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

This Article Abstract Full Text (PDF) All Versions of this Article: 90/1/89    most recent 00612.2002v1 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 ISI Web of Science 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 Thio, L. L. Articles by Yamada, K. Search for Related Content PubMed PubMed Citation Articles by Thio, L. L. Articles by Yamada, K.

Benzodiazepines Block ₂-Containing Inhibitory Glycine Receptors in Embryonic Mouse Hippocampal Neurons

Departments of ¹Neurology, ²Pediatrics, and ³Psychiatry and the ⁴Center for the Study of Nervous System Injury, Washington University School of Medicine; ⁵Division of Pediatric Neurology and the ⁶Pediatric Epilepsy Center, St. Louis Children's Hospital, St. Louis, Missouri 63110

Submitted 29 July 2002; accepted in final form 18 March 2003

	ABSTRACT

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION ACKNOWLEDGMENTS REFERENCES

 
Inhibitory glycine receptors (GlyRs) in the mammalian cortex probably contribute to brain development and to maintaining tonic inhibition. Given their presence throughout the cortex, their modulation likely has important physiological consequences. Although benzodiazepines potentiate -aminobutyric acid_A receptors (GABA_ARs), they may also modulate GlyRs because binding studies initially suggested that they act at GlyRs. Furthermore, their diminished ability to potentiate neonatal GABA_ARs suggests that they may exert their beneficial clinical effects at another site in the developing brain. Therefore we examined the effect of benzodiazepines on whole cell currents mediated by GlyRs in cultured embryonic mouse hippocampal neurons. First, we determined the GlyR subunit composition in this preparation. Glycine, -alanine, and taurine activate strychnine-sensitive chloride currents in a dose-dependent manner. Maximal concentrations of the three agonists produce equal, nonadditive responses as expected of full agonists. The pharmacological properties of the GlyR currents including their pattern of modulation by picrotoxinin, picrotin, and tropisetron indicate that GlyRs consist of ₂ heteromers and ₂ homomers. Reverse transcriptase polymerase chain reaction (RTPCR) studies confirmed the presence of ₂ and subunits. Second, we found that micromolar concentrations of some benzodiazepines, including chlordiazepoxide and nitrazepam, inhibit GlyR currents. Nitrazepam inhibition of GlyRs is noncompetitive, is not voltage dependent, and does not reflect enhanced desensitization. Thus benzodiazepines allosterically inhibit ₂-containing GlyRs in embryonic mouse hippocampal neurons via a "low"-affinity site.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION ACKNOWLEDGMENTS REFERENCES

 
The inhibitory glycine receptor (GlyR) likely is a pentameric ligand-gated ion channel composed of three and two subunits (reviewed in Legendre 2001; Rajendra et al. 1997). The subunit has four isoforms (_1–4) while the subunit only has one. Although subunit isoforms show both developmental and regional variation, the CNS expresses GlyRs diffusely throughout life. Along with -aminobutyric acid_A receptors (GABA_ARs), they mediate fast inhibitory synaptic transmission in the brain stem and spinal cord. While their function in the cortex remains unclear, they may have a role in cortical development (Flint et al. 1998) and may provide tonic inhibition rather than participating in fast inhibitory synaptic transmission (Mori et al. 2002). Nevertheless, their presence in the mammalian forebrain throughout life is well established based on physiological (Chattipakorn and McMahon 2002; Krishtal et al. 1988; Ye et al. 1999), immunocytochemical (Becker et al. 1993) and in situ hybridization (Malosio et al. 1991) studies. Thus GlyR modulation is likely to have profound effects on cortical function.

Various agents positively and/or negatively modulate GlyRs. For example, ethanol (Ye et al. 2001a,b), steroids (Maksay et al. 2001), dihydropyridines (Chesnoy-Marchais and Cathala 2001), tropeines (Supplisson and Chesnoy-Marchais 2000), and zinc (Chattipakorn and McMahon 2002; Laube et al. 1995) modulate GlyRs positively and negatively. The type of modulation exerted can depend on the subunit composition as with GABA_ARs. Indeed, many modulators of GABA_ARs are also modulators of GlyRs as might be expected from the homology between these receptors, which both belong to the cysteine loop superfamily of ligand-gated ion channels (Karlin and Akabas 1995).

Benzodiazepines are a clinically important class of GABA_ARs potentiators frequently used to treat a variety of neuropsychiatric disorders including anxiety, spasticity, and seizures in patients of all ages. Although benzodiazepines were initially hypothesized to exert their clinical effects through GlyRs because they displace strychnine in binding assays (Young et al. 1974 but see Hunt and Raynaud 1977), subsequent electrophysiological studies demonstrated no effect of benzodiazepines on GlyRs (Choi et al. 1981; Macdonald and Barker 1978). However, this discrepancy between the binding and the electrophysiological studies with respect to GlyRs remains unresolved. In addition, benzodiazepine potentiation of GABA_ARs in the neonatal brain is much less pronounced than in the adult (Brooks-Kayal et al. 2001; Rovira and Ben Ari 1993; Zhang et al. 1993). This observation raises the possibility that other receptors, such as GlyRs, may mediate the effects of benzodiazepines in the neonate. Therefore we tested the hypothesis that benzodiazepines modulate GlyRs.

We examined the effect of benzodiazepines on GlyRs in cultured embryonic mouse hippocampal neurons using whole cell patch-clamp recordings. First, we demonstrated that cultured mouse embryonic hippocampal neurons express ₂ heteromeric and ₂ homomeric GlyRs using pharmacological and reverse transcription-polymerase chain reaction (RT-PCR) techniques. We found that several benzodiazepines inhibit ₂-containing GlyRs and that nitrazepam acts noncompetitively.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION ACKNOWLEDGMENTS REFERENCES

 
Embryonic mouse hippocampal cultures

Timed pregnant Swiss Webster mice (Taconic Farms, Germantown, NY) at day 16 of gestation were anesthetized with 5% halothane and killed by cervical dislocation. Embryos were removed by cesarean section and decapitated. The brains of five embryos were removed. The hippocampi were dissected and sliced in Leibovitz's L-15 media (Gibco BRL, Grand Island, NY) containing 0.4 mg/ml bovine serum albumin before being enzymatically digested at 37°C for 20 min in the same solution to which 1 mg/ml papain was added. Then the slices were gently triturated with a glass pipette in plating media containing 1x minimum essential medium with Earle's salts (Gibco BRL), 10% Nu-Serum (Collaborative Biomedical Products, Bedford, MA), 5 mg/ml glucose, 2.2 mg/ml NaHCO₃, 20 units/ml penicillin, and 20 µg/ml streptomycin. The resulting single cell suspension was centrifuged for 5 min at 1,000 rpm in the presence of 0.25 mg/ml BSA and 0.25 mg/ml trypsin inhibitor. The pellet was resuspended in plating media and then plated on a monolayer of passaged cortical astrocytes at a density of 2.5 x 10⁵ cells/ml. To inhibit glial proliferation, 5-fluoro-2'-deoxyuridine and uridine were added to a final concentration of 15 and 35 µg/ml, respectively, at 48 h. Recordings were made from neurons incubated for 4–14 days.

Whole cell patch-clamp electrophysiology

Voltage-clamp recordings from embryonic mouse hippocampal neurons were made using an Axopatch 200A amplifier (Axon Instruments, Union City, CA) in the whole cell patch-clamp mode. All experiments were performed at a holding potential of –65 mV unless otherwise indicated. All holding potentials were corrected for junction potentials, which were measured empirically for each pipette solution (Neher 1992). Ramp current-voltage relationships were obtained by subjecting neurons to a voltage ramp of 0.1 V/s. Series resistance compensation was set at 90–95%, and the 4-pole low-pass filter on the amplifier was set at 1–5 kHz.

Neurons were bathed in an extracellular solution containing (in mM) 140 NaCl, 5 KCl, 1.5 CaCl₂, 1 MgCl₂, 10 D-glucose, 2.5 x 10^–4 tetrodotoxin, and 10 N-[2-hydroxyethyl]piperazine-N'-[2-ethanesulfonic acid] (HEPES) (pH 7.35–7.38). Patch pipettes had resistances of 2–4 M after fire-polishing when filled with a solution containing (in mM) 140 CsCl, 4 NaCl, 0.5 CaCl₂, 5 ethylene glycol-bis(2-aminoethylether)-N,N,N',N'-tetraacetic acid (EGTA), and 10 HEPES (pH 7.36). In some experiments, the pipette solution contained 140 mM Cs methanesulfonate (CsCH₃SO₃) or 70 mM CsCl + 70 mM CsCH₃SO₃ rather than 140 mM CsCl. Solutions had osmolarities of 283–293 mosM.

The bath was continuously perfused with extracellular solution at 0.5 ml/min. The bath temperature was monitored continuously with a thermister and ranged from 22 to 24°C. In addition, the neuron being studied was continuously perfused with either extracellular solution or a test solution at 2 ml/min using a multibarrel, gravity-driven, flowtube system developed for studying rapidly desensitizing glutamate currents. Drugs were dissolved in the extracellular solution and applied with this system. Drugs, which were not water soluble, were first dissolved in dimethyl sulfoxide (DMSO) before being added to the extracellular solution. When DMSO was used, the same DMSO concentration was present in all solutions. The highest final concentration of DMSO used was 0.1% except for those experiments using chlordiazepoxide concentrations >500 µM in which the final DMSO concentration was 2%. In the presence of 0.1% DMSO, 50 µM glycine currents were 94 ± 4% (mean ± SE, n = 4) of control. Chlordiazepoxide concentrations >500 µM were obtained by diluting a 100 mM stock solution made with 100% DMSO. The experiments were performed before the chlordiazepoxide precipitated out of solution, which occurred 30 min after it was prepared. In the presence of 2% DMSO, 50 µM glycine currents were 89 ± 7% (n = 3) of control.

Data analysis

Whole cell currents were digitized at 5–10 kHz using pCLAMP 8 (Axon Instruments) and were analyzed using pCLAMP 8, Origin 6 and 7 (OriginLab, Northampton, MA), and Microsoft Excel 2000 (Microsoft, Redmond, WA). Applications of agonist in the absence and presence of a modulator were interleaved. Experimental peak currents were compared with the average of the bracketing control peak currents. Control currents generally changed by <10%. Typically, two to three trials from each neuron were included in the analysis. Exponential processes were fit to the function I(t)

(1)

where t is time, A_j is the amplitude of the jth component having a time constant _j, and C is a constant. Exponential fits to desensitizing currents started at the point the peak current had decayed by 5%. k was determined using an F test and by visual inspection of the residual plots. The weighted mean time constant, _mean, was calculated by

(2)

where

(3)

Agonist dose-response curves were fit to the logistic equation

(4)

where R(A) is the response to a given concentration of the agonist [A], R_max is the response to a saturating concentration of the agonist, EC₅₀ is the concentration producing a half-maximal response, and n is the Hill coefficient. The Hill coefficient was fixed during the fit as the slope of a log{R(A)/[R_max – R(A)]} versus log[A] plot. Antagonist dose-response curves were fit to the logistic equation

(5)

where R(B) is the response of the agonist in the presence of an antagonist concentration [B], R_peak is the response to the agonist in the absence of the antagonist, IC₅₀ is the concentration of the antagonist producing a half-maximal inhibition, and n is the Hill coefficient. The Hill coefficients for antagonist dose-response plots were determined in a manner analogous to that used for the agonist dose-response plots. The Levenberg-Marquardt algorithm was used for all nonlinear fits.

Data from the fits are presented as mean ± SE along with 95% confidence intervals when necessary, whereas all other data are presented as mean ± SE with n being the number of cells examined. Means from two samples were compared using a t-test while means from three or more samples were compared using a one-way ANOVA. Statistical significance was set at P < 0.05.

RT-PCR

Previously published methods (Heck et al. 1997) were adapted as follows for RT-PCR detection of GlyR subunit and gephyrin mRNA. RNA from embryonic mouse hippocampal neurons cultured for 7 days and adult (60-day-old) mouse brain was harvested and quantified spectrophotometrically. Equal quantities of hippocampal culture RNA and whole brain RNA were converted to cDNA in single reactions. Then the cDNA was aliquotted for each specific PCR reaction. Individual primer pairs designated P1 and P2 by Heck et al. (1997) specific for GlyR ₁, ₂, ₃, or subunits or gephyrin were used. First-strand synthesis was primed with oligo-dT. PCR was performed at an annealing temperature of 56°C and amplified with a modified Taq polymerase (Expand, Roche Applied Science, Indianapolis, IN) according to the manufacturer's protocol. Products were electrophoretically separated on a 1.5% agarose gel and stained with ethidium bromide.

Materials

Flumazenil was a gift from F. Hoffman-La Roche (Basel, Switzerland), RO 15-4513 was a gift from Charles F. Zorumski (Washington University, St. Louis, MO), and GYKI 52466 was a gift from Istvan Tarnawa (IVAX Drug Research Institute, Budapest, Hungary). All other chemicals were obtained from Sigma (St. Louis, MO) unless stated otherwise.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION ACKNOWLEDGMENTS REFERENCES

 
Glycine activates both slowly and rapidly desensitizing currents

We classified the responses of cultured embryonic mouse hippocampal neurons into three categories. Of 268 neurons exposed to saturating (100–1,000 µM, see following text) glycine concentrations, 246 (92%) showed a slowly desensitizing current (Fig. 1A), whereas 20 (7%) showed no detectable response, 2 (<1%) showed only a rapidly desensitizing response (Fig. 1B), and 2 (<1%) showed both a slowly and rapidly desensitizing response. Because two neurons showed a rapidly desensitizing current with subsaturating glycine concentrations, a total of six or <1% of all neurons (6/937) examined in this study showed a rapidly desensitizing current.

View larger version (26K): [in this window] [in a new window]   FIG. 1. Glycine evokes slowly (A) and rapidly (B) desensitizing currents in cultured embryonic mouse hippocampal neurons. A1: 300 µM glycine induced a slowly desensitizing current (—) in a neuron voltage-clamped at –65 mV. , a 2-exponential fit with time constants of 0.68 and 3.4 s accounting for 41 and 59% of the decay, respectively. Only every 1,800th point of the fitted curve is plotted. A2: 5 superimposed traces showing the current evoked by a 20-s application of 300 µM glycine followed by a 1-s test pulse of 300 µM glycine applied 1, 3.5, 11, 31, and 51 s later () in another neuron held at –65 mV. A3: the slowly desensitizing glycine current recovers along a 2-exponential time course. , the mean test peak current expressed as a percentage of the control peak current as determined from experiments as shown in A2 (n = 4–16). Error bars represent SE. —, a 2-exponential function having time constants of 1.6 ± 0.6 and 17 ± 3.2 s accounting for 40 and 60% of the recovery, respectively. The weighted mean time constant is 11 s. B: 100 µM glycine induced a rapidly desensitizing current (—) in another neuron held at –65 mV. , a single exponential fit with a time constant of 43 ms. Only every 250th point of the fitted curve is plotted. Note the difference in the time scale from A1. Bars in all figures indicate the duration of agonist application.

The slowly desensitizing current elicited by a 20-s application of 300 µM glycine decayed to 92 ± 1% (n = 12) of its peak along a two-exponential time course in 92% (11/12) of neurons. The faster time constant was 1.0 ± 0.9 s (340–1,800 ms, n = 11) and the slower time constant was 5.0 ± 0.5s (2.6–9.1s, n = 12) with the faster time constant accounting for 48 ± 8% (14–78%, n = 11) of the desensitization (Fig. 1A1). We have not identified the reason for the variability in the desensitization parameters, but it may reflect the variability in GlyR density (Legendre et al. 2002) or phosphorylation (Gentet and Clements 2002), two factors that modulate GlyR desensitization. Recovery followed a two-exponential time course with a weighted mean time constant of 11s (Fig. 1A2 and 1A3). Both the rate and degree of desensitization depended on glycine concentration. With 10 µM glycine, the current decayed to 75 ± 3% (n = 6) of its peak along a single exponential in 83% (5/6) neurons with a weighted mean time constant of 14.9 ± 2.8 s (n = 6). Thus glycine currents in >90% of the neurons studied desensitized with kinetics similar to those in other neurons (Akaike and Kaneda 1989; Fatima-Shad and Barry 1995; Krishtal et al. 1988; Melnick and Baev 1993) and to the slower components of desensitization in outside-out patches from brain stem neurons (Harty and Manis 1998; Legendre 1998).

The rapidly desensitizing current elicited by 20–100 µM glycine decayed to 83 ± 7% (n = 6) of its peak along a single-exponential time course (Fig. 1B). When elicited by 100 µM glycine, the time constant was 39 ± 5 ms (n = 4), which is similar to the faster components of desensitization in outside-out patches (Harty and Manis 1998; Legendre 1998). The rapidly desensitizing current did not recover. We did not study the rapidly desensitizing current further because it was rare and did not recover under the conditions used in this study. The results presented in the following text apply only to the slowly desensitizing current. Specifically, subsequent references to fast and slow components of desensitization refer only to the slowly desensitizing current.

Glycine, -alanine, and taurine are full agonists of a strychnine-sensitive chloride conductance

Glycine, -alanine, and taurine evoked dose-dependent currents (Fig. 2A). Table 1 shows the EC₅₀ and Hill coefficient for each agonist, which are similar to previous reports using dissociated neurons (Krishtal et al. 1988; Lewis et al. 1991; McCool and Botting 2000). Because we derived the dose-response curve for each agonist by normalizing peak currents to the peak current elicited by 100 µM glycine, we can directly compare the maximum responses. All three agonists evoked comparable maximal responses with the order of potency being glycine > -alanine > taurine (Fig. 2A2). Prior studies using both native neurons and expression systems reported the same order of potency (Harvey et al. 2000; Krishtal et al. 1988; Lewis et al. 1991; McCool and Botting 2000; Schmieden et al. 1992, 1999).

View larger version (27K): [in this window] [in a new window]   FIG. 2. Properties of glycine, -alanine, and taurine currents in cultured embryonic mouse hippocampal neurons. A: glycine, -alanine, and taurine elicited dose-dependent currents. A1: currents evoked by increasing concentrations of glycine in a neuron voltage-clamped at –65 mV. A2: dose-response curves for currents evoked by glycine (), -alanine (), and taurine (*). Data points represent the mean peak current for each agonist concentration tested expressed as a percentage of the peak current elicited by 100 µM glycine in the same neuron at –65 mV (n = 3–8). When larger than the symbol, error bars show SE. — (glycine), · · · (-alanine), and - - - (taurine) show fits to a logistic equation. The fit yielded an EC50 of 57 ± 8.0 µM using an n of 1.2 ± 0.2 for glycine. Respective values for -alanine were 180 ± 34 µM and 1.2 ± 0.2. The values for taurine were 580 ± 52 µM and 1.6 ± 0.3. The maximum responses of the 3 agonists were compared using the 2 highest concentrations tested for each agonist. They were not significantly different (ANOVA, P = 0.06). Raw current amplitude ranges for saturating glycine, -alanine, and taurine concentrations were 4,020–11,400, 3,580–14,300, and 1,950–11,590 pA, respectively. B: -alanine gates a Cl– conductance. B1: current-voltage relationship for the 150 µM -alanine current in a neuron voltage-clamped at potentials between –125 and +10 mV in 15 mV increments using the CsCH3SO3 internal solution. Inset: raw current traces. B2: semilogarithmic plot of the reversal potential for 150 µM -alanine currents vs. pipette chloride concentration. , the mean reversal potential determined from plots as shown in B1 (n = 4–6). SE was smaller than the symbol for all points. · · ·, a linear fit with a slope of 49 ± 4 mV. C: strychnine reversibly inhibited glycine, -alanine, and taurine currents. C1: currents evoked by 600 µM taurine (control), 600 µM taurine +5 nM strychnine (+strych) after preapplying 5 nM strychnine for 40 s, and 600 µM taurine alone after washout of strychnine (recovery) in a neuron voltage-clamped at –65 mV. The spikes in the traces prior to the onset of drug application in this and all other figures are an artifact of the drug perfusion system. C2: dose-response curves for strychnine inhibition of currents induced by EC50 concentrations of glycine (50 µM, ), -alanine (150 µM, ), and taurine (600 µM, *) at –65 mV. Strychnine was preapplied as in C1. Points represent mean peak currents in the presence of varying concentrations of strychnine expressed as a percentage of the peak current evoked by the agonist alone (n = 4–5). When larger than the symbol, error bars show the SE. — (glycine), · · · (-alanine), and - - - (taurine) curves show fits to a logistic equation. The fit yielded an IC50 of 8.4 ± 2.1 nM using an n of 1.0 ± 0.3 for glycine. Respective values for -alanine were 8.1 ± 2.3 nM and 0.8 ± 0.1. The values for taurine were 3.4 ± 0.9 nM and 0.8 ± 0.2. D: bicuculline reversibly blocked glycine currents. Currents evoked by 50 µM glycine (control), 50 µM glycine + 100 µM bicuculline (+Bicuc) after preapplying 100 µM bicuculline for 15 s, and 50 µM glycine alone after washout of bicuculline (recovery) in a neuron voltage-clamped at –65 mV. E: saturating concentrations of -alanine and taurine alone or combined with a saturating concentration of glycine produced currents equal in amplitude to a saturating concentration of glycine. Peak currents induced by 5 mM taurine, 1 mM -alanine, 300 µM glycine + 5 mM taurine, and 300 µM glycine + 1 mM -alanine expressed as a percentage of the peak current induced by 300 µM glycine alone at –65 mV (n = 3). Peak currents were not significantly different (ANOVA, P = 0.46). and error bars represent the mean and SE.

To prove that these agonists activate a Cl^– conductance, we obtained current-voltage plots (Fig. 2B1) for each agonist using three different internal [Cl^–]. When the internal and external [Cl^–] were near 140 mM, the current-voltage plots were nearly linear as confirmed by the ratio of the chord conductance at +55 mV to that at –65 mV being near one for each agonist (Table 1). Semi-logarithmic plots showed that the reversal potentials for each agonist vary linearly with the logarithm of the internal [Cl^–] with slopes near 50 mV (Fig. 2B2, Table 1). These values are less than the expected value of 58.6 mV for a chloride selective conductance. Because the deviation occurs primarily at low internal [Cl^–], one explanation for this difference may be intracellular Cl^– accumulation by the K⁺-Cl^– cotransporter at low internal [Cl^–] (DeFazio et al. 2000). Theoretically, CH₃SO₃^– permeation through the GlyR channel may also contribute.

Strychnine and bicuculline are the antagonists often used to define GlyRs and GABA_ARs, respectively. Strychnine reversibly inhibited the current elicited by an EC₅₀ concentration of each agonist with a low nanomolar IC₅₀ when strychnine was preapplied (Fig. 2C, Table 1). Strychnine was more potent when preapplied as previously reported (McCool and Botting 2000). Bicuculline also reversibly inhibited the current elicited by an EC₅₀ concentration of each agonist with a low micromolar or greater IC₅₀ (Fig. 2D, Table 1). It was significantly more potent against glycine than -alanine and taurine currents (Table 1). GlyRs in other preparations have a similar pharmacological profile (Harvey et al. 2000; Jonas et al. 1998; Krishtal et al. 1988; Lewis et al. 1991; McCool and Botting 2000; O'Brien and Berger 1999; Schmieden et al. 1992, 1999; Shirasaki et al. 1991), which is distinct from GABA_ARs (Table 1).

The preceding results suggest that glycine, -alanine, and taurine are full agonists at GlyRs. To provide further evidence for this hypothesis, we compared peak currents elicited by saturating concentrations of each agonist alone to those elicited by a saturating concentration of glycine + -alanine and glycine + taurine. Saturating concentrations of -alanine, taurine, glycine + -alanine, and glycine + taurine produced responses equal to glycine alone (Fig. 2E). Taken together, these data demonstrate that glycine, -alanine, and taurine are full agonists at GlyRs in cultured embryonic mouse hippocampal neurons.

Cultured embryonic mouse hippocampal neurons express ₂-containing GlyRs

In situ hybridization studies indicate that both ₂ and subunits are expressed in the embryonic hippocampus with ₃ being expressed postnatally (Malosio et al. 1991). ₁ is not expressed in the hippocampus at any age. ₄ is not expressed in the brain at high levels (Harvey et al. 2000; Matzenbach et al. 1994). Three variants of ₂ exist: ₂A, ₂B, and (Kuhse et al. 1990, 1991). ₂A and ₂B are formed by alternative splicing. differs from ₂A by one amino acid, and the two may be allelic variations of the same gene. In the hippocampus, ₂B transcripts decline postnatally while ₂A transcripts persist. is only expressed in neonates. Thus we expected to find ₂ and subunits in our preparation. The next series of experiments tested this hypothesis.

GlyR subunits when expressed alone or in pairs form functional channels. Although the pharmacological profile of the glycine current does not correlate with a unique subunit composition, it eliminates some possibilities. We will begin by considering the results already presented. First, strychnine is not a potent inhibitor of and homomers (Grenningloh et al. 1990; Kuhse et al. 1990). Second, strychnine is not a potent inhibitor of ₂ homomers when activated by taurine (Schmieden et al. 1992). Third, homomers require 100-fold higher agonist concentrations for activation than subunit containing GlyRs (Grenningloh et al. 1990). In contrast, we found that strychnine is a potent antagonist of GlyR currents regardless of the agonist used. In addition, the EC₅₀ for each agonist is comparable or lower than that obtained in other preparations known or thought to contain subunits (Harvey et al. 2000; Jonas et al. 1998; Krishtal et al. 1988; Lewis et al. 1991; McCool and Botting 2000; Schmieden et al. 1992, 1999). These results suggest that the majority of the GlyRs in our preparation are not ₂, , or homomers.

We examined the effect of picrotoxinin to better define the subunit composition (Fig. 3). Picrotoxinin blocks ₁, ₂, and ₃ homomers with an IC₅₀ of 5–9 µM, both ₁ and ₃ heteromers with an IC₅₀ over 1 mM, and ₂ heteromers with an IC₅₀ of 20 or 300 µM using glycine concentrations at or above the EC₅₀ (Bloomenthal et al. 1994; McCool and Farroni 2001; Pribilla et al. 1992). In our neurons, picrotoxinin reversibly inhibited glycine currents with an IC₅₀ that depended on glycine concentration as shown previously (Lynch et al. 1995; Yoon et al. 1998) (Fig. 3A). Picrotoxinin inhibited 20 µM (EC₂₀), 50 µM (EC₅₀), and 300 µM (EC₁₀₀, see Fig. 2A2 and Table 1) glycine currents with a corresponding IC₅₀ and Hill coefficient of 21 ± 5 µM and 0.9 ± 0.1, 82 ± 13 µM and 0.9 ± 0.1, and 190 ± 40 µM and 0.7 ± 0.1, respectively. Notably, picrotoxinin does not block 20 µM glycine currents with an IC₅₀ <10 µM. In addition, the picrotoxinin inhibition curves are not obviously biphasic. Yoon et al. (1998) obtained biphasic curves that they attributed to the presence of two types of GlyRs each having a different subunit composition and EC₅₀ for activation by glycine. However, a biphasic curve would not be seen if the two receptor types have a similar EC₅₀. Nevertheless, this intermediate sensitivity to picrotoxinin suggests the presence of both ₂ heteromers and ₂ homomers (Chattipakorn and McMahon 2002).

View larger version (30K): [in this window] [in a new window]   FIG. 3. GlyRs contain 2 subunits based on their modulation by picrotoxinin, picrotin, and tropisetron. A1–C1: currents evoked by glycine alone (control); glycine + 100 µM picrotoxinin (+pctxnin in A1), 1 mM picrotin (+picrotin in B1), or 100 nM tropisetron (+tropisetron in C1) after preapplying the modulator tested for 30 s; and glycine alone after washout of the modulator (recovery) in neurons voltage-clamped at –65 mV. 50 µM glycine (A1 and B1), 20 µM glycine (C1). A2–C2: dose-response curves for the modulation of glycine currents by picrotoxinin (A2), picrotin (B2), or tropisetron (C2) at –65 mV. Symbols show the mean peak current in the presence of various modulator concentrations expressed as a percentage of the peak current evoked by 20 µM (A2, ), 50 µM (A2, ), 300 µM (A2, *), 50 µM (B2, ), or 20 µM (C2, ) glycine alone (n = 4–18). Error bars represent SE. · · · (20 µM glycine), — (50 µM glycine), and - - - (300 µM glycine) curves in A2 and B2 illustrate the fit to a logistic equation. For picrotoxinin inhibition, the IC50 using 20 µM glycine was 21 ± 5 µM when n was 0.9 ± 0.1; the IC50 using 50 µM glycine was 82 ± 13 µM when n was 0.9 ± 0.1; and the IC50 using 300 µM glycine was 190 ± 40 µM when n was 0.7 ± 0.1. For picrotin inhibition, the IC50 was 330 ± 190 µM when n was 0.6 ± 0.2. — in C2 is a spline fit to the data demonstrating potentiation and inhibition by tropisetron.

To provide further evidence for the presence of subunits, we examined the effect of picrotin and tropisetron. Picrotin (100 µM) reduces ₁ and ₃ homomer-mediated responses by 90%, whereas it reduces ₁ and ₃ heteromer mediated responses by ≤25% (Steinbach et al. 2000). ₂ subunits were not examined in this study. We found that picrotin reversibly inhibited 50 µM glycine currents with an IC₅₀ of 330 ± 190 µM and a Hill coefficient of 0.6 ± 0.2 (Fig. 3B). The relatively weak inhibition by picrotin suggests that subunits exist in our neurons. Tropisetron, a tropeine, biphasically modulates ₁ homomers, ₁ heteromers, and ₂ heteromers while it only inhibits ₂ homomers (Supplisson and Chesnoy-Marchais 2000). Tropisetron concentrations <1 µM augmented, whereas those >10 µM inhibited 20 µM glycine (EC₂₀, see Fig. 2A2) currents (Fig. 3C). We used an EC₂₀ concentration to facilitate observing both potentiation and block. This finding suggests that ₂ homomers do not predominate.

Therefore the pharmacological profile indicates that embryonic hippocampal neurons contain both ₂ and subunits. While both ₂ heteromers and ₂ homomers exist, our results also indicate that GlyRs in these neurons are primarily ₂ heteromers. The presence of ₄ cannot be excluded (Harvey et al. 2000).

Although the pharmacological profile of glycine currents has correctly identified the GlyR subunit composition in other preparations (McCool and Farroni 2001), the conclusion that cultured embryonic mouse hippocampal neurons express ₂-containing GlyRs assumes that native and expressed GlyRs have the same pharmacological properties. To confirm the presence of ₂ and subunits in our cultures, we turned to RT-PCR (Heck et al. 1997). Our cultures expressed mRNA transcripts for ₂ subunits, subunits, and gephyrin but not ₁ and ₃ subunits (Fig. 4). As expected, the adult mouse brain expressed ₁, ₂, ₃, and subunit as well as gephyrin mRNA.

View larger version (29K): [in this window] [in a new window]   FIG. 4. Cultured embryonic mouse hippocampal cells express 2 and GlyR subunit and gephyrin mRNA. Concurrent RT-PCR of mRNA from cultured embryonic hippocampal neurons (left lanes) and adult mouse brain (right lanes) was performed for GlyR 1, 2, 3, and subunits as well as for gephyrin (geph). Numbers to the left indicate the migration of the marker DNA (middle lane, designated M). Amplification of adult mouse brain mRNA produced products of the expected sizes: 1, 300 and 324 bp; 2, 330 bp; 3, 309 bp; , 223 bp and 271 bp (the latter minor product is more clearly visible in lane 4); and gephyrin, 279 bp. Cultured hippocampal cells expressed 2, , and gephyrin mRNA; no 1 or 3 mRNA was detected. Similar results were obtained in 2 other trials using different platings.

Some benzodiazepines inhibit glycine currents

We screened several benzodiazepines for their effect on 50 µM glycine (EC₅₀, see Fig. 2A2 and Table 1) currents. At 20–200 µM, the 1,4-benzodiazepines, chlordiazepoxide, nitrazepam, and lorazepam; the 1,4-benzodiazepines with a fused triazolo ring, alprazolam and triazolam; the competitive benzodiazepine antagonist, flumazenil; and the partial inverse benzodiazepine agonist, RO 15–4513, inhibited 50 µM glycine currents by ≥20%. For example, chlordiazepoxide inhibited 50 µM glycine currents in a dose-dependent manner with an IC₅₀ of 230 ± 38 µM and a Hill coefficient of 0.8 ± 0.1 (Fig. 5). Young et al. (1974) found that chlordiazepoxide blocked strychnine binding to rat brainstem and spinal cord synaptic membranes with an identical potency. In contrast, 100 µM diazepam (a 1,4-benzodiazepine), 100 µM clobazam (a 1,5-benzodiazepine), and 30 µM GYKI 52466 (a 2,3-benzodiazepine) inhibited 50 µM glycine currents by <20%.

View larger version (17K): [in this window] [in a new window]   FIG. 5. Chlordiazepoxide inhibits GlyR-mediated currents. A: currents produced by 50 µM glycine alone (control), 50 µM glycine + 200 µM chlordiazepoxide (CDPX), and 50 µM glycine alone after washout of chlordiazepoxide (recovery) in a neuron voltage-clamped at –65 mV. Chlordiazepoxide was preapplied for 15 s. B: dose-dependent inhibition of 50 µM glycine currents by chlordiazepoxide at –65 mV. , the mean peak current in the presence of various chlordiazepoxide concentrations expressed as a percentage of the peak current evoked by 50 µM glycine alone (n = 3–8). When larger than the symbol, error bars represent SE. —, the fit to a logistic equation with an IC50 of 230 ± 38 µM using an n of 0.8 ± 0.1.

We examined nitrazepam in detail because it was one of the more potent benzodiazepines tested, and it did not precipitate out of an aqueous solution at higher concentrations. Nitrazepam inhibited 50 µM glycine responses with an IC₅₀ of 83 ± 14 µM and a Hill coefficient of 1.3 ± 0.2 (Fig. 6, A and B). Nitrazepam blocks strychnine binding to rat brain stem and spinal cord synaptic membranes with a similar potency (Hunt and Raynaud 1977; Young et al. 1974). Because some dihydropyridines such as nitrendipine and nicardipine block responses to high glycine concentrations and potentiate responses to low glycine concentrations (Chesnoy-Marchais and Cathala 2001), we examined the effect of nitrazepam on 20 µM (EC₂₀) and 200 µM (EC_100, see Fig. 2A2) glycine currents. Unlike nitrendipine and nicardipine, nitrazepam inhibited both 20 and 200 µM glycine currents (Fig. 6B). Chlordiazepoxide also inhibited 20 µM glycine currents (data not shown). The similarity of the dose-response curves for nitrazepam inhibition of 20, 50, and 200 µM glycine currents suggests that the inhibition is noncompetitive. Further support for noncompetitive inhibition comes from the rightward shift and decreased maximum in the glycine dose-response curve when nitrazepam is present (Fig. 6C). In the presence of 80 µM nitrazepam, glycine had an EC₅₀ of 120 ± 23 µM and a Hill coefficient of 1.4 ± 0.1 with a 22% decrease in the maximum current. Nitrazepam block was not voltage-dependent because the degree of block at –65 and +35 mV was the same (Fig. 6D). Finally, nitrazepam did alter the reversal potential of glycine currents. Using symmetric [Cl^–] solutions, 50 µM glycine currents reversed at 1.8 ± 0.8 mV (n = 4) and 3.0 ± 0.9 mV (n = 4) in the absence and presence of 100 µM nitrazepam as determined by subjecting steady-state currents to voltage ramps.

View larger version (37K): [in this window] [in a new window]   FIG. 6. Nitrazepam inhibits GlyR-mediated currents. A: currents evoked by 50 µM glycine alone (control), 50 µM glycine + 100 µM nitrazepam (NZP) after preapplying 100 µM nitrazepam for 30 s, and 50 µM glycine alone after washout of nitrazepam (recovery) in a neuron voltage-clamped at –65 mV. B: dose-response curves for the inhibition of 20, 50, and 200 µM glycine currents by nitrazepam at –65 mV. Points show the mean peak glycine current in the presence of various nitrazepam concentrations expressed as a percentage of the current evoked by 20 µM (), 50 µM (), or 200 µM (*) glycine alone in the same neuron (n = 3–13). When larger than the symbol, error bars show SE. The · · · (20 µM glycine), — (50 µM glycine), and - - - (200 µM glycine) curves are fits to a logistic equation. The fits yielded an IC50 of 90 ± 15, 83 ± 14, and 82 ± 4 µM using an n of 1.4 ± 0.2, 1.3 ± 0.2, and 1.3 ± 0.1 for 20, 50, and 200 µM glycine, respectively. C: nitrazepam inhibited glycine currents noncompetitively. , the mean peak current elicited by varying concentrations of glycine + 80 µM nitrazepam expressed as a percentage of the peak current elicited by 100 µM glycine alone in the same neuron at –65 mV. —, the fit to a logistic equation with an EC50 of 120 ± 23 µM using an n of 1.4 ± 0.1. The glycine dose-response curve from Fig. 2A2 with an EC50 of 57 ± 8.0 µM in the absence of nitrazepam ( and · · ·) is shown for comparison. The maximum responses in the presence and absence of nitrazepam were compared using the three highest glycine concentrations tested. All three pairs were significantly different (P = 0.0001–0.03). D: nitrazepam inhibition of glycine currents was not voltage-dependent. and , the mean peak 50 µM glycine current + 20 µM or 300 µM nitrazepam, respectively, expressed as a percentage of the peak 50 µM glycine current alone at –65 and +35 mV (n = 3–5). are not significantly different (P = 0.16); are not significantly different (P = 0.17). Error bars show the SE.

Nitrazepam may inhibit GlyRs by enhancing desensitization. Indeed, chlordiazepoxide enhances GABA_AR desensitization (Mierlak and Farb 1988). In the presence of 100 µM nitrazepam, 300 µM glycine currents had peak amplitudes that were 56 ± 4% (n = 8) of control and desensitized along a two-exponential time course as control currents did. While nitrazepam increased the fast time constant by 50%, it did not alter the slow time constant, the relative contributions of each time constant, or the weighted mean time constant (Fig. 7, A–C). The time course of recovery from desensitization for 300 µM glycine currents followed a single exponential with a time constant of 7.6 ± 1.1s in the presence of 100 µM nitrazepam (Fig. 7D). Nitrazepam may accelerate recovery from desensitization slightly given a weighted mean time constant of 11 s in the absence of nitrazepam. Similarly, 50 µM nitrazepam did not alter the desensitization parameters for 10 µM glycine currents (data not shown). Thus nitrazepam did not enhance GlyR desensitization.

View larger version (30K): [in this window] [in a new window]   FIG. 7. Nitrazepam does not alter GlyR desensitization. A–C: nitrazepam prolonged the fast (fast) but did not alter the slow (slow) or weighted mean (mean) time constant of GlyR desensitization (A). Nitrazepam did not alter the relative contributions of the fast (fast) and slow (slow) components of GlyR desensitization (B) or the percent desensitization (C). Fitting the desensitizing phases of the currents evoked by 300 µM glycine () or 300 µM glycine + 100 µM nitrazepam () to a 2-exponential function as in Fig. 1A1 yielded values for fast and slow. Bars show SE (n = 6–7). *P = 0.04 compared with 300 µM glycine alone. P = 0.28–0.85 when pairwise comparisons in the presence and absence of nitrazepam are made for the other parameters in A–C. D: GlyR recovery from desensitization follows a single exponential in the presence of 100 µM nitrazepam. Recovery was determined as in Fig. 1A2 except that 100 µM nitrazepam was present throughout the experiment. , the raw data (n = 3–9) with error bars representing SE. —, a single exponential function having a time constant of 7.6 ± 1.1 s. , the time course of recovery from desensitization for 300 µM glycine currents in the absence of nitrazepam with a weighted mean time constant of 11 s (from Fig. 1A3). Data in this figure came from neurons voltage-clamped at –65 mV.

We preapplied the benzodiazepines for 30 s before coapplying with glycine because greater inhibition occurred with preapplication. In addition, the inhibition required the presence of the benzodiazepine during the glycine application. For example, the peak amplitude of 50 µM glycine currents was 14 ± 6% (n = 6) of control when 100 µM nitrazepam was both preand co-applied with glycine. Peak 50 µM glycine currents were 33 ± 10% (n = 6) of control when 100 µM nitrazepam was co-applied but not preapplied. When 100 µM nitrazepam was preapplied but not co-applied with glycine, 50 µM glycine currents were 82 ± 2 (n = 4) of control. Similar results were obtained with alprazolam.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION ACKNOWLEDGMENTS REFERENCES

 
The principal finding of this study is that some benzodiazepines inhibit ₂-containing GlyRs in cultured embryonic mouse hippocampal neurons. This study provides electrophysiological evidence for an inhibitory benzodiazepine site on GlyRs. Although Young et al. (1974) predicted that benzodiazepines interact with GlyRs based on binding studies, subsequent electrophysiological studies in spinal neurons found no interaction (Choi et al. 1981; Macdonald and Barker 1978). Perhaps these negative results reflect the inability of benzodiazepines to modulate spinal GlyRs, which have a different subunit composition. Alternatively, observing this interaction may require using voltage-clamp recordings and fast perfusion systems. When Rigo et al. (2002) screened several antiepileptic drugs for their effects on GlyRs using these methods, they found that clonazepam slightly inhibits GlyR currents in spinal cord neurons.

Cultured embryonic mouse hippocampal neurons express ₂-containing GlyRs

GlyR subunit composition changes during development in the spinal cord, which expresses ₂ homomers prenatally and ₁ heteromers in the adult (Rajendra et al. 1997). Location is a second factor influencing GlyR subunit composition based on in situ hybridization results (Malosio et al. 1991). For example, the embryonic hippocampus expresses both ₂ and GlyR subunits, which our RT-PCR results confirm. However, in situ hybridization cannot determine whether GlyRs are ₂ heteromers or ₂ and homomers. The pharmacological properties of these subunit combinations can help determine which ones exist (see RESULTS). Because the pharmacological properties of GlyRs in embryonic hippocampal neurons have not been reported, this was one goal of our study. The pharmacological properties indicate that the GlyRs are primarily ₂ heteromers but ₂ homomers also exist. We cannot exclude the presence of low levels of other subunit combinations such as ₃, which may account for the rare rapidly desensitizing current. Alternatively, the rapidly desensitizing current may depend on second-messenger modulation (Gentet and Clements 2002). ₂ heteromers may persist into adulthood because the juvenile hippocampus (Chattipakorn and McMahon 2002) and adult amygdala (McCool and Farroni 2001) express the same subunit combination. Furthermore, in situ hybridization fails to detect ₁ expression in the adult hippocampus (Malosio et al. 1991). Thus GlyR subunit composition depends on both location and age, and the effect of GlyR modulators will vary with location and age.

Although -alanine's role in the CNS remains unknown, taurine plays an important role during brain development (Palackal et al. 1986). Because -alanine and taurine activate both GlyRs and GABA_ARs (Choquet and Korn 1988; del Olmo et al. 2000), the receptor mediating their actions in different regions of the CNS during development is unclear. Glycine, taurine, and -alanine currents have similar pharmacological and physiological profiles in embryonic hippocampal neurons. Indeed, based on their interaction with glycine at saturating concentrations, they appear to activate GlyRs exclusively at the concentrations tested. Thus GlyRs, which may be the primary receptor for taurine in the developing neocortex (Flint et al. 1998) and postnatal hippocampus (Mori et al. 2002), may also be the primary receptor for taurine and -alanine in the embryonic hippocampus. The difference in the sensitivity of glycine currents to bicuculline compared with -alanine and taurine currents may reflect a difference in the binding site and/or gating mechanism for these agonists (Han et al. 2001; Schmieden et al. 1992).

Taurine activates GlyRs as either a full or a partial agonist depending on the preparation. Taurine may be a partial agonist in some cases because it acts as an antagonist via a specific domain in the ₁ subunit (Schmieden et al. 1999). However, other factors may be involved because taurine ranges from a full agonist to a partial agonist as the glycine EC₅₀ increases for ₁ and ₂ homomers expressed in oocytes (de Saint et al. 2001). We found that taurine is a full agonist on ₂-containing GlyRs in embryonic hippocampal neurons. In contrast, it is a partial agonist in adult amygdala neurons (McCool and Botting 2000), which express the same subunit combination (McCool and Farroni 2001). Together, these results indicate that factors in addition to subunit composition determine whether taurine acts as a full or partial agonist.

Benzodiazepine inhibition of GlyRs

Benzodiazepine concentrations similar to those that block ₂-containing GlyRs modulate GABA_ARs and inhibit some voltage-gated ion channels. GABA_ARs have both a "high-" and "low-"affinity benzodiazepine site (Walters et al. 2000). Diazepam at 20–100 µM potentiates ₁₂₂ GABA_ARs via a low-affinity site. Interestingly, these same concentrations acting at the high-affinity site inhibit ₁₂₂ GABA_ARs (Rovira and Ben Ari 1993; Walters et al. 2000). Walters et al. (2000) postulate that the low-affinity site mediates the anesthetic effects of benzodiazepines. We hypothesize that GlyRs have a low-affinity benzodiazepine site homologous to the low-affinity benzodiazepine site on GABA_ARs because both receptors belong to the same family. Benzodiazepines at 10–200 µM also inhibit neuronal voltage-gated sodium, low- and high-threshold calcium, and delayed rectifier potassium channels (Backus et al. 1991; Yang et al. 1987). The presence of a low-affinity benzodiazepine site on both ligand- and voltage-gated channels suggests that the site influences channel gating rather than ligand binding or the voltage sensor. This lowaffinity benzodiazepine site may be analogous to the conserved inhibitory pentobarbital site on various ligand- and voltage-gated channels hypothesized by Akk and Steinbach (2000). Furthermore, our findings indicate that GlyRs lack a high-affinity, potentiating benzodiazepine site.

Nitrazepam does not inhibit ₂-containing GlyRs by a simple open channel block mechanism or by promoting desensitization. The lack of greater inhibition with higher agonist concentrations excludes a simple open channel block mechanism. Nitrazepam does not alter the slow time constant of desensitization, the weighted mean time constant of desensitization, or the degree of desensitization, but it significantly prolongs the faster time constant of desensitization. It also may slightly accelerate recovery from desensitization overall despite eliminating the faster component of recovery. The net effect of slowing the entry of GlyRs into the desensitized state and increasing the rate of recovery from desensitization would not result in smaller currents. Thus these effects on desensitization do not account for its inhibition of GlyR currents, and we conclude that nitrazepam does not block GlyR currents by enhancing desensitization.

Nitrazepam inhibition is noncompetitive because it shifts the glycine dose-response curve to the right and diminishes the maximum response. A noncompetitive inhibitor should allow glycine to occupy all binding sites at saturating concentrations. Therefore nitrazepam inhibits GlyRs by altering channel gating and/or other transitions between ligand bound channel states through an allosteric mechanism (Colquhoun 1998). However, we cannot exclude the possibility that nitrazepam also alters glycine binding allosterically.

Clinical relevance

The low-affinity benzodiazepine site on GlyRs and GABA_ARs may have several important clinical roles. First, Rigo et al. (2002) recently demonstrated that levetiracetam, a novel antiepileptic drug, may modulate GlyRs and GABA_ARs to a lesser extent via a low-affinity benzodiazepine site. Second, the low-affinity site on the GlyR may mediate some of the effects of anesthetic doses of benzodiazepines, the hypothesized role for the site on GABA_ARs (Walters et al. 2000). As expected from our results, mice with diminished GlyR function are more sensitive to the anesthetic effects of midazolam (Quinlan et al. 2002). Third, GlyR inhibition via this site may be one mechanism by which benzodiazepines stop neonatal seizures because GlyR activation produces a depolarizing rather than a hyperpolarizing response in neonatal neurons (Ito and Cherubini 1991). Thus the relative inability of benzodiazepines to augment neonatal GABA_ARs may be advantageous clinically. While the benzodiazepine concentrations needed to inhibit GlyRs are higher that those needed to augment GABA_ARs, the cerebrospinal fluid concentrations achieved clinically in neonates are unknown. Finally, this site may mediate the effects of the high dose benzodiazepine regimens used to treat electrical status epilepticus during slow wave sleep (De Negri et al. 1995) and nonketotic hyperglycinemia, a condition associated with elevated cerebrospinal fluid glycine concentrations and severe neonatal seizures (Matalon et al. 1983).

In summary, benzodiazepines inhibit ₂-containing GlyRs in embryonic mouse hippocampal neurons via a low-affinity benzodiazepine site that may have several important roles clinically.

	ACKNOWLEDGMENTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION ACKNOWLEDGMENTS REFERENCES

 
We thank S. M. Rothman for helpful comments and discussion. We thank N. Barron and N. Rensing for preparing and maintaining the neuronal cultures. We thank T. Fu for technical assistance.

This work was supported by National Institutes of Neurological Disorders and Stroke Grants 5K12NS-0169004 and 1K02NS-43278-01.

	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: Correspondence: Liu Lin Thio, M.D., Ph.D., Pediatric Epilepsy Center, Suite 12E47, St. Louis Children's Hospital, One Children's Place, St. Louis, MO 63110 (E-mail: thio{at}kids.wustl.edu).

	REFERENCES

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION ACKNOWLEDGMENTS REFERENCES

 
Akaike N and Kaneda M. Glycine-gated chloride current in acutely isolated rat hypothalamic neurons. J Neurophysiol 62: 1400–1409, 1989.[Abstract/Free Full Text]

Akk G and Steinbach JH. Activation and block of recombinant GABA_A receptors by pentobarbitone: a single-channel study. Br J Pharmacol 130: 249–258, 2000.[ISI]

Backus KH, Pflimlin P, and Trube G. Action of diazepam on the voltage-dependent Na⁺ current. Comparison with the effects of phenytoin, carbamazepine, lidocaine and flumazenil. Brain Res 548: 41–49, 1991.[ISI][Medline]

Becker CM, Betz H, and Schröder H. Expression of inhibitory glycine receptors in postnatal rat cerebral cortex. Brain Res 606: 220–226, 1993.[ISI][Medline]

Bloomenthal AB, Goldwater E, Pritchett DB, and Harrison NL. Biphasic modulation of the strychnine-sensitive glycine receptor by Zn²⁺. Mol Pharmacol 46: 1156–1159, 1994.[Abstract]

Brooks-Kayal AR, Shumate MD, Jin H, Rikhter TY, Kelly ME, and Coulter DA. -Aminobutyric acid_A receptor subunit expression predicts functional changes in hippocampal dentate granule cells during postnatal development. J Neurochem 77: 1266–1278, 2001.[ISI][Medline]

Chattipakorn SC and McMahon LL. Pharmacological characterization of glycine-gated chloride currents recorded in rat hippocampal slices. J Neurophysiol 87: 1515–1525, 2002.[Abstract/Free Full Text]

Chesno