HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Corrigendum All Versions of this Article: 96/2/846    most recent 01199.2005v1 Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via ISI Web of Science (18) Citing Articles via Google Scholar Google Scholar Articles by Prenosil, G. A. Articles by Vogt, K. E. Search for Related Content PubMed PubMed Citation Articles by Prenosil, G. A. Articles by Vogt, K. E.

Specific Subtypes of GABA_A Receptors Mediate Phasic and Tonic Forms of Inhibition in Hippocampal Pyramidal Neurons

University of Zurich, Institute of Pharmacology and Toxicology, Zurich, Switzerland

Submitted 10 November 2005; accepted in final form 8 May 2006

	ABSTRACT

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION GRANTS ACKNOWLEDGMENTS REFERENCES

 
The main inhibitory neurotransmitter in the mammalian brain, GABA, mediates multiple forms of inhibitory signals, such as fast and slow inhibitory postsynaptic currents and tonic inhibition, by activating a diverse family of ionotropic GABA_A receptors (GABA_ARs). Here, we studied whether distinct GABA_AR subtypes mediate these various forms of inhibition using as approach mice carrying a point mutation in the -subunit rendering individual GABA_AR subtypes insensitive to diazepam without altering their GABA sensitivity and expression of receptors. Whole cell patch-clamp recordings were performed in hippocampal pyramidal cells from single, double, and triple mutant mice. Comparing diazepam effects in knock-in and wild-type mice allowed determining the contribution of 1, 2, 3, and 5 subunits containing GABA_ARs to phasic and tonic forms of inhibition. Fast phasic currents were mediated by synaptic 2-GABA_ARs on the soma and by synaptic 1-GABA_ARs on the dendrites. No contribution of 3- or 5-GABA_ARs was detectable. Slow phasic currents were produced by both synaptic and perisynaptic GABA_ARs, judged by their strong sensitivity to blockade of GABA reuptake. In the CA1 area, but not in the subiculum, perisynaptic 5-GABA_ARs contributed to slow phasic currents. In the CA1 area, the diazepam-sensitive component of tonic inhibition also involved activation of 5-GABA_ARs and slow phasic and tonic signals shared overlapping pools of receptors. These results show that the major forms of inhibitory neurotransmission in hippocampal pyramidal cells are mediated by distinct GABA_ARs subtypes.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION GRANTS ACKNOWLEDGMENTS REFERENCES

 
In the mammalian CNS, GABA is the main inhibitory neurotransmitter, activating GABA_A receptors (GABA_ARs) on target neurons either in a phasic or a tonic fashion (Farrant and Nusser 2005; Mody and Pearce 2004). GABA_ARs are composed of five subunits, most frequently two , two , and one subunit. Different subunit isoforms (1–6, 1–4, 1–4, , , , ) give rise to a considerable diversity of GABA_ARs (Barnard et al. 1998; Mohler et al. 2002; Sieghart and Sperk 2002) that are differentially expressed in the brain and localized in different cell types and subcellular areas (Fritschy and Mohler 1995; McKernan and Whiting 1996; Pirker et al. 2000). Although much is known about the distribution and subcellular location of major GABA_AR subtypes, the functional significance of this diversity is less well understood, chiefly because of a lack of pharmacological tools that distinguish between the different subtypes. The presence of a 2 or 3 subunit is required for the formation of a benzodiazepine-binding site (Knoflach et al. 1991; Pritchett et al. 1989). Additionally, the subunit variant determines activation and deactivation kinetics of GABA_ARs (Banks and Pearce 2000; Bosman et al. 2002; Hutcheon et al. 2000; Vicini et al. 2001) and their affinity for classical benzodiazepines (Scholze et al. 1996; Smith et al. 2001) such as diazepam (DZ). Mutation of a conserved histidine residue (His 101 in the 1 subunit) into an arginine renders the corresponding GABA_AR DZ-insensitive (Benson et al. 1998; Rudolph et al. 1999; Wieland et al. 1992). Using gene targeting to introduce this point mutation in various subunit variants in vivo, it has been shown that different GABA_AR subtypes mediate distinct effects of DZ (Rudolph and Mohler 2004). In particular, its sedative properties are mediated by 1-GABA_ARs (McKernan et al. 2000; Rudolph et al. 1999) and its anxiolytic effects by 2-GABA_ARs (Low et al. 2000). It can therefore be assumed that specific neuronal circuits use distinct GABA_AR subtypes differing in subunit variant.

Two major modes of GABA_AR-mediated inhibitory transmission can be observed in mammalian CNS: phasic inhibition mediated by synaptic receptors and tonic inhibition mediated by extrasynaptic receptors (Farrant and Nusser 2005; Mody and Pearce 2004). In the hippocampal formation, phasic inhibitory postsynaptic currents (IPSCs) can further be subdivided into GABA_A,fast and GABA_A,slow IPSCs based on kinetics and amplitude (Pearce 1993). GABA_A,fast IPSCs are mediated mostly by somatic and proximal dendritic synapses (Freund and Buzsaki 1996), whereas GABA_A,slow IPSCs are thought to originate at distal dendritic sites (Banks et al. 1998). GABA_A,slow IPSCs exhibit on average a larger charge transfer than GABA_A,fast IPSCs; their prolonged time course and their sensitivity to GABA reuptake inhibitors (Banks et al. 2000) suggested that they are activated by GABA spillover.

The subunit composition and function of tonically activated GABA_ARs varies across brain regions and cell types. In cerebellum, dentate gyrus, neocortex, and thalamic relay nuclei, tonic inhibition is mediated by DZ-insensitive GABA_ARs containing 6 or 4 subunits together with the subunit (Brickley et al. 1996; Cope et al. 2005; Drasbek and Jensen 2005; Mtchedlishvili and Kapur 2006; Nusser and Mody 2002; Porcello et al. 2003; Stell et al. 2003; Sun et al. 2004). In the CA1 area, tonic inhibition is mediated by DZ-sensitive GABA_ARs in interneurons and, to a lesser extent, pyramidal cells, when ambient GABA concentration is increased (Scimemi et al. 2005; Semyanov et al. 2003). At low GABA concentrations activation of GABA_ARs containing the 4 subunit was found to dominate (Scimemi et al. 2005). Thus far, however, no single GABA_AR subtype has been found to have an exclusive synaptic or extrasynaptic location, but several lines of evidence suggest that specific GABA_ARs selectively participate in different forms of inhibition. Unraveling this selectivity would help understanding how distinct behaviors and DZ effects are linked with the diverse forms of GABAergic inhibition.

The aim of this study was to investigate whether DZ-sensitive GABA_ARs differing in subunit composition mediate distinct modes of GABAergic inhibition in the hippocampal formation. Whole cell patch-clamp recordings of CA1 and subicular pyramidal cells were performed on acute slices from knock-in mice carrying a histidine-to-arginine point -mutation in either the 1, 2, or 3 subunit gene to render the respective GABA_AR DZ-insensitive. Using this approach in single (1, 2, 3), double (12), and triple (123) mutant mice to pharmacologically isolate GABA_ARs containing different subunits, we studied the contribution of identified DZ-sensitive GABA_AR subtypes to evoked and spontaneous IPSCs and to tonic inhibition in the hippocampus. A particular focus was put on characterization of GABA_A,slow IPSCs, because they share elements of both phasic and tonic type of GABAergic inhibition.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION GRANTS ACKNOWLEDGMENTS REFERENCES

 
Generation of mutant mice

Experiments were performed in 129/SvJ knock-in mice carrying DZ-insensitive GABA_AR subtypes obtained by a histidine-to-arginine point mutation in the 1, 2, or 3 subunit gene [1(H101R), 2(H101R) and 3(H126R)] (Low et al. 2000; Rudolph et al. 1999) introduced into the mouse genome by homologous recombination in embryonic stem cells. Mice carrying a point mutation in both 1 and 2 subunits (12) were obtained by crossing double heterozygous mice born from homozygous 1(H101R) and 2(H101R) intercrosses. Homozygous 1(H101R)/2(H101R) offspring were identified by PCR analysis from tail biopsy. Likewise, triple mutants, carrying point-mutated 1, 2, and 3 subunits (123) were generated by crossing double homozygous 1(H101R)/2(H101R) and 1(H101R)/3(H126R) mice, which yielded offspring homozygous for 1(H101R) but heterozygous for 2(H101R) and 3(H126R) subunits. These were crossed again with each other to obtain mice homozygous for all three point-mutations, as confirmed by PCR analysis. All experiments were approved by the cantonal Veterinary Office of Zurich and were performed in accordance with the European Community Council Directive (86/609/EEC).

Immunohistochemistry

Mice (P21) were anesthetized with pentobarbital sodium (Nembutal, 50 mg/kg ip, Abbot Laboratories, Chicago, IL) and were perfused transcardially with 4% paraformaldehyde dissolved in 0.15 M sodium phosphate buffer containing 15% saturated picric acid solution. The brain was extracted, postfixed for 4 h in the same solution, incubated overnight in sodium citrate buffer (pH 4.5), and boiled for 60 s in a microwave oven for antigen retrieval (Fritschy et al. 1998). They were then cryoprotected with 30% sucrose in phosphate-buffered saline, frozen with dry-ice, and cut in 40-µm parasagittal sections using a freezing microtome. Free-floating sections were processed for immunoperoxidase staining using antibodies against the 1, 2, 3, and 5 subunit, as described (Fritschy et al. 1998). For antibody characterization, see Fritschy and Mohler (1995).

Slice preparation

Mice from both sexes (P18-24) were anesthetized with inhaled isoflurane and decapitated. The brain was quickly removed and placed in ice-cold artificial cerebrospinal fluid (ACSF, composition in mM: 125 NaCl, 26 NaHCO₃, 1.25 NaH₂PO₄, 2.5 KCl, 1 MgCl₂, 2.5 CaCl₂, and 11 glucose, oxygenated with 95% O₂-5% CO₂). The brain was affixed to a vibratome stage (Microm HM 650 V, Microm International AG, Volketswil, Switzerland) with cyanoacrylate and kept in the ice-cold ACSF for slicing. Parasagittal 300- to 350-µm-thick hippocampal slices were prepared and incubated at 33°C for 20 min before being stored at room temperature (25°C) in oxygenated ACSF. For the recording of GABA_A,slow IPSCs, mice (P24–P30) were anesthetized with Nembutal and were perfused transcardially with 50 ml ice-cold sucrose-ACSF (composition in mM: 87 NaCl, 26 NaHCO₃, 1.25 NaH₂PO₄, 2.5 KCl, 9 MgCl₂, 0.5 CaCl₂, 75 sucrose, and 25 glucose). Parasagittal slices (310 µm thick) were cut in sucrose-ACSF, transferred to normal ACSF, and incubated and stored as above.

Electrophysiological recordings and data analysis

Slices were visualized with a CCD camera (PCO Vx45, Till Photonics) mounted on an upright microscope (BX51WI, Olympus), equipped with a long working distance water-immersion objective (Xlumplan FI 20X, 0.95 numerical aperture), a fourfold magnification changer, Nomarski-type differential interference contrast, and infrared illumination. Patch electrodes were pulled from borosilicate glass (GC150TC, Clark Instruments) and had an open tip resistance of 3–4 M when filled with the internal solution. Kynurenic acid (2 mM) was added to the external ACSF solution to block excitatory synaptic transmission. Recordings were made using a Multiclamp 700A patch-clamp amplifier (Axon Instruments), filtered at 4 kHz, digitized at 20 kHz, stored, and analyzed using IGOR Pro software (Wave Metrics). Access resistance was monitored for all experiments, and they were not included in further analysis if it changed by >20% during the recording.

Evoked IPSCs

Whole cell voltage-clamp recordings from CA1 pyramidal cells of wildtype (WT), 1, 2, 3, 12, and 123 mutant mice were made at room temperature (25°C) with continuous superfusion (1–2 ml/min) of ACSF. External stimulation (0.1–10 µA) was delivered every 10 s with a constant current stimulus isolator (WPI) through a bipolar, custom-made electrode from polytetrafluoroethylene-insulated platinum-iridium wire of 50 µM diam (Advent). Evoked IPSCs (eIPSC) were recorded at a holding potential of 0 mV with a low chloride containing internal solution (in mM: 130 CsGlu, 1 EGTA, 10 HEPES, 5 MgATP, 0.5 NaGTP, and 5 NaCl, pH 7.3, 295 mOsm). Access resistance was 10–12 M.

To assess the contribution of different GABA_AR subtypes to eIPSCs in slices from WT and knock-in mice, DZ (1 µM, dissolved in DMSO) was bath-applied after 15 min of baseline recordings, and eIPSCs were recorded for another 15–30 min. The average eIPSC amplitude after application of DZ was normalized to the peak amplitude of the average baseline eIPSC. No significant difference in the DZ effect was observed between mono- and biexponential fittings; therefore, we chose the monoexponential method. The increase in amplitude and after DZ application in each line of mutant mice was compared with WT mice using one-way ANOVA followed by Bonferroni’s post hoc multiple comparisons tests (SPSS 11.5, Lead Technology). In all experiments, n refers to the number of cells recorded. On average, one to two cells were recorded from one animal.

Spontaneous GABA_A,fast and GABA_A,slow IPSCs

We recorded spontaneous GABAergic events to reliably distinguish between GABA_A,slow and GABA_A,fast and to facilitate the comparison of results from two hippocampal regions. Whole cell voltage-clamp recordings of spontaneous IPSCs (sIPSC) from CA1 and subiculum pyramidal cells from WT and 123 mutant mice were obtained at room temperature with a holding potential of –60 mV and a high chloride containing internal solution (in mM: 100 CsCl, 2 MgCl₂, 1 EGTA, 2 ATP, 0.3 GTP, and 40 HEPES, pH 7,2, 300 mOsm). Experiments with the GABA reuptake inhibitor NO711 (2 µM, dissolved in DMSO) were performed in the presence of the GABA_B receptor antagonist CGP 55845 (1 µM, dissolved in DMSO). Continuous recordings started after the holding current had stabilized. Spontaneous events were recorded for a 10- to 15-min baseline period and for the same amount of time in the presence of each drug tested. Banks et al. (2002) reported an increase in frequency of GABA_A,slow sIPSCs with age in rats, whereas other parameters such as amplitude and kinetics remained unchanged. We checked for the age dependency of GABA_A,slow sIPSCs in WT mice by recording from 34 animals ranging from P15 to P35. As in rats, the frequency of GABA_A,slow sIPSCs was higher in older animals. However, we observed a steep increase around P19 (results not shown), whereas in rats, the mean frequency of GABA_A,slow sIPSCs was reported to increase gradually with age. Based on this result, all subsequent sIPSC recordings were performed in mice between P20 and P25.

Spontaneous events were detected off-line automatically with the ‘Mini Analysis’ software (Synaptosoft) with the detection threshold being set 5 times higher than the RMS level of baseline noise. All detected events were counted for analysis of frequency. For analysis of kinetics and amplitude only currents without subsequent contaminating events in the decaying phase were considered. Spontaneous events having an onset-to-peak rise time >5 ms were classified as GABA_A, _slow sIPSCs; remaining events were classified as GABA_A,fast sIPSCs (Banks et al. 2000). Amplitude, rise time, and decay time constants for single GABA_A,slow sIPSCs were calculated by the Mini Analysis software. Kinetics of GABA_A,fast sIPSCs was also calculated individually, whereas their amplitude was determined from an average trace for each recorded cell. This minimized contamination by noise as, in general, the signal-to-noise ratio was smaller for GABA_A,fast than for GABA_A,slow sIPSCs. The effects of DZ and NO711 on sIPSCs were calculated from the average value of pooled sIPSCs from single experiments before and after drug application. Statistical significance was determined by paired Student’s t-test.

Tonic inhibition recordings

Tonic inhibition was determined by the change in holding current after application of the GABA_AR antagonist picrotoxin (2 or 100 µM dissolved in DMSO). Sensitivity to DZ (1 µM) or L-655,708 (5 µM) was measured by applying the drug to the bath solution and assessing the effect of picrotoxin. Whole cell recordings from CA1 pyramidal cells in WT and 123 mutant mice were made at a holding potential of –60 mV with a high chloride containing internal solution. GABA_B receptors were blocked with the antagonist CGP 55845 (1 µM). Recordings were analyzed with the Mini Analysis software with the detection threshold set 5 times higher than the level of baseline noise. The holding current was measured during 10-ms segments preceding spontaneous events to avoid contamination with phasic currents. The average holding current was calculated in recordings taken after chloride equilibration (baseline), 200 s after DZ application, 120–150 s after picrotoxin application (t1), and 240–300 s after picrotoxin application (t2). For each experiment, changes in holding current after drug application were statistically compared with baseline with paired Student’s t-test. Differences between genotypes were analyzed with one-way ANOVA and Bonferroni’s post hoc multiple comparisons tests.

Drugs

Chemicals were purchased from Sigma/Fluka or Tocris, and DZ was provided by Hoffmann-La Roche. We tested DMSO alone at the appropriate concentrations on IPSCs to rule out direct effects when it was used as a solvent.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION GRANTS ACKNOWLEDGMENTS REFERENCES

 
GABA_A receptors subtypes expressed in CA1 area

The distribution of the four subunit variants (1, 2, 3, 5) contributing to DZ-sensitive GABA_ARs was analyzed in the hippocampal formation by immunoperoxidase staining in P21 WT mice, revealing clear-cut differences in staining intensity and regional distribution (Fig. 1). Although staining intensity cannot be compared across antibodies, comparisons can be made relative to other brain regions. Thus the 2 and 5 subunit exhibited the strongest immunoreactivity in the hippocampal formation, with 2 being more intense in the dentate gyrus than in CA1, whereas the 5 subunit was strongest in CA1, especially in the pyramidal cell layer. Staining for both subunits was homogeneous across dendritic and cell body layers of the hippocampal formation, suggesting the presence of the corresponding receptors on the soma and dendrites of principal cells. As reported previously (Brünig et al. 2002), the 1 subunit antibody strongly labeled a population of interneurons throughout the hippocampal formation and produced only a moderate staining in the dendritic layers of CA1, CA3, and the dentate gyrus, whereas the cell body layers appeared almost devoid of staining. The 3 subunit was expressed at low levels in CA1 and was present in a few isolated interneurons, mainly found in the stratum oriens and in the hilus. The boundary between CA1 and subiculum was particularly evident for the 5 subunit, which is almost not detectable in the latter region.

View larger version (74K): [in this window] [in a new window]   FIG. 1. Differential distribution of subunit variants contributing to diazepam (DZ)-sensitive GABAA receptors in the hippocampal formation. Parasagittal sections from P21 wildtype mice were processed for immunoperoxidase staining with subunit-specific antibodies. 1, 2, and 5 subunits have overlapping, but distinct, distribution in CA1, CA3, and the dentate gyrus (DG). In contrast, staining for the 3 subunit is much weaker in the hippocampus than deep cortical layers (top right) or superior colliculus (bottom right). Both the 1 and 3 subunit label interneurons in various subfields. In subiculum (S), staining for the 1 and 2 subunit is prominent, whereas 3 is weakly stained and 5 absent, marking a sharp boundary with the CA1 subfield. Scale bar, 300 µm.

Differential contribution of 1- and 2-GABA_A receptors to evoked synaptic inhibition in CA1 pyramidal cells

To assess the major GABA_AR subtypes mediating phasic inhibition, the effects of DZ on eIPSCs were compared in WT mice and five lines of mutant mice carrying DZ-insensitive GABA_A receptors (1, 2, 3, 12, and 123 knock-in mice). Current pulses were delivered every 10 s through an extracellular stimulus electrode and their intensity was adjusted to generate eIPSCs of similar amplitude in all experiments (Table 1). There was no difference in rise time and decay time constants of eIPSCs in the different genotypes investigated (Table 1).

View larger version (19K): [in this window] [in a new window]   FIG. 2. Effect of DZ (1 µM) on evoked inhibitory postsynaptic currents (IPSCs) in wild-type (WT) and knock-in mice. A: samples of averaged evoked IPSCs (eIPSCs) obtained before (baseline) and after DZ application in WT and in 123 knock-in mice; note the lack of contribution of 5-GABAARs. B and C: normalized averaged DZ effect on peak amplitude (B) and decay time constant (C) in the different genotypes tested: WT, 1, 2, and 3 knock-in mice showed a significant increase in amplitude and decay time after application of DZ (*P < 0.05, paired Student’s t-test). ANOVA tests indicated a significant difference between genotypes (P < 0.001, n = 66) for both amplitude and decay time after DZ. Post hoc analysis with Bonferroni multiple comparisons tests revealed significantly larger responses in WT and 3 knock-in mice than in 12 and 123 knock-in mice (P 0.001 for amplitude and P 0.026 for decay time). Numbers of experiments per genotype are shown in Table 1.

View larger version (22K): [in this window] [in a new window]   FIG. 3. Effect of DZ (1 µM) on eIPSCs reveals a genotype effect according to the area of stimulation. A: examples of averaged eIPSCs obtained before (baseline) and after DZ application in 1 (left) and 2 (right) knock-in mice according to the area of stimulation: either proximal (top) or distal (bottom) to the cell soma. B: in WT mice, DZ-induced amplitude increase of eIPSCs was similar in proximal and distal evoked responses (n = 14, not significant). In 1 knock-in mice, the DZ-induced amplitude increase of eIPSCs was similar to WT in proximally evoked responses, but significantly reduced in distally evoked responses (P < 0.02, n = 15, unpaired Student’s t-test, equal variances assumed). In 2 knock-in mice, DZ application affected proximally evoked responses significantly more than distally evoked responses (P < 0.002, n = 10, unpaired Student’s t-test, equal variances assumed). *Significant changes from baseline (P < 0.05, paired Student’s t-test).

GABA_A,fast and GABA_A,slow IPSCs represent two distinct modes of phasic synaptic inhibition in CA1 pyramidal cells, best distinguished in recordings of spontaneous events. GABA_A,slow IPSCs have been suggested to involve a mixture of synaptic and perisynaptic or extrasynaptic receptors and to be mediated by a specialized, yet not identified, type of interneuron (Banks et al. 2000). Although 5-GABA_ARs do not seem to be much involved in eIPSCs (Fig. 2), a contribution to GABA_A,slow IPSCs is conceivable. Therefore to determine whether different GABA_ARs mediate GABA_A,fast and GABA_A,slow sIPSCs, the effects of DZ were assessed in 123 knock-in and WT mice, using continuous whole cell voltage-clamp recordings from CA1 and subicular pyramidal cells.

Baseline properties of sIPSCs (peak amplitude, rise time, and decay time constant) did not differ significantly between genotypes and hippocampal regions (CA1, subiculum; Tables 2 and 3). Examples of recordings from a CA1 pyramidal cell are shown in Fig. 4, A and D. sIPSCs were analyzed and subdivided into GABA_A,slow (Fig. 4, B and E) and GABA_A,fast sIPSCs (Fig. 4, C and F).

View larger version (38K): [in this window] [in a new window]   FIG. 4. In CA1, but not subicular, pyramidal cells 5-GABAA receptors (GABAARs) mediate slow spontaneous IPSCs (sIPSCs). Samples of continuous whole cell voltage-clamp recordings from CA1 (A and D) and subicular [sub] (G and K) pyramidal cells in slices obtained from WT (A and G) and 123 (D and K) knock-in mice. Baseline conditions in black; gray traces after bath application of 1 µM DZ. GABAA,slow sIPSCs are marked with asterisks. Corresponding isolated and averaged GABAA,slow sIPSCs (B, E, H, and L) and GABAA,fast sIPSCs (C, F, J, and M) are shown below with baseline conditions drawn in black and sIPSCs after DZ application in gray. GABAA,slow sIPSCs can be unambiguously detected in subicular pyramidal cells. In CA1 pyramidal cells, GABAA,slow sIPSC amplitude was increased in WT and 123 knock-in mice after DZ. In contrast, DZ had no significant effect on the amplitude of GABAA,slow sIPSCs in the subiculum in 123 knock-in mice, whereas it caused a significant increase in WT mice. N and O: average normalized effect on amplitudes and decay time constants of sIPSCs in 3 conditions: DZ in WT (black bars, nCA1 = 7, nsub = 6), DZ in 123 knock-in mice (gray bars, nCA1 = 9, nsub = 7), and the 5-GABAARs selective inverse agonist L-655,708 (5 µM) in WT mice (open bars, nCA1 = 7, nsub = 5). *Significant changes from baseline (DZ, P < 0.03; L-655,708, P < 0.05, paired Student’s t-test).

The effect of DZ in 123 knock-in mice indicates that 5-GABA_ARs participate in GABA_A,slow sIPSCs. However, the increase in GABA_A,slow sIPSCs in CA1 pyramidal cells was significantly larger in WT than in 123 knock-in mice (2-tailed, unpaired Student’s t-test, n_WT = 7, n₁₂₃ = 8), suggesting that 1- and/or 2-GABA_ARs also contribute to GABA_A,slow sIPSCs. In contrast, GABA_A,fast sIPSCs were not affected by DZ in 123 knock-in mice, unlike in WT (Fig. 4, N and O). These findings are consistent with the observations made on eIPSCs and suggest that the contribution of 5-GABA_ARs to phasic inhibition is restricted to GABA_A,slow IPSCs.

As shown in Fig. 1, 5 subunit immunoreactivity is almost undetectable in subiculum, just adjacent to the CA1 area, allowing to characterize GABA_A,slow sIPSCs in neurons lacking 5-GABA_ARs and to validate the findings obtained in 123 knock-in mice. Experiments were performed applying the same protocol used for the CA1 area (Fig. 4, G and K). GABA_A,slow sIPSCs were observed in baseline recordings of subicular pyramidal cells, albeit at a lower frequency than in CA1 (f_Subiculum = 0.039 ± 0.005/s; f_CA1 = 0.099 ± 0.012/s). Bath-application of DZ (1 µM) increased the amplitude and decay time constant of both GABA_A,slow and GABA_A,fast sIPSCs, in cells from WT mice (Fig. 4, H, J, N, and O), whereas in 123 knock-in mice, these values remained unchanged (Fig. 4, L–O). In addition, application of the inverse agonist at the benzodiazepine binding site L-655,708 (5 µM) only affected the amplitude of GABA_A,slow in WT CA1 pyramidal cells (–74.1 ± 12.4 to –58.1 ± 9.5 pA, n = 7, P < 0.05), but not in subicular pyramidal cells (n = 5). Amplitude and kinetics of GABA_A,fast sIPSCs in CA1 and subiculum (n = 7 and n = 5, respectively) remained unchanged. L-655,708 exhibits 50 times higher affinity for 5-GABA_ARs over GABA_ARs containing other subunits (Quirk et al. 1996). These results further confirmed that 5-GABA_ARs are absent in the subiculum and thus are not required for the generation of GABA_A,slow and that these receptors do not contribute to GABA_A,fast sIPSCs.

Because of their slow time-course and sensitivity to GABA reuptake inhibitors, GABA_A,slow IPSCs have been proposed to involve, at least in part, extrasynaptic receptors activated by GABA spillover (Banks et al. 2000). To compare the subunit profile of synaptic and extrasynaptic receptors contributing to GABA_A,slow sIPSCs in CA1 pyramidal cells, we enhanced the spillover component using the selective GABA reuptake inhibitor NO711. Recordings from CA1 pyramidal cells from WT and 123 knock-in mice (Fig. 5, A and D) were obtained under baseline condition, in presence of 2 µM NO711 (Table 4), and finally after application of DZ (1 µM). To prevent tonic activation of GABA_B receptors (Le Feuvre et al. 1997; Scanziani 2000), 1 µM of the GABA_B antagonist CGP55845 was present throughout the experiment.

View larger version (19K): [in this window] [in a new window]   FIG. 5. GABA spillover enhances GABAA,slow currents and increases the contribution 5-GABAARs in the CA1 area. A–F: sample traces of recordings in the CA1 area in slices from 123 knock-in (A) and WT (D) mouse in the presence of 1 µM CGP 55845. After baseline recording (black), 2 µM NO711 was added, and the recording was continued (light gray); finally 1 µM DZ was applied (dark gray). GABAA,slow sIPSCs are marked with an asterisk. Averaged sIPSCs from the 123 knock-in (B) and WT (E) mouse show typical responses of GABAA,slow sIPSCs to NO711 and the subsequent application of DZ. Corresponding isolated and averaged GABAA,fast sIPSCs are depicted in C and F at an enlarged scale to reveal their kinetics. GABA reuptake inhibition increases the amplitude and alters the kinetics only of GABAA,slow sIPSCs. G and H: comparison of GABAA,slow sIPSCs mean amplitudes (G) and mean decay time constants (H) normalized to base-line conditions after addition of 2 µM NO711 (light gray) and 1 µM DZ (dark gray). *Significant difference to preceding condition. Note that DZ in the presence of NO711 now significantly affects decay time constant of GABAA,slow sIPSCs in 123 knock-in mice (n = 8 cells from each genotype, paired Student’s t-test; P < 0.05).

Diazepam-activated tonic inhibition in CA1 pyramidal cells is mediated by 5-GABA_ARs

GABA_A,slow sIPSCs share elements of tonic GABAergic transmission such as involving perisynaptic GABA_ARs. Therefore we assessed which DZ-sensitive GABA_AR subtypes contribute to tonic inhibition in CA1 pyramidal cells. Whole cell voltage-clamp recordings were made in slices from WT and 123 knock-in mice using a high chloride-containing intracellular solution. Using a perfusion rate of 1 ml/min, tonic inhibition was measured as the change in holding current after the application 100 µM picrotoxin in the presence of the GABA_B antagonist CGP 55845 (1 µM). No GABA reuptake inhibitor was used. Under these conditions, application of 100 µM picrotoxin produced an outward shift in the baseline current of 33 ± 4 pA (n = 7) at room temperature and 35 ± 9 pA (n = 5) at physiological temperature. The tonic conductance was mediated by GABA_ARs because it was also blocked by bicuculline (10 µM; data not shown). Because it has been reported that low doses of picrotoxin selectively (e.g., more rapidly) block tonic inhibition (Semyanov et al. 2003), we performed the same experiments using 2 µM picrotoxin and measured the outward shift of the holding current during two distinct time windows: at 120–150 (t1) and 240–300 s (t2) after picrotoxin application. The holding current at t1 changed by 22 ± 5 pA (n = 6; Fig. 6, A and B, middle, and C) and by 33 ± 4 pA at t2 (n = 6; Fig. 6, A and B, right, and C). Root mean square noise decreased from 5.4 ± 0.27 to 4.4 ± 0.27 pA (P < 0.05) at t1 and to 3.9 ± 0.18 pA (not significant) at t2. The amplitude of fast events remained constant at t1, whereas a tendency for smaller GABA_slow amplitudes was observed at that time-point. At t2, both fast and slow events were significantly reduced in amplitude (–44.8 ± 1.1 to –28.6 ± 0.65 pA, n = 6, P < 0.01 in _fast sIPSCs and –61.1 ± 7.5 to –34.4 ± 1.9 pA, n = 6, P < 0.01 in _slow sIPSCs; Fig. 6D). At t1, the frequency and decay time constant of GABA_A,fast sIPSCs were unaltered compared with baseline (3.8 ± 0.26 to 4.0 ± 0.17 Hz, n = 6; and 21.7 ± 0.7 to 23.1 ± 0.9 ms, respectively; Fig. 6, E and F, left trace), whereas the frequency and decay time of GABA_A,slow sIPSCs were significantly reduced (from 0.09 ± 0.01 to 0.06 ± 0.01 Hz; n = 6, P < 0.05 and from 27.6 ± 2.3 to 20.2 ± 2.8 ms, P < 0.05, respectively; Fig. 6, E and F, right). The selective change in decay time and frequency of GABA_A,slow sIPSCs that correlated with the change in holding current provides further indication that slow events and tonic inhibition share a common pool of extrasynaptic and thus most likely 5-containing GABA_ARs (Caraiscos et al. 2004a).

View larger version (31K): [in this window] [in a new window]   FIG. 6. Effect of picrotoxin (PIC) on holding current and GABAA,fast and GABAA,slow sIPSCs. A: representative trace from a continuous voltage-clamp recording at room temperature from a CA1 pyramidal cell of a WT mouse. PIC (2 µM) application to bath solution resulted in a decrease in the inward holding current and a reduction in the frequency and decay time of sIPSCs. The arrows marked t1 and t2 represent the 2 time windows during which holding current and spontaneous events were analyzed. B: magnified 1-s traces of baseline (left), t1 (middle), and t2 (right). GABAA,slow sIPSCs are marked with asterisks. C: average change in holding current measured at t1 and at t2; holding current was significantly reduced at t1 (*P < 0.05, n = 6, paired Student’s t-test). D: average peak amplitude in fast (left) and slow sIPSCs (right); amplitude of fast and slow events was significantly reduced at t2 (*P < 0.006, n = 6, paired Student’s t-test). E: averaged change in frequency of fast (left) and slow (right) sIPSCs. At t1, as the holding current was significantly reduced, fast sIPSC frequency was unaltered (left), whereas frequency of slow sIPSCs was significantly reduced (right; *P < 0.05, n = 6, paired Student’s t-test). ANOVA shows significant differences between baseline, t1, and t2 (n = 6, P 0.01). F: averaged change in decay time of fast (left) and slow (i) sIPSCs. Changes follow the pattern seen in D and E. At t1, fast sIPSC decay time was unaltered (left), whereas decay time of slow sIPSCs was significantly reduced (right; *P < 0.05, n = 6, paired Student’s t-test). ANOVA reveals significant differences between baseline, t1, and t2 (n = 6, P 0.01).

View larger version (15K): [in this window] [in a new window]   FIG. 7. Effect of DZ and L-655,708 on tonic current from CA1 pyramidal cells. A: representative trace segments recorded in a CA1 pyramidal cell from WT mouse. Note the inward drift in the holding current after the application of 1 µM DZ as well as the increase in amplitude and frequency from spontaneous events (middle). After 5 min of PIC (100 µM) application to bath solution, holding current had an outward drift equal to that observed with 2 µM PIC. B: representative traces from a recorded CA1 pyramidal cell in 123 knock-in mouse. After application of 1 µM DZ to the bath solution, holding current exhibited an inward drift, but frequency and amplitude of fast spontaneous events remained unchanged (middle). Outward drift in holding current after application of PIC (100 µM) was similar as in WT mice. C: average shift in holding current in percentage of baseline after drug applications: In WT mice, holding current increased after DZ application by 12 ± 7 pA (n = 10), whereas in 123 knock-in mice, it increased by 7 ± 2 pA (n = 6). Increase in holding current after DZ application showed no significant difference between WT and 123 knock-in mice (unpaired Student’s t-test, not significant). When using 5 µM L-655,708 instead of DZ, no effect on holding current from CA1 pyramidal cells was observed (n = 4). Subsequent application of PIC (100 µM) in these cells caused a change in holding current comparable with DZ experiments. L-655,708 was not used on 123 knock-in mice. *Significant changes from baseline (P < 0.05, 2-tailed paired Student’s t-test).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION GRANTS ACKNOWLEDGMENTS REFERENCES

Technical considerations

The conclusions of this study are derived from a pharmacological distinction of GABA_AR subtypes in knock-in mutant mice. There are two major prerequisites to validate these conclusions. First, the point-mutation should not alter the organization and functional properties of GABA_ARs in mutant mice. Several lines of evidence indicate that this prerequisite is fulfilled: Animals containing one or more point-mutations develop normally, have no overt phenotype and the cellular and subcellular location of GABA_ARs is unchanged (Benke et al. 2004; Mohler et al. 2004; Rudolph and Mohler 2004; Rudolph et al. 1999; Yee et al. 2004). The functional properties of receptors containing mutated subunits are unaltered in vivo (Bacci et al. 2003; Fagiolini et al. 2004; Marowsky et al. 2004). The present results reveal no difference across genotypes in the kinetics of sIPSCs or eIPSCs, and in the stimulus strength required to evoke IPSCs of similar amplitude, confirming that functional properties of GABA_ARs are normal in mutant mice. Second, the effects of DZ on peak amplitude and/or decay time constant should be detectable on all sensitive GABA_AR subtypes and should not be masked by agonist saturation. At least for the 1, 2, and 5-GABA_ARs, these conditions were fulfilled, confirming that mouse CA1 pyramidal cells have an incomplete postsynaptic GABA_AR occupancy at room temperature (Hajos et al. 2000; Perrais and Ropert 1999, 2000). Failure to detect a contribution of 3-GABA_ARs to eIPSCs likely reflects the low abundance of these receptors in CA1 pyramidal cells (Laurie et al. 1992) (Fig. 1). Therefore with the knock-in strategy, the function of GABA_ARs containing specific subunits can be analyzed with great selectivity (Rudolph and Mohler 2004) and without inducing compensatory changes, as occurs in gene deletion experiments. To date, no drugs are available to discriminate between 1-, 2-, and 3-GABA_ARs in vivo with a comparable degree of selectivity. One obvious limitation of the approach lies in the fact that it is limited to DZ-sensitive GABA_ARs.

Somatic and dendritic inputs are mediated differentially by 2- and 1-GABA_ARs

The results from the analysis of eIPSCs are largely in agreement with the known subcellular distribution of GABA_AR subtypes in CA1 pyramidal cells. 2-GABA_ARs located predominantly on the axon-initial segment and on the soma, facing terminals from chandelier cells and from basket cells expressing cholecystokinin, respectively (Nusser et al. 1996; Nyiri et al. 2001). The 1-GABA_ARs on the soma are contacted by parvalbumin-positive basket cell terminals, and they predominate in the dendritic layers of CA1, as seen by immunohistochemistry (Fig. 1) (Brunig et al. 2002). This segregation of different GABA_ARs to different subcellular compartments is of functional relevance, as shown in vivo with the contribution of distinct interneurons signaling through these receptors during specific behavioral states (Klausberger 2003; Klausberger et al. 2002). Additionally, Pouille and Scanziani (2001, 2004) have recently shown that feedforward and feedback inhibition are mediated predominantly by somatic and dendritic receptors, respectively. The differential modulation by DZ of IPSCs evoked by proximal and distal stimulation of inhibitory inputs in 1 and 2 knock-in mice confirms this functional segregation of 2- and 1-GABA_ARs, thereby validating the use of this knock-in model to pharmacologically isolate specific GABA_AR subtypes. The agreement between the present results and the known segregation of 1- and 2-GABA_ARs in CA1 neurons makes it unlikely that the difference in the effects of DZ on amplitude of eIPSCs in 1 and 2 knock-in mice are caused by differential occupancy of receptors located on the soma and dendrites. To explain these results, the degree of occupancy of these receptors should be opposite on the soma and dendrites (1 high on soma and low on dendrites, and vice versa for 2), which seems rather unlikely. Finally, it is well established that 2-GABA_ARs deactivate more slowly than 1-GABA_ARs, as shown in recombinant systems (McClellan and Twyman 1999) and in vivo (Bosman et al. 2005; Goldstein et al. 2002; Vicini et al. 2001). This distinction was not apparent with eIPSCs, reflecting the detection of compound, nonsynchronous events and the variable effects of dendritic filtering on proximal and distal stimulation.

GABA_A,slow IPSCs involve activation of synaptic and perisynaptic receptors

GABA_A,slow IPSCs in CA1 pyramidal cells have been observed in evoked, spontaneous, and miniature IPSCs (Banks et al. 1998; Pearce 1993) GABA_A,slow eIPSCs exhibit the same properties as GABA_A,slow sIPSCs and probably emanate from the same synapse (Banks et al. 1998). Our data extend the previous findings by showing the presence of GABA_A,slow sIPSCs also in the subiculum. It was not possible to determine whether GABA_A,slow IPSCs in CA1 and subicular pyramidal cells are produced by similar mechanisms. The question whether GABA_A,slow IPSCs are produced by one specific cell type common to both brain areas therefore remains unresolved.

Slow events have been proposed to involve phasic activation of perisynaptic or even extrasynaptic GABA_ARs on GABA spillover from the synaptic cleft (Banks et al. 1998). The sensitivity of GABA_A,slow to DZ in 123 knock-in mice and their depression by L-655,708 in WT mice show that 5-GABA_ARs contribute to GABA_A,slow sIPSCs. More importantly, this contribution is amplified and likely dominates the current in conditions of enhanced spillover, because the DZ-induced increase in GABA_A,slow currents became indistinguishable between WT and 123 knock-in mice in presence of a GABA uptake inhibitor. An enhanced diffusion volume covered by the neurotransmitter and the prolonged gating kinetics of the receptors by the elevated concentration of GABA can explain the change in rise and decay time constant of GABA_A,slow in the presence of a GABA uptake inhibitor. Which particular mechanism prevails would most likely be determined by the amount of spilled out GABA.

Our results clearly show that the spillover component of slow phasic currents is mediated by 5-GABA_ARs and nicely agree with the predominantly extrasynaptic distribution of 5-GABA_ARs (Brunig et al. 2002; Caraiscos et al. 2004a,b; Crestani et al. 2002). Interestingly, however, GABA_A,slow sIPSCs also occurred in subiculum where the 5 subunit is virtually absent. The potentiation by DZ in WT mice provides direct evidence that GABA_A,slow sIPSCs there are also mediated by other DZ-sensitive GABA_AR subtypes. These IPSCs are otherwise indistinguishable from those recorded in CA1, suggesting that the function of 5-containing GABA_ARs is taken over by other GABA_ARs with similar peri- and extrasynaptic localization in subicular neurons.

The stronger effect of DZ on GABA_A,slow sIPSCs in WT compared with 123 knock-in mice in CA1 pyramidal cells and the partial effect of L-655,708 (Fig. 4) implicate the involvement of additional GABA_AR subtypes in slow currents, with 1-GABA_ARs as likely coparticipants, because this subunit has been shown sometimes to colocalize with the 5 subunit in the same clusters (Hutcheon et al. 2004). The lack of DZ effect on GABA_A,fast sIPSCs in 123 knock-in mice (Fig. 4) indicates that 5-GABA_ARs do not participate in fast phasic inhibition. We did not detect any spillover component in fast GABAergic sIPSCs in the CA1 region and even the combination of NO711 and DZ on cells from 123 mice revealed no contribution of 5-GABA_ARs to GABA_A,fast. Therefore these GABA_ARs are targeted selectively to sites mediating GABA_A,slow sIPSCs in CA1 pyramidal cells, where they are probably located both in the synaptic cleft and perisynaptically. Using paired intracellular recordings; Thomson et al. (2000) provided indirect pharmacological evidence for the existence of 5-GABA_ARs mediating fast phasic inhibition from bistratified cells, which innervate distal pyramidal cell dendrites. The failure to detect these receptors in point-mutated mice might reflect the minor contribution of these events to the overall number of sIPSCs received by pyramidal cells. Banks et al. (1998) provided evidence for blockade of GABA_A,fast sIPSCs by furosemide, suggesting a contribution of 4-GABA_ARs to these events. We did not assess the contribution of 4-GABA_ARs, which are DZ-insensitive, but their low expression level in CA1 area (Pirker et al. 2000) makes it unlikely that they play a major role in mediating GABA_A,fast sIPSCs.

Our findings emphasize the proposed role of 5-GABA_ARs as detectors for extrasynaptic GABA (Caraiscos et al. 2004a; Glykys and Mody 2006), and reveal their dual assignment to the generation of slow GABAergic events and the mediation of DZ-sensitive tonic inhibition. In both cases, GABA is not confined to synaptic clefts and the extent to which the neurotransmitter release underlying GABA_A,slow sIPSCs contributes to ambient GABA remains to be examined.

Tonic inhibition

Thus far we have shown that in the CA1 area, GABA_A,slow events partially arise from the activation of 5-GABA_ARs receptors that are predominantly located extrasynaptically. Because GABA_A,slow sIPSCs are rare, their early block by the activity-dependent antagonist picrotoxin, which parallels the early decrease of tonic GABAergic currents, most likely reflects crosstalk between the receptor populations mediating these two types of GABAergic currents. The pronounced sensitivity of tonic inhibition and GABA_A,slow for picrotoxin is in line with a higher opening probability of tonically activated, extrasynaptic receptors.

The increase in holding current after application of DZ in both WT and 123 knock-in mice confirms that 5-GABA_ARs can generate both tonic inhibition and slow events in CA1 pyramidal cells. Several lines of evidence indicate that most 5-GABA_ARs contain the 3 subunit in pyramidal neurons and 3-containing GABA_ARs are present in extrasynaptic regions (Pirker et al. 2000). It would be of major interest to know whether these receptors have the required high affinity for GABA. The fact that in our experiments the inverse agonist L-655,708 in contrast to DZ failed to exert any effect on the holding current of CA1 pyramidal cells raises questions about the role of 5-GABA_ARs detecting ambient GABA concentrations in the absence of DZ. DZ increases the probability of GABA_ARs to be activated after GABA binding. This can lead to tonic opening of receptors under low ambient GABA concentration. Consequently, an inverse agonist fails to produce any effect on previously not activated receptors. Scimemi et al. (2005) have been the first to address the dependency of 5-GABA_ARs activation on GABA concentrations. The authors found that L-655,708 decreases the holding current only when ambient GABA concentrations are elevated. Under physiological conditions, synchronous neural network activity might lead to such heightened ambient GABA concentration and subsequent activation of extracellular 5-GABA_ARs (Towers et al. 2004). Nevertheless our findings are relevant to explain the physiological effects of DZ, because this drug experiences widespread pharmaceutical use.

Significance of the work

We have shown that the subunit composition of GABA_ARs not only affects their subcellular distribution but also dictates their functional role. The clear segregation of GABA_AR subtypes demonstrated here shows the complex organization of the inhibitory system at the molecular level and provides a compelling explanation for the specific effects of different GABA_ARs in distinct behaviorally significant signaling pathways.

	GRANTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION GRANTS ACKNOWLEDGMENTS REFERENCES

 
G. A. Prenosil, E. M. Schneider Gasser, and K. E. Vogt were supported by Swiss National Science Foundation Grant 631–066012.

	ACKNOWLEDGMENTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION GRANTS ACKNOWLEDGMENTS REFERENCES

 
We thank C. Sidler and F. Parpan for excellent technical assistance.

	FOOTNOTES

 
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Address for reprint requests and other correspondence: K. E. Vogt, Inst. of Pharmacology and Toxicology, Univ. of Zurich, Winterthurerstrasse 190, 8057 Zuric