar twisted nichrome wire electrode (0.2-ms pulses of 7–30 V) positioned within 250 μm of the recording pipette and placed directly over stratum pyramidale, with an interstimulus interval of 30 to 60 s, as previously described (proximal stimulation; Weiner et al., 1997a). All cells were clamped to −65 mV (after correction for the liquid junction potential) and recorded in the voltage-clamp mode. After superfusion with DNQX and APV to block the glutamatergic components of the synaptic current, the strength of the stimulation pulse was adjusted so that the peak amplitude of the residual GABAergic component (GABA_A IPSC) was about 50 to 100 pA.

All drugs were purchased from Sigma (St. Louis, MO) unless otherwise indicated. Drugs applied to slices were made up as 100-fold concentrates and added to the superfusion buffer via calibrated syringe pumps (Razel Scientific Instruments, Stamford, CT). A 4 M solution of ethanol (Aaper, Shelbyville, KY; diluted in deionized water) was prepared immediately before each experiment from a 95% stock solution kept in a glass storage bottle.

Drug effects were quantified as the percentage of change in amplitude or area under the curve of IPSCs relative to the mean of control and washout values. Statistical analyses of drug effects were carried out using two-tailed student's paired and unpaired t tests, or two-way ANOVAs as indicated, with a level of significance ofP < .05.

Results

Recordings were made from CA1 pyramidal neurons in hippocampal slices from ILS and ISS mice, and from the replicate lines of HAS₁ and HAS₂ and LAS₁ and LAS₂ rats. In most cases, the responses evoked in the presence of DNQX + APV were mediated primarily via GABA_A receptors, but in some instances there was a small, late component on the falling phase of the IPSC (Fig. 1A) that could be blocked by superfusion with the GABA_B receptor antagonist CGP 35348 (data not shown). However, this secondary component did not overlap significantly with the peak of the GABA_A response, so it was not blocked pharmacologically in the instances where it was observed. To examine the ethanol sensitivity of GABA_A responses, slices were superfused with 80 mM ethanol, and changes in the amplitude and area of the IPSC response were determined. This concentration of ethanol was used because this is the mean blood ethanol concentration at regain of righting reflex in the heterogeneous stock of animals from which the HAS and LAS rat lines were originally selected (Draski et al., 1992). At generation 12 of selection, the LAS and HAS animals regained their righting reflexes with mean blood ethanol concentrations of 87 and 75 mM, respectively (Draski et al., 1992).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Effects of bath superfusion with 80 mM ethanol on GABA_A IPSCs evoked from selected lines of rats and mice. IPSCs were evoked by proximal stimulation in the presence of DNQX + APV in each of the six lines of animals tested. Each set of traces was obtained from a single pyramidal neuron from the indicated line of animals. Averaged responses were recorded under control conditions (C), in the presence of 80 mM ethanol (E), and after ethanol washout (W), and the control and washout responses were generally indistinguishable. Responses from HAS₁, HAS₂, and ILS mice all showed enhancement comparable with what we have previously observed in the Sprague-Dawley rats (Weiner et al., 1997a), whereas the ethanol-insensitive lines (LAS₁, LAS₂, and ISS) were all unaffected. A, illustrates one example of a response where there was a small GABA_B component to the evoked response, seen as a slowly decaying tail on the GABA_A response; this component could be blocked by the GABA_B antagonist CGP 35348 (data not shown). There was also considerable variability in the time course of the GABA_A component (note the different time scales in B and C versus the other records), but there did not appear to be any relationship between the time course and the ethanol sensitivity of the response.

Under these experimental conditions, bath superfusion with ethanol did not significantly change either the holding current or the input resistance of CA1 pyramidal neurons in any of the lines of animals tested (Table 1). However, 80 mM ethanol significantly enhanced the GABA_A IPSC in both of the lines of rats selected for behavioral sensitivity to ethanol (HAS₁, HAS₂; Fig. 1, A and C) compared with the baseline response amplitude (average of pre-ethanol and washout periods). This was observed both as an increase in the amplitude of the response, as well as an increase in the area under the curve (Table 1). The falling phases of the IPSCs with no GABA_B component were fitted to single exponential functions, and the time constants for the decay were compared for control and ethanol responses. Ethanol had no significant effect on the time constants in any of the individual lines tested, but across all the animals there was a significant 22 ± 7.9% (n= 23; P < .02) increase in the average time constant, suggesting that ethanol did prolong the IPSCs to a limited extent.

The time course of the effect of bath superfusion with ethanol on GABA_A responses in HAS animals was similar to what we have previously reported in Sprague-Dawley rats, i.e., it usually took 5 to 10 min of superfusion with ethanol to achieve the maximal enhancement of the IPSC (Fig.2A). There did not appear to be any short-term tolerance occurring during the superfusion with ethanol, and recovery to baseline was observed within 5 to 10 min of ethanol washout (Fig. 2). The enhancement of GABA_A IPSCs induced by ethanol was also repeatable, and in some slices, potentiation was observed with as many as three successive applications of ethanol with no apparent decrement in the ethanol enhancement (Fig. 2B). Unlike the HAS rats, GABA_A IPSCs elicited in hippocampal slices from the two independent strains of rats with low behavioral sensitivity to ethanol (LAS₁, LAS₂) showed no significant change in response to ethanol superfusion (Fig. 1, B and D). In addition to there being no effect on the peak amplitude of the IPSC, this concentration of ethanol had no effect on the area under the curve for the IPSC as well (Table1). When the changes in the peak IPSC induced by ethanol were compared between the pooled HAS and LAS animals, there was a highly significant difference between the ethanol-sensitive and ethanol-insensitive groups of rats (LAS: 107 ± 4.7%, n = 29; HAS: 141 ± 9.6%, n = 30; P < .01)

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

Time course and repeatability of ethanol responses. A, illustrates the time course of ethanol action on GABA_A IPSCs evoked from cells from a HAS₂ and a LAS₂ rat. The rate of onset and washout of the effect in the slice from the HAS₂ animal was comparable with what we have previously observed in Sprague-Dawley rats (Weiner et al., 1997a; Fig.3). As shown in these examples, there was no significant effect of ethanol on the LAS₂ IPSC, but a significant enhancement of the HAS₂ IPSC. B, shows the repeatability of the ethanol response in a single cell from a HAS₁ animal that was superfused repeatedly with the same concentration of ethanol. Although there was some decline in the absolute magnitude of the GABA_A response over time, the relative enhancement of the IPSC by ethanol was maintained throughout the experiment.

Because it is not known whether there are proximal and distal GABA_A responses in mice such as we have previously described in rats (Weiner et al., 1997a), all of the experiments in ILS and ISS mice were conducted with two independent stimulating electrodes positioned to activate each of these pathways selectively. IPSC responses evoked by stimulation in stratum radiatum (distal stimulation) were unaffected by 80 mM ethanol in both lines of mice. The amplitude of the IPSCs in ethanol were 109 ± 13% (n = 17) of control in ISS mice, and 95 ± 6.8% (n = 17) of control in ILS mice. As far as the proximal pathway was concerned, the ethanol-insensitive mouse strain (ISS) did not show any change either in the amplitude or the area of the hippocampal GABA_A IPSC during ethanol superfusion (Fig. 1F; Table 1), but IPSCs in the ethanol-sensitive ILS line were significantly enhanced relative to their baseline and washout values (Fig. 1E). When results from proximal stimulation experiments in the two lines of animals were compared with each other, there was a highly significant difference in the effects of ethanol on IPSC amplitude in the two groups (Fig. 3;P < .003).

• Download figure
• Open in new tab
• Download powerpoint

Figure 3

Effects of ethanol on peak IPSC amplitude in selected lines of mice and rats. The average effect of 80 mM ethanol is illustrated on the peak amplitude of the IPSC evoked by proximal stimulation for each of the six lines of animals tested. Although the magnitude of the enhancement in the ILS and HAS₁animals was somewhat less than in the HAS₂ rats, the effect of ethanol was statistically significant in each of the ethanol sensitive lines, whereas it was not in any of the lines selected for ethanol insensitivity (∗, indicates P < .05 compared with control/washout responses in the same line of animals, paired t test).

Discussion

Ethanol Effects in Selected Lines of Rodents.

Ethanol activates or inhibits the activity of a number of different ligand-gated ion channels, and a key issue in understanding the cellular mechanisms of ethanol action is to determine which of these effects underlie specific pharmacological actions of ethanol. One approach to making such determinations is to use animals developed using classical selective breeding approaches to segregate genes that are associated with specific behavioral phenotypes, such as ethanol sensitivity. Comparisons of ethanol responses in animals developed by this approach can then be used to identify the specific molecular targets of ethanol that are linked to the behavioral phenotype.

One model that has been used to explore the basis for ethanol sensitivity differences is the long sleep (LS) and short sleep (SS) mice, which were selected based on the duration of loss of righting reflex in response to ethanol (Heston et al., 1973). Martz et al. (1983) proposed based on behavioral experiments with LS and SS mice that the sensitivity of GABA responses to ethanol might be one factor responsible for genetic differences in ethanol sensitivity. Subsequent biochemical studies showed that ethanol potentiates muscimol-stimulated chloride flux in LS mice, but has no effect in preparations from SS mice (Allan and Harris, 1986; Harris and Allan, 1989). A similar potentiating effect of ethanol has been demonstrated in cortical and cerebellar microsacs, as well as membrane vesicles from whole brain in HAS rats, but not in LAS rats (Allan et al., 1988, 1991; Liu and Deitrich, 1998). However, the ethanol sensitivity of synaptically mediated GABA_A IPSPs has never been determined in these lines of animals. Because the ethanol enhancement of GABA_A receptor-mediated³⁶Cl⁻ flux is primarily observed at low concentrations of GABA, and is negligible at high GABA concentrations (Mihic et al., 1994), this suggests that ethanol might not affect synaptic responses, because of the high concentrations of synaptic GABA after release.

Nevertheless, the present studies demonstrated a direct relationship between the behavioral sensitivity to ethanol in selectively bred rodent strains, and the in vitro ethanol sensitivity of hippocampal GABAergic synapses. This constitutes the first evidence for differences in the sensitivity of GABA_A receptor-mediated synaptic potentials that correspond to the behavioral sensitivity of selected lines of animals. The present data suggest that there is a robust association between synaptic sensitivity to ethanol and behavioral sensitivity, based on the occurrence of such differences in six selected lines of animals from two different species. Using different lines of selectively bred rodents reduces the probability that the differences between lines of animals with high and low ethanol sensitivity are due to random fixation of genes associated with GABA_A receptor sensitivity. Although the differences in synaptic GABA_A receptor sensitivity are correlated with the behavioral phenotype, it seems unlikely that hippocampal sensitivity to ethanol is causally linked with the behavior because the hippocampus probably has little influence over the loss of the righting reflex. Nevertheless, the gene(s) that regulates the sensitivity of these hippocampal synapses to ethanol is likely to affect the sensitivity of other GABA_Areceptors in other brain regions that may play such a role.

In the present studies, there was essentially no effect of 80 mM ethanol on GABA_A responses in the ISS/LAS animals. This might only reflect a relative difference in sensitivity, in that the threshold for ethanol effects in these lines might be higher than 80 mM. However, because behavioral effects of ethanol (such as motor incoordination) are clearly apparent in all of these lines of animals at blood ethanol concentrations of 80 mM (Draski et al., 1992), the lack of any effect on synaptic potentials at this concentration demonstrates that there are GABA_A synapses where concentrations of ethanol sufficient to produce a loss of the righting reflex have little effect. The behavioral effects that are observed in these insensitive lines of animals must depend either on GABA_A receptors in other brain regions that have greater ethanol sensitivity, or on receptors for other neurotransmitters that are more sensitive to ethanol. Previous studies have estimated that approximately six to eight genes play a significant role in the heritable component of ethanol sensitivity (Dudek and Abbott, 1984), whereas quantitative trait loci studies have identified at least five quantitative trait loci that are associated with the sleep time phenotype (Bennett et al., 1994; Markel et al., 1997). Although some of these genes may be related to GABA_A receptors, it seems likely that others are associated with other neurotransmitter systems.

Ethanol Effects on Synaptic GABA_A Responses.

Studies of ethanol effects on GABA_A responses in nonselected lines of rodents have found considerable variability in ethanol sensitivity, even in studies of what would seem to be the same receptors in the same population of cells (Proctor et al., 1992a,b;Peoples and Weight, 1994; Wan et al., 1996; Weiner et al., 1997a;Peoples and Weight, 1999). This suggests that ethanol sensitivity must depend on a rather specific combination of factors. One such variable is the subpopulation of GABA_A receptors activated by synaptic stimulation. Previous work by Pearce and colleagues has demonstrated that there are populations of GABA_Asynapses on CA1 pyramidal neurons that differ in a variety of respects, including their kinetic properties, and sensitivity to pharmacological agents such as furosemide (Pearce, 1993). Our studies in Sprague-Dawley rats have shown that distal GABA_A-mediated IPSCs are less sensitive to ethanol action than are proximal IPSCs (Weiner et al., 1997a). This study demonstrated such differential ethanol sensitivity of GABA_A receptors even in single hippocampal pyramidal neurons, confirming that the differences in sensitivity were not merely due to unknown variables in the experimental approach. The present studies found that proximal GABA_A IPSCs are sensitive to ethanol in the HAS₁ and HAS₂ lines of rats, and in the ILS mice. Further support for the conclusion that somatically located GABA_A receptors are particularly sensitive to ethanol comes from local application experiments in which GABA was applied either directly to the somata of CA1 pyramidal cells, or to their dendrites in stratum radiatum. Ethanol enhanced GABA_A responses mediated by the somatic receptors, while having no effect on dendritic currents (W. R. Proctor, unpublished data). Similarly, Soldo et al. (1998) concluded based upon local application experiments that ethanol had a selective effect on somatic GABA_A receptors in cortical neurons, whereas dendritic receptors showed little or no sensitivity. These experiments support the conclusion that at least a part of the action of ethanol is postsynaptic, either on the receptor itself, or on other cellular constituents that can modulate GABA_A receptor sensitivity.

Identifying the factors that are responsible for the differences in the sensitivity of proximal and distal GABA_Areceptors to ethanol is clearly a critical issue, but one that will be difficult to resolve with electrophysiological approaches. One possibility is that these differences are due to varying subunit composition of the receptors, although expression studies to date have not established a clear subunit dependence in ethanol sensitivity. The furosemide sensitivity of proximal responses (Pearce et al., 1995) suggests that these receptors incorporate the α4 subunit of the GABA_A receptor because furosemide selectively antagonizes responses mediated by GABA_A receptors incorporating the α6, which is not found in hippocampus, and α4 subunits (Wafford et al., 1996). An association between the α4 subunit and ethanol sensitivity could also explain why ethanol appears to have no effect on ³⁶Cl⁻flux in hippocampus (Proctor et al., 1992a) because the α4 subunit is thought to be a relatively uncommon constituent of GABA_A receptors in this brain region (Wisden et al., 1992), and thus such receptors would make a relatively small contribution to overall³⁶Cl⁻ flux in this brain region. Another possibility is that there are differences in post-translational mechanisms such as phosphorylation, which might differ in different parts of the cell. Phosphorylation by PKC appears to positively modulate the interaction between ethanol and GABA_A receptors, and might play a role in regulating such sensitivity (Weiner et al., 1997b). Possibly related to this mechanism, blockade of GABA_B receptors in hippocampal slices sensitizes distal GABA_Aresponses to ethanol, and this has been proposed to be due to a GABA_B-mediated reduction of PKA and/or PKC activation (Wan et al., 1996). However, in our own studies we have found that a GABA_B antagonist affects the sensitivity of the distal but not the proximal pathway (W. Poelchen and T. V. Dunwiddie, unpublished data), so it seems unlikely that this mechanism could account for differences in the sensitivity of the proximal pathway reported here. Finally, differences in proteins that interact with GABA_A receptors, such as gephyrin (Sassoè-Pognetto et al., 1999), GABA_Areceptor-associated protein (Wang et al., 1999), and the receptor for activated C kinase (Brandon et al., 1999), might also account for differences in sensitivity.

In summary, the present experiments demonstrate that synaptically evoked GABA_A responses in the rat and mouse hippocampus are enhanced by superfusion with the same concentrations of ethanol that produce behavioral responses in vivo. Moreover, the sensitivity of these synapses is under the control of genes that can be selected for by classical genetic selection techniques. We hypothesize that these genes regulate the sensitivity of GABA_A receptors in the hippocampus, a brain region that is unlikely to be involved in the loss of righting reflex, as well as in other brain regions that are directly involved in the regulation of the ethanol-sensitive or -insensitive behavioral phenotype. Although the GABA_A receptor is unlikely to be the only receptor involved in determining behavioral sensitivity (DeFries et al., 1989), the fact that the ethanol sensitivity of these GABA_A synapses is reduced in three lines of independently selected “insensitive” animals suggests that GABA_A receptors must play a significant role in this behavior.