Mode of action of ICS 205,930, a novel type of potentiator of responses to glycine in rat spinal neurones

Abstract

1. The effect of a novel potentiator of glycine responses, ICS 205,930, was studied by whole-cell recordings from spinal neurones, and compared with that of other known potentiators, in an attempt to differentiate their sites of action.

2. The ability of ICS 205,930 (0.2 μM) to potentiate glycine responses persisted in the presence of concentrations of Zn²⁺ (5–10 μM) that were saturating for the potentiating effect of this ion.

3. Preincubation with 10 μM Zn²⁺ before application of glycine plus Zn²⁺ had an inhibitory effect, which did not result from Zn²⁺ entry into the neurone, since it persisted with either 10 mM internal EGTA or 10 μM internal Zn²⁺. To test whether the potentiating effects of ICS 205,930 and Zn²⁺ interact, both compounds were applied without preincubation.

4. The potentiating effect of ICS 205,930 was similar for responses to glycine and for responses to glycine plus Zn²⁺, provided the concentrations of agonist were adjusted so as to induce control responses of identical amplitudes.

5. ICS 205,930 remained able to potentiate glycine responses in the presence of ethanol (200 mM).

6. ICS 205,930 also retained its potentiating effect in the presence of the anaesthetic propofol (30–90 μM), which strongly potentiated glycine responses but, in contrast with ICS 205,930, also markedly increased the resting conductance.

7. The anticonvulsant chlormethiazole (50–100 μM) neither potentiated glycine responses nor prevented the effect of ICS 205,930, even though it increased the resting conductance and potentiated GABA_A responses.

8. The mechanism of action of ICS 205,930 appears to be different from those by which Zn²⁺, propofol or ethanol potentiate glycine responses.

Introduction

In a previous study performed in ventral spinal cord neurones in primary culture, the chloride conductance increase induced by glycine was shown to be potentiated by three molecules previously known as 5-HT₃ antagonists (ICS 205,930, MDL 72222, LY 278,584). The ability of these compounds to potentiate glycine responses, however, appeared to be independent of their 5-HT₃ antagonist properties, and to result from their capacity to increase the apparent affinity of the receptors for glycine (Chesnoy-Marchais, 1996). Responses to glycine have also been reported to be potentiated by other compounds, such as Zn²⁺ ions (Akagi et al., 1993; Bloomenthal et al., 1994; Laube et al., 1995; Zhang & Berg, 1995; Kumamoto & Murata, 1996; Trombley & Shepherd, 1996), various types of anaesthetics, such as propofol (Hales & Lambert, 1991; Mascia et al., 1996a; 1997; Pistis et al., 1997) or inhalation anaesthetics (Wakamori et al., 1991; Harrison et al., 1993; Downie et al., 1996; Mihic et al., 1997), and alcohols (Celentano et al., 1988; Aguayo & Pancetti, 1994; Mascia et al., 1996a,1996b; Mihic et al., 1997; Ye et al., 1998). The identification of the site(s) of action of potentiators of glycine receptors is currently under study in several laboratories. In the case of Zn²⁺, the role of some extracellular domains has been demonstrated (Laube et al., 1995), and in the case of volatile anaesthetics and high concentrations of ethanol, a residue present in the second transmembrane segment of α subunits has been shown to play a critical role (Mihic et al., 1997). It is well-known that different potentiators of a given receptor can involve different sites of action (for reviews concerning the potentiations of GABA_A responses by benzodiazepines, barbiturates, steroids and other agents, see MacDonald & Olsen, 1994 or Whiting et al., 1995). The mode of action of propofol on GABA_A receptors, for example, has been shown to be different from that of many other potentiators, including ethanol (Mihic et al., 1997) and some other general anaesthetics, such as volatile anaesthetics (Mihic et al., 1997; Krasowski et al., 1998) or etomidate (Hill-Venning et al., 1997). In the present paper, whole-cell recordings from ventral spinal cord neurones and combined applications of ICS 205,930 with either Zn²⁺, ethanol or propofol were used in order to determine whether the mechanism of potentiation of glycine responses by low concentrations of ICS 205,930 can be discriminated from the mechanisms involved in the effects of these other modulators. The effects of chlormethiazole, another molecule reported to potentiate glycine responses (Gent & Wacey, 1983; Hales & Lambert, 1992), were also reinvestigated.

Methods

The experiments were performed at room temperature (20–23°C) on primary cultures of ventral spinal cord neurones from rat embryos, using the whole-cell configuration of the patch-clamp technique.

Cell preparation

Female rats (OFA, Iffa Credo) carrying E15 or E16 embryos were killed by CO₂-induced anoxia. No anaesthetic was used. The embryos were rapidly removed and decapitated. Their spinal cords were taken out and dissected under a microscope in a phosphate buffer saline (PBS without Ca and Mg, Gibco) supplemented with glucose (33 mM). After removal of the meninges, the ventral part of the spinal cord was excised and cut in small pieces in order to dissociate the neurones as previously described (Chesnoy-Marchais, 1996). Recordings from cells were performed between day 11 and day 15 in culture.

Experimental solutions

Before recording, the culture medium was replaced by the external solution to be used during the recording which contained (in mM): NaCl 150, KCl 2.5, CaCl₂1.8, MgCl₂ 1, glucose 20 and HEPES-NaOH 10 pH 7.4. The internal solution used to fill the recording electrode in most experiments, internal solution 1, contained (in mM): Cs methanesulphonate 145, CsCl 15, MgCl₂ 1, EGTA 0.1, ATP-Mg 3, GTP-Na 0.3 and HEPES-CsOH 10 pH 7.2. A few experiments were also performed either with internal solution 2, containing (in mM): Cs methanesulphonate 130, CsCl 15, MgCl₂ 1, EGTA 10, CaCl₂ 1, ATP-Mg 3, GTP-Na 0.3 and HEPES-CsOH 10 (pH 7.2) or with internal solution 3, identical to solution 1 except that it did not contain EGTA and was supplemented with 10 μM ZnCl₂.

Drugs

Stock solutions of ICS-205,930 (3-tropanyl-indole-3-carboxylate hydrochloride, RBI) were prepared before use at 1 mM in distilled water. Stock solutions of propofol were prepared in DMSO, at either 60 or 180 mM. In all the experiments where its modulatory effects were tested, all the solutions applied by the fast perfusion system contained the same amount of DMSO (1/1000). A stock solution of ZnCl₂ (Sigma) was prepared at 50 mM in dilute HCl (2 mM) and kept at −20°C. Chlormethiazole edisylate was a gift from Astra and was diluted before use at 10 mM in water. A given stock solution of glycine (Sigma), prepared at 10 mM in distilled water and kept at −20°C, was used for several weeks.

Perfusion system

The culture dish was continuously perfused with the external solution. In addition, a fast perfusion system was used for rapid application of glycine and modulators. All solutions applied via this system contained tetrodotoxin (TTX) 0.2 μM. As previously explained, the fast perfusion system was made of two glass barrels and lateral movements of the two barrels were controlled by a computer-driven motor in order to apply the solution of the desired barrel to the cell; only one of the barrels contained glycine and the cell was continuously perfused with one of the solutions that could be applied by this system (either the control solution containing TTX or a solution also containing a modulator and/or glycine, see Chesnoy-Marchais, 1996 for more details concerning this perfusion system). All tubings were in Teflon, and traces of chemicals previously used were eliminated by extensive washing.

At the low concentrations of glycine used in most experiments (below 20 μM), successive responses recorded in the same solution were usually quite stable when tested every 40 s. Repetitive measurements of successive glycine responses at a fixed interval have been systematically performed in order to separate true modulatory effects from possible slow spontaneous changes in the response. When a high concentration of glycine was used (100–200 μM), the interval between successive tests was longer (100 s) in order to allow recovery from desensitization.

When a modulator of glycine responses was applied ‘with preincubation', it was applied continuously between and during the successive glycine applications, that is in both barrels of the fast perfusion system; when applied ‘without preincubation', it was present only in the glycine-containing barrel. In several experiments, ICS 205,930 was applied only during glycine applications, since its potentiating effect was previously shown to be identical whether it was applied without or with preincubation (see Figure 5b in Chesnoy-Marchais, 1996). ICS 205,930 has been preferred to the two other 5-HT₃ antagonists tested previously (LY 278,584 and MDL 72222; Chesnoy-Marchais, 1996) because it is active at much lower concentrations than LY 278,584 and because it is water-soluble and has faster effects than MDL 72222.

Figure 5.

High concentrations of chlormethiazole, increasing the leak current and potentiating GABA responses, do not affect glycine responses and do not prevent the potentiating effect of ICS 205,930. (a), (b) and (c) from a same experiment. (a) Amplitude of successive peak response to 10 μM glycine recorded at −20 mV in an experiment during which 50 μM chlormethiazole was applied with preincubation. (b) Mean of four glycine responses recorded with or without chlormethiazole (two before, two after the chlormethiazole application). (c) Current recorded in the absence of glycine at −20 mV; note the small reversible increase in outward current induced by chlormethiazole (that was also visible in (b) before or after the glycine application). (d) and (e) Experiment showing that the same chlormethiazole concentration potentiated the responses to 2.5 μM GABA (same culture, −30 mV). (d) Amplitude of successive peak GABA responses. (e) Mean of two GABA responses recorded without or with chlormethiazole. (f) Amplitude of successive peak responses to 10 μM glycine recorded at −30 mV in another cell continuously perfused with 5 μM bicuculline; neither of the two applications of 100 μM chlormethiazole (with preincubation) affected the response, whereas 0.2 μM ICS 205,930 applied with preincubation during the second test of chlormethiazole did potentiate the response, without affecting the current recorded in the absence of glycine (not illustrated).

Recording

Patch-clamp micropipettes were made from hard glass (Kimax 51); the shank of each pipette was covered with Sylgard and the tip was fire-polished. The resistance of these electrodes filled with the usual internal solution was between 5 and 10 MΩ. The cells were voltage-clamped by an EPC7 List amplifier, controlled by a TANDON 38620 computer, via a Cambridge Electronic Design (CED) 1401 interface, using CED patch- and voltage-clamp software. The current monitor output of the amplifier was filtered at 0.3 kHz before being sampled on-line at 0.6 kHz. The bath was connected to the ground via an agar bridge. Membrane potentials were corrected for the junction potential of 10 mV amplitude that was measured between the recording pipette and the usual external solution.

The series resistance (Rs) was systematically measured several times during each experiment. Particular care was taken to eliminate experiments in which Rs changed suddenly. Rs was between 10 and 20 MΩ. These values are high enough to introduce a difference of a few mV between the voltage applied to the electrode and that actually applied to the inside of the cell (error of maximum 10 mV for the largest responses). Usually, no correction has been used to compensate for these errors. However, the current modulations observed occurred without any simultaneous change in Rs and thus cannot result from changes in the applied voltage.

The zero indicated on current traces is the absolute zero current level.

All values are expressed as mean±s.d. (number of observations).

Results

Additivity of the potentiating effects of ICS 205,930 and Zn²⁺ on glycine responses

The effect of ICS 205,930 (0.2 μM) on the response to a low concentration of glycine (10 μM) was first tested during continuous application of Zn²⁺ at concentrations (5 or 10 μM) known to induce the maximal potentiation that can be evoked by this ion (Laube et al., 1995). Such an experiment is illustrated in Figure 1a and b and shows that coapplication of ICS 205,930 with glycine potentiated the response in the continuous presence of 5 μM Zn²⁺; in the absence of ICS 205,930, increasing the concentration of Zn²⁺ during glycine applications from 5 to 10 μM did not affect the response, which confirmed that 5 μM was saturating for the rapid potentiating effect of these ions. The persistence of the potentiating effect of ICS 205,930 in the presence of Zn²⁺ was confirmed in experiments such as that illustrated in Figure 1c, where the effect of ICS 205,930 was successively tested on the same cell in the absence of Zn²⁺ and in the presence of Zn²⁺. In three such experiments, using either 5 or 10 μM Zn²⁺, the percentage of potentiation induced by ICS 205,930 was 98±29% and 40±9%, in the absence and presence of Zn²⁺, respectively.

Figure 1.

Potentiating effect of ICS 205,930 in the continuous presence of Zn²⁺ and inhibitory effect of Zn²⁺ preincubation on glycine responses. (a) Amplitude of successive peak responses to 10 μM glycine recorded at −30 mV, without or with 0.2 μM ICS 205,390, Zn²⁺ being initially applied at 5 μM both between and during glycine applications; then, the concentration of Zn²⁺ applied together with glycine was increased to 10 μM (which did not affect the response). (b) Mean of two glycine responses recorded in the continuous presence of 5 μM Zn²⁺, without or with ICS 205,930 (same cell as in (a)). (c) Amplitude of successive peak responses to 10 μM glycine recorded at −30 mV in another cell, first in the absence of Zn²⁺, then when Zn²⁺ was applied at 10 μM, either without or with preincubation; when indicated, ICS 205,930 was added to the corresponding glycine-containing solution. (d) Amplitude of successive peak responses of another neurone to simultaneous applications of 10 μM glycine and 10 μM Zn²⁺ at −30 mV using internal solution 3 (containing no EGTA and supplemented with 10 μM Zn²⁺); 10 μM Zn²⁺ was applied in the external solution without or with preincubation, as indicated. (e) Mean of two current traces recorded in the same neurone without or with Zn²⁺ preincubation. Note that under the conditions used here, Zn²⁺ preincubation did not affect the current recorded in the absence of glycine (as shown in (e) by the superposition of the holding currents before glycine applications).

The experiment of Figure 1c also revealed an unexpected result: an inhibitory effect of Zn²⁺ preincubation. In six such experiments (performed at −30 mV with the usual internal solution 1) the response to the simultaneous application of glycine (10 μM) and Zn²⁺ (10 μM) was reduced by 45±8% when Zn²⁺ (10 μM) was also added between the glycine pulses. This inhibitory effect of Zn²⁺ preincubation was slowly reversible. It was also observed below the reversal potential of glycine responses, at a holding potential of −90 mV (reduction by 29±5% (4), using again 10 μM glycine, 10 μM Zn²⁺ and internal solution 1). In order to test the hypothesis according to which this inhibitory effect could result from an entry of Zn²⁺ into the cell, two types of experiments were performed. First, the concentration of EGTA in the internal solution was increased up to 10 mM (internal solution 2) in order to better buffer Zn²⁺ possibly entering into the cell. The inhibitory effect of preincubation with Zn²⁺ (10 μM) persisted (not illustrated), the degree of inhibition being of 36±8% (4) after 9–15 min of cell dialysis with 10 mM EGTA. A second series of experiments was performed using internal solution 3 that contained no buffer of divalent ions and was already supplemented with 10 μM Zn²⁺. Again, the inhibitory effect of Zn²⁺ preincubation persisted, as illustrated by Figures 1de: reduction of 44±6% (4) by Zn²⁺ preincubation after 5–9 min of cell dialysis. These results indicate that the inhibitory effect of Zn²⁺ preincubation is not due to an entry of Zn²⁺ into the cells. Preincubation with 10 μM Zn²⁺ also reduced the peak responses to the simultaneous application of a high concentration of glycine (100 or 200 μM) and 10 μM Zn²⁺ (responses that were close to saturating, not shown); however the degree of inhibition induced in these experiments, 16±6% (3) for responses to 200 μM glycine and Zn²⁺, 22±5% (6) for responses to 100 μM glycine and Zn²⁺, was significantly smaller (Student's t-test P<0.0005) than in the experiments using only 10 μM glycine.

In order to avoid the inhibitory effect of Zn²⁺ preincubations, the additional experiments designed to look for possible interferences between the potentiating effects of ICS 205,930 and Zn²⁺ were performed by applying both modulators without preincubation.

In a first series of experiments (performed at −30 mV), the effect of ICS 205,930 was tested on each cell both on the control response to glycine (10 μM) and when the same concentration of glycine was applied with Zn²⁺ (5 or 10 μM). The results from six cells (not illustrated) showed that ICS 205,930 potentiated the control response by 100±31% and potentiated the response to the simultaneous application of glycine and Zn²⁺ by 41±14%. Thus, even though the potentiation was clear, it was smaller than that observed in the same cells in the absence of Zn²⁺. The ratio of the percentage of potentiation induced by ICS 205,930 in the presence of Zn²⁺ over the percentage measured in the same cell in the absence of Zn²⁺ was 0.41±0.05 (6), which is significantly different from 1 (Student's t-test, P<0.000005).

Both the potentiating effect of Zn²⁺ (Laube et al., 1995) and the potentiating effect of ICS 205,930 (Chesnoy-Marchais, 1996) are known to result from decreases in the EC₅₀ for glycine without changes in the maximal response. In agreement with these published observations, it was confirmed in four experiments that the responses to a high concentration of glycine (200 μM) were very little affected by the simultaneous application of ICS 205,930 (0.2 μM) and Zn²⁺ (10 μM): peak responses were enhanced by only 8±4%. In order to compare the potentiating effect of ICS 205,930 on responses of comparable amplitude without or with Zn²⁺, a second series of experiments was performed in which the concentration of glycine was lower when applied with Zn²⁺ than in the absence of Zn²⁺. The reduction was such that the response in the presence of Zn²⁺ had a size similar to that induced in the same cell by application of 10 μM glycine alone (Figure 2). In four such experiments, in which the responses obtained by applying 5 μM glycine together with 10 μM Zn²⁺ were very similar to the control responses obtained by applying 10 μM glycine alone (90±14% of these control responses), ICS 205,930 potentiated the responses to 10 μM glycine by 57±11% whereas it potentiated the responses to 5 μM glycine plus 10 μM Zn²⁺ by 72±16%. Thus, the ability of ICS 205,930 to potentiate glycine responses of similar amplitudes was not significantly different in control and in the presence of a saturating concentration of Zn²⁺ (paired Student's t-test P>0.1).

Figure 2.

Comparison of the effects of ICS 205,930 on responses of similar amplitudes recorded in the absence or presence of Zn²⁺. (a) Amplitude of successive peak responses to glycine recorded at −30 mV without or with 0.2 μM ICS 205,930, glycine being first applied at 10 μM without Zn²⁺, then at 5 μM with 10 μM Zn²⁺. (b) Mean of two responses to 10 μM glycine recorded in the absence of Zn²⁺, without and with ICS 205,930. (c) Mean of two responses to 5 μM glycine recorded in the presence of 10 μM Zn²⁺, without ICS 205,930 (before and after its application) and with ICS 205,930. (a), (b) and (c) from the same cell.

In four experiments, the potentiating effects of Zn²⁺ (1 or 10 μM) on responses to glycine were tested on the same cell at two membrane potentials (−90 mV and either −30 or −10 mV) and no voltage-dependence was detected (not illustrated). In contrast, the potentiating effect of ICS 205,930 was slightly but reproducibly more pronounced at −90 mV than at −30 mV: in 14 experiments (see also Chesnoy-Marchais, 1996) the response to glycine (15 or 20 μM) was alternately recorded on each cell at −90 and −30 mV, before, during and after a given application of ICS 205,930 (at either 0.1, 0.2 or 1 μM according to the experiment). The ratio of the response recorded in the presence of ICS 205,930 over the control response was 1.14±0.07 (14) times higher at −90 mV than at −30 mV (which is significantly different from 1, paired Student's t-test, P<0.000005).

Potentiating effect of ICS 205,930 in the presence of 200 mM ethanol

As already observed by other authors (see Introduction) and shown by Figures 3a and b, ethanol can potentiate responses to low concentrations of glycine. The potentiation of glycine responses has been reported to increase with the concentration of ethanol (Aguayo & Pancetti, 1994; Mascia et al., 1996a,1996b) and it does not seem possible to saturate this effect. Despite this limitation, a few experiments were performed in order to determine whether or not the presence of a high concentration of ethanol (200 mM) would impair the potentiating effect of ICS 205,930 on glycine responses. As illustrated by Figures 3c and d, in the continuous presence of 200 mM ethanol, the responses to 10 μM glycine were still potentiated by 0.2 μM ICS 205,903, by 78±28% (5). The mean potentiation observed in the absence of ethanol from a large number of other cells was of 91±34% (27) (see Discussion).

Figure 3.

High concentrations of ethanol potentiate glycine responses and do not prevent the potentiating effect of ICS 205,930. (a) and (b) Potentiating effect of 200 mM ethanol (applied continuously) on the response to 10 μM glycine recorded at −10 mV. (a) Amplitude of successive peak responses. (b) Mean of four responses recorded with or without ethanol (two before, two after ethanol application). (c) and (d) Persistence of the potentiating effect of 0.2 μM ICS 205,930 on the response to 10 μM glycine recorded in another cell at −30 mV in the continuous presence of 200 mM ethanol. (c) Amplitude of successive peak responses. (d) Mean of two responses recorded with or without ICS 205,930 (one before, one after ICS 205,930 application).

Potentiating effect of ICS 205,930 in the presence of high concentrations of propofol

The intravenous general anaesthetic propofol is known to induce several effects in neurones. Propofol potentiates the chloride responses to subthreshold concentrations of GABA (see Sanna et al., 1995 and references therein) and is one of the few agents known as able to potentiate glycine responses (see Hales & Lambert, 1991 for spinal neurones; see Mascia et al., 1996a and Pistis et al., 1997 for cloned glycine receptors). In chromaffin cells (Hales & Lambert, 1991), hippocampal neurones (Hara et al., 1993; Orser et al., 1994) and hypothalamic neurones (Adodra & Hales, 1995), propofol has also been shown to activate a chloride conductance in the absence of GABA and glycine, and since this effect could be partly blocked by addition of bicuculline, it was interpreted as resulting, at least partly, from the direct activation of GABA_A receptors.

Initial experiments performed in ventral spinal cord neurones confirmed the ability of propofol to potentiate glycine responses and to activate a chloride conductance. The lowest concentrations of propofol (between 0.5 and 1 μM) inducing a detectable potentiation of responses to 10 μM glycine already induced an increase in chloride conductance in the absence of glycine (not shown). Thus, in order to estimate the potentiating effect of higher concentrations of propofol on glycine responses, propofol had to be applied both in the absence and presence of glycine. Initially, glycine responses were measured at a constant holding potential, different from the chloride equilibrium potential. It appeared that, under such conditions, the reversal potential of glycine responses could be appreciably modified by chloride fluxes through the large conductance activated by propofol in the absence of glycine, which prevented a reliable estimation of the potentiating effect of propofol on the glycine-induced conductance. Therefore, the protocol was modified in order to minimize the propofol-activated chloride fluxes and to better control the reversal potential. In order to reduce the conductance directly activated by propofol, bicuculline (10 μM) was added in all the solutions applied by the fast perfusion; this was not sufficient however to induce a complete blockade (see Discussion). Thus, in addition, the holding potential was set close to the reversal potential of the propofol-induced current, and the conductance in the absence of glycine and during the glycine responses was measured by using either voltage-jumps (see Figure 4a) or brief voltage ramps (see Figures 4def) which allowed direct measurement of reversal potentials. Using these methods, it was possible to reliably measure glycine responses in the continuous presence of high concentrations of propofol (30–90 μM) and to study the effects of the addition of ICS 205,930 (0.2 μM) after a maximal potentiation by propofol. Note that in these experiments, ICS 205,930 was applied both between and during glycine applications, in order to look for possible modifications of the conductance recorded with propofol in the absence of glycine. In several experiments, the addition of ICS 205,930 in the presence of propofol induced a very small conductance increase in the absence of glycine (not shown), which was taken into account when measuring glycine responses in the simultaneous presence of propofol and ICS 205,930.

Figure 4.

High concentrations of propofol, which increase the leak currents and glycine responses, do not prevent the potentiating effect of ICS 205,930. (a) Current traces recorded in the absence of propofol and in the continuous presence of either 30 or 60 μM propofol, during voltage jumps from −52 to −92 mV (regularly applied for 4 s every 40 s), 10 μM glycine being systematically applied at −92 mV. The conductance of the neurone in the absence of glycine is shown by the instantaneous current change induced by the voltage jump. (b) Current traces recorded at −95 mV in the absence or continuous presence of 0.2 μM ICS 205,930 in another cell continuously perfused with 90 μM propofol, 5 μM glycine being regularly applied as in (a) (that is every 40 s, 900 ms after the beginning of each voltage jump, from −55 to −95 mV for this cell). (c) and (d) Persistence of the potentiating effect of 0.2 μM ICS 205,930 in the presence of a saturating concentration of propofol. (c) Amplitude of successive responses to 10 μM glycine measured close to −90 mV (after subtraction of the leak current) in a cell held at −50 mV to which voltage ramps from −90 to −30 mV were alternately applied in the absence and presence of glycine; propofol was continuously applied at either 30 μM or 60 μM as indicated; ICS 205,930 was also continuously applied during the arrow. (d) Records corresponding to the points indicated in (c), obtained during the voltage ramps, after subtraction of the corresponding traces recorded in the absence of glycine; the application of glycine began 1.3 s before the ramp and, as expected, the response was not detectable at the holding potential (that was close to the chloride equilibrium potential). (e) and (f) Leak current traces recorded in the same cell in the absence of glycine, (e) at the beginning of the experiment either without or with 30 μM propofol, (f) in the presence of either 30 or 60 μM propofol (two traces superimposed). In these experiments, all the solutions applied by the fast perfusion contained TTX, a final dilution of 1/1000 DMSO and 10 μM bicuculline.

Figure 4a illustrates the effects of 30 and 60 μM propofol on the response to glycine and on the conductance measured in the absence of glycine (see figure legend). The glycine response was potentiated by 362±43% by 30 μM propofol in three similar experiments and increasing the propofol concentration from 30 to 60 μM did not have a significant additional effect either on glycine responses (see also Figure 4c) or on the conductance measured in the absence of glycine (see also Figure 4f). The experiment illustrated by Figure 4b shows that ICS 205,930 clearly potentiated the response to a low concentration of glycine in the continuous presence of a very high concentration (90 μM) of propofol (for clarity only the currents recorded at the test potential are displayed).

The results of an experiment using voltage ramps, alternately applied in the absence and presence of glycine in different solutions, are illustrated by Figures 4cdef. Glycine responses were measured during the ramps (after subtraction of the corresponding traces obtained in the absence of glycine) first in the absence of any modulator, then in the presence of 30 μM propofol (trace 1 in Figure 4d), then in the presence of both 30 μM propofol and 0.2 μM ICS 205,930, which further potentiated the response without changing the reversal potential (trace 2 in Figure 4d); finally, after wash of ICS 205,903, the propofol concentration was increased to 60 μM, and this did not affect the response. Current traces recorded during the ramps applied in the absence of glycine are also illustrated (Figures 4e and f); addition of 30 μM propofol induced a clear conductance increase (despite the presence of bicuculline, Figure 4e), and increasing the propofol concentration to 60 μM did not further affect the conductance in the absence of glycine (Figure 4f, two traces superimposed).

The results (confirmed in several other experiments using 5 μM glycine and a propofol concentration of either 30 μM (three cells) or 90 μM (four cells including that of Figure 4b) clearly show that ICS 205,930 remains capable of potentiating glycine responses in the presence of concentrations of propofol which were saturating both for the potentiation of glycine responses and for the direct conductance activation by this anaesthetic.

Chlormethiazole does not affect glycine responses

In order to study possible interferences between the effects of ICS 205,930 and chlormethiazole, experiments were first performed to confirm the ability of this anticonvulsant to potentiate glycine responses. Surprisingly (see Introduction), 50 μM chlormethiazole (applied either continuously or only with glycine) did not affect the responses to glycine (10 μM) in six different cells (Figures 5a and b). However, the leak current recorded in these cells in the absence of glycine was systematically increased by chlormethiazole, as indicated by the slight reversible increase in outward current observed above E_Cl (Figures 5b and c). Chlormethiazole potentiated the responses of the same neurones to a low concentration (2.5 μM) of GABA (Figures 5d and e; potentiation of 124±6% in three similar experiments). The inability of chlormethiazole to affect glycine responses (on either side of E_Cl) was confirmed by using a higher concentration (100 μM) in seven additional cells, continuously perfused with bicuculline (5 μM) in order to block GABA responses and to prevent the leak current increase usually induced by chlormethiazole (Figure 5f). Figure 5f also shows that, as expected, the continuous presence of chlormethiazole and bicuculline did not affect the ability of ICS 205,930 to potentiate the responses of these cells to glycine, a result confirmed in four similar experiments.

Discussion

In the first part of the present paper, it has been shown that the potentiating effect of ICS 205,930 on responses to low concentrations of glycine persists in the presence of high concentrations of Zn²⁺ (5 or 10 μM), whatever the method used for Zn²⁺ applications (with or without preincubation). From the study of Laube et al. (1995), these concentrations were known to be saturating for their potentiating effect, either in spinal neurones (Zn²⁺ ions being then applied without preincubation) or in Xenopus oocytes expressing α1, α1β or α2 receptors (Zn²⁺ being then applied with a preincubation of 15 s). This has been confirmed by the present results (Figure 1a). Even though the potentiating effect of ICS 205,930 was shown to persist in the presence of Zn²⁺, the percentage of increase of the response to a given concentration of glycine was lower in the presence than in the absence of Zn²⁺. This apparent reduction of the potentiating effect of ICS 205,930 results from the fact that both ICS 205,930 and Zn²⁺ potentiate glycine responses by increasing the apparent affinity of the receptors for glycine. Indeed, the percentage of increase induced by ICS 205,930 became very similar in the absence and presence of Zn²⁺ when the controls used were responses of identical amplitudes, either to glycine alone (10 μM) or to an equipotent mixture of a lower concentration of glycine plus 10 μM Zn²⁺ (Figure 2). These results show that the sites of action of ICS 205,930 and Zn²⁺ are different. This conclusion is also in agreement with the observation that potentiating effect of ICS 205,930 shows a slight (but reproducible) voltage-sensitivity, which is lacking in the potentiating effect of Zn²⁺ (despite the presence of a larger positive charge), as previously noticed in other neurones (Kumamoto & Murata, 1996; Trombley & Shepherd, 1996) and as expected if the site involved is extracellular (Laube et al., 1995).

During the comparison of the potentiating effects of ICS 205,930 and Zn²⁺, the existence of an inhibitory effect of Zn²⁺ on glycine responses (Bloomenthal et al., 1994; Laube et al., 1995) was confirmed. In previous investigations, when these ions were applied on spinal neurones without preincubation, their inhibitory effect had been detected only with very high concentrations (above 100 μM: see Figure 1b in Laube et al., 1995). The present study has shown that preincubation with Zn²⁺ already reveals a pronounced inhibitory effect of these ions at 10 μM (see for example Figure 1c). The strong effect of preincubation with Zn²⁺ ions might have resulted from their slow entry into the neurones, since various routes of entry for these ions have been described in neurones (see for example Sensi et al., 1997). Several experiments, using different internal solutions (containing either a buffer for divalent ions or a high concentration of Zn²⁺, see Figures 1d and e) have shown that the inhibitory effect of Zn²⁺ preincubation does not result from an entry of Zn²⁺. It probably results from the existence of a second binding site for these ions that is only slowly accessible: this site would not be occupied by Zn²⁺ during the short applications of low concentrations of glycine and Zn²⁺ used in the present study (of a few seconds duration), but would be occupied when Zn²⁺ is applied between these applications (here, during at least 30 s). Zn²⁺ preincubation also showed a slight inhibitory effect on the peak response to high concentrations of glycine applied with Zn²⁺; however, since this effect was smaller than the inhibitory effect observed when using only 10 μM glycine and Zn²⁺, some competition between glycine and Zn²⁺ could account, at least partly, for the inhibitory effect of Zn²⁺ preincubation. Further studies will be necessary to better characterize the mechanism responsible for this inhibitory effect. The Zn²⁺-sensitivity of responses to glycine recorded in some neuronal preparations, such as septal cholinergic neurones (Kumamoto & Murata, 1996) or olfactory bulb neurones (Trombley & Shepherd, 1996), in particular the absence of inhibitory effects of Zn²⁺ below 100 μM in these studies, has been used to support the idea that the glycine receptors expressed in the corresponding neurones are different from those previously studied in spinal neurones and from α1- or α2-containing receptors expressed in Xenopus oocytes. These data, obtained without Zn²⁺ preincubation, should now be reconsidered taking into account the strong influence of the protocol used (presence or absence of Zn²⁺ preincubation) on the Zn²⁺-sensitivity of glycine responses (as revealed by the present study).

Various compounds which potentiate glycine responses also act, like Zn²⁺ and ICS 205,930, by shifting the dose-response curve for glycine to the left. This has been shown for some alcohols, including ethanol (Mascia et al., 1996a,1996b; Ye et al., 1998) and for an intravenous general anaesthetic, propofol (Pistis et al., 1997). Note that the behavioural action of propofol has been shown to involve its effect on glycine receptors: for example Mascia et al. (1997) showed that the difference in sleep time induced by administration of propofol between two lines of mice, the ‘long-sleep' mice and the ‘short-sleep' mice, is correlated with a genetic difference in propofol-sensitivity of their glycine receptors (without any difference in propofol-sensitivity of their GABA_A receptors). As discussed below, the results obtained in the present study indicate that ICS 205,930 acts on glycine receptors at a site different from the site(s) of action of either ethanol or propofol.

In the presence of 200 mM ethanol, ICS 205,930 remained able to increase the response to 10 μM glycine (Figure 3). It can be argued that the concentration of ethanol used was not saturating for the potentiation of glycine responses and indeed, one cannot exclude that higher concentrations of ethanol would interfere with the site of action of ICS 205,930. It was shown above, in the case of Zn²⁺, that the degree of potentiation induced by ICS 205,930 (for a given concentration of glycine) can be lowered by another potentiator that also shifts to the left the dose-response curve for glycine. One might have expected such a result in the case of ethanol. However, the experiments concerning the effect of ICS 205,930 in the presence of ethanol did not include tests of the effect of ICS 205,930 on the same cells in the absence of ethanol; thus, because of the variability of the effect of ICS 205,930 among different cells (already reported in Chesnoy-Marchais, 1996), these experiments did not exclude the possible reduction by ethanol of the degree of potentiation induced by ICS 205,930. These experiments merely showed the persistence of a significant potentiating effect of ICS 205,930 in the presence of ethanol. The strong variability of the effect of ICS 205,930 among different cells is unlikely to result from variations between individual dose-response curves for glycine (in view of previous results obtained from these neurones and in view of the absence of pronounced difference in the EC₅₀ for glycine between different types of recombinant glycine receptors; see for example Pribilla et al., 1992); preliminary results obtained in Xenopus oocytes expressing recombinant glycine receptors (given by Pr. H. Betz) indicate that this variability is more likely to result from a different ICS 205,930-sensitivity of the different types of glycine receptors expressed in cultured neurons (unpublished results obtained in collaboration with S. Supplisson).

In the case of propofol, it was shown that even at very high concentrations, which were saturating for the potentiation of glycine responses, this anaesthetic did not prevent the potentiating effect of ICS 205,930 (Figure 4). Thus, propofol and ICS 205,930 also seem to potentiate glycine responses by different mechanisms. The results obtained in the present study have confirmed in native neurones the initial observation by Hales & Lambert (1991) of the potentiating effect of propofol on glycine responses. In agreement with the recent demonstration (Pistis et al., 1997) that propofol induces a shift to the left of the dose-response curve for glycine, the potentiations induced by propofol in the present study, using 5–10 μM glycine, were much larger than those observed in the initial study (Hales & Lambert, 1991) using 100 μM glycine. It was also confirmed that whereas ICS 205,930 does not affect the basal conductance in the absence of glycine (Chesnoy-Marchais, 1996), propofol markedly increases this conductance, even in the presence of bicuculline, known to block most responses involving GABA_A receptors. The bicuculline-resistant conductance increase induced by propofol could result from the activation of bicuculline-resistant GABA_A receptors in the absence of GABA, such as some homomeric β receptors (Sanna et al., 1995; Krishek et al., 1996; see also Davies et al., 1997) and/or from the activation of some glycine receptors in the absence of glycine (Pistis et al., 1997). This last hypothesis could explain why in several experiments performed in the presence of propofol, ICS 205,930 induced a small conductance increase in the absence of glycine: this compound might have potentiated the basal activity of glycine receptors induced by propofol in the absence of glycine. In agreement with the results obtained in the present study, even in cases where bicuculline was reported to ‘suppress' the propofol-induced current, the blockade induced by 10 or 20 μM bicuculline was not complete (see Figure 3 in Hara et al., 1993, where the propofol concentration was only 20 μM). As explained in the Results section, the large conductance increase induced by propofol in the absence of neurotransmitter made the experiments concerning propofol particularly difficult. Some of the difficulties encountered, such as the progressive change of the baseline current and the imperfect stability of the responses to glycine applications in the presence of propofol, could not be avoided. These changes may result partly from desensitization-like effects, as shown in hippocampal neurones for the baseline current induced by propofol (Hara et al., 1993) and as suggested by the fact that GABA_A responses can be desensitized by preincubations with high concentrations of propofol (Orser et al., 1994). Thus, with the methods used in the present study, it would not have been reasonable to try to perform a more quantitative study of possible interferences between the potentiating effects of propofol and ICS 205,930. However, the data obtained indicate that these two drugs act by different mechanisms.

Contrary to some previous observations (Gent & Wacey, 1983; Hales & Lambert, 1992), chlormethiazole could not potentiate responses to glycine in the present study. Several possibilities could explain this apparent discrepancy. As already pointed out (Hales & Lambert, 1992), the potentiating effect of chlormethiazole previously reported might result from an effect on glycine uptake mechanisms rather than from a direct effect on glycine receptors. It is also possible that chlormethiazole has a direct potentiating effect on some subtypes of glycine receptors that would be different from those which are predominant in the preparation used in the present study. Such a subunit selectivity could explain the previously noticed variability in the degree of potentiation induced by chlormethiazole on the electrical activity of brainstem neurones (Gent & Wacey, 1983).

In conclusion, it has been shown that ICS 205,930 is a novel type of potentiator of glycine responses. Not only is it efficient at remarkably low concentrations without affecting the resting conductance, but it also uses a site that is different from the site(s) responsible for the potentiating effects of several other modulators.

Abbreviations

ATP

adenosine triphosphate

DMSO

dimethyl sulphoxide

EGTA

ethylenediaminetetraacetic acid

GABA

γ-amino-n-butyric acid

GTP

guanosine triphosphate

HEPES

N-[2-hydroxyethyl]piperazine-N′-[2-ethane-sulphonic acid]

TTX

tetrodotoxin