European Journal of Pharmacology, 210 (1992) 239-246

239

© 1992 Elsevier Science Publishers B.V. All rights reserved 0014-2999/92/$05.00

EJP 52231

Modulation of GABA A and glycine receptors by chlormethiazole T i m G. H a l e s

a

a n d J e r e m y J. L a m b e r t b

a Department of Anesthesiology, Medical Center, UCLA, LosAngeles, CA 90024, U.S.A. and b Neuroscience Research Group, Department of Pharmacology and Clinical Pharmacology, Ninewells Hospital and Medical School, University of Dundee, Dundee DD1 9SY, U.K.

Received 6 June 1991,revised MS received 17 September 1991, accepted 22 October 1991

The influence of chlormethiazole, on currents evoked by y-aminobutyric acid (GABA) and glycine, was investigated under voltage-clamp conditions, in bovine chromaffin cells and murine spinal neurones, respectively. Chlormethiazole (30 and 100/xM) dose dependently potentiated currents activated by either inhibitory neurotransmitter. The potentiation of the GABA-evoked response occurred without altering the reversal potential and was not influenced by the benzodiazepine receptor antagonist Ro 15-1788 (300 nM). GABA-gated channels, recorded from outside-out membrane patches, showed increased probability of being in the conducting state in the presence of chlormethiazole. High concentrations of chlormethiazole (3 mM) activated bicuculline (1 /zM)-sensitive whole-cell currents with a reversal potential similar to the chloride equilibrium potential. Chlormethiazole potentiates GABA- and glycine-activated currents and at higher doses, directly activates the G A B A A receptor. Chlormethiazole; GABA A receptors; Glycine receptors; Patch-clamp; Anaesthetics (general); Anticonvulsants

I. Introduction

Chlormethiazole has been used clinically as an induction agent in situations when deep sedation, short of surgical anaesthesia, was required (Wilson et al., 1969). However, chlormethiazole is more frequently used for its anxiolytic and anticonvulsant properties which have been demonstrated in both experimental animal models and human studies (0gren, 1986). Chlormethiazole is particularly effective in the treatment of status epilepticus (Harvey et al., 1975) and delirium tremens (Athen et al., 1977), in patients failing to respond to barbiturates and benzodiazepines. Recently, chlormethiazole has been demonstrated to be neuroprotective in experimental animals following ischaemic episodes (Cross et al., 1991). Electrophysiological and radioligand binding studies have provided evidence for the G A B A A receptor as a site of action for a number of intravenous (i.v.) anaesthetics, anticonvulsants and anxiolytics (e.g. Simmonds and Turner, 1987; Olsen, 1988; Peters et al., 1988). The anaesthetic steroids (Harrison and Simmonds, 1984; Barker et al., 1987; Lambert et al., 1987), depressant barbiturates (Barker and Ransom, 1978; Simmonds, 1981; Peters et al., 1988), propofol (Hales and Lain-

Correspondence to: T.G. Hales, Department of Anesthesiology, Medical Center, UCLA, Los Angeles, CA 90024, U.S.A. Tel. 1.310.206 6227, fax 1.310.206 5779.

bert, 1991), propanidid (Peters et al., 1989), etomidate (Robertson, 1989) and the benzodiazepines (Macdonald and Barker, 1978; Vicini et al., 1988) have all been demonstrated, using electrophysiological techniques, to potentiate responses mediated by the G A B A A receptor. Extracellular recordings from the rat medullary reticular formation (Gent and Wacey, 1983) and from slices of rat cuneate nucleus (Harrison and Simmonds, 1983), together with radioligand binding data (Cross et al., 1989; Moody and Skolnick, 1989) have demonstrated that chlormethiazole also modulates the GABA A receptor. However, chlormethiazole has additionally been reported to modulate the strychnine-sensitive glycine receptor (Gent and Wacey, 1983; Harrison and Simmonds, 1983). The mechanism of action of chlormethiazole, on the G A B A A receptor of bovine chromaffin cells, was investigated u n d e r voltage-clamp conditions using whole-cell and outside-out patch configurations of the patch-clamp technique. This preparation provides G A B A A receptors pharmacologically similar to those present on central neurones (Peters et al., 1989). The whole-cell configuration was also used to investigate the effect of chlormethiazole on glycine-evoked currents recorded from murine spinal neurones. In these studies the actions of chlormethiazole on inhibitory neurotransmitter responses were compared to those of pentobarbitone. Preliminary accounts of a part of this work have appeared in abstract form (Hales and Lambert, 1988; Hales et al., 1990).

240 2. Materials and methods

2.1. Preparation of bovine chromaffin cell cultures Bovine chromaffin cells were isolated and cultured by the method of Fenwick et al. (1982) with minor modifications (Peters et al., 1988). Cells were used in electrophysiological experiments at 17-21 ° C, 1-7 days after plating.

2.2. Preparation of murine spinal neurone cultures The method of dissociation and maintenance in culture, of embryonic spinal neurones was similar to that of Barker and Ransom (1978) with minor modifications. Briefly, spinal cords of embryonic mice (13-14 days) were removed and dissociated in Hank's balanced salt solution containing 0.25% trypsin and NaHCO 3 (1.5 mg ml -~) at 37°C for 30 min (90% air-10% CO2). During dissociation tissue was transferred to incubation medium consisting of Dulbecco's Modified Eagles Medium (with pyruvate), supplemented with 10% v/v heat-inactivated foetal calf serum, 10% v / v heat-inactivated horse serum, penicillin (50 IU m l - 1 ) / streptomycin (0.05 mg ml-1), NaHCO 3 (1.5 mg ml-l) and glutamine (2 mM). Sterile distilled water was added to reduce the osmolarity of the incubation medium to 330 mOsm. Flame polished pasteur pipettes were used to triturate the tissue. Complete dissociation was achieved by additional trituration using pipettes flame polished further to reduce their aperture. During this process dispersed cells were removed from the top of the suspension and transferred to a centrifuge tube containing fresh incubation medium. Cells were plated at a density of 1 x 106 cells per 35 mm diameter tissue culture dish. Cultures were incubated in incubation medium (1.5 ml) and an humid atmosphere of 10% CO2-90% air. Neurones and supporting cells were allowed to proliferate for 5 days after which the incubation medium was replaced by one supplemented with glucose (5 nag ml-~). Initially, the mitotic inhibitors 2-deoxy-5-fluoridine (15/zg ml- 1) and uridine (35 /zg ml -~) were also present. After 2 days the incubation medium was replaced by one lacking the mitotic inhibitors. The medium was renewed every third day and electrophysiological recordings were made from cultured neurones between weeks 6 and 12, at 17-21° C.

2.3. Electrophysiological recordings An EPC-7 patch-clamp amplifier (List) was used to record, under voltage-clamp conditions, currents from whole-cells (low-pass filtered at 500 Hz) and outside-out membrane patches (low-pass filtered at 1 kHz). Unless otherwise stated the holding potentials were - 6 0 and

- 1 0 0 mV, respectively. Data were recorded with a Racal (Store 4DS) FM tape recorder on to magnetic tape (Ampex) for subsequent analysis (described below). Cells and membrane patches were continuously superfused (3-5 ml min -1) with saline comprising (in mM): NaCI 140; MgC12 2.0; CaCI 2 1.0 and HEPESNaOH 10 (pH 7.2). The internal solution for chromaffin whole-cells and excised outside-out membrane patches consisted of (in mM): CsC1 140; MgC12 2.0; CaC12 0.1; EGTA 1.1 and HEPES-NaOH 10 (pH 7.2). Cs ÷ ions were employed to minimise outward currents through potassium channels. The internal solution for spinal neurones comprised (in mM): KCI 140; MgC12 2.0; CaC12 0.1; EGTA 1.1 and HEPES-NaOH 10 (pH 7.2). Where stated, corrections for liquid junction potentials were made by the method of Fenwick et al. (1982). Drugs were applied either locally, by pressure ejection (General Valve Picospritzer II) from modified patch pipettes (at a pressure of 1.4 x 105 Pa, a fiequency of 0.05 Hz and a duration of 10-50 ms), or by addition to the superfusate.

2.4. Data analysis Whole-cell currents were either replayed onto a pen recorder (Lectromed Multitrace 2) and measured by hand, or analysed using a PDP 11-73 minicomputer (Dempster, 1988), which enabled currents to be averaged a n d / o r superimposed. Whole-cell currents were digitised into 512 points at 100 Hz using an analog to digital converter (Cambridge Electronic Design 502). Currents were plotted for illustration on a Hewlett Packard plotter (7470A). y-Aminobutyric acid (GABA)-gated single channel currents were digitised at 10 kHz and an automated procedure was used to analyse records. The signal was inspected on a display screen (Electronic Visual 8060 oscilloscope) and a threshold level was set during a quiescent section of the recording, just below the baseline 'noise'. The program analysed the digital recording, using the threshold to determine the location of each channel opening and closure. The probability of at least one channel being in the conducting state (Popen) was determined for 30-120 s segments of data. The number of channels on a particular patch was not estimated, Therefore direct comparisons of Pope, between patches were not made. Percentage changes in Pope. within individual patches were calculated in order to determine the effects of chlormethiazole on GABA-gated channels.

2.5. Drugs used Drugs used in this study were: GABA, Na-pentobarbitone, bicuculline, strychnine HCI, glycine (Sigma),

241

Ro 15-1788 (ethyl-8-fluoro-5,6-dihydro-5-methyl-6-oxo4H-imidazo[1,5-a][1,4]benzodiazepine-3-carboxylate) (Roche) and ehlomethiazole edisilate (chlormethiazole) (Astra).

3. Results

3.1. Potentiation of GABA and glycine-evoked whole-cell currents by chlormethiazole Submaximal whole-cell currents evoked by locally applied GABA (100/zM), were potentiated by pentobarbitone (30 and 100/~M) applied to chromaffin cells by superfusion (fig. 1A). Bath-applied chlormethiazole (30 and 100 /~M) also potentiated GABA-activated currents recorded from chromaffin cells (fig. 1B). The potentiation of responses to GABA by both compounds was dose-dependent, pentobarbitone being more potent than chlormethiazole at both doses tested (fig. 1C).

GABA

j

Currents, evoked by the local application of glycine (100 /xM) to spinal neurones voltage-clamped at - 6 0 mV, were abolished by bath applied strychnine (100 nM; n = 5). This dose of strychnine had no effect on GABA-activated currents recorded from spinal neurones (n = 3, data not shown). Chlormethiazole (10, 30 and 100 /zM) when bath applied, caused a dose-dependent potentiation of glycine-evoked currents (fig. 1D and E). In contrast, pentobarbitone (10 and 100 /xM) had no effect on glycine-activated currents (fig. 1E). Doses of chlormethiazole and pentobarbitone above 100 /xM were not tested against glycine responses because application of high concentrations of either drug induced an inward current in the absence of pressure applied agonist. This was probably due to the direct activation of GABA n receptors and will be discussed below. Chlormethiazole (10 and 100 /xM) produced a greater enhancement of GABA-activated currents than of glycine-evoked currents (fig. 1). The mechanism of action of chlormethiazole on the GABA A receptor was further investigated.

GABA

/

/..-~o..,~.~,...ltL--~o..Chio,...,.,..o. --V/ -O..~d OOpA

~/

--IOOuM Chio--thiazolo

ls (50pA

% of Control Response

's '150pA

% of Control Response

C

E

200 -

800 150 600 100 400

50'

200 -

i

0 - 7//'"

30

50 Conceflt ration (pM)

100

;,/J

, 10

i 30

100

Concentration (uM)

Fig. 1. Chlormethiazole modulates GABA and glycine-induced currents. (A, B) Superimposed traces illustrating the concentration-dependent potentiation of GABA (100 p.M)-evoked currents by bath-applied pentobarbitone and chlormethiazole, respectively. Recordings were made at - 60 mV from bovine chromaffin cells. (C) Plot of the effect of chlormethiazole (e) and pentobarbitone ( • ) on the amplitude of currents evoked by GABA. (D) Potentiation of glycine (100 ~M)-activated currents by chlormethiazole recorded from a mouse spinal neurone clamped at - 6 0 mV. (E) Plot of the effect of chlormethiazole (o) and pentobarbitone (E3) on the amplitude of currents evoked by glycine. In the graphs, current amplitudes (ordinate) are expressed as a percentage of the amplitudes in the absence of the drugs. The abscissa indicates the concentration of chlormethiazole or pentobarbitone on a logarithmic scale. Each data point is the mean of at least four observations and vertical lines indicate the S.E.M. Current traces depict computer generated averages of four responses to GABA or glycine (applied by pressure ejection).

242

Chlormethiazole lOOlJm Chlormethiazole Ro15-1788 + lOOpM 30OHM Wash

-160

ilIllIll

IllIrrl-i

vh (mY)

--80 IlpA)

Diazepam

Ilf ? lrr rllrrl IrrIll lll[I 1~

, -160

Fig. 2. Chlormethiazole potentiates GABA-evoked currents without affecting their reversal potential. The average current amplitudes (I) of three responses to locally applied GABA (100 /,M) were determined at holding potentials ranging between +60 mV (20 mV increments) and were plotted versus the holding potential (Vh). The GABA-activated current-voltage relationship in the absence (e) and presence (111) of chlormethiazole (30 /*M) demonstrates that the drug potentiates the GABA response at all potentials. Curves were fitted by eye and the reversal potential was estimated by extrapolation to be 4 mV. The liquid junction potential was subtracted from the data prior to plotting. All data were obtained from the same bovine chromaffin cell.

3.2. Mechanism of potentiation by chlormethiazole of GABA-activated currents The current-voltage relationship of the GABAevoked response was determined under control conditions and in the presence of chlormethiazole ( 3 0 / , M )

Diazepam lpM + Ro15-1788 31~nM

~1200

pA

lmin

Fig. 3. The benzodiazepine receptor antagonist Ro 15-1788 does not affect chlormethiazole-evoked modulation of the G A B A A receptor. Potentiation of GABA (100/,M)-activated currents by chlormethiazole (100 tzM) was resistant to bath application of Ro 15-1788 (300 nM). In contrast, potentiation of GABA-evoked currents by diazepam (1/,M) was abolished by Ro 15-1788. Currents were recorded from the same chromaffin cell voltage-clamped at - 60 inV.

(fig. 2). Chlormethiazole had no effect on the reversal potential (4 mV) of GABA-activated currents and the potentiation was independent of voltage. In an attempt to determine whether chlormethiazole binds to the benzodiazepine recognition site on the GABA A receptor, experiments were performed using

~"lr i 'lr""llll r,iT-lrlr,,tfl,,,, ,I'1Pll'I i~r~l ~-ipl i~ ~ 'I'F r'741'-'"rllI~-~l i i'rl"ri '-~"li'~7 r 17:-i,,~ ...mr i,ri.~r~i.~.F1j r.rl ri ,.r~ r.i i llil i ~iSell,.,iii~l~,~i.r~lijrli,., Ulift~LVII,lli~lr., lli~l. ~r-.Til~[ll~r, .c,y~ il ,~-