INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

GABA_A receptor channels are responsible for inhibitory synaptic transmission in the majority of neurons in the CNS and have considerable heterogeneity of constituting subunits (Olsen and MacDonald 2002). The pivotal role that GABA_A receptor subunits play in brain functions has been demonstrated by the large array of behavioral effects caused by deletions of individual constituting subunits (Mohler 2002). The deletion of the 3 subunit produces developmental and neurological impairments including premature death, epileptic convulsion, and Angelman syndrome-like features (Homanics et al. 1997). In contrast, deletion of the 2 subunit has little behavioral consequence in spite of the dramatic decrease (~50%) in GABA_A receptor expression in these mice (Sur et al. 2001). How can one reconciliate these contrasting results? Compensatory alteration of subunit expression, assembly, or targeting may underlie some of these differences. However, no compensatory upregulation of other  subunit isoforms has been observed in 3 / mice (Homanics et al. 1997). An intriguing alternative possibility is that particular subunits need to be associated to be expressed at the neuronal membrane surface. Indeed a pivotal role for  subunit in surface expression and localization has been demonstrated (Connolly et al. 1996, Connor et al. 1998).

At inhibitory synapses, the duration of GABA action is reflected in the time course of the inhibitory postsynaptic currents (IPSCs). The IPSCs time course determines synaptic strength and is the target of the action of several commonly prescribed tranquilizer, sedative, and antiepileptic drugs (MacDonald and Olsen 1994). Ultrarapid agonist applications studies with recombinant receptors revealed that the presence of specific  subunits determines the time course of GABA responses. Specifically, receptor isoforms containing the 1 subunit have a faster time course than those with 2 or 3 subunits. (Gingrich et al. 1995; Lavoie et al. 1997; McClellan and Twyman 1999; Verdoon 1994). Studies in neurons of transgenic mice lacking the 1 subunit (Vicini et al. 2001) as well as in neurons over expressing this subunit (Okada et al. 2000) indicate that the presence of the 1 subunit can indeed shorten IPSCs duration. The presence of the 1 subunit can be confirmed by prolongation of sIPSCs with zolpidem, a selective imidazopyridine for benzodiazepine type 1 receptors that contain this subunit (MacDonald and Olsen 1994, Pritchett et al. 1989).

Electrophysiological studies from neurons derived from 3 / mice have demonstrated a dramatic reduction of maximal GABA evoked whole cell currents in acutely dissociated DRG neurons and to a lesser extent in cultured hippocampal neurons (Homanics et al. 1997, Krasowski et al. 1998). In addition, smaller and faster sIPSCs were recorded both in neurons of the reticular thalamic nucleus (Huntsman et al. 1999) and granule neurons in the olfactory bulb (Nusser et al. 2001) in slices from 3 / mice.

We investigated the properties of sIPSCs and the action of zolpidem on their duration in cultured cortical neurons from 3 / mice. Our electrophysiological results, supported by immunocytochemical studies, suggest that the observed shortening of sIPSC duration in 3 / mice is the result of an altered expression of  subunits.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

All studies were conducted with an approved protocol from Georgetown University and University of Pittsburgh Animal Care and Use Committees in compliance with the National Institutes of Health guidelines for the care and use of experimental animals.

Mutant mice production

3-subunit-deficient mice were produced at the University of Pittsburgh and Georgetown University as described (Homanics et al. 1997). Matings between 3 ± mice (>F10 generation; mixed C57BL/6J × Strain 129Sv/SvJ genetic background) were set up in the late afternoon and checked for vaginal plugs the next morning. The day of plug detection was designated embryonic day 0.5 (E0.5).

Embryos (E16.5-17.5) from heterozygous intercrosses were genotyped by PCR amplification of genomic DNA extracted from liver using the hot sodium hydroxide and Tris (HotSHOT) protocol (32) (Truett et al. 2000). Briefly, embryonic liver samples were lysed with 75 µl of an alkaline lysis reagent (25 mM NaOH, 0.2 mM disodium EDTA; pH 12) at 95°C for 40 min. Subsequently, samples were cooled to 4°C for 30 min, and 75 µl of a neutralizing reagent (40 mM Tris-HCl; pH 5) was added. One to 5 µl of the final preparation was used per 12 µl PCR reaction. The amplification mixture contained genomic DNA, 2.0 mM MgCl₂, 0.2 mM dNTP, PCR buffer, 0.02 U/µl Ampli. Taq Polymerase (Perkin Elmer No. 808-0161) 0.8 µM oIMR 372 primer (primer for 3 exon 3: 5'-GCATCGACATGGTTTCTGAAGTC-3'), 0.8 µM oIMR 374 primer (primer for NEO cassette: 5'-CAGAAAGCGAAGGAACAAAGCTG), and 1 µM oIMR 373 primer (common 3 reverse primer: 5'-GGGCTACTGATCTCCTCTTTCCAC-3'). The PCR conditions were according to the Jackson Laboratory's genotyping protocol. The assay used a "touchdown cycling" to reduce nonspecific amplification products. For the first 12 cycles, the annealing temperature (starting at 64°C) was reduced by 0.5°C/cycle. For the remaining 25 cycles, the annealing temperature remained at 58°C. PCR amplification from 3 +/+ mice produced a 690-bp product, 3 ± produced 690- and 490-bp products, and 3 / produced a 490-bp product and had a pink eye phenotype.

Primary cortical neuronal cultures.

Cortical neuronal cultures were prepared from individual embryonic mice (E17-18) from both genotypes. Briefly, the cortex were chopped after careful dissection and digested in 0.28% trypsin (Sigma, St Louis, MO) for 15 min at 37°C with gentle shaking. Dissociated cells were inoculated at a density of 0.75 × 10⁶ in a 35-mm dish on poly-L-lysine-coated coverslips in basal Eagle medium (BME, Invitrogen, Carlsbad, CA) containing 10% FBS, 2 mM glutamine, 100 µg/ml gentamicin (all from Invitrogen), and 25 mM KCl, and maintained at 37°C in 5% CO₂. After 24 h in vitro, the medium was replaced with 50:50 mixture of BME and Neurobasal medium containing 2% B27 supplement, 1% antibiotic, and 0.25% glutamine (Invitrogen). At 5 days in vitro (DIV5), cytosine arabinofuranoside was added at final concentration of 10 µM. Thereafter, half of the medium was replaced twice a week with Neurobasal medium containing 2% B27 supplement, 1% antibiotic, and 0.25% glutamine.

CORTICAL NEURON TRANSFECTION. We transfected primary cultures of 3 / mouse cortical neurons with 1 µg of GABA_A receptor rat 3 subunit (a gift of Dr. Peter H. Seeburg, University of Heidelberg, Germany) subcloned into the expression vector pCDM8 (Invitrogen) and the enhanced green fluorescent protein (EGFP) plasmid (0.3 µg, Clontech, Palo Alto, CA, No. 6077-1) to allow visualization of successfully transfected cells. A modification of the calcium phosphate precipitation technique (Chen and Okayama 1987) was used. Briefly, cultured neurons at DIV5 on glass coverslips were transferred to a well in a four-well plate with 500 µl transfection medium, a MEM medium (Invitrogen) with pH adjusted to 7.85 by 5 M NaOH. Then 30 µl of the DNA/Ca²⁺ mixture was added and incubated for 30 min at room temperature. After two washes with transfection medium, the original culture medium was returned and neurons were maintained at 37°C in 5% CO₂ for several days.

IMMUNOCYTOCHEMISTRY. Live cultured neurons were incubated in anti-1 (4 µg/ml, Upstate, Waltham, MA), anti-2 (1:3000), anti-3 (1:3000), anti 5 (1:2000), kind gifts from Dr. Jean Marc Fritschy (University of Zurich, Switzerland), and anti 2/3 (4 µg/ml, Upstate) antibodies for 90 min at room temperature. Cells were washed and fixed in 4% paraformaldehyde, 4% sucrose in PBS for 15 min at room temperature. Cells were then incubated in 10% goat serum for 30 min to block nonspecific staining. After washing with PBS several times, cells were incubated with indocarbocyanine (Cy3) conjugated secondary antibodies (1:500, Jackson ImmunoResearch Laboratories, West Grove, PA) for 1 h at room temperature. Nikon band-pass filter cubes were used for Cy3 or EGFP fluorescence. Neurons were imaged on a Nikon EN600 microscope equipped with a ×20 Ph2 0.5 N.A. and a ×60, 1.0 N.A. objective. Digital images were acquired with a Hamamatsu Orca-100, 12-bit (4,096 gray-scale intensity level) cooled CCD digital camera, 1,392 × 1,040 pixel array. The expression of GABA_A receptor subunits was investigated in DIV15 cortical neurons. Fluorescence intensity was measured from region of interest transferred with MetaMorph image analysis software (Universal Imaging, Downingtown, PA) from phase contrast microphotographs taken with a ×20 objective to mark the outline of individual neurons and to avoid bias. Cells were considered positive when staining intensity was greater that twice the background. To count dendritic clusters, a single level of focus was maintained throughout each recording. Recording at a single focal plane was usually sufficient to capture GABA receptor clusters throughout the full thickness of small distal dendrites but not of proximal dendrites or cell body. Thus for our measurements, we estimated the number of clusters only in distal dendrites. To quantitate changes in clustering, we measured the average pixel intensity of all clusters along a dendritic segment. Measurement of average fluorescence intensity was performed using MetaMorph. Several neurons from two to three coverslips per culture were randomly selected on the basis of healthy morphology and scored to determine the percentage of clusters in segments of dendrites of 50 µm length. CCD images were background subtracted, and to define clusters, a single threshold was chosen manually so that clusters corresponded to puncta of at least twofold greater intensity than the diffuse fluorescence on the dendritic shaft. A minimum size of 3 pixels was considered to define a cluster, pixel size at ×60 magnification was 0.109 µm.

ELECTROPHYSIOLOGY. Cortical neurons were voltage-clamped at room temperature. The recording chamber was continuously perfused at 5 ml/min with an extracellular medium composed of (in mM) 145 NaCl, 5 KCl, 1mM MgCl₂, 2 CaCl₂, 5 glucose, and 5 HEPES at pH 7.4 with NaOH. Osmolarity was adjusted to 325 mosM with sucrose. Electrodes were pulled in two stages on a vertical pipette puller from borosilicate glass capillaries (Wiretrol II, Drummond). Typical pipette resistance was 5-7 M. Intracellular (patch pipette) solutions contained (in mM) 145 CsCl, 5 MgATP, 10 bis-(o-aminophenoxy)-N,N,N',N'-tetraacetic acid (BAPTA), 0.2 NaGTP, and 10 HEPES at pH 7.2 with CsOH. Whole cell recordings were performed with a patch-clamp amplifier (Axopatch 200, Axon Instrument, CA) after capacitance and series resistance compensation. Series resistance was typically <15 M and was checked for constancy throughout the experiments. GABA was applied directly by gravity-fed Y-tubing delivery system (Murase et al. 1989) placed within 100 µm of the recorded cell. GABA application had fast onset (<20ms) and achieved a completely local perfusion of the recorded cell. Repeated applications were followed each time by a recovery period of 2 min. The peak amplitude of responses to individual GABA concentration was measured and normalized to the maximal GABA current. Dose-response curves were fitted with the logistic equation
where I_max is the maximal Cl current, elicited by GABA, EC₅₀ is the GABA concentration eliciting the half-maximal response, and n_h is the Hill coefficient. Drugs were also coapplied with GABA after a short period of preperfusion before coapplication with GABA to observe maximal potentiation. Zolpidem (Sigma) and Loreclezole (a gift from Janssen Research Foundation, Beerse Belgium) were dissolved in dimethylsulfoxide (DMSO, <0.001% final concentration, Sigma) and diluted in the extracellular medium. They were superfused through a parallel input to the perfusion chamber until effective replacement of the solution was obtained for synaptic current recordings. Miniature IPSCs (mIPSCs), a subset of spontaneous IPSCs (sIPSCs), were recorded in the presence of tetrodotoxin (TTX 1 µM, Sigma). Currents were filtered at 2 kHz with an 8-pole low-pass Bessel filter (Frequency Devices, Haverhill, MA), digitized using a PC-compatible microcomputer equipped with a Digidata 1200 data-acquisition board (Axon Instruments) and Pclamp8 (Axon Instrument) software. Off-line data analysis, curve fitting, and figure preparation were performed with Origin (MicroCal Software, Northampton, MA), PClamp 8.0 (Axon Instrument) and Mini Analysis (Synaptosoft, www.synaptosoft.com, Decatur, GA) softwares. For each cell, mIPSC were averaged from 100 events aligned on the point of steepest rise. Peak amplitudes were measured at the absolute maximum of the currents, taking into account the noise of the baseline and noise around the peak. Rise times were measured as the time elapsed from 10 to 90% of the peak amplitude of the response. Curve fitting was performed using simplex algorithm least-squares exponential fitting routines with double exponential equations of the form I(t) = I_fe ^(t/_f) + I_se ^(t/_s) where I_f and I_s are the amplitudes of the fast and slow decay components, and _f and _s are their respective decay time constants. To compare decay times between different experimental conditions, we used a weighted mean decay time constant _w = [I_f/(I_f + I_s)]_f + [I_s/(I_f + I_s)]_s. Drug effects were assessed based on averages of 100 events in each neuron by statistical comparisons. Unless otherwise indicated, data are expressed as mean ± SE; P values represent the results of independent t-test with Bonferroni corrections, or Newmann-Kuel test with prior ANOVA for repeated measures as appropriate.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

GABA responses in 3 / neurons

We investigated GABA activated currents in cortical neurons at day 7 in vitro (DIV7) from 3 +/+ and 3 / mice. To allow a comparison between neurons with different size and dendritic arborization, we used excised nucleated patches (Sather et al. 1992) that have a considerable amount of membrane and at the same time allow better quality voltage clamping and more reliable drug application. The application of increasing GABA concentrations elicited inward currents in these patches when symmetrical Cl concentrations were present in the intracellular recording pipette solution and in the extracellular bath solution (Fig. 1, A and B). Analysis of the dose response (Fig. 1C) yielded slightly different EC₅₀, but current density values were significantly smaller for patches from 3 / neurons (Fig. 1D).

View larger version (21K): [in this window] [in a new window]   Fig. 1. Dose-response of GABA-elicited Cl currents in cortical neurons from 3 / mice. Representative Cl currents (voltage-clamped at 60 mV holding potential) were elicited in nucleated patches from neurons at day 7 in vitro (DIV7) derived from 3 +/+ (A) and 3 / (B) mice. GABA applications at increasing concentrations were performed for the duration indicated by the bars. The average of GABA dose-response curves normalized to the maximal current in individual neurons (C) and the current density, maximal current normalized by the cell capacitance (D), derived from neurons in 3 3 +/+ and 3 3 / mice. EC50 values were 7.4 ± 1.1 µM (+/+) and 15.8 ± 1.2 µM (/). Number of cells is indicated in parenthesis. *P < 0.005.

To verify the expression of distinct functional GABA_A receptors between 3 +/+ and 3 / mice, we compared the allosteric modulation of GABA_A receptors by applying the imidazopyridine zolpidem. To test the degree of potentiation by zolpidem, the peak current elicited by a GABA concentration that was at the EC₂₀ derived from the respective GABA dose-response curve was considered as control, and then the drug was coapplied with GABA (Fig. 2, A and B). As shown in Fig. 2C, for GABA_A receptors in 3 +/+ mice cortical neurons at both DIV7 and DIV15, the GABA-induced currents were slightly potentiated by zolpidem (100 nM). On the other hand, this drug induced a threefold larger potentiation of GABA response in neurons from 3 / than from 3 +/+ mice at both ages in culture tested (Fig. 2C).

View larger version (17K): [in this window] [in a new window]   Fig. 2. Whole cell patch-clamp recordings of GABA currents in cortical neurons from genetically altered mice (3 /). Whole cell inward Cl currents elicited by GABA at the EC20 in voltage-clamped cortical neurons at DIV7 are differentially potentiated by zolpidem (100 nM) in a neuron from a culture from 3 +/+ mice (A) and a neuron from a culture from 3 / mice (B). In C, the summary of the experiments is shown demonstrating the potentiation of GABA activated currents in these neurons at 2 distinct days in vitro (DIV) in 3 3 +/+ and 3 3 / mice. EC20 for GABA was similar in neurons at DIV7 and DIV15. Number of cells is indicated in parenthesis. *P < 0.005.

IPSCs in 3 / neurons

We studied mIPSCs at DIV15 in 31 cortical neurons in cultures from four 3 +/+ mice and 21 cortical neurons from four 3 / mice. While some neurons without mIPSCs were found in both genotypes, their occurrence was not significantly different. mIPSCs amplitude and frequency of occurrence were extremely variable between 3 +/+ neurons [45 ± 28 pA; 1.3 ± 0.9 (SD) Hz] and were not significantly different from 3 / neurons (40 ± 39 pA; 1.2 ± 1.3 Hz). In contrast, the weighted time constant resulting from the double-exponential fitting of the mIPSCs decay (_w) was 2.5-fold larger in the 3 +/+ than in 3 / neurons (Fig. 3). The coefficient of variation of the mIPSCs decay for both genotypes was 0.45. At DIV7, mIPSCs could not be investigated because the strong depression of event frequency with TTX in these young neurons prevented reliable measurement. However, the _w from sIPSCs (without TTX) was faster in 3 / than in 3 +/+ neurons (Fig. 3C).

View larger version (23K): [in this window] [in a new window]   Fig. 3. Whole cell patch-clamp recordings of synaptic currents in 3 / cortical neurons. A: the occurrence of minitaure inhibitory postsynaptic currents (mIPSCs; in TTX) is illustrated on a slow time scale in cortical neurons from DIV15 cultures of 3 +/+ mice (top) and 3 / mice (bottom). B: averages of 50 currents recorded from individual cortical neurons from DIV15 cultures of 3 +/+ mice (top) and 3 / mice (bottom) are illustrated with a superimposed double-exponential curve with an indication of the weighted decay time constant (w) from the exponential fitting of these currents. C: the summary data on weighted time constant from the exponential fitting of these currents are compared between mice and distinct days in vitro. Data from DIV7 mice are from sIPSCs (no TTX). Bars indicate the average values. Data derive from 3 distinct mice in each group with the total number of cell indicated in parenthesis over the bars. * P < 0.01.

We then compared the effects of zolpidem (100 nM) on sIPSCs and mIPSCs recorded, respectively, from neurons at DIV7 and DIV15 (Fig. 4). In cells from 3 / mice, zolpidem prolongation of both sIPSCs and mIPSCs was significantly larger compared with cells from 3 +/+ mice (Fig. 4). In 11 3 +/+ and 8 3 / cortical neurons in culture at DIV15, we also compared the prolongation of mIPSCs decay by the 2/3 selective drug loreclezole (Fisher et al. 2000; Wafford et al. 1994). Bath perfusion with 10 µM loreclezole, prolonged the _w by 115 ± 19 versus 149 ± 32%, in neurons from the 3 +/+ and 3 / mice, respectively. The prolongation of mIPSCs decay was not statistically different between the two groups.

View larger version (23K): [in this window] [in a new window]   Fig. 4. Effects of the imidazopyridine zolpidem on mIPSCs in cortical neurons from genetically altered mice. mIPSCs recorded in cortical neurons in primary culture (DIV15) are illustrated before and after zolpidem perfusion in a 3 +/+ mouse (A) and in a homozygote 3 / mice (B). C: the percent prolongation of the decay time constants of mIPSCs (DIV15) and sIPSC (DIV7) by 100 nM zolpidem is reported for cultures from different mice and a distinct days in vitro. Data derive from 3 mice in each group. *, significant with respect to 3 +/+ mice (P < 0.05).

subunit expression in 3 / neurons

Our electrophysiological data suggest a change in  subunit composition of GABA_A receptors resulting from the 3 subunit deletion. For instance, a larger participation of the 1 subunit or a reduction of 2/3 subunits could both enhance zolpidem sensitivity and decrease synaptic current duration. We therefore used surface immunolabeling with specific antibodies against the 1, 2, 3, and 5 subunits to verify the extent of expression of these subunits in cortical neurons in primary cultures from 3 +/+ and 3 / mice. We also used for comparison an antibody specific for the 2/3 subunits. We followed the protocol suggested in Brunig et al. (2002) with surface staining with primary antibodies in live cells followed by fixation and secondary antibody staining to minimize antigen capping by the secondary antibodies in living cells.

A comparison of various  and 2/3 subunits staining in cortical neurons from both genotypes was assessed from fluorescent and phase contrast microphotographs as percent of the total neurons expressing these subunits counted in at least six coverslips from DIV15 cultures deriving from three distinct mice in each group. A significant reduction of the number of cells expressing 2 (from 81 ± 7 to 46 ± 7%, n = 202 neurons) as well as 3 subunits (from 75 ± 5 to 32 ± 10%, n = 140) was detected, whereas no change was observed for 1 subunit (from 76 ± 6 to 61 ± 9%, n = 192) in 3 / mice. As expected, deletion of the 3 subunit resulted in decreased staining with a 2/3-specific antibody (from 89 ± 7 to 50 ± 7%, n = 138), indicating that the 2 subunit is still present and that it does not compensate for the lack of 3 subunit, as previously reported (Homanics et al. 1997).

In Fig. 5 examples of neurons stained live revealed clusters of GABA_A receptors subunits in dendrites from cortical neurons. In most cells from 3 +/+ mice, punctate staining on the membrane was more prominent for the 2 than for the 1 and 3 subunits. Staining for the 5 subunit was very weak and it was not different between genotypes (not shown). In Table 1, we report the results of an assessment of density of receptor clusters (number of clusters per 10 µm) for  subunit isoforms expressed on the surface of dendrites of neurons from 3 +/+ and 3 / mice. There was a significant reduction in cluster density for both 2 and 3 subunits, whereas, no significant change in density of clusters was seen when surface expression was evaluated with an antibody for 2/3 subunits. A striking effect of 3 deletion was observed when the relative fluorescence intensity of clusters was compared between neurons from 3 +/+ and 3 / mice. As shown in Table 1, while the average cluster intensity did not change for the 1 subunit, it was significantly decreased for both 2, 3, and 2/3 subunits.

View larger version (96K): [in this window] [in a new window]   Fig. 5. Expression of distinct  subunits of the GABAA receptor in 3 +/+ and 3 / neurons. Mouse cortical neurons were immunostained live at DIV15 with specific antibodies against 1, 2, and 3 subunits. Cells were then fixed and stained with Cy3-conjugated secondary antibodies. Each panel shows micrograph of Cy3 fluorescence for cells derived from 3 +/+ (left) or 3 / (right) mice. Insets with partial segments of dendrites are shown in each panel illustrating cluster of receptor subunits. Scale bar 20 and 7 µm for insets.

View this table: [in this window] [in a new window]   Table 1. GABAA receptor subunit clusters in dendrites of cortical neurons in culture from 3 +/+ and 3 / mice

3 subunit transfection in 3 / neurons

We used a modification of calcium phosphate mediated DNA transfection to introduce cDNA for the 3 subunit and EGFP into cortical neurons from 3 / mice. Three coverslips of cultures from a 3 / mouse were transfected at DIV5, and electrophysiological and immunocytochemical studies were performed at DIV15. In Fig. 6, a cortical neuron expressing EGFP has been surface stained with 2/3 antibody. Surface expression of the subunit was similar to that seen in neurons from 3 +/+ mice and contrasts with the weaker staining seen in neurons from 3 / mice (Fig. 6). In 28 transfected neurons from the three coverslips, 91% of cells had 2/3 staining significantly higher than twice the background. In 10 3 transfected neurons from two coverslips, surface staining for the 2 and 3 subunit revealed that all these cells were positive for both subunits. In 2/3 antibody-stained neurons, when average dendritic cluster intensity was compared between neurons in culture from 3 +/+ mice and 3 transfected neurons in cultures from 3 / mice, no significant differences were observed. Similar results were obtained when cluster intensity was compared for 2 and 3 antibody staining in neurons in culture from 3 +/+ mice and 3 transfected neurons in cultures from 3 / mice. Decay time of mIPSCs recorded in five 3 transfected cells (_w, 51 ± 7 ms) were not different from those from 3 +/+ neurons. Zolpidem (100 nM) prolonged the _w of mIPSC by 26 ± 5%, a similar extent to that observed in 3 +/+ neurons.

View larger version (80K): [in this window] [in a new window]   Fig. 6. Transfection of the 3 subunit in 3 / neurons. A: surface anti-3-cy3 staining in (+/+) and (/) neurons. B: EGFP fluorescence (left) and anti-3-Cy3 surface staining (right) in 3 / neurons cotransfection with 3 cDNA and EGFP. Scale bar 20 µm.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

We have studied GABA_A receptors in cortical neurons in primary cultures from embryonic mice with and without the deletion of the 3 subunit. Our results show a reduced GABA current density, enhanced sensitivity to zolpidem, and shorter duration of m- and sIPSCs in neurons from 3 / mice. Current density reduction of a similar extent was observed in cultured hippocampal neurons from these mice together with the small changes we observed in GABA sensitivity (Krasowski et al. 1998). We complement these results by showing that the enhancement of whole cell current with the 1-subunit-selective drug zolpidem is considerably larger in neurons from 3 / mice than in 3 +/+ mice. Smaller and faster sIPSCs were previously reported both in neurons of the reticular thalamic nucleus (Huntsman et al. 1999) and in granule neurons in the olfactory bulb (Nusser et al. 2001) in slices from 3 / mice. We have observed shortening of sIPSCs and mIPSCs as result of the 3 subunit deletion in cortical neurons. However, the heterogeneity between neurons and the large variability in size of mIPSCs has prevented rigorous assessment of differences in mIPSCs amplitude between 3 +/+ and 3 / neurons.

We studied the prolongation of mIPSC duration by zolpidem and loreclezole, drugs selective for the 1 and 2/3 subunits, respectively (Pritchett et al. 1989; Wafford et al. 1994). Zolpidem action was considerably enhanced in neurons from 3 / mice relative to 3 +/+, whereas the effect of loreclezole was unchanged. These results suggest that the 3 subunit deletion has produced a selective retention of 1 subunit containing GABA_A receptors. They also suggest that no other  subunit is able to substitute for the 3 subunit and that the remaining functional GABA_A receptors are 2 rather than 1 containing as previously suggested (Homanics et al. 1997). Quantification of surface immunocytochemical staining with subunit selective antibodies revealed that both the number of positive neurons and dendritic subunit cluster density and fluorescent intensity for the 2 and 3 subunits were significantly reduced in cultures from 3 / mice. A similar reduction was not observed for the 1 subunit. Furthermore, the fluorescent intensity of dendritic GABA receptor clusters stained with 2/3 antibody, was significantly lower in 3 / neurons. Taken together our results suggest that 2 and 3 subunit form separate pools of receptors with specific  subunit. The immunocytochemical results support the electrophysiological finding of a selective retention of 1 subunit containing GABA_A receptor in 3 / mice and imply that these receptors are co-assembled with the 2 subunit whereas the majority of 2 and/or 3 subunits are assembled with the 3 subunit.

The density of dendritic GABA_A receptor clusters stained with 2/3 antibody, was not decreased in neurons from 3 / mice compared with 3 +/+ mice. This may suggest that the deletion of the 3 subunit in mice does not affect the formation of clusters unlike what is observed with the 2 subunit deletion (Essrich et al. 1998). The reduction in cluster density of 2 and 3 subunits may relate to the decreased surface expression of those subunits rather than to the inability of those subunits to form clusters. This supports the pivotal role for  subunit in surface expression and localization of GABA_A receptor subtypes (Connolly et al. 1996; Connor et al. 1998). Our data suggest specificity for the 3 subunit in allowing surface expression of 2 and 3 subunit receptor subtypes. This suggestion is further supported by the results that show that 3 cDNA transfection in neurons from 3 / mice restored 2/3, 2, and 3 subunit surface staining, slowed mIPSC decay and decreased zolpidem prolongation.

The similar cluster density between 1, 2, and 2/3 subunits stain in 3 +/+ neurons and the lack of reduction of cluster density with 2/3 subunit stain in 3 / neurons indicates that receptors pools containing these subunits coexist in the same puncta in 3 +/+ neurons. This suggestion is supported by the similar coefficient of variation measured for mIPSCs decays in neurons from 3 +/+ and 3 / mice. In fact, if synapses were made on a separate population of 1 and 2/3 subunit containing receptor clusters in 3 +/+ mice, one would expect a larger variation of decay than in 3 /.

Native and recombinant GABA_A receptors containing the 1 subunit have a faster time course than those with 2 or 3 subunits. (Gingrich et al. 1995; Lavoie et al. 1997; McClellan and Twyman 1999; Verdoon 1994). This result suggests that deletion of the 3 subunit left 1 containing GABA_A receptor isoforms endowed with faster decay kinetics. Alternatively, the acceleration in the deactivation kinetics could be explained by the presence of 2 rather than 2/3 receptors. Our finding of increased zolpidem sensitivity of sIPSCs duration in the 3 / mice cortical neurons favors the 1 subunit as being responsible for short mIPSCs in cortical neurons from these mice. Further work that will characterize the deactivation properties of recombinant receptors with distinct  subunit isoforms will ultimately address this issue.

The suggestion that l-containing receptors are selectively retained in 3 / neurons, implies that they are likely co-assembled with the 2 subunit. This lend further support to the hypothesis that two major types of GABA_A receptor can be found throughout the CNS, one with slow kinetics composed of 2/3 subunits together with the 3 subunit and another with fast kinetics where 1 together with 2 are the major players. Support for this hypothesis comes from anatomical and biochemical data that show the preferential association of the 2 with the 3 subunit and the 1 with the 2 subunit (Benke et al. 1994; Fritschy et al. 1994). Because benzodiazepines affecting 1 subunits are sedative and those affecting 2 are anxiolytic (Mohler et al. 2002), it is possible that slow decaying IPSCs allow inhibitory neurons to produce stronger and longer-lasting inhibition perhaps associated with a decrease in vigilance and in generalized anxiety while those associated with fast IPSCs are target for sedation. Fast IPSCs in 3 / mice may underlie the generalized seizures reported in behavioral studies of these mice (DeLorey et al. 1998; Homanics et al. 1997). Additionally, given the pivotal role of GABA in early development (Maric et al. 2001) and the high expression of 2- and 3-containing receptors at early stages in life (Hornung and Fritschy 1996; Laurie et al. 1992; Poulter et al. 1992), the reduction of expression of receptors containing these subunits may possibly be related to the severe developmental malformation and behavioral impairment observed in these mice.