Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

Considerable experimental attention in research on the central effects of ethanol has been directed toward examining its modulatory effects on the GABA_A receptor complex. Early biochemical studies demonstrated an ethanol-induced potentiation of GABA-mediated ³⁶Cl flux in synaptoneurosomes and microsacs derived from rat and mouse cerebral cortex (Allan and Harris, 1986; Mehta and Ticku, 1988; Suzdak et al., 1986). This was largely corroborated by electrophysiological studies indicating that physiologically-relevant concentrations of ethanol potentiated GABA-mediated currents in cultured cerebral cortical and spinal cord neurons (Aguayo, 1990; Celentano et al., 1988; Nishio and Narahashi, 1990; Reynolds and Prasad, 1991). Other electrophysiological studies, however, failed to observe modulatory effects of ethanol on neuronal responses to GABA (Sigel et al., 1993; Siggins et al., 1987; White et al., 1990). One prevailing postulate to reconcile the diverse and apparently conflicting results related to sensitivity to modulation by ethanol takes into account the diversity of the known GABA_A receptor subunits, of which 17 have been uncovered and classified into five families [reviewed in (Macdonald and Olsen, 1994; McKernan and Whiting, 1996; Sieghart 1995; Tyndale et al., 1995; Yeh and Grigorenko, 1995)]. The subunits display discrete yet overlapping patterns of distribution in the brain, and recombinant GABA_A receptors of defined subunit combinations exhibit different yet predictable functional properties. These considerations, taken together, could account in theory for the observed modulation of GABA_A receptors by ethanol among different types of neurons and brain regions.

Several studies have attempted to define the subunit composition of GABA_A receptors sensitive to modulation by ethanol through the use of recombinant receptors assembled in a variety of expression systems. Thus, differences in sensitivity to ethanol have been found to depend on the  subunit isomer, because recombinant GABA_A receptors expressing either the 1 or 6 subunit differed in rates of desensitization in response to co-application of GABA and ethanol (Marszalec et al., 1994). Zolpidem, aside from its demonstrated high-affinity binding to Benzodiazepine type I receptors, also appears to be selective for receptors which are sensitive to potentiation by ethanol (Criswell et al., 1993).

The 2 subunit exists as either a long or a short alternatively spliced variant (2_L or 2_S) (Whiting et al., 1990). The 2_L splice variant contains in the intracellular loop between transmembrane domains TM3 and TM4 an additional 8-amino acid mini-exon that also harbors a consensus sequence for phosphorylation by PKC. In a series of studies using 1/2/2-containing recombinant GABA_A receptors expressed in Xenopus oocytes, a dependence on subunit specificity was demonstrated in that the ethanol-induced potentiation was observed only when the GABA_A receptor assembly included the 2_L subunit (Wafford et al., 1991; Wafford and Whiting, 1992). However, a lack of effect by ethanol on either 2_L- or 2_S-containing recombinant GABA_A has also been reported (Sigel et al., 1993), and the degree to which sensitivity to modulation by ethanol depends on the 2_L subunit remains an outstanding issue.

In this study we compared the effect of ethanol on whole-cell GABA-activated current responses under identical patch clamp electrophysiological recording conditions in two immortalized cell lines expressing endogenous functional GABA_A receptors, two cell lines stably transfected with 1/x/2 GABA_A receptor subunits and in cerebellar Purkinje cells either maintained in long-term primary culture or after acute dissociation from the neonatal rat cerebellum. We chose to examine cerebellar Purkinje cells because they express GABA_A receptor subunit profiles similar to those of the stably transfected cell lines examined (Laurie et al., 1992). We report that ethanol potentiated GABA-mediated current in Purkinje cells but had no effect on GABA responses recorded in any of the cell lines tested. Furthermore, an ethanol-induced potentiation of GABA responses in Purkinje cells could be observed at a time in development before the expression of the 2_L GABA_A receptor subunit mRNA.

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

Maintenance of cell lines in culture. The RINm5F, IMR-32, WSS-1 and PA3/X25 cell lines were each maintained under conditions according to published reports (Hadingham et al., 1992; Noble et al., 1993; Tyndale et al., 1994; Wong et al., 1992). The RINm5F cell line was maintained in RPMI medium, the IMR-32 cell line in -minimum essential medium, and the WSS-1 cell line in DMEM. These media were supplemented with 10% fetal bovine serum and fungizone/penicillin/streptomycin. The PA3 and X25 cell lines were cultured on 0.01% poly-D-lysine coated plates and maintained in DMEM supplemented with 5% fetal calf serum. After 2 days in culture, 1 µM dexamethasone was added to the PA3 and X25 cell cultures to induce GABA_A receptor expression and electrophysiological recordings were conducted 3 to 5 days thereafter.

Preparation of primary and acutely dissociated cerebellar cells. Primary cultures of cerebellar neurons were derived from cerebella of PD 0 Sprague-Dawley rats. Cerebella were removed, minced and then incubated in Ca⁺⁺-Mg⁺⁺-free phosphate-buffered saline containing 0.1% DNase and 0.25% trypsin for 15 min at 37°C. The tissue was then gently triturated in medium containing DMEM, 10% heat-inactivated fetal bovine serum, 2 mM glutamine and fungizone/penicillin/streptomycin. Dissociated cells in suspension were added at a density of 3 × 10⁴ cells/cm² to glass coverslips coated with 0.01% poly-D-lysine and placed in an incubator for 1 hr to allow the cells to adhere. Before the seeding of cells, silicone droplets were attached to the periphery of each cover slip. Once the cells have adhered, the coverslips were placed inverted in 35-mm culture dishes containing a bed of cerebellar glial cells. Then 48 hr after plating, the serum-containing medium was replaced with a DMEM-based serum-free growth medium supplemented with B-27 nutrient mixture. The cultures were maintained in humidified 95% oxygen/5% carbon dioxide at 37°C and replenished with serum-free growth medium every 2 days. On selected days after plating, cerebellar cultures were fixed and processed immunohistochemically according to previously established procedures (Cheun and Yeh, 1992) to reveal calbindin-like immunoreactive cells. In vivo, only Purkinje cells express calbindin-like immunoreactivity in the cerebellum (Sequier et al., 1990). Calbindin-like immunoreactivity has also been used to identify Purkinje cells in culture (Brorson et al., 1991). Thus, calbindin-like immunoreactivity and cell morphology was used to identify Purkinje cells in culture for electrophysiological recording (see fig. 5A).

Electrophysiology. Unless stated otherwise, whole-cell patch clamp recording of GABA-activated current responses were performed at room temperature in a bath solution containing (in mM): 140 NaCl, 5.4 KCl, 1.8 CaCl₂, 1 MgCl₂, 5 HEPES, (pH 7.4). Some experiments involving PA3 and X25 cells employed bath solution consisting of (in mM): 130 NaCl, 5 KCl, 2 CaCl₂, 1 MgCl₂, 5 HEPES, 11 D-glucose, 13 sucrose (pH 7.2) in accordance with a previously published protocol (Harris et al., 1995a). Recording pipets were fire-polished to an input resistance of 5 to 10 M. Seal formation and recordings were conducted using conventional whole-cell patch clamp recording procedures (Hamill et al., 1981) using an EPC-7 amplifier (Darmstadt, Germany). The recording solution contained (in mM): 140 KCl, 1.8 CaCl₂, 1 MgCl₂, 5 HEPES, 3 Mg⁺⁺-ATP, 0.1 leupeptin (pH 7.4). Leupeptin was included to inhibit proteolysis and Mg⁺⁺-ATP was included to prevent possible rundown of the GABA-activated response (Chen et al., 1990). Membrane currents were amplified and filtered through a 4-pole Bessel filter. The analog signals were also monitored throughout the experiments using a Gould Brush 260 chart-recorder. GABA-activated currents were filtered, digitized and analyzed off-line using a data acquisition and analysis program (DATAQ/DATANAL, JPM Programming). Data were analyzed as previously described (Yeh and Kolb, 1997). Ethanol was determined to have a potentiating effect when there was a 20% increase in the GABA-ethanol response amplitude over mean peak control response.

Preparation and delivery of drugs. GABA, bicuculline methiodide and ethanol were dissolved bath solution. Diazepam was first dissolved in DMSO and then serially diluted in bath solution so that the final concentration of DMSO was <0.01%. Drugs were loaded into separate barrels of a multi-barrel pipet assembly and delivered by brief pulses of regulated pressure (2 PSI). The drug pipet assembly was mounted onto a micromanipulator so that it could be navigated under visual control to within 10 µm of the cell under study. This method facilitated brief focal applications of multiple test substances. One of the drug pipets within the multibarrel pipet assembly was routinely filled with bath solution that was applied continuously between epochs of drug application to clear drugs from the immediate vicinity of the cell and to control for possible mechanical artifacts due to bulk flow.

RT-PCR-based profiling of GABA_A receptor subunit mRNAs. Candidate GABA_A receptor mRNA profiles were obtained from the different cell line cultures as well as from selected single cells by means of an RT-PCR-based protocol (Yeh et al., 1996). For cultures, RNA was extracted with TRI reagent (Molecular Research Center), solubilized in RNase-free water and cDNA was synthesized by the addition of AMV reverse transcriptase (10 U) (Seikagaku, Ijamsville, MD), 1X first strand buffer, 2.5 mM dNTPs (Promega, Madison, WI), 10 ng oligo (dT) (Genosys, Woodlands, TX), RNasin inhibitor (20 U) and 10 mM dithiothreitol, incubated for 1 hr at 42°C. For single cells, first-strand cDNA was synthesized as above except that the oligonucleotide contained an oligo-dT region and the T7 RNA polymerase promotor. This facilitated one round of aRNA amplification using the T7 RNA polymerase before amplication by PCR (Eberwine et al., 1992). PCR amplification of the reversed transcribed cDNA template derived from either cultures or single cells was then performed in a solution containing thermal buffer, 2.5 mM MgCl₂, 0.25 mM dNTPs, 2 U of Taq DNA polymerase (Promega), 50 pmol of primers, and 1 µl cDNA in a final volume of 20 µl. Samples were amplified for 30 or 40 cycles at 94°C for 45 sec, 60°C for 45 sec, 72°C for 1 min and a final extension was applied at 72°C for 7 min using a programmable thermocycler (Perkin-Elmer Cetus, Cupertino, CA). The amplified products were separated on a 2% agarose gel and visualized under UV illumination after staining with ethidium bromide. Whole-brain rat cDNA was used as positive control. Samples of reversed transcribed template to which no RNA was added served as negative control. Southern blot analysis, such as that illustrated in figure 6C, was performed as described previously (Yeh et al., 1996) using cDNA probes specific for the 2_S and 2_L GABA_A receptor subunits.

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

IMR-32

IMR-32 (human neuroblastoma) cells possess functional GABA_A receptors, as revealed by TBPS binding autoradiography and GABA-stimulated chloride flux assays (Anderson et al., 1993). Western blot analysis demonstrated the presence of at least the 3 subunit (Noble et al., 1993), but the full complement of GABA_A receptor subunits has yet to be characterized. In our study, we profiled the expression of GABA_A receptor subunit mRNAs in cultured IMR-32 cells. An example of such an expression profile is illustrated in figure 1A. In addition to mRNA encoding the 3 subunit, 1, 4, 1, 3, 2 and  subunit mRNAs were also detected. Thus, numerous GABA_A receptor subunit transcripts were found to be expressed by IMR-32 cells. However, pharmacological characterization of GABA-activated current response properties of individual IMR-32 cells suggests that not all IMR-32 cells express the full profile of subunits (D.W. Sapp and H.H. Yeh, manuscript in preparation). For example, individual IMR-32 cells could be shown to display GABA responses that were either sensitive or insensitive to potentiation by diazepam (fig. 2, A and B).

View larger version (47K): [in this window] [in a new window]   Fig. 1.   IMR-32 cells express functional GABAA receptors. A, Ethidium bromide-stained agarose gel showing PCR-amplified products generated in the presence of GABAA receptor subunit-specific primers and corresponding to the 1, 3, 4, 1, 3, 2 and  subunits. The cDNA template used for PCR was reverse transcribed from mRNA isolated from IMR-32 cells that had been grown to confluence in two 60-mm Petri dishes. -actin (A) was routinely included as a positive control. B, The penwriter record (upper panel) shows that the GABA-activated current response is markedly reduced in the presence of bicuculline (100 µM). Digitized and averaged traces corresponding to GABA responses before (control), during and after (recovery) exposure to bicuculline are superimposed to facilitate quantitative analysis; bicuculline reduced the GABA response by 91%. C, Peak GABA response current-voltage relationship obtained from an IMR-32 cell. The inset illustrates a set of superimposed current traces obtained at various holding potentials, as indicated, and this was used to determine the peak amplitude of the GABA response at a given holding potential.

View larger version (27K): [in this window] [in a new window]   Fig. 2.   Effect of ethanol on GABA-activated current response in IMR-32 cells. A, Continuous penwriter record (A1) and averaged digitized traces corresponding to periods before (control), during and after (recovery) application of ethanol (100 mM; A2) or diazepam (0.5 µM; A3). This cell was insensitive to ethanol but was sensitive to positive modulation by diazepam. B, Continuous penwriter record (B1) and averaged digitized traces from another IMR-32 cell that was insensitive to modulation by either ethanol (100 mM; B2) or diazepam (0.5 µM; B3). C, Summary of data from all IMR-32 cells (n = 31) in which interaction between GABA and ethanol (10-200 mM; open bars) as well as diazepam (shaded bars) were tested in the same cell. For this and all subsequent figures, the numbers included in the bar graphs reflect the number of cells tested in the corresponding experiments.

GABA (100 µM) delivered by regulated pressure induced whole-cell inward currents in IMR-32 cells held near their resting membrane potential. The GABA-activated current was blocked by 100 µM bicuculline methiodide (fig. 1B) and reversed at approximately 0 mV with equimolar [Cl]_in and [Cl]_out (fig. 1C). GABA was applied for variable lengths of time in order to derive an operationally defined range of response amplitudes that would be suitable for testing the effect of ethanol (fig. 1D, inset). With a 5-sec application of 100 µM GABA, the maximal response (I_max) typically ranged between 200 to 300 pA (274 ± 31 pA, mean ± S.E.M.; n = 42). Figure 1D also illustrates data obtained from an IMR-32 cell in which amplitudes of the peak current response to GABA were plotted as a function of duration of agonist application. A "duration-response" relationship could be readily established. For each cell tested, the duration of agonist application was thus adjusted so as to produce current responses of constant amplitude approximately 20% of the I_max in response to 100 µM GABA, and this was subsequently used to assess the modulatory effect of ethanol and diazepam on the same cell. As determined by data obtained from the 31 IMR-32 cells included in this study, the duration of GABA application ranged between 40 and 200 msec.

Ethanol at any concentration examined in this study (10-200 mM) had no effect on the GABA-mediated current responses in IMR cells (fig. 2C). In figure 2A1 (top panel), the continuous penwriter record shows GABA-activated current responses monitored on-line before, during and after exposure to ethanol (100 mM) and diazepam (0.5 µM). The averaged digitized traces (fig. 2, A1 and A2) indicate that, although exposure to ethanol did not affect the amplitude of the GABA-activated current response, exposure to diazepam resulted in a reversible 81% potentiation of the GABA response in the same cell. Overall, in IMR-32 cells (n = 10) that could be shown to be sensitive to diazepam, the benzodiazepine-induced potentiation averaged 59 ± 16%, although ethanol exerted no effect in any of these cells. IMR-32 cells displaying GABA responses that were insensitive to modulation by diazepam were also encountered in this study (fig. 2B). Ethanol also did not affect the GABA responses in such cells (n = 21). The data obtained from diazepam-sensitive and -insensitive cells were pooled and are graphically summarized in figure 2C.

RINm5F

RINm5F cells display GABA-activated current responses that are blocked by bicuculline and zinc but are insensitive to modulation by the benzodiazepines, suggesting the predominant expression of a GABA_A receptor that lacks a  subunit (Hales and Tyndale, 1994). Pressure application of 100 µM GABA (1 sec in duration) produced an I_max of 64 ± 9 pA (mean ± S.E.M.; n = 8), similar to those reported previously (Hales and Tyndale, 1994). To assess the effect of ethanol, the duration of GABA application was adjusted for each cell so as to produce an approximately half-maximal level of the GABA-activated current response. The protocol for testing the effect of ethanol or diazepam on RINm5F responses to GABA was identical to that described above for testing IMR-32 cells. Neither a low (10 mM) nor a higher (100 mM) concentration of ethanol modulated the GABA-activated currents (fig. 3). The GABA-activated current responses monitored in RINm5F cells were also insensitive to potentiation by diazepam (fig. 3).

View larger version (24K): [in this window] [in a new window]   Fig. 3.   Effect of ethanol and diazepam on GABA responses elicited in RINm5F cells. This bar graph summarizes data obtained from RINm5F cells that were tested for both diazepam-GABA and ethanol (10 or 100 mM)-GABA interactions. Error bars reflect the S.E.M.

PA-3, X-25 and WSS-1. The PA-3 and X-25 cells are derived from a mouse fibroblast cell line that had been stably transfected with the 1/1/2_L and 1/1 combinations of GABA_A receptor subunits, respectively (Hadingham et al., 1992). In both ³⁶Cl flux assays and electrophysiological studies (Harris et al., 1995a), GABA-mediated responses of PA3 cells, but not those of X-25 cells, were potentiated by 10 mM ethanol. In PA-3 cells maintained on a poly-L-lysine substrate, the ethanol effect was observed at low GABA concentrations (3 and 10 µM). In our study, the interaction between ethanol and GABA-activated current responses was examined in PA-3 cells and X-25 cells under similar conditions using GABA at a concentration of 10 µM (fig. 4A). As was observed in every PA-3 cell tested (n = 28), a robust and readily reversible potentiation of the GABA response could be elicited upon exposure to diazepam but not to ethanol. In contrast, X-25 cells (12 of 12) were neither sensitive to modulation by diazepam nor ethanol. Ethanol tested at a range from low to high concentrations (10, 30 and 100 mM) yielded the same results (fig. 4A).

View larger version (51K): [in this window] [in a new window]   Fig. 4.   Effect of ethanol and diazepam on GABA-activated current responses in cell lines expressing recombinant 122S (WSS1 cells), 112L (PA3 cells) and 11 (X25 cells) GABAA receptors. A, Data obtained from each cell tested were normalized such that the averaged amplitudes of GABA (10 µM) responses recorded during exposure to ethanol (10, 30, 100 mM) or diazepam (0.5 µM) were expressed as a percentage change from those obtained during control (pre-ethanol or -diazepam exposure) periods. Inset, Examples of averaged, superimposed current traces taken from a PA3 cell (left panel) and an X25 cell (right panel) showing positive modulation by and no effect of diazepam, respectively. B, RT-PCR profiling of 2L and 2S mRNA expression in single IMR-32, PA3, X25 and WSS-1 cells. Each cell illustrated had been recorded electrophysiologically, and the data were incorporated into the summary graphs shown in figure 2C (IMR-32 cell) and figure 4A (WSS-1, PA3, X25 cells).

Cerebellar Purkinje cells

Cultured Purkinje cells. The outcome of an interaction between ethanol and GABA (25 µM) was examined in cultured Purkinje cells, identified based on size and morphology of calbindin-immunopositive profiles in long-term primary cultures derived from postnatal day 0 rat cerebellum (fig. 5A). The electrophysiological recordings were performed between 2 and 10 days in culture. In sharp contrast to the consistent lack of an effect in cell lines, ethanol (25 mM) potentiated the amplitude of the GABA-activated current responses elicited in cultured Purkinje cells. An ethanol-induced potentiation (34 ± 7%, mean ± S.E.M.; range = 20-42%) was seen in 10 of 14 cases (71%). In figure 5B, data derived from the 14 cells tested are summarized in the form of a scattergram and an example taken from an individual Purkinje cell is given in the inset.

View larger version (59K): [in this window] [in a new window]   Fig. 5.   Ethanol-induced potentiation of GABA-activated current response in cultured cerebellar Purkinje cells. A, Calbindin-like immunoreactivity in cerebellar neurons in vivo (A1) and maintained 4 days in a culture derived from PD 0 rat (A2). Calbindin-like immunoreactive profiles (solid arrow), presumably Purkinje cells, display a large soma and dendritic branching that is more elaborate than the smaller cells (open arrow) displaying background level of staining. Bars = 50 µm. B, "Scattergram" summarizing data derived from 14 cultured Purkinje cells tested for ethanol-GABA interaction. Each data point represents the mean peak amplitude of the GABA (25 µM) response during exposure to 25 mM ethanol plotted as a function of that monitored during the control (pre-ethanol exposure) period. The shaded area on either side of the 45° "equivalence" line (Yeh and Kolb, 1997) predicts no significant change due to ethanol exposure using 20% as an operationally defined level of significance for ethanol potentiation of GABA-mediated currents. Of the 14 cells, data points derived from 10 cells lie above this area, indicating potentiation of the GABA response in the presence of ethanol.

Acutely dissociated Purkinje cells. The ethanol-induced potentiation of GABA responses observed under long-term culture conditions was also found in Purkinje neurons acutely dissociated from neonatal cerebella (fig. 6A). Data from a total of 18 Purkinje cells, acutely dissociated from PD-3 (n = 4), PD-5 (n = 6), PD-7 (n = 6) and PD-10 (n = 2) rat cerebella, were included for analysis in this study (fig. 6B). The degree of potentiation induced by ethanol was highly variable and, given the limited sampling size for each postnatal age, it was not possible to attribute this variability to the age of the cerebellum from which the acutely dissociated Purkinje cells were derived. Acute exposure to ethanol resulted in a 20% increase in the GABA (10 µM) response amplitude in 83% (15/18) of the cells tested, reflecting a 39.0 ± 5.2% (mean ± S.E.M.) augmentation of the control GABA response.

View larger version (25K): [in this window] [in a new window]   Fig. 6.   Ethanol potentiates GABA-activated current responses in acutely dissociated rat cerebellar Purkinje cells. A, Continuous penwriter record (top) and digitized current traces (bottom) illustrate current transients in response to GABA (10 µM) monitored corresponding to periods before (control), during and after (recovery) exposure to 50 mM ethanol. The 200-msec pressure pulses of GABA were applied every 30 sec. The bar above the penwriter record indicates an 18-sec period of continuous application of ethanol. During the control and recovery periods, bath solution was applied from a separate drug pipet in lieu of ethanol. B, Agarose gel and corresponding Southern blot of PCR-amplified products derived from individual electrophysiologically-studied Purkinje cells indicate that the 2L subunit mRNA is not detected on PD-5 but is present by PD-7 although the 2S subunit transcript can be detected on both PD-5 and PD-7.

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

This study addresses two fundamental issues related to the relationship between GABA_A receptor subunit combination and sensitivity to modulation by ethanol. The first issue relates to the discrepant results in the literature on ethanol-recombinant GABA_A receptor interactions that have often been attributed to differences in experimental techniques and methods of analysis. To this end, ethanol-GABA_A receptor interaction was examined under identical recording conditions in several cell lines and in cerebellar Purkinje cells. The second issue relates to whether there is a specific requirement of the 2_L subunit in conferring ethanol sensitivity to native GABA_A receptors, as has been reported for recombinant GABA_A receptors, and this was examined in primary rat cerebellar cell cultures and in acutely-dissociated Purkinje cells.

Subunit combination alone does not determine sensitivity to GABA_A receptor modulation by ethanol. Taking the view that subunit composition is a critical determinant in the neuroactive nature of ethanol on the native GABA_A receptor, sensitivity to modulation by ethanol can be expected to vary, depending on the GABA_A receptor isoform. Given the diversity of GABA_A receptor subunits that can assemble to form functionally distinct receptor complexes, an intuitively logical approach toward resolving the issue of whether sensitivity to ethanol may depend on subunit composition has been to examine recombinant GABA_A receptors of defined heterologous subunit combinations in expression systems. However, although several studies have reported subunit-specific dependence of the ethanol effect (Wafford et al., 1991), others have not (Marszalec et al., 1994; Korpi et al., 1995), and these discrepant outcomes have been commonly attributed to differences in the electrophysiological methods employed by individual laboratories. In this study, identical patch clamp recording conditions were employed to compare the effect of ethanol on GABA-mediated current responses in immortalized cells, stably transfected cells and rat cerebellar Purkinje cells either maintained in primary culture or isolated after acute dissociation. The PA-3 and WSS-1 cell lines express 1/1/2 and 1/2/2 recombinant GABA_A receptors, respectively, similar to that expected to be expressed in cerebellar Purkinje cells because they express a limited set of 1/2/3/2 transcripts (Hadingham et al., 1992). Our results clearly indicate that ethanol had no effect on GABA-mediated currents elicited in either the PA-3 or WSS-1 cell line, but potentiated GABA responses elicited in Purkinje cells. A systematic analysis of sensitivity to ethanol under varying concentrations of GABA was not performed in this study. Nonetheless, aside from this consideration, our results rule out differences in other electrophysiological recording conditions as confounding factors. Thus, we are led to conclude that subunit combination alone cannot account for sensitivity to modulation by ethanol.

Ethanol modulation of native GABA_A receptor function is independent of the expression of the 2_L subunit. In this study ethanol-induced potentiation of GABA-mediated responses was manifest in acutely dissociated immature Purkinje cells prior to the detection of 2_L mRNA expression, indicating a lack of stringency for the expression of this subunit and the development of sensitivity to GABA_A receptor modulation by ethanol. Consistent with this finding, ethanol potentiated GABA-mediated responses in cultured Purkinje cells as early as 2 days in vitro, well before the onset of 2_L expression (H.H. Yeh, unpublished observation).